8 THE SWAHILI LANGUAGE AND ITS EARLY HISTORY Martin Walsh From modest beginnings as the speech of a small group of mixed farmers, Swahili has become the lingua franca of millions of people in Eastern Africa and beyond (Lewis et al. 2015). How did this extraordinary transformation begin? This chapter outlines what is known (and not known) about the origins and initial development of the Swahili language and its dialects, and what this tells us in turn about the history of its speakers. It is based largely on research in historical and comparative linguistics undertaken since the 1970s and continuing through to the present. The potential contribution of research of this kind to understandings of the African past is well established (Nurse 1997; Blench 2006), and the Swahili-speaking world has already been the subject of important studies (including Nurse and Spear 1985; Nurse and Hinnebusch 1993). Needless to say, this work is neither exhaustive nor unproblematic. As we shall see, many gaps in our knowledge remain, while current research suggests that major revisions might be made to existing reconstructions of Swahili linguistic and cultural history. Taylor and Francis Not for distribution The origins of Swahili Swahili (Kiswahili) is a Bantu language, and so a member of the wider Niger-Congo family, one of the four main African language phyla (Nurse and Philippson 2003: 1–7). Its name is derived from Arabic sawāhil ‘coast’, hence Waswahili ‘people of the coast’ and Kiswahili ‘language (and culture) of the coast’. Swahili is the conventional English name, Kiswahili the autonym in Standard Swahili and the dialects on which it is based. Its use as an ethnic and linguistic label for the people and their language in the unrecorded past is a convenient and accepted anachronism. While linguists have long been aware of its Bantu affiliation, the presence of a large number of Arabic loanwords in Swahili has led many commentators over the years to claim that it is (or was) a pidgin, creole, or some other kind of mixed language. Like crude neo-racist theories about the Asian and Middle Eastern origins of Swahili ‘civilisation’, this is a view that has now been thoroughly repudiated (Nurse and Hinnebusch 1993: 36–7; Hinnebusch 1996). Swahili has the typological characteristics of a Bantu language (Nurse and Philippson 2003: 7–10), the most obvious perhaps being its system of noun classes (also found in some Niger-Congo languages outside of Bantu). Where it lacks such features – for example, tone – the processes that led to their loss can be readily identified. 121 Martin Walsh The genetic classification of Swahili within Bantu has been progressively refined by linguists as more data have become available on the language and its closest relatives. Research undertaken from the 1970s onwards culminated in the publication of Derek Nurse and Tom Hinnebusch’s Swahili and Sabaki: A Linguistic History (1993), one of the most extensive studies of any Bantu subgroup. Nurse and Hinnebusch trace the origins of Swahili back to one particular group of East African Bantu languages, Northeast Coast (NEC) Bantu. In their classification NEC has four members; one of these, Sabaki, being the subgroup to which Swahili belongs. The full list of subgroups and their principal constituent languages is given by them as follows (1993: 4–19): Sabaki: Swahili, Mwani, Elwana (Ilwana), Pokomo, Mijikenda, Comorian Seuta: Bondei, Shambala (Shambaa, Sambaa), Ngulu (Nguu), Zigula (Zigua) Ruvu: Gogo, Kaguru, Kami, Kutu, Doe, Nhwele, Luguru, Sagara, Vidunda, Zaramo Pare: Pare (Asu), Tuȕeta (Taveta) The non-Sabaki languages listed above are spoken primarily in the east of Tanzania, and it is generally presumed that Proto-Northeast Coast (PNEC) and its early dialects arose somewhere in this area, north of the Rufiji Basin. Although the precise origins of PNEC remain uncertain, the available evidence suggests a close affiliation with the Southern Highlands group of Bantu languages in southwest Tanzania. The implication of this is that PNEC and its early dialects arose at the time Bantu languages spread northwards and eastwards from the Nyasa-Tanganyika corridor towards the East African coast. A comparison of linguistic geography and ecology suggests a historical movement along the line of the Eastern Arc Mountains and onto the coastal plain to the east, a trajectory that demanded gradual adaptation to increasingly drier environments, especially when the ancestors of the Sabaki languages pushed northwards (Walsh 2003: 6). Nurse and Hinnebusch did not make anything of these and other ecological connections. They did, however, speculate on the correlations between their linguistic findings and the known archaeological record, assuming that the incoming Bantu speakers were responsible for the first Early Iron Age (EIA) sites on the coast and in its hinterland. As a result, they judged that ‘an approximate date around or slightly later than 1 ad would seem reasonable for PNEC’, with Proto-Sabaki (PSA) emerging ‘perhaps five hundred years later’ and Proto-Swahili ‘shortly after that’ (1993: 23; also 493). It should be emphasised that these were broad estimates based on the archaeological dates that were then available. They were not derived independently, but were considered to be plausible in terms of Nurse and Hinnebusch’s linguistic reconstruction. In his ambitious synthesis of linguistic and archaeological evidence, An African Classical Age, Ehret (1998: 192) similarly surmised that speakers of NEC Bantu must have settled on the central Tanzanian coast in the early centuries ad. Unlike Nurse and Hinnebusch, however, Ehret argues that they were preceded by ‘Upland Bantu’, who moved to parts of (what is now) northeastern Tanzania and eastern Kenya from eastern Nyanza, to the east of Lake Victoria. These people, ancestors of the Chaga-Dabida and the Thagicu (Central Kenya Bantu, CKB), are supposed by Ehret to have been the producers of Kwale Ware pottery, while the NEC Bantu speakers who replaced them from the south were responsible for Tana Tradition ceramics (1998: 184–9; 2002: 175–6). However, the evidence for Ehret’s scenario is very weak: the distribution of Kwale and EIA/EIW (Early Iron Working) pottery is a much better fit with the NEC languages, and there is a much more parsimonious explanation for the presence of CKB loanwords on the coast (Helm 2000: 281–2; Spear 2000: 274; Walsh 2003: 6; 2013: 25–30). Taylor and Francis Not for distribution 122 he Swahili language and its early history More recently, one of Ehret’s students (Gonzales 2009) has reprised his thesis, arguing that Upland Bantu (she calls them ‘Azania Bantu’) had spread down the Tanzanian coast around 200 bc, while PNEC speakers arrived in northeastern Tanzania much later, around ad 300. On Gonzales’ account the PNEC period was short-lived, and by about ad 500–600 ProtoSabaki and other NEC subgroups had begun to diverge. She identifies three initial subgroups, Proto-Sabaki, Proto-Asu (Nurse and Hinnebusch’s Proto-Pare), and Proto-Wami, the latter subsequently splitting into Proto-Seuta and Proto-Ruvu (compare with Figure 1 below). Gonzales tries to maintain the correlation with pottery types proposed by Ehret, with NEC Bantu speakers as the makers of Tana Tradition ceramics. But, as we now know, the dates do not fit: early Tana Tradition ceramics are dated c. 600–900 ce, and are most closely associated with the coastal sites that became Swahili stonetowns, as well as some of their inland neighbours (Fleisher and Wynne-Jones 2011: 246–8). This is clearly too late for PSA, and a much earlier date for its divergence from PNEC is suggested by Kwale/EIW dates from the third century ce, in the hinterland of the southern and central parts of the Kenya coast (Helm 2000: 69–71). Let us return, though, to Nurse and Hinnebusch’s findings on the genetic classification of PNEC and especially PSA. Perhaps surprisingly, they did not summarise them in a tree diagram. Their Swahili and Sabaki (1993) is a highly technical and densely argued work of historical linguistic reconstruction that gives careful consideration to different possible interpretations of the data and is correspondingly cautious in its conclusions. Because of this, relatively few nonspecialists are conversant with the detail of their arguments or their principal conjectures, let alone in a position to assess them critically. Archaeologists and historians working on the East African coast are generally much more familiar with Nurse’s earlier and more popular collaboration, which includes a much simpler tree diagram of the Sabaki languages (Nurse and Spear 1985: 54, Figure 7) than the ones that can be drawn from his joint work with Hinnebusch. My own version of the latter, based on a close reading of the 1993 text, is shown in Figure 8.1 (for an alternative interpretation of Nurse and Hinnebusch’s reconstruction of PSA, see Helm 2000: 45, Figure 2.6: this omits Upper Pokomo but retains Mwani as a primary branch of Sabaki). Nurse and Hinnebusch’s reconstruction of Sabaki history encompassed a number of significant findings. As can be seen from Figure 8.1, the genetic classification of the Sabaki languages differs from their typological description. Viewed in terms of their historical development and Taylor and Francis Not for distribution Swahili Elwana ProtoSabaki Upper Pokomo ProtoNortheast Coast Bantu ProtoSeuta Lower Pokomo ProtoPare Mijikenda ProtoRuvu Comorian Figure 8.1 Genetic classification of the Sabaki languages. (Based on Nurse and Hinnebusch 1993.) 123 Martin Walsh relationships, Upper and Lower Pokomo are quite different languages, not the dialects of a single language that their names suggest. Nurse and Hinnebusch had some difficulty in classifying Upper Pokomo, which is spoken along the Tana River between Elwana and Lower Pokomo, and shares features of both. Lower Pokomo, though, has evidently shared much of its history with Mijikenda, and they both form a readily identified group together with Comorian. Moreover, whereas Comorian had traditionally been treated as a dialect or very close relative of Swahili, it was now clearly identified as a separate language with much more in common with Lower Pokomo and Mijikenda, though there was also evidence to suggest that the early Comorians had interacted with different Swahili-speaking communities (1993: 494–6). Although this might seem to bring us closer to solving the problem of Comorian origins, it still leaves a lot of questions unanswered, including the precise location of their last homeland and point of departure on the East African coast. Based on a pattern of shared phonological innovations, one linguist has suggested recently that a South Bantu substratum can be identified in Comorian, implying that it originated on the Mozambican coast as a language related to the group that now includes Makhuwa, Chuwabo, Sena, Ndau, Ronga, Gitonga and others, before shifting to become a Sabaki language (Wills 2013, 2015). This is an intriguing proposal whose details remains to be worked out and subjected to critical scrutiny. However, a similar language shift almost certainly gave rise to Mwani. This was demoted in Nurse and Hinnebusch’s (1993: 527) classification from its position as the southernmost branch of Sabaki and reanalysed as the result of historical interaction between a variety of Southern Swahili and the Maviha dialect of the non-Sabaki language Makonde in northern Mozambique. Swahili and Elwana remain as primary branches in Nurse and Hinnebusch’s Sabaki tree, together with the Pokomo/Mijikenda/Comorian group. In their synthesis they argued that following the breakup of the PNEC community, ‘the PSA moved north to a homeland somewhere in the area north of the Tana River, and probably into southern Somalia’ (1993: 493). Nurse and Hinnebusch came to this conclusion after considering both linguistic geography (the principle of ‘least moves’) and the historical traditions of the Sabaki groups that claim a northern origin in a place called Shungwaya, usually located on the southern Somali coast. Traditions like this are notoriously unreliable as guides to the distant past, and the PSA community was more likely located somewhere to the south of where Nurse and Hinnebusch placed it (as argued in Walsh 1992, and Helm 2000). It is easier to accept their conclusion that the PSA period must have been relatively short, lasting less than three centuries (1993: 493) – though the date range that they proposed (500 ce or earlier to sometime before 800 ce) might be adjusted to line up with current archaeological dates, which suggest that a distinctive Swahili community had already emerged by the seventh century ce. Taylor and Francis Not for distribution The development of dialects Nurse and Hinnebusch assumed that the Proto-Swahili community, located in the same general area as their PSA predecessors, must also have been short-lived (1993: 297). The early dispersal of Swahili speakers from their northern homeland and establishment of settlements along more than a thousand miles of the eastern African coast and islands represents an extraordinary social and economic transformation that we are still only just beginning to understand (Ray, this volume). It was matched by the development of many diferent dialects and local varieties of Swahili. The main dialects as they survive and are known today are listed in Table 8.1, together with an indication of the principal sources of lexical information on each of them. I have included the Comorian dialects, as well as the ‘mixed languages’ in Mozambique whose classiicatory status is equivocal (I will say more about this later). 124 he Swahili language and its early history Table 8.1 Main Swahili dialects as they are known today. Dialect/dialect cluster SWAHILI Mwiini (Mbalazi) Bajuni (Tikuu, Gunya) Siu, Pate, Amu Jomvu, Ngare Mvita Chifundi Vumba Mrima Pemba Tumbatu Kae (Makunduchi, ‘Hadimu’) Unguja Mafia Mgao Makwe (Maraba) Mwani Koti COMORIAN Ngazija Mwali Nzwani Maore Principal lexical sources Kisseberth and Abasheikh 2004; Kisseberth 2016 Sacleux 1939; Nurse 2010 Sacleux 1939 Lambert 1958a Krapf 1882; Binns 1925 Lambert 1958b Lambert 1957 Sacleux 1939; Whiteley 1956 Whiteley 1958; BAKIZA 2012a BAKIZA 2012b Chum 1962, 1994; BAKIZA 2012c Sacleux 1939; Johnson 1939 Kipacha 2004 Kipacha 2010 Devos 2008 Rzewuski 1979; Floor 1998; Abudo et al. 2009 Schadeberg and Mucanheia 2000 Fischer 1949; Ahmed Chamanga and Gueunier 1979; Lafon 1991 Ahmed-Chamanga 1992, 1997 Blanchy 1996 Taylor and Francis Figure 8.2 summarizes Nurse and Hinnebusch’s (1993) genetic classification of Swahili dialects, together with some of the additional detail provided by Nurse’s earlier work (1982, 1984/85), including his collaboration with Spear (Nurse and Spear 1985). This work demonstrated a clear historical split between Northern and Southern Swahili dialects. The Northern dialects themselves split into two main subgroups, with Mwiini on the Somali coast being the most divergent member of the first of these, and Chifundi on the south Kenya coast of the second. The dialects of the Lamu archipelago are close enough to one another to be described as local varieties, likewise the Mombasa dialects. The Southern dialects proved much harder to classify, and the resulting arrangement is rather more uncertain. In all of these cases linguistic analysis provides no more than relative timings, and for dates we have to rely on assumptions about the relation between known settlement histories and dialect development. Since the publication of Nurse and Hinnebusch’s work a lot more information has become available on some of the Swahili dialects as well as other Sabaki languages. Understandings of language and dialect change have also evolved in tandem with developments in historical linguistics. This has not yet resulted in anything approaching a comprehensive revision of Nurse and Hinnebusch’s Swahili and Sabaki, which remains the standard text on the emergence and development of the Swahili language. It is possible, however, to outline some of the more recent observations and findings that might form part of such a revision. Some of these have been made by Nurse and Hinnebusch themselves, writing in the wake of their 1993 study. Hinnebusch (1996: 76), for example, in a subsequent summary of their work, treated Mwiini as a distinct language rather than as a dialect of Swahili, implying that they might be considered to be coordinate members of the same branch of PSA (but see Nurse 1991; 2000: 234). Researchers working specifically on Mwiini have also raised doubts about its classification Not for distribution 125 Martin Walsh Mwiini Bajuni Siu Lamu archipelago Northern dialects Pate Amu Jomvu Mombasa dialects Ngare Mvita Chifundi ProtoSwahili Vumba Pemba Tumbatu Mtang’ata Unguja Kae Southern dialects Mafia Mgao Mwani Taylor and Francis Figure 8.2 Genetic classification of Swahili dialects. (Based on Nurse 1982; 1984/85; Nurse and Spear 1985; Nurse and Hinnebusch 1993.) Not for distribution (for example, Kisseberth and Abasheikh 1977: 179; 2011: 1988), while sometimes continuing to describe it as a Swahili dialect (Henderson 2010: 75–6). Nurse’s subsequent research focused on the role of contact phenomena and their consequences for understanding language change, work that was inspired by methodological developments in historical linguistics (Thomason and Kaufman 1988). The potential significance of language shift in the creation of Swahili dialects was clearly demonstrated in studies of Chifundi and Vumba (Möhlig 1992; Nurse and Walsh 1992). Although these two dialects sit on either side of the Northern/Southern dialect divide, they share a number of features in common that cannot be ascribed simply to borrowing between them. It appears instead that they originate in Digo or an earlier form of Mijikenda, and that these distinctive Swahili dialects were formed when Mijikenda speakers shifted to speaking the language of their Swahili neighbours. Closer inspection indicates that there is evidence for contact phenomena and similar processes taking place throughout Sabaki and sometimes involving contact with non-Sabaki languages. Nurse (2000) has reanalysed Elwana (Ilwana) from the perspective of its interactions with the Cushitic language Orma, and Wills’ (2013; 2015) suggestion that there might be a South Bantu substratum in Comorian would also fall under this heading. Further evidence for the role of contact in the development of Swahili dialects comes from the southern end of the Swahili spectrum. Since the publication of Nurse and Hinnebusch’s study in 1993, much more research has been undertaken and data published on Mwani and other ‘mixed languages’ on the coast and islands of northern Mozambique. The best known of these studies is Schadeberg and Mucanheia’s monograph on Ikoti (2000), which is spoken 126 he Swahili language and its early history on the island of the same name and some villages on the adjacent mainland (Schadeberg and Mucanheia 2000). More often than not, Ikoti has been described as a dialect of Swahili, and a lot of its vocabulary suggests as much. Structurally, however, it is clearly derived from Makhuwa, and can be analysed as being the product of a partial language shift from this South Bantu language to Swahili. Schadeberg and Mucanheia hypothesise that this process began relatively recently, in the fifteenth century, when speakers of Swahili – possibly the Mgao dialect – are said to have established the town of Angoche (2000: 7–8). As noted earlier in this chapter, Mwani can be analysed similarly (Petzell 2002: 88–9), as can other coastal idioms like Makwe (Maraba) that appear to have Makonde origins. It is quite likely that more examples of such admixture are waiting to be identified and described among the languages and dialects of the East African coast and islands, adding further complexity to the language trees that have been presented here. Such cases of past interaction are hardest to analyse when they involve closely related dialects of the same language: this applies in particular to Swahili, given what we already know and traditions tell us about the movements of groups of speakers from one community to another. This is made more difficult by the lack of reliable linguistic data on many dialects. While new information continues to emerge, the possibility of collecting more is diminished by the accelerating decline of local varieties in response to the deleterious impacts of globalisation, including the spread of Standard Swahili and other dominant dialects. The focus of much linguistic research on Swahili literature and urban speech habits does little to help in this context either. Towards a new synthesis? There remains a lot about the history of Swahili before the modern era that we do not know. In addition to pursuing the revision of Nurse and Hinnebusch’s model along the lines that have been sketched above, there are a number of other strategies of historical linguistic research that have barely been followed to date. The most obvious perhaps of these is the in-depth study and analysis of particular lexical and cultural domains (cf. Walsh 2003: 63–5; 2007: Appendix), particularly those relevant to understanding the history of distinctive social and economic practices among the speakers of Swahili and neighbouring languages. This relies in turn on the availability of good lexical data. Although a great deal can be done with existing dictionaries and wordlists, there are many gaps in our lexical knowledge of Sabaki and Swahili that remain to be illed and act as a constraint on this kind of research. Even when we possess good vocabularies, like the published lists of Swahili nautical terminology (Prins 1970; Miehe and Schadeberg 1979: 80–93), we often lack comparative material from elsewhere (in this case from the Comoros and other maritime communities in the region). In the case of this particular lexical ield, it would also be good to know more about the Arabic and other sources of much of the vocabulary. While a fair amount is known already about the influence of Arabic on Swahili in recent centuries, relatively little is known about earlier patterns of word borrowing from the East African coast and its hinterland as well as from different places across the Indian Ocean. Nurse and Hinnebusch have outlined the main sources of lexical innovation in Swahili and its dialects, but some sets of loanwords deserve much more thorough investigation. These include words of Cushitic origin that appear to derive from contact between the early Swahili and their neighbours on the coast, among them hunting, foraging and fishing communities (Nurse 1988). Further research is also required to understand the history of interactions between Swahili and other languages spoken in and around the Indian Ocean. The nature and extent of early contacts with the speakers of Austronesian languages, including the Proto-Malagasy and their descendants, is only one of a number of unsolved problems in this context for both Swahili and Comorian history (Adelaar 2009). Taylor and Francis Not for distribution 127 Martin Walsh We are some way off from fully understanding how and why Swahili society and culture emerged and developed from the relatively simple beginnings that they had. How did this group of mixed farmers adapt to life on the coast, and the new opportunities for subsistence and exchange that it must have offered? What prompted the migration of both the Swahili and the Comorians from their homelands, and what was the connection between these movements and other developments in the western Indian Ocean, including the trade in slaves and other African goods? What role, if any, did the settlers of Madagascar play in this history? These and other unanswered questions invite a multidisciplinary approach that includes historical and comparative linguistics. Research along the lines suggested in this chapter has the potential to deliver not only a better understanding of the history of Swahili, but also to make an important contribution to the synthesis that will surely emerge. References Abudo, F., Buana, C., Bacar, S., Aquimo, M. and de Sousa, A. F. 2009. Vocabulário de Kimwani. Nampula: SIL Mozambique Adelaar, A. 2009. ‘Towards an integrated theory about the Indonesian migrations to Madagascar’. In Ancient Human Migrations: A Multidisciplinary Approach, edited by P. N. Peregrine, I. Peiros and M. Feldman, pp. 149–172. Salt Lake City: The University of Utah Press Ahmed-Chamanga, M. 1992. Lexique comorien-(shindzuwani)-français. Paris: Editions l’Harmattan Ahmed-Chamanga, M. 1997. Dictionnaire français-comorien (dialecte shindzuwani). Paris: CEROIINALCO/L’Harmattan Ahmed Chamanga, M. and Gueunier, N.-J. 1979. Le dictionnaire comorien-français et français-comorien du R. P. Sacleux (2 Vols.). Paris: SELAF Baraza la Kiswahili la Zanzibar (BAKIZA) 2012a. Kamusi la Lahaja ya Kipemba. Nairobi: Oxford University Press Baraza la Kiswahili la Zanzibar (BAKIZA) 2012b. Kamusi la Lahaja ya Kitumbatu. Zanzibar: Jamana Printers Ltd Baraza la Kiswahili la Zanzibar (BAKIZA) 2012c. Kamusi la Lahaja ya Kimakunduchi. Zanzibar: Jamana Printers Ltd Binns, H. K. 1925. Swahili-English Dictionary: Being Dr. Krapf’s Original Swahili-English Dictionary Revised and Re-arranged. London: Society for Promoting Christian Knowledge Blanchy, S. 1996. Dictionnaire mahorais-français, français-mahorais. Paris: L’Harmattan Blench, R. M. 2006. Archaeology, Language, and the African Past. Lanham: Altamira Press Chum, H. 1962. ‘A vocabulary of the Kikae (Kimakunduchi, Kihadimu) dialect: with examples illustrating the morphology’. Swahili: Journal of the East African Swahili Committee, 33 (1): pp. 51–68 Chum, H. 1994. Msamiati wa Pekee wa Kikae: Kae Speciic Vocabulary. Uppsala: Nordic Association of African Studies Devos, M. 2008. A Grammar of Makwe (Palma, Mozambique). München: LINCOM Europa Ehret, C. 1998. An African Classical Age: Eastern and Southern Africa in World History, 1000 BC to AD 400. Charlottesville: University Press of Virginia Ehret, C. 2002. The Civilizations of Africa: A History to 1800. Oxford: James Currey Fischer, F. 1949. Grammaire-dictionnaire comorien. Strasbourg: Société d’Edition de la Basse-Alsace Fleisher, J. B. and Wynne-Jones, S. 2011. ‘Ceramics and the Early Swahili: deconstructing the Early Tana Tradition’. African Archaeological Review, 28(4): pp. 245–78 Floor, S. 1998. Vocabulário mwani-português. Pemba: SIL Mozambique Gonzales, R. M. 2009. Societies, Religion, and History: Central-East Tanzanians and the World They Created, c. 200 BCE to 1800 CE. New York: Columbia University Press. Available online at: http://www. gutenberg-e.org/gonzales/ Henderson, B. 2010. ‘Chimwiini: endangered status and syntactic distinctiveness’. Journal of West African Languages, 37 (1): pp. 75–91 Helm, R. 2000. ‘Conlicting histories: the archaeology of the iron-working, farming communities in the central and southern coast region of Kenya’. PhD diss., University of Bristol Hinnebusch, T. J. 1996. ‘What kind of language is Swahili?’ Afrikanistiche Arbeitspapiere, 47: pp. 73–95 Taylor and Francis Not for distribution 128 he Swahili language and its early history Johnson, F. (Ed.) 1939. A Standard Swahili-English Dictionary. London: Oxford University Press Kipacha, A. 2004. ‘Kingome-English lexicon’. Swahili Forum, 11: pp. 179–209 Kipacha, A. 2010. ‘The quest for Kimgao Coastal Swahili: latest indings’. In Bantu Languages: Analyses, Description and Theory, edited by K. Legère and C. Thornell, pp. 85–98. Köln: Rüdiger Köppe Kisseberth, C. W. 2016. ‘Chimiini-English dictionary and chrestomathy’. http://swahiliendangered.cdh. ucla.edu/wp-content/uploads/2016/05/Chimiini-Lexicon-and-Chrestomathy-February-2016.pdf Kisseberth, C. W. and Abasheikh, M. I. 1977. ‘The object relationship in Chi-Mwi:ni, a Bantu language’. In Syntax and Semantics, Volume 8: Grammatical Relations, edited by P. Cole and J.N. Sadock, pp. 179–218. New York: Academic Press Kisseberth, C. W. and Abasheikh, M. I. 2004. The Chimwiini Lexicon Exempliied. Tokyo: Research Institute for Languages and Cultures of Asia and Africa (ILCAA), Tokyo University of Foreign Studies Kisseberth, C. W. and Abasheikh, M. I. 2011. ‘Chimwiini phonological phrasing revisited’. Lingua, 121: pp. 1987–2013 Krapf, L. 1882. A Dictionary of the Suahili Language. London: Trubner & Co. Lafon, M. 1991. Lexique français-comorien (shingazidja). Paris: Editions L’Harmattan Lambert, H. E. 1957. Ki-Vumba: A Dialect of the Southern Kenya Coast. Kampala: East African Swahili Committee Lambert, H. E. 1958a. Chi-Jomvu and Ki-Ngare: Subdialects of the Mombasa Area. Kampala: East African Swahili Committee Lambert, H. E. 1958b. Chi-Chifundi: A Dialect of the Southern Kenya Coast. Kampala: East African Swahili Committee Lewis, M. P., Simons, G. F. and Fennig, C. D. (Eds) 2015. Ethnologue: Languages of the World (18th Edition). Dallas, Texas: SIL International. Available online at: http://www.ethnologue.com Miehe, G. and Schadeberg, T. C. (Eds) 1979. Sambo ya Kiwandeo: The Ship of Lamu-Island by Ahmed Sheikh Nabhany. Leiden: Afrika-Studiecentrum Möhlig, W. J. G. 1992. ‘Language death and the origin of strata: two case studies of Swahili dialects’. In Language Death: Factual and Theoretical Explanations with Special Reference to East Africa, edited by M. Brenzinger, pp. 157–79. Berlin and New York: Mouton de Gruyter Nurse, D. 1982. ‘A tentative classiication of the primary dialects of Swahili’. Sprache und Geschichte in Afrika, 4: pp. 165–205 Nurse, D. 1984/85. ‘A historical view of the southern dialects of Swahili’. Sprache und Geschichte in Afrika, 6: pp. 225–50 Nurse, D. 1988. Alternative historical sources for Swahili vocabulary’. In Languages and Cultures: Studies in Honor of Edgar C. Polomé, edited by M. A. Jazayery and W. Winter, pp. 475–87. Berlin: Mouton de Gruyter Nurse, D. 1991. ‘Language contact, creolization, and genetic linguistics: the case of Mwiini’. In Proceedings of the Seventeenth Annual Meeting of the Berkeley Linguistics Society: Special Session on African Language Structures, pp. 177–87. Berkeley: University of California Nurse, D. 1997. ‘The contribution of linguistics to the study of history in Africa’. Journal of African History, 38 (3): pp. 359–91 Nurse, D. 2000. Inheritance, Contact, and Change in Two East African Languages. Köln: Rüdiger Köppe Verlag Nurse, D. 2010. ‘Lexicon for Ki-Bajuni = Ki-Gunya = Ki-Thikhuu’. Available online at: http://www.ucs. mun.ca/~dnurse/bajuni_database/wordlist.pdf Nurse, D. and Hinnebusch, T. J. 1993. Swahili and Sabaki: A Linguistic History. Berkeley: University of California Press Nurse, D. and Philippson, G. 2003. ‘Introduction’. In The Bantu Languages, edited by D. Nurse and G. Philippson, pp. 1–12. London and New York: Routledge Nurse, D. and Spear, T. 1985. The Swahili: Reconstructing the History and Language of an African Society, 800–1500. Philadelphia: University of Pennsylvania Press Nurse, D. and Walsh, M. T. 1992. ‘Chifundi and Vumba: partial shift, no death’. In Language Death: Factual and Theoretical Explanations with Special Reference to East Africa, edited by M. Brenzinger, pp. 181–121. Berlin and New York: Mouton de Gruyter Petzell, M. 2002. ‘A sketch of Kimwani (a minority language of Mozambique)’. Africa & Asia: Göteborg Working Papers on Asian and African Languages and Literatures, 2: pp. 88–110.3 Prins, A. H. J. 1970. A Swahili Nautical Dictionary. Dar es Salaam: Chuo cha Uchunguzi wa Lugha ya Kiswahili Taylor and Francis Not for distribution 129 Martin Walsh Rzewuski, E. 1979. Vocabulário da lingua mwani (quimuane). Maputo: Eduardo Mondlane University (Departamento de Letras Modernas) Sacleux, C. 1939. Dictionnaire Swahili-Français. Paris: Institut d’Ethnologie Schadeberg, T. C. and Mucanheia, F. U. 2000. Ekoti: The Maka or Swahili Language of Angoche. Köln: Rüdiger Köppe Spear, T. 2000. ‘Early Swahili history reconsidered’. The International Journal of African Historical Studies, 33 (2): pp. 257–90 Thomason, S. G. and Kaufman, T. 1988. Language Contact, Creolization, and Genetic Linguistics. Berkeley and Los Angeles: University of California Press Walsh, M. T. 1992. ‘Mijikenda origins: a review of the evidence’. TransAfrican Journal of History, 21: pp. 1–18 Walsh, M. T. 2003. ‘Languages, cultures and environments: historical linguistics between the African Great Lakes and the western Indian Ocean’. In Second Platina Workshop 17–19 October 2002, Usa River, Arusha, Tanzania, edited by A. Dahlberg, H. Öberg, S. Trygger, K. Holmgren and P. Lane, pp. 53–74. Stockholm: Environment and Development Studies Unit, Stockholm University Walsh, M. T. 2007. ‘Island subsistence: hunting, trapping and the translocation of wildlife in the western Indian Ocean’. Azania, 42: pp. 83–113 Walsh, M. T. 2013. ‘The Segeju Complex? Linguistic evidence for the precolonial making of the Mijikenda’. In Contesting Identities: The Mijikenda and Their Neighbors in Kenyan Coastal Society, edited by R. Gearhart and L. Giles, pp. 25–51. Trenton, New Jersey: Africa World Press Whiteley, W. H. 1956. Ki-Mtang’ata: A Dialect of the Mrima Coast. Kampala: East African Swahili Committee Whiteley, W. H. 1958. The Dialects and Verse of Pemba: An Introduction. Kampala: East African Swahili Committee Wills, J. 2013. ‘“Strong Y-” as an Isogloss for Southern Bantu’. Paper presented to the 5th International Conference on Bantu Languages (Bantu 5), Paris, 12–15 June Wills, J. 2015. ‘The South Bantu Stage of Comorian’. Unpublished ms. Taylor and Francis Not for distribution 130