Printer Friendly

Sentential complementation in Akkadian.


GUY DEUTSCHER'S RECENT BOOK comprises a number of studies that treat different facets of sentential complementation in Akkadian. Following a short introductory chapter, the second chapter is a general discussion of the term "sentential complementation." The third is an introduction to Akkadian for the non-specialist reader, including a short but useful section devoted to the language of the letters. The fourth chapter is an attempt to prove that finite complements marked by kima originate in historically adverbial clauses of the same structure. Chapter 5 describes the evolution of the quotative marker umma X-ma. Chapter 6 introduces the functional domain of complementation. Chapters 7 and 8 describe the complementation of verbs of perception and knowledge and verbs of manipulation, respectively. The ninth chapter treats indirect questions. Chapter 10 adduces cross-linguistic parallels to the Babylonian development. The last chapter discusses putative reasons for the development of finite complementation in Babylonian.

The book as a whole is a thorough, large-scale investigation of several interrelated issues in the diachronic syntax of Babylonian. The resulting diachronic statements are obtained only after the synchronic facts regarding the issues in question have been recovered and stated. Some synchronic facts of Old Babylonia syntax are actually revealed for the first time in this book. The comments to follow stem mainly from the reviewer's different interpretation of the material or are the result of a disparate linguistic point of view.


What is referred to by Deutscher as "sentential complementation" should more precisely be regarded as substantival clauses, for such clauses, in addition to their function as objects, can commute (i.e., they can serve as subject, object, appositives, etc., at least in principle) with an infinitive anywhere in the text. Viewing these clauses as arguments is perhaps a bit too extreme (since verbal arguments may just as well be adverbial, e.g., "behave well," "se comporter comme il faut"), and does not account for that-clauses adjoining a substantive, viz., "the fact that you came," etc. (if we remember that the substantive fact has no parallel verb synchronically requiring such complementation, nor is it a sentence).

Chapter 4 is an attempt to show that complement kima-clauses emerged from adverbial kima-clauses via a semantic bleaching process of kima. Such proposed bleaching is a mirror of the same ever-debated issue in regard to Indo-European languages. Deutscher derives the conjunction kima from the preposition kima, taking the latter as diachronically preceding the conjunction (p. 38), probably on semantic grounds. (Note that while the element k- in Semitic is indeed comparative, it also has a co-temporal value.)

The hard evidence indeed shows that kima-clauses are not attested as substantival clauses in Old Akkadian. This conforms well to another view stating that there had been, in pre-Old Akkadian times, a particle *ma / ma having just this function. (1)


Deutscher says (p. 50) that verba dicendi are weakly transitive or intransitive. It is true that verba dicendi do occur with no nominal objects (e.g., direct speech, or with no object at all) but they do just as frequently appear with objects; note, in past chains, their co-occurrence with infinitives. It is normal for a verbum dicendi to take a direct citation as its object, (2) which does not make it any less transitive. Transitivity in Akkadian means the capacity, rather than the requirement, to have an object. The verb nadanum ("to give"), a natural double-object verb, often comes with merely one, at times even no object at all. Nevertheless it is not to be considered an intransitive. A paratactic structure like qibisum-ma lillik ("tell him he should go," lit. "tell him and let him go") does not render the verb qabum ("tell, order") any less transitive, just because it does not have a formal object. (What is said is quite clear from the second clause.) Transitivity in Old Babylonian is about (in)completeness of information. Other strategies do fill in the content of this allegedly "missing information."

Another historical path is illustrated by verbs of proof, where a kima-clause functioning as object allegedly starts as a comparative particle. In the first example, kima warassu ukangu-ma ("he will prove that (he is) his slave" (3) CH [section] 282), warassu is a sentence, for if it were not (in other words: if it were, as Deutscher claims, a preposition meaning as), we would have had kima wardisu (= "as his slave"), rather than the full sentence. The rest of the examples are no different from the first; they all exhibit a kima-clause.


The quotative construction starts as an independent clause, becoming a quotative marker and finally (Neo-Babylonian) a general complementizer (pp. 66-67). In Old Akkadian it is a marker of (direct) speech, not occurring with verbs of speech. In early Old Babylonian the construction begins to co-occur with speech verbs (usually, but not always, paratactically). In Old Babylonian it is still independent, but becomes almost obligatory after speech verbs. It is still a direct-speech marker, as is shown below. Deutscher's use of the term "reported speech" to refer to direct speech is confusing, for in other linguistic works it is rather the parallel of indirect speech. (4)

Deutscher (p. 68) holds that the translation of umma as "thus" is erroneous. The problem stems from different points of view: diachronically umma, stemming from Old Akkadian enma, is probably a presentative, (5) or at least very similar to one. Synchronically, however, it is nothing but a part of the quotative construction, and therefore does not have an individual value. Incidentally, the particle -mi is different in all respects from the quotative construction, for it does not mark the construction as quotative but is rather, at least in Old Babylonian, an alloform of the focal -ma in direct speech. (6)

Deutscher mentions that in Old Akkadian the construction does not co-occur with verba dicendi (p. 69), and that in early Old Babylonian (7) the construction occurs after verbal lexemes other than qabam (p. 73). He reasons that it is redundant (for both the verb and the construction mean supposedly "to say"). It seems, however, more plausible that qabum, much like the roots [square root of bhl] (Eth.), [square root of mr] (Heb.) (8) and [square root of qwl] (Arab.) in the various ancient Semitic languages, initially required direct speech, whereas the other verbs mentioned (saparum "write, order in writing," apalum "answer") did not. At later times, when qabum started to take kima-clauses as well, there was reason for it to take the quotative construction as well, as because by then it was no longer a verb denoting only direct speech. Once qabum is capable of occurring with a kima-clause, it is susceptible to being accompanied by the by the umma X-ma construction.

Deutscher claims that "When [...] follows speech verbs its independence is clearly eroded" (p, 76). An example of indirect speech following -ma is in order, to show that there was, beginning at some point in Ok Babylonian, a choice between direct and indirect speech both connected via -ma to the verb qabum:
 ana PN aqbi = ma umma anaku = ma inika lullik "I told PN
 lilt. 'and']: 'Let me go with you,'" Deutscher, example 115

 (1. 11) alpa arkia ... addinsum--ma ... kanik simatim
 nustezib ... (1.18) kiam aqbisum iqabbu--ma alpa arkia ...
 iddinnnikku u kanik simatim taddin "I gave him one rear
 ox ... and we had a sales document drawn up... I spoke to
 him as follows: 'they say that (lit. and) they gave you a rear
 ox ... and that you handed over a sealed document." AbB
 12, 5:11-21

Here the writer had the choice between qabum + -ma + direct speech and qabam+ -ma + indirect speech This choice means that the quotative construction retain the value of direct speech + speaker, for it is opposed to indirect speech in the same syntactic conditions. Hence there is no semantic erosion.

The construction corresponds--whenever it is no obligatory--to our combination of colon (:) and quotation marks ("--"); in addition, the speaker is referred to by a (pro)nominal element. This construction has another function: it makes it possible for the direct speech to fit into the chain. Actually the construction is a kind of converter which enables this, often in opposition to indirect speech.

Deutscher (pp. 88-89) attempts to show a similar development in Biblical Hebrew regarding the grammaticalization of [square root of mr]. However, the roots [square root of mr], [square root of ngd], and [square root of dbr] co-exist, belonging to the same system, rather than to different phases. The difference between them is whether they take direct or indirect speech as complement. (9) The first ([square root of mr]) is a direct speech verb. The rest are not necessarily so, and to use them with direct speech one had to add some form of [square root of mr] as a marker. The grammaticalization process mentioned as regard Biblical Hebrew is therefore not compatible with the linguistic facts.


Deutscher claims that the difference between manipulative and regular verbs is in the construction employed (p. 97). However, wasabsu ina alim iqbunim could mean either manipulation ("they told me he should stay in town") or report ("they told me of his staying in town"). Abstract nominalization with -ut- is said to serve in the same context as infinitives. It is true that in examples 229 (wasabsu ... iqbunim) and 230 (wasbussu iqbunim "they told me of his staying") the context is identical, but the former could serve in manipulative contexts whereas the latter could not, and this is the most notable difference between them.

Two different structures, viz., infinitive construction and parataxis, are divided (according to Deutscher) on a temporal basis: past actions take infinitive whereas non-past manipulations take parataxis (p. 124). However, since the praesens does occasionally take infinitive complement (ana durim erebam mannum iqabbikum "Who tells yon to enter the wall?" AbB 9, 140:13-14. Note, in addition, that except for example 266 Deutscher's examples are all directives, i.e., precatives, imperatives, and cohortatives), it would be better, so it seems, to draw the line between directives (which tend with rare exceptions to interconnect paratactically) and non-directives (which tend to take infinitive constructions).

Deutscher states (p. 130) that the use of nominal constructions with the infinitive complement tends to have non-manipulative meaning (wasabsu ... iqbanim), whereas manipulative meaning is effected via verbal complementation (... se am makasam iqbi"... he said to collect the barley"). He admits that manipulation occurs with nominal (i.e., genitival) rection, but more rarely. The adduced examples are not always well chosen: in example 292, b[al]assu iqtabi, lit., "he ordered its li[v]ing," and in ex. 293, alaksunu ... iqtabi, lit., "he ordered their going ...," both lexemes are intransitive, and the genitival ending has no alternative. (A substantive in the nominative as the agent of the infinitive does not exist, as far as I know, in Old Babylonian). (10) In discussing transitive verbs, another problem comes up: the accusative rection is always the object of the infinitive whereas the genitival bound pronoun neutralizes both subject and object. So there is no actual symmetry between the two options.

Further it is mentioned that the assum infinitive construction, in opposition to the accusative infinitive construction, is a paraphrase of the actual command. same in fact holds for any infinitive construction serving as an object in Akkadian. That is exactly the difference between uttering a directive ("He ordered: 'go!' ") and reporting the utterance ("He ordered [...] to go").

Contra Deutscher (p. 132), there are some rare examples of a manipulative directive with an infinity complement:
 eqlam suati la epesam PN sudki "Convince PN not to cultivate
 that field." AbB 3, 2:41-42
 ana bit tuppim alakam suhissu "Instruct him to go to
 school." AbB 2, 81:29


Deutscher views the development of sentential complementation as a trend, showing that there are parallel examples in other languages. This trend, however, could be reversed. That is, such use could decline just as well. (11) Even more complex systems, such as verbal systems, can deteriorate as well as be rebuilt anew. (Note for example, the path from Northwest Semitic, with a prolific verbal system, to Syriac, which is relativity poor in forms, to Eastern Neo-Aramaic, which has a particularly complex verbal system.) Further, attributing this development to the growing need for more complex patterns of communication could turn out to be a problem: it inherently precludes every discussion about such previous evolution.

A simple example to rebut this reasoning would be contemporary Turkish--a modern language, certainly coping with modern needs and complexity of communication--which has relatively little finite sentential complementation (via ki), and whose exponents of embedding complementation are rather various participials (by way of -dig- and -eceg,-). Languages do cope with intricate communicative contents by means of direct speech or participials, which are felt to be less suitable (so Deutscher p. 167) only from a modern, Northwest Indo-European perspective.


Deutscher does not differentiate, in glossing, between the connective and the focal -ma. The latter is not given any consideration in translation (example 174, p. 105, and other places as well).

Deutscher, on a few occasions (pp. 75, 181-82), discusses "the colloquial language of the letters," but we do not really know what spoken style was, and no spoken style would have remained almost frozen for three centuries. (12)

Transcription of Biblical Hebrew (e.g., pp. 89, 133) could have been done much more carefully.

Deutscher states that "in emphatic contexts ... where the verb 'know' is preceded by the emphatic particle lu only asyndetic parataxis occurs ..." (p. 106). First of all, lu is not an emphatic particle. In example 180 it is part of a precative syntagm. This is important because precative forms do not generally tend to take kima clauses as objects. (There are very few examples, such as 204 and 211, given here.)

As to indirect questions in the corpus, the following might serve as a representative, a precursor to this phenomenon in Middle Babylonian:
 assum beli atta warkat bitika parasam tele u ul nide "We
 do not know whether you, my lord, can investigate the case
 regarding your house." AbB 6, 153:14-17


This book is a most important contribution to the historical description of the Babylonian branch of Akkadian, with regard to the syntax of complementation and concurrent strategies. Despite minor technical flaws, Deutscher's book is the most exhaustive statement of this complex issue so far, both diachronically and synchronically, Since large-scale treatments of Akkadian syntax are quite few, it is hoped that more works of this scale and quality are planned by the author.

This is a review article of Syntactic Change in Akkadian The Evolution of Sentential Complementation. By GUY DEUTSCHER. Oxford: OXFORD UNIVERSITY PRESS, 2000. Pp. xv + 204. $70.

(1) E. Cohen, "Akkadian -ma in Diachronic Perspective," ZA 90 (2000): 207-26.

(2) See F. R. Palmer, Mood and Modality (Cambridge, 1986), 134.

(3) J. Huehnergard, Key to a Grammar of Akkadian (Atlanta, 1997), 70; and cf. CAD K, 168a.

(4) Cf. Palmer, Mood and Modality, 134.

(5) Cf. inni in R. M. Whiting, Jr., Old Babylonian Letters from Tell Asraar (Chicago, 1987), 78-79.

(6) Already A. Finet, L'accadien des lettres de Mari (Brussels, 1956), [subsection] 50g, 100h, regarding Mari texts.

(7) E.g., in the texts studied in Whiting, Tell Asmar.

(8) See G. Goldenberg, "On Direct Speech and the Hebrew Bible," in Studies in Hebrew and Aramaic Syntax Presented to Professor J. Hoftijzer on the Occasion of his Sixty-fifth Birthday, ed. K. Jongeling, H. L. Murre-van den Berg, and L. van Rompay (Leiden, 1991), 79-96, esp. 81-87 (rpt. in Goldenberg, Studies in Semitic Linguistics [Jerusalem, 1998], 197-214).

(9) See Goldenberg, "Direct Speech," 85.

(10) Two examples of accusative as agent are listed in J. Aro, Die akkadischen Infinitivkonstruktionen (Helsinki, 1961), 89, as objects of the auxiliary nadanum "to let."

(11) See Cohen, "Akkadian -ma," regarding the prehistoric use of -ma as a substantivizing converter, much like substantivizing kima in Old Babylonian.

(12) W. Sallaberger, "Wenn du Mein Bruder bist ...": Interaktion und Textgestaltung in altbabylonischen Alltagsbriefen (Groningen, 1999), 4-6.


COPYRIGHT 2002 American Oriental Society
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2002 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Cohen, Eran
Publication:The Journal of the American Oriental Society
Date:Oct 1, 2002
Previous Article:Notes on a new volume of old Assyrian texts.
Next Article:A new Coptic grammar.

Terms of use | Copyright © 2017 Farlex, Inc. | Feedback | For webmasters