The Syntax of Neo-Aramaic: The Jewish Dialect of Zakho 9781463234737

This monograph provides an extensive syntactic description of the rather well-known but not previously described Jewish

161 77 9MB

English Pages 490 Year 2012

Report DMCA / Copyright


Polecaj historie

The Syntax of Neo-Aramaic: The Jewish Dialect of Zakho

Citation preview

The Syntax of Neo-Aramaic: The Jewish Dialect of Zakho

Gorgias Neo-Aramaic Studies

13 Series Editors Geoffrey Khan Hezy Mutzafi

The purpose of the Neo-Aramaic Studies series is to publish volumes relating to the modern dialects of Aramaic or the literature of the Christian and Jewish speakers of these dialects. The volumes will include linguistic studies relating to any of the dialects or to groups of dialects.

The Syntax of Neo-Aramaic: The Jewish Dialect of Zakho

Eran Cohen


34 2012

Gorgias Press LLC, 954 River Road, Piscataway, NJ, 08854, USA Copyright © 2012 by Gorgias Press LLC

All rights reserved under International and Pan-American Copyright Conventions. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning or otherwise without the prior written permission of Gorgias Press LLC. 2012



ISBN 978-1-60724-048-8

ISSN 1935-4428

Library of Congress Cataloging-in-Publication Data Cohen, Eran, 1967The syntax of neo-Aramaic : the Jewish dialect of Zakho / by Eran Cohen. p. cm. -- (Gorgias neo-Aramaic studies) Includes bibliographical references. 1. Syriac language, Modern--Syntax. 2. Syriac language, Modern--Dialects--Iraq--Zakhu--Grammar. I. Title. PJ5802.C64 2012 492'.35--dc23 2012018425 Printed in the United States of America

TABLE OF CONTENTS Table of Contents Preface


0 Introduction 0.1 History of research 0.2 Method 0.3 The corpus 0.4 Structure 0.5 General practices 0.6 Abbreviations 0.7 A short glossary of terms

1 1 3 5 8 9 10 11

Part I: Microsyntax: intraclausal syntactic relationships


1 The predicative relationship 1.1 Verbal forms 1.2 Thematic nominal groups 1.2.1 Nouns and determination 1.2.2 Pronouns 1.2.3 Clauses and infinitives 1.2.4 Adjectives and adjective clauses 1.3 Copular expressions 1.3.1 The present copula 1.3.2 the verb hwya 1.3.3 The presentative copula 1.3.4 Other copular expressions: pšle 1.4 Dependent nexus 1.4.1 Object nexus 1.4.2 Adjunct nexus 1.4.3 Subjunct nexus 1.5 Complex rhematic groups 1.5.1 Existential expressions 1.5.2 Presentative function 1.5.3 Other thetic expressions

17 17 20 20 27 28 29 30 31 48 63 65 69 70 72 74 77 77 85 88



2 The attributive relationship 2.1 Nucleus marking and nucleus groups 2.1.1 Pronominal nuclei 2.1.2 Substantival nuclei 2.1.3 The infinitive as nucleus 2.1.4 The adjective as nucleus 2.1.5 Adverbial nuclei 2.2 Attributive groups 2.2.1 Substantival attribute group 2.2.2 Pronominal attribute group 2.2.3 Adverbial attribute group 2.2.4 Predicative attribute group 2.3 Attributive complexes 2.3.1 Substantival complexes 2.3.2 Adjectival complexes 2.3.3 Pronominal complexes 2.3.4 Adverbial complexes

91 92 93 97 98 100 102 105 105 112 116 118 123 123 131 138 138

3. Completive relationship 3.1 Object complements 3.1.1 Direct object paradigms The substantive group as direct object Pronouns as objects Objects preceding the verbal form Object nexus: external syntax Clausal objects: external syntax 3.1.2 Indirect/prepositional object syntagms 3.1.3 Representative valency patterns: parameters Giving Ability, will, obligation and permission Fear Asking vs. not-knowing Knowing and thinking Seeing Saying 3.1.4 Causative and passive 3.1.5 Reciprocal and reflexive 3.2 Adverbial complements 3.2.1 Circumstantial adverbials 3.2.2 Qualifying adverbials 3.2.3 Adverbials of manner Purpose clauses Temporal syntagms Comparative adverbials

141 141 141 142 147 150 153 154 159 163 163 165 171 172 176 177 179 179 182 185 186 191 193 193 195 199

TABLE OF CONTENTS Causal expressions Concessive expressions Other adverbial types Presupposed degree or amount Spatial adverbials Without Result adverbials


200 204 207 207 209 210 212

4 Apposition 4.1 Adjectival syntagms 4.1.1 Morphological adjectives 4.1.2 Adnominal prepositional syntagms 4.1.3 Adjective syntagms and clauses 4.1.4 mare syntagms 4.1.5 Non-clausal adjectival nexus 4.2 Appositive content clauses

213 214 214 216 219 224 225 226

Part II: Macrosyntax Chaining

229 232

5 Narrative syntax 5.0 Narrative events 5.1 Narrative FSP effects of focus and topic 5.1.1 Focus particles 5.1.2 Topic 5.1.3 Rhematization 5.1.4 Peak marking signals 5.1.5 ‘Subordinate’ events 5.2 Point of view 5.2.1 Character’s point of view 5.2.2 Narrator’s point of view 5.2.3 Direct thought representation 5.3 Cohesion 5.3.1 Episode markers 5.3.2 basr hdax 5.3.3 qmle 5.4 Grounding 5.4.1 šqlwle, qam šqlwle and šqla wle 5.4.2 Presentative constructions 5.4.3 Textually marked circumstantials 5.4.4 Negative preterites: la šqlle, la qam šqlle 5.4.5 Setting 5.4.6 Descriptions 5.4.7 Passive forms

237 237 239 239 242 249 251 257 258 259 270 282 284 284 287 290 293 293 297 300 310 314 317 324



5.5 Aspectual oppositions 5.5.1 The gerund bšqla vs. simple forms 5.5.2 pšle bšqla 5.6. Tense oppositions in narrative 5.6.1 šql for šqlle 5.6.2 kšql in narrative 5.6.3 pšql and pšqlwa in narrative 5.6.4 Tense in narrative: subordinate clauses 5.7 Concluding remarks 6 Dialogue syntax 6.1 Functional Sentence Perspective 6.1.1 Topic 6.1.2 Questions and answers 6.1.3 Focus Syntactic patterns Subject focus: the post-verbal position Cleft pattern Object focus: the pre-verbal position Object focus: nominative pronouns for object Particle-marked focus Simple inclusive focus Scalar inclusive Specifying focus Exclusive focus: Contrastive focus Conclusive remarks 6.2 The modal system 6.2.1 Deontic modality šud Particles 6.2.2 Epistemic Modality Epistemic particles and expressions Verbal forms denoting epistemic modality Conditional Conditional forms and values Conditional types Paratactic (concessive-)conditional pattern Counter-factual conditional patterns 6.3 The tense-aspect system 6.3.1 šql-w-le, qam šql-w-le and šqla w(w)le 6.3.2 šql-le and qam šql-le 6.3.3 k-šql-wa and b-šqla w(w)le 6.3.4 wal šql-le (or qam šql-le) and wle~wn~k-we šqla

330 330 333 334 335 337 343 346 350 353 354 355 360 365 365 365 369 374 374 375 375 377 378 380 381 383 384 385 388 391 393 393 397 399 400 404 405 408 413 414 419 422 427

TABLE OF CONTENTS 6.3.5 k-šql: generic/aoristic present 6.3.6 wle b-šqla and wal k-šql: actual or specific present 6.3.7 p-šql~lak-šql form 6.3.8 Conclusions 6.4 Space 6.4.1 Vocative 6.4.2 Concrete presentatives 6.4.3 Anchored deixis


431 435 438 440 443 443 445 446

7 Appendix: the morphology of JZ 7.1 Pronouns 7.1.1 Personal pronouns 7.1.2 Demonstrative pronouns 7.1.3 Interrogative pronouns 7.1.4 Other pronouns 7.1.5 Determiners 7.2 The noun 7.3 The predicative forms 7.3.1 Copulas 7.3.2 Verbs Verbal inflection: the sound verb Weak verbs

449 449 450 452 452 453 454 454 455 455 456 457 461



PREFACE This monograph is a complex, systematically organized collection of innumerable details and facts, whose infinitude is well known to anyone who has had the experience of trying to achieve such description; the bulk of material is partially manifest in the physical and digital materials accumulated during the years of the project. The complexity of production has been in the hands of the author, but actually many people have helped, both directly and indirectly: H.J. Polotsky, who collected the texts and investigated them. By leaving us his Jewish Zakho notes, as well as by teaching the language to many students over many years, he left us a significant heritage; Gideon Goldenberg, who taught me this language (inter alia), trusted me with these texts, which made this project possible, and answered an infinite number of questions, thus helping to resolve many qualms; Yona Sabar, the most prominent figure in the study of this dialect, who willingly sat with me throughout the Polotsky Zakho text collection and helped clarify many incomprehensible expressions; Olga Kapeliuk, for some helpful pieces of advise; My advanced students Michal (Schwartzbart) Marmorstein and Ya’ar Hever, who participated in the project and offered their invaluable professional opinion many times throughout the writing process; Noa Tal, who participated in the first stages of the project, and was the one in charge of the digitization process of the Polotsky material as well as its initial translation; The Israel Science Foundation (grant 1074/05), who supported this research and actually made it possible;



My colleagues from the department of linguistics: Ariel Shisha Halevy, Moshe Taube and Lea Sawicki, for their willingness to discuss many issues over the years; Eitan Grossman, with whom I conducted an countless linguistic conversations, who also read through the manuscript, pointed out many difficulties and suggested many valuable solutions and corrections (pertaining both to the language used and various linguistic issues); My two editors, Geoffrey Khan and Hezy Mutzafi, who accepted the book to the series and, drawing from their experience, offered their corrections, suggestions and thoughts, for which I am thankful. In addition, Geoffrey Khan has been, over the years, a constant and unique source of assistance, especially with regard to various syntactic issues. My many Zakho students (five two-year cycles) have asked a plethora of questions, which no doubt have made me stand on my toes and indirectly have made this description a better one. And, last but not least, my family, who are my constant source of energy, inspiration and raison d’être.

0 INTRODUCTION The Jewish Neo-Aramaic dialect of Zakho is considered a well-known dialect, possibly for its long service—it is one of the longest-taught NeoAramaic dialects (along with the literary dialect of Urmi), or perhaps because of the breadth of the extant material (see below, corpus definition). The impression of the actual profound linguistic knowledge about this dialect is not commensurate with the actual amount of linguistic publications devoted to it: the amount of knowledge is remarkable, but very little of it is actually published; it has largely remained an oral tradition in Jerusalem. This was the impetus for the project and the present monograph (as well as several papers which have come to accompany it) constitutes an attempt to remedy this situation.

0.1 HISTORY OF RESEARCH H.J. Polotsky was the pioneer in working with data from the Jewish dialect of Zakho (henceforth JZ), which he collected firsthand. The existence of these data constitutes the incentive for this description. Polotsky’s attention was centered, judging from his few but thorough publications, on the Christian dialect of Urmi, whereas the grammar of Zakho was given but occasional, brief comments in his publications. Thus, the linguistic information pertaining to this language has remained by and large within the confines of oral tradition; the language has no published grammar. However, Polotsky had compiled a dossier of grammatical facts backed by examples, pertaining to phonology, morphology and possibly syntax. This dossier has remained unpublished, but it had been, for a long time, accessible to anyone interested, as it was on the shelf in the library of the Hebrew University’s linguistics department until it disappeared. There are two more dossiers—one is mostly type-written, organized as a grammar, perhaps meant to be a replacement for the hand-written dossier in the library; and the other is a folio-sized notebook, dated January 1945. Both files contain various lists of a



morphological and syntactic nature. The syntax part, which is more relevant to us, is in both cases a collection of examples with hardly any comments. Reconstructing Polotsky’s thoughts from such bare examples, without having personally studied with him, is quite impossible. Whenever there are some clues to Polotsky’s train of thought, I either accept them or argue with them, always mentioning him. In addition, some terminology is adopted off his notes (for instance, the term “presentative”). Note, however, that the concept of the syntactic structure of JZ presented in this monograph is my own, and is similar in more than one point to my syntactic description of Literary Old Babylonian (Izre’el and Cohen 2004:62–114), although it is much more detailed and nuanced, being based on a far larger corpus of a modern language. Polotsky’s impact on JZ research was immense, but indirect; his only publication specifically related to JZ was Polotsky 1967:73–77 (JZ text) and 104–111 (JZ Glossary). His impact was achieved mainly through teaching, whereby his students (most notably Y. Sabar and G. Goldenberg) published JZ material and related research. No doubt, the material collected by him was, and still is, the hidden goldmine in the research of JZ. Following Polotsky’s work,1 JZ has been given due consideration in further published texts (see §0.3 below, corpus definition), some materials have to do with the lexicon (see below), a little has to do with phonology, which has been given particular consideration in Nakano 1969, Hoberman 1989:149–156, 1997:314–333 and Sabar 2002b:22–37. Zakho morphology is given a short sketch in Avinery 1988:12–15, and a special consideration in Sabar 2002b:38–54. The quadriliteral verb is treated in Sabar 1982. The field of lexicography in Zakho Neo-Aramaic has received much attention, and has been dealt with by many, most notably by Y. Sabar (1974, 1975), whose name has come to be intimately associated with the language of Zakho, as a key figure in the research of Neo-Aramaic in general and of Zakho in particular. Sabar’s dictionary (2002b) is now the standard authority on JZ and the Nerwa Homilies, it is based on Polotsky’s JZ vocabulary card collection. As for syntax, it is truly a vast territory; many works dealing with syntactic issues in Neo-Aramaic in general have bearing on Zakho as well, but articles or monographs entirely devoted to Zakho syntax are rare at best.

An enormous impact on Neo-Aramaic studies in general was achieved through his first publications related to Urmi Neo-Aramaic—Polotsky 1961 and 1962. 1



Nominal syntax is addressed in Tsereteli 1965, 1968a; Goldenberg 1993:296–298, 2000:78–81; Khan 2001 and 2008; verbal syntax in Tsereteli 1968b, 1972; Polotsky 1984–6, 1994, 1996; Hoberman 1989; Hopkins 1989, 2002; Goldenberg 1991, 1992:119–131, 2000; Kapeliuk 1998, 2006, 2008; Pennacchietti 1996; and finally, Khan 2006, 2007, 2008c and more. Valuable syntactic information is also found in extensive footnotes in various text editions, e.g., Meehan and Alon 1979. Noteworthy in this context is Goldenberg 1988, which describes several features of the homiletic language of the early Neo-Aramaic of Nerwa (Sabar 1984), which can be considered the closest we have to an earlier diachronic stage of Zakho. Another research domain (relevant to all levels of linguistic analysis) is the one dealing with influence exerted by neighboring languages, e.g., Kapeliuk 1996, 2002:315– 318 and Pennacchietti 1988. The syntactic discourse in the field of Neo-Aramaic has been greatly enhanced along the last decade; this is manifest primarily in G. Khan’s publications: 2001 (indefiniteness), 2008a (definiteness), 2008b (monumental treatment of the Barwar dialect and indeed one of the richest syntactic descriptions in Semitic linguistics), 2009a (discourse structure), 2009b (genitive), 2010 (deontic modality). These publications often show a wide comparative perspective. My own published work pertains to 1. information structure: 2008b (focus), 2008c (syntactic focus marking) and 2008d (copular clauses), and 2. the nominal group: 2010 (nucleus and attribute) and one which is forthcoming (the determination system). Work in progress investigates circumstantial clause combining. Three papers were written by students of mine: Gutman 2008 (bare preterites), Hever 2006 (discussing the difference between the two preterite formations) and Schwartzbart 2008 (the exposition unit of folktales).

0.2 METHOD The linguistic method followed throughout this monograph is structural, which is merely a way to look at things rather than imposing an analysis on the data (see below for more detail). This description is more or less consistent with the ideas behind non-aprioristic syntactic theory (Frajzyngier 2010), or framework-free grammatical theory (Haspelmath 2010). The method followed in this study is empirical, based on a defined corpus. The results of the analysis are system-oriented rather than featureoriented. The description refers to exponents, that is, formal features, whether morphological, syntactic or macrosyntactic, which have a con-



sistent relationship to certain values or functions. The term ‘empirical’ means, in addition, that every statement formulated in this monograph is not impressionistic, but rather rests on synthesis of many details of data, that is, it is retraceable and hence it is refutable. The system of interrelationships between linguistic signs is not directly accessible, and hence analysis starts with the parole, the concrete manifestation of language, the primary material at our disposal, represented by the text. The linguist’s task is to consider anything found in the text as long as it is lingustically pertinent, viz., deemed as having a function in the langue (the abstract language-system shared by a linguistic community). Such linguistic pertinence exists with regard to an entity as long as it is found in opposition with at least one other entity. Since the parole is not a system but only a local actualization thereof, an analysis should be performed using the syntagmatic and paradigmatic dimensions, in order to expose the system behind the parole. The syntagmatic dimension is the one present in the text – the linear, or sequential, relationships between the entities, phrased in terms of (in)compatibility, order, etc. The paradigmatic dimension is the relationship between all the entities that may figure in the same syntactic location in absentia, without changing the syntactic structure of the segment. This kind of relationship is categorial—all linguistic entities, of whatever order or size, which may be found in a given syntactic slot, belong in and constitute part of the same functional category, or the same paradigm.2 Thus it is possible to determine and define at each stage of the analysis the exact interrelationships existing between different paradigms, as well as the relationships between the entities within a paradigm. Moreover, the commutation (interchangeability) test teaches us important details about the syntagmatic dimension of entities—for example, their boundaries, i.e., where entities begin or end. The value of entities is defined as the meaning, or function of the entity. This function is definable only in opposition to the other members of the paradigm and in comparison between paradigms. Note that each paradigm is by definition environment-sensitive; for instance, two externally identical forms may actually belong to different paradigms and have distinct functions. These tools do not foretell the findings; they merely supply the linguist with a relatively objective way of approaching the material, observing it,

To be sharply distinguished from the traditional sense of ‘paradigm,’ in morphological word-lists. See glossary, §0.7. 2



and uncovering the extant entities and the interrelationships found among them. The description does not follow any design, other than that which is presented, or dictated, by the language itself. It strives to formulate JZ syntax, so as to provide its mode d’emploi. Much attention is paid to syntactic patterns, often termed constructions: a syntactic pattern is recognized whenever a syntactic structure consistently signals a function.



Unlike many modern treatments, based upon modern collection of materials from old native speakers, the language of the Zakho corpus at our disposal is in general the fruit of old-style fieldwork, often without recordings. For this reason it sometimes must be treated as a written language (for instance, because there is no more information about the informants who contributed the material, or when a long time has passed and there is no one to approach for further inquiries). This entire enterprise has sometimes earned the scorn of field linguists, as if the way the data was collected makes JZ less real or worthy of study. However, the depth of treatment, viz., analysis and description, does not entirely depend on the methodological purity of the way the material had been obtained, and depth may be attained even when data is not perfect. Incidentally, the material, coming from speakers who were much closer temporally to their natural speech conditions in the original habitat, is considerably less influenced by linguistic interference than the material which is collectible nowadays. The corpus consists of the material published up to this point and of the Polotsky Zakho texts. When available, text sizes (assessed by the number of words) is given in parentheses: Polotsky 1967:73–77 (text, no edition) and 104–111 (Glossary) and Zaken 1997 are parts of the entire Polotsky text material. They are not given any special reference. Meehan and Alon 1979 is an excellent edition of a story with a detailed linguistic commentary (abbreviated MA) (2700 words)



Avinery 1988 is a monograph containing, inter alia3 1. a few stories and story-like texts in modern Zakho Neo-Aramaic (abbreviated A+number. line); 2. most of the book is devoted to a translation of a selection from the two thousand sentences in Frei 1966. These translations have no context, yet they expose an often unfamiliar, colloquial side of this language (abbreviated AS+number). (3860+13000 words) Sabar 1976b, Lēl Hūza, a text traditionally recounted on the Ninth of Ab, reflects somewhat older language (abbreviated LH). (1430 words) Sabar 2002b is a reconsideration and a detailed edition of a text recorded in 1870 (Socin 1882:159–166), both texts have a detailed sociolinguistic as well as linguistic commentary (abbreviated SS). (1300 words) Fassberg 2003 and 2004 are different, having been elicited not too long ago (2000 and 2003 respectively) from a woman who is not, strictly speaking, a native speaker of Zakho Neo-Aramaic, since she was born in Israel and learned the language from her grandparents with whom she had lived as a child. These texts belong to the genre termed “personal experience narrative”, which is considerably different from the more usual, folktale genre. Two more extensive texts belong to this genre (acronym PEN), Sabar 2005 and Sabar 2007 (abbreviated SYG and SAG respectively). These texts have not been given separate consideration, although as a different genre they probably deserve their own description. (7000+5700 words) The main corpus upon which this description is based consists of texts recorded in writing by H. J. Polotsky between the end of 1944 and mid1947. Polotsky had prepared for this material collection; he systematically learned a great deal from his informants before he started writing down stories: the basic vocabulary, the basic grammatical forms, etc. By the time he started recording the stories, he was quite well versed in the language. There is an impressive indirect evidence for this: throughout the material, nine notebooks, there is an astonishingly small number of corrections or

It also contains translations of Biblical texts (Book of Jonah and the haftarah for the Ninth of Ab, Jeremiah 8:13–9:23); the hymn ʿeṯ šaʿăre rāṣōn and a glossary at the end of the book. These materials are not part of the corpus: the Bible translations are explained below, and hymns are analogous to other poetry types. 3



erasures. The material was collected from three informants about whom we know next to nothing, and it comprises 55,000 words, about 150 pages.

an exemplary page from notebook 7



It has more than double the extant amount of text from other sources (to the exclusion of Avinery’s sentences). Originally the text was put down in Polotsky’s large handwriting and extends over almost 968 of these pages (abbreviated PT, but being the main source for this description, just the original page number is enough). The material consists exclusively of stories (folktales, novellas and a farce) and the informants seem to be very good storytellers. This material was examined and the result, combined with that of the other texts, is this monograph. These texts are planned to be published shortly after this syntactic description. The Polotsky Zakho texts were related by three informants—Zion Levy, Simha Levy and Avraham Levy, I am not sure whether they were related, as there is very little available information about them. The last two were young people, around twenty years old. There are minor idiolectal differences between them, which are mentioned occasionally. Other genres, such as epic poetry (e.g., Rivlin 1959) or Bible translations (Sabar 1983, 1988, 1990, 1993, 1994; Goldenberg and Zaken 1990) are not considered in this description. The poetry is evidently different syntactically from the rest of the material, whereas the translations are very literal (as is explained in Sabar 2000) and their syntax is heavily influenced by accommodation to the Hebrew source; therefore, these texts have little value for syntactic analysis.

0.4 STRUCTURE The monograph consists of two parts—the first describes microsyntax, namely, the basic syntactic relationships at sentence-level: the predicative relationship, the attributive relationship and the completive relationship. A small chapter on apposition, not considered a relationship in itself, concludes this part. The second part is devoted to macrosyntax—relationships and syntactic entities beyond sentence-level. The chapters describe the narrative and the dialogue. The last chapter, which is really an appendix, succinctly outlines the morphology for anyone without a prior knowledge of Neo-Aramaic.















0.5 GENERAL PRACTICES The transcription used is a compromise: it is not quite the original text itself, which is written in an extremely narrow transcription, nor is it a purely phonemic transcription. As a phonemic analysis has not been carried out, I use the prevalent transcription system for JZ, as used, e.g., by Yona Sabar. This means that many differences, kept in the original super-narrow transcription, are neutralized in this transcription. There is a certain inconsistency, since no unification took place; consequently, the way a unit is written could be different across the text, basically keeping the original (e.g., ʾaw vs. aw) to avoid an arbitrary decision. Additionally, in this transcription, the accent was kept whenever 1. it was originally placed in a different place than the usual penultima; 2. when it is not clear why it is there (e.g., on the determiners: is there a difference perhaps between aw vs. áw?). The verbal forms in this description are generally not named after their putative most common function, for the simple reason that such habit may produce an occasional clash: for instance, the form p-šāqıl-wa is usually found in counter-factual apodoses ‘he would have taken, il aurait pris’ (§ in the dialogue. On the other hand, in narrative, it has two functions (§5.6.3): one is an intentive circumstantial (‘intending to X’), where a finite verbal form interchanges with the gerund or the perfect participle. The second is where p-šāqıl-wa occurs instead of k-šāqıl-wa as an imperfective ‘he



was taking, il prenait’. The practice then is to use the form itself, šqıl-wā-le, k-šāqıl, wēle bı-šqāla, etc. This practice is adopted from Damourette and Pichon 1911–52, who use throughout their monumental book the various forms of savoir ‘know’ in its various 2CPL forms—for instance, le sachiez (=present subjunctive), le sauriez (=conditionnel), le sûtes (=passé simple), etc.




= Avinery 1988:48–75 (texts)

A = Avraham Levy


= Avinery 1988, 76–209 (sentences)

S = Simha Levy


= Sabar 1976b

Z = Zion Levy


= Meehan and Alon 1979




H. J. Polotsky, unpublished Zakho Texts, Jerusalem 1944–1947 H. J. Polotsky, unpublished notes on Zakho grammar

SAG = Sabar 2007 SYG = Sabar 2005 SS

= Sabar 2002a



= copula

= determiner = pronoun = quantifier

= rheme = substantive = adjective

= adverb = indirect object = direct object


1 2 3

= masculine singular = feminine singular = common singular = common plural = nucleus = attribute = 1st person

= 2nd person = 3rd person



0.7 A SHORT GLOSSARY OF TERMS The following table provides a short definition of the most important terms used throughout the book: term




A syntactic conditioned variant: a different actualization of a syntactic entity in changing environments.

the syntagm x–īle (predicate+COP) is realized as īle–x (COP+predicate) only in subordination.


An expansion of a nucleus (which may be nominal or adverbial), which may consist of a (pro)noun, an adverbial or even a clause.

xabra ‘word’: xabr-ox ‘your word’; xabır xōr-e ‘word(s) of his friend’; xabır mxēla ‘word(s) she said’.


An exponent specialized to signal the nexus (or predicative link) between non-verbal entities. The ancient non-verbal clause is obsolete.




ʾāhet you

brōni wētēn my-son are2MS

(‘It is I’=)

A grammatical exponent mar- ʾēma nāša king the obligatory actualizču nāša ation framework of the noun; flāna nāša may express degrees of defi- ʾaw nāša niteness, specificity, etc.

rheme copula ʾāna wın I am1MS

‘which ‘no ‘a certain ‘the

man’ man’ man’ man’

A concrete, consistent signal daw nāša (the exponent to for some grammatical function. mark the nominal attribute)


The domain accounting for the informational status of linguistic units, for instance, whether they are new, presupposed, contrasted, etc.

āna bıdamxen ūʾāhıt nṭōr ‘I will sleep whereas you guard’ (contrastive topic);


An entity marked for a special informational prominence. It is marked this way for reasons of contrast, exclusivity, restriction, etc.

mqabıl mınne ši ‘just before him; xa mınnu ši la gēwızwa ‘even one of them he would not do’; qımla ʾápāya bxēla ʾımmu ‘so she too wept with them’.

ppēšan ʾāna baxtox ‘Let me be your wife’ (focus).




The relationship between the main components of the clause, theme and rheme (which is also termed predicative link).

The term nexus stands, in addition, for the entire combination of theme and rheme.

rheme p-šaql FUT-take

nexus theme ın ͜ 1MS

{I ͜ will take}

theme nex. rheme kxāzē le ͜ bīzāla they-see him going They see {him ͜ going}


The ‘grammatical center’ of a construction, which determines the syntactic status of the entire construction.

baxtıt mıtla ‘(the) wife (who) died’; baxtıt tre ‘(the) second wife’; ayFS d tre ‘the second (lit. the-that-of-two)’.


A group of forms which may šqıl-le xa (indef.) nāša interchange with one another aw (def.) at a given slot. The members of he took a/the man such group genrally have a function in common. (the object DET paradigm)


A complex form, often realized by an abstraction often consisting of several variants, which has a consistent function.

#šāqıl-wa—p-šāqıl-wa# (SUBJ+-wa—FUT+-wa)= counter-factual conditional (#had he …ed, he would have …ed#).


Past perfect, perfect in the past. šqīla wēle ‘he had taken’.


Past of past, without the no- šqıl-wā-le, qam šāqıl-wā-le tions of perfect. ‘he took (a long time ago)’.


The entity holding the new information in a clausal setting; the rheme is often found inside the verbal form. Otherwise it is marked by the copula or by other exponents. For instance, the (minimal) answer is always the rheme.

1. bēsa dīdi–le ‘the house is mine’. 2. mani mírrēlox ʿazmıtti? mırre ʾō gōra: baxti mírrāli ʿazmınnox ‘Who told you to invite me? My wife told me to invite you’.


A form with two main fun- xá yāʾe ‘someone (who) ctions: mark subordination and knows’; (šud) yāʾe ‘let him modality. know’.


Any combination of two or more entities, an ordered sequence of entities.

aw nāša ‘the man’; xızyāle baxta ‘he saw the woman’.




The presupposed entity, about which new information is pre- ṭalb–ın dicated within clause level. ṭḷıb–li


Above clause-level, the topic is the thematic entity, the entity about which one predicates new information (termed ‘comment’). It often serves as a discourse anchor, containing presupposed or backgrounded information, whose function is to maintain cohesion between parts of the text.

theme rheme

(that)skI ask I asked

ıswa xa gōra ... aw gōra… ‘There was a man... the man...’;

ūʾāhet mayla qaḥra dīdox? ‘As for you, what is your grief?’

PART I: MICROSYNTAX: INTRACLAUSAL SYNTACTIC RELATIONSHIPS The number of basic syntactic relationships is a debated issue. Nevertheless, it seems that three basic syntactic relationships are tenable for JZ, which correspond, by and large, to the same relationships in the ancient Semitic languages, where they are symbolized by the three-case system. This scheme is presented in Becker (1841 §60) and discussed, with reference to the Semitic languages, by Goldenberg (1987=1998:138–47), where the general linguistic literature treating these relationships is presented. It seems that, despite the lack of case marking, the syntactic behavior of JZ is similar, in several respects, to the ancient Semitic case languages, such as Arabic and Akkadian. These basic relationships are 1. the predicative relationship (=nexus), which obtains between any theme and rheme (for which see §1 below); 2. the attributive relationship, which is realized syntactically by any genitive construction; and 3. the completive relationship, found between the nexus and the object or adverb. The following table summarizes the basic information regarding these relationships: syntactic relationship

morphological realization

syntactic realization

predicative attributive

verbal forms adjectives

copulas and their rheme


pronominal object suffixes

independent object or an adverb

construct state and its attribute

Each line in the table represents a relationship. The idea of these basic relationships is the classificatory principle used throughout the description of the JZ microsyntactic system. Consequently, each relationship has a chapter in the first part of this syntactic description. It must be stressed that apposition is not in itself a relationship, but rather a repetition of a syntagm, and



occasionally, of the relationship itself, in the same syntactic conditions (Becker 1841 §60). Apposition has an entire chapter as well. These relationships, taken together, constitute microsyntax, and this entire domain will be described with reference to them.

1 THE PREDICATIVE RELATIONSHIP Theme and rheme are the basic functions of a clause: the theme is the given, or known entity in the clause, whereas the rheme is the new piece of information that is predicated on the theme. Note that these terms, which concern information structure, have little to do with morphological marking and can be encoded in various morphological, syntactic, and other ways. The relationship obtaining between the theme and the rheme is the predicative link, or the nexus. In JZ an independent nexus, or a clause, is a syntagm that has to have an exponent of the nexus. The most common exponents of nexus are the verbal forms and the copulas. Besides them, one encounters various types of dependent nexus (such as object nexus and adjunct nexus)

1.1 VERBAL FORMS The morphological constitution of verbal forms is displayed in chapter 7. Their internal syntactic constitution is discussed in this section, whereas the individual values of these forms is discussed under macrosyntax (chapters 5 and 6). Verbal forms are described in Goldenberg 1985 as a basically “built-in clause”. It is shown ibid. to be made up of 1. a subject index; 2. a verbal lexeme; and 3. the nexus between them. The subject index is generally the theme, the verbal lexeme is the rheme. The individual values of each verbal form are determined by form in the wide sense, that is, by the combination of morphological class and syntactic environment. The following table shows this analysis on all extant verbal bases:





1. imperative


2. subjunctive šaql-

3. active preterite


4. passive preterite


5. pres. perfect (as well as passive state)






ø (2CS)



2nd (SG.

-un (2CP)




-ın (1MS)


(that) I ask

-ıt (2MS)


(that) you ask

-a (3FS)


(that) she ask

-li (1MS)


I asked

-lox (2MS)


youMSG asked

-la (3FS)


she asked

ø/-ın (3MS)


he was killed

-a (3FS)


she was killed

-i (3CS) (copula) nexus - theme w—ın (1MS)


they were killed

wın qṭīla

I have killed

w—at (2FS)

wat qṭılta

you have killed

w—ax (1CP)

wax qṭīle

we have killed






The analysis into the base, representing the verbal lexeme (which is the rheme), and the subject index affix (which represents the theme) is rather simple. Note that with the perfect tense, the copula, which represents the nexus, is an explicit exponent of the nexus. It is important to note, however, that the subject indices for each base are different (for instance, 1SG may be -ın or -li). The presentation of the system, mainly because of the special diachronic origins of the NA preterite, was discussed by Polotsky, as well as by others after him—Goldenberg (1992), Hopkins (1989) and Khan (passim). Polotsky examines the following representative pair from the Urmi literary dialect: nşiq


‘she kissed me’



‘that I kiss her’

He arrives at the conclusion that “since the two sets of pronominal morphemes are the same and occur in the same order, their inherent meaning and their syntactic relation to the participial base must be the same in both tenses ... the contrast between the two participial bases must concern their respective ‘voice’: pti(j)x- must be passive. In nşijq-in-nə we have a passive base,



a nominatival undergoer (-in), and a datival actor (-lə)...” (Polotsky 1996:16, the emphasis is mine).

Polotsky does not say that the form nşijq-in-nə is passive; he merely refers to the base nşijq-. The same could be said about the perfect or passive participle in Romance languages—je les ai sorties implies that the participle is passive, since it agrees with the object, but the verbal expression does not signal any passivity but rather past tense. A different view may lead to a different presentation. We may compare similar forms to those that Polotsky compares, since in JZ the inventory is not identical to Urmi’s. We could compare the following forms: subjunctive

našq-ī-lax ‘let them kiss youFS’


nšīq-ī-lax ‘youFS kissed them’

From a synchronic point of view, these forms seem to be analogous to the ones compared by Polotsky. However, when taking all of the information into consideration, the results obtained are somewhat different. The following table shows the syntagmatic (that is, linear) conditions and paradigmatic constitution (i.e., the forms which figure in posse in each slot). base

verbal prefix qam-, -ø





base forms šāqıl(-)(_C) šaql- (_V) šqıl- (_C) šqīl- (_V)

1st set prons.

2nd set prons.

3 pers.

2 pers.

1 pers.















-lox -le

pl. -lan -lṓxun -lu

1. The syntagmatic conditions are not the same for both bases—that is, the verbal prefix slot occurs only with the subjunctive base; hence, according to the linguistic principles used here, the paradigms are different (even though they may look the same). 2. The paradigms of the first set pronouns are also explicitly different: the preterite base occurs only with 3rd pers. forms, while the subjunctive base occurs with all persons of this set. Such difference in paradigmatic constitution is enough to account for a difference in function. 3. The occurrence of both pronominal sets is different for each base: with the subjunctive base the first set is always present; with the preterite base it



is often absent (e.g., when the object is indefinite). On the other hand, the second set (l- pronouns) is facultative with the subjunctive base,1 whereas it is obligatory with the preterite base. What is obligatory is here viewed as subject index, representing the external theme. Another form is the relatively rare passive preterite (see e.g., Polotsky 1996:17–18): base

verbal prefix

base forms


1st set prons. 3rd

2nd set prons.










(passive) none


-a -i

In this case, the 1st set (only 3rd pers.) is obligatory, functioning as subject index. This form has different syntagmatics, and there is no need to view it as a preterite without the 2nd set, but rather as a distinct form. This form incidentally strengthens the case made by Polotsky—it occurs without an expression of the agent and we tend to view it as either passive or as impersonal, which corroborates Polotsky’s concept of the passive base.

1.2 THEMATIC NOMINAL GROUPS Nominal groups occur in many syntactic functions, the theme being merely one of them. Nevertheless, the nominal group will be described here in full, and below, in the context of other functions, only the relevant facts peculiar to that syntactic function are described.

1.2.1 Nouns and determination The nominal group includes any entities co-occurring with the substantive, that is, determiners, quantifiers (henceforth DET and QUANT respectively) or entities which occur instead of it, representing it pronominally, that is, pronouns (PRON) as well as other entities which exhibit pronominal behavior (as quantifiers often do), and finally, substantival clauses. In addition, the adjective (for which see §4.1) is also a part of this group. The basic structure of


Except in the formation qam-šāqıl-le, where it is obligatory as object.



the nominal group is DET+(QUANT+)SUBST, each representing an entire paradigm, a substitution group. The DET group2 occurs obligatorily with members of the nominal groups — SUBST, pronominals and ADJ (but not with real pronouns): determiner type





±def., generic

flāna ‘a certain’


ʾēma ‘which’ ču ‘no’

aw ō(ha)


dir. and indir. interrogative





ʾyā(ha) xa ‘a’

xakma ‘some’ xapča ‘some’

negative definite distal indef. indef.


What makes this group a paradigm is that one of them always heads a noun group. It is a syntactic function rather than a semantic one; semantically these determiners have a variety of functions—they may be non-specific (flāna ‘a certain’), negative (ču ‘no’), demonstrative (ʾwāha), etc. What they have in common is that they all identify the noun phrase as such even if this is not transparent: ʾan ksēmi mınnan ‘th(os)e (who) hate us’ SAG 3.17


In ex. (1), the construction consist of a definite determiner (no definite pronoun ever occurs in this function with a verbal form) and an asyndetic adjective clause consisting of a finite verbal form (ksēmi mınnan ‘they hate us’). What tells us that this verbal form functions as noun is the nominal envelope provided by the determiners, in this case an (plural definite DET). The functional importance of the determiner is further illustrated by a comparison with the following example: anya žġıllu ‘These were busy’ 808



For a wider linguistic context see Cohen forthcoming.



anya ‘these’ is often thought to be of the same group. It does occasionally occur with nouns; however, when it precedes a verbal form, it is clearly an independent pronoun, ‘these’, which functions as the theme. In fact, all the entities in the table function as determiners; but the ones shaded grey can only be determiners, they do not occur alone, e.g., cannot represent the theme. They do, however, mark the category of the entire phrase, which makes them the nucleus of the noun phrase. The thematic noun phrase, the given information, is generally both familiar and specific. Not at first, however, when an entity is being presented for the first time: (3)

ıswa xa gōra ... aw gōra… ‘There was a man... the man...’ 663


sēle xa gōra kıslōhun ... uʾōha gōra pıšle kıslu ‘A man came to them ... and the/this man stayed with them...’ 224

In both examples xa gōra is the first, introductory occurrence; the second time it is marked with a DET (aw and ʾōha), where gōra is henceforth both referential and retrievable. Thereafter, there is a fluctuation in tracking function between the definite and ø determiner: (5)

sēle aw gōra ıl bēsa ‘The man came home’ 240


sēle ø gōra ʾıl bēsa ‘The man came home’ 210

ø determiner, then, shows here definiteness. Note that in some slots it may signal indefiniteness (e.g., with plural nouns). It is important to note that the indefinite noun in ex. (4) is not the theme of the clause; in fact, the entire clause consists of new information. Such expressions are termed thetic expressions (see §1.5 below). That is why exx. (3) and (4) are comparable, since both introduce new referents. The negative determiner plays a similar function, but does not introduce a new referent: (7)

ču nāša mın mıjlıs lá fhımle mā qıṣṭa wēla ‘No man from the council understood what the story was (about)’ 476


la sēle axxa ču gōra ‘No man came here’ 263

The subtler functional difference between the various definite determiners needs further examination. Several issues are known: First, ø determiner is a pure determiner: its function is only anaphoric, without any demonstrative force, whereas the other definite determiners occasionally do have this force:



ø baxta lıbba +rāba pıšle rıẓya mın ʾīsāya dīda ldḗ dūka ‘The woman was very satisfied that she came to that place’ 862


(10) sēle ø ʾarya uqrūle ‘The lion came and approached’ 665

The cases in exx. (6) and (10), where ø occurs with a definite substantive, are rather common. Ø occurs with indefinite substantives but rarely: (11) ø baxta mōnıxla bbaxta ‘A woman looked at (another) woman’


This is a peculiar case. We know who the women are, and this expression is analogous to a reciprocal expression. The following example is arguable, as the lion could be definite: (12) xá bēná šmeʾle mı raḥūqa xa hırímhırímta dīd ʾ arya ūbasır xakma

daqīqe muxwēle ø ʾarya mın raḥūqa ‘Suddenly he heard from a distance a lion’s roar, and after a few minutes a/the lion appeared from afar’ 664

In addition, generic nouns are marked by this ø determiner: (13) ø ʾarya lág dāmıx lá-hōya rıš xızēna ‘A lion (any lion) does not

lie down, unless it is on top of a treasure’ 918

(14) ø baxtıt3 mpıqla mım bḗs gōra ūpıšla gō kōlāne ... lág daʾra lbēsa

bčú-mūj́ ıb ‘A woman (=any woman, all women) (who) left her husband’s house and remained in the streets ... will not return home under any circumstances’ 848

(15) ø baxta kmád hāye gōra sqīla ū xazya xa xḗt ápāwa gıbāle ‘A

woman, however beautiful her husband may be, if she sees another one, she wants him too’ 807

(16) bale yāʾıt lā kxāšex ø nāša manēlu pāre dīde ‘but you should

know it is not proper for a man to count his money’ 7

A further characteristic of the definite DETs is that they occur quite readily with substantives in the construct state:

In fact, in the remainder of this maxim, the woman occurs with a definite DET: …uʾáy baxta ḥaqqa qıṭla–le ‘and that woman (=meaning, the generic one we introduced before, such woman), deserves death’. 3



(17) bıd dáw kōlānıt lēʾal wēle qaṣrıt wazīrıt ḥakōma. bax wazīra skınta

wēla go ʿōlīya ʾılēsa. ḗ bax wazīra baxta +rāba spahin wēla ‘On that street was the castle of the king’s vizier. The vizier’s wife lived in the upper floor. The vizier’s wife was a very beautiful woman’ 481

The definite suffix:


may co-occur with a substantive followed by the genitive

(18) ḗha brāti waʿda dīda–īle ta gwāra ‘This daughter of mine, it is

time for her to marry’ 172

(19) mjōwıble ʾaw brōne rūwa ūmırre ‘This older son of his answered

saying’ 646

The indefinite


occur but rarely with construct state:

(20) xá mbínnōke sēle xá ṭērıt maḥkōye ūmırre ‘One morning there

came some speaking bird and said’ 278

However, a substantive with genitive suffixes is considered definite. This creates no problems with the the definite DET, but the indefinite DET with it would need an explanation: (21) xá yōmá basır yalunke dīde rwēlu qımle xá brōne ūmırre ta bābe

‘One day, after his children grew up, one of his sons stood up and said to his father’ 632

This xá is not an indefinite DET, but something else. Another problem associated with this case is that occasionally it is difficult for us to decide which xa it is, the indefinite DET or the number. Examining related cases seems to explain the problem: (22)









‘the two angels’


(23) (25)



trḗ —

malaxīne malaxīne

‘two angels’ ‘the angels’

362 364



These examples put together prove that there are in fact two slots preceding the substantive itself, as is explicit in ex. (24). The same happens below, especially ex. (29):4 (26)



‘(±def) soldier(s)’






‘(±def) ten soldiers’






‘the ten people’












askar gūre

‘(±def) two soldiers’ ‘some ten soldiers’ ‘the ten men’

113 133 (obj!) 131

In these examples we have no trouble determining which is the determiner and which is the quantifier. Exx. (24) and (29)–(31) exhibit both DET and QUANT groups. In fact, so do exx. (23), (27) and (28) except here the DET is ø. That is what we have in ex. (21)—The DET is ø and the QUANT is xa. Except cardinal numbers, other members of the QUANT group are the collective numerals kutr-+gen. pron. (‘both of them’) as well as kúṭḷāhūn (‘all three of them’), kud ʾ arbu (‘all four of them’ 864), kud šōʾu (‘all seven’ 521). kull- ‘all of’, although semantically very close, works different syntactically and is placed outside the substantive group—the syntagm kullu ʾan gūre ‘all these men’ (952) proves that kullu, being compatible with a determiner, cannot be one. kma ‘few’ (row 8a below) mostly does not occur independently of xa and as an indefinite, and hence xakma is shared between both DET and QUANT groups. The next table represents various combinations of the entire substantival group: Syntagmatic behavior of the three paradigms source

Eng. parallel


‘a woman’













‘one woman’





‘no woman’

2 4 6

xa ø ø

— —


baxta baxta



862 772

‘about one month’

‘a/the woman’ ‘two women’

Note that this syntagm is not thematic but rather objective; however, the principle works everywhere. 4

26 7



8a ø 9




‘about twenty rooms’




‘several/some men’



‘the(se) two women’






‘few hours’ (rare)

10 an




‘both (of the) women’

12 ay



‘the/this woman’

14 ø



‘(the/some) women’

11 ø


13 an



836 A 6-7

‘both women’

‘the(se) women’

The table illustrates many possible combinations, thereby explaining how xa could signify, with a substantive, both indefinite DET or the QUANT ‘one’. Of the quantifiers, the cardinal numbers may occur in partial pronominal function, representing the substantive itself (brought for comparison in line 1), while still occurring with a DET (rows 2 and 3): Pronominal behavior of the quantifiers DET


Eng. Parallel


‘the(se) three things’


‘the(se) three’


‘this(f) one/thing’







3 4




ay xa


xa kma


ču xa


ø mındi


xa mındi


ču mındi




ō, aw mındi

‘(some) three’




(some)thing, anything



360, 493

‘no one’ something the thing

passim 746, 750 307, 380

The pronominal complexes ayxa (or ēxa; masc. awxa), 5 xakma and čuxa (lines 4, 5 and 6 respectively) are in fact independent pronouns, but analyzing their structure reaveals that structurally they are some kind of amalgams

In other dialects this pronoun is used for anaphora in memory; see Khan 2008b:485–487. 5





and PRON (or QUANT). The pronominal mındi occurs with any singular

1.2.2 Pronouns True pronouns are considered only those entities which can represent the entire substantive group; the fact that some of these pronouns look like some DETs or QUANTs is of course not accidental, they are related diachronically. Synchronically, however, the DET xa, the QUANT xa and the pronoun xa are distinct entities syntactically, at times this is clear semantically as well. The following table lists the most common pronouns: Pronouns


Eng. parallel










‘both of them’






‘each one’





‘(the) three of them’



ō(ha), ē(ha) anya


‘this/that, these/those’

687, 808


āwa, āya, āni


‘he, she, they’


5 7 9

kull-+gen. pron. ʾwā(ha), ʾyā(ha) 6


10 xakma 11 ču(x)xa

collective distal

‘all of+PRON’ ‘that yonder’


passim 799, SYG 2.79


‘this (F) one/thing’



‘no one’


plural indefinite



A short survey of the pronoun xa can illustrate the differences from xa in the other groups: (32) xáʾ brāʾıš ʾıllan ‘Someone will detect us’ 656

The two sets of pronouns (rows 6 and 8) are in variation when functioning as theme. The differences between them are more prominent in other functions, e.g., when standing for the object, when rhematic, when focal and when topical. See below. 7 The 1st and 2nd personal pronouns are perfectly capable of functioning as theme, however, they are different in that they do not represent a substantive, but rather a ‘discourse role’, namely, the speaker and the addressee respectively. 6



(33) xā ́ lak tāfıq ʾ ıbbe ‘No one will meet him’ (lit. ‘one won’t meet

him’) 929

xa takes the place of the entire substantive group, when occurring with a negative clause, it is analogous to the pronounn čuxa ‘no one’. xa occurs as theme in reciprocal constructions in which the verb occurs in the singular:8 (34) xā mırre ta daw xēt ‘One said to the other’ 52

Another example is of gyān-, a usually reflexive pronoun, when used thematically: (35) mattaw lēʾal gyāni dīdox–īla ūmtuslamta wāna ṭālox ‘From now on

my very self is yours, and I am given to you’ 815

1.2.3 Clauses and infinitives Other entities that may occur in the thematic group are substantive clauses and infinitives. Both are by no means common in this function. The position of substantive clauses is restricted, occurring only after the rhematic entity, and only with special expressions: (36) bāš īla ksıbli ō qadda pāre ‘It is good (that) I earned so much

money’ 335

(37) ḗha lēwa brīsa ūlá gbarya ınnu āna ḥakōma wēna ūhatxa bāre

go gınsa dīdi ay kıma šınne ‘It is absolutely impossible that I am a king and such a thing would happen in my garden for several years’ 185

(38) ūkulla mınnox brēla dīd qam qaṭlınnu an raqıt māya ʾan

papūkin ‘and everything happened because of you, that I killed the poor frogs’ 248–249

(39) ḗha lá kısya (ı)lqbāla dīd jwanqa mın ġēr bāžer ā se ūpā yeš

ḥakōma dēni ‘This will not be accepted, that a youngster from another city would come and become our king’ 153

The substantive clauses occur asyndetically (ex. (36)), or with ınnu (ex. (37)) and dīd (exx. (38)–(39)). Infinitives in thematic function are very rare: It is not certain whether xa in cases with verbs in the plural should be regarded as thematic: uʾāwa uʾaxōne yımḗwālu xā ta daw xet ‘And he and his brother had sworn one to the other’ (276). 8



(40) bıš xlēla ʾılli ʾwāza xōrūsax mın ʾwāź xōrūś bnās mīre ūṗāšāye

‘Befriending you pleases me more than befriending emirs’ and pashas’ daughters’ 896

Note that the thematic infinitive takes an object (xōrūsax). (41) uʾō ʾī sāyox lá kīʾe ʾ ıbbe čuxxa ‘And your being here (lit.

coming), nobody knows about it...’ 798

(42) ūmaḥkōye dōhun la fhımli mınne čū xabra ‘and (as for) their

speech, I did not understand a word of it’ 422

The last example shows us that substantive clauses are not exactly like infinitives; the former does not take a DET, whereas the latter can. This is not selfevident, since adjective clauses generally occur with DETs.

1.2.4 Adjectives and adjective clauses Adjectives (simple or complex) are different from substantives in that they are complex entities that contain, in addition to the quality they signal, some kind of pronominal representation of the substantive they describe (for an in-detail description, see below, §4.1). Both morphological and syntactic adjectives mostly occur in JZ following a substantive. Their less common function, as an independent theme without a preceding substantive, is the main point here. Note that the adjective, whether morphological or syntactic, generally occurs with a DET. Unlike a substantive, however, it never occurs with DET ø: (43) ha ʾatta wal ʾurre xá nuxrāya ‘Just now a stranger has passed

by’ 811

This really is a thetic expression, where the free adjective is not really the theme, being indefinite. The next example has the full package: (44) ıswa trē axawāsa xā ʿāqıl uxā ́ šızāna. aw ʿāqıl gwīra wēle uʾaw

šızāna lá wēle gwīra ‘There were two brothers, one wise and one an idiot. The wise one was married and the idiot was not married’ 241

Note that the brothers are introduced together with their most prominent characteristic. This done, the text can now refer to them by these qualities, ‘the wise one’ and ‘the idiot (one)’. These free adjectives both represent each of the brothers and function as themes in their respective clauses. The following case is the usual way in JZ to express the notion of ‘the third one’:



(45) áw dṭlāha pıšle bıṭrāya basru ‘The third one was riding behind

them’ 887

It is clear that all adjectival syntagms in the examples represent a substantival entity. The same occurs when the adjective is a clause, syndetic or asyndetic: (46) balki tafqi ʾıbba ʾán d(ı)kṭāʾe ʾılla ‘Perhaps those who look for

her may find her’ 885


ūmıttūla trē kāšīye xazır ḥakumta.


áy dʾāla yamme mṣūra wēle ʾılla warda ūṭarpe dīde,

‘and she put two cups near the queen. On the one on the right were painted a rose and its leaves


uʾáy ttréʾ laswa ʾılla čū́ ṣurta


ʾáy qamēsa dʾıswa ʾılla ṣurtıt warda qam yāwāla ḥakumta ta jwanqa

and on the second one there was no picture


uʾay-xḗt dwīqāla bīza ūštēlu mızġas

the first one, on which there was the painting of the rose, the queen gave to the youngster

the other one she held in her hand, and they drank together’ 813–814

Exx. (45)–(47) all contain adjectival syntagms, each preceded by a determiner, which shows the referent’s gender and number. The free adjective is a nominal, and hence capable of functioning as theme. As far as its external syntax is concerned, it is not different from the substantive in this case and in object function as well. Its internal syntax is treated under §2.3.2 (for the syntactic adjective) and §4.1.1 (simple adjective).

1.3 COPULAR EXPRESSIONS JZ independent clauses generally occur with an exponent for a nexus (namely, the predicative link), be it a verbal form or a copular expression. The line separating the two is often very thin, meaning that the various copulas and the various forms of √hwy are often interchangeable with one another. The so-called ‘presentative copula’ is described here (§1.3.3), as well as below (§1.5.2), becauses it functions both as a copula (without any presentative function whatsoever) and as presentative. Verbal copulas (like √hwy or pıšle) behave like intransitive verbal forms whose most prominent feature is the fact that they do not occur with the qam preterite formation, since this formation necessitates a direct object. It



is true that verbs that have only an indirect complement do not occur in the qam preterite as well, but then the prepositional phrase that serves as object is consistent (e.g., mōnıxle b-/ıbb-). In the case of all these copular expressions, as a complement we get a diverse group. This group, the rheme paradigm, occurs with other, non-verbal copular expressions (the copula and the presentative) as well, and is different than any object paradigm (for which see §3.1.1). This paradigm consists of anything which can function as rheme, i.e., with the copula—(pro)nominals, adverb(ial syntagm)s, gerunds and perfect participles. It is this paradigm that unifies all these expressions, verbal or not. Dividing the various copula types between different sections is done for the sake of order and comprehension. Nevertheless, all these forms are part of one sub-system, which is characterized by a series of complementary distributions. In addition, when a copula is paired with a verbal rheme (that is, a perfect participle or a gerund), the resulting copular verbal forms also form oppositions inside the verbal system, they are part of this system as well.

1.3.1 The present copula The present copula (for forms see §7.3.1; named after its most common function, opposed to the presentative copula, for which see below, §1.1.3) occurs mainly in dialogue, but also in subordinate clauses in narrative. The following survey refers to the common features of the copula in all textemes and genres. The copula is not quite considered a verbal form, although it has some features in common with other verbal forms, and peculiarly, at least in the 1st and 2nd person affirmative, and in the negative, its origin is most probably verbal.9 The common features with verbal forms are as follows: First, the copula is in continuous opposition with verbal copulas (e.g., īle [indicative] vs. hāwe [subjunctive]). Second, when occurring with a verbal rheme, the copula has oppositions inside the tense-aspect system (for instance, wın ısya

In Goldenberg 1992:123 it is suggested that the origin of the copula is similar to that of other III-y verbs which occur, in the Nerwa homilies, in a special formation otherwise deemed as denoting the perfect (mṣın, mṣıt, i.e., preterite base and the 1st set of pronominal endings, reserved in JZ for passive preterites). It is suggested that the copular form īwın originates in *hwın, etc. 9



‘I have come’ vs. sēli ‘I came’). All this is described below under macrosyntax, chapters 5 and 6. The copula consists of a thematic index and a representation of the nexus, and always has to be supplied with a rheme. As such, it is essential in JZ to mark the predicative link between the theme and the rheme, since the Semitic concept of the non-verbal clause no longer exists in JZ as an independent clause (except for rare exclamatory expressions, e.g., axnan nāše mare-nāmus ‘we are respectable people’ 848. See Khan 2008b:854). The internal partition inside the present copula group is a. between 3rd person and 1st/2nd persons; b. between affirmative and negative; and c. between indpendent and subordinate. Further parameters are the type of rheme joining the copula—on one hand, verbal rhemes, i.e., the gerund and the perfect participle, on the other hand, (pro)nominal and adverbial rhemes. The affirmative copula (3rd pers.) The affirmative 3rd pers. copula is the most common, occurring basically with non-verbal rhemes (namely, substantives, adjectives, adverbials, etc.) and only seldom with verbal rhemes. In complementary distribution with this copula we have the presentative copula (see §1.1.3 below), which occurs more readily with verbal rhemes. The basic feature of the order (always x-īle) makes it always clear beyond any doubt what is the rheme, since the copula consistently follows it immediately. The copula has an allotagm, dīle-x, occurring only immediately following construct states (whether pronominal, nominal or adverbial), in other terms, it is the genitive, or attributive form of the copula (to be further discussed §2.2.4, exx. (116)–(133)). There is a tight juncture between the copula and the rheme, that is, the new information: it is expressed phonetically in the following examples: (48) ba ta ʾīman–(n)īla hatxa bāre ıl ḥakō ma ‘but till when is it

(that) thus will happen to the king?’ 646

(49) yassaq–(q)īlu ʾanya šoʾāle ta yıhāwa ‘These things are for-

bidden to give’ 697

(50) ūčú ḥ akim lá zeʾle mā ́ ʾēš–(š)īle ʾılle ‘No doctor knew what

illness he had (lit. is on him)’ 703

(51) hawēlu ʾanya ḥaqqıt dṓ jawāhar–(r)īlu ‘Here they are, they are

the price of the gem’ 670



What these cases have in common is an ad hoc doubling of the last consonant in the rheme, where it touches the copula. Another indication for this intimate relationship is found in a relic: (52) atta badle dīd dīla ʾılli bāmer dīde–h–īla ‘now the suit which is on

me, he’ll probably say it is his’ 220

(53) ūlá rʾıšle mā qıṣṭu–h–īla ‘and did not understand nor sense what

their story is’ 877

The h is a relic of the old (and otherwise obsolete) h in 3rd person suffix genitive pronouns. It is preserved only here, enclosed within this tight link. The last, as well as most common expression of this close juncture is the fact that an allomorph of the copula (-le/a/u) is the one that follows a vowel, see exx (54)–(58). Substantival rhemes: One normally finds common nouns functioning as rhemes: (54) ūxa baxta dīd hōyāwa smıxta kīʾēwa ʾınkan brōna–le uʾınkan

brāta–la ‘And (in case of) a woman who would be pregnant he would know whether it is a boy or whether it is a girl’ 226– 227

(55) de atta si kıs flāna qaṣāba xōri–le ‘So now, go to a certain

butcher, he is my friend’ 358

Proper nouns, especially in naming constructions, occur as rhemes as well: (56) lalxın sēle kısli xa ʾ arxa ʿ azīza u+rāba muḥsın ūšımme Mamo–le

‘Yesterday a dear and beautiful guest came to me, his name is Mamo’ 322

Pronouns: Several types of pronouns occur as rhemes. Personal pronouns are thematic/topical in nature and must be marked as rheme or as focus: (57) mmandın ʾızwa ūmād nāpıq ta kutxa mınnōxūn, āwa–le ḥaqqe ‘I

will cast a lot and what(ever) comes out for each of you, that (lit. he) is his portion’ 427



Demonstrative pronouns: ́ dīd gıbe kābısla bāžer dīdi ‘Look, it is (58) mēnxun ṓhā–le qaramān him, Qaraman who wants to conquer my city’ 127

Interrogative pronouns always occur preceding the copular expression: (59) mayle šūlax gō dē barīya ‘What is your business in this

wilderness?’ (lit. ‘what is it (namely) your business...’) 858

mayle is different—the copula following it has its full form, īle. Note the difficulty in translation: the nuclear clause is ‘what is it’. The rest (‘your business in this wilderness’) is an expansion of the theme (‘it’), which is represented by the theme index in the form of the copula (that is, in this case, the 3FS signal). Such an expansion is often a noun of some kind, or a pronoun. In the following example it is an adjectival clause: (60) āna gıbēna yāʾēna mani–le bıkābısla bāžer dīdi ‘I want to know,

who is it (who) will conquer my city’ 123

This is probably a direct interrogative clause (an interpretation as an indirect question is called for, but the structure explicitly indicates that mani ‘who’ is the rheme), where the thematic index in the copula (3MS) is expanded by a clause. The pattern is a cleft sentence (see §, whose purpose is to mark something other than the verbal form (in this case it is the interrogative) as rheme or as focus. Another example for a cleft pattern in a question is found in ex. (48) above. There is some difficulty in differentiating between direct and indirect questions: (61) ma bēsa–le ōha? ‘What (kind of a) house is this?’ 54 (62) bxāzin ēma bḗsa–le ‘I will see which house it is’ 162 (63) drēle ṭīna ıl kullu dargā(y)e dīd lá yaʾāx bēse ēma–le ‘He put dirt

on all doors so that we cannot know which is his house (lit. his house, which is it)' 165

Ex. (61) is a direct question. Since the question is ‘what (kind of) house’ the copula marks both as the rheme. Ex. (62) looks like an indirect question; note that the sequence interrogative–substantive, analyzed in (61) as one syntagm in rhematic function, should now be analyzed differently: ēma here introduces an indirect clause. The third example makes it even more complicated. The clausal object of not-knowing is often put in the form of an indirect question, that is, one would normally expect ēma bēsa–le, as in ex. (62). However, what we get in ex. (63) is bēse ēma–le ‘his house, which is it’,



perhaps to be analyzed as an asyndetic indirect question, in which bēse is the theme and what looks like the interrogative is marked as rheme10. The difficulty is that ēma is explicitly marked as rheme by the copula which immediately follows it, even though it is not an interrogative. Pronominal Nuclei are a group whose members represent a noun and are always in the construct state (and consequently function as nuclei of their constructions), like the pronoun dīd(-) ‘that of’. Besides signalling possession (with the pronominal suffixes), such a pronominal nucleus can adjectivize a following syntagm (whether substantival, pronominal or verbal):11 (64) qımle ōha mare bēsa ūṣrıxle ūmırre: bēsa dīdi–le ūsūse dīdi–la ‘The

landlord got up and cried out, saying: The house is mine and the mare is mine’ 220

In ex. (64), dīdi is a rhematic possessive pronominal adjective. In the following example it has a somewhat more complex function: (65) mbōqırru: dīd mani–le ōqadda māl? gımri: dīd flāna nāša–le

‘They asked: So much property, whose (lit. of whom) is it? They say: It is of some man’ 20

dīd, as a pronominal nucleus here adjectivizes, first the pronoun ‘who’, marking it as an adjectival ‘of whom’ or ‘whose’, and second, adjectivizing a substantival syntagm.12 The concord of the copula is in most cases with the theme (e.g., in ex. (65), with the substantive māl [MS]). In a few cases there is no theme, and then the concord is with the rheme (e.g., ex. (54)). Ex. (67) below has a substantive clause for a theme, and the concord is with the impersonal (whose default marking in JZ is feminine singular).

10 Compare la kīʾax kutxa ʾēma šāqıl ‘we do not know which (thing) each one should take’ (lit. ‘each which should take’ 427), where kutxa is the theme. 11 For this function of dīd, see §2.1.1. 12 The view of marking one entity as functioning differently is shown in Tesnière 1959 (termed “translation”). However, endorsing Tesnière’s views to their full extent entails the presupposition that there are tranformations in synchrony. This is not how synchrony is viewed in this framework, and hence, one can merely discuss the function of this construct state pronoun inside its syntagm.



Adjectives: The following example has an adjective as rheme, whose expansion (mın dīdi ‘than mine’) further identifies it as comparative: (66) māxulla qaḥra dīdox zōdanta–la mın dīdi? ‘Your grief, is it

indeed greater than mine?’ 239

In the following example the adjective occurs with a content clause (‘that I earned so much money’), which is the theme of the clause: (67) bāš īla ksıbli ō qadda pāre ‘It is good (that) I earned so much

money’ 335

The issue of rhematic expansion is somewhat difficult. The rheme slot, immediately preceding the copula, can house only relatively short syntagms; when the rhematic information exceeds a certain level of complexity or length, it is continued by other syntagms, see ex. (66) above, for a prepositional expansion (mın dīdi ‘than mine’). Ex. (68) has a relative clause for an expansion: (68) ay dammıt xūwe ġzēle ʾınnu aw gōra–le dīd mírwāle ṭāle: …

‘When the snake saw that it is the (same) man who had told him...’ 35

This might look like a cleft sentence (cf. ex. (60) above and §, but it is not; the dīd clause is an expansion of the rheme and hence part of it, whereas in a cleft it is an expansion of the theme. The same idea recurs in ex. (69): (69) marri ʾēma ʾurxa–la dʾāzın ʾıbba elqaṣır ḥakōma ‘Tell me which

way is it that I should walk to the king’s palace’ 757

That is, the rheme here is ‘which way that I should walk to the king’s palace’. It is broken in the middle by the copula for technical reasons, and continued immediately thereafter. Adverbial syntagms: These are not particularly common as rhemes. The following example shows the correspondence in information structure between a question and an answer: (70) ūnohāla kmá–ʾīle? hīl tıhōm īle ‘And the valley, how (deep) is it

(lit. how much is it)? It is (all the way) to the abyss’ 235



Note that the interrogative, obligatorily marked as rheme in the question, corresponds to the answer, marked as the rheme as well. The following example too has prepositional syntagms. In this example, two prepositional phrases in a row are marked as rheme: (71) ūla xašwıt ʾınni anya pāre ṭāli–lu. ġēr ta yatūme ūta gawāye–lu

‘Do not think that this money is for me. It is only for the orphans and the beggars’ 4–5

Ex. (72) has an adverbial syntagm consisting of the negative particle (in this case the semantic equivalent of ‘without’) and a kind of a compound noun: (72) xšūla balki ʾanya jwanqe lá yımma ūbāba –lu ‘She thought

perhaps these youngsters are orphans (lit. without mother and father)’ 858

The last example has a temporal adverb for rheme: (73) atta qarwāwıt palgıdyom īla ‘Now it is almost noon’ 75

Verbal rhemes: Although verbal rhemes are very rare with the affirmative present copula (less so with other copular expressions), this domain is essential in rendering transparent the verbal components. The latter are discussed above (§1.1) based upon Goldenberg 1985 and said to consist of 1. a subject index; 2. a verbal lexeme; and 3. the nexus between them. Here this analysis becomes transparent. Take a look at the following, unique example of the copula with a gerund as rheme: (74) baġ-wazīra mʿōjıbla ṭamā ́ mʿōwıqle ṓ gōra xūna hēš mxallōle–le

ʾīze ‘The vizier’s daughter wondered why the man delayed; (could it be that) he is still washing his hands?’ 494–495

This example comes from a stretch of free indirect discourse (for a description of which see § The copular/verbal complex mxallōle–le ʾī ze ‘he is washing his hands’ is perfectly analogous to the English structure, except that in JZ the theme is to be found inside the copula. The verbal components can be observed directly: 1. the subject index is the 3MS index inside the copula; 2. the verbal lexeme is represented by the gerund; and 3. the nexus between them is found in the copula, whose most important function is being an explicit exponent of this predicative link. Similar in principle, but ambiguous with regard to diathesis, is the perfect participle functioning as rheme:



(75) uʾāna gımʿajbın dlá-kīʾıt

cí-mındi mın faʿle ūxabır baxtāsa. ay dōhun mırta–la ‘And I am surprised that you do not know anything of the (evil) deeds and words of women. (So) one (lit. the) of them has said’ 933 +

(76) mırta–la bani-ʾādam lēbe emūnā. āna qam ʾamnínnōxun ‘It is said

(that) a human has no faith. I had faith in you’ 637

In ex. (75) the copula and the perfect participle mırta (√ʾmr ‘say’, FS) form an active present perfect ‘she has said’. Note that the auxiliary is different than what we usually find in most European languages, where we normally have the auxiliary ‘have’ when it is an active construction of a transitive verb. Here it is ‘be’, which creates ambiguity between passive and active. Ex. (76) is passive because the subject index, which in ex. (75) refers to a woman, represents in ex. (76) what is said, hence it is passive. So far we have examined the syntagm R-(ī)le. The latter has an allotagm, a syntactic alloform, dīle-R, which occurs immediately following any entity in the construct state. There are considerably less cases of this kind, but on the other hand, the syntactic attributive status being different from nominative status, we have a greater variety, viz., there are some examples for a verbal rheme with the present copula in the 3rd person. This situation is explained first and foremost by the fact that the presentative copula (§1.3.3), which otherwise complements the present copula as far as verbal rhemes are concerned, does not figure in this syntactic environment, that is, in attributive status. The main difference is the order—the rheme now follows the copula, which itself has an attributive alloform, dīle: ́ sēle bıd waʿıd dīd dīlu bıġlāqa tarʾāye ‘Šamʿān came (77) ṓha šamʿān at the time that they were (lit. are)13 closing the gates’ 149

(78) ūmbōqırru ʾıstāze hakan dīd dīle zīla ʾáw yōma ıl knıšta ‘and they

asked his teacher whether he has gone that day to the synagogue’ 589

(79) ūdımma kāweš rıš daw gōra dīd dīle pīša kēpa ‘and her blood

flows on the man who has become a stone’ 102–103

The tense issues of all predicative forms are discussed §6.3. There is no doubt that the temporal values in subordinate clauses are different than those, e.g., in the narrative (§5.6.4). 13



́ , dīd dīle mīsa qam day ēna ‘We now go to (80) bāsax kıs šamʿān Šamʿān who has died in front of this fountain’ 147

Ex. (77) has a gerund of a transitive lexeme, which occurs with a direct object. Exx. (78), (79) and (80) all have intransitive lexemes and hence must be analyzed as active. Except for ex. (80) (‘who has died’), we have typical complements—in ex. (78) it is the allative complement of the lexeme ‘go’ (‘whether he has gone today to the synagogue’), and in ex. (79) it is a rhematic complement14 (‘who has become a stone’). The allotagm dīle-R equally occurs with non-verbal rheme types: (81) dammıt ʾurre qam ʾēne taxmin dʾḗ baxta ddīla baxte ‘when the

thought of this woman who is his wife passed in front of his eyes’ 850

(82) ūxakma mınnu kīʾēlu ddīlu xakma nāše dgzavri ʾımmıt ṭōv

muxwāse ‘And some of them knew them to be (lit. that they are) people who hang out with his kind’ 940

The substantival rheme poses no problems, the rheme, as expected, immediately follows the copula. However, this is not always the case. Compare the following pair, featuring adjectival rhemes: (83) čımmāt́ dīle ṣāx ‘as long as he is well’ 770 (84) bale láswāle hāyi mın xōre ddīla kāse ṣrīsa ‘but he was not aware

of his friend whose belly was slit (lit. that his belly is slit)’ 956

Note that in ex. (84) the rheme does not immediately follow the copula, and the theme (kāse ‘his belly’) occurs in between them. This loose juncture is not attested with R-īle (it is attested with the presentative copula, however). It should be noted that in ex. (84) the result of this copular combination is a passive state rather than present perfect. This is arrived at either by context (the belly is not normally the agent of slitting), by the presence of a concrete object, or by an intransitive lexeme, which does not normally allow passive reading. The adverbial rheme is normal in this respect: (85) yōmıt tre dīd dīlu go barīya ’the second day that they are in the

wilderness’ 136


For the rhematic complement of √pyš see below, §1.3.4.



(86) ūʾanya nāse dīdīlu (sic) rıš dāra ‘and those people who are on

the tree’ 268

In the following examples, what seems like the verbal rheme nevertheless follows the copula, rather than precede it (as is expected from the basic pattern R-īle): (87) bābe rāba sqīla wēle, bıd bābe–hīle zīla ‘His father was beautiful,

it is by (the looks of) his father (that he has) gone’ SAG 5.73

PN ʿarabāye–lu mumṣe dīda ‘fPN, it is (lit. they were) Arabs (who have) nursed her’ SAG 5.61



The perfect participles in both examples, zīla ‘gone’ and mumṣe ‘nursed’ follow the copula. What immediately precedes the copula, i.e., the prepositional syntagm in ex. (87) and the nominal in ex. (88), is the rheme or rather (in this case) the focus. Therefore, the copula is combined with these syntagms rather than with the following perfect participles in both examples, which are consequently not the rhemes. What we have here are in fact cleft patterns, which are discussed below, § The affirmative copula (1st and 2nd pers.) The 1st and 2nd persons are inherently different from the 3rd person: they need not be introduced into discourse—they are presupposed—whereas the 3rd person has to be introduced and presupposed for it to be represented by a pronoun. In other words, for a 3rd person to be operative, one needs to know what it refers to. This is reflected in, e.g., the fact that dōlamánt wın ‘I am rich’ is a complete utterance, whereas dōlamánt īle is not, it would be complete (as far as information is concerned) only with the explicit referent of the 3rd person. Another angle is the difference between dialogue and narrative; in the former all persons are at play, whereas in the latter it is never the 2nd person. The affirmative 1s and 2 nd person copula is attested with both nonverbal rhemes (adverbial or nominal) or verbal-based rhemes (perfect participle or the gerund). It is different from the 3rd person in both form and syntax. Unlike the latter, there are two different patterns here: 1. R-wın15 (occurring with non-verbal rhemes) and 2. wın-R (occurring with all rheme wın here represents the entire 1st and 2nd persons copula (for the entire list see the morphological appendix, §7.3.1). 15



types). The form dīwın-R is an attributive allotagm of R-wın (occurring with non-verbal rhemes in attributive status, i.e., following the construct state). The following survey is arranged in a similar manner to the one given above with the 3rd person copula. Substantival rheme: This group often occurs with an explicit 1st and 2nd person pronouns as topics. R-wın: (89) āna brat ḥakōma wāna ‘I am a princess’ 114 (90) mādam ʾāhet brōni wētēn ‘Since you are my son... ’ 649 (91) āna bronōxūn wēn dīd zaʿli ‘I am your son who got lost’ 605

This is not needed when it is clear from a close reference to that person: (92) uḥakōma lágjāgır ʾılli čukūn brōne wēna ‘And the king will not be

angry with me since (I) am his son’ 648

(93) ūrāba gbannox čukūn bır ʾamōyi wētēn ūdımmi ūpısri wētēn ‘and I

love you very much because (you) are my cousin, (you) are my flesh and blood’ 685

wın-R: This pattern seems to occur when the substantive group is longer: (94) axōni atta waxni xūre bāš ‘My brother, now we are good

friends’ 249

(95) axnan waxni xamša nāše wal qam dāwıqlan ṓ kapōra pxá ʾēna

‘We are five people, (whom) the giant with one eye has caught us...’ 396–397

Four examples seemingly answer to the same criterion of length, but are nevertheless different: (96) āna wēna aw gōrıt ʾaqle ṣaqat wēlu ‘I am the man whose feet

were handicapped’ 55

(97) āna wēna šaqālıt kısta dīd pāre ‘I am the taker of the money bag’


(98) āna wāna dīd šqılli parrōke mınnox ‘I am the one who took the

textiles from you’ 13



In exx. (96), (97) and (98) the salient piece of information is āna ‘I’, as the speaker in all these cases identifies him- or herself with regard to an already given piece of information (a man with a handicap, somebody who took the bag money, someone who took the textiles). That is, they seem to belong in fact with R-wın, and not, as previously supposed, with wın-R. A similar situation is to be found in ex. (99): ́ ‘The king said: so you (99) mırre ḥakōma: ba ʾāhet wētēn axōni šamʿān are my brother Šamʿān!’ 158

The king’s brother had died, and now a stranger tells him his story from which the king understands that the stranger is his presumably dead brother. Hence ʾāhıt is the salient piece of information in this example. Note that, in addition to the fact that exx. (96)–(99) all answer to R-wın, the rheme in them is in fact a personal pronoun. This whole group will be discussed further below, since these cases clearly belong with focus (§6.1.3). Another related example is the following: (100) ba ʾāna wın, ču xa xın lēsın ‘But it is I (lit. I am me), there is no

one else’ LH1.12

A lost girl is telling a man about her long lost brother, and ex. (100) is his response, since the concept of ‘I’ is always a given in any conversation, the marking of ʾāna here has more to it than just being the rheme: it is focal (as the cases above), for it is not just new information (see §6.1.3), as the context tells us— ‘me and not another’. (101) yōna nāḥum manīle? mırri: ʾāna wēna ‘Who is Yona Nahum? I

said: it is I’ SYG 4.17

Here too, ʾāna is the answer and the rheme. To sum this up, with a substantival rheme, the unmarked order is Rwın. The other order, wın-R, occurs when the substantival syntagm is longer, and hence could be referred to as an allotagm. The order R-wın with a personal pronoun as rheme is marked for another reason—the unmarked function of the personal pronoun is thematic, and when formally rhematic, it is often focal, as in exx. (96)–(99) (but not in an answer, as in ex. (101)). dīwın-R, the attributive allotagm of R-wın, is not attested with substantives or pronouns. Adjectival rheme: The adjective (including its comparative and superlative functions) occurs both with and without an explicit personal pronoun:



(102) kxašwētūn ʾāna šızāna wēna, bale ʾāna bıž ʿāqıl wēn mınnōxun

‘You think I am crazy, but I am smarter than you’ 271

(103) ʾāna ʾay zurta wāna ‘I am the small(est of all the rest)’ AS 144

The comparative is marked lexically by bıš ‘more’, whereas the superlative is marked syntactically, in this case by the definite determiner. The following pair of examples shows R-wın order: (104) ʾāna ṣīwāya wēna uʾıtli baxta ūyalunke daqīqe ıl bēsa urāba faqir

waxni ‘I am a lumberjack and I have a wife and small children and (we) are very poor’ 666

(105) ḥaqqāna wın yān naḥaqqāna? ‘Am (I) right or wrong?’ 477

The order wın-R occurs with no apparent distribution: (106) āna ʾıdyo wın žġīla ‘Today I am busy’ 449 (107) uʾāna wın rāba čıhya ‘And I am very tired’ 627

The attributive allotagm (dīwın-R) occurs primarily with adjectival rhemes: (108) axni gıbax āzax ıl xā bāžer dīd dīwax kırye gāwa ‘We need to go

to a city in which we are hired laborers’ 53

(109) ūmayla sıbbe ddīwıt hatxa pēšt-rāṣt bıt šūlox ‘And what is the

reason that you are so sure in your work’ 514

(110) uʾāhet čukūn dīwıt qarīwa llubbi ṣurr dīd gōri pkašfánnēlox ‘And

you, since you are close to my heart, I will tell you the secret of my husband’ 815–816

́ (111) ūbımqaḷṭın basrax čımmaddīwın ṣāx ‘And I will be lovesick for you as long as I am well’ 883

Adverbial rheme: The order is basically wın-R: (112) xšūle hēš āna wēna go škafta ‘He thought I was (lit. am) still in

the cave’ 391

(113) xá yōma wın bıd ʾurxa dīd qṭāʾıt ṣīwe ‘One day I am on the way

to cut trees...’ 673

(114) wın qāmax ‘I am in front of you’ 896



Unlike above, this order here has nothing to do with the syntactic complexity of the adverbial syntagm. R-wın occurs with few adverbs as well as adverbial interrogative particles: (115) āna hatxa wın ‘I am like this’ 622 (116) mēka wētēn (wıt) āhet? ‘Where are you from?’ 421

Some ambiguity occurs when both sides of the copula could be in principle identified as rhemes: (117) ṭamā ́ ya baxti wat bıd julle čīqe? ‘Why, my wife, are you in torn

clothes?’ 693

Potentially, both the interrogative ‘why’ and the prespositional phrase at the end may be rhemes. In fact, however, it seems that the interrogative has priority over everything. The peculiarity here is that a vocative element (ya baxti) separates between the interrogative and the copula. This, however, is a widespread phenomenon in many languages, and hence viewed here as normal. Other similar cases are discussed below. Verbal-based rhemes: The affirmative 1st and 2nd person copula is different from the 3rd person mainly in freely joining verbal-based rhemes to create complex verbal forms. This difference is probably due to fact that the presentative copula is the one that combines with verbal-based rhemes in the 3rd person, and thus neither interferes nor competes with 1st and 2nd forms. With verbal-based rhemes only the order wın-R is attested. With the gerund this formation has the value of bounded or actual present: (118) wın mʿājōbe hīl ʾatta ‘I have been (lit. am) wondering till now’


(119) ʾaxnan kulla šāta wax bınṭāra ıl dē dalīsa ‘We have been (lit. are)

guarding this vine all year long’ 655

(120) ūʾıdyo wēna bımyāsa mın kıpna ‘but today I am dying of

hunger’ 292–3

With the perfect participle it gives the present perfect: (121) wın musya ʾımmi ṭ állasmá gūre ‘I have brought with me three

hundred men’ 954

(122) uʾāna uʾaxōni wax yımye ʾınnu ınkan xáʾ mınnēni māyes... ‘I and

my brother have sworn that if one of us dies...’ 278



The perfect is a linguistic term denoting an action that is basically referential to two points in time: the time it takes place and the time of reference (in this case it is the point of the utterance, but it could be another point, e.g., with the plusquamperfect, this reference point is the past). The perfect gives us both the preceding action and the resulting state of affairs. For instance, ex. (121), actually bringing along the soldiers takes place before the conversation, but the resulting state is that now (at the time of convesation) there are three hundred men at the speaker’s disposal, which is highly relevant to the current power balance. The perfect can be found in any texteme, not necessarily with tense-denoting auxiliaries but also with forms which denote various moods or aspects (e.g., subjunctive perfect), where the same idea of two reference points is found independently of a specific time. In several cases one encounters entities on both sides of the copula, each of which may in principle be regarded as the rheme: (123) āna mın bāžır raḥuqta wın ʾısya 698

Ex. (123) calls initially for two possible analyses: ‘I have come (=wın ʾısya) from a remote city’ (as a realization of the only order found with verbal rheme, wın-R) OR ‘I am from a remote city (mın bāžır raḥuqta wın)’ (realization of the usual order for adverbial rhemes R-wın) when the last part, the perfect participle ısya ‘gekommen’ is left outside. In this analysis, the example could be interpreted as some kind of a cleft pattern (see the end of §, viz., ‘it is from a remote city (that I have) come’. In fact, a solution for this locus is still impossible. The following examples are rather similar, containing an interrogative which precedes the copula: (124) ha, ıl ma wıt ʾısya? ‘Hey, what have you come for?’ 637 (125) ṭamā́ wētın ġzīʾa gıyānox ıl dē brāta sqılta? ‘Why have you

wrapped yourself around this pretty girl?’ 35

(126) ṭamā ́ wētın bīzāla burxa ūmēnōxe bıd ʾarʾa ‘Why are you walking

the road and looking at the ground?’ 520

In these three examples it seems plausible that the rheme is the interrogative, but it can be proved only indirectly, by showing that an interrogative, except for special circumstances, is aways the rheme.


CHAPTER ONE The negative copula (3rd person) This copula is not the syntactic mirror image of the affirmative copula:16 unlike the latter, it occurs with verbal rhemes, and always precedes the rheme. Moreover, its functions are divided between those of the present copula and those of the presentative copula. The presentative function is evident after verbs of perception: (127) bale mōnıxle bıt ʾasıqse lēwa ʾasıqsa dīde wēla mxōlapta ‘But he

looked at his ring: it is not his ring, it is exchanged’ 309–310

In ex. (127) the syntactic parallelism between lēwa and the presentative copula wēla makes this clear beyond any doubt. The present copula never occurs in these syntactic circumstances. Any type of rheme can join the negative copula in this function: (128) kxazya jwanqa dīda lē ʾaw qamāya ‘She sees her youngster (and

lo) he is not the former one’ 905

(129) kpasxın ēni kxāzın zīne lēwa xazri ūlēwe xılma ‘I open my eyes, I

see (lo) Zine is not around me and it is not a dream’ 320

In exx. (128) and (129) we see as rheme an adjective and a prepositional syntagm respectively. Ex. (130) has an adverb as rheme: (130) gxazya—ma gxazya?—nērīya wēle mīsa ūṭḷība skīna dīde wēla

tāma ūʾāwa lēwe tāma ‘She sees—what does she see?—(lo) the goat is dead, and her fiancé’s knife is there but he is not there’ 235

Here, like in ex. (127) above, there is a syntactic parallelism between lēwe and the presentative copula wēle, this time, however, with exactly the same rheme (tāma 'there'). Ex. (131) has a gerund for rheme: (131) māt wēle bıṭʾāya, lēwe bıġzāya ču ʿattāla ‘As much as he was

searching, he cannot find any porter’ 571

Indeed in this example (as well as another, ex. (137) above) there is no verb of perception, but the function is rather clear. Note also that the present copula does not normally join the narrative line in an independent clause, Historically and externally this is obvious, of course: the affirmative in JZ is constructed with -īl- while the negative form is similar to the 1st and 2nd persons, consisting of *īwe. 16



whereas lēwe can and does. This tells us something about its temporal range, which is more like the presentative copula’s rather than that of the present copula. Other than its presentative function, the negative copula is also the counterpart of the affirmative copula in subordinate clauses: (132) yāʾet ʾıbba ʾınnu ōha lēwe xulma ‘You should realize that this is

not a dream’ 363

(133) uxšūle xılma–le ‘he thought it was (lit. is) a dream’ 309 (134) bale mın tōs uʿıjaj la muxwēlu mán-īlu ūmani lēwu ‘But because of

the dust they did not show who they were (lit. are) and who they were (lit. are) not’ 846

(135) ūmbōqırru ʾıstāze hakan dīd dīle zīla ʾáw yōma ıl knıšta. mjōwıble

ʾıstāza dīd lēwe zīla brōne ıl knıšta ‘and they asked his teacher whether he had (lit. has) gone that day to the synagogue. The teacher answered that his son had (lit. has) not gone to the synagogue’ 589

In independent clauses the negative copula readily joins verbal rhemes: (136) hēž lē bırya mındi ‘Nothing has happened yet’ 850 (137) zıllu, ʾurru go knıšta, čuxa lēs ʾısya, lēwu jmīʿe nāše ‘So, they went

into the synagogue, no one had come, people had (lit. have) not (yet) gathered’ MA 7.4

(138) yōma, hah, gnēle, lēwe gınya, gnēle, lēwe gınya ‘The sunlight was

just about to disappear any minute lit. ended and has not ended)’ SYG 2.43

(139) ʿabdi wēle ʾısya. ʾī de baxte hēš lēwa zılta, ‘(I heard that) Abdi

has come (back). Ide his wife had (lit. has) not yet gone (with him)’ SYG 2.49

(140) hēš lēwu mpīqe ‘they have not come out yet’ SYG 4.6 The negative copula (1st and 2nd person) This copula mostly occurs with verbal rhemes. Occurrences with other rhemes are very rare: (141) ya brōni, ā het ma lēwet brōni ūʾāna bābox ‘My son, are you not

my son and I (am) your father?’ 501



(142) ūʾāna lēwen bġammıt māl dīdi ‘I do not care about my property’


Within the verbal rhemes, the gerund is very rare as well: (143) ṃāṭo la gzadʾıt mınni dhatxa ḥmıllox barqūli ulēt bīzāla ‘How

come you are not afraid of me that you stand thus in front of me and you are not going away?’ A 1.2

We are left, then, with one prominent function, signalling negative present perfect: ́ dṓ (144) ʾaxnan kīʾax qārax kullu līšāne, bale lēwax xızye muġ līšān

ḥammaš ‘We can read all languages, but we have not seen (anything) like the language of this book’ 591

(145) ūhīĺ ʾatta hēž lēwın mzubna kıtte mınnōhun ‘And until now I have

not yet sold any of them’ 431

(146) wınne ya gōri mın zūna lēwet zīla kıs xāsox “Indeed, husband, for

a long time you have not gone to your sister’ 200

The meaning of the perfect in JZ is discussed at the end of § above. Here something needs to be added about the negative perfect, which is somewhat different from the affirmative: it is still referential to two temporal points—the time of reference (in this case, the moment of utterance), and the event, except that the event has never happened. The form then covers the timespan from one point, which is often abstract, but not always: (147) mın ʿumri lēwın xızya... ‘In my life I have never seen... ’ AS 817

(compare 282)

(148) bābo, qṭılālox šōše? gēmır: la, lē(wı)n qṭīla dīda ‘Father, did you

kill Shoshe? He says: No, I have not killed her’ LH 3.4

In ex. (147), the point is specified, the span is the speaker’s entire life. In ex. (148), the range is even more limited, as this killing could only have happened when the first speaker had left home for a short time. As such, this copula functions mostly within the verbal system.

1.3.2 the verb hwāya All the forms of this verb are part of the predicative system. The most common is the form wēle, the preterite, due to the narrative nature of our texts. wēle needs to be discussed also because it is homonymous with the presenta-



tive copula, but it can be distinguished on syntactic grounds. In addition, we have the following forms: the present kāwe, the future/modal pāwe and the subjunctive hāwe. All these forms also occur with the ‘one step back’ morpheme -wa(-). These forms are in a paradigmatic relationship with the copula when different tenses and modi are expressed. R-wēle The verbal syntagm R-wēle (to be sharply distinguished from the copular presentative wēle-R) is the preterite of the verb hwāya. It is basically the past/narrative form17 of the present copula, but is not identical to the latter in its rheme distribution, occurring primarily with nominal rhemes (very few adverbials) and with perfect participles (gerunds occur almost exclusively in subordination). The form itself is complicated by one by-form (wēwāle) and one allotagm (wēle-R), which occurs only in attributive function. The byform is in use in the material coming from one informant, Z (see §§ 0.3 and 0.5 in the Introduction). In that material both R-wēle and wēwāle are found, with no appreciable difference in their distribution. There are, however, several important differences between them: while wēwāle has only 3rd person forms (not in Fassberg 2003, where the entire personal spectrum is attested), R-wēle occurs in 1st and 2nd persons as well. We meticulously insist on using the syntagm R-wēle since the order is linguistically pertinent, distinguishing it from the presentative copula (wēle-R). The form wēwāle, however, is not homonymous with any other entities, and is hence distinctly the preterite of √hwy. The following table summarizes all these differences and the distribution: syntactic environment

personal domain

independent clause

attributive clause

past √hwy 1


wēle-R (~R-wēle)

all persons



3rd person

past √hwy 2


3rd person

Although the morphological structure of wēwāle is identical to the form šqılwāle, it does not have plusquampreteritum value, and it is a free variant of the syntagm R-wēle.


These differences are discussed below, under macrosyntax.



The syntagm R-wēle is differentiated from other opposing syntagms in the following examples: (149) hawwa šıtqe smıxta wēla. garxıt ʾ e wēla smıxta ‘Well, well, she

was pregnant last year. (There) she is pregnant again’ SAG 4.3

Ex. (149) is a textbook example for the semantic difference between R-wēle and the presentative copula wēle-R; the first, as testified by the adverb šıtqe ‘last year’, has a past reference. The second refers to the point of utterance, rather than the past (and this is its value in dialogue outside special presentative envronments like verbs of perception, etc.). The following example opposes the affirmative R-wēle to the negative la wēle-R: (150) ıswa trē axawāsa xā ʿāqıl uxā ́ šızāna. aw ʿāqıl gwīra wēle uʾaw

šızāna lá wēle gwīra ‘There were two brothers, a smart one and a crazy one. The smart one was married and the crazy one was not married’ 241

In attributive position, most (but not all!) occurrences of this wēle are in the allotagm wēle-R, thus there is now symmetry with la wēle-R: (151) wēla mın māt́ ṭʾēle rāba, tfıqle rıš xá nunīsa dīd lá wēla xasta ūdīd

množ gyāna wēla muḥẓarta ta mzabōne ‘So much was he looking, he hit upon one fish which was not new and which alone was ready for sale’ 727

The following pair shows, using the same syntagm, the complementary distribution between the independent R-wēle and the attributive wēle-R: (152) ṓ šūla pıšle maraq go lıbbıd dṓ nāša dwēle ḥmīla mımbínnōke hīl

barbar ʿāṣırta ‘This thing became an obsession in the heart of the man who was standing from morning till early evening’ 512

(153) škılle bābe bımxāya bbıllūre hīl ʿāṣırta ūbrōne ḥmīla wēle ‘His

father began playing his flute and his son was standing...’ 447

The vast majority of the cases where R-wēle occurs in attributive position it is as the allotagm wēle-R. As for the other oppositions of R-wēle, they often depend on the kind of rheme, and even on the kind of lexeme. For instance, the participles of ytāwa ‘to sit’ and ḥmāla ‘to stand’ are mostly statives, not denoting any perfect (these lexemes are not attested with a gerund). They are hence opposable to the form kšāqılwa. Compare ex. (153) to ex. (154) above:



(154) kúdyom ō gōra gēzılwa … ukḥāmılwa hīl gnapqīwa ʾán talmīze mın

knıšta ‘Everyday the man would go ... and would stand until the students were leaving school’ 576

The difference is (for now; a more detailed discussion is found below, under macrosyntax) a temporary state (ḥmīla wēle) vs. description of a fixed feature. The same opposition is found in the following pair of examples: (155) rāba nāše ʿızīme wēlu ūgawōhun tīwa wēle Bakko ‘Many people

were invited and among them was sitting Bakko’ 327

(156) kudyom gēzīwa ḗ naqla kútrū-́ kutru ldáy dūka, bāba gēmırwa bıd

bıllūre ū brōne gyātūwa xazre hīl ʿ āṣırta ‘Everyday they would go, now both of them, to that place, his father would play (lit. say) his flute and his son would sit near him till the evening’ 448

With other lexemes, notably intransitive ones, the participle denotes a perfect, and with R-wēle, the result is plusquamperfectum. As such, it is opposable to šqılwāle: (157) ūšımmıt zīne mpīqa wēle uzvīra bıt sqīlūsa dīda hīl bāžēre raḥūqe

‘Zine’s name had gone out and spread for her beauty to faraway cities’ 315

(158) ūʾēha kulla brēla bıd ʿājıbwāsa ūšımmıt dan nāše mpíqwāle go kullu

dan bāžēre ‘And all this happened wonderfully and the name of these people had gone out in all these cities’ 300

(159) bale bābox la wēle lbēsa. ʾamrannox, bābox bıš kfāhımwa, bābox

zīla lʿaskarīya wēle ‘but your father was not at home... your father was more knowledgeable [about these things], (but) he had gone to the army’ SAG 3-34

(160) ınnu yōmıt qam kardatti mın bēsa zílwāli go bēsıt qōrāsa ‘Well, the

day you chased me out of the house I had gone to a cemetery’ 65

Further value differences between these forms are discussed under macrosyntax (§§5.4.1 and 6.3.1). Another, rare opposition is found when the participle is transitive and passive; in that case the result is basically a passive state:



(161) an tmanya yōme dīd wēlu go ḥuppa ʿzīme wēlu ıl bē bāb ṭḷ ıpte

‘ṭhese eight days that they were in the bridal chamber they were invited to the house of the bride’s father’ 546

(162) xá yōma sēle ʿızāma bahlūl ldaʿwa ‘One day Bahlul was invited

to a wedding’ 938

The difference is that in ex. (161) we have a passive state (‘eingeladen sein’), whereas in ex. (162) it is a dynamic passive (‘eingeladen werden’). The same idea applies to the following pair: (163) qāqıbat mutwe wēlu rıž nūra ʾısri uʾarbá saʿe lágbāṭílwa mın xēsu

‘The caldrons were placed (waren ... gestzt) twenty four hours on fire which would not stop (burning) under them’ 466

(164) ūʾō jwanqa sēle matōwe rīš kursi ‘and this youngster was seated

(wurde ... gesetzt) on a chair’ 797

The foregoing examples of wēle occurred mostly with the participle. The rheme paradigm of R-wēle and wēwāle (both in independent clauses) is mainly nominal, with a relatively small number of participles and adverbial syntagms and rarely with the gerund. Its tendency towards joining nominal rhemes could be explained by the fact that many verbal forms are available as exponents of past events, but there are no other neutral copulative expressions in this temporal domain.18 However, in subordinate clauses, the distribution is more even and one comes across verbal and adverbial rhemes. The reason for this might be found in the characterization of the subordinate clause (discussed under macrosyntax, §5.6.4). That is, in view of the different distribution inside subordinate clause, one should expect slightly different values for virtually the same forms. R-wēle with gerundial rheme signals, as expected, a bounded continuous past. It is differentiated from the form kšāqılwa in that the latter denotes habitual, unbounded action: (165) kulle waʿda mṣalōye wēlu ūbıḅxāya ta ʾarxa dōhūn ‘All the time

they were praying and weeping for their guest’ 618

The verbal copula pıšle shows various nuances of Aktionsart and the presentative copula has other functions, viz., as presentative (for which see §§1.3.3 and below) and as a circumstantial clause (§3.2). 18



kulle waʿda ‘all the time’/ kulle yōma ‘all day (long)’ are used as time boundaries with the gerunds. (166) basır kxāḷıṣwa mın qrāya gdārḗwāle ʾaw ḥammaš xe xá baḷāṭa ıl xá

ʾal dīd ʾōḍa dīd la xāzḗwāle ʾ an nāše nuxrāye ū gbāxēwa ta bābe ūyımme ʾan sāwōne ‘After he would finish reading he would put the book under a floor tile in one side of the room so that these strangers do not find it, and would weep about his elderly parents’ 587

In this example weeping takes place on a regular basis. The same is true for the following example: (167) kuḍyōm gımṣālēwa ta ʾīlāha ūkṭālıbwa xa yāla uʾīlāha la hullēle ču

yāla ‘Everyday he would pray to god and would ask for a child but god did not give him any child’ 576

Other rheme types are illustrated in the following examples. Adverbial rheme: (168) uʾēha mın ʾīlā ́ wēla ūlēwe sūji ‘And it was because of god, and is

not my fault’ 711

(169) kulla mın baxti wḗwāla ‘It was all because of my wife’ 540 (170) uḥakṓm day bāžer nuxrēsa dīd wēle brōnu gāwa pıšle naṣax ‘and

the king of that foreign city which their son was in (it) became sick’ 590

The examples all show an adverbial rheme. Exx. (168) and (169) are opposed, with no appreciable difference in value. Ex. (170) has an adverbial rheme as well inside an adjective clause. Note the allotagm wēle-R. The same phenomenon is illustrated in the following three examples, this time, however, the rheme is substantival: (171) ōha ḥakō ma íswāle ṭḷāha yalunke: xā rūwa, šımme aḥmad wēle,

uxā ́ palgāya, šımme maḥamad wēle, ūxā ́ zōra, šımme mērzá maḥamad wēle ‘This king had three children: an older one, his name was Aḥmad, and a middle one, his name was Maḥammad, and a small one, his name was Merzá Maḥammad’ 69

́ , ūʾaw xet šımme wḗwāle baršamān ́ ‘One (172) xa šımme wḗwāle bāšmān his name was Bašmān, the other his name was Baršamān’ 121



(173) gmaxwēwa ʾılle dwēle bır xa mıllítta nuxrēsa ‘It was apparent

that he was of (lit. son) a foreign nation’ 790


occurs in 1st and 2nd person as well:

(174) bır ṭmānıʾsar wēli yōmıt mpıqli mın balad dēni ‘I was eighteen the

day I left our country’ 424

Pronominal rheme: With pronouns, R-wēle seems to be limited to demonstrative and interrogative pronouns: (175) ōha wēle šūl daw gōra kudyom ‘This was the business of the man

everyday’ 626

(176) ṓha wḗwāle ḥālu kud šāta ušāta ‘This was their situation every

year’ 184

(177) ēka wēlox hīl dō waʿda? ‘Where were you until now?’434 (178) ʾē ma wēla sēlan bıd dō +mıṭra ‘Why (lit. what) was it we came in

this rain?’ A 6.8

And finally with adjectives: (179) ʾāna rāba dōlamánt wēli ‘I was very rich’ 702 (180) nāše dīda ʾēnu yarīxāne wēlu

elongated’ 422

‘Its people, their eyes were

This verbal form participates in many oppositions, in the copular system as well as in the verbal system, since most of the events naturally occur in either past or narrative realm. The other forms of hwāya have considerably less occurrences. Present kāwe The present form of hwāya is not common. It is most frequently opposed to the copula. In the first pair, where the rheme is adverbial, the difference seems to be aspectual: (181) kud yōm āhet ay dammıt napqēten mın bēsa lıbbi kāwe bıd zdoʾsa

ūlák ṭōʾen hīl lá ʾasēten ‘Everyday, when you leave the house, my heart is afraid (lit. at fear) and I do not fall asleep until you (lit. do not) come’ 355



(182) qū́ dmṓx ʾēnox wēlu pšınsa ‘Go to sleep, your are sleepy (lit.

your eyes are asleep)’ 867

The function of the form kāwe, like the form kšāqılwa, hints at something regular or habitual whereas the copula wēle-x may refer to something which happens at the moment of utterance: (183) ılmūjıb kīʾın ʾāna, kapōre ay dammıt kṭōʾi, ʾēnu kāwe psīxe ‘As far

as I know, when giants fall asleep, their eyes are open’ 414

(184) čukun kīʾe nāše rūwe lıbbu kāwe rwīxa ‘Because he knew (that)

great people, their heart is wide’ 483

rāba mapṣūṭ́ ūrıẓya wın mın baxti, īs ʾahawā ́ šalóm gawēni uʾāna rıẓya wın bıt ḥaqqi mad mıšadırri ʾīlāha ʾāwa–le rısqi (rız-) ūbaxti qnıʿta–la bıt ʿīš dīdi ‘I am very happy and pleased with my wife, there is love and peace between us and I am pleased with my share, whatever God may send me—that is my livelihood, and my wife is satisfied with my livelihood’ 515

(185) ʾāna


Exx. (183), (184) and (185) all have adjectives as rhemes. Exx. (183) and (184) refer to a characteristic of giants and of great people respectively, that is, they have generic theme and generic rheme, which together make up a generic statement. On the other hand, ex. (185) has a specific theme predicated with a permanent, general state, together making up a habitual statement (see Khan 2008b:641–643). The point here, however, is that kāwe is marked for genericity (representing, for instance, characteristic feature(s) of a group—giants, rich people), whereas the copula may denote the entire spectrum, from generic down to specific. With an adverbial interrogative: (186) marri walox ēka gēzıt kudyom uʾēka kāwıt hīl ʿāṣırta ūmēka

gmēsıtte kudyom ō dıhwa ‘Tell me where do you go everyday and where are you till the evening and where do you bring the gold coin evryday’ 446

(187) marri, yā ́ naša xšīwa, mēka-wıt ʾāhet ūmani-wıt ‘Tell me, o

important man, where are you from and who are you’ 943

(188) ıl mā ́ wıt ʾısya? ‘What is it (that you have) come for?’ 637

Whereas kāwıt in ex. (186) refers to something happening on a regular basis, ex. (187) does not have to, as is clear from ex. (188), probably a cleft construction with the present copula referring to a one-time occurrence. It is important to mention that kāwe is different in this respect from other kšāqıl



form, which can denote this habituality, but are not marked for it. kāwe, being opposed to the present copula, is marked for this habitual or generic function, which is part of its value (see further on §6.3.5). Quite different are the negative forms of kāwe. Like lakšāqıl forms, they function as the negative counterpart of the form pšāqıl: (189) mattaw lēʾal lá kāye naqōṣa

cí mındi ıl brātox uʾīxālu lá kāye ṭḷāhā ́ dāne bıd yōma ṭarpısılqe ‘From now on, nothing will be missing for your daughter, and her food will not be mangold three times a day’ 573 fut +

(190) mḗsīle waʿıd palgızlal. dunye pōya karr ‘Bring him at the time of

midnight. The world will be quiet’ 929

Ex. (190) has the form pōya which here denotes the future. Ex. (189) exhibits the negative future, perhaps with a touch of habituality. This habituality does not exist at all in the following example, occuring with the participle: (191) walax štōq, la šāmeʾ, hēž la kāye zīla ‘Shut up, you! Let him not

hear, he has probably still not gone’ A 1.9

(192) atta pāwe baxti ū yalunke dīdi ḥmīle hīvīti ‘Now my wife and

children are probably waiting for me’ 410

(193) atta ši yısá dʾōr lbēsa. hēž lē bırya mındi. Even now, come (and)

return home. Nothing has happened yet’ 850

The form pšāqıl has another notion, which is modal, rather than temporal: it denotes probability. This is the case with pāwe in ex. (192). The same thing occurs in ex. (191), in which the negative form denotes this probability. Ex. (193), on the other hand, signals indicative present perfect. Note that all three cases are basically referential to the same time, the difference between them being modal. The form kāwēwa denotes genericity or unboundedness, as any kšāqılwa form: (194) áy-dammıt kāyēwa mnōše aw yāla gō ʾōḍa kšāqílwāle ḥammaš

berēšıt dīde ū(q)qārēwa ʾıbbe ‘Whenever the child would be alone in the room he would take his bible (lit. Genesis book) and read in it’ 587

(195) ṓha faqir mnōše wḗwāle, láswāle lá baxta ūlá yalunke, bas mnōše

wḗwāle ‘This poor man was alone, he did not have a wife nor children, he was just by himself’ 519



In ex. (194) there is unbounded recurrent (and hence non-specific) action. On the other hand, the wē(wā)le preterite forms in ex. (195) reflect an unchanging state. The opposite in a way is illustrated in the following pair of examples: (196) xá yōma ıswa trē gūre. xá gōra kulle waʿda tīwa wēle ıl bēsa

ūbıqrāya ūxá gōra kulle yōma bımxāya kāwēwa gēra ūqašṭa bıd šımme ‘Once there were two men. One man was sitting all the time and reading, and one man would shoot (lit. was being shooting) arrow and bow to the sky all day (long)’ 301

(197) kulle waʿda mṣalōye wēlu ūbıbxāya ta ʾarxa dōhun ‘All the time

they were praying and weeping for their guest’ 618

We have shown above (ex. (165), reproduced here as ex. (197)) that R-wēle with the gerund shows a bound continuous past. When the auxiliary is kāwēwa, the value of the form is marked as describing a characteristic, nonspecific action (ex. (196)), whereas bıšqāla wēle could either describe a characteristic, non-specific action (ex. (196)), or describe a more specific situation (as in ex. (197)), which has no regular basis. This means that wēle is unmarked vis-à-vis the form kāwēwa. Textemic differences between these forms are resumed under macrosyntax, §5.5.1 and fn. 24 in § The order of the elements tends to be kāwe-R, except when the rheme is an interrogative (ex. (186)), and note ex. (196), where the gerund precedes the verbal form. Negative forms always precede the rheme. Future/modal pāwe The double function of pāwe has been discussed in the previous section. The various rheme types are illustrated: (198) ʾatta ʾāhat pōyat baxti ʿazızta ‘Now you will be my dear wife’


(199) yā ́ nāša, xa mındi wal gyāsıq qam ʾēn lıbbi kxašwāna yāla pāwe ‘O

man, something comes up in front of my heart, I think it may be a child’ 750

bāš́ pāye ta qzál dḗ kapurta ‘I thought that the king’s sword with which I killed the second giant would/might be good for this giantess’ neck’ 415

(200) āna mtuxmınni dıʾṓ sēpa dıqṭıḷḷi ʾıbbe ʾṓ kapōrıt trḗ




(201) pāwıt axxa čımmāt́ dīwıt ṣāx ‘You will be here as long as you are

well’ 410

(202) ēma gōra ʾōha pāye? ‘What man would this one be?’ 572

́ pāwe bīmāra ṓha pāsuq: ‘From now (203) mın atta ūhīl ʾabade ʾabadān to eternity they will be saying this verse:’ 58

(204) baxtıt gōra ʾṓ šıkıl gwāra pōya xšūta mā ́ xāye mnabla (//

mmáʾūra) gō dunye ‘A wife (in) this kind of marriage has probably thought what (kind) of life she will lead in the world’ 821–822

The order of the elements is like kāwe, but there are, here too, examples with rheme-first order, as exx. (200) and (202). Included in this order, but somewhat different in being focal (and hence discussed under FSP below, §6.1.3), is the following example: ́ a pāwe ‘Any(one who) (205) kud ā se šāqel āfıllu xá mallīn mınnu, āw comes and takes even one cent from them, will be him’ 169

The form pāwe occurs with only one argument, which is analyzed as the rheme. Both pāwe and kāwe occur with all persons. Subjunctive hāwe The subjunctive form hāwe occurs, like kāwe and pāwe, in all persons and basically with a following rheme. hāwe is much more common than kāwe and pāwe, and different from them in that it usually has no temporal, but rather modal values. In various subordinate clauses these values are somewhat different, and the form could either be devoid of any value (e.g., when obligatory) or acquire temporal values (when in a paradigm). The rhemes which join hāwe are mostly nominal and adverbial. Verbal rhemes are not very common, active perfect participles are few and the gerund is attested once. Nominal rhemes constitute the majority of cases: (206) baxtāsa dēni baxtāś bēsa wēlu ūlá-wēlu līpe ṭāʾe mnōšu, lá gōra

dhāwe šıvāna dōhun ‘Our women were domestic wives and were not used to search by themselves, without a man who would be their shepherd’ 845

(207) mġabīne ʾıllox hāwıt bızvāra go barīye. lak šākıllox hāwıt mšúttıta,

bale kēfox–īle ‘It is a pity for you to be wandering in the



wilderness. It does not suit you to be exiled, but as you wish’ 950 Note that within certain expressions, when obligatory, the subjunctive is not opposed to any other form, and hence has no value. Some aspectual value is possible when coupled with perfect participles and gerunds, however (ex. (207)). Within the nominal group pronouns are very rare; there are no interrogatives occurring with the subjunctive, and in fact the only examples we have are the following pair of examples: ́ a ‘The king’s (208) mısāsıt ḥakōma bıž durust–īla. lá qablın la hōya āy

scale is more precise. I will not agree (that) it should not be it’ 253

The pronoun āya in the example functions as a rheme. The importance of this example is that it corroborates the special category of the rheme. The personal pronouns āwa, āya and āni do not normally occur as object of transitive verbs.19 Here, in the function of rheme, they do, and constitute one of the distinctive features of the rheme paradigm. The indefinite pronoun xa is more flexible in this respect and occurs in many functions, the rheme being merely one of them: (209) balkin hāwēla brōna, gıbe hāwe ta brōna xaʾ ‘If perchance she

would have a boy, it was necessary to have (lit. that there be) one for a boy SAG 3.2

xa is not the theme, but rather the rheme of hāwe: hāwe ta is the subjunctive form parallel to the expression ıtl- ‘x has’. Unlike English, where the complement of the verbal form of ‘have’ is its object,20 here this expression is based on the existential expression īs and a dative marker, when the possessed, or the existant (in our case xa), is the rheme. Adjectives as rheme basically occur, like substantives, immediately following the verbal form: (210) bale hāwıt ʿāqıl ‘but be smart’ 802 (211) yā ḥakumta hōyat ṣāx ūbassımta ‘O queen, be alive and well’ 802 Except when focal, see § Although even in English there are various points of view regarding these expressions, e.g., Lambrecht 2000. 19 20



The following example has the form štāwe, which is an amalgam of the particle šud and the form hāwe, whose function together is an explicitly independent or marked as modal: (212) ūštāwe lıbbox skīna ‘and may your heart be calm’ 546

Occasionally this immediacy is interrupted by the theme, wedged between the verbal form and the rheme: (213) si hāwıt mburxıt ʾīlāha uhōya ʾurxox ṣlıḥta ‘Go, be blessed by God,

and may your way be successful’ 583–584

(214) lak šākıllox hāwıt ʾāhıt faqīr ‘It does not fit you to be (lit. that you

be) poor’ 485

In the following example, the adjective precedes the verbal form. The reason for this difference is not clear as yet: (215) ūyımēle ʾınnu la hōya ēha brāta ṭāle, ḥaramta hōya ılle ġēr brāta

‘and he swore that (if) this girl would not be his, any other girl would be forbidden for him’ 83–84

Adverbial rheme is not very common. A prepositional syntagm: (216) dı-si uʾīlāha hāye ʾımmox ‘So go and God be with you’ 603 (217) pšātax ʾāna uʾāwa mux tre dōste kan gıbe štōyan mux xā ́ tıjjar

muxwāse ‘We will drink, him and me, like two friends, if he wants, let me be like a merchant, like him’ 805

Another adverbial syntagm is made up of a substantive preceded by the negative particle la, together denoting ‘without x’: (218) mġabīne ʾıllōxun hāwētūn la sármıyan ‘It is a pity that you should

be without a person in charge’ 862

The verbal rheme group is not very common. Included in this group are perfect participles as well: (219) balki ʾīlā ́ lá-ʾāwıs wal sēle brēše xa maqaddır, lakun hāye mšulxa

ūmpīla bdunye ūlá21 hāwe qṭīla gō xā gali yan xa nohāla ‘Perhaps, God forbid, an accident has befallen him, that (lit. lest) he be

This is not the negative particle, but an asseverative particle (see Sabar 2002b:203a). 21



robbed and left out there (lit. fallen in the world) and probably killed in some valley’ 843–844 Despite the fact that these participles are related to a verbal lexeme, their passive function makes them, in a certain way, very similar syntactically to any adjective. Yet, they are comparable to passive constructions, which makes them, in a wider sense, a part of the opposition system of verbal forms: (220) balki gıba ʾōzāli xa bala dʾāsın qıṭla ʾıbba ‘Perhaps she wants to

cause me some misfortune that I may be killed for’ 801

Ex. (220) contains an occurrence of the passive construction in the dependent subjunctive (ʾāsın qıṭla), which is quite comparable to ex. (219) (ūla hāwe qṭīla). There are several differences between the two examples, but the most important and consistent one here is stative or resultative passive in ex. (219) vs. the dynamic passive in ex. (220). There are a handful of examples which denote the perfect with the participle: (221) xšūlu lakūn hāye bırya ʾılle xa maqaddır ‘They thought that (lit.

lest) an accident may have happened to him’ 873

(222) qamāye xšūlu la hāye ʾán mšalxāne mṭušye gyānu ūmpīqe ʾıllu gō

ʾurxa žnīgva ‘First they thought these robbers may have hidden and come out on them all of a sudden’ 892

These examples show the notion of perfect very clearly, when the auxiliary supplies the modal circumstances: the verbal lexeme xšāwa allows a choice between modal and indicative substantive clauses. In both cases here the subjunctive denotes probability, similar to the form pāwe in independent clauses. Note that in asyndetic clauses subjunctive verbal forms tend to occur immediately following the matrix verb (xšūlu in both cases). Unlike the preceding examples, the following one has a subjunctive form, which is obligatory in this construction, and consequently has no modal value: (223) ūzılle basre mux xā ́ dhāwe pīša yaqsir uʾāzıl mın naʿlājūse lakun

ʾāwız ʾ ılle xa maqaddır ‘So he followed him like someone who has become a captive and walks out of necessity, lest (the captor) bring a disaster upon him’ 911

Adjective clauses that describe an generic, non-specific entity consists of a subjunctive form, which is obligatory. The following example shows a subjunctive form in independent clause, expressing the will of the speaker:



(224) bıne bāzın hāwēt xīpa ulwīša xa badle sqılta ūhā wēt tīwa go xa

sanduq ‘Tomorrow I am going, be washed (lit. te sois lavé) and have (lit. aie mis) a suit on and be sitting in a box’ 181

The translation is awkward in English, the point being a near-imperative value of the subjunctive auxiliary and a perfect or resultative value of the verbal lexemes. The last case in the example, hāwēt tīwa, although occurring with the participle, does not convey a perfect or resultative value, as it is not opposed to bītāwa in the same function (that is, with any kind of copular expression). The only case found with the gerund is the following: (225) ya flankas. mġabīne ʾıllox hāwıt bızvāra go barīye ‘O stranger, it

is a pity that you are wandering in the desert’ 950

Here too the subjunctive is obligatory after mġabīne, and the actual present meaning (rather than any modal value) is clear. The form hāwēwa is an important part of the irrealis conditional. It is basically part and parcel of every irrealis protasis: (226) ūkan la hōyāwa ḥakumta, hōyāwa xa baxta ʿāde, gbarya ʾıbēwāla

‘and if she had not been the queen, (if she) had been a plain woman, it is possible (that) he might have wanted her’ 809

The rhemes are substantives in the preceding example, and an adjective in the following: (227) xa hāwēwa naṣax, kīʾēwa ınkan māyes uʾınkan bıṭāreṣ ‘if someone

would have been ill, he would know whether he will die or whether he will heal’ 226

This example has hāwēwa, but it is not an irrealis protasis continued by kāwēwa forms, but rather a characterization of a person, which is generally expressed by kšāqılwa forms. The subjunctive form is a common form in adjective clauses that describes a non-specific entity (see §4.1.4). Here the subjunctive occurs with -wa in agreement with kšāqılwa forms. In the following example, however, all the forms occur with the participle and denote an irrealis perfect: (228) ʾanya pıšlu sabab dīda. ʾanya la hāwewa ʾısye, ʾ asqad nāše la

kāwēwa jmīʿe ʾaxxa, ʾasqad nāše la kāwewa ḥsīde dēni, uʾe brāta la kōyāwa mısta. kulle waʿda kxašwanwa lday xaʾ ‘these people (=visiting relatives) became the cause of her (death). Had they not come, so many people would not have gathered here (to



watch them), so many people would not have envied us, and this girl would not have died’ SAG 7.32 Only the first clause constitutes the protasis, the form la hāwewa ʾısye is the mark of irrealis protases. The la kāwēwa forms are the negative counterpart of pāwēwa, the apodotic form in irrealis conditions.

1.3.3 The presentative copula The presentative copula22 has several syntactic functions: atemporal copula, abstract presentative (§ and a concrete presentative (§ In this section are discussed only its copular functions. The copular functions are like those found in the affirmative present copula, with three major differences: The presentative copula has only 3rd person forms, it is compatible with any rheme type (but most commonly verbal rhemes), and it is atemporal, viz., it is concomitant with the surrounding temporal (or other, e.g., narrative) frame. In the absence of such frame, the form is referential to the point of the utterance. When the rheme is a substantive the syntagm sometimes functions as a concrete presentative (to be discussed below, § (229) wēle ō jwanqa dīd gbēla brātox, rāba dōlamánt īle ūmın mıšpāḥā ́

xšūta ‘There is this youngster who wants your daughter, he is very rich and from a good family’ 544

Other than this, other rheme types work as expected. The first example contrasts two similar syntagms, R-wēle and wēle-R, both with an adjective for rheme: (230) hawwa šıtqe smıxta wēla. garxıt ʾ e wēla smıxta ‘Well, well, she

was pregnant last year. (There) she is pregnant again’ SAG 4.3

The difference (already mentioned briefly above) is that, while the preterite of √hwy has past value, the presentative is concomitant with time of refer-

The diachronic constitution thereof is some element *wa, perhaps still found in the vocative expression wa-lox (vocative ‘youMS’), perhaps also in the far demonstrative ʾ-wā-ha (‘thatMS yonder’), appended to the copula īwın, īwıt (which are older forms), and īle to produce wēle, wēla, etc. The 1st person does not occur in JZ save for relics (hawēwın ‘here I am’ [606]), but it does exist in the Nerwa homilies (Sabar 1984). 22



ence, which is the present. This is not the only option, compare the following pair of examples, both with an adverbial rheme: (231) xmāra dīdi hár wēle ʾ ımmi lá ṭeʾna ‘My donkey is with me

without a load anyway’ 921–922

rāba zaʿif–īle, qēmın šōqınne hīl ṭārıṣ +bāš́ ūxarāye wēle gō ʾīzi, dammıt ʾājıbli baxlınne ‘He is still very weak, let me leave him until he gets fatter, and then he will be in my hand, when I feel like it, I will eat him’ 387

(232) hēš


The presentative copula in ex. (231) is referential to the time of utterance, but the one in ex. (232) is referential to the temporal framework of the adverb xarāye ‘then’ (said with reference to the time of utterance) and the form baxlınne referring to the future. It hence follows that the presentative copula is truly devoid of any tense of its own. The following example has a gerund for a rheme: (233) bale ʾo hozāya dīdi faqīr-īle uwēle bıbxāya ‘but this Jew of mine

is poor, and he is crying’ SYG 2.55

Note the interesting, and quite regular, distribution: the adjective normally goes with the present copula, while the gerund occurs with the presentative copula. The opposition between gerund and perfect participle holds here too: (234) rēši wal ṭrıṣle, bale ʾaw xabrıt mırra baxta hēž wēle bıʾzāla gō nāvi

ūmux xa gurrīyıt nūra bıšʿāla gō lıbbi ‘my head has healed, but the word the wife said is still going through my guts and burning in my heart like a flame of fire’ 937

(235) wēle ʾwīza

SYG 1.12


rāba hawūsa ʾımman ‘He has done us a great favor’

(236) yımmox wēla xdımta baxti ‘Your mother has served my wife’

SYG 2.19

That is, in dialogue, the combination yields respectively an actual present and a present perfect (see §6.3). In the narrative of personal experience narrative, the presentative keeps in line with the story-line: (237) sēli, š -gār maré-bēsa. maré-bēsa wēle bimāra ta baxte: ‘I came to

the roof of the landlord. The landlord was saying to his wife:’ SYG 2.55



(238) pıšle sıswa. ʿabdi wēle zīla lmōṣıl ‘It was winter. ʿabdi had gone

to Mosul’ SYG 2.123

1.3.4 Other copular expressions: pıšle There is another verb, pyāša ‘remain, be(come)’, whose lexeme makes a natural candidate for copular function expressing various shades of Aktionsart which are often absent from the other copular expressions (ingressive, inchoative). The working definition of Aktionsart in this framework refers to the same functional category as aspect (the internal constitution of the action), but in microsyntax, rather than macrosyntax. A representative example for this categorization is the function of the perfect participle, in active function, with various copular expressions: the texteme type by and large does not matter, it basically adds the same nuance (linguistic perfect) wherever it occurs. Aspect will be used in macrosyntax to refer to this functional category when texteme-sensitive, or, in other words, is different between, e.g., dialogue and narrative. This is the reason why also the category of tense is not discussed here, although predicative forms do participate in temporal oppositions; the values of virtually the same form are sometimes very different depending upon the syntactic domain in which they occur. The valency of pyāša is like the other copular expressions, viz., the rheme paradigm. The first and most common function of pıšle, together with the gerund, is to express ingressive action: (239) qımle mxēle lʾurxa ūpıšle bıthāya go barīya ‘So he hit the road

and started wandering in the wilderness’ 529

(240) mpıqle mın tāma ūzılle. pıšle bıthāya hīl mṭēle qam xá dāra rāba

rūmāna “He went out and left. He was wandering until he came to a very tall tree’ 530

Ex. (239) shows the problem very well; the vast majority of the examples of pıšle+gerund are indeed ingressive, and this default value is reflected in the example. Nevertheless, one occasionally finds it functioning somewhat differently, in this case (ex. (240)), it is a bounded continuous narrative tense. This second value is found when we know that the beginning had already taken place: (241) škılli sāxēna bıt day qarsa ūkp īna. xá yarxa pıšli bısxāya rıš dan

māya ‘I began swimming in this cold weather (being) hungry. (For) one month I was swimming in this water’ 373–374



In ex. (241) the beginning of the action is made explicit by škılli ‘I began’. Occurring again almost immediately, the pıšli syntagm is less likely to express ingressivity. Moreover, the time-span adverbial (one month) makes it quite plain it is not an ingressive, but rather, as ex. (240) above, it is a bounded, progressive narrative tense. The auxiliary pıšle bıšqāla may be compared to wēle bıšqāla: (242) ūtūle ūškılle bıd ʾīxāla. āwa pıšle bīxāla ūʾēne nzırra ıl xā čūčıksa

‘so he sat down and began eating. (While) he was eating, his eye caught a glance of a bird’ 222–223

(243) ūtūle ūškılle bıd ʾīxāla. wēle bīxāla ūsēlu ʾılle mšalxāne ‘so he sat

down and began eating. (While) he was eating, there came robbers to him’ 524–525

Exx. (242) and (243) are completely parallel in structure, and the difference between pıšle (non-ingressive, since the beginning is made explicit by škılle) and presentative wēle, both functioning as preposed circumstantial clause is very small indeed. The use of pıšle with an active perfect participle is attested rarely: (244) pıšlu ʾān ṭḷāha jwanqe mšupye ıl malaxīne ‘The three youngsters

became like (lit. resembled) angels’ 556

(245) muqdar xá-saʿʿa pıšla ṭweʾta ‘she was asleep for one hour’ 899 (246) ūpıšlu ḥmīle hādax ‘and they remained standing thus’ 888

Note that the trio become–be–remain is very difficult to delineate according to formal criteria. The neutral notion of being with a non-verbal rheme is confined to certain fixed expressions, where there seems to be no opposition with other copular expressions: (247) qamāye marri, ma pıšle sabab dō ʾīzāla dīdax mnōž gyānax ‘First

tell me, what is the reason for your walking alone’ 860

(248) ūmayla sabab snīqūsox ʾıllı, ʾāhıt ūxūrāsox? ‘and what is the

reason of you and your friends’ needing me?’ 859

Ex. (247) is ambiguous: it could be analyzed as a direct question (as we have in ex. (248)) or as an indirect question, the object of marri. All the other cases where pıšle occurs with sabab ‘reason’ are in indirect questions. In ex. (248) it is a direct question, this time with the copula (and the real



rheme is ma ‘what’). The preterite of √hwy may occur in these circumstances: (249) mani mınnōxūn bʾımṣe ʾāmırri ma wēla sıbbıt dō šūla? Who of you

can tell me what was the reason of this matter?’ 476

(250) ūbamrínnōxūn ma wēla sıbbıttıt qıṭlat danya nāše ‘and I will tell

you what was the reason of killing these men’ 475

The difference between wēla and pıšle (compare ex. (247) with exx. (249) and (250)) is not the time frame: in ex. (248) the present copula expresses the present domain. Rather, it is mechanical, and has to do with the word sabab itself: in the past domain the verb which occurs with sabab is always pıšle, whereas with sıbbe it is always wēla. Other than these special cases the clear notion of neutral being is very rare with pıšle. We are left, then, with the two basic notions, which are in a way opposites, of ‘become’ and ‘remain’. These notions, unlike the ingressive use of pıšle with the gerund, are not confined to the preterite and occur with other tense forms. With substantival rhemes, pıšle basically denotes ingressive being (becoming): (251) ḥakōma mın nıxpūse ūmın qaḥrīte qṭīlāle gıyāne. ō ha bır ḥākīm

šqīlāle brat ḥakōma ūgurre ʾımma ūpıšle āwa ḥakōma go day bāžer ‘The king killed himself out of shame and grief. The son of the healer took the princess and married her and he became the king in that town’ 120

(252) ınkan gıbētūn āna pēšen ṭālōxūn ḥakōma ‘If you wish, I will (or:

let me) become king for you’ 342

Ex. (251) contains the explicit circumstances which compel us to interpret pıšle as ‘became’. Ex. (252) demonstrates this notion in future tense as well. The following pair of examples opposes pıšle and wēle with an adjectival rheme: (253) dammıt kušle dımma rıš daw gōra, ṭ rıṣle pıšle ṣāx ūqımle rıš ʾaqle

‘The moment her blood flowed over the man, he healed, became well and stood on his feet’ 104

́ wēle ‘Until the time you (254) hīl dammıt qam nablıtte ṣāx salīm brought him he was alive and well’ 923



pıšle signals ingressive being, whereas wēle signals a continuous (rather than punctual) being. The notion of ‘remain’ is rare with adjectives: (255) mōnıxle pıšli āna xarāyıt kullu “He looked, I remained the last

one’ 387

(256) zılla upummıt (pımmıt) škafta ġ lıqla. pıšli ḥbīsa tāma hīl ʿāṣırta

‘She went and closed the opening of the cave. I remained imprisoned there until the evening’ 411–412

(257) uʾāni mıtlu upıšli yatumta ‘And they died and I became/remain-

ed an orphan’ 702

(258) uʾapāwa yatūma wēle la bāba ula yımma ‘And he too was an

orphan without parents’ 684

With adverbial rhemes one cannot state a clean distribution: when the rheme is temporal (mbínnōke ‘morning’, ʿāṣırta ‘evening’, palgızlal ‘midnight’ etc.) we have an unmarked, indeterminate (ingressive or continuous) value: (259) pıšla bazıl mbínnōke ‘It became/was morning’ 381 (260) bazıl mbínnōke wēla xūrāse ḥ mīle hīvīte “It was morning, his

friends waiting for him’ 432

The opposition with wēle, whose value is continuous, is clear. With other adverbial syntagms the value seems to be ‘remain’. Locative adverbials are the most common: (261) ūtre parre dīd ṭēra pıšlu go ʾīzıt daw bır ḥakōma ‘And two feathers

of the bird remained in the prince’s hand’ 188

(262) uʾōha gōra pıšle kıslu kma yarxe ʾarxa ‘And that man remained

with them several months as guest’ 224

(263) ō yala pıšle ʾaxxa ‘The child remained there’ 588 (264) zıllu ṭḷāha yōme ūṭḷāha lēlwāsa wāwa pıšle rıš daw dāra ‘Three

days and nights passed, and he remained on the tree’ 236

(265) qam šadrāla ġulamta dīda lšūl bēsa uʾāya pıšla ʾımmıt dō gōra go

ʾōḍıt dmāxa dīda ‘She sent her maid to her housework, and she herself remained with the man in her bedroom’ 483–484

Rarely the notion is nevertheless ‘be’: (266) pıšla mtaxmōne ūlıbba pıšle kıs xawırsa mın xá-ʾāla ūtaxmīne dīda

ūbāžūra wēlu kız gōra, kız daw jwanqa ‘She began thinking but



her heart remained with her friend, on one hand, and her thoughts and longings were with her man, with the youngster’ 855–856 Ex. (266) illustrates some kind of an equation between pıšle and wēle, there being very little diference between the two. (267) baxtunta dēni lubba mhurhırre dammıt pıšla bēn ʾīzāse ‘Our

woman’s heart had impure thoughts when she was between his hands’ 861

Dative: (268) uʾāni mıtlu ... ū kulle māl pıšle ṭāli ‘And they died... and all the

property remained for me’ 702

(269) qam ḥ absīle ūb ēsa kulle pıšle ta do jīran ‘They imprisoned him

and the house remained for the neighbor/the neighbor’s’ 221

(270) kēfu rāba sēle dīd la rʾıšlu ʾıbbu ūdīd pıšle kulle māl dōhūn ta danya

'They were very happy that they did not notice them and that all the property remained for them/theirs’ 270–271

adverbial: (271) ūbaxte uyalunke dīde pıšlu la ʾī xāla ‘And his wife and children

remained without food’ 627

(272) ūzılle. pıšli mnōš gyāni go dē škafta ‘And he left. I remained

alone in the cave’ 388

To sum up, precisely circumscribing the specific value of pıšle is not always possible, but its being the copular expression specialized to express various shades of Aktionsart, which is not texteme-sensitive, is certainly its first and most important function.

1.4 DEPENDENT NEXUS Nexus may also occur in reduced form, that is, without the exponent otherwise supporting it, such as the copula or a verbal form. This basically happens as part of the clause, when functioning as object, adverbially (as ḥāl, that is, circumstantial expression) and probably as an adnominal nexus too. An abundance of examples from the literary Urmi dialect of all these types are brought and discussed in Polotsky 1996:23–29, collectively referred to



as “non-finite predicative group”. Similar cases are described in Khan (2008b:849-855) and regarded as omission of the verb ‘to be’.

1.4.1 Object nexus Object nexus is a term borrowed from Jespersen (1924:122–126), referring to a non-clausal nexal object (that is, the “object” construction basically does not constitute a full clause), occurring most commonly with perception verbs. In this section only the nexus interests us, i.e., the internal syntax of this complex object. The external syntax is to be resumed below, under §3.1.3. (273) mōrımle ġulāma sēpıt gyāne ūqam qāṭılle. tʾırre (rʾıšle) bır ḥakōma

ūġzēle ġulāma murma sēpa mın ʾılʾḗl rēše ‘The servant lifted his own sword and struck him. The prince woke up and saw the servant lifting (lit. [having] lifted) a sword above his head’ 95

The two entities ġulāma ‘servant’ and murma ‘lifted’ (active perfect participle) form a non-clausal dependent nexus. This analysis follows two principles; first, that the verbal lexeme ‘see/find’ takes but one object (unlike, e.g., verbs of giving, which habitually take two objects) and second, that this object can be put into the form of a clause (in our case: he saw that the servant was lifting a sword’). The object nexus is often substitutable with such an object clause. Two directions can be taken to show that the relationship between two entities is indeed predicative: (274) kxāze, ma kxāze, bēsa spīqa ūsrīqa ūbaxtāse kutru zīle mbēsa

‘He sees, what does he see? The house empty and both his wives gone from the house’ 844

In this case, it cannot be claimed that he sees his wives; what he sees is that they have left the house, they are gone. (275) kxāzēwa xa līra naqōṣa mın pāre dīde ‘he used to find one coin

missing from his money’ 1

It is clear that the coin itself was not found at all, since it was missing. Moreover, it is not a missing coin that is found but rather (the fact or situation) that the coin was missing. The following example shows the alternative construction: (276) ūʾay dammıt kxāzēwa ḥakōma dīd wēle nāqōṣa xa quṭeʾfa... ‘And

whenever the king used to see that one cluster was missing...’ 641–642



The other way of looking at these constructions is comparing active with passive perception: (277) kxāzēla gyāne go namūsīye dīde ugo bēs gyāne ‘He sees himself

in his bed and in his own home’ 568

(278) dammıt muxwēla gō nāše... ‘When she was seen among

people...’ 840

Ex. (277) features our object nexus construction, where a nexus exists between gyāne ‘himself’ and the prepositional syntagm go namūsīye dīde ‘in his bed’. In ex. (278), on the other hand, we find another verbal lexeme, ‘to be seen, seem’. It is not formally the passive form of ‘see’, but it does function as such.23 This lexeme, together with others (appear, become etc.) all function as semi-predicatives, and sometimes show a special distinction: (279) rāba spahīn ūbāš gmaxūya ‘She looks very handsome and

good’ 543–544

(280) rāba spahīn wēla ‘She was very handsome’ 778

Exx. (279) and (280) are quite similar, predicating a quality of a woman. Ex. (279) reflects the more subjective character of this predication. The same predicative relationship, as habitually found with maxwōye ‘seem’ (exx. (278) and (279)) is found in the active expression of xzāya ‘see’, between the two parts of the object nexus (exx. (275) and (277)). Object nexus constructions consist of two members: one theme-like, the second a rheme. The paradigmatic constitution of the theme-like paradigm is similar to the thematic function discussed above, however, it may contain indefinite substantives as well as pronouns. The object nexus itself often constitutes a thetic expression, which contains all new information, and is hence unanalyzable into theme and rheme (and see below, §§1.5.2–1.5.3). Unlike the regular thematic function, here even object pronouns can have thematic function—the reflexive object pronoun gyān- (see above, ex. (277)), as well as suffixed object pronouns: (281) qam xāzē-le bīzāla ‘they saw him going’ 494

The rhematic slot is very similar to any rhematic slot in an independent clause (to be discussed below, §1.5)—prepositional syntagms, adjectives, This curious use is found in several NENA dialects. The form is also used in its original function, namely, as a causative of the lexeme ‘see’, namely. ‘show’. 23



perfect participles and gerunds, the difference is that no substantives are attested so far in this slot. As for a finite verbal form, there are a few candidates, but they are ambiguous: ́ ġzēli xá-baxta xēt gʾōrāwa pkōlāna dēni ‘Whenever I saw (282) čımmād another woman passing in our street’ 829

The potential object nexus can equally be analyzed as an asyndetic object clause, ‘I saw {another woman was passing...}’.

1.4.2 Adjunct nexus This type of dependent nexus immediately follows a substantive, describing an inherent feature thereof. Compare the following pair of examples: (283) mpıqle mın xá-ʾāl xa ʾarya ʾēne smōqe ‘From one side came out a

lion, his eyes red’ 908

(284) mpıqle xa ʾarya mare xakma ʾēne smōqe ‘There came out a lion

(possessor) of some red eyes’ A 1.1

Both examples occur in parallel stories, but whereas ex. (283) contains our construction, where ‘red’ is predicated on ‘his eyes’, ex. (284) has another strategy, where a general nucleus (mare ‘possessor of’, see §4.1.4) is attributed by the syntagm ‘red eyes’; here we have no predication. The following examples are in the same vein: (285) xa yōma musēlu tre jandurme dīda xa jwanqa bōle murpıya

ūkwīša basır kapāne ūdıqne yrıqta ū la yrāqa ‘One day her gendarmes brought a young man, his hair loose and falling behind his shoulders, and his beard barely sprouted’ 789

(286) ... ūqṭıḷḷa gyāna rıš xá ʿ abd kōma mux šaxōra, sıppāse garūse

ūmšúmbıle ‘... and she killed herself over some slave black as charcoal, his lips full and mustached’ 807

(287) basya ʾurīsa xā xızya mıḷya xalwa bāš ūxā ́ xızya wēle marūʾa

umʾobya ‘A lioness will come, one of her teats full of good milk, and one teat of her will be painful and swollen’ 76

Ex. (287) is different from the rest in two points; first, the thematic substantive is preceded by the quantifier xa, and second, the dependent nexus is followed by a presentative expression (to be discussed below, §1.5.2). The presentative clause and the construction in question create a sort of equation—two types of nexus, one after the other.



This type is not common, and seems to be heavily restricted: 1. the described entity is indefinite; 2. the thematic entity inside the dependent nexus is an inalienable part of the body, with a pronominal suffix referring back to the described entity; 3. the rhematic entity is made up of an adjective or a perfect participle; 4. the nexus immediately follows the nominal group, taking the place kept for adjectival syntagms. This is an adjectival dependent nexus. The following example is somewhat different: ́ ś (sic) dīde qčīn (288) mpıqle mın xe daw qaṭra xa xūwe kōma xlīmū nāfṭangıt xá ġalāma ‘From under that rock came out a black snake, its width like a boy’s loins’ 440-441

The snake in ex. (288) is doubly qualified: first, by the adjective kōma ‘black’, and then by the dependent nexus where the size (‘like a boy’s loins’) is predicated on its width. Note that, semantically, the width of the snake has nothing to do with the current situation; it is rather a constant feature. However, the width is an exception in this group, as it is marked syntactically by dīde ‘his’, rather than by a suffixed form, as all the rest. Additionally, whereas the rest of the examples have either a perfect participle or an adjective for rheme, ex. (288) has a prepositional syntagm. One exception to the characteristics of this structure is the following example: (289) ṓha nāša īlāha mburxa ūmpuṣna šımme qam pāsıxlu nısyāsıt dō

gōra ūšmeʾle ūfhımle ‘God, blessed and praised be his name, opened this man’s ears and he heard and understood’ 279

The differences are that the described entity is definite (īlāha) and the thematic entity is the substantive šımma ‘name’, which may however belong with inalienable body parts. One last thing—the syntagm is commonly understood as a wish (praised be…’), which then makes it impossible as an inherent quality. At any rate, the syntactic association is clear. As for the relationship with adjective clauses, take a look at the two last examples: (290) axni waxni ḥ mīle hīvīt xā gōra dīd ʾaqle wēlu go xā qōqıt ʾıpra

‘We are waiting for a man whose feet were in a clay pot’ 55

(291) ō nāša dwēla kāse ṣrīsa ‘The man whose belly was slit...’ 959

These example predicate information on an inalienable part of the body within an adjective clause; however, the described entity is definite, and, more important, these two clauses describe non-inherent qualities. Other



than that, as far as I can tell, there are no other examples of adjective clauses with an inalienable part of the body as theme, which means that our construction and adjective clauses are in some kind of complementary distribution in the conditions stated above.

1.4.3 Subjunct nexus This section deals with dependent nexus in circumstantial function. The external syntax of this function is to be discussed below, §3.2.1. This dependent nexus is characterized, as the one discussed above, by two parts, one basically being the theme, the other the rheme, together providing the circumstances for the main clause. The syntactic link with one of the actants in the main clause is mostly obtained by a pronominal entity (not necessarily a genitive one, see below, exx. (299) and (300)): (292) ūʾanya nāše kēsēwa lbēsa, lıbbu bassīma ūrıẓye pḥaqqu ‘These

people were coming home, (with) their heart content and satisfied with their portion’ 449

In the dependent nexus lıbbu bassīma, lıbbu is the theme and bassīma the rheme. The nexus refers back to nāše by the genitive pronoun -u suffixed to lıbbu ‘their heart’. There is another form in circumstantial function, the adjective form rıẓye. The difference is obvious, as rıẓye does not form a nexus, but as a circumstantial it refers back nāše by its plural ending -e. (293) yıʾılle ū mpıqle ūʾaqle ṭrīse ‘He came in and out (with) his feet

healed’ 54

Note that, compared to the preceding type, this nexus does not modify a substantive, but rather refers to the situation, and in syntactic terms, to the main nexus. ‘with his feet healed’ describes adverbially his coming out. As is clear from ex. (293), the nexus in question is not introduced by any specialized exponent, but the (otherwise connective) u- is often found at the head of the nexus, as in exx. (294) and (295): (294) qımle ž gyāne uhādax rēše bızvāra uʾēne mdamōʾe ūbehne kīta

‘He stood up, his head spinning, his eye(s) shedding tears, his breath dried out’ 761

Ex. (294) contains three such nexus constructions, each introduced by u-. The first u-, seemingly connecting the main clause with the first dependent nexus, is actually not a connective, since connectives interconnect entities of the same function, and it is obviously not the case here. The main clause is



an independent clause, whereas the nexus under discussion is dependent, functioning as part of the main clause, as is well testified to by the absence of the copula between its theme and rheme. Other possible introducing exponents are the adverb hādax (otherwise ‘thus’), which is very common with a circumstantial function: (295) ūthēla gō daw čōl hādax lıbba bıčrāpa mızdoʾsa ‘She wandered

in that desert, (with) her heart terrified with fear’ 901

hādax here seems as if it introduces the dependent nexus, but comparing this occurrence to ex. (294) above, we see that hādax and u co-occur, leading us to think that they do not belong to the same category. A third (rare) element introducing circumstantials is the preposition bıd ‘with’ (in this case): (296) kullu gımšadrālu bıd rēšu kīpa (múkūša) ūnxīpe ıl bēs gyānu ‘She

sends them all back home with their head bent (down) and ashamedPL’ 684

(297) qbılle zērıngır ūhul le kmāḍ ṭḷ ıble mare nunīsa ūšqīlāle nunīsa bıt

kēfe ʾısya ūzılle lbēsa ‘The goldsmith agreed and paid what the owner of the fish asked and took it happily (lit. with his joy come [=perfect participle]) and went home’ 727–728

This is not a regular prepositional syntagm containing a substantive and an adjective: note that the substantive, as in many other cases of this dependent nexus, has a genitive pronoun referring to one of the actants, and the second part is the perfect participle. The theme can be a definite substantive,24 not necessarily an inalienable body part, as well as a pronoun: (298) ūpıšlu mēnōxe bıd dō naša hādax tīwa rıš xa taxta ūṣīwa lbırke

‘They started looking at this man sitting on a chair, with the (piece of) wood on his knee’ 510

ṣīwa is not a body part, but serves just as well as the theme of the dependent nexus; note that the reference to this man is on the rheme ‘his knee’ (as it is in ex. (305)). In the following pair of examples, there is no special pronoun Except one case, which is an existential (where we might have no theme at all): mpıqle garxet aw xūwe ūppumme xa dehwa ‘The snake came out once more with’ 4 a coin in his mouth (443). 24



to create this link back; it is the pronominal theme that is enough to create this reference: (299) pıšla maxmōxe ūbınšāqa dīde ūbıšyāpa pāsa bıt pāse ūqzāle ūʾāwa

ṭwīʾa xa šınsa qwīsa (yāqurta) ‘She started sniffing and kissing him and rubbing her face against his face and his neck (while) he (has) fallen asleep deeply (lit. he fallen asleep one strong sleep)’ 870–871

(300) qam šaʿwıṭle lıbbi bıd marʾıt dūmıki uʾāna bīhāwa ṭāle jawāhar

mqabıl māxēle narʾa ıl dumıki bıt xá daqīqa ‘He burned my heart by the ache of my tail, with me giving him a jewel (lit. I giving him) just one minute before he struck my tail with an axe’ 637– 638

As for the rhematic group, it is somewhat more flexible than than that of the adjunct nexus described above, perhaps due to the larger number of examples. One finds adjectival as well as adverbial syntagms, but perfect participles and gerunds are the most common: (301) ʾay dammıt gmāxḗwāle jaʾōza čār nıkāŕ ṣīwa hādax múḥumla rıž

bırke ūʾaw jaʾōza hādax xarūpa muġ nūra ‘Whenever he would strike with the chopper around the (piece of) wood thus rested on his knee, (when) the chopper is sharp just as fire’ 509-510 (adjective)

(302) wēlu tīwe brēza ... ūkutxa ʾāman dīde qāme ‘They are sitting in a

row ... each one his plate in front of him’ 939 (prepositional syntagm)

(303) uʾōqat ġzēle, ʾarya sēle ū rēše la ysāra ‘Hardly did he look than

the lion came and his head without bandages’ 682 (adverbial syntagm)

(304) mpılla dıbba lʾarʾa hēš ʾīza duqta bıt bābıt dō yāla. ‘The she-bear

fell to the ground, her hand still holding the boy’s father’ 752 (active perfect participle)

(305) zılle ʾarya dēni burxe qḥīra ūḅarḅāṭıt nūra bıfyāra mın ʾēne ‘The

lion went away sad and sparks of fire flying out of his eye(s)’ 936 (gerund)

Polotsky (1996:23 n. 32) defines finiteness as involving person and tense, ascribing these to the functions of the copula. It is true that these cases of



dependent nexus have no temporal reference of their own, and it is also true that the rhemes, unlike verbal forms, do not show person. All these circumstantial dependent nexuses follow the main verbal form (not necessarily immediately). Their dependency is clear beyond any doubt because they occur without a copula.

1.5 COMPLEX RHEMATIC GROUPS This section is devoted to rhematic groups, that is, to various complex groups which are judged rhematic in their entirety. Within this group one can isolate three subgroups: existential expressions, presentatives, and other thetic expressions.

1.5.1 Existential expressions Existential proper The specialized existential expressions (ıswa ‘there was’, īs ‘there is’) occur in two basic versions—story-initial and story-internal. These are special expressions where an entity is intoduced for the first time into the story, in a way somewhat similar to presentative expressions (§1.5.2) and other thetic expressions (§1.5.3). Existentials are also expressed by the various copulas. Story-internal existential expression25 are characterized by an exponent (ıswa ‘there was’) and an existant (xa rūvīka ūbaxte ‘a fox and his wife’ 460). More often than not, an adverbial, usually locative in nature, is present as well, whose location within the expression is rather flexible. The existant is never explicitly definite, which means it is all-new in the discourse, and therefore judged to be rhematic. In addition to introducing a new entity into discourse, in about eighty percent of the cases one also finds an expansion of the existant. The exact syntactic nature of this expansion varies, but in general it is adjectival in nature—half of the expansions consist of an adjective, a fifth of them are adjectival dīd syntagms (containing either a substantive, a pronoun or a clause, the latter being an adjective clause), and the rest are clauses analyzed as asyndetic adjective clauses. These adjectival expansions are rhematic as well:26 Story-initial existential expressions are discussed in Schwartzbart 2008. Asyndetic adjective clauses are rare in JZ, occurring, by and large, after indefinite antecedents. See §2.3.1. 25 26



(306) ūgo muġrub ıswa xa jwanqa ham muxwāsa sqīla ūmaʿqul ‘In the

west there was a young man as handsome and smart as her’ 304

(307) ımmıt day karwan ıswa rāba māl ūsūswāsa múṭʾıne rāba īxāla

‘With this caravan there were lots of money and horses loaded with lots of food’ 268

(308) ıswa tre axawāsa dīd mpíqwālu mın bēsohūn ‘There were two

brothers who had left their home...’ 154

(309) gō karma ʾıswa xa nāša gnāṭırwa ʾılle lēle uyōma ‘In the vineyard

there was a man (who) used to guard it night and day’ 772

All these four examples have a complex rheme, then, comprising both the existant and the information packaged as an adjective, whether simple or complex. As regards the latter, ex. (306) contains a simple adjective; ex. (307) an adjective with an adverbial expansion; ex. (308) a syndetic relative clause introduced by dīd; and ex. (309) an asyndetic relative clause. Those examples represent the entire inventory. The fact that the expansion of the existant can only be adjectival in nature sets this group apart from, e.g., presentatives, where one can find a much fuller paradigm, similar to the paradigms occurring with copular clauses. This adjectival group together with the existant preceding it are both considered rhematic, containing allnew information. A curious case is when existentials do not introduce a new referent into the narrative, but rather into the dialogue; these cases are marked by ʾīs ‘there is’ and lēs ‘there is no’, laswa ‘there was no’, mostly followed by an indefinite entity. This indefiniteness is apparent from the indefinite article (e.g., ču ‘no’) or from article ø. In other cases the introduced existant looks formally definite—lēs ġēra ‘there is no other (lit. her other)’ or lēs muxwāsa ‘there is no one like her’, but in fact refers to a non-specific entity. In case of an affirmative existant, it seems that the article is rather ø and the substantive is actually preceded by a quantifier, when countable: īs xá mıllıtta ‘there is one nation’; īs +rāba fıhım ‘there is a lot of understanding’ (noncountable); ʾīs trḗ ʾurxāsa ‘there are two roads’. In these cases it is better to analyze a quantifier rather than an article for two reasons: 1. the element xa is stronger prosodically (lengthened, accented or both), when occurring as a quantifier. 2. more than one item takes an explicit quantifier.



The existant can either follow or precede a locative adverbial: īs axxa xā baxta ‘there is one woman here’, īs gāwe xā mġāra ‘there is a cave in it’. Other existential and existential-like expressions In addition to the specialized existential expressions (such as īs, ıswa and their negative counterparts) one comes across other related expressions: ībe (lit. ‘there is in it’), ıswābe (lit. ‘there was in it’) and their negative counterparts, and the possessive existentials ıtl- (lit. ‘there is to x’) and ıswāl- (lit. ‘there was to x’) and their negative counterparts. These expressions are historically related, all containing the historical core ʾīṯ with various additions: (la) ʾīṯ









+ +

+ +




lēs …

‘there is no …’

existential expression

laswa …

‘there was no …’

ʾīs …


‘there is … in x’


ʾıswāb- …

‘there was … in x’


latl- …

‘x doesn’t have …’

lēb- …

ʾıtl- …



‘there was …’

ʾīb- …



‘there is …’


+ +


ʾıswa …

+ +


(hwā) (prep.)

ʾıswāl- …


laswāl- …

‘there isn’t … in x’ ‘x has …’

‘x had …’

existential (or ability) predicative possession

‘x did not have …’

The construction laswāb- is attested only once as existential, in ex. (310). Its other occurrences are in fact expressions of ability (§ The following table summarizes the various existant expansions attested with the different existentials discussed below: exp.




























+ +

+ +

+ +


def S





S preceding



+ +



The existential expression with b- is externally similar to the basic expression īs/ıswa, except it occurs with that locative element (which is otherwise optional with existentials) and an anaphoric genitive pronoun: (310) ūqam nābılla qam mandēla go čōl go xa bıhra dīd láswaba māya,

ūsēle ḷbēsa. ıswa go day bıhra xā nuqba ıltḗx bıd gūda dīd íswābe xā xūwe ‘So he took her and threw her in the desert in a well which had no water (lit. there was no water in it). Down the well there was a cavity in the wall where there was (lit. that in it there was) a snake’ 24–25

These b- existentials27 always occur with a substantival existant while the rest of the information (which is not obligatory, but mostly occurs) is packaged, like the basic forms, as an adjectival unit or as a genitive attribute: (311) gıba xazya ēma darga ībe ṭīna ‘She wanted to see which door has

mud on’ 165

(312) ubabēni ıtle xa gınsa dīd ība xā dārıt xabūše “And our father has

a garden in which there is an apple tree (lit. which there is an apple tree in it)’ 197

(313) bale ʾībe xa mıʿūz rūwa dīd gımjangır ūkkā yım ‘But there is a

great disadvantage in it, that it rusts and becomes black’ 466

(314) ūġzēle gāwa xā menōra dīd dehwa ūʾíswāba šōʾa š amʿe dīd dūšıt

dıbbōre ‘And he saw in it a lamp of gold and it had (there were in it) seven candles of bee-honey’ 112

No adjectival verbal dīd clauses are found expanding such an existant (the dīd clause in ex. (313) is not adjective but substantive clause). In addition to the basic temporal difference found between ībe and ıswābe (denoting in general past and present) one finds, albeit rarely, verbal forms which supply future by a subjunctive form of √hwy: (315) ūnablūle go xa barīya dīd la hayēba la ʿıns wala jıns, la ʾīxāla ūla

māya. zıllu anya ʾısra gūre ... ūqam nablīle go xa barīya dlēs tāma la ʿ ıns ula jıns, la ʾī xāla ūla štāya ‘...and bring him to a desert which does not have (lit. where in it there is not) a living creature, neither food nor drink. These ten men went ... and brought

Note that a superficially identical expression but with different valency is used to convey ability. 27



him to a desert where there was not a living creature, neither food nor drink’ 131–132 Note the parallel clauses, containing once a subjunctive and once the negative existential lēs.

Another example is ambiguously interpreted as existence as well:

(316) ūgbāxēwa ta bābe ū yımme ʾan sāwōne dīd wēla ruḥāyu ʾımme ux-

šūle mın basır mınne lákāyēbu xāye rıž dunye ‘And he would weep for his father and mother whose soul was with him and thought (that) after him (gone) they will not have life in the world (OR will not be able to live...)’ 587–588

This ambiguity stems in the fact that all these expressions, when complemented by a subjunctive form, are a suppletive form of the verbal lexeme mṣāya ‘to be able’ (see § Since ex. (316) has the form xāye which is intepretable as either ‘life’ (an existant) or as subj. 3PL. of xyāya ‘to live’. A definite existant rarely occurs: (317) rāba sqīla ūjındāya-le ūkšākıl ta ḥukūm dīd gyāne. bale xapča rīxa

kēse mın pumme ḥēf ʾılle ūʾıl maʿqulūsa dīde dīd ʾībe ʾō nuqṣāni ‘He is very pretty and kind and fits his own authority. However, a little smell comes out of his mouth. Too bad for him and to his nobleness that he has (lit. that there is in him) this shortcoming’ 678–679

Another occurrence of a definite existant is in the expression ībe tēr-x ‘there is enough for x’: ūkud yōm baxlēna mın danya ʾurwe (ʾırwe), ību tēri ‘and everyday I will eat from these sheep, they will suffice for me’ (lit. ‘there is my fill in them’ 388). This expression always occurs with these genitive pronouns. The existential expression with l- is used as the normal expression of possession. Like ībe it has negative forms of both present and past and, in addition, suppletive forms with the verb hwāya. The existant paradigm is somewhat more like the one found with the basic existentials—a substantive with an adjectival expansion: (318) ıtli xa tawırta ‘I have a cow’ 243 (319) kud ıtle pısır gumla, ṭlōb xapča bıd xa ḥu šta ‘Anyone who has

camel-meat, ask for some with some excuse’ 162

(320) ʾaxxa ʾıtlox mād ʿājıble lıbbox ‘Here you have what(ever) pleases

you’ 795



(321) mírrūli: ʾıtlan ō maḥfūra ūʾō ʿaṣa uʾē kaffīya ‘They told me: we

have this rug, and this stick and this scarf’ 426

The preceding examples have various forms of an unqualified existant: ex. (318) a simple substantive; ex. (319) a genitive construction (in which the nucleic substantive is qualified, but not by an adjectival syntagm); and ex. (320) has a substantival clause. Ex. (321) is unique in having a definite existant, which perhaps is no existant at all (although, as far as the context is concerned, there is no textual justification for the definiteness of these substantives, since they are presented to the addressee for the first time). The following pair of examples have a quantified substantive and the quantifier pronoun as the existants: (322) āna ū baxti ıtlan ıčʾa yalunke ‘Me and my wife, we have nine

children’ 49

(323) āna ıtli ısri ūtre ‘I have twenty two’ 47

In this case the ‘existant’ is no doubt referential. The existant can also have a substantival apposition as its expansion: (324) xā xazāla ıtla xā brōna banī-ādam ‘a gazelle (who) has a human

son’ 82

The following examples all have various types of adjectival expansion: (325) wal ʾıtle tre baxtāsa xēta ‘he has two other women’ 815 (326) āna ıtli xā dıkkāna dīd bdāʿa ‘I have a store of merchandise’ 2 (327) mírrēla ıtli arbi ūxá ʾōda mıḷye kullu šıkıl māl ‘He told her: I have

fourty one rooms filled (with) all kinds of property’ 173

Ex. (325) has a simple adjective, ex. (326) has a dīd syntagm and ex. (327) a participle. The past/narrative form ıswāl- has more or less the same inventory, but additionally it has an asyndetic adjective clause expanding the existant: (328) uʾē baxta ʾíswāla ḥıjūlkıt dehwa ktaqlīwa kutru qarwāwıt xá ḥuqqa

‘And this woman had gold anklet(s) (which) used to weigh both almost one ḥuqqa’ 835

(329) wal zwīnāli ūláswāli xā ́ mēsḗlāli ‘I have bought it but did not

have anyone (who) would bring it to me’ 917



The asyndetic adjective clauses occur in JZ basically following indefinite entities, and since the existant is generally indefinite, these clauses are rather common here. Another issue is the order of elements. In the examples presented so far, the existant follows the existential expression. There are a few exceptions, illustrated by the following examples: (330) ōha gōra ūbaxte rāba pāre íswālu ‘This man and his wife had lots

of money’ 331

(331) āna xá yōma rāba pāre íswāli, ō qadda íswāli, dīd íswābi kullu

anya čōle ū kulla ēha bāžer zōnínwāla ‘I once had lots of money, so much did I have, that I could have bought all these deserts and all this town’ 291

Clearly, the preposed existant has more syntactic emphasis, namely, it is almost focal, which is generally treated under §6.1.3. Another order issue comes up when the existant is expanded by a perfect or passive participle. The first example shows the common order, existant—perfect participle: (332) anya nāše íswālu xa dūka xpırta go barīya rāba raḥūqa mın bāžır

́ dōhūn qam xaprīwāla ay dūka ʾalpá draʾe xe ʾarʾa ‘These men had a place dug in the desert very far from their town, they had dug this place a thousand yards underground’ 462

The following example shows the other order, perfect participle—existant: (333) íswālu mjumʿe

rāba ʿıdde +urāba qıṭʾe dīd šuʾāle ‘They had collected many tools and pieces of things’ 462 +

The difference seems to do with the relative degree of similarity to verbal expressions; the opposite order seems more verbal than truly existential: (334) pıšle dmīxa qam daw qōra. go daw qōra ıswa qwīre trē xaswāsa

‘He remained lying in fornt of this tomb. In that tomb were buried two sisters’ 60

Another related example is the following: (335) ʾurru go knıšta, čuxa lēs ʾısya, lēwu jmīʿe nāše ‘They entered the

synagogue, nobody has come, people had not (yet) assembled’ MA 7.4

Besides the existentials which are based on īs, one also comes across a couple of patterns, including the presentative copula, which denote exist-



ence as well. These patterns are special in that they mark a substantive as the rheme, and have, additionally, a locative expression. The most important pattern is #wēle-RSUBST+locative#: (336) mōnıxli wēle xa dāra go daw palgūś baḥar ‘I looked (and lo) there

is a tree in the middle of the sea’ 37

(337) ġzēli wēla xa škafta rapsa barqūli ‘I saw (and lo) there was a big

cave ahead of me’ 396

(338) hēš wēle xapča dımma ʾıbbu ‘There was still a little bit of blood

on them’ SYG 2.84

(339) wēlu +tmāne čarxīye lbēsa ‘There are eighty charxis at home’ SYG


(340) xa wēle xa-sēpa bʾīze ‘One had (lit. there was) a sword in his

hand’ SYG 3.20

The important thing is the fact that the substantive analyzed as existential in these constructions is marked as rheme, which can be regarded as evidence of its being an existential. A similar case is sporadically observed with R-wēle: (341) rāba kēf wēle go daʿwa dōhun ‘There was a lot of happiness in

their wedding’ 547

A few cases, which seem to belong to the same pattern, are nevertheless different, because the substantive is definite (either explicitly or by context): (342) wēle yāla bıd ʾurxa ‘The child is on the road’ 584 (343) ūʾatta wēlu parre dīde ımmi ‘And now I have his money with me’


(344) wēle ō maxzan dīdi axxa ‘I have this storage room of mine here’


The stipulation, then, is that the existant needs to be indefinite. A rare alternative pattern is #locative+wēle-RSUBST#: (345) bıd daw kōlānıt lēʾal wēle qaṣrıt wazīrıt ḥakō ma ‘In that street

above there was the palace of the king’s vizier’ 481

(346) qam pasxālu ʾan tre čanṭát ūkxazya—ma kxazya—go xá č anṭa

wēlu xakma xūwıwāsa ʾwīze mın dehwa ‘She opened the two bags



and sees — what does she see — in one bag there were a few snakes made of’ 204 The following example might belong with either patterns since there are two locative syntagms on each side of the wēle-R construction: (347) ōha gōra fāhımwa blīšāna dēnı, āzılwa atta go ʾōdıt bābe wēla xā

kāwa go gūdıt yamme, ība xā čūčıksa ‘Had this man understood our language, he would have gone now, in his father’s room there is a hole in the righthand wall, there is a sparrow in it’ 102

1.5.2 Presentative function We distinguish two presentation types, viz., an abstract one, introducing an entity into the discourse, and a concrete one, introducing something into the actual situation common to the speaker and the addressee inside the text. These two types are mostly distinct in JZ. Abstract presentative The abstract presentative function is not idiomatic to modern English (when all else fails, the archaic ‘lo’ or ‘behold’ is used to express this), but it is occasionally an important function in other languages, for instance, voilà in French,28 wəhinne in Biblical hebrew.29 To be able to consider this function of presentative wēle in JZ, one can look into a special narrative construction, containing the structure kxāze—ma kxāze(?) ‘he sees—what does he see(?)’. This structure is the intersection of two cardinal points—answers and complements of a perception verb. Both points can be isolated, as they occur separately as well. The question—answer mechanism identifies the rheme— the minimum answer is always rhematic. That in fact is the function of various ‘narrative questions’—to rhematize what follows, be it a part of a clause or an entire clause (see §5.1.3). The second point could be examined separately investigating the various complements of perception verbs (see §, which are often nexal in some way (see §1.4.1 above). From the À peine suis-je dans la rue, voilà un violent orage qui éclate (Alphonse Daudet, Lettres de mon moulin, cited in Lambrecht 2000:63. 29 wattāḇōʾ ʾelāyw hay-yōnā lə-ʿeṯ ʿɛrɛḇ wəhinnēh ʿăle-zayiṯ ṭārāp̄ bə-p̄ īhā ‘And the dove came in to him in the evening; and, lo, in her mouth was an olive leaf pluckt off’ (Gen. 8:11). 28



two points put together we get a complex rheme. This is conceived as true for any occurrence of the presentative in its special environments— introducing new information consisting of an entire nexus. In the majority of cases of kxāze—ma kxāze(?), there follows a presentative construction. In a small number of cases similar constructions follow: an existential, a non-verbal expression of surprise, a nexal resumption of an entire incident and an object nexus, which could be accounted for as surprise or something similar. Besides this mechanism, the presentative function is common after perception verbs in general30. It also occurs in other circumstances: (348) wēla bıt šātıt šōʾa wēla tūta xá yōma xa rēše marōḥe dīde ūbıbxāya

mux kullu yōmāsa. xá bēna kšamʾa xáʾ wēle bıdyāqa tarʾa ‘(And lo) in the seventh year, she was sitting near him fanning him and crying like everyday. Suddenly she hears (lo,) someone was knocking at the door’ 687

The first wēla introduces an adverbial syntagm. This use is found with one informant (S). This occurs occasionally with more complex adverbials as well, e.g., uwēla māt ṭpēla ʾıbbe kšıfle ṭāla ‘so much did she nag him, he (finally) told her’ (673). Note that the second case does not reflect any presentation, and is rather copular in nature (and more specifically, circumstantial to what follows). The third case occurs after a perception verb. It is important to note that in all these cases the presentative clause is never subordinate. The following examples have presentative function independently of perception verbs: (349) ūqam maḥmılla kud qazāne xá ʾāla ū mxēle ʾılla xakma fīšake

ūxrāye qam māšēla dūkıt fīšaka ūqazāne wēla muġ gyāna ‘and he placed each pot at one side and shot several bullets at it and then wiped off the place that the bullet (hit) and lo, the pot is as it used to be’ 469–470

(350) go daw qōra ıswa qwīre tre xaswāsa. wēle bıšmāʾa xā xāsa wal

gımra ta xāsa xēta: ‘In that grave there were buried two sisters. Suddenly he is hearing one sister saying to the other sister’ 60

Very similar in fact to Biblical Hebrew, where one finds the presentative particle wəhinnēh after perception verbs. 30



(351) muxwēlu asıqyāsa dōhūn xa ta dō xet ūġzēlu wēlu mxulpe. asıqsıt

zīne wēla bıt ʾīẓ mamo ūʾasıqsıt mamo wēla bıt ʾīẓ zīne ‘They showed their rings to each other and saw (lo,) they were exchanged. (All of a sudden) Zine’s ring is on Mamo’s hand and Mamo’s ring is on Zine’s hand’ 323–324

In ex. (349) the presentative introduces a surprising result as perceived by those who look at the pot. In ex. (350) again, a somewhat surprising occurrence is introduced. In ex. (351), the second and third presentatives introduce the outcome of the exchange, exactly as seen by the characters. Note that the presentative copula joins both verbal (exx. (348) and (350)) and non verbal rhemes (exx. (348), (349) and (351)). The narrative function of these presentative expressions is discussed under macrosyntax (§§ and 5.4.2). As in its copular function, the presentative copula is not referential to any fixed time, and borrows its time-frame from the environment. In the narrative, it is the narrative time, in dialogue it could be any time. In other words it could be viewed as atemporal or concomitant. This concomitance is often translated, in narrative, into some kind of a present tense in the target language. Concrete presentative The concrete presentative is used to introduce an object or a fact (rather than an occurrence) into the actual situation. It is marked by wēle as well as by a specialized means, hawēle: (352) ya bābi, wēle musēli xa dıhwa ḥaqqıt zmāra dīdi ʾıdyo ‘Father,

here, I brought one gold coin, the price of my singing today’ 450

Indeed a finite verbal form follows the form wēle, but it seems that the gold coin, or its new existence in the actual situation is the center of such presentation (note the agreement of wēle with it). The following pair of examples has the same parallel structure: (353) wēle šulḥan dīdox uhwḗlēle xa šulḥan xēta, uwēlu kutru šulḥāne

‘Here is your table, and another table was born to it, so here are the two tables’ 213

(354) hawēla hāwan dīdox uhwḗlēla xa xet uwēlu kutru hāwāne dīdox

‘Here is your mortar, and another (mortar) was born to it, so here are both your mortars’ 214



In the second example hawēla occurs instead of wēlu. The former is more common, and like wēle generally shows agreement with the introduced object: (355) hawēla tawırta yısırta bıddwa dāra ‘Here is the cow, tied to that

tree’ 243

(356) uhawēlu tre parre dīde pıšlu go ʾīzi ‘Here are two feathers of his

remained in my hand’ 197

The following example shows no agreement with the center of presentation: (357) hawēla musḗlīlox tre-ga ıl dan pāre dīd mzabnítwāla āhet ‘Here, I

brought you twice the money (for) which you would have sold it’ 248

Possibly tre ga is not felt as the correct center and pāre (PL) is remote from the beginning of the clause, so the choice is the default feminine. One example occurs inside a subordinate clause: (358) ūlēwe xılma, čıkūn hawēla ʾasıqsa dīda ʾ ımmi ‘And it is not a

dream, since here is her ring with me’ 320

It should be noted, finally, that, unlike the abstract presentatives, which occur in the 3rd person only, the concrete presentative occurs in other persons as well: (359) babi uyımmi hawēwın āna bronōxūn wēn dīd zaʿli ‘Father and

mother, here I am, I am your son who got lost’ 605

1.5.3 Other thetic expressions Thetic expressions are special clause types that do not contain the usual theme–rheme partition. Typically, there is a lack of a sense of aboutness between the theme and the rheme, as both entities are newly presented. One well-known example is the existential clause, where both the existant and what is said about it are equally new to the discourse. This is clear by virtue of the fact that existants in such expressions are always indefinite. The clearest case of thetic expression in JZ, existential expressions aside, is when we get an explicitly indefinite theme (which is rare to begin with): (360) zıllu yomāsa sēlu yomāsa, sēle xa gōra kıslōhun mın xa bāžēr

rāḥuqta uʾōha gōra pıšle kıslu kma yarxe ʾarxa ‘Days came and



went, there came a man to them from a faraway city and this man stayed with them several months as guest’ 224 Note that both the man and his coming are introduced for the first time in narrative, and they are analyzed as one new package. Note that the English pattern used here to render this notion has a special form as well. The special pattern (verbal form first, indefinite actant) is strictly kept, only in one case do we find the opposite order.31 The verbal forms are mostly intransitive (īsāya ‘come’, mpāqa ‘come out’, qrāwa ‘come close’) and only marginally do we come across transitive forms (tfāqa ‘meet’, šqāla ‘take’). The package of verbal form and actant is separated but rarely by an object or an adverbial: (361) xá yōma ō ha faqīr wēle bīzāla bıṭʾāya čıngırre, sēle barakūse xa

gōra ‘One day, (while) this poor man was going looking for rugs, there came a man towards him’ 290

(362) xá yōma ō faqīr wēle bīzāla ūtfıqle ʾıbbe xá gōra ‘One day he was

walking and a man met him’ 519

Such thetic expressions can be found in JZ containing a presentative; here too, the indefinite actant is normal but the special order is not: (363) ūʾōha faqīr wal kxāze—ma kxāze—xá kapōra wēle bīsāya mın

raḥūqa ‘And this poor man sees—what does he see—a giant is coming from afar’ 534

Our thetic pattern can also occur after a perception verb, as in the next example: (364) ūpıšla palgıdlal wal kxāze sēle xā gōra ūpsıxle dıkkāna wīʾılle

ūšqılle xā līra ū mpıqle ‘It was midnight, he sees (that) there came a man, opened the store, entered, took one coin and came out’ 3–4

In the preceding example it is not quite certain whether the pattern is subordinate or not. Its occurrence with explicit subordination tells us that this is quite possible:

mṭēle qarwāwıt day bāžer dīd dīla xāse skınta tāma. xá gōra qam xazēle ūqam bāqırre: ‘He arrived at the outskirts of the town where his sister lives. A man saw him and asked him:’ (202). 31



(365) qam jōbīle dıkkandāre ʾınnu sēle xa gōra rāba dōlımánt ‘The

shopkeepers answered him that there came a very rich man...’ 571

The tense used in our pattern is the preterite. One example, coming from dialogue, has the present tense: (366) ūmırru ta ḥakōma: xa nāša faqīr gıbe maḥke ımmox ‘They said to

the king: a poor man wishes to talk to you’ 370–371

Like the narrative, both the presented man and his wishing something are new to the discourse. Another environment where such thetic expressions occur is after the expression ōqat32 ġzēle ‘no sooner/hardly did he look...’ (lit. ‘so much did he see’): (367) xšıkla dunye, ō qat ġzēli, qrūle xa ʾarya ʾılli ‘It became dark, I

hardly looked, and a lion approached me’ 379

(368) ōqat ġzēle mpıqle mın xe daw qaṭra xa xūwe kōma “He hardly

looked, and there came out from under the rock a black snake’ 439

It is important to note that ōqat ġzēle occurs with other verbal patterns, not only with the thetic pattern; its functions are discussed below, under macrosyntax.

The lexical use of ōqat ‘so much’ is illustrated in the following example: ūʾōqat ṣrıxli ṭamāha la qam jōbítūli ‘I shouted so much, why did you not answer me?’ (265). 32

2 THE ATTRIBUTIVE RELATIONSHIP The genitive construction (also termed annexion, Heb. smīḵūṯ, Arabic ʾiḍāfa) is a trademark of the ancient Semitic languages as well as of some of the neo-Semitic languages. The relationship obtaining between the entity in the so-called construct state and the substantive in genitive status (and its syntactic equivalents) is the topic of this chapter. An important detail must be added: the Semitic genitive is merely a syntactic marking of the attribute, unlike, e.g., the Indo-European genitive, which is found in other functions as well, such as object and possibly others. This never happens with the ancient Semitic genitive. The extension of this genitive construction by substituting this genitive substantive with a syntactically-equivalent clause is also found in ancient Semitic (in full use in Akkadian and Ethiopic, less so in Arabic and Hebrew). This extension is still in full use in JZ (Goldenberg 1993:296–298). Like the most ancient Semitic languages, this extension in JZ is found at the basis of the most important strategy of subordination, which is fully described in this chapter. The construct state is by and large explicitly marked in JZ. The function of all these entities is to serve as a syntactic nucleus. The genitive status is often explicitly marked in JZ as well, but to be able to cover terminologically the entire paradigm, it is henceforth referred to as the attributive function.1 Since the nucleus, which determines the function of the entire syntagm, can function in any syntactic status, these attributive complexes are found everywhere. For this reason, in this chapter only the internal syntax of the attributive syntagms is discussed, that is, their structure, whereas their external syntax, viz. interrelationships with the surrounding environment is


A detailed description of this relationship is found in Cohen 2010.



dealt with where appropriate: object complements and adverbial complexes under §§3.1.1 and 3.2 respectively, adjectival complexes under §4.1. The following table previews the various combinations which have to do with the attributive relationship and their combined result: attribute⇒ nucleus ⇓


substantive annexion pronoun adverb

syntactic adjective

prepositional syntagm



‘possessive’ attribute

substantive+clausal attribute

possessive or pronominal adjective; pronouns

adjectival, substantival and adverbial clauses

inflecting preposition

adverbial clause

Not all attested combinations are given representation here, as some have limited syntagmatics, for instance, an adverb as attribute occurs only with an adverbial nucleus to produce complex adverbs (e.g., axxa ‘here’ ⇒ l-axxa ‘hither’; ʾēka ‘where’ ⇒ m-ēka ‘whence’, etc.), an adjective as nucleus occurs only with substantives and only certain pronouns producing a superlative pattern with kullu (aw zōr kullu ‘the youngest’).

2.1 NUCLEUS MARKING AND NUCLEUS GROUPS The nucleus of attributive constructions can be a nominal, a pronoun or an adverb. What makes it special is that it is a bound form, which never occurs on its own and must have some kind of attribute. The nucleus syntactically represents the entire syntagm and may occur in any syntactic status. There is a complementary distribution between an unbound entity and its nucleus, or construct form. The former occurs when there is no attributive expansion, or when the expansion is appositive (as with an adjective), the latter occurs only with an attribute. Nucleus marking in JZ is alive and fully productive thanks to the renewal of this marking (as is the case also in other NENA dialects). There are two basic forms: the ancient status absolutus,2 e.g., in bēs (‘house’, a nucleus) vs. bēsa (‘house’, unbound) and the renewed form, which is marked by -ıd or -ıt, e.g., bēsıd (‘house’, a nucleus). Many entities have both forms, some only one. Few other items, mainly borrowed ones, are not marked at all. In some cases; in others, it is probably a case of backformation, e.g., bax from baxta (and not from baxtıd). 2



The first marking, the ancient status absolutus, is not contested. The renewed form, -ıd, may be analyzed in two ways—the one presented above, that it is merely a mark of the construct state, or in our terminology, the nucleus, and second—that it is still equivalent syntactically to the diachronic function of d-. What seems to corroborate this view is that in JZ, d- with nouns is truly rare (the pronoun dīd occurs instead). It is perfectly plausible that this renewed ending issues diachronically from the pronoun d-. Nevertheless, it will be demonstrated that the ending -ıd is synchronically different from the pronoun d-. The most important evidence for the first analysis has to do with the places we find the form -ıd. One would expect, if -ıd is indeed the equivalent of d-, to see full parallelism in their functions, that is, to find -ıd only where d- is expected. This is hampered by the occurrence of -ıd with prepositions. While the majority of the prepositions occur in their apocopate form only (that is, their synthetic forms, e.g., mın ‘from’, ıl ‘to, on’, kıs ‘chez’, xe ‘under’, rıš ‘on’) when preceding a substantive, other prepositions occur only in their analytic nucleus form (e.g., ımmıd ‘with’, qarwāwıt ‘near’). A third group of prepositions have two variants: b-~bıd ‘in’, ta~tad ‘to’, ‘to, on’.3 These analytic nucleus endings (-ıd) on a preposition have no motivation whatsoever, nor an explanation, if analyzed as d-: As far as I know, a preposition is not linked to its genitive rectum by d-.4 One could claim that this marking had begun with other nuclei—pronominal or substantival, and only then extended to prepositions, still, the diachronic origin of this -ıd is not contested, only its synchronic value. The motivation is clear, however, if these endings are viewed as nucleus markers, which fits very well the entire picture. A related discussion will be presented below with regard to the attributive marking on definite DETs, various pronouns and the present copula (§§ and 2.3).

2.1.1 Pronominal nuclei That a pronoun serves as nucleus, namely, take the position of the nominal entity in construct-state, can be deduced from: In Sabar 2002b one finds listed both mın~mınnıd ‘from’ and ıl-~ıllıd ‘to’, since the dictionary covers ENA as well. 4 Mengozzi 2005 explained that phrases such as ımmıd gōra ‘with the man’ originate in ʿamm–ēh d–ḡaḇrā glossed ‘with him, of the man’, based upon historicallyoriented writing of Christian NA as well as actual examples from Syriac. 3



1. the explicit mark of nucleus when applied to a recognized pronoun: māt (nucleus form of ma ‘what’ = ‘whatever, as much as’), ēmıt (523, nucleus form of ēma ‘what/which’ = ‘whichever’) and kud (nucleus form of the obsolete *kul, ‘each, (the one) who/which’ = ‘whoever’); 2. an explicit mark of the attributive function (discussed below), e.g., mād dīla ‘what(ever) it is’ (615); or 3. the ability to precede the attributive bound pronouns. The latter point can be illustrated by d(īd), which joins the attributive pronouns (in the singular dīd-i, -ox, -ax, -e, -a, and in the plural d-ēni, -ōxun, -ōhun). This morphosyntactic behavior also reveals the allomorphy between dīd and d-. Whether or not it is a pronoun or not can be examined by an actual commutation with a substantive (this being the basic definition for a pronoun), e.g., mın bēsi ‘from my house’ (240) vs. mın dīdi ‘from mine’ (239). The pronouns are kud ‘each’, mād ‘what(ever)’ dīd ‘that (which, of), manīd and ʾēmıd ‘who(ever)’. The first pair of examples below shows both criteria—nucleus marking and pronominal representation: (1)

kud ʾāwızlox hawūsa ‘anyone (who) does you a favor’ 458


nāš ʾāmer ‘a man (who) says’ 440

In ex. (2) a substantive, nāša, is marked as nucleus and expanded by a clause. The very same structure is found in ex. (1), where instead of the substantive one finds the nucleus form of *kull.5 The fact that it stands for a substantive (indeed, for a closed set representing humans) it is treated henceforth as a pronoun. The second pair reflects the same idea, with another pronoun: (3)

bıt šūl muḥkēla ʾılle ‘the business she spoke to him about’ 778


mın mād mırrīlox ‘from what I told you’ 616

šūl is the nucleus form of šūla ‘business’, expanded by an attributive clause. mād, as already mentioned, is the nucleus form of ma ‘what’. It too is ex-

5 In JZ the unbound form of the common Semitic *kull is not attested, and what is left are the forms kull-, which occurs only with suffixed attributive pronouns (kulle ‘all of it’, etc.) and kud ‘each, any’ (with an attributive substantive or pronoun), ‘(any)one who’ (with a clause, always in nominative status, see §2.3.2).



panded by a clause. mād is a pronoun since it can occur instead of a substantive in the same syntagm, as illustrated by exx. (3) and (4). An entity is a pronoun if it actually represents a substantive, not only paradigmatically: (5)

māxulla qaḥra dīdox zōdanta-la mın dīdi ? ‘Is your grief superior to mine?’ 239

The pronominal entity dīd- represents qaḥra ‘grief’ twice; most clearly in the second occurrence where it is found alone (dīdi ‘mine’), but also in its first occurrence (dīdox ‘yours’), where it represents qaḥra as well—lit. ‘grief(,) that of you’. The following set of examples illustrates that the syntactic nature of dīd is pronominal also when not followed by an attributive pronoun, but by a numeral: DET







(9) (10)


(11) ʾō


nucleus baxtıt d



‘theFS second’ 813


‘the second wife’ 818



‘the third wife’ 831



‘theMS second’ 884

gabārıt ʾo



ṭḷāha tre

‘the third hero’ 962 ‘the second hero’ 961

The most important opposition for the present question is in exx. (6) vs. (7) and (9) vs. (10) (tinted grey). Again, the entity d(īd) is in a paradigmatic relationship with a substantive, and hence it represents one. This becomes even more obvious when we examine the third example in each cluster, exx. (8) and (11). The analysis of these examples is modeled after the one conducted above for ex. (5): baxta dīd ṭḷāha is thus glossed ‘the woman, that of three (i.e., the third one’). The following example shows that mani ‘who’ is a (non-nucleic) pronoun as well: (12) dīd mani–le ōqadda māl? gımri: dīd flāna nāša–le ‘So much

property, whose (lit. of whom) is it? They say: It is of a certain man’ 20

Note that here too, although not in exactly the same syntactic setting, but rather in some kind of parallelism between question and answer, mani represents, inside the question, what turns out to be, in the answer, flāna nāša ‘a certain man’.



The following pair of examples present the unbound and nucleus forms of mani. This time, however, mani is not an interrogative pronoun: (13) ınnu kud mani dīd yāwıl pāre ta ṣıdāqa ‘After all, each one who

gives money for charity...’ 292

(14) ay damma manīt hāwe ḥ mīla qāmox bāse lqṭāla ‘At that time,

whoever is/will be standing in front of you will be killed’ 522

It should be mentioned that the pronouns originating in interrogatives and functioning as signals of indirect questions are different from their nucleus forms: (15) ūbxāzin ēma bḗsa–le ‘I will see which house this is’ 162 (16) ēmıt julle dʾājıble lıbbox ṭḷōb mın de kwīna ‘Ask for whatever

clothes which please you from that tent’ 523

(17) ūla zeʾla mánīle mamo ūmani qam nābılla ıl tām ‘but she did not

know who is Mamo and who brought her there’ 308

(18) ūʾíswāle xa ṭēra dīd ḥukū m, rıš rēšıt manīt yātūwa, ā wa pāyıšwa

ḥakōma ‘and he had a royal bird, on whosever’s head it would sit, he would become king’ 145–146

(19) ılla gıban xazyan ma ʾīs gāwa ‘I do want to see what is in it’ 174 (20) šaqlaxni māt ʾīban mṭāšax go jēbēni ‘Let’s take whatever we can

hide in our pockets’ 259

In the first example in each pair (exx. (15), (17) and (19)), the signal for indirect question is not marked as nucleus. On the other hand, the second in each pair of examples (exx. (16), (18) and (20)) has a nucleus form of that pronoun, and then we typically do not have an indirect question but rather a syntactic indefinite.6

6 Such syntactic indefinite constructions are found, by the same analysis, in Biblical Hebrew: wayyōmɛr YHWH ’ɛl mōšɛ mī ʾăšɛr ḥāṭā’ lī ʾɛmḥɛnnū mis-sip̄rī ‘And the Lord said unto Moses, whosoever hath sinned against me, him will I blot out of my book’ (Ex. 32:33; lit. ‘who that sinned’), as well as in Arabic: wa-huwa maʿakum



2.1.2 Substantival nuclei Substantives are the largest group attested as nuclei of attributive constructions. Polotsky in his unpublished notes of Zakho grammar (PUG) distinguished between the two types of nucleus marking, viz. between the original (and renewed) apocopate form and the new formation with -ıd, calling them strong and weak annexion (or juncture) respectively. However, the functional distinction between them is not clear, and for now it must be regarded as a mere variation. Compare the following pairs: (21) pxabır xōre ‘by the word(s) of his friend’ 717 (22) xabrıt xōre ‘the word(s) of his friend’ 657

Exx. (21) and (22) are synonymous, so far as I can tell. The same applies to the following pair, where both syntagms occur in exactly the same syntactic setting, which corroborates this variation even further: (23) fhımle ʾaw gōra dīd šmeʾle ʾarya ʾıl xabır mxēla baxte bıt ʾarya

‘The man understood that the lion had heard the word(s) his wife said about the lion’ 683

(24) baxtunta pıšla bıthāša ʾıbbu ū mtaxmōne bıt xabrıt mírrēla ‘The

woman began pondering about them and thinking about the word(s) he told her’ 858–859

The following example contains two almost identical syntagms, each with a different nucleus marking: (25) kud nāšıt hāwe xīzan mād mande ʾīze, kısya bīze, ūnāš hōya ʾēne

kpınta gımjāmıʿ čāŕ nıkāŕ e (sic) rāba ʾīxāla zōdāna ‘Any man (who) is poor, whatever he may cast his hand (upon), he succeeds, but a(ny) man (whose) eye is hungry gathers (from) all around him lots of excessive food’ 822

Here too, no difference can be discerned. Note that both marking types occur with both a substantive (exx. (21) and (22)) and a clause (exx. (23)– (25)) as their attribute.

ʾayna–mā kuntum ‘and he is with you wheresoever you may be’ (Qur. 57:4; lit. ‘where that you were’).



Unlike ancient Semitic, where the entity in construct state could not occur with various DETs,7 in JZ this group is compatible with both the DET and QUANT groups (described above, §1.2.1, esp. exx. 13–15): (26) ʾō šūl nāṭōre dīde hīle ‘That matter of the guards is his’ 642 (27) ō šūl gımrat daʿwīd de ʾarsax lēwe bırya ‘That matter you tell

about your rival wife has not happened’ 822–823

It makes no difference whether the attribute is a substantive or a clause. This is the case even when the attribute is a pronoun, which renders the substantive definite in itself: (28) ēha brāti waʿda dīda īle ta gwāra ‘This daughter of mine, it is

time for her to marry’ 172

(29) mjōwıble ʾaw brōne rūwa ‘This older son of his answered’ 646

The inherent definiteness of these substantives with genitive pronominal suffixes excludes indefinite DETs. When xa occurs in such a syntagm, it is rather a QUANT: (30) qımle xa brōne ūmırre ta bābe ‘One son of his got up and said to

his father’ 632



occur as well:

(31) ūla rʾıšle dammıt ʾurru tre baxtās ʾıstāze ‘but he did not notice

when his master’s two wives entered’ 818

2.1.3 The infinitive as nucleus As a sub-group one also finds infinitives as nuclei. Infinitives of intransitive lexemes are only capable of having their agent added as an attribute: (32) mṭēle waʿıd ʾīsāyıt danya ṭḷāha xūrāsa ‘The time arrived for the

three friends’ coming’ 474

(33) mṭēle waʿıd mpāqıt 8 xūwe ‘The time arrived for the snake’s

coming out’ 453

Of course there are exceptions, as the construct adjective in Arabic when found in an indirect attribution—ʾaṭ-ṭawīlu l-qāmati ‘the one whose stature is tall (lit. the one long of stature)’. See Goldenberg 2002. 7



In exx. (32) and (33) the nucleus forms of the infinitives are expanded by an attributive substantive. In ex. (34), it is the attributive pronoun -i which represents the agent: (34) uzaʿli bıd ʾīzāli ıl knıšta ‘and I got lost while going (lit. in my

going)9 to the synagogue’ 600

The following pair of examples shows the same lexeme, with two different nucleus forms, the absolute and the analytic, followed by the agent: (35) pıšla waʿıd ʾīzal jwanqe lnēčīr ‘It was the time of the youngsters’

going out for a hunt’ 885

(36) mōnıxli čūča ıl ʾīzālıt ʾırba ūʾuzli muxwāsu ‘I looked a little at the

sheep’s walking, and I did like them’ 388

As seen above with substantives, infinitives show no difference between both types of nucleus marking. On the other hand, when the lexeme is transitive, the attribute in principle represents the undergoer: (37) ūškıllu bıt ʾwāzıt ṣandūqa ta ḥakōma ‘They begun preparing (of) a

box for the king’ 718

Here, however, the expression of the undergoer is not limited to attributive entities: (38) bıš xlēla ʾılli ʾwāza xōrūsax mın ʾwāz xōrūs bnās mīre ūṗāšāye

‘Befriending you pleases me more than befriending (of) emirs’ and pashas’ daughters’ 896

Ex. (38) has two occurrences of the same infinitive: the second (ʾwāz xōrūs bnās mīre) has an attributive representation of the undergoer, while the first (ʾwāza xōrūsax) has the same relationship with its undergoer as a verbal form with its object. The following pair seems similar: 8 A special and rare case is the paronomastic infinitive construction with an infinitive as nucleus (§ mpāqıt mpıqla ġzēla xá rakāwa malōze ‘Just upon getting out (lit. getting out [which] she got out), she saw one rider hurrying’ (905). The function of this construction is very similar to ʾōqat ġzēle (§ Such expressions’ narrative functions are described under §5.1.4. 9 This is an infinitive, rather than a gerund, since gerunds of the simple stem always occur with bı- rather than bıd. In addition, gerunds do not occur with attributive pronouns.



(39) ġzēle ʾīza mundēsa lxpāqe ‘He saw her hand cast to his lap’ 871 (40) ūpıšla ṭāla maṛaq nḥāqa ʾılle ū xpāqa dīde ‘and it became her

desire to touch him and hug him’ 796

However, ex. (39) in fact has a substantive (xpāqa ‘bosom’) rather than an infinitive and the attributive pronoun represents the possessor, while ex. (40), comprising an infinitive, has a pronominal syntagm (dīde) which is its object. These verbal-type objects essentially do not belong with attributive constructions and are hence described in the following chapter.

2.1.4 The adjective as nucleus Adjectives as nuclei of attributive constructions are quite rare. There are three prototypes. One resembles an ancient Semitic mechanism to produce superlative adjectives:10 (41) rūwıt ganāwe ‘the biggest/head thief’ (lit. ‘big of thieves’) 160 (42) ʾaw zōr kullu ‘the smallest of them all’ 791 (43) hīl tūle xazır dan xšīwıt kullu ‘until he sat down with the most

important of them all’ 941

(44) ʾaw xılyıt lıbbe ‘his favorite (lit. the sweet of his heart)’ 593

This is the only way, to the best of my knowledge, to express the superlative in JZ (there is only one exception, with dīd, see n. 21) The second is found in several languages, where the logical order (otherwise nucleus—attribute) is virtuallly reversed: (45) uʾē pappukıt ḥakumta ‘but the poor queen’ (lit. mais cette pauvre

de reine’) 797

(46) pappūkıt xmāra ‘poor of a donkey’ 919 (47) mquṭmantıd ʿarabēsa ‘the accursed (lit. ashen) Arab (woman)’

SAG 5.62

The last example belongs to the same group, despite the fact that the nucleus is a subtantive:

qəṭōn bānāw ‘(the) young(est) of his sons’ (2 Chr. 21:17); ḥaḵmōṯ śārōṯɛhā ‘(the) smart(est) of her ministers’ (Judg. 5:29). 10



(48) kumta, xakma šırmāsa–lu ʾ ılla – ʾihihh! – ṭōpıd šırmāsa ‘a black

(woman), it is some arse (that) she’s got! Wow! A cannon of an arse!’ Hopkins 2009:380

This attributive construction is found in many languages (ibid.), and named ‘the emotive genitive’. It is characterized by 1. a reverse of the ‘normal’ order; 2. the nucleus, whether a substantive or an adjective, carries some emotional weight; 3. there is often an agreement between nucleus and attribute (which is clearly absent otherwise), e.g., ex. (47), where the adjective mquṭmanta and the attributive substantive ʿ arabēsa are both FS (the other cases cannot show this agreement for various reasons: the adjective in exx. (45) and (46) do not show gender difference, and ex. (48) has a substantive, whose gender does not change). The relationship between this so-called emotive annexion and the elative in Arabic (ibid. 382) does not obtain here because the “attribute” in Arabic (ʾakbaru r-rajulin lit. ‘bigger of a man’) is in the singular while in our case (as well as in BH) it is in the plural. The third group consists of one entity, ġēr, which is borrowed from Arabic. When preposed to a substantive, it excludes any DETs or QUANTs, yet, it is not quite one of them (it stands, like kud, as a construct nucleus), but displays its Arabic syntax. Compare our ġēr yōma (611) to Iraqi Arabic ġēr yōm (Woodhead and Beene 1967:341b) ‘another day’. That it is the nucleus of the entire syntagm can be deduced from its tendency to occur with attributive pronouns, e.g., ġēre ‘except him, another’. It is analogous in this behavior to various pronominal nuclei, like kud, rather than to an adjective. In addition to the three groups described, there is another marginal phenomenon: (49) gōra qamāyıt danya baxtāsa hulle jwāb ‘The first husband of these

women answered...’ 838

(50) ūpıšlu bıṭʾāya gō xá māsa qarūtıt dē māsa dwēlu skīne ʾō jwanqa

uʾanya baxtāsa ‘They began looking in a village nearby to (lit. of) the village where the youngster and the women had lived’ 838

This is not quite the same thing as other adjectival nuclei, since the mark as nucleus actually applies to the entire nominal syntagm, i.e., {gōra qamāya}+-ıt danya baxtāsa. In other words, the attributive substantive group danya baxtāsa ‘(of) these women’ does not apply only to the adjective but to the entire syntagm, that is, {gōra qamāya}. qarūtıt in ex. (50) is not an adverbial (which would be qarwāwıt), but an adjective showing FS marking,



which represents pronominally the preceding māsa ‘village’. An alternative, more frequent way is to represent the nominal syntagm by dīd: (51) ʾáy ʾasıqsa turta dīd ḥakōma ‘The broken ring of the king’ 732 (52) āna

rāba dōlāmánt wēna ū brōna yakāna dīd yımmi ‘I am very rich and my mother’s only son’ 541 +

2.1.5 Adverbial nuclei This section deals with adverbials rather than adverbs since not only adverbs function in adverbial function, but other entities as well, the commonest in many languages are such entities that represent a place or a time span (e.g., ‘day’) and function adverbially. Such adverb(ial)s, even when marked as nuclei, still function as such. But when they occur with a clausal attribute we refer to them as conjunctions, and as prepositions when followed by a (pro)nominal syntagm. The latter are often marked as nuclei as well (although admittedly sometimes lacking an unbound form). One can start with an item such as waʿda ‘time’. It can function substantivally as well as adverbially: (53) hatxa ʾuzle rāba waʿda ‘thus he did (for) a long time’ 579 (54) mḗsīle waʿıd palgızlal ‘bring it by (lit. [(at the) time of])

midnight’ 929

(55) brat ḥakō ma mḥōṣırra waʿıd ġzēla ʾō jwanqa ḥ mılle barqūl daw

gabbara ‘The princess sighed when (lit. [(at the) time]) she saw that the youngster stood in front of the hero’ 765

When used adverbially as a nucleus with an attribute, it has a tendency to grammaticalize as a temporal preposition or conjunction (depending only on the type of attribute following, be it a substantive or a clause). The following pair of examples shows this grammaticalization very well. The entity damma ‘time’, borrowed from Kurdish, is hardly attested in substantival function anymore, and, together with its definite DET (ay), it is synchronically only an adverbial: (56) ay damma


mırre ‘(At) that moment, he said:’

(57) ay dammıt šmeʾle hatxa, ṣrıxle ūmırre ‘When (lit. the moment) he

heard thus, he screamed and said:’ 288–289



The only difference between the respective structures of exx. (56) and (57) is whether the adverbial ay damma occurs expanded by a clausal attribute (ex. (57)) or not (ex. (56)). In the examples, the function of waʿda and damma is the same, regardless of whether they are further attributed or not (although waʿda is attested in substantive function). The same idea, although less common, is found with locative entities: (58) bıd waʿdıd gnāpıqwa mın bēsa, āzılwa xa dūka ‘At the time he was

leaving the house to go (to) some place...’ 24

(59) šqīlāle lxāṣe ūnublāle dūkıt zıllu ‘He took her on his back and

carried her where (lit. place) they went’ 272

dūkıt and a substantival attribute is for some reason not used adverbially, but dūkıt with a clause is simply the equivalent of a locative conjunction in other languages. Locative adverbs are occasionally used as nuclei, just like a preposition: (60) ġzēle gıyāne dīd wēle ḥmīla rıš xa ṭūra umın ʾıltḗx īsın baḥar ‘He

saw himself standing on the top of a mountain and below there is a sea’ 111

(61) mōnıxlu mıltḗx qaṣra, qam xāzēle bīzāla ‘They looked below the

palace, (and) saw him running’ 494

(62) gıbınwa mandınwa gyāni mın ılʾḗl, bale zdeʾli ‘I felt like throwing

myself from above, but I feared...’ 378

(63) ūġzēle ġulāma murma sēpa mın ʾılʾḗl rēše ‘and he saw a lad lifting

a sword above his head’ 95

In all these cases the adverbial originally consists of a preposition and an adverb. The first example in each pair (exx. (60) and (62)) illustrates the use of the adverb unexpanded, while the second example (exx. (61) and (63)) has these adverbs functioning as nuclei of an entire attributive construction. Other similar examples are qamāye ‘before(hand)’ (451), whose nucleus form is qamāyıt ‘before-...’ (391). The semantic link between the adverbial nucleus preceding a substantival or clausal attribute is not always transparent: (64) ūmın rızlox mın dūkox pšōqın ʾatta rēšox ʾāse mafōre ‘If/as soon as

you move from your place, I will immediately (lit. now) have your head cut off’ 414



(65) mın yāʾe, rēšēni pqaṭeʾle ‘If/as soon as he learns, he will cut our

head off’ 256

mın (otherwise ‘from’) is here used as a conjunction, either conditional or temporal. Another source for nucleic adverbials are adverbial interrogative particles, such as kma ‘how much’ and māṭo ‘how’: (66) la kīʾın māṭo zılla ‘I do not know how she went’ 310 (67) uzle ōha ʾırba māṭut mírrāle ‘The sheep did as (lit. how) she told

him’ 206

(68) āna gıban yāʾan kma māl ıtlox ‘I want to know how much

property you have’ 173

(69) halli ḥaqqe kmad dīle ‘give me its price, as much as it is’ 669–670

The unbound pronoun māṭo ‘how’ is opposed with the nucleus form māṭut ‘as’, and the unbound kma ‘how much/many’ with kmad ‘as much as’. The prepositions, which are the most common form of nucleic adverbials are elaborated and discussed above, §2.1. Their nucleus status is shown by comparing them with substantives in the construct state (analyzed here as nuclei) and their common genitive attribute (§2.2.1, found under exx. (75), (76), and (77)). One preposition only stands out in JZ as not being a nucleus: (70) qam nablīla namūsīye dīde bēb ʾāwa ıl bēse ‘They brought his bed

with him (lit. with he) to his home’ 568

bēb is borrowed,11 and in JZ it may occur, unlike any other preposition, with the non-attributive personal pronoun āwa. The exceptional transparency of JZ enables us to see that what in other languages would often be considered to be three distinct groups (adverbs, prepositions and conjunctions), are here, at least in principle, one group with two alloforms. This is summarized in the following table: adverb(ial) unbound

bound (nucleus)



= resulting function



2. prepositional phrase




1. simple adverbial 3. adverbial clause

(53) (55)

See Mutzafi 2008:121–122 for a possible Aramaic etymology.



These syntactic qualities, rare even in Semitic languages (let alone neoSemitic), faithfully reproduce the original Semitic state of affairs, fully represented by the early phases of Akkadian Old Babylonian, ca. 1800 BC): adverbial allomorphs kīʾam (unbound) kīma (bound) asNUC


— ab-ī-šu fatherATTR+3MSATTR īpuš-u he-madeATTR



‘as his father’


‘as he made’




JZ is similar in this respect (as well as in other aspects of this syntactic relationship) to Old Babylonian Akkadian.

2.2 ATTRIBUTIVE GROUPS The attributive function, as said above, corresponds to the ancient Semitic genitive slot; this ancient genitive slot was not limited to morphological substantives and could house clauses as well (which are marked as such in Akkadian by the so-called subjunctive form of the verb, see īpuš-u ‘he made’ in the table above). JZ has no case marking of course, but there are nevertheless some formal indications that make this syntactic status crystal-clear. The function of this slot is to attribute information to the nucleus, whatever its nature may be. Like ancient Semitic, this slot in JZ can be filled by a clause as well. This function is marginally marked as such. Unlike the nucleus function, the attribute is a clear and constant syntactic status, different, e.g., from the thematic status in several parameters which are clearly observable in its markedly distinct paradigmatic constitution (i.e., the group of forms which may figure in the attribute slot). The differences are found in various forms (in DETs and pronouns, the copula, the syntax of the verbal form wēle) as well as in different behavior: the ease with which substantives are substitutable with verbal forms; the occurrence of all levels of definiteness in substantives. Note that in thematic status the substantive is, at least in principle, definite, while in attributive status it may be indefinite as well.

2.2.1 Substantival attribute group Whereas Akkadian and Classical Arabic mark the attribute by the genitive ending, in JZ the substantive itself in the attributive slot is not marked by a



case marking. Yet, it obligatorily occurs with a special series of DETs. The definite DETs are marked for the attributive function.12 In other words, JZ has a clear signal analogous to the Semitic genitive: an attributive allomorph of the definite DETs. In the following section it is shown that the rest of the demonstrative pronouns behave the same way: non-attributive MS DET


ʾaw ʾay


attributive daw day


This marking is not, as one might conjecture at first glance, the nucleic pronoun d- (§2.1.1), since it also occurs where such d- is not called for, most notably after prepositions (for instance ta daw gōra ‘to the man’, see below). Moreover, normally, when the DET follows the pronoun d(īd), this mark is found as well: (71) mın do bıhna rwīxa d do jwanqa ‘from the patience of this

youngster’ 784

(72) ʾōda dīd daw gōra ‘the room of this man’ 334 (73) ʾīxāla dīda udīd ø gōra ‘her and her husband’s food (lit that [of

her husband])’ 553

(74) ḥāl ūquṣta dīdi ūdīd xa ṭēra ‘my and a bird’s story food (lit that

[of a bird)’ 189

The first d- is analyzed as a pronominal nucleus (parallel to dīd in ex. (72)), whereas the second d (which, in ex. (71), occurs after the preposition mın as well) is the attributive marking on the DET. Exx. (73) and (74) are adduced in order to show that whenever the DETs are different (e.g., ø or xa) no such marking occurs. This is illustrated in the following table:

This is comparable with German, where most case marking is found in the DET system. 12



pronominal nucleus (as any nucleus)

attributive slot






(72) (74)


do ø



substantive jwanqa gōra gōra ṭēra

If this had been indeed the pronoun d- in the attributive slot, we would have expected it with the other DETs as well. We can deduce, based upon the foregoing examples, that this d is an inseparable part of the DETs, and has no existence outside this domain. This attributive marking is consistent across all types of nucleus. The illustration of this last point starts with the synthetic form of the nucleus, whose nature is not contested, after which we find the attributive alloform of the DET: (75) bēs daw gōra ‘the man’s house’ (557); bēs day brāta ‘the

daughter’s house’ (544) vs. bēs xa brāta ‘the house of one daughter’ (542); xās dan kapōre ‘these giants’ sister’ (408); bax dōha gōra ‘this man’s wife’ (231); līšāń do ḥammaš ‘this book’s language’ (591); ʾaxṓn de baxta ‘the woman’s brother’ (853); ʾīzás danya kapōre ‘these giants’ hands’ (758)

This is the normal situation; there is only one exception to this group.13 Note that in this syntactic situation as well (viz., when a nucleus precedes a substantive), d- was not normally called for either, since the use of d- was altogether another strategy of attribution.14 Analytic nucleus marking (-ıt/d) works just the same as far as the attributive marking is concerned: (76) pummıd daw nāša ‘this man’s mouth’ (924); gūrıd day māsa ‘this

village’s men’ (956); šımmıt dan nāše ‘these men’s name’ (301) šımmıt daw gōra the name of the man’ (563) vs. šımmıt xa gōra ‘the name of one man’ (559); pārıt do gōra ‘this man’s money’ (283); ʾīzıt de baxta ‘this woman’s hand’ (222); dūkıt danya nāše

bēs ʾaw sımsar ‘the matchmaker’s house’ (568). That is, one finds, e.g., in Syriac, rūḥ qudšā along rūḥā dqudšā, but not *rūḥ dqudšā. 13 14


CHAPTER TWO ‘these men’s place’ (463); lıbbıt dōha ṭēra ‘this bird’s heart’ (378); bābıt dēha xamsa ‘this girl’s father’(232)

This group, which is substantially larger than the former (there are considerably less synthetic substantival nucleus forms), has only a few exceptions.15 Synthetic prepositions (which have the original, historic form) join the nominal group in an identical way: (77) mın daw xalwıt arya ‘from the lion(ess)’s milk’ (80–81); ta daw

xēt ‘to the otherMS’(66); mın day xzēna ‘from the treasure’ (251); mın day xet ‘from the otherFS’ (836); mın dan nāšıt karwan ‘from the caravan’s people’ (268); ta dan ṭl āha nāše-xēt ‘to the other three people’ (557); xazır dan xšīwıt kullu ‘around the (most) important (people) of all’ (941); mın do xılma ‘from this dream’(434); mın do xet ‘from the otherMS’ (324); mın dōha gōra ‘from this man’ (531); mın de dūka ‘from this place’ (767); mın dēha dūka ‘from this place’ (28); mın danya juwāhırre ‘from these gems’ (256); mın danya ṭḷāh́ á šuʾāle ‘from these three things’ (427); mın danya ʾarbi ‘from these forty’ (385)

There are very few exceptions to this attributive marking.16 The behavior after prepositions ending in -ıd is identical. In view of these special alloforms of the definite DETs and their consistent occurrence following a nucleus, it can be stated that this group regularly marks the attributive (or genitive) status. Moreover, since substantives as well as adjectives often occur together with these DETs, their importance is immense—a simple noun always occurs with one of these DETs. The definite DETs are especially common, which means that a great number of nouns are unambiguously marked as attributive. šımmıt ay brāta ‘the daughter’s name’ (304) rıš xāṣıt an qōzınne ‘on the mule’s back’ (252) baxtıt ō faqīr ‘this poor man’s wife’ (199); ūgo day bāžır dīd bābe ūyımmet aw brōna ‘and in the town of his father and mother of the child’ (589); ṣrıxlu ḥakīmet kulla ʾay bāžır ‘they called the healers of the entire town’ (590). In the last case, the pronoun kulla is located between the nucleus and the unmarked attribute. This buffer might be the reason why ʾay bāžır is not marked as attribute. 16 ta aw ʿāqıl ‘to the smart one’ (256); ta gōra qamāya ūʾaw dīd tre ‘to the first husband and (to) the second’ (874): ʾaw dīd tre is appositive to gōra and is hence expected to be marked as attributive. bıd ʾaw nıqba ‘in that hole’ (111) belongs to the following group, analytically marked prepositions. 15



A further point to corroborate the precision of these attributive alloforms is that, when the substantive follows a nucleic pronoun, but does not constitute its attribute, it is not marked as such: (78) ʾāw nāša dīd ʾō ʾ arxa skúnwāle kısle ‘the man that this guest

lived with’ 617

(79) uxa lákīʾēwa d–ʾanya trēʾ baxtāś do gōra–lu ‘but no one knew

that these two are the wives of one man’ 772

In both examples the substantive following the nucleic pronoun is the theme of the clause (namely, it is in nominative status). The entire clause is the attribute. Note that, in both Akkadian and Arabic, the first term would not have been marked as genitive when functioning as the theme of a clausal attribute. Such clausal attributes are marginally marked as such as well (see below, §2.2.4). Since the attributive status occurs not only after nucleic nominals, but also following prepositions, the group of nouns in this status is as large as as in the thematic status. The general characteristics of the noun group are discussed above (§1.2.1). Consequently, here we only need specify the differences between the two nominal paradigms, namely, the thematic and the attributive: Attributive


determiners MS


type PL


dyā(ha) —






danya dan

distal definite dir. and indir. interrogative

Yet another DET which shows an attributive alloform is the interrogative determiner ēma ‘which (one)’. Whenever the alloform dēma occurs, it is in attributive status: (80) ġzēle xe dēma balāṭa gdārēlu pāre ‘He saw under which floor tile

he puts the money’ 341

(81) mbōqırre ʾıstāź brōne bıd dēma ḥammaš brōne pqāre ‘He asked his

son’s teacher (in) which book his son will read’ 580

(82) mırre ḥakōma ta yāla: mın dēma mıllıtta wētın? ‘Said the king to

the boy: From which persuasion are you?’ 596



Unlike the DETs and demonstrative pronouns, this is consistent only in one direction: whenever we find dēma, it is an attribute (following a preposition). However, not all cases of a preposition preceding ēma are marked this way: (83) la zeʾle ʾēka ʾāzıl basru ū -b-ēma–ʾal ṭāʾe ʾ ıllu ‘He did not know

where he should go after them and in which direction he should look for them’ 844–845

This occurs only with the preposition b- and only in the material originating in the informant A. The second table illustrates, in a somewhat more detailed format, the entire substantival group, the paradigmatic differences from the theme paradigm are shaded grey: Syntagmatic behavior of the attributive nominal group


1 muqdar 2 muqdar





Eng. parallel



‘a period of a few minutes’





xa kma


4 mın




‘from some woman’

6 mın





‘from (the) two women’

8 bıd




‘at the woman’

7 ıl


9 brāt


11 mın

xa —


ḥakōma 196 baxta


‘the hand of (the) woman’ ‘children of one king’ ‘to no woman’



‘the daughter of the woman’

danya —



‘from these women’

12 bēn

danya kutru



‘between those two women of his’

13 basır



‘after both his women’



‘on that table yonder’

10 šoʾāl

14 rıž

15 mın


dwā ́



‘a period of about one month’

3 īzıt

5 yalunkıt ø







‘women matters’

‘from which place’

muqdar ‘measure’ is a substantive, often functioning as an adverbial, which is expandable by an attribute. This is corroborated by the attributive DET in the following example: muqdar danya ṭḷāha yōme ‘(the) period of these three days’ (946). 17



Any substantive can figure in the attributive slot. Contrary to the thematic substantive, however, as an attribute it is many times explicitly indefinite: (84) with a substantival or pronominal nucleus: ıl-bēs xa brāta ‘to the

house of a daughter’ (542); mın ṭaraf xa jwanqa ‘on behalf (lit. from side) of a youngster’ (542); bır xa rāv ‘(the) son of a rabbi’ (604); dīd xā xamsa ‘of a maiden’ (497)

(85) with a preposition: mın xa bāžēr rāḥuqta ‘from a faraway town’

(224) contr. mın day bāžır ‘from this town’ (791); bıd xā nıqbıt xmāṭa ‘in an eye of a needle’ (302); go xá barīya ‘in some desert’ (276) xe xa dāra ‘under a tree’ (276) vs. xe daw dāra ‘under the tree’ (530); ʾımmıt xakma šıhāre ‘with some blind men’ (343)

Unlike other Semitic languages, where it is often not clear which member of the attributive construction is definite, in JZ it can be explicitly marked: (86) ōha sawōyıt dōha yāla ‘the grandfather of this child’ 91–92

The possibility to have each of the slots, namely, the nucleus and the attribute, each explicitly and separately marked as (in)definite, is very convenient, but it is not frequently used. Infinitives, although rare as themes, are found after all types of nuclei: (87) waʿıd ʾīsāyıt danya ṭḷāha xūrāsa ‘time of the three friends’

coming’ (474); waʿdıt īsāya dīde ‘time of his coming’ (199); sabab ʾīsāya dīdi ‘the reason of my coming’ (882); sabab do ʾīzāla dīdax ‘the reason of this going of yours’ (860); urxıt īzāla–la ūla dʾāra ‘it is a road of going and not returning’ (193); ʾımmıt maʾōre dīda ‘while entering her’ (273) húllēlu šūla dīd mēsōye dīd nāše qam mıšpaṭ ‘he gave them the task of bringing the people in front of court’ 557

Temporal substantives such as waʿda or causal such as sabab function very much like prepositions. Following a substantival nucleus the attributive infinitive tends to exhibit substantival behavior: (88) ṭērıt maḥkōye ‘a speaking bird’ (lit. bird of speaking’ 278); ʾōḍıt

dmāxa ‘bedroom’ (lit. room of sleep, 484–5); dūkıt dmāxa ‘a place to sleep’ 409

Free adjectives occur in the attributive slot as they do anywhere, i.e, with a DET:



(89) mın zōra ūhīl rūwa ‘both small and great’ (lit. ‘from small to

great’, generic and hence DET ø) (90); ta daw šızāna ‘to the crazy one’ (252); mux daw qamāya ‘like the first one’ (398) mux xā jıḥıl ‘like an inexperienced (one)’ (831)’ mux xa ṣkırta ‘like a drunkFS’ (893); kāpıt daw xēt ‘the other’s shoulder’ (122); lxakma xıt ‘to some other (ones)’ (958); mın pummıt xā dla ʾāwız gāzında mın baxte ‘from the mouth of (some)one who does not complain about his wife’ (517)

Adjectival clauses, as in the last instance (see below, §2.3.2), occur in the same slot as simple adjectives. We have seen above (§§1.2.1–1.2.2), that numerals can function (pro)nominally, with DETs: (90) mın danya xamša ‘from the(se) five’ (399); kut xamša ‘all five’

(399); dūkıt dan tre ‘the place of the(se) two’ (869)

In these cases they function pronominally, representing substantives; in the following cases they are true numbers functioning as attributes (a simple adjective for ‘second’ or ‘third’, etc. does not exist in JZ): (91) gōr trḗʾ ‘(the) second man’ (lit. man of two’, 561); baxtıt tre

‘(the) second wife’ (lit. ‘wife of two’ 141); naqıl ṭḷāha ‘(the) third time’ (89); ʾaw dīd tre ‘the second one’ (874)

2.2.2 Pronominal attribute group The first set of attributive pronouns are the bound attributive pronouns (-i, -ox, -ax, -e, etc. See §7.1.1 in the morphologial appendix), which regularly occur with substantives (bēs-ox ‘yourMS house’), certain pronouns (dīd-ax ‘of youFS, yoursFS’) and prepositions (ımm-u ‘with them’). There is a problem with analyzing them as pronouns, since they only interchange with one another in the exact same morphosyntactic environment. However, when examined somewhat more freely, ignoring the difference between degrees of juncture,18 one can arrive at the following paradigmatic interchange:

18 Levels of juncture are sometimes critical, but not in this case, where the morpho-phonological differences do not count much in the syntax of these constructions: ıl ḥakōma ‘to the king’ (646) vs. ılle ‘to him’ (21), where we see a different allomorph of the preposition (ıl and ıll- respectively) in each case.


preposition (92)




substantival nucleus




bēs e

‘to his house’ passim

bēs ḥakōma

‘to the king’s house’ 253

What we see is a paradigm where -e ‘his’ commutes with ḥakōma ‘king’. Using the principles of the structural method, the attributive pronouns can safely be referred to as such, representing a substantival entity. Likewise, if the pronouns are attributive, then so are the substantives in the same syntactic slot. From PUG (§5.3.1) we know that these pronouns occur habitually with substantives that are relatively inalienable. Polotsky divides these substantives into four groups (the examples given here are not exhaustive): (94) Kinship: ʾaxōni ‘my brother’ (815), bābox ‘your father’ (759),

baxti ‘my wife’ (371),19 brōna ‘her son’ (751), gōra ‘her husband’ (785), xāsox ‘your sister’ (200), yımmi ‘my mother’ (430);

(95) Social relationship: ʾıstatti ‘my mistress’ (777), ʾıstāze ‘his master’

(772), ġulāmax ‘your servant’ (791), xōre ‘his friend’ (771), xūrāsox ‘your friends’ (433);

(96) Personality related: ʿāqılox ‘your mind’ (788), bālu ‘their mind’

(888), dımmi ‘my blood’ (685), hıšši ‘my wits’ (383), šımme ‘his name’ (434);

(97) Parts of the body: ʾaqlāse ‘his feet’ (786), ʾēnōxūn ‘yourPL eye(s)’

(313), bōla ‘her hair’ (868), gawde ‘his body’ (931), kāse ‘his abdomen’ (954), lašše ‘his body’ (381), mēʾēwāse ‘his guts’ (963), pummi ‘my mouth’ (921), ṣadru ‘their chest’ (901).

There are, of course, counter-examples (some are mentioned in PUG ibid.), in most cases without any appreciable difference: ́ am, wal gıʾarqa ūbrōna (98) uġzēle xā xazāla ıtla xā brōna bánī-ād gıʾārıq basra. ʾrıqle basru hīl qam dāwıqle brōna dīda ‘and he saw a gazelle (which) had a human son running and its son running after her. He pursued them until he caught up with its son’ 82


Cf. baxti dīdi ‘my own wife’ (lit. ‘my wife, that of me’ 501).



Rarely does this make a difference, in terms of meaning. See, in the examples below, the difference between ‘husband’ and ‘man’: (99) āhat pēšat baxti ūʾāna pēšēna gōrax ‘You will be my wife, and I

will be your husband’ 499

(100) ō gōra dēni fhımle ma qıṣṭa–la ‘The (lit. our) man understood

what was going on’ 827

dēni is a narrative strategy to mark specificity (see further § Other substantives, which relate to Polotsky’s group above, are attested only with the pronominal adjective dīd-: (101) yalunke dīdox ‘your kids’ (492, 662, 673), ʾōda dīdax ‘your

servant’ (685) (vs. ʾōḍax ‘your room’ (815)); fıkır dīdox ‘your thought’ (788)

There are other pairs where the difference between the attributive bound morpheme and the analytic construction with dīd- is here elusive: (102) xa xmārox ‘one donkey of yours’ (917) vs. xmāra dīdox ‘your

donkey’ (921, 924); pārox ‘your money’ (3, 6) vs. pāre dīdi ūdīdax ‘my and your money’ (15); bēsi (486) vs. bēsa dīdi (220) both ‘my house’

The substantives in ex. (102) denote property, there is not much difference between the members in each pair, except the fact that pronominal adjective dīd- does not entail definiteness, whereas the bound attributive pronouns do. The following two groups of examples again show hardly any difference between the two strategies: (103) ḥālox ‘your situation’ (515) vs. do ḥāl dīdox ‘(of) your situation’

(415); ʿīšox (514) vs. ʿīš dīdox (374) both ‘your life’; qıṣṭe (631) vs. qıṣṭa dīde (716) both ‘his story’

(104) baqartox 541 vs. baqarta dīdox (317, 654) both ‘your question’;

bıd zmārox ‘at your playing’ (459) vs ḥaqqıt zmāra dīdox ‘the price of your playing’ (459); ʾō ʾī sāyox ‘this coming of yours’ (798) vs. īsāya dīdi ‘my coming’ (882 )

In the last example in this section, the two strategies occur in the same sentence: (105) bale mōnıxle bıt ʾasıqse lēwa ʾasıqsa dīde ‘but he looked at his

ring, (and lo,) it is not his ring’ 309



The first occurrence of ʾasıqse turns out to be the ring that is currently on him, rather than actually his own ring (=ʾasıqsa dīde). Related to these distinctions is the possession expressed by the otherwise reflexive pronoun gyān- followed by a personal attributive suffix: (106) šımmıt gyāne ‘his own name’ 672; bāžer dīd gyāne ‘his own city’

538; sēpıt gyāne ‘his own sword’ 95; bēsıt gyāne ‘his own house’ 151; mandōyıt gyāne ‘throwing himself’ 495; ımmıt gyāne ‘(together) with itself’ 18

The group of bound attributive pronouns, however, is not all we have in order to establish a sound attributive status in JZ. There is yet another group of demonstrative pronouns (see §1.2.2) analogous to the attributive DETs (described in the previous section) that have an attributive alloform, that is, they are positively marked as attributive: (107) basır dōha ‘after this’ (762); ıl dēha ‘to this’ (19); mın danya

‘from these’ (359)

Analogous to basır dōha is basır hādax (passim), which however, although an adverbial or clausal pro-form, has no attributive mark. The table below summarizes these data: Attributive










distal dem.



(The rest of the PRONs do not have a special form for this position)

The same forms that habitually form part of the DET paradigm occur, albeit rarely, alone, representing a substantive. The following table summarizes the pronominal data in the attributive slot: Attributive pronoun paradigm nucleus




Eng. parallel

1 ʾaqlās


863, 386


‘the feet of one (of them)’

2 ʾaqlās




‘his feet’

3 ta




‘to each one’

4 l-




‘to both of them’

5 mın

kull-+ pron 807


‘from all of+pron’



6 basır




‘after this/that’

7 ıl




‘to this’

8 mın




‘from these’

9 bıd

dayxa, dēxa 290, 360


‘thisFS one/thing’

10 basır


AS 1761


‘after that one yonder’

11 ımmıt




‘with no one’

Several remarks are in order. The differences vis-à-vis the thematic pronominal group are found in the demonstrative group. In addition, there are less obvious differences. First, the thematic personal pronouns (ʾāwa, ʾāya and ʾāni) do not occur in this slot. Second, contrary to the thematic slot, where indefinite pronouns (unlike indefinite substantives!) are attested, in the attributive slot these pronouns are difficult to come by, e.g., the indefinite plural pronoun xakma does not figure here at all and xa is very rare as well, basically occurring only with expansions: mınn-, e.g., xa mınnu ‘one of them’. To conclude, the pronoun is frequently marked, as is the substantival group, in the attributive function. The analogy to the Arabic and Akkadian genitive is twofold—both the fact that there is some kind of marking (which is absent in the other Semitic languages), as well as the purely syntactic value of this marking, which applies to Semitic, but not, e.g., to Indo-European languages. In Indo-European languages, the genitive is not only a syntactic status, since it is used for other syntactic functions, for instance, as the object of thinking verbs.

2.2.3 Adverbial attribute group This group is rather limited, and the adverbs that occur here (spatial and temporal) occur only following prepositions and dīd: (108) hīl ʾatta ‘till now’ (passim); mın ʾatta ‘from now’ (32) as well as

matta (in mattaw lēʾal ‘from now on’ 828); maxxa ‘from here’ (278); mın tam ‘from there’ (435); mın tımmal ‘from yesterday’ (535); mux tımmal ‘like yesterday’ (548)

(109) kud žaġīl dīd tam ‘any worker (who was) there’ (lit. ‘of there’)


(110) ūṭamāha wēla ksūta go ḥammaš dīd bēz-dīn dīd ḷʾēl la xullox kullu

šıkıl ʾīxāle ‘Then why is it written in the book of the heavenly court (lit. court of above) (that) you did not eat all kinds of food’ 559



The “adverbs” in this slot function as attributes, rather than adverbially. Another marginal—but more interesting case—is where a prepositional syntagm functions in its entirety as an attribute to a substantival or infinitival nucleus: (111) ʾarya dēni pıšle mēnōxe pšōʾa ʾēne ʾāl basre ūʾāl qāme ‘Our lion

began looking very carefully (lit. with seven eyes) behind him and in front of him (lit. side of behind him and side of in front of him) 930–931 (see also 826, 856 and 818)

In ex. (111), two prepositional syntagms are the attributes of the nucleus āl ‘side’. These attributes are not a sentence, but an adverbial syntagm which can figure in the attributive slot. The analysis of this example is not in itself certain, because ʾāl is sometimes used as an unbound form (e.g., 825), but the following examples show this phenomenon unequivocally: (112) ūʾan nāš ʾımme ‘and the people (which are) with him’ 855 (113) rubʿıt mınnu ‘a quarter of them’ (lit. from them’) 618

The translation compromises the original structure of ex. (112) (lit. ‘the people of with him’, rather than a clause). The same idea is found in ex. (113) (lit. ‘a quarter of from them’). The prepositional syntagm in ex. (112) is no doubt originally adverbial, whereas the one in ex. (113) is partitive which can be judged otherwise. In the attributive slot entities are attributes, rather than adverbial.20 The prepositional syntagms in the following example looks just like the regular complement of the verb: (114) ʾēn nāša lagsōʾa mmēnōxıd ʾıbbu ‘the man’s eye is not satiated of

looking at him’ AS 1197 (also AS 453, 680)

The object of mēnōxe always includes the preposition b-. Here, despite the ‘substantival’ behavior of the infinitive (occurring with an attribute rather than an object), the object keeps its form, and is put, as is, in the attributive slot. In ex. (115) the attributive prepositional syntagm is partitive: (115) ūla mṣēle ṣābır bıžrab ıl ġzāyıt daw quṭeʾfa ūla ʾī xālıt mınne ‘But

he could not keep patient anymore for looking at the cluster (of grapes) but not eating from it’ 647 (also AS 1357)

This could perhaps be compared with morphological derivation of a preposition into an adjective: Contemporary Hebrew taxat ‘under’ vs. taxt-i / taxt-on ‘(that which is) low(er)’. In the adjectival formation it is no longer considered an adverbial. 20



The partitive is not adverbial, but, like all “adverbial” attributes, it is adnominal (as in the frequent xa mınnu ‘one of them’).21

2.2.4 Predicative attribute group The attributive slot often contains a clausal attribute. This occurs, as attested sporadically in many examples presented above, not only after adverbial nuclei (i.e., conjunctions) and pronominal nuclei, but following any nucleus. The clausal attribute following substantival nuclei is a special feature here, not attested in any other neo-Semitic languages. Moreover, a special marking for the predicative group in attributive function is found only in Akkadian on a regular basis—and in JZ, to a lesser extent. The nature of this clause, which is basically regarded as a paradigm following all nuclei (there are minute differences between the various clause types, to be discussed in the following sections), has as its most salient characteristic the commutability with substantives: verbal forms in this slot are the syntactic equivalent of nouns. This is their most prominent characteristic. First and easiest to state are those forms which are in principle excluded from this paradigm. Two modal forms belong to this group of excluded items, namely the imperative and subjunctive forms preceded by the prefix šud-~št-. That imperative forms tend to be syntactically rigid is very common cross-linguistically. As for subjunctive forms, it is clear that they have a very important function in subordination in JZ. These forms, when preceded by šud- instead of ø, are a mark of non-attributive status.22

An example which illustrates the adnominal status of the partitive is the following: ʾaw quṭeʾfa rūwa usqīla dīd mın go kullu quṭeʾfe ‘the big(gest) and pretti(est) cluster of all clusters’ (644). Note that the partitive syntagm (mın go kullu quṭeʾfe) is preceded by dīd which formally nominalizes the entire syntagm. Another occasional adnominal is the preposition muxwās- ‘like’ (in sqılta muxwāsi ‘pretty like me’ 542) and this too can be seen in cases such as ʾarxe ʿazīze, dīd muxwāsox ‘precious guests, the likes of you’ (lit. ‘that (of) like you’ 316). 22 There are cases where subordination could be contested: mār ta bābax šud čuxxa la bāxe ‘Tell your father, let no one weep’ (or [that] no one should weep)(MA 16.6); húllīla ʾīxāla ūmuyʾılāli lʿōya, mutwāli nēxa ūmírrīla šut ʾaxla ʾılʿōya uxarāye ʾāza ‘I gave her food and brought her inside, I seated her so that she calms down and told her she should eat (lit. let her eat) inside and then go away’ (704). However, there is not one case where šut and the subjunctive follow a nucleus. On 21



Another group that is excluded from the attributive slot is the presentative copula. The cases where, following a nucleus, the pattern looks like the presentative pattern, are in fact the allotagm of R-wēle (see § above). All the rest of the forms can occur in this slot (their actual temporal and other values will be dealt with under §5.6.4). There is one specialized attributive marking, found with the affirmative copula, viz., the attributive alloforms dīle-R/dīwın-R that were mentioned in §1.3.1 above. The first group of examples show a substantival nucleus followed by dīl-: (116) bale bıd ḥaqqıd dīlu ʾısye mın mōṣıl ṭālox hallūle ‘But give them to

him at the price which they had (lit. are) come to you from Mosul’ SYG 2.29

(117) qam mesēli ʾıl day qaṣır maṭmıryam dīla ʾılle ‘They brought me to

the castle on which (the statue of) the Virgin Mary stands (lit. is on it)’ SYG 4.16

(118) bıd daw waʿdıt dīle raḥūqa mınna jwanqa xet ‘at that time (which)

the other youngster is faraway from her’ 869

In all three cases a nucleus (ḥaqqıd ‘price’, qaṣır ‘castle’ and waʿdıt ‘time’) precedes a copular clause. The following example, like ex. (117), features the synthetic form of the nucleus, the apocopate:23 ́ dīle mutwa (119) bıyaʾlıt go kullu ōde hīl go ʾōdıt arbi ūxā ́ psōx xā sandūq

go qurnīsıt čappe ‘You will enter all of the rooms up to the forty first room, open a chest that is placed in the left corner’ 107

The third person of the copula is fully illustrated. Note that the copula dīl- is often close to the nucleus, but need not follow it immediately (e.g., ex. (117)). This attributive copula is much more common following a pronominal nucleus: (120) ō gōra dīd dīle go namūsīye ‘the man who is in his bed’ 562

the difference between subordination and attributive constructions see the following section. 23 Many substantives which have this ending -ūCa show a stressed apocopate ́ : šūla ‘matter, job’ vs. šūĺ ; palgūsa ‘middle’ vs. palgūś , etc. form -ūC



This is the protoype, and it can be compared with (approximately) the same syntagm, which contains, however, another copular form, namely, wḗwāle: (121) ta daw gōra dīd wḗwale faqīr ‘to the man who was poor’ 34

The comparison yields the result (discussed above with regard to attributive DETs) that the d in dīle is inseparable, and that that the attributive copula is parallel in its entirety to wēwāle, which has no special attributive status alloform. (122) axni gıbax āzax ıl xā bāžer dīd dīwax kırye gāwa ‘We want to go

to a town in which we are hired laborers’ 53

(123) ūdımma kāweš rıš daw gōra dīd dīle pīša kēpa ‘and its blood is

flowing on the man who had (lit. is) become stone’ 102–103

(124) rūwānıt ḥukūm dīdi ddīlu tīwe bıd yamme ūčappe dīdi ‘The

important (ones) of my regime who are sitting on my left and right’ 798–799

Note that ex. (122) contains the 2nd person of the attributive copula. There are other cases, e.g., appositive content clause (see §4.2 below): (125) ūmay–la sıbbe ddīwıt hatxa pēšt-rāṣt bıt šūlox ‘But what is the

reason that you are so sure in your work?’ 514

The clause (ddīwıt ... šūlox) is not an adjectival, but rather a substantival clause, which indicates the contents of the substantive sıbbe ‘reason’. Another example is of a substantival dīd clause: (126) ay dammıt ġzēlu ʾōdıt ḥakōma dīd dīle ʾaw najāra črīpa mın zdoʾse

mírrūle ‘When the king’s servants saw that the carpenter is startled from his fear, they told him:’ 717–718

Nevertheless, in both cases, the substantival nucleus (dī)d- takes the attributive copula. Adverbial nuclei also occur with the attributive copula: (127) yımkın mın hīngıd dīla hwīsa lēwa xıpta ‘perhaps from the time

she was (lit. is) born, she has (lit. is) not bathed’ SAG 5.58

hīnge ‘then, at that time’ is an adverb of Kurdish origin. In ex. (127) it is marked as nucleus and followed by dīla. The following group of examples contains the attributive copula following an adverbial nucleus: (128) halli ḥaqqe kmad dīle ūbāzın ıl bēsi ‘give me my due as much as it

is and I will go to my home’ 669–670



(129) čımmād dīle ṣāx ‘as long as he is well’ (857, 865) čımmāt́ dīwıt

ṣāx ‘as long as you are well’ (409)

(130) čukūn dīwıt qarīwa llıbbi ‘since you are close to my heart’ 815–


(131) mōrımle sēpe tqāṭıllu kutru muġ dīlu šṭīḥe mızġas ‘He lifted his

sword to kill both as they are lying down together’ 827

(132) atta ptafqi ʾıbbu dīddīlu rıž dāra ‘Now they will find them when

they are on the tree’ 269–270

In view of these data, we can sometimes infer the existence of a nucleus from examples where the attributive copula is attested: (133) uʾatta la kīʾēn hakan dīlu ṣāx basri ‘And now I do not know

whether they are well after I left (lit. after me)’ 600 (similarly 720, 731 and 790)

These examples tell us that (ha)kan ‘whether’ is a nucleus. However, there are exceptions, i.e., not every nucleus is followed by dīle, and rarely, even when we are certain it is a nucleus, we nevertheless find, especially at some distance from the nucleus, the regular allotagm, Rīle. Compare the preceding and the following examples: (134) ūlákīʾın hakan kōma–le ya xwāra–le ‘But I do not know whether

it is black or white’ 660

One could say that the language still hesitates as to the nucleic nature of (ha)kan. But take a look at the following pair: (135) ukīʾēwa dīd ʾō šūl nāṭōre dīde hīle ‘and he knew that this affair

with his guards’ is his own’ 642

(136) uxa lákīʾēwa dʾanya trēʾ baxtāś dō gōra–lu ‘But nobody knew that

these two are wives of the same man’ 772

By now, there is no doubt that (dī)d is a pronominal nucleus. Nevertheless, the attributive copula is not used following them in exx. (135) and (136). There are other exceptions, some of which are quite problematic: (137) ay ʾurxa dīla msukarta mın qam mšalxāne ūganāwe ‘the road

which is closed because of robbers and thieves’ 30



In this example, dīla follows an explicitly unbound substantive, after which our attributive copula is not called for. There are more examples of this type: (138) ʾāna gıbın madʾrıtti ta bābi uyımmi dīlu sawōne ulatlu ču xa mjīt

mınni ‘I want you to return me to my father and mother, who are old and have no one except me’ A 5.8

Here too, yımmi is unbound. One way to look at these examples is to reason that ddīl- (that is d+ dīl-, which is what we expect here, based on most of the occurrences, typified by ʾē baxta ddīla baxte ‘that woman who is his wife’ [850]) is occasionally simplified phonetically into dīlu. This ‘simplification’ occurs, albeit rarely, with pronouns and DETs as well: (139) hama ʾapāni dīlu š-kapānax dı mbārıxlu lxawxıt! ‘They too, who

are on your shoulders, just marry them off to each other!’ SAG 7.28

(140) kıs daw dīle tīwa qam ṭayyāra ‘to the person who was sitting near

the plane’ SYG 4.8

(141) ʾan dīlu ʾısye mqabıl mınnan ‘those who came before us’ SYG


This could perhaps be ascribed to a later phase of the language (ex. (138), coming from Avinery’s texts) or to a different genre (exx. (139)–(141), coming from two distinct personal experience narratives, told in spoken language). It needs to be mentioned, however, that the second of these texts (SYG), has more examples of the first type (that is, attributive dīle e.g., exx. (116) and (117)). Another way to look at these occurrences is to infer the existence of a nucleus ø. This option is discussed below, §2.3.1. The allotagm wēle-R, discussed above under §, is found in subordination, but it is not synonymous with a clausal attribute: the latter is always subordinate, whereas not every case of subordination is a case of a clausal attribute (see presently). So much for clauses. There is however a group containing a non-clausal nexus (treated under §1.4.3 above) in attributive status: (142) qbılle zērıngır ūhulle kmāḍ ṭḷıble mare nunīsa ūšqīlale nunīsa bıt kēfe

ʾısya ūzılle lbēsa ‘The goldsmith agreed and paid what the owner of the fish asked and took it happily (lit. with his joy come) and went home’ 727-728



This rare nexal group, which is not a clause, occurs only after the preposition bıd.

2.3 ATTRIBUTIVE COMPLEXES The focus so far has been on each of the components of the attributive construction, viz. nucleus and attribute seperately. In this section the syntactic status of the entire combination is treated. This is not always self-evident, for instance, a pronominal head and an attribute combine at times into an adjective and at other times into a pronoun. The following table is reproduced from §2 above. The various combinations associated with the attributive relationship are presented in full: attribute⇒ nucleus ⇓






‘possessive’ attribute


syntactic adjective

substantive+clausal attribute

possessive or pronominal adjectival, adverbial adjective; pronouns and substantival clauses


prepositional inflecting preposition syntagm

adverbial clause

This section deals with the more important resulting combinations, as far as their internal syntax is concerned. Their external environment is dealt with where appropriate. The attributive complex is probably the main strategy of subordination in JZ—the principle unifying all the cases is a nucleic entity preceding an attributive clause. There are one or two additional strategies, which are mentioned in passing in this chapter, but are described more fully elsewhere: the asyndetic substantival (§§2.3.1,, adjectival (§4.6.3) and adverbial clauses (e.g., § There is yet another type, containing a head-like exponent, but the relationship obtaining between this exponent and the rest of the clause is not evident (e.g., ınnu as well as indirect question exponent).

2.3.1 Substantival complexes Substantival clauses as well as appositive content clauses (e.g., ‘the fact that…’) are headed primarily by ınnu (a specialized mark for substantival clauses in general) and d(īd) (which is not). d(īd), otherwise a pronoun, is



nevertheless the nucleus of the construction (see, in addition, ex. (126) above): (143) hādax žġīla bıt taxmīne hīl la rʾıšle ddīla ḥmılta mılʾḗl mınne ‘... so

preoccupied in his thought that (lit. until) he did not feel that she is standing above him’ 902

Ex. (143) is a clear-cut substantival clause. Compare the following pair, this time with the conjunction ınnu following the same verbal form, rʾıšle: (144) wēla basır xakma yōme rʾıšle ʾınnu bıštum īle ‘After several days he

felt that he is better’ 312

(145) ōha ġulāma rʾıšle ʾınnu ḥakōma gıbe nāʾısle brōne ūšāqılla bax brōne

‘The servant felt that the king wants to sting his son and take his son’s wife’ 94–95

Whereas d(īd) (ex. (143)) is the nucleus of an attributive construction, ınnu cannot be proved to be one: the regular, rather than the attributive copula in ex. (144), testifies to that. That is the only option attested with ınnu).24 These are substantival clauses (“abstract relative clauses”, in Polotsky’s early terminology (Polotsky 1944)). With respect to the external syntax of these clauses, they commute with the infinitive, at least in principle.25 The other related term, “concrete relative” (‘the one who, the thing which’), mostly belongs, so it seems, with adjectival complexes. Another instance of a substantival attributive construction is the topical syntagm #[mād]NUC+[copular expression+substantive]ATTR#: (146) ıswa xa gōra gımrīwāle šēxe bardāne. mād dīle ōha šēxe bardāne

kīʾēwa bıd ṣurrıt nāše ‘There was a man, one would call him ŠB. As for this ŠB, he knew peoples’ secrets’ 226

(147) ʾíswa xá xwāja ʿáli ... mā-dwḗle xwāja ʿáli, ʾíswāle xá brṓna ‘There

was one (whose name was) Mr. ʿali... as for Mr. ʿali, he had a son’ SS 1.2

Actually, ınnu is more complicated than that, and is possibly compatible with the presentative copula as well, in which case it does not even subordinate. See §, following ex. [80]. 25 E.g., with the verbal lexeme yzāʾa: basır yāla xapča fhımle ūzeʾle mahjōye ‘after the child understood a bit and knew (how) to pronounce...’ (580). 24



The external syntax of topical constructions is dealt with under §§5.1.2 and 6.1.1. Yet another group of clausal substantival attributive constructions involves appositive content clauses: (148) mın dḗ xaṭa

rabsa dīd dmıxle ʾımmıt baxta marē-gōra ‘...of this great sin (that is,) that he slept with a married woman’ 787 +

(149) ūdırye wēlu rāyi ūtakbīr lxawxēta tčuxxa mınnu la gāwır baxta ṭūl

ʿumre ‘...and they had made with each other a plan (that is,) that no one of them shall marry a wife his entire life’ 857

(150) bale ʾíswālu naṣīb dīd nāš karwan la mōnıxlu ıl ču ʾāla ūla lʾēl ‘but

they had luck (that is,) that the caravan’s people did not look anywhere nor up(wards)’ 270

All these examples have a substantive whose content is given via a clausal substantival attributive construction. These are not adjectival clauses (the substantive plays no role inside the clause, as it would in adjectival clauses, see §2.3.2). These clauses can be marked by ınnu (whereas adjectival clauses cannot): (151) bıyāwaxlox nīšan ʾınnu mxōlíplōxun bıt ʾ asıqyāsa ūbıt kaffīye ‘We

will give you a sign that you exchanged rings and scarves’ 310

This type is described under §4.2 below. Other clausal entities, which are not in themselves attributive constructions, may function as substantival complexes. First and foremost is the attributive clause itself (that is, just the clause which follows the nucleus, which is the syntactic equivalent of a substantive; see §2.2.4 above, where the external syntax is discussed as well) which is substantival, but does not in itself constitute an attributive construction; it is merely a part of one. Other such clausal entities are indirect questions and asyndetic subject and object clauses. Both groups are mentioned in passing, their external syntax is discussed in § Indirect questions are compared above (end of §2.1.1, exx. (15)–(20)) to similar examples where the (in)direct question signal is marked as nucleus to produce syntactic indefinite constructions. The indirect question signals (often identical in form to the corresponding interrogative signal) are hence not analyzed as nuclei. The resulting clause, in case of a pronominal indirect question (nexus questions are somewhat different in this respect) often look like embedded questions mutatis mutandis:



(152) ya axōni, ma brēle ıllox? ‘O my brother, what happened to you?’


(153) mbōqırre dıkandāre ma brēle ıl kullu ʿattāle ‘He asked the

shopkeepers what happened to all (the) porters’ 571

The differences are mostly in the analysis: in ex. (152) the interrogative pronoun ma is the rheme of the clause, representing the new information yet to come, whereas the indirect question signal ma in ex. (153), although originally pronominal, is not the rheme, since it is more of an exponent to introduce and mark the indirect question. Such exponents tend to carry little information of their own. Asyndetic substantival clauses are not considered to be attributive constructions. Asyndetic clauses can, at least in principle, be analyzed as having a conjunction ø (or nucleus ø). This option needs to be examined. The first thing to be discussed is the question whether an attributive entity can occur without a material nucleus. Since the attribute is positively marked in JZ (e.g., by the attributive alloforms of the DETs), one could check whether we ever come across such an attributive alloform without such a material nucleus.26 The answer is negative. It follows that the nucleus of an attributive construction cannot be ø. This statement can be extended to clausal attributes by induction. This could be shown in the following pair of examples: (154) ġzēlu ʾōdıt ḥakōma dīd dīle ʾaw najāra črīpa ‘... the king’s servants

saw that the carpenter is startled ...’ 717–718

(155) kīʾın gabāra wıt ‘I know you are a hero’ 954

In ex. (154), which is a reduced version of ex. (126) above, we find a substantival object with dīd as nucleus and a clausal attribute marked by dīle. In ex. (155), on the other hand, the object clause has no overt nucleus, and in turn, the copula occurs in its non-attributive form. However, the actual picture is more complicated. We have seen above, in exx. (137)–(141) (§2.2.4), that there are exceptions to the generalization regarding the use of the form dīl-. The examples could all be explained away 26 English, for example, allows a free genitive: 1. And Bush’s is not a single-issue problem ( =1549959); 2. In that sense his appeal on issues can be said to be broader than Bush’s ( wisconsin_poll_040922.html).



by positing a ø nucleus (e.g., ex. (137): ay ʾurxa ø dīla msukarta ... ‘the road (which) is closed ...’ [30]). All these exceptions, however, are adjectival clauses, and not substantival clauses. These five exceptions do not seem to be enough to change the entire picture, and the explanation given above, viz., that dīle in these exceptions is the phonetic outcome of the far-morecommon d-dīle seems to be in conformity with what the vast majority of the attributive constructions in JZ tell us.27 These considerations lead us to state that there is no nucleus ø with clausal attributive. This conclusion leads to the following analysis: matrix verb (156) xšūle...

substantival clause nucleus attribute dīd

sēlu ʾōdıt ḥakōma

(157) xšūle

xūwe mpıqle

(158) xšūle

ʾınnu xulma–le

translation ‘He thought ... that the king’s servants came’ 717

‘He thought the snake came out’ 813

‘He thought that it is a dream’ 363

Ex. (156) represents the attributive construction which functions as an object clause, whereas ex. (157) is an asyndetic object clause, which, unlike the former example, has no nucleus. Ex. (158) is analyzed similarly, although it is marked by ʾınnu, and is hence not asyndetic. The relationship between ʾınnu and the clause following it are not regarded as attributive. Another related and important type of a substantival attributive construction is the one containing a substantival nucleus and a clausal attribute: (159) ʾıl xabır mxēla baxte ‘the word(s) his wife said’ (683); nāš hōya

ʾēne kpınta ‘a(ny) man (whose) eye is hungry (822); šūl muḥkēla ʾılle ‘the matter she spoke about’ (778); gōr lḗ ʾaxxa ‘a man (who) is not here’ (379)

To these we can match examples with the analytic nucleus form:

G. Khan suggests (P.C.) that Polotsky may have developed a convention of transcribing the constructions which made the ‘normal’ nucleus form appear more frequent than it really is. However, Polotsky transcribed these folktales while hearing them for the first time (there was no tape recording involved), and it hardly seems that he would have much time for such analysis. The indirect evidence for this view is that the transcription ddīle is not consistent throughout. 27



(160) xabrıt mírrēla ‘the word(s) he said to her’ (859); ʾaw nāšıt pāyıš

gōr tre ṭḷā baxtāsa ‘the man (who) is the husband of two (or) three wives’ (838); šūlıt ʾuzle ‘the job he did’ (455); aw gōrıt ʾaqle ṣaqat wēlu ‘the man (whose) feet are crippled’ (55)

These examples are translated, in the absence of a better alternative,28 as substantives followed by adjectival clauses. However, the analysis of the actual examples clearly shows that there is no adjective clause here, but rather a substantival nucleus and its clausal attribute which is rather, as already stated above, the syntactic equivalent of a substantive. Despite the differences, this construction and the adjectival clauses share a similar domain. The adjectival clauses with d(īd) (described in the following section) are much more common in fact. The clausal attribute following a substantive (as in adjectival clauses) bears some kind of relationship with the substantival nucleus, or in other words, the latter, except for being the nucleus of the construction, has some function inside the attributive clause: (161) ē naqla zılle gōr lḗ ʾaxxa ‘This time he is doomed (lit. went) the

man (who) is not here’ 379 (see also exx. (2) and (25))

Like adjective clauses, the described entity is never explicitly repeated in the clause, but in certain cases it is resumed by a pronominal entity. For instance, in the first example, the entity which is marked as nucleus (gōra) also participates as theme in the attributive clause (lē axxa ‘(he) is not there’), the negative copula lē shows 3MS which pronominally resumes, by way of agreement with gōr, the nucleus. In the next example the nucleus also participates as the object in the attributive clause: (162) qam bāqırru hakān ʾībe maḥke ʾımmıt daw gōr duqlu ‘He asked

them whether he can talk with the man they caught’ 613 (also 512, 606)

The objecthood of gōra is merely implied, there is no explicit resumptive pronoun (this absence might be a matter of chance, due to the relative paucity of examples—in the related adjective clauses such resumption The famous Akkadian parallel of these constructions awāt iqbû ‘the word he spoke’ (Laws of Hammurabi §3) was literally translated by H. Schuchardt ‘das Wort seines Gesagt-habens’ (cited in Goldenberg 1993:298). 28



occasionally occurs). However, when the nucleus has attributive function in the attributive clause, it is always resumed pronominally:29 (163) āna wēna aw gōrıt ʾaql-e ṣaqat wēlu ‘I am the man (whose) feet

are handicapped’ 55 (see also ex. (25))

The attributive 3MS pronoun -e, occurring with ʾaqle in the attributive clause, resumes the nucleus, gōra. This resumption is the only one truly required, whereas the object resumption is optional and thematic resumption occurs in some kind of default (the predicative forms, viz., verbal forms and copulas, obligatorily resume the theme, at least in principle). Adverbial function is usually left unresumed: (164) mın waʿıd ʾēni ṇẓırra ʾıllox ‘from the time my eye looked at you’


The function of waʿda inside the attributive clause is merely inferred. In the following example there is no resumption to the locative idea subsumed in dūka: (165) ūtxırra dūkıt wēla mın qabel waʿda kırya ‘and she remembered the

place (in which) she was a short while before’ 308

The function of dūka ‘place’ in the attributive clause may be either adverbial or rhematic (she was in a place). The last example in this group has dūka as nucleus. This time, however, it is resumed by the adverb tāma: (166) qımle ō jōtıyara ū zılle rıž daw gırıt kēpe dūkıt šqılle ʾay baxta mın

tāma ‘So the peasant walked on the summit of stones, the place he took the woman from (lit. from there)’ 504

tāma ‘there’ is not quite a pronoun, yet it is similar to one in representing a place name. The following table summarizes the resumption data pertaining to the function of the nucleus inside the attributive clause following a substantival nucleus:


In conformity with the results obtained in Keenan and Comrie 1977.




clausal attribute

nucleus function in the clause resumption


lḗ ʾaxxa



ʾaqle ṣaqat wēlu



dūkıt dūkıt

duqlu wēla




šqılle ʾay baxta mın tāma adverbial

+ –


Bearing a resemblance to the foregoing clausal attribute are numeral attributes. As already mentioned (§2.2.1, ex. (91)), ordinal counting in JZ is not effected (except for qamāya ‘first’ and xarāya ‘last’) by adjectives, but rather by an attributive construction with either a substantive as nucleus (which belongs to this section) or a pronoun (belonging to the following section). The conceptual difference obtaining between adjective clauses and the clausal attributive following a substantive is exactly the same as between the ordinal adjective and the attributive construction with a numeral attribute: clausal attribute

substantival attribute

substantival complex

appositive adjectival complex

gōr trḗʾ ‘(the) second man’ (lit. man of two’) (561)

gōra = qamāya ‘the first man’ (841)

gōr lḗ ʾaxxa ‘(the) man (who) is not here’ (379)

ō gōra = dīddīle go namūsīye ‘the man who is in his bed’ (562)

Further details about these constructions and the interrelations between them are revealed in the following section as well as in §4.1. Attributive complexes whose nucleus is māt30 usually function as nominals, but there is some hesitation whether these complexes are substantival or adjectival. Their internal syntax, viz. the pronominal nucleus, makes it, syntactically speaking, an adjective-like entity (viz. ‘that which happens’ etc.). However, there are no minimal pairs with simple adjectival entities, and their external syntax rather has to do with substantives such as xabra ‘word’, qıṣṭa ‘story’, maqaddır ‘event’, ʾıjbōna ‘desire’, takbīr ‘plan’. The function of mād in the clausal attribute is either theme or object:

30 One case only of māt is followed by a noun: mēsun māt faqīre ūhál(l)ūlu ʾīxāla ‘bring whatever poor and give them food’ (225). Incidentally, in that same place, kmat precedes a noun: ūʾōzūn kmat mıṣwāye dīd ʾīs go lıbbōxun ‘and do according to the (or: whatever) charity which is in your heart’.



(167) tu ūmaḥkınnox māt sēla brēši bıt waʿdet wēli jwanqa ‘Sit and let me

tell you what happened to me when I was young’ 372

māt ‘that which’ is the nucleus and functions as theme inside the clausal attribute. The syntagm māt sēla brēši lit. ‘what(ever) came in my head’ occurs where a substantive of similar content (rather than an adjective) occurs: (168) ūqam maḥkēla kulla qıṣṭa ‘and he told the entire story’ 258

Another such pair would be the following: (169) zōnın ṭālax ūta yalunke māt ʾājíblōxun ‘Let me buy you and your

children what(ever) you want’ 446

(170) hīl marāda maḥıslıttu ‘until you fulfill (your) wishes’ 821

That is, māt ʾājíblōxun, may be represented both syntactically and semantically by the substantive marāda ‘wishes’. In the following examples the function of mād in the clausal attribute is object: (171) mād gımret kulla ṭrōṣa–la ‘What you say is all true’ 358

The syntagm may be represented by the substantive xabra ‘word’, rather than by any adjectival entity such as the participle: (172) ūxabre dīde musēlu ṭīme ūqam ġālıbla ‘But his words made sense,

and he won her over’ 884

The syntagm in the last example is interchangeable with takbīr ‘plan’: (173) āna txırri māt drḗwāli go lıbbi ‘I remembered what I had planned’


2.3.2 Adjectival complexes This section is devoted to adjectival attributive complexes with a pronominal nucleus.31 The link between adjectives and such adjectival complexes is briefly mentioned under §1.2.4 above, and is discussed at length below, under §4.1. It is important to mention that not every complex attributive construction whose nucleus is a pronoun is deemed adjectival. We have seen above that such constructions can be substantival clauses or complexes (e.g., see māt at the end of the previous section), and we further show below that Adjectival attributive complexes with an adjectival nucleus are described above, §2.1.4. 31



these complexes can also be pronominal as well as adverbial. Their internal constitution can only partially determine their ultimate function; it is rather the external syntax, viz. their syntagmatic and paradigmatic relationships, which actually determine their function. The pronominal nuclei d(īd) and kud are at the center of this section. Other, more occasional pronominal nuclei are manīt ‘whoever’ (exemplified above, exx. (14) and (18) §2.1.1) and ēmıt ‘who(ever)’.32 Adjectival clauses are thus named because they are analogous both internally and externally (§4.1) to simple adjectives. The term ‘relative clauses’ is not used because JZ (as well as Semitic in general) does not have relative pronouns, that is, pronouns that are capable of incorporating simultaneously two functions—representing both the antecedent in the main clause and showing its function in the relative clause. dīd (or any other pronominal nucleus for that matter) cannot show the function of the antecedent inside the clausal attribute, which is effected by selective resumption of the antecedent. dīd occurs in non-clausal adjectival complexes. What it can do is nominalize the entity in the attribute, make it syntactically suitable for adnominal expansion: (174) ūšımmıt ay brāta dīd go mušruq ‘the name of the girl in the east’


(175) aw gōra dīd go māya ‘the man in the water’ 537 (176) sēle ıl bāžer dīd gyāne ‘he came to his own town (lit. town of

himself)’ 538

(177) narʾa dīd qṭāʾıt dāre ‘an axe to cut trees’ 633

In both exx. (174) and (175) a prepositional syntagm, which does not normally function adnominally,33 is marked as attribute. The entire syntagm is this way naturally adnominal. In ex. (176) the otherwise reflexive pronoun

Compare the almost identical story endings: 1. ʾḗmıd šmiʾāle xāye, uʾḗmıd la-šmiʾāle xāye SS 66 2. kud šmīʾāle xāye, ūkud la šmīʾāle xāye 159 Both have the same translation: ‘whoever heard it (=the story) may he live, and whoever did not hear it, may he live (as well)’. 33 A prepositional syntagm does, rarely, join the substantive as a direct attribute, See §2.2.3 above. 32



gyān- can occur only as an attribute34 if it is to expand a substantive. Ex. (177) illustrates an infinitive used adnominally by being made an attribute whose nucleus is dīd. Adjectival clauses whose nucleus is d(īd) constitute the majority of these adjective clauses. d(īd) represents pronominally the described entity and is also the nucleus of the attributive construction. The entity represented by d(īd) (the antecedent) has some function in the attributive clause—it may be a theme, an object, an adverb or an attribute, and the examples are organized according to this principle. When the antecedent participates as a theme in the attribute clause, it is never explicitly repeated, but on the other hand, it is always represented by the subject index in the predicative form: (178) ʾanya qazānatCP dīd laktōri3CP ūla(k)kēmi3CP ūla gımjángıri3CP ‘the

pots which do not break and do not become black and do not rust’ 469

The verbal forms show 3CP, which represents, in the verbal form, qazānat. The same happens in the following example: (179) mōnıxle bıd dūmıkeMS dīd sēle3MS qṭāʾa ‘he looked at his tail which

got cut’ 454–455

The antecedent here is 3MS, which is duly represented in the verbal form (note that the verbal complex is passive). The antecedent can be represented by an independent pronoun, if the need arises: (180) ūsēle ʾaw sımsar dīd wēle ʾapāwa ʿızīma ‘and the matchmaker who

was, he too, invited’ 574

In ex. (180) the antecedent is represented in the form wēle (3MS), as well as by the focal personal pronoun ʾápāwa. The adjectival syntagm as a rule immediately follows the antecedent. There have been found, in this group, two exceptions—one, when the adjective clause follows an attributive construction but actually refers to the nucleus in the latter: (181) fēkıt ḥakōma dīd dīlu go karmāne dīde ‘the fruits of the king

which are in his vineyards’ 602


E.g., narʾıt gyāne ‘his own axe (lit. axe of himself)’ (680).



The adjective clause expands fēke (PL), as is clear from the 3CP index in the copula. The second exception is when a rhematic substantive is expanded, in which case the adjectival clause cannot precede the copula and hence follows it, rather than the antecedent: (182) āna bronōxūn wēn dīd zaʿli ‘I am your son who got lost’ 605

Note, in addition, that the verbal form inside the attribute is in full agreement (1CS) with the thematic pronoun āna. This entire group is also characterized by verbal forms or copular expressions that immediately follow the nucleus. The adjective clause occurs, like a simple adjective, without an explicit antecedent as well; still, like a simple adjective, such adjective clause represents an entity: (183) ʾay dgıbáwāle ‘theFS (one) who wantedFS him’ 825 (184) aw dla xṭḗwāle hēš ʾımmıd day baxta ‘theMS (one) who had not yet

sinned with the woman’ 892–893

(185) ʾan dıkṭāʾe ʾılla ‘theCP (ones) who look for her’ 885

The DETs ay, aw and an are neither pronouns nor do they function as such, and hence cannot be suspected to be antecedents. In all these examples, what we have are in effect determined complex nominal syntagms. When the antecedent has the function of attribute in the clausal attribute it is mostly resumed by an attributive pronoun: (186) mın day wēša dīd kúšwālu gāw-a ‘’from that dry land into which

they decended’ 351

(187) qımli šqılli aw šabūza dīd mṭōwēle ʾıbb-e xūrāsi ‘So I took the

skewer with which he roasted my friends’ 390

(188) ʾāqılte mṭēla hīl qam ṭūra dīd wēli mṭušya basr-e ‘His foot reached

the mountain behind which I had hidden’ 393–394

(189) rīš kursi dwēla tūta rēš-e ‘on the chair on which she was sitting’


The function of attribute is clearly not only possessive; it occurs also whenever the antecedent’s function in the attributive clause is to be the attribute of a prepositional phrase, whether adverbial (as in the preceding examples) or valency-related:



(190) ūtrḗ sūsıwāsa dīdi dīd ʾāna grakwen ʾıllu ‘and two of my horses

which I ride’ 601

There are, however, examples where the resumptive attributive pronoun does stand for possession: (191) ay brāta dīd qam xazēla ṣurt-a ‘the girl whose picture he saw’ 89

The attributive resumptive pronoun can also be suffixed to dīd-: (192) dalīsa dīde dīd drēle ta nāṭurūsa dīd-a brōnıt tre ‘his vine on whose

guard he put his second son’ 651

A subgroup that needs some consideration is where the antecedent (explicit or not) has ‘dative’ function inside the attributive clause. Dative is not a separate syntactic status, but rather expressed by a prepositional syntagm whose nucleus is the preposition (ı)l-. This is why it is dealt with here, since the resumptive pronoun in the attributive clause is always an attributive pronoun as well:35 (193) kut ḥakō ma dıʾıtle

him) a hero’ 755


pālavan ‘each king who has (lit. there is to

The predicative complex ıtle ‘he has’ (lit. ‘there is to him’, § contains the prepositional syntagm l-, whose pronominal attribute is the resumptive pronoun. In the following example -le is the second object of hulle ‘he gave’: (194) ʾaw dʾīla hullēle tānj uʿāqıl ‘the (one) to whom God gave crown

and wits’ 800

Note that there is no explicit antecedent, and that the theme precedes the verbal form in the attributive clause. So far, whether the function of the antecedent in the attributive clause was theme or attribute, we have seen that resumption always occurs. In the following groups it does not. When the antecedent has the function of direct object in the attributive clause, explicit resumption is not very common. It is not visible when the preterite in the clause follows a definite MS entity: (195) bıd daw nuqba dīd xpır(-ø-)ri ‘in the hole which I dug’ 405 35 The thematic pronominal indices in the preterite as well as the object pronomina of the subjunctive base, both looking like a prepositional syntagm whose nucleus is l-, are not considered in this framework prepositional syntagms, but rather as pronominal entities, of nominative or accusative status.



In other cases it is easier to tell whether there is a resumptive pronoun or not: (196) pāre dkšaqlın bıt parrōke dīdi ‘the money which I take for my

textiles’ 791

(197) ʾan gūre xēt dīd mjōmıʿle ‘the other men he gathered’ 556 (198) mōʾarta wēla mın dan kapōre dīd qṭıḷḷi ‘she was better (lit.

entered) than the giants which I killed’ 412

The antecedents in all three examples are plural, two of them are explicitly definite, yet, no resumption whatsoever occurs. Compare these with the following examples, where resumption is found: (199) ʾan ṭlāha šoʾāle dīd šqīl-ī-le ‘the three things that he took’ 701 (200) mjıd ḥammaš dīde dīd gımṭāšḗwā-le mın qam dáw rakāwa ‘except

his book which he would hide from the rider’ 587

(201) ʾaw dʾıban-ne ‘theMS (one) whom I desire’ 800

These are harder to come by. Besides these object pronouns, we also find, in case of an a analytic verbal form, the pronoun dīd- (for which see the following section) representing the antecedent as object in the attributive clause: (202) buhrāyıt dunye dīd la wēli ġzīsa dīde šōʾá š ınne ‘daylight which I

had not seen for seven years’ 705

(203) ʾan jullıt ḥakōme dīd ču nāša láwēwāle xızya dōhūn ‘the king’s

clothes which no one had seen’ 527

Note that in ex. (203), besides the resumption of the antecedent by dōhun, the thematic substantive (ču nāša) precedes the verbal form. Adverbial syntagms can also occur in that place: (204) bēsa dīdi dīd bıd ʾīzi qam bānınne ‘my house which I built with

my own hands’ 220

Adverbial function of the antecedent in the attributive clause often occurs without resumption: (205) ıl xa dūka dīd ıswa xakma nāše ‘to a place (in) which there were

several people’ 360

(206) lday dūka dīd wēle tımmal ʾımmıt bābe ‘to the place (in) which he

was with his father the day before’ 449



(207) bıddaw waʿda dīd tūle yāla xazır ḥakōma ‘at the time (in) which

the child sat with the king’ 596

In other cases, which are brought above, a prepositional syntagm with an attributive pronoun performs the resumption: (208) go day bāžer dīd gımzabnīwa gāwa ‘in the town in which they

would sell’ 51

Only very rarely indeed do we find a true adverbial resumption (209) ıl day bāžēr dīd dīle bır dō tıjjar skīna tāma ‘to the town in which

the merchant’s son lives’ (lit. ‘which... there’) 353

The other pronominal nucleus which forms adjectival complexes is kud. Unlike d(īd), it is very limited in that its function inside the attributive clause can only be thematic: (210) kud šāqılle, āwa pāwe ‘Any(one who) takes it, will be him’ 161

That the attributive complex with kud is adjectival can be deduced from its exchangeability with the rare active nomen agentis (the pattern CaCāCa): (211) šaqālıt kısta dīd pāre ‘the taker of the purse’ 284 (212) kud ʾāse mın tāma ‘any(one who) comes from there’ 536

(compare ʾasāya 855)

(213) ʾurxıt ṭḷāha kud āzel ıbba ‘the third road, any(one who) walks

it...’ 193 (vs. ʾazāla 855)

The adjectival kud syntagm does not have an antecedent, yet, it represents pronominally a generic entity. Like the asyndetic substantival clauses mentioned in the previous section, asyndetic adjective clauses are viewed as adjectival complexes whose nucleus is not ø: (214) uʾāna gıbēn ṭāʾıttı xa xamsa hōya mın mıšpāḥá bāš ‘I want you to

look for a maid (who) is from a good family’ 541–542

(215) šrēla mın nāfṭanga xá mındi gımšāpēwa lxarxāṣa ‘She took off her

loins something (which) resembled a belt’ 833

́ ‘They did not find one (whose) (216) la ġzēlu xa šımme hāwe qaramān name is Qaraman’ 124

Most asyndetic adjective clauses, which are not very common, describe an indefinite entity, whether a substantive (ex. (214)) or a pronoun (exx. (215)



and (216)), perhaps modeled after the Arabic (this phenomenon also occurs in NENA dialects that are clearly outside the Arabic speaking area). The very rare exception is the nominalization of a verbal form preceded by a definite DET: (217) an ksēmi mınnan ‘theCP (ones who) hate us’ SAG 3.7 (218) aw gmāẓıʿ ‘theMS (one who) loses’ SAG 5.65

These examples are found only in SAG, and do not occur with an explicit antecedent. The external syntax of such clauses is described in §4.1. They usually reflect some kind of genericity, referring to a group.

2.3.3 Pronominal complexes Few attributive constructions form a pronoun: kut+xa ‘each one’, kutran (kut+tre+attr. pron.) ‘us two’ (as well as the analogous kud ʾarbu ‘the four of them’ etc.), and dīda (dīd+attr. pron.). The latter is a pronoun, rather than an adjective (it hardly ever occurs with DETs, for example),36 which is clearly seen when it is the object pronoun of analytic verb formations, since in this context it commutes with substantives, e.g., bıd ġzāya dīda ‘seeing her’ (701) vs. bıġzāya xılma ‘seeing a dream’ (871), where both dīda and xılma take the substantival slot of the object.

2.3.4 Adverbial complexes This section deals with attributive complexes that have an adverbial function. Unlike the adjective clauses, which show complex co-reference between antecedent and its resumption, the relationships here come down to an attribute, whether simple or clausal, which expands the adverbial nucleus. Since the external syntax and the values of these construtions are dealt with below (§3.2.2), this is merely an overview. The various complexes are grouped according to their nucleus and its function: purpose: ṣapxāṭır ‘for, in order that’ and ta ‘to, in order that’, commute with dīd (la). All these signals are followed by the subjunctive form. ṣapxāṭır and ta also



ay dōhun mırta–la ‘(so) one of them has said’ (933), where it is used as an ad-



come with substantival or pronominal attributes in roughly the same value. lakūn ‘lest’, borrowed from the Arabic, is arguably a nucleus. result: hīl ‘until’, hadax d(īd) ‘so that’, and occasionally d- ‘that’, all commute with each other when denoting result. In this group only hīl takes a substantival attribute. indefinite quantification: mād ‘as much as, so much’, kmad ‘as much as’ and čımmāt́ ‘as much as’, all take only a clausal attribute. comparative: mux ‘as (if), like’ (modally colored), qčın ‘as, like’, (ı)lmūjıb ‘as, according to’ and kudax ‘as’ are all nuclei, all occurring with clausal and substantival nuclei, to the exclusion of kudax which occurs only with a clausal attribute. temporal: mqabıl ‘before’, basır ‘after’, hīl ‘until’ and mın ‘from, as soon as’, occur with substantival or clausal attributes, whereas ay dammıt ‘when’, čımmād ‘as long as’ and hīl la ‘as long as ...not’, occur only with a clausal attribute. causal: čıkun, čınki and mādam ‘because’ are arguably nuclei, being all borrowings from neighboring languages, all occurring with a following clause. This topic continues under §3.2.3 below.

3. COMPLETIVE RELATIONSHIP The completive relationship comprises the relationship of both objects and adverbials to the nexus. It is more complex and less tangible than the previously described relationships (viz., the predicative relationship in the first chapter and the attributive relationship in the second), and actually, to this day nobody has undertaken the task of describing it fully. In the ancient Semitic languages, where one still had cases, this relationship was made more transparent by the identical marking of both functions by the accusative. In JZ this relationship is less apparent. Yet, what is nevertheless common to both object and adverb is their relationship to a nexus or a reduced form thereof, that is, a relationship between an entity and a relationship (the nexus). What illustrates this link between objects and adverbials is the frequent difficulty, in certain cases, in deciding whether a given element is an object or an adverbial complement (occasionally termed, together, complements, or goals). This complex chapter, as befits a complex variety, describes on the one hand the various aspects of verbal valency—different complementation patterns and strategies and accordingly various object paradigms; and on the other hand, different prototypes of adverbial complementation.

3.1 OBJECT COMPLEMENTS Objects have two basic forms. The most common form is the direct object, but there are also prepositional objects. The latter are not very different from the former, yet the dissimilarity is enough to have grammatical conseqences, e.g., in passivization possibilities.

3.1.1 Direct object paradigms Direct objects are structurally different from indirect objects. The latter sometimes formally resemble adverbial complements in more than one aspect. Direct objects are easy to identify in JZ, since they may be represented



by the object pronominal element in synthetic verbal forms, viz., with both the preterite and the subjunctive bases. For example qam šaql-lan ‘they took us’ and šql--lan ‘we took them’. The qam verb formation goes even farther than that; it obligatorily occurs with a pronominal representation of the object. This situation, however, is stretched even further: there is a small number of cases where verbal lexemes which would typically be considered intransitive formally behaves as transitive lexemes: štq le ‘he shut (it) up’ (877), 1rq le ‘he ran (it) away’ (893) 1rAqla ‘(that) he run (it) away’ (437). qam gam$la ‘they smiled (it)’ (470) and even bAgm$a dda ‘smiling (it)’ (867). In some of these cases there is variation between transitive and intransitive behavior (which boils down to a semantic differentiation). What nevertheless makes these lexemes intransitive (functionally, rather than formally) is the fact that these formal object pronouns do not represent, nor are they referential to, any concrete substantival entity in the text, unlike real transitive lexemes. The direct object is not limited to pronouns or substantives. It may be a complex entity, e.g., an adjective (simple or complex, see §4.1), or a nexus, which takes the position (and hence the function) of the object. For this type of object there are two varieties—non-clausal nexus (§ and clauses (§ The substantive group as direct object The substantive group, as already seen twice (as theme, in §1.2.1, and as attribute, in §2.2.1), is presented inseparably with its DETs and QUANTs. In the case of direct object, however, it is often paired up with the object pronouns. This is the reason why these pronouns are described together with the substantive. Direct object pronouns come in three sets, which are almost alloforms, according to the formation with which they occur: bases


object pronoun

form (3FS)





šql- -le


subjunctive imperative

šqAl šq!l

šqAl-X šqul-X

šqAl-la šqul-la

participle gerund

wAn šqla wAn bAšqla

wAn šqla X wAn bAšqla X

wAn šqla dda wAn bAšqla dda


The allomorphy of these pronouns is not perfect: the set of object pronouns occurring with the preterite base are capable of representing only 3rd person referents, whereas the object pronouns of the subjunctive base occur in all



persons. The adjectival/adverbial forms of the verb take the pronoun dd-, which, compared with the other two sets, shows a slightly different compatibility with a substantival object; only in few cases does it occur with one, and otherwise, the pronoun is in complementary distribution with the substantive. The common behavior of these pronouns is represented in the following example: (1)

gmAr: bbo, q&Al- -lox šše? gmAr: la, l(w)An q&la dda. gmAr: q&Al- -lox!! ‘He says: Father, did you kill Shoshe? He says: No, I have not killed her. He says: You killed her!’ LH 3.4

The first occurrence is of the preterite in a question, containing a pronoun (--) in proleptic function (being but a specific case of cataphoric representation), which is basically appositive to the explicit object. The answer has a present perfect, whose object is dda, which represents anaphorically (but is not in apposition with) the same referent in the previous clause. The response to the answer in the example is in the preterite again. This time, however, there is no mention of the explicit object following the preterite, only a pronominal representation thereof inside the preterite form. The preterite base object pronouns occur in two distinct functions, viz., cataphoric or anaphoric. These functions are elaborated below. The object pronoun of the analytic tense formations, however, usually has anaphoric function, resuming a previous mention of the object. However, there are several exceptions to this statement, which are adduced in the following examples: (2)

la hye 1awha mra dde  xabra ‘Perhaps that one has said it, this thing’ AS 1027


pAšla mand!ye dde aw tma ‘She began throwing (scattering) it, the garlic’ A 2.6

Literally translating ‘it, the garlic’ or said ‘it, this thing’ is really mechanical, and what we actually have here is fully analogous to the cases dealt below of a proleptic pronoun and substantive group which are in apposition with each other: (4)

škAlla baxte mqal!we dda ay nunsa ‘His wife began cleaning the fish’ 728


pAšlu 1anya tre bA)yra dda ay baxta ‘The three began to examine the woman’ 877


wle bAr 1am!yi psxa 1ne 'xpqa dda ay abde bAd dan julle dda šAxtne


CHAPTER THREE ‘... (and lo) my cousin has opened his eyes and hugged the negress with her dirty clothes’ 706 (compare xpqle ‘he hugged her’ 831)


&amha wlu mukme dda ay b ž7r ‘why they had blackened the town’ 604

These cases imply that, although proleptic function is not very common, the behavior of the object pronoun dd- is nevertheless very similar functionally to the bound object morphemes in the synthetic verbal forms, in co-occurring with the substantival object. With regard to the occurrence and function of these pronouns with synthetic forms of the verb, we mentioned there were two distinct functions, anaphoric and cataphoric, the difference between which has to do with whether the object is mentioned before or occurs after the verbal form. The latter is treated here, the former further below, with the pronominal group. The following examples all have the same verbal lexeme, yhwa: A. Substantival object without a proleptic pronoun 1. with




kušle go x!ra ywAl alq ta mahne ‘He went down to the stable to give fodder to the mare’ 83


mamo hulle r ba p re ta day s!tAnta ‘M. gave lots of money to the old lady’ 324

(10) halli nšanqe ‘give me signs’ 322 (11) hulle ta gyne jur7tta ‘He gave himself courage’ 492

2. with indefinite



(12) gm&lu 1anya nše ywAttu xá xz na ‘These people deserve that

you give them a treasure’ 472

(13) 'basAr hdax bywannox xa qra lre ‘And after that I will give

you a handful of coins’ 487

(14) hulle xakma q7et 7rba m&'ya &li ‘He gave a few pieces of

roasted mutton (lit. sheep) to me’ 389

(15) '1aw 1lhAt hulle ! fukAr bAywAl xa b7šf ‘And this God who

gave this thought will give a better (one)’ 466–467



The two preceding groups contain an indefinite susbtantive. We have seen, however, above (§1.2.1) that DET ø also occurs with a contextually definite substantive. So, sequences identical to those of exx. (8)–(11) may also occur with a contextually definite substantive: (16) aw ax!na šAzna šqAlle juw h7rre ‘The crazy brother took (the)

gems’ 257

The gems are mentioned previously and so are definite. In the remaining groups the substantive is definite. 3. with suffixed genitive pronoun (17) 'ta xá mAnn!x'n bAywAn brti ‘And to one of you I will give

my daughter’ 470

(18) hál'li aqqi ‘Give me my due’ 244 (19) aw jwanqa hulle 1ze ta day xamsa ‘the youngster gave his hand

to the maiden’ 545

Groups 3 and 4, despite the fact that there is no proleptic object pronoun, have a definite substantive. This is effected by either an attributive pronoun (ex. (19)) or by a definite DET (ex. (20)): 4. with a definite


(20) hulle an jullAt ak!me ‘He gave the royal clothes’ 527

B. Substantival object with a proleptic object pronoun In all of the following cases, containing a proleptic object pronoun, the substantival object is always definite, no matter what DETs precede it. 1. with



(21) '1!ha šAhra qam ywxla kAstAt pre 1ze ‘And this blind man,

we gave the money bag to his hand’ 335

(22) hál'l li ammaš ‘Give me the book’ 594

Both substantival objects are mentioned before in the text. The proleptic object pronoun marks both substantives as definite. The rest of the groups actually show double definite signals:



2. with suffixed genitive pronoun (23) uháll li xa1 xmrox ‘Give me one donkey of yours’ 917

3. with a definite


(24) hál'lli bas  gra ‘Give me just this arrow’ 524 (25) bAywíll le ha fstna ‘He will give him this gown’ 215

The table summarizes the exemplified data: Definiteness of substantival object substantive no object pronoun with object pronoun





+attributive pronoun








* save for contextually-definite plural substantives

Marking of (in)definiteness is, therefore, shared among the proleptic object pronouns and the various DETs. This leads to the logical conclusion that the proleptic object pronouns have to be included syntactically with the object substantive group despite the fact that they morphologically belong with the verbal form. Take a look at the following table, illustrating the actual object substantive group, this time of the verbal lexeme šqla: object pronoun šqAlle



The substantival object syntagm DET

ø ø ø






qam šaqAl-lu -lu šql- -le

- -





substantive pare




group gloss

indefinite (some) money definite both boxes definite (the) gems

locus 71

202 257


definite these gems



definite the money


axaws-e tawArta

definite both his brothers 80 definite the cow


The grey shade in the table above could be summarized as follows: the object substantive group is larger and has to include the object pronoun:


This is not the DET but rather the synonymous QUANT ‘one’.


object pronoun





It follows that the substantive group functioning as object is different than the one described for the theme in several details. One of which is that the current group allows indefinite substantives, whereas the thematic group does not. Another is the inseparable syntactic link between the bound object pronoun in the verb and the substantive group. Pronouns as objects In addition to the designated bound object pronouns occurring on their own (see below), one sporadically finds various unbound pronouns representing the object. The following groups of examples (to the exclusion of group 9 and 11) are exhaustive: 1. xa ‘someone, somebody’: no proleptic pronoun (26) hl la mandet x  go danya mya, bžnt bnapqet ‘Until you

throw someone in the water, only then you will come out’ 537

(27) mxi walox x  lqzli ‘You, strike one (blow) to my neck’ 853

2. kutxa ‘each’: +proleptic pronouns (28) ubaxtsa ddi mutwli kutxa g! xá 1!a ‘And my wives, I put

them each in one room’ 830

3 kutr- ‘both of’: +proleptic pronouns (29) m!rAmle spe tq&Allu kutru ‘He lifted his sword to kill them

both’ 827

(30) atta si 'mjrAblu kutru ‘Now go and test them both’ 302

4. kull- ‘all’: +proleptic pronouns (31) hda 1uzle hl muxurjle kullu “he did thus till he sold them all’


(32) bale lbi &a1nAnnu kullu pxá-ga mn!ž-gyni ‘But I cannot load

them all at once by myself’ 915

(33) 'qam mayle kulle xAly'sa dda ‘Her sweetness filled all of

him’ 808



(34) '1an lre hwle kullu ta daw $wya ‘...and these coins, I gave

them all to the woodcutter’ 670

5. ha/ha ‘this’: +proleptic pronouns (35) lzAm qa&lAtta  ha lan ‘It is necessary (that) you kill this one

immediately’ LH 1.10

(36) la xzlan u mAndi xta bas ha ‘We have seen nothing else but

this’ A 6.8 (u mAndi xta functions as proleptic pronoun)

6. wa, ya, ni ‘he, she, them’: never as neutral object: no proleptic pronoun (37) mbaqr'n  wa, balkAn 1wa k1e ‘Ask him, perhaps he knows’

MA 9.6

(38) 'mt šam1et dd 1mer, ya 1!zet ‘and whatever you hear him

say, that you should do’ 534

7.  xa ‘this one’: +proleptic pronouns (39) ! x!š šar& le, bale gAbAn 1het mn!šox 1!zAtta  xa ‘This is a good

covenant, but I want you to do this one by yourself’ 722

(40) na ha la qablanna xá mAnnox ‘I do not agree to this thing

from you!’ 58–59

8. ux(x)a ‘no one’: no proleptic pronoun (41) lagba uxxa ‘she does not want anyone’ 684;

expanded: (42) la šqAlli uxxa mAnn!h'n ‘I did not take anyone of them’ 702

9. 7m7ndi ‘nothing’ is one of two amply attested pronouns which show no proleptic pronoun: (43) na la g!zAn 1Allox i m7ndi ‘I will not do anything to you’ 665

This is due to the fact that AmAndi is indefinite. 10. xa m7ndi: no proleptic pronoun (44) lakun 1wAz 1Alle 1! 1arya xa m7ndi ‘...lest the lion do anything to

him’ 912

11. gy n- ‘—self’: The NENA verbal system has lost, as a group, the historical formation of the reflexive. The reflexive in JZ, as in some NENA dialects



(e.g., Christian Urmi), is marked by the reflexive pronoun gyn-, which takes the object slot. It is optionally represented by a pronoun in the verbal form. The preterite base generally does not show a proleptic pronoun with the reflexive pronoun, but sporadically it does, compare the following examples: (45) qAmla apya &Ate mundAy la g7y na kAsle “She too, his fiancée,

got up and threw herself at him’ 236

(46) /AbAqla bbAxya 'mundla gy na lxpqe ‘She choked crying and

threw herself in his lap’ 883

(47) 'mAn qarte q&l le g7y ne ‘and because of his grief he killed

himself’ 120

(48) šqAlla xanjar 'q&Alla g7y na ‘She took a dagger and killed

herself’ 329

(49) na š1Ašli gy ni ... š1š li gy ni xá-gar xta ‘I rocked myself... I

rocked myself once more’ 391–392

No difference is detectable. Of the forms in the subjunctive base, the form qam šqAlle always occurs with a bound object pronoun; the forms kšqAl and bšqAl rarely occur with the reflexive pronoun, but when they do, they occur either with or without the proleptic pronoun. The subjunctive form šqAl, when reflexive, mostly occurs without a proleptic pronoun. The same applies to the imperative forms. The unbound object pronouns are different than other nominals with regard to the DET system: these pronouns do not occur inside the DET envelope as other nominals (substantives, adjectives and adjective clauses). Yet they are compatible with a proleptic object pronoun whose mere presence indicates their definiteness. The following table summarizes the preceding data: Unbound object pronouns PRONOUNS



proleptic pron. gloss


537, 853 indefinite





‘each one’


827, 302 collective


‘both of them’


926, 915 collective


‘all of+pron’

!ha, ha

LH 1.10





(focal) ‘him, her, them’

wa, ya, ni MA 9.6





722, 59




‘this(F) one/thing’

684, 702 negative

‘no one’

A mAndi




xa mAndi









The rather small inventory of pronouns functioning as object teaches us that most of this functional load lies with the designated bound object pronouns. In addition, here too, the definite pronouns basically co-occur with proleptic bound object pronouns and the indefinite without it. The only exception is the personal pronouns (exx. (37) and (38)), which no doubt represent definite entities. They are marked as focal. Objects preceding the verbal form In § we have discussed the co-occurrence of the bound object pronouns with a following substantive. However, in many cases the entity which seemingly functions as the object in the clause precedes the verb, either in the same clause or in previous parts of the text. The latter case is the most common—an entity is presented at a certain point in the text, and then, whenever it is needed again, it can either be explicitly repeated or alternatively, represented by a pronoun: (50) xá y!ma !ha g!ra sle kAs do xamri '1Aswa 1Amme x kAstAt pre.

qam mat'la bn tre 1nAt xamra 'qam nšla tma ‘Once this man came to a wine dealer, he had with him a purse. He put it between two wine jars and forgot it there’ 281

(51) kud brta m!pAqla x kAstAt pre mAn go jba 'qam matw la ‘Each

girl took out a purse from her pocket and put it down’ 113

(52) musle xakma 1Arwe nš&e 'm&'ye '1wze wa$le-wa$le, mAtwle

qamni ‘He brought a few sheep, skinned, roasted and cut to slices, (and) put them in front of us’ 397

In the preceding examples (exx. (50)–(52)), the substantival referent subsequently represented by the bound object pronoun is introduced for the first time in the preceding context (all marked by indefinite DETs). Thereafter it is known, and hence definite, and as such can be represented in the verbal form. In these cases, the pronoun is basically used to maintain the link with the preceding text (minimally the preceding clause, sometimes farther away), no matter which verbal form may be involved—qam šqAlle



(exx. (50) and (51)), šqAlle (ex. (52)), or the modal šqAl and šq!l (both in ex. (53)): (53) 'maxwi gynox dd qam matwAtta 1asAqsa kAsle 'la matwAtta kAsle

mándla go palgs baar ‘Pretend as if (lit. show yourself that) you place the ring with him, but do not place it with him, throw it in the sea’ 722

Needless to say, the referent of the object pronoun may function in a variety of functions in the preceding text (e.g., object in exx. (51) and (52), existant in ex. (50), etc.). The other case of an ostensibly preceding object with pronominal resumption in the verbal forms turns out, in fact, to be related to topical extraposition; unlike the cases viewed above (§3.1.1), where the substantive deemed as object followed the verbal form, here it precedes the verbal form: (54) ubaxtsa ddi mutwli kutxa g! xá-1!a ‘As for my wives, I put

them each in one room’ 830

The substantival syntagm baxtsa ddi is here analyzed not as object, but rather as an extrapositional topic (the topic paradigm, as a group, is different from the object paradigm), being the discourse anchor about which the rest is predicated. For this reason it is, in a way, outside the kernel clause, and hence not really its object. It is resumed by the bound pronoun in the verbal form, which is the real and formal object. Note that the preceding substantive in the topic group is not appositive to this bound pronoun, as it is when following the verbal form: (55) (Every year one would send the king a lad and a maiden)

jwanqa q&ílwle 'xamsa gywílwla ta xa mAn bžer dde ‘The lad he would kill, whereas the maiden he would give to someone from his town’ 152

The substantives jwanqa ‘lad’ and xamsa ‘maiden’, although mentioned and hence definite, are an excellent example of generic topical entities (referring to any item from a group). Another example: (56) 'mx!lAplu bAt 1asAqysa, asAqsAt zne qam yawla ta mamo

'1asAqsAt mamo hw le ta zne ‘So they exchanged rings, Z.’s ring she gave to M. and M.’s ring he gave to Z.’ 305–306

These examples show a classic rationale for topics—contrastive extraclausal thematic entities. Both sets of topics are introduced before, and on both some new information is predicated in the comment clause. Note that in ex.



(56) there is a perfect syntactic symmetry between the two clauses, despite the different preterite formations. The same thing occurs in the following example:  (57) 1an julle qam z!náwlu (zwnwla) xaga mAn xa baxta ‘These clothes, she had bought them once from a woman...’ 778–779

 The gloss zwnwla for qam z!náwlu (both ‘she had bought them’) is in the original manuscript. (58) hl zAlla x šabsa, kullu mšalxne uganwe qam q&Allu ‘Within a

week, all robbers and thieves, he killed them’ 331

Ex. (58) can be opposed, similarly, to qmAn qa&lAnne x'we ‘Let me kill the snake’ (634), where the object group is {-ne x'we}. On the contrary, in all these examples illustrating a topic (exx. (54)–(58)), the object group consists of the bound pronoun alone. So far only one example turned up in which what looks like the topic has no resumption: (59) anya xabre la m$la mbaqra mAn xAddamta ‘She could not ask

these things from her maid’ (795).

The substantival group that precedes the verbal form but is not resumed later could be analysed as a preposed object, perhaps functioning as focus (see Givon 2001 vol. II:225 and § below). The last case is an adjective clause whose antecedent also functions as an object inside the attributive clause. We have shown above (§2.3.2, exx. (194)–(202)) that in this case resumption is optional: (60) húlllu mamo ta dan nše tre-ga Al dan pre dd hulle bAd ll

1arx'sa dde ‘M. gave the men two times the money which he gave the night of his stay’ 315

(61) mAn dan kap!re dd q&Ai ‘than the giants that I killed’ 412 (62) 1an &lha šo1le dd šql--le ‘the three things that he took’ 701 (63) 'xle ta dan raqAt mya dd qam q&Al-lu ‘so he wept over the

frogs that he killed’ 248

(64) 'bsa ddi dd bAd 1zi qam bnAn-ne ‘and my house that I built

with my own hands’ 220



Exx. (60) and (61) have no resumption, despite the fact that the antecedent is definite. Exx. (62), (63) and (64) do show the resumption of the definite antecedent. To conclude this issue, resumption by object pronouns (bound or not) is primarily determined by the pattern: 1. when the entity needed as object is previously mentioned in the preceding context it is either resumed or explicitly repeated in the clause. If it occurs in the preceding context as focus it is basically not resumed in the clause; 2. when it is topical, it is always resumed; and 3. when it functions as an antecedent of an adjective clause and the object of the attribute clause, its resumption is optional (when object!) but it never occurs with an independent pronoun. Object nexus: external syntax Object nexus has been discussed above, §1.4.1, mainly for its internal syntax: the form of the thematic entity and the various forms for the rhematic entity. Its external syntax, which is that of the object, is treated here. Object nexus and object clause (§ are related in that in both a nexus takes the functional slot of the direct object. That the object nexus is really a nexus is shown above (§1.4.1), but it is useful to give further illustration of this point here. The lexeme 1wza ‘make’ is occasionally used in such a way that something is made into something else (‘he made x [into] y’): (65) qam 1awAzla šaqq-šaqqe ‘He made her (into) pieces’ 937 (66) uzla g7y nox ne ‘Make yourself (into) pieces’ 205

This lexeme does not take two direct objects, but rather the slot of the object is here occupied by a nexus in reduced form, obtaining between the two entities (her and pieces in ex. (65), yourself and pieces in ex. (66)). The passive formation of this verbal lexeme is often related to a group of expressions functioning as semi-predicatives (e.g., maxw!ye ‘seem’, §1.4.1). This is illustrated in the following example: (67) t'lu kud 1arbu 'an  re... 1wzi wa$le ‘All four of them sat and

the birds... were made (into) pieces’ 864

The birds and the pieces are visibly in nexus with each other in the passive formation. The external syntax of these constructions is that of the object, in this case, a nexal object. Compare the following examples; the first is a simple object:



(68) la xš'le {u 1nAtta xr'ta} g! lAbbe ‘He did not have (lit. think in

his heart) {any evil intention} 819

The second group has object nexus: (69) 'qam xš'{la 1Astatti} ‘He thought {her (to be) my mistress}’


(70) lak'n mtáxmAnat 'xašwat{ti nša d1Abe mu)Arra d!x'n} ‘Don’t

you think and consider {me (to be) a man who wants your harm}’ 791

(71) $adli xašwAt{ti dAžmAn ddox} ‘I am afraid you will consider {me

your enemy}’ 923–924

The third group has a clausal object: (72) xš'le mšzinne 1! g!ra ‘He thought the man went crazy’ 765 (73) kxašwt'n 1na šAzna wna ‘You think I am crazy’ 271 (74) ax!ne xš'le Annu ganwa–le ‘His brother thought that he was a

thief’ 238

(75) xš'lu d1urre bar1a 1! pap'ka ‘They thought that the poor thing

entered the ground’ 768

All these examples share the same verbal lexeme, xšwa ‘think’. What is different is the type of object complement: in ex. (68) it is a substantive (evil intention). This substantive is equivalent syntactically to more complex forms of object complements, object nexus, in ex. (69)–(71) and substantival clauses in exx. (72)–(75). Exx. (72) and (73) have asyndetic object clauses, whereas exx. (74) and (75) have syndetic object clauses (with 1Annu and drespectively). The fact that all these different syntagms occur in one slot, namely, the object function, makes them each a part of a single paradigm, the object paradigm. The current issue in question, the object nexus, is not very common, except with the verbal lexemes xzya and mn!xe, strictly ‘to see, look’. Clausal objects: external syntax The internal syntax of various types of substantival clauses is found above, §2.3.1. This section is concerned with the external syntax as well as with the various types of clausal objects.



In this domain, dealing with any clause which figures inside the object slot—namely, ‘subordinate’ substantival clauses as well as ‘non-subordinate’ clauses containing direct speech—the analysis of the material can be presented efficiently and symmetrically only by putting to use another wellknown dichotomy, viz., syndetic and asyndetic constructions. Starting from the verbum dicendi par excellence, 1mra, we first ascertain that it indeed takes a direct object. This is best viewed in the substantival object slot (discussed above, § (76) mr- -le málabe dde ta daw g!ra ‘He told the man his request’


(77) mr- -le kulla q7a ta dan nše ‘He told the entire story to the

people’ 943

(78) mar-ra baqartox ‘Say your question’ 541

In these examples the definite object is also represented inside the verbal form. This means that this verbal lexeme takes in principle a direct object, which is syntactically and semantically equivalent to what is said. We may still have difficulty how to view and analyze the direct speech that occurs often with this verbal lexeme. The verbal lexeme 1mra ‘say’ occurs first and foremost with direct speech. However, it is compatible with other strategies as well. Only rarely do we find an object clause introduced by d(d) with this lexeme: (79) 'mArre ta daw g!ra dd 1be 1zAl hš kudy!m ušqAl lra dde mAn

kAsle ‘And he said to the man that he (=the man) may still come (lit. go) everyday and take his coin from him (=the speaker)’ A 1.8

(80) lak'n y br!ni 1amrtAn t-slox2 mux kullu

plavne ta a$'$a ‘My son, don’t you say that you came like all the heroes for fight’ 759 +

Exx. (79) and (80) are object clauses in which we find indirect speech. This occurs more often with this lexeme introduced by 1Annu:

The t- in ex. (80) is in fact the phonetic realization of the pronoun d- in the vicinity of unvoiced consonants. 2



(81) het gAmret 7nnu latlox hyi mAn prAt d! g!ra '1wa gmer 7nnu

qam nšlu ‘You say that you are not aware of this man’s money and he says that he forgot it (=the money)’ 283

(82) 1ay s!tAnta mArra ta d! faqr 7nnu m&še gyne ‘The old lady told

the poor (man) that he should hide’ 534

(83) mArru 7nnu 1Atlu 1Ammox xá maksAt tAjjr'sa ‘They said that

they have some trade issue with you’ 346

Some differences between dd and 1Annu are discussed above, §2.3.1. One more difference is that only 1Annu is compatible with direct speech in its scope: (84) slu xakma nše 'mArru &le 7nnu 1ítlni xá maksAt tAjr'sa

1AmmAt bbox ‘There came some people and said that “we have some trade issue with your father”’ 345

Note that exx. (83) and (84) have the same contents inside the scope of 1Annu, except that in the former it is indirect and in the latter direct. One quite readily compares this 1Annu to the Arabic 1Anna which basically occurs with direct speech. However, despite a small number of cases in JZ which remind one of Arabic,3 this entity is judged as an exponent of substantival clauses, generally containing indirect speech, but not exclusively. The direct speech in its scope occurs only after 1mra, which is, generally speaking, specialized for direct speech. An excursus about (in)direct speech is necessary here. The essential difference between direct and indirect speech could be summarized as follows: in direct speech the (putative) original utterance is (re)produced, whereas in indirect speech, the deictic elements (e.g., pronouns, adverb(ial)s of time and place as well as tenses) are different from the putative corresponding direct speech, conforming with the current relative situation (see, for instance, Palmer 1986:134). Another, a mixed type termed free indirect discourse which has specific narrative functions, is discussed below, § (In)direct speech can occur in syndetic clauses (as in exx. (83) and (84)). The following pair of examples has asyndetic clauses:

Showing some kind of explicative function: xá y!ma, tre 'mírrlu: 7nnu ls 1Ammi pre ‘One day, two, and he said to them: ‘but there isn’t any money with me’ (365) and similarly (65). 3



(85) si, marre: hawwa bywAnnlox brti ‘Go, tell him: “Fine, I will

give you my daughter”’ MA 10.4

(86) gzn h vi amr tn ta stzi šf t ape naqla msmé li ‘I beg that

you say to my lord the Judge (that) this time too he should forgive me’ 566

It is worthwhile to mention that the distinctive feature proving indirectness in ex. (86) is merely the place of the stress, which, in case of the imperative  eli ‘forgive me’).4 Asyndetic object would occur on the first syllable (msm clauses are very common in JZ, in fact, with certain verbal lexemes (e.g., ones which express the will of the speaker) they are the rule. The four options are illustrated in the following table: direct speech

indirect speech

syndetic clause

asyndetic clause

ex. (83)

ex. (86)

ex. (84)

ex. (85)

So, despite the fact that indirect speech commonly occurs in syndetic clauses and direct speech in asyndetic clauses, we still have the other options. The next group of examples represents the embedded question object clause. The first pair contains embedded direct and indirect nexus questions: (87) 'mArre ak!ma &le: ya g!ra k1At makAt lšnAt x'we? ‘So the

king said to him: “O man, do you know (how) to speak the language of the snake?”’ 34

(88) 'mb!qArre yla hakan k1e qre 1Abbe ‘So he asked the child

whether he knows (how) to read in it’ 598

These two examples (which are not strictly a minimal pair—the lexeme of the matrix verb is different in each) exemplify an embedded asyndetic direct question (in ex. (87)) and an embedded syndetic indirect question (in ex. (88)). Note that these are the only options attested: direct question

indirect question

syndetic clause

asyndetic clause

ex. (87)

ex. (88)

Compare: de, rba, mbš lla gynox. qam bšílla gyne ‘”O sheep, cook yourself”. It cooked itself’ (209). The first form is the imperative, the second is the qam form, based on the subjunctive base. 4



The following pair is not included in the table above, because the embedded pronominal direct questions are introduced by an interrogative pronoun, yet it would be odd to refer to a direct question by the term ‘syndetic’: (89) 'mírrle: ya ax!ni, ma brle Allox? ‘So he told him: “O my

brother, what happened to you?”’ 79

(90) mb!qArre dAkandre ma brle Al kullu /attle ‘He asked the

storekeepers what happened to all of the porters’ 571

Nevertheless, the object position is occupied in ex. (89) by an embedded direct pronominal question and in ex. (90) by an embedded indirect pronominal question (more on this issue is found above, §1.3.1). It is important to note that indirect questions are not really questions, and their internal structure is different: whereas in a real question the interrogative is normally the rheme of the clause, in an indirect question what looks like the interrogative (viz., the seemingly identical form whose function is, inter alia, to introduce these clauses) is not the rheme, but rather an exponent (see §2.3.1 above). The internal grammar of these clauses that involve indirect speech is somewhat different from its environment (e.g., the narrative), most notably as far as tense choice is concerned, and it is dealt with below, under macrosyntax (§ To sum up, these various clauses are embedded in the slot of the object, regardless whether they are traditionally subordinate or not. Hence the terms (non-)subordinate, used until now, is less practical in this domain. The following table shows why: general framework traditional terms

more relevant terms marking

speech content type content

nexus question**

embedded clauses in the object slot subordinate




by exponent*

by syntax


ex. (83)

ex. (88)




ex. (84)

ex. (86)

ex. (85)

ex. (87)

* Annu, d(d), markers of indirect questions (ma, ka, mani, etc.) and indefinite pronouns (mostly md). The term “reported speech” is not used because all speech, whether direct or indirect, is reported by the narrator in our case, and occupies basically the same syntactic positions. 5



** Questions are difficult, since direct pronominal questions are always introduced by an interrogative exponent, and thus the term ‘asyndetic’ does not quite fit them. For this reason, the question part in the table actually refers to nexus questions. 6

3.1.2 Indirect/prepositional object syntagms Verbal forms also take prepositional syntagms as their object. In principle, the difference is not significant, as can be shown in Hebrew, a language which has always marked its definite direct objects by a preposition, while indefinite direct objects occur without one (e.g., 1% hab-bayi% ‘the houseACC’ VS. bayi% ‘a house(ACC)’). The implications are that, despite the obvious difference in definiteness, both syntagms occupy the same syntactic position, that of the object. In JZ it is somewhat more complicated: unlike direct objects, which are potentially represented by a pronominal element in the verbal form, the prepositional object syntagms are not, and are represented by a prepositional syntagm whose attribute is only a pronoun.7 The prepositions b- and l- are occasionally suffixed to the verbal form. One finds pure a prepositional syntagm as object in some verbal lexemes. The first lexeme, &pya ‘stick (to), be glued’ always occurs with the preposition b(Ad): (91) '&'be &plu bAd 1ar1a ‘And the bricks stuck to the floor’ 209 (92) kud š!1a yr!ne dda &plu 1Abba ‘All her seven lovers stuck to

her’ 210

Other lexemes are r1ša ‘feel, notice’ and tfqa ‘meet, notice’, occurring primarily with b(Ad), but occasionally with All-, with no appreciable difference: (93) r1Ašle 1Abbi ‘he noticed me’ 393

In fact, in JZ even rhetorical questions are occasionally marked by a particle: xula bšwAqlan mn!šan? ‘Will he leave us alone?’ (SYG 1.20). 7 One should remember that some historical prepositional phrases in NA came to function as pronouns, as for instance the form representing the agent in the verbal form šqAl-la or the object in qam šqAl-la. By structural standards, these l- pronouns are no longer analyzable as prepositional phrases, since no other preposition may interchange with them. This may be illustrated by comparing the forms šql--le ‘he took her’ and šql- ‘she was taken’, where the sole function of the syntagm -le is to represent the agent. 6



(94) 'la r1Ašle Atti bAt kApna ‘And he did not even feel the hunger

(lit. even the hunger)’ 510-511

(95) r1Ašle 1Alla 1Astza ‘Her master noticed her’ 551 (96) tfAqle 1Abbe xá g!ra ‘A man met him’ 519 (97) tfAqla bAd bAr 1am!ya ‘She met her cousin’ 684 (98) kan xzlox ak!ma yn tafqa 1Allox brte ‘If the king sees you

or his daughter meets you’ 758–759

It is important to illustrate that these prepositional syntagms occur as objects. The first thing to do is to compare them to the synonymous syntagms functioning adverbially: (99) 'pre xta psAxle 7- ! $wya xa dAkkán tAjjr'sa ‘And the

other money, the woodcutter opened with it a merchant store’ 434

(100) t'ma ma b!zna 7? ‘The garlic, what will I do with it?’ 299 (101) zle xa 1man dd gxaprA 7 ‘He saw a dish which one digs

with’ 110

The lexemes psxa, 1wza and xpra all take direct object. The 1Abb- syntagm functions here as an instrumental adverbial. So, in order to judge precisely the syntactic status of these syntagms, we need to compare them, in the very same syntactic conditions, with a direct object. Such pairs, or groups, do exist in JZ. The following two pairs of examples feature the lexeme $rxa ‘call, scream’: (102) '$rAxla * ‘So she called her husband’ 728 (103) $rAxle $0* ‘He called his friend’ 959 (104) $rAxle &,*%$0& ‘He called the two angels’ 564 (105) $rAxle 7$&,*)  ‘He called the two messengers’ 561

It is plain that the object could either be a direct one or a prepositional one. These data allow us to configure a paradigm, where, in the object position, a prepositional phrase and a direct object have exactly the same function— despite many examples of both possibilities, no consistent distinction was found to exist between them. Having discussed above prepositional attributive constructions, we know that the syntactic status of the nucleus is the



same for the entire construction. Hence, if the prepositional syntagm interchanges with a direct object, they are syntactically equivalent to one another. However, this (in)direct object position is not fully identical to the common direct object, for which a direct object bound pronoun occurs in or with the verbal form. A similar situation is found with other lexemes. The case of &1ya ‘look for’ is a special one. Like $rxa above, it is capable of occurring with either direct or prepositional object. However, here one does find a regular difference between the two with the subjunctive base: (106) 1na gAbn &1Atti xa xamsa h!ya mAn mAšpá bš ‘I want you to

look for a maiden (who) is from a good family for me’ 541– 542

(107) p&1axni xa šahra ‘We will look for a blind man’ 654 (108) !ha faqr gzAlwa g&1wa šaqfat $we AngArre ‘The poor man

used to go (and) look for (some) piece(s) of wood (and) rugs’ 290

(109) na gAbAn &1t'n xá1 y1e qrle 1! ammaš ‘I want you to look

for one (who) knows to read this book’ 595

(110) slu kutru ak!me k&1e Al yalunku ‘The two kings came looking

for their children’ 126

(111) uk'n k1en (k1An) k&1At Al qA&la ddi ‘because I know (that) you

are planning to kill me (lit. you look for my killing)’ 457

(112) dx '&1ax Al maammad axonni ‘Let’s go and look for M. our

brother’ 79

The opposition direct object (ex. (106)–(109)) vs. prepositional syntagm (exx. (110)–(112)) with the subjunctive base of &1ya is linguistically pertinent: the former is used with indefinite entities, the latter with definite. This functional difference, which is very similar to the regular difference in Hebrew, is local and occurs only with this lexeme and only in the subjunctive base. Note that, when the object of a lexeme which takes a prepositional syntagm as object (e.g., r1ša ‘feel’ is a clause, the preposition does not precede the clause, and the clause is the same as it would be for a regular direct object lexeme, that is, without a preposition:



(113) basAr xakma y!me r1Ašle 7nnu bAštum le ‘After a few days he felt

that he was (lit. is) better’ 312

(114) r1Ašla d-lAbba rba zAlle ld! jwanqa ‘She felt that her heart went

after this youngster’ 763

This commutation between a prepositional syntagm and an object clause is yet another way to show the syntactic equivalence between the syntagm in question and a direct object. Another case of prepositional syntagm as object is found in the second object position. Various verb types are ditransitive, i.e., occur with two objects. The clearest case is the (rather rare) causative stem of a transitive lexeme. The causative stem, when occuring with a root which is found with the simple stem as well, always appears to have one more argument compared with the same root in the simple stem, and that is the main difference between both lexemes (compare Eng. eat vs. feed; see vs. show; die vs. kill; etc.), although, as in the Semitic languages, the formal relationship between them is transparent, for instance, d1ra ‘return (vi)’ vs. mad1!re ‘return (vt)’; ytwa ‘sit’ vs. matt!we ‘set’; txra ‘remember’ vs. matx!re ‘remind’. Thus, when a causative pattern occurs with a root for which a simple formation taking a direct object exists, the same root in the causative formation takes two direct objects. Compare the following pairs: (115) 1na la lApli 1! xabra(1) ‘I did not learn this thing(1)’ MA 2.8 (116) urba m!lAple br!ne(2) t!ra(1) ‘and he taught his son(2) plenty

Torah(1)’ 608

(117) &A'+A+.!.%(/A##1%&)'A*A#((1) ‘The youngster’s bride ate

the beet leaves(1)’ 548

(118) * . !)  "/A#. '$/A#. /+(2) & %

&)'A*A#((1) ‘The house of the man who used to eat and feed his wife(2) three times beet leaves(1)’ 557–558

The lexemes #0' and 1/# are plain transitives, whose object are what is learned (1! /) in ex. (115)) or eaten (1% &)'A*A#( in ex. (117)), respectively. When the same roots are found in the morphological causative pattern, they have yet another object—the actual agent of studying ()!% in ex. (116)) or eating (/+in ex. (118)). Thus exx. (116) and (118) have in common two direct objects each.



The issue of double direct object potentially has something in common with prepositional objects—the second direct object in JZ cannot occur as subject index in the corresponding passive pattern. When a non-causative lexeme has two objects, a direct one and a prepositional one, it is basically the direct object which is not an object anymore in the corresponding passive pattern, while the prepositional object remains intact: (119) na hwli brti ta daw g!ra ‘I gave my daughter to the man’

554 (active)

(120) umd 1s gwu wlu hwe &li ‘and what is in them is (lit. are)

given to me’ 606

(121) ay dammAt slu yAhwa pre &le ‘When (his) money was given

to him’ 289

(122) 'pAsrox ta wne bse yAhwa ‘and your flesh will be given to

the animals’ 365

Ex. (119) is an example for the active behavior of the lexeme yhwa ‘give’, occurring with a direct object (-- brti) ‘my daughter’ and the prepositional syntagm headed by ta~&l-. The latter seems, in exx. (120), (121) and (122), all occurring with different passive exponents and various tenses, not to change at all.

3.1.3 Representative valency patterns: parameters The objective of this section is to describe some of the most prominent valency types. Verbal lexemes are often viewed semantically (and named accordingly, e.g., verba dicendi, verba movendi, etc.), however, this classification can be carried out using syntactic criteria as well. Syntactic and semantic criteria overlap here, but not entirely. Giving This valency pattern is given ample consideration in Polotsky 1979:213– 226, including lexemes of similar valency (the corresponding lexemes in JZ are mšad!re ‘send’, mzab!ne ‘sell’, mes!ye ‘bring’, 1mra ‘say, tell’, šqla ‘take’, mak!ye ‘tell’, etc. Of course the valency patterns of these lexemes are not identical to yhwa ‘give’). Polotsky terms the direct object ‘gift’ and the indirect object ‘recipient’. In JZ the situation is somewhat less complicated than it is in literary Christian Urmi: Any object can be represented by a



bound pronoun if it is definite. Thus, if both gift and recipient happen to be definite, the two pronominal sets can and do co-occur: Both definite: (123) 'qam ywíl–l'–le ‘He gave them (=the stones) to him (=the

king)’ 361

(124) bAywín–n–lox ‘I will give her (=my daughter) to you’ 359 (125) bAywAn–ne &lox ‘I will give it (=the greater part) to you’ 253

Note that that the opposition definite vs. indefinite is relevant in the 3rd person only (the 1st and 2nd persons are definite by definition). The basic pronominal order is direct object—indirect object.8 The recipient (indirect object), when definite, can be represented by an external prepositional syntagm, unlike the gift (direct object): (126) lág bwa ywílw–la brte ta d!ha jwanqa ‘He did not want to

give his daughter to the youngster’ 232

Recipient only definite:  ma e ‘He told them to give him (127) mírrlu yw–le xá darmn some anesthetic drug’ 567

(128) gm&lu 1anya nše ywAt–tu xá xzna ‘These people deserve

(that) you give them a treasure’ 472

Since the bound signals for (in)direct objects in the subjunctive and imperative bases are identical, when only one occurs with the verbal form it could in fact be either. Gift only definite: (129) 'xamsa gywílw–la ta xa mAn bžer dde ‘and as for the girl, he

used to give her to one of his town’ 152

The definite direct object need not be represented by a bound pronoun: (130) 'ta xá mAnn!x'n bAywAn brti ‘And to one of you I will give my

daughter’ 470

There is one exception: hál–l–le 1!ha šúlan ‘give me this table’ (212), as opposed to the regular order, e.g., uhál–l–li xa xmrox ‘give me one donkey of yours’ (917). 8



Gift indefinite, recipient unknown or unmentioned: (131) kud mani dd ywAl pre ta $Adqa ‘anyone who gives money for

charity’ 292

(132) kud šta !ha dra gAywAlwa bas &ha xab'še ‘Each year the tree

woud only yield three apples’ 184

(133) '1aw 1lhAt hulle ! fukAr bAywAl xa bAš&!f ‘And that God who

gave this thought will give a better one’ 466–467

Some of the lexemes that follow this valency pattern are enumerated above. Some of them (e.g., 1mra, see below, § are more complex in taking embedded clauses in their direct object position. Ability, will, obligation and permission The verbal lexemes in this group (‘be able’, ‘want’, ‘order, request’ and ‘allow’) are not identical; they have in common an obligatory subjunctive clause in one of their object positions, but can differ from one another in other parameters, namely the number of object positions, or the nature of the other objects. From the structural perspective, whereby functional values are derived from oppositions, an obligatory form has no functional value. This is plain below, where the subjunctive form is the only option. The diachronic rationale that led to this synchronic state is explicable by the evident connection between a verbal lexeme which expresses, e.g., will, and the verbal form whose purpose is to reflect such will. However, synchronically speaking, the obligatory subjunctive form in the clausal object is nothing but what Marouzeau 1951 called mode grammaticale, devoid of its own synchronic, systemic value. Ability is expressed by a variety of forms, forming some kind of suppletion: the verbal lexeme m$ya, the verbal expression hwya b-, and the expression 1(s) b-: m$ya

hwya b-





lakwb- (neg. only)


pwb- (aff.) lakwb- (neg.)

imperfect future


subjunctive 1Am$e (affirm. only) “past” subj. 1Am$wa

1s b1Aswb(neg. only)





The lexeme m$ya is almost specific, having two related, allo-values—‘be able’ (with a clausal object) and ‘overpower’ (with a substantival one). The other two expressions could mean ‘be in’ and ‘exist in’ respectively. What nevertheless constitutes their unity when having the value of ‘be able’ is their complement—an asyndetic object clause featuring, as a rule, subjunctive forms. The other value, ‘overpower’, is always followed by a prepositional syntagm headed by Al-. The following examples all feature the lexeme ‘to be able’: (134) la m$lu nablle ‘they could not bring him’ 617 (135) u1wa lbe npeq ‘while he cannot come out’ 533 (136) dd hybi yasqAn rAš kapnox ‘so I can go up on your shoulders’


(137) la kybu 1!zle ‘they cannot (or will not be able to) do it’ 721 (138) g1Am$wa 1swa 'mbqArwa 1Abbe ‘he could come and ask about

it’ 693

Each of the three suppletive forms shows the very same valency: (139) '1uzli xapa y 'qahwa dd hybi amrAnnox m&o pybox msAt

xalwAt arya ta bbox ak!ma ‘and make me some tea and coffee, so that I can tell you how you could bring lion’s milk to your father the king’ 72

(140) ! qadda íswli, dd íswbi kullu anya le kulla ha bžer

z!nínwla ‘I once had plenty of money, I had so much that I could buy all this wilderness and all this city’ 291

Want or wish are represented by the root 1Abya ‘want, be necessary, love’, which also occurs with a direct object: (141) 1atta gAbanne dehwi ‘Now I want my gold’ 13



The first two, tightly connected values have a simple and consistent complement strategy—an asyndetic clause, always occurring with a subjunctive form:9 (142) apna 1Abli 1!zAn 1Ammox haw'sa ‘I too wanted to do (lit. [that]

I do) you a favor’ 662

(143) gAbAt 1zAt ‘You want to go (lit. [that] you go)’ 749 (144) gAbAn 1het mn!šox 1!zAtta 1xa ‘I want you yourself to do (lit.

[that] you yourself do) this’ 722

(145) na gAbAn msafren 1Adyo Al xá bžAr rauqta ‘I want to travel

today to a faraway town’ 695

(146) gAbwa šqAlwa mahne dde ‘He wanted to take his mare’ 238

(147) qam d!qle 'g   ‘They caught him and wanted to kill (lit. [that] they kill) him’ 196

This asyndetic subjunctive clause is the syntactic equivalent of the simple substantival direct object. In exx. (146)–(147), the subjunctive clause shows a kind of concord with the past in the matrix clause, marked by the -wa morpheme. In addition to will, gAbe has an impersonal function, meaning roughly ‘it is necessary’, always occurring in the present tense form: (148) kud baxtAt 1Ablox gAbe 1ya masylox bAjabryi l1urxAt g!ra

'baxt'sa ‘Any woman who wants you, it is necessary that she (herself) bring you by coercion to have sex (lit. to the way of {man-and-woman}hood)’ 821

(149) nšli xá mAndi z!nna; gAbe da1ren 1Al š'qAt bžAr ‘I forgot to buy

something; it is necessary (that) I return to the town’s market’ 696

(150) gAbe msafraxni 1Al bsa ‘It is necessary that we travel home’ 696

Other expressions of obligation are identical in their valency:

What looks like an exception, when it occurs with the future indicative, is in fact a conditional structure: 'gAbat ppšan mAzas ‘and (should) you want, we will stay together’ (904). 9



(151) xá yma pšle  jwanqa na lj šwAqlu kutru baxtse ‘One day

the youngster was obliged to leave (lit. [that] he leave) both his wives 842

(152) pšla (majbr ) na lj amra ay dlubba ‘She was obliged to say

(lit. [that] she say) what is in her heart’ 861

The root w$y ‘order’, despite the similar semantic domain, is somewhat more complicated, having in principle two possible slots (unlike the preceding lexemes). One is the basically obligatory object clause, always containing the subjunctive, but which can be introduced either asyndetically or syndetically; the other is an optional position for the addressee of the order. Clear cases of the former are illustrated by the following pair of examples: (153) m li x la mamra lday xet ‘So I ordered (that) one should

not command the other’ 830

(154) m!$le 1zi xakma mAn sardre dde ‘He ordered (that) some of

his chiefs should go’ 260

These examples feature an asyndetic object clause. The addressee of the order can take the position preceding the clause: (155) m!$le ak!ma Al sardre dde ywi xa dAyri ta d! g!ra ‘The

king ordered his chiefs (that) they/one give a present to the man’ 662

The small difference between exx. (154) and (155) is found in both structure and meaning: in the former, the addressee is not mentioned (the position is left empty), and the order is to be carried out by the chiefs; in the latter, the addressees are explicitly the chiefs, and they are (possibly) the ones to carry out the order. (156) m!$le sardre dde dd 1zi $arxlu kullu nš day bžAr ‘He

ordered his chiefs that they go and call all of the people of the town’ 594

(157) m!$le 1aw ak!ma sardre dde 1Anni 1zi basAr r'wAt najre ‘The

king ordered his chiefs that they go after carpenter’ 713

the greatest

The cases in exx. (155)–(157) clearly show the first, addressee position, by marking off either the substantival syntagm (ex. (155) Al sardre dde) or the substantival clause itself (as in exx. (156) and (157), by dd or 1Anni). Note



that as far as the information is concerned, there is little difference between (154) and (155), but the pattern is of course different. The first, addressee position can be occupied by an object pronominal entity as well: (158) 'qam m!$–le dd la lqA&le 1aw qu&e1fa ‘and he ordered him not

to (lit. that he should not) glean the (grape)cluster’ 650

(159) 'qam m!$–le dd la 1asya Al xal!pe ‘and he ordered him not to

have it (lit. that it (=the ring) should not be) exchanged’ 723

(160) 'qam m!$–lu mxle ‘and he ordered them to (lit. [that] they)

beat him’ 564–565

This corroborates what we stated above: that the first position is an object position. This is not clear when, e.g., the clause is asyndetic and the first position is occupied by a substantive: (161) um!$la mare daw 1ax!ra mxAlla 1xla basma ‘So she ordered

the owner of the stable (that) he feed it tasty food’ OR ‘(that) the owner ... feed it...’ 318

(162) m!$le xarakarat dde qrlu 1an ammše qam ak!ma ‘He

ordered his magicians (that) they read the books to the king’ OR ‘he ordered (that) the magicians...’ 590

That is, by just looking at the two preceding examples, we cannot tell whether the underlined entities constitute another object position, or are part of the clausal object, as is clearly the case in the following example: (163) $rAxle 'm!$le 1se kullu xarakarat dde ‘He called and ordered

(that) all of his magicians come’ 590

The clause, in one occasion, is in the form of direct speech: (164) m!$la ulamwsa dda “mšawli d'ka 'ywli /ašya” ‘She

ordered her servants(:) “let them prepare a place for me and give me supper”’ 409

l ba ‘request, ask for’ has a basic valency of one direct object (the thing requested) and a prepositional syntagm (mAn) introducing the addressee:



(165) k&lAbwa mAn 1lha bas xa yla ‘He would ask God (for) just one

child’ 575

(166) 'k&lAbwa xa yla ‘He would ask (for) just one child’ 576

Comparing the preceding examples, we can see the object is more or less obligatory, while the other position, marked by mAn, is not. This situation is the same when in the direct object position we find a clausal object: (167) '&abi maxwlu mAn dma d'ka šqlle 1anya jawArre ‘and that

they should request (that) he show them from what place he took those gems’ 260

(168) &lAble mAn baxte pasxle tar1a ‘He requested of his wife (that) she

open the door’ 717

In addition to the asyndetic clauses in the direct object position, we find syndetic clauses as well: (169) '&Abla mAnna dd yatwa muqdar xá sa/a ‘and she requested her

that she sit about an hour’ 688

(170) &Abli mAnnox dd la 1zAt bAd d 1urxa ‘I asked you not to go (lit.

that you do not go) in this way’ 614

Two special cases show other markers for the clausal object: (171) hl &lAb mAnna kan qabla gwArra ‘until he would ask her

whether she agrees (that) he marry her’ 809

(172) k&abAn mAnnax lakn karbat mAn xabre ddi ‘I ask you not to be

(lit. lest you be) angry because of my words’ 881

These different markers might occur because of a slightly different value of the lexeme, ‘ask’ rather than request. The lexeme šw qa ‘let, allow’,10 like all the lexemes discussed in this section, has a clausal object with the subjunctive:11

Rather than ‘leave’, which is a simple transitive: šwqli xparta ddi ‘I left my digging’ (403–404). 11 Occasionally, this lexeme is found with an object nexus: šuqlu tar1a psxa ‘they left the door open’ 678 hama š!qAtti $x ‘just leave me alive’ 748, 852 10



(173) pqa&Anne mamo 'lak š!qAn mtahne bAd zne ‘I will kill Mamo and

not allow (that) he enjoys Zine’ 327

(174) 'kan h!ywa 1Abbu lak š!qwa 1!zwa 1Alla u mu)Arra ‘and if they

could, they would not have allowed (that) one cause her any harm’ 850

(175) la š!qAt barli u mu)Arra ‘Do not let any harm happen to me’


These object clauses are asyndetic, as is basically the norm in this group. A somewhat different form of complementation is found in the following group of examples: (176) la šuqla !ha x'we npAq ‘She did not allow (that) this snake

come out’ 25

This example could be analyzed as if there is an asyndetic object clause in which the agent (=the snake) takes the first position (whereas all the preceding cases it is either not there or following the verbal form). However, in view of the rest of the examples, this is impossible: (177) šuqli 1zen Al bsi ‘let me (lit. [that] I should) go home’ 375–376 (178) 'qam š!qle 1zAl ‘she let him (lit. [that] he should) go’ 490 (179) šwAqla baxte ma&ya ld mA&'we ‘(that) he let his wife (lit. [that]

she should) reach such low level’ 851

Exx. (177)–(179) show us a different strategy. The pronouns, structurally in the same position as the snake in ex. (176) above, all turn out to be object pronouns. This situation indicates that we should not analyze the structure as a clause but rather treat it as a construction very similar to object nexus, but whose rheme is an entire asyndetic subjunctive clause. Note that in JZ object clauses in general are finite, in conformity with the linguistic area consisting of Arabic and Kurmanji. Fear Fear is expressed by the lexeme zd1a as well as by the expression $adl-. The common clausal object is an asyndetic subjunctive clause: (180) 'zde1li damxAn go daw !l ‘but I was afraid to sleep (lit. [that] I

sleep) in the desert’ 406

(181) gzd1wa npAqwa ‘He was afraid to get out’ 25



(182) $adle š!qle kAz gyna xakma y!me 'basAr hdax kardle ‘He was

afraid (that) she should leave him with her a few days and then drive him away’ 803

(183) $adli xašwAtti dAžmAn ddox ‘I am afraid (that) you consider me

your enemy’ 923–924

The lexeme zd1a is, however, capable of taking a clausal object introduced by lak'n ‘lest’, which is the distinctive feature of this group: (184) zde1le lakn hwe xAlma ‘He was afraid lest it is a dream’ 433 (185) zde1li lakn 1se xa wan yan xa &ra ‘I was afraid lest an

animal or a bird would come’ 379

This lexeme can also occur with dd: (186) na gzad1an dd yla 1zAl bAt r 1urxa 'zyA/ ‘I am afraid that

the child would go in some other way and get lost’ 584

(187) uk'n zde1le dd nšle zwAn mAn kullu šAkle ‘since he was afraid

that he forgot to buy of all kinds’ 570

Another strategy altogether is juxtaposition: (188) la zad1t, la ksAt Al qA&a ‘Do not fear, you will not be killed’ 718 (189) la $adlox 1na la g!zAn i mAndi Al x'we ‘Do not fear, I will not do

anything to the snake’ 633

The connection between the two clauses is not embedding: the verbal form in the second clause is not the subjunctive, but the (future) indicative. Asking vs. not-knowing The basic valency of mbaq!re ‘ask’ is as follows: no.







zli xakma nše ... mbuqr--li: ‘I saw some people... I asked them:’ 423


mb!qArre šahra ham 1ay baqarta ‘He asked the blind man the same question too’ 660




'qam bqArre 1Al kfe ‘he asked him about his health’ 679




mb!qArre ld! /arja ‘he asked the lame man:’ 657

prep. Al



mb!qArre mAn d!ha g!ra ‘he asked the man...’ 531 prep. mAn




That is, the addressee is presented either as direct object (exx. (190)–(192)), or prepositional object (exx. (193) and (194)). The object question, when non-clausal, is similarly either direct or prepositional object (exx. (191) and (192) respectively). When clausal, the object question is an embedded direct or indirect question. The former is by far the most common, and takes the object position: (195) 'qam baqrli: amn wet yan kfer? ‘They asked me: “Are you a

believer or a disbeliever?”’ 381

(196) qam baqrli: mka wten het? ‘They asked me: “Where are you

from?”’ 421

Embedded direct questions can be either about the nexus (termed ‘nexus questions’, ‘yes-no questions’ or ‘polar questions’, ex. (195)) or about various functions in the clause—object, adverb, etc., which are represented by an interrogative pronoun in the question (ex. (196)). Indirect questions too are related to these two varieties, nexus questions and pronominal questions. The extent to which they can really be considered true questions at all is still to be determined: the only indirect clausal object the lexeme mbaq!re takes is what we identify as an indirect question. For this to be structurally the case, viz., for such clausal object to have this value, we need it to be opposed to some other type of clausal object. This is done below, in discussing the valency of not-knowing. However, take a look at the following pair: (197) qam baqrle bbe 'yAmme 1ka wle kulle wa/da ‘His father and

mother asked him where he was all (that) time’ 607

(198) qam baqrle ulumwse hakan zunne mt nšwle ‘His servants

asked him whether he bought what he had forgotten’ 700

If we compare ex. (195) to ex. (198), on the one hand, and ex. (196) to ex. (197), on the other hand, we get the following: ex.

matrix verb

(195) (198) (196)


embedded clause

amn wet...? qam baqrl-

hakan zunne

mka wten...? 1ka wle

question type embedded clause nexus pronominal

direct question ?

direct question ?



The direct questions (in exx. (195) and (196)) are in the same paradigm and are opposed to indirect questions (in exx. (197) and (198)). It seems, then, that indirect questions are nevertheless a linguistically pertinent entity here. The following example illustrates the two types of indirect question when connected: (199) basAr hdax mb!qArra lkfi (1) 'm&o gma1'rAnne y!ma go d

škafta (2); kan rAzya wen mAn 1xla 'štya dgywli (3) ‘Thereafter she asked me how I was (lit. for my happiness) (1) and how I passed the day in the cave (2), whether I am satisfied with the food and drink that she gives me (3)’ 412

This example has, in the question position, a prepositional object (1); an indirect pronominal question (2); and an indirect nexus question (3), which means they all belong to one paradigm, performing the same function. The following tables show both the possible exponents, as well as the inventory of tenses in the clausal objects in question: exponent variants

Asking: indirect nexus question

(1)skn, (1A)zakan, (ha)kan ‘if, whether’





588, 700


335, 412, 598

perfect future

Asking: indirect pronominal question

ma ‘what’, bAd dma ‘in which’, 1ka ‘where’, m&o ‘how’, &amha ‘why’, 1man ‘when’, etc.






607, 671

present future

290–291, 592, 695 580, 861–862

Note that in this group, unlike the preceding ones, the subjunctive is not obligatory, and, in fact, it is doubtful whether it occurs here at all. This implies that, unlike the places where subjunctive is obligatory, and hence no temporal or modal differences12 could be discussed, here we have an almost One such difference may be pertinent, namely the -wa(-) form, which, when it occurs with subjunctives is either 1. modal congruence (“past subjunctive”) or 2. hypothetical modality (as past subjunctives in the Romance languages). 12



normal spectrum of temporality.13 Additionally, the indirect questions are all syndetic, introduced by some exponent. Another way found in the texts is where dd is preposed to the signal of indirect question: (200) 'qam bqArre ak!ma 1! g!ra dd m&o m$le la lqA&le 1! qu&e1fa

‘So the king asked the man how he succeeded (in) not gleaning the (grape)cluster’ 661

(201) qam bqArre 1aw sardar ak!ma dd bAd dma y!ma pxa$a mAn

mal!me ‘The king’s chief asked him in what day the soldering will be done’ 723

(202) 'mb!qArru 1Astze hakn dd dle zla 1aw y!ma Al knAšta ‘They

asked his teacher whether he has gone that day to the synagogue’ 589

Such examples are difficult to analyze; it is as if these indirect pronominal questions are not felt to be syndetic by the speaker (S), since they are supplied with the exponent dd. The last example (ex. (202)) is even more difficult, since dd occurs after the indirect question exponent. Related, but not identical, is the lexeme yz1a. When negated, it is related to mbaq!re in that it predominantly takes indirect questions as its object (but never direct questions!). It is different in that it is not obligatory, that is, it also takes a regular clausal object, introduced either asyndetically or syndetically: (203) ma la ze1lox br!nox jAAl le ‘Did you not know your son is ill-

mannered?’ 461

(204) het la k1ten m&šten pre ‘You do not know (how) you

should hide the money’ 339

(205) 'la ze1lox dd wan t'ta xa ršox ‘But you did not know that I

was (lit. am) sitting near you’ 690

(206) la ze1le dxa-baxta xt wla mAlta 'bAthša pkutru ‘He did not

know that another woman was standing (and) gazing at both of them’ 820

The values of these forms in embedded clauses and the constraints in these environments are discussed below, under macrosyntax, §5.6.4. 13



And most often we find a clausal object which is an indirect question: (207) u1atta la k1n hakan dlu $x basri ‘and now I do not know

whether they are well after me (leaving)’ 600

(208) uk'n nša la k1e ma basya brše ‘since a man does not know

what will happen to him’ 835

(209) la ze1la ka wla ay d'ka 'la ze1la mánle mamo 'mani qam

nbAlla Al tm ‘She did not know where was this place, and did not know who is M. and who brought her there’ 308

(210) la k1An m o zAlla ‘I do not know how she went’ 310

If we compare the regular clausal object (exx. (203)–(206)) to the indirect questions (exx. (207)–(210)), we get a clear opposition: in the former type all of the information is there, while in the latter one piece of information is represented by an entity most closely associated with an indefinite pronoun. Knowing and thinking These two lexemes are similar in their syntactic valency. The differences are tabulated below: strategy






ze1le mahj!ye

object nexus asyndetic

1Annu embedded d(d) clauses direct speech

indirect question

presentative clauses


ex. (183)

most common uncommon uncommon 618, 648, 713 803



The kernel of similarity between the two lexemes lies first and foremost with their compatibility with various clausal objects, whether asyndetic or syndetic (including relative frequency). The differences lie in the less common strategies—object nexus, embedded direct speech and indirect questions. The latter is illustrated below: (211) xa hwwa (hywa) na$ax, k1wa 7nkan myes u7nkan bA&re$

'xa baxta dd h!ywa smAxta k1wa 7nkan br!na-le u7nkan brta-la ‘Someone (who) was sick, he would know whether he



would die or recover. And (as for) a woman who was pregnant, he would know whether it is a boy or a girl’ 226–227 (212) xš'le m of gbarya h ya  baxta hatxa musn kulle mndi x

 za uya bd d umr gbarya h ya pšta rauqta mn g ra lá h!ya &mA/ta 'mAšt!hsa ‘He thought how it is possible (that) this woman is so beautiful and has everything in reach, but she has remained, at this age, without a man and does not desire and long’ 803 Seeing Lexemes such as xzya ‘see’ and men!xe ‘see, look’ have a special and rich array of valency strategies. It is impossible to name this group “perception” for the simple reason that other “natural” perception verbs such as šm1a ‘hear’ behave differently. As far as a simple object goes, these lexemes are not the same: xzya takes a direct object, whereas men!xe a prepositional object (1ibb-~bAd). This is, as is shown above, insignificant. As far as the various strategies are concerned, they are almost identical (mutatis mutandis, of course, since men!xe is much less frequent). One very typical valency strategy with this group is the object nexus. The strategy has been explained and described above (§§1.4.1 and (213) zlu Ašti rubwsa talmze dwqe ammše bzu ‘They saw an

infinite number (lit. sixty myriads) of his followers holding a book in their hand(s)’ 55

(214) zli xakma nše bzla ‘I saw some people going’ 420 (215) m n xlu m ltx qa ra, xz–le bzla ‘They looked beneath the

castle, to see him going’ 494

(216) 1! g!ra $wya 1ay dammAt zle 1arya qarwa 1Alle rba rAple ‘The

woodcutter was very startled when he saw the lion close to him’ 665

(217) kxzla gyne go nam'sye dde ‘He finds himself in his own

bed’ 568

The principle can be established by comparing exx. (214) and (215), where we see that the theme in these constructions looks like the object of the verb. The rheme can be a verbal lexeme (perfect participle or gerund), an adjective (ex. (216)) or a prepositional syntagm (ex. (217)).



The link between object nexus and a clausal object is made plain comparing the following pair of examples: (218) kxzwa xa lra naq!$a mAn pre dde ‘He would find one coin

missing from his money’ 1

(219) '1ay dammAt kxzwa ak!ma dd wle nq!$a xa qu&e1fa... ‘And

whenever the king would see that one cluster was missing...’ 641–642

In both examples there is a nexus—a coin missing (he did not find the missing coin, but rather a coin missing) and a cluster missing. Both constructions occupy the same position, namely, the object position. Embedded clauses occur asyndetically (among those it is the most frequent), but also syndetically with d(d) and 1Annu: (220) wa/Ad zla {1! jwanqa mAlle barq'l daw gabbara} ‘when she

saw (that) the youngster stood in front of the giant’ 765

(221) zla {g!ra qr'le} ‘she saw (that) her husband came close’ 826 (222) 'zle qarwwAt mbínn!ke, {sle xa &ra 'gAbe q&é1wle xab'šAt

&ha} ‘Toward morning he saw (that) a bird came and wanted to cut the third apple’ 188

(223) mle qam x  mra uzle {mpAqle xa x'we} ‘He arrived at a

cave and saw (that) a snake came out ’ 627

(224) 'zle ax!ne amad aw r'wa {7&&- pre dde wlu x$e} ‘His

older brother A. saw that his money was finished’ 79

(225) sle ak!ma 'kxze {7&&- br!ne wle t1ra 'xab'ša wle

(A)ld'ke} ‘The king came and saw (lit. sees) that his son was awake and the apple was in its place’ 188

(226) 'gmnAxwa { wle nq$a ham 1aw qu&e1fa r'wa usqla} ‘He

saw that the great beautiful (grape)cluster was missing as well’ 644

Yet another strategy, which is mentioned here on semantic rather than syntactic grounds, is the presentative clause (for the forms as well as differences from the copular wle, see above, §1.5.2). These clauses are not embedded, and may occur independently as well. In other words, they do not occupy the object position syntactically, and are better described as occurring in sequence with our ‘main’ clause. However, semantically, they are



very common with seeing, semantically complementing it and vividly specifying what is being seen by the beholder: (227) kse bbe mbínn!ke 'kxze {br!ne wle &w1a 'xá xab'ša wle

nq!$a} ‘The father comes and sees: his son is asleep and one apple is missing’ 185–186

(228) 'zle {ham 1aw qu&e1fe r'wAt kullu qu&e1fe dd gnqA$wa kuššat

wle garxAt q$a} ‘and he saw: also the largest cluster, which was missing every year, is missing again’ 649

Comparing ex. (225) to ex. (227), and ex. (226) to ex. (228), each pair occurring in the same story and involving the same or similar participants, makes it clear that we are dealing with competing strategies, which are not, however, found in the same syntactic level. The function of the presentative constructions is essentially a narrative function, zooming in to the point of view of the characters themselves, whereas the embedded clauses basically refer to a general, or objective, point of view. Presentative constructions are discussed below, § Saying This lexeme is already discussed above, §§ (basic valency) and (various clausal objects).

3.1.4 Causative and passive By choosing to describe microsyntax according to syntactic relationships, passive expressions do not belong under the completive relationship, unless they take a complement (for instance, the passive of the causative). Their basic form, after all, shows only one relationship, the predicative link. However, since these expressions have a lot to do with their active counterparts as well as with causative constructions, they will be briefly considered here as well. There are three patterns for passive constructions in JZ: 1. The pattern #1sya (l-)+infinitive (or nomen actionis)#: (229) 's la w za 1All!h'n šar/etta ‘Justice was done with them’


2. The analytic formation copular expression and passive participle: (230) u1! xarx$a kulle wza w le mn alaqys dAhwa ‘and this belt

was made of gold links’ 833



3. The bare preterite base:14 (231) t'lu kúd 1arbu '1án &re mpúr 7ki '1án wne nši 'wzi wa$le

‘The four of them sat down and the birds were plucked and the animals were skinned and sliced’ 864

The 2nd construction type is Zustandspassiv, depicting a state, whereas the other types (1 and 3) are Vorgangspassiv, describing a process, namely a dynamic or eventive passive.These constructions are not very common; in addition, another similar strategy, the impersonal, namely, the non-referential 3CP, is used more often: (232) pAšle mux xa dh ye wzeCP &le sAAr ‘He was like someone on

whom magic has been performed (lit. one whom they [Fr. on] have performed magic on)’ 785

(233) 'g! xAlma qam makšla mAn kursAt uk'm 'qam zla mux xá

jerye ‘and in (the) dream they (Fr. on) took her off the throne and made her a maid’ 797

All these have one thing in common—the agent is not mentioned.15 This means that if the lexeme in its active pattern can have two participants (e.g., agent and recipient), when in the passive formation it can have in turn one less, i.e., only the recipient. In the same vein, when a causative pattern is compared with the simple pattern, we have a similar result: the former will have one more participant than the latter, namely, the instigating agent (the one causing the action), the recipient (the object) and a second agent (the one actually performing the action, or on whom the action is imposed). Take a look at the following table:

See Gutman 2008. Only rarely is the agent mentioned in a passive construction by the prepositional syntagm bAt 1z- ‘by (the) hand of’. The following are the only examples I know of: 'sle dwqa 'bse Al maq!ze b7d z mušulmne ‘he was caught and will be burned by the Moslems’ (612–613); 'la ksAt Al qA&la b7t z daw ak!ma ‘You will not be killed by the king’ (716). This agent marking is found with other types of clauses as well: uk'n xá1 br1Aš 1Allan 'maduqlan bAt 1z ak!ma ‘puisque quelqu’un peut nous remarquer et nous faire attraper par le roi’ (656). 14 15



example (234) kut hwe bAš &r$a bse l1xla bAš hayya ‘Anyone (1)


who is healthier will be eaten earlier’ 384





(235) &AptAt daw jwanqa xAlla(1) 1an &arpAsAlqe(2) ‘The transitive

youngster’s bride(1) ate the beet leaves(2)’ 548

(236) g!ra dd ... gmxAl-ø(1)-wa baxte(2) &ha dne

&arpAsAlqe(3) ‘the man who used to(1) ... feed his causative wife(2) three times beet leaves(3)’ 557–558



Exx. (234)–(236) all have the same root (not the same lexeme, since a lexeme is here regarded as root in a given stem which has a certain valency pattern), and differ by their valency. Ex. (235), the active form, can be regarded as our point of reference, having two participants: an agent (=bride) and a recipient (=mangold). Compared with the passive construction in ex. (234), the latter has only one: a recipient (=the one who is healthier). Ex. (236), on the other hand, has three participants: the instigating agent (=man), the recipient (=mangold) and yet another one—the second agent (=his wife). The same idea is found to exist in comparing an intransitive lexeme to its causative construction: example (237) ay dammAt mpAlle !ha kap!ra

fell...’ 89

pattern (1)

‘when the giant


(238) bAr ak!ma(1) qam mampAlle !ha kap!ra(2) ‘The

prince(1) caused the giant(2) to fall’ 89






While ex. (237) has only one participant (=giant), ex. (238) has an additional one, the one who causes the falling (=prince). The last group is similar, but the passive construction with the causative lexeme is adduced as well: example (239) d1Arre







returned to his home ’ (2)



1 or 2

(240) qam mad1Ar-ø(1)-re(2) 1Al d'ke(3) ‘He(2) returned it(1)


2 or 3

(241) sAn(1) mad1!re ‘Let me(1) be returned (here)’ 566



to its place(3)’ 388

1 (or 2)

Exx. (239) and (240) are analogous to the preceding pair, except perhaps the issue of whether the habitually occurring locative prepositional syntagm



should be counted as part of its valency. Ex. (241) is an example of the passive pattern with a causative lexeme, where the agent, the one who causes the returning, is taken out.

3.1.5 Reciprocal and reflexive Reciprocal expressions are found where an action is mutually performed in such a way that the agent is also the recipient and vice versa. They are not expressed morphologically anymore (as in older Aramaic, where there were specialized stems for this), and are constructed using various combinations of xa... xet ‘one... another’, and rarely xa... x!re ‘one... his friend’. The first version of this expression is constructed with a verbal form in the singular, whose agent is xa and its object is xet: (242) (baxtsa d di...) m li x  la mamra lday xet ‘(My wives...) so

I ordered that one should not command the other FS’ 830

(243) (nše...) ux pšle bngza  xt ‘(People...) and one began

biting the otherMS’ 942

Analysis seemingly poses no problems in this version, each part of the reciprocal signal is found in its due position, enveloping the verbal form (analytic constructions as well) as it were, the entire structure resembling any verbal clause, and there is a choice between masculine and feminine in the object position. Yet, these constructions always have a plural referent (my wives, people), and hence we have some logical incompatibility with reference in the following singular verbal form. In the other version, the verbal form is in the plural and the two parts of the reciprocal signal xa... xet occur after the verbal form: (244) Aswa... xa jwanqa ux xamsa dd rba gmabwa x 1aw xet

‘There were... a lad and a maiden who loved one another very much’ 231–232

(245) 'kullu nše pAšlu bAz1pa x aw xet ‘And all of the people began

pushing one another’ 118

This version has indeed a plural verbal form, in agreement with its plural referents; still, how the reciprocal signal is to be explained is not easy: whereas the reflexive pronoun often simply occupies the object slot, this syntagm does not, and despite the fact that similar constructions occur in many languges (e.g., ‘one another’), it is quite impossible to state what ar-



gument slot they take. Consider the following cases, where the two versions of this construction occur with prepositional objects as well: (246) (three wise men) x mArre ta daw xt ‘One said to the other’ 52 (247) (prince and his friend) m rru x  ta daw xt ‘They said to each

other (lit. one to the other) 86

(248) (three wise men) mArru xa ta x!re ‘They said to each other (lit.

one to his friend) 57

Ex. (246) is the first version. In exx. (247) and (248), which belong to the second, the problem seems to be that formally, at least, only the second part of the reciprocal signal is preceded by the preposition, which we could gloss ‘one to the other’. This structure clearly leaves the first element, xa, unanalyzable—it is not quite in agreement with the plural referent, nor is it formally included in the prepositional object. These difficulties are somewhat resolved by a diachronic path common to many languages, where this complex becomes ‘an inseparable whole’ (Jespersen 1924:224): (249) yAmma ubr!na xpAqlu xawxta ‘The mother and her son hugged

each other’ 771

(250) hama mpl'n basAr xawxta mux kalwe 'kalApysa ‘Just go after

each other like dogs and bitches’ 828

(251) mattaw l1al xA$lan mAn xawxta ‘From now on we are finished

with each other’ 828

(252) 'kfu rba sle 'mArru ta xawxta: ‘They were very happy and

said to each other:’ 467–468

This new inseparable whole in the preceding group of examples (xawxta) is made distinct first and foremost by the position of the prepoposition, which here precedes it. It now clearly occupies the position of the object, but at the same time signals reciprocity. Reflexive expressions, where the agent is also the recipient, are simpler. A pronoun, gyn- (~Kurd. jan ‘soul’, see also §, takes the place of the object. The simplest case is when the reflexive pronoun is in the position of the direct object: (253) mundle 1ze lxpqa ‘He cast his hand to her lap’ 880 (254) 'mundla gyna lxpqe ‘She threw herself in his lap’ 883



The object ‘his hand’ in ex. (253) is syntactically equivalent to ‘herself’ in ex. (254). The following pair of examples shows the same for a prepositional object: (255) 'mArra ta g!ra ‘She said to her husband’ 690 (256) 'mArra ta gyna ‘She said to herself’ 884

In exx. (255) and (256) again, ‘husband’ takes the same slot as does the pronoun ‘herself’. In the last pair of examples we see object nexus: (257) 'qam xazy–le bAt julle 1atqe ‘And she saw him with old

clothes’ 684

(258) kxz–la gyne go nam'sye dde ‘He found himself in his bed’


The nexus in ex. (257) obtains between ‘him’ and ‘with old clothes’, whereas in (258) it obtains between ‘himself’ and ‘in his bed’.



3.2 ADVERBIAL COMPLEMENTS Complements that are not inherent to or required by the verbal lexemes are to be considered adverbial complements. The line between object complement and adverbial complements is not easily drawn. Unlike other syntactic positions in a clause, which are limited (as, e.g., the number of objects), numerous adverbial expressions can co-occur in one clause: (259) barbar

palgAdy!m sle 1aw jwanqa mAl'ša jullAt xa mAn xAddamysa 1AmmAt day xAddamta ‘Towards noon came the youngster clothed (with) the clothes of one of her maids (together) with that maid’ 794

In JZ there is no specialized adverbial formation. Besides a small group of ‘natural adverbs’, namely, forms which have only this function (e.g., tAmmal ‘yesterday’, hayya ‘early’), the majority of adverbial expressions are adverbial because of their function. For example, a prepositional syntagm can function adverbially, but also as an object (see above, §3.1.2), adnominally (e.g., xa mAnnu ‘one of them’) and as rheme. The same is true for the syntagm aw y!ma ‘that day’ which can function adverbially (dmAxle 1aw y!ma ‘he slept that day’ [686]) as well as substantivally (1urre 1aw y!ma ‘that day passed’ [723]). For this reason, there is hardly any justification for the term ‘adverb’, and hence in this framework, ‘adverbial’ (syntagm, clause, etc.) will refer to the function of an entity in a given syntactic setting only. The other functions (e.g., adnominal and rhematic), where these so called adverbs do not function adverbially, are discussed where they belong, §§4.1.4 and 1.3 respectively. Adverbial complements essentially modify a nexus or a reduced version thereof. There are many types, groups and subgroups and the general linguistic study of this domain is based, by and large, on semantic criteria. The only syntactically-based classification I know of is found in Goldenberg 1985 §17. Goldenberg draws a line between three basic types of adverbial complementation, based upon their relationships to the components of the verbal complex, namely, the subject index, the verbal lexeme and the nexus. All types are necessarily related to the nexus (as stated above), but two are linked further: the first type, l in the Arab grammatical terminology, circumstantial syntagms or clauses, which describe the state of one of the participants in the clause (subject index or one of the objects; termed “participant oriented adjuncts” as well as “depictives” in the cross-linguistic literature, e.g., Himmelmann and Schultze-Berndt 2005); the second is the



group of qualifying adverbials, which refer to the verbal lexeme. The third (and largest) group, adverbials of manner, have ties only with the nexus. This classificational key is followed here as well. Semantic sub-classification is used secondarily only.

3.2.1 Circumstantial adverbials These adverbial entities are mostly not clauses in JZ, although a subtype thereof does contain a nexus. Thus they are always marked as part of a clause. An altogether different competing strategy, characterized by the presentative clause, is described below as well. There are two types of these adverbials, one containing a simple entity in this function, the other containing an explicit nexus: (260) nše kafn wa bbe ay dammt gmx wle ja za r nkr  wa

hdax múumla rAž bArke '1aw ja1!za hdax xar'pa mu n'ra ‘People were looking at him when he used to hew with the ax around the wood rested on his knee, the ax (being) as sharp as fire’ 509–510

The first adverbial syntagm contains only a participle (múumla), referring to the piece of wood and describing its state. The second syntagm is a nexal complex, in which the ax is explicitly mentioned as the theme, and the rheme is xar'pa. The following example contrasts two adjectives in this function: (261) dt day na wwla kud gšte mnna qamya gmyes. nš t day

bžer, ay dammt k š wa l day na štwa mya, gnabl wa mmt gynu xa qa sa, gmašt wla qamsa ‘The custom of this spring was (that) anyone who drinks from it firstMS dies. The people of that town, when they would come down to the spring to drink water, they would bring a cat with them (and) would let it drink firstFS’ 144

Note that the adverbial ‘first’ shows, by its gender morphemes, the reference to one of the participants—in the first case it is ‘anyone who’, which is conceived as masculine (qamya), whereas in the second case, it is the cat, which is feminine (qamsa). The most representative type of examples consists of perfect participles as well as other adjectives, which, unlike the gerund, show gender and number:



(262) aw jwanqa qam d qwle g bžr 1A*0 1A$$A+ '))!" ‘The

youngster had been caught (lit. they had caught him) in town, (having) comePP MS with textiles’ 789–790

(263) /"$ $A%%, 8###, š'# %š /A&0 ,1A*0 A# 1% ($ $Aš'&

‘To some of them he gave the role of sinners, (having) come PP CP to court for judgement’ 556–557

(264) pAšlu mn!xe bAd d! naša hdax twa rAš xa taxta '$wa lbArke

'bsya 1Alle bAd daw ja1!za hadax mux br' ‘They began looking at this man, sitting on a chair and the (piece of) wood on his knee, coming at it (=the wood) with the ax like gunpowder’ 510

The preceding examples all contain circumstantial adverbials with the same lexeme, ‘come’. Exx. (262) and (263) have a perfect participle, masculine singular and common plural respectively, whereas ex. (264) has a gerund. All three have the same function,16 the difference being the actual reference to one of the participants: in ex. (262) the participle is referential to the youngster, in ex. (263) to the people, whereas in ex. (264), we can only guess the reference to the man. The following example shows very clearly the tie between the adverbial and the participants in the main clause: (265) 'lAbbe pAšle hš mharh!re '1ne mn!xe bAt gawdAt 1Astatte hdax

šulxya 'pur1ya (šulxsa 'pur1sa) ‘He (lit. his heart) was still having improper thoughts and his eyes were looking at his mistress’s body, him/she/it (being) stark naked’ 786

šulxya 'pur1ya lit. ‘naked and uncovered’ (masculine) occurs originally in the text, while šulxsa 'pur1sa (feminine) is added on the left of the text. The former could only refer to the man looking at his mistress (in both senses), that is, ‘him (being) naked and uncovered’. The latter, on the other hand, could refer to either the mistress or her body (which is feminine as well). The adverbial syntagm contains two adjectival forms in adverbial function, and the syntactic link is made clear by the form of the adjective. The negative counterpart of these circumstantial syntagms occurs only with the infinitive form, and is to be considered the negative form of the gerund, because, much like the latter, it does not occur in clear-cut substantival positions: 16

The syntagm in ex. (263) can also be analyzed as adnominal.



(266) hama lušle mt sle qam 1ze 'md wle qadAr dde 'b!le hama

qam srAqle la mn!xe bAt u n!ra ‘So he put on whatever he found and whatever fit him, and combed his hair, not looking at any mirror’ 796

(267) bale x  mnnu dw le bž ra qa ršle lgyne žn gva hama la

mn!xe 1l basre mxle lgyne '1rqle ‘But one of them who was farther away suddently became aware and, not looking behind him, he upped and ran away’ 900

(268) 1wa pAšle bAkxka 'bAgm$a dda la šqal &AmAt xabra 'la fham

ma/ne dde ‘He began laughing and smiling, not taking in the essence of the word(s) and not understanding its meaning’ 867

(269) dammAt &li lbsa, hž la šryAt karta, kxzAn nhAjle xmra

'mpAlle ‘When I arrived at home, not yet untying the load, I see (that) the donkey panted and fell’ 922

It is important to mention that this syntagm, la+infinitive (as well as other substantives and pronouns), also has the meaning ‘without’ (see below, § Ex. (269) is somewhat different due to the temporal adverbial included in it, but it is nevertheless part of the group. The relationship between the gerund and its putative negative form, which also means ‘without’ is perhaps best illustrated by the following example: (270) mandla 1atta  s!tAnta go baar. hakan la, barqtAn 1AmmAt

ma1!re dda go baar Al1la xet ‘Throw now the old woman to the sea. If not, you will sink while (lit. with) transfering her across the sea’ 273

The preposition 1AmmAt does not otherwise occur with the infinitive form. In addition to the simple circumstantial syntagm, we also encounter nexal syntagms (see above, ex. (260)). Compare the following pairs: (271) 'msupyla hdax mAlta ‘And she died standing’ 853 (272) 1ya hdax mAlta, ham mírrle: ‘She standing, nevertheless she

told him:’ 852

(273) mpAqle x'we mAn d'ke dwqa ppumme xá dAhwa ‘The snake came

out from its place, holding a gold coin in his mouth’ 453

(274) mpAlla dAbba l1ar1a hš 1za duqta bAt bbAt d! yla ‘The she-bear

fell to the ground, its paw still holding the child’s father’ 752



(275) wa pAšle bA&1ya 'pša pk!van d!hun u1ni bA&1ya 'mans!re

dAm1e. wa zAlle xá-1l dunye u1ani bA&1ya b1la xet ‘He started searching, still missing them, (while) they (were) searching and shedding tears. He went to one side of the world while they (were) searching the other side’ 845

Exx. (271) and (273) both have a simple syntagm consisting of a participle (and its complements), whereas exx. (272) and (274) are complex, a whole nexus obtaining between a (pro)nominal entity (‘she’ in the former and ‘her paw’ in the latter) and the participle. The introduction of a body part with a pronominal attribute that is referential to the participant is another way to compensate for the fact that the gerund is not explicitly marked as referential: (276) 1urri 'lubbi bArpa ‘I entered (with) my heart startled’ 407 (277) 'thla g! daw !l hdax lAbba bArpa ‘So she wandered in the

desert (with) her heart startled’ 901

(278) pAšla mn!xe 1Abbe umírrle hdax 1na mans!re dAm1e ‘She began

looking at him and told him, her eyes shedding tears’ 895

The position of the circumstantial syntagm is usually following the main clause, but we have seen it positioned initially (ex. (272)), and it can also occur inside the clause; in this case, immediately following the topic: (279) '1 baxta hdax qArta 'mtaxm!ne taxmne... zídlla xa taxmn

xsa ‘But the woman, grieving and thinking thoughts... a new thought surged in her’ 878

More examples of this type are treated above, §1.4.3. Note in particular the special connection the adverb hdax (otherwise ‘so’) has with circumstantial syntagms: it is not obligatory, but it is nevertheless typical. Another, rare type of construction, with kšqAl or pšqAl forms, is similar to the Arabic l construction: (280) slu kutru ak!me k&1e Al yalunku ‘Both kings came looking (lit

they look) for their children’ 126

(281) pAšla bsqa 'bAkwsa &'rne 'gAre u1na la g/alqa 'lak tafqa bAt

u 1azla '1asya ‘She began ascending and descending mountains and summits, (while) her eye does not meet any passer(s)-by’ 855



(282) pAšla bAkypa g! p!)e 'b!la bA/lqa bA pse 'gub1ne ‘She began

bending over his face, (while)17 her hair (is) touching his face and his forehead’ 868

Comparing exx. (281) and (282), we see that the syntagms u1na la g/alqa ‘her eye does not meet’ and 'b!la bA/lqa ‘her hair (is) touching’ both occupy the same position and perform the same function. The difference being that the former syntagm is a clause, while the latter is a nexus, but not a clause. Such interchangeability of a clause with a non-clausal entity is one of the definitions of subordination. The form otherwise used to express the future is used in this function as well: (283) ay dammAt m!rAmli bmatwAnne lqzla, t1Arra 'mírrli ‘When I

lifted (it) intending to put it to her throat, she woke up and told me:’ 415

(284) 'sle go xa qurnsa 'qam twArre bd lAbbe bwez n'ra ‘He took it

to a corner and broke it, intending to make fire’ 361

(285)  brta wla t'ta g! xne, šApp!lta murme hl nf&anga 'ha xá-

bAhna yn tr1 bxa&ya 1Amme ‘(And lo) the girl is sitting in his lap, the hems (of her dress) lifted up to her loins, and she is about to sin with him at any moment’ 811

The kšqAl forms in the preceding examples (exx. (280) and (281)) are concomitant circumstantial clauses, made conspicuous by their disagreement with the story line, which itself generally consists of preterites. The pšqAl forms (exx. (283)–(285)), in the same syntactic conditions, are circumstantial clauses, which in addition signal intention. A completely different strategy is the presentative clause (see above, §1.3.3). It is different because it does not occupy the same position as the circumstantial adverbials inside the clause, but rather is joined to the clause, either preceding or following it: (286) ax!na r'wa wle bzla bAthya go barya, zle x saw!na...

‘(While) the older brother is walking wandering in the desert, he saw an old man...’ 71

Theoretically, the substantive and gerund could be analyzed as ellipsis of the copula pAšla; however, the substantive b!la is masculine, so the probability that this is an ellipsis is not very high. 17



(287) wla bAš&xa julle 'bAr ak!ma qímwle mAn šAnsa ‘(While) she is

hanging clothes, the prince had woke up from sleep’ 499

(288) 'škAlle bAd 1xla. wle bxla 'slu 1Alle mšalxne ‘So he began

eating. (while) he is eating, robbers came upon him’ 524

(289) xá y ma zlle o yla l n  r '&ra '1Azza wlu Amme ‘One day the

boy went hunting (while) the bird and the goat are with him’ 124

(290) zAlle basre 1aw $wya uwle bAr1la mAn zdo1sa la 1xAlle 1arya

‘The woodcutter went after him (while) he is shaking with fear lest the lion eat him’ 667

The presentative clause in this function is not a temporal clause as the translation with ‘while’ might suggest, nor is it subordinate (as one could claim in similar cases in Arabic). Note that these clauses are interconnected, regardless of the relative order in which they occur.

3.2.2 Qualifying adverbials The second syntactic group of adverbials consists of adverbial expansions that refer to the (verbal) lexeme. This link can be typically shown by the use of internal object. The latter is a syntactic mechanism by which a cognate infinitive, or nomen actionis, co-occurs with the finite verbal form, and is further expanded by an adjective: (291) qam n1Asle x'we n1sa rba r'wa ‘The snake bit him a very

strong bite’ 635

(292) mukle q'ya ‘He spoke mightily’ AS 1337 (293) šq!l 1! nar1a mxle go rši pkulla q'wAtox ‘Take this ax and bang

it on my head with all your might’ A 1-10

(294) z1pli 1Azza u1na ši grAšli 1Amma pkulla q'wAtti ‘I pushed the

she-goat and I too pulled together with her with all my might’ 402

Ex. (291) has an internal object, ‘a very strong bite’. This syntagm is interchangeable with other adverbial syntagms, as we see in exx. (292) (‘mightily’) and (293) and (294) (‘with all your/my might’). This interchangeability makes it clear that this so-called internal object is in fact an adverbial syntagm rather than an object (the normal object of n1sa being the victim of biting). The following examples are similar, at least in principle:



(295) mundla kutru 1zsa lqzle 'm!$Arra xa a$ra gamo1ta ‘She

threw her hands to his neck and sighed a deep sigh’ 819

(296) 'kut gnpAqwa go 1urxe láswle flta qq&ílwle xa q&la kAret

‘and anyone who would come out in his way, there was no escape for him, he would kill him an ugly killing’ 964

In both preceding examples the syntactic as well as semantic corresponding constructions would be ‘she sighed deeply’ and ‘he would kill him in an ugly fashion’. In the following pair the so-called internal object is not strictly of the same lexeme: (297) zAlle 1aw g!ra $wya 'š&Ale xe daw dra behne d1Arre '&we1le xá

šAnsa qwsa ‘The woodcutter went and lied down under the tree, his breath returned and he fell asleep deeply (lit. one strong sleep)’ 627

(298) pAšla maxm!xe 'bAnšqa dde 'bAšypa psa bAt pse 'qzle '1wa

&w1a xa šAnsa qwsa (yqurta) 'latle hyi mAn bye falake ‘She began feeling and kissing him, and rubbing her face in his face and neck, he having fallen asleep deeply (lit. one heavy sleep), unaware of anything going on’ 871

Note that while ex. (297) has a verbal form, ex. (298) has a circumstantial nexus ('1wa &w1a ‘he (having) fallen’), but the internal object occurs just the same, applying to both the nexus as well as the verbal lexeme. The following case is considered to belong here, although the expansion of the lexeme is effected by different means: (299) mAn qarte $rAxle xa $raxta upqe1le ‘Out of his grief he screamed

one scream and exploded’ 395

(300) kmle dd zlan br!nni xa naqla xt ‘It is enough that we saw

our son once more’ 606

The form $raxta can be considered a true object, but also an internal object. If we choose the latter, we can say that it is expanded by the QUANT xa, making it syntactically and semantically equivalent to the idea subsumed in ‘once’, as it is in the second example. Having adduced a small (but sufficient) number of examples for the internal object, we can look at common adverbs and state that they belong to this group of adverbs, which modify the lexeme: (301) 'rba zde1le ‘and he was very afraid’ passim



(302) la brlle i mAndi r jrAla 'a sApse bAt xa kutwa ‘Nothing

happened to him except (that) his lip was hurt a little by a thorn’ 493

In both examples the adverbs rba (‘very, a lot’) and 'a (little) theoretically correspond to an internal object construction, such as ‘a great fear’ or ‘a small injury’.

3.2.3 Adverbials of manner The internal syntax of those adverbial syntagms which are constructed as an attributive complex is discussed above under §§2.1.5 and 2.3.4. In this section, the external syntax of adverbial constructions is evaluated as well as the distinctive features of each adverbial group. This group is subdivided on semantic grounds, which however often conform with syntactic criteria. There are formal differences between the following groups: purpose clauses; temporal expressions and clauses; comparative expressions and clauses; indefinite quantifier expressions and clauses and locative expressions. The groups themselves are not always classifiable on a syntactic basis. Purpose clauses Purpose clauses are semantically a combination of causality and will or necessity. They are sporadically constructed with the infinitive: (303) mš!dArre bakko ulumwse Al q&l mamo ‘B. sent servants to kill

M.’ 328–329 (see also 664)

(304) na la sli ta 1xla ‘I did not come (with intention) to eat’ 317

However, in most cases they occur as a finite clause, always with a subjunctive form (occasionally with -wa). They are introduced either asyndetically or syndetically, by a number of different exponents. The first one is ta (otherwise ‘to, for’): (305) bsAn kAsl!xun ta xzAnna 1ay brtAt muklax 1Alla ‘I will come to

you (so that) I (can) see the girl you talked about’ 777

The unity of the purpose clause, regardless of the particular exponent used, is made apparent in the following pair of examples: (306) gnablwa 1Ammu zawde dd lle uy!ma ta la sanqi l1xla 'štya

'dla 1zi u1se rba ‘They would bring provisions of night and


CHAPTER THREE day so they do not need food and drink and that they do not go and return too much’ 462

(307) mAlle hl nše mbárbAzi ta 1se (d1se) ‘He waited until the

people scatter to so he (may) come (that he come) 840

In ex. (306) we see that two purpose clauses are interconnected. The first is introduced by ta while the second by d-, with no appreciable difference except perhaps the fact that d- potentially has more than just one function (see above, §2.3). In ex. (307) the informant himself supplied an alternative form to ta 1se—d1se. (308) kem mš!dArre x n&!ra dd šme1 ma wlu mak!ye ‘The ruler

sent a guard so that he hear what they were talking about’ 334

dd~d- is quite common in this function, but its multifunctionality may cause ambiguity. The following example contains three purpose clauses: (309) gzAlwa g!ra q&e1wa $we... sapx&Ar mzbínwlu 'dd 1axlwa

wa ubaxte ‘The man would go to (lit. so that he) cut trees, in order that he sell them and that he and his wife would eat’ 199

The third and second purpose clauses in ex. (309) are headed by dd and sapx&Ar (to, for) respectively. The following example has the combination of ta and sapx&Ar: (310) durtyom bbe zAlle Al $wAt gyne ta $apx&Ar xze ma brle mAn

br!ne ‘The following day he went to his tree in order to see what happened with his son’ 635–636

The last group contains asyndetic expression of purpose, which consists of a bare subjunctive clause (as already seen in the beginning of ex. (309) above): (311) qAmla 'zAlla xazya kle ‘She got up and went to (lit. [so that]

she) see where he is’ 232

(312) t'lu xe xa dra nxi ‘They sat under a tree to (lit. [so that]

they) rest’ 266

(313) marpli 1zen ‘Free me (so) I (may) go’ 410

The table below summarizes the various exponents participating in this group. Note that the clear, consistent structure and syntactic characteristics



of this group make it a distinct syntactic group. Additionally, these expressions in principle follow the main clause: purpose clauses and expressions type

nucleus A#

attributive construction



finite form


(ta) sapx&Ar asyndetic clause




example # (303)


(305) (306) (307) (306) (307) (308) (309) (309) (310)

(311) (312) (313) Temporal syntagms The following survey covers the more common temporal adverbials in JZ. The simple ones are forms such as xá y!ma ‘one day’, aw y!ma (=bAd daw y!ma) ‘that day’, as well as other time spans such as yarxa ‘month’, šta ‘year’. Similar is the entity naqla as well as wa/da both meaning ‘time’. These entities can function in substantival positions as well as (perhaps even mainly) in adverbial function. Other entities are more restricted and are found only in adverbial function, such as xa ga, tre ga ‘once, twice’, tAmmal ‘the day before’, bAne ‘the day after’, atta ‘now’, hš ‘still’, etc. The fact that many of these can have more than one function corroborates the claim raised above that adverbiality is merely a syntactic position. Some of these simple adverbials (ex. (314)) are expandable by an attribute (simple or complex, ex. (315)) or even by a relative clause (ex. (316)): (314) y!m tre garxAt musla xmsAt day (xamsa) baxta 1an &arpAsAlqe

mux tAmmal ‘(The) second day (lit. day of two), the mother-inlaw of this girl brought mangold again, like the day before’ 548.

(315) y!mAd &li go d balad, /omri 1Asri 'xamšá šAnne wle ‘(The) day

I arrived in this country, I was twenty five years old’ 424

(316) xá y!ma dd wwlu go day bžer ‘one day (in) which they

were in town…’ 51

Besides these, one finds some specialized temporal clause nuclei such as mqabAl ‘before’, basAr ‘after’, ay dammAt and wa/Ad~wa/dAt ‘when’:



(317) mqabAl 1zAl mírrle 1arya ‘Avant qu’il s’en aille, le lion lui dit:

(Before he left, the lion told him:)’ 677

(318) basAr zAllu kullu nše mírrle ‘Après que tous les gens s’en sont

allés, elle lui dit: (After everybody has left, she told him:)’ 173

This pair of temporals is interesting: like the Romance languages, the one whose value is ‘before’ takes the subjunctive (which, since it is obligatory, it has no value of its own, note that the English parallel would have been ‘before he left’), while the other, whose value is ‘after’, takes the indicative. This is important because the indicative has a temporal spectrum that does not exist with the subjunctive, so basAr occurs mostly with a preterite, but occasionally also with the present. There are other small differences: while basAr is compatible with many attribute types (pronouns and substantives as well as clauses), mqabAl is compatible only with infinitive and clauses, and barbar ‘towards’ is almost its alloform de facto, occurring with temporal attributes (barbar palgAzyom ‘towards noon’). ay dammAt ‘the time when’ is the generalized head for temporal clauses, occurring with all verbal forms, including the subjunctive. In such contexts, unlike after mqabAl, the subjunctive has its own value since it is opposed in this case to the other verbal forms, viz., it signifies the future. The form that normally signifies future, pšqAl, basically does not figure after ay dammAt (the entire issue of tense in temporal as well as other embedded clauses is treated under macrosyntax, § Compare the following pair (already discussed above, §2.2): (319) {ay damma}

said:’ 527

mArre ‘{(At) that moment} he

(320) {ay dammAt šme1le hatxa}, $rAxle 'mArre ‘{When (lit. the

moment) he heard thus}, he screamed and said:’ 288–289

The adverbial ay damma occurs alone (ex. (319)) or expanded by a clausal attribute (ex. (319)). Another temporally-related adverbial nucleus is hl ‘until’, which poses some problems. In its simplest function, it occurs with the preterite forms šqAlle and qam šqAlle, imperfect kšqAlwa, present kšqAl and subjunctive šqAl (whose value is future in this paradigm, as with basAr): (321) Alla xapa bAtaxnAtti hl amran y 'xala$ ‘Grind me just a bit,

until I say that’s it’ 175



(322) qam mazvArra hl mArra: y ‘He turned her until she said: that’s

it’ 175

(323) ukmAlwa hl gnapqwa 1an talmze mAn knAšta ‘He used to wait

until the students would come out of the synagogue’ 576

(324) kma Ahwa khet hl g/lAqlox xa pra ‘How exhausted you

(must) get until you earn a coin?’ 487

Negation is not very common after hl, and when it occurs, it means something slightly different: (325) la gzAt hl la makínnlox mt sla brši ‘You will not go before

I (lit. until I do not) tell you what happened to me’ 375

(326) hl la 1se ak!ma d!x'n mn!še 'nbAlli bAd 1ze 1na laksAn

1AmmAt uxxa mAnn!x'n ‘I will not come with anyone (lit. no one) of you before your king comes and takes me by his (own) hand(s)’ 617

This construction, however, occurs mostly with the subjunctive, very much like mqabAl. There is only one such example with the preterite: (327) qAmle gra, pAšle mamp!le kallše hl 1! faqr la mArre kmle ! gra

la mAlle ‘So the arrow began pushing corpses, as long as the poor man did not (=before) say “enough”, the arrow did not stop’ 526

Another temporal nucleus is  ‘as long as, while, whenever’: (328)                 anta bAd bani

1dam ‘Whenever I (will) look at my severed tail (or: my tail severed) I will remember that there is no trust in humans’ 638

 w le hdax q ra lbbe tw ra, zídl le xa (329) axn day baxta mmd k!van xet ‘This woman’s brother, while he was grieving and broken-hearted, one more pining pain grew in him (lit. added)’ 854

The last examples in this group are possibly for asyndetic temporal clauses. The first has an explicit background: (330) 1ay dammAd &we1, hama šqulla... q&u1le rše ‘When he falls

asleep, just take it (=the knife)... cut his head off’ MA 10.5



Initially, temporality is explicitly marked by 1ay dammAt. The following time the same issue is mentioned in this text, there is only a subjunctive form in initial position: (331) gAmra: 1wa &we1, b)abanne ‘she said: let him fall asleep

(=when he falls asleep?), I will slaughter him’ MA 12.2

This is very irregular; this pattern (#šqAl—pšqAl#) basically denotes (concessive-)conditionality (see § below as well as Cohen 2007:168– 174). However, the subjunctive in JZ is sometimes used asyndetically in all sorts of embedded clauses. The following example is similar: (332) mírrle 1ay brta 1se bbi makn 1Amme 'ysál!xun xa-šapsa-xt

‘The girl said to him: let my father come (=when my father comes?), let me speak to him, so come next week’ 543

For the time being, since there are merely two such examples of a potential “temporal” subjunctive, we cannot be sure nor can we characterize this subjunctive vis-à-vis the subjunctive form in the (concessive-)conditional pattern (which is clear beyond any doubt). In both exx. (331) and (332) it is not even established whether we have an embedded clause or a part of a chain. In the case of the [concessive-]conditional pattern, it is a sequenced, rather than an embedded clause. The point of this entire discussion can be summed up in comparing ex. (319) to ex. (320) above, or in the following table: example

type of syntagm

(333) hatxa 1uzle {rba wa/da} ‘He did this (for) {a long time}’ 579

simple adverbial

(334) msle {wa/Ad palgAzlal} ‘Bring it (at) {midnight time}’ 929

attributive construction

(335) qyAmwa y!m by!m {wa/Ad $rxAt šammaš} ‘He used to get up everyday (at) {sunset time}’ 18 772

infinitival attr. construction

(336) {wa/Ad zlu nš mAjlAs} qam d!qlu zdo1sa ‘{When the people of the council saw}, fear seized them’ 769

clausal attr. construction


Lit. the call of the caretaker of the synagogue.



The adverbial syntagm shows various degrees of complexity, but all, regardless of their complexity, share the same function and relationship with the nexus. We have seen presentative constructions competing with clause-level functions, e.g., with object nexus and object clauses (§ above), or with circumstantial adverbials (§3.2.1). Here too we come across a similar phenomenon. In the material told by the informant S one finds initial temporal (as well as other) adverbials occurring as the rheme of a presentative copula: (337) basAr xá yarxa 1uzlu xA&be ‘After a month they had a betrothal

ceremony’ 547

(338) wla basAr xakma y!me r1Ašle 1Annu bAštum le ‘Lo, after several

days, he felt that he is better’ 312

(339) bAt sa/a tre1sar bAt palgAzlal mš!dArre ‘At twelve midnight he

sent...’ 557

(340) wla bAt štAt š!1a wla t'ta xá-y!ma xa-rše ‘Lo, in the seventh

year, she is sitting one day near his bed...’ 687

(341) dammAt m&lu kxze mAjlAs

rapsa jmA/ta ‘When they arrived, they see a large assembly gathered’ 759 +

(342) wla ay dammAt t'lu Al štyAt da/wa rba nše /Azme wlu ‘Lo,

when they sat at the ball, many people were invited’ 327 (cf. ex. (320))

Since in the material coming from the same informant the two options occur, we assume there must be some functional difference. At this point, however, we cannot determine what it is. Comparative adverbials Comparative adverbials are introduced by a nucleus—mux ‘like, as (if)’, m&ut ‘as’ and kudax ‘as’. The only one of them that is compatible with an attributive pronoun (in other terms, capable of functioning as a full fledged preposition) is mux, having an alloform for these occasions, muxws-: (343) m!nAxli 'a Al 1zlAt 1Arba '1uzli muxwsu ‘I watched for a little

bit the walk of the sheep’s walk and did like them’ 388

mux is a nucleus for substantives as well as pronouns:



(344) na baxta wna gAbna axlna mux kullu nše ‘I am a woman, I

want to eat like all people’ 59

(345) qam &a1nli mux xa &1en xa tallsa ‘They loaded me like

(some)one (who)19 loads a sack’ 382

Finally, it occurs with a clause: (346) qam 1!zla gynu xašim 'štAqlu mu la brle A-mAndi ‘They

pretended (lit. made themselves) to be naïve and kept quiet as if nothing happended’ 823

In this case there is a clear notion of an imaginary situation (‘as if’), but this notion is not expressed by a consistent exponent, and therefore cannot be described as a value (which by definition comes along with a consistent exponent). The same structure could also have the neutral notion of as: (347) zAlle u1uzle mux mírrle bbe ‘He went and did as his father told

him’ 358

m&ut is the nucleus form of m&o ‘how’. It is attested only with clauses: (348) uzle !ha 1Arba m&ut mírrle ‘The sheep did as she told him’ 206

(also 207)

Note that exx. (347) and (348) are practically synonymous. kudax is very similar, attested with clauses only: (349) lha mabhArra Allox kudax qam mabAhrAtta Alli ‘May God shine

upon you as you shone upon me’ 193

(350) b n baqle kuda hya muqdar sr  šnne har m en lb s b bi ‘(If)

I go by foot, as much as it may be, for twenty years, I will arrive at my father’s house anyway’ 425

The difference between the two examples ((349) and (350)) seems to be due to the opposition between the subjunctive and preterite forms. Causal expressions Causal expressions have several exponents; the most common is uk'n. Other specialized, less common exponents are mdam and unki. NonThe syntagm xa &1en is analyzed as a nominal syntagm (pronoun and adjectival clause) for various reasons, see §4.1.3. 19



specialized exponents (guest-exponents) are the free will quantifiers and the otherwise purpose exponent, $apx&Ar. All of the above are nuclei of a clausal construction. Simple attributes can be found with $apx&Ar, but they are not distinct from purpose-denoting syntagms—the gloss for both is ‘for’. uk'n occurs with a variety of forms. The subjunctive form does not figure in these clauses: (351) hle uk'n zAlle 1urxa rba rauqta ‘He got tired because he

walked a very long (lit. far) way’626

(352) Adyo 1na bzAn mn!ži Al knAšta uk'n rwli 'pAšli yaq'ra ta &1na

‘Today I will go by myself to the synagogue, because I grew up and became too heavy to carry’ 583

(353) nox Al gAynox, uk'n bakko gAbe q&Aox ‘Be careful, because

Bakko wants to kill you’ 326

(354) kan zamrAt hl m1bAt lag napqAn xá-gar xet mAn d'ki uk'n k1en

k&1At Al qA&la ddi ‘(Even) if you sing until you die, I will not come out again from my place, since I know (that) you are looking to kill me’ 457

(355) lak qa&axle Adlal uk'n xAtna xsa–le ‘We will not kill him

tonight because he is a fresh bridegroom’ 329

As becomes clear from the examples, a variety of verbal forms are attested in causal clauses. Despite the fact that the subjunctive form does not figure in these clauses, some kind of epistemic modality may nevertheless be found in them. Contrast the following, all pointing to the future: (356) uk'n qam mahli Alla bzabAnne ‘Because he made me tired, I

will slaughter him’ 234

(357) 'škAlli mxakp!re xa d'ka ta haybi napqAn ... uk'n la kwbi

marmAnne aw fta &'ra drAš pummAt škafta mn!š gyni ‘I began digging a place so I could get out... since I would not be able to lift the big rock off the opening of the cave all by myself’ 402–403

(358) har bAd y!ma lágbarya laq&axle uk'n xá1 br1Aš 1Allan 'maduqlan

bAt 1) ak!ma ‘Anyway, during daytime it is impossible for us to pick it, since someone may notice us and have us caught by the king’ 656



While the causal clauses in ex. (356) points to the past, exx. (357) and (358) point to the future. The first shows resolution, the second (ex. (358)) points to a mere possibility. A slightly different case of uk'n occurs with a prohibitive main clause: (359) /AjAzli mAnna gAbAn mandAnna. mírr'le: la mandAtta, uk'n bsax Al

qA&a ‘I got tired of her, I want to throw her (away). They told him: “Do not throw her (away) or we will be killed”’ 269

(360) la msAtta 1Ammox  baxta s!tAnta, uk'n la kwbox 1!rAtte

xaw!ra ‘Do not bring her with you, this old woman, or you will not be able to cross the river’ 272

Note that the causal clauses in the last three examples are somewhat different in that the information contained in the causal clause is relevant only if the lexeme in the prohibitive (or in the negative expression, as in ex. (358)) does take place. It is as if the causal clause contains an ellipsis of a conditional protasis, e.g., for ex. (360): ‘do not bring her with you, because [if you do] you will not be able...’, but the logical sequence is not harmed: the fear that someone might notice them is the reason for not being able to pick the grape cluster during daylight in ex. (358), the fear that they may be killed is the reason for not throwing the dead old woman in ex. (359), etc. Rendering this causality by ‘or (else)’ seems to capture what they do following a prohibitive. The next group of causal expressions consists of the indefinite nuclei md ‘whatever’ or kmt ‘however’ and a clause. In adverbial function, this syntactic form has three different notions: causal, concessive (see § and the presupposed degree or amount type (see § The causal and concessive notions are basically distinguishable on the basis of context, rather than by formal features. Together with their attributive clause these indefinite nuclei constitute what is termed ‘free choice quantification’, which in this particular case is causal in nature: (361) ! jwanqa 1ne )A""a, bale kmad wle lAbbe $Apya la xš'le u 1nAtta

xr'ta g! lAbbe ‘The youngster looked, but so pure was his heart (that) he did not have any evil intention’ 819

(362) na zli flna baxta 'md mp lla qam l bbi md sq lta wwla,

xr'li “I saw some woman, and so much did I fall for her, and so beautiful was she (that) I fainted’ 500–501

(363) 'mahne dde md agd wwla 1urxAt &ha llwsa '&ha

y!msa qam 1!zla bAd xa lle 'bAd xa y!ma ‘His mare was so



fast (that) it did a way of three nights and days in just one night and one day’ 276 The adverbial clause whose nucleus is (k)md cannot be rendered with a parallel structure in English, since English uses an entirely different structure in which the idea of so much is the main clause and the rest is not.20 In JZ it is the other way around: the free choice quantifier is the nucleus of the adverbial clause in question. The adverbial syntagm with this notion can occur not only relatively initially (ex. (363) has it after the topical segment), but also following the main clause: (364) gAmšpwa 1anya tre baxtsa mux xamsa kmad wlu spahAn ‘The

two women looked like maidens, so beautiful were they’ 772

(365) 1!qad bxle 1ay dammAd hwle, Asse ldargAd s&ambul zAlle, mad

wle zqa ‘He cried so much after he was born, his voice reached the gates of Istanbul, so strong was it’ SAG 3.18

Exx. (361)–(365) all have an adjectival rheme in the range of (k)mat. No such rule, however, can be formulated in view of ex. (362), which has an additional verbal form. The following example type is somewhat different: (366) uwla mt &pla 1Abbe kšAfle &la ‘And lo, so much did she nag

him (that) he told her (the secret)’ 673

(367) wla mAn mt &1le rba, tfAqle rAš xá nunsa ‘And lo, so much did

he search (that) he found a fish’ 727 (and similarly 577)

The main peculiarity of this type is that it occurs as the rheme of a presentative, which makes it different from any other adverbial. It is nevertheless adduced here because, as we have seen above (§§3.2.1 as well as above), this is a competing strategy, attested specifically in the texts coming from informant S, where initial adverbials (mostly temporals) are sometimes preceded by a presentative. Here it is comparable to the cases in exx. (361)– (363).

Curme 1931 §29.2 classifies these as ‘degree clause of modal result’. However, the English structure is different from our JZ material, the result clause being introduced by ‘that’. Contemporary Hebrew, on the other hand, has a mechanism very similar syntactically—merov (še) lit. ‘for much/many of (that)’ which always heads a causal adverbial. 20



mdam is a specialized causal exponent,21 just like uk'n, but much less common: (368) m dam bz!nAtti julle sqle 1slox 1Ammi maxwínnlox ‘Since you

intend to buy me nice clothes, come with me, I will show you’ 250

(369) bale fhAmle, m dam la d1Arre br!ne hl mbAn!ke Al bsa, pye

m1uzya bAt 1azyAt x'we ‘But he understood (that) since his son did not return home till morning, he may have been hurt by the snake’s harm’ 636

The clause with mdam always precedes the main clause. There is one example with $apxa&Ar that is causal: (370) rba jgArre Al xarakara dde '1Al kullu sardre dde apx 7r la

ze1lu qrle 1aw ammaš ‘He raged at his magician and at all his chiefs because they did not know (how) to read the book’ 598

This is a special case, which however is based not upon Sprachgefühl, but rather on syntax: this is the only example where $apx&Ar is not followed by a subjunctive form. The putative reason for this flowing from one group to the other is that in the non-clausal domain, $apx&Ar actually neutralizes the difference between causality and purpose, as does for in English.22 The exponent Anki is very rare: (371) gnablwa 1Ammu zawde d lle uy!ma ta la sanqi l1xla 'štya

'dla 1zi u1se rba, 7nki wa/da d!h'n rba gAran wle 1Allu ‘They would bring provisions of night and day not so they do not need food or drink and that they do not go and return too much, since their time was precious to them’ 462 Concessive expressions The connection between free choice quantification and concessivity is mentioned in König 1988:153 and described in Maraldi 2002. In JZ, most of the cases of a free choice quantification adverbial are concessive. To show the 21 Sabar 2002:209a translates, in conformity with Arabic m dma ‘as long as’; however, all of the examples in PT are rather causal. 22 bbi wle &w1a u1na lágmate1rAnne sapx&Ar!xun ‘My father has fallen asleep and I will not wake him up for your sake’ 345



difference from the former type, denoting causality, the following example has both types, causality and concessivity: (372) kmád wla  akumta ql fahumta  d ra mt šw šla ala  lla

mAn kmát xlle 1Alla 1áw jwanqa ‘As much as this queen was wise and smart, this time she got confused and lost, so much did she like the youngster’ 796

The fact that both notions can cohabit the same sentence is evidence that these are indeed two distinct types. With this type, the adverbial clause is usually initial, but there is one case in which it occurs at the end: (373) hš la šme1le mAm pummAt x dla 1wAz gzAnda mAn baxte kmad

h!ya bš '1šAt-yAl ‘He had never heard from (some)one who does not complain about his wife, however good and wonderful she may be’ 517

It occurs, as we have seen above with causality, following the topic: (374) !ha ulma mt gwAzwa haw'sa AmmAt bAr ak!ma, br akma

g xš wa  nnu l xr w sa d de wwle ‘This servant, whatever favor he would do for the prince, the prince would think that he was out to harm him’ 93–94

But the most common form is that where the adverbial clause precedes: (375) 'mt zavret, la gxzet u darga dd ya1let Alla ‘As much as you

turn around, you will not find a door (in) which you (can) enter’ 194

(376) md 1uzla, la t'le ‘Whatever she did, he did not sit down’ 204

day  a la zle ammaš dde ‘As much (377) mt  le go kullu dukn

as he looked in all of the places of that room, he could not find his book’ 592

The adverbial clause could be complex: (378) md 1uzle la 1uzle 'mt pAšle bAnza dde, la-rAzle mAn d'ke

‘Whatever he did, and as much as he was poking it (the donkey), it did not move’ 922–923

A peculiarity of this type only is that the adverbial clause is often separated from the main clause by the connective u-. The following example contains two parallel sections, one with, the other without u-:



(379) mt &1le go kulle š'qa la zle u /attla. m/!jAble. mb!qArre

dAkandre ma brle Al kullu /attle: mt wle bA&1ya, lwe bAzya u /attla ‘As much as he searched the entire market, (lit. and) he did not find any porter. He wondered. He asked the shopkeepers what happened to all of the porters: as much as he is searching, he cannot find any porter’ 571

Besides the fact that the first occurrence is from a narrative and the second constitutes free indirect discourse, the same content is presentable in both ways: (380) 'q&e1le $we 'nublle Al mzab!ne 'uxxa la qam zwAnnu $we dde,

md zvArre zvArre hl /$Arta la slu mzab!ne $we dde ‘He cut woods and brought them in order to sell them, but no one bought his woods, as much as he walked around, he walked till the evening, (lit. and) his woods were not sold’ 664

(381) mt 1uzlu 'muslu kullu akme ukullu jam/At ak!ma uxa

láswabe mqímwle ! x'we mAn qzla ‘Whatever they did and brought all (of the) kings and the kings’ people, no one could lift the snake off her neck’ 33

This u- is difficult to judge, reminding one of the particle fa- separating the protasis from the apodosis in Arabic. The concessive adverbial is also attested preceding both conditional and concessive-conditional patterns: (382) baxta kmad hye g!ra sqla xazya xa xt ápwa gAble ‘A

woman, as much as her husband may be beautiful, (if) she sees another one, she wants him too’ 807

(383) kmad h!ya sqAlta 'kan h!ya ruyi ši kAsla pqa&Anna ‘However

beautiful she may be, (even) if she has my very soul, I will kill her’ 851

This type of concessive clauses is by far more common than the specialized concessive exponent š!1 ga d- ‘although’ (lit. ‘seventy times that’, Sabar 2002: 118a, 295b): (384) š!1gAt la sle mA&ra rba bb'ne pqAxlu ‘Although not much rain

came (down), the flowers have blossomed’ Yona 1999:417a

(385) šo1ga dg!zanwa 1axxa š'la, la wle mux daw š'l /irq ‘Although

I used to work here, it was not like the work in Iraq’ SAG 2.1



Another way to mark a concessive clause is by focus particles, whose scope in this case is the entire clause: (386) mdam {qam msAnwal!xun kAsli} ši dAd xyetwa (sic plural),

uqam qa&lAnna xsi! ‘Since even (though) {I had brought you to me} so that you may live, nevertheless I killed my sister!’ LH 1.13

The particle ši is basically a focus particle (see § As such it is used to denote, like ‘even’, factual concessivity. Non-factual concessives (concessive-conditionals) are introduced by Atti and afíllu, both of which are focus particles when they refer to a single element in the clause. When referring to an entire clause, they function as concessive conjunctions: (387) 7tti {xuwwe hwe xz1a 1Al g!zakAt 1aqle} la q&e1le $l!se ‘Even

(if) {a snake should be wound around his ankle}, he should not stop his prayer’ Yona 1999:145b

(388) 1af7llu {br!ni myAs 1 naqla}, b!zAnn!x'n 1fu ‘Even if {my son

dies this time}, I will pardon you’ MA 3.8

Note that in the preceding examples the curly brackets mark the scope of the focus particle. However, unlike ši, which is still a particle, Atti and afAllu are most probably nuclei. Concessive-conditionals are discussed further below, with conditional clauses (§ The concessive-conditional pattern, as well as conditional structures in general, are judged in this framework to be a macrosyntactic issue, which does not follow the basic syntactic relationships, and more specifically, the completive relationship. For this reason they are treated below, under macrosyntax (§ Other adverbial types Presupposed degree or amount This type is not very common, but it is essential to complete the picture of free choice quantification already discussed above (§§–5). It is expressed in various ways, first establishing some given amount or degree and then referring back to it: (389) mnA 1!ha nša bbe qalz wle, 1wa xa utre1 pAšle mard bAž mAn

bbe ‘Look, this man, his father was cheap, (but) he came out (lit. became) twice as generous as (lit. than) his father’ AS 568



(390) kma kxarjAt 'gz!nAt 1awye ta šapsa, &ha ga Al 1!qa máAzrAt ta

šapsAt 1asya ‘As much as you sell and buy provisions for Sabbath, you will prepare three times as much for the following Sabbath’ 615

(391) kma 1le maksab d!x'n tre sa/e? na bywínn!xun tr-ga l1!qat

xt ‘How much is your fee (for) two hours? I will give you twice as much’ 555

The presupposed amount or degree is introduced in different ways—in ex. (389) it is a declarative clause (‘his father was cheap’) while in exx. (390) and (391) we find kma ‘how much/as much as’ in an indefinite expression and in a question respectively. In the structure discussed here, the indefinite quantifiers are used to introduce the amount or degree, making it presupposed for what follows: (392) 'kmad wla /qAl, xa utre wla spahn ‘As smart as she was, she

was twice as beautiful’ 781

(393) kud žal dd tam md gnbAlwa, qaramn gnblwa x tre bžrab

mAnne ‘Any worker at that place, whatever he would bring, Qaramn would bring twice as much’ 129

(394) kmad wle nax!pa u1ne kApta 'mare nm's hdax pAšle mAštaq

dda ‘However shy, humble and concientious he was, that much did he desire her’ 784

The presupposed amount is referred to in all cases by multiplying it (twofold in exx. (389), (391), (392) and (393); threefold in ex. (390)) or by insisting on the presupposed amount or degree (exx. (394) and (395)). Note that degree and amount are expressed either by x-ga ‘x times’ or by xa utre ‘twice as much’. Similar notion is also conveyable by qn ‘as, like’: (395) qn wle bbe qus'r hdax le brne mf arf tna ‘As (much as)

his father was cheap, so23 is his son wasteful‘ AS 568

The syntax gives us the adverb hdax ‘so’ marked as the rheme (immediately preceding the copula le), namely ‘it is to this degree (that) his son is wateful’, or, in better English, ‘his son is SO wasteful’. 23


209 Spatial adverbials This group includes specialized spatial adverbials such as tma ‘there’, axxa ‘here’,  ‘down’, Al1l ‘up’, as well as several nucleic adverbials such as xazAr ‘around’, qam, barq'l ‘in front’, basAr ‘behind’, xe ‘under’, rAš ‘on’, go ‘in’, etc. The following table shows their placement with the same lexeme (ytwa ‘sit’) and in the same syntactic function: no.


(396) t'le (397) t'lu (398) t'le (399) t'lu (400) t'lu

spatial syntagm qam dAkknAt daw tAjjar

Translation in front of the store he sat of the merchant

rAš ršet ax!na z!ra

he sat

Al bsa

they sat


they sat

xe dan dre

they sat

barq'l mamo

she sat

(401) t'le


(403) t'lu (404) t'lu

xazAr xáwxét

go d d'ka

they sat

(405) t'le

lxá-1l rAš xa qa&ra

he sat

(402) t'la

he sat

they sat



at home

near the head of his younger brother

41 43



at her place (chez elle)


under the trees


in front of Mamo


at that place


near each other aside on a rock



This group can, as any adverbial modification, occur with other adverbial groups. A case such as we find in ex. (405), where there are in fact two locative modifications, is perhaps to be explained by a complex spatial syntagm, in which one part specifies the other. The substantive d'ka ‘place’ can have an adverbial function: (406) bAd wa/dAd gnpAqwa mAn bsa, zAlwa xa d'ka ‘At the time he

was leaving the house to go (to) some place...’ 24

(407) xá y!ma zAlle Al xa d'ka ‘One day he went to some place...’ 360

The fact that xa d'ka (ex. (406)) and Al xa d'ka (ex. (407)) occupy the same slot means that they have the same function as well. We have seen a few such cases in the object slot (§3.1.2), where the conclusion was that the slot was essentially a substantival slot. Here the opposite conclusion is drawn, that in this slot even what looks like a substantive actually functions adverbially. Another way to evaluate d'ka is to see which entities can commute with it:



(408) ! maf'ra d'kAt 1jAblox mnbAllox kan amlet rše ‘This rug will

bring you where(ver) it pleases you if you stand on it’ 426

(409) rk'le ls'se dde 'húllla das&'r za kAt 1jAbla ‘He rode his mare

and gave it permission to go wherever it pleases it’ 313–314

In the preceding example we see that d'kAt 1jAblox ‘wherever it please you’ (ex. (408)) and kAt 1jAbla ‘wherever it pleases it’ (ex. (409)) occur in the same slot. The first is the nucleus form of d'ka ‘place’, whereas the second is the nucleus form of the pronoun ‘where’. The difference is found in the degree of definiteness—the second expression is indefinite by definition. The following pair confronts d'kAt and kúd'kAt ‘any placeNUC’: (410) šqlle lx$e 'nublle d'kAt zAllu ‘He took her on his back and

brought her where (lit. place) they went’ 272

(411) ta hatxa k&awla drAnna 1Al x$i 'nablAnna kúd'kAt 1zna ‘For

this (reason) she deserves (that) I put her on my back and bring her anywhere I go’ 266

The difference, again, is only in the degree of indefiniteness, marked in the case of kúd'kAt. Now, if we confront exx. (410) above and (412) below, we see a similar difference which, however, is marked differently: (412) kan sle wa/di mmsAn d'kAt hwAn mla ‘If my time comes, I

will die wherever I am standing’ 766

The difference between exx. (410) and (412) is signalled by the local opposition between the indicative form zAllu and the subjunctive perfect form hwAn mla. Without A specialized exponent for this notion does not occur in JZ. There is a borrowed item, bala (qqāyem>qāyem), which is discussed presently.

5.6.2 kšāqıl in narrative This form, used in dialogue as an aoristic present, is found in narrative as well although, unlike the dialogue, it is not the unmarked form. It is found in a couple of functions in narrative: one is the case of kšāqıl form seemingly standing for šqılle, the other is when kšāqıl forms seem to occur instead of expected kšāqılwa forms. kšāqıl for šqılle The most important type of kšāqıl forms, that occur where šqılle is expected, is illustrated in the following examples: (312) ṣrıxlu ḥakīmet kulla ʾáy bāžır ū lá ġzēlu čū́ darmāna ta naṣaxūsıt

ḥakōma. mjōrıble kullu darmāne ū lá ṭ rıṣle. mın yōma lyōma naṣaxūse bıž gyaqra. ṣrıxle ūmōṣēle ʾāse kullu xaračkarat dīde ‘The doctors of the entire city were called but they did not find any medicine for the king’s illness. He tried all the medicines but did not get better. From day to day his illness gets worse. He called out and ordered all his magicians to come’ 590

The facts pertaining to the sickness of the king are narrated by the preterites (underlined). However, the detail about the sickness getting worse, instead of occurring as a preterite (cf. yıqırre 570), appears with gyaqra. There is no difference in temporal deixis, since everything that happens in the example relates to the same temporal line. In fact, gyaqra is expressed from the sick king’s perspective, reflecting his own ‘now’. (313) ha ġzāun +kōpāla dīdi atta napṣınne +kōpāla dīdi ūġzāun kmā ́ pāre

bıkōši mınne. kšāqılle +kōpāla dīde ūgnāpıṣle bıd ʾarʾa ūlá kēse afıllu ḥıs pāre. ṣrıxle ūmırre ‘Here, look at my walking stick, now


CHAPTER FIVE I shall shake off my walking stick and see how much money will fall from it. He takes his walking stick and shakes it off on the ground but not even a sound of money comes out. He screamed…’ 338–339

The blind man brags about his ability to hide money. What he does to show his friends, from his own point of view, is expressed by kšāqıl forms. Note that his screaming, which follows, is conveyed via a preterite, that is, objectively. (314) anya ṭḷāha ʾınwe māṭo māle ta kulle dīwān qēmın ʾaxlınnu. kēxıl

xā,́ gnapqi trḗ xēt šwīna. kēxıl xa-xēt, gnapqi ṭḷāha xēt šwīna ‘How can these three grapes be enough for the entire council? Let me eat them. He eats one, two other come out instead of it. He eats another one, three other come out instead of it’ 504– 505

A plowman is given a task by the king, involving a bunch of grapes that should feed the entire king’s council. He is given three grapes. What happens is narrated from his own point of view. The following example is selfexplanatory: (315) ūxrūle ūmpılle ıl dūkıt gıyāne. ṓha ḥakōma gıṭāʾe ıl brōne ūlag

xāzēle. pıšla palgıdyōm brōne la sēle ‘he (=the prince) fainted and fell there. The king looks for his son but does not find him. It became noon, his son did not come back’ 499–500

The king’s now is a token of his own angle. In the next example, a midwife is taken by strange people to attend a delivery: (316) qam šaqlīla ūqam nablīla. gō ʾurxa gnablīla hayya-hayya ūgzamri

ūkṣarxi qāma qāma ‘They took her and carried her. On the way, they carry her quickly, sing and call out louder and louder’ 298

Note that the fact that they carry her is given twice—once as šqılle and then as kšāqıl forms, which reflect the midwife’s own experience. Some verbal lexemes, such as √ ʾmr, √yzʾ, √ʾsy, √qym and √xzy, show free variation between the formations šqılle and kšāqıl. So gēmır ‘he says’ for mırre ‘he said’ is normal (it is actually usual in many languages, for instance Classical Syriac and Old-Babylonian Akkadian). In some of these cases, namely, where the lexeme is but a framework for the content, serving as an exponent, the reason is clear—it does not represent an event. However, this



phenomenon with verbal lexemes such as √ʾsy, when they do not have any rhematic complement, still remains to be explained. Consider the following pair of examples: (317) aw lēle nṭırre axōna rūwa hīl palgızlal ūṭweʾle. kēse bābe mbínnōke

ūkxāze brōne wēle ṭwīʾa ‘That night the big brother guarded till midnight and fell asleep. His father comes in the morning and sees: his son has fallen asleep’ 185–6

This situation repeats with another son: (318) aw lēle pıšle bınṭāra axōna palgāya hīl palgızlal, ū ṭweʾle. pıšla

qarwāwıt bhāra ūsēle xa ṭēra ūqṭeʾle xa xabūša ūzılle. pıšla mbínnōke, sēle ḥakōma ūġzēle ınnu brōne wēle ṭ wīʾa ‘That night the middle brother was guarding till midnight, but he fell asleep. It was almost dawn and a bird came and plucked an apple and went away. It became morning, the king came and saw that his son was asleep’ 186–187

The cases seem to be in variation. kšāqıl for kšāqılwa The form kšāqıl occasionally occurs where descriptive kšāqılwa is usually found. Under § recurrent or distributive incidents are mentioned as occurring with kšāqılwa forms: (319) zılle ṓha xūwe ıl day ʾurxa, kud ganāwa ūkud mšalxāna qātílwalu.

hīl zılla xā šabsa, kullu mšalxāne uganāwe qam qāṭıllu ūʾurxa psıxla ‘The snake went to that road (and) was killing every thief and every robber. By the time a week passed, he killed all the robbers and the thieves and the road opened’ 31

In this example, the recurrent distributive action of killing each and every robber is recounted by kšāqılwa (the form qātílwalu is probably a simplification of kqātílwalu. The subjunctive in asyndetic purpose clauses tends to occur first in the clause). Then, when a conclusive overview is given, kud ‘each’ is replaced by kullu ‘all’, and the verbal form is preterite qam qāṭıllu. The following examples have to do with hunting as well: (320) kud

gʾāwırwa mın qāme, ham lák šāwıqwa fālıtwa ṣāx ‘Nevertheless, he would not let anyone who was passing near him get away alive’ 959



(321) sēpe š līpa bīze ū ku(t) gnāpıqwa go ʾurxe láswāle flāta qqāṭílwāle

xa qṭāla kıret ‘His sword drawn in his hand, anyone who was coming out in his way had no escape, he would kill him in an ugly fashion’ 964

However, in other cases one finds kšāqıl forms instead, in the same syntactic and semantic environments: ́ gyāʾel ūkṭāḷıblu ʾanya ṭḷāhā ́ šoʾāle, gmádeʾri ʾılle ʾán (322) gō kút dıkkān

mare dıkkāne bıt jıgra: yassaq (q)īlu ʾanya šoʾāle ta yıhāwa ‘In each store he enters and asks for these three items, the store owners answer him angrily: Giving these things is forbidden!’ 697

The difference seems to be the same as reported above, where kšāqıl forms occur where šqılle forms are expected, namely, the perspective of the referent, rather than an objective or narratorial one. In this case it is a man who is desperately looking for three items his angry wife asked him to bring along. This is his hic et nunc. The same applies in the following example where a young lady’s courting manners are described: (323) uʾáy xamsa rāba qam ʾıbāle ūkut jwanqıt kēse mın bāte xšīwe ūsqīla

ıl ṭ aḷabāye dīda lágba čuxxa ūlákxālḗla čuxxa mın basır bır ʾamōya. kullu gımšadrālu bıd rēšu kīpa ūnx īpe ıl bēs gyānu ‘and the maiden loved him very much, but (of) any fine youngster who comes from important families to ask for her hand, she does not want anyone and does not like anyone except her cousin. She dismisses them all with their heads down and ashamed to their homes’ 684

This is just as recurrent and distributive as the actions found in exx. (319)– (321) above. The difference again is in a specific point of view: (324) zılle musēle xā barmīlıt qīra ūkud naqla gıṭāmıšlu kundarre go day

qīra ūgēzıl ıbbu rıš dan pāre. ī ze lágmande ıl pāre dīd lá raʾši nāṭōre. hama gēzılwa rıš dan pāre gıṭāpēwa bıd kundarre ūgēzılwa ıl bēsa ‘He went (and) brought a barrel of tar and each time he dips his shoes in the tar and walks with them on the money. He does not extend his hands to the money so that the guards would not notice. He would just walk on that money, it would stick to his shoes and he would go home’ 169–170



The boy in the example is picking up money from the floor without touching it with his hands, so as not to frame himself. The recurrent action is first described from his own point of view by kšāqıl forms. Then an explanation follows, using ‘normal’ descriptive kšāqılwa forms. Generic statements are yet another slot where kšāqıl forms are found in narrative. Such forms report a fact that is valid anytime (at least with regard to the story): (325) ʿādıt day ʾēna wḗwāla kud gšāte mınna qamāya gımāyes. nāšıt day

bāžer, ay dammıt kōšīwa ıl day ʾēna šātēwa māya, gınablīwa ımmıt gyānu xa qaṭūsa, gmaštḗwāla qamēsa: āya gmēsāwa ūxarāye kullu gıšātēwa ū gēzīwa ‘The custom of this spring was (that) whoever drinks from it first, dies. The people of that city, when they would come down to the spring to drink water, they would bring a cat with them, let it drink first; it would die and then everyone would drink and go’ 144 generic

This example serves as a key: both kšāqıl and kšāqılwa forms are found in it. The former provide an explanation as to the nature of that fountain, whereas the latter describe the local (but habitual) solution. In the following examples it seems that kšāqıl provides a general explanation while kšāqılwa a local one: (326) ʾarya gımšāpe lbani-ʾādam. kfāhım kúd mındi mın hawwa hīl

xrīwa ‘The lion resembles human beings. He understands everything, good and bad’ 910

(327) batāne ūqaṣrāre dīda ġēr šıkıl wēlu; gımšāpēwa ıl kusyāś ḥāji lōkat

‘its houses and castles were (of) a different kind; they were similar to mushrooms’ 421

In the following pair, olfactory abilities are described: (328) uhayya-hayya šqılle mahīne dīde ūrkūle ılla ūzılle. ūmahīnat kšaqli

rīxıt urxa xā dīd daw xēt ‘So he quickly took his mare and mounted it and left. Now mares pick up each other’s smell on the way’ 84

(329) lák xāzēwa, bale gmāyıxwa ūlmūjıb myāxa gmandēwa gāve ‘He

could not see, but he would sniff and take a step according to the smell’ 394



The first case is general, while the second is local (the giant had just been blinded). General cases may be found in a saying: (330) kxāze wēle štīḥa ʾáw ʾarya tāma. rūwāne–lu mīre: ʾarya lág +dāmıx

lá-hōya rıš xızēna ‘He sees: the lion is lying there. It is great men (who have) said:30 A lion does not sleep, unless it is atop a treasure’ 918

The last example in this group is very useful and constitutes a key example for the prolem treated in this section: (331) ʾō yāla rwēle; kudyōm, gēzıl gzāvır, gēzıl go kōlāne; qāṭıl ʾō yāla

dīd dō nāša uqqāṭıl ʾwā ́ yāla, uqāṭıl yalunkıd maḥalle. kudyōm kēse gımṭárṭımi ıl bāba-yımme, nāšıd, nāšıd maḥalle, nāšıd bāžer. qam qāṭıllu yalunke dōhun. ma ʾōzi? kšatqi, lēbu ʾ ılle ‘the boy grew up; everyday he goes, wanders about in the streets; he beats this man’s boy, and beats that boy, and beats the boys of the neighborhood. Everyday the neighborhood’s people, the town’s people, come complaining to his parents. He beat their children. What are they supposed to do? They keep silent, they cannot cope with him’ MA 2.17–19

The preterites are underlined, but do not quite represent any events: the first, rwēle ‘he grew up’ is descriptive and the other, qam qāṭıllu ‘he beat them’ is part of the complaint. The rest of the forms (in boldtype) are given away by kudyom as kšāqıl forms which relate to backgrounding kšāqılwa forms, while reflecting the point of view of the parents. Note that there is a major difference between the cases presented so far and others where kšāqıl forms may be analyzed as circumstantial: (332) pıšla bīsāqa ūbıkwāša ṭūrāne ūgıre uʾēna lá gıʿalqa ū lák tafqa bıt

čū́ ʾazāla ūʾasāya ‘she kept going up and down mountains and hillocks (while) not meeting any passer(s)-by’ 855

The circumstantial kšāqıl is basically a subordinate clause and has nothing to do with point of view.


This structure is a cleft sentence, see §



5.6.3 pšāqıl and pšāqılwa in narrative pšāqılwa functions in dialogue mainly as a counter-factual apodosis. In narrative there are very few of these, and they function, marginally, in two slots: one is the circumstantial slot, where a finite verbal form interchanges with the gerund or the perfect participle. The second is where it occurs instead of kšāqılwa. In the circumstantial slot the forms are subordinate, and already discussed above (§3.2.1, and see further below). However, the relationships between pšāqıl and pšāqılwa still need to be examined. pšāqıl is opposed to both kšāqıl and šāqıl in this slot: (333) muzvırru lgyānu dʾāse (//bāse) lbēsa ūṭrēlu blazzi ‘They turned

around [intending] to go home and drove hastily’ 889

The double slash in the translation indicates another, usually close possibility inside the original text. While the subjunctive (also when asyndetic) signals a purpose clause, the form pšāqıl denotes intention; opposed to kšāqıl, which in these circumstances is co-temporaneous, pšāqıl shows relative posteriority: (334) qam dārēla ʾáy dıste ʾápāý a rıž dán qapye ūṭrēle xmāra bāse lbēsa

‘He put the pot as well on top of the logs of wood and led the donkey, intending to go home’ 919

pšāqıl occurs, with similar function, with the expression bēd lıbbu: (335) aw yōma sēlu kud arbi ganāwe bēd lıbbu bāse šāte kıs rūwa dōhun

‘That day, all of the forty thieves came, intending to come to drink with their leader’ 177

The following example is somewhat more difficult: (336) gmẽxa má gmẽxa: ḗ brāta wēla tūta gō xāne, šıppōlāla murme hīl

nafṭanga ūhā ́ xá-bıhna yān trḗʾ bxaṭya ʾımme ‘She sees, what does she see? The girl is sitting in his lap, her hems (of the skirt) lifted up to her waist, just about to sin with him in a moment or two’ 811

It is hard to determine whether the circumstancial expression ūhā ́ xá-bıhna yān trḗʾ bxaṭya ʾımme (lit. ‘and there, in a minute or two, she will sin with him’) is indeed subordinate. It could rather be compared with ex. (332), both cases showing a full clause introduced by the connective u. In that case



these syntagms can be compared with non-subordinate circumstantial presentative clauses (§5.4.2 above). pšāqılwa occurs in circumstantial clauses as well. The line between it and pšāqıl in narrative is slight: ́ ıqwa, mpıqlu qāme arbi bınāsa ‘(while) he was going (337) sēle bınāp out, forty girls came out in front of him’ 109

(338) ay dammıt mpılle ṓ ha kapōra mōrumle bır ḥakō ma sēpa dīde

bqāṭílwāle. kšıfle gyāne ṓ ha kapōra ū mpıqla ay brāta dīd qam xazēla ṣurta qam dargıt āxō ra ‘When the giant fell down, the prince raised his sword, intending to kill him. (However) the giant revealed himself, and the girl, whose figure he saw in front of the stable’s gate, came out’ 89

(339) rkūle gō pāpor-baḥar uʾṓ pāpor-baḥar zılle, zılle, bāzılwa ıl flāna

bāžer ūʾurxa zaʿla ʾıl dan nāše nablānıt pāpor-baḥar ūthēlu bıd baḥar ‘He embarked a ship and the ship went (and) went, it was going to a certain city but the people who led the ship lost the(ir) way, and they wandered in sea’ 348–349

In all three cases (and in a few other), pšāqılwa forms are circumstancial and prospective, like pšāqıl. However, in all these cases, the intention expressed immediately turns out to be a mistake—in ex. (337) his going out is obstructed; in ex. (338) the killing does not take place; and in ex. (339) the ship does not get to its original destination. The implications are that there is a modal difference between the two forms. The following examples are of independent clauses in narrative: (340) hādax wēle ḥāl dīde ṓ gōra kúdyṓm bāsēwa ūbzāmırwa hīl ʿāṣırta

ūmmāpıqwa ō xūwe xa dıhwa bıt pumme ūbyāwílwāle ṭāle ḥaqqıt zmāra dīde ‘That was the man’s condition: Every day he would come and play until the evening and the snake would come out, a gold coin in his mouth, and would give it to him, as payment (for) his playing’ 445

(341) kudyom gēzīwa ḗ naqla kútrū-́ kutru ldáy dūka, bāba gēmırwa bıd

bıllūre ūbrōne gyātūwa xazre hīl ʿāṣırta uʾṓ xūwe gmattūwa qāmu trē dıhwe ūkxāyıšwa ldūke ūʾanya nāše kēsēwa lbēsa, lıbbu bassīma ū rıẓye (mın / p) ḥaqqu ) ‘Everyday they would go, now both of them, to that place, the father would play his flute and his son would sit next to him until the evening, then the snake would put in front of them two gold coins and creep to his



place, and the people would come home, with good heart and satisfied’ 448–449 Exx. (340) and (341) describe roughly the same things, with a few slight differences. Still, the forms in ex. (340) constituting this description are, extraordinarily, pšāqılwa forms instead of the more common kšāqılwa forms. In the next example we find both: (342) baryāwa uxā ́ mın dan gundıkat rūwa dīda lā ́ yāwılwa xarj ta

ḥakōma bnāpıqwa ʾṓ gabāra ʾımmıd ʿaskar dīde dgmānēwa ʾımmá gūre ūkutxa mınnu mare qūwıta wēle tēr gyāne. gēzīwa lday gundıke ‘(When) it would happen that some chief of one of the big villages was not paying taxes to the king, the mighty (man) would go out with his army, which numbered a hundred men, and each of them was strong by his own right. They would go to that village’ 945

The paucity of these examples leaves this case a bit mysterious. However, there are similar cases: (343) kudyom ʾāna uyımmi, uʾaxōni frāyim, ʾilāha mānıxle, fēḥıl ḥalāl

hāwe, mnablax ʿabāya, bāzax. mād-īlu ʾ axōni uyımmi bdōqi ʿabāya rrēšıt, qam pımmıd jalal, uʾāna bāzēna, mkaškıšınnu nunyāsa, gır, gır, gır, gmēsaxlu, gčēkaxlu go ʿabāya. +gmarmaxla ʿabāya, +ʾısra, xamšaʾsar nunyāsa kāwe gāwa. kudyom gdōqaxlan xa karta nunyāsa ‘Everyday I, my mother, and my brother Ephraim, may he rest in peace and be forgiven, would take an abaya (large robe) and go (to the brook). My brother and my mother would hold the abaya at the beginning, at the mouth, of the brook, and I would stir the fish (in their direction). (Thus), we would bring them and push them into the abaya. (Then) we would lift the abaya (from the water and see), ten, fifteen, fish would be in it. Everyday we would catch a (whole) load of fish’ SYG 1.2–3

This description comes from personal experience narrative. In this text specifically, kšāqılwa forms for the description of habitual actions do not abound, but are rather sporadic. The passage in ex. (343) has, instead, pšāqıl and kšāqıl forms having the same—imperfective—function as kšāqılwa forms elsewhere. This use of pšāqıl forms seems to be related to the imperfective function of pšāqılwa forms.



5.6.4 Tense in narrative: subordinate clauses True tense oppositions exist in JZ narrative only in subordinate clauses—in adjective, substantive and adverbial clauses. What is found in these clauses is not the system of abslolute tense found in dialogue, but rather a tense system relative to the matrix clause, or in other words, to its reference point, which is the ‘story-line now’. The paradigm in each type of subordinate clause, sometimes in each type of conjunction, is different, but the general picture arising is that except for clauses that originally express modality (i.e., modal content, or purpose clauses), subordinate clauses basically exhibit a relative tense system, even in narrative. Adjective clauses Clauses introduced by d(īd) exhibit the following paradigm: šāqıl (subjunctive) occurs as the unmarked form mainly in cases of an indefinite referent, it is therefore a-temporal: (344) xa gōra dhāwe ġṛība mım baxte ū xāzēla ... ‘a man who is far

away from his wife and sees her...’ 784

(345) xā ́ baxta gurta ūdhōya rḥıqta mın gōra ‘a married woman who is

far away from her husband’ 787

(346) xā dʾīlā ́ šāqılla mınne, mır rēše ‘someone who lost his mind (lit.

whom god took it from his head)’ 887

(347) xā ́ dšāwıq māĺ dīde ūʾāzıl ‘someone who leaves his property and

goes away’ 812

Note that this subjunctive form may reflect the entire temporal spectrum (e.g., past in ex. (346), present in exx. (344)–(345) and (347)), and that it does not have a modal meaning in this environment. It is a conditioned form, occurring inside indefinite adjective clauses, of the more common wēle: (348) ʾáw xá dʿāqıle wēle brēše ‘one whose wits were in his head’ 888 (349) uġzēlu xā gōra dīd aqle wēlu go xā qōqıt ıpra ‘They saw a man

whose feet were in a clay pot’ 52



wēle 31 denotes co-temporality with the line of events, ‘story now’, most commonly with a characteristic feature. The same obtains with complex verbal forms such as dīd wēlu tīwe gāwe ‘in which they were staying’ (264). So does, in fact, the ‘present’ copula, denoting the same ‘story-line now’: (350) ṓha gōra faqir mbōqırre mın dṓha gōra dīd dīle go dan +māya ‘the

poor man asked the man who was (lit. is) in the river’ 531

(351) mṭēle qarwāwıt day bāžer dīd dīla xāse skınta tāma ‘He arrived

near the town where his sister was (lit. is) living’ 202

The form kšāqıl denotes a fixed characteristic here as well: (352) zılle ō faqir yıʾılle go day lokanda dīd ḥakōma kēxıl gāwa ‘The

poor man went (and) entered that restaurant in which the king eats’ 18

(353) ūġzēla xa makīna dīd ktaxni ıbba nāše ‘and she saw a machine

with which one grinds people’ 174

The king eats there regularly (ex. (352)), and the machine is habitually used to grind people (ex. (353)). There is only one example of wal kšāqıl, which serves to denote an actual, precise, co-temporal: (354) mṭēle ıl day bāžer dīd axōne wal gıḥākım rēša ‘He arrived at the

city which his brother (now) rules’ 149

That his brother rules the city is quite recent. kšāqılwa is very similar to kšāqıl, and the temporal difference does not stand out: ́ ımwa Aharón Rašīd́ xā sōtınta ‘There (355) ıswa go day māsa dīd gıḥāk was an old woman in the town which Aharon Rashid used to rule’ 8

(356) mdōyınne trē tawıryāsa mın xā gōra dīd kıʾḗwale ‘He borrowed

two cows from a man whom he knew’ 62

The preterite forms šqılle and qam šāqıl-le are different: (357) sēlu ham ʾan malaxīne dīd qam mēsēla Zīne kıs Mamo, qam šaqlīla

ūfırru ʾımma ıl bāžer dīda ‘the angels who (had) brought Zīne to Mamo came again, took her and flew with her to her city’ 307

It is the preterite of √hwy. The rheme follows rather than precedes because it is in subordination. This is actually the allotagm of x-wēle used in subordination. 31



They may reflect anteriority vis-à-vis the narrative line, very much like the pluperfect forms šqılwāle and šqīla wēle: (358) ūʾáw gōra mıṭēle go xá barīya dīd tāma axōne míswāle ‘The man

reached a wilderness, where his brother had died’ 276

(359) ūqam šāqılla ṓha ʾaxōna dīd mīsa wēle ēha jērīye ta baxta ṭāle

‘The brother who had died took this maid as his wife’ 280

Note that dīd mīsa wēle does not mean ‘who was dead’ (by this time, he was not dead anymore). To sum up, adjective clauses have a binary opposition expressing cotemporaneity vs. anteriority. Substantive clauses Substantive clauses occur with ınnu, dīd, as well as asyndetically. ınnu is the most difficult, because its nature is not quite obvious. However, the distribution of forms following it is similar to that of dīd. (360) tūle uxšūle ʾınnu mād ʾurra ʾılle kulla bıd dáw lēle xılma wēle ‘He

sat and thought that whatever happened to him that night was a dream’ 568–569

(361) axōne xšūle ınnu ganāwa–le ‘His brother thought that it is/was a

thief’ 238

(362) uʾṓha mare-bēsa xšūle go lıbbe ınnu bıyāwíllēle ṓha fīstāna ūmēse

ṭāle trē fīstāne ‘and the landlord thought to himself that he would give him this gown and he would bring him two gowns’ 215

There is a functional convergence between ‘past’ and ‘present’ expressions of being, both referring to ‘storyline now’ like adjective clauses. Here, unlike the latter, one finds pšāqıl forms which function as future relative to the storyline. dīd is similar; note that šqılle is attested with ambiguous value— either co-temporaneity or anteriority to ‘storyline now’: (363) xšūle ʾáw najāra dīd sēlu ʾōdıt ḥakōma ta nabōle dīde ıl qıṭḷa ‘The

carpenter thought that the king’s servants came/had come to take him to die’ 717

Here too one finds pšāqıl forms as future relative to the storyline:



(364) lá xšūle dīd ʾṓ zērıngır byāwılle ʾasıqsıt ḥakōma ‘By no means did

he think that this goldsmith would give him the king’s ring’ 730

Asyndetic substantive clauses (which are not deemed to be zero, see §2.3.1 above) show the richest variety of forms: (365) axōne zeʾle qam mazıʿla ‘His brother knew he (had) lost it’ 245 (366) xšūla xılma wēle ‘She tought it was a dream’ 307 (367) uxšūle xılma–le ‘so he tought it was (lit. is) a dream’ 309 (368) uxšūle wal gāxık ʾıbbe ʾṓ zērıngır ‘he thought (that) the goldsmith

was making fun of him’ 670

(369) bale zeʾle ʾōha zērıngır gmaḥke mın kulle lıbbe ‘but he knew (that)

the goldsmith speaks the truth’ 670

Exx. (368) and (369) have kšāqıl forms, used in dialogue as an aoristic present. However, in ex. (368), as the use of wal hints, this is not a co-temporaneous, general state of affairs (as depicted, e.g., by wēle), but rather an actual, precise co-temporal (compare ex. (354), where the brother rules the city right that moment, that is, story-line now). (370) xšūlu atta ptafqi ʾıbbu ‘they thought (that) now one would

notice them’ 269

(371) xšūle barbar ʿāsırta mnāpıq ʾaw xūwe ‘he thought (that) towards

evening the snake would come out’ 456

These clauses seem to present us with a richer temporal spectrum: there are pšāqıl forms in the paradigm that allows three relative tenses. One is anterior to the storyline (šqılle, qam šāqılle), another is co-temporaneous with it (kšāqıl, šqılle, wēle) and the third refers to a point posterior to the storyline (pšāqıl). Adverbial clauses The adverbial clause type chosen here to demonstrate a temporal system is causal clauses, headed by čıkun ‘since, because’. The preterite (šqılle, qam šāqılle) represents events that chronologically precede the matrix clause, i.e., before the narrative ‘now’:



(372) rāba čhēle čukun zılle ʾurxa rāba raḥuqta ūqṭeʾle ṣīwe ‘He got very

tired because he (had) walked a very long way and (had) cut trees’ 626

(373) qḥırre ūbxēle (čukun) qam qāṭılla ē sōtınta ‘He grieved and wept

(because) he (had) killed this old woman’ 264

(374) pıšle nāṭōrıt tarʾıt ḥakō ma, čukun zeʾle ınnu anya arbi gannāwe

bāse kıs ḥakōma ‘he became a guard of the king’s gate, since he knew that these forty thieves would come to the king’ 176–177

The difference between zeʾle and kīʾēwa may be temporal here—the former reflecting some kind of anteriority (e.g., ‘he had found out/come to know’). (375) rāba gjāgırwa čukun kīʾēwa kmá quṭeʾfe–lu bıd dáy dalīsa ‘he

would be very angry, because he would know how many bunches there are in the vine’ 642

(376) kēfe

rāba sēle … čukun kīʾe nāše rūwe lıbbu kāwe rwīxa ‘He was very happy … since he knew important people were generous’ 482 +

On the other hand, no difference is detectable in comparing kīʾēwa and kīʾe. The last example shows the difference between a co-temporal state of affairs (‘is the last one’) and a posterior one (prospective ‘would not give’): (377) uxšūle mın basır mınne lákāyēbu xāye rıž dunye, čukun ṓ yāla

xarāya–le uʾīlā ́ lágyāwıl ġēre ‘he thought (that) after him (gone) they would not be able to live in the world, because this child is/was the last one and God would not give another one for him’ 587–588

5.7 CONCLUDING REMARKS This chapter attempted to cover the function of forms inside the narrative. The organization proceeded from meaning to form, and the arrangement, accordingly, was by function and then by specific exponent. The issue of FSP effects (§5.1) is discussed apart from its consideration in the dialogue because the functions and the effect topic and especially focus have in narrative are quite different. The phenomenon of narrative point of view (§5.2) is given due consideration, since many forms, including verbal forms and presentative expressions, take part in conveying it. The sections on cohesion (§5.3) covered boundary markers and various connective devices. The issue



of grounding (§5.4), a central theme in any narrative, accounted for many matters—the plupreterites, presentative constructions, the setting paradigm, descriptions, passive forms—all these and more take part in expressing the background. The chapter ends with a focussed view on the verbal system— the aspectual and temporal oppositions found within the narrative.

6 DIALOGUE SYNTAX The aim of this chapter is to describe in detail the DIALOGUE texteme in order to arrive at a precise syntactic definition of this macrosyntactic unit. Like the clause, it can have various sizes, but the syntactic features which characterize it are consistent. A priori, every segment of direct speech is dialogic in nature, coming from a character in the tale. However, difficulties immediately arise; for example, a story told in the 1st person (most notably in text xxxv, 372–431, but also in two other texts, SYG and SAG), which doubtlessly consist of a segment of direct speech, may nevertheless contain narrative parts as well. The safest way to proceed then is to start with text parts which are unquestionably dialogic, to differentiate them syntactically, and only then examine, using this syntactic characterization, the questionable parts. DIALOGUE is characterized by several distinctive features that immediately stand out when compared with narrative: unmediated deictic categories, all directly referring first and foremost to the here and now of the (al)locutor(s), whereas the narrative refers first and foremost to narrative now, but this reference is relative rather than absolute. Many categories are influenced by this reference: • Person occurs to its full capacity, namely, all three persons, most prominently the allocutive, the second person (and functions associated with it, such as vocative 1 and imperative). Narrative is confined to the delocutive (third person), occasionally occurring in the locutive (first person). • Modality is fully expressed here: some kinds of modality are speakeroriented, being performative in nature (Palmer 1979:58–59). These can only occur in dialogue. A case in point is the imperative—it has to do with the 1

Of course not a ‘case’, but rather a function: yā baxta ‘O woman’. See §6.4.1.



(al)locutive sphere, and the expression of will subsumed in it is not reported but expressed. The mere act of saying is the expression of locutor’s will, hence ‘performative’. The same applies to true interrogatives, which ask for an answer. • Tense (stricto sensu) shows the fullest spectrum possible, because verbal forms are here deictic to various temporal points; it means that many tense distinctions are found here but not in the narrative. For instance, present related functions such as the present perfect, or the actual present are found only here, and so is the future tense. • Space-time is reflected directly—now (vs. indirect then), yesterday (vs. indirect the day before), here (vs. indirect there), as well as concrete presentatives, which also have to do with here and now. • Personal involvement: Other categories, which are not necessarily deictic, are influenced as well, especially those which show personal involvement. A prominent member of this class is the focus system which is much more developed. The dialogue seems to be more “normal” for the speakers, the narrative often regarded as subsidiary, artistic, optional part. This may be observed in reading or watching a play being performed: it consists entirely of a dialogue, the narrative is replaced by a varying combination of acting and dramatic instructions. It is these differences, or these categories, that this part of the syntax sets on to describe.

6.1 FUNCTIONAL SENTENCE PERSPECTIVE Functional sentence perspective (FSP) and information structure are viewed in this framework as synonymous. It is about new vs. given, or salient vs. non-salient information. Some of these issues are dealt with above, under chapter 1, where new information (the RHEME) has to do with various copular expressions, which serve as its exponents. The given information (the THEME) was discussed as well under §1.2. This section describes topicality and focality. TOPIC is discussed above, §, having to do with the case of objects preceding verbal forms. The JZ focus system is described in Cohen 2008a and 2008b. FSP is examined mainly syntagmatically, because clauses with focus marking, for instance, do not commute with plain clauses in the same paradigms (for the simple reason that they are not found in the same context).



They do, occasionally, interchange among themselves. The circumstances for this marking are often clear from the syntagm only: we can gather whether entities are newly presented, repeated, reintroduced, made salient, etc., only by weighing the information already given.

6.1.1 Topic The clearest cases of TOPIC are those that cannot be explained away by the simple division used above of THEME and RHEME: (1)

ūʾāhet mayla qaḥra dīdox? ‘And as for you, what is your grief?’ 240


āna ʾáy kmā ́ šınne baxti lēwa smıxta ūʾatta smıxla ‘As for me, for several years now, my wife has not been pregnant, but now she got pregnant’ 227


āna +rāba bıt +ʿızāb́ wēle ʾısya ṓ yāla ṭāli ‘As for me, this child has come to me with great torment’ 5802


uʾāhet čukun dīwıt qarīwa llıbbi ṣúrr dīd gōri pkašfánnēlox ‘And you, since you are close to my heart, I will tell you the secret of my husband’ 815–816


(The king said: ‘I came to ask for your wife, I want to take her for my son’. The hunter replied:) āna baxti dīdi māṭo bšaqlıtta mınni, bıd dēma mūjeb ūbyāwıtta ta brōnox? ‘I, as for my own wife, how will you take her from me, in what way, and give her to your son?’ 501–502


atta āna la qam zaxıtti go kolāne ‘Now, as for me, you did not find me in the street’ MA 19.3

In the entire group of foregoing examples, one finds, in addition to the basic predicative link, external participants in the form of 1st and 2nd personal

This example may have another interpretation, as a cleft (see below, §5.1.3: ‘As for me, it was with great torment (that he had) come to me, this child’. This interpretation is possible because wēle can be either a presentative copula, or the past of √hwy. In the latter case, the rheme is the preceding entity (+rāba bıt +ʿızāb́ ), rather than ısya. This interpretation does not change anything with regard to the topic interpretation (ʾāna is still the topic, ṭāli still resumes it). 2



pronouns,3 which are analyzed here as topics. In other words, in ex. (2), the syntagm baxti lēwa smıxta ‘my wife has not been pregnant’ is an independent, full clause. The 1st person pronoun preceding it is an example of topic. It is linked to the clause following it by resumptive elements, in this case, by the 1st person attributive pronoun -i in baxti. The resumptive element can also be an object pronoun (ex. (6)) and even a subject index. (7)

́ aw dla qṭeʾle ʿāqıle la hullēle ta dēmıt qṭeʾle ʿāqıle hulle ṭāle kartīs, ču kartīs ‘To anyone who wanted he gave a ticket, (to) one who did not want, he did not give a ticket’ A 6.14


kud xa uxā halle xá kartīs ‘Each one, give him one ticket’ A 6.1

The case in exx. (7) and (8) is rare, and does not come from regular dialogue. JZ generally avoids topics whose function in the comment is dative, that is, the propensity is towards constructions as found in the beginning of ex. (7) (ta dēmıt...), that is, leaving the topical entity still marked by a preposition. Another similar example is the following: tūma ma bōzāna ʾıbbe? ‘The garlic, what should I do with it?’ 299


Here the resumptive element is the attributive pronoun appended to the preposition ib-. The topic functions as an anchor on which the following clause is predicated. In ex. (3), +rāba bıt +ʿızāb́ wēle ʾısya ṓ yāla ṭāli is said about the child. The entire construction, however, ties this entire clause to āna ‘I’. In contrast to the modern trend of excluding this type of relation from syntax and classifying it under ‘pragmatics’, it seems more efficient here to relate the relation to syntax; the exponents are syntactic (the extraposition of the topic, as well as the resumptive element in the comment, that is, the clause following the topic). Moreover, the relationship between the topical entity and the comment is here viewed as parallel, or very similar to the predicative link, In fact, the independent 1st and 2nd persons exponent cannot commute directly with any substantive, which makes it a priori different from other entities deemed pronominal. They can nevertheless be linked to the group of 3rd person personal pronouns, which can and do commute with substantives in the same syntactic setting: āya baxta–la (‘she is a woman’ 676) vs. āna baxta wāna (‘I am a woman’ 59). The interchange is not ‘clean’, because the entity in question commutes together with its resumptive element. In the paradigm consisting of these examples, the difference is 1st. vs. 3rd person. 4 A 6 is not a folktale, and is similar in many parameters to the dialogue. 3



the nexus, discussed in chapter 1 above. The difference from the constructions dealt with there is that here we do not have any copular expression whatsoever; an entire clause functions as the rheme (what is here termed ‘comment’), and that the relationship between the topic and the comment is marked by different exponents, because the construction functions on a different scale. The construction can consist of two topics, that is, contain several cases of nexus-like constructions. One case is ex. (5), where we have āna ‘I’ as the primary topic, and baxti dīdi ‘my wife’ as secondary. The table below illustrates the analysis of ex. (4), which is similar structurally: level

ʾāhet (...)



2 3


resumed by -ox

ṣúrr dīd gōri




resumed by -ne-

theme is 1st pers. (-an-)



That is, in this example three predicative links are found (not including the causal clause), in three different levels of hierarchy: 1. between ʾāhet and its comment (ṣurr... pkašfánnēlox); 2. between ṣurr and its comment (pkašfánnēlox); and 3. within the verbal clause between the 1st person and the lexeme kšāfa ‘reveal’. TOPIC then is here viewed as the parallel of the THEME above clause level. The foregoing examples all contain 1st and 2nd personal pronouns for a reason: the 1st and 2nd persons are different in this respect from the 3rd person, in that they need not be introduced into the dialogue, nor explained—‘I’ is always the speaker, ‘you’ is always the addressee, both well marked in verbal forms and in copular expressions. Hence, the addition of the independent pronoun is not part of the minimal predicative expression, but something different, which necessitates motivation. The motivation for topical constructions in dialogue (narrative is somewhat different) is first and foremost contrastivity with other themes or topics: while the idea of ‘topic continuity’ is reflected in the verbal form itself, topic shift needs to be clearly marked: (10) āna bıdamxen ū ʾāhıt nṭōr la ʾāse xa nāša ‘I will sleep whereas

you guard (so that) a man should not come’ 86–87

(11) mırra baxte: ṭayyıb, ʾāna bāzan uʾāya štōya mın gēbi ‘His wife

said: fine, I will go, and she, let her be in my stead’ 691



The contrast obtains between ‘I’ and ‘you’ in ex. (10) and between ‘I’ and ‘she’ in ex. (11). These examples are different from the examples above in that the topic is correferential with the theme, whose exponent is the subject index in the verbal form. The correferential subject index functions as the resumptive element of the topical construction, which makes both structures equivalent. The following pair of examples opposes the same referent in nontopical and topical function: (12) lákīʾētun ēka mṭāšítūlu pāre dōxun. āna gımṭāšınnu pāre dīdi go

kōpāla ‘You do not know where you should hide your money, whereas I hide my money in a walking-stick’ 338

(13) āna pāre dīdi gımṭāšınnu xē xā balāṭa gō bēsi ‘I, on the other hand,

my money, I hide it under a tile at my house’ 340–41

In both examples we have āna as first topic. The difference lies in the syntagm pāre dīdi ‘my money’, which, in ex. (12) is the object, while in ex. (13), where it precedes the verbal form, it is the second topic. In ex. (12) the object, ‘my money’, which is in potential contrast with ‘your money’, goes unmarked, whereas the contrast between 2nd person plural and ‘I’ is marked. In ex. (13) both ‘I’ and ‘my money’ are marked for contrast, standing as topics. Another set of examples seems to testify to a related, but different phenomenon. The 1st and 2nd person plural category is different from the singular: whereas the latter is at any point referential to a single entity only, the former are not: the idea subsumed by ‘we’ and ‘youCP’ consists, in addition to the speaker or addressee respectively, of additional 2nd or 3rd person referent. For this reason, these plural indices are sometimes further specified: (14) uʾāna uʾaxōni wax yımye ‘Me and my brother, we have sworn...’


(15) mā dām āhet hatxa gōzıt bıd danya pāre āna ūʾāhıt pēšax mux

axawāsa ‘As long as you are doing thus with the money, you and I, we will be like brothers’ 5

(16) mpıqlan āna ūxūrāsi bızvāra bıt dunye ‘We went out, me and my

friends, touring the world’ 373

(17) pšātax ʾāna uʾāwa mux trē dōste ‘We will drink, me and him, like

two friends’ 805

The explicit specification of the referents of the subject indices occur on either side of the verbal form: in exx. (14) and (15) it precedes the verbal



form, whereas in exx. (16) and (17) it follows, unlike the cases of topic discussed so far. As is shown below (§, the independent personal pronoun, when following the verbal form in the dialogue, is generally focal, which is clearly not the case here. This case (exx. (16)–(17)) is not topical, but rather an expansion of the subject index (termed antitopic). The difference in analysis originates in the different relationships between the unit and the subject index in the verbal form: while the topic has a relationship similar to the predicative link with the comment, this specfication does not, and it is deemed to be appositive to the subject index in the verbal form. In similar examples, the specification is even farther from the verbal form: (18) ʾāse bābi maḥkēn ʾımme ūysálōxun xa-šapsa-xēt ʾāhet ūʾáw jwanqa

kıslēni ıl bēsa ‘(When) my father comes I will speak to him, so come next week, you and this youngster, to our house’ 5435

(19) lá kēsētun gō bēsi bıžrab, kútrōxun baxtāsa uʾāhet yā ́ jwanqa ‘You

will not come into my house anymore, both of you, wives and you, O youngster’ 828

Yet another related phenomenon is exemplified next: (20) ʾaxnan huzāye

rāba gdōqax gyānan qlīwe mqabıl qarwax ıl baxta ‘Us Jews, we keep ourselves clean before we get close to a woman’ 489 +

In this case, huzāye ‘Jews’ is an expansion of ʾaxnan, as well as the reason for the contrastive topic, ‘us Jews’ as opposed to other groups, similar to the explicit exclusive French expression ‘nous autres’. There is one more way to mark contrasive topic, with the particle ši: (21) qṭeʾlox dūmıki uʾāna-ši qam naʾsınnox ‘You severed my tail while

I bit you’ 455

(22) qam zaʾpíttūli ıl warya uāna ši māxınna go pāsōxun ē msāsa ‘You

pushed me out while I threw the scales at your face’ 258

This particle is basically a focus particle (see below), and only in very few cases does it function as topic marker. mnōš- (followed by the genitive suffixed pronoun) is, like the particle ši, used (apart from its lexical value, ‘alone’)6 mainly as focal particle (see 5 6

This may be an asyndetic temporal clause, see n. 32. ūʾāna msōfírwāli mnōši ʾıl barīya ‘And I travelled alone to the desert’ (239).



below). However, there are a couple of cases where it is used in a topical construction together with ʾāna: (23) ġēr ta yatūme ūta gawāye–lu. āna mınōši go xā čādırre gdamxēna

‘It is only for orphans and beggars. I myself sleep in a tent’ 4–5

Note that in certain cases there is no motivation for what seems like a topic. Compare the following example, with parallel constructions: ́ ınkan āhet qam dōqıtti šımmox ıllox īle (24) yā ḥakōma Aharón Rašīd,

ḥakōma. ūʾınkan la qam dōqıtti šımmox hāwe xa xmāra ‘O king ́ 7 if you catch me, your name befits (lit. on) you, Aharón Rašīd, (o) king. But if (you) do not catch me, let your name be a donkey’ 11

Note the symmetry between both conditional structures: ‘you’ is not called for, since the vocative and the subject index in the verbal form take care of the retrievability factor, who is this 2nd person stands for. The vocative function, often marked, like in Arabic, by yā is not the topic (see §6.4.1).

6.1.2 Questions and answers Part of the field termed information structure includes asking for information and supplying it. In chapter 1, dealing with the predicative link, there were sporadic references to this issue in general. However, since this issue belongs by and large to dialogue only (occurrences of what looks like questions and answers in narrative are in fact exponents of various narrative functions, and are discussed above, §5.1.3). Questions are traditionally divided into nexus questions, which ask for yes–no answers, and pronominal questions, which contain an interrogative pro-form. The former type is exemplified in the following pair of examples: (25) ēhā–la ʿ ılbe dīdax? gımra he ‘Is this your box? She says: Yes’ 15

(and similarly 34, 103, 155, 156, 182)

(26) xa laxma ʾaxlēn? laʾ. tre? laʾ. ʾarba? laʾ. kud šōʾu baxlınnu ‘Should

I eat one bread? No. Two? No. Four? No. I will eat all seven’ 521


Aharón Rašīd́ is obviously a Hebrew conversion of Harūn ar-Rašīd.



The answer can in fact house both the idea of ‘yes’ (he) and a repetition of the questioned nexus: (27) ınkan atta mapqınna ʿılbe dīdax bıyāʾatta? gımra: he bıyāʾanna ‘If I

take out your box, will you recognize it? She says: Yes, I will recognize it’ 15 (and similarly 57, 316, 616)

(28) gēmır: bābo, qṭılālox šōše? gēmır: la, lē(wı)n qṭīla dīda ‘He says:

Father, did you kill Šōše? He says: No, I have not killed her’ LH 3.4

In a few cases, repeating the verb form in order to affirm the nexus seems to be enough: (29) zılle kıs ʾaxōne. gēmır: sēlox? gēmır: sēli ‘He went to his brother.

He says: You came? He says: I came’ SS 28

(30) ʾīs tām xa gōra huzāya. mírrūle: īs ‘Is there a Jewish man there?

They said: There is’ 613

This behavior assists in understanding the syntactic nature of he ‘yes’: it occupies the same slot, the answer, as a clause (exx. (29) and (30)), hence it is capable of representing an entire clause in this syntactic environment, very much like the pro-clausal adverb hādax ‘thus’ in other environments.8 The fact that they are compatible with one another in the same answer does not change this analysis—they co-occur in some kind of apposition. The answer, whether consisting of yes/no or of a clause, merely affirms (or denies) the nexus at the base of the question. Pronominal questions are somewhat more complicated. The interrogative pronoun in the question corresponds to, or represents the new information (that is, the rheme) in the answer: (31) mani mírrēlox ʿazmıtti? mırre ʾō gōra: baxti mírrāli ʿazmınnox

‘Who told you to invite me? My wife told me to invite you’ 675–676

(32) kma bıyāwıtti ınkān ā na pasxınna ṭālox ay ʾurxa ...? mōjıble

ḥakōma ū mırre: palgıt ḥ ukūm dīdi bıyāwınna ṭālox ‘How much will you give me if I open the road for you...? The king answered: I will give you half of my kingdom’ 30

Indeed, in some languages the word for ‘yes’ is derived from such a pronominal entity, e.g., Romance si (