Arabic and
contact-induced
change
Edited by
Christopher Lucas
Stefano Manfredi
Contact and Multilingualism 1
language
science
press
Contact and Multilingualism
Editors: Isabelle Léglise (CNRS SeDyL), Stefano Manfredi (CNRS SeDyL)
In this series:
1. Lucas, Christopher & Stefano Manfredi (eds.). Arabic and contact-induced change.
Arabic and
contact-induced
change
Edited by
Christopher Lucas
Stefano Manfredi
language
science
press
Lucas, Christopher & Stefano Manfredi (eds.). 2020. Arabic and contact-induced
change (Contact and Multilingualism 1). Berlin: Language Science Press.
This title can be downloaded at:
http://langsci-press.org/catalog/book/235
© 2020, the authors
Published under the Creative Commons Attribution 4.0 Licence (CC BY 4.0):
http://creativecommons.org/licenses/by/4.0/
ISBN: 978-3-96110-251-8 (Digital)
978-3-96110-252-5 (Hardcover)
DOI:10.5281/zenodo.3744565
Source code available from www.github.com/langsci/235
Collaborative reading: paperhive.org/documents/remote?type=langsci&id=235
Cover and concept of design: Ulrike Harbort
Typesetting: Christopher Lucas, Felix Kopecky
Proofreading: Alys Boote Cooper, Amir Ghorbanpour, Amr El-Zawawy, Anna
Shea, Andreas Hölzl, Carla Bombi, Chams Bernard, Edalat Shekari, Ikmi Nur
Oktavianti, Jeroen van de Weijer, Jean Nitzke, Sean Stalley, Tom Bossuyt,
Waldfried Premper, Varun deCastro-Arrazola & Yvonne Treis
Fonts: Libertinus, Arimo, DejaVu Sans Mono, SIL Scheherazade
Typesetting software: XƎLATEX
Language Science Press
Xhain
Grünberger Str. 16
10243 Berlin, Germany
langsci-press.org
Storage and cataloguing done by FU Berlin
Contents
1
Introduction
Christopher Lucas & Stefano Manfredi
1
I
Contact-induced change in varieties of Arabic
2
Pre-Islamic Arabic
Ahmad Al-Jallad
37
Classical and Modern Standard Arabic
Marijn van Putten
57
4 Arabic in Iraq, Syria, and southern Turkey
Stephan Procházka
83
3
5
6
7
Khuzestan Arabic
Bettina Leitner
115
Anatolian Arabic
Faruk Akkuş
135
Cypriot Maronite Arabic
Mary Ann Walter
159
8 Nigerian Arabic
Jonathan Owens
175
9 Maghrebi Arabic
Adam Benkato
197
10 Moroccan Arabic
Jeffrey Heath
213
Contents
11 Andalusi Arabic
Ángeles Vicente
225
12 Ḥassāniyya Arabic
Catherine Taine-Cheikh
245
13 Maltese
Christopher Lucas & Slavomír Čéplö
265
14 Arabic in the diaspora
Luca D’Anna
303
15 Arabic pidgins and creoles
Andrei Avram
321
II Language change through contact with Arabic
16 Modern South Arabian languages
Simone Bettega & Fabio Gasparini
351
17 Neo-Aramaic
Eleanor Coghill
371
18 Berber
Lameen Souag
403
19 Beja
Martine Vanhove
419
20 Iranian languages
Dénes Gazsi
441
21 Kurdish
Ergin Öpengin
459
22 Northern Domari
Bruno Herin
489
23 Jerusalem Domari
Yaron Matras
511
ii
Contents
24 Mediterranean Lingua Franca
Joanna Nolan
533
III Domains of contact-induced change across Arabic varieties
25 New-dialect formation: The Amman dialect
Enam Al-Wer
551
26 Dialect contact and phonological change
William M. Cotter
567
27 Contact and variation in Arabic intonation
Sam Hellmuth
583
28 Contact-induced grammaticalization between Arabic dialects
Thomas Leddy-Cecere
603
29 Contact and calquing
Stefano Manfredi
625
30 Contact and the expression of negation
Christopher Lucas
643
Index
669
iii
Chapter 1
Introduction
Christopher Lucas
SOAS University of London
Stefano Manfredi
CNRS, SeDyL
This introductory chapter gives an overview of the aims, scope, and approach of
the volume, while also providing a thematic bibliography of the most significant
previous literature on Arabic and contact-induced change.
1 Rationale
With its lengthy written history, wide and well-studied dialectal variation, and involvement in numerous heterogeneous contact situations, the Arabic language
has an enormous contribution to make to our understanding of how language
contact can lead to change. Until now, however, most of what is known about the
diverse outcomes of contacts between Arabic and other languages has remained
inaccessible to non-specialists. There are brief summary sketches (Versteegh 2001;
2010; Thomason 2011; Manfredi 2018), as well as a recent collection of articles on a
range of issues connected with Arabic and language contact in general (Manfredi
& Tosco 2018), but no larger synthesis of the kind that is available, for example,
for Amazonian languages (Aikhenvald 2002).
Arabic has thus played little part in work to date on contact-induced change
that is crosslinguistic in scope (though see Matras 2009; Trudgill 2011 for partial
exceptions). By providing the community of general and historical linguists with
the present collaborative synthesis of expertise on Arabic and contact-induced
change, we hope to help rectify this situation. The work consists of twentynine chapters by leading authorities in their fields, and is divided into three
Christopher Lucas & Stefano Manfredi. 2020. Introduction. In Christopher Lucas &
Stefano Manfredi (eds.), Arabic and contact-induced change, 1–33. Berlin: Language
Science Press. DOI:10.5281/zenodo.3744501
Christopher Lucas & Stefano Manfredi
Parts: overviews of contact-induced change in individual Arabic varieties (Part I);
overviews of the outcomes of contact with Arabic in other languages (Part II); and
overviews of various types of changes across Arabic varieties, in which contact
has played a significant role (Part III). Chapters in each of the three Parts follow
the fixed broad outlines detailed below in §5, in order to maximize coherence
and ease of reference. All authors have also been encouraged a) to ensure their
chapters contain a rich set of (uniformly glossed and transcribed) linguistic data,
including original data where appropriate, and b) to provide as much sociohistorical data as possible on the speech communities involved, framed where possible
with reference to Van Coetsem’s (1988; 2000) distinction between changes due
to borrowing (by agents dominant in the recipient language (RL)) and imposition (by agents dominant in the source language (SL); see §4 for further details).
These features are aimed at ensuring that the data presented in the volume can
be productively drawn upon by scholars and students of linguistics who are not
specialists in Arabic linguistics, and especially those working on the mechanisms,
typology, outcomes, and theory of contact-induced change cross-linguistically.
The rest of this introductory chapter is structured as follows. We begin by providing a thematic bibliography of existing work on Arabic and contact-induced
change in §2. The overall scope of the present volume is then detailed in §3. §3.1
locates and classifies the different varieties of what is called “Arabic” according to
Jastrow’s (2002) three geographic zones and Labov’s (2007) concepts of transmission and diffusion in language change, while §3.2–§3.4 provide an overview of
the content of each of the three Parts into which the present volume is divided. In
§4 we give details of Van Coetsem’s (1988; 2000) framework, and in §5 we outline
the common structure and transcription and glossing conventions of the volume.
This introductory chapter then finishes with §6, in which we discuss some of the
challenges to Van Coetsem’s framework posed by the data in this volume, how
these challenges can be addressed, and how the data and analyses collected in
the present work can be built on by others.
2 Previous work
As noted in §1, there is a reasonably large existing literature focusing on specific aspects of Arabic and contact-induced change. For reviews of much of this
literature, readers are referred to the relevant chapters of the present volume.
Here we simply list some key works for ease of reference in the following (noncomprehensive) bibliography, organized by linguistic variety.
2
1 Introduction
) Old Arabic and Middle Aramaic: Retsö (2011), Weninger (2011), Owens (2016).
) Arabic and Neo-Aramaic: Arnold & Behnstedt (1993), Arnold (2007), Coghill
(2010; 2012; 2015), Jastrow (2015).
) Arabic and Hebrew: Blau (1981), Yoda (2013), Horesh (2015).
) Arabic and (Modern) South Arabian languages: Diem (1979), Lonnet (2011),
Zammit (2011), Watson (2018).
) Arabic and Indo-Iranian languages: Tsabolov (1994), Matras (2007), Asbaghi
(2011), Gazsi (2011), van der Wal Anonby (2015), Herin (2018).
) Arabic and Turkish: Procházka (2002; 2011), Haig (2014), Taylan (2017), Akkuş
& Benmamoun (2018).
) Arabic and Berber: Taine-Cheikh (1997; 2018), Brahimi (2000), Corriente (2002),
Lafkioui & Brugnatelli (2008), Kossmann (2009; 2010; 2013), El Aissati (2011),
Lafkioui (2013a), Souag (2013), van Putten & Souag (2015).
) Arabic and (sub-)Saharan languages: Owens (2000a; 2015), Lafkioui (2013b),
Souag (2016).
) Arabic and Latin/Romance languages: Brunot (1949), Benoliel (1977), Corriente
(1978; 1992), Talmoudi (1986), Heath (1989; 2015), Cifoletti (1994), Vicente (2006),
Sayahi (2014).
) Contact influences on Classical and Modern Standard Arabic (MSA): Jeffery
(2007) [1938], Blau (1969), Hebbo (1984).
) Contact influence in Mesopotamian Arabic: Masliyah (1996; 1997), Matras &
Shabibi (2007), El Zarka & Ziagos (2019).
) Contact influence in Central Asian Arabic: Jastrow (2005), Ratcliffe (2005), Ingham (2011).
) Contact influence in Levantine Arabic: Barbot (1961), Neishtadt (2015).
) Contact influence in Cypriot Maronite Arabic: Newton (1964), Tsiapera (1964),
Borg (1997; 2004).
) Contact influence in Maltese: Colin (1957), Aquilina (1958), Krier (1976), Mifsud
(1995), Brincat (2011), Souag (2018).
) Arabic pidgins and creoles: Owens (1985), Miller (1993), Luffin (2014), Avram
(2017), Bizri (2018), Owens (2018).
) Contact between Arabic dialects: Gibson (2002), Miller et al. (2007), Cotter &
Horesh (2015), Leddy-Cecere (2018).
3
Christopher Lucas & Stefano Manfredi
3 Scope
3.1 Where and what is Arabic?
Arabic is one of the most widely spoken languages in the world, and the first
language of around 350 million speakers spread throughout the Middle East and
North Africa. There are twenty-five sovereign states in which Arabic is an official
language. In addition, Arabic is widely spoken as a lingua franca (i.e. vehicular
language) for a range of communicative interactions between different linguistic
communities in Asia and Africa. Following Jastrow (2002; see also Watson 2011;
Manfredi forthcoming), the present-day Arabic-speaking world can be broadly
subdivided into three geographic zones (cf. Figure 1): Zone I covers the regions of
the Arabian Peninsula where Arabic was spoken before the beginning of the Islamic expansion in the seventh century; Zone II includes the Middle Eastern and
North African areas into which Arabic penetrated during the Islamic expansion,
and where it is today spoken as a majority language; and Zone III encompasses
isolated regions where Arabic is spoken today by minority bilingual communities (see also Owens 2000b). Further to this, following successive waves of mass
emigration in recent centuries, Arabic is also spoken as a heritage language by
diasporic communities around the world (Rouchdy 1992; Boumans & de Ruiter
2002; D’Anna, this volume). Against the backdrop of this complex geo-historical
distribution, the question that arises is what unites all the varieties that fall under
the glottonym “Arabic” and, more generally, what should count as Arabic from
a linguistic point of view?
After all, the term “Arabic” encompasses a great deal of internal variety, whose
origins can be traced back to both internally and externally motivated (i.e. contact-induced) changes. One way of understanding these different patterns of language change is through Labov’s (2007) distinction between transmission and
diffusion. If transmission refers to change through an unbroken sequence of
first-language acquisition (Labov 2007: 346), diffusion rather implies the transfer of features across languages via language/dialect contact (Labov 2007: 347).
Change through transmission is said to be regular because it is incremented by
young native speakers, whereas diffusion is thought to be more irregular and
unpredictable because it is typically produced by adult bilingual speakers. Both
mechanisms contribute to long-term language change even though, according
to Labov, transmission is the foremost mechanism by which linguistic diversity
is produced and maintained. In a recent study, Owens (2018) tests the generality of the Labovian distinction between transmission and diffusion against the
complex linguistic and sociohistorical patchwork of Arabic. He concludes that
4
1 Introduction
Figure 1: Approximate distribution of languages and Arabic varieties
discussed in this volume
change through diffusion cannot be said to be more irregular than change via
transmission and that, other than for Arabic-based creoles (see Avram, this volume), there are no clear-cut criteria for distinguishing the two mechanisms of language change. The reason for this is that most of the linguistic varieties that are
commonly referred to under the heading of “Arabic” are the result of a longstanding series of multi-causal changes encompassing both internal drift and convergence, as well as contact-induced change via diffusion. What we do not see, however, in any of the varieties usually referred to as Arabic, are the atypical kinds
of changes produced by the disruption of language transmission as observed in
pidgin and creole languages (but see below). Thus Part I of this volume primarily
(but not exclusively) deals with contact-induced change in spoken varieties of
Arabic that have gone through an unbroken chain of language transmission, the
so-called “Arabic dialects”.
5
Christopher Lucas & Stefano Manfredi
3.2 Overview of Part I: Contact-induced change in varieties of Arabic
The survey chapters in Part I of this volume offer an extensive overview of
contact-induced change in first Eastern (mašriqī) and then Western (maɣribī )
Arabic dialects (to use the terminology of the traditional geographical classification of modern Arabic dialects; cf. Palva 2009; Benkato, this volume). The majority of chapters dealing with types of Eastern Arabic describe varieties spoken
by bilingual minorities affected to different degrees by language shift towards
local dominant languages. For instance, the Arabic-speaking Maronite community of Kormakiti is involved in an asymmetric pattern of bilingualism resulting in a gradual and inexorable language shift towards Cypriot Greek (Walter,
this volume). In contrast, speakers of Nigerian Arabic (Owens, this volume), despite considerable proficiency in Kanuri and/or Hausa, maintain transmission of
their ancestral language to the younger generations. As far as it is possible to
tell, a similar situation holds for the Mesopotamian dialects of Anatolia (Akkuş,
this volume) and Khuzestan (Leitner, this volume), which are in intense contact
with Turkish and Persian respectively (among other languages), but without (yet)
showing signs of definitive language shift. Procházka (this volume), on the other
hand, describes the effects of contact-induced change in a continuum of Eastern Arabic dialects dispersed across Lebanon, Syria, Iraq, and southern Turkey.
In this broader geographical context, Arabic represents the main vernacular language, affected to different degrees by long-term bi- or multilingualism with Aramaic, Kurdish, and Turkish.
As far as Western Arabic dialects are concerned, Benkato (this volume) describes a history of contact-induced change in different Maghrebi dialects from
the beginning of the Arabization of North Africa until the colonial period. Four
further chapters then take a closer look at contact-induced changes in specific
varieties of Western Arabic. Heath (this volume) covers Moroccan, while TaineCheikh (this volume) covers Ḥassāniyya – two majority varieties of Arabic historically affected by contact with Berber and Romance languages. Lucas & Čéplö
(this volume) then provide an overview of contact-induced change in Maltese –
a variety which is no longer usually considered to be a subtype of “Arabic”, but
which, as Lucas & Čéplö show, is nevertheless historically part of the Western
group of Arabic dialects. Indeed, despite the far-reaching lexical and grammatical effects of contact with Italo-Romance and English, Maltese remains largely a
product of transmission in the Labovian sense. We would not therefore classify
it as a contact (i.e. mixed) language (cf. Stolz 2003 and see further below). Lastly,
D’Anna (this volume) offers a linguistic account of different varieties of Arabic
in diasporic settings, with particular focus on the Tunisian community of Mazara
6
1 Introduction
del Vallo in Sicily. Unlike the Western varieties described in the aforementioned
chapters, in this latter context Arabic is involved in an unbalanced contact situation, resulting in moderate language shift towards Sicilian and Italian.
As well as the aforementioned spoken varieties of Arabic, Part I of the volume also includes three chapters analysing the outputs of language contact in
different varieties of written Arabic. First of all, Al-Jallad (this volume) describes
a number of likely instances of contact-induced change in pre-Islamic Arabic
documentary sources (primarily inscriptional), and postulates the existence of
different patterns of bilingualism between Arabic and Akkadian, Aramaic, Old
South Arabian, and Greek (among other languages). Van Putten (this volume)
then focuses on contact influences on the later Classical and MSA, examining
both early influences from Aramaic, Greek, Persian, Ethio-Semitic and Old South
Arabian, as well as later influence from Ottoman Turkish and twentieth-century
journalism in European languages. Since these written varieties of Arabic are
rather artificial constructs, van Putten also examines the influence of the native
Arabic dialects of the authors of texts in Classical Arabic and MSA. The third and
final written Arabic variety analysed in this volume is Andalusi Arabic. Attested
as a form of Middle Arabic (Lentin 2011) between the tenth and seventeenth centuries, Andalusi Arabic displays significant grammatical and lexical input from
both Romance and Berber languages (Vicente, this volume). As evidence for the
Arabic varieties described in these three chapters is exclusively written, they
cannot be treated in the same manner as spoken varieties which emerged in a
context of first language acquisition. They are, however, representative of a longstanding and uninterrupted written tradition that goes back to the pre-Islamic
period, and that has always been in a multi-faceted relationship of mutual influence with different varieties of spoken Arabic. In this sense, despite their rather
artificial nature, written varieties of Arabic may also be considered the product
of language transmission.
In the final chapter of Part I, on the other hand, Avram (this volume) describes
a number of Arabic-based pidgins and creoles, which contrast with modern Arabic dialects (including Maltese) in that they have emerged in contact situations
where the available language repertoires did not constitute an effective tool for
communication (Bakker & Matras 2013: 1). These contact languages are thus the
product of partial or full interruption of language transmission, and for this reason they fall outside the range of what is usually considered Arabic (i.e. they are
not straightforwardly classifiable as genetically related to it; cf. McMahon 2013).
In such contexts, the effects of language diffusion via second language acquisition are obviously more evident. The varieties discussed by Avram include the
so-called Sudanic pidgins and creoles (i.e. Juba Arabic, Kinubi, and Turku), which
7
Christopher Lucas & Stefano Manfredi
emerged in Sudan in the nineteenth century and are today scattered across East
Africa, as well as a number of contact languages that have recently emerged in
the context of labour migration to the Middle East: Gulf Pidgin Arabic, Pidgin
Madame, Romanian Pidgin Arabic, and Jordanian Pidgin Arabic. Despite their
different sociohistorical and ethnolinguistic backgrounds, the contact languages
included in this chapter share many formal features as a result of the strong impact of second language acquisition of Arabic in extreme contact situations.
In sum, Part I of the present volume aims at a comprehensive overview of
contact-induced changes in both spoken and written varieties of Arabic, as well
as in Arabic-based contact languages (but see §3.5).
3.3 Overview of Part II: Language change through contact with
Arabic
Throughout its history, Arabic has not only been subject to contact influence
from other languages, but has also itself induced profound changes in the languages with which it has come into contact (see Versteegh 2001 for a general
overview). The latter topic is the focus of the chapters included in Part II of the
present volume. Let us note in this regard that, thanks to its religious function as
the language of Islam, the linguistic influence of (Classical) Arabic has of course
travelled well beyond the traditional borders of the Arabic-speaking world, and
has affected linguistic communities that have never acquired Arabic as a second
language. Such is the case, for example, of Indonesian and Swahili, whose lexica are characterized by a high proportion of Arabic-derived loanwords. In the
present volume we largely disregard this kind of influence, however, as our focus is rather on the effects of language contact in communities characterized by
a relatively high degree of societal bilingualism in Arabic. These bilingual communities typically fall within Jastrow’s Zone II (see §3.1 and Figure 1), and are
therefore affected to varying degrees by language shift towards Arabic.
Accordingly, the first two chapters of Part II focus on the structural effects
of language contact with Arabic in two Semitic languages of the Middle East.
First of all, Bettega & Gasparini (this volume) provide an overview of Arabic influence on the Modern South Arabian languages (i.e. Mehri, Hobyōt, Ḥarsūsi,
Baṭḥari, Śḥerɛt/Jibbāli and Soqoṭri) of Oman and Yemen. These minority languages are used in an asymmetric pattern of bilingualism with Arabic, and have
been strongly affected by contact with the dominant language, both in their lexicon and grammar. A similar situation is described by Coghill (this volume) for
North-Eastern Neo-Aramaic (NENA), a group of closely related languages whose
speakers are scattered across Iraq, Turkey, Syria, and Iran (as well as in several
8
1 Introduction
diasporic communities around the world). Unlike for the Modern South Arabian
languages, however, Arabic has only recently become the dominant language in
much of the region where NENA languages are spoken, with Kurdish being the
primary historical contact language. Nevertheless, the intensity of this contact,
despite its relatively short duration, has been sufficient to result in significant
influence on the grammar and lexicon of NENA languages, as Coghill demonstrates. Being closely related to Arabic, NENA and Modern South Arabian languages are incidentally particularly relevant to the question of the role played
by language contact (i.e. diffusion) as opposed to internal drift (i.e. transmission)
in the reconstruction of the Semitic language family.
The next two chapters in Part II deal with languages that are also genetically
related to Arabic, though much more distantly. First of all, Souag (this volume)
surveys some of the most prominent examples of the influence of Arabic on the
numerous Berber languages spoken across North Africa and the Sahara. Though
many Berber-speaking communities are in the process of language shift, different communities present different patterns of bilingualism. Tuareg, for example,
has been least affected by contact with spoken Arabic, whereas smaller varieties,
such as that of Awjila in Libya, are severly endangered, with language shift to
Arabic being rather far advanced (van Putten & Souag 2015). Berber as a whole
thus represents a particularly rich source of data for the typology of changes
brought about by contact with Arabic (see also Kossmann 2013). Vanhove (this
volume), on the other hand, describes the influence of Arabic on Beja, a Northern
Cushitic language mainly spoken in eastern Sudan. Probably due to their constituting a large proportion of the population in this region, and in spite of their
high degree of bilingualism with Sudanese Arabic, Beja speakers continue robust
transmission of their ancestral language to younger generations and are therefore not involved in a process of language shift. Against this background, Beja
offers interesting hints for the analysis of the morphological effects of contact
with Arabic, especially in relation to the transfer of roots and patterns (see also
Vanhove 2012).
Part II of the volume also provides data for the analysis of contact-induced
changes that occurred in languages with no genetic link with Arabic. These are
all Indo-Iranian languages, spoken in a large area stretching from Iran in the
east to Israel in the west. Gazsi (this volume) offers a wide-ranging survey of the
mostly lexical influence of Arabic on Iranian languages, with a particular focus
on New Persian and Modern Persian dialects spoken in Iran. Öpengin (this volume) then describes the effects of contact with Arabic in Northern and Central
Kurdish languages spoken in Turkey, Syria, and Iraq. Due to the longstanding bilingualism with Arabic since the early phases of the Islamic expansion, Kurdish
9
Christopher Lucas & Stefano Manfredi
has been profoundly affected in its phonology and lexicon by contact with both
Mesopotamian dialects and Classical Arabic. Lastly, two further chapters assess
the changes produced by contact with Arabic in different varieties of Domari,
an Indic language spoken by itinerant linguistic minorities in the Middle East.
Matras (this volume) analyses the Southern variety of Domari, spoken in Jerusalem, which is reported to be extremely endangered, while Herin (this volume)
focuses on the Northern varieties of Domari, spoken in Syria, Lebanon, Jordan,
and Turkey, which exhibit different degrees of linguistic vitality. In this overall
situation, Domari has been thoroughly affected in all lexical and grammatical
domains by contact with Arabic, with dialects of Syria and Turkey showing a
lower degree of linguistic interference, while more southerly dialects are on the
verge of extinction due to language shift.
In the final chapter of Part II, Nolan (this volume) discusses another contact
language with significant input from Arabic: Mediterranean Lingua Franca, a vehicular language spoken from the sixteenth to nineteenth centuries on the North
African Barbary coast as an interethnic means of communication between various populations, including pirates and captured slaves. The lexicon and grammar
of Mediterranean Lingua Franca were apparently drawn from a wide range of
Italo-Romance, Spanish, Portuguese, Franco-Provençal, Turkish, Greek and Arabic varieties. Although the contribution of Arabic to this language was relatively
slight, a substantial proportion of its speakers had Arabic as their first language
and inevitably therefore transferred Arabic features into this contact language.
3.4 Overview of Part III: Domains of contact-induced change across
Arabic varieties
Parts I and II of the present volume offer overviews of contact-induced changes
in individual languages and Arabic varieties. Part III, by contrast, presents studies examining contact-induced change in various domains, across a number of
relevant languages and Arabic varieties. Some of these chapters focus on the processes producing contact-induced change in Arabic (e.g. dialect contact, contactinduced grammaticalization), while the others describe the outcomes of language
contact in specific grammatical domains (e.g. intonation, negation) in a crossdialect perspective. Taken together, the chapters included in Part III provide a
broader framework for understanding the dynamics and results of language contact involving Arabic.
First of all, drawing on the concepts of koinéization and focusing, as defined
by Trudgill (2004), Al-Wer (this volume) describes the process of new dialect
formation in Amman, resulting from the contact there between Palestinian and
10
1 Introduction
Jordanian dialects. Through examination of a number of morphophonological
variants, Al-Wer assesses the relative contributions of different social factors in
the formation of the Amman dialect, concluding that gender and style are the
major organizing factors, while ethnicity plays only a secondary role.
In the following chapter, Cotter (this volume) addresses the closely related
topic of phonetic and phonological changes, affecting both consonant and vowel
systems, resulting from contact between Arabic dialects. Cotter’s analysis emphasizes the role of large-scale migration within and between Arabic-speaking
countries in the emergence of phonological diversity in Arabic, as in the case of
the dialect of Gaza City, which presents both Bedouin and sedentary phonological features.
Though far less often considered from a historical linguistic perspective than
segmental changes, supra-segmental change also appears to be particularly liable to be caused by language contact. In this vein, Hellmuth (this volume) explores the hypothesis that variation in the intonation systems of Arabic dialects
is largely a product of language contact. Describing a series of dialect-specific
prosodic features in Tunisian, Moroccan and Egyptian Arabic, Hellmuth proposes different contact scenarios with Berber in the Maghreb and with Greek
and Coptic in Egypt as the cause, though without excluding the possibility of
purely internal prosodic change.
As evidenced by almost every contribution to the present volume, contactinduced change is certainly not limited to lexicon and phonology, with the impact of language contact clearly felt also in the morphosyntax and semantics
of Arabic varieties. Accordingly, Leddy-Cecere (this volume) adopts the theoretical framework of contact-induced grammaticalization proposed by Heine &
Kuteva (2003; 2005) for an analysis of the outcome of contact between Arabic
dialects in the domain of future tense markers. Though traditionally situated
in the context of contact between genetically unrelated languages, this model
of contact-induced change proves useful for explaining the development and
distribution of a range of morphosyntactic features across Arabic varieties (cf.
Leddy-Cecere 2018). In his contribution, Leddy-Cecere identifies five prototypical paths of grammaticalization of future markers whose spread, he argues, is
best explained as the outcome of dialect contact.
Manfredi (this volume), for his part, focuses exclusively on the process of
calquing, understood as the transfer of semantic and morphosyntactic patterns
without accompanying morphophonological matter. He thus analyses several instances of lexical and grammatical calquing in a range of Arabic varieties, and
explains their distribution in terms of different degrees of bilingual proficiency.
This perspective permits an explanation of why narrow grammatical calquing
11
Christopher Lucas & Stefano Manfredi
tends to be limited to communities with a high degree of bilingual proficiency,
whereas lexical calquing can occur also in largely monolingual communities.
In the final contribution to Part III, Lucas (this volume) presents a diachronic
overview of the development of different negation patterns in Arabic and a number of its contact languages. While recognizing that conclusive evidence of diffusion as opposed to transmission in this domain is hard to come by, Lucas argues
that the geographical distribution of preverbal, bipartite, and postverbal clausal
negation in Arabic and its contact languages (i.e. Modern South Arabian, Berber
and Domari among others) is a product of transfer, rather than of internal parallel
developments (see also Lucas & Lash 2010).
3.5 Limitations
Inevitably with a project of this scale, it has not been possible to cover every
aspect of the topic that we would have liked to, and the chapters included necessarily represent a compromise between several different academic and practical
considerations (not least the availability of contributors with the relevant expertise). Thus, while we have aimed for blanket coverage of languages and varieties
of Arabic that have been significantly affected by contact, a number of omissions
should be noted.
For example, Central Asian Arabic (see Seeger 2013), a minority variety strongly affected by contact with Tajik and Uzbek, though it is cited a number of times
for comparative purposes by several contributors, is not thoroughly analysed
in a dedicated chapter in Part I. Similarly, the influence of Modern Hebrew on
Palestinian Arabic in Israel (see Horesh 2015) is not analysed in detail here. Furthermore, with the exception of Nigerian Arabic, the volume has regrettably little
to say about the range of vernacular and vehicular varieties of Arabic spoken in
sub-Saharan Africa (see Lafkioui 2013b).
Similarly, the languages discussed in Part II are certainly not the only ones
to have been affected by direct contact with Arabic. For instance, several NiloSaharan languages found in central and eastern Africa have historically been in
contact with different varieties of Arabic. This is the case of Nubian, an Eastern Sudanic language spoken on the Egypt–Sudan border (Rouchdy 1980), for
example. The same applies to a number of Niger-Kordofanian languages spoken
in the Nuba Mountains region of Sudan, and among which we can mention the
case of Koalib (Quint 2018). As far as the Middle East is concerned, the influence
of Arabic on the Armenian varieties spoken in Lebanon unfortunately remains
unstudied, and the same is true for the Turkmen dialects of Iraq and Syria.
12
1 Introduction
There are also several phenomena that can be observed in multiple Arabic varieties and for which explanations in terms of language contact have been made,
but on which it was not possible to include a chapter in the present volume. To
cite a single example, several works (including Coghill 2014; Döhla 2016; Souag
2017) have investigated the possible role of contact between varieties of Arabic
and other languages in the development of differential object marking and clitic
doubling (see also Lucas & Čéplö, this volume).
Despite these descriptive gaps, the chapters included in the present volume
have the collective merit of discussing a wide range of contact situations involving Arabic (balanced bilingualism, unbalanced bilingualism, pidginization and
creolization), covering a broad geographical area and lengthy timespan, and thus
giving a near-comprehensive picture of the currently known facts of Arabic and
contact-induced change.
4 Framework
4.1 Overview
The majority of works cited in §2 (like the majority of work generally on contactinduced changes in specific languages) describe a set of linguistic outcomes of
language contact, without addressing the cognitive and acquisitional processes
that lead speakers to introduce and adopt changes of this kind. In the present volume, we have encouraged authors wherever possible to go beyond mere itemization of contact-induced changes, and to give consideration to the processes which
are likely to have brought them about. Specifically, we have asked authors to
analyse changes wherever possible in terms of the framework (and terminology)
developed by Frans Van Coetsem (1988; 2000).
While there are various models of contact-induced change available (see e.g.
Thomason & Kaufman 1988; Johanson 2002; Matras 2009), Van Coetsem’s is
preferable for our purposes, in that it allows us to distinguish the major types
of contact-induced change, based on the cognitive statuses of the source and recipient languages in the minds of the bilingual speakers who are the agents of
the changes in question. This model, which has gained greater prominence following Winford’s (2005; 2007; 2010) work to popularize it (see also Ross 2013 for
a broadly similar approach), makes a fundamental distinction between borrowing and imposition as the two major types of transfer (i.e. contact-induced
change that has the effect of making the RL more closely resemble the SL in
some respect).1 The distinction between borrowing and imposition boils down
1
Note that not all contact-induced changes involve transfer in this sense. See §4.4 for details.
13
Christopher Lucas & Stefano Manfredi
to whether the agents of a particular change (i.e. the bilingual speakers who first
introduce it) are cognitively (not sociolinguistically) dominant in the SL or the
RL. Lucas (2012; 2015) argues that this notion of dominance (which Van Coetsem
himself does not define precisely) can be reduced to nativeness, and is thus not
equivalent to temporary accessibility: borrowing (also referred to as change under RL agentivity) is when a speaker for whom the RL is a native language introduces changes to the RL based on an SL model; imposition (also referred to as
change under SL agentivity) is when changes of this sort are made by a speaker
for whom the RL is not a native language. Imposition occurs essentially because
adults, with their impoverished language acquisition abilities relative to young
children, consciously or unconsciously draw on the resources of their native language(s) to fill the gaps in their knowledge of the non-native RL. Borrowing, on
the other hand, occurs either as a deliberate enrichment of the native language
with material drawn from a second language, or otherwise as a result of the
“inherent cognitive tendency to minimize the processing effort associated with
the use of two (or more) languages” (Lucas 2012: 291). Imposition thus prototypically transfers more abstract structural features (e.g., for German native speakers
speaking second-language English, syllable-final devoicing and lack of preposition stranding), whereas borrowing is prototypically associated with transfer of
lexical and constructional material.
This approach neatly complements Labov’s distinction between transmission
and diffusion. Labov (2007: 349) points out that “transmission is the product of
the acquisition of language by young children” whereas “most language contact is largely between and among adults” and that the fundamental differences
between child first language and adult second language acquisition (cf. BleyVroman 1989; 2009; Meisel 2011) explain the characteristically different types of
change associated with transmission versus diffusion. We can go further and say
that diffusion changes are of two main types – borrowing versus imposition –
and it is similarly because borrowing is carried out by native speakers and imposition by second language learners that these two types of diffusion typically
have different results (see §6 for further discussion).
Moreover with this approach we even have a prospect, at least in certain specific cases, of addressing one of the hardest problems in historical linguistics,
Weinreich et al.’s (1968) “actuation problem”:
For even when the course of a language change has been fully described and
its ability explained, the question always remains as to why the change was
not actuated sooner, or why it was not simultaneously actuated wherever
identical functional conditions prevailed. (Weinreich et al. 1968: 112)
14
1 Introduction
If the change in question involves diffusion, understood in the above terms, then
we have a straightforward answer to this question. Prior to contact with the
SL, the change did not occur because the linguistic conditions were such that it
could not occur in a normal language transmission scenario. Once the RL comes
into contact with the SL, however, the landscape of language acquisition and
use is drastically altered, such that the linguistic conditions are now sufficient to
trigger the change, which can then, potentially, spread throughout and beyond
the bilingual speech community (see Lucas, this volume; Lucas & Lash 2010 for
further discussion of this point in the context of the contact-induced spread of
bipartite negation in the languages of North Africa and southern Arabia).
To illustrate these concepts, the following subsections give some examples of
borrowing and imposition (as well as some problematic cases that do not fit easily
into either of these categories), drawn from the contributions to this volume.
4.2 Borrowing
As noted above, borrowing most typically and saliently targets lexical items. Every chapter in Parts I and II testifies to the large number of loanwords in the
varieties discussed. While borrowing prototypically involves content words, it
can also result in transfer of function words, idiomatic structure, and derivational
and inflectional morphology. For example, Vanhove (this volume) notes that Beja
has borrowed the Arabic conjunction wa ‘and’ as an enclitic which coordinates
noun phrases and nominalized clauses, as in (1).
(1)
Beja (BEJ_MV_NARR_01_shelter_057)2
bʔaɖaɖ=wa
i=koːlej=wa
sallam-ja=aj=heːb
sword=coord def.m=stick=coord give-pfv.3sg.m=csl=obj.1sg
‘Since he had given me a sword and the stick…’
Leitner (this volume) shows that Khuzestan Arabic has borrowed a phrasal
verb constructional frame from Persian, as illustrated in (2), consisting of an Arabic light verb (a calque of the Persian source verb) and a noun borrowed from
Persian.
(2)
2
a. Khuzestan Arabic (Leitner’s field data)
kað̣ð̣
īrād
take.prf.3sg.m nagging
‘to pick on someone’
See Vanhove (this volume) for details of the source of this example.
15
Christopher Lucas & Stefano Manfredi
b. Persian
īrād
gereftan
nagging take.inf
‘to pick on someone’
As an example of the borrowing of derivational morphology, Benkato (this
volume) cites the Moroccan Arabic circumfix tā-...-t, borrowed from Berber, as
the regular means of deriving nouns of professions and traits, as in tānǝžžāṛt
‘carpentry’ (< nǝžžāṛ ‘carpenter’).
Finally, in the domain of verbal inflection, we can point to the contact-induced
grammaticalization in NENA of a prospective future marker zi-, as in (3), on the
model of Arabic raḥ- with the same function, both deriving from elements with
the basic meaning of ‘going’.
(3) Christian Telkepe NENA (Coghill’s field data)
zi-napl-ɒ
prsp-fall.pres-3sg.f
‘She’s going to fall.’
4.3 Imposition
As well as changes due to borrowing, the contributions to this volume cite numerous instances of changes due to imposition, which are typically more abstract
and less lexical–constructional than changes due to borrowing.
In the domain of phonology, we can point to the example of conditioned
monophthongization found only in the Arabic dialects of coastal Syria and northern Lebanon, almost certainly as a result of imposition from Aramaic, older layers
of which shared this feature. As Procházka (this volume) shows, in the dialect of
the island of Arwad *ay and *aw are preserved only in open syllables. Elsewhere
they merge to /ā/, as illustrated in (4).
(4)
Arwad Arabic, western Syria (Procházka 2013: 278)
*bayt, *baytayn > bāt, baytān ‘house, two houses’
*yawm, *yawmayn > yām, yawmān ‘day, two days’
*bayn al-iθnayn > bān it-tnān ‘between the two’
In the domain of morphosyntax, van Putten (this volume) cites Wilmsen’s
(2010) example of imposition in the treatment of direct and indirect pronominal
objects in MSA. As Wilmsen shows, native speakers of Egyptian Arabic writing
MSA tend to impose their native system, such that the direct object cliticizes to
16
1 Introduction
the verb, as in (5), whereas native speakers of Lebanese Arabic tend to impose
their native system, such that it is the indirect object that cliticizes to the verb,
as in (6).
(5)
Egyptian-style MSA (Wilmsen 2010: 100)
llatī
sallamat-hā
la-hu
ʔarmalat-u
al-ʔawrāq-i
def-papers-obl rel.sg.f give.prf.3sg.f-3sg.f dat-3sg.m widow-nom
ʕabdi l-wahhāb
pn
‘the papers, which Abdel Wahhab’s widow had given him’
(6) Lebanese-style MSA (Wilmsen 2010: 99)
al-ʔawrāq-i
llatī
sallamat-hu
ʔiyyā-hā ʔarmalat-u
def-papers-obl rel.sg.f give.prf.3sg.f-3sg.m acc-3sg.f widow-nom
ʕabdi l-wahhāb
pn
‘the papers, which Abdel Wahhab’s widow had given him’
Taken together, the above examples give an impression of the nature and variety of changes that are reported on in this volume, and which can be understood
as having occurred via either borrowing or imposition.
4.4 Problematic cases
Not all changes due to contact can be classified as either borrowing or imposition in Van Coetsem’s terms, however. First of all, there is the rather frequent
case of communities in which the norm is not monolingual native acquisition
followed by acquisition of a second language later in life, but the simultaneous
acquisition, from early childhood, of two (or more) native languages. While Van
Coetsem (2000) acknowledges such cases, the data from studies of bilingual individuals of this type do not bear out his suggestion (2000: 86) that these situations
lead to “free transfer” of elements from any linguistic domain between the two
languages. Instead, what we see in both the speech of (young) individuals of this
kind, as well as communities in which multiple native languages are the norm, is
typically little phonological transfer but often considerable syntactic reorganization (Lucas 2009: 96–98; 2012: 279). The traditional term for the process by which
languages (typically in so-called “linguistic areas” such as the Balkans) become
more similar over time is convergence. Lucas (2015) extends the use of this term
to specifically those contact-induced changes brought about by individuals who
are native speakers of both the RL and the SL.
17
Christopher Lucas & Stefano Manfredi
Language situations described in this volume in which convergence in this
sense, rather than borrowing, is the likely mechanism underlying the changes
described include the Modern South Arabian languages, especially Baṭḥari, as
described by Bettega & Gasparini (this volume), as well as both Northern and
Jerusalem Domari, as described by Herin (this volume) and Matras (this volume).
As several authors point out, however, for some historical contact situations we
simply do not have enough sociolinguistic information to be able to infer what
kind of agentivity must underlie a given change. In such cases we must content
ourselves with merely identifying the changes that are (likely) due to contact
and, for the time being at least, give up on the goal of actually explaining how
and why they were actuated.
Finally, a word is required here on changes, such as reduction or elimination of
inflectional distinctions, which are characteristic of the usage of second-language
speakers, but which do not necessarily have the effect of making the RL more
closely resemble the native language of those speakers, and are not therefore
properly classified as instances of transfer. Lucas (2015) gives the label “restructuring” to changes of this kind, which presumably occur in almost any contact
situation where imposition is also taking place, though they will usually go undetected, being indistinguishable after the fact from purely internally caused
changes. One circumstance where restructuring changes are clearly identifiable,
however, concerns pidgins and creoles. Where these show a reduction in morphological complexity relative to the lexifier language that also does not represent
transfer from the substrate(s), this can only have been caused by restructuring.
See Avram (this volume) for several cases of this kind involving Arabic-based
pidgins and creoles.
5 Layout of chapters
5.1 Structure
Chapters in each Part of the present volume follow a fixed basic structure. In
Part I chapters, the first section gives sociolinguistic, demographic, and other
relevant background information on the current state and/or historical development of the dialect(s) or varieties of Arabic under discussion. The second section then details the languages which the variety under discussion is or was in
contact with, and describes the nature of those contacts. The third and main section then provides the data on the most noteworthy contact-induced changes in
the variety under discussion. In general, changes described in this third section
are ordered: phonology, morphology, syntax, lexicon. All chapters finish with a
18
1 Introduction
concluding section that includes an outline of what we still do not know about
contact-induced change the variety in question, as well as the most urgent issues
for future research. Part II chapters on language change through contact with
Arabic follow the same structure, with the second section focusing on the nature of the contact between Arabic and the language under discussion, as well as
any other significant contacts in the case of those languages which have had contact influence from multiple languages. Since Part III focuses on contact-induced
changes in specific, rather distinct, linguistic domains, the structure of chapters
in this Part is less uniform, but each chapter begins with an introduction to the
topic from a general linguistic point of view, followed by an overview of contactinduced changes in the domain in question, and finally a conclusion which again
includes discussion of what remains unclear about the topic of the chapter, as
well as the most promising avenues for future research.
5.2 Transcription and glossing
All chapters in the present volume adhere as far as possible to a single consistent
system of transcription and glossing of numbered examples. In this subsection
we summarize key elements of these two systems.
Examples from any language which has an official standardized Latin-script orthography (such as English, French, or Maltese) are transcribed in that orthography. Other than Arabic, any languages with no official standardized orthography,
or only one which is not based on the Latin script, are transcribed according to
a consistent scholarly system of each contributor’s choosing. The International
Phonetic Alphabet (IPA) is used only when the specific focus of discussion is
points of phonological or phonetic detail. All Arabic examples in the volume are
transcribed in accordance with the system for consonants laid out in Table 1.
In this table, voiced/voiceless pairs appear with the voiced sound immediately
below its voiceless counterpart. Emphatic sounds (i.e. sounds with a secondary
pharyngeal/uvular/velar articulation) appear immediately to the right of their
plain counterparts, and are distinguished from them with a dot below.3 This
broad phonemic system only distinguishes sounds which express meaningful
contrasts (and vowels are transcribed following the same principles). For subphonemic contrasts that cannot be captured with the symbols in Table 1, the IPA
is used. Gemination is signalled by doubling consonant symbols, vowel length by
3
Note that /ḥ/ does not, however, represent an emphatic version of /h/. We have chosen to retain
the use of the traditional symbol 〈ḥ〉 (rather than 〈ħ〉) for the voiceless pharyngeal fricative,
despite this unwanted implication that it represents an emphatic sound, so as to avoid confusion with the use of the symbol 〈ħ〉 in the Maltese orthography (for details of which, see Lucas
& Čéplö, this volume).
19
Christopher Lucas & Stefano Manfredi
Nasal
Vibrant
Lateral
Approximant
a
f
v
m
w
θ
ð ð̣
sṣ
zẓ
n
rṛ
l
č
ǧ
š
ž
ḫ (x)a
ɣ
ttal
q
Glo
k
g
P ha
r yn
Uvu
lar
gea
l
lar
Pal
ata
l
tṭ
dḍ
Affricate
Fricative
Pos
talv
eo
Alv
eola
r
tal/
Den
rde
nt
p
b
Vel
ar
Plosive
Inte
Lab
ial
al
Table 1: Transcription system for Arabic consonants
ʔ
ḥ
ʕ
h
y
〈ḫ〉 represents the voiceless velar fricative phoneme in all Arabic varieties where this contrasts
with pharyngeal and glottal fricative phonemes. In Walter’s (this volume) chapter on Cypriot
Maronite Arabic, however, the symbol 〈x〉 is used to represent the single phoneme in that
variety that is the outcome of the merger of the voiceless fricatives at all three of these places
of articulation.
a macron above the long vowel. Stress is only marked for Arabic (with an acute
accent on the nuclear vowel) where it marks a meaningful contrast, or where it
is otherwise the focus of discussion in a particular passage.
Glossing of linguistic examples in the volume is handled similarly to transcription. The Leipzig Glossing Rules are followed throughout, with extensions where
necessary. Every chapter includes at the end a list of glossing and other abbreviations used in that chapter. Within these parameters authors make their own
choices for precisely how they wish to gloss languages other than Arabic. For
all Arabic examples in the volume, we have tried to ensure that way they are
glossed is completely consistent. Some of the key choices we have made in this
regard are as follows.
As is well known, regular verbs in Arabic varieties have two basic conjugations: one in which the person–number affixes are exclusively suffixal, and one
in which they are mainly prefixal. The suffix conjugation typically (but not al-
20
1 Introduction
ways) functions to express past tense and/or perfective aspect, while the prefix
conjugation typically (but not always) functions to express non-past tense and/or
imperfective aspect. Since our aim with all glossing in the volume is to have one
consistent gloss per morpheme, regardless of the precise temporal or aspectual
functions in context, we have chosen to use the traditional Arabist labels of perfect and imperfect for these two conjugations, as opposed to alternatives such
as past/non-past or perfective/imperfective. The abbreviations used are prf for
perfect and impf for imperfect.
Related to the issue of how best to label these two conjugations is the question
of how best to analyse the distribution of person, number, gender, tense–aspect,
and mood features across the verb stem and any affixes. The details need not concern us here, but finding an intuitive way of assigning each of these features to
an appropriate morpheme, in a way that is consistent across all cells in the relevant paradigms, is extremely challenging. For this reason, in the present volume
we make no attempt at morphological decomposition in the glossing of a word
such as MSA yaktubūna ‘they (m.) write’. This is glossed simply as: yaktubūna
‘write.impf.3pl.m’. Accordingly, sallamat in (5) is glossed as ‘give.prf.3sg.f’. It follows from this that the absence of a hyphen in a string of Arabic in a numbered
example cannot be taken to imply that that string is monomorphemic. Relatedly,
we make no attempt to distinguish between clitics and affixes in the glossing
of Arabic examples in the present volume: a morpheme boundary of any sort is
signalled by a hyphen.
The overarching principle we have followed in all of these decisions on glossing and transcription is to try to present the relevant linguistic data in as clear,
plain, and unambiguous a format as possible.
6 Problems and prospects
As discussed in §4, Van Coetsem’s framework, with its basic distinction between
borrowing and imposition, has the merit of enabling us not only to coherently categorize many contact-induced changes according to the processes of language acquisition and use that produced them, but also, at least in some cases, to attempt
to address Weinreich et al.’s (1968) actuation problem, and so provide a genuinely
explanatory account of the genesis of individual contact-induced changes.
This is certainly not to claim, however, that Van Coetsem’s framework, in the
way that he himself presents it, is without its weaknesses. We have already discussed in §4.4 some instances of contact-induced change which are not easily
accommodated by the neat dichotomy between the two main transfer types: this
21
Christopher Lucas & Stefano Manfredi
is why Lucas (2015) proposes extending Van Coetsem’s model to accommodate
convergence and restructuring as additional transfer types.
A more fundamental problem is that, for many of the changes discussed in this
volume and elsewhere, there is simply not enough sociohistorical information
available to be able to infer with confidence what precise mechanisms underlie
the changes in question. In such cases Van Coetsem (1988; 2000) and, following
him, Winford (2005) suggest that the type of transfer that was operative in a
given change can be diagnosed from its results. That is, for example, if a change
involves word order, we can assume that it was due to imposition, while loanwords can be assumed to have been introduced via borrowing. Van Coetsem
(1988: 25) argues that this is so because “language does not offer the same degree of stability in all its parts, in particular […] there are differences in stability
among language domains, namely among vocabulary, phonology and grammar
(morphology and syntax).” He labels this observation the stability gradient,
and suggests that it is this supposed fact about language that underlies the observed discrepancies between the types of change characteristically associated
with borrowing and imposition respectively. As argued by Lucas (2012; 2015),
however, there is no a priori or empirical reason to believe that the whole of
“grammar” – a term which covers a range of highly heterogeneous phenomena
– should necessarily behave similarly in language contact situations, with any
contact-induced grammatical changes necessarily being due to imposition. This
argument does not of course deny the strong tendency, already pointed out in
§4.1, for imposition to be systematic and to target abstract structural features,
while borrowing is more sporadic and centred on lexicon. But if the stability gradient only reflects a tendency, not an exceptionless law, then its usefulness as a
diagnostic tool is greatly reduced. Indeed, several authors have pointed out that
there are clear cases of contact-induced grammatical change for which only RL
agentivity is plausible. For example, Kossmann (2013: 430) points out that, though
the predictions of the stability gradient tend to be borne out in cases in which
phonological and morphological change are mediated through borrowed lexical
items, there are however also cases in which elements of Arabic structure (e.g.
the syntax of clausal coordination and relativization) have been transferred into
Berber under RL agentivity, without obviously being related to lexical transfer.
Further challenges to the idea of the stability gradient are provided by several
of the contributions to the present volume. For example, Leitner (this volume)
points to the transfer of verb–auxiliary order from Persian to Khuzestan Arabic as an instance of abstract structural transfer (not the transfer of a specific
construction) in a context in which only borrowing, not imposition, can be the
cause (cf. (2) in §4.2). Similarly, Walter (this volume) points out that in Cypriot
22
1 Introduction
Maronite Arabic there has been systematic abstract phonological (as well as syntactic) transfer from Greek, in a sociolinguistic situation in which RL-dominant
individuals must have been the agents of change. In the contribution of Manfredi
(this volume), the necessity for a fine-grained approach to how transfer interacts with the different types of agentivity is brought into sharp relief, thanks
to Manfredi’s distinction between three types of grammatical calquing, two of
which involve the calquing of polyfunctionality of lexical or grammatical items
with or without syntactic change, while the third is a “narrow” type, producing
syntactic change without calquing of the functions of lexical/grammatical items.
A simplistic approach that sees lexicon and grammar as wholly distinct, internally homogeneous entities is clearly inadequate for an understanding of the
mechanisms underlying changes of this sort.
A final challenge to a straightforward application of Van Coetsem’s framework
to problems in contact linguistics concerns the emergence of new languages in
extreme contact situations. According to Winford (2005: 396; 2008: 128), the processes that create contact languages are the same as those that operate in ordinary cases of contact-induced language change. Thus he identifies three broad categories of contact languages: those that arise through RL agentivity (i.e. borrowing); those that arise primarily through SL agentivity (i.e. imposition); and those
that arise through a combination of SL and RL agentivities (see also Manfredi
2018: 414). From the perspective of this classification, Winford points out that
creole languages, since they emerge in a context of second language acquisition,
are essentially a product of SL agentivity. But if we take a closer look at Arabicbased pidgins and creoles (Avram, this volume), the picture is more complex.
For example, a number of phonological features of Juba Arabic (e.g. loss of pharyngeal and pharyngealized consonants; loss of consonant and vowel length)
are clearly attributable to imposition from Bari, the main substrate language,
during the first phases of its emergence. In the same manner, the lexical and
grammatical semantics of Juba Arabic are strongly affected by those of Bari, as
shown by several cases of calquing (Manfredi, this volume). However, a number of phonological and morphological innovations (e.g. presence of implosive
sounds and integration of nominal prefixes and suffixes) must instead be seen
as the result of borrowing enacted by Juba Arabic-dominant speakers latterly exposed to Bari as an adstrate language. What this shows is that creolization, being
necessarily multicausal, cannot be straightforwardly reduced to a single type of
linguistic transfer. Instead, it is essential that we combine the linguistic dominance approach with fine-grained sociohistorical criteria for typologizing contact
languages.
23
Christopher Lucas & Stefano Manfredi
As is evident from our decision to adopt Van Coetsem’s model as this volume’s
basic analytical framework, we believe that its focus on agentivity and dominance must be central to any attempt understand the cognitive factors that actually cause contact-induced change, as opposed to the sociolinguistic factors that
promote it. We do not consider, therefore, that the challenges for this framework
that we have explored in the current section are insurmountable (see Lucas 2012;
2015 for a detailed defence, revision, and application of the framework). Rather
our hope is that the ideas explored in this introduction, together with the wealth
of data presented in the following chapters, will serve as a stimulus for the wider
community of Arabists and historical linguists to push forward understanding
both of the history of the Arabic language, and of the nature of contact-induced
change in general.
Acknowledgments
The publication of this work would not have been possible without Leadership
Fellows grant AH/P014089/1 from the UK Arts and Humanities Research Council,
whose support is hereby gratefully acknowledged. We would also like to thank
Sebastian Nordhoff and Felix Kopecky of Language Science Press for their kind
assistance in bringing the project to fruition, as well as all those who so generously donated their time and expertise in the writing, reviewing, and proofreading of the chapters.
Abbreviations
*
1, 2, 3
acc
coord
csl
dat
def
f
impf
inf
IPA
m/m.
MSA
24
reconstructed form
1st, 2nd, 3rd person
accusative
coordination
causal
dative
definite
feminine
imperfect
infinitive
International Phonetic
Alphabet
masculine
Modern Standard Arabic
NENA
nom
obj
obl
pfv
pn
pres
prf
prsp
rel
RL
sg
SL
North-Eastern Neo-Aramaic
nominative
object
oblique
perfective
personal name
NENA Present Base
perfect
prospective
relative
recipient language
singular
source language
1 Introduction
References
Aikhenvald, Alexandra Y. 2002. Language contact in Amazonia. Oxford: Oxford
University Press.
Akkuş, Faruk & Elabbas Benmamoun. 2018. Syntactic outcomes of contact in
Sason Arabic. In Stefano Manfredi & Mauro Tosco (eds.), Arabic in contact, 38–
52. Amsterdam: John Benjamins.
Aquilina, Joseph. 1958. Maltese as a mixed language. Journal of Semitic Studies
3(1). 58–79.
Arnold, Werner. 2007. Arabic grammatical borrowing in Western Neo-Aramaic.
In Yaron Matras & Jeanette Sakel (eds.), Grammatical borrowing in crosslinguistic perspective, 185–195. Berlin: Mouton de Gruyter.
Arnold, Werner & Peter Behnstedt. 1993. Arabisch-aramäische Sprachbeziehungen
im Qalamūn (Syrien). Wiesbaden: Harrassowitz.
Asbaghi, Asya. 2011. Persian loanwords. In Lutz Edzard & Rudolf de Jong (eds.),
Encyclopedia of Arabic language and linguistics, online edn. Leiden: Brill.
Avram, Andrei A. 2017. Sources of Gulf Pidgin Arabic features. In Simone Bettega
& Fabio Gasparini (eds.), Linguistic studies in the Arabian Gulf, 131–151. Turin:
Università di Torino.
Bakker, Peter & Yaron Matras. 2013. Introduction. In Peter Bakker & Yaron Matras
(eds.), The mixed language debate: Theoretical and empirical advances, 1–14.
Berlin: Mouton de Gruyter.
Barbot, Michel. 1961. Emprunts et phonologie dans les dialectes citadins syrolibanais. Arabica 8(2). 174–188.
Benoliel, José. 1977. Dialecto judeo-hispano-marroquí o hakitia. Madrid: Varona.
Bizri, Fida. 2018. Contemporary Arabic-based pidgins in the Middle East. In Reem
Bassiouney & Elabbas Benmamoun (eds.), The Routledge handbook of Arabic
linguistics, 421–38. London: Routledge.
Blau, Joshua. 1969. The renaissance of Modern Hebrew and Modern Standard
Arabic: Parallels and differences in the revival of two Semitic languages. Berkeley:
University of California Press.
Blau, Joshua. 1981. The emergence and linguistic background of Judaeo-Arabic: A
study of the origins of Neo-Arabic and Middle Arabic. 3rd edn. Jerusalem: BenZvi Institute.
Bley-Vroman, Robert. 1989. What is the logical problem of foreign language learning? In Susan M. Gass & Jacquelyn Schachter (eds.), Linguistic perspectives on
second language acquisition, 41–68. Cambridge: Cambridge University Press.
Bley-Vroman, Robert. 2009. The evolving context of the Fundamental Difference
Hypothesis. Studies in Second Language Acquisition 31. 175–198.
25
Christopher Lucas & Stefano Manfredi
Borg, Alexander. 1997. Cypriot Arabic phonology. In Alan S. Kaye (ed.), Phonologies of Asia and Africa, 219–244. Winona Lake, IN: Eisenbrauns.
Borg, Alexander. 2004. A comparative glossary of Cypriot Maronite Arabic (Arabic–
English): With an introductory essay. Leiden: Brill.
Boumans, Louis & Jan Jaap de Ruiter. 2002. Moroccan Arabic in the European
diaspora. In Aleya Rouchdy (ed.), Language contact and language conflict in
Arabic: Variations on a sociolinguistic theme, 259–285. London: Routledge.
Brahimi, Fadila. 2000. Loanwords in Algerian Berber. In Jonathan Owens (ed.),
Arabic as a minority language, 371–382. Berlin: Mouton de Gruyter.
Brincat, Joseph M. 2011. Maltese and other languages: A linguistic history of Malta.
Malta: Midsea Books.
Brunot, Louis. 1949. Emprunts dialectaux arabes à la langue française depuis 1912.
Hespéris 36. 347–430.
Cifoletti, Guido. 1994. Italianismi nel dialetto di Tunisi. In Dominique Caubet &
Martine Vanhove (eds.), Actes des premières journées internationales de dialectologie arabe de Paris, 451–458. Paris: Langues’O.
Coghill, Eleanor. 2010. The grammaticalization of prospective aspect in a group
of Neo-Aramaic dialects. Diachronica 27(3). 359–410.
Coghill, Eleanor. 2012. Parallels in the grammaticalisation of Neo-Aramaic ziland Arabic raḥ- and a possible contact scenario. In Domenyk Eades (ed.), Grammaticalization in Semitic, 127–144. Oxford: Oxford University Press.
Coghill, Eleanor. 2014. Differential object marking in Neo-Aramaic. Linguistics
52(2). 335–364.
Coghill, Eleanor. 2015. Borrowing of verbal derivational morphology between
Semitic languages: The case of Arabic verb derivations in Neo-Aramaic.
In Francesco Gardani, Peter Arkadiev & Nino Amiridze (eds.), Borrowed
morphology, 83–108. Berlin: De Gruyter Mouton.
Colin, Georges S. 1957. Mots «berbères» dans le dialecte arabe de Malte. In
Mémorial André Basset (1895–1956), 7–16. Paris: Adrien Maisonneuve.
Corriente, Federico. 1978. Los fonemas /p/, /č/ y /g/ en árabe hispánico. Vox
Romanica 37. 214–218.
Corriente, Federico. 1992. Arabe andalusí y lenguas romances. Madrid: Editorial
MAPFRE.
Corriente, Federico. 2002. The Berber adstratum of Andalusi Arabic. In Otto
Jastrow, Werner Arnold & Hartmut Bobzin (eds.), “Sprich doch mit deinen
Knechten aramäisch, wir verstehen es!”: 60 Beiträge zur Semitistik: Festschrift
für Otto Jastrow zum 60. Geburtstag, 105–111. Wiesbaden: Harrassowitz.
Cotter, William M. & Uri Horesh. 2015. Social integration and dialect divergence
in coastal Palestine. Journal of Sociolinguistics 19(4). 460–83.
26
1 Introduction
Diem, Werner. 1979. Studien zur Frage des Substrats im Arabischen. Der Islam 56.
12–80.
Döhla, Hans-Jörg. 2016. The origin of differential object marking in Maltese. In
Gilbert Puech & Benjamin Saade (eds.), Shifts and patterns in Maltese, 149–174.
Berlin: De Gruyter Mouton.
El Aissati, Abderahman. 2011. Berber loanwords. In Lutz Edzard & Rudolf de Jong
(eds.), Encyclopedia of Arabic language and linguistics, online edn. Leiden: Brill.
El Zarka, Dina & Sandra Ziagos. 2019. The beginnings of word order change in
the Arabic dialects of southern Iran in contact with Persian: A preliminary
study of data from four villages in Bushehr and Hormozgan. Iranian Studies.
1–24. DOI:10.1080/00210862.2019.1690433
Gazsi, Dénes. 2011. Arabic–Persian language contact. In Stefan Weninger,
Geoffrey Khan, Michael P. Streck & Janet C. E. Watson (eds.), The Semitic
languages: An international handbook, 1015–1021. Berlin: De Gruyter Mouton.
Gibson, Maik. 2002. Dialect levelling in Tunisian Arabic: Towards a new standard. In Aleya Rouchdy (ed.), Language contact and language conflict in Arabic:
Variations on a sociolinguistic theme, 24–30. London: Routledge.
Haig, Geoffrey L. J. 2014. East Anatolia as a linguistic area? Conceptual
and empirical issues. In Lale Behzadi, Patrick Franke, Geoffrey L. J. Haig,
Christoph Herzog, Birgitt Hofmann, Lorenz Korn & Susanne Talabardon (eds.),
Bamberger Orientstudien, 13–36. Bamberg: University of Bamberg Press.
Heath, Jeffrey. 1989. From code-switching to borrowing: A case study of Moroccan
Arabic. London: Kegan Paul.
Heath, Jeffrey. 2015. D-possessives and the origins of Moroccan Arabic.
Diachronica 32(1). 1–33.
Hebbo, Ahmed. 1984. Die Fremdwörter in der arabischen Prophetenbiographie des
Ibn Hischām (gest. 218/834). Frankfut am Main: Peter Lang.
Heine, Bernd & Tania Kuteva. 2003. On contact-induced grammaticalization.
Studies in Language 27(3). 529–572.
Heine, Bernd & Tania Kuteva. 2005. Language contact and grammatical change.
Cambridge: Cambridge University Press.
Herin, Bruno. 2018. The Arabic component in Domari. In Stefano Manfredi &
Mauro Tosco (eds.), Arabic in contact, 19–36. Amsterdam: John Benjamins.
Horesh, Uri. 2015. Structural change in Urban Palestinian Arabic induced by contact with Modern Hebrew. In Aaron Michael Butts (ed.), Semitic languages in
contact, 198–233. Leiden: Brill.
Ingham, Bruce. 2011. Afghanistan Arabic. In Lutz Edzard & Rudolf de Jong (eds.),
Encyclopedia of Arabic language and linguistics, online edn. Leiden: Brill.
27
Christopher Lucas & Stefano Manfredi
Jastrow, Otto. 2002. Arabic dialectology: The state of the art. In In Shlomo Izre’el
(ed.), Semitic linguistics: The state of the art at the turn of the twenty-first century,
347–363. Winona Lake, IN: Eisenbrauns.
Jastrow, Otto. 2005. Uzbekistan Arabic: A language created by Semitic–Iranian–
Turkic linguistic convergence. In Éva Ágnes Csató, Bo Isaksson & Carina
Jahani (eds.), Linguistic convergence and areal diffusion: Case studies from
Iranian, Semitic and Turkic, 133–139. London: Routledge.
Jastrow, Otto. 2015. Language contact as reflected in the consonant system of
Ṭuroyo. In Aaron Michael Butts (ed.), Semitic languages in contact, 234–250.
Leiden: Brill.
Jeffery, Arthur. 2007. The foreign vocabulary of the Qurʾān. Leiden: Brill. Originally published in 1938 by Benoytosh Bhattacharyya.
Johanson, Lars. 2002. Structural factors in Turkic language contacts. Oxford:
Curzon Press.
Kossmann, Maarten. 2009. Loanwords in Tarifiyt. In Martin Haspelmath & Uri
Tadmor (eds.), Loanwords in the world’s languages: A comparative handbook,
191–214. Berlin: De Gruyter.
Kossmann, Maarten. 2010. Parallel system borrowing: Parallel morphological
systems due to the borrowing of paradigms. Diachronica 27(3). 459–487.
Kossmann, Maarten. 2013. The Arabic influence on Northern Berber. Leiden: Brill.
Krier, Fernande. 1976. Le maltais au contact de l’italien: Etude phonologique, grammaticale et sémantique. Hamburg: Helmut Buske.
Labov, William. 2007. Transmission and diffusion. Language 83(2). 344–387.
Lafkioui, Mena. 2013a. Reinventing negation patterns in Moroccan Arabic. In
Mena Lafkioui (ed.), African Arabic: Aprroaches to dialectology, 51–93. Berlin:
De Gruyter Mouton.
Lafkioui, Mena (ed.). 2013b. African Arabic: Approaches to dialectology. Berlin: De
Gruyter Mouton.
Lafkioui, Mena & Vermondo Brugnatelli (eds.). 2008. Berber in contact: Linguistic
and sociolinguistic perspectives. Cologne: Rüdiger Köppe.
Leddy-Cecere, Thomas. 2018. Contact-induced grammaticalization as an impetus
for Arabic dialect development. Austin: University of Texas. (Doctoral
dissertation).
Lentin, Jérôme. 2011. Middle Arabic. In Lutz Edzard & Rudolf de Jong (eds.),
Encyclopedia of Arabic language and linguistics, online edn. Leiden: Brill.
Lonnet, Antoine. 2011. Modern South Arabian. In Lutz Edzard & Rudolf de Jong
(eds.), Encyclopedia of Arabic language and linguistics, online edn. Leiden: Brill.
Lucas, Christopher. 2009. The development of negation in Arabic and Afro-Asiatic.
Cambridge: University of Cambridge. (Doctoral dissertation).
28
1 Introduction
Lucas, Christopher. 2012. Contact-induced grammatical change: Towards an explicit account. Diachronica 29. 275–300.
Lucas, Christopher. 2015. Contact-induced language change. In Claire Bowern &
Bethwyn Evans (eds.), The Routledge handbook of historical linguistics, 519–536.
London: Routledge.
Lucas, Christopher & Elliott Lash. 2010. Contact as catalyst: The case for Coptic
influence in the development of Arabic negation. Journal of Linguistics 46. 379–
413.
Luffin, Xavier. 2014. The influence of Swahili on Kinubi. Journal of Pidgin and
Creole Languages 29(2). 299–318.
Manfredi, Stefano. 2018. Arabic as a contact language. In Reem Bassiouney &
Elabbas Benmamoun (eds.), The Routledge handbook of Arabic linguistics, 407–
20. London: Routledge.
Manfredi, Stefano. Forthcoming. The Arab world. In Miriam Meyerhoff &
Umberto Ansaldo (eds.), The Routledge handbook of pidgin and creole languages.
London: Routledge.
Manfredi, Stefano & Mauro Tosco (eds.). 2018. Arabic in contact. Amsterdam: John
Benjamins.
Masliyah, Sadok. 1996. Four Turkish suffixes in Iraqi Arabic: -li, -lik, -siz and -çi.
Journal of Semitic Studies 41(2). 291–300.
Masliyah, Sadok. 1997. The diminutive in spoken Iraqi Arabic. Zeitschrift für
Arabische Linguistik 33. 68–88.
Matras, Yaron. 2007. Grammatical borrowing in Domari. In Yaron Matras &
Jeanette Sakel (eds.), Grammatical borrowing in cross-linguistic perspective, 151–
164. Berlin: Mouton de Gruyter.
Matras, Yaron. 2009. Language contact. Cambridge: Cambridge University Press.
Matras, Yaron & Maryam Shabibi. 2007. Grammatical borrowing in Khuzistani
Arabic. In Yaron Matras & Jeanette Sakel (eds.), Grammatical borrowing in
cross-linguistic perspective, 137–149. Berlin: Mouton de Gruyter.
McMahon, April. 2013. Issues in the genetic classification of contact languages.
In Peter Bakker & Yaron Matras (eds.), The mixed language debate: Theoretical
and empirical advances, 333–362. Berlin: Mouton de Gruyter.
Meisel, Jürgen. 2011. First and second language acquisition: Parallels and differences. Cambridge: Cambridge University Press.
Mifsud, Manwel. 1995. Loan verbs in Maltese: A descriptive and comparative study.
Leiden: Brill.
Miller, Catherine. 1993. Restructuration morpho-syntaxique en Juba-Arabic et
Ki-Nubi: A propos du débat universaux/substrat et superstrat dans les études
créoles. Matériaux Arabes et Sudarabiques (GELLAS) 5. 137–174.
29
Christopher Lucas & Stefano Manfredi
Miller, Catherine, Enam Al-Wer, Dominique Caubet & Janet C. E. Watson (eds.).
2007. Arabic in the city: Issues in dialect contact and language variation. London:
Routledge.
Neishtadt, Mila. 2015. The lexical component in the Aramaic substrate of
Palestinian Arabic. In Aaron Butts (ed.), Semitic languages in contact, 280–310.
Leiden: Brill.
Newton, Brian. 1964. An Arabic–Greek dialect. Word 20(sup2). 43–52.
Owens, Jonathan. 1985. The origins of East African Nubi. Anthropological
Linguistics 27(2). 229–271.
Owens, Jonathan. 2000a. Loanwords in Nigerian Arabic: A quantitative approach.
In Jonathan Owens (ed.), Arabic as a minority language, 259–346. Berlin: De
Gruyter.
Owens, Jonathan (ed.). 2000b. Arabic as a minority language. Berlin: Mouton de
Gruyter.
Owens, Jonathan. 2015. Idioms, polysemy, context: A model based on Nigerian
Arabic. Anthropological Linguistics 57(1). 46–98.
Owens, Jonathan. 2016. Dia-planar diffusion: Reconstructing early Aramaic–
Arabic language contact. In Manuel Sartori, Manuela Giolfo & Philippe
Cassuto (eds.), Approaches to the history and dialectology of Arabic in honor
of Pierre Larcher, 77–101. Leiden: Brill.
Owens, Jonathan. 2018. Why linguistics needs an historically oriented Arabic
linguistics. In Stefano Manfredi & Mauro Tosco (eds.), Arabic in contact, 207–
232. Amsterdam: John Benjamins.
Palva, Heikki. 2009. From qǝltu to gilit: Diachronic notes on linguistic adaptation
in Muslim Baghdad Arabic. In Enam Al-Wer & Rudolf De Jong (eds.), Arabic
dialectology in honour of Clive Holes on the occasion of his sixtieth birthday, 17–
40. Leiden: Brill.
Procházka, Stephan. 2002. Contact phenomena, code-copying, and codeswitching in the Arabic dialects of Adana and Mersin (southern Turkey). In
Abderrahim Youssi (ed.), Aspects of the dialects of Arabic today: Proceedings of
the 4th conference of the International Arabic Dialectology Association (AIDA),
Marrakesh, April 1-4, 2000. In honour of Professor David Cohen, 133–139. Rabat:
Amapatril.
Procházka, Stephan. 2011. Turkish loanwords. In Lutz Edzard & Rudolf de Jong
(eds.), Encyclopedia of Arabic language and linguistics, online edn. Leiden: Brill.
Procházka, Stephan. 2013. Traditional boatbuilding: Two texts in the Arabic
dialect of the island of Arwād (Syria). In Renaud Kuty, Ulrich Seeger &
Shabo Talay (eds.), Nicht nur mit Engelszungen: Beiträge zur semitischen
30
1 Introduction
Dialektologie. Festschrift für Werner Arnold zum 60. Geburtstag, 275–288.
Wiesbaden: Harrassowitz.
Quint, Nicolas. 2018. An assessment of the Arabic lexical contribution to
contemporary spoken Koalib. In Stefano Manfredi & Mauro Tosco (eds.),
Arabic in contact, 189–207. Amsterdam: John Benjamins.
Ratcliffe, Robert R. 2005. Bukhara Arabic: A metatypized dialect of Arabic in
Central Asia. In Éva Ágnes Csató, Bo Isaksson & Carina Jahani (eds.), Linguistic
convergence and areal diffusion: Case studies from Iranian, Semitic and Turkic,
141–51. London: Routledge.
Retsö, Jan. 2011. Aramaic/Syriac loanwords. In Lutz Edzard & Rudolf de Jong
(eds.), Encyclopedia of Arabic language and linguistics, online edn. Leiden: Brill.
Ross, Malcolm. 2013. Diagnosing contact processes from their outcomes: The
importance of life stages. Journal of Language Contact 6(1). 5–47.
Rouchdy, Aleya. 1980. Languages in contact: Arabic-Nubian. Anthropological
Linguistics 22(8). 334–344.
Rouchdy, Aleya (ed.). 1992. The Arabic language in America. Detroit: Wayne State
University Press.
Sayahi, Lotfi. 2014. Diglossia and language contact: Language variation and change
in North Africa. Cambridge: Cambridge University Press.
Seeger, Ulrich. 2013. Zum Verhältnis der zentralasiatischen arabischen Dialekte.
In Renaud Kuty, Ulrich Seeger & Shabo Talay (eds.), Nicht nur mit Engelszungen: Beiträge zur semitischen Dialektologie. Festschrift für Werner Arnold
zum 60. Geburtstag, 313–322. Wiesbaden: Harrassowitz.
Souag, Lameen. 2013. Berber and Arabic in Siwa (Egypt): A study in linguistic
contact. Cologne: Rüdiger Köppe.
Souag, Lameen. 2016. Language contact in the Sahara. In Mark Aronoff (ed.),
Oxford research encyclopedia of linguistics. Oxford: Oxford University Press.
Souag, Lameen. 2017. Clitic doubling and language contact in Arabic. Zeitschrift
für Arabische Linguistik 66. 45–70.
Souag, Lameen. 2018. Berber etymologies in Maltese. International Journal of
Arabic Linguistics 4(1). 190–223.
Stolz, Thomas. 2003. Not quite the right mixture: Chamorro and Malti as
candidates for the status of mixed language. In Yaron Matras & Peter Bakker
(eds.), The mixed language debate: Theoretical and empirical advances, 271–315.
Berlin: De Gruyter Mouton.
Taine-Cheikh, Catherine. 1997. Les emprunts au berbère zénaga: Un soussystème vocalique du ḥassāniyya. Matériaux Arabes et Sudarabiques (GELLAS)
8. 93–142.
31
Christopher Lucas & Stefano Manfredi
Taine-Cheikh, Catherine. 2018. Ḥassāniyya Arabic in contact with Berber: The
case of quadriliteral verbs. In Stefano Manfredi & Mauro Tosco (eds.), Arabic
in contact, 135–159. Amsterdam: John Benjamins.
Talmoudi, Fathi. 1986. A morphosemantic study of Romance verbs in the Arabic dialects of Tunis, Sūsa, and Sfax. Gothenburg: Acta Universitatis Gothoburgensis.
Taylan, Eser. 2017. Language contact in Anatolia: the case of Sason Arabic.
In Ramazan Korkmaz & Doğan Gürkan (eds.), Endangered languages of the
Caucasus and beyond, 209–225. Leiden: Brill.
Thomason, Sarah G. 2011. Language contact. In Lutz Edzard & Rudolf de Jong
(eds.), Encyclopedia of Arabic language and linguistics, online edn. Leiden: Brill.
Thomason, Sarah G. & Terrence Kaufman. 1988. Language contact, creolization
and genetic linguistics. Berkeley: University of California Press.
Trudgill, Peter. 2004. New-dialect formation: The inevitability of colonial Englishes.
Oxford: Oxford University Press.
Trudgill, Peter. 2011. Sociolinguistic typology: Social determinants of linguistic
complexity. Oxford: Oxford University Press.
Tsabolov, Ruslan. 1994. Notes on the influence of Arabic on Kurdish. Acta Kurdica
1. 121–124.
Tsiapera, Maria. 1964. Greek borrowings in the Arabic dialect of Cyprus. Journal
of the American Oriental Society 84(2). 124–126.
Van Coetsem, Frans. 1988. Loan phonology and the two transfer types in language
contact. Dordrecht: Foris.
Van Coetsem, Frans. 2000. A general and unified theory of the transmission process
in language contact. Heidelberg: Winter.
van der Wal Anonby, Christina. 2015. A grammar of Kumzari. Leiden: University
of Leiden. (Doctoral dissertation).
van Putten, Marijn & Lameen Souag. 2015. Attrition and revival in Awjila Berber:
Facebook posts as a new data source for an endangered Berber language.
Corpus (14). 23–58.
Vanhove, Martine. 2012. Roots and patterns in Beja (Cushitic): The issue of
language contact with Arabic. In Martine Vanhove, Thomas Stolz, Hitmoi
Otsuka & Aina Urdze (eds.), Morphologies in contact, 311–326. Berlin: Akademie
Verlag.
Versteegh, Kees. 2001. Linguistic contact between Arabic and other languages.
Arabica (48). 470–508.
Versteegh, Kees. 2010. Contact and the development of Arabic. In Raymond
Hickey (ed.), The handbook of language contact, 634–651. Oxford: WileyBlackwell.
32
1 Introduction
Vicente, Ángeles. 2006. El proceso de arabización de Alandalús: Un caso medieval de interaccíon de lenguas. Zaragoza: Instituto de Estudios Islámicos y del
Oriente Próximo (IEIOP).
Watson, Janet C. E. 2011. Arabic dialects. In Stefan Weninger, Geoffrey Khan,
Michael P. Streck & Janet C. E. Watson (eds.), The Semitic languages: An international handbook, 851–896. Berlin: De Gruyter Mouton.
Watson, Janet C. E. 2018. South Arabian and Arabic dialects. In Clive Holes (ed.),
Arabic historical dialectology: Linguistic and sociolinguistic approaches, 316–334.
Oxford: Oxford University Press.
Weinreich, Uriel, William Labov & Marvin Herzog. 1968. Empirical foundations
for a theory of language change. In Winfred Lehman & Yakov Malkiel (eds.),
Directions for historical linguistics, 95–195. Austin: University of Texas Press.
Weninger, Stefan. 2011. Aramaic–Arabic language contact. In Stefan Weninger,
Geoffrey Khan, Michael P. Streck & Janet C. E. Watson (eds.), The Semitic languages: An international handbook, 747–755. Berlin: De Gruyter Mouton.
Wilmsen, David. 2010. Dialects of written Arabic: Syntactic differences in the
treatment of object pronouns in Egyptian and Levantine newspapers. Arabica
57. 99–128.
Winford, Donald. 2005. Contact-induced change: Classification and processes.
Diachronica 22(2). 373–427.
Winford, Donald. 2007. Some issues in the study of language contact. Journal of
Language Contact 1(1). 22–39.
Winford, Donald. 2008. Processes of creole formation and related contactinduced language change. Journal of Language Contact 2(1). 124–145.
Winford, Donald. 2010. Contact and borrowing. In Raymond Hickey (ed.), The
handbook of language contact, 170–187. Oxford: Wiley-Blackwell.
Yoda, Sumikazu. 2013. Judeo-Arabic, Libya, Hebrew component in. In Geoffrey
Khan (ed.), Encyclopedia of Hebrew language and linguistics. Leiden: Brill.
Zammit, Martin R. 2011. South Arabian loanwords. In Lutz Edzard & Rudolf de
Jong (eds.), Encyclopedia of Arabic language and linguistics, online edn. Leiden:
Brill.
33
Part I
Contact-induced change in
varieties of Arabic
Chapter 2
Pre-Islamic Arabic
Ahmad Al-Jallad
The Ohio State University
This chapter provides an overview of Arabic in contact in the pre-Islamic period,
from the early first millennium BCE to the rise of Islam. Contact languages include Akkadian, Aramaic, Ancient South Arabian, Canaanite, Dadanitic, and Greek.
The chapter concludes with two case studies on contact-induced development: the
emergence of the definite article and the realization of the feminine ending.
1 Preliminaries
1.1 Language contact in the pre-Islamic period
[I]n the Djāhiliyya, “the Age of Ignorance” […], the Arabs lived to a great extent in almost complete isolation from the outer world… [t]his accounts for
the prima facie astonishing fact that Arabic, though appearing on the stage
of history hundreds of years after the Canaanites and Aramaeans, nevertheless in many respects has a more archaic character than these old Semitic
languages. The Arabs, being almost completely isolated from outer influences and living under the same primitive conditions of their ancestors preserved the archaic structure of their language. (Blau 1981: 18).
This is the image of Arabic’s pre-Islamic past that emerges from Classical Arabic sources. For writers such as Ibn Khaldūn, contact-induced change in Arabic
was a by-product of the Arab conquests, and served to explain the differences
between the colloquial(s) of his time and the literary language. More than a century and a half of epigraphic and archaeological research in Arabia and adjacent
areas has rendered this view of Arabic’s past untenable. Arabic first appears in
the epigraphic record in the early first millennium BCE, and for most of its preIslamic history, the language interacted in diverse ways with a number of related
Ahmad Al-Jallad. 2020. Pre-Islamic Arabic. In Christopher Lucas & Stefano Manfredi
(eds.), Arabic and contact-induced change, 37–55. Berlin: Language Science Press.
DOI:10.5281/zenodo.3744503
Ahmad Al-Jallad
Semitic languages and Greek. This chapter will outline the various foci of contact between Arabic and other languages in the pre-Islamic period based on documentary evidence. Following this, I offer two short case studies showing how
contact-induced change in the pre-Islamic period may explain some of the key
features of Arabic today.
1.2 Old Arabic
Old Arabic is an umbrella term for the diverse forms of the language attested in
documentary and literary sources from the pre-Islamic period, including inscriptions, papyri, and transcriptions in Greek, Latin, and cuneiform texts. The present
usage does not refer to Classical Arabic or the linguistic material attributed to
the pre-Islamic period collected in the eighth and ninth centuries CE, such as
poetry and proverbs, as we cannot be sure about their authenticity, especially
with regard to their linguistic features. Al-Jallad (2017) defines the corpus of Old
Arabic as follows: Safaitic, an Ancient North Arabian script concentrated in the
Syro-Jordanian Ḥarrah (end of the 1st millennium BCE to 4th c. CE), Hismaic, an
Ancient North Arabian script spanning from central Jordan to northwest Arabia (chronology unclear, but overlapping with Nabataean), the substratum of
Nabataean Aramaic, along with a few Arabic-language texts carved in this script
(2nd c. BCE to 4th c. CE), the Nabataeo-Arabic inscriptions (3rd c. CE to 5th c. CE),
pre-Islamic Arabic script inscriptions (5th c. CE to early 7th c. CE) and isolated
inscriptions in the Greek, Dadanitic (the oasis of Dadān, modern-day al-ʕUlā,
northwest Ḥiǧāz), and Ancient South Arabian alphabets (varied chronology).
In geographic terms, Old Arabic is attested mainly in the southern Levant, the
Sinai, and northwestern Arabia, as far south as Ḥegrā (Madāʔin Ṣāleḥ). Within
this area a variety of non-Arabic languages were spoken and written, with which
Old Arabic interacted. The main contact language was Imperial Aramaic, which
served as a literary language across North Arabia in the latter half of the first
millennium BCE until, perhaps, the rise of Islam. Since contact must be viewed
through the lens of writing, it is in most cases difficult to determine how extensive multilingualism was outside of literate circles.
2 Contact languages
2.1 Arabic and Akkadian
The first attestations of Arabic are preserved in cuneiform documents. While
no Arabic texts written in cuneiform have yet been discovered, isolated lexical
38
2 Pre-Islamic Arabic
items survive in this medium. Livingstone (1997) identified an example of the Old
Arabic word for ‘camel’ with the definite article in the inscriptions of Tiglathpileser III (744–727 BCE): a-na-qa-a-te = (h/ʔ )an-nāq-āte ‘the she-camels’. Aside
from this, almost all other Arabic material consists of personal and divine names.
There are reports of “Arabs” in Mesopotamia – inhabiting walled towns in western Babylonia – as early as the eighth century BCE (Eph’al 1974: 112). While
we cannot be sure that the people whom the Babylonians called Arabs were in
fact Arabic speakers, a few texts in dispersed Ancient North Arabian scripts hail
from this region. So far, all seem to contain only personal names with Arabic
or Arabian etymologies.1 These facts can only suggest the possibility of contact
between speakers of Arabic and Akkadian in the early first millennium BCE.
2.2 Arabic and Canaanite
Contact between Arabic speakers and speakers of Canaanite languages is documented in the Hebrew Bible (Eph’al 1982: ch.2; Retsö 2003: ch.8), and there is
one inscription directly attesting to contact between both groups. An Ancient
North Arabian inscription from Bāyir, Jordan contains a prayer in Old Arabic
to three gods of the Iron Age Canaanite kingdoms of Moab, Ammon, and Edom
(Hayajneh et al. 2015). The text is accompanied by a Canaanite inscription, which
remains undeciphered. The reading of the Arabic according to the edition is as
follows:
(1)
Bāyir inscription (Hayajneh et al. 2015)
h mlkm w-kms w-qws b-km
ʕwðn
h-ʔsḥy
voc pn
conj-pn conj-pn prep-2pl.m protect.prf.1pl dem-well.pl
m-mdwst
prep-ruin
‘O Malkom, Kemosh, and Qaws, we place under your protection these
wells against ruin.’
2.3 Arabic and Aramaic
Evidence for contact between Arabic and Aramaic spans from the middle of the
first millennium BCE to the late sixth century CE, and is concentrated in the
southern Levant and northwest Arabia.2 Perhaps one of the earliest examples
1
“Dispersed Ancient North Arabian” is a temporary term given to the Ancient North Arabian inscriptions on seals, pottery, bricks, etc. which have been found in various parts of Mesopotamia
and elsewhere (Macdonald 2000: 33).
2
See Stein (2018) on the role of Aramaic in the Arabian Peninsula in the pre-Islamic period.
39
Ahmad Al-Jallad
of Arabic speakers using Aramaic as a written language comes from the fifthcentury-BCE Nile Delta. A king of Qedar, Qayno son of Gośam,3 commissioned
an Aramaic votive inscription dedicated to hn-ʔlt ‘the goddess’ (Rabinowitz 1956).
Arabic names can be found in transcription across the Levant in Aramaic inscriptions (Israel 1995), and in most cases names with an Arabic etymology terminating in the characteristic final -w, reflecting an original nominative case (Al-Jallad
forthcoming).
Arabic and Aramaic language contact reaches a climax in the written record
at the end of the first millennium BCE with the arrival of inscriptions in the
Nabataean script. The Nabataeans established a kingdom in the region of Edom
in the fourth century BCE, which at its greatest extent spanned from the Ḥawrān
to the northern Ḥiǧāz. While they, like their contemporaries across the Near East,
wrote in a form of Imperial Aramaic, the spoken language of the royal house and
large segments of the population was Arabic. Unlike other examples of Aramaic
written by Arabic speakers so far, Nabataean incorporated Arabic elements into
its writing school, such as the optative use of the perfect, the negator ɣayr, and
a significant number of lexical items relating to daily life (Gzella 2015: 242–243).
Perhaps one of the most interesting examples of contact between the two languages is found in Nabataean legal papyri from the Judaean desert (1st–2nd c. CE).
These Aramaic-language legal documents contain a number of glosses in Arabic, for example: ʕqd /ʕaqd/ ‘contract’; mʕnm /maɣnam/ ‘profit’; prʕ /faraʕ/ ‘to
branch out’; ṣnʕh /ṣanʕah/ ‘handiwork’, etc. (Yardeni 2014). Macdonald (2010: 20)
has suggested, based on this evidence, that Nabataean legal proceedings would
have taken place in Arabic, while all written records were made in Aramaic.
In addition to the use of Arabic within Aramaic, a unique votive inscription
from ʕEn ʕAvdat (Negev, Israel) contains three verses of an Arabic hymn to the
deified Nabataean king ʕObodat embedded within an Aramaic text. While undated (but likely earlier than 150 CE), the text is certainly the earliest example
of continuous Arabic language written in the Nabataean script, as before this
almost all examples are isolated words and personal names.
3
40
The symbol ś denotes the Old Arabic reflex of Classical Arabic 〈 〉ش, which is usually transcribed
š. /ś/ was likely realized as a voiceless lateral fricative [ɬ].
2 Pre-Islamic Arabic
(2)
ʕEn ʕAvdat inscription4
a. Aramaic
dkyr
b-ṭb
q[r]ʔ
qdm ʕbdt ʔlhʔ
remember.ptcp.pass prep-good read.ptcp.act prep pn god.def
grmʔlhy br
mn ktb
w-dkyr
son.cs
conj-remember.ptcp.pass rel write.prf.3sg.m pn
tymʔlhy šlm
lqbl ʕbdt ʔlhʔ
pn
be_secure.prf.3sg.m prep pn god.def
‘May he who reads this aloud be remembered for good before
ʕObodat the god, and may he who wrote be remembered. May
Garmallāhi son of Taymallāhi be secure in the presence of ʕObodat
(the god).’
b. Arabic
p-kn
ʔtrʔ
w-lʔ
lʔ pdʔ
p-ypʕl
conj-act.impf.3sg.m neg ransom.acc conj-neg scar.acc conj-be.inf
lʔ ʔbʕ-h
ʔl-mwtw
hnʔ ybʕ-nʔ
here seek.impf.3sg.m-1pl def-death.nom neg make.obtain.inf-3sg.m
lʔ
grḥw
hnʔ ʔrd
p-kn
conj-be.inf here want.prf.3sg.m wound.nom neg
yrd-nʔ
want.impf.3sg.m-1pl
‘May he act that there be neither ransom nor scar; so be it that death
would seek us, may he not aid its seeking, and so be it that a wound
would desire (a victim), let it not desire us!’
c. Aramaic
yd-h
grmʔlhy ktb
writing.cs hand.3sg.m
pn
‘Garmallāhi, the writing of his hand.’
The presence of Aramaic is much more lightly felt in the desert hinterland to
the east and north of Nabataea. A small handful of Safaitic–Aramaic bilingual
inscriptions are known (Hayajneh 2009: 214–215). In one Safaitic text, produced
by a Nabataean, the author gives his name and affiliation to social groups in a
type of Aramaic, but then writes the remainder of the inscription in Old Arabic,
suggesting that this individual may have been bilingual.
4
This is my translation; the editio princeps is Negev, Naveh & Shaked (1986); it is discussed most
recently in Fiema et al. (2015: 399–402) and Kropp (2017).
41
Ahmad Al-Jallad
(3) Nabataean Safaitic (Al-Jallad 2015: 19; C 2820)
l
ʔʔsd bn
rbʔl bn
ʔʔsd bn
rbʔl nbṭwy
slmy
w
prep pn son.cs pn son.cs pn son.cs pn Nabataean Salamite conj
brḥ
ḫlqt
śty
h-dr
w
depart.prf.3sg.m period.cs winter def-region conj
h-smy
tð̣r
keep_watch.prf.3sg.m def-sky
‘By ʔAʔsad son of Rabbʔel son of ʔAʔsad son of Rabbʔel, the Nabataean
Salamite, and he set off from this place for the period of winter and kept
watch for the rains.’
A handful of Aramaic loans are found in the Safaitic inscriptions: sfr ‘writing’;
ʔsyt ‘hide, trap’; lṣṭ ‘thief’, ultimately from Greek lēistḗs. Other words, such as
mdbr /madbar/ ‘the Hamad, wilderness’ and nḫl /naḫl/ ‘valley’, are absent in
Classical Arabic yet appear in the Northwest Semitic languages. These do not
appear to be loans, however, as their meanings and phonologies are local and
Arabic, respectively. They should instead be regarded as genuine cognates that
did not make it into the Islamic-period lexica.
2.4 Provincia Arabia and the Nabataeo-Arabic script
In 106 CE, under circumstances that remain poorly understood, the Romans annexed the Nabataean Kingdom and established their Province of Arabia. While
Nabataean political independence ended, their script, writing tradition and language continued to thrive and evolve. This is exemplified by the famous tomb
inscription of Raqōś bint ʕAbd-Manōto from Madāʔin Ṣāliḥ. Dated to 267 CE,
the text is a legal inscription associated with the grave of a woman who died
in al-Ḥegr. Unlike other grave inscriptions at this site, the Raqōś inscription is
composed almost entirely in Arabic, with the Aramaic components restricted to
the introductory demonstrative dnh ‘this’, the words for ‘son’ and ‘daughter’, the
dating formula, and the name of the deity. The Aramaic components are bolded
below:
(4)
5
42
Madāʔin Ṣāliḥ inscription (JSNab 17)5
kʕbw br ḥrtt l-rqwš brt
dnh qbrw ṣnʕ-h
dem grave build.prf.3sg.m-3sg.m pn son pn prep-pn daughter
fy ʔl-ḥgrwy šnt
hlkt
w-h
ʕbdmnwtw ʔm-h
mother-3sg.m conj-3sg.f die.prf.3sg.f prep def-pn year
pn
For the latest discussion of this text, see Macdonald’s contribution to Fiema et al. (2015: 402–
405).
2 Pre-Islamic Arabic
mʔh
w-štyn
w-tryn b-yrḥ
tmwz
hundred conj-sixty conj-two prep-month Tammūz
mryʕlmʔ mn yšnʔ
ʔl-qbrw d[ʔ]
w-lʕn
conj-curse.prf.3sg.m pn
rel desecrate.impf.3sg.m def-grave dem
w-mn
yftḥ-h
ḥšy w wld-h
conj-rel open.impf.3sg.m-3sg.m except w children-3sg.m
w-lʕn
mn yqbr
w-yʕly
conj-curse.prf.3sg.m rel bury.impf.3sg.m conj-remove.impf.3sg.m
mn-h
prep-3sg.m
‘This is a grave that Kaʕbo son of Ḥāreθat constructed for Raqōś
daughter of ʕAbd-Manōto, his mother, and she perished in al-Ḥegro year
one hundred and sixty two in the month of Tammūz. May the Lord of
the World curse anyone who desecrates this grave and anyone who
would open it, with the exception of his children, and may he curse
anyone who would bury or remove from it (a body).’
During the same period, the classical Nabataean script continues to evolve towards what we consider the Arabic script (Nehmé 2010). Its letter forms take on
a more cursive character, and the connecting element of each letter goes across
the bottom of the text. Nehmé considers the letter forms typical of the Arabic
script to have evolved from Nabataean between the third and fifth centuries CE.
In inscriptions from this period, the Arabic component begins to increase at the
expense of Aramaic (Nehmé 2017). This trend may suggest that knowledge of Aramaic was waning in these centuries, or that the writing tradition itself was transforming – Aramaic was slowly being replaced by Arabic. If we think in terms of
writing schools, there may not have been much Arabic–Aramaic bilingualism in
Arabia outside of the scribal class – indeed, scholars have continued to debate
whether Nabataean Aramaic was ever a colloquial, and there are good arguments
to doubt that it was (Gzella 2015: 240). The remnants of Aramaic in the latest
phases of the Nabataeo-Arabic inscriptions, however, most certainly functioned
as a code, grams for Arabic words, a situation comparable to the Aramaeograms
of Pahlavi (cf. Nyberg 1974).
2.5 The Arabic inscriptions of the sixth century CE
In Arabic inscriptions of the sixth century, written Arabic and Aramaic continue
the stable situation of contact witnessed in the Nabataeo-Arabic period. Aramaic
fossils are employed in dating formulae and the word for ‘son’, and possibly the
43
Ahmad Al-Jallad
first person pronoun. But otherwise, the language of these texts is entirely Arabic.
Perhaps the most famous among these is the inscription of Jebel Usays, given in
(5), in which the Aramaic components are bolded.
(5) Jebel Usays inscription6
ʔnh7 rqym br mʕrf ʔl-ʔwsy
ʔrsl-ny
ʔlḥrt ʔl-mlk ʕly
1sg pn son pn def-Awsite send.prf.3sg.m-1sg pn def-king prep
ʔsys mslḥh snt 423
Usays outpost year 423
‘I, Ruqaym son of Muʕarrif the Awsite, al-Ḥāriθ the king sent me to
Usays as an outpost, year 423.’ [= 528/9 CE]
2.6 Arabic, Greek and Aramaic in sixth-century Petra
In 1993, a corpus of carbonized Greek papyri – some 140 rolls – was discovered at
the Byzantine church of Petra.8 These documents attest to a trilingual situation
at the city: Greek served as the official administrative language, while Arabic and
Aramaic appear to have been spoken languages. The microtoponyms (names of
small plots of lands and vineyards) are in both Arabic and Aramaic, and oftentimes the same word is expressed in both languages, as in Table 1.
Table 1: Arabic–Aramaic equivalents in the Petra Papyri (Al-Jallad
2018a: 41)
Translation
Arabic
Aramaic
‘land markers’
‘farm’
‘canal’
Αραμ /ārām/
αλ-Ναϲβα /al-naṣbah/
αλ-Κεϲεβ /al-qeṣeb/
Εραμαεια /eramayyā/
Ναϲβαθα /naṣbatā/
Κιϲβα/Κειϲβα /qiṣbā/
This naturally suggests that, alongside literacy in Greek, there was spoken
bilingualism in Arabic and Aramaic, perhaps a stable situation extending back
to Nabataean times.
6
For the latest discussion of this text, see Macdonald’s contribution to Fiema et al. (2015: 405).
While it has been suggested that the spelling ʔnh reflects a pausal form (Larcher 2010), it seems
more likely in light of the Thaʕlabah Nabataeo-Arabic inscription (Avner et al. 2013), which
spells ‘I’ as ʔnh, that this form reflects the Aramaic spelling of the pronoun rather than an
Arabic variant.
8
These papyri are edited in a five-volume series: the Petra Papyri I–V (2002–2018), various
editors, Amman: American Center of Oriental Research. See Arjava et al. (2018) for the last
volume.
7
44
2 Pre-Islamic Arabic
2.7 Arabic and Ancient South Arabian
Classical Arabic sources note a situation of close contact between Arabic and
“Ḥimyaritic”, a term used for a language they associated with the pre-Islamic
kingdom of Ḥimyar in what is today Yemen. The pre-Islamic inscriptions from
the northern Yemeni Jawf, the so-called Haram region, attest to a similar situation. These texts are composed in Sabaic, but contain a significant admixture of
non-Sabaic linguistic material. Some scholars (e.g. Robin 1991) have considered
Arabic to be the contributing source, but in most cases the non-Sabaic linguistic
features are not specific to Arabic, such as the use of the causative verb ʔaCCaC,
which is attested in Aramaic and Gəʕəz for example, rather than haCCaC as in
Sabaic. As Macdonald (2000: 55) rightly puts it, these inscriptions are basically
Sabaic, with a small admixture from North Arabian languages, but not necessarily Arabic. Four texts from this region, however, exhibit the Arabic isogloss
of lam for past-tense negation, suggesting that some form of Arabic may have
contributed to their mixed character.9
Mixed North/South Arabian texts can be found further to the north, in Naǧrān
and Qaryat al-Fāw. The most famous is perhaps the grave inscription of Rbbl
bn Hfʕm. This unique text attests features that can be attributed to both nonSabaic and Sabaic sources. On the non-Sabaic side, it uses the definite article
ʔl, the causative morpheme ʔ- rather than h-, and occasionally the 3rd person
pronoun h rather than hw. At the same time, the text employs mimation, clitic
pronouns with long vowels, e.g. -hw, and prepositions not known in Arabic (AlJallad 2018b: 30). At Naǧrān, one occasionally encounters Arabic lexical items,
such as ldy ‘at’ and ʕnd ‘with’ in otherwise perfectly good South Arabian texts.
So then, how are we to interpret the mixed character of these texts? For Qaryat
al-Fāw, Durand (2017: 95, fn.32) has suggested, based on the significant amount of
Petraean pottery, that a sizable Nabataean colony existed at the oasis. It could be
the case that Nabataean colonists introduced Arabic to the oasis, where it naturally gained prestige as a trade language given its links with the north. The mixed
nature of some of the inscriptions of this site could therefore be interpreted in
two ways. If they reflect a spoken variety, then perhaps they are the result of convergence between the Arabic introduced by the Nabataeans and Sabaic, similar to
the modern dialects of Yemeni Arabic today, which are essentially Arabic with a
significant South Arabian admixture.10 If we are dealing with an artificial scribal
register, then the language may be the result of a scribe attempting to produce
9
For a list of the Haram inscriptions, see Macdonald (2000: 61), who labels these texts SabaeoNorth-Arabian.
10
On these varieties, see Watson (2018).
45
Ahmad Al-Jallad
a text in Arabic, for an Arabic-speaking customer, but inadvertently introducing
Sabaicisms from the language he is more used to writing. A similar phenomenon
might be at play in the Aramaic–Hasaitic tomb inscription from Mleiha.11 There,
the scribe – seemingly unintentionally – uses the Aramaic word for son, br, in
the Hasaitic portion of the text, suggesting perhaps that he was bilingual and
more used to writing in Aramaic (Overlaet et al. 2016).
2.8 Arabic in the Ḥiǧāz
Before the arrival of the Nabataeans, the written language of the oasis of al-ʕUlā
and associated environs in the northern Ḥiǧāz was Dadanitic, a non-Arabic Central Semitic language. A few texts, however, display features that are unambiguously Arabic. The best known of these is JSLih 384. This short text is written in
the Dadanitic script but seems to be, in other respects, produced in a dialect of
Old Arabic, notably making use of the relative pronoun ʔlt /ʔallatī/. Two other
Dadanitic texts make use of the Arabic construction ʔn yfʕl, that is, the use of the
subordinator ʔan with a modal verb. In addition to this, one occasionally finds the
ʔ(l) definite article employed in these inscriptions. The interpretation of this contact situation, like that in South Arabia, is unclear. Do these few texts represent
the writings of travelers or immigrants from the north, whose spoken language
influenced the dictation of text to the scribe? Or do they reflect unique points
on a dialect continuum? The complex linguistic situation at ancient Dadan is the
subject of a fascinating study by Kootstra (2019).
2.9 Arabic and the languages of the Thamudic inscriptions
Even more difficult to distill is the possible contact situation between Arabic and
the more shadowy pre-Arabic Semitic languages of north and central Arabia. We
are afforded a small glimpse of these languages by the laconic Thamudic inscriptions, mainly those classified in the C, D, and F scripts.12 While it is difficult to
say much about the languages these scripts express, they are clearly distinct from
Arabic (Al-Jallad 2017: 321–322). The only evidence for contact between Arabic
and any of these languages is found in the tomb inscription of Raqōś at Madāʔin
Ṣāliḥ, illustrated in (4). This text, as discussed in §2.4, is written mainly in Arabic, with a few fossilized Aramaic components. Alongside the main inscription,
there is a short text inscribed in the Thamudic D script stating: ʔn rqś bnt ʕbdmnt
‘This is Raqōś, daughter of ʕAbdo-Manōto’. The use of the introductory element
11
12
46
Hasaitic is the name given to the pre-Islamic script and language of East Arabia.
Thamudic B, C, and D are discussed in Macdonald (2000) and Al-Jallad (2017; 2018b); Thamudic
F is outlined in Prioletta & Robin (2018).
2 Pre-Islamic Arabic
ʔn ‘this’ or perhaps ‘for’, rather than the Arabic demonstrative dʔ /ðā/ or perhaps its feminine equivalent dy /ðī/, employed in the Nabataean text, indicates
that we are dealing with a third language.13 Did Raqōś originally hail from a nomadic community who spoke a non-Arabic Semitic language expressed in the
Thamudic D script? And did she later come to live in Arabic-speaking Ḥegrā?
Was the use of this script on her grave a tribute to her heritage? These questions
are impossible to answer with the data available to us now, but they widen the
scope of investigation when examining Arabic’s history. The available fragments
of evidence support the suggestion put forth recently by Souag (2018): we must
consider the possibility of unknown Semitic substrate(s) in the development of
early Arabic.
2.10 Arabic and Greek
The nexus of Arabic–Greek contact, based on the inscriptions known so far, is
the Syro-Jordanian Harrah, the basalt desert that stretches from the Hawrān to
northern Arabia. Greek inscriptions are occasionally found throughout this region, interacting with the local Arabic dialects in diverse ways. The commonest
type of bilingual text consists of simple signatures in Safaitic and Greek. These
texts, illustrated in (6), only prove that the author knew how to write his name
in Greek, and do not constitute evidence for genuine bilingualism.
(6) Graeco-Arabic inscription A1 (Al-Jallad & al-Manaser 2016: 56)
a. Greek
Θαιμος Γαφαλου
Taym Gaḥfal
‘Taym, son of Gaḥfal’
b. Arabic
bn gḥfl
l-tm
prep-Taym son Gaḥfal
‘for/by Taym, son of Gaḥfal’
The second inscription discussed by Al-Jallad & al-Manaser (2016), illustrated
in (7), provides more insight into the different degrees of Arabic–Greek bilingualism. The author carves a short text in both Greek and Old Arabic, indicating that
he knew both languages but that his command of Arabic was obviously better.
13
While it is tempting to interpret ʔn as the first-person singular pronoun ʔanā, such a formula would indeed be strange in a grave epitaph. Perhaps ʔn is cognate with the demonstrative/presentative element *han, or perhaps it should be construed as a dative ‘to, for’ cognate
with East Semitic ana.
47
Ahmad Al-Jallad
(7)
Graeco-Arabic inscription 2 (Al-Jallad & al-Manaser 2016: 58)
a. Arabic
l-ɣθ
w tḥll
ʔfwh ʕql
sr
prep-Ghawth conj go.prf.3sg.m prep protected_area Sayr
‘By Ghawth and he went into the protected area of Sayr.’
b. Greek
Γαυτος
ἀπῆλθεν
[ε]ἰς τόν
Ακελον
Σαιρου
Ghawth.nom depart.aor.3sg prep def.m.acc.sg ʕaql.acc.sg Sayr.gen
‘Ghawth, he went away to the ʕaql of Sayr.’
The author translates the Arabic into Greek effectively, but seems not to have
known the Greek word for the culturally specific term ʕaql, ‘a protected area of
pasturage’. In this case, he simply wrote the word out in Greek: Ακελον.
There is evidence that some nomadic Arabic speakers did master the Greek
language, as one sometimes comes across very well-composed texts in Greek, attesting to full-scale bilingualism, at least in writing (for example A2 in Al-Jallad
& al-Manaser (2015). This level of bilingualism, however, must have been rare.
There is no appreciable influence from Greek on the Arabic of the Safaitic inscriptions. A few loanwords are known, e.g. qṣr ‘Caesar’, lṣṭ ‘thief’, but these
more likely come through Aramaic.
2.11 Arabic in eastern Arabia
The inscriptional record of eastern Arabia is relatively poor when compared to
the western two-thirds of the Peninsula. Nevertheless, the extant texts point towards contact between Aramaic and the local Arabian language, called Hasaitic
by scholars. This language, however, cannot be regarded as a form of Arabic, and
there are no pre-Islamic attestations of Arabic from eastern Arabia yet (Al-Jallad
2018b: 260–261).
3 Grammatical features arising from contact
This section offers a contact-based explanation for two linguistic features found
in Old Arabic: the definite article, and the realization of the feminine ending.
3.1 Definite article
It has long been established that the overt marking of definiteness in the Semitic
languages is a relatively late innovation (Huehnergard & Rubin 2011: 260–261).
48
2 Pre-Islamic Arabic
All varieties of Arabic today attest some form of the definite article – most commonly variants of ʔal but other forms exist as well, mainly in southwest Arabia,
including am, an, and a-, with gemination of the following consonant. In light
of the comparative evidence, did Arabic innovate this feature independently or
was contact with other Semitic languages involved?
The evidence suggests that the prefixed article *han- emerged in the central
Levant sometime in the late second millennium BCE, after the diversification of
Northwest Semitic (Tropper 2001; Gzella 2006; Pat-El 2006). It seems clear that by
the early first millennium BCE, the article had spread across the southern Levant
and to North Arabia, as it is found in Taymanitic, Thamudic B, and Dadanitic, as
well as in the Old Arabic of the Safaitic inscriptions. In the latter case, contact
with Canaanite is substantiated in the inscriptional record in the form of the
Bāyir inscription (see §2.2 above).
All of these languages, including the earliest Old Arabic, took over the form
of the article unchanged; that is h- with the assimilation of the /n/ before a consonant, the exception being Dadanitic, which preserves the /n/ before laryngeal
consonants, e.g. h-mlk /ham-malk/ ‘the king’ vs. hn-ʔʕly ‘the upper’ /han-ʔaʕlay/.
We cannot, however, argue for the spread of the definite article to Proto-Arabic.
The original, article-less situation is attested in the inscriptions of Central Jordan
stretching down to the Hismā, known as Hismaic (Graf & Zwettler 2004). These
texts are in unambiguously Arabic language, but they lack the definite article.
The h-morpheme exists, but it has a strong demonstrative force. Indeed, in a few
Nabataean–Hismaic bilingual inscriptions, the definite article ʔl of the Nabataean
component is rendered as zero in the Hismaic text (Hayajneh 2009). A minority
of Safaitic inscriptions also lack the definite article (Al-Jallad 2018b), showing
that it had not spread to all varieties of Arabic even as late as the turn of the Era.
Thus, like Hebrew and Aramaic, the earliest linguistic stages of Arabic – and indeed Proto-Arabic – lacked a fully grammaticalized definite article. Contact with
Canaanite then seems to be the likeliest explanation for the appearance of the
h-article in Old Arabic.
While the h- article is the commonest form in Old Arabic, whence the ʔal form?
The ʔal article appears to be a later development from the original han article,
through two irregular sound changes: h > ʔ and n > l.14 The former is well attested in Arabic (e.g. the causative ʔaCCaCa from haCCaCa), while the latter is
not uncommon in loans (e.g. finǧān vs. finǧāl ‘cup’). The ʔal article appears to
have developed in the western dialects of Old Arabic, attested first in the Nile
Delta (cf. the famous αλιλατ al-ʔilat ‘the goddess’ mentioned in Herodotus, Histories I: 131), and is the regular form of the article in the dialect of the Nabataeans,
14
The origins of the al-article are discussed in detail in Al-Jallad (forthcoming).
49
Ahmad Al-Jallad
who were situated in ancient Edom, stretching south to the Ḥiǧāz. The ʔal-article
is attested sporadically at Dadān in the western Ḥiǧāz as well. Based on the inscriptional record, the ʔal-article was a typical linguistic feature of settled, rather
than nomadic groups, being attested most frequently in the Nabataean dialect,
and in cities and oases like Petra and Ḥegrā. The nomads used a variety of definite article forms. It was perhaps not until the rise of Islam, and the resulting
prestige given to official Arabic of the Umayyad state, that the ʔal article began
to dominate at the expense of other forms.
3.2 The feminine ending
In most modern Arabic dialects, the feminine ending *-at is realized as -a(h) in
all contexts except the construct state, where it retains its original form -at. In
Classical Arabic, it is -at in all situations, except for in utterance-final position,
where it is realized as -ah. The Quranic Consonantal Text resembles the situation
in the modern dialects, as do the transitional Nabataeo-Arabic and sixth-century
Arabic script inscriptions (Nehmé 2017). Yet, if we go back further to the first
century CE, it seems that varieties of Arabic written in the Hismaic and Safaitic
script never experienced the sound change -at > -ah in any position – the feminine ending is always written as 〈t〉. In the Arabic of the Nabataeans, however,
the sound change of -at to -ah seems to have operated as early as the third century BCE (Al-Jallad 2017: §5.2.1).
The sound change -at > -ah is common in the Central Semitic languages, but
the distribution can vary. In Phoenician, it applies to verbs but not nouns, while
in Hebrew it applies equally to nouns and verbs (Huehnergard & Rubin 2011:
265–266). The most common Arabic distribution matches Aramaic: it applies to
nouns but not verbs. I would suggest that, since this sound change is first attested in a dialect of Arabic for which we have abundant evidence of heavy contact with Aramaic, it is likely a contact-induced change (see also van Putten, this
volume). Contact, or the lack thereof, may explain its absence in the ancient nomadic dialects, where, as we have seen above, there is little evidence for contact
with Aramaic. Thus, like the ʔal article, the -at to -ah change would have been
a typical feature of Arabic dialects of settled groups in the pre-Islamic period. In
later forms of Arabic, the change spreads even to nomadic dialects, as we find it
operational today across the Arabian Peninsula. Yet, the chronology of this diffusion is not quite clear. In an important study by van Putten (2017), the Dosiri
dialect of Kuwait appears to preserve the archaic situation where the feminine
ending is realized as -at in all positions.
50
2 Pre-Islamic Arabic
4 Conclusion
Contact must be factored into our understanding of language change for Arabic
at every attested stage. A summary of the facts above show that Arabic was in
most intense contact with Aramaic, a situation that persisted for over a millennium prior to the rise of Islam, which may explain the high number of Aramaic
loanwords into Arabic, and indeed some striking structural parallels, such as the
distribution of the sound change -at > -ah. At the same time, there is very little
evidence for contact with Sabaic (Old South Arabian), a contact situation only
represented by a small number of mixed texts. This nicely matches the absence
of South Arabian influence on Old Arabic and later forms of the language, with
the exception of those dialects spoken in southwest Arabia.
Further reading
) Al-Jallad (2018b) provides a comprehensive outline of the languages and scripts
of pre-Islamic North Arabia.
) Macdonald (2003) gives a description of the multilingual environment of ancient Nabataea.
) Nehmé (2010) outlines the development of the Arabic script based on the newest
Nabataeo-Arabic inscriptions from Northwest Arabia.
) Stein (2018) gives an outline of the use of Aramaic in pre-Islamic Arabia.
Abbreviations
1, 2, 3
acc
act
aor
BCE
c.
CE
conj
def
dem
f
impf
1st, 2nd, 3rd person
accusative
active
aorist
before Common Era
century
Common Era
conjunction
definite
demonstrative
feminine
imperfect (prefix conjugation)
inf
m
nom
pass
pl
pn
prep
prf
ptcp
rel
sg
voc
infinitive
masculine
nominative
passive
plural
proper noun
preposition
perfect (suffix conjugation)
participle
relative
singular
vocative
51
Ahmad Al-Jallad
References
Al-Jallad, Ahmad. 2015. An outline of the grammar of Safaitic inscriptions. Leiden:
Brill.
Al-Jallad, Ahmad. 2017. The earliest stages of Arabic and its linguistic classification. In Elabbas Benmamoun & Reem Basiouney (eds.), The Routledge handbook
of Arabic linguistics, 315–331. London: Routledge.
Al-Jallad, Ahmad. 2018a. The Arabic of Petra. In Antti Arjava, Jaakko Frösén &
Jorma Kaimio (eds.), The Petra papyri V. Amman: American Center of Oriental
Research.
Al-Jallad, Ahmad. 2018b. What is Ancient North Arabian? In Daniel Birnstiel
(ed.), Re-engaging comparative Semitic and Arabic studies, 1–44. Wiesbaden:
Harrassowitz.
Al-Jallad, Ahmad. Forthcoming. One wāw to rule them all: The origins and fate
of wawation in Arabic. In Fred M. Donner & Rebecca Hasselbach-Andee (eds.),
Scripts and scripture. Chicago: Oriental Institute Publications.
Al-Jallad, Ahmad & Ali al-Manaser. 2015. New epigraphica from Jordan I: A preIslamic Arabic inscription in Greek letters and a Greek inscription from northeastern Jordan. Arabian Epigraphic Notes 1. 57–70.
Al-Jallad, Ahmad & Ali al-Manaser. 2016. New epigraphica from Jordan II: Three
Safaitic–Greek partial bilingual inscriptions. Arabian Epigraphic Notes 2. 55–
66.
Arjava, Antti, Jaakko Frösén & Jorma Kaimio (eds.). 2018. The Petra papyri V.
Amman: American Center of Oriental Research.
Avner, Uzi, Laïla Nehmé & Christian Robin. 2013. A rock inscription mentioning Thaʿlaba, an Arab king from Ghassān. Arabian Archaeology and Epigraphy
24(2). 237–256.
Blau, Joshua. 1981. The emergence and linguistic background of Judaeo-Arabic: A
study of the origins of Neo-Arabic and Middle Arabic. 3rd edn. Jerusalem: BenZvi Institute.
Durand, Caroline. 2017. Banqueter pour mieux régner? A propos de quelques
assemblages céramiques provenant de Pétra et du territoire Nabateen. Syria
74. 85–98.
Eph’al, Israel. 1974. “Arabs” in Babylonia in the 8th century B.C. Journal of the
American Oriental Society 94(1). 108–115.
Eph’al, Israel. 1982. The ancient Arabs: Nomads on the borders of the Fertile
Crescent, 9th–5th centuries B.C. Jerusalem: Magnes Press.
52
2 Pre-Islamic Arabic
Fiema, Zbigniew T., Ahmad Al-Jallad, Michael C. A. Macdonald & Laïla Nehmé.
2015. Provincia Arabia: Nabataea, the emergence of Arabic as a written
Language, and Graeco-Arabica. In Greg Fisher (ed.), Arabs and empires before
Islam, 373–433. Oxford: OUP.
Graf, David F. & Michael J. Zwettler. 2004. The North Arabian “Thamudic E”
inscription from Uraynibah West. Bulletin of the American Schools of Oriental
Research 335. 53–89.
Gzella, Holger. 2006. Die Entstehung des Artikels im Semitischen: Eine
“phönizische” Perspektive. Journal of Semitic Studies 51(1). 3–18.
Gzella, Holger. 2015. A cultural history of Aramaic: From the beginnings to the
advent of Islam. Leiden: Brill.
Hayajneh, Hani. 2009. Ancient North Arabian–Nabataean bilingual inscriptions
from southern Jordan. Proceedings of the Seminar for Arabian Studies 39. 203–
222.
Hayajneh, Hani, Mohammad Ababneh & Fawwaz Khraysheh. 2015. Die Götter
von Ammon, Moab und Edom in einer neuen frühnordarabischen Inschrift
aus Südost-Jordanien. In Viktor Golinets, Hanna Jenni, Hans-Peter Mathys &
Samuel Sarasin (eds.), Neue Beiträge zur Semitistik: Fünftes Treffen der Arbeitsgemeinschaft Semitistik in der Deutschen Morgenländischen Gesellschaft vom
15.–17. Februar 2012 an der Universität Basel, 79–105. Münster: Ugarit-Verlag.
Huehnergard, John & Aaron D. Rubin. 2011. Modern Standard Arabic. In Stefan Weninger, Geoffrey Khan, Michael P. Streck & Janet C. E. Watson (eds.),
The Semitic languages: An international handbook, 259–278. Berlin: De Gruyter
Mouton.
Israel, Felice. 1995. L’onomastique arabe dans les inscriptions de Syrie et de
Palestine. In Hélène Lozachmeur (ed.), Présence arabe dans le croissant fertile
avant l’Hégire, 47–57. Paris: Editions Recherche sur les civilizations.
Kootstra, Fokelien. 2019. The writing culture of ancient Dadān: A description
and quantitative analysis of linguistic variation. Leiden: Leiden University.
(Doctoral dissertation).
Kropp, Manfred. 2017. The ʿAyn ʿAbada inscription thirty years later: A reassessment. In Ahmad Al-Jallad (ed.), Arabic in context, 53–74. Leiden: Brill.
Larcher, Pierre. 2010. In search of a standard: Dialect variation and new Arabic
features in the oldest Arabic written documents. In Michael C. A. Macdonald
(ed.), The development of Arabic as a written language, 103–112. Oxford:
Archaeopress.
Livingstone, Alasdair. 1997. An early attestation of the Arabic definite article.
Journal of Semitic Studies 42(2). 259–262.
53
Ahmad Al-Jallad
Macdonald, Michael C. A. 2000. Reflections on the linguistic map of pre-Islamic
Arabia. Arabian Archaeology and Epigraphy 11. 28–79.
Macdonald, Michael C. A. 2003. Languages, scripts, and the uses of writing
among the Nabataeans. In Glenn Markoe (ed.), Petra rediscovered: The lost city
of the Nabataeans, 37–56. New York: Harry N. Abrams/The Cincinnati Art
Museum.
Macdonald, Michael C. A. 2010. Ancient Arabia and the written word. In Michael
C. A. Macdonald (ed.), The development of Arabic as a written language, 5–27.
Oxford: Archaeopress.
Negev, Avraham, J. Naveh & S. Shaked. 1986. Obodas the god. Israel Exploration
Journal 36. 56–60.
Nehmé, Laïla. 2010. A glimpse of the development of the Nabataean script into
Arabic based on old and new epigraphic material. In Michael C. A. Macdonald
(ed.), The development of Arabic as a written language, 47–88. Oxford: Archaeopress.
Nehmé, Laïla. 2017. Aramaic or Arabic? the Nabataeo-Arabic script and the language of the inscriptions written in this script. In Ahmad Al-Jallad (ed.), Arabic
in context, 75–89. Leiden: Brill.
Nyberg, Henrik Samuel. 1974. A manual of Pahlavi. Vol. 2: Ideograms, glossary,
abbreviations, index, grammatical survey, corrigenda to Part I. Wiesbaden:
Harrassowitz.
Overlaet, Bruno, Michael C. A. Macdonald & Peter Stein. 2016. An Aramaic−Hasaitic bilingual inscription from a monumental tomb at Mleiha, Sharjah,
UAE. Arabian Archaeology and Epigraphy 27(1). 127–142.
Pat-El, Na’ama. 2006. The development of the Semitic definite article: A syntactic
approach. Journal of Semitic Studies 51(1). 19–50.
Prioletta, Alessia & Christian Julien Robin. 2018. New research on the ‘Thamudic’
graffiti from the region of Ḥimā (Najrān, Saudi Arabia). In Michael C. A. Macdonald (ed.), Languages, scripts and their uses in ancient North Arabia: Papers
from the special session of the Seminar for Arabian Studies held on 5 August 2017,
53–70. Oxford: Archaeopress.
Rabinowitz, Isaac. 1956. Aramaic inscriptions of the fifth century B. C. E. from a
North-Arab shrine in Egypt. Journal of Near Eastern Studies 15(1). 1–9.
Retsö, Jan. 2003. The Arabs in antiquity: Their history from the Assyrians to the
Umayyads. London: Routledge.
Robin, Christian-Julien. 1991. La pénétration des arabes nomades au Yémen.
Revue des mondes musulmans et de la Méditerranée 61(1). 71–88.
Souag, Lameen. 2018. Yaqṭīn as substratum vocabulary. http://lughat.blogspot.
com/2018/06/yaqtin-as-substratum-vocabulary.html. Accessed 19/9/2019.
54
2 Pre-Islamic Arabic
Stein, Peter. 2018. The role of Aramaic on the Arabian Peninsula in the second
half of the first millennium BC. In Michael C. A. Macdonald (ed.), Languages,
scripts and their uses in ancient North Arabia, 39–49. Oxford: Archaeopress.
Tropper, Josef. 2001. Die Herausbildung des bestimmten Artikels im Semitischen.
Journal of Semitic Studies 46(1). 1–31.
van Putten, Marijn. 2017. The archaic feminine ending -at in Shammari Arabic.
Journal of Semitic Studies 62(2). 357–369.
Watson, Janet C. E. 2018. South Arabian and Arabic dialects. In Clive Holes (ed.),
Arabic historical dialectology: Linguistic and sociolinguistic approaches, 316–334.
Oxford: Oxford University Press.
Yardeni, Ada. 2014. A list of the Arabic words appearing in Nabataean and Aramaic legal documents from the Judaean Desert. Scripta Classica Israelica 33.
301–324.
55
Chapter 3
Classical and Modern Standard Arabic
Marijn van Putten
University of Leiden
The highly archaic Classical Arabic language and its modern iteration Modern
Standard Arabic must to a large extent be seen as highly artificial archaizing registers that are the High variety of a diglossic situation. The contact phenomena
found in Classical Arabic and Modern Standard Arabic are therefore often the result of imposition. Cases of borrowing are significantly rarer, and mainly found in
the lexical sphere of the language.
1 Current state and historical development
Classical Arabic (CA) is the highly archaic variety of Arabic that, after its codification by the Arab Grammarians around the beginning of the ninth century,
becomes the most dominant written register of Arabic. While forms of Middle
Arabic, a style somewhat intermediate between CA and spoken dialects, gain
some traction in the Middle Ages, CA remains the most important written register for official, religious and scientific purposes.
From the moment of CA’s rise to dominance as a written language, the whole
of the Arabic-speaking world can be thought of as having transitioned into a state
of diglossia (Ferguson 1959; 1996), where CA takes up the High register and the
spoken dialects the Low register.1 Representation in writing of these spoken dialects is (almost) completely absent in the written record for much of the Middle
Ages. Eventually, CA came to be largely replaced for administrative purposes by
Ottoman Turkish, and at the beginning of the nineteenth century, it was functionally limited to religious domains (Glaß 2011: 836). During the nineteenth-century
1
Diglossic situations are often seen as consisting of a high register (often called H) and a low
register (L). These two are seen to be in complementary distribution, where each register is
used in designated environments, where the H register takes up such domains like formal
speeches and writing, while the L register is used in personal conversation, oral literature etc.
Marijn van Putten. 2020. Classical and Modern Standard Arabic. In Christopher Lucas
& Stefano Manfredi (eds.), Arabic and contact-induced change, 57–82. Berlin: Language
Science Press. DOI:10.5281/zenodo.3744505
Marijn van Putten
Arabic literary revival known as the Nahḍa, CA goes through a rather amorphous and decentralized phase of modernization, introducing many neologisms
for modern technologies and concepts, and many new syntagms became part of
modern writing, often calqued upon European languages. After this period, it is
customary in scholarly circles to speak of CA having transitioned into Modern
Standard Arabic (MSA), despite the insistence of its authors that CA and MSA
are one and the same language: al-ʕarabiyya l-fuṣḥā ‘the most eloquent Arabic
language’ (Ryding 2011: 845).
2 Contact languages
Considering the significant time-depth of CA and MSA, contact languages have
of course changed over time. Important sources of linguistic contact of the preIslamic varieties of Arabic that come to form the vocabulary for CA are Aramaic,
Greek and Ethio-Semitic. While there are already some Persian loanwords in the
very first sources of CA, this influence continues well into the Classical period,
and ends up having a marked effect on CA and MSA alike.
2.1 Aramaic
Aramaic becomes the dominant lingua franca in much of the Achaemenid empire,
and both written and spoken varieties of Aramaic continue to play an essential
role all throughout Arabia, Syria and Mesopotamia right up until the dawn of Islam. As such, a not insignificant amount of vocabulary has been borrowed from
Aramaic into Arabic, which shows up in CA. Moreover, Aramaic was an important language of Christianity and Judaism, and a noticeable amount of religious
vocabulary from Aramaic has entered CA (§3.4.2). There may even be some structural influence on the phonology of pre-Classical Arabic that has made it into CA
(§3.1).
2.2 Greek
Greek was the language of state of the Byzantine Empire, which, when not directly ruling over Arabic-speaking populations, was at least in close contact with
them. This can be seen in the significant amount of Greek vocabulary that can
be detected in CA. Aramaic, however, has often borrowed the same terms that
we find in CA, and it is usually difficult, if not impossible, to decide whether a
Greek word entered Arabic directly from Greek or through the intermediary of
Aramaic (§3.4.3).
58
3 Classical and Modern Standard Arabic
2.3 Persian
After the rise of Islam, Greek and Aramaic quickly lose the central role they once
played in the region, and they do not continue to influence CA significantly in
the Islamic period. Persian, however, of which a number of words can already be
detected in the Quran, continues to have a pronounced influence on Arabic, and
many more Persian words enter CA throughout its history (§3.4.5).
2.4 Ethio-Semitic and Old South Arabian
It is widely recognized that some degree of influence from Ethio-Semitic can be
identified within CA (§3.2.3; 3.4.1). Many of the Ethio-Semitic words that have
entered into Quranic Arabic presumably arrived there through South Arabian
contact after the invasion of Yemen by Christian Ethiopia in the sixth century.
Also previous South Arabian contact must probably be assumed, and the divine
epithet ar-Raḥmān is usually thought to be a borrowing from South Arabian,
where it in turn is a borrowing from Aramaic (Jeffery 2007 [1938]: 140–141).
While Ethio-Semitic contact has been fairly well-researched, research into contact with Ancient South Arabia is still in its infancy. The exact classification of
the Old South Arabian languages and their relation to Modern South Arabian
and Ethio-Semitic is still very much under debate. A simple understanding of
this highly multilingual region seems impossible. Due to the extensive contact
within South Arabia and the South Arabian languages, it is not always easy to
pin down the exact vector of contact between CA and these languages of South
Arabia and Ethiopia (§3.4.4).
2.5 Arabic dialects
The spoken Arabic dialects, of course, have had and continue to have a noticeable
influence on CA and MSA (§2.5; 3.2.1; 3.2.2; 3.3; 3.5). It seems that from the very
moment CA became canonized as an official language, it was already a highly
artificial register that nobody spoke in the form in which it was canonized. Especially the Ḥiǧāzi conquerors had a noticeable effect on the language – no doubt
through mediation of the Quranic text. Noticeable irregularities in the treatment
of the glottal stop, for example, have entered the language, and have influenced
the treatment of certain morphological features (§3.2).
2.6 Ottoman Turkish
In the Ottoman period, Ottoman Turkish becomes the official language in use in
the Middle East, and replaces many of the sociolinguistic functions that CA had
59
Marijn van Putten
previously had. The imposition of this official language had a significant effect
on the Arabic vernaculars throughout the Middle East (even outside the borders
of the Ottoman Empire), but also had a noticeable impact on the vocabulary of
CA, especially in the eighteenth and nineteenth centuries, which feeds into MSA
(§3.4.6).
3 Contact-induced changes in Classical and Modern
Standard Arabic
3.1 Phonology
Due to the highly conservative nature of CA, finding any obvious traces of contact in phonological change is very difficult. From the period in which Sibawayh
describes the phonology of the ʕarabiyya until today, only minor changes have
taken place in the phonology of CA. The most obvious example of this is the
loss of the lateral realization of the ḍād, which in Sibawayh’s description is still
a lateral, while today it is generally pronounced as [dˤ]. Blau (1969: 162–163) convincingly attributes this development to influence from the modern dialects. In
most modern Arabic dialects, the reflexes of ḍ [ɮˤ] and ð̣ [ðˤ] merged to ð̣ [ðˤ].2
In sedentary dialects that lose the interdentals, this merged sound subsequently
shifts to ḍ [dˤ]. As such, original ð̣ and ḍ are either both pronounced as an emphatic interdental fricative or both as an emphatic dental stop. As virtually all
modern dialects, however, have lost the lateral realization of ḍ, the sedentary
stop realization was repurposed for the realization of ḍ, to introduce the phonemic distinction between ð̣ and ḍ in MSA.
As this is a case where the speakers influencing the phonology of the RL are
SL-dominant, this change in pronunciation of the ḍ from a lateral to a stop realization can be seen as a form of imposition on the phonology of MSA. It should
be noted, however, that the type of imposition we are dealing with in this case is
of quite a different character than what is traditionally understood as imposition
within the framework of Van Coetsem (1988; 2000). In this case, we see a conscious effort to introduce a phonemic distinction lost in the SL between original
ḍ and ð̣ by using two different dialectal outcomes of the merger of these two
phonemes.
Other cases of phonetic imposition on MSA from the modern dialects may especially be found in the realization of the ǧīm. While Sibawayh’s description of
the ǧīm was probably a palatal stop [ɟ], today the realization that seems to carry
2
60
Not all dialects, however, see Behnstedt (2016: 16ff.).
3 Classical and Modern Standard Arabic
the most prestige and is generally adhered to in Quranic recitation is [ʤ]. However, here too we often find imposition of the local pronunciation of this sound
in MSA. In spoken MSA of Egyptians the ǧīm is regularly pronounced as [ɡ], the
realization of the ǧīm in Egyptian Arabic. Likewise, Levantine Arabic speakers
whose reflex of the ǧīm is [ʒ] will often use that realization when speaking MSA.
If we shift our focus to developments that began in the pre-Classical period and
continue in CA, we find that there are several phonetic developments that bear
some similarity to developments of Aramaic. It has therefore, not unreasonably,
been suggested that such developments are the result of contact with Aramaic.
The first of these similar phonetic developments shared between CA and Aramaic is the shift of the semivowels w and y to ʔ between a preceding ā and a following short vowel i or u. This can be seen, for example, in the similar outcomes
of the active participles of hollow roots. This similarity was already remarked
upon and described by Brockelmann (1908: 138–139), e.g.:
(1)
a. CA *qāwimun > qāʔimun ‘standing’
b. Aram. *qāwim > qāʔem ‘standing’
However, it is clear that, at least in Nabataean Arabic, this development had
not yet taken place (Diem 1980: 91–93). This is a dialect that was certainly in
contact with Aramaic, as most of the writing of the Nabataeans was in a form
of Aramaic. As such, we may plausibly suggest that this development took place
after the establishment of linguistic contact between Aramaic and Arabic. It is
quite difficult to decide whether this development, if we are correct to interpret
it as the result of contact-induced change, is the result of imposition, borrowing
or convergence. We do not have a clear enough picture of the sociolinguistic
relations between Aramaic and pre-Classical Arabic to identify the type of contact situation that would have caused it. One is tempted to see it as the result of
imposition simply because of the fact that phonological borrowing seems to be
uncommon (Lucas 2015: 526).3
As proposed by Al-Jallad (this volume), another possible case of contact induced phonological change between Aramaic and pre-Classical Arabic is the shift
of pausal -at to ah, found only in nouns and not in verbs. Huehnergard & Rubin
(2011: 267–268) already suggested that this development, which cannot be due to
a development in a shared ancestor, may have been the result of areal diffusion.
3
We cannot discount the possibility of parallel development, however. Akkadian seems to have
undergone an almost identical development (Huehnergard 1997: 196), where it is not likely to
have been the result of contact.
61
Marijn van Putten
Whether we can really interpret the development of Aramaic as similar to
that of CA, however, depends somewhat on the interpretation of the Aramaic
evidence. While we can indeed see a development of the original Aramaic feminine ending *-at that is written with 〈-h〉 in consonantal writing, which might
suggest it has shifted to /ah/, one also finds that all other cases of word-final
nominal t have been lost, while not leaving a consonantal -h, e.g.:
(2)
a.
b.
c.
d.
*ṣalōt > ṣlō 〈ṣlw〉 ‘prayer’
*zakūt > zkū 〈zkw〉 ‘merit, victory’
*ešāt > ʔešā 〈ʔšʔ〉 ‘fire’
*bayt > bay 〈by〉 ‘house’
For this parallel loss of final t in all other environments, Beyer (1984: 96, fn.
4) prefers to interpret the 〈-h〉 as a mater lectionis for final /ā/ or /a/. In this
interpretation, the development of Aramaic compared to Arabic is quite different,
since in Arabic the 〈-h〉 is clearly consonantal, and the loss of final t does not
happen after long vowels in Arabic:
(3) Aramaic
a. *kalbat > kalbā 〈klb〉 ‘bitch’ (-at# > -a/ā)
b. *ʔešāt > ʔešā 〈ʔšʔ〉 ‘fire’ (-āt# > -ā)
(4)
Arabic
a. *kalbat > kalbah
b. *kalbāt > kalbāt remains unchanged
However, if one takes the 〈-h〉 of the feminine to originally represent *-at >
-ah, and the loss of t in other word-final positions to be a different development,
one could reasonably attribute the development in Arabic to the result of contact
with Aramaic, as it is clear that in many varieties of pre-Islamic Arabic, the *-at
> -ah shift had not yet taken place.4
3.2 Morphology
3.2.1 Imposition of the taCCiʔah stem II verbal noun for glottal-stop-final
verbs
A well-known feature of Ḥiǧāzī Arabic in the early Islamic period, and a feature
that is found in many of the modern dialects, is the (almost) complete loss of the
4
62
For a discussion on the development of the *-at > -ah shift in pre-Islamic Arabic see Al-Jallad
(2017: 157–158).
3 Classical and Modern Standard Arabic
glottal stop (Rabin 1951: 130–131; van Putten 2018). This loss has usually caused
glottal-stop-final roots to be reanalyzed as final-weak verbs, e.g. Cairene ʔara,
ʔarēt ‘he read, I read’ (< *qaraʔa, *qaraʔtu).
A typical feature of final-weak verbal noun formations in CA is their formation
of the verbal noun of stem II verbs. Sound verbs form verbal nouns using the pattern taCCīC, e.g. taslīm ‘greeting’ from sallama ‘to greet’. Final-weak verbs, however, regularly use the pattern taCCiyah instead (Fischer 2002: 44), for example,
tasmiyah ‘naming’ from sammā ‘to name’.5
In CA, the ʔ generally functions as a regular consonant. Thus a verb like
qaraʔa/yaqraʔu ‘to read, recite’ does not differ significantly in its behavior from
any other triconsonantal verb such as fataḥa/yaftaḥu ‘to open’.
However, verbs with ʔ as final root consonants unexpectedly frequently side
with the final-weak verbs when it comes to the verbal noun of stem II verbs
(Fischer 2002: 128). For example, hannaʔa/yuhanniʔu ‘to congratulate’ does not
have the expected verbal noun **tahnīʔ, but instead tahniʔah ‘congratulation’.
Other examples are:
(5)
a.
b.
c.
d.
nabbaʔa v.n. tanbiʔah (besides tanbīʔ ) ‘to inform’
barraʔa v.n. tabriʔah ‘to acquit’
hayyaʔa v.n. tahyiʔah (besides tahyīʔ ) ‘to make ready’
naššaʔa, v.n. tanšiʔah (besides tanšīʔ ) ‘to raise (a child)’
Some other verbs with the same pattern do have the expected CA form such
as baṭṭaʔa v.n. tabṭīʔ ‘to delay’.
This behaviour can plausibly be attributed to the fact that in many (if not
most) spoken varieties of Arabic, from early on the final-glottal-stop verbs had
already merged completely with the final-weak verbs, and as such a verb like
hannaʔa had come to be pronounced as hannā, and was thus reanalyzed as a finalweak verb. Like original final-weak verbs, their regular verbal noun formation
would be tahniyah. When verbs of this type were employed in CA, the weak root
consonant y was replaced with the etymological glottal stop ʔ, rather than completely converting the verbal noun to the regular pattern. This is a clear example
of the imposition of a morphological pattern onto CA grammar by speakers of
Arabic dialects.
5
This is an ancient idiosyncrasy of final-weak verbs. While the taCCīC formation is not a regular
formation in other Semitic languages, when it does occur, the final-weak verbs have a feminine
ending, e.g. Hebrew tarmi-ṯ ‘betrayal’, toḏå ‘praise’ (< *tawdiy-ah), see Brockelmann (1908:
385–387).
63
Marijn van Putten
3.2.2 Imposition of the ʔaCCiyāʔ broken plural pattern
A similar case of imposition, where the morphological categories of glottal-stopfinal roots behave in the grammar as if they are final-weak, may be found in the
broken-plural formation of CaCīʔ nouns and adjectives. The broken-plural formation most generally used for final-weak adjectives with the pattern CaCiyy (<
*CaCīy) is ʔaCCiyāʔ. For example, ɣaniyy pl. ʔaɣniyāʔ ‘rich’, waliyy pl. ʔawliyāʔ
‘close associate’, daʕiyy pl. ʔadʕiyāʔ ‘bastard’, sawiyy pl. ʔaswiyāʔ ‘correct’, ḫaliyy
pl. ʔaḫliyāʔ ‘free’.
For sound nouns of this type, it is much more typical to use the plural formations CiCāC (kabīr pl. kibār ‘big’) or CuCaCāʔ (faqīr pl. fuqarāʔ ‘poor’), although
there are a couple of sound nouns that do use this plural, such as qarīb pl. ʔaqribāʔ
‘relative’ and ṣadīq pl. ʔaṣdiqāʔ ‘friend’ (Ratcliffe 1998: 106–107).6
CaCīC formations where the last root consonant is ʔ, however, behave in rather
unexpected ways in CA, usually following the pattern of final-weak nouns, often
even replacing the final ʔ with y, for example: barīʔ pl. ʔabriyāʔ ‘free’, radīʔ pl.
ʔardiyāʔ ‘bad’. These nouns have plurals that are proper not to the Classical form
they have, but rather to the colloquial form without ʔ, i.e. bariyy, radiyy. Once
again this can be seen as a clear case of imposition of the colloquial Arabic forms
onto the classical language.7
3.2.3 Borrowing of the broken plural pattern CaCāCiCah
CA, like the modern Arabic dialects, is well-known for its broken-plural patterns.
This is a feature it shares especially with Old South Arabian (Stein 2011: 1050–
1051), Modern South Arabian languages (Simeone-Senelle 2011: 1085) and EthioSemitic (Weninger 2011a: 1132). The use of broken plurals has caused somewhat
6
The pattern (with metathesis) is also regular for geminated CaCīC adjectives, e.g. šadīd pl.
ʔašiddāʔ ‘severe’.
7
These two cases of imposition of glottal stop-less morphology onto CA are two of the more
clear and systematic cases, but close observation of CA morphology reveals many more of
these somewhat more isolated cases, e.g. ḫaṭīʔah ‘sin’ with a plural ḫaṭāyā, for which the expected singular would rather be ḫaṭiyyah; bariyyah pl. barāyā ‘creature’ which is a derivation
from baraʔa ‘to create’; ðurriyyah, ðirriyyah pl. ðarāriyy ‘progeny, offspring’, derived from
ðaraʔa ‘to sow, seed’. Another example of irregular treatment of ʔ that is presumably the result of impositition is found in verbal nouns of stem VI verbs, and mafāʕil plurals of hollow
roots, which modern textbooks say should not have a ʔ despite having the environment that
is expected to undergo the shift āwu/i, āyi > āʔu/i, āʔi as discussed in §3.1. The lexicographical
tradition and Quranic reading traditions often record disagreements on the application of the
hamzah in such cases. For example, we find both tanāwuš and tanāʔuš ‘reaching one another’,
and maʕāyiš and maʕāʔiš ‘ways of living’.
64
3 Classical and Modern Standard Arabic
of a controversy in the subgrouping of the Semitic language family. Scholars
who consider broken plurals a shared retention do not view their presence as
important for grouping Arabic, Old South Arabian, the Modern South Arabian
languages and Ethio-Semitic together (Huehnergard 2005: 159–160); while those
who consider their presence an innovation in a subset of Semitic languages see
this as a strong indication that these languages should be grouped together into
a South Semitic branch (e.g. Ratcliffe 1998).
While most scholars today seem to agree that the broken-plural system is a
shared retention (Weninger 2011b: 1116), it seems clear that the retention of a
highly productive broken-plural system is to be considered an areal feature that
clusters around South Arabia and the Horn of Africa. CA partakes in this areal
feature.
A possible case of influence from Old South Arabian (and/or Ethio-Semitic)
into Arabic is the introduction of the CaCāCiCah plural formation. In the South
Arabian languages,8 the equivalent plural formation CaCāCiCt is extremely productive, and numerous words with four consonants form their plural in this way.
For example in Sabaic, mCCCt is the regular plural formation to mCCC nouns of
location, e.g. mḥfd pl. mḥfdt ‘tower’ (Beeston 1962: 34). It is likewise common in
Gəʕəz, e.g. tänbäl pl. tänabəlt ‘ambassador’ (Dillmann 2005 [1907]: 309), and occurs occasionally in Modern South Arabian, e.g. Mehri məlēk pl. məlaykət ‘angel’
(Rubin 2010: 68).
While this pattern exists in CA, it is much rarer than the other broken plural
formations of four consonantal forms, i.e. CaCāCiC and CaCāCīC. In the Quran,
malak pl. malāʔikah ‘angel’ is the only plural with this pattern. This noun is
widely recognized as being a loanword from Gəʕəz malʔak, malāʔəkt (Jeffery
2007 [1938]: 269), in part on the basis that it shares this plural formation: the
word seems to have been borrowed together with its plural formation. Considering the rarity of this pattern in Arabic and how common it is in South Arabian,
it seems possible that the pattern was introduced into Arabic through South Arabian contact. However, the absence of other clearly identifiable South Arabian
loanwords with this plural pattern makes it rather difficult to make a strong case
for this identification.
Another possible word of South Arabian origin with this plural pattern is
tubbaʕ pl. tabābiʕah ‘a Yemenite king’, but evidence that this word is indeed
of Old South Arabian origin is missing. The word does not occur as a separate
word in Old South Arabian, and instead is only the first part of several Old South
Arabian theophoric names such as tbʕkrb, tbʔʕl. Such names should probably be
8
South Arabian is used here as a purely geographical descriptive term, not one of classification.
65
Marijn van Putten
understood as being related to the root √tbʕ which, like in Arabic, may have had
the meaning ‘following’, so such names likely mean ‘follower of the deity KRB’
and ‘follower of the deity ʔL’. Such names being associated with Yemenite kings
may have led to the Arabic meaning of tubbaʕ as ‘Yemenite king’, but in Old
South Arabian itself it does not seem to have carried a meaning of this kind.
All in all, the evidence for this being a pattern that is the result of South Arabian influence is rather slim, although the rarity of the pattern in CA does make
it look unusual. If the interpretation of this plural pattern as being a borrowing
from South Arabian is correct, it seems that some South Arabian nouns were
borrowed along with their respective plural. This would be a case of morphological borrowing rather than the more common type of morphological influence
through imposition.9
Note that this plural pattern has become the productive plural pattern for
quadriconsonantal loanwords regardless of them being of South Arabian origin
or elsewhere, e.g. biṭrīq pl. baṭāriqah ‘patrician’ (< Latin patricius), ʔusquf pl.
ʔasāqifah ‘bishop’ (< Greek epískopos), ʔustāð pl. ʔasātiðah ‘master’ (< Middle
Persian ōstād), tilmīð pl. talāmiðah ‘student’ (< Aramaic talmīḏ).
3.3 Syntax
Due to CA being the High register in a diglossic situation for centuries, we should
presumably consider the majority of the written material produced in this language to be written exclusively by non-native speakers. Moreover, a large proportion of its writers all throughout its written history must have been speakers not
only of Arabic vernaculars but also of entirely different languages such as Persian and Turkish. It seems highly unlikely that such a multilingual background
of authors of CA would have been completely without effect on the syntax of
the language; however, as it is difficult to decide from what moment onward we
can speak of true diglossia, and what the syntax was like before that period, it
has not yet been possible to trace such influences in detail.
There is, however, promising research being done on influence on MSA syntax from the speakers of modern Arabic dialects. Wilmsen (2010) convincingly
describes one such point of influence in a paper on the treatment of object pronouns in Egyptian and Levantine newspapers.
Wilmsen (2010: 104) shows that, in the case of ditransitive verbs, Egyptian
and Levantine have a different natural word order. In Egyptian Arabic, the direct
9
66
This can be seen as a type of “Parallel System Borrowing” similar to that which we find in
Berber languages. Berber languages, like Arabic, have apophonic plurals; but Arabic nouns
are simply borrowed along with their own Arabic broken plurals (Kossmann 2010).
3 Classical and Modern Standard Arabic
object must precede the indirect object as in (6), while in Levantine Arabic the
indirect object preceding the direct object is preferred, as shown in (7):
(6) Egyptian
rabbi-na yḫalli-hū-l-ak
Lord-1pl keep.impf.3sg.m-3sg.m-dat-2sg.m
‘Our lord keep him for you.’
(7)
Levantine
aḷḷa yḫallī-l-ak
iyyā
God keep.impf.3sg.m-dat-2sg.m acc.3sg.m
‘God keep him for you.’
Wilmsen argues that the following two variant sentences in a Reuters news
story written in MSA, the original in (8), likely written by an Egyptian, and the
slightly altered version in (9), which appeared in a Lebanese newspaper, show
exactly this difference of word order found in the respective spoken dialects:
(8)
MSA (Egyptian)
al-ʔawrāq-i
llatī
sallamat-hā
la-hu
ʔarmalat-u
def-papers-obl rel.sg.f give.prf.3sg.f-3sg.f dat-3sg.m widow-nom
ʕabdi l-wahhāb
pn
‘the papers, which Abdel Wahhab’s widow had given him’
(9)
MSA (Lebanese)
al-ʔawrāq-i
llatī
sallamat-hu
ʔiyyā-hā ʔarmalat-u
def-papers-obl rel.sg.f give.prf.3sg.f-3sg.m acc-3sg.f widow-nom
ʕabdi l-wahhāb
pn
‘the papers, which Abdel Wahhab’s widow had given him’
Wilmsen (2010: 114–115) goes on to examine three newspapers (the Londonbased, largely Lebanese, al-Ḥayāt of the years 1996–1997; the Syrian al-Θawra of
the year 2005 and the Egyptian al-ʔAhrām), and shows that with the two most
common verbs in the corpus with such argument structure (manaḥa ‘to grant’
and ʔaʕṭā ‘to give’), the trend is consistently in favour of the pattern found. The
recipient–theme order is overwhelmingly favoured in the Levantine newspapers,
while the theme–recipient order is clearly favoured by the Egyptian newspaper.
The results are reproduced in Tables 1 and 2.
67
Marijn van Putten
Table 1: Occurences of theme–recipient and recipient–theme order
with manaḥa ‘to grant’
Database
al-Ḥayāt 96
al-Ḥayāt 97
al-Θawra
al-ʔAhrām
theme–recipient
recipient–theme
29
27
27
44
56
52
66
8
Table 2: Occurrences of theme–recipient and recipient–theme order
with ʔaʕṭā ‘to give’
Database
al-Ḥayāt 96
al-Ḥayāt 97
al-Θawra
al-ʔAhrām
theme–recipient
recipient–theme
11
8
9
33
23
22
38
2
From this data it is clear that the dialectal background of the author of an MSA
text can indeed play a role in how its syntax is constructed, despite both resulting
sentences being grammatically acceptable in CA/MSA.10
This (and any contact phenomenon in MSA–dialect diglossia) should be seen
as a case of imposition, where the dialect SL, in which the speakers/writers are
dominant, has influenced the MSA RL.
It stands to reason that such syntactic research could be undertaken with CA
works as well. Taking into account the biographies of authors, it might be possible to find similar imposition effects that can be connected to different dialects
and languages in former times. To my knowledge, however, this work has yet to
be undertaken.
3.4 Lexicon
In terms of lexicon, Jeffery’s indispensable (2007 [1938]) study of the foreign vocabulary in the Quran allows us to examine some of the important sources of
lexical influence on pre-Classical Arabic. Influence from Greek, Aramaic, Gəʕəz
and Persian are all readily recognizable.
10
68
Other works that discuss clear cases of country-specific language use of MSA include Ibrahim
(2009), Parkinson (2003), Parkinson (2007) and Parkinson & Ibrahim (1999).
3 Classical and Modern Standard Arabic
3.4.1 Gəʕəz
Nöldeke (1910) is still one of the most complete and important discussions of
Gəʕəz loanwords in CA. Both Gəʕəz and Arabic display a significant amount of
religious vocabulary that is borrowed from Aramaic. It is quite often impossible
to tell whether Arabic borrowed the word from Gəʕəz or from Aramaic. Such
examples are ṭāɣūt ‘idol’, Gz. ṭaʕot, Aram. ṭāʕū ‘error, idol’ (Nöldeke 1910: 48);
tābūt ‘ark; chest’, Gz. tabot ‘ark of Noah, ark of the covenant’, Aram. tēḇō ‘chest;
ark’ (Nöldeke 1910: 49).
There is religious vocabulary that is unambiguously borrowed from Gəʕəz, e.g.
ḥawāriyyūn ‘disciples’ < Gz. ḥäwarəya ‘apostle’ and muṣḥaf ‘book (esp. Quran)’
< Gz. mäṣḥäf ‘scripture’, but there is also religious vocabulary borrowed unambiguously from Aramaic, e.g. zakāt ‘alms’ < Aram. zāḵū ‘merit, victory’; sifr
‘large book’ < Aram. sp̄ar, sep̄rā. It is therefore just as likely that Arabic would
have borrowed such Aramaic loanwords via Gəʕəz as directly from Aramaic.
Some religious vocabulary from Aramaic and Hebrew can be shown to have
arrived in Arabic through contact with Gəʕəz, since these words have undergone specific phonetic developments shared between CA and Gəʕəz but absent
in the source language. As these often involve core religious vocabulary, and
the Christian Axumite kingdom was established centuries before Islam, it seems
reasonable to assume such words to be borrowings from Gəʕəz into CA, e.g. CA
ǧahannam ‘hell’ < Gz. gähännäm (but Hebrew gehinnom and Syriac gehannā)
and CA šayṭān ‘Satan’ < Gz. śäyṭan (but Hebrew śåṭån and Syriac sāṭānā).11
3.4.2 Aramaic
As already remarked upon by Retsö (2011), Aramaic loanwords in CA often have
an extremely archaic character. The Aramaic variety that influenced Quranic and
pre-Classical Arabic had not undergone the famous bəḡaḏkəp̄aṯ lenition of postvocalic simple stops, nor had it lost short vowels in open syllables. This necessarily means that the form of Aramaic that influenced Quranic and Classical Arabic,
even the religious vocabulary, cannot be Syriac, which almost certainly underwent both shifts before becoming a dominant religious language. The bəḡaḏkəp̄aṯ
spirantization can be dated between the first and third centuries CE, and the syncope of short vowels in open syllables takes place sometime in the middle of the
third century (Gzella 2015: 41–42). However, Classical Syriac itself, as an important vehicular language of Christianity, only emerges in the fourth century CE,
well after these developments had taken place (Gzella 2015: 259).
11
Leslau (1990) often reverses the directionality of such borrowings, though without an explanation as to why he thinks a borrowing from CA into Gəʕəz is more likely.
69
Marijn van Putten
Had bəḡaḏkəp̄aṯ taken place, we would expect Syr. ḡ, ḏ, ḵ, and ṯ to be borrowed
with their phonetic equivalents in CA: ɣ, ð, ḫ, and θ respectively.12 This, however,
is not the case; instead these consonants are consistently borrowed with the stop
equivalents ǧ, d, k, and t, and without the loss of vowels in open syllables, clearly
showing that these Aramaic loanwords predate the phonetic developments in
Classical Syriac.
(10)
a. malakūt ‘kingdom’, Syr. malkūṯ-ā ‘kingdom’ < *malakūt-ā
b. malik ‘king’, Syr. mleḵ ‘king’ < *malik13
c. masǧid ‘place of worship, mosque’, Syr. masgeḏ-ā ‘place of worship’ <
*masgid-ā
Even the proper names of Biblical figures have a markedly un-Syriac form.
(11)
a. zakariyā, zakariyāʔ, Syr. Zḵaryā < *zakaryā
b. mīkāʔīl, mīkāʔil,14 Syr. mīḵāʔel < *mīkāʔēl
In other words, far from Syriac being “undoubtedly the most copious source
of Qurʾānic borrowings” (Jeffery 2007 [1938]: 19), the Aramaic vocabulary in the
Quran seems to not be Syriac at all.15 Any isogloss that would allow us to identify
it as such is conspicuously absent. This has important historical implications, as
the presence of supposed Syriac religious vocabulary in the Quran is viewed as an
important indication that Syriac Christian thought had a pronounced influence
on early Islam (e.g. Mingana 1927: 82–90; Jeffery 2007 [1938]: 19–22).16 While
12
Retsö (2011) suggests that ḇ could also be borrowed as w. This might be true, but at least the
phonetic match in this case is not perfect.
13
This word is not recognized as an Aramaic loanword by Jeffery (2007: 270), but it likely is.
All the Semitic cognates of this noun are derived from a form *malk, which should have been
reflected in CA as malk. However, we find it with an extra vowel between the last two root
consonants. This can be best understood as the epenthetic vowel insertion as it is attested
in Aramaic which was then subsequently borrowed with this epenthesis into Arabic. I thank
Ahmad Al-Jallad for pointing this out to me.
14
Most readers of the Quran read either mīkāʔīl or mīkāʔil, only the most dominant tradition
today, that of Ḥafṣ, reads it in the highly unusual form mīkāl (Ibn Muǧāhid no date: 166).
15
Note that Jeffery (2007 [1938]: 19) explicitly states that by Syriac he means any form of Christian
Aramaic, so, besides Syriac, most notably also Christian Palestinian Aramaic. However, this
caveat hardly solves the chronological problem, as the latter rises to prominence even later.
16
Even if we were to accept the possibility that the dating of the lenition and syncope is somehow
off by several centuries, the suggestion that “it is possible that certain of the Syriac words we
find in the Qurʾān were introduced by Muḥammad himself” (Jeffery 2007 [1938]: 22) must
certainly be rejected. In the grammatical works of Jacob of Edessa (640–708 CE) we have an
unambiguous description of the lenition of the consonants (Holger Gzella p.c.). It seems highly
unlikely that a wholesale lenition took place in only a few decades between the composition
of the Quran and the time of his writings.
70
3 Classical and Modern Standard Arabic
this is of course still a possibility, this has to be reconciled with the fact that
the majority of clearly monotheistic religious vocabulary was already borrowed
from a form of Aramaic before the rise of Syriac as a major religious language.
This does not mean that CA is completely devoid of Aramaic loanwords that
have undergone the lenition of the consonants, and several post-Quranic loanwords have been borrowed from a variety which, like Syriac, had lenited its stops,
e.g.:
(12)
a.
b.
c.
d.
e.
tilmīð ‘student’ < Syr. talmīḏā (Fraenkel 1886: 254)
tūθ, tūt ‘mulberry’ < Syr. tūṯā (Fraenkel 1886: 140)
ḥiltīθ, ḥiltīt ‘asa foetida’ < Syr. ḥeltīṯā (Fraenkel 1886: 140)
kāmaḫ, kāmiḫ ‘vinegar sauce’ < Syr. kāmḵā (Fraenkel 1886: 288)
karrāθ, kurrāθ ‘leek’ < Syr. karrāṯā (Fraenkel 1886: 144)
It is interesting to note that Aramaic loanwords in Gəʕəz reflect a similar archaicity, in those cases where this is detectable. The expected lenited ḵ is not represented with Gəʕəz ḫ but with k, and short vowels in open syllables are retained. This might suggest that, when looking for religious influences on Islam,
we should rather shift our focus to the south, where during the centuries before
Islam both Judaism and Christianity were introduced, presumably through the
vector of Gəʕəz. Some examples of such similarly archaic Aramaic loanwords in
Gəʕəz are cited by Nöldeke (1910: 31–46), e.g.:
(13)
Gəʕəz
a.
b.
c.
d.
e.
f.
g.
h.
mälʔäk ‘angel’, cf. CA malak, Syr. malʔaḵ-ā < *malʔak-ā
mäläkot ‘kingdom’, cf. CA malakūt, Syr. malkūṯ-ā < *malakūt-ā
ḥämelät ‘mantle, headcloth’, Syr. ḥmīlṯ-ā < *ḥamīlat-ā
näbīy ‘prophet’, cf. CA nabiyy, nabīʔ, Syr. nḇīyyā < *nabīʔ-ā
mäsīḥ ‘Messiah’, cf. CA al-masīḥ, Syr. mšīḥ-ā < *masīḥ-ā
siʔol ‘hell’, cf. Syr. siwūl < *siʔūl (cf. Hebr. səʔol)
ʔärämi, ʔärämāwi, ʔärämay ‘heathen’, cf. Syr. ʔarmāy-ā < *ʔaramāy-ā
mänarät, mänarat ‘candlestick’, cf. CA manārah, Syr. mnārṯ-ā <
*manārat-ā
As of yet, there is not a clear historical scenario that helps us better understand how both CA and Gəʕəz, and, from the scanty information that we currently have, also Old South Arabian, ended up with similarly archaic forms of
71
Marijn van Putten
Aramaic. This seems to suggest an as yet unattested, very archaic form of Aramaic in South Arabia. Alternatively, the syncope and lenition so well-known in
Syriac may have had a much less broad distribution across the written Aramaic
dialects than previously thought.
3.4.3 Greek (and Latin)
Besides this noticeable cluster of Aramaic and Gəʕəz words, there are of course
also Greek loanwords in CA, generally in the semantic fields of economy and
administration. Very often Aramaic likewise has these words, and it is usually
not possible to decide whether Arabic borrowed the word from Aramaic or directly from Greek. The former direction is presumably more likely considering
the broad presence of Aramaic as a lingua franca. Some examples are e.g. dīnār
‘dinar’, Aram. dēnār, Gk. dēnárion, Lat. denarius; zawǧ ‘spouse, pair’, Aram. zōḡ
‘id.’, Gk. zeûgos ‘yoke’; ṣirāṭ ‘way’, Aram. ʔesṭrāṭ ‘street’, Gk. stráta, Lat. (via)
strata; qirṭās ‘parchment, papyrus’, Aram. qarṭīs, Gk. kʰártēs; qaṣr ‘castle’, Aram.
qaṣrā, Gk. kástron, Lat. castrum; qalam ‘reed-pen’, Gk. kálamos ‘reed-pen’.17
A new influx of mostly philosophical and scientific Greek vocabulary entered
CA during the early Abbasid period (mid 8th–10th centuries), at the time of the
Graeco-Arabic translation movement (Gutas 1998). Once again, these words seem
to have entered the language through Syriac (Gutas 2011). From this translation
movement, we have words such as ǧins ‘genus’ < Syr. gensā < Gk. génos; faylasūf
‘philosopher’ < Syr. pīlōsōp̄ā < Gk. pʰilósopʰos; kīmyāʔ ‘alchemy’ < Syr. kīmīyā <
Gk. kʰēmeía; and ʔistāðiyā ‘stadium’18 < Syr. estaḏyā < Gk. stádion.
3.4.4 Old South Arabian
It is often difficult to establish from which of the South Arabian languages a
certain word originates. As Old South Arabian retained all the Proto-Semitic
consonants, a borrowing from Old South Arabian or an inheritance from ProtoSemitic is often difficult to distinguish in CA. While Jeffery (2007 [1938]: 305)
identifies a fair number of possible words of South Arabian origin, hardly ever
17
Nöldeke (1910: 50) argues that the CA qalam must come from Greek through Gz. qäläm. While
this is possible, there is nothing about this word that requires us to assume this directionality,
nor is it particularly unlikely that CA and Gəʕəz independently borrowed this word without
its Greek ending -os.
18
Note here the apparent application of the Syriac lenition being borrowed as such in Arabic,
unlike earlier loans. But it may also be possible that the lenition is part of the Greek lenition
of the delta instead, as we see it today in Modern Greek.
72
3 Classical and Modern Standard Arabic
does this seem the only possibility. Another issue with identifying South Arabian loanwords is that we have very scanty knowledge of its vocabulary or its
linguistic developments. As a result, Old South Arabian identifications can be
quite difficult to substantiate.
In recent years several lexical studies have tried to draw connections between
Old South Arabian and Arabic vocabulary, but this is often based on certain semantic extensions or uses of words as described in CA dictionaries. While these
observations may eventually be proven correct, it is somewhat difficult to evaluate whether we are truly dealing with borrowings in these cases, and the extremely limited knowledge that we have of the vowel system of the different Old
South Arabian languages makes it difficult to evaluate this in detail. Several interesting suggestions are given by Weninger (2009), Hayajneh (2011) and Elmaz
(2014; 2016).
To illustrate the difficulties we run into when trying to identify Old South
Arabian borrowings in Arabic, let us examine the word tārīḫ pl. tawārīḫ ‘date’.
From the perspective of CA morphology, tārīḫ could only be a hypocorrect form
of taʔrīḫ – which is indeed an attested biform of tārīḫ. The existence of the plural
tawārīḫ rather than taʔārīḫ, however, seems to suggest that taʔrīḫ is rather a
hypercorrect insertion of hamzah from an original form tārīḫ, which certainly
looks foreign in its formation.
Both Hebbo (1984: 27) and Weninger (2009: 399) have suggested that this word
is to be connected with the the widespread Semitic root √wrḫ, related to ‘month’
or ‘moon’ (cf. Hebrew yɛraḥ < *warḫ ‘month’), which exists in Old South Arabian
but not in CA.19 The verb ʔarraḫa ‘to date’ would then reasonably be taken as a
backformation from tārīḫ.
However, this explanation still leaves us with many problems. There is perhaps
some reason to suppose that in Old South Arabian *aw would have collapsed to
an unknown monophthong (Early Sabaic ywm ‘day’; Late Sabaic ym). This might
explain why the word is tārīḫ and not **tawrīḫ, but tārīḫ is not actually attested
in Old South Arabian. So while the suggestion is certainly possible, it seems that
another of the many non-Arabic Ancient northern Arabian epigraphic languages
could likewise have been an origin. Barring further discoveries, many such proposed etymologies remain highly speculative, and drastically simplify the rather
complex multilingual situation of pre-Islamic Arabia, where many other sources
besides Old South Arabian remain possible (Al-Jallad 2018).
19
Note, however, that the root √wrḫ ‘month’ is attested unambiguously in the singular (wrḫ),
dual (wrḫn) and plural (ʾrḫ) in the Old Arabic corpus of Safaitic inscriptions (Al-Jallad 2015:
353).
73
Marijn van Putten
3.4.5 Persian
Whereas with the advent of Islam the influence of Aramaic, Greek and Gəʕəz
on CA quickly diminished and disappeared, the influence of Persian actually increased. While the Quran already contains a sizeable number of Persian borrowings, this only increases in the following centuries.
Some clear Persian borrowings in the Quran include: ʔistabraq ‘silk brocade’,
cf. New Persian istabra (Eilers 1962: 204); numruq ‘cushion’ < Middle Persian
namrag; kanz ‘treasure’ < Middle Persian ganz/ganǧ ‘treasury’ (Eilers 1962: 206).
Outside of the Quran many other Persian words may be found in Arabic, e.g.
dīwān ‘archive, collected writings’ < Early New Persian dīwān (Eilers 1962: 223),
banafsaǧ, manafsaǧ ‘violet’ < Middle Persian banafš (Eilers 1971: 596); barnāmaǧ
‘program’ < Middle Persian bārnāmag (Eilers 1962: 217-218); wazīr ‘minister’ <
Middle Persian wizīr (Eilers 1962: 207).20
3.4.6 Ottoman Turkish
The influence of Ottoman Turkish on MSA is significantly less than on the modern Arabic dialects, largely due to linguistic purism (Procházka 2011). Words that
have entered MSA are words related to administration, technology and food,
but also several other origins are found. For example: damɣa ‘stamp’ < damga;
ǧumruk ‘customs’ < gümrük (ultimately from Latin commercium); bāšā ‘pasha’
< paşa; bābūr < vapur ‘steam ship’ (ultimately from French [bateau à] vapeur);
quṣāǧ ‘pliers’ < kıskaç; balṭa ‘axe’ < balta; šāwurma, šāwirma ‘lamb, etc., roasted
on a spit’ < çevirme; qāwurma, qāwirma ‘fried meat’ < kavurma; kufta ‘meatballs’
< köfte.
Of some interest is the -ci suffix that denotes professions and characterizations
in Turkish. This suffix has developed some amount of productivity in modern
dialects (especially in Iraq, Syria and Egypt), where it may even be suffixed to
nouns of non-Turkish origin. In MSA the suffix is attested not infrequently, although it would probably go too far to say that it is productive. Some examples
are nawbatǧī ‘on duty; command of the guard’ < nawba ‘shift, rotatation’ + -ci;
qahwaǧī ‘coffeehouse owner’ < qahwa ‘coffee’ + -ci; xurdaǧī ‘dealer in miscellaneous smallwares’ < hordaci ‘id.’; balṭaǧī ‘sapper, pioneer’ < baltaci ‘sapper’;
būyāǧī ‘painter, bootblack’ < boyaci ‘painter’.
3.4.7 Influence of Standard Average European
A rather different, but nevertheless important factor of language contact for MSA,
especially in the journalistic style, was described by Blau (1969). Blau argues
20
74
I thank Chams Bernard for updating the transcription of the Middle Persian forms.
3 Classical and Modern Standard Arabic
that, under the influence of what he dubs “Standard Average European” (SAE;
cf. Whorf 1956), MSA (as well as Modern Hebrew) has taken on a large amount
of vocabulary,21 phraseology, and syntax similar to the journalistic language use
of European languages, though the actual languages of influence could be quite
different in different countries (e.g. Russian and Yiddish for Modern Hebrew;
English for Egyptian MSA, French for Lebanese, Moroccan, Tunisian and Algerian MSA).22 Examples of such influence take up over a hundred pages in Blau’s
pioneering work.
Blau identifies examples of lexical expansion of existing words to include lexical associations present in SAE, e.g. saṭḥī ‘flat’ is extended in meaning towards
‘superficial’ due to influence of, e.g. French superficiel and German oberflächlich
(Blau 1969: 65); ǧaww ‘air, atmosphere’ comes to be used in a metaphorical sense
in the same way English uses ‘atmosphere’, e.g. ǧawwu s-siyāsati mukahrabun
‘the political atmosphere is electrified’ (Blau 1969: 69).
Even whole phrases may show up as loan translations, such as MSA ʔanqaða
l-mawqifa ‘to save the situation’, cf. French sauver la situation, German die Situation retten; MSA qatala l-waqta ‘to kill time’, cf. French tuer le temps, German die
Zeit totschlagen (Blau 1969: 76). Even such highly specific metaphorical expressions as ‘to miss the train’, in the meaning of missing an opportunity, appears
in MSA ʔasriʕ wa-ʔillā fātaka l-qiṭāru ‘hurry, otherwise you will miss the train’
(Blau 1969: 101).
Such linguistic influence, of course, does not lend itself particularly well to be
classified within the framework of Van Coetsem (1988; 2000), as the writers of
MSA in these cases are dominant in neither the source language(s) nor the recipient language, a situation which is a rather unique result of the Arabic diglossia
in combination with the influence of foreign journalistic styles that have transformed the way in which MSA is written.
3.5 Influence of the early Islamic vernaculars
While, as a general rule, CA retains its archaic features, such as the retention
of glottal stop in all positions and the lack of vowel harmony and syncope, we
occasionally find single lexical items which optionally allow innovative forms
which presumably stem from spoken vernaculars before the standardization of
the classical language. This tends to be visible especially for words that have lost
21
For further discussion of the development of Modern Standard Arabic technical vocabulary
see Dichy (2011) and Jacquart (1994).
22
The influence of French in terms of borrowings and adaptations is especially salient in literary
Arabic as used in the Maghreb. Kropftisch (1977) is an excellent study on this topic.
75
Marijn van Putten
the glottal stop, a feature usually attributed to the Ḥiǧāzī variety of the early
Islamic period. For example, CA has nabiyy ‘prophet’, nubuwwah ‘prophethood’
from the root √nbʔ ;23 likewise bariyyah ‘creature’ from the root √brʔ.24
The likely loss of postconsonantal ʔ in Ḥiǧāzī Arabic has influenced the way
the verb raʔā ‘to see’ (√rʔy) is conjugated. Its imperfect irregularly loses the ʔ :
yarā ‘he sees’. Similarly the verb saʔala ‘to ask’ (√sʔl) has two different imperatives, either the regular isʔal or the Ḥiǧāzī sal (< *sʔal). The imperative ʔalik
‘send!’ must be the imperative of an otherwise unattested verb *ʔalʔaka ‘to send’,
which has likewise irregularly lost its postconsonantal ʔ. Besides verbs, we may
also see the irregular lack of representation of post-consonantal ʔ in other nouns,
e.g. malak ‘angel’, which, considering its plural malāʔikah and etymological origin, was presumably originally *malʔak.
The pseudo-verbs niʕma ‘what a wonderful …’ and biʔsa ‘what an evil …’, are
presumably originally from *naʕima and *baʔisa, with vowel harmony and syncope. These original forms have disappeared from the classical language in their
pseudo-verbal use, only retaining their verbal meaning: naʕima ‘to be happy,
glad’ and baʔisa ‘to be miserable, wretched’. However, other pseudo-verbs retain both unharmonized and unsyncopated forms as optional variants even in
their pseudo-verbal use: ḥasuna, ḥusna, ḥasna ‘how beautiful, magnificent’, and
ʕað̣uma, ʕuð̣ma, ʕað̣ma ‘how powerful, mighty’. Such syncopated and harmonized forms are claimed by the Arab grammarians themselves to be part of the
eastern dialects, and absent in the Ḥiǧāzī dialects (Rabin 1951: 97), but surprisingly are retained for such pseudo-verbs.
Syncopated forms, while reported for regular verbs as well by the Arab grammarians (e.g. šihda or šahda for šahida), never occur in the Classical language.
For some CaCiC nouns, syncopated forms are reported by lexicographers (e.g.
katf and kitf besides katif ), but it is not clear whether these syncopated forms
are used in CA outside of these lexicons.
These kinds of dialectal forms that appear to have been incorporated into CA
are indicative of the artificial amalgam that makes up the language, and require
a much more in-depth discussion than the present chapter allows. It seems clear
that the vast amount of dialectal variation that is described by the Arab grammarians, judiciously collected by Rabin (1951), does not end up in CA, but some
amount of variants are either allowed, or are the only possible form present in
the standard. The exact parameters that determine how and why such dialectal
forms were incorporated into the language are currently unclear.
23
In several Quranic reading traditions these are still read nabīʔ and nubūʔah, as expected (Ibn
Muǧāhid no date: 106–107).
24
Read as barīʔah in several Quranic reading traditions (Ibn Muǧāhid no date: 693).
76
3 Classical and Modern Standard Arabic
4 Conclusion
Due to CA and MSA being almost exclusively High literary registers, with no true
native speakers, the type of language contact that we see in the Islamic period is
rather different from what we may see in more natural language contact situations. We mostly see imposition of certain dialectal forms onto the Classical ideal.
An interesting exception to this is the calquing of MSA words and phraseology
upon “Standard Average European”, where the speakers are dominant in neither
the recipient nor the source language.
Borrowing can be detected in phonology, morphology and vocabulary from
Greek, Aramaic and Ethio-Semitic from the pre-Islamic period, which were then
inherited by CA. In the Islamic period, it is mostly vocabulary that is borrowed,
with a significant number of loans coming from Greek, Persian and Ottoman
Turkish into CA.
Examining these pre-Islamic borrowings, it has become clear that the Aramaic
that has primarily influenced CA, contrary to what is popularly believed, was not
a form of Syriac, but rather a more archaic variety. The historical implications
of this have not yet been well-integrated into our understanding of pre-Islamic
linguistic diversity in Arabia and neighbouring regions.
While some studies have looked at syntactic imposition of the spoken dialects
onto MSA with promising results, this has not yet been applied to medieval texts
written in CA. Nevertheless, considering the clear ethnic and geographic diversity of writers of CA, it seems likely that future work should be able to detect
such influences even in the medieval period.
Further reading
) Jeffery (2007) [1938] is still one of the most comprehensive books on loanwords
in Quranic Arabic.
) Hebbo (1984) is an in-depth study of foreign words as they appear in the Sīrah
of Ibn Hišām.
) Fraenkel (1886) is an in-depth discussion of Aramaic loanwords in Arabic, but
in some respects outdated.
) Nöldeke (1910) contains an important section on loanwords both from Arabic
to the Ethio-Semitic languages and the other way around.
) Blau (1969) is a pioneering work researching the interaction between European literary languages and the effects they have on the literary style of Modern Standard Arabic and Modern Hebrew.
77
Marijn van Putten
) The chapters on language contact in the Encyclopaedia of Arabic Language
and Linguistics are also highly useful and informative, and contain many up
to date references for contact with Greek (Gutas 2011), Persian (Asbaghi 2011),
Aramaic Retsö (2011), and Turkish (Procházka 2011).
Acknowledgements
I thank Stefan Procházka, Christopher Lucas, Maarten Kossmann and Ahmad AlJallad for providing me with important references, comments and suggestions.
Abbreviations
*
**
1, 2, 3
acc
Aram.
CA
CE
dat
f
Gk.
Gz.
impf
Lat.
reconstructed form
unattested form
1st, 2nd, 3rd person
accusative
Aramaic
Classical Arabic
Common Era
dative
feminine
Greek
Gəʕəz
imperfect (prefix conjugation)
Latin
m
MSA
nom
obl
pl
pn
prf
rel
RL
sg
SL
Syr.
v.n.
masculine
Modern Standard Arabic
nominative
oblique
plural
personal name
perfect (suffix conjugation)
relative pronoun
recipient language
singular
source language
Syriac
verbal noun
References
Al-Jallad, Ahmad. 2015. An outline of the grammar of Safaitic inscriptions. Leiden:
Brill.
Al-Jallad, Ahmad. 2017. Graeco-Arabica I: The southern Levant. In Ahmad AlJallad (ed.), Arabic in context: Celebrating 400 years of Arabic at Leiden University, 99–186. Leiden: Brill.
Al-Jallad, Ahmad. 2018. What is Ancient North Arabian? In Daniel Birnstiel
(ed.), Re-engaging comparative Semitic and Arabic studies, 1–44. Wiesbaden:
Harrassowitz.
Asbaghi, Asya. 2011. Persian loanwords. In Lutz Edzard & Rudolf de Jong (eds.),
Encyclopedia of Arabic language and linguistics, online edn. Leiden: Brill.
78
3 Classical and Modern Standard Arabic
Beeston, Alfred F. L. 1962. A descriptive grammar of Epigraphic South Arabian.
London: Luzac.
Behnstedt, Peter. 2016. Dialect atlas of North Yemen and adjacent areas. Leiden:
Brill. (Translated by Gwendolin Goldbloom).
Beyer, Klaus. 1984. Die aramäischen Texte vom Toten Meer samt den Inschriften
aus Palästina, dem Testament Levis aus der Kairoer Genisa, der Fastenrolle und
den alten talmudischen Zitaten. Vol. 1. Göttingen: Vandenhoeck & Ruprecht.
Blau, Joshua. 1969. The renaissance of Modern Hebrew and Modern Standard
Arabic: Parallels and differences in the revival of two Semitic languages. Berkeley:
University of California Press.
Brockelmann, Carl. 1908. Grundriss der vergleichenden Grammatik der semitischen
Sprachen. Band I: Laut- und Formenlehre. Berlin: Reuther & Reichard.
Dichy, Joseph. 2011. Terminology. In Lutz Edzard & Rudolf de Jong (eds.), Encyclopedia of Arabic language and linguistics, online edn. Leiden: Brill.
Diem, Werner. 1980. Untersuchungen zur frühen Geschichte der arabischen
Orthographie, II: Die Schreibung der Konsonanten. Orientalia 49. 67–106.
Dillmann, August. 2005. Ethiopic grammar. 2nd edn, revised by Carl Bezold and
translated by James A. Crichton. Eugene, Oregon: Wipf & Stock. Originally
published in 1907 by Williams & Norgate.
Eilers, Wilhelm. 1962. Iranisches Lehngut im arabischen Lexikon: Über einige
Berufsnamen und Titel. Indo-Iranian Journal 5(3). 203–232.
Eilers, Wilhelm. 1971. Iranisches Lehngut im Arabischen. In Actas, IV Congresso
de Estudos Árabes e Islâmicos, Coimbra-Lisboa 1 a 8 de Septembro de 1968, 581–
657. Leiden: Brill.
Elmaz, Orhan. 2014. Investigating South Arabian words in al-Khalīl’s Kitāb alʿAyn. Supplement to the proceedings of the Seminar of Arabian Studies 44. 29–
42.
Elmaz, Orhan. 2016. ʿarim – A Sabaic word in the Qurʾān? In Roswitha G. Stiegner
(ed.), Süd-Arabien / South Arabia: A great “lost corridor” of mankind, vol. 1, 213–
244. Münster: Ugarit-Verlag.
Ferguson, Charles A. 1959. Diglossia. Word 15. 325–40.
Ferguson, Charles A. 1996. Epilogue: Diglossia revisited. In Alaa Elgibali (ed.),
Understanding Arabic: Essays in contemporary Arabic linguistics in honor of ElSaid Badawi, 49–67. Cairo: American University in Cairo Press.
Fischer, Wolfdietrich. 2002. A grammar of Classical Arabic. 3rd, revised edn.
Translated by Jonathan Rogers. New Haven: Yale University Press. Originally
published in 1972 by Harrassowitz.
Fraenkel, Siegmund. 1886. Die aramäischen Fremdwörter im Arabischen. Leiden:
Brill.
79
Marijn van Putten
Glaß, Dagmar. 2011. Creating a modern standard language from medieval tradition: The Nahḍa and the Arabic academies. In Stefan Weninger, Geoffrey
Khan, Michael P. Streck & Janet C. E. Watson (eds.), The Semitic languages: An
international handbook, 835–44. Berlin: De Gruyter Mouton.
Gutas, Dimitri. 1998. Greek thought, Arabic culture: The Graeco-Arabic translation
movement in Baghdād and early ʿAbbāsid society. London: Routledge.
Gutas, Dimitri. 2011. Greek loanwords. In Lutz Edzard & Rudolf de Jong (eds.),
Encyclopedia of Arabic language and linguistics, online edn. Leiden: Brill.
Gzella, Holger. 2015. A cultural history of Aramaic: From the beginnings to the
advent of Islam. Leiden: Brill.
Hayajneh, Hani. 2011. The usage of Ancient South Arabian and other Arabian
languages as an etymological source for Qurʾānic vocabulary. In Gabriel Said
Reynolds (ed.), New perspectives on the Qurʾān: The Qurʾān in its historical context 2, 117–146. London: Routledge.
Hebbo, Ahmed. 1984. Die Fremdwörter in der arabischen Prophetenbiographie des
Ibn Hischām (gest. 218/834). Frankfut am Main: Peter Lang.
Huehnergard, John. 1997. A grammar of Akkadian. Atlanta: Scholars Press.
Huehnergard, John. 2005. Features of Central Semitic. In Agustinus Gianto (ed.),
Biblical and Oriental essays in memory of William L. Moran, 155–203. Rome:
Pontificio Istituto Biblico.
Huehnergard, John & Aaron D. Rubin. 2011. Modern Standard Arabic. In Stefan Weninger, Geoffrey Khan, Michael P. Streck & Janet C. E. Watson (eds.),
The Semitic languages: An international handbook, 259–278. Berlin: De Gruyter
Mouton.
Ibn Muǧāhid, Abū Bakr. no date. Kitāb al-sabʕah fī al-qirāʔāt. Šawqī Ḍayf (ed.).
Cairo: Dār al-Maʕārif.
Ibrahim, Zeinab. 2009. Beyond lexical variation in Modern Standard Arabic: Egypt,
Lebanon and Morocco. Newcastle upon Tyne: Cambridge Scholars Publishing.
Jacquart, Danielle (ed.). 1994. La formation du vocabulaire scientifique et intellectuel dans le monde arabe. Turnhout: Brepols Publishers.
Jeffery, Arthur. 2007. The foreign vocabulary of the Qurʾān. Leiden: Brill. Originally published in 1938 by Benoytosh Bhattacharyya.
Kossmann, Maarten. 2010. Parallel system borrowing: Parallel morphological
systems due to the borrowing of paradigms. Diachronica 27(3). 459–487.
Kropftisch, Lorenz. 1977. Der französische Einfluß auf die arabische Schriftsprache im Maghrib. Zeitschrift der Deutschen Morgenländischen Gesellschaft
128. 39–64.
Leslau, Wolf. 1990. Arabic loanwords in Ethiopian Semitic. Wiesbaden:
Harrassowitz.
80
3 Classical and Modern Standard Arabic
Lucas, Christopher. 2015. Contact-induced language change. In Claire Bowern &
Bethwyn Evans (eds.), The Routledge handbook of historical linguistics, 519–536.
London: Routledge.
Mingana, Alphonse. 1927. Syriac influence on the style of the Ḳurʾān. John Rylands Library Bulletin 11. 77–98.
Nöldeke, Theodor. 1910. Neue Beiträge zur semitischen Sprachwissenschaft. Strasbourg: Karl J. Trübner.
Parkinson, Dilworth B. 2003. Future variability: A corpus study of Arabic future
particles. In Dilworth B. Parkinson & Samira Farwaneh (eds.), Perspectives on
Arabic linguistics XV: Papers from the fifteenth Annual Symposium on Arabic
linguistics, Salt Lake City 2001, 191–211. Amsterdam: John Benjamins.
Parkinson, Dilworth B. 2007. Sentence subject agreement variation in Arabic. In
Zeinab Ibrahim & Sanaa A. M. Makhlouf (eds.), Linguistics in an age of globalization: Perspectives on Arabic language and teaching, 67–90. Cairo: American
University in Cairo Press.
Parkinson, Dilworth B. & Zeinab Ibrahim. 1999. Testing lexical differences in regional standard Arabic. In Elabbas Benmamoun (ed.), Perspectives on Arabic
linguistics XII: Papers from the twelfth Annual Symposium on Arabic linguistics,
183–202. Amsterdam: John Benjamins.
Procházka, Stephan. 2011. Turkish loanwords. In Lutz Edzard & Rudolf de Jong
(eds.), Encyclopedia of Arabic language and linguistics, online edn. Leiden: Brill.
Rabin, Chaim. 1951. Ancient West-Arabian. London: Taylor’s Foreign Press.
Ratcliffe, Robert R. 1998. The broken plural problem in Arabic and comparative
Semitic: Allomorphy and analogy in non-concatenative morphology. Amsterdam: John Benjamins.
Retsö, Jan. 2011. Aramaic/Syriac loanwords. In Lutz Edzard & Rudolf de Jong
(eds.), Encyclopedia of Arabic language and linguistics, online edn. Leiden: Brill.
Rubin, Aaron D. 2010. The Mehri language of Oman. Leiden: Brill.
Ryding, Karin C. 2011. Modern Standard Arabic. In Stefan Weninger, Geoffrey
Khan, Michael P. Streck & Janet C. E. Watson (eds.), The Semitic languages: An
international handbook, 844–50. Berlin: De Gruyter Mouton.
Simeone-Senelle, Marie-Claude. 2011. Modern South Arabian. In Stefan
Weninger, Geoffrey Khan, Michael P. Streck & Janet C. E. Watson (eds.), The
Semitic languages: An international handbook, 1073–1113. Berlin: De Gruyter
Mouton.
Stein, Peter. 2011. Ancient South Arabian. In Stefan Weninger, Geoffrey Khan,
Michael P. Streck & Janet C. E. Watson (eds.), The Semitic languages: An international handbook, 1042–1073. Berlin: De Gruyter Mouton.
81
Marijn van Putten
Van Coetsem, Frans. 1988. Loan phonology and the two transfer types in language
contact. Dordrecht: Foris.
Van Coetsem, Frans. 2000. A general and unified theory of the transmission process
in language contact. Heidelberg: Winter.
van Putten, Marijn. 2018. Hamzah in the Quranic consonantal text. Orientalia
87(1). 93–120.
Weninger, Stefan. 2009. Der Jemen als lexikalisches Ausstrahlungszentrum in
der Antike. In Werner Arnold, Michael Jursa, Walter W. Müller & Stephan
Procházka (eds.), Philologisches und Historisches zwischen Anatolien und
Sokotra: Analecta semitica in memoriam Alexander Sima, 395–410. Wiesbaden:
Harrassowitz.
Weninger, Stefan. 2011a. Old Ethiopic. In Stefan Weninger, Geoffrey Khan,
Michael P. Streck & Janet C. E. Watson (eds.), The Semitic languages: An
international handbook, 1124–1142. Berlin: De Gruyter Mouton.
Weninger, Stefan. 2011b. Ethio-Semitic in general. In Stefan Weninger, Geoffrey
Khan, Michael P. Streck & Janet C. E. Watson (eds.), The Semitic languages: An
international handbook, 1114–1123. Berlin: De Gruyter Mouton.
Whorf, Benjamin Lee. 1956. Language, thought and reality: Selected writings of
Benjamin Lee Whorf. John B. Carroll (ed.). Cambridge, Massachusetts: Technology Press of MIT.
Wilmsen, David. 2010. Dialects of written Arabic: Syntactic differences in the
treatment of object pronouns in Egyptian and Levantine newspapers. Arabica
57. 99–128.
82
Chapter 4
Arabic in Iraq, Syria, and southern
Turkey
Stephan Procházka
University of Vienna
This chapter covers the Arabic dialects spoken in the region stretching from the
Turkish province of Mersin in the west to Iraq in the east, including Lebanon and
Syria. The area is characterized by a high degree of linguistic diversity, and for
about two and a half millennia Arabic has come into contact with various other
Semitic languages, as well as with Indo-European languages and Turkish. Bilingualism, particularly with Aramaic, Kurdish, and Turkish, has resulted in numerous contact-induced changes in all realms of grammar, including morphology and
syntax.
1 Current state and historical development
The region discussed in this chapter is linguistically extremely heterogeneous:
in it three different Arabic dialect groups, plus several other languages, are spoken. The two main Arabic dialect groups are Syrian and Iraqi, the distribution
of which does not exactly correspond to the political boundaries of those two
countries. Syrian-type dialects are also spoken in Lebanon, in three provinces
of southern Turkey (Mersin, Adana,1 Hatay), and in one village on Cyprus (cf.
Walter, this volume). In Iraq, Arabic is mainly spoken in Mesopotamia proper,
whereas considerable parts of the mountainous parts of the country are Kurdishspeaking. Arabic dialects which are very akin to the Iraqi ones extend into northeastern Syria and southeastern Anatolia (for the latter see Akkuş, this volume).
These two groups are geographically divided by a third dialect group, which arrived in the region with an originally (semi-) nomadic population from northern
1
The dialects spoken in Mersin and Adana provinces will henceforth referred to as Cilician
Arabic.
Stephan Procházka. 2020. Arabic in Iraq, Syria, and southern Turkey. In Christopher
Lucas & Stefano Manfredi (eds.), Arabic and contact-induced change, 83–114. Berlin:
Language Science Press. DOI:10.5281/zenodo.3744507
Stephan Procházka
Arabia. Today, this variety preponderates in all villages and most towns between
the eastern outskirts of Aleppo and the right bank of the Tigris, and stretching
north into the Turkish province of Şanlıurfa.
The total number of native Arabic speakers in the whole region is estimated
to be 54 million (see Table 1). The dialects of large urban centers like Beirut,
Damascus, Aleppo, and Baghdad have become supra-regional prestige varieties
that are also used in the media and therefore understood by most inhabitants of
the respective countries. The situation is very different in Turkey, where the local
Arabic is in sharp decline and public life is exclusively dominated by Turkish.
Only recently has the position of Arabic in Turkey been socially enhanced by the
influx of more than 3.5 million Syrian refugees fleeing the civil war that started
in 2011.2
Table 1: Speaker populations for dialects of Arabic
Country
Speakers
Syria
Lebanon
Iraq
Turkey
17,000,000
6,000,000
30,000,000
1,000,000
Arabic was spoken in the region long before the advent of Islam (Donner 1981:
95), but became the socially dominant language in the wake of the Muslim conquests in the seventh century CE. From that time until the end of the tenth
century, when Bedouin tribes seized large parts of central and northern Syria,
there was probably a continuum of sedentary-type dialects that stretched from
Mesopotamia to the northeastern Mediterranean (Procházka 2018: 291). During
the Mongol sacking of Iraq in 1258, much of the population was killed or expelled. This resulted in far-reaching demographic and linguistic changes as the
original sedentary-type dialects were only able to hold ground in Baghdad and
the larger settlements to its north. Further south they persisted only among the
non-Muslim population. Most of today’s Iraq was re-populated by people who
spoke Bedouin-type dialects (mostly coming from the Arabian Peninsula), which
over the centuries have heavily influenced the speech of even most large cities
(Holes 2007). Very similar dialects are spoken further south and in the Iranian
province of Khuzestan (see Leitner, this volume). The foundation of nation states
2
84
See UNHCR figures at https://data2.unhcr.org/en/situations/syria/location/113.
4 Arabic in Iraq, Syria, and southern Turkey
after World War One caused a significant decrease in contact between the different dialect groups and an almost complete isolation of the Arabic dialects spoken
in Turkey.
2 Contact languages
During its two-and-a-half-millennia presence in the region, Arabic has come into
contact with many languages, both Semitic and non-Semitic. Those most relevant for the topic will be treated in more detail below (for Syria, see also Barbot
1961: 175–177). Akkadian was spoken in southern Iraq until about the turn of
the eras, i.e. the first century CE.3 Greek was the language of administration in
Greater Syria until the Arab conquest (Magidow 2013: 185–187) and continued
to play a role for Orthodox Christians.4 During Crusader times, Arabic speakers in Syria came into contact with various medieval European languages; and
along the Mediterranean coast the so-called Lingua Franca (see Nolan, this volume) was an important source for the spread of particularly nautical vocabulary
for many centuries (Kahane et al. 1958). Since the nineteenth century, locally restricted contacts between Arabic and Armenian and Circassian have existed in
parts of Syria and Lebanon.
2.1 Aramaic
Aramaic is a Northwest Semitic language and thus structurally very similar to
Arabic. Different varieties of Aramaic were the main language in Syria and Iraq
from the middle of the first millennium BCE and it can be assumed that some
contact with Arabic existed even at that time. From the first century CE onwards,
the southern fringes of the Fertile Crescent became largely Arabic-dominant and
there was significant bilingualism with Aramaic, particularly in the towns along
the edge of the steppe, such as Petra, Palmyra, Hatra, and al-Ḥīra (Procházka 2018:
260–262). Though after the Muslim conquests Arabic eventually became the majority language, it did not oust Aramaic very quickly: the historical sources suggest that Aramaic dominated in the larger towns and the mountainous regions
of Syria and Lebanon for a long time. In Iraq, by contrast, the massive influx of
Arabs into the cities fostered their rapid Arabization, while Aramaic continued
to be spoken in the countryside (Magidow 2013: 184; 188). But over the centuries,
3
4
For Akkadian lexical influence on Arabic, see Holes (2002) and Krebernik (2008).
The enormous influence of Modern Greek on the Arabic spoken in the Kormakiti village of
Cyprus is discussed by Walter (this volume). For a detailed study, see also Borg (1985).
85
Stephan Procházka
the diverse Aramaic dialects became marginalized and, with very few exceptions,
were finally relegated to non-Muslim religious minorities, particularly Christians
and Jews, in peripheral regions like Mount Lebanon and the Anti-Lebanon Mountains, where Aramaic was prevalent until the eighteenth century (Retsö 2011).
Western Aramaic is still spoken in three Syrian villages, the best known of which
is Maaloula.5 There also remain speakers of Neo-Aramaic in northern Iraq.6
It is hard to establish the degree of bilingualism in the past, but it can be assumed that it was mostly Aramaic L1 speakers who had a command of Arabic
and not vice versa. In the present time, nearly all remaining Aramaic speakers in
Syria are fluent in Arabic. In Iraq this is mainly true of those living in the plain
just north of Mosul (Arnold & Behnstedt 1993; Coghill 2012: 86). The influence
of different strata of Aramaic on spoken Arabic is a long debated issue, various
scholars rating it from considerable to negligible (Hopkins 1995: 39; Lentin 2018:
199–204).
2.2 Persian and Kurdish
For many centuries, Arabic and the two Western Iranian languages Persian and
Kurdish have influenced each other on different levels. Persian-speaking communities existed in medieval Iraq, and economic and cultural contacts between
Mesopotamia and Iran have continued to the present (cf. Gazsi 2011). An important factor of language contact are the holy shrines of the Imams in Kerbela, Najaf,
and other Iraqi cities, which have always attracted tens of thousands of Persianspeaking Shiites every year. Intensive contacts between speakers of Kurdish and
Arabic have existed since at least the tenth century, particularly in Northern Iraq,
northeast Syria, and southeast Anatolia (see Akkuş, this volume). Until their exodus in the early 1950s, the Arabic-speaking Jewish communities which existed
in Iraqi Kurdistan usually had a native-like command of Kurdish (Jastrow 1990:
12). Due to the multilingual character of the region, bilingualism in Kurdish and
Arabic is still relatively widespread, particularly in urban settings, though with
Kurds usually much more fluent in Arabic than the other way around.7 However,
for obvious reasons, little linguistic research has been done in Iraq for decades,
which makes it impossible to give up-to-date information about the linguistic
situation in ethnically-mixed cities like Kirkuk.
5
The village heavily suffered from the jihadist occupation of 2013–2014, but after government
troops had retaken control over the region, many inhabitants returned and began its reconstruction (cf. the reports collected at http://friendsofmaaloula.de/).
6
See Coghill (2012) and http://glottolog.org/resource/languoid/id/nort3241.
7
With significant exceptions in some parts of southeast Anatolia; see Akkuş (this volume).
86
4 Arabic in Iraq, Syria, and southern Turkey
2.3 Ottoman and Modern Turkish
Contacts between spoken Arabic varieties and various Turkic languages existed
from the ninth century onwards. These early contacts, however, left hardly any
traces in Arabic except for a handful of loanwords. In the sixteenth century, the
Ottomans established their rule over most Arab lands, including Syria, Lebanon,
and Iraq. This domination lasted four hundred years, until World War One. Particularly in the provinces of Aleppo and Mosul, there was a relatively high percentage of Turkish speakers and probably a significant degree of bilingualism.8
As the language of the ruling elite, Turkish had high prestige and therefore was
at least rudimentarily spoken by many inhabitants of those regions, especially
urban men. The collapse of the Ottoman Empire put an abrupt end to Turkish–
Arabic contacts, which today remain intensive only among the Arabic varieties
spoken within the borders of Turkey itself, where most Arabic speakers are fluent in Turkish, the dominant language in all contact settings.
In some areas of Syria and in northern Iraq, the Arabic-speaking population
lives side by side with several hundred thousand speakers of Turkish and Azeri
Turkish, who call themselves Turkmens. Unfortunately, no reliable data on the
sociolinguistic settings and the degree of bilingualism exist for those areas. Again,
it can be assumed that most of the Turkmens in both countries are dominant in
Turkish, but use Arabic as a second language.
2.4 French and English
After World War One, Syria and Lebanon stayed under the French mandate and
Iraq under the British mandate until they reached independence.9 French is still
widely spoken as a second language in Lebanon, especially by Christians. In Iraq,
English has maintained its position as by far the most important foreign language
– a fact which was reinforced by the US military occupation from 2003 to 2010.
2.5 Intra-Arabic contacts
Contacts between different Arabic varieties, for instance between speakers of rural and urban dialects, happen on an everyday basis and often trigger short-term
accommodation without leading to long-lasting changes. The situation is different with regard to the enduring contacts between the Bedouin and the sedentary
8
See Wilkins (2010: xv) for Aleppo. Koury (1987: 103) maintains that Aleppo’s hinterland was
culturally even more Turkish than Arab. For Mosul, see Shields (2004: 54–55).
9
Iraq in 1932, Lebanon in 1943, Syria in 1946.
87
Stephan Procházka
populations, whose dialects differ from each other considerably.10 Such contacts
are most intense at the periphery of the Syrian steppe and along the middle Euphrates, where scattered towns with sedentary dialects like Palmyra, Deir ez-Zor
and Hit are surrounded by an originally nomadic population. Though the nomadic way of life has been abandoned by most of them, they still speak Bedouintype Arabic dialects. As the nomads were, for many centuries, both socially and
economically dominant, speakers of sedentary dialects often adopted linguistic
features from more prestigious Bedouin (though reverse instances are also found;
cf. Behnstedt 1994a: 421). Due to the historical circumstances mentioned in §1,
Bedouins also had a strong linguistic impact on Iraqi dialects. In Baghdad the
sedentary dialect of the Muslim population has been gradually Bedouinized due
to massive migration from the countryside to the city (Palva 2009). The Christian and, in former times, Jewish inhabitants, on the other hand, preserved their
original sedentary-type dialects because they had much less contact with the
Muslim newcomers.
3 Contact-induced changes
Change induced by contact with Aramaic almost exclusively happened through
imposition, that is, by Aramaic speakers who had learned Arabic as a second language and later often completely shifted to Arabic. This explains the relatively
numerous phonological changes and pattern replications in syntax. Lexical transfers from Aramaic certainly were also made by Arabic-dominant speakers, particularly in semantic fields like agriculture that included novel concepts for the
mostly animal-breeding Arabs.
The same is true for transfers from Greek, for which a very low level of bilingualism can be assumed. Thus we find only matter replication (in the sense of
Sakel 2007) in the form of loanwords, mostly in domains where lexical gaps in
older layers of spoken Arabic are likely.
In the case of transfer from Kurdish, bilingualism is much more widespread
among speakers of the source language, suggesting imposition. This might explain some of the phonological changes discussed in §3.1.2, as speakers dominant in the source language tend to preserve its phonological features (Lucas
2015: 532). The relatively small number of instances of lexical matter replication
is probably the result of the fact that Arabic has long been regarded as the more
prestigious by speakers of both the source and the recipient language.
10
88
Since these two speech communities differ from each other in so many ways, it is a relatively
robust approach to rate the following features as results of dialect contact and not mere variation (cf. Lucas 2015: 533).
4 Arabic in Iraq, Syria, and southern Turkey
The numerous loanwords from Persian into Iraqi Arabic may well be the result of matter replication by agents who were dominant in the recipient language
Arabic. Starting with the rule of the Abbasid caliphs in the eighth century CE
and continuing to the present, Iranian material culture and cuisine often had a
great impact on neighbouring Mesopotamia. There were also many intellectuals, among them praised writers of Arabic prose, who were actually Iranians and
hence knew both languages. Frequent contacts on the everyday level caused additional borrowing of ordinary vocabulary and the retention of sounds that are
replaced in Persian loans found in Classical Arabic or other dialects.11
Changes induced by contact with Ottoman Turkish may have happened mostly through Arabic-dominant speakers. The current situation of Arabic speakers
in Turkey is, however, very different, because at least the last two generations
have acquired Turkish as an L2 or even as a second L1 at very young age. Thus,
at least some of the contact phenomena described in the following paragraphs
may be examples of linguistic convergence (see Lucas 2015: 525).
French and English have largely remained typical “foreign languages” learned
at school or in business with a considerable amount of bilingualism only in some
urban settings of Lebanon, particularly Beirut. The agents of change are certainly
dominant in the recipient language.
The distinction between the two transfer types is not always clearly discernible in case of intra-Arabic contact-induced changes. In the towns of the Syrian
steppe and the middle Euphrates the agents of change were mostly the sedentary population who adapted their speech towards the norms of the socially
more prestigious Bedouin. However, there has always been inter-marriage, and
Bedouins often settled in towns and may well have adopted features from the
local sedentary variety. Especially in cases like Muslim Baghdadi (see §1), we
may assume with good reason that the Bedouin character of today’s variety developed through both imposition and borrowing.
3.1 Phonology
3.1.1 Aramaic-induced changes
It has been hypothesized that several phonological features of the Syrian and
Lebanese dialects are due to the contact-induced influence of Aramaic. But in
the case of the shift from interdental fricatives to postdental plosives (/ð/ > /d/;
/θ/ > /t/; /ð̣/ > /ḍ/) this is unlikely because: (i) this sound change is common
crosslinguistically; (ii) it does not occur in all dialects of the region; and (iii) it is
found in many other Arabic dialects without an Aramaic substrate.
11
The phonological changes are not, however, only the result of Persian influence (cf. §3.1.2).
89
Stephan Procházka
A phonotactic characteristic of most dialects spoken along the Mediterranean,
from Cilicia in the north to Beirut in the south, is that all unstressed short vowels
(including /a/) in open syllables are elided,12 whereas in other dialects east of
Libya only /i/ and /u/ in this position are consistently dropped.
(1)
Cilician Arabic (Procházka 2002a: 31–32; 130)
Old Arabic (OA) *raṣāṣ > rṣāṣ ‘lead, plumb’
OA *miknasa > mikinsi ‘broom’13
*fataḥ-t > ftaḥt ‘I opened’
Because this rule corresponds to the phonotactics of Aramaic and is otherwise
not found to the same degree except in Maghrebi dialects (cf. Benkato, this volume), pattern replication is likely, though cannot be proved.14
In roughly the same region, except Cilicia and many dialects of Hatay,15 the
diphthongs /ay/ and /aw/ are only preserved in open syllables, but monophthongized to /ē/ and /ō/ respectively in closed syllables. In some regions, for instance on
the island of Arwad, both diphthongs merge to /ā/ in closed syllables (Behnstedt
1997: map 31).
(2) Arwad, western Syria (Procházka 2013: 278)
OA *bayt, *baytayn > bāt, baytān ‘house, two houses’
OA *yawm, *yawmayn > yām, yawmān ‘day, two days’
OA *bayn al-iθnayn > bān it-tnān ‘between the two’
Likewise, in older layers of Aramaic, diphthongs were usually monophthongized in closed syllables (for Syriac see Nöldeke 1904: 34), which makes imposition
by L1 speakers of Aramaic rather likely (Fleisch 1974a: 227).
Another striking phenomenon is the split of historical /ā/ into /ō/ and /ē/ that
is found in scattered areas of the Levant, particularly northern Lebanon, around
the Syrian port of Tartous, the Qalamūn Mountains, and the exclusively Christian
town of Maḥarde on the Orontes River.16 Because in many varieties of Aramaic
the old Semitic /ā/ is reflected as /ō/, it could be assumed that Aramaic speakers transferred their peculiar pronunciation to Arabic when learning it. Fleisch
12
Therefore, Cantineau (1960: 108) called them parlers non différentiels – a term still very often
applied in Arabic dialectology – as they make no distinction in the treatment of the three short
vowels.
13
With insertion of an epenthetic /i/ to avoid a sequence of three consonants.
14
Cf. Diem (1979: 47); Arnold & Behnstedt (1993: 69–71); Weninger (2011: 748).
15
Where this phenomenon occurs only in Alawi villages (Arnold 1998: 84).
16
For details cf. Behnstedt (1997: map 32). The conditioned shift /ā/ > /ō/ is also found in and
around Tarsus in Turkey (Procházka 2002a: 37–38).
90
4 Arabic in Iraq, Syria, and southern Turkey
(1974b: 49) rejected the hypothesis of an Aramaic influence, arguing that the conditioned distribution of the two allophones is merely a further development of the
[ɒ] : [æ] split widely attested for Lebanon and parts of western Syria. However,
in the Syrian Qalamūn Mountains there are dialects with an unconditioned shift
(Behnstedt 1992), and this is precisely the region where the shift from Aramaic to
Arabic occurred relatively late, probably after a long phase of bilingualism. In the
town of Nabk, for instance, one can infer that the former Aramaic speaking inhabitants would have simply turned every /ā/ into /ō/ – except those which long
before had become [ē] (or [ɛ̄]) as a result of the so-called conditioned imāla (i.e.
the tendency of long /ā/ to be raised towards [ē] or even [ī] if the word contains
an /i/ or /ī/).17 Example (3) clearly shows that the distribution of the allophones
is not conditioned by the consonantal environment.
(3)
Nabk, Syria (Gralla 2006: 20)
OA *ṭābiḫ > ṭɛ̄beḫ ‘cooking’ vs. OA *ṭālib > ṭōleb ‘student’
OA *ḥāmil > ḥɛ̄mel ‘pregnant’ vs. OA *ḥāmiḍ > ḥōmeḍ’ ‘sour’
In these cases Aramaic influence seems plausible. For the region of Tripoli it
may be assumed that Aramaic bilinguals from the adjacent mountains used [ō]
instead of [ā] when speaking Arabic and thus reinforced the already existing [ɒ]
: [æ] split.18
3.1.2 The “new” phonemes /č/, /g/, and /p/
Consonantal phonemes that are originally alien to Arabic are found in all Arabic dialects spoken in Turkey (see also Akkuş, this volume), northern Syria, and
Iraq. These are the unvoiced affricate /č/, the voiced /g/,19 and the unvoiced /p/,
the latter mainly used in Iraq. The emergence of these sounds was very likely
contact-induced, but it is often impossible to discern which language triggered
each development: all three sounds are found in Persian, Kurdish, Turkish, and
the Lingua Franca. For the dialects of Cilicia, Hatay and Syria, the main source
language doubtless was Turkish. The sound /p/ in the Iraqi dialects was probably
first introduced through contact with Persian and Kurdish, and then reinforced
17
Cf. Arnold & Behnstedt (1993: 68).
For discussion see Fleisch (1974b: 48–50; 1974a: 133–136), Diem (1979: 45–46); Behnstedt (1992);
Arnold & Behnstedt (1993: 67–68); Weninger (2011: 748).
19
The sound [g] is prevalent in whole Syria and Lebanon but seems to have phonemic status
only in the north (Sabuni 1980: 26). For further examples and discussion see Ferguson (1969).
This “foreign” /g/ must therefore be differentiated from the /g/ which is the regular reflex of
OA *q. The latter development is found in many Bedouin-type dialects.
18
91
Stephan Procházka
by Ottoman Turkish. In the Bedouin-type dialects of the region, the phonemes
/č/ and /g/ are not products of contact-induced change but occur due to internal
sound changes, unvoiced /č/ as a conditioned affricated variant of /k/ and /g/ as
the ordinary reflex of OA *q.
Thus, it can be assumed that over the centuries speakers of the sedentary dialects of Iraq and Syria borrowed either from other languages or from Bedouin
Arabic varieties words that possess these two sounds, which subsequently were
fully incorporated into the phonemic inventory. This development may have
been facilitated by the fact that the three sounds /č/, /p/, and /g/ are not fundamentally unfamiliar to Arabic, but are the voiceless/voiced counterparts of the
well-established phonemes /ǧ/, /b/, and /k/. It seems no accident that the new
sound /č/ is much more often found in dialects that have preserved the affricate
/ǧ/ than in those where it has shifted to /ž/, as illustrated in examples (4) and (5).
(4)
Aleppo (Sabuni 1980: 205–210)
čanṭāye ‘handbag’ (Turkish çanta)
čwāl ‘sack’ (Turkish çuval)
čāy ‘tea’ (Turkish çay)
gaǧaleg ‘nightgown’ (Turkish gecelik)
The words given in (4) are usually pronounced with [š] instead of [č] in the
central Syrian and Lebanese dialects where contact with Turkish was less intense
and /ǧ/ is reflected as /ž/.20
(5) Mosul (own data)
ṣūč ‘fault’ (Turkish suç)
pāča ‘stew of sheep and cow legs and innards’ (Kurdish/Persian pāče)
zangīn ‘rich’ (Turkish zengin)
Once integrated into the phonological system, these sounds not only enabled
easier integration of loanwords from other languages like French and English
(see §2.4), but sometimes also resulted in the spread of assimilation-induced allophones from single words to the whole paradigm or even root. In Aleppo one
finds *yəkdeb > yəgdeb ‘he lies’, due to assimilation. The g subsequently was
transferred to other words derived from the root: gadab ‘he lied’, gədbe ‘lie’, and
gaddāb ‘liar’ (Sabuni 1980: 26, 209).
20
92
Cf. Behnstedt (1997: maps 18, 19, 25). For details and more examples see Sabuni (1980: 205–210),
who lists all words with č/g in Aleppo, and Procházka (2002b: 185) for Cilician Arabic.
4 Arabic in Iraq, Syria, and southern Turkey
Speakers of sedentary dialects who had everyday contact with Bedouins – for
example the inhabitants of Deir ez-Zor and Khatuniyya – first integrated /č/ and
/g/ into their phonemic inventory through the borrowing of typically Bedouin
vocabulary such as dabča ‘a Bedouin dance’ (Khawetna; Talay 1999: 29) and ṭabga
‘milk-bowl’ (Soukhne; Behnstedt 1994b: 310). These sounds then entered other
fields of the lexicon, which led to unpredictable distribution, including doublets,
as in (6)–(8).
(6) Khawetna (Talay 1999: 28–31)
gəṣṣa ‘forehead’, but qəṣṣa ‘story’ (OA *quṣṣa / *qiṣṣa)
dīč ‘rooster’ (OA *dīk)
(7)
Deir iz-Zor (Jastrow 1978: 42–43).
gāʕ ‘soil’ (OA *qāʕ)
čam ‘how much?’ (OA *kam)
(8)
Baghdad (Palva 2009: 18–19)
guffa ‘large basket’ (OA *quffa), but quful ‘lock’ (OA *qufl)
ʕigab ‘to pass’, but ʕiqab ‘to follow’ (both OA *ʕaqab)
The opposition /k/ : /č/ has even entered morphology, particularly with the 2sg
suffixes: ʔabū-k ‘your (sg.m) father’ vs. ʔabū-č ‘your (sg.f) father’. In the Syrian
oasis of Soukhne, long-term contact with speakers of Bedouin dialects caused a
chain of phonetic changes: first /k/ shifted to /č/, which originally was the reflex
of OA /ǧ/; then /č/ (< /ǧ/) shifted further to /ts/, which has become a unique
feature of the local dialect. The unconditioned shift from /k/ > /č/, which is not
found in the Bedouin dialects, in turn caused a shift from /q/ > /k/.21
(9)
Soukhne (Behnstedt 1994b: 226, 344, 357, 360)
kirbi ‘water-skin’ (< OA *qirba, Bedouin girba)
čalb ‘dog’ (< OA *kalb, Bedouin čalib)
čurr ‘donkey foal’ (< OA *kurr, Bedouin kuṛṛ)
tsubn ‘cheese’ (< OA *ǧubn, Bedouin ǧubun)
3.2 Morphology
3.2.1 Diminutive
The Aramaic diminutive suffix -ūn has become restrictedly productive in Iraqi
Arabic (Masliyah 1997: 72), as illustrated in (10). In Syria and Lebanon it is only
21
See Behnstedt (1994b: 4–11) for details.
93
Stephan Procházka
found in fossilized forms such as šalfūn ‘young cockerel’ and qafṣūne ‘little cage’.
Such kinds of morphological transfer are usually triggered by lexical borrowing.
Thus, it may be assumed that this suffix spread from loanwords like šalfūne ‘small
knife blade’ < Aramaic šelpūnā ‘little knife’ (cf. Féghali 1918: 82).22
(10) Iraq (Masliyah 1997: 72)
darb ‘road’ > darbūna ‘alley’
gṣayyir ‘short’ < gṣayyrūn ‘very short’
mḥammdūn hypocoristic form of the name Muḥammad
3.2.2 Morphological templates
Syrian and Lebanese dialects exhibit a few word patterns (templates) that are
attested for OA (and other dialects) but seem to have become widespread through
contact with Aramaic due to their frequency in the latter. These are the verbal
pattern šaC1 C2 aC3 and the (primarily diminutive) nominal patterns C1 aC2 C2 ūC3
and C1 aC2 C3 ūC4 .23
An example of the first is šanfaḫ ‘to puff up’, related to nafaḫ ‘to blow up’
(Féghali 1918: 83; cf. Lentin 2018: 201 for further discussion); the nominal forms
are illustrated in (11) and (12).
(11) Aleppo (Barthélemy 1935: 104, 158, 851)
ǧaḥḥūš ‘little donkey’ (related to ǧaḥš ‘young donkey’)
ḥassūn ‘goldfinch’ (related to the personal name ḥasan)
namnūme ‘small louse’ (cf. naml ‘ants’)
The pattern C1 aC2 C2 ūC3 (i) is still productive in the whole region, including
the Bedouin dialects, to derive hypocoristic forms from personal names:
(12)
fāṭma > faṭṭūma
ḥalīme > ḥallūma
aḥmad/mḥammad > ḥammūdi
3.2.3 Pronouns
In all Syrian and Lebanese dialects, as well as in Anatolia, the 2pl and 3pl pronouns exhibit an /n/ in place of the /m/ that is found in other Arabic dialects,
which makes them look as if they were reflexes of OA feminine forms (Table 2).
22
This must be a very old borrowing because the suffix is also found in the Gulf dialects (e.g.
ḥabbūna ‘a little’ Holes 2002: 279) and even in Tunisian Arabic (Singer 1984: 496), where direct
Aramaic influence can be excluded.
23
For the latter two see Corriente (1969) and Procházka (2004).
94
4 Arabic in Iraq, Syria, and southern Turkey
Table 2: 2pl and 3pl pronouns
2pl
3pl
Damascus
Jerusalem
OA pl.f
Syriac pl.m
ʔəntu / -kon
hənne(n) / -hon
ʔintu/ -kom
humme / -hom
ʔantunna / -kunna
hunna / -hunna
ʔatton / -kon
hennon / -hon
Because generalization of the feminine is unlikely,24 these forms have often
been explained as a contact-induced change. In Aramaic the corresponding pronouns also have /n/ (for Syriac see Muraoka 2005: 18). In particular, the 3rd person forms with final -n exactly mirror the Aramaic pattern, but lack a plausible
intra-Arabic etymology. Thus imposition seems plausible. Nevertheless, substratum influence has been doubted, particularly because of the infrequent evidence
of n-pronouns in other regions.25
3.2.4 Vocative suffixes
The suffixes -o (in the west of the region) and -u (in the east) can be attached
to various kinship terms and given names when used for direct address, usually
hypocoristically.26
(13)
Urfa (own data)
šnōnak ḫayy-o? ‘Brother, how are you?’
ǧidd-o ‘Grandfather!’
ʕamm-o ‘(paternal) Uncle!’
ḫāl-o ‘(maternal) Uncle!’
In Syria the suffix is also added to female nouns: ʕamm-t-o ‘(paternal) Aunt!’
and ḫāl-t-o ‘(maternal) Aunt!’, whereas in Iraq the corresponding forms end in
-a: ʕamm-a, ḫāl-a.
Since this suffix has no overt Arabic etymology, it has been assumed to be a
borrowing of the Kurdish vocative -o (e.g. Grigore 2007: 203). The Persian suffix -u also forms affective diminutives,27 which would make Persian influence
24
This is mainly because the feminine forms are only used for addressing groups of females,
whereas the masculine forms may also refer to a mixed group. Therefore, the masculine forms
are certainly more frequent. In all Arabic dialects except those mentioned above, the genderneutral plural forms are clearly derived from the historical masculine.
25
See Owens (2006: 244–245) and Procházka (2018: 283–284) for details.
26
See also Ferguson (1997: 187).
27
E.g. pesar-u ‘kid’; ʕamm-u is even the common word for ‘uncle’ (Perry 2007: 1011).
95
Stephan Procházka
possible, at least for Iraq.28 However, the distribution of this feature extends far
beyond even indirect contact with Kurdish or Persian,29 though reinforcement
and influence on the phonology may be possible for certain regions. Similar endings in Aramaic (Fassberg 2010: 88–89) and Ethiopian (Brockelmann 1928: 122)
suggest a common Semitic origin (see also Pat-El 2017: 463–465).
3.2.5 Turkish derivational suffixes
All dialects of the region have incorporated the Turkish suffix -ci [ʤi] into their
nominal morphology, as illustrated in (14) and (15). This suffix has become productive and is therefore a good example of morphological matter borrowing (Gardani et al. 2015). It is widely used for expressing professions, occupations, and
habitual actions – the latter overwhelmingly pejorative, or at least humorous. In
Iraqi dialects the suffix is reflected as -či, which corresponds to its pronunciation in the regional Turkish varieties. In the other varieties, it follows the usual
development of *ǧ, which means that it is realized as -ǧi or -ži.
(14) Syria/Damascus (own data)
kahrab-ži ‘electrician’ (kahraba ‘electricity’)
nəswān-ži ‘womanizer’ (nəswān ‘women’)
maškal-ži ‘troublemaker’ (məšəkle ‘problem’)
(15)
Iraq (Procházka-Eisl 2018: 40–44)
pančar-či ‘tire repairman’ (pančar ‘puncture’)
mharrib-či ‘human trafficker’ (mharrib ‘one who helps s.o. to escape’)
ʕarag-či ‘drunkard’ (ʕarag ‘aniseed brandy’)
The suffix clearly fills a morphological gap, because it enables morphologically transparent derivation even from loanwords, by preserving the basic, immediately recognizable word – in contrast to the Arabic C1 aC2 C2 āC3 pattern or
participles, which are derived from the root (for details see Procházka-Eisl 2018).
To a lesser extent other Turkish suffixes have enhanced the morphological
devices of the dialects treated here,30 specifically the relative suffix -li, the privative suffix -siz, and the abstract suffix -lik, which is reflected as -loɣiyya in Iraq,
28
In the Iraqi dialects the vowel is -u, e.g. ʕamm-u, ḫāl-u and ǧidd-u (Abu-Haidar 1999: 145).
The suffix is, for instance, attached to given names for endearment in the Gulf dialects, cf.
Holes (2016: 128). The address forms ya ʕamm-u, ya ḫāl-u ‘uncle’, gidd-u ‘grandfather’, sitt-u
‘grandmother’ are used in Cairo, where hypocoristic variants of given names are likewise attested, e.g. mīšu for hišām (Woidich 2006: 109). The suffix -o/-u in address forms is also attested
in eastern Sudan (Stefano Manfredi, personal communication), and in the Maghreb; Prunet &
Idrissi (2014: 184) provide a list of such nouns for Morocco.
30
See Halasi-Kun (1969: 68–71); Sabuni (1980: 168); Masliyah (1996); Procházka (2002a: 186).
29
96
4 Arabic in Iraq, Syria, and southern Turkey
i.e. with the Arabic abstract morpheme affixed. For the most part these suffixes
appear in Turkish loanwords, e.g. Cilicia ṣiḥḥat-li (< Turkish sıhhatlı) ‘healthy’,
raḥaṭ-ṣīz (< Turkish rahatsız) ‘uncomfortable’. Only in Iraq have they gained a
certain degree of productivity, particularly -sizz and -loɣiyya:
(16)
Iraq (Masliyah 1996: 293–294)
muḫḫ-sizz ‘stupid, brainless’
ḥaya-sizz ‘shameless’
ḥaywān-loɣiyya ‘ignorance’ (lit. ‘animal-ness’)
zmāl-loɣiyya ‘stupidity’ (lit. ‘donkey-ness’)
3.2.6 Light-verb constructions
Arabic dialects spoken in Turkey not infrequently use light-verb constructions
(in Turkish grammar mostly called phrasal verbs) which consist of the verb ‘to
do’ plus a following noun (see also Akkuş, this volume). Such compound verbs
are very frequent in Turkish (and Kurdish) and enable easy integration of foreign
vocabulary into the verbal system. The light verbs found in the Arabic dialects
show that this formation is a case of selected pattern replication because, first,
not all examples are exact copies of the Turkish model, and second, the word
order follows the Arabic VO rather than the Turkish OV pattern:
(17)
Harran–Urfa (own data)
sāwa qaza (Turkish kaza yapmak) ‘to have an accident’
sāwa ʕēš (in Turkish not a phrasal verb, but pişirmek) ‘to cook’
(18)
Cilician Arabic (Procházka 2002a: 198)
sawwa zarar (Turkish zarar vermek) ‘to harm’
sawwa ḫayir (Turkish hayır işlemek) ‘to do a good deed’
3.2.7 Intra-Arabic dialect contact
Concerning intra-Arabic contact, here we see that this has led to the adoption
of typical Bedouin-type pronouns into sedentary dialects (cf. Palva 2009: 27–29),
e.g.:
(19) Baghdad, Deir ez-Zor, Soukhne
ʔəḥna for nəḥna 1pl
(20)
Baghdad
ʔāni for ʔana 1sg
97
Stephan Procházka
In addition, as shown in Table 3, virtually all the eastern sedentary dialects of
Syria have copied the typical Bedouin-type active participles of the verbs ‘to eat’
and ‘to take’, which exhibit initial m- (Behnstedt 1997: map 175).
Table 3: Active participles of the verbs ‘to eat’ / ‘to take’
Bedouin
Soukhne
Palmyra
Damascus
māčil / māḫið
mīčil / mīḫið
mākil / māḫið
ʔākel / ʔāḫed
Finally, in a few places intensive mutual contact has resulted in an interdialect
(Trudgill 1986: 62) with completely new forms, such as the 3pl.m inflectional
suffix -a in the Syrian village of Ṣōrān (Behnstedt 1994a: 423–425), as shown in
Table 4.
Table 4: 3pl.m inflectional suffixes – ‘they said’
Bedouin
Sedentary
Ṣōrān
gāḷ-am
qāl-o
qāl-a
3.3 Syntax
3.3.1 Changes due to contact with Aramaic
3.3.1.1 Clitic doubling
In all but the Bedouin-type dialects of the region, two constructions exist which
both use an anticipatory pronoun and the preposition l- ‘to’: (i) a construction
involving analytical marking of a definite direct object, as in (21–23); and (ii) a
construction involving analytic attribution of a noun, as in (24). The frequency
and constraints of these two cases of clitic doubling show great variety, but in
general the usage of construction (i) is restricted to specific objects, particularly
elements denoting human beings, and construction (ii) is mostly found with inalienable possession, particularly kinship. A detailed discussion of both features
is found in Souag (2017).
98
4 Arabic in Iraq, Syria, and southern Turkey
(21)
Damascus (Berlinches 2016: 144)
ḥabbēt-o
la-ʕamər
love.prf.1sg-3sg.m to-Amr
‘I loved Amr.’
(22)
Baghdad, Christian (Abu-Haidar 1991: 116)
qaɣētū-nu
l-əl-əktēb
read.prf.1sg-3sg.m to-def-book
‘I read the book.’
(23) Cilician Arabic (ʕalā instead of l-; Procházka 2002a: 158)
biyḥibb-u
ʕala ḫāl-u
love.impf.ind.3sg.m-3sg.m on uncle-3sg.m
‘He loves his (maternal) uncle.’
(24)
Baghdad, Christian (Abu-Haidar 1991: 116)
maɣt-u
l-aḫū-yi
wife-3sg.m to-brother-obl.1sg
‘my brother’s wife’
Though the preposition l- is sometimes attested in Classical Arabic for introducing direct objects and is common even in Modern Standard Arabic for analytic
noun annexation, there are good arguments that the two constructions are pattern replications of an Aramaic model.31 For one thing, they do not have direct
parallels either in OA or in dialects which lacked contact with Aramaic. Example
(25) shows that both constructions have striking parallels in especially the later
eastern varieties of Aramaic (Rubin 2005: 94–104).
(25)
a. Syriac (Rubin 2005: 100)
bnā-y
l-bayt-ā
build.prf.3sg.m-3sg.m to-house-def
‘He built the house.’
b. Syriac (Hopkins 1997: 29)32
šm-ēh
l-gabr-ā
name-3sg.m to-man-def
‘the name of the man’
31
Not discussed here are two variants of construction (i), one without the suffix and the other
without the preposition (cf. Lentin 2018: 203). Among the many studies that are in favor of Aramaic influence are Contini (1999: 105); Blanc (1964: 130); and Weninger (2011: 750). Diem (1979:
47–49) and Lentin (2018) are more skeptical. Souag (2017: 52) suggests that at least “the initial
stages of the development of clitic doubling in the Levant derive from Aramaic substratum
influence, but the current situation also reflects subsequent Arabic-internal developments”.
32
The same pattern using the linker d- is more common.
99
Stephan Procházka
3.3.1.2 fī ‘can’
In the entire western part of the region including southern Turkey, the preposition fī ‘in’, together with a pronominal suffix, is used to express a capability, as
in (26). This has a striking parallel in the modern Aramaic ʔīθ b- ‘there is in’ ~
‘be able’ (Borg 2004: 52).
(26) Damascus (Cowell 1964: 415)
fī-ni sāʕd-ak
əb-kamm lēra
in-1sg help.impf.1sg-2sg.m with-some pound
‘Can I help you with a few pounds?’
3.3.1.3 Specific indefinite šī
A final example of possible Aramaic influence is the Syrian particle šī that mainly
indicates partial specifity, as in (27). It might be a pattern replication of the Western Neo-Aramaic form mett, used with the same function (Diem 1979: 49). What
reduces the likelihood of imposition by Aramaic speakers is the existence of a
cognate in Moroccan Arabic which is used with almost the same function.33
(27) Damascus (own data)
hnīk fī šī
ʕamūd
there exs indf column
‘There is some column.’
3.3.2 Changes due to contact with other languages
3.3.2.1 Indefiniteness
A hallmark of both sedentary and Bedouin-type Iraqi dialects is that reflexes of
the noun fard ‘individual (thing or person)’ are used to mark different kinds of indefiniteness (Blanc 1964: 118–119). The same form with the same indefinite article
function is found in in the Iranian province of Khuzestan, and in all Arabic speaking language islands of Central Asia, i.e. Khorasan, Uzbekistan, and Afghanistan,
as illustrated in (28).
(28)
33
Kirkuk (own data)
taʕrif-lak
fadd ṭabīb bāṭiniyye
know.impf.2sg.m-dat.2sg.m indf doctor internal
‘Do you know a doctor of internal medicine?’
Cf. Brustad (2000: 19, 26–27) and Wilmsen (2014: 51–53).
100
4 Arabic in Iraq, Syria, and southern Turkey
It is very likely that the noun fard has developed into a kind of indefinite article under the influence of other areal languages, particularly Turkish, Turkmen,
Persian, and Neo-Aramaic. However, in contrast to all contact languages, Iraqi
Arabic has not grammaticalized the numeral ‘one’ (wāḥəd), but fard. This clearly
indicates that this feature is a case of pattern replication. There are many parallels in the functions of the indefinite articles (such as marking pragmatic salience,
semantic individualization, approximation with numerals). Moreover, in all languages they are not fully systematized as a grammatical category as their usage
is often optional.
In the dialects of the Jews of Kurdistan the definite article is often omitted in
subject position – a flagrant imitation of the Kurdish model (see also Akkuş, this
volume, for some Anatolian dialects).
(29)
Kurdistan Arabic (Jastrow 1990: 71)
baʕdēn mudīra baʕatət
ḫalf-na
then director send.prf.3sg.f after-1pl
‘Then the director sent for us.’
3.3.2.2 m-bōr ‘because, in order to’
An interesting case of calquing which shows the difficulty of distinguishing between borrowing and imposition (see Manfredi, this volume) is the conjunction
m-bōr ‘because, in order to’. It exhibits both matter and pattern transfer, as it is a
copy of Kurdish ji ber (ku). In the actual form the Kurdish ji ‘from’ was replaced
by the Arabic equivalent m- (Jastrow 1979: 64).
3.3.2.3 Evidentiality
Syntactic change because of contact with Turkish is restricted to the Arabic dialects spoken in Turkey. In Cilicia and the Harran–Urfa region, active participles
express evidentiality, that is, they are used in utterances where a speaker refers to
second-hand information. As evidentiality is not a common category in Semitic,
it is very likely that the bilingual Arabic speakers of those regions copied this linguistic category from Turkish. In Turkish, any second-hand information is obligatorily marked by the verbal suffix -mış, whose second function besides evidentiality is to express stativity and perfectivity. The latter two functions are assumed
by the active participle in many Arabic dialects, including those in question here.
Thus, we can suppose that the stative/perfective function, which is shared by
both Arabic active participles and the Turkish suffix -mış, was likely the starting
101
Stephan Procházka
point of the development that led to the additional evidential function of Arabic
participles. The fact that evidentials seem to spread readily through language contact (Aikhenvald 2004: 10) makes Turkish influence even more probable.34 The
example in (30) illustrates how the speaker uses perfect forms for those parts of
the narrative he witnessed himself, and participles for secondhand information
(perfect forms italic, participles in bold face).
(30)
Harran–Urfa (Procházka & Batan 2016: 465)
ʔiḥne b-zimānāt čān ʕid-na ǧār b-al-maḥalle huwwa māt ərtiḥam
əngūl-lu šēḫ mǝṭar […] nahāṛ rabīʕ-u wāḥad ʕāzm-u ʕala stanbūl rāyiḥ
maʕzūm ʕala stanbul māḫið šēḫ mǝṭar əb-sāgt-u
‘Once we had a neighbor in our quarter. He died; he passed away. We
called him Sheikh Mǝṭar. One day somebody invited his friend to Istanbul.
As he was invited he went to Istanbul and he took Sheikh Mǝṭar with
him.’
3.3.2.4 Comparative and superlative
In most Arabic dialects that are spoken in Turkey, comparatives and superlatives
may be expressed by means of the Turkish particles daha and en, respectively,
followed by the simplex instead of the elative form of the adjective (cf. Akkuş, this
volume). As for comparatives, the use of such constructions is rather restricted,
while, at least in Cilician Arabic, they are relatively frequent for the superlative.
(31)
Harran–Urfa (own data)
daha zēn ṣārat
more good become.prf.3sg.f
‘It has become better.’
(32)
Cilician Arabic (Procházka 2002a: 155)
mīn en zangīl bi-d-dini
who sup rich in-def-world
‘Who is the richest (person) in the world?’
In Cilicia, comparison is often expressed by the elative pattern of an adjective,
which is preceded by the particle issa. This clearly reflects a calque: the Turkish
equivalent of the adverb issa ‘still, yet’ is daha, which in Turkish is also used as
the particle of the comparative.
34
For more examples and further details see Procházka (2002a: 200–201) for Cilicia, and
Procházka & Batan (2016: 464–465) for the Bedouin-type dialects in the Harran–Urfa region.
102
4 Arabic in Iraq, Syria, and southern Turkey
(33)
Cilician Arabic (Procházka 2002a: 202)
ṣāyir
issa aḥsan
become.ptcp more good.ela
‘It became better’.
(34)
Turkish
Daha iyi ol-du.
more good become-prf.3sg
‘It became better.’
3.3.2.5 Valency
Sometimes a change in verb valency occurs as a consequence of the copying of
Turkish models. A case found throughout these dialects is the verb ʕaǧab ‘to like’:
usually in Arabic the entity that is liked is the grammatical subject and the person
who likes something is the direct object of the verb; but in the Arabic dialects
in question, the construction of this verb reflects its Turkish (and English) usage
with the person doing the liking being the grammatical subject.
(35)
a. Cilicia (Procházka 2002a: 200)
ʕǧabt
bayt-ak
like.prf.1sg house-2sg.m
b. Damascus (own data)
bēt-ak
ʕažab-ni
house-2sg.m like.prf.3sg.m-1sg
‘I liked your house.’
3.4 Lexicon
Apart from the Aramaic loanwords also found in Classical Arabic (see Retsö 2011;
van Putten, this volume) – often in the realms of religion and cult – the dialects
of this region exhibit a large number of Aramaic lexemes. They are particularly
common in Lebanon and western Syria, but also found in Iraq and even in the
Bedouin-type dialects (Féghali 1918; Borg 2004; 2008). A large percentage of these
words belong to flora and fauna, agriculture, architecture, tools, kitchen utensils,
and other material objects:35
35
See also Neishtadt (2015: 282). Note that, unless otherwise indicated, lexemes cited in this
section are taken from Barthélemy (1935) for Syrian dialects, and Woodhead & Beene (1967)
and al-Bakrī (1972) for Iraqi dialects.
103
Stephan Procházka
(36) ṣumd ~ ṣimd ‘plough’ < Syriac ṣāmdē ‘yoke’
qālūz ‘bolt (of a door)’ < Syriac qālūzā
nāṭūr ‘guard (of a vineyard etc.)’ < Syriac nāṭūrā
šaṭaḥ ‘to spread’ < Syriac šeṭaḥ
šōb ‘heat, hot’ < Syriac šawbā
Many nautical terms and words denoting agricultural products and tools were
borrowed by Arabic from Greek, often via other languages, especially Aramaic,36
the Lingua Franca, and Turkish:
(37)
brāṣa < Greek práson ‘leek’
laḫana < Greek láḫana ‘cabbage’
dərrāʔen < Greek dōrákinon ‘peaches’
ʔabrīm/brīm ‘keel’ < Greek prýmnē ‘stern, poop’
sfīn < Greek sfēn ‘wedge’
Kurdish borrowings are mainly restricted to northern Iraq, where bilingualism
is widespread:
(38)
Mosul
pūš ‘chaff’ < Kurdish pûş
hēdi hēdi ‘slowly’ < Kurdish hêdî (Jastrow 1979: 68)
The intensive cultural and economic contacts between Iraq and Iran led to
many Persian loanwords in various domains of the Iraqi dialects.
(39) mēwa ‘fruit’ < Persian mīva ~ mayva
baḫat ‘luck’ < Persian baḫt
čariḫ ‘wheel’ < Persian čarḫ
gulguli ‘pink’ < Persian gol ‘rose’
yawāš ‘slow’ < Persian yavāš
puḫta ‘mush’ < Persian poḫte ‘(well) cooked’
Ottoman Turkish contributed a great deal to culinary vocabulary and the terminology of clothing and (technical) tools of Syria and Iraq.37 It was even the
source of several adverbs and even verbs in the local Arabic varieties (HalasiKun 1969; 1973; 1982).
36
This is especially true for words related to Christian liturgy and ritual, which constitute about
twenty per cent of the Greek vocabulary that entered the dialects of Syria.
37
The same loanwords are, of course, often found in other regions that were under Ottoman rule,
above all in Egypt, but also in Tunisia, Yemen and other regions.
104
4 Arabic in Iraq, Syria, and southern Turkey
(40)
Syria (Damascus)
šāwərma ‘shawarma’ < Turkish çevirme
ṣāž ‘iron plate for making bread’ < Turkish saç
yalanži ‘vine-leaves stuffed with rice’ < Turkish yalancı ‘liar’ (as they
pretend to be “real” dolma stuffed with meat)
šīš ṭāwūʔ ‘spit-roasted chicken’ < Turkish şiş tavuk
kǝzlok ‘glasses’ < Turkish gözlük
ʔūḍa ‘room’ < Turkish oda
ballaš ‘to begin’ < Turkish başla-mak by metathesis.
(41)
Iraq (Muslim Baghdadi, cf. Reinkowski 1995)
qūzi ‘a dish with roasted mutton’ < Turkish kuzu ‘lamb’
tēl ‘wire’ < Turkish tel
yašmāɣ ‘kerchief (for men)’ < Turkish yaşmak ‘veil (for women)’
bōš ‘empty; neutral’, which yielded also the verb bawwaš ‘to put into
neutral (gear)’ < Turkish boş ‘empty’
qačaɣ ‘smuggled goods’ < Turkish kaçak
During the last century, the Arabic dialects in Turkey38 have incorporated
numerous Turkish words in addition to loanwords from Ottoman times. Among
them are terms in education, medicine, sports, media, and technology. Besides
these, kinship terms, the vocabulary of everyday life, and structural words like
adverbs and discourse markers have infiltrated the dialects from Turkish.
(42)
Cilician Arabic
qāyin … ‘-in-law’ (< Turkish kayın)
ṭōrūn ‘grandchild’ (< Turkish torun)
bīle ‘even’ (< Turkish bile)
qāršīt ‘opposite from’ (< Turkish karşı)
The cases of semantic extension of an Arabic word result from the wider semantic range of its Turkish equivalent which has been transferred into Arabic.
Thus, in both Cilician and Harran–Urfa Arabic sāq/ysūq ‘to drive’ also occurs
with the meaning of ‘to last’ like the Turkish verb sürmek. In Harran–Urfa barð̣ ‘on the place/ground (of)’ has become a preposition/conjunction meaning
‘instead’. This can be seen as an instance of contact-induced grammaticalization
(Gardani et al. 2015: 4) under the influence of Turkish yerine ‘instead, in its place’.
38
For Cilicia see Procházka (2002a; 2002b: 187–199).
105
Stephan Procházka
(43) Harran–Urfa (own data)
al-mille
tākl-u
b-arð̣ al-laḥam
def-people eat.impf.3sg.f-3sg.m in-place def-meat
‘The people eat it instead of meat.’
(44)
Harran–Urfa (own data)
b-arð̣-in
tibči
ʔigir
āya
in-place-link cry.impf.2sg.m read.imp.sg.m verse
‘Instead of crying recite a (Koranic) verse!’
In Iraq, many English words related to Western culture and technology have
been, and still are, borrowed into the dialects. The same is true for French in
Syria and (particularly) Lebanon (cf. Barbot 1961: 176).
(45) Iraq (words of English origin)
kitli < kettle
buṭil < bottle
glāṣ < glass
pančar ‘flat tire’ (< puncture)
pāysikil < bicycle
māṭōrsikil < motorcycle
lōri < lorry
igzōz < exhaust (pipe)
brēk < brake
(46)
Syria and Lebanon (words of French origin)
gātto ~ gaṭō < gâteau ‘cake’
garsōn < garçon ‘waiter’
sēšwār < séchoir ‘hair drier’
kwaffēr < coiffeur ‘hair-dresser’
ʔaṣanṣēr < ascenseur ‘elevator’
grīb < grippe ‘influenza’
Due to long-term contacts, there are mutual borrowings between the Bedouin
and sedentary dialects of the region. This affects not only specific vocabulary
of the respective cultures but also basic lexical items. Historically, the sedentary
dialects have been much more influenced by the Bedouin-type dialects than vice
versa.
106
4 Arabic in Iraq, Syria, and southern Turkey
4 Conclusion
The sociolinguistic history of the regions treated here suggests that the conditions for imposition were relatively restrictive and mainly found in contact settings with Aramaic, which, over the centuries, has been given up by most of its
speakers in favor of Arabic. Thus, it is not surprising that so many features beyond the lexicon for which contact-induced change can be assumed are related
to Aramaic influence.
Morphological borrowing is in general relatively rare because it presupposes
a high intensity of contact (Gardani et al. 2015: 1). Practically all cases presented
in §3.2 corroborate the universal tendencies that: (i) derivational morphology
is more prone to borrowing than inflectional morphology; and (ii) nominalizers and diminutives are very frequently represented in instances of borrowed
derivational morphology (Gardani et al. 2015: 7; Seifart 2013). On the whole, the
Bedouin-type dialects exhibit significantly fewer contact-induced changes than
the sedentary dialects. This may be the result of both the Bedouin groups’ nomadic way of life at the fringes of the desert and their tribally organized society,
which impedes intense contact with outsiders.
The relative infrequency of contact-induced changes in morphology and syntax found in the Arabic varieties spoken in Turkey have two main explanations:
first, the high degree of complete bilingualism is a very recent phenomenon that
only pertains to the last two generations; and second, and probably more importantly, the great structural differences between the two languages, which have
impeded both matter and pattern replications.
What is still relatively unclear is the degree of historical bilingualism between
Arabic on the one hand and Ottoman Turkish, Kurdish, and Persian on the other.
Future research would be particularly desirable with regard to Iraq, providing
interesting new data on contact-induced changes in multilingual regions like
Mosul and Kirkuk, where Arabic, Turkmen, and Kurdish speakers have been in
contact for a long time. Also, studies like that of Neishtadt (2015) for Palestine
should be carried out for Syrian and especially Iraqi dialects with regard to lexical borrowings from Aramaic. Another completely under-researched topic is
idiomatic constructions, in which the mutual influence of most languages in the
region may be assumed.
107
Stephan Procházka
Further reading
There are no studies which treat the subject of contacts between Arabic and the
other languages of the whole region covered in this chapter. However:
) Arnold & Behnstedt (1993) is an in-depth study of the mutual contacts between
Western Neo-Aramaic and the local Arabic dialects in the Anti-Lebanon Mountains of Syria.
) Diem (1979) is a pioneer study of substrate influence in the modern Arabic
dialects, though with focus on South Arabia, i.e. outside of the region treated
in this chapter.
) Palva (2009) is a very good case study of the diachronic relations between
sedentary and Bedouin-type dialects in the Iraqi capital Baghdad.
) Weninger (2011) is a concise overview of contact between different varieties
of Aramaic and Arabic.
Acknowledgements
I am grateful to my colleagues Bettina Leitner and Veronika Ritt-Benmimoun for
their valuable comments on earlier drafts of this paper. I warmly thank Jérôme
Lentin for extensive discussion of the possible origin of the hypocoristic -o suffix
(§3.2.4) and his help in finding important sources.
Abbreviations
1, 2, 3
BCE
CE
comp
def
f
ela
exs
imp
impf
indf
108
1st, 2nd, 3rd person
before Common Era
Common Era
complementizer
definite
feminine
elative degree
existential
imperative
imperfect (prefix conjugation)
indefinite
L1
L2
link
m
OA
obl
pl
prf
sg
sup
first language
second language
linker
masculine
Old Arabic
oblique
plural
perfect (suffix conjugation)
singular
superlative
4 Arabic in Iraq, Syria, and southern Turkey
References
Abu-Haidar, Farida. 1991. Christian Arabic of Baghdad. Wiesbaden: Harrassowitz.
Abu-Haidar, Farida. 1999. Vocatives as exclamatory nouns in Iraqi Arabic. In Yasir
Suleiman (ed.), Arabic grammar and linguistics, 144–160. London: Routledge.
Aikhenvald, Alexandra Y. 2004. Evidentiality. Oxford: Oxford University Press.
al-Bakrī, Ḥāzim. 1972. Dirāsāt fī l-alfāð̣ al-ʕāmmiyya al-mawṣiliyya. Baghdad:
Maṭbaʕat Asʕad.
Arnold, Werner. 1998. Die arabischen Dialekte Antiochiens. Wiesbaden:
Harrassowitz.
Arnold, Werner & Peter Behnstedt. 1993. Arabisch-aramäische Sprachbeziehungen
im Qalamūn (Syrien). Wiesbaden: Harrassowitz.
Barbot, Michel. 1961. Emprunts et phonologie dans les dialectes citadins syrolibanais. Arabica 8(2). 174–188.
Barthélemy, Adrien. 1935. Dictionnaire arabe-français, dialectes de Syrie: Alep,
Damas, Liban, Jérusalem. Vol. 5. Paris: Paul Geuthner. Volumes 4–5 edited by
H. Fleisch.
Behnstedt, Peter. 1992. Zur Spaltung von *ā in ō/ē (ǭ/ę̄ ) im Qalamūn (Antilibanon).
Die Welt des Orients 23. 77–100.
Behnstedt, Peter. 1994a. Dialektkontakt in Syrien. In Dominique Caubet &
Martine Vanhove (eds.), Actes des premières journées internationales de dialectologie arabe de Paris, 417–430. Paris: Langues’O.
Behnstedt, Peter. 1994b. Der arabische Dialekt von Soukhne (Syrien). Teil 2: Phonologie, Morphologie, Syntax. Teil 3: Glossar. Wiesbaden: Harrassowitz.
Behnstedt, Peter. 1997. Sprachatlas von Syrien. Vol. I: Kartenband. Wiesbaden:
Harrassowitz.
Berlinches, Carmen. 2016. El dialecto árabe de Damasco (Siria): Estudio gramatical
y textos. Zaragoza: Prensas de la Universidad de Zaragoza.
Blanc, Haim. 1964. Communal dialects in Baghdad. Cambridge, MA: Harvard
University Press.
Borg, Alexander. 1985. Cypriot Arabic: A historical and comparative investigation
into the phonology and morphology of the Arabic vernacular spoken by the
Maronites of Kormakiti village in the Kyrenia district of north-west Cyprus.
Stuttgart: Deutsche Morgenländische Gesellschaft.
Borg, Alexander. 2004. A comparative glossary of Cypriot Maronite Arabic (Arabic–
English): With an introductory essay. Leiden: Brill.
Borg, Alexander. 2008. Lexical Aramaisms in nomadic Arabic vernaculars. In
Stephan Procházka & Veronika Ritt-Benmimoun (eds.), Between the Atlantic
and Indian Oceans: Studies on contemporary Arabic dialects. Proceedings of the
109
Stephan Procházka
7th AIDA Conference, held in Vienna from 5-9 September 2006, 89–112. Münster:
LIT-Verlag.
Brockelmann, Carl. 1928. Deminutiv und Augmentativ im Semitischen.
Zeitschrift für Semitistik und verwandte Gebiete 6. 109–134.
Brustad, Kristen. 2000. The syntax of spoken Arabic: A comparative study of
Moroccan, Egyptian, Syrian, and Kuwaiti dialects. Georgetown: Georgetown
University Press.
Cantineau, Jean. 1960. Etudes de linguistique arabe. Paris: Klincksieck.
Coghill, Eleanor. 2012. Parallels in the grammaticalisation of Neo-Aramaic ziland Arabic raḥ- and a possible contact scenario. In Domenyk Eades (ed.), Grammaticalization in Semitic, 127–144. Oxford: Oxford University Press.
Contini, Ricardo. 1999. Le substrat araméen en néo-arabe libanais: Préliminaires
à une enquête systématique. In Marcello Lamberti & Livia Tonelli (eds.), Afroasiatica Tergestina, 101–122. Padova: Unipress.
Corriente, Federico. 1969. Qalqūl en Semitico. Sefarad 29. 3–11.
Cowell, Mark W. 1964. A reference grammar of Syrian Arabic. Washington, DC:
Georgetown University Press.
Diem, Werner. 1979. Studien zur Frage des Substrats im Arabischen. Der Islam 56.
12–80.
Donner, Fred McGraw. 1981. The early Islamic conquests. Princeton: Princeton
University Press.
Fassberg, Steven Ellis. 2010. The Jewish Neo-Aramaic dialect of Challa. Leiden:
Brill.
Ferguson, Charles A. 1969. The /g/ in Syrian Arabic: Filling a gap in a phonological pattern. Word 25. 114–119.
Ferguson, Charles A. 1997. Arabic baby talk. In Kirk Belnap & Niloofar Haeri
(eds.), Structuralist studies in Arabic linguistics: Charles A. Ferguson’s papers,
1954–1994, 179–187. Leiden: Brill.
Féghali, Michel. 1918. Etude sur les emprunts syriaques dans les parlers arabes du
Liban. Paris: Champion.
Fleisch, Henri. 1974a. Le parler arabe de Kfar-Ṣghâb, Liban. In Henri Fleisch (ed.),
Etudes d’arabe dialectal, 221–262. Beirut: Dar el-Machreq.
Fleisch, Henri. 1974b. Le changement ā > ō dans le sémitique de l’ouest et en arabe
dialectal libanais. In Henri Fleisch (ed.), Etudes d’arabe dialectal, 45–50. Beirut:
Dar el-Machreq.
Gardani, Francesco, Peter Arkadiev & Nino Amiridze (eds.). 2015. Borrowed morphology. Berlin: De Gruyter.
110
4 Arabic in Iraq, Syria, and southern Turkey
Gazsi, Dénes. 2011. Arabic–Persian language contact. In Stefan Weninger,
Geoffrey Khan, Michael P. Streck & Janet C. E. Watson (eds.), The Semitic
languages: An international handbook, 1015–1021. Berlin: De Gruyter Mouton.
Gralla, Sabine. 2006. Der arabische Dialekt von Nabk (Syrien). Wiesbaden:
Harrassowitz.
Grigore, George. 2007. L’arabe parlé à Mardin: Monographie d’un parler arabe
“périphérique”. Bucharest: Editura Universitãţii din Bucureşti.
Halasi-Kun, Tibor. 1969. The Ottoman elements in the Syrian dialects. Archivum
Ottomanicum 1. 14–91.
Halasi-Kun, Tibor. 1973. The Ottoman elements in the Syrian dialects, II.
Archivum Ottomanicum 5. 17–95.
Halasi-Kun, Tibor. 1982. The Ottoman elements in the Syrian dialects, II (continuation). Archivum Ottomanicum 7. 117–267.
Holes, Clive. 2002. Non-Arabic Semitic elements in the Arabic dialects of eastern Arabia. In Otto Jastrow, Werner Arnold & Hartmut Bobzin (eds.), “Sprich
doch mit deinen Knechten aramäisch, wir verstehen es!”: 60 Beiträge zur Semitistik: Festschrift für Otto Jastrow zum 60. Geburtstag, 269–279. Wiesbaden:
Harrassowitz.
Holes, Clive. 2007. Colloquial Iraqi Arabic. In J. N. Postgate (ed.), Languages of
Iraq, ancient and modern, 123–134. Baghdad: British School of Archaeology in
Iraq.
Holes, Clive. 2016. Dialect, culture, and society in eastern Arabia. Vol. 3: Phonology,
morphology, syntax, style. Leiden: Brill.
Hopkins, Simon. 1995. ṣarār ‘pebbles’: A Canaanite substrate word in Palestinian
Arabic. Zeitschrift für Arabische Linguistik 30. 37–49.
Hopkins, Simon. 1997. On the construction šmēh l-gabrā ‘the name of the man’
in Aramaic. Journal of Semitic Studies 42. 23–32.
Jastrow, Otto. 1978. Die mesopotamisch-arabischen qəltu-Dialekte. Vol. 1: Phonologie und Morphologie. Wiesbaden: Steiner.
Jastrow, Otto. 1979. Zur arabischen Mundart von Mossul. Zeitschrift für Arabische
Linguistik 2. 76–99.
Jastrow, Otto. 1990. Die arabischen Dialekte der irakischen Juden. In Werner
Diem & Abdoldjavad Falaturi (eds.), XXIV. Deutscher Orientalistentag vom 26.
bis 30. September 1988 in Köln: Ausgewählte Vorträge, 199–205. Wiesbaden:
Steiner.
Kahane, Henry Romanos, Renée Kahane & Andreas Tietze. 1958. The Lingua
Franca in the Levant: Turkish nautical terms of Italian and Greek origin. Urbana,
IL: University of Illinois Press.
111
Stephan Procházka
Khoury, Philipp S. 1987. Syria and the French Mandate: The politics of Arab nationalism, 1920–1945. London: Tauris.
Krebernik, Manfred. 2008. Von Gindibu bis Muḥammad: Stand, Probleme und
Aufgaben altorientalistisch-arabistischer Philologie. In Otto Jastrow, Shabo Talay & Herta Hafenrichter (eds.), Studien zur Semitistik und Arabistik: Festschrift
für Hartmut Bobzin zum 60. Geburtstag, 247–279. Wiesbaden: Harrassowitz.
Lentin, Jérôme. 2018. The Levant. In Clive Holes (ed.), Arabic historical dialectology: Linguistic and sociolinguistic approaches, 170–205. Oxford: Oxford University Press.
Lucas, Christopher. 2015. Contact-induced language change. In Claire Bowern &
Bethwyn Evans (eds.), The Routledge handbook of historical linguistics, 519–536.
London: Routledge.
Magidow, Alexander. 2013. Towards a sociohistorical reconstruction of pre-Islamic
Arabic dialect diversity. Austin: University of Texas. (Doctoral dissertation).
Masliyah, Sadok. 1996. Four Turkish suffixes in Iraqi Arabic: -li, -lik, -siz and -çi.
Journal of Semitic Studies 41(2). 291–300.
Masliyah, Sadok. 1997. The diminutive in spoken Iraqi Arabic. Zeitschrift für
Arabische Linguistik 33. 68–88.
Muraoka, Takamitsu. 2005. Classical Syriac. 2nd, revised edn. Wiesbaden:
Harrassowitz.
Neishtadt, Mila. 2015. The lexical component in the Aramaic substrate of
Palestinian Arabic. In Aaron Butts (ed.), Semitic languages in contact, 280–310.
Leiden: Brill.
Nöldeke, Theodor. 1904. Neue Beiträge zur semitischen Sprachwissenschaft. Strasbourg: Trübner.
Owens, Jonathan. 2006. A linguistic history of Arabic. Oxford: Oxford University
Press.
Palva, Heikki. 2009. From qǝltu to gilit: Diachronic notes on linguistic adaptation
in Muslim Baghdad Arabic. In Enam Al-Wer & Rudolf De Jong (eds.), Arabic
dialectology in honour of Clive Holes on the occasion of his sixtieth birthday, 17–
40. Leiden: Brill.
Pat-El, Na’ama. 2017. Digging up archaic features: Neo-Arabic and comparative
Semitics in the quest for proto-Arabic. In Ahmad Al-Jallad (ed.), Arabic in context: Celebrating 400 years of Arabic at Leiden University, 441–475. Leiden: Brill.
Perry, John R. 2007. Persian morphology. In Alan S. Kaye (ed.), Morphologies of
Asia and Africa, 975–1019. Winona Lake, IN: Eisenbrauns.
Procházka, Stephan. 2002a. Die arabischen Dialekte der Çukurova (Südtürkei).
Wiesbaden.
112
4 Arabic in Iraq, Syria, and southern Turkey
Procházka, Stephan. 2002b. Contact phenomena, code-copying, and codeswitching in the Arabic dialects of Adana and Mersin (southern Turkey). In
Abderrahim Youssi (ed.), Aspects of the dialects of Arabic today: Proceedings of
the 4th conference of the International Arabic Dialectology Association (AIDA),
Marrakesh, April 1-4, 2000. In honour of Professor David Cohen, 133–139. Rabat:
Amapatril.
Procházka, Stephan. 2004. Unmarked feminine nouns in modern Arabic dialects.
In Martine Haak, Rudolf de Jong & Kees Versteegh (eds.), Approaches to Arabic
dialects: A collection of articles presented to Manfred Woidich on the occasion of
his sixtieth birthday, 237–282. Leiden: Brill.
Procházka, Stephan. 2013. Traditional boatbuilding: Two texts in the Arabic
dialect of the island of Arwād (Syria). In Renaud Kuty, Ulrich Seeger &
Shabo Talay (eds.), Nicht nur mit Engelszungen: Beiträge zur semitischen
Dialektologie. Festschrift für Werner Arnold zum 60. Geburtstag, 275–288.
Wiesbaden: Harrassowitz.
Procházka, Stephan. 2018. The northern Fertile Crescent. In Clive Holes (ed.),
Arabic historical dialectology: Linguistic and sociolinguistic approaches, 257–292.
Oxford: Oxford University Press.
Procházka, Stephan & Ismail Batan. 2016. The functions of active participles in
Šāwi Bedouin dialects. In George Grigore & Gabriel Biţună (eds.), Arabic varieties: Far and wide. Proceedings of the 11th International Conference of Aida,
Bucharest 2015, 457–466. Bucharest: Editura Universității din București.
Procházka-Eisl, Gisela. 2018. A suffix on the move: Forms and functions of the
Turkish suffix -ci in Arabic dialects. Mediterranean Language Review 25. 21–52.
Prunet, Jean-François & Ali Idrissi. 2014. Overlapping morphologies in Arabic
hypocoristics. In Sabrina Bendjaballah, Noam Faust, Mohamed Lahrouchi &
Nicola Lampitelli (eds.), The form of structure, the structure of form: Essays in
honor of Jean Lowenstamm, 177–192. Amsterdam: John Benjamins.
Reinkowski, Maurus. 1995. Eine phonologische Analyse türkischen Wortgutes im
Bagdadisch-Arabischen mitsamt einer Wortliste. Folia Orientalia 31. 89–115.
Retsö, Jan. 2011. Aramaic/Syriac loanwords. In Lutz Edzard & Rudolf de Jong
(eds.), Encyclopedia of Arabic language and linguistics, online edn. Leiden: Brill.
Rubin, Aaron D. 2005. Studies in Semitic grammaticalization. Winona Lake, IN:
Eisenbrauns.
Sabuni, Abdulghafur. 1980. Laut- und Formenlehre des arabischen Dialekts von
Aleppo. Frankfurt am Main: Peter Lang.
Sakel, Jeanette. 2007. Types of loan: Matter and pattern. In Yaron Matras, Georg
Bossong & Bernard Comrie (eds.), Grammatical borrowing in cross-linguistic
perspective, 15–29. Berlin: Mouton de Gruyter.
113
Stephan Procházka
Seifart, Frank. 2013. AfBo: A world-wide survey of affix borrowing. Leipzig: Max
Planck Institute for Evolutionary Anthropology. http://afbo.info.
Shields, Sarah. 2004. Mosul questions: Economy, identity, and annexation. In
Reeva S. Simon & Eleanor Harvey Tejirian (eds.), The creation of Iraq: 1914–
1921, 50–60. New York: Columbia University Press.
Singer, Hans-Rudolf. 1984. Grammatik der arabischen Mundart der Medina von
Tunis. Berlin: De Gruyter.
Souag, Lameen. 2017. Clitic doubling and language contact in Arabic. Zeitschrift
für Arabische Linguistik 66. 45–70.
Talay, Shabo. 1999. Der arabische Dialekt der Khawētna. Vol. 1: Grammatik.
Wiesbaden: Harrassowitz.
Trudgill, Peter. 1986. Dialects in contact. Oxford: Wiley-Blackwell.
Weninger, Stefan. 2011. Aramaic–Arabic language contact. In Stefan Weninger,
Geoffrey Khan, Michael P. Streck & Janet C. E. Watson (eds.), The Semitic languages: An international handbook, 747–755. Berlin: De Gruyter Mouton.
Wilkins, Charles L. 2010. Forging urban solidarities: Ottoman Aleppo 1640–1700.
Leiden: Brill.
Wilmsen, David. 2014. Arabic indefinites, interrogatives and negators: A linguistic
history of western dialects. Oxford: Oxford University Press.
Woidich, Manfred. 2006. Das Kairenisch-Arabische: Eine Grammatik. Wiesbaden:
Harrassowitz.
Woodhead, Daniel R & Wayne Beene. 1967. A dictionary of Iraqi Arabic: Arabic–
English. Washington, DC: Georgetown University Press.
114
Chapter 5
Khuzestan Arabic
Bettina Leitner
University of Vienna
Khuzestan Arabic is an Arabic variety spoken in the southwestern Iranian province
of Khuzestan. It has been in contact with (Modern) Persian since the arrival of
Arab tribes in the region before the rise of Islam. Persian is the socio-politically
dominant language in the modern state of Iran and has influenced the grammar of
Khuzestan Arabic on different levels. The present article discusses phenomena of
contact-induced change in Khuzestan Arabic and considers their limiting factors.
1 Current state and historical development
1.1 Historical development
Arab settlement in Iran preceded the Arab destruction of the Sasanian empire
with the rise of Islam. Various tribes, such as the Banū Tamīm, had settled in
Khuzestan prior to the arrival of the Arab Muslim armies (Daniel 1986: 211). In
the centuries after the spread of Islam in the region, large groups of nomads from
the Ḥanīfa, Tamīm, ʕAbd-al-Qays, and other tribes crossed the Persian Gulf and
occupied some of the territories of southwestern Iran (Oberling 1986: 215). The
Kaʕb, still an important tribe in the area,1 settled there at the end of the sixteenth
century (Oberling 1986: 216). During the succeeding centuries many more tribes
moved from southern Iraq into Khuzestan. This has led to a considerable increase
of Arabic speakers in the region, which until 1925 was called Arabistan (see Gazsi
2011: 1020; Gazsi, this volume). Today Khuzestan is one of the 31 provinces of the
Islamic Republic of Iran, situated in the southwest, at the border with Iraq.
There has been considerable movement to and from Iraq, to Kuwait, Bahrain,
and Syria, and from villages into towns. Many of these migrations were a consequence of the Iran–Iraq war (1980–1988), but some were due to socio-economic
1
Cf. Oberling (1986: 218) for an overview of the Arab tribes in Khuzestan.
Bettina Leitner. 2020. Khuzestan Arabic. In Christopher Lucas & Stefano Manfredi
(eds.), Arabic and contact-induced change, 115–134. Berlin: Language Science Press.
DOI:10.5281/zenodo.3744509
Bettina Leitner
reasons. The settlement of Persians in the region over the past decades (Gazsi
2011: 1020) is another important factor in its demographic history. From the early
twentieth century on, Khuzestan has attracted international, especially British,
interest because of its oil resources.
1.2 Current situation of Arabs in Khuzestan
Information about the exact number of Arabic-speaking people in Iran, and in
Khuzestan in particular, is hard to find. Estimates in the 1960s of the Arabicspeaking population in Iran ranged from 200,000 to 650,000 (Oberling 1986: 216).
Today it is estimated that around 2 to 3 million Arabs live in Khuzestan (Matras
& Shabibi 2007: 137; Gazsi 2011: 1020).
Many Arabs and Persians living in Khuzestan work in the sugar cane or oil
industries, but few of the former hold white-collar or managerial positions (De
Planhol 1986: 55–56). This is one of the reasons why many Arabs in Khuzestan
feel strongly disadvantaged in society and politics in comparison to their Persian
neighbours.2
2 Language contact in Khuzestan
Currently, the main and most influential language in contact with Khuzestan
Arabic (KhA) is the Western Iranian language Persian. Among the other (partly
historically) influential languages in the region the most prominent are English,
Turkish/Ottoman (cf. Ingham 2005), and Aramaic (see Procházka, this volume).
Persian and different forms of Arabic share a long history of contact in the
region of Khuzestan, implying a long exchange of language material in both directions.
KhA belongs to the Bedouin-type south Mesopotamian gələt-dialects.3 Therefore, it shows great similarity to Iraqi dialects such as Basra Arabic, as well as to
other dialects in the Gulf, such as Bedouin Bahraini Arabic – that is, the Arabic
spoken by the Sunni Arab population descended from Najd.
2
The most common Khuzestan Arabic terms for the Persian people and their language are
ʕaǧam ‘Persian’ (people and language; lit. ‘non-Arab’), and əl-ǧamāʕa ‘Persians’ (lit. ‘group
of people’). Both are often used pejoratively.
3
There is as yet no comprehensive grammar of the dialects of Khuzestan. The main source of
information on these dialects is the collection of data made in the 1960s by the Arabist and linguist Bruce Ingham (1973; 1976; 2011). The article by Yaron Matras and Maryam Shabibi, “Grammatical borrowing in Khuzistani Arabic” (Matras & Shabibi 2007), is based on Shabibi’s unpublished dissertation “Contact-induced grammatical changes in Khuzestani Arabic” (Shabibi
2006).
116
5 Khuzestan Arabic
The dialects of Khuzestan can be considered “peripheral” dialects of Arabic
because they are spoken in a country where Arabic is not the language of the
majority population and is not used in education or administration. Therefore,
there is practically no influence of Modern Standard Arabic. However, because
it shares a long geographically-open border with Iraq, Khuzestan is not isolated
from the Arabic-speaking world. Moreover, since around 2000 it has had access
to Arabic news, soaps, etc. via satellite TV. Intra-Arabic contact is limited to the
linguistically very similar (southern) Iraqi dialects4 through, for example, religious visits to Kerbala.
Persian is the only official language in Iran, it is the only language used in
education, and is sociolinguistically and culturally dominant, especially in the domains of business and administration. Persian consequently enjoys high prestige
in society. For Persian speakers, and sometimes also for KhA speakers, the KhA
varieties have very low prestige and are not associated with the highly prestigious Arabic of the Quran, which is taught in schools. KhA speakers who acquire
KhA as a first language usually acquire Persian at school. Later, the opportunities
for KhA speakers to use Persian are restricted to certain social settings outside
the family, e.g. school, work (employment in a large company would probably
require communication in Persian), contact with Persian friends, or through the
Persian media.
Accordingly, the command of Persian or the degree of bilingualism among
KhA speakers varies greatly due to such factors as level of education, affiliation,
age, gender, and urban or rural environment. The older generation and women
have far less access to education and jobs and consequently less contact with
people outside the family, which implies less exposure to contact situations and
a lower degree of bilingualism. Among some members of the younger generation we may notice a certain intentional reinforcement of Arabic words alongside a resistance to recognizable Persian lexical borrowings, plus a preference for
the Arabic over the Persian names for the cities in Khuzestan. This is of course
consistent with nationalist ideas and the separatist movement taking place in
present-day Khuzestan, and also shows the impact of intentionality in language
contact situations.
In sum, one might find very different degrees of Persian influence among the
speakers of KhA (cf. Matras & Shabibi 2007: 147). For that reason, all statements
on Persian–KhA contact phenomena must be seen in relation to the above factors,
which are decisive for any speaker’s command of Persian.
4
KhA is often differentiated from its neighboring Iraqi dialects by the number of Persian borrowings that are employed (Gazsi 2011: 1020). Although the greatest influence has occurred in
lexicon, Persian influence also extends to grammar (see below).
117
Bettina Leitner
3 Contact-induced changes in KhA
3.1 General remarks
The main aim of the present chapter is to highlight the most striking phenomena
and trends in KhA language change due to contact with Persian.5
All phenomena of contact-induced change in KhA can be considered as transfer of patterns or matter6 from the source language (SL) Persian to the recipient
language (RL) KhA under RL agentivity (i.e. borrowing rather than imposition).
The agents of transfer are cognitively dominant in the RL KhA, the agents’ L1.
Even though Persian is generally acquired during childhood and today is spoken
by most speakers, it usually is the speakers’ L2. Cases of convergence (cf. Lucas
2015: 530–531) are possible in the present contact situation among speakers with
a very high (L1-like) command of Persian, for example university students. But
of course it is hard to draw an exact line between L1 and L2 proficiency and thus
between convergence and borrowing (cf. Lucas 2015: 531).
3.2 Phonology
As in other Bedouin Arabic dialects, the presence of the phonemes /č/ and /g/ is
ultimately the result of internal development from original *k and *q, rather than
borrowing from Persian (see Procházka, this volume).
The phoneme /p/, e.g. perde ‘curtain’ < Pers. parde,7 is also common in all
Iraqi dialects and probably emerged in this region due to contact with Persian
and Kurdish (see Procházka, this volume).
An interesting phonological feature of KhA is that /ɣ/ often reflects etymological *q,8 which is otherwise realized as /g/ and /ǧ/. It is most likely that the
shift /ɣ/ < *q first occurred in KhA forms borrowed from Persian but ultimately
of Arabic origin, e.g. ɣisma ‘part, section’ (cf. Pers. ɣesmat), taṣdīɣ ‘driving licence’ (cf. Pers. taṣdīɣ ‘approval’), taɣrīban ‘approximately’ (cf. Pers. taɣrīban
5
The data used for the present analysis was collected mainly in Aḥwāz, Muḥammara
(Khorramshahr), Ḥamīdiyye and Ḫafaǧiyye (Susangerd) in 2016. The male and female informants were bilingual as well as monolingual KhA speakers from 25 to over 70 years old.
6
Sakel (2007: 15) defines matter replication as the replication of “morphological material and its
phonological shape”.
7
For convenience, and due to the lack of sources on other spoken varieties of Persian, in this
and all following lexical references “Persian” refers to Contemporary Standard Persian. This
should not be taken to suggest that the relevant form in KhA was necessarily borrowed from
this variety of Persian. The transcription and translation of all Persian lexical items is based on
the forms as given by Junker & Alavi (2002) and/or information provided by native speakers.
8
This phenomenon is also documented for the Arabic dialects of Kuwait, Qatar, and the United
Arabic Emirates (Holes 2016: 54, fn. 5).
118
5 Khuzestan Arabic
‘idem’), bəɣri ‘electronic’ (cf. Pers. barɣi with the same meaning but ultimately
going back to CA barq ‘lightning’). This feature is either an internal development,9 or a transfer from Persian, in which both *q and *ɣ in Arabic loanwords
are always pronounced /ɣ/ (Matras & Shabibi 2007: 138).10 Later, this phonological change further affected lexemes which have no cognate forms in Persian, e.g.
baɣra ‘cow’, a borrowing from Modern Standard Arabic (the KhA dialectal form
being hāyša ‘cow’). There are, however, certain lexemes, especially those that do
not have a cognate form in Persian, which are not affected by this rule, e.g. gāl
‘he said’, gēð̣ ‘summer’, or marag ‘sauce’. Other lexemes show free variation in
the pronunciation of /g/, e.g. gabul ~ ɣabul ‘formerly, before’.
Lexical borrowings are often adapted to Arabic phonology. For example, speakers of the older generation usually pronounce the phoneme /p/ as [b], e.g. berde
‘curtain’ < Pers. parde.
Negative structures bear stress on the first syllable,11 e.g. KhA mā́ arūḥ ‘I don’t
go’. This is a feature shared with some Persian and Turkish varieties and other
North East Arabian dialects (Ingham 2005: 178–179). This common phonological characteristic therefore seems to be a Sprachbund phenomenon of the Mesopotamian region, which reflects the long history of contact and migration across
language boundaries due to trade, war, shared cultural practices, nomadism, etc.
(Winford 2003: 70–74). Though the directions and mechanisms of borrowing
within the languages of a Sprachbund are often hard to categorize (Winford 2003:
74), we can probably assume that KhA, being spoken by a minority group, has
borrowed and adapted this phonological stress pattern under RL agentivity.
3.3 Syntax
3.3.1 Replication of Persian phrasal verbs
The replication of phrasal verbs is a contact phenomenon also found in the Arabic
varieties of Turkey (Grigore 2007: 157–159; Procházka, this volume). As shown in
examples (1–4), KhA replicates Persian phrasal verbs by substituting the Persian
light verbs with KhA equivalents and directly replicating the Persian nouns (cf.
9
Cf. Holes (2016: 53–54), who explains the /ɣ/–/q/ merger among the Najd-descendent Bahraini
Arabic speakers as an internal development.
10
In Modern Standard Persian with Tehran “standard” pronunciation (cf. Paul 2018: 581) the
phoneme /ɣ/ (corresponding to CA /q/) has two allophones, [ɢ] and [ɣ] (Majidi 1986: 58–60).
There are, however, some varieties of Spoken Modern Persian, for instance Yazdi Persian, that
maintain a difference between *q and *ɣ (Chams Bernard personal communication; cf. Paul
2018: 582).
11
Ingham (1991: 724) describes this phenomenon also for KhA wh-interrogatives and prepositions.
119
Bettina Leitner
Matras & Shabibi 2007: 142). The noun in example (1) is Arabic in its origins but its
usage in a phrasal verb construction with a new meaning is a Persian innovation.
(1)
a. Aḥwāz, Khuzestan, male, 26 years (own data)
ṭəgg
muḫḫ
hit.prf.3sg.m brain
b. Persian
muḫḫ zadan
brain hit.inf
‘to brainwash, convince someone’12
(2)
a. Aḥwāz, Khuzestan, male, 39 years (own data)
kað̣ð̣
īrād
take.prf.3sg.m nagging
b. Persian
īrād
gereftan
nagging take.inf
‘to pick on someone’
As examples (3) and (4) show, Persian nouns are sometimes adapted morphophonologically.
(3)
a. Aḥwāz, Khuzestan, male, 50 years (own data)
sawwa
ʔōmāde
make.prf.3sg.m ready
b. Persian
āmāde kardan
ready make.inf
‘to prepare sth.’
(4)
a. Aḥwāz, Khuzestan, male, 26 years (own data)
ṭalaʕ
ɣabūli13
emerge.prf.3sg.m acceptance
b. Persian
ɣabūl
šodan
acceptance become.inf
‘to pass (an exam), be accepted’
12
All Persian translations are given in the modern spoken Tehrani variety of Persian, and were
provided by Hooman Mehdizadehjafari, a native speaker of this variety. They are presented in
a broad phonemic transcription.
120
5 Khuzestan Arabic
The pattern for phrasal verbs – transferred into the RL KhA under RL agentivity
– provides KhA with an easy way to convert foreign nouns into verbs.
As illustrated in examples (5) and (6), the pattern is adapted according to Arabic syntactic rules: (i) the verb is moved into the initial position; and (ii) a direct
object is introduced between verb and nominal element (post-verbally). In Persian, however, the verb always remains in final position following the nominal
element and a direct object would be introduced before both elements (see e.g.
Majidi 1990: 447–448).
(5)
a. Aḥwāz, Khuzestan, male, 26 years (own data)
ṭəggi
ɣandart-i
wāks
hit.imp.2sg.f shoe-obl.1sg wax
b. Persian
kafš-am-o
vāks be-zan
shoe-obl.1sg-obj wax imp-hit.prs
‘Polish my shoes!’
(6)
a. Aḥwāz, Khuzestan, female, 35 years (own data)
yṭəggūn
əṭ-ṭamāṭe rande
hit.impf.3pl.m def-tomato grater
b. Persian
gūǧe_farangi-ro rande mī-zanan
tomato-obj
grater ind-hit.prs.3pl
‘They grate some tomato.’
This structure has become productive in KhA. For example, in the phrasal
verb ṭəgg dabbe ‘to cheat’ (lit. ‘to hit a water canister’) both the verb and noun
are taken from KhA and only the construction’s syntactic pattern is taken from
Persian.
3.3.2 Definiteness marking
Matras & Shabibi (2007: 141–142) see KhA relative clauses without definite heads
as evidence for the decline of overt definiteness marking in KhA, based on a Persian model with generally unmarked definiteness, e.g. mara lli šiftū-ha ḫābarat
‘The woman that you saw called’ (2007: 142). However, this pattern is also documented in Arabic dialects which have had no contact with Persian (Pat-El 2017:
454–455; cf. Procházka 2018: 269).
13
The final -i in ɣabūli probably originates from the Persian indefiniteness marker -i (see Majidi
1990: 309–314) and has become part of this word in KhA, so that ɣabūli is monomorphemic.
121
Bettina Leitner
Matras & Shabibi (2007: 140) further postulate that the Persian ezāfe pattern
in adjectival attribution is replicated in KhA.14 According to their theory, the
construct state marker -t (with an indefinite head) and/or the definite article (of
the attribute) are reanalysed as markers of attribution matching the Persian ezāfe
marker -(y)e, as in (7).
(7)
a. KhA (Matras & Shabibi 2007: 140)
ǧazīra-t
l-ḫað̣ra
island-con def-green.f
b. Persian
ǧazīre-ye sabz
island-ez green
‘the green island’
However, this pattern is also observed in other modern Arabic dialects which
have not been exposed to Persian influence as well as in older forms of Arabic.15
Consequently, it is highly unlikely that this phenomenon has developed due
to Persian influence, although it cannot be ruled out that contact with Persian
has fostered the preservation of this apparently old feature.
3.3.3 Word order changes
KhA shows no changes due to contact in basic word order.16 The only attested
word order changes concern the position of the verbs čān ‘to be’ and ṣār ‘to
become’, both of which can appear in final position as an unmarked construction.
This sentence-final position in no case functions as the default, and is in fact
14
See e.g. Ahadi (2001: 103–109) for the usage of the Persian ezāfe.
See Pat-El (2017: 445–449) and Stokes (2020) for numerous examples from different varieties
of Arabic and other Central Semitic languages. See also Retsö (2009: especially 21–22) and
Procházka (2018: 267–269), who also proves that this is an old feature already found in Old
Arabic and points out that it is mainly found among dialects which are spoken in regions with
no or only marginal influence from Modern Standard Arabic.
16
Ingham (1991: 715) states that in KhA neither VSO nor SVO word order is particularly dominant.
Matras & Shabibi (2007: 147) postulate that the usage of OV order in KhA is increasing as “the
beginning of a shift in word order” on the basis of the Persian type, where OV prevails. In both
of their examples the objects are topicalized (with pronominal resumption), which is a common
phenomenon in spoken Arabic (Brustad 2000: 330–333; 349), and as such not obviously the
result of Persian influence (cf. El Zarka & Ziagos 2019, who in their recent description of the
beginnings of word order changes in some Arabic dialects spoken in southern Iran, show that
these dialects, like KhA, have still retained VO as their basic word order despite the strong
influence of Persian).
15
122
5 Khuzestan Arabic
less frequent than its non-final position.17 čān or ṣār in final position are never
stressed.
The sentence-final position of čān or ṣār (see examples 8–10) is likely a pattern
replication of the Persian model, i.e. sentences with final būdan ‘to be’ or šodan
‘to become’.
(8)
a. ʕAbbādān, Khuzestan, male, 35 years (own data)
šuɣul-hum b-əl-bandar čān
work-3pl.m in-def-port be.prf.3sg.m
b. Persian
kār-ešūn tū-ye bandar būd
job-obl.3pl in-ez port be.pst.3sg
‘Their job was at the port.’
(9)
a. Muḥammara, Khuzestan, male, 30 years (own data)
əǧdād-i
mallāk-īn čānaw
grandparents-obl.1sg owner-pl be.prf.3pl.m
b. Persian
aǧdād-am
mālek būdan
grandparents-obl.1sg owner be.pst.3pl
‘My grandparents were owners [of land].’
(10)
a. Aḥwāz, Khuzestan, female, 40 years (own data)
hassa šway l-māy
bārəd ṣār
now a_bit def-water cold become.prf.3sg.m
b. Persian
alʔān yekam ʔāb sard šod
now a_bit water cold become.pst.3sg
‘The water has become a bit cold now.’
The next example might show a tendency to use a present-tense copula with
human subjects, expressed with the verb ṣār ‘to become’:
(11)
17
a. Aḥwāz, Khuzestan, female, 35 years (own data)
əhya mart uḫū-y
əṣṣīr
3sg.f wife brother-obl.1sg cop.impf.3sg.f
b. Persian
ūn zan-dādāš-am-e
3sg wife-brother-obl.1sg-cop.prs.3sg
‘She is the wife of my brother.’
In my data, čān appears 23 of 152 times in sentence-final position, ṣār 11 of 165 times. The
additional examples are taken from my questionnaire.
123
Bettina Leitner
In the KhA construction for pluperfect tense, čān can also appear in sentencefinal position, after the active participle. This construction, although not very
frequent, is very likely a direct transfer of the Persian structure, in which the
auxiliary būdan also follows the participle.18
(12)
a. Aḥwāz, Khuzestan, male, 26 years (own data)
lamman əyēna
l-əl-bīət,
əhma mākl-īn
čānaw
when come.prf.1pl to-def-house 3pl.m eat.ptcp-pl.m be.prf.3pl.m
b. Persian
vaɣti-ke mā bargaštīm
ḫūne, ūnhā ɣazā-ro ḫorde
when-rel 1pl come_back.pst.1pl home 3pl food-obj eat.ptcp
būdan
be.pst.3pl
‘When we came home, they had (already) eaten.’
This word order change has probably been triggered by the high frequency in
speech of Persian sentences with forms of būdan in final position. Lucas (2012:
295) explains the usage of foreign patterns as the result of the human cognitive
tendency to minimize the high processing efforts associated with the extensive
use of two languages.19
čān is also used in sentence-final positions after the main verb in the imperfect in KhA constructions expressing the continuous past. In spoken Persian, the
continuous past is formed without a sentence-final būdan.20 This case is not a
direct transfer of the Persian pattern, but perhaps a construction analogous to
the pluperfect and other Persian forms with būdan in final position.
(13)
18
a. Aḥwāz, Khuzestan, male, 55 years (own data)
hāda
ham mən zuɣur
yəštəɣəl
čān
dem.sg.m also from childhood work.impf.3sg.m be.prf.3sg.m
‘This one has also been working from childhood on.’
Matras and Shabibi (2007: 142–143) describe the use of this construction as a change in the
KhA tense system. However, the pattern kān + active participle is also commonly used in
other Arabic dialects to express pluperfect meaning or to describe completed actions which
have an impact on the present, see for example Denz (1971: 92–94; 115–116) for Iraqi (Kwayriš)
and Grotzfeld (1965: 88) for Syrian Arabic.
19
Connections between units of a neural network associated with certain syntactic patterns can
be strengthened from repeated exposure to and use of that pattern (Lucas 2012: 291). Hence,
the employment of a Persian syntactic structure in KhA needs less processing effort because
the same strengthened neural network is activated.
20
The Modern Iranian Persian continuous past is formed with the particle mī prefixed to the
simple past of the respective main verb and can (for the progressive form) be preceded by the
simple past of dāštan ‘to have’: e.g. (dāšt) mī-raft ‘he was going’ (Majidi 1990: 232, 235).
124
5 Khuzestan Arabic
b. Persian
īn-am
az kūdaki
kār mī-kard
dem.sg-also from childhood work ind-do.pst.3sg
‘This one has also been working from childhood on.’
Example (14) shows both syntactic variants in one sentence, i.e. čān before and
after the main verb.
(14)
a. Muḥammara, Khuzestan, female, 40 years (own data)
umm-i
čānat
tətḥaǧǧab,
eh, əb-zamān əš-šāh,
mother-obl.1sg be.prf.3sg.f veil.impf.3sg.f yes in-time def-shah
čānat
bass tətbawwaš
only veil.impf.3sg.f be.prf.3sg.f
b. Persian
mādar-am (dāšt)
neqāb mī-zad,
āre, dar zamān-e
mother-1sg (have.pst.3sg) veil ind-hit.pst.3sg yes in time-ez
šāh, hamīše neqāb mī-zad
shah always veil ind-hit.pst.3sg
‘My mother used to veil her face (with a būšiyye),21 yes, during the
times of the shah, she always used to veil her face.’
Because all the above examples equally work with čān/ṣār in non-final position, this process of word-order-related pattern replication in KhA is still ongoing. Indeed, all informants, when asked for the correct structure in the above
examples, preferred the verb čān in non-final position.22
Lucas (2015: 530–531) explains the basic word order changes (from VSO to
SOV) in Bukhara Arabic (cf. Ratcliffe 2005: 143–144; and Versteegh 2010: 639) as
a result of convergence with Uzbek.23 Although a clear division between convergence and borrowing is hard to make, I consider the contact-induced word order
changes that occur in KhA to be instances of borrowing because most speakers
are clearly native speakers of, and therefore dominant in, KhA only.
21
būšiyye or pūšiyye ‘veil’ is also documented for Iraqi Arabic (Woodhead & Beene 1967: 53).
My informants from Baghdad considered all constructions with čān in final position to be
wrong. However, this structure is used in Basra Arabic (Qasim Hassan, personal communication, January 2018).
23
Lucas (2015: 525) defines convergence as changes made to a language under the agentivity of
speakers who are native speakers of both the SL and the RL.
22
125
Bettina Leitner
3.3.4 ḫōš preceding verbs and nouns
In Persian, ḫoš ‘good, well’ is used as a prefixed (lexicalized) element preceding
some nouns and verbs to coin compound adjectives, nouns, and verbs (Majidi
1990: 411, 413): e.g. Pers. ḫoš-andām ‘handsome’ (< andām ‘shape; body’), ḫošnevīs ‘calligrapher’ (< present stem nevīs- ‘to write’).
KhA has borrowed some of these Persian compound adjectives: e.g. KhA ḫōšbū ‘nice-smelling’ (< Pers. bū ‘smell, scent’), ḫōš-tīp ‘handsome’ (< Pers. tīp ‘type’),
and ḫōš-aḫlāq ‘(with) good manners’ (< Pers. aḫlāq ‘decency; ethics, morality’,
pl. of ḫolq ‘character, nature’). However, in KhA the use of this element has been
further developed. It is productively used as an attributive adjective preceding
nouns, but not agreeing in gender or number with them, e.g. ḫōš walad ‘a good
boy’, ḫōš əbnayya ‘a good girl’, ḫōš banāt ‘good girls’, ḫōš əwlād ‘good kids’, and
as and adverb meaning ‘well’, e.g. hәyya ḫōš təsʔal ‘she asks good questions’ (lit.
‘she asks well’; speaker: Aḥwāz, Khuzestan, male, 27 years).24
3.4 Lexicon
3.4.1 Lexical transfer
The greatest influence from Persian on KhA has occurred in lexicon. Many Persian lexemes were borrowed generations ago. The most frequently borrowed elements are nouns denoting cultural or technological innovations which have
filled lexical gaps in the RL KhA. Verbs, adverbs, adjectives, and many discourse
particles have also been borrowed from the SL Persian.
The majority of the examples below are cases of transfer of morphophonological material (matter) and semantic meaning (pattern) under RL agentivity.
Many of the Persian borrowings have been phonologically and morphologically integrated into the RL. For instance, for many borrowed Persian nouns
Arabic internal plural forms are created, e.g. ḫətākīr ‘ball-point pens’ (sg. ḫətkār
< Pers. ḫod-kār ‘ball-point pen’), or banādər ‘ports’ (sg. bandar < Pers. bandar
‘port’).
Again, the borrowing of foreign (L2) elements into the speakers’ L1 might be
explained by the human cognitive tendency to minimize the processing effort
in lexical selection between two languages (Lucas 2012: 291; see §3.3.3). So if a
certain Persian word is frequently used and often heard (for example at school),
the connections of a neural network associated with this word are strengthened
(Lucas 2012: 291), which makes it easier to employ the word in one’s L1.
24
This construction is also found in Iraqi Arabic (cf. Erwin 1963: 256), which might prove that
the element ḫōš is an older borrowing.
126
5 Khuzestan Arabic
3.4.2 Semantic fields
The following illustrative list of Persian loans in KhA shows the most important
semantic fields of lexical borrowing.
Administration and military:
čārra ‘crossroad’ < Pers. čahār-rāh; sarbāz ~ šarbāz ‘soldier’ < Pers. sarbāz;
farmāndāri ‘governorship’ < Pers. farmāndāri.
Agriculture:
kūd ‘dung’ < Pers. kūd; ʕalafkoš ‘pesticide’ (lit. weed-killer) < Pers. ʔalafkoš.
Dress and textiles:
dāmen ‘skirt’ < Pers. dāman; šīəla ‘head covering’ < Pers. šāl ‘Kashmir
shawl’ (Ingham 2005: 174).
Education:
klāṣ ‘class, grade’ < Pers. kelās; ḫətkār ‘ball-point pen’ < Pers. ḫod-kār;
dānišga ‘university’< Pers. dānišgāh.
Food:
ǧaʕfari ‘parsley’ < Pers. ǧaʔfari; češmeš ‘raisins’ < Pers. kešmeš; serke ‘vinegar’ < Pers. serke; šalɣam ‘turnip’ < Pers. šalɣam.
Material culture:
šīše ‘bottle’ < Pers. šīše; ǧām ‘(window) glass’ < Pers. ǧām ‘(window) glass;
goblet, cup’; tīɣe ‘blade’ < Pers. tīɣe; yəḫčāle ‘refrigerator’ < Pers. yahčāl;
sīm buksel ‘towrope’ < Pers. sīm-e boksol; perde ~ berde ‘curtain’ < parde;
gīre ‘hair barrette’ < Pers. gīre-ye sar/mūy; mīz ‘table’ < Pers. mīz; darīše
‘window’ < Pers. darīče; pənǧara ‘window’ < Pers. panǧare.
Other:
ɣīme ‘price’ < Pers. ɣīmat; bandar ‘port’ < Pers. bandar; nāmard ‘brute’ <
Pers. nāmard ‘coward; brute, rascal’.
Some items ultimately of Arabic origin have been re-borrowed into KhA from
Persian, preserving the Persian meaning, e.g. KhA bərɣi ‘electronic’ < Pers. barɣ
‘electricity; lightning’ < Arabic barq ‘lightning’.
127
Bettina Leitner
3.4.3 Verbs and adverbs
KhA verbs and adverbs resulting from language contact are always morphologically integrated. These are either directly borrowed Persian verbs, e.g. bannad ‘to
close (e.g. the tap)’ < Pers. imperfect and present stem band- ‘close’;25 gayyər ‘to
get stuck’ < Pers. gīr šodan ‘to get stuck’; ʕammәr ‘to repair’ < Pers. taʕmīr kardan ‘to repair’; čassәb ‘to glue’ < Pers. časb zadan ‘to glue’; gəzar ‘to pass (time)’
< Pers. present stem gozar- ‘to pass (time)’ (see example (15) below);26 zaḥəm ‘to
bother’ (transitive) < Pers. zahmat dādan ‘to bother, cause trouble’ (transitive)
(see examples (16) and (17) below);27 or Persian nouns turned into KhA (ad)verbs,
e.g. əb-zūr ‘by force’ < Pers. zūr ‘power; violence; force’.
(15)
Aḥwāz, Khuzestan, male, 26 years (own data)
čā hāy əl-ḥayāt lō la? təgzar
baʕad, təmši
dm dem.f def-life or no pass.impf.3sg.f after_all go.impf.3sg.f
‘See, that is how life is, right? It passes by (quickly), it goes.’
(16) Aḥwāz, Khuzestan, male, 26 years (own data)
ʕafwan
zaḥmīət-kum,
bother.prf.1sg-2pl.m sorry
‘Sorry, I must have bothered you.’28
(17) Aḥwāz, Khuzestan, male, 25 years (own data)
mumkin azaḥm-ək
əb-šuɣla
possible bother.impf.1sg-2sg.m with-issue
‘May I bother you with something (i.e. ask you a favour)?’
3.4.4 Discourse elements
A range of Persian discourse elements have been borrowed by KhA (cf. Matras &
Shabibi 2007: 143–145),29 e.g. KhA ham ~ hamme ‘also, as well’ < Pers. ham and
25
Also common in the Gulf region and in Yemen (Behnstedt & Woidich 2014: 290).
The verb gəzar is used only in phrases that refer to the “passing by” of life.
27
The KhA noun zaḥme ‘shame’ is also used for a rebuke, e.g. zaḥme ʕalīək! ‘Shame on you!’,
which would be expressed in a different way in Persian: ḫeǧālat ne-mī-keši? ‘Shame on you!’
(lit. ‘Are you not ashamed?’).
28
A phrase often used when leaving, for example after an invitation for dinner, cf. Pers. ḫeyli
zahmat dādīm lit. ‘We have caused (you) a lot of trouble’.
29
Matras & Shabibi (2007: 144) claim that the Persian conjunctions agarče and bāīnke, both meaning ‘although, even though’, and the Persian factual complementizer ke ‘that’ have also been
borrowed by KhA. However, I have found no evidence for their usage in my data.
26
128
5 Khuzestan Arabic
KhA ham…ham ‘(both)…and’ < Pers. ham…ham;30 or KhA hīč ‘nothing; no(t)…
at all’ < Pers. hīč.31
The KhA discourse elements ḫō/ḫōš ‘well; okay’ < Pers. ḫo(b)/ḫoš are often
used phrase-initially, (18).32 They are of Persian origin, but have partly adopted
a different form and function in KhA.33
(18)
Aḥwāz, Khuzestan, male, 55 years (own data)
əhna baba
ḫōš, š-ʕəd-na, taʕay
dm what-at-1pl come.imp.sg.f here father
‘Okay, what (else) do we have, come here, dear!’
Both ḫō and ḫōš are also often used in stories following the verb gāl ‘to say’.
(19) Aḥwāz/Fəllāḥiyya, Khuzestan, female, 50 years (own data)
lamman ɣada
mən ʕəd-hum, gāl-la
ḫō,
when leave.prf.3sg.m from at-3pl.m say.prf.3sg.m-dat.3sg.m dm
hāy
ər-rummānāt
š-asawwi
bī-hən
dem.sg.f def-pomegranate.pl what-make.impf.1sg with-3pl.f
‘When he left them, he said to him, “Well, what shall I do with these
pomegranates?”’
4 Conclusion
Because of the dominance of Persian in the Iranian educational system and work
environment, the lack of influence from Modern Standard Arabic, and the long
period of geographical proximity, the Persian-speaking society of southwest Iran
has left many linguistic traces in the language of the Arabic-speaking community
of Khuzestan.
30
This discourse element is also known for Iraq (Malaika 1963: 36) and, like KhA hast ~ hassət
‘there is’ < Pers. hast (Ingham 1973: 25, fn.27), is probably an older borrowing.
31
Shabibi (2006: 176–177) further derives KhA balkət ‘maybe, hopefully’ from Pers. balke ham,
which can mean ‘maybe’. A Turkish origin of this word seems more likely: cf. Aksoy (1963: 620)
for the existence of belke ~ belkit in Eastern Turkish dialects. Malaika (1963: 35) also derives
the Baghdadi Arabic belki ‘rather, maybe’ from Turkish, as does Seeger (2009: 28) for balki,
balkīš, balkin ‘maybe; possibly; probably’ in Ramallah Arabic.
32
According to my informants and data, the form ḫōb is not used in KhA (contrast Matras &
Shabibi 2007: 143).
33
In Persian, ḫob is a discourse particle and related to the adjective and adverb ḫūb, ḫo is also
a discourse particle used in less formal situations (Mehrdad Meshkinfam, Erik Anonby and
Mortaza Taheri-Ardali, personal communication), and ḫoš is an adjective (see §3.3.4; Shabibi
2006: 160; Mohammadi 2018: 104–105). Thus the Persian adjective ḫoš has been desemanticized
in KhA to function as a discourse particle with the meaning ‘well, okay’ (Shabibi 2006: 160).
129
Bettina Leitner
Van Coetsem (2000: 59; cf. Lucas 2015: 532) suggests that lexical, but not syntactic and phonological transfer is to be expected under RL agentivity. However,
KhA phonology and syntax have been influenced by the SL Persian under RL
agentivity, albeit to a much lesser extent than the lexicon.
KhA does not show transfer of patterns from Persian in either inflectional or
derivational morphology. However, we do find an adapted pattern replication of
Persian phrasal verbs (with preservation of the Arabic word order).
As for syntax and contact-induced word order changes, the alternative sentence construction with čān in sentence-final position can be explained as a result
of Persian influence on KhA. This change might have been triggered by the similar and very frequent Persian constructions with sentence-final būdan. Thus, we
do have some syntactic change due to transfer under RL agentivity, which Van
Coetsem considered to be unexpected (see above).
Persian lexical items have often been borrowed in KhA for novel concepts
(lexical gaps), which is why semantic fields relating to technical or cultural innovations, education, and administration show the greatest amount of Persian
borrowing. This also explains why nouns are generally more often transferred
than verbs (cf. Lucas 2015: 532). Persian words are regularly integrated into KhA
phonology and morphology, for example the Arabic internal plural is formed for
Persian nouns. Also, many discourse particles have been transferred from Persian into KhA. Some of them, e.g. ham ‘also’, had been in use generations ago
among Arabic speakers in Khuzestan and beyond (Iraq, Gulf).
Of course, contact between KhA and Persian has always been limited to certain
social contexts (outside the family), especially for women, who had and still have
much less access to education and employment and thus to the Persian-speaking
world. This fact, and some structural differences between the languages, explain
the limits of contact-induced language change in KhA, especially in morphology
and syntax.
Hopefully, future research on the dialects of Khuzestan will provide more empirical data on instances of contact-induced change. An enlarged database should
especially provide further evidence concerning the development and extent of
word order changes.
Further reading
) Ingham (2011) provides a sketch grammar of KhA.
) Ingham (2005) discusses Turkish and Persian borrowings in KhA and northeastern Arabian dialects.
130
5 Khuzestan Arabic
) Matras & Shabibi (2007) is an article on contact-induced changes in KhA based
on Shabibi (2006).
) Shabibi (2006) is an unpublished doctoral dissertation on contact-induced change
in KhA.
Acknowledgements
I would like to thank Stephan Procházka and Dina El-Zarka for their critical
remarks and bibliographical suggestions on this chapter. I would additionally
like to thank my informant and good friend Majed Naseri for all his help on the
transcription and translation of my recordings.
Abbreviations
1, 2, 3
CA
cop
dat
def
dem
dm
ez
f
imp
impf
ind
inf
1st, 2nd, 3rd person
Classical Arabic
copula
dative
definite
demonstrative
discourse marker
Persian ezāfe
feminine
imperative
imperfect (prefix conjugation)
indicative
infinitive
KhA
m
obj
obl
Pers.
pl/pl.
ptcp
prf
prog
prs
pst
rel
sg/sg.
Khuzestan Arabic
masculine
object
oblique
Persian
plural
participle
perfect (suffix conjugation)
progressive
present
past
relative particle
singular
References
Ahadi, Shahram. 2001. Verbergänzungen und zusammengesetzte Verben im
Persischen: Eine valenztheoretische Analyse. Wiesbaden: Reichert.
Aksoy, Ömer A. (ed.). 1963. Türkiyede halk ağzından derleme sözlüğü. Ankara:
Türk Tarih Kurumu Basımevi.
Behnstedt, Peter & Manfred Woidich. 2014. Wortatlas der arabischen Dialekte III:
Verben, Adjektive, Zeit und Zahlen. Leiden: Brill.
131
Bettina Leitner
Brustad, Kristen. 2000. The syntax of spoken Arabic: A comparative study of
Moroccan, Egyptian, Syrian, and Kuwaiti dialects. Georgetown: Georgetown
University Press.
Daniel, Elton L. 1986. Arab settlements in Iran. In Ehsan Yarshater (ed.), Encyclopaedia Iranica, vol. 2, 210–214. London: Routledge & Kegan Paul.
De Planhol, Xavier. 1986. Ābādān, ii. Modern Ābādān. In Ehsan Yarshater (ed.),
Encyclopaedia Iranica, vol. 2, 53–57. London: Routledge & Kegan Paul.
Denz, Adolf. 1971. Die Verbalsyntax des neuarabischen Dialektes von Kwayriš
(Irak): Mit einer einleitenden allgemeinen Tempus- und Aspektlehre. Wiesbaden:
Steiner.
El Zarka, Dina & Sandra Ziagos. 2019. The beginnings of word order change in
the Arabic dialects of southern Iran in contact with Persian: A preliminary
study of data from four villages in Bushehr and Hormozgan. Iranian Studies.
1–24. DOI:10.1080/00210862.2019.1690433
Erwin, Wallace M. 1963. A short reference grammar of Iraqi Arabic. Washington,
DC: Georgetown University Press.
Gazsi, Dénes. 2011. Arabic–Persian language contact. In Stefan Weninger,
Geoffrey Khan, Michael P. Streck & Janet C. E. Watson (eds.), The Semitic
languages: An international handbook, 1015–1021. Berlin: De Gruyter Mouton.
Grigore, George. 2007. L’arabe parlé à Mardin: Monographie d’un parler arabe
“périphérique”. Bucharest: Editura Universitãţii din Bucureşti.
Grotzfeld, Heinz. 1965. Syrisch-arabische Grammatik (Dialekt von Damaskus).
Wiesbaden: Harrassowitz.
Holes, Clive. 2016. Dialect, culture, and society in eastern Arabia. Vol. 3: Phonology,
morphology, syntax, style. Leiden: Brill.
Ingham, Bruce. 1973. Urban and rural Arabic in Khūzistān. Bulletin of the School
of Oriental and African Studies 36. 533–553.
Ingham, Bruce. 1976. Regional and social factors in the dialect geography of southern Iraq and Khuzistan. Bulletin of the School of Oriental and African Studies
39. 62–82.
Ingham, Bruce. 1991. Sentence structure in Khuzistani Arabic. In Alan S. Kaye
(ed.), Semitic studies: In honor of Wolf Leslau on the occasion of his eighty-fifth
birthday, November 14th, 1991, vol. 1, 714–728. Wiesbaden: Harrassowitz.
Ingham, Bruce. 2005. Persian and Turkish loans in the Arabic dialects of North
Eastern Arabia. In Éva Ágnes Csató, Bo Isaksson & Carina Jahani (eds.), Linguistic convergence and areal diffusion: Case studies from Iranian, Semitic and
Turkic, 173–180. London: Routledge.
Ingham, Bruce. 2011. Khuzestan Arabic. In Lutz Edzard & Rudolf de Jong (eds.),
Encyclopedia of Arabic language and linguistics, online edn. Leiden: Brill.
132
5 Khuzestan Arabic
Junker, Heinrich F. J. & Bozorg Alavi. 2002. Persisch-Deutsch Wörterbuch. 9th edn.
Wiesbaden: Harrassowitz.
Lucas, Christopher. 2012. Contact-induced grammatical change: Towards an explicit account. Diachronica 29. 275–300.
Lucas, Christopher. 2015. Contact-induced language change. In Claire Bowern &
Bethwyn Evans (eds.), The Routledge handbook of historical linguistics, 519–536.
London: Routledge.
Majidi, Mohammad-Reza. 1986. Strukturelle Grammatik des Neupersischen (Fārsi).
Band I: Phonologie. Hamburg: Buske.
Majidi, Mohammad-Reza. 1990. Strukturelle Grammatik des Neupersischen (Fārsi).
Band II: Morphologie. Hamburg: Buske.
Malaika, Nisar. 1963. Grundzüge der Grammatik des arabischen Dialektes von
Bagdad. Wiesbaden: Harrassowitz.
Matras, Yaron & Maryam Shabibi. 2007. Grammatical borrowing in Khuzistani
Arabic. In Yaron Matras & Jeanette Sakel (eds.), Grammatical borrowing in
cross-linguistic perspective, 137–149. Berlin: Mouton de Gruyter.
Mohammadi, Ariana Negar. 2018. Discourse markers in colloquial and formal
Persian: A corpus-based discourse analysis approach. Gainesville, FL: University
of Florida. (Doctoral dissertation).
Oberling, Pierre. 1986. Arab tribes of Iran. In Ehsan Yarshater (ed.), Encyclopaedia
Iranica, vol. 2, 215–219. London: Routledge & Kegan Paul.
Pat-El, Na’ama. 2017. Digging up archaic features: Neo-Arabic and comparative
Semitics in the quest for proto-Arabic. In Ahmad Al-Jallad (ed.), Arabic in context: Celebrating 400 years of Arabic at Leiden University, 441–475. Leiden: Brill.
Paul, Ludwig. 2018. Persian. In Geoffrey L. J. Haig & Geoffrey Khan (eds.), The
languages and linguistics of Western Asia: An areal perspective, 569–624. Berlin:
De Gruyter Mouton.
Procházka, Stephan. 2018. The northern Fertile Crescent. In Clive Holes (ed.),
Arabic historical dialectology: Linguistic and sociolinguistic approaches, 257–292.
Oxford: Oxford University Press.
Ratcliffe, Robert R. 2005. Bukhara Arabic: A metatypized dialect of Arabic in
Central Asia. In Éva Ágnes Csató, Bo Isaksson & Carina Jahani (eds.), Linguistic
convergence and areal diffusion: Case studies from Iranian, Semitic and Turkic,
141–51. London: Routledge.
Retsö, Jan. 2009. Nominal attribution in Semitic: Typology and diachrony. In
Janet C. E. Watson & Jan Retsö (eds.), Relative clauses and genitive constructions in Semitic, 3–33. Oxford: Oxford University Press.
133
Bettina Leitner
Sakel, Jeanette. 2007. Types of loan: Matter and pattern. In Yaron Matras, Georg
Bossong & Bernard Comrie (eds.), Grammatical borrowing in cross-linguistic
perspective, 15–29. Berlin: Mouton de Gruyter.
Seeger, Ulrich. 2009. Der arabische Dialekt der Dörfer um Ramallah. Vol. 2: Glossar.
Wiesbaden: Harrassowitz.
Shabibi, Maryam. 2006. Contact-induced grammatical changes in Khuzestani
Arabic. Manchester: University of Manchester. (Doctoral dissertation).
Stokes, Phillip W. 2020. gahwat al-murra: The N + DEF-ADJ syntagm and the *at
> ah shift in Arabic. Bulletin of the School of Oriental and African Studies 83(1).
75–93.
Van Coetsem, Frans. 2000. A general and unified theory of the transmission process
in language contact. Heidelberg: Winter.
Versteegh, Kees. 2010. Contact and the development of Arabic. In Raymond
Hickey (ed.), The handbook of language contact, 634–651. Oxford: WileyBlackwell.
Winford, Donald. 2003. An introduction to contact linguistics. Malden: Blackwell.
Woodhead, Daniel R & Wayne Beene. 1967. A dictionary of Iraqi Arabic: Arabic–
English. Washington, DC: Georgetown University Press.
134
Chapter 6
Anatolian Arabic
Faruk Akkuş
University of Pennsylvania
This chapter investigates contact-induced changes in Anatolian Arabic varieties.
The study first gives an overview of the current state and historical development
of Anatolian Arabic. This is followed by a survey of changes Anatolian Arabic
varieties have undergone as a result of language contact with primarily Turkish
and Kurdish. The chapter demonstrates that the extent of the change varies from
one dialect to another, and that this closely correlates with the degree of contact a
dialect has had with the surrounding languages.
1 Current state and historical development
Anatolian Arabic is part of the so-called qəltu-dialect branch of the larger Mesopotamian Arabic, and essentially refers to the Arabic dialects spoken in eastern Turkey.1 In three provinces of Turkey – Hatay, Mersin and Adana – Syrian
sedentary Arabic is spoken (see Procházka, this volume, for discussion of these
dialects). Other than these dialects, in Jastrow’s (1978) classification of Mesopotamian qəltu dialects, Anatolian Arabic dialects are subdivided into five groups:
Diyarbakır dialects (spoken by a Jewish and Christian minority, now almost extinct); Mardin dialects; Siirt dialects; Kozluk dialects; and Sason dialects. In his
later work, Jastrow (2011a) classifies Kozluk and Sason dialects under one group
along with Muş dialects – investigated primarily by Talay (2001; 2002). The two
larger cities where Arabic is spoken are Mardin and Siirt, although in the latter
Arabic is gradually being replaced by Turkish.
1
This group represents an older linguistic stratum of Mesopotamia as compared to the gələt
dialects. The terms qəltu vs. gələt dialects are due to Blanc (1964), who distinguished between
the Arabic dialects spoken by three religious communities, Muslim, Jewish, and Christian, in
Baghdad. He classified the Jewish and Christian dialects as qəltu dialects and the Muslim dialect
as a gələt dialect, on the basis of their respective reflexes of Classical Arabic qultu ‘I said’.
Faruk Akkuş. 2020. Anatolian Arabic. In Christopher Lucas & Stefano Manfredi
(eds.), Arabic and contact-induced change, 135–158. Berlin: Language Science Press.
DOI:10.5281/zenodo.3744511
Faruk Akkuş
The linguistic differences between these various Arabic-speaking groups are
quite considerable. Thus, given the low degree of mutual intelligibility, speakers of different varieties resort to the official language, Turkish, to communicate. Jastrow (2006) reports an anecdote, wherein high school students from
Mardin and Siirt converse in Turkish, since they find it difficult to understand
each other’s dialects. Expectedly, mutual intelligibility is at a considerably higher
level among different varieties of a single group, despite certain differences. For
instance, speakers of Kozluk and Muş Arabic have no difficulty in communicating with one another in Arabic.
The existence of Anatolian dialects closely relates to the question of the Arabicization of the greater Mesopotamian area. Although the details largely remain
obscure, a commonly-held view is that it took place in two stages: the first stage
concerns the emergence of urban varieties of Arabic around the military centers,
such as Baṣra or Kūfa, during the early Arab conquests. Later, the migration of
Bedouin dialects of tribes added another layer to the urban dialects (see e.g. Blanc
1964; Versteegh 1997; Jastrow 2006 for discussion). According to Blanc (1964), the
qəltu dialects are a continuation of the medieval vernaculars that were spoken
in the sedentary centers of Abbasid Iraq. Blanc (1964) also noted that the qəltu
dialects did not stop at the Iraqi–Turkish border, but in fact continued into Turkish territory. He mentioned the towns of Mardin and Siirt as places where qəltu
dialects were still spoken.
Despite being a continuation of Mesopotamian dialects, Anatolian dialects of
Arabic have been cut off from the mainstream of Arabic dialects. How exactly
this cut-off and separation between dialects happened, given the lack of specific
barriers, is largely unknown and remains at a speculative level. Regarding this
topic, Procházka (this volume) suggests “the foundation of nation states after
World War One entailed significant decrease in contact between the different
dialect groups and an almost complete isolation of the Arabic dialects spoken in
Turkey”.
Like Central Asian Arabic and Cypriot Maronite Arabic (Walter, this volume),
Anatolian Arabic dialects are characterized by: (i) separation from the Arabicspeaking world; (ii) contact with regional languages, which has affected them
strongly; and (iii) multilingualism of speakers.
The Anatolian dialects have diverged much more from the Standard type of
Arabic compared to the other qəltu dialects, such as the Tigris or Euphrates
groups (Jastrow 2011b). One of the hallmarks of Anatolian Arabic is the suffix -n
instead of -m in the second and third person plural (e.g. in Mardin Arabic baytkən
‘your (pl) house’, baytən ‘their house’) and the negation mō with the imperfect. In
136
6 Anatolian Arabic
addition to many interesting properties like the ones just mentioned, Anatolian
Arabic has acquired a large number of interesting contact-induced patterns.
These dialects are spoken as minority languages by speakers belonging to
different ethnic or religious groups. As noted by Jastrow (2006), not all of the
Anatolian Arabic varieties are spoken in situ however, and in fact some may no
longer be spoken at all. Jastrow notes that some of the dialects were exclusively
spoken by Christians and almost died out during World War One as a result of
the massacres of the Armenians and other Christian groups. A few thousand
speakers of these dialects survive to this day, most of whom have migrated to
big cities, starting from the mid-1980s, particularly Istanbul. Some speakers of
these dialects also live in Europe. Nevertheless, these dialects are very likely to
face extinction in a few decades.
The Jews who spoke Anatolian Arabic varieties (mainly in Diyarbakır, but also
in Urfa and Siverek; cf. Nevo 1999) migrated to Israel after the foundation of the
State of Israel in 1948. These dialects also face a serious threat of extinction.
Today Anatolian Arabic dialects are predominantly spoken by Muslims (although there are a few hundred Arabic-speaking Christians, particularly in some
parts of Istanbul, such as Samatya). These dialects are still found in situ, however
they are also subject to constant linguistic pressure from Turkish (the official
language) and Kurdish (the dominant regional Indo-Iranian language), and social pressure to assimilate. The quote from Grigore (2007a: 27) summarizes the
overall context of Anatolian Arabic: “il se situe dans un microcontexte kurde,
situé à son tour dans un macrocontexte turc, étant isolé de la sorte de la grande
masse des dialectes arabes contemporains.”2
The total number of speakers is around 620,000 (Procházka 2018: 162), most
of whom are bi- or trilingual in Arabic, Kurdish and Turkish. As Jastrow (2011a:
88) points out, the phenomenon of diglossia is not observed in Anatolia; instead
Turkish occupies the position of the ‘High variety’, and Anatolian Arabic, the
‘Low variety’, occupies a purely dialectal position. In addition, speakers of different dialects may speak other minority languages as well. For instance, a considerable number of Sason Arabic speakers know the local variety of the Iranian
language Zazaki, and those of Armenian origin speak an Armenian dialect.
Anatolian Arabic varieties are in decline among the speakers of these varieties,
and public life is dominated primarily by Turkish (and Kurdish). The presence of
Arabic in Turkey has increased due to Syrian refugees who fled to Turkey, yet this
increased presence primarily concerns Syrian Arabic, rather than Anatolian Arabic (see Procházka, this volume). In addition to the absence of awareness about
2
“It is situated in a Kurdish microcontext, which is in turn situated in a Turkish macrocontext,
thus being isolated from the vast majority of contemporary Arabic dialects”.
137
Faruk Akkuş
Anatolian Arabic dialects in the Arab states, the Anatolian dialects also suffer
from a more general lack of interest. The speakers generally do not attribute any
prestige to their languages, calling it “broken Arabic”, and often making little effort to pass it on to the next generations. It should, however, be noted that there
has been increasing interest in these dialects in recent years, especially at the
academic level. To this end, several workshops have been organized at universities in the relevant regions, aimed at promoting these dialects and discussing
possible strategies for their preservation.
The data referenced in this chapter come from various Anatolian Arabic dialects. The name of each variety and its source(s) are as follows: Āzəḫ (Wittrich
2001); Daragözü (Jastrow 1973); Ḥapəs (Talay 2007); Hasköy (Talay 2001; 2002);
Kinderib (Jastrow 1978); Mardin (Jastrow 2006; Grigore 2007b; Grigore & Biţună
2012); Mutki-Sason (Akkuş 2016; 2017; Isaksson 2005); Siirt (Biţună 2016; Grigore
& Biţună 2012); Tillo (Lahdo 2009).
2 Contact languages
2.1 Overview
Anatolia, especially the (south)eastern part, has been home to many distinct linguistic groups (as well as ethnic and religious groups). Up until the beginning
of the twentieth century, speakers of the largest Anatolian languages – Kurdish, Zazaki, Armenian, Aramaic and Arabic – had been co-existing for almost
a thousand years. This has naturally resulted in extensive contact among these
languages.
Contact influence on Anatolian Arabic has arisen mainly through long-term
bi- and multi-lingualism rather than through language shift (in which speakers of
other languages shifted to Arabic; Thomason 2001).3 As a result, when applicable,
the changes seem to be primarily through borrowing, rather than imposition (in
the sense of Van Coetsem 1988; 2000).
2.2 Turkish
Turkish, as the official language of Turkey, currently dominates public life in
most Arabic-speaking areas. However, as noted by Haig (2014: 14), “the current omnipresent influence of Turkish in the region is in fact a relatively recent
phenomenon, fueled by compulsory Turkish-language state education, the massmedia, and large-scale military operations carried out by the Turkish army in the
3
But note also the case of the Mhallamiye near Midyat, who most likely were Aramaic speakers
and shifted to Arabic after adopting Islam as their religion (thanks to Stephan Procházka for
bringing this to my attention).
138
6 Anatolian Arabic
conflict against militant Kurdish groups. But prior to the twentieth century, the
influence of Turkish in many parts of rural east Anatolia was negligible.”
Although Turkish is the dominant language in the public sphere, there are
still many people, particularly in rural parts of (south-)eastern Turkey, who do
not speak Turkish, including speakers of Anatolian Arabic varieties. It is usually
women over forty years old that fall into this category. They tend to speak the
local Arabic variety along with the dominant language in that geographic area.
Moreover, the amount of Turkish influence is greater on the Arabic speakers
who have migrated to bigger cities such as Istanbul, compared to those who still
speak their dialects in situ.
2.3 Kurdish and Zazaki
Anatolian Arabic has been in intensive contact with two Western Iranian languages: Kurmanji Kurdish and Zazaki. These languages have influenced each
other on different levels. As noted by Procházka (this volume) and Öpengin (this
volume), Kurdish and Arabic, including the region of south-eastern Anatolia,
have experienced extensive contact since at least the tenth century.
Due to the multi-ethnic (and to a lesser extent multi-religious) nature of the
regions, bilingualism between Arabic and Kurdish (or Zazaki) is very widespread.
The speakers of the non-dominant languages tend to have a stronger command
of the dominant languages than the reverse situation. For instance, in Mutki,
Bitlis province, where Kurdish is the dominant regional language, Arabic speakers have a native-like command of Kurdish, whereas not many Kurdish speakers
speak the local Arabic variety. In some parts of Sason, Batman province, on the
other hand, Arabic is the dominant language, and Kurdish speakers learn Arabic
as a second language.
2.4 Aramaic
Aramaic and Arabic have for centuries lived side by side, so that it is possible to
speak of both substrates (from Syriac/Neo-Aramaic to Arabic), and of adstrates,
or rather, of superstrates (from Arabic to Aramaic). In the context of Anatolian
Arabic, Aramaic has been in contact mainly with the Mardin dialect group.
These two languages have influenced each other in many ways. For instance,
the many dialects constituting Modern Eastern Aramaic show considerable diversity as to choice of verbal particles. Some dialects use particles similar in form and
function to those of the qəltu-dialects (see e.g. Jastrow 1985; as well as Coghill,
this volume, for North-Eastern Neo-Aramaic dialects).
139
Faruk Akkuş
Finally, it is worth mentioning that, given the existence of Arabic speakers
of Armenian origin, Armenian might have influenced certain Anatolian Arabic
varieties. However, the influence of Armenian is hardly known, apart from the
fact that many villages in the further eastern part of Anatolia, in which Arabic
was spoken or is still spoken, bear Armenian names. This requires further investigation in its own right.
3 Contact-induced changes in Anatolian Arabic
Anatolian Arabic dialects manifest considerable variation, and have also come
to exhibit interesting patterns due to language contact in every linguistic aspect.
This section surveys these changes and features in turn.
3.1 Phonology
Anatolian Arabic has undergone significant changes in its consonant and vowel
inventories due to language contact (as well as language internal developments).
These changes include the introduction of new consonantal phonemes, loss or
weakening of emphatic consonants, and introduction of new vowels. In addition
to these changes, it is possible to count word-final devoicing as a contact-induced
change.
This section first introduces the consonant inventory in varieties of Anatolian
Arabic. It should be noted that not all consonants are present in every variety,
but the chart serves as the sum of consonants available across Anatolian Arabic
varieties. For instance, the phonology of Sason Arabic (and other varieties of the
Kozluk–Sason–Muş group) is characterized by the (near) absence of pharyngeal
and emphatic (pharyngealized) consonants,4 which have fused with their plain
counterparts, e.g. pasal ‘onions’ in Sason < Old Arabic (OA) baṣal.5
Table 1, with information largely taken from Jastrow (2011a), demonstrates that
Anatolian Arabic has several consonants that were originally alien to Arabic (see
§3.1.1 for discussion). With respect to the inventory of vowels, the noteworthy
development is the introduction of /ē/ and /ō/ for some lexical items. Note that
4
These sounds, whose emphatic quality is indicated in Table 1 and throughout with a subscript
dot are only nearly absent for two reasons: (i) it is possible to detect them in the speech of
elderly speakers in some lexical items, while the younger generations have lost them, (ii) Talay
(2001) reports their availability in Hasköy, Muş province to a certain extent.
5
Compare Cypriot Maronite Arabic (Walter, this volume), Maltese (Lucas & Čéplö, this volume)
and Nigerian Arabic (Owens, this volume).
140
6 Anatolian Arabic
Nasal
Vibrant
Lateral
Approximant
f
v
m
θ
ð ð̣
sṣ
z
n
rṛ
l
w
ḫ
ɣ
ttal
q
Glo
k
g
P ha
r yn
lar
gea
l
č
ǧ
š
ž
Uvu
ata
l
Pal
talv
eola
r
tṭ
dḍ
Affricate
Fricative
Pos
al
Den
t
rde
nt
p
b
Vel
ar
Plosive
Inte
Lab
ial
al
Table 1: Inventory of consonants. Marginal or doubtful phonemes
within parentheses
(ʔ)
ḥ
ʕ
h
y
the Old Arabic diphthongs *ay and *aw have largely been preserved in these varieties: Jastrow (2011a: 89) notes that one of the processes by means of which these
mid long vowels entered the inventory of Anatolian Arabic is via loanwords from
Turkish and Kurdish, e.g. commonly used items, čōl ‘desert’, tēl ‘wire’ (Turkish,
probably through the intermediary of Kurdish), ḫōrt ‘young man’ (Kurdish).
3.1.1 New phonemes /p, č, ž, g, v/
The Anatolian Arabic varieties, as well as the varieties in (northern) Syria and
Iraq, have certain phonemes that were not originally familiar to these varieties
of Arabic. These phonemes include the voiceless bilabial stop /p/, the voiceless
affricate /č/, the voiced post-alveolar fricative /ž/,6 the voiced velar stop /g/, and
the voiced labiodental fricative /v/.7 The emergence of these phonemes is most
likely due to the massive contact with Turkish, Kurdish and Aramaic. That is,
the most likely scenario is that the centuries-long borrowing of words which
contained these sounds ultimately resulted in them getting incorporated into
the phonemic inventory.
6
Cf. Jastrow (2011a) and Grigore & Biţună (2012) regarding the status of /ž/: this sound is largely
restricted to borrowed words. The reflex of Arabic 〈 〉جin Anatolian Arabic is /ǧ/.
7
Blanc (1964: 6–7) considers /p/ and /č/ as characteristic of Mesopotamian varieties.
141
Faruk Akkuş
With regard to /v/, it is likely that there are two paths of emergence: (i) as an internal evolution of the voiced interdental fricative /ð/ and (ii) via loan-words from
Turkish and Kurdish. The forms vīp and zīp ‘wolf’ (cf. OA ðiʔb ‘wolf’) represent
a language internal development, whereby the interdental fricatives have shifted
to sibilants in Kozluk–Sason–Muş, and to labiodental fricatives in Āzəx (Şırnak
province, Wittrich 2001), whereas they have been retained in most Mardin group
dialects.8
In many cases, it is impossible to pinpoint which language these sounds were
(initially) borrowed from. However, as also noted in Procházka (this volume), /p/
was probably introduced via contact with Kurdish, followed by influence from
Ottoman and Modern Turkish.9 Some illustrations are as follows:
(1)
pīs ‘dirty’, cf. Kurdish/Turkish pîs, pis
parčāye ‘piece’, cf. Turkish parça
pūz ‘nose’ (Ḥapəs), cf. Kurdish poz
davare ‘ramp’, cf. Kurdish dever fem. ‘place’
čuvāle ‘sack’, cf. Turkish çuval
pēlāv (Ḥasköy) ‘shoe’, cf. Kurdish pêlav
čāy ‘tea’, cf. Turkish çay
čaqmāq ‘lighter’, cf. Turkish çakmak
rēnčbarī (Hasköy), rēžbarī (Sason) ‘husbandry’, cf. Kurdish rêncberî
žīžo (Āzəḫ) ‘hedgehog’, cf. Kurdish jîjo
ṭāži ‘greyhound’, cf. Kurdish tajî
gōmlak ‘shirt’, cf. Turkish gömlek
magzūn, mazgūn (in Sason) ‘sickle’, cf. Syriac magzūnā; Ṭuroyo magzūno
Talay (2007) suggests that the loss of the phonemic status of the emphatic
consonants and the weakness of the pharyngeal in Kozluk–Sason–Muş group is
likely due to the influence of Turkish, which does not have them. Examples are
from the Hasköy dialect, and are taken from Talay (2007: 181):
(2) ata ‘he gave’ (< *ʔaʕṭā), cf. adā in Sason
sēbi ‘boy’ (< *ṣabiyy)
zarab ‘he hit’ (< *ð̣arab < *ḍarab)
Thus, changes of this kind can be seen as a quasi-adaptation of the consonant
inventory to that of the superstrate and adstrate languages.
8
For more discussion, see Wittrich (2001), Jastrow (2011a), Grigore (2007b), Talay (2011), Akkuş
(2017), and Biţună (2016) among others.
9
For further illustrations and discussion, see e.g. Vocke & Waldner (1982), Jastrow (2011a), Talay
(2002; 2007) and Grigore & Biţună (2012).
142
6 Anatolian Arabic
3.1.2 Word-final devoicing
Certain voiced stops in Anatolian Arabic /b, d, ǧ, g/ have a tendency to become
devoiced [p, t, č, k] when they occur word-finally, probably due to Turkish influence, which is well-known for this property.
For instance, /b/ is mainly realized as the voiceless [p] in final pre-pausal position, e.g.: anep ‘grape(s)’, cf. OA ʕinab; ɣarīp ‘stranger’, cf. OA ɣarīb. This might
reflect a change in progress, as Lahdo (2009) points out that the incidence of
devoicing in other Anatolian dialects is also increasing over time. Note that the
devoicing process does not take place in all instances, supporting the claim that
the language is undergoing a transition in this regard. Moreover, the lack of a
written form removes a possible brake on this process. Further illustrations are
as follows:
(3)
axa[θ] ‘he took’
kata[p] ‘he wrote’ (Mardin; Jastrow 2011a: 90)
ktē[p] ‘book’ (Mardin), cf. OA kitāb
baʕī[t] ‘far’ (Āzəḫ), cf. OA baʕīd
aṭya[p] ‘nicer’ (Tillo), cf. OA ʔaṭyab
azya[t] ‘more’, cf. OA ʔazyad (Lahdo 2009: 106)
Devoicing is not limited to word-final position, however, but is also attested before voiceless consonants, e.g. haps ‘prison’, cf. OA ḥabs.
3.2 Morphology
The influence of language contact is also observable in the domain of morphology.
For example, as discussed by Prochazka (2018: 182–183), the numerals 11–19 in the
Kozluk–Sason region show inversion of the unit and decimal positions, e.g. ʕašṛa
sətte (and not sətt ʕašra) ‘sixteen’. See also Procházka (this volume) for discussion
of the personal pronouns.
Some other cases of contact-induced changes such as reduplication, degree in
adjectives and compounds are discussed below.
3.2.1 Reduplication
A type of reduplication due to contact with Turkish produces doublets with /m/.
The consonant /m/ may be added initially to vowel-initial words, as in (4a), or
replaces the initial consonants in consonant-initial words, as in (4b) (see Akkuş
2017; Lahdo 2009). The reduplication conveys vagueness, with a meaning paraphrasable with ‘et cetera’ or ‘something like that’.
143
Faruk Akkuş
(4)
Sason Arabic
a. ažīn m-ažīn
dough m-dough
‘dough or something like that’
b. hās m-ās
lā təso
sound m-sound neg make.impf.2pl
‘Don’t make any noise!’ (Lit. ‘Don’t make sound or something like
that.’)
Following the same restriction in Turkish, if a word starts with /m/, this type of
reduplication is disallowed, e.g. māse ‘table’ cannot be reduplicated in a way that
would result in māse māse.
3.2.2 Degree in adjectives
Adjectives in Anatolian Arabic follow the noun directly, agreeing with it in gender, number, and definiteness. In this respect, the situation is similar to most
Arabic varieties. Degree, on the other hand, is not an inflectional category in
Sason Arabic. Instead, this dialect has adopted the Turkish adverbs daha ‘more’
and en ‘most’ for comparative and superlative, respectively. Both these items
precede the adjectival constituent, as shown in (5a) and (5b).
(5) Sason Arabic
a. mənn-i
daha koys-e
ye
from-obl.1sg more beautiful-f cop.3sg
‘She is more beautiful than me.’
b. en gbīr
most big
‘the biggest’
The Tillo variety also uses the Turkish-derived an ‘most’ in superlative forms,
with both Arabic-derived adjectives (in the elative form) and Turkish-derived
adjectives (which lack an elative form), as in (6a) and (6a) respectively.
(6) Tillo Arabic (Lahdo 2009: 198)
a. an10 aṭyap
most delicious.ela
‘the most delicious’
10
b. an yāqən
most close
‘the closest’
Lahdo (2009) describes this vowel as “short front-to-back unrounded” in Tillo.
144
6 Anatolian Arabic
On the other hand, the comparative in the Tillo variety is formed through the
elative alone (which functions in other Arabic varieties as both comparative and
superlative). The standard of comparison is introduced by the preposition mən
‘from’.
(7)
Tillo Arabic (Lahdo 2009: 162)
təllo iyy
aṭyap
mən əṣṭanbūl
Tillo cop.3sg good.ela than Istanbul
‘Tillo is better than Istanbul.’
3.2.3 Derivational affixes
Through numerous loanwords, a few derivational suffixes have been introduced
into Anatolian Arabic. These suffixes include the agentive morpheme -ǧi/-či, and
the abessive suffix -səz, which translates as ‘without’. Ingham (2011: 178) points
out that these suffixes, especially the former, are also found in the dialects of Iraq,
Syria, and elsewhere (see also Procházka-Eisl 2018 for further details).
(8)
a. Sason Arabic
gahwa-ǧi
coffee-agt
‘coffee maker’
b. Sason Arabic
viǧdan-səz
conscience-abess
‘unconscientious’
c. Tillo Arabic (Lahdo 2009: 199)
kəlla kānu
mṭahhər-či-yye
all be.prf.3pl circumcizer-agt-pl
‘They all were circumcizers.’
The presence of these suffixes on lexemes of the local Arabic varieties, e.g. ḫāserǧi ‘yogurt maker, yogurt seller’ (Sason Arabic) or mṭahhər-či ‘circumsiser’ (Tillo
Arabic), suggests that the forms above are not necessarily adopted as a whole.
Rather, Arabic speakers may decompose the word and apply the suffix to other
lexemes in some cases.
145
Faruk Akkuş
3.2.4 Compounds
Anatolian Arabic has borrowed the N+N compounding strategy from Turkish,
where the right-hand member carries the compound linker morpheme -i. This
pattern is not generally found in other varieties of Arabic and it is most likely
due to contact with Turkish. This type of compound is often used with whole
Turkish phrases. The examples are as follows (note that the buffer consonant -s
appears between the linker morpheme and the noun when the noun ends in a
vowel):
(9)
Sason Arabic (Akkuş & Benmamoun 2018: 41)
a. lisa
mudur-i
high_school director-link
‘high school director’
b. qurs oratman-i
course teacher-link
‘course teacher’
(10) Tillo Arabic (Lahdo 2009: 199)
fabriqa-si
fəstəq
pistachio factory-link
‘pistachio factory’
This compounding strategy is found in other Arabic varieties spoken in Turkey
as well, for instance, buz dolab-i ‘refrigerator’ (lit. ‘ice cupboard-link’) in the
Adana dialect. Whether compounding has been borrowed as a productive process
as opposed to borrowing of the whole phrase requires further investigation.11
3.2.5 Vocative ending -o
Another morphological feature that Anatolian Arabic has acquired is the vocative particle -o. When addressing a person directly, -o is commonly affixed to
kinship terms and given names. This appears to be available in the whole area.
Unlike the situation in Syria and Iraq (see Procházka, this volume), this form of
address is not usually used hypocoristically. Some examples are below:
(11) amm-o ‘(paternal) uncle!’
ǧemāl-o ‘Cemal!’
ḫāl-o ‘(maternal) uncle!’
11
Thanks to Stephan Procházka for the discussion and the example from the Adana dialect.
146
6 Anatolian Arabic
The corresponding forms of feminine nouns end in -ē, as in ḥabībt-ē ‘darling!’.
Grigore (2007a: 203) suggests that this vocative -o is borrowing of a morphological form from Kurdish (cf. Haig & Öpengin 2018), since the suffix, with masculine
and feminine forms, is not historically available in Arabic. Note, however, the
existence of cognates in other Semitic languages and -u in the whole of North
Africa, where Kurdish influence is not likely (see Prochazka, this volume).
In brief, contact with Turkish and other neighboring languages has led to various noticeable changes in the morphology of Anatolian Arabic, particularly the
more easterly varieties.
3.3 Syntax
Research on the syntax of Anatolian Arabic varieties, let alone work on contactinduced syntactic changes, lags significantly behind the research conducted on
other aspects of these languages. Several factors might have contributed to this
situation. Researchers’ tendency to focus on phonological or lexical aspects and
the lack of sufficient data from which to draw conclusions are two possible factors. Another possibility that Ingham (2005) raises for contact-induced syntactic
change is that since the languages in contact are so typologically different, it is
difficult for them to adopt syntactic features from each other without extensive
language change taking place.
This section introduces several syntactic phenomena that can be attributed to
language contact, including copulas, marking of indefiniteness, light verb constructions and the periphrastic causative. Although the details are not elaborated on here, the conclusion we can arrive at is in line with Ingham (2005), in
that the degree and intensity of contact with the neighboring languages leads
to differences among Anatolian Arabic dialects. The more easterly varieties, e.g.
the Kozluk–Sason–Muş group, appear to be the most innovative, and the dialect
group(s) most influenced by the language contact, whereas the Mardin group
appears to be the most conservative (see Akkuş 2017; Jastrow 2011a for further
discussion).
3.3.1 Copula
One of the most distinctive features of Anatolian Arabic is the existence of the
copula in nominal sentences, based on the independent pronouns. This copula
is realized as an enclitic suffix in most Anatolian dialects. Although researchers
seem to differ with respect to the degree of the influence, they converge on the
view that it is a matter of language contact, and that at least the development
147
Faruk Akkuş
and the proliferation of the obligatory copula is under the influence of the neighboring languages – Turkish, Kurdish, Zazaki and Aramaic – which all have copulas in nonverbal clauses (see Lahdo 2009; Grigore 2007b; Palva 2011; Talay 2007;
Jastrow 2011a; Akkuş 2016; 2017; Akkuş & Benmamoun 2018, for more discussion
and illustrations).
Although the copula forms themselves are not imported, the way they are used
in Anatolian Arabic is exactly the same as it is in Kurdish, Turkish and Ṭuroyo
(Aramaic), which have copula in the present tense. The copula is placed after the
predicate (examples from Grigore 2007b).
(12)
a. Kurdish
bav-ê
min
şivan-e
father-ez poss.1sg shepherd-cop.3sg
‘My father is a shepherd.’
b. Turkish
baba-m
çoban-dır
father-poss.1sg shepherd-3sg
‘My father is a shepherd.’
c. Ṭuroyo
bab-i
rəʕyo-yo
father-poss.1sg herder-3sg
‘My father is a herder.’
Some examples from Anatolian Arabic are illustrated in (13).12
(13)
12
a. Kinderib Arabic (Jastrow 1978: 131)
malīḥ-we
beautiful-3sg.m
‘He is beautiful.’
b. Sason Arabic
raḫw-īn nen
sick-pl 3pl
‘They are sick.’
It should be noted that the copula is not necessarily realized as an enclitic in some dialects. For
instance, in the dialect of Siirt (Jastrow 2011a) the copula precedes the predicate. Moreover, the
copula is identical to the personal pronoun in Siirt, whereas other Anatolian varieties use the
shortened version of the pronoun in the 3sg and 3pl. See Akkuş (2016) for some discussion.
148
6 Anatolian Arabic
c. Daragözü Arabic (Jastrow 1973: 40)
nā ḅāš-nā
1sg good-1sg
‘I am good.’
In negative sentences as well, the same order of morphemes is attested. The negative morpheme (and the copula if there is one) follows the predicate in the neighboring languages, as the sentences in (14) show.
(14)
a. Turkish
hasta değil-ler
sick neg.cop-3pl
‘They are not sick.’
b. Kurdish
kemal xwendekar nîn-e
neg-cop.3sg
Kemal student
‘Kemal is not a student.’
c. Zazaki
cinya niwaş ni-yo
child sick neg-cop.3sg
‘The child is not sick.’
The same order is found in Sason Arabic, in that the neg+cop follows the predicate.13
(15)
Sason Arabic
nihane me-nnen
here neg-cop.3pl
‘They are not here.’14
Given that the copula is almost unknown in other Arabic speaking areas (but see
Blanc 1964; also Lucas & Čéplö, this volume; Walter, this volume), it is safe to assume that the development of a full morphological paradigm for the copula along
with its syntactic function is at least facilitated by contact with the neighboring
languages.
13
This is not the most common order in Anatolian Arabic varieties, however. For more discussion,
see Jastrow (2011a) and Akkuş (2016; 2017).
14
In Sason Arabic, the 3pl personal pronoun can be innen or iyen. A shortened version of this
pronoun is used both in affirmative, as in (13b) and negative, as in (15), non-verbal clauses.
149
Faruk Akkuş
3.3.2 Light verb construction
Light verb constructions are another domain where the influence of contact is
clearly manifested. In surrounding languages, particularly Turkish, Kurdish and
Zazaki, a light verb construction consists of a nominal part followed by the light
verb, which is usually ‘to do’ or ‘to be’, e.g. Kurdish pacî kirin (lit. ‘kiss do’) ‘to
kiss’, Turkish motive etmek (lit. ‘motivation do’) ‘to motivate’.
There are a relatively large number of compound verbs constructed with Arabic sāwa – ysāwi ‘to do’ and a nominal borrowed from Turkish or Kurdish, as
illustrated in (16). In the majority of the cases, the construction is a complete
calque of its Turkish or Kurdish counterparts (see e.g. Versteegh 1997; Lahdo
2009; Grigore 2007b; Talay 2007; Jastrow 2011a; Akkuş 2016; 2017; Akkuş & Benmamoun 2018 and Biţună 2016 for more examples).
(16)
a. Āzəḫ/Mardin Arabic (Talay 2007: 184)
sāwa brīndār ‘to injure’, cf. Kurdish brîndar kirin
sāwa ǧāmērtīye ‘to act generously’, cf. Kurdish camêrtî kirin
sāwa ɣōt ‘to mow’, cf. Kurdish cot kirin
b. Tillo Arabic (Lahdo 2009: 202)
sāwa yārdəm ‘to help’, cf. Turkish yardım etmek
ysāwaw dawām ‘they continue …’, cf. Turkish devam etmek
nsayy qaḥwaltə ‘we have breakfast’, cf. Turkish kahvaltı etmek
In Sason Arabic, the default order in this construction has reversed, in that in
most cases the nominal is followed by the light verb. Thus, Sason manifests headfinal order, undoubtedly due to contact with Turkish and Kurdish. Similarly, the
nominal part of the construction can be borrowed from Turkish as in (17), including instances of reborrowing of an originally Arabic word, (17b), or Kurdish as
in (18). In fact, the nominal part might also be Arabic, as in (19).
(17) Sason Arabic (Turkish borrowing)
a. qazan sāwa
win do.prf.3sg.m
‘to win’
(18) Sason Arabic (Kurdish borrowing)
ser
asi
watch do.impf.1sg-do
‘I watch ...’
150
b. išāret sāwa
sign do.prf.3sg.m
‘to sign’
6 Anatolian Arabic
(19) Sason Arabic
a. gerre/hās sāwa
noise/sound do.prf.3sg.m
‘to make noise/sound’
b. šəɣle lā təsi,
aməl si!
talk neg do.impf.2sg.m work do.imp.2sg.m
‘Don’t talk, do work!’
c. huǧūm sinna
attack do.prf.1pl
‘We attacked.’
Anatolian Arabic usually resorts to the same periphrastic construction when borrowing verbs from Turkish; it creates a complex predicate, rather than adapting
a foreign verb directly to Arabic verbal morphology, a borrowing strategy seen
also in the other languages in the region, such as Kurdish, Zazaki. In many cases,
the complex predicate comprises of sāwa + the Turkish verbal form of the indefinite past (i.e. miş-verb), rather than the bare form of the verb, as illustrated in
(20).
(20)
Anatolian Arabic (Talay 2007: 184)
sawa gačənməš ‘to manage’, cf. Turkish geçinmiş
bašlaməš sawa ‘to begin’, cf. Turkish başlamış
Despite the widespread use of this process for loanwords, some borrowed verbal forms have been totally assimilated to the Arabic verbal system; the majority
of these verbs are formed according to verbal measures (stems) II or III, as can
be seen in example (21).
(21)
Āzəḫ (examples from Talay 2007)
Stem II qappat – īqappət ‘to close’
Stem II qayyad – īqayyəd ‘to register’
Stem III ḍāyan – īḍāyən ‘to be patient, to bear up’
Stem III tēlan – ītēlən ‘to rob’
cf. Tr. kapatmak
cf. Tr. kayıt etmek
cf. Tr. dayanmak
cf. Kr. talan kirin
3.3.3 Marking of (in)definiteness
In Classical Arabic and in modern varieties spoken in the Arab world, the indefinite noun phrase is unmarked or is preceded by an independent indefinite particle, whereas an NP becomes definite by prefixing the definite article al-/əl-/l-
151
Faruk Akkuş
etc. (Brustad 2000). However, Kozluk–Sason–Muş group dialects have adopted
the reverse pattern (see also Uzbekistan Arabic; Jastrow 2005), which is found
in the neighboring languages Turkish and Kurdish. That is, the definite NP is
left unmarked, and the enclitic -ma is used to mark the indefiniteness of an NP
(Talay 2007; Akkuş 2016; 2017; Akın et al. 2017; Akkuş & Benmamoun 2018), as
illustrated in (22).
(22)
Sason Arabic mara ‘the woman’ > mara-ma ‘a woman’
bayt ‘the house’ > bayt-ma ‘a house’
The parallel constructions in Kurdish and Turkish are illustrated in (23) and (24)
respectively.
(23)
Kurdish
derɪ̂ ‘the door’ > derɪ́ -yek ‘a door’
(24) Turkish
kadın ‘the woman’ > bir kadın ‘a woman’ (Turkish)
3.3.4 Periphrastic causative
Sason Arabic resorts to periphrastic causative constructions rather than the root
and pattern strategy found in other non-peripheral Arabic varieties. In this respect it is on a par with Kurdish, which uses the light verb bıdın ‘to give’ to form
the causative, as in (25).
(25)
Adıyaman Kurmanji Kurdish (Atlamaz 2012: 62)
mı
piskilet do
çekır-ın-e
obl.1sg bicycle give.ptcp repair.ptcp-ger-obl
‘I had the bicycle repaired.’ (Lit: ‘I gave the bicycle to repairing.’)
Sason Arabic exhibits the same pattern for causative and applicative formation,
as shown in (26), which is most likely as a result of extensive contact with Kurdish.15
15
Sason Arabic also has another periphrastic construction that is formed with the verb sa ‘to
do/make’, which may embed a finite clause (i.a) or a verbal-noun phrase (i.b).
(i) Sason Arabic (adapted from Taylan 2017: 221)
a. doḫtor məša ali ku
isi
fiy-u
(le
yaddel)
doctor to
Ali cop.3sg.m make.impf.3sg.m in-3sg.m (comp make.impf.3sg.m)
sipor
sports
‘The doctor is making Ali do sports.’
152
6 Anatolian Arabic
(26)
Sason Arabic (Taylan 2017: 221)
əmm-a
məša fatma ši adəd-u
addil
mother-obl.3sg.f to
Fatma food give.prf.3sg.f-3sg.m make
‘Her mother made Fatma cook (Lit: Her mother gave food making to
Fatma).’
3.4 Lexicon
Anatolian Arabic dialects have borrowed single words and whole phrases or expressions mainly from (Ottoman and Modern) Turkish and Kurdish. The influence of these two languages on the Arabic lexicon is enormous. Aramaic words
also survive in Anatolian Arabic to a lesser degree. A few illustrations are given
in (27).16
(27)
bōš ‘much’, cf. Kurdish boş
bōšqa ‘different’, cf. Turkish başka
ṛūvi ‘fox’, cf. Kurdish rûvî
hič ‘none, whatsoever’, cf. Turkish hīç
səpor ‘sport’, cf. Turkish spor
magzūn, mazgūn (in Sason) ‘sickle’, cf. Syriac magzūnā; Ṭuroyo magzūno
As Jastrow (2011a: 95) mentions, while more Turkish borrowings are found in
bigger cities such as Mardin, Diyarbakır or Siirt, Kurdish borrowings constitute
a bigger part of the lexicon of rural dialects. Anatolian Arabic dialects which
have preserved the emphatics, pharyngeals or interdentals adapt borrowings into
their phonology. For instance, Turkish halbuki ‘however’ is borrowed as ḥālbūki.
In most cases, the velar k is turned into the uvular q, e.g. čaqmāq ‘lighter’, cf.
Turkish çakmak. Also, Kurdish feminine nouns (and even some Turkish nouns)
are suffixed with the Arabic feminine morpheme -e/-a, e.g. tūre ‘shoulder’ (cf.
Kurdish tûr).
There are several function words that are copied from Turkish into Arabic,
e.g. Turkish ama ‘but’ is realized as hama in Sason, and as aṃa in Tillo Arabic.
b. aɣa
sa
hazd hašīš
headman make.prf.3sg.m cut.inf grass
‘The village headman had the grass cut.’
Although the origin of these constructions is not clear, they do not appear to be contactinduced.
16
See Vocke & Waldner (1982: xxxix–li) for detailed statistics on Kurdish/Turkish/Aramaic loanwords. See also Lahdo (2009: 207–223) for a comprehensive glossary of Turkish and Kurdish
loanwords in Tillo Arabic, most of which are found in other Anatolian Arabic varieties as well.
153
Faruk Akkuş
The conjunction çünkü ‘because’ from Turkish is attested in many Anatolian
varieties, with the same function. Lahdo (2009: 179) notes that it expresses causal
clauses in Tillo, as in (28), and Biţună (2016: 213) reports the same role for Siirt.
Jastrow (1981: 278) and Grigore (2007a: 261) also confirm its existence in Ḥalanze
and Mardin, respectively.
(28)
Tillo Arabic (Lahdo 2009: 179)
mā ʕaṭaw-ni
əzan
čünki ǧītu
əl-anqara
neg give.prf.3pl-1sg permission because come.prf.1sg to-Ankara
‘They did not give me permission because I had come to Ankara.’
Procházka (2005) notes that particles such as bīle < bile ‘even’, or zātan < zaten
‘already’ in the Adana region are also borrowed from Turkish (see also Isaksson
2005).
4 Conclusion
This chapter has dealt with contact-induced changes in the Anatolian Arabic
dialects. We have seen that Anatolian Arabic has been primarily in contact with
Turkish, Kurdish and Aramaic, and the influence of these neighboring languages
on Anatolian Arabic is evident. We have surveyed some contact-induced changes
at the phonological, morphological, syntactic and lexical level.
Mardin and Siirt dialects have been covered much more comprehensively than
other dialects in the literature. It is desirable to have more comprehensive investigations carried out for the dialects around the Bitlis, iliMuş and Diyarbakır areas.
This research has the potential to fill the gaps in our current state of knowledge
about these dialects.
Similarly, in terms of the linguistic features investigated, phonological and
morphological properties (along with lexicon) have received more attention in
the literature, whereas syntax, in particular, has been understudied. This situation, however, might change once we are at a point where we have enough
recordings and transcriptions to investigate syntactic properties of the dialects.
154
6 Anatolian Arabic
Further Reading
) Jastrow (1978) is a seminal work, which provided the classification for Anatolian Arabic varieties.
) Jastrow (2011a) is a concise, yet comprehensive encyclopedia entry on characteristic features of Anatolian Arabic.
) Talay (2011) is a good source for an overview of Arabic dialects in the Mesopotamian region.
Acknowledgements
I would like to thank Stephan Procházka and Mary Ann Walter for sharing their
work, and to Stephan Procházka for reading another version of the paper, which
also helped with this version. Thanks to Gabriel Biţun̆ a for directing my attention
to several important sources. I also thank the editors Chris Lucas and Stefano
Manfredi for their help and patience. All remaining errors are of course my own.
Abbreviations
1, 2, 3
I, II etc.
abess
agt
comp
cop
def
ela
ez
f
ger
impf
inf
1st, 2nd, 3rd person
1st, 2nd etc. verbal derivation
abessive
agentive
complementizer
copula
definite article
elative degree
ezāfe
feminine
gerund
imperfect (prefix conjugation)
infinitive
Kr.
link
m
neg
OA
obl
pl
poss
prf
ptcp
sg
Tr.
Kurdish
linker
masculine
negation
Old Arabic
oblique
plural
possessive
perfect (suffix conjugation)
participle
singular
Turkish
155
Faruk Akkuş
References
Akın, Cahit, Otto Jastrow & Shabo Talay. 2017. Nōṛšēn: Ein anatolisch-arabischer
Dialekt der Sason–Muş–Gruppe. Zeitschrift für Arabische Linguistik 65. 73–96.
Akkuş, Faruk. 2016. The Arabic dialect of Mutki–Sason areas. In George Grigore
& Gabriel Biţună (eds.), Arabic varieties: Far and wide. Proceedings of the 11th
International Conference of Aida, Bucharest 2015, 29–41. Bucharest: Editura Universității din București.
Akkuş, Faruk. 2017. Peripheral Arabic dialects. In Elabbas Benmamoun & Reem
Bassiouney (eds.), The Routledge handbook of Arabic Linguistics, 454–471. London: Routledge.
Akkuş, Faruk & Elabbas Benmamoun. 2018. Syntactic outcomes of contact in
Sason Arabic. In Stefano Manfredi & Mauro Tosco (eds.), Arabic in contact, 38–
52. Amsterdam: John Benjamins.
Atlamaz, Ümit. 2012. Ergative as accusative case: Evidence from Adıyaman Kurmanji. Istanbul: Bogaziçi University. (Master’s dissertation).
Biţună, Gabriel. 2016. Morfo-sintaxa dialectului Arab Nord-Mesopotamian din Siirt,
Turcia. Bucharest: University of Bucharest. (Doctoral dissertation).
Blanc, Haim. 1964. Communal dialects in Baghdad. Cambridge, MA: Harvard
University Press.
Brustad, Kristen. 2000. The syntax of spoken Arabic: A comparative study of
Moroccan, Egyptian, Syrian, and Kuwaiti dialects. Georgetown: Georgetown
University Press.
Grigore, George. 2007a. L’arabe parlé à Mardin: Monographie d’un parler arabe
“périphérique”. Bucharest: Editura Universitãţii din Bucureşti.
Grigore, George. 2007b. L’énoncé non verbal dans l’arabe parlé à Mardin.
Romano-Arabica. 51–63.
Grigore, George & Gabriel Biţună. 2012. Common features of North Mesopotamian Arabic dialects spoken in Turkey (Şırnak, Mardin, Siirt). Bilim, Düşünce ve
Sanatta Cizre (Uluslararası Bilim Düşünce ve Sanatta Cizre Sempozyumu Bildirileri). 545–555.
Haig, Geoffrey L. J. 2014. East Anatolia as a linguistic area? Conceptual
and empirical issues. In Lale Behzadi, Patrick Franke, Geoffrey L. J. Haig,
Christoph Herzog, Birgitt Hofmann, Lorenz Korn & Susanne Talabardon (eds.),
Bamberger Orientstudien, 13–36. Bamberg: University of Bamberg Press.
Haig, Geoffrey L. J. & Ergin Öpengin. 2018. Kurmanji Kurdish in Turkey:
Structure, varieties and status. In Christiane Bulut (ed.), Linguistic Minorities in
Turkey and Turkic-speaking minorities of the peripheries, 157–230. Wiesbaden:
Harrassowitz.
156
6 Anatolian Arabic
Ingham, Bruce. 2005. Persian and Turkish loans in the Arabic dialects of North
Eastern Arabia. In Éva Ágnes Csató, Bo Isaksson & Carina Jahani (eds.), Linguistic convergence and areal diffusion: Case studies from Iranian, Semitic and
Turkic, 173–180. London: Routledge.
Ingham, Bruce. 2011. Afghanistan Arabic. In Lutz Edzard & Rudolf de Jong (eds.),
Encyclopedia of Arabic language and linguistics, online edn. Leiden: Brill.
Isaksson, Bo. 2005. New linguistic data from the Sason area in Anatolia. In Éva
Ágnes Csató, Bo Isaksson & Carina Jahani (eds.), Linguistic convergence and
areal divergence: Case studies from Iranian, Semitic and Turkic, 181–190. London:
Routledge.
Jastrow, Otto. 1973. Daragözü: Eine arabische Mundart der Kozluk-Sason-Gruppe
(Südostanatolien). Nürnberg: Hans Carl.
Jastrow, Otto. 1978. Die mesopotamisch-arabischen qəltu-Dialekte. Vol. 1: Phonologie und Morphologie. Wiesbaden: Steiner.
Jastrow, Otto. 1981. Die mesopotamisch-arabischen qəltu-Dialekte. Vol. 2: Volkskundliche Texte in elf Dialekten. Wiesbaden: Steiner.
Jastrow, Otto. 1985. Laut- und Formenlehre des neuaramäischen Dialekts von Mīdin
in Ṭūr ʿAbdīn. Wiesbaden: Harrassowitz.
Jastrow, Otto. 2005. Uzbekistan Arabic: A language created by Semitic–Iranian–
Turkic linguistic convergence. In Éva Ágnes Csató, Bo Isaksson & Carina
Jahani (eds.), Linguistic convergence and areal diffusion: Case studies from
Iranian, Semitic and Turkic, 133–139. London: Routledge.
Jastrow, Otto. 2006. Arabic dialects in Turkey: Towards a comparative typology.
Türk Dilleri Araştırmaları 16. 153–164.
Jastrow, Otto. 2011a. Anatolian Arabic. In Lutz Edzard & Rudolf de Jong (eds.),
Encyclopedia of Arabic language and linguistics, online edn. Leiden: Brill.
Jastrow, Otto. 2011b. Iraq: Arabic dialects. In Lutz Edzard & Rudolf de Jong (eds.),
Encyclopedia of Arabic language and linguistics, online edn. Leiden: Brill.
Lahdo, Ablahad. 2009. The Arabic dialect of Tillo in the region of Siirt (south-eastern
Turkey). Uppsala: Uppsala University.
Nevo, Moshe. 1999. Notes on the Judaeo-Arabic dialect of Siverek. Zeitschrift für
Arabische Linguistik 36. 66–84.
Palva, Heikki. 2011. Dialects: Classification. In Lutz Edzard & Rudolf de Jong
(eds.), Encyclopedia of Arabic language and linguistics, online edn. Leiden: Brill.
Procházka, Stephan. 2005. The Turkish contribution to the Arabic lexicon. In Éva
Ágnes Csató, Bo Isaksson & Carina Jahani (eds.), Linguistic convergence and
areal diffusion: Case studies from Iranian, Semitic and Turkic, 191–202. London:
Routledge.
157
Faruk Akkuş
Procházka, Stephan. 2018. The Arabic dialects of eastern Anatolia. In Geoffrey
L. J. Haig & Geoffrey Khan (eds.), The languages and linguistics of western Asia:
An areal perspective, 159–189. Berlin: De Gruyter Mouton.
Procházka-Eisl, Gisela. 2018. A suffix on the move: Forms and functions of the
Turkish suffix -ci in Arabic dialects. Mediterranean Language Review 25. 21–52.
Talay, Shabo. 2001. Der arabische Dialekt von Hasköy (Dēr Khāṣ), Ostanatolien,
I: Grammatische Skizze. Zeitschrift für Arabische Linguistik 40. 71–89.
Talay, Shabo. 2002. Der arabische Dialekt von Hasköy (Dēr Khāṣ), Ostanatolien,
II: Texte und Glossar. Zeitschrift für Arabische Linguistik 41. 46–86.
Talay, Shabo. 2007. The influence of Turkish, Kurdish and other neighbouring
languages on Anatolian Arabic. Romano-Arabica. 179–189.
Talay, Shabo. 2011. Arabic dialects of Mesopotamia. In Stefan Weninger, Geoffrey
Khan, Michael P. Streck & Janet C. E. Watson (eds.), The Semitic languages: An
international handbook, 909–919. Berlin: De Gruyter Mouton.
Taylan, Eser. 2017. Language contact in Anatolia: the case of Sason Arabic.
In Ramazan Korkmaz & Doğan Gürkan (eds.), Endangered languages of the
Caucasus and beyond, 209–225. Leiden: Brill.
Thomason, Sarah G. 2001. Language contact: An introduction. Edinburgh: Edinburgh University Press.
Van Coetsem, Frans. 1988. Loan phonology and the two transfer types in language
contact. Dordrecht: Foris.
Van Coetsem, Frans. 2000. A general and unified theory of the transmission process
in language contact. Heidelberg: Winter.
Versteegh, Kees. 1997. The Arabic language. 1st edn. Edinburgh: Edinburgh University Press.
Vocke, Sybille & Wolfram Waldner. 1982. Der Wortschatz des anatolischen
Arabisch. Erlangen: W. Waldner.
Wittrich, Michaela. 2001. Der arabische Dialekt von Āzǝx. Wiesbaden:
Harrassowitz.
158
Chapter 7
Cypriot Maronite Arabic
Mary Ann Walter
Middle East Technical University, Northern Cyprus Campus
Cypriot Maronite Arabic is a severely endangered variety that has been in intensive
language contact with Greek for approximately a millennium. It presents an interesting case of a language with extensive contact effects which are largely limited
to the phonological domain.
1 Current state and historical development of Cypriot
Maronite Arabic
Cypriot Maronite Arabic (CyA) is a minority language spoken by a small community on the island of Cyprus. Although essentially moribund, it is currently
the focus of preservation and revitalization efforts.
1.1 Historical development of Cypriot Maronite Arabic
The time of arrival of this community of Arabic speakers to Cyprus is unknown.
The island was occupied by an Arab garrison subsequent to Muʕāwiya’s invasion of 649 CE, but the garrison was then removed and, presumably, the Arabic
speakers left as well. More likely, a permanent presence dates back to the population movements of the ninth and tenth centuries during disruptions to Byzantine
rule.1 Subsequent waves of Arab emigration to Cyprus are documented during
the early crusading period. Such movements also quite likely took place during
Lusignan (French crusader) rule in Cyprus (1192–1489), for some portion of which
the Anatolian city of Adana, where Arabic is still widely spoken (see Procházka,
1
See §2 for a discussion of where the CyA-speaking community originated from and the dialectological affiliation of this variety of Arabic.
Mary Ann Walter. 2020. Cypriot Maronite Arabic. In Christopher Lucas & Stefano
Manfredi (eds.), Arabic and contact-induced change, 159–174. Berlin: Language Science
Press. DOI:10.5281/zenodo.3744513
Mary Ann Walter
this volume), was also held by Lusignan rulers. Speakers of not only Arabic but
a locally distinct version of Arabic in Cyprus are mentioned by Arab historians
beginning in the thirteenth century, thereby providing a terminus ad quem to its
dialectal development (Borg 2004).
As fellow communicants in the Catholic church, the Maronite community was
granted certain privileges of independent worship during the Lusignan period,
which were later lost during Venetian (1489–1571) and Ottoman rule (1571–1878),
at which time some retaliation occurred on the part of the Orthodox community
(Gulle 2016). After the Ottoman conquest of Cyprus in 1571, the Maronite community was at first placed under the administration of the Orthodox bishop, but
regained religious autonomy shortly thereafter.
The number of Maronite villages underwent a steady decline during the Ottoman period, from over thirty to only five at the time of British occupation of the
island in 1878 (Baider & Kariolemou 2015; though it is unclear if this is associated
with any actual population decline). The remaining five villages are all located in
the northwestern area of the island. However, as of the twentieth century at least,
only one of them was home to speakers of CyA, the others having linguistically
assimilated to Cypriot Greek entirely. The CyA-speaking village is Kormakiti(s)
(also known as Kormacit and Koruçam in CyA and in Turkish, respectively).
Both the Cypriot liberation struggle of the 1950s against the British, and the
years after independence was attained in 1960, saw increased communal conflict between the Turkish and Greek communities on the island. This period witnessed increasing separation of communities, as Turkish Cypriots withdrew into
ethnic enclaves, and culminated in the 1974 conflict between Greece-sponsored
coup plotters, military forces of Turkey, and local Cypriots on various sides, the
result of which was a de facto division of the island between the Republic of
Cyprus-controlled territory in the south, which was majority Greek Orthodox
and Greek-speaking, and the Muslim and Turkish-speaking northern part of the
island. This northern area subsequently declared independence, but remains unrecognized by any other country except the Republic of Turkey to this day.
It is important to note that the relative geographical separation between Greek
Cypriots and Turkish Cypriots dates only from this recent period, as refugees
sought safety within their own communities. This entailed a radical change in the
social circumstances of CyA speakers, who moved to the capital city of Nicosia
essentially en masse. Thus, they went from living in a Maronite village in which
community life could be conducted in CyA, to being a tiny percentage of a large
urban population. Not only that, but the pre-1974 population surrounding the
CyA-speaking Maronite village of Kormakiti was composed of Greek speakers,
whereas the current local population around the village is comprised of Turkish
160
7 Cypriot Maronite Arabic
speakers (many of whom also know Greek, but no longer use it as a language of
public life).
Since 1974 the permanent population of the village of Kormakiti has amounted
to at most a couple of hundred residents, with the rest of the Maronite community residing primarily in the capital city Nicosia. The Maronite community has
occupied a special place in Cypriot society, as for three decades they alone had
the ability to freely cross the UN-monitored “Green Line” (buffer zone) dividing
the island. Thus connections with the village have been maintained throughout
this period, and weekend visits are common. Since 2003 the line has been crossable for all Cypriots.
1.2 Current situation of Cypriot Maronite Arabic
The Cypriot Maronite community currently numbers roughly 5,000 individuals.
However, only approximately one thousand are CyA speakers (estimates range
from 900 to perhaps 1300; Council of Europe 2017).
All CyA speakers are bilingual in Cypriot Greek, with Greek as their dominant
language, and currently living in a heavily Greek-dominant urban area. There are
currently no fluent native speakers under the age of thirty. Due to these factors,
the CyA language was designated as severely endangered by UNESCO in 2002.
However, the accession of the Republic of Cyprus to the European Union in
2004 has led to an influx of both institutional and financial support for CyA. In its
2004 initial report on its implementation of the European Charter for Regional or
Minority Languages (ECRML), which it ratified in 2002, the Republic of Cyprus
declared Armenian as such a language in Cyprus. Although CyA was explicitly
excluded as being “only” a dialect and therefore in no need of protection, this
formulation was not accepted by ECRML, and CyA was thenceforth officially
recognized as a minority language of Cyprus as well. Since 2008 Maronites have
been officially recognized as a separate community within Cyprus, and are no
longer required to identify themselves as Greek Cypriots (or Turkish Cypriots)
on government documents.
The change in designation of the Cypriot Maronites as a linguistic as well as
religious minority community led to associated changes in the linguistic rights
legally accorded to them. After decades of waiting, one state school in Nicosia
is now designated as Maronite and offers optional after-school classes in CyA
for its approximately 100 Maronite students, the majority of whom have now
joined the classes. Adults may also study CyA now at the new community center. Funding was also made available for a one-to-two week summer language
immersion camp for Maronite youth in Kormakiti village, attendance at which
161
Mary Ann Walter
has risen to approximately 100. For the first time, training seminars for teachers have also been organized, concomitantly with codification efforts towards a
written version of CyA. Sporadic writing in CyA has been carried out using the
Greek alphabet. (See the community websites in the Further reading section at
the end of this chapter).
Outside the government, there is also an NGO Hki fi Sanna (‘speak in our language’) with the goal of promoting CyA use. Usage remains community- and
home-based, as Standard Greek (and English) is the language of written and
broadcast media. The Cyprus Center of the Peace Research Institute Oslo (PRIO)
has undertaken a project entitled The protection and revival of Cypriot Maronite
Arabic. The scope of the project included a variety of community activities, as
well as meetings with Sami (Norway) community members for sharing revitalization strategies, described in the resulting publication (PRIO 2009). Finally, a
project at the University of Cyprus titled The creation of an archive of oral tradition for Cypriot Maronite Arabic is currently underway under the supervision
of Dr. Marilena Karyolemou, though with no web presence or published deliverables to date. There is thus some reason for optimism regarding the future of
CyA.
2 Contact languages
CyA has undergone intensive language contact with Cypriot Greek for the entirety of its presence in Cyprus, which may extend to a millennium (see §1.1). This
contact has intensified since the removal of the population from the traditionally
Maronite and CyA-speaking village of Kormakiti to the capital city, Nicosia.
This move has also resulted in a concomitantly larger social role for Standard Greek. Cyprus is a diglossic society in which Cypriot Greek coexists with
Standard Greek, the language of education and formal domains.2 In moving to
Nicosia, the children of the community also began attending schools with Greek
Cypriot children, rather than their own village schools. Only in the last few years
has a primary school been designated specifically for Maronite children. Most
of them still attend other schools, and the Maronite school is in any case also
(Standard-)Greek-medium and follows the same national curriculum (with the
addition of optional after-school weekly CyA language classes).
2
Some in fact refer to triglossia, encompassing Standard Greek, koinéized Cypriot Greek, and
various other local varieties, with the island-wide koine taking a mesolectal position (Arvaniti
2010).
162
7 Cypriot Maronite Arabic
Therefore, the influence of Greek has increased radically through contact with
Greek classmates and neighbors, as well as intermarriages with Greek Cypriots
and Maronites from other, non-CyA speaking villages. Such situations are common due to the small size of the Maronite community, and typically CyA is not
used in these households.
In comparison, contact with Turkish has been limited. Although remaining
residents of Kormakiti are now surrounded by Turkish speakers, the village remains quite set apart socially, to the extent that all water supplies are trucked in
rather than plumbing systems being shared. In Borg’s (1985) texts, speakers do
mention using Turkish with some speakers employed as farm workers, however.
Contact with Turkish speakers in Nicosia is, of course, rare.
Cyprus is “double-diglossic”: the same situation as with Cypriot and Standard
Greek holds also with respect to Cypriot and Standard Turkish. To the extent that
contact with Turkish does occur, it is with Cypriot Turkish rather than Standard
Turkish, unlike Greek, where both varieties are prominent in the lives of CyA
speakers.
There is next to no contact with other varieties of Arabic. The Maronite clergy
in Cyprus often come from Lebanon, and some intermarriage occurred in the
more distant past between the Cypriot and Lebanese Maronite communities, but
this no longer occurs. Roth (2004) refers to the “double minoritization” of CyA
speakers with respect to both the Cypriot context and the wider Arabophone
context – in both, their speech variety is considered deviant and unintelligible.
While early research on CyA identifies it as a Levantine variety of Arabic
(Tsiapera 1969), Borg (1985; 2004) argues strongly for an Anatolian origin with
significant Aramaic substrate influence. Because the Aramaic influence, if any,
must have occurred in the pre-Cyprus period, contact with Aramaic will not be
considered further here, despite its putative influence. A substantial discussion
can be found in Borg (2004).
Another Semitic language, Syriac, is the liturgical language of the Maronite
community. However, no instruction is available in Syriac in Cyprus, so its use
is limited to rote recitation during (very sparsely attended) church services, at
which transliterations and Greek translations are also provided.
English is the third official language of the Republic of Cyprus (along with
Greek and Turkish) and is widely spoken. Instruction in English begins in primary school in the national curriculum, and private English-medium schools
are also widespread. However, contact with English postdates contact with Greek
and Turkish (beginning only after 1878 and intensifying in the twentieth century)
and appears to have had no effects on CyA language structures.
163
Mary Ann Walter
The French school in Nicosia is traditionally a popular choice for Maronite
families, so that competence in French has also been common in the community
– a shared characteristic with Lebanese Maronite society. However, like English,
this appears not to have influenced CyA grammar in any significant way.
Remaining minority languages of Cyprus include Armenian and a variety of
Romani locally called Kurbetça/Gurbetça. The reports of ECRML specify that
there has been no contact requested or arranged between the Armenian and
Maronite community institutions, however. The small size of the communities
(each less than 1% of the population) no doubt also reduces the chances of contact.
As for Kurbetça, it is unclear whether or not it is still actually spoken on the
island. Members of this community are Turkish-speaking and interact little if at
all with the Maronite community.
Finally, the most common immigrant language after English is Russian, which
occupies an increasingly prominent place in the linguistic landscape of Cyprus.
There are now several Russian-medium schools on the island. However, these
are primarily located outside the capital, and its recent appearance means that it
also has not influenced CyA.
Therefore, the next section will focus on contact effects from Greek on CyA.
3 Contact-induced changes
According to Borg, the doyen of CyA studies, “linguistic acculturation to Greek
in [CyA] is fairly extensive…and involves transfer of allophonic rules, function
words, and virtually unrestricted borrowing of content words in the context of
codeswitching” as well as “a significant degree of calquing on Greek idioms”
(2004: 64). This occurs to such an extent that he describes CyA as “Greek in
transparent Arabic garb”, although “the degree of hellenization…tends to be concealed…the inflectional pattern of [CyA] having largely resisted significant intrusion of Greek morphological elements” (2004: 65).
In the remainder of this section, we will examine examples of such Greek influence, particularly in the phonological domain. At the same time, the remarkable
persistence of CyA language patterns in the face of intensive contact, especially
in the morphological domain, will be discussed.
3.1 Phonology
CyA phonology has been heavily restructured in comparison with other varieties of Arabic, resulting in what Roth (2004: 55) calls “total convergence” of the
164
7 Cypriot Maronite Arabic
phonological system with Cypriot Greek. Similarly, Gulle (2016: 47) refers to the
“complete adoption of Greek phonology.”
Like other varieties of Arabic in intense contact with non-Semitic languages,
CyA has lost the series of so-called emphatic, guttural or pharyngealized consonants. The obstruents have merged with their non-emphatic counterparts, and
the pharyngeal fricative ḥ has merged with the original glottal fricative h, which
in turn is now pronounced as a velar fricative [x] under the influence of Greek,
as in the examples in Table 1.3
Table 1: Reflexes of emphatic and guttural consonants in CyA
CyA
Arabic
Gloss
taraf
txin
pakar
axsen
ṭaraf
ṭaḥīn
baqar
aḥsan
‘end’
‘flour’
‘cattle’
‘better’
The sole survivals among the Arabic consonants that have no counterparts in
Greek are the interdental consonants and the pharyngeal glide /ʕ/ (see example
5b below). It is interesting that the pharyngeal glide, perhaps the most typologically unusual, remains as a sort of iconic survivor of the Arabic phonemic inventory. The retention of this phoneme, alongside the loss of so many others, implies
that the radical changes to the consonant inventory of CyA, though clearly linked
to Greek influence, cannot be wholly attributed to imposition in the sense of Van
Coetsem (1988; 2000) – or at least, is evidence of significant resistance to such
imposition. In any case, imposition would presumably be due to late learners of
CyA, and it is doubtful that CyA was ever acquired in this way by speakers from
outside the community.
As for the vowels, the Arabic vowel length contrast has also been lost, unstressed (formerly) short vowels deleted, and mid vowels have joined the inventory, resulting in a five-vowel inventory matching that of Greek, as illustrated in
Table 2.
This unsurprising result also occurred in other contact varieties such as Maltese and Andalusi Arabic, although may have evolved without the influence of
contact, as in some Levantine varieties.
3
Examples are taken from Borg’s (2004) glossary except where noted otherwise. CyA forms
are given in his orthography. “Arabic” forms are the presumed etymological source forms,
typically shared by Standard Arabic as well as other varieties.
165
Mary Ann Walter
Table 2: Illustration of the innovative vowel system of CyA
CyA
Arabic
CyA Gloss
ipn
umm
tarp
klep
yaxtop
ten
ibn
umm
darb
kilāb
yaktub
yadayn
‘son’
‘mother’
‘road’
‘dogs’
‘he writes’
‘hands.du’
Phonotactically speaking, CyA remains more permissive than Cypriot Greek,
in that it “allows a wider range of final consonants and is alone [relative to
Cypriot Greek] in allowing final clusters” (Newton 1964: 51).
The effect of (Cypriot) Greek has not been limited to the phonemic inventory.
CyA also conforms in the realm of alternations. Like Cypriot Greek, CyA has
absolute neutralization of voicing in stop consonants, as illustrated in Table 3.
Table 3: Voicing neutralization in CyA stop consonants
CyA
Arabic
Gloss
sipel
ʕates
pakar
sabal
ʕadas
baqar
‘stubble’
‘lentils’
‘cattle’
It also has the same palatalization and spirantization rules (with the latter applying to the first member of consonant clusters), as well as epenthesis of transitional occlusives in clusters (Tsiapera 1969; Borg 1985; Roth 2004), as illustrated
in Table 4.
Table 4: Greek-derived phonological processes in CyA
166
CyA
Arabic
Gloss
Phonological process
kʲilp
xtuft
pkyut
kalb
katabt
buyūt
‘dog’
‘I wrote’
‘houses’
Palatalization
Spirantization
Consonant epenthesis
7 Cypriot Maronite Arabic
As with changes in the phoneme inventory, these additions to the phonological
rules of CyA imply considerable L2 pronunciation effects of Cypriot Greek, even
though it was presumably typically acquired later in life than CyA, a puzzling
apparent contradiction.
3.2 Morphology
According to Newton (1964: 43), “words of Arabic […] origin retain the full morphological apparatus of Arabic while those of Cypriot-Greek […] origin appear
exactly as they do in the mouths of monolingual speakers of the Greek dialect.”
He goes on to state that “the exceptions to the rule that the morphemes of any
one word are either exclusively [Cypriot Greek] or exclusively [Arabic] in origin would seem to be few,” and that Greek verbs “are conjugated exactly as they
are when they occur in [Greek].” Example sentences that he provides contain
multiple code-switches between Arabic and Greek-origin words, as in (1), where
Greek words are highlighted in bold.
(1)
CyA Newton 1964: 49
paxsop
na enicaso
xamse kamares
intend.impf.1sg sbjv rent.prs.1sg five room.pl
‘I intend to rent five rooms.’
Newton (1964: 50) concludes that neither source “would be in a position to
claim an undisputed majority [of words/morphemes].” Gulle (2016) also discusses
examples of “loss of systemic integration” morphologically, with respect to noun
plurals, meaning that Greek-origin nouns are used with Greek affixal morphology rather than being integrated into the CyA morphological system. The example in (2) illustrates the use of Greek-origin nouns with Greek plural morphology
intact (in bold) in a CyA matrix sentence.
(2)
CyA Borg 1985: 183, 193
allik p-petrokop-i
n-tammet
l-ispiriðk-ya ta
dem.pl def-stonecutter-pl pass-end.prf.3sg.f def-match-pl comp
kan-yišelu
fayy-es
prog.pst-light.impf.3pl dynamite.hole-pl
‘While those stonecutters were igniting sticks of dynamite, the matches
got used up.’
On the whole, the picture is of a language somewhat similar to Maltese (see
Lucas & Čéplö, this volume), in that we have two morphological systems operating in parallel, depending on the etymological origin of the root (Romance or
167
Mary Ann Walter
Arabic, in the case of Maltese; Greek or Arabic, in the case of CyA). Alternatively,
we could say that speech in CyA is replete with code-switching, and the use of
such Greek forms says nothing about the system of CyA itself.
The main exception to morphological non-interaction between CyA and Greek
is the use of the Greek diminutive suffix -ui (feminine -ua) with native CyA words,
noted by all three of the major authors on CyA (Borg, Tsiapera, and Newton). For
example, this suffix is used with Arabic nouns such as xmara ‘female donkey’ and
pint ‘girl’, yielding xmarua ‘small donkey’ and pindua ‘girl’ (Newton 1964: 43–
44). Tsiapera (1964) additionally notes the borrowing of two adjectival suffixes,
-edin (which makes nouns into adjectives) and nominal masculine singular -o.
Relatedly, Gulle (2016) observes that CyA lacks marking for directive and locative, unlike other Arabic varieties but like spoken Greek. Accusative case marking is used in spoken Greek for this purpose, but due to the lack of overt case
marking in CyA, such constructions are unmarked entirely.
(3) CyA Gulle 2016: 44
a. k-kafene
def-cafe
‘(in) the cafe’
b. fi-l-lixkali
in-def-field
‘in the field’
Occasional use of Arabic fi ‘in,’ as in other varieties and example (3b), was attributed by some CyA speakers of Gulle’s acquaintance to the influence of Levantine Arabic. For at least one speaker, the usage of locative/directional fi appeared
to be influenced by calquing from Standard Greek.
However, Borg (2004: 3) notes similar usage in Old Arabic and Hebrew, such
that Greek is not necessarily the source of this pattern. Gulle (2016: 47) concludes
that “the tense–aspect–modality (TAM) system [of CyA] is surprisingly almost
completely intact”, adding only the exception of the use of the Greek modal verb
prepi in necessitative constructions.
Finally, the occasional borrowing of the Greek plural morpheme is observed.
However, this is sporadic, and a quantitative investigation of pluralization based
on Borg’s (2004) glossary (Walter 2017) reveals that native non-suffixal plurals
are still used for over half of all pluralizable nouns, at percentages even higher
than those posited for other Arabic varieties. Greek plurals were given for only
8 of the 251 nouns.
Therefore, although the typically-Arabic use of non-concatenative plural morphology is indeed subject to some degree of suffixal regularization (17% of cases)
and somewhat more restricted in terms of the variety of plural forms in CyA, the
effect of Greek plural forms has been negligible.
168
7 Cypriot Maronite Arabic
Plural formation, perhaps the most distinctive and cross-linguistically idiosyncratic morphological characteristic of CyA, thus appears remarkably robust in
the face of contact. This echoes the retention of the pharyngeal glide in the
phonological domain.
On the whole, as Borg (1994: 57) states, “the external impact on the native
morphological patterns of [CyA] is slight.”
3.3 Syntax
According to Roth (2004: 70), “syntax is a linguistic domain particularly permeable to interference from Greek” (author’s translation). By this she means that
function words are doubled with loans from Greek, in particular with relative
clause markers and more complex constructions, as well as the use of Greek
and Arabic-origin negation markers in combination. The example in (4) demonstrates the CyA use of the native ma negation morpheme concurrently with
Greek me…me. In this case, phonetic similarity may have aided the adoption
of me.
(4)
a. CyA Borg 1985: 149
me pira me mpit
ma-pišrap
neg-drink.impf.1sg neg beer neg wine
‘I don’t drink either beer or wine.’
b. Cypriot Greek
em-pinno
me piran
me krasin
prog-drink.prs.1sg neg beer.acc neg wine.acc
‘I don’t drink either beer or wine.’
It is unclear, however, whether all or most of this is simply code-switching and
whether it should be termed syntactic rather than lexical influence.
A syntactic change which does not involve code-switching or lexical borrowing is the development of a predicative copula (lacking in the present tense in
most varieties of Arabic) from Arabic pronouns, discussed by both Roth (2004)
and Borg (1985), and illustrated in (5).
(5)
CyA Borg 1985: 134
a. l-iknise
e
maftux-a
def-church 3sg.f open-f
‘The church is open.’
b. p-pkyara enne maʕak
def-well.pl 3pl deep.pl
‘The wells are deep.’
In example (5a), the copula corresponds to the third-person feminine pronoun
‘she’ (also e, < hiya). Likewise, the copula enne in (5b) corresponds to the third169
Mary Ann Walter
person plural pronoun ‘they’ (also enne, < hunna). The development of this copula presumably replicates the obligatory present-tense copula found in Greek.
See Lucas & Čéplö (this volume) for a similar phenomenon in Maltese.
Finally, both Roth (2004) and Newton (1964) document variable placement of
adjectives, according to both Arabic and Greek norms, as illustrated in (6).
(6)
a. CyA Roth 2004: 72
m-mor-a
li-zʕar
def-child-pl def-small.pl
‘small children’
b. Lebanese Arabic Newton 1964: 48
l-ikbīr
l-bēt
def-house def-big
‘the big house’
c. CyA Newton 1964: 47
li-kbir payt
def-big house
‘the big house’
d. Cypriot GreekNewton 1964: 48
to meálo spítin
def big house
‘the big house’
However, Borg (2004) notes that so-called “peripheral” varieties of colloquial
Arabic have been said to employ freer word order than others, so the variation
in noun–adjective ordering may be an independent internal development (or alternatively, perhaps peripheral varieties are by nature more subject to contact,
which leads to this pattern of variation).
In summary, syntax, like morphology, shows relatively little influence of language contact, especially in contrast to the phonological system. As word order
is already relatively flexible in both CyA and Cypriot Greek (e.g. with respect to
subject–verb ordering; Newton 1964: 48–49), this is perhaps to be expected.
3.4 Lexicon
According to Newton (1964), of the 630 common lexical items which he elicited,
38% were Greek in origin. However, he goes on to say that the percentage is lower
in running speech, in which typically the most common (and therefore native
Arabic origin) vocabulary was used. Newton raises the possibility (1964: 51) that
170
7 Cypriot Maronite Arabic
CyA consists of “Arabic plus a large number of Cypriot [Greek] phrases thrown
in whenever [a speaker’s] Arabic fails him or the fancy takes him.” Tsiapera
(1964: 124) concurs, stating that “any speaker of [CyA] has a minimum of about
thirty per cent of Greek lexical items in his speech which are not assimilated into
the phonological and morphological system of his native language.” She identifies the semantic fields of government and politics, numerical systems including weights and measures, and adverbial particles as particularly dominated by
words of Greek origin.
This percentage contrasts with the relatively small number of Greek-origin
items appearing in Borg’s (2004) glossary. However, the difference in elicitation
contexts must be kept in mind – Newton’s work occurring in the Cypriot context and himself being competent in Greek, versus Borg’s work occurring partly
overseas and himself an Arabist rather than a scholar of Greek.
Roth (2004) refers to the drastic reduction of the lexicon, and estimates that it
includes at most 1300 items. Borg’s (2004) glossary contains roughly 2000 entries
(corresponding to 720 lexical consonantal roots), which he considers to be a “substantial portion” (though not all) of the “depleted” Arabic-origin CyA lexicon.
Gulle (2016: 45) notes suppletion in the paradigm of the verb ‘to come’, with
imperative forms borrowed from Greek. The consonantal root of the verb ‘to
come’, in CyA as elsewhere in Arabic, is √žy, as seen in the form ža ‘he came’.
However, CyA imperative forms of this verb (ela, eli, elu, in masculine singular,
feminine singular, and plural forms, respectively) are clearly based on Greek ela,
elate (singular and plural, respectively). This particular case seems to reflect a
pan-Balkan spread of this item, as ela/elate are also used in Bulgarian (personal
knowledge).
In summary, universal bilingualism and Greek dominance among CyA speakers results in widespread use of code-switched Greek vocabulary and associated
morphology, with marginal lexical suppletion. However, there is very little loan
material integrated into the CyA grammatical system.
As a final note, Hadjidemetriou’s (2009) doctoral dissertation examines language contact between CyA and Cypriot Greek (as well as Armenian and Cypriot
Greek), in the opposite direction, to identify any effects of CyA on Cypriot Greek.
Unsurprisingly, however, given the current dominance of Cypriot Greek for these
speakers, no such effects were found, in any of the above domains.
4 Conclusion
CyA appears to present a counterexample to Van Coetsem’s notion of the stability gradient, which claims that phonology (and syntax) are more stable than
171
Mary Ann Walter
other domains (the lexicon). It is clear that for CyA, phonology has been the least
stable domain. The observed phonological convergence to Greek is of the type
that suggests pervasive effects of L2 pronunciation (except for the retention of
the pharyngeal glide). Yet it is difficult to imagine any sociolinguistic scenario
in which CyA was taken up in any significant numbers by Greek speakers from
outside the community, and the typical acquisition scenario (when CyA was still
acquired by children) has been use of CyA as a home language, and Greek as
a school language, thereby generating sequential (though eventually probably
Greek-dominant) bilinguals. The historical record is unfortunately lacking any
relevant information that could shed light on the situation.
The most urgent issue for future research on CyA is undoubtedly the need
for additional documentation efforts. In particular, naturalistic texts and audio
recordings are a desideratum. It is to be hoped that the documentation and revitalization efforts currently underway will remedy this situation.
Further reading
) Tsiapera’s (1969) work is the only one so far to consider CyA in its totality as
a spoken language, although not at great length, and it drew subsequent criticism of the author’s lack of background knowledge of the Arabic language.
This monograph does, however, have the additional advantage of a publication
date very close in time to the radical change in the sociolinguistic circumstances of CyA speakers due to ethnic tensions in the island, culminating in
their near-unanimous relocation from traditionally Maronite villages to the
capital city Nicosia.
) Borg’s (1985) foundational work on morphophonology is still the most extensive resource on CyA grammar. He takes a historical perspective on changes
from earlier Arabic to contemporary CyA, both contact-driven and otherwise,
and also includes substantial textual material in CyA at the end. These texts
are currently the only published ones available.
) The follow-up volume by Borg (2004) includes a substantial introductory essay situating CyA within the range of Arabic dialects and elucidating the influences of the main contact language, Cypriot Greek. The lexical entries are
enriched by comparisons with dialectal forms from other varieties of Arabic,
as well as Greek, Aramaic, and other contact languages where relevant.
) The most up-to-date and reliable information regarding CyA and its speakers,
including documentation, preservation and revival efforts, may be found in
the Council of Europe (2017) report.
172
7 Cypriot Maronite Arabic
) The following two community websites contain information on CyA institutions and activities in both Greek and English, including contact information,
historical background, archived copies of the monthly (Greek-language) community newsletter, and so on.
• http://www.maronitesofcyprus.com (in both Greek and English)
• http://kormakitis.net/portal/ (in Greek)
Acknowledgements
My sincere thanks to Christopher Lucas and Stefano Manfredi for organizing the
workshop on Arabic and language contact at the 23rd International Conference
on Historical Linguistics, and their editing work on this volume. I would also like
to thank the audience at ICHL for their insightful comments, as well as those of
the members of the Processing and Acquisition of Language Lab at Cambridge
University. Remaining errors and infelicities are of course my own.
Abbreviations
acc
CE
CyA
comp
def
dem
du
ECRML
f
impf
pass
accusative
Common Era
Cypriot Maronite Arabic
complementizer
definite
demonstrative
dual
European Charter for
Regional or Minority
Languages
feminine
imperfect (prefix
conjugation)
passive
prf
pl
PRIO
prog
prs
pst
sbjv
sg
TAM
UN
UNESCO
perfect (suffix conjugation)
plural
Peace Research Institute
Oslo
progressive
present
past tense
subjunctive
singular
tense–aspect–modality
United Nations
United Nations Educational,
Scientific and Cultural
Organization
References
Arvaniti, Amalia. 2010. Linguistic practices in Cyprus and the emergence of
Cypriot Standard Greek. Mediterranean Language Review 17. 15–45.
173
Mary Ann Walter
Baider, Fabienne & Marilena Kariolemou. 2015. Linguistic Unheimlichkeit: The
Armenian and Arab communities of Cyprus. In Ulrike Jessner-Schmid & Claire
Kramsch (eds.), The multilingual challenge: Cross-disciplinary perspectives, 185–
209. Berlin: De Gruyter Mouton.
Borg, Alexander. 1985. Cypriot Arabic: A historical and comparative investigation
into the phonology and morphology of the Arabic vernacular spoken by the
Maronites of Kormakiti village in the Kyrenia district of north-west Cyprus.
Stuttgart: Deutsche Morgenländische Gesellschaft.
Borg, Alexander. 1994. Some evolutionary parallels and divergences in Cypriot
Arabic and Maltese. Mediterranean Language Review 8. 41–67.
Borg, Alexander. 2004. A comparative glossary of Cypriot Maronite Arabic (Arabic–
English): With an introductory essay. Leiden: Brill.
Council of Europe. 2017. Fifth periodical report presented to the Secretary General
of the Council of Europe in accordance with Article 15 of the European Charter
for Regional or Minority Languages. Strasbourg: Council of Europe.
Gulle, Ozan. 2016. Kormakiti Arabic: A study of language decay and language
death. In Vera Ferreira & Peter Bouda (eds.), Language documentation and conservation in Europe, 38–50. Honolulu: University of Hawai’i Press.
Hadjidemetriou, Chryso. 2009. The consequences of language contact: Armenian
and Maronite Arabic in contact with Cypriot Greek. Colchester: University of
Essex. (Doctoral dissertation).
Newton, Brian. 1964. An Arabic–Greek dialect. Word 20(sup2). 43–52.
PRIO. 2009. The protection and revival of Cypriot Maronite Arabic. https://www.
prio.org/Publications/Publication/?x=7371.
Roth, Arlette. 2004. Le parler arabe maronite de Chypre: Observations à propos
d’un contact linguistique pluriséculaire. International Journal of the Sociology
of Language 168. 55–76.
Tsiapera, Maria. 1964. Greek borrowings in the Arabic dialect of Cyprus. Journal
of the American Oriental Society 84(2). 124–126.
Tsiapera, Maria. 1969. A descriptive analysis of Cypriot Maronite Arabic. The
Hague: Mouton.
Van Coetsem, Frans. 1988. Loan phonology and the two transfer types in language
contact. Dordrecht: Foris.
Van Coetsem, Frans. 2000. A general and unified theory of the transmission process
in language contact. Heidelberg: Winter.
Walter, Mary Ann. 2017. A quantitative investigation of noun pluralization in
Cypriot Maronite and Maltese Arabic. Paper presented at the 23rd International Conference on Historical Linguistics, San Antonio, Texas, July 31 –
August 4.
174
Chapter 8
Nigerian Arabic
Jonathan Owens
University of Bayreuth
Nigerian Arabic displays an interesting interplay of maintenance of inherited structures along with striking contact-induced innovations in a number of domains.
This chapter summarizes the various domains where contact-based change has
occurred, concentrating on those less studied not only in Arabic linguistics, but
in linguistics in general, namely idiomatic structure and an expanded functionalization of demonstratives. Methodologically, comparative corpora are employed to
demonstrate the degree of contact-based influence.
1 Historical and linguistic background
Nigerian Arabic (NA) is spoken by perhaps – there are no reliable demographic
figures from the last 50 years – 500,000 speakers. These are found mainly in
northeast Nigeria in the state of Borno where their homeland is concentrated
along the Cameroon–Chad border as far south as Banki, spreading westwards towards Gubio, and south of Maiduguri towards Damboa. Mirroring a larger trend
in Nigerian demographics, the past 40 years have seen a considerable degree of
rural–urban migration. This has seen, above all, the development of large Arab
communities in cities in Borno – the capital Maiduguri has at least 50,000 alone1
– though they are now found throughout cities in Nigeria.
Arabs in Nigeria are traditionally cattle nomads, part of what the anthropologist Ulrich Braukämper (1994) has called the “Baggara belt”, named after the Arab
1
A report in the 1970s by an urban planning company, the Max Lock Group (1976), estimated
that 10% of the then estimated population of 200,000 Maidugurians were Arabs. Today the
population of Maiduguri is not less than one million and may be considerably larger, which
proportionally would estimate an Arab population in Maiduguri alone of at least 100,000. Of
course, if one included the refugee camps today, the number would be much higher.
Jonathan Owens. 2020. Nigerian Arabic. In Christopher Lucas & Stefano Manfredi
(eds.), Arabic and contact-induced change, 175–196. Berlin: Language Science Press.
DOI:10.5281/zenodo.3744515
Jonathan Owens
tribe in the western Sudan (Kordofan, Darfur; see Manfredi 2010) whose culture
and dialect are very similar to those of the Nigerian Arabs. Until the very recent
Bokko Haram tragedy, besides nomadism, Arabs practiced subsistence farming.
As of the writing of this chapter, nearly all rural Nigerian Arabs have been forced
to flee their home villages and cattle camps, and are living mainly in refugee
camps in northeast Nigeria and neighboring countries.
Arabs first came to the Lake Chad area – whether territorial Nigeria is at this
point undetermined – in the late fourteenth century. They were part of what
initially was a slow migration out of Upper Egypt towards the northern Sudan
beginning in the early thirteenth century, which gained momentum after the
fall of the northern Nubian kingdom of Nobadia (or Maris) in the fourteenth
century. All in all, NA exhibits a series of significant isoglosses which link it
to Upper Egypt, via Sudanese Arabic, even if it displays interesting “archaisms”
linking it to regions far removed from Africa (Owens 2013). Its immediate congeners are found in what I have termed Western Sudanic Arabic (WSA; Owens
1994a,b), stretching between northeast Nigeria in the west and Kordofan in the
east (Manfredi 2010). When properties of NA are contrasted with other varieties
of Arabic, it is implicitly understood that these do not necessarily include other
WSA varieties. Much more empirical work is necessary in this regard, but, to
give one example, many of the extended functions of the NA demonstrative described in §3.3.2 below are also found in Kordofanian Arabic (Manfredi 2014).
Moreover, where thoughout the Sudanic region as a whole any given isogloss
lies is also an open question, as is the issue of the degree to which the contactinduced changes suggested here represent broad areal phenomena. As my own in
many cases detailed data derives from NA, I limit most observations to this area.
NA itself divides into two dialect areas, a western and an eastern one that I have
also termed Bagirmi Arabic, since it is spoken by Arabs in the Bagirmi-speaking
region.
In Borno, Arabs are probably the largest minority ethnic group, though still
a minority. The entire area bordering Lake Chad, both to the east and to the
west, is dominated by Kanuri-speaking peoples (Kanembu in Chad). This was a
domination which the Arabs already met in their first migrations into the region,
both a political and a linguistic domination. As will be seen, this has left dramatic
influences in some domains of NA, while leaving others untouched.
While until about 1970 Kanuri was the dominant co-territorial language, Arabs
in the Lake Chad area have been in close contact with other languages and
ethnic groups as well, for instance Fulfulde, Kotoko (just south of Lake Chad) and
Bagirmi (south of Ndjammena in Chad). Furthermore, Kanuri established itself
in Borno in an area already populated by speakers of Chadic languages, so it as
176
8 Nigerian Arabic
well was probably influenced by some of the co-territorial languages Arabs met.
Since 1970, Hausa has become the dominant lingua franca in all urban areas in
northeastern Nigeria (indeed throughout the north of the country). In a sample of
58 Maiduguri speakers for instance (Owens 1998; Owens 2000: 324), 50 professed
knowing Hausa, and 46 Kanuri. In the only study of its type, Broß (2007) shows
that urban Maiduguri Nigerian Arabs have a high degree of accuracy for a number of complex variables in Hausa, while, using a similar sample, in one of the
few interactional studies available, Owens (2002) also documents a high multilingual proficiency between Arabic and Hausa, and for some speakers, English.
How such micro-studies can be interpreted against the over 400 years of NA
contact with area languages remains a question for the future. Rural areas have
not yet experienced such a high penetration of Hausa. In a second, rural sample
consisting of 48 individuals, only sixteen self-reported knowing Hausa versus
forty Kanuri. Note that as of the 1990s, there were still a considerable number
of monolingual Arabic speakers, particularly in the area along the Cameroonian
border which among Nigerian Arabs is known as the Kala–Balge region.
While Standard Arabic (Classical Arabic) has always been a variety known
among a small educated elite in Borno (of all ethnic backgrounds), along with
Hausa it has gained considerable momentum in recent years. Whereas traditionally Classical Arabic, as a part of Koranic memorization, has always been a part
of Arabs’ linguistic repertoire, it is only since about 1990 that the teaching of
Standard Arabic as a school subject has spread oral fluency in this variety.
To this point, conditions have been described which, on paper at least, would
favor influence via borrowing under RL-agentivity (in the terminology of Van
Coetsem 1988; 2000). Nigerian Arabs as a linguistic minority tend to be bilingual,
and, it may be assumed, have had a history of bilingualism in Kanuri and locally
other languages going back to their first migrations into the region. Equally, however, Nigerian Arabic society has itself integrated other ethnic groups creating
conditions of shift to Arabic. According to Braukämper’s (1994) thesis, the very
basis of Nigerian Arab nomadism is cattle nomadism based on a Fulani model.
This is said to have arisen around the mid-seventeenth century as Arabs coming
from the east met Fulani moving west. Today there is very little Fulfulde spoken in Borno or Chad, so it may be surmised that the result of the Fulani–Arab
contact was language shift in favor of Arabic. Furthermore, slavery was a wellestablished institution which incorporated speakers from other ethnic groups
(see recording TV57b-Mule-Hawa in Owens & Hassan 2011, as an instance of
a slave descendant). Intermarriage is another mechanism by which L1 speakers
would switch to Arabic. In contemporary Nigeria, intermarriage in fact tends to
favor Arab women marrying outside their group, rather than marriage into Arab
177
Jonathan Owens
society, though there is no cultural proscription of the practice, and such practices tend, inter alia, to be influenced by the relative prestige and power of the
groups involved. Today Arabs are dominated politically by the Kanuri, though
there are eras, for example the period of Kanemi in the mid-nineteenth century,
or the rule of Rabeh at the beginning of the twentieth, when Arabs were more
dominant and perhaps had greater access to marriage from outside groups. I will
return to these summaries in §4.
The data for this chapter comes from long years of working on Arabic in the
Lake Chad region. More concretely, a large oral corpus of about 400,000 words
(Owens & Hassan 2011) forms the basis of much of the research, and this corpus
will be referred to in a number of places in the chapter. When a form is said to be
rare, frequent, etc., these evaluations are made relative to what can be found in
the corpus. All examples come from this corpus. The source of the recording in
the data bank is indicated by the number in brackets at the end of the example.
2 Contact and historical linguistics
Language contact is an integral part of historical linguistics. In the case of Arabic, the history of Arabic has different interpretations, so it is relevant here to
very briefly reiterate my own views (Owens 2006). All varieties of contemporary
Arabic derive from a reconstructed ancestor or ancestors. Whether singular or
plural is a crucial matter, but one answered legitimately only within historical
linguistic methodology (see e.g. Retsö 2013, who appears to favour the plural).
As is usually accepted (perhaps not by some working within grammaticalization
theory, e.g. Heine & Kuteva 2011), historical linguistics operates at the juncture
of inheritance and contact, and examines change due to internal developments
and change due to contact. In the case of Arabic, contact extends well into the
pre-Islamic era (Owens 2013; 2016a; forthcoming).
Furthermore, it operates at the level of the speech community, and Arabic
has and had many speech communities, each with its own linguistic history. The
history of speech communities is not co-terminous with political history, usually
not with the history of individual countries, or even with cultural entities such as
a nomadic lifestyle. It follows that Arabic linguistic history is quite complicated,
its large population being the product of and reflecting many individual social
entities.
Any individual contemporary Arabic speech community therefore lies at the
end of many influences. Interpreting whether and when a particular change
occurred due to contact is anything but straightforward, as I will discuss very
briefly in the following phonological issue.
178
8 Nigerian Arabic
Ostensibly NA shows the loss of *θ:
(1)
*θ > t, *θawr > tōr ‘bull’
or in the eastern area:
(2)
*θ > s, *θawr > sōr ‘bull’
There is no space to go into the detailed historical linguistic arguments here,
but it would be incorrect to assert that these changes, quite plausibly originally
due to contact, took place in the territorial NA or WSA region. This can be seen
inter alia in the fact that all of Egyptian Arabic (EA) and all of the Sudanic region
including the WSA area has (1). Whenever the shift occurred, it was well before
Arabs came to the Sudanic region, let alone Nigeria. The changes in (1) and, I
would argue, (2) as well, are part of the historical linguistics of ancestral Sudanic
Arabic, but the changes themselves are antecedent to Arabic in the Sudanic region and therefore are not treated here.
3 Contact-induced changes
3.1 Phonology
Excluding cases like (1–2) on methodological grounds, other than marginal effects due to borrowing, discussed briefly in §3.2, there are no significant instances
of contact-induced phonological change limited only to NA. Two changes confined to all or part of the WSA region can be suspected, however.
Throughout Nigeria, Cameroon, and most of Chadian Arabic, *ḥ/ʕ have depharyngealized.
(3)
*ḥ/ʕ > h/ʔ
ḥilim ‘dream’ > hilim
gaʕad ‘stay, sit’ > gaʔad
As a set, the change is attested only in this region. Moreover, the area it is
attested in begins by and large in the region where Arabic fades into minority
status.
A second candidate for a local WSA innovation is the reflex of *ṭ, which is a
voiced, emphatic implosive /ɗ/. The implosive /ɗ/ is also found in Fulfulde, as
well as other possible contact languages such as Bagirmi, which, as noted above,
are one source of shifters to Arabic. Manfredi (2010: 44; and personal communication) notes that /ɗ/ is an allophonic variant in Kordofanian Baggara Arabic.
179
Jonathan Owens
The status of one phoneme, /č/, is still open. It is fairly frequent (about 100
entries out of about 8,500 (excluding proper names) in a dictionary currently in
preparation begin with /č/). In a minority of cases an Arabic origin is certain
or likely, e.g. čāl ‘come’ (eastern variant) < *tāl and perhaps čatt ‘all’, < *šattā
‘various’, with [š + t] > /č/ recalling some Gulf dialects ičūf ‘you see’. /č/ is never
a reflex of *k. However, most instances of /č/ are still unaccounted for (e.g ču
‘very red’, čāqab ‘wade through’).
All in all then, there has not been a great deal of fundamental phonological
change due to contact. Note that NA maintains all inherited emphatics, and probably inherited its phonemically contrastive emphatic /ṃ/, /ṛ/ and perhaps its /ḷ/
as well.
3.2 Loanwords
Despite its long period as a minority language in the Lake Chad region, NA has
only a modest number of loanwords (see Owens 2000 for a much more detailed
treatment of all aspects of loanwords in the classical sense). In a token count
based on about 500,000 words, only about 3% of all words were loans. On a
type basis the percentage rises considerably, though still is far from overwhelming. Table 1 presents loanword provenance data from the dictionary currently in
progress.
Table 1: Loanwords in NA, types, 𝑁 = 1263
Language
English
Hausa
Kanuri
Standard Arabic
French
Fulfulde
Kotoko
Types
509
255
252
212
21
12
2
The figures in Table 1 are probably a slight underestimation, as there are about
sixty words, like bazingir ‘soldier of Rabeh’ which clearly are not of Arabic origin
but whose precise origin has not been found.
180
8 Nigerian Arabic
There are many interesting issues in understanding the loanwords, a few of
which I mention very cursorily here. The semantic domains differ from source
to source. Standard Arabic, for instance, has mainly learned words. Kanuri covers a fairly wide spectrum, and strikingly includes a large number of discourse
markers and conjunctions, on a token basis. dugó ‘then, so’ (< dugó) for instance
has something in the range of 630 occurrences and yo, yō, iyō ‘so, okay, aha’ has
938. In Owens (2000: 303), discourse particles and conjunctions are shown to
make up no less than 23.3% of all loanword tokens in the sample. It is noticeable
that although a few Hausa discourse marker tokens (to ‘right, okay, so’) do occur,
there are hardly twenty in all, this being indicative of the much shorter time span
Hausa has been in large-scale contact with NA as compared to Kanuri.
The question of origin has two aspects, one the ultimate origin, the other how
it got into NA. bel ‘belt’ is ultimately of English origin, but the same word is also
found in Hausa (bel) and in Kanuri (bêl). Given that both of these languages are
dominant ones, it is likely that bel entered NA from one of these, not directly from
English. The statistics above are the ultimate origin. The medial origin (travel
words) is much harder to trace. Using the corpus, it is possible to discern likely
paths. For instance, NA sanāʔa ~ saɲa ‘trade, occupation, profession’ is cognate
with both Standard Arabic ṣināʕa ‘art, occupation, craft’ and Hausa sanāʔā ‘trade,
craft, profession’. Considering the distribution of saɲa among speakers who have
no knowledge of Standard Arabic, it is likely that the word reached NA via Hausa.
Non-Arabic phonology will often be maintained in the loanword. However, as
can be discerned from loanwords of higher frequency, usually there is variation
between retention of the source phoneme and adaptation. For instance ‘police’
comes in two forms, polīs and folīs (Owens 2000: 278). The [p] variant occurs
in 19 tokens distributed among eight speakers, the [f] in 18 tokens among six
speakers. Inspection of the statistics shows only a tendential bias towards [f]
among women and villagers. Both variants appear therefore to be widespread.
Note in this case that variation between [p] and [f] is also endemic to Kanuri, so
it is likely here that the variation itself was borrowed.
3.3 Syntax
There are three strong candidates for contact-induced change in the syntactic
domain: word order, ideophones and an expansion and realignment of demonstratives.
181
Jonathan Owens
3.3.1 Word order and ideophones
NA has only two pre-noun modifiers, gōlit ‘each’, kunni ‘each’.
(4)
gōlit ʔīd
nulummu
each holiday gather.impf.1pl
‘We would gather at each festival.’
Otherwise NA is head-n-initial, which means that čatt ‘all’ and kam ‘how
many’ are post-n, while demonstratives only have a post-n position (as in EA).
(5) numšu
be ʔaḫuwāt-na čatt-ina
go.impf.1pl with sisters-1pl all-1pl
‘We go with our sisters, all of us.’
(6) taǧīb
ḍahaḅ kam
bring.impf.2sg.m gold how.much
‘How much gold do you bring?’
The post-nominal-only demonstrative would have been inherited from EA.
čatt ‘all’ mirrors the post-nominal alternative for kull, both taking a pronoun
cross-referencing the head noun. Therefore, strictly speaking, the only innovation is the post-nominal position of kam ‘how many’, and an argument could be
made that internal analogies lead NA towards a more consistent head-first nounphrase order. By the same token, Kanuri is also consistently head-first order in
the np, so it could be that contact with Kanuri accelerated an inherited trend.
The numeral phrase has undergone considerable re-structuring. From ‘twenty’
upwards, the order is decade–ones.
(7)
talātīn haw wāhid
thirty and one
‘thirty one’
Though inherited teens do occur, the usual structure is ten–ones.
(8)
ʔasara haw wāhid
ten
and one
‘eleven’
This order mirrors that of Kanuri (Hutchison 1981: 203), and indeed that of
most languages in the immediate Lake Chad area. Uzbekistan Arabic has the
182
8 Nigerian Arabic
same numeral order and structure as NA, and in these cases independent contact
events are likely the reason for the shift from an inherited structure.
A new syntactic category (for Arabic), that of ideophones, is described in detail
in Owens & Hassan (2004) (see tul in (11b) below). To date in the dictionary of
NA in progress there are 342 ideophones, about 4% of the lemma total.
3.3.2 Demonstratives
Formally, NA demonstratives reproduce their inherited forms, and therefore are
virtually identical to paradigms found in various Egyptian dialects, except that,
in consonance with NA morphology, feminine plural has a distinct form, which
most Egyptian dialects have neutralized (see Table 2).
Table 2: NA demonstratives
Near
Masculine
Feminine
Far
Singular
Plural
Singular
Plural
da
di
dōl
dēl
ɗāk
ɗīk
ɗōlak
ɗēlak
As with all Arabic demonstratives, NA demonstratives are used both as modifiers and pronominally. The traditional, inherited functions are entity referential
(al-bēt da ‘this house’), and propositional anaphoric (ʔašān da ‘because of this’,
where ‘this’ references an introduced proposition).
Additionally, however, the demonstratives occur in several contexts which
either are not attested at all, or are attested only on an extremely infrequent
basis in other Arabic dialects. I summarize these here.
1. Marking the end of dependent clauses, whether relative, conditional or adverbial.
Usually da is the default form in this function, though in the case of relative
clauses the demonstratives often agrees with the head noun.
(9) Conditional clause
[kan gul
balkallam
kalām-hum
da]
ma
[if say.prf.1sg speak.impf.1sg language-3pl.m dem.sg.m] neg
bukūn
possible
‘If I said I speak their language, it is not possible.’
183
Jonathan Owens
(10) Relative clause
balkallam
le-əm be l-luqqa
speak.impf.1sg to-3pl.m with def-language
l-biyarifū-ha
di
rel-know.impf.3pl.m-3sg.f dem.sg.f
‘I speak to them in the language which they know.’
2. Text referential, cataphoric.
da is used cataphorically to foreshadow a propositional expansion. In (11) the
speaker is asked how he farms. Instead of answering directly, he introduces
his answer with the cataphoric use of da, which is then expanded upon in the
following independent proposition.
(11)
a. kēf tihērit
how farm.impf.2sg.m
‘How do you farm?’
b. baharit
da,
al-hirāta
l-wād-e tul di
farm.impf.1sg dem.sg.m def-farming def-one-f only dem.sg.f
d-duḫun
def-millet
‘How I farm? The one type of farming is only millet.’
3. Deictics.
A number of deictic words, mainly adverbs, are marked by demonstratives,
in this case nearly always da. The deictics include hassa ‘now’, dugut ‘now’,
wakit ~ waqit ‘now’, tawwa ‘previously, formerly’, hine/hinēn ‘here’, awwal
‘first, before’, gabul ‘previously, before, baʔad ‘afterwards’, alōm/alyōm ‘today’, bukura ‘day after tomorrow’, amis ‘yesterday’, albāre ‘yesterday evening’,
ambākir ‘tomorrow’, mǝṇṇaṣabá ‘in the morning’, qādi ‘there’, hināk ‘there’,
haǧira ‘(a place) away from here’, bilhēn ‘much’.
(12)
haǧira da
ma mašēt
away dem.sg.m neg go.prf.1sg
‘I didn’t go away anywhere.’
(13)
albāre
da
as-sarārīk daḫalo
yesterday dem.sg.m def-thieves enter.prf.3pl.m
‘Yesterday evening thieves broke in.’
184
8 Nigerian Arabic
4. Demonstratives mark pronouns, in this case often agreeing with the pronoun
in terms of number and gender, and other demonstratives, where usually da
occurs.
(14)
a. inti di
ǧībi
le-i
š-suqúl da
2sg.f dem.sg.f bring.imp.sg.f to-obl.1sg def-thing dem.sg.m
‘You there bring me the watchamacallit.’
b. ʔard gaydam dōla
da
kula ʔarab
land Geidam dem.pl.m dem.sg.m also Arab
‘In the land around Geidam and the like are also Arabs.’
Basic attributes of these expanded functions can be given in cursory manner.
Concerning frequency, the occurrence of demonstratives in these functions
on a token basis is high. For instance, there are 887 tokens of qādi ‘there’ in the
corpus, of which 108 or 12% are marked by da. The highest percentages of demonstratives in these functions occur with dependent clauses and the 3sg pronouns
hu ‘he’ and hi ‘she’. For hu, nearly 25% of all tokens occur with da (586/2407
24.3%). As far as the four innovative functions summarized above are concerned,
a sample of 1318 tokens of da gathered from an arbitrary selection of 45 texts in
the corpus reveals the data presented in Table 3. While the inherited referential
functions constitute the largest single class, they make up only 53% of the total.
The remaining 47% are functionally innovative.
Table 3: Functions of da in NA
Function
Percentage of total
Inherited functions
Entity referential
Proposition-anaphoric
53.4%
42.3%
11.1%
Innovative functions
Cataphoric-propositional
Dependent clause
Adverbs/deictic
Pronouns, demonstratives
46.7%
7.2%
18.7%
12%
8.8%
The syntactic, pragmatic and semantic nuances of using or not using the demonstratives in these innovative contexts have yet to be worked out. The two
185
Jonathan Owens
examples in (15) and (16) illustrate different ways the innovative functions are
integrated with other elements of the grammar.
Syntactically, for instance, based on the sample described above, da marks the
end of about 30% of all conditional clauses. When it does not occur, its final clause
boundary marking position commutes with an alternative pragmatically-marked
element, such as the discourse marker kula ‘even’. (No tokens of *kula da closing
a conditional clause occur in the corpus).
(15)
kan qayyart-a
kula
if change.prf.2sg.m-3sg.m dm
‘Even if you changed it.’
Pragmatically there are many instances where da has a focusing function, as in
the following, where a mixed linguistic region ‘here’ is contrasted with another
‘there’, which is linguistically homogeneous.
(16) nās
gadé
gadé
kula hinēn katīrīn fi [qādi da]
nafar-na
people different different dm here many exs there dem.sg.m type-1pl
nafara wāhid
type one
‘Here there are a lot of different (types) but [over there] there is just our
one ethnic group.’
The functions outlined in Table 3 are therefore both of high frequency and are
systematically embedded in the syntax and pragmatics.
It should be intuitively clear that the functions in examples (9–16) are innovative in their systematicity relative to other varieties of Arabic. To show this in
detail it would, however, be necessary to look at large-scale corpora of other Arabic dialects. This can very briefly be done with EA, which, as noted above, is an
ancestral homeland of NA. The EA corpus is from LDC Callhome (Canavan et al.
1997), Nakano (1982), Behnstedt & Woidich (1987), and Woidich & Drop (2007),
comprising about 417,000 words. It is thus of comparable size to the NA corpus. In
this corpus there do occasionally occur collocations of pronoun + demonstrative
in the same contexts as illustrated in (14), in particular as in (17).
(17) hiwwa da
lli mawgūd ʕandi-na
3sg.m dem.sg.m rel present at-1pl
‘That is what we have.’
186
8 Nigerian Arabic
It clearly, however, has a different functionality from NA pronoun + demonstrative. In EA the construction consistently is anaphoric to a previous proposition or situation, as in (17), where it introduces a previously-established topic to
a following descriptive qualification. In 11 of the 58 tokens in the EA corpus it
is followed by a relative clause, as in (17). Most tellingly, there are 2,677 huwwa
(~ hu, hū, hiwwa, hūwa) tokens, of which only 58, or 2% are followed by da (~ dah,
dih, deh, dī ). This compares to the nearly 25% hu + da tokens in NA noted above.
Moreover, in the NA sample, no tokens of hu da are followed by a relative clause.
In this same statistical vein, the total number of singular proximal demonstratives in NA amounts to 16,774 tokens (14,591 da, 2,183 di). In the EA corpus there
are only 8,239 (4,996 da, 3,243 di). Given that the corpora sizes are comparable
(EA in fact a little larger), the demonstratives in NA are vastly over-proportional.
This preponderance is due to da. Clearly there is a case to be answered: what
accounts for the vastly higher frequency of the 3sg.m demonstrative in NA? Recall, in answering this question, that behind the simple statistical comparison is
a fundamental historical one as well. Ancestral NA came from ancestral EA. The
initial populations, it needs to be assumed, had a demonstrative system like that
of EA, and the majority of NA demonstrative tokens (see Table 3) still reflect
this system. A blunt historical linguistic question is what caused the vast shift in
frequencies.
From these initial, basic observations, it does not appear that the greatly expanded functionality of the demonstrative in NA can be explained by an increasing grammaticalization of the demonstrative.2 This follows from two observations. First, the expanded functions of the demonstrative in Table 3 are, with
the exception of the boundary-marking of dependent clauses (10), not those associated with the grammaticalization of demonstratives (e.g. the 17 trajectories of
demonstratives in Diessel 1999). Secondly, NA and EA split over 400 years ago.
One of the branches, represented by NA, underwent the considerable changes
outlined here, whereas the other branch, EA, probably did not change at all (i.e.
sentences such as (17) were probably present in EA in 1200, and before).3 There is
thus no natural or inherent tendency for demonstratives to expand as in Table 3.
It can thus be safely assumed that the expanded functionality of the NA demonstrative was due to contact.
2
I do not at all agree with Heine & Kuteva (2011) and Leddy-Cecere (this volume) that changes
due to contact can be assimilated to a type of grammaticalization process, so the following
contact-based account is independent of grammaticalization. Grammaticalization, in Meillet’s
original sense, pertained only to internally-motivated changes.
3
Cf. Damascus, which has an identical construction to that of EA. There are parallels also in
Classical Arabic, so this type of construction is probably proto-Arabic. If so, it only heightens
the degree to which NA has innovated away from an original, stable structure.
187
Jonathan Owens
In fact, there is a good deal of prima facie evidence supporting this supposition.
However, as is so frequently the case when one suspects pattern (metatypical)type contact influence which is probably centuries old, support for the position
will be indexical. Moreover, in the current case one is most probably dealing
with a large-scale areal phenomenon in the Lake Chad area (and perhaps beyond)
which encompasses well over a hundred languages. In this summary chapter it
will therefore have to suffice to rather peremptorily indicate that throughout the
region there is a referential marker, sometimes a demonstrative, sometimes an
article-like element, sometimes an element with both demonstrative and articlelike properties, which consistently has the distribution of (9–16). Some languages
have a better fit than others, and, of course, they will differ in detail in their
language-internal functionality. A basic pattern is illustrated in (18) with Kanuri
(Hutchison 1981: 47, 207, 218, 234, 241, 270), and summary references are made
for Bagirmi, Wandala and Fali. So far as is known, Fali and Wandala had no
significant contact with NA or its WSA relatives.
The Kanuri determinative -də has the following functions.
(18) Kanuri (Nilo-Saharan/Saharan)
Anaphoric entity reference
a. obligatorily ends RC and optionally many adverbial clauses; = (9), (10)
b. pronoun focus; = (14)
c. marks adverbs; = (12), (13)
The only Kanuri structure missing from the list appears to be the propositional
cataphoricity illustrated in (11).
Wandala (Frajzyngier 2012: 507–34, 603) has two morphemes: -na which is
broadly glossed as a determiner and -w ‘that’. -na, besides marking entity reference, obligatorily marks the ends of a relative clause, and optionally a conditional
(=9, 10); it occurs as an obligatory element in certain time/place adverbs (=12, 13);
it is part of the previous mention marker ŋán-na; ŋán itself is said to originally
be a third person singular pronoun, so there is a structural parallel to hu + da. -w
functions as a topic marker that marks pronouns (=14).
In Fali (Adamawa; Niger-Congo) the demonstratives gi/go also obligatorily
mark the end of relative and conditional clauses (=9, 10), subject focus (=14), and
occur with some adverbs (=12, 13).
In Bagirmi a “determiner particle” -na is a constitutive part of the demonstrative enna < et-na ‘this’, and -na alone obligatorily marks the end of relative
clauses, and can emphasize pronouns, adverbs and entire sentences (Stevenson
1969: 40, 51, 54).
188
8 Nigerian Arabic
Areal features typically are not sensitive to language family, and this appears
to be the case in this brief exemplification. Kanuri and Bagirmi are Nilo-Saharan,
Wandala is Chadic, Fali is Niger-Congo, and Arabic is Semitic. Only Wandala and
Arabic are very distantly related genetically. Nonetheless, in all of the languages
there is a deictic–referential marker (demonstrative, determiner, demonstrative–
determiner) which, besides a classic deictic or anaphoric function, surfaces in
an extended range of identical (cf. marking boundary of dependent clause) or
similar (pronouns, adverbs)4 functions. These extended functions are precisely
those which distinguish NA from other varieties of Arabic. The case for contact
follows from two directions: in certain (not all) respects, NA deviates markedly
from a putative ancestral source shared with EA, and where it does, its deviation
corresponds broadly to analogous categories in co-territorial languages.
3.4 Semantics
The innovative distribution of the NA demonstratives is striking for the degree
to which it appears to have raised the overall demonstrative token count, relative
to EA. Discerning its presence in a text, however, is a straightforward matter. A
much subtler, but no less pervasive instance of contact-based change pertains
to idiomaticity. Like the demonstrative, this has a semantic and a formal aspect.
Semantically, meanings emerge which are, for Arabic, unique, as in the following.
(19)
a. rās al-bēt
head def-house
‘roof’
b. nādim rās-a
person head-3sg.m
‘an independent person, person of his own means’
(20)
a. tallafo
gaḷb-i
spoil.prf.3pl.m heart-1sg
‘They angered me.’
b. gaḷb-a
helu
heart-3sg.m sweet
‘He is happy.’
4
The comparativist is limited to the extant reference grammars. These are in many instances
excellent. Still, I suspect that they understate the flexibility of distribution of elements such as
the deictic marker discussed here. Mea culpa, in Owens (1993: 88, 221, 235) the extended functions of the demonstrative described in this chapter for NA were treated in disparate sections,
with no overall focus.
189
Jonathan Owens
Formally the idioms are distinctive (as Arabic collocations) in bringing together lexemes which in other dialects would hardly co-occur, like [tallaf + gaḷb]
or [gaḷb + helu]. The idiomatic meanings of the keywords (e.g. tallaf, gaḷb) are,
in usage terms, often the typical usage for a given lexeme. In the NA corpus, for
instance, of 101 tokens of gaḷb ‘heart’ all of them, 100%, are idiomatic. There is
no reference to a physical heart. Similarly, rās is 80% idiomatic (247/308 tokens;
Ritt-Benmimoun et al. 2017: 53). Thus, while idiomaticity has been consistently
ignored as a theoretical issue in historical linguistics in general and in Arabic in
particular, on a usage basis it is an integral aspect of understanding the lexical
texture of the language.
Here as well NA is strikingly different from EA, as again can be determined
from corpora-based comparison. In general, though both NA and EA share idiomatic keywords (gaḷb/ʔalb and rās are frequent in both, for instance), their
meanings and their collocational environments hardly overlap. For instance, in
the EA corpus there are 110 tokens of gaḷb/ʔalb ‘heart’, of which 102 or 93% are
idiomatic. This percentage closely parallels that of NA idiomatic gaḷb. The typical EA collocate of idiomatic ʔalb, however, is very different. The most frequent
meaning is ‘center of X’, ʔalb il-baḥr ‘middle of the sea’. This meaning is entirely
lacking in NA, and consequently collocates like !gaḷb al-bahar (! = collocationally/semantically odd) are also lacking.
How different NA idiomaticity (meaning and collocational environment) is
from EA was shown recently in Ritt-Benmimoun et al. (2017). There a three-way
comparison was conducted between EA, southern Tunisian Arabic and NA, looking at three idiomatic keywords frequent in all three dialects: ṛās, gaḷb ‘heart’,
and ʕēn ‘eye’. EA and southern Tunisian, though separated by a longer period of
time (ca. 1035–present) than EA–NA (ca. 1300–present), showed a much higher
identity of idiomatic structure than EA–NA (or NA–southern Tunisian). Both EA
(21a) and Tunisian Arabic (21b), for instance, maintain the same lexemes, same
structure, same idiomaticity in a highly specific meaning.
(21)
190
a. Egyptian Arabic
ṛās-u
fi t-turāb
ḥaṭṭ
put.prf.3sg.m head-3sg.m in def-ground
‘He humiliated him.’
b. Tunisian Arabic
ḥaṭṭ-l-a
ṛās-a
fi t-tṛāb
put.prf.3sg.m-to-3sg.m head-3sg.m in def-ground
‘He humiliated him.’
8 Nigerian Arabic
These are nonsensical, or literal collocations in NA.
The comparison between EA and southern Tunisian Arabic serves as a similar
baseline to comparing the overall demonstrative frequencies between EA and
NA. The same question occurs. Why is NA different?
In this case the answer is even clearer than with the demonstrative. Essentially,
NA has calqued its idiomatic structure (meaning and collocation) from Kanuri.
The Kanuri of (19a) and (20b), for instance, are as in (22).
(22)
a. kəla fato-be
head house-gen
‘roof’
b. kam kəla-nzə-ye
person head-3sg.m-gen
‘an independent person, person of his own means’
A ‘roof’ in both languages is the ‘head of a house’, an independent person is
a ‘person of his head’, and so on, for something in the range of 70–80% of all the
approximately 340 idioms studied (see Owens 1996; 2014; 2015; 2016b for details).
In summary, a large part of NA lexical structure is, as it were, not Arabic, but
rather, as termed in Owens (1998), part of the Lake Chad idiomatic area. This
identity, however, exists only at a semantic and collocational level. In their basic
meaning, and their phonology, morphology and syntax, even in the context of
idioms (Owens & Dodsworth 2017), the constituent lexemes rās, bēt, tallaf, gaḷb
etc. in NA are indistinguishable from any variety of Arabic at all.
There doubtless remains a good deal more systematic, contact-based correspondence between NA and languages of the Lake Chad area to be explored.
The influence on NA is significant.
4 Conclusion
According to the historico-demographic background to NA, this variety did and
does live with co-territorial languages, particularly Kanuri, today increasingly
with Hausa, and in the past, Fulfulde and other smaller languages. NA bilingualism should, presumably, manifest itself in borrowing. Equally, NA speech
communities have incorporated speakers of other languages into its fabric. The
expectation here is that NA would be influenced via shift (imposition) from other
languages.
In the domains summarized here, it is hard to discern a clear correlation between linguistic outcome and type of contact. There has been some phonological
191
Jonathan Owens
change, which in Van Coetsem’s (1988; 2000) model is suggestive of change via
shift (imposition), but the influence is limited to the features discussed in §3.1.
What I believe is more striking than the contact-induced phonological change
is the maintenance of inherited structures. NA still maintains a robust series
of emphatics, has a non-reductive syllable structure reminiscent of, inter alia,
Tihāma varieties, has classic distinguishing syllable structure attributes such as
the gahawa syndrome (ahamar ‘red’) and the bukura syndrome (bi-ǧiri ‘he runs’),
to mention but a few. If the changes in (9–16) are due to imposition, it is equally
clear that the “imposers” otherwise learned/learn a very normal Arabic.
Classic borrowing is moderate. The fact that discourse markers and conjunctions are token-wise frequent suggests that speakers were/are conversant in both
Kanuri and Arabic. This does not, however, indicate whether these loans arose
through imposers or borrowers. Moreover, to complicate matters even more, assuming Kanuri to have been the widespread lingua franca in the past, it would
not need to have been native Kanuri speakers who imposed the Kanuri into
Arabic. Speakers of Fulfulde, Kotoko, Malgwa or other languages would have
been involved as well. As shown in Owens & Hassan (2010), discourse markers
are prevalent in code-switching, which here would be conducted by Arabs codeswitching between Arabic and Kanuri. From this scenario the discourse markers
entered as borrowed elements.
The interpretation of demonstratives and idiomatic structure is equally ambiguous. The easiest development to envisage is L2 Arabic speakers imposing
their L1 Kanuri, Fulfulde etc. usage onto their L2 Arabic. What makes this interpretation attractive is that it explains why in both cases such a massive importation of non-Arabic structure came into Arabic. As the name implies, these speakers could simply have imposed their own semantics and collocational alignment
onto Arabic. Equally, however, it is not impossible that L1 Arabic speakers, fully
bilingual in Kanuri and/or other languages simply shifted their Arabic usage to
accommodate to their L2. Full fluency implies knowing idiomatic structure and
the use of demonstratives, which the Arab borrowers could eventually incorporate into their own Arabic.
The only obvious common denominator to these musings is that the speakers
would have been highly fluent in their respective L2s, whether L2 Arabic speakers shifting to Arabic or L1 Arabic speakers fluent in Kanuri or other languages
borrowing from their L2. The issue is only partly who the L1 and L2 speakers are.
It is equally how well the populations knew/know Arabic/other languages, and
how the high level of fluency produces the results shown.
192
8 Nigerian Arabic
Adding to the interpretive problem is that neither of the domains, idiomaticity or the expansion of demonstratives as it occurred in NA, have a comparative
basis. Idiomaticity in the recent western linguistic tradition has been all but entirely subordinated to metaphor theory (Lakoff & Johnson 1999; see Haser 2005
for one critical perspective). It has received very little principled historical interpretation, and what work has been done (e.g. Sweetser 1990) tends to follow a
Lakoffian paradigm and to be confined to European languages and to societies
quite different from that of Nigerian Arabs. As far as demonstratives go, the little work that has been done on the languages co-territorial with NA (e.g. Kramer
2014: 141 on Fali), assume a grammaticalization of demonstrative usage ab novo
via grammaticalization processes. Assuming such a perspective for the development of NA gives the lie to this simple assumption for the following reason. It
would need to explain why the grammaticalization process did not take place in
EA or other Arabic varieties, but did in NA, which is spoken in an area where
the co-territorial languages, historically antecedent to Arabic, have the structures
which NA acquired. If change via contact is the only plausible explanation for
NA, it equally needs to be entertained for any language in the Lake Chad region.
Given so many open variables, it might be interesting to approach the issue
from the opposite perspective, namely, what parts of language were not influenced by contact. Most of phonology was not, morphology hardly at all, syntax
to a degree, basic vocabulary little.5 This minimally implies that if the contact
changes were due to shift, the shifters in other domains (those where they did
not impose idiomaticity or demonstrative usage) acquired a native-like competence in Arabic. In this respect it might be easier to envisage L1 Arabic borrowers
maintaining these structures, and borrowing idiomaticity/demonstrative usage
via their L2.
At the end of the day I think the range of questions evoked far surpasses the
ability of currently-formulated linguistic theories of contact or language change,
whether based on sociolinguistic or on cognitive perspectives (Lucas 2015: 523)
to provide profound insight into how the obvious, and in some cases pervasive
influence on NA via contact came about. It would be more fruitful to turn the
question around and ask how rich databases such as exist for NA, EA and some
other Arabic dialects inform the overall issue of change via contact.
5
A Swadesh 100-word list gives something in the range of 79–83% cognacy with other varieties
of Arabic.
193
Jonathan Owens
Abbreviations
1, 2, 3
def
dem
dm
EA
exs
f
gen
impf
m
N
1st, 2nd, 3rd person
definite article
demonstrative
discourse marker
Egyptian Arabic
existential
feminine
genitive
imperfect (prefix conjugation)
masculine
number
n
NA
neg
np
obl
pl
prf
rel
sg
WSA
noun
Nigerian Arabic
negative
noun phrase
oblique
plural
perfect (suffix conjugation)
relative
singular
Western Sudanic Arabic
References
Behnstedt, Peter & Manfred Woidich. 1987. Die Ägyptisch-arabischen Dialekte.
Vol. 3: Delta Dialekte. Wiesbaden: Reichert.
Braukämper, Ulrich. 1994. Notes on the origin of Baggara Arab culture with
special reference to the Shuwa. Sprache und Geschichte in Afrika (SUGIA) 14.
13–46.
Broß, Michael. 2007. L2 speakers, koineization and the spread of language norms:
Hausa in Maiduguri, Nigeria. In Gudrun Miehe, Jonathan Owens & Manfred
von Roncador (eds.), Languages in African urban contexts, 51–74. Hamburg:
LIT.
Canavan, Alexandra, George Zipperlen & David Graff. 1997. CALLHOME
Egyptian Arabic Speech LDC97S45. Web Download. Philadelphia, Linguistic
Data Consortium.
Diessel, Holger. 1999. The morphosyntax of demonstratives in synchrony and
diachrony. Linguistic Typology 3. 1–46.
Frajzyngier, Zygmunt. 2012. Wandala. Berlin: Mouton de Gruyter.
Haser, Verena. 2005. Metaphor, metonymy and experientialist philosophy:
Challenging cognitive semantics. Berlin: Mouton de Gruyter.
Heine, Bernd & Tania Kuteva. 2011. The areal dimension of grammaticalization.
In Bernd Heine & Heiko Narrog (eds.), The Oxford handbook of grammaticalization, 291–301. Oxford: Oxford University Press.
Hutchison, John. 1981. The Kanuri language. Madison: African Studies Program.
Kramer, Raija. 2014. Die Sprache der Fali in Nordkamerun. Cologne: Rüdiger
Köppe.
Lakoff, George & Mark Johnson. 1999. Philosophy in the flesh. New York: Basic
Books.
194
8 Nigerian Arabic
Lucas, Christopher. 2015. Contact-induced language change. In Claire Bowern &
Bethwyn Evans (eds.), The Routledge handbook of historical linguistics, 519–536.
London: Routledge.
Manfredi, Stefano. 2010. A grammatical description of Kordofanian Baggara
Arabic. Naples: Università degli Studi di Napoli “L’Orientale”. (Doctoral
dissertation).
Manfredi, Stefano. 2014. Demonstratives in a Bedouin Arabic dialect of western
Sudan. Folia Orientalia 51. 27–50.
Max Lock Group. 1976. Max Lock Group report: Maiduguri: Surveys and planning
reports for Borno, Bauchi and Gongola State governments. London: Westminster.
Nakano, Aki’o. 1982. Folktales of Lower Egypt: Texts in Egyptian Arabic. Tokyo:
Institute for the Study of Languages, Cultures of Asia & Africa.
Owens, Jonathan. 1993. A grammar of Nigerian Arabic. Wiesbaden: Harrassowitz.
Owens, Jonathan (ed.). 1994a. Arabs and Arabic in the Lake Chad region. Special
issue of Sprache und Geschichte in Afrika (SUGIA) 14.
Owens, Jonathan. 1994b. Nigerian Arabic in comparative perspective. Sprache
und Geschichte in Afrika (SUGIA) 14. 85–176.
Owens, Jonathan. 1996. Idiomatic structure and the theory of genetic relationship.
Diachronica 13(2). 283–318.
Owens, Jonathan. 1998. Neighborhood and ancestry: Variation in the spoken Arabic
of Maiduguri (Nigeria). Amsterdam: John Benjamins.
Owens, Jonathan. 2000. Loanwords in Nigerian Arabic: A quantitative approach.
In Jonathan Owens (ed.), Arabic as a minority language, 259–346. Berlin: De
Gruyter.
Owens, Jonathan. 2002. Processing the world piece by piece: Iconicity, lexical insertion and possessives in Nigerian Arabic codeswitching. Language
Variation and Change 14(2). 173–209.
Owens, Jonathan. 2006. A linguistic history of Arabic. Oxford: Oxford University
Press.
Owens, Jonathan. 2013. The historical linguistics of the intrusive -n in Arabic and
West Semitic. Journal of the American Oriental Society 133(2). 217–248.
Owens, Jonathan. 2014. Many heads are better than one: The spread of motivated
opacity via contact. Linguistics 52(1). 125–165.
Owens, Jonathan. 2015. Idioms, polysemy, context: A model based on Nigerian
Arabic. Anthropological Linguistics 57(1). 46–98.
Owens, Jonathan. 2016a. Dia-planar diffusion: Reconstructing early Aramaic–
Arabic language contact. In Manuel Sartori, Manuela Giolfo & Philippe
Cassuto (eds.), Approaches to the history and dialectology of Arabic in honor
of Pierre Larcher, 77–101. Leiden: Brill.
195
Jonathan Owens
Owens, Jonathan. 2016b. The lexical nature of idioms. Language Sciences 57. 49–
69.
Owens, Jonathan. Forthcoming. Arabic. In Salikoko S. Mufwene & Anna María
Escobar (eds.), The Cambridge handbook of language contact. Cambridge:
Cambridge University Press.
Owens, Jonathan & Robin Dodsworth. 2017. Semantic mapping: What happens
to idioms in discourse. Linguistics 55(3). 641–682.
Owens, Jonathan & Jidda Hassan. 2004. Remarks on ideophones in Nigerian
Arabic. In Martine Haak, Rudolf de Jong & Kees Versteegh (eds.), A Festschrift
for Manfred Woidich, 207–220. Leiden: Brill.
Owens, Jonathan & Jidda Hassan. 2010. Conversation markers in Arabic–Hausa
codeswitching: Saliency and language hierarchies. In Jonathan Owens &
Alaa Elgibali (eds.), Information structure in spoken Arabic, 207–242. London:
Routledge.
Owens, Jonathan & Jidda Hassan. 2011. In their own voices, in their own words:
A corpus of spoken Nigerian Arabic. http : / / www . neu . uni - bayreuth . de /
de / Uni _ Bayreuth / Fakultaeten / 4 _ Sprach _ und _ Literaturwissenschaft /
islamwissenschaft / arabistik / en / Idiomaticity _ _lexical _ realignment _ _and _
semantic_change_in_spoken_arabic/Nigerian_Arabic/index.html.
Retsö, Jan. 2013. What is Arabic? In Jonathan Owens (ed.), The Oxford handbook
of Arabic linguistics, 433–450. Oxford: Oxford University Press.
Ritt-Benmimoun, Veronika, Smaranda Grigore, Jocelyne Owens & Jonathan
Owens. 2017. Three idioms, three dialects, one history: Egyptian, Nigerian
and Tunisian Arabic. In Veronika Ritt-Benmimoun (ed.), Tunisian and Libyan
Arabic dialects: Common trends – recent developments – diachronic aspects, 43–
84. Zaragoza: Prensas de la Universidad de Zaragoza.
Stevenson, R. C. 1969. Bagirmi grammar. Khartoum: Sudan Research Unit, University of Khartoum.
Sweetser, Eve. 1990. From etymology to pragmatics: Metaphorical and cultural
aspects of semantic structure. Cambridge: Cambridge University Press.
Van Coetsem, Frans. 1988. Loan phonology and the two transfer types in language
contact. Dordrecht: Foris.
Van Coetsem, Frans. 2000. A general and unified theory of the transmission process
in language contact. Heidelberg: Winter.
Woidich, Manfred & Hanke Drop. 2007. ilBaħariyya – Grammatik und Texte.
Wiesbaden: Harrassowitz.
196
Chapter 9
Maghrebi Arabic
Adam Benkato
University of California, Berkeley
This chapter gives an overview of contact-induced changes in the Maghrebi dialect
group in North Africa. It includes both a general summary of relevant research on
the topic and a selection of case studies which exemplify contact-induced changes
in the areas of phonology, morphology, syntax, and lexicon.
1 The Maghrebi Arabic varieties
In Arabic dialectology, Maghrebi is generally considered to be one of the main
dialect groups of Arabic, denoting the dialects spoken in a region stretching from
the Nile delta to Africa’s Atlantic coast – in other words, the dialects of Mauritania, Morocco, Algeria, Tunisia, Libya, parts of western Egypt, and Malta. The
main isogloss distinguishing Maghrebi dialects from non-Maghrebi dialects is
the first person of the imperfect, as shown in Table 1 (cf. Lucas & Čéplö, this
volume).1
Table 1: First-person imperfect ‘write’ in Maghrebi and non-Maghrebi
Arabic
Non-Maghrebi
Singular
Plural
1
Maghrebi
Classical Arabic
Baghdad Arabic
Casablanca Arabic
Maltese
aktub
naktub
aktib
niktib
nəktəb
nkətbu
nikteb
niktbu
More about the exact distribution of this isogloss can be found in Behnstedt (2016).
Adam Benkato. 2020. Maghrebi Arabic. In Christopher Lucas & Stefano Manfredi
(eds.), Arabic and contact-induced change, 197–212. Berlin: Language Science Press.
DOI:10.5281/zenodo.3744517
Adam Benkato
This Maghrebi group of dialects is in turn traditionally held to consist of two
subtypes: those spoken by sedentary populations in the old urban centers of
North Africa, and those spoken by nomadic populations. The former of these,
usually referred to as “pre-Hilali” (better: “first-layer”) would have originated
with the earliest Arab communities established across North Africa (~7th–8th
centuries CE) up to the Iberian Peninsula. The latter of these, usually referred to
as “Hilali” (better: “second-layer”), is held to have originated with the westward
migration of a large group of Bedouin tribes (~11th century CE) out of the Arabian
Peninsula and into North Africa via Egypt. Their distribution is roughly as follows.2 First-layer dialects exist in cities such as Tunis, Kairouan, Mahdia, Sousse,
Sfax (Tunisia), Jijel, Algiers, Cherchell, Tlemcen (Algeria), Tangier, Tetuan, southern Rif villages, Rabat, Fez, Taza, so-called “northern” dialects (Morocco), Maltese,
and formerly Andalusi and Sicilian dialects; most Judeo-Arabic dialects formerly
spoken in parts of North Africa are also part of this group. Second-layer dialects are spoken by populations of nearly all other regions, from western Egypt,
through all urban and rural parts of Libya, to the remaining urban and rural parts
of Algeria and Morocco. Though some differences between these two subtypes
are clear (such as [q, ʔ, k] vs. [g] for *q), there have probably been varying levels of
interdialectal mixture and contact since the eleventh century CE. In many cases,
first-layer varieties of urban centers have been influenced by neighboring secondlayer ones, leading to new dialects formed on the basis of inter-dialectal contact.
It is important to note that North Africa is becoming increasingly urbanized and
so not only is the traditional sedentary/nomadic distinction anachronistic (if it
was ever completely accurate), but also that intensifying dialect contact accompanying urbanization means that new ways of thinking about Maghrebi dialects
are necessary. It is also possible to speak of the recent but ongoing koinéization
of multiple local varieties into supralocal or even roughly national varieties—
thus one can speak, in a general way, of “Libyan Arabic” or “Moroccan Arabic”.
This chapter will not deal with contact between mutually intelligible varieties of
a language although this is equally important for the understanding of both the
history and present of Maghrebi dialects.3
2
More will not be said about the subgroups of Maghrebi dialects that have been proposed. For
more details about the features and distribution of Maghrebi dialects see Pereira (2011); for
more detail on the complex distribution of varieties in Morocco see Heath (2002).
3
The emergence of new Maghrebi varieties resulting from migration and mixture is discussed
in Pereira (2007) and Gibson (2002), for example. The oft-cited distinction between urban and
nomadic dialects is also problematized by the existence of the so-called rural or village dialects
(though this is also a problematic ecolinguistic term), on which see Mion (2015). Dialect contact
outside of the Maghreb is discussed by Cotter (this volume).
198
9 Maghrebi Arabic
2 Languages in contact
Contact between Arabic and other languages in North Africa began in the late
seventh century CE, when Arab armies began to spread westward through North
Africa, reaching the Iberian Peninsula by the early eighth century CE and founding or occupying settlements along the way. Their dialects would have come
into contact with the languages spoken in coastal regions at that time, including varieties of Berber and Late Latin, and possibly even late forms of Punic and
Greek. The numbers of Arabic speakers moving into North Africa at the time of
initial conquests were likely to have been quite small.4 By the time of the migration of Bedouin groups beginning in the eleventh century, it is doubtful that languages other than Berber and Arabic survived in the Maghreb. The Arabization
of coastal hinterlands and the Sahara increased in pace after the eleventh century. Berber varieties continue to be spoken natively by millions in Morocco and
Algeria, and by smaller communities in Libya, Tunisia, Mauritania, and Egypt.
Any changes in an Arabic variety due to Berber are almost certainly the result
of Berber speakers adopting Arabic rather than Arabic speakers adopting Berber
– the sociolinguistic situation in North Africa is such that L1 Arabic speakers
rarely acquire Berber.
Beginning in the sixteenth century, most of North Africa came under the
control of the Ottoman Empire and thus into contact with varieties of Turkish, although the effect of Turkish is essentially limited to cultural borrowings
(see §3.4). The sociolinguistic conditions in which Turkish was spoken in North
Africa are poorly understood.
The advent of colonialism imposed different European languages on the region, most prominently French (in Mauritania, Morocco, Algeria, and Tunisia),
Italian (in Libya), and Spanish (in Morocco). Romance words in dialects outside of
Morocco may also derive from forms of Spanish (via Andalusi refugees to North
Africa in the 16th–17th centuries) or from the Mediterranean Lingua Franca.5
The effects on Maghrebi Arabic of contact with Chadic (e.g. Hausa) or NiloSaharan (e.g. Songhay, Tebu) languages is largely unstudied since in most cases
data from the relevant Arabic varieties is lacking. Yet some borrowings from
these languages can be found in Arabic and Berber varieties throughout the region (Souag 2013).6 Lastly, Hebrew loans are present in most Jewish Arabic dialects of North Africa (Yoda 2013), though unfortunately these dialects hardly exist
anymore.
4
See Heath (this volume) for discussion of Late Latin influence in Moroccan Arabic dialects.
On the Lingua Franca see Nolan (this volume).
6
See also Souag (2016) for an overview of contact in the Sahara region not limited to Arabic.
5
199
Adam Benkato
To restate these facts in Van Coetsem’s (1988; 2000) terms, there are two major
contact situations at work in Maghrebi Arabic in general, though the specifics
will of course differ from variety to variety. The first is change in Arabic driven
by source-language (Berber) dominant speakers; this transfer type is imposition.
The second is change in Arabic driven by recipient-language (Arabic) dominant
speakers where the source language is a European colonial language; this transfer type is called borrowing.7 So far, “dominance” describes linguistic dominance, that is, the fact that a speaker is more proficient in one of the languages
involved in the contact situation. However, social dominance, referring to the
social and political status of a language (Van Coetsem 1988: 13), is also important,
especially in North Africa.
3 Contact-induced changes in Maghrebi dialects
3.1 Phonology
Changes in Maghrebi Arabic phonology due to contact with Berber are difficult
to prove. There are several cases, for example, where historical changes in Arabic
phonology may be argued to be the result of contact with Berber or the result
of internal developments. These include the change of *ǧ to /ž/ in many varieties, or the emergence of phonemic /ẓ/ (Souag 2016). Another example, the pronunciation /ṭ/ in some first-layer varieties where most Arabic varieties have /ð̣/,
has also been explained as a result of Berber influence, or as unclear directionality (Kossmann 2013a: 187), while Al-Jallad (2015) argues that it is actually an
archaism within Arabic.
The merger in Arabic of the vowels *a and *i (and even *u) to a single phoneme
/ǝ/ in some, especially first-layer, varieties, is often attributed to Berber influence,
as many Berber varieties have only a single short vowel phoneme /ǝ/. However
Kossmann (2013a: 171–174) points out that Berber also merged older *ă and *ǝ to
a single phoneme /ǝ/ and that it cannot be proven that the reduction happened in
Berber before it happened in Arabic. Hence, again the directionality of influence
is difficult to show.
Related to this development is also that many Maghrebi varieties disallow
vowels in light syllables (often described as the deletion of short vowels in open
syllables), such that *katab ‘he wrote’ > Tripoli ktǝb or *kitāb ‘book’ > Algerian
ktāb.8 Meanwhile, second-layer varieties often do allow vowels in light syllables
7
Another good illustration of the two transfer types in the Van Coetsemian framework can be
found in Winford (2005: 378–381).
8
Since the short vowels merge to schwa in many Moroccan and Algerian varieties, vowel length
is no longer contrastive and it is common to transcribe e.g. ktab rather than ktāb.
200
9 Maghrebi Arabic
(e.g. Benghazi kitab ‘he wrote’, Douz mi šē ‘he went’). While proto-Berber and
some modern varieties allow vowels in light syllables, most Berber varieties of
Algeria and Morocco do not. This is another example of a similar development
wherein the directionality of influence is unclear (see Souag 2017: 62–65 for further discussion).
In the Arabic variety of Ghomara, northwest Morocco, *d and *t are spirantized to /ð/ and /θ/ initially (*d only), postvocalically and finally (Naciri-Azzouz
2016): e.g. māθǝθ ‘she died’ (*mātat), warθ ‘inheritance’ (although etymologically *warθ, dialects of the wider Jbala region of Morocco have no interdentals so *wart), ðāba ‘now’ (*dāba), ḫǝðma ‘work’ (*ḫidma), wāḥǝð ‘one’ (*wāḥid).
Naciri-Azzouz points out that the distribution of spirantization is the same as in
Ghomara Berber, a variety spoken by groups in the same region.9
New phonemes have been borrowed into Maghrebi varieties through contact
with European languages: for example, /p/ and nasalized vowels in more recent
French loans in Tunisian Arabic, or /v, č, ǧ/ in Italian loans in Libyan Arabic
(grīǧū ‘gray’ < grigio).
3.2 Morphology
In the realm of morphology, changes in Arabic varieties due to contact vary depending on whether the relationship between Arabic and the contact language
is substratal, adstratal, or superstratal.
Morphological influence from Berber on the Arabic varieties of the northern
Maghreb is not overly common.10 In some places where Berber–Arabic bilingualism is or was more common, contact has led to the borrowing of Berber nouns
into Arabic together with their morphology, a phenomenon known as “parallel
system borrowing”.11 In Ḥassāniyya, for example, many nouns have been transferred together with their gender and number marking.12 In the dialect of Jijel,
Berber singular nouns are transferred together with their prefixes (āwtūl ‘hare’,
cf. Kabyle āwtūl); plurals are then formed in a way which resembles Berber but is
9
The Berber variety of Ghomara exhibits an extreme amount of influence from dialectal Arabic,
see Mourigh (2015). Kossmann (2013a: 431) writes that given the existence of parallel morphological systems for virtually all grammatical categories (nominal, adjectival, pronominal and
verbal morphology) and a high loanword count (more than 30% of basic lexicon is Arabic) it
would be possible to call Ghomara Berber a mixed language.
10
Documentation of the varieties where such influence would be more expected, such as Arabicspeaking towns in the otherwise Berber-speaking Nafusa Mountains in Libya, is lacking.
11
For a closer look at parallel system borrowing in the context of Arabic and Berber contact, see
Kossmann (2010), mostly discussing the borrowing of Arabic paradigms into Berber.
12
See Taine-Cheikh (this volume).
201
Adam Benkato
not identical (Jijel āsrǝf, āsǝrfǝn ‘bush(es)’, cf. Kabyle Berber āsrǝf, īsǝrfǝn); moreover, the prefix ā- is also used with nouns of Arabic origin (āfḫǝd ‘thigh’, Arabic
*faḫað) (Marçais 1956: 302–318).
In Algeria and Morocco the circumfix tā-...-t, which occurs on feminine nouns
in Berber, can derive abstract nouns (e.g. Jijel tākǝbūrt ‘boasting’, tāwǝḥḥūnt ‘having labor pains’) and in Moroccan Arabic tā-...-t is the regular way of forming
nouns of professions and traits (e.g. tānǝžžāṛt ‘carpentry’) (Kossmann 2013b).
The verbal morphology of Arabic dialects is much less affected by Berber,
though Ḥassāniyya again provides an interesting example. It has a causative prefix sä- used with both inherited Arabic verbs and borrowed Berber verbs, and
most likely to be borrowed from Berber causative forms in s-/š- (Taine-Cheikh
2008).
Turkish influence on morphology is restricted to the suffix -ği/-ži (< -ci) used
to indicate professions and borrowed widely into Arabic dialects in general. In
Tunisia, its use has been extended to derive adjectives of quality from nouns
(sukkārži ‘drunkard’) and has also even been added to borrowed French nouns
(bankāži ‘banker’ < French banque). As Manfredi (2018: 410) points out, the productivity of this borrowed derivational morpheme constitutes one example of
how recipient-language agentivity can introduce morphological innovations via
borrowing.
French (and other Romance) verbs are also routinely borrowed into Maghrebi
varieties. Talmoudi (1986) discusses their integration into different forms of the
verbal system of Tunisian Arabic, e.g. mannak ‘to be absent’ < French manquer
(1986: 81–82) or (t)rānā ‘to train’ < French entrainer (1986: 21–24).
3.3 Syntax
Syntax is often the least documented aspect of the grammar of Maghrebi Arabic
varieties and research on contact-induced changes in syntax is still in its infancy.
Much attention has been devoted recently to explaining the rise of bipartite negation in Arabic and Berber; in varieties of both languages the word for ‘thing’
(Arabic šayʔ, Berber *ḱăra) has been grammaticalized postverbally in a marker
of negation:
(1)
202
Arabic (Benghazi)
mā-šift-hā-š
neg-see.prf.1sg-3sg.f-neg
‘I didn’t see her’.
(2)
Berber (Tarifit)
wā t-ẓṛiɣ
ša
neg 3sg.f-see.prf.1sg neg
‘I didn’t see her’.
9 Maghrebi Arabic
Although some accounts give no attention to Berber, while others attribute
the Arabic development solely to Berber, the development in both languages
in the same contexts is probably not a coincidence, though there is no current
consensus on the direction of transfer – see Lucas (this volume) for discussion.13
However, it must be noted that not all Berber varieties have double negation (e.g.
Tashelhiyt ur nniγ ak ‘I didn’t tell you’ where the only negator is ur).
In another area, recent work on the variety of Tunis has yielded interesting
conclusions: while possessives with French nouns are overwhemingly analytic
(l-prononciation mtēʕ-ha ‘her pronunciation’) and those with Arabic nouns are
almost as overwhelmingly synthetic (nuṭq-u ‘his pronunciation’), the frequent
occurence of French loan nouns may be triggering an increase in the overall frequency of analytical possessives over syntactic ones, including those with Arabic
nouns (Sayahi 2015).
The remainder of this section will discuss one particularly interesting case:
the first-layer dialect of Jijel, a city in eastern Algeria. At the time of its description (Marçais 1956), it showed little influence from second-layer varieties, but
displayed wide-ranging influence from Berber in multiple domains. In a recent
article, Kossmann (2014) has demonstrated how a Berber marker of non-verbal
predication was adopted into the Arabic dialect of Jijel as a focus marker. Here I
will briefly summarize Kossmann’s arguments with a few examples. In the Jijel
dialect, as described by Marçais and reanalyzed by Kossmann, a morpheme d occurs in the following syntactic contexts (examples (3–7) are all from Kossmann
2014: 129–131, who retranscribes from Marçais’ texts): before non-verbal predicates (3), in clefts with a noun/pronoun in the cleft (4), in secondary predication
with a specific noun (5), as a marker of subject (or object) focus (6), and in leftmoved focalizations (7).
(3)
l-lila
d-ǝl-ʕid
def-night d-def-feast
‘Tonight is the feast.’
(4)
d-hum ǝddǝ šraw-ǝh
qbǝl-ma
nǝzdad
d-3pl.m rel buy.prf.3pl.m-3sg.m before-comp be.born.impf.1sg
‘It is them who bought it before I was born.’
13
See Lucas (2007; 2010; 2018) and Souag (2018) for further discussion of the grammaticalization of ‘thing’ for indefinite quantification and polar question marking in Arabic and Berber.
Kossmann (2013a: 324–334) surveys the situation in the Berber languages. See Lafkioui (2013)
for an overview of negation in especially Moroccan Arabic, as well as discussion of a variety
of Moroccan Arabic which features the discontinuous morpheme mā- ... -bū, where the latter
part has been borrowed from Tarifit.
203
Adam Benkato
(5) ṛa-na
nqǝṭṭʕu-č
d ǝṭ-ṭraf
prst-1pl cut.impf.1pl-sg d def-pieces
‘We will cut you (into) pieces.’
(6) tkǝṣṣṛǝt
d l-idura
break.prf.3sg.f d def-bowl
‘The bowl has broken.’
(7)
qalu
d ǝṛ-ṛbiʕ
dǝḫlǝt
say.prf.3pl d def-spring enter.prf.3sg.f
‘They say spring has come.’
Although previous analyses attempted to explain d within Arabic, Kossmann
notes that an Arabic-internal derivation of d is impossible. However, Kabyle, the
Berber language neighboring the Jijel area has an element d (realized [ð] due
to spirantization in Kabyle) which is used in (pro)nominal predicates (8), cleft
constructions (9), and secondary predication when non-verbal (10). Examples (8–
10) are all Kabyle Berber, taken from Kossmann (2014: 135–136). This element d is
attested in Berber more widely, too, and is likely reconstructible to older stages
of the language.
(8)
d-yǝlli-m
d-daughter-2sg.f
‘Is it your daughter?’
(9)
d-ay-ǝn
i d-tǝnna
abrid amǝnzu
d-this-deict rel hither-say.3sg.f road first
‘This is what she said the first time.’
(10) ad nǝǧʕǝl iman nn-ǝɣ d-inǝbgiwǝn n ṛǝbbi
mod make.1pl self gen-1pl d-guests
gen lord
‘We shall pretend to be beggars (lit. guests of God).’
Thus Berber d is the best candidate for the origin of Jijel Arabic d, though its
usage in (Kabyle) Berber (where it is primarily a marker of syntactic organization) differs from that of Jijel Arabic (where it is mainly a marker of information
structure). In a simplified scenario with a Berber variety as source language and
Jijel Arabic as recipient, d would likely have been imposed into Jijel Arabic with
its exact Berber functions. As Kossmann notes, though, speech communities are
full of variation and language contact is a “negotiation between the frequency
of non-native speech and the prestige of the native way of speaking” (Kossmann
2014: 138). Kossmann thus proposes a scenario in which larger groups of Berber
speakers switched to a variety of Jijel Arabic and began imposing their own d; the
204
9 Maghrebi Arabic
native Jijel Arabic speakers, fewer in number, began adopting d but understood
it differently and interpreted it as a focus marker, introducing it into new contexts; eventually the variety of Jijel Arabic with d in all these functions became
nativized. Per Kossmann (2014: 138–139), two processes would have taken place:
the transfer of a source-language feature by speakers dominant in the source language (Berber), followed by the borrowing of this feature by speakers dominant
in the recipient language (Arabic), and its eventual regularization in that variety.
Jijel Arabic is an excellent example of what may happen when large numbers of
Berber speakers switch to Arabic.
3.4 Lexicon
Much work on contact and Maghrebi Arabic has focused on loanwords, the most
salient effects of borrowing, with secondary attention to their phonological or
morphological adaptation. The concept of social dominance has particular relevance for borrowing: in the North African context, the colonial languages, and especially French, have high social status for both Arabic and Berber native speakers. One also must modify the idea of linguistic dominance to include those who
acquire two languages natively (2L1 speakers; see Lucas 2015: 525), definitely the
case for certain speakers of Berber and Arabic in North Africa.
Unsurprisingly, we see firstly that the majority of words borrowed into Arabic varieties are nouns, and secondly that the lexical domains into which these
borrowings fall are often restricted. Social dominance seems to play a role in the
nature of the nouns borrowed.
Berber loans are found in most Maghrebi Arabic varieties, though their number ranges from only a handful of words in the east to many more in the west
(cf. §3.2 above). Almost all Maghrebi varieties have borrowed the words ž(i)ṛāna
‘frog’ and fakrūna ‘turtle’, while in some oases Berber influence in agricultural
terminology can be seen. Again, the documentation of the relevant varieties is
often insufficient.
Several studies on contact between Maghrebi Arabic varieties and European
languages exist. For French in Morocco, Heath (1989) argues that code-switching
and borrowing are essentially the same in a bilingual community which has
established borrowing routines.14 For French in Tunisia, Talmoudi (1986) analyzes the phonological and morphological adaptation of French verbs into Arabic.
14
Van Coetsem (1988: 87) notes that for bilingual speakers who have a balance in linguistic dominance between the two languages, the separation between the two transfer types (borrowing
and imposition) will be weaker. Hence, either of the two dominant languages can serve as
the recipient language in code-switching behavior. Winford (2005, esp. 394–396), expanding
on Van Coetsem’s framework, points out that code-switching is inherently linked to the borrowing transfer type. In the Maghreb, this scenario is possible for Berber–Arabic bilinguals
as well as for some French–Arabic bilinguals. See Ziamari (2008) for an insightful and more
205
Adam Benkato
Sayahi (2014: 127–151) gives a broader view of lexical borrowing in diglossic or
bilingual communities, focusing on French in Tunisia and Spanish in Morocco.
Vicente (2005) studies Arabic-Spanish code-switching in Ceuta, a Spanish enclave in northern Morocco. Italian in Tunisia is studied briefly by Cifoletti (1994).
Studies of contact with Turkish are limited to discussion of lexical borrowing:
on Morocco see Procházka (2012); on Algeria, see Ben Cheneb (1922), to be read
with the review by G. S. Colin (Colin 1999: 21–30).
The remainder of this section will consider the influence of Turkish and Italian
on Libyan Arabic (henceforth LA), a hitherto under-researched topic. Uniquely in
the Maghreb region there is at present no superstratum language spoken widely
by Arabic speakers in Libya, while there are also fewer Berber speakers than in
Algeria or Morocco. As far as documented varieties of LA (Tripoli and Benghazi)
go, contact situations are historical and not active.
There seems to be an impression among dialectologists that LA varieties have
the largest number of Turkish loans, though there is not a published basis for this.
Procházka (2005: 191) suggests that the number of (Ottoman) Turkish loans in a
given Arabic dialect is proportional to the length and intensity of Ottoman rule.
By this criterion Libya should have quite a few, as the regions now constituting
Libya were under control of the Ottoman Empire from 1551 to 1911, but Procházka
estimates that the dialect would show 200 to 500 surviving loans, fewer than in
other dialects. Another important factor is likely to be that Libya’s population
was very small during the period of Ottoman rule so that the long-term presence
of even a few thousand Turkish speakers could have had a significant effect. However, I cannot yet offer a statistical analysis of Turkish words in LA.15 It is clear so
far, though, that the effects of Turkish on LA can mainly be seen in the lexicon
and, in my data, almost entirely in nouns. In terms of their semantic domains,
Procházka (2005: 192) points out that the majority of Turkish loans in Arabic dialects in general fall into three categories, roughly described as: private life; law,
government, social classes; and army, war. By far the majority of surviving loans
would belong to the first of these classes (such as šīšma ‘tap’ < çeşme, dizdān ‘wallet’ < cüzdan), or the second (such as fayramān ‘order’ < ferman, ḥafð̣a ‘week’ <
hafte) while I suspect that words from the third class are increasingly rarer. Outside of these, only a few words other than nouns seem to be present, such as
recent analysis of Moroccan Arabic in contact with French using a “matrix language frame”
analysis.
15
The only study dedicated to Turkish loans in LA is Türkmen (1988), who lists 90 words. However, the basis for his wordlist seems unclear and several items are either spurious or incorrect
(e.g. there is no word kabak ‘pumpkin’ in Benghazi Arabic but there is bkaywa ‘pumpkin’,
identified by Souag (2013) as a loan from Hausa). Turkish words in LA cited here are from the
Benghazi variety, author’s data.
206
9 Maghrebi Arabic
duġri ‘straight ahead’ and balki ‘maybe’. The length of time since Turkish was
last actively spoken in Libya no doubt means that the number of Turkish loans
actively used by speakers has been decreasing.
LA is unique among Maghrebi varieties in having had Italian as the main European contact language. Italian had a presence in what is now Libya from the
1800s, but this was mainly limited to the Tripolitanian Jewish community and
wealthy merchant families. The Italian colonization of Libya officially began in
1911; though the majority of the region was not brought under Italian control until the early 1930s, large numbers of Italian colonists had begun to settle in Libya
in the 1920s. From that period until 1970, when the remaining Italian citizens were
expelled from the country, Italians made up 15% or more of the population and
the language was in widespread use. From the 1970s on, Italian was scarcely used
in Libya, and the teaching of foreign languages was banned in 1984, not to return
again until 2005.16 Many of the postwar generation spoke (and still speak) Italian, though they rarely use it anymore, but few Libyans of younger generations
do. The 1920s to the 1970s can thus be regarded as the main period of contact
between LA and Italian.17 However, the concentration of Italians differed from
region to region and thus may have influenced local varieties differently. The primary study devoted to analyzing Italian loans in LA is that of Abdu (1988) who,
focusing on the variety of Tripoli, draws up a list of nearly 700 items (a few are
misidentified), of which about 50% were recognized by a majority of those surveyed. Some 93% of these are nouns and the remainder are practically all derived
from nouns or adjectives, such as bwōno ‘well done!’ < buono ‘good’ or faryaz
‘to go out of order’ < Italian fuori uso.18 Abdu’s study (1988: 248–268) groups
Italian loans into some 22 semantic categories, the vast majority of which relate
to material culture. Examples of these from the Benghazi variety are byāmbu
‘lead’ < piombo, bōskō ‘zoo’ < bosco ‘wood’, furkayta ‘fork’ < forchetta, maršabīdi
‘sidewalk’ < marciapiede (author’s data).
As D’Anna (2018) points out, the adaptation of Italian words to LA phonology
varies: new phonemes, particularly [v] and [č], sometimes occur but are sometimes adapted to the dialects’ pre-existing phonologies, an indication of “subsidiary phonological borrowing” (Van Coetsem 1988: 98). Of course, the maintenance of new phonemes often depends on speakers continuing to have access
16
For more information on the return of Italian instruction to Libya, see D’Anna (2018).
The Italian words in Yoda’s (2005) study of Tripoli Judeo-Arabic need to be seen slightly differently than Italian words in non-Jewish dialects, owing to a different history of contact between
the Tripolitanian Jewish community and Italy.
18
See Abdu (1988: 271) and D’Anna (2018). Some denominal verbs are cited by Abdu, but more
extensive data might reveal several more in use: for example in the variety of Benghazi, I
identified fuṛan ‘to brake (intransitive)’ < frayno ‘brake’ < Italian freno, not listed by Abdu.
17
207
Adam Benkato
to the source language; as this is no longer the case in Libya, Italian borrowings in LA are traversing a different trajectory than French borrowings in other
Maghrebi varieties, where only the oldest borrowings have been phonologically
integrated.
The overwhelming majority of surviving Turkish and Italian loans in LA are
nouns, widely acknowledged to be the most easily-borrowed word class due to
their being the least disruptive of the recipient language’s argument structure
(Myers-Scotton 2002), though a few verbs derived dialect-internally do exist. Furthermore, almost all the nouns are cultural borrowings — “lexical content-words
that denote an object or concept hitherto unfamiliar to the receiving society, terminology related to institutions that are the property of the neighboring [or colonizing] culture, and so on” (Matras 2011: 210). Cultural borrowings are to be
differentiated from core borrowings, the latter being words that more or less duplicate already existing words and which originate in a bilingual code-switching
context. These facts lead us to conclude that Turkish and Italian borrowings in
Libyan varieties would be from (1) to (2) on the borrowing scale proposed by
Thomason & Kaufman (1988: 78–83). While (1) of the scale involves lexical borrowing of non-basic vocabulary only, (2) includes some function words as well
as new phones appearing in those loanwords. Colonial language contact situations are typically ones of recipient-language agentivity, as the number of indigenous people learning the colonial language is many times more than the
number of colonizers learning indigenous languages. Without a longer period
of sustained bilingualism or language education motivated by continued contact
with the metropole, Italian has affected LA to a much smaller degree than French
has Libya’s Maghrebi neighbours.
4 Conclusion
The general parameters of the Maghrebi linguistic landscape and contact situations are relatively well understood. However, more documentation of Maghrebi
varieties is needed, and more specifically, of those where contact situations – especially with Berber – may have existed. Additionally, further research into the
sociolinguistic factors affecting bilingualism in Berber and Arabic, or regarding
the intersection of diglossia with bilingualism, will no doubt add to our knowledge of the parameters of contact-induced change more generally. Finally, interdialectal contact as well as the gradual rise of national or at least supra-local
varieties certainly merits continuing attention.
208
9 Maghrebi Arabic
Further reading
) Kossmann (2013a) is the most extensive study so far of Berber–Arabic contact,
written from a Berberological point of view but important for Arabists.
) Sayahi (2014) studies the intersection of dialects, Standard Arabic, French and
Spanish in Tunisia and Morocco.
) Souag (2016) summarizes contact in the Saharan region among Arabic, Berber,
Hausa, Songhay, Chadic, etc.
) Ziamari (2008) is the most up-to-date work discussing code-switching and
borrowing strategies between Moroccan Arabic and French.
Acknowledgements
The research for this article was supported by a grant from the Alexander von
Humboldt-Stiftung.
Abbreviations
1, 2, 3
comp
def
deict
f
gen
impf
LA
1st, 2nd, 3rd person
complementizer
definite
deictic
feminine
genitive
imperfect (prefix conjugation)
Libyan Arabic
m
mod
neg
pl
prf
prst
rel
sg
masculine
modal
negative
plural
perfect (suffix conjugation)
presentative
relative
singular
References
Abdu, Hussein Ramadan. 1988. Italian loanwords in colloquial Libyan Arabic as
spoken in the Tripoli region. Tucson: University of Arizona. (Doctoral dissertation).
Al-Jallad, Ahmad. 2015. On the voiceless reflex of *ṣ and *ṯ̣ in pre-Hilalian
Maghrebian Arabic. Zeitschrift für Arabische Linguistik 62. 88–95.
Behnstedt, Peter. 2016. The niktib-niktibu issue revisited. Wiener Zeitschrift für
die Kunde des Morgenlandes 106. 21–36.
Ben Cheneb, Mohammed. 1922. Mots turks et persans conservés dans le parler
algérien. Algiers: Université d’Alger.
209
Adam Benkato
Cifoletti, Guido. 1994. Italianismi nel dialetto di Tunisi. In Dominique Caubet &
Martine Vanhove (eds.), Actes des premières journées internationales de dialectologie arabe de Paris, 451–458. Paris: Langues’O.
Colin, Georges S. 1999. Arabe marocain: Inédits de Georges S. Colin. Dominique
Caubet & Zakia Iraqui-Sinaceur (eds.). Paris: INALCO.
D’Anna, Luca. 2018. Phonetical and morphological remarks on the adaptation of
Italian loanwords in Libyan Arabic. In Stefano Manfredi & Mauro Tosco (eds.),
Arabic in contact, 172–187. Amsterdam: John Benjamins.
Gibson, Maik. 2002. Dialect levelling in Tunisian Arabic: Towards a new standard. In Aleya Rouchdy (ed.), Language contact and language conflict in Arabic:
Variations on a sociolinguistic theme, 24–30. London: Routledge.
Heath, Jeffrey. 1989. From code-switching to borrowing: A case study of Moroccan
Arabic. London: Kegan Paul.
Heath, Jeffrey. 2002. Jewish and Muslim dialects of Moroccan Arabic. London:
Routledge.
Kossmann, Maarten. 2010. Parallel system borrowing: Parallel morphological
systems due to the borrowing of paradigms. Diachronica 27(3). 459–487.
Kossmann, Maarten. 2013a. The Arabic influence on Northern Berber. Leiden: Brill.
Kossmann, Maarten. 2013b. Borrowing. In Jonathan Owens (ed.), The Oxford
Handbook of Arabic Linguistics, 349–368. Oxford: Oxford University Press.
Kossmann, Maarten. 2014. On substratum: The history of the focus marker d in
Jijel Arabic (Algeria). In Carole de Féral, Maarten Kossmann & Mauro Tosco
(eds.), In and out of Africa: Languages in question. In honour of Robert Nicolaï,
vol. 2: Language contact and language change in Africa, 127–140. Leuven:
Peeters.
Lafkioui, Mena. 2013. Reinventing negation patterns in Moroccan Arabic. In
Mena Lafkioui (ed.), African Arabic: Aprroaches to dialectology, 51–93. Berlin:
De Gruyter Mouton.
Lucas, Christopher. 2007. Jespersen’s cycle in Arabic and Berber. Transactions of
the Philological Society 105(3). 398–431.
Lucas, Christopher. 2010. Negative -š in Palestinian (and Cairene) Arabic: Present
and possible past. Brill’s Annual of Afroasiatic Languages and Linguistics 2. 165–
201.
Lucas, Christopher. 2015. Contact-induced language change. In Claire Bowern &
Bethwyn Evans (eds.), The Routledge handbook of historical linguistics, 519–536.
London: Routledge.
Lucas, Christopher. 2018. On Wilmsen on the development of postverbal negation
in dialectal Arabic. Zeitschrift für Arabische Linguistik 67. 44–70.
210
9 Maghrebi Arabic
Manfredi, Stefano. 2018. Arabic as a contact language. In Reem Bassiouney &
Elabbas Benmamoun (eds.), The Routledge handbook of Arabic linguistics, 407–
20. London: Routledge.
Marçais, Philippe. 1956. Le parler arabe de Djidjelli (Nord Constantinois, Algérie).
Paris: Maisonneuve.
Matras, Yaron. 2011. Universals of structural borrowing. In Peter Siemund (ed.),
Linguistic universals and language variation, 204–233. Berlin: De Gruyter
Mouton.
Mion, Giuliano. 2015. Réflexions sur la catégorie des «parlers villageois» en arabe
tunisien. Romano Arabica 15. 269–279.
Mourigh, Khalid. 2015. A grammar of Ghomara Berber. Leiden: University of
Leiden. (Doctoral dissertation).
Myers-Scotton, Carol. 2002. Contact linguistics: Bilingual encounters and grammatical outcomes. Oxford: Oxford University Press.
Naciri-Azzouz, Amina. 2016. Les variétés arabes de Ghomara? s-saḥǝl vs. ǝǧ-ǧbǝl
(la côte vs. la montagne). In George Grigore & Gabriel Biţună (eds.), Arabic
varieties: Far and wide. Proceedings of the 11th International Conference of Aida,
Bucharest 2015, 405–412. Bucharest: Editura Universității din București.
Pereira, Christophe. 2007. Urbanization and dialect change: The Arabic dialect of
Tripoli (Libya). In Catherine Miller, Enam Al-Wer, Dominique Caubet & Janet
C. E. Watson (eds.), Arabic in the city: Studies in dialect contact and language
variation, 77–96. London: Routledge.
Pereira, Christophe. 2011. Arabic in the North Africa region. In Stefan Weninger,
Geoffrey Khan, Michael P. Streck & Janet C. E. Watson (eds.), The Semitic languages: An international handbook, 954–969. Berlin: De Gruyter Mouton.
Procházka, Stephan. 2005. The Turkish contribution to the Arabic lexicon. In Éva
Ágnes Csató, Bo Isaksson & Carina Jahani (eds.), Linguistic convergence and
areal diffusion: Case studies from Iranian, Semitic and Turkic, 191–202. London:
Routledge.
Procházka, Stephan. 2012. Les mots turcs dans l’arabe marocain. In Alexandrine
Barontini, Christophe Pereira, Ángeles Vicente & Karima Ziamari (eds.),
Dynamiques langagières en Arabophonies: Variations, contacts, migrations et
créations artistique. Hommage offert à Dominique Caubet par ses elèves et
collègues, 201–222. Zaragoza: Instituto de Estudios Islámicos y del Oriente
Próximo (IEIOP).
Sayahi, Lotfi. 2014. Diglossia and language contact: Language variation and change
in North Africa. Cambridge: Cambridge University Press.
211
Adam Benkato
Sayahi, Lotfi. 2015. Expression of attributive possession in Tunisian Arabic: The
role of language contact. In Aaron Michael Butts (ed.), Semitic languages in
contact, 333–347. Leiden: Brill.
Souag, Lameen. 2013. Sub-Saharan lexical influence in North African Arabic and
Berber. In Mena Lafkioui (ed.), African Arabic: Approaches to dialectology, 211–
236. Berlin: Mouton de Gruyter.
Souag, Lameen. 2016. Language contact in the Sahara. In Mark Aronoff (ed.),
Oxford research encyclopedia of linguistics. Oxford: Oxford University Press.
Souag, Lameen. 2017. Syllable weight and morphophonologically induced resyllabification in Maghrebi Arabic. In Paul Newman (ed.), Syllable weight in
African languages, 49–67. Amsterdam: John Benjamins.
Souag, Lameen. 2018. Arabic–Berber–Songhay contact and the
grammaticalization of ‘thing’. In Stefano Manfredi & Mauro Tosco (eds.),
Arabic in contact, 54–71. Amsterdam: John Benjamins.
Taine-Cheikh, Catherine. 2008. Arabe(s) et berbère en contact: Le cas
mauritanien. In Mena Lafkioui & Vermondo Brugnatelli (eds.), Berber in
contact: Linguistic and sociolinguistic perspectives, 113–38. Cologne: Rüdiger
Köppe.
Talmoudi, Fathi. 1986. A morphosemantic study of Romance verbs in the Arabic dialects of Tunis, Sūsa, and Sfax. Gothenburg: Acta Universitatis Gothoburgensis.
Thomason, Sarah G. & Terrence Kaufman. 1988. Language contact, creolization
and genetic linguistics. Berkeley: University of California Press.
Türkmen, Erkan. 1988. Arapça’nın Libya lehçesindeki türkçe kelimeler [Turkish
words in the Libyan dialect of Arabic]. Erdem 4–10. 211–225, 227–243.
Van Coetsem, Frans. 1988. Loan phonology and the two transfer types in language
contact. Dordrecht: Foris.
Van Coetsem, Frans. 2000. A general and unified theory of the transmission process
in language contact. Heidelberg: Winter.
Vicente, Ángeles. 2005. Ceuta: Une ville entre deux langues. (Une étude sociolinguistique de sa communauté musulmane). Paris: L’Harmattan.
Winford, Donald. 2005. Contact-induced change: Classification and processes.
Diachronica 22(2). 373–427.
Yoda, Sumikazu. 2005. The Arabic dialect of the Jews of Tripoli (Libya): Grammar,
text and glossary. Wiesbaden: Harrassowitz.
Yoda, Sumikazu. 2013. Judeo-Arabic, Libya, Hebrew component in. In Geoffrey
Khan (ed.), Encyclopedia of Hebrew language and linguistics. Leiden: Brill.
Ziamari, Karima. 2008. Le code switching au Maroc: L’arabe marocain au contact
du français. Paris: L’Harmattan.
212
Chapter 10
Moroccan Arabic
Jeffrey Heath
University of Michigan
Morocco, even if the disputed Western Sahara is excluded, is rivaled only by Yemen
in its variety of Arabic dialects. Latin/Romance sub- and ad-strata have played
crucial roles in this, especially 1. when Arabized Berbers first encountered Romans;
2. during the Muslim and Jewish expulsions from Iberia beginning in 1492; and 3.
during the colonial and post-colonial periods.
1 History and current state
1.1 History
Moroccan Arabic (MA) initially took shape when Arab-led troops, probably Arabized Berbers from the central Maghreb who spoke a contact variety of Arabic,
settled precariously in a triangle of Roman cities/towns consisting of Tangier,
Salé, and Volubilis, starting around 698 AD. Mid-seventh-century tombstones
from Volubilis, inscribed in Latin, confirm that Roman Christians were present,
though in small numbers, when the Arabs arrived. Shortly thereafter, in 710–
711, an Arab-led army from Morocco began the conquest of southern Spain, a
richer and more secure prize that drew away most of the Arab elite. In Morocco,
turnover of the few Arabs and of their Arabized Berber troops was high; they
were massacred or put to flight in the Kharijite revolt of 740. The eighth and
ninth centuries had perfect conditions for the development of a home-grown
Arabic in the Roman triangle in Morocco, and in the emerging Andalus, with a
strong Latinate substratum.
The first true Arab city, Fes, was not founded until approximately 798, a century after the first occupation of Morocco, and its population did not bulk up until immigration from Andalus and the central Maghreb began around 817. With
a cosmopolitan population, and located outside of the old Roman triangle, its
Jeffrey Heath. 2020. Moroccan Arabic. In Christopher Lucas & Stefano Manfredi
(eds.), Arabic and contact-induced change, 213–223. Berlin: Language Science Press.
DOI:10.5281/zenodo.3744519
Jeffrey Heath
Andalusi and non-Andalusi quarters may have maintained their respective dialects for a long time. The remainder of Morocco was occupied by Berber tribes
until much later.
During the eleventh century, the Arabian Bedouin often called Banu Hilāl
entered the central Maghreb in large numbers (cf. Benkato, this volume). They
partially bedouinized the Arabic dialects in Tunisia and Algeria, producing hybrid varieties that combined pre- and post-Hilalian features. They also gradually
pushed their way south and west across the Sahara, bringing their distinctively
Bedouin Arabic, known as Ḥassāniyya, into the southern Maghreb, including
some oases of southern Morocco proper and the entire Western Sahara. Meanwhile, hybridized Algerian dialects, also reflecting a Berber substratum, were
spreading into western Morocco, taking root in new farming villages in the central plains around Fes, and in the younger cities such as Meknes and Marrakesh
(Heath 2002).
In 1492, the Catholic Kings abruptly expelled Spanish Jews from Spain, followed by expulsions through 1614 of Muslims from Spain and Portugal (see also
Vicente, this volume). Jewish deportees, whose predominant home language was
Judeo-Spanish, flooded into the Jewish quarters (mellahs) of Moroccan cities,
constituting a new Jewish elite. Muslim deportees, variably speaking Arabic or
Romance, arrived in several waves and were more easily assimilated. The Jewish
presence in Morocco was strong until 1951, when most Jews left for Israel and
other destinations.
Moroccan ports participated in growing Mediterranean and Atlantic maritime
activity, associated linguistically with Lingua Franca (cf. Nolan, this volume) and
various Romance languages along with Turkish, in the seventeenth and eighteenth centuries. European precolonial penetration into coastal Morocco in the
late nineteenth century later expanded during the French and much smaller Spanish protectorates which lasted from 1912 to 1956. Exposure to French increased
dramatically in this period.
Also of linguistic relevance is the fact that the Moroccan–Algerian border has
been virtually closed for decades, due mainly to political disputes. This has partially sealed off Morocco from the central Maghreb and allowed a specifically
Moroccan koiné to flourish.
1.2 Current situation
Of the 33 million Moroccans recorded in a 2014 census, nearly all are fluent L1 or
(among the Berber-speaking minority) L2 speakers of some form of MA. Moreover, except in the thinly populated Western Sahara, the once-robust dialectal
214
10 Moroccan Arabic
variation within MA has now been greatly compressed. The MA that one is likely
to hear in cafés in Rabat, Fes, Meknes, Marrakesh, Oujda, and even Tangier is the
Moroccan koiné, a hybridized variety mixing pre- and post-Hilalian features and
showing heavy Berber influence in prosody and vocalism.
Many Berber dialects, commonly (but inaccurately) classified into three languages (Tarifiyt, Tamazight, and Tashelhiyt), are still widely spoken in the mountain ranges and in the Souss valley along the Atlantic coast near Agadir. However, these Berber languages are full of Arabic loans, and they are slowly losing
ground to Arabic in all of the cities and large towns.
2 Contact languages
2.1 General
This chapter focuses on contact between MA and European languages. Punic
(Phoenecian) had probably died out locally before the Arab conquest, and Greek
was a non-factor in spite of nominal Byzantine suzerainty after the fall of the
Roman empire. Berber–Arabic contact is covered elsewhere (see Souag, this volume and Benkato, this volume). Diglossic borrowing from literary Arabic would
take us far afield; on this, see Sayahi (2014) and Heath (1989).
The hallmark of abrupt language shift is powerful substratal influence in phonology and prosody. Some calquing of grammatical constructions may occur, but
this can be difficult to tease apart from morphosyntactic simplification. There
may be little or no carryover of core vocabulary and of concrete grammatical
morphemes. The profile of language shift contrasts with that of adstratal borrowing during prolonged bilingualism, whose manifestations are mainly lexical,
and whose complexities involve the morphological and semantic nativization of
foreign-source inflected forms (cf. Manfredi, this volume).
2.2 Late Latin
The best-kept secret about MA is that, unlike the case elsewhere in the Maghreb,
its oldest forms originated by language shift (probably rapid) from Late Latin (LL)
to a contact Arabic spoken by Berber troops.
There are no written records of colloquial LL of the relevant period, either in
North Africa or in Europe, but we can surmise that the LL spoken in the Roman
triangle was intermediate between Classical Latin and early Medieval Romance,
e.g. Medieval Spanish. This implies either five or possibly seven vowel qualities,
phonemic stress, no vowel length, and probably some affricates č [ʧ] and ǧ [ʤ].
215
Jeffrey Heath
2.3 Medieval Judeo-Spanish
The major injection of Medieval Spanish into the Moroccan heartland was the
arrival of expelled Spanish Jews in 1492. They joined existing Jewish communities in the large cities, but a cultural divide between the newcomers (megorashim)
and incumbents (toshavim) quickly emerged. We know from rabbinical responsa
that Judeo-Spanish was still spoken in the central cities for two centuries after 1492 (Chetrit 1985). In far northern Morocco, a form of Arabic- and Hebrewinfluenced Judeo-Spanish called Hakitia or Haketia remained in vernacular use
until the early twentieth century (Benoliel 1977), after which it merged with Modern Spanish.
2.4 Modern French and Spanish
Spanish and to some extent Portuguese and Catalan remained contact influences
chiefly in ports through the late nineteenth century, when direct Spanish involvement in northern Morocco became more significant. Iberian loanwords figure prominently in the early twentieth-century maritime vocabulary provided
by Brunot (1920). During the Protectorates, French became a major language of
education and administration in most of Morocco, especially in the west-to-east
Casablanca–Rabat–Fes–Meknes–Taza corridor, while Spanish consolidated its
position in the far north. French loanwords during the early Protectorate are in
Brunot (1949). MA–French and MA–Spanish bilingualism has increased in the
postcolonial period due to media and mass education. English influence is increasing, mainly through tourism, science education, and finance.
3 Contact-induced changes in MA
3.1 Phonology
MA dialects – archaic Pre-Hilalian, hybridized Post-Hilalian, and in the far south
the unhybridized Ḥassāniyya – differ sharply in vocalic systems, reflecting their
different histories (Heath 2018).
Classical Arabic (CA) had short {ĭ ă ŭ} versus long {ī ā ū}, diphthongs {ăy ăw},
no syncope, and no phonemic stress.
Of the three main types of MA, Ḥassāniyya is closest to CA. It has short vowels
limited to closed syllables: {ə ă} with ə < {*ĭ *ŭ}, in some dialects (e.g. Mali) also
some cases of ŭ. It distinguishes long {ī ā ū} from diphthongs {ăy ăw}, and has
no phonemic stress, but unlike CA it does allow syncope of short vowels (cf.
Taine-Cheikh 1988). Ḥassāniyya shows limited effects of language contact in the
phonology of Berber loanwords (cf. Taine-Cheikh 1997).
216
10 Moroccan Arabic
By contrast, the koiné and some other hybrids reduce all three short vowels
to just one short vowel ə with various allophones, contrasting with full vowels
{i a u}. The hybrid dialects monophthongize {*ăy *ăw} to merge with {i u}. The
rounding of original short *ŭ often survives next to a velar/uvular consonant,
even after syncope (which is productive), suggesting an ongoing feature transfer
that, if and when fully implemented, would result in underlying labiovelars {kʷ gʷ
qʷ ḫʷ ɣʷ } next to ə (which becomes phonetic [ʊ]) or before a consonant. Again
there is no phonemic stress. This is a Berber-like system, reflecting deep longterm substratal/adstratal contact.
A more archaic Berber-like system, still preserving at least the opposition
of short *ĭ ~ *ă versus *ŭ and likely at least some diphthongs, was brought to
Morocco by the early Arabized Berber troops. There it was overlaid on an LL substratum that had five to seven vowel qualities, phonemic stress, no syncope, and
no vowel length. The resulting Pre-Hilalian MA has: three regular vowels {i a u}, a
subset of which (the original short vowels) syncopate in weak metrical positions;
phonemic stress; and a schwa vowel ə confined to posttonic final closed syllables.
The leveling of vowel length distinctions, and the re-splitting of the previously
merged *i ~ *a into i and a based on consonantal environment, were disruptive
to the morphology (see §3.2). Both the leveling, and the new phonemic stress,
were shared with speakers of early Andalusi Arabic, which had a similar LL substratum and whose first invaders came from Morocco. This points to an original
dialect area in the eighth and ninth centuries, including coastal Andalus and at
least the Tangier–Salé axis in Morocco (after Volubilis was abandoned in favor
of Fes), differing significantly from even Pre-Hilalian central Maghrebi dialects,
which likely never had major LL substratal effects.
The differences among MA dialect types can be illustrated by forms of ‘big’
(Table 1). The suggested proto-forms are close to CA but show some adjustments
to short *ĭ and *ŭ. Acute accent marks stress in Pre-Hilalian. Observe especially
that the two homophonous Pre-Hilalian kbír forms behave differently when a
vowel-initial suffix is added. The morphological consequences of length merger
in Pre-Hilalian are considered below. Emphatic /ṛ/ is phonemically distinct from
plain /r/ in all varieties.
Later adstratal borrowings from Spanish and French, as well as from CA, predictably required adjustments to MA phonology. The most disruptive changes
affected French borrowings into MA (our data are best for the hybrid koiné). The
rich array of French vowel qualities had to be squeezed into three MA qualities.
French {i ü e ɛ} merge as MA i. French {u o ɔ œ} merge as MA u. French a becomes
MA a. This compression has had considerable morphological consequences (see
§3.2 below).
217
Jeffrey Heath
Table 1: The word-family ‘big’ in MA dialect types
Gloss
Proto
Pre-Hilalian
Hybrid
Ḥassāniyya
‘big’ (sg.m)
‘big’ (sg.f)
‘he got big’
‘she got big’
‘bigger’
‘big (pl)’
*kăbīr
*kăbīr-a
*kăbĭr
*kăbĭr-at
*ăkbăr
*kŭbār
kbír
kbír-a
kbír
kíbr-ət
kbáṛ
kbaṛ-ín
kbir
kbir-a
kbər
kbr-ət
kbəṛ
kʷbaṛ
kbīr
kbīr-a
kbər
kəbr-ət
(ă)kbăṛ
kbāṛ
The main contribution of Romance to MA consonantism is the affricate č [ʧ].
In the current koiné, this is present as a phoneme (if at all) in the loanword lččina ~
ltšina ‘orange (fruit)’ < Spanish la China, as brought out in the diminutive which
breaks up the čč cluster, hence lčičin ~ ltišin and further variants (Heath 1999).
Archaic northern dialects have more examples of č, and these dialects pronounce
geminated ž as affricate ǧ [ʤ].
3.2 Morphology
Direct borrowing of bound function morphemes is rare in MA as in other languages. A notorious exception is ta-…-t in abstract nouns of profession, from the
Berber feminine singular, likely extrapolated from specific Berber borrowings
like ta-šəffaṛ-t ‘thief’.
Another glaring exception is the set of D-possessives: d (archaic di) before
nouns, dyal- (Pre-Hilalian dyál-) primarily before pronominal suffixes (e.g. dyali ‘mine’, dyal-u ‘his’). The obvious etymology (Latin dē > LL *de or unstressed *di)
presents no phonological or semantic difficulties, but it was rejected by a century
of Maghrebi Arabists, who favored various far-fetched Arabic-internal etymologies. However, an LL source is also indicated by its dialectal distribution: PreHilalian MA, regional colloquial Andalusi Arabic, and certain coastal enclaves
in Algeria that were likely settled by Andalusi merchants. The mysterious prepronominal variant dyál- was generalized from LL *di él(l)u ‘its; his’ and LL *di
él(l)a ‘hers’, which are near-exact matches to the still extant Pre-Hilalian dyál-u
‘his’ and dyál-a ‘hers’. The motivation for this admittedly unusual morphemic
borrowing was the need for a new possessive morpheme as Arabic dialects gradually abandoned the compound-like CA “construct” possessive (Heath 2015). The
fact that possessive morphemes are not immune from borrowing is also shown
218
10 Moroccan Arabic
by possessed forms of certain kin terms, with a Berber nasal suffix, before nominal possessors in hybrid dialects, as in (koiné) ḅḅa-yn ḥamid ‘Hamid’s father’, cf.
ḅḅa ‘father’.
Verbs as well as nouns are readily borrowed from Romance languages into MA.
This raises the question of which Romance inflected form is borrowed, and what
value it is assigned to within the MA aspectual system, which groups 1st/2nd persons versus 3rd person subject splits in the perfect of some verb types. Most Spanish verb borrowings look like Spanish infinitives, e.g. fṛinaṛ ‘to brake’ (< frenar),
but more likely reflect a cluster of forms based on this stem shape in Spanish itself. In addition to the infinitive, this set also includes future frenar-é, conditional
frenar-ía, and forms with d instead of r, namely participle frenado and imperative
plural frenad. Consonant-final borrowed verbs like fṛinaṛ behave like native MA
quadriliteral verbs, and have identical perfect and imperfect forms.
By contrast, French verbs are regularly borrowed as weak (i.e. vowel-final)
verbs, with imperfect and 1st/2nd perfect i, versus 3rd-person perfect a. An example is ‘declare’: imperfect -ḍiklaṛi matching perfect 1st/2nd ḍiklaṛi-, versus 3rd
ḍiklaṛa(-). The likely crosslinguistic bridge is the conspicuous cluster of French
forms ending in orthographic -er (infinitive), -ez (2pl subject), -ais/-ait/-aient (imperfect), and -é(e)(s) (participle). All of these are phonetic [e] or [ɛ] and therefore
merge as MA i, interpretable in MA as the imperfect and 1st/2nd perfect of weak
verbs. The marked 3rd-person perfect with final a is then easily formed by analogy (cf. Lucas & Čéplö, this volume: §4.2 for a parallel development in Maltese).
The merger of vowel length in Pre-Hilalian MA set off a chain reaction of
morphophonological restructurings, most notably in the verbal system. The CA
three-way vocalic opposition of hollow verbs, e.g. for ‘to be’ imperfect kūn-, preconsonantal perfect kŭn-, and prevocalic (or word-final) perfect kān-, is largely
preserved in hybrid and Post-Hilalian dialects. By contrast, in Pre-Hilalian MA,
after the momentous vowel-length merger, the hollow paradigm was reorganized
into a binary opposition of kún (imperfect and 1st/2nd perfect) versus kán (3rd
perfect). This paradigmatic reorganization, which makes no sense semantically
and is apparently unique to Pre-Hilalian MA, then spread analogically to other
verb types, including strong triliterals that have three consonants and no long
vowels, e.g. ‘enter’: imperfect -tḫul matching 1st/2nd perfect tḫul-, but 3rd perfect
tḫal.
3.3 Syntax
Before reaching Morocco, spoken Arabic had prepositions, possessum–possessor,
and def–n–adj order within NPs, preverbal negation (cf. Lucas, this volume)
219
Jeffrey Heath
and complementizers, a perfect/imperfect split in verbs, and pronominal-subject
agreement on verbs (expressed, in part, by suffixes). Romance languages like
Spanish, and presumably eighth-century LL, were already close to this profile,
so opportunities for syntactic influence were limited. Some minor French complementizers are common in educated MA, as in au lieu d’igulu… ‘instead of them
saying’, from French au lieu de ‘instead of’ plus MA igulu ‘they say’.
3.4 Lexicon
While the LL substratum had a profound effect on early MA phonology and
morphophonemics, and also left behind a morphemic souvenir in the form of
D-possessives, not a single basic LL lexical item can be shown to have been preserved in any archaic MA dialect. The most promising candidate for such a retention is dialectal MA qbṭal and variants ‘elbow’. The likely etymon is LL *cubitellu (later LL *kubtɛllu), diminutive of Latin cubitu(s) ‘elbow’, cf. Modern Spanish codillo. The other possibility, less straightforward semantically, is a reflex of
the related adjective, Latin cubitāle, cf. Modern Spanish codal. In Morocco, qbṭal
‘elbow’ survives in several Judeo-Arabic dialects. For Muslims, it was recorded
in an unspecified location in the unpublished fichier of colonial-period linguist
Georges Colin (Iraqui Sinaceur 1993: 1525; de Prémare 1998: 224), and by me in the
1980s in archaic varieties of the Fes–Sefrou area. qbṭal is completely unknown
to the great majority of Moroccan Muslims. Preservation of b shows that qbṭal
is not a recent borrowing from any form based on Modern Spanish codo. The b
was still present in (very) Old Spanish cobdo, its diminutive cobdillo, and cobdal.
“Cubtíll” ‘elbow’ is recorded for late Andalusi Arabic (Corriente 1997: 412; Dozy
1967: 302). The geographic and communal distribution of qbṭal, especially among
Muslims, suggests that it was introduced into Morocco by late Medieval Jewish
refugees.
There are, however, hundreds of well-established Spanish loanwords, especially in northern Morocco. There, Spanish is ubiquitous in schools and broadcast media, Spanish tourists are common, and many Moroccans serve as daylaborers in Spanish enclaves Ceuta and Melilla. While Spanish got a precolonial
head-start, French has long since overtaken it in the rest of Morocco. Of special interest are cases where an original Spanish borrowing was later gallicized,
sometimes only in part. Examples are MA antiris ‘(monetary) interest’, a hybrid
of Spanish interés and French intérêt, and MA gṛabaṭa ‘necktie’ from Spanish
corbata and French cravate. Nonsynonymous mergers also occur, as with gaṛṣun,
attested both as ‘waiter’ (French garçon) and ‘underpants’ (Spanish calzón). ‘To
sign’ is now usually -siɲi/siɲa or -sini/sina (< French signer), but an obsolescent
220
10 Moroccan Arabic
Judeo-Arabic variant siɲaṛ with (pseudo-)Spanish infinitival ending is attested.
Since the Spanish synonym is the unrelated firmar, MA siɲaṛ must have been
formed by applying a borrowing routine “add -aṛ to the stem” to French stems,
probably early in the colonial period when still-abundant Spanish borrowings
were being replaced or hybridized under the influence of the newly dominant
French.
The process is now coming full circle, as English influence expands. The weak
verb alternation of final a/i is productive for verbs borrowed from French, as
noted above (cf. again the close parallels in Maltese; Lucas & Čéplö, this volume: §4.2). A borrowing routine “add final a/i to the stem” extrapolated from
French/MA pairs, is now extended to English, where it has no basis in English
inflectional paradigms. Examples are the comical ka-y-spiki mzyan ‘he speaks
(English) well’, and junkie slang like tt-ṣṭuna ‘he got stoned’ (participle m-ṣṭuni
‘stoned’).
And then there are the many playful translinguistic inventions, concocted
among groups of men sitting in cafés, sipping mint tea or smoking… whatever.
Nearly all such inventions are ephemeral, but a few have caught on (Heath 1987).
Consider the fairly common koiné noun ḫwadri ‘pal, buddy’. Unbeknownst to
those who now use it, it must have arisen via two successive transformations.
First, Spanish padre and madre were playfully combined with the CCaCCi template for denominal occupational derivatives, as though derived from MA ḅḅa ~
bu ‘father’ and MA ṃṃ(ʷ)- ‘mother’. Templatic CCa… is realized as Cwa… when
based on a CV… input, as in ṣwabni ‘seller of soap’ (< ṣabun). Combining CCaCCi
with padre and madre produces the slang terms (attested but rare) ṗwaḍṛi and
ṃwaḍṛi. The final and most ingenious step was to combine the sub-template
Cwadr-i, emergent from these ‘father/mother’ forms, to ḫa- ~ ḫu- ‘brother’, outputting ḫwadri, which then acquires the same ‘buddy’ sense as American English
bro.
4 Conclusion and prospects
The broad outlines of historical language contact in Morocco are becoming reasonably clear. The most urgent need is for more material and analysis of Moroccan Judeo-Arabic (MJA), in forms accessible to international audiences. Ideally
we would want to tease apart the original LL influence on Pre-Hilalian MJA, as
preserved by the toshavim, from the medieval Judeo-Spanish brought to Morocco
in 1492 by the megorashim.
Significant Moroccan Arab and Berber expat communities exist in France, the
Netherlands, Belgium, Germany, Switzerland, and Spain. These vacanciers return
221
Jeffrey Heath
to Morocco in large numbers during summer vacations and on Muslim holy days.
There are opportunities to study them both in Europe (Nortier 1990) and in their
interactions with other Moroccans.
Another promising topic for investigation is a semi-pidginized form of MA
used by monolingual maids in large cities as a kind of foreigner talk to their
expat French employers.
Further reading
) Heath (1989) is a study of lexical and phrasal borrowing/code-switching from
European languages and from Standard Arabic in Moroccan Arabic.
) Nortier (1990) examines language contact phenomena among Moroccans in
the Netherlands.
) Sayahi (2014) is a regional study of Arabic sociolinguistics and language contact from Spain through Morocco to Tunisia.
Acknowledgements
Fieldwork in Morocco was supported by grants from the National Science Foundation (especially BNS 79-04779 in 1979–1981) and by a Fulbright research fellowship in 1986. For support while working on MA material in the 1980s, thanks
also to the National Endowment for the Humanities, the Deutscher Akademischer Austauschdienst, the Alexander von Humboldt Stiftung, and the Hebrew
University of Jerusalem.
Abbreviations
CA
f
L2
LL
Classical Arabic
feminine
second language
Late Latin
MA
m
pl
sg
Moroccan Arabic
masculine
plural
singular
References
Benoliel, José. 1977. Dialecto judeo-hispano-marroquí o hakitia. Madrid: Varona.
Brunot, Louis. 1920. Notes lexicologiques sur le vocabulaire maritime de Rabat &
Salé. Paris: Ernest Leroux.
222
10 Moroccan Arabic
Brunot, Louis. 1949. Emprunts dialectaux arabes à la langue française depuis 1912.
Hespéris 36. 347–430.
Chetrit, Joseph. 1985. Judeo-Arabic and Judeo-Spanish in Morocco and their
sociolinguistic interaction. In Joshua Fishman (ed.), Readings in the sociology
of Jewish languages, 261–279. Leiden: Brill.
Corriente, Federico. 1997. A dictionary of Andalusi Arabic. Leiden: Brill.
Dozy, Reinhart. 1967. Supplément aux dictionnaires arabes. 3rd edn. Vol. 2.
Leiden/Paris: Brill & Maisonneuve et Larose.
Heath, Jeffrey. 1987. Hasta la mujerra! and other instances of playful language
mixing in Morocco. Mediterranean Language Review 2. 113–116.
Heath, Jeffrey. 1989. From code-switching to borrowing: A case study of Moroccan
Arabic. London: Kegan Paul.
Heath, Jeffrey. 1999. Sino-Moroccan citrus: Borrowing as a natural scientific experiment. In Lutz Edzard & Mohammed Nekroumi (eds.), Tradition and innovation: Norm and deviation in Arabic and Semitic linguistics, 168–176. Wiesbaden:
Harrassowitz.
Heath, Jeffrey. 2002. Jewish and Muslim dialects of Moroccan Arabic. London:
Routledge.
Heath, Jeffrey. 2015. D-possessives and the origins of Moroccan Arabic.
Diachronica 32(1). 1–33.
Heath, Jeffrey. 2018. Vowel-length merger and its consequences in archaic
Moroccan Arabic. Zeitschrift für Arabische Linguistik 6. 12–43.
Iraqui Sinaceur, Zakia. 1993. Le dictionnaire Colin d’arabe dialectal marocain.
Vol. 6. Rabat: Institut d’Etudes et de Recherches pour l’Arabisation.
Nortier, Jacomine. 1990. Dutch–Moroccan Arabic code switching. Dordrecht: Foris.
de Prémare, André-Louis. 1998. Dictionnaire arabe-français. Vol. 10. Paris:
l’Harmattan.
Sayahi, Lotfi. 2014. Diglossia and language contact: Language variation and change
in North Africa. Cambridge: Cambridge University Press.
Taine-Cheikh, Catherine. 1988. Métathèse, syncope, épenthèse: A propos de la
structure prosodique du hassaniyya. Bulletin de la Société de Linguistique de
Paris 83(1). 213–252.
Taine-Cheikh, Catherine. 1997. Les emprunts au berbère zénaga: Un soussystème vocalique du ḥassāniyya. Matériaux Arabes et Sudarabiques (GELLAS)
8. 93–142.
223
Chapter 11
Andalusi Arabic
Ángeles Vicente
University of Zaragoza
This chapter covers an ancient contact language situation: Andalusi Arabic with
two other languages – the Romance varieties spoken by the local population, and
the Berber varieties brought by different Berber speakers arriving in al-Andalus
during its existence. The situation of bilingualism whereby the Romance language
was sociolinguistically dominant for most of the population over the course of
several centuries resulted in numerous contact-induced changes in all areas of
grammar. In addition, interaction between Arabic-speaking and Berber-speaking
populations constituted a second locus of language contact with consequences for
Andalusi Arabic.
1 Historical development of Andalusi Arabic
A dialect of the Western Neo-Arabic type, Andalusi Arabic is currently a dead
language. It was spoken from the eighth to the seventeenth century in a changing
territory following historical vicissitudes.
Arabic arrived in the Iberian Peninsula in the eighth century with Arabicspeaking tribes coming from different zones at various stages.1 According to historical sources, the number of Muslims initially arriving was small, most of them
probably partially Arabized Berber-speakers from North Africa.2 Over time, the
society of al-Andalus (the name given to the territory in the Iberian Peninsula
1
Historians have long argued for the ethnic variety of the Arabs who invaded the Iberian Peninsula, particularly referring to the presence of Syrian and Yemeni tribes. See Terés Sádaba
(1957), Al-Wasif (1990) and Guichard (1995).
2
Historians agree that it is extremely difficult, if not impossible, to establish what the level of
Arabization of this population was. According to Manzano Moreno (1990: 399), it seems that
linguistic Arabization was not widespread among Andalusi Berbers at least during the eighth
century.
Ángeles Vicente. 2020. Andalusi Arabic. In Christopher Lucas & Stefano Manfredi
(eds.), Arabic and contact-induced change, 225–244. Berlin: Language Science Press.
DOI:10.5281/zenodo.3744521
Ángeles Vicente
under different Muslim–Arab systems of rule for eight centuries) would eventually come to use a distinctive variety of Maghrebi Arabic known as Andalusi
Arabic.3 This variety evolved through dialectal levelling and changes resulting
from contact with other languages present in the zone, and had become a reasonably unified variety by the tenth century. The political success of the Umayyad
dynasty and the establishment of their caliphate in the year 929 CE may have
contributed to language levelling, though dialect variation continued to exist in
the form of diatopical variants from various regions; scholars thus refer to the
existence of an Andalusi “dialect bundle” (e.g. Corriente 1977: 6; 1992a: 446). For
instance, the Granadian variety seems to have been more conservative than dialects spoken in other regions.4 The regional Andalusi variety spoken in Valencia
was the last to disappear with the expulsion of the moriscos (Muslims forced to
convert to Christianity) in the seventeenth century (Barceló & Labarta 2009: 117).
Even though Andalusi Arabic was a vernacular variety, the few extant sources
are always written, and therefore reflect a higher register than that of the language used for daily communication. In fact, hardly any material reflecting the
everyday dialectal level is available, since most of the sources consist of texts
written in Middle Arabic (i.e. a written form intermediate between Classical and
spoken dialectal Arabic; see Lentin 2011). Furthermore, complications arise due
to the use of Arabic script to record dialect variants.5
Consequently, a comprehensive view of all the periods and places where this
language was spoken is lacking. For instance, sources are scarce regarding the
use of the language in the eighth and ninth centuries. As Wasserstein (1991: 3)
puts it: “A linguistic map of Islamic Spain for any period between the middle of
the eighth century and the middle of the thirteenth century would be extremely
difficult to draw.”
Nevertheless, written documents in Andalusi Arabic are available from the
tenth century until the expulsion of the moriscos in the seventeenth century. The
oldest documented and preserved Andalusi text is an early form of zaǧal poetry
dating from 913 CE, illustrated in (1).6
3
Andalusi Arabic features the only common discriminating trait of Maghrebi varieties, that
is, the n- and n-…-u desinences for the first person singular and plural of the imperfect (cf.
Benkato, this volume).
4
According to Corriente (1998a: 56), this is because Granada was relatively isolated from the
Andalusi mainstream, and played a secondary political role, at least initially. An example that
Corriente gives of this conservatism is the retention of strong imāla (raising of originally low
front vowels) found in Granadian Arabic, since this feature was eliminated or reduced in other
Andalusi varieties with written attestation.
5
An overview of sources for the description of Andalusi Arabic can be found in Corriente et al.
(2015: xxiii–xxiv).
6
It consists of a verse by one of the supporters of ʕUmar ibn Ḥafsūn, insulting the caliph ʕAbd
ar-Raḥmān III. It appears in the historical chronicle al-Muqtabis V, by Ibn Ḥayyān.
226
11 Andalusi Arabic
(1)
Tenth-century Andalusi Arabic (Corriente et al. 2015: 237).7
a. labán úmm-u
fi fúmm-u
milk mother-3sg.m in mouth-3sg.m
‘His mother’s milk is in his mouth.’
b. rás ban ḥafṣún fi ḥúkm-u
head Ban Ḥafṣún in power-3sg.m
‘Ban Ḥafṣún’s head is at his disposal.’
The latest attestations of this language consist of private documents written
by moriscos from Valencia from the seventeenth century, in which interesting
instances of Romance dialectalisms and influence of Catalan, the Romance language spoken in the region, Aragonese and Castilian can be observed (Barceló
& Labarta 2009: 119).
Andalusi Arabic continued to be spoken in the Iberian Peninsula after the end
of al-Andalus as a Muslim–Arab state in 1492 CE, as some of the Arabic-speaking
population remained in certain regions up until the seventeenth century, when
the last moriscos were expelled. This language was therefore taken by the migrant
population to various places in North Africa in different periods from the Middle
Ages up to the Modern Era.8
Initially a second language (L2) for most of the population, after a two-century
gestation process (from around the time of the conquest in 711 until the beginning
of the caliphate in 929), Andalusi Arabic gradually became the first language (L1)
of the majority of the population, overtaking the Romance dialect spoken by the
original local population. The main reason for this was the growing social prestige attached to Arabic in an Islamic society, in contrast to the lower social status
of Andalusi Romance, which became the local L2 and eventually disappeared.9
Andalusi Arabic became the dominant language (regardless of religion) thanks
to the political and social situation of al-Andalus. Furthermore, the advent of an
Arabic-speaking population from the east, especially in the Umayyad caliphate
(929–1031), played a major role in the expansion of Arabization. According to
some scholars such as Fierro Bello (2001) and Corriente (2008: 104), al-Andalus
became a society largely monolingual in Andalusi Arabic around the eleventh
7
Acute accents on vowels in transcription of Andalusi Arabic represent stress rather than vowel
length. See §3.1.1 for further details.
8
This is the reason why Andalusi Arabic has played a very important role in the formation of
Moroccan Arabic (cf. Vicente 2010; Heath, this volume).
9
Mixed marriages between Muslims and Christian women constituted a significant factor in
the propagation of Andalusi Arabic amongst Christians until it also became their L1 (Guichard
1989: 82–83; 1995: 456–457; Chalmeta 2003).
227
Ángeles Vicente
century, though communities using other languages did exist, especially in rural
areas (see §2.1 for more details).
The vernacular Arabic variety spoken in al-Andalus even reached the status of
a literary language, appropriating part of the domain of Classical Arabic through
proverbs and a number of stanza-based poetic forms (including some ḫaraǧāt
and the azǧāl). Andalusi Arabic poetry reached the circles of the court and the
palaces of Taifa kings. Such social and cultural prestige reveals the extent to
which Andalusi Arabic had become the dominant language in this society, and it
is for this reason that it is the best-documented vernacular Arabic variety of all
those spoken in the Middle Ages.
Andalusi Arabic does not conform neatly to either the Bedouin or the preHilali sedentary type of dialect in the classification usually applied nowdays to
Maghrebi Arabic dialects (cf. Benkato, this volume). It shares features of both
types of dialects. For instance, in the phonological system, the three interdental
phonemes are the same as those in Old Arabic, as is the case in Bedouin-type
Maghrebi dialects;10 however, /q/ is realized using the voiceless variant [q] as in
sedentary-type dialects, rather than the voiced variant [ɡ], as in Bedouin-type
dialects.11
According to Corriente (1992b: 34), the number of speakers of Andalusi Arabic
was at its largest between the eleventh century – a time when the Andalusi koiné
reached maturity – and the twelfth century.
2 Contact languages
Andalusi Arabic developed in the Iberian Peninsula through the interaction of
various different Arabic dialects along with two contact languages.12 This situ10
In sedentary-type Maghrebi dialects these are typically pronounced as occlusives. The data do
show that the occlusive pronunciation of interdentals was known in Andalusi Arabic, though
it was considered vulgar and was repressed (Corriente et al. 2015: 29).
11
That said, /q/ may have been realized as a voiced [ɡ] in some registers, regions or periods in
Andalusi Arabic (see Corriente et al. 2015: 64).
12
Besides Eastern Neo-Arabic varieties brought by invaders in the eighth century, from which
Andalusi Arabic emerged, this language continued to evolve in interaction with Maghrebi
dialects, particularly with Moroccan Arabic. Owing to this, it is possible to find intra-Arabic
contact-induced language change, for instance in the Andalusi variety of Granada. Some instances of transfer from Moroccan are the verbs šāf ‘to see’ and ǧāb ‘to bring’, and the second
element in the negative ma šāf ši ‘he did not see’ (cf. Corriente 1998a: 57). For example, the
particle lás or lís (a variant of lás with imāla) was the most frequently used negation particle
in Andalusi Arabic, while the ma... ši construction was generally exceptional in older sources,
though not in the work of aš-Šustarī, a Granadian author, due to his travels to North Africa,
228
11 Andalusi Arabic
ation spanned a long period of time, resulting in a significant amount of transfer. This has been analysed by various authors (e.g. Ferrando 1995; 1997; Vicente
2006), and particularly by Corriente (e.g. Corriente 1981; 1992b; 2000; 2002).
The languages with which Andalusi Arabic was in contact were the Romance
varieties spoken by the Andalusi population and the Berber varieties brought by
different Berber speakers arriving in al-Andalus during its existence.
2.1 Andalusi Romance
Andalusi Romance is a dialect bundle originating in the Romance varieties that
were spoken in the Iberian Peninsula when the Islamic invasion occurred in 711,13
and which underwent a particular evolution through interaction with Arabic.
This Ibero-Romance dialect was the L1 of a large proportion of Andalusi society regardless of their religion. It is also the oldest documented variety of IberoRomance: according to Corriente, the language of the ḫaraǧāt (see below) reflects
the Romance dialect bundle used in al-Andalus between the ninth and eleventh
centuries (Corriente 1995; 1997a; 2000).
The language is not well known: only a few written sources are available, transmitted by copyists who may have had limited knowledge of the language. These
sources are written both in Arabic and Latin scripts.
Sources in Arabic script consist of bilingual dictionaries and botanical, agronomical and medical glossaries. These evidence a limited number of Andalusi
Romance loanwords in Andalusi Arabic, constituting less than 5% of the lexicon
according to Corriente (1992b: 142).
Another source in Arabic script are ḫaraǧāt, the final refrains of each stanza
of the muwaššaḥāt, one of the two types of Andalusi strophic poetry. A few of
these refrains were partially written in Andalusi Romance.14 In addition to these
ḫaraǧāt, loanwords of Andalusi Romance origin were also transmitted in the
zaǧal poems of Ibn Quzman.
Latin-script sources also exist, in toponymy, for instance, as well as in loanwords from Andalusi Romance in more northerly Romance dialects, though the
data these contribute need to be treated with caution, since adaptation to other
according to Corriente et al. (2015: 212–215). In addition, Classical Arabic had an influence, especially on the lexicon. The migration of the Bedouin population into North Africa, however,
did not have an influence on the evolution of Andalusi Arabic.
13
These varieties in turn descended from Iberian Vulgar Latin, with substrate influence from
pre-Romance Iberian languages and Visigoth lexical borrowings.
14
Up to 68 ḫaraǧāt in Andalusi Romance have been found (42 in Arabic script and 26 in Hebrew
script) with one or more words in this language (Corriente 1997a: 268–323), all of them dating
from the tenth–eleventh centuries (Corriente 1997a: 343).
229
Ángeles Vicente
Romance dialects blurs features of the source language, making them of limited
use from a linguistic point of view.
Andalusi Romance has been analysed by Corriente (1995; 2000; 2012); who
has compiled lists of lexical borrowings from Andalusi Romance into Andalusi
Arabic in botanical glossaries and in ḫaraǧāt poetry.
In the first centuries of the history of al-Andalus, Andalusi Romance was
the L1 used by the majority of Andalusi society, even by some Muslims, such
as the muwalladūn (converted Muslims), who would learn Arabic as their L2
for self-promotion in society. In time, however, as an Arabic variety became
the dominant language, diastratic differences become noticeable. Thus, Andalusi
Romance was the L1 used by the rural population and lower classes, whereas
the urban Andalusi population underwent more rapid Arabization due to increased exposure to Arabic through mosques, schools, trade, pilgrimages, and
so on. Thus, the inhabitants of cities and, above all, leading members of society
always had Andalusi Arabic as their L1.
No concrete evidence exists as to when monolingualism in Andalusi Arabic
became established. The most commonly accepted date for the disappearance
of Romance as a common means of communication in al-Andalus is the late
twelfth century, under Almoravid rule. This period saw migrations north out of
al-Andalus of the Christian Mozarabs, although most of these were in fact Arabic
speakers, as instances of lexical borrowings from Andalusi Arabic in Romance
languages from the north reveal. Corriente (1997b; 1992b: 443; 2005) suggests that
bilingualism no longer existed by the thirteenth century, and that in the eleventh
and twelfth centuries it was merely vestigial. In contrast, Galmés de Fuentes and
Menéndez Pidal have defended the existence of bilingualism in Andalusi society
up until the thirteenth century (Galmés de Fuentes 1994: 81–88; Menéndez Pidal
& de Fuentes 2001).15
2.2 Berber
The arrival of a Berber-speaking population in al-Andalus occurred in the eighth
and thirteenth centuries, first as auxiliary troops and later as conquerors, though
many of them may have already become Arabic-speaking and used an early form
of North African Arabic as L2 or even as L1 in the case of those arriving later.
Modern historiography (e.g. Manzano Moreno 1990; Guichard 1995; Chalmeta
2003) reveals that a significant number of Berbers played a major role in the
conquest of al-Andalus, a population which grew larger with the later arrival of
15
While some Romance-speaking communities may indeed have lasted up until the thirteenth
century, note that this circumstance does not imply the existence of a wider bilingual Andalusi
society.
230
11 Andalusi Arabic
the Almoravid and Almohad dynasties in the twelfth and thirteenth centuries.
Interaction between Arabic-speaking and Berber-speaking populations on both
sides of the Strait of Gibraltar facilitated lasting language contact.
The role of Berber in the language development of al-Andalus has not been
analysed in depth, however. This is due to data being scarce regarding not only
the state of Berber varieties at the time, but also their impact on Andalusi Arabic and the speed of their disappearance from the language scene in the Iberian
Peninsula. No sources exist written directly in Berber, plus interpretation issues
arise due to the transmission of Berber loanwords in Arabic or Latin script, as
the phonological systems of these languages do not fully coincide.
Berber varieties had no social prestige in al-Andalus, and were associated with
lower registers, a fact which had obvious effects on the direction of transfers in
contact-induced changes. According to scholars such as Chalmeta (2003: 160) and
Guichard (1995) the reason behind this could be the Berbers’ social organization,
who tended to settle in rural zones.
As a result of all of the above, plus the fact that the number of local Romance
speakers was much higher, there is far less transfer into Andalusi Arabic from
Berber than there is from Romance.
These transfers basically consist of lexical borrowings, which are mainly to be
found in Arabic-script botanical glossaries, and have been analysed by various
authors, including: Ferrando (1997),16 Corriente (1981; 1998b; 2002) and Corriente
et al. (2017; 2020).
3 Contact-induced changes in Andalusi Arabic
3.1 Contact with Andalusi Romance
A special feature of the linguistic history of al-Andalus is that, within a few centuries, a situation of bilingualism, whereby the Romance language was the L1 for
most of the population while Andalusi Arabic was L2, was reversed, eventually
leading to a third phase of monolingualism using only Andalusi Arabic.
Transfers from Romance to Andalusi Arabic probably took place during the
first of the bilingualism phases, a situation which, according to Corriente (2005;
2008), must have lasted two hundred years, from the eighth to the tenth century.
It is difficult to diagnose what type of transfer took place in such an ancient
contact situation. When the agents of change used Romance (the source language; SL) as L1 and Andalusi Arabic (the recipient language; RL) as L2, the type
16
This work includes a previously unpublished analysis conducted by G. S. Colin.
231
Ángeles Vicente
of change was imposition, according to the framework of Van Coetsem (1988;
2000). As we have seen, however, this situation would evolve, and the agents of
change would come to have Andalusi Arabic (the RL) as their L1 and Romance
(the SL) as their L2, meaning that transfer in this situation would be classified as
borrowing in Van Coetsem’s framework.
However, in cases such as this where the precise sociolinguistic situation at a
given time is impossible to judge, it is difficult to establish whether the agents
of change had two L1s or one L1 and one L2. Thus, the possibility exists that
the contact-induced language changes taking place are a convergence type of
transfer (in the terms of Lucas 2015).
3.1.1 Phonology
One contact-induced language change from Romance concerned the prosodic
rhythm of Andalusi Arabic. The quantitative rhythm of Old Arabic was replaced
by the intense stress system of early Romance languages in the Iberian Peninsula.17 Thus, while all Old Arabic and Neo-Arabic varieties feature a prosodic
rhythm that distinguishes long and short syllables, Andalusi Arabic is the only
variety where this quantitative rhythm was replaced by a system where there is
no phonemic vowel length (Corriente 1977; 1992a; Corriente et al. 2015: 75–78).
In this case, the agents of change were presumably L1 speakers of Andalusi
Romance, making the transfer a case of imposition on the L2, Andalusi Arabic.
The altered use of the matres lectionis in the Arabic script constitutes graphemic evidence of this change in prosodic rhythm. Thus, in Andalusi sources, the
graphemes which traditionally mark the Old Arabic long vowels are sometimes
used to mark etymologically short vowels, to indicate that these are stressed. For
instance: مقاصmuqāṣ = /muqáṣṣ/ ‘pair of scissors’ (OA miqaṣṣ), أسقوفusqūf =
/usqúf/ ‘bishop’ (OA usquf ), قنفودqunfūd = /qunfúd/ ‘hedgehog’ (OA qunfud).
Moreover, historically long vowels that were not stressed are often represented
without the regular matres lectionis, for instance: فرانfirān = /firán/ ‘mice’, عمʕam
= /ʕam/ ‘year’.
Another instance is the very name al-Andalus, pronounced by its inhabitants
as /alandalús/, a fact known due to the matres lectionis for /ū/ which appears
in the final syllable, indicating that this syllable is stressed: الاندلوسal-andalūs =
/alandalús/.
In addition, lexical borrowings from Andalusi Arabic currently found in IberoRomance languages also attest to this change of prosodic rhythm. For instance,
the Spanish word andaluz (stressed on the last syllable) can only originate in
17
A change which had taken place in Latin about one thousand years earlier. This language
evolved from a quantitative stress system to an intense stress system in some of its daughter
languages. The same process took place later in Andalusi Arabic.
232
11 Andalusi Arabic
the Andalusi word /alandalús/, while the Spanish word azahar ‘orange blossom’
(also stressed on the last syllable) comes from the Andalusi word /azzahár/, rather
than directly from Old Arabic zahr ‘flower’.
The use of matres lectionis in this way was by no means systematic, since less
cultivated scribes inserted or suppressed them arbitrarily; a fact which could be
interpreted as indicative of an incipient evolution towards the loss of the phonological value of stress in Andalusi prosody (Corriente et al. 2015: 76, fn. 213), a
phenomenon that today characterizes Moroccan Arabic, perhaps the last step of
this evolution in Maghrebi Arabic dialects.
In some cases, a graphic gemination of the following consonant instead of the
grapheme of the vocal quantity is an alternative means of indicating a stressed
vowel, for instance: أسقففusquff = /usqúf/ ‘bishop’, ثققةθiqqa = /θíqa/ ‘trust’,
(Corriente et al. 2015: 77).
Andalusi Arabic also features the appearance of three marginal phonemes /p/,
/g/ and /č/ as transferred from Andalusi Romance, which, however, may not
have existed in some Andalusi sub-dialects. Bearing in mind that these phonemes
were incorporated through loanwords (Corriente 1978), we can assume that the
agents of change had Andalusi Arabic as L1 and that therefore this is a borrowing
type of transfer. Examples include: čípp ‘trap’, čiqála ‘cicada’, čírniya ‘blackbird’
(Corriente et al. 2015: 57). As these phonemes exist even in late toponymy it may
be concluded that they were part of the Andalusi phonological system.
Another example of a contact-induced phonological change was the partial
loss of contrastive velarization in some phonemes. As velarization does not exist in Romance languages, we can assume that this was a case of phonological
imposition by L1 Romance speakers on their L2 Andalusi Arabic.
The effects of this change are visible, for instance, in the frequent interchangeability of /s/ and /ṣ/. Recurrent permutations between both realizations exist and
pseudo-corrections are also in evidence. For example: /sūr, ṣūr/ ‘wall’, /nāqūs,
nāqūṣ/ ‘bell’, /qaswa, qaṣwa/ ‘cruelty’. This is not, however, a very common feature and took place only in the early stages of the Arabization process (Corriente
et al. 2015: 82).
The spirantization of occlusives is another example of contact-induced phonological change in Andalusi Arabic, due to imposition from Andalusi Romance.
According to Romanists, this phenomenon was commonly found in Romance
languages since the Latin period.18
18
The spirantization of the occlusives is also a feature of some Arabic varieties spoken in
Morocco, especially, though not exclusively, in the north (Sánchez & Vicente 2012: 235–236).
In this case, the agents of change were Arabic–Berber bilingual speakers who imposed the
phonology of their L1 Berber on their L2 Arabic. This may have also happened in Andalusi
society, though data to corroborate it is insufficient.
233
Ángeles Vicente
For instance, spirantization of /d/ > [ð] can be observed. Authors of Andalusi
Arabic would write 〈( 〉ذð) rather than 〈( 〉دd) for both *d and *ð because they
considered both sounds to be allophones of /d/, particularly in postvocalic position.19 The realization of the /d/ phoneme clearly changed through contact
with Andalusi Romance. This is a widespread feature noted in various authors,
regions, ages and social groups. For instance: جذول/ǧaðwal/ ‘creek’ < ǧadwal,
حفيذ/ḥafīð/ ‘nephew’ < ḥafīd, ألحذ/al-ḥaðð/ ‘Sunday’ < al-ḥadd, سيذي/sīði/ ‘my
lord’ < sīdi. This phenomenon seems to have been more common in lower and
middle registers of Andalusi.
Another example is the spirantized allophone of /b/, [β], which could constitute a borrowing from Romance or Zenati Berber. This may be confirmed by
the use of 〈f〉 to represent /b/ (as in قسفورىqasfūrā < kuzbara ‘coriander’, فشfiš <
baš/biš ‘in order to’), or by confusion between both phonemes: baysāra/faysāra
‘a dish of cooked beans’ (Corriente et al. 2015: 19).
3.1.2 Morphology
A noteworthy contact-induced morphological change concerns the elimination
of a gender distinction in the second person singular of both pronouns and verbs,
as in taqtúl ‘you kill’, tikassár ‘you break’, taḥtarám ‘you respect’, taḫriǧ ‘you
throw’ (Corriente et al. 2015: 154–155).
The addition of Romance suffixes to Arabic words to produce hybrid terms
was another example of morphological transfer. These suffixes are numerous. For
instance, the augmentative suffix -ūn, as in ǧurrún ‘big jar’ < ǧarra ‘jar’, raqadún
‘sleepyhead’ < rāqid ‘asleep’, and the agentive suffix -áyr, as in ǧawabáyr ‘cheeky’
< ǧawāb ‘answer’ (cf. Corriente 1992b: 126–131; Corriente et al. 2015: 230–231).
3.1.3 Syntax
Changes in gender agreement also arguably result from contact-induced change:
ʕáyn ‘eye’, šáms ‘sun’, and nár ‘fire’ are generally feminine in Arabic but were
occasionally treated as masculine in Andalusi Arabic, as their translation equivalents are in Romance. Likewise, má ‘water’ and dwá ‘medicine’ are masculine
in Arabic but were sometimes considered feminine in Andalusi Arabic, again
on a Romance model (Corriente et al. 2015: 232). This was presumably a case of
imposition, where the agents of change were L2 speakers of Andalusi Arabic.
There are cases of concordless determination constructions in qualifying syntagms following the Romance construction, for instance: alʕaqd θānī ‘the second contract’ instead of more typical al-ʕaqd aθ-θānī (Corriente et al. 2015: 186).
19
This spirantization is also realized in other positions, however.
234
11 Andalusi Arabic
These examples come from texts written by bilingual Mozarabs from Toledo;
since they were either dominant in Andalusi Arabic or had both Andalusi Arabic
and Andalusi Romance as L1s, this change must have been either an instance of
borrowing or of convergence.
There are instances of a construction using the analytic genitive with the preposition min ‘of’ as well as innovative uses of li ‘for’. These are found particularly
in late texts with strong influence from Andalusi Romance (cf. Corriente 2012). As
in the previous case, we are dealing here with agents of change who are either
dominant in Andalusi Arabic and thus borrowing from Andalusi Romance, or
this is an instance of convergence brought about by speakers of both languages
as L1s.
(2)
Late Andalusi Arabic (Corriente et al. 2015: 233–234)
a. mudda min ʕām-ayn
period from year-du
‘a two-year period’
b. min ʕām
from year
‘one year old’
c. naḫruǧ
li wild-ī
go_out.impf.1sg to father-obl.1sg
‘I look like my father.’
The examples in (2) are clearly calqued on Romance expressions: un periodo
de dos años, de un año and salgo a mi padre, respectively.
3.1.4 Lexicon
Lexical borrowings from Romance in Andalusi Arabic constitute less than 5%,
according to Corriente (1992b: 142).20
20
The number of lexical borrowings from Andalusi Arabic into Romance languages spoken in
Spain is larger. According to Corriente (2005), its number is close to two thousand, not counting
the lexical derivations and place names included by other authors, who have put the number
at four thousand or even five thousand. Many of the terms in question are nowadays obsolete
(Corriente 2005: 203, fn. 59). We must not forget that these languages had a different social
status during the period of bilingualism, a major element in contact-induced language changes.
In such situations, less prestigious languages always receive a larger number of transfers (cf.
Corriente et al. 2019).
235
Ángeles Vicente
The most common semantic fields are botanical terms of species endemic to
the Iberian Peninsula, as in ūliyā ‘olive’, amindāl ‘almond’, blāṭur ‘water lily’,
bulmuš ‘elm tree’, and zoological terms, as in burrays ‘lamb’, poḫóta ‘whiting’,
buṭrah ‘mule’, ṭābaraš ‘capers’. For more examples, see Corriente et al. (2017).
Other semantic fields are parts of the body, as in imlīq ‘navel’ and muǧǧa ‘breast’,
family relations, as in šuqru ‘father-in-law’, šubrīn ‘nephew’, and household items
and technicalities of various professions, as in šuqūr ‘axe’ and šayra ‘basket’,
(Corriente et al. 2015: 224).
Some words even adapted to the pattern of broken plural in Andalusi Arabic,
for instance š(u)nyūr ‘sir’, pl. šanānīr, though most used the regular plural suffix
-āt.
3.2 Contact with Berber
As with the Arabic–Romance contact situation, lack of information regarding
the sociolinguistic status of Berber speakers in al-Andalus in the relevant period
makes it difficult to classify the relevant changes according to the types of agentivity involved. That said, since we have no reason to think that significant numbers of native Arabic speakers would have acquired Berber languages as L2s, the
changes described here seem most likely to be the result of imposition by L1
Berber speakers.
3.2.1 Phonology
Available data is always from written sources and it is therefore hard to be certain
about the existence of contact-induced phonological changes.
The realization of *k as [ḫ] has been considered a Zenati Berber influence
(Corriente 1981: 7). For instance: aḫθar ‘more’, aḫṭubar ‘October’ (Corriente et al.
2015: 61).
The replacement of /l/ with /r/, as in Tarifit Berber, could be another instance of
transfer from Berber. Thus, the following spellings in documents written in Latin
script could be instances of possible assimilation-induced allophones: Huaraç,
Hurad, Uarat < walad ‘boy’. The late source where these spellings are found, documents written by Valencian moriscos in the second half of the sixteenth century
(Labarta 1987), suggests that this change could have been introduced through contact with the last Berber immigration waves into al-Andalus (thirteenth century).
However, this trait may not have been generalized in the speech of the wider
community, and could merely represent idiolectal variation or even misspelling.
236
11 Andalusi Arabic
3.2.2 Lexicon
While contact-induced changes in Andalusi Arabic from Berber were initially
considered very scarce, more comprehensive analyses of the sources have revealed that changes may not have been so insignificant.21 In fact, the list compiled by Corriente in 1981 contained 15 Berber loanwords in Andalusi Arabic
(1981: 28–29), the list in his dictionary of 1997 listed 62 (Corriente 1997b: 590),
and the compilation made by Ferrando the same year included 82, of which 39
corresponded to an unpublished study by G. S. Colin and 43 were compiled from
proposals made by various other scholars (Ferrando 1997: 133). The most recent
list contains 115 Berber loanwords (Corriente et al. 2017: 1432–1433).
As Ferrando (1997: 140) points out, these borrowings appear mostly in earlier
sources, and their number decreases considerably in later sources. This fact could
be put down to the social and cultural prestige Andalusi Arabic achieved in later
centuries, even contributing to social cohesion and, therefore, linguistic cohesion.
Most lexical transfers must have taken place in the early centuries of the existence of al-Andalus, prior to the arrival of new Berber speakers, the Almoravids
and the Almohads. For obvious geographical reasons it is quite likely that the
Berber-speaking Muslims (already Arabized) who reached the Iberian Peninsula
with the first Muslim troops came from an area in modern northwestern Morocco,
the region known as Jbala. Ghomara and Senhaja are the vernacular Berber varieties from this region. These non-Zenati varieties are different from those spoken
in the Rif (Kossmann 2017). It is therefore probable that Ghomara and Senhaja
Berber were the sources of a good deal of these borrowings, though any attempt
at classifying them is hindered by the lack of detailed phonetic or morphological
data.
Semantically, most of these lexical borrowings correspond to phytonyms and
zoological terms, socio-political symbols and names of weapons, clothing, food,
and household goods. The number of Berber loanwords that were regularly used
by the Andalusi population is not easily determined, as many are names of plants
that probably only occurred in Berber botanical treatises.
The following are some examples from Corriente et al. (2017):22 azarūd ‘sweet
clover’ < azrud/aẓrud, aṭṭifu ‘take him’ < ǝṭṭǝf ‘take’, āwurmī ‘garden street’ <
21
For instance, linguistic analyses of some sources, such as the botanical glossaries written in
al-Andalus, have yielded a large number of Berber loans in Andalusi Arabic (cf. al-Išbīlī 2004;
2007; Corriente 2012).
22
The Berber origin of some of the lexical borrowings from these lists is only probable, not
certain. Due to the characteristics of the sources, written in Arabic or Romance by possibly
non-Berber-speaking scribes, the available information sometimes does not allow us to go
beyond mere working hypotheses. It is also difficult to decide which Berber variety they belong
237
Ángeles Vicente
awurmi/iwurmi, aɣlāl ‘snails’ < aɣlal, tamaɣra ‘banquet’ < tamǝɣra ‘wedding
party’, zuɣzal (with agglutination of the preposition s- ‘with’) ‘half-pike (Berber
weapon)’ < ugzal, tāqra ‘terrine’ < tagra ‘wooden dish to make couscous’, aqrūn
‘pancakes cut into squares and eaten with honey’ < aɣrum ‘bread’.23
Some of these loans present a chronological problem. The problematic items
are those which have an ungeminated /š/ or /q/, phonemes that were transferred
to the Berber varieties through contact with Arabic.24 These would appear, therefore, to be later loans that arrived with the Berber already Arabized or through
Moroccan Arabic, for instance: išir < iššir ‘boy’, finniš ‘mule’ < afǝnniš ‘snubnosed’,25 barqī < abǝrqi ‘slap’.26
Some of these loans do not appear in modern dictionaries of Berber varieties,
such as arɣīs ‘barberry’ < arɣis,27 āðiqal ‘watermelon’ < adigal, maqaqūn ‘stallion’ < amaka.28
In some cases we have loans that come from Vulgar Latin to Andalusi Arabic
via Berber, for instance: fullūs ‘chicken’ < afǝllus (Berber) < pullus (Vulgar Latin),
bāqya ‘large clay dish’ < tabaqit/θabǝqqišθ (Tarifit) ‘great dish of superior quality’29 < bacchia (Vulgar Latin) ‘goblet, water jug’, hirkāsa ‘rustic leather shoe’ <
arkasǝn (Kabyle) or arkas, ahǝrkus (Tarifit) perhaps < calcĕus (Vulgar Latin), tirfās
‘truffles’ < tǝrfas (Berber) < tuferas (Vulgar Latin), zabzīn ‘low-quality couscous’
< zabazin (Berber, with agglutination of the preposition s- ‘with’) < pisellum (Vulgar Latin, diminutive of pisum ‘pea’).30 These transfers are very likely to have
to: Tarifit, Taqbaylit and Tashelhiyt have all been found. Note also that all Arabic items in
this section are rendered as transliterations of their Arabic-script orthography, rather than
transcriptions of their (assumed) phonology.
23
This item exists in Taqbaylit with the meaning ‘unleavened cooked pasta cookie’ (Dallet 1982).
The ending -um becomes -un due to a metanalysis that associates it with the Romance suffix
-on, which is highly productive in Andalusi Romance.
24
I thank Maarten Kossmann for this and other valuable comments on the section of this work
dealing with contact between Andalusi Arabic and Berber varieties.
25
In Moroccan Arabic fǝnnīš/fənnūš (de Prémare 1998: 167).
26
According to de Prémare (1993: 5), the Moroccan Arabic word ābāṛ ǝq ‘slap’ is also a loanword
of Berber origin.
27
The Berber origin of this item has nevertheless been affirmed by Colin and Ferrando, based on
the data provided by Ibn al-Bayṭār (Ferrando 1997: 110–111). It is documented in Moroccan Arabic, ārɣīs ‘barberry’ (de Prémare 1995: 151), and in Spanish it has become alargue and alguese,
and in Portuguese largis (Corriente et al. 2020). A fall into disuse in the SL is perhaps the reason
for its absence from the current dictionaries.
28
The last two lexical borrowings are documented in the Andalusi source kitābu ʕumdati ṭ-ṭabīb,
by Abu l-Ḫayr al-ʔIšbīlī (2004; 2007), a botanist of the eleventh century. However, their Berber
origin is quite doubtful for M. Kossmann (personal communication).
29
See Ibáñez (1949: 272) whose transcription is zabeqqixz.
30
The word exists in Moroccan Arabic as ābāzīn (de Prémare 1993: 5), and in Kabyle Berber as
tabazint (augmentative of abazin).
238
11 Andalusi Arabic
first taken place in North Africa (the northern part of present-day Morocco),
since we know that some variety of Vulgar Latin was in contact there with the
Berber variety of the region before the arrival of Muslim troops (cf. Heath, this
volume). The Berber-speaking Andalusians would have then later transferred
these items to Andalusi Arabic.31
Some of these lexical borrowings have certain characteristics that demonstrate
greater integration than others in Andalusi Arabic:
1. Morphophonemic adaptations.
a) Phonemic adaptation to Arabic (although this may simply be a problem of orthography, since the Arabic script lacks a means of representing the Berber phonemes /g/ and /ẓ/). /g/ is represented as 〈k〉,
〈q〉 or 〈ǧ〉: akzal/aqzal ‘pike (characteristic weapon of the Berbers)’
< agzal;32 āðiqal ‘watermelon’ < adigal; arǧān ‘argan tree’ < argan,
qillīd ‘Berber prince’ < agǝllid, while /ẓ/ is represented as 〈z〉: zawzana
‘mutism’ < aẓiẓun, lazāz ‘werewolf’ < aẓẓaẓ.
b) Elimination of typically Berber morphemes: e.g., the loss of prefix aof masculine nouns: bāzīn ‘a dish of couscous, meat and vegetables’
< abazin, dād abyaḍ ‘white chameleon’ < addad, mizwār ‘manager,
commander’ < amǝzwaru ‘first’, finniš ‘mule’ < afǝnniš ‘snub-nosed’,
mazad ‘Quranic school’ < amzad. Likewise the loss of prefix and suffix t-…-t of feminine nouns: zaɣnaz ‘brooch, buckle’ < tisǝɣnǝst (Tarifit),33 muzūra ‘horse braid’ < tamzurt (Tarifit and Kabyle), sarɣant
‘root of the orpine plant’ < tasǝrɣint, as well as elimination of prefix
t-, as in abɣā ‘wild bramble’ < tabɣa .
2. Another process for the integration of lexical borrowing involves fitting
Berber words to Arabic patterns, as in zawzana ‘mutism’ (with the Arabic
pattern CawCaCa) < aẓiẓun, harkama ‘tripe stew’ (with the Arabic pattern
CaCCaCa) < urkimen, hirkāsa ‘rustic leather shoe’ (with the Arabic pattern
CiCCāCa) < arkasǝn (Kabyle) or arkas, ahǝrkus (Tarifit).
31
A number of these Berber loans have then gone on to reach the Romance languages through
Andalusi Arabic. The most recent list includes forty of these borrowings in Romance languages
(Corriente et al. 2019).
32
Andalusi Arabic seems to have had a diminutive form of this item: tagzalt (modern dictionaries
give the diminutive tagǝzzalt ‘small stick’; Taïfi 1991). This could then be the source of Castilian
tragacete and Portuguese tragazeite ‘dart’ (Corriente et al. 2020).
33
This is a noun of instrument derived from the verb ɣnǝs ‘to tie with a brooch’. Corriente derives
it from asǝgnǝs ‘needle’, see (Corriente et al. 2020), but the phoneme /ɣ/ makes the first option
more likely (M. Kossmann, personal communication).
239
Ángeles Vicente
4 Conclusion
Andalusi Arabic developed in the Iberian Peninsula through intra-Arabic leveling and contact with two other language types: Romance and Berber. This situation spanned a long period of time and resulted in a good deal of contact-induced
change.
Initially the L2 of most of the population, after a two-century gestation process, Andalusi Arabic gradually became the dominant language, overtaking the
Romance dialect spoken by the local population. The main reason was the growing social prestige attached to Arabic in an Islamic society, in contrast to the
lower social status of Andalusi Romance, which first became an L2, before the
bilingual situation eventually disappeared. This contact situation resulted in a
number of contact-induced changes in all areas of grammar, but it is often difficult to diagnose what type of transfer took place in such an ancient contact
situation.
Concerning Berber varieties, modern historiography reveals that the interaction between Arabic-speaking and Berber-speaking populations on both sides of
the Strait of Gibraltar facilitated lasting language contact. The role of Berber in
the language development of al-Andalus, however, has not yet been analysed in
depth. The nature of the available data is such that lexical borrowings are the
only transfers that have been well described at present.
Future research would be particularly desirable with regard to contact-induced
changes in Andalusi Arabic due to the presence of Berber varieties in the Iberian
Peninsula. This should involve collaboration between scholars of Berber and of
Arabic.
Further reading
) Corriente (1997a) provides a linguistic analysis of Andalusian strophic poetry.
) Corriente (2005) offers valuable information concerning the impact of Andalusi
Arabic on Ibero-Romance.
) Corriente et al. (2015) is the most up-to-date book-length description of Andalusi
Arabic grammar. It contains a section dealing with transfer from Romance and
Berber.
) Ferrando (1997) offers an etymological description of some Berber loanwords
in Andalusi Arabic.
) Vicente (2010) details the Andalusi influence on the dialects of northern Morocco.
240
11 Andalusi Arabic
Abbreviations
1, 2, 3
du
impf
L1
L2
m
1st, 2nd, 3rd person
dual
imperfect (prefix conjugation)
first language
second language
masculine
OA
obl
RL
sg
SL
Old Arabic
oblique
recipient language
singular
source language
References
al-Išbīlī, Abū l-Ḫayr. 2004. Kitābu ‘umdati ṭṭabīb fī ma‘rifati nnabāt likulli labīb:
Libro base del médico para el conocimiento de la botánica por todo experto.
Joaquín Bustamante Costa, Federico Corriente & Mohand Tilmatine (eds.).
Vol. 1: Texto árabe. Madrid: Consejo Superior de Investigaciones Científicas.
al-Išbīlī, Abū l-Ḫayr. 2007. Kitābu ‘umdati ṭṭabīb fī ma‘rifati nnabāt likulli
labīb: Libro base del médico para el conocimiento de la botánica por todo experto. Mohand Tilmatine, Federico Corriente & Joaquín Bustamante Costa
(eds.). Vol. 2: Traducción anotada. Madrid: Consejo Superior de Investigaciones
Científicas.
Al-Wasif, Muhammad-Fajri. 1990. La inmigración de árabes yemeníes a alAndalus desde la conquista islámica (92/711) hasta fines del siglo II/VIII.
Anaquel de Estudios Árabes 1. 203–219.
Barceló, Carmen & Ana Labarta (eds.). 2009. Archivos moriscos: Textos árabes de
la minoría islámica valenciana, 1401–1608. Valencia: Universidad de Valencia.
Chalmeta, Pedro. 2003. Invasión e islamización: La sumisión de Hispania y la formación de al-Andalus. Jaén: Universidad de Jaén.
Corriente, Federico. 1977. A grammatical sketch of the Spanish Arabic dialect
bundle. Madrid: Instituto Hispano-Arabe de Cultura.
Corriente, Federico. 1978. Los fonemas /p/, /č/ y /g/ en árabe hispánico. Vox
Romanica 37. 214–218.
Corriente, Federico. 1981. Notas de lexicología hispanoárabe (III y IV). III: Los
romancismos del Vocabulista. IV: Nuevos berberismos del hispano-árabe.
Awrāq 4. 5–30.
Corriente, Federico. 1992a. Linguistic interferences between Arabic and the
Romance languages of the Iberian Peninsula. In Salma Khadra Jayyusi (ed.),
The legacy of Muslim Spain, 443–456. Leiden: Brill.
Corriente, Federico. 1992b. Arabe andalusí y lenguas romances. Madrid: Editorial
MAPFRE.
241
Ángeles Vicente
Corriente, Federico. 1995. El idiolecto romance andalusí reflejado por las xarajāt.
Revista de Filología Española 75. 5–33.
Corriente, Federico. 1997a. Poesía dialectal árabe y romance en Alandalus: Cejeles
y xarajāt de muwaššaḥāt. Madrid: Gredos.
Corriente, Federico. 1997b. A dictionary of Andalusi Arabic. Leiden: Brill.
Corriente, Federico. 1998a. On some features of late Granadian Arabic (mostly
stress). In Jordi Aguadé, Patrice Cressier & Ángeles Vicente (eds.), Peuplement
et arabisation au Maghreb occidental: Dialectologie et histoire, 53–57. Madrid,
Zaragoza: Casa de Velázquez, Universidad de Zaragoza.
Corriente, Federico. 1998b. Le berbère à Al-Andalus. Études et Documents Berbères
15–16. 269–275.
Corriente, Federico. 2000. El romandalusí reflejado por el glosario botánico de
Abulxayr. Estudios de Dialectología Norteafricana y Andalusí 5. 93–241.
Corriente, Federico. 2002. The Berber adstratum of Andalusi Arabic. In Otto
Jastrow, Werner Arnold & Hartmut Bobzin (eds.), “Sprich doch mit deinen
Knechten aramäisch, wir verstehen es!”: 60 Beiträge zur Semitistik: Festschrift
für Otto Jastrow zum 60. Geburtstag, 105–111. Wiesbaden: Harrassowitz.
Corriente, Federico. 2005. El elemento árabe en la historia lingüística peninsular:
Actuación directa e indirecta. Los arabismos en los romances peninsulares
(especialmente en castellano). In Rafael Cano Aguilar (ed.), Historia de la
lengua española, 185–206. Barcelona: Ariel.
Corriente, Federico. 2008. Romania arábica: Tres cuestiones básicas: Arabismos,
“mozárabe” y “jarchas”. Madrid: Editorial Trotta.
Corriente, Federico. 2012. Une première approche des mots occidentaux
(arabes, berbères et romans) dans l’ouvrage botanique d’As-Sarif Al-Idrisi. In
Alexandrine Barontini, Christophe Pereira, Ángeles Vicente & Karima Ziamari
(eds.), Dynamiques langagières en Arabophonies: Variations, contacts, migrations et créations artistique. Hommage offert à Dominique Caubet par ses elèves
et collègues, 57–63. Zaragoza: Instituto de Estudios Islámicos y del Oriente Próximo (IEIOP).
Corriente, Federico, Christophe Pereira & Ángeles Vicente. 2015. Aperçu
grammatical du faisceau dialectal arabe andalou: Perspectives synchroniques,
diachroniques et panchroniques (Encyclopédie Linguistique d’Al-Andalus, vol.
1). Berlin: De Gruyter.
Corriente, Federico, Christophe Pereira & Ángeles Vicente. 2017. Dictionnaire du
faisceau dialectal arabe andalou: Perspectives phraséologiques et étymologiques
(Encyclopédie Linguistique d’Al-Andalus, vol. 2). Berlin: De Gruyter.
242
11 Andalusi Arabic
Corriente, Federico, Christophe Pereira & Ángeles Vicente. 2019. Dictionnaire des
emprunts ibéro-romans: Emprunts à l’arabe et aux langues du Monde Islamique
(Encyclopédie Linguistique d’Al-Andalus, vol. 3). Berlin: De Gruyter.
Corriente, Federico, Christophe Pereira & Ángeles Vicente. 2020. Le substrat
roman et l’adstrat berbère dans le faisceau dialectal andalou (Encyclopédie
Linguistique d’Al-Andalus, vol. 4). Berlin: De Gruyter.
Dallet, Jean-Marie. 1982. Dictionnaire kabyle-français, français-kabyle. Paris:
SELAF.
Ferrando, Ignacio. 1995. Los romancismos de los documentos mozárabes de
Toledo. Anaquel de Estudios Árabes 6. 71–86.
Ferrando, Ignacio. 1997. G. S. Colin y los berberismos del árabe andalusí. Estudios
de Dialectología Norteafricana y Andalusí 2. 105–146.
Fierro Bello, María Isabel. 2001. Al-Andalus: Saberes e intercambios culturales.
Barcelona: Icaria.
Galmés de Fuentes, Álvaro. 1994. Las jarchas mozárabes. Barcelona: Crítica.
Guichard, Pierre. 1989. Otra vez sobre un viejo problema: Orientalismo y
occidentalismo en la civilización de la España Musulmana. In Generalitat
Valenciana (ed.), En torno al 750 aniversario: Antecedentes y consecuencias de
la conquista de Valencia, 73–96. Valencia: Generalitat Valenciana.
Guichard, Pierre. 1995. Al-Andalus: Estructura antropológica de una sociedad
islámica en Occidente. Granada: Servicio de Publicaciones de la Universidad.
Facsimile edition prepared by Antonio Malpica Cuello. Originally published
in 1976 by Barral.
Ibáñez, Esteban. 1949. Diccionario rifeño-español (etimológico). Madrid: Instituto
de Estudios Africanos.
Kossmann, Maarten. 2017. La place du parler des Senhaja de Sraïr dans la
dialectologie berbère. In Ángeles Vicente, Dominique Caubet & Amina NaciriAzzouz (eds.), La région du Nord-Ouest marocain: Parlers et pratiques sociales et
culturelles. Zaragoza: Prensas de la Universidad de Zaragoza.
Labarta, Ana. 1987. La onomástica de los moriscos valencianos. Madrid: Consejo
Superior de Investigaciones Cientificas.
Lentin, Jérôme. 2011. Middle Arabic. In Lutz Edzard & Rudolf de Jong (eds.),
Encyclopedia of Arabic language and linguistics, online edn. Leiden: Brill.
Lucas, Christopher. 2015. Contact-induced language change. In Claire Bowern &
Bethwyn Evans (eds.), The Routledge handbook of historical linguistics, 519–536.
London: Routledge.
Manzano Moreno, Eduardo. 1990. Beréberes de al-Andalus: Los factores de una
evolución histórica. Al-Qantara 11(2). 397–428.
243
Ángeles Vicente
Menéndez Pidal, Ramón & Alvaro Galmés de Fuentes. 2001. Islam y cristiandad:
España entre las dos culturas. Málaga: Universidad de Málaga.
de Prémare, André-Louis. 1993. Dictionnaire arabe-français. Vol. 1. Paris:
l’Harmattan.
de Prémare, André-Louis. 1995. Dictionnaire arabe-français. Vol. 5. Paris:
l’Harmattan.
de Prémare, André-Louis. 1998. Dictionnaire arabe-français. Vol. 10. Paris:
l’Harmattan.
Sánchez, Pablo & Ángeles Vicente. 2012. Les mots turcs dans l’arabe marocain. In Alexandrine Barontini, Christophe Pereira, Ángeles Vicente & Karima
Ziamari (eds.), Dynamiques langagières en Arabophonies: Variations, contacts,
migrations et créations artistique. Hommage offert à Dominique Caubet par ses
elèves et collègues, 223–252. Zaragoza: Instituto de Estudios Islámicos y del
Oriente Próximo (IEIOP).
Taïfi, Miloud. 1991. Dictionnaire tamazight-français (Parlers du Maroc central).
Paris: L’Harmattan.
Terés Sádaba, Elías. 1957. Linajes árabes en Al-Andalus según la “Yamhara” de
Ibn Ḥazm. Al-Andalus 22. 55–111.
Van Coetsem, Frans. 1988. Loan phonology and the two transfer types in language
contact. Dordrecht: Foris.
Van Coetsem, Frans. 2000. A general and unified theory of the transmission process
in language contact. Heidelberg: Winter.
Vicente, Ángeles. 2006. El proceso de arabización de Alandalús: Un caso medieval de interaccíon de lenguas. Zaragoza: Instituto de Estudios Islámicos y del
Oriente Próximo (IEIOP).
Vicente, Ángeles. 2010. Andalusi influence on northern Morocco following various centuries of linguistic interference. In Juan Pedro Monferrer-Sala & Nader
Al Jallad (eds.), The Arabic language across the ages, 141–159. Wiesbaden:
Reichert.
Wasserstein, David. 1991. The language situation in al-Andalus. In Studies on the
Muwaššaḥ and the Kharja: Proceedings of the Exeter international colloquium,
1–15. Oxford: Ithaca Press.
244
Chapter 12
Ḥassāniyya Arabic
Catherine Taine-Cheikh
CNRS, LACITO
The area where Ḥassāniyya is spoken, located on the outskirts of the Arab world, is
contiguous with those of several languages that do not belong to the Afro-Asiatic
phylum. However, the greatest influence on the evolution of Ḥassāniyya has been
its contact with Berber and Classical Arabic. Loanwords from those languages are
distinguished by specific features that have enriched and developed the phonological and morphological system of Ḥassāniyya. In other respects, Ḥassāniyya and
Zenaga are currently in a state of either parallel evolution or reciprocal exchanges.
1 Current state and historical development
1.1 Historical development of Ḥassāniyya
The arrival in Morocco of the Banī Maʕqil, travelling companions of the Banī
Hilāl and Banī Sulaym, is dated to the thirteenth century. However, the gradual
shift to the territories further south of one of their branches – that of the Banī
Ḥassān, the origin of the name given to the dialect described here – began closer
to the start of the subsequent century.
At that time, the Sahel region of West Africa was inhabited by different communities: on the one hand there were the “white” nomadic Berber-speaking tribes,
on the other hand, the sedentary “black” communities.
Over the course of the following centuries, particularly during the seventeenth
and eighteenth centuries, the sphere of Zenaga Berber gradually diminished, until it ceased to exist in the 1950s, other than in a few tribes in the southwest of
Mauritania. At the same time, Ḥassāniyya Arabic became the language of the nomads of the west Saharan group, maintaining a remarkable unity (Taine-Cheikh
2016; 2018a). There is virtually no direct documentation of the region’s linguistic
Catherine Taine-Cheikh. 2020. Ḥassāniyya Arabic. In Christopher Lucas & Stefano
Manfredi (eds.), Arabic and contact-induced change, 245–263. Berlin: Language Science Press. DOI:10.5281/zenodo.3744523
Catherine Taine-Cheikh
history during these centuries. This absence of information itself suggests a very
gradual transformation and an extended period of bilingualism.
Despite the lack of documentation of the transfer phenomenon, it seems highly
likely that bilinguals played a very important role in the changes described in this
chapter.
1.2 Current situation of Ḥassāniyya
The presence of significant Ḥassāniyya-speaking communities is recognized in
six countries. With the exception of Senegal and especially of Niger, the regions
occupied by these communities, more or less adjacent, are situated primarily in
Mauritania, in the north, northeast and east of the country.
The greatest number of Ḥassāniyya speakers (approximately 2.8 out of a total
of four million) are found in Mauritania, where they constitute the majority of
the population (approximately 75%). The Ḥassāniyya language tends to fulfil the
role of the lingua franca without, however, having genuine official recognition
beyond, or even equal to, that which it has acquired (often recently) in neighbouring countries.
2 Contact languages
2.1 Contact with other Arabic varieties
The Islamization of the Ḥassāniyya-speaking population took place at an early
date, and Ḥassāniyya has therefore had lengthy exposure to Classical Arabic.
For many centuries this contact remained superficial, however, except among
the Marabout tribes, where proficiency in literary Arabic was quite widespread
and in some cases almost total. The teaching of Islamic sciences in other places
reached quite exceptional levels in certain mḥāð̣əṛ (a type of traditional desert
university).1 In the post-colonial era, the choice of Arabic as official language, and
the widespread Arabization of education, media and services, greatly increased
the Ḥassāniyya-speaking population’s contact with literary Arabic (including in
its Modern Standard form), though perfect fluency was not achieved, even among
the young and educated populations.
Excluding the limited influence of the Egyptian and Lebanese–Syrian dialects
used by the media, the Arabic dialects with which Ḥassāniyya comes into contact
1
These may be referred to as universities both in terms of the standard of teaching and the
length of students’ studies. They were, however, small-scale, local affairs, located either in
nomadic encampments or in ancient caravan cities.
246
12 Ḥassāniyya Arabic
most often today are those of the neighbouring countries (southern Moroccan
and southern Algerian). Most recently Moroccan koiné Arabic has established a
presence in the Western Sahara, since the region came under the administration
of Morocco.
2.2 Contact with Berber languages
Ḥassāniyya has always been in contact with Berber languages. Currently, speakers of Ḥassāniyya are primarily in contact with Tashelhiyt (southern Morocco),
Tuareg (Malian Sahara and the Timbuktu region) and Zenaga (southwest Mauritania). In these areas, some speakers are bilingual in Ḥassāniyya and Berber.
In Mauritania, where Zenaga previously occupied a much larger area, Berber
clearly appears as a substrate.
2.3 Contact with languages of the Sahel
Contacts between Ḥassāniyya speakers and the languages spoken in the Sahel
have varied across regions and over time, but have left few clearly discernible
traces on Ḥassāniyya.
The contact with Soninke is ancient (cf. the toponym Chinguetti < Soninke
sí-n-gèdé ‘horse well’), but the effects are hardly noticeable outside of the old
cities of Mauritania. The contact with Songhay is both very old and still ongoing,
but is limited to the eastern part of the region in which Ḥassāniyya is spoken
(especially the region of Timbuktu).
The influence of Wolof, albeit marginal, has always been more substantial
in southwestern Mauritania, especially among the Awlād Banʸūg of the Rosso
region. It peaked in the years 1950–70, in connection with the immigration to
Senegal of many Moors (e.g. gordʸigen ‘homosexual’, lit. ‘man-woman’). In Mauritania, the influence of Wolof can still be heard in some areas of urban crafts (e.g.
mechanics, electricity), but it is primarily a vehicle for borrowing from French.
Although Pulaar speakers constitute the second-largest linguistic community
of Mauritania, there is very limited contact between Ḥassāniyya and Pulaar, with
the exception of a few bilingual groups (especially among the Harratins) in the
Senegal River valley.
Certain communities (particularly among the Fulani) were traditionally known
for their perfect mastery of Ḥassāniyya. As a result of migration into major cities
and the aggressive Arabization policy led by the authorities, Ḥassāniyya has
gained ground among all the non-Arabic speakers of Mauritania (especially in
the big cities and among younger people), but this has come at the cost of a
sometimes very negative attitude towards the language.
247
Catherine Taine-Cheikh
2.4 Contact with Indo-European Languages
Exposure to French has prevailed in all the countries of the region, the only exception being the Western Sahara, which, from the end of the nineteenth century
until 1975, was under Spanish occupation.
In Mauritania the French occupation came relatively late and was relatively
insignificant. However, the influence of the colonizers’ language continued well
after the country proclaimed its independence in 1960. That said, French has
tended to regress since the end of the twentieth century (especially with the rise
of Standard Arabic, e.g. minəstr has been replaced by wazīr ‘minister’), whilst
exposure to English has become somewhat more significant, at least in the better
educated sections of the population.
3 Contact-induced changes in Ḥassāniyya
3.1 Phonology
3.1.1 Consonants
3.1.1.1 The consonant /ḍ/
As in other Bedouin dialects, /ð̣/ is the normal equivalent of the 〈 〉ضof Classical
Arabic (e.g. ð̣mər ‘to have an empty stomach’ (CA ḍamira) and ð̣ḥak ‘to laugh
(CA ḍaḥika). Nonetheless, /ḍ/ is found in a number of lexemes in Ḥassāniyya.
The form [ḍ] sometimes occurs as a phonetic realization of /d/ simply due to
contact with an emphatic consonant (compare ṣḍam ‘to upset’ and ṣadma ‘annoyance’, CA √ṣdm). However, /ḍ/ generally appears in the lexemes borrowed
from Standard Arabic, either in all words of a root, or in a subset of them, for
example: staḥḍaṛ ‘to be in agony’ and ḥaḍari ‘urbanite’ but ḥð̣aṛ ‘to be present’
and maḥəð̣ṛa ‘Quranic school’. The opposition /ḍ/ vs. /ð̣/ can therefore distinguish a classical meaning from a dialectal meaning: compare staḥḍaṛ to staḥð̣aṛ
‘to remember’.
/ḍ/ is common in the vocabulary of the literate. The less educated speakers
sometimes replace /ḍ/ with /ð̣/ (as in qāð̣i for qāḍi ‘judge’), but the stop realization
is stable in many lexemes, including in loanwords not related to religion, such as
ḍʕīv ‘weak’.
The presence of the same phoneme /ḍ/ in Berber might have facilitated the
preservation of its counterpart in Standard Arabic loans, even though in Zenaga
/ḍ/ is often fricative (intervocalically). Moreover, the /ḍ/ of Berber is normally
devoiced in word-final position in Ḥassāniyya, just as in other Maghrebi dialects,
for example: ṣayvaṭ ‘to say goodbye’, from Berber √fḍ ‘to send’.
248
12 Ḥassāniyya Arabic
3.1.1.2 The consonant /ẓ/
/ẓ/ is one of the two emphatic phonemes of proto-Berber. This emphatic sibilant
sound regularly passes from the source language to the recipient language when
Berber words are used in Ḥassāniyya. For example: aẓẓ ~ āẓẓ ‘wild pearl millet’
(Zenaga īẓi).
However, /ẓ/ is also present in lexemes of a different origin. Among Ḥassāniyya roots also attested in Classical Arabic, *z often becomes /ẓ/ in the environment of /ṛ/ (e.g. ṛāẓ ‘to try’, CA rāza; ṛaẓẓa ‘lightning’, CA rizz; ẓəbṛa ‘anvil’, CA
zubra). Sometimes /ẓ/ appears in lexemes with a pejorative connotation, e.g. ẓṛaṭ
‘fart; lie’ (CA ḍaraṭa), ẓagg ‘make droppings (birds)’ (CA zaqq).
3.1.1.3 The consonant /q/
The normal equivalent of the 〈 〉قof Classical Arabic is the velar stop /g/, as in
other Bedouin dialects (e.g. bagṛa ‘cow’, CA baqara). However, /q/ is in no way
rare.
First of all, /q/ appears, like /ḍ/, in a number of words borrowed from Classical Arabic by the literate: ʕaq əd ‘religious marriage contract’; vassaq ‘to pervert’.
The opposition /g/ vs. /q/ can therefore produce two families of words, such as
qibla ‘Qibla, direction of Mecca’ and gəbla ‘one of the cardinal directions (south,
southwest or west, depending on the region)’. It can also create a distinction between the concrete meaning (with /g/) and the abstract meaning (with /q/): θgāl
‘become heavy’, θqāl ‘become painful’.
Next, /q/ is present in several lexemes of non-Arabic origin, such as bsaq ‘silo’,
mzawṛaq ‘very diluted (of tea)’, (in southwest Mauritania) səṛqəlla ‘Soninke people’, (in Néma) sasundaqa ‘circumcision ceremony’, (in Walata) raqansak ‘decorative pattern’, asanqās ‘pipe plunger’, sayqad ‘shouting in public’, and (in the
southeast) šayqa ‘to move sideways’. These lexemes, often rare and very local in
use, seem to be borrowed mostly from the languages of the Sahel region.2
Finally, /q/ is the outcome of *ɣ in cases of gemination, (/ɣɣ/ > [qq]): compare
raqqad ‘to make porridge’ to rɣīda ‘a variety of porridge’ (CA raɣīda). This correlation, attested in Zenaga and more generally in Berber, can be attributed to
the substrate.
Insofar as the contrast between /ɣ/ and /q/ is poorly established in Berber, the
substrate could also explain the tendency, sometimes observed in the southwest,
to velarize non-classical instances of /q/ (or at least instances not identified as
2
I am currently unable to specify the origin of these terms except that bsaq (attested in Zenaga)
could be of Wolof origin.
249
Catherine Taine-Cheikh
classical): hence ɣandīr ‘candle’ for qandīr < CA qandīl – this is despite the fact
that the shift /ɣ/ > /ʔ/ is very common in Zenaga. However, the influence of
Berber does not explain the systematic shift of /ɣ/ to /q/ throughout the eastern part of the Ḥassāniyya region (including Mali): thus eastern qlab ‘defeat’ for
southwestern ɣlab (CA ɣalaba).3
3.1.1.4 Glottal stop
The glottal stop is one of the phonemes of Zenaga (its presence in the language
is in fact a feature that is unique among Berber varieties), however it is not found
in Ḥassāniyya, with the exception of words borrowed from Standard Arabic, e.g.
tʔabbad ‘to live religiously’, danāʔa ‘baseness’ and taʔḫīr ‘postponement’. Very
rarely the glottal stop is also maintained when it occurs at the end of a word as
in baṛṛaʔ ‘to declare innocent’.
3.1.1.5 Palatalized consonants
There are three palatalized consonants: two dental (/tʸ/ and /dʸ/) and a nasal
/nʸ/. Unlike the phonemes discussed above, these are very rare in Ḥassāniyya,
especially /nʸ/.
The palatalized consonants are also attested in certain neighbouring languages
of the Sahel, as well as in Zenaga (but these are not phonemes of Common
Berber). They are rather infrequent in the Zenaga lexicon, occurring especially in
syntagmatic contexts (-d+y-, -n+y-) and in morphological derivation (formation
of the passive by affixation of a geminate /tʸ/).
In Ḥassāniyya, the palatalized consonants mostly appear in words borrowed
from Zenaga or languages of the Sahel. Interestingly, certain loanwords from
Zenaga are ultimately of Arabic origin and constitute examples of phonological
integration, as in tʸfāɣa, a given name and, in the plural, the name of a tribe
< Zenaga atʸfāɣa ‘marabout’ < CA al-faqīh, and ḫurūdʸ ‘leave (from Quranic
school)’ < Zenaga ḫurūdʸ < CA ḫuruǧ ‘exit’.
One should also note the palatization of /t/ in certain lexemes from particular
semantic domains (such as the two verbs related to fighting tʸbəl ‘to hit hard’ and
kawtʸam ‘boxer’). This may suggest the choice of a palatalized consonant for its
expressive value (and would then be a marginal case of phonosymbolism).
3
The regular passage from /ɣ/ to /q/ is a typical Bedouin trait, related to the voiced realization
(/g/) of *q. It occurs especially in southern Algeria, in various dialects of the Chad–Sudanese
area, and in some Eastern dialects (Cantineau 1960: 72).
250
12 Ḥassāniyya Arabic
3.1.1.6 Labial and labiovelar consonants
The labiovelar consonants /mʷ, bʷ, fʷ, vʷ/ or /ṃ, ḅ, f̣, ṿ/) are common in Ḥassāniyya, as they are in Zenaga. In both cases, they often come in tandem with a
realization [u] of the phoneme /ə/.
This phenomenon may have originally arisen in Zenaga, since the Ḥassāniyya of Mali (where it was most likely in contact with other languages) exhibits
greater preservation of a [u] vowel sound and, at the same time, less pronounced
labiovelarization of consonants.
The Ḥassāniyya of Mali also has a voiceless use of the phoneme /f/, where the
Ḥassāniyya of Mauritania is characterized by the use of /v/ in its place (Heath
2004; an observation that my own studies have confirmed). This phonetic trait
does not come directly from Zenaga (in which /v/ exists but is very rare). However, it could be connected with the preference for voiced phonemes in Berber
generally and in Zenaga in particular.
3.1.2 Syllabic structures
In Ḥassāniyya, Arabic-derived syllabic structures do not contain short vowels in
word-internal open syllables, with the exception of particular cases such as passive participles in mu- (mudagdag ‘broken’) and certain nouns of action (ḥašy >
ḥaši ‘filling’). However, loanwords from literary Arabic and other languages (notably Berber and French) display short vowels quite systematically in this context: abadan ‘never’ and ḥazīn ‘sad’ (from Standard Arabic); tamāt ‘gum’ (from
Zenaga taʔmað); taṃāta ‘tomato’. In fact, it may be noted that, unlike the majority of Berber varieties (particularly in the north), Zenaga has a relatively substantial number of lexical items with short vowels (including ə) in open syllables:
kaṛað̣ ‘three’, tuðuṃaʔn ‘a few drops of rain’ awayan ‘languages’, əgəðih ‘necklace made from plants’.4
Furthermore, a long vowel ā occurs word-finally in loaned nouns which in
Standard Arabic end with -āʔ : vidā/vidāy ‘ransom’. In other cases, underlyingly
long word-final vowels are only pronounced long when non-final in a genitive
construct.
4
It is precisely for this reason that, regarding the loss of the short vowels in open syllables,
I deem the hypothesis of a parallel evolution of syllabic structures in Maghrebi Arabic and
Berber to be more convincing than the frequently held alternative hypothesis of a one-way
influence of the Berber substrate on the Arabic adstrate.
251
Catherine Taine-Cheikh
3.2 Morphology
3.2.1 Nominal morphology
3.2.1.1 Standard forms
Nouns and adjectives borrowed from Standard Arabic may often be identified
by the presence of: a) open syllables with short vowels, e.g. vaḍalāt ‘rest of a
meal’, ɣaḍab ‘anger’, vasād ‘alteration’, ḥtimāl ‘possibility’, b) short vowels /i/
(less frequently /u/) in a closed syllable: miḥṛāb ‘mihrab’, muḥarrir ‘inspector;
editor’.
Some syllables are only attested in loanwords, such as the nominal pattern
CVCC, where the pronunciation of the double coda necessitates the insertion of
a supporting vowel, in which case the dialect takes on the form CCVC: compare
ʕaq əd ‘religious marriage’ with ʕqal ‘wisdom’.
The most characteristic loanword pattern, however, is that of taḥrīr ‘liberation; verification (of an account)’. In Ḥassāniyya the equivalent of the pattern
taCCīC is təCCāC. For the root √ḥrr, this provides a verbal noun for other meanings of the verb ḥaṛṛaṛ: təḥṛāṛ ‘whipping of wool (to untangle it); adding flour to
make dumplings’. As for the form taCaCCuC, the /u/ is sometimes lengthened:
taḥammul ‘obligation’, but tavakkūr ‘contemplation’.
3.2.1.2 Berber affixes
Nouns borrowed from Berber are characterized by the frequent presence of the
vowels /a, ā, i, ī, u, ū/. These are of varying lengths, except that in a word-final
closed syllable they are always long and stressed. Since these vowels appear in all
types of syllables – open and closed – this results in much more varied syllabic
patterns than in nouns of Arabic origin.
These loans are also characterized by the presence of affixes which, in the
source language, are markers of gender and/or number: the prefix a/ā- or i/ī- for
the masculine, to which the prefix t- is also added for the feminine or, more frequently (especially in the singular), a circumfix t-...-t. Compare iggīw ~ īggīw
‘griot’ with the feminine form tiggiwīt ~ tīggīwīt. A suffix in -(ə)n characterizes the plurals of these loanwords which, moreover, differ from the singulars
in terms of their vocalic form: iggāwən ~ īggāwən ‘griots’, feminine tiggawātən
~ tīggawātən. The presence of these affixes generally precludes the presence of
the definite article.
Though these affixes pass from the source language to the target language
along with the stems, the syllabic and vocalic patterns of such loans are often particular to Ḥassāniyya: compare Ḥassāniyya āršān, plural īršyūn ~ īršīwən ‘shallow
pit’ with Zenaga aʔraš, plural aʔraššan (see Taine-Cheikh 1997a).
252
12 Ḥassāniyya Arabic
Ḥassāniyya speakers whose mother tongue is Zenaga have most likely played
a role in the transfer of these affixes and their affixation to nouns of all origins (including those of Arabic origin: a possible example being tasūvra ‘large decorated
leather bag for travelling’, cf. sāvər ‘to travel’). The forms that these speakers use
can also be different from those used by other Ḥassāniyya speakers – especially
if the latter have not been in contact with Berber speakers for a long time.
It is not proven that Berber speakers are the only ones to have created and
imposed these forms which are more Berberized than authentically Berber. However, it may be noted that the gender of nouns borrowed from Berber is generally
well preserved in Ḥassāniyya, even for the feminine nouns losing their final -t,
other than in special cases such as the collective tayšəṭ ‘thorny tree (Balanites
aegyptiaca)’ with a final -ṭ (< Zenaga tayšaḌ for tayšaḍt).5 In fact, this indicates
a deep penetration of the meaning of these affixes and of Berber morphology in
general (up to and including the incompatibility of these affixes with the definite
article).
The borrowing of the formants ən- ‘he of’ and tən- ‘she of’ (quasi-equivalents
of the Arabic-derived bū- and ūṃ(ṃ)-) is fairly widespread, in particular in the
formation of proper nouns. It is also mostly in toponyms and anthroponyms
that the diminutive form with prefix aɣ- and suffix -t is found, e.g. the toponym
Agjoujt (< aɣ-žoʔž-t ‘small ditch’).
3.2.2 Verbal Morphology
3.2.2.1 The derivation of saThe existence of verb forms with the prefix sa- is one of the unique characteristics of Ḥassāniyya (Cohen 1963; Taine-Cheikh 2003). There is nothing, however,
to indicate that the prefix is an ancient Semitic feature that Ḥassāniyya has preserved since its earliest days. Instead, the regular correspondences between the
three series of derived verb forms (causative–factitive vs. reflexive vs. passive)
and the specialization of the morpheme t as a specific marker of reflexivity underlie the creation of causative–factitives with sa-. Neologisms with sa- generally appear when forms with the prefix sta- have a particular meaning: staslaʕ
‘to get worse (an injury)’ – saslaʕ ‘to worsen (injury)’; stabṛak ‘to seek blessings’
– sabṛak ‘to give a blessing’; stagwa ‘to behave as a griot’ – sagwa ‘to make someone a griot’; staqbal ‘to head towards the Qibla’ – saqbal ‘to turn an animal for
slaughter in the direction of the Qibla’.
5
In Zenaga, non-intervocalic geminates are distinguished not by length, but rather by tension,
and it is this that is indicated by the use of uppercase for the final Ḍ.
253
Catherine Taine-Cheikh
Furthermore, the influence of Berber has certainly played a role, since the prefix s(a)- (or one of its variants) very regularly forms the causative–factitive structure in this branch of the Afro-Asiatic language family.
In Zenaga, the most frequent realization of this prefix is with a palato-alveolar
shibilant, but a sibilant realization also occurs, particularly with roots of Arabic
origin. For example: Hass. sādəb (variant of ddəb) – Zen. yassiʔðab ‘to train an
animal (with a saddle)’ < CA √ʔdb (cf. ʔaddaba ‘educate, carefully bring up’);
Hass. sasla – Zen. yassaslah ‘to let a hide soak to give it a consistency similar to
a placenta’ and Hass. stasla – Zen. staslah ‘start to lose fur (of hides left to soak)’
< CA √sly (cf. salā ‘placenta’).
Parallel to these examples where the Berber forms (at least those with the prefix st(a)-) are most likely themselves borrowed, we also find patterns with sa-/šawhich are incontestably of Berber origin: compare Ḥassāniyya niyyər ‘to have a
good sense of direction’, sanyar ‘to show the way’, stanyar ‘to know well how
to orient oneself’ and Tuareg ener ‘to guide’, sener ‘to make guide’. Typically,
however, when Ḥassāniyya borrows causative forms from Berber, it usually integrates the Berber prefix as part of the Ḥassāniyya root, making it the first radical of a quadriliteral root, e.g. Hass. sadba – Tuareg sidou ‘to make s.o. leave in the
afternoon’ and Hass. ssadba (< tsadba) – Tuareg adou ‘to leave in the afternoon’.
The parallelism between Arabic and Berber is not necessarily respected in all
cases, but the forms with initial s-/š- are usually causative or factitive in both
cases. The only exception concerns certain Zenaga verbal forms which have become irregular upon contact with Ḥassāniyya: thus yassəðbah ‘to leave in the
afternoon’ or yišnar ‘to orient oneself’ (a variant of yinar), of which the original
causative value is now carried by a form with a double prefix (ž+š): yažəšnar ‘to
guide’.
3.2.2.2 The Derivation of uThe existence of a passive verbal prefix u- for quadrilateral verbs and derived
forms constitutes another unique feature of Ḥassāniyya. For example: udagdag,
passive of dagdag ‘to break’; uṭabbab, passive of ṭabbab ‘to train (an animal)’;
udāɣa, passive of dāɣa ‘to cheat (in a game)’.
The development of passives with u- was most likely influenced by Classical
Arabic, since here the passives of all verbal measures feature /u/ in the first syllable in both the perfect and the imperfect, e.g. fuʕila, yufʕalu; fuʕʕila, yufaʕʕalu;
and fūʕila, yufāʕalu, the respective passives of faʕala, faʕʕala and fāʕala.
However, influence from Berber cannot be excluded here since, in Zenaga, the
formation of passives with the prefix Tʸ is directly parallel to those of the passives
254
12 Ḥassāniyya Arabic
with u- in Ḥassāniyya. Moreover, this prefix is t(t)u- or t(t)w- in other Berber
varieties (especially those of Morocco) and this could also have had an influence
on the emergence of the prefix u-.
3.3 Syntax
3.3.1 Ḥassāniyya–Zenaga parallelisms
Ḥassāniyya and Zenaga have numerous common features, and this is especially
true in the realm of syntax. In general, the reason for these common traits is that
they both belong to the Afro-Asiatic family and remain conservative in various
respects; for example, in their lack of a discontinuous negative construction.
There are, however, also features of several varieties of both languages documented in Mauritania that represent parallel innovations. Thus, corresponding to the diminutive forms particular to Zenaga, we have in Ḥassāniyya mutatis mutandis a remarkably similar extension to verbs of the diminutive pattern
with infix -ay-, e.g. mayllas, diminutive of mallas ‘to smooth over’ (Taine-Cheikh
2008a: 123–124).
In the case of aspectual–temporal forms, there are frequent parallels, such
as Ḥassāniyya mā tla and Zenaga war yiššiy ‘no longer’, Ḥassāniyya ma-zāl
and Zenaga yaššiy ‘still’, Ḥassāniyya tamm and Zenaga yuktay ‘to continue to’,
Ḥassāniyya ʕgab and Zenaga yaggara ‘to end up doing’. One of the most notable
parallel innovations, however, concerns the future morpheme: Ḥassāniyya lāhi
(invariable participle of an otherwise obsolete verb, but compare ltha ‘to pass
one’s time’) and Zenaga yanhāya (a conjugated verb also meaning ‘to busy oneself with something’, in addition to its future function). In both cases we have
forms related to Classical Arabic lahā ‘to amuse oneself’, with the Zenaga form
apparently being a borrowing. It seems, therefore, that this borrowing preceded
the lāhi of Ḥassāniyya and likely then influenced its adoption as a future tense
marker. Note also that in the Arabic dialect of the Jews of Algiers, lāti is a durative
present tense marker (Cohen 1924: 221; Taine-Cheikh 2004: 224; Taine-Cheikh
2008a: 126–127; Taine-Cheikh 2009: 99).
Ḥassāniyya and Zenaga also display common features with regard to complex phrases. For example, concerning completives, Zenaga differs from other
Berber languages in its highly developed usage of ad ~ að, and in particular in the
grammaticalized usage of this demonstrative as a quotative particle after verbs of
speaking and thinking (Taine-Cheikh 2010a). This may have had an influence on
the usage of the conjunctions an(n)- and ʕan- (the two forms tend to be confused)
in the same function in Ḥassāniyya.
255
Catherine Taine-Cheikh
Finally, regarding the variable appearance of a resumptive pronoun in Ḥassāniyya object relative clauses, if influence from Berber (where a resumptive pronoun is always absent) has played any role here, it has simply been to reinforce
a construction already attested in the earliest Arabic, whereby the resumptive
pronoun is absent if the antecedent is definite, as in (1).
(1)
nṛədd
ʕlī-kum əṛ-ṛwāye lli ṛadd-∅
ʕlī-ya
tell.impf.1sg on-2pl def-story rel tell.prf.3sg.m-∅ on-obl.1sg
muḥammad
Mohammed
‘I am going to tell you the story that Mohammed told me.’
3.3.2 Regional influence of Maghrebi Arabic
The Ḥassāniyya spoken in the south of Morocco is rather heavily influenced by
other Arabic varieties spoken in the region. Even among those who conserve
virtually all the characteristic features of Ḥassāniyya (preservation of interdentals, synthetic genitive construction, absence of the pre-verbal particle kā- or tā-,
absence of discontinuous negation, absence of the indefinite article), particular
features of the Moroccan Arabic koiné appear either occasionally or regularly
among certain speakers. The most common such features are perhaps the genitive particle dyal (Taine-Cheikh 1997b: 98) and the preverbal particle kā (Aguadé
1998: 211, §37; 213, §42).
In the Ḥassāniyya of Mali, usage of a genitive particle remains marginal, although Heath (2004: 162) highlights a few uses of genitive (n)tāʕ in his texts.
3.4 Lexicon
3.4.1 Confirmed loanwords
3.4.1.1 Loanwords from Standard Arabic
Verbs loaned from Standard Arabic are as common as nominal and adjectival
loans. Whatever their category, loans are often distinctive in some way (whether
because of their syllabic structure, the presence of particular phonemes or their
morphological template), since the lexeme usually (though not always) has the
same form in both the recipient language and the source language. Examples of
loans without any distinctive features are baṛṛaṛ ‘to justify’ and ðahbi ‘golden’.
A certain number of Standard Arabic verbs with the infix -t- or the prefix staare borrowed, but these verbal patterns can be found elsewhere in Ḥassāniyya.
256
12 Ḥassāniyya Arabic
Certain lexical fields exhibit a particularly high degree of loans from Standard
Arabic: anything connected with Islamic studies or abstract concepts (religion,
rights, morality, feelings, etc.) and, more recently, politics, media and modern
material culture. These regularly retain the meaning (or one of the meanings) of
the source-language item.
3.4.1.2 Loanwords from Berber
There are many lexical items that are probable loans from Berber, with a number
of certain cases among them.
Here we may point to several non-Arabic-origin verbs with cognates across
a wide range of Berber languages, such as kṛaṭ ‘to scrape off’ (Zenaga yugṛað̣);
šayð̣að̣ ‘to make a lactating camel adopt an orphaned calf from another mother’
(Zenaga yaṣṣuð̣að̣ ‘to breastfeed’, yuḍḍað̣ ‘to suckle’); santa ‘to begin’ (Zenaga
yassanta ‘to begin’, Tuareg ent ‘to be started, to begin’); gaymar ‘to hunt from a
distance’ (Berber gmər ‘to hunt’).
Other verbs are derived from nouns loaned from Berber. Hence, ɣawba ‘to
restrain a camel, put it in an aɣāba’ (Tuareg aɣaba ‘jaws’). Sometimes there is
both a verb and an adjective stemming from a loaned root, as in gaylal ‘to have
the tail cut’ and agīlāl ‘having a cut tail’ (Tuareg gilel and agilal).
Some loaned Ḥassāniyya nouns are found with the same root (or an equivalent root) in Berber languages other than Zenaga. For example: agayš ‘male bustard’ (Tuareg gayəs); āškəṛ ‘partridge’ (Kabyle tasekkurt in the feminine form);
tayffārət ‘fetlock (camel)’ (Zenaga tiʔffart, Tuareg téffart); azāɣər ‘wooden mat
ceiling between beams’ (Zenaga azaɣri ‘lintel, beam (of a well)’, Tuareg ǝzgər ‘to
cross’, ăzəgər ‘crossbeam’); talawmāyət ‘dew’ (Zenaga tayaṃut, Tuareg tălămut);
(n)tūrža ‘Calotropis procera’ (Zenaga turžah, Tuareg tərza).
Most of the loanwords cited above are attested in Zenaga (sometimes in a more
innovative form than is found in other Berber varieties, such as yaggīyay ‘to have
a cut tail’ where /y/ < *l). However, there are numerous cases where a corresponding Berber item is attested only in Zenaga. In such cases it is difficult to precisely
identify the source language, even if the phonology and/or morphology seems
to indicate a non-Arabic origin.
Loanwords from Berber seem to be particularly common in the lexicon of
fauna, flora, and diseases, as well as in the field of traditional material culture
(objects, culinary traditions, farming practices, etc.; Taine-Cheikh 2010b; 2014).
Unlike the form of the loans, which is often quite divergent from that of the
source items, their semantics tends to remain largely unchanged. However, there
are some exceptions, notably when the verbs have a general meaning in Berber
257
Catherine Taine-Cheikh
(cf. above ‘to breastfeed’ vs. ‘to make a lactating camel adopt an orphaned calf
from another mother’).
3.4.1.3 Loanwords from Sahel languages
Rather few Ḥassāniyya lexical items seem to be borrowed directly from African
languages, and the origin of those that are is rarely known precisely. We may
note, however, in addition to gadʸ ‘dried fish’ (< Wolof) and dʸəngra ‘warehouse’
(< Soninke), a few terms which appear to be borrowed from Pulaar: tʸəhli ‘roof
on pillars’ and kīri ‘boundary between two fields’.
In some regions we find a concentration of loans in particular domains in relation to specific contact languages. For example, in the ancient town of Tichitt,
we find borrowings from Azer and Soninke (Jacques-Meunié 1961; Monteil 1939;
Diagana 2013): kā ‘house’ (Azer ka(ny), Soninke ká) in kā n laqqe ‘entrance of the
house’; killen ‘path’ (Azer kille, Soninke kìllé); kunyu ~ kenyen ‘cooking’ (Azer
knu ~ kenyu, Soninke kìnŋú).
A significant list of loanwords from Songhay has been compiled by Heath
(2004) in Mali, including e.g.: ṣawṣab (< sosom ~ sosob) ‘pound (millet) in mortar
to remove bran from grains’; daydi ~ dayday (< deydey) ‘daily grocery purchase’;
ākāṛāy (< kaarey) ‘crocodile’; sari (< seri) ‘millet porridge’. Only sari has been
recorded elsewhere in Mauritania (in the eastern town of Walata). On the other
hand, all authors who have done field work on the Ḥassāniyya of Mali (particularly in the region of Timbuktu and the Azawad), have noted loanwords from
Songhay. This is true also of Clauzel (1960) who, as well as a number of Berber
loanwords, gives a small list of Songhay-derived items used in the salt mine of
Tāwdenni, such as titi ‘cylinder of saliferous clay used as a seat by the miners’ (<
tita) and tʸar ‘adze’ (< tʸara).
3.4.1.4 Loanwords from Indo-European languages
The use of loanwords from European languages tends to vary over time. Thus, a
large proportion of the French loanwords borrowed during the colonial period
have more recently gone out of use, such as bəṛṭmāla or qoṛṭmāl ‘wallet’ (< portemonnaie), dabbīš ‘telegram’ (< dépêche ‘dispatch’) or ṣaṛwaṣ ‘to be very close to
the colonizers’ (< service ‘service’). This is true not only of items referring to obsolete concepts (such as the currency terms sūvāya ‘sou’ or ftən/vəvtən ‘cent’, likely
< fifteen), but also of those referring to still-current concepts which are, however,
now referred to with a term drawn from Standard Arabic (e.g. minəstr ‘minister’,
258
12 Ḥassāniyya Arabic
replaced by wazīr). This does not, however, eliminate the permanence of some
old loanwords such as wata ‘car’ (< voiture) or maṛṣa ‘market’ (< marché).6
Although not unique to Ḥassāniyya, the frequency of the emphatic phonemes
(especially /ṣ/ and /ṭ/) in loans from European languages is notable. Consider, in
addition to the treatment of service, porte-monnaie and marché as noted above,
that of baṭṛūn ‘boss’ (< patron), which gives rise to tbaṭṛan ‘to be(come) a boss’
ṭawn ‘ton’ (< tonne).
3.4.2 More complex cases
3.4.2.1 Wanderwörter
Various Arabic lexical items derive from Latin, Armenian, Turkish, Persian, and
so on. In the case of, for example, the names of calendar months, or of items
such as trousers (sərwāl), these terms are not borrowed directly from the source
language by Ḥassāniyya and are found elsewhere (e.g. balbūẓa ‘eyeball’ < Latin
bulbus, attested throughout the Maghreb). The history of such items will not be
dealt with here. We can, however, mention the case of some well-attested terms
in Ḥassāniyya that appear to have been borrowed from sub-Saharan Africa.
One such is māṛu ‘rice’, which seems to come from Soninke (máarò), although
it is also attested in Wolof (maalo) and Zenaga (mārih). Another term, which is
just as emblematic, is mbūṛu ‘bread’, whose origin has variously been attributed
to Wolof, Azer, Mandigo, and even English bread.
To these very everyday terms, we may also add ṃutri ‘pearl millet’ and makka
‘maize’, which have the same form both in Ḥassāniyya and in Zenaga. The first
is a loanword from Pulaar (muutiri). The second is attested in many languages
and seems to have come from the placename Mecca.
As for garta ‘peanut’, ḷāḷo ~ ḷaḷu ‘pounded baobab leaves that serve as a condiment’ (synonym of taqya in the southwest of Mauritania) and kəddu ‘spoon’,
these appear to be used just as frequently in Pulaar as they are in Wolof.
3.4.2.2 Berberized items
Despite the absence of any Berber affixes in the loanwords listed in §3.4.2.1, only
kəddu ‘spoon’ is regularly used with the definite article. In this regard, these
loanwords act like words borrowed from Berber, or more generally, those with
Berber affixes.
6
Ould Mohamed Baba (2003) gives an extensive list of loanwords from French and offers a
classification by semantic field.
259
Catherine Taine-Cheikh
It is, in fact, difficult to prove that a noun with this kind of affix is definitely of
Berber origin, since we find nouns of various origins with Berber affixes. Some
of them are loanwords from the languages of the sedentary people of the valley, such as adabāy ‘village of former sedentary slaves (ḥṛāṭīn)’ (< Soninke dèbé
‘village’); iggīw ~ īggīw ‘griot’ (Zenaga iggiwi, borrowed from Wolof geewel or
from Pulaar gawlo). Others are borrowed from French: agāṛāž ‘garage’; təmbīskit ‘biscuit’. Even terms of Arabic origin are Berberized, as is likely the case
with tasūvra ‘large decorated leather bag for travelling’ (cf. sāvər ‘to travel’) or
tāẓəẓmīt ‘asthma’ (cf. CA zaǧma ‘shortness of breath when giving birth’).
3.4.2.3 Reborrowings
Instances of back and forth between two languages – primarily Ḥassāniyya and
Zenaga – seem to be the reason for another type of mixed form, illustrated previously in §3.2.2.1 by the Zenaga verbs yassəðbah ‘to leave in the afternoon’ and
yišnar ‘to orient oneself’.
Ḥassāniyya saɣnan ‘to mix gum with water to make ink’ provides another
example, where this time the points of departure and arrival seem to be from
the Arabic side. In fact, this loanword is a borrowing of Zenaga yassuɣnan ‘to
thicken (ink) by adding gum’, a verb formed from əssaɣan ‘gum’. This noun in
turn appears to be an adaptation of the Arabic samɣa ‘ink’.
In the case of sla ‘placenta’, there is a double round-trip between the two languages, this time without metathesis: after a passage from Arabic to Zenaga (>
əs(s)la), there is return to Ḥassāniyya with the causative verb sasla ‘to soak a
hide’, and a second loan into Zenaga with the reflexive form (yə)stasla ‘to start
to lose fur (of soaked hides)’.
3.4.2.4 Calques
Calques are undoubtedly common, but they are particularly frequent in locutions
such as rəggət əž-žəll ‘susceptibility’ and bū-damʕa ‘rinderpest’ (literally ‘thinness of skin’ and ‘the one with a tear’). These are exact calques of their Zenaga
equivalents taššəddi-n əyim and ən-anḍi (Taine-Cheikh 2008a).
3.4.2.5 Individual variation
Receptivity to loanwords differs from one individual to another. This is natural
when we are dealing with bilingual speakers and this probably explains the special features of the Ḥassāniyya of the Awlād Banʸūg (often bilingual speakers of
Ḥassāniyya and Wolof) or the Ḥassāniyya of Mali (where Arabic speakers often
260
12 Ḥassāniyya Arabic
speak Songhay and sometimes Tamasheq). However, it also depends on the individuals in question in terms of what we might call their “loyalty” to the language,
whether the language is under pressure from Moroccan Arabic koiné in Morocco
(Taine-Cheikh 1997b; Heath 2002; Paciotti 2017), or whether it is imposed as a
lingua franca in Mauritania (Dia 2007).
4 Conclusion
The principal domain affected by contact in Ḥassāniyya is that of the lexicon
(though an assessment in percentage terms is not at present possible). However,
the integration of loanwords – in particular those from Standard Arabic and
Berber – has resulted in a significant enrichment of the phonological system and
of the inventory of nominal patterns. The effects of contact on the verbal morphology and syntax of the dialect are more indirect. The major developments
in Ḥassāniyya seem most likely to instead be a product of internal evolution. In
certain cases, Zenaga has probably had an influence; in others, we rather witness
instances of parallel evolution.
In future, by studying the vehicular Ḥassāniyya of Mauritania and of the border regions (southern Morocco, southern Algeria, Senegal, Niger, and so on) we
will perhaps discover new developments as a result of contacts triggered by the
political and societal changes of the twenty-first century.
Further reading
Links between Ḥassāniyya and other languages are particularly complex at the
level of semantics and lexicon. On these topics, beyond the available Ḥassāniyya
and Zenaga dictionaries (Heath 2004; Taine-Cheikh 1988–1998; 2008b), readers
may consult the available studies of specific fields (Monteil 1952; Taine-Cheikh
2013) or particular templates (Taine-Cheikh 2018b).
Abbreviations
1, 2, 3
CA
def
Hass.
impf
m
1st, 2nd, 3rd person
Classical Arabic
definite
Ḥassāniyya
imperfect
masculine
obl
pl
rel
sg
Zen.
oblique
plural
relativizer
singular
Zenaga
261
Catherine Taine-Cheikh
References
Aguadé, Jordi. 1998. Relatos en hassaniyya recogidos en Mħāmīd (Valle del Dra,
sur de Marruecos). Estudios de Dialectología Norteafricana y Andalusí 3. 203–
215.
Cantineau, Jean. 1960. Etudes de linguistique arabe. Paris: Klincksieck.
Clauzel, Jean. 1960. L’exploitation des salines de Taoudenni. Institut de recherches
sahariennes de l’Université d’Alger: Imprimerie Protat Frères.
Cohen, David. 1963. Le dialecte arabe ḥassānīya de Mauritanie. Paris: Klincksieck.
Cohen, Marcel. 1924. Le système verbal sémitique et l’expression du temps. Paris:
Ernest Leroux.
Dia, Alassane. 2007. Uses and attitudes towards Hassaniyya among Nouakchott’s
Negro-Mauritanian population. In Catherine Miller, Enam Al-Wer, Dominique
Caubet & Janet C. E. Watson (eds.), Arabic in the city: Issues in dialect contact
and language variation, 325–344. London: Routledge.
Diagana, Ousmane Moussa. 2013. Dictionnaire soninké-français (Mauritanie).
Paris: Karthala.
Heath, Jeffrey. 2002. Jewish and Muslim dialects of Moroccan Arabic. London:
Routledge.
Heath, Jeffrey. 2004. Hassaniya Arabic (Mali)–English–French dictionary.
Wiesbaden: Harrassowitz.
Jacques-Meunié, Djinn. 1961. Cités anciennes de Mauritanie. Paris: Klincksieck.
Monteil, Charles. 1939. La langue azer. In Théodore Monod (ed.), Contributions à
l’étude du Sahara occidental, 213–341. Paris: Larose.
Monteil, Vincent-Mansour. 1952. Essai sur le chameau au Sahara Occidental. SaintLouis, Senegal: Centre IFAN–Mauritanie.
Ould Mohamed Baba, Ahmed-Salem. 2003. Emprunts du dialecte ḥassāniyya à
la langue française. In Ignacio Ferrando & Juan José Sanchez Sandoval (eds.),
AIDA 5th conference proceedings, Cádiz, September 2002, 61–74. Cadiz: Universidad de Cádiz, Servicio de Publicaciones.
Paciotti, Luca. 2017. Conflitti linguistici e glottopolitica in Marocco: La situazione
della Ḥassāniyya nella regione di Guelmime-Oued Noun. Naples: Università
degli Studi di Napoli “L’Orientale”. (Master’s dissertation).
Taine-Cheikh, Catherine. 1988–1998. Dictionnaire ḥassāniyya–français. Paris:
Geuthner.
Taine-Cheikh, Catherine. 1997a. Les emprunts au berbère zénaga: Un soussystème vocalique du ḥassāniyya. Matériaux Arabes et Sudarabiques (GELLAS)
8. 93–142.
262
12 Ḥassāniyya Arabic
Taine-Cheikh, Catherine. 1997b. Les hassanophones du Maroc: Entre affirmation
de soi et auto-reniement. Peuples méditerranéens 79. 85–102.
Taine-Cheikh, Catherine. 2003. Les valeurs du préfixe s- en hassaniyya et les
conditions de sa grammaticalisation. In Ignacio Ferrando & Juan José Sanchez
Sandoval (eds.), AIDA 5th conference proceedings, Cádiz, September 2002, 103–
118. Cadiz: Universidad de Cádiz, Servicio de Publicaciones.
Taine-Cheikh, Catherine. 2004. Le(s) futur(s) en arabe: Réflexions pour une typologie. Estudios de Dialectología Norteafricana y Andalusí 8. 215–238.
Taine-Cheikh, Catherine. 2008a. Arabe(s) et berbère en contact: Le cas
mauritanien. In Mena Lafkioui & Vermondo Brugnatelli (eds.), Berber in
contact: Linguistic and sociolinguistic perspectives, 113–38. Cologne: Rüdiger
Köppe.
Taine-Cheikh, Catherine. 2008b. Dictionnaire zénaga–français: Le berbère de
Mauritanie par racines dans une perspective comparative. Cologne: Rüdiger
Köppe.
Taine-Cheikh, Catherine. 2009. Les morphèmes de futur en arabe et en berbère:
Réflexions pour une typologie. Faits de Langue 33. 91–102.
Taine-Cheikh, Catherine. 2010a. The role of the Berber deictic and TAM markers
in dependent clauses in Zenaga. In Isabelle Bril (ed.), Clause linking and clause
hierarchy: Syntax and pragmatics, 355–398. Amsterdam: John Benjamins.
Taine-Cheikh, Catherine. 2010b. Aux origines de la culture matérielle des
nomades de Mauritanie: Réflexions à partir des lexiques arabes et berbères.
The Maghreb Review 35(1–2). 64–88.
Taine-Cheikh, Catherine. 2013. Des ethnies chimériques aux langues fantômes:
L’exemple des Imraguen et des Nemâdi de Mauritanie. In Carole de Féral (ed.),
In and out of Africa: Languages in question. In honour of Robert Nicolaï, vol. 1:
Language contact and epistemological issues, 137–164. Leuven: Peeters.
Taine-Cheikh, Catherine. 2014. Les voies lactées: Le lait dans l’alimentation des
nomades de Mauritanie. Awal – Cahiers d’études berbères 42. 27–50.
Taine-Cheikh, Catherine. 2016. Etudes de linguistique ouest-saharienne. Vol. 1:
Sociolinguistique de l’aire hassanophone. Rabat: Centre des Etudes Sahariennes.
Taine-Cheikh, Catherine. 2018a. Historical and typological approaches to
Mauritanian and West Saharan Arabic. In Clive Holes (ed.), Arabic historical
dialectology: Linguistic and sociolinguistic approaches, 293–315. Oxford: Oxford
University Press.
Taine-Cheikh, Catherine. 2018b. Ḥassāniyya Arabic in contact with Berber: The
case of quadriliteral verbs. In Stefano Manfredi & Mauro Tosco (eds.), Arabic
in contact, 135–159. Amsterdam: John Benjamins.
263
Chapter 13
Maltese
Christopher Lucas
SOAS University of London
Slavomír Čéplö
Institute of Oriental Studies, Slovak Academy of Sciences/IMAFO Abteilung Byzanzforschung, Österreichische Akademie der Wissenschaften
This chapter presents an overview of the most prominent contact-induced developments in the history of Maltese, a language which is genetically a variety of Arabic,
but which has undergone significant changes, largely as a result of lengthy contact
with Sicilian, Italian, and English. We first address the precise affiliation of Maltese
and the nature of the historical and ongoing contact situations, before detailing
relevant developments in the realms of phonology, inflectional and derivational
morphology, syntax, and lexicon.
1 Maltese and Arabic
From a historical point of view, Maltese is a variety of spoken Arabic, albeit one
that has undergone far-reaching changes as a result of sustained and intensive
contact with Italo-Romance varieties, and more recently also with English. This
is a fact about which there is no controversy among contemporary linguists. It
should be noted, however, that a mix of social, cultural, historical, political, and
indeed linguistic factors has led to a situation in which many Maltese people
today view their language as Semitic, but not a type of Arabic. Since we are
concerned here only with the historical perspective, we will not dwell on the
vexed question of whether or not contemporary Maltese should be classified as
an “Arabic dialect”.1 Suffice it to say that the idea, first popularized by de Soldanis
1
Note that Maltese itself has a number of different dialects, one of which – that of the major
towns, and the variety used in media, literature and administration – is referred to as Standard
Maltese. Except where specified, this chapter deals exclusively with the standard variety of
Maltese.
Christopher Lucas & Slavomír Čéplö. 2020. Maltese. In Christopher Lucas & Stefano Manfredi (eds.), Arabic and contact-induced change, 265–302. Berlin: Language
Science Press. DOI:10.5281/zenodo.3744525
Christopher Lucas & Slavomír Čéplö
(1750) and Vassalli (1791), that Maltese is a variety of Phoenician or Punic, has
been shown since at least since Gesenius (1810) and de Sacy (1829) to be entirely
without merit.
Since the Phoenicians and then the Carthaginians occupied Malta for much
of the first millennium BCE, followed by Roman and Byzantine occupation for
much of the first millennium CE, it would seem prima facie likely that elements
of the languages of these occupiers would survive into contemporary Maltese.
Brincat (1995) shows, however, based on the account of al-Ḥimyarī, that Malta
was to all intents and purposes uninhabited in the period between its conquest
by the Arabs in 870 CE and the first concerted efforts at colonization by Arabicspeaking Muslims in 1048–1049 CE. It is for this reason that the Semitic component of Maltese phonology, morphology, syntax and lexicon is Arabic and Arabic
only (see also Grech 1961).
As for the provenance of the Arabic component of contemporary Maltese,
there is no doubt that the most important source is a variety of Maghrebi (Western) Arabic. This is evident from grammatical features such as: the pan-Maghrebi
extension to the singular of the first-person n- prefix of the imperfect verbal
paradigm (see Table 1); the loss of a gender distinction in the second person singular, in pronouns and both perfect and imperfect verbs, as in urban Tunisian
Arabic varieties (Gibson 2011); variable rearticulation of the definite article on
postnominal adjectives in definite noun phrases, as in (1) (cf. Gatt 2018), found
also in Casablanca Arabic (Harrell 2004: 205); and the -il suffix of the numerals
‘eleven’ to ‘nineteen’ in determiner use, as in (2), which also occurs in the Arabic
dialects of Casablanca (Caubet 2011) and Tlemcen (Taine-Cheikh 2011).2
Table 1: First-person imperfect ‘write’ in Eastern and Western Arabic
Eastern
Singular
Plural
(1)
2
Western
Classical Arabic
Baghdad Arabic
Casablanca Arabic
Maltese
ʔaktub
naktub
aktib
niktib
nəktəb
nkətbu
nikteb
niktbu
il-kelb (l-)abjad
def-dog (def)-white
‘the white dog’
(2)
it-tnax-il
appostlu
def-twelve-dep apostle
‘the twelve apostles’
Unless otherwise specified, all numbered examples present data from Maltese. All Maltese
examples in this chapter are rendered using Standard Maltese orthography.
266
13 Maltese
Narrowing matters down further, Zammit’s (2014) study of lexicon shared between Maltese and the Arabic dialect of Sfax offers yet more support (see also
Vanhove 1998) for the geographically unsurprising conclusion that Maltese is
more closely related to the traditional (so-called pre-Hilalian; see Benkato, this
volume) urban Tunisian dialects than to any other extant Arabic variety. This
is not to suggest, however, that the Arabic component of Maltese resembles
these dialects in all respects. Borg (1996) lists a number of areas in which Maltese accords more closely with Levantine Arabic dialects than with those of the
Maghreb. But the social and political history of Malta after the end of direct Arab
rule in 1127 CE is such that most or all of these similarities should be understood
as the failure of Maltese to participate in innovations that later spread through
the mainland Maghrebi varieties, and not as evidence of influence of Eastern
Arabic on the formation of Maltese.
2 Contact with Italo-Romance and English
2.1 Italo-Romance
A comprehensive history of immigration to Malta in the medieval period is yet
to be written (if indeed such a history is possible at all, given the apparently
scarce documentary evidence). It is therefore impossible to give precise details
of the sociolinguistic conditions under which the Arabic variety spoken in Malta
came into contact with varieties of Italo-Romance in the course of the second
millennium. We can, however, sketch the broad outlines of this process, and make
some reasonable inferences.
The Arabic-speaking settlers who colonized Malta in 1048–1049 CE can be
assumed to have come from either Sicily or southern Italy or both (Brincat 1995:
22), but in any case it seems likely that at least some of these came speaking
a variety of Sicilian in addition to Arabic. Even after Malta was brought under
Norman control in 1127 CE by Roger II of Sicily, and went on to be part of the
Kingdom of Sicily, there does not seem to have been a large-scale immigration of
non-Arabic speakers to Malta at any point, a fact which is of course consistent
with the survival of the Maltese language until today. Unsurprisingly from a
geographical and political perspective, what immigration there was appears to
have come overwhelmingly from Sicily and southern Italy, with lesser numbers
coming also from Spain (Ballou 1893: 134, 289; Blouet 1967: 43–46; Fiorini 1986;
Goodwin 2002: 26–32).
Comprising mostly soldiers, craftsmen and churchmen of various types, it
would appear that this immigration was disproportionately male. In addition to
267
Christopher Lucas & Slavomír Čéplö
families in which the only language spoken was Maltese, there must, therefore,
have been significant numbers of families in medieval Malta in which the father
spoke only Sicilian natively and the mother spoke only Maltese natively, with
communication necessarily involving second-language speech by one or both
parents. Children of such families would therefore have been exposed minimally
to native and non-native Maltese speech and native Sicilian speech.
From the perspective of Van Coetsem’s (1988; 2000) framework for understanding contact-induced change, therefore, it seems highly likely that transfer from Sicilian to Maltese occurred both through imposition under sourcelanguage agentivity (by L1 Sicilian speakers) and borrowing under recipientlanguage agentivity (by L1 Maltese speakers).
There is no doubt that, alongside Sicilian, (Tuscan) Italian had an important
place in Maltese life over many centuries, starting at the latest in 1530, when
it became the official language of government under the regime of the Knights
of Malta. But as Comrie & Spagnol (2016: 316) point out, Italian did not gain a
foothold at the expense of Sicilian among bilingual Maltese until the later eighteenth century, and given its social function as a vehicle for government, education and high culture, rather than the native language of a significant proportion
of ordinary Maltese, it is reasonable to say that transfer from Italian will have
been mediated predominantly by borrowing under recipient-language agentivity.
2.2 English
Starting in 1800, when Malta became a protectorate of the British Empire, English gradually began to supplant Italian as the language of government, education and high culture, being joined in that role by the Maltese language itself
only in the last few decades. English is now widely spoken in Malta: according
to 2011 census data (National Statistics Office 2014: 149), 94.6% of the population
of Malta reported speaking Maltese “well” or “average[ly]”, while 82.1% reported
the same for English. English is a native language for only a very small percentage of Maltese residents, however: Sciriha & Vassallo (2006) put the figure at
2%. As with Italian, then, transfer from English to Maltese will overwhelmingly
have occurred through borrowing under recipient-language agentivity. With the
Maltese variety of English, the reverse is true of course: here the transfer from
English to Maltese will have been almost exclusively imposition under sourcelanguage agentivity by native speakers of Maltese, resulting in such hallmark
features of Maltese English as word-final obstruent devoicing (cf. §3.1.1.2 below),
and the use of but in clause-final position (Lucas 2015: 527).
268
13 Maltese
Given that transfer from English was and is restricted to borrowing in Van
Coetsem’s sense, while the more extensive and long-lasting contact with Sicilian
will have involved both borrowing and imposition, it is not surprising that a
picture will emerge in the following sections whereby Italo-Romance dominates
as a source of contact-induced changes across all linguistic domains, with English
playing a much more modest role, largely restricted to lexicon and associated
inflectional morphology.
3 Contact-induced changes
3.1 Phonology
3.1.1 Consonants
3.1.1.1 Additions to the native phonemic inventory
One of the most salient – and uncontroversially contact-induced – innovations
in Maltese phonology is the addition of at least five (arguably seven) consonant
phonemes.3 This came about through the transfer (presumably borrowing) of
Italo-Romance and English lexical items without subsequent adaptation to the
original native inventory (compare, e.g., Maltese pulizija with unadapted initial
[p] and Cairene Arabic bulīs ‘police’). The five uncontroversial additions are /p/,
/v/, /ʦ/, /ʧ/ and /g/ (orthographically: 〈p〉, 〈v〉, 〈z〉, 〈ċ〉 and 〈g〉; see Table 2), as
in evaporazzjoni ‘evaporation’ and granċ ‘crab’. One can also make a case for an
innovative borrowed phoneme /ʣ/. There are no minimal pairs demonstrating
a phonemic distinction between /ʣ/ and /ʦ/ (and both are represented by 〈z〉 in
the orthography), but Borg & Azzopardi-Alexander (1997: 301) point out that /ʣ/
occurs in environments not requiring a voiced obstruent, as in gazzetta /gɐˈʣːɛtːɐ/
‘newspaper’. More marginal is /ʒ/, which Mifsud (2011) and Borg & AzzopardiAlexander (1997: 303) point out can be found in recent loanwords from English,
such as televixin ‘television’ and bex ‘beige’, though whether all speakers voice
the 〈x〉 in these items is uncertain.
Proto-Semitic *g, represented as 〈 〉جin Arabic script, and usually rendered [ʤ]
when Standard Arabic is spoken, is reflected as /ʤ/ (orthographic 〈ġ〉) in Maltese. This appears to be a retention of the original Maghrebi realization of this
phoneme, other Maghrebi varieties having in general deaffricated it to /ʒ/ (cf.
Heath 2002: 136). Unlike some other Maghrebi varieties, however, the Maltese reflex of 〈 〉جdoes not become /g/ before sibilants (cp. Maltese ġewż vs. Casablanca
3
For useful overviews of the phonology of Maltese, see Borg (1997) and Cohen (1966; 1970).
269
Christopher Lucas & Slavomír Čéplö
Affricate
Fricative
Nasal
Trill
Lateral
Approximant
f
v
m
t
d
z
s
ż
n
r
l
w
gea
l
ata
l
Pal
talv
eola
r
Pos
ċ
ġ
x
Lar
yn
p
b
Vel
ar
Plosive
Alv
eola
r
Lab
ial
Table 2: Inventory of consonants. Symbols are Maltese orthography.
k
g
q
ħ
j
gūz ‘walnuts’).4 Similarly, Proto-Semitic *q (on which more below), is never reflected as /g/ (orthographic 〈g〉) in Maltese (cf. Vanhove 1998: 99), meaning that
the presence of /g/ in the Maltese phonemic inventory is certainly due to its
occurence in numerous lexical borrowings. The majority of these are from ItaloRomance (e.g. gwerra ‘war’), but some are from Berber (e.g. gendus ‘calf’ < Berber
agenduz; Naït-Zerrad 2002: 827), suggesting that /g/ as an independent phoneme
has been present in Maltese since the earliest days of Arabic speech on the Maltese islands.5
3.1.1.2 Losses, mergers and shifts
Alongside these additions, the Maltese consonant phoneme inventory has also
witnessed a number of losses and mergers. Clearly it is not possible to establish
with certainty whether or not these changes were due to contact, but various considerations make it reasonable to assume that contact at least accelerated these
changes. For example, the inherited emphatic (pharyngealized/uvularized) consonants – *ṣ, *ṭ and *ð̣ – have all merged with their non-emphatic counterparts,
4
An exception is gżira ‘island’ < Arabic ǧazīra, perhaps to be explained by direct contiguity
with the sibilant.
5
There are also some sporadic examples of /g/ < *k in Arabic roots, e.g. gideb ‘to lie’. See Cohen
(1966: 14–15) for further details.
270
13 Maltese
as in sħab /sħɐːb/ ‘clouds’ < saḥāb, and also ‘companions’ < ʔaṣḥāb. Note in this
connection that among other Arabic varieties, it is only a handful of those most
strongly affected by contact (such as pidgins and creoles, as well as Cypriot Maronite Arabic; see Avram, this volume; Walter, this volume) that have merged the
emphatic consonants in this way. This suggests that non-native acquisition of
Maltese by Italo-Romance speakers precipitated this change (i.e. that it involves
source-language agentivity in Van Coetsem’s 1988; 2000 terms).
In addition to the loss of the emphatic consonants, Maltese has undergone
significant losses and mergers among the velar and laryngeal phonemes.
Perhaps most saliently, an earlier version of what is today Standard Maltese
merged and then lost the voiced uvular/velar fricative *ɣ and the voiced pharyngeal fricative *ʕ. In Maltese’s rather etymologizing orthography, these historic
phonemes are given the digraph symbol 〈għ〉. In general, this symbol either has
no phonetic correlate, as in għajn /ɐɪn/ ‘eye, spring’ and għonq /ɔnʔ/ ‘neck’, or
otherwise corresponds to the lengthening of a vowel in morphological patterns
where the vowel would ordinarily be short, as in the stem I CaCeC verb għamel
/ˈɐːmɛl/ ‘to do’. That the two original phonemes first merged and were then lost
in Standard Maltese can be inferred from the behaviour of 〈għ〉 + 〈h〉 sequences.
These are realised as /ħː/ in roots where 〈għ〉 reflects *ʕ (e.g. semagħ-ha /sɛˈmɐħːɐ/
hear.prf.3sg.m-3sg.f, ‘he heard it’ < samaʕ ‘to hear’), where other Arabic varieties behave similarly (cf. Woidich 2006: 18), but also, unlike other Arabic varieties, in roots where 〈għ〉 reflects *ɣ (e.g. ferragħ-ha /fɛrˈrɐħːɐ/ pour.prf.3sg.m3sg.f, ‘he poured it out’ < farraɣ ‘to empty’). This merger and subsequent loss
did not take place in all varieties of Maltese. To this day, there are apparently
speakers of dialectal Maltese whose speech preserves both *ɣ as a velar fricative,
and *ʕ as a pharyngeal fricative (Klimiuk 2018). The fact that the merger and loss
of these two phonemes is more advanced in the standard language of the major
conurbations and less so in the dialects of more isolated villages suggests that
contact-induced change played an important role here, with non-native speakers of Maltese presumably being the principal agents of change.
Arguably the most interesting set of mergers and losses concerns the voiceless
fricatives, which represent a case of considerable phonemic reorganization despite relatively little change at the phonetic level. The phonemic changes in this
domain are as follows. First, *h, while maintained in the orthography (as 〈h〉), has
merged with /ħ/ in codas (e.g. ikrah /ɪkˈrɐħ/ ‘ugly’) and sporadically in onsets (e.g.
naħaq /ˈnɐħɐʔ/ < nahaq ‘to bray (of donkeys)’), and is otherwise lost altogether
(e.g. hemm /ɛmː/ ‘there’). The Maltese phoneme /ħ/ thus represents the continuation of the voiceless pharyngeal fricative *ḥ, as well as the partial merger of *h.
Moreover, original *ḫ, the voiceless uvular/velar fricative, has also merged with
271
Christopher Lucas & Slavomír Čéplö
/ħ/, as in ħajt ‘thread’ < ḫayṭ, and also ‘wall’ < ḥāyiṭ. Strikingly, however, the single Maltese phoneme /ħ/ exhibits considerable inter- and intra-speaker variation
in its precise realization, such that glottal, pharyngeal, and velar/uvular voiceless fricative realizations may commonly be heard (Borg & Azzopardi-Alexander
1997: 301), and it is in this sense there has been little phonetic change despite the
considerable phonological reorganization.
Like the loss of the emphatic consonants, the loss or merger of *h (as well as
one or more of the pharyngeal and velar/uvular fricatives) is restricted to a handful of Arabic varieties that have been very strongly affected by contact (see, e.g.,
Walter, this volume). As such, these changes too are suggestive of imposition by
non-native speakers lacking these sounds in their native phonemic inventory (as
was the case for speakers of the Romance varieties with which Maltese has had
the most intense contact, cf. Loporcaro 2011: 141–142). On the other hand, the
preservation of the glottal and pharyngeal fricatives as allophones of /ħ/ complicates this picture, such that the role of contact in bringing about these particular
changes must remain uncertain for now.
It is similarly hard to diagnose the causes of the shift of *q to glottal stop (nevertheless written as 〈q〉 in Maltese orthography) and the stopping of the interdental
fricatives *θ and *ð. In both cases, however, we can at least rule out with confidence any suggestion that these are ancient changes that predate the arrival of
Arabic in Malta, or are historically connected to similar realizations in the Arabic
dialects of urban centres in the Maghreb, Egypt, and the Levant. Written records
of earlier Maltese clearly show that a dorsal realization of *q, as well as the interdental fricative realization of *θ and *ð, survived until at least the late eighteenth
century (Avram 2012; 2014). It is at least plausible, therefore, that contact with
Italo-Romance played a role in these changes too, but firm evidence on this point
is so far lacking.
Finally, a well-known feature of contemporary Maltese (and Maltese English)
phonology is the devoicing of word-final obstruents, as in ħadd [ħɐtː] ‘nobody’.
Avram (2017) shows that devoicing gradually diffused across the Maltese lexicon
over the course of about two centuries from the late sixteenth century onwards,
and he makes a strong case that the initial trigger for this development was imposition by native speakers of Sicilian and Italian, since word-final obstruent devoicing has been shown by various studies (e.g. Flege et al. 1995) to be a frequent
feature of the L2 speech of L1 speakers of Romance languages.
272
13 Maltese
3.1.2 Vowels
Maltese has a much richer vowel phoneme inventory than typical Maghrebi Arabic dialects, with, among the monophthongs, five short-vowel qualities /ɪ, ɛ, ɐ, ɔ,
ʊ/ (orthographic 〈i, e, a, o, u〉), and six long-vowel qualities /iː, ɪː, ɛː, ɐː, ɔː, uː/ (orthographic 〈i, ie, e, a, o, u〉), as well as seven distinct diphthongs (with a number
of different orthographies – see Borg & Azzopardi-Alexander 1997: 299 for details): /ɪʊ, ɛɪ, ɛʊ, ɐɪ, ɐʊ, ɔɪ, ɔʊ/. Compare this with the three-vowel-quality system
of Tunis Arabic, which also lacks diphthongs (Gibson 2011).
Since the Italo-Romance languages have vowel systems of a similar richness to
Maltese, one might assume that this proliferation of vowel phonemes is a straightforward case of transfer. This is, in general, not the case, however. The majority
of new phonemic distinctions are at least partially the result of the loss of emphatic consonants and of *ʕ,6 which led to the phonemicization of vowel qualities
that were previously merely allophonic. Note also that the innovative lax close
front long vowel /ɪː/ is apparently an entirely internal development – the outcome of an extreme raising of the front allophone of *ā (so-called imāla), as in
ktieb /ktɪːb/ ‘book’ < kitāb.
Following Krier (1976: 21–22), we can nevertheless point to three innovations
in this domain which do seem to be the direct result of lexical borrowing from
Italo-Romance.
Krier (1976: 21) points out first of all that, of the five short vowels, only four /ɪ,
ɛ, ɐ, ɔ/ appear in all positions in Arabic-derived lexicon. In contrast, /ʊ/ occurs
only in final position in unstressed syllables in this portion of the lexicon, with
the single exception of kull ‘all’. Were it not for the (extensive) Italo-Romance
component of the Maltese lexicon, therefore, we can say that the distinction between [ɔ] and [ʊ] would remain allophonic, as it is in Tunis Arabic. As it is, the
two sounds should probably be considered phonemically distinct in Maltese. Although minimal pairs are hard to find, possible examples include punt ‘point’ vs.
pont ‘bridge’ and lotto ‘lottery’ vs. luttu ‘mourning’.7
Among the long vowels, the presence of /ɛː/ and /ɔː/ phonemes in Maltese is
also largely attributable Italo-Romance loans containing these sounds. Although
/ē/ and /ō/ do occur in certain Tunisian Arabic varieties (Gibson 2011; Herin
& Zammit 2017), these are the result of historical monophthongization of the
original *ay and *aw diphthongs. The Maltese reflexes of these sounds remain
diphthongs, as in sejf /sɛɪf/ ‘sword’ and lewn /lɛʊn/ ‘colour’. Other than in cases
of compensatory lengthening in items where the consonants represented by 〈għ〉
6
7
These latter changes are themselves, however, arguably contact-induced – see §3.1.1.2.
Our thanks to Michael Spagnol for suggesting these examples.
273
Christopher Lucas & Slavomír Čéplö
and 〈h〉 have been lost (see §3.1.1.2), /ɛː/ and /ɔː/ only occur in the non-Arabic
component of the Maltese lexicon, as in żero /ˈzɛːrɔ/ ‘zero’ and froġa /ˈfrɔːʤɐ/
‘omelette’.
To these three contact-induced monophthongal innovations we can add one
new contact-induced diphthong: /ɔɪ/. Mifsud (2011) points out that this occurs
only in non-Arabic lexical items (e.g. vojt /vɔɪt/ ‘empty space’) in Standard Maltese.
In summary, then, the majority of innovative vowel phonemes in Maltese are
not the direct result of transfer, but the three new monophthongal phonemes
whose emergence is (at least partially) contact-induced, combine to create a nearsymmetrical system in which all five short vowel phonemes have a long counterpart.
3.1.3 Intonation
Despite pioneering work by Alexandra Vella (e.g. Vella 1994; 2003; 2009; Grice
et al. 2019), the study of intonation in Maltese, as in most non-Indo-European
languages, remains in its infancy (cf. Hellmuth, this volume). Impressionistically
speaking, the tunes that can be heard in Maltese (and Maltese English) speech
are highly distinctive, and often quite unlike those of the Mediterranean Arabic
dialects. Several studies have demonstrated that intonation patterns are highly
susceptible to transfer in language contact situations, especially through imposition by source-language-dominant speakers (see the studies of Spanish intonation by O’Rourke 2005; Gabriel & Kireva 2014). Interestingly, however, this
appears to be less true for the tunes associated with polar interrogatives, at least
in the varieties of Spanish described by the aforementioned authors, presumably
because of the importance of intonation in establishing interrogative force in
the absence of syntactic cues in this language. What data we have on this issue
for Maltese fits rather neatly into this larger picture. According to Vella (2003),
the intonational patterns of Maltese late-focus declaratives on the one hand, and
wh-interrogatives on the other, pattern with Palermo Sicilian and Tuscan Italian
respectively, while that of Maltese polar interrogatives more closely resembles
counterparts in Arabic dialects.
It seems safe to assume that imposition by native speakers of Italo-Romance
varieties is the primary cause of the similarities in intonation between Maltese
and Italo-Romance, but borrowing by Maltese-dominant bilinguals should not
be ruled out as an additional factor.
274
13 Maltese
3.2 Morphology
3.2.1 Nouns and adjectives
3.2.1.1 Inflection
It has been shown (e.g. Gardani 2012; Seifart 2017) that plural affixes are, with case
affixes, the most widely transferred inflectional morphemes. Maltese conforms
neatly to the general crosslinguistic picture: it has acquired plural morphemes
from Sicilian and English and little in the way of other inflectional morphology
(but see §3.2.2).8
In addition to a rich array of stem-altering (so-called “broken”) plural patterns,
most of which also serve as the plurals of at least some items of Italo-Romance or,
more rarely, English origin (see Spagnol 2011 for details), Maltese has six plural
suffixes: -in, -a, -iet, -ijiet, -i, and -s.9 Of these, -in, and -iet are straightforward
retentions from Arabic (nevertheless extended to numerous non-Arabic items),
-i and -s are straightforward cases of indirect affix borrowing (in the sense of
Seifart 2015), and -a, and -ijiet arguably involve a subtle interplay of internal and
externally-caused developments.
The most recently borrowed plural suffix is the English-derived -s. This occurs
exclusively with bases borrowed from English, and may be considered only partially integrated into monolingual Maltese (to the extent that such a thing exists;
see §2.2), in that it often alternates optionally with -ijiet in items such as kejk
‘cake’ (pl. kejkijiet ~ kejks). There are, however, a number of reasonably frequent
items (e.g. friżer ‘freezer’) which appear never to take a plural suffix other than
-s.
The Sicilian-derived suffix -i can mark the plural of a far higher proportion of
Maltese nouns than can -s, and is demonstrably better integrated into the Maltese
inflectional system. In addition to marking the plural of Sicilian-derived nouns
which also take -i, e.g. xkupa ‘broom’ < Sicilian scupa (pl. scupi), fjakk ‘weak’ <
Sicilian fiaccu (pl. fiacchi), it has also been extended to: Italian-derived nouns,
including those with a plural in -e in Italian, e.g. statwa ‘statue’ < Italian statua
(pl. statue); nouns from other Romance languages, e.g. pitrava ‘beetroot’ < French
betterave with ∅-plural (orthographic -s); English-derived nouns, e.g. jard ‘yard
(unit of distance)’); and even a few Arabic-derived nouns, e.g. saff ‘layer’ < ṣaff
‘row’, samm ‘very hard’ < ʔaṣamm ‘deaf, hard’.
8
One should note also, however, the appearance in a couple of items of a singulative suffix -u,
apparently borrowed from Sicilian. Borg (1994: 57) cites wiżż-u ‘geese-sing’, dud-u ‘wormssing’, and ful-u ‘beans-sing’.
9
There are also one or two examples of zero plurals, e.g. martri ‘martyr(s)’.
275
Christopher Lucas & Slavomír Čéplö
Arabic and Sicilian coincidentally have an identical less frequently used plural (or collective) suffix -a, as in Arabic mārra ‘passers-by’ (singular mārr) and
Sicilian libbra ‘books’ (singular libbru). A plural suffix of this form also occurs
in Maltese, with nouns of both Arabic and Italo-Romance origin (e.g. kittieba
‘writers’ < Arabic kattāb; nutara ‘notaries’ < Italian notaro). Evidence that this
is perceived and treated as a single morpheme rather than two homophonous
items comes from the fact that the restriction of this suffix to groups of people
in Arabic applies also to the Italo-Romance part of the Maltese lexicon (Mifsud
2011).
A curious feature of Maltese plural morphology from a comparative Arabic
perspective is the very frequent suffix -ijiet (-jiet after certain vowel-final stems),
as in postijiet ‘places’ (singular post) and ommijiet ‘mothers’ (singular omm).
While clearly based on the Arabic-derived suffix -iet (< Arabic -āt, with characteristic Maltese imāla), the provenance of the initial -ij- is not obvious. Mifsud
(2011) plausibly suggests that -ijiet as a whole is “derived from the plural of verbal
nouns with a weak final radical, like tiġrijiet ‘races’, tiswijiet ‘repairs’”, but Geary
(2017) makes a strong case that the large influx into Maltese of Italo-Romance
nouns whose singulars ended in -i (e.g. affari ‘affair, matter’ < Sicilian affari
or Italian affare) was instrumental in the emergence of this morpheme. On this
account Maltese speakers originally pluralized such words with -iet, with glideinsertion an automatic phonological consequence of the juncture of a vowel-final
stem and a vowel-initial suffix. Later, according to Geary, the whole string -ijiet
was reanalysed as constituting the marker of plurality, and this new plural suffix
was extended to consonant-final stems, including Arabic-derived items of basic
vocabulary such as omm ‘mother’ and art ‘land’.10
3.2.1.2 Derivation
Maltese displays a rich array of derivational suffixes borrowed (presumably initially as part of polymorphemic lexical items) from Italo-Romance. A definitive
list of these has not been provided to date, but Saade (2019) offers a detailed typology of such items, of which we present a simplified version here, drawing also
10
Geary’s contact-induced scenario for the emergence of this suffix may not be the whole story,
however. Evidence on this point comes from Arabic loanwords in Siwa Berber. Souag (2013:
74) lists a number of examples of Arabic-origin nouns whose plural is formed by adding a
suffix -iyyat (e.g. sḥilfa ‘turtle’, pl. sḥilfiyyat), despite the fact that both Classical Arabic and
present-day Egyptian Arabic lack plurals of this type. Siwa Berber must therefore have borrowed these items and their pluralization strategy from some early form of (eastern) Maghrebi
Arabic, suggesting that the presence in Maltese of the -ijiet suffix is, at least to some extent, an
Arabic-internal development that predates the large-scale borrowing of Italo-Romance nouns
into Maltese.
276
13 Maltese
on examples from Brincat & Mifsud (2015), and focusing just on the nominal,
adjectival and adverbial domains (see §3.2.2.2 for borrowed participial morphology).
First of all, there are at least twenty suffixes, such as the nominalizer -zzjoni,
which, though relatively frequent, only occur in items clearly borrowed wholesale from Italo-Romance (e.g. dikjarazzjoni ‘declaration’ < Italian dichiarazione)
or in coinages which, in a process that is relatively common in Maltese, represent
borrowings from English that are adapted to fit the phonology and morphology of Romance-influenced Maltese, as in esplojtazzjoni ‘exploitation’ (cf. Gatt
& Fabri 2018). Given this restriction, there must be some doubt as to whether
one can regard the suffixes themselves as borrowed, or only the polymorphemic
items in which they occur.
Secondly, there are a number of borrowed suffixes which are sufficiently well
integrated that they can attach to Arabic-derived bases. Examples include:
-ata, e.g. xemxata ‘sunstroke’ (xemx ‘sun’)
-ezza, e.g. mqarebezza ‘naughtiness’ (mqareb ‘naughty’)
-un (< Sicilian -uni, Italian -one), e.g. ħmarun ‘great fool’ (ħmar ‘donkey’)
Finally, there is at least one borrowed suffix: -tura, which forms single-instance
verbal nouns. The integration of this morpheme can be seen from the fact that it
attaches to productively to English bases, as in ċekkjatura ‘an instance of checking’ or weldjatura ‘an instance of welding’.
3.2.2 Verbs
3.2.2.1 Loaned verbs
Maltese has borrowed a large number of verbs from Sicilian and Italian, and more
recently a smaller number from English. The chief interest in these borrowings
lies in the way in which they have been integrated into the Maltese inflectional
and derivational verbal paradigms. An in-depth study of this phenomenon was
provided by Mifsud (1995), who distinguished the following four types of loaned
verbs:
Type A:
Type B:
Type C:
Type D:
Full integration into Semitic Maltese sound verbs
Full integration into Semitic Maltese weak-final verbs
Undigested Romance stems with a weak-final conjugation
Undigested English stems
277
Christopher Lucas & Slavomír Čéplö
Mifsud (1995: 58) points out that most (perhaps all) Type A verbs are so-called
“second generation” loans, whereby a nominal or adjectival form has been borrowed, a root extracted from it, and a verb formed on this root, as in pitter ‘to
paint’ – a denominal derivation from pittur ‘painter’, borrowed from Sicilian
pitturi (and supported by Italian pittore). Such items do not, therefore, represent
genuine cases of transfer of verbs, and are reminiscent of similar coinages in
other Arabic varieties (e.g. fabrak ‘to fabricate’). In Arabic as in Maltese, such
items are overwhelmingly restricted to the denominal verbal stems II and V of
triliteral roots and I and II of quadriteral roots (CVCCVC and tCVCCVC).
In contrast to Type A, Mifsud’s Types B and C are genuine cases of loaned
verbs. Mifsud (1995: 110–116) shows that the imperative (rather than the homophonous 3sg present, or any other verb forms) was the most likely base form
of the Romance models on which the Maltese loaned verbs were created.11 In
both Italian and Sicilian all verbs in the imperative end in either -i or -a. As it
happens, Maltese weak-final verbs (in which the final radical element is a vowel
rather than a consonant) also all end in either /ɪ/ or /ɐ/ in the imperfect and imperative singular, depending on which of the two weak-final conjugation classes
they fall into. This coincidence resulted in borrowed Romance verbs being integrated into one of the these two weak-final classes, as in kanta ‘he sang’, jkanta
‘he sings’ (< Sicilian/Italian imperative canta); and serva ‘he served’, jservi ‘he
serves’ (< Sicilian/Italian imperative servi).
The difference between verbs of Types B and C is that the former are analysed
as having root-and-pattern morphology, with a triliteral or quadriliteral root,
whereas Type C are borrowed as a concatenative stem without a root. This can
be seen from the fact that Type B verbs can give rise to new verbs with the same
root in other verbal stems, as in kompla ‘to continue’, tkompla ‘to be continued’
(< Sicilian cumpliri ‘to finish’), whereas Type C verbs cannot.
Another difference between Types B and C is that no Type C verb begins with
a single (ungeminated) consonant, whereas most Type B verbs do. In fact, apart
from certain well-defined exceptions (see Mifsud 1995: 152), all Type C verbs begin with a geminate consonant, as in ffolla ‘to crowd’ < Italian affollare. What
exactly was the combination of historical factors that gave rise to this synchronic
state of affairs is a complex matter (see Mifsud 1995: 158–168 for discussion), but
the key point to note is that at least some of the instances of initial gemination
in Type C verbs are apparently not attributable to phonological properties of the
source item (e.g. pprova ‘to try’ < Italian provare). It seems that speakers of Maltese came to feel that all loan verbs must have an initial geminate consonant,
whether or not this was actually true of the item being borrowed.
11
This parallels the situation in Arabic-based pidgins and creoles, for which Versteegh (2014)
shows that verbs generally appear to derive from imperatives in the lexifier varieties.
278
13 Maltese
This state of affairs manifests itself rather spectacularly in more recent borrowings from English (Type D verbs), in which initial consonants are duly geminated
(despite this never being the case in the English source items), but which also
fall into the conjugation class of weak-final verbs, as in ddawnlowdja ‘to download’. What underlies this treatment of loans from English seems to be a type
of reanalysis, which we can sketch as follows. In the initial stage, verbs without roots (not necessarily identifiable to speakers as loans from Italo-Romance)
are analysed as falling into the weak-final conjugation class because they have
a stem-final vowel. But since all verbs without roots (at this pre-English stage)
have a stem-final vowel, it is possible to view the lack of a root, not the presence
of a stem-final vowel, as the reason that loan verbs obligatorily fall into the weakfinal conjugation class; and it seems that speakers indeed made this reanalysis.
In a parallel development, initial consonant gemination also came to be seen an
obligatory feature of the class of verbs lacking a root. As a result of these developments, when a verb is borrowed from English, because it lacks a root its initial
consonant is geminated and it is conjugated as a weak-final verb, regardless of
whether it has a stem-final vowel.12
3.2.2.2 Participles
Unsurprisingly, one of the additional ways in which Type A verbs differ from the
remaining three classes of loaned verbs is the formation of passive participles: in
Type A verbs, passive participles are formed in accordance with the Semitic pattern for the respective derived stem, e.g. pejjep ‘to smoke’ (stem II, from Italian
pipa ‘pipe’) produces mpejjep ‘smoked’ (Mifsud 1995: 70). In contrast, some Type
B verbs allow for the formation of a passive participle using Romance suffixes
(Mifsud 1995: 127–133), and this is the sole option for Type C and even Type D
verbs: for Type C verbs, the choice of the actual suffix depends on the original
form of the verb and, in some cases, the path of transfer (see below). For Type D
verbs borrowed from English, the suffix -at is the only productive way to form
a passive participle (e.g. inxurjat ‘insured’) with spellut ‘spelled’ as the only exception (Mifsud 1995: 248).
And finally, there are two distinct classes of Type B and C verbs which can
each derive two passive participles. In the first class, one participle is derived
from the weak (regular) form root and the other derived from the strong one, e.g.
12
In addition, virtually all Type D verbs insert a palatal glide between the borrowed stem and
the added weak-final vowel, as in pparkja ‘to park’. Similarly to the initial gemination and
weak-final inflection of Type D verbs, this glide insertion must be the result of analogical
extension from numerous glide-final borrowed Romance verbs, e.g. rdoppja ‘to double’ < Italian
raddoppiare. See Mifsud (1995: 225–236) for a detailed discussion.
279
Christopher Lucas & Slavomír Čéplö
konfondut ‘confused’ vs. konfuż (Mifsud 1995: 134). In the second class, one participle is derived using the Sicilian suffix -ut, the other using the Italian-derived
suffix -it, e.g. preferut ‘preferred’ vs. preferit (Mifsud 1995: 230). The reason for
these doublets is largely sociolinguistic: the variability of the first class echoes
a similar situation in Italian dialects (Mifsud 1995: 134); that of the second class
reflects a situation whereby the loaned verb effectively has two sources, spoken
Sicilian and Standard (Tuscan) Italian.
3.3 Syntax
3.3.1 Phrase syntax
3.3.1.1 Word order
The expansion of Maltese lexicon with items borrowed from Sicilian and Italian
had a profound effect on the syntax of Maltese. The primary example of this is
word order within the noun phrase, involving the order of adjectives and their
heads. In Arabic, adjectives (with the exception of comparatives, superlatives
and a number of specific cases) follow their heads. This is largely true of Italian
adjectives as well, with the exception of a small subclass some grammars term
“specificational adjectives” (e.g. Maiden & Robustelli 2007: 55–56), such as stesso
‘same’ and certo ‘certain’, which precede their head. Such adjectives borrowed
into Maltese retained their syntactic properties, as with the pre-nominal ċertu (<
Sicilian certu) in (3).
(3) [BCv3: it-torca.8685]
Kien
bniedem ta’ ċerta
personalità.
be.prf.3sg.m person gen certain.f personality
‘He was a person with a certain personality.’
In Italian, specificational adjectives to a large extent overlap with a class of
adjectives that perform double duty as quantifiers (or perhaps determiners) and
vary their position according to their respective roles: Adj–N for quantifiers, N–
Adj for adjectives. One could argue that it is in the former function that they were
borrowed into Maltese and thus should be considered quantifiers or determiners
rather than adjectives, especially in light of the fact that they are (for the most
part) in complementary distribution with the definite article, as determiners and
quantifiers are. Determiners and quantifiers in Maltese precede their heads (as
with the definite article il-, kull ‘all’, xi ‘some’ etc.).
There are three arguments against such an account: first of all, borrowed prenominal specificational adjectives actually fall into two classes, where members
280
13 Maltese
of the first, such as ċertu ‘certain’, diversi ‘diverse’ (< Italian diverso) or varju
‘various’ (< Sicilian varju), do not (for the most part) allow the definite article.
In contrast, words in the second class such as stess ‘same’ (< Italian stesso) or
uniku ‘unique’ (< Sicilian uniku) predominantly co-occur with the definite article
when pre-nominal. The same, incidentally, is true of the etymologically Arabic
pre-nominal quantifier ebda ‘no, none’.
Secondly, there are morphological considerations: pre-nominal specificational
adjectives of both types mark gender and/or number (varju for the first, uniku
for the second) like Maltese adjectives do; Maltese determiners and quantifiers
do not inflect for either gender or number.13
The final argument against considering borrowed pre-nominal specificational
adjectives as being borrowed into the slot for determiners involves ordinal numerals. In Italian, these also fall into the subclass of prenominal specificational
adjectives (Maiden & Robustelli 2007: 55) and thus precede their head. The same
is invariably true of Maltese ordinal numerals, as with ewwel in (4).
(4)
[BCv3: l-orizzont.64586]
wara l-ewwel sena
after def-first year
‘after the first year’
In North African Arabic, ordinal numerals can either precede or follow their
heads, but when they precede them, they never take the definite article, even
when the noun phrase is semantically definite (see e.g. Ritt-Benmimoun 2014:
284 for Tunisian Arabic). In contrast, Maltese never allows its ordinal numerals
to follow their heads, and the definite article is obligatory.
All these arguments, including the comparison with related Arabic varieties,
suggest that the pre-nominal position of some adjectives and ordinal numerals
in Maltese is due to transfer under recipient-language agentivity from Italian.
3.3.1.2 The analytical passive
As with adjectives (§3.3.1.1), lexical borrowings from Italo-Romance have also had
a significant impact on the syntax of Maltese verbs. One of the most conspicuous
consequences of this development involves the passive voice: as Romance-origin
verbs cannot generally form one of the passive derived verbal stems (but see
§3.2.2.1), they brought with them their Romance syntax and thus a new type of
passive construction arose in Maltese – the analytical passive.
13
With the exception of the very specific category of demonstrative pronouns where gender and
number are marked not by affixes, but rather a form of suppletion.
281
Christopher Lucas & Slavomír Čéplö
In Maltese, there are two types of analytical passive construction containing a passive participle: the so-called “dynamic passive” (Vanhove 1993: 321–
324; Borg & Azzopardi-Alexander 1997: 214), which combines passive participles
with the passive auxiliary ġie ‘to come’; and the so-called “stative passive” (Borg
& Azzopardi-Alexander 1997: 214, Vanhove 1993: 318–320), which has the same
structure as copular clauses (see §3.3.2.3), the only difference being that stative
passive constructions can feature an agentive NP introduced by the preposition
minn ‘from’ (see Čéplö 2018: 104–107 for a detailed analysis).
The stative passive can be viewed as an extension of the structurally identical
construction which is sporadically attested already in Classical Arabic (Ullmann
1989: 76–84), but becomes quite prominent in Christian Arabic documents at least
as early as tenth century, where, incidentally, it gained prominence under influence from Aramaic and Greek (Blau 1967: 424).
The dynamic passive (5), on the other hand, is a straightforward calque on
either Italian or Sicilian, where a construction featuring a verb semantically equivalent to ġie ‘to come’ – venire in Italian – combines with a past participle (see
also Manfredi, this volume).
(5) [MUDTv1: 30_01P05]
Kif diġà
għedt,
ġie
ppreżentat
il-kuntratt.
as already say.prf.1sg come.prf.3sg.m present.ptcp.pass def-contract
‘As I already said, the contract was presented.’
While the dynamic passive must have originally functioned to fill a hole in the
verbal system of Maltese by providing a way to passivize Romance verbs, it has
meanwhile spread to include native verbs as well, as with ta ‘to give’ (< √ʕṭy) in
(6).
(6) [BCv3: inewsmalta-ott.29.2013.1257-11045]
It-tagħrif
ġie
mogħti
mill-Ministru
def-information come.prf.3sg.m give.ptcp.pass from.def-minister
Konrad Mizzi.
Konrad Mizzi.
‘The information was given by Minister Konrad Mizzi.’
3.3.1.3 Modality
Another clear-cut example of grammatical calquing comes from the domain of
modality and involves the pseudoverb għand-. In Maltese, its primary function
is that of a possessive (7), as is the case with its cognates ʕind-/ʕand- in many
Arabic varieties.
282
13 Maltese
(7)
[MUDTv1: 22_02J03]
M’ għandi xejn
kontri-hom.
neg have.1sg nothing against-3pl
‘I have nothing against them.’
In addition to this, however, the Maltese għand- has also taken on a function
as a deontic modal of weak obligation ‘should, ought’ taking verbal complement,
as in (8).14
(8)
[MUDTv1: 22_02J03]
Naqbel
li
għandhom jivvutaw
aktar nies.
agree.impf.1sg comp have.3pl vote.impf.3pl more people.
‘I agree that more people should vote.’
The use of għand- in this kind of modal function appears to be unique to Maltese; not even Cypriot Maronite Arabic with its many parallels to Maltese (on
which see below) exhibits the same behavior for its cognate ʕint- (Borg 2004:
346) and uses a different verb, salaḫ/pkyislaḫ (Borg 2004: 323), as the default deontic modal. The Maltese development must therefore be another calque, since
the basic possessive verb of Sicilian, aviri, also doubles as a deontic modal, as in
(9).
(9)
Sicilian (Piccitto 1977: 340)
Cci
l’
àiu
a-ddiri
a-tto
patri.
dat.3sg.m obj.3sg.m have.pres.1sg to-say.inf dat-2sg.m father
‘I have to say it to your father.’
3.3.2 Sentence syntax
3.3.2.1 Differential object marking
Differential object marking (DOM) is a phenomenon whereby direct objects are
marked according to some combination of the semantic and pragmatic properties of the object in question. In Spanish, for example, objects denoting humans
(and equivalent entities) are marked by the particle a, originally a directional
preposition. DOM is a phenomenon attested cross-linguistically (see Khan 1984
for Semitic languages), including in varieties of Arabic such as Levantine, Iraqi
(Coghill 2014 and references therein), and Andalusi (University of Zaragoza 2013:
108).
14
għand- is the only Maltese pseudoverb (and verb) which exhibits a three-way distinction between present (għand-), past (kell-) and future/habitual (ikoll-) forms; all can occur in the modal
function.
283
Christopher Lucas & Slavomír Čéplö
DOM is a well-documented feature of Maltese morphosyntax and largely conforms to the Spanish prototype: in general, both pronominal and nominal direct
objects denoting entities high in the “animacy hierarchy” (Borg & AzzopardiAlexander 1997: 55) take the object marker lil (10), which also does double duty
as the indirect object marker for all objects. Inanimate direct objects do not take
lil (11).
(10) [BCv3: ilgensillum.2011-Mejju-22.8230]
Min jara
lili
jara
lil Missier-i.
who see.impf.3sg.m obj.1sg see.impf.3sg.m obj father-obl.1sg
‘Who looks at me, looks at my Father.’
(11) [BCv3: l-emigrant]
orrizzonti ġodda u min baħħ.
Min jara
who see.impf.3sg.m horizon.pl new.pl and who void.
‘Some see new horizons, some see a void.’
Döhla (2016) examines DOM in Maltese in some detail and arrives at the conclusion that while there is “a certain predisposition for object marking in general within pan-Arabic grammar” (2016: 169), Maltese DOM cannot be ascribed
to purely internal developments within Arabic. A striking feature of the Arabic varieties that exhibit DOM is that they were all in prolonged contact with
other languages: Aramaic for Levantine and Iraqi Arabic (and, by extension, for
Cypriot Maronite Arabic, cf. Borg 2004: 412), Romance for Andalusi Arabic and
Maltese. In the case of Maltese, the Romance variety in question is Sicilian, where
the object marker a performs the same double duty as the Maltese lil, and DOM
in both languages shows a number of remarkable similarities: in both Sicilian
and Maltese, DOM is primarily triggered “by humanness along with definiteness/referentiality” (Iemmolo 2010: 257, in reference to Sicilian), it is obligatory
with personal pronouns, but optional with plural “kinship terms and human common nouns” and disallowed with “(in)animate and indefinite non-specific nouns”
(Iemmolo 2010: 257, again in reference to Sicilian), as exemplified by the nonspecific Maltese nies ‘people’ in (12).
(12)
[BCv3: l-orizzont.41390]
hekk?
nies jgħixu
jara
Min irid
who want.impf.3sg.m see.impf.3sg.m people live.impf.3pl.m thus
‘Who wants to see people live like that?’
In Maltese DOM, then, we have an instance of what Manfredi (this volume)
labels “calquing of polyfunctionality of grammatical items inducing syntactic
284
13 Maltese
change”: Maltese acquired a rule of DOM as a result of the indirect object marker
lil inheriting the dual function of its Sicilian equivalent a. It is clear that this is a
contact-induced change. But since with this and the similar changes discussed below there is no transfer of lexical matter, it seems impossible at present to judge
whether they are the result of borrowing or imposition, or whether they were
actuated by speakers for whom neither the source language nor the recipient
language were dominant, in the process that Lucas (2015) calls “convergence”.
3.3.2.2 Clitic doubling (proper)
The existence of various reduplicative phenomena associated with direct and indirect clitic pronouns in Maltese has been noted at least since Sutcliffe (1936: 179),
who identifies what classical tradition refers to as nominativus pendens. This analysis has been elaborated on by Fabri (1993), Borg & Azzopardi-Alexander (1997)
and Fabri & Borg (2002), primarily in the context of pragmatically determined
constituent order variation, especially topicalization. Building on these works
and the analysis of Maltese clitics by Camilleri (2011), Čéplö (2014) notes that
in addition to these phenomena, which in one way or another entail dislocation, there exists in Maltese another related phenomenon, where lexical objects
and clitic pronouns co-occur, but without the dislocation of the lexical object.
This phenomenon, termed Clitic Doubling Proper to distinguish it from similar
constructions (see Krapova & Cinque 2008 for a detailed analysis), involves the
co-occurrence of a lexical object and the clitic with the object in situ, which in
Maltese is after the verb (see Čéplö 2018). Maltese Clitic Doubling Proper occurs
with both direct (13) and indirect objects (14).
(13)
[BCv3: l-orizzont.36758]
Ftit nies jafu-ha
l-istorja
marbuta
ma’
few people know.impf.3pl.m-3sg.f def-history connected.sg.f with
il-proġett tant sabiħ.
dan
dem.sg.m def-project such beautiful
‘Few people know the history connected with such a beautiful project.’
(14)
[BCv3: 20020313_714d_par]
f’ pajjiż-na.
lil dawn in-nies
Hekk qed ngħidu-lhom
thus prog say.impf.1pl-dat.3pl dat dem.pl def-people in country-1pl
‘This is what we say to these people in our country.’
Unlike various types of dislocation with resumptive clitic pronouns which are
quite common in European languages (see e.g. de Cat 2010), Clitic Doubling
285
Christopher Lucas & Slavomír Čéplö
Proper is a much rarer phenomenon; in Europe, it is largely confined to the
Balkan Sprachbund (Friedman 2008) and some Romance languages outside of
the Balkans, like Spanish (Zagona 2002: 7) and varieties of Italian (Russi 2008:
231–233). The phenomenon is also attested in Semitic languages (Khan 1984), including Arabic, where it was studied in detail by Souag (2017). Comparing Clitic
Doubling Proper in various varieties of Arabic including Maltese, Souag (2017:
57) notes parallels between Maltese and some varieties of Algerian Arabic, especially in regard to the doubling of indirect objects. Ultimately, however, he
arrives at the conclusion that Maltese Clitic Doubling Proper “has little in common with any other Arabic variety examined, but closely resembles that found in
Sicilian” (Souag 2017: 60). This suggests that here too we have a contact-induced
change, this time of the sort that Manfredi (this volume) labels “narrow syntactic
calquing”, that is, without any accompanying calque of lexical items.
3.3.2.3 Copular constructions
In Maltese, there are four types of copular clauses (Borg & Azzopardi-Alexander
1997: 53):15
Type 1:
Type 2:
Type 3:
Type 4:
No copula
The verb kien as the copula
Personal pronoun as the copula
Present participle qiegħed as the copula
Type 1 describes what traditional grammars of Semitic languages refer to as nominal sentences; copular clauses with an explicit verbal copula (kien) then fall into
Type 2. Types 3 and 4, while not without parallel in other varieties Arabic,16 feature much more prominently in Maltese. This is especially true of Type 3 copular
clauses, which involve the use of a personal pronoun as the copula (15).
(15)
15
[BCv3: 2010 Immanuel Mifsud - Fl-Isem tal-Missier (U tal-Iben)]
omm-ok.
Din hi
this.f 3sg.f mother-2sg
‘This is your mother.’
In addition to these, Borg (1987–1988) and Borg & Spagnol (2015) also describe the copular
function of the verb jinsab ‘to be found’. This being a finite verb, both Borg & AzzopardiAlexander (1997: 53) and Čéplö (2018: 99–104) exclude this type of clause, as well as similar
ones, such as those featuring the verb sar ‘to become’, from the category of copular clauses.
16
See the analysis of Type 4 copulas in Camilleri & Sadler (2019).
286
13 Maltese
Similar copular constructions to that illustrated in (15) have been described for
several Maghrebi varieties (cf. Vanhove 1993: 355), but Maltese stands apart in
terms of the frequency with which Type 3 constructions occur: in MUDTv1, for
example, 110 non-negative copular clauses are of Type 1; 181 are Type 3. In this,
Maltese Type 3 copular clauses are comparable to equivalent copular constructions in Anatolian Arabic (see Lahdo 2009: 172–173 for Tillo Arabic and the
references therein, as well as Akkuş, this volume), Andalusi Arabic (University
of Zaragoza 2013: 105), and especially Cypriot Maronite Arabic (Borg 1985: 135;
Walter, this volume), where they are but one piece of evidence linking Cypriot
Maronite Arabic to qəltu dialects (Borg 2004: 31). The conclusion to be drawn
here is the same as for DOM and Clitic Doubling Proper above: it is no coincidence that these copular constructions are in wide use and the copular construction of choice especially in varieties of Arabic which have been under contact
influence from languages with a mandatory copula – Turkish for Anatolian Arabic, Spanish for Andalusi Arabic, Greek for Cypriot Maronite Arabic, and Italian
for Maltese. Whether the origin of such constructions can be traced to a feature in
(one of) these dialects’ Old Arabic ancestors, or whether they came about through
parallel development, contact undoubtedly triggered the widespread adoption of
such constructions in these varieties of Arabic.
3.4 Lexicon
3.4.1 Major sources
That Maltese contains large numbers of loanwords from Romance and English
is a fact immediately obvious to even the most casual observer. Over the years,
there have been a number of attempts to quantify the influence of other languages on Maltese by providing a classification of lexemes by their origin. The
earliest, Fenech (1978: 216–217), compiled such statistics for journalistic Maltese,
but also provided a comparison to literary and spoken Maltese (albeit using a
very small data sample). Brincat analyzed the etymological composition of entries in Aquilina’s dictionary, first examining the origin of 34,968 out of all 39,149
headwords (Brincat 1996: 115) and then applying the same analysis to the entire
list (Brincat 2011: 407); Mifsud & Borg (1997) did the same with the vocabulary
contained in an introductory textbook of Maltese as a foreign language. In 2006,
Bovingdon & Dalli (2006) analyzed the etymology of lexical items in a 1000-word
sample obtained from a corpus of Maltese and, most recently, Comrie & Spagnol
(2016: 318) did the same on a list of 1500 “lexical meanings” within the framework
of the Loanwords in the world’s languages project (Haspelmath & Tadmor 2009).
Figure 1 summarizes all these findings.
287
Christopher Lucas & Slavomír Čéplö
100
Origin
%
75
Semitic
Romance
English
Other
50
25
Fen
ech
(19
78:
140
) (jo
urn
alis
tic)
Fen
ech
(19
78:
140
) (li
tera
Fen
ry)
ech
(19
78:
140
) (s
pok
en)
Br i
nca
t (1
996
: 11
5)
Mif
sud
&B
org
Bov
(19
97)
ing
don
&D
alli
(20
06:
68−
69)
Br i
nca
t (2
011
: 40
Com
7)
rie
&S
pag
nol
(20
16:
325
)
0
Figure 1: A summary of previous studies of the composition of the Maltese lexicon
The primary explanation for the sharp differences between these analyses is
methodology: while Fenech (1978) analyzes entire texts and thus counts tokens,
Brincat (1996) (including its updated version in Brincat 2011) and Bovingdon &
Dalli (2006) analyze lists of unique words, i.e. types. The later is also true of
Mifsud & Borg (1997) and Comrie & Spagnol (2016), except where Brincat (1996)
uses dictionary data and Bovingdon & Dalli (2006) corpus data, Mifsud & Borg
(1997) employ a list of lexical items with high frequency of use in daily communication and Comrie & Spagnol (2016) base their analysis on a list compiled for
the purposes of cross-linguistic comparison. The high ratio of words of Semitic
origin in token-based analyses is thus due to the prevalence of function words,
which are overwhelmingly Arabic. The type-based analyses then provide a somewhat more accurate picture of the lexicon as a whole, even though they are not
without their problems. Chief among these is the issue of what exactly counts
as type, especially with regard to productive derivational affixes, e.g. whether all
the words with the prefix anti- count as distinct types or not.
In addition to general analyses, both Bovingdon & Dalli (2006) and Comrie
& Spagnol (2016) also provide breakdowns for individual parts of speech. Unfortunately, these analyses are not comparable, as each has a different focus:
Bovingdon & Dali (2006: 71) are interested in the composition of each etymological stock by word class (Table 3).
288
13 Maltese
Table 3: Source language component of Maltese by word class
(Bovingdon & Dalli 2006: 71).
Origin
Function words
V
Adj
N
Adv
Prn
3%
0%
0%
70%
38%
29%
2%
11%
8%
21%
48%
63%
2%
3%
0%
2%
0%
0%
Semitic
Romance
English
In contrast, Comrie & Spagnol (2016: 328) focus on the composition of individual word classes by their origin (Table 4).17
Table 4: Word class composition by source language (Comrie & Spagnol 2016: 328)
Word class
Function words
Verbs
Adjectives
Nouns
Arabic
Romance
English
Misc.
84.7%
75.3%
65.2%
44.7%
6.2%
14.1%
28.5%
39.6%
0%
1.3%
0.3%
7.2%
9.1%
9.2%
6.0%
8.6%
Comrie & Spagnol (2016) also provide a breakdown of their data by semantic
field, permitting a comparison of the domains in which Romance versus English
loans are more or less prominent. A number of generalizations can be made here
(see Table 5 for a summary), though ultimately they all follow naturally from the
fact that contact with English was more recent, and less intensive, than contact
with Sicilian and Italian.
Unsurprisingly, English is best represented in the category of items relating to
the modern world, but even here Romance dominates. Examples include Englishderived televixin ‘television’ and Italian-derived kafè ‘coffee’.
The domain of animals divides rather neatly as follows. Common animals (especially land animals) of the Mediterranean area are largely Arabic-derived (e.g.
17
The details of Comrie & Spagnol’s (2016) methodology mean that loans in their dataset come
from Romance and English but not from any other languages. The category we label “Misc.”
in Tables 4 and 5 encompasses those meanings in the Loanwords in the world’s languages 1500item set which have no corresponding single-word Maltese lexical item, and those where the
etymology is at present unknown, or where the item in question is an innovative Malteseinternal coinage.
289
Christopher Lucas & Slavomír Čéplö
Table 5: Composition of semantic fields by source language (Comrie &
Spagnol 2016: 327)
Semantic field
Modern world
Animals
Clothing and grooming
Warfare and hunting
Law
Social and political relations
Arabic
Romance
English
Misc.
3.0%
47.8%
38.7%
28.8%
36.0%
48.4%
65.3%
29.1%
47.2%
65.0%
50.0%
48.4%
22.8%
13.9%
10.4%
2.5%
0.0%
0.0%
9.0%
9.1%
3.8%
3.8%
14.0%
3.2%
fenek ‘rabbit’ < Maghrebi Arabic fanak ‘fennec fox’), while well-known nonindigenous animals are largely Romance-derived (e.g. ljunfant ‘elephant’ < Sicilian liufanti, the additional /n/ perhaps the result of influence from ljun ‘lion’).
More exotic animals, if there is a corresponding Maltese item at all, derive from
English (e.g. tapir ‘tapir’). Clothing and grooming presents a similar picture,
with Arabic-derived suf ‘wool’, Sicilian-derived ngwanta ‘glove’, and Englishderived fer ‘fur’, as does warfare and hunting, with Arabic-derived sejf ‘sword’,
Sicilian-derived xkubetta ‘gun’, and English-derived senter ‘shotgun’ (< centrebreech-loading shotgun).
The total lack of English loans in the domains of law and social and political relations, at least in Comrie and Spagnol’s sample, is remarkable, given the extent
to which the English language dominated public life in Malta in the twentieth
century. A generalization that underlies this finding is that while English influence is strongest in the spheres of commerce, consumerism and, especially in the
twenty-first century, popular culture (e.g. vawċer ‘voucher’, ċċettja ‘to chat’),18
at least as far as Maltese lexicon is concerned, it has not supplanted Italian in the
domains of high culture and the affairs of state (e.g. gvern ‘government’ < Italian
governo, poeżija ‘poem’ < Italian poesia).
18
Until at least 1991, when the Maltese government opened up television broadcasting rights to
more than just the single state broadcaster TVM, Italian television stations, whose broadcasts
from Sicily could be received in Malta, were very widely watched, and there was consequently
considerable Italian influence on Maltese popular culture (Sammut 2007). This influence has
waned considerably at the expense of English and American culture since the advent of broadcast pluralism in Malta, and especially with the rise of cable television and online video streaming.
290
13 Maltese
3.4.2 Minor sources
Considering its location and the nature of population movements in the Mediterranean, it is hardly surprising that the Maltese lexicon also contains borrowings from languages other than Sicilian, Italian and English. The most obvious
of these are borrowings from other Romance languages. First among them, as in
other European languages, stands Latin, which provided a large chunk of Maltese
scientific and technical vocabulary, whether as terminology (e.g. ego, rektum or
sukkursu ‘underground water’), biological nomenclature (fagu ‘European beech,
Fagus sylvatica’, mirla ‘brown wrasse, Labrus merula’) or set phrases and expressions (ex cathedra, ibidem). Curiously for a Catholic country, Latin is the source of
very little religious vocabulary in Maltese; in this area, Maltese continues to rely
almost exclusively on words of Arabic origin. Those Latin words related to religious matters employed in modern Maltese therefore typically refer to minutiae
of Catholic Church rituals and procedures, such as ekseat ‘a bishop’s permission
for a priest to leave the diocese’ (< exeat) or indult ‘a Pope’s authorization to perform an act otherwise not allowed by canon law’. Of the few Latin terms related
to religion still in common use, nobis stands out as a rather curious lexical item:
in Maltese, it is used as a (post-nominal) modifier indicating intensity or size, as
in tkaxkira nobis ‘a sound thrashing’ or tindifa nobis ‘a thorough cleaning’.
Before the Order of Saint John gained control of Malta, the islands were for
more than two centuries a part (whether officially or not) of the Crown of Aragon.
As such, one would expect that speakers of Maltese during that era found themselves exposed the languages of the Crown like Catalan, Spanish and Occitan,
and that this was then reflected in the Maltese lexicon. In truth, however, there
are only a few Maltese words that can clearly be traced to Ibero-Romance. Biosca
& Castellanos (2017) identify a number of lexical items with Catalan or Occitan
origins, but note that many of them can also be found in Sicilian, which in most
cases can be clearly determined as the origin of the loan. On the other hand,
there are Maltese words of obviously Romance origin whose current shape cannot be easily explained by any of the processes by which Sicilian or Italian words
were made to conform to Maltese phonology, and where the Catalan or Occitan
origin postulated by Biosca & Castellanos (2017) may offer a better explanation
than that of “local formation” resorted to by previous works. These may include:
boxxla ‘compass’ < Catalan búixola vs. Italian bussola; frixa ‘pancreas’ < Catalan
freixura ‘entrails’ and even the very frequent żgur ‘certain’, which, due to its
phonology, especially the /g/ (see §3.1.1.1), points to an origin in Catalan segur or
Spanish seguro, rather than to its (Tuscan) Italian or Sicilian cognates, which both
feature a /k/ in its place. These and other lexical items, onomastics (see Biosca &
291
Christopher Lucas & Slavomír Čéplö
Castellanos 2017: 46), and even usage (such as the ubiquitous Maltese swear word
l-ostja, literally ‘the host, sacramental bread’, which is very atypical for Italian or
Sicilian, but has a counterpart in the Spanish la hostia) suggest some influence
of Ibero-Romance on Maltese which is yet to be thoroughly researched.
The much shorter French occupation of the Maltese islands left very little linguistic trace, and so it is internationalisms in the semantic field of culture (bonton
‘high society’, etikett ‘etiquette’), fashion (manikin ‘manequin’) and the culinary
arts (fundan ‘fondant’, ragu ‘ragout’) where French borrowings in Maltese can
be found. The few notable exceptions include berġa (< auberge), the term used
for the residences of langues (chapters) of the Order of Saint John. The most
prominent of these palaces, Berġa ta’ Kastilja, now houses the office of the Prime
Minister of Malta, for which the term Berġa is often used metonymically. The
other two Maltese words of French origin still in frequent daily use both happen
to be connected to transportation: xufier (< chauffeur) ‘driver’ and xarabank (<
char à bancs) ‘bus’. The latter is particularly interesting due to its pronunciation
/ʃɐrɐˈbɐnk/, which indicates that it was borrowed directly from French and not
from English (which would give /ʃɛrɛˈbɛnk/, as well as for its connection to the
French-speaking Maghreb, where the same word was in use; this indicates the
possibility that it was brought from there by Maltese expatriates.
In addition to Romance languages, post-classical Greek, with its ubiquitous
presence all across the Mediterranean (including the neighboring Sicily), could
not help but leave a trace on Maltese vocabulary, small though it is. Aquilina
(1976: 23) gives Lapsi ‘Feast of Ascension’ (< análipsi) as the solitary example of
a Maltese religious term not inherited from Christian Arabic or borrowed from
Romance languages. The other two examples of Greek loanwords involve a completely different sphere. The first is ħamallu ‘lewd, vulgar person’, from Greek
xamális (Dimitrakou 1958: 7781). This word may ultimately be traceable to Arabic
(through Turkish), as is evident from its other meaning in Greek, namely ‘porter’
(< ḥammāl). However, the meaning in which it appears in Maltese is unique to the
Greek word, indicating that it was borrowed into Maltese from Greek. The other
such term is vroma ‘complete failure, fiasco’ which is quite straightforwardly
traceable to the Greek vróma ‘dirt, filth’ (Dimitrakou 1958: 1506, 1516).
With regard to the debates on the origin and history of Maltese, borrowings
from other Afro-Asiatic languages have long been at the centre of attention of
Maltese etymological research. Berber is perhaps the most notorious example
here, with a number of items cited as having Berber origins by Colin (1957) and
Aquilina (1976: 25–39). Aquilina’s list is an expansion of Colin’s and thus both
feature the same conspicuous items, which for the most part involve zoology,
such as fekruna ‘tortoise’ (< fekrun; Naït-Zerrad 2002: 553) and gendus ‘bull’ (<
292
13 Maltese
agenduz; Naït-Zerrad 2002: 827). Additionally, Aquilina postulates a Berber origin for a number of lexical items where this seems questionable. In some cases the
items in question are obviously Arabic loanwords in Berber (as with bilħaqq ‘by
the way’, quite transparently from Arabic b-il-ḥaqq ‘in truth’). In other cases subsequent research has argued against a Berber origin. For example, while Aquilina
identifies żenbil ‘a large carrying basket’ as having a Berber origin, Borg (2004:
261) notes that it can also be found in the Arabic dialect of Aleppo and Arbil,
and traces its ultimate origin to Akkadian through Aramaic. A large group of
similarities between Maltese and Berber identified by Aquilina involve “Berber
nursery language”, containing items like Berber papa ‘bread’ and Maltese pappa,
Berber ppspps or ppssi ‘urine’ and Maltese pixxa, and Berber kakka/qaqah and
Maltese kakka (both having to do with defecation). These forms are actually attested cross-linguistically (Ferguson 1964) at least as far north as Slovak (Ondráčková 2010) and cannot thus be considered loans from Berber. Nevertheless,
the fact that there is a Berber lexical component in Maltese is well established,
and Souag (2018) has shown that it may be larger than previously thought (e.g.
his case for the Berber etymology of the frequent adjective ċkejken ‘small’).
Finally, in addition to Berber, Maltese also contains a small number of words
that can be reasonably traced back to Aramaic. Along with obsolescent lexical
items such as żenbil given above or andar ‘threshing floor’ (Behnstedt 2005: 116–
117), this small list includes the frequent verb xandara ‘to broadcast, to spread
(news)’, otherwise unattested in any other variety of Arabic (Borg 1996: 46). This
verb is presumably derived from the common Aramaic root √šdr ‘to dispatch,
send’ with cognates in Mandaic (Drower & Macúch 1963: 450), Jewish Babylonian Aramaic (JBA; Sokoloff 2002: 1112-1113) and Christian Neo-Aramaic (Khan
2008: 1179). The insertion of [n] reflects the dissimilation of the geminated [dd]
into [nd] (Lipiński 1997: 175–176); the same phenomenon involving the original
geminated [bb] can also account for żenbil (cf. JBA zabbīlā; Sokoloff 2002: 397).
These borrowings could on the one hand strengthen the case for a Levantine substrate in (if not origin of) Maltese, as Borg (1996) insists; on the other hand, some
of them can also be found in other North African varieties (Behnstedt 2005).
4 Conclusion
This chapter has reviewed the extensive changes that have taken place in Maltese
as a result of contact with Sicilian, Tuscan Italian, English, and other languages.
The changes due to contact with Italo-Romance languages are so striking, especially but by no means only with respect to lexicon, that it is almost misleading
293
Christopher Lucas & Slavomír Čéplö
to speak of these contacts having changed “Maltese”. Rather it might be argued
that it was a Maghrebi Arabic dialect like any other that was subjected to these
changes, and that Maltese, the distinct language that its speakers now feel it to
be, was what emerged only once these changes were complete. The result is a
language in which typically Semitic and typically Indo-European elements exist
side-by-side at all linguistic levels.
The elements of contemporary Standard Maltese that are the result of contact,
summarized in this chapter, are now relatively well understood. But the language
has naturally also evolved in numerous ways that owe little or nothing to the effects of contact with other languages. With a few notable exceptions (e.g. Borg
1978; Vanhove 1993), these changes have received far less attention. A desideratum for future historical linguistic work on Maltese is therefore to redress this
imbalance.
Concerning contact-induced change specifically, future research could fruitfully include comparative work on the differential effects of contact on standard
versus dialectal Maltese. And to the extent that it is possible, the field would
benefit greatly from a detailed history of the sociolinguistic effects of language
contact in Malta in the early modern period.
Further reading
) Krier (1976) is a short monograph on the influence of Italo-Romance on Maltese
phonology, morphology, syntax, and lexicon.
) Mifsud (1995) gives an in-depth description of Maltese loaned verbs.
) Comrie & Spagnol (2016) examine lexical borrowing in Maltese in the context
of loanword typology crosslinguistically.
) Drewes (1994) and Stolz (2003) explore the question of whether Maltese is
properly labeled a “mixed language”.
Acknowledgements
The research presented in this chapter was partly funded by Leadership Fellows
grant AH/P014089/1 from the UK Arts and Humanities Research Council, grant
number APVV-15-0030 from the Slovak Research and Development Agency
(APVV), and ERC Starting Grant number 679083 from the European Research
Council, whose support is hereby gratefully acknowledged.
294
13 Maltese
Abbreviations
1, 2, 3
Adj
Adv
BCE
BCv3
CE
comp
dat
def
dem
dep
DOM
f
JBA
gen
impf
inf
prg
1st, 2nd, 3rd person
adjective
adverb
before Common Era
Bulbulistan corpus malti v3
Common Era
complementizer
dative
definite article
demonstrative
dependent form
differential object marking
feminine
Jewish Babylonian Aramaic
genitive
imperfect (prefix conjugation)
infinitive
pragmatic marker
L1, L2
m
MUDTv1
N
neg
NP
obj
obl
pass
pl
prf
Prn
prog
ptcp
sg
sing
V
1st, 2nd language
masculine
Maltese Universal
Dependencies Treebank v1
noun
negative (particle)
noun phrase
object
oblique
passive
plural
perfect (suffix conjugation)
pronoun
progressive
participle
singular
singulative
verb
Primary sources
Maltese examples above are primarily cited from the general corpus of Maltese
bulbulistan corpus malti v3 (accessible at www.bulbul.sk/bonito2, login: guest,
password: Ghilm3), as well as from the Maltese Universal Dependencies Treebank
v1 (accessible at www.bulbul.sk/annis-gui-3.4.4/), both described as to their composition and annotation in Čéplö (2018). Each citation is accompanied by an abbreviation identifying the source (BCv3 and MUDTv1, respectively), as well as
the specific document where it can be found.
References
Aquilina, Joseph. 1976. Maltese linguistic surveys. Malta: The University of Malta.
Avram, Andrei A. 2012. Some phonological changes in Maltese reflected in
onomastics. Bucharest Working Papers in Linguistics 14. 99–119.
Avram, Andrei A. 2014. The fate of the interdental fricatives in Maltese. RomanoArabica (14). 19–32.
295
Christopher Lucas & Slavomír Čéplö
Avram, Andrei A. 2017. Word-final obstruent devoicing in Maltese: inherited,
internal development or contact-induced? Academic Journal of Modern
Philology 6. 23–37.
Ballou, Maturin M. 1893. The story of Malta. Cambridge, MA: The Riverside Press.
Behnstedt, Peter. 2005. Voces de origen sirio y yemení en el árabe magrebí /
andalusí. In Jordi Aguadé, Ángeles Vicente & Leila Abu Shams (eds.), Sacrum
Arabo-Semiticum, 115–122. Zaragoza: Instituto de Estudios Islámicos y del
Oriente Próximo (IEIOP).
Biosca, Carles & Carles Castellanos. 2017. Aspects of the comparison between
Maltese, Mediterranean Lingua Franca and the Occitan-Catalan linguistic
group (13th–15th centuries). In Benjamin Saade & Mauro Tosco (eds.), Advances in Maltese linguistics, 39–64. Berlin: De Gruyter.
Blau, Joshua. 1967. A grammar of Christian Arabic based mainly on SouthPalestinian texts from the first millennium, Fasc. ii: §§170–368. Louvain:
Secrétariat du CorpusSCO.
Blouet, Brian. 1967. The story of Malta. London: Faber & Faber.
Borg, Albert. 1987–1988. To be or not to be: A copula in Maltese. Journal of Maltese
Studies 17–18. 54–71.
Borg, Albert & Marie Azzopardi-Alexander. 1997. Maltese. London: Routledge.
Borg, Albert & Michael Spagnol. 2015. Nominal sentences in Maltese: A corpusbased study. Paper presented at the 5th International Conference on Maltese
Linguistics, University of Turin, June 24–26.
Borg, Alexander. 1978. A historical and comparative phonology and morphology
of Maltese. Jerusalem: The Hebrew University of Jerusalem. (Doctoral dissertation).
Borg, Alexander. 1985. Cypriot Arabic: A historical and comparative investigation
into the phonology and morphology of the Arabic vernacular spoken by the
Maronites of Kormakiti village in the Kyrenia district of north-west Cyprus.
Stuttgart: Deutsche Morgenländische Gesellschaft.
Borg, Alexander. 1994. Some evolutionary parallels and divergences in Cypriot
Arabic and Maltese. Mediterranean Language Review 8. 41–67.
Borg, Alexander. 1996. On some Levantine linguistic traits in Maltese. Israel
Oriental Studies 16. 133–152.
Borg, Alexander. 1997. Maltese phonology. In Alan S. Kaye (ed.), Phonologies of
Asia and Africa, vol. 1, 245–285. Winona Lake, IN: Eisenbrauns.
Borg, Alexander. 2004. A comparative glossary of Cypriot Maronite Arabic (Arabic–
English): With an introductory essay. Leiden: Brill.
296
13 Maltese
Bovingdon, Roderick & Angelo Dalli. 2006. Statistical analysis of the source
origin of Maltese. In Andrew Wilson, Dawn Archer & Paul Rayson (eds.),
Corpus linguistics around the world, 63–76. Amsterdam: Rodopi.
Brincat, Joseph M. 1995. Malta 870–1054: Al-Ḥimyarī’s account and its linguistic
implications. Valetta: Said International.
Brincat, Joseph M. 1996. Maltese words: An etymological analysis of the Maltese
lexicon. In Jens Lüdtke (ed.), Romania Arabica. Festschrift für Reinhold Kontzi
zum 70. Geburtstag, 111–115. Tübingen: Gunter Narr.
Brincat, Joseph M. 2011. Maltese and other languages: A linguistic history of Malta.
Malta: Midsea Books.
Brincat, Joseph M. & Manwel Mifsud. 2015. Maltese. In Peter O. Müller, Ingeborg
Ohnheiser, Susan Olsen & Franz Rainer (eds.), Word-formation: An international handbook of the languages of Europe, 3349–3366. Berlin: De Gruyter
Mouton.
Camilleri, Maris. 2011. On pronominal verbal enclitics in Maltese. In Sandro
Caruana, Ray Fabri & Thomas Stolz (eds.), Variation and change: The dynamics
of Maltese in space, time and society, 131–156. Berlin: Akademie Verlag.
Camilleri, Maris & Louisa Sadler. 2019. The grammaticalisation of a copula in
vernacular Arabic. Glossa 4(1: 137). 1–33.
Caubet, Dominique. 2011. Moroccan Arabic. In Lutz Edzard & Rudolf de Jong
(eds.), Encyclopedia of Arabic language and linguistics, online edn. Leiden: Brill.
Coghill, Eleanor. 2014. Differential object marking in Neo-Aramaic. Linguistics
52(2). 335–364.
Cohen, David. 1966. Le système phonologique du maltais: Aspects synchroniques
et diachroniques. Journal of Maltese Studies 3. 1–26.
Cohen, David. 1970. Le système phonologique du maltais: Aspects synchroniques
et diachroniques. In David Cohen (ed.), Etudes de linguistique arabe et sémitique, 126–149. The Hague: Mouton.
Colin, Georges S. 1957. Mots «berbères» dans le dialecte arabe de Malte. In
Mémorial André Basset (1895–1956), 7–16. Paris: Adrien Maisonneuve.
Comrie, Bernard & Michael Spagnol. 2016. Maltese loanword typology. In Gilbert
Puech & Benjamin Saade (eds.), Shifts and patterns in Maltese, 315–329. Berlin:
De Gruyter Mouton.
Čéplö, Slavomír. 2014. An overview of object reduplication in Maltese. In Albert
Borg, Sandro Caruana & Alexandra Vella (eds.), Perspectives on Maltese linguistics, 201–222. Berlin: De Gruyter.
Čéplö, Slavomír. 2018. Constituent order in Maltese: A quantitative analysis.
Prague: Charles University. (Doctoral dissertation).
297
Christopher Lucas & Slavomír Čéplö
de Cat, Cécile. 2010. French dislocation: Interpretation, syntax, acquisition. Oxford:
Oxford University Press.
de Sacy, Silvestre. 1829. [no title]. Journal des Savants 4. 195–204.
de Soldanis, Giovanni Pietro Francesco Agius. 1750. Della lingua punica presentemente usata da Maltesi. Rome: Generoso Salomoni.
Dimitrakou, Dimitriou. 1958. Méga leksikón tés Ellénikés glósses. Athens: Ekdoseis
Domi.
Döhla, Hans-Jörg. 2016. The origin of differential object marking in Maltese. In
Gilbert Puech & Benjamin Saade (eds.), Shifts and patterns in Maltese, 149–174.
Berlin: De Gruyter Mouton.
Drewes, A. J. 1994. Borrowing in Maltese. In Peter Bakker & Maarten Mous (eds.),
Mixed languages: 15 case studies in language intertwining, 83–111. Amsterdam:
Institute for Functional Research into Language & Language Use (IFOTT).
Drower, Ethel Stefana & Rudolf Macúch. 1963. A Mandaic dictionary. Oxford:
Clarendon Press.
Fabri, Ray. 1993. Kongruenz und die Grammatik des Maltesischen. Tübingen: Max
Niemeyer.
Fabri, Ray & Albert Borg. 2002. Topic, focus and word order in Maltese.
In Abderrahim Youssi, Fouzia Benjelloun, Mohamed Dahbi & Zakia IraquiSinaceur (eds.), Aspects of the dialects of Arabic today, 354–363. Rabat: AMAPATRIL.
Fenech, Edward. 1978. Contemporary journalistic Maltese. Leiden: Brill.
Ferguson, Charles A. 1964. Baby talk in six languages. American Anthropologist
66(6). 103–114.
Fiorini, Stanley. 1986. The resettlement of Gozo after 1551. Melita Historica 9. 203–
244.
Flege, James Emil, Murray J. Munro & Ian R. A. MacKay. 1995. Effects of age
of second-language learning on the production of English consonants. Speech
Communication 16. 1–26.
Friedman, Victor A. 2008. Balkan object reduplication in areal and dialectological
perspective. In Dalina Kallulli & Liliane Tasmowski (eds.), Clitic doubling in the
Balkan languages, 35–63. Amsterdam: John Benjamins.
Gabriel, Christoph & Elena Kireva. 2014. Prosodic transfer in learner and contact
varieties: Speech rhythm and intonation of Buenos Aires Spanish and L2 Castilian Spanish produced by Italian native speakers. Studies in Second Language
Acquisition 36. 257–281.
Gardani, Francesco. 2012. Plural across inflection and derivation, fusion and agglutination. In Lars Johanson & Martine Robbeets (eds.), Copies versus cognates
in bound morphology, 71–97. Leiden: Brill.
298
13 Maltese
Gatt, Albert. 2018. Definiteness agreement and the pragmatics of reference in the
Maltese NP. STUF – Language Typology and Universals 71(2). 175–198.
Gatt, Albert & Ray Fabri. 2018. Borrowed affixes and morphological productivity:
A case study of two Maltese nominalisations. In Patrizia Paggio & Albert Gatt
(eds.), The languages of Malta, 143–169. Berlin: Language Science Press.
Geary, Jonathan. 2017. The historical development of the Maltese plural suffixes
-iet and (i)jiet. Paper presented at the 23rd International Conference on Historical Linguistics, San Antonio, TX, 31 July – 4 August.
Gesenius, Wilhelm. 1810. Versuch über die maltesische Sprache zur Beurtheilung
der neulich wiederhohlten Behauptung, dass sie Ueberrest der altpunischen sey,
und als Beytrag zur arabischen Dialektologie. Leipzig: Fr. Chr. Wilh. Vogel.
Gibson, Maik. 2011. Tunis Arabic. In Lutz Edzard & Rudolf de Jong (eds.), Encyclopedia of Arabic language and linguistics, online edn. Leiden: Brill.
Goodwin, Stefan. 2002. Malta: Mediterranean bridge. Westport, CT: Bergin &
Garvey.
Grech, Paul. 1961. Are there any traces of Punic in Maltese? Journal of Maltese
Studies 1. 130–138.
Grice, Martine, Alexandra Vella & Anna Bruggeman. 2019. Stress, pitch accent,
and beyond: Intonation in Maltese questions. Journal of Phonetics 76. 100913.
Harrell, Richard S. 2004. A short reference grammar of Moroccan Arabic.
Washington, DC: Georgetown University Press.
Haspelmath, Martin & Uri Tadmor (eds.). 2009. Loanwords in the world’s
languages: A comparative handbook. Berlin: De Gruyter Mouton.
Heath, Jeffrey. 2002. Jewish and Muslim dialects of Moroccan Arabic. London:
Routledge.
Herin, Bruno & Martin R. Zammit. 2017. Three for the price of one: The dialects of
Kerkennah (Tunisia). In Veronika Ritt-Benmimoun (ed.), Tunisian and Libyan
Arabic dialects: Common trends – recent developments – diachronic aspects, 135–
146. Zaragoza: Prensas de la Universidad de Zaragoza.
Iemmolo, Giorgio. 2010. Topicality and differential object marking: Evidence
from Romance and beyond. Studies in Language 34(2). 239–272.
Khan, Geoffrey. 1984. Object markers and agreement pronouns in Semitic languages. Bulletin of the School of Oriental and African Studies 43(3). 468–500.
Khan, Geoffrey. 2008. The Neo-Aramaic dialect of Barwar. Leiden: Brill.
Klimiuk, Maciej. 2018. Preliminary remarks on the Gozitan dialect of Għarb,
Malta. Paper presented at the 12th conference of the Association Internationale
de Dialectologie Arabe, Marseille, 30 May – 2 June.
299
Christopher Lucas & Slavomír Čéplö
Krapova, Iliyana & Guglielmo Cinque. 2008. Clitic reduplication constructions
in Bulgarian. In Dalina Kallulli & Liliane Tasmowski (eds.), Clitic doubling in
the Balkan languages, 257–287. Amsterdam: John Benjamins.
Krier, Fernande. 1976. Le maltais au contact de l’italien: Etude phonologique, grammaticale et sémantique. Hamburg: Helmut Buske.
Lahdo, Ablahad. 2009. The Arabic dialect of Tillo in the region of Siirt (south-eastern
Turkey). Uppsala: Uppsala University.
Lipiński, Edward. 1997. Semitic languages: Outline of a comparative grammar.
Leuven: Peeters.
Loporcaro, Michele. 2011. Phonological processes. In Martin Maiden, John
Charles Smith & Adam Ledgeway (eds.), The Cambridge history of the Romance
languages, vol. 1: Structures, 109–154. Cambridge: Cambridge University Press.
Lucas, Christopher. 2015. Contact-induced language change. In Claire Bowern &
Bethwyn Evans (eds.), The Routledge handbook of historical linguistics, 519–536.
London: Routledge.
Maiden, Martin & Cecilia Robustelli. 2007. A reference grammar of modern Italian.
London: Routledge.
Mifsud, Manwel. 1995. Loan verbs in Maltese: A descriptive and comparative study.
Leiden: Brill.
Mifsud, Manwel. 2011. Maltese. In Lutz Edzard & Rudolf de Jong (eds.), Encyclopedia of Arabic language and linguistics, online edn. Leiden: Brill.
Mifsud, Manwel & Albert Borg. 1997. Fuq l-Għatba tal-Malti. Strasbourg: Council
of Europe Publishing.
National Statistics Office. 2014. Census of population and housing 2011: Final report.
Valletta: National Statistics Office.
Naït-Zerrad, Kamal. 2002. Dictionnaire des racines berbères (formes attestées).
Vol. 3: Ḍ–GƐY. Leuven: Peeters.
O’Rourke, Erin. 2005. Intonation and language contact: A case study of two varieties of Peruvian Spanish. Urbana-Champaign: University of Illinois. (Doctoral
dissertation).
Ondráčková, Zuzana. 2010. Komparatívny výskum detskej lexiky. Prešov: Prešovská University.
Piccitto, Giorgio (ed.). 1977. Vocabolario siciliano. Vol. 1: A–E. Catania: Centro di
Studi Filologici e Linguistici Siciliani.
Ritt-Benmimoun, Veronika. 2014. Grammatik des arabischen Beduinendialekts der
Region Douz (Südtunesien). Wiesbaden: Harrassowitz.
Russi, Cinzia. 2008. Italian clitics: An empirical study. Berlin: Mouton de Gruyter.
Saade, Benjamin. 2019. Assessing productivity in contact: Italian derivation in
Maltese. Linguistics 57(1).
300
13 Maltese
Sammut, Carmen. 2007. Media and Maltese society. Plymouth: Lexington Books.
Sciriha, Lydia & Mario Vassallo. 2006. Living languages in Malta. Malta: Print IT
Printing Services.
Seifart, Frank. 2015. Direct and indirect affix borrowing. Language 91. 511–532.
Seifart, Frank. 2017. Patterns of affix borrowing in a sample of 100 languages.
Journal of Historical Linguistics 7. 389–431.
Sokoloff, Michael. 2002. A dictionary of Jewish Babylonian Aramaic of the
Talmudic and Geonic periods. Ramat Gan: Bar Ilan University Press.
Souag, Lameen. 2013. Berber and Arabic in Siwa (Egypt): A study in linguistic
contact. Cologne: Rüdiger Köppe.
Souag, Lameen. 2017. Clitic doubling and language contact in Arabic. Zeitschrift
für Arabische Linguistik 66. 45–70.
Souag, Lameen. 2018. Berber etymologies in Maltese. International Journal of
Arabic Linguistics 4(1). 190–223.
Spagnol, Michael. 2011. A tale of two morphologies: Verb structure and argument
alternations in Maltese. Konstanz: Universität Konstanz. (Doctoral dissertation).
Stolz, Thomas. 2003. Not quite the right mixture: Chamorro and Malti as
candidates for the status of mixed language. In Yaron Matras & Peter Bakker
(eds.), The mixed language debate: Theoretical and empirical advances, 271–315.
Berlin: De Gruyter Mouton.
Sutcliffe, Edmund. 1936. A grammar of the Maltese language: With chrestomathy
and vocabulary. Oxford: Oxford University Press.
Taine-Cheikh, Catherine. 2011. Numerals. In Lutz Edzard & Rudolf de Jong (eds.),
Encyclopedia of Arabic language and linguistics, online edn. Leiden: Brill.
Ullmann, Manfred. 1989. Adminiculum zur Grammatik des klassischen Arabisch.
Wiesbaden: Harrassowitz.
University of Zaragoza (ed.). 2013. A descriptive and comparative grammar of
Andalusi Arabic. Leiden: Brill.
Van Coetsem, Frans. 1988. Loan phonology and the two transfer types in language
contact. Dordrecht: Foris.
Van Coetsem, Frans. 2000. A general and unified theory of the transmission process
in language contact. Heidelberg: Winter.
Vanhove, Martine. 1993. La langue maltaise: Etudes syntaxiques d’un dialecte arabe
«périphérique». Wiesbaden: Harrassowitz.
Vanhove, Martine. 1998. De quelques traits pré-hilaliens en maltais. In Jordi
Aguadé, Patrice Cressier & Ángeles Vicente (eds.), Peuplement et arabisation
au Maghreb occidental: Dialectologie et histoire, 97–108. Zaragoza: Casa de
Velázquez.
301
Christopher Lucas & Slavomír Čéplö
Vassalli, Michelantonio. 1791. Mylsen phoenico-punicum sive grammatica
melitensis. Rome: Antonio Fulgoni.
Vella, Alexandra. 1994. Prosodic structure and intonation in Maltese and its
influence on Maltese English. Edinburgh: University of Edinburgh. (Doctoral
dissertation).
Vella, Alexandra. 2003. Language contact and Maltese intonation: Some parallels
with other language varieties. In Kurt Braunmüller & Gisella Ferraresi (eds.),
Aspects of multilingualism in European language history, 261–283. Amsterdam:
John Benjamins.
Vella, Alexandra. 2009. On Maltese prosody. In Bernard Comrie, Ray Fabri,
Elizabeth Hume, Manwel Mifsud, Thomas Stolz & Martine Vanhove (eds.),
Introducing Maltese linguistics, 47–68. Amsterdam: John Benjamins.
Versteegh, Kees. 2014. Pidgin verbs: Infinitives or imperatives. In Isabelle
Buchstaller, Anders Holmberg & Mohammad Almoaily (eds.), Pidgins
and Creoles beyond Africa–Europe encounters, 141–170. Amsterdam: John
Benjamins.
Woidich, Manfred. 2006. Das Kairenisch-Arabische: Eine Grammatik. Wiesbaden:
Harrassowitz.
Zagona, Karen. 2002. The syntax of Spanish. Cambridge: Cambridge University
Press.
Zammit, Martin R. 2014. The Sfaxi (Tunisian) element in Maltese. In Albert Borg,
Sandro Caruana & Alexandra Vella (eds.), Perspectives on Maltese linguistics,
23–44. Berlin: Akademie Verlag.
302
Chapter 14
Arabic in the diaspora
Luca D’Anna
Università degli Studi di Napoli “L’Orientale”
This paper offers an overview of contact-induced change in diasporic Arabic. It provides a socio-historical description of the Arab diaspora, followed by a sociolinguistic profile of Arabic-speaking diasporic communities. Language change is analyzed
at the phonological, morphological, syntactic and lexical level, distinguishing between contact-induced change and internal developments caused by reduced input
and weakened monitoring. In the course of the description, parallels are drawn between diasporic Arabic and other contemporary or extinct contact varieties, such
as Arabic-based pidgins and Andalusi Arabic.
1 Current state and historical development
The terms Arabic in the diaspora and Arabic as a minority language have
been used to designate two distinct linguistic entities, namely Arabic Sprachinseln outside the Arabic-speaking world and Arabic in contemporary migration
settings. The two situations correspond to the two major social processes that
give rise to language contact: conquest and migration. In the former case, speakers of Arabic were isolated from the central area in which the Arabic language is
spoken, exposed to a different dominant language, and consequently underwent
a slow process of language erosion (and eventually shift) usually spanning across
several generations. This situation often gives rise to long periods of relatively
stable bilingualism, where contact-induced change is more noticeable (Sankoff
2001: 641). In migration contexts, on the contrary, language shift occurs at a faster
pace, sometimes within the lifespan of the first generation and usually no later
than the third (Canagarajah 2008: 151).
This chapter analyzes contact-induced change in migration contexts. Arab migration to the West started in the late nineteenth century, with the first wave of
Luca D’Anna. 2020. Arabic in the diaspora. In Christopher Lucas & Stefano Manfredi
(eds.), Arabic and contact-induced change, 303–320. Berlin: Language Science Press.
DOI:10.5281/zenodo.3744527
Luca D’Anna
migrants who left Greater Syria to settle in the United States and Latin America.
The first migrants were mostly Christian unskilled workers, followed by more
educated Lebanese, Palestinians, Yemenis and Iraqis after World War II. During
the 1950s and 1960s, more migrants continued to settle in the US, while the unstable political situations in Palestine, Lebanon and Iraq resulted in a fourth wave in
the 1970s and 1980s (Rouchdy 1992a: 17–18). Because of the events that took place
during the last two decades and that resulted in a further destabilization of the
entire Middle East, immigration toward the US has never stopped, even though
recent American policies have considerably reduced the intake of refugees and
immigrants. In 2016, however, 84,995 refugees were resettled in the US, with two
Arabic-speaking countries (Syria and Iraq) featuring among the top five states
that make 70% of the total intake.1
Large-scale migration to western Europe from Arabic-speaking countries began in the wake of the decolonization process during the 1960s and mainly involved speakers from North Africa (Morocco, Algeria and Tunisia). Following a
common trend in labor migration, men arrived first, followed by their wives and
children. In 1995, a total of 1,110,545 Moroccans, 655,576 Algerians and 279,813
Tunisians lived in Europe, mostly in France, the Netherlands, Belgium, Germany
and Italy (Boumans & de Ruiter 2002: 259–260). The socioeconomic profile of the
first immigrants mainly consisted of unskilled laborers, usually with low education rates. After six decades from the first wave of immigration, however, most
communities consist today of a first, second and third generation, while the political upheaval which started at the end of 2010 resulted in a new wave of young
immigrants. Both old and new immigrants had to face the economic crisis that
hit Europe in the early 1990s and, again, in 2007, with particularly harsh consequences for the immigrant population (Boumans & de Ruiter 2002: 261).
The sociolinguistic profile of Arabic-speaking communities in the diaspora is
quite diverse in different parts of the world and can be analyzed using the ethnolinguistic vitality framework, according to which status, demographics, and institutional support shape the vitality of a linguistic minority (Giles et al. 1977; Ehala
2015). Arabic-speaking immigrants do not usually enjoy a particularly high status, while the level of institutional support is variable. The first waves of immigration to the US, for instance, had to face an environment that was generally hostile
to foreign languages. The English-only movement actively worked to impose the
exclusive employment of English in public places, while the immigrants themselves committed to learning and using English to integrate into mainstream
1
Data come from the US Department of State. https://www.state.gov/j/prm/releases/factsheets/
2017/266365.htm, accessed April 2, 2019.
304
14 Arabic in the diaspora
American life. Only in the aftermath of 9/11 did American policymakers begin to
re-evaluate the importance of Arabic (and other heritage languages), considering it a resource for homeland security (Albirini 2016: 319–320). Other countries,
such as the Netherlands, provided higher levels of formal institutional support,
including Arabic in school curricula. These efforts did not achieve the desired
goals, however, mostly because the great linguistic diversity of the Moroccan
community living in the Netherlands cannot be adequately represented in the
teaching curricula. Moroccans in the Netherlands, in fact, speak different Arabic
dialects, alongside three main varieties of Berber, namely Tashelhiyt, Tamazight
and Tarifiyt (Extra & de Ruiter 1994: 160–161). The voluntary home language instruction program, however, provides instruction in Modern Standard Arabic,
even though writing skills are only taught starting from third grade (Extra & de
Ruiter 1994: 163–165). This is not, of course, the language students are exposed
to at home, but attempts to introduce Moroccan dialect or Berber are generally
opposed by parents, who value Classical Arabic for its religious and cultural relevance. Similar Home Language Instruction programs are found in most European countries, even though their implementation is sometimes carried out by
local governments (in the Netherlands and Germany), private organizations (in
Spain) or even by the governments of the origin country (in France) (Boumans
& de Ruiter 2002: 264–265). The Italian town of Mazara del Vallo in Sicily represents an extreme case, since the members of the Tunisian community obtained
from the Tunisian government the opening of a Tunisian school, where a complete Arabic curriculum is offered and Italian is not even taught as a second
language. Until the end of the 1990s, this school, opened in 1981, was the first
choice for Tunisian families, who hoped for a possible return to Tunisia. When
it eventually became clear that this was unlikely to happen, enrollments consequently declined, which means that Arabic teaching is no longer available to the
community in any form (D’Anna 2017a: 73–77). Issues of diglossia and language
diversity thus undermine Home Language Instruction programs, which usually
occupy a marginal position within school curricula.
Given the generally low status of, and insufficient institutional support for,
Arabic-speaking communities in the diaspora, demographic factors are often decisive in determining the ethnolinguistic vitality of the community. While speakers of Arabic are usually scattered in large areas where the dominant language is
prevalently spoken, in some Dutch towns Moroccan youth make up 50% of the
population of certain neighborhoods (Boumans 2004: 50). At the other end of the
continuum, we find closely-knit communities, living in the same neighborhood,
such as in Mazara del Vallo, where Tunisians hailing from the two neighboring towns of Mahdia and Chebba constitute up to 70% of the population of the
305
Luca D’Anna
old town (D’Anna 2017a: 27). All things being equal, given the low status of the
Tunisian community and the mediocre institutional support they receive, it is primarily demographic factors which have resulted in the preservation of Arabic in
this community beyond the threshold of the third generation.2
In the light of what has been said above, and despite some notable exceptions,
Arabic diasporic communities are characterized by relatively rapid processes
of language shift, both in the US (Daher 1992: 29) and in Europe (Boumans &
de Ruiter 2002: 282). This means that the processes of contact-induced change
observed in diasporic communities of Arabic are generally the prelude to language loss. The importance of studying language change in migrant languages,
however, also resides in the fact that the same changes usually take place, at a
much slower rate, in the standard spoken in the homeland. Internally motivated
change in diasporic varieties, from this perspective, often represent an accelerated version of language change in the homeland. Contact-induced change, on
the other hand, sometimes suggests parallels with the socially different process
of pidginization (Gonzo & Saltarelli 1983: 194–195). The study of Arabic-speaking
diasporic communities, thus, can help us shed light on the more general evolution
of the language, with regard to both contact-induced and internally-motivated
change.
2 Contact languages
Contact languages for diasporic Arabic-speaking communities include, but are
not restricted to, American (Rouchdy 1992b) and British English (Abu-Haidar
2012), Portuguese in Brazil (Versteegh 2014: 292), French (Boumans & Caubet
2000), Dutch (Boumans 2000; 2004; 2007; Boumans & Caubet 2000; Boumans
& de Ruiter 2002), Spanish (Vicente 2005; 2007) and Italian (D’Anna 2017a; 2018).
Some contact situations are better described than others, as in the case of English,
French and Dutch. At the other end of the continuum, research on the outcome of
contact between Italian and Arabic is extremely recent, and data on Portuguese
are scarce.
In the following sections, we will draw from the sources so far cited to describe
the main phenomena of language change occurring in diasporic Arabic at the
phonological, morphological, syntactic and lexical level, highlighting possible
parallels with comparable changes in other non-diasporic varieties of Arabic.
2
Other factors also played a minor role in the preservation of Arabic in Mazara del Vallo (D’Anna
2017a: 80–81).
306
14 Arabic in the diaspora
3 Contact-induced changes in diasporic Arabic
Despite the great variety of contact languages, it is possible to individuate a number of phenomena that predictably occur in diasporic Arabic-speaking communities. It is not always easy, however, to assess whether an individual phenomenon
is due to contact or whether it is, on the contrary, the result of internal development (Romaine 1989: 377). Gonzo & Saltarelli (1977: 177) put the matter as follows:
While it seems clear that some types of changes are due to interference
from the dominant language, and others may be attributable to sociological
and other external pressures, there are some changes which are languageinternal. The latter type is in accordance with a principle of regularization
and code reduction which one might expect when the language is acquired
in a weakly monitored sociolinguistic environment.
The concept of weakened monitoring, a situation in which a generally accepted standard and the reinforcement of correct norms are lacking, is an effective tool of analysis when investigating language change in diasporic communities (Gonzo & Saltarelli 1977; 1983). In a situation of weakened monitoring,
processes of language change that are occurring slowly in other varieties of the
language can be sped up.
In the following sections, interference between languages will be referred to
as transfer, which occurs from the source language (SL) to the recipient
language (RL). If the speaker is dominant in the SL, transfer is more specifically
defined as imposition. If, on the contrary, the speaker is dominant in the RL,
transfer is defined as borrowing (Van Coetsem 1988; 2000; Lucas 2015). While
the concept of linguistic dominance will be extensively used in this paper, one final caveat concerns the difficulty of individuating the dominant language (which
may actually shift) in second-generation speakers. Lucas identifies a category of
2L1 speakers, who undergo the simultaneous acquisition of two distinct native
languages (Lucas 2015: 525). The linguistic trajectory of most second-generation
speakers, however, usually involves two consecutive stages in which first the
heritage and then the socially dominant language function as the dominant language. While the heritage language is almost exclusively spoken at home during
early childhood, in fact, second-generation speakers gradually shift to the socially dominant language when they start school and consequently expand their
social network.
307
Luca D’Anna
3.1 Phonology
In the domain of phonology, diasporic varieties of Arabic generally go in the direction of the loss of marked phonemes (Versteegh 2014: 293). It is generally the
emphatic and post-velar phonemes that undergo erosion, though the loss is usually not systematic, featuring a great deal of inter and intra-individual variation.
In non-diasporic communities, adults, peers and institutions provide corrective
feedback to children during their process of language acquisition, while in immigrant communities, due to the weakened monitoring mentioned above, the
chain of intergenerational transmission is less secure. Some phenomena of phonetic loss thus have a developmental origin, and are equally common in pidgins
and dying languages (Romaine 1989: 372–373). Consider the following example:
(1)
Tunisian Arabic, Mazara del Vallo (D’Anna 2017a: 85)
ʕala ḫāṭr-i
ʕarbi u nnəžžəm
naʕrəf
aktər wāəd
on thought-obl.1sg Arab and can.impf.1sg know.impf.1sg more one
mia
lingua
poss.1sg.f language
‘Because I’m an Arab and I can know above all my language.’
The speaker in sample (1) realizes the voiced pharyngeal fricative /ʕ/, one of the
phonemes that are usually lost, but then fails to realize its voiceless counterpart
/ḥ/ in wāəd < wāḥəd ‘one’.3 Similar phenomena also occur, as noted above, in
Arabic-based pidgins and creoles, such as Juba Arabic (Manfredi 2017: 17, 21; cf.
Avram, this volume).
In the process of phonological erosion, therefore, contact languages seem to
have a limited impact. If the dominant language does not feature, in its phonemic inventory, the phoneme that is being eroded, it fails to reinforce whatever
input young bilingual speakers receive in the other L1 in the contexts of primary
socialization. Reduced input and weakened monitoring, however, play a bigger
role, allowing forms usually observed in the earliest stages of language acquisition by monolingual children to survive and spread. It is relatively common,
for instance, to observe the presence of shortened or reduced forms, such as qe
< lqe ‘he found’, ḥal < nḥal ‘bees’, ləd < uləd ‘kid’, which sometimes give rise
to phenomena of compensation, such as in uləd > ləd > lədda ‘kid’ (Tunisian
3
Similar phenomena of phonetic simplification occur in peripheral varieties of Arabic and
Sprachinseln, such as Nigerian Arabic (Owens 1993: 19–20; this volume), Cypriot Maronite
Arabic (Borg 1985; Walter, this volume), Uzbekistan Arabic (Seeger 2013) and Maltese (Borg
& Azzopardi-Alexander 1997: 299; Lucas & Čéplö, this volume). The single varieties here mentioned vary with regard to the phonological simplification they underwent.
308
14 Arabic in the diaspora
diasporic Arabic, Mazara del Vallo, Italy; D’Anna 2017a: 85). In diasporic communities, reduced forms are more easily allowed to survive and spread, occurring in
the speech of teenagers, as in the examples reported here. Once again, the same
phenomenon also occurs in pidgin and dying languages:
In the case of dying and pidgin languages it may be that children have greater
scope to act as norm-makers due to the fact that a great deal of variability exists
among the adult community (Romaine 1989: 372–373).
In conclusion, the phonology of diasporic Arabic does not seem to be heavily influenced by borrowing from contact languages. The combined action of
reduced input and weakened monitoring, on the other hand, is responsible for
the unsystematic loss of marked phonemes and for the survival and spread of
reduced forms.
3.2 Morphology
The complex mixture of concatenative and non-concatenative morphology in the
domain of Arabic plural formation has been one of the main focuses of research in
situations of language contact resulting from migration. Once again, borrowing
from contact languages and independent developments occur side by side.
In Arabic, both concatenative and non-concatenative morphology contribute
to plural formation. Concatenative morphology, which consists in attaching a
suffix to the singular noun, yields the so-called sound plurals, that is, in spoken Arabic, the plural suffixes -īn and -āt respectively. It has been argued that
sound feminine plural is the default plural form according to the morphological
underspecification hypothesis, even though masculine is the default gender in
all other domains of plural morphology (Albirini & Benmamoun 2014: 855–856).
While sound masculine plural is specified for [+human], in fact, sound feminine
plural has the semantic feature [±human]. Non-concatenative, or broken, plurals require a higher cognitive load, since they involve the mapping of a vocalic
template onto a consonantal root.4 Sound feminine plurals are acquired by children by the age of three, while broken plurals involving geminate and defective
roots are not mastered until beyond the age of six (Albirini & Benmamoun 2014:
857–858). After the age of five, however, heritage speakers of Arabic become increasingly exposed to their L2, which encroaches upon their acquisition of broken plurals. It has thus been convincingly demonstrated that heritage speakers
display a better command of sound plurals and that, in the domain of broken
4
The notion of root and pattern, which has long been at the core of the morphology of Arabic, has recently been criticized (Ratcliffe 2013), even though psycholinguistic studies seem to
confirm the existence of the root in the mental lexicon of native speakers (Boudelaa 2013).
309
Luca D’Anna
plurals, some are more affected by language erosion than others (Albirini & Benmamoun 2014: 858–859). Across different varieties of diasporic Arabic, therefore,
plural morphology displays both contact phenomena due to borrowing and internal developments that are akin to what might be called restructuring, that
is:
changes that a speaker makes to an L2 that are the result not of imposition
but of interpreting the L2 input in a way that a child acquiring an Ll would
not (Lucas 2015: 525).5
Borrowing from the contact languages can take two forms. In rare cases, the
suffix plural morpheme of the contact language is directly borrowed, as in the
examples ḥuli-s ‘sheep-PL’, ḥmar-s ‘donkeys’ and l-ʕud-s ‘the horses’6 collected
from one Moroccan informant in the Netherlands (Boumans & de Ruiter 2002:
274). Sometimes, however, transfer works in a subtler way, which consists in the
generalization of the sound masculine plural suffix -īn,7 by analogy with the default form of the contact language, yielding ḥul-in ‘sheep-PL’, ḥmār-in ‘donkeys’,
ʕewd-in ‘horses’ (Boumans & de Ruiter 2002: 274). A study conducted by Albirini
& Benmamoun (2014: 866–867) shows that L2 learners of Arabic usually tend to
overgeneralize the sound masculine plural, wrongly perceived as a default form,
while heritage speakers more often resort to the Arabic-specific default, i.e. sound
feminine plural. The cases of borrowing reported above, therefore, represent an
idiosyncratic exception.
On the other hand, the non-optimal circumstances under which Arabic is
learned in diasporic communities often result in overgeneralization processes
that cannot be directly attributed to contact. One of them is, as noted above, the
generalization of the sound feminine plural -āt. In the domain of broken plurals,
moreover, not all patterns are equally distributed. The iambic pattern, consisting of a light syllable followed by one with two moras (CVCVVC), is the most
common among Arabic broken plurals (Albirini & Benmamoun 2014: 857). As a
consequence, it is often generalized by heritage speakers of Levantine varieties
(Syrian, Lebanese, Palestinian and Jordanian) living in the US, yielding forms
such as: fallāḥ ‘farmer’, pl. aflāḥ/fulūḥ (target plural fallāḥ-īn); šubbāk ‘window’,
5
In this case, of course, the speaker would not be re-interpreting an L2, but an L1 learned under
reduced input conditions and subject to language erosion.
6
The target form here is ʕewd-an, so that also vowel quality is not standard.
7
The suffix for masculine plural -īn is realized with a short vowel in the diasporic Moroccan
varieties that are being discussed.
310
14 Arabic in the diaspora
pl. šubūk (target plural šabābīk); ṭabbāḫ ‘cook’, pl. ṭabāʔiḫ (target plural ṭabbāḫīn) (Albirini & Benmamoun 2014: 865).8
Borrowing does not involve plural morphemes only, but other classes as well.
In Mazara del Vallo, for instance, young speakers occasionally use the Sicilian
diminutive morpheme -eddru with Arabic names, creating morphological hybrids of the kind illustrated in (2):
(2)
Tunisian Arabic, Mazara del Vallo (D’Anna 2017a: 107)
Grazie safwani-ceddruu9
thanks Safwan-dim
‘Thanks little Safwan.’
This type of borrowing, quite widespread among young speakers, seems to
replicate another instance of contact-induced change that occurred in an extinct variety of Arabic. Andalusi Arabic, in fact, borrowed from Romance the
diminutive morpheme -el (e.g. tarabilla ‘mill-clapper’ < ṭarab+ella ‘little music’),
incidentally etymologically cognate with the Sicilian -eddru (Latin -ellum > Sicilian -eddru/-eddu) (University of Zaragoza 2013: 60). The behavior of the young
Tunisian speakers of Mazara del Vallo, who use these Sicilian diminutives in a
playful mode, might represent the first stage of the same process that resulted in
in the transfer of this morpheme into Andalusi Arabic (D’Anna 2017a: 108).
While plurals represent one of the most common areas of change in diasporic
Arabic, morpheme borrowing is a much rarer phenomenon, which probably occurs in situations of more pronounced bilingualism. The above two examples,
however, provide a representative exemplification of the effect of language contact in the domain of morphology.
3.3 Syntax
Borrowing and restructuring also happen in the domain of syntax. As has been
noted both for Moroccans in the Netherlands (de Ruiter 1989: 99) and Tunisians in
Italy (personal research), second-generation speakers tend to use simpler clauses
than monolingual speakers, namely main or subordinate clauses to which no
other clause is attached, as evident from the following sample:
8
The overgeneralization of some broken plural patterns indicates that the root and pattern system is still productive in heritage speakers, as opposed, for instance, to speakers of Arabicbased pidgins and creoles. Recent studies, however, have advanced the hypothesis that the
iambic pattern involves operations below the level of the word, but without necessarily entailing the mapping of a template onto a consonantal root (Albirini et al. 2014: 112).
9
The utterance appeared as a Facebook post in the timeline of one of my informants and was
transcribed verbatim.
311
Luca D’Anna
(3) Tunisian Arabic, Mazara del Vallo (personal research)
m-baʕd əl-uləyyəd rqad
u l-kaləb zāda u l-žrāna
from-after def-boy.dim sleep.prf.3sg.m and def-dog also and def-frog
ḫaržət
mən əl-wāḥəd ēh
dabbūsa
exit.prf.3sg.f from def-one hesit bottle
‘Then the little boy slept and also the dog and the frog escaped from the
hum bottle.’
Accordingly, they also display the effects of language erosion in establishing
long-distance dependencies typical of more complex clauses (Albirini 2016: 305).
Palestinian and Egyptian speakers born in the US have also been found to realize overt pronouns in sentences that opt for the pro-drop strategy in the speech
of monolinguals, which is probably due to the influence of English (Albirini et al.
2014: 283). Preliminary observations on second-generation Tunisians in Italy, in
fact, do not show the same phenomenon. Since Italian is, like Arabic, a pro-drop
language, the use of overt pronouns in American diasporic Arabic can be considered as a case of syntactic borrowing or convergence (Lucas 2015), depending on
the speakers’ degree of bilingualism.
The syntax of negation is another area in which language erosion triggers
phenomena that seem to be happening, albeit at a slower rate, in non-diasporic
communities. Egyptian speakers in the US, for instance, seem to overgeneralize
the monopartite negatior miš/muš at the expense of the default discontinuous
verbal negator ma…-š:
(4)
Egyptian Arabic in the US (Albirini & Benmamoun 2015: 482)
huwwa miš rāḥ
l-kaftiria
3sg.m neg go.prf.3sg.m to-cafeteria
‘He didn’t go to the cafeteria.’
Example (4) represents a deviation from the standard Cairene dialect spoken
by monolinguals. In Egypt, however, the negative copula miš~muš represents a
pragmatically marked possibility to negate the b- imperfect (Brustad 2000: 302),
while in Cairo it is now the standard negation for future tense (miš ḥa-…, contrasting with ma-ḥa-…-š in some areas of Upper Egypt (Brustad 2000: 285). More
generally, therefore, miš~muš is gaining ground at the expense of the discontinuous negation (Brustad 2000: 285), so that what we observe in diasporic Egyptian
Arabic might just be an accelerated instance of the same process.
Another major area of language change, documented in most diasporic languages, is the erosion of complex agreement systems (Gonzo & Saltarelli 1983:
312
14 Arabic in the diaspora
192). In diasporic Arabic, heritage speakers show relatively few problems with
subject–verb agreement, but struggle with the subtleties of noun–adjective agreement (Albirini et al. 2013: 8). While subject–verb agreement involves a verbal
paradigm with a relatively large number of cells, it is nevertheless simpler than
noun–adjective agreement, since plural nouns can trigger adjective agreement
in the sound or broken plural or in the feminine singular, depending on factors
involving humanness, individuation, and the morphological shape of both the
noun and the adjective, with marked dialectal variation (D’Anna 2017b: 103–104).
Heritage speakers thus perform significantly better when default agreement in
the masculine singular is required (Albirini et al. 2013: 8), but display evident
signs of language erosion when more complex structures are involved:
(5)
Egyptian Arabic in the US (Albirini 2014: 740)
wi-kamān baḥibb
arūḥ
l-Detroit ʕašān ʕinda-ha
and-also love.impf.ind.1sg go.impf.1sg to-Detroit because at-3sg.f
maṭāʕim
*mumtaz-īn
restaurant.pl excellent-pl.m
‘And I also like to go to Detroit because it has excellent restaurants.’
In (5), the speaker selects the sound masculine plural, while non-human plural nouns require either the broken plural or the feminine singular in Egyptian
Arabic. Once again, language change in diasporic Arabic, where the language is
learned under reduced input conditions, tends to replicate processes of language
change that happened or are happening in the Arabic-speaking world. In the
case of agreement, the standardization that the agreement system underwent
in the transition from pre-Classical to Classical Arabic has been convincingly
explained as emerging from the overgeneralization of frequent patterns by L2
learners (Belnap 1999).
Finally, isolated cases show syntactic borrowing or convergence10 at the level
of word order, which is usually preserved in diasporic contexts, as in the example
in (6).
(6) Moroccan Arabic in the Netherlands (Boumans 2001: 105)
u ʕṭat
l-u
dyal-u
l-lḥem
and give.prf.3sg.f to-3sg.m gen-3sg.m def-meat
‘And she gave it [i.e. the dog] its meat.’
10
Once again, considering this phenomenon as syntactic borrowing or convergence depends
on the speaker’s language dominance, which is not clear from the source and is not easily
ascertained in second-generation speakers, whose dominant language is often subject to shift.
313
Luca D’Anna
This example illustrates an extreme case of word order change, in which the
possessive dyal-u ‘its’ precedes the head. Overgeneralization of permissible (but
sometimes pragmatically marked) word orders, however, occur much more frequently. Egyptian heritage speakers in the US, for instance, use SVO order 77.65%
of the time, vs. 52.64% for Egyptian native speakers (Albirini et al. 2011: 280–281).
In situations of stable bilingualism, such as in some Arabic Sprachinseln, convergence with contact languages can result in permanent alterations to word
order. In Buxari Arabic, for instance, transitive verbs feature a mandatory SOV
word order, with optional resumptive pronoun after the verb. Cleft sentences
such as the following one are quite common in all Arabic dialects:
(7)
Egyptian Arabic (Ratcliffe 2005: 145)
il-fustān gibt-u
def-dress get.prf.1sg-3sg.m
‘I got the dress.’
In Bukhari Arabic, which has long been in contact with SOV languages (such
as Persian and Tajik), this structure became the standard for transitive verbs, so
that the resumptive pronoun can also be dropped, as in the following sample:
(8)
Bukhari Arabic (Ratcliffe 2005: 144)
fāt
ʕūd ḫada
indef stick take.prf.3sg.m
‘He took a stick.’
3.4 Lexicon
In the domain of lexical borrowing, which has attracted considerable interest
among scholars, the situation of bilingualism in diasporic contexts poses some
methodological issues in the individuation of actual loanwords. The production
of heritage speakers, in fact, is inevitably marked by frequent phenomena of
code-switching, which makes difficult to distinguish between nonce-borrowings
(Poplack 1980) and code-switching. If we define lexical borrowing as “the diachronic process by which languages enhance their vocabulary” (Matras 2009:
106), in fact, it is not clear which language is here enhancing its vocabulary, since
diasporic varieties of Arabic are not discrete varieties and feature the highest degree of internal variability. A possible solution to this impasse consists in looking exclusively at the linguistic properties of the alleged loanword. In this vein,
Adalar & Tagliamonte (1998: 156) have shown that, when foreign-origin nouns appear in contexts in which they are completely surrounded by the other language,
they are treated like borrowings (in this case, nonce-borrowings) at the phonological, morphological and syntactic level. When, on the contrary, they appear
314
14 Arabic in the diaspora
in bilingual (or multilingual) utterances, they represent cases of code-switching,
patterning with the language of their etymology. The domain of lexical borrowing in diasporic varieties of Arabic, however, is an area that needs further research.
4 Conclusion
This chapter has offered an overview of the main phenomena of contact-induced
change observed in Arabic diasporic communities, distinguishing them from internal developments due to reduced input and weakened monitoring. Diasporic
communities rarely feature situations of stable bilingualism, so that language
change usually corresponds to language attrition and is followed by the complete shift to the dominant language. The study of language change in diasporic
communities, however, constitutes an interesting field of investigation, both in
itself and for the insight it can give us into language change in monolingual communities. Change at the phonological, morphological and syntactic level finds
parallels in comparable phenomena that have occurred in the history of Arabic
(such as in the case of agreement) or that are occurring as we speak (such as in
the case of the spread of the negator miš in Egyptian Arabic). Not by chance, similar phenomena also occur(red) in the Arabic-based pidgins of East Africa, such
as Juba Arabic. Various scholars, in fact, have maintained that the mechanisms
of change differ in the degree of intensity, but not in their intrinsic nature, from
those operating in less extreme situations of contact (e.g. Miller 2003: 8; Lucas
2015: 528).
On the other hand, the analysis of contact phenomena in diasporic communities poses some methodological issues with regard to the categories of borrowing,
imposition and convergence (Van Coetsem 1988; 2000). These categories, in fact,
imply the possibility to define clearly the speaker’s dominant language or, at
least, to define him as a stable 2L1 speaker. This is rarely the case with heritage
speakers, whose repertoires follow trajectories in which language dominance
shifts, usually from the heritage language to the socially dominant one. This process is usually concomitant with the beginning of school education, but we lack
theoretical and methodological tools to determine with accuracy the speaker’s
position on the trajectory.
Further avenues of research on this topic thus include a more rigorous investigation of emerging and shifting repertoires and the analysis of the complex relation between diasporic languages, pidginization and creolization, which has
already been the object of a number of contributions (e.g. Gonzo & Saltarelli
1983; Romaine 1989).
315
Luca D’Anna
Further reading
) Rouchdy (1992b) is the first description of Arabic in the US.
) Rouchdy (2002) analyzes more broadly language contact and conflict, with a
section devoted to Arabic in the diaspora.
) Owens (2000) collects essays on Arabic as a minority language, focusing on
both Spracheninseln and diasporic Arabic, but introducing also historical and
cross-ethnic perspectives.
Acknowledgements
I am grateful to the University of Mississippi, which generously funded this research and my fieldwork in Mazara del Vallo. To Adam Benkato, for reading the
manuscript and providing, as always, his valuable feedback. To all my informants
in Mazara del Vallo, whose patience during the interviews was only matched by
their warm hospitality.
Abbreviations
1, 2, 3
def
dim
f
hesit
gen
impf
ind
1st, 2nd, 3rd person
definite article
diminutive
feminine
hesitation
genitive
imperfect (prefix conjugation)
indicative
indef
m
neg
obl
pl
poss
prf
sg
indefinite
masculine
negative
oblique
plural
possessive pronoun
perfect (suffix conjugation)
singular
References
Abu-Haidar, Farida. 2012. Arabic and English in conflict: Iraqis in the UK. In Aleya
Rouchdy (ed.), Language contact and language conflict in Arabic: Variations on
a sociolinguistic theme, 286–297. London: Routledge.
Adalar, Nevin & Sali Tagliamonte. 1998. Borrowed nouns, bilingual people:
The case of the “Londrali” in Northern Cyprus. International Journal of
Bilingualism 2. 139–159.
Albirini, Abdulkafi. 2014. Toward understanding the variability in the language
proficiencies of Arabic heritage speakers. International Journal of Bilingualism
18(6). 730–765.
316
14 Arabic in the diaspora
Albirini, Abdulkafi. 2016. Modern Arabic sociolinguistics: Diglossia, variation,
codeswitching, attitudes and identity. London: Routledge.
Albirini, Abdulkafi & Elabbas Benmamoun. 2014. Concatenative and nonconcatenative plural formation in L1, L2, and heritage speakers of Arabic. The
Modern Language Journal 98(3). 854–872.
Albirini, Abdulkafi & Elabbas Benmamoun. 2015. Factors affecting the retention
of sentential negation in heritage Egyptian Arabic. Bilingualism: Language and
Cognition 18(3). 470–489.
Albirini, Abdulkafi, Elabbas Benmamoun & Brahim Chakrani. 2013. Gender
and number agreement in the oral production of Arabic heritage speakers.
Bilingualism: Language and Cognition 16(1). 1–18.
Albirini, Abdulkafi, Elabbas Benmamoun, Silvina Montrul & Eman Saadah. 2014.
Arabic plurals and root and pattern morphology in Palestinian and Egyptian
heritage speakers. Linguistic Approaches to Bilingualism 4(1). 89–123.
Albirini, Abdulkafi, Elabbas Benmamoun & Eman Saadah. 2011. Grammatical
features of Egyptian and Palestinian Arabic heritage speakers’ oral production.
Studies in Second Language Acquisition 33(2). 273–303.
Belnap, R. Kirk. 1999. A new perspective on the history of Arabic variation in
marking agreement with plural heads. Folia Linguistica 33(1–2). 169–186.
Borg, Albert & Marie Azzopardi-Alexander. 1997. Maltese. London: Routledge.
Borg, Alexander. 1985. Cypriot Arabic: A historical and comparative investigation
into the phonology and morphology of the Arabic vernacular spoken by the
Maronites of Kormakiti village in the Kyrenia district of north-west Cyprus.
Stuttgart: Deutsche Morgenländische Gesellschaft.
Boudelaa, Sami. 2013. Psycholinguistics. In Jonathan Owens (ed.), The Oxford
handbook of Arabic Linguistics, 369–392. Oxford: Oxford University Press.
Boumans, Louis. 2000. Topic pronouns in monolingual Arabic and in codeswitching. In Manwel Mifsud (ed.), Proceedings of the Third International Conference
of l’Association Internationale de Dialectologie Arabe held in Malta 29 March–2
April 1998, 89–95. Malta.
Boumans, Louis. 2001. Moroccan Arabic and Dutch: Languages of Moroccan
youth in the Netherlands. Langues et Linguistique: Revue Internationale de
Linguistique 8. 97–120.
Boumans, Louis. 2004. L’arabe marocaine de la generation ayant grandi aux PaysBas. In Dominique Caubet, Jacqueline Billiez, Thierry Bulot, Isabelle Léglise &
Catherine Miller (eds.), Parlers jeunes ici et là-bas: Pratiques et representations,
49–68. Paris: L’Harmattan.
317
Luca D’Anna
Boumans, Louis. 2007. The periphrastic bilingual verb construction as a marker
of intense language contact: Evidence from Greek, Portuguese and Maghribian
Arabic. In Everhard Ditters & Motzki Harald (eds.), Approaches to Arabic
linguistics presented to Kees Versteegh on the occasion of his sixtieth birthday,
291–311. Leiden: Brill.
Boumans, Louis & Dominique Caubet. 2000. Modelling intrasentential
codeswitching: A comparative study of Algerian/French in Algeria and
Moroccan/Dutch in the Netherlands. In Jonathan Owens (ed.), Arabic as a
minority language, 113–181. Berlin: Mouton de Gruyter.
Boumans, Louis & Jan Jaap de Ruiter. 2002. Moroccan Arabic in the European
diaspora. In Aleya Rouchdy (ed.), Language contact and language conflict in
Arabic: Variations on a sociolinguistic theme, 259–285. London: Routledge.
Brustad, Kristen. 2000. The syntax of spoken Arabic: A comparative study of
Moroccan, Egyptian, Syrian, and Kuwaiti dialects. Georgetown: Georgetown
University Press.
Canagarajah, Suresh A. 2008. Language shift and the family: Questions from the
Sri Lankan Tamil diaspora. Journal of Sociolinguistics 12. 143–176.
D’Anna, Luca. 2017a. Italiano, siciliano e arabo in contatto: Profilo sociolinguistico
della comunità tunisina di Mazara del Vallo. Palermo: Centro di studi filologici
e linguistici siciliani.
D’Anna, Luca. 2017b. Agreement with plural controllers in Fezzānī Arabic. Folia
Orientalia 54. 101–123.
D’Anna, Luca. 2018. Dialect contact in the Tunisian diaspora: Chebba speakers
in Mazara del Vallo (Sicily). In Perspectives on Arabic linguistics XXXI: Papers
from the Annual Symposium on Arabic Linguistics, Norman, Oklahoma, 2017.
Amsterdam: John Benjamins.
Daher, Nazih Y. 1992. A Lebanese dialect in Cleveland: Language attrition in
progress. In The Arabic language in America, 25–36. Detroit: Wayne State
University Press.
de Ruiter, Jan Jaap. 1989. Young Moroccans in the Netherlands: An integral approach
to their language situation and acquisition of Dutch. Utrecht: Rijksuniversiteit.
(Doctoral dissertation).
Ehala, Martin. 2015. Ethnolinguistic vitality. In Karen Tracy, Cornelia Ilie & Todd
Sandel (eds.), The international encyclopedia of language and social interaction,
1–7. Hoboken, NJ: Wiley-Blackwell.
Extra, Guus & Jan Jaap de Ruiter. 1994. The sociolinguistic status of the Moroccan
community in the Netherlands. Indian Journal of Applied Linguistics 20(1–2).
151–176.
318
14 Arabic in the diaspora
Giles, Howard, Richard Y. Bourhis & Donald M. Taylor. 1977. Towards a theory of
language in ethnic group relations. In Howard Giles (ed.), Language, ethnicity
and intergroup relations, 307–348. London: Academic Press.
Gonzo, Susan & Mario Saltarelli. 1977. Migrant languages: Linguistic change in
progress. In Marcel De Greve & Eddy Rossel (eds.), Problèmes linguistiques des
enfants des travailleurs migrants, 167–187. Brussels: Didier AIMAV.
Gonzo, Susan & Mario Saltarelli. 1983. Pidginization and linguistic change in
emigrant languages. In Roger Andersen (ed.), Pidginization and creolization
as language acquisition, 181–198. Rowley: Newbury House.
Lucas, Christopher. 2015. Contact-induced language change. In Claire Bowern &
Bethwyn Evans (eds.), The Routledge handbook of historical linguistics, 519–536.
London: Routledge.
Manfredi, Stefano. 2017. Arabi Juba: Un pidgin-créole du Soudan du Sud. Leuven:
Peeters.
Matras, Yaron. 2009. Language contact. Cambridge: Cambridge University Press.
Miller, Catherine. 2003. The relevance of Arabic-based pidgins-creoles for Arabic
linguistics. Al-Lugha 4. 7–46.
Owens, Jonathan. 1993. A grammar of Nigerian Arabic. Wiesbaden: Harrassowitz.
Owens, Jonathan (ed.). 2000. Arabic as a minority language. Berlin: Mouton de
Gruyter.
Poplack, Shana. 1980. Sometimes I’ll start a sentence in Spanish y termino en
español. Linguistics 18. 581–618.
Ratcliffe, Robert R. 2005. Bukhara Arabic: A metatypized dialect of Arabic in
Central Asia. In Éva Ágnes Csató, Bo Isaksson & Carina Jahani (eds.), Linguistic
convergence and areal diffusion: Case studies from Iranian, Semitic and Turkic,
141–51. London: Routledge.
Ratcliffe, Robert R. 2013. Morphology. In Jonathan Owens (ed.), The Oxford handbook of Arabic linguistics, 71–92. Oxford: Oxford University Press.
Romaine, Suzanne. 1989. Pidgins, creoles, immigrant and dying languages.
In Nancy C. Dorian (ed.), Investigating obsolescence: Studies in language
contraction and death, 369–385. Cambridge: Cambridge University Press.
Rouchdy, Aleya. 1992a. Introduction. In Aleya Rouchdy (ed.), The Arabic language
in America, 17–21. Detroit: Wayne State University Press.
Rouchdy, Aleya (ed.). 1992b. The Arabic language in America. Detroit: Wayne
State University Press.
Rouchdy, Aleya (ed.). 2002. Language contact and language conflict in Arabic:
Variations on a sociolinguistic theme. London: Routledge.
319
Luca D’Anna
Sankoff, Gillian. 2001. Linguistic outcomes of language contact. In Jack Chambers,
Peter Trudgill & Natalie Schilling-Estes (eds.), The handbook of language
variation and change, 638–668. Oxford: Blackwell.
Seeger, Ulrich. 2013. Zum Verhältnis der zentralasiatischen arabischen Dialekte.
In Renaud Kuty, Ulrich Seeger & Shabo Talay (eds.), Nicht nur mit Engelszungen: Beiträge zur semitischen Dialektologie. Festschrift für Werner Arnold
zum 60. Geburtstag, 313–322. Wiesbaden: Harrassowitz.
University of Zaragoza (ed.). 2013. A descriptive and comparative grammar of
Andalusi Arabic. Leiden: Brill.
Van Coetsem, Frans. 1988. Loan phonology and the two transfer types in language
contact. Dordrecht: Foris.
Van Coetsem, Frans. 2000. A general and unified theory of the transmission process
in language contact. Heidelberg: Winter.
Versteegh, Kees. 2014. The Arabic language. 2nd edn. Edinburgh: Edinburgh
University Press.
Vicente, Ángeles. 2005. Ceuta: Une ville entre deux langues. (Une étude sociolinguistique de sa communauté musulmane). Paris: L’Harmattan.
Vicente, Ángeles. 2007. Two cases of Moroccan Arabic in the diaspora. In
Catherine Miller, Enam Al-Wer, Dominique Caubet & Janet C. E. Watson (eds.),
Arabic in the city: Studies in dialect contact and language variation, 123–145.
London: Routledge.
320
Chapter 15
Arabic pidgins and creoles
Andrei Avram
University of Bucharest
The chapter is an overview of eight Arabic-lexifier pidgins and creoles: Turku,
Bongor Arabic, Juba Arabic, Kinubi, Pidgin Madame, Jordanian Pidgin Arabic, Romanian Pidgin Arabic, and Gulf Pidgin Arabic. The examples illustrate a number
of selected features of these varieties. The focus is on two types of transfer, imposition and borrowing, within the framework outlined by Van Coetsem (1988; 2000;
2003) and Winford (2005; 2008).
1 Introduction
This chapter aims to illustrate the emergence of Arabic-lexifier pidgins and creoles for which the contact situation – i.e. socio-historical context, the agents of
change, and the languages involved – is at least relatively well known.
The varieties considered can be classified into two groups, in geographical,
historical and developmental terms: the Sudanic pidgins and creoles, and the
immigrant pidgins in various Arab countries. Geographically, the Sudanic varieties developed in Africa – in present-day South Sudan, Chad, Uganda, and
Kenya. Historically, the varieties derive from a putative common ancestor, a pidgin that emerged in southern Sudan, in the first half of the nineteenth century.
Various Turkish–Egyptian military expeditions between 1820 and 1840 opened
southern Sudan for the slave trade. Permanent camps were set up soon after by
slave traders in the White Nile Basin, Bahr el-Ghazal and Equatoria Province, inhabited by an Arabic-speaking minority and a huge majority of slaves from various ethnic and linguistic backgrounds. After 1850, the slave traders’ settlements
were turned into military camps in which a military pidgin emerged, which is
traditionally referred to as “Common Sudanic Pidgin Creole Arabic” (Tosco &
Andrei Avram. 2020. Arabic pidgins and creoles. In Christopher Lucas & Stefano
Manfredi (eds.), Arabic and contact-induced change, 321–347. Berlin: Language Science
Press. DOI:10.5281/zenodo.3744529
Andrei Avram
Manfredi 2013: 253). Two subgroups of Sudanic varieties are recognized: the western branch, consisting of Turku and Bongor Arabic (in Chad), and the eastern one,
made up of Juba Arabic (in Sudan) and Kinubi (spoken in Uganda and Kenya).
Immigrant pidgins emerged in the eastern part of the Arab World, in Lebanon,
Jordan, Iraq and the countries of the Arab Gulf. Historically, these do not go back
more than 50 years. All these varieties are incipient pidgins.
The contact situations illustrated presuppose: (i) a source language (SL) and
a recipient language (RL); (ii) agents of contact-induced change, who may be
either SL or RL speakers; (iii) a psycholinguistically dominant language, which
is not necessarily a socially dominant language (Van Coetsem 1988; 1995; 2000;
2003; Winford 2005; 2008). A distinction is made between two types of transfer:
imposition and borrowing (Van Coetsem 1988; 2000; 2003). Imposition involves
SL-dominant speakers as agents (SL agentivity), is typical of second-language
(L2) acquisition, and induces changes mostly in phonology and syntax, although
it may also include transfer of lexical items from the dominant SL into the nondominant RL (Van Coetsem 1995: 18; Winford 2005: 376). Borrowing normally
involves RL-dominant speakers as agents (RL agentivity), typically targets lexical items, but may also include transfer of morphological material from a nondominant SL into the dominant RL.
In light of their sociolinguistic history, the varieties considered all emerged
under conditions of untutored, short-term L2 acquisition by adults dominant
in their socially subordinate SLs. L2 acquisition, a fortiori with adults, triggers
processes such as imposition via SL agentivity (i.e. substrate influence), simplification (Trudgill 2011: 40, 101) – also known as restructuring (Lucas 2015: 529)
– as well as language-internal (i.e. non-contact-induced) developments such as
grammatical reanalysis (Winford 2005: 415).
As in Manfredi (2018), the focus of this chapter is on imposition and borrowing.
It does not illustrate restructuring which does not involve any kind of transfer,
but often involves a reduction in complexity (Lucas 2015: 529). In the case of Arabic pidgins and creoles, restructuring is manifest in the domain of morphology,
in, for example, the loss of the Arabic verbal affixes and of the nominal and verbal
derivation strategies (Miller 1993).
The examples are illustrative only of selected contact-induced features of Arabic pidgins and creoles and their number has been kept to a reasonable minimum.
The examples from Arabic and the pidgins and creoles considered appear in a
uniform system of transliteration.
The chapter is organized as follows. §2 and §3 are concerned with Sudanic
pidgins and creoles. §4, on the other hand, deals with Arabic immigrant varieties.
§5 summarizes the findings and introduces issues for further research.
322
15 Arabic pidgins and creoles
2 Turku and Bongor Arabic
2.1 Current state and historical development
Turku is an extinct pidgin, formerly spoken in the Chari–Bagirmi region in western Chad (Muraz 1926). After the abolition of slavery by the Turkish–Egyptian
government in 1879, the Nile Nubian trader Rabeh withdrew with his slave soldiers into Chad. From a sociolinguistic point of view, Turku was initially a military pidgin. However, it later became one of three trade languages in what was
then French Equatorial Africa, along with Sango and Bangala (Tosco & Owens
1993: 183). Turku was a stable pidgin which does not appear to have creolized
(Tosco & Owens 1993).
Bongor Arabic is spoken in southwestern Chad, in and around the town of
Bongor, the capital of the Mayo–Kebbi Est region, at the border with Cameroon
(Luffin 2013). Given the many structural features it shares with Turku, it is plausible to assume that Bongor Arabic developed from the former. Sociolinguistically,
Bongor Arabic is a trade pidgin, used by the local Masa and Tupuri populations
with Arabic-speaking traders. It is currently a stable pidgin, but it exhibits features indicative of depidginization under the influence of Chadian Arabic (ChA).
No information about the number of speakers is available.
2.2 Contact languages
The lexifier language of Turku and Bongor Arabic is Western Sudanic Arabic.
The substratal input was provided by languages of various genetic affiliations:
Nilo-Saharan – e.g. Bagirmi, Mbay, Ngambay, Sar, Sara (Central Sudanic), Kanuri
(Western Saharan); Afro-Asiatic – Hausa (West Chadic); Niger-Congo – Fulfulde.
In the case of Turku, an additional contributor was the creole language Sango.
Both in Turku and in Bongor Arabic there is also adstratal input from French.
The adstrate of Bongor Arabic additionally includes two languages: Masa (NiloSaharan, Western Chadic) and Tupuri (Niger-Congo).
2.3 Contact-induced changes
2.3.1 Phonology
The substrate languages do not have /ḫ/, which is generally replaced by [k]:
Turku kamsa ‘five’ < ChA ḫamsa; Bongor Arabic kídma ‘work’ < ChA ḫidma.
Many of the substrate languages do not have /f/, which is substituted with [p]
323
Andrei Avram
or perhaps [ɸ],1 e.g. Turku pfil ‘elephant’ < ChA fīl. In French loanwords, the
reflexes of /v/ are either [b] or [w]: Bongor Arabic boté ‘to vote’ < French voter,
wotír ‘car’ < French voiture.
The consonants [ɲ] and [ŋ] occur only in loanwords: Bongor Arabic ngambáy
‘Ngambay’ < Ngambay ngàmbáy; Turku ngari ‘manioc’ < Mbay ngàrì, konpanye
‘company’ < French compagnie; [v] and [ʒ] occur only in phonologically nonintegrated words of French origin: Turku sivil ‘civilian’ < French civil; Bongor
Arabic žurnalíst ‘journalist’ < French journaliste.
Variation affects several consonants. For instance, [f] occurs in variation with
[b] or [p]: Turku fišan ~ bišan ‘because’; Bongor Arabic máfi ~ mápi ‘neg’ <
ChA mā fī, sofér ~ sopér ‘driver’ < French chauffeur. Most of the substrate languages do not have /š/, which accounts for [ʃ] ~ [s] variation, in words with
either etymological /s/ or /ʃ/: Turku gasi ~ gaši ‘expensive’ < ChA gāsī, biriš ~
biris ‘mat’ < ChA birīš; Bongor Arabic máši ~ mási ‘go’. The usual reflexes of
French /ʒ/, absent from the phonological inventories of the substrate languages,
are [z], [ʤ] and [s] respectively: Turku ǧinenal ‘general’ < French général, suska
‘until < French jusqu’à; Bongor Arabic zúska ‘when, during’ < French jusqu’à
‘until’.
Finally, vowel length is not distinctive: Turku, Bongor Arabic kalam ‘speech;
speak’ < ChA kalām ‘speech’.
2.3.2 Morphology
On current evidence (Luffin 2013: 180–181), Bongor Arabic exhibits signs of depidginization under the influence of Chadian Arabic. The most striking instance
of this is the use of pronominal suffixes, unique among Arabic-lexifier pidgins
and creoles:
(1)
Bongor Arabic (Luffin 2013: 180)
índi gáy árifu úsum-i
2sg impf know name-poss.1sg
‘You know my name.’
Also, verbal affixes are sporadically used:
(2) Bongor Arabic (Luffin 2013: 181)
a. ána ma n-árfa
1sg neg 1sg-know
‘I don’t know.’
1
Transcribed as 〈pf〉 by Muraz (1926: 168).
324
15 Arabic pidgins and creoles
b. anína rikíb-na
wotír da sáwa
1pl ride.prf-1pl car prox together
‘We took the car together.’
These cases might be analyzed as borrowing under sui generis RL agentivity,
whereby morphological material from a non-dominant SL is imported into a nondominant (second) RL.
2.3.3 Lexicon
A part of the non-Arabic vocabulary of Turku can be traced back to its substrate languages (Avram 2019). Most of the loanwords are from Sara-Bagirmi
languages: adinbang ‘eunuch’ < Bagirmi ádim mbàŋ ‘servant of the sultan’; gao
‘hunter’ < Sar gáw; ngari ‘manioc’ < Mbay ngàrì. The second most significant
important contributor is Sango: kay ‘paddle’ < Sango kâî, tipoy ‘carrying hammock’ < típóí. A few words can be traced to Fulfulde and Kanuri: kelkelbanǧi
‘golden beads’ < Fulfulde kelkel-banja; wélik ‘lightning’ < Kanuri wulak ‘flash of
lightning’. In a number of cases, the exact SL cannot be established: koporo ‘0.10
Francs’ < Fulfulde, Sango, Sara koporo ‘coin’; gurumba ‘hat’ < Hausa gurúmba,
Kanuri gurumbá. As for Bongor Arabic, its African adstrate languages have contributed only a few loanwords, such as bursdíya ‘Monday’. There are also loanwords from French. In Turku most of these relate to the military (Tosco & Owens
1993: 262–263), e.g. Turku itenan ‘lieutenant’ < French lieutenant, permišon ‘permission’ < French permission. In addition to nouns, French loanwords include
some verbs, such as Bongor Arabic komandé ‘order’ < French commander, and
at least one function word, Turku suska, Bongor Arabic zúska ‘when, during’ <
French jusqu’à ‘until’.
The substratal influence on Turku can also be seen in a number of compound
calques (Avram 2019; Manfredi, this volume). Some of these are modelled on SaraBagirmi languages: bahr gum ‘rising water’, cf. Ngambay màn-kàw, lit. ‘river
goes’; nugra ana asal ‘beehive’, cf. Ngambay bòlè-tǝnji, lit. ‘hole (of) honey’. Other
calques have equivalents in several SLs, such as nugra haǧer ‘cave’, lit. ‘hole
mountain/stone’, cf. Kanuri kûl kau-be lit. ‘cavity mountain-of’, Ngambay bòlòmbàl lit. ‘hole mountain’, Sango dûtênë lit. ‘hole stone’.
325
Andrei Avram
3 Juba Arabic and Kinubi
3.1 Current state and historical development
Juba Arabic is mainly spoken in South Sudan; there are also diaspora communities, mostly in Sudan and Egypt. Two main reasons make it difficult to estimate its
number of speakers. Firstly, while Juba Arabic is spoken as a primary language
by 47% of the population of Juba, the capital city of South Sudan, it is also used
as a second or third language by the majority of the population of the country
(Manfredi 2017: 7). Secondly, the long coexistence of Juba Arabic with Sudanese Arabic, its main lexifier language, has led to the emergence of a continuum
ranging from basilectal, through mesolectal, to acrolectal varieties; delimiting
acrolectal Juba Arabic from Arabic is no easy task, particularly in the case of the
large diaspora communities in Khartoum and Cairo.
Juba Arabic emerged as a military pidgin. Sociolinguistically, it is today an
inclusive identity marker for the ethnically and linguistically diverse population
of South Sudan (Tosco & Manfredi 2013: 507). Developmentally, Juba Arabic is a
pidgincreole.2
The Mahdist revolt, which started in 1881, eventually brought about the end
of Turkish–Egyptian control over Equatoria, in southern Sudan. Following an
invasion by Mahdist rebels, the governor fled to Uganda, accompanied by slave
soldiers loyal to the central government. These soldiers subsequently became the
backbone of the British King’s African Rifles. While some of the troops remained
in Uganda, others were moved to Kenya and Tanzania. This led to the dialectal
division between Ugandan and Kenyan Kinubi. Like Juba Arabic, therefore, Kinubi started out as a military pidgin, then underwent stabilization and expansion.
Today, however, Kinubi is the only Arabic-lexifier fully creolized variety, that is,
a native language for its entire speech community.
Kinubi is spoken in Uganda and in Kenya. The number of speakers of Kinubi
is a matter of debate. Ugandan Kinubi was spoken by some 15,000 people, according to the 1991 census, and Kenyan Kinubi by an estimated 10,000 in 2005.
However, other estimates put the combined number of speakers at about 50,000.
The largest communities of Kinubi speakers are in Bombo (Uganda), Nairobi (the
Kibera neighbourhood) and Mombasa (Kenya).
2
A pidgincreole is “a former pidgin that has become the main language of a speech community
and/or a mother tongue for some of its speakers” (Bakker 2008: 131).
326
15 Arabic pidgins and creoles
3.2 Contact languages
The main lexifier language of Juba Arabic is Sudanese Arabic (SA), with some
input from Egyptian Arabic (EA) and Western Sudanic dialects as well. The substrate is represented by a relatively large number of languages, belonging to
super-phylums, Nilo-Saharan and Niger-Congo. The former includes Eastern
Sudanic languages, such as Bari, Lotuho (Eastern Nilotic), Acholi, Belanda Bor,
Dinka, Jur, Nuer, Päri, Shilluk (Western Nilotic), Didinga (Surmic), and Central
Sudanic languages, such as Avokaya, Baka, Bongo, Ma’di, Moru; the Niger-Congo
super-phylum is represented by, for example, Zande and Mundu. The main substrate language is considered to be Bari, including its dialects Kakwa, Kuku, Pojulu, and Mundari.3
Given its sociolinguistic history, Kinubi shares much of its substrate with Juba
Arabic. However, the substrate of Ugandan Kinubi additionally includes Eastern Sudanic languages, such as Alur, Luo (Western Nilotic), and Central Sudanic
languages such as Mamvu, Lendu and Lugbara (Owens 1997: 161; Wellens 2003:
207), spoken in Uganda. Unlike Juba Arabic, Kinubi also exhibits the effects of
the adstratal influence exerted by two Bantu languages, Luganda – particularly
in Ugandan Kinubi – and Swahili – particularly in Kenyan Kinubi. One other
language that should be mentioned is English, official both in Uganda and in
Kenya.
3.3 Contact-induced changes
3.3.1 Phonology
A number of consonants found in Arabic, but absent from the phonological inventories of the substrate languages, are either deleted or substituted. Consider the
reflexes of pharyngeals: háfla ‘feast’ < SA ḥafla; árabi ‘Arabic’ < SA ʕarabī. The
pharyngealized consonants are replaced by their plain counterparts: towíl ‘long’
< SA ṭawīl; dul ‘shadow’ < SA ḍull; súlba ‘hip’ < SA ṣulba; zúlum ‘to anger’ < SA
ẓulum. The velar fricatives of Arabic are always replaced by velar stops: kábara
‘piece of news’ < SA ḫabar; šókol ‘work’ < SA šoɣol, gárib ‘west’ < SA ɣar(i)b.
As in Juba Arabic, the pharyngeals of Arabic are either replaced or lost in Kinubi (Owens 1985: 10; Wellens 2003: 209–212). The earliest records of Ugandan
Kinubi4 are replete with illustrative examples (Avram 2017a): haǧa ‘thing’ < SA
ḥāǧa, aram ‘thief’ < SA ḥarāmi, līb < ‘to play’ < SA liʕib. The pharyngealized
3
4
Sometimes considered to be separate languages (Wellens 2003: 207).
The main ones are: Cook (1905), Jenkins (1909), Meldon (1913), and Owen & Keane (1915).
327
Andrei Avram
consonants are replaced by their plain counterparts, as in these examples from
early Ugandan Kinubi: towil ‘long’ < SA ṭawīl; dulu ‘shadow’ < SA ḍull, hisiba
‘measles’ < SA ḥiṣba; zulm ‘to anger’ < SA ẓulum. Like Juba Arabic, Kinubi substitutes velar stops for the Arabic velar fricatives. Consider the following early
Ugandan Kinubi forms: kidima ‘work’ < SA ḫidma; šokolo ‘work’ < SA šoɣol,
balago ‘commandment’ < SA balāɣ ‘message’. Substratal influence also accounts
for consonant degemination, given that the substrate languages “lack these in all
but a few morphonologically determined contexts” (Owens 1997: 162).
Substratal influence can also be seen in the occurrence of certain consonants.
Consider first /ɓ/ and /ɗ/: Juba Arabic d’éngele ‘liver’ < Bari denggele; Juba Arabic
b’ónǧo ‘pumpkin’ < Bongo b’onǧo. The other consonants which occur only in
loanwords from the substrate and/or adstrate languages are [p] [v], [ʧ], [ɲ], and
[ŋ]: Kinubi lípa ‘to pay’ < Swahili -lipa; Kinubi camp ‘camp’ < English camp;
Kinubi víta ‘war’ < Swahili vita; Juba Arabic čam ‘food’ < Acholi, Belanda Bor,
Jur čama, Juba Arabic čayniz < English Chinese, Kinubi čay ‘tea’ < Swahili chai;
Juba Arabic nyékem, Kinubi nyékem ‘chin’ < Bari nyékem, Kinubi nyánya ‘tomato’
< Swahili nyanya; Juba Arabic ŋun ‘divinity’ < Bari ngun. The integration of
these phonemes is thus a result of borrowing (under RL agentivity) rather than
of imposition.
The following instances of consonant variation are more common in Juba Arabic (Manfredi 2017: 25–27). The most frequent is [ʃ] ~ [s]: geš ~ ges ‘grass’. Further,
[z] is in variation with [ʤ] before /o/ and /a/: zówǧu ~ ǧówǧu ‘to marry’, záman
~ ǧáman ‘time; when’. There is also [p] ~ [f] variation in word-initial position, including in loanwords: poǧúlu ~ foǧúlu ‘Pojulu’, prótestan ~ frótestan ‘protestant’.
Finally, the phoneme /f/ may also be phonetically realized as [p]: nédifu ~ nédipu
‘to clean’. Of these cases of variation, the latter has been specifically attributed
to substratal influence from Bari (Miller 1989; Manfredi 2017). It might be argued,
however, that all these instances of consonant variation reflect the influence of
the substrate languages, regardless of their genetic affiliations. The following do
not have /ʃ/: Acholi, Avokaya, Baka, Bari, Belanda Bor, Bongo, Dinka, Jur, Lotuho,
Ma’di, Moru, Mundu, Nuer, Päri, Shilluk, Zande. Of these, Acholi, Belanda Bor,
Bongo, Dinka, Jur, Nuer, Päri and Shilluk do not have /s/ either. A number of substrate languages do not have /z/: Acholi, Bongo, Belanda Bor, Dinka, Jur, Lotuho,
Nuer, Päri, and Shilluk. All of these, however, have /ʤ/. Finally, /f/ is not part of
the phonological inventory of Acholi, Bongo, Dinka, Jur, Nuer, Päri, and Shilluk,
which do, however, have /p/. Given the intricacies of the distribution of /ʃ/, /s/, /z/,
/ʤ/, /f/, and /p/ across the substrate languages, the types of variation illustrated
are not surprising.
328
15 Arabic pidgins and creoles
As in Juba Arabic, [ʃ] is in variation with [s] in Kinubi (Owens 1985: 237; Owens
1997: 161; Wellens 2003: 38; Luffin 2005: 62; Avram 2017a): early Ugandan Kinubi
šabaka ~ sabaka ‘net’). Although it is etymological /š/ which is typically subject
to variation, occasionally this also applies to etymological /s/: early Ugandan
Kinubi sikin ~ šekin ‘knife’ < SA sikkīn (Avram 2017a) and modern Kenyan Kinubi
fluš ~ flus ‘money’ < SA fulūs (Luffin 2005: 63). Note that [š] ~ [s] variation also
extends to loanwords from Swahili (Wellens 2003: 80; Luffin 2005: 63; Avram
2017a): early Ugandan Kinubi šamba ~ samba ‘field’ < Swahili shamba. Like Juba
Arabic, Kinubi exhibits [z] ~ [ʤ] variation (Owens 1985: 235; Owens 1997: 161;
Wellens 2003: 215; Luffin 2005: 63; Avram 2017a): early Ugandan Kinubi ǧalan ~
zalan ‘angry’ < SA zaʕlān. However, unlike Juba Arabic, in Kinubi the [z] ~ [ʤ]
variation also occurs before the two front vowels /i/ and /e/: ze ~ ǧe ‘as’, early
Ugandan Kinubi anǧil ~ enzil ‘descend’. According to Owens (1997: 161), this “is
due perhaps to Bari substratal influence, since Bari has only j, not z.” In fact, as
in the case of Juba Arabic, the same is true of several other substrate languages.
Lastly, there are instances of [l] ~ [r] variation (Wellens 2003: 214; Luffin 2005: 65),
affecting both etymological /l/ and etymological /r/ in Arabic-derived words, e.g.
tále ~ táre ‘go out’, gerí ~ gelí ‘near’, and in borrowings, e.g. Ugandan Kinubi čálo
~ čáro ‘village’ < Luganda e-kyalo; Kenyan Kinubi tumbíli ~ tumbíri ‘monkey’ <
Swahili tumbili. This variation seems to reflect the influence of Luganda and
Swahili. In the former, [l] and [r] are in complementary distribution, with [r]
occurring after the front vowels /i/ and /e/, and [l] elsewhere (Wellens 2003: 214),
while in the latter [l] and [r] are in free variation (Luffin 2014: 79).
As in the substrate languages, there is no distinction between short and long
vowels: Juba Arabic sudáni ‘Sudanese’ < SA sudānī, Kinubi kabír ‘big’ < SA kabīr.
3.3.2 Morphology
Apart from the Arabic-derived plural suffixes -at and -in, Juba Arabic uses the
plural marker of Bari origin -ǧín (Nakao 2012: 131; Manfredi 2014a: 58), which is
attached only to loanwords from local languages:
(3)
Juba Arabic (Manfredi 2014a: 58)
a. kɔrɔpɔ-ǧín (< Bari kɔrɔpɔ)
leaf-pl
‘leaves’
b. beng-ǧín (< Dinka beng)
chief-pl
‘chiefs’
329
Andrei Avram
c. b’angiri-ǧín (< Zande b’angiri)
cheek-pl
‘cheeks’
Another phenomenon worth mentioning is the occurrence in the speech of
young urban speakers of hybrid forms, which consist of the Bari relativizer loand a noun either from Arabic or from one of the substrate/adstrate languages
(Nakao 2012: 131). Note, however, that there is a functional overlap between Bari
lo- and Sudanese Arabic abu.
(4)
Juba Arabic (Manfredi 2017: 46)
a. lo-beléde
(< Bari lo- + SA beled)
rel-country
‘peasant’
b. lo-pómbe (< Bari lo- + Swahili pombe)
rel-alcohol
‘drunkard’
Given that a relatively large number of Bari-derived words contain lo- (Miller
1989; Manfredi 2017: 46), the examples in (4) confirm the fact that morphological
innovations are typically introduced through lexical borrowings via RL agentivity, and subsequently become productive in the RL.
Note, finally, that most of the speakers who use the plural marker -ǧín and the
relativizer lo- are dominant in Juba Arabic. These cases therefore confirm the
fact that RL monolinguals can be agents of borrowing (Van Coetsem 1988: 10).
A small number of Kinubi nouns borrowed from Swahili exhibit the Bantu
nominal classifiers:
(5) Kinubi (Wellens 2003: 57)
a. mu-zé wa-zé
nc1-old.man nc2-old.man
‘old man, old men’
b. mu-zukú
wa-zukú
nc1-grandchild nc2-grandchild
‘grandchild, grandchildren’
c. m-zúngu
wa-zúngu
nc1-European nc2-European
‘European, Europeans’
330
15 Arabic pidgins and creoles
3.3.3 Syntax
Bureng Vincent (1986: 77) first noted the similarity between the prototypical passive construction in Juba Arabic and its Bari counterpart:
(6) Juba Arabic (Bureng Vincent 1986: 77)
a. bágara áyinu ma Wáni
cow see.pst with Wani
‘The cow has been seen by Wani.’
b. Bari (Bureng Vincent 1986: 77)
kítɜŋ a mɛtà kɔ̀ Wànì
cow pst see with Wani
‘The cow has been seen by Wani.’
As can be seen, in both Juba Arabic and Bari the agent is introduced by the
comitative preposition ‘with’. This is a case of lexico-syntactic imposition via
identification of SL and RL lexemes (Manfredi 2018: 415): the Juba Arabic lexical
entry ma is derived from Sudanese Arabic maʕ, but its semantics reflects the
influence of Bari kɔ̀. The same is true of Kenyan Kinubi:
(7)
Kinubi (Luffin 2005: 230)
yal-á al akulú
ma nas tomsá
child-pl rel eat.pst.pass with pl crocodile
‘the children who were eaten by a crocodile’
Consider next the syntax of numerals in Kinubi (Wellens 2003: 90; Luffin 2014:
309). Their post-nominal placement is calqued on Swahili:
(8)
Kinubi (Luffin 2014: 309)
wéle kámsa ma baná árba
boy five with girl.pl four
‘five boys and four girls’
(9)
Swahili (Luffin 2014: 309)
miti mia
tatu
tree hundred three
‘three hundred trees’
With cardinal numerals, the order is hundred + unit and thousand + unit respectively:
331
Andrei Avram
(10) Kinubi (Luffin 2014: 309)
elf
wáy
thousand one
‘one thousand’
Kinubi thus follows the Swahili model:
(11) Swahili (Luffin 2014: 309)
elfu
moja
thousand one
‘one thousand’
Consider also a case of syntactic change induced by lexical calquing. Juba Arabic (fu)wata ‘ground’ functions as an impersonal subject in weather expressions:
(12)
Juba Arabic (Nakao 2012: 141)
(fu)watá súkun
ground hot
‘It is hot.’
Nakao (2012: 141) shows that this is also the case in Acholi and Ma’di:
(13)
Acholi (Nakao 2012: 141)
piiny lyeet
ground warm
‘It is warm.’
(14) Ma’di (Nakao 2012: 141)
aci
vu
ground hot
‘It is hot.’
In fact, these types of sentences are widespread in Western Nilotic substrate
languages, such as Dinka, Jur, Päri, and Shilluk:
(15)
Dinka (Nebel 1979: 202)
piny a-tuc
ground 3sg-warm
‘It is warm.’
In both Juba Arabic and Kinubi ras ‘head’ also occurs in the complex preposition fi ras ‘on’:
332
15 Arabic pidgins and creoles
(16)
a. Juba Arabic (Nakao 2012: 141)
merísa fí fi ras terebéza
beer exs on head table
‘The beer is on the table.’
b. Kinubi (Wellens 2003: 159)
fi rás séder
on head tree
‘on top of the tree’
Nakao (2012: 141) attributes this function of ras to substratal influence from
Acholi and Ma’di:
(17)
Acholi (Nakao 2012: 141)
cib wi-meja
put head-table
‘Put it on the table.’
However, other possible sources include Western Nilotic languages such as
Belanda Bor, Jur, Päri and Shilluk:
(18)
Jur (Pozzati & Panza 1993: 342)
kedh ŋo wi tarabesa
put 3sg head table
‘Put it on the table.’
Moreover, a preposition ‘on’ derived from the noun ‘head’ is also attested in
Bongo (Central Sudanic) and Zande (Niger-Congo):
(19) Bongo (Moi et al. 2014: 39)
ba do mbaa
3sg on car
‘He is on a car.’
(20)
Zande (De Angelis 2002: 288)
mo mai he ri ngua
2sg put 3sg on wood
‘Put it on the wood.’
The verb gal/gale/gali ‘say’ is used in Juba Arabic and Ugandan Nubi as a
complementizer, with verba dicendi and verbs of cognition:
333
Andrei Avram
(21)
a. Juba Arabic (Miller 2001: 469)
úwo kélem gal úwo bi-ǧa
3sg speak comp 3sg irr-come
‘He said that he would come.’
b. Ugandan Kinubi (Wellens 2003: 204)
úmon áruf gal fí difan-á al gi-ǧá
3pl know comp exs guest-pl rel prog-come
‘They know that there are are guests who are coming.’
The use of a verbum dicendi as a complementizer resembles the situation in
Bari,5 where adi ‘say’ introduces direct speech (Owens 1997: 163; Miller 2001:
469):
(22)
Bari (Miller 2001: 469)
mukungu a-kulya adi nan d’ad’ar kakitak merya-mukanat
sub-chief pst-say comp 1sg want worker fifty
‘The sub-chief spoke saying: I want fifty workers.’
3.3.4 Lexicon
Since Bari is the main substrate language of Juba Arabic, unsurprisingly it contributes most of its African-derived words: gúgu ‘granary’ < Bari gugu; kení
‘co-wife’ < Bari köyini; loɲumég ‘hedgehog’ < Bari lónyumöng; tóŋga ‘pinch’ <
Bari toŋga. In several cases, the Juba Arabic form can be traced to a specific dialect: d’oŋóŋ ‘back of head’ < Pojulu doŋoŋ; láŋa ‘wander’ < Mundari laŋa ‘travel’;
nyéte vs ŋéte ‘black-eyed pea leaf’ < Bari nyete vs Kakwa, Pojulu ŋete. Moreover,
“more Bari lexical items are being borrowed” in Youth Juba Arabic (Nakao 2012:
131): kapaparát ‘butterfly’ < Bari kapoportat; lukulúli ‘bat’ < Bari lukululi. Several other substrate and adstrate languages have contributed to the lexicon of
Juba Arabic (Nakao 2012; 2015): adúngú ‘harp’ < Acholi aduŋu; b’ónǧo ‘pumpkin’ < Bongo b’onǧo; báfura ‘cassava’ < Dinka bafora ‘manioc, (sweet) cassava’;
káwu ‘cowpea’ < Ma’di kau; malangí < bottle’ < Bangala/Lingala molangi; kámba
‘belt’ < Swahili kamba; imbíró ‘palm tree’ < Zande mbíró. Some sixty lexical items
found in the earliest records of Ugandan Kinubi can be traced back to various
substrate languages (Avram 2017a): lawoti ‘neighbours’ < Acholi lawoti ‘fellow,
friend’; korufu ‘leaf’ < Bari korofo ~ kɔrɔpɔ ‘leaves’; lwar ‘abscess’ < Dinka luär
5
Unsurprisingly, in Juba Arabic “the use of adi as in Bari [is] the most frequent […] in particular
among speakers of Bari origin” (Miller 2001: 470; author’s translation).
334
15 Arabic pidgins and creoles
‘pain of a swelling’; seri ‘fence’ < Lugbara seri ‘plant used for fencing’; mukuta
‘key’ < Päri mukuta.
The influence of Luganda and Swahili as adstrate languages is already documented in early Ugandan Kinubi (Avram 2017a): Ugandan Kinubi kibra ~ kibera
‘forest’ < Luganda e-kibira, nyinveza ‘fix’ < Luganda nyweza ‘make firm, hold
firmly’; dirisa ‘window’ < Swahili dirisha; kibanda ‘shed’ < Swahili kibanda ‘small
shed’. The lexicon of modern Ugandan Kinubi is characterized by a large number of loanwords from Luganda and Swahili (Wellens 2003; Nakao 2012: 133–
134), such as: mé(é)mvu ‘banana’ < Luganda amaemvu ‘bananas’; ntulége ‘zebra’
< Luganda e-ntulege; karibísha ‘welcome’ < Swahili karibisha ‘welcome’; sangá ~
šangá ‘be surprised’ < Swahili shangaa. In some cases, these loanwords have replaced previously attested compounds consisting of Arabic-derived elements:6
early Ugandan Kinubi mária bitá murhúm ‘widow’, lit. ‘wife of the deceased’
vs. modern Ugandan Kinubi mamwándu ‘widow’ < Luganda nnamuwandu. As
for the lexicon of modern Kenyan Kinubi, it is strongly influenced by Swahili.
Luffin (2004) lists some 170 loanwords from Swahili (out of approximately 1,400
words recorded), from a wide range of domains, for example: barabára ‘highway’
< Swahili barabara; serikáli ‘government’ < Swahili serikali; tafaúti ‘difference’
< Swahili tafauti; úza ‘sell’ < Swahili ku-uza. Swahili has also contributed several function words: badáye ‘after’ < Swahili baadaye ‘afterwards’; íle ‘these’ <
Swahili ile; na ‘and, with’ < Swahili na. Kenyan Kinubi lexical items have occasionally undergone semantic shift or semantic extension under the influence of
the meanings of their Swahili counterparts (Luffin 2014: 315): destúr ‘tradition’,
cf. Swahili desturi ‘tradition’; fáham ‘to understand, to remember’, cf. Swahili
-fahamu ‘to understand, to remember’.
In some cases, the exact origin of loanwords found in Juba Arabic cannot be
established: búra ‘cat’ < Acholi, Bongo, Dinka, Päri bura, Didinga buura; daŋá
‘bow’ < Bari, Jur daŋ, Didinga d’anga, Dinka dhaŋ; pondú ‘cassava leaf’ < Bangala,
Kakwa, Lingala pondu, Pojulu pöndu. The same holds for a number of loanwords
attested in early Ugandan Kinubi (Avram 2017a): bongo ‘cloth’ < Acholi, Lendu,
Lugbara, Zande bongo, Bari boŋgo; godogodo ‘thin from illness’ < Acholi, Avokaya,
Bari, Baka, Lotuho, Moru, Zande godogodo ‘thin, sick(ly)’; mukungu ‘headman’ <
Acholi mukuŋu, Bari mʊkʊŋgʊ, Luganda o-mukungu, Lugbara mukungu ‘(sub-)
chief’. This is also true of several Kinubi words attested in more recent sources
(Wellens 2003; Nakao 2012: 133–134): júju ‘shrew’ < Bari juju, Ma’di juju; kingílo
‘rhinoceros’ < Avokaya kiŋgili, Moru kingile. In some cases, the occurrence of
alternative forms is due to their different SLs: banǧa ‘debt’ < Bari banja, Lugbara
banja, Luganda e-bbanja vs. banya ‘debt’ < Acholi banya.
6
See also Tosco & Manfredi (2013: 509).
335
Andrei Avram
Under the influence of the substrate and adstrate languages, some Arabicderived lexical items have undergone semantic extension, thereby becoming polysemous in Juba Arabic (Nakao 2012: 136), e.g. gówi ‘hard; difficult’, cf. Acholi
tek, Bari logo’, Lotuho gol, Ma’di okpo, Swahili kali.
Juba Arabic “compensates its lexical gaps through the lexification of Arabic
morphosyntactic sequences” (Tosco & Manfredi 2013: 509). A case in point are
Juba Arabic compounds, formed via juxtaposition or with their two members
linked by the possessive particle ta (Manfredi 2014b: 308–309). These include
calques after several substrate languages (Nakao 2012: 136), e.g. ída ta fil ‘elephant trunk’, cf. Acholi ciŋ lyec, Bari könin lo tome, Dinka ciin akɔɔn, Jur ciŋ lyec,
Lotuho naam tome, Shilluk bate lyec, lit. ‘arm (of) elephant’. Kinubi also exhibits
a number of calques (Nakao 2012; Avram 2017a; Manfredi, this volume). Some
of these compounds and phrases can be traced to several SLs, as in the following early Ugandan Kinubi examples (Avram 2017a): gata kalam ‘decide, judge’,
cf. Acholi ŋɔlɔ kop ‘decide, give judgment’, Bongo ad’oci kudo, Jur ŋɔl lubo, Päri
ŋondi lubo, lit. ‘cut word/speech’; Dinka wèt tèm ‘decide, give the sentence’, lit.
‘word cut’; jua bita ter ‘nest’, cf. Acholi ot winyo, Bari kadi-na-kwen, Belanda
Bor kwɔt winy, Shilluk wot winyo, Zande dumô zirê, lit. ‘house (of) bird’. Other
calques, presumably more recent ones, reflect the growing influence of Swahili
on Kenyan Kinubi (Luffin 2014: 315): bakán wáy ‘together’, cf. Swahili pamoja
‘together’, lit. ‘place one’, mára wáy wáy ‘seldom’, cf. Swahili mara moja moja
‘seldom’, lit. ‘time one one’.
To conclude, SL agentivity accounts for the small number of loanwords and
calques recorded in the earliest stage (i.e. pidginization) of Juba Arabic and Kinubi. At a later stage (i.e. after nativization), the larger number of loanwords and
calques is a result of borrowing under RL agentivity.
4 Arabic-lexifier pidgins in the Middle East
4.1 Current state and historical development
Several Arabic-lexifier pidgins have emerged in the Middle East. These include
Romanian Pidgin Arabic, Pidgin Madame, Jordanian Pidgin Arabic, and Gulf Pidgin Arabic. The first three can be classified as work force pidgins.7 Gulf Pidgin
Arabic also started out as work force pidgin (Smart 1990: 83), but it is now an
interethnic contact language (Avram 2014: 13).8
7
8
These are pidgins which “came into being in work situations” (Bakker 1995: 28).
That is, one which is “used not just for trade, but also in a wide variety of other domains”
(Bakker 1995: 28).
336
15 Arabic pidgins and creoles
Romanian Pidgin Arabic (Avram 2010) was a short-lived pidgin, formerly used
on Romanian well sites in Iraq, in locations in the vicinity of Ammara, Basra, Kut,
Nassiriya, Rashdiya and Rumaila. Romanian Pidgin Arabic emerged after 1974,
when Romanian well sites started operating in Iraq. Romanians typically made
up two thirds of the oil crews, with Arabs making up the final third. The first
Gulf War and the subsequent withdrawal of the Romanian oil rigs put an end to
the use of Romanian Pidgin Arabic.
Immigration of Sri Lankan women to Arabic-speaking countries is reported to
have started in 1976 (Bizri 2010: 16), but the large influx into Lebanon came later,
in the early 1990s. Pidgin Madame is spoken in Lebanon by Sri Lankan female
domestic workers and their Arab employers, mostly in the urban centres of the
country.
Jordanian Pidgin Arabic (Al-Salman 2013) is used in the city of Irbid, in the ArRamtha district in the north of Jordan, in interactions between Jordanians and
Southeast Asian migrant workers of various linguistic backgrounds. However,
only Jordanian Pidgin Arabic as spoken by Bengalis is documented.
Gulf Pidgin Arabic is a blanket term designating the varieties of pidginized
Arabic used in Saudi Arabia and the countries on the western coast of the Arab
Gulf, i.e. Kuwait, the United Arab Emirates, Oman, Bahrain, and Qatar.
4.2 Contact languages
The main languages involved in the emergence of Romanian Pidgin Arabic are
Romanian, Egyptian Arabic (spoken by immigrant workers), and Iraqi Arabic
(IA). A small minority of the participants in the language-contact situation had
some knowledge of English.
The other pidginized varieties of Arabic in the Middle East share the characteristic of having various Asian languages as their substrate.9 For Pidgin Madame,
the main contact languages are Lebanese Arabic, as the lexifier language, and
Sinhalese. Another language, with a much smaller contribution, is English. In
the case of Jordanian Pidgin Arabic, the contact languages are mainly Jordanian
Arabic (JA) and Bengali. The contribution of English is very limited. As for Gulf
Pidgin Arabic, it emerged in a contact situation of striking complexity. On the
one hand, Arabic, the lexifier language, is represented by several dialects, which
are not all subsumed under what is known as Gulf Arabic (GA), in spite of what
the name of the pidgin suggests. On the other hand, the number of languages
spoken by the immigrant workers is staggering: for instance, in the United Arab
9
Bizri (2014: 385) therefore suggests the cover term “Asian Migrant Arabic pidgins”.
337
Andrei Avram
Emirates the 200 nationalities and 150 ethnic groups speak some 150 languages.
Adding to the complexity of the language-contact situation is the fact that these
languages are typologically diverse. Last but not least, English also plays a role
in interethnic communication, particularly in the service sector.
4.3 Contact-induced changes
4.3.1 Phonology
The phonology of all the pidginized varieties of Arabic in the Middle East exhibits the outcomes of SL agentivity, which also accounts for the occurrence of
considerable intra- and inter-speaker variation (Avram 2010: 21–22; Bizri 2014:
393; Avram 2017b: 133).
Consider first Romanian Pidgin Arabic. The following are features characteristic of speakers with Romanian as their first language (L1). The phrayngeals are
either replaced or deleted: habib ‘friend’ < IA/EA ḥabīb; mufta ‘key’ < IA/EA
muftāḥ; saa ‘hour’ < IA/EA sāʕa. Plain consonants are substituted for pharyngealized ones: halas ‘ready’ < IA/EA ḫalāṣ. Both velar fricatives are replaced:
hamsa ‘five’ < IA/EA ḫamsa; šogol ‘work (n)’ < IA šuɣ(u)l. Geminate consonants
are degeminated: sita ‘six’ < IA/EA sitta. There is no distinction between short
and long vowels, either in lexical items of Arabic origin or in those from English:
lazim ‘must’ < IA/EA lāzim; slip ‘sleep’ < English sleep. A feature typical of speakers with Iraqi or Egyptian Arabic as L1 is the substitution of /b/ for Romanian
or English /p/ and /v/: bibul ‘people, men’ < English people; gib ‘give, bring’ <
English give.
Consider next several selected features, generally typical of Pidgin Madame,
Jordanian Pidgin Arabic, and Gulf Pidgin Arabic. Pharyngeals are either replaced:
Pidgin Madame hareb ‘war’ < LA ḥareb; Jordanian Pidgin Arabic bisallih ‘repair’
< JA biṣalliḥ ‘repair.impf.3sg.m’; Gulf Pidgin Arabic aksan ‘best’ < GA aḥsan, hut
‘put’ < GA ḥuṭṭ ‘put.imp.2sg.m’; or deleted: Pidgin Madame ēki ‘cry’ < LA əḥki
‘cry.imp.2sg.f’; Jordanian Pidgin Arabic arabi ‘Arabic’ < JA ʕarabi; Gulf Pidgin
Arabic araf ‘know’ < GA ʕaraf. The pharyngealized consonants are replaced by
plain counterparts: Pidgin Madame sarep ‘envelope’ < LA ẓaref ; Jordanian Pidgin
Arabic bandora ‘tomato’ < JA banḍōra; Gulf Pidgin Arabic halas ‘finish’ < GA
ḫalāṣ; or they are realized as retroflex: Pidgin Madame ʈawīle ‘long’ < LA ṭawīle
‘long.f.sg’. The velar fricatives are replaced by velar stops or, less frequently, by
/h/: Pidgin Madame sokon ‘warm’ < LA suḫun ‘warm’, sogol < LA šəɣəl ‘work’;
Jordanian Pidgin Arabic kamsa ‘five’ < JA ḫamsa, sukul ‘work (n)’ < JA šuɣl, zagīr
‘small’ < JA ṣaɣīr; Gulf Pidgin Arabic kubus ‘bread’ < GA ḫubz; halas ‘finish’ <
338
15 Arabic pidgins and creoles
GA ḫalaṣ; yistokol ‘work’ < GA yištuɣul ‘work.impf.3sg.m’, šugl ‘work’ < GA šuɣl.
Geminate consonants generally undergo degemination (Næss 2008: 36; Avram
2014: 15): Jordanian Pidgin Arabic sitin ‘sixty’ < JA sittīn; Gulf Pidgin Arabic sita
‘six’ < GA sitta.
Moreover, consonants not found in the L1s of the users of Gulf Pidgin Arabic
may also be replaced. For instance, Indonesian, Javanese, Sinhalese and Tagalog
speakers may substitute [p] for /f/: Pidgin Madame palēpil ‘falafel’ < LA falēfil;
Jordanian Pidgin Arabic pi ‘in’ < JA fī ; Gulf Pidgin Arabic napar ‘person’ < GA
nafar; Indonesian and Sinhalese speakers may realize /z/ as [s] or [ʤ]: Pidgin
Madame esa ‘if’ < LA iza; Gulf Pidgin Arabic sēn ~ ʤēn ‘good’ < GA zēn (Bizri
2014: 393; Avram 2017b: 133). Bengali and Sinhalese speakers may replace /š/ with
[s]: Pidgin Madame sū ‘what’ < LA šū; Jordanian Pidgin Arabic su ‘what’ < JA
šū.
Finally, although phonetically long vowels do occur, vowel length is not distinctive, as shown by the occurrence of variation, e.g. Gulf Pidgin Arabic baden
~ badēn ‘then’ < GA baʕdēn.
4.3.2 Syntax
There is relatively little that can be attributed to SL agentivity in the syntax of
the Arabic-lexifier pidgins in the Middle East (Almoaily 2013; Al-Salman 2013;
Avram 2014; Bizri 2014; Avram 2017b; Bakir 2017).
Since the substrate of these varieties, with the exception of Romanian Pidgin
Arabic, consists of many SOV languages, e.g. Bengali, Hindi/Urdu, Malayalam,
Punjabi, Persian, Sinhalese, Tamil, this word order is occasionally attested (Avram 2017b: 133–134; Bizri 2014: 403). For instance, direct objects may occur in
pre-verbal position:
(23)
a. Pidgin Madame (Bizri 2010: 227)
misʈer kilot
sīli
mister underwear take off
‘Mister takes off his underwear.’
b. Gulf Pidgin Arabic (Avram 2017b: 133)
ana čiko sūp
1sg child see
‘I will see my children.’
In attributive possession constructions the order of constituents is possessor–
possessee:
339
Andrei Avram
(24)
a. Pidgin Madame (Bizri 2010: 198)
kullu māmā benet
all mother girl
‘All mother’s girls.’
b. Gulf Pidgin Arabic (Næss 2008: 87)
ana jawd
bādēn ysīr Jakarta stokol
1sg husband then go Jakarta work
‘Then my husband went to work in Jakarta.’
Adjectives generally precedes the nouns they modify:
(25)
Pidgin Madame (Bizri 2010: 119)
bīr bēt
big house
‘A big house.’
Similarly, adverbs precede the adjectives they modify:
(26)
a. Pidgin Madame (Bizri 2010: 119)
ʈīr gūɖ
very good
‘very good’
b. Gulf Pidgin Arabic (Avram 2014: 25)
sem-sem kalām
same
speak
‘They speak in the same way.’
Occasional instances of postpositions are attested:
(27)
a. Pidgin Madame (Bizri 2010: 132)
mister mayik masārē
mister with money
‘Mister has the money.’
b. Gulf Pidgin Arabic (Avram 2014: 25)
zamal fok
camel above
‘Above the camel.’
Interestingly, Pidgin Madame has a focalized negative copula, derived etymologically from English no:
340
15 Arabic pidgins and creoles
(28)
Pidgin Madame (Bizri 2010: 133)
māmā bīrūt no
mother Beirut neg.foc
‘It’s not in Beirut that my mother is.’
This resembles the Sinhalese negator nemiyi, which “is used only in focalized
phrases” (Bizri 2010: 69):
(29)
Pidgin Madame (Bizri 2010: 69)
bat kāve mama nemeyi
rice ate 1sg neg.foc
‘It is not I who ate the rice.’
4.3.3 Lexicon
Imposition under SL agentivity accounts for the fact that there are few instances
of transfer of lexical items from the various SLs into the non-dominant RL (i.e.
the pidgin).
The lexicon of Romanian Pidgin Arabic includes words of Romanian and English origin (Avram 2010: 32): mašina ‘car’ < Romanian maşină, sonda ‘oil rig’ <
Romanian sonda; spik ‘speak, say, tell’ < English speak, tumač ‘much, many’ <
English too much. Occasionally, non-Arabic words undergo semantic extension
under the influence of phonetically similar Arabic words (Avram 2010: 32): gib
‘give; bring’ < English give, cf. EA gīb ‘bring.imp.2sg.m’.
The lexicon of all the other pidginized varieties of Arabic spoken in the Middle East includes loanwords from English: Pidgin Madame ambasi ‘embassy’ <
English embassy; go ‘go’ < English go, kam ‘come’ < English come, no gūɖ ‘bad’
< English no good, oké ‘OK’ < English OK; Jordanian Pidgin Arabic bēbi ‘child’ <
English baby, finiš ‘finish’ < English finish, fisa ‘visa’ < English visa; Gulf Pidgin
Arabic hazband ‘husband’ < English husband, pēšent ‘patient’ < English patient.
However, as noted by Smart (1990: 113) concerning Gulf Pidgin Arabic, “it is difficult to say […] whether they are a true part of the pidgin” or rather nonce
borrowings.
Given the extreme diversity of the substrate, it is not surprising that only a
few words from the SLs have made it into the lexicon of Gulf Pidgin Arabic
(Avram 2017b: 134–135): ača ‘fine’ < Urdu achā ‘good, very well’, ǧaldi ~ ǧeldi
< Hindi/Urdu jaldī ‘quick’.
Jordanian Pidgin Arabic and Gulf Pidgin Arabic exhibit light-verb constructions which may well be calques on Bengali (noun/adjective + kara ‘make’) and/
341
Andrei Avram
or Hindi/Urdu (noun/adjective + karnā ‘make’) and/or Persian – noun/adjective
+ kardan ‘make’): Jordanian Pidgin Arabic sawwi zadīd ‘renew’, lit. ‘make new’;
Gulf Pidgin Arabic sawwi suāl ‘ask’, lit. ‘make a question’, sawwi zalān ‘upset’,
lit. ‘make angry’.
5 Conclusion
This chapter has shown that Arabic-lexifier contact languages emerged primarily through imposition under SL agentivity, in line with the typology of contact
languages (Winford 2005: 396; 2008: 128).
The effects of imposition are most obvious in the phonology, syntax and the
syntax-semantics interface, and to a lesser extent in the morphology and the lexicon. In the phonology, SL agentivity induces the loss or replacement of certain
phonemes not found in the SLs. However, there are also instances of imposition
in the sense of transfer from the SLs. As seen, for example, in Turku and Bongor
Arabic, some consonants occur only in loanwords from the substrate languages.
The occurrence of such loanwords confirms the fact that imposition under SL
agentivity may include transfer of lexical items into the RL. Borrowing under
RL agentivity has generally played a far less significant role in the development
of Arabic pidgins and creoles. As expected, it mostly involves transfer of lexical
items; these may lead to the borrowing of certain consonant phonemes, as seen
in, for example, Juba Arabic and Kinubi. Finally, borrowing has been shown to
include transfer of morphological material as well.
A notable difference between Juba Arabic and Kinubi on the one hand, and
the Arabic-lexifier pidgins in the Middle East on the other hand, resides in the
relative weight of imposition under SL agentivity and borrowing under RL agentivity. As we have seen, Juba Arabic and Kinubi exhibit the effects of both imposition in their earliest stage (i.e. pidginization), and of borrowing in their latest
stage (i.e. nativization). In contrast, imposition is pervasive in the Arabic-lexifier
pidgins in the Middle East, given that these varieties have not undergone nativization.
There are still a number of issues awaiting resolution. For instance, the identification of the SLs is rendered difficult by their number and typological diversity.
This difficulty is further compounded by the fact that some substrate languages
are still under-researched. This is particularly the case of the substrate languages
of Juba Arabic and Kinubi. Also, the distinction between substrate and adstrate
languages is blurred (Nakao 2012: 132), particularly when varieties emerge and
develop in situ, as, for example, with Juba Arabic. Further research also needs
342
15 Arabic pidgins and creoles
to consider the effects of the existence of a creole continuum in Juba Arabic as
well as of bilingual and monolingual speakers of the language on the relative importance of restructuring, imposition and borrowing. The extent of restructuring
and imposition, for instance, is presumably much greater in basilectal and L2 varieties, as opposed to acrolectal and monolingual varieties of the language. The
same holds for Bongor Arabic, which, as shown, appears to be undergoing depidginization. Last but not least, further investigations are necessary to establish
whether Gulf Pidgin Arabic is evolving towards stabilization, possibly becoming
closer to its lexifier via borrowing of morphological material, or is rather undergoing constant repidginization, essentially via imposition.
Further reading
) Miller (1993), Nakao (2012), and Luffin (2014) illustrate in detail substratal and
adstratal influence on Juba Arabic and Kinubi.
) Avram (2019) analyzes the substratal input in the lexicon of Turku.
) Avram (2017b) and Bakir (2017) discuss the various sources of Gulf Pidgin Arabic.
Abbreviations
1, 2, 3
ChA
EA
exs
foc
GA
IA
impf
JA
L1
L2
n
nc
1st, 2nd, 3rd person
Chadian Arabic
Egyptian Arabic
existential
focus
Gulf Arabic
Iraqi Arabic
imperfect (prefix conjugation)
Jordanian Arabic
first language
second language
noun
noun class
neg
pass
pst
pl
poss
prf
prox
rel
RL
SA
SL
sg
negative
passive
past
plural
possessive
perfect (suffix conjugation)
proximal
relative
recipient language
Sudanese Arabic
source language
singular
343
Andrei Avram
References
Al-Salman, Ibrahim. 2013. Jordanian Pidgin Arabic. Irbid: Yarmouk University.
(Master’s dissertation).
Almoaily, Mohammad. 2013. Language variation in Gulf Pidgin Arabic. Newcastle:
Newcastle University. (Doctoral dissertation).
Avram, Andrei A. 2010. An outline of Romanian Pidgin Arabic. Journal of
Language Contact 3. 19–37.
Avram, Andrei A. 2014. Immigrant workers and language formation: Gulf Pidgin
Arabic. Lengua y Migración 6(2). 7–40.
Avram, Andrei A. 2017a. Evidence of variation in early records of Nubi. Paper presented at the Eleventh Creolistics Workshop “Assessing old assumptions: New
insights on the dynamics of languages”, Justus-Giebig University, Giessen,
March 22–24.
Avram, Andrei A. 2017b. Sources of Gulf Pidgin Arabic features. In Simone
Bettega & Fabio Gasparini (eds.), Linguistic studies in the Arabian Gulf, 131–
151. Turin: Università di Torino.
Avram, Andrei A. 2019. Africanisms in Turku. In Catherine Miller, Alexandrine
Barontini, Marie-Aimée Germanos, Jairo Guerrero & Christophe Pereira (eds.),
Studies on Arabic dialectology and sociolinguistics: Proceedings of the 12th conference of AIDA held in Marseille from May 30th to June 2nd 2017, 15–24. Aix-enProvence: Institut de recherches et d’études sur le monde arabe et musulman.
Bakir, Murtadha J. 2017. GPA genesis: The extent of substratal influence. Paper
presented at the Annual Summer Conference of the Society for Pidgin and
Creole Linguistics, University of Tampere, 19–22 June.
Bakker, Peter. 1995. Pidgins. In Jacques Arends, Pieter Muysken & Norval
Smith (eds.), Pidgins and creoles: An introduction, 130–157. Amsterdam: John
Benjamins.
Bakker, Peter. 2008. Pidgins versus creoles and pidgincreoles. In Silvia
Kouwenberg & John Victor Singler (eds.), The handbook of pidgin and creole
studies, 130–157. Malden, MA: Wiley-Blackwell.
Bizri, Fida. 2010. Pidgin Madame: Une grammaire de la servitude. Paris: Geuthner.
Bizri, Fida. 2014. Unity and diversity across Asian migrant Arabic pidgins in the
Middle East. Journal of Pidgin and Creole Languages 29(2). 385–409.
Bureng Vincent, George. 1986. Juba Arabic from a Bari perspective. In Gerrit
J. Dimmendaal (ed.), Current approaches to African linguistics, vol. 3, 71–78.
Dordrecht: Foris.
Cook, Albert R. 1905. Dinka and Nubi Vocabulary. Unpublished manuscript.
344
15 Arabic pidgins and creoles
De Angelis, Pietro. 2002. Vocabolario zande. Trieste: Edizioni Università di
Trieste.
Jenkins, Edward V. 1909. English–Arabic vocabulary with grammar and phrases
representing the language as spoken by the Uganda Sudanese in the Uganda and
British East Africa Protectorate. Kampala: The Uganda Company.
Lucas, Christopher. 2015. Contact-induced language change. In Claire Bowern &
Bethwyn Evans (eds.), The Routledge handbook of historical linguistics, 519–536.
London: Routledge.
Luffin, Xavier. 2004. Un créole arabe: Le kinubi de Mombasa. Etude descriptive.
Brussels: Université Libre de Bruxelles. (Doctoral dissertation).
Luffin, Xavier. 2005. Un créole arabe: Le kinubi de Mombassa, Kenya. Munich:
Lincom Europa.
Luffin, Xavier. 2013. Some new information about Bongor Arabic. In Mena
Lafkioui (ed.), African Arabic: Approaches to dialectology, 161–183. Berlin: De
Gruyter Mouton.
Luffin, Xavier. 2014. The influence of Swahili on Kinubi. Journal of Pidgin and
Creole Languages 29(2). 299–318.
Manfredi, Stefano. 2014a. Qualche osservazione sull’espressione del plurale
nominale nel creole arabo di Juba. RiCognizioni 2(1). 53–61.
Manfredi, Stefano. 2014b. Strategie di rilessificazione in Juba Arabic. In Alberto
Manco & Domenico Silvestri (eds.), L’etimologia: Atti del XXXV Convegno della
Società di Glottologia, 305–311. Rome: Il Calamo.
Manfredi, Stefano. 2017. Arabi Juba: Un pidgin-créole du Soudan du Sud. Leuven:
Peeters.
Manfredi, Stefano. 2018. Arabic as a contact language. In Reem Bassiouney &
Elabbas Benmamoun (eds.), The Routledge handbook of Arabic linguistics, 407–
20. London: Routledge.
Meldon, J. A. 1913. English–Arabic dictionary of words and phrases used by the
Sudanese in Uganda. Unpublished manuscript.
Miller, Catherine. 1989. Bari interference in Juba Arabic. In Lionel Bender
(ed.), Proceedings of the 4th Nilo-Saharan Symposium, 1–10. Hamburg: Helmut
Buske.
Miller, Catherine. 1993. Restructuration morpho-syntaxique en Juba-Arabic et
Ki-Nubi: A propos du débat universaux/substrat et superstrat dans les études
créoles. Matériaux Arabes et Sudarabiques (GELLAS) 5. 137–174.
Miller, Catherine. 2001. Grammaticalisation du verbe ‘dire’ et subordination en
Juba Arabic. In Robert Nicolaï (ed.), Leçons d’Afrique: Filiation, rupture et reconstitution des langues. Un hommage à G. Manessy, 455–482. Leuven: Peeters.
345
Andrei Avram
Moi, Daniel R., Mario L. B. Kuduku, Mary M. Michael, Simon H. John, Raphael Z.
Mafoi & Nyoul G. Kuduku. 2014. Bongo grammar book. Juba: SIL-South Sudan.
Muraz, Gaston. 1926. Vocabulaire du patois arabe tchadien ou “Tourkou” et des
dialectes sara-madjinngaye et sara-m’baye (S.-O. du Tchad). Paris: Charles
Lavauzelle.
Nakao, Shuichiro. 2012. Revising the substratal/adstratal influence on Arabic
creoles. In Osamu Hieda (ed.), Challenges in Nilotic linguistics: Phonology,
morphology, and syntax, 127–49. Tokyo: ILCAA.
Nakao, Shuichiro. 2015. Another story of Bangala: Language contact with South
Sudanese Arabic Creole. Paper presented at the 8th World Congress of African
Linguistics, Kyoto University, 20–24 August.
Næss, Unn Gyda. 2008. “Gulf Pidgin Arabic”: Individual strategies or a structured
variety? A study of some features of the linguistic behaviour of Asian migrants
in the Gulf countries. Oslo: University of Oslo. (Master’s dissertation).
Nebel, Arthur. 1979. Dinka–English, English–Dinka dictionary. Bologna: Editrice
Missionaria Italiana.
Owen, H. R. & G. J. Keane. 1915. An abbreviated vocabulary in Hindustani Luganda
Lunyoro Swahili Nubi: Designed for the use of the Uganda Medical Service.
Bukalasa: White Father’s Printing Press.
Owens, Jonathan. 1985. The origins of East African Nubi. Anthropological
Linguistics 27(2). 229–271.
Owens, Jonathan. 1997. Arabic-based pidgins and creoles. In Sarah G. Thomason
(ed.), Contact languages: A wider perspective, 125–172. Amsterdam: John
Benjamins.
Pozzati, Aurelio & G. Panza. 1993. Dizionario Giur. Trieste: Edizioni Università di
Trieste.
Smart, Jack R. 1990. Pidginization in Gulf Arabic: A first report. Anthropological
Linguistics 32(1–2). 83–119.
Tosco, Mauro & Stefano Manfredi. 2013. Pidgins and creoles. In Jonathan Owens
(ed.), The Oxford handbook of Arabic linguistics, 496–519. Oxford: Oxford
University Press.
Tosco, Mauro & Jonathan Owens. 1993. Turku: A descriptive comparative study.
Sprache und Geschichte in Afrika 14. 177–267.
Trudgill, Peter. 2011. Sociolinguistic typology: Social determinants of linguistic
complexity. Oxford: Oxford University Press.
Van Coetsem, Frans. 1988. Loan phonology and the two transfer types in language
contact. Dordrecht: Foris.
Van Coetsem, Frans. 1995. Outlining a model of the transmission phenomenon
in language contact. Leuvense Bijdragen 88. 63–85.
346
15 Arabic pidgins and creoles
Van Coetsem, Frans. 2000. A general and unified theory of the transmission process
in language contact. Heidelberg: Winter.
Van Coetsem, Frans. 2003. Topics in contact linguistics. Leuvense Bijdragen 92.
27–99.
Wellens, Inneke. 2003. The Nubi language of Uganda: An Arabic creole in Africa.
Nijmegen: Catholic University of Nijmegen. (Doctoral dissertation).
Winford, Donald. 2005. Contact-induced change: Classification and processes.
Diachronica 22(2). 373–427.
Winford, Donald. 2008. Processes of creole formation and related contactinduced language change. Journal of Language Contact 2(1). 124–145.
347
Part II
Language change through contact
with Arabic
Chapter 16
Modern South Arabian languages
Simone Bettega
Università degli Studi di Torino
Fabio Gasparini
Freie Universität Berlin
In the course of this chapter we will discuss what is known about the effects that
contact with Arabic has had on the Modern South Arabian languages of Oman
and Yemen. Documentation concerning these languages is not abundant, and even
more limited is our knowledge of the history of their interaction with Arabic. By
integrating the existing bibliography with as yet unpublished fieldwork materials,
we will try to provide as complete a picture of the situation as possible, also discussing the current linguistic and sociolinguistic landscape of Dhofar and eastern
Yemen.
1 History of contact between Arabic and the Modern
South Arabian languages
Much to the frustration of modern scholars of Semitic, the history of the Modern
South Arabian languages (henceforth MSAL) remains largely unknown.1 To this
day, no written attestation of these varieties has been discovered, and it seems
safe to assume that they have remained exclusively spoken languages throughout
all of their history. Since European researchers became aware of their existence
in the first half of the nineteenth century (Wellsted 1837), and until very recently,
the MSAL were thought by many to be the descendants of the Old (epigraphic)
1
§1 was authored by Simone Bettega, while §2 was authored by Fabio Gasparini. §3 and §4 are
the result of the conjoined efforts of both authors. In particular, Gasparini was responsible for
analyzing most of the primary sources and raw linguistic data, while Bettega worked more
extensively on the existing bibliography.
Simone Bettega & Fabio Gasparini. 2020. Modern South Arabian languages. In
Christopher Lucas & Stefano Manfredi (eds.), Arabic and contact-induced change, 351–
369. Berlin: Language Science Press. DOI:10.5281/zenodo.3744531
Simone Bettega & Fabio Gasparini
South Arabian languages (Rubin 2014: 16). This assumption has been conclusively
disproven by Porkhomovsky’s (1997) article, which also contributed significantly
to the re-shaping of the proposed model for the Semitic family tree. This modified version of the family tree (which finds further support in the recent works
of Kogan 2015 and Edzard 2017) sets the MSAL apart as an independent branch
of the West Semitic subgroup, one whose origins are therefore of considerable
antiquity. This brings us to the question of when it was that the MSAL (or their
forebears) first came into contact with Arabic. This might have happened at any
time since Arabic-speaking people started to penetrate into southern Arabia, a
process that – as we know from historical records – began in the second half of
the first millennium BCE (Robin 1991; Hoyland 2001: 47–48). Roughly one thousand years later, almost the whole population of central and northern Yemen was
speaking Arabic, and possibly a considerable portion of the southern population
as well (Beeston 1981: 184; Zammit 2011: 295). It is therefore possible that Arabic and the MSAL have been in contact for quite some time, and it seems likely
that the intensity and effects of such contact grew stronger after the advent of
Islam (Lonnet 2011: 247). It is also possible, as some scholars have written, that
the MSAL “represent isolated forms that were never touched by Arabic influence
until the modern period” (Versteegh 2014: 127). Admittedly, evidence to support
either one of these hypotheses is scarce, and at present it is probably safer to
say that our knowledge of the history of contact between Arabic and the MSAL
before the twentieth century is fragmentary at best. This is why studies on the
outcomes of such contact are of particular interest, since they could help to shed
light on parts of that history. This is also why, in the course of this chapter, we
will refrain from addressing the question of how contact with the MSAL affected
the varieties of Arabic spoken in Oman and Yemen, and focus solely on the influence of Arabic on the MSAL. Although there is plenty of evidence that South
Arabian exerted a powerful influence on the Arabic of the area (see for instance
Retsö 2000 and Watson 2018),2 it is often difficult to assess whether this influence is the result of contact with forms of Old South Arabian or more recent
interaction with the MSAL. Such a discussion, also because of space constraints,
is beyond the scope of the present article.
As far as the interaction between Arabic and the MSAL in the twentieth and
twenty-first centuries is concerned, Morris (2017: 25) provides a good overview
of the multilingual environment in which the MSAL were and are spoken:
2
To the point that so-called mixed varieties are reported to exist, whose exact linguistic nature
seems difficult to pinpoint. See Watson et al. (2006) and Watson (2011) for discussion.
352
16 Modern South Arabian languages
Speakers of [a Modern South Arabian] language always had to deal with
speakers of other MSAL, as well as with speakers of various dialects of Arabic. The Baṭāḥirah, for instance, did nearly all their trade with boats from
Ṣūr and other Arabic-speaking ports; they lived and worked with the Arabicspeaking Janaba, while being in contact with speakers of Ḥarāsīs and Mahra.
The Ḥarāsīs interacted with the Arabic speakers surrounding their Jiddat alḤarāsīs homeland, traded in the Arabic-speaking markets of the north, and
in the summer months went to work at the northern date harvest. Mehri
speakers lived beside and traded with Arabic-speaking Kathīri tribesmen
in the Nejd region, Śḥerɛt speakers in the mountains, and Arabic speakers in the coastal market towns of Dhofar. Śḥerɛt speakers interacted with
the Mahra, some of whom settled among them, and with Arabic-speaking
peoples of the coast as well as the desert interior […] There was marriage
between Arabic-speaking men of the coastal towns and MSAL-speaking
women of the interior, and over time, families of Mehri and Śḥerɛt speakers
settled in or near the towns, with the result that even more Arabic speakers
became familiar with these languages. (Morris 2017: 25)
2 Current state of contact between Arabic and the
Modern South Arabian languages
Today, six Modern South Arabian languages exist, spoken by around 200,000 people in eastern Yemen (including the island of Soqotra) and western Oman. These
six languages are: Mehri, Hobyōt, Ḥarsūsi, Baṭḥari, Śḥerɛt/Jibbāli and Soqoṭri.
They are all to be regarded as endangered varieties, though the individual degree of endangerment varies remarkably. No exact census concerning the number of speakers is currently available (Simeone-Senelle 2011: 1075), but we know
that Mehri is the most spoken language, with an estimated 100,000 speakers. It
is followed by Soqoṭri (about 50,000 speakers), Śḥerɛt (25,000), Ḥarsūsi (a few
hundred), Hobyōt (a few hundred) and Baṭḥari (less than 20 speakers). The main
causes of endangerment are reckoned to be shift to Arabic and the disappearance
of traditional local lifestyles. In addition, the current political situation in Yemen
is having effects on the linguistic landscape of the region which are difficult to
document or foresee: the area is currently inaccessible to researchers, and there
is no way to know how the conflict will affect the local communities.
353
Simone Bettega & Fabio Gasparini
As far as Oman is concerned, the city of Salalah undoubtedly represents the
major locus of contact between Arabic and the MSAL. The rapid growth the city
has witnessed in recent years, and the improved possibilities of economic development that came with it, have led many Śḥerɛt speakers from the nearby
mountains to settle in the city or its immediate surroundings, where they now
employ Arabic on a daily basis as a consequence of mass education and media,
neglecting other local languages. This has led to a split, in the speakers’ perception, between “proper” Śḥerɛt, spoken in the mountains, and the “city Śḥerɛt” of
Salalah, often regarded as a sort of “broken” variety of the language in which,
among other things, code-switching with Arabic is extremely frequent. Unfortunately, data on this subject are virtually non-existent, given the extreme difficulty
of documenting such an episodic phenomenon (aggravated by speakers’ understandable reluctance to having their imperfect language proficiency evaluated
and recorded).
Even outside the urban centers, however, contact with Arabic is on the rise.
Even the most isolated variety, Soqoṭri, is apparently undergoing rapid change
under the influence of Arabic: the existence of a koinéised variety of Soqoṭri,
heavily influenced by Arabic, has been recently reported in Ḥadibo (Morris 2017:
27). This is not to say, of course, that all MSAL are being affected to the same
degree: Watson (2012: 3), for instance, notes how “Mahriyōt [the eastern Yemeni
variety of Mehri] […] exhibits structures unattested in Mehreyyet [Mehri Omani
variety] […] and shows greater Arabic influence both in terms of the number of
Arabic terms used, and the length and frequency of Arabic phrases within texts.”
However, no MSAL seems at present to be exempt from the effects of contact.
The case of Baṭḥari, which, as we have seen, is the most severely endangered of
all the MSAL, exemplifies well the processes of morphological loss and erosion
that a language undergoes in the final stages of endangerment. Morris (2017)
reports how already in the 1970s Baṭḥari seemed to display many of the signs of
a moribund language. In recent times:
[t]he younger generations showed little interest in their former language;
they were eager to embrace Arabic and to feel themselves part of the wider
Arabic Islamic community; and they were proud to call themselves ‘ʕarab’,
with all that word’s overtones of Bedouin ancestry and code of honour.
(Morris 2017: 11)
In the following sections we will discuss several types of contact-induced
changes in the MSAL. Although we will use material taken from all varieties,
Baṭḥari will be in particular focus due to its singular status.
354
16 Modern South Arabian languages
3 Contact-induced changes in the MSAL
As already noted, in the course of this chapter we will focus solely on the effects that contact with Arabic has had on the various MSAL. Therefore, Arabic
will always be the source language of all the transfer phenomena considered in
the next pages, while the recipient language will be, depending on the different
examples, one or the other of the six MSAL. Obviously, this poses the question
of who the agents of change are and were in the case of these particular phenomena, and what type(s) of transfer are we confronted with (cf. Van Coetsem
1988; 2000; Winford 2005). According to the overview of the MSAL’s sociolinguistic status presented above, it should be clear by now that, while the two cases
are extremely common of (a) mono- or multilingual MSAL speakers who acquire
Arabic as an L2 and (b) bilingual MSAL–Arabic speakers, the opposite is not true
(that is, monolingual Arabic speakers who come to acquire one or more MSAL
as L2s later in life). In other words, all the transfer phenomena we will be considering in the next paragraphs are either instances of borrowing (brought about
by speakers who are dominant in one or more MSAL) or convergence (brought
about by speakers who are native speakers of Arabic and at least one MSAL; see
Lucas 2015 for a definition of convergence).
3.1 Phonology
3.1.1 Phonetic adaptation of loanwords
As illustrated in §3.4, lexical borrowings from Arabic are extremely common in
the MSAL. As Morris (2017: 13) remarks, such loanwords are often altered in order
for them to acquire a “South Arabian flavour”, so to speak. The phenomenon
is not one of simple adaptation dictated by difficulty of articulation, since the
sounds that are replaced are present in the phonological inventory of the MSAL.
In fact, the opposite appears to be true, these sounds normally being replaced
by others which are typical of South Arabian but absent in Arabic. For Baṭḥari,
Morris gives the example of Arabic pharyngealised dental fricative /ð̣/ (IPA [ðʕ ])
being replaced by the pharyngealised alveolar lateral fricative /ṣ́/ (mostly realised
as IPA [ɮʕ ], see §3.1.4), as in raṣ́ṣ́ ‘bruise’ (from Janaybi Arabic rað̣ð)̣ , or Arabic /š/
(IPA [ʃ]) being replaced by /ś/, as in men śān-k ‘for you, for your sake’, in place
of men šān-k, śarray ‘buyer’ for šarray, or śəmāl ‘inland, north’ for šəmāl (while
Baṭḥari śēməl(i) is normally used to refer to the left hand only).
Lexical borrowing can also be the cause of variation in the realisation of certain
sounds, as is the case with the phonemes /g/ and /y/ (IPA [ɡ] and [j] respectively),
which represent different reflexes of Proto-Semitic *g in different Omani Arabic
355
Simone Bettega & Fabio Gasparini
dialects. It is possible to find traces of this variation in those MSAL that are in
contact with more than one variety of Arabic, as is the case with Ḥarsusi: see for
instance fagr and fayr, both meaning ‘dawn’, or the opposition between yann
‘madness’ and genni ‘jinni’, both from the same etymological root (Lonnet 2011:
299).
3.1.2 Affrication of /k/ > [ʧ]
It can also be the case that some phonetic processes regularly taking place in
the local Arabic varieties but otherwise unknown to MSAL phonology are transferred to original MSAL vocabulary. This is what happens in Baṭḥari, where
some speakers may show an affricate realisation of the voiceless occlusive [k]
> [ʧ], which resembles the Janaybi Arabic realisation of the phoneme /k/ (whose
complementary distribution with the voiceless plosive realisation [k] is still unclear). For example, some speakers regularly produce /yənkaʕ/ ‘come.3sg.m.sbjv’
as [jənˈʧaʕ] instead of [jənˈkaʕ].
3.1.3 Stress
The structural similarity of Arabic and the MSAL can sometimes cause stress
patterns which are typical of the former to be applied to the latter, as is the case
with ‘she began’: Soqoṭri bédʔɔh, (local) Arabic bədáʔat, Soqoṭri with an Arabic
stress bədɔ́ʔɔh (Lonnet 2011: 299).
3.1.4 Realisation of emphatics
This is a topic that has attracted the attention of several scholars since the publication of Johnstone’s (1975) article on the subject, because of the realisation
of the so-called Semitic “emphatics” as glottalised consonants. Glottalisation is
a secondary articulatory process in which narrowing (creaky voice) or closure
(ejective realisation) of the glottis takes place: the action of the larynx compresses
the air in the vocal tract which, once released, produces a greater amplitude in
the stop burst (Ladefoged & Maddieson 1996: 78).
Lonnet (2011: 299) notes a tendency for speakers of various MSAL to replace
the ejective articulation of certain consonants (especially fricatives, see Ridouane
& Gendrot 2017) with a pharyngealised realisation, typical of Arabic emphatics. Pharyngealisation is a kind of secondary articulation involving a constriction of the pharynx usually realised through tongue-root retraction, resulting
in a backed realisation (Ladefoged & Maddieson 1996: 365). This process is welldocumented across Semitic languages. Naumkin & Porkhomovsky (1981) note for
356
16 Modern South Arabian languages
Soqoṭri an ongoing process of transition from a glottalised to a pharyngealised
realisation of emphatics, with only stops being realised as fully glottalised items.
Work by Watson & Bellem (2010; 2011) and Watson & Heselwood (2016) shows the
co-occurrence of pharyngealisation and glottalisation in relation to pre-pausal
phenomena in Ṣanʕāni Arabic, Mahriyōt and Mehreyyet (respectively the westernmost Yemeni and Omani varieties of Mehri). Dufour (2016: 22) states that
“le caractère éjectif des phonèmes emphatiques ne fait aucun doute, en jibbali
comme en mehri” (“the nature of the emphatic phonemes is undoubtedly ejective, in Jibbali as much as in Mehri”).3 Finally, in Baṭḥari only /ḳ/ is realised as
a fully ejective consonant [k’]. /ṭ/ and the fricative emphatics, on the the other
hand, are described as mainly pharyngealised (and partially voiced, in the case
of fricatives; Gasparini 2017).
Unfortunately, since there is no thorough phonetic description of any MSAL
that predates the 1970s, it is impossible to ascertain whether these realisations
(which, again, range from fully glottalised to fully pharyngealised) are the result
of the influence of Arabic, or have arisen as the consequence of internal and typologically predictable developments. It is likely, though, that bilingualism and
constant contact with Arabic have at least favoured this phonetic change. Evidence in support of this view may come from the fact that speakers who are
poorly proficient in Arabic and live in rural and more isolated areas are more
likely to preserve a glottalised realisation of the emphatics (as emerges from direct fieldwork observations).
3.2 Morphology
3.2.1 Nominal morphology
Morphological patterns which are typical of Arabic can enter a language through
borrowing, as is the case with the passive participle pattern for simple verbs,
which is mVCCūC in Arabic and mVCCīC in MSAL. Soqoṭri maḫlɔḳ, for instance,
is clearly derived from Arabic maḫlūq ‘human being’ (lit. ‘created’), while this is
not the case for Ḥarsusi mḫəlīḳ (Lonnet 2011: 299). Also, in the realm of verbal
derivational morphology, certain phenomena can be introduced into the recipient language through lexical borrowing: this is the case with gemination and
prefixation of t- in Ḥarsusi, as in the participle mətḥaffi ‘barefoot’ (from Omani
Arabic mitḥaffi; Lonnet 2011).
In general, Arabic loanwords are normally well integrated in MSAL morphology, probably because of the high degree of structural similarity that exists be3
Authors’ translation.
357
Simone Bettega & Fabio Gasparini
tween these languages. One example, reported by Lonnet (2011), is that of bəḳerēt
‘cow’, a fully integrated loan from Arabic used in Ḥarsusi and Western Yemeni
Mehri, which possesses its own plural and diminutive form (bəḳār and bəḳərēnōt,
respectively).
Arabic loans in several MSAL stand out because of their characteristic feminine ending in -V(h) instead of -(V)t, as in Śḥehri saʕah ‘watch’ and ṭorəh ‘revolution’ (but consider the more adapted ris͂t ‘trigger’ from Omani Arabic rīšah;
Lonnet 2011).
It is also worth noting that the Arabic ending -V(h) is replaced by its MSAL
equivalent when the noun is in the construct state, that is, final -t reappears. This
would also happen in Arabic, but the alteration in the quality of the vowel is a
clear signal that the suffix is to be considered an MSAL morpheme. Consider the
following example from Morris’ Baṭḥari recordings:4
(1)
Baṭḥari
mʕayš-it-həm
bəss mʕayš-it-həm
ḥawla ʕār ḥāmis
sustenance-f.cs-3pl.m only sustenance-f.cs-3pl.m once only turtle
ḥawīl
mʕayš-ah
w-ṣayd śālā
and-fish nothing sustenance-f once
‘Their sustenance, only that! Their sustenance was once only turtle and
fish, there was nothing to eat in the past.’
The word mʕayšah ‘sustenance, food’ is a loanword from Arabic (as the -ah
ending suggests). When suffixed with the possessive 3pl.m pronoun -həm, however, Baṭḥari -it replaces -ah/-at (note also, in the example, the use of the restrictive adverbial particle bəss ‘only’, which is a well-integrated loan from dialectal
Arabic and occurs in alternation with Baṭḥari ʕār).
Finally, in Baṭḥari the Arabic definite article (a)l- is occasionally used instead
of the MSAL definite article a-: bə-l-ḫarifēt ‘during the rainy season’.
3.2.2 Pronouns
The influence of Arabic can be observed, to an extent, even in the pronominal
system, especially in those MSAL that are more exposed to contact due to the
limited size of their speech communities. Lonnet (2011), for instance, reports how,
4
Audio file 20130929_B_B02andB04_storyofcatchingturtle recorded, transcribed and kindly
shared with Fabio Gasparini by Miranda Morris. The recording was produced in the context
of Morris’ and Watson’s “Documentation and Ethnolinguistic Analysis of the Modern South
Arabian languages” project, funded by the Leverhulme Trust. More recordings are accessible
at the ELAR archive of SOAS University of London. The transcription has been adapted.
358
16 Modern South Arabian languages
despite the fact that in the MSAL the first person suffix pronoun is normally an
invariable -i, in Ḥarsusi this can be replaced by -ni after verbs and prepositions (as
is the case with Arabic; see also §3.3 for another interesting example concerning
the marking of pronominal direct objects).
In addition, Baṭḥari relative pronouns (sg: l-, lī pl: əllī) are close to their equivalent in Janaybi Arabic (and diverge from the rest of MSAL, where a ð- element
can be found). Baṭḥari has also borrowed the reflexive pronoun ʕamr- ‘oneself’
from the Arabic dialect of the Janaba, despite the existence of an original Baṭḥari
term with the same meaning, ḥanef - (note that both terms must always be followed by a suffix personal pronoun). ʕamr- has also been given a plural form in
Baṭḥari, based on MSAL derivational patterns, ḥaʕmār- (Morris 2017: 14).
3.2.3 Baṭḥari verbal plural marker -uw
Baṭḥari differs from the rest of the MSAL in that all 2/3pl.m verbal forms are
marked by an -uw suffix, while in the other languages of the group these persons are marked by a -Vm ending and/or by internal vowel change (e.g. Mehri
and Ḥarsusi -kə(u)m for the 2pl.m and -ə(u)m/ umlaut for the 3pl.m of the perfective conjugation; t-…-ə(u)m and y/i-…-ə(u)m respectively for 2 and 3pl.m of the
imperfective conjugation; Simeone-Senelle 2011: 1093–1094).
The origin of this suffix is uncertain. Its presence might well be connected to
contact with Arabic (neighboring dialects have an -u or -ūn suffix in the 3pl.m
person of the verb in both the perfective and imperfective conjugation) or to
otherwise unattested stages of development internal to the MSAL verbal system.
In this regard, Rubin (2017: 5) suggests for Mehreyyet the presence of a subjacent
-ə- in 2nd/3rd plural masculine verbal suffixes which could therefore be somehow
related to the Baṭḥari -uw marker. However, the optional simultaneous presence
of apophony within the stem of 3pl.m verbal forms (similarly to what happens
elsewhere in the MSAL), together with scarcity of data, prevents any conclusive
assessment of the topic.
3.3 Syntax
At present, the syntax of the various MSAL has not been made the object of detailed investigation. The only scientific work dealing with this topic is Watson’s
(2012) in-depth analysis of Mehri syntax. However, Watson’s thorough description provides only sporadic insights into the issue of language contact (as for
instance the use in Mahriyōt, the eastern Yemeni variety of Mehri, of a swē ~
amma… yā construction to express polycoordination, probably to be regarded as
359
Simone Bettega & Fabio Gasparini
the result of Arabic influence; Watson 2012: 298). In general, though, the topic is
left unaddressed in the literature, and more research is needed.
Gasparini’s data on Baṭḥari offer an interesting example of Arabic influence on
MSAL syntax. In Baṭḥari, as in the other MSAL, pronominal direct and indirect
objects may require a particle t- to be inserted between them and the verb, depending on the morphological form of the verb itself. Masculine singular imperatives, for instance, require the presence of the marker, as shown in the following
example:
(2) Baṭḥari (Gasparini, unpublished data):
zum
t-ī
t-ih
give.imp acc-1sg acc-3sg.m
‘Give it to me.’
Example (3), in contrast, shows that the pronominal indirect object -(ə)nī is
suffixed directly to the verb as it would be in Arabic (see §3.2.2).
(3) Baṭḥari (Gasparini 2018: 66):
zɛm-ənī
θrɛh
give.imp-1sg two.m
‘Give me both (of them).’
In other words, the introduction of the Arabic form of the object pronoun has
caused the Baṭḥari object marker to disappear. Note that informants judged the
alternative construction zum t-ī θrɛh, (with the use of the object marker t- and the
1sg object pronoun marker -ī ) to be acceptable, but this form was not produced
spontaneously.
A peculiarity of the MSAL spoken in Oman is the use of circumstantial qualifiers, a type of clausal subordination well attested in Gulf Arabic (Persson 2009).
Baṭḥari regularly introduces predictive and factual conditional clauses asyndetically by using the structure [sbj.pro w-sbj.pro]. Consider (4):
(4)
360
Baṭḥari (Gasparini, unpublished data)
hēt w-hēt
aṣbaḥ-k
aḫayr saḥīr-e
t-ōk
2sg.m and-2sg.m wake_up.prf-2sg.m better brand.ptcp-pl.m acc-2sg.m
w-marað̣ zēd
aḫass hāmā-k?
lā w ham aṣbaḥ-k
neg and if wake_up.prf-2sg.m worse hear.prf-2sg.m and-illness huge
l-ōk
nḥā saḥīr-e
t-ōk
śkīl-e
t-ōk
to-2sg.m 1pl brand.ptcp-pl.m acc-2sg.m scar.ptcp-pl.m acc-2sg.m
16 Modern South Arabian languages
mən
a-gab
because_of def-infection.
‘In case you wake up feeling better / (we) do not brand you but in case
you wake up feeling worse / do you understand? And you are seriously
ill / we brand you and scar you because of the infection.’
The first clause hēt w-hēt aṣbaḥ-k aḫayr is an asyndetical circumstantial qualifier functioning as a predictive conditional clause. It contrasts with w ham aṣbaḥk aḫass, in which the conjunction ham introduces a counterfactual conditional
clause.
In Omani Mehri conditional clauses are commonly introduced through conjunction of pronouns (Watson et al. forthcoming: 211). This structure is unattested in Yemeni Mehri:
(5)
Mehri (Watson et al. forthcoming: 211)
sēh wa-sēh t-ḥam-ah
lā ḥib-sa
yi-ḳal-am
3sg.f and-3sg.f 3sg.f-want.impf-3sg.m neg parents-3sg.f 3m-let.impf-pl
t-ēs
ta-ghōm
š-ih
lā
acc-3sg.f 3sg.f-go.sbjv with-3sg.m neg
‘If she doesn’t want him, her parents won’t let her go with him.’
These uses closely resemble those of Gulf Arabic, where circumstantial qualifiers are widely attested to codify predictive and factual conditional and consecutive clauses.
3.4 Lexicon
In the case of the MSAL, it can often be difficult to clearly set apart the effects
of Arabisation from those of modernisation and lifestyle changes (which is not
surprising, since the two phenomena are interrelated). According to what the
speakers themselves report,
it was only since the introduction of formal education, and the awareness
of [Modern Standard Arabic] via the media, that Arabic became the second
language for many of the MSAL speakers in Dhofar, and, in the case of
younger speakers, often to the detriment of their proficiency in their MSAL
variety (Davey 2016: 11).
As a consequence, phenomena of borrowing (such as code-switching and loanwords) are particularly common, especially in those varieties (and in the idiolects
361
Simone Bettega & Fabio Gasparini
of individuals) that are more exposed to Arabic. The following is a good example
of code-switching in Baṭḥari (note that the speaker in question tended to employ
Janaybi forms more than other informants):
(6) Baṭḥari (Gasparini, unpublished data)
mɛ̄t
məssəlīm nə-šāhəd
l-ōk
die.prf.3sg.m muslim 1pl-say_šahada.impf for-2sg.m
w-y-sabbah-uw
w-y-kabbər-uw w-y-hālul-uw
and-3m-pray.impf-pl and-3m-pray-pl and-3m-praise_allah-pl
‘(If) a Muslim dies / we say the šahada for you / and they pray and say
‘allāhu ʔakbar’ and praise Allah.’
In (6) the speaker makes use of several Arabic verbs related to the semantic field of religious practices, which are not lexically encoded in Baṭḥari. This
might indirectly show the introduction of new ritual practices at a certain point
of the history of the tribe. Note that C2 -geminate stems such as ysabbahuw and
ykabbəruw represent verbal patterns not attested in MSAL morphology, and are
therefore easily identifiable as loans.
Morris (2017: 15) makes the important remark that lexical erosion is directly
connected with the loss of importance of a language in the eyes of its speakers.
She gives the example of the Baṭḥari word for ‘home, living quarters’, for which
speakers nowadays frequently resort to some version of Arabic bayt, while the
many possible original synonyms are falling into disuse. Many of these (kədōt,
mōhen, mašʕar, mōḫayf, and ḫader) are connected to traditional ways of living
which have all but disappeared in the course of the last 40–50 years, so that
speakers probably judge them inadequate to refer to modern built houses.
3.4.1 Numerals
Watson (2012: 3) reports that “[w]hile Mehri cardinal numbers are typically used
for both lower and higher cardinals in Mehreyyet, Mahriyōt speakers, in common
with speakers of Western Yemeni Mehri, almost invariably use Arabic numbers
for cardinals above 10.” This type of lexical substitution connected to numerals
higher than ten is also mentioned by Lonnet (2011) and Simeone-Senelle (2011:
1088), who states that “[n]owadays the MSAL number system above 10 is only
known and used by elderly Bedouin speakers.” Watson & Al-Mahri (2017: 90)
note that it is mostly younger generations (especially in urban settings) who have
lost the ability to count beyond ten. Interestingly, they point out that telephone
numbers are given exclusively in Arabic, “possibly due to the lack of a singleword MSAL equivalent to Arabic ṣufr ‘zero’.”
362
16 Modern South Arabian languages
3.4.2 Spatial reference terms
According to Watson & Al-Mahri (2017: 91) the MSAL employ topographically
variable absolute spatial reference terms. In other words, these terms can differ
depending on the language employed, the moment of the utterance and the position of the speaker in relation to absolute points of reference. For instance, in and
around the city of Salalah in Dhofar, both in the mountains and on the coastal
plain, the equivalents of the words for ‘sea’ and ‘desert’ are used to indicate south
and north, respectively, in both Mehri (rawram and nagd) and Śḥehri (ramnam
and fagir). This is because the sea lies to the south and the desert lies to the north
(beyond the mountains). In other parts of the coastal plain, however, the word for
‘mountains’ (śḥɛr) is used to indicate north instead. Another common way to describe south and north is to refer to the direction in which the water flows, with
the result that the same word that means ‘south’ on the sea-side of the mountains can be used to indicate ‘north’ on the desert-side. However, all these rather
complex sets of terms are being rapidly replaced, particularly in the speech of
the younger generations and among urban populations, with the Arabic words
for south and north (ǧanūb and šimāl respectively).
3.4.3 Colour terms
The MSAL lexically encode four basic colour terms: white, black, red and green
(Bulakh 2017: 261–262). For example, in Śḥehri one can find lūn for ‘white’, ḥɔr
for ‘black’, ʕɔfər for ‘red’ (and warm colours in general, including brown) and
śəẓ́rɔr for ‘green’ (and everything from green to blue). A fifth colour term, ṣɔfrɔr
‘yellow’ (Mehri ṣāfər), is most probably an adapted borrowing from Arabic already present at the common MSAL level (Bulakh 2017: 271).
A preliminary field inquiry on the subject was conducted by Gasparini in 2017,
with 6 young speakers from the city of Salalah and its immediate surroundings,
all between 20 and 35 years old and all bilingual in Śḥehri and Arabic. The results of the tests showed a remarkable degree of idiolectal variation in the colour
labeling systems employed by the informants, with different levels of interference from Arabic. Remarkably, when asked to label colours in Śḥehri from a
printed basic colour wheel, which was shown to them during interviews, all the
speakers used the Arabic word for ‘blue’, azraq, which seems to have replaced
śəẓ́rɔr (traditionally used for both blue and green, but now confined to the latter).
Two speakers also used aḫḍar for ‘green’, claiming that they could not recall the
Śḥehri term. In addition, only one speaker used ʕɔfər for ‘brown’, Arabic bunnī
being preferred by the other interviewees. The three basic colours ‘white’, ‘black’
363
Simone Bettega & Fabio Gasparini
and ‘red’, however, were regularly referred to using the Śḥehri forms by all speakers. Summing this up, it would seem that the Śḥehri colour system (at least in urban environments, but see below) is undergoing a radical process of restructuring.
The three typologically fundamental colour terms are retained in most contexts,
and a distinction between blue and green is being introduced through reduction
of the original semantic spectrum of śəẓ́rɔr, adoption of the Arabic word for blue,
and subsequent replacement of śəẓ́rɔr with aḫḍar (which indicates only green
in Arabic). Further distinctions are either being replaced with the corresponding
Arabic terms, or introduced if not part of the original semantic inventory of the
language.
On this matter, Watson & Al-Mahri (2017: 90) argue that colour terms (together
with numbers) are often among the first lexical items to be lost in contexts of
linguistic endangerment, and that this is precisely the case with the MSAL. They
write that even children in rural communities are now employing Arabic terms
to refer to the different breeds of cattle (which traditionally used to be referred to
by use of the three basic colour terms ‘white’, ‘black’ and ‘red’). This is probably a
result of the fact that even in villages younger generations are no longer involved
in cattle herding. Examples include aḥmar ‘bay’ in place of Mehri ōfar or Śḥehri
ʕofer, aswad ‘black’ in place of Mehri ḥōwar, and abyað̣̣ ‘white’ in place of Mehri
ūbōn.
3.4.4 Other word classes
Watson & Al-Mahri (2017: 90) note that, since the introduction of a public school
system in Arabic in the 1970s, a number of common lexical items and expressions in Mehri and Śḥehri have been replaced by the corresponding Arabic ones.
Lonnet (2011) also remarks that borrowings from Arabic are particularly common among particles and function words, Examples include nafs aš-šī ‘the same
thing’ for Śḥehri gens, Mehri gans; lākin ‘but’ in place of Śḥehri duʰn and min
duʰn, Mehri lahinnah; yaʕnī ‘that is to say’ and ʕabārah in place of Śḥehri yaḫīn,
Mehri (y)aḫah; tamām ‘fine’ in place of Śḥehri ḥays̃ōf and Mehri hīs taww ~
histaww; Mehri and Ḥarsusi vocative yā ‘oh’ in place of MSAL ʔā-; Śḥehri bdan,
Mehri ʔabdan ‘never, not at all’, against Mehri and Ḥarsusi bəhawʔ, Śḥehri bhoʔ.
Consider also the case of Arabic bəss ‘only’, already mentioned in §3.2.1. In Mehri
as in Baṭḥari, this particle appears now to be interchangeable with its equivalent
ār, as example (7) shows:
(7)
364
Mehri (Sima 2009: 328, cited in Watson 2012: 371; transcription adapted)
bass ta-ṭʕam-h
ḳād aḫah ār ṭʕām ð-maḥḥ
only 2sg.m-taste.impf-3sg.m int fine only taste of-clarified_butter
‘Just taste it, like it is just the taste of clarified butter.’
16 Modern South Arabian languages
As is predictable, also in this field Baṭḥari is the language most affected by Arabic: besides those already cited, we might add the expressions zēn ‘well’, (a)barr
‘outside’ (also in Mehri, as opposed to Soqoṭri ter), ḫalaṣ ‘and this is it’ (used to
end a narrative). Finally, Watson (2012: 3) remarks how “Mahriyōt also exhibits
structures unattested in Mehreyyet such as ‘What X!’ phrases reminiscent of
Arabic, e.g. maṭwalk ‘How tall you (sg.m) are!’.
4 Conclusions
Throughout this chapter we have repeatedly pointed out how research on the
MSAL, and in particular on the effects that contact with Arabic has had on their
evolution, is still far from reaching its mature stage. Much remains to be done,
in particular, in terms of sheer documentation, especially in the case of the most
endangered varieties (Hobyōt, Ḥarsūsi, Baṭḥari). In addition to this, and although
Watson’s (2012) work has greatly contributed to expanding our knowledge in
this area, MSAL syntax remains a strongly neglected field of inquiry. Finally,
our knowledge of the history of the MSAL prior to the twentieth century (and
therefore the history of their contact with Arabic) is extremely poor.
It must also be remarked that, although the most widely spoken among the
MSAL are undoubtedly better documented, very little is known about the effects
that urbanisation has had on their speech communities in recent years. In particular, anecdotal evidence suggests that the varieties of Śḥehri and Soqoṭri spoken
in Salalah and Ḥadibo are undergoing rapid change under the influence of Arabic (both the standard variety of the language, which children learn in school,
and the dialects). Fieldwork conducted in the two abovementioned urban centres
could provide extremely valuable information concerning the effects of contact
between Arabic and Modern South Arabian.
Despite the far-from-complete state of research in this field, what we currently
know is sufficient to say that contact has had a strong impact on the MSAL.
Though this is more evident in the area of lexicon, where borrowings are legion,
phonetics and phonology have also been affected (though to a different extent
from one language to another). Morphology and syntax, on the contrary, appear
to be more resistant to contact-induced change, though in the most endangered
varieties one can notice a partial disruption of the original pronominal system
and verbal paradigm, and though the seemingly high degree of resistance to external influence shown by MSAL syntax could actually be due to our limited
knowledge of the subject.
One last note is due concerning another heavily neglected topic, namely the
effects that contact with the MSAL have had on spoken Arabic. Though we have
365
Simone Bettega & Fabio Gasparini
not addressed the question in the course of this paper, evidence drawn from the
existing literature (Simeone-Senelle 2002) suggests that this influence, too, is not
completely absent, and that further research in this direction could produce interesting results.
Further reading
) Morris (2017) can be thought of as a general introduction to contact between
MSAL and Arabic.
) Watson & Al-Mahri (2017) offer an intriguing account of how language change,
contact with Arabic and changes to the traditional environment are all deeply
interrelated.
) Lonnet (2011) – although limited in scope and extension due to its nature as
an encyclopedic entry – offers interesting highlights on the effects of contact
on the MSAL.
Abbreviations
1, 2, 3
acc
BCE
cs
f
imp
impf
int
m
1st, 2nd, 3rd person
accusative
before Common Era
construct state
feminine
imperative
imperfect (prefix
conjugation)
intensifier
masculine
neg
MSAL
ptcp
pro
prf
pl
sg
voc
negative
Modern South Arabian
languages
participle
pronoun
perfect (suffix conjugation)
plural
singular
vocative
References
Beeston, Alfred F. L. 1981. Languages of pre-Islamic Arabia. Arabica 28(2/3). 178–
186.
Bulakh, Maria. 2017. Color terms of the Modern South Arabian languages: A diachronic approach. In Leonid Kogan, Natalia Koslova, Sergey Loesov & Sergey
Tishchenko (eds.), Babel und Bibel 1: Annual of Ancient Near Eastern, Old
Testament and Semitic studies, 269–282. Winona Lake, IN: Eisenbrauns.
Davey, Richard J. 2016. Coastal Dhofari Arabic: A sketch grammar. Leiden: Brill.
366
16 Modern South Arabian languages
Dufour, Julien. 2016. Recherches sur le verbe sudarabique moderne. Paris: Ecole
pratique des hautes études. (Habilitation dissertation).
Edzard, Lutz. 2017. On the role of Modern South Arabian within a comparative
Semitic lexicographical project. Brill’s Journal of Afroasiatic Languages and
Linguistics 9(1–2). 119–138.
Gasparini, Fabio. 2017. Phonetics of emphatics in Bathari. In Simone Bettega
& Fabio Gasparini (eds.), Linguistic studies in the Arabian Gulf, 69–85. Turin:
Dipartimento di Lingue e Letterature Straniere e Culture Moderne, Università
di Torino.
Gasparini, Fabio. 2018. The Baṭḥari language of Oman: Towards a descriptive
grammar. Naples: Università degli studi di Napoli “L’Orientale”. (Doctoral
dissertation).
Hoyland, Robert G. 2001. Arabia and the Arabs: From the Bronze Age to the coming
of Islam. London: Routledge.
Johnstone, Thomas M. 1975. The Modern South Arabian languages. Afro-Asiatic
Linguistics 1. 93–121.
Kogan, Leonid. 2015. Genealogical classification of Semitic: The lexical isoglosses.
Berlin: Walter de Gruyter.
Ladefoged, Peter & Ian Maddieson. 1996. The sounds of the world’s languages.
Oxford: Blackwell.
Lonnet, Antoine. 2011. Modern South Arabian. In Lutz Edzard & Rudolf de Jong
(eds.), Encyclopedia of Arabic language and linguistics, online edn. Leiden: Brill.
Lucas, Christopher. 2015. Contact-induced language change. In Claire Bowern &
Bethwyn Evans (eds.), The Routledge handbook of historical linguistics, 519–536.
London: Routledge.
Morris, Miranda J. 2017. Some thoughts on studying the endangered Modern
South Arabian languages. Brill’s Journal of Afroasiatic Languages and
Linguistics 9(1–2). 9–32.
Naumkin, Vitaly & Victor Porkhomovsky. 1981. Očerki po etnolingvistike sokotry.
Moscow: Nauka.
Persson, Maria. 2009. Circumstantial qualifiers in Gulf Arabic dialects. In Bo
Isaksson, Heléne Kammensjö & Maria Persson (eds.), Circumstantial qualifiers
in Semitic: The case of Arabic and Hebrew, 206–286. Wiesbaden: Harrassowitz.
Porkhomovsky, Victor. 1997. Modern South Arabian languages from a Semitic
and Hamito-Semitic perspective. In Janet Starkey (ed.), Proceedings of the
Seminar for Arabian Studies vol. 27, 219–223. Oxford: Archaeopress.
Retsö, Jan. 2000. Kaškaša, t-passives and the ancient dialects in Arabia. Oriente
Moderno 19(1). 111–118.
367
Simone Bettega & Fabio Gasparini
Ridouane, Rachid & Cédric Gendrot. 2017. On ejective fricatives in Omani Mehri.
Brill’s Journal of Afroasiatic Languages and Linguistics 9(1–2). 139–159.
Robin, Christian-Julien. 1991. La pénétration des arabes nomades au Yémen.
Revue des mondes musulmans et de la Méditerranée 61(1). 71–88.
Rubin, Aaron D. 2014. The Jibbali (Shaḥri) language of Oman: Grammar and texts.
Leiden: Brill.
Rubin, Aaron D. 2017. A re-analysis of Mehri grammar. Paper presented at the
Cinquièmes journées d’études sur les langues sudarabiques modernes, September 2017, Paris.
Sima, Alexander. 2009. Mehri-Texte aus der jemenitischen Sharqīyah: Transkribiert
unter Mitwirkung von Askari Hugayran Saad. Wiesbaden: Harrassowitz.
Simeone-Senelle, Marie-Claude. 2002. L’arabe parlé dans le Mahra (Yémen). In
Otto Jastrow, Werner Arnold & Hartmut Bobzin (eds.), “Sprich doch mit deinen
Knechten aramäisch, wir verstehen es!”: 60 Beiträge zur Semitistik: Festschrift für
Otto Jastrow zum 60. Geburtstag, 669–684. Wiesbaden: Harrassowitz.
Simeone-Senelle, Marie-Claude. 2011. Modern South Arabian. In Stefan
Weninger, Geoffrey Khan, Michael P. Streck & Janet C. E. Watson (eds.), The
Semitic languages: An international handbook, 1073–1113. Berlin: De Gruyter
Mouton.
Van Coetsem, Frans. 1988. Loan phonology and the two transfer types in language
contact. Dordrecht: Foris.
Van Coetsem, Frans. 2000. A general and unified theory of the transmission process
in language contact. Heidelberg: Winter.
Versteegh, Kees. 2014. The Arabic language. 2nd edn. Edinburgh: Edinburgh
University Press.
Watson, Janet C. E. 2011. South Arabian and Yemeni dialects. Salford Working
Papers in Linguistics and Applied Linguistics 1. 27–40.
Watson, Janet C. E. 2012. The structure of Mehri. Wiesbaden: Harrassowitz.
Watson, Janet C. E. 2018. South Arabian and Arabic dialects. In Clive Holes (ed.),
Arabic historical dialectology: Linguistic and sociolinguistic approaches, 316–334.
Oxford: Oxford University Press.
Watson, Janet C. E. & Abdullah Musallam Al-Mahri. 2017. Language and nature
in Dhofar. In Simone Bettega & Fabio Gasparini (eds.), Linguistic studies in the
Arabian Gulf, 87–103. Turin: Dipartimento di Lingue e Letterature Straniere e
Culture Moderne, Università di Torino.
Watson, Janet C. E., Abdullah Musallam Al-Mahri & Ali Al-Mahri. Forthcoming.
Taghamk afyat: A course in Mehri. Wiesbaden: Harrasowitz.
368
16 Modern South Arabian languages
Watson, Janet C. E. & Alex Bellem. 2010. A detective story: Emphatics in Mehri.
In Janet Starkey (ed.), Proceedings of the Seminar for Arabian Studies vol. 40,
345–355. Oxford: Archaeopress.
Watson, Janet C. E. & Alex Bellem. 2011. Glottalisation and neutralisation in
Yemeni Arabic and Mehri: An acoustic study. In Majeed Hassan Zeki &
Barry Heselwood (eds.), Instrumental studies in Arabic phonetics, 235–256.
Amsterdam: John Benjamins.
Watson, Janet C. E. & Barry Heselwood. 2016. Phonation and glottal states in
Modern South Arabian and San’ani Arabic. In Youssef A. Haddad & Eric
Potsdam (eds.), Perspectives on Arabic linguistics XXVIII: Papers from the Annual
Symposium on Arabic Linguistics, Gainesville, Florida, 2014, 3–36. Amsterdam:
John Benjamins.
Watson, Janet C. E., Bonnie Glover Stalls, Khlaid Al-Razihi & Shelagh Weir. 2006.
The language of Jabal Rāziḥ: Arabic or something else? In Rob Carter & St
John Simpson (eds.), Proceedings of the Seminar for Arabian Studies vol. 36, 35–
41. Oxford: Archaeopress.
Wellsted, James R. 1837. Narrative of a journey into the interior of Omán, in 1835.
Journal of the Royal Geographical Society of London 7. 102–113.
Winford, Donald. 2005. Contact-induced change: Classification and processes.
Diachronica 22(2). 373–427.
Zammit, Martin R. 2011. South Arabian loanwords. In Lutz Edzard & Rudolf de
Jong (eds.), Encyclopedia of Arabic language and linguistics, online edn. Leiden:
Brill.
369
Chapter 17
Neo-Aramaic
Eleanor Coghill
Uppsala University
This paper examines the impact of Arabic on the North-Eastern Neo-Aramaic dialects, a diverse group of Semitic language varieties native to a region spanning Iraq,
Turkey, Syria and Iran. While the greatest contact influence comes from varieties
of Kurdish, Arabic has also had considerable influence, both directly and indirectly
via other regional languages. Influence is most apparent in lexicon and phonology,
but also surfaces in morphology and syntax.
1 Current state and historical development
The Aramaic language (Semitic, Afro-Asiatic) has nearly three thousand years
of documented history up to the present day. Once widely used, both as a first
language and as a language of trade and officialdom, since the Arab conquests of
the seventh century it has steadily shrunk in its geographical coverage. Today its
descendants, the Neo-Aramaic dialects, only remain in pockets, especially in remoter regions, and are spoken almost exclusively by religious–ethnic minorities.
Four branches of the language family exist today: due to diversification these cannot be considered a single language. Indeed, the largest branch, North-Eastern
Neo-Aramaic (NENA), which is treated in this chapter, itself consists of many
mutually incomprehensible dialects. Its closest relation is Ṭuroyo/Ṣurayt, which
is spoken by Christians, known as Suryoye, indigenous to the area immediately
west of NENA’s western edge in Turkey. Another member of this branch (Central Neo-Aramaic) was Mlaḥso, but this was nearly wiped out during the First
World War, and its last speaker apparently died in the 1990s.
The NENA dialects are, or were, spoken in a contiguous region stretching
across northeastern Iraq, southeastern Turkey, northeastern Syria and northwestern Iran. The majority ethnicity in this region is the Kurds. NENA’s native
Eleanor Coghill. 2020. Neo-Aramaic. In Christopher Lucas & Stefano Manfredi
(eds.), Arabic and contact-induced change, 371–402. Berlin: Language Science Press.
DOI:10.5281/zenodo.3744533
Eleanor Coghill
speakers are exclusively from Christian and Jewish communities. The Christians
belong to a variety of churches: the Church of the East, the Chaldean Catholic
Church (which split off from the Church of the East when it came into communion with Rome), and (in fewer numbers) the Syriac Orthodox Church and
its uniate counterpart, the Syriac Catholic Church. The Christians’ traditional
religious–ethnic endonym is Surāye and they call their language Sūraθ or Sūrət
(depending on dialectal pronunciation). In other languages, and sometimes in
their own, they identify mainly as Assyrians or Chaldeans.
The Jews are called hudāye or hulāʔe (depending on dialectal pronunciation),
and they call their language lišāna deni/nošan ‘our language’ or hulaula ‘Jewishness’. In Israel, where most now live, they are known as kurdím, reflecting their
geographical origin in the Kurdish region, rather than their ethnic identity.
Historically, the NENA-speaking Christians usually lived in rural mono-ethnic
villages and predominantly practiced agriculture, animal husbandry and crafts.
Jews lived in both villages and towns, alongside other ethnic groups such as
Kurds. They had diverse professions: tradesmen (pedlars, merchants and shopkeepers), craftsmen, peasants and landowners (Brauer & Patai 1993: 205, 212).
The region to which NENA is indigenous was, until, the twentieth century,
highly diverse in terms of ethnicity, religion and language. Some of this diversity remains, but a great deal has been lost, due to the persecutions and ethnic
cleansing that went on during that century and which were not unknown prior
to it. During the First World War, Christian communities in Anatolia, being
viewed as a fifth column in league with Russia, suffered murderous attacks and
deportations. This affected not only Armenians and Greeks, but also the Sūraθspeaking Surāye and Ṭuroyo-speaking Suryoye, as well as the many Arabicspeaking Christian communities in the region (the extirpation of some of these is
documented in Jastrow 1978: 3–17).1 By the 1920s, the Hakkari province of Turkey
had been emptied of its many communities of Surāye: survivors ended up in Iraq
and Iran. Some Sūraθ-speaking villages remained in the neighbouring Şırnak and
Siirt provinces, but in the late twentieth century these too were mostly emptied
of their inhabitants, during the conflict between the Turkish state and the Kurds.
In Iraq too the twentieth century was far from peaceful for the NENA-speaking
communities. After a massacre in the 1930s, a proportion of the survivors of the
genocide moved from Iraq to Syria, where they settled along the Khabur river,
still in their tribal groups. Others remained in Iraq, in some places in their original
1
The relationship between language and ethno-religious identity was and remains complex.
Many Christians belonging to the Syriac churches spoke and continue to speak yet other regional languages, including varieties of Turkish, Armenian and Kurdish.
372
17 Neo-Aramaic
communities, in other places in mixed communities, where a koiné form of Sūraθ
arose. After the founding of Israel, there was a backlash against Jews in Iraq, and
almost all Jews left the country for Israel during the 1950s. In Israel their heritage
and language were for the most part not appreciated and the language was not
passed on to younger generations. Most remaining speakers are now elderly and
some dialects have already died out.
From the 1960s onwards, conflicts between Kurdish groups and the Iraqi state
resulted in the destruction of numerous northern Iraqi villages, including many
Christian ones. Other villages were appropriated by Kurdish tribes. The war in
1990–1991, the international sanctions and the invasion of 2003 and subsequent
instability further affected these communities, as they did all Iraqis, and resulted
in a dramatic shrinking of the Christian community in Iraq. In 2014, when ISIS
captured large swathes of northern Iraq, many Christians and other non-Sunni
minorities had to leave their villages overnight. These villages were later recaptured, but, in the absence of extensive rebuilding and due to fears of a recurrence, many inhabitants have not returned and seek to leave the country. The
outlook is therefore bleak for these communities and for their language.
2 Contact languages
The main contact language for NENA is – and has been for long time – Kurdish
(Iranian, Indo-European), in its many varieties, as Kurds are by far the largest
ethnic group in the region as a whole, excepting Iranian Azerbaijan, where Azeris predominate.2 Kurds have also been politically dominant: during the Ottoman period, Christians and Jews were in the power and under the protection
of local Kurdish rulers, the aghas (see Sinha 2000: 11–12; Brauer & Patai 1993:
223). Most NENA speakers in the Kurdish-speaking areas at this time seem to
have spoken the local Kurdish dialect.3 It is not surprising, therefore, that there
is more influence from Kurdish than from any other language across most if not
all of the NENA dialects, even if its extent varies from dialect to dialect.
2
Small communities of Turkic-speaking Turkmens are also found within northern Iraq. Their
dialects share features with both Anatolian Turkish varieties and Iranian Azeri (Bulut 2007).
3
For such information we rely mainly on statements in grammatical descriptions, where the
researcher asked their informants about this. For instance, Hoberman (1989: 9) states, “All
my informants who grew to adulthood in Kurdistan report that they spoke fluent Kurdish
(Kurmanji)”. Other references for Jews’ competence in Kurdish are: Sabar (1978: 216), Mutzafi
(2004: 5), Khan (2007: 198) and Khan (2009: 11); for the Christians see Sinha (2000: 12–13) and
Khan (2008: 18).
373
Eleanor Coghill
What role, then, has Arabic played? To summarize: there has been longstanding direct contact with small Arabic-speaking communities in what are otherwise Kurdish-speaking regions; there has been indirect contact through loans
transmitted via Kurdish and Azeri varieties; finally, there has been intense contact more recently due to the establishment of states with Arabic as the national
language, as well as various other modern developments. In the remainder of
this section, we will go through these three types of contact in turn.
Although the region is not majority Arabic-speaking, there have been longstanding Arabic-speaking communities in certain parts of it: moreover many of
these were Jewish and Christian, like the NENA-speakers, so one might well
expect more social contacts with them. The Arabic dialects across the region are
overwhelmingly of the qəltu Mesopotamian–Anatolian type (contrasted with the
southern Iraqi/Bedouin gələt type).4
Christian qəltu Arabic speakers could be found in the city of Mosul (alongside qəltu Arabic speakers of other religions) on the edge of the NENA-speaking
Nineveh Plain (also known as the Mosul Plain). They are also present in two
villages on the Nineveh Plain, namely Bəḥzāni and Baḥšiqa. Arabic-speaking
Yazidis5 also live in these villages, as well as (in Baḥšiqa) some Muslim Arabs
(Jastrow 1978: 24). The Christian NENA speakers of the Nineveh Plain, therefore,
had ample opportunity to come into contact with Arabic. To find more Christian
Arabic-speaking communities in or near the NENA region, we have to travel
quite far, to what are now the Turkish provinces of Şırnak, Siirt and Mardin. In
this region there were many Christian qəltu Arabic-speaking communities living in villages and towns until the First World War; fewer afterwards. The settlements with such communities included Āzəḫ (Turkish İdil) and Ǧazīra (Cizre)
in Şırnak province, as well as provincial centres Siirt and Mardin (Jastrow 1978:
1–23). Thus, Christian Arabic speakers were in close proximity to speakers of
NENA dialects in the Bohtan and Cudi regions of Şırnak province, as well as to
speakers of Ṭuroyo/Ṣurayt in Mardin Province.
Jewish qəltu Arabic-speaking communities were also found in both northern
Iraq and southeastern Turkey. In Iraq, Arabic was spoken by the Jews of Mosul,
ʕAqra (Kurdish Akre) and Arbil (Erbil; Kurdish Hawler), as well as of the village
4
The two types of Mesopotamian–Anatolian Arabic dialects are labelled by scholars according
to the shibboleth of the form ‘I said’: qəltu vs. gələt (Blanc 1964: 5–8). qəltu dialects realize *q as
/q/, while gələt dialects (such as Muslim Baghdadi), which are Bedouin or Bedouin-influenced,
realize it as /g/. Qəltu dialects also preserve the 1sg inflection -u on the suffix-conjugation
verb. See Talay (2011) for an overview of Mesopotamian–Anatolian Arabic varieties. Note that
some Bedouin influence may be seen in the Muslim qəltu dialects spoken on the plain south
of Mardin (Jastrow 1978: 30).
5
Elsewhere, Yazidis are Northern Kurdish-speaking.
374
17 Neo-Aramaic
of Ṣəndor, near Duhok (Hoberman 1989: 9). These all left in the 1950s. Further
afield, there were also some Jewish Arabic speakers in Urfa, Diyarbakır, Siverek
and Çermik (Jastrow 1978: 4), who also migrated to Israel. There are known to
have been contacts between NENA-speaking and Arabic-speaking Jews, through
family connections and commerce. Mutzafi (2004: 6) reports such contacts involving the Jewish men of Koy Sanjaq and the Arabic-speaking Jews of Kurdistan.
Sabar (1978: 216–217) relates that the Jews of Zakho would visit relatives who
had moved to Mosul and Baghdad. On the other hand, Hoberman (1989: 9) stated
that the Jews of ʕAmədya knew no more than a few words of Iraqi Arabic.
To sum up, historically, Christian NENA speakers only had direct local contact
with Arabic speakers (of their own faith) in Mosul and the Nineveh Plain in Iraq
and Şırnak province in Turkey. The NENA-speaking Jews, on the other hand, had
Arabic-speaking co-religionists not only in Mosul, but also within Iraqi Kurdistan
itself.
While most NENA dialects show greatest influence from the majority languages of the region – Kurdish and (in Iranian Azerbaijan) Iranian Azeri – these
also played a role in transferring Arabic influence to NENA. Arabic, as the language of Islam, has had a great influence on Kurdish varieties and Azeri, especially in the lexicon, and many originally Arabic words have been transmitted
to NENA via these languages. Sometimes it is difficult to identify the immediate
donor of such words, but phonetics and morphology can help (see §3.1.1).
During the twentieth century, with the founding of the states of Iraq and Syria,
Arabic became the language of the states that most NENA-speakers found themselves in. They came into contact with it through education, officialdom, military
service, radio and trade. Many Christians from the north of Iraq moved south to
the major (Arabic-speaking) cities, Mosul, Baghdad and Basra, where, in some
cases, they shifted to speaking Arabic, while keeping in close contact with relatives back in the north. By the end of the twentieth century most NENA speakers in Iraq and Syria would have been at ease in Arabic. Naturally these later developments did not affect speakers in Turkey and Iran, who, instead, developed
greater competence in Turkish and Persian, respectively. Jewish speakers from
Iraq, who had left the region by the end of the 1950s, would have had less exposure to Arabic through these means.
It should be mentioned that there has also been influence from European languages, namely from French (via the influence of the Catholic Church among the
Chaldean Catholic communities) and from English (dating to the British Mandate
period, as well as the period of globalization from the late twentieth century),
though some lexical borrowings from these languages may have been mediated
by Arabic.
375
Eleanor Coghill
3 Contact-induced changes in North-Eastern
Neo-Aramaic
Contact influence on NENA6 seems to have arisen mainly through long-term biand multi-lingualism, rather than language shift. Indeed, if any shift has taken
place, it is more likely to have involved NENA speakers who converted to Islam
and shifted to Kurdish.7 Furthermore, much of Iraq was in earlier times Aramaicspeaking, so it can be assumed that over the centuries a shift took place from
Aramaic to Arabic. Some Aramaic substrate features can indeed be seen in Iraqi
Arabic dialects, such as a kind of differential object marking (Coghill 2014: 360–
361).
Using Van Coetsem’s (1988; 2000) distinctions between changes due to borrowing (by agents dominant in the recipient language) and imposition (by agents
dominant in the source language), the contact influences from Arabic attested in
NENA are clearly of the first kind, namely borrowing.
Borrowing from Arabic into NENA is of interest particularly as a case of transfer between related and typologically similar languages, as both are Semitic. Like
Arabic and other Semitic languages, NENA has in its verbal morphology, and to
a lesser extent in its nominal morphology, a non-concatenative root-and-pattern
system, complemented by affixes. Thus, with the triradical root √šql, we get such
forms as k-šāqəl ‘he takes’, šqəl-lə ‘he took’, šqāla ‘taking’, šaqāla ‘taker’, šqila
‘taken’, and so on.
6
Sources for the main contact languages, if not indicated, are as follows: Iraqi Arabic (specifically Muslim Baghdadi): Woodhead & Beene (1967); Northern Kurdish (i.e. Kurmanji/Bahdini):
Chyet (2003). Although Muslim Baghdadi Arabic is not the dialect in closest contact with
NENA, as a Mesopotamian dialect it shares much lexicon with more northerly varieties (which
do not have a dictionary). The transcription of Northern Kurdish words is based on the conventional orthography, as given in Chyet (2003: xxxix–xl): an IPA transcription is also given.
The source for the Christian Alqosh and Christian Telkepe data is the author’s own fieldwork.
Other sources are referenced in the text. The author’s own NENA data is transcribed in IPA
except as follows: č [ʧ], j [ʤ] (equivalent to Arabic ǧ), y [j], ḥ [ħ], x between [x] and [χ],
and ġ between [ɣ] and [ʁ]. Apart from ḥ, consonants with a dot under are the emphatic (velarized/pharyngealized) versions of the undotted consonant; for instance, the symbol ð̣ represents
[ðˤ]. Some dialects have emphasis extending across whole words: such words are conventionally indicated with a superscript cross, e.g. + sadra (equivalent to ṣạḍṛạ). The schwa symbol ə is
used to transcribe a NENA vowel that is, in non-emphatic contexts, typically pronounced as [ɪ].
Phonemically contrastive length in vowels is indicated with a macron, e.g. ā [aː]. The vowels
/i/, /e/ and /o/ are usually realized long: [iː], [eː] and [oː]. NENA words from other sources have
had their transcription adjusted in some cases to bring them closer to this system: the original
transcription may be checked in the referenced sources.
7
It often happened that Christian girls were (occasionally by arrangement, but often unwillingly) kidnapped by Kurds for the purpose of marriage. Any children would have been considered Kurds.
376
17 Neo-Aramaic
Arabic influence in NENA is considerable in the realm of the lexicon, but this
has very often occurred via other contact languages, rather than directly. (All
the contact languages show great influence from Arabic, at least in the lexicon).
Direct lexical borrowing or morphological and structural borrowing from Arabic are less common: they are however well attested in the Christian dialects of
the Nineveh Plain, as well as some Jewish dialects of the Lišāna Deni branch in
northern Iraq, including the dialects of Zakho, Nerwa and ʕAmədya (Kurdish
Amêdî, Arabic al-ʕAmādiyya).
It is difficult to establish with any certainty which contact influences entered
the dialects at which time. The earliest Christian and Jewish NENA texts (from
the 16th and 17th centuries)8 already show considerable contact influence from
Kurdish and Arabic. The extent of Arabic influence in the early Jewish Lišāna
Deni texts (Sabar 1984) is quite surprising. The towns in which these texts originate lie deep in Kurdistan, relatively far from the Arabic speaking part of Iraq.
As we have seen in §2, however, Jews in Kurdistan had contacts with Arabicspeaking co-religionists. Some contact influence in the NENA dialects is clearly
of recent date, such as loanwords from English, which probably date to the twentieth century. The prospective construction of the Christian Nineveh Plain dialects, which appears to be a structural borrowing from vernacular Arabic (see
§3.4), seems to have developed only in the last hundred years or so (Coghill 2010:
375).
By the end of the twentieth century, Arabic was having an immense influence
on the speech of Christian Aramaic-speaking communities living in northern
Iraq, expecially those close to Mosul, such as the town of Qaraqosh. Khan (2002:
9) found that most people from Qaraqosh introduced Arabic words and phrases
into their Neo-Aramaic without adaptation. Khan attributes this to the policy
of Arabicization in Iraq, which meant that schoolchildren were only educated
in Arabic. He found significantly greater influence from Arabic in the younger
generation’s speech. In Christian Qaraqosh, as in the neighbouring dialects of
Christian Alqosh and Christian Telkepe (author’s fieldwork), a large number of
Arabic loanwords have recently been absorbed into the lexicon. Nevertheless,
as Khan remarks, “the proportion of Arabic loans that have penetrated the core
vocabulary of the dialect and replaced existing Aramaic words are relatively few.”
This may, however, not be the case with speakers who have grown up in Arabmajority cities such as Baghdad. In my admittedly limited experience with such
8
The Jewish manuscripts date to the 17th century, but the texts may have been composed earlier
(Sabar 1976: xxix, xliii–xlvi). The Christian manuscripts date to the 18th century but the composition of the texts can be dated to the 16th and 17th centuries (Mengozzi 2002: 16).
377
Eleanor Coghill
speakers, they use a noticeably greater proportion of Arabic loanwords, even
sometimes for basic vocabulary, e.g. Iraqi Arabic ð̣ēʕa for māθɒ ‘village’ (heard
from a Christian Telkepe speaker who grew up in Baghdad before settling in the
US).
3.1 Lexicon
3.1.1 Introduction
All NENA dialects have adopted a large number of loanwords. While Kurdish predominates among these, Arabic loanwords are also common, especially among
the Christian dialects of the Nineveh Plain and the Jewish Lišāna Deni dialects.
Khan (2002: 516) makes a useful distinction for Christian Qaraqosh between
“(i) loan-words that do not have any existing Aramaic equivalent and (ii) those
for which a native Aramaic substitute is still available in the dialect.”9 These
two types seem to reflect two layers of borrowing, an earlier one and a recent
one, which, in many cases, is akin to code-switching. Most Kurdish loans belong
to the first type, while Arabic loans are most common in the second, though
earlier loans do exist. Borrowed Arabic nouns of the second type show little or no
adaptation to native morphology, Khan finds. Verbs, however, are always adapted
to NENA verbal morphology. Most are slotted into the existing NENA verbal
derivations (see §3.1.4).
Khan (2002: 516) remarks that speakers of Christian Qaraqosh are generally
aware of the Aramaic alternatives to these Arabic loans and can give them if
asked. It could be, however, that subsequent generations will have had little exposure to the older synonyms.10 Khan notes that some of these older synonyms
are themselves loanwords, in some cases from Arabic, but so integrated and longstanding that many speakers may not be aware of this. Examples include the recent Arabic loan fəkr (< Arabic fikr) and the older loan taxmanta (f. infinitive of
NENA √txmn Q ‘to think’, denominal < Arabic taḫmīn ‘estimation’; see §3.1.4),
both meaning ‘thought’.
Many loanwords are common to several languages of the region, especially
words specific to local culture or to technologies. While the ultimate source can
usually be identified, it can sometimes be hard to determine the immediate donor
of the loan.
9
Note, however, that apparent synonyms are not always identical in meaning. Christian Alqosh
šəbbakiyə (< Ar. šubbāk) is used for a modern glass window, while the inherited lexeme kāwə
is used for the traditional type of window.
10
The fieldwork for the monograph on this dialect was carried out around the year 2000.
378
17 Neo-Aramaic
Nevertheless, there is sometimes evidence that can establish the immediate
donor. This is the case, for example, for Arabic words ending in the feminine suffix tāʔ marbūṭa (Standard Arabic -a(t)). The Arabic morpheme is realized with
the final /t/ in suffixed forms and in the construct (i.e. followed by a possessor).
When borrowed into NENA, the /t/ is not realized in the absolute (isolated) form
of the word, as in Arabic, e.g. Alqosh sāʕa ‘hour’ (Ar. sāʕa). This contrasts with
Kurdish, which has the /t/ in all forms, e.g. N. Kurd. sa‘et [sɑːˈʕæt] ‘hour’. In
some NENA dialects, in certain words, the /t/ appears as -ət- in suffixed forms,
replicating a pattern in (qəltu) Arabic. Sometimes this leads to back-formations
(see §3.3.1). In other items the tāʔ marbūṭa is realized as -at in all contexts, as
it typically is in Kurdish, and this suggests it was borrowed via Kurdish. An example of the latter is Jewish Betanure/Jewish Challa ʕaširat ‘tribe’, pl. ʕaširatte
(Mutzafi 2008: 103; Fassberg 2010: 270). This is borrowed from Northern Kurdish
‘eşîret [ʕæʃiːˈræt], which borrowed it from Ar. ʕašīra(t) ‘tribe’, almost certainly
via Persian and/or Ottoman Turkish. Another example, ʕādat ‘custom’, is given
by Maclean in his grammar of “Vernacular Syriac” (Maclean 1895: 35), where he
states that nouns ending in -at are feminine.11 Fox (2009: 91), writing of Christian Bohtan, also views Arabic loans ending in -at as having been borrowed via
Kurdish. Examples in this dialect are: sahat ‘hour’, hakowat ‘tale’, qəṣṣat ‘story’,
kəflat ‘family’ (< N. Kurd. kuflet [kʊfˈlæt] ~ k’ulfet [kʰʊlˈfæt] ‘wife, family’ < Ar.
kulfa ‘trouble’) and məllat ‘nation’ (< N. Kurd. milet [mɪˈlæt] < Ar. milla). Some
of the same examples (məḷḷat and qəṣṣat) may also be found in Christian ʕUmra:
Hobrack (2000: 108) takes these to have been borrowed via Turkish, but, given
the overwhelming influence of Kurdish in the region, it seems more plausible
that they were borrowed via Kurdish.12
Sometimes there are other indications in the word’s form that it was borrowed
via Kurdish: the common NENA word šūla ‘work’ derives ultimately from Arabic
11
In Maclean’s dictionary (Maclean 1901: 235), he gives ʕādat (orthography adjusted) as the form
in the Christian Urmia dialect and as one of the variants in “Alqosh”, by which he means
the Nineveh Plain dialects (the other variant being ʕāde, which, lacking the final /t/, appears
to be directly borrowed from Arabic). He gives ʕādəta, on the other hand, for his “Ashirat”
dialect group, which was spoken in “central Kurdistan” (today’s Hakkari province of Turkey).
This looks like the back-formations from direct Arabic loans discussed in §3.3.1, which is a
little surprising, as one would not expect much direct contact with Arabic in that region. It is,
however, a large and diverse group of dialects, and he does not specify in which precise dialect
it was attested.
12
The Kurdish forms attested in dictionaries are not always what we would expect as the sources
of these forms, however. Thus we find ḧekyat [ħækjɑːt] ~ ḧikyet [ħɪkˈjæt] ‘story’ and qise [qɪˈsæ]
‘story’ (not qiset). A variant of the latter ending in /t/, however, is found in a nineteenth-century
dictionary cited in Chyet (2003: 490–491).
379
Eleanor Coghill
šuɣl. Northern Kurdish has also borrowed this word, as şuxul [ʃuˈxul] with a
variant şûl [ʃuːl]. It is perhaps the latter which is the immediate origin of the
NENA word.
The gender in NENA can also suggest the immediate source of a loanword. For
instance, qalam ‘pen’ in Arabic has masculine gender, but, loaned into Northern
Kurdish as qelem, it may have feminine or masculine gender (Rizgar 1993: 322;
Chyet 2003: 478). That qalāma ‘pen’ has feminine gender in certain NENA dialects (e.g. Alqosh; Coghill 2004: 199) suggests that it was borrowed via Kurdish,
not directly from Arabic.
It is difficult to date loanwords in a predominantly unwritten language. Nevertheless, we do have written texts in both the Christian Nineveh Plain and the
Jewish Lišāna Deni dialects going back at least four hundred years, and even in
early texts the proportion of lexemes that were borrowed was high. Arabic loans
are conspicuous in both sets of texts. Sabar (1984: 208) found that in a typical
Jewish text from Nerwa, 30% of lexemes are ultimately of Arabic origin (whether
directly or via another language).
Loanwords may be adapted to varying degrees and in varying ways to the
recipient language. §§3.1.2–3.1.5 deal with the ways in which loans in different
word classes may be integrated, as well as the ways in which they retain characteristics of the donor language, focusing on Arabic loans.
3.1.2 Integration of nouns
Most NENA nouns end in the nominal suffix -a (usually, but not exclusively,
masculine nouns) or -ta~-θa (feminine nouns). Older borrowed nouns usually
have one of these endings, e.g. Christian Alqosh ʕamma ‘paternal uncle’ (< Ar.
ʕamm), ʕašāya ‘dinner’ (< Iraqi Ar. ʕaša) ḥadāda ‘blacksmith’ (< Ar. ḥaddād),
ʕāṣərta ‘early evening’ (< Iraqi Ar. ʕaṣir) and maʕwəlta ‘axe (or similar tool)’ (<
Iraqi Ar. maʕwal ‘pickaxe’). Even if they do not, they are adapted to NENA stress
patterns. Thus Ar. ḥayawā́n ‘animal’ is borrowed (possibly via N. Kurd. ḧeywan
[ħɛjˈwɑːn]) as ḥɛwan in Christian Alqosh, which has penultimate stress (Coghill
2004: 81).
More recent loans, on the other hand, may be used without any such modifications, e.g. Christian Alqosh ʕamal ‘thing’ (< Ar. ʕamal ‘work’), xām ‘linen’
(Iraqi Ar. ḫām ‘raw; cotton cloth’), and sāʕa ‘hour’ (f., < Ar. sāʕa f.). They often
occur also in their original Arabic plural forms, e.g. Christian Alqosh fallāḥín
‘farmers’ and ʔaʕdād ‘(large) numbers’.
Many Arabic loanwords come with the Arabic feminine marker tāʔ marbūṭa
(Standard Arabic -a). In qəltu Arabic dialects this usually has two realizations:
380
17 Neo-Aramaic
-a after emphatic or back consonants, otherwise a high vowel such as -e or -i.13
Such loans in NENA usually also have the same distribution, that is -e (or the
dialectal variant -ə), except after an emphatic or back consonant, when it is -a
(Telkepe -ɒ), e.g. Christian Alqosh baṭālə ‘idleness’ and rawð̣a ‘kindergarten’ and
Christian Telkepe ʕādə ‘custom’ and qəṣṣɒ ‘story’ (see also §3.3.1).
Some loans appear to have come from Standard Arabic and have the -a regardless, e.g. Christian Telkepe lahjɒ ‘dialect’ and madrasɒ ‘school’. Christian
Qaraqosh seems to always represent the tāʔ marbūṭa as -a (Khan 2002: 204).
Borrowed nouns are quite commonly given Aramaic derivational suffixes. For
instance, Jewish Azerbaijani amona ‘paternal uncle’ has a borrowed stem, am-,
from Ar. ʕamm ‘paternal uncle’ via Kurdish or Azeri, but an Aramaic derivation,
-ona, originally with diminutive function (Garbell 1965: 165). An example from
the early Lišāna Deni texts is ġaribūθa ‘foreignness’, from Arabic ɣarīb ‘foreign,
strange’ and the NENA abstract ending -ūθa (Sabar 1984: 205).
NENA often adopts the gender of the donor language, where that language
has nominal genders (as in the case of Arabic and Northern Kurdish, which both
have masculine–feminine gender systems). Thus, the following Christian Alqosh
words share the same gender as their Arabic source: ʕašāya ‘dinner’ (m., like Iraqi
Ar. ʕaša) and daʔwa ‘wedding party’ (f., like Arabic daʕwa ‘invitation, party’).
The loanword ʕāṣərta ‘early evening’ is, however, feminine (as indicated by the
NENA feminine ending -ta), while the Arabic source (Iraqi Ar. ʕaṣir) is masculine.
In Northern Kurdish, however, it is feminine (’esir [ʕæˈsɪɾ]), and this may have
influenced the gender, which, in turn, motivated the adding of the feminine suffix.
In Christian Telkepe, some Arabic loanwords of the structure *CaCC have,
when not suffixed, an epenthetic vowel between the final two consonants. This is
absent when a suffix beginning with a vowel is added, i.e. the construct suffix -əd
or a possessive pronominal suffix. This follows the rules in the donor language:
those Arabic dialects which have the epenthetic vowel (including Baghdadi and
some qəltu dialects, such as Mosul) also lose it under similar conditions.14 Examples include ʕaqəl ‘mind’: ʕaql-əd=baxtɒ [mind-cstr=woman] ‘a woman’s
mind’; and ḥarub ‘war’: p-ḥarb-əd=sawāstipūl [in-war-cstr=Sebastopol] ‘in the
Crimean war’. It is interesting to note that the same rule is also found for Arabic
loanwords in Kurdish (Thackston 2006: 5).
Occasionally, loanwords are adapted to the native root-and-pattern templates,
following the selection of a root. This frequently occurs when the root is also bor13
See Jastrow (1979: 40) for the conditioned imāla (raising of a-vowels) in the tāʔ marbūṭa in
the Arabic dialect of Mosul, and Jastrow (1990: 70) for the same in the Jewish Arabic dialect of
ʕAqra and Arbīl.
14
For Baghdadi Arabic, see Erwin (1963: 56–58).
381
Eleanor Coghill
rowed as a verb. Thus we find Christian Qaraqosh ʔəjbona ‘a will, wish’ (Khan
2002: 517), alongside the verb √ʔjb I ‘to please’ (< Ar. √ʕǧb IV), by analogy with
native words on the pattern CəCCona, e.g. yəqðona ‘a burn’ (< √yqð I ‘to burn’).
Sabar (1984: 205) gives further examples from the early Lišāna Deni texts. More
often, however, borrowed nouns are not adapted to native templates, e.g. Alqosh
ḥanafiya ‘tap’ (< Ar. ḥanafiyya), or only coincidentally follow a native noun pattern (Arabic and NENA share many similar patterns), e.g. Alqosh qahwa ‘coffee’
(< Ar. qahwa), which fits into the common Aramaic pattern CaCCa.
NENA dialects all have a variety of plural suffixes, the most common being
perhaps -e (or its dialectal variant -ə). Loanwords, like inherited words, take a
wide variety of native plural suffixes, but certain suffixes may be preferred or
dispreferred for loanwords, in combination with other factors. For instance in
Christian Alqosh feminine loanwords are not attested with the Aramaic plural
suffixes -wāθa and -awāθa, while the loan-plural -at (< Ar. -āt) is almost exclusively found with loanwords (Coghill 2005: 347). Recent Arabic loans in Christian
Nineveh Plain dialects often occur, unadapted, in their Arabic plural form (see
§3.3.1).
3.1.3 Integration of adjectives
Like nouns, loan adjectives may occasionally be adapted to the native root-andpattern templates, after the selection of a root. For instance, Arabic ʔazraq ‘blue’
(√zrq) is borrowed by Christian Alqosh as zroqa ‘blue’, by analogy with certain inherited colour adjectives of the form CCoCa, such as smoqa ‘red’. Another example is Christian Alqosh ʕadola ‘straight’ (cf. Iraqi Ar. ʕadil ‘straight’ and Christian
Qaraqosh which has borrowed it simply as ʕadəl).15 More often the stem of the
loan adjective is borrowed more or less unchanged, as in Christian Alqosh faqira
‘poor’ (Ar. faqīr), coincidentally fitting the inherited adjectival pattern CaCiCa.
Adapted loan adjectives tend to take NENA inflection (e.g. f. -ta~-θa, pl. -ə). Unadapted loan adjectives usually take no inflection at all, e.g. Christian Telkepe
qə́rməzi ‘purple’ (Ar. qirmizī m. ‘crimson’) and ð̣aʕíf ‘thin’ (Iraqi Ar. ð̣aʕīf m.
‘thin, weak’).
Loan-adjectives of a certain group including colours and bodily traits behave
in a special manner in some NENA dialects: they take Aramaic inflection for
masculine and plural, but a special inflection -ə (identical to the plural ending)
for the feminine. This occurs in Christian Qaraqosh particularly with Arabic loan
15
Attested inherited words of the pattern CaCoCa are all in fact nouns in Christian Alqosh, e.g.
ʔalola ‘street’. The pattern CaCūCa might be more expected, being found with several common
adjectives, e.g. xamūṣa ‘sour’.
382
17 Neo-Aramaic
adjectives, e.g. ṭarša ‘deaf’ (f./pl. ṭaršə, < Ar. m. ʔaṭraš, f. ṭaršāʔ ) and zarqa ‘blue’
(f./pl. zarqə, < Ar. m. ʔazraq, f. zarqāʔ ), see Khan (2002: 219). It appears to come
from a dialectal reflex (-ē) of the Arabic -āʔ feminine ending, found especially
with adjectives of these semantic groups.16 In Christian Alqosh it is also found
with loanwords of Northern Kurdish origin, e.g. kačal-a ‘bald’ (f./pl. kačal-ə, from
N. Kurd. k’eçel [kʰæˈʧæl]).
In Arabic and Kurdish, adjectives normally follow the head noun, as in NENA.
There are, however, a few pseudo-adjectival modifiers borrowed from Arabic
which precede the noun in Arabic and are uninflected. These show the same behaviour when borrowed into NENA. One is ʔawwal ‘first’ in Christian Alqosh (a
synonym to the inherited adjective qamāya ‘first’), as in ʔawwal꞊ga ‘the first time’
– compare Arabic ʔawwal marra ‘the first time’. Another is ġer ‘other’ (< Iraqi
Ar. ɣēr), which is attested in Jewish Betanure, e.g. ġer꞊məndi ‘something else’
(Mutzafi 2008: 105) – compare Iraqi Arabic ɣēr yōm ‘another day’. Another loanword, xoš ‘good’, invariably precedes the noun, e.g. Christian Telkepe xoš꞊ʔixālɒ
‘good food’. This seems to originate in Iranian (Persian or Kurdish), but is also
common in Iraqi and Anatolian Arabic dialects (as ḫōš), as well as in Turkic varieties (as hoş [hoʃ] or xoş [xoʃ]). In all these languages it precedes the noun,
regardless of the usual word order.
3.1.4 Integration of verbs
The borrowing of verbs has been identified as potentially more complicated than
the borrowing of other lexemes, due to their tendency to be morphologically complex (Matras 2009: 175). The borrowing of verbs in a Semitic language presents
particular issues, due to the unusual root-and-pattern system. In Semitic languages verb lexemes are composed of a root (typically consisting of three –
occasionally four – consonants or semi-vowels) and a derivation (also known
as “stem”, “form”, “measure”, “binyan” or “theme”). NENA dialects mostly have
three triradical derivations (I, II and III) and at least one quadriradical derivation
(Q). A borrowed verb will usually be integrated into this system. Three main
strategies have been identified for the borrowing of verbs in NENA. One, common also in other Semitic languages (Wohlgemuth 2009: 173–180), is root extraction, whereby from the phonological matter of the source verb a tri- or quadriradical root is selected. This is usually then allocated to a verbal derivation. A
16
Oddly enough, however, the realization as -ē seems to be restricted to Anatolian qəltu Arabic dialects (where it is stressed, e.g. Āzəḫ lālḗ ‘dumb’), and not found in the dialects in Iraq
(Jastrow 1978: 76). Other words ending in *-āʔ have -ē (unstressed) in qəltu Arabic dialects, but
only as cases of imāla (raising of a-vowels) conditioned by a neighbouring high vowel.
383
Eleanor Coghill
second is the borrowing of not only the root but also some of the morphology
of the Arabic derivation: see below and §3.3.2. A third is the light verb strategy,
whereby the loanverb consists of a light verb (with meanings such as ‘become’
or ‘make’) and a (verbal) noun, the latter containing the main semantic content.
The light verb strategy is found in some NENA dialects, but usually with Kurdish or Turkish verbs, which already consist of a light verb plus noun. It is not used
to integrate Arabic loanverbs, although sometimes the noun in the predicate ultimately comes from Arabic.
The root-extraction strategy is well attested across NENA dialects and is particularly common with Arabic loanverbs. This is unsurprising, as these already
have a root, which in many cases can simply be adopted as it is. For instance, Arabic √ɣlb I ‘to win’ (ɣalaba ‘he won’) is borrowed as Christian Telkepe √ġlb I ‘to
win’. Sometimes the root is adapted, to conform to the rules of root formation in
NENA. For instance, ‘geminate’ roots, where the final two radicals are identical
(√C1 C2 C3 , where C2 =C3 ), are rare in NENA, and apparently absent altogether in
derivation I. Just as inherited geminate roots were converted into middle-y roots
(√C1 yC3 ), so too are Arabic geminate roots. Thus, Arabic √sdd I ‘to close, stop up’
is borrowed as Christian Alqosh √syd I ‘to close, seal’ (compare inherited √qyṛ I
‘to be cold’ < √qrr).
Sometimes derivational affixes are adopted as radicals, often replacing a weak
radical. For instance, Arabic derivation VIII verb ittafaqa (√wfq) is borrowed by
Christian Alqosh as √tfq I ‘to meet’, with the VIII derivational infix -t- reanalysed as a radical. Frequently the root is borrowed not from a true verb but from
a (verbal) noun or adjective. Thus, the NENA verb √txmn Q (found, e.g., in Jewish
Betanure and Christian Qaraqosh, and as √txml Q in Alqosh) is borrowed from
the Arabic noun taḫmīn (possibly via Northern Kurdish t’exmîn [tʰæxˈmiːn] ‘supposition, guess’), itself a derivation of Arabic √ḫmn II ‘to guess’ (ḫammana ‘he
guessed’). The /t/ of the NENA root is not found in the Arabic root, but can only
come from the verbal noun. This is an extension of an inherited Semitic strategy
of deriving verbs from nouns. See Sabar (1984; 2002: 52) and Garbell (1965: 166)
for more on the creation of verbal roots from non-Aramaic verbs.
The process of integration does not end with the establishment of a root, however. Every verb lexeme must also have a derivation. Tendencies can also be
identified for this (Coghill 2015). Arabic loanverbs already have a derivation, but
the majority of Arabic derivations have no cognate or functional equivalent in
NENA. Where there is a cognate, there are also some formal and functional similarities, and thus such cases are usually loaned into the cognate derivation. Thus,
for instance, Arabic √ʕdl II (ʕaddala) ‘to put in order’ is borrowed as Christian
Telkepe √ʕdl II ‘to fix, tidy’ (e.g. mʕudəlli ‘I tidied’), Telkepe derivation II being
the cognate of the Arabic derivation of the same number.
384
17 Neo-Aramaic
Verbs in Arabic derivations that have no cognate are sometimes allocated to
derivations that bear some similarity in form or function to the original derivation. For instance, the NENA derivation most closely resembling Arabic derivation III in form is derivation II (the two share the template -CvCvC-, as opposed
to -vCCvC-). Thus Arabic √hğr III (hāğara) ‘to emigrate’ is borrowed as Christian
Telkepe √hjr II ‘to emigrate’ (e.g. mhujera ‘they emigrated’).
Arabic derivations VIII and X may be treated differently: in Christian Iraqi
dialects, in particular those of the Nineveh Plain, the derivational morphology
may itself be borrowed along with the lexeme (see §3.3.2).
3.1.5 Grammatical words and closed classes
NENA has freely borrowed grammatical words such as prepositions, conjunctions and particles of various functions, and some of these are Arabic, though
most are Kurdish. In some cases, the original Arabic items may have been borrowed via Kurdish. In Christian Alqosh we find the preposition ṣob ‘towards,
near’ (< Ar. ṣawba ‘towards’, cf. Iraqi Ar. ṣōb ‘direction’) and baḥás ‘about, concerning’ (< N. Kurd. beḧs [bæħs] ‘discussion (about)’ < Ar. baḥθ). Another example is m-badal ‘instead of’ (< m- ‘from’ + Iraqi Ar. badāl; Coghill 2004: 300). In
Jewish Challa we also find m-badal and, in addition, mābayn ‘between, among’
(< Ar. mā bayn; Fassberg 2010: 149, 151). Even in Jewish Arbel, which generally
shows less Arabic influence, we find ḍidd ‘against’ (< Ar. ḍidd; Khan 1999: 188).
Loan prepositions are not a new phenomenon in NENA, but are already attested in the early Jewish Lišāna Deni texts (Sabar 1984: 208), e.g. ʕann-ɩd ‘about’
(< Ar. ʕan ‘about’), ṣōb ‘beside’ (< Ar. ṣawba). By analogy with certain native
prepositions, some have been extended with the construct suffix -əd, e.g. ʕann-ɩd.
A particle that has been commonly borrowed is bas ‘only; but’ (cf. Iraqi Ar.
bass ‘enough; only; but’). This may have been borrowed via Northern Kurdish
bes [bæs] ‘enough; but’.
Many dialects, including Christian Alqosh and Christian Telkepe, use kabira
to express ‘much’ or ‘very’. This derives from Arabic kabīr ‘big’. In Christian
Qaraqosh (Khan 2002: 284–5) they use another Arabic loan for the same meaning:
ḥel ~ ḥelə (cf. Iraqi Ar. ḥēl ‘with force’).
Other particles commonly borrowed are fa (roughly ‘and so’ in both Arabic
and NENA) and lo ‘or; either’ (Iraqi Ar. lō). The adverb baʕdén ‘then; later’ (< Ar.
baʕdēn) is attested frequently in the Christian dialects of Alqosh, Telkepe and
Qaraqosh, despite the presence of an inherited synonym, baθər꞊dəx [after꞊how]
‘then; later’.
In Christian Alqosh and Christian Qaraqosh, a particle də- is used with imperatives to give the command a sense of urgency or encouragement. This is already
385
Eleanor Coghill
attested in the early Jewish Lišāna Deni texts (Sabar 1976: xl). This appears to
come from Northern Kurdish de [dæ] with the same function. A similar participle (dē-, də-) is found in both qəltu and Baghdadi Arabic (Jastrow 1978: 310–311).
3.2 Phonology
Two types of phonological contact influences in NENA will be considered here:
new phonemes adopted through contact, and allophonic alternations influenced
by contact.
3.2.1 New phonemes
NENA dialects have gained several new phonemes through language contact.
These phonemes have entered the dialects via loanwords that were not fully
adapted to Aramaic phonology. Some new phonemes are restricted to loanwords,
while others have developed also in native words, through processes such as combination (creating affricate phonemes) and assimilation. As might be expected,
Kurdish loanwords are responsible for the majority of the borrowed phonemes,
but Arabic has also played a role, especially in those dialects closest to the Arabicspeaking region, i.e. the Christian dialects of the Nineveh Plain. The examples
given below are from the Christian Alqosh dialect of this group (Coghill 2004:
11–25, with adapted transcription).
Some of the borrowed phonemes in NENA dialects have been introduced by
both Kurdish and Arabic loanwords. These include /j/ [ʤ] and /č/ [ʧ]. The latter
is not found in Standard Arabic, but is found in Mesopotamian dialects of Arabic.
The phoneme /f/ seems to be borrowed predominantly from Arabic, although
this phoneme also exists in Kurdish. Examples of loanwords with these three
phonemes are: ješ ‘army’ (< Iraqi Ar. ǧēš), jullə ‘clothes’ (< N. Kurd. cil [ʤɪl]),
čārək ‘quarter’ (< N. Kurd. čarêk [ʧɑːˈreːk]) √čyk I ‘to pierce’ (< Iraqi Ar. √čkk I),
and faqira ‘poor’ (< Ar. faqīr).
The phoneme /č/ is also found in certain native Aramaic words, as a result of
the combination of /t/ and /š/, e.g. čeri in čeri qamāya ‘October’ (< *tšeri, cognate
with Christian Qaraqosh təšri and CSyr tešri ~ tešrin ‘Tishrin’).
The Arabic phoneme /ð̣/ [ðˤ] is found in many loanwords in Iraqi NENA dialects, e.g. √ḥð̣r III ‘to prepare’ (< Iraqi Ar. √ḥð̣r II). In most Mespotamian dialects of Arabic in contact with NENA, /ḍ/ is rarely found, as it has merged with
/ð̣/. Nevertheless, one loanword in Alqosh and Qaraqosh has the /ḍ/ phoneme,
namely ʔoḍa ‘room’, which originally comes from Turkish oda. While Turkish is
not considered to have emphatic consonants, it does have vowel harmony, and
386
17 Neo-Aramaic
words with back vowels have been interpreted as having emphatic consonants,
when borrowed into qəltu (and other) Arabic dialects (Jastrow 1978: 51–52). Thus
the qəltu dialect of Qarṭmin, in which *ḍ and *ð̣ have merged as /ð̣/, also has
ʔōḍa ‘room’ (Jastrow 1978: 70). NENA ʔoḍa was borrowed from Turkish either
via a local Arabic variety or directly, in which case its speakers must have also
interpreted back-voweled Turkish words as emphatic.17
The pharyngeals /ʕ/ and /ḥ/, which in most inherited Aramaic lexemes have
shifted to /ʔ/ and /x/ respectively, have been reintroduced through loanwords
from both Arabic and the Classical Syriac used in the church. Examples for /ʕ/
are: ʕamma ‘uncle’ (< Ar. ʕamm), √ʕyš I ‘to live’ (< Ar. √ʕyš I), ʕəddāna ‘time’ (CSyr
ʕeddānā). Examples for /ḥ/ are: √jrḥ I ‘to get injured’ (< Ar. √ǧrḥ I ‘to injure’), √ḥð̣r
III ‘to prepare’ (< Iraqi Ar. √ḥð̣r II), mšiḥa ‘Christ’ (< CSyr mšiḥā), and ḥaṭṭāya
‘sinner’ (< CSyr ḥaṭṭāyā). In some Arabic loans, however, /ʕ/ has shifted to /ʔ/,
perhaps indicating that they belong to an earlier stratum, e.g. Christian Alqosh
daʔwa ‘wedding party’ (Ar. daʕwa). Some cases of /ʕ/ and /ḥ/ in Alqosh, as in
other NENA dialects, are original: the shift to /ʔ/ and /x/ respectively has been
blocked in certain phonetic environments, particularly in the neighbourhood of
emphatic consonants or /q/, e.g. raḥūqa ‘far’ (< *raḥḥūqa), see Khan (2002: 40–
41). Furthermore, /ḥ/ has arisen in the third person singular possessive suffixes,
as a shift from original *h. This appears to be a strategy of disambiguating these
suffixes from the phonetically similar nominal endings (see Coghill 2008: 96–97).
The voiced uvular fricative was an allophone of the voiced velar stop /g/ in
earlier Aramaic. In NENA it merged with *ʕ and shifted to a glottal stop /ʔ/. Like
the pharyngeals, it has been reintroduced into NENA through loanwords from
both Arabic and Classical Syriac, e.g. √ġlb I ‘to win, defeat’ (< Ar. √ɣlb I) and
paġra ‘body’ (< CSyr paḡrā). It has also arisen in native words through regular
assimilation of /x/ to a following voiced consonant. In the case of the verb √ġẓd
I ‘to reap’ (< *√xẓd < *√xṣd < *√ḥṣd), the voiced allophone, originally only found
in certain forms, has spread by analogy throughout the paradigm (Coghill 2004:
20).
The cases of /č/, the pharyngeals, and /ġ/ show how new phonemes may arise
through borrowing, while being assisted by internal developments.
17
Northern Kurdish also has this word, but Chyet’s (2003) dictionary only gives variants without emphasis (e.g. ode), although Iraqi Kurdish dialects do often preserve emphasis in Arabic
loanwords (Chyet 2003: viii; see also Öpengin, this volume).
387
Eleanor Coghill
3.2.2 Allophonic sound alterations
Some NENA dialects, such as Christian Alqosh (Coghill 2004: 27), exhibit wordfinal devoicing of consonants, e.g. mjāwəb [mˈdʒæup] ‘answer!’ (cf. mjawobə ‘to
answer’ with [b]) and qapaġ [ˈqɑpɐχ] ‘lid’ (cf. qapaġəd-dəstiθa ‘saucepan lid’, with
[ʁ]). There is also a strong tendency towards word-final devoicing in both qəltu
Arabic (Jastrow 1978: 98) and the Kurdish dialects of Iraq (MacKenzie 1961: 49), so
it seems to be an areal feature (see also Akkuş, this volume, on contact-induced
devoicing in Anatolian Arabic, and Lucas & Čéplö, this volume, on the same
phenomenon in Maltese).
3.3 Morphology
NENA dialects have borrowed a variety of morphemes from regional languages
via lexical loans. As these become more integrated into the language, they may
be found not only in the original loanwords but also with new words, including
inherited lexemes. NENA being a Semitic language, it is possible for morphological borrowings to be a templatic pattern rather than a single phonetic chunk:
indeed, some verbal derivational patterns have been borrowed from Arabic, as
will be shown in §3.3.2.
3.3.1 Nominal inflection
A grammatical suffix that has been borrowed by some Iraqi dialects is the Arabic
feminine sound plural suffix -āt. In Christian Alqosh and Christian Qaraqosh, as
well as the Jewish Lišāna Deni dialects of northern Iraq, it has been integrated
into the native morphology: as these dialects have penultimate stress in nouns,
the suffix itself is not stressed in these dialects as it is in Arabic (Coghill 2004:
272–273; 2005; Khan 2002: 193–194). Accordingly it has also been shortened to
-at, e.g. Christian Alqosh makina ‘machine’, pl. makinat, maḥallə ‘town quarter’,
pl. maḥallat. In Alqosh and Qaraqosh it is only attested with feminine nouns. It is
not, however, restricted to Arabic loans, but has been extended to other foreign
words, e.g. Alqosh pošiya ‘turban’ (N. Kurd. pʼoşî [pʰoːˈʃiː]) pl. pošiyat. In Alqosh
and Qaraqosh it is even found with some native Aramaic words, e.g. Christian
Qaraqosh ʔarnuwa ‘rabbit’, pl. ʔarnuwat ‘rabbits’; ʔilāna ‘tree’, pl. ʔilānat ‘trees’.
In some words, probably borrowed during the more recent and more intense
period of contact with Arabic, the original stress and length of the ending is preserved, e.g. Christian Alqosh holā́t ‘halls’ and Christian Qaraqosh badlā́t ‘suits’
and gadlā́t ‘tresses’ (Khan 2002: 194). (Note, however, that the latter is an Aramaic word). This is always the case in Telkepe, e.g. jəddɒ ‘midwife’, pl. jəddā́t
388
17 Neo-Aramaic
and traktar ‘tractor’, pl. traktarā́t. Note that in Telkepe, as in Arabic, this plural is
sometimes found with masculine nouns, e.g. mez (m.) ‘table’, pl. mezā́t or primuz
(m.) ‘primus stove’, pl. primuzā́t.
Apart from the Christian Nineveh Plain dialects, -at is attested regularly as a
plural in some of the Jewish Lišāna Deni dialects, spoken further to the north. As
mentioned in §2, these Jewish communities would have had contact with spoken
Arabic through connections with their co-religionists.
In the modern Jewish dialect of Zakho, -at is used with the following types of
nouns (Sabar 2002: 44–45): feminine Arabic loans ending in -a or -e (i.e. the dialectal version of the Arabic feminine suffix tāʔ marbūṭa; see §3.1.2), some nouns
of Kurdish origin ending in -e (perhaps by analogy with Arabic loans ending in
-e), and nouns ending in certain borrowed suffixes, namely the diminutive suffix -ka (f. -ke) borrowed from Kurdish, the professional suffix -či borrowed from
Turkish, and the ending -o. It is also one of the two most common plurals for
European loanwords, e.g. + pākētat ‘packets (of cigarettes)’ (Sabar 1990: 57). This
suggests it is particularly associated with loanwords, regardless of origin. In Jewish Duhok (also Lišāna Deni), however, it is attested with a native Aramaic word,
raʔolat ‘brooks’ (Sabar 2002: 45). It seems therefore that the morpheme has been
extended far beyond its original distribution.
The plural -at does not seem to have spread to all Lišāna Deni dialects, however: it is not mentioned in the grammars of Jewish Challa (Fassberg 2010) and
Jewish Betanure (Mutzafi 2008). It has, nevertheless, an early origin: it is found in
the late seventeenth-century manuscripts originating in the towns of ʕAmədya
and Nerwa. I found one example of it in the grammar of the modern ʕAmədya
dialect (Greenblatt 2011: 70), namely maymonke (f.) ‘monkey’, pl. maymonkat,
probably because it has the Kurdish diminutive suffix (see above).
Across the border in Turkey, another Christian dialect has this plural ending,
that is the dialect of ʕUmra (Turkish name Dereköyü), close to the town of Cizre.
In this region of Turkey there are or were several Arabic-speaking communities, including Christian Arabic speakers in Cizre (until the First World War; see
Jastrow 1978: 17), so it is not surprising that there should be influence from Arabic. In this dialect, -at is mostly attested with borrowed feminine nouns ending
in -e, though there are also a couple ending in -a, both masculine and feminine
(Hobrack 2000: 114). The majority have the Kurdish diminutive suffix -ka (f. -ke)
mentioned above in relation to Jewish Zakho.
In the Christian dialects of Iraq, as spoken currently, it is common to use Arabic words with their original plural morphology, probably because almost all
speakers speak Arabic with native or near-native competence and many con-
389
Eleanor Coghill
cepts are more familiar or only available to them in this language.18 Thus, apart
from the -āt plural, we also find the masculine sound plural suffix -in and the
non-concatenative broken plurals, e.g. Christian Alqosh fallāḥ-ín ‘farmers’, and
barāmíl ‘barrels’ (sg. barmíl) (Coghill 2004: 273). We even find such examples in
the late seventeenth-century manuscripts written in Jewish Lišāna Deni dialects,
e.g. ġāfılīn ‘fools’ and ʔarwāḥ ‘spirits’ (Sabar 1984: 205–206).
Many Arabic loanwords come with the Arabic feminine marker tāʔ marbūṭa,
either the qəltu Arabic variants or the Standard Arabic -a (§3.1.2). In some dialects
of the Nineveh Plain, the tāʔ marbūṭa is borrowed along with its connecting
allomorph -ət. In Arabic the /t/ is only realized in construct state (as the head of
a genitive phrase) or before possessive suffixes.
In Christian Qaraqosh the isolated form of such loans ends in -a, like inherited
masculine nouns, although the gender is feminine (as in the source words). When
possessive suffixes are added, however, the /t/ is realized, as in Arabic (Khan
2002: 204–206). Thus Qaraqosh badla ‘suit of clothes’ (cf. Iraqi Arabic badla)
becomes badl-ətt-əḥ [suit-f-3sg.m] ‘his suit of clothes’. The gemination of the
/t/ is not found in the Arabic forms, but can be explained as follows. In Mosul
Arabic, unlike in many Arabic dialects, the tāʔ marbūṭa takes the stress, when
any possessive suffix is added: báṣali ‘onion’, baṣal-ә́t-ak [onion-f-2sg.m] ‘your
onion’ (Jastrow 1983: 105). It is likely that the /ə/ vowel in the NENA morpheme
-ətt- imitates the vowel of the Arabic morpheme. The stress pattern fits well into
NENA, which has penultimate stress. However, in NENA /ə/ is dispreferred in an
open syllable, especially when stressed. The /t/ is probably geminated in order
to close the syllable so as to conform to this preference.19 This mechanism has
parallels elsewhere in NENA.
These same loanwords take the Arabic plural -at discussed above. Even some
Aramaic feminine words in Christian Qaraqosh have acquired both -ətt- and -at,
e.g. ʔarnuwa (f.) ‘rabbit’, ʔarnuwəttəḥ ‘his rabbit’, ʔarnuwat ‘rabbits’. But -ətt- is
also found with some Aramaic feminine words that have native plurals, e.g. bira
(f.) ‘well’, birāθa ‘wells’, birəttəḥ ‘his well’. In exceptional cases -ətt- may also be
used with feminine words with the Aramaic f. ending -ta~-θa, e.g. šwiθa ‘bed’,
šwiyāθa ‘beds’, šwiθəttəḥ ‘his bed’. It seems, therefore, that in Qaraqosh this is
now a morphological borrowing independent of the loanwords it was originally
borrowed with.
18
Younger NENA speakers who have grown up in the Kurdish-controlled region since 1991 may
have less competence in Arabic, however.
19
Khan (2002: 206) gives two other possible derivations: a combination of Arabic f. -ət and Aramaic f. -ta (though the latter is not found on the isolated form) or the NENA independent
genitive particle did-. The explanation above seems to me to be simpler, however.
390
17 Neo-Aramaic
In Christian Telkepe, vernacular Arabic nouns with tāʔ marbūṭa are borrowed
ending in either -ə or -ɒ, matching the two realizations of the tāʔ marbūṭa in
qəltu Arabic (§3.1.2). As in Qaraqosh, these nouns retain their feminine gender
in Telkepe. They also have the -ətt- allomorph before possessive suffixes, e.g.
ṣəḥḥɒ (f.) ‘health’, ṣəḥəttux [ṣəḥ-ətt-ux health-f-2sg.m] ‘your (m.) health’; qubbə
(f.) ‘room’, qubbətte [qubb-ətt-e room-f-3sg.m] ‘his room’. The suffix seems to be
used productively with Arabic words, as and when they are used. One example
in Telkepe is not borrowed from a feminine with tāʔ marbūṭa, namely čāyi (f.)
‘tea’ (cf. Iraqi Ar. čāy (m.)). This word is, however, feminine in Northern Kurdish
(çay [ʧɑːj]), whence it may have been borrowed.
Christian Alqosh seems to have gone a step further, creating back-formations
from the suffixed forms. Thus the unsuffixed forms also have -ətt-, e.g. ṣaḥətta
‘health’, qaṣətta ‘story’ and məllətta ‘religious community’. When the plural suffix (always the feminine plural -yāθa) is added, one /t/ alone is preserved, suggesting that the second is now analysed as part of the feminine singular ending
-ta, while -ət- is analysed as part of the stem: qaṣət-ta ‘story’, qaṣət-yāθa ‘stories’;
məllət-ta ‘community’, məllət-yāθa ‘communities’.
Similar forms are also attested in Jewish Challa (Lišāna Deni), but without
the gemination of the /t/, e.g. məlləta ‘ethnic group’, ʕādəta ‘custom’ (Fassberg
2010: 52). Rather than explaining the /t/ as originating in the Arabic suffixed
stem, as I have done above, Fassberg suggests that the /t/ is present because the
words were borrowed via (Northern) Kurdish, which realizes the tāʔ marbūṭa as
a final /t/ even when the noun is unsuffixed: milet [mɪˈlæt] and ʕadet [ʕɑːˈdæt]
(Chyet 2003: 387). Khan (2002: 206) also suggests this route for Qaraqosh. This
explanation would not explain why the unaffixed forms in Qaraqosh do not end
in /t/, nor why the preceding vowel in all these dialects is /ə/ rather than /a/ (the
nearest phonetic equivalent to Kurdish 〈e〉). In fact, there are some clear loans of
Arabic words via Kurdish which end in -at in the singular unsuffixed form (see
§3.1.1). The Kurdish route would furthermore not explain the close association
in Qaraqosh of this morpheme with words taking an -at plural, which seems
to have been borrowed directly from Arabic. It seems more likely, therefore, that
the Qaraqosh, Telkepe, Alqosh and Challa feminine nouns with suffixed -ət- have
been borrowed directly from Arabic and are influenced by the Arabic suffixed
forms, which have a similar form.
3.3.2 Verbal derivation
The NENA verbal system consists of both synthetic and analytic verb forms. The
synthetic verb forms are formed from two stems, the Present Base and the Past
Base, e.g. Christian Alqosh k-šaql-i [ind-take.pres-3pl] ‘they take’ and šqəl-lɛ
391
Eleanor Coghill
[take.past-3pl] ‘they took’. Analytic forms involve auxiliary verbs or verboids
combined with non-finite verb forms, such as the infinitive or participles, or, less
often, with finite verb forms. Like Arabic, NENA has a verbal system based on the
root-and-pattern system. As also in Arabic, a verb lexeme typically has a triconsonantal root and a verbal derivational class (see §3.1.4). While Standard Arabic
has ten fairly common triradical verbal derivations, NENA dialects typically have
only three or four inherited verbal derivations.
Morphological loans may be found in the verbal system. Christian NENA dialects of the Nineveh Plain and elsewhere have partially borrowed Arabic verbal
derivations along with borrowed verb lexemes. NENA and Arabic have some
cognate verbal derivations and the relationships are relatively transparent. Most
Arabic loanverbs are allocated to a NENA derivation that is formally or functionally similar to the donor derivation (and often cognate). See §3.1.4 for discussion
of this. In the case of Arabic verbal derivations VIII and X, however, this is not
possible, as no NENA derivations have the characteristic affixes -t- and (i)st-.
In some cases, the affix may instead be analysed as a radical (§3.1.4). In others,
loanverbs in these derivations are borrowed with this derivational morphology,
i.e. with the affixes. This has, in effect, created new derivations, the Ct- and Stderivations.
Table 1 gives all hitherto attested examples of verbs in the new derivations
from Christian Telkepe, but additional verbs are attested in Christian Qaraqosh
(Khan 2002: 130).
Table 1: Arabic loanverbs borrowed into the new NENA derivations
NENA verb
Source verb
√ḥrm Ct- ‘to respect’
√xlf Ct- ‘to differ’
√ḥfl Ct- ‘to celebrate’
√ʕml St- ‘to use’
√ġll St- ‘to exploit’
Ar. √ḥrm VIII (iḥtarama)
Ar. √ḫlf VIII (iḫtalafa)
Ar. √ḥfl VIII (iḥtafala)
Ar. √ʕml X (istaʕmala)
Ar. √ɣll X (istaɣalla)
When Arabic verbs in derivations VIII and X are borrowed as they are, their
characteristic consonantal clusters -Ct- and -st- are preserved and not broken up
by an epenthetic vowel, even if this results in a syllabic structure that is dispreferred in the NENA dialect (such as a stressed short vowel in an open syllable),
e.g. k-maḥtarəm [ind-respect.pres.3sg.m] ‘he respects’. This may be in order to
preserve a salient characteristic of the original Arabic forms.
392
17 Neo-Aramaic
The vowel pattern in these derivations is, on the other hand, variable, even
within the speech of one speaker. For instance, in the Present Base of the Stderivation, we find məstaCaCC-, məstaCCəC- and məstaCəCC- (e.g. məstaʕaml-,
məstaʕməl-, məstaʕəml- ‘use’) as variants of one and the same form. What are
the reasons for this variability? Firstly, Arabic derivations VIII and X are morphophonemically more complex than the native Aramaic derivations. The consonant
clusters bring the necessity of epenthetic vowels: this leads to at least one short
vowel in an open syllable, which is disfavoured in Telkepe. Where the epenthetic
vowel is placed is still optional and in flux. Secondly, there is a conflict between
the characteristic vowels of the Iraqi Arabic source and the vowels typical of Aramaic derivations. Sometimes the former may be more influential and sometimes
the latter.
The new Ct- and St- derivations in NENA have not been extended to inherited
roots nor used productively, unlike some Arabic derivations in Western NeoAramaic. See Coghill (2015) for full details of the new derivations found in NENA,
Western Neo-Aramaic and other Neo-Aramaic varieties.
3.4 Syntax and pattern borrowings
A syntactic borrowing attested only in the Christian Nineveh Plain dialects is the
grammaticalization of a prospective auxiliary (and, as a further step, uninflected
particle) on the model of the vernacular Arabic prospective future particle raḥ-,
which is attested in nearby Mosul Arabic (author’s fieldwork), as well as more
widely across the Syrian and Mesopotamian Arabic dialects (Jastrow 1978: 304).
Example (1) shows the Neo-Aramaic construction (with the particle) and example
(2) shows the Arabic construction.20
(1)
Christian Telkepe NENA (author’s fieldwork)
zi-napl-ɒ
prsp-fall.pres-3sg.f
‘She’s going to fall.’
(2)
Christian Mosul Arabic (author’s fieldwork)
ɣāḥ-təqaʕ
š-šaǧaɣa!
prsp-fall.impf.3sg.f def-tree
‘The tree’s going to fall!’
20
All glosses in the present chapter are the author’s own.
393
Eleanor Coghill
In both cases the gram has developed from a verb ‘to go’ in a form with imperfective or imperfective-like functions.21 Such a development is of course extremely common in the world’s languages and does not need a contact explanation. Nevertheless, there is evidence that contact played a role. The construction is only found in NENA dialects close to the Arabic-speaking zone of Iraq,
i.e. near to Mosul. Furthermore, the most mature versions of the gram (formally
and functionally) are found in the villages closest to Mosul. The gram seems to
have developed only in the last 100 years or so, as it is not attested in texts or
mentioned in grammars of those dialects before then. See Coghill (2010; 2012) for
more details.
NENA shares a number of idiomatic expressions with neighbouring languages.
Among these are formulae used regularly in specific contexts, such as telling a
story or expressing thanks, congratulations or condolences. One that is widespread in NENA dialects, as well as several neighbouring languages, is the opening formula to a fictional story, which begins ‘there was (and) there wasn’t’: see
also Chyet (1995: 236–237). It is attested in various dialects of NENA, Ṭuroyo,
Kurdish, Azeri, Persian and Arabic, e.g.:
(3)
21
a. Christian Alqosh NENA (Coghill 2009: 268)
ʔəθwa꞊w laθwa
b. Christian Bohtan NENA (Fox 2009)
ətwa lətwa
c. Akre Kurdish (MacKenzie 1962: 288)
hebo nebo [hæˈboː næˈboː]
d. Iranian Azeri (Garbell 1965: 175)
(bir) vármɨš (bir) jóxmuš
e. Christian Bəḥzāni Arabic (Jastrow 1981: 404)
kān w ma kān22
In the case of the Nineveh Plain dialects, it originates in a verb that originally had perfect
aspect, e.g. zil-ən ‘I have gone’, possibly with the implication of ‘I am on my way’. It had also
acquired a meaning of imminent future ‘I am about to go’, in effect ‘I am in the process of just
leaving’, hence “imperfective-like functions”.
22
This is a variant (along with kān ma kān, attested in Palestinian Arabic) of the well-known
formula kān yā ma kān ‘once upon a time’. While kān w ma kān clearly means ‘there was and
there was not’, kān yā ma kān has been interpreted in different ways both by scholars and
native speakers. Taking yā ma in its meaning of ‘how much’, it can be understood as ‘there
was, how much there was!’ Alternatively, the ma is understood as a negator, as is found in the
formula in the other languages. See Lentin (1995) for a discussion of kān yā ma kān and similar
expressions.
394
17 Neo-Aramaic
When such formulae are shared by multiple regional languages, it is difficult
to say for certain which language NENA borrowed them from. Kurdish is usually the assumed donor, simply because it is the language most in contact with
NENA and which has had the greatest influence at all levels. Given, however,
that many speakers knew other regional languages as well, they may have heard
such expressions in several languages.
Proverbs are another area in which there are shared expressions across the regional languages (Segal 1955; Garbell 1965: 175; Chyet 1995: 234–236). An example
is ‘He who knows, knows. He who doesn’t know, says “a handful of lentils”.’ This
stems from a folktale and means something like ‘looks can be deceiving’ (Chyet
1995: 235–236). It is attested in Kurdish, Iraqi Arabic, and NENA, as illustrated in
(4–5).
(4)
Iraqi Arabic (Chyet 1995: 235)
il-yidrī
yidrī
w-il
ma yidrī
rel-know.impf.3sg.m know.impf.3sg.m and-rel neg know.impf.3sg.m
gað̣bit
ʕadas
handful.cs lentils
‘He who knows knows, he who doesn’t know (says) “a handful of lentils”.’
(5)
Jewish Zakho NENA (Segal 1955: 262, adapted transcription)
aw d-lá
k-īʔe
aw d-k-īʔe
3sg.m rel-ind-know.pres.3sg.m ind-know.pres.3sg.m 3sg.m rel-not
bi-ṭloxe
g-mēnüx
k-īʔe
ind-know.pres.3sg.m ind-look.pres.3sg.m at-lentils
‘He who knows knows, he who doesn’t know looks at a handful of lentils.’
Sabar (1978), who lists proverbs used by the Jews of Zakho, states also that
many proverbs were not translated into NENA, but used in the original language,
whether Kurdish or Arabic.
There are also some areas of structural convergence in the region’s languages,
where the donor language cannot be definitely identified. For instance, all the
languages (NENA, Sorani, Northern Kurdish, Persian, Turkish, Azeri, Iraqi Turkmen and qəltu Arabic) have enclitic copulas, as illustrated in (6–8).
(6) Akre Kurdish (MacKenzie 1961: 175)
ew kî꞊e
[æw ˈkiːæ]
dem who꞊prs.cop.3sg
‘Who is that?’
395
Eleanor Coghill
(7)
Christian Telkepe NENA (author’s fieldwork)
man꞊ilə
who꞊prs.cop.3sg.m
‘Who is he?’
(8)
Jewish Arbel Arabic (Jastrow 1990: 37, 46)
mani꞊we
who꞊3sg.m
‘Who is he?’
Another shared structure is the use of finite subordinate clauses in subjunctive
mood, rather than infinitives, as complements. In earlier Aramaic varieties, such
as Classical Syriac, both were used (Nöldeke 1904: 224–226), but in NENA only
finite verbs are used, as in example (9).
(9)
Christian Telkepe NENA
k-əbə
d-āxəl
ind-want.pres.3sg.m comp-eat.pres.3sg.m
‘He wants to eat.’
Finite verbs in an irrealis mood are also used in such subordinate clauses
in qəltu (and other vernacular) Arabic (e.g. Jastrow 1990: 65), Northern Kurdish (MacKenzie 1961: 208–209), Sorani (MacKenzie 1961: 134–135), Iraqi Turkmen
(Bulut 2007: 175–176), and Iranian Azeri (Fariba Zamani, personal communication). The development in Turkic is attributed to Iranian influence (Bulut 2007:
175–176). This parallels the loss of the infinitive and its replacement by finite verb
forms in the Balkan Sprachbund (see, e.g., Joseph 2009).
The existence of markers in the noun phrase to specify for indefiniteness (and
in many cases specificity, e.g. ‘a certain man’) is widespread in the area, being
found in NENA (xa- ‘one, a (certain)’), Northern Kurdish (-ek [ɛk] < yek ‘one’),
Sorani (-ēk [eːk] < yek), qəltu Arabic (faɣəd < fard ‘individual’), Baghdadi Arabic
(fadd < fard) and Turkish/Azeri (bir ‘one’).
4 Conclusion
Though not the dominant contact language, Arabic has influenced NENA dialects
considerably, especially those in close contact with Arabic-speaking population
centres, namely the Christian Nineveh Plain dialects, the Jewish Lišāna Deni dialects and the Christian dialects in Şırnak province in Turkey.
396
17 Neo-Aramaic
The influence from Arabic is manifested mostly in lexicon, phonology and
morphology, and less in syntax.
Arabic influence has occurred in different phases. Earlier Arabic influence was
mostly indirect, via Kurdish loans, but direct borrowing seems to have occurred
too.
In the twentieth and twenty-first centuries, Arabic influence has increased
dramatically in the dialects spoken in Iraq, due to mass education exclusively
in Arabic, as well as national media, military service, improved transport, and
migration to the Iraqi cities. Most NENA speakers are bilingual and speak Arabic with native competence, and this has affected how they use Arabic words
within their own language. Typically, recent loans are unadapted and close to
code-switching.
As much of the fieldwork on which this description depends was undertaken
in the late twentieth century or first few years of the twenty-first century, in
future research it would be interesting to look at the speech of young people
today and see whether much has changed. It would also be worth comparing
the speech of communities in their ancestral villages with diaspora communities
living in (or who have recently left) Baghdad or Basra.
Further reading
Most work on NENA and language contact has focused on contact with Kurdish.
To my knowledge, only three works are dedicated to contact with Arabic, none
of which is an overview: Sabar’s (1984) study of Arabic influence in the early
texts in Jewish Lišāna Deni; Coghill’s (2010; 2012) research into a prospective
construction found in the Christian Nineveh Plain dialects, which has apparently
grammaticalized under influence from Arabic; and Coghill’s (2015) study of new
verbal derivations borrowed from Arabic into various Neo-Aramaic languages,
including NENA.
Khan’s (2002) grammar of Christian Qaraqosh contains a great deal of information, scattered through the volume, about contact influences from Arabic,
Qaraqosh being one of the dialects most affected by such influence.
Acknowledgements
I would like to thank all the Neo-Aramaic speakers who have generously given
of their time in my fieldwork and the fieldwork of other scholars which is cited
in this chapter. Much of my research on NENA and language contact took place
397
Eleanor Coghill
in the German Research Foundation-funded project, Neo-Aramaic morphosyntax
in its areal-linguistic context. I would also like to thank the editors of this volume
for their helpful comments.
Abbreviations
1, 2, 3
Ar.
comp
cop
cs
CSyr
dem
f/f.
impf
ind
m/m.
1st, 2nd, 3rd person
Arabic
complementizer
copula
construct state
Classical Syriac
demonstrative
feminine
Imperfect (prefix conjugation)
indicative
masculine
N. Kurd.
NENA
neg
past
pl/pl.
pres
prs
prsp
rel
sg
Northern Kurdish
North-Eastern Neo-Aramaic
negator
NENA Past Base
plural
NENA Present Base
present
prospective
relativizer
singular
Symbols
I, II, III etc.
I, II, III, Q
=
꞊
Arabic verbal derivations
NENA verbal derivations
links two words or morphemes in a phrase with a single stress on
the second component (including but not limited to proclitics)
links two words or morphemes in a phrase with a single stress on
the first component (including but not limited to enclitics)
References
Blanc, Haim. 1964. Communal dialects in Baghdad. Cambridge, MA: Harvard
University Press.
Brauer, Erich & Raphael Patai. 1993. The Jews of Kurdistan. Detroit: Wayne State
University Press.
Bulut, Christiane. 2007. Iraqi Turkman. In J. Nicholas Postgate (ed.), Languages
of Iraq, ancient and modern, 159–187. London: British School of Archaeology in
Iraq.
398
17 Neo-Aramaic
Chyet, Michael L. 1995. Neo-Aramaic and Kurdish: An interdisciplinary consideration of their influence on each other. In Shlomo Izre’el & Rina Drory (eds.),
Language and culture in the Near East, 219–252. Leiden: Brill.
Chyet, Michael L. 2003. Kurdish–English dictionary. New Haven, CT: Yale
University Press.
Coghill, Eleanor. 2004. The Neo-Aramaic dialect of Alqosh. Cambridge: University
of Cambridge. (Doctoral dissertation).
Coghill, Eleanor. 2005. The morphology and distribution of noun plurals in the
Neo-Aramaic dialect of Alqosh. In Alessandro Mengozzi (ed.), Studi afroasiatici: XI Incontro italiano di linguistica camitosemitica, 337–348. Milan: FrancoAngeli.
Coghill, Eleanor. 2008. Some notable features in North-Eastern Neo-Aramaic dialects of Iraq. In Geoffrey Khan (ed.), Neo-Aramaic dialect studies: Proceedings of
a workshop on Neo-Aramaic held in Cambridge, 2005, 91–104. Piscataway, NJ:
Gorgias Press.
Coghill, Eleanor. 2009. Four versions of a Neo-Aramaic children’s story. ARAM
Periodical 21. 251–280.
Coghill, Eleanor. 2010. The grammaticalization of prospective aspect in a group
of Neo-Aramaic dialects. Diachronica 27(3). 359–410.
Coghill, Eleanor. 2012. Parallels in the grammaticalisation of Neo-Aramaic ziland Arabic raḥ- and a possible contact scenario. In Domenyk Eades (ed.), Grammaticalization in Semitic, 127–144. Oxford: Oxford University Press.
Coghill, Eleanor. 2014. Differential object marking in Neo-Aramaic. Linguistics
52(2). 335–364.
Coghill, Eleanor. 2015. Borrowing of verbal derivational morphology between
Semitic languages: The case of Arabic verb derivations in Neo-Aramaic.
In Francesco Gardani, Peter Arkadiev & Nino Amiridze (eds.), Borrowed
morphology, 83–108. Berlin: De Gruyter Mouton.
Erwin, Wallace M. 1963. A short reference grammar of Iraqi Arabic. Washington,
DC: Georgetown University Press.
Fassberg, Steven Ellis. 2010. The Jewish Neo-Aramaic dialect of Challa. Leiden:
Brill.
Fox, Samuel Ethan. 2009. The Neo-Aramaic dialect of Bohtan. Piscataway, NJ:
Gorgias Press.
Garbell, Irene. 1965. The impact of Kurdish and Turkish on the Jewish NeoAramaic dialect of Persian Azerbaijan and the adjoining regions. Journal of
the American Oriental Society 85(2). 159–177.
Greenblatt, Jared R. 2011. The Jewish Neo-Aramaic dialect of Amǝdya. Leiden:
Brill.
399
Eleanor Coghill
Hoberman, Robert D. 1989. The syntax and semantics of verb morphology in
modern Aramaic: A Jewish dialect of Iraqi Kurdistan. New Haven, CT: American
Oriental Society.
Hobrack, Sebastian. 2000. Der neuaramäische Dialekt von Umra (Dereköyü): Lautund Formenlehre – Texte – Glossar. Erlangen–Nürnberg: Friedrich-AlexanderUniversität. (Magister dissertation).
Jastrow, Otto. 1978. Die mesopotamisch-arabischen qəltu-Dialekte. Vol. 1: Phonologie und Morphologie. Wiesbaden: Steiner.
Jastrow, Otto. 1979. Zur arabischen Mundart von Mossul. Zeitschrift für Arabische
Linguistik 2. 76–99.
Jastrow, Otto. 1981. Die mesopotamisch-arabischen qəltu-Dialekte. Vol. 2: Volkskundliche Texte in elf Dialekten. Wiesbaden: Steiner.
Jastrow, Otto. 1983. Tikrit Arabic verb morphology in a comparative perspective.
Al-Abhath 31. 99–110.
Jastrow, Otto. 1990. Der arabische Dialekt der Juden von ʿAqra and Arbīl.
Wiesbaden: Harrassowitz.
Joseph, Brian D. 2009. The synchrony and diachrony of the Balkan infinitive:
A study in areal, general and historical linguistics. Cambridge: Cambridge
University Press.
Khan, Geoffrey. 1999. A grammar of Neo-Aramaic: The dialect of the Jews of Arbel.
Leiden: Brill.
Khan, Geoffrey. 2002. The Neo-Aramaic dialect of Qaraqosh. Leiden: Brill.
Khan, Geoffrey. 2007. Grammatical borrowing in North-eastern Neo-Aramaic. In
Yaron Matras & Janet Sakel (eds.), Grammatical borrowing in cross-linguistic
perspective, 197–214. Berlin: Mouton de Gruyter.
Khan, Geoffrey. 2008. The Neo-Aramaic dialect of Barwar. Leiden: Brill.
Khan, Geoffrey. 2009. The Jewish Neo-Aramaic dialect of Sanandaj. Piscataway,
NJ: Gorgias Press.
Lentin, Jérôme. 1995. kān ya ma kān: Sur quelques emplois de ma dans les dialectes arabes du Moyen-Orient. Studia Orientalia Electronica 75. 151–162.
MacKenzie, David N. 1961. Kurdish dialect studies. Vol. 1. Oxford: Oxford
University Press.
MacKenzie, David N. 1962. Kurdish dialect studies. Vol. 2. Oxford: Oxford
University Press.
Maclean, Arthur John. 1895. Grammar of the dialects of vernacular Syriac as spoken
by the eastern Syrians of Kurdistan, north-west Persia, and the Plain of Mosul:
With notices of the vernacular of the Jews of Azerbaijan and of Zakhu near Mosul.
Cambridge: Cambridge University Press.
400
17 Neo-Aramaic
Maclean, Arthur John. 1901. A dictionary of the dialects of vernacular Syriac as
spoken by the Eastern Syrians of Kurdistan, northwest Persia, and the plain of
Moṣul. Oxford: Clarendon press.
Matras, Yaron. 2009. Language contact. Cambridge: Cambridge University Press.
Mengozzi, Alessandro. 2002. “Israel of Alqosh and Joseph of Telkepe: A Story in
a Truthful Language”: Religious poems in vernacular Syriac (North Iraq, 17th
century). Vol. 2: Introduction and translation. Leuven: Peeters.
Mutzafi, Hezy. 2004. The Jewish Neo-Aramaic dialect of Koy Sanjaq (Iraqi
Kurdistan). Wiesbaden: Harrassowitz.
Mutzafi, Hezy. 2008. The Jewish Neo-Aramaic Dialect of Betanure (Province of
Dihok). Wiesbaden: Harrassowitz.
Nöldeke, Theodor. 1904. Compendious Syriac grammar. London: Williams &
Norgate.
Rizgar, Baran. 1993. Dictionary Ferheng Kurdish–English English–Kurdish
(kurmancî). London: M. F. Onen.
Sabar, Yona. 1976. pəšaṭ wayəhî bəšallaḥ: A Neo-Aramaic midrash on Beshallaḥ
(Exodus): Introduction, phonetic transcription, translation, notes, and glossary.
Wiesbaden: Harrasowitz.
Sabar, Yona. 1978. Multilingual proverbs in the Neo-Aramaic speech of the Jews
of Zakho, Iraqi Kurdistan. International Journal of Middle East Studies 9(2). 215–
235.
Sabar, Yona. 1984. The Arabic elements in the Jewish Neo-Aramaic texts of Nerwa
and ʿAmādīya, Iraqi Kurdistan. Journal of the American Oriental Society 104(1).
201–211.
Sabar, Yona. 1990. General European loanwords in the Jewish Neo-Aramaic dialect of Zakho, Iraqi Kurdistan. In Wolfhart Heinrichs (ed.), Studies in NeoAramaic, 53–66. Atlanta, GA: Scholars Press.
Sabar, Yona. 2002. A Jewish Neo-Aramaic dictionary: Dialects of Amidya, Dihok,
Nerwa and Zakho, northwestern Iraq. Wiesbaden: Harrassowitz.
Segal, Judah B. 1955. Neo-Aramaic proverbs of the Jews of Zakho. Journal of Near
Eastern Studies 14(4). 251–270.
Sinha, Jasmin. 2000. Der neuostaramäische Dialekt von Bēṣpən (Provinz Mardin,
Südosttürkei): Eine grammatische Darstellung. Harrassowitz.
Talay, Shabo. 2011. Arabic dialects of Mesopotamia. In Stefan Weninger, Geoffrey
Khan, Michael P. Streck & Janet C. E. Watson (eds.), The Semitic languages: An
international handbook, 909–919. Berlin: De Gruyter Mouton.
Thackston, Wheeler M. 2006. Kurmanji Kurdish: A reference grammar with selected readings. Harvard University.
401
Eleanor Coghill
Van Coetsem, Frans. 1988. Loan phonology and the two transfer types in language
contact. Dordrecht: Foris.
Van Coetsem, Frans. 2000. A general and unified theory of the transmission process
in language contact. Heidelberg: Winter.
Wohlgemuth, Jan. 2009. A typology of verbal borrowings. Berlin: Mouton de
Gruyter.
Woodhead, Daniel R & Wayne Beene. 1967. A dictionary of Iraqi Arabic: Arabic–
English. Washington, DC: Georgetown University Press.
402
Chapter 18
Berber
Lameen Souag
CNRS, LACITO
Arabic has influenced Berber at all levels – not just lexically, but phonologically,
morphologically, and syntactically – to an extent varying from region to region.
Arabic influence is especially prominent in smaller northern and eastern varieties,
but is substantial even in the largest varieties; only in Tuareg has Arabic influence
remained relatively limited. This situation is the result of a long history of largescale asymmetrical bilingualism often accompanied by language shift.
1 Current state and contexts of use
1.1 Introduction
Berber, or Tamazight, is the indigenous language family of northwestern Africa,
distributed discontinuously across an area ranging from western Egypt to the
Atlantic, and from the Mediterranean to the Sahel. Its range has been expanding
in the Sahel within recent times, as Tuareg speakers move southwards, but in
the rest of this area, Berber has been present since before the classical period
(Múrcia Sánchez 2010). Its current discontinuous distribution is largely the result
of language shift to Arabic over the past millennium.
At present, the largest concentrations of Berber speakers are found in the
highlands of Morocco (Tashelhiyt, Tamazight, Tarifiyt) and northeastern Algeria
(Kabyle, Chaoui). Tuareg, in the central Sahara and Sahel, is more diffusely spread
over a large but relatively sparsely populated zone. Across the rest of this vast
area, Berber varieties constitute small islands – in several cases, single towns –
in a sea of Arabic.
This simplistic map, however, necessarily leaves out the effects of mobility
– not limited to the traditional practice of nomadism in the Sahara and transhumance in parts of the Atlas mountains. The rapid urbanisation of North Africa
Lameen Souag. 2020. Berber. In Christopher Lucas & Stefano Manfredi (eds.),
Arabic and contact-induced change, 403–418. Berlin: Language Science Press.
DOI:10.5281/zenodo.3744535
Lameen Souag
over the past century has brought large numbers of Berber speakers into traditionally Arabic-speaking towns, occasionally even changing the town’s dominant language. The conquests of the early colonial period created small Berberspeaking refugee communities in the Levant and Chad, while more recent emigration has led to the emergence of urban Berber communities in western Europe
and even Quebec.
1.2 Sociolinguistic situation of Berber
In North Africa proper, the key context for the maintenance of Berber is the village. Informal norms requiring the use of Berber with one’s relatives and fellow
villagers, or within the village council, encourage its maintenance not only there
but in cities as well, depending on the strength of emigrants’ (often multigenerational) ties to their hometowns. In some areas, such as Igli in Algeria (Mouili
2013), the introduction of mass education in Arabic has disrupted these norms,
encouraging parents to speak to their children in Arabic to improve their educational chances; in others, such as Siwa in Egypt (Serreli 2017), it has had far less
impact. Beyond the village, in wider rural contexts such as markets, communication is either in Berber or in Arabic, depending on the region; where it is in
Arabic, it creates a strong incentive for bilingualism independent of the state’s
influence. For centuries, Berber-speaking villages in largely Arabic-speaking areas have sporadically been shifting to Arabic, as in the Blida region of Algeria
(El Arifi 2014); the opposite is also more rarely attested, as near Tizi-Ouzou in
Algeria (Gautier 1913: 258).
In urban contexts, on the other hand, norms enforcing Berber have no public presence – quite the contrary. There one addresses a stranger in Arabic, or
sometimes French, but rarely in Berber, except perhaps in a few Berber-majority
cities such as Tizi-Ouzou (Tigziri 2008). Even within the family, Arabic takes
on increasing importance; in a study of Kabyle Berbers living in Oran (Algeria),
Ait Habbouche (2013: 79) found that 54% said they mostly spoke Arabic to their
siblings, and 10% even with their grandparents. In the Sahel, Arabic is out of
the picture, but there too family language choice is affected; 13% of the Berber
speakers interviewed by Jolivet (2008: 146) in Niamey (Niger) reported speaking
no Tamasheq at all with their families, using Hausa or, less frequently, Zarma
instead.
Bilingualism is widespread but strongly asymmetrical. Almost all Berber speakers learn dialectal Arabic (as well as Standard Arabic, taught at school), whereas
Arabic speakers almost never learn Berber. There are exceptions: in some contexts, Arabic-speaking women who marry Berber-speaking men need to learn
404
18 Berber
Berber to speak with their in-laws (the author has witnessed several Kabyle examples), while Arabic speakers who settle in a strongly Berber-speaking town –
and their children – sometimes end up learning Berber, as in Siwa (Egypt). Nevertheless, most Arabic speakers place little value on the language, and some openly
denigrate it; in Bechar (Algeria), anyone expressing interest in Berber can expect
frequently to hear the contemptuous saying əš-šəlḥa ma-hu klam wə-d-dhən ma
hu l-idam ‘Shilha (Berber) is no more speech than vegetable oil is animal fat’. To
further complicate the situation, French remains an essential career skill (except
in Libya and Egypt), since it is still the working language of many ministries and
companies; in some middle-class families, it is the main home language spoken
with children.
On paper, Berber (Tamazight) is now an official language of Morocco (since
2011) and Algeria (since 2016), while Tuareg (Tamasheq/Tamajeq) is a recognised
national language of Mali and Niger. In practice, “official language” remains a
misleading term. Official documents are rarely, if ever, provided in Berber, and
there is no generalised right to communicate with the government in Berber.
However, Berber is taught as a school subject in selected Algerian, Moroccan,
and (since 2012 or so) Libyan schools, while some Malian and Nigerien ones even
use it as a medium of education. It is also used in broadcast media, including
some TV and radio channels. Both Morocco and Algeria have established language planning bodies to promote neologisms and encourage publishing, with a
view towards standardisation. The latter poses difficult problems, given that each
country includes major varieties which are not inherently mutually intelligible.
Berber varieties have been written since before the second century BC (Pichler
2007) – although the language of the earliest inscriptions is substantially different from modern Berber and decipherable only to a limited extent – and southern Morocco has left a substantial corpus of pre-colonial manuscripts (van den
Boogert 1997); many other examples could be cited from long before people such
as Mammeri (1976) attempted to make Berber a printed language. Nevertheless,
writing seems to have had very little impact on the development of Berber as yet.
Awareness of the existence of a Berber writing system – Tifinagh – is widespread,
and often a matter of pride. However, most Berber speakers have never studied
Berber, and do not habitually read or write in it in any script – with the increasingly important exception of social media and text messages, typically in Latin or
Arabic script depending on the region. Efforts to create a standard literary Berber
language have not so far been successful enough to exert a unifying influence on
its dispersed varieties. In the North African context, this is often understood as
implying that Berber is not a language at all – “language” (Arabic luɣa) being
popularly understood in the region as “standardised written language”.
405
Lameen Souag
1.3 Demographic situation of Berber
No reliable recent estimate of the number of Berber speakers exists; relevant data
is both scarce and hotly contested. The estimates brought together by Kossmann
(2011: 1; 2013: 29–36) suggest a range of 30–40% for Morocco, 20–30% for Algeria,
8% for Niger, 7% for Mali, about 5% for Libya, and less than 1% for Tunisia, Egypt,
and Mauritania. Selecting the midpoint of each range, and substituting in the
mid-2017 populations of each of these countries (CIA 2017) would yield a total
speaker population of about 25 million, 22 million of them divided almost evenly
between Morocco and Algeria.
2 Contact languages and historical development
2.1 Across North Africa
Berber contact with Arabic began in the seventh century with the Islamic conquests. For several centuries, language shift seems to have been largely confined
to major cities and their immediate surroundings, probably affecting Latin speakers more than Berber speakers. The invasion of the Banū Hilāl and Banū Sulaym
in the mid-eleventh century is generally identified as the key turning point: it
made Arabic a language of pastoralism, rapidly reshaping the linguistic landscape of Libya and southern Tunisia, then over the following centuries slowly
transforming the High Plateau and the northern Sahara in general. This rural expansion further reinforced the role of Arabic as a lingua franca, while the recruitment of Arabic-speaking soldiers from pastoralist tribes encouraged its spread
further west to the Moroccan Gharb.
The resulting linguistic divide between rural groups and towns remained a key
theme of Maghrebi sociolinguistics until the twentieth century. In several cases,
a town spoke a different language than its hinterland; in much of the Sahara,
Berber-speaking oasis towns such as Ouargla or Igli formed linguistic islands in
regions otherwise populated by Arabic speakers, and in the north, towns such as
Bejaia or Cherchell constituted small Arabic-speaking communities surrounded
by a sea of Berber-speaking villages. Even in larger cities such as Algiers or
Marrakech, the dominance of Arabic was counterbalanced by substantial regular
immigration from Berber-speaking regions further afield.
Today all Berber communities are more or less multilingual, usually in Arabic
and often also in French; outside of the most remote areas, monolingual speakers
are quite difficult to find. Even in the nineteenth century, however, monolingual
Berber speakers were considerably more numerous (Kossmann 2013: 41).
406
18 Berber
Alongside the coexistence of colloquial Maghrebi Arabic with Berber, Classical Arabic also had a role to play as the primary language of learning and in particular religious studies. Major Berber-speaking areas such as Kabylie (northern
Algeria) and the Souss (southern Morocco) developed extensive systems of religious education, whose curricula consisted primarily of Arabic books (van den
Boogert 1997; Mechehed 2007). The restriction of Classical Arabic to a limited
range of contexts, and the relatively small proportion of the population pursuing higher education, gave it a comparatively small role in the contact situation;
even in the lexicon, its influence is massively outweighed by that of colloquial
Arabic, and it appears to have had no structural influence at all.
2.2 In Siwa
Examples of contact-induced change in this chapter are often drawn from Siwi,
the Berber language of the oasis of Siwa in western Egypt. Sporadic long-distance
contact with Arabic there presumably began in the seventh or eighth century
with the Islamic conquests, and increased gradually as Cyrenaica and Lower
Egypt became Arabic-speaking and as the trade routes linking Egypt to West
Africa were re-established. During the eleventh century, the Banū Sulaym, speaking a Bedouin Arabic dialect, established themselves throughout Cyrenaica.
In the twelfth century, al-Idrīsī reports Arab settlement within Siwa itself,
alongside the Berber population. Later geographers make no mention of an Arab
community there, suggesting that these early immigrants were integrated into
the Berber majority. Several core Arabic loans in Siwi, such as the negative copula qačči < qaṭṭ šayʔ and the noon prayer luli < al-ʔūlē, are totally absent from surrounding Arabic varieties today; such archaisms are likely to represent founder
effects dating back to this period (Souag 2009).
The available data gives nothing close to an adequate picture of the linguistic
environment of medieval Siwa. We may assume that, throughout these centuries,
most Siwis – or at least the dominant families – would have spoken Berber as
their first language, and more mobile ones – especially traders – would have
learned Arabic (but whose Arabic?) as a second language. Alongside these, however, we must envision a fluctuating population of Arabic-speaking immigrants
and West African slaves learning Berber as a second language. In such a situation, both Berber-dominant and Arabic-dominant speakers should be expected
to play a part in bringing Arabic influences into Siwi.
The oasis was integrated into the Egyptian state by Muhammad Ali in 1820,
but large-scale state intervention in the linguistic environment of the oasis only
took effect in the twentieth century; the first government school was built in
407
Lameen Souag
1928, and television was introduced in the 1980s. An equally important development during this period was the rise of labour migration, taking off in the 1960s
as Siwi landowners recruited Upper Egyptian labourers, and Siwi young men
found jobs in Libya’s booming oil economy. It has then grown further since the
1980s with the rise of tourism and the growth of tertiary education. The effects of
this integration into a national economy include a conspicuous generation gap
in local second-language Arabic: older and less educated men speak a Bedouinlike dialect with *q > g, while younger and more educated ones speak a close
approximation of Cairene Arabic.
3 Contact-induced changes in Berber
3.1 Introduction
As noted above, bilingualism in North Africa has been asymmetrical for many
centuries, with Berbers much more likely to learn Arabic than vice versa. This
suggests the plausible general assumption that the agents of contact-induced
change were typically dominant in the (Berber) recipient language rather than
in Arabic. However, closer examination of individual cases often reveals a less
clear-cut situation; as seen above in §2.1, the history of Siwi suggests that Berberand Arabic-dominant speakers both had a role to play, and post facto analysis
of the language’s structure seems to confirm this assumption. The loss of feminine plural agreement, for example (§3.3 below), can more easily be attributed to
Arabic-dominant speakers adopting Berber than to Berber-dominant speakers.
In the absence of clear documentary evidence, caution is therefore called for in
the application of Van Coetsem’s (1988; 2000) model to Berber.
3.2 Phonology
The influence of Arabic on Berber phonology is conspicuous; in general, every
phoneme used in a given region’s dialectal Arabic is found in nearby Berber varieties. Almost all Northern Berber varieties have adopted from Arabic at least the
pharyngeals /ʕ/ and /ḥ/, a series of voiceless emphatics: /ṣ/, /ḫ/, non-geminate
/q/, and either /ḍ/ or /ṭ/. These phonemes presumably reached Berber through
loanwords from Arabic, but have been extended to inherited vocabulary as well,
through reinterpretation of emphatic spread or through their use in “expressive
formations” (Kossmann 2013: 199), e.g. Kabyle θi-ḥəðmər-θ ‘breast of a small animal’ < iðmar-ən ‘breast’.
408
18 Berber
In Siwi (Souag 2013: 36–39; Souag & van Putten 2016), at least nine phonemes
were clearly introduced from Arabic. The pharyngealised coronals /ṣ/, /ḷ/, /ṛ/ and
/ḍ/ have no regular source in Berber, and occur in inherited vocabulary almost
exclusively as a result of secondary emphasis spread (with the isolated exception
of ḍəs ‘to laugh’). The order of borrowing appears to be ḷ, ṛ > ṣ > ḍ; in a few older
loans, Arabic ṣ is borrowed as ẓ (e.g. ẓəffaṛ ‘to whistle’ < ṣaffar), and in all but
the most recent strata of loans, Arabic ḍ/ð̣ is borrowed as ṭ (e.g. a-ʕṛiṭ ‘broad’ <
ʕarīḍ). The pharyngeals /ḥ/ and /ʕ/ (e.g. ḥəbba ‘a little’ < ḥabba ‘a grain’, ʕammi
‘paternal uncle’ < ʕamm-ī ‘uncle-obl.1sg’) likewise have no regular source in
Berber, although 1sg -ɣ- has become -ʕ - for some speakers (an irregular sound
change specific to this morpheme). ʕ is lost in a number of older loans (e.g. annaš
‘bier’ < an-naʕš), but ḥ is always retained as such rather than being dropped or
adapted (unlike Tuareg, where it is typically adapted to ḫ). This suggests that
Siwi continued to adapt Arabic loans to its phonology by dropping ʕ up to some
stage well after the beginning of significant borrowing from Arabic, but started
accepting Arabic loans with ḥ too early for any adapted to survive, implying an
order of borrowing ḥ > ʕ. Among the glottals, /h/ (e.g. ddhan ‘oil’ < dihān ‘oils’)
appears in inherited vocabulary only in the distal demonstratives, where comparison to Berber languages that do have h suggests that it is excrescent, while
/ʔ/ only rarely appears even in recent loanwords (e.g. ʔəǧǧəṛ ‘to rent’ < ʔaǧǧar).
The mid vowel /o/ has been integrated into Siwi phonology as a result of borrowing from Arabic; having been established as a phoneme, however, it went
on to emerge by irregular change from original *u in two inherited words (allon
‘window’, agṛoẓ ‘palm heart’), and from irregular simplification of *aɣu in some
demonstratives (e.g. wok ‘this.sg.m’ < *wa ɣuṛ-ək ‘this.sg.m at-2sg.m’). The interdentals /θ/ and /ð̣/ have a more marginal status, but are used by some speakers
even in morphologically well-integrated loans, e.g. a-θqil or a-tqil ‘heavy’ < θaqīl.
Arabic influence may also be responsible for the treatment of [ʒ] and [dʒ] as free
variants of the same phoneme /ǧ/ (Vycichl 2005), so that e.g. /taǧlaṣt/ ‘spider’ is
variously realised as [tʰæʒlˤɑsˤt] ~ [tʰædʒlˤɑsˤt] (Naumann 2012: 152); other Berber
languages with phonemic ž normally have [dʒ] as a conditioned allophone (e.g.
when geminated) or as a cluster.
Arabic influence has also massively affected the frequency of some phonemes.
/q/ and /ḫ/ were marginal in Siwi before Arabic influence, while *e had nearly
disappeared due to regular sound changes, but all three are now quite frequent.
Conversely, the influx of Arabic loans has helped make labiovelarised phonemes
such as gʷ and qʷ rare.
409
Lameen Souag
3.3 Morphology
Berber offers numerous examples of the borrowing of Arabic words together
with their original Arabic inflectional morphology, a case of what Kossmann
(2010) calls Parallel System Borrowing. This phenomenon is most prominent for
nominal number marking, but sometimes attested in other contexts too.
In Berber, most nouns are consistently preceded by a prefix marking gender
(masculine/feminine), number (singular/plural), and often case/state. Nouns borrowed from Arabic normally either get assigned a Berber prefix, or fill the prefix slot with an invariant reflex of the Arabic definite article: compare Figuig agʕud vs. Siwi lə-gʕud ‘young camel’ (< qaʕūd). The Berber plural marking system
prior to Arabic influence was already rather complex, combining several different
types of affixal marking with internal ablaut strategies; many Arabic loans are integrated into this system, e.g. Kabyle a-bellar ‘crystal’ > pl. i-bellar-en (< billawr),
Siwi a-kəddab ‘liar’ > pl. i-kəddab-ən (< kaððāb). However, in most Berber varieties, Arabic loans have further complicated the system by frequently retaining
their original plurals, e.g. Kabyle l-kaɣeḍ ‘paper’ > le-kwaɣeḍ (< kāɣid), Siwi əlgənfud ‘hedgehog’ > pl. lə-gnafid (< qunfuð). (The difference correlates fairly
well with the choice in the singular between a Berber prefix and an Arabic article, but not perfectly; contrast e.g. Siwi a-fruḫ ‘chick, bastard’ < farḫ, which takes
the Arabic-style plural lə-fraḫ.) Berber has no inherited system of dual marking,
instead using analytic strategies. Nevertheless, for a limited number of measure
words, duals too are borrowed, e.g. Kabyle yum-ayen ‘two days’ < yawm-ayn
(although ‘day’ remains ass!), Siwi s-sən-t ‘year’ > sən-t-en ‘two years’ < san-atayn. Arabic number morphology may sporadically spread to inherited terms as
well, e.g. Kabyle berdayen ‘twice’ < a-brid ‘road, time’, Siwi lə-gʷrazən ‘dogs’ <
a-gʷərzni ‘dog’ (Souag 2013).
Whereas nouns are often borrowed together with their original inflectional
morphology, verbs almost never are. The only attested exception is Ghomara,
a heavily mixed variety of northern Morocco. In Ghomara, many (but not all)
verbs borrowed from Arabic are systematically conjugated in Arabic in otherwise
monolingual utterances, a phenomenon which seems to have remained stable
over at least a century: thus ‘I woke up’ is consistently faq-aḫ, but ‘I fished’ is
equally consistently ṣṣað-iθ (Mourigh 2016: 6, 137, 165). However, the borrowing
of Arabic participles to express progressive aspect is also attested in Zuwara, if
only for the two verbs of motion mašəy ‘going’ (pl. mašy-in) and žay ‘coming’ (pl.
žayy-in), contrasting with inherited fəl ‘go’, asəd ‘come’ (Kossmann 2013: 284–
285).
410
18 Berber
Prepositions are less frequently borrowed; in some cases where this does occur,
however – including Igli mənɣir- ‘except’, Ghomara bin ‘between’ (Kossmann
2013: 293) – they too occasionally retain Arabic pronominal markers, e.g. Siwi
msabb-ha ‘for her’ < min sababi-hā ‘from reason.obl-obl.3sg.f’ (Souag 2013: 48).
In Awjila, more unusually, two inherited prepositions somewhat variably take
Arabic pronominal markers, e.g. dit-ha ‘in front of her’ (van Putten 2014: 113).
A rarer but more spectacular example of morpheme borrowing is the borrowing of productive templates from Arabic. Such cases include the elative template
əCCəC in Siwi, used to form the comparative degree of triliteral adjectives irrespective of etymology – thus əmləl ‘whiter’ < a-məllal alongside əṭwəl ‘taller’
< a-ṭwil < Arabic ṭawīl (Souag 2009) – and the diminutive template CCiCəC in
Ghomara (Mourigh 2016), e.g. aẓwiyyəṛ ‘little root’ < aẓaṛ alongside ləmwiyyəs
‘little knife’ < l-mus < Arabic al-mūsā ‘razor’ (gemination of y is automatic in
the environment i_V). As the latter example illustrates, borrowed derivational
morphology sometimes becomes productive.
The effects of Arabic on Berber morphology are by no means limited to the
borrowing of morphemes. There is reason to suspect Arabic influence of having
played a role in processes of simplification attested mainly in peripheral varieties,
such as the loss of case marking in many areas. In Siwi, where Arabic influence
appears on independent grounds to be unusually high, the verbal system shows
a number of apparent simplifications targeting categories absent in sedentary
Arabic varieties: the loss of distinct negative stems, the near-complete merger of
perfective with aorist, the fixed postverbal position of object clitics, and so on. It
is tempting to explain such losses as arising from imperfect acquisition of Siwi
by Arabic speakers.
Structural calquing in morphology is also sporadically attested. Siwi has lost
distinct feminine plural agreement on verbs, pronouns, and demonstratives, extending the inherited masculine plural forms to cover plural agreement irrespective of gender. Within Berber, this is unprecedented; plural gender agreement is
extremely well conserved across the family. However, it perfectly replicates the
usual sedentary Arabic system found in Egypt and far beyond.
3.4 Syntax
Syntactic influence is often difficult to identify positively. Nevertheless, Berber
offers a number of examples, and relative clause formation is one of the clearest
(Souag 2013: 151–156; Kossmann 2013: 369–407). Relative clauses in Berber are
normally handled with a gap strategy combined with fronting of any stranded
prepositions, as in (1).
411
Lameen Souag
(1)
Awjila (Paradisi 1961: 79)
ərrafəqa-nnəs
wi
ižin-an-a
nettin id-sin
ksum
friend.pl-gen.3sg rel.pl.m divide-3pl.m-prf 3sg.m with-obl.3pl.m meat
‘his friends with whom he divided the meat’
In subject relativisation, a special form of the verb not agreeing in person (the
so-called “participle”) is used, as in (2); such a form is securely reconstructible
for proto-Berber (Kossmann 2003).
(2) Awjila (Paradisi 1960: 162)
amədən wa
tarəv-ən
nettin ʕayyan
man
rel.sg.m write.ipfv-ptcp 3sg.m ill
‘The man who is writing is ill.’
In several smaller easterly varieties apart from Awjila, however, both of these
traits have been lost. The strategy found in varieties such as Siwi – resumptive
weak (affixal) pronouns throughout, and regular finite agreement for subject relativisation – perfectly parallels Arabic:
(3) Siwi (Souag 2013: 151–152)
tálti
tən
dəzz-ɣ-as
ǧǧəwab
woman rel.sg.f send-1sg-dat.3sg letter
‘the woman to whom I sent the letter’
(4)
Siwi (field data)
iməǧran
ággʷid wənn i-ʕəṃṃaṛ
man rel.sg.m 3sg.m-make.ipfv sickle.pl
‘the man who makes sickles’
In the case of verbal negation, an originally syntactic calque has often been
morphologised in parallel in Arabic and Berber. A number of varieties – especially the widespread Zenati subgroup of Berber, ranging from eastern Morocco
to northern Libya – have developed a postverbal negative clitic -š(a) from *ḱăra
‘thing’, apparently a calque on Arabic -š(i) from šayʔ ; however, some instead use
the direct borrowings ši or šay (Lucas 2007; Kossmann 2013: 332–334).
3.5 Lexicon
Lexical borrowing from Arabic is pervasive in Berber. Out of 41 languages around
the world compared in the Loanword Typology Project (Tadmor 2009), Tarifiyt
412
18 Berber
Berber was second only to (Selice) Romani in the percentage of loanwords –
more than half (51.7%) of the concepts compared. More than 90% of loanwords
examined in Tarifiyt were from Arabic, almost all from dialectal Maghrebi Arabic. There is little reason to suppose that Tarifiyt is exceptional in this respect
among Northern Berber languages; to the contrary, Kossmann (2013: 110) finds
its rate of basic vocabulary borrowing to be typical of Northern Berber, whereas
Siwi and Ghomara go much higher. The rate of borrowing from Arabic, however,
is considerably lower further south and west; on a 200-word list of basic vocabulary, Chaker (1984: 225–226) finds 38% Arabic loans in Kabyle (north-central Algeria) vs. 25% in Tashelhiyt (southern Morocco) and only 5% in Tahaggart Tuareg
(southern Algeria).
This borrowing is pervasive across the languages concerned, rather than being restricted to particular domains. Every semantic field examined for Tarifiyt,
including body parts, contained at least 20% loanwords, and verbs or adjectives
were about as frequently borrowed as nouns were (Kossmann 2009). Numerals
stand out for particularly massive borrowing; most Northern Berber varieties
have borrowed all numerals from Arabic above a number ranging from ‘one’ to
‘three’ (Souag 2007).
The effects of this borrowing on the structure of the lexicon remain insufficiently investigated, but appear prominent in such domains as kinship terminology. Throughout Northern Berber, a basic distinction between paternal kin
and maternal kin is expressed primarily with Arabic loanwords (ʕammi ‘paternal uncle’ vs. ḫali ‘maternal uncle’ etc.), whereas in Tuareg that distinction is not
strongly lexicalised. Nevertheless, borrowing does not automatically entail lexical restructuring; Tashelhiyt, for example, kept its vigesimal system even after
borrowing the Arabic word for ‘twenty’ (ʕšrin), cf. Ameur (2008: 77).
The borrowing of analysable multi-word phrases – above all, numerals followed by nouns – stands out as a rather common outcome of Berber contact
with Arabic. Usually this is limited to the borrowing of numerals in combination
with a limited set of measure words, such as ‘day’; thus in Siwi we find forms
like sbaʕ-t iyyam ‘seven days’ rather than the expected regular formation *səbʕa
n nnhaṛ-at (Souag 2013: 114). In Beni Snous (western Algeria), the phenomenon
seems to have gone rather further: Destaing (1907: 212) reports that numerals
above ‘ten’ systematically select for Arabic nouns. Souag & Kherbache (2016),
however, explain this as a code-switching effect, rather than a true case of one
language’s grammar requiring shifts into another.
413
Lameen Souag
4 Conclusion
The influence of Arabic on Berber has come to be better understood over the
past couple of decades, but much remains to be done. Synchronically, Berber–
Arabic code-switching remains virtually unresearched; rare exceptions include
Hamza (2007) and Kossmann (2012). Sociolinguistic methods could help us better understand the gradual integration of new Arabic loanwords; the early efforts
of Brahimi (2000) have hardly been followed up on. Diachronically, it remains
necessary to move beyond the mere identification of loanwords and contact effects towards a chronological ordering of different strata, an approach explored
for some peripheral varieties by Souag (2009) and van Putten & Benkato (2017).
While linguists are belatedly beginning to take advantage of earlier manuscript
data to understand the history of Berber (van den Boogert 1997; 1998; Brugnatelli
2011; Meouak 2015), this data has not yet been used in any systematic way to
help date the effects of contact at different periods. For many smaller varieties,
especially in the Sahara, basic documentation and description are still necessary
before the influence of Arabic can be explored. The unprecedented degree of Arabic influence revealed in Ghomara by recent work (Mourigh 2016), extending to
the borrowing of full verb paradigms, suggests that such descriptive work may
yet yield dividends in the study of contact.
Despite all these gaps, the work done so far is more than sufficient to establish
a general picture of Arabic influence on Berber. Throughout Northern Berber,
Arabic influence on the lexicon is substantial and pervasive, bringing with it
significant effects on phonology and morphology. Structural effects of Arabic on
morphology, and Arabic influence on Berber syntax, are less conspicuous but
nevertheless important, especially in smaller varieties such as Siwi. Looking at
these results through Van Coetsem’s (1988; 2000) framework, this suggests that
speakers dominant in the recipient languages have had an especially prominent
role in Arabic–Berber contact in larger varieties, whereas the role of speakers
dominant in the source language is more visible in smaller varieties. However,
this a priori conclusion should be tested against directly attested historical data
wherever possible.
Further reading
) The key reference for Arabic influence on Northern Berber is Kossmann (2012),
frequently cited above; this covers all levels of influence including the lexicon,
phonology, nominal and verbal morphology, borrowing of morphological categories, and syntax.
414
18 Berber
) The most extensive in-depth study of Arabic influence on a specific Berber
variety is Souag (2013), effectively a contact-focused grammatical sketch of
Siwi Berber.
) Mourigh (2016) is a thorough synchronic description of by far the most strongly
Arabic-influenced Berber variety, Ghomara, giving a uniquely clear picture of
just how far the process can go without resulting in language shift.
Acknowledgements
The author thanks his consultants in Siwa, especially the late Sherif Bougdoura,
for their help with studying Siwi.
Abbreviations
1, 2, 3
dat
f
gen
ipfv
m
1st, 2nd, 3rd person
dative
feminine
genitive
imperfective
masculine
obl
pl/pl.
prf
ptcp
sg
oblique
plural
perfect (suffix conjugation)
participle
singular
References
Ait Habbouche, Khadidja. 2013. Language maintenance and language shift among
Kabyle speakers in Arabic speaking communities: The case of Oran. Oran:
University of Oran. (Master’s dissertation).
Ameur, Meftaha. 2008. Les contacts arabe-amazighe: Le cas des noms de nombre.
In Mena Lafkioui & Vermondo Brugnatelli (eds.), Berber in contact: Linguistic
and sociolinguistic perspectives, 63–80. Cologne: Rüdiger Köppe.
Brahimi, Fadila. 2000. Loanwords in Algerian Berber. In Jonathan Owens (ed.),
Arabic as a minority language, 371–382. Berlin: Mouton de Gruyter.
Brugnatelli, Vermondo. 2011. Some grammatical features of Ancient Eastern
Berber (the language of the Mudawwana). In Luca Busetto (ed.), Studies on
language and African linguistics in honour of Marcello Lamberti, 29–40. Milano:
Qu.A.S.A.R.
Chaker, Salem. 1984. Textes en linguistique berbère: Introduction au domaine
berbère. Paris: Editions du Centre national de la recherche scientifique.
415
Lameen Souag
CIA. 2017. The world factbook 2017. Washington, DC: Central Intelligence
Agency.
Destaing, Edmond. 1907. Etude sur le dialecte berbère des Beni-Snous. Paris: Ernest
Leroux.
El Arifi, Samir. 2014. Tachelhit de l’Atlas blidéen. 4th edn. Algiers: Haut Commissariat à l’Amazighité.
Gautier, E.-F. 1913. Répartition de la langue berbère en Algérie. Annales de géographie 22(123). 255–266.
Hamza, Belgacem. 2007. Berber ethnicity and language shift in Tunisia. Brighton:
University of Sussex. (Doctoral dissertation).
Jolivet, Remi. 2008. La fidélité au tamashek des berbérophones dans le contexte
multilingue du Niger. In Mena Lafkioui & Vermondo Brugnatelli (eds.), Berber
in contact: Linguistic and sociolinguistic perspectives, 139–149. Cologne: Rüdiger
Köppe.
Kossmann, Maarten. 2003. The origin of the Berber “participle”. In M. Lionel
Bender, Gábor Takács & David L. Appleyard (eds.), Selected comparative–
historical Afrasian linguistics in the memory of Igor M. Diakonoff, 27–40.
Munich: LINCOM Europa.
Kossmann, Maarten. 2009. Loanwords in Tarifiyt. In Martin Haspelmath & Uri
Tadmor (eds.), Loanwords in the world’s languages: A comparative handbook,
191–214. Berlin: De Gruyter.
Kossmann, Maarten. 2010. Parallel system borrowing: Parallel morphological
systems due to the borrowing of paradigms. Diachronica 27(3). 459–487.
Kossmann, Maarten. 2011. A grammatical sketch of Ayer Tuareg (Niger). Cologne:
Rüdiger Köppe.
Kossmann, Maarten. 2012. Berber–Arabic code-switching in Imouzzar du Kandar.
STUF – Language Typology and Universals 65(4). 369–382.
Kossmann, Maarten. 2013. The Arabic influence on Northern Berber. Leiden: Brill.
Lucas, Christopher. 2007. Jespersen’s cycle in Arabic and Berber. Transactions of
the Philological Society 105(3). 398–431.
Mammeri, Mouloud. 1976. Tajerrumt n tmazight (tantala taqbaylit). Paris:
Maspéro.
Mechehed, Djamel-Eddine. 2007. L’organisation des notices de catalogage des
manuscrits arabes ou berbères: Cas de la collection Ulahbib, Kabylie. In Les
manuscrits berbères au Maghreb et dans les collections européennes, 129–138. La
Fresquière: Atelier Perrousseaux.
Meouak, Mohamed. 2015. La langue berbère au Maghreb médiéval. Leiden: Brill.
Mouili, Fatiha. 2013. The Berber language of Igli: Language towards extinction.
Humanities and Social Sciences Review 2(2). 563–580.
416
18 Berber
Mourigh, Khalid. 2016. A grammar of Ghomara Berber (north-west Morocco).
Cologne: Rüdiger Köppe.
Múrcia Sánchez, Carles. 2010. La llengua amaziga a l’antiguitat a partir de les
fonts gregues i llatines. Barcelona: Universitat de Barcelona.
Naumann, Christfried. 2012. Acoustically based phonemics of Siwi (Berber).
Cologne: Rüdiger Köppe.
Paradisi, Umberto. 1960. Il berbero di Augila: Materiale lessicale. Rivista degli
Studi Orientali 35. 157–177.
Paradisi, Umberto. 1961. Testi berberi di Augila (Cirenaïca). Annali dell’Istituto
Orientale di Napoli 10. 79–91.
Pichler, Werner. 2007. Origin and development of the Libyco-Berber script. Cologne:
Rüdiger Köppe.
Serreli, Valentina. 2017. Language and identity in Siwa Oasis: Indexing belonging,
localness and authenticity in a small minority community. In Reem Bassiouney
(ed.), Identity and dialect performance: A study of communities and dialects, 226–
242. London: Routledge.
Souag, Lameen. 2007. The typology of number borrowing in Berber. In Naomi
Hilton, Rachel Arscott, Katherine Barden, Sheena Shah & Meg Zellers (eds.),
Proceedings for the Fifth Cambridge Postgraduate Conference in Linguistics, 237–
244. Cambridge: Cambridge Institute of Language Research.
Souag, Lameen. 2009. Siwa and its significance for Arabic dialectology. Zeitschrift
für Arabische Linguistik 51. 51–75.
Souag, Lameen. 2013. Berber and Arabic in Siwa (Egypt): A study in linguistic
contact. Cologne: Rüdiger Köppe.
Souag, Lameen & Fatma Kherbache. 2016. Syntactically conditioned codeswitching? The syntax of numerals in Beni-Snous Berber. International Journal
of Bilingualism 20(2). 97–115.
Souag, Lameen & Marijn van Putten. 2016. The origin of mid vowels in Siwi.
Studies in African Linguistics 45(1–2). 189–208.
Tadmor, Uri. 2009. Loanwords in the world’s languages: Findings and results. In
Martin Haspelmath & Uri Tadmor (eds.), Loanwords in the world’s languages:
A comparative handbook, 55–75. Berlin: De Gruyter.
Tigziri, Nora. 2008. Le kabyle au contact des langues en présence en Algérie:
Entre codeswitching et parler hybride? In Mena Lafkioui & Vermondo Brugnatelli (eds.), Berber in contact: Linguistic and sociolinguistic perspectives, 175–
185. Cologne: Rüdiger Köppe.
Van Coetsem, Frans. 1988. Loan phonology and the two transfer types in language
contact. Dordrecht: Foris.
417
Lameen Souag
Van Coetsem, Frans. 2000. A general and unified theory of the transmission process
in language contact. Heidelberg: Winter.
van den Boogert, Nico. 1997. The Berber literary tradition of the Sous, with an
edition and translation of “The Ocean of Tears” by Muḥammad Awzal (d. 1749).
Leiden: Nederlands Instituut voor het Nabije Osten.
van den Boogert, Nico. 1998. “La révélation des énigmes”: Lexiques arabo-berbères
des XVIIe et XVIIIe siècles. Aix-en-Provence: Institut de recherches et d’études
sur le monde arabe et musulman (IREMAM).
van Putten, Marijn. 2014. A grammar of Awjila Berber (Libya), based on Umberto
Paradisi’s material. Cologne: Rüdiger Köppe.
van Putten, Marijn & Adam Benkato. 2017. The Arabic strata in Awjila Berber.
In Ahmad Al-Jallad (ed.), Arabic in context: Celebrating 400 years of Arabic at
Leiden University, 476–502. Leiden: Brill.
Vycichl, Werner. 2005. Sprachskizzen: Die berberischen Sprachinseln in
Tunesien. In Dymitr Ibriszimow & Maarten Kossmann (eds.), Berberstudien &
A sketch of Siwi Berber (Egypt), 136–147. Cologne: Rüdiger Köppe.
418
Chapter 19
Beja
Martine Vanhove
LLACAN (CNRS, INALCO)
This chapter argues for two types of outcomes of the long-standing and intense contact situation between Beja and Arabic in Sudan: borrowings at the phonological,
syntactic and lexical levels, and convergence at the morphological level.
1 Current state and historical development
1.1 Historical development of Beja
Beja is the sole language of the Northern Cushitic branch of the Afro-Asiatic
phylum. Recent archaeological discoveries show growing evidence that Beja is
related to the extinct languages of the Medjay (from which the ethnonym Beja is
derived; Rilly 2014: 1175), and Blemmye tribes, first attested on Egyptian inscriptions of the Twelfth Dynasty for the former, and on a Napatan stela of the late
seventh century BCE for the latter. For recent discussions, see Browne (2003);
El-Sayed (2011); Zibelius-Chen (2014); Rilly (2014); and Rilly (2018). The Medjays were nomads living in the eastern Nubian Desert, between the first and
second cataracts of the River Nile. The Blemmyes invaded and took part in defeating the Meroitic kingdom, fought against the Romans up to the Sinai, and
ruled Nubia from Talmis (modern Kalabsha, between Luxor and Aswan) for a few
decades, before being defeated themselves by the Noubades around 450 CE (Rilly
2018). In late antiquity, the linguistic situation involved, in northern Lower Nubia, Cushitic languages, Northern Eastern Sudanic languages, to which Meroitic
and Nubian belong, also Coptic and Greek to some extent, and in the south, EthioSemitic. It is likely that there was mutual influence to an extent that is difficult
to disentangle today.
Martine Vanhove. 2020. Beja. In Christopher Lucas & Stefano Manfredi (eds.),
Arabic and contact-induced change, 419–439. Berlin: Language Science Press.
DOI:10.5281/zenodo.3744537
Martine Vanhove
1.2 Current situation of Beja
The Beja territory has shrunk a lot since late antiquity, and Beja (biɖawijeːt) is
mainly spoken today in the Red Sea and Kassala States in eastern Sudan, in the
dry lands between the Red Sea and the Atbara River. The 1993 census, the last
one to include a language question, recorded some 1,100,000 Beja speakers, and
there is probably at least double that figure today. There are also some 60,000
speakers in northern Eritrea, and there may be still a few speakers left in Egypt,
in the Nile valley at Aswan and Daraw, and along the coast towards Marsa Alam
(Morin 1995; Wedekind 2012). In Sudan today, Beja speakers have also settled in
Khartoum and cities in central and western Sudan (Hamid Ahmed 2005a: 67).
All Bejas today are Muslims. They consider themselves Bedouins, and call
themselves arab ‘Arab’;1 they call the ethnic Arabs balawjeːt. Before the introduction of modern means of transportation, they were traditionally the holders
of the caravan trade in the desert towards the west, south and north of their
territory, and they still move between summer and winter pastures with their
cattle. They also produce sorghum and millet for daily consumption, and fruits
and vegetables in the oases. The arrival of Rashaida migrants from Saudi Arabia
in the nineteenth century created tensions in an area with meagre resources, but
the first contemporary important social changes took place during the British
mandate with the agricultural development of the Gash and Tokar areas, and the
settlement of non-Beja farmers. The droughts of the mid-1980s brought about a
massive exodus towards the cities, notably Port Sudan and Kassala, followed by
job diversification, and increased access to education in Arabic, although not generalized, especially for girls, who rarely go beyond primary level (Hamid Ahmed
2005a).
Beja is mostly an oral language. In Eritrea, a Latin script was introduced in
schools after independence in 1993, but in Sudan no education in Beja exists.
Attempts made by the Summer Institute of Linguistics and at the University of
the Red Sea to implement an Arabic-based script did not come to fruition. On
the other hand, in the last few years school teachers in rural areas have begun to
talk more and more in Beja in order to fight illiteracy (in Arabic) and absenteeism
(Onour 2015).
1
In Sudan the term ʕarab is widely used for referring to nomad groups in general, and not only
to ethnically defined Arabs. Thanks to Stefano Manfredi for this information..
420
19 Beja
2 Arabic–Beja contact
Contact between Bejas and Arabs started as early as the beginning of Islamization, and through trade relations with Muslim Egypt, as well as Arab incursions
in search of gold and emerald. Evidence of these contacts lies in the early Arabicization of Beja anthroponyms (Záhořík 2007). The date of the beginning of
Islamization differs according to authors, but it seems it started as early as the
tenth century, and slowly expanded until it became the sole religion between the
eighteenth and nineteenth centuries (Záhořík 2007).
We have no information concerning the onset and spread of Beja–Arabic bilingualism. It is thus often impossible to figure out if a transfer occurred through
Beja-dominant speakers or was imposed by fluent Beja–Arabic bilingual speakers, and consequently to decide whether a contact-induced feature belongs to the
borrowing or to the imposition type of transfer as advocated by Van Coetsem
(1988; 2000) and his followers. What is certain though, is that socio-historical as
well as linguistic evidence speaks in favour of Beja–Arabic bilingualism as an
ancient phenomenon, but in unknown proportion among the population. With
the spread of Islam since the Middle Ages, contact with Arabic became more and
more prevalent in Sudan.
In this country, which will be the focus of this chapter, bilingualism with Sudanese Arabic is frequent, particularly for men, and expanding, including among
women in cities and villages, but to a lesser extent. Bejas in Port Sudan are also
in contact with varieties of Yemeni Arabic. Rural Bejas recently settled at the
periphery of the big cities have the reputation of being more monolingual than
others, which was still the case fifteen years ago (Vanhove 2003).
The Beja language is an integral part of the social and cultural identity of the
people, but it is not a necessary component. Tribes and clans that have switched
to Arabic, or Tigre, such as the Beni Amer, are considered Bejas. Beja is prestigious, since it allows its speakers to uphold the ethical values of the society, and is
considered to be aesthetically pleasing due to its allusive character. The attitude
towards Arabic is ambivalent. It is perceived as taboo-less, and thus contrary to
the rules of honour, nevertheless it is possible to use it without transgressing
them. Arabic is also prestigious because it is the language of social promotion
and modernity (Hamid Ahmed 2005b). Language attitudes are rapidly changing,
and there is some concern among the Beja diaspora about the future of the Beja
language, even though it cannot be considered to be endangered. Some parents
avoid speaking Beja to their children, for fear that it would interfere with their
learning of Arabic at school, leaving to the grandparents the transmission of Beja
421
Martine Vanhove
(Wedekind 2012; Vanhove 2017). But there is no reliable quantitative or qualitative sociolinguistic study of this phenomenon. Code-switching between Beja and
Arabic is spreading but understudied.
This sketch of the sociolinguistic situation of Beja speaks for at least two types
of transfer: (i) borrowing, where the agents of transfer are dominant in the recipient language (Beja); (ii) convergence phenomena, since the difference in linguistic dominance between the languages of the bilingual speakers tends to be really
small (at least among male speakers today, and probably earlier in the history of
Beja; see Van Coetsem 1988: 87). Imposition has probably also occurred of course,
but it is not always easy to prove.
3 Contact-induced changes in Beja
3.1 Phonology
The few contact-induced changes in Beja phonology belong to the borrowing
type.
The phonological system of Beja counts 21 consonantal phonemes, presented
in Table 1.
Affricate
Fricative
Nasal
Trill
Lateral
Approximant
m
w
s
n
r
l
ʤ
ʃ
eal
Lar
yng
Lab
io-v
elar
Vel
ar
l
ʈ
ɖ
ata
t
d
Pal
rofl
ex
r
Ret
f
b
Alv
eola
Plosive
Lab
Bila
bial
io-d
ent
al
Table 1: Beja consonants
k
g
kʷ
gʷ
ʔ
h
j
The voiced post-alveolar affricate ʤ (often realized as a voiced palatal plosive
[ɟ] as in Sudanese Arabic) deserves attention as a possible outcome of contact
422
19 Beja
with Arabic. Since Reinisch (1893: 17), it is usually believed that this affricate is
only present in Arabic loanwords and is not a phoneme (Roper 1928; Hudson
1976; Morin 1995). The existence of a number of minimal pairs in word-initial position invalidates the latter analysis: ʤiːk ‘rooster’ ~ ʃiːk ‘chewing tobacco’; ʤhar
‘chance’ ~ dhar ‘bless’; ʤaw ‘quarrel’ ~ ɖaw ‘jungle’ ~ ʃaw ‘pregnancy’ ~ gaw
‘house’ (Vanhove 2017). As for the former claim, there are actually a few lexical
items such as bʔaʤi ‘bed’, gʷʔaʤi ‘one-eyed’ (gʷʔad ‘two eyes’), that cannot be
traced back to Arabic (the latter is pan-Cushitic; Blažek 2000). Nevertheless, it
is the case that most items containing this phoneme do come from (or through)
(Sudanese) Arabic: aːlaʤ ‘tease’, aʤiːn ‘dough’, aʤib ‘please’, ʔaʤala ‘bicycle’,
ʔiʤir ‘divine reward’, ʤaːhil ‘small child’, ʤabana ‘coffee’, ʤallaːj ‘because of’,
ʤallab ‘fish’, ʤanna ‘paradise’, ʤantaːji ‘djinn’, ʤarikaːn ‘jerrycan’, ʤeːb ‘pocket’,
ʤhaliː ‘coal’, ʤimʔa ‘week’, ʤins ‘sort’, ʤuwwa ‘inside’, faʤil ‘morning’, finʤaːn
‘cup’, hanʤar ‘dagger’, hiʤ ‘pilgrimage’, maʤaʔa ‘famine’, maʤlis ‘reconciliation meeting’, siʤin ‘prison’, tarʤimaːl ‘translator’, waʤʤa ‘appointment’, and
xawaʤa ‘foreigner’. It is clear that ʤ is not marginal anymore. However ʤ is
unstable: it has several dialectal variants, ʧ, g and d, and may alternate with the
dental d or retroflex ɖ, in the original Beja lexicon (ʤiwʔoːr/ɖiwʔoːr ‘honourable
man’) as well as in loanwords (aʤiːn/aɖiːn ‘dough’) (Vanhove & Hamid Ahmed
2011; Vanhove 2017). In my data, which counts some 50 male and female speakers
of all age groups, this is rarely the case, meaning that there is a good chance that
this originally marginal phoneme will live on under the influence of (Sudanese)
Arabic.
There are two other consonants in Arabic loanwords that are regularly used
by the Beja speakers: z and x, neither of which can be considered phonemes since
there are no minimal pairs.
Blažek (2007: 130) established a regular correspondence between Beja d and
Proto-East-Cushitic *z. In contemporary Beja z only occurs in recent loanwords
from Sudanese Arabic such as ʤaza ‘wage’, ʤoːz ‘pair’, rizig ‘job’, wazʔ ‘offer’,
xazna ‘treasure’, zamaːn ‘time’, zirʔa ‘field’, zuːr ‘visit’. It may alternate with d,
even within the speech of the same speaker as free variants, e.g. damaːn, dirʔa,
duːr. The fricative alveolar pronunciation is more frequent among city dwellers,
who are more often bilingual. It is difficult to ascertain whether Beja is in the process of re-acquiring the voiced fricative through contact with Sudanese Arabic,
or whether it will undergo the same evolution to a dental stop as in the past.
A few recent Arabic loanwords may also retain the voiceless velar fricative
x (see also Manfredi et al. 2015: 304–305): xazna ‘treasure’, xawaʤa ‘foreigner’,
xaddaːm ‘servant’, xaːtar ‘be dangerous’, aːxar ‘last’. In my data, this is usually the
423
Martine Vanhove
case in the speech of fluent bilingual speakers. We thus have here a probable imposition type of transfer. In older borrowings, even among these speakers, Arabic
x shifted to h (xajma > heːma ‘tent’). It may be because these older loans spread in
a community which was at that time composed mainly of Beja-dominant speakers, but we have no means of proving this hypothesis.
3.2 Morphology
3.2.1 General remarks
Most Cushitic languages only have concatenative morphology, the stem and pattern schema being at best highly marginal (Cohen 1988: 256). In addition to Beja,
Afar and Saho (Lowland East-Cushitic branch), Beja’s geographically closest sisters, are exceptions, and all three languages use also non-concatenative morphology. In Afar and Saho it is far less pervasive than in Beja; in particular they do
not use vocalic alternation for verbal derivation, this feature being restricted to
the verb flexion of a minority of underived verbs.
Even though Beja and Arabic share a similar type of morphology, the following overview shows that each language has developed its own system. Although
they have been in contact for centuries, neither small-scale nor massive borrowing from Arabic morphological patterns can be postulated for the Beja data. An
interpretation in terms of a convergence phenomenon is more relevant, both in
terms of semantics and forms.
Non-concatenative morphology concerns an important portion of the lexicon:
a large part of the verb morphology (conjugations, verb derivations, verbal noun
derivations), and part of the noun morphology (adjectives, nouns, “internal” plurals, and to a lesser extent, place and instrument nouns). In what follows, I build
on Vanhove (2012) and Vanhove (2017), correcting some inaccuracies.
3.2.2 Verb morphology
Only one of the two Beja verb classes, the one conjugated with prefixes (or infixes), belongs to non-concatenative morphology. This verb class (V1) is formed
of a stem which undergoes ablaut varying with tense–aspect–mood (TAM), person and number, to which prefixed personal indices for all TAMs are added (plural and gender morphemes are also suffixes). V1 is diachronically the oldest pattern, which survives only in a few other Cushitic languages. In Beja V1s are the
majority (57%), as against approximately 30% in Afar and Saho, and only five
verbs in Somali and South Agaw (Cohen 1988: 256). Table 2 provides examples
in the perfective and imperfective for bi-consonantal and tri-consonantal roots.
424
19 Beja
Table 2: Perfective and imperfective patterns
pfv
ipfv
Bi-consonantal dif ‘go’
Tri-consonantal kitim ‘arrive’
i-dif ‘he went’
i-dif-na ‘they went’
i-n-diːf ‘he goes’
eː-dif-na ‘they go’
i-ktim ‘he arrived’
i-ktim-na ‘they arrived’
k<an>tiːm ‘he arrives’
eː-katim-na ‘they arrive’
Prefix conjugations are used in Arabic varieties and South Semitic languages
but their functions and origins are different. In South Semitic, the prefix conjugation has an aspectual value of imperfective, while in Cushitic it marks a particular morphological verb class. The Cushitic prefix conjugation (in the singular)
goes back to auxiliary verbs meaning ‘say’ or ‘be’, while the prefix conjugation
of South Semitic has various origins, none of them including a verb ‘say’ or ‘be’
(Cohen 1984). Although different grammaticalization chains took place in the two
branches of Afro-Asiatic, this suggests that the root-and-pattern system might
have already been robust in Beja at an ancient stage of the language. It is noteworthy that there are at least traces of vocalic alternation between the perfective
and the imperfective in all Cushitic branches (Cohen 1984: 88–102), thus reinforcing the hypothesis of an ancient root-and-pattern schema in Beja. In what
proportion this schema was entrenched in the morphology of the proto-Cushitic
lexicon is impossible to decide.
Verb derivation of V1s is also largely non-concatenative. Beja is the only Cushitic language which uses qualitative ablaut in the stem for the formation of semantic and voice derivation. The ablaut can combine with prefixes.
Table 3 presents the five verb derivation patterns with ablaut, and Table 4
shows the absence of correspondence between the Beja and Arabic (Classical
and Sudanese) patterns. Sudanese patterns are extracted from Bergman (2002:
32–34), who does not provide semantic values.
Among the Semitic languages, an intensive pattern similar to the Beja one is
only known in some Modern South Arabian languages spoken in eastern Yemen
(not in contact with Beja), where it is also used for causation and transitivization
(Simeone-Senelle 2011: 1091). The Modern South Arabian languages are close relatives of Ethio-Semitic languages and it is usually considered that the latter were
brought to the Horn of Africa by South-Arabian speakers (Ullendorf 1955). However, this ablaut pattern was not retained in Ethio-Semitic. It is also unknown in
Cushitic. In Classical Arabic, the plurisyllabic pattern does not have an intensive
value, but a goal or sometimes reciprocal meaning.
425
Martine Vanhove
Table 3: V1 derivation patterns with ablaut
Monosyllabic V1
Plurisyllabic V1
int
mid
pass
boːs (bis ‘burry’)
faf (fif ‘pour’)
aːtoː-maːn (min ‘shave’)
recp
caus
amoː-gaːd (gid ‘throw’)
kaːtim (kitim ‘arrive’)
rimad (rimid ‘avenge’)
at-dabaːl (dibil ‘gather’)
am-heːjid (haːjid ‘sew’)
am-garaːm (girim ‘be inimical’)
si-katim (kitim ‘arrive’)
Table 4: Comparison between Beja and Arabic derivation patterns
int
mid
Beja Plurisyllabic V1
Classical Arabic
Sudanese Arabic
CaːCa kaːtim < kitim
‘arrive’
CiCaC rimad < rimid
‘avenge’
CaCCaCa, CaːCaCa
CaCCaC, CaːCaC
CuCiCa,
iC<t>aCaCa,
ta-CaCCaCa,
ta-CaːCaCa,
in-CaCaCa,
tas-CaCaCa,
ista-CaCaCa
CuCiCa,
iC<t>aCaCa,
ta-CaCCaCa,
ta-CaːCaCa,
in-CaCaCa,
ista-CaCaCa
CaːCaCa ,
ta-CaːCaCa,
iC<t>aCaCa
CaCCaCa, ʔa-CCaCa
it-CaCCaC, it-CaCaC
pass
at-CaCaːC at-dabaːl
< dibil ‘gather’
am-CeːCiC am-heːjid
< haːjid ‘sew’
recp
am-CaCaːC
am-garaːm < girim
‘be inimical’
si-CaCiC si-katim <
kitim ‘arrive’
caus
426
it-CaːCaC, it-CaCaC,
CiCiC, in-CaCaC
CaːCaC, it-CaːCaC
CaCCaC, a-CCaC
19 Beja
Beja is the sole Cushitic language which differentiates between active and middle voices by means of vocalic alternation. Remnants of this pattern exist in some
Semitic languages, among them Arabic, in a fossilized form.
In Cushitic, qualitative ablaut for the passive voice only occurs in Beja. Passive
formation through ablaut exists in Classical and Sudanese Arabic, but with different vowels. Bergman (2002: 34) mentions that “a handful of verbs in S[udanese]
A[rabic]” can be formed this way. For Stefano Manfredi (personal communication) it is a productive pattern in this Arabic variety.
Like the passive voice, the reciprocal is characterized by a qualitative ablaut in
aː in the stem, but the prefix is different and consists of am(oː)-. m is not used for
verbal derivation in Arabic, which uses the same ablaut, but for the first vowel
of disyllabic stems, to express, marginally, the reciprocal of the base form. Most
often the reciprocal meaning is expressed by other forms with the t- prefixed or
infixed to the derived form or the base form. In some other Cushitic languages -m
is used as a suffix for passive or middle voice (without ablaut). In Beja m- can also
marginally be used as a passive marker, together with ablaut, for a few transitive
intensive verbs: ameː-saj ‘be flayed’, ameː-biɖan ‘be forgotten’.
Although a suffix -s (not a prefix as in Beja) is common in Cushitic, Beja is once
more the only Cushitic language which uses ablaut for the causative derived form.
Neither ablaut nor the s- prefix exist in Arabic. Arabic uses different patterns for
the causative: the same as the intensive one, i.e. with a geminated second root
consonant, and the (Ɂ)a-CcaC(a) pattern.
This brief overview shows that Beja has not borrowed patterns from (Sudanese) Arabic, but has at best similar, but not exact, cognate patterns which are
marginal in both Classical and Sudanese Arabic.
Beja also has four non-finite verb forms. The simultaneity converb of V1s is
the only one with non-concatenative morphology. The affirmative converb is
marked for both verb classes with a suffix -eː added to the stem: gid ‘throw’,
gid-eː ‘while throwing’; kitim ‘arrive’, kitim-eː ‘while arriving’. In the negative,
the negative particle baː- precedes the stem, and V1s undergo ablaut in the stem
(CiːC and CaCiːC), and drop the suffix; it has a privative meaning: baː-giːd ‘without throwing’; baː-katiːm ‘without arriving’. No similar patterns exist in Arabic
or other Cushitic languages.
3.2.3 Verbal noun derivation
In the verbal domain, non-concatenative morphology concerns only V1s. With
nouns, it applies to action nouns (maṣdars) and agent nouns.
427
Martine Vanhove
There are several maṣdar patterns, with or without a prefix, with or without
ablaut, depending mostly on the syllabic structure of the verb. The most frequent
ones with ablaut are presented below.
The pattern m(i(ː)/a)-CV(ː)C applies to the majority of monosyllabic verbs. The
stem vowel varies and is not predictable: di ‘say’, mi-jaːd ‘saying’; dir ‘kill’, madar ‘killing’; sʔa ‘sit down’, ma-sʔaː ‘sitting’; ak ‘become’, miː-kti ‘becoming’; hiw
‘give’, mi-jaw ‘gift, act of giving’. A few disyllabic V1s comply to this pattern:
rikw ij ‘fear’, mi-rkʷa:j ‘fearing’; jiwid ‘curl’, miː-wad ‘curling’. Some V1s of the
CiC pattern have a CaːC pattern for maṣdars, without a prefix: gid ‘throw’, gaːd
‘throwing’. In Classical Arabic, the marginal maṣdars with a prefix concern trisyllabic verbs, none showing a long vowel in the stem or the prefix, nor a vowel
i in the prefix.
CiCiC and HaCiC2 disyllabic verbs form their maṣdars by vocalic ablaut to
uː: kitim ‘arrive’, kituːm ‘arriving’; ʔabik ‘take’, ʔabuːk ‘taking’; hamir ‘be poor’,
hamuːr ‘being poor’. CiCaC V1s, and those ending in -j, undergo vocalic ablaut
to eː: digw agw ‘catch up’, digw eːgw ‘catching up’; biɖaːj ‘yawn’, biɖeːj ‘yawning’.
In Classical Arabic, the maṣdar pattern with uː has a different vowel in the first
syllable, a (in Beja a is conditioned by the initial laryngeal consonant), and it is
limited almost exclusively to verbs expressing movements and body positions
(Blachère & Gaudefroy-Demombynes 1975: 81).
Bergman (2002: 35) provides no information about verbal nouns of the base
form in Sudanese Arabic except that they are “not predictable”.
As for agent nouns of V1s, they most often combine ablaut with the suffix -aːna,
the same suffix as the one used to form agent nouns of V2 verbs, whose stems
do not undergo ablaut. The ablaut pattern is the same as with the verbal intensive derivation: bir ‘snatch’, boːr-aːna ‘snatcher’; gid ‘throw’, geːd-aːna ‘thrower,
a good shot’; dibil ‘pick up’, daːbl-ana ‘one who picks up’. Some tri-consonantal
stems have a suffix -i instead of -aːna: ʃibib ‘look at’, ʃaːbb-i ‘guard, sentinel’. Some
have both suffixes: kitim ‘arrive’, kaːtm-aːna/kaːtim-i ‘newcomer’.
These patterns are unknown in Arabic.
3.2.4 Noun morphology
3.2.4.1 General remarks
The existence of verbal noun derivation patterns and nominal plural patterns are
well recognized in the literature about Beja morphology; for a recent overview,
see Appleyard (2007). It is far from being the case for adjective and noun patterns.
2
Where H stands for the laryngeals ʔ and h.
428
19 Beja
All noun and adjective patterns linked to V1s are listed below. Vanhove (2012)
provides an overview of these patterns which are summed up below.
3.2.4.2 Adjective patterns
There are eight adjective patterns, two of which are shared with nouns. Most are
derived from V1 verbs, but the reverse is also attested. A corresponding verb form
is inexistent in a few cases. All patterns are based on ablaut, in two cases with
an additional suffix -a, or gemination of the medial consonant. Arabic has no
dedicated adjective pattern (but the active participle pattern of the verbal base
form CaːCiC may express properties). Table 5 provides the full list of patterns
with examples. It is remarkable that none of them is similar to those of Classical
Arabic or colloquial Sudanese Arabic (Bergman 2002: 17).
Table 5: Adjective patterns
Pattern
Adjective
Verb form
aCaːC
CaCCa
CaːCi(C)
amaːg ‘bad’
marɁa ‘wide’
naːkw is ‘short’
daːji ‘good’
dawil ‘close’
takw aːkw ‘prepared’
ragaːg-a ‘long’
ʃikw aːn ‘aromatic’
ʃallik ‘few’
mig ‘do evil’
mirɁ ‘be wide’
nikw is ‘be short’
∅
diwil ‘be close’
tikw ikw ‘prepare’
rigig ‘stand up’
ʃikw an ‘emit pleasant odour’
ʃilik ‘be few’
CaCi(C)
CaCaːC
CaCaːC-a
CiCaːC
CaCCiC
3.2.4.3 Nouns
There are eleven basic noun patterns related to V1 verbs. Most of the patterns for
triconsonantal roots resemble those of Arabic (but are not strictly identical), a
coincidence which is not surprising since both languages have a limited number
of vowels. Table 6 provides the full list of these patterns. The CaCi pattern is
shared with adjectives. The CiCi(C) pattern does not undergo ablaut.
429
Martine Vanhove
Table 6: Noun patterns
Pattern
Noun
Verb form
CaC
CiCa
CaCi
CaCa
CiCi(C)
nakw ‘pregnancy’
nisa ‘advise’
sari ‘wakefulness’
nada ‘dew’
mirɁi ‘width’
riʃid ‘wealth’
ʃaɖiːɖ ‘strip’
ʃakw iːn ‘fragrance’
raʃiːd-a ‘cattle’
tiːlaːl ‘stride’
taboːk ‘double handful’
tiluːl-a ‘exile’
nikw i ‘become pregnant’
nisa ‘advice’
sir ‘keep awake’
nidaj ‘sweat, exude water’
mirɁ ‘be wide’
riʃid ‘raise, tend crops or cattle’
ʃiɖiɖ ‘strip off’
ʃikw an ‘emit pleasant odour’
riʃid ‘raise, tend crops or cattle’
tilil ‘stride far away from home’
tiboːk ‘fill scoop with cupped hands’
tilil ‘stride far away from home’
CaCiːC
CaCiːC-a
CiːCaːC
CaCoːC
CiCuːC-a
3.2.4.4 Nouns with prefix m(V)A few other semantic types of nouns, mostly instrument and place names, are
formed through ablaut and a prefix m(V)-, like in Arabic. Contrary to Arabic
where these patterns are productive, they are frozen forms in Beja (some are
not loanwords from Arabic, see the last three examples): Ɂafi ‘prevent, secure’,
m-Ɂafaj ‘nail, rivet, fastener’; himi ‘cover’, m-himmeːj ‘blanket’; ginif ‘kneel’, mignaf ‘camp’; moːk ‘take shelter’, ma-kw a ‘shelter’; rifif ‘drag an object on the
ground’, mi-rfaf ‘reptile’.
3.2.5 Plural patterns
The so-called “internal” plural patterns are common and frequent in Arabic (and
Ethio-Semitic). Beja also has a limited set of internal plural patterns, but it has
developed its own system. Ablaut patterns for plural formation mainly concern
non-derived nouns containing either a long vowel or ending in a diphthong. Both
iː and uː turn to i in the plural, and aː, eː and oː turn to a, sometimes with the addition of the plural suffix -a; nouns ending in -aj turn to a long vowel -eːj: angw iːl,
pl. angw il ‘ear’; luːl, pl. lil ‘rope’; asuːl, pl. asil ‘blister’; hasaːl, pl. hasal/hasal-a
‘bridle’; meːk, pl. mak ‘donkey’; boːk, pl. bak ‘he-goat’; ganaj, pl. ganeːj ‘gazelle’
(Vanhove 2017).
430
19 Beja
Even though internal plurals can be considered as a genetic feature, the fact
that they are very rare or absent in other Cushitic languages (Zaborski 1986)
speaks for a possible influence of Arabic (in Sudan) upon Beja.
3.3 Syntax
3.3.1 General remarks
As far as we know, there are no syntactic calques from Arabic in Beja. There
are nevertheless a few borrowed lexical and grammatical items that gave rise to
constructions concerning coordination and subordination.
3.3.2 Coordination
One of the three devices that mark coordination is borrowed from Arabic wa. It is
only used for noun phrases or nominalized clauses (deranked, temporal and relative clauses), whereas the Arabic source particle can be used with noun phrases
and simple sentences. wa is preposed to the coordinated element in Arabic, but
in Beja it is an enclitic particle =wa, a position in line with the favoured SOV
word order. =wa follows each of the coordinated elements. (1) illustrates the coordination of two noun phrases.
(1)
Beja (BEJ_MV_NARR_01_shelter_057)3
bʔaɖaɖ=wa
i=koːlej=wa
sallam-ja=aj=heːb
sword=coord def.m=stick=coord give-pfv.3sg.m=csl=obj.1sg
‘Since he had given me a sword and the stick…’
Deranked clauses with non-finite verb forms, which partly have nominal properties (Vanhove 2016), are also coordinated with =wa. (2) is an example with the
manner converb, and (3) with the simultaneity converb.
(2)
3
Beja (BEJ_MV_NARR_14_sijadok_281-284)
winneːt si-raːkʷ-oːm-a=b=wa
plenty
caus-fear.int-pass-cvb.mnr=indf.m.acc=coord
far-iːni
ʔas-ti
gadab-aː=b=wa
be_sad-cvb.mnr=indf.m.acc=coord be_up-cvb.gnrl jump-ipfv.3sg.m
‘Very frightened and sad, he jumps up.’
The sources of the examples are accessible online at http://corporan.huma-num.fr/Archives/
corpus.php; the indications in parenthesis refer to the texts they are extracted from.
431
Martine Vanhove
(3) Beja (BEJ_MV_NARR_13_grave_126-130)
afirh-a=b
aka-jeː=wa
be_happy-cvb.mnr=indf.m.acc become-cvb.smlt=coord
i=dheːj=iːb
hawaː-jeː=wa
rh-ani
def.m=people=loc.sg play-cvb.smlt=coord see-pfv.1sg
‘I saw him happy and playing among the people.’
Relative and temporal subordinate clauses also have nominal properties: the
relative markers derive from the articles, and the temporal markers go back to
nouns. (4) illustrates the coordination with a relative clause which bears the coordination marker, and (5) the coordination of two temporal clauses.
(4)
Beja (06_foreigner_22-24)
uːn
ani
t=ʔarabijaːj=wa
oː=maːl
prox.sg.m.nom 1sg.nom def.f=car=coord def.sg.m.acc=treasure
jʔ-a=b
w=haːj
def.sg.m/rel=com come-cvb.mnr=indf.m.acc
a-kati=jeːb=wa
kass=oː
a-niːw=hoːk
1sg-become\ipfv=rel.m=coord all=poss.3sg.acc 1sg-give.ipfv=obj.2sg
‘I’ll give you a car and all the fortune that I brought.’
(5) Beja (BEJ_MV_CONV_01_rich_SP2_136-138)
naː=t
bi=i-hiːw=oː=hoːb=wa
thing=indf.f opt=3sg.m-give\neg.opt=obj.1sg=when=coord
i-niːw=oː=hoːb=wa
3sg.m-give.ipfv=obj.1sg=when=coord
‘Whether he gives it to me or not…’ (lit. when he does not give me
anything and when he gives me)
Adversative coordination between two simple clauses is also expressed with a
borrowing from Arabic: laːkin ‘but’.
3.3.3 Subordination
The reason conjunction sabbiː ‘because’ is a borrowing from the Arabic noun
sabab ‘reason’. Like most balanced adverbial clauses, it is based on one of the
relative clause types, the one nominalized with the noun na ‘thing’ in the genitive
case. sabbiː functions as the head of the relative clause.
432
19 Beja
(6) Beja (03_camel_192)
ʔakir-a
ɖab
ɖaːb-iːn=eː=naː-ji
sabbiː
be_strong-cvb.mnr run.ac run-aor.3pl=rel=thing-gen because
‘Because it was running so fast…’
sabbiː can also be used after a noun or a pronoun in the genitive case: ombarijoːk
sabbiː ‘because of you’.
Terminative adverbial clauses are expressed with a borrowing from Arabic,
hadiːd ‘limit’. Again the borrowing is the head of the relative clause.
(7)
Beja (BEJ_MV_NARR_51_camel_stallion_026-030)
oːn
i=kaːm=oːk
heː=heːb
prox.sg.m.acc def.m=camel=poss.2sg.acc give[.imp.sg.m]=obj.1sg
oː=buːn
eːn baruːk
i-ndi
3sg.m-say.ipfv dm 2sg.m.nom def.sg.m.acc=coffee
gʷʔa-ti=eːb
hadiːd
drink-aor.2sg.m=rel.m until
‘Leave your camel with me, he says, until you have drunk your coffee!’
hadiːd can also be used as a postposition after a noun, in which case it can be
abbreviated to had: faʤil-had ‘until morning’.
3.4 Lexicon
The study of the Beja lexicon lacks research on the adaptation of Arabic loanwords and their chronological layers. There are no statistics on the proportions
of lexical items borrowed from Arabic or Ethio-Semitic as compared to those inherited from Cushitic, not to mention Afro-Asiatic as a whole or borrowed from
Nilo-Saharan. Phonetic and morphological changes are bound to have blurred
the etymological data, but what is certain is that massive lexical borrowings from
Arabic for all word categories took place at different periods of time, and that the
process is still going on. Lexicostatistical studies (Cohen 1988: 267; Blažek 1997)
have shown that Beja shares only 20% of basic vocabulary with its closest relatives, Afar, Saho and Agaw.
In this section I mainly concentrate on verbs, because they are often believed
to be less easily borrowed in language contact situations (see Wohlgemuth 2009
for an overview of the literature on this topic), which obviously is not the case
for Beja.
Cohen (1988) mentions that tri-consonantal V1s contain a majority of Semitic
borrowings. I conducted a search of Reinisch’s (1895) dictionary, the only one to
433
Martine Vanhove
mention possible correspondences with Semitic languages. It provided a total of
225 V1s, out of which only nine have no Semitic cognates (four are cognates with
Cushitic, one is borrowed from Nubian, and one cognate with Egyptian). Even if
some of Reinisch’s comparisons are dubious, the overall picture is still in favour
of massive borrowings from Semitic (96%). It is not easy to disentangle whether
the source is an Ethio-Semitic language or Arabic, but until a more detailed study
can be undertaken, the following can be said: 55 verbs (20%) have cognates only
in Ethio-Semitic (Tigre, Tigrinya, Amharic, and/or Ge’ez); out of the remaining
161 (72%), 85 are attested only in Arabic, 76 also in Ethio-Semitic. Because of the
long-standing contact with Arabic for a large majority of Beja speakers in Sudan, and the marginality of contact with Tigre limited to the south of the Beja
domain, it is tempting to assume that almost 3/4 of the 76 verbs are of Arabic
origin. They may have been borrowed at an unknown time when the new suffix conjugation was still marginal. However, there are also tri-consonantal verbs
(V2) which are conjugated with suffixes, albeit less numerous: 164. 141 have cognates with Semitic languages (95 Arabic, 31 Ethio-Semitic, and 15 attested in both
branches), six are pan-Cushitic, one is pan-Afro-Asiatic, one Nubian, six are of
dubious origin, and nine occur only in Beja. Does this mean that these borrowings occurred later than for V1s? In the current state of our knowledge of the
historical development of Beja, it is not possible to answer this question.
On the other hand, Cohen (1988: 256), in his count of consonants per stem
in eight Cushitic and Omotic languages, showed that biconsonantal stems are
predominant in six of the languages. By contrast, they form 52.8% of the 770 Beja
stems in Roper’s (1928) lexicon, and 42.7% of the 611 Agaw stems, almost on a par
with bi-consonantal stems (42.2%). What this shows is that massive borrowings
from Arabic (or from Ethio-Semitic for Agaw) helped to preserve tri-consonantal
stems, which still form a majority of the stems in Beja, unlike in other Cushitic
languages.
4 Conclusion
This overview has shown that massive lexical borrowings from Arabic in Beja
have helped to significantly entrench non-concatenative morphology in this language. Whether this is a preservation of an old Cushitic system, or a more important development of this structure than in other Cushitic languages under
the influence of Arabic, is open to debate, but what is certain is that it is not
incidental that this system is so pervasive in Beja, the only Cushitic language
to have had a long history of intense language contact with Arabic, the Semitic
434
19 Beja
language where non-concatenative morphology is the most developed. What is
important to recall is that Beja non-concatenative morphology shows no borrowings of Arabic patterns (unlike in Modern South Arabian languages; see Bettega
and Gasparini, this volume), leading to the conclusion that we are dealing with a
convergence phenomenon. Lexical borrowings and morphological convergence
are not paralleled in the phonological and syntactic domains where Arabic influence seems marginal.
Much remains to be done concerning language contact between Beja and Arabic, and we lack reliable sociolinguistic studies in this domain. We also lack a
comprehensive historical investigation of the Beja lexicon, as well as a sufficiently elaborated theory of phonetic correspondences for Cushitic (Cohen 1988:
267). Even though important progress has been made, in particular for Beja in
the comparison of its consonant system with other Cushitic languages and concerning the etymology of lexical items in some semantic fields, thanks to Blažek
(2000; 2003a; 2003b; 2006a; 2006b), the absence of a theory of lexical borrowings
in Beja (and other Cushitic languages) is still an impediment for a major breakthrough in the understanding of language contact between Beja and Arabic.
Further reading
) Apart from Vanhove (2012) on non-concatenative morphology already summarized in §3.2, we lack studies on contact-induced changes in Beja.
) Vanhove (2003) is a brief article on code-switching in one tale and two jokes
based on conversational analysis.
) Wedekind (2012) is an appraisal of the changing sociolinguistic situation of
Beja in Egypt, Sudan and Eritrea.
Acknowledgements
My thanks are due to my Sudanese consultants and collaborators, in particular Ahmed Abdallah Mohamed-Tahir and his family in Sinkat, Mohamed-Tahir
Hamid Ahmed in Khartoum, Yacine Ahmed Hamid and his family who hosted
me in Khartoum. I am grateful to the two editors of this volume, and wish to acknowledge the financial support of LLACAN, the ANR projects CorpAfroAs and
CorTypo (principal investigator Amina Mettouchi). This work was also partially
supported by a public grant overseen by the French National Research Agency
(ANR) as part of the program “Investissements d’Avenir” (reference: ANR-10LABX-0083). It contributes to the IdEx Université de Paris – ANR-18-IDEX-0001.
435
Martine Vanhove
Abbreviations
ac
acc
aor
BCE
caus
CE
com
coord
csl
cvb
def
dm
f
gen
gnrl
imp
indf
int
ipfv
action noun, maṣdar
accusative
aorist
before Common Era
causative
Common Era
comitative
coordination
causal
converb
definite
discourse marker
feminine
genitive
general
imperative
indefinite
intensive
imperfective
loc
m
mid
mnr
neg
nom
obj
opt
pass
pfv
pl
poss
prox
recp
rel
sg
smlt
TAM
locative
masculine
middle
manner
negation
nominative
object
optative
passive
perfective
plural
possessive
proximal
reciprocal
relator
singular
simultaneity
tense–aspect–mood
References
Appleyard, David. 2007. Beja morphology. In Alan S. Kaye (ed.), Morphologies of
Asia and Africa, 447–480. Winona Lake, IN: Eisenbrauns.
Bergman, Elizabeth. 2002. Spoken Sudanese Arabic: Grammar, dialogues, and
glossary. Springfield, VA: Dunwoody Press.
Blachère, Régis & Maurice Gaudefroy-Demombynes. 1975. Grammaire de l’arabe
classique. 3rd edn. Paris: Maisonneuve et Larose.
Blažek, Václav. 1997. Cushitic lexicostatistics: The second attempt. In Afroasiatica
italiana, 171–188. Napoli: Instituto Universitario Orientale.
Blažek, Václav. 2003a. Beja kinship and social terminology. In Monika R. M.
Hasitzka-Johannes & Diethart-Günther Dembski (eds.), Das alte Ägypten und
seine Nachbarn: Festschrift zum 65. Geburtstag von Helmut Satzinger, 307–340.
Krems: Österreichisches Literaturforum.
436
19 Beja
Blažek, Václav. 2003b. Fauna in Beja lexicon: A fragment of a comparativeetymological dictionary of Beja. In Leonid Kogan & Alexander Militarev (eds.),
Studia Semitica: Festschrift for Alexander Militarev, 230–294. Moscow: Russian
State University for Humanities.
Blažek, Václav. 2006a. Natural phenomena, time and geographical terminology
in Beja lexicon: fragment of a comparative and etymological dictionary of Beja:
I. Babel und Bibel (Memoriae Igor M. Diakonoff) 2. 365–407.
Blažek, Václav. 2006b. Natural phenomena, time and geographical terminology in
Beja lexicon: Fragment of a comparative and etymological dictionary of Beja:
II. Babel und Bibel 3. 383–428.
Blažek, Václav. 2007. Beja historical phonology: Consonantism. In Azeb Amha,
Graziano Savà & Maarten Mous (eds.), Omotic and Cushitic language studies:
Papers from the Fourth Cushitic Omotic Conference, Leiden, 10–12 April 2003,
125–156. Cologne: Rüdiger Köppe.
Blažek, Václav. 2000. Fragment of a comparative and etymological dictionary of
Beja: Anatomical lexicon. Unpublished manuscript.
Browne, Gerald M. 2003. Textus Blemmycus: Aetatis Christianae. Champaign, IL:
Stipes.
Cohen, David. 1984. La phrase nominale et l’évolution du système verbal en
sémitique: Etude de syntaxe historique. Leuven: Peeters.
Cohen, David. 1988. Bédja. In David Cohen (ed.), Les langues dans le monde ancien
et moderne, vol. 3: Langues chamito-sémitiques, 270–277. Paris: Editions du
CNRS.
El-Sayed, Rafed. 2011. Afrikanischstämmiger Lehnwortschatz im älteren Ägyptisch:
Untersuchungen zur ägyptisch-afrikanischen lexikalischen Interferenz im dritten
und zweiten Jahrtausend v. Chr. Leuven: Peeters.
Hamid Ahmed, Mohamed-Tahir. 2005a. “Paroles d’hommes honorables”: Essai
d’anthropologie poétique des Bedja du Soudan. Leuven: Peeters.
Hamid Ahmed, Mohamed-Tahir. 2005b. Ethics and oral poetry in Beja society
(Sudan). In Catherine Miller (ed.), Land, ethnicity and political legitimacy in
eastern Sudan: Kassala and Gedaref States, 475–504. Lawrenceville, NJ: Africa
World Press/The Red Sea Press.
Hudson, Richard A. 1976. Beja. In Lionel Bender (ed.), The non-Semitic languages
of Ethiopia, 97–132. East Lansing, MI: African Studies Center, Southern Illinois
University.
Manfredi, Stefano, Marie-Claude Simeone-Senelle & Mauro Tosco. 2015.
Language contact, borrowing and codeswitching. In Amina Mettouchi,
Martine Vanhove & Dominique Caubet (eds.), Corpus-based studies of lesser-
437
Martine Vanhove
described languages: The CorpAfroAs corpus of spoken AfroAsiatic languages,
283–308. Amsterdam: John Benjamins.
Morin, Didier. 1995. “Des paroles douces comme la soie”: Introduction aux contes
dans l’aire couchitique (bedja, afar, saho, somali). Leuven: Peeters.
Onour, ʕAbdallah. 2015. ʔasbāb al-ʔummiyya ʕind al-Biǧā: dirāsāt ḥāla, wilāyat
Kasala, maḥalliyyat Šamāl ad-Daltā, manṭiqat Wagar [The reasons for illiteracy
among the Bejas: Case study, Kassala State, Shamal al-Dalta region, Wagar
locality]. Khartoum: University of Khartoum. (Doctoral dissertation).
Reinisch, Leo. 1893. Die Beḍauye-Sprache in Nordost-Afrika. Vol. 1–2. Vienna:
Buchhändler der kaiserlichen Akademie der Wissenschaften.
Reinisch, Leo. 1895. Wörterbuch der Beḍawiye-Sprache. Vienna: Alfred Hölder.
Rilly, Claude. 2014. Language and ethnicity in ancient Sudan. In Julie R. Anderson
& Derek A. Welsby (eds.), The Fourth Cataract and beyond: Proceedings of the
12th International Conference for Nubian Studies, 1169–1188. Leuven: Peeters.
Rilly, Claude. 2018. Languages of Ancient Nubia. In Dietrich Raue (ed.), Handbook
of Ancient Nubia, 129–152. Berlin: De Gruyter.
Roper, E. M. 1928. Tu Beḍawiɛ: An elementary handbook for the use of Sudan
government officials. Hertford: Stephen Austin.
Simeone-Senelle, Marie-Claude. 2011. Modern South Arabian. In Stefan
Weninger, Geoffrey Khan, Michael P. Streck & Janet C. E. Watson (eds.), The
Semitic languages: An international handbook, 1073–1113. Berlin: De Gruyter
Mouton.
Ullendorf, Edward. 1955. The Semitic languages of Ethiopia: A comparative
phonology. London: Taylor’s.
Van Coetsem, Frans. 1988. Loan phonology and the two transfer types in language
contact. Dordrecht: Foris.
Van Coetsem, Frans. 2000. A general and unified theory of the transmission process
in language contact. Heidelberg: Winter.
Vanhove, Martine. 2003. Bilinguisme et alternance bedja-arabe au Soudan. In
Ignacio Ferrando & Juan José Sánchez Sandoval (eds.), AIDA 5th conference proceedings, Cádiz, September 2002, 131–142. Cádiz: Universidad de Cádiz, Servicio
de Publicaciones.
Vanhove, Martine. 2012. Roots and patterns in Beja (Cushitic): The issue of
language contact with Arabic. In Martine Vanhove, Thomas Stolz, Hitmoi
Otsuka & Aina Urdze (eds.), Morphologies in contact, 311–326. Berlin: Akademie
Verlag.
Vanhove, Martine. 2016. The Manner converb in Beja (Cushitic) and its
refinitization. In Claudine Chamoreau & Zarina Estrada-Fernández (eds.),
Finiteness and nominalization, 323–344. Amsterdam: John Benjamins.
438
19 Beja
Vanhove, Martine. 2017. Le bedja. Leuven: Peeters.
Vanhove, Martine & Mohamed-Tahir Hamid Ahmed. 2011. Le bedja. In Emilio
Bonvini, Joëlle Bussutil & Alain Peyraube (eds.), Dictionnaire des langues, 362–
368. Paris: PUF.
Wedekind, Klaus. 2012. Sociolinguistic developments affecting Beja dialects. In
Matthias Brenzinger & Anne-Maria Fehn (eds.), Proceedings of the 6th World
Congress of African Linguistics, Cologne, 17–21 August 2009, 623–633. Cologne:
Rüdiger Köppe.
Wohlgemuth, Jan. 2009. A typology of verbal borrowings. Berlin: Mouton de
Gruyter.
Zaborski, Andrzej. 1986. The morphology of nominal plural in the Cushitic
languages. Vienna: Afro-pub.
Záhořík, Jan. 2007. The islamization of the Beja until the 19th century. In Marc
Seifert, Markus Egert, Fabian Heerbaart, Kathrin Kolossa, Mareike Limanski,
Meikal Mumin, Peter André Rodekuhr, Susanne Rous, Sylvia Stankowski &
Marilena Thanassoula (eds.), Beiträge zur 1. Kölner Afrikawissenschaftlichen
Nachwuchstagung (KANT I). Cologne: Institut für Afrikanistik, Universität zu
Köln.
Zibelius-Chen, Karola. 2014. Sprachen Nubiens in pharaonischer Zeit. Lingua
Aegyptia: Journal of Egyptian Language Studies 22. 267–309.
439
Chapter 20
Iranian languages
Dénes Gazsi
Iranian languages, spoken from Turkey to Chinese Turkestan, have been in language contact with Arabic since pre-Islamic times. Arabic as a source language has
provided phonological and morphological elements, as well as a plethora of lexical
items, to numerous Iranian languages under recipient-language agentivity. New
Persian, the most significant member of this group, has been a prominent recipient of Arabic language elements. This study provides an overview of the historical
development of this contact, before analyzing Arabic elements in New Persian and
other New Iranian languages. It also discusses how Arabic has influenced Modern
Persian dialects, and how Persian vernaculars in the Persian Gulf region of Iran
have incorporated Arabic lexemes from Gulf Arabic dialects.
1 Current state and historical development
1.1 Iranian languages
Iranian languages, along with Indo-Aryan and Nuristani languages, constitute
the group of Indo-Iranian languages, which is a sizeable branch of the IndoEuropean language family. The term “Iranian language” has historically been
applied to any language that descended from a proto-Iranian parent language
spoken in Asia in the late third to early second millennium BCE (Skjærvø 2012).
Iranian languages are known from three chronological stages: Old, Middle, and
New Iranian. Persian is the only language attested in all three historical stages.
New Persian, originally spoken in Fārs province, descended from Middle Persian,
the language of the Sasanian Empire (third–seventh centuries CE), which is the
progeny of Old Persian, the language of the Achaemenid Empire (sixth–fourth
centuries BCE). New Persian is divided into Early Classical (ninth–twelfth centuries CE), Classical (thirteenth–nineteenth centuries) and Modern Persian (from
the nineteenth century onward), the latter considered to be based on the dialect
of Tehran (Jeremiás 2004: 427).
Dénes Gazsi. 2020. Iranian languages. In Christopher Lucas & Stefano Manfredi
(eds.), Arabic and contact-induced change, 441–457. Berlin: Language Science Press.
DOI:10.5281/zenodo.3744539
Dénes Gazsi
Today, Iranian languages are spoken from the Caucasus, Turkey, and Iraq in
the west to Pakistan and Chinese Turkestan in the east, as well as in a large
diaspora in Europe and the Americas. New Iranian languages are divided into
two main groups: Western and Eastern Iranian languages. The focus of this study
is New Persian, the most significant member among Iranian languages, but a
brief overview of Arabic influence on other New Iranian languages will also be
provided. Below is a list of the most important members and their geographical
distribution (Schmitt 1989: 246).
1.1.1 Western Iranian languages
1.1.1.1 Southwestern group
Persian (Fārsī) (spoken throughout Iran and adjacent areas), Tajik (the variety of
New Persian in Central Asia), Darī Persian (Afghanistan), Kumzārī (Musandam
Peninsula). Persian dialects in this group include Dizfūlī (Khuzestan province),
Lurī (ethnic group along the Zagros mountain range), Baḫtiārī (nomadic tribe in
the Zagros mountains), Fārs dialects (Fārs province), Lāristānī dialects (Lāristān
region of Fārs province), Bandarī (dialects spoken around Bandar ʕAbbās in the
Persian Gulf region, to which Fīnī also belongs).
1.1.1.2 Northwestern group
Kurdish, Zazaki (in eastern Turkey), Gurānī (in eastern Iraq and western Iran),
Balūčī (Balochi, spoken chiefly in Iranian and Pakistani Baluchistan, and parts of
Oman). Non-literary languages and dialects: Tātī, Tālišī and Gīlakī (on the shores
of the Caspian Sea), Central dialects (spoken in a vast area between Hamadān,
Kāšān and Iṣfahān), Kirmānī (south of the Dašt-i Kawīr).
1.1.2 Eastern Iranian languages
1.1.2.1 Southeastern group
Pashto (Afghanistan, Pakistan, eastern border region of Iran), Pamir languages
(Pamir Mountains along the Pānj River).
1.1.2.2 Northeastern group
Yaɣnōbi (Zarafšān region of Tajikistan), Ossetic (central Caucasus).
442
20 Iranian languages
1.2 Historical development of Arabic–Persian language contact
Language contact between Arabic and Persian has been a reciprocal process for
the past 1500 years. During the pre-Islamic and early Islamic era (sixth–seventh
centuries CE), Middle Persian, being embedded in the well-established and sophisticated Iranian culture, provided many loanwords to pre-Classical and Classical Arabic (Gazsi 2011: 1015; see also van Putten, this volume) under RL (recipientlanguage) agentivity (Van Coetsem 1988; 2000). With the collapse of the Sasanian
Empire and expansion of Islam and the Arabic language over vast territories outside Arabia, Classical Arabic began to exercise an unprecedented impact on the
emerging New Persian language. Arabic never took root in the everyday communication of the ethnically Persian population, although it gained some dominance
as a written vehicle in the administrative, theological, literary and scientific domains in the eastern periphery of the Abbasid Caliphate. Instead, spoken Middle
Persian (Darī ) flourished as a vernacular language. In the middle of the ninth
century CE, it was in this part of Iran, specifically in Fārs province, that Darī
emerged in a new form as it repositioned itself in the culture and literature of
the local populace. This new literary language, the revitalization of the Persian
linguistic heritage, would be called New Persian. Since its earliest phase, New Persian has borrowed a staggering number of loanwords. Initially, these loanwords
were borrowed from various northwestern and eastern Iranian languages, such
as Parthian and Sogdian. Despite this relatively large group of loans, the most
versatile lenders were the Arabs. Whereas in the pre-Islamic era Arabic had almost exclusively taken lexical items from Middle Persian (in the fields of religion,
botany, science and bureaucracy among others), New Persian also incorporated
Arabic morphosyntactic elements.
The first Arabic loanwords began to permeate New Persian in the ninth–tenth
centuries CE (20–30%). This process was not even diminished by the Iranian
šuʕūbiyya movement, the major output of which was all conducted in Arabic. In
subsequent centuries, Persian continued to absorb an ever-expanding set of Arabic lexemes. By the turn of the twelfth century, the proportion of Arabic loans
increased to approximately 50%. The majority of Arabic loans had already been
integrated into New Persian by that time and have shown a remarkable steadiness until recently.
After the fall of Baghdad in 1258 CE, Arabic lost its foothold in the eastern
provinces of the Caliphate, thereby drawing the final boundary between the
use of Arabic and Persian (Danner 2011). The Mongol Ilkhānids, who as nonMuslims were not dependent on Arabic, introduced Persian as the language of
education and administration in Iran and Anatolia. Despite the significant de-
443
Dénes Gazsi
struction the Mongols caused to northern Iran during their conquest, this period
(thirteenth and fourteenth centuries CE) is considered to be the zenith of Persian
literature. This is also the epoch when literary Persian is, in an excessive way,
inundated with Arabic language elements. This phenomenon is easily detectable
in the works of one of the most significant personalities in Classical Persian literature, and a pre-eminent poet of thirteenth-century Persia, Saʕdī of Shiraz. Following the norms of Persian prose writing and poetry of his time, Saʕdī flooded
his writings with a bewildering array of Arabic language elements. To illustrate
this, here is a typical sentence from Saʕdī’s Gulistān ‘Rose Garden’ (completed in
1258 CE), where words of Arabic origin are highlighted in boldface (Yūsifī 2004:
77).
(1)
به جانب ما التفات كند در رعايت خاطرش هرچه، أحسن الله خلاصه، اگر راى عزيز فلان
تمامترسعى كرده شود واعيان اين مملكت به ديدار او مفتقرند و جواب اين حروف را منتظر
agar rāy-i ʕazīz-i fulān, aḥsana allāhu ḫalāṣahu, ba ǧānib-i mā iltifāt
kunad dar riʕāyat-i ḫāṭiraš har či tamāmtar saʕī karda šawad wa
aʕyān-i īn mamlakat ba dīdār-i ū muftaqirand wa ǧawāb-i īn ḥurūf rā
muntaẓir.
‘If the precious mind of that person, may God make the end of his affairs
prosperous, were to look in our direction, the utmost efforts would be
made to please him, because the nobles of this realm would consider it an
honor to see him, and are waiting for a reply to this letter.’1
It is easy to ascertain that, apart from verbs and adverbs, almost every lexical
item in the sentence is of Arabic origin. But writers of this era, such as Saʕdī,
not only inundated their works with Arabic elements, but even used Arabic morphology and semantics freely by coining new and innovative meanings, e.g. ṣaʕqa
‘lightning’ < MSA/MSP ṣāʕiqa or baṭṭāl ‘liar’ < MSA/MSP ‘inactive, unemployed
person’,2 < MSA mubṭil ‘liar’. The Persian and Arabic language use of Saʕdī and
other literary figures in the Classical Persian period came closest to what Lucas
(2015) calls convergence under the language dominance principle. As reflected in
the purely Arabic and Arabic-infused Persian segments of his oeuvre, Saʕdī was
equally dominant in both Classical Arabic and Classical Persian along with the
dialect of Shiraz.
1
Persian transcription in this chapter follows the Arabic phonological conventions to avoid
using two disparate systems.
2
In this chapter, references are made to Modern Standard Arabic (MSA) and Modern Standard
Persian (MSP) as a comparison to dialectal forms in both languages. This seemed more straightforward as it is not always feasible to determine at what point in time a lexeme was borrowed
from Arabic into Persian.
444
20 Iranian languages
Modern Persian is still deeply rooted in Arabic. Arabic loanwords constitute
more than 50% of its vocabulary, but in elevated styles (religious, scientific, literary) Arabic loans may exceed 80% (Jeremiás 2011). Although the proportion of
these loanwords fluctuates according to age, genre, social context or idiolect, any
style in Modern Persian deprived of Arabic influence is almost impossible. An endeavor similar to Atatürk’s to purge Turkish of foreign language elements would
be unrealistic in Modern Persian, even with recurring efforts by linguistic purists
and the Academy of Persian Language and Literature (Farhangistān-i zabān wa
adab-i fārsī ).3 It is noteworthy that when the need arose for new terminology
to describe fledgling political concepts in Iran, for instance during the Constitutional Revolution in the early twentieth century, as Elwell-Sutton (2011) phrased
it, “politicians and journalists instinctively turned to Arabic rather than Persian”.
Frequently, however, these “Arabic” words were new coinages in the recipient
language, e.g. mašrūṭa ‘constitution’, mawqiʕiyyat ‘situation, position’. After the
Islamic Revolution in 1979, another wave of Arabic lexemes related to the new
religious governing system surfaced, e.g. mustaẓʕifīn ‘the needy, the enfeebled’
(< MSA mustaḍʕafūna/mustaḍʕafīna).
Primary and secondary schools in contemporary Arabic-speaking countries
do not offer language education in Persian. In Iran, compulsory Classical Arabic
instruction is part of the curriculum. However, the language is taught for religious purposes only, with no intention to utilize MSA as a means of acquiring
communication skills.
2 Contact languages
This section briefly describes the linguistic impact of Standard Arabic on several
New Iranian languages. A more detailed analysis of contact-induced language
change in New Persian will follow in §3.
2.1 Arabic influence on New Iranian languages
2.1.1 Tajik (Tōǧīkī)
Tajik, written in a modified Cyrillic script, is the variety of New Persian spoken throughout Central Asia, most notably in Tajikistan, Uzbekistan, and northern Afghanistan. Similarly to all varieties of Persian, Arabic borrowings constitute the earliest layer of foreign vocabulary in Tajik (Perry 2009). This lex3
An example of their activity is the publication (by Rāzī 2004) of a dictionary that lists “pure”
Persian words.
445
Dénes Gazsi
icon was transferred under RL agentivity. Although Arabic lexical items have
a firm hold in Tajik, their pattern of distribution differs from that of New Persian. For instance, Tajik uses pēš ‘before’ and pas ‘after’ rather than MSA/MSP
qabl and baʕd, but ōid ba-/ōid-i (< MSA ʕāʔid ‘returning’) ‘concerning, relating
to’ in lieu of MSP rāǧiʕ ba- (< MSA rāǧiʕ ‘recurring’). Also, madaniyyat ‘civilization’ (< MSA madaniyya ‘civilization’; cf. MSA/MSP tamaddun ‘civilization’),
hōzir ‘now’ (< MSA ḥāḍir ‘present; ready’, MSP ḥāẓir ‘present’), ittifōq ‘(labor)
union’ (< MSA ittifāq ‘agreement; contract’; cf. MSP ittiḥād ‘[labor] union’).
Arabic plural forms, both sound feminine plural and broken plural, were lexicalized with collective or singular meanings: hašarōt ‘insect’, with regular plural
ending hašarōthō ‘insects’ (< MSA/MSP ḥašarāt ‘insects’), talaba ‘student’, pl. talabagōn (< MSA/MSP ṭalaba ‘students’), šarōit ‘condition, stipulation’ (< MSA/
MSP šarāʔiṭ ‘conditions’).
2.1.2 Kurdish
A characteristic feature of Kurdish, the change of postvocalic /m/ > /v/ or /w/,
also occurs frequently in words of Arabic origin: silāv ‘greeting’ (< MSA/MSP
salām; Paul 2008).
2.1.3 Gurānī
The phonological system of Gurānī dialects is similar to Kurdish in the occurrence of Arabic pharyngeal and emphatic sounds /ʕ/, /ḥ/, /ṣ/ (MacKenzie 2012).
2.1.4 Ossetic
Ossetic has incorporated terms related to Islam from Arabic and Persian through
neighboring Caucasian languages (Thordarson 2009).
2.2 Arabic-speaking communities in Iran
Arabic-speaking communities are known to be present within the boundaries
of the Islamic Republic of Iran, but their exact number is not readily discernible
from official statistics. It is estimated that 3% of Iran’s 80 million citizens are
Arabs, which would put the Arab population at approximately 2.5 million. The
majority of Arabs live in the western parts of Khuzestan province (see Leitner,
this volume), but also along Iran’s Persian Gulf coast and parts of Khorasan in
eastern Iran (Oberling 2011). Already during the Sasanian era, several Arab tribes,
including the Bakr ibn Wāʔil and Banū Tamīm, settled in the area stretching from
446
20 Iranian languages
the Šaṭṭ al-ʕArab to the Zagros Mountains (Daniel 2011). At the end of the sixteenth century, the Banū Kaʕb, originating from present-day Kuwait, settled in
Khuzestan. During subsequent centuries, more Arab tribes moved from southern Iraq to the province. As a result, Khuzestan, which until 1925 was called
ʕArabistān, became extensively Arabized. Members of these Arab tribes live on
either side of the Iran–Iraq border. In the same way as Iraqi Arabic vernaculars,
Khuzestan Arabic has been influenced by Persian. However, Khuzestan Arabic
can most easily be distinguished from Iraqi dialects by its wide-ranging transfer
of Persian lexemes (Ingham 1997: 25; see also Leitner, this volume).
Arab presence has a well-documented history on the Iranian coastline of the
Persian Gulf, in what now constitutes Būšihr and Hurmuzgān provinces. According to travelogues from the eighteenth to the twentieth centuries CE, as well as
British archival materials dating back to the British Residency of the Persian Gulf,
Arab tribes inhabited most fishing and pearling villages, as well as islands and
coastal towns with strategic importance (e.g. Bandar ʕAbbās). The most significant tribes in this area were, and in some cases still are, the Qawāsim, Marāzīq,
Āl Ḥaram, Āl ʕAlī, Āl Naṣūr, Banī Tamīm, Banī Ḥammād, Banī Bišr, among others. In contrast to most Persians and Khuzestani Arabs who are primarily Shiite,
these tribes are Sunni Muslims. A widespread exonym to designate Arabs on the
Iranian coast, but shunned by the local population, is hōla (variously referred to
as hula, huwala or hawala). Local tribes prefer the endonym ‘Arabs of the Coast’
(ʕarab as-sāḥil) (Gazsi 2017: 110).
Most Khuzestani and Iranian Persian Gulf Arabs are bilingual, speaking Arabic
as their mother tongue and Persian as a second language. Although Khuzestan
and the two Persian Gulf provinces are geographically part of Iran, linguistically
their Arab populations form a continuum with the southern Mesopotamian Muslim gilit dialects, and the dialects of the eastern coast of the Arabian Peninsula,
respectively. In the spoken and written code, ‘Arabs of the Coast’ often engage in
tetra-glossic switching between MSA, Gulf Arabic (GA), MSP, Colloquial Persian
and one of its local dialects such as Bandarī. In their speech, Persian phonological
and lexical elements are borrowed into GA under RL agentivity.
3 Contact-induced changes in New Persian and modern
Persian dialects
Language contact between Arabic and New Persian is most evidently detectable
in the New Persian lexicon, and to a lesser extent in phonology and morphosyntax. This section summarizes the characteristics of this contact. In addition to
447
Dénes Gazsi
standard New Persian, and its contemporary variant MSP, Arabic has also influenced modern Persian dialects. This influence is slightly different, and in several
ways more far-reaching, particularly in the realm of phonology and lexicon.
Persian dialects developed separately from and parallel to Classical Persian
and MSP. Modern Persian dialects retain several Early Classical and Classical
Persian phonological and morphosyntactic features that are not present in MSP.
Additionally, they were in direct contact with the Arabic language through Arab
tribes that settled across Persia immediately after the Islamic conquest or in later
centuries. Although most Arab tribes have long been integrated into the Persianspeaking population, the Arabic language in the areas currently dominated by
ethnic Arabs is still in contact with the surrounding Persian dialects. Unlike Arabic influence on the standard version of New Persian, Arabic influence on modern Persian dialects is an understudied field that does not allow for providing
an exhaustive list of contact-induced changes at this point. Instead, below is a
preliminary description of salient examples of how Arabic phonological and lexical elements were transferred to New Persian, both its standard and dialectal
variations.
3.1 Phonology
3.1.1 New Persian
The initial step in the adoption of Arabic lexemes was the adoption of the Arabic
script. New Persian began to use a modified Arabic script in the ninth century CE;
it has 32 letters, 28 acquired from Arabic and 4 new letters added to represent
Persian phonemes (/p/, /č/, /ž/, /g/). Arabic /θ/ and /ṣ/ collapse to Persian /s/,
Arabic /ð/, /ḍ/, /ð̣/ collapse to Persian /z/, and Arabic /ṭ/ becomes Persian /t/. The
phonemic inventory of Early Classical Persian was augmented with the glottal
stop, which originated in the two separate Arabic phonemes /ʔ/ and /ʕ/.
3.1.2 Modern Persian dialects
This section highlights phonological features of modern Persian dialects that
were the result of contact-induced language change under RL agentivity, either
with Arabic or with Classical Persian, and subsequently MSP.
3.1.2.1 Adoption of Arabic pharyngeal sounds
The two Arabic pharyngeal sounds undergo phonological integration in New
Persian: the voiceless pharyngeal fricative /ḥ/ is pronounced as a voiceless glottal fricative /h/, and the voiced pharyngeal fricative /ʕ/ as a glottal stop /ʔ/. The
448
20 Iranian languages
dialects of Dizfūl and Šūštar have acquired pharyngeal sounds directly from Arabic, which occur in Arabic loanwords: ʕaǧīb ‘strange’, baʕd ‘after’ (MacKinnon
2015). The dialect of Jarkūya shares this feature: ḥasüd ‘jealous’, ǧimʕa ‘Friday’
(Borjian 2008).
The dialect of Kulāb in Tajikistan also borrows Arabic pharyngeal sounds in
words of Arabic origin: ʕaib ‘flaw’, daʕvō ‘claim’, mıʕalim ‘teacher’, ḥıkımat ‘wisdom’, sōḥib ‘owner’. Arabic pharyngeal sounds also occur in a few Persian/Tajiki
words (ʕasp ‘horse’, ḥamsōya ‘neighbor’). Interestingly, the pharyngealized form
for ‘horse’ occurs far and wide within the Iranian linguistic domain, as ʕasb in
the Lurī dialect of Šūštar, in Ḫānsāri and Caucasian Tātī. In the Arab Gulf states,
the ʕAǧam, ethnic Persians holding Kuwaiti, Emirati and other Gulf citizenship,
pronounce Arabic loanwords in their Persian speech with pharyngeal sounds.
3.1.2.2 Dropping of Arabic pharyngeal sounds
In several modern Persian dialects, the voiceless pharyngeal fricative /ḥ/ is absent. The preceding vowel is lengthened or the subsequent vowel disappears
too, e.g. mūtāǧ ‘in need, destitute’ < MSA/MSP muḥtāǧ (Īzadpanāh 2001: 190),
ṣārā ‘desert’ < MSA/MSP ṣaḥrā (Sarlak 2002: 15), ṣāb ‘owner’ < MSA/MSP ṣāḥib
(Ṣarrāfī 1996: 135), mulāẓa ‘consideration, observation’ < MSP mulāḥiẓa, cf. MSA
mulāḥað̣a (Ṣarrāfī 1996: 188), ṣul ‘peace’ < MSA/MSP ṣulḥ (Stilo 2012), ēsās ‘feeling’ < MSA/MSP iḥsās (Salāmī 2004: 160–161). In Kirmān, the sound change /uḥ/ >
/ā/ is attested, e.g. fāš ‘insult’ < MSA/MSP fuḥš (Borjian 2017).
The voiced pharyngeal fricative /ʕ/, pronounced as a glottal stop in MSP, can
also be dropped. This may result in vowel lengthening: māṭal ‘idle’ < MSA/MSP
muʕaṭṭal (Ṣarrāfī 1996: 184), māmila ‘transaction’ < NewP muʕāmila, cf. MSA
muʕāmala (Ṣarrāfī 1996: 184; Sarlak 2002: 15), rubbi sāt ‘quarter hour’ < MSP rubʕ
sāʕat, cf. MSA rubʕ sāʕa (Ṣarrāfī 1996: 108), mānī ‘meaning’ < MSP maʕnī, cf. MSA
maʕnā (Sarlak 2002: 15), mōǧiza ‘miracle’ < MSA/MSP muʕǧiza (Īzadpanāh 2001:
190), tāǧub ~ tāǧuv ‘surprise, wonder’ < MSA/MSP taʕaǧǧub (Salāmī 2004: 162–
163), rāyat ‘regard’ < MSP riʕāyat, cf. MSA riʕāya (Ṣarrāfī 1996: 107).
3.1.2.3 Dropping of the Arabic voiceless glottal fricative /h/
The voiceless glottal fricative disappears in closed syllables in many Persian dialects, resulting in occasional vowel lengthening: ṭārat ‘cleanliness’ < MSP ṭahārat,
cf. MSA ṭahāra (Sarlak 2002: 76), nāal ‘impolite’ < MSP nāahl (Īzadpanāh 2001:
192).
449
Dénes Gazsi
3.1.2.4 Miscellaneous sound changes
A range of additional consonant developments and shifts can be attested in Persian dialects. Some of these developments include:
/ʕ/ > /ḥ/:
In Lurī and the dialect of Jarkūya, a shift occurs from the voiced to the
voiceless pharyngeal: ḥilāǧ ‘cure’ < MSA/MSP ʕilāǧ (Īzadpanāh 2001: 207),
ṭaḥna ‘sarcasm’ < MSA/MSP ṭaʕna (Borjian 2008).
/ḥ/ > /ʔ/ occurring with occasional metathesis:
ṭaʔr ‘plan’ < MSA/MSP ṭarḥ (Ṣarrāfī 1996: 137), maʔla ‘city quarter’ < MSA/
MSP maḥalla (Ṣarrāfī 1996: 188), maʔala ‘city quarter’ (Naǧībī Fīnī 2002:
133).
/h/ > /ʔ/:
muʔlat ‘deadline, respite’ < MSP muhlat, cf. MSA muhla (Ṣarrāfī 1996: 190).
/θ/ > /t/:
This shift is also common in several Arabic dialects, e.g. in Egypt and
Morocco: mīrāt ‘heritage’ < MSP mīrās, cf. MSA mīrāθ (Īzadpanāh 2001:
190).
Word-final /b/ and /f/ > /m/:
naǧīm ‘noble’ < MSA/MSP naǧīb (Īzadpanāh 2001: 193), niṣm ‘half’ < MSA/
MSP niṣf (Īzadpanāh 2001: 195).
/r/ > /l/:
in Kirmān, zilar ~ zilal ‘damage, loss’ < MSP ẓarar, cf. MSA ḍarar (Ṣarrāfī
1996: 136; Dānišgar 1995: 163), ḥaṣīl ‘straw mat’ < MSA/MSP ḥaṣīr (Ṣarrāfī
1996: 85), qulfa ‘small room for summer resting’ < MSA ɣurfa ‘room’ (Fāẓilī
2004: 151).
Arabic voiceless dental emphatic /ṭ/ > /d/:
mudbaḫ ~ madbaḫ ‘kitchen’ < MSA maṭbaḫ (Ṣarrāfī 1996: 186; not attested
in MSP), mudbaq in Baḫtiārī (Sarlak 2002: 251).
/b/ > /f/:
muftilā ‘afflicted’ < MSP mubtilā, cf. MSA mubtalā (Borjian 2017).
Medial and word-final /b/ > /v/:
in Baḫtiārī, ādāv ‘customs’ < MSA/MSP ādāb (Sarlak 2002: 15), ʕajīv
450
20 Iranian languages
‘strange’ < MSA/MSP ʕajīb (Sarlak 2002: 25), qavīla ‘tribe’ < MSA/MSP
qabīla (Sarlak 2002: 199).
Word-initial /ḫ/ > /h/:
in northern Lurī and Baḫtiārī, hāla ‘aunt’ < MSA/MSP ḫāla (Īzadpanāh
2001: 204).
/q/ > /k/:
kabīla ‘tribe’ < MSA/MSP qabīla (Naǧībī Fīnī 2002: 21).
/ɣ/ > /q/:
šuql ‘occupation’ < MSP/MSA šuɣl (Stilo 2012).
/ǧ/ > /y/:
direct borrowing from Khuzestan Arabic dialects, mailis ‘council’ < MSA/
MSP maǧlis (Sarlak 2002: 260; Fāẓilī 2004: 165).
Metathesis:
qulf ‘lock’ < MSA/MSP qufl (Salāmī 2004: 84–85; Imām Ahwāzī 2000: 146),
ṣuḥb ‘morning’ < MSA/MSP ṣubḥ (Dānišgar 1995: 161; Naǧībī Fīnī 2002: 23).
The full /t/ of the tāʔ marbūṭa appears on words where it is absent in MSP:
ḥalmat ‘attack’ < MSA/MSP ḥamla (Īzadpanāh 2001: 207), ḥaǧāmat ‘cupping’ <
MSA/MSP ḥaǧāma (Salāmī 2004: 92–93). This was a typical feature of Classical
Persian literature.
3.2 Morphosyntax
Several Arabic morphosyntactic features were transferred to New Persian in the
realm of nominal morphology under RL agentivity. These features encompass
sound and broken plural forms (musāfirīn ‘passengers’, tablīɣāt ‘propaganda’,
dihāt ‘villages’, ḥuqūq ‘rights’), possessive constructions (fāriɣ ut-taḥṣīl ‘graduate’, wāǧib ul-iǧrā ‘peremptory’) and occasional gender agreement in lexicalized
expressions (quwwa-yi darrāka ‘perceptive power’). Word formation has been
an active method of transferring Arabic lexical elements into New Persian from
early on, either by way of derivation (diḫālat ‘interference’ < MSA mudāḫala,
awlā-tar ‘superior’ < MSA awlā, raqṣīdan ‘to dance’ < MSA raqṣ, aks̱aran ‘most,
generally’ < MSA akθar ‘more, most’) or compounding. Compounding is a highly
developed process of enlarging the New Persian vocabulary. It is manifest in
lexical compounds (taɣẕia-šinās ‘nutritionist’, ḫiānat-kārāna ‘perfidiously’) and
phrasal compounds (iṭāʕat kardan ‘to obey’, ʕadam-i wuǧūd ‘non-existence’, ʕala
l-ḫuṣūṣ ‘particularly’).
451
Dénes Gazsi
3.3 Lexicon
3.3.1 Arabic lexicon in New Persian
Contact-induced language change manifests itself most strikingly in the lexicon
transferred from Arabic to New Persian under RL agentivity. The earliest loanwords entered New Persian during the ninth–tenth centuries. This process occurred smoothly, as the phonological inventory of Early Classical Persian was
likely close to that of Middle Persian and also close to that of Classical Arabic.4
The influx of Arabic loanwords has unabatedly continued over the centuries until now. To showcase a recent example of Arabic vocabulary in Modern Persian,
below are titles of articles from Hamšahrī ‘Fellow Citizen’, a major Iranian national newspaper, taken from its 29th January 2018 edition. Arabic words are
highlighted in boldface:
(2)
a. kulliyyāt-i lāyiḥa-yi būdǧa-yi sāl-i
97-i
kull-i
total.pl-gen bill-gen budget-gen year-gen 97-gen whole-gen
kišwar radd šud
country reject be.pst.3sg
‘The total budget bill of the year 2018 for the whole country was
rejected.’
b. daʕwat az tihrānīhā barā-yi ihdā-yi
ḫūn asāmī-yi
call
from Tehrani.pl for-gen donation-gen blood name.pl-gen
marākiz-i
faʕʕāl
center.pl-gen active
‘Calling the residents of Tehran to donate blood. Names of active
centers.’
c. iḥrāz-i
huwiyyat dar muʕāmilāt-i
milkī
authentication-gen identity in transaction.pl-gen proprietary
bā kārt-i
hūšmand-i millī
anǧām mī-šaw-ad
with card-gen smart-gen national complete prs-be-3sg
‘Personal authentication in real estate transactions is done with the
national smart card.’
In the Arabic lexicon of New Persian, further characteristics can be observed,
such as phonetic changes (NewP maʔnī ‘meaning’ < MSA maʕnā, NewP madrisa
‘school’ < MSA madrasa, NewP šikl ‘shape, form’ < SA šakl), where in some cases
4
In Early Classical Persian, short vowels were likely pronounced as /u/ and /i/, and the alif as
/ā/. In MSP, the pronunciation is /o/, /e/ and /ɒ/.
452
20 Iranian languages
the Persian pronunciation may follow Arabic dialectal forms, semantic changes
(NewP kitābat ‘writing’ and kitāba ‘inscription’ < MSA kitāba ‘writing’, NewP
ṣuḥbat ‘speech’ < MSA ṣuḥba ‘companionship’), and occasional imāla in elevated
or poetic style (NewP ḥiǧīz < MSA ḥiǧāz).
3.3.2 Arabic lexicon in Persian dialects
Arabic loanwords affect Persian dialects in two ways that differ from MSP: i)
semantic changes, where Arabic lexemes assume new meanings unattested in
both MSA and MSP: in Kirmān ðāt ‘age’ (Ṣarrāfī 1996: 106) < MSA/MSP ‘self,
soul, essence, nature’, ðātī ‘old’ < MSA/MSP ‘own, personal’; ii) lexemes and
expressions directly borrowed from Arabic, and not attested in MSP: in Šūštar,
ḥaya ‘snake’ < MSA ḥayya, MSP mār (Fāẓilī 2004: 140), ṭayyāra ‘airplane’ < Arabic dialects ṭayyāra, MSA ṭāʔira, MSP hawāpaimā (Fāẓilī 2004: 150), ṣaḥn ‘bowl,
dish’ < MSA ṣaḥn, MSP bušqāb (Fāẓilī 2004: 150), ṭabaq ‘plate, tray’ < MSA ṭabaq,
MSP sīnī (Fāẓilī 2004: 150), in Fīn, mismāl ‘nail’ < MSA mismār, MSP mīḫ ‘nail’
(Naǧībī Fīnī 2002: 133), in Kirmān, aḥad un-nās ‘nobody, somebody’ < MSA aḥad
un-nās, MSP hīčkas ‘nobody’, kasī ‘somebody’ (Ṣarrāfī 1996: 33).
On the Persian Gulf coast of Iran, due to linguistic, economic and commercial
connections with the Arabian Peninsula, Persian dialects have incorporated from
Gulf Arabic a number of Arabic technical terms relating to pearling, fishing and
traditional shipbuilding: muḥār ‘shellfish, oysters’ (cf. MSA maḥār), giyās ‘measure, gauge’ (< GA giyās, cf. MSA qiyās), mīdāf ‘helm (boat)’ (< GA mīdāf, cf.
MSA miǧdāf ), māčila ‘meal (on a boat)’ (< GA māčila, cf. MSA maʔkūl). Two
neighborhoods in the town of Bandar Linga (opposite Dubai, 180 km west of
Bandar ʕAbbās) are called Maḥalla-yi Baḥrainī ‘Bahraini Quarter’ and Maḥallayi Sammāčī ‘Fishers’ Quarter’ (< GA sammāč, cf. MSA sammāk) (Baḫtiyārī 1990:
137–138).
4 Conclusion
Although Arabic–Persian language contact has been a well-known phenomenon
for centuries, academic research dedicated to this topic is far from abundant.
Throughout the centuries, Persian writers and poets used Arabic lexical elements
in new meanings or coined non-standard Perso-Arabic lexemes based on Arabic
derivational patterns. Idiosyncratic features of individual Persian writers should
be examined separately before compiling a comprehensive review of this contactinduced language change. Substantial fieldwork needs to be conducted to describe the bilingualism of ethnic Arab communities of Iran and ethnic Persians
453
Dénes Gazsi
in Arabic-speaking countries. Additionally, it is essential for linguists to look
into Arabic influence on Modern Persian dialects and Iranian languages other
than New Persian. This will help scholars understand the scale and depth of how
Arabic has shaped Iranian languages for the past thousand years.
Contact-induced language change in New Iranian languages primarily transpired under RL agentivity. It should be noted, however, that medieval Persian
literati were so well-versed in Arabic due to its prestige and dominance, that
their bilingualism may have enabled convergence in Arabic–Persian language
contact.
Further reading
) Asbaghi (1987) gathers eight hundred Persian words of Arabic origin in twentythree groups and analyzes the semantic changes they underwent when transferred from Arabic to New Persian.
) Gazsi (2011) gives an overview of Arabic–Persian language contact from preIslamic times up to the modern era, also touching on Arabic dialects in Iran.
A brief analysis of Arabic morphosyntactic features in New Persian is also
provided.
) Ṣādeqī (2011) discusses a range of Arabic phonological, grammatical and semantic elements in New Persian.
Acknowledgements
I would like to express my gratitude to Prof. Éva Jeremiás, Prof. Werner Arnold
and Prof. Ali Ashraf Sadeqi for their support while I was working on Arabic–
Persian language contact. I am thankful to members of the ‘Arabs of the Coast’ in
the UAE, especially Sheikh Ibrahim, Sheikh Abdulrahman, Dr. Abdullah, Walid,
Ahmed, and many others for providing language data in Gulf Arabic.
Abbreviations
BCE
CE
GA
gen
MSA
MSP
454
before Common Era
Common Era
Gulf Arabic
genitive
Modern Standard Arabic
Modern Standard Persian
NewP
pl
prs
pst
RL
New Persian
plural
present
past
recipient language
20 Iranian languages
References
Asbaghi, Asya. 1987. Die semantische Entwicklung arabischer Wörter im Persischen.
Stuttgart: Steiner.
Baḫtiyārī, Maǧīd. 1990. Ustān-i hurmuzgān (rāhnamā-yi mufaṣṣal-i Īrān). Tehran:
Gītāšināsī.
Borjian, Habib. 2008. Jarquya ii: The dialect. In Ehsan Yarshater, Mohsen
Ashtiany & Mahnaz Moazami (eds.), Encyclopædia iranica, online edn. New
York: Encyclopædia Iranica Foundation.
Borjian, Habib. 2017. Kerman xvi: Languages. In Ehsan Yarshater, Mohsen
Ashtiany & Mahnaz Moazami (eds.), Encyclopædia iranica, online edn. New
York: Encyclopædia Iranica Foundation.
Daniel, Elton L. 2011. ʿArab iii: Arab settlements in Iran. In Ehsan Yarshater,
Mohsen Ashtiany & Mahnaz Moazami (eds.), Encyclopædia iranica, online edn.
New York: Encyclopædia Iranica Foundation.
Dānišgar, Aḥmad. 1995. Farhang-i wāžahā-i rāyiǧ-i turbat-i ḥaidarīya. Mašhad:
Āstān-i Quds-i Rażawī.
Danner, Victor. 2011. Arabic language iv: Arabic literature in Iran. In Ehsan
Yarshater, Mohsen Ashtiany & Mahnaz Moazami (eds.), Encyclopædia iranica,
online edn. New York: Encyclopædia Iranica Foundation.
Elwell-Sutton, L. P. 2011. Arabic language iii: Arabic influences in Persian literature. In Ehsan Yarshater, Mohsen Ashtiany & Mahnaz Moazami
(eds.), Encyclopædia iranica, online edn. New York: Encyclopædia Iranica
Foundation.
Fāẓilī, Muḥammad Taqī. 2004. Farhang-i gūyiš-i šūštarī. Tehran: Pāzīnah.
Gazsi, Dénes. 2011. Arabic–Persian language contact. In Stefan Weninger,
Geoffrey Khan, Michael P. Streck & Janet C. E. Watson (eds.), The Semitic
languages: An international handbook, 1015–1021. Berlin: De Gruyter Mouton.
Gazsi, Dénes. 2017. Language and identity among the ‘Arabs of the coast’ in Iran
and the Arab Gulf states: Social media and beyond. RiCognizioni 7. 105–129.
Imām Ahwāzī, Sayyid Muḥammad ʕAlī. 2000. Čīstānnāma-yi dizfūlī. Tehran:
Daftar-i Pažūhišhā-yi Farhangī.
Ingham, Bruce. 1997. Arabian diversions: Studies on the dialects of Arabia. Reading:
Ithaca Press.
Īzadpanāh, Ḥamīd. 2001. Farhang-i lurī. Tehran: Intišārāt-i Asāṭīr.
Jeremiás, Éva M. 2004. New Persian. In Peri Bearman, Thierry Bianquis,
C. Edmund Bosworth, Emeri Johannes van Donzel & Wolfhart P. Heinrichs
(eds.), Encyclopaedia of Islam, 2nd edn., vol. 12: Supplement, 426–448. Leiden:
Brill.
455
Dénes Gazsi
Jeremiás, Éva M. 2011. Iran. In Lutz Edzard & Rudolf de Jong (eds.), Encyclopedia
of Arabic language and linguistics, online edn. Leiden: Brill.
Lucas, Christopher. 2015. Contact-induced language change. In Claire Bowern &
Bethwyn Evans (eds.), The Routledge handbook of historical linguistics, 519–536.
London: Routledge.
MacKenzie, David N. 2012. Gurāni. In Ehsan Yarshater, Mohsen Ashtiany
& Mahnaz Moazami (eds.), Encyclopædia iranica, online edn. New York:
Encyclopædia Iranica Foundation.
MacKinnon, Colin. 2015. Dezfūlī and Šūštarī dialects. In Ehsan Yarshater, Mohsen
Ashtiany & Mahnaz Moazami (eds.), Encyclopædia iranica, online edn. New
York: Encyclopædia Iranica Foundation.
Naǧībī Fīnī, Bihǧat. 2002. Barrasī-yi gūyiš-i fīnī. Tehran: Našr-i Ās̱ār-i
Farhangistān-i Zabān wa Adab-i Fārsī.
Oberling, Pierre. 2011. ʿArab iv: Arab tribes of Iran. In Ehsan Yarshater, Mohsen
Ashtiany & Mahnaz Moazami (eds.), Encyclopædia iranica, online edn. New
York: Encyclopædia Iranica Foundation.
Paul, Ludwig. 2008. Kurdish language i: History of the Kurdish language. In
Ehsan Yarshater, Mohsen Ashtiany & Mahnaz Moazami (eds.), Encyclopædia
iranica, online edn. New York: Encyclopædia Iranica Foundation.
Perry, John R. 2009. Tajik ii: Tajik Persian. In Ehsan Yarshater, Mohsen Ashtiany
& Mahnaz Moazami (eds.), Encyclopædia iranica, online edn. New York:
Encyclopædia Iranica Foundation.
Rāzī, Farīda. 2004. Farhang-i wāžahā-yi fārsī-yi sara barā-yi wāžahā-yi ʕarabī dar
fārsī-yi muʕāṣir. Tehran: Našr-i Markaz.
Salāmī, ʕAbdulnabī. 2004. Ganǧīna-yi gūyiššināsī-yi fārs. Tehran: Našr-i Ās̱ār-i
Farhangistān-i zabān wa adab-i Fārsī.
Sarlak, Riẓā. 2002. Wāžanāma-yi gūyiš-i baḫtiārī-yi čahārlang. Tehran: Ās̱ār.
Schmitt, Rüdiger. 1989. Compendium linguarum iranicum. Wiesbaden: Reichert.
Skjærvø, Prods Oktor. 2012. Iran vi: Iranian languages and scripts. In Ehsan
Yarshater, Mohsen Ashtiany & Mahnaz Moazami (eds.), Encyclopædia iranica,
online edn. New York: Encyclopædia Iranica Foundation.
Stilo, Donald. 2012. Gīlān x: Languages. In Ehsan Yarshater, Mohsen Ashtiany &
Mahnaz Moazami (eds.), Encyclopædia iranica, online edn. New York: Encyclopædia Iranica Foundation.
Ṣādeqī, ʕAlī Ašraf. 2011. Arabic language i: Arabic elements in Persian. In Ehsan
Yarshater, Mohsen Ashtiany & Mahnaz Moazami (eds.), Encyclopædia iranica,
online edn. New York: Encyclopædia Iranica Foundation.
Ṣarrāfī, Maḥmūd. 1996. Farhang-i gūyiš-i kirmānī. Tehran: Surūš.
456
20 Iranian languages
Thordarson, Fridrik. 2009. Ossetic language i: History and description. In Ehsan
Yarshater, Mohsen Ashtiany & Mahnaz Moazami (eds.), Encyclopædia iranica,
online edn. New York: Encyclopædia Iranica Foundation.
Van Coetsem, Frans. 1988. Loan phonology and the two transfer types in language
contact. Dordrecht: Foris.
Van Coetsem, Frans. 2000. A general and unified theory of the transmission process
in language contact. Heidelberg: Winter.
Yūsifī, Ɣulāmḥusain. 2004. Gulistān-i Saʕdī. Tehran: Širkat-i Sahāmī-yi Intišārāt-i
Ḫārazmī.
457
Chapter 21
Kurdish
Ergin Öpengin
University of Kurdistan-Hewlêr
This chapter provides an overview of the influence of Arabic on Kurdish, especially on its Northern and Central varieties spoken mainly in Turkey–Syria–Iraq
and Iraq–Iran, respectively. It summarizes and critically assesses the limited research on the contact-induced changes in the phonology and syntax of Kurdish,
and proposes several new dimensions in the morphology and syntax, in addition
to providing a first treatment of lexical convergence in Kurdish through borrowings from Arabic.
1 Kurdish and its speech community
Kurdish is a Northwestern Iranian language spoken by 25 to 30 million speakers
in a contiguous area of western Iran, northern Iraq, eastern Turkey and northeastern Syria. There are also scattered enclaves of Kurdish speakers in central
Anatolia, the Caucasus, northeastern Iran (Khorasan province) and Central Asia,
with a large European diaspora population. The three major varieties of Kurdish
are: (i) Southern Kurdish, spoken under various names near the city of Kermanshah in Iran and across the border in Iraq; (ii) Central Kurdish (also known as
Sorani), one of the official languages of the autonomous Kurdish region in Iraq,
also spoken by a large population in western Iran along the Iraqi border; (iii)
Northern Kurdish (also known as Kurmanji), spoken by the Kurds of Turkey,
Syria and the northwestern perimeter of Iraq, in the province of West Azerbaijan
in northwestern Iran and in pockets in the west of Armenia (cf. Haig & Öpengin
2014 for a discussion on defining “Kurdish”). Of these three, the largest group
in terms of speaker numbers is Northern Kurdish. The Kurdish population in
respective states is difficult to reliably determine since none of the sovereign
countries make the relevant census information available. Table 1 provides some
Ergin Öpengin. 2020. Kurdish. In Christopher Lucas & Stefano Manfredi (eds.),
Arabic and contact-induced change, 459–487. Berlin: Language Science Press.
DOI:10.5281/zenodo.3744541
Ergin Öpengin
cautious estimates based on various sources (especially Sirkeci 2005; Zeyneloğlu
et al. 2016; and Ethnologue).1,2
Table 1: Estimates of Kurdish population numbers
Country
Turkey
Iraq
Iran
Syria
Population size
c. 15,000,000
c. 6,000,000
c. 8,000,000
c. 2,000,000
2 The history of Kurdish–Arabic contact
Information about the pre-Islamic history of the Kurds and their language is
scarce. According to early Islamic sources, at the time of the Islamic conquest
of the Near East (Upper Mesopotamia, Iran, and Armenia) in the seventh century (Bois et al. 2012: 451), the communities designated with the term Kurd were
already living in most of the present-day Kurdish-inhabited areas, namely from
Mosul to the north of Lake Van, and from Hamadan to the Jazira region situated
around the intersection of present-day Syria, Iraq and Turkey (James 2007: 111).
The Kurds have thus been living in contact with various Aramaic-speaking Christian and Jewish communities as well as Arabic-speaking communities since at
least the early Islamic period, though the contact of Iranian-speaking populations
with Aramaic dates back to the fifth century BCE (cf. Utas 2005: 69, citing also
Folmer 1995 and Kent 1953). Kurdish differs from other Iranian languages such as
Persian in sharing the same or close geographical spaces with Arabic-speaking
populations, especially in Upper Mesopotamia. The historical socio-cultural contact between Kurdish and Arabic-speaking communities requires a more refined
treatment than is currently possible, but there are a number of medieval Arabic
sources which attest to the interaction and mobility of Kurdish and Arabic communities in some regions (e.g. Erbil, Mosul), as well as language shift of some
Kurdish communities to Arabic and vice versa (cf. Bois et al. 2012: 449, 452, 456;
James 2007: 115–120).
1
2
See https://www.ethnologue.com/language/kur (accessed 31/01/2020; Eberhard et al. 2019).
The population figures should not be taken as equivalent to “number of speakers”, since especially in Turkey a significant portion of the Kurdish population grow up with no or very
limited knowledge of Kurdish (cf. Öpengin 2012; Zeyneloğlu et al. 2016).
460
21 Kurdish
Given the unquestionably prestigious status of Arabic in administration and
sciences in the Islamicized Near East, consolidated especially under Abbasid rule
(which included most of the Kurdish-inhabited areas), Kurdish was heavily dominated by Arabic. Even in several of the important medieval Kurdish dynasties
such as that of the Marwānids (10th–11th centuries), Arabic enjoyed the high status of being the administrative and literary language (cf. James 2007: 112), since
the coins bore Arabic script, while qaṣīda reading ceremonies or contests would
feature primarily Arabic, but to a limited extent also Persian pieces (Ripper 2012:
507–528). With the conquest of the Kurdish-inhabited regions by Turkic peoples
and Mongols from tenth century onwards, which led also to the final overthrowing of the Abbasid state in 1248 by the Mongols, the Arabic-speaking populations
may have started to diminish and retreat. Although at this stage Persian attained
a firm status as the literary language in the Islamic East (Perry 2012: 73), Arabic
preserved its higher status in administration and, later on, especially in education, well into the end of the nineteenth century. Thus, Kurdish developed a
literary tradition only starting from the sixteenth century, but its limited usage
was largely restricted to writing verse throughout the following several centuries.
The literature in this period is heavily dominated by the vocabulary and literary
formulas and metaphors of the two dominant languages, Arabic and Persian (cf.
Öpengin forthcoming).
In the early twentieth century, with the dissolution of the Ottoman Empire,
Kurdish in Iraq and Syria again came into primary direct contact with Arabic. In
Iraq, up until 1991, with the establishment of a Kurdish autonomous region, the
language configuration was one in which Arabic was the prestigious language of
higher domains. Not being in possession of any official status, the Kurds in Syria
have been in a highly asymmetric language-contact situation with Arabic. In
Turkey, especially in Mardin and Siirt provinces, Kurds have been in contact with
Arabic-speaking communities, but as the lingua franca of the communities of
cultural–historical Kurdistan (cf. Edwards 1851: 121), Kurdish must have been the
dominant language of interaction between these communities (cf. Lentin 2012),
and it is indeed possible to observe important influences from Kurdish on the
local Arabic dialects (cf. Jastrow 2011 and §3.1 below.).
As a result of these differing degrees and modalities of contact with Arabic,
the influence of Arabic should be viewed as consisting of at least two layers, and
viewed separately for different country contexts where Kurdish is spoken. Of
the two layers, there should be assumed a deeper contact influence, shared in
larger portions of Kurdish-speaking areas, dating to before the twentieth century; and a more shallow layer that is the result of the more recent societal bilingualism in Iraq and Syria. Likewise, while in Syria and Iraq the Arabic influence
461
Ergin Öpengin
on Kurdish continues, this influence is largely replaced by influence from the
dominant state languages in Turkey and Iran. Naturally, the intensity of Arabic
influence on Kurdish shows a great deal of variation across Kurdish varieties
and dialects within varieties. Accordingly, the historically deeper-layer Arabic
influence on Kurdish is characterized by its being restricted mostly to lexicon
and being shared in the majority of Kurdish dialects. This has been the result
of borrowing under recipient-language agentivity in the sense of Van Coetsem
(1988; 2000). On the other hand, the relatively advanced Arabic influence on the
Kurdish spoken in the historical Jazira region (including Mosul, northeast Syria,
and Mardin province in southeast Turkey), as well as the more recent Arabic influence on the Kurdish spoken in Syria, but also – albeit more restrictedly – in
Iraq, concerns also grammatical constructions and at least some of that contact
influence could be due to imposition under source-language agentivity.
3 Contact-induced changes in Kurdish
3.1 Phonology
The consonant inventory of Kurmanji is given in Table 2.3
In cells of doublets/triplets, the voiceless phonemes come first. The apostrophe
on plosive and fricative phonemes indicates aspiration, which marks a phonemic
distinction in Kurmanji. In addition to these consonants with indisputable phonemic status, there are the so-called emphatic or pharyngealized variants of the
obstruents /p, b, t, d, s, z/. These variants are transcribed in the text with a dot
beneath the characters.
The consonant inventory of Sorani is basically identical with Table 2, except:
(i) it does not have unaspirated stop phonemes; and (ii) it has velar nasal and
velarized lateral phonemes (Öpengin 2016: 27).
Arabic (or more generally Semitic) influence on the phonology of Kurdish is
most clearly observed in the presence of the two pharyngeal phonemes ḥ [ħ] and
ʿ [ʕ] (cf. Kahn 1976; Haig 2007; Anonby 2020; Barry 2019), as well as the series
of emphatic obstruants ṭ, ḍ, ṣ, and ẓ (Haig & Öpengin 2018), respectively. The
precise Semitic source language for these sounds cannot be determined, since
Kurdish (or rather its ancestor languages) must have been in close contact with
3
Kurdish data are transcribed in the standard Kurdish Latin alphabet with some additions for
emphatics and pharyngeals, mostly consonant with the Library of Congress approach for
the romanization of Kurdish: https://www.loc.gov/catdir/cpso/romanization/kurdish.pdf (accessed 31/01/2020).
462
21 Kurdish
m
ḥʿ
’
h
ttal
q
Glo
gea
l
k’ k g
xẍ
P ha
r yn
lar
Uvu
şj
ç’ ç c
Vel
ar
t’ t d
sz
ata
l
fv
p’ p b
Pal
-de
Alv
eola
r
Plosive
Fricative
Affricate
Nasal
Trill
Flap
Approximant
Lateral
Lab
io
Bila
b
ial
nta
l
Table 2: Consonant phonemes in Kurmanji
n
r̄
r
w
y
l
various Semitic languages for more than two millennia (Utas 2005: 69). However, these phonemes set the consonant inventory of Kurdish clearly apart from
other West Iranian languages such as Persian, with the only other West Iranian
languages possessing both pharyngeals and emphatic consonants being Zazaki,
and the Kumzari language spoken mainly in Oman (Anonby 2020). In what follows, I illustrate the presence and interactions of the pharyngeal and emphatic
consonants in Kurdish, and provide a brief discussion of their paths of development.4
The pharyngeal phonemes are found in varying degrees in both Central Kurdish and Northern Kurdish. They are retained in most of the Arabic loanwords
originally bearing them, a list of which is given in Table 3.5
Some loanwords with original pharyngeals are reanalysed as containing their
non-pharyngeal counterparts. Such is the word haq from Arabic ḥaqq ‘right’, or
4
The Kurmanji lexical items presented in this section are based on my native-speaker knowledge of the Şemdînan (Şemdinli) dialect, and my knowledge of Kurmanji-internal dialectal
variation, drawing also on (Chyet 2003), (Öpengin & Haig 2014), and the Manchester Database
of Kurdish Dialects presented in Matras & Koontz-Garboden (2016). The Sorani lexical items
are from Öpengin (2016) and the popular press.
5
Note that all through the article, unless stated otherwise, the Arabic data represents Classical
Arabic, giving an approximation of the ultimate Arabic etyma of the items without necessarily
implying that these are the immediate source of the Kurdish items (as they may have been
borrowed from local Arabic dialects as well as through the intermediary languages such as
Persian or Ottoman). Furthermore, the glosses in tables are for Kurdish items, as sometimes
the meanings of the Arabic etyma are not completely identical.
463
Ergin Öpengin
Table 3: Loanwords with retained pharyngeals in Kurdish
Arabic
Northern Kurdish
Central Kurdish
Gloss
ʕarab
maʕlūm
ʕadāla(t)
ṭābiʕ
maḥall
maḥšar
ḥākim
ḥammām
baḥr
ʿereb
meʿlûm
ʿedalet
tabiʿ/ṭabiʿ
miḥele
meḥşer
ḥakim
ḥemam
beḥr
ʿereb
meʿlûm
ʿedaḷet
ṭabiʿ
meḥel
meḥşer
ḥakim
ḥemam
beḥr
‘Arab’
‘evident’
‘justice’
‘dependent’
‘neighborhood’
‘resurrection (day)’
‘judge, governor’
‘bath’
‘sea’
the Arabic word ṭaʕm ‘taste’ that is seen in eastern dialects of Northern Kurdish
and in Central Kurdish without the voiced pharyngeal as ṭam and tam, respectively.
Furthermore, an original pharyngeal in a loanword may be substituted with
the alternative pharyngeal sound, so, for example, the voiced pharyngeal of the
Arabic ṭamaʕ ‘greed’ may be realized as either of the pharyngeals in different
Kurdish dialects. Such indeterminate or alternative use of pharyngeals may exist
within a single dialect (cf. Kahn 1976: 25). For instance, in the Mukri dialect of
Central Kurdish, (Öpengin 2016: 41–42) the following Arabic-origin words can
be found in both of the form pairs: saʿib ~ saḥib ‘owner’, ʿerz ~ ḥerz ‘honour’,
cemaʿet ~ cemaḥet ‘community’.
Finally, a pharyngeal may develop in loanwords that have no pharyngeal in the
source language. Thus, in most of Northern Kurdish the Arabic word ʔarḍ ‘earth’
appears with a non-etymological pharyngeal as ʿerd, while the Arabic word ǧāhil
‘naïve, young’ is seen with a pharyngeal as caḥêl (but also cahil).
Although the pharyngeals in Kurdish occur mostly in Arabic loanwords, they
have expanded also into inherited native Iranian lexicon, especially in Northern
Kurdish. However, unlike in Arabic loanwords, fluctuation between pharyngeal
and non-pharyngeal uses of such words among the dialects (sometimes in immediate geographic proximity) is readily apparent. Table 4 presents some native
Iranian words of this kind. Where relevant, the non-pharyngeal forms are also
noted, while Persian cognates are included for comparison.
More striking, however, is the emergence of a voiced pharyngeal in a subset of
words with similar structure in the northern dialects of Northern Kurdish that are
464
21 Kurdish
Table 4: Pharyngeal sounds in native Iranian lexical items
Persian
Northern Kurdish
Central Kurdish
Gloss
abr
zabān
āsemān
ḫošk
haft
hašt
bahašt
ʿewr
ʿezman ~ ziman
ʿesman
ḥişk ~ hişk
ḥeft ~ heft
ḥeşt ~ heşt
biḥeşt ~ bihişt
hewr
ziman
asman ~ hasman
wişk
ḥewt
heşt
beheşt
‘cloud’
‘language’
‘sky’
‘dry, hard’
‘seven’
‘eight’
‘paradise’
geographically farthest from direct Arabic/Semitic contact but close to Caucasian
languages which also possess pharyngeals. Thus, the native words such as masî
‘fish’, çav ‘eye’, mar ‘snake’ (in Central Kurdish and in central and southern dialects of Northern Kurdish) appear in the northern dialects of Northern Kurdish
with a pharyngeal, as meʿsî, çeʿv, meʿr. These are obviously the result of languageinternal processes, though nested in an initial introduction of the phonemes into
the language via contact with either Arabic or Caucasian languages, or both.
As for their distribution, the pharyngeal phonemes are most robustly present
in the central areas of the Northern and Central Kurdish speech zones. Their
presence in Arabic loanwords is weakened towards the extreme northern and
southern peripheries in heavy contact with Turkish and Persian (cf. Map 1.27 in
the Manchester Kurdish Database, which illustrates such weakening of pharyngeals at the peripheries through the distribution of the Arabic loanword ḥeywan
‘animal’).6
We turn now to the series of emphatic (pharyngealized) obstruents ṭ, ḍ and ṣ, ẓ.
Table 5 gives a list of Arabic loanwords in which the original emphatic consonant
is retained in Kurdish.
In the deeper-layer loanwords, the Arabic interdental and voiced alveolar emphatics are merged into the voiced emphatic alveolar phoneme ẓ in Kurdish. But
in present-day Iraqi and Syrian Kurdish speech, especially those speakers with
formal education may also pronounce the interdental phoneme, especially in the
case of nonce borrowings and code-mixing.
On the other hand, quite a number of Arabic loanwords are pronounced without their original emphatic consonants, and thus reanalysed as the corresponding
plain phonemes (similarly to Persian), as in the items in Table 6.
6
See http://kurdish.humanities.manchester.ac.uk/pharyngeal-retentionloss-animal/ (accessed
31/01/2020).
465
Ergin Öpengin
Table 5: Arabic loanwords with emphatic consonants in Kurdish
Arabic
Northern Kurdish
Gloss
ṭaʕm
ṭāʔir
baṭṭāl
ð̣ulm
ḍābit
ṣūfī
ṣāfī
ṭam ~ ṭeʿm
ṭeyr
beṭal
ẓulm
ẓabit
ṣofî
ṣaf ~ ṣafî
‘taste’
‘bird’
‘empty, cancel’
‘oppression’
‘clerk’
‘devotee, Sufi’
‘clear’
Table 6: Arabic loanwords with lost emphatics in Kurdish
Arabic
Northern Kurdish
Central Kurdish
Gloss
ḫāṭir
ṭaraf
šayṭān
ḍaʕīf
ḥāḍir
qaṣṣāb
fasīḥ
ṣabr
xatir
teref
şeytan
zeʿîf
ḥazir
qesab
fesîḥ
sebr ~ ṣebr
xatir
teref
şeytan
zeʿîf
ḥazir
qesab
fesîḥ
sebr
‘mind’
‘side, direction’
‘devil’
‘weak’
‘ready’
‘butcher’
‘clear’
‘patience’
On the reverse side, some Arabic loanwords with no original emphatic consonants are pronounced with emphatic consonants in Kurdish, such as ẓełał (~
ẓelal and zelal) from Arabic zulāl ‘clear’ (dialectal zalāl), or ẓelam ‘man’ from
Syrian Arabic zalame.
Finally, just as with the pharyngeal consonants, emphatic sounds also appear
in inherited native Iranian words, as illustrated in Table 7.
Of the emphatic obstruents, the fricative pair (ṣ, ẓ) are found both in Northern
and Central Kurdish (though less often in the latter), while the stops (ṭ, ḍ) are
found only in Northern Kurdish, with the voiced counterpart being extremely
rare. The fact that the voiceless emphatic stop is widespread only in Northern
Kurdish most probably has to do with the presence of two series of aspirated
and unaspirated voiceless stops in the language (cf. Table 2). The unaspirated
stops are probably intermediary in the development of emphatics. This is fur-
466
21 Kurdish
Table 7: Emphatic consonants in native Iranian lexical items
Northern Kurdish
Central Kurdish
Gloss
meẓin
ẓiman
ẓik ~ zik
aẓad
ẓava
beẓîn
peẓ
ṣal
ṣed ~ sed
ṣe
beṣ ~ bes
ṣawa
ṣotin
ṣiṣt
ṭarî
ṭezî
ṭeng
ṭerm
ṭirş
ḍaṣ(ik)
ḍiṙî
ziman
zig
azad
zawa
pez
ṣał
ṣed
ṣeg
bes
sawa
sûtan
tarîk
tezî
teng
term
tirş
das
-
‘big’
‘language’
‘stomach’
‘free’
‘groom’
‘to run’
‘sheep’
‘year’
‘hundred’
‘dog’
‘enough’
‘very young, newborn’
‘to burn’
‘loose’
‘dark’
‘cold’
‘narrow’
‘dead body’
‘sour’
‘sickle’
‘blackberry bush’
ther reinforced by the fact that in Northern Kurdish the bilabial voiceless stop
p also has an emphatic version, as in the native words ṗeẓ ‘sheep’ and ṗenîr
‘cheese’ (in some dialects; cf. Kahn 1976: 27). Within Northern Kurdish, they are
found in more southerly dialects, and are noted to be particularly frequent in
both the Kurdish and Neo-Aramaic of Duhok and Hakkari provinces (Blau 1989:
329). They tend to be less present moving northwards (Erzurum–Kars) while
MacKenzie (1961: 43) notes that they are altogether absent in the Yerevan dialect.
This distribution is of course consistent with a language-contact scenario, in the
sense that in the northern dialects away from Semitic influence the language
either did not develop emphatics or lost them as a result of contact with and
bilingualism in Armenian, Turkic and Caucasian languages that do not possess
such emphatics.
467
Ergin Öpengin
Given the shallow history of written Kurdish, it is not possible to determine the
historical period of the introduction of the emphatics and pharyngeals into the
language. However, they are found abundantly even in the earliest Kurdish texts,
especially in the Arabic component, but also in inherited lexical items, such as
ṣal ‘year’, ṣar ‘cold’, ṣed ‘hundred’, meẓin ‘big’, ḥemyan ‘all of them’ (items taken
from Şêxê Senʿaniyan by the early seventeenth-century poet Feqiyê Teyran, cf.
Teyran 2011).
Three studies have treated the pharyngeals and emphatics in Kurdish, namely
Kahn (1976), Anonby (2020) and Barry (2019). Barry (2019) suggests that the pharyngeal sounds (including emphatics) in Kurdish are the result of contact influence from Arabic with a phonetic basis. The phonetic basis consists in the recategorization of vowels and the h sound within syllables with “flat” consonants
(including pharyngeals, rhotics, grooved fricatives, and labials). Thus, initially,
through extensive language contact with and bilingualism in Arabic, the speakers attained an active category of pharyngeals. Then the (inherited or loan) vocabulary with sounds that have pharyngeal-like effects on neighbouring vowels
led to the reanalysis of the given vocabulary items as pharyngeal. In this account, the whole syllable is pharyngeal rather than individual sound segments.
This account is particularly appropriate since, while it acknowledges the role of
language contact with Arabic in the initial stage, it posits a phonetic mechanism
of language-internal development of pharyngealization that captures an expansion of pharyngeals into historically non-pharyngeal lexical items that would
be impossible to explain on purely language-contact grounds. It is, for instance,
consistent with the fact that, in the above-presented data, the emphatics, but
not pharyngeals in loan words, are restricted to the environment of more open
vowels: e, a, o, and i [ɪ]. Furthermore, although not stated in the source study, the
assumed subsequent development of a phonetic basis for the propagation of the
pharyngeals into items originally without pharyngeal sounds is consonant with
the facts of different stages or layers of borrowing. For instance, from the Arabic
root √ǧmʕ we have three forms in Kurmanji: civat ‘community, company’, cimat
‘the assembly of prayers in a funeral’, and cemaʕet ‘community’. The first form is
probably the result of an early borrowing right after the initial Islamicization of
the Kurds, as the fricativization of the bilabial nasal was active then (as seen also
in silav ‘greeting’ from Arabic salām; Paul 2008). The second form with a slightly
specialized semantic difference may have originated in a dialect where the mentioned fricativization did not occur. In any case, the first two forms, which are
clearly early borrowings, did not retain the original pharyngeal, whereas in a
later borrowing from the same root, when one can assume that the pharyngeals
468
21 Kurdish
were better tolerated in the language (and that the fricativization of bilabial nasal
was not active), the pharyngeal sound did survive.
However, this account fails to explain why, in the great majority of the vocabulary with the relevant phonetic environment (syllables with “flat” consonants
and low and back vowels), pharyngealization has not occurred. If the phonetic
mechanism is integrated into the phonological system of the language, then pharyngealization would be expected in all relevant contexts. In this sense, although
there is a phonetic ground to the propagation of the pharyngeals and emphatics
in Kurdish, it may be safer not to postulate it as integrated into the phonological system of the language. Rather, the pharyngeals and emphatics should still
be considered as peripheral to the phonological system (cf. Haig 2007; Anonby
2020), since, as noted by Haig (2007: 167), they are restricted to individual lexical items, their functional load is very limited, and there is considerable crossspeaker and cross-dialectal variability in the extent of their presence.
Although it is not the main focus of this chapter, a note on the reverse direction of contact influence is in order at this point. The Arabic dialects of Anatolia or Upper Mesopotamia (Mardin, Siirt, Kozluk, Sason, and the plain of Muş)
have adopted some consonant and vowel phonemes via loanwords from Kurdish
and Turkish, which do not exist in mainstream Arabic dialects (Jastrow 2011: 84;
Akkuş, this volume: §3.1.1). The phonemes and example words with their sources
are given in Table 8.
These additions into the phoneme inventory of the Anatolian Arabic are evidently the result of contact with Kurdish and Turkish. The introduction of these
new phonemes has, as noted by Jastrow (2011: 84), on the one hand re-established
the lacking symmetry caused by historical sound changes in Old Arabic, while on
the other hand causing further sound shifts in the inherited Arabic vocabulary.
3.2 Morphology
It is usually assumed that Arabic influence on Kurdish is absent in the grammar
(e.g. Edwards 1851), being largely restricted to phonology and lexicon. This is
indeed to a large extent true. There are, however, several potential grammatical
features that may be related to such contact influence.
Matras (2010: 75) suggests that the presence of aspect–mood prefixes in the languages of the Eastern Anatolian linguistic zone, namely Persian, Kurdish, NeoAramaic, Arabic and Western Armenian, is an outcome of language contact. Accordingly, all of these languages have a progressive–indicative aspectual prefix
(in turn: mī-, di-, gǝ-, ko-, ba-/-a-), while subjunctive is marked either by the absence of the indicative prefix (Armenian and Neo-Aramaic) or by a specialized
469
Ergin Öpengin
Table 8: Borrowed phonemes in Arabic dialects of Anatolia
Phoneme
Example
bilabial stop /p/
voiced labio-dental fricative /v/
voiceless affricate /č/
voiced palatal fricative /ž/b
voiced velar stop /g/
mid long front vowel /ē/
mid long back vowel /ō/c
parčāye ‘piece’ < Tr. parça
davare ‘ramp’ < Kr. dever (f) ‘place’
čǝqmāq ‘lighter’ < Tr. çakmak a
ṭāžī ‘greyhound’ < Kr. ṭajî
gōmlak ‘shirt (modern)’ < Tr. gömlek
tēl ‘wire’ < Tr. tel (via Kr. têl)
ḫōrt ‘young man’ < Kr. xort
a
It is more probable that this word (and others attributed to Turkish) is borrowed via Kurdish,
since the uvularization (/k/ > /q/) in loanwords and the change in the vowel of the first syllable
(cf. also qeymaẍ ‘cream’, from Tr. kaymak) are typical of Kurmanji spoken in the region.
b
Note that the reflex of Arabic 〈 〉جin this variety is /ǧ/, not /ž/.
c
Note also that the original Arabic diphthongs *ay and *aw are preserved in this variety, not
monophthongized to /ē/ and /ō/.
subjunctive prefix (Persian, Kurdish, Arabic). Since such aspect–mood prefixes
are considered typical of Iranian languages of the region, they would have diffused from Kurdish and Persian into the other languages of the zone, including
Arabic (which in its standard grammar does not have such forms; cf. Ryding 2014:
46–47). However, assessing the validity of Eastern Anatolia being a linguistic
area, Haig (2014: 20–25) casts doubt on this claimed contact scenario, primarily
since: (i) the feature exists in Arabic dialects outside the region; and (ii) it is absent in the two major languages of Anatolia, namely Turkish and Zazaki. Jastrow
(2011: 92), on the other hand, although acknowledging the source of such verbal
prefixes grammaticalizing from Old Arabic verb forms, hypothesizes – though
without providing supporting arguments – that they may have developed under
Turkish and Kurdish influence. Assessing also the grammaticalization of such
formatives in various languages and rejecting a contact scenario behind their
frequent occurrence in the languages of Anatolia, Haig (2014: 26) concludes that
the present indicative prefixes found in Kurmanji, and in certain varieties of Aramaic and Arabic in Anatolia, could be interpreted as reflexes of an inherited
morphological template, which is well-attested in the related Northwest Iranian
and Semitic languages outside Anatolia.
Another (not previously discussed) candidate for Arabic influence on Kurmanji Kurdish relates to gender assignment in more recent loanwords from European languages. In Kurmanji, like Arabic, nouns are assigned to feminine and
470
21 Kurdish
masculine genders. The gender of inanimate nouns is largely arbitrary, with limited morpho-phonological basis in both languages. In Arabic, words carrying the
-a ending are feminine, while in Kurmanji abstract nouns ending in -î are feminine, while the rest may be of either gender. Now, when Arabic borrows modern
vocabulary items from European languages, items ending in -a are assigned to
feminine gender, while the rest are assigned to masculine gender (Ibrahim 2015:
5). The default gender assigned to new lexical borrowings is masculine in Arabic.
There is as yet no research on the gender assignment of borrowings in Kurmanji.
However, it is easily observed that Kurmanji spoken in Turkey mostly favors feminine, while the Kurmanji of Iraq uses masculine gender for integrating modern
vocabulary items into the language. The modern lexical borrowings (boldface) in
(1) are all assigned to masculine gender in Badini Kurmanji of Iraq. Note that the
gender of the nouns is visible in the ezāfe (see §3.3) and oblique case suffixes.
(1)
Badini dialect of Kurmanji in Iraq (from media outlets)
a. sîstem-ê
endroyd-ê
system-ez.m android-obl.f
‘Android system’
parleman-î
b. serok-ê
president-ez.m parliament-obl.m
‘the president of the parliament’
c. formê
têgehiştin-ê
form-ez.m understanding-obl.f
‘the form of understanding’
d. moral-ê
diyalog
moral-ez.m dialogue
‘the moral of dialogue’
e. proj(e)-ê
av-ê
project-ez.m water-obl.f
‘the water project’
f. prensîp-ê
hevwelatîbûn-ê
principle-ez.m citizenship-obl.m
‘the principle of citizenship’
All of these lexical borrowings exist also in Kurmanji as spoken in Turkey, but
they are systematically used with feminine gender. For instance the phrase in
471
Ergin Öpengin
(1b) would be realized as serok-ê parleman-ê (president-ez.m parliament-obl.f),
with the feminine form of the oblique case suffix.
As was stated above, the majority of such modern lexical borrowings in Arabic
are assigned to masculine gender. The masculine gender assignment in Kurmanji
in Iraq is thus most probably motivated by the Arabic gender assignment pattern.
This is all the more plausible when we consider that Arabic, as the dominant state
language for the Iraqi Kurds for almost a century, serves also as the intermediary
language via which such lexical items are normally borrowed into Kurmanji in
Iraq. However, this contact influence must have been established relatively recently, since earlier technical borrowings in Kurmanji in Iraq such as têlevizyon
and radyo are treated as feminine nouns, despite being masculine in Arabic.
3.3 Syntax
Although several studies have dealt with the outcomes of language contact between Kurdish and (Neo-)Aramaic in the grammar of these languages – especially on such topics as alignment (Coghill 2016), word order (Haig 2014), and
noun phrase morphology (Noorlander 2014) – as far as I am aware, the only study
on Arabic–Kurdish contact in grammar is the short note of Tsabolov (1994) about
the distinctive position of the possessor in a multiple-modifier noun phrase in
Northern Kurdish.
As is well known, a number of West Iranian languages (Middle and contemporary Persian, Kurdish, Zazaki, etc.) employ a bound morpheme for linking posthead modifiers in a noun phrase, called ezāfe or izāfe (from Arabic ʔiḍāfa ‘joining,
addition’), as in (2) and (3).
(2) Persian (personal knowledge)
ḫāna-e bozorg
house-ez big
‘(the) big house’
(3) Northern Kurdish (personal knowledge)
xanî-yê
mezin
house-ez.m big
‘the big house’
The ezāfe in Northern Kurdish differs from its cognates in, for instance, Central
Kurdish and Persian, as it inflects for gender (masculine -ê vs. feminine -a) and
number (singular -ê/-a and plural -ên/-êd), in addition to having secondary or
pronominal forms used in chain ezāfe constructions with multiple modifiers (and
472
21 Kurdish
some other predicative functions; cf. Haig 2011; Haig & Öpengin 2018). In most
West Iranian languages, noun phrases with multiple modifiers have their head
noun first, followed by qualitative then possessive modifiers, as in (4) and (5). This
is also the order in Middle Persian, as in (6), where Tsabolov (1994: 122) considers
such constructions may be regarded as prototypes of the ezāfe constructions of
modern West Iranian languages.
(4)
Persian (personal knowledge)
ḫāna-e bozorg-e Malek
house-ez big-ez pn
‘Malek’s big house’
(5)
Central Kurdish (personal knowledge)
kurr-î gewre-y Karwan
son-ez big-ez pn
‘my friend’s beautiful daughter’
(6) Middle Persian (Tsabolov 1994: 122)
pus ī mas ī Artavān
son ez big ez pn
‘Artavan’s elder son’
However, in Northern Kurdish the order of modifiers is reversed, such that a
possessor of the head noun in the noun phrase comes before attributive modifiers,
as in (7), where the secondary linking element is glossed as sec.
(7)
Northern Kurdish (personal knowledge)
xanî-yê
Malik-î (y)ê
mezin
house-ez.m pn-obl.m ez.m.sec big
‘Malik’s big house’
Tsabolov observes that these syntactic particularities of Northern Kurdish have
no parallels in other Kurdish varieties and Iranian languages as a whole, but that
they correspond to the word order in noun phrases in Arabic, as can be seen in
the comparison of (8) and (9).
(8)
Arabic (Tsabolov 1994: 123)
l-ǧadīdatu
miḥfað̣atu ṭ-ṭālibi
bag.nom def-student.gen def-new.f.nom
‘the student’s new bag’
473
Ergin Öpengin
(9)
Northern Kurdish (Tsabolov 1994: 123)
çent-ê şagirt-î
taze
bag-ez.m student-ez.m.sec new
‘the student’s new bag’
Note that although in standard Kurmanji (Northern Kurdish) the forms of the
primary and secondary ezāfes are identical, with the difference being in the latters’ status either as enclitics or independent particles, in the northern dialect
of Northern Kurdish considered by Tsabolov, the singular forms of the secondary ezāfe are different (with masculine -î and feminine -e). In Tsabolov’s view,
the centuries-old close contacts between Kurdish and Semitic dialects, especially
Arabic, have not only resulted in the above-described change of noun-phraseinternal word order (syntactic) but also in the development of secondary forms
of ezāfe through the “weakening” of the primary ones (morphological), because,
he argues, such distinct forms “were necessary for correlating each attribute in
an [ezāfe] chain with the ruling noun they refer to” (1994: 123).
On closer scrutiny, however, the motivation Tsabolov puts forward for the
morphological change may not be entirely correct, since, on the one hand, ezāfe
forms in Northern Kurdish distinguish gender and number, which already correlate the modifiers with their head nouns, and on the other hand, in the majority
of Northern Kurdish dialects the primary and secondary ezāfes are formally identical. The change in form is an instance of vowel raising (a > e, ê > î) that is also
observed elsewhere in the morphology of noun phrase (cf. Haig & Öpengin 2018).
As for Tsabolov’s main claim regarding word-order change leading to the initial positioning of a possessor modifier in the noun phrase, here too the role of
language contact might require revision, since it might have more to do with
language-internal organization of morphological material: Zazaki (geographically contiguous with Kurmanji but from a separate historical source to Kurdish), which, like Kurmanji, has gender/number-marking ezāfe forms and a case
distinction in its nominal system, follows precisely the same word order pattern
as Kurmanji in the noun phrase (cf. Todd 2002: 95), while Sorani, which has lost
gender/number-marking in ezāfes and case distinctions in its nominal system,
differs from them and instead follows the Persian and Middle Persian pattern
(cf. Öpengin 2016: 61–64). That is, the determining factor seems to be the presence or absence of gender/number-marking ezāfe forms, which enable reference
tracking between heads and dependents in a noun phrase independently of word
order.
Despite the scepticism one may have towards Tsabolov’s hyopthesis, there
is a rather parallel more recent syntactic change in progress stemming from the
474
21 Kurdish
Arabic influence on the Kurdish of Iraq. This change concerns especially the naming of institutions, such as schools and airports. Recall that in Central Kurdish
the possessor in a chain ezāfe construction is positioned at the end of the noun
phrase, as illustrated in (5). However, in the case of these examples, the proper
name occurs right after the head noun and before the qualitative modifier, as in
(10) and (11).
(10)
Central Kurdish (official signage)
qutabxane-y Qemeryan-î seretayî
school-ez pn-ez
primary
‘Qamaryan primary school’
(11)
Central Kurdish (official signage)
firokexane-y Hewlêr-î nêwdewletî
airport-ez pn-ez
international
‘Hawler international airport’
If the proper name is understood as having the function of possessor here,
this is an order that is rather different from the typical Central Kurdish syntax of
chain ezāfe constructions. But this is precisely the order described for multiple
modifier noun phrases of Arabic, as in (8). Thus the order in (11) is the exact
replication of the Arabic version of the same name illustrated in (12).
(12)
Arabic (official signage)
maṭār arbīl ad-dawlī
airport pn def-international
‘Erbil international airport’
This is clearly a recent imposition from Arabic which does not seem to have
gone much beyond naming institutions, especially official signage: the Arabiclike ordering of the name of the airport appears only half as frequently as the
inherited order in a Google search. Furthermore, there is no trace of such a word
order pattern in the use of Central Kurdish in Iran.
3.4 Lexicon
Arabic influence on Kurdish and all other Near Eastern languages is observed
most clearly and abundantly in the vast number of loanwords. According to Perry
(2005: 97), the process of lexical convergence initially took place in Persian between the ninth and thirteenth centuries, when a large number of learned terms
475
Ergin Öpengin
were borrowed into literary Persian, and thence transmitted to the other languages of the region. This scenario explains some of the similarities of loanword
integration in the two languages (e.g. the borrowing of tāʔ marbūṭa as -at/-et
(rather than a) in a number of words, such as hukūmat ‘government’, Persian
hokūmat, and quwet ‘strength’, Persian qovvat). However, being spoken in a region that is closer geographically to Arabic-speaking communities, and having
had its own educational and religious institutions where Arabic served as the
high literary language, Kurdish must have also followed its own course of contact
with Arabic. Despite this, there are no studies of lexical borrowing from Arabic
into Kurdish. Given the vastness of the topic, with its layers of time-depth and
subsantial extra-linguistic aspects, I can only propose here to sketch the major
lexical domains of borrowing, and note some observations on the word class and
morpho-phonological integration of the borrowings.
The three major varieties differ in their proportions of borrowing from Arabic.
Impressionistically, Northern Kurdish seems to have borrowed most extensively.
There is, however, a deeper layer of lexical borrowings shared throughout Kurdish (some of which are common to all or most of the Near Eastern languages),
such as the following (cited in their Northern Kurdish forms):7
(13)
7
xerab ‘bad’
< Ar. ḫarāb ‘ruins’
xelk/xelq ‘people’
< Ar. ḫalq (√xlq ‘to create’)
xiyanet ‘betrayal’
< Ar. ḫiyāna
xizêm ‘nose-ring’
< Ar. ḫizām
xizmet ‘service’
< Ar. ḫidma
ʿeql/aqil ‘reason’
< Ar. ʕaql (qəltu Ar. ʕaqəl)
qelem ‘pen’
< Ar. qalam
quwet ‘strength’
< Ar. quwwa
kitêb ‘book’
< Ar. kitāb
xiyal ‘thought, grief’ < Ar. ḫayāl ‘imagination’
hevîr ‘dough’
< Ar. ḫamīr
fikr ‘thought, idea’ < Ar. fikr
fêkî/fêqî ‘fruit’
< Ar. fākiḥa
ḥal ‘condition’
< Ar. ḥāl
ḥazir ‘ready’
< Ar. ḥāḍir
şol/şuẍul ‘work’
< Ar. šuɣl
terk ‘abandonment’ < Ar. √trk ‘to abandon’
The main source for the lexical items in this section, together with the information regarding their Arabic origin, is Chyet (2003). However, I have supplied the interpretation and the
discussion of the material and as such only I am responsible for any shortcomings.
476
21 Kurdish
Within varieties too, the dialect zones where the communities have had historically closer contact with Arabic-speaking areas show greater Arabic influence in
vocabulary. Thus, the dialect of Northern Kurdish named as Southern Kurmanji
by Öpengin & Haig (2014), spoken around Mardin and Diyarbekir provinces in
Turkey, the Jazira province of northeast Syria, and the Sinjar region of Iraq, is the
dialect with most extensive Arabic lexical borrowings. Thus, the following items
are restricted to this dialect of Northern Kurdish: tefa-ndin ‘extinguish-tr.inf’
(from dialectal Ar. ṭafa or standard ṭafiʔa), şiteẍl-în ‘speak-intr.inf’ (from dialectal Ar. ištaɣal ‘to work’), hersim ‘unripe and sour grapes’ (from Ar. ḥiṣrim),
siʿûd ‘good luck’ (Ar. suʕūd, pl. of saʕd), şîret and şêwr ‘advice, counsel’ (Ar. √šwr).
Arabic loanwords in Kurdish belong to various semantic fields, such as kinship,
body parts, animals, agriculture, basic tools, temporal concepts and religion. Regarding kinship terms, while the terms for the members of the nuclear family are
all inherited, the four second-degree kin terms are all borrowed from Arabic: met
‘paternal aunt’ (cf. Ar. ʕamma(t); this item does not exist in Sorani), xalet/xaltî
‘maternal aunt’ (Ar. ḫāla), mam ~ am ‘paternal uncle’ (Ar. ʕamm), xal ‘maternal uncle’ (Ar. ḫāl). Considering that the language had its own kin terms before
its contact with Arabic, the borrowing of such kin terms constitutes a case of
prestige borrowing, probably motivated by the use of such kin words as address
forms in the cultures of the region (cf. Haig & Öpengin 2015).
Similarly, while words for basic animals are inherited, the animals not indigenous to the mountainous region of core Kurdistan are borrowed from Arabic,
such as tîmseḥ ‘crocodile’ (Ar. timsāḥ), fîl ‘elephant’ (Ar. fīl), xezal ‘gazelle, deer’
(Ar. ɣazāl). Likewise, the generic term for ‘bird’ or ‘large birds’ is the Arabic
loanword ṭeyr (Ar. ṭayr), while the category word ferx ‘young bird/chicken’ is
also from Arabic farḫ. Several agricultural terms are also borrowed from Arabic, such as ẓad ‘grain, food’ (Ar. zād ‘provisions’), simbil ‘spike (of corn or
wheat)’ (Ar. sunbul), xox ‘peach’ (Ar. ḫawḫ), dims ‘grape molasses’ (Ar. dibs).
Various terms for spaces and tools of daily life are also borrowed from Arabic,
such as saʿet ‘hour’ (Ar. sāʕa), sifre ‘tablecloth’ (Ar. sufra ‘dining table’), qefes
‘cage’ (Ar. qafaṣ), ḥubr ‘ink’ (Ar. ḥibr), ḥemam ‘bath’ (Ar. ḥammām), ḥewş ‘yard’
(Ar. ḥawš), meẍmer ‘velvet’ (Ar. muḫmal). Some occupational terms from Arabic are neqş ‘embroidery’ (Ar. naqš ‘painting, drawing’), ḥedad ‘blacksmith’ (Ar.
ḥaddād), ʿesker ‘soldier’ (Ar. ʕaskar ‘army’), tucar and its older form têcirvan (Ar.
tuǧǧār ‘traders’, sg. tāǧir).
The older layer of administrative and legal terms are predominantly derived
from Arabic – though they may have mostly entered via Persian and Ottoman
Turkish – such as sultan ‘monarch’ (Ar. sulṭān), walî ‘provincial governor’ (Ar.
wālī), muxtar ‘village chief’ (Ar. muḫtār), ḥukûmet ‘government’ (Ar. ḥukūma),
477
Ergin Öpengin
meḥkeme ‘court’ (Ar. maḥkama), deʿwā ‘request, court case’ (Ar. daʕwa ‘request,
invitation’ and daʕwā ‘court case’), qanûn ‘law’ (Ar. qānūn), mekteb ‘school’ (Ar.
maktab ‘office, desk’).
As for religious terms, similar to the Persian case (cf. Perry 2012: 72), a number of basic Islamic concepts are inherited, such as the words for god, prophet,
angel, devil, heaven, purgatory, prayer, fasting, and sin. In some instances, the
Arabic equivalents of these terms exist alongside the inherited ones, restricting
the use of the latter, as in the cases of şeytan ‘devil’ and cehnem ‘hell’, from Arabic šayṭān and ǧahannam, replacing the Iranian dêw and dojeh. Many other basic
and more peripheral concepts are borrowed from Arabic, such as the following:
xêr ‘good’ (Ar. ḫayr), xezeb ‘wrath’ (Ar. ɣaḍab), civat/cemaʿet ‘society, gathering’ (Ar. ǧamāʿa), ḥec ‘pilgrimage’ (Ar. ḥaǧǧ), şeytan ‘devil’ (Ar. šayṭān), weʿz
‘(Islamic) sermon’ (Ar. waʿð̣), ḥelal ‘permitted’ (Ar. ḥalāl), ḥeram ‘forbidden’ (Ar.
ḥarām), ruḥ ‘soul, spirit’ (Ar. rūh), tizbî (Sorani tezbêḥ) ‘prayer beads’ (Ar. tasbīḥ).
Finally, there are also a large number of concepts (temporal, moral, cosmological) that originate from Arabic roots, such as sibe(h) ‘morning, tomorrow’
(Ar. ṣabāḥ), heyam ‘period’ (Ar. ayyām ‘days’), hêsîr ‘prisoner’ (Ar. ʔasīr), dinya
‘world’ (Ar. dunyā), ḥesab ‘count, calculation’ (Ar. ḥisāb), ḥîle ‘trick, ruse’ (Ar.
ḥīla), ḥel ‘solution’ (Ar. ḥall), eşq ‘love’ (Ar. ʕišq), ʿerz ‘honor, esteem’ (Ar. ʕirḍ).
Note also that the word dinya is used corresponding to the English expletive subject it in time and weather expressions, as in dinya esr e ‘it is late afternoon’ or
dinya ewr e ‘it is cloudy’. This usage is noted to exist also in colloquial Arabic
(Chyet 2003: 155).
Some other interesting developments with Arabic material in Kurdish lexicon
may be noted here. The Arabic daʕwa ‘invitation’ has resulted in two related
but different concepts: dawet/dewat ‘wedding ceremony’ and deʿwet ‘invitation’.
While the latter meaning is shared in Ottoman/Turkish and Persian, the former
is a Kurdish-internal semantic expansion of the source meaning. The Kurdish
(in all three varieties) word for ‘home’ mal, in the sense of family and familial
belongings, rather than the house as a structure, is probably derived from the
Arabic word māl ‘goods, property’. The generic term in Kurdish that designates
Christians regardless of their ethnicity and confession is fileh/file which derives
from Arabic fallāḥ ‘peasant, farmer’. Finally, there is the word mixaletî ‘the son of
the maternal uncle or aunt’ in the southern Kurmanji dialect of Northern Kurdish
that can probably be analysed as mi (< ben ‘son’) + xalet ‘aunt’ (< Ar. ḫāla) + î
‘my’.
Turning now to the word class categories of the loanwords, as has been seen
from the presentation of semantic domains above, most Arabic loanwords in
Kurdish are nouns. However, many Arabic noun loans are incorporated into
478
21 Kurdish
Kurdish verb forms. This takes place either through morphological integration
or syntactic composition. In morphological integration, the Arabic root (whether
nominal or verbal) is taken as the stem onto which the Kurdish verbal suffixes
-în/-îyan for instransitives and -andin for transitives are added. Thus the Arabic
noun ʕilm ‘knowledge’, apart from being used in its nominal sense, serves as the
stem for the derivation of the intransitive ʿelimîn (ʿelim-în) ‘to learn’ and transitive ʿelimandin ‘to teach, educate’. The following verbs are further examples
of using Arabic roots (whether the original borrowings are nouns or verbs is
not always clear) in the derivation of verbs in Kurdish: tefandin ‘to extinguish’
(Ar. ṭafa/ṭafiʔa), fetisandin ‘to suffocate’ (Ar. faṭṭas), fetilîn ‘to turn around’ (?Ar.
fatala ‘to twist together’), qulibîn ‘to be overturned’ (Ar. qalaba ‘to overturn’),
sekinîn ‘to stand, stop’ (Ar. √skn ‘calm, rest’), fikirîn ‘to think; to look at’ (Ar. fikr
‘thought’).8 The verb qelandin ‘to roast; to uproot’ has two sources as Ar. qalā
and qalaʕa, respectively, which explains its polysemy in Kurdish.
In syntactic composition, on the other hand, a compound verb9 is obtained by
combining an Arabic root with an inherited auxiliary light verb, such as kirin
‘do’ or dan ‘give’ for transitives, and bûn ‘to be’ for instansitives. Thus, the combination of Arabic adjective loanword xerab ‘bad’ (< Ar. ḫarāb ‘ruin’) with kirin
yields the verbal meaning ‘to destroy’ while its combination with bûn means ‘to
go bad, be spoiled’. Some example compound verbs with Arabic roots are given
in (14).
(14)
qedr ‘respect’ (Ar. qadr) + girtin ‘to hold’ = ‘to respect’
silav ‘greeting’ (Ar. salām ‘peace’) + dan ‘to give’ = ‘to greet’
teʿn/ṭan ‘scolding’ (Ar. ṭaʕn ‘piercing’) + dan = ‘to criticize’
qedeẍe ‘forbidden’ (Ar. qadaḥa ‘to rebuke’) + kirin ‘to do’ = ‘to forbid’
qesd ‘intention’ (Ar. qaṣd) + kirin = ‘to head for’
zeʿîf ‘weak’ (Ar. ḍaʕīf ) + bûn ‘to be’ = ‘to become slim’
What motivates the choice between the morphological versus syntactic technique in the integration of Arabic loan roots in forming verbs in Kurdish is not
8
Kurdish possesses a number of preverbs such as ve- and ra-. When inflected with tense–aspect–
mood prefixes, these preverbs are detached from the verb stem, as with the verb ve-kirin ‘to
open’ in ve-di-ki-m (pvb-ind-do.prs-1sg) ‘I open (it)’. Now, the initial syllable of the verbs
sekinîn and fekirîn, which are based on Arabic loan roots, resemble such Kurdish preverbs. As
a result, in some dialects, they are treated as preverbal elements detaching from the verb stem,
as with fe-di-ki-m-ê ‘I look at it’ (own data, Şirnak area) or se-di-kin-e ‘s/he stands’ (own data,
from Gevaş), where the initial syllables of originally Arabic roots are reanalysed as preverbs.
9
Here the term compound verb is employed in a pre-theoretical sense, regardless of whether or
not the given complex verb is considered to form a compound. See Haig (2002) for a discussion
of complex verbs in Kurdish.
479
Ergin Öpengin
yet clear. While a few such verbs are found to be used in both synthetic and analytic forms, such as ceribandin and cerebe kirin ‘to try’ (Ar. < ǧarraba), most verbs
are used in just one of the two forms. However, there is a great deal of dialectal
differentiation as to whether a verb is analytically or synthetically integrated.
Thus, the morphologically integrated verbs of most Northern Kurmanji dialects
such as emilandin ‘to use’ (dialectal Ar. ʕimil ‘to do’), şuẍulîn (Ar. šuɣl ‘work’),
fikirîn (Ar. fikr ‘thought’) are seen in the southeastern Badini dialect in synthetic
form, with a nominal base combining with a light verb, as emel kirin, şol kirin,
fikr kirin.
There are also various function words (discourse markers, conjunctions, adverbs) which are either borrowed from Arabic or developed in Kurdish based on
material borrowed from Arabic. Thus, the conjunction xeyrî (also seen as xeyr ji
and xêncî) ‘apart from, besides’ is based on Arabic ɣayr ‘other than’, while the adversative emā ‘but’ is dervied from Arabic ʔammā ‘however’. The similative şibî
(also şubhetî and şitî) is derived from the Arabic root √šbh ‘resemblance’. The
classifiers ḥeb (and the adverbial hebekî ‘a little’) and lib are derived from Arabic
ḥabb ‘grain(s)’ and lubb ‘kernels’, respectively. Finally, some discourse and verbal
adverbs resulting from Arabic sources are as follows: meselen ‘for instance’ and
helbet ‘of course’ are from Arabic maθalan and al-batta; in the eastern section
of the Badini dialect of Kurmanji, there is the use of the discourse marker seḥî
‘apparently, that means’, which is derived from the Arabic aṣaḥḥ ‘more correct’
– which separately exists in wider Kurdish as esseḥ ‘certainly’; while, finally, the
Arabic adjective qawī ‘strong’ has evolved into an adverb qewî ‘very; very much’
(though this is more literary than spoken).
All of these lexical borrowings illustrate matter transfer (in the sense of Matras
& Sakel 2007). In the following we have two instances of pattern transfer. First,
there is a particular adverbial form nema ‘no longer’, found only in the southeastern dialect of Kurmanji, spoken in the Mardin region of Turkey and Jazira
region of northeast Syria. This can be analysed as ne-ma, consisting of the negative prefix ne- and the past tense 3sg conjugation of the verb man ‘to stay’, as
in (15).
(15)
10
Southern dialect of Northern Kurdish (Media)10
nema
di-kar-im
veger-im
welêt
no.longer ind-be.able.prs-1sg return.prs.sbjv-1sg country.obl
‘I can no longer return to the homeland’
From a poem by an author from Syria, available online at: http://avestakurd.net/blog/2016/10/
26/romanivs-kurd-jan-dost-lal-b-ye-vdyo/ (accessed 31/01/2020).
480
21 Kurdish
There is an immediately-corresponding adverbial form mā ʕād ‘no longer’ in
Arabic, which is based on the negative form of the semantically similar verb ʕād
‘to return, keep doing’. This is obviously not a very recent development as it is
shared in the whole dialect area across country borders, but seemingly not so
deep either as to be shared by all Kurdish varieties, not even by all Northern
Kurdish dialects, further strengthening the particular status of the Jazira region
in Arabic–Kurdish language contact.
Second, there is a particular lexical construction bi X rabûn ‘to do; to complete;
to achieve’ in Northern Kurdish and hellsan be X in Central Kurdish, where X
stands for any activity or task (usually in the form of an infinitive verb). The
construction is based on the verb for ‘to rise, stand’ and a preposition in both
varieties, as illustrated in (16) and (17).
(16)
Central Kurdish (Media)11
polîs hellsa
be kokirdinewe-y zanyarî
police rise.pst.3sg with collecting-ez information
‘The police undertook (the task of) collecting information.’
(17)
Northern Kurdish (Media)12
Mîr Celadet (…) bi kar-ê
dewlet-ek-ê
rabû
Emir Celadet
with work-ez.m state-indf-obl.f rise.pst.3sg
‘Emir Celadet undertook the work of a state.’
This lexical construction also has a parallel in Modern Standard Arabic, based on
the verb qāma ‘to stand (up)’ and the preposition bi ‘with’, with the collocation
qāma bi meaning ‘to undertake’. This is obviously a recent influence on Kurdish,
as it is seen only in Iraq and Syria, and in a manner cross-cutting the broad variety
borders between Sorani and Kurmanji.
4 Conclusion
Contact with Arabic, which started in the early medieval period (approx. 7th–8th
centuries) with the arrival of Islam in the Near East, has had a profound impact
on Kurdish, particularly on its lexicon and phonology. Given the total absence
of any substantial previous study on the matter, the present chapter provides a
11
URL of article: http://www.kurdistan24.net/so/news/5ca67132-7a7f-4840-bfb4-dea5bf25ea2e
(accessed 31/01/2020).
12
URL of article: http://portal.netewe.com/mir-celadet-bedirxan-bi-tene-sere-xwe-bi-karedewleteke-rabu/ (accessed 31/01/2020).
481
Ergin Öpengin
first assessment of the influence of Arabic on Kurdish, primarily as represented in
Kurdish phonology and lexicon but also, albeit more restrictedly, in morphology
and syntax. Kurmanji Kurdish seems to be the variety that is most affected by
contact with Arabic, which is understandable considering the geographical continuity of the Kurdish and Arabic communities, especially in the historical Jazira
region and more widely in Upper Mesopotamia (in Mardin–Diyarbekir, Mosul–
Sinjar, and Haseke province). There are thus areas which show more intensive
Arabic influence within the speech zones of major Kurdish varieties, while the
outcomes of the contact reflect different layers in terms of time depth. Accordingly, the deeper-layer influence comes in the form of lexical convergence with
Arabic, sometimes through the intermediary of Persian and/or Ottoman Turkish.
This contact has repercussions in the expansion of the phonological inventory of
the language, and is shared across most Kurdish varieties. There are no unquestionably demonstrated changes in the morphosyntax resulting from contact with
Arabic at this layer. At the relatively shallower layer, the influence is mainly seen
in Syria and Iraq, and in the form of further expansion of the phonological inventory and a vocabulary heavily lexified by Arabic roots incorporated also into
the verbal domain. There are also several cases illustrating morphosyntactic and
lexicosyntactic change, such as the default gender assignment and word order in
complex noun phrases, as well as certain phrasal and adverbial lexical items.
In terms of “cognitive dominance”, in the sense of Van Coetsem (1988; 2000)
and Lucas (2015), in these instances of contact influence, the deeper-layer influence, which is restricted to, or related to, lexical borrowing, takes place with
the speakers being cognitively dominant in the recipient language, Kurdish. The
more recent instances of heavy lexification, and morphosyntactic and lexicosyntactic changes may, however, be the result of imposition, where the speakers are
dominant in the source language.
These outcomes may also be related to bilingualism and language configuration in historical perspective. That is, the absence of imposition (in the form
of morphosyntactic changes) in the deeper historical layer, and the restriction
of the influence to lexicon, point to the absence of widespread Arabic–Kurdish
bilingualism among the speakers of Kurdish at those historical stages. Some imposition of this kind is observed in the Kurmanji of the Jazira region, which is
known to have had greatest speaker contact between Kurdish and Arabic speech
communities. By contrast, the widespread bilingualism and Arabic-dominant linguistic configuration in Syria and Iraq for at least a century has led to instances
of imposition where the morphosyntactic and lexical patterns of Arabic are replicated in Kurdish. These outcomes are also mostly consonant with the predictions
of Van Coetsem’s (1988; 2000) “stability gradient”, which argues that lexicon is
482
21 Kurdish
less stable than syntax and phonology, which require dominance in the source
language in order to be affected by contact-induced change.
Given the limitations of a first attempt, much is yet to be explored regarding
Kurdish–Arabic language contact. In particular, the precise paths of development
of pharyngeals and emphatics in Kurdish should be analysed through fieldworkbased comparative dialect data, while, in the domain of lexicon, it is important
to analyse the morphophonological integration of borrowings into Kurdish. It is
also of interest to be able to develop diagnostics to disentangle direct Arabic influence on Kurdish from influence via other major languages such as Persian and
Ottoman Turkish. Finally, a detailed account of the history of Kurdish–Arabic
socio-political and cultural contact is required in order to complement the linguistic data and enable a more fine-grained analysis of the agentivity of contactinduced change in Kurdish.
Further reading
) Barry (2019) is a comprehensive and theoretically grounded treatment of the
introduction and further propagation of pharyngeal sounds in Kurdish.
) Chyet (2003) is the most comprehensive Kurdish–English dictionary, providing information on the source language of most loanwords in Kurdish, including those from Arabic.
) Tsabolov (1994) is the only work published so far on Arabic influence on the
grammar of Kurdish.
Acknowledgements
I would like to thank the editors of the volume and an anonymous reviewer for
their helpful and detailed feedback. Thanks also to Adam Benkato for his help
with Arabic data. Only I am responsible for any remaining shortcomings and
errors.
Abbreviations
Ar.
BCE
ca.
def
drct
Arabic
before Common Era
circa
definite
directional
dial.
ez
f
gen
Kr.
dialectal
ezāfe
feminine
genitive
Kurdish
483
Ergin Öpengin
ind
indf
inf
intr
ipfv
m
neg
nom
obl
pl/pl.
pn
indicative
indefinite
infinitive
intransitive
imperfective
masculine
negative
nominative
oblique
plural
proper noun
poss
prs
pst
pvb
sbjv
sec
possessive
present
past
preverb
subjunctive
secondary or pronominal
ezāfe/linking element
sg/sg. singular
tr
transitive
Tr.
Turkish
References
Anonby, Erik. 2020. Emphatic consonants beyond Arabic: The emergence and
proliferation of uvular–pharyngeal emphasis in Kumzari. Linguistics 58(1). 1–
54.
Barry, Daniel. 2019. Pharyngeals in Kurmanji Kurdish: A reanalysis of their
source and status. In Songül Gündogdu, Ergin Öpengin, Geoffrey L. J. Haig
& Erik Anonby (eds.), Current issues in Kurdish linguistics, 39–71. Bamberg:
University of Bamberg Press.
Blau, Joyce. 1989. Le kurde. In Rüdiger Schmitt (ed.), Compendium linguarum
iranicarum, 327–335. Wiesbaden: Reichert.
Bois, Thomas, Vladimir Minorsky & David N. MacKenzie. 2012. Kurds, Kurdistān.
In Peri Bearman, Thierry Bianquis, C. Edmund Bosworth, Emeri Johannes van
Donzel & Wolfhart P. Heinrichs (eds.), Encyclopaedia of Islam, 2nd edn (online).
Leiden: Brill.
Chyet, Michael L. 2003. Kurdish–English dictionary. New Haven, CT: Yale
University Press.
Coghill, Eleanor. 2016. The rise and fall of ergativity in Aramaic: Cycles of alignment change. Oxford: Oxford University Press.
Eberhard, David M., Gary F. Simons & Charles D. Fennig (eds.). 2019. Ethnologue:
Languages of the world. 22nd edn. Dallas, TX: SIL International.
Edwards, Bela B. 1851. Note on the Kurdish language. Journal of the American
Oriental Society 2. 120–123.
Folmer, Margaretha L. 1995. The Aramaic language in the Achaemenid period: A
linguistic study in variation. Leuven: Peeters.
484
21 Kurdish
Haig, Geoffrey L. J. 2002. Complex predicates in Kurdish: Argument sharing,
incorporation, or what?. STUF – Language Typology and Universals 55(1). 15–
48.
Haig, Geoffrey L. J. 2007. Grammatical borrowing in Kurdish (Northern group).
In Yaron Matras & Janet Sakel (eds.), Grammatical borrowing in cross-linguistic
perspective, 165–184. Berlin: Mouton de Gruyter.
Haig, Geoffrey L. J. 2011. Linker, relativizer, nominalizer, tense-particle: On the
ezafe in West Iranian. In Foong Yap, Karen Grunow-Hårsta & Janick Wrona
(eds.), Nominalization in Asian languages: Diachronic and typological perspectives, vol. 1: Sino-Tibetan and Iranian languages, 363–390. Amsterdam: John
Benjamins.
Haig, Geoffrey L. J. 2014. East Anatolia as a linguistic area? Conceptual
and empirical issues. In Lale Behzadi, Patrick Franke, Geoffrey L. J. Haig,
Christoph Herzog, Birgitt Hofmann, Lorenz Korn & Susanne Talabardon (eds.),
Bamberger Orientstudien, 13–36. Bamberg: University of Bamberg Press.
Haig, Geoffrey L. J. & Ergin Öpengin. 2014. Kurdish: A critical research overview.
Kurdish Studies 2(2). 99–122.
Haig, Geoffrey L. J. & Ergin Öpengin. 2015. Gender in Kurdish: Structural and
socio-cultural dimensions. In Marlis Hellinger & Heiko Motschenbacher (eds.),
Gender across languages, vol. 4, 247–276. Amsterdam: John Benjamins.
Haig, Geoffrey L. J. & Ergin Öpengin. 2018. Kurmanji Kurdish in Turkey:
Structure, varieties and status. In Christiane Bulut (ed.), Linguistic Minorities in
Turkey and Turkic-speaking minorities of the peripheries, 157–230. Wiesbaden:
Harrassowitz.
Ibrahim, Iman. 2015. Gender assignment to lexical borrowings by heritage speakers of Arabic. Western Papers in Linguistics / Cahiers linguistiques de Western
1(2).
James, Boris. 2007. Le «territoire tribal des Kurdes» et l’aire iraqienne (Xe-XIIIe
siècles): Esquisse des recompositions spatiales. Revue des mondes musulmans
et de la Méditerranée 10(117–118). 101–126.
Jastrow, Otto. 2011. Turkish and Kurdish influences in the Arabic dialects of
Anatolia. Türk Dilleri Araştırmaları 21(1). 83–94.
Kahn, Margaret. 1976. Borrowing and regional variation in a phonological description of Kurdish. Ann Arbor, MI: Phonetics Laboratory of the University of
Michigan.
Kent, Roland. 1953. Old Persian. 2nd, revised edn. New Haven, CT: American
Oriental Society.
485
Ergin Öpengin
Lentin, Jérôme. 2012. Du malheur de ne parler ni araméen ni kurde: Une complainte en moyen arabe de l’évêque chaldéen de Siirt en 1766. In Johannes Den
Heijer, Paolo La Spisa & Laurence Tuerlinck (eds.), Autour de la langue arabe:
Etudes présentées à Jacques Grand’Henry à l’occasion de son 70e anniversaire,
229–251. Louvain-la-Neuve: Institut Orientaliste.
Lucas, Christopher. 2015. Contact-induced language change. In Claire Bowern &
Bethwyn Evans (eds.), The Routledge handbook of historical linguistics, 519–536.
London: Routledge.
MacKenzie, David N. 1961. Kurdish dialect studies. Vol. 1. Oxford: Oxford
University Press.
Matras, Yaron. 2010. Contact, convergence, and typology. In Raymond Hickey
(ed.), The handbook of language contact, 66–85. Oxford: Wiley-Blackwell.
Matras, Yaron & Andrew Koontz-Garboden. 2016. The dialects of Kurdish. Manchester: University of Manchester. http://kurdish.humanities.manchester.ac.uk/.
Accessed 31/01/2020.
Matras, Yaron & Jeanette Sakel. 2007. Investigating the mechanisms of pattern
replication in language convergence. Studies in Language 31. 829–865.
Noorlander, Paul M. 2014. Diversity in convergence: Kurdish and Aramaic
variation entangled. Kurdish Studies 2(2). 201–224.
Öpengin, Ergin. 2012. Sociolinguistic situation of Kurdish in Turkey: Sociopolitical factors and language use patterns. International Journal of the Sociology of
Language 2012(217). 151–180.
Öpengin, Ergin. 2016. The Mukri variety of Central Kurdish: Grammar, texts, and
lexicon. Wiesbaden: Reichert.
Öpengin, Ergin. Forthcoming. The history of Kurdish and the development of
literary Kurmanji. In Hamit Bozarslan, Cengiz Gunes & Veli Yadirgi (eds.),
Cambridge history of the Kurds. Cambridge: Cambridge University Press.
Öpengin, Ergin & Geoffrey L. J. Haig. 2014. Regional variation in Kurmanji: A
preliminary classification of dialects. Kurdish Studies 2(2). 143–176.
Paul, Ludwig. 2008. Kurdish language i: History of the Kurdish language. In
Ehsan Yarshater, Mohsen Ashtiany & Mahnaz Moazami (eds.), Encyclopædia
iranica, online edn. New York: Encyclopædia Iranica Foundation.
Perry, John R. 2005. Lexical areas and semantic fields of Arabic loanwords in
Persian and beyond. In Éva Ágnes Csató, Bo Isaksson & Carina Jahani (eds.),
Linguistic convergence and areal divergence: Case studies from Iranian, Semitic
and Turkic, 97–110. London: Routledge.
Perry, John R. 2012. New Persian: Expansion, standardization, and inclusivity. In
Brian Spooner & William L. Hanaway (eds.), Literacy in the Persianate world:
486
21 Kurdish
Writing and the social order, 70–94. Philadelphia: University of Pennsylvania
Museum of Archaeology & Anthropology.
Ripper, Thomas. 2012. Diyarbekir Merwanileri: İslami ortaçağ’da bir Kürt
hanedanı. Istanbul: Avesta. Translated by Bahar Şahin Fırat. Originally published in 2000 by Ergon.
Ryding, Karin C. 2014. Arabic: A linguistic introduction. Cambridge: Cambridge
University Press.
Sirkeci, Ibrahim. 2005. War in Iraq: Environment of insecurity and international
migration. International Migration 43(4). 197–214.
Teyran, Feqiyê. 2011. Dîwana Feqiyê Teyran. Seîd Dêreşî (ed.). Diyarbakir: Lîs.
Todd, Terry L. 2002. A grammar of Dimili (also known as Zaza). 2nd edn. Stockholm: Iremet.
Tsabolov, Ruslan. 1994. Notes on the influence of Arabic on Kurdish. Acta Kurdica
1. 121–124.
Utas, Bo. 2005. Semitic in Iranian: Written, read and spoken language. In Éva
Ágnes Csató, Bo Isaksson & Carina Jahani (eds.), Linguistic convergence and
areal divergence: Case studies from Iranian, Semitic and Turkic, 65–77. London:
Routledge.
Van Coetsem, Frans. 1988. Loan phonology and the two transfer types in language
contact. Dordrecht: Foris.
Van Coetsem, Frans. 2000. A general and unified theory of the transmission process
in language contact. Heidelberg: Winter.
Zeyneloğlu, Sinan, Ibrahim Sirkeci & Yaprak Civelek. 2016. Language shift among
Kurds in Turkey: A spatial and demographic analysis. Kurdish Studies 4(1). 25–
50.
487
Chapter 22
Northern Domari
Bruno Herin
Inalco, IUF
This chapter provides an overview of the linguistic outcomes of contact between
Arabic and Northern Domari. Northern Domari is a group of dialects spoken in
Syria, Lebanon, Jordan and Turkey. It remained until very recently largely unexplored. This article presents unpublished first-hand linguistic data collected in
Lebanon, Syria, Jordan and Turkey. It focuses on the Beirut/Damascus variety, with
references to the dialects spoken in northern Syria and southern Turkey.
1 Current state and historical development
Domari is an Indic language spoken by the Doms in various countries of the
Middle East. The Doms are historically itinerant communities who specialize
in service economies. This occupational profile led the lay public to call them
the Middle Eastern Gypsies. Common occupations are informal dentistry, metalwork, instrument crafting, entertainment and begging. Most claim Sunni Islam
as their religion, with various degrees of syncretic practices. Although most have
given up their semi-nomadic lifestyle and settled in the periphery of urban centres, mobility is still a salient element in the daily lives of many Doms.
The ethnonym Dom is mostly unknown to non-Doms, who refer to them with
various appellations such as nawar, qurbāṭ or qarač. The Standard Arabic word
ɣaǧar for ‘Gypsy’ is variably accepted by the Doms, who mostly understand
with this term European Gypsies. All these appellations are exonyms and the
only endonym found across all communities is dōm. Only the Gypsies of Egypt,
it seems, use a reflex of ɣaǧar to refer to themselves.
From the nineteenth century onwards, European travellers reported the existence of Domari in the shape of word lists collected in the Caucasus, Iran, Iraq and
the Levant (see Herin 2012 for a discussion of these sources). The first full-length
Bruno Herin. 2020. Northern Domari. In Christopher Lucas & Stefano Manfredi
(eds.), Arabic and contact-induced change, 489–509. Berlin: Language Science Press.
DOI:10.5281/zenodo.3744543
Bruno Herin
grammatical description of a dialect of Domari is by Macalister (1914), who described the dialect spoken in Palestine in the first years of the twentieth century.
At present, the language is known to be spoken in Palestine, Jordan, Lebanon,
Syria and Turkey. No recent account can confirm that it is still spoken in Iraq and
Iran. There are roughly two dialectal areas: Southern Domari, spoken in Palestine and Jordan, and Northern Domari, spoken in Lebanon, Syria and southern
Turkey. This geographical division is not clear cut, as I have recorded speakers of
Southern varieties in Lebanon and speakers of Northern dialects in Jordan. The
main isogloss separating these two groups is the maintenance of a two-way gender system. Southern dialects have maintained the gender distinction, whereas
it has mostly disappeared in the north. Compare Northern gara ‘(s)he went’ vs.
Southern gara ‘he went’ and garī ‘she went’. These are sufficiently different to
allow us to posit an early split. Mutual intelligibility appears to be very limited.
A case in point is kinship terminology, which is largely divergent in both groups.
Within Northern Domari, the Beirut/Damascus dialect stands out because of the
glottal realization [ʔ] of etymological /q/ and the loss of the differential subject
marker -ən.
No general statement can be made about language endangerment. Jerusalem
Domari is reported to have only one fluent speaker left (Matras, this volume),
but the presence of speakers of Palestinian Domari in other places may not be
excluded. Young fluent speakers of Southern dialects are easy to find in Jordan. As
far as Northern Domari is concerned, the language is no longer transmitted to the
young generation in Beirut but it is in Damascus. In northern Syria, intergenerational transmission is quite solid. The situation in southern Turkey is, according
to some consultants, more precarious, but I have personally witnessed quite a
few children fully conversant with the language. In any case, bilingual Doms acquire both Domari and Arabic in early childhood, making both languages equally
“dominant” in Van Coestem’s (1988; 2000) terms.
Many Dom groups are also found in Eastern Anatolia. These groups have
shifted to Kurdish but maintained an in-group lexicon based on Domari, locally
called Domani. According to what I could personally observe on the ground and
what well-informed local actors reported to me, full-fledged Domari is not spoken beyond Urfa. East of Urfa, the shift to Kurdish is complete and even the
in-group lexicon is only remembered by elderly individuals.
There are no reliable figures on the number of speakers of Domari. The language has often been mistaken for a variety of Romani but this claim has no
linguistic grounds, except that they are both classified as Central Indo-Aryan
Languages with a possible Dardic adstrate.
490
22 Northern Domari
2 Contact languages
Besides a Central Indic core and a Dardic adstrate, the language exhibits various
layers of influence. Easily identifiable sources of contact are Persian, Kurdish,
Turkish and finally Arabic. This suggests, quite logically, that the ancestors of
the Doms left the Indian subcontinent, and then travelled into Persian-speaking
lands, before reaching Kurdish- and Turkish-speaking areas (most probably in
eastern Anatolia), before venturing into Arab lands. It is striking to see that the
Iranian and Turkic elements in Domari are not uniform across Northern and
Southern varieties, which suggests an early split in eastern Anatolia between
speakers of both groups. The impact of Arabic is also not uniform both across
Southern and Northern Domari, nor even within Northern Domari. What this
means is that the validity of any discussion of the Arabic component in Domari
is limited to the varieties considered.
The Beirut/Damascus dialect is undoubtedly the most Arabized one within the
Northern group, pointing to an earlier settlement of the community in an Arabicspeaking environment. Bilingualism (Domari–Arabic) is general in Lebanon and
Syria. Except perhaps for very young children who have not yet acquired any
other language, monolinguals in Domari are not to be found.
As far as Turkey is concerned, trilingualism in Domari, Turkish and Kurdish
is not uncommon, especially in southern Turkey around Gaziantep. In Hatay
province, many speakers above the age of forty are trilingual Domari–Arabic–
Turkish. The generations born here in the eighties onwards did not acquire Arabic.
According to personal recollection from various consultants, the community
of Beirut/Damascus used to spend the winter in Lebanon, and would go back to
Damascus in the summer. This semi-nomadic way of life seems to have stopped
when the civil war in Lebanon began. Although movements between Beirut and
Damascus remained frequent, this phenomenon ceased to be seasonal. In Damascus, they settled in the area of Sayyida Zaynab, in the suburbs of the city, and
in Beirut many of them settled in Sabra. Since the civil war started in Syria, virtually all the Damascus community have moved to Lebanon and settled in refugee
camps in the Bekaa Valley close to the Syrian border.
491
Bruno Herin
3 Contact-induced changes in Northern Domari
As noted above, Domari speakers in Lebanon and Syria are also fully proficient in
Arabic, to the point that I have never encountered or heard of any monolingual
adult. The Dom community, although largely endogamous and socially isolated,
cannot afford monolingualism, primarily because of their peripatetic profile. As
far as one can judge, their proficiency in Arabic is that of any monolingual native speaker of Arabic. Their pronunciation, however, is often not fully congruent with the local dialect spoken in the immediate vicinity of their settlements.
This is, as usual, due to the variety of inputs and migration after acquisition.
The Doms of Beirut for instance, do not speak Beirut Arabic and their speech
is immediately perceived as Syrian by Lebanese because they do not raise /ā/.
Raising of /ā/ towards [eː] is the hallmark of Lebanese Arabic in perceptual dialectology. Proficient speakers of Domari all exhibit Arabic–Domari bilingualism.
On the whole, there is a general license to integrate any Arabic lexeme in Domari speech, even when a non-Arabic morpheme exists. Code-switching is also
very common and there seems to be no conservative ideology about linguistic
practices, leading to a very permissive environment for language mixing.
3.1 Phonology
All the segmental phonology of Arabic has made its way into Domari. Arabic
stands out cross-linguistically because of its series of back consonants such as
the pharyngeals /ḥ/ and /ʕ/, the post-velars /q/, /ḫ/ and /ɣ/, and a set of velarized
consonants whose number varies from dialect to dialect. Typically, sedentary
varieties in the Levant minimally exhibit contrast between /ḍ/, /ẓ/, /ṭ/ and /ṣ/. In
Domari, the pharyngeals /ḥ/ and /ʕ/ are commonly found in loans from Arabic:
ḥḍər h- ‘watch’ (from Levantine Arabic ḥiḍir ‘he watched’). The same goes for /ʕ/:
ʕammər kar- ‘build’ (from Arabic ʕammar ‘he built’). An oddity surfaces in the
word for coffee, realized ʔaḥwa from Arabic ʔahwe. These pharyngeals are also
common in Kurdish-derived items such as ḥazār ‘thousand’, moʕōri ‘ant’ and also
in the inherited (Indic) stock in ʕaqqōr ‘nut’. Post-velar /q/, /ḫ/ and /ɣ/ are found
in all the layers of the language: qāla ‘black’ (inherited), qāpī ‘door’ (Turkish),
sāɣ ‘alive’ (Kurdish), ɣarīb ‘strange’ (Arabic). The most striking innovation of
the Beirut/Damascus dialect is the glottal realization [ʔ] of /q/: ʔər ‘son’ (< qər),
ʔāyīš ‘food’ (< qāyīš). This innovation is very likely contact-induced because it is
commonly found in the Arabic dialects of both Damascus and Beirut and beyond.
Velarized consonants mostly surface in the Arabic-derived stock as in naḍḍəf
kar- ‘clean’ (< Arabic naḍḍaf ‘he cleaned’), but also in pre-Arabic items: ḍāwaṭ
492
22 Northern Domari
‘wedding’ (borrowed from Kurdish but ultimately from Arabic daʕwa ‘invitation’), pạ̄ṣ ‘at him’ (< Old Indo-Aryan pārśvá ‘side’). It is still unclear to what
extent velarization in Domari continues Indo-Aryan retroflexion (Matras 2012:
64). Domari also kept a contrast between /p/ and /b/, not found in Arabic: bīrōm
‘I feared’ vs. pīrōm ‘I drank’.
As far as vowels are concerned, Levantine Arabic exhibits either a two-way
distinction in the short vowel system (/a/ and /ə/) or a three-way distinction (/a/,
/i/ and /u/). In Northern Domari, only the two short vowels /a/ and /ə/ are contrastive: kərī ‘house’ vs. karī ‘pot’. Such a paucity of contrastive short vowels is
probably due to contact with Arabic varieties which exhibit a two-way system
(/a/ vs. /ə/), such as many Lebanese and Syrian dialects. Most Arabic dialects in
the area have a five-way system of long vowels because of the monophthongization of /ay/ and /aw/: /ā/, /ī/, /ū/, /ē/ and /ō/. In addition to these long vowels,
Domari displays another contrast between /ā/ and a back /ạ̄/ (IPA [ɑː]): māsī
[maːsiː] ‘meat’ (< Old Indo-Aryan māṁsá) vs. mạ̄s-ī [mɑːsiː] ‘month-pl’ (< Old
Indo-Aryan mā́sa).
Domari has also preserved distinct suprasegmental features, such as final syllable stress assignment. Arabic-derived items are fully integrated into this pattern
and bear final primary stress, whether common nouns or proper nouns: Domari
[faːˈdja] vs. Arabic [ˈfaːdja] (personal name Fādya). An interesting phenomenon is
that Arabic epenthetic vowels in final-syllable position are reinterpreted as plain
vowels and bear primary stress. Compare Domari [sˤaˈʕab] and Arabic [ˈsˤaʕəb]
‘difficult’; Domari [waˈdˤaʕ] and Arabic [ˈwadˤəʕ] ‘situation’.
3.2 Morphology
Northern Domari has not borrowed any derivational or inflectional morphemes
from Arabic. This is of course due to the fact that Arabic morphology is mostly
non-concatenative. Borrowed morphology mostly comes from Kurdish and Turkish, whose morpheme segmentation is much more transparent. These borrowed
morphemes must have entered Domari when Kurdish and Turkish were contact
languages of Domari. A case in point is the Kurdish diminutive -ək, which has
made its way into all layers of the lexicon: panč-ək ‘tail’, ḫar-ək ‘bone’ (both
Indic), taḫt-ək ‘wood’, qannīn-ək ‘bottle’ (both derived from Arabic: taḫt ‘bed’
and qannīne ‘bottle’). The dialects of northern Syria and southern Turkey have
also borrowed from Kurdish the comparative suffix -tar, the Turkish conditional
marker -sa and the Turkish superlative marker ān. These constructions are not
available in the Beirut/Damascus dialect, which relies entirely on Arabic-derived
material. Compare the translation of the Arabic sentence inte aḥsan minni ‘you
are better than me’ into Sarāqib Domari (1) and Beirut/Damascus Domari (2):
493
Bruno Herin
(1)
Sarāqib Domari
tō dēšōm bḫēz-tar ištōre
2sg 1sg.abl good-cmpr cop.2sg
‘You are better than me.’
(2) Beirut/Damascus Domari
tō aḥsan wēšōm ištōr
2sg better 1sg.abl cop.2sg
‘You are better than me.’
Sarāqib is located in northern Syria and the dialect spoken by the Doms of
Sarāqib is a good representative of the Domari of northern Syria and southern
Turkey. Three differences are immediately apparent. The first is morphological,
whereby there are different forms for the ablative of the first-person pronoun.
The second difference is syntactic: in (1) the standard of comparison precedes the
comparative adjective (dēšōm bḫēz-tar) and in (2) it follows it (aḥsan wēšōm).1 The
Beirut/Damascus Domari syntax exhibits full congruence with the Arabic syntax.
The third difference is lexical. Because Beirut/Damascus Domari does not have
at its disposal the morpheme -tar, speakers are obliged to draw on Arabic for
the comparative. This phenomenon, labelled “bilingual suppletion” by Matras, is
described at length for Jerusalem Domari (Matras 2012: 379–382; see also Matras,
this volume: §3.5).
Beirut/Damascus Domari also relies entirely on Arabic material for the expression of time and date, as shown in (3). In northern Syria, speakers favour the use
of inherited numerals, as exemplified in (4).
(3) Beirut/Damascus Domari
sabʕa tmāne ōtanta sa
ʕašra la s-sāʕa
mi-s-sāʕa
mānane
stay.ipfv.1pl from-det-hour ten to def-hour seven eight there all
čāɣ-an-sa
children-obl.pl-com
‘We stay there with the all the kids from ten o’clock to seven or eight
o’clock.’
(4)
1
Sarāqib Domari
čōrt-ə-ma
ē
ḥatta saʕat štār ēwar mānde
until hour four evening stay.pfv.3sg dem.obl wasteland-obl-in
‘He stayed until 4pm in this wasteland.’
Comparative constructions typically involve two noun phrases (NPs). Stassen (2013) labels the
object of comparison the “comparee NP” and the other the “standard NP”.
494
22 Northern Domari
Some speakers of Beirut/Damascus Domari also extend the use of Arabic to
higher numerals because, according to their own judgment, they have difficulties retrieving the pre-Arabic options. A look at their distribution reveals that
the main parameter that triggers the use of Arabic items is not so much high
numerals, but rather the complexity of the numeral. Compare in this regard (5)
and (6). In (5), the speaker uses Arabic for the more complex numeral ‘95000’ but
uses Domari items for simpler ‘2000’, ‘3000’ and ‘4000’.
(5)
Beirut/Damascus Domari
pārda
abōs
šaʔʔ-āka ši
ḫamse u tisʕīn alf
dolar
buy.pfv.3sg 3sg.ben flat-indf about five and ninety thousand dollar
‘He bought a flat for her, about ninety-five thousand dollars.’
(6) Beirut/Damascus Domari
načīš-a-ki
dī ḥazār
trən ḥazār
štār ḥazār
dfaʕ
dancing-obl-abl two thousand three thousand four thousand pay
kaštand
dādōs
kē
do.prog.3pl her.mother ben
‘They give two, three, four thousand (dollars) to her mother from
dancing.’
As noted above, it appears that the use of Arabic numerals is closely linked to
language dominance. Speakers themselves are aware of it and when asked why
they do not use Domari numerals, they justify it claiming a lack of proficiency.
Looking at the distribution of inherited and Arabic numerals is therefore a good
way to assess whether language attrition is incipient or not.
The impact of Arabic is also apparent in some morphological differences between the Beirut/Damascus variety and the dialects of northern Syria. For instance, the verb sək- means ‘to learn’. The Beirut/Damascus dialect adds the passive suffix -yā/-ī. The corresponding verb in Arabic tʕallam is marked with the
valency-decreasing prefix t-. What the speakers of the Beirut/Damascus dialect
have done is to replicate the valency-decreasing prefix t- of tʕallam by means
of the Domari passive suffix yā/-ī : skə-rd-ōm (learn-pfv-1sg; northern Syria) vs.
sk-ī-r-ōm (learn-pass-pfv-1sg; Beirut/Damascus) ‘I learnt’.
Unlike Southern Domari, Northern Domari does not normally transfer Arabic
plurals. Speakers simply use the singular form and add the Domari plural marker
-ī(n): azʕar-īn ‘thugs’ instead of the Arabic plural zuʕrān. Arabic plurals do surface at times, but only when they exhibit a high degree of independence within
the lexicon. Examples are ʔarāyb-ē-mā (relatives-pl-1pl) ‘our relatives’, ǧīrān-ēmā (neighbors-pl-1pl) ‘our neighbors’, from Arabic qarāyib and ǧīrān. Although
495
Bruno Herin
these items have singular forms (respectively qarīb and ǧār), they are arguably
lexicalized plurals and independent entries in the Arabic lexicon.
3.3 Syntax
3.3.1 Constituent order
The impact of Arabic in the realm of syntax is not uniform across Domari dialects. Dialects of northern Syria and southern Turkey show a strong tendency towards a head-final constituent-order typology, both within the NP and the clause.
This feature is areal, so its presence in Domari may well be contact-induced.
The canonical syntax of the NP is (demonstrative) (numeral) (adjective) (noun)
noun. Complex NPs could only be retrieved through elicitation (examples (7) to
(10)) and hardly occur in spontaneous speech. Example (7) illustrates the canonical syntax, where all the modifiers appear to the left of the head. Speakers of
Beirut/Damascus Domari, however, tend to dislocate some modifiers to the right
of the head, converging towards the Arabic syntax, as in (8), (9) and (10).
(7)
Sarāqib Domari
ē
štār lāfty-ən-ki
dād-ō-sā
dem four girl-obl.pl-abl mother-sg-3pl
‘the mother of these four girls’
(8)
Beirut/Damascus Domari
dād-ō-sā
štār lāfty-an-ki
mother-sg-3pl four girl-obl.pl-abl
‘the mother of the four girls’
(9)
Beirut/Damascus Domari
nām-ē-sā
ǧəwr-an-ki
tərn-an-ki
name-pl-3pl woman-obl.pl-abl three-obl.pl-obl
‘the names of the three girls’
(10) Beirut/Damascus Domari
dōm-an-sa
ēr-an-sa
štār-an-sa
dom-obl.pl-com dem-obl.pl-com four-obl.pl-com
‘with these four Doms’
In (9), the speaker also dislocates to the right the numeral trən ‘three’ which
normally appears to the left giving the expected order trən ǧəwr-an-ki nām-ēsā (three woman-obl.pl-abl name-pl-3pl). The numeral remains unmarked for
496
22 Northern Domari
case when it appears to the left of the head. When it is placed to the right, it agrees
in case with the head. This is also the case with the demonstrative in (10). Here the
normal order would be ē štār dōm-an-sa (dem four Dom-obl.pl-com). The fact
that speakers replicate case marking on right-dislocated modifiers suggests that
they feel the need to strengthen constituency in case of non-canonical syntax.
The influence of Arabic also surfaces in the Beirut/Damascus dialect in the
syntax of the quantifier sa ‘all’. This is normally located to the right of the head:
ammat sa ‘all the people’ (‘people all’). In Beirut/Damascus, sa consistently surfaces to the left, like the Arabic quantifier kull: sa ammat (Arabic kull in-nās).2
3.3.2 Internal object
Domari speakers regularly replicate Arabic constructions and idioms, but tend to
do so by recruiting inherited or pre-Arabic material – they do not borrow Arabic
material. For instance, all dialects have replicated the so-called internal object
construction, commonly used in Arabic as a predicate-modifying construction.
Consider for instance (11) in Jordanian Arabic, where the speaker narrows the
scope of the predication using the verbal noun ʕirəf ‘knowledge’, derived from
the verb ʕirif ‘he knew’, and modifies it with the adjective ṭayyib ‘good’. In (12),
the speaker has used the deverbal derivation kūš from the root kū- ‘throw’ and
coded it as an object, as evident from the accusative marker -əs. This replicates
the Arabic internal object construction as illustrated in (11).
(11)
Jordanian Arabic
baʕrif-hum
ʕirəf
ṭayyib
know.impf.1sg-3pl knowledge good
‘I know them well.’
(12)
Sarāqib Domari
dād-ōs
ibnḥarām e ē
kūš-əs
mother-3sg son.of.illicit cop dem throwing-acc
ktōs-s-e
throw.pfv.3sg-obj.3sg-prs
‘His mother is heartless for having thrown (her baby) in such a way.’
3.3.3 Impersonal construction
Speakers also replicate the Arabic impersonal construction with the indefinite
pronoun il-wāḥad by way of the inherited noun mānəs ‘individual, people’. Exam2
Arabic kull can also appear to the right as in in-nās kull-ha ~ kull-hum ‘all the people’ but this
is a marked syntax.
497
Bruno Herin
ple (13) illustrates the use of il-wāḥad in (Jordanian) Arabic. In (14), the sequence
gzare māns-as corresponds to Arabic biʕiḍḍ il-wāḥad, literally ‘it bites one’. The
fact that māns-as replicates il-wāḥad is also apparent from the accusative marking in Domari, which normally surfaces only with definite objects. The referent
here is by nature indefinite and non-referential, so accusative marking in Domari
can only be explained by the presence of the definite article il- in Arabic il-wāḥad.
(13)
Jordanian Arabic
kān
ʕēb
il-wāḥad yrūḥ
ʕala ʔutēl
be.prf.3sg.m shameful def-one go.impf.sbjv.2sg.m to hotel
‘One was ashamed to spend the night in a hotel.’
(14) Beirut/Damascus Domari
ašti ši hana lli baḥr-a-ma e gzare
māns-as
exs too dem rel sea-obl-in cop bite.ipfv.3sg man-acc
‘There is this thing in the sea, it bites you.’
3.3.4 Auxiliaries
Probably the most striking difference between Southern and Northern Domari
as far as the Arabic component is concerned is the absence of Arabic inflected
material in the latter. Only the dialect of Beirut/Damascus has borrowed the auxiliaries kān (with its imperfect form bikūn), ṣār and ḫalli.
(15)
Beirut/Damascus Domari
lāfty-a kē bḫēr
ǧahhəz lakand
ṣār
become.prf.3sg prepare do.sbjv.3pl girl-obl ben well
‘They prepare the girl well now (for the wedding).’
(16) Beirut/Damascus Domari
ḫaḍra kān
məǧnār-a
Khadra be.prf.3sg.m breastfeed.ipfv.3sg-pst
‘Khadra was breastfeeding.’
(17) Beirut/Damascus Domari
āwande
bikūn
krēnde
mā kē kyāmōr
come.ipfv.3pl be.impf.3sg do.prf.3pl 1sg ben something
‘(My kids) would come and they would have done something (naughty).’
In (15), the subject is in the 3pl but ṣār remains invariable, as the 3pl is ṣāru.
In (16), the subject is feminine so if there was agreement one would expect kānat,
498
22 Northern Domari
not masculine kān. A further intriguing feature in (16) is the redundancy in past
marking, first with kān and second with the past suffix -a, which in northern
Syria and southern Turkey Domari suffices to mark past tense. The same invariability is apparent in (17) where the 3pl of bikūn should be bikūnu. These
auxiliaries have the same semantic load as in Arabic. The morpheme ṣār puts
emphasis on the inception of the event, kān followed by the imperfect places the
event in the past and gives it an iterative/habitual aspect and bikūn describes a
possible state of affairs not attested at the time of utterance. Arabic ṣār, kān and
bikūn are absent in the dialects of northern Syria and southern Turkey. The only
auxiliary that has been replicated here is ṣār. These dialects, however, have only
replicated the structure, not the substance, that is they rely on inherited morphemes, as exemplified in (18). The speaker simply translates Arabic ṣār with
the Domari equivalent hra, replicating the Arabic structure ṣār + subjunctive
(see Manfredi, this volume). A further difference is word order, with the verb
placed clause-finally in the subordinate clause.
(18)
Sarāqib Domari
wārsīndạ lwār
hər
hit.sbjv.3sg
become.pfv.3sg rain
‘It started raining.’
As noted above, in these dialects the functions of Arabic kān are expressed by
the inherited past suffix -a. The functions covered by Arabic, bikūn, however do
not seem to be encoded in the grammar of these dialects.
In Levantine Arabic, the imperative form ḫalli ‘let’ of ḫalla ‘he let’ is often used
to soften an order and allows the speaker to avoid using an imperative, flagging
a suggestion or an invitation, as shown in (19):
(19) Jordanian Arabic
ḫalli ibn-ak
yrūḥ
la ǧ-ǧēš
let son-2sg.m go.impf.3sg.m to def-army
‘Let your son serve in the army.’
This auxiliary has been borrowed into Beirut/Damascus Domari with the exact
same function, as illustrated in (20). In this case too, ḫalli remains invariable and
does not surface as ḫallī-(h)un (let.imp.2sg-3pl) as it would in Beirut/Damascus
Arabic. Here again, the dialects of northern Syria and southern Turkey have borrowed the structure, but not the substance, and use the inherited root mək ‘let’,
as exemplified in (21).
499
Bruno Herin
(20)
Beirut/Damascus Domari
ḫalli ǧānd
dfən lakrand-əs
let go.sbjv.3pl bury do.sbjv.3pl-3sg
‘Let them go and bury him.’
(21)
Aleppo Domari
mək pāʋər
pạ̄sōr
let come.sbjv.3sg 2sg.ad
‘Let him come to your place.’
3.3.5 Negation
Only two Arabic negators have made their way into the grammar of Northern Domari: Damascus Arabic mū and the contrastive negative coordination
markers lā...walā ‘neither…nor’. Arabic mū is only available in the dialect of
Beirut/Damascus. Its distribution and functions, however, only partially match
those of Damascus Arabic. The primary function of mū in Damascus Arabic is
to negate non-verbal predicates. This is not attested in Domari, which relies for
this purpose only on inherited nye. mū surfaces first when negation has scope
over non-clausal constituents, as shown in (22), and second when the predicate
is in a non-indicative mood (subjunctive, jussive and imperative) as in (23):
(22)
Beirut/Damascus Domari
wāšya mū wāšōm
səff (h)ra
side become.pfv.3sg 3pl.com neg 1sg.com
‘He took sides with them, not with me.’
(23)
Beirut/Damascus Domari
biǧūz mū māntyar
wāš
məṣrī
possible neg stay.sbjv.3sg 3sg.com money
‘He might not have any money left.’
The Arabic structure lā…walā is readily available in all varieties, but whereas
it is the only option in Beirut/Damascus, it competes with the inherited structure
nə…nə in northern Syria and southern Turkey. Interestingly, this clash has led to
a mixed form nə…walā, as shown in (24). The Domari syntax is also reminiscent
of the Turkish possessive predication syntax with possessive marking on the
noun and an existential morpheme.
500
22 Northern Domari
(24)
Antioch Domari (southern Turkey)
nə lawr-ōs ašti wala šarš-ōs ašti
neg tree-3sg exs neg root-3sg exs
‘It doesn’t grow on a tree nor has it roots.’
3.3.6 Complex sentences
Complex sentences minimally include coordinated and subordinate clauses. The
Arabic coordinators w ‘and’, aw ‘or’, walla ‘or’, bass ‘but’ and others have all
made their way into Domari. Originally, Domari seems to have distinguished
clausal coordination from phrasal coordination, a not so frequent feature from a
typological point of view. Nominal categories are coordinated with the Turkishderived morpheme la and clauses are coordinated with the Kurdish-derived enclitic -ši. The intrusion of Arabic w, which in Arabic is used indiscriminately for
both kinds of coordination, has led to the marginalization of the original system
in Beirut/Damascus Domari, which now tends to favour the use of Arabic w.
(25) Beirut/Damascus Domari
illi mangar
tōre
māṣṭ-a-ma
w illi mangar
rel want.ipfv.3sg put.ipfv.3sg yoghurt-obl-in and rel want.ipfv.3sg
ʔār-s-e
nāšif
eat.ipfv.3sg-obj.3sg-prs dry
‘Some eat it in yoghurt, some eat it dry.’
As far as phrasal coordination is concerned, some alternation between Arabic w and Turkish-derived la is still observed: dōmwārī w ṭāṭwārī ‘Domari and
Arabic’ ~ dōm la ʕarabi ‘Domari and Arabic’.
Virtually all the conjunctions of subordination found in Domari are borrowed
from Arabic. This includes the relativizer illi, the complementizer inno and potentially all the adverbial conjunctions found in Levantine Arabic: lamma ‘when’,
qabəl-mā ‘before’, baʕəd-mā ‘after’, ʕa-bēn-mā ‘by the time’, and many more. PreArabic constructions are attested for relativization and conditional clauses, but
these only survive in the dialects of northern Syria and southern Turkey, and
tend to be replaced by Arabic material (except in the varieties spoken in Turkey).
A case in point is conditional clauses. Arabic iza and law are available everywhere, even in Turkey, as shown in (26), recorded in Antioch. In this example,
the speaker uses the Arabic morpheme aza (< iza) in the first sentence of the utterance, and no overt marking in the protasis, making parataxis a possible means to
express condition. As far as counterfactual conditions are concerned, it appears
501
Bruno Herin
that the dialect of Beirut/Damascus is fully congruent with Arabic in having borrowed also the morpheme kān in both the protasis and the apodosis, as shown
in (27). The dialects of northern Syria and southern Turkey exhibit a native strategy using subjunctive mood and past marking in the protasis and perfective and
past marking in the apodosis. The two clauses are coordinated with the Kurdish
derived enclitic ši (28).
(26) Antioch Domari
aza kām karne
qāne
kām nə-karne
nə-qāne
if work do.ipfv.1pl eat.ipfv.1pl work neg-do.ipfv.1pl neg-eat.ipfv.1pl
‘If we work, we eat, (if) we don’t work, we don’t eat.’
(27) Beirut/Damascus Domari
law kān
nəčnār-sā
bāb-ōm kān
if be.prf.3sg make.dance.ipfv.3sg-obj.3pl father-1sg be.prf.3sg
abṣar
kaki (h)re
not.know what become.pfv.3sg
‘If my father had put them to dance, I don’t know what would have
happened.’
(28)
Sarāqib Domari
aḷḷ-əs
byātyənd-a
nə-ktēnd-s-a
ši
God-acc fear.sbjv.3pl-pst neg-throw.pfv.3pl-obj.3sg-pst and
‘Had they feared God, they would not have thrown him.’
3.4 Lexicon
3.4.1 Function words
Arabic prepositions do occur in Domari, but these are mostly non-core prepositions such as qabəl ‘before’, baʕad ‘after’, minšān ‘for’, ɣēr ‘other’. Some have
made their way into Domari only recently, and still alternate with pre-Arabic
options, such as the Iranian equative morpheme war, which tends to be replaced
by Arabic mitəl ‘as, like’ especially in the dialect of Beirut/Damascus. Currently,
war and mitəl are in a quasi-complementary distribution, with war being used
with full NPs and mitəl with pronouns, as shown below in (29) and (30):
(29) Beirut/Damascus Domari
tō ʔr-ōm war ištōr
you son-1sg like cop.2sg
‘You are like my son.’
502
22 Northern Domari
(30)
Beirut/Damascus Domari
tāni ʔər gēna mitl-ōs kām karre
second son also like-3sg work do.ipfv.3sg
‘My second son has the same job.’
The Arabic core preposition b- ‘in, with’ occurs in Domari, but it appears to be
restricted to certain constructions and idioms such as gāl b-gāl ‘discussion’ (word
in-word), ārāt əb-dīs ‘night and day’ (night in-day), b-rəbʕ-āk ‘for a quarter of a
pound’ (with-quarter-indf). The preposition min ‘from’ also sporadically occurs
in Beirut/Damascus Domari:
(31)
Beirut/Damascus Domari
min ši
šēš mạ̄s ǧərsa
krōm
ḍāwaṭ-ōs
from about six month wedding do.pfv.1sg wedding-3sg
‘Some six months ago I married him off.’
Domari also borrows high-frequency adverbs, fillers, connectors and all kinds
of discourse-structuring devices, such as masalan ‘for instance’, abadan ‘at all,
never’, yaʕni ‘I mean’, aywa ‘yes, so’, waḷḷa ‘I swear’, inno (complementizer and
discourse marker) and many more. One finds also common adverbial phrases
such as ṭūl in-nhār ‘all day long’, ṭūl il-waʔət ‘all the time’, and ʕala ṭūl ‘immediately’. The very common Domari phrase tīka tīka ‘slowly’ replicates Arabic
šwayy əšwayy.
3.4.2 Content words
In Syria and Lebanon, Arabic is the de facto lexical reservoir of Domari, so there
is a general licence to integrate any element from Arabic if no pre-Arabic option exists. The issue is the replacement of pre-Arabic options with Arabic material. There is of course a certain amount of variation in lexical knowledge across
speakers, but it seems possible to differentiate several levels of replaceability.
Some items have long been replaced by Arabic words, and only a handful of
speakers are able to retrieve them, such as lōrga ‘tomato’ or pīsənga ‘bulgur’, replaced respectively by Arabic bandōra and bərɣəl. Other items tend to be replaced
by Arabic equivalents but may still surface in the speech of some speakers, such
as čatīn ‘hard’, čirkī ‘bird’, alčāḫ ‘low’ replaced by Arabic ṣaʕab, ṭēr/ʕaṣfūr and
wāṭi. Some items seem stable but are sporadically replaced with Arabic-derived
items such as drəs kar- ‘study’ instead of inherited sək-. Finally, other items such
as ǧawwəz h- ‘get married’ and ǧirsāwī h- freely alternate. It appears therefore
that every pre-Arabic item is somewhere on a continuum of replaceability from
503
Bruno Herin
“very unlikely” to “completely disappeared”. To illustrate the variablility in replaceability judgment, I remember an elicitation session in Aleppo with a father
and his son. One of the sentences contained the Arabic word baṣal ‘onion’. The
son simply translated the sentence with the Arabic word baṣal but the father
strongly objected to this answer, stating that the proper Domari word is pīwāz.
As noted above, Arabic nouns are integrated in their singular form, except in
the case of lexicalized plurals. Adjectives are borrowed in their masculine form
and never agree in gender, as shown in (32). Other than the past copula a, all the
words in this example are Arabic. Two features, however, allow its identification
as Domari. First, ḥāla is realized without raising (also stressed on the last syllable [ħaːˈla]), unlike Levantine Arabic ḥāle, and second taʕbān does not agree in
gender with ḥāla and surfaces in its masculine form, instead of feminine taʕbāne,
as it would normally occur in Arabic.
(32)
Beirut/Damascus Domari
ʔabəl ḥāla
taʕbān a
Before situation tired cop.pst
‘Before, the situation was bad’
Arabic verbs are easily integrated into Domari, because Domari has a light
verb strategy. Roughly, transitive verbs tend to be integrated with the light verb
kar- ‘do’: rabbī kar- ‘raise’ from Arabic rabba, yrabbi ‘raise’. Intransitive verbs
are integrated with h- ‘become’: ʕīš h- ‘live’ from Arabic ʕāš, yʕīš ‘live’. While all
the verbs that are integrated with kar- are transitive, some verbs integrated with
h- are not intransitive: lməs (h)rōs-s-e ‘he has touched it’ (touch become.pfv.3sg3sg-prs) from Arabic lamas, yilmis ‘touch’. This seems to happen with transitive
verbs that are lower on the transitivity scale, or at least perceived to be so. In
the case of lamas, yilmis, its integration into Domari by way of the light verb
h- suggests that speakers perceive it as less transitive. Formally, speakers isolate
the imperfect stem of the verb, and apply a vocalism in /i/: nsī kar- ‘forget’ and
stannī kar- ‘wait’, from the Arabic imperfect stems of nsa ‘forget’ and stanna
‘wait’.3 An exception to this tendency occurs with the so-called hollow roots in
Arabic whose imperfect stem is CūC. In this case, speakers simply extract the
imperfect stem and leave it unchanged: zūr h- ‘visit’, dūr h- ‘turn’, ʕūz h- ‘need’,
from the Arabic imperfect stems zūr, dūr and ʕūz.
Some English-derived items were also recorded in the Beirut/Damascus dialect, such as mōmari ‘memory card’, hambarga ‘hamburger’ and, more surprisingly, tōmanǧīre ‘Tom and Jerry’ [toːmanʤiːˈre], expectedly stressed on the last
syllable.
3
These verbs are only available in Beirut/Damascus, other dialects use respectively ziwra karand akī kar-.
504
22 Northern Domari
3.4.3 Speech sample
Probably the best way to capture how Arabic integrates into Domari is to consider a piece of spontaneous speech, reproduced below in (33). It is part of a
recorded discussion I had with a consultant in her mid-thirties in Beirut. It illustrates the level of endangerment of Beirut/Damascus Domari. The consultant
belongs to the last generation of fluent speakers. Her children did not acquire
the language. According to what she reports, she was unable to speak to her
children in their early childhood because her husband, who is a semi-speaker
of Domari, prevented her from transmitting the language. Her daughter-in-law,
aged twenty-one at that time, is also a fluent speaker of Domari because she grew
up in Damascus, where language transmission was more solid than in Lebanon.
Both of them use Domari in the household. Her son reacts negatively when he
hears it, and even labels it aǧnabi ‘foreign, non-Arabic’. Linguistically, the text illustrates some of the features discussed above. Arabic-derived items are marked
in boldface.
(33)
Beirut/Damascus Domari
nā n-ǧib
karre
pānǧī gāl karre
gāl karre
no neg-tongue do.ipfv.3sg 3sg word do.ipfv.3sg word do.ipfv.3sg
dōm wāšōm mā gāl kame
wāšī
ʕādi bass əʔr-ōm ʔzīn
Dom 1sg.com 1sg word do.ipfv.1sg 3sg.com normal but son-1sg shout
karre
wat
ftyare
ma-gāl
ka
aǧnabí
do.ipfv.3sg 3sg.supr say.ipfv.3sg neg-word do.sbjv.2sg foreign
nə-fəmm
(h)ōme
watōr, gāl karse
neg-understand become.ipfv.3sg 2sg.supr word do.ipfv.2pl
ʕarabiy-a-ma yaʕni ma-gāl
k(a)
ēhānī laʔanno
Arabic-obl-in I.mean neg-word do.sbjv.2sg so
because
n-fəmm
(h)ōre
watī
bass mā l pānǧī ǧib
neg-understand become.ipfv.3sg 3sg.supr but 1sg and 3sg tongue
kane
ṭūl
il-waʔət kəry-a-ma
yaʕni iza mā l pānǧi
do.ipfv.1pl length def-time house-obl-in I.mean if 1sg and 3sg
štēn
kəry-a-ma
ṭūl
in-nhār gāl kane
dōm-a-ma
cop.1pl house-obl-in length def-day word do.ipfv.1pl Dom-obl-in
yaʕni ʔr-ōm wāri ʕəmr-ōs wāḥad u ʕišrīn sane akbar
I.mean son-1sg bride age-3sg.f one
and twenty year bigger
ʔr-ōm-ki
b-trən
wars mū ʕādi ʕādi nye
amīn lāzim
son-1sg-abl with-three year neg normal normal cop.neg 1pl must
505
Bruno Herin
lpāran
azɣar wēšōma bass bxēz e u ādami e u
take.sbjv.1pl smaller 1pl.abl but good cop and humane cop and
kē pārdōm-əs
ʔr-ōm
maḥšūm e mā ēhāny-a xr-a
respectful cop so so-obl heart-obl ben take.pfv.1sg-obj.3sg son-1sg
kē u ǧamāʕt-ēm kē skīr(a)
ēta baʕdēn skīra
ben and folks-1sg ben learn.pfv.3sg here then learn.pfv.3sg
mahná
baʕdēn kām əkra
wars-ā
wars-ā
nīm makanīk
profession then work do.pfv.3sg year-indf year-indf half mechanic
baʕdēn wəndrārda u īsa nə-kām kištar
wala kkyā
then fire.pfv.3sg and now neg-work do.prog.3sg nor thing
wēsre
kəry-a-ma
sit.pfv.3sg house-obl-in
‘No, [my son] doesn’t speak [Domari], [my daughter-in-law] does, she
speaks with me, I speak with her normally but my son shouts at her and
tells her: “Don’t speak foreign, I don’t understand you, you all speak in
Arabic, don’t speak like this”, because he doesn’t understand her. But me
and her we speak all the time in Domari, that is, if both of us are in the
house, all day long we speak in Domari. The bride of my son, she is
twenty-one years old, three years older than my son, it’s not usual, we
[women] have to take someone older, but she is a good person, humane
and respectful. That’s why I took her for my son and my family. [My son]
studied here [in the school]. After that he went for vocational training
and worked for a year a year and a half as a mechanic – then he quit.
And now he doesn’t do anything, he stays at home.’
4 Conclusion
Multilingualism seems to have been a normal state of affairs amongst the Doms
for a very long time, probably since the genesis of the community. The reason for
this is mostly because the sociolinguistics of Domari has in likelihood remained
unchanged throughout the centuries: Domari is a community language whose
use is restricted to in-group communication. Out-group interactions imply the
use of the majority language. Due to the very nature of their occupational profile,
peripatetic groups are forced to have frequent interactions with outsiders. This
involves de facto high levels of bilingualism. Although it is hard to assess whether
the dominant language is the insider code or the outsider code, it makes sense to
suspect that balanced bilingualism was the norm, as much in the past as in the
present.
506
22 Northern Domari
Van Coetsem (1988; 2000) uses the term “transfer” generically for any kind of
contact-induced phenomenon. If the transfer is triggered by speakers who are
dominant in the source language, he uses the term “imposition”. If it originates
from recipient-language dominance, it is called “borrowing”. Lucas (2015: 525)
further introduces two categories, the first of which he calls “restructuring”, defined as a “type of change […] brought about by speakers for whom the changing
language is an L2, but it does not involve transfer”. He notes that for individuals
who acquired two languages simultaneously (in early childhood), “the distinction between borrowing and imposition breaks down”. In this case, both languages typically undergo “convergence”, that is the fourth category of contactinduced change. Because I posit balanced Arabic–Domari bilingualism as the
norm, the question that needs to be answered is whether all the contact-induced
changes happening in Domari are the product of convergence, or whether there
are changes that can be attributed to Arabic dominance (source-language agentivity or imposition). Another problem concerns the sociolinguistic limits of the
model. Speakers with two first languages are expected to initiate changes that target both languages. When languages exhibit unbalanced sociolinguistic statuses
(minority versus majority), one wonders how changes originating from minority language agentivity can diffuse to the majority. Although it cannot be ruled
out, it remains very unlikely. Consequently, convergence will always happen in
the direction of the minority language. And this is indeed what is happening between Arabic and Domari: they become more and more similar at all levels, but
only Domari is moving towards Arabic.
In the realm of phonology, it was shown that Domari has kept a distinct inventory from Arabic, although convergence with Arabic is almost complete for
short vowels. A possible consonantal imposition is found in Beirut/Damascus
Domari where etymological /q/ is realized as /ʔ/, as in neighbouring Arabic dialects. As far as morphology is concerned, eligible candidates for imposition are
the Kurdish diminutive -ək, the Turkish conditional clitic sa and superlative ān.
An evident case of imposition is the phenomenon that seems the most sensitive
to dominance: so-called “bilingual suppletion” (Matras 2012). Bilingual suppletion in Northern Domari can be observed only in the dialect of Beirut/Damascus
in the case of comparatives, and incipiently in the case of numerals. As far as
syntax is concerned, cases of imposition are probably the transfer of Arabic auxiliaries and the negator mū. The transfer of utterance modifiers such as fillers,
adverbs, conjunctions and virtually all discourse structuring devices is so prone
to replication in contact situations (Matras 1998) that it is difficult to assess the
source of agentivity. Other features discussed in this paper, such as constituent
507
Bruno Herin
order, the internal object and the impersonal construction are clear instances of
convergence.
As noted above, the main direction of change in Domari is towards convergence with Arabic, as expected in cases of absence of dominance. The dialect
of Beirut/Damascus is the most convergent of all the Northern dialects, which
in itself suggests that Arabic–Domari bilingualism is older in that variety. The
Arabic component in Domari is largely uneven cross-dialectally and no overall
statement about its nature can be made. The general picture that arises is that
the impact of Arabic gradually increases from north to south, with the dialects of
northern Syria and southern Turkey being the least Arabized, the Southern dialects spoken in Palestine and Jordan being the most influenced by Arabic, and the
dialect of Beirut/Damascus exhibiting an intermediary stage. It was also shown
that the main difference between Northern and Southern Domari as far as Arabic
is concerned is the reluctance in Northern Domari to transfer Arabic inflections
and the general tendency to favour the transfer of structures without substance.
Further reading
) For a general account of the Arabic component in all the varieties of Domari
documented so far, see Herin (2018). The paper discusses the Arabic component in Southern and Northern dialects. This is the only paper that tackles
extensively the issue of contact-induced change in Domari from a global perspective.
) For a description of the Domari dialect of Aleppo, readers can refer to Herin
(2012).
) Herin (2014) identifies the grammatical features that make Northern Domari
a coherent dialectal group.
) Herin (2016) investigates the full extent of variation in Domari as a whole,
drawing on data from both Northern and Southern Domari.
) Readers can refer to Matras (this volume) for a number of references relating
to Jerusalem Domari.
Abbreviations
1, 2, 3 1st, 2nd, 3rd person
abl ablative
acc accusative
ad adessive
ben benefactive
508
cmpr
com
cop
def
dem
comparative
comitative
copula
definite article
demonstrative
22 Northern Domari
exs
f
impf
IPA
existential
feminine
imperfect (prefix conjugation)
International Phonetic
Alphabet
in
inessive
ind indicative
indf indefinite
ipfv imperfective
m
masculine
neg negation
NP noun phrase
obj object
obl
pfv
pl
prs
prf
prog
pst
rel
sbjv
sg
supr
oblique
perfective
plural
present
perfect (suffix conjugation)
progressive
past
relative
subjunctive
singular
superessive
References
Herin, Bruno. 2012. The Domari language of Aleppo (Syria). Linguistic Discovery
10. 1–52.
Herin, Bruno. 2014. The Northern dialects of Domari. Zeitschrift der Deutschen
Morgenländischen Gesellschaft 164. 407–450.
Herin, Bruno. 2016. Elements of Domari dialectology. Mediterranean Language
Review 23. 33–73.
Herin, Bruno. 2018. The Arabic component in Domari. In Stefano Manfredi &
Mauro Tosco (eds.), Arabic in contact, 19–36. Amsterdam: John Benjamins.
Lucas, Christopher. 2015. Contact-induced language change. In Claire Bowern &
Bethwyn Evans (eds.), The Routledge handbook of historical linguistics, 519–536.
London: Routledge.
Macalister, Robert Alexander Stewart. 1914. The language of the Nawar or Zutt,
the nomad smiths of Palestine. Edinburgh: Edinburgh University Press.
Matras, Yaron. 1998. Utterance modifiers and universals of grammatical borrowing. Linguistics 36(2). 281–331.
Matras, Yaron. 2012. A grammar of Domari. Berlin: De Gruyter Mouton.
Stassen, Leon. 2013. Comparative constructions. In Matthew S. Dryer & Martin
Haspelmath (eds.), The world atlas of language structures online. Leipzig: Max
Planck Institute for Evolutionary Anthropology. https://wals.info/chapter/121.
Van Coetsem, Frans. 1988. Loan phonology and the two transfer types in language
contact. Dordrecht: Foris.
Van Coetsem, Frans. 2000. A general and unified theory of the transmission process
in language contact. Heidelberg: Winter.
509
Chapter 23
Jerusalem Domari
Yaron Matras
University of Manchester
Jerusalem Domari is the only variety of Domari for which there is comprehensive
documentation. The language shows massive influence of Arabic in different areas
of structure – quite possibly the most extensive structural impact of Arabic on any
other language documented to date. Arabic influence on Jerusalem Domari raises
theoretical questions around key concepts of contact-induced change as well as
the relations between systems of grammar and the components of multilingual
repertoires; these are dealt with briefly in the chapter, along with the notions of
fusion, compartmentalisation of paradigms, and bilingual suppletion.
1 Historical development and current state
Domari is a dispersed, non-territorial minority language of Indo-Aryan origin
that is spoken by traditionally itinerant (peripatetic) populations throughout the
Middle East. Fragmented attestations of the language place it as far north as Azerbaijan and as far south as Sudan. The self-appellation dōm is cognate with those
of the řom (Roma or Romanies) of Europe and the lom of the Caucasus and eastern
Anatolia. All three populations show linguistic resources of Indo-Aryan origin
(which in the case of the Lom are limited to vocabulary), as well as traditions of
a mobile service economy, and are therefore all believed to have descended from
itinerant service castes in India known as ḍom. Some Domari-speaking populations are reported to use additional names, including qurbāṭi (Syria and Lebanon),
mıtrıp or karači (Turkey and northern Iraq) and bahlawān (Sudan), while the
surrounding Arabic-speaking populations usually refer to them as nawar, ɣaǧar
or miṭribiyya. The language retains basic vocabulary of Indo-Aryan origin, and
shows elements of lexical phonology that place its early development within the
Central Indo-Aryan group of languages. It retains conservative derivational as
well as present-tense inflectional verb morphology that goes back to late Middle Indo-Aryan, alongside innovations in nominal and past-tense verb inflection
Yaron Matras. 2020. Jerusalem Domari. In Christopher Lucas & Stefano Manfredi
(eds.), Arabic and contact-induced change, 511–531. Berlin: Language Science Press.
DOI:10.5281/zenodo.3744545
Yaron Matras
that suggest that the language was contiguous with the Northwestern frontier
languages (Dardic) during the transition to early modern Indo-Aryan (cf. Matras
2012).
The first attestation of Palestinian Domari is a list of words and phrases collected by Ulrich Jasper Seetzen in 1806 in the West Bank and published by Kruse
(1854). It was followed by Macalister’s (1914) grammatical sketch, texts and lexicon, collected in Jerusalem in a community which at the time was still nomadic,
moving between the principal West Bank cities of Nablus, Jerusalem and Hebron.
This community settled in Jerusalem in the early 1920s, the men taking up wage
employment with the British-run municipal services. In the 1940s they abandoned their makeshift tent encampment and moved into rented accommodation
within the Old City walls, where the community still resides today. Between 1996
and 2000 I carried out fieldwork among speakers in Jerusalem and published a
series of works on the language, including two descriptive outlines (Matras 1999;
2011), annotated stories (Matras 2000), an overview of contact influences (Matras
2007), and a descriptive monograph (Matras 2012).
A number of sources going back to Pott (1907), Newbold (1856), Paspati (1870),
Patkanoff (1907), and Black (1913) provide language samples collected among the
Dom of Lebanon, Syria, Iraq and the Caucasus. These are supplemented by a
few more samples collected by ethnographers (cf. Matras 2012: 15ff.) and subsequently by data collected in Syria and Lebanon by Herin (2012). That documentation allowed me to identify a number of differences that appeared to separate a
Northern group of Domari dialects from a Southern group, which latter includes
the data recorded in Palestine as well as a sample from Jordan (see Matras 2012:
15ff.). That tentative classification has since been embraced by Herin (2014), who
goes a step further and speculates about an early split between two branches of
the language. To date, however, published attestation of Northern varieties remains extremely fragmented, notwithstanding recent work by Herin (2016; this
volume), while the only comprehensive overview of a Southern variety remains
that from Jerusalem.
Outside of Jerusalem and its outskirts there are known communities of Palestinian Doms in some of the refugee camps on the West Bank and Gaza, as well
as in Amman, where a few families sought refuge in 1967. Numbers of speakers
were very low in all these communities already in the mid 1990s and the language was only in use among the elderly. During my most recent visit to the
Jerusalem community, in January 2017, it appeared that there was only one single fluent speaker left, who, for obvious reasons, no longer had any practical use
for the language, apart from flagging the odd phrase to younger-generation semispeakers. Jerusalem Domari, and most likely Palestinian Domari in general, must
therefore now be considered to be nearly extinct.
512
23 Jerusalem Domari
2 Contact languages
Given the migration route that the Dom will have taken to reach the Middle
East from South Asia, it is plausible that the language was subjected to repeated
and extensive contact influences. Kurdish influences on Jerusalem Domari, some
of them attributable specifically to Sorani Kurdish, and some Persian items, are
apparent in vocabulary, while some of the morpho-syntactic structures (such
as extensive use of person affixes, and the use of a uniform synthetic marker
of remote tense that is external to the person marker) align themselves with
various Iranian languages. There is also a layer of Turkic loans, some of which
may be attributable to Azeri varieties, while others are traceable to Ottoman
rule in Palestine; such items are numerous in the wordlists compiled by Seetzen
and Macalister during the Ottoman period, but are much less frequent in the
materials collected a century later (for a discussion of etymological sources see
Matras 2012: 426–429).
The circumstances under which speakers of Domari first came into contact
with Arabic are unknown. There are some indications of a layered influence:
Domari tends to retain historical /q/ in Arabic-derived words, as in qahwa ‘coffee’,
qabil ‘before’, qaddēš ‘how much’, as found in the rural dialects of the West Bank
(and elsewhere), whereas contemporary Jerusalem Arabic (also used by Doms
when speaking Arabic) shows a glottal stop, as in ʔahwe, ʔabl, ʔaddēš; the word
for ‘now’ is hessaʕ, while Jerusalem Arabic has hallaʔ. It appears that the community has been fully bilingual in Arabic and Domari at least since the early
1800s, with knowledge of Turkish having been widespread among adults during the Ottoman rule. Due to the nature of the Doms’ service economy, Arabic
was an essential vehicle of all professional life, whether metalwork, hawking,
begging, or performance, but Domari remained the language of the household
until the introduction of compulsory school education under Jordanian rule in
the 1950s–60s, at which point parents ceased to pass on the language to children.
By the 1990s, use of Domari was limited to a small circle of perhaps around forty–
fifty elderly people. Due to the multi-generational structure of households it was
rare even then for conversations to be held exclusively among Domari speakers.
Domari–Arabic bilingualism has always been unidirectional, with Arabic being
the language of commerce and public interactions for all Doms, and more recently also of education and media, eventually replacing Domari as a home and
community language.
513
Yaron Matras
3 Contact-induced changes in Jerusalem Domari
As a result of ubiquitous bilingualism among all Domari speakers, Domari talk
is chequered not only with expressions that derive from Arabic, but also with
switches into Arabic for stylistic and discourse-strategic purposes such as emphasis, direct quotes, side remarks, and so on. The structural intertwining of
Domari and Arabic, and the degree to which active bilingual speakers maintain
a license to incorporate Arabic elements into Domari conversation, pose a potential challenge to the descriptive agenda. In the following I discuss those structures
that derive from Arabic, and are shared with Arabic (in the sense that they are
employed by speakers both in the context of Domari conversation and in interactions in Arabic) but constitute a stable and integral part of the structural inventory of Domari without which Domari talk cannot be formed, and for which
there is no non-Arabic Domari alternative. All examples are taken from the Jerusalem Domari corpus described in Matras (2012). Examples from Arabic are
based on colloquial Palestinian Arabic as spoken in Jerusalem.
3.1 Phonology
The entire inventory of Palestinian Arabic phonemes is available in Domari;
Arabic-derived words that are used in Domari conversation (whether or not
they have non-Arabic substitutes) do not undergo phonological or phonetic integration, except for the application of Domari grammatical word stress on caseinflected nouns (e.g. lambá ‘lamp.acc’, from Arabic lámba). The pharyngeals [ḥ]
and [ʕ] are limited to Arabic-derived vocabulary. The sounds [q], [ɣ] and [ḷ] as
well as [z] and [f] appear primarily in Arabic-derived vocabulary, but there is
evidence that they entered the language already through contact with Turkic
and Iranian languages. Less clear is the status of the pharyngealised dental consonants /ḍ, ṭ, ṣ/. These are largely confined to Arabic-derived vocabulary, but
they can also be found in inherited words of Indo-Aryan stock, where they often
represent original (Indo-Aryan) retroflex sounds (cf. ḍōm ‘Dom’, pēṭ ‘belly’). An
ongoing phonological innovation that is shared with Jerusalem Arabic is the simplification of the affricate [ʤ] to the fricative [ʒ] in inherited lexemes, e.g. džami
‘I go’ > žami. This triggers a corresponding simplification of [ʧ] to [ʃ], as in lači
‘girl’ > laši.
3.2 Morphology
Domari has not adopted productive word-derivational templates from Arabic.
Arabic inflectional morphology, however, is productive with some Arabic-de-
514
23 Jerusalem Domari
rived word forms, resulting, in effect, in a compartmentalised morphological
structure. Arabic-derived plural nouns tend to retain Arabic plural inflection,
but indigenous (inherited, Indo-Aryan) plural inflections are added to the word:
thus muslim ‘Muslim’, plural musilmīn-e Muslims-pl ‘muslims’; madrase ‘school’,
dative plural madāris-an-ka (schools-pl.obl-dat) ‘to the schools’. While Jerusalem Domari retains inherited plural marking with nouns derived from both IndoAryan and Arabic, in the closely related variety of the nomadic Doms of Jordan
the Arabic plural ending -āt is often used with inherited nouns: thus putur ‘son’,
Jerusalem Domari plural putr-e, Jordanian Domari plural putr-āt.
Arabic person agreement inflection is retained with Arabic-derived modal and
aspectual auxiliaries. The auxiliaries kān ‘be’, ṣār ‘begin’, and baqa ‘continue’
take Arabic verbal inflection, while bidd- ‘want’, ḍall- ‘continue’, and ḫallī- ‘allow’
take Arabic nominal-possessive marking:
(1)
a. kān-at
par-ar-m-a
wāšī-s
be.prf-3sg.f take-3sg-1sg-pst with-3sg
‘She used to take me with her.’
b. dōm-e kān-u
kam-k-ad-a
ḥaddādīn-e
dom-pl be-.prf-3pl work-tr-3pl-pst blacksmiths-pl
‘The Dom used to work as blacksmiths.’
(2)
a. ṣār
qaft-ar-i
min bɔy-os
begin.prf.3sg.m steal-3sg-prog from father-3sg
‘He started to steal from his father.’
b. ṣār-u
kar-and-i
ḥafl-e
begin.prf-3pl do-3pl-prog party-pl
‘They started to have parties.’
(3)
a. š-ird-i
ama-ke bidd-ha
qumn-ar
say-pfv-f 1sg-ben want-3sg.f eat-sbjv.3sg
‘She said to me that she wants to eat.’
b. bidd-i
par-am
itžawwiz-om-is
want-1sg take-1sg.sbjv marry-1sg.sbjv-3sg.obl
‘I want to take her and marry her.’
(4)
a. ḫallī-hum
naḍḍif-k-ad-i
ehe
marn-an
let.imp.2sg-3pl clean-tr-3pl-prog these.pl dead-obl.pl
‘Let them clean up these corpses.’
515
Yaron Matras
b. ḫallī-h
rʕi-k-ar
hundar
let.imp.2sg-3sg graze-tr-3sg.sbjv there
‘Let it graze there.’
Inflected Arabic-derived auxiliaries include the existential verb kān- ‘to be’,
which is used in Domari, as in Arabic, as a past- and future-tense copula, supplementing the Domari remoteness or external past-tense marker -(y)a, which
follows the lexical predication or predicate object:
(5) ihi illi par-d-om-is
kān-at
yatīm-ēy-a
this.f rel take-pst-1sg-3sg.obl be.prf-3sg.f orphan-pred.sg-pst
‘The one [woman] whom I married [her] was an orphan.’
Arabic-derived auxiliaries are also inflected for tense following Arabic paradigms:
(6) lāzem tkūnu
itme mišaṭṭaṭ-hr-es-i
must be.impf.sbjv.2pl 2pl dispersed-itr-2pl-prog
‘You must remain dispersed.’
This amounts, in effect, to a functional compartmentalisation in verbal morphology: both inherited and Arabic-derived lexical verbs take inherited IndoAryan inflection, while Arabic-derived modal and aspectual auxiliaries take Arabic inflection (for further discussion see Matras 2015).
Arabic person inflection is also found with the Arabic-derived secondary pronominal object marker iyyā-, complementiser inn-, and conjunction liʔann- ‘because’:
(7)
ple
illi t-or-im
iyyā-hum
money.pl rel give.pst-2sg-1sg.obl obj-3pl
‘the money that you gave [it] to me’
(8)
inn-hom min šamāl-os-ki
aɣlabiyy-osan š-ad-i
majority-3pl say-3pl-prog comp-3pl from north-3sg-abl
hnūd-an-ki
india-obl.pl-abl
‘Most of them say that they are from northern India.’
(9)
wars-ar-i
barra liʔann-ha
na kil-d-om
neg exit-pfv-1sg out because-3sg.f rain-3sg-prog
‘I did not go out because it was raining.’
516
23 Jerusalem Domari
(10)
payy-os
liʔinn-o
ṭāṭ-i
kān
husband-3sg because-3sg.m Arab-pred.sg be.prf.3sg.m
‘Because her husband was an Arab.’
Note that in example (9) the agreement is in the feminine singular, corresponding to the grammatical mapping of the Jerusalem Arabic construction ‘it rains’
where the (underlying) subject is the feminine noun dunya ‘the world’, while in
(7), resumptive pronoun agreement with ‘money’, a plural noun, is in the plural.
Domari is seemingly an exception to the frequently cited generalisation that
derivational morphology is more likely to be borrowed than inflectional morphology (cf. Moravcsik 1978; Field 2002; Matras 2009: §6.2.2). In fact, the constraint on
the borrowing of word-derivational morphology results from the clash with the
principle of the transparency of morphemes (cf. Matras 2009: §6.2.2): Arabic has
few if any word-derivational morphemes that can be isolated, relying instead on
complex morphological templates into which lexical roots are inserted. Nominal
plural morphemes have both inflectional function (relevant to other elements
in the clause) and derivational function (having independent meaning in standalone expressions). As shown above, they are replicated in Jerusalem Domari
as an integral part of Arabic plural word forms. On the other hand, the replication of inflectional material on auxiliaries is not productive, in that it is not
incorporated into the general lexicon, not even with lexical words of Arabic origin, but remains confined to the near-wholesale adoption of modal and aspectual
auxiliaries from Arabic. In this respect, Arabic-derived inflectional paradigms in
Domari constitute a case of both fusion as defined in Matras (2009) – the wholesale non-separation of language systems around a particular functional category
– and at the same time a case of functional compartmentalisaton as defined in
Matras (2015) – the distinct treatment of functional sub-components of a category, here the verbal category, in regard to grammatical inflection.
3.3 Syntax
Generally, Jerusalem Domari shows full congruence with Palestinian Arabic in
most syntactic functions. This includes word order rules and the formation of
both simple and complex clauses. It also includes configurations such as mapping
of tenses and modality to complement and conditional clauses, and the mapping
of semantic relations onto case markers. The latter can be adpositional or inflectional. For nominal possessive constructions, Domari has two options. The first
of those options, illustrated in (11a), is what we might call canonical Domari. It
corresponds to the inherited Indo-Aryan pattern. The second option, illustrated
517
Yaron Matras
in (11b), corresponds to the common Palestinian Arabic construction, which is
presented in (11c). Here Domari replicates the role of the Arabic dative preposition la by means of the inherited Domari ablative/possessive inflectional ending
-ki:
(11)
a. Canonical Domari
bɔy-im
kuri
father-1sg house
b. Convergent Domari
kury-os bɔy-im-ki
house-3sg father-1sg.obl-abl
c. Arabic
bēt-o
la-ʔabū-y
house-3sg.m to-father-obl.1sg
‘my father’s house’
The canonical position of adjectives in Domari is, as in other Indo-Aryan languages, before the noun (12a), while in Arabic adjectives follow the noun. However, speakers show an overwhelming preference for avoiding pre-posed adjectives and instead make use of the non-verbal predication marker in order to allow
the adjective to follow the noun (12b), thereby replicating Arabic word order patterns (12c):
(12)
a. Canonical Domari
qišṭoṭ-i šōni
er-i
come.pfv-f little-f girl
‘A little girl arrived.’
b. Convergent Domari
er-i
šōni qišṭoṭ-ik
come.pfv-f girl little-pred.sg.f
‘A little girl arrived.’ [= ‘A girl arrived, being little.’]
c. Arabic
bint zɣīre
ʔižat
come.prf.3sg.f girl little.f
‘A little girl arrived.’
The emergence of nominal clauses, facilitated by the availability of non-verbal
predication markers, might be regarded as an innovation for an Indo-Iranian language, which reinforces sentence-level convergence between Arabic and Domari:
518
23 Jerusalem Domari
(13)
a. Domari
wuda bizzot-ēk
old.m poor-pred.sg
b. Arabic
l-ḫityār
miskīn
def-old.man poor
‘The old man is poor.’
Domari, like Arabic, shows a strong tendency toward SVO word order in categorical sentences in which a thematic perspective is established by linking to a
known topical entity:
(14)
mām-om putur yāsir gar-a
swēq-ē-ta
uncle-1sg son Yassir go.pfv-m market-obl.f-dat
‘My (paternal) cousin Yassir went to the market.’
By contrast, as seen in example (12), Domari shows consistent convergence
with Arabic in regard to the position of the subject after the verb when new
topical entities are introduced, especially with verbs that convey movement and
change of state and in presentative constructions. Drawing on inherited morphology, this convergence in word order patterns also allows for the encoding
of the pronominal experiencer–recipient through a person affix that is attached
to an intransitive verb in presentative constructions, matching the Arabic construction:
(15)
a. Domari
er-os-im
ḫabar
come.pfv-3sg-1sg.obl notice
‘I received notification’
b. Arabic
ʔažā-ni
ḫabar
come.prf.3sg.m-1sg notice
Complex clauses are also congruent with Arabic. Like Arabic, Domari shows
three distinct co-temporal adverbial constructions. In the first, the subordinate
clause is introduced by the conjunction ‘and’ and the verb is finite and indicative:
(16)
a. Domari
kahind-ad-i ū pandži našy-ar-i
look-3pl-prog and 3sg
dance-3sg-prog
b. Arabic
b-yitfarražu
w hiyye b-turʔuṣ
ind-look.impf.3pl and 3sg.f ind-dance.impf.3sg.f
‘They watch her dance.’
519
Yaron Matras
In the second, the subordinated predicate appears in the present participle:
(17)
a. Domari
lah-erd-om-is
mindir-d-ēk
see-pfv-1sg-3sg.obl stand-pfv-pred.sg.m
b. Arabic
šuft-o
wāʔef
see.prf.1sg-3sg.m standing
‘I saw him standing.’
The final option shows a nominalised verb, whose possessive inflection indicates the subject/agent, introduced by the preposition ‘with’ in the subordinate
position alongside a finite main clause:
(18)
a. Domari
maʕ šuš-im-ki
tiknaw-ar-m-i
gurg-om
with sleep-1sg.obl-abl hurt-3sg-1sg-prog neck-1sg
b. Arabic
maʕ nōmt-i
b-tūžaʕ-ni
raʔbt-i
with sleep-obl.1sg ind-3sg.f-hurt.impf.3sg.f-1sg neck-obl.1sg
‘As I sleep, my neck hurts.’
Relative clauses follow the format of Arabic relative clauses: they employ the
Arabic-derived post-nominal relativiser illi and show the same distribution rules
for pronominal resumption as in Arabic:
(19) ihi illi par-d-om-is
kān-at
yatīm-ēy-a
this.f rel take-pst-1sg-3sg.obl be.prf-3sg.f orphan-pred.sg-pst
‘The one [woman] whom I married [her] was an orphan.’
Factual (indicative) complements are introduced by the Arabic-derived complementiser inn-, which carries Arabic-derived inflection (as in example 8 above),
and show comparable clause structure as in Arabic:
(20)
520
a. Domari
džan-ad-i
in-na
dōm
know-3pl-prog comp-1pl Dom
b. Arabic
b-yiʕrafu
in-na
dōm
ind-know.impf.3pl comp-1pl Dom
‘They know that we are Dom.’
23 Jerusalem Domari
Modal complements and same-subject purpose clauses show, as in Arabic, a
subjunctive complement, without a complementiser:
(21)
a. Domari
bidd-i
dža-m
ḥaram-ka ṣalli-k-am
want-1sg go-1sg.sbjv mosque-dat pray-tr-1sg.sbjv
b. Arabic
bidd-i
arūḥ
ʕa-l-ḥaram
aṣalli
want-1sg go.impf.sbjv.1sg to-def-mosque pray.impf.sbjv.1sg
‘I want to go to the mosque to pray.’
Adverbial clauses employ Arabic-derived adverbial subordinators, including
lamma ‘when’, as in (22), or composite conjunctions consisting of a preposition
and complementiser, such as baʕd mā ‘after’ and qabil mā ‘before’, as in (23) and
(24), and generally follow Arabic sentence organisation and tense and modality
distribution patterns.
(22)
lamma lak-ed-a ḫāl-os
inǧann-ahr-a bɔy-om
when see-pfv-m uncle-3sg crazy-tr.pfv-m father-1sg
‘When he saw his uncle, my father went crazy.’
(23) baʕd mā ḫaḷḷaṣ-k-ed-a kam-os gar-a
kury-is-ta
after comp finish-tr-pfv-m work-3sg go.pfv-m house-3sg.obl-dat
‘After he finished his work he went home.’
(24)
qabil mā dža-m
ḫaḷḷaṣ-k-ed-om kam-as
before comp go-1sg.sbjv finish-tr-pfv-1sg work-obl.m
‘Before I left I finished my work.’
Conditional clauses similarly draw on the Arabic conjunctions iza and law,
both ‘if’, and show similar distribution of tense and aspect categories, including
the Arabic-derived impersonal marker of counter-factuality kān, literally ‘was’:
(25)
a. Domari
law er-om
ḫužoti
kān lah-erd-om-s-a
if come.pfv-1sg yesterday was see-pfv-1sg-3sg-pst
b. Arabic
law žīt
mbāreḥ kān
šuft-o
if come.prf.1sg yesterday be.3sg.m see.prf.1sg-3sg.m
‘If I had come yesterday, I would have seen him.’
521
Yaron Matras
3.4 Lexicon
Jerusalem Domari shows extensive impact of Arabic on the grammatical lexicon,
including almost wholesale reliance on Arabic-derived material for entire categories. In the pronominal domain, Domari employs, in additional to the secondary pronominal object marker iyyā- discussed above, also the Arabic reflexive
pronoun ḥāl-, derived from the word ‘state’, combined with person/possessive
inflection, and the Arabic reciprocal pronoun baʕḍ-:
(26) naḍḍif-k-ad-a ḥāl-os
clean-tr-pfv-m refl-3sg
‘He cleaned himself.’
(27) tʕarraf-h-r-ēn baʕḍ-ē-man-ta
meet-tr.pfv-1pl recp-pl-1pl-dat
‘We met one another.’
Indefinite expressions draw on Arabic-derived forms of category determination including negative wala, free choice ayy and universal kull, which may
be combined with inherited ontological markers, as well as on the ontological
specifiers ḥāǧ- for thing and maḥall for location. Indefinite expressions that derive entirely from Arabic include temporal wala marra ‘never’, dāyman ‘always’,
and universal-thing kullši ‘everything’. Arabic-derived focus particles are barḍo
‘also, too’ and ḥatta ‘even’ and quantifiers are kull ‘every, each’ and akamm ‘a
few’. Interrogatives are generally inherited (Indo-Aryan), with the exception of
qaddēš ‘how much’. Numerals are all derived from Arabic with the exception of
the lowest numeral forms (‘one’ to ‘five’ in citation function and ‘one’ to ‘three’
in attributive role) (see Tables 1–2); all ordinal numerals (awwal ‘first’, tāni ‘second’ etc.) are from Arabic.
Alongside a very small number of inherited prepositions that are used exclusively with pronominal (person-inflected) forms, most prepositions are derived
from Arabic (Table 3).
Arabic-derived grammatical operators at verbal clause level include a series of
modality adverbs such as masalan ‘for example’, yimken ‘perhaps’, atāri ‘well’,
time adverbs such as hessaʕ ‘now’ and baʕdēn ‘then, afterwards’, and the phasal
adverbs lissa and lāyzāl, both ‘still’. As discussed above, Domari adopts Arabic
modal and aspectual auxiliaries wholesale, i.e. along with their Arabic-derived
inflection. This covers almost the full category of modal and aspectual auxiliaries
including habitual/iterative kān ‘be’, ṣār ‘begin’, and baqa ‘continue’, bidd- ‘want’,
ḍall- ‘continue’, and ḫallī- ‘let’, as well as the impersonal form lāzem ‘must’. The
522
23 Jerusalem Domari
Table 1: Jerusalem Domari numerals
Numeral
Citation
Attribute
1
2
3
4
5
6
7
8
9
10
20
21
22
23
24
100
1000
ikak
diyyes
taranes
štares
pʌndžes
sitt-ēk-i
sabʕ-ak-i
tamāni-ak-i
tisʕ-ak-i
das ‘ten’, ʕašr-ak-i
ʕišrīn-i, wīs-i
ʕišrīn ū ekak-i
ʕišrīn-i ū diyyes-i
ʕišrīn-i ū taranes-i
ʔarbaʕ ū ʕišrīn
miyyēk hi, siyy-ak-i
alf-ak-i
-ak
di
taran
ʔarbaʕ
ḫamis
sitt
sabaʕ
tamānye
tisʕa
ʕašr
ʕišrīn
wāḥed w ʕišrīn
tnēn w ʕišrīn
talāte w ʕišrīn
ʔarbaʕ w ʕišrīn
miyye
alf
Table 2: Jerusalem Domari higher numerals
Numeral
Form
30
40
50
60
70
80
90
talātīn
ʔarbaʕīn
ḫamsīn
sittīn
sabʕīn
tamanīn
tisʕīn
523
Yaron Matras
Table 3: Arabic-derived prepositions in Jerusalem Domari
ʕan
maʕ
min
la, ʕala
fi
zayy
ʕind (ʕand)
badāl
‘on, about’
‘with’
‘from’
‘to’
‘in’
‘like’
‘at’
‘instead of’
ʕašān
minšān
min ɣēr
min/bi dūn
bēn
ḥawāli
min ḍamn
ʔilla ɣēr
‘because’
‘for’
‘without’
‘without’
‘between’
‘around’
‘among’
‘except for’
nawāḥi
qabil
baʕd
laɣāyet
bi
ḍiḍḍ
žamb
‘toward’
‘before’
‘after’
‘until’
‘in, for’
‘against’
‘next to’
only modal for which an Indo-Aryan form is retained is sak- ‘to be able to’.
Past-tense finite predications take the Arabic negator mā (alongside inherited
na) while in non-finite predications the Arabic negation particle miš is used:
(28)
mā lak-ed-om-is
neg see-pfv-1sg-3sg.obl
‘I didn’t see him/her.’
(29) bay-os miš kury-a-m-ēk
wife-3sg neg house-obl.f-loc-pred.sg
‘His wife is not at home.’
Clause combining relies exclusively on Arabic-derived material (connectors
and conjunctions) (see Table 4).
Likewise, the inventory of discourse particles and interjections is adopted in
its entirety from Arabic: We find the interjection, tags and filers yabayyi, yaḷḷa,
xaḷaṣ, waḷḷa, and yaʕni, as well as segmental markers with a lexical meaning such
as l-muhimm ‘anyway’, l-ḥāṣil ‘finally’, ṭayyib ‘well’, w ʔiši ‘and the like’, w hāda
‘and so on’, abṣar ‘whatever’, and the filler hāy ‘that’. The quotation particle
qal/ḫal, from Arabic ‘say’, is not found in Jerusalem Arabic and appears to represent an older layer of Arabic influence (as indicated also by its phonological
structure; see §2).
The content lexicon equally shows massive impact of Arabic. In the Jerusalem
Domari corpus of narrational and conversational talk as well as sentence elicitation recorded in the 1990s (Matras 2012), almost two thirds of lexical items are
Arabic-derived; the count includes single-word insertions from Arabic, including
attributive nominal compounds (noun–possessor and noun–adjective), but excludes phrases containing a finite lexical verb that is Arabic-derived (which latter
524
23 Jerusalem Domari
Table 4: Arabic-derived conjunctions in Jerusalem Domari
w
wala
yā
willa
bass
illi
innliʔann
lamma
kull mā
‘and’
‘and not’, ‘(n)either’
‘or’
‘or (else)’
‘but’, ‘only’
relative pronoun
‘that’
‘because’
‘when’
‘whenever’
qabil mā
baʕd mā
min-yōm-mā
iza
law
bi-r-raɣim
ʕašān
minšān
ta
‘before’
‘after’
‘since’
‘if’
‘if’
‘despite’
‘for’, ‘in order to’
‘for’, ‘in order to’
‘in order to’
are regarded as optional code-switches). Both Arabic-derived nouns and adverbs
outnumber inherited (Indo-Aryan) counterparts by around 65% to 35%, while for
verbs and adjectives the numbers are roughly equal. Around 26% of items of both
the Swadesh 100-item list and the Leipzig–Jakarta 100-item list (Haspelmath &
Tadmor 2009) are Arabic-derived. This puts Domari in the range of languages
considered to be “high borrowers” by the Leipzig Loanword Typology Project
(Haspelmath & Tadmor 2009). Meanings on the list that are replaced by Arabic loans in Domari include a number of animals (‘ant’, ‘bird’, ‘fish’), activities
(‘to run’, ‘to fly’, ‘to crush’), elements of nature (‘star’, ‘soil’, ‘shade’, ‘ash’, ‘leaf’,
‘root’), and some body parts (‘knee’, ‘navel’, ‘liver’, ‘thigh’; also ‘wing’, ‘tail’). On
the whole, the meaning and usage of Arabic-derived lexemes matches that of
Jerusalem Arabic. Creative processes are marginal and include such processes
as the phonological volatility of /q/ (as [q], [x], [qx] and [ɡ]), the alternation between farǧik- ‘to show’ (Arabic √frǧ) and warǧik-, and the occasional creative
derivation such as bisawahr- ‘to get married’, from Arabic bi-sawa ‘together’.
Arabic verbs are integrated into Domari through a light verb construction that
draws on the inherited verb stems -k- ‘to do’ and -h- ‘to become’, which are grammaticalised into loan-verb adaptation markers (see Matras 2012: 240–244) that
are sensitive to valency. This follows a strategy for the adaptation of loan verbs
that is widespread across a geographical area stretching from the Balkans and
the Caucasus through Anatolia and Western Asia and on to the Indian Subcontinent. For some verbs, alternating adaptation markers can indicate change in valency: ǧawwiz-h-r-i (marry-itr-pfv-f) ‘she got married’, ǧawwiz-k-am-is (marrytr-1sg.sbjv-3sg.obl) ‘I shall marry her off’. The core of integrated Arabic verbs
generally derives from the Arabic subjunctive–imperative form, which in Arabic
525
Yaron Matras
never occurs in isolation from its person inflection in the prefix conjugation, as
in ǧawwiz- ‘marry’, from *ǧawwiz ‘marry (off)!’ or *tǧawwiz ‘get married!’. Note,
however, that the vowel structure of the core does not always correspond to the
subjunctive–imperative form of contemporary Palestinian Arabic, which is quite
possibly a further indication of the layered historical influence of Arabic. Thus
we find s’il-k-ed-om (ask-tr-pfv-1sg) ‘I asked’, from *s’il- ‘ask’, while Palestinian
Arabic has isʔal ‘ask!’, and rawwaḥ-ah-r-a (go-itr-pfv-m) ‘he travelled’, while
Palestinian Arabic has rawwiḥ ‘go away!’.
3.5 Cross-category interplay
A typologically curious case of contact-induced change is offered by the use in
Jerusalem Domari of three construction types that cut across structural categories. The first pertains to the comparative form of adjectives. In the absence of a
structurally transparent, isolated and replicable marker of adjective comparison
(comparative and superlative), Domari draws on Arabic word forms for all comparative adjective forms, even when an inherited (non-Arabic) word form is used
for the positive form of the adjective, as illustrated in (30) (cf. Herin, this volume:
§3.2).
(30)
a. Domari
atu
qaštot-ik
you.sg small-pred.sg.f
‘You are small.’
b. Domari
atu
azɣar mēšī-m-i
you.sg smaller from-1sg-pred.sg
‘You are smaller than I.’
c. Arabic
inti zɣīre
2sg.f small.f
‘You are small.’
d. Arabic
inti azɣar minn-i
2sg.f smaller from-obl.1sg
‘You are smaller than I.’
This formation involves essentially the recruitment of an alternative, Arabicderived item from the category of lexical items in order to carry out a grammatical procedure that is derivational–inflectional by nature (derivational in that it
modifies meaning, inflectional in that it is inherently embedded into a syntactic
relationship at the phrase level); thus we have a case of cross-category interplay.
A further case is that of lexical suppletion around Arabic-derived numerals.
Domari and Arabic differ typologically in respect of numeral agreement: with
Indo-Aryan numerals, the Domari noun appears in the default singular form,
while in Arabic, numerals up to ‘ten’ take plural agreement. The clash is resolved
in Domari in such a way that Arabic-derived numerals under ‘ten’ invariably
trigger an Arabic-derived lexical item even when an inherited form of the corresponding lexeme is available:
526
23 Jerusalem Domari
(31)
a. ḥkum-ke-d-os
taran wars maḥkame
sentence-tr-pfv-3sg three year court
‘The court sentenced him to three years.’
b. eh-r-a
ʕumr-om sitte snīn
become-pfv-m age-1sg six year.pl
‘I turned six years old.’
Such alternation is systematic (see further examples in Table 5) and might
be regarded as a case of bilingual suppletion, where every countable noun in
the language for which an inherited (Indo-Aryan) word form exists also has an
Arabic-derived counterpart that is used with numerals between ‘three’ and ‘ten’.
Table 5: Some phrases from the corpus containing numerals and nouns
Inherited numeral and singular noun
Arabic numeral and plural noun
di dīs taran dīs ‘two days three days’
taran mas ‘three months’
taran wars ‘three years’
taran zard ‘three pounds’
sabaʕ-t-iyyām ‘seven days’
ḫamas-t-ušhur ‘five months’
sitte snīn ‘six years’
ḫamas līrāt ‘five pounds’
Finally, while Domari lacks a definite article, the Arabic definite article l- is
employed with definite noun phrases where both the noun and the numeralattribute are derived from Arabic:
(32) mar-d-e
l-ʔarbaʕ ḫurfān
kill-pfv-3pl def-four lamb.pl
‘They slaughtered the four lambs.’
(33)
dīr-os
it-tānye
eh-r-i
muhandis-ēk
daughter-3sg def-second.f become-pfv-f engineer-pred.sg.f
‘Her other daughter became an engineer.’
4 Conclusion
The comparison with Macalister’s (1914) materials offers some scope for observations in respect of the historical development of contact-induced change over
the past century in at least two areas of structure, namely the loss of Turkishderived vocabulary as well as of some of the inherited Indo-Aryan vocabulary
(around 55 words are attested in Macalister’s materials that were not familiar
527
Yaron Matras
to the speakers I interviewed), and the adoption of fully-inflected modal and aspectual auxiliaries, compared to their use as impersonal forms in Macalister’s
material. One has to bear in mind, however, that Macalister’s corpus is based on
work with just a single speaker. Nevertheless, these changes provide some indication that the impact of Arabic continued to expand during the last century in
which the language was spoken, a period during which the Doms lost much of
their distinct culture and lifestyle as a result of the shift from a semi-nomadic service economy to a settled, wage-based but still socially isolated and stigmatised
community.
The impact of Arabic on Domari prompts a theoretical challenge around identifying a form of the language that is structurally inseparable from Arabic. This
can be illustrated by the following two examples:
(34)
a. Domari
aktar min talātīn ḫamsa w talātīn sana mā lak-ed-om-is
more from thirty five and thirty year neg see-pfv-1sg-3sg.obl
b. Arabic
aktar min talātīn ḫamsa w talatīn sana mā šuft-ha
more from thirty five and thirty year neg see.prf.1sg-3sg.f
‘I haven’t seen her for more than thirty, thirty five years.’
(35)
a. Domari
ʕumr-om yimken sitte snīn sabʕa snīn
kān
be.prf.3sg.m age-1sg maybe six years seven years
b. Arabic
kān
ʕumr-i
yimken sitte snīn sabʕa snīn
be.prf.3sg.m age-obl.1sg maybe six years seven years
‘I was maybe six or seven years old.’
Both (34a) and (35a) are unambiguously identifiable to speakers as Domari
utterances; moreover, their meaning cannot be conveyed in Domari in any other
way. Yet they each differ in just one single element from their respective counterpart Arabic utterances in (34b) and (35b): the use of the lexical verb with subject
and object agreement (Domari lak-ed-om-is ‘I saw her’, Arabic šuf-t-ha) in the
first, and the use of the 1sg possessive marker (Domari -om, Arabic -i) with the
word ʕumr ‘age’ in the second. Despite being isolated examples, (34)–(35) illustrate the considerable extent of structural overlap between the two languages.
Furthermore, the examples discussed above of bilingual suppletion in number
agreement and adjective comparison, and the productive use of Arabic person
528
23 Jerusalem Domari
agreement inflection with auxiliaries and with some complementisers and secondary object markers, mean in effect that active command of Arabic is a prerequisite for speaking Domari.
It follows that Domari provides us with an opportunity to reconsider the taxonomy of contact-induced language change phenomena. It is not a Mixed Language by conventional definitions (cf. Bakker & Matras 2013; Matras 2009: chapter 10) since the Indo-Aryan source of grammatical inflection in all word classes
is overwhelmingly consistent with the source of basic lexical vocabulary and of
deictic and anaphoric elements (demonstrative and personal pronouns, interrogatives, and spatial adverbs). Impressionistically speaking, it is a language with
“heavy borrowing” in that it shows the adoption of Arabic-derived material in a
wide range of different structural categories. But the distribution of some of this
material, taking into account the ubiquitous active bilingualism among Domari
speakers, lends itself to the postulation of several particular types of contactinduced structural change, which I have labeled above fusion (wholesale nonseparation of languages around a particular structural category, e.g. clause connectors and modal auxiliaries), inflectional compartmentalisaton (the use of Arabic inflectional paradigms with particular functional categories, notably modal
and aspectual auxiliaries), and bilingual suppletion (activation of speakers’ full
command of Arabic vocabulary and inflection for creative formations around
number agreement and adjective comparison).
Further Reading
) Matras (2007) outlines contact influences on Jerusalem Domari in the context
of a collection of chapters on contact-induced change in a sample of different
languages.
) Matras (2012) provides a descriptive and historical overview of Jerusalem Domari and includes extensive discussion of contact-induced change in the individual chapters as well as a chapter devoted to the impact of Arabic.
) Matras (2009) is a general theoretical discussion of contact-induced change in
functional-typological perspective and includes many examples from Jerusalem Domari.
) Finally, Matras (2015) discusses patterns of morphological borrowing and their
theoretical implications and gives as one of the examples the compartmentalisaton of modal and aspectual auxiliaries in Jerusalem Domari.
529
Yaron Matras
Abbreviations
1, 2, 3
abl
ben
comp
dat
f
imp
impf
ind
itr
loc
m
obl
1st, 2nd, 3rd person
ablative
benefactive
complementiser
dative
feminine
imperative
imperfect (prefix conjugation)
indicative
intransitive
locative
masculine
oblique
pfv
pred
prs
prf
prog
pst
recp
refl
rel
sbjv
sg
tr
perfective
predication (non-verbal)
present
perfect (suffix conjugation)
progressive
past
reciprocal
reflexive
relativiser
subjunctive
singular
transitive
References
Bakker, Peter & Yaron Matras (eds.). 2013. The mixed language debate: Theoretical
and empirical advances. Berlin: Mouton de Gruyter.
Black, George Fraser. 1913. The gypsies of Armenia. Journal of the Gypsy Lore
Society (new series) 6. 327–330.
Field, Frederick. 2002. Linguistic borrowing in bilingual contexts. Amsterdam:
John Benjamins.
Haspelmath, Martin & Uri Tadmor (eds.). 2009. Loanwords in the world’s
languages: A comparative handbook. Berlin: De Gruyter Mouton.
Herin, Bruno. 2012. The Domari language of Aleppo (Syria). Linguistic Discovery
10. 1–52.
Herin, Bruno. 2014. The Northern dialects of Domari. Zeitschrift der Deutschen
Morgenländischen Gesellschaft 164. 407–450.
Herin, Bruno. 2016. Elements of Domari dialectology. Mediterranean Language
Review 23. 33–73.
Kruse, Friedrich. 1854. Ulrich Jasper Seetzen’s Reisen durch Syrien, Palästina,
Phönicien, die Transjordan-Länder, Arabia Petraea und Unter-Aegypten. Vol. 2.
Berlin: Reimer.
Macalister, Robert Alexander Stewart. 1914. The language of the Nawar or Zutt,
the nomad smiths of Palestine. Edinburgh: Edinburgh University Press.
Matras, Yaron. 1999. The state of present-day Domari in Jerusalem. Mediterranean
Language Review 11. 1–58.
530
23 Jerusalem Domari
Matras, Yaron. 2000. Two Domari legends about the origin of the Doms. Romani
Studies 10. 53–79.
Matras, Yaron. 2007. Grammatical borrowing in Domari. In Yaron Matras &
Jeanette Sakel (eds.), Grammatical borrowing in cross-linguistic perspective, 151–
164. Berlin: Mouton de Gruyter.
Matras, Yaron. 2009. Language contact. Cambridge: Cambridge University Press.
Matras, Yaron. 2011. Domari. In Tatiana I. Oranskaia, Yulia V. Mazurova, Andrej
A. Kibrik, Leonid I. Kulikov & Aleksandr Y. Rusakov (eds.), Languages of the
world: New Indo-Aryan languages, 775–811. Moscow: Academia.
Matras, Yaron. 2012. A grammar of Domari. Berlin: De Gruyter Mouton.
Matras, Yaron. 2015. Why morphological borrowing is dispreferred. In Francesco
Gardani, Peter Arkadiev & Nino Amridze (eds.), Borrowed morphology, 47–80.
Berlin: Mouton de Gruyter.
Moravcsik, Edith A. 1978. Universals of language contact. In Joseph H. Greenberg, Charles A. Ferguson & Edith A. Moravcsik (eds.), Universals of human
language, vol. 1: Method and theory, 94–122. Stanford: Stanford University
Press.
Newbold, F. R. S. 1856. The Gypsies of Egypt. Journal of the Royal Asiatic Society
of Great Britain and Ireland 16. 285–312.
Paspati, Alexandre. 1870. Etudes sur les Tschinghianés ou Bohémiens de l’Empire
Ottoman. Osnabrück: Biblio.
Patkanoff, K. P. 1907. Some words on the dialects of the Transcaucasian Gypsies.
Journal of the Gypsy Lore Society (new series) 1. 229–257.
Pott, August F. 1907. Über die Sprache der Zigeuner in Syrien. Zeitschrift für die
Wissenschaft der Sprache 1. 175–186.
531
Chapter 24
Mediterranean Lingua Franca
Joanna Nolan
SOAS University of London
This chapter explores the effect of Arabic contact on Lingua Franca, an almost exclusively oral pidgin spoken across the Mediterranean and along the North African
coastline from the seventeenth to the nineteenth centuries. The chapter highlights
the phonological and lexical impact Arabic appears to have had on Lingua Franca.
1 Overview and historical development
Today, lingua franca is a term describing a language used by two or more linguistic groups as a means of communication, often for economic motives. Typically, none of the groups speak the chosen language as their native tongue. The
original and eponymous Lingua Franca, however, was a trading language, used
among and between Europeans and Arabs across the Mediterranean (Kahane &
Kahane 1976). Its exact historical and geographical roots – as well as its precise
lexifier languages – prove elusive. Hall (1966) dates Lingua Franca’s birth to the
era of the crusades, while other linguists (Minervini 1996; Cifoletti 2004) suggest
that it took root on the North African Barbary Coast (in the Regencies of Algiers,
Tunis and Tripoli) at the close of the sixteenth century.
Contention extends to its very name. There are several discrete etymological suggestions for the term Franca. Some linguists interpret Franca as meaning ‘French’ (e.g Hall 1966: 3). Hall claims that France’s regional significance in
the medieval era meant that its languages, specifically Provençal, were adopted
across the Mediterranean, and were a key constituent of the original Lingua
Franca. In their etymological study of Lingua Franca, Kahane & Kahane (1976:
25), on the other hand, assert that the name Lingua Franca is rooted in the East
and the Byzantine tradition, stemming from the Greek word phrangika, which denoted Venetian as much as Italian, or indeed, as any Western language (Kahane
Joanna Nolan. 2020. Mediterranean Lingua Franca. In Christopher Lucas & Stefano
Manfredi (eds.), Arabic and contact-induced change, 533–548. Berlin: Language Science Press. DOI:10.5281/zenodo.3744547
Joanna Nolan
& Kahane 1976: 31). An alternative etymology for Lingua Franca, espoused by
Schuchardt (1909: 74) among others, is from the Arabic, lisān al-faranǧ ‘language
of the Franks’. This initially referred to Latin and then to describe a trading
language employed largely by Jews across the Mediterranean. It later came to
encompass the languages of all Europeans, but particularly Italians (Kahane &
Kahane 1976: 26).
Evolving from its maritime origins, by the late sixteenth century Lingua Franca
was the language of pirates of the North African Barbary coast and their captured slaves, and, as such, the subject of legend and myth. Indeed, the variation
found in the accounts of Lingua Franca, and descriptions of its linguistic makeup
lead some linguists (Minervini 1996; Mori 2016) to suggest that there may have
been multiple Lingua Francas or that it was simply second-language Italian. As
Schuchardt (1909: 88), identified, Lingua Franca – perhaps above all in its resistance to theoretical classification – adheres to Heraclitus’ philosophy of panta
rei ‘everything is in flux’.
Contemporaneous descriptions of Lingua Franca detailing its lexifiers and, in
some cases, its salient features, come mostly from the North African Barbary
Regencies and from the Levant. While the writers of these descriptions often
identify Italian and Spanish as lexifiers, there are also, if fewer, mentions of Portuguese, French, Provençal, Arabic, Turkish and Greek (see below for further
detail). This speaks to the hypothesis that there were multiple Lingua Francas,
or perhaps more appropriately lingua francas. It also raises the frequent subjectivity of the source’s writer and their consequent interpretation of Lingua Franca.
Their native language appears to have a bearing on the makeup of the Lingua
Franca recorded. It may influence the lexicon they hear, as well as the orthography they employ in their account. Equally, there is the subjectivity of the researcher to bear in mind. The assumption that a French source, for example, has
represented Lingua Franca in a particular manner overlooks the fact that the European residents, particularly of port cities across the Mediterranean, most likely
would have been multilingual, with an ability to adapt their lexicon to maximize
understanding and communication with their interlocutor.
The most widespread documentation of Lingua Franca comes from the Levant and northwest Africa. Algiers, and to a lesser extent, Tunis and Tripoli, had
long been the crucible of Mediterranean piracy, and as the slave trade of Barbary
pirates increased – with over a million European slaves held there between the
sixteenth and nineteenth centuries (Davis 2004: 23) – so too did the domains
and usage of Lingua Franca. The sixteenth–seventeenth century Spanish Abbot
Diego del Haedo described it as follows:
534
24 Mediterranean Lingua Franca
La que los Moros e Turcos llaman Franca … siendo todo una mexcla de lenguas
cristianas y de vocablos, que son por la mayor parte Italianos e espanoles y algunos portugueses … Este hablar Franco es tan general que non hay casa do
no se use
‘that which the Arabs and Turks call Franca … being a mix of Christian languages and words, which are in the majority Italian and Spanish and some
Portuguese, this Franca speech is so widespread that there isn’t a house [in
Algiers] where it isn’t spoken’ (Haedo 1612: 24; author’s translation).
Despite its alleged profusion in the Barbary coast, and numerous references
in various contemporary sources, the corpus of Lingua Franca is remarkably limited. The exclusively European documentary sources (from diplomats, travellers,
priests and slaves) provide mostly phrases and individual words and a handful of
short dialogues. The most fulsome examples come from literature, and, as such,
only provide indirect, and potentially less authentic, evidence of the contact vernacular.
For example, the alleged earliest record of Lingua Franca comes from an anonymous poem, Contrasto della Zerbitana, found by Grion (1890) in a fourteenthcentury Florentine codex, and apparently written in the late thirteenth or early
fourteenth century on the island of Djerba, off the coast of Tunisia. The sixteenth
century La Zingana (Giancarli 1545) has an eponymous Arabic-speaking heroine
whose language features hallmarks of Lingua Franca, while the speeches of the
Turkish characters in Molière’s Le Sicilien (1667) and Le Bourgeois gentilhomme
(1798 [1670]) also appear to share a number of its defining linguistic traits.
The first detailed description and documentation of Lingua Franca comes in
Haedo’s (1612) Topographia, a comprehensive study of Algiers, with a chapter devoted to the languages spoken there. Haedo spent several years in Algiers at the
close of the sixteenth century and was even imprisoned for a number of months.
His Topographia details the urban features, social makeup and linguistic mix of
Algiers, creating an impression of Lingua Franca’s ubiquity across multiple domains, and indispensability to daily commercial, and even domestic, life.
Other early sources are predominantly French. A Trinitarian priest, Pierre Dan,
was almost contemporary with Haedo in Algiers; in the mid-seventeenth century
the diplomat Savary de Brèves travelled to Tripoli; and Chevalier D’Arvieux,
King Louis XIV’s envoy to the region, and advisor to Molière on Turkish and
Arabic matters, visited both Algiers and Tunis. All these men offered excerpts
of Lingua Franca in their writings, as well as descriptions of its character and
lexifiers (Savary de Brèves 1628; Dan 1637; D’Arvieux 1735).
535
Joanna Nolan
Certainly seventeenth-century Algiers and the other two Barbary Regencies
of Tunis and Tripoli provided the conditions for what had previously been a prepidgin – with limited lexicon and a lack of stability – to evolve into the language
of daily life, permeating all echelons of society and facilitating contact among and
between the plurilingual populations of the urban centres. The Barbary states
were, from the late sixteenth century, under the de jure but not de facto control of
the Ottoman empire, whose support was needed to shore up the rule of the Greek
Barbarossa brothers who had ousted Spanish forces from North Africa. The two
brothers (named for their red beards), Aruj and Hizir, gradually brought much
of North Africa under Turkish sovereignty through a series of naval challenges
and, later, city sieges securing power over coastal areas. The indigenous population rallied to the brothers’ cry and although the elder, Aruj, was slain, Hizir
assumed control of Algiers in the early sixteenth century (Tinniswood 2010: 8;
Weiss 2011: 10). He immediately offered the Ottoman Empire control over the
brothers’ conquests in order to bolster his own position and ward off threats
from Spain. Ottoman rule was compounded over the following decades (Clissold
1977: 27).
In reality, however, the Regencies had unstable political systems with local
elites vying for power. Their economy was driven by corsairing, and the real
power lay in the hands of the mostly European renegades who carried out raids
on land and at sea, seizing cargo and most importantly human booty, sold as
slaves on their return (Plantet 1889).
The huge influx of captured Europeans swelled the urban population and created multinational, multidenominational and notably multilingual societies. The
Flemish diplomat D’Aranda, imprisoned in Algiers in the 1660s, wrote of hearing
twenty-two languages in the slave quarters of the city (D’Aranda 1662). Lingua
Franca emerged as a contact language accessible to the majority of slaves (though
not all), given its Romance-influenced lexicon. Although the elites were predominantly Arabic-speaking, Europeans permeated the upper levels of Barbary society through their economic sway as corsairs and high levels of inter-marriage of
Arabs and Europeans. Lingua Franca quickly became the default language within
the slave quarters, known locally as bagnios, seemingly a Lingua Franca term, and
in master–slave relationships. Authors who detail the use of Lingua Franca across
more than 250 years and throughout the regencies, including Pananti (1841) and
Broughton (1839), report the regular use of Lingua Franca by Arabic-speaking
slave owners, including the Pashas, Beys and various dignitaries of the ruling
households.
As noted above, Lingua Franca also elicits various opinions regarding its key
lexifiers, though one common point of agreement among its contemporary wit-
536
24 Mediterranean Lingua Franca
nesses and speakers is that Italian1 and Spanish are mentioned repeatedly as
principal sources. Descriptions mostly include at least three lexifying languages,
though not always the same three: while Italian and Spanish are consistently
named, Provençal also features (D’Arvieux 1735: vol. 5, p.235) as does Portuguese
(Haedo 1612: 24; D’Aranda 1662: 22). A much later account from the Italian merchant Pananti, briefly imprisoned in Algiers, mentions Arabic as a lexifier: la
Lingua franca è una mistura d’italiano, di arabo e di spagnolo ‘Lingua Franca is a
mix of Italian, Arabic and Spanish’ (Pananti 1841: 201). Within the same memoir,
however, Pananti refers to African rather than Arabic as one of three lexifiers of
Lingua Franca (the other two being Italian and Spanish). Such inconsistency is
the hallmark of many of the sources, compounding an already confused and often contradictory picture of the language. As Selbach (2008: 18) observes, “lexical
variants were as much a part of the language as variant lexifiers.”
The contribution of multiple languages to Lingua Franca is borne out by its
lexicon: there are often several alternatives for a single meaning, as listed in the
Dictionnaire (Anonymous 1830), the sole comprehensive lexical record of the language. For example, ‘to do’ is translated by far (from the Italian fare), fazir (from
Portuguese fazer), and counchiar (likely from Sicilian cunzari; Cifoletti 2004: 316).
Lingua Franca – as substantiated by its corpus of written attestations – is always labelled as such by a non-speaker. The European authors of descriptions of
Lingua Franca, its diffusion and its usage, present it as, if not foreign, certainly
removed and remote from their own languages, and attribute the speaking of
Lingua Franca to the Arabs and Turks (even if it is clearly a language that lexically is much closer to European, and specifically Romance, languages).2 In her
comprehensive if subjective account and an analysis of Lingua Franca, Dakhlia
highlights how citations, often expressing insults and aggressive warnings, and
usually introduced into the texts by witnesses to Lingua Franca in direct speech,
punctuated with exclamation marks, underline what she terms le choc linguistique de l’altérité et de la barbarie ‘the linguistic shock (or jolt) of otherness and
barbarism’ (Dakhlia 2008: 351). This choice of words is important. Dakhlia’s association of otherness and barbarism suggests that Lingua Franca, according to
the European documentary sources, was the language of the Arab oppressor.
While this may apply to the corsairs and slave-masters, speaking Lingua Franca
1
Italian is a catch-all term, used by contemporary authors in Barbary, as well as linguists today, as identified by Trivellato (2009: 178): “I write “Italian”, “Portuguese” and “Spanish”, but
recall that European written languages in the epoch were not fully standardized.” Venetian and
Tuscan were both described as Italian, for example.
2
See, for example, the exchange between Louis Bonaparte and Hyde Clarke across several issues
of the periodical Athenaeum in 1877.
537
Joanna Nolan
to their European captives, there are other instances within the corpus of written attestations where Arabic-speaking elites use Lingua Franca in diplomatic,
even philosophical, exchanges. For example, Louis Frank, the Bey of Tunis’ doctor, comments on the deemed impropriety of the Bey speaking formal Italian,
and his consequent use of Lingua Franca, which permeated all levels of society
(Frank 1850: 70):
la langue franque, c’est à dire cet italien ou provençal corrompu qu’on parle
dans le Levant, lui est également familière; il avait même voulu essayer d’apprendre à lire et à écrire l’italien pur-toscan: mais les chefs de la religion l’ont
détourné de cette etude, qu’ils prétendaient être indigne d’un prince musulman.
‘Lingua Franca, or rather this bad Italian or Provençal spoken in the Levant,
is equally familiar to him; he had actually wanted to learn to read and write
pure Tuscan Italian; but his religious chiefs had warned him off such study,
which they claimed was unworthy of a Muslim prince’ (author’s translation).
Frank wrote of the intriguing linguistic, socio-political, cultural and even religious conflation evidenced in Lingua Franca, describing an encounter with a
Muslim beggar, who implored: “Donar mi meschino la carità d’una carrouba3 per
l’amor della Santissima Trinità e dello gran Bonaparte”, ‘Please to give miserable
me the charity of a penny for the love of the most holy Trinity and the great
Bonaparte’ (Frank 1850: 101; author’s translation). In just this one Lingua Franca
sentence, multiple lexifiers are represented: meschino is from the Arabic, miskīn,
and there is Spanish in carrouba ‘penny’, donar ‘give’, and amor ‘love’, with the
Italian Catholic reference of Santissima Trinità ‘most holy Trinity’, and French
Bonaparte. The latter would have still been Emperor and possibly at the height
of his power. Bonaparte is qualified by gran ‘great’, from the Italian or even Venetian. It suggests how cosmopolitan, multicultural and multilingual Tunis and
its population had become that a beggar should speak this way. Even Frank was
struck by the incongruity of the beggar’s words: “sa supplique en ces termes, bien
étranges dans la bouche d’un Musulman”, ‘his petition in these terms, very odd in
the mouth of a Muslim’ (Frank 1850: 101; author’s translation).
Lingua Franca’s demise dates from 1830, as a consequence of the outlawing of
slavery and the start of the French colonization of North Africa. Lingua Franca
3
Until 1891 a carrouba was worth 1/16 of a Tunisian piaster, according to Rossetti (1999).
538
24 Mediterranean Lingua Franca
became known alternatively as Sabir (HSA 1882: Letter 6-7473), with a later incarnation which Schuchardt dubbed Judeo-Sabir (Schuchardt 1909: 87). Residual elements seem to persist, however, in other contemporary jargons and languages.
The pidgin spoken in Algeria, Pataouète, while largely lexified by French and
Arabic, also features a significant number of words that are identified by Lanly
(1962) as Lingua Franca in origin. Duclos’ (1992) Pataouète dictionary enumerates
at least thirty words whose etymology she specifies as Lingua Franca. These include baroufa ‘quarrel’, fantasia ‘pride, delusion’, mercanti ‘merchant’, and rabia
‘rage’.
2 Contact with Arabic
As mentioned above, a substantial proportion of Lingua Franca speakers appear
to have had Arabic as their first language. Inevitably, there would have been
transfer when they spoke Lingua Franca, although, given the shared history
of Arabic and European cultures in Sicily, Spain and other parts of the Mediterranean, it is perhaps hard to state unequivocally whether lexical influences
stem from the contact of Arabic with Lingua Franca directly, or its earlier contact with Romance languages. Pellegrini (1972) identified the many Arabic loanwords integrated into Italian, particularly in the realms of trade, conflict and
exploration. A number of these are included in the Dictionnaire (1830), including
magazino ‘shop’ from the Arabic, maḫāzīn ‘storage facility’, and fondaco ‘trading
post’ from the Arabic funduq ‘hotel, inn’. Both would already have been in use
in Italian, thus complicating further an etymological study of Lingua Franca’s
lexicon.
3 Contact-induced changes in Lingua Franca
Contact-induced changes in Lingua Franca with regard to Arabic are relatively
limited, evident predominantly in its lexicon but also, to some extent, in its phonology. It is perhaps even an overstatement to consider Arabic’s influence as
contact-induced change; rather it might be viewed simply as an additional lexifier.
3.1 Phonology
The relative lack of written record and potential unreliability of the sources’ excerpts of Lingua Franca makes the identification of a definitive phonemic inventory both difficult, and at times, inconclusive. Overall, Lingua Franca follows the
539
Joanna Nolan
phonology of Romance languages, predominantly Tuscan Italian, though with elements of Venetian and Spanish. Venetian influence is also evident in the Lingua
Franca tendency to omit final vowels following sonorants /l/, /n/ and /r/, as in
colazion instead of colazione ‘breakfast’. Both Venetian and Lingua Franca exhibit examples of degemination (e.g. tuto ‘all’ vs. Tuscan tutto) and voicing of
intervocalic stops – segredo ‘secret’ rather than segreto (Ursini 2011). The voicing of intervocalic *t is also consistent with Spanish, which appears also to have
had an influence on elements of Lingua Franca phonology. An example from
the Dictionnaire (1830: 63) that illustrates both the plosive voicing and the final
vowel omission is padron ‘master’, an epithet that recurs throughout the corpus
of attestations. However, in terms of the language’s vocalic system, Arabic appears to exert some influence. Cifoletti (2004) suggests that Arabic influence on
the realization of vocalic elements can be seen in the Dictionnaire: bonou from
the Italian buono ‘good’ evidences a simplification of the diphthong uo. Camus
Bergareche (1993: 444) confirms this simplification; for example Italian uovo ‘egg’,
duole ‘hurt’, buono ‘good’ are reduced in their Lingua Franca counterparts: obo,
dole, bono.
There is some evidence in the Dictionnaire of a reduction in the number of
qualities for short vowels in Lingua Franca from a typical five-vowel Romance
system to the more impoverished systems found in North African Arabic varieties, seen, for example in the frequent realization of final /e/ as 〈a〉, as in scoura
from scure ‘axe’ or gratzia from grazie ‘thanks’, or 〈i〉, as in sempri from sempre
‘always’ or grandi from grande ‘big’. Camus Bergareche (1993: 444) reinforces
this, citing the Lingua Franca words mouchou ‘much, many’, poudir, ‘to be able’,
and inglis ‘English’, with their roots in Spanish (mucho, poder, inglés) as evidence
of a reduction in vowel qualities as a result of contact with Arabic.
Unlike most (non-sonorant-final) words in Lingua Franca that have a typical
Romance vowel ending, Arabic-derived words generally retain their consonant
ending, as in rouss from ruzz ‘rice’ and maboul from mahbūl ‘stupid’ (Cifoletti
2004: 38). Non-Romance influence on Lingua Franca is also evidenced in the regular substitution of /b/ for /p/, which is lacking in the phonemic inventory of
most Arabic varieties. The Dictionnaire features a number of replacements of
this type (Cifoletti 2004: 38), as in esbinac ‘spinach’ and nabolitan ‘Neapolitan’.
Minervini (1996: 257–60) analyses the speech of the eponymous heroine of Giancarli’s (1545) La Zingana, and comments on the frequent substitution of /b/ for
/p/ and /v/, offering examples such as cattiba (cattiva ‘nasty’ in Italian), bericola
(pericolo ‘danger’ in Italian), the native Arabic of the character allegedly influencing her pronunciation. Given, however, that these examples come from a work
of fiction, they do not provide conclusive evidence of influence.
540
24 Mediterranean Lingua Franca
Perhaps given that Lingua Franca as attested is replete with abbreviation, ellipsis and omissions, it predictably features examples of aphaeresis. For example,
many Romance-derived items beginning with a syllable that resembles the Arabic definite article see this omitted in Lingua Franca. Examples include bassiador
for ambasciatore ‘ambassador’, bastantza for abbastanza ‘enough’, and rigar for
irrigare ‘to water’. As with many other linguistic features, however, the similarity between Lingua Franca and Venetian dialect must be considered, as some of
these words exist in a similarly abbreviated form in Venetian (Schuchardt 1909).
3.2 Lexicon
From my quantitative analysis of the material in the Dictionnaire and other available sources, it is apparent that there were very few Arabic lexemes in Lingua
Franca’s lexicon. Of the more than 2,100 entries in the Dictionnaire, 32 are of
Arabic origin, and of the 176 additional Lingua Franca lexemes identified in the
corpus of attestations, only nine have an Arabic etymology. However, a number
of the individual items are regularly repeated in the corpus, and, as such, Arabic
appears a more influential lexifier on a token than on a type basis.
Romance/non-Romance (often Arabic) doublets feature particularly in terms
of place names, and officialdom within the Regencies. The port of Tunis was
known by its French, Italian and Arabic names, seemingly interchangeably: La
Goulette, La Goletta, and Wādi l-Ḥalq ‘the gullet’. In his manual for future consuls, the outgoing English consul of Tripoli, Knecht, enumerates the hierarchies
within the Pasha’s household and city administration. Many of these involve a
combination of Arabic or Turkish and Italian, or perhaps Lingua Franca. Key positions include a hasnadawr Grande ed un hasndawr Piccolo – ‘a senior treasurer
and a junior treasurer (Pennell 1982: 97; author’s translation). Hasnadawr comes
from the Ottoman Turkish hazinedar or haznadar (from Arabic ḫāzin ad-dār)
‘Lord treasurer of the household’ (Gilson 1987).
Another example is Kecchia Grande and Kecchia Piccolo ‘chief administrator
and assistant administrator’ (Pennell 1982: 104). In the letters written by members
of the household of Richard Tully, British Consul to Tripoli at the close of the
eighteenth century, there is a reference to the “Great Chiah and the Little Chiah”
(Tully 1819: 70), surely an anglicization of the title. Kecchia appears to derive from
the Tunisian Arabic kāhiya ‘chief officer of an administrative district’ – kecchia
is an italianised (or, again, possibly Lingua Franca-influenced) orthography and
pronunciation. Similarly, sotto rais (from the Italian and Arabic literally meaning
‘under captain’) denoted the second in command of the harbour (Pennell 1982: 97,
100). The commander is referred to separately as the rays de la marina ‘chief of the
541
Joanna Nolan
port’ (Pennell 1982: 92). Again, one finds the combination of Arabic and Italian.
(Rays is spelt in two different ways,4 which highlights how, pre-standardisation
of European languages, orthography was erratic, even within a single document.)
Another example is a proverb regarding the eradication of plague by the Day of
St. John. Two variations on the proverb are cited – by Poiret (1802) and Rehbinder
(1800):
(1)
(Poiret 1802)
Saint Jean venir, Gandouf andar
‘[The day of] St. John comes, the plague leaves.’
(2) (Rehbinder 1800)
Saint Jean venir, buba andar
‘[The day of] St. John comes, the plague leaves.’
Selbach (2008: 44) remarks on how such varied nomenclature in Lingua Franca
“allowed for much room to manoeuver, and for speakers to mark their religious,
political and cultural identity”. Buba ‘plague’ appears to come originally from the
Greek, βουβών (boubōn) ‘groin’, suggesting yet another potential lexifying influence on Lingua Franca, while Gandouf plausibly derives from Arabic ɣunduba ~
ɣundūb ‘swollen tonsils’ (Schuchardt 1909: 72; cf. Selbach 2008: 45).
This example raises the issue of words already common to Arabic and Romance
languages, since contact between them, as discussed above, had been prolonged
and extensive. Similarly, the French (and possibly Italian or even Lingua Franca)
word avanie ‘fine, insult, affront’ occurs often in the corpus of attestations (e.g.
Pananti 1841 and Grandchamp 1920). Grandchamp (1920: xiii) defines it thus:
les avanies étaient des sommes d’argent que les pachas réclamaient aux
marchands des échelles sous les prétextes les plus divers, prétextes la plupart du temps injustes, parfois extrêmement bizarres
‘the fines were sums of money the Pashas demanded of the Levant merchants on various pretexts, pretexts that were for the most part unfair, and
at times extremely strange’ (author’s translation).
Although this word would appear to be derived from French, or at least a
Romance language, given that it was the creation of Ottoman elites, it seems
more likely that its origins are Turkish. This is confirmed by Pihan (1847) who
suggests that it actually originally derives from the Arabic hawān ‘contempt’,5
but who also states (Pihan 1847: 46):
4
5
The standard Arabic form is raʔīs.
This etymology is also favoured by Le Trésor de la langue francaise informatisé (Dendien 1994).
542
24 Mediterranean Lingua Franca
se dit également des impôts énormes que les Turcs font peser sur les Chrétiens dans le but de les humilier
‘it applies equally to the enormous taxes the Turks impose on Christians
with the goal of humiliating them’ (author’s translation).
Additionally, there are words that appear to have etymologies in multiple languages that are rarely translated, at least by English sources such as Tully (1819).
Two terms with similar meanings, firman ‘pass, decree’ and teschera ‘pass, edict’,
both issued by Ottoman or Arabic rulers, also bear remarkable similarity to Italian words with comparable meanings. Firman is from Arabic: firmān ~ faramān,
though originally Persian, and would have come into Arabic through Ottoman
Turkish, once again reinforcing how the languages spoken in the region were far
from discrete entities. However, firmare in Italian means ‘to sign’, and the Lingua
Franca translation of the French seing ‘signature’ and signature ‘signature’ in the
Dictionnaire is firmar. A decree or pass (allowing free passage or safe conduct)
would necessarily require an official signature. Teschera ‘pass, edict’ might appear to come from the Italian tessera ‘pass, ticket’ but there is also the Arabic
word taðkira ~ taðkara ~ tazkira ~ tazakara (all variant realizations of the same
item), which also means ‘permit’ or ‘ticket’. Both words seem integral to Barbary
life, and are not translated. Tully (1819: 258) writes:
It is still affirmed that he has a teskerra, or firman, with him for this unfortunate Bashaw. A teskerra is a written order from the Grand Signior, and is
held so sacred that every Musulman who receives it must obey its mandate,
even to death.
Perhaps the most iconic word in Lingua Franca is fantasia. It is mentioned by
multiple sources spanning more than two centuries (e.g. Haedo 1612; Broughton
1839). Although it appears to be a Romance word, Schuchardt (1909: 71) points
out that it is used in the Arabic sense of pride, arrogance, as in, for example,
Egyptian Arabic itfanṭaz ‘to give oneself airs’.
Collections in the UK National Archives provide some limited evidence of borrowings from Lingua Franca (as opposed to other Romance languages) into Arabic. Hopkins’ (1982) research in these archives focuses on two sets of English
(and later British) state papers relating to the Barbary regencies and Morocco,
including correspondence in Arabic, from the late sixteenth to late eighteenth
centuries.
Hopkins adds a glossary to his translations that demonstrates the extent to
which the Arabic letters from the Barbary States disproportionately feature loans
543
Joanna Nolan
from Romance languages. Hopkins (1982: x) comments that “[f]oreign words are
very common and seem to be used quite unselfconsciously”, and many of those
words Hopkins isolates are also listed in the Dictionnaire (1830) as Lingua Franca
words. Three such items – justisiya ‘justice’, markānti ‘merchant’ and zabantut
(sbendout in Lingua Franca, presumably from the Italian bandito) ‘pirate’ – occur
in a single letter of unspecified provenance but written to the King of England
in 1730 by a man claiming to be an Algerine trader in Tripoli (TNA: SP 71/23/51).
The incidence of three non-Arabic, Romance words, is noteworthy and it seems
plausible that these were Lingua Franca terms in such common usage that they
would be borrowed as a native language alternative, an example of mot juste
switching (Gardner-Chloros 2009: 32).
4 Conclusion
As demonstrated, the available corpus of writing in Lingua Franca – both documentary texts written by European visitors to the Barbary states and the dramatic works produced by contemporary authors – offers limited evidence of both
lexicon and grammar. This makes description of Lingua Franca challenging, and,
likewise, any concrete and substantiated analysis of its relationship with other,
particularly non-European, languages.
Nevertheless, this chapter has suggested how Arabic and Romance languages
influenced the emergence of Lingua Franca, specifically in terms of its lexicon
and phonology. Authors throughout the era of Lingua Franca’s existence, from
Haedo (1612) to Broughton (1839) and Frank (1850) reiterate that, despite its overwhelming Romance base, Lingua Franca was spoken predominantly by the often
Arabic-speaking slavemasters and rulers of the Barbary states. The plurilingual
character of the population of this region, both collectively and individually, compounds an already unclear picture, however, as the fluidity of Barbary society
led to European (and often Romance-language speaking) corsairs and diplomats
alike permeating its upper echelons (Haedo 1612: 9; Garcès 2011: 129).
The lexical influence of Arabic is most evident in the Romance/Arabic doublets
used in official terms and place names. These are often compound terms, such as
ra’ïs de la marina ‘captain of the port’. Warrington, English Consul to Tripoli in
the late eighteenth century, uses an Anglicised version of the phrase, rays marina,
suggesting the ubiquity of such doublets (TNA: FO 161/9) Further evidence from
the National Archives (Hopkins 1982) demonstrates that Lingua Franca words
were borrowed in correspondence from Arabic-speaking dignitaries in the Barbary Regencies to the English Secretary of State, showing evidence of Lingua
Franca borrowings in the written as well as the oral domain.
544
24 Mediterranean Lingua Franca
In terms of phonology, there seems to be evidence offered by the Dictionnaire
(1830) and the observation of Haedo (1612) of the influence of Arabic on Lingua
Franca. Haedo (1612: 24) also stated, with regard to the “Moors and Turks” that
“no saben ellos variar los modos, tiempos y casos”, ‘they don’t know about gender,
tenses and cases’ (author’s translation). Given that Lingua Franca lacks for the
most part any verbal inflection and an absence of cases, this might be, as Haedo
says, a result of contact, but it is typical of most pidgins and cannot be attributed
solely to contact with Arabic. Such lack of certainty applies more generally. Arabic evidently exerted some influence on the evolution of Lingua Franca in North
Africa, but not to the extent that it can straightforwardly be classified as contactinduced change.
Further reading
) Nolan (2018) gives a comprehensive English-language introduction to Lingua
Franca.
) Corré (2005) compiles a substantial number of texts featuring Lingua Franca,
offering translations from the original European languages and a selection of
articles analysing and contexualising Lingua Franca. He also provides a glossary, largely derived from the Dictionnaire (1830) and Schuchardt’s seminal
(1909) work.
) Schuchardt (1909) offers the earliest detailed historical and linguistic analysis
of Lingua Franca.
) More recent in-depth texts include Dakhlia’s (2008) rather subjective book,
Minervini’s (1996) comprehensive article that focuses predominantly on Italian and French sources, and Cifoletti’s (2004) forensic biography and analysis
of Lingua Franca, again based almost exlusively on Romance language sources.
) The key source authors are inevitably Haedo (1612), Broughton (1839), Pananti
(1841), and Frank (1850).
Abbreviations
HSA
TNA
Hugo Schuchardt Archiv
The National Archives
545
Joanna Nolan
References
Anonymous. 1830. Dictionnaire de la Langue Franque ou petit Mauresque, suivi de
quelques dialogues familiers, et d’un vocabulaire de mots arabes les plus usuels;
à l’usage des français en afrique. Marseille: Feissat et Demonchy.
Broughton, Elizabeth. 1839. Six years’ residence in Algiers. London: Saunders &
Otley.
Camus Bergareche, Bruno. 1993. El estudio de la lingua franca: Cuestiones pendientes. Revue de linguistique romane 57(227–228). 433–454.
Cifoletti, Guido. 2004. La lingua franca barbaresca. Rome: Il Calamo.
Clissold, Stephen. 1977. The Barbary slaves. London: Edek.
Corré, Alan D. 2005. A glossary of Lingua Franca. https://minds.wisconsin.edu/
bitstream/item/3920/go.html. (Accessed 01/01/2020).
D’Aranda, Emanuel. 1662. Relation de la captivité et libération de Sieur Emanuel
d’Aranda, jadis esclave à Alger. Brussels: Jean Mommart.
D’Arvieux, Laurent. 1735. Mémoires du Chevalier d’Arvieux. Paris: Jean-CharlesBaptiste Delespine.
Dakhlia, Jocelyne. 2008. Lingua Franca. Arles: Actes Sud.
Dan, Pierre. 1637. Histoire de la Barbarie, et de ses corsaires divisée en six livres.
Paris: P. Rocolet.
Davis, Robert C. 2004. Christian slaves, Muslim masters: White slavery in the
Mediterranean, the Barbary Coast and Italy, 1500–1800. London: Palgrave
Macmillan.
Dendien, Jacques. 1994. Trésor de la langue française informatisé. http://stella.atilf.
fr/. (Accessed 01/01/2020).
Duclos, Jeanne. 1992. Dictionnaire du français d’algérie: Français colonial,
pataouète, français des Pieds-Noirs. Paris: Editions Bonneton.
Frank, Louis. 1850. Algérie, états tripolitains, Tunis. Paris: Firmin Dido.
Garcès, Maria Antonia. 2011. An early modern dialogue with Islam: Antonio de
Sosa’s Topography of Algiers (1612). Trans. by Diana de Armas Wilson. Indiana:
University of Notre Dame.
Gardner-Chloros, Penelope. 2009. Code-switching. Cambridge: Cambridge University Press.
Giancarli, Gigio Artemio. 1545. La zingana. Mantova: V. Ruffinello.
Gilson, Erika H. 1987. The Turkish grammar of Thomas Vaughan: Ottoman Turkish
at the end of the XVIIth century. Wiesbaden: Harrassowitz.
Grandchamp, Pierre. 1920. La France en Tunisie à la fin du XVIe siècle (1582–1600).
Tunis: Société Anonyme de l’Imprimerie Rapide.
546
24 Mediterranean Lingua Franca
Grion, G. 1890. Farmacopea e lingua franca del Dugento. Archivio Glottologico
Italiano 12. 181–186.
Haedo, Diego de. 1612. Topographia e historia general de Argel. Valladolid.
Hall, Robert A. Jr. 1966. Pidgin and creole languages. Ithaca, NY: Cornell University
Press.
Hopkins, J. F. P. 1982. Letters from Barbary 1576–1744: Arabic documents in the
Public Record Office. Oxford: Oxford University Press.
Kahane, Henry Romanos & Renée Kahane. 1976. Lingua franca: The story of a
term. Romance Philology 30. 25–41.
Lanly, André. 1962. Le français d’Afrique du nord. Paris: Presses universitaires de
France.
Minervini, Laura. 1996. La lingua franca mediterranea: Plurilinguismo, mistilinguismo, pidginizzazione sulle coste del Mediterraneo tra tardo medioevo
e prima età moderna. Medioevo Romanzo 30. 231–301.
Molière, Jean-Baptiste Poquelin. 1667. Le sicilien. Paris: Jean Ribov.
Molière, Jean-Baptiste Poquelin. 1798. Le bourgeois gentilhomme. Marseille: Jean
Mossy.
Mori, Laura. 2016. Plurilinguismo, interferenza e marche acquisizionali in
“italiano di contatto” nella comunicazione transculturale del Mediterraneo
moderno. In Cristina Muru & Margherita Di Salvo (eds.), Dragomanni, sovrani
e mercanti: Pratiche linguistiche nelle relazioni politiche e commerciale del
Mediterraneo moderno, 23–72. Pisa: Edizione ETS.
Nolan, Joanna. 2018. Fact and fiction: A re-evaluation of Lingua Franca. London:
SOAS University of London. (Doctoral dissertation).
Pananti, Filippo. 1841. Avventure e osservazioni sopra le coste di Barberia. Milan:
Leonardo Ciardetti.
Pellegrini, Giovanni Battista. 1972. Gli arabismi nelle lingue neolatine con speciale
riguardo all’Italia. Brescia: Padeia.
Pennell, Richard. 1982. Tripoli in the mid eighteenth century: A guidebook to the
city in 1767. Revue d’histoire maghrébine 25–26. 91–121.
Pihan, Antoine Paulin. 1847. Glossaire des mots français de l’arabe, du persan et du
turc. Paris: Benjamin Duprat.
Plantet, Eugène. 1889. Correspondance des deys d’Alger avec la cour de France,
1579–1833. Paris: F. Alcan.
Poiret, Jean Louis Marie. 1802. Voyage en Barbarie ou lettres écrites de l’ancienne
Numidie. Paris: Callixte Voland.
Rehbinder, Johan von. 1800. Nachricten und Bemerkungen über den algerischen
Staat 1798–1800. Altona: Hammerich.
547
Joanna Nolan
Rossetti, Roberto. 1999. An introduction to lingua franca. Englishes: letterature
inglesi contemporanee 8. 42–62.
Savary de Brèves, François. 1628. Relation des voyages de monsieur de Brèves, tant
en Grèce, Terre Saincte et Aegypte, qu’aux Royaumes de Tunis et Arger. Paris:
Nicolas Gasse.
Schuchardt, Hugo. 1909. Pidgin and creole languages. Cambridge: Cambridge
University Press. Trans. by Glen G. Gilbert, 1980.
Selbach, Rachel. 2008. The superstrate is not always the lexifier: Lingua Franca
in the Barbary Coast 1530–1830. In Suzanne Michaelis (ed.), Roots of creole structures: Weighing the contribution of substrates and superstrates, 29–58.
Amsterdam: John Benjamins.
Tinniswood, Adrian. 2010. Pirates of Barbary: Corsairs, conquests and captivity in
the 17th century. London: Vintage Books.
Trivellato, Francesca. 2009. The familiarity of strangers: The Sephardic diaspora,
Livorno, and cross-cultural trade in the early modern period. New Haven: Yale
University Press.
Tully, Miss. 1819. Ten years’ residence at the court of Tripoli: Letters written during a
ten years’ residence at the court of Tripoli 1783–1995. Vol. 1. London: H. Colburn.
Ursini, Flavia. 2011. Dialetti veneti. In Raffaele Simone (ed.), Enciclopedia italiana.
Rome: Treccani.
Weiss, Gillian. 2011. Captives and corsairs: France and slavery in the early modern
Mediterranean. Stanford: University of California Press.
548
Part III
Domains of contact-induced
change across Arabic varieties
Chapter 25
New-dialect formation: The Amman
dialect
Enam Al-Wer
University of Essex
One fascinating outcome of dialect contact is the formation of totally new dialects
from scratch, using linguistic stock present in the input dialects, as well as creating
new combinations of features, and new features not present in the original input
varieties. This chapter traces the formation of one such case from Arabic, namely
the dialect of Amman, within the framework of the variationist paradigm and the
principles of new-dialect formation.
1 Contact and new-dialect formation
1.1 Background and principles
The emergence of new dialects is one of the possible outcomes of prolonged and
frequent contact between speakers of mutually intelligible but distinct varieties.
The best-known cases of varieties that emerged as a result of contact and mixture of linguistic elements from different dialectal stock are the so-called colonial varieties, namely those varieties of English, French, Spanish and Portuguese
which emerged in the former colonies in the Southern Hemisphere and the Americas.1 In addition to colonial situations, the establishment of new towns can also
lead to the development of new dialects; a case in point is Milton Keynes (UK),
which was investigated by Paul Kerswill.2 For Arabic, similar situations of contact are abundant, largely due to voluntary or forced displacement of populations,
growth of existing cities and the establishment of new ones. To date, however,
1
Among the studies that investigated such varieties are: Trudgill (2004), Gordon et al. (2004),
Sudbury (2000) and Schreier (2003) for English; Poirier (1994; cited in Trudgill 2004) for French;
Lipski (1994) and Penny (2000) for Spanish; and Mattoso (1972) for Portuguese.
2
See Kerswill & Williams (2005).
Enam Al-Wer. 2020. New-dialect formation: The Amman dialect. In Christopher
Lucas & Stefano Manfredi (eds.), Arabic and contact-induced change, 551–566. Berlin:
Language Science Press. DOI:10.5281/zenodo.3744549
Enam Al-Wer
the only study of a brand new dialect is the on-going investigation of the dialect
of Amman, the capital city of Jordan, which is anticipated to provide a model for
the study of dialect contact and koinéization in other burgeoning conurbations
elsewhere in the Arab World. The bulk of this chapter will be dedicated to the
details of this case.
Several other studies in Arab cities have focused on contact as a primary
agency through which innovations permeate the speech of migrant groups. Although no new dialects emerge in such situations, new patterns and interdialectal forms are common. For instance, Al-Essa (2009) reports that among the residents of the city of Jeddah, those who originally emigrated from various locations in Najd generally converge to the dialect of Jeddah, but also use innovations
that do not occur in the target dialect, such as the second person singular feminine suffix -ki in words ending in a consonant, as in ʔumm-ki for Najdi ʔummits and Jeddah ʔumm-ik ‘your (sg.f) mother’. Similarly, Alghamdi (2014) found
interdialectal forms of the diphthongs /ay/ and /aw/ (viz. narrow diphthongal
variants [ɛi], [ɔʊ]), as well as the monophthongs [ɛː] and [ɔː], in the speech of
Ghamdi migrants who originally came to Mecca from Al-Bāḥa in the southwest
of Saudi Arabia. In Casablanca, rapid urbanization led to immigration of large
numbers of groups from all over Morocco, and subsequent contact between different dialects. Hachimi (2007: 97) suggests that this situation resulted in “the
disruption of the rural/urban dichotomy that once dominated Moroccan dialects
and identities”, and the emergence of new categories of identification, which are
symbolized through the usage of a mixture of features from different dialects.
In this context, it is worth pointing out some methodological challenges concerning the measurement of contact as an independent variable in quantitative
sociolinguistics, and some improvements that have been made in research on
Arabic. Contact is often invoked as an explanatory factor in contact linguistics
in general, and has indeed been incorporated in theoretical formulations (e.g.
Thomason & Kaufman 1988). In quantitative sociolinguistics, however, analysis
of contact as a constraint on linguistic variation requires treating it as a variable from the outset of research, and finding ways to quantify it, in esssentially
the same way that social categories such as age, gender and class are factored
into the analysis. But how can contact be quantified? Recognizing the crucial
role that (dialect) contact plays in the structure of variation and mechanisms of
change, a number of quantitative studies have tested various methods of quantification. Al-Essa’s (2009) study, mentioned above, was the first known quantification of contact in studies of this sort. In order to do this, she measured the
speakers’ level of exposure to the target features through an index, consisting
of a four-point scale, which gave a numerical value to each speaker’s level of
552
25 New-dialect formation: The Amman dialect
contact. Four criteria were used to determine the numerical value assigned to
each speaker: friendships at school and work; involvement in neighbourhood
affairs; friendship with speakers of the target dialect; kinship and intermarriage
in the family (Al-Essa 2009: 208). Alghamdi’s (2014) study in Mecca utilized and
adapted Chambers’ (2000) concept of regionality, by devising a regionality index based on the speakers’ date of arrival in the city and place of residence. In
Al-Wer (2002a), I suggested that in some cases level of education may be treated
as an indication of level of contact with outside communities; and Horesh (2014)
elicited information that was indicative of levels of contact between the speakers’
L1 Arabic and L2 Hebrew, which were later converted into factor groups, one of
which was language of education, thus demonstrating that type of education can
also be used to measure contact.3
1.2 Theoretical framework
The study of the formation of new dialects is credited particularly to the work of
Peter Trudgill. In his Dialects in contact (1986) he laid the theoretical foundations
of research in the field, arguing that “face-to-face interaction” is a prerequisite
for linguistic adaptation and diffusion of linguistic innovations.4 Focusing on
the formation of New Zealand English, Trudgill (2004) suggests a three-stage
approach to dialect formation, which roughly corresponds to three successive
generations of speakers.5 These stages are very briefly summarized below, and
illustrated using examples from Amman in §2.3.6
Stage I (first generation): rudimentary leveling.
This stage stipulates that at the initial point of contact and interaction between adult speakers of different regional and social varieties, minority
and very localized linguistic features are leveled out.
Stage II (second generation): variability and mixing.
At this stage, the first locally-born generation of children are presented
with a plethora of features to choose from. Their speech contains considerable inter-individual and intra-individual variability, and new combinations of features.
3
Several additional doctoral theses completed at the University of Essex address this issue.
Trudgill (1986) integrated insights from Accommodation Theory (Giles 1973) in the study of
dialect contact.
5
In the same year, and based on the same data, the team working on the Origins of New Zealand
English (ONZE) project, in which Peter Trudgill participated, also published a co-authored
book on the topic (see Gordon et al. 2004).
6
Trudgill (2004: 83–128) discusses and illustrates each stage with data from the ONZE corpus.
4
553
Enam Al-Wer
Stage III (third generation): emergence of stable and relatively uniform dialect.
At this stage, focusing (Le Page & Tabouret-Keller 1985; see §1.3.2) gives
rise to a crystalized dialect.
Trudgill (2004: 149)7 concludes that the processes of dialect mixture and newdialect formation are not haphazard but “deterministic in nature”, “mechanical
and inevitable”, and that, in tabula rasa situations, social and attitudinal factors
do not play a role in the formation of new dialects.8 “Determinism” in new-dialect
formation and “the minor role that social factors, such as identity, play in tabula
rasa situations” have instigated a wide and interesting debate among scholars.
For instance, Tuten (2008: 261) proposes that “community identity formation and
koiné formation are simultaneous and mutually dependent processes”. Mufwene
(2008: 258) agrees that common identity “is not part of the processes that produce
new dialects”; but rather a by-product of it. Schneider (2008) elaborates on two
issues: the relationship between accommodation and identity, and “the changing
role of identity” in different colonial and postcolonial phases (2008: 262), pointing out cases of features from colonial varieties where the origins and spread of
these features coincided with “a heightened national or social awareness” (2008:
266). Bauer (2008) contests Trudgill’s implicit suggestion that accommodation
leads directly to dialect mixing, on the basis that individuals vary in the extent
to which they accommodate to others, and vary depending on the context; and
in some cases no accommodation takes place, that is, accommodation is sporadic.
He maintains that “it is not the accommodation as such that leads to dialect mixing; rather, it is the use that accommodation is put to by the next generation that
leads to dialect mixing” (2008: 272). On the role of identity, Bauer contends that
the very choice of a particular variant over another is indirectly an expression of
“complex kinds of identity” (2008: 273).9
1.3 Mechanisms
The mechanisms involved in new-dialect formation fall under two broad headings: koinéization and focusing. Below are brief explanations of these mechanisms, to be followed by illustrations from data from Amman in the relevant
sections.
7
See also Trudgill et al. (2000) and Trudgill (2008).
Cf. Labov’s (2001) principle of density.
9
For more details, see Bauer (2008); and for Trudgill’s responses to these points, see the discussion and rejoinder in Language in Society, 2008, vol. 37.
8
554
25 New-dialect formation: The Amman dialect
1.3.1 Koinéization
Trudgill (2004: 84–88) uses koinéization as an umbrella term to refer to five processes, which operate at the same or different stages in the formation of new dialects: (i) mixing, which, as the name suggests, involves the use of features which
originally came from different dialects; (ii) leveling, which involves gradual reduction and ultimate loss of minority features, that is, features that have least
representation in the dialect mix; (iii) unmarking, a sub-type of leveling, which
refers to the survival of unmarked and more regular forms even if they are not
the majority forms; (iv) interdialect development, which are forms that arise out
of interaction between different forms in the original mix, and can include phonetically, morphologically and syntactically intermediate forms; (v) reallocation,
which refers to the survival of more than one variant of the same feature, which
then undergoes reallocation in the new system; reallocation can be linguistic,
social or stylistic.
1.3.2 Focusing
This term was introduced into sociolinguistics by Le Page & Tabouret-Keller
(1985) to refer to the process whereby the new system “acquires norms and stability”. A focused dialect contrasts with a diffuse (or non-focused) linguistic situation, where there is no consensus over norms, and no stability of usage.10
2 Dialect formation in Amman
2.1 History and demographics
Amman has no traditional dialect simply because until relatively recently it had
no indigenous inhabitants. Though an important centre in ancient times, it remained largely deserted until the early years of the twentieth century.11
In 1921, it was designated as the capital of Transjordan (the land east of the
River Jordan), which became the Kingdom of Jordan in 1946. It thus attracted
migrants from other parts of the country, as well as from Palestine, Syria and
Lebanon. By the 1930s, the population had grown to 10,000 inhabitants, and by
10
11
See Le Page & Tabouret-Keller (1985: 181–182).
Amman’s ancient history is traced to the Ammonites (eighth century BCE), who called it Rabbath Ammon ‘the great (or royal) city of the Ammonites’; the Romans changed its name to
Philadelphia; the Arab Ummayads took over in the seventh century CE and restored its Semitic
name, Amman.
555
Enam Al-Wer
1946 it stood at 65,000. The early migrants consisted of two groups: (i) the majority were economic migrants (traders and shop keepers as well as labourers) or
civil servants, who were appointed in the state administration; (ii) the rest were
political activists (mostly individuals from Syria and Lebanon, which were then
still under French colonialist rule). The first group included families from both
sides of the River Jordan, namely indigenous Jordanians from the east side, and
Palestinians from all parts of historical Palestine. Statistics regarding numbers
from each group are unavailable, but I was able to collect fairly reliable information, through ethnographic interviews, about the provenance of a large sector
of the first generation of migrants. According to my research, the vast majority came from two particular locations: the Jordanian city of Sult (20 kilometres
northwest of Amman), and the Palestinian city of Nablus (110 kilometres from
Amman).
The city continued to receive waves of migrants from other locations in Jordan
and from Palestine, especially following the two wars in 1948 and 1967, which resulted in the occupation of historical Palestine, and the displacement of well over
three million Palestinians over the years, most of whom sought refuge in Jordan.
Between 1950 and 1990, the population of Amman doubled more than fifteen
times, to reach approximately two million by 2004. According to the 2018 census
estimate, the city is home to 2,554,923 Jordanian nationals, and 1,452,603 nonJordanians, that is, a total of over four million people live in the city currently.12
Given the political situation in the region, the population of Amman is forecast
to reach six million by 2025.
Against this demographic background, there are three important points to
note:
1. There is no geographically neutral variety of spoken Jordanian Arabic. All
speakers therefore use some form of local dialect, regardless of social class.
2. Whereas in neighbouring countries (Syria, Egypt, Lebanon, Palestine), the
dialect of the capital acts as a standard prestigious norm, Jordan never had
a linguistic centre of its own.
3. Jordanians, and Palestinians, generally identify themselves with the area
in which their forebears lived, rather than the locality in which they were
born and bred. However, recently a growing number of inhabitants of
Amman (particularly third and fourth generations) have begun to identify
themselves as “Ammanis”.
12
Department of Statistics, Jordan: http://dosweb.dos.gov.jo/DataBank/Population_Estimares/
PopulationEstimates.pdf (accessed 06/01/2020).
556
25 New-dialect formation: The Amman dialect
The emergence of a distinctive and focused dialect of Amman, in tandem with
the emerging Ammani identity, represents a radical shift in the sociolinguistic
patterns from a plethora of local varieties to a situation similar to that described
for neighbouring states in §1.
2.2 The Amman Project
This research traces the formation of this new dialect from inception to stabilization over three generations, spanning a period of approximately the last eighty
years. It initially focused on generational differences, by investigating the developments in the speech of three generations of families who originally came from
the Jordanian city of Sult and the Palestinian city of Nablus; this initial investigation confirmed the following hypotheses:
) A new dialect has emerged and its usage has stabilized.
) This dialect is unique – it grew as an outcome of the contact between Jordanian and urban Palestinian dialects, but is distinct from these input varieties.
) The formation of a distinctive dialect in Amman is closely associated with
relative stabilization in the population, possibly during 1970–1990, and the
development of an Ammani community with its own identity.13
The second phase of the research focused on the younger generation from affluent West Amman; and the final phase, ongoing, expands the sample to include
speakers from less affluent East Amman. Altogether, the research aims to collect
data from approximately 120 speakers, from both sides of the city. The project
on Amman itself is complemented by past and ongoing research (by myself and
others) on sociolinguistic trends in areas outside Amman, which provide two
valuable types of relevant information: (i) further evidence of the input varieties;
(ii) spreading of innovative features of the Amman dialect to other parts of the
country.
The framework of analysis adopted here is the Variationist Sociolinguistic
Paradigm, as described in Labov’s trilogy (1994; 2001; 2010). More specifically, the
project is guided by the principles of dialect contact and new-dialect formation,
as outlined in §1. As discussed earlier, one of the dominant issues in the study
of the formation of new dialects is the debate over the types of factors which determine it. The Amman project offers an opportunity to investigate these issues
13
Full details can be found in Al-Wer (2002a, 2003).
557
Enam Al-Wer
in detail, particularly because it is still possible to trace the different stages of
formation over the three generations of native inhabitants.
2.3 Formation over three generations
Based on the analysis of speech samples from three generations of Ammanis, the
formation of the dialect is a textbook case. Many of the processes of koinéization
explained above are operative, as will be demonstrated presently.
2.3.1 Stage I: first generation
The first generation arrived in Amman during the 1930s as adults. The most noticeable aspect of their speech is that it can easily be identified with the original
dialects of the places from which they migrated, while localized features are leveled out (cf. rudimentary leveling; Trudgill 2004). The features which are lost at
this stage are summarized below.
Jordanian input. Traditional Jordanian dialects, including the dialect of Sult,
are known to affricate /k/ to [ʧ] in front-vowel environments generally, as in
/keːf/ > [ʧeːf] > ‘how’. This feature is still widely used, especially in northern
varieties, as well as in Sult14 – where most of the early migrants in Amman came
from. Already in the first generation, this feature is completely lost; all instances
of this variable were rendered with [k]. In other words, first-generation speakers
deaffricate /k/.15 Although conditional affrication of /k/ is fairly widespread in
the region’s rural dialects, and is certainly not a minority feature in Jordanian
dialects, its use is heavily stigmatized, and none of the urban dialects have it.
Stigmatization is the likely reason that motivates the loss of this feature.
Also characteristic of the traditional dialects is the maintenance of a gender distinction in the second and third person plural pronouns, and pronominal, verbal
and nominal suffixes. For example: ʔintu ‘you (pl.m)’, ʔintin ‘you (pl.f)’; ʔummhum ‘their (m) mother’, ʔumm-hin ‘their (f) mother’; rāḥu ‘they (m) have left’,
rāḥin ‘they (f) have left’; ḥilwīn ‘pretty (pl.m)’, ḥilwāt ‘pretty (pl.f)’. What we
find in Amman is gender neutralization in these forms, such that the masculine
form is used to refer to both genders. The traditional system is currently variable in all major Jordanian cities, and seems to be giving way to a neutralized
form, as in Amman, which is an indication that Amman has become a focal point
from which linguistic innovations radiate. No particular social value is attached
14
On this feature in Traditional Jordanian dialects, see Al-Hawamdeh (2015), Herin (2010) and
Herin & Al-Wer (2013).
15
See Al-Wer (2007) for more details.
558
25 New-dialect formation: The Amman dialect
to the traditional feature (maintenance of gender distinction), although there is
awareness that it is characteristically found in provincial towns and villages. The
observation that it has become variable in many cities and towns means that it
is also becoming a minority feature in urban areas in particular. The change affecting this feature in Jordanian dialects in general may be described as a form of
simplification, where the number of distinct forms in the paradigm as a whole is
reduced. Additionally, none of the urban Palestinian dialects maintain a gender
distinction in these categories. In contact situations especially, the direction of
change is normally towards the simpler system (the “Simplification Preference”;
Lass 1997: 253).16
Palestinian input. In urban Palestinian, the high-frequency terms mbāriḥ ‘yesterday’ and sāʕa ‘hour/time’ are pronounced with raised vowels: mbēriḥ and sēʕa.
This is an extremely marked pronunciation in the context of Jordan, as no Jordanian dialect has it. It is also a feature that is overtly commented upon, and often
used to mimic dialects that have it. Extreme raising of /ā/ generally is a hallmark
of many urban Palestinian dialects, most notably in the dialect of Jerusalem; as
will be explained later, third-generation speakers with urban Palestinian heritage
use considerably lower variants than first- and second-generation speakers from
the same group. It is possible that lowering in these high-frequency items in the
first generation is the onset of the change that escalated in successive generations. In the first generation of this group, the speakers change their pronunciation in these items only, but continue to use noticeably higher variants of /ā/ in
other items, for example, ‘Amman’ is pronounced as [ʕəmmɛːn]; fālit ‘loose’ is
pronounced as [fɛːlɪt].
2.3.2 Stage II: second generation
This is the first locally-born generation; the majority of the speakers in the sample fall in this category, or arrived as very young children (under ten). The speech
of members of this generation shows extreme inter-speaker and intra-speaker
variability, and a mixture of features from both norms (Jordanian and urban
Palestinian). For example, the same speaker is found to use the second person
plural pronominal suffixes -ku (Jordanian) and -kon (Palestinian), e.g. kēf ḥāl-ku
~ kīf ḥāl-kon ‘how are you (pl)’. In this example, we also find alternation in the
vowel of the item kēf ~kīf ‘how’; the former is Jordanian while the latter is typical
16
It should be pointed out that the urban Palestinian dialect, similarly to all city dialects in the
region, has the -on and -kon endings, which are used with both genders, rather than masculine
-um and -ku, which are the koiné forms in modern Jordanian dialects; see Al-Wer (2003) for
more details about this feature.
559
Enam Al-Wer
of urban Palestinian (and urban Levantine in general). The data also contained a
mixture of Jordanian and Palestinian third person plural suffixes -hum and -hon,
e.g. šift-hum ~ šift-hon ‘I have seen them’. At the level of phonology, speakers in
this generation use a mixture of Jordanian [ɡ] and urban Palestinian [ʔ], which
are variants of historical /q/; and a mixture of interdental and stop counterparts
of /θ/, /ð/ and /ð̣/. Importantly, in this generation there is a complication in sociolinguistic correlations: whereas in the first generation there is a one-to-one relationship between origin and the dialect used, in the second generation certain
groups from both backgrounds use features characteristic of the other group’s
dialect. The particular sub-groups that do this are the Jordanian women, who in
this generation use Palestinian [ʔ] almost consistently, as well as a high rate of
the stop variants of interdentals (see above), and use both Jordanian ʔiḥna and
Palestinian niḥna ‘we’. The second most divergent group (from their heritage
dialect) is Palestinian men; they use Jordanian [ɡ] at a rate of 50%, or more in
some cases. The remaining groups, Jordanian men and Palestinian women, are
considerably more conservative with respect to their heritage variants, although
they too are variable. What this pattern shows is that gender emerges as an important social factor in this generation, in addition to dialectal heritage, which
continues to influence individuals’ behaviour, but interacts with gender at the
same time.
2.3.3 Stage III: third generation
Third-generation Ammanis were all born in the city (in the 1970s). They diverge
from their parents’ and grandparents’ dialects, and speak a clearly distinct dialect,
regardless of their own dialectal heritage. The mixture and variability we saw in
the second generation is much reduced in the third generation; there is, instead,
stability in the usage of many features, including intermediate fudged forms, new
patterns, and new features that were not present in the input varieties. The third
generation agree on the characteristics of Ammani, and have intuitions as to
what you can and cannot say in this dialect. Importantly, they express affiliation
with the city; for instance, they identify themselves as “Ammanis”, by which they
mean that they are native to the city. In other words, the formation of the dialect
is simultaneously a formation of a community.
In this generation, gender emerges as a major organizing category; for instance, all the women in this generation, regardless of dialectal heritage, use
[ʔ] consistently, while the men continue to use both [ɡ] and [ʔ]. The variability in men’s speech is constrained by context and interlocutor for the most part;
560
25 New-dialect formation: The Amman dialect
whereas the speech of women is not subject to these constraints.17 The development in the use of variants of the variable (q), as explained above, is a clear example of a variable that has undergone social and stylistic reallocation (see §1
above) in the sense that both variants [ɡ] and [ʔ], which originally come from
different dialects in the input varieties, have survived the koinéization process
but no longer signify ethnicity or dialectal background straightforwardly; the
use of one or the other is now subject to layers of constraints. As far as the interdental sounds are concerned, both gender groups use the stop variants more often than the interdental variants. But while the men vary between affricate [ʤ]
and fricative [ʒ] of the variable (ǧ), the women use [ʒ] almost consistently.
In addition to the features listed under stage III, the following features are at
an advanced stage of focusing in the new dialect:
) The feminine ending -a. The input varieties differ in the phonology and
phonetics of the realization of the feminine ending in the unbound state.
In traditional Jordanian, the low vowel [a] is the default choice, except after
coronal sounds, where it is raised to [ɛ]. In urban Palestinian dialects, the
default choice is [e], or raised [e̝], except after pharyngeal and emphatic
consonants in general.18 In Amman, the third generation consistently use
a fudged form made up of urban Palestinian phonology and the Jordanian
phonetic property of the raised vowel, such that they raise /a/ to [ɛ] except
after back sounds, e.g. mukinsɛ ‘broom’, ḥilwɛ ‘pretty’, ṣaʕbɛ ‘difficult’, but
rāyḥa ‘has gone/is going’; žāmʕa ‘university’; ṣulṭa ‘authority’.
) In morphophonology, a new form of the second person plural suffix has
emerged, and is used consistently. The input forms are: Jordanian -ku, as
in gultilku ‘I have told you (pl)’; and urban Palestinian -kon, as in ʕmiltilkon
‘I have made for you (pl)’. The form that has been focused in Amman is
-kum, thus ʕindkum ‘you (pl) have’; ḥakētilkum ‘I have told you (pl)’. The
success of this form, rather than one from the input varieties, is explained
with reference to markedness and simplification.19
) In morphology, the input varieties differ in the conjugation of the third person masculine imperfect verb form. In both dialects, Jordanian and Palestinian, the imperfect takes a b- prefix, but whereas in Jordanian dialects
17
Details about this feature can be found in Al-Wer & Herin (2011).
A preceding /r/ blocks raising in general unless there is an /i/-type vowel in the environment;
for a complete account of the phonology of the feminine ending, see Al-Wer et al. (2015).
19
For analysis of this development, see the full details in Al-Wer (2003).
18
561
Enam Al-Wer
yod is dropped from the stem in the b-imperfect in all environments, in
Palestinian dialects it is dropped in open syllables only. For example: Jordanian biḥki ‘he talks’, binuṭṭu ‘they jump’; urban Palestinian byiḥki, binuṭṭu. Ammanis (third generation) drop yod everywhere except where it
carries person information, namely in glottal-initial verbs ʔakal ‘to eat’,
and ʔaxað ‘to take’; thus we get biḥki, binuṭṭu, but byākul ‘he eats’ (stem
ʔakal ‘to eat’), byāḫdu ‘they take’ (stem ʔaḫað ‘to take’).20
3 Conclusion
The formation of the Amman dialect is simultaneously the formation of a community; and the social factors involved in the formation of the dialect evolve and
realign accordingly. One of the most interesting aspects of this process is that
none of the factors become totally irrelevant. For instance, dialectal heritage –
which, in the case in hand, coincides with ethnicity (Jordanian/Palestinian) – is
the most important predictor in the speech of the first generation. In the second
generation, gender emerges as an important factor, but the linguistic developments at this stage can only be understood as an interaction between the old
and new social constraints; for instance, in stage II, it is not merely women who
use [ʔ] rather than [ɡ], but it is Jordanian women who diverge from their heritage variant; and it is not the behaviour of men in general that explains the
evolution in the re-distribution of these variants, but specifically the behaviour
Palestinian men. These two sub-groups (Jordanian women and Palestinian men)
are responsible for the diversification of, firstly, their respective group’s linguistic repertoire and consequently the repertoire of the linguistic system that is
passed on to the next generation. In stage III, the third generation’s behaviour
responds to two riders: the system inherited from their parents and the changes
in the socio-political environment around them. A further realignment of social
factors occurs in response, and new constraints are added to the old pile; at this
stage, the inherited identifications of the variants involved – that is, [ɡ] is Jordanian and appropriate for men, [ʔ] is Palestinian and appropriate for women –
are reformulated through the addition of further new constraints, namely context and interlocutors. Consequently, the usage of the variants involved is redistributed according to style,21 they acquire additional identifications and social
20
There are further complications and variations in the conjugations of these verbs; for these
details see Al-Wer (2014).
21
Style as a correlate of linguistic usage can mean different things; here I use it to refer to context
(as in Labov 1972), and audience or interlocutor (as in Bell 1984). For details of how style evolved
as a sociolinguistic correlate, see Eckert & Rickford (2001).
562
25 New-dialect formation: The Amman dialect
meanings, and the social constraints are realigned, such that the role of ethnicity
becomes subsidiary, while gender and style are the major organizing factors. The
younger generation now define [ʔ] as “Ammani”, and [ɡ] as “authentic Jordanian”. The meaning of “Jordanian” itself is often negotiated and expanded beyond
the limits of ethnicity to denote a regional identity, recognizing citizenship as
the primary defining component of membership in this group, although the old
meaning (those whose roots lie on the east side of the river) is not obliterated
altogether.22 A further realignment of social factors in Amman involves type of
profession, which is emerging as a constraint. This may have been precipitated
by the expansion of the private sector over the past two decades or so, especially
banking and the service industry in general, and the tourism industry. According
to preliminary analysis of recently collected data, different types of employment,
within and across the two sectors, fall within the realms of different linguistic
markets.
The context in which the Amman dialect was formed was tabula rasa in the
sense that there was no pre-existing Amman dialect. The obvious difference from,
say the tabula rasa colonial situations, is that the early settlers in Amman were
not isolated from their original communities or from Arabic speakers in the surrounding areas; social factors definitely play a role in the formation of the dialect
in this case. The question therefore is not whether social and attitudinal factors
are involved, but rather which social factors, how they evolved, and their relative
importance.
Further reading
) Al-Wer (2011) provides a brief description of the Amman dialect.
) Al-Wer (2007) summarizes the processes and dynamics of the formation of the
dialect of Amman, along with a list of thirteen linguistic features that have
been focused in this dialect.
) Al-Wer (2002b) focuses on the long vowels and the realization of the feminine
ending in the newly formed dialect of Amman.
22
The question of “who is Jordanian” is, for many, a sensitive issue, which has often caused
heated debates on various media platforms.
563
Enam Al-Wer
Abbreviations
1, 2, 3
f
L1
L2
m
ONZE
pl
sg
1st, 2nd, 3rd person
feminine
first language
second language
masculine
Origins of New Zealand English project
plural
singular
References
Al-Essa, Aziza. 2009. When Najd meets Hijaz: Dialect contact in Jeddah. In Enam
Al-Wer & Rudolf De Jong (eds.), Arabic dialectology: In honour of Clive Holes
on the occasion of his sixtieth birthday, 203–222. Leiden: Brill.
Al-Hawamdeh, Areej. 2015. A sociolinguistic investigation of two Horani features
in Souf, Jordan. Colchester: University of Essex. (Doctoral dissertation).
Al-Wer, Enam. 2002a. Education as a speaker variable. In Aleya Rouchdy (ed.),
Language contact and language conflict in Arabic: Variations on a sociolinguistic
theme, 41–53. London: Routledge.
Al-Wer, Enam. 2002b. Jordanian and Palestinian dialects in contact: Vowel raising
in Amman. In Mari Jones & Edith Esch (eds.), Language change: The interplay of internal, external and extra-linguistic factors, 63–79. Berlin: Mouton de
Gruyter.
Al-Wer, Enam. 2003. New dialect formation: The focusing of -kum in Amman. In
David Britain & Jenny Cheshire (eds.), Social dialectology: In honour of Peter
Trudgill, 59–67. Amsterdam: Benjamins.
Al-Wer, Enam. 2007. The formation of the dialect of Amman: From chaos to order.
In Catherine Miller, Enam Al-Wer, Dominique Caubet & Janet C. E. Watson
(eds.), Arabic in the city: Issues in dialect contact and language variation, 55–76.
London: Routledge.
Al-Wer, Enam. 2011. Amman Arabic. In Lutz Edzard & Rudolf de Jong (eds.), Encyclopedia of Arabic language and linguistics, online edn. Leiden: Brill.
Al-Wer, Enam. 2014. Yod-dropping in b-imperfect forms in Amman. In Reem
Khamis-Dakwar & Karen Froud (eds.), Perspectives on Arabic Linguistics XXVI:
Papers from the Annual Symposium on Arabic Linguistics, New York, 2012, 29–
44. Amsterdam: Benjamins.
564
25 New-dialect formation: The Amman dialect
Al-Wer, Enam & Bruno Herin. 2011. The lifecycle of Qaf in Jordan. Langage et
société 138. 59–76.
Al-Wer, Enam, Uri Horesh, Bruno Herin & Maria Fanis. 2015. How Arabic
regional features become sectarian features: Jordan as a case study. Zeitschrift
für Arabische Linguistik 62. 68–87.
Alghamdi, Najla Manie. 2014. A sociolinguistic study of dialect contact in Arabia: Ghamdi immigrants in Mecca. Colchester: University of Essex. (Doctoral
dissertation).
Bauer, Laurie. 2008. A question of identity. Language in Society 37(2). 270–273.
Bell, Alan. 1984. Language style as audience design. Language in Society 13(2).
145–204.
Chambers, Jack. 2000. Region and language variation. English World-Wide 21. 1–
31.
Eckert, Penelope & John R. Rickford. 2001. Style and sociolinguistic variation.
Cambridge: Cambridge University Press.
Giles, Howard. 1973. Accent mobility: A model and some data. Anthropological
Linguistics 15(2). 87–105.
Gordon, Elizabeth, Lyle Campbell, Jennifer Hay, Margaret MacLagan, Andrea
Sudbury & Peter Trudgill. 2004. New Zealand English: Its origin and evolution.
Cambridge: Cambridge University Press.
Hachimi, Atiqa. 2007. Becoming Casablancan: Fessis in Casablanca as a case
study. In Catherine Miller, Enam Al-Wer, Dominique Caubet & Janet C. E.
Watson (eds.), Arabic in the city: Issues in dialect contact and language variation,
97–122. London: Routledge.
Herin, Bruno. 2010. Le parler arabe de Salt (Jordanie). Brussels: Université Libre
de Bruxelles. (Doctoral dissertation).
Herin, Bruno & Enam Al-Wer. 2013. From phonological variation to grammatical
change: Depalatalisation of /č/ in Salti. In Clive Holes & Rudolf De Jong (eds.),
Ingham of Arabia: A collection of articles presented as a tribute to the career of
Bruce Ingham, 55–73. Leiden: Brill.
Horesh, Uri. 2014. Phonological outcomes of language contact in the Palestinian
Arabic dialect of Jaffa. Colchester: University of Essex. (Doctoral dissertation).
Kerswill, Paul & Ann Williams. 2005. New towns and koineization: Linguistic
and social correlates. Linguistics 43. 1023–1048.
Labov, William. 1972. Sociolinguistic patterns. Oxford: Blackwell.
Labov, William. 1994. Principles of linguistic change. Vol. 1: Internal factors. Oxford:
Blackwell.
Labov, William. 2001. Principles of linguistic change. Vol. 2: Social factors.
Chichester: Wiley-Blackwell.
565
Enam Al-Wer
Labov, William. 2010. Principles of linguistic change. Vol. 3: Cognitive and cultural
factors. Chichester: Wiley-Blackwell.
Lass, Roger. 1997. Historical linguistics and language change. Cambridge:
Cambridge University Press.
Le Page, R. B. & Andrée Tabouret-Keller. 1985. Acts of identity: Creole-based
approaches to language and ethnicity. Cambridge: Cambridge University Press.
Lipski, John. 1994. Latin American Spanish. Harlow: Longman.
Mattoso, Joaquim. 1972. The Portuguese language. Chicago: University of Chicago
Press.
Mufwene, Salikoko S. 2008. Colonization, population contacts, and the
emergence of new language varieties: A response to Peter Trudgill. Language
in Society 37(2). 254–258.
Penny, Ralph. 2000. Variation and change in Spanish. Cambridge: Cambridge
University Press.
Poirier, Claude. 1994. La langue parlée en Nouvelle-France: Vers une convergence
des explications. In Raymond Mougeon & Édouard Beniak (eds.), Les origines
du français québécois, 237–273. Sainte-Foy, Québec: Presses de l’Université
Laval.
Schneider, Edgar W. 2008. Accommodation versus identity? A response to Peter
Trudgill. Language in Society 37(2). 262–267.
Schreier, Daniel. 2003. Isolation and language change: Contemporary and sociohistorical evidence from Tristan da Cunha English. London: Macmillan.
Sudbury, Andrea. 2000. Dialect contact and koineisation in the Falkland Islands:
Development of a Southern Hemisphere variety? Colchester: University of Essex.
(Doctoral dissertation).
Thomason, Sarah G. & Terrence Kaufman. 1988. Language contact, creolization
and genetic linguistics. Berkeley: University of California Press.
Trudgill, Peter. 1986. Dialects in contact. Oxford: Wiley-Blackwell.
Trudgill, Peter. 2004. New-dialect formation: The inevitability of colonial Englishes.
Oxford: Oxford University Press.
Trudgill, Peter. 2008. Colonial dialect contact in the history of European
languages: On the irrelevance of identity to new-dialect formation. Language
in Society 37. 241–254.
Trudgill, Peter, Elizabeth Gordon, Gillian Lewis & Margaret Maclagan. 2000.
Determinism in new-dialect formation and the genesis of New Zealand
English. Journal of Linguistics 36. 299–318.
Tuten, Donald. 2008. Identity formation and accommodation: Sequential and
simultaneous relations. Language in Society 37(2). 259–262.
566
Chapter 26
Dialect contact and phonological change
William M. Cotter
University of Arizona
This chapter examines phonological and phonetic changes that have been documented and analyzed in spoken Arabic varieties, occurring as a result of dialect
contact. The factors contributing to dialect contact in Arabic-speaking communities vary, from economic migration which has encouraged individuals to move into
new dialect areas seeking work, to migration that stems from political violence
and upheaval. These diverse factors have contributed to the large-scale migration
of Arabic speakers to other parts of the Arabic speaking world. As a result, dialect
contact is rampant, and decades of Arabic sociolinguistic research have shown that
the phonological and phonetic effects of these contact situations have been quite
profound.
1 Introduction
In this chapter, I discuss research that has examined the outcomes of Arabic
dialect contact and the influence of contact on phonological change in spoken
Arabic varieties. This chapter also discusses the interface between phonology
and phonetics, and the effect of contact on these areas of the linguistic system.
Given space constraints, I discuss only a portion of the published work in these
areas, giving some priority to recent doctoral dissertations that have contributed
to this body of research. Further, I exclude work that has investigated the effects of contact on the morphology and syntax of Arabic (e.g. Al-Wer et al. 2015;
Gafter & Horesh 2015; Leddy-Cecere, this volume; Lucas, this volume; Manfredi,
this volume).
Although Arabic sociolinguistics is an increasingly robust area of linguistic
research, limiting my discussion to cases of contact-induced phonological and
phonetic change is perhaps unsurprising, given the scholarly history of dialectcontact research and its place within sociolinguistics. Sociolinguistics has made
William M. Cotter. 2020. Dialect contact and phonological change. In Christopher
Lucas & Stefano Manfredi (eds.), Arabic and contact-induced change, 567–581. Berlin:
Language Science Press. DOI:10.5281/zenodo.3744551
William M. Cotter
great progress towards the goal of analyzing the full scope of variation in languages around the world. However, historically, and to some extent still today,
examinations of variation and change in the realms of phonology and phonetics
have been the meat and potatoes of sociolinguistic work. I would argue that this
is true of Arabic sociolinguistic work as well.
From Labov’s (1963) early work on Martha’s Vineyard, phonetics and phonology have been at the heart of analyses of dialect contact. As a result, much of
what we know about Arabic dialect contact has stemmed from earlier foundational research on dialect contact in the English-speaking world. Within this
work on English, research by Milroy (1987), Trudgill (1986; 2004), Britain (2002),
and Britain & Trudgill (2009), among many others, has shown how dialect contact often plays out, and how that contact influences language variation and
change.
However, research on Arabic has moved beyond simply testing the hypotheses put forward by scholars of English dialect contact, playing its own role in
refining sociolinguistic theory. Notably, Arabic sociolinguistics has refined our
understanding of diglossia (Ferguson 1959). Ibrahim (1986) and Haeri (2000) have
reoriented our understanding of Arabic diglossia from Ferguson’s High–Low dichotomy to one that draws on locally meaningful understandings of linguistic
prestige. In doing so, this work has moved our discussion away from analyzing
Arabic through the lens of “standard” or “nonstandard” varieties or variants, setting the stage for decades of research that has examined contact-induced change
in Arabic varieties.
Before moving on to a discussion of a number of specific cases of Arabic dialect contact, I briefly address the potential limitations of Van Coetsem’s (1988;
2000) framework for discussions of dialect contact, as opposed to language contact. After discussing Van Coetsem’s approach, I shift my focus to discuss Arabic
dialect contact through a theoretical lens that has proven productive in earlier
sociolinguistic work (Trudgill 1986; 2004).
In analyzing phonological change as a result of dialect contact, Van Coetsem’s
framework presents a number of possible challenges. One specific issue is that in
many cases, a clear distinction between the borrowing or imposition of linguistic forms is challenging to establish in cases of Arabic dialect contact. Scholars
may encounter challenges in attempting to assert the agentivity of the recipient language in making a case for the borrowing of, for example, aspects of a
dialect’s phonology into the phonology of another dialect. Asserting the agentivity of the source language in making the case for imposition is similarly challenging. These challenges stem from the cognitive orientation of Van Coetsem’s
568
26 Dialect contact and phonological change
framework, which, as Lucas (2015: 521) notes, is not based on social realities or
variation in the power and prestige that a given dialect or language may hold.
The approach that many scholars within sociolinguistics and allied fields like
linguistic anthropology have taken is, in contrast, inherently social. We concern
ourselves with the social life of language, and although we do not discredit cognitive approaches to language acquisition and use, in much of the work on dialect
contact, we have foregrounded social factors in our analyses of language change.
However, it is worth noting that within sociolinguistic research on second dialect acquisition, researchers have highlighted the role of social factors, as well
as the constraints placed on acquisition by the linguistic system (e.g. Nycz 2013;
2016).
With the above discussion in mind, I argue that Van Coetsem’s framework
is less readily applicable to the cases that I describe in this chapter. Instead, I
suggest that outcomes of Arabic dialect contact are better analyzed through the
framework advocated for within sociolinguistics. It is to that framework that I
now turn.
As Trudgill (2004) notes in discussing new dialect formation, dialect contact
often progresses in stages. One of the earliest stages in this process is leveling
(Trudgill 2004: 83), which results in the reduction of forms from a given dialect.
These forms may be, but do not have to be, socially marked, e.g. affrication of /k/
to [č] in certain Arabic dialects. Most importantly for Trudgill, during leveling
certain variants of a given feature will supplant others (Trudgill 2004: 85). As
a result, forms that are socially marked may be leveled out, while unmarked
forms may survive even if they were not a majority variant. In those cases where
socially marked forms are present, they are often reduced across generations.
Trudgill also describes processes of interdialect development, where forms arise
out of the interaction between dialects, such as reallocation, where surviving
forms are reallocated in some way, and focusing, whereby a new variety born
out of contact begins to stabilize (Trudgill 2004).
What I feel that this framework offers in discussions of Arabic dialect contact
is an acknowledgement of the social issues that may influence linguistic change,
especially in situations of contact. In the remainder of this chapter I discuss cases
of contact-induced change in Arabic varieties. In doing so, I draw on sociolinguistic understandings of how contact-induced changes take hold and progress.
569
William M. Cotter
2 Contact-induced changes in the phonology of Arabic
dialects
When discussing Arabic dialect contact, a brief discussion of the typology of
Arabic is useful, as it provides a shared lexicon for discussing the outcomes of
contact. Cadora (1992) offers an ecolinguistic taxonomy of Arabic, describing a
continuum of Arabic varieties containing linguistic features ranging from what
he describes as Bedouin in provenance, to those that can be considered sedentary.
In presenting a related contrast, Cadora describes features that situate dialects
as being urban versus those that are rural.
However, what Cadora offers is not a hard and fast classification of Arabic
varieties. Instead, his typology highlights linguistic features that typically group
together within dialect types, providing a way to conceptualize the similarities
and differences across these varieties. Importantly for this chapter, sites of contact between Arabic dialects are also often sites of contact between types of dialects as well.
In my own work on Palestinian Arabic this has been the case, with Gaza City
offering one example of contact between different Palestinian Arabic dialects. Today, the dialect of Gaza City has both Bedouin and urban sedentary features, dialect types that likely came into contact in Gaza as a result of Palestinian refugee
migration (see de Jong’s 2000 discussion of Gaza City). This contact is undeniable
given Gaza’s current demographic reality, which suggests that its population is
roughly 70% refugee.1 It is also unsurprising given that Gaza has long been a site
of contact. This history of contact has resulted in a city dialect that looks different than other urban Palestinian varieties spoken in major cities like Jerusalem
or Nablus.
The above example serves as a way of framing the linguistic discussion of
contact-induced phonological change provided below. I begin by covering documented consonantal changes that have grown out of contact, before moving on
to vocalic changes and the need for additional research in this area as studies of
Arabic contact move forward.
1
This figure has been reported by the United Nations Relief Works Agency for Palestinian
Refugees but only reflects refugees that have actually registered with the U.N. (https://www.
unrwa.org/where-we-work/gaza-strip, accessed 07/01/2020). Other estimates place the percentage of Gaza’s population that are refugees as closer to 80%.
570
26 Dialect contact and phonological change
2.1 Consonantal changes
One of the most widely discussed linguistic features within work on Arabic dialect contact has been the variable realization of the voiceless uvular stop /q/.
Motivation for the scholarly interest in /q/ likely stems from a number of factors.
First, the phoneme has a wide range of dialectal variation, with dialectal realizations including a true voiceless uvular [q], as well as [k, ɡ, ʔ] and an additional
[k] variant articulated between a velar and uvular (Shahin 2011). Second, interest in /q/ is also likely due to the high social salience of its variation in many
Arabic-speaking communities (see Hachimi 2012; Cotter & Horesh 2015).
The result is that /q/ has been one of the most heavily studied features in Arabic sociolinguistics. Variation and change in /q/ has been discussed in a number of
different communities throughout the Arabic speaking world, including: Palestine (Abd El-Jawad 1987; Al-Shareef 2002; Cotter & Horesh 2015; Cotter 2016);
Egypt (Haeri 1997); Iraq (Blanc 1964; Abu-Haidar 1991);2 Jordan (Abd El-Jawad
1981; Al-Wer 2007; Al-Wer & Herin 2011); Morocco (Hachimi 2007; 2012); and
Bahrain (Holes 1987), among others.
What these cases suggest are robust processes of linguistic change in the realization of /q/ coming as a result of factors such as migration and dialect contact.
While the social patterning of these changes (e.g. stratified along age, gender, or
sectarian lines) has been as diverse as the communities in which /q/ has been
analyzed, across these contexts we see regular patterns of change in /q/ over
time.
Taking the case investigated by Cotter (2016) as an example, we can see how
patterns of change in /q/ may progress over time. In the speech of Jaffa Palestinian refugees in the Gaza Strip, Cotter (2016) showed that across three generations Jaffa refugees in Gaza showed progressively lower use of their traditional
[ʔ] realization of /q/, instead beginning to favor the voiced velar [ɡ] variant that
is common in Gaza City Arabic. Within the oldest generation of this community,
Jaffa refugees showed near categorical retention of the glottal variant, and little
rudimentary leveling. However, the second generation of Jaffa refugees showed
substantial variability between [ʔ] and [ɡ], while in the third generation in the
study, speakers showed higher rates of usage of the [ɡ] variant that is native to
the Gaza City dialect.
However, as Cotter & Horesh (2015) discuss, variability in /q/ is often situated within broader identity projects that speakers and communities have under2
Both Abu-Haidar and Blanc’s analyses are dialectological and descriptive in scope, however
both ultimately discuss what appear to be processes of change taking place for /q/ within what
Blanc termed “communal” (i.e. religio-sectarian) varieties of Arabic in Baghdad.
571
William M. Cotter
taken. It is important then that analyses of Arabic dialect contact also consider
the broader ethnographic context in which this contact takes place.
Another area of interest for researchers examining dialect contact has been
the interdental fricatives /θ, ð, ð̣/. Across Arabic varieties, these phonemes quite
often vary between realizations as true interdental fricatives [θ, ð, ð̣] and their
stop counterparts [t, d, ḍ] (Al-Wer 1997; 2003; 2011). In addition to descriptive
work that has documented the realization of the interdentals across Arabic varieties, they have also been examined as sociolinguistic variables in cases of dialect
contact.
For example, Holes (1987; 1995) investigated sociolinguistic variation in the
realization of /θ, ð, ð̣/ roughly split along sectarian lines in the speech of Arab
and Baḥārna speakers in Bahrain. In Bahrain, in the dialect of Sunni Arabs these
phonemes are traditionally pronounced as [θ, ð, ð̣], whereas in the dialect of
Shi’i speakers they are pronounced as [f, d, ḍ]. Holes (1995: 275) details that in
the speech of young literate speakers in Manama, intercommunal dialect realizations of the interdentals have emerged that are generally centered on the Sunni
Arab realizations of these phonemes. More recently, Al-Essa (2008) examined the
interdentals in the speech of Najdi Arabic speakers living in Jeddah, Saudi Arabia, an Urban Hijazi Arabic dialect area. Although Najdi Arabic typically retains
the interdental realization of these phonemes, Al-Essa concluded that degree of
contact with Urban Hijazi speakers was a significant factor influencing whether
Najdi speakers adopted the stop realizations common in Urban Hijazi Arabic.
Additionally, Alghamdi (2014) investigated the interdentals through the lens
of migration and contact in the Saudi Arabian city of Mecca. Alghamdi describes
what may be the beginning of a change from the traditional interdental realization of these phonemes in the direction of their stop counterparts. As Alghamdi
(2014: 112) notes, if it is the case that an incipient change in the interdentals exists
in Mecca, the results of her study suggest that female speakers may be leading
this change. This finding supports earlier sociolinguistic work, which has highlighted that female speakers are often at the vanguard of linguistic change.
Change in the interdentals has also been examined as part of new dialect
formation in the Jordanian capital of Amman. As Al-Wer (2007) describes (see
also Al-Wer, this volume), Amman Arabic has grown out of contact between
speakers of two different dialect types: urban Palestinian and traditional Jordanian varieties, which differ in their realizations of the interdentals. Urban Palestinian dialects typically favor non-interdental realizations [t, d, ḍ], while, in contrast, traditional Jordanian dialects retain the interdentals [θ, ð, ð̣]. Al-Wer describes the case of the interdentals in Amman as a process of focusing (Trudgill
2004) that has arisen out of contact. In Trudgill’s terms (drawing on Le Page
572
26 Dialect contact and phonological change
& Tabouret-Keller 1985), focusing is one part of the process of new-dialect formation, whereby features of input dialects are leveled and stability emerges, resulting in new shared linguistic norms. Al-Wer describes that, in Amman, focusing of the interdentals in the direction of their stop counterparts [t, d, ḍ] has taken
place (Al-Wer 2007: 66). In addition, Al-Wer notes that, as a result of contact,
Amman Arabic has also focused towards the common Palestinian [ž] realization
of /ǧ/ at the expense of the traditional Jordanian [ǧ] (Al-Wer 2007: 66).
In addition to the work by Al-Essa (2008) and Alghamdi (2014) discussed above,
a more recent case of contact-induced change in Saudi Arabia has been identified: the voiced lateral fricative [ɮˤ] realization of 〈〉ض. Al-Wer & Al-Qahtani
(2016) investigate /ɮˤ/ as a variable in the dialect of Tihāmat Qaḥtān. What this
work shows is that in the Tihāmat Qaḥtān variety, the lateral [ɮˤ] represents a
conservative, traditional variant of the phoneme, whereas the voiced interdental
fricative [ð̣] represents the innovative variant. As a result of dialect contact, AlWer & Al-Qahtani (2016) describe an intergenerational process of change towards
the voiced emphatic interdental [ð̣], with use of the historic [ɮˤ] variant receding
over time.
Another area of interest in dialect contact research has been affrication. As
descriptive work has shown, affrication of certain phonemes, notably /k/ in the
direction of [č], is common in Arabic dialects. As an example of this process,
Shahin (2011) notes that in rural varieties of Palestinian Arabic, /k/ palatalizes to
become an affricate [č] (e.g. čīfak ‘how are you (sg.m)?’ < kīfak). While typologically this affrication is common, processes of affrication or de-affrication have
also been noted as the outcome of contact.
Al-Essa (2008) investigated affrication of /k/ and /g/ in the speech of Najdi Arabic migrants in Jeddah, and found that the affrication that is a common feature
of this variety had been almost completely undone in this migrant community.
Examining this change in light of dialect contact, Al-Essa concludes that this deaffrication represents the leveling out of marked regional dialect forms as a result
of contact (Trudgill 1986; Kerswill & Williams 2000). More recently, Al-Wer et al.
(2015) note that the conditional, root-based distribution of the affricate [č] for /k/
in the Sult variety of Arabic in Jordan, which, although it has receded (Al-Wer
1991), now interacts with other innovative features in Sult that show potential
stratification along religious lines.
Elsewhere in Jordan, notably in Amman, Al-Wer (2007) describes the leveling of the affricate [č] across generations. The city dialect that has emerged in
Amman, which has Sulti Arabic as one of its input varieties, underwent rudimentary leveling (Trudgill 2004) within the first generation. This leveling resulted in
the loss of this affricate variant of /k/ in the speech of Sulti migrants. In this case,
573
William M. Cotter
Al-Wer describes the deaffrication of [č] as stemming from its status as a marked
feature of Horani Arabic varieties like that of Sult. This marked status makes it
a primary candidate for the kinds of leveling that sociolinguists have identified
in other cases of contact.
2.2 Vocalic changes
In general, the Arabic vocalic system remains understudied within research on
Arabic varieties. However, multiple cases of change linked to dialect contact have
been identified. One of the most well studied cases of contact-induced vocalic
change in Arabic is perhaps better thought of as a morphophonological change:
the Arabic feminine gender marker. The feminine gender marker is a word final
vocalic morpheme that is realized variably across Arabic varieties. The realization of this vowel varies from an unraised [a] to [æ, ɛ, e], or even as high as [i]
(e.g. Al-Wer 2007; Naïm 2011; Shahin 2011; Woidich 2011).
Even within one region, the full range of variation in this morpheme can be
seen. Taking the Levant as an example, the Lebanese capital, Beirut, is known for
raising this vowel to [e] or even [i] (Naïm 2011). The Syrian capital, Damascus, is
known to raise to [e] (Lentin 2011). Urban Palestinian (Rosenhouse 2011; Shahin
2011) is also often described as raising to [e], while Amman (Al-Wer 2007) raises
this vowel to [ɛ]. These city varieties can be contrasted with, for instance, the
variety of Cairo (Woidich 2011), which does not raise this vowel, leaving it as [a].
This morpheme is particularly interesting within a discussion of dialect contact because raising of this vowel is phonologically conditioned. The phonological factors that constrain raising vary across dialects, with urban Levantine Arabic (e.g. Syria, Palestine, Lebanon) providing one example of these factors. In urban Levantine, the following rules constrain the raising of this vowel (Grotzfeld
1980: 181; Levin 1994: 44–45; Al-Wer 2007: 68):
1. The default realization of the vowel is raised: [e]
2. The vowel is unraised, realized as [a], when:
a) it occurs after back consonants (i.e. pharyngeal, glottal, post-velar,
emphatic/pharyngealized): /ḥ, ʕ, ʔ, h, ṣ, ḍ/ð̣, ḫ, ɣ, q/;
b) it occurs after /r/, but only when preceding /r/ there is no high front
vowel. In cases where a high front vowel does precede /r/, raising is
allowed, e.g. [kbi:re] ‘big (f)’.
Below I provide two specific documented examples of contact and change in
the feminine gender marker. First, Cotter & Horesh (2015) investigated change in
574
26 Dialect contact and phonological change
the feminine gender marker in the speech of refugees originally from the Palestinian city of Jaffa who now live as refugees in the Gaza Strip. This sample included both speakers who were expelled from Jaffa after the creation of the state
of Israel in 1948 and their descendants. Their traditional urban Palestinian dialect
(Horesh 2000; Shahin 2011) is one that raises the feminine gender marker to [e],
subject to the phonological conditioning mentioned above. In contrast, based on
the available dialectological information, the dialect of Gaza City does not raise
this vowel (Bergsträßer 1915).
Cotter & Horesh (2015) highlight a process of contact-induced change that
has taken place in this community. Across generations, the realization of this
vowel appears to be lowering and backing, moving from [e] in the direction of
[a]. The result is that younger Jaffa refugee speakers realize the vowel closer to
the [a] common in Gaza City. This type of change is perhaps unsurprising in a
city like Gaza, given that the population of Gaza is overwhelmingly comprised
of refugees, including large communities who are of [a] dialect types for this
feature. This diversity and the high numbers of refugees in Gaza means that
the city, and the territory generally, is a site where many dialects of Palestinian
Arabic are in intimate contact. What remains to be determined is whether or not
new linguistic norms are emerging in the dialect of Gaza City more generally as
a result of this contact.
One other case, which is discussed in more detail by Al-Wer (this volume),
provides a succinct example of the intersection between phonetics, phonology,
and Arabic dialect contact. In discussing the formation of Amman Arabic, AlWer (2007) notes the centrality of vocalic change to the formation of the dialect.
The feminine gender marker represents one feature that has helped to define the
variety of Amman.
As Al-Wer (2007) describes, through contact between Palestinian and Jordanian Arabic dialects in Amman, the realization of the feminine gender marker
has focused on [ɛ], the indigenous Jordanian realization (as in e.g. the Horani
dialect of Sult, see Herin 2014). However, although Amman Arabic has focused
on the Jordanian phonetic realization of this vowel, it has retained urban Palestinian phonology, which blocks raising in the environment of back consonants
as defined above (Al-Wer 2007: 69). This is less restrictive than in Horani Arabic,
where raising is also blocked in the environment of velar consonants such as /k/
and the labiovelar /w/ (Al-Wer et al. 2015: 77).
Finally, I mention one other case of vocalic change that has been documented
as an outcome of dialect contact: the diphthongs [ay] and [aw]. Alghamdi (2014)
investigated monopthongization of the traditional Arabic diphthongs /ay/ and
575
William M. Cotter
/aw/ in the speech of Ghamdi migrants in Mecca. Alghamdi found that the diphthongs common in the dialect spoken by this migrant community were monophthongizing, reflecting a change towards the norms of Mecca Arabic, which
lacks diphthongs. Alghamdi’s analysis of the diphthongs provides an example of
dialect leveling borne out of contact, noting two additional aspects of this variable in the speech of this migrant community: i) Alghamdi describes the high
degree of social salience that the diphthongs have in this community and their
possible stigmatization in Mecca, and ii) that retention of the diphthongs is uncommon in Saudi Arabic, making the Ghamdi realization a minority realization
in Saudi Arabia generally. These two facts create an environment conducive to
change.
3 Conclusion
In examining Arabic dialect contact, a growing body of research highlights that
the phonology and phonetics of Arabic represent rich sites for linguistic change.
As the examples that I have provided throughout this chapter, and those discussed elsewhere throughout this volume suggest, we can identify a number of
cases where dialect contact has influenced the directionality and extent of change
in Arabic dialects. With the findings of this selection of work in mind, a number
of areas remain open for future investigation.
Perhaps the most pressing of these is the reality that, although I have highlighted work here that investigates vocalic change, the vocalic system of Arabic
varieties remains drastically understudied. Although phonetic research on the vocalic system of Arabic varieties continues to grow (see e.g. Hassan & Heselwood
2011; Khattab & Al-Tamimi 2014; Al-Tamimi & Khattab 2015), we know little
about sociophonetic changes that may take place in cases of contact like those discussed in this chapter. Given the scope of dialect contact in the Arabic-speaking
world, much of which has come as a result of mass migration throughout the
region, investigating the potential for processes such as vocalic chain-shifting
(Al-Wer 2007) represents an important next step for research on language variation and change in Arabic. I would argue that more robust investigation of vocalic change in Arabic dialects represents a pressing area of concern for Arabic
sociolinguistics.
In addition, examples like the feminine gender marker in Amman (Al-Wer
2007) open the door for future work that investigates the potential for blending
of the phonetics and phonology of different Arabic varieties as a result of contact.
Although Amman is a somewhat different case, given that it represents an example of new dialect formation, a close examination of phonetics and phonology
576
26 Dialect contact and phonological change
together in contact situations will provide us with an opportunity to examine
how dialect focusing and leveling takes place, and how the linguistic systems of
multiple different Arabic varieties interact and regularize through contact.
Additional research that looks more closely at change in the vocalic system of
Arabic dialects will go a long way towards enriching the depth of Arabic sociolinguistic research. This is especially true of work that examines cases of dialect
contact. However, beyond sociolinguistics, a closer examination of the vocalic
system will contribute to the description and documentation of Arabic dialects,
which will further enrich linguistic research that investigates the varieties of
Arabic spoken around the world.
Further reading
) As I have discussed throughout this chapter, Al-Wer’s (2007) work details a
number of aspects of how the dialect of Jordan’s capital, Amman, emerged
as a result of contact between Jordanian and Palestinian dialects. It offers a
clear picture of how contact has played out in Amman with respect to a number of linguistic features, and how the city’s dialect emerged over successive
generations.
) Cotter & Horesh’s (2015) article examines contact in the Gaza Strip, one of the
more understudied areas within Arabic linguistic research. The article draws
on sociolinguistic fieldwork conducted in Gaza City, as well Jaffa and the West
Bank, to analyze change in three specific features of Palestinian Arabic.
) Holes’ (1987) work on Bahrain provided one of the early accounts of sociolinguistic variation and change in Arabic. In addition, this work provides a
clear example of variation in Arabic that has been stratified on sectarian, or
religious, lines.
Acknowledgements
I would like to thank Christopher Lucas and the anonymous reviewer for their
comments on this chapter. In addition, I would like to thank Uri Horesh and Enam
Al-Wer for their feedback, as well as the attendees of the Arabic and ContactInduced Change workshop at the 23rd International Conference on Historical
Linguistics for their feedback on my own research as it appears in this chapter.
577
William M. Cotter
References
Abd El-Jawad, Hassan. 1981. Lexical and phonological variation in spoken Arabic
in Amman. Philadelphia: University of Pennsylvania. (Doctoral dissertation).
Abd El-Jawad, Hassan. 1987. Cross-dialectal variation in Arabic: Competing
prestigious forms. Language in Society 10. 359–67.
Abu-Haidar, Farida. 1991. Christian Arabic of Baghdad. Wiesbaden: Harrassowitz.
Al-Essa, Aziza. 2008. Najdi speakers in Hijaz: A sociolinguistic investigation
of dialect contact in Jeddah. Colchester: University of Essex. (Doctoral
dissertation).
Al-Shareef, Jamal. 2002. Language change and variation in Palestine: A case study
of Jabalia refugee camp. Leeds: University of Leeds. (Doctoral dissertation).
Al-Tamimi, Jalal & Ghada Khattab. 2015. Acoustic cue weighting in the singleton vs geminate contrast in Lebanese Arabic: The case of fricative consonants.
Journal of the Acoustical Society of America 138(1). 344–60.
Al-Wer, Enam. 1991. Phonological variation in the speech of women from three
urban areas in Jordan. Colchester: University of Essex. (Doctoral dissertation).
Al-Wer, Enam. 1997. Arabic between reality and ideology. International Journal
of Applied Linguistics 7(2). 251–65.
Al-Wer, Enam. 2003. Variability reproduced: A variationist view of the [ḏ̣]/[ḍ]
oppostion in modern Arabic dialects. In Kees Versteegh, Martine Haak &
Rudolf De Jong (eds.), Approaches to Arabic dialects, 21–31. Leiden: Brill.
Al-Wer, Enam. 2007. The formation of the dialect of Amman: From chaos to order.
In Catherine Miller, Enam Al-Wer, Dominique Caubet & Janet C. E. Watson
(eds.), Arabic in the city: Issues in dialect contact and language variation, 55–76.
London: Routledge.
Al-Wer, Enam. 2011. Phonological merger. In Lutz Edzard & Rudolf de Jong (eds.),
Encyclopedia of Arabic language and linguistics, online edn. Leiden: Brill.
Al-Wer, Enam & Khairiah Al-Qahtani. 2016. Lateral fricative ḍād in Tihāmat
Qaḥtān: A quantitative sociolinguistic investigation. In Stuart Davis & Usama
Soltan (eds.), Perspectives on Arabic linguistics XXVII: Papers from the Annual
Symposium on Arabic Linguistics, Bloomington, Indiana, 2013, 151–69. Amsterdam: John Benjamins.
Al-Wer, Enam & Bruno Herin. 2011. The lifecycle of Qaf in Jordan. Langage et
société 138. 59–76.
Al-Wer, Enam, Uri Horesh, Bruno Herin & Maria Fanis. 2015. How Arabic
regional features become sectarian features: Jordan as a case study. Zeitschrift
für Arabische Linguistik 62. 68–87.
578
26 Dialect contact and phonological change
Alghamdi, Najla Manie. 2014. A sociolinguistic study of dialect contact in Arabia: Ghamdi immigrants in Mecca. Colchester: University of Essex. (Doctoral
dissertation).
Bergsträßer, Gotthelf. 1915. Sprachatlas von Syrien und Palästina. Leipzig:
Hinrichs.
Blanc, Haim. 1964. Communal dialects in Baghdad. Cambridge, MA: Harvard
University Press.
Britain, David. 2002. Diffusion, levelling, simplification and reallocation in past
tense BE in the English Fens. Journal of Sociolinguistics 6(1). 16–43.
Britain, David & Peter Trudgill. 2009. New dialect formation and contact-induced
reallocation: Three case studies from the English Fens. International Journal of
English Studies 5(1). 183–209.
Cadora, F. J. 1992. Bedouin, village and urban Arabic: An ecolinguistic study. Leiden:
Brill.
Cotter, William M. 2016. (Q) as a sociolinguistic variable in the Arabic of Gaza
City. In Youssef A. Haddad & Eric Potsdam (eds.), Perspectives on Arabic
linguistics XXVIII: Papers from the Annual Symposium on Arabic Linguistics,
Gainesville, Florida, 2014, 229–246. Amsterdam: John Benjamins.
Cotter, William M. & Uri Horesh. 2015. Social integration and dialect divergence
in coastal Palestine. Journal of Sociolinguistics 19(4). 460–83.
De Jong, Rudolf. 2000. A grammar of the Bedouin dialects of the northern Sinai
littoral: Bridging the linguistic gap between the eastern and western Arab world.
Leiden: Brill.
Ferguson, Charles A. 1959. Diglossia. Word 15. 325–40.
Gafter, Roey & Uri Horesh. 2015. When the construction is axla, everything is
axla: A combined lexical and structural borrowing from Arabic into Hebrew.
Journal of Jewish Languages 3(1–2). 337–48.
Grotzfeld, Heinz. 1980. Das syrisch-palästinesische Arabisch. In Wolfdietrich
Fischer & Otto Jastrow (eds.), Handbuch der arabischen Dialekte, 174–90.
Wiesbaden: Harrassowitz.
Hachimi, Atiqa. 2007. Becoming Casablancan: Fessis in Casablanca as a case
study. In Catherine Miller, Enam Al-Wer, Dominique Caubet & Janet C. E.
Watson (eds.), Arabic in the city: Issues in dialect contact and language variation,
97–122. London: Routledge.
Hachimi, Atiqa. 2012. The urban and the urbane: Identities, language ideologies,
and Arabic dialects in Morocco. Language in Society 41(3). 321–41.
Haeri, Niloofar. 1997. The sociolinguistic market of Cairo: Gender, class and
education. London: Routledge.
579
William M. Cotter
Haeri, Niloofar. 2000. Form and ideology: Arabic sociolinguistics and beyond.
Annual Review of Anthropology 29(1). 61–87.
Hassan, Zeki Majeed & Barry Heselwood. 2011. Instrumental studies in Arabic
phonetics. Amsterdam: John Benjamins.
Herin, Bruno. 2014. Salt, dialect of. In Lutz Edzard & Rudolf de Jong (eds.),
Encyclopedia of Arabic language and linguistics, online edn. Leiden: Brill.
Holes, Clive. 1987. Language variation and change in a modernising Arab state: The
case of Bahrain. London: Kegan Paul.
Holes, Clive. 1995. Community, dialect and urbanization in the Arabic-speaking
Middle East. Bulletin of the School of Oriental and African Studies 58(2). 270–87.
Horesh, Uri. 2000. Toward a phonemic and phonetic assessment of Jaffa Arabic:
Is it a typical urban Palestinian dialect? In Manwel Mifsud (ed.), Proceedings
of the third international conference of AIDA: Association Internationale de
Dialectologie Arabe, held in Malta, 29 March – 2 April 1998, 15–20. Malta: AIDA.
Ibrahim, Muhammad H. 1986. Standard and prestige language: A problem in
Arabic sociolinguistics. Anthropological Linguistics 28(1). 115–26.
Kerswill, Paul & Ann Williams. 2000. Creating a New Town Koine: Children and
language change in Milton Keynes. Language in Society 29(1). 65–115.
Khattab, Ghada & Jalal Al-Tamimi. 2014. Geminate timing in Lebanese
Arabic: The relationship between phonetic timing and phonological structure.
Laboratory Phonology 5(2). 231–269.
Labov, William. 1963. The social motivation of a sound change. Word 19(3). 273–
309.
Le Page, R. B. & Andrée Tabouret-Keller. 1985. Acts of identity: Creole-based
approaches to language and ethnicity. Cambridge: Cambridge University Press.
Lentin, Jérôme. 2011. Damascus Arabic. In Lutz Edzard & Rudolf de Jong (eds.),
Encyclopedia of Arabic language and linguistics, online edn. Leiden: Brill.
Levin, Aryeh. 1994. A grammar of the Arabic dialect of Jerusalem. Jerusalem:
Magnes Press.
Lucas, Christopher. 2015. Contact-induced language change. In Claire Bowern &
Bethwyn Evans (eds.), The Routledge handbook of historical linguistics, 519–536.
London: Routledge.
Milroy, Lesley. 1987. Language and social networks. Oxford: Wiley.
Naïm, Samia. 2011. Beirut Arabic. In Lutz Edzard & Rudolf de Jong (eds.),
Encyclopedia of Arabic language and linguistics, online edn. Leiden: Brill.
Nycz, Jennifer. 2013. Changing words or changing rules? Second dialect
acquisition and phonological representation. Journal of Pragmatics 52. 49–62.
580
26 Dialect contact and phonological change
Nycz, Jennifer. 2016. Awareness and acquisition of new dialect features. In
Anna Babel (ed.), Awareness and Control in Sociolinguistic Research, 62–79.
Cambridge: Cambridge University Press.
Rosenhouse, Judith. 2011. Jerusalem Arabic. In Lutz Edzard & Rudolf de Jong
(eds.), Encyclopedia of Arabic language and linguistics, online edn. Leiden: Brill.
Shahin, Kimary. 2011. Palestinian Arabic. In Lutz Edzard & Rudolf de Jong (eds.),
Encyclopedia of Arabic language and linguistics, online edn. Leiden: Brill.
Trudgill, Peter. 1986. Dialects in contact. Oxford: Wiley-Blackwell.
Trudgill, Peter. 2004. New-dialect formation: The inevitability of colonial Englishes.
Oxford: Oxford University Press.
Van Coetsem, Frans. 1988. Loan phonology and the two transfer types in language
contact. Dordrecht: Foris.
Van Coetsem, Frans. 2000. A general and unified theory of the transmission process
in language contact. Heidelberg: Winter.
Woidich, Manfred. 2011. Cairo Arabic. In Lutz Edzard & Rudolf de Jong (eds.),
Encyclopedia of Arabic language and linguistics, online edn. Leiden: Brill.
581
Chapter 27
Contact and variation in Arabic
intonation
Sam Hellmuth
University of York
Evidence is emerging of differences among Arabic dialects in their intonation patterns, along known parameters of variation in prosodic typology. Through a series
of brief case studies, this chapter explores the hypothesis that variation in intonation in Arabic results from changes in the phonology of individual Arabic varieties,
triggered by past (or present-day) speaker bilingualism. If correct, variation in intonation should reflect prosodic properties of the specific languages that a particular
regional dialect has had contact with.
1 Introduction
1.1 Rationale
The hypothesis explored in this chapter is that observed synchronic variation
in intonation across Arabic dialects is contact-induced. In this scenario, differences between dialects would result from changes in the intonational phonology
of individual varieties triggered by speaker bilingualism in Arabic and one or
more other languages (Lucas 2015), either in the past, or up to and including the
present day. To achieve this, I outline a framework for analysis of variation in intonation (§1.2), summarise recent research on the effects of bilingualism on the
intonational phonology of bilingual individuals and the languages they speak
(§1.3), and sketch the types of language contact scenario which may be relevant
for Arabic (§1.4). In §2 I present case studies of prosodic features which appear to
be specific to a particular dialect, on current evidence at least, and discuss which
of the potentially relevant contact languages might have served as the potential
Sam Hellmuth. 2020. Contact and variation in Arabic intonation. In Christopher
Lucas & Stefano Manfredi (eds.), Arabic and contact-induced change, 583–601. Berlin:
Language Science Press. DOI:10.5281/zenodo.3744553
Sam Hellmuth
source of the feature in question, considering also possible endogenous (internal) sources of the change. The chapter closes (§3) with suggestions for future
research.
1.2 Cross-linguistic variation in intonation
Any attempt to delimit the nature and scope of variation in intonation depends on
the model of intonational phonology adopted. The analyses explored in §2 below
are framed in the Autosegmental-Metrical (AM) theory of intonation (Ladd 2008),
and the parameters of intonational variation explored are thus influenced by this
choice.
A basic debate in the analysis of intonation is whether the primitives of the
system are whole contours (defined over an intonational phrase), or some subcomponent of those contours (Ladd 2008). In AM theory, intonation is modelled
as interpolation of pitch between tonal targets; these tonal targets are the primitives of the system and are of two types: pitch accents are associated with the
heads of metrical domains (e.g. stressed syllables), boundary tones are associated
with the edges of metrical domains (e.g. prosodic phrases). In AM, tonal targets
are transcribed using combinations of high (H) or low (L) targets, which reflect
significant peaks and valleys, respectively, in the pitch contour of the utterance;
association of these events to landmarks in the metrical structure is marked using “*” for pitch accents (associated with stressed syllables) and “%” for boundary tones (associated with the right edge of prosodic phrases of different sizes).
A typical AM analysis yields an inventory of the pitch accents and boundary
tones needed to model the contours in a corpus of speech data, supported by a
description of the observed contours (Jun & Fletcher 2015).
Ladd’s (2008) taxonomy of possible parameters of cross-linguistic variation in
intonation (based on Wells 1982) envisages four broad (inter-related) categories
of variation: systemic (differences in the inventory of pitch accents or boundary
tones); semantic (differences in the meaning or function associated with a particular contour, pitch accent or boundary tone); realisational (differences in the
phonetic realisation of otherwise parallel pitch accents or boundary tones; and
phonotactic (differences in the distribution of pitch accents and boundary tones,
or in their association to metrical structure).
Comparison of AM analyses across a typologically distinct set of languages
(Jun 2005; 2015) has highlighted systematic cross-linguistic variation of a systemic and/or phonotactic nature, in terms of prosodic phrasing (with relatively
smaller or larger domains involved in structural organization of intonation patterns), the distribution of tonal events relative to prosodic constituents (marking
584
27 Contact and variation in Arabic intonation
either the edges or the metrical heads of phrases or both), and the size and composition of the inventory of tonal events regularly observed (pitch accents and
boundary tones). There is also a large body of research on cross-linguistic variation in the phonetic realisation of pitch accents, in particular on peak alignment
(Atterer & Ladd 2004; Ladd 2006) and scaling (Ladd & Morton 1997), confirming
the existence of realisational cross-linguistic variation. The most advanced work
on semantic variation to date has been on Romance languages, facilitated by a
concerted effort to develop descriptions of these languages’ intonation patterns
within a common annotation system (Frota & Prieto 2015).
For Arabic, evidence is emerging of variation along similar lines. Recent review
articles have highlighted clear differences in the size and composition of the inventory of pitch accents and boundary tones across Arabic dialects (Chahal 2011;
El Zarka 2017), and in the association of pragmatic meanings with contours (cf.
the case study in §2.1). Initial evidence suggests a difference between Jordanian
and Egyptian Arabic in the mapping of prosodic phrases to syntax (Hellmuth
2016) similar to that reported across Romance languages (D’Imperio et al. 2005).
Recent research suggests that Moroccan Arabic is a non-head-marking language
in contrast to other Arabic dialects which are head-marking (see §2.2), mirroring
the cross-linguistic variation captured in Jun’s (2005) typology, and among the
head marking dialects, there appears to be variation in the density of distribution
of pitch accents (Chahal & Hellmuth 2015; see §2.3).
1.3 Contact-induced variation in intonation
A growing body of research has explored contact-induced prosodic change in the
speech of bilingual communities and individuals. The initial focus of most studies
was on second-language (L2) learners’ intonation patterns, or studies of individual bilinguals (Queen 2012), and early L2 studies focused on realisational effects
of a speaker’s L1 on their L2, and vice versa (Atterer & Ladd 2004; Mennen 2004).
More recent studies reveal a complex array of prosodic effects, both in terms of
the features involved in the change (taking in all four of Ladd’s categories of
possible variation), and also in the directionality of effects (L1 on L2, L2 on L1, or
hybrid effects).
Bullock (2009) characterizes the general contact-induced language change literature (e.g. Weinreich et al. 1968; Thomason & Kaufman 1988) as having made
the assumption that segmental effects would “precede” prosodic effects, thus predicting prosodic effects would be seen only in contexts of widespread or sustained community bilingualism. As Bullock notes, however, there is no logical
585
Sam Hellmuth
structural reason why this should be the case; her own study of English-like prosodic patterns in heritage French speakers in Pennsylvania confirms an effect of
the dominant language in the prosodic domain (specifically in the realisation of
focus) in speakers who in other respects maintain French segmental patterns.
Another example of prosodic properties of a dominant language affecting prosodic realisation of a heritage or second language, is that of immersion Gaelic
learners in Scotland (Nance 2015). Nance demonstrates a structural change in
progress in Gaelic, from lexical pitch accent – still used by older English–Gaelic
bilinguals – to a purely post-lexical system, used by younger bilinguals in immersion education who produce Gaelic with English-like intonation. Similar effects
of the dominant language on the non-dominant language are reported for Spanish in contact with Quechua (O’Rourke 2004).
The reverse effect has also been found in a number of studies, however, where
prosodic properties of a non-dominant or heritage language have an effect on
the prosodic realisation of the dominant language, in the speech of an individual
or of the whole community. Fagyal (2005) studied a group of bilingual French–
Arabic adolescents in Paris; instead of a typical French phrase-final rise, these
speakers produce a phrase-final rise–fall contour in declaratives, similar to the
contour observed in Moroccan Arabic (MA) in parallel contexts. Simonet (2011)
shows that the steep (“concave”) final fall in Majorcan Catalan declaratives is
now widely observed in Majorcan Spanish, replacing the typical gradual (“convex”) fall in Majorcan Spanish, but that each individual bilingual’s usage closely
mirrors their reported language dominance. Colantoni & Gurlekian (2004) observe patterns of peak alignment in pre-nuclear accents and pre-focal downstep
in Buenos Aires Spanish which differ from neighbouring varieties of Spanish but
resemble those in Italian, and ascribe them to high levels of Spanish–Italian bilingualism in the city in the late nineteenth and early twentieth centuries. In this
last case, the period of community bilingualism which triggered the change is
now long past, but the effect on the prosodic patterns of the dominant language
(in this case, Spanish) persists.
Finally, Queen (2001; 2012) reports a case of “fusion”: Turkish–German bilinguals in Germany display phrase-final intonation contours which are never used
by monolinguals in either language, but are found only in the speech of bilinguals, in a new variety of German known as “Türkisch-Deutsch”.
This emerging literature suggests that contact-induced prosodic change is a
frequent phenomenon, arising in varied forms and across diverse contact situations. Bullock (2009) suggests that prosody and intonation are especially prone to
change for three reasons. First, because the acoustic parameters involved – pitch,
intensity and duration – are part of the linguistic encoding of all languages, albeit
586
27 Contact and variation in Arabic intonation
in different constellations, and are thus readily adapted. Second, because, perceptually, all languages make use of prosodic parameters to convey some aspects of
utterance-level meaning, thus the mapping of form to meaning is also readily
adapted. Third, and perhaps most persuasively, because the form–meaning mapping in intonation is generally not fixed, but displays considerable inter- and
intra-speaker variation (Cangemi et al. 2015; 2016) as well as contextual variation (cf. Walker 2014), and it is pockets of structural ‘indeterminacy’ of this kind
which are prone to change in bilingual grammars (Sorace 2004). Queen (2001: 57)
also suggests that the intertwining of form and function in intonation makes it a
fruitful sphere for investigation of contact-induced change, because “intonation
is one of the few linguistic elements that comments simultaneously on grammar,
context and culture”. Indeed, Simonet’s (2011) work shows that speakers are able
to adopt the intonation of a contact language without actually being proficient
in the source language. Finally, Matras (2007) argues that the separate nature of
prosody, which is processed separately from segmental phonology, and can be
interpreted independently of the propositional content of the utterance, renders
prosody more “borrowable” than other aspects of the grammar.
In sum, there is strong evidence that intonation patterns are highly porous,
being transferred between dominant and non-dominant languages in either direction; intonation is thus a fruitful area for investigation of contact-induced
language change. Among the literature reviewed here, the paper by Colantoni
& Gurlekian (2004) most closely resembles the type of work which is needed in
future for Arabic; they investigate present-day intonational variation in closely
related varieties, and provide evidence from historical migration patterns to support the claim that the present-day variation can be ascribed to an earlier period
of widespread bilingualism, in a language which is a plausible source of the feature in question. In the next section we outline a similar line of investigation for
Arabic.
1.4 Contact-induced variation in Arabic
The time depth of descriptions of intonation patterns is shallow, due to the lack
of historical audio recordings, and a general tendency that traditional grammars
do not include detailed descriptions of prosody. It is thus difficult to reliably
determine when changes in intonation may have happened, and the range of
languages to be considered as the source of any putative intonational change is
rather broad.
One set of potential source languages is the substrate languages spoken in
a particular region before the arrival of Arabic, for example Amazigh (Berber)
587
Sam Hellmuth
in North Africa, and Egyptian–Coptic in Egypt. We might also see influence
from the external languages which these indigenous languages were in contact
with prior to the arrival of Arabic, such as Greek and Latin, or of other external
languages whose influence was felt throughout the Arab world in later periods,
such as Persian and Ottoman. Other possible source languages are European languages spoken along the northern coast of the Mediterranean, since large areas
of southern Europe were under Arab rule for extended periods (7th–15th centuries CE), and contact through sea-borne trade is likely to have continued after
that time. Conversely, large areas of the Arab world were under direct or indirect European control also (19th–20th centuries CE), and the influence of these
languages is still felt today. Finally, we might also consider the potential effects
of contact with global languages such as English, and with the L1 languages of
migrant workers and long-term displaced language communities.
The decision to treat observed present-day variation as the result of change
does not entail assuming that any one variety of Arabic was the ancestor of all
dialects. Instead, the approach here will be to identify prosodic features which
are seen in one Arabic dialect (or group of dialects), but not (yet) documented in
any other dialects, as the most likely cases of potential contact-induced change.
In each case study we evaluate the hypothesis by looking for evidence of the
same feature in the relevant contact languages for the dialect in question, with
comparison to possible language-internal sources of the change.
2 Contact-induced variation in Arabic intonation
2.1 Tunisian Arabic question marking
Tunisian Arabic (TA) polar questions are typically associated with a salient rise–
fall pitch contour at the end of the utterance: speakers from southeast Tunisia
produce a complete rise–fall in which pitch rises over the stressed syllable of
the last word in the utterance to a peak, then falls to low; in contrast, speakers
from Tunis produce a rise–plateau contour, in which, after the peak, pitch falls
slightly then levels out. These patterns are illustrated in Figure 1 (Bouchhioua
et al. 2019). The rise–fall prosodic contour is frequently accompanied by a segmental question marker, in the form of a vowel added to the end of the last word
in the utterance. The quality of the epenthesised vowel is influenced partly by
vowels earlier in the word (in a form of vowel harmony) and partly by regional
dialect, though these vowel quality patterns require further investigation.
The rise–fall yes/no-question contour in TA differs from the rise seen in yes/no
questions in most Arabic dialects (Hellmuth in preparation) and, in terms of dis-
588
27 Contact and variation in Arabic intonation
si naˈbil /mawˈʒud/ [mawˈʒudə/u]
Mr Nabil present
‘Is Mr Nabil there?’ (tuno-arc1-f1/tuse-arc1-f1)
Figure 1: Pitch trace of yes/no questions from a Tunisian Arabic speaker
from Tunis (left) and from the southeast (right).
tribution, from the rise–fall contour observed in Moroccan Arabic (MA) across
all utterance types (not only in yes/no questions). The vowel epenthesis marker
appears to be unique to TA, thus far.
A pattern of utterance-final vowel epenthesis has been observed in a number
of Romance languages spoken along the northern edge of the Mediterranean, including Bari Italian (Grice et al. 2015), and different varieties of Portuguese (Frota
et al. 2015). These cases of utterance-final vowel epenthesis are interpreted as
text–tune adjustment, where segmental material is added to accommodate a complex prosodic contour. For example, in Standard European Portuguese, a more
general rule of utterance-final vowel deletion is blocked in utterances bearing a
complex prosodic contour, such as the fall–rise (H+L* LH%) on yes/no questions
(Frota et al. 2015). In Bari Italian, epenthesis is seen on a range of utterance types,
but – like Portuguese – its occurrence can be ascribed to tonal crowding (i.e. that
the complex contour requires more segmental material to be realised). This is reflected in higher incidence of epenthesis on utterance-final monosyllables than
on longer words, and on words in which the final sound is an obstruent than on
words with a final sonorant (Grice et al. 2015).
Investigation of utterance-final vowel epenthesis in TA yes/no questions, in
a corpus of data collected in Tunis, shows a very different pattern, however. In
TA the incidence of epenthesis is not affected by the number of syllables in the
utterance-final word nor by the type of final sound. In addition, whereas in the
other Romance languages epenthesis occurs on a range of utterance types, in
TA epenthesis occurs only in yes/no questions, and predominantly in yes/no
questions which are produced with a complex rise–plateau or rise–fall contour
(Hellmuth forthcoming). The effects which in the Romance languages are taken
589
Sam Hellmuth
as evidence of text–tune adjustment are lacking in TA, which appears to rule out
a language-internal (endogenous) source of the TA pattern of vowel epenthesis.
The TA epenthetic vowel is in fact best characterized as an optional question
marker comprising the vowel itself plus an accompanying fall in pitch. The segmental marker is well-known among Tunisian linguists, being described as “the
pan-Tunisian question marker clitic [ā]” (Herin & Zammit 2017: 141), but the accompanying prosodic contour has received little attention in the literature until recently. This traditional question-marking strategy may however be in decline, since it now alternates with realisation of a yes/no question using a simple
rise contour similar to that found in most other Arabic dialects, and without
an utterance-final epenthetic vowel. The picture is complicated by the fact that
there is a somewhat higher incidence of epenthesis among young female speakers, who might be expected to use the traditional form less, rather than more
(Hellmuth forthcoming).
The epenthesis + complex contour strategy in TA yes/no questions stands out
from other Arabic dialects and may thus be due to contact-induced prosodic
change. Italian was spoken in Tunisia more widely than French, in the late nineteenth century (Sayahi 2011), and is thus a potential source of the contour, since
rise–falls occur in yes/no questions in a number of Italian dialects (Gili Fivela
et al. 2015). However, the conditioning environments of epenthesis reported for
Bari Italian are very different, suggesting that contact with Italian is not a likely
source of the epenthesis component of the TA pattern.
An alternative source of the vowel epenthesis pattern is French, since Tunisia
has seen very high levels of bilingualism in TA and French from the late nineteenth century up to the present day, despite concerted efforts to reduce usage
of French (Daoud 2007). Utterance-final schwa epenthesis has been reported as
an emerging phenomenon in French (Hansen 1997), but its distribution is again
much broader, being seen across a range of utterance types, and not restricted
to yes/no questions. Despite clear evidence of contact-induced effects of French
on TA in other domains, such as lexical borrowing, and a general trend towards
use of French by female speakers (Walters 2011), the different distribution of final
epenthesis in French suggests it is not the most likely source of the TA segmental
question-marking strategy.
The other major contact language with TA is Tunisian Berber (TB). Although
levels of TA–Berber bilingualism in Tunisia are now low, other than in certain
regions (Gabsi 2011), there was a sustained period of TA–TB bilingualism from
the eleventh century, and TB is an important substrate of TA (Daoud 2007). Although there are no studies of the prosody of TB, to our knowledge, a recent
detailed study of Zwara Berber, spoken close to the Tunisian border in western
590
27 Contact and variation in Arabic intonation
Libya, documents a polar question marking clitic /a/ which is obligatorily accompanied by a rise–fall contour (Gussenhoven 2017). The match of this description
to the TA pattern is so close that it seems plausible that the TA question-marking
pattern arose due to contact with TB during the period of sustained TA–TB bilingualism. The greater use of the epenthesis + contour strategy by female speakers
than male speakers, as well as regional variation, makes this feature of TA ripe
for further detailed sociolinguistic study.
2.2 Moroccan Arabic word prosody
Variation in word stress patterns across Arabic dialects has inspired much phonological investigation (Watson 2011), but the Moroccan Arabic (MA) stress system
has defied analysis until recently. Mitchell (1993: 202) notes that “in contrast with
all the other vernaculars […], the place of prominence in a word in isolation is not
carried over to its occurrence in the phrase and sentence”, and this characterization was confirmed experimentally by Boudlal (2001). A range of positions have
emerged, with some authors claiming that MA does have word stress (Benkirane
1998; Burdin et al. 2014), and others that it does not (El Zarka 2012; Maas 2013).
It is now clear that MA is indeed typologically different from most other Arabic dialects in its word prosody. Whereas the majority of Arabic dialects have
salient word-level stress and are thus clearly “head-marking” languages, in the
typology proposed by Jun (2005), MA is a non-head-marking language in which
tonal events mark the edges of prosodic phrases only. Bruggeman (2018) provides
acoustic evidence that there are no consistent cues to lexical prominence in MA,
and perceptual evidence that MA listeners display the same type of “deafness”
to stress as has been reported for listeners in languages which also lack head
marking such as French (Dupoux et al. 2001) and Persian (Rahmani et al. 2015).
Can this stark variation in prosodic type between MA and other dialects of
Arabic be attributed to contact-induced change? The Arabic language has been
in sustained contact with Amazigh (Moroccan Berber, MB) since the seventh
century, but also with Latin, French and Spanish (Heath, this volume). Maas &
Procházka (2012) argue from corpus data that MA and MB share a common phonology, across a range of segmental and suprasegmental features. Bruggeman
(2018) confirms that there is no difference between MA and Tashlhiyt MB: both
lack acoustic cues to word-level prominence in production and both groups of
listeners display stress deafness.
Since French is also an edge-marking language, without lexical stress, can we
rule out French as an alternative source of this prosodic feature of MA? The
main evidence comes from the fact that MA and MB also share other prosodic
591
Sam Hellmuth
features which are not found in French, such as the shape of the tonal contour
used to mark the edges of phrases, which is a rise in French (Delais-Roussarie
et al. 2015), but a rise–fall in both Tashlhiyt MB (Grice et al. 2015; Bruggeman
et al. 2017) and MA (Benkirane 1998; Hellmuth in preparation). The contrast is
also exemplified in Faygal’s (2005) study of French–MA bilinguals in Paris who
use an MA rise–fall contour in French.
2.3 Egyptian Arabic accent distribution
Cairene Egyptian Arabic (EA) displays a rich distribution of sentence accents,
with a pitch accent typically observed on every content word. This has been
noted independently by different authors (Rifaat 1991; Rastegar-El Zarka 1997),
and is observed in both read and spontaneous speech styles (Hellmuth 2006).
Initial studies suggest that the same may be true also of some other dialects, such
as Emirati (Blodgett et al. 2007) or Ḥiǧāzi (Alzaidi 2014), but these observations
await corroboration across different speech styles.
Dense accent distribution has been noted in some languages on the northern coast of the Mediterranean also, including Spanish and Greek (Jun 2005),
although, in Spanish, the rich accent distribution seen in laboratory speech is
reduced in spontaneous speech (Face 2003). Portuguese dialects vary in accent
distribution: most varieties typically have an accent on every content word, but
Standard European Portuguese shows an accent on the first and last words in an
utterance only (Frota et al. 2015).
Rich accent distribution is not observed in Moroccan Arabic (Benkirane 1998),
nor in Tunisian Arabic (Hellmuth in preparation). If the EA accent distribution
pattern were due to contact between EA and the southern European languages
on the other side of the Mediterranean which share the tendency towards rich
accent distribution, we might expect the pattern to be found all across North
Africa.
There is strong documentary evidence from written sources of historical sustained multilingualism in Egypt. Greek arrived in Egypt in the fourth century
BCE, serving as a formal administrative language alongside Egyptian for several centuries, and with the country reaching a state of “balanced societal bilingualism” in Greek and Egyptian in the sixth and seventh centuries CE (Papaconstantinou 2010: 6). Egyptian evolved into Coptic, and its prestige continued
to increase from the sixth century CE onwards. After the Arab conquest in the
seventh century CE, Arabic began to take over from Greek as the language of administration, eventually replacing Coptic in daily use (Papaconstantinou 2012).
592
27 Contact and variation in Arabic intonation
Is it possible that Egyptian–Coptic or Greek is the source of the rich accent
distribution observed in EA (and indeed in Romance languages in southern Europe)?
The distribution of full and long vowels in Coptic indicates that it had wordlevel prominence (Peust 1999: 270), but it is not possible to determine from written texts the nature or distribution of any tonal contours which may have been
associated with prominent syllables. Anecdotal evidence suggests that the intonation patterns used in surviving liturgical forms of Coptic are very different
from those in EA (Peust 1999: 32), though this difference may owe more to the
liturgical setting than to properties of the languages in spoken form.
Ancient Greek is generally thought to have had a pitch accent system in which
the primary marker of culminative accent in each word was pitch (Devine &
Stephens 1985). The Koiné Greek dialect used in Egypt is thought to have lost
pitch accent in favour of a stress accent system, however, by the fourth century
BCE (Benaissa 2012).
Support for the hypothesis that Greek is the original source of the rich accent
distribution would come from a match between the historical spread of Koiné
Greek around the Mediterranean with the location of languages in which rich
accent distribution is also found. This would predict that eastern varieties of
Libyan (Cyrenaican) Arabic might also be found to display rich accent distribution. If rich accent distribution is confirmed in dialects of Arabic (such as Emirati
or Ḥiǧāzi) which did not have sustained contact with Greek, or with EA more
recently, this would argue against Greek as the original source. Although Nubi
(Gussenhoven 2006) and Juba Arabic (Nakao 2013) display hybrid properties between stress and lexical tone, the most likely explanation of their prosodic patterns is direct contact with local tonal languages. A potential endogenous trigger
for development of rich accent distribution would be the absence of other forms
of phonological marking of word domains, which are indeed somewhat reduced
in EA, in comparison to other dialects (Watson 2002), though the direction of
causality of this correlation is not easily determined.
Accent distribution has only recently been added to the parameters of variation explored in work on prosody (Hellmuth 2007), and thus included in descriptions of the intonation systems of languages (e.g. Frota & Prieto 2015). As
further descriptions emerge of more dialects of Arabic it will be important to
include documentation of accent distribution, across genres and speaking styles,
in future research.
593
Sam Hellmuth
3 Conclusion
There is much that we do not yet know about variation in intonation in Arabic,
which leaves scope for investigation of further potential cases of contact-induced
prosodic change. One such case may be the Syrian Arabic utterance-final rising
intonation, sometimes known as “drawl”, which is found in yes/no questions
but also across other utterance types (Cowell 1964), and which is an identifiable
feature of the Damascus dialect (Kulk et al. 2003). Although the full geographical
range of the pattern has not been investigated in detail, and may be diffused to
other dialects in the Levant, this rising declarative intonation pattern stands out
from most other Arabic dialects, and is thus another potential case of contactinduced change.
Another potential outlier pattern is the rise–fall intonation contour seen in
yes/no questions in Yemeni Arabic from Ṣanʕāʔ (Hellmuth 2014). The full areal
reach of this prosodic question-marking strategy is also not yet fully known,
and may extend into Ḥiǧāzi Arabic and western dialects of Oman. However, we
do know that a rise–fall is seen in both Tunisia and Morocco, though in these
places the pattern may be due to contact with varieties of Berber. Nevertheless,
it is tempting to speculate how a pattern found in Yemen might also be found in
Tunisia and Morocco, and thus to explore the potential role of contact-induced
variation due to ancient migrations between the eighth and fourteenth centuries
(Holes 2018).
Finally, the intonation of Modern Standard Arabic (MSA) and of other formal
registers may prove to be a fruitful domain of future research. As our knowledge
of the intonational phonology of spoken Arabic dialects improves, this will facilitate investigation of the extent to which the intonation patterns of a speaker’s
regional dialect can be observed and/or perceived in their MSA speech, building
on the findings of prior studies (El Zarka & Hellmuth 2009). An important goal
would be to determine the extent to which a separate intonational system can
be described for MSA, and to document the differential contribution to this system of specific genres of MSA discourse versus contact-induced influence due to
widespread community mastery of multiple registers of the language.
All these investigations would benefit from improved documentation of the
time depth of present-day surface intonation patterns. For the quasi-unique features explored in §2, we do not know whether these are the result of recent or
much more distant historical change. This situation might be rectified through
analysis of archive audio materials, though dialect studies have often worked
on oral narratives, which yield only a limited range of prosodic expression (i.e.
usually few questions, and no information about turn-taking). A more viable
594
27 Contact and variation in Arabic intonation
strategy to gauge the time depth of contact-induced variation in Arabic intonation would be for future sociolinguistic studies to include prosodic features as
variables in apparent-time studies with participants in different age ranges, or
for pre-existing corpora of apparent-time data to be made available for prosodic
analysis.
Further reading
There are two key reference works, so far, on intonation in Arabic dialects, based
on secondary analysis of prior published work: Chahal (2011) and El Zarka (2017).
Hellmuth (2019) suggests prosodic variables for inclusion in studies of variation
and change in Arabic.
Acknowledgements
The Intonational Variation in Arabic corpus (Hellmuth & Almbark 2017) was
funded by an award to the author by the UK Economic and Social Research Council (ES/I010106/1).
Abbreviations
AM
BCE
CE
EA
H
L
Autosegmental-Metrical theory
before Common Era
Common Era
Egyptian Arabic
high tone
low tone
L1, L2
MA
MB
MSA
TA
TB
1st, 2nd language
Moroccan Arabic
Moroccan Berber
Modern Standard Arabic
Tunisian Arabic
Tunisian Berber
References
Alzaidi, Muhammad. 2014. Information structure and intonation in Hijazi Arabic.
Colchester: University of Essex. (Doctoral dissertation).
Atterer, Michaela & D. Robert Ladd. 2004. On the phonetics and phonology of
“segmental anchoring” of F0: Evidence from German. Journal of Phonetics 32.
177–197.
Benaissa, Amin. 2012. Greek language, education, and literary culture. In
Christina Riggs (ed.), The Oxford handbook of Roman Egypt, 526–542. Oxford:
Oxford University Press.
595
Sam Hellmuth
Benkirane, Thami. 1998. Intonation in Western Arabic (Morocco). In Daniel Hirst
& Albert Di Cristo (eds.), Intonation systems: A survey of twenty languages, 345–
359. Cambridge: Cambridge University Press.
Blodgett, Allison, Jonathan Owens & Trent Rockwood. 2007. An initial account of
the intonation of Emirati Arabic. In Jürgen Trouvain & William J. Barry (eds.),
Proceedings of the 16th International Congress of Phonetic Sciences, Saarbrücken,
Germany, 1137–1140. Saarbrücken: Universität des Saarlandes.
Bouchhioua, Nadia, Sam Hellmuth & Rana Almbark. 2019. Variation in prosodic
and segmental marking of yes–no questions in Tunisian Arabic. In Catherine
Miller, Alexandrine Barontini, Marie-Aimée Germanos, Jairo Guerrero &
Christophe Pereira (eds.), Studies on Arabic dialectology and sociolinguistics:
Proceedings of the 12th conference of AIDA held in Marseille from May 30th to
June 2nd 2017, 38–48. Aix-en-Provence: Institut de recherches et d’études sur
le monde arabe et musulman.
Boudlal, Abdelaziz. 2001. Constraint interaction in the phonology and morphology
of Casablanca Moroccan Arabic. Rabat, Morocco: Université Mohammed V.
(Doctoral dissertation).
Bruggeman, Anna. 2018. Lexical and postlexical prominence in Tashlhiyt Berber
and Moroccan Arabic. Cologne: Universität zu Köln. (Doctoral dissertation).
Bruggeman, Anna, Timo B. Roettger & Martine Grice. 2017. Question word
intonation in Tashlhiyt Berber: Is “high” good enough? Laboratory Phonology:
Journal of the Association for Laboratory Phonology 8(1: 5). 1–27.
Bullock, Barbara E. 2009. Prosody in contact in French: A case study from a
heritage variety in the USA. International Journal of Bilingualism 13. 165–194.
Burdin, Rachel Steindel, Sara Phillips-Bourass, Rory Turnbull, Murat Yasavul,
Cynthia G. Clopper & Judith Tonhauser. 2014. Variation in the prosody of focus
in head- and head/edge-prominence languages. Lingua 165. 254–276.
Cangemi, Francesco, Martine Grice & Martina Krüger. 2015. Listener-specific
perception of speaker-specific production in intonation. In Susanne Fuchs,
Daniel Pape, Caterina Petrone & Pascal Perrier (eds.), Individual differences
in speech production and perception, 123–145. Frankfurt Am Main: Peter Lang.
Cangemi, Francesco, Dina El Zarka, Simon Wehrle, Stefan Baumann & Martine
Grice. 2016. Speaker-specific intonational marking of narrow focus in Egyptian
Arabic. In Jon Barnes, Alejna Brugos, Stefanie Shattuck-Hufnagel & Nanette
Veilleux (eds.), Proceedings of Speech Prosody 2016, 31 May – 3 Jun 2016, Boston,
USA, 335–339. Boston: Boston University.
Chahal, Dana. 2011. Intonation. In Lutz Edzard & Rudolf de Jong (eds.),
Encyclopedia of Arabic language and linguistics, online edn. Leiden: Brill.
596
27 Contact and variation in Arabic intonation
Chahal, Dana & Sam Hellmuth. 2015. Comparing the intonational phonology
of Lebanese and Egyptian Arabic. In Sun-Ah Jun (ed.), Prosodic typology II:
The phonology of intonation and phrasing, 365–404. Oxford: Oxford University
Press.
Colantoni, Laura & Jorge Gurlekian. 2004. Convergence and intonation:
Historical evidence from Buenos Aires Spanish. Bilingualism: Language and
Cognition 7. 107–119.
Cowell, Mark W. 1964. A reference grammar of Syrian Arabic. Washington, DC:
Georgetown University Press.
D’Imperio, Mariapaola, Gorka Elordieta, Sónia Frota, Pilar Prieto & Marina
Vigario. 2005. Intonational phrasing in Romance: The role of syntactic and
prosodic structure. In Sónia Frota, Marina Vigario & Maria João Freitas (eds.),
Prosodies, with special reference to Iberian languages, 59–98. Berlin: De Gruyter.
Daoud, Mohamed. 2007. The language situation in Tunisia. In Robert B. Kaplan
& Richard B. Jr. Baldauf (eds.), Language planning and policy in Africa, 256–307.
Clevedon: Multilingual Matters.
Delais-Roussarie, Elisabeth, Brechtje Post, Caroline Buthke M. Avanzi, Albert
Di Cristo, Ingo Feldhausen, Sun-Ah Jun, Philippe Martin, Trudel Meisenburg,
Annie Rialland, Rafèu Sichel-Bazin & Hiyon Yoo. 2015. Intonational phonology
of French: Developing a ToBI system for French. In Sónia Frota & Pilar Prieto
(eds.), Intonation in Romance, 63–100. Oxford: Oxford University Press.
Devine, A. M. & Laurence D. Stephens. 1985. Stress in Greek? Transactions of the
American Philological Association (1974–2014) 115. 125–152.
Dupoux, Emmanuel, Sharon Peperkamp & Núria Sebastián-Gallés. 2001. A robust
method to study stress “deafness”. Journal of the Acoustical Society of America
110. 1606–1618.
El Zarka, Dina. 2012. One ra – many meanings: Syntax, semantics and prosody
of the Moroccan modal particle ra and its Egyptian Arabic counterparts. STUF
– Language Typology and Universals 65(4). 412–431.
El Zarka, Dina. 2017. Arabic intonation. In Oxford Handbooks Online. DOI:
10.1093/oxfordhb/9780199935345.013.77.
El Zarka, Dina & Sam Hellmuth. 2009. Variation in the intonation of Egyptian
formal and colloquial Arabic. Langues et Linguistique 22. 73–92.
Face, Tim. 2003. Intonation in Spanish declaratives: Differences between lab
speech and spontaneous speech. Catalan Journal of Linguistics 2. 115–131.
Fagyal, Zsuzsanna. 2005. Prosodic consequences of being a Beur: French in
contact with immigrant languages in Paris. University of Pennsylvania Working
Papers in Linguistics 10. 91–104.
597
Sam Hellmuth
Frota, Sónia, Marisa Cruz, Flaviane Fernandes-Svartman, Gisela Collischonn,
Aline Fonseca, Carolina Serra, Pedro Oliveira & Marina Vigário. 2015.
Intonational variation in Portuguese: European and Brazilian varieties. In
Sónia Frota & Pilar Prieto (eds.), Intonation in Romance, 235–283. Oxford:
Oxford University Press.
Frota, Sónia & Pilar Prieto (eds.). 2015. Intonation in Romance. Oxford: Oxford
University Press.
Gabsi, Zouhir. 2011. Attrition and maintenance of the Berber language in Tunisia.
International Journal of the Sociology of Language 211. 135–164.
Gili Fivela, Barbara, Cinzia Avesani, Marco Barone, Giuliano Bocci, Claudia
Crocco, Mariapaola D’Imperio, Rosa Giordano, Giovanna Marotta, Michelina
Savino & Patrizia Sorianello. 2015. Intonational phonology of the regional
varieties of Italian. In Sónia Frota & Pilar Prieto (eds.), Intonation in Romance,
140–197. Oxford: Oxford University Press.
Grice, Martine, Michelina Savino, Alessandro Caffo & Timo B. Roettger. 2015. The
tune drives the text: Schwa in consonant-final loan words in Italian. In The
Scottish Consortium for ICPhS 2015 (ed.), Proceedings of the 18th International
Congress of Phonetic Sciences, Glasgow, UK. Glasgow: University of Glasgow.
Paper number 0381.
Gussenhoven, Carlos. 2006. Between stress and tone in Nubi word prosody.
Phonology 23. 192–223.
Gussenhoven, Carlos. 2017. Zwara (Zuwārah) Berber. Journal of the International
Phonetic Association 48(3). 371–387.
Hansen, Anita Berit. 1997. Le nouveau [ǝ] prépausal dans le français parlé à Paris.
In Jean Perrot (ed.), Polyphonie pour Iván Fonágy, 173–198. Paris: L’Harmattan.
Hellmuth, Sam. 2006. Intonational pitch accent distribution in Egyptian Arabic.
London: SOAS. (Doctoral dissertation).
Hellmuth, Sam. 2007. The relationship between prosodic structure and pitch
accent distribution: Evidence from Egyptian Arabic. The Linguistic Review 24.
289–314.
Hellmuth, Sam. 2014. Dialectal variation in Arabic intonation: Motivations for
a multi-level corpus approach. In Samira Farwaneh & Hamid Ouali (eds.),
Perspectives on Arabic linguistics XXIV–XXV: Papers from the annual symposia
on Arabic linguistics, Texas, 2010 and Arizona, 2011, 63–89. Amsterdam: John
Benjamins.
Hellmuth, Sam. 2016. Explorations at the syntax–phonology interface in Arabic.
In Stuart Davis & Usama Soltan (eds.), Perspectives on Arabic linguistics XXVII:
Papers from the Annual Symposium on Arabic Linguistics, Bloomington, Indiana,
2013, 75–97. Amsterdam: John Benjamins.
598
27 Contact and variation in Arabic intonation
Hellmuth, Sam. 2019. Prosodic variation in Arabic. In Enam Al-Wer & Uri Horesh
(eds.), The Routledge handbook of Arabic sociolinguistics, 169–184. London:
Routledge.
Hellmuth, Sam. Forthcoming. Text–tune alignment in Tunisian Arabic yes-no
questions. In Marisa Cruz & Sónia Frota (eds.), Prosodic variation (with)in
languages: Intonation, phrasing and segments. London: Equinox.
Hellmuth, Sam. In preparation. Intonation in spoken Arabic dialects. Oxford:
Oxford University Press.
Hellmuth, Sam & Rana Almbark. 2017. Intonational Variation in Arabic Corpus.
UK Data Archive http://reshare.ukdataservice.ac.uk/852878/.
Herin, Bruno & Martin R. Zammit. 2017. Three for the price of one: The dialects of
Kerkennah (Tunisia). In Veronika Ritt-Benmimoun (ed.), Tunisian and Libyan
Arabic dialects: Common trends – recent developments – diachronic aspects, 135–
146. Zaragoza: Prensas de la Universidad de Zaragoza.
Holes, Clive. 2018. Introduction. In Clive Holes (ed.), Arabic historical dialectology:
Linguistic and sociolinguistic approaches, 1–28. Oxford: Oxford University
Press.
Jun, Sun-Ah. 2005. Prosodic typology. In Sun-Ah Jun (ed.), Prosodic typology:
The phonology of intonation and phrasing, 430–458. Oxford: Oxford University
Press.
Jun, Sun-Ah. 2015. Prosodic typology: By prominence type, word prosody, and
macro-rhythm. In Sun-Ah Jun (ed.), Prosodic typology II: The phonology of
intonation and phrasing, 520–539. Oxford: Oxford University Press.
Jun, Sun-Ah & Janet Fletcher. 2015. Methodology of studying intonation: From
data collection to data analysis. In Sun-Ah Jun (ed.), Prosodic typology II: The
phonology of intonation and phrasing, 493–519. Oxford: Oxford University
Press.
Kulk, Friso, Cecilia Odé & Manfred Woidich. 2003. The intonation of colloquial
Damascene Arabic: A piot study. Institute of Phonetic Sciences, University of
Amsterdam, Proceedings 25. 15–20.
Ladd, D. Robert. 2006. Segmental anchoring of pitch movements: Autosegmental
phonology or gestural coordination? 18. 19–38.
Ladd, D. Robert. 2008. Intonational phonology. Cambridge: Cambridge University
Press.
Ladd, D. Robert & Rachel Morton. 1997. The perception of intonational emphasis:
Continuous or categorical? Journal of Phonetics 25. 313–342.
Lucas, Christopher. 2015. Contact-induced language change. In Claire Bowern &
Bethwyn Evans (eds.), The Routledge handbook of historical linguistics, 519–536.
London: Routledge.
599
Sam Hellmuth
Maas, Utz. 2013. Die marokkanische Akzentuierung. In Renaud Kuty, Ulrich
Seeger & Shabo Talay (eds.), Nicht nur mit Engelszungen: Beiträge zur
semitischen Dialektologie. Festschrift für Werner Arnold zum 60. Geburtstag,
223–234. Wiesbaden: Harrassowitz.
Maas, Utz & Stephan Procházka. 2012. Moroccan Arabic in its wider linguistic
and social contexts. STUF – Language Typology and Universals 65. 329–357.
Matras, Yaron. 2007. The borrowability of structural categories. In Yaron Matras
& Jeanette Sakel (eds.), Grammatical borrowing in cross-linguistic perspective,
31–73. Berlin: Mouton de Gruyter.
Mennen, Ineke. 2004. Bi-directional interference in the intonation of Dutch
speakers of Greek. Journal of Phonetics 32. 543–563.
Mitchell, T. F. 1993. Pronouncing Arabic 2. Oxford: Clarendon Press.
Nakao, Shuichiro. 2013. The prosody of Juba Arabic: Split prosody, morphophonology and slang. In Mena Lafkioui (ed.), African Arabic: Approaches to
dialectology, 95–120. Berlin: De Gruyter Mouton.
Nance, Claire. 2015. Intonational variation and change in Scottish Gaelic. Lingua
160. 1–19.
O’Rourke, Erin. 2004. Peak placement in two regional varieties of Peruvian
Spanish intonation. In Julie Auger, J. Clancy Clements & Barbara Vance (eds.),
Contemporary approaches to Romance linguistics: Selected papers from the 33rd
Linguistic Symposium on Romance Languages (LSRL), 321–342. Amsterdam:
John Benjamins.
Papaconstantinou, Arietta. 2010. Introduction. In Arietta Papaconstantinou (ed.),
The multilingual experience in Egypt, from the Ptolemies to the Abbasids, 1–16.
London: Ashgate Publishing.
Papaconstantinou, Arietta. 2012. Why did Coptic fail where Aramaic succeeded?
Linguistic developments in Egypt and the Near East after the Arab conquest.
In Alex Mullen & Patrick James (eds.), Multilingualism in the Graeco-Roman
Worlds, 58–76. Cambridge: Cambridge University Press.
Peust, Carsten. 1999. Egyptian phonology: An introduction to the phonology of a
dead language. Göttingen: Peust und Gutschmidt.
Queen, Robin M. 2001. Bilingual intonation patterns: Evidence of language
change from Turkish–German bilingual children. Language in Society 30. 55–
80.
Queen, Robin M. 2012. Turkish–German bilinguals and their intonation:
Triangulating evidence about contact-induced language change. Language 88.
791–816.
600
27 Contact and variation in Arabic intonation
Rahmani, Hamed, Toni Rietveld & Carlos Gussenhoven. 2015. Stress “deafness”
reveals absence of lexical marking of stress or tone in the adult grammar. PLOS
One 10(12: e0143968).
Rastegar-El Zarka, Dina. 1997. Prosodische Phonologie des Arabischen. Graz: KarlFranzens-Universität. (Doctoral dissertation).
Rifaat, Khalid. 1991. The intonation of Arabic: An experimental study. Alexandria:
University of Alexandria. (Doctoral dissertation).
Sayahi, Lotfi. 2011. Introduction: Current perspectives on Tunisian
sociolinguistics. International Journal of the Sociology of Language 211.
1–8.
Simonet, Miquel. 2011. Intonational convergence in language contact: Utterancefinal F0 contours in Catalan–Spanish early bilinguals. Journal of the
International Phonetic Association 41. 157–184.
Sorace, Antonella. 2004. Native language attrition and developmental instability
at the syntax–discourse interface: Data, interpretations and methods.
Bilingualism: Language and Cognition 7. 143–145.
Thomason, Sarah G. & Terrence Kaufman. 1988. Language contact, creolization
and genetic linguistics. Berkeley: University of California Press.
Walker, Traci. 2014. Form (does not equal) function: The independence of prosody
and action. Research on Language and Social Interaction 47. 1–16.
Walters, Keith. 2011. Gendering French in Tunisia: Language ideologies and
nationalism. International Journal of the Sociology of Language 211. 83–111.
Watson, Janet C. E. 2002. The phonology and morphology of Arabic. Oxford:
Oxford University Press.
Watson, Janet C. E. 2011. Word stress in Arabic. In Marc van Oostendorp, Colin
J. Ewen, Elizabeth Hume & Keren Rice (eds.), The Blackwell companion to
phonology, vol. 5, 2990–3018. Oxford: Blackwell.
Weinreich, Uriel, William Labov & Marvin Herzog. 1968. Empirical foundations
for a theory of language change. In Winfred Lehman & Yakov Malkiel (eds.),
Directions for historical linguistics, 95–195. Austin: University of Texas Press.
Wells, John C. 1982. Accents of English. Cambridge: Cambridge University Press.
601
Chapter 28
Contact-induced grammaticalization
between Arabic dialects
Thomas Leddy-Cecere
Bennington College
This chapter describes the phenomenon of contact-induced grammaticalization
between Arabic dialects and its significance in accounting for the development of
future tense markers across modern Arabic varieties. After an introduction to theoretical aspects of general grammaticalization theory and contact-induced grammaticalization in particular, discussion shifts to the identification of specific contact-induced grammaticalization processes leading to the modern distribution of
future tense-marking forms across the Arabic-speaking world. Finally, the significance of these findings to broader inquiry in Arabic dialectology and theoretical
contact linguistics is considered.
1 Introduction & theory
1.1 Overview
This chapter presents evidence for the occurrence of contact-induced grammaticalization processes between dialects of Arabic over the course of the language’s
history. The critical role of dialect contact as a source of synchronic variation and
diachronic change across Arabic varieties is well recognized, and the description of its outcomes a long-standing occupation of Arabic dialectologists (e.g.
Behnstedt & Woidich 2005; Miller et al. 2007). Representing a fairly recent theoretical development in the field of contact linguistics – following largely from
the proposals of Heine & Kuteva (2003; 2005) and Dahl (2001) – contact-induced
grammaticalization as a model has not been applied to the analysis of Arabic
dialect data on a large scale. As will be seen, however, the phenomenon displays
significant merit as an account for the evolution and diffusion of a number of
morphosyntactic features across the modern Arabic dialects.
Thomas Leddy-Cecere. 2020. Contact-induced grammaticalization between Arabic
dialects. In Christopher Lucas & Stefano Manfredi (eds.), Arabic and contact-induced
DOI:10.5281/zenodo.3744555
change, 603–623. Berlin: Language Science Press.
Thomas Leddy-Cecere
In the following subsections, I begin with a review of the current state of research in grammaticalization theory and in contact-induced grammaticalization
(CIG) specifically. I then proceed to an illustrative example of CIG in the Arabic
context, demonstrating the model’s power as an explanatory mechanism in interpreting the distribution of future tense markers across the modern dialects. I
conclude the chapter with a brief discussion of the broader significance of CIG in
the analysis of Arabic and the potential role for Arabic data in advancing general
theoretical knowledge of the phenomenon at large.
1.2 Grammaticalization
Most linguists agree that it is possible to synchronically classify the majority
of linguistic forms along a cline from “more lexical” to “more grammatical”, in
a manner roughly consistent with the progression as conceived by Hopper &
Traugott (2003):
content word > grammatical word > clitic > inflectional affix
Historical linguists would add to this synchronic observation the diachronic
reflection that it is common to observe a single etymological item advancing
through the successive stages of this cline as it develops as part of a linguistic
system over time. In fact, the sheer frequency of examples indicating such a trajectory of evolution has led to the identification of a cross-linguistically attested
phenomenon known as grammaticalization. The following definition provided
by Hopper and Traugott is indicative of several currently referenced in the field,
which – though differing in emphasis and points of detail – broadly subscribe to
a similar central principle:
[Grammaticalization is] the change whereby lexical items and constructions
come in certain linguistic contexts to serve grammatical functions and, once
grammaticalized, continue to develop new grammatical functions (Hopper
& Traugott 2003: 18).
Though useful for purposes of general definition, this largely intuitive formulation of grammaticalization and the cline which it follows must be further deconstructed if they are to be operationalized as part of a rigorous analysis. Andersen
(2008) summarizes the issue succinctly in observing that the grammaticalization
cline so articulated conflates numerous discrete dimensions of language change
by presenting them as unified steps in a chain: the shift from lexical to grammatical word is one of semantic content, while that from word to clitic to affix
604
28 Contact-induced grammaticalization between Arabic dialects
involves morphosyntax and any associated loss of phonological material is best
understood as phonological change. Since the early stages of grammaticalization research, more complex approaches to the description of the phenomenon
have been proposed based on the concurrent evaluation of multiple parameters
(e.g. Lehmann 1985). Other authors opt instead to define analogous parameters
in terms of diachronic processes, thereby rendering them more directly relatable
to the modes of historical linguistic analysis which underlie the bulk of investigations in grammaticalization research. The latter approach is adopted here, largely
following the account proposed by Heine (2004).
Heine views grammaticalization as defined by the simultaneous progression
of four distinct but interrelated diachronic processes: desemanticization, extension, decategorialization, and erosion. Desemanticization involves the loss of concrete lexical (“content”) meaning and a corresponding rise in abstract grammatical function associated with the use of an item in particular contexts. This often
represents the first observable stage of grammaticalizing change, and, as its name
suggests, primarily concerns the semantic content of the item rather than its distribution, form, or syntactic behavior. Closely coupled with desemanticization
is extension, namely the novel use of the grammaticalizing item in contexts in
which it was not previously employed; extension is thus defined as a change in
incidence. The hand-in-hand advance of these two processes is demonstrated in
the evolution of the French negative element pas: having shed its content semantics as a noun meaning ‘step’ and developed grammatical function as a marker
of verbal negation, pas is extended in contemporary usage to contexts involving none of the implied motion of its lexical source (Hansen & Visconti 2009:
137–138).
The third process described by Heine, that of decategorialization, consists of
the changes by which a grammaticalized item comes to lose the morphosyntactic
properties characteristic of its source’s original word class, such as word order
freedom or agreement inflection; an example may be found in the gradual development of the English adverbial marker -ly from a morphosyntactically free
substantive meaning ‘body, form’ to a bound derivational suffix (Ramat 2011: 505).
Erosion, the fourth process considered by Heine, refers to the gradual reduction
and lenition of phonological form beyond what is accounted for by regular sound
change, as observed in the irregular changes deriving the Jewish Babylonian Aramaic continuous aspect marker qā ~ kā from earlier *qāʔē ‘standing’ (Rubin 2005:
134).
Theories of grammaticalization have also been strongly linked to the notion
of unidirectionality, the proposal that change along the above-described cline occurs only from more lexical to more grammatical and not vice versa (Lehmann
605
Thomas Leddy-Cecere
2015). Though the absolute formulation of this hypothesis has been the subject of
much debate (e.g. Norde 2009), recognition of a strong unidirectional tendency
remains integral to understandings of grammaticalization on both empirical and
theoretical grounds (Haspelmath 1998; Heine 2004). It has been proposed that
the impetus for such a tendency lies in a universal set of cognitive and communicative principles common to the human mental faculty (Claudi & Heine 1986;
Bybee 2003; Lehmann 2015); these would provide an account for the pervasive
occurrence of grammaticalization as a worldwide phenomenon, and may be seen
to bias the results of grammatical change in the directions entailed by the four
processes described above.
The concomitant advancement of these processes is discernible in one of the
few cases of Arabic grammaticalization for which a reasonably complete chain
of historical development is attested: that of the Egyptian Arabic future tense
marker ḥa-. Documented in sixteenth- and seventeenth-century sources as rāyiḥ,
this item already shows evidence of desemanticization and extension, departing
from the semantics of its lexical source as an active participle ‘going’ to indicate
intention and imminent futurity of action, consequently allowing its extension
to usage contexts devoid of actual motion: ʔanā rāyiḥ aɣannī ʕalēh ‘I am going to sing about it [and proceeds to sing]’ (Davies 1981: 241). In its nineteenthcentury incarnation rāḥ ~ raḥ ~ ḥa-, the form shows further desemanticization
and extension as its value changes from an imminent to a general future and
it comes to be employed in previously unacceptable circumstances, such as in
the presence of a non-immediate temporal adverb: rāḥ yīgi bukra ‘he’ll come tomorrow’ (Elias & Elias 1981: 157; for earlier usage constraints, see Davies 1981:
241-243). These increasingly modern forms also attest decategorialization, as the
once-obligatory adjectival agreement marking of the participial original becomes
optional – raḥ (sg.m)/raḥa (sg.f)/raḥīn (pl) ~ raḥ (invar.) (Vollers 1895: 40) – and
eventually ceases to exist altogether in the tightly bound modern clitic ḥa- ~ ha(Abdel-Massih et al. 2009: 268). Fourthly, progressive phonetic erosion is visible
throughout the item’s history, as none of the loss or lenition of phonetic material
through the stages rāyiḥ > rāḥ > raḥ > ḥa- > ha- attested above is attributable to
regular sound change. Taken together, these combined processes chart the grammaticalization of lexical rāyiḥ ‘going’ through its gradual development into the
modern future tense clitic ḥa-.1
1
On sources referenced in the preceding paragraph: Davies (1981) is a study of colloquial elements in the seventeenth century Egyptian text Hazz al-quḥūf fī šarḥ qaṣīd Abī Ṣādūf ; Vollers
(1895) is a descriptive grammar of Egyptian Arabic at the close of the nineteenth century, and
Elias & Elias (1981) an English–Egyptian Arabic vocabulary and phrasebook first released ca.
1899; Abdel-Massih et al. (2009) is a reference grammar of modern Egyptian Arabic.
606
28 Contact-induced grammaticalization between Arabic dialects
Having established these understandings of grammaticalization and its component processes, we now turn to the proposal that specific grammaticalization
pathways are able to be shared between interacting languages and dialects: the
aforementioned CIG.
1.3 Contact-induced grammaticalization
The most fully elaborated theoretical model proposed for CIG is that of Heine &
Kuteva (2003; 2005). This model represents the phenomenon by which “a grammaticalization process […] is transferred from the model (M) to the replica language (R),” without corresponding transfer of any actual phonological form (2003:
539). As paraphrased and clarified by Law (2014: 215), this occurs when one language, “the ‘replica language,’ develops a feature observed in another language,
the ‘model’ language, but goes through a path of universal development using
resources internal to the replica language.” The result is such as that seen in the
Basque innovation of an allative preposition from the noun buru ‘head’ and a
perfect tense formed with the lexical verb ukan ‘have’, apparently influenced by
parallel grammaticalizations in neighboring varieties of Romance (Haase 1992
apud Heine & Kuteva 2003: 550). Such an effect is proposed to be actuated according to the following model (Heine & Kuteva 2003: 539):
1. Speakers of language R notice that in language M there is a grammatical
category Mx .
2. They develop an equivalent category, Rx , using material available in their
own language (R).
3. To this end, they replicate a grammaticalization process they assume to
have taken place in language M, using an analogical formula of the kind
[My > Mx ] = [Ry > Rx ].
4. They grammaticalize Ry to Rx .
This proposal for the diffusion of parallel grammaticalization trajectories between linguistic varieties is presaged by Bisang’s (1998) observation of the potential synergy between grammaticalization, which he considers to be a primarily
construction-based process, and previously observed forms of contact-induced
structural convergence. The phenomenon has also been influentially described
by Dahl in the form of “gram families”, consisting of groups of “grams [grammaticalized items] with related functions and diachronic sources that show up
607
Thomas Leddy-Cecere
in genetically and/or geographically related groups of languages, in other words,
what can be assumed to be the result of one process of diffusion” (Dahl 2001:
1469). Heine & Kuteva draw heavily on Dahl’s theorizations, though they diverge
from him in a few critical ways. First, they are significantly more conservative
than Dahl in identifying examples of the phenomenon, insisting on corroborating evidence of language contact in order to posit CIG rather than inductively
inferring its occurrence given genetic relatedness or proximity. Second, they do
not necessarily attempt to link multiple replications of the same grammaticalization pathway into “one process of diffusion,” but instead prefer to treat them as
individual instances of contact between participating languages.
Further, Heine & Kuteva’s model is primarily situated in the context of contact between genetically distinct languages. Regarding the occurrence of CIG
between related language varieties or dialects, Dahl sees such scenarios as generating the bulk of evidence for the phenomenon: “in the majority of all such
cases [of areally diffused grams], the languages involved are more or less closely
related” (2001: 1469). Heine & Kuteva are wary of such identifications. Critically,
however, their reasons for being so are methodological rather than theoretical.
In their analysis, they choose to rely on the principle of genetic patterning as “an
empirically well-founded tool for identifying cases of contact-induced linguistic
transfer” (2005: 33–34), meaning that examples of CIG between unrelated languages are often easiest to identify and defend and thus have been favored in the
effort to present an unambiguous account. Regarding the broader occurrence of
the phenomenon, however, they state that “genetic relationship is entirely irrelevant” (2005: 184) and that CIG may occur between related languages just as it
does between unrelated ones. They remain, though, more careful than Dahl to
set apart cases attributable to inheritance of any stage of the grammaticalization
chain from a common ancestor, which could lead to a superficially similar result
not in fact dependent on any degree of contact. Along the same lines, Law (2014)
reminds us that when dealing with closely related languages the possibility of
drift or typological poise precipitating parallel development rises dramatically
in likelihood. Thus, the analyst must be stringent in linking proposed cases of
CIG to cross-linguistically attested paths and parameters of grammaticalization
and not to the local idiosyncrasies of a given language family or subgroup.
To Heine & Kuteva, CIG is unambiguously situated in terms of Van Coetsem’s
(1988; 2000) dichotomy between source language (SL) and recipient language
(RL) agentivity: the four-stage model of replication presented above clearly identifies speakers of the RL as the agents of contact-induced change in this instance.
This judgment has opened the proposal to major critique, as several key theorists maintain that structural pattern replication of the kind required for CIG
608
28 Contact-induced grammaticalization between Arabic dialects
is only possible in a scenario of SL agentivity. Ross (2007), for example, views
the phenomenon as part of a broader process of bilingual calquing (cf. Manfredi,
this volume), involving the subconscious imposition of the functional range of
an SL item onto its RL equivalent, followed secondarily by the processes Heine
& Kuteva attribute to grammaticalization but which Ross views as the natural
result of increases in frequency and automization stemming from the RL item’s
new functionality. Ross asserts that “one cannot reasonably argue” for Heine &
Kuteva’s construal of CIG as an RL-agentive direct replication of a grammaticalization process because of his conviction that the phenomenon is “largely driven
by effort reducing practices of which speakers are only marginally aware” (2007:
135).
Matras (2009; 2011), however, supports Heine & Kuteva’s initial characterization by arguing for RL-agentivity in his own recent accounts of CIG, which
provide more attention to the role played by the individual bilingual in the phenomenon’s actuation. He cites the individual’s communicative imperatives and
creative impulse as the primary force driving the replication of grammaticalization processes, as speakers actively borrow from constructions they control in
one of their languages, as a source of expressive innovation in the other, limiting this transfer solely to “pattern” out of respect for the norms of the distinct
speech communities in which they operate. Matras’ account thus has the benefit
of aligning with the motivating forces theorized to obtain for grammaticalization processes more generally. As described by Lehmann in his consideration of
grammaticalization’s communicative/pragmatic dimension:
To the degree that language activity is truly creative, it is no exaggeration to
say that languages change because speakers want to change them … they
do not want to express themselves the same way they did yesterday, and
in particular not the way that somebody else did yesterday (Lehmann 1985:
315).
Building upon this position, it holds that in scenarios of language or dialect
contact innovating speakers may very well wish to express themselves the same
way somebody else did yesterday if the means of expression involved are novel
to a distinct speech community with which they are interacting today. This synergy with less controversial understandings of grammaticalization outside the
context of language/dialect contact provides a viable counterpoint to the skepticism voiced by Ross, and strongly recommends the association of CIG with
RL-agentivity.
609
Thomas Leddy-Cecere
In the case of contact between closely related varieties, this characterization
of CIG may be further qualified. Under Matras’ RL-agentive formulation, pattern replication occurs in the presence of constraints against the usage of an
SL’s forms due to speaker expectation and “language loyalty” among members
of the RL community (2011: 283). In the broader language contact literature, such
sociolinguistic constraints have been noted to play a role in pattern diffusion (e.g.
Epps 2005), and presumably interact with speakers’ judgments of interlocutors’
perceived bilingual competency in favoring or disfavoring matter-based mixing
or borrowing (cf. Grosjean 2001). In contexts where mutual comprehensibility is
a less salient concern, the drivers of pattern vs. matter-based innovation would
be expected to be almost purely sociolinguistic and pragmatic, and are perhaps
most fruitfully understood through the lenses of indexicality (Silverstein 2003)
and focus at the level of the speech community (Le Page & Tabouret-Keller 1985)
rather than as a desire to adhere to a reified linguistic code. Such is the state of
affairs most likely to obtain in the case of CIG between neighboring varieties of
Arabic, to which we now turn in detail.
2 CIG in the development of Arabic future tense markers
2.1 Methods of investigation
In the following subsections, I present evidence for the role of CIG as the primary
mechanism underlying the development and distribution of future tense markers in the modern Arabic dialects. The data considered is drawn from a survey of
eighty-one geographic sample points spanning the contiguous Arabic-speaking
world, based on a total of eighty-eight descriptive sources.2 This sample was
constructed as part of the broader investigation of CIG between Arabic dialects
presented by Leddy-Cecere (2018), which investigates the role of CIG in the development of a number of morphosyntactic features in modern Arabic varieties,
including future tense markers, genitive exponents, and temporal adverbs meaning ‘now’. A discussion of data and findings for the first of these features is the
focus of this chapter, and these shall be seen to argue strongly for the identification of CIG as a key force in shaping the evolution of modern Arabic dialects.
Readers are encouraged to refer to Leddy-Cecere (2018) for additional examination and expansion of the points to follow.
2
Details of sample composition and sources consulted, as well as the selection criteria for each,
may be found in Leddy-Cecere (2018: 34–35, 43–46).
610
28 Contact-induced grammaticalization between Arabic dialects
To begin, I first examine the complete set of specific grammaticalizations of
future tense markers attested by the Arabic dialect data. These have been identified via the observation of concurrent processes of desemanticization, extension,
decategorialization and erosion (as described in §1.2). These individual instances
are further sorted into higher level groupings by grammaticalization path: future tense markers deriving from motion verbs meaning ‘go’, for example, as
against those deriving from purposive constructions, etc. Special attention is paid
to those specific grammaticalization paths represented by multiple evolutions involving distinct etyma, thus identifiable as potential candidates for the products
of replication through CIG. As a final step in the evaluation, the geographic incidence of forms representing such multiply attested paths is considered to assess
whether their modern distribution is consistent with a historical account of diffusion via contact. This latter portion of the analysis will be presented in §2.3,
following the complete accounting of grammaticalized forms provided in §2.2
immediately below.3
2.2 Grammaticalizations of Arabic future tense markers by
grammaticalization path
2.2.1 Futures from ‘go’ (fut < go)
Grammaticalizations of future tense markers from forms of lexical verbs meaning ‘go’ are well represented in the Arabic data. This grammaticalization path is
widely attested cross-linguistically, providing one of the major sources for the development of future tense markers worldwide (Bybee et al. 1994; Heine & Kuteva
2002). Grammaticalizations of specific items observed in the cross-dialectal Arabic sample are described below.
*rāyiḥ:
Future markers representing grammaticalized forms of an active participle *rāyiḥ ‘going’ are found across a broad east-west swath of the Arabicspeaking world, extending from southern Iraq in the east to Algerian territory in the west. Differing degrees of grammaticalization are attested,
with some forms maintaining full phonological integrity and categorial
membership (e.g. Basra rāyiḥ; Mahdi 1985, which displays adjectival gender/number agreement with its subject) and others showing dramatic erosion and of loss of morphosyntactic autonomy (including Cairo ḥa- ~ ha-,
3
For further discussion and justification of each stage of this heuristic, see Leddy-Cecere (2018:
36–43, 209–214).
611
Thomas Leddy-Cecere
as described in §1.2). Semantically, some forms, such as Algiers rāḥ and
Jerusalem rāyiḥ ~ rāḥ ~ ḥā-, are recorded as expressing a value of immediate future or future intent (Boucherit 2011; Rosenhouse 2011), while the
majority are associated with a meaning of general futurity.
*ɣādī:
Grammaticalized forms of the active participle *ɣādī ‘going’ are common
future markers throughout much of Morocco and adjacent regions. Reduced and invariant forms often co-exist alongside less grammaticalized
reflexes, thereby attesting discrete links in an increasingly advanced grammaticalization chain, as seen in Casablanca ɣādī ~ ɣa- (Caubet 2011).
*māšī:
Grammaticalized future markers deriving from the active participle *māšī
‘going’ occur in two distinct geographic pockets, one centered in northcentral Morocco and the other in Tunisia and eastern Algeria. In addition
to more predictable effects of phonetic erosion and decategorialization, several forms from the latter area display a further example of irregular sound
change with the sporadic denasalization of */m/ > /b/, as in Sousse māš ~
bāš (Talmoudi 1980).
*sāyir:
Alone in the sample, Maltese attests a future marker deriving from a grammaticalized form of the active participle *sāyir ‘going’. This can be found in
both an inflecting form sejjer (sg.m)/sejra (sg.f)/sejrin (pl) and as the more
grammaticalized, invariant forms ser and se (Vanhove 1993).
2.2.2 Futures from ‘want’ (fut < want)
Grammaticalizations from source constructions indicating desire or volition are
another cross-linguistically common origin for future tense markers (Bybee et al.
1994; Heine & Kuteva 2002), and are similarly widespread to their fut < go counterparts in the Arabic dialect data. Specific grammaticalizations are discussed
below.
*yabɣā ~ yabɣī:
Grammaticalized forms of the imperfect verb *yabɣā ~ yabɣī ‘want’ serve
as future markers across a large portion of the Arabic-speaking world,
stretching from the Arabian Peninsula across the Red Sea to the greater
Sudanic area, and then northward through present-day Libya. While many
Arabic varieties attest only a highly grammaticalized, reduced form of the
612
28 Contact-induced grammaticalization between Arabic dialects
item (e.g. Abu Dhabi b-; Qafisheh 1977), other dialects display direct evidence of multiple stages of the grammaticalization chain, e.g. Ḥarb yabɣā
~ yabā ~ ba- (Il-Hazmy 1975).
*biddu ~ widdu:
Future tense markers arising from grammaticalizations of *biddu ~ widdu
‘want’ are found throughout the broader Levantine area. In their most phonetically reduced forms (e.g. Soukhne b-; Behnstedt 1994), they are often
superficially indistinguishable from the highly grammaticalized products
of *yabɣā ~ yabɣī discussed above; several dialects, however, provide clear
evidence for a distinct chain of development, such as Jebel Ansariye baddo
~ bado ~ b- (Lewin 1969) and Cilicia baddu ~ baddi- ~ bad- (Procházka 2011).
In some varieties, grammaticalizations of *biddu ~ widdu operate alongside
other markers of future tense to designate a more specified value: Damascus bǝddo ~ b- is often reported to denote a modal value of possible or
planned future, as opposed to the *rāyiḥ-derived forms raḥ ~ ḥa- which indicate a higher degree of certainty or expectation (cf. Lentin 2011). In other
dialects, these forms would appear to have further desemanticized and extended to a value of more general futurity. Future investigation is needed
into the degree to which reduced reflexes of *biddu may have merged in
mental representation with the continuous aspect marker b- present in
many of the same varieties; relevant parallels might be drawn with scenarios of near homophony like that found in the dialect of Dhofar, in which
continuous bi- exists alongside future bā- (< *yabɣā ~ yabɣī; Davey 2016).
*yišā:
In a number of Yemeni dialects, the future tense marker may be traced
to a grammaticalized form of *yišā ‘want’. It is notable that in cases such
as Sana’a ša- this form is used only with the first person singular verb
(Watson 1993); in such circumstances, it is possible that its ultimate source
should be more properly identified with *ašā ‘I want’.
*ydawr:
Varieties belonging to the Ḥassāniyya dialect complex of Mauritania and
neighboring Mali are recorded as utilizing a grammaticalized form of the
verb *ydawr ‘want’ with a following imperfect verb to denote a value of
intentional future. This grammaticalization is relatively light, consisting
primarily of desemanticization and extension with little in the way of decategorialization or erosion: Nouakchott ydoːr, for example, denotes future
intent while continuing to operate morphosyntactically as a fully inflected
finite verb (Taine-Cheikh 2011).
613
Thomas Leddy-Cecere
*bɣā:
Dialects of southern Morocco and southwestern Algeria occasionally attest grammaticalized forms of *bɣā ‘want’ expressing a future tense value.
Though lexically similar in origin to the grammaticalizations based on
*yabɣā ~ yibbā ~ yibbī discussed above, the phonological shape of these
items (e.g. Marrakech bɣa: ~ ba-; Sánchez 2014) recommends an identification of their source in the perfect stem *bɣā, which is the typical means for
expressing ‘want’ in this area.
2.2.3 Futures from ‘come’ (fut < come)
Another cross-linguistically common path of future tense grammaticalization,
that involving verbs meaning ‘come, return’ (Bybee et al. 1994; Heine & Kuteva
2002), is represented in the Arabic data by markers originating from a single
source etymon, *ʕād ‘return’.
*ʕād: Future tense markers traditionally identified as grammaticalized forms of
*ʕād ‘return’ are attested in three locations in the cross-dialectal survey:
Yemen, Upper Egypt and interior Tunisia. The forms found in Tunisia and
Egypt, Tozeur ʕa- and Aswan ʕa- (Saada 1984; Schroepfer 2019), are highly
reduced, and thus difficult to ascribe definitively to a specific source. It is
notable that in both of these dialects the markers in question vary with
a ‘go’-derived future ḥa- and could thus plausibly represent an erosion of
the latter in the form of a sporadic lenition of /ḥ/ > /ʕ/ (not to mention
that Aswan ʕa- on its own might be linked to local ʕāyiz ~ ʕāwiz ‘want’).
At least in the case of the Yemeni forms, however, an origin in *ʕād seems
clear, as reduced forms such as Sana’a ʕā- display an allomorph ʕad- in
prevocalic contexts (Watson 1993).
2.2.4 Futures from purposive constructions (fut < purp)
A further source of future tense markers in the Arabic data involves the grammaticalization of purposive operators. This path is not widely discussed in the
cross-linguistic grammaticalization literature, though intriguingly the reverse
trajectory, that of purp < fut, is noted (Bybee et al. 1994). The primary difficulty would seem to rest in the identification of a clear process of desemanticization, as it is difficult to judge precisely which function between fut and purp is
more concrete/abstract than the other. Despite this, the occurrence of extension,
decategorialization and erosion in the Arabic forms seems to recommend their
identification as products of a grammaticalization process.
614
28 Contact-induced grammaticalization between Arabic dialects
*ḥattā:
Grammaticalizations of *ḥattā ‘in order to’ are used to indicate future tense
in areas of northern Mesopotamia, the coastal Levant, and Oman. In terms
of geographic distribution and the specific path of phonetic erosion followed, it may be possible to recognize Levantine and Mesopotamian forms
like Cypriot Maronite tta- and Mosul də- (Jastrow 1979; Borg 1985) as representing a single historical innovation, though Oman ḥa- ~ ha- is more
likely an independent development. In the Omani case, the attested use of
ḥa- with purposive meaning recommends a source in *ḥattā rather than gofuture *rāyiḥ: šrab ḥa-turwe! ‘Drink so your thirst be quenched!’ (Reinhardt
1894: 276).
2.2.5 Futures from ‘to busy oneself with’ (fut < verb of
activity/preparation)
A small number of Arabic dialects utilize a future tense marker seeming to derive
from a grammaticalized form of a verb meaning ‘to busy oneself’. Such a path of
development is not discussed in the cross-linguistic literature on grammaticalization, but perhaps has a counterpart in the use of grammaticalized Southern
American English fixing to ~ fixin’ a ~ fi’na to express proximate futurity (cf.
Wolfram & Schilling-Estes 1998). In any case, obvious desemanticization, extension, decategorialization and erosion of the source form indicate a clear example
of grammaticalization here.
*lāhī:
In the Ḥassāniyya dialects of Mauritania, Mali and far southern Morocco,
the future tense marker derives from a grammaticalized form of *lāhī, itself
the active participle form of the verb lha ‘to busy oneself’. Decategorialization is attested in all cases by the lack of adjectival agreement marking
predicted for the original participial, and in at least some varieties phonetic erosion is evidenced as well: Mali lāhi ~ lā (Heath 2003).
2.3 Evidence of replication and diffusion via contact
Of the five grammaticalization paths for Arabic future markers presented above,
two merit closer examination in the search for evidence of CIG: those of fut
< go and fut < want. These paths are identified due to the fact that each is
represented in the data by multiple, parallel realizations arising from etymologically distinct but semantically and functionally analogous sources. Such a state of
615
Thomas Leddy-Cecere
affairs plausibly reflects the result of continued processes of replication, whereby
a grammaticalization process occurring in one Arabic variety is transferred to another and recreated using native etymological material.
Both paths identified, however – together representing the great majority of
future tense markers attested in the sample – are also extremely common crosslinguistically, and could, in principle, have fed multiple independent developments instantiated across the modern Arabic dialect continuum. Key to selecting
between an analysis of CIG and one of repeated, internally-motivated grammaticalization is the factor of geography, as in the absence of fine-grained historical
sociolinguistic data (see §1.3) this is perhaps the most reliable proxy in positing
the feasibility of historical contact between dialects. In the case of CIG, analogous grammaticalization processes ought to be positioned in a geographically
contiguous (or near contiguous) bloc, consistent with a history of diffusion via
contact between speakers of neighboring dialects. In a scenario of independent
development, on the other hand, one should expect the various grammaticalizations to be more or less randomly distributed across the map, equally likely to
occur in any individual dialect considered.
The geographic incidences of the members of the fut < go and fut < want
paths both clearly align with the contiguous profile anticipated for the results
of CIG. All realizations of fut < go future tense markers described connect geographically with other members of the bloc. The large eastern and central zone
of *rāyiḥ futures, encompassing southern Mesopotamia, much of the Levant, and
the Nile Valley, stretches westward across Libya (where *rāyiḥ-derived forms
are recorded alongside *yibbī-based want-futures) to include most of Algeria.
Directly adjacent to this North African arm of the *rāyiḥ forms are found grammaticalizations of *ɣādī in Morocco and of *māšī in Tunisia. Further neighboring
or co-territorial with the latter two areas are a second set of *māšī-based forms in
northern Morocco and the Maltese *sāyir-derived future tense marker, thus completing the connected geographic trend. Future markers representing the path
fut < want display a similar spatial contiguity. Grammaticalizations of *biddu
~ widdu in the Levant stretch to reach those of *yabɣā ~ yabɣī present in the
Arabian Peninsula. These in turn span the Red Sea across to the greater Sudanic
area and northward through the central Sahara into Libya. Moving to the west
and southwest of this zone, the next future markers encountered include grammaticalizations of *bɣā and *ydawr, respectively. Rounding out the set, forms
derived from *yišā exist in close proximity to analogous *yabɣā ~ yabɣī futures
in Yemeni territory. While the integrity of this want-future bloc may seem to
be challenged by natural features such as the Red Sea and the Sahara Desert, historical and anthropological investigations of the regions in question have rather
616
28 Contact-induced grammaticalization between Arabic dialects
shown persistent social and cultural connectivity across these would-be barriers
(Lydon 2009; Power 2012). This evaluation is supported by the distribution of
additional Arabic dialectological isoglosses extending beyond the discussion of
CIG.
The geographic contiguity displayed by the representatives of both the fut <
go and fut < want pathways favors an interpretation of areal diffusion over one
of independent, internally motivated occurrence (of the type perhaps evidenced
by the more scattered distributions of the sole representatives of fut < come
and fut < purp). The optimal account for the development of the modern Arabic go- and want-futures, together representing the greater part of future tense
markers attested in the data, is thus one by which grammaticalization processes
leading to the development of new future tense markers have repeatedly been
subject to transfer and replication between speakers of neighboring dialects. A
CIG-driven analysis such as this has the benefit of accounting for both the development of individual dialect forms and more global trends in source semantics
and geographic incidence, and offers a theoretically unified interpretation of the
Arabic data obtaining on multiple scales.
3 Conclusion
The analysis summarized above has demonstrated the significant explanatory
power of CIG as an account for the development of Arabic future tense markers. Additional proposed occurrences of CIG between Arabic dialects, pertaining
to genitive exponents and temporal adverbs meaning ‘now’, are identified and
examined in Leddy-Cecere (2018). Together with the future tense data discussed
here, these call for corroboration and refinement at the hands of future investigators.
Should further examination provide evidence for a widespread history of CIG
between Arabic dialects, this finding could prove instrumental in satisfactorily
accounting for a number of so-called “pluriform” developments which have repeatedly vexed students of Arabic dialectology. Defined by Versteegh as functionally analogous but etymologically disparate developments for which “a general
trend … has occurred in all Arabic dialects, and an individual instantiation of
this trend in each area,” dialect contact has most often been dismissed as a causal
mechanism for these innovations due to a belief that “typically dialect contact
leads to the borrowing of another dialect’s markers, not to the borrowing of a
structure, which is then filled independently” (2014: 146). CIG provides a theoretical mechanism by which precisely such borrowing and filling may occur, and as
such offers the dialectologist a novel analytical tool in the elucidation of structural transfer and diffusion between Arabic varieties.
617
Thomas Leddy-Cecere
A critical open question in the application of CIG to the Arabic context, as
well as in study of the phenomenon more generally, lies in the problem of agentivity and actuation (as discussed in §1.3). Here, too, further accrual of Arabic
data has the potential to inform broader domains of inquiry. If Arabic is established as a productive ground for the study of CIG and significant cases of the
historical transfer of grammaticalization pathways between dialects are brought
to light, it stands to reason that the same societal and linguistic forces which
have motivated these to take place may still be in force, and that observation of
synchronic Arabic dialect interaction represents a singular opportunity to catch
newly occurring instances of CIG “in the act” and to observe their progress in
real time (for at least one such attempt already presented, see Abuamsha 2016).
Studies of this type will enable linguists to add critically lacking synchronic data
to their sociolinguistic and psycholinguistic analyses of CIG, and so elaborate
and strengthen ongoing theorizations of a revelatory new dimension of contactinduced language change.
Further reading
For a complete theoretical discussion of contact-induced grammaticalization, see
Heine & Kuteva (2003; 2005). The former work is an article-length sketch of the
proposal and is valuable as a direct and concise reference, while the latter provides a more elaborated description with additional linguistic examples. Matras
(2009) provides valuable commentary and critique of Heine & Kuteva’s work
while simultaneously extending exploration to the psycholinguistic and sociolinguistic dimensions of CIG.
For an overview of grammaticalization processes in the development of Arabic
future tense markers (though without reference to contact) see Stewart (1998). A
more detailed treatment of CIG processes in the development of the Arabic future markers and other morphosyntactic features may be found in Leddy-Cecere
(2018).
Acknowledgements
I am grateful to Drs. Kristen Brustad and Daniel Law for their contribution to the
broader work which this chapter reflects, as well to the participants and organizers of the workshop “Arabic and Contact-Induced Change” at the 23rd International Conference on Historical Linguistics for their insightful feedback relating
to the ideas discussed here.
618
28 Contact-induced grammaticalization between Arabic dialects
This material is based upon work supported by the National Science Foundation Graduate Research Fellowship under Grant No. DGE-1110007. Any opinion, findings, and conclusions or recommendations expressed in this material
are those of the author and do not necessarily reflect the views of the National
Science Foundation.
Abbreviations
CIG
f
fut
invar.
m
contact-induced grammaticalization
feminine
future
invariant
masculine
pl
purp
RL
sg
SL
plural
purposive
recipient language
singular
source language
References
Abdel-Massih, Ernest, Zaki Abdel-Malek, El-Said Badawi & Ernest McCarus.
2009. A reference grammar of Egyptian Arabic. Washington, DC: Georgetown
University Press.
Abuamsha, Duaa. 2016. The future marker in Palestinian Arabic: An internal or
contact-induced change? In Lindsay Hracs (ed.), Proceedings of the 2016 Annual
Conference of the Canadian Linguistic Association.
Andersen, Henning. 2008. Grammaticalization in a speaker-oriented theory of
change. In Thórhallur Eythórsson (ed.), Grammatical change and linguistic
theory: The Rosenthal Papers, 11–44. Amsterdam: John Benjamins.
Behnstedt, Peter. 1994. Der arabische Dialekt von Soukhne (Syrien). Teil 2: Phonologie, Morphologie, Syntax. Teil 3: Glossar. Wiesbaden: Harrassowitz.
Behnstedt, Peter & Manfred Woidich. 2005. Arabische Dialektgeographie: Eine Einführung. Leiden: Brill.
Bisang, Walter. 1998. Grammaticalization and language contact, constructions
and positions. In Ana Giacalone-Ramat & Paul Hopper (eds.), The limits of
grammaticalization, 13–58. Amsterdam, Philadelphia: John Benjamins.
Borg, Alexander. 1985. Cypriot Arabic: A historical and comparative investigation
into the phonology and morphology of the Arabic vernacular spoken by the
Maronites of Kormakiti village in the Kyrenia district of north-west Cyprus.
Stuttgart: Deutsche Morgenländische Gesellschaft.
Boucherit, Aziza. 2011. Algiers Arabic. In Lutz Edzard & Rudolf de Jong (eds.),
Encyclopedia of Arabic language and linguistics, online edn. Leiden: Brill.
619
Thomas Leddy-Cecere
Bybee, Joan. 2003. Cognitive processes in grammaticalization. In Michael
Tomasello (ed.), The new psychology of language, vol. 2, 145–167. Mahwah, NJ:
Lawrence Erlbaum Associates, Inc.
Bybee, Joan, Revere Perkins & William Pagliuca. 1994. The evolution of grammar:
Tense, aspect, and modality in the languages of the world. Chicago, London:
Chicago University Press.
Caubet, Dominique. 2011. Moroccan Arabic. In Lutz Edzard & Rudolf de Jong
(eds.), Encyclopedia of Arabic language and linguistics, online edn. Leiden: Brill.
Claudi, Ulrike & Bernd Heine. 1986. On the metaphorical basis of grammar.
Studies in Language 10(2). 297–335.
Dahl, Östen. 2001. Principles of areal typology. In Martin Haspelmath, Ekkehard
König, Wulf Oesterreicher & Wolfgang Raible (eds.), Language typology and
language universals: An international handbook, 1456–1470. Berlin: Walter de
Gruyter.
Davey, Richard J. 2016. Coastal Dhofari Arabic: A sketch grammar. Leiden: Brill.
Davies, Humphrey. 1981. 17th-century Egyptian Arabic: A profile of the colloquial
material in Yūsuf al-Širbīnī’s Hazz al-quhūf fī šarḥ qaṣīd Abī Ṣādūf. Berkeley:
University of California. (Doctoral dissertation).
Elias, Elias & Edward Elias. 1981. Egyptian Arabic manual for self study. 19th edn.
Cairo: Elias Modern Publishing House.
Epps, Patience. 2005. Areal diffusion and the development of evidentiality:
Evidence from Hup. Studies in Language 29(3). 617–650.
Grosjean, François. 2001. The bilingual’s language modes. In Janet Nicol (ed.), One
mind, two languages: Bilingual language processing, 1–22. Malden: Blackwell.
Haase, Martin. 1992. Sprachkontakt und Sprachwandel im Baskenland: Die Einflüsse des Gaskognischen und Französischen auf das Baskische. Hamburg:
Buske.
Hansen, Maj-Britt Mosegaard & Jacqueline Visconti. 2009. On the diachrony of
“reinforced” negation in French and Italian. In Corinne Rossari, Claudia Ricci &
Adriana Spiridon (eds.), Grammaticalization and pragmatics: Facts, approaches,
theoretical issues, 137–171. Bingley: Emerald Group.
Haspelmath, Martin. 1998. Does grammaticalization need reanalysis? Studies in
Language 22(2). 271–287.
Heath, Jeffrey. 2003. Hassaniya Arabic (Mali): Poetic and ethnographic texts.
Wiesbaden: Harrassowitz.
Heine, Bernd. 2004. Grammaticalization. In Brian Joseph & Richard Janda (eds.),
The handbook of historical linguistics, 575–601. Oxford: Wiley-Blackwell.
Heine, Bernd & Tania Kuteva. 2002. World lexicon of grammaticalization.
Cambridge: Cambridge University Press.
620
28 Contact-induced grammaticalization between Arabic dialects
Heine, Bernd & Tania Kuteva. 2003. On contact-induced grammaticalization.
Studies in Language 27(3). 529–572.
Heine, Bernd & Tania Kuteva. 2005. Language contact and grammatical change.
Cambridge: Cambridge University Press.
Hopper, Paul & Elizabeth Traugott. 2003. Grammaticalization. 2nd edn.
Cambridge: Cambridge University Press.
Il-Hazmy, Alayan Mohammed. 1975. A critical and comparative study of the spoken
dialect of the Ḥarb tribe in Saudi Arabia. Leeds: University of Leeds. (Doctoral
dissertation).
Jastrow, Otto. 1979. Zur arabischen Mundart von Mossul. Zeitschrift für Arabische
Linguistik 2. 76–99.
Law, Daniel. 2014. Language contact, inherited similarity and social difference: The
story of linguistic interaction in the Maya Lowlands. Amsterdam, Philadelphia:
John Benjamins.
Le Page, R. B. & Andrée Tabouret-Keller. 1985. Acts of identity: Creole-based
approaches to language and ethnicity. Cambridge: Cambridge University Press.
Leddy-Cecere, Thomas. 2018. Contact-induced grammaticalization as an impetus
for Arabic dialect development. Austin: University of Texas. (Doctoral
dissertation).
Lehmann, Christian. 1985. Grammaticalization: Synchronic variation and diachronic change. Lingua e Stile 20. 303–318.
Lehmann, Christian. 2015. Thoughts on grammaticalization. Berlin: Language
Science Press. DOI:10.17169/langsci.b88.98
Lentin, Jérôme. 2011. Damascus Arabic. In Lutz Edzard & Rudolf de Jong (eds.),
Encyclopedia of Arabic language and linguistics, online edn. Leiden: Brill.
Lewin, Bernhard. 1969. Notes on Cabali: The Arabic dialect spoken by the Alawis
of “Jebel Ansariye”. Stockholm: Almqvist & Wiksell.
Lydon, Ghislaine. 2009. On trans-Saharan trails: Islamic law, trade networks,
and cross-cultural exchange in nineteenth-century western Africa. Cambridge:
Cambridge University Press.
Mahdi, Qasim R. 1985. The spoken Arabic of Baṣra, Iraq: A descriptive study of
phonology, morphology and syntax. Exeter: University of Exeter. (Doctoral
dissertation).
Matras, Yaron. 2009. Language contact. Cambridge: Cambridge University Press.
Matras, Yaron. 2011. Grammaticalisation and language contact. In Bernd Heine &
Heiko Narrog (eds.), Handbook of grammaticalisation, 279–290. Oxford: Oxford
University Press.
621
Thomas Leddy-Cecere
Miller, Catherine, Enam Al-Wer, Dominique Caubet & Janet C. E. Watson (eds.).
2007. Arabic in the city: Issues in dialect contact and language variation. London:
Routledge.
Norde, Muriel. 2009. Degrammaticalization. Oxford: Oxford University Press.
Power, Timothy. 2012. The Red Sea from Byzantium to the Caliphate, AD 500–1000.
Cairo: The American University in Cairo Press.
Procházka, Stephan. 2011. Cilician Arabic. In Lutz Edzard & Rudolf de Jong (eds.),
Encyclopedia of Arabic language and linguistics, online edn. Leiden: Brill.
Qafisheh, Hamdi. 1977. A short reference grammar of Gulf Arabic. Tucson, Arizona:
University of Arizona Press.
Ramat, Paolo. 2011. Adverbial grammaticalization. In Bernd Heine & Heiko
Narrog (eds.), The Oxford handbook of grammaticalization, 502–510. Oxford:
Oxford University Press.
Reinhardt, Carl. 1894. Ein arabischer Dialekt gesprochen in ‘Omān und Zanzibar.
Stuttgard & Berlin: W. Spemann.
Rosenhouse, Judith. 2011. Jerusalem Arabic. In Lutz Edzard & Rudolf de Jong
(eds.), Encyclopedia of Arabic language and linguistics, online edn. Leiden: Brill.
Ross, Malcolm. 2007. Calquing and metatypy. Journal of Language Contact 1(1).
116–143.
Rubin, Aaron D. 2005. Studies in Semitic grammaticalization. Winona Lake, IN:
Eisenbrauns.
Saada, Lucienne. 1984. Eléments de description du parler arabe de Tozeur (Tunisie):
Phonologie, morphologie, syntaxe. Paris: Librairie Orientaliste Paul Geuthner.
Sánchez, Pablo. 2014. El árabe vernáculo de Marrakech: Análisis lingüístico de un
corpus representativo. Zaragoza: Prensas de la Universidad de Zaragoza.
Schroepfer, Jason. 2019. Notes on Aswan Arabic. Unpublished manuscript,
University of Texas.
Silverstein, Michael. 2003. Indexical order and the dialectics of sociolinguistic life.
Language and Communication 23. 193–229.
Stewart, Devin. 1998. Clitic reduction in the formation of modal prefixes in the
post-classical Arabic dialects and Classical Arabic “sa-/sawfa”. Arabica 45(1).
104–128.
Taine-Cheikh, Catherine. 2011. Ḥassānīyya Arabic. In Lutz Edzard & Rudolf de
Jong (eds.), Encyclopedia of Arabic language and linguistics, online edn. Leiden:
Brill.
Talmoudi, Fathi. 1980. The Arabic dialect of Sūsa (Tunisia). Göteborg: Acta
Universitatis Gothoburgensis.
Van Coetsem, Frans. 1988. Loan phonology and the two transfer types in language
contact. Dordrecht: Foris.
622
28 Contact-induced grammaticalization between Arabic dialects
Van Coetsem, Frans. 2000. A general and unified theory of the transmission process
in language contact. Heidelberg: Winter.
Vanhove, Martine. 1993. La langue maltaise: Etudes syntaxiques d’un dialecte arabe
«périphérique». Wiesbaden: Harrassowitz.
Versteegh, Kees. 2014. The Arabic language. 2nd edn. Edinburgh: Edinburgh
University Press.
Vollers, Karl. 1895. The modern Egyptian dialect of Arabic: A grammar with exercises, reading lessons and glossaries. Translated by Francis Burkitt. Cambridge:
Cambridge University Press.
Watson, Janet C. E. 1993. A syntax of Ṣanʿānī Arabic. Wiesbaden: Harrassowitz.
Wolfram, Walt & Natalie Schilling-Estes. 1998. American English: Dialects and
variation. Malden: Blackwell.
623
Chapter 29
Contact and calquing
Stefano Manfredi
CNRS, SeDyL
The notion of calquing refers to the transfer of semantic and syntactic patterns deprived of morphophonological matter. By providing examples of lexical and grammatical calques in a number of Arabic dialects and Arabic-based contact languages,
this chapter identifies ways to relate the process of calquing to Van Coetsem’s psycholinguistic principle of language dominance.
1 Introduction
In its simplest definition, calquing is a type of contact-induced change in which
a word or sentence structure is transferred without actual morphemes (Thomason 2001: 260). Calques are sometimes called loan translations, as they typically
represent a word-by-word (or morpheme-by-morpheme) translation of a lexeme
or a sentence from another language. Heath (1984: 367) labels this process “pattern transfer” and distinguishes it from “matter borrowing” which is instead
linked to the integration of morphophonological material. Ross (2007), for his
part, points out that that calquing can also have important grammatical effects,
and he considers it a necessary precondition for contact-induced morphosyntactic restructuring (what Ross calls “metatypy”).
Broadly speaking, we can distinguish two types of calquing: lexical calquing,
which entails the transfer of semantic properties of lexical items, and grammatical calquing, which instead implies the transfer of the functional properties of
morphemes and syntactic constructions. Using Ross’s words (2007: 126), lexical
calquing consists of remodelling lexical “ways of saying things”, whereas grammatical calquing consists of remodelling grammatical “ways of saying things”.
Despite this fundamental difference, lexical and grammatical calquing share a
Stefano Manfredi. 2020. Contact and calquing. In Christopher Lucas & Stefano
Manfredi (eds.), Arabic and contact-induced change, 625–641. Berlin: Language Science Press. DOI:10.5281/zenodo.3744557
Stefano Manfredi
single cause: bilingual speakers’ need to express the same meaning in two languages (Sasse 1992: 32). This also means that everything that expresses meaning
(i.e. morphemes, lexemes, and constructions) can, in principle, be a source of
calquing.
Focusing mainly on the transfer of linguistic matter, Van Coetsem (1988) does
not overtly mention the possibility of transfer of lexical and grammatical meanings through calquing. The present chapter thus aims at relating contact-induced
changes produced by calquing to the principle of language dominance as postulated by Van Coetsem.
2 Contact-induced changes and calquing
2.1 Lexical calquing
According to Haspelmath (2009: 39), a lexical calque is a lexical unit that was
created by an item-by-item translation of the source unit. This type of contactinduced change occurs as bilingual speakers reorganise the lexicon of one of
their languages to match the semantic organisation of the other (Ross 2007: 132).
Adopting the psycholinguistic standpoint of language dominance, Winford (2003:
345) regards lexical calquing as a subtype of lexical borrowing, which is a combination of recipient language (RL) lexemes in imitation of source language (SL)
semantic patterns. In contrast, I will show that, though lexical calquing can easily be triggered by RL-dominant speakers, it can also be a product of imposition via SL agentivity. In order to do this, I will mainly focus on calquing of
compound nouns. A compound noun is here defined as a series of two or more
lexemes, which is semantically conceived as a single unit. Each component of
the compound can function as a lexeme independent from the other(s), and may
show some phonological and/or morphological constraints within the compound
when compared to its isolated syntactic usage (Bauer 2001). Against this backdrop, I will specifically discuss noun–noun compounds, as they represent the
more uniform phenomenon of nominal compounding in the world’s languages
Pepper (forthcoming). As we will see, the transfer of the semantics of compound
nouns does not imply any morphosyntactic change in Arabic, as calqued compounds are typically adjusted to fit RL morphosyntactic patterns.
Generally speaking, lexical calquing through borrowing can occur in indirect
contact situations characterized by a very low degree of bilingualism. This is because RL monolinguals can also be agents of lexical borrowing (Van Coetsem
1988: 10). Typical instances of lexical calquing via RL agentivity are related to the
transfer of the semantic patterns of English compound nouns in modern Arabic
626
29 Contact and calquing
dialects. This kind of transfer is linked to the expansion of the non-core Arabic
lexicon for expressing previously unknown concepts. A prime example is the
English calque lōḥit il-mafatīḥ ‘keyboard’ (lit. ‘the board of keys’) in Egyptian
Arabic (Wilmsen & Woidich 2011: 9). Here, it can be clearly seen that the transfer of the semantic organization of the SL compound noun does not affect the
morphosyntax of the RL, as the word order of the English nominal juxtaposition
is reversed to fit the Arabic construct state.
Lexical calquing can also take place in prolonged contact situations, as testified
by numerous Italian compounds in Maltese. A singular case of mixed calquing
is that of wiċċ tost ‘shameless person’ (lit. ‘tough face’) deriving from the Italian compound faccia tosta ‘shameless person’ (lit. ‘tough face’) (Aquilina 1987).
On the one hand, the first lexical item of the compound presents an Arabic
phonological form while expressing semantic properties associated with the lexeme ‘face’ in Italian. On the other hand, the second lexical item clearly results
from the borrowing of the adjective tosto ‘hard, tough’ retaining both the Italian phonological matter and semantic properties. The mixed nature of this compound brings to the fore the complementary relationship between RL and SL
agentivity and shows that it is not always a trivial matter to distinguish between
imposition and borrowing. However, Maltese also gives evidence of genitive compounds in which both lexical components have an Arabic phonological form coupled with Italian semantic properties. This is the case of the compound nouns
saba’ ta’ sieq ‘toe’ calqued on the Italian dito del piede ‘toe’ (Pepper forthcoming). Such instances of lexical calquing clearly mirror semantic properties of SL
lexemes and they most plausibly result from borrowing via RL agentivity (cf.
Lucas & Ćeplö, this volume).
Ḥassāniyya Arabic, for its part, presents many compound nouns that are traditionally analysed in terms of substratum interference from Zenaga Berber (TaineCheikh 2008; 2012). Also in this case, the transfer of the semantic properties of
the SL does not produce any morphosyntactic change in Arabic, as we can see
in the pairs of examples in (1) and (2).
(1)
a. Ḥassāniyya Arabic (Taine-Cheikh 2008: 126)
kṛaʕ lә-ɣṛab
foot def-crow
‘aquatic herbaceous plant’ (lit. ‘crow’s foot’)
b. Zenaga Berber (Taine-Cheikh 2008: 126)
að̣aʔṛ әn tayyaḷ
foot gen crow
‘aquatic herbaceous plant’ (lit. ‘crow’s foot’)
627
Stefano Manfredi
(2)
a. Ḥassāniyya Arabic (Taine-Cheikh 2008: 126)
sayllāl lә-ʕrāgib
ripper def-ankle.pl
‘honey badger’ (lit. ‘ripper of ankles’)
b. Zenaga Berber (Taine-Cheikh 2008: 126)
amәssäf әn ūržan
ripper gen ankle.pl
‘honey badger’ (lit. ‘ripper of ankles’)
Taine-Cheikh (2008: 126) stresses that it is somewhat difficult to trace back
the origin of these compounds. Accordingly, she speaks of a process of convergence between the two languages, rather than determining the direction of the semantic transfer. However, it should be observed that these compound nouns are
not attested in other spoken varieties of Arabic. Furthermore, since at least the
mid-twentieth century, Berbers in Mauritania have been gradually loosing competence in Zenaga, in favour of Arabic (Taine-Cheikh 2012: 100), while Zenaga
is rarely acquired as second language by Ḥassāniyya Arabic speakers. In such a
context, the most probable agents of contact-induced change were former Berberdominant speakers who gradually shifted to Arabic. Thus, it seems plausible that
the transfer of the semantic properties of Zenaga compounds has been achieved
through imposition, rather than through borrowing.
Nigerian Arabic also shows interesting instances of lexical calquing as a consequence of a longstanding contact with Kanuri, a Nilo-Saharan language widely
spoken in the Lake Chad area. Owens (2015; 2016) gives evidence of the transfer
of the semantic properties of numerous compound nouns including the lexeme
ṛās ‘head’. Similar to the previous instances of compound calquing, the integration of Kanuri semantic patterns does not affect the Arabic morphosyntax, as we
can see in the pairs of examples in (3) and (4).
(3)
628
a. Nigerian Arabic (Owens 2016: 69)
ṛās al-bēt
head def-house
‘roof’ (lit. ‘head of house’)
b. Kanuri (Owens 2016: 69)
kǝla fato-be
head house-gen
‘roof’ (lit. ‘head of house’)
29 Contact and calquing
(4)
a. Nigerian Arabic (Owens 2016: 65)
ṛās al-qalla
head def-corn
‘tassel’ (lit. ‘head of corn’)
b. Kanuri (Owens 2016: 65)
kǝla argǝm-be
head corn-gen
‘tassel’ (lit. ‘head of corn’)
According to Owens (2016: 65), Kanuri–Arabic bilingualism, with Arabic being a minority language, would have been the foremost factor underlying the
transfer of these compound nouns into Nigerian Arabic. He further stresses that
Kanuri is the main source of compound nouns in a number of other minority
languages in the area (e.g. Kotoko, Glayda, and Fulfulde) and that there is little evidence of Kanuri to Arabic shift in the region (Owens 2014: 147). However,
the fact that Kanuri represents the majority language of northeastern Nigeria,
does not shed light on the transfer mechanism lying behind lexical calquing in
Nigerian Arabic. This is because speakers can be linguistically dominant in a socially subordinate language (Winford 2005: 376). In fact, such contact settings
are closely tied to SL agentivity, as the youngest bilingual generations tend to
impose semantic features from their dominant language (i.e. Kanuri) onto the
ancestral language (i.e. Arabic). It is only at a later stage that these innovations
are borrowed by older bilingual speakers who are still dominant in Arabic.
The fact that Nigerian Arabic speakers have gradually developed a high bilingual proficiency in Kanuri is also testified by the transfer of a number of idiomatic
expressions. In this regard, Ross (2007: 122) observes that calquing of meaning
is not only reflected in word compounding, but also in lexical collocations of
idiomatic expressions. These are combinations of lexical items that are semantically idiosyncratic as they have a pairing of form and meaning that cannot be
predicted from the rest of the grammar. The pair of examples in (5) provide evidence of an idiomatic Kanuri calque in Nigerian Arabic.
(5)
a. Nigerian Arabic (Ritt-Benmimoun et al. 2017: 77)
šuqul
šāl
ṛās-i
something carry.prf.3sg.m head-obl.1sg
‘Something distracted me.’ (Lit. ‘Something carried my head.’)
b. Kanuri (Ritt-Benmimoun et al. 2017: 77)
awode
kǝla gō-zǝ-na
something head carry-3sg-prf
‘Something distracted me.’ (Lit. ‘Something carried head.’)
629
Stefano Manfredi
Given that idiomatic expressions are syntactically compositional (i.e. their lexical components behave syntactically as they do in non-idiomatic expressions),
it is not only the meanings expressed by the lexeme ‘head’ which correspond between Nigerian Arabic and Kanuri, but also their idiomatic collocations, which
align between the two languages (Owens 2014: 157). Besides, it is worth noting
that also idiomatic expressions are adjusted to fit RL morphosyntactic patterns.
This is evidenced by the inalienable possession of body parts in Nigerian Arabic
(ṛās-i ‘my head’), which is instead unattested in the SL (kǝla ‘head’). Even if we
cannot exclude the possibility that these kinds of calques are a product of borrowing, it is evident that their integration needs a high proficiency in the SL for
individuating the single idiomatic collocations of lexical items. Furthermore, differently from borrowed calques, imposed idiomatic expressions can significantly
affect the lexical semantics of the RL created by SL-dominant bilinguals, and thus
produce grammatical changes in the long run.
Finally, lexical calquing via SL agentivity can also take place in extreme contact situations such as creolization. For instance, Juba Arabic, the Arabic-based
pidgincreole spoken in South Sudan, shows numerous calques in which Arabicderived lexemes are compounded according to the semantic patterns of Bari, the
main substrate language of Juba Arabic (Nakao 2012; Manfredi 2017: 50; Avram,
this volume) As we can see in (6) and (7), the word order in Juba Arabic compounds follows the order of Bari compounds. However, this cannot be seen as
an innovative morphosyntactic development, as the possessed–possessor order
matches also with the Arabic lexifier.
(6)
a. Juba Arabic (Nakao 2012: 136)
éna ta séjera
eye gen tree
‘fruit’ (lit. ‘eye of tree’)
b. Bari (Nakao 2012: 136)
koŋe lo-ködini
eye gen-tree
‘fruit’ (lit. ‘eye of tree’)
(7)
a. Juba Arabic (Nakao 2012: 137)
ída ta fil
hand gen elephant
‘trunk’ (lit. ‘hand of elephant’)
b. Bari (Nakao 2012: 137)
könin lo-tome
hand gen-elephant
‘trunk’ (lit. ‘hand of elephant’)
Given that the asymmetric contact situation leading to creole formation limits
access to the superstrate language (i.e. Sudanese Arabic), the semantic patterns
of substrate languages (i.e. Bari) can be easily carried over into the creole in ways
peculiar to imposition via SL agentivity.
630
29 Contact and calquing
All things considered, unlike lexical borrowing, lexical calquing allows for a
semantic overlapping of RL and SL lexical entries and it can also produce important structural changes.
2.2 Grammatical calquing
Grammatical calquing brings about a match between the grammatical categories
of two languages and the memberships of these categories (Ross 2007: 132). Heine
& Kuteva (2005) suggest that the grammatical changes induced by calquing can
be better analysed in terms of contact-induced grammaticalization (cf. LeddyCecere, this volume). In fact, the calquing of the semantic properties of lexical
and grammatical items may lead to the grammaticalization of innovative syntactic structures in the RL matching with those of the SL. From the traditional
sociohistorical perspective of contact-induced change (Thomason & Kaufman
1988), grammatical calquing is basically seen as a product of language shift. In
contrast, Ross (2007: 131) argues that grammatical calques can widely occur in
situations of language maintenance. Actually, the different grammatical outputs
of calquing mainly depend on the way in which they are transferred from the SL
into the RL and, by extension, on different kinds and degrees of bilingualism.
For the purposes of this chapter, I distinguish between three different types of
grammatical calquing:
1. Calquing of polyfunctionality of lexical items without syntactic change;
2. Calquing of polyfunctionality of grammatical items leading to syntactic
change;
3. Narrow syntactic calquing (without calquing of polyfunctionality of lexical/grammatical items).
Being lexical in nature, the first of these three types of grammatical calquing
can be triggered by both imposition via SL agentivity and borrowing via RL agentivity, whereas the two latter types are likely to result only from imposition via
SL agentivity.
Calquing of polyfunctionality patterns of lexical items is by far the most common type of grammatical calquing, and it can be exemplified by the comparison
of reflexive anaphors in different Arabic dialects. As is well known, Classical and
Standard Arabic express a reflexive meaning either by means of agent-oriented
derived verbs lacking an overtly expressed patient (e.g. istaḥamma ‘he washed
himself’) or by anaphoric constructions in which the syntagm nafs-pro.poss
631
Stefano Manfredi
‘soul-pro.poss’ marks coreferentiality between the agent and the patient of the
predicate (e.g. qatala nafsa-hu ‘he killed himself’). Nevertheless, as a result of
contact with different languages, a number of modern Arabic dialects have grammaticalized other lexical sources for expressing a reflexive meaning. Western
Maghrebi dialects are a case in point. As we can see in (8)–(9), both Moroccan and
Ḥassāniyya Arabic have grammaticalized the nominal syntagm ṛāṣ=pro.poss
‘head-pro.poss’ as default reflexive anaphor.
(8)
Moroccan Arabic (D. Caubet, personal communication)
qtәl
ṛās-o
kill.prf.3sg.m head-3sg.m
‘He killed himself.’ (Lit. ‘He killed his head.’)
(9)
Ḥassāniyya Arabic (Taine-Cheikh 2008: 16)
ktәl
ṛāṣ-u
kill.prf.3sg.m head-3sg.m
‘He killed himself.’ (Lit. ‘He killed his head.’)
This reflexive use of the lexeme ‘head’ has generally been interpreted as substrate interference from Berber languages (El Aissati 2011: 197), in which the same
grammaticalization path is attested, as shown in the following examples from
Tarifiyt and Zenaga:
(10) Tarifiyt Berber (Kossmann 2000: 95)
yәtšaθ
iḫәf nnәs
beat.prf.3sg.m head poss.3sg.m
‘He beats himself.’ (Lit. ‘He beats his head.’)
(11) Zenaga Berber (Taine-Cheikh 2008: 126)
yәʔna
iʔf-әn-š
kill.prf.3sg.m head-gen-3sg.m
‘He killed himself.’ (Lit. ‘He killed his head.’)
Lexemes for ‘head’ are the second most common source of grammaticalization of reflexive anaphors worldwide (König et al. 2013) and its occurrence is
particularly common in West Africa (Heine 2011: 50). In this scenario, it should
be stressed that the reflexive function of the lexeme ‘head’ is an innovative feature of both Arabic and Berber varieties of northwestern Africa. Other Berber
languages typically use the reflexive anaphor iman-poss ‘soul-poss’, as we can
see in the following example from Kabyle.
632
29 Contact and calquing
(12)
Kabyle (Mettouchi 2012)
n-səlk-dd
iman-ntəɣ
1pl-spare.prf-prox soul.abs.sg.m-poss.1pl.f
‘We saved ourselves.’
In addition, the known Arabic–Berber contact situation, in which second language acquirers of Berber only played a marginal role in triggering contactinduced change in Arabic, suggests that the contact induced grammaticalization
of ‘head’ in westernmost Arabic dialects resulted from an imposition enacted by
former Berber-dominant speakers.
A similar instance of calquing in the domain of anaphoric reflexive constructions is found in Kordofanian Baggara Arabic, a Western Sudanic dialect spoken
in the Nuba Mountains area, in central Sudan. In this case, the source of the
reflexive anaphor is the lexeme for ‘neck’, as we can see in (13).
(13)
Kordofanian Baggara Arabic (Manfredi 2010: 176)
abrahīm gaṣṣa
ragabt-a
Ibrahim cut.prf.3sg.m neck-3sg.m
‘Ibrahim cut himself.’ (Lit. ‘Ibrahim cut his neck.’)
Different from ‘head’, the grammaticalization of ‘neck’ as a reflexive anaphor
is quite rare in Africa (Heine 2011: 50), but it is attested in a number of NigerKordofanian languages spoken in the same region. Such is the case of Tagoi (14)
and Koalib (15).
(14)
Tagoi (Alamin 2015: 26)
ínní
t-áɡám t-ùrúŋ
nc-neck nc-poss.3 kill.prf.3
‘He killed himself.’ (Lit. ‘He killed his neck.’)
(15)
Koalib (N. Quint, personal communication)
ɛ̀ ɽnyɛ́ r-ɔ́kwɽɔ̀ r-ùŋwún
kill nc-neck nc-poss.3
‘to kill oneself’ (lit. ‘to kill one’s neck’)
Similarly to the situation described with reference to western Maghrebi dialects, Arabic-speaking groups in the Nuba Mountains have hardly developed any
bilingual competence in local Niger-Kordofanian languages. Therefore, it seems
likely that the calquing of the polyfunctionality patterns of ‘neck’ has been imposed by Arabized populations who were dominant in the SL.
633
Stefano Manfredi
Maltese also provides remarkable examples of calquing of polyfunctionality
of lexical items. This is particularly evident in the domain of auxiliary verbs
(Vanhove 1993; Vanhove et al. 2009). A well-known example is that of the lexical verb ġie ‘come’ used as an auxiliary for expressing a dynamic passive (16) in
the same way as Italian (17).
(16) Maltese (Borg & Azzopardi-Alexander 1997: 214)
it-tabib
ġie
afdat
bi-l-każ
def-doctor come.prf.3sg.m trust.ptcp.pass with-def-case
‘The doctor was entrusted with the case.’ (Lit. ‘The doctor came entrusted
with the case.’)
(17) Italian (own knowledge)
creduto
non venne
neg come.pst.3sg.m trust.ptcp.pst
‘He was not trusted.’ (Lit. ‘He did not come trusted.’)
Even if imposition played a role in the emergence of Maltese (Lucas & Ćeplö,
this volume), it is generally accepted that intertwined languages emerge mainly
from a widespread process of borrowing in Van Coetsem’s terminology (Winford
2005: 397; Manfredi 2018). This suggests that, unlike the aforementioned grammaticalization of reflexive anaphors in Arabic dialects, the calquing of polyfunctionality of lexical verb ‘come’ in Maltese was most likely triggered by agentivity
of RL dominant speakers.
Regardless of the different contact situations, what unites all the previous instances of grammatical calquing is the fact that the transfer of patterns of grammaticalization did not produce any syntactic change in Arabic. In contrast to
the above, the calquing of polyfunctionality of grammatical items can be accompanied by important typological changes. This is the case of the grammaticalization of prototypical passive constructions in Juba Arabic (Manfredi 2017: 92;
2018: 415). As we can see in (18), the South Sudanese pidgincreole presents an
innovative passive construction in which the patient occupies the syntactic slot
of a preverbal subject, whereas the oblique-marked agent is introduced by the
comitative preposition ma- ‘with’.
(18) Juba Arabic (Manfredi 2017: 86)
bab de
kasurú
ma-jón
door prox.sg break.pass with-John
‘This door has been broken by John.’ (Lit. ‘This door has been broken
with John.’)
634
29 Contact and calquing
Interestingly, this prototypical passive construction is not attested in the lexifier language of Juba Arabic (i.e. Sudanese Arabic), which instead makes use of
impersonal passive constructions with a default 3pl.m subject.
(19) Sudanese Arabic (own knowledge)
kassaru-hu
break.prf.3pl.m-3sg.m
‘It got broken.’ (Lit. ‘They have broken it.’)
Indeed, the grammaticalization of this complex syntactic structure is the result of the calquing of the functional properties associated with the comitative
preposition of the main substrate language, Bari. Bari presents the same kind
of prototypical passive construction in which an oblique-marked agent is introduced by the preposition ko- ‘with’.
(20)
Bari (Owen 1909: 65)
niena wuret a-wur-ö
ko-nan
prox.sg book 3sg.pst-write-pass with-1sg
‘This book has been written by me.’ (Lit. ‘This book has been written with
me.’)
If we assume that the emergence of creole languages is always induced by the
disruption of the transmission of the lexifier language (Comrie 2011), we can conclude that Bari speakers have imposed the semantics of their dominant language
on a grammatical item derived from Arabic, and thus induced profound changes
in the word order of the creole when compared to its lexifier language.1 In light
of the above, the contact dynamics lying behind the calquing of polyfunctionality of grammatical items are quite restrictive as they are most likely a product of
imposition via SL agentivity.
The third kind of grammatical calquing is linked to the transfer of syntactic patterns without transfer of polyfunctionality of either lexical or grammatical items.
This narrow type of syntactic calquing can be exemplified by possessor doubling
in Central Asian Arabic (Ratcliffe 2005). Clitic doubling is a construction in which
a clitic co-occurs with a full nominal phrase in argument position, forming a discontinuous constituent with it. Various forms of clitic doubling have arisen in a
number of Arabic varieties as a result of contact with different substrate/adstrate
languages (Souag 2017). In regard to possessive constructions, Arabic typically
presents a possessed–possessor order. In contrast, Central Asian Arabic (21) gives
evidence of the opposite order with obligatory possessor doubling in the same
way as Tajik (22).
1
This kind of syntactic change accompanied by the calquing of semantic properties of substrate
items in creole languages is traditionally labelled “relexification” (Lefebvre 1998).
635
Stefano Manfredi
(21)
Central Asian Arabic (Ratcliffe 2005; Souag 2017: 56)
amīr wald-u
prince son-3sg.m
‘the prince’s son’
(22)
Tajik (Souag 2017: 56)
buḫoro universitet-aš
Bukhara university-3sg
‘Bukhara University’
Souag (2017: 157) states that double possessor constructions in Central Asian
Arabic are instances of grammatical calquing, accommodated through the reinterpretation of pre-existing topicalized constructions. This means that, unlike
the syntactic changes induced by the calquing of polyfunctionality of morphemes, the emergence of double possessor constructions in Bukhara Arabic
would have been favoured by a formal congruence between SL and RL syntactic
structures. As such, this instance of contact-induced morphosyntactic restructuring (metatypy) does not derive from a direct copying of a double possessor
construction. Rather, it consists in speakers expressing a possessive meaning in
Arabic by using a construction which they equate with the construction in adstratal languages (Ross 2007: 128). If we consider that the youngest speakers of
Central Asian Arabic are gradually losing competence in their ancestral language
in favour of socially dominant languages (Chikovani 2005: 128), it is plausible to
assume that this kind of syntactic restructuring can only be a result of imposition via SL agentivity. Still, given our limited diachronic knowledge, we cannot exclude the hypothesis of an early process of borrowing enacted by former
Arabic-dominant speakers.
3 Conclusion
Van Coetsem (1988: 20) suggests that the variable outcomes of language contact
are primarily a reflex of the “stability gradient” of language, which induces speakers to preserve the domains of their dominant language that are less affected
by change. As lexicon is the most unstable linguistic domain, it is likely to be
transferred via RL agentivity. In contrast, morphosyntax and phonology are considered to be relatively stable domains and they are expected to be transferred
only via SL agentivity. Against this background, it is unclear how the transfer of
semantic features deprived of morphophonological matter should be understood
in relation to the linguistic dominance of the agents of contact-induced change.
636
29 Contact and calquing
If we look at the previously analysed instances of lexical calquing (§2.1), it
is evident that the transfer of the semantic features of nominal compounds can
take place within speech communities with a very low degree of bilingualism, as
in the case of Egyptian-Arabic-dominant speakers borrowing the semantics of
English compounds. But it is also true that compound calquing can be a product
of imposition resulting from ongoing language shift or pidginization, and the
transfer of semantic features of single lexical items within idiomatic expressions
always requires a widespread proficiency in the SL, as in the case of Arabic–
Kanuri bilingualism in northern Nigeria.
As far as grammatical calquing is concerned (§2.2), I have shown that calquing
of the polyfunctionality of lexical items can be triggered either by imposition, as
in the case of substrate interference in Ḥassāniyya and Baggara dialects, or by
borrowing in the emergence of intertwined languages such as Maltese. Calquing
of polyfunctionality of grammatical items, for its part, requires a higher degree
of linguistic abstraction for the identification of a functional overlap between
morphemes. Accordingly, this type of transfer will typically occur via imposition
by SL-dominant speakers in deep contact situations such as creolization. In the
same manner, narrow syntactic calquing requires high bilingual proficiency, as it
necessitates the recognition of some formal congruence between the SL and the
RL, as shown by the emergence of possessor doubling in Central Asian Arabic.
To stay somewhat in line with the stability gradient principle, we could argue that, in absence of the transfer of linguistic matter, the semantic properties
of morphemes and syntactic constructions are more stable than those of lexical
items. However, such a generalization would be misleading without an in-depth
knowledge of the sociolinguistic circumstances underlying a specific instance of
second language acquisition (i.e. symmetric bilingualism, asymmetric bilingualism, multilingualism, pidginization/creolization). Thus, it becomes evident that
the recognition of different patterns of bilingualism within the same community
remains the only way to identify the transfer type at play in a given contact
situation, regardless of its different structural outputs.
Drawing on the available literature, this chapter has surveyed only a few instances of calquing induced by contact between Arabic and other languages. This
is mainly because we lack information about calquing in dialect contact situations. Indeed, it is regrettable that studies dealing with dialect contact and new
dialect formation are still exclusively focused on the diffusion of few lexical and
morphophonological features, while disregarding the transfer of semantic and
syntactic patterns. Fine-grained analyses of semantic changes induced by dialect contact thus remain a major desideratum for the development an aggregate
variationist Arabic dialectology.
637
Stefano Manfredi
Further reading
) Keesing (1988) adopts the notion of calquing and describes the transfer of semantic properties of Oceanic morphemes in Melanesian Pidgin.
) Meyerhoff (2009) focuses on the notions of replication, transfer, and calquing,
thereby strengthening connections between variationist sociolinguistics and
contact linguistics.
) Zuckermann (2009) provides numerous instances of calquing in Modern Hebrew and analyses them in the light of the Congruence Principle.
Abbreviations
1, 2, 3
def
f
gen
m
nc
obl
pass
pst
pl
1st, 2nd, 3rd person
definite article
feminine
genitive
masculine
noun class
oblique
passive
past
plural
poss
prg
prf
pro
prox
RL
refl
sg
SL
abs
possessive pronoun
pragmatic marker
perfect (prefix conjugation)
pronoun
proximal
recipient language
reflexive
singular
source language
absolute state
References
Alamin, Suzan. 2015. The Tagoi pronominal system. Occasional Papers in the
Study of Sudanese Languages 11. 17–30.
Aquilina, Joseph. 1987. Maltese–English dictionary. Malta: Midsea Books.
Bauer, Laurie. 2001. Compounding. In Martin Haspelmath & Wolfgang
Oesterreicher (eds.), Language typology and universals, 695–707. Berlin: De
Gruyter.
Borg, Albert & Marie Azzopardi-Alexander. 1997. Maltese. London: Routledge.
Chikovani, Guram. 2005. Linguistic contacts in Central Asia. In Éva Ágnes Csató,
Bo Isaksson & Carina Jahani (eds.), Linguistic convergence and areal diffusion:
Case studies from Iranian, Semitic and Turkic, 127–36. London: Routledge.
Comrie, Bernard. 2011. Creoles and language typology. In Claire Lefebvre (ed.),
Creoles, their substrates and language typology, 599–611. Amsterdam: John
Benjamins.
638
29 Contact and calquing
El Aissati, Abderahman. 2011. Berber loanwords. In Lutz Edzard & Rudolf de Jong
(eds.), Encyclopedia of Arabic language and linguistics, online edn. Leiden: Brill.
Haspelmath, Martin. 2009. Lexical borrowing: Concepts and issues. In Martin
Haspelmath & Uri Tadmor (eds.), Loanwords in the world’s languages: A
comparative handbook, 35–54. Berlin: De Gruyter.
Heath, Jeffrey. 1984. Language contact and language change. Annual Review of
Anthropology 13. 367–84.
Heine, Bernd. 2011. Areas of grammaticalization and geographical typology. In
Osamu Hieda, Christa König & Hirosi Nakagawa (eds.), Geographical typology
and linguistic areas with special reference to Africa, 41–66. Amsterdam: John
Benjamins.
Heine, Bernd & Tania Kuteva. 2005. Language contact and grammatical change.
Cambridge: Cambridge University Press.
Keesing, Roger. 1988. Melanesian Pidgin and the Oceanic substrate. Stanford, CA:
Stanford University Press.
Kossmann, Maarten. 2000. Esquisse grammaticale du rifain oriental. Leuven:
Peeters.
König, Ekkehard, Peter Siemund & Stephan Töpper. 2013. Intensifiers and
reflexive pronouns. In Matthew Dryer & Martin Haspelmath (eds.), The
world atlas of language structures online. Leipzig: Max Planck Institute for
Evolutionary Anthropology.
Lefebvre, Claire. 1998. Creole genesis and the acquisition of grammar. Cambridge:
Cambridge University Press.
Manfredi, Stefano. 2010. A grammatical description of Kordofanian Baggara
Arabic. Naples: Università degli Studi di Napoli “L’Orientale”. (Doctoral
dissertation).
Manfredi, Stefano. 2017. Arabi Juba: Un pidgin-créole du Soudan du Sud. Leuven:
Peeters.
Manfredi, Stefano. 2018. Arabic as a contact language. In Reem Bassiouney &
Elabbas Benmamoun (eds.), The Routledge handbook of Arabic linguistics, 407–
20. London: Routledge.
Mettouchi, Amina. 2012. Kabyle corpus. Corpus recorded, transcribed
and annotated by A. Mettouchi. In Amina Mettouchi & Christian
Chanard (eds.), The CorpAfroAs corpus of spoken AfroAsiatic languages.
doi.org/10.1075/scl.68.website.
Meyerhoff, Miriam. 2009. Replication, transfer, and calquing: Using variation as a
tool in the study of language contact. Language Variation and Change 21. 297–
317.
639
Stefano Manfredi
Nakao, Shuichiro. 2012. Revising the substratal/adstratal influence on Arabic
creoles. In Osamu Hieda (ed.), Challenges in Nilotic linguistics: Phonology,
morphology, and syntax, 127–49. Tokyo: ILCAA.
Owen, Roger. 1909. Bari grammar and vocabulary. London: Bumpus.
Owens, Jonathan. 2014. Many heads are better than one: The spread of motivated
opacity via contact. Linguistics 52(1). 125–165.
Owens, Jonathan. 2015. Idioms, polysemy, context: A model based on Nigerian
Arabic. Anthropological Linguistics 57(1). 46–98.
Owens, Jonathan. 2016. The lexical nature of idioms. Language Sciences 57. 49–
69.
Pepper, Stephen. Forthcoming. The typology of binominal lexemes: Topic maps,
railways and the nature of assocative thought. Oslo: University of Oslo.
(Doctoral dissertation).
Ratcliffe, Robert R. 2005. Bukhara Arabic: A metatypized dialect of Arabic in
Central Asia. In Éva Ágnes Csató, Bo Isaksson & Carina Jahani (eds.), Linguistic
convergence and areal diffusion: Case studies from Iranian, Semitic and Turkic,
141–51. London: Routledge.
Ritt-Benmimoun, Veronika, Smaranda Grigore, Jocelyne Owens & Jonathan
Owens. 2017. Three idioms, three dialects, one history: Egyptian, Nigerian
and Tunisian Arabic. In Veronika Ritt-Benmimoun (ed.), Tunisian and Libyan
Arabic dialects: Common trends – recent developments – diachronic aspects, 43–
84. Zaragoza: Prensas de la Universidad de Zaragoza.
Ross, Malcolm. 2007. Calquing and metatypy. Journal of Language Contact 1(1).
116–143.
Sasse, Hans-Jürgen. 1992. Language decay and contact-induced change:
Similarities and differences. In Matthias Brenzinger (ed.), Language death:
Factual and theoretical explorations with special reference to East Africa, 59–
80. Berlin: Mouton de Gruyter.
Souag, Lameen. 2017. Clitic doubling and language contact in Arabic. Zeitschrift
für Arabische Linguistik 66. 45–70.
Taine-Cheikh, Catherine. 2008. Arabe(s) et berbère en contact: Le cas
mauritanien. In Mena Lafkioui & Vermondo Brugnatelli (eds.), Berber in
contact: Linguistic and sociolinguistic perspectives, 113–38. Cologne: Rüdiger
Köppe.
Taine-Cheikh, Catherine. 2012. Arabe(s) et berbère en Mauritanie. In Gunvor
Mejdell & Lutz Edzard (eds.), High vs. low and mixed varieties: Status, norms
and functions across time and languages, 88–108. Wiesbaden: Harrassowitz.
Thomason, Sarah G. 2001. Language contact: An introduction. Edinburgh: Edinburgh University Press.
640
29 Contact and calquing
Thomason, Sarah G. & Terrence Kaufman. 1988. Language contact, creolization
and genetic linguistics. Berkeley: University of California Press.
Van Coetsem, Frans. 1988. Loan phonology and the two transfer types in language
contact. Dordrecht: Foris.
Vanhove, Martine. 1993. La langue maltaise: Etudes syntaxiques d’un dialecte arabe
«périphérique». Wiesbaden: Harrassowitz.
Vanhove, Martine, Catherine Miller & Dominique Caubet. 2009. The grammaticalisation of modal auxiliaries in Maltese and Arabic vernaculars of the Mediterranean area. In Björn Hansen & Ferdinand de Haan (eds.), Modals in the
languages of Europe: A reference work, 325–362. Berlin: Mouton de Gruyter.
Wilmsen, David & Manfred Woidich. 2011. Egypt. In Lutz Edzard & Rudolf de
Jong (eds.), Encyclopedia of Arabic language and linguistics, online edn. Leiden:
Brill.
Winford, Donald. 2003. An introduction to contact linguistics. Malden: Blackwell.
Winford, Donald. 2005. Contact-induced change: Classification and processes.
Diachronica 22(2). 373–427.
Zuckermann, Ghil’ad. 2009. Hybridity versus revivability: Multiple causation,
forms and patterns. Journal of Language Contact VARIA 2. 40–67.
641
Chapter 30
Contact and the expression of negation
Christopher Lucas
SOAS University of London
This chapter presents an overview of developments in the expression of negation
in Arabic and a number of its contact languages, focusing on clausal negation, with
some remarks also on indefinites in the scope of negation. For most of the developments discussed in this chapter, it is not possible to say for certain that they are
contact-induced. But evidence is presented which, cumulatively, points to widespread contact-induced change in this domain being the most plausible interpretation of the data.
1 Overview of concepts and terminology
1.1 Jespersen’s cycle
Historical developments in the expression of negation have been the subject of
increasing interest in the past few decades, with particular attention given to
the fact that these developments typically give the appearance of being cyclical
in nature. We can date the beginning of this sustained interest to Dahl’s (1979)
typological survey of negation patterns in the world’s languages, in which he
coined the term Jespersen’s cycle1 for what is by now the best-known set of
developments in this domain: the replacement of an original negative morpheme
with a newly grammaticalized alternative, after a period in which the two may cooccur, prototypically resulting in a word-order shift from preverbal to postverbal
negation. The best-known examples of Jespersen’s cycle (both supplied, among
1
The name was chosen in recognition of the early identification of this phenomenon by the
Danish linguist Otto Jespersen in a (1917) article, though others did identify the same set of
changes earlier: Meillet (1912), for example, but also, significantly for the present work, Gardiner (1904), who observed a parallel set of changes in Coptic and Arabic as well as French (cf.
van der Auwera 2009).
Christopher Lucas. 2020. Contact and the expression of negation. In Christopher
Lucas & Stefano Manfredi (eds.), Arabic and contact-induced change, 643–667. Berlin:
Language Science Press. DOI:10.5281/zenodo.3744559
Christopher Lucas
others, by Jespersen himself in his 1917 work) come from the history of English
(1), and French (2).
(1)
English (Jespersen 1917: 9)
a. Stage I – Old English
ic ne secge
1sg neg say.prs.1sg
‘I do not say.’
b. Stage II – Middle English
I ne seye
not.
1sg neg say.prs.1sg neg
‘I do not say.’
c. Stage III – Early Modern English
I say not.
(2) French (Jespersen 1917: 7)
a. Stage I – Old French
jeo ne di
1sg neg say.prs.1sg
‘I do not say.’
b. Stage II – contemporary written French
Je ne dis
pas.
1sg neg say.prs.1sg neg
‘I do not say.’
c. Stage III – contemporary colloquial French
Je dis
pas.
1sg say.prs.1sg neg
‘I do not say.’
More recently, Jespersen’s cycle has come to be the subject of intensive investigation, especially in the languages of Europe (e.g. Bernini & Ramat 1992;
1996; Willis et al. 2013; Breitbarth et al. 2020), but also beyond (e.g. Lucas 2007;
2009; 2013; Lucas & Lash 2010; Devos & van der Auwera 2013; van der Auwera
& Vossen 2015; 2016; 2017), with a picture emerging of a marked propensity for
instances of Jespersen’s cycle to be areally distributed, as we will see below in
the discussion of Jespersen’s cycle in Arabic and its contact languages (§2).
644
30 Contact and the expression of negation
While Jespersen’s cycle is the best known, best studied, and perhaps crosslinguistically most frequently occurring set of changes in the expression of negation, two other important types of changes must also be mentioned here: Croft’s
cycle, and changes to indefinites in the scope of negation.
1.2 Croft’s cycle
In a typologically-oriented (1991) article, Croft reconstructs from synchronic descriptions of a range of languages a recurring set of cyclical changes in the expression of negation. Unlike Jespersen’s cycle, in which the commonest sources of
new negators are nominal elements expressing minimal quantities, such as ‘step’
or ‘crumb’, or generalizing pronouns like ‘(any)thing’, Croft’s cycle (named
for Croft by Kahrel 1996), involves the evolution of new markers of negation
developed from negative existential particles. Croft (1991: 6) distinguishes the
following three types of languages:
Type A: the verbal negator is also used to negate existential predicates.
Type B: there is a special negative existential predicate distinct from the verbal
negator.
Type C: there is a special negative existential predicate, and this form is also
used to negate verbs.
For Type A, Croft (1991: 7) cites the example of Syrian Arabic mā fī ‘there is
not’ and mā baʕref ‘I do not know’ among others. For Type B he cites (1991: 9),
among other examples, the contrast between the Amharic negative existential
yälläm (affirmative existential allä) and regular verbal negation a(l)…-əm. For
Type C he cites (1991: 11–12) Manam (Oceanic) among other languages, giving
the example in (3).
(3)
Manam (Croft 1991: 11–12; Lichtenberk 1983: 385, 499)
a. Verbal negation
tágo
u-lóŋo
neg(.exs) 1sg.real-hear
‘I did not hear.’
b. Negative existential predicate
anúa-lo tamóata tágo
[*i-sóaʔi]
village-in person neg.exs [3sg.real-exs]
‘There is no one in the village.’
645
Christopher Lucas
A number of languages also exhibit variation between two of the types: A ~
B, B ~ C, and C ~ A. This indicates a cyclical development A > B > C > A, in
which a special negative existential predicate arises in a language (A > B), comes
to function also as a verbal negator (B > C), and is then felt to be the negator
proper, requiring supplementation by a positive existential predicate in existential constructions (C > A).
While Croft’s cycle is less common than Jespersen’s cycle, and has not been
shown to have occurred in its entirety in the recorded history of any language, I
mention it here because recent work by Wilmsen (2014: 174–176; 2016), discussed
below in §2.1.2, argues for several instances of Croft’s cycle in the history of
Arabic.
1.3 Changes to indefinites in the scope of negation
The final major set of common changes to be dealt with here involve indefinite
pronouns and quantifiers in the scope of negation. Here too cyclical patterns
are commonplace, and these changes have been labelled “the argument cycle”
(Ladusaw 1993) or “the quantifier cycle” (Willis 2011). What we find is that certain items, typically quantifiers such as ‘all’ or ‘one’ or generic nouns such as
‘person’ or ‘thing’, are liable to develop restrictions on the semantic contexts in
which they can occur, namely what are referred to as either downward-entailing
or non-veridical contexts (see Giannakidou 1998 for details and the distinction between the two). In essence, this means interrogative, conditional, and negative
clauses, as well as the complements of comparative and superlative adjectives,
but not ordinary affirmative declarative clauses. Items that are restricted to appearing in such contexts, such as English ever (consider the ungrammaticality
of, e.g., *I’ve ever been to Japan), are generally termed negative polarity items.
Often, however, we find negative polarity items whose appearance is restricted
to a subset of these contexts, and much the most common restriction is to negative contexts only. Items with this narrower distribution, such as the English
degree-adverbial phrase one bit, are generally termed strong negative polarity
items and those with the wider downward-entailing/non-veridical distribution
may be termed weak negative polarity items in contrast.
A commonly recurring diachronic tendency of such items is that they become
stronger over time. That is, an item goes from having no restrictions, to being a
weak negative polarity item, to being a strong negative polarity item, to eventually being itself inherently negative. The best-known instance of this progression comes from French personne ‘nobody’ and rien ‘nothing’. These derive from
the ordinary, unrestricted Latin generic nouns persona ‘person’ and rem ‘thing’
and still behaved as such in medieval French, as in (4).
646
30 Contact and the expression of negation
(4)
Medieval French (Hansen 2013: 72; Buridant 2000: 610)
Et si vous dirai
une
rien.
and so 2pl say.fut.1sg indf.sg.f thing
‘And so I’ll tell you a thing.’
In later medieval French they grammaticalized as indefinite pronouns and began to acquire a weak negative polarity distribution, as in the interrogative example in (5).
(5)
Thirteenth-century French (Hansen 2013: 72; Buridant 2000: 610)
As
tu rien
fet?
aux.2sg 2sg anything do.ptcp.pst
‘Have you done anything?’
In present-day French these items have become essentially inherently negative, as shown in (6). They can no longer appear in interrogative, conditional
or main declarative clauses with an affirmative interpretation (Hansen 2013: 73),
though an affirmative interpretation remains possible in comparative complements, albeit largely in frozen expressions, as in rien au monde ‘anything in the
world’ in (7).
(6) Contemporary French (Hansen 2013: 68)
Qui t’
a
vu?
Personne!
who 2sg.obj aux.3sg see.ptcp.pst nobody
‘Who saw you? Nobody!’
(7)
Contemporary French (Hansen 2013: 73)
monde.
au
vin mieux que rien
J’ aime le
1sg like.prs def.sg.m wine better than anything in+def.sg.m world
‘I like wine better than anything in the world.’
Note that French rien ‘nobody’ and personne ‘nothing’, like their equivalents in
many other Romance varieties (e.g. Italian niente and nessuno), are not straightforward negative quantifiers like English nobody and nothing, even disregarding
their behaviour in contexts such as (7). This is because French, like many other
languages but unlike Standard English, Standard German, Classical Latin etc.,
exhibits negative concord. This refers to the fact that when two (or more) elements which express negation on their own co-occur in a clause, the result is not
logical double negation (i.e. a positive) but a single logical negative, as illustrated
in (8).
647
Christopher Lucas
(8)
Contemporary French (Hansen 2013: 69)
Personne n’ a
rien
dit.
nobody neg aux.prs.3sg nothing say.ptcp.pst
‘Nobody said anything.’
Items which have this unstable behaviour are distinguished from straightforwardly negative items by the term n-word (coined by Laka 1990; see also
Giannakidou 2006). We will see in §3 that these distinctions and terminology
are helpful in understanding developments in varieties of Arabic and its contact
languages that directly parallel those described above for French.
2 Developments in the expression of clausal negation
2.1 Arabic
2.1.1 Synchronic description
One of the most striking ways that a number of spoken Arabic varieties differ
from Classical and Modern Standard Arabic is in the expression of negation. In
Classical and Modern Standard Arabic, and in the majority of varieties spoken
outside of North Africa, negation is exclusively preverbal, with the basic verbal
negator in the spoken varieties being mā, as in the Damascus Arabic example in
(9).
(9)
Damascus Arabic (Cowell 1964: 328)
hayy masʔale mā bəḍḍaḥḥək
dem.f matter neg laugh.caus.impf.ind.3sg.m
‘This is not a laughing matter.’ (lit. ‘does not cause laughter’)
But in the varieties spoken across the whole of coastal North Africa and into
the southwestern Levant, as well as in parts of the southern Arabian Peninsula
(see Diem 2014; Lucas 2018 for more precise details), negation is bipartite, with
preverbal mā joined by an enclitic -š which follows any direct or indirect pronominal object clitics, as in (10).
(10) Cairo Arabic (advertising slogan)
laʔ
banda ma yitʔal-lahā-š
Panda neg say.pass.impf.3sg.m-dat.3sg.f-neg no
‘You don’t say “no” to Panda.’ (lit. ‘Panda, “no” is not said to it.’)
648
30 Contact and the expression of negation
Finally, in a subset of the varieties that permit the bipartite construction in
(10), a purely postverbal construction is also possible, as in the Palestinian Arabic
example in (11).
(11)
Palestinian Arabic (Seeger 2013: 147)
badaḫḫini -š
smoke.impf.ind.1sg-neg
‘I don’t smoke.’
2.1.2 Jespersen or Croft?
There is near unanimous agreement among those who have considered the matter that the bipartite construction illustrated in (10) arose from the preverbal construction via grammaticalization, phonetic reduction, and cliticization of šayʔ
‘thing’, and that the purely postverbal construction in (11) in turn arose from the
bipartite construction via omission of the original negator mā. As such, Lucas
(2007; 2009; 2018) and Diem (2014), among many others, view this as a paradigmatic case of Jespersen’s cycle.
The only dissenting voice is that of Wilmsen (2013; 2014), who describes the
parallels between the Arabic data and that of well known cases of Jespersen’s
cycle such as French as being “dutifully mentioned by all” (2014: 117) who write
on the topic. Wilmsen (2014) turns the agreed etymology of negative -š on its
head by arguing: (i) that the original form in Arabic was šī, not šayʔ ;2 (ii) that
at an early stage this form had the full range of functions that we observe for it
in different Arabic dialects today (existential predicate, indefinite determiner, interrogative particle; see Wilmsen 2014: ch. 3, 122–123); (iii) that this element was
then reanalysed as a negative particle; and (iv) šī/šayʔ as a content word ‘thing’ is
a later development of the function word – an instance of degrammaticalization.
For a discussion of some of the numerous difficulties with these proposals, see
Al-Jallad (2015), Pat-El (2016), Souag (2016) and Lucas (2018).
A specific element of Wilmsen’s proposals that we need to consider in some
detail here before we proceed is his suggestion that, while in his view we should
not see the developments in Arabic as an instance of Jespersen’s cycle, we can
discern in them an instance of Croft’s cycle. As we will see below, this suggestion
involves a distortion or misunderstanding of both the Arabic data and the sorts
2
Wilmsen (2014) also attempts to trace his etymology back further to the Proto-Semitic thirdperson pronouns. Apart from the implausibility of the putative semantic shift from definite
pronoun to indefinite determiner, this reconstruction is untenable on phonological grounds
(see Al-Jallad 2015 for details).
649
Christopher Lucas
of patterns that constitute genuine instances of Croft’s cycle, but the proposal
has some prima facie plausibility, because of the existence in some dialects of
the south and east of the Arabian Peninsula of an existential predicate šī/šē/šay,
as in (12).
(12)
Northern Omani Arabic (Eades 2009: 92)
ḥmīr
šē l-ḥmīr
barra
donkey.pl exs def-donkey.pl outside
‘There were donkeys… the donkeys were outside.’
Note that a similar element śī [ɬiː], with the same existential function, is found
in the Modern South Arabian languages (MSAL) of Yemen and Oman, as in (13),
from Mehri of Yemen.
(13)
Mehri of Yemen (Watson 2011: 31)
śī fśē
exs lunch
‘Is there any lunch?’
Though Wilmsen (2014: 126; 2017: 298–301) seems to view Arabic šī and Modern South Arabian śī as cognates, it is more likely that the presence of this item in
the one set of varieties is the result of transfer from the other (cf. Al-Jallad 2015).
The direction of transfer is unclear, however. At first glance, the fact that śī as
an affirmative existential is found in essentially all of the MSAL spoken on the
Arabian Peninsula, which have a long history of intensive contact with Arabic,
but not in Soqoṭri, spoken on the island of Soqotra, where contact with Arabic is
more recent and less intensive (Simeone-Senelle 2003), would appear to suggest
that this is an innovation within Arabic originally, which was then transferred
to just those MSAL with which there was most contact. On the other hand, the
precise situation in Soqoṭri is perhaps instructive. Here the affirmative existential predicate is a unique form ino, while the negative existential predicate is biśi
(Simeone-Senelle 2011: 1108). It is conceivable that the latter is a borrowing from
Arabic, since affirmative existentials in b- are widespread in the Arabic dialects
of Yemen. But a negative existential predicate bīši or similar is completely unattested in the Yemeni data provided by Behnstedt (2016: 346–348). This suggests,
therefore, that: (i) existential śī is an original feature of MSAL; (ii) Soqoṭri is an
example of a Type B language in Croft’s typology, having innovated a new affirmative existential predicate ino, such that there is a special negative existential
predicate that is neither identical to the verbal negator, nor simply a combination
650
30 Contact and the expression of negation
of the verbal negator with the affirmative existential predicate; and (iii) šī as an
existential predicate in Arabic dialects is the result of transfer of MSAL śī.
This scenario is supported by the distribution of existential šī within Arabic
varieties: the only clear cases are in dialects of Yemen and Oman with a history
of contact with MSAL, and dialects of the Gulf whose speakers are known to
have migrated there from Yemen or Oman (such as Šiḥḥī, §2.4). In various places
Wilmsen tries to make a case for existential uses of šī outside this region, but this
appears to be the result of confusion on his part between šī as a bona fide existential predicate and the existential presupposition that will inevitably be associated
with the use of šī as an indefinite determiner (see, e.g., Heim 1988 on the semantics of indefinite noun phrases). For example, Wilmsen (2014: 123) cites Caubet’s
(1993a: 123, 1993b: 280) Moroccan Arabic examples in (14) as evidence of an existential use of šī as far west as Morocco. But there is no justification for Wilmsen’s
contradicting Caubet’s uncontroversial analysis of šī as an indefinite determiner
here: there are no existential predicates in these examples – the existence of the
referents of the indefinite noun phrases is presupposed, not asserted.
(14)
Moroccan Arabic (Caubet 1993a: 123, Caubet 1993b: 280)
a. ši
nās
kayāklu-ha
indf people eat.impf.real.3pl-3sg.f
‘Some people eat it.’
əl-lbən
kaybɣēw
b. ši
nās
indf people like.impf.real.3pl def-milk
‘Some people like milk.’
Nevertheless, šī does function as an existential predicate in a few Arabic varieties. The question, then, is whether a negated form of this predicate participates
in a version of Croft’s cycle, as Wilmsen maintains.
For the vast majority of Arabic varieties the answer is a clear no: these varieties
straightforwardly belong to Type A of Croft’s typology. The verbal negator (mā,
mā…-š, or -š) is also used to negate existential predicates, as illustrated in (15) for
Cairo Arabic.
(15)
Cairo Arabic, personal knowledge
a. ma ʕamalti -š
ḥāga
neg do.prf.1sg-neg thing
‘I didn’t do anything.’
ḥāga
b. ma fī-š
neg exs-neg thing
‘There is nothing.’
651
Christopher Lucas
Wilmsen (2014: 173–175) suggests that Type B and Type C constructions can
also be found, however. For Type B (“there is a special negative existential predicate, distinct from the verbal negator”; Croft 1991: 6), he cites Sana’a māšī and
Moroccan māši. Sana’a māšī is certainly a negative existential predicate. But
there is nothing special about it – it is a paradigmatic Type A construction, with
the negation of the existential predicate (šī) performed by the verbal negator
(mā). Moroccan māši, on the other hand, is the negator for nominal predicates
(equivalent to muš/miš/mū/mub in dialects east of Morocco). It is not a negative
existential predicate at all, and, as discussed above, the /ši/ component of this
item does not function as an existential in Moroccan, unlike in Sana’a and other
southern Arabian varieties. The existence of māši in Moroccan Arabic is thus irrelevant to the question of whether this constitutes a Type B variety.3 Moroccan
is a Type A variety: the positive existential predicate is kāyn and it is negated
with the ordinary Moroccan verbal negator ma…-š (Caubet 2011).
Wilmsen’s identification of Arabic varieties of Type C (“there is a special negative existential predicate, which is identical to the verbal negator”; Croft 1991:
6) depends on the idea that the Arabic predicate negator māši/muš/miš/mū/mub
is a negative existential predicate, which, as we have seen, it is not. If it were, it
would be true that there are Arabic varieties that are optionally of Type C, since
in Cairo Arabic, among other varieties, it is possible to negate verbs with miš
instead of the usual ma…-š, as Mughazy (2003) and others have pointed out. But
Cairo miš (and Moroccan māši) are not negative existential predicates, and there
is no evidence to suggest they ever were. Moreover, since the Sana’a negative
existential predicate māšī also does not seem to be able to function as a verbal
negator, there is little apparent merit in Wilmsen’s (2014) attempt to recast the
history of negation in Arabic as an instance of Croft’s cycle.4
3
Van Gelderen (2018) argues that the definition of Croft’s cycle should be expanded to encompass cases in which new negators arise from the univerbation of verbal negators with copulas
and auxiliaries, as well as existentials. Wilmsen’s (2014) presentation of Croft’s cycle makes no
mention of any predicates other than existentials participating in the cycle, however.
4
This is not to deny, however, that some Arabic dialects show some incipient Type B tendencies
of a different kind. For example, Behnstedt (2016: 347) cites the northern Yemeni dialects of Rās
Maḥall as-Sūdeh, Ḥammām ʕAlī and Afk, as varieties in which different morphemes are used in
positive and negative existentials, albeit the negative construction used in each case is identical
to that used for ordinary verbal negation. In a different context, Stefano Manfredi (personal
communication) points out that many urban speakers of Sudanese Arabic use the item māfīš,
borrowed from Egyptian Arabic, as a negative existential, while ordinary verbal negation is
performed with preverbal mā alone (without postverbal -š).
652
30 Contact and the expression of negation
2.1.3 Internal or external?
It is clear from the above discussion that there is no reason to doubt the majority view of the emergence of negative -š as an instance of Jespersen’s cycle.
What is less clear and more controversial is the question of whether language
contact played a role in triggering these developments, or whether this was a
purely internal phenomenon (cf. Diem 2014: 11–12). This is an issue about which
it is impossible to be certain given our present state of knowledge. Lucas & Lash
(2010) make the case that contact did play a triggering role, however, and also
provide arguments against the widely held view that, in the words of Lass (1997:
209), “an endogenous explanation of a phenomenon is more parsimonious [than
one invoking contact – CL], because endogenous change must occur in any case,
whereas borrowing is never necessary” (cf. also Lucas 2009: 38–43). Aside from
this generalized reluctance to invoke contact in explanations of linguistic change
unless absolutely necessary, another factor that is likely operative in the preference for seeing the Arabic developments as a purely internal phenomenon is
ignorance of the wider picture of negative developments in Arabic and its contact languages. It is scarcely an exaggeration to say that everywhere an Arabic
variety with bipartite negation is spoken, there is (or was) a contact language that
also has bipartite negation, and – just as importantly – wherever Arabic dialects
have only a single marker of negation, the local contact languages do too. The
picture is similar in Europe, Ethiopia (Lucas 2009), Vietnam (van der Auwera &
Vossen 2015), and many other places besides. There can therefore be no doubt that
negative constructions, and especially bipartite negation (and hence Jespersen’s
cycle more generally), are particularly prone to diffusing through languages in
contact. In the following sections I will briefly survey apparent instances of transfer of bipartite or postverbal negation in Arabic and Coptic, Arabic and MSAL,
Arabic and Kumzari, Arabic and Berber, and Arabic and Domari. For more details
see Lucas (2007; 2009; 2013) and Lucas & Lash (2010).
2.2 Arabic and Coptic
Based on an examination of evidence from Judaeo-Arabic documents preserved
in the Cairo Genizah, among other sources of evidence, Diem (2014) comes to the
conclusion that the Arabic bipartite negative construction found across coastal
North Africa originated in Egypt between the tenth and eleventh centuries. This
chronology and point of origin conforms closely with the conclusions I have
drawn on this point in my own work (Lucas 2007; 2009; Lucas & Lash 2010),
except that I have argued that what triggered the development of bipartite negation in Egypt was contact with Coptic (the name for the Egyptian language from
653
Christopher Lucas
the first century CE onwards), which, at the relevant period, had a frequently
occurring bipartite construction ən…an, as illustrated in (16).
(16) Coptic (Lucas & Lash 2010: 389)
en ti-na-tsabo-ou
an e-amənte
neg 1sg-fut-teach-3pl neg on-hell
‘I will not teach them about hell.’
The argument made in Lucas & Lash (2010) is that native speakers of Coptic
acquiring Arabic as a second language must have encountered sentences negated
with preverbal mā only, but which also contained after the verb šī/šāy, functioning either as an argument ‘(any)thing’ or an adverb ‘at all’,5 and interpreted
this as the second element of the bipartite negative construction that their firstlanguage Coptic predisposed them to expect. If this is correct, then the initial
transfer of bipartite negation from Coptic to Arabic in Egypt should be understood as an instance of imposition under source-language agentivity, in the terms
of Van Coetsem (1988; 2000), while the presence of bipartite negation in the dialects spoken across the rest of coastal North Africa, and the southwestern Levant,
should be understood as the result of contact between neighbouring dialects of
Arabic.
2.3 Arabic and Modern South Arabian
Diem (2014: 73) – like Obler (1990: 148) and, following her, Lucas (2007: 416) – suggests that bipartite negation in the southern Arabian Peninsula must have spread
there from Egypt. This is conceivable, but historical evidence of significant early
migration flows in this direction is lacking. The alternative explanation offered
by Lucas & Lash (2010) is that bipartite negation in the Arabic dialects of this
region is an independent parallel development, here triggered by contact with
MSAL, all mainland varieties of which have a bipartite negative construction of
their own (or once had – some, such as Ḥarsūsi, have largely progressed to stage
III of Jespersen’s cycle and lost the original preverbal negator), as illustrated in
(17) for Omani Mehri.
(17) Omani Mehri (Johnstone 1987: 23)
laʔ
b-ɛy
əl təhɛləz
neg nag.impf.2sg.m with-1sg neg
‘Don’t nag me!’
5
Diem (2014) makes the case that šī/šāy had already developed an adverbial use at a very early
stage, and that it is this adverbial use that should be seen as the form that was reanalysed as a
negator.
654
30 Contact and the expression of negation
If this is correct, then here too, exactly as with the Coptic–Arabic contact in
the previous section, we must have had an instance of transfer under sourcelanguage agentivity, with MSAL-dominant acquirers of Arabic imposing a bipartite construction on their second-language Arabic by reanalysing šī/šay as a
negator. The key point is that in all dialects in which šī/šay functioned as an indefinite pronoun or adverb ‘at all’, the potential was there for reanalysis as the
second element in a bipartite negative construction. But aside from in the dialects of Egypt and the southern Arabian Peninsula (and latterly dialects adjacent
to Egyptian) this reanalysis never took place. Why the reanalysis did take place
in Egypt and the southern Peninsula can be understood as being the result of
the catalysing effect of contact with languages which themselves had a bipartite
negative construction.6
2.4 Arabic and Kumzari
Kumzari is an Iranian language with heavy influence from both Arabic and MSAL
that has only recently been described in detail (see van der Wal Anonby forthcoming). It is spoken on the Musandam Peninsula of northern Oman, where its
primary contact language of recent times has been the Šiḥḥī variety of Arabic (see
Bernabela 2011 for a sketch grammar), which is clearly of the originally southern
Arabian type described by Holes (2016: 18–32).
Šiḥḥī Arabic has no Jespersen stage II (bipartite) negative construction, but it
has both a typical eastern Arabic stage I construction with mā, as in (18a), perhaps
due to recent influence from other Gulf Arabic varieties, alongside a unique (for
Arabic) stage III postverbal construction with -lu, as in (18b). The latter construction is apparently a straightforward transfer of the postverbal negator laʔ/lɔʔ of
MSAL (17).
(18)
Šiḥḥī Arabic (Bernabela 2011: 87)
ḫaṣāb əl-yōm
a. mā mšēt
neg go.prf.1sg Khasab def-day
‘I didn’t go to Khasab today.’
b. yqōl-lu
bass il-kilmatēn
say.impf.3sg.m-neg only def-words.du
‘He doesn’t just say the two words.’
6
For further discussion of the details of these changes, including the issues of the semantics and
positioning in the clause of the second negative element in each of the three languages, see
Lucas & Lash (2010: 395–401).
655
Christopher Lucas
The Kumzari negator is the typical Iranian (and Indo-Iranian) na. What is less
typical is that na occurs postverbally in Kumzari, as shown in (19).
(19) Kumzari (van der Wal Anonby forthcoming: 211)
mām-ō
kōr bur
na
mother-def blind become.3sg.real neg
‘The mother didn’t become blind.’
It seems very likely that contact with Šiḥḥī Arabic has played a role in this
shift to postverbal negation, though not enough is known about the historical
sociolinguistics of these two speech communities to say with confidence which
of the two languages the agents of this change were dominant in.
2.5 Arabic and Berber
Berber languages are spoken from the oasis of Siwa in western Egypt in the
east, across to Morocco and as far south as Burkina Faso. The most southerly of
the Berber varieties – Tashelhiyt, spoken in southern Morocco, Zenaga, spoken
in Mauritania, and Tuareg, spoken in southern Algeria and Libya, Niger, Mali
and Burkina Faso – have only preverbal negation, as illustrated by the Tuareg
example in (20).
(20)
Tuareg (Chaker 1996: 10)
ur igle
neg leave.pfv.3sg.m
‘He didn’t leave.’
These languages have, until recently, either had little significant contact with
Arabic, or otherwise only with varieties such as Ḥassāniyya that have only preverbal negation with mā. All other Berber varieties which are in contact with
Arabic varieties with bipartite negation also themselves have bipartite negation,
illustrated for Kabyle (Algeria) in (21), or, in a few cases, purely postverbal negation, as in Awjila (Libya), illustrated in (22). The one exception is Siwa (23),
which negates with preverbal lā alone – clearly a borrowing from a variety of
Arabic, though which variety is not clear (see Souag 2009 for further discussion).
(21)
656
Kabyle (Rabhi 1996: 25)
kra
ul ittaggad
neg fear.aor.3sg.m neg
‘He is not afraid.’
30 Contact and the expression of negation
(22)
Awjila (Paradisi 1961: 82)
akellim iššen-ka
amakan
servant know.pfv.3sg.m-neg place
‘The servant didn’t know the place.’
(23) Siwa (Souag 2009: 58)
lā gā-nūsd-ak
neg fut-come.1pl-dat.2sg
‘We won’t come to you.’
Different Berber varieties have postverbal negators with a range of different
forms, but in most cases they either derive from two apparently distinct ProtoBerber items *kʲăra and *(h)ară(t), both meaning ‘thing’ (Kossmann 2013: 332),
or are transparent loans of Arabic šay/ši. This fact, when combined with the respective geographical distributions of single preverbal and bipartite negation in
Arabic and Berber varieties, is sufficient to conclude that the presence of bipartite negation in Berber is in large part a result of calquing the second element
of the Arabic construction, pace Brugnatelli (1987) and Lafkioui (2013a) (see also
Kossmann 2013: 334; and see Lucas 2007; 2009 for more detailed discussion).7
Given that, until recently, native speakers of Arabic in the Maghreb acquiring
Berber as a second language will always have been greatly outnumbered by native speakers of Berber learning Arabic as a second language, we must assume
that the agents of this change were Berber-dominant speakers who made the
change under recipient-language agentivity in a process akin to what Heine &
Kuteva (2005) call polysemy copying and contact-induced grammaticalization
(see also Leddy-Cecere, this volume; Manfredi, this volume; Souag, this volume).
2.6 Arabic and Domari
The final instance of contact-induced changes to predicate negation to be mentioned here concerns the Jerusalem variety of the Indo-Aryan language Domari,
as described by Matras (1999; 2007; 2012; this volume).
Matras (2012: 350–351) describes two syntactic contexts in which negators borrowed from Palestinian Arabic are the only options in this variety of Domari. The
first is with Arabic-derived modal auxiliaries that take Arabic suffix inflection,
as in bidd- ‘want’ in (24). Here negation is typically with the Palestinian Arabic
stage III construction -š (without mā), as it is would be also in Palestinian Arabic.
7
Another postverbal negator – Kabyle ani – derives from the word for ‘where’ (Rabhi 1992),
and so should perhaps be seen as more of an internal development, or at least less directly
contact-induced. Tarifiyt also has a postverbal negator bu, whose etymology is uncertain, but
which has also been transferred to the Moroccan Arabic dialect of Oujda (Lafkioui 2013b).
657
Christopher Lucas
(24) Jerusalem Domari (Matras 2012: 351)
ben-om bidd-hā-š
žawwiz-hōš-ar
sister-1sg want-3sg.f-neg marry-vitr.sbjv-3sg
‘My sister doesn’t want to marry.’
The second is when the negated predicate is nominal, as in (25a), or, to judge
from Matras’s examples, when we have narrow focus of negation with ellipsis,
as in (25b). Here the negator that would be used in these contexts in Arabic – miš
– is transferred to Domari and functions in the same way.
(25)
Jerusalem Domari (Matras 2012: 350)
a. bay-os
mišš kury-a-m-ēk
mother-3sg neg house-obl.f-loc-pred.sg
‘His wife is not at home.’
mišš min ʕēl-oman-ki
b. day-om
min ʕammān-a-ki
mother-1sg from Amman-obl.f-abl neg from family-1pl-abl
day-om
mother-1sg
‘My mother is from Amman, she’s not from our family, my mother.’
In addition to these straightforward borrowings, Domari has a bipartite negative construction in which both elements involve inherited lexical material, as
illustrated in (26).
(26) Jerusalem Domari (Matras 2012: 117)
ʕašān ihne ama n-mang-am-san-eʔ
l-ʕarab
because thus 1sg neg-want-1sg-3pl-neg def-Arabs
‘Because of this I don’t like the Arabs.’
In Lucas (2013: 413–414) I pointed out that the second element of this construction – -eʔ – was apparently not attested in varieties of Domari spoken outside of
Palestine, and suggested that its presence in Jerusalem Domari could therefore
be the result of influence from the Palestinian bipartite negative construction.
Herin (2016; 2018), however, has since convincingly shown that this is incorrect,
and that the Jerusalem Domari bipartite construction is an internal development
with cognates in more northerly varieties, the latter being in contact with Arabic
varieties that lack the bipartite negative construction. What is unique about the
Jerusalem variety of Domari is that here a stage III construction with -eʔ alone is
possible, omitting the original preverbal negator n(a) that appears in (25b). Herin
658
30 Contact and the expression of negation
(2018: 32) argues that it is this stage III construction, not the stage II bipartite construction, that should be seen as the result of contact with Palestinian Arabic.
Overall, therefore, while the details naturally vary from one contact scenario
to another, we see that negative constructions appear just as liable to be transferred between varieties of Arabic and neighbouring languages as they are between the languages of Europe and beyond.
3 Developments in indefinites in the scope of negation
3.1 Loaned indefinites
The organization and behaviour of indefinites in the scope of negation seem to
be much more resistant to transfer between languages than is the expression
of clausal negation, at least in the case of Arabic and its contact languages.8
Direct borrowing of individual indefinite items is rather common, however. I
make no attempt at an exhaustive list here, but note the following two examples
for illustrative purposes.
First, Berber varieties stand out as frequent borrowers of Maghrebi Arabic
indefinites. The negative polarity item ḥadd/ḥədd ‘anyone’ is borrowed by at
least Siwa (Souag 2009: 58), Kabyle, Shawiya, Mozabite (Rabhi 1996: 29), and
Tashelhiyt (Boumalk 1996: 41). The n-word walu ‘nothing’ is borrowed by at least
Tarifiyt (Lafkioui 1996: 54), Tashelhiyt, and Central Atlas Tamazight (Boumalk
1996: 41). ḥətta, in its function as an n-word determiner, is borrowed by at least
Tashelhiyt (Boumalk 1996: 41). qāʕ, in its function as a negative polarity adverb ‘at
all’, is borrowed by at least Tarifiyt and Central Atlas Tamazight (Boumalk 1996:
42). And the negative polarity adverb *ʕumr ‘(n)ever’ (< ‘age, lifetime’) is borrowed by at least Kabyle, Mozabite (Rabhi 1996: 30), and Tarifiyt (Lafkioui 1996:
72). Why these items should have been so freely borrowed, when each of them,
with the possible exception of ḥətta, have direct native equivalents, is unclear.
But it is perhaps to be connected with the high degree of expressivity typically
associated with negative statements containing indefinites, which therefore creates a constant need for new and “extravagant” (in the sense of Haspelmath 2000)
means of expressing these meanings.
Second, while Arabic itself seems to have been much more constrained in its
borrowing of indefinites from other languages, we can here point at least to the
8
Though for recent discussion of a related case – namely the acquisition of a determiner function by the Berber indefinite kra ‘something, anything’ via a calque of the polyfunctionality of
Maghrebi Arabic ši – see Souag (2018).
659
Christopher Lucas
n-word hīč ‘nothing’, borrowed from Persian, which Holes (2001: 549) includes
in his glossary of pre-oil era Bahraini Arabic, citing also Blanc (1964: 159) and
Ingham (1973: 547) for its occurrence in Baghdadi and Khuzestan Arabic respectively. It remains in use in the latter (cf. Leitner, this volume), but consultations
with present-day speakers of Baghdadi Arabic indicate that, in this variety at
least, this item has since dropped out of use.
3.2 The indefinite system of Maltese
While most or perhaps all Arabic varieties have at least some items that qualify
as n-words according to the definition in §1.3, it is only Maltese that has developed into a straightforward negative-concord language with a full series of
n-word indefinites in largely complementary distribution with a separate series
of indefinites that cannot appear in the scope of negation, as is the situation in
French, described in §1.3. These two series are shown in Table 1, adapted from
Haspelmath & Caruana (1996: 215).
Table 1: Maltese indefinites
Determiner
Thing
Person
Time
Place
n-words
non-n-words
ebda
xejn
ħadd
qatt
imkien
xi
xi ħaġa
xi ħadd
xi darba
xi mkien
All the lexical material that makes up the Maltese indefinite system illustrated
in Table 1 is inherited from Arabic, but the neat paradigm of n-words for determiner, ‘thing’, ‘person’, ‘time’, and ‘place’ is much more typical of European
Romance languages than of Arabic. The extent to which, for example, xejn ‘nothing’ (deriving from šayʔ ‘thing’)9 is felt by Maltese speakers to be inherently
negative, is shown by the existence of the denominal verb xejjen meaning ‘to
nullify’, as illustrated in (27).10
9
As pointed out in Lucas (2009: 83–84) and argued in greater detail in Lucas & Spagnol (forthcoming), the final segment of this item represents a fossilized retention of the indefinite suffix
(so-called nunation or tanwīn), as found in Classical Arabic.
10
This is despite the fact that it may also occur in interrogatives with non-negative meaning (cf.
Camilleri & Sadler 2017). Compare the French n-word rien, which, as illustrated in (7), retains
a non-negative interpretation in a restricted set of negative-polarity contexts.
660
30 Contact and the expression of negation
(27)
Maltese (Lucas 2013: 441)
Iżda xejjen
lil-u
nnifs-u
but nullify.prf.3sg.m obj-3sg.m self-3sg.m
‘But he made himself nothing.’
As such, it seems likely that the intensive contact that occurred over several
centuries between Maltese and the negative-concord languages Sicilian and Italian (cf. Lucas & Čéplö, this volume) played a role in these developments in the
Maltese indefinite system. Precisely how this influence was mediated is hard to
say, since both borrowing under recipient-language agentivity and imposition
under source-language agentivity were likely operative in the Maltese–Romance
contact situation, and either are possible here. See Lucas (2013: 439–444) for further discussion.
4 Conclusion
As we have seen, the overall areal picture of bipartite clausal negation in Arabic
and its contact languages (and also, to a lesser extent, indefinites in the scope of
negation) strongly suggests a series of contact-induced changes, and not a series
of purely internally-caused independent parallel developments. What is required
in future research on this topic, to the extent that textual and other historical evidence becomes available, is a detailed, case-by-case examination of the linguistic
and sociolinguistic conditions under which these constructions emerged in the
languages in question. Such investigations would serve to either substantiate
or undermine the contact-based explanations for these changes advanced in the
course of this chapter. Ideally, they would also allow to understand in more detail the mechanisms of bilingual language use and acquisition that give rise to
changes of this sort.
Further reading
) Chaker & Caubet (1996) is an edited volume providing a wealth of descriptive
data on the expression of negation in a number of Berber and Maghrebi Arabic
varieties.
) Diem (2014) is a detailed study of the grammaticalization of Arabic šayʔ as a
negator, with particular attention paid to early sources of textual evidence for
this development.
) Willis et al. (2013) and Breitbarth et al. (2020) are two volumes of a work examining in detail the history of negation in the languages of Europe and the
Mediterranean.
661
Christopher Lucas
Acknowledgements
The research presented in this chapter was partly funded by Leadership Fellows
grant AH/P014089/1 from the UK Arts and Humanities Research Council, whose
support is hereby gratefully acknowledged. I am also very grateful to Stefano
Manfredi, Lameen Souag and Bruno Herin for their comments on an earlier draft
of the chapter. Responsibility for any failings that remain is mine alone.
Abbreviations
1, 2, 3
abl
aor
aux
caus
dat
def
dem
du
exs
f
fut
impf
ind
indf
m
1st, 2nd, 3rd person
ablative
aorist
auxiliary
causative
dative
definite article
demonstrative
dual
existential
feminine
future
imperfect (prefix conjugation)
indicative
indefinite
masculine
neg
obj
MSAL
obl
pass
pfv
pst
pl
ptcp
pred
prf
prs
real
sbjv
sg
vitr
negative
object
Modern South Arabian
oblique
passive
perfective
past
plural
participle
predicate
perfect (suffix conjugation)
present
realis
subjunctive
singular
intransitive marker
References
Al-Jallad, Ahmad. 2015. What’s a caron between friends? A review article of
Wilmsen (2014), with a special focus on the etymology of modern Arabic ši.
Bibliotheca Orientalis 72(1/2). 34–46.
Behnstedt, Peter. 2016. Dialect atlas of North Yemen and adjacent areas. Leiden:
Brill. (Translated by Gwendolin Goldbloom).
Bernabela, Roy S. 2011. A phonology and morphology sketch of the Šiħħi Arabic
dialect of əlǦēdih, Musandam (Oman). Leiden: Leiden University. (Master’s
dissertation).
Bernini, Giuliano & Paolo Ramat. 1992. La frase negativa nelle lingue d’Europa.
Bologna: Il Mulino.
662
30 Contact and the expression of negation
Bernini, Giuliano & Paolo Ramat. 1996. Negative sentences in the languages of
Europe. Berlin: Mouton de Gruyter.
Blanc, Haim. 1964. Communal dialects in Baghdad. Cambridge, MA: Harvard
University Press.
Boumalk, Abdallah. 1996. La négation en berbère marocain. In Salem Chaker &
Dominique Caubet (eds.), La négation en berbère et en arabe maghrébin, 35–48.
Paris: L’Harmattan.
Breitbarth, Anne, Christopher Lucas & David Willis. 2020. The history of negation
in the languages of Europe and the Mediterranean. Vol. 2: Patterns and processes.
Oxford: Oxford University Press.
Brugnatelli, Vermondo. 1987. La negazione discontinua in berbero e in arabomagrebino. In Giuliano Bernini & Vermondo Brugnatelli (eds.), Atti della 4a
giornata di Studi Camito-semitici ed Indoeuropei, 53–62. Milan: Unicopli.
Buridant, Claude. 2000. Grammaire nouvelle de l’ancien français. Paris: Sedes.
Camilleri, Maris & Louisa Sadler. 2017. Negative sensitive indefinites in Maltese.
In Miriam Butt & Tracy Holloway King (eds.), Proceedings of the LFG ’17
conference, 146–166. Stanford, CA: CSLI Publications. http : / / web . stanford .
edu/group/cslipublications/cslipublications/LFG/LFG-2017.
Caubet, Dominique. 1993a. L’arabe marocain. Vol. 1: Phonologie et morphosyntaxe.
Leuven: Peeters.
Caubet, Dominique. 1993b. L’arabe marocain. Vol. 2: Syntaxe et catégories grammaticales. Leuven: Peeters.
Caubet, Dominique. 2011. Moroccan Arabic. In Lutz Edzard & Rudolf de Jong
(eds.), Encyclopedia of Arabic language and linguistics, online edn. Leiden: Brill.
Chaker, Salem. 1996. Remarques préliminaires sur la négation en berbère. In
Salem Chaker & Dominique Caubet (eds.), La négation en berbère et en arabe
maghrébin, 9–22. Paris: L’Harmattan.
Chaker, Salem & Dominique Caubet (eds.). 1996. La négation en berbère et en arabe
maghrébin. Paris: L’Harmattan.
Cowell, Mark W. 1964. A reference grammar of Syrian Arabic. Washington, DC:
Georgetown University Press.
Croft, William. 1991. The evolution of negation. Journal of Linguistics 27(1). 1–27.
Dahl, Östen. 1979. Typology of sentence negation. Linguistics 17. 79–106.
Devos, Maud & Johan van der Auwera. 2013. Jespersen cycles in Bantu: Double
and triple negation. Journal of African Languages and Linguistics 34. 205–274.
Diem, Werner. 2014. Negation in Arabic: A study in linguistic history. Wiesbaden:
Harrassowitz.
663
Christopher Lucas
Eades, Domenyk. 2009. The Arabic dialect of a Šawāwī community of northern
Oman. In Enam Al-Wer & Rudolf De Jong (eds.), Arabic dialectology in honour
of Clive Holes on the occasion of his sixtieth birthday, 77–98. Leiden: Brill.
Gardiner, Alan H. 1904. The word iwn3. Zeitschrift für Ägyptische Sprache und
Altertumskunde 41. 130–135.
Giannakidou, Anastasia. 1998. Polarity sensitivity as (non)veridical dependency.
Amsterdam: John Benjamins.
Giannakidou, Anastasia. 2006. N-words and negative concord. In Martin Everaert
& Henk van Riemsdijk (eds.), The Blackwell companion to syntax, vol. 3, 327–
391. Oxford: Blackwell.
Hansen, Maj-Britt Mosegaard. 2013. Negation in the history of French. In David
Willis, Christopher Lucas & Anne Breitbarth (eds.), The history of negation
in the languages of Europe and the Mediterranean, vol. 1: Case studies, 51–76.
Oxford: Oxford University Press.
Haspelmath, Martin. 2000. The relevance of extravagance: A reply to Bart Geurts.
Linguistics 38. 789–798.
Haspelmath, Martin & Josephine Caruana. 1996. Indefinite pronouns in Maltese.
Rivista di Linguistica 8. 213–227.
Heim, Irene. 1988. The semantics of definite and indefinite noun phrases. New York:
Garland Publishing.
Heine, Bernd & Tania Kuteva. 2005. Language contact and grammatical change.
Cambridge: Cambridge University Press.
Herin, Bruno. 2016. Elements of Domari dialectology. Mediterranean Language
Review 23. 33–73.
Herin, Bruno. 2018. The Arabic component in Domari. In Stefano Manfredi &
Mauro Tosco (eds.), Arabic in contact, 19–36. Amsterdam: John Benjamins.
Holes, Clive. 2001. Dialect, culture, and society in eastern Arabia. Vol. 1: Glossary.
Leiden: Brill.
Holes, Clive. 2016. Dialect, culture, and society in eastern Arabia. Vol. 3: Phonology,
morphology, syntax, style. Leiden: Brill.
Ingham, Bruce. 1973. Urban and rural Arabic in Khūzistān. Bulletin of the School
of Oriental and African Studies 36. 533–553.
Jespersen, Otto. 1917. Negation in English and other languages. Copenhagen: A. F.
Høst.
Johnstone, T. M. 1987. Mehri lexicon and English–Mehri word-list. London:
Routledge.
Kahrel, Peter. 1996. Aspects of negation. Amsterdam: University of Amsterdam.
(Doctoral dissertation).
Kossmann, Maarten. 2013. The Arabic influence on Northern Berber. Leiden: Brill.
664
30 Contact and the expression of negation
Ladusaw, William A. 1993. Negation, indefinites, and the Jespersen Cycle. In
Joshua S. Guenter, Barbara A. Kaiser & Cheryl C. Zoll (eds.), Proceedings of the
nineteenth Annual Meeting of the Berkeley Linguistics Society, 437–46. Berkeley:
Berkeley Linguistics Society.
Lafkioui, Mena. 1996. La négation en tarifit. In Salem Chaker & Dominique
Caubet (eds.), La négation en berbère et en arabe maghrébin, 49–78. Paris:
L’Harmattan.
Lafkioui, Mena. 2013a. Reinventing negation patterns in Moroccan Arabic. In
Mena Lafkioui (ed.), African Arabic: Aprroaches to dialectology, 51–93. Berlin:
De Gruyter Mouton.
Lafkioui, Mena. 2013b. Negation, grammaticalization and language change in
North Africa: The case of the negator NEG_____*bu. In Giorgio Francesco
Arcodia, Federica Da Milano, Gabriele Iannàccaro & Paolo Zublena (eds.),
Tilelli: Scritti in onore di Vermondo Brugnatelli, 113–130. Rome: Caissa Italia.
Laka, Itziar. 1990. Negation in syntax: On the nature of functional categories
and projections. Cambridge: Massachusetts Institute of Technology. (Doctoral
dissertation).
Lass, Roger. 1997. Historical linguistics and language change. Cambridge:
Cambridge University Press.
Lichtenberk, Frantisek. 1983. A grammar of Manam. Honolulu: University of
Hawaii Press.
Lucas, Christopher. 2007. Jespersen’s cycle in Arabic and Berber. Transactions of
the Philological Society 105(3). 398–431.
Lucas, Christopher. 2009. The development of negation in Arabic and Afro-Asiatic.
Cambridge: University of Cambridge. (Doctoral dissertation).
Lucas, Christopher. 2013. Negation in the history of Arabic and Afro-Asiatic. In
David Willis, Christopher Lucas & Anne Breitbarth (eds.), The development of
negation in the languages of Europe and the Mediterranean, vol. 1: Case studies,
399–452. Oxford: Oxford University Press.
Lucas, Christopher. 2018. On Wilmsen on the development of postverbal negation
in dialectal Arabic. Zeitschrift für Arabische Linguistik 67. 44–70.
Lucas, Christopher & Elliott Lash. 2010. Contact as catalyst: The case for Coptic
influence in the development of Arabic negation. Journal of Linguistics 46. 379–
413.
Lucas, Christopher & Michael Spagnol. Forthcoming. Proceedings of lingwistika
maltija 2019. In Przemysław Turek & Benjamin Saade (eds.). Berlin: De Gruyter
Mouton.
Matras, Yaron. 1999. The state of present-day Domari in Jerusalem. Mediterranean
Language Review 11. 1–58.
665
Christopher Lucas
Matras, Yaron. 2007. Grammatical borrowing in Domari. In Yaron Matras &
Jeanette Sakel (eds.), Grammatical borrowing in cross-linguistic perspective, 151–
164. Berlin: Mouton de Gruyter.
Matras, Yaron. 2012. A grammar of Domari. Berlin: De Gruyter Mouton.
Meillet, Antoine. 1912. L’évolution des formes grammaticales. Scientia 12. 384–
400.
Mughazy, Mustafa. 2003. Metalinguistic negation and truth functions: The case
of Egyptian Arabic. Journal of Pragmatics 35(8). 1143–1160.
Obler, Lorraine. 1990. Reflexes of Classical Arabic šay’un ‘thing’ in the modern
dialects. In James Bellamy (ed.), Studies in Near Eastern culture and history
in memory of Ernest T. Abdel-Massih, 132–152. Ann Arbor, MI: University of
Michigan.
Paradisi, Umberto. 1961. Testi berberi di Augila (Cirenaïca). Annali dell’Istituto
Orientale di Napoli 10. 79–91.
Pat-El, Na’ama. 2016. David Wilmsen, Arabic indefinites, interrogatives, and
negators: A linguistic history of western dialects. Journal of Semitic Studies 61(1).
292–295.
Rabhi, Allaoua. 1992. Les particules de négation dans la Kabylie de l’Est. Etudes
et documents berbères 9. 139–145.
Rabhi, Allaoua. 1996. De la négation en berbère: Les donnés algériennes. In
Salem Chaker & Dominique Caubet (eds.), La négation en berbère et en arabe
maghrébin, 23–34. Paris: L’Harmattan.
Seeger, Ulrich. 2013. Der arabische Dialekt der Dörfer um Ramallah. Vol. 3: Grammatik. Wiesbaden: Harrassowitz.
Simeone-Senelle, Marie-Claude. 2003. Soqotri dialectology and the evaluation
of the language endangerment. In S. A. Ba-Angood, M. O. Ba-Saleem & M. O.
Hussein (eds.), Proceedings of the second International Symposium on the Socotra
Archipelago, 63–73. Aden: Aden University Printing.
Simeone-Senelle, Marie-Claude. 2011. Modern South Arabian. In Stefan
Weninger, Geoffrey Khan, Michael P. Streck & Janet C. E. Watson (eds.), The
Semitic languages: An international handbook, 1073–1113. Berlin: De Gruyter
Mouton.
Souag, Lameen. 2009. Siwa and its significance for Arabic dialectology. Zeitschrift
für Arabische Linguistik 51. 51–75.
Souag, Lameen. 2016. Werner Diem: Negation in Arabic: A study in linguistic
history. Linguistics 54(1). 223–229.
Souag, Lameen. 2018. Arabic–Berber–Songhay contact and the
grammaticalization of ‘thing’. In Stefano Manfredi & Mauro Tosco (eds.),
Arabic in contact, 54–71. Amsterdam: John Benjamins.
666
30 Contact and the expression of negation
Van Coetsem, Frans. 1988. Loan phonology and the two transfer types in language
contact. Dordrecht: Foris.
Van Coetsem, Frans. 2000. A general and unified theory of the transmission process
in language contact. Heidelberg: Winter.
van der Auwera, Johan. 2009. The Jespersen cycles. In Elly van Gelderen (ed.),
Cyclical change, 35–71. Amsterdam: John Benjamins.
van der Auwera, Johan & Frens Vossen. 2015. Negatives between Chamic and
Bahnaric. Journal of the Southeast Asian Linguistics Society 8. 24–38.
van der Auwera, Johan & Frens Vossen. 2016. Jespersen cycles in Mayan,
Quechuan and Maipurean languages. In Elly van Gelderen (ed.), Cyclical
change continued, 189–218. Amsterdam: John Benjamins.
van der Auwera, Johan & Frens Vossen. 2017. Kiranti double negation: A copula
conjecture. Linguistics of the Tibeto-Burman Area 40(1). 40–58.
van der Wal Anonby, Christina. Forthcoming. A grammar of Kumzari: A mixed
Perso-Arabian language of Oman. Berlin: De Gruyter Mouton.
van Gelderen, Elly. 2018. The negative existential and other cycles: Jespersen,
Givón, and the copula cycle. Unpublished manuscript, Arizona State
University. https://ling.auf.net/lingbuzz/004051.
Watson, Janet C. E. 2011. South Arabian and Yemeni dialects. Salford Working
Papers in Linguistics and Applied Linguistics 1. 27–40.
Willis, David. 2011. Negative polarity and the quantifier cycle: Comparative
diachronic perspectives from European languages. In Pierre Larrivée &
Richard Ingham (eds.), The evolution of negation: Beyond the Jespersen cycle,
285–323. Berlin: De Gruyter.
Willis, David, Christopher Lucas & Anne Breitbarth (eds.). 2013. The history of
negation in the languages of Europe and the Mediterranean. Vol. 1: Case studies.
Oxford: Oxford University Press.
Wilmsen, David. 2013. The interrogative origin of the Arabic negator -š: Evidence
from copular interrogation in Andalusi Arabic, Maltese, and modern spoken
Egyptian and Moroccan Arabic. Zeitschrift für Arabische Linguistik 58. 5–31.
Wilmsen, David. 2014. Arabic indefinites, interrogatives and negators: A linguistic
history of western dialects. Oxford: Oxford University Press.
Wilmsen, David. 2016. Another Croft cycle in Arabic: The laysa negative
existential cycle. Folia Orientalia 53. 327–367.
Wilmsen, David. 2017. Grammaticalization and degrammaticalization in an
Arabic existential particle šay. Folia Orientalia 54. 279–307.
667
Name index
ʕAbd ar-Raḥmān III, 2266
Abd El-Jawad, Hassan, 571
Abdel-Massih, Ernest, 606
Abdu, Hussein Ramadan, 207
Abu-Haidar, Farida, 96, 99, 306, 571
Abuamsha, Duaa, 618
Aguadé, Jordi, 256
Aikhenvald, Alexandra Y., 1, 102
Akın, Cahit, 152
Akkuş, Faruk, 3, 6, 83, 86, 867 , 91, 97,
101, 102, 138, 142, 143, 146–
150, 152, 287, 388, 469
Alamin, Suzan, 633
al-Bakrī, Ḥāzim, 103
Albirini, Abdulkafi, 305, 309–314
Al-Essa, Aziza, 552, 553, 572, 573
Alghamdi, Najla Manie, 552, 572, 573,
575
Al-Hawamdeh, Areej, 558
al-Ḥimyarī, 266
al-Idrīsī, Muḥammad, 407
al-Išbīlī, Abū l-Ḫayr, 237
Al-Jallad, Ahmad, 7, 38, 40, 42, 44–51,
61, 73, 200, 649, 650
Al-Mahri, Abdullah Musallam, 362,
366
al-Manaser, Ali, 47, 48
Almbark, Rana, 595
Almoaily, Mohammad, 339
Al-Qahtani, Khairiah, 573
Al-Salman, Ibrahim, 337, 339
Al-Shareef, Jamal, 571
Al-Tamimi, Jalal, 576
Al-Wasif, Muhammad-Fajri, 225
Al-Wer, Enam, 10, 553, 558, 559, 561–
563, 567, 571–576
Alzaidi, Muhammad, 592
Andersen, Henning, 604
Anonby, Erik, 462, 463, 468, 469
Anonymous, 537
Appleyard, David, 428
Aquilina, Joseph, 3, 627
Arjava, Antti, 44
Arnold, Werner, 3, 86, 90, 108
Arvaniti, Amalia, 162
Asbaghi, Asya, 3, 78, 454
Atatürk, 445
Atlamaz, Ümit, 152
Atterer, Michaela, 585
Avner, Uzi, 44
Avram, Andrei A., 3, 5, 7, 18, 23, 271,
272, 308, 325, 327, 329, 334–
341, 343, 630
Azzopardi-Alexander, Marie, 272, 273,
282, 284–286, 308, 634
Baḫtiyārī, Maǧīd, 453
Baider, Fabienne, 160
Bakir, Murtadha J., 339, 343
Bakker, Peter, 7, 326, 336, 529
Ballou, Maturin M., 267
Barbarossa, 536
Aruj, 536
Hizir, 536
Name index
Barbot, Michel, 3, 85, 106
Barceló, Carmen, 226, 227
Barry, Daniel, 462, 468, 483
Barthélemy, Adrien, 94, 103
Batan, Ismail, 102
Bauer, Laurie, 554, 626
Beene, Wayne, 103, 125, 376
Beeston, Alfred F. L., 65, 352
Behnstedt, Peter, 3, 86, 88, 90, 91, 93,
98, 108, 128, 186, 197, 293, 603,
613
Bell, Alan, 562
Belnap, R. Kirk, 313
Ben Cheneb, Mohammed, 206
Benaissa, Amin, 593
Benkato, Adam, 6, 16, 90, 214, 215, 2263 ,
228, 267, 414
Benkirane, Thami, 591, 592
Benmamoun, Elabbas, 3, 146, 148, 150,
152, 309–312
Benoliel, José, 3, 216
Bergman, Elizabeth, 429
Bergsträßer, Gotthelf, 575
Berlinches, Carmen, 99
Bernabela, Roy S., 655
Bernard, Chams, 11910
Bernini, Giuliano, 644
Bettega, Simone, 8, 18, 435
Biosca, Carles, 291
Biţună, Gabriel, 138, 141, 142, 150
Bizri, Fida, 3, 337–341
Blachère, Régis, 428
Black, George Fraser, 512
Blanc, Haim, 100, 135, 136, 149, 374,
571
Blau, Joshua, 3, 37, 74, 75, 77, 282
Blau, Joyce, 467
Blažek, Václav, 423, 433
Bley-Vroman, Robert, 14
670
Blodgett, Allison, 592
Blouet, Brian, 267
Bois, Thomas, 460
Borg, Albert, 272, 273, 282, 284–288,
308, 634
Borg, Alexander, 85, 100, 103, 160, 163,
166, 167, 169, 170, 172, 267,
269, 283, 284, 287, 293, 294,
308, 615
Borjian, Habib, 449, 450
Boucherit, Aziza, 612
Bouchhioua, Nadia, 588
Boudelaa, Sami, 309
Boudlal, Abdelaziz, 591
Boumalk, Abdallah, 659
Boumans, Louis, 4, 304–306, 310, 313
Bovingdon, Roderick, 287–289
Brahimi, Fadila, 3, 414
Brauer, Erich, 372, 373
Breitbarth, Anne, 644, 661
Brincat, Joseph M., 3, 266, 267, 277,
287, 288
Brockelmann, Carl, 96
Broß, Michael, 177
Broughton, Elizabeth, 543–545
Browne, Gerald M., 419
Bruggeman, Anna, 591, 592
Brugnatelli, Vermondo, 3, 414, 657
Brunot, Louis, 3, 216
Brustad, Kristen, 122, 152, 312
Bulakh, Maria, 363
Bullock, Barbara E., 585, 586
Bulut, Christiane, 373, 396
Burdin, Rachel Steindel, 591
Bureng Vincent, George, 331
Buridant, Claude, 647
Bybee, Joan, 606, 611, 612, 614
Cadora, F. J., 570
Name index
Camilleri, Maris, 285, 286, 660
Canagarajah, Suresh A., 303
Canavan, Alexandra, 186
Cangemi, Francesco, 587
Cantineau, Jean, 250
Castellanos, Carles, 291
Caubet, Dominique, 266, 306, 612, 651,
652, 661
Čéplö, Slavomír, 6, 13, 193 , 140, 149,
167, 170, 197, 219, 221, 282,
285, 295, 3083 , 388, 627, 634,
661
Chahal, Dana, 585, 595
Chaker, Salem, 656, 661
Chalmeta, Pedro, 227, 230
Chetrit, Joseph, 216
Chikovani, Guram, 636
Chyet, Michael L., 376, 380, 387, 391,
395, 463, 476, 478, 483
Cifoletti, Guido, 3, 206, 533, 537, 540
Cinque, Guglielmo, 285
Claudi, Ulrike, 606
Clauzel, Jean, 258
Clissold, Stephen, 536
Coghill, Eleanor, 8, 13, 86, 139, 283,
376, 377, 380, 382, 384–388,
390, 393, 394, 472
Cohen, David, 253, 424, 425, 433, 435
Cohen, Marcel, 255
Colantoni, Laura, 586, 587
Colin, Georges S., 3, 206, 23116 , 237,
23827 , 292
Comrie, Bernard, 288–290, 294, 635
Cook, Albert R., 327
Corré, Alan D., 545
Corriente, Federico, 3, 94, 220, 226–
229, 232–240
Cotter, William M., 3, 11, 198, 571, 574,
575
Council of Europe, 161
Cowell, Mark W., 100, 594, 648
Croft, William, 645, 652
D’Anna, Luca, 4, 6, 207, 305, 306, 308,
309, 311, 313
D’Aranda, Emanuel, 536, 537
D’Arvieux, Laurent, 535, 537
D’Imperio, Mariapaola, 585
Daher, Nazih Y., 306
Dahl, Östen, 603, 608
Dakhlia, Jocelyne, 537
Dallet, Jean-Marie, 238
Dalli, Angelo, 287–289
Dan, Pierre, 535
Daniel, Elton L., 115, 447
Dānišgar, Aḥmad, 450, 451
Danner, Victor, 443
Daoud, Mohamed, 590
Davey, Richard J., 361, 613
Davies, Humphrey, 606
Davis, Robert C., 534
De Angelis, Pietro, 333
de Cat, Cécile, 285
de Fuentes, Alvaro Galmés, 230
De Planhol, Xavier, 116
de Prémare, André-Louis, 220, 238
de Ruiter, Jan Jaap, 4, 304–306, 310,
311
de Sacy, Silvestre, 266
de Soldanis, Giovanni Pietro Francesco
Agius, 265
Delais-Roussarie, Elisabeth, 592
Dendien, Jacques, 542
Devine, A. M., 593
Devos, Maud, 644
Dia, Alassane, 261
Diagana, Ousmane Moussa, 258
Dichy, Joseph, 75
671
Name index
Diem, Werner, 3, 61, 100, 108, 648, 649,
653, 654, 661
Diessel, Holger, 187
Dillmann, August, 65
Dimitrakou, Dimitriou, 292
Dodsworth, Robin, 191
Döhla, Hans-Jörg, 13, 284
Donner, Fred McGraw, 84
Dozy, Reinhart, 220
Drewes, A. J., 294
Drop, Hanke, 186
Drower, Ethel Stefana, 293
Dupoux, Emmanuel, 591
Eades, Domenyk, 650
Eberhard, David M., 460
Eckert, Penelope, 562
Edwards, Bela B., 461, 469
Edzard, Lutz, 352
Ehala, Martin, 304
Eilers, Wilhelm, 74
El Aissati, Abderahman, 3, 632
El Arifi, Samir, 404
El Zarka, Dina, 3, 585, 591, 594, 595
Elias, Edward, 606
Elias, Elias, 606
Elmaz, Orhan, 73
El-Sayed, Rafed, 419
Elwell-Sutton, L. P., 445
Eph’al, Israel, 39
Epps, Patience, 610
Erwin, Wallace M., 126
Extra, Guus, 305
Fabri, Ray, 277, 285
Face, Tim, 592
Fagyal, Zsuzsanna, 586
Fassberg, Steven Ellis, 96, 379, 385,
389, 391
672
Fāẓilī, Muḥammad Taqī, 450, 451, 453
Féghali, Michel, 94, 103
Fenech, Edward, 288
Ferguson, Charles A., 57, 91, 293, 568
Ferrando, Ignacio, 229, 237, 238, 240
Field, Frederick, 517
Fiorini, Stanley, 267
Fischer, Wolfdietrich, 63
Flege, James Emil, 272
Fleisch, Henri, 90
Fletcher, Janet, 584
Folmer, Margaretha L., 460
Fox, Samuel Ethan, 394
Fraenkel, Siegmund, 71, 77
Frajzyngier, Zygmunt, 188
Frank, Louis, 538, 544, 545
Friedman, Victor A., 286
Frota, Sónia, 585, 589, 592, 593
Gabriel, Christoph, 274
Gabsi, Zouhir, 590
Gafter, Roey, 567
Galmés de Fuentes, Álvaro, 230
Garbell, Irene, 381, 394, 395
Garcès, Maria Antonia, 544
Gardani, Francesco, 96, 105, 107, 275
Gardiner, Alan H., 643
Gardner-Chloros, Penelope, 544
Gasparini, Fabio, 8, 18, 357, 360, 435
Gatt, Albert, 266, 277
Gaudefroy-Demombynes, Maurice, 428
Gautier, E.-F, 404
Gazsi, Dénes, 3, 9, 86, 115–117, 443,
447, 454
Geary, Jonathan, 276
Gendrot, Cédric, 356
Gesenius, Wilhelm, 266
Giancarli, Gigio Artemio, 535
Giannakidou, Anastasia, 646, 648
Name index
Gibson, Maik, 3, 198, 266, 273
Giles, Howard, 304, 553
Gili Fivela, Barbara, 590
Gilson, Erika H., 541
Glaß, Dagmar, 57
Gonzo, Susan, 306, 307, 312, 315
Goodwin, Stefan, 267
Gordon, Elizabeth, 551, 553
Gośam, 40
Graf, David F., 49
Gralla, Sabine, 91
Grandchamp, Pierre, 542
Grech, Paul, 266
Greenblatt, Jared R., 389
Grice, Martine, 274, 589, 592
Grigore, George, 95, 119, 138, 141, 142,
148, 150
Grosjean, François, 610
Grotzfeld, Heinz, 574
Guichard, Pierre, 225, 227, 230
Gulle, Ozan, 160, 167, 168
Gurlekian, Jorge, 586, 587
Gussenhoven, Carlos, 591, 593
Gutas, Dimitri, 72, 78
Gzella, Holger, 40, 43, 49, 69
Haase, Martin, 607
Hachimi, Atiqa, 571
Haedo, Diego de, 535, 537, 543–545
Haeri, Niloofar, 568, 571
Haig, Geoffrey L. J., 3, 147, 459, 462,
463, 469, 472–474, 477, 479
Halasi-Kun, Tibor, 104
Hall, Robert A. Jr, 533
Hamid Ahmed, Mohamed-Tahir, 420,
421, 423
Hamza, Belgacem, 414
Hansen, Anita Berit, 590
Hansen, Maj-Britt Mosegaard, 605, 647,
648
Harrell, Richard S., 266
Haser, Verena, 193
Haspelmath, Martin, 287, 525, 606, 659
Hassan, Jidda, 177, 178, 183, 192
Hassan, Qasim, 12522
Hassan, Zeki Majeed, 576
Hayajneh, Hani, 39, 41, 49, 73
Heath, Jeffrey, 6, 198, 199, 205, 214–
216, 218, 221, 222, 2278 , 239,
251, 258, 261, 269, 591, 615
Hebbo, Ahmed, 3, 77
Heim, Irene, 651
Heine, Bernd, 178, 187, 605–607, 611,
612, 614, 631–633, 657
Hellmuth, Sam, 11, 274, 585, 588–590,
592–595
Herin, Bruno, 3, 10, 18, 273, 489, 508,
512, 526, 558, 561, 571, 575,
590
Herodotus, 49
Heselwood, Barry, 357, 576
Hoberman, Robert D., 375
Hobrack, Sebastian, 389
Holes, Clive, 84, 85, 94, 118, 571, 594
Hopkins, J. F. P., 544
Hopkins, Simon, 86, 99
Hopper, Paul, 604
Horesh, Uri, 3, 12, 553, 567, 571, 574,
575
Hoyland, Robert G., 352
Hudson, Richard A., 423
Huehnergard, John, 48, 50, 61, 65
Hutchison, John, 182, 188
Ibn al-Bayṭār, 23827
Ibn Ḥayyān, 2266
Ibn Khaldūn, 37
673
Name index
Ibn Muǧāhid, Abū Bakr, 70, 76
Ibrahim, Iman, 471
Ibrahim, Muhammad H., 568
Ibrahim, Zeinab, 68
Iemmolo, Giorgio, 284
Il-Hazmy, Alayan Mohammed, 613
Imām Ahwāzī, Sayyid Muḥammad ʕAlī,
451
Ingham, Bruce, 3, 116, 119, 127, 129,
130, 147, 447
Iraqui Sinaceur, Zakia, 220
Isaksson, Bo, 138, 154
Israel, Felice, 40
Īzadpanāh, Ḥamīd, 449–451
Jacquart, Danielle, 75
Jacques-Meunié, Djinn, 258
James, Boris, 460, 461
Jastrow, Otto, 3, 86, 93, 101, 104, 135–
143, 147–150, 152, 155, 372,
374, 375, 383, 386–390, 393,
394, 396, 461, 469, 615
Jeffery, Arthur, 3, 59, 65, 70, 77
Jenkins, Edward V., 327
Jeremiás, Éva M., 441, 445
Jespersen, Otto, 644
Johanson, Lars, 13
Johnson, Mark, 193
Johnstone, T. M., 654
Joseph, Brian D., 396
Jun, Sun-Ah, 584, 591, 592
Kahane, Henry Romanos, 85, 533, 534
Kahane, Renée, 533, 534
Kahn, Margaret, 462, 464, 467, 468
Kahrel, Peter, 645
Kariolemou, Marilena, 160
Kaufman, Terrence, 13, 552, 585, 631
Keane, G. J., 327
674
Keesing, Roger, 638
Kent, Roland, 460
Kerswill, Paul, 551, 573
Khan, Geoffrey, 283, 286, 293, 381, 382,
385, 388, 390, 392
Khattab, Ghada, 576
Kherbache, Fatma, 413
Kireva, Elena, 274
Klimiuk, Maciej, 271
Knecht, 541
Kogan, Leonid, 352
König, Ekkehard, 632
Koontz-Garboden, Andrew, 463
Kootstra, Fokelien, 46
Kossmann, Maarten, 23824 , 23828 , 23933
Kossmann, Maarten, 9, 66, 200–204,
209, 237, 406, 408, 410–414,
632, 657
Kramer, Raija, 193
Krapova, Iliyana, 285
Krebernik, Manfred, 85
Krier, Fernande, 3, 294
Kropftisch, Lorenz, 75
Kruse, Friedrich, 512
Kulk, Friso, 594
Kuteva, Tania, 178, 187, 607, 611, 612,
614, 631, 657
Labarta, Ana, 226, 227, 236
Labov, William, 4, 562
Ladd, D. Robert, 584, 585
Ladefoged, Peter, 356
Ladusaw, William A., 646
Lafkioui, Mena, 3, 12, 203, 657, 659
Lahdo, Ablahad, 138, 143–146, 148, 150,
154, 287
Laka, Itziar, 648
Lakoff, George, 193
Lanly, André, 539
Name index
Larcher, Pierre, 44
Lash, Elliott, 12, 15, 644, 653, 654
Lass, Roger, 559
Law, Daniel, 608
Le Page, R. B., 554, 555, 572, 610
Leddy-Cecere, Thomas, 3, 11, 1872 , 567,
610, 617, 618, 631, 657
Lefebvre, Claire, 635
Lehmann, Christian, 605, 606, 609
Leitner, Bettina, 6, 15, 22, 84, 446, 447,
660
Lentin, Jérôme, 7, 86, 94, 99, 226, 394,
461, 574, 613
Leslau, Wolf, 69
Levin, Aryeh, 574
Lewin, Bernhard, 613
Lichtenberk, Frantisek, 645
Lipiński, Edward, 293
Lipski, John, 551
Lonnet, Antoine, 3, 352, 356–358, 362,
364, 366
Loporcaro, Michele, 272
Lucas, Christopher, 6, 12–15, 17, 18,
193 , 22, 24, 61, 88, 89, 118, 124,
126, 130, 140, 149, 167, 170,
193, 197, 203, 205, 219, 221,
232, 268, 307, 3083 , 310, 312,
315, 322, 355, 388, 412, 444,
482, 567, 583, 627, 634, 644,
648, 649, 653, 654, 657, 660,
661
Luffin, Xavier, 3, 323, 324, 329, 331,
332, 335, 336, 343
Lydon, Ghislaine, 617
Maas, Utz, 591
Macalister, Robert Alexander Stewart, 490, 513
Macdonald, Michael C. A., 39, 51
MacKenzie, David N., 388, 394–396,
446
MacKinnon, Colin, 449
Maclean, Arthur John, 379
Macúch, Rudolf, 293
Maddieson, Ian, 356
Magidow, Alexander, 85
Mahdi, Qasim R., 611
Maiden, Martin, 280, 281
Majidi, Mohammad-Reza, 119, 121, 124,
126
Malaika, Nisar, 129
Mammeri, Mouloud, 405
Manfredi, Stefano, 1, 4, 11, 23, 176, 215,
282, 284, 286, 308, 321, 322,
325, 326, 328–331, 336, 4201 ,
423, 427, 499, 567, 609, 630,
633, 634, 657
Manzano Moreno, Eduardo, 230
Marçais, Philippe, 202, 203
Masliyah, Sadok, 93, 94, 97
Matras, Yaron, 1, 3, 7, 10, 13, 18, 116,
117, 119, 120, 122, 128, 129, 131,
208, 314, 383, 463, 480, 490,
493, 494, 507, 508, 512–514,
516, 517, 524, 525, 529, 587,
618, 657, 658
Mattoso, Joaquim, 551
McMahon, April, 7
Mechehed, Djamel-Eddine, 407
Meillet, Antoine, 643
Meisel, Jürgen, 14
Meldon, J. A., 327
Menéndez Pidal, Ramón, 230
Mengozzi, Alessandro, 377
Mennen, Ineke, 585
Meouak, Mohamed, 414
Mettouchi, Amina, 633
Meyerhoff, Miriam, 638
675
Name index
Mifsud, Manwel, 3, 269, 276–280, 287,
288, 294
Miller, Catherine, 3, 315, 322, 328, 330,
334, 343, 603
Minervini, Laura, 533, 534
Mingana, Alphonse, 70
Mion, Giuliano, 198
Mohammadi, Ariana Negar, 129
Moi, Daniel R., 333
Monteil, Charles, 258
Monteil, Vincent-Mansour, 261
Moravcsik, Edith A., 517
Mori, Laura, 534
Morin, Didier, 420, 423
Morris, Miranda J., 353, 354, 359, 366
Morton, Rachel, 585
Mouili, Fatiha, 404
Mourigh, Khalid, 410, 411, 414, 415
Muʕāwiya, 159
Mughazy, Mustafa, 652
Muraoka, Takamitsu, 95
Muraz, Gaston, 323
Múrcia Sánchez, Carles, 403
Mutzafi, Hezy, 379, 383, 389
Myers-Scotton, Carol, 208
Naciri-Azzouz, Amina, 201
Næss, Unn Gyda, 339, 340
Naǧībī Fīnī, Bihǧat, 450, 451, 453
Naïm, Samia, 574
Naït-Zerrad, Kamal, 270, 292, 293
Nakano, Aki’o, 186
Nakao, Shuichiro, 329, 330, 332–336,
342, 343, 593, 630
Nance, Claire, 586
National Statistics Office, 268
Naumann, Christfried, 409
Naumkin, Vitaly, 356
Nebel, Arthur, 332
676
Nehmé, Laïla, 43, 50, 51
Neishtadt, Mila, 3, 107
Nevo, Moshe, 137
Newbold, F. R. S, 512
Newton, Brian, 3, 166–168, 170
Nolan, Joanna, 10, 85, 199, 214, 545
Nöldeke, Theodor, 69, 77, 90, 396
Noorlander, Paul M., 472
Norde, Muriel, 606
Nortier, Jacomine, 222
Nyberg, Henrik Samuel, 43
Nycz, Jennifer, 569
O’Rourke, Erin, 274, 586
Oberling, Pierre, 115, 116, 446
ʕObodat, 40
Ondráčková, Zuzana, 293
Onour, ʕAbdallah, 420
Öpengin, Ergin, 9, 139, 147, 387, 459–
464, 473, 474, 477
Ould Mohamed Baba, Ahmed-Salem,
259
Overlaet, Bruno, 46
Owen, H. R., 327
Owen, Roger, 635
Owens, Jonathan, 3, 4, 6, 140, 176–
178, 180, 181, 183, 191, 192, 308,
3083 , 316, 323, 325, 327–329,
334, 628–630
Paciotti, Luca, 261
Palva, Heikki, 6, 88, 93, 97, 108, 148
Pananti, Filippo, 537, 542, 545
Panza, G., 333
Papaconstantinou, Arietta, 592
Paradisi, Umberto, 412, 657
Parkinson, Dilworth B., 68
Paspati, Alexandre, 512
Pat-El, Na’ama, 49, 96, 121, 649
Name index
Patai, Raphael, 372, 373
Patkanoff, K. P, 512
Paul, Ludwig, 119, 446, 468
Pellegrini, Giovanni Battista, 539
Pennell, Richard, 541, 542
Penny, Ralph, 551
Pepper, Stephen, 626, 627
Pereira, Christophe, 198
Perry, John R., 95, 445, 461, 478
Persson, Maria, 360
Peust, Carsten, 593
Piccitto, Giorgio, 283
Pichler, Werner, 405
Pihan, Antoine Paulin, 542
Plantet, Eugène, 536
Poiret, Jean Louis Marie, 542
Poplack, Shana, 314
Porkhomovsky, Victor, 356
Pott, August F., 512
Power, Timothy, 617
Pozzati, Aurelio, 333
Prieto, Pilar, 585, 593
PRIO, 162
Procházka, Stephan, 6, 16, 74, 78, 84,
85, 90, 94, 97, 99, 102, 103,
116, 118, 119, 121, 135–137, 139,
142, 146, 154, 160, 206, 591,
613
Procházka-Eisl, Gisela, 96, 145
Qafisheh, Hamdi, 613
Qayno, 40
Queen, Robin M., 585
Quint, Nicolas, 12
Rabeh, 178, 180, 323
Rabhi, Allaoua, 656, 657, 659
Rabin, Chaim, 63, 76
Rabinowitz, Isaac, 40
Rahmani, Hamed, 591
Ramat, Paolo, 605, 644
Raqōś bint ʕAbd-Manōto, 42, 43, 46,
47
Rastegar-El Zarka, Dina, 592
Ratcliffe, Robert R., 3, 64, 65, 125, 309,
314, 635, 636
Rāzī, Farīda, 445
Rehbinder, Johan von, 542
Reinhardt, Carl, 615
Reinkowski, Maurus, 105
Retsö, Jan, 3, 39, 69, 70, 78, 86, 103,
178, 352
Rickford, John R., 562
Ridouane, Rachid, 356
Rifaat, Khalid, 592
Rilly, Claude, 419
Ripper, Thomas, 461
Ritt-Benmimoun, Veronika, 190, 281,
629
Rizgar, Baran, 380
Robin, Christian-Julien, 45, 352
Robustelli, Cecilia, 280, 281
Roger II of Sicily, 267
Romaine, Suzanne, 307–309, 315
Roper, E. M., 423
Rosenhouse, Judith, 574, 612
Ross, Malcolm, 13, 609, 625, 626, 631,
636
Rossetti, Roberto, 538
Roth, Arlette, 163, 166, 169–171
Rouchdy, Aleya, 4, 12, 304, 306, 316
Rubin, Aaron D., 48, 50, 65, 99, 352,
605
Russi, Cinzia, 286
Ryding, Karin C., 58, 470
Saada, Lucienne, 614
Saade, Benjamin, 276
677
Name index
Sabar, Yona, 377, 381, 385, 386, 389,
390, 395
Sabuni, Abdulghafur, 91, 92
Ṣādeqī, ʕAlī Ašraf, 454
Sadler, Louisa, 286, 660
Sakel, Jeanette, 88, 480
Salāmī, ʕAbdulnabī, 449, 451
Saltarelli, Mario, 306, 307, 312, 315
Sammut, Carmen, 290
Sánchez, Pablo, 233, 614
Sankoff, Gillian, 303
Sarlak, Riẓā, 449–451
Ṣarrāfī, Maḥmūd, 449, 450, 453
Sasse, Hans-Jürgen, 626
Savary de Brèves, François, 535
Sayahi, Lotfi, 3, 203, 206, 209, 215, 222,
590
Schilling-Estes, Natalie, 615
Schmitt, Rüdiger, 442
Schneider, Edgar W., 554
Schreier, Daniel, 551
Schroepfer, Jason, 614
Schuchardt, Hugo, 539, 541, 542, 545
Sciriha, Lydia, 268
Seeger, Ulrich, 12, 308, 649
Seetzen, Ulrich Jasper, 512, 513
Segal, Judah B., 395
Seifart, Frank, 107, 275
Selbach, Rachel, 537, 542
Serreli, Valentina, 404
Shabibi, Maryam, 3, 116, 117, 119, 120,
122, 128, 129, 131
Shahin, Kimary, 571, 573–575
Silverstein, Michael, 610
Sima, Alexander, 364
Simeone-Senelle, Marie-Claude, 64, 353,
359, 362, 366, 425, 650
Simonet, Miquel, 586
Singer, Hans-Rudolf, 94
678
Sinha, Jasmin, 373
Sirkeci, Ibrahim, 460
Skjærvø, Prods Oktor, 441
Smart, Jack R., 336, 341
Sokoloff, Michael, 293
Sorace, Antonella, 587
Souag, Lameen, 3, 9, 13, 98, 199–201,
203, 206, 209, 215, 286, 293,
407, 409–415, 635, 636, 649,
656, 657, 659
Spagnol, Michael, 275, 286, 288–290,
294, 660
Stassen, Leon, 494
Stein, Peter, 39, 51, 64
Stephens, Laurence D., 593
Stevenson, R. C., 188
Stewart, Devin, 618
Stilo, Donald, 449, 451
Stokes, Phillip W., 122
Stolz, Thomas, 6, 294
Sudbury, Andrea, 551
Sweetser, Eve, 193
Tabouret-Keller, Andrée, 554, 555, 572,
610
Tadmor, Uri, 287, 412, 525
Taïfi, Miloud, 239
Taine-Cheikh, Catherine, 6, 201, 202,
216, 245, 252, 253, 255–257,
260, 261, 266, 613, 627, 628,
632
Talay, Shabo, 93, 138, 140, 142, 148,
150–152, 155, 374
Talmoudi, Fathi., 3, 202, 205, 612
Taylan, Eser, 3, 152, 153
Terés Sádaba, Elías, 225
Teyran, Feqiyê, 468
Thackston, Wheeler M., 381
Name index
Thomason, Sarah G., 1, 13, 138, 552,
585, 625, 631
Thordarson, Fridrik, 446
Tigziri, Nora, 404
Tinniswood, Adrian, 536
Todd, Terry L., 474
Tosco, Mauro, 1, 321, 323, 325, 326,
336
Traugott, Elizabeth, 604
Tropper, Josef, 49
Trudgill, Peter, 1, 10, 98, 322, 551, 553,
554, 558, 568, 569, 572, 573
Tsabolov, Ruslan, 3, 472–474, 483
Tsiapera, Maria, 3, 163, 166, 168
Tully, Miss, 541, 543
Türkmen, Erkan, 206
Ullendorf, Edward, 425
Ullmann, Manfred, 282
ʕUmar ibn Ḥafsūn, 2266
University of Zaragoza, 283, 287, 311
Ursini, Flavia, 540
Utas, Bo, 460, 463
Van Coetsem, Frans, 17, 21–24, 138,
177, 200, 205, 207, 269, 307,
315, 322, 330, 355, 443, 482,
626
van den Boogert, Nico, 405, 407, 414
van der Auwera, Johan, 643, 644, 653
van der Wal Anonby, Christina, 3, 655,
656
van Putten, Marijn, 3, 7, 9, 16, 50, 63,
409, 411, 414, 443
Vanhove, Martine, 9, 15, 267, 270, 282,
287, 294, 421–424, 429–431,
435, 612, 634
Vassalli, Michelantonio, 266
Vassallo, Mario, 268
Vella, Alexandra, 274
Versteegh, Kees, 1, 8, 125, 136, 150, 278,
306, 308, 352
Vicente, Ángeles, 3, 7, 206, 214, 227,
229, 233, 240, 306
Visconti, Jacqueline, 605
Vocke, Sybille, 142
Vollers, Karl, 606
Vossen, Frens, 644, 653
Vycichl, Werner, 409
Waldner, Wolfram, 142
Walker, Traci, 587
Walter, Mary Ann, 6, 20, 22, 83, 854 ,
136, 140, 149, 168, 271, 272,
287, 3083
Walters, Keith, 590
Warrington, 544
Watson, Janet C. E., 3, 4, 45, 352, 357,
360–362, 364, 366, 591, 593,
613, 614, 650
Wedekind, Klaus, 420, 422, 435
Weinreich, Uriel, 14, 585
Weiss, Gillian, 536
Wellens, Inneke, 327, 329–331, 333–
335
Wells, John C., 584
Wellsted, James R., 351
Weninger, Stefan, 3, 64, 65, 73, 108
Whorf, Benjamin Lee, 75
Williams, Ann, 551, 573
Willis, David, 644, 646, 661
Wilmsen, David, 17, 66, 627, 649, 662
Winford, Donald, 22, 119, 205, 322, 342,
355, 629, 634
Wittrich, Michaela, 138, 142
Wohlgemuth, Jan, 383, 433
Woidich, Manfred, 96, 128, 186, 271,
574, 603, 627
679
Name index
Wolfram, Walt, 615
Woodhead, Daniel R, 103, 125, 376
Yardeni, Ada, 40
Yoda, Sumikazu, 3, 199
Yūsifī, Ɣulāmḥusain, 444
Zaborski, Andrzej, 431
Zagona, Karen, 286
Záhořík, Jan, 421
Zammit, Martin R., 3, 273, 352, 590
Zeyneloğlu, Sinan, 460
Ziagos, Sandra, 3
Ziamari, Karima, 205, 209
Zibelius-Chen, Karola, 419
Zuckermann, Ghil’ad, 638
Zwettler, Michael J., 49
680
Language index
Acholi, 327, 328, 332–336
Afar, 424, 433
Afro-Asiatic, 245, 254, 255, 292, 323,
371, 419, 425, 433, 434
Agaw, 424, 433, 434
Akkadian, 7, 39, 613 , 85, 853 , 293
Alur, 327
Amharic, 434, 645
Arabic (Ar.)
Abu Dhabi, 613
Aleppo, 84, 92, 9220 , 94
Algiers, 198, 255, 612
Amman, 11, 572–577, 658
Anatolian, 101, 163, 287, 374, 3744 ,
383, 38316 , 388, 469
Andalusi, 7, 165, 198, 214, 217, 218,
220, 283, 284, 287, 311
Aswan, 614
Āzəḫ, 138, 142, 143, 150, 151, 374,
38316
Baggara, 179, 633, 637
Baghdad, 84, 88, 89, 93, 97, 99,
105, 108, 12522 , 12931 , 1351 , 197,
266, 3744 , 3766 , 381, 38114 , 386,
396, 5712 , 660
Bahrain, 116, 1199 , 453, 572, 577,
660
Basra, 116, 12522 , 611
Bedouin, 11, 84, 87–89, 9119 , 92–
94, 97, 98, 100, 10234 , 103, 106–
108, 116, 118, 136, 214, 228,
248, 249, 2503 , 374, 3744 , 407,
408, 570
Beirut, 84, 90, 492, 574
Bongor, 322–325, 342, 343
Bukhari, 314
Cairo, 63, 9629 , 269, 312, 408, 574,
611, 648, 651, 652
Casablanca, 197, 266, 269, 552, 612
Central Asian, 3, 12, 136, 635–637
Chadian (ChA), 179, 323, 324
Cilician, 831 , 90, 9220 , 97, 99, 101,
102, 10234 , 103, 105, 10538 , 613
Classical (CA), 3, 7, 8, 10, 37, 38,
403 , 42, 45, 50, 89, 99, 103,
119, 11910 , 1351 , 151, 177, 1873 ,
197, 216–219, 226, 228, 22912 ,
246, 248–250, 254, 255, 260,
266, 27610 , 282, 305, 313, 407,
425, 427–429, 443–445, 452,
4635 , 631, 648, 6609
Cypriot Maronite (CyA), 3, 20,
22, 136, 1405 , 271, 283, 284,
287, 3083 , 615
Damascus, 84, 95, 96, 98–100, 103,
105, 1873 , 491, 492, 500, 574,
594, 613, 648
Daragözü, 138, 149
Dhofar, 613
Diyarbakır, 135, 153, 154
Eastern, 6, 22812 , 2503 , 267
Egyptian (EA), 11, 16, 61, 66, 179,
182, 186, 187, 1873 , 189–191,
Language index
193, 27610 , 312–315, 327, 337,
338, 341, 543, 585, 592, 593,
606, 6061 , 627, 6524
Gaza, 11, 570, 571, 575
Ghamdi, 576
Gulf (GA), 8, 9422 , 9629 , 115, 116,
12825 , 130, 180, 337–339, 360,
361, 442, 446, 447, 449, 453,
651, 655
Gulf Pidgin, 336–343
Ḥapəs, 138, 142
Hasköy, 138, 1404 , 142
Ḥassāniyya, 6, 201, 202, 214, 216,
627, 628, 632, 637, 656
Mali, 216, 250, 251, 256, 258,
260, 615
Ḥiǧāzi, 59, 592–594
Iraqi (IA), 83, 88, 89, 91, 93, 96,
9628 , 100, 101, 10335 , 104, 107,
116, 117, 1174 , 118, 12418 , 12521 ,
12624 , 283, 284, 337, 338, 374–
376, 3766 , 378, 380–383, 385–
387, 390, 391, 393, 395, 397,
447
Janaybi, 355, 356, 359, 362
Jebel Ansariye, 613
Jerusalem, 95, 513, 514, 517, 524,
525, 559, 612, 657, 658
Jordanian (JA), 11, 310, 337–339,
497–499, 556–558, 55814 , 559,
55916 , 560–563, 572, 573, 575,
577, 585
Jordanian Pidgin, 8, 336–339, 341,
342
Juba, 7, 23, 308, 315, 322, 326–
334, 3345 , 335, 336, 342, 343,
593, 630, 634, 635
Khuzestan, 6, 15, 22, 84, 100, 442,
446, 447, 451, 660
682
Kinderib, 138, 148
Kinubi, 7, 322, 326–336, 342, 343
Kenyan, 326, 327, 329, 331, 335,
336
Kozluk, 135, 136, 143
Kozluk–Sason–Muş, 140, 142, 147,
152
Kurdistan, 101
Lebanese, 17, 89, 92, 94, 170, 310,
337, 492, 493, 574
Levantine, 3, 61, 66, 67, 163, 165,
168, 267, 283, 284, 293, 310,
492, 493, 499, 501, 504, 560,
574, 613, 615
Libyan (LA), 198, 201, 206, 20615 ,
207, 208, 593
Maghrebi, 6, 90, 217, 218, 226, 2263 ,
228, 22810 , 22812 , 233, 248,
2514 , 266, 267, 269, 273, 27610 ,
287, 290, 294, 406, 407, 413,
632, 633, 659, 6598 , 661
Mardin, 135, 136, 138, 139, 142, 143,
147, 150, 153, 154, 374, 3744 ,
461, 462, 469, 477, 480, 482
Marrakech, 614
Mecca, 572, 576
Mesopotamian, 3, 6, 10, 116, 119,
135, 136, 1417 , 374, 3744 , 3766 ,
386, 393, 447, 615
Modern Standard (MSA), 3, 7, 16,
17, 21, 99, 117, 119, 11910 , 12215 ,
129, 1653 , 177, 180, 181, 209,
222, 246, 248, 250–252, 256–
258, 261, 269, 305, 361, 379–
381, 386, 390, 392, 404, 444,
4442 , 445–447, 449–453, 481,
489, 594, 631, 648
Moroccan (MA), 6, 11, 16, 75, 100,
198, 1994 , 2008 , 202, 20313 ,
Language index
20614 , 209, 213, 222, 2278 , 22812 ,
233, 238, 23825 , 23826 , 23827 ,
23830 , 247, 256, 261, 305, 310,
3107 , 313, 406, 552, 585, 586,
589, 591, 592, 632, 651, 652,
6577
Mosul, 86, 878 , 92, 104, 107, 381,
390, 393, 460, 462, 482, 615
Mutki-Sason, 138
Muş, 135, 136
Najdi, 552, 572, 573
Nigerian, 6, 12, 1405 , 3083 , 628–
630
Old, 3, 38, 39, 403 , 41, 46–49, 51,
7319 , 90, 91, 9119 , 92–95, 99,
12215 , 140–143, 168, 228, 232,
233, 287, 469, 470
Omani, 352, 355, 357, 358, 594,
615, 650, 651
Palestinian, 10, 12, 107, 310, 39422 ,
514, 517, 518, 526, 557, 559,
55916 , 560–562, 570, 572–575,
577, 649, 657–659
Pidgin Madame, 8, 336–341
Post-Hilalian, 216, 219
Pre-Hilalian, 216–219, 221
Romanian Pidgin, 8, 336–339, 341
Sana’a, 613, 614, 652
Sason, 135, 137, 140, 142–146, 148,
149, 14914 , 150–152, 15215 , 153
Sfax, 198, 267
Sicilian, 198
Šiḥḥī, 651, 655, 656
Siirt, 135, 136, 138, 14812 , 153, 154,
372, 374, 461, 469
Soukhne, 93, 97, 98, 613
Sousse, 198, 612
Sudanese (SA), 9, 176, 326–331,
422, 423, 425, 427–429, 630,
635, 6524
Sudanic pidgins, 7, 321, 322
Syrian, 83, 89, 92, 94, 100, 10335 ,
107, 12418 , 135, 137, 310, 393,
466, 492, 493, 594, 645
Tillo, 138, 143, 144, 14410 , 145, 146,
150, 153, 15316 , 154, 287
Tlemcen, 198, 266
Tozeur, 614
Tunis, 198, 203, 273, 588, 589
Tunisian (TA), 6, 11, 75, 9422 , 190,
191, 201, 202, 266, 267, 273,
281, 305, 306, 308, 311, 312,
541, 588–592
Uzbekistan, 100, 152, 182, 3083 ,
445
Western, 6, 7, 225, 266
Western Sudanic (WSA), 176, 179,
188, 323, 327, 633
Yemeni, 45, 352, 421, 594, 613, 614,
616, 650, 651, 6524
Aragonese, 227
Aramaic (Aram.), 6, 7, 16, 39, 392 , 40–
44, 447 , 45, 46, 48–51, 58, 59,
61, 62, 66, 68–70, 7013 , 7015 ,
71, 72, 74, 77, 78, 85, 86, 88–
91, 93, 94, 9422 , 95, 96, 99,
9931 , 100, 103, 104, 107, 108,
116, 138, 1383 , 139, 141, 148,
153, 15316 , 154, 163, 172, 282,
284, 293, 460, 470, 472
Christian Palestinian, 7015
Eastern, 139
Imperial, 38, 40
Jewish Babylonian (JBA), 293, 605
Mandaic, 293
Middle, 3
Mlaḥso, 371
Nabataean, 38, 43, 49, 50, 61
683
Language index
Neo-Aramaic, 3, 86, 101, 139, 293,
467, 469
North-Eastern (NENA), 8, 9, 16,
139, 371–376, 3766 , 377–388,
390, 39018 , 39019 , 391–397
Western, 100, 108
Syriac (Syr.), 69, 70, 7015 , 7016 , 71,
72, 7218 , 77, 90, 95, 99, 104,
139, 142, 153, 163, 372, 3721 ,
379, 386, 387, 396
Ṭuroyo, 142, 148, 153, 371, 372, 374,
394
Western, 86
Armenian, 12, 85, 137, 138, 140, 161,
164, 171, 259, 3721 , 467, 469
Arwad, 16, 90
Avokaya, 327, 328, 335
Azer, 258, 259
Azeri, 87, 3732 , 374, 375, 381, 394–396,
513
Bagirmi, 176, 179, 188, 189, 323, 325
Baka, 327, 328, 335
Balochi, 442
Bari, 23, 327–331, 334, 3345 , 335, 336,
630, 635
Basque, 607
Beja, 9, 15
Belanda Bor, 327, 328, 333, 336
Bengali, 337, 339, 341
Berber, 3, 6, 7, 9, 11, 12, 16, 22, 669 ,
199–201, 20110 , 20111 , 2019 , 202,
203, 20313 , 204, 205, 20514 ,
206, 208, 209, 214–219, 221,
225, 229–231, 23318 , 234, 236,
237, 23721 , 23722 , 238, 23824 ,
23826 , 23827 , 23828 , 23830 , 239,
23931 , 240, 245, 247–251, 2514 ,
252–261, 270, 27610 , 292, 293,
684
305, 587, 590–592, 594, 627,
628, 632, 633, 653, 656, 657,
659, 6598 , 661
Awjila, 9, 411, 412, 656, 657
Beni Snous, 413
Ghomara, 201, 2019 , 237, 410, 411,
413–415
Kabyle, 201, 202, 204, 238, 23830 ,
239, 257, 403–405, 408, 410,
413, 632, 633, 656, 6577 , 659
Mozabite, 659
Northern, 408, 413, 414
Senhaja, 237
Shawiya, 403, 659
Siwa, 27610 , 404, 405, 407, 412,
415, 656, 657, 659
Tamazight, 215, 305, 403, 659
Tarifiyt, 215, 305, 403, 412, 413,
632, 6577 , 659
Tashelhiyt, 203, 215, 23822 , 247,
305, 403, 413, 656, 659
Tuareg, 9, 247, 254, 257, 403, 405,
409, 413, 656
Tunisian (TB), 590, 591
Zenaga (Zen.), 245, 247–249, 2492 ,
250–253, 2535 , 254, 255, 257,
259–261, 627, 628, 632, 656
Zenati, 234, 236, 237, 412
Zuwara, 410, 590
Bitlis, 154
Bongo, 327, 328, 333–336
Canaanite, 39, 49
Castilian, see Spanish
Catalan, 216, 227, 291
Majorcan, 586
Chadic, 176, 189, 199, 209, 323
Chaoui, see Shawiya
Circassian, 85
Language index
Coptic, 11, 419, 588, 592, 593, 6431 , 653–
655
Cushitic, 9, 419, 423–425, 427, 431, 433–
435
Dadanitic, 38, 46, 49
Dardic, 490, 491, 512
Didinga, 327, 335
Dinka, 327–329, 332, 334–336
Domari, 10, 12, 18, 653, 657, 658
Antioch, 501, 502
Beirut/Damascus, 490–504, 5043 ,
505, 507, 508
Jerusalem, 490, 494
Jordanian, 515
Northern, 512
Southern, 10, 490, 491, 495, 498,
508, 512
Dutch, 305, 306
Egyptian (Ancient), 419, 434, 588, 592,
593, 653
English, 6, 14, 19, 75, 87, 89, 92, 103,
106, 116, 162, 163, 177, 180, 181,
216, 221, 248, 259, 265, 268,
269, 272, 274, 275, 277, 279,
287, 289, 28917 , 290, 29018 ,
291–293, 304, 306, 312, 327,
328, 337, 338, 340, 341, 375,
377, 504, 551, 5511 , 568, 586,
588, 605, 6061 , 626, 627, 637,
644, 646, 647
American, 221, 306, 615
British, 306
Early Modern, 644
Maltese, 268
Middle, 644
New Zealand, 553, 5535
Old, 644
Fali, 188, 189, 193
French, 19, 74, 75, 7522 , 87, 89, 92, 106,
164, 180, 199, 201–203, 205,
20514 , 206, 20614 , 208, 209,
214, 216, 217, 219–221, 247,
248, 251, 258, 2596 , 260, 275,
292, 306, 323–325, 375, 404–
406, 533–535, 538, 539, 541–
543, 545, 551, 5511 , 586, 590–
592, 605, 6431 , 644, 646–649,
660, 66010
Old, 644
Fulfulde, 176, 177, 179, 180, 191, 192,
323, 325, 629
Gaelic, 586
German, 14, 75, 398, 586, 647
Greek (Gk.), 7, 10, 11, 23, 38, 42, 44,
47, 48, 58, 59, 66, 68, 72, 7217 ,
7218 , 74, 77, 78, 85, 854 , 88,
104, 10436 , 160–173, 199, 215,
282, 287, 292, 419, 533, 534,
542, 588, 592, 593
Ancient, 593
Cypriot, 6, 160–162, 1622 , 163, 165–
167, 169–172
Koiné, 593
Standard, 162, 1622 , 163, 168
Gurbetça, see Kurbetça
Gurānī, 442, 446
Gīlakī, 442
Gəʕəz (Gz.), 45, 65, 68, 69, 6911 , 71, 72,
7217 , 74
Ḫānsāri, 449
Hasaitic, 46, 4611 , 48
Hausa, 6, 177, 180, 181, 191, 199, 20615 ,
209, 323, 325, 404
685
Language index
Hebrew, 3, 12, 39, 49, 50, 635 , 69, 73,
75, 77, 168, 199, 216, 22914 ,
553, 638
Hindi, 339, 341, 342
Hismaic, 38, 49, 50
Iberian, 216, 22913
Igli, 404, 406, 411
Indo-Aryan, 441, 490, 493, 511, 512,
514–518, 522, 524–527, 529,
657
Indo-Iranian, 3, 9, 137, 518, 656
Indonesian, 8, 339
Iranian, 9, 86, 116, 137, 139, 373, 383,
396, 459, 460, 463, 464, 466,
470, 472, 473, 478, 491, 502,
513, 514, 655, 656
Italian, 7, 199, 201, 206, 207, 20716 , 20717 ,
20718 , 208, 268, 272, 274–279,
27912 , 280–282, 286, 287, 289–
293, 305, 306, 312, 533–535,
537, 5371 , 538–545, 586, 590,
627, 634, 647, 661
Bari dialect, 589, 590
Javanese, 339
Jibbāli, see Śḥerɛt
Jur, 327, 328, 332, 333, 335, 336
Kakwa, 327, 334, 335
Kanembu, 176
Kanuri, 6, 176–178, 180–182, 188, 189,
191, 192, 323, 325, 628–630,
637
Kirmānī, 442
Koalib, 12, 633
Kotoko, 176, 180, 192, 629
Kuku, 327
Kumzari, 442, 463, 653, 655, 656
Kurbetça, 164
Kurdish (Kr.), 6, 9, 83, 86, 88, 91, 92,
95–97, 101, 104, 107, 118, 137,
1372 , 138, 139, 141, 142, 147–
153, 15316 , 154, 372, 3721 , 373,
3733 , 374–379, 37912 , 380, 381,
383–386, 38717 , 388, 389, 39018 ,
391, 394, 395, 397, 442, 446,
490–493, 501, 502, 507, 513
Central, 459, 463–467, 472, 473,
475, 481
Kurmanji, 139, 152, 3733 , 3766 , 459,
462, 4634 , 468, 470–472, 474,
477, 478, 480–482
Badini, 471, 480
Northern, 3745 , 3766 , 379–381, 383–
386, 38717 , 388, 391, 395, 396,
459, 463–467, 472–474, 476–
478, 480, 481
Sorani, 395, 396, 459, 462, 4634 ,
474, 477, 478, 481, 513
Latin, 3, 19, 38, 66, 72, 74, 1994 , 213,
218, 220, 229, 22913 , 231, 23217 ,
233, 236, 238, 239, 259, 291,
304, 311, 405, 406, 420, 4623 ,
534, 588, 591, 646
Classical, 215, 647
Late (LL), 199, 1994 , 215, 217, 218,
220, 221
Vulgar, 22913 , 238, 239
Lendu, 327, 335
Lingua Franca, 10, 85, 91, 104, 199, 1995 ,
214
Lotuho, 327, 328, 335, 336
Luganda, 327, 329, 335
Lugbara, 327, 335
Luo, 327
Mahriyōt, 354, 357, 359, 362, 365
686
Language index
Malayalam, 339
Malgwa, 192
Maltese, 3, 6, 7, 19, 193 , 1405 , 165, 167,
168, 170, 197, 198, 219, 221,
3083 , 388, 612, 616, 627, 633,
634, 637, 660, 661
Mamvu, 327
Manam, 645
Masa, 323
Ma’di, 327, 328, 332–336
Mbay, 323–325
Meroitic, 419
Moru, 327, 328, 335
Mundari, 327, 334
Mundu, 327, 328
Ngambay, 323–325
Niger-Congo, 323, 327, 333
Nilo-Saharan, 12, 188, 189, 199, 323,
327, 433, 628
Nubian, 12, 176, 323, 419, 434
Nuer, 327, 328
Nuristani, 441
Occitan, 291
Omotic, 434
Ossetic, 442, 446
Pamir, 442
Parthian, 443
Persian (Pers.), 6, 7, 9, 15, 22, 58, 59,
66, 68, 74, 7420 , 77, 78, 86,
89, 8911 , 91, 92, 95, 96, 101,
104, 107, 115, 116, 1162 , 117,
1174 , 118, 1187 , 119, 11910 , 120,
12012 , 121, 12113 , 122, 12214 ,
12216 , 123, 124, 12419 , 12420 ,
126–128, 12827 , 12828 , 12829 ,
129, 12930 , 12931 , 12933 , 130,
259, 314, 339, 342, 375, 379,
383, 394, 395, 460, 461, 463,
4635 , 464, 465, 469, 470, 472–
478, 482, 483, 491, 513, 543,
588, 591, 660
Bandarī, 442, 447
Baḫtiārī, 442, 450, 451
Classical, 444, 448, 451, 452, 4524
Darī, 442
Dizfūlī, 442
Fīnī, 442
Lurī, 442, 449–451
Lāristānī, 442
Modern Standard (MSP), 444, 4442 ,
446–451, 4524 , 453
New (NewP), 9, 441–443, 445–
449, 451–454
Phoenician, 50, 266
Pojulu, 327, 328, 334, 335
Portuguese, 10, 216, 23827 , 23932 , 306,
534, 535, 537, 5371 , 551, 5511 ,
589, 592
Provençal, 533, 534, 537, 538
Punic, 199, 215, 266
Punjabi, 339
Päri, 327, 328, 332, 333, 335, 336
Quechua, 586
Romance, 3, 6, 7, 10, 167, 199, 202, 214,
215, 218–220, 227, 229, 230,
23015 , 231–235, 23520 , 236,
23722 , 23823 , 23931 , 240, 265,
267, 269–276, 27610 , 277–279,
27912 , 281, 282, 284, 286, 287,
289, 28917 , 290–294, 311, 536,
537, 539–545, 585, 589, 593,
607, 647, 660, 661
Andalusi, 227, 229, 22914 , 230, 232–
235, 23823 , 240
687
Language index
Romani, 164, 413, 490
Romanian, 337, 338, 341
Russian, 75, 164
Safaitic, 38, 41, 42, 47–50, 7319
Saho, 424, 433
Sami, 162
Sango, 323, 325
Sar, 323, 325
Sara, 323, 325
Semitic, 8, 9, 37, 38, 42, 46, 47, 4713 ,
48, 49, 635 , 65, 7013 , 72, 73,
85, 90, 96, 101, 147, 163, 165,
189, 253, 265, 266, 269, 270,
277, 279, 283, 286, 288, 289,
294, 351, 352, 355, 356, 371,
376, 383, 384, 388, 419, 425,
427, 430, 433, 434, 462, 463,
465, 467, 470, 474, 55511 , 6492
Central, 46, 50, 12215
Ethiopian, 7, 58, 59, 64, 65, 77
Northwest, 85
Shilluk, 327, 328, 332, 333, 336
Sicilian, 7, 267–269, 272, 274, 275, 2758 ,
276–278, 280–286, 289–293,
311, 537, 661
Sinhalese, 337, 339, 341
Slovak, 293
Sogdian, 443
Somali, 424
Songhay, 199, 209, 247, 258, 261
Soninke, 247, 249, 258–260
South Arabian, 3, 9, 18, 38, 45, 51, 59,
65, 658 , 66, 72, 73, 425, 435
Modern (MSAL), 8, 12, 64, 65, 425,
650, 651, 653–655, 662
Baṭḥari, 8, 18, 353–360, 362, 364,
365
Ḥarsūsi, 8, 353, 365, 654
688
Hobyōt, 8, 353, 365
Mehri, 8, 65, 353, 354, 357–359,
361–365, 650, 654
Śḥerɛt, 8, 353, 354
Soqoṭri, 8, 353, 354, 356, 357,
365, 650
Old, 7, 51, 59, 64–66, 71–73, 352
Spanish, 10, 199, 206, 209, 214–221,
227, 232, 233, 23827 , 23932 ,
274, 283, 284, 286, 287, 291,
292, 306, 534, 535, 537, 5371 ,
538, 540, 551, 5511 , 586, 591,
592
Buenos Aires, 586
Hakitia, 216
Majorcan, 586
Old, 220
Sudanic, 12, 321, 323, 327, 333, 419
Ṣurayt, see Ṭuroyo
Swahili, 8, 327–332, 334–336
Tagalog, 339
Tagoi, 633
Tajik, 12, 314, 442, 445, 446, 635, 636
Tamil, 339
Thamudic, 46, 4612 , 47, 49
Tigre, 421, 434
Tigrinya, 434
Tupuri, 323
Turkic, 87, 3732 , 383, 396, 461, 467,
491, 513, 514
Turkish, 3, 6, 10, 66, 74, 78, 84, 87, 89,
91, 92, 96, 97, 101–105, 116,
119, 12931 , 130, 135–137, 1372 ,
138, 139, 141–144, 146–153, 15316 ,
154, 160, 163, 164, 199, 202,
206, 20615 , 207, 208, 214, 259,
287, 292, 3721 , 3732 , 374, 375,
379, 384, 386, 387, 389, 395,
Language index
396, 445, 465, 469, 470, 491–
493, 500, 501, 507, 513, 527,
534, 541, 542, 586
Ottoman, 7, 57, 59, 74, 77, 89, 92,
104, 107, 116, 142, 153, 206,
379, 4635 , 477, 478, 482, 483,
541, 543, 588
Turkmen, 12, 101, 107, 395, 396
Turku, 7, 322–325, 342, 343
Tālišī, 442
Tātī, 442, 449
Urdu, 339, 341, 342
Uzbek, 12
Venetian, 533, 5371 , 538, 540, 541
Wandala, 188, 189
Wolof, 247, 2492 , 258–260
Yaɣnōbi, 442
Yiddish, 75
Zande, 327, 328, 333–336
Zarma, 404
Zazaki, 137–139, 148–151, 442, 463, 470,
472, 474
689
Did you like this
book?
This book was brought to you for
free
Please help us in providing free access to
linguistic research worldwide. Visit
http://www.langsci-press.org/donate to
provide financial support or register as a
community proofreader or typesetter
at http://www.langsci-press.org/register.
language
science
press
Arabic and contact-induced change
This volume offers a synthesis of current expertise on contact-induced change in Arabic and its neighbours, with thirty chapters written by many of the leading experts on
this topic. Its purpose is to showcase the current state of knowledge regarding the diverse outcomes of contacts between Arabic and other languages, in a format that is both
accessible and useful to Arabists, historical linguists, and students of language contact.
ISBN 978-3-96110-251-8
9 783961 102518