JOURNAL OF SEMITIC STUDIES
VOLUME
17
NUMBER I
SPRING
I972
ON T H E USE OF
AKKADIAN INFINITIVE AFTER
OR CONSTRUCT STATE
I. N O M I N A L I Z A T I O N O F COMMAND,
POTENTIALITY
WISH,
The relative clause Ja Qzls'zl in an Akkadian sentence like
bitzlm ;a pzls'zl i m p t "the house which he built fell down"
can be analysed as representing a finite sentence
bitam pas' "he built a house ".
This process is called nominalization, because a sentence is
thereby correlated to a noun phrase or its equivalent: in the
example given, bitam pzlf is transformed into the relative clause
s'a pzlfzl which has the same syntactical role as an attribute or
other similar noun phrase. The notion, if not the term, of
nominalization' is well known to traditional grammar, for
instance in the case of the subjective and objective genitive. A
noun phrase of the type
eris'ti Sarrim "the king's request"
is accordingly explained as the equivalent of some such sentence
as
farrzlm irris' "the king asks for ",
i.e. the genitive Jarrim is explained by resolving the noun phrase
into a finite sentence where farrm appears as the subject (hence
the term "subjective genitive"). Conversely, a noun phrase of
the type
eris'ti kzspim "a request of silver "
is explained as the equivalent of
,&@am irris' "he asks for silver ",
I On the notion of nominalization as applied to Akkadian, and in general
on the method followed in this article, see my forthcoming book A StrwturaI
Grammar of Babylonian.
THE AKKADIAN INFINITIVE AFTER c ' S ~ ' J
where h q a m occurs as the object (hence "objective genitive").'
The nominalization of a sentence with the verb in the indicative occurs regularly by introducing a relative pronoun and
inflecting the verb in the subjunctive - as in the example quoted
at the beginning. The same procedure, however, cannot be
applied if the predicate of the underlying sentence is in the
precative: a sentence like
bitam'i.@Z
"let him build the house"
cannot be nominalized by means of a relative clause because
bitam 5-z *@aSa "the house which he ought to build "
is impossible in Akkadian, as there is no subjunctive of the
precative. The question then is: can a sentence with the precativez be nominalized, and if so, how? I wish to suggest here
that the answer is in the affirmative, and that the device used in
Akkadian consists in inflecting the infinitive in the genitive after
a noun in the construct state, or after the determinative pronoun
.fa. Continuing with the paradigmatic example introduced above,
bftam Zpz/s'
"let him build the house"
is regularly norninalized in Akkadian as
bitum fa epdSim "the house which he ought to build".
As is well known, the precative exhibits various shades of
meaning, from command ("let him. . ."), to wish ("may he. . .")
and to potentiality especially in interrogatives("he could. . . .' 7 . 3
1 A similar procedure is also found in J. Aro, "Die akkadischen Infinitivkonstruktionen", Sttldia Orientaha XXVI (1961)~$ 2 . 3 5, p. 41 (abbreviated
henceforth as Aro), for a construction which will be studied below in this
article, namely er~et13 tdri "the land of no returning" : this he equates with
iSin er~etid itnrr# "one will not return from the land".
I refer here only to the precative for brevity's sake, but in effect the
formula proposed applies to all moods of command, both positive and
negative, i.e. in addition to the precative, also the cohortative, imperative,
vetitive and prohibitive. Note that even though the prohibitive uses the
indicative (type ld teppef "do not build!"), nominalization by means of a
relative clause would be ambiguous (at least in Babylonian) since l d f subjunctive is used for a statement. Hence: b i t m fa ld teppeSir "a house you will
not build", but b&m !a hi ep&&ka"a house you ought not to build". For the
combination of all moods of command into one paradigm see E. Reiner,
A Lingtlistic AnaIysis of Akkadian (The Hague, 1966)~5.4.2.4.
3 Cf. G A G $ I 5 3 g; AHw 5 5 9 A I d. The potential value of the precative
is found normally in interrogative, but occasionally also in affirmative,
sentences; cf. AHw 5 59 A I f, and the formula h i r lit?&"it could be more,
it could be less",passim in OB, C A D A12 488. Also potential in nature may
be considered the concessive usage of the precative, cf. G A G $ I 5 8 c; AHw
5 59 A I e, e.g. pzi&tim Ziddinzinikkim d damiq "they could give you a substi-
THE AKKADIAN INFINITIVE AFTER "$A"
Accordingly, the meaning of the correlative noun phrase will
also vary within a considerable range : in adltion to a translation
of the type "which he ought to. . ." (corresponding to the
expression of command), there will be translations such as
"which he should. . " (wish) and "which he might. . ."
(potential). In general, to arrive at a choice among the various
possibilities one can try to understand the nominal constructions
in terms of one of three underlying sentences. In the first case,
there is an implicit command on the part of the speaker to the
effect that the subject has to perform a certain action, e.g.
fa epds'im "which he ought to build"
I@zis' "let him build "
-- epds'am aqabbis'zimI "I order him to build".
For expression of wish also there is implicit the point of view of
the speaker, who in this case formulates precisely a wish, rather
than a command, that an action be performed:
s'a epds'im "which he should build"
I@& " may he build"
(epds'am errisszi)z "I wish that he builds", "I want him to
build ".
In the third alternative, the potential, there is no implicit participation of the speaker, but simply a statement that the subject
will - if he can, if he wants, or generally if some condition is
met - perform a certain action:
s'a epds'im "which he could build"
-- IipzIs' "he could build "
-- s'zimma. . .ippes' "he will build if. . ."
An alternative finite form to express the potential in Akkadian
besides the precative is with the present and the enclitic -man
(-min in Old Assyrian) appended to any part of the sentence:
ippesinan3 "he could build".
.
N
-
N
tute, but to no avail" ABPh 40: 14-16 (OB). Cf. also the rare combination
of precative and enclitic -man to express potential, G A G § 170 h.
For various examples of constructions of this type see Aro, 3.10,3.64, etc.
2 Normally with a verb of wish the subject of the main verb is the same
as the subject of the subordinate clause, e.g. Sham EaSeh "he wishes to buy"
L E A iii 24 = B iii 8 (OB). For a rare example in which the subjects are different see at8 turra Ja ?&tii irriSRk " why do they ask that you return a
favour ?" A B L 29 I Rev. 3-4 = Aro, 3.102 (NB). The paradigmatic sentence
in the text is built on this example; the parenthesis in the text is meant to
indicate that the type is rare (even when the subjects are the same) and that
the synthetic form (the precative) is in fact the regular Akkadian construction
instead of the periphrastic form.
3 For constructions of this type cf. G A G § 152 d; AHw 601, 654 f.
T H E A K K A D I A N I N F I N I T I V E A F T E R "SA"
(If the implicit condition is conceived as unrealizable at the
same time it is posited, we have the so-called "irrealis"
which is formally differentiated in Akka&an by means of
the enclitic -man, or -min in Old Assyrian, and the verb in the
preterite :
$tdwan*"he could have built (if he wanted, but he didn't)".
If the condition is unrealizable, and the main action is desired
by the speaker, Akkadian uses diman with the present or 1.with
the preterite :
1iman ippes'2 "oh, could he build!"
1
. @& "he should have built ".
Note also that the potential is different from a positive statement
concerning the actual ability of the subject to perform a certain
action, without reference to limiting conditions of sort; such a
positive statement is expressed in Akkadian by the verb 1e'.
"to be able to" construed with the accusative of the infinitive:
ej,Cam ile"i4 "he is able to build".)
It appears clearly from the correspondences listed above that
there is an inherent ambiguity in the system, because the same
formal devices are used to express different shades of meaning.
On the range and effect of this ambiguity we will return briefly
in the last section of the present article. Here it should be stressed,
as a limitation to my own reasoning, that my analysis is based
not on formal, but on notional or contextual considerations : the
precise meaning of a clause like .fa epi%m, no less than that of a
sentence like I p d , can be gauged only from the connections
which either one of them has with the rest of the discourse.
The classification which follows is therefore based not on formal,
but merely on notional, criteria.
We will now see, then, how the proposed interpretation of the
infinitive after noun or pronoun serves to a better understanding
of several textual passages, even though their general meaning is
normally clear.5 The research is based on the extensive collection
of data to be found in the work by Jussi Aro on the Akkadian
For constructions of this type cf. the references in the preceding note.
For constructions of this type cf. G A G 5 I 5 4 e; AHw 5 63.
3 For constructions of this type cf. G A G (and G A G 2 ) 5 I 5 2 f; AHw
5j9A6.
4 For constructions of this type cf. Aro, 3.14,3.69, etc.
5 In fact, current translations often render the form with the infinitive
precisely as if it were a subordinate clause with a precative, see for instance
Aro, 2.39, who translates a model phrase Sa ept3h as "was zu tun ist, was
getan werden m&/kann/darf
I
2
".
T H E A K K A D I A N I N F I N I T I V E AFTER "SA"
infinitive.' He has gathered some 470 pertinent examples from
all dialects and periods - a corpus which provides a safe basis
for interpretative elaborations. Only a few of the most indicative
passages will be quoted here. In the listing of the examples I will
indicate the role which the noun in the construct state, or respectively the determinative pronoun, has in the correlative finite
sentence. Thus in
bitnm fa epifim "the house he ought to build"
the determinative pronoun corresponds to the object of a
sentence
bitam ZQzd "let him build a house ",
hence the construction will be labelled: construct as object.
Similarly
bit erzbim "the house he should enter"
will be labelled: construct as complement, because in
ana bitim Zirnb "let him enter the house"
the noun bitim is part of the complement of place. Paronomastic
and negative constructions will also be expressly indicated.
Command
Construct as object:
MiZik bziZim bBZi ZimZik, fnmma fa fz2bnrim ana ka~im,annitam ki
annitam beZz Zi@nram (0B)Z "Let my lord decide about the
cattle, let my lord tell me one thing or the other, whether it
ought to be brought over to the steppe" ;
the correlative finite sentence is an advice (miZhm) from the
lord to his servant, i.e. to the writer of this letter: the advice
will be either positive (Zifibir, "let the servant bring over the
cattle ") or negative (ayyiEbir, "let him not. . .").
Construct as object, paronomastic infinitive:
Ja fitdim Ziftd(OB)3 "Let him ask what he ought to ask",
"let him ask what he must".
Aro considers this and a few other phrases as peculiar,4 but they
1 Aro, pp. 30-67. The important review by K. Deller, Or. N.S. x x x ~
( I 962),
zzj-3j adds several NA examples. See also the earlier work by 0. Ravn,
The So-called Relative Clazirses in Accadian or the Accadian Particle fa (Copenhagen, 1941), pp. 9-32, with the important review by A. Goetze, J.C.S. I
(1947)~73-80. On the determinative pronoun see recently F. A. Pennacchietti, " Studi sui pronomi determinativi semitici", Ricerche IV (Napoli,
1968).
2 A R M v 81 : 23-7 = Aro, 2.60.
3 Syria XIX, I I 2: I 7 (Mari letter) = Aro, 2.72.
4 Aro, 2.72.
T H E AKKADIAN I N F I N I T I V E AFTER "SAY'
can in fact be explained within the general framework of the
nominalization of the precative. The pronoun fa is the object of
both the finite form lift21 and of the infinitive fitdim which in
fact stands for another liftd,but nominalized. A good translation
of the phrase with the infinitive is thus the one given by the
original editor, Dossin, who understands fa fitilim simply as
"necessary " ("qu'il prenne les dkcisions ndcessaires").
Construct as object, negative phrase:
Ina af[ta1i i@ziti d s'a s"t;i[m] (OB)' "Among those singers
there is no one whom one should exclude",
or possibly better, given the context of the letter, with the command expressed in the second person: "whom you should
exclude". In either case, our clause would constitute a special
type of nominalization, i.e. it would be the predcate of a cleft
sentence,= which in turn stands for simple: mamman ina as'tali
a_yjiii;i"let them exclude no one among those singers", or in
the second person: mamman 18 tnfegi "exclude no one. . .!"
Construct as complement :
Arnam hbtam fa ina ap1itim nas8him (OB)3 " A serious offence
because of which he ought to be removed from the inheritance" ;
here, as often in similar sentences, there is a consecutive sense
which can be rendered more explicitly in English as : " so serious
an offence that he ought to be disinherited". The correlative
finite sentence would be: ina aph7tim Zissn&v "let (the father)
disinherit him !"
Ah3 k p i m s'aqdim iktas'danni (OB)4 "The time in which I
ought to pay the silver has arrived for me."
An equally good English translation would preserve the use of
the noun phrase for nominalization: "the date for the payment
of the silver "; but the point being made here is that the Akkadian
formulation contains an element of pressure and command
which corresponds to the use of a mood of command in a
correlative finite sentence (here it would be the cohortative:
ZnJqnZ "I ought to pay"). Note that English "the date for the
payment of the silver" would also be the translation of a daerent
A R M I 83: 10-11 = Aro, 2.60.
On the term and the notion applied to an ancient Near Eastern language
see H. J. Polotsky, "Nominalsatz und Cleft Sentence im Koptischen", Or.
N.S. XXXI (1962)~413-30; for Akkadian see a forthcoming article by the
writer, "Of Emphasis in Akkadian ".
3 CH Rev. xii 18-20 = Aro, 2.77.
4 VABVI 217: 7-8 = Aro, 2.5.
I
2
6
T H E AKKADIAN INFINITIVE AFTER "$A"
Akkadian noun phrase, namely a&-nzm ana kaspim faq2Zim.I The
first translation given here is especially meaningful if one looks
at the general context: the sender of the letter has a deadline to
meet for the payment of a debt, and tries in turn to collect a
certain amount of silver from a debtor of his. Hence the reference
to the deadline is not in terms of a statement concerning the
future ("the time in which I will pay", i.e. a&-nzm fa ahqqalz)
but precisely in terms of an obligation which he is forced to
meet ("the time when I ought to pay").
$a taridiya. . .idb[zb]i(OB)2 "They said that by which one
should send me," "they say that I should be sent";
i.e. they say lipzdifz "let one send him ! "
fa dzlppriyama ina yha(x)annitim kalima idabbzb (OB)3
"Absolutely everyone says that by which one should totally
expel me from the office of mayor ", "they say that I should be
expelled ";
i.e. they say : lida#rifz
"let one expel him !"
Construct as subject, negative phrase:
Ninzma zd fa aridi (SB)4 "We are such that we may not go
down " :
this too can be understood as a type of cleft sentence for a
simple sentence with the prohibitive : 12 nzrrad "we must not go
down!" The use of the cleft sentence, and the corresponding
norninalization of the prohibitive, appears in much clearer light
if we consider the context. We are at the beginning of the poem
of Nergal and Ereshkigal, and in the words of Anu the text
wishes to juxtapose the dei szperi with the dei inferi; hence a
special emphasis is placed on the subjects in the descriptions of
the two groups. The "you" of Ereshkigal is opposed emphatically to the "we" of Anu and the celestial gods, and the result
is a sequence of two cleft sentences: att[im]aJ fa elis 'tyoz are
such that you cannot come up, and we are such that we cannot
come down". Obviously this formulation obtains a stronger
effect than a corresponding sequence of two prohibitive sentences: attima l i telli, ninzma l2 nzrrad "you must not come up,
we must not come down".
1 See for instance heddnam ana alak sabim "a deadline for the coming of
the troops" A R M I 43 : 10. On the contrast between the two constructions
see below, pp. 23-7.
2 A R M 11 137: 20-2 = Aro, 2.44.
3 A R M 11 137:34-5 = Aro, 2.44.
4 STT I 28 (An. St. 10, p. rro) I 33' = Aro, 2.64.
5 S T T I 28 (An. St. ro, p. rro) I 31'.
THE AKKADIAN INFINITIVE AFTER %A"
Wish
Construct as complement:
Gabara) yhahqis'zl (0B)I " A revolt as a consequence of which
he may end up in ruin", "whch may bring about his ruin":
a resolution of the infinitive as a finite form Zi& "may he end
up in ruin" is particularly fitting in this context of the Code
where Hammurapi registers his curses against whoever may
damage his stela, and hence uses frequently the precative.
$a ~ibitim.fkiti ZeqEm ep.f (0B)Z "Do that by which you may
obtain t h s wish",
which can be understood consecutively as "act so that you may
fulfil this wish"; one may compare the sentence Sibiti ZzlMud3
"may I attain my wish !"
IsimmZn kearitim, .faadi BZbiZim h.fidim,Jzlrkibam (OB)4 "Load
on the boat the provisions for the hierodules, with which
they may arrive as far as Babylon ", " . . .enough provisions
that they may arrive. . ."
Nambzlrbi. . .Zzlmzln ittiSzl i.tqi (SB)5 "A namburbi rite. . .by
means of which one may put off its evil omen."
Construct as complement, negative clause:
[Namlbzlrbi Zzlmzm ~Eri.. .ana awiZim. . .lZ fe& (SB)6 "A namburbi rite by means of which the evil of a snake may not
approach a man."
Ana biti aiar ZZ a;E Zi.firibiizc(SB)7 "Let them take him to a
house from which he may not come out."
Potential
Construct as object:
Seyhram fa iapzrim tis'zl (OB)8 "You have a boy whom you
could send."
The precative does not have a form for the second person (for
command and wish the imperative is used instead), but an
analogous periphrastic form is constituted by Z . and the present ;9
hence the equivalent finite sentence for the clause with the
infinitive given above could probably be reconstructed as Z .
tatappar "you could send ".
CH Rev. xxvi 60-1
ABPh 115: 21.
KAR 72 : Obv.
=
Aro, 2.30.
4
1-2
=
VS XVI 109: I 2-14 = Aro, 2.43.
VABVI 2: 17-19 = Aro, 2.76.
Aro, 2.23.
O ~ . N . S . X X X V I ~ . ~ I : I - ~ , 7C CTxv113s:j1
~ . A ~ O , ~ . ~=Aro,r.37.
I.
OECT 111 67 (= AbB IV 141) 17-18 = Aro, 2.76.
GAG §81 e ; AHw 5,592.
8
T H E A K K A D I A N I N F I N I T I V E A F T E R "SA"
Jabdram s'a takdlim .hi i .(OA) I " I do not have a boy whom I
could trust."
Construct as object, paronomastic infinitive :
Sdbam !a tarddim abil$md(OB)2 "Let my father send the troops
which he may be able to send"; a more concise translation
would utilize an adjective:3 "let my father send (all) available
troops ".
This translation retains the nuance of potentiality which is inherent
in the construction with the infinitive better than a translation
" . . .all the troops he can "(which properly would be in Akkadian :
s'a t a r d h m ilenZ);4 in practice, however, the latter translation may
often be the best when Akkadian uses a construction with the
infinitive.
$%am s'apaqddim piqdanni (OB)s "Entrust me the sheep which
you may be able to entrust", "give me the available sheep".
$a laqd'e alaqqi (OA)6 "I will take what I may be able to take ",
"what is available".
$a hzld'im akalla (OA)7 "I will save what I may be able to
save", "I will make all possible savings ".8
Construct as complement :
ElepEt ebEriSan8 zll ibafs'i (0B)Q " There are no boats with
which they could cross (the river)."
Note that a little earlier in the same letter there is a slightly
different formulation of what is essentially the same concept:
eleppitzlm ana [ebEri]nizd i[baf!i] '0 "there are no boats for our
crossing", "that we may cross the river with".
Though practically synonymous, the two constructions are
nevertheless different, as I shall try to elucidate below.I1
S m m a ina bitis'a J% paflrim ibas's'i (OB)I2 "If in his house there
is that by means of which he could be redeemed."
$a apilis'zl. . .~ a b a(OB)
t
13 " Take that by which you may pay
him", "take what you need to pay him".
TCL XIX 4: 22-3 = Aro, 2.87.
A R M 11 63: 28 = Aro, 2.86.
O n this use of English adjectives for the construction with the infinitive
see below, p. 27
4 See for instance: fa. . .kIlafu iileS'ti "(a man) who is able to hold it",
A R M IV 16-17.
5 A R M 11 66: 19-20 = Aro, 2.86.
6 CCT 111 7 a: 22-3 = Aro, 2.61.
7 CCT 111 12 b : 6-7 = Aro, 2.72.
8 For this interpretation see B. Landsberger, ZA XXXVIII,
279: "Dieser
verspricht. . .so vie1 s u sparen, als er sparen kann."
9 A R M IV 6: 17-18 = Aro, 2.19.
10 Loc. cit. 10-1 I .
" See pp. 26 f.
12 C H xi 20-2
= Aro, 2.42.
13 TCL XVIII 147: 14-1 5 = Aro, 2.45.
1
3
THE AKKADIAN INFINITIVE AFTER "$A"
11. N O M I N A L I Z A T I O N O F P R E S E N T - F U T U R E
ACTION
The use of the infinitive in the environments indicated, i.e.
after a noun in the construct state and after a determinative
pronoun, is not reserved to the nominalized expression of command, wish or potentiality - the three verbal moods which are
expressed in a finite form by a single mood, the precative. It also
serves for the nominalized expression of a statement of presentfuture action - which is expressed in a finite form by a specific
tense, the present indicative or subjunctive. Following the scheme
utilized above, we may say that
bitzlm Sa epds'im
can also serve as the nominalized equivalent of
bitam @eS "he builds/will build a house".
As a result, then, there are two nominalized constructions which
can be placed side by side,
bitzlmmSa epdSim
}"the house whch he builds",
bitzlm s'a ippeh
and the question is whether the two constructions are wholly
synonym& or whether there is a difference in meaning. In my
judgement, the evidence is normally in favour of the second
alternative, but for a few exceptions where both forms seem
indeed to be equivalent. We shall start by considering the
differences.
The most distinctive one is that nominalization with the
infinitive is clearly preferred when the action of the verb is
considered as generic. The meaning of the term "generic" may
best be explained with some examples. Thus the clause
Sammi dimiparifi (SB)' "the plant with which one stops blood "
is different from
s'ammzl s'a &-ma iparraszl "the plant with which he stops the
blood"
because the latter envisages a specific situation, or at least a
specific subject, whereas the former has no reference to specific
subjects or situations, and simply states the quality of the plant
almost outside of time. In translation, t h s aspect of atemporality
is often best rendered, when one wishes to use a finite verb, by
introducing an adverb such as "normally " or the auxiliary "can ":
warah a r d Gibil (SB)2 "the month in which the fire god
normally descends ";
1
2
CT xrv 36, 79-7-8, 22: Rev. 3 = Aro, 2.22.
Lyon, KeiIschrifttexte Sargons (Leipzig, I 88j), pp. I o, 3 6 : 6I = Aro, 2.3
I.
T H E AKKADIAN INFINITIVE AFTER
"
SA"
jammi da'ini parisi (SB) "the plant with which one can stop
blood ".
Note that the latter is different from potentiality in that the
quality of the plant (in our example) is considered capable of its
effect without emphasizing possible limitations such as the wish
of the subject ("with which one could stop blood, if desired").
The Merence between generic action and potentiality would
probably be expressed by specifying the subject in the latter.
Thus to the clause:
a h l i amiri "a place one (can) not (normally) see"
one would oppose the clause:
ajar l i amiris'u "a place he (could) not see (if he wanted) ".
In turn, both would be different from a clause
aJar li immarzl "a place which he does not see"
because the latter simply gives a statement, without potential or
generic nuances, about both a specific situation and a specific
subject.
The construction corresponding to generic action is used
frequently, with an almost gnomic value, in formulaic expressions, from frozen and stereotyped phrases such as
aban er; (SB)I "stone (amulet) through which one can become
pregnant "
to more original formulations such as
namJaru ~ a g t ufa epC t i b a ~(SB)2
i
"a sharp sword with which
one can give battle".
Especially common is this usage of the construction with negative particles, e.g.
gabal li mabirim (OB)3 "a battle one cannot withstand ";
ajar li amiri (SB)4 "a place which cannot be found ".
An instance of nominalization of generic action may also be
seen, perhaps, in constructions with a deverbal noun instead
of an infinitive. See for instance:
.&adimmati (SB)5 "the one of moaning", "the one who
normally moans ", "the moaner" ;
inzi buqimim (OB)6 "the days in which one can pluck (th,
WOO])
,'.
It is difficult, however, to distinguish between examples where
C A D E 325 = Aro, 2.20.
Streck, AswbanipaZ, VAB VII I 16: 5 6 = Aro, 2.82.
3 CHiii 71-2 = Aro, 2.36.
4 Several examples of this and other similar formulas listed in Aro,
2.37-8.
5 C A D D 143.
6 A R M 11 140:27; apparently taken as an infinitive D by Aro, 2.5.
THE AKKADIAN INFINITIVE AFTER "$A"
the noun has preserved the function of a true deverbal through
which the action of the verb comes to the fore (as presumably
with dimmatu and buqimm), and other examples where the noun
has acquired a lexicalized, i.e. more specific, meaning. The
latter is certainly the case, for instance, in
mdr s'ipri "the son of the mission ", "the messenger" (rather
that "the son whom one can send");
eqeZ iZki "the field of the iZ&-duty" (rather than "the field in
which one goes ").
Clear examples of a "specific" present corresponding to the
construction with the infinitive, and which could be opposed to
the "generic" present just discussed, seem to be very few. As one
such case one may cite:
E
!m aldkiki Suprimma Zzlbdu (0B)I "send me news about your
coming that I may rejoice".
Here the infinitive cannot be interpreted as corresponding to a
generic present ("that you normally come"), nor to a command
("that you ought to come"), nor to a conditional ("that you
might come"). It can be explained only as corresponding to a
precise and specific statement in the present or better in the future:
"that you will (in fact) come". But examples of this type are
indeed rare; normally one would find a subordinate clause with
a finite verb in the subjunctive, such as:
ktma taZZakam s'uprim "write that you will come ".
What remains to be seen are ambiguous cases, in which more
than one meaning may obtain. First, there are instances in which
the ambiguity which is formally inherent in the construction
appears to be intentionally put to use to achieve a certain stylistic
effect. The speaker expresses, as it were, both nuances at the
same time, the nuance of wish/command and that of a statement.
One can consider for example the following passage which is
part of a series of curses appended by Hammurapi to his code:
Stram Zimnam Sa . . .BaZdq mdtisu ZiShnSum (OB)2 "may (DN)
provide him with a bad omen through which his land will
go/may go to ruin" :
the ruin of the land can be envisaged as both a desire on the part
of the speaker, i.e. Hammurapi (ZibZiq "may he go to ruin"), and
a generic statement describing the nature of the curse through the
effects which normally ensue from it (ibaZZiq "one goes to ruin").
Second, there are instances in which the ambiguity is perhaps
I VAB vr 160: I 3-1 5
than Ungnad.
=
Aro, 2.1 I ; translation following Aro rather
CH Rev. xxvii 27-30 = Aro, 2.77.
THE A K K A D I A N I N F I N I T I V E A F T E R c c S ~ l B
only apparent because it seems in fact better to assume that the
speaker wanted to convey a univocal meaning. In the following
sentence
s'a qabifu epuf (OB),I
a translation "do what he says" appears at face value to be the
most logical and simple. In so doing we take the construction
with the infinitive as wholly equivalent to one with a finite verb we would not distinguish, in other words, between s'a qabis'n
and fa iqabbi. If, on the other hand, we consider the two clauses
as a syntactical minimal pair, and try to see a difference between
the two, we may wish to see in :a qabis'u a potential value. We
would then translate :
s'a qabis'n "what he might decide",
a formulation which leaves open the possibility that no decision
be taken; while on the other hand
fa iqabbi "what he will decide"
implies that a decision will in any case be taken, and thus
expresses a greater determination on the part of the speaker.
This interpretation is perhaps supported by the fact that occasionally the expression with the infinitive is accompanied by an
explicit reference to a possible alternative, e.g. :
annitam Z i annitam bBZi Ziijhiram; SVd qabB b?Zga Zdpuf (0B)Z "let
my lord write to me one (decision) or the other; what(ever)
my lord might decide, I will do".
Shortly afterwards, in the same letter, the writer repeats the
same alternative, and then restates his expression of loyalty, but
using this time a finite verbal form, as if to lift any uncertainty
about the fact that the lord will, in fact, send a decision:
s'a bdi iqabbi Zipns'(OB)s "what my lord will say, I will do".
111. N O M I N A L I Z A T I O N O F P A S T A C T I O N
A N D OF CONDITION
There are two significant limitations in the corpus gathered by
Aro: the infinitive after noun or pronoun is used only rarely,
and then under special circumstances, for the nominalization of
either past action or condition. Following our scheme, the phrase
bitnm s'a epdfim
does not normally have the meaning "the house which he built ",
serving, that is, as the nominalization of a sentence
I
2
ARM
ARM
VI 2 6 :
11 2 9 :
Rev. 4' = Aro, 2.69.
Rev. 3'-4', cf. Aro, 2.69.
3
ARM
11 2 9 : Rev.
7'.
THE AKKADIAN INFINITIVE AFTER "SA"
bitam @s' "he built a house",
nor does
bttzim :a damiqim
occur with the meaning "the house which is good" as the
nominalization of
bttzim damiq "the house is good ".
There are, to be sure, exceptions to these general statements,
but they are few, and often susceptible of a special interpretation.
To begin with the verbs of condition, it should be noted
that properly they express conltion only in the stative; when,
on the other hand, they occur as finite verbal forms, they acquire
an ingressive meaning and thus are no longer to be considered
verbs of condition, at least qzia finite forms. The verbs of
condition which occur in the infinitive after noun or pronoun
are few in number. Among the most frequent are such verbs as
badi "to be/become happy" and Qamatzi "to be quick/to act
quickly ". For both verbs, the dictionaries show that finite verbal
forms (with ingressive meaning) are much more common than
statives (expressing condition). It seems plausible therefore that
an ingressive meaning should also be preferred for the construction with infinitive after noun or pronoun. Accordingly, a phrase
like
awit baddka (OB) I
is more likely to mean properly "the words by which you may
rejoice" rather than "the words by which you are happy".
Note, in favour of this interpretation, the comparison between
tdmzim a n n h .fab[adi]ya(0B)z "this information is such that I
rejoice "
and
tt?m alikiki fziprimma Z.,jdzi (OB)3 "send me information that
you will come, that I may rejoice":
in the second example the cohortative Zzibdu, in coordmation
with -ma (virtual subordmation), seems to have the same meaning
as the nominalization with the infinitive in the first example.
Similar is the case for bamgtzi as in the following example:
tdmzim fis'a hamatim (OB)4 "this information is such that it
ought to arrive quickly ",
a cleft sentence for simpler
~imz/m.fiZibmz/t "let this information arrive quickly".
Instead of the construction with the infinitive, nominalization
1
3
VS XVI 57: 36 = Aro, 2.12.
VAB VI 160: 13-15 = Aro,
2
2.1 I.
4
ARM
ARM
IV 29: 32 = Aro, 2.67.
VI 53: 7 = Aro, 2.67.
THE AKKADIAN INFINITIVE AFTER "SAM
for verbs of condition is common with other deverbal nouns,
such as those based on the patterns pirs or pars (e.g. damqa),
or the feminine of the adjective (damiqh), or a formation with
the abstract afformative -it- (dannzita). In all these cases a stative
meaning for the underlying sentence seems proper. Thus while
the infinitives damiqa or danim* do not occur after noun or
pronoun (though not impossible: awit damzqi would mean
"words by which one becomes good"), other expressions quite
common are, e.g. :
amit damiqtimz "a word which is good", "a good word" ;
idZt damqi3 "a sign whch is good", "a good sign";
21 dannzSti4 "a city which is strong", "a strong city".
Nominalization by means of a deverbal noun in the genitive
also occurs for roots whch easily admit finite forms with an
ingressive value, such as badi and barn@, for which we have
already seen attested the use of the infinitive in the genitive; for
deverbal nouns from these roots see for example:
in binis'a s'a biditim (OB)s "with his countenance which is
joyful ";
narkabta fa bamattim (OB)6 "a chariot which is fast ".
In addition, of course, and even more frequently than with a
deverbal noun in the genitive (attributive genitive), nominalization with a verb of condition is obtained by means of the formation which is most characteristic of the attribute of condition,
namely the verbal adjective, e.g.
awitm damiqtzlm7 "the good word".
Incidentally it may be noted that, dependmg on the context,
nominalization of a verb of condition through the use of a
deverbal noun may also refer to condition projected in the past,
e.g.
warkat m[itzi]lz_ya zi baltitga a1 taprasi (OB)8 "you did not
check whether I was dead or alive".
1 The occurrence in a lexical text of the clause mardram fa danlini "to be
bitter (in the sense) of to be strong" (CAD D 83) represents obviously a
special case with a technical, lexicalized meaning, and is only superficially
similar to the cases we are considering here.
CADD65.
3 C A D D I 8 I. Note the interesting contrast between deverbal noun for a
verb of condition and infinitive for a verb of action in the same context:
i&t damqi Ja leqg kiffitz'"a good sign (signifying) that I would attain universal
rule
4 C A D D roo.
5 YOS IX 3 5 i 28.
6 A R M VII 161: 16, a list of objects.
7 C A D D 69. For the difference between nominalization by means of an
adjective and an attributive genitive, see below, pp. 28 f.
8 AbB I 5 3 :8-10.
".
T H E A K K A D I A N I N F I N I T I V E A F T E R "SA"
The second limitation apparent in Aro's corpus is with respect
to past action. Clear examples in which the infinitive after noun
or pronoun is used as nominalization for past action are rare,
and from late periods, e.g. :
s'ipirti epiS ardtsti Sa PN arbis' ina pan Sarri i .takizlda (NA)I
"the news that PN has submitted should arrive quickly in
front of the king".
In the other cases (and these too are few) in which infinitive
after noun or pronoun refers to the past, a special nuance is
present, whereby the construction serves to express either posteriority in the past, or command, or wish, or potentiality.
Some examples follow.
Posteriori0 in the past
Construct as complement :
Mdr S'ipriSzl Ja epZJ arditi z/ naSi bii'ti. . .iQzlra (SB)2 "He sent
his messenger (with promises) according to whlch he would
pay homage and bring tribute."
Command referring t o the past
Construct as object :
Se'am fa ieqika itbaii (OB)3 "They took away the barley you
should have received";
Construct as object, paronomastic infinitive:
Sa naqbn'im iqqebi (OA)4 "What ought to have been said, was
said ";
Construct as object, paronomastic infinitive, negative clause:
Titerlami. . .eriSti hi erESi (SB)5 "You asked of me a question
which one ought not to have asked."
This example shows more clearly than any other how the
formulation with the infinitive is in fact the only one which
allows, in Akkacban, the expression of a negative command in
the past; for, presumably, there is no literal Akkadian rendering
of an English sentence "you should not have asked". In other
words, this is a case where a given linguistic feature, which
from the viewpoint of meaning (deep structure) could also
appear in a finite sentence, appears instead in a nominalized
transform only; a similar, and more important, case is that of the
I ABL 896: Rev. I 5-17 = Aro, 2.33.
2
3
5
Winckler, Sargon, Prunkinschrift 75 : I 52-3 = Aro, 2.83.
TCL XVII 7: 13-14 = Aro, 2.86.
4 BIN IV 79: 10' = Aro, 2.61.
CT xv 47 (Descent of Ishtar) : Rev. 22 = Aro, 2.3 7.
T H E AKKADIAN INFINITIVE AFTER "SA"
agentive, which in Akkadian cannot be expressed after a finite
passive verb, but only in a nominalized transform ("he was
struck down by the wall" in Akkadian can only occur as
imma& "he was struck down", necessarily omitting the reference to the agent; but the nominalization mak~amigirirn' "hit by
the wall" is possible).
Construct as complement :
Affzlm tim ipir ;dim watriti (?) nadi~imfa tafpurinim (OB)2
"Concerning the notice you sent me, accordmg to which I
should pay the rations of the extra workers. . .";
Ja dfi'ikika ti$zlf(OA)3 "You did (such a thing) for which one
should have killed you."
Wish referring to the past
Construct as complement :
[Ilna panItim fa li balgtoa [awillm. . .idbd (OB)4 "In the
past that man said (such things) by whch I may not live",
"spoke so that I may not survive".
So ekime mi;ri_ya i@wa m&- f$ri (SB)5 " H e sent a messenger
according to whom they should conquer my territory."
Potentialio in the past
Construct as object:
Mimma fa Leq; PN .uL ibfi (OB)6 "There was nothing which
PN could take" ;
Construct as object, paronomastic Infinitive :
Ward& fa dikim iddzlki (OB)' "They have killed (all) the
servants of his they could" ;
A@ zl immeritim. . .Lz mafibim imfzlhi (0B)g "They stole the
oxen and sheep they could" ;
Construct as complement, paronomastic infinitive:
Atanbarn fa tzlirima atiram (OA)9 "I became tired and came
back in whatever manner I could", "as best I could".
As with nominalization of present action, so for past action too
one can find deverbal nouns as a nominalizing device, e.g.:
I
2
3
4
5
6
8
Cf. von Soden, J.N.E.S. XIX,165.
UCP IX, p. 364, 30: 6-7 = Aro, 2.11.
CCT IV gb : 24 = Aro, 2.68.
A R M v 4 : 9-10 = Aro, 2.44.
Winckler, Jargon, Prunkinschrift 65 : 3 I = Aro, 2. 5 I.
TCL x 34: 14-15 = Aro, 2.42. 7 A R M 11 74: Rev. 7' = Aro, 2.86.
A R M IV 80: 4-5 = Aro, 2.86.
9 BIN IV 70: 11-16 ='Are, 2.72.
THE A K K A D I A N I N F I N I T I V E AFTER "SA"
[uItu] imi rzlqzlti ~ibitAffur (SB)' "since the remote days when
ASSur had been taken. . .".
But even expressions of this type are rare. The conclusion then
seems inescapable that the only regular type of nominalization
for past action is by means of relative and subordinate clauses.
IV. S U M M A R Y
The cases of nominalization considered in the present article
have all one feature in common, namely that in the resulting
nominalized transform the predicate of the correlative finite
sentence never appears as the construct, or head of the construction. This may best be explained by considering the following
examples as representative of the material studed:
Finite sentence
-
Correlative
- noun phrase
/
Head Modifier
bit(m .La) ippek "the house which he
I.
bitam ippes'
builds "
bitum fa epifim "the house of building"
bit(zcm fa) irrubu "the house in which he
2. ana bitim irrub
enters "
bit(m .I%)eribim "the house of entering "
(bitum dannm "the strong house"
bitam dan 1bit
3.
dunnim "the house of strength "
The terms "head" and "modifier" have been introduced to
serve as uniform labels for the two constituents of the noun
phrase: they refer to the morphological characteristics (surface
structure) of the noun phrase, whereby "head" is the governing
substantive or pronoun, and "modifier" the subjunctive, genitive or adjective which is governed by the c < head". The feature,
then, which is common to all cases of nominalization studied in
this article is that the modifier in the noun phrase corresponds
always to either the object, or the complement, or the subject of
the correlative finite sentence, never to the predicate. We have
not been considering, in other words, the inverse type, which is
also possible and quite common, namely:
I.
bitam ippei epif bjtim "the building of the house"
2. ana bitim i m b
erib bitim "entering in the house"
bitum
dan
dunni bitim "the strength of the house"
3.
N
I
1
N
N
N
Sargon, VS I 7 I : left side 3 2, cf. J. Lewy, H.U.C. A. XIX,466, nn. 203-4.
THE AKKADIAN I N F I N I T I V E AFTER
"SA-
Here the head of the noun phrase corresponds in each case to
the predcate of the correlative finite sentence. It may be noted
in passing that with this type of nominalization there is no
morphological device to express a correlation with a precative
in the corresponding finite sentence, i.e. there is no construction
similar in structure and function to
bitz/ma: epBfim "a house which should be built".
Instead, a new lexical item in the form of a substantive has to
be introduced; for example, in the phrase
eris'ti bitim epzs'im "the desire to build a house" I
the substantive in the construct state eris'ti governing the infinitive epBs'im could be considered as the nominal equivalent of the
precative in the sentence
bitam lqz/s'"may he build the house".
Considering now only the constructions which have been
studied here, i.e. those in which the construct (or "head")
corresponds to object, complement or subject of the correlative
finite sentence, the following conclusions emerge (dsregardng
here the exceptions which have appeared to be of minor importance).
(I) A sentence expressing command, wish or potentiality
(hence employing the precative or other command moods as a
finite verb) is regularly norninalized as a noun phrase consisting
of an infinitive after noun or pronoun, e.g. ad& s'aqdim "the
time in which I ought to pay".
(2) The same type of nominalization is also used when the
predicate of the correlative finite sentence refers to presentfuture action with generic or universal meaning, e.g. qabal l2
mahirim "a battle one cannot withstand".
(3) If the predicate of the correlative finite sentence is a verb
of condtion, the resulting noun phrase consists of either a
substantive and an adjective (e.g. Jarrum dannm "a strong king "),
or a substantive and a deverbal noun ( d danngti "a city of
strength ","a strong city ").
(4) If the predicate of the correlative finite sentence is a verb
referring to a specific present-future action, or to past action,
the resulting noun phrasenormally consistsofa relative clause, e.g.
bit ippe:ztCc the house hebuilds "and bit pz/s'i/"the house he built ".
I For attested examples see, e.g., eriJti t i F iabfm "(divine) desire that
water be drawn (for a libation)" YOS x 5 I i 30 (OB omen); bi~ihtih ~ ~
~amddim"the wish that a chair be prepared" VS XVI 167: 7-8 = Aro, 2.29
(OB); on the second example see however CAD 8 204.
Constituents
of finite
sentence
Nominalization
Finite sentence
Object;
transitive
predicate
bitam ipuS
he built the house
spec.
bitam lipuS
may he build the house
ana bitim irub
he entered the house
Deverbal noun after
construct or Ia
bit(zrm Sa) r?pzrSu
the house he built
bit(um Sa) ippeSu
the house he builds
bitam @eY
he builds the house
Complement ;
intransitive
fientive
Infinitive after
construct or fa
Relative clause
Adjective and
participle
bitum epSum
the built-up house***
bitum Sa epdh'm*
the house one can build
the house he should build
I
b f t ( m Sa) frubu
the house he entered
spec. bit(zrm Sa) irrzrbzr
the house he enters
ana bitim irrub
he enters the house
bit(zrm Sa) eribim
the house one can enter
the house he should enter
ana bitim h u b
may he enter the house
Subject; stative
bit dunnim
the house of strength
bittrm dun
the house is strong
Subject ;fientive b i t m zrballit
the temple gave life
bitum uballat
the temple gives life
bitum liball$
may the temple give life
bit(t/m Sa) uballitzr
the temple which gave life
bit(tim Sa) zrballafu
the temple which gives life
b i t m dannm
the strong house
bitum mubalIi.tum
the life-giving temple
-**
T H E AKKADIAN I N F I N I T I V E AFTER "SA-
The various data may be tabulated in the form of a syntactical
paradigm, whch faces. As often in paradgms, not all constructions give good sense; but it seems useful to keep the same
lexical items throughout, so as to place in better relief the
variables where they occur. The first column includes indcations concerning the nature of "head" and "modifier" in the
correlative finite sentence, and more precisely whether the
"head" corresponds to object, complement or subject, and
whether the "modfier " corresponds to a transitive, intransitive
fientive or intransitive stative predicate.
Some considerations are appended in notes marked by asterisks
on the chart. The single asterisk (*) calls attention to the fact,
already noted by Are,' that the infinitive does not occur after a
noun in the construct state when this stands for the object of the
correlative finite sentence; in such instances, the pronoun fa is
always used - hence we do not find
bit epifirn
but rather only
biturn s'a epijirn
in the sense of "the house which he ought to build". The only
exception, which numbers, however, many examples, is with
negation, e.g.
afar 12am&i "a place one cannot find ".
The second note (**) is meant to emphasize the lack, or at
least the rarity, of the construction with the infinitive when the
construct corresponds to the subject of the correlative finite
sentence and the verb is fientive, either transitive or intransitive.
Even though in principle there seems to be nothmg against a
construction of the type
bft(m fa) buZZz$i "a temple such that it ought to give life",
it is in fact very seldom attested.2 (I exclude of course the subject
of a passive verb, since this is actually the equivalent of the
object of a transitive verb.) For examples with and without
fa, both with a verb of condition used ingressively, see:
.ternurn fz2 la yharngtirn (OB)3 "this information is such that it
ought to arrive quickly " ;
Aro, 2.2.
I t is interesting to note that while the formula Ia ddu7dkiand& "I am one
whom one should kill" (construct as object) is frequent(Ar0~2.66)~
the semantically equivalent Ia mdti and& "I am one who should die" (construct as
subject) does not seem to be attested.
3 ARM vr j 3 : 7 = Aro, 2.67.
I
2
THE AKKADIAN INFINITIVE AFTER "SA"
tim bam<tim (OB)' "the undertaking which ought to take
place quickly
Constructions of this type are mostly used for emphasis in cleft
sentences, as in the first of the two examples just quoted and in
aw&m fzi zd fa bahtim (0B)Z "that man is not such that he
should live", "he is not worthy to continue livingy7.
Finally it should be noted that the expression bitz/mepfiim(***)
has actually acquired a special, lexicalized meaning of "built-on
house plot"; the meaning given in the paradigm is therefore
not attested as such,3 and is simply meant to convey the basic
meaning of the construction, and its connection with the
correlative finite sentence on the left in the paradigm.
The distribution of the data in the preceding parabgm shows
that there is relatively little ambiguity in the use of the forms.
The only overlaps are between the generic present and the
precative as the correlative finite forms of the infinitive after
noun or pronoun. In turn, the precative itself can be interpreted
as referring to command, wish or potentiality, without any
differentiation in form. Such ambiguity, however, can normally
be lifted on the basis of lexical and contextual considerations.
These could perhaps be formalized through the analysis of the
iexical features of the constituents of a sentence - a procedure,
however, which would be rather complex. Intuitively, one can
see at a glance that the difference in interpretation between the
following two sentences is due to lexical, rather than morphological features :
ahr li amiri "a place which cannot be seen",
erijti hi erGi "a request which shodd not be made".
The English rendering introduces a formal differentiation where
for Akkadian the lexical connotation of the constituents and the
broader context in which they occur are sufficient. The broader
context, on the other hand, remains the only criterion for a
choice when Akkadian says :
ajar l i amirj
in the sense of "a place which shodd not be seen", since in this
case both form and lexical connotation remain the same as when
the same phrase is used in the sense of "a place which cannot be
seen". Ambiguities of this type are common in all languages
".
ARM 11 48 : 21 = Aro, 2.12.
ARM v 72 : 5 = Aro, 2.67. See also STT I 28 : i 3 1l.3j', above p. 7.
3 See however bittl @ltl adi gtlltlrdltl adi dakitiJ[tl"a house in good repair,
with its beams and its doors" A D D 324: 6-7 (NA) and passim in ADD,
d.CAD G 141.
2
22
T H E AKKADIAN INFINITIVE AFTER "SA"
and certainly so in Akkadian. One may dunk, for instance, of
the multiple functions served by the genitive (the objective and
subjective functions mentioned at the beginning of this article
are only two among many), or the use of undifferentiated coorlnation with -ma for a variety of subordinating relationships
(temporal, concessive, consecutive, etc.),I or the role played by
emphatic devices, such as inversion of word order, to express
in a veiled manner what the language can otherwise express in a
clearer and more differentiated way.2
In some cases the construction with the infinitive after noun
or pronoun seems equivalent to other constructions equally, or
even more, frequent in the language. We will consider here only
one (ana with the infinitive), to which reference has been made
in the course of the article. Thus it has been noted that the following two pairs seem practically synonymous:
adan kaspim faqiZim iktafad "the time in which I ought to pay
the silver has arrived";
adanam ana k a q i m faqdim iktafad "the time for the paying of
the silver has arrived";3
and
eZeppit ebirts;ni .d ibaffi "there are no boats with which they
may cross (the river) " ;
elefiBtzlm ana ebdrZfz1nz7d ibafjd "there are no boats for their
crossing ".4
Two alternatives present themselves in the analysis of the second
member of each pair.
(I) In the first alternative, the prepositional phrase with ana
and the i h t i v e is adnominal in character, i.e. it constitutes a
single noun phrase together with the noun whch precedes. In
&us case the correlation between the two formulations is particularly close, but they remain, none the less, daerent. The
formulation with ana expresses finality, consequence, etc., but
not properly or specifically obligation, wish or potentiality as is
true of the infinitive after noun or pronoun. Thus in the first
example the phrase
adznzlm ana h p i m faqdim
I See on the subject R. D. Patterson, Old Babylonian Parataxis as Exhibited
in the Royal Letters of the Middle Old Bablonian Period and in the Code of Hammurapi, Ph.D. dissertation, University of California at Los Angeles, 1971.
2 On this interpretation see my forthcoming article "Of Emphasis in
Akkadian". There I also deal more in extenso with the stylistic value of
ambiguity.
4 See above, p. 9.
3 See above, pp. 6 f.
THE AKKADIAN INFINITIVE AFTER "SA"
is best understood as the nominalization of a sentence which
introduces the idea of finality, such as
aaZnum ana kaspim faqdim iffakin1 "a deadline was set in order
that the silver be paid".
It seems dfficult that a different understanding of the phrase be
possible; at any rate it does not seem possible to introduce the
notion of obligation which is instead proper to
adan kaspim s'aqdim,
a phrase which, according to the argumentation proposed here,
is to be resolved as
ina ahinim fu&i kaspam ZifquZ "he ought to pay the silver at
the specified time".
Note especially that a temporal interpretation of
adanam ana kaspim faqdim
is not possible, since ana+infinitive does not occur in a temporal
sense, though ana+noun does.2 If a temporal meaning were
possible, then the meaning of the two phrases in our first pair
would in fact be identical: "the time in whch I ought to pay the
silver" and "the time when I ought to pay the silver7'. The
interpretation in the sense of finality proposed above implies
that the phrase
ah-num ana hspim faqdim
be considered elliptic for a common type of nominalization with
relative clause :
adZnum (fa) ana kaspim faqiZim (ifs'ahu) "the deadline (which
was set) in order that I pay the silver".
In point of fact, this is practically synonymous with
adin hspim JaqdZim "the time in which I ought to pay the
silver ",
but it is important that we become aware of the real difference in
structure and of the potential difference in meaning, since the
latter may become relevant in a given context.
And similarly for the second pair noted. The phrase
eleppitum ana ebirifzlnzi
is best understood as elliptic for some such sentence as
eleppitum Az ana ebt%YunZwasmZ3 "boats which are suitable for
the purpose of their crossing".
In practice, this may well be synonymous with
el@& ebM..unzi "boats with which they may cross ";
I
2
3
For constructions with adzntlm and Yak-ntlm cf. C A D A/I 98 f.
Cf. AHw 47 E 2 ; Aro, 6.1-4.
For constructions with wasdmtl and ana cf. C A D A / z 328 f.
T H E A K K A D I A N I N F I N I T I V E A F T E R "SA"
however, the first formulation puts the stress on the quality of
the boat, the second on the potentiality of the action of crossing.
(The introduction of distinct lexical items such as fahinzl in the
first, and wasbmu in the second example is arbitrary and may seem
disturbing, especially without the availability of living informants. Yet the procedure is justified if the elements deleted have
an ample attestation in environments without deletion. The same
procedure underlies some of the conclusions of traditional
grammar; for example the term and the notion of "possessive
genitive"' implies an understanding of a phrase such as bit
awidim "the man's house" on the basis of some such sentence as
awidum bitam is'zlz "the man owns the house".)
(2) The second alternative consists in taking the complement
with ana as adverbial, rather than adnominal. In other words, the
phrase ana h s p i m faqbdim is governed directly by the predicate
iktafad:
ad&wm ana h s p i m faqbdim iktafad "the deadline has come so
that I have to pay the silver",
and the phrase ana ebirifuni by the predicate ibaffi:
edeppizhm ana ebirifzlni zld ibafii "there are no boats so that
they cannot cross (the river) ".
If so, the difference in meaning with respect to the construction
with the infinitive after construct is more noticeable. In fact the
sentence
edeppit ebirifzlni d ibajfi "boats with which they may cross
(the river) are not here"
stresses the potentiality and desire of crossing on the part of the
subject, whereas the sentence with adverbial complement states
objectively the impossibility of crossing due to the lack of means.
A stronger formulation of the sentence with adverbial complement, which emphasizes the lack of means over the impossibility
of crossing, is with coordination by means of -ma (virtual
subordination) :
edeppitzlm d iba.Gima HZ ibbiri "there are no boats and thus they
will not cross ".
We could set up the following as a syntactical paradigm whch
shows the gradation of meanings from one formulation to the
other :
I . eleppit ebirifzlni d ibai'i " there are no boats with which
they may cross"
emphasis on potentiality and wish of crossing;
1
Cf. CAD I-]
Cf. GAG S136a.
2f
291.
T H E AKKADIAN I N F I N I T I V E AFTER "SAW
eleppitum ana ebirifuni / d ibas'fi (adnominal complement)
"there are no boats suitable for their crossing"
emphasis on the type of boats (there may be boats suitable for
other purposes) ;
3. eleppitzm / ana ebiris'zvni d ibas'fi (adverbial complement)
"there are no boats so that they cannot cross"
emphasis on the impossibility of crossing (for lack of means);
4. eleppitzm ul ibas'fima ul ibbirg "since there are no boats
they cannot cross "
emphasis on the lack of means (which prevents the crossing).
Similarly in the sentence
ad& kaspim faqdim iktafad "the time in which I ought to
pay the silver has arrived"
the adnominal complement with the i h t i v e defines the nature
of the deadline, whereas the use of an adverbial complement
in
ad2num ana kaspim faqdim iktas'ad "the time has come so that
I must pay the silver"
stresses the need of paying as the result of a given circumstance,
which, however, is not described further (of it the speaker could
say by whom it was set, for when, etc.).
A choice between the two alternatives is difficult on the basis
of only written documents, but it seems probable that living
speech differentiated between the two by means of pauses and
intonation :
eleppitz/m-ana-ebirifuni / al ibaffi (adnominal complement)
eleppitum / ana ebirifuni / ul ibaffi (adverbial complement).
There are in the language two other devices which, when used,
remove the ambiguity otherwise inherent in the writing system,
namely word order and the determinative pronoun s'a:I
eleppitzm fa ana ebiris'uni d ibajfi (adnominal)
ana ebirifuni eleppitzlm d ibafji (adverbial).
When neither one is used, the ambiguity seems insurmountable
for the reader of a written text, unless the context is sufficient
as a clue. Because of the existence of formally defined adnominal
complements (types eleppit ebirifuni, eleppitzlm fa ana ebirifuni) 1
would tend to think that when these devices are not used, and
2.
I The latter device has also been pointed out by J. Aro, "Prapositionale
Verbindungen als Bestimmungen des Nomens im Akkadischen", Or. N.S.
X X X I I (1963)~402. He also mentions word order, pp. 399-401, but only in
connection with partitives, where inversion can occur even within the noun
phrase.
T H E A K K A D I A N I N F I N I T I V E A F T E R "$A"
the context gives no clue, we have in fact an adverbial complement ; I I would like in other words to read normally:
eleppztum / ana ebIris'unI / d ibas's'e.
But the basis for such hypothesis is admittedly very small. In
concrete cases, a choice between the two alternatives may not
always be necessary, since the nuances which I have tried to
elucidate may in most cases have but a minimal influence on the
meaning. In any case, what mattered here was to describe
various structures which are in fact formally different.
A good reason for the productivity and specialization of the
construction with infinitive after noun or pronoun is, as we
have seen, the impossibility of using the precative in a relative
or subordinate clause: a phrase like bitztm s'a *l@us'u is in fact
impossible. But there is another reason for the productivity of
the type s'a epzs'im, and this is the lack, in Akkadian, of gerundive
adjectives. Akkadian is generally poor in adjectival formations,
whether derived through internal (type paris, pums) or external
inflection (type -in-, -&); and none of these formations expresses
the command, wish, potentiality or possibility that a certain
action be performed. There is no equivalent, in other words, of
such formations as English -able (e.g. in "acceptable") or Latin
-end- (e.g. delendum " to be destroyed ") - no equivalent, that is,
other than precisely the periphrastic construction with infinitive
after noun or pronoun. Thus
;u@ram s'a takilim li i.u (0A)Z "I do not have a boy whom I
could trust"
can equally well be translated as "I do not have a trustwort&y
boy" (though indirectly a "trusted" boy, Akkadian taklum,
can also be considered "trustworthy", since experience in the
past is taken as a warranty of future performance); or again:
s'arrit la s'anin (SB), "a reign which cannot be duplicated",
"an incomparable reign ".
In effect, many of Aro7stranslations in his book on the infinitive
use precisely adjectives of the type mentioned, and quite
properly so, e.g. :
'la1 abikim (OA)4 "hin~uschaffend";
s'a lZ akili (SB)5 "unel3bar ";
s'a li nakdr (SB)6 "unabanderlich".
Also to Aro, "Verbindungen", quoted, p. 402, this seems a tempting
hypothesis.
TCL XIX 4: 22-3 = Aro, 2.87.
3 See several examples in Aro, 2.38.
4 BIN VI 109: 29 = Aro, 2.87.
5 IV R 63 iii 41 = Aro, 2.92.
IV R 62 Rev. No. 2: 45 = Aro, 2.96. See also above, p. 9.
27
T H E AKKADIAN I N F I N I T I V E AFTER %A"
It may be noted in t h s connection that a similar function is
also served by the "attributive" genitive of the type ddannz7ti.
This type is practically the only one available in the language
when the "modifier" is a noun derived from a nominal, rather
than a verbal root - for in this case the adjective cannot be
derived by means of normal patterns through internal inflection.
For example, the noun ilnm "god", being a nominal root, does
not admit of adjectives derived throughinternal inflection(such as
paris). There is, it is true, an adjective derived through external
afformatives, il&i (i.e. il-&-i-n), but this is rare and with a special,
lexicalized meaning, "prosperous, lucky". Hence the normal
way of expressing an adjectival relationship is by means of the
abstract noun appended as a genitive, e.g. M a t i l i t i ~"&vine
abode". The same construction is also found when a regular
adjective is available through internal inflection, as with Zl
danniti "strong city" next to which we find Jarrnm dannm
"strong king". The Merence between the two constructions
is not immediately clear. As a hypothesis one may suggest the
following distinction. The type .&arrmdannwz corresponds to a
finite sentence in which the head (Jarrum) appears as the subject
and the modifier (dannm) as the predicate, i.e. s'arrm dan "the
king is strong". The type i Z danniti, on the other hand, corresponds to a finite sentence in which the head (dzim) appears as a
complement, and the modifier (danniti) as any other element of
the sentence - in our case, also the predcate: ina alim danni
"in the city they are strong". If so, then d danniti would
properly mean "the city in which one is strong, one feels
secure" (and iZ.m dannm "the foruhed city"). Similarly with
other roots:
aw&m damiqtzlmz "a good word ", "a good thing"
vs. sii- dmqi3 "an omen according to which a good thing (will
happen) ";
Sarrzlm kinzim4 "true, legitimate king "
vs. d q y h kitti[m]s "a judge throngh whom justice (takes place) ";
awilzlm _had9 "a happy man"
vs. 'im _hiditil7 "a day in which one is happy".
1
3
4
5
C A D I-J 105.
2 E.g. YOS x 47: 7 (OB).
CT xxx~v3 I ii 5 6 (NB), and often with words for omen, C A D D I 8 I.
TCL 111 114 (SB, Sargon).
VAB VI 218: 27 (OB).
Oppenheim, Dreambook, p. 3 I 3 : ix x+ 6 (SB).
KAR I 77 Rev. ii 41 (SB).
THE AKKADIAN INFINITIVE AFTER "SA"
Needless to say, in English, which is much more liberal in the
use of adjectives than Akkadian, we would translate on most
occasions with an adjective, not word by word with a substantive: "a good omen" (not "an omen of goodness "); "a just
judge" (not "a judge of justice"); "a happy day" (here also
possibly "a day of happiness"). The genitives from nominal
roots also fall in this category: in fact the type discussed above,
Subat iZiti, does not mean "the dwelling is god ", but rather "an
abode in which god dwells ", hence, in English, "a &vine abode".
And similarly, with another primary noun:
tabtzi. . .?a abbz7tiI "goodness which is proper for a father",
hence, in good English, "fatherly goodness".
How far these considerations would apply I cannot say without a thorough analysis of the evidence. But this type of norninalization is essentially different from the one with which we have
been concerned in this article - properly only the norninalization
with the infinitive in the genitive - and thus can be left aside for
another study.
CT XXII N. 43 : 21-4 (NB).
JOURNAL O F SEMITIC STUDIES
Extract from
JOURNAL OF SEMITIC STUDIES
Vol. 17, No.
I,
Spring, I 972