Neo-aramaic and Its Linguistic Context 9781463204105, 1463204108

Citation preview

Neo-Aramaic and its Linguistic Context

Gorgias Neo-Aramaic Studies

14 Series Editors Geoffrey Khan Hezy Mutzafi

This series contains monographs and edited collections relating to the modern dialects of Aramaic, including linguistic studies and grammatical descriptions of the dialects, as well as the literature of the Christian and Jewish communities that speak them.

Neo-Aramaic and its Linguistic Context

Edited by

Geoffrey Khan Lidia Napiorkowska

9

34 2015

Gorgias Press LLC, 954 River Road, Piscataway, NJ, 08854, USA www.gorgiaspress.com Copyright © 2015 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.

2015

‫ܐ‬

ISBN 978-1-4632-0410-5

9 ISSN 1935-4428

Library of Congress Cataloging-in-Publication Data A Cataloging-In-Publication record is available from the Library of Congress. Printed in the United States of America

TABLE OF CONTENTS

PREFACE .........................................................................................................VII ABBREVIATIONS ................................................................................................ IX STUDIES IN THE ṬUROYO VERB .............................................................................1 Yulia Furman and Sergey Loesov TOWARDS A DESCRIPTION OF WRITTEN ṢURAYT/ṬUROYO: SOME SYNTACTIC FUNCTIONS OF THE PARTICLE KAL ...............................................................29 Maciej Tomal PRELIMINARY REMARKS ON DISCOURSE PRAGMATICS IN ṬUROYO ........................... 53 Michael Waltisberg

THE TURKISH LEXICAL INFLUENCE ON ṢŪRAYT/ṬŪRŌYO: A PRELIMINARY SELECTION OF EXAMPLES .......................................................................... 69 Aziz Tezel NEOLOGISMS IN ṢŪRAYT/ṬŪRŌYO ................................................................... 100 Sina Tezel THE THREE /R/S OF BAṚTƏḶḶA ....................................................................... 110 Kristine Mole A DIACHRONIC PERSPECTIVE ON EMPHASIS IN CHRISTIAN DIYANA -ZARIWAW ........ 130 Lidia Napiorkowska

IV

DOMAINS OF EMPHASIS, SYLLABLE STRUCTURE AND MORPHOLOGICAL BOUNDARIES IN THE CHRISTIAN URMI DIALECT OF NEO-ARAMAIC ...................................145 Geoffrey Khan THE HISTORY OF THE FUTURE ..........................................................................162 Samuel Ethan Fox THE ORIGIN OF THE PERIPHRASTIC PRETERITE KƏM/QAM-QĀṬƏLLE IN NORTHEASTERN NEO-ARAMAIC .........................................................................172 Steven E. Fassberg THE PRESENT CONTINUOUS IN THE NEO-ARAMAIC DIALECT OF ʿANKAWA AND ITS AREAL AND TYPOLOGICAL PARALLELS .......................................................187 Roberta Borghero CAUSATIVE-INCHOATIVE ALTERNATION IN NORTH-EASTERN NEO-ARAMAIC ..........207 Kathrin Göransson SPLIT ERGATIVITY IN THE NENA DIALECTS........................................................232 Alessandra Barotto THE PARTICLE WAL IN THE JEWISH NEO-ARAMAIC DIALECT OF ZAKHO..................250 Eran Cohen MORE FROM H. J. POLOTSKY’S NACHLAẞ ON THE VERB IN URMI .........................271 Olga Kapeliuk CHRISTIAN SALAMAS AND JEWISH SALMAS: TWO SEPARATE TYPES OF NEOARAMAIC ..............................................................................................289 Hezy Mutzafi SOME FEATURES OF THE GAZNAX DIALECT (SOUTH-EAST TURKEY).......................305 Ariel Gutman

V

HOW

A NEO-ARAMAIC SPEAKER (AVIDANI OF ʿAMIDYA) COPES WITH A BIBLICAL ARAMAIC TEXT (BOOK OF DANIEL): A SURVEY OF MISTRANSLATIONS OF VERBAL AND NOMINAL FORMs ................................................................ 322 Yona Sabar

FOLK-NARRATIVES IN THE JEWISH NEO-ARAMAIC DIALECT OF ZAKHO: THE CASE OF YOSEF VE-ʾEḤAV ............................................................................... 331 Oz Aloni SONGS OF THE ASSYRIAN MOUNTAINEERS ......................................................... 345 Nineb Lamassu NEO-MANDAIC IN MANDAEAN MANUSCRIPT SOURCES ........................................ 367 Matthew Morgenstern TENSE, ASPECT AND MOOD IN THE DOCTRINE OF JOHN ....................................... 397 Charles G. Häberl VERB-GOAL (VG) WORD ORDER IN KURDISH AND NEO-ARAMAIC: TYPOLOGICAL AND AREAL CONSIDERATIONS .................................................................. 407 Geoffrey Haig ON

THE

CONVERGENCE OF VERBAL SYSTEMS OF ARAMAIC AND ITS NEIGHBOURS. PART I: PRESENT-BASED PARADIGMS ....................................................... 426 Paul M. Noorlander and Donald Stilo

ON

THE

CONVERGENCE OF VERBAL SYSTEMS OF ARAMAIC AND ITS NEIGHBOURS. PART II: PAST PARADIGMS DERIVED FROM PRESENT EQUIVALENTS .............. 453 Donald Stilo and Paul M. Noorlander

INDEX........................................................................................................... 485

PREFACE This volume contains papers that were given at two conferences on Neo-Aramaic. The first conference took place at Lucy Cavendish College, Cambridge, in July 2011, which was organized by Lidia Napiorkowska, Illan Gonen and myself. The second, which was organized by Steve Fassberg, Hezy Mutzafi and Simon Hopkins, was held at the Israel Institute for Advanced Studies, Jerusalem, in June 2013. The collection of papers has been supplemented by a few additional ones that were not presented at these conferences. The Cambridge conference in 2011 was concerned in particular with features in Neo-Aramaic that have been brought about by language contact. We were privileged to have as our guests Geoffrey Haig and Don Stilo, who are specialists in the nonSemitic languages of western Asia that are or have been historically in contact with Aramaic. Their papers are presented at the end of this volume, those of Don Stilo coauthored with Paul Noorlander, who is currently researching contact-induced features in Neo-Aramaic verbal systems for his Ph.D. thesis. These papers show convincingly that many of the features of syntax and morphosyntax in Eastern NeoAramaic dialects, particularly those of the North-Eastern Neo-Aramaic (NENA) subgroup, have parallels in many non-Semitic languages of the area. Their papers are clear testimony to the fact that the historical development of Neo-Aramaic cannot be fully understood without taking into account the structures of the languages with which the dialects have been in contact. These papers also show that the parallels have developed in the Neo-Aramaic dialects by varying degrees of convergence with other languages, which makes their relationship with these languages complex. The other papers in the volume concern a wide range of topics in the Eastern Neo-Aramaic dialects, including Ṭuroyo (also know as Ṣurayt), NENA and NeoMandaic. They all make important contributions to the documentation of these dialects and to the understanding of their historical development. There is a great diversity across the Eastern Neo-Aramaic dialect area, which, compounded by the fact that many of the dialects are now highly endangered, is a great challenge for researchers engaged in the task of their documentation and in the investigation of their historical background. This diversity, however, also enriches the field and as

VIII more details of the dialects are documented, their historical development and relationship gradually become clearer. This volume advances the field of Neo-Aramaic in the various dimensions mentined above, namely issues of language contact, documentation and language history. Much of the data published here and the accompanying theoretical discussions are also of significance for General Linguistics, especially for Language Typology. The examples cited in most of the papers are accompanied by morpheme-bymorpheme glosses. These glosses follow the general principles of the Leipzig Glossing Rules. There are, however, variations in their application across the papers. It was an editorial decision to retain this variation, since it often reflects differences in interpretation of the data by the respective authors. I acknowledge here with gratitude the support given to the Cambridge conference by the Faculty of Asian and Middle Eastern Studies and the Golden Web Foundation and the support given to the organization of the Jerusalem conference by the Israel Institute for Advanced Studies. I should also like to express my thanks to Lidia Napiorkowska for her help in editing the volume. Geoffrey Khan Cambridge, July 2014

ABBREVIATIONS I, II, III—class, type I, II, III etc. 1,2, etc.—morphemes of Set 1, Set2, etc. A—agent ABL—ablative ABS—absolutive ACC—accusative ADD—additive ADJ—adjective ADP—adposition AGR—agreement marker AGT—agentive ALLT—allative AOR—aorist Ar.—Arabic ART—article ATR—advanced tongue root feature AUG—augment AUX—auxiliary C/c.—common CLC—clitic CL—noun class CNST—genitival construction or construct state COMP—complement CONJ—conjunction CONV—past convertor COP—(present) copula COP.NEG—negative copula COP.PST—past copula DAT—dative DEIC—deictic

DEM—demonstrative DET—determinator DETRANS—detransitive DIR—directive DIRC—directional particle DO—direct object DUR—durative EMPH—emphatic ERG—ergative EXIST—existential EXP—experiencer EZ—ezafe (-like suffix) F—feminine FUT—future marker GER—gerund HAB —habitual IMP—imperative IMPRF—imperfect INDF—indefinite IND—indicative prefix/particle INETR—interrogative INFECT—infectum INF—infinitive INS—instrumental INTER—interrogative INTR—intransitive INTRJ—interhection J—Jewish Kurd.—Kurdish

X L—Neo-Aramaic past person agreement marker formant LNK—linking particle LV—light verb MAL—malefactive MARK—marker M—Middle Persian M—masculine NC—numeral classifier NEG—negator NOM.ACT—nomen actionis NOM—nominative NUC—nucleus marking OA—Old Aramaic OBL—oblique OP—Old Persian OPT—optative PAM—person agreement marker PART—partitive PAS—passive PAST—preverbal past tense particle Pers.—Persian PL—plural PN—proper name POSS—possessive suffix P—patient PEJ—pejorative PREFIX—prefix PRES—presentative PRET—preterite PRF—perfect(ive)

PROG—progressive (marker) PROH—prohibitive PRON—pronoun PRP—preposition PRS—present PST —past tense/form PTC—particle PTCP—participle PUNCT—punctual Q—question word/particle QUANT—quantifier REFL—reflexive REL—relative morpheme/subordination particle RES—resultative participle RTR—retracted tongue root feature SET.POSS—oblique person agreement marker SG—singular SOURCE—source STAT—stative SUBJ—subjunctive SUBOR—subordinator SUBS—substantive SUB—subject TV—thematic vowel V.EMPH—verbal emphatic particle VBD—verboid VB—verb VERB.NOUN—verbal noun VOC—vocative

The form qaṭəl (appearing elsewhere as the Present Base, subjunctive, non-past etc.) has a wide range of functions, which makes it difficult to ascribe to it a single label that captures all its uses. For this reason, rather than appearing with a different gloss in each article, it is left unmarked in the volume.

STUDIES IN THE ṬUROYO VERB YULIA FURMAN AND SERGEY LOESOV

1

For Otto Jastrow and Shabo Talay The first part of the paper is a report on the Verb Glossary in progress. In the second part, we use the Glossary database for a study of the historical grammar of Ṭuroyo, in which we discuss the two-place experiencer verbs of Ṭuroyo that are qatǝl-shaped, i.e., morphologically intransitive.

I The present writers’ interest in Ṭuroyo was first stimulated by their work on a history of the Aramaic language. As is well-known, the key problem in the history of Aramaic is to identify the causes of the development from the Middle to the Modern form of the language in the Eastern group of dialects, i.e., “the morphological revolution,” the reshaping of the finite verb in what is now North-Eastern Neo-Aramaic (NENA) and Ṭuroyo-Mlaḥsô. Ṭuroyo is the most grammatically conservative among the modern Eastern Aramaic varieties (if we exclude Neo-Mandaic from consideration). The Ṭuroyo verb appears to represent more closely than other dialects the form of the Aramaic verb that arose at the period of the “revolution.” The claims of the last paragraph can be justified by the following facts:

1

Sergey Loesov presented an earlier version of this paper at the conference “Neo-Aramaic Dialectology: Jews, Christians and Mandaeans” held at the Hebrew University of Jerusalem, June 26–27, 2013. We are grateful to Shabo Talay who corrected some errors in the draft of the paper. Our thanks go to Ilya Arkhipov who discussed with us some details of ergative alignment in Ṭuroyo. The research has been supported by RFH grant No. 14-04-00374.

2

YULIA FURMAN AND SERGEY LOESOV

1) The Ṭuroyo verb keeps intact two different shapes of the G-stem Preterite depending on the transitivity of the root, e.g. nšəqle ‘he kissed (him)’ vs. nafəq ‘he went out/he left’. We are assuming that the ancestors of all Eastern Neo-Aramaic dialects used to have two Preterite shapes (cf. in particular Hopkins 1989). This evidence, along with some vestiges of object agreement, is traditionally described in terms of split ergativity (and see Khan 2007 on the situation in NENA). 2) Unlike in NENA, all TENSE-ASPECT-MOOD-DIATHESIS meanings of Ṭuroyo verbs are still expressed by synthetic conjugations, sometimes in combination with particles (prefixed, suffixed or infixed), which have no inflection of their own. There are no analytic verb forms as in English, Standard Literary Arabic, or NENA. In other words, there are no verbal conjugations consisting of an auxiliary inflected for person + a form of the lexical verb. In particular, against the background of NENA, Ṭuroyo and Mlaḥsô are special in having synthetic passives (Jastrow 1996). These features are retentions and must have been characteristic also of the unattested ancestors of NENA. 3) All nine participles of Middle Eastern Aramaic (see the table in Brockelmann 1962: 126-129) have become bases of Ṭuroyo finite inflections, cf. Jastrow 1967 with previous literature, in particular Siegel 1923. Thus, the Ṭuroyo verb is the key to the understanding of the crucial period in the history of Eastern Aramaic, i.e., the transition from the Middle to the Modern period. At a certain point in our work, it became clear to us that a morphosyntactic description of the Ṭuroyo verb in the form of a glossary would produce results important for further historical research, especially given that Ṭuroyo lexicography is still rudimentary. The state of the art is as follows: Hellmut Ritter’s Ṭuroyo-Wörterbuch, a typewritten manuscript with handwritten additions and corrections, was published posthumously (Ritter 1979) and was recently digitized by Veronika Betzold, an MA student of Shabo Talay (Bezold 2012). It has everything but verbs. It constitutes a solid base for a future comprehensive dictionary of Ṭuroyo, as far as nouns and uninflected words are concerned. Hellmut Ritter created the list of lemmata for this dictionary drawing principally upon his own fieldwork. Yet our reading experience shows that in the three volumes of oral texts Ritter had published there occur dozens of nouns that do not appear in the Ṭuroyo-Wörterbuch. A second book-length contribution to lexicography is Jan Beṯ-Şawoce’s “Xëzne d xabre Ordlista: Şurayt-Swedi” (Beṯ-Şawoce 2012). Verbs are represented in the “Xëzne d xabre” in the form of action nouns (=infinitives) of the G-stem, with only one to three glosses. Derived stems are not mentioned. The corpus upon which “Xëzne d xabre” is based is not indicated.

STUDIES IN THE ṬUROYO VERB

3

There are also glossaries in Otto Jastrow’s Lehrbuch (Jastrow 2002) and Shabo Talay’s Lebendig Begraben (a testimony of a Syriac Orthodox priest who was taken hostage by Hisbollah, Talay 2004). All verb roots and pertinent examples from these two books are now part of our Verb Glossary. As a preliminary step towards the Verb Glossary, Loesov created a pattern of the entry. This entry organization will facilitate a future study of the Ṭuroyo verbal morphosyntax, which is of great importance both for the general linguist and the student of the Semitic verb. The lemma of the entry is the verbal root. It is immediately followed by etymological information, e.g.

nḥt MEA nḥt SL 909f.: ‘go down, descend’

[MEA is Middle Eastern Aramaic, SL is Sokoloff’s Syriac Lexicon. For the abbreviations used in this study, see a list at the end of the paper.] There follows a bold face Roman numeral (I, II or III) which stands for the respective stem. Then follow the basic (i.e., 3sg.m.) shapes of the Preterite and the Infectum, Jastrow’s “Präsens.” We use the label “Infectum,” following the model of Latin grammar, because by itself the form does not express the present tense. It is, in fact, a conjugated base used to encode PRESENT, FUTURE, VOLITIVE, IMPERFECT and more Tense–Mood–Aspect (TAM) meanings in combination with various clitics and affixes. In almost all cases, the shapes of the Preterite and the Infectum are predictable, yet we believe it will be useful for the reader who is not familiar with the language to specify them in each entry. The paragraph that follows the basic forms contains numbered lexical meanings of the stem. The glosses are currently all in German, because they are borrowed from Hellmut Ritter’s Grammatik (Ritter 1990) and the published field corpus, where all texts are accompanied by German translations, e.g.

I: naḥǝt/noḥǝt 1. herab-, hinab, ab-, aussteigen. 2. auf freien Platz, freies Feld hervortreten, hinausziehen, auf den Kampfplatz treten, zum Kampfe antreten, losgehen auf (l-); 3. ausspielen (Schachfigur, Spielkarte), einen Zug machen. 4. fallen (der Regen). 5. laufen, strömen, rieseln (die Erde, das Wasser, die Tränen).

There follow boxes with verbal inflections: in case of transitive verbs, these are the Preterite and the Infectum of both Active and Detransitive, the latter being our label for the traditional Passive. We use this alternative label, “Detransitive,” to highlight the fact that the respective verb bases encode various valence-decreasing

4

YULIA FURMAN AND SERGEY LOESOV

operations, not just the agent backgrounding (= passivization). It appears that the agent backgrounding/passivization is not the most frequent reading of the traditional “Passive.” Indeed, the detransitive form is attested even for intransitive roots. Since it is not always possible to predict diathetic (and even lexical) meanings of the Ṭuroyo detransitive forms on the basis of the direct voice meanings, we provide the “Detransitive” box with its own glosses (see the sample entry below). The direct voice bases are the default ones and therefore are presented first and are unmarked, while detransitive bases bear the label Detransitive. Within the basic stem, the nafəq-shaped Preterite is flagged as Preterite Intransitive, while the nšəqle-shaped Preterite is simply “Preterite”. We have also introduced “subordinate” boxes for Preterite-wa (Jastrow’s Plusquampreteritum) and Infectum-wa (Ritter’s Habitualis/Irrealis). The main bulk of the dictionary is constituted by textual examples. The numbering of the examples within each box follows the running numbers of the glosses at the head of the stem entry (as shown above for naḥǝt/noḥǝt). Thus, because we have singled out five meanings for naḥǝt/noḥǝt, each of the boxes for finite forms will, ideally, include no less than five textual examples, although currently this cannot be achieved for each and every verb, given the limited volume of the corpus. We have also provided boxes for the Infinitive, which is a fully productive action noun, and for the Part act (= Active Participle). The latter is an umbrella concept for two different morphological shapes of deverbal adjectives. One of them is formed only for the G-stem static (or change-of-state) verbs (e.g. yatiwo ‘sitting,’ derived from yatu/yotu ‘sit/ sit down’), and this is diachronically the *qattīl pattern. The other is formed only for the G-stem verbs of intransitive motion, both telic and atelic, e.g. azolo ‘going’ (derived from azzé/əzzé ‘go’) and nafoqo ‘coming out’ (derived from nafəq/nofəq ‘come out’), and this is diachronically the *qattāl pattern. Otherwise, no “active participles” exist in the language, see Jastrow (2002: 141ff). We have included these two kinds of adjectives because, according to our preliminary observations, they play a part in the coding of the present time sense when used as nominal predicates. There are also boxes for the Part pass (= Passive Participle) of all three stems. They are formed freely from transitive verbs. As a sample, we present below the draft of a short entry, which has only two closely related lexical meanings of the G-stem and a small number of textual examples. The second of the two meanings is metaphorical and lexically bound, i.e. darbo ‘the way’ is an obligatory surface argument. Note that in the etymological sections of the Glossary the G-stem of the assumed etymon is treated as the default one: ʿwğ Ar. ʿwğ AWSG 891: krumm, gekrümmt, gebeugt, gebogen sein; Anat. ʿwğ VW II 76: biegen, beugen; Kinderib 99: 1. (intrans.) schief sein od. werden. 2. (trans.) biegen

STUDIES IN THE ṬUROYO VERB

5

I: ʿwǝğle/ʿowǝğ 1. schief, krumm machen. 2. (mit darbo) vom rechten Wege abweichen Preterite 1. ʿwǝğle u=basmoro ʻer schlug den Nagel krummʼ 467

2. faqaṭ ramḥǝl mǝ-d-hawina aṯto w-gawro, d-oṯe naqla d-ʿuğo u=darbo, Infectum

ġēr m-u=qaṭlo layt bayn l-ŭno w-liya ʻAber wenn wir morgen Mann und Frau geworden sind,wenn es vorkommt, dass sie vom geraden Wege abweicht, dann gibt es nur Töten zwischen mir und ihrʼ 467; 26/253

Detransitive ʿwiğ/mǝʿwǝğ

1. schief sein/werden. 2. (mit darbo) vom rechten Wege abweichen 1. ǝšmo d-mǝʿwǝğ u=qamyon-ste muḥaqqaq g-qulbína-wo ʻWenn der LastInfectum

wagen sich ein wenig schief gelegt hätte, wären wir bestimmt heruntergekipptʼ 11/145. 2. zlām d-mǝʿwǝğ m-u=darbo ʻein Mann, der vom Wege abweichtʼ 467

Part. Pass. ʿwiğo ʻkrumm, schief, zusammengedrücktʼ 467 The sample happens to show that, in the way characteristic of the language, under certain conditions the Detransitive may be similar to the direct voice in its lexical meaning. The next important stage of the project was taken in 2010–2011 by Loesov’s student Andrey Ontikov, who, in his MA thesis, entered alphabetically all verbal roots mentioned into Hellmut Ritter’s 800-page-long Grammatik and arranged all verb forms attested in this book in the entry pattern described above, along with some textual examples. Ontikov also provided preliminary etymologies of the roots. In this way, a draft of the Glossary came into being. Since then, the members of the Glossary team have been entering into the Glossary database new textual material from the corpus and improving on lexical meanings and etymologies. The text transcriptions of the published corpus are heterogeneous, consider especially the sub-phonemic and sometimes impressionistic renderings of the first text edition based on field research (Prym-Socin 1881). For the moment, there are no two publications of Ṭuroyo texts which follow the same rules for transcription. Yet

6

YULIA FURMAN AND SERGEY LOESOV

what we need for our lexicographic and grammatical purposes is a standardized phonological rendering, as simple/simplified as reasonable, especially because the Glossary may turn out useful for readers who are not well-versed in the language. As a result of our ongoing conversations with Otto Jastrow and Shabo Talay, we decided to use the rules of Jastrow’s Lehrbuch for all village Mundarten, with two kinds of simplifications: 1) We write dukṯo rather than dŭkṯo ‘Stelle, Ort’: since the CVCC environment does not tolerate long vowels, one is allowed to drop the breve. This simplification is valid for all short u-vowels in the CVCC context, most importantly for the Infectum of the Midyat dialect: ko-nušqi ‘they kiss,’ etc. 2) We write griš rather than grīš ‘er wurde gezogen,’ sim rather than sīm ‘es wurde getan,’ because in this morphological context the simple i-sign is an unambiguous indication of the vowel length, cf. grəšno ‘ich wurde gezogen’ and səmno ‘ich wurde getan’. This simplification is valid only for derivations from the *qtīl base. As for the Midyat variety, we follow Otto Jastrow’s description of the Midyat vocalism (a manuscript written in 2013). Jastrow formulates the most salient difference in the vocalism of the two dialects as follows: “All long vowels in open syllables which lose the stress are preserved as long in the villages. In Midyat only those long vowels which go back to a short vowel + geminated consonant remain long; if they go back to a long vowel they are shortened.” The shortening may involve a change in the vowel quality; the reader will learn the details from Jastrow’s study when it is published. At this early stage of the project, the corpus of the Glossary is almost entirely limited to published field research. At the moment, the corpus is divided into two parts. The most important part consists of the field research that forms the basis for Ritter’s Grammatik. It includes in particular three volumes of texts recorded and published by Hellmut Ritter (for details, see References). Ritter’s collection contains 117 texts, the numbering running through the three volumes. We refer to this collection by text and verse, with no additional sigla, e.g. 82/57. If the text in question is cited in Ritter’s Grammatik as well, we mention first the page number of the Grammatik: 94; 82/57, the number 94 referring to Ritter 1990: 94. In the Grammatik, there occur important verb forms presented in sentences out of context, which had been elicited by Ritter from his informants. We refer to these sentences by the page number of the Grammatik only. Prym-Socin 1881 is quoted by page and line (e.g. PrS 57/19), the page number of Ritter 1990 coming first if the PrS text is mentioned there. The textual Anhang of Jastrow 1967 (= LuF) is cited by page and verse. The second part of the corpus consists of texts published after the completion of the Grammatik, which still remains (and will probably always be) the main source of verb roots for our Glossary. These are texts recorded by Otto Jastrow for his Lehr-

STUDIES IN THE ṬUROYO VERB

7

buch (Jastrow 2002 = JL, cited by lesson-paragraph-verse, e.g. JL 9.9.11), Talay 2004 (LB, cited by verse), and various texts published by Otto Jastrow, of which only Jastrow 1968 (= MM, cited by verse) has been quoted in this paper. There are a number of literary texts written in Ṭuroyo at our disposal. They include among other things translations and free renderings of the Bible and other classical texts (e.g. Alice in Wonderland), an ABC book for native speakers, and essays by Jan Beṯ-Şawoce related to current politics. For various reasons, we have not yet included them into the Glossary corpus.

II The Glossary now includes a total of around 2000 verbal roots. According to the preliminary etymological analysis, some 600 roots are of Aramaic origin, some 700 roots are of Arabic origin, about 100 roots are of Kurdish origin, and 27 roots are of Turkish origin. There are, in addition, about 250 roots of unknown origin, for which further etymological research is needed. Though the Glossary is still in preparation, its draft can be used as a research tool, as we shall now show. The Glossary has 209 roots with qatǝl (or “nafǝq-shaped”) Preterites, 116 of them are of Aramaic origin, 81 are of Arabic origin, 12 have not yet been etymologized. Let us now look at the two-place experiencer verbs with qatǝl-Preterites. The case study is of interest both synchronically and also for the history of Aramaic. We have found ten verbs of this type that constitute a tightly knit semantic group: yalǝf ‘learn,’ aḏǝʿ ‘to know,’ abǝʿ ‘want,’ ṭaʿi ‘forget,’ šamǝʿ ‘hear,’ fahəm ‘understand,’ ʿayǝz ‘need,’ qadǝr ‘can, be able,’ ṭamǝʿ ‘desire,’ lazǝm ‘need, require’. The first five of them are of Aramaic origin and were singled out as qatǝl-“transitives” in Jastrow 1967: 71. The last five roots were borrowed from Arabic. In what follows, we reproduce and analyze some of the examples gathered in the Glossary. We deal with our textual examples in the following way. The transcription is normalized according to the standards described above, and the original German translations of the editions are maintained. We then add morphological glossing and English translations of the Aramaic, which results in a four-level construction. Abbreviated grammatical category labels are those of the Leipzig Glossing Rules. Additional abbreviations, of our own creation, are listed at the end of the article. Looking at the evidence of the corpus, one notices that the way these ten verbs link to their Stimuli may vary. In particular, the data gathered below show that the Infectum and the Preterite of the same verb within the group tend to use different methods of governing their Stimuli expressed by substantives. The distinction mir-

8

YULIA FURMAN AND SERGEY LOESOV

rors the split ergativity inherent in the G-stem of the Ṭuroyo verb. Time and again, the qatǝl-shaped Preterite of the ten verbs is syntactically intransitive, i.e., its Stimulus is governed by a preposition (in our sample, by l-, b- and ʿal-), while the Infectum of the same verb is in most cases syntactically transitive. That is to say, the qatǝl Preterite tries to avoid taking its Stimulus in the zero-marked form, which would result in both Subject and Stimulus arguments being assigned the same absolutive case. For detailed observations, see the syntactic notes following the textual evidence for individual verbs, where counterexamples are also recorded. ylf ‘learn’:

(1) yalǝf u=kurrǝko qroyo, msǝkle as=saḥrat bi=qrayto ‘Der Junge lernte lesen, und durch das Lesen empfing er Zauberkräfte’. (PrS 157/25). yalǝf-∅ learn.PRET.INTR-3SG.M 3SG.M as=saḥrat ART.PL=magic.PL

u=kurrǝk-o

ART.SG.M=boy-SG

qroyo, read.INF

msǝk-le acquire.PRET-

b-i=qrayt-o in-ART.SG.F=reading.SG

‘The boy learned how to read, [and] acquired [the art of] magic through reading’. (2) ilǝfla qroyo? omǝr ilǝfla mayiṯo ‘“Hat es (das Kamel f.) lesen gelernt?”—“Ja, es hat’s gelernt und ist gestorben.”’ (PrS 201/6) ilǝf-la learn.PRET-3SG.F

qroyo? read.INF

omǝr-∅ say.INFECT-3SG.M

ilǝf-la learn.PRET-3SG.F

mayiṯ-o-∅ die.PRET.INTR-F-3SG ‘“Did it (the she-camel) learn how to read?” He says: “It learned [to read] and died.”’ (3) bi=qriṯayḏan kǝtlan rawmo rabṯo, eba yalǝfno li=soba ‘In unserem Dorf haben wir einen grossen Teich, in dem habe ich schwimmen gelernt’. (JL 9.9.11) b-i=qriṯ-ayḏan in-ART.SG.F=village-POSSII.1PL

kǝt-l-an there_is-to-POSSI.1PL

rawm-o pond[F]-SG

STUDIES IN THE ṬUROYO VERB rab-ṯo big-F

eb-a in-POSSI.3SG.F

9

yalǝf-∅-no learn.PRET.INTR-M-1SG

l-i=soba to-ART.SG.F=swimming ‘In our village we have a big pond where I learned to swim’. For the Preterite of ylf both morphological shapes are attested, yalǝf and ilǝfle, while Shabo Talay tells us that ilǝfle is unusual; in (3) the target of learning is introduced as an indirect object (via the preposition l-), unlike in (1) and (2). In our view, the prepositional government of yalǝfno li=soba in (3) betrays the tendency of qatǝl-shaped experiencer Preterites to avoid absolutivus duplex. One has to bear in mind that in Ṭuroyo the genuine direct object does not take the preposition l-, unlike in Classical Syriac and other MEA varieties. ʾḏʿ ‘know’: (4) omǝr kāyso. aḏǝʿ i=emo-yo ‘“Schön,” sagte jener und erkannte, daß es die Mutter war’. (PrS 58/21) omǝr-∅ say.INFECT-3SG.M

kāys-∅-o. good-M-SG

aḏǝʿ-∅ know.PRET.INTR-3SG.M

i=em-o=yo ART.SG.F=mother-SG=COP.PRS.3SG ‘He says, “Good!” He realized that she was the mother’. (5) lo=kuḏʿət d-kəthət ṭayo ‘Du weißt nicht, daß du Muslim bist’. (JL 19.6.15) lo=k-uḏʿ-ət NEG=PRS-know.INFECT-2SG.M

d-kət-hət ṭay-∅-o that-COP-SUBJ.PRON.2SG.M Muslim-M-SG

‘You do not know that you are a Muslim’. (6) mak-koḏǝʿ b-lišone d-an=nune ‘Wer versteht die Sprache der Fische?’ (PrS 57/19) mak-k-oḏǝʿ-∅ who-PRS-know.INFECT-3SG.M ‘Who understands the language of fish?’ (7) edí tǝ=nošo lǝ=koḏǝʿ abxun

b-lišon-e-d-an=nun-e in-language[M]-EZ-of-ART.PL=fish-PL

10

YULIA FURMAN AND SERGEY LOESOV

‘Niemand erfährt dann (etwas) von euch’. (47/17) edí then

tǝ=noš-o none=person[M]-SG

lǝ=k-oḏǝʿ-∅ NEG=PRS-know.INFECT-3SG.M

ab-xun in-POSSI.2PL ‘Then no one will learn [anything] about you’. (8) hiye w-hiya aḏiʿiwa ḥḏoḏe bu=ṭuro ‘Die beiden hatten sich im Gebirge kennengelernt’. (MM 126) hiye he

w-hiya and-she

aḏiʿi-∅-wa know.PRET.INTR-PL-3-CONV

ḥḏoḏe each_other

b-u=ṭur-o in-ART.SG.M=mountain-SG ‘They came to know each other in the mountains’. (9) uʿdo hani tarte=ǝšne aṯina l-Ṣṭambǝl w-aḏiʿina lu=profesor Ritter, w-uʿdo komʿawnínale ǝšmo bu=lišono ṭuroyo ‘Jetzt vor zwei Jahren sind wir nach Istanbul gekommen und haben Professor Ritter kennengelernt, und jetzt helfen wir ihm ein wenig bei der Ṭuroyosprache’. (94/1) uʿdo now

hani tarte=ǝšn-e these two.F=year[F]-PL

w-aḏiʿ-i-na and-know.PRET.INTR-PL-1PL

aṯi-na go.PRET.INTR-1PL

l-u=profesor to-ART.SG=professor

ko-mʿawn-í-na-le PRS-help.INFECTII-PL-1PL-3SG.M.P

l-Ṣṭambǝl to-Istanbul Ritter Ritter

w-uʿdo and-now

ǝšmo b-u=lišon-o little with-ART.SG.M=language-SG

ṭuroy-o Ṭuroyo-SG ‘Now it is two years that we have come to Istanbul and met professor Ritter. And now we help him a little bit with the Ṭuroyo language’. As expected for a verb of propositional attitude, ʾḏʿ often takes sentential complements, cf. (4) and (5). The argument encoding the content of knowledge may be introduced by the preposition b- in the Infectum, example (6) and probably (7), and cf. Ritter 1990: 721: ‘Auch mit b- “verstehen, Bescheid wissen, sich verstehen auf.”’

STUDIES IN THE ṬUROYO VERB

11

This fluctuation in the coding of the Stimulus between zero and a preposition is typical of experiencer verbs in general, quite independent of the syntactic alignment problems, cf. an instance of prepositional government of ydʿ in Biblical Hebrew: (10) ʾīš lō yēḏaʿ bad-dəḇārīm hā-ʾēllɛ ‘Let no one know (about) these words’. (Jer 38: 24) Note the l-government of a qatǝl-Preterite in (9): aḏiʿina lu=profesor Ritter ‘we came to know professor Ritter’. ʾbʿ ‘want’: (11) u=zlāmano l=abǝʿ gowǝr ʿal a=tre=naʿimani d-kǝtne gabe ‘Der Mann wollte zu den beiden Kindern, die bei ihm waren, nicht wieder eine Frau nehmen’. (75/3) u=zlām-ano ART.SG.M=man-this.M gowǝr-∅ marry.INFECT-3SG.M

l=abǝʿ-∅ NEG-want.PRET.INTR-3SG.M ʿal because

d-kǝt-ne REL-COP-SUBJ .PRON.3PL

a=tre=naʿim-ani ART.PL=two=child-these

gab-e with-POSSI.3SG.M

‘This man did not want to (re)marry because of these two children who were at his side (lit. ‘with him’)’. (12) ono mi=naqla d-ḥzélilux, l=abǝʿno mqatanno aʿmux ‘Ich habe schon, als ich dich sah, nicht mit dir kämpfen wollen’. (78/115) ono I

m-i=naqla from-ART.SG.F=time

l=abǝʿ-∅-no NEG=want.PRET.INTR-M-1SG

d-ḥzé-li-lux when-see.PRET-1SG-2SG.M.P mqatan-∅-no fight.INFECTII-M-1SG

aʿm-ux with-POSSI.2SG.M

‘As soon as I saw you, I did not want to fight with you’. (13) bu=gabano pano abǝʿ d-ʿorǝq mu=ğăwāb ‘Er suchte der Antwort aus dem Wege zu gehen’. (MM 27) b-u=gab-ano in-ART.SG.M-side-this.M

p-ano in-this.M

abǝʿ-∅ want.PRET.INTR-3SG.M

12

YULIA FURMAN AND SERGEY LOESOV d-ʿorǝq-∅ that-escape.INFECT-3SG.M

m-u=ğăwāb from-ART.SG.M=answer

‘By all means (lit. on this side and that [side]), he tried (lit. wanted) to escape answering’. (14-15)

kobaʿno d-mšadrətli. bi=dukṯo d-aḏ= ̣ ḏ̣arbe

lo=kobaʿno

d-ono-ze

fašno

‘Ich will, daß Sie mich wegschicken. Auch ich will nicht da bleiben, wo es immer Schläge gibt (wörtl.: am Ort der Schläge)’. (JL 19.6.19) k-obaʿ-∅-no PRS-want.INFECT-M-1SG

d-mšadr-ət-li that-send.INFECTII-2SG.M-1SG.P

lo=k-obaʿ-∅-no NEG=PRS-want.INFECT-M-1SG b-i=dukṯ-o in-ART.SG.F=place-SG

d-ono-ze that-I-also

faš-∅-no stay.INFECT-M-1SG

d-aḏ̣=ḏ̣arb-e of-ART.PL=blow-PL

‘I want you to send me away [since] I do not want to stay at the place of blows anymore’. (16) kubʿi ḥalwo d-aryo ‘Sie wollen Löwenmilch.ʼ (75/276) k-ubʿ-i-∅ PRS-want.INFECT-PL-3

ḥalw-o milk[M]-SG

d-ary-o of-lion[M]-SG

‘They want lion’s milk’. (17) ṭr-oṯe l-arke uʿdo kobaʿne ‘Er soll hierher kommen, ich brauche ihn jetzt’. (MM 140) ṭr-oṯe-∅ OPT-come.INFECT-3SG.M

l-arke to-here

uʿdo now

k obaʿ-∅-n-e PRS-want.INFECT-M-1SG-3SG.M.P ‘Let him come here, I need him now’. (18) abiʿi lu=qaṭlayḏe, mazǝʿíwole ‘Sie wollten ihn töten, machten ihm Angst’. (21/39)

STUDIES IN THE ṬUROYO VERB abiʿ-i-∅ want.PRET.INTR-PL-3

13

l-u=qaṭl-ayḏe to-ART.SG.M=killing-POSSII.3SG.M

mazǝʿ-í-∅-wo-le frighten.INFECTIII-PL-3-CONV-3SG.M.P ‘They wanted to kill him, [and therefore] frightened him’. A sentential complement encoding the object of volition may be expressed by either an asyndetic jussive whose subject is coreferential with that of the main clause (11-12), or by a d-jussive whose subject does not have to coincide with that of the main clause (14-16). The object of volition is encoded by a substantive and an object pronoun in (16-17), while in (18) it is encoded by an substantival indirect object introduced by l-. In the latter instance, the predicate is once more a morphologically intransitive Preterite: abiʿi lu=qaṭlayḏe. ṭʿy ‘forget’: (19) haṯe ṭaʿínowayla d-ománnanxuyo meqǝm ‘Ich habe vorher vergessen, euch das zu erzählen’. (57/210) haṯe this.F

ṭaʿi-∅-no-way-la forget.PRET.INTR-M-1SG-CONV-3SG.F.P

d-omán-∅-na-nxu-yo that-say.INFECT-M-1SG-2PL.DAT-3SG.P

meqǝm before

‘I forgot to tell you this earlier’. (20) ŭno-stene mi=šrolo ḥšuli ruḥi d-kǝtno bu=qayṭo. băle ṭaʿino d-kǝtno bṢṭambul ‘Ich fühlte mich wirklich im Sommer. Aber ich vergaß, daß ich in Istanbul bin’. (8/10) ŭno-ste I-also

nem-i=šrol-o from-ART.SG.F=truth-SG

ḥšu-li think.PRET-1SG

ruḥ-i REFL-POSSI.1SG

d-kǝt-no that-COP-SUBJ.PRON.1SG

b-u=qayṭ-o in-ART.SG.M=summer-SG

băle but

ṭaʿi-∅-no forget.PRET.INTR-M-1SG

d-kǝt-no that-COP-SUBJ.PRON.1SG

b-Ṣṭambul in-Istanbul

‘As for me, I did think that I found myself in summer. But I forgot that I was in Istanbul’.

14

YULIA FURMAN AND SERGEY LOESOV

(21) ma hēš lo=ṭaʿǝt u=mḥoyayḏi? ‘Hast du denn noch immer nicht vergessen, mich zu verprügeln?’ (62/343) ma Q

hēš still

lo=ṭaʿ-∅-ǝt NEG=forget.PRET.INTR-M-2SG.M

u=mḥoy-ayḏi ART.SG.M=whip.INF-POSSII.1SG ‘You still have not forgotten to whip me?’ (22) a=bnoṯayḏe maṭǝn lu=gworo. lǝ=mšayele hiye aʿlayye xud ṭaʿǝlle ‘Die Töchter kamen ins Heiratsalter. Er fragte aber nicht nach ihnen, als ob er sie vergessen hätte’. (77/18) a=bn-oṯ-ayḏe ART.PL=daughter-PL-POSSII.3SG.M l-u=gworo to-ART.SG.M=marry.INF aʿl-ayye about-POSSI.3PL

maṭ-ǝn reach.PRET.INTR-3PL

lǝ=mšaye-le NEG-ask.PRETII-3SG.M

xud as

hiye he

ṭaʿǝ-∅-lle forget.PRET.INTR-3SG.M-3PL.P

‘His daughters reached marriage age. [However], he did not inquire about them as if he had forgotten them’. What is forgotten may be introduced by a d-clause, whether a verbal one with a jussive (19) or a copular one with kit (20). It may be encoded by an infinitive (21) or a bound pronoun as well (22). Note that in (18) abiʿi lu=qaṭlayḏe lit. ‘they wanted to his killing’ the government is different from (21) lo=ṭaʿǝt u=mḥoyayḏi lit. ‘you did not forget my whipping’. The reason for this is unclear. šmʿ ‘hear’: (23) šamǝʿ u=šulṭono sǝmle hawxa l-Abu-Zed ‘Hierauf hörte der Sultan, daß Abu-Zed solches getan habe’. (PrS 15/11) šamǝʿ-∅ hear.PRET.INTR-3SG.M hawxa so

u=šulṭon-o ART.SG.M=sultan-SG

l-Abu-Zed A-Abu-Zed

‘The sultan heard that Abu-Zed had done so’.

sǝm-le do.PRET-3SG.M

STUDIES IN THE ṬUROYO VERB

15

(24) băle aḥna šamiʿina d-komaḥarmǝt niše, gawre, flān ‘Doch, wir haben gehört, daß du Frauen und Männer und so weiter verdammst’. (LB 103) bale but

aḥna we

niš-e woman-PL

šamiʿ-i-na hear.PRET.INTR-PL-1PL gawr-e man-PL

d-ko-maḥarm-ǝt that-PRS-curse.INFECTIII-2SG.M

flān somebody

‘But we have heard that you are in the habit of cursing women, men, and so on’. (25) hano u=abro du=šēx, hēš, lu=ḥzéwayle lo=šamǝʿwa xabro me-fema lo=ṭowo ‘Der Sohn des Scheichs hatte bis dahin noch nie ein ungutes Wort aus ihrem Munde erlebt und gehört’. (94; 82/57) hano this.M

u=abr-o

ART.SG.M=son-SG

lu=ḥzé-way-le NEG=see.PRET-CONV-3SG.M xabr-o word[M]-SG

d-u=šēx of-ART.SG.M=sheikh

hēš yet

lo=šamǝʿ-∅-wa

NEG=hear.PRET.INTR-3SG.M-CONV

me-fem-a from-mouth[SG.M]-POSSI.3SG.F

lo=ṭow-∅-o NEG=good-M-SG

‘The sheikh’s son had not yet seen this; he had hardly heard a bad word from her lips’. (26) u=xabrano man šmǝʿle baʿ=ʿašayir w-baq=qabayl? ‘Wer hat je von dergleichen bei den Stämmen gehört?’ (29/92) u=xabr-ano ART.SG.M=word-this.M

man who

šmǝʿ-le hear.PRET-3SG.M

b-aʿ=ʿašayir among-ART.PL=tribe.PL

w-b-aq=qabayl and-among-ART.PL=clan.PL

‘Who among tribes and clans heard of this custom (lit. word)?’ The Stimulus participant is encoded either by a sentential complement (asyndetic or introduced by d-, 23-24) or as a noun in the absolutive case (25), the cause probably being the form ḥzéwayle, an ergative “Plusquampreteritum”.

16

YULIA FURMAN AND SERGEY LOESOV

Note that we have found one example of the l-Preterite for this root (26). šmǝʿle is definitely an innovation vis-à-vis šamǝʿ, it is a NENA-like development. fhm ‘understand’: (27) i=šariʿaṯxu ṣafro lǝ=fahimina mede mena ‘Heute Morgen haben wir von eurem Rechtshandel nichts begriffen’. (26/57) i=šariʿ-aṯxu ART.SG.F=case-POSSII.2PL

ṣafr-o morning[M]-SG

lǝ=fahim-i-na NEG=understand.PRET.INTR-PL-1PL

mede something[SG.M]

men-a of-POSSI.3SG.F

‘This morning, we have not understood anything about your case’. (28) i=aṯto du=tağǝr ḥzela u=maktub du=gawrayḏa, qrela fahímole ‘Die Frau des Kaufmanns sah den Brief ihres Mannes, las ihn und verstand ihn’. (23/47) i=aṯt-o ART.SG.F=woman-SG

d-u=tağǝr of-ART.SG.M=merchant[M]

u=maktub ART.SG.M=letter[M] qre-la read.PRET-3SG.F

ḥze-la see.PRET-3SG.F

d-u=gawr-ayḏa of-ART.SG.M=husband-POSSII.3SG.F fahím-o-le understand.PRET.INTR-3SG.F-3SG.M.P

‘The merchant’s wife saw the letter of her husband, read and understood it’. (29) fahǝm mʿašarle i=kačkayo kmo waxt, fahǝm ʿal i=lǝġaṯṯe ‘Er verstand, [was sie sagten]; er war ja eine Zeitlang mit jenem Mädchen zusammengewesen und hatte die Sprache [der Tauben] zu verstehen gelernt’. (57/178) fahǝm-∅ understand.PRET.INTR-3SG.M i=kačk-ayo ART.SG.F=girl-this.F ʿal about

kmo some

mʿašar-le live_together.PRETII-3SG.M waxt time[M]

fahǝm-∅ understand.PRET.INTR-3SG.M

i=lǝġ-aṯṯe ART.SG.F=language-POSSII.3PL

‘He understood. He spent with this girl some time, so he had come to under-

STUDIES IN THE ṬUROYO VERB

17

stand their language’. (30) fhǝmle lu=šēx ki hani b-lebo ṣafyo aṯǝn aʿme ‘Der Scheich merkte, daß diese Leute in aufrichtiger Gesinnung mit ihm gekommen waren’. (29/195) fhǝm-le understand.PRET-3SG.M b-leb-o with-heart[M]-SG

l-u=šēx A-ART.SG.M=sheikh

ṣafy-∅-o pure-M-SG

ki that

aṯ-ǝn come.PRET.INTR-3PL

hani these

aʿm-e with-POSSI.3SG.M

‘The Sheikh understood that these came with him without bad intentions’. (31) u=muxtar fhǝmle i=masale omǝr «ṭayyǝb uʿdo l-mǝn aṯutu l-gabi?» ‘Der Schultheiß verstand, was gespielt wurde, und sprach: “Schön! Und warum seid ihr jetzt zu mir gekommen?”’ (54/59) u=muxtar ART.SG.M=head

fhǝm-le understand.PRET-3SG.M

i=masala ART.SG.F=matter

omǝr-∅ say.INFECT-3SG.M

ṭayyǝb good

l-mǝn to-what

aṯ-utu to-side[M]-POSSI.1SG

uʿdo now l-gab-i come.PRET.INTR-2PL

‘The mayor understood the problem. He said, “Fine. Now then, why have you come to me?”’ (32) u=ḥakǝm ğwamēr b-ʿaqǝlyo fhǝmle ‘Der Fürst war ein Edelmann und klug; er verstand’. (102/44) u=ḥakǝm ğwamēr ART.SG.M=governor noble[M]

b-ʿaqǝl-yo with-reason[M]-COP.PRS.3SG

fhǝm-le understand.PRET-3SG.M ‘The ruler, being noble in mind, understood’. For the Preterite of fhm, both morphological shapes are attested, fahǝm and fhǝmle. In the corpus, we have found five tokens of fahǝm and thirteen tokens of fhǝmle. In (29) the object of understanding is introduced via the preposition ʿal, unlike in (27 [mede]), where it is introduced by ∅. Speculatively, fhǝmle looks younger than fahǝm, and see above on šamǝʿ/šmǝʿle.

18

YULIA FURMAN AND SERGEY LOESOV

ʿyz ‘need’: (33) ʿayizat mo=lirat ‘Du brauch[te]st 100 Pfund’. (513) ʿayiz-at need.PRET.INTR-2SG.F

mo=lir-at hundred=pound-PL

‘You needed one hundred pounds’. (34) lo=ʿayə́znole qǝrš ‘Ich brauchte keinen Groschen von ihm’. (513) lo=ʿayə́z-∅-no-le NEG=need.PRET.INTR-M-1SG-SOURCE.3SG.M

qǝrš penny[SG.M]

‘I did not need a penny from him’. Both sentences are taken from the same informant from the village ʿIwardo, and we have found no more examples in the published corpus. It seems the G-stem verbs bʿy and ʾbʿ can express similar notions, as well as lzm (see below). Be this as it may, (33) and (34) are counterexamples to our suggestion above that the qatǝlshaped Preterite of the nine verbs is syntactically intransitive, i.e., its Stimulus is governed by a preposition. qdr ‘can, be able’: (35) bi=ḥarayto qadǝrina d-mʿadlínala [i=gōro] ‘Finally we managed to fix it [= the roof]’. (1/19) bi=ḥaray-t-o in-ART.SG.F-last-F-SG

qadǝr-i-na can.PRET.INTR-PL-1PL

d-mʿadl-í-na-la that-repait.INFECTII-PL-1PL-3SG.F.P

[i=gōro] [ART.SG.F=roof-SG]

‘Finally we managed to repair it [= the roof]’. (36) u=admo knŭfaḥḥe w-kmawqarre, w-lo=fayǝš kqudri ruhṭi xayifo ‘Das Blut schwellt sie (die Wanzen) auf und macht sie schwer, und sie können dann nicht mehr so schnell laufen’. (1/26) u=adm-o ART.SG.M=blood-SG

k-nŭfaḥ-∅-ḥe PRS-make_swell.INFECT-3SG.M-3PL.P

STUDIES IN THE ṬUROYO VERB

19

w-k-mawqar-∅-re and-PRS-make_heavier.INFECTIII-3SG.M-3PL.P w-lo=fayǝš-∅ and-NEG=remain.PRET.INTR-3SG.M ruhṭ-i-∅ run.INFECT-PL-3

k-qudr-i-∅ PRS-can.INFECT-PL-3

xayif-∅-o quick-M-SG

‘Blood makes them (bugs) swell, and they grow heavier, and as a result they are unable to move quickly’. (37) lo=qadǝṇṇo oxanno arbʿo ḥappoṯe maz=zaytunanǝk ‘Ich konnte aber nicht einmal vier dieser Oliven essen’. (LB 73) lo=qadǝṇ-∅-ṇo neg=can.PRET.INTR-M-1SG ḥapp-oṯe piece-PL

oxan-∅-no eat.INFECT-M-1SG

arbʿo four

m-az=zaytun-anǝk from-ART.PL-olive-those

‘I was unable to eat four of these olives’. (38) k-ŭmalle lu=abro: “qay damixǝt?” lo=qadǝr d-obe ğŭwāb ‘Er spricht zum Sohne: “Warum hast du geschlafen?” Er konnte keine Antwort geben’. (21/6) k-ŭmal-∅-le PRS-say.INFECT-3SG.M-to.3SG.M

l-u=abr-o to-ART.SG.M=son-SG

damix-ǝt sleep.PRET.INTR-2SG.M

lo=qadǝr-∅ NEG=can.PRET.INTR-3SG.M

d-obe-∅ that-give.INFECT-3SG.M

ğŭwāb answer[SG.M]

qay why

‘He speaks to his son: “Why have you fallen asleep?” He was unable to give an answer’. This Arabic loan is the basic verb for the concept ‘can, be able’ in the language. An alternative is the non-verbal, etymologically existential predication kibe ‘he can,’ kibi ‘I can,’ etc., and laybe ‘he cannot,’ laybi ‘I cannot,’ etc., see Jastrow (2002: 107). As is natural, the subject of the sentential complement is coreferential with that of the main clause. The sentential complement is introduced either by d- or asyndetically.

20

YULIA FURMAN AND SERGEY LOESOV

ṭmʿ ‘crave’: (39) w-uʿdo, ʿal d-ṭamiʿat ebi, d-mǝlli ono ḥoṯo elux, madām hawxa-yo, laybux howatlan mede ‘Und jetzt, wo du mich begehrst, wo ich dir doch gesagt habe, ich würde für dich Schwester sein, und da es nun so ist, kannst du uns gar nichts mehr sein’. (52/176) w-uʿdo and-now

ʿal d-ṭamiʿ-at because-of-crave.PRET.INTR-2SG.M

d-mǝl-li that-say.PRET-1SG madam since

ono I

ḥoṯ-o sister.F-SG

hawxa-yo so-COP.PRS.3SG

how-at-lan be.INFECT-2SG.M-to.1PL

eb-i in-POSSI.1SG el-ux to-POSSI.2SG.M

layb-ux PROH-POSSI.2SG.M

mede something[SG.M]

‘Now then, because you have craved for me, while I said “I am going to be your sister,”—since this is so, you are not allowed to be anything whatsoever for us’. According to a personal communication from Jan Beṯ-Şawoce, one can also say ‘w-uʿdo, ʿal d-ṭamiʿat-li,’ i.e., interpreting the Stimulus syntactically as Direct Object. Shabo Talay doubts if this is a good usage. We have found so far no examples of the Infectum in the corpus. lzm ‘require’: (40) húlele u=mede d-lazǝmle, i=xarğiyayḏe ‘Er gab ihm das Reisegeld und das, was er sonst brauchte’. (97/48) hu-le-le give.PRET-3SG.M-to.3SG.M

u=mede ART.SG.M=something

d-lazǝm-∅-le, REL-be_needed.PRET.INTR-3SG.M-to.3SG.M i=xarğiy-ayḏe ART.SG.F=travel_money-POSSII.3SG.M ‘[His father] gave him whatever was necessary for him, (including) the travel money’.

STUDIES IN THE ṬUROYO VERB

21

According to Ritter (1990: 206), the Preterite lazǝm is used only “unpersönlich,” as in (40), while the Infectum lozǝm appears both as an impersonal predicate (ko-lozǝm d- = ‘it is necessary that,’ and cf. Jastrow 2002:169) and as a finite verb inflected for person, e.g. ayna yawmo d-luzmat-li ‘if you need me one day’ (60/173), note the Direct Object in this finite verb phrase. To complicate matters, lazǝm is also used as a present-tense impersonal modal predicate ‘it is necessary that,’ i.e., it is synonymous to the impersonal ko-lozǝm. Ritter (1990: 207) terms this usage of lazǝm “arabisches Part.” Consider the following examples: (41) kolozǝm d-mafṭanno ‘Ich muß frühstücken’. (9/25) ko-lozǝm-∅ PRS-be_needed.INFECT-3SG.M

d-mafṭan-∅-no that-breakfast.INFECTIII-M-1SG

‘It is necessary that I have breakfast’. (42a) lazǝm ḥa minan d-foyǝš harke ‘Einer von uns muß hier bleiben’. (60/114) lazǝm-∅ ḥa be_needed.PRET.INTR-3SG.M one[M] d-foyǝš-∅ that-stay.INFECT-3SG.M

min-an from-POSSI.1PL

harke here

‘It is necessary that one of us stay here’. (42b) lo lazǝm ḥa marḥǝm aʿle ‘Man darf kein Mitleid mit ihm haben’. (69/574) lo=lazǝm-∅ NEG=be_needed.PRET.INTR-3SG.M

ḥa one[M]

marḥǝm-∅ pity.INFECTIII-3SG.M

aʿl-e on-POSSI.3SG.M ‘It is not necessary that anybody have pity on him’. The root lzm is of Arabic origin, as so many other roots of the language which nonetheless display, unlike lzm, trivial morphosyntactic behaviour. It is likely that, as Ritter hints, lazǝm in some of its usages corresponds to the impersonal participial lazǝm of Anatolian Arabic (for the evidence, see Kinderib 131; VW II 172). The personal Infectum is clearly transitive, as in ayna yawmo d-luzmat-li ‘if you need me one day,’ and in this usage the verb is Infectum tantum. The picture looks a little

22

YULIA FURMAN AND SERGEY LOESOV

bizarre. One may surmise that the derivatives of lzm had been borrowed from spoken Arabic more than once, with different outcomes. We shall now summarize our review of the evidence. The five verbs of Aramaic origin above are the basic exponents of the respective concepts. As for the five Arabic loans (fahəm ‘understand,’ ʿayǝz ‘need,’ qadǝr ‘can, be able,’ ṭamǝʿ ‘crave,’ and lazǝm ‘need, require’), only qadǝr is the basic exponent of a verbal notion in Ṭuroyo. Though fahǝm appears several dozens of times in the texts, the basic exponent of ‘understand’ is most probably the Aramaic root ʾḏʿ ‘know’. Lazǝm is not quite rare, but its severe paradigmatic restrictions make it the weakest member of the list. Nevertheless, its “modal” (propositional attitude) lexical meaning fits well into the group. How do we explain the emergence of these two-place nafǝq-shaped Preterites in the language? This must be a semantically conditioned innovation. We propose that, for example, šmʿ ‘hear,’ the verb that had been inherited by proto-Ṭuroyo from its Middle Eastern Aramaic ancestor, used to have the l-Preterite *šmǝʿle, only to forfeit it in favour of šamǝʿ. This is not only because of the famous šmʿ ly ‘I have heard’ (TADAE A6.10:3) and šmʿ ln ‘we had heard’ (TADAE A3.3:13) forms, which are probably due to an Eastern Aramaic adstrate in the Imperial Aramaic corpus, and not just because this periphrasis with šmʿ is known in JBA (Bar-Asher Siegal 2011) and Syriac (see šmʿ ln cited in Nöldeke 1966: 210, and additional examples of the same collocation šmīʿ lan in the Peshitta for Act 15: 24 and 19: 2, both corresponding to the active transitive aorist ἠκούσαμεν ‘we heard’ of the Greek original). The innovative nature of šamǝʿ follows also from the absence of the *šammīʿ adjective from the corpus of Classical Syriac. According to Barsky and Loesov (in preparation), in Syriac, predicative deverbal qattīl-adjectives were formed only for intransitive verbs, in particular for verbs of intransitive motion (see a sample of examples in Loesov 2013: 101ff.), and the same is true of other Middle Eastern Aramaic varieties. Thus the predicative qattīl, an innovative pattern of deverbal adjectives, had been becoming increasingly popular since the early days of Aramaic until the advent of Modern Aramaic. At a certain point of Eastern Aramaic evolution, the predicative *qattīl (inflected via suffixes that had developed from cliticized subject pronouns) became the default pasttime tense for intransitive G-stem verbs, most probably having passed the stages of RESULTATIVE and PERFECT (i.e., while the old Preterite *qatal was still alive). In protoṬuroyo, semantic attraction of this new intransitive Preterite was so strong that a number of two-place experiencer verbs changed their Preterite shape from nšəqle to nafəq. Moreover, Ṭuroyo used the nafəq pattern for adaptation of Arabic borrowings with appropriate meanings. Note that in Maʿlula, a modern Western Aramaic dialect, the verbs yḏʿ ‘know,’ šmʿ ‘hear,’ ḥmy ‘see,’ along with some other experiential transitive roots, have *qattīl-

STUDIES IN THE ṬUROYO VERB

23

rather than *qtīl-shaped RESULTATIVE, and this evidence corresponds to the situation in Ṭuroyo: in Maʿlula, transitive G-stem verbs normally have *qtīl-RESULTATIVE, while intransitive ones have *qattīl.2 In proto-Ṭuroyo, there may have been an additional reason for the above nšǝqle > nafǝq shift. “In einigen Fällen (LuF 76)”, nafǝq-shaped verbs are able to form detransitive shapes. Otto Jastrow (ibid.) mentions those of ʾḏʿ ‘know’ and ylf ‘learn’. Yet the verbal concepts such as ‘learn,’ ‘know,’ ‘want,’ ‘forget,’ ‘hear,’ ‘need,’ ‘be able,’ ‘desire’ probably have little need for passivization. Be this as it may, the only detransitive forms of these verbs we have found in the corpus are as follows: (43) a=kṯowe d-gkuṯwutu bu=lišono ṭuroyo, l-kulayna xaṭǝra d-lǝ=kmǝṭʿoyo hul lu=mawto gdowǝn ‘Die Bücher, die Ihr in der Turöyosprache schreiben werdet, werden für uns alle ein Andenken sein, das bis zum Tode nicht vergessen wird’. (11/331) a=kṯow-e ART.PL=book-PL

d-g-kuṯw-utu REL-FUT-write.INFECT-2P

ṭuroyo-∅ Ṭuroyo-SG

l-kul-ayna for-all-POSSI.1PL

b-u=lišon-o in-ART.SG.M=language-SG

xaṭǝra matter.SG.F

d-lǝ=k-mǝṭʿoy-o REL-NEG=PRE-be_forgotten.DETRANS.INFECT-3SG.F hul-l-u=mawt-o till-to-ART.SG.M=death-SG

gd-ow-ǝn FUT-be.INFECT-3PL

‘The books that you will write in the Ṭuroyo language will become for all of us a matter that will not be forgotten till death’. (44) u=qǝsǝm d-lawġul d-owe kayiwo w-ṣăqaṭ, lǝ=kmǝšmǝʿ aw kyuqro i=aḏno ‘Wenn der innere Teil krank und gelähmt wird, hört man nicht, oder das Ohr wird schwer(hörig)’. (3/44) u=qǝsǝm ART.SG.M=part

2

d-lawġul of-inside

d-owe-∅ if-be.INFECT-3SG.M

See Loesov (2012: 429, with reference to standard works on Maʿlula), where this remarkable isogloss was dealt with in relation to the Sapirian “common drift” in the evolution of Aramaic.

24

YULIA FURMAN AND SERGEY LOESOV kayiw-∅-o sore-M-SG

w-ṣăqaṭ-∅ and-corrupted-SG.M

lǝ=k-mǝšmǝʿ-∅ NEG=PRS-be_heard.DETRANS.INFECT-3SG.M k-yuqr-o PRS-become_hard.INFECT-3SG.F

aw or

i=aḏn-o ART.SG.F=ear-SG

‘If the inner part [of the ear] becomes sore and corrupted, one cannot hear, or develops poor hearing’. (45) kmiḏǝʿ b-ǝšme du=Qanda d-Daywān ‘Er war bekannt unter Namen Der Qanda von Deiwan’. (45/3) k-miḏǝʿ-∅ PRS-be_known.DETRANS.INFECT-3SG.M b-ǝšm-e-d-u=Qanda in-name[SG.M]-EZ-of-ART.SG.M=Qanda

d-Daywān of-Daywān

‘He is known by the name Qanda from Daywān’. (46) kmiḏoʿo d-kǝtyo walaye ʿatǝqto mu=săbab d-hēš kfoyǝš biya aṯarat w-baniyat rabe ‘Man erkennt, daß es eine alte Stadt ist, weil es darin noch Ruinen und grosse Gebäude gibt’. (11/157) k-miḏoʿ-o PRS-be_known.DETRANS.INFECT-3SG.F ʿatǝq-t-o old-F-SG

m-u=săbab from-ART.SG.M=reason

bi-ya in-POSSI.3SG.F

aṯar-at ruin-PL

d-kǝt-yo that-COP-COP.PRS.3SG d-hēš that-still

walaye town.SG.F

k-foyǝš-∅ PRS-stay.INFECT-3SG.M

w-baniy-at and-building-PL

rab-e big-PL

‘The town is known to be old because there still remain in it ruins and big buildings’. (47) ya u=malkayḏi kul mede hani am=medonani lo=kobaʿʿe d-ḥa ġắlabe mfakǝr appe. komiḏoʿi l-ruḥayye komibayni ‘Mein König, über alle diese Dinge braucht man nicht viel nachzudenken. Sie werden durch sich selber erkannt und klar’. (81/111)

STUDIES IN THE ṬUROYO VERB ya voc

u=malk-ayḏi ART.SG.M=king-POSSII.1SG

hani these

am=med-on-ani ART.PL=thing-PL-these

d-ḥa that-one[M]

ġắlabe much

kul all

25 mede thing[SG.M]

lo=k-obaʿ-∅-ʿe NEG=PRES-want.INFECT-3SG.M-3PL

mfakǝr-∅ think.INFECTII-3SG.M

ko-miḏoʿ-i-∅ PRS-be_known.DETRANS.INFECT-PL-3

ap-pe about-POSSI.3PL

l-ruḥ-ayye for-REFL-POSSI.3PL

ko-mibayn-i-∅ PRS-be_understood.DETRANSII.INFECT-PL-3 ‘Oh, my king! One does not have to think a lot about these things. They get known and understood by themselves’. (48) u=zlamano d-maqimatla čadǝre li=barṯayḏux mi=walaye w-larwal, gǝzzela li=čadǝre. gǝzzela li=čadǝre, u=yawmawo gǝmmiḏǝʿ ‘Wenn du außerhalb der Stadt für deine Tochter ein Zelt aufschlagen lässt, so wird der Mann zu dem Zelte kommen. An dem Tag wird er erkannt warden’. (101/16) u=zlam-ano ART.SG.M=man-this.M

d-maqim-at-la when-set_up.INFECTIII-2SG.M-to.3SG.F

čadǝre tent.SG.F

l-i=barṯ-ayḏux to-ART.SG.F=daugther-POSSII.2SG.M

w-larwal and-outside

g-ǝzz-e-la FUT-go.INFECT-3SG.M.-to.3SG.F

u=yawm-awo to-ART.SG.M=day-that.M

m-i=walaye from-ART.SG.F=town

l-i=čadǝre to-ART.SG.F=tent

gǝm-miḏǝʿ-∅ FUT-be_known.DETRANS.INFECT-3SG.M

‘When you set up a tent for your daughter outside the town, the man will come to the tent and be recognized on that day’. According to Ritter (1990: 727), there exists the detransitive Preterite iḏiʿ ʻer wurde gewusst, bekannt,’ the form (and the whole Preterite detransitive paradigm) having been elicited by Ritter from an informant (cf. also LuF 76). We have not found it in the corpus. Ritter (1990: 727) has even the detransitive Imperative paradigm miḏíʿ/miḏiʿu [no translation], which was also elicited. Shabo Talay tells us that iḏiʿ in the sense ‘es wurde bekannt’ does exist, while “nowadays many people use mtawḏǝʿ instead” (i.e., the village form of the III-stem Preterite detransitive).

26

YULIA FURMAN AND SERGEY LOESOV

Thus we have found in the corpus detransitive forms of three of the ten verbs: ṭʿy ‘forget,’ šmǝʿ ‘hear,’ ʾḏʿ ‘know’. The form k-milǝf ʻes wird gelerntʼ is represented only by elicited examples (Ritter 1990: 659). All six available tokens are Infectum forms. The form d-lǝ=k-mǝṭʿoyo (43) has a genuine passive reading (‘a matter that will not be forgotten’ by the speaker). The form lǝ=k-mǝšmǝʿ of (44) is impersonal (‘one cannot hear’). The four tokens of miḏǝʿ are different with regard to diathesis. The construction k-miḏǝʿ b-ǝšme ‘he was known (under a certain name)’ (45) is nonpassive and probably idiomatic, cf. German ‘wie heissen Sie?’, ǝmmi-le ‘they say to him’ = ‘they call him/his name is’ in Ṭuroyo (JL 5.10.6), ‘he is known as’ = ‘his name is’ in English, etc. The verb k-miḏoʿo (46) agrees in the feminine gender with the walaye ‘city’ of the subordinate clause and therefore has to have a passive meaning: the city can be recognized as old by an observer because of its visible peculiarities. In (47), ko-miḏoʿi l-ruḥ-ayye ‘they get known by themselves’ must be nonpassive due to the use of the reflexive pronoun in the construction. In the co-text of (48), gǝm-miḏǝʿ means ‘he will be recognized (by those who will spot him),’ it is a clear-cut PASSIVE. Thus some of these verbs do have detransitive forms with passive readings, and this fact may constitute another piece of evidence that their Preterite had once had the nšǝqle rather than the nafǝq shape. The assumed (pragmatically conditioned) paucity of their detransitive tokens may have been an additional reason for the shift. Before closing this paper on two-place verbs with nafǝq-shaped Preterites, we have to mention some roots that keep their l-Preterite though they belong to the semantic group of perception/propositional attitude. These are in particular bʿy (bʿele) ‘want, need,’ which must be etymologically related to ʾbʿ ‘want’ discussed above, and rḥm (rḥǝmle) ‘love’. Why the verbs ʾbʿ and bʿy have got different Preterite shapes is unclear.

REFERENCES Abbreviations and sources Anat. = Anatolian Arabic AWSG = Wehr 1985 JL = Jastrow 2002 Kinderib = Jastrow 2005 LB = Tayal 2004 LuF = Jastrow 1967 MM = Jastrow 1968 PrS = Prym-Socin 1881 Ritter Texte Bd.1 = Ritter 1967

STUDIES IN THE ṬUROYO VERB

27

Ritter Texte Bd.2 = Ritter 1969 Ritter Texte Bd.3 = Ritter 1971 SL = Sokoloff 2009 TADE = Porten-Yardeni 1986-1999 VWII = Vocke-Waldener 1982

Bibliography Beṯ-Şawoce, Jan. 2012. Xëzne d xabre Ordlista: Şurayt-Swedi [mëḏyoyo]. Stockholm: Författares Bokmaskin. Bar-Asher Siegal, Elizur A. 2011. “On the Passiveness of One Pattern in Jewish Babylonian Aramaic—a Linguistic and Philological Discussion.” Journal of Semitic Studies 56: 111–143. Barsky, Evgeny, and Loesov, Sergey. in preparation. “Predicative Adjectives of Syriac in the Neo-Aramaic Retrospective.” Bezold, Veronika. 2012. Überarbeitung und Digitalisierung des Ṭuroyo-Wörterbuchs von Hellmut Ritter. Master’s Dissertation. Philosophische Fakultät II, FriedrichAlexander University. Nürnberg - Erlangen. Brockelmann, Carl. 1962. Syrische Grammatik, mit Paradigmen, Literatur, Chrestomathie und Glossar. Leipzig: Enzyklopädie. Hopkins, Simon. 1989. “Neo-Aramaic Dialects and the Formation of the Preterite.” Journal of Semitic Studies 34 (2): 413–432. Jastrow, Otto. 1994. “Erlebnisse eines Lastwagenfahrers. Ein neuer Ṭuroyo-Text im Dialekt von Midən.” In Festschrift Ewald Wagner zum 65. Geburtstag. Band 1 Semitische Studien unter besonderer Berücksichtigung der Südsemitistik, edited by Wolfhart Heinrichs and Gregor Schoeler, 221–233. Beiruter Texte und Studien 54. Beirut: Steiner (in Kommission). —. 1993. Laut- und Formenlehre des neuaramäischen Dialekts von Mīdin im Ṭūr ʿAbdīn. Semitica Viva 9. Wiesbaden: Harrassowitz. —. 1968. “Ein Märchen im neuaramäischen Dialekt von Mīḏin (Ṭūr ʿAbdīn).” Zeitschrift der Deutschen Morgenländischen Gesellschaft 118: 29–61. —. 1996. “Passive Formation in Turoyo and Mlaḥsô.” Israel Oriental Studies 16: 49– 57. —. 2002. Lehrbuch der Ṭuroyo-Sprache. Wiesbaden: Harrassowitz. —. 2005. Glossar zu Kinderib (Anatolisches Arabisch). Wiesbaden: Harrassowitz. Khan, Geoffrey. 2007. “Ergativity in the North Eastern Neo-Aramaic Dialects.” In Studies in Semitic and General Linguistics in Honor of Gideon Goldenberg, edited by Eran Cohen and Tali Bar, 147–157. Münster: Ugarit. Loesov, Sergey. 2012. “A New Attempt at Reconstructing Proto-Aramaic (Part I).” Babel und Bibel 6: 421–456.

28

YULIA FURMAN AND SERGEY LOESOV

—. 2013, “A New Attempt at Reconstructing Proto-Aramaic (Part II).” In Proceedings of the 14th Italian Meeting of Afroasiatic Linguistics, edited by Alessandro Mengozzi and Mauro Tosco, 91–106. Nöldeke, Theodor. 1966. Kurzgefasste syrische Grammatik. (Reprint). Darmstadt: Wissenschaftliche Buchgesellschaft. Porten, Bezalel, and Yardeni, Ada. 1986–1999. Textbook of Aramaic Documents from Ancient Egypt, 4 vols. Jerusalem. Prym, Eugene, and Socin, Albert. 1881. Der neu-aramaeische Dialekt des Ṭûr ‘Abdîn. Erster Teil. Die Texte. Göttingen: Vandenhoeck & Ruprecht’s. Ritter, Helmut. 1969. Ṭūrōyo. Die Volkssprache der syrischen Christen des Ṭūr ‫ع‬Abdîn. Texte Band II. Beirut: Steiner (in Kommission). —. 1979. Ṭūrōyo. Die Volkssprache der syrischen Christen des Ṭūr ‫ع‬Abdîn. Wörterbuch. Beirut: Steiner (in Kommission). —. 1990. Ṭūrōyo. Die Volkssprache der syrischen Christen des Ṭūr ‫ع‬Abdîn. Grammatik: Pronomen, „sein, vorhanden sein“, Zahlwort, Verbum. Stuttgart: Steiner. Siegel, Adolf. 1923. Laut- und Formenlehre des neuaramäischen Dialekts des Tûr Abdîn. Hannover: Heinz Lafaire. Sokoloff, Michael. 2009. A Syriac Lexcion: A Translation from the Latin, Correction, Expansion, and Update of C. Brockelmann’s Lexicon Syriacum. Winona Lake: Eisenbrauns-Piscataway, NJ: Gorgias Press. Talay, Shabo. 2004. Lebendig begraben. Die Entführung des syrisch-orthodoxen Priesters Melki Tok von Midən in der Südosttürkei. Einführung, Aramäischer Text (Turoyo), Übersetzung und Glossar. Studien zur Orientalischen Kirchengeschichte. Münster: Lit. Vocke, Sybille, and Waldner, Wolfram. 1982. Der Wortschatz des Anatolischen Arabisch. MA Dissertation. Philosophische Fakultät II (Sprach- und Literaturwissenschaften), Friedrich-Alexander Universitӓt. Nürnberg-Erlangen. Wehr, Hans. 1985. Arabisches Wörterbuch für die Schriftsprache der Gegenwart: Arabisch-Deutsch (5th ed.). Wiesbaden: Harrassowitz.

TOWARDS A DESCRIPTION OF WRITTEN

ṢURAYT/ṬUROYO: SOME SYNTACTIC FUNCTIONS OF THE PARTICLE KAL1

MACIEJ TOMAL This paper will deal with some syntactic phenomena of a literary variety of the NeoAramaic dialect called Ṭuroyo.2 Recently another term, Ṣurayt, has been coined to refer to both the spoken and literary variety of this dialect. It would, in fact, be more appropriate to refer to Ṣurayt/Ṭuroyo as a language rather than a dialect. The most conspicuous feature of the written texts that are the object of this paper is that they are written in Latin script. Since the 16th century there have been various attempts to commit NeoAramaic dialects to writing. These have resulted mostly in texts with cultural relevance and importance for the religious communities of the Neo-Aramaic speakers. Some varieties of Neo-Aramaic, therefore, have been written down in the form of literary texts. We may mention here the Jewish Neo-Aramaic adaptations or translations of Biblical texts, Biblical commentaries, Midrashim and homilies (Sabar 1983), and also written versions of hagiographic stories in Christian dialects, such as books of Jean Bedjan (1912), or folktales (cf. Lidzbarski 1896). An attempt to write down a vernacular version of religious texts was undertaken by a Nestorian priest who 1

I would like to express my gratitude to Jan Beṯ-Ṣawoce for his efforts to propagate the heritage of Ṣurayt/Ṭuroyo texts and his kind support in preparing the present text. 2 In this paper we use for citations the Latin transliteration employed in numerous publications under the auspices of Yusuf Ishaq and subsequently modified by Jan Beṯ-Ṣawoce. In the system of transcription adopted by them, ş stands for ṣ, ţ stands for ṭ, h stands for ḥ, ğ stands for ġ, ë stands for ə, x stands for h and c stands for ʿ. For the purpose of the present article, I adopted the standard transcription except citations of the sources.

30

MACIEJ TOMAL

wished to preserve or offer to the local community different texts relevant to their religious life, as is testified by David Stoddard (1855: 3): The first attempt worthy of record to reduce the Modern Syriac to writing was made by Rev. Justin Perkins, a Missionary of the American Board of Commissioners for Foreign Missions, at Tabreez, in the winter of 1834–5, in connection with the study of the language, under the instruction of the Nestorian Bishop Mar Yohannan. The first attempt to write it in a permanent and useful form, was made by Dr. Perkins in the construction of school-cards, in the winter 1836, after he and Dr. Grant had settled in Oormiah.

The supra-dialectal koiné based on Christian Urmi speech—Literary Urmian Aramaic—became widely used in Christian Neo-Aramaic-speaking communities.3 There was an especially productive development in the first half of the 20th century in the Assyrian community of the former USSR, which resulted in a wide range of written publications, including newspapers, educational material and literary texts. This was in fact an attempt at the standardization of the language. A new Latin script, the so called Novyj Alfavit, was created and used for the purpose of publishing these texts.4 Numerous texts published in this initiative were not connected with the cultural heritage of the speakers but rather were translations of contemporary Russian literature, including, for example, Maxim Gorki, Lev Tolstoy and Alexander Pushkin. The language of the Christian communities originated in Ṭurʿabdin known as Ṭuroyo5 and subsequently as Ṣurayt has survived mostly in diaspora communities in Sweden, Germany and Holland. The language of Ṭurʿabdin is in fact a cluster of dialects from the so-called central group of Neo-Aramaic, which were spoken in southern Turkey in the province Midyat. Since 1980s this Neo-Aramaic dialect cluster has been undergoing the process of standardization among the diaspora communities in European countries especially through the production of written texts. The importance of the process was pointed out by Otto Jastrow in his introduction to a new 3

An exemplary work describing the standardization of the Neo-Aramaic dialect of Urmia is of H.L. Murre-van den Berg (1999; the author discusses also the texts of J. Bedjan). 4 “Re-edition of Assyrian texts in the Latin script, published in the Soviet Union (mainly in Moscow) in the 1930’s in roman characters, has recently been undertaken, by a well-known Semitist Johannes Friedrich (…) [T]hat author gives two short texts (extracts from two stories: L. Tolstoy, After the Ball, and Pushkin’s, The Stationmaster)” ”, Tsereteli (1965: xvi; see Polotzky (1961: 1-35)). 5 The term introduced to the field of Neo-Aramaic studies especially due to a monographic work of Helmut Ritter (1967–1971).

TOWARDS A DESCRIPTION OF ṢURAYT/ṬUROYO

31

Swedish-Ṣurayt dictionary, compiled by Jan Beṯ-Ṣawoce: “Today I am very happy to say that despite many obstacles, the written form of Ṣurayt is gradually gaining ground” (Beṯ-Ṣawoce 2012: x). It should be noted that this was achieved due to the initiatives of staff at the National Swedish Institute for teaching Material with Dr. Yusuf Ishaq acting as scientific director. This initiative produced books in standardized Ṣurayt, including educational material Toxu qorena, as well as folktales and stories. The subsequent development of the standard written language was due to a large extent to the wide-ranging activities of Jan Beṯ-Ṣawoce. In the present paper, I would like to discuss one syntactic phenomenon of written or literary Ṣurayt/Ṭuroyo. In Ṣurayt/Ṭuroyo, as in many other Neo-Aramaic dialects, there are three sets of particles that fill the slot of the copula in nominal clauses.6 These are usually referred to as enclitic copula, independent copula and emphatic independent copula. The non-enclitic copula, therefore, is expressed by two sets of particles, frequently with two different stems. Some syntactical functions of the two independent particles will be discussed here. The difference in function between these two copulas has still not been exhaustively analysed. I shall here make an attempt to examine this aspect of literary Ṣurayt/Ṭuroyo syntax, by looking for regularities of usage in a corpus of texts published in the 1980s and 90s. One may expect a standardization of the usage of the forms under discussion in written texts or at least an attempt at standardization.7 The existential particle in Ṣurayt/Ṭuroyo usually is based on the ʾit element and has the form kit. This may be inflected with a pronominal element and, therefore, some scholars call the form a pseudo-verb (Jastrow 1985: 119–122). As far as the syntactic function is concerned, this form performs the function of copula. In Ṣurayt/Ṭuroyo, side by side with the copula kit, there is also a “presentative”/emphatic form of the copula, which has the base kal:

6

këtyo këtyo këtne ~ kënne

kale ~ kaleyo kala ~ kalayo kalan ~ kalanne

këlle ~ këlleyo këlla ~ këllayo këllën ~ këllënyo

këthët ~ këtat këthat këthatu ~ këtatu

kalox kalëx kaloxu

këllox këllëx këloxu

For the general study of copulas in nominal clauses see Frajzyngier et al. (2002). A description of the functions of the kal particle, especially within the tense system, has been attempted in Tomal (2008–2009). 7

32

MACIEJ TOMAL këtno ~ kënno këtna ~ kënna

kali ~ kalënno kalan ~ kalënna

këlli ~ këllino këllen ~ këllënne

(Beṯ-Ṣawoce 2008: 86, 133–134) An important difference between the two types of particles is that kit has a form of past tense with the element -wo > këtwo. Syntactically, the past tense copula appears in the same type of clauses as its present tense equivalent. On the other hand, only the present tense form of the kal form exists. Moreover, although this copula is inflected and functions independently, it may also have the role of a verb modifier, which is not the case with the kit form. Therefore, it is questionable whether the notion “copula” may be applied to the syntactic functions of the kal form or not. It may be more appropriated to refer to it as “particle”. For convenience, however, I propose to use the terms copula and emphatic copula for the kit and kal forms respectively when they are used in non-verbal clauses. As has been mentioned already, many NENA dialects use an independent copula side by side with an enclitic one. The NENA dialects, moreover, often an independent copula that has an emphatic function. The emphatic copula has been shown to have a distinct communicative function. In the Christian Qaraqosh dialect, for example, clauses with the emphatic copula express contingent rather than permanent situations (Khan 2002: 128). While clauses with the enclitic copula may express permanent features or states, those with the emphatic copula express rather a contingent state or property. The difference between the two clause types, therefore, is connected with the semantics of the predicate. Thus there are morphological, as well as syntactic and semantic differences between the two copulas. However, we also find that the functions of both copulas may overlap. In the literary variety of Ṣurayt/Ṭuroyo the forms kit and kal appear as copulas in all types of nominal clauses, i.e. with three types of nominal predicates— adjective, nominal and locative. In an attempt to find distinctive features in the usage of the two copulas, we shall analyse their functions in all three types of nonverbal clauses. We shall, then, attempt to identify the motivations for using the kal particle usage as a verb modifier.

1. PRELIMINARY SYNTACTICAL OBSERVATIONS AS TO THE FUNCTIONS OF COPULA KIT IN ṢURAYT

The functions of the kit copula frequently overlap with those of the enclitic copula -yo. The difference between the two forms, i.e. kit and –yo, lies in their syntactic position. While the former is independent, the latter is attached to the predicate as

TOWARDS A DESCRIPTION OF ṢURAYT/ṬUROYO

33

an enclitic element. Both may be used in all types of clauses with non-verbal predicates.8 The kit form performs the function of copula—or fills the slot of copula—in non-verbal clauses. It occurs in clauses with adjective, nominal and locative predicates. The second category of clauses includes also equative ones, while the third category overlaps with that of existential clauses. An example of the kit copula occurring in a clause with a locative predicate is as follows: (1) U and cal on

Fawlus Fawlus I

ART

barëm slip.PST.3SG

cal aẖ on ART

ẖeţe wheat

d

REL

kit

COP.PRS

adro. threshing_floor

‘And Fawlus slipped on the wheat that is on the threshing floor’. (Bëğël, 21) The state referred to here is non-permanent. In (2), the kit copula occurs in a clause with an adjectival predicate. The property expressed by the predicate is a permanent one: (2) [ẖeţe d] kët-wayye9 [wheat REL] COP-PST.3PL

kayës, good

nadëf pure

w and

šafër fine

‘[wheat that] was good, pure and fine’. (Bëğël, 21) In (3), in which predicate expresses a material, the property is also permanent:

8

It may be argued that the enclitic particle tends to be used in equative clauses rather than the kit one. We may add here that in Ṭuroyo in the equative clauses the copula slot is fulfilled mostly by enclitic -yo and its counterparts, see Frajzyngier et al. (2002). Accordingly, the enclitic particle -yo may express a permanent feature. Consequently, even if it may perform the function of copula in locative clauses, it may refer rather to a permanent than contingent localisation: I Kaše du Cafso b falgë du darbo yo, ‘Kašo du Cafso is in half of the way’ (Bëğël, 25). When the predicate is temporal the particle yo is preferred. What is said above may be true only with regard to clauses with locatives as a predicates (maybe also adjectives). The contingency or permanence of the state referred to may be the motivation for the choice of one of the two particles in nominal clauses, as is the case in Neo-Aramaic dialect of Qaraqosh, for instance (Khan: 2002). It seems, however, not to be the case in Ṭuroyo. 9 In the present text the hyphen is introduced for the sake of consistency in Ṭuroyo glossing with other texts in this volume. In the cited original texts the hyphen is not used at all.

34

MACIEJ TOMAL

(3) U and

babo father

qamẖo flour

mër-le say.PST-3SG d

d

REL

ketyo COP.PRS.3SG

ẖeţe wheat

REL

‘And the father said that there is flour [made] of wheat’. (Bëğël, 24) In general clauses with the kit copula and adjectival predicate express a permanent state rather than a contingent one. The copula may be sometimes fronted, especially in dialogue: (4) Laẖdo, Lahado

b for

did-ox POSS-you

Aloho God

cal on

i

ART

kito COP.3SG

ẖeţe wheat

šafër nice

m than

adro treshing_floor

‘Laẖdo, for God’s sake, the wheat is nicer than yours on the threshing floor’. (Bëğël, 21) The kit copula may occur in purely existential clauses. (5) Kët-wo COP-PST.3SG

ğalabe plenty

ẖeţe wheat

‘There was plenty of wheat [there]’ (Bëğël, 21) Note that the state referred to here is likewise permanent. In all the aforementioned cases, i.e. (1)–(3) clauses, the copula is the theme or—more precisely—may be regarded as the textual theme, i.e. a non-referential element that plays a role of signalling relations within a text (cf. Ping 2004: 82, see also idem 2001; Forey 2002: 73).10 Therefore, the function of the kit element seems to be twofold: on the one hand it fills the slot of copula on the level of the clause, and on the other hand, on the level of the text it plays a role of a linkage. The most frequent usage of the copula kit is in the clauses with a locative predicate. The situation expressed by the predicate may be permanent: (6) E, yes 10

lat-wo COP.NEG-PST.3SG

gubone wells

d

REL

maye water

qariwe lan close to.ART

The problem is discussed whether lexemes, as ‘there is’, may be treated as a theme, since they are referentially empty. Among the proposals for how to resolve the problem is that of Ping (2004) and Foyer (2001); Ping proposes to treat it as a “Textual Theme”.

TOWARDS A DESCRIPTION OF ṢURAYT/ṬUROYO adroṯo, bale floor but

cam haṯe with this

këtne COP.PRS.3PL

šënoṯo Stones

adro earth

basëmtër better

kula all

rabe great mu than.ART

an

ART

adroṯo threshing_floor

35 mëd for.REL

w ftiẖe, u šloqo d and flat and cooking REL šloqo cooking

du

REL.ART

cal i on ART

bayto kowe house possible

‘Indeed, there were not wells close to the threshing floor, but on the surface of this entire threshing floor as there are big blocks of rocks, and cooking on the threshing floor is better than in the house’. (Bërğël, 33) In the narrative in (6), the kit clause introduces a setting of the narrative and appears in a chain of coordinate clauses. The clauses with locative predicates very frequently supply background or incidental information in relative clauses: (7) Aẖ

ART

ẖmore donkeys

kito COP.PRS

cal on

baţëli-wo free.PL-COP.PST

mu from.ART

ţacno load

d

REL

ẖaş-ayye necks-their

‘The donkeys were freed from the load that was [lit. is] on their necks’. (Bëğël, 27) (8) I

ART

naqqoyo now

yarixo hul long till

d

REL

gëd

FUT

qëdarwo can.3SG.CONV

mëţaniwo come.3PL.CONV

fašwo zabno remain.3SG.CONV time hani d they REL

këtne laxalf COP.3PL behind

‘Now, a long time could remain till they that were behind could come’. (Bëğël, 27) (9) U and

Musa Musa

maẖak-le throw.PST-3SG.M

ẖër-le see.PST-3SG.M

bu in.ART

ak

d

ART

kefe stone

REL

babo door këtwo COP.PST.3SG

mnakëf feared b in

w and iḏe hand

‘Musa looked at the father with fear and threw the stone that was in his hand’. (Bëğël, 29) The state referred to in these relative clauses is usually contingent. In general, it may be said that nominal clauses with the kit copula and locative predicate usual-

36

MACIEJ TOMAL

ly express a contingent state supplying background information when the clause is a relative one. At the same time, the kit clause with a locative predicate may be coordinated in a chain within a setting and in such cases it may express a permanent situation. The syntactic position of the copula is different and, consequently, the information content differs. While in the case of locative and adjectival predicates the kit particle plays the role of a textual theme, it may be placed in the rhematic position when the pronoun is extraposed: (10) Aẖno we

bu in.ART

aṯrano place.this

diḏ-an POSS-we

kito COP

ağawiye lords

‘We, in our place, there are the lords’. (Bëğël, 27)

In equative clauses, the position of the copula may be occupied by the particle yo.

In our corpus clauses with the nominal predicates occur only when someone’s age is presented: (11) Nacimo child

d

REL

ketyo COP.3SG.M

tmënyo eight

yarhe months

‘A child who is eight months old’.(Bërël, 49) In general, clauses with adjective predicates express a non-contingent state and supply background information. Locative predicates generally express contingent states. If the locative predicate occurs in a relative clause, it generally supplies background information. If, however, the clause is independent the situation expressed is often permanent, especially within of the preliminary setting of a narrative. Finally, in the case of clauses with nominal predicates, the usage of the copula is often associated with strong assertion or contrastive focus.

2. KAL AS A COPULA IN CLAUSES WITH A NON-VERBAL PREDICATE

The most conspicuous feature of the kal copula is that it never occurs in relative clauses. It does not, therefore, introduce background information. Rather, this copula appears in independent clauses, usually constituting elements of a preliminary narrative setting and introducing new referents in the narrative, as in the following personal account: (12) Yawmo day

këlli

COP.PRS.1SG

bu in.ART

karmo. vineyard

Šamceno ẖës. hear.PST.1SG sound

TOWARDS A DESCRIPTION OF ṢURAYT/ṬUROYO Mşanat-li kayiso, këlle make.PST-1SG well COP.3SG

ẖës sound

d

REL

37

dënnaga drum

w d and

REL

zërnaye. U ẖës mëqqayës d ëṯewo mqarawwo oboe. And sound from_side REL come.3SG.CONV approach.3SG.CONV lafeli. Hawwël to_me. Behold

naqqa ẖzewo-li bahro d suddenly see.PST.CONV-1SG light REL

bi in.ART

diḏ-an POSS-we

mcarṯayo cave_that

d

REL

qadëẖ fire

këtyo baq COP.PRS.3SG in.ART

qalačat dak karm-eyḏan. U ẖës de dënnaga uncultivated_area REL_ART vineyard-our and voice REL drum w and

zërniye oboe

latyo COP.NEG

ẖešsabo Sunday

dënnaga drum

w and

kayiso. B well in

ẖdo one

Këlla ẖḏo COP.PRS.3SG.F one Këmmo say.3SG.F

ëṯewo come.3SG.CONV

mlaf i mcrṯo. Aloh from_direction ART cave god

d

mëštuṯo. feast

REL

zërnaye-yo oboe-COP

dënnaga drum

mën šëfro what beauty

u

ART

kito

COP.PRS

nẖat descend

Këlla këṯyo AUX stand.3SG.F

acma with_her

Haṯe this

mën from

mër-li. Huw-li aḏni say.PST-1SG. give.PST-1SG ear

darbo šamëno way hear.PST.1SG

lu yawmo for.ART day

Hawxa šaferto. so beautiful

howe be.3SG.M

gëd

FUT

ẖës voice

sid-i. by-me

acla mën šëfro. on.her what beauty yësqono go_up.1SG

l dëkṯ-ox. to place-your

lafel-i. Hës di in_front-me voice REL.ART

freq-li notice.PST-1SG

d

REL

ketyo COP.3SG

mdëštaw sickness

mën-[c]ayn-a from-eyes-her ‘One day I was in the vineyard. I heard a sound. I gave an ear—it was the sound of drum and oboe. And the sound is approaching me. Suddenly, I saw a light of fire in our cave that is on the uncultivated area of our vineyard. And the sound is approaching from the cave. God! There is no Sunday to hold a feast. This is like drum and oboe, I said. I gave an ear. At

38

MACIEJ TOMAL some point I heard a voice by me. There is a beauty; there is a beauty there. She says: when the sun sets I will ascend to you. How beautiful she is. She is standing before me. And the sound of drum is with her [accompanies her]. I noticed there is a sort of sickness’. (Bëğël, 64)

What is interesting in the passage above is that in all cases where a new referent appears, the kal copula is preferred—‘This is a sound of drum and oboe’. Additionally, the appearance of the referent is unexpected. In another case, strong assertion is expressed: ‘How beautiful she is!’. The kit copula, on the other hand, occurs in the dependent clauses. What is striking is the equivalence between copula kal and the existential particle kito, a non-inflected variant of the copula kit. When a clause with a new referent is repeated, the copula kal of the previous statement is replaced by the kito particle—‘There is a beauty; there is a beauty there’. When the element of unexpected appearance is not present, the copula is changed from an emphatic to a non-emphatic one. The distribution of the copula -yo is more variable—it occurs in main, as well as in relative clauses. In clauses that are coordinated with verba sentiendi the copula kal is preferred. Generally speaking, we may conclude that kal tends to fill the slot of the copula in independent clauses, when a new referent is introduced and when its appearance is unexpected by the participant in the situation. The particle kal fulfils the niche of the copula in all types of nominal clauses, i.e. clauses with the three types of predicates—adjective, locative and noun. As far as the state referred to is concerned, clauses with the copula of the kit type tend to express a permanent state or non-contingent state with adjectival predicate, as is the case of equative clauses, while with locative predicate the state is contingent rather than permanent. But the category of the state referred to seems not to be sufficient. Thus we must look for some additional conditions which motivate the usage of kal rather than kit. We will now look more closely at clauses with adjective and locative predicates. The kal particle is usually a textual theme in these types of clauses. Therefore, there does not seem to be any syntactic or informative difference between clauses with the two copulas kit and kal. The question arises: what is the textual significance of kal? The usage of the kal particle in the function of copula in clauses with adjectival predicates may to be motivated by a tendency for it to express unexpected events

TOWARDS A DESCRIPTION OF ṢURAYT/ṬUROYO

39

with a negative connotation. Moreover, the clause is coordinated with a verbum sentiendi:11 (13) ẖdo one

aṯto wife

quryayto of_deacon

šqi-la bring.PST-3SG.F dë

REL

bece. eggs

gab to

u

k-ẖayro PRS-see.3SG.F

baqono shepard

ART

Azz-a lu bayto come.PST-3SG.F to.ART home

mbašla-lle, cook.SG.F-them

man-ne from_them

azza go.PST.3SG.F

k-ubco PRS-want.3SG.F

këlën COP.PRS.3PL

hën some

xalye empty

‘A wife of a deacon went to a shepherd and bought eggs. She went home [and] wants to boil them and sees—some of them are empty’. (Pësseqat Ṣarëke, 27) The state referred to is contingent; moreover, it is contrary to the expectations of the participant of the situation. Therefore, unlike the usage of the kit copula in this type of non-verbal clause, the usage of kal seems to be connected with strong assertion. As far as the informative structure of the clause is concerned, the kal copula is a textual theme, as is the case of the kit, expressing an adversative clause—‘but some of them are empty!’ The kal particle plays the role of the copula in locative clauses. In fact, the occurrence of the kal copula in that type of clause is the most frequent among all occurrences of this copula. The state or feature described in the clause is generally a contingent one. Frequently it is contingent in a sense that it is perceived as unexpected by a participant of the situation. The clause with a locative predicate is usually combined with a verbum sentiendi that expresses an act of perceiving by the participant of the situation: (14) Aṯy-o go.PST-3SG.F Qarca head

11

taẖto down

b in

gawe inside w and

du

REL.ARt

rağloṯa legs

lacël. up

sanduqo bin

du

REL.ART

Aṯi come.PST.3SG

zawlo. rubbish ẖa one

dë

REL

Formally, the relation between two clauses is that of dependence, but the lack of connection brings the relation closer to that of independent speech, i.e. “practical” coordination.

40

MACIEJ TOMAL mhalëq throw.PST.3SG.M du

k-ẖoyër PRS-see.3SG.M

sanduqo. ẖër-le bin see.PST.3SG.M

REL.ART

këlla COP.3SG.F biyya, on_her

ẖḏo one

b in

gawe inside

mcayn-o-le check.PST-OBJ.3SG.M-3SG.M

kayiso well ‘It went into the rubbish bin. It is upside-down. One came to throw [something] away and sees: something is in the bin. He looked at it, he checked it well’. (Pësseqat Ṣarëke, 14) (15) Cas possibly cël over

lo

ẖze-lën see.PST-3PL

qarc-i. head-my

U and

su at.ART

taxt makeshift_bed

NEG

m from

këlle COP.3PL

at

tfënak rifles

ART

taxt makeshift_bed

biḏ-ayye in_hands-their dëṯṯe POSS.their

diḏ-i POSS-I

‘They didn’t see that their rifles in their hands were over my head. And their makeshift bed was [lit. is] by that of mine’. (Bërǧël, 64) The state is contingent for the participant of the situation, the witness. In (14) the copula kal plays a role of an interjection ‘Behold!’ or ‘Oh!’ In (15) the subject is in the thematic position, while kal constitutes a part of the rheme. In general, therefore, greater immediacy is gained through the usage of the kal copula. Moreover, kal as a copula in the existential clause appears in exclamatory clauses in dialogue paragraphs: (16) A

ART

ẖarmaye thieves

omar: “[…] say.3SG.M. A

ART

cabr-i l bayto. come.PST-3PL to house

More owner

[G]ëd

a

FUT

oman-naxu say.1SG.M-you

ẖaramiye mër-re thieves say.PST-3PL

Mar say.IMP.SG

ẖozina see.1PL

omar: say.3PL.M

“Këllëne COP.3PL

ART

“Cafarëm! Mën bravo what ẖayko-ne?” where.COP.3PL

bi in.ART

banqa! bank

du

bayto house

REL.ART

zuze hayko-ne.” money where.COP3.PL xayifo fast

More Owner

mër-le!? say.PST-3SG.M du

REL.ART

Këllëne bi COP.3PL.M in.ART

bayto house banqa!” bank

TOWARDS A DESCRIPTION OF ṢURAYT/ṬUROYO

41

‘The thieves came to the house. The owner of the house [the landlord] said: … I will tell you where the money is. The thieves said: Great! So quick he revealed [where money is]! Say and we will see where it is? The landlord said: It is in the bank! It is in the bank!’ (Pësseqat Ṣareke, 9) The thematized copula signals a contradiction or contrast within the paragraph. Therefore, the relevant clauses should be rendered English adversative clauses: ‘But it is in the bank!’ An extreme example of exclamatory function is found a dialogue containing solely exclamatory clauses expressing location: (17) Qumuuu rise_up.IMP.PL

laẖ for.ART

Msaaaku-nnee! wait.IMP.PL-for_them Këllënnn COP.3PL

ẖaramiyeeeee! thieves

Quuuumuuu rise_up.IMP.PL

Quumu-nneeee! rise_up.IMP-for_them

laẖ ẖaramiyeee! […] for.ART thievs

këllënnn! COP.3PL

Ayko-neee! where-COP.3PL Këllën COP.3PL Musa Musa

taẖt under aydarbo-hat how.COP.2SG

Toxuu! … come.IMP.PL

Këllën

Kë li li li liiii COP.1SG.

Kë li li liiiiii COP.1SG

AUX

u

ART

şaqfo. […] rock

abr-i? brother-my k-mahzëmiii!12… PRS-escape.3PL

‘Rise up against the thieves! Rise up against them! Wait for them! Rise up against the thieves! They are here, they are here! […] Where are they? They are under the rock. […]

12

This specific usage will be discussed later on in the section on the usage of the kal particle as a verb modifier.

42

MACIEJ TOMAL Musa, where are you? Come! They are escaping! I am here …’ (Bërǧël, 69-72)

In this particular story the peak-episode is marked by the expressive forms of the kal copulas. As has been already pointed out above, the additional condition motivating the usage of kal in the function of copula in clauses with locative predicate is coordination with a verbum sentiendi. This construction is intended to express the perception of the situation by a participant of the situation. The verb ‘to see’ may, indeed, appear in the imperative form within of a dialogue: (18) I

ART

sqëfto du darbo ayko dëkṯo-yo? plain REL.ART threshing where place.COP

ëmmi: say.3PL

Qay well

lë

ẖzëlat-la, see.2SG.M-her

NEG

diḏ-ox […] POSS-you

I

ẖzay! look.IMP.SG

Këlla COP.3SG.F

ART

sqëfto plain

du

REL.ART

Mëral-le say.PST.3PL-him

këlla COP.3SG.F

mqabël in_fornt

darbo threshing

ayko-yo? […] where.COP

maqbël in_front

‘Where is the place of the threshold? They said: Can’t you see it? It is before you […] Where is the threshold? […] Look! It is before you!’ (Pësseqat, 25) (19) Zox go.IMP.SG Ştayfo! Ştayfo

k-ëmarno-lox! PRS-say.1SG-to_you Azza w go.PST.3SG.F and

Zox go.IMP.SG

ţacyo forget

lë

NEG

nëfqo go_out.PST.3SG.F

mi from.ART

Mëd when

nafiko go_out.PST.3SG.M

lë

NEG

këlla COP.3SG.F

ruẖa herself

tamo. there

dërto courtyard

du

k-ëḏco PRS-know.3SG.F

REL.ART

sbë in_house Ţër keep bayto. house

dëciro come_back.3SG.F

‘Go! I am telling you! Go! She is in the house of Ştayfo! She went there and lost her way [she has forgotten her way back] there. She can’t go out from the courtyard. When she goes out, she is not able to come back’. (Bërǧël, 12)

TOWARDS A DESCRIPTION OF ṢURAYT/ṬUROYO

43

The clauses quoted above contain a strong assertion. The choice of that particular copula may be connected with—broadly speaking—emotive meaning. The copula kal may be used with a locative predicate together with a focusing particle: (20) Këlli COP.1SG

haž.

Hat you

FOC

haw-li give.PST-1SG

xabro word

gëd

FUT

ëteno come.1SG

‘I am [here], indeed. I promised you to come’. (Bërǧël, 37) The two conditions of the kal usage—strong assertion and eye-witnessing— may be combined. The appearance of the kal copula may be motivated by a contingency, in the form of an unusual situation, connected with eye-witnessing: (21) Bac och

mën d what REL

ẖaşe shoulder traco door

du

REL.ART

qay why

cal on

ẖëzyo see.3SG.F abro, son

b cayn-a! U on eye-her art

mër-la-le: say.PST-3SG.F-to_him

tarco këlle cal door COP on Abr-i! son-my

Hano this

haşo-yo? shoulder-COP

‘What she sees with her own eyes! The door is on the back [shoulders] of [her] son. She told him: Why is this door on your back?’ (Pësseqat, 29) In that case, however, the copula is part of rheme, while the subject is the theme of the clause. This word order seems to be marked. In all the examples above we are dealing with contingent states. The contingency of the state may be connected with the unexpected appearance of the referent in a certain place: (22) I

ART

këlle COP

naqqla suddenly

d

Malla qadi

Ğëzo Ğëzo

REL

caber go.PST.3SG.M

l to

bu in.ART

gawe, inside

k-ẖoyër PRS-see.3SG.M

karmo vineyard

‘When he went into it [the vineyard], he sees the qadi [mufti] Ğëzo is in the vineyard’. (Pësseqat, 30) (23) Aṯ-tox, wife-your

hani FOC

taqne deep

këlla COP.3SG.F

‘Your wife is deep under the water’. (Pësseqat, 31)

taẖt under

am

ART

mayew water

44

MACIEJ TOMAL

(24) Šabo Šero Šabo Šero

mer-le: say.PST-3SG

Abu-na father-our

këllen COP.3PL

nune qum-ox fish before-you

‘Šabo Šero said: O father, fish are before you’. (H̠ ëlme, 30) The copula kal occurs in clauses expressing possession. The motivation for the usage may be a strong assertion: (25) abre son

d

REL

deṯxun. POSS.you.PL

ẖëlti aunt

Bëţrës Bëţrës

Aydarbo how

gëd

FUT

omër: say.3SG

Këlle COP.3SG.M

u

ARt

mal property

mëblitul-le? bring.2PL-it

‘The son of my aunt Bëţrës says: The property is yours [pl.]. How will you bring it?’ (H̠ ëlme, 19) The copula plays also here the role of a textual theme and the clause may be probably better rendered in English by the expression: ‘Look! The property is yours’. The conditions for the particle’s appearance are, therefore, the following: eyewitnessing of a participant of the situation or of a narrator, unexpected appearance of the referent, or an unusual situation, i.e. contrast. As a textual theme, it signals contrastive or adversative meaning. The situation or feature referred to is contingent rather than permanent, which is may also be the case of the kit clauses of the same type. Another motivation for the choice of the kal copula may be the desire to express greater immediacy, in contrast to less prominent background information. The latter condition motivates some authors to use kal in the setting of the narrative, when the referents are introduced. (26) U Tammuz maẖkewo […] “Yawmo And Tammuz say.3SG.CONV day Šamëno […] ẖës hear.1SG voice

hawxa so

šafërto. beautiful

këlli COP.1SG Këlla

AUX

bu in.ART

këṯyo stand.3SG.F

karmo. vineyard lafel-i” in_front-me

‘Tammuz said: One day I am in a vineyard. I hear a voice […]. What a beautiful [girl]. She is standing in front of me’. (Pësseqat, 37) (27) Këlla COP.3SG.F griši pull.PST.3PL

haṯe this

ste

FOC

këllën COP.3PL

i

ART

cël upon

sëfro table

daf

REL

man from

faqire. Ad poor ART

droce handle

adroṯo threshing_floor

‘It is also the table of the poor. They pulled the handle of the door; they are on the threshing floor’. (Bëğël, 39)

TOWARDS A DESCRIPTION OF ṢURAYT/ṬUROYO

45

Accordingly, the copula kal appears with some regularity in the collection of short narratives of Jan Beṯ-Șawoce. The collection, which is called Aṯri Beṯ-Naẖrin baẖ Hëlme di Goluṯo, is a special kind of diary—a somnambular one. It reports the dreams of the author. (28) ẖze-li see.PST-1SG Swed. Sweden

ruẖi myself

këllan COP.1PL

b in

şënëf classroom

bu in.ART

Talmiḏe-wan pupils-COP.PST.1PL

‘I saw myself; we were in the classroom in Sweden. We were pupils’. (H̠ ëlme, 8) (29) ẖze-li see.PST-1SG

ruẖi myself

këlli COP.1SG

talmiḏo pupil

bi in.ART

madrašto school

‘I saw myself; I was a pupil in the school’. (H̠ ëlme, 15) (30) ẖze-li ruẖi see.PST-1SG myself

këllan bu bayt-ayḏan u catiqo b Mëḏyaḏ COP.1PL in.ART house-our ART old in Mëḏyed

‘I saw myself; we were in our old house in Mëḏyaḏ’. (H̠ ëlme, 21) (31) ẖze-li see-1SG i

ART

ruẖi këlli myself COP.1SG

Astël Astel

cam with

u

ART

bu darbo in.ART road abro son

d,

REL

azolo far_away camm-i uncle-my

bi in.ART Yusëf Yusëf

caraba car Karim Karim

‘I saw myself; I was on a road driving a car of Astël together with my uncle Yusëf Karim’. (H̠ ëlme, 22) (32) Cas lo maybe NEG m from

ẖze-lën see.PST-3PL

at

ART

tfënak rifles

qarc-i. U taxt head-my and makeshift_bed

taxt makeshift_bed

biḏ-ayye cël in_hands-their upon dëṯṯe POSS.their

këlle su COP.3PL by.ART

diḏ-i POSS-I

‘Maybe they didn’t see that their rifles were under my head. And their makeshift bed was [lit. is] by that of mine’. (Bërǧël, 64) Equative clauses with the copula kal express an emotive meaning, such as strong assertion, which is reinforced by the adverb šrolo ‘truly’ in the following case:

46

MACIEJ TOMAL

(33) Alo God du

REL.ART

l for

šuro w row and

ëmmi say.PST.3PL soz” swear

ẖa one

manne from_them

mqadam-le go_forth.PST-3SG

omar: say.3SG

“Mar say.IMP.SG

“Ee, ee yes yes

ëmmi “Soz!” say. PST.3PL swear

şuroyena! Syrians.COP.1PL

l for

qume in_front

şuroye Syrians

hatu?” COP.2PL

Omar say.3SG.M

“Maru say.IMP.PL

ẖër-le këlle see.PST-3SG.M COP

šrol

FOC

şuroyene Syrians.COP.1PL

‘One of them appeared before the row [of the Syrians] and says: Are you Syrians? They said: Yes, we are Syrians. He says: Say, we swear! We swear. He saw: they are Syrians indeed’. (Pësseqat Ṣarëke, 13) One can, however, hardly describe the situations as contingent ones. We may say, generally speaking, that in this kind of non-verbal clauses the contingency or permanency of the state is not relevant for the usage of the kal copula. In this type of clause the kal form may lose the function of the copula—as the slot is filled by the enclitic -yo form—and it may [re]gain the status of an exclamatory, presentative particle: (34) Qay? what këlle COP

Cal on

mën what

gaẖixët?” laugh.2SG.M

u

Cisa Cisa

b in

ART

omër: Ëno say.3SG.M I

ẖašawnowo think.1SG.CONV

caqël-yo reason-COP.3SG

‘What is that?! Why are you laughing? I thought that Cisa is right [is wise]’. (Pësseqat, 25) The particle may be used as an existential particle to express contrasting themes: (35) Këlla COP.3SG.F

i

ART

jazwe load

nacëmto light

w and

këlla COP.3SG.F

i

ART

rapṯo heavy

‘There is a light load and there is a heavy one’. (Bërǧël, 39) The conditions for the usage of the copula kal are, therefore, the following: unusual, unexpected situations, a strong assertion up to the degree of exclamation, and coordination with verba sentiendi. As a textual theme, it may express contrastive or adversative meanings. The contingency of the state is appears to be only a secondary criterion since the opposition: contingent vs. permanent is relevant only in some cases.

TOWARDS A DESCRIPTION OF ṢURAYT/ṬUROYO

47

3. KAL AS A VERB MODIFIER (INCLUDING THE PARTICIPLE)

The emphatic or presentative particle plays the function of verb modifier in some Neo-Aramaic dialects. Such usage is restricted to some contexts, not necessarily connected with tense and aspect categories. It may be said that the default reference of the construction is the actual present. It seems, however, that some additional textual and aspectual conditions must be fulfilled. The particle kal plays the function of textual theme, together with a time referent marker: (36) këlla AUX.3SG.F

kiṯyo stand.3SG.F

lafel-i in_front-me

‘Behold! She is standing in front of me’. (Bëğël, 64) tive:

The construction kal + participle may appear within the settings of the narra-

(37) Malke we Malke and yatiwe sit.PTCP.PL

u

ART

Yuẖanun, Yuhanun

bi in.ART

a

ART

lokanda café

tre two

ẖawrone friends

kël[l]ën AUX.3PL

ko-mëjğoli PRS-talk.3PL

‘The two friends Malke and Yuẖanun were sitting in the café and talking’. (Pësseqat Ṣarëke, 9) (38) ẖze-li see.PST-1SG

ruẖi myself

këlli AUX.1SG

yatiwo sit.PTCP.SG

b in

bayto house

‘I saw myself: I was sitting in the house’. (H̠ ëlme, 22) The same construction may be coordinated with a verbum sentiendi: (39) Qašo priest

ma from

qaše priests

REL

k-uzze PRS-go.3SG.M

li to.ART

mḏiṯo town

laqi meet.PST.3SG.M yatiwo sit.PTCP.SG

b on w and

d

ẖa one iḏe hand

ẖeḏër region d

REL

faqiro poor

Mëḏyaḏ Medyad Mërde. Mërde

ko-qoyëm PRS-stand_up Bi in.ART

ẖzë-le see.PST-3SG.M

šuqo market_place

kële

AUX.3SG.M

ftëẖto open

‘One of the priests of the Mëḏyaḏ region starts going to the town of Mërde. In the market place he met a poor man. He [the poor] is sitting with an open

48

MACIEJ TOMAL hand’. (Pësseqat Ṣareke, 24)

(40) k-ẖoyër, PRS-keep.3SG.M u

bac mën what

PEJ

ẖawrayḏ-e friend-his

REL.ART

d

ẖoze, see.3SG.M

REL

këllën COP.3PL

aj

ART

jule garments

tarye wet.PTCP.PL

‘He sees: the garments of the friend are wet’. (Pëssekat, 36) (41) Bramšël yesterday

ẖze-li see.PST-1SG

qëm at

cam with

du

REL.ART

masa table Swed Sweden

b in

i

ART

Margaretha Margaretha

ẖëlm-i, dream-my

këlli AUX.1SG

wazërto ministress af af

yatiwo sit.PTCP.SG

barayto foreign

Uggalas Uggalas

‘Yesterday I saw in my dream: I was sitting at the table together with the minister of foreign affairs Sweden Margharetha af Uglass’. (H̠ ëlme, 37) The kal particle may be preceded by the exclamatory particle ha. The kal particle in such cases may be treated as a textual theme, the kal + participle construction being part of the theme. The combination of locative/exclamatory ha and kal appears in adversative clauses, expressing a reservation about the relevance of a previous utterance, often without questioning the general validity of the preceding statement (Sawicki 2008: 132), in (42) without questioning the relevance of the subject of the participant-narrator’s fear: (42) Qay why

lalyo, cal night because

d

REL

kito COP.3SG

nose ste people FOC

ẖdor-i around-me

k-zacno? PRS-scare.1SG

Qay why

hawxa-yo? so-COP.3SG

Ha

bab-i father-my

w and

falgë di half REL.ART

adro

b in

threshing_floor

këlla AUX.3SG.F

damixe.

sleep.PTCP.PL

Cal because

PRES

qriṯo

village

mën what

l for

këlle AUX.3SG.M cal on zayëno? scare.1SG

‘Why in the night, because people are around me, am I scared? Why it is so? Yet, my father and half the village are sleeping on the threshing floor. Why am I afraid?’ (Bëğël, 46)

TOWARDS A DESCRIPTION OF ṢURAYT/ṬUROYO

49

The verb has actual present time reference. Additionally, the participant of the situation or the speaker is frequently an eye-witness of the situation: (43) K-ẖarno PRS-see.1SG u

hno so

ART

mën k-ẖarno, what PRS-see.1SG bë and

këlle AUX

ẖa one

k-maẖët PRS-stand.3SG

hno so

‘I look: one [of them]—so and so—is standing’. (Pësseqat, 24) Sometimes the particle is used in order to gain greater immediacy and then, again, it is connected with eye-witnessing of the situation: (44) Xëd when

ẖa one

d

REL

cobër i come.3SG.M ART

fabriqa factory

ẖoze see.3SG.M

mi badaye from beginnig

mede, something

d

ah art

mişawli put.3PL

harke here

këllen

k-mëšloqi, PRS-boil.3PL

ẖeţe wheat

tfayo fireplace i

ART

REL

k-saymo PRS-make.3SG.F

AUX

danoke pot

li to

ẖarayto end i

ART

si at.ART

rabṯo great du

REL.ART

fabriqa. factory tfayo, fireplace lu to.ART

Ha

PRES

cal on

w and

d

REL

syomo making këllen AUX.3PL i

ART

gabo inside

k-mëfroso. PRS-squeeze.PASS.3SG.F

‘[It is as] when someone comes into a great factory, sees the beginning and the end of the production of something, whatever the factory produces. They put the wheat on the fireplace, on the fireplace they boil it slightly; in the pot it is squeezed’. (Bërǧël, 62) (45) ẖzeli see.PST.1SG lu to.ART

këlla AUX.3SG.F

aṯt-i wife-my

k-dëwqo PRS-put_into.3SG.F

laẖmo bread

tanuro oven

‘I see my wife putting the bread into the oven’. (H̠ ëlmo, 48) Accordingly, the subordinate clause with kal + verb form expresses an activity simultaneous with that referred to in the main clause, while in the latter the verb belongs to the set of verba sentiendi. The coordination with a verbum sentiendi im-

50

MACIEJ TOMAL

plies that the situation referred to is going on in front of the participant of the situation: (46) U Musa and Musa ẖmoro d donkey REL

mnakëf afraid.PST.3SG.M qamuṯë in_front

du

REL.ART

u

ART

waqt d ẖze-le time REL see.PST-3SG.M

u

ART

babo këllë k-nocër father AUX.3SG.M PRS-bray.3SG.M

‘And Musa got afraid when he saw the donkey that was braying in front of the father’. (Bëğël, 29) The particle kal occurs as a verb modifier also in exclamatory clauses: (47) Hat, you

haw not

ko-ruẖmat-li! PRS-like.2SG.F-me

Meqëm, first

cayno, amţatwa eye bring.2SG.F.CONV

u

W and

ko-ẖozat, PRS-see.2SG.F

cudo, kalox now AUX.2SG.F

ART

ţalab-ayḏi requests-my

bi in.ART

ẖiyarto blink

li to.ART

di

REL.ART

dukṯo. place

falitat! eascape.2SG.F

‘[She says:] “You don’t like me anymore! First you brought what I wanted in a blink of an eye to the place. And now, you see, you are escaping!’ (Pësseqat Ṣarëke, 24) The particle has an actualizing force that gives the verb an actual present reference. That is why it is very frequently coordinated with verba sentiendi. The construction makes it possible to express greater immediacy when the content of the perception is reported. The construction may be used for the sake of immediacy or for expressing a habitual activity. Also, the construction may be used in adversative clauses, expressing reservation as to the previous utterance. It bears, then, a contrastive sense, also as is the case of the copula kal.

4. CONCLUSIONS

The kal copula appears in all three types of verbless clauses—with three types of predications. First, in contrast to the kit copula, the kal one never occurs in relative clauses introducing background information. Second, the position—theme or textual theme—and contingency of the state are not exhaustive conditions for the usage of copula kal in the function of the copula. In contrast to the kit predications, clauses with kal may have either emotive value or be coordinated with the verba sentiendi. In the first case it plays the role of the textual theme that introduces new referents, elements of setting etc. When the copula is a theme, different degrees of assertion or

TOWARDS A DESCRIPTION OF ṢURAYT/ṬUROYO

51

contrast are expressed; when it is an element of the rheme, rather only emotive meaning is expressed thereby. All the aforementioned functions may account for its usage as a verb modifier. Therefore, the complex verbal forms—kal + ko-qatël— express not only actual aspect, but also emotive meaning.

REFERENCES

Sources Bërğël—Can, Adnan. 1989. Bërğël. SiL.Minv, Statens Institut för Laromedel: Stockholm. Pësseqat—Mirza, Abrohom Gabriel. 1997. “Maẖkay H̠ ḏo H̠ reto …” Pësseqat Me Aṯri Beṯ-Nahrin. H̠ ëḏoro Jan Beṯ-Ṣawoce, Frose dë Nisibin: Södertälje. H⃭ëlme—Beṯ-Ṣawoce, Jan. 1994. Aṯri Beṯ-Nahrin baẖ H⃭ëlme di Goluṯo. Beṯ-Froso Nisibin: Södertälje. Pëssekat Ṣarëke, Ṣarëke—Caziz, 1994. Pësseqat. Më Aṯri: Beṯ-Nahrin (b Sëryoyo Macërboyo). Frose dë Nisibin: Södertälje. Bibliography Bedjan, Jean. 1912. Xayyi də-Qaddīš—Vies des Saints. Paris-Leipzig : Harrassowitz. Beṯ-Ṣawoce, Jan. 2012. Xëzne d xabre. Ordlista. Şurayt-Swedi [mëḏyoyo]. Stockholm: Beṯ-Froso & Beṯ-Prasa Nisibin. [in Ṣurayt/Ṭuroyo] —. 2008. [Utveclikng & förberedelse], Svensk-nyvästsyrisk Lärobok. Swedi-Şurayt [Ţuroyo]: Bokförlaget Nisibin. Tsereteli, Konstantin G. 1965. Materialy po aramejskoj dialektologii, I. Urmijskij dialekt [Materials on Aramaic Dialectology, vol. I. The Urmia Dialect], Tbilisi. Frajzyngier, Zygmunt, Krech, Holly, and Mirzayan, Armik. 2002. “Motivation for Copulas in Equational Clauses.” Linguistic Typology 6 (2): 155–198. Forey, Gail. 2002. Aspects of Theme and their Role in Workplace Texts. Work submitted for the Doctor of Philosophy, Department of English Language, Faculty of Arts, University of Glasgow. Khan, Geoffrey. 2002. A Neo-Aramaic Dialects of Christian Qaraqosh. Leiden: Brill. Jastrow, Otto. 1985. Laut- und Formenlehre des neuaramäischen Dialekts von Mīdin im Ṭūr ‘Abdin. Wiesbaden: Harrassowitz. Lidzbarski, Mark. 1896. Die neu-aramäischen Hanschriften der Königlishen Bibliothek zue Berlin, Bd. 1–2, Berlin: E. Faber. Murre-van den Berg, Heleen, L. 1999. From a Spoken to a Written Language. The Introduction and Development of Literary Urmian Aramaic in the Nineteenth Century, Leiden: “De Goeje Fund,” Nederland Instituut Voor Het Nabije Osten. Ping, Alvin L., 2000. “Identifying the Theme of Existential Clauses: A Suggested Approach.” Folia Linguistica 24 (3–4): 307–331.

52

MACIEJ TOMAL

—. 2004. Theme and Rheme: An Alternative Account. Bern: Peter Lang AG, European Academic Publishers. Polotzky, Hans. 1961. “Studies in Modern Syriac.” Journal of Semitic Studies 6: 1–35. Ritter, Helmut. 1967–1971. Ṭūrōyo. Die Voklssprache der syrischen Christen des Ṭūr ‫ﻉ‬Abdin, Band 1–3, Beirut. —. 1971. Volssprache der syrischen Christen des Ṭūr ‫ﻉ‬Abdīn. Wörterbuch. BeirutWiesbaden: Rudolf Selheim. Yona, Sabar. 1983 The Book of Genesis in Neo-Aramaic in the Dialect of Jewish Community of Zakho, Including Selected Texts in Other Neo-Aramaic Dialects, and Glossary. Jerusalem: Hebrew University, Language Tradition Project. Sawicki, Lea. 2009. Toward a Narrative Grammar of Polish. Warszawa: Wydawnictwa Uniwersytetu Warszawskiego. Stoddard, David. 1855. Grammar of the Modern Syriac Language as Spoken in Oormia and Koordistan. New Haven. Tomal, Maciej. 2008–2009. Studies in Neo-Aramaic Tenses. Kraków-Budapeszt: Austeria.

PRELIMINARY REMARKS ON DISCOURSE PRAGMATICS IN ṬUROYO1

MICHAEL WALTISBERG 1. INTRODUCTORY REMARKS

Discourse pragmatics is an approach not without its methodological problems (cf., inter alia, Van Valin and LaPolla 1997: 199ff.; Lambrecht 1998; Ariel 2008). There are different approaches to pragmatic issues, and the terminology used is not as clear-cut and well defined as one would wish. The available secondary literature is extensive and diverse. Still, for linguistic analysis, the relationship between context and meaning is of vital importance and plays a major part in functionally-oriented linguistic approaches. As for Neo-Aramaic, see the relevant sections in the grammars by Khan (e.g. 2002 or 2008). Due to the lack of the necessary data, it is not possible at this time to write a comprehensive account of the discourse pragmatics of Ṭuroyo. Instead, two minor aspects of discourse pragmatics in Ṭuroyo will be discussed: firstly, the seemingly redundant use of the independent subject pronoun and, secondly, the various techniques that a speaker of Ṭuroyo may use to make his narrative livelier or more interesting, in order to catch the listener’s attention. To my knowledge, common terminology for the latter phenomenon is unclear. In addition, this phenomenon borders on stylistics, insofar as the speaker deliberately chooses his linguistic means for his pragmatic and/or stylistic purposes. Since both phenomena are matters of the parole in the de Saussurian sense (see Glück 1993: 452; de Saussure 1994), it is clear that they are optional and that there 1

An earlier version of this paper was presented at The Neo-Aramaic Conference in Cambridge, UK, on July 7th 2011. The text has been slightly revised.

54

MICHAEL WALTISBERG

is a certain fluctuation in the use of their means of linguistic expression. The exact identification of the reasons for their occurrence, though usually quite straightforward, may in some case be difficult or even impossible. The various Ṭuroyo sources are cited in this paper with abbreviations, for which see the references below. The transcription of the Ṭuroyo examples has been standardized.

2. THE USE OF THE INDEPENDENT PERSONAL PRONOUN

As a pro-drop language, i.e. a language that does not need an overt subject pronoun on account of its being incorporated in the copula or the verb, Ṭuroyo does not need an independent pronoun in subject function (inter alia JG 129, 130). But it is still quite common and has some discourse pragmatic relevance (also Ritter 1990: 2). The same phenomenon occurs in other Neo-Aramaic varieties as well (see, for example, the remarks in Khan 2008: 865ff.). The following account bears similarities with Khan’s description, but deviates in some points. As mentioned above, such a pronoun is certainly optional and belongs to the freedom of the parole. As far as is discernible at the moment, in Ṭuroyo, the distinction between copula clauses and verbal clauses is not relevant to the use of an independent subject pronoun. In addition, in most cases, it seems to be irrelevant whether the pronoun is preposed or postposed. One exception is its use next to the imperative (on this see below). The different uses of such a pronoun in Ṭuroyo may be categorized into three groups. As will become clear, some functions may overlap. The overall functional link between these uses seems to be the marking of a change in the discourse structure of the utterance. The first function is contrastive focus (see, inter alia, Drubig and Schaffar 2001; Molnár 2002), usually with two different pronouns (or one pronoun and a noun) in adjacent clauses: (1) a.

ăṭ-ṭaye ART-Muslims hĭnnĭk they

d-ḥeḏor-ăn, ăḥna REL-around-us we

b-u-noquṣo in-the-least

tleṯi-qriyawoṯe thirty-villages u-hĭnǧ-ăṯṯe ART-pretence-their

kĭp-pe comprise.VBD-3PL kŭlle-ṭaye all.PL-Muslims ăʿl-ăn on-us

ḥḏo-qriṯo-na, făqăt one-village-COP.1PL but ʿĭsri twenty ḥeḏor-ăn, around-us

w-layim-i and-gather.PST-3PL

hŭl, until w-sĭm-me and-do.PST-3PL

PRELIMINARY REMARKS ON DISCOURSE PRAGMATICS IN ṬUROYO

55

‘The Muslims around us used a pretence—we are only one village, but they have at least 20 to 30 villages, they are all Muslims around us—and they gathered’. (JL 85.1) b.

deri next

yăwmo day

ṣăfro morning

mĭḷ-ḷe-le say.PST-3SG.M-to_him ono I

omĭr say.3SG.M

g-noḥătno FUT-go_down.1SG.M

u-ḥolo ART-uncle

d-u-zʿuro REL-ART-little

fĭš stay.IMP.SG

hăt you

l-i-wălaye to-ART-town

hărke, here

zobăṭno catch.1SG.M

dukano shop

‘The next day in the morning, the uncle of the little one said to him: You stay here, while I go down to town and obtain a shop’. (JM 43.34) c.

hiye he

ko-doyĭq PRS-beat.3SG.M

b-i-ʿărbanak-ăyḏe in-ART-drum-his

w-u-măymun and-ART-monkey

ko-roqĭḏ PRS-dance.3SG.M ‘He beats his drum, and the monkey dances’. (JM 44.41) The second use is to indicate a change of subject (German Neueinsatz). This is quite common at the beginning of an utterance: (2) k-omĭr PRS-say.3SG.M omăṇṇo say.1SG

qăy why ono I

dĭš-lŭx run_over.PST-2SG.M b-ḥămd in-purpose

ăʿl-a. on-her

w-lo-dĭš-li and-NEG-run_over.PST-1SG

ruḥa herself

l-i-qamăyt-ăyḏi to-the-front-my

diḏi mine

ʿal on

i-kăčk-aṯe? ART-girl-this

lo-dĭš-li NEG-run_over.PST-1SG ăʿl-a, on-her

maye-li swerve.PST-1SG

mḥalăq-la throw.PST-3SG.F m-ăʿl-a from-on-her

‘He asked: Why did you run into this girl? I said: I did not deliberately run into her. I indeed did not run into her, she threw herself in front of me, and I swerved from her’. (JL 119.14) In addition, the pronoun sometimes serves to resume a previously mentioned subject after the insertion of clauses with other constituents:

56

MICHAEL WALTISBERG

(3) a.

haṯe this

i-xodĭmto d-i-bărṯo ART-servant REL-ART-daugher

d-u-tagoro REL-ART-merchant

d- ĭmmi-na FUT-say.1PL

nafiq-o go_out.PST-3SG.F

l-u-šaboko, to-ART-window

ḥĭḷ-ḷa, look.PST-3SG.F

ḥze-la see.PST-3SG.F

u-zlam-ano ART-man-this

ko-moḥe PRS-beat.3SG.M

ʿal on

w-u-măymun-ăyḏe and-ART-monkey-his

ko-roqĭḏ, PRS-dance.3SG.M

hăwxa thus

ăʿl-e, on-him

ko-mfărǧ-i PRS-watch.3PL

b-u-šaboko in-ART-window

ko-ḥăyro PRS-look.3SG.F

kit

i-ʿărbane ART-drum

kmo-zʿure some-little.PL

COP

fayiš-o stay.PST-3SG.F

hiya he

ăʿl-e on-him

‘The servant of the merchant’s daughter stepped towards the window and looked. She saw this man beating the drum and his monkey dancing. Some children stood there watching him. She stayed at the window looking at him’. (JM 45.47) b.

w-qayĭm and-stand_up.PST.3SG.M d-u-Farănsa REL-ART-France

ǧĭdd-i grandfather-my

tamo there

b-i-ʿăskăr by-the-army

măzbăṭ-le cause_seize.PST.him-3SG.M

w-măslăm-me l-Tĭrkiya, and-hand_over.PST.him-3PL to-Turkey

w-mămṭă-lle, and-bring.PST.him-3PL

kĭtwăy-lăn have.VBD.PST-1PL

ăḥna we

qaraqol police_station

mămṭă-lle bring.PST.him-3PL

l-i-qriṯo-ze, to-ART-village-FOC

w-i-ʿăskăr and-ART-army

mbe-la bring.PST.him-3SG.F

hŭl until

b-Mărde in-Mardin

hawi become.PST.3SG.M

ifada interrogation

Mărde, Mardin

b-i-qriṯo in-ART-village

‘My grandfather caused him to be arrested by the French army there. They extradited him to Turkey and brought him even to the village—we had a po-

PRELIMINARY REMARKS ON DISCOURSE PRAGMATICS IN ṬUROYO

57

lice station in the village. The army brought him to Mardin, (where) he was questioned’. (JL 132.31) The most problematic term that will be employed here to describe the use of an independent subject pronoun is emphasis for the third group. This occurs in various contexts, some notion of contrast or the like may also be involved: (4) a.

ĭmmo say.3SG.F

ya-zlam o-man

hăt you

zlam man

nŭxroyo-hĭt foreign-COP.2SG.M

w-aṯ-ĭt and-come.PST-2SG.M

qĭm before

u-tărʿo, ART-door

i-sĭsto ART-horse

măkle-la-lŭx stop.PST-3SG.F-you

qĭm before

u-tărʿo, ART-door

nafiq-o-no, come_out.PST-F-1SG

mănḥăt-li-lŭx cause_dismount.PST-1SG-you

măʿbăḷ-ḷi-lŭx usher_in.PST-1SG-you

w-mšaḥăl-li-lŭx and-make_warm.PST-1SG-you

w-manĭḥ-li-lŭx and-restore.PST-1SG-you

maqĭm-li-lŭx, cause_stand.PST-1SG-you lo-ṭĭlbĭt NEG-ask.2SG.M bărṯo girl

mede something

min-i, from-me

d-săymono REL-do.1SG.F

ḥăram illicit

ono I

mede something

lălġŭl, inside

lăn-no COP.NEG-1SG ḥăramtiye illicitness

‘She said: O man, you are a foreign man. You have come to my door, the horse stopped with you in front of the door, I came out, let you dismount, took you inside and made you warm. I restored you and made you stand up again. Do not ask something illicit from me. I am not a girl that does something illicit’. (JG 278.15) b.

ĭmmi-wa say.3PL-CONV

mădam since

mšiḥoyo-hĭt Christian-COP.2SG.M

băynoṯ-ăn? among-us

ko-qŭṭlutu PRS-kill.2PL

b-ăwrupa in-Europe

mĭk-ko-săymĭt what-PRS-do.2SG.M

ăq-qŏnṣolos-ăyḏăn ART-consuls-our

kŭl-le all-PL

58

MICHAEL WALTISBERG hatu you

ko-qŭṭlut-ne PRS-kill.2PL-them

u-šŭġl-ắtxu ART-work-your

w-hatu and-you

w-lătyo and- COP.NEG

hărke here

ġam-ăyxu-ze. worry-your-FOC

ko-šŭġlutu PRS-work.2PL ăʾ-ʾăwraq-ắṯxu ART-papers-your

lĭ-g-săymină-nne NEG-FUT-do.1PL-them ‘They said: Since you are a Christian, what are you doing among us? You kill all our consuls in Europe, you kill them and here, you do your job, and you don’t mind. We will not process your papers’. (JE 224.31) c.

ăzz-ino go.PST-1SG ḥĭtta even

nafil-o-no fall.PST-F-1SG

ă-ġzalăt ART-gazelles

ko-zăyʿono

PRS-fear.1SG.F

ko-fĭhmono PRS-understand.1SG.F hăwy-o become.PST-3SG.F

b-u-ṭuro anăqla in-ART-mountain then

d-kĭt-no REL-COP-1SG i-ṭabiʿa ART-nature

mĭn-ne, from-them isan, human

d-ă-ġzalăt REL-ART-gazelles

k-ĭzzino, PRS-go.1SG ono I băs but ăʿm-i with-me

‘I went away to the mountains (where) I was roaming. I was even afraid of the (other) gazelles, (for) I understood that I was a human being, but I had been transformed into a gazelle’. (JM 50.83) In example (4b), the subject does not change, the statement is repeated and therefore stressed. The use of the pronoun underlines this. In the last example above (4c), a notion of contrast may be involved: unlike the other gazelles, she is aware of her human nature. On many occasions a pronoun occurs for emphasis next to an imperative. Khan (2008: 870) speaks of “added force” which is indicated by the pronoun. In the majority of cases, such a pronoun precedes the imperative: (5) k-ĭmmi PRS-say.3PL mŭnasĭb, suitable

ba

PTC

hăt you

ḥzi, see.IMP.SG

sĭm-la do.IMP.SG-to_her

ăyk where

ko-ḥozăt-la PRS-see.2SG-to_her

dŭkṯo place

‘They said: Look yourself! Make her location, where you find something suitable for her!’ (R1 412.20 = RH 81.[20])

PRELIMINARY REMARKS ON DISCOURSE PRAGMATICS IN ṬUROYO

59

When particular focus is put on the pronoun, it is placed after the imperative. The crucial discourse-pragmatic factor in the following example is contrast: (6) u-fŭrzli ART-Furzuli omĭr say.3SG.M

mĭḷ-ḷe-le say.PST-3SG.M-to_him

l-u-băʿlbăkki to-ART-Baʿlbakki

de

l-u-qašo! to-ART-priest

PTC

u-băʿlbăkki ART-Baʿlbakki

ăḥkiy-a tell.IMP.SG-it

omĭr say.3SG.M

hăma

ăḥkiy-a tell.IMP.SG-it

PTC

hăt you

‘The man from Furzul told the one from Baalbek: Tell it to the priest! The man from Baalbek replied: You tell it!’ (R1 498.35) Although most instances of an independent subject pronoun can be subsumed under one of the three headings mentioned, it is sometimes difficult or even impossible to determine its exact meaning. A case in point is the following somewhat problematic passage: (7) more owners

d-ă-ʿwone REL-ART-sheep

mamĭṣ-ṣe let_suckle.PST-3PL

ă-ʿwon-ăṯṯe, ART-sheep-their

u-rĭʿyo aṯi ART-shepherd come.PST.3SG.M

lăff-el-i toward-to-me

ko-qolĭ PRS-goad_on.3SG.M

ono-ste I-FOC

he still

me Azĭx from Azekh

ă-ʿwone. ART-sheep

b-kis-i in-pocket-my

kĭt-wa ṭăwʿo naʿimo COP-PST ṭăwʿo small

twiro-wa, broken-be.PST.3SG

mĭd-li take.PST-1SG

qătte chunk

m-u-ṭăwʿ-ano, from-ART-ṭăwʿo-this

l-i-qamăyto to-ART-front

d-ă-ʿwone. REL-ART-sheep

ono-ste I-FOC

hăwxa, thus

u-rĭʿyo ART-shepherd

ko-ḥoyĭr PRS-look.3SG.M

ono-ste I-FOC

ăzz-ino go.PST-1SG k-omăṇṇo dbrdbr PRS-say.1SG.M dbrdbr ăʿl-i on-me

‘The sheep owners let their sheep suckle. The shepherd came in my direction, goading on the sheep. I still had a small ṭăwʿo—i.e. a small piece of bread used for the Eucharist—from Azekh in my pocket which was broken. I took a chunk of this ṭăwʿo and went in front of the sheep. (When) I made the sound of ‘dbrdbr’, the shepherd looked at me’. (LB 108.244ff.)

60

MICHAEL WALTISBERG

Whereas the first use of ono-ste can be easily explained (change of subject), the next two instances are hard to explain. My informant told me he did not understand the use of ono-ste here, for him, it is “pointless”. The repeated use of -ste is another problem in these instances. It should be noted that the marking of a discourse pragmatically relevant change in the narrative, e.g. change of subject, by means of an independent pronoun as discussed above is just one of several strategies that Ṭuroyo uses for this purpose. Further strategies are the use of demonstratives and the particle -ze ~ -ste(ne), leftdislocation, the verb qoyĭm lit. ‘stand up’ at the beginning of a clause or combinations of these means. Their description will form a part of a more comprehensive account of Ṭuroyo syntax.

3. TECHNIQUES WHICH RENDER A NARRATIVE IN ṬUROYO LIVELIER

How can a speaker of Ṭuroyo make his narrative livelier? Of course, “livelier” is a matter of degree and sometimes hard to fathom. Again, it belongs to the parole, and such usage is optional. There are several techniques that I think serve this purpose, and it may well be that there are still others which are not mentioned in this article. One possible example is the predominant use of direct speech instead of reported speech. This is certainly a stylistic device, but since it is so common, not only in Ṭuroyo, but in Semitic languages in general, it will not further be discussed here. In European languages, the so-called ‘historical/narrative present’ is a common device to give the sense of immediacy to segments of a narrative, to make them foregrounding events (cf. inter alia Glück 1993: 480; Comrie 1976: 73ff.). Both available grammars of Ṭuroyo mention this as one of the functions of the present tense. Ritter (1990: 54) states that the speaker narrates the events in the present tense as if he witnesses them (“dass er das erzählte geschehen gleichsam miterlebend im präsens erzählt”). Jastrow (JG 147) interprets the function of the present tense in narrative as a means of expressing a greater vividness and immediacy (“Das Präsens hat gegenüber dem Präteritum den Vorzug grösserer Anschaulichkeit und Unmittelbarkeit und ist infolgedessen im lebhaften Erzählstil sehr beliebt”). But the use of the present tense in Ṭuroyo narrative is so common (see also Ritter 1990: 54, “ziemlich regellos”) that it hardly warrants the introduction of the notion of ‘historical present’. In fact, it may well be that the present tense in narrative can be interpreted as having its basic aspectual-temporal function (for now cf. Wilke 2008: 82ff.). On the whole then, the use of the present tense in Ṭuroyo narrative is not merely a stylistic device. Therefore, it is quite possible that a

PRELIMINARY REMARKS ON DISCOURSE PRAGMATICS IN ṬUROYO

61

narrative present in Ṭuroyo simply does not exist, although imperfective aspect admittedly may be suited for this purpose. One possible instance, though, may be the following passage which may be explained in this fashion: (8) dwĭṣ-la b-răġl-e bite.PST-3SG.F in-leg-his qoyĭm stand.3SG.M dŭkṯe place

u-găldo ART-skin ăʿl-a on-it

gorĭš i-xănǧăr, draw.3SG.M ART-dagger

d-feme REL-mouth

b-iḏ-e in-hand-his

lăltăḥ unterneath

d-i-kŭrfo REL-ART-snake

w-disane and-again

hawo-ste that-FOC

zobĭṭ grab.3SG.M d-i-dwăṣto REL-ART-biting

w-b-i-xănǧăr and-in-ART-dagger w-maḥĭt and-put.3SG.M

m-i-bărko, from-ART-knee

qoṭăʿ-la, cut.3SG.M-it hawo-ste that-FOC

qoṭʿĭ cut.3SG.M

dărmono powder

qṭile kill.PST.3SG.M

i-kŭrfo ART-snake

‘(The snake) bit his leg underneath his knee. He draws his dagger, grabs the spot with his hand where the snake had bitten, and cuts it with the dagger. He cuts the skin and puts (musket) powder on it. He again killed the snake’. (JG 302.63)2 There are techniques that certainly serve to make a text livelier. One of these is the sequence of several quick actions one after another in order to accelerate the narrative. This is the case, for example, in the following passage concerning the speaker’s abduction: (9) năqla nafiq-i suddenly go_out.PST-3PL d-kĭtwa REL-COP.PST

ḥewărto white

w-nafiq-i and-go_out.PST-3PL 2

silaḥli tre armed two

m-i-tăqsiye from-ART-taxi

w-pilaq and-number_plate

l-qamuṯ-i to-front-my

hăma PCL

lăyto COP.NEG

ăʿl-a. on-it

yarix-i make_oneself_long.PST-3PL

According to my informant, the use of the bare imperfective stem in this passage may be problematic. The present prefix ko- is expected.

62

MICHAEL WALTISBERG l-i-dŭkṯo d-ono to-ART-place REL-I ftĭḥ-ḥe open.PST-3PL

u-tărʿo ART-door

l-abĭʿ-no NEG-want.PST-1SG b-u-zor, in-ART-force

kĭtwăy-no

w-mĭd-de and-take.PST-3PL

d-noḥăn-no REL-get_out-1SG

mḥă-llă-lli beat.PST-3PL-me

hŭl until

tre-ne w-silaḥli two-COP.3PL and-armed nḥir-ăyye nose-their

qĭm before

w-maṣĭṛ-ṛe ʿăyn-i, and-bind.PST-3PL eyes-my

mĭḥalli. seat l-druʿ-i, to-arm-my

grĭš-šăl-li draw.PST-3PL-me

b-ăd-dăḥfăt in-ART-hits

băyn l-ă-tre b-i-ḥarăyto between to-ART-two in-ART-back hĭnĭk they

b-šufer in-driver

COP.PST-1SG

yani that_is

w-maḥăt-tă-lli and-put.PST-3PL-me d-u-tăqsi REL-ART-taxi

w-maṣĭṛ-ṛe ĭšmo and-bind.PST-3PL something

ʿăyn-ăyye eyes-their

w-maḥăt-tă-lli and-put.PST-3PL-me

maṣĭṛ-ṛe bind.PST-3PL

iḏ-i laxălf hands-my backward

‘Suddenly two armed men got out of the white taxi without a number-plate and came toward me. They rushed to the passenger seat, where I was. They opened the door, took my arm, I did not want to get out, they drew me out with force. They beat me and put me between the two men on the backseat—that is to say of the taxi. They were two and armed, and had bound something over their face. They put me there and bound my eyes, they bound my hands on my back’. (LB 48.29ff.) Another device is the common use of presentatives. There are several presentatives in Ṭuroyo. These include kalé, kăl, ha, ga, as well as demonstrative pronouns in presentative function (see JL 106 and Ritter 1990: 38ff.). (10) a.

čĭk-no b-băxt d-alo w-diḏŭx, enter.PST-1SG in-mercy REL-god and-yours

ḥŭr, look.IMP.SG

omĭr, say.3SG.M

i-ḥĭrm-ăyḏi ART-wife-my

i-bărṯo ART-girl

w-u-ḥăwr-ăyḏa and-ART-friend-her

hărke here

d-u-ḥăram REL-ART-illicit maṣĭr-ră-lli bind.PST-3PL-me

hiya she

mḥalăq-qă-lli throw.PST-3PL-me

PRELIMINARY REMARKS ON DISCOURSE PRAGMATICS IN ṬUROYO l-ărke w-kal-ĭn to-here and-PRES-3PL

b-ḥăramtiye in-illicitness

yatiwe ʿăm sitting.PL with

ḥḏoḏe each_other

‘I ask God’s mercy and yours! Look, the whore’s child, my wife, and her boyfriend have put me in fetters and thrown me here, and there they sit together illicitly’. (JG 282.25) b.

ʿariq-i hĭnne flee.PST-3PL they

ko-mhălxi anăqla PRS-walk.3PL now

d-u-ăbro d-ʿămm-i ko-ḥĭzyo REL-ART-son REL-uncle-my PRS-see.3SG.F

Šmuni i-ăṯto Šmuni ART-wife mede something

yarixo long

ko-folĭt bĭṯr-a, ko-ḥăyro laxăf, PRS-move.3SG.M behind-her PRS-look.3SG.F backwards ga

PRES

kăl

PRES

kit

COP

mede ko-rohĭṭ something PRS-run.3SG.M

bĭṯr-a behind-her

xd-i-kŭrfo as-ART- snake ‘They ran away. Šmuni, the wife of my cousin, saw something long moving behind her, so she looked back—there was something like a snake running behind her’. (JG 290.12) c.

mĭḷ-ḷe l-u-ăbro say.PST-3SG.M to-ART-son aṯi ḥĭs come.PST.3SG.M sound lo

NEG

rabo old

nafĭq go.out.PST.3SG.M

ăbr-i son-my

m-i-dărga, zŭx from-ART-door go.IMP.SG

mede qalĭb something turn.PST.3SG.M

l-u-remono to-ART-pomegranate_tree

omĭr say.3SG.M

l-u-kărmo to-ART-vineyard

omĭr say.3SG.M

ăzz-e go.PST-3SG.M

u-ăbro ART-son

l-qŭl to-before

ḥzăy see.IMP.SG ăzze go.PST.3SG.M rabo old

i-dărga ART-door

kal-a ftĭḥto PRES-it open ‘He said to the eldest son: My son, a sound came from the door, go and take a look whether something went over to the vineyard and went to the pomegranate tree! The eldest son went out and went in front of the door, there

63

64

MICHAEL WALTISBERG it was open’. (R3 552.381)

As in example (10b), presentatives are quite common after verbs of perception such as ḥoyĭr ‘look’ or ḥoze ‘see’. The occurrence of the presentative in such cases may be conditioned primarily by the preceding verb. Probably the most important device in a lively narrative is the word năqla ~ năqqa ‘suddenly’. This is used in several ways, it may even be repeated. Năqqa generally serves to indicate surprise, the sudden occurrence of a new event. Its use can already be seen in number (9) at the beginning of the narrative of the speaker’s abduction: năqla nafiqi silaḥli tre etc. Other instances are the following passages: (11) a.

yăwmo day

kal-ăn PRES-we

ʿal on

i-ăyno, ART-well

w-raʿi-wăy-na, and-graze-PST-1PL

sawiʿe-wăy-na, full-COP.PST-1PL

i-ărʿo, hăwxa ART-earth thus

maḥăt-lăn put.PST-1PL

ĭmm-o say-3SG.F qole voice ĭʿl above

năqla suddenly

maġe-lăn settle_down.PST-1PL

riš-ăn head-our

u-mede

ART-something

d-tfĭnge nafĭq REL-rifle go_out.PST.3SG.M me from

ko-šte-lăn

PRS-drink.PST-1PL

qĭm before

ḥḏoḏe each_other

d-ḥze-li REL-see.PST-1SG

w-u-bŭġro and-ART-bullet

ʿal on

ono, I

šapĭʿ pass.PST.3SG.M

qărʿ-i-va head-my-along

‘We were at the well one day, (where) we had drunk and grazed. We were full. We had settled down on the ground and put our heads next to each other, when suddenly—to my horror—the sound of a rifle resonated, and a bullet passed over my head’. (JM 51.87) b.

hano this

kŭ-mḥalĭq ʿăl PRS-throw.3SG.M on

w-năqla and-suddenly ko-midăwdlo, PRS-swing.3SG.F

kŭ-mʿalăq-le

gŭšro m-d-u-qṭoro beam from-REL-ART-ceiling

PRS-hang.3SG.M-it

năqqa suddenly

w-ftiḥ and-be_opened.PST.3SG

ḥăwlo. rope

b-qḏol-a, w-năqla in-neck-her and-suddenly ḥniq-o choke.PST-3SG.F

fem-a. mouth-her

omĭr say.3SG.M

ḥur-u look.IMP-PL

PRELIMINARY REMARKS ON DISCOURSE PRAGMATICS IN ṬUROYO

hĭngi dŭkṯ-a because_so_much place-her haṯe this

basĭmto-yo kal-a nice-COP.3SG PRES-she

sim-o-le do.PST-it.3SG.F-3SG.M

d-lo-foyĭš REL-NEG-remain.3SG.M

fṣiḥăyto. cheerful

l-u-Qănda PRP.ART-Qănda

noše person

fohĭmtĭr more_intelligent

ḥătta until men-e from-him

b-i-qriṯ-ayo in-ART-village-that ‘(Qănda) throws a rope over one of the beams of the ceiling. Suddenly, he fastens it on her neck and at once she swings to and fro. She was asphyxiated, and her mouth opened. He said: Look, since her location is so nice, she is cheerful! This Qănda did, until there was no one more intelligent than himself in the village’. (R1 412.22.23 = RH 81[.22.23]) c.

w-ăzz-ehĭn b-lălyo and-go.PST-3PL in-night

dlo ḥĭs without sound

w-maṭ-ĭn and-come.PST-3PL

l-qŭm-u-šuro d-i-kozke d-yatiw-i-wa eb-a before-ART-wall REL-ART-raised_hide REL-sit.PST-3PL-CONV in-it tlĭṯ-mo three-hundred

zlamăt men

m-ăṭ-ṭaye from-ART-Muslims

w-năqla and-suddenly

mḥă-lle ṣadr-ăyye b-u-šuro d-i-kozk-ayo beat.PST-3PL chest-their in-ART-wall REL-ART-raised_hide-that w-măzʿăq-qe ʿal and-shout.PST-3PL on aṯi-nă-nxu come.PST-1PL-to_you ăṭ-ṭaye ART-Muslims mĭd when

ĭmm-i say-3PL

w-dră-lle u-šuro and-overturn.PST-3PL ART-wall

d-kĭtwăy-ne REL- COP.PST-3PL

ḥză-lle see.PST-3PL

zayiʿ-i fear.PST-3PL

ăṭ-ṭaye ART-Muslims

b-gawe in-interior

ʿal on

d-i-kozke

REL-ART-raised_hide

u-med-ano l-ăṭ-ṭaye ART-thing-this PRP.ACT-ART-Muslims

măhzăm-me flee.PST-3PL

b-i-xăṣra in-ART-side

ḥreto-da other-along

‘And they went by night without a sound and came to the wall of the raised

65

66

MICHAEL WALTISBERG hide in which three hundred Muslim men had sat down. Suddenly, they threw their chest against the wall of the raised hide and shouted at the Muslims: We have come over you! They overturned the wall onto the Muslims who were inside the raised hide. When the Muslims saw this, they were afraid and fled out of the other side’. (R1 336.30)

It should be noted, however, that the occurrence of năqqa does not necessarily indicate strong surprise in every instance. An example with a presumably rather weak degree of surprise may be the following passage: (12) omĭr say.3SG.M

aḥuno brother

w-gĭd-qŭṭʿăt and-FUT-cut_off.2SG d-obăt-li REL-give.2SG-to_me solăqno go_up.1SG sĭm do.IMP.SG hano this

mădam since qărʿ-i head-my

mĭhla respite

l-u-nigoro to-ART-roof

hăwxa-yo thus-COP.3SG k-obăʿno

PRS-want.1SG

hăma PTC

d-noqăḥno

u-mede d-k-ĭbʿăt. e, yes

d-saʿaye

REL-hour

hawa air

REL-smell.1SG

ART-thing REL-PRS-want.2SG

omĭr say.3SG.M

ruʿe quarter

ġamo concern

ḥăq right

lăyt, COP.NEG

w-edi and-then

diḏŭx-yo. yours-COP.3SG isăq lălʿăl go_up.IMP.SG above

hŭl ruʿe d-saʿaye w-bĭṯĭr nḥăt omĭr until quarter REL-hour and-after go_down.IMP.SG say.3SG.M qayĭm salĭq ʿal get_up.PST.3SG.M go_up.PST.3SG.M on

i-nigoro ART-roof

w-maqĭṯ-le-le qăllĭn w-năqqa and-light.PST-3SG.M-to_him pipe and-suddenly i-pĭrtke ART-feather

d-u-ṭăyro REL-ART-bird

ʿal on

maḥăt-le put.PST-3SG.M

i-nuro ART-fire

‘He said: Brother, since this is the case and you are going to cut off my head, I would like to ask you to give me a respite of a quarter of an hour, so that I can go up on the roof and get some fresh air. Then do what you want to, it is your right. The other said: Yes, fine, go upstairs for a quarter of an hour, and then come down. He went up on the roof, lit himself a pipe, and (lit. suddenly) put the feather of the bird onto the fire’. (R1 466.153)

PRELIMINARY REMARKS ON DISCOURSE PRAGMATICS IN ṬUROYO

67

This ends our short survey of two aspects of discourse pragmatics in Ṭuroyo. It is clear that there are more discourse pragmatic aspects to be investigated and discussed. Despite its methodological problems, this is an approach to Ṭuroyo that promises to reveal many functional features that complement the description of the syntactic features of the language.

REFERENCES Sources with their abbreviations

JE = Jastrow, Otto. 1994. “Erlebnisse eines Lastwagenfahrers. Ein neuer ṬuroyoText im Dialekt von Midən.” In Festschrift Ewald Wagner zum 65. Geburtstag. Band 1 Semitische Studien unter besonderer Berücksichtigung der Südsemitistik, edited by Wolfhart Heinrichs and Gregor Schoeler, 221–233. Beiruter Texte und Studien 54. Beirut: Steiner (in Kommission). JG = Jastrow, Otto. 1993. Laut- und Formenlehre des neuaramäischen Dialekts von Mīdin im Ṭūr ʿAbdīn. Semitica Viva 9. Wiesbaden: Harrassowitz. JL = Jastrow, Otto. 1992. Lehrbuch der Ṭuroyo-Sprache. Semitica Viva, Series Didactica. Wiesbaden: Harrassowitz. JM = Jastrow, Otto. 1968. “Ein Märchen im neuaramäischen Dialekt von Mīḏin (Ṭūr ʿAbdīn).” Zeitschrift der Deutschen Morgenländischen Gesellschaft 118: 29– 61. LB = Talay, Shabo. 2004. Lebendig begraben. Die Entführung des syrisch-orthodoxen Priesters Melki Tok von Midən in der Südosttürkei. Einführung, Aramäischer Text (Turoyo), Übersetzung und Glossar. Studien zur Orientalischen Kirchengeschichte 29. Münster: Lit. R1 = Ritter, Helmut. 1967. Ṭūrōyo. Die Volkssprache der syrischen Christen des Ṭūr ‫ﻉ‬Abdîn. Texte Band I. Beirut: Steiner (in Kommission). R3 = Ritter, Helmut. 1971. Ṭūrōyo. Die Volkssprache der syrischen Christen des Ṭūr ‫ﻉ‬Abdîn. Texte Band III. Beirut: Steiner (in Kommission). RH = Ritter, Helmut. 1967. “Ṭūrōyo (Ṭōrānī).” In An Aramaic Handbook. Part II/1 Texts, edited by Franz Rosenthal, 78–81. Porta Linguarum Orientalium – Neue Serie X. Wiesbaden: Harrassowitz.

Secondary literature

Ariel, Mira. 2008. Pragmatics and Grammar. Cambridge textbooks in linguistics. Cambridge: Cambridge University Press. Comrie, Bernard. 1976. Aspect. Cambridge: Cambridge University Press. Drubig, Hans Bernhard, and Schaffar, Wolfram. 2001. “Focus Constructions.” In Language Typology and Language Universals. An International Handbook, edited by

68

MICHAEL WALTISBERG

Martin Haspelmath et al., 1079–1104. Handbooks of Linguistics and Communication Science, 2002. Berlin: de Gruyter. Glück, Helmut, ed. 1993. Metzler Lexikon Sprache. Stuttgart: Metzler. Khan, Geoffrey. 2002. The Neo-Aramaic Dialect of Qaraqosh. Studies in Semitic Languages and Linguistics XXXVI. Leiden: Brill. —. 2008. The Neo-Aramaic Dialect of Barwar. Volume One: Grammar. Handbook of Oriental Studies. Leiden: Brill. Lambrecht, Knud. 1998. Information Structure and Sentence Form: Topic, Focus, and the Mental Representations of Discourse Referents. Cambridge Studies in Linguistics, 71. Cambridge: Cambridge University Press. Molnár, Valéria. 2002. “Contrast—From a Contrastive Perspective.” In Information Structure in a Cross-linguistic Perspective, edited by Hilde Hasselgård et.al., 147– 161. Amsterdam: Rodopi. Ritter, Helmut. 1990. Ṭūrōyo. Die Volkssprache der syrischen Christen des Ṭūr ‫ﻉ‬Abdîn. Grammatik. Pronomen, „sein, vorhanden sein“, Zahlwort, Verbum, edited by Rudolf Sellheim. Schriften der Wissenschaftlichen Gesellschaft an der JohannWolfgang-Goethe-Universität Frankfurt am Main: Geisteswissenschaftliche Reihe, 6. Stuttgart: Steiner. de Saussure, Ferdinand. 1994. Cours de linguistique générale. Ed. crit. Paris: Payot. Van Valin, Robert D., and LaPolla Randy J. 1997. Syntax. Structure, Meaning and Function. Cambridge Textbooks in Linguistics. Cambridge: Cambridge University Press. Wilke, Oliver. 2008. Verwendung und Funktion der verbalen Kategorien des Ṭuroyo. Unpublished Master’s Dissertation. Philipps-University of Marburg.

THE TURKISH LEXICAL INFLUENCE ON

ṢŪRAYT/ṬŪRŌYO : A PRELIMINARY SELECTION OF 1

EXAMPLES

2

AZIZ TEZEL This article deals with a considerable number of Turkish lexical elements which have entered the Ṣūrayt/Ṭūrōyo language at different periods, mostly from 1940s onwards, some of the elements being taken over from Kurdish and Arabic dialects. The article also gives an account of some of the Ṣūrayt/Ṭūrōyo lexical elements which had been in use before the Turkish elements were borrowed, demonstrating thus some important changes that took place within the Ṣūrayt/Ṭūrōyo lexicon. Further, the article suggests some etymological solutions, which are important not only for Ṣūrayt/Ṭūrōyo, but also for some other languages in the region. This is especially true of some cases of homonymy.

1. SOME WORDS PERTAINING TO CURRENCY

The word for ‘money’ in Ṣūrayt/Ṭūrōyo is källä (m.). I do not know of such a word for ‘money’ in the neighbouring languages. Yet the probability is that we have to do with a borrowing, as is clear from the form of the word in question. The Turkish word for ‘head’, kelle, which also means ‘a sugar loaf; a cake (of cheese)’, may be of 1

The name Ṣūrayt/Ṭūrōyo replaces Western Neo-Syriac, which I have used previously. In connection with the preparation of this article I have some informants to thank: Zeki Aydin (ʿĪwardo), Gabriel dbē-Malke Masʿəd (Mīdən), my brothers Görgis and Muqsi, the brothers Aziz and Yaʿqub dbē-Masʿəd (Mzīzaḥ), Šamʿən Türkmen (Bsōrīno) and Zeki Demir (Ḥabsəs). I made a tape-recorded interview with Zeki Aydin and Gabriel. From the others, I have recorded many words and phrases when they spoke spontaneously. 2

70

AZIZ TEZEL

significance.3 To be sure, the Turkish kelle does not provide such a sense as that of our källä, but it may have been used in combination with the Turkish word for ‘money’, para, thus: *kelle parası. Its meaning could then have been transferred to kelle; cf. Turkish başlık parası ‘a sum paid to the brideʼs father by the bridegroom’, where başlık may be used independently with the same meaning. It should be pointed out, however, that the stem baş- is the normal Turkish word for ‘head’ and the semantic connection between ‘head’ and ‘capital’ is known from other languages, for example, Arabic raʾsu l-māli ‘capital’, not to mention the English ‘capital’ itself.4 An alternative solution may be sought in an Arabic loanword in Turkish, namely galle ‘income, rent, yield’, a term used in Ottoman Turkish within canon law. It is a borrowing from Arabic ġalla(t);5 cf. Syriac ʿǝlaltā ‘ingathering; harvest; yield’. Their respective roots ġ-l-l and ʿ-l-l are cognates in a Semitic context. An eventual change galle > källä is, after all, not impossible; cf. Kurdish kalle in İzoli, where it is said to have the meaning ‘tithe’ in the ašīrat-system.6 The word for ‘money’ in general among the purists nowadays is zūze (pl.), introduced recently into the language from Syriac zūzē. A compound, namely manqalōt,7 which denotes a kind of ‘Turkish money’, reflects a borrowing either via Kurdish or directly from Turkish banknot, from English banknote.8 It is mainly known to the older generation. Its second part occurs independently and unchanged only in the plural, thus: nōtat, in which -at is a plural morpheme of Arabic origin. So also the Ṣūrayt/Ṭūrōyo word for ‘a piaster’, qǝrš/qurūš goes back to Turkish kuruş, found also in some Arabic dialects and in Kurdish. The Turkish kuruş, occuring also in the form guruş, itself is said to be a borrowing from German Groschen.9 The Turkish ‘lira’ (of Italian origin) is of course also used by Ṣūrayt/Ṭūrōyospeakers as līra. However, Ṣūrayt/Ṭūrōyo also has another word for denoting ‘a lira of 100 piasters’, namely wāraqto, coming ultimately from Arabic waraqa(t) (actually

3

The Turkish kelle is of Persian origin, according to Türkçe Sözlük (2005: 1131a). For details, see Tezel (2003: 226). 5 See Redhouse (1999: 382b). 6 See İzoli (1992: 220a). Note that the author in his transcription makes a distinction between kalle ‘tithe’, and kelle ‘head’ (for the latter, see ibid.: 227a). 7 The change banknot > manqalōt is suggestive of b  m, by regressive assimilation at distance, and n  l, by progressive dissimilation at distance, see Tezel (2003: 216). 8 See Türkçe Sözlük (2005: 197b). 9 For such a statement, see Türkçe Sözlük (2005: 1266a). For the form guruş, see Redhouse (1999: 417a). 4

THE TURKISH LEXICAL INFLUENCE ON ṢŪRAYT/ṬŪRŌYO

71

‘a single leaf’), which in the form varaka was also borrowed into Turkish. Its root w-r-q is in a Semitic context cognate with Ṣūrayt/Ṭūrōyo y-r-q ‘to become green’. An older word mäǧīdi ‘a Turkish silver coin of 20 piasters, coined under Sultan Abdul-Mejid’ is to be found in Turkish, Arabic and Kurdish. Its forms (with the same meaning) in Turkish are mecit, mecidiye, but cf. also Turkish mecidî ‘pertaining to the Sultan Abdul-Mejid’.10 Arabic in general seems to have maǧīdī, with the same meaning.11 So too Ṣūrayt/Ṭūrōyo qämäri ‘a piaster, penny’12 is a borrowing either from Turkish kamerî or Arabic qamarī ‘of or pertaining to the moon, moon-shaped, lunar’, the Turkish form being a borrowing from the Arabic one.13 In Sabar, qamari ‘an Ottoman coin’, which is the same word, is considered Arabic.14 Neither in my sources of Turkish nor in those of literary Arabic could I find the sense in question under Turkish kamerî and literary Arabic qamarī. However, the sense concerned is indicated sub Levantine Arabic qamarī in Barthélemy.15 There remains the Ṣūrayt/Ṭūrōyo čärxi ‘a coin of five piasters’, which occurs also in Kurdish and Jewish Neo-Aramaic. On the Kurdish çerxî, Chyet quotes Sir James W. Redhouse’s Turkish and English Dictionary saying: “(provincial, perhaps Kurdish) a five piasters piece”.16 According to Sabar, it is a Turkish word, but he does not indicate its Turkish etymon.17 I have failed to find a Turkish word with such a form and meaning in my sources of Turkish, which points to a changed word. To be sure, in form this word in the region agrees with Persian čarxī, which, among other things, means ‘spherical; celestial’, but these and other meanings indicated under this word do not provide us with the sense we need. I, for my part, am inclined to seek the origin of this word rather in the Persian loanword in Turkish çeyrek (older çaryek < Persian čahār ‘four + yak ‘one’) ‘a quarter, one fourth; quarter of an hour; five piastres’.18 The last sense gives us the solution we need. Its literal meaning is ‘one

10

See Redhouse (1999: 743b). See Wehr (1976: 893b). 12 The Ṣūrayt/Ṭūrōyo qämäri today occurs only in some expressions, for example, lō kōṭōwe qämäri yāqīḏo ‘he is not worth a burnt penny’. 13 Cf. also qrān or qirān, which, according to Maclean (1901: 286a), denoted ‘five piaster’. 14 See Sabar (2002: 280b). 15 See Barthélemy (1935–1969: 681). 16 See Chyet (2003: 106b). 17 See Sabar (2002: 133a). 18 See Redhouse (1999: 250b). 11

72

AZIZ TEZEL

fourth’,19 actually ‘one fourth’ of the Turkish word mecidiye quoted above.20 Its literal meaning has thus nothing to do with ‘five’. Provided that we are on the right track as regards the origin of this word, the word in question would occur as a doublet in Ṣūrayt/Ṭūrōyo, Jewish Neo-Aramaic and Kurdish, having also čārīk, c̆ārık, çarêk ‘quarter of an hour’, respectively; cf. Iraqi Arabic čārak, with the same meaning. If not a reflex of the Persian loanword in Turkish çaryek, then the above-mentioned Persian čarxī might be its etymon.

2. SHOES

Certain words denoting ‘shoes’ in Ṣūrayt/Ṭūrōyo are either of Turkish origin, or came into the language directly or indirectly via Turkish. An old word is čārǝx/čārōx/čārōxo, which denotes ‘a sandal made of rawhide and tied up by a shoelace round about’ (itself made of rawhide). It goes back to Turkish çarık21 with the same meaning, which is also to be found in Kurdish and some neighbouring Arabic dialects.22 The normal word for ‘the modern shoe’ among Ṣūrayt/Ṭūrōyo-speakers in or from Turkey nowadays is qundara/qundrāye, coming from Turkish kundura. It is considered a Greek loanword in Turkish.23 Turkish has also another more common word for ‘shoe’, namely ayakkabı, which has not been borrowed into Ṣūrayt/Ṭūrōyo, to my knowledge. The purists among Ṣūrayt/Ṭūrōyo-speakers today make use of msōne (pl.) (< s-ʾ-n) for denoting ‘the modern shoe’. It has recently been introduced into Ṣūrayt/Ṭūrōyo from Syriac. It is derived from the same root as that of Ṣūrayt/Ṭūrōyo ṣʿūno (< *sʾūno), ‘a sandal’, which seems to be one of the few genuine Ṣūrayt/Ṭūrōyo words among the words denoting different sorts of shoes.24

19

Losing its literal meaning, the word was corrupted, perhaps for the first time in Kurdish, from which then it was introduced into Ṣūrayt/Ṭūrōyo. The loss of the literal meaning may also be behind the Turkish changed form çeyrek. 20 For such a statement, see Türkçe Sözlük (2005: 421a). 21 Compare Arabic širāk ‘shoelace’? 22 See, among others, Chyet (2003: 100a); Barthélemy (1935–1969: 133). 23 See Redhouse (1999: 685b). 24 In Syriac, too, several words denoting various sorts of shoes appear to be of foreign origin. Thus Syriac ṭellārā ‘sandals’ is assumed to be a borrowing from Latin talaria. With Latin cortex is connected Syriac qarqā ‘thin sandals’, see Brockelmann (1982: 700a). A Syriac word with different spellings sedlā/sadlā/sandālā/sendālā ‘a sandal’ ultimately comes from Greek

THE TURKISH LEXICAL INFLUENCE ON ṢŪRAYT/ṬŪRŌYO

73

The normal word for ‘a boot’ among the same speakers is bōṭ, which has been taken over from Turkish bot, borrowed from French botte.25 From Turkish also comes the word ǧäzme/čǝzma (the latter in the dialect of Mǝḏyaḏ) ‘top-boot; Wellington’, which reflects Turkish çizme. The form ǧäzme may have been taken over from Kurdish. We also note the occurrence of poṭīne ‘soldiers’ boot’, which came into the language from Turkish potin or via Kurdish, where the same word is to be found. The Turkish potin itself is a borrowing from French bottine.26 Many Ṣūrayt/Ṭūrōyo-speakers in or from Turkey nowadays make use of tärlǝk/e for denoting ‘slippers’, from Turkish terlik. The older Ṣūrayt/Ṭūrōyo words, with the same meaning, are mǝ(š)šāye and šǝmǝke.27

3. FABRICS

All words denoting different sorts of ‘fabrics’ in Ṣūrayt/Ṭūrōyo seem to be of foreign origin. The word for ‘fabric’ in general is qumaš, which is of Arabic origin, from Arabic qumāš, found also in Turkish, Persian and Kurdish. Both the meaning and phonological shape of this word make it difficult to determine whether it came into Ṣūrayt/Ṭūrōyo directly from Arabic or via some other neighbouring language.

σανδάλιον. Of Persian origin is Syriac mūqā ‘a shoe, slipper’; hence Arabic mūq, according to Brockelmann (1982: 377a). 25 See Türkçe sözlük (2005: 304a). 26 See Türkçe sözlük (2005: 1623b). 27 The words mǝ(š)šāye and šǝmǝke, too, are of foreign origin, as is evident from their forms. The former occurs mainly in the dialect of Mǝḏyaḏ and the latter in the village dialects. The former is a borrowing from Arabic maššāya(t), with the same meaning, see Baalbaki (1991: 1045a); see also Thesaurus Syriacus sub Syriac kssʾ (1981: col. 1778). If it is derived from the Arabic root m-š-w/m-š-y ‘to go on foot, walk’, it has then a good etymology. The latter, šǝmǝke, which also occurs in Kurdish, is of obscure origin. Considering its phonological shape and the identity in meaning between it and mǝ(š)šāye it is tempting to consider it to be one and the same word. However, the question remains whether it could have undergone such sweeping changes. An interesting case, which is met with in Thesaurus Syriacus (1981: col. 1382) sub Syriac ḥraftā, is šmšk (unvocalized). It seems to be a synonym of Arabic xuff ‘a slipper’. It is presented as an Arabic word al-šmšk. Could this tentatively be read as šamšik/šimšik? A loss of one of two occurrences of š would give us the exact solution we need as far the form of the word is concerned, with the -e in šǝmǝke as the Kurdish oblique ending.

74

AZIZ TEZEL

The same is true of another word, namely qädīfa ‘velvet’, which ultimately comes from the literary Arabic qaṭīfa(t), by dissimilation of ṭ > d under the influence of q.28 It also occurs in Turkish, Kurdish and Iraqi Arabic in the forms kadife, qedife and qadīfa, respectively, with more or less the same pronunciation; cf. Syriac qaṭṭīftā ‘a shaggy cloak or rug’ and Syriac maqṭǝfā, maqṭaftā ‘a covering, wrapper’. The word for ‘chintz’, čīt, is to be found in several neighbouring languages, where Turkish has çit, Iraqi Arabic çīt, Kurdish çît, Jewish Neo-Aramaic çīti,29 and Persian čīt. The Turkish çit is considered a Persian loanword in Redhouse Sözlüǧü, a statement not found in Türkçe Sözlük, which notes it as a colloquial word.30 The Persian čīt is considered a Hindi word.31 The same seems to be true of the English ‘chintz’, which comes from Hindi chint, from Sanskrit citra ‘gaily-coloured’, according to Collins Dictionary.32 It seems that the word in question found its way into the region via Persian, where the -n in the borrowed Hindi form chint was dropped in the doubly closed syllable. The loss of the n in this word came to lead to a homonymy in the Turkish lexicon with the word çit ‘hedge; fence of hurdles; fence’; hence Kurdish çît and Ṣūrayt/Ṭūrōyo čīt, where, too, the word constitutes a homonym. So also the word for ‘satin’, Ṣūrayt/Ṭūrōyo ʾaṭlaṣ/ʾačlaṣ, is to be found in several neighbouring languages, where its form in Arabic and Persian is ʾaṭlas and in Turkish atlas, the Arabic form being the original one. It is traditionally considered a derivative of the Arabic root ṭ-l-s. In Ṣūrayt/Ṭūrōyo we also meet with the word čāpān, which denotes a sort of ‘American linen’. Such or a similar word is to be found in several languages in the region. A dialectal Turkish word, namely çapan, is to be found in Derleme Sözlüǧü;33 cf. also Turkish çepken ‘a kind of short overcoat with wide sleeves’?34 In Anatolian 28

The dissimilation in question seems to have taken place in some neighbouring language, from which the word concerned was then taken over into Ṣūrayt/Ṭūrōyo. As the literary Arabic qaṭīfa(t) manifests the change ṭ > d in several neighbouring languages, such as Turkish, Kurdish and Arabic dialects, it is difficult to determine the immediate source of the Ṣūrayt/Ṭūrōyo qädīfa. Perhaps the change in question first took place in Ottoman Turkish, from which then it reached the neighbouring languages. 29 Sic! with final -i in Sabar (2002: 131b), connecting it with Iraqi Arabic and Turkish, derived from Hindi. 30 See Redhouse (1999: 257b); Türkçe Sözlük (2005: 440a). 31 See Steingass (2006: 405b). 32 See Collins (1991: 284a). 33 See Derleme Sözlüǧü (1963–1977: III, 1073b). 34 For the Turkish çepken, see Hony (1957: 64b). The word in question is not indicated in Türkçe Sözlük.

THE TURKISH LEXICAL INFLUENCE ON ṢŪRAYT/ṬŪRŌYO

75

Arabic we note čāpūn ‘grobes Leinen, leinernes Tuch’,35 with which Sabar connects Jewish Neo-Aramaic ǧāpūn, čāpān ‘coarse linen (used for traditional underwear)’, suggesting as to its origin: ‘EN Japan?’36 The right solution seems, however, to be found in Barthélemy, where we find čabbōn and ǧabbōn ‘jupon’, which he considers a French word.37 Finally, we shall mention an older word in Ṣūrayt/Ṭūrōyo, šūqo, which denotes a kind of ‘linen fabric’, known to the older generation. There is in the neighbouring languages a similar word. Thus, Arabic has ğūx, with the meaning ‘broadcloth’, and Turkish possesses çuha/çuka, explained by Turkish tüysüz, ince, sık dokunmuş yün kumaş, which approximately means ‘smooth, fine, and closely woven woollen fabric’.38 The Arabic word itself is said to be a borrowing from Turkish one.39 According to Bustānīʼs Muḥīṭ al-muḥīṭ, the original form of the Arabic form ğūxa(t) is çūqa,40 whose pronunciation seems to agree with the Turkish form çuka. Our sources on Turkish consider the Turkish çuha a borrowing from Persian čūxa.41 The Kurdish lexicon indicates čox/čoxe/čox, with the same or similar meaning. Both the meaning and the form of this word in the neighbouring languages bear thus a close resemblance to those of Ṣūrayt/Ṭūrōyo šūqo, which came to coincide with another Ṣūrayt/Ṭūrōyo word, namely šūqo ‘a street; the centre of the city’.42

4. MEANS OF CONVEYANCE, ‘STOP’, ‘HALT’ AND ‘STAND’

The majority of modern words denoting means of conveyance among the Ṣūrayt/Ṭūrōyo-speakers are of foreign origin. To the older words belongs Ṣūrayt/Ṭūrōyo käläk, which denotes ‘raft of inflated skins’. It is the same word as Arabic kalak, Turkish kelek and Kurdish kelek, with the same meaning. The Arabic kalak is the original form, while the others are borrowings. The Arabic itself seems to be a borrowing from Syriac kalkā/klakkā,43 which in

35

See Vocke and Waldner (1982: 99). See Sabar (2002: 126a). 37 See Barthélemy (1935–1969: 103, 134). 38 For the explanation in Turkish, see Türkçe Sözlük (2005: 454a), having only the form çuha. For the form çuka, see Tarama Sözlüǧü (1974: II, ). 39 See Kazimirski de Biberstein (1860: I, ); Barthélemy (1935–1969: 139). 40 See Bustānī (1983: ). 41 See Redhouse (1999: 262a); Türkçe Sözlük (2005: 454a). 42 For details, see Tezel (2003: 120–122). 43 See Fraenkel (1886: 220). 36

76

AZIZ TEZEL

turn is a borrowing from Akkadian kalakku.44 As the Ṣūrayt/Ṭūrōyo käläk does not agree in form with the Syriac kalkā/klakkā, it is to be considered a foreign word, whose immediate donor language is difficult to determine. The normal word for ‘a ship’ among the Ṣūrayt/Ṭūrōyo-speakers in or from Turkey is gämīye, coming from Turkish gemi, from which is borrowed also Kurdish gemî, with the same meaning. Syriac ʾelpā and sfī(n)tā, which both denote ‘a ship’, have not survived in Ṣūrayt/Ṭūrōyo, to my knowledge. The purists nowadays occasionally make use of these words, although in the forms ʾelfo and sfīto, but they are of rare occurrence. The modern word wāpūr/wāpōr ‘steamer’ probably came into the language via Turkish vapur, itself of French origin; cf. semantically Arabic bāxīra, which occurs among Ṣūrayt/Ṭūrōyo-speakers in or from Arabic-speaking countries. From Turkish is also borrowed qayīke ‘caïque’, from Turkish kayık.45 Western Syriac qarqūrō (cf. Greek ϰέρϰουρος and Arabic qūrqūr46) has the same or a similar meaning. I do not, however, know of its occurrence among the purists. An importation of this word into Ṣūrayt/Ṭūrōyo would lead to a homonym, as the language already possesses qarqūro, which denotes ‘a lamb’; cf. dialectal Arabic qarqūr, with the same meaning, which al-Munğid considers a borrowing from Syriac.47 Going on to the word for ‘a car’ we note that Ṣūrayt/Ṭūrōyo-speakers in or from Turkey use ʿaraba (f.), which is the same word as the Turkish araba and Arabic ʿaraba(t). It is hard to determine whether Ṣūrayt/Ṭūrōyo ʿaraba is a direct borrowing from Arabic or a borrowing via Turkish. The fact that it begins with the sound ʿ, as does the Arabic ʿaraba(t), is no guarantee for its being a direct borrowing from Arabic.48 One reason for regarding it as a borrowing from Turkish is that the Turkish 44

See von Soden (1965–1981: I, 423b). According to Collins (1991: 226b), the English caïque is a borrowing from French, from Italian caicco, from Turkish kayık. 46 See Fraenkel (1886: 217). 47 I have failed to find such a word with such a meaning in Syriac. 48 It is not quite clear from our sources whether the Turkish araba is a borrowing from the Arabic ʿaraba(t), or, vice versa, the latter is a borrowing from the former. As for the former, both Redhouse (1999: 67b), and Türkçe Sözlük (2005: 110a), indicate it as it were a Turkish word. Redhouse Sözlüǧü makes a distinction between Turkish araba 1. ‘car’ and araba 2. ‘pontoon used in bridges on the Tigris’. The latter is considered Arabic. The question is whether such a distinction is correct as far as the etymology of the word is concerned. According to Lane (2003: 1994a), the meaning ‘a wheel-carriage of any kind’ under the Arabic word ʿaraba(t), with a pl. ʿarabāt ‘certain stationary vessels that used to be in the Tigris’, is postclassical. From this, one may infer that the Arabic word in question is etymologically the same, 45

THE TURKISH LEXICAL INFLUENCE ON ṢŪRAYT/ṬŪRŌYO

77

araba is the normal word for ‘a car’ in Turkish, while Arabic besides ʿaraba(t) nowadays also has sayyāra(t); hence the Ṣūrayt/Ṭūrōyo sa/ǝyyāra, used by the Ṣūrayt/Ṭūrōyo-speakers in or from Arabic-speaking countries. The purists make use of a neologism rāḏ̣ayto (root: Syriac r-d-y ‘to go’). Also words such ʾōtōbus ‘bus’, dolmuš ‘minibus’, q/kamyōn ‘lorry’, which are in use among Ṣūrayt/Ṭūrōyo-speakers in or from Turkey, are borrowings from Turkish or via some other neighbouring language.49 The same is true of the word for ‘a train’, trēn, which is derived from the French loanword tren in Turkish. The corresponding word among Ṣūrayt/Ṭūrōyospeakers in or from Arabic-speaking countries is qiṭār, from Arabic qiṭār. From this is ultimately derived a Ṣūrayt/Ṭūrōyo neologism qṭōro, which is the word used by the purists. The word for ‘an airplane, aircraft’, ṭa(y)yāra, ultimately comes from Arabic ṭayyāra(t), found also in Turkish as a borrowing in the form tayyare, which nowadays has been replaced by the native uçak. The latter has not been borrowed into Ṣūrayt/Ṭūrōyo, as far as I know. The word for ‘an airplane, aircraft’ among the purists is ṭāyosto/ṭāyəsto, a neologism formed from Syriac ṭ-w/y-s ‘to fly’. Also almost all words denoting ‘stop, halt and stand’, being in use among Ṣūrayt/Ṭūrōyo-speakers in or from Turkey, have been taken over from Turkish. Thus, the normal word for ‘a stop, stand’ (of a car, bus, taxi) is dūraq, from Turkish durak. Its equivalent among Ṣūrayt/Ṭūrōyo-speakers in or from Arabic-speaking countries is mawqǝf (root: w-q-f). Both words may easily be replaced by the Ṣūrayt/Ṭūrōyo dukṯo dū-klōyo or dukṯo dū-nṭōro.

with a new meaning. According to al-Munǧid (1975: 495c), this word, with the meanings quoted, in Arabic is of Greek origin, without an indication of the Greek etymon in question. In this connection I can also note Syriac ʿarbā, with four meanings, of which ‘a water-wheel’ or ‘mill’, see Payne Smith (1903: 427a), is interesting as compared to the meanings of the Arabic ʿaraba(t). An eventual relationship between the words quoted in the region demands a further thorough study. 49 Another word for ‘lorry; automobile’, namely Ṣūrayt/Ṭūrōyo ṭrǝmbēl, is not attested in Turkish in the form in question but rather in some other neighbouring languages; cf. Iraqi Arabic ṭrumbēl, ṭrāmbēl, Woodhead and Beene (2003: 289), Kurdish trimbêl, Chyet (2003: 628a), and Jewish Neo-Aramaic ṭrambēl, Sabar (2002: 175b). The origin of this is the word ‘automobile’, as has already been stated in Chyet. It seems that the first syllable au- was dropped, whereupon a secondary r has been added, thus: automobile > *tomobile > ṭrumbēl ~ ṭrǝmbēl.

78

AZIZ TEZEL

The word for ‘port, harbour’ among the same speakers is līman, which is a borrowing from Turkish liman, itself from Greek λιμήν; hence also Syriac lǝmīnā (lmʾnʾ) and Arabic mīnāʾ, where the first syllable *li- is dropped. Ṣūrayt/Ṭūrōyo-speakers in or from Arabic-speaking countries use the latter. The use, if any, of Syriac lǝmīnā (Western Syriac lmīnō) among the purists is rare. It may without any difficulties be introduced into Ṣūrayt/Ṭūrōyo, as it bears close resemblance to Turkish liman and the Arabic mīnāʾ.50 The normal word for ‘an airport’ among Ṣūrayt/Ṭūrōyo-speakers in or from Turkey is hawā-ʾalāni, which in meaning corresponds to maṭār, used by the Ṣūrayt/Ṭūrōyo-speakers in or from Arabic-speaking countries. The former is borrowed from Turkish hava alanı, and the latter from Arabic maṭār. The equivalent word among the purists is bēṯ ṭawṣo, which is a neologism. The word for ‘station’ among Ṣūrayt/Ṭūrōyo-speakers in or from Turkey is stāsyōn/stīyāson, which is a borrowing from Turkish istasyon, itself being a borrowing from French. Another word with the same meaning is mḥaṭṭa (< Arabic maḥaṭṭa(t), root ḥ-ṭ-ṭ), used by Ṣūrayt/Ṭūrōyo-speakers in or from Arabic-speaking countries. No common neologism with the same meaning among the purists is familiar to me. One may introduce a neologism such as bēṯ māšǝryo dū-qṯōro or only bēṯ qṭōro. The normal word for ‘a way’ and ‘a motor road’ is the Ṣūrayt/Ṭūrōyo word for ‘way’, namely därbo, and därbo dī-ʿaraba or dī-rāḏ̣ayto. The Turkish word for ‘way’, yol, is borrowed only in compounds, for example, the word for ‘crossroads’ is termed dörtyol (known from dörtyol in Mǝḏyaḏ), as is the case in Turkish. I do not know of any use among the purists of Western Syriac fōlšaṯ ʾūrḥōṯō, with the same or a similar meaning. The word for ‘a bridge’, köpri,51 is a borrowing from Turkish köprü. The corresponding term used by Ṣūrayt/Ṭūrōyo-speakers in or from Arabic-speaking countries is ǧǝsǝr, which is to be compared to Western Syriac gīšrō; hence today Ṣūrayt/Ṭūrōyo gəšro, used by the purists to denote ‘a bridge’; cf. the genuine Ṣūrayt/Ṭūrōyo gušro ‘a plank’, a sense found already with Western Syriac gīšrō.

50

Note that there are two homonyms in the form mīnāʾ in the Arabic lexicon. The other mīnāʾ, which has meanings such as ‘glaze, glazing; enamel’ (of the teeth), is of Persian origin, according to Fraenkel (1886: 232). The same word is to be found in Turkish in the form mine. I have noted its occurrence also among Ṣūrayt/Ṭūrōyo-speakers in Istanbul. 51 Also Kurdish pǝr, with the same meaning, is used, especially among the older generation.

THE TURKISH LEXICAL INFLUENCE ON ṢŪRAYT/ṬŪRŌYO

79

5. SOME TURKISH LOANWORDS PERTAINING TO THE ELECTORAL SYSTEM

The normal word for the ‘election’ among the Ṣūrayt/Ṭūrōyo-speakers in Turkey nowadays is säčīm, from Turkish seçim. In meaning, it corresponds to intixāb, which is an Arabic borrowing, used mainly by Ṣūrayt/Ṭūrōyo-speakers in or from Arabicspeaking countries. As a replacement for both foreign terms, the purists have introduced g/ġūbōyo from Syriac into Ṣūrayt/Ṭūrōyo. The normal word for ‘vote’ among the same speakers nowadays is ʾōy, coming from Turkish oy. The older word for signifying the same meaning is rāy, found also in Turkish in the form rey, from Arabic raʾy.52 The purists nowadays use the word for ‘sound, voice’ qōlo, also in the sense ‘vote’ within the electoral system; cf. Arabic ṣawt, with the same senses. The word for ‘a candidate’ among Ṣūrayt/Ṭūrōyospeakers in Turkey nowadays is the same as Turkish aday. The words for ‘majority’ and ‘minority’ among the same speakers nowadays are čōġunluq and ʾazǝnlǝq, from Turkish çoğunluk and azınlık, respectively. The equivalent words used by the purists are sāgīyūṯo and qālīlūṯo which have been introduced into the language from Syriac. They are, however, not of frequent occurrence. The word for ‘party’ is partīye, which has been taken over either from Turkish parti, or has entered the language via Kurdish. This word may easily be replaced by Ṣūrayt/Ṭūrōyo gābo ‘side; party’, which is also the word used by the purists in the diaspora for denoting ‘a party’.

6. SOME TURKISH LOANWORDS PERTAINING TO SCHOOL

Before introducing the Turkish school system into Ṭūr ʿAbdīn, the school system in the area consisted of a religious Christian school. This was and still is termed mädräsä/e, the same word as Turkish medrese, ‘a Muslim theological school, a religious boarding school associated with a mosque’, from Arabic madrasa(t), which is the normal word for ‘a school’. The normal word for ‘a modern school’ in Turkey today is okul, which compared to an older word mekteb/p, is a neologism, the former

52

In the dialect of Mīdǝn one finds rā ‘a vein (of body)’ as a synonym of the native wārīḏo. When occurring in genitive constructions, it becomes a homonym with rāy ‘a vote’. The word rā seems to have been taken over from Kurdish ra, reh, which is to be connect with Persian reg, which is also to be found in Ottoman Turkish as a borrowing (for reg in Ottoman Turkish, see Redhouse (1999: 952.)). If the Persian represents a more original form, this means that the sound -g has been dropped in the Kurdish forms.

80

AZIZ TEZEL

being a native Turkish word and the latter an Arabic loanword in the Turkish language. Ṣūrayt/Ṭūrōyo-speakers in Turkey usually make use of mäktäb as the normal word denoting ‘a Turkish school’. The Turkish word okul alone has not got a foothold in Ṣūrayt/Ṭūrōyo. However, it is used in compounds such as ilk okul ‘primary school’, and orta okul ‘middle school’, in which the sound k is pronounced as q in Ṣūrayt/Ṭūrōyo. The purists in the diaspora have introduced a Syriac word madrašto ‘a school’. Going from the word for ‘a school’ on to the word for ‘a teacher’, we find a situation similar to that of the word for ‘a school’. ‘A teacher’, who teaches in a mädräsä/e is termed ma/ǝlfōno, which is a native word, while ‘a teacher’, who teaches in a Turkish school is termed mʿa(l)lǝm. This, to be sure, ultimately comes from Arabic muʿallim, but influence from the Turkish form muallim is a possibility that cannot be ruled out. That the Ṣūrayt/Ṭūrōyo mʿa(l)lǝm has the sound ʿ, as does the Arabic, it is not unequivocal evidence for it being a direct borrowing from Arabic.53 The Turkish muallim is the older word for ‘a schoolteacher’, being today replaced by a native Turkish word öğretmen, which has not got a foothold in Ṣūrayt/Ṭūrōyo. The word for ‘class’ (in a school) is ṣǝnǝf, which has been taken over from Turkish sınıf. This itself is taken from Arabic ṣanf, ṣinf ‘kind, sort, class’, which is to be kept separate from Arabic ṣaff, which in Arabic today is the normal word for ‘a form, grade (in a school)’. The former is derived from the root ṣ-n-f and the latter from the root ṣ-f-f. The purists in the diaspora use sədro, with the same meaning, from Western Syriac seḏrō. It is natural that many of the words denoting school equipment have come into Ṣūrayt/Ṭūrōyo directly from Turkish or via some other neighbouring language. Thus the word for ‘a pen, pencil’, qalam, and that for ‘a penknife; pencil-sharpener’, qalamtǝrāš, that for ‘exercise book’, däftär, and that for ‘a rubber, eraser’, sǝlgīye, have their equivalents in the Turkish kalem, kalemtıraş (< Arabic qalam + Persian terāš),54 defter and silgi, respectively. The words kalem and defter themselves are foreign words in Turkish. They are also to be found in Arabic and other neighbouring languages. It is thus difficult to determine the immediate donor language. Traditionally, they are considered to be of Greek origin. However, the immediate sources of

53

This is of course about the situation among Ṣūrayt/Ṭūrōyo-speakers in or from Turkey. The word mʿalləm among Ṣūrayt/Ṭūrōyo-speakers in or from Arabic-speaking countries is a direct borrowing from dialectal Arabic. 54 See Türkçe Sözlük (2005: 1044a).

THE TURKISH LEXICAL INFLUENCE ON ṢŪRAYT/ṬŪRŌYO

81

the Turkish kalem and defter are Arabic qalam and daftar, according to our sources of Turkish. Also the word for ‘blackboard’, and that for ‘chalk’, täxtä/täxtāye and täbäšīr, respectively, found their way into Ṣūrayt/Ṭūrōyo either via Turkish or via some other neighbouring language. In Turkish they are termed tahta and tebeşir, both being of Persian origin. Words such as qänyo ‘pen, pencil’, kǝrko ‘exercise book’, lāḥayto ‘rubber, eraser’, lūḥo ‘the blackboard’, are used by the purists. They have been introduced into the language from Western Syriac. The words taḥṭīl ‘holiday’ and ʾǝzǝn ‘permission’ have probably been taken over from Turkish tatil and izin, respectively. Both are of Arabic origin, deriving from Arabic taʿṭīl and ʾiḏn.

7. SOME LOANWORDS PERTAINING TO BUILDING AND HOUSES

We note that some words pertaining to building and houses are Turkish borrowings of a recent date. The linchpin of the older building materials consist of kēfe ‘stones’, kalšo ‘lime’, nqarto/nqurto ‘limestone’, ṭīno ‘clay’, which—besides nqarto/nqurto55— have their equivalents in Western Syriac kēfē, kalšō and ṭīnō, respectively. Among the words pertaining to building materials of today we also can note čmǝnto ‘cement’, qūm ‘sand’, brīkēt ‘briquette’, tuġla ‘brick’, which are taken over from Turkish çimento (from Italian), kum, briket (briket, from French) and tuğla (from Greek), respectively. ‘A mason, bricklayer; a master workman’ is termed hosta, which may have been taken over from Kurdish hoste rather than from Turkish usta,56 which ultimately comes from Persian ustād, found also in Arabic as a loanword in the form ʾustāḏ, although rather with the meaning ‘high schoolteacher’. A word bānōyo ‘bricklayer’ has its equivalent in Western Syriac bannōyō; cf. Arabic bannāʾ. The immediate source of word for ‘great buildings under construction’, inšāat, appears to be Turkish inşaat, which in turn, comes from Arabic inšāʾāt, the plural form of inšāʾ.57 55

I have altogether failed to determine the origin of the word nqarto/nqurto, unless it is a derivative of the root n-q-r ‘to hollow out’. 56 Note that the word in question occurs as a doublet in Turkish, namely the changed form usta ‘master; master workman’, and üstad ‘master; teacher; expert’, in which the -d is preserved. 57 The Arabic root is n-š-ʾ, from which is also formed nāšiʾ, which, among other things, means ‘a beginner’. Ṣūrayt/Ṭūrōyo has a debatable word, namely nāši, which occurs in certain

82

AZIZ TEZEL

So also the word for ‘maintenance, care’ (of a house), bāqǝm,58 and that for ‘whitewashing; plaster’, ṣǝwa, and that for ‘repair’, taʿmīr, are recent borrowing from Turkish bakım, siva, tamir respectively, although the word taʿmīr in form and phonological shape is more reminiscent of Arabic taʿmīr, which is the etymon of the Turkish form. The word for ‘plaster’, ṣǝwa, is quite a late borrowing. The native words such as šyōʿo and ṣwōʿo were and still are in use. So too words such as tawān ‘ceiling’, čāti ‘gable roof’ are recent Turkish borrowings, mainly used by Ṣūrayt/Ṭūrōyo-speakers in Istanbul. The native word for ‘ceiling’ is qṭōro. The word for ‘iron’, dämǝr, from Turkish demir, is in the process of superseding ḥadīd and hēsǝn among Ṣūrayt/Ṭūrōyo-speakers in or from Turkey. The former is of Arabic origin and the latter has been taken over from Kurdish; cf. Persian āhen, which was borrowed into Ottoman Turkish. Among other words pertaining to building it should be noted that the word for ‘room’, ʾōda or ʾōdāye, is a recent Turkish loanword, the older word, with the same meaning, was/is mäzäle, which has been taken over from Kurdish, but it is probably ultimately from Arabic manzil.59 No genuine Ṣūrayt/Ṭūrōyo word denoting ‘a room’ seems to have survived. To be sure, ‘an upper-room’ is termed ʿōlīke in some village dialects,60 where its meaning and also partly its form is reminiscent of Syriac ʿellīṯā (hence Arabic ʿilliyya(t)), but ʿōlîk is also the word used in the Kurdish spoken in Ṭūr ʿAbdīn. Its present form in Ṣūrayt/Ṭūrōyo is thus not genuine. The original form must have been *ʿōlīṯō. A native form related to the Syriac word in question is atphrases, for example, latno nāši, ‘I am not gullible’. Ritter (1979: 359) derives it from the same root as Syriac n-š-y (cf. Hebrew n-š-h, Arabic n-s-y) ‘to forget’, which makes a good sense as far as the meaning of the word is concerned. However, there are difficulties in assuming that the form of the word in question is genuine. This is in all probability the same word as nāši in Maclean (1901: 219a), where it is said to have the meaning ‘inexperienced, foolish’. According to the author, it also occurs in Arabic, Persian and Kurdish. The question is whether the Ṣūrayt/Ṭūrōyo nāši is not the same word as Arabic nāšiʾ in the sense ‘a beginner’, used figuratively. Several persons, with whom I have discussed the word in question, consider it a derivative of the root n-š-y, but I am doubtful, for I do not know of any native adjective surviving only in the absolute state in Ṣūrayt/Ṭūrōyo. 58 It is of frequent occurrence among Ṣūrayt/Ṭūrōyo-speakers in Istanbul. 59 See Tezel (2003: 216, n. 92). 60 ‘An upper-room’ in the dialect of Mǝḏyaḏ is termed məḏ̣ḏ̣āra, which comes from Anatolian Arabic manḏ̣ara (for details, see Tezel (2003: 178)), the root of which is Anatolian Arabic n-ḏ̣-r. This word has nothing to do with the Western Syriac word meḏyōrō ‘a story (of a building) a flat’, the root of which is d-w-r/d-y-r.

THE TURKISH LEXICAL INFLUENCE ON ṢŪRAYT/ṬŪRŌYO

83

tested in the plural in the dialect of Mǝḏyaḏ as ʿalyōṯo, which rather denotes a ‘hayloft’.61 Today the purists have introduced Western Syriac qellōytō, although in the form qēlayto, to denote ‘a room’ in general. However, the Syriac word concerned itself seems to be a foreign word of Greek origin, the Arabic qilliyya(t) being a borrowing from Aramaic/Syriac The word for ‘storey’, qāṭ, as in kmō qāṭāṭyo ‘how many stories is it?’ is borrowed from Turkish kat. This became a synonym of ṭābōqo, coming from Arabic ṭābiq. In meaning, they correspond to Western Syriac meḏyōrō, which in Ṣūrayt/Ṭūrōyo occurs among the purists, being introduced into the language from Syriac.

8. SOME BORROWED PARTICLES FROM TURKISH

Ṣūrayt/Ṭūrōyo-speakers in or from Turkey have incorporated into their language several Turkish particles, especially in the form of adverbs and conjunctions, but also in a few cases prepositions. Some examples are as follows. A word of frequent occurrence is ʾanǧaq (< Turkish ancak), with meanings such as ‘hardly, only, at the earliest, at the very most’, as in ʾanğaq kībe ʿǝsri ʾǝšne ‘at the very most he is 20 years old’, qaymīna, ʾanǧaq mōṭīna ‘let us go, we hardly arrive in time’. The same meanings may in Ṣūrayt/Ṭūrōyo be expressed by yā kībe ʿǝsri ʾǝšne yā lō; qūmu, yā mōṭīna yā lō, respectively. Of frequent occurrence is also ʾaṣlǝnda (< Turkish aslında) ‘actually; by rights’, as in ʾaṣlǝnda kul šāto kō-lōzamle bāqǝm ‘actually, it needs care every year’. The base of the word ʾasl- is of Arabic origin, from Arabic ʾaṣl, which also occurs in Ṣūrayt/Ṭūrōyo in forms such as ʾaṣǝl and ʿasǝl, meaning ‘origin; descent’.62 The for-

61

See Beṯ-Ṣawoʿe (1997: 87b). A Turkish loanword kök in Ṣūrayt/Ṭūrōyo, with meanings such as ‘root, fang (of tooth); base, foundation; origin, descent’, is in the process of superseding the word ʾaṣǝl in the sense ‘origin, descent’. It belongs to the Turkish words getting a foothold in the language. In its several senses, it corresponds to Ṣūrayt/Ṭūrōyo wārīḏo (cf. Western Syriac warīḏō), for example, ʾū-kökaṯṯe nāfǝq larwal ‘their (teeth) roots are bare’ may be expressed by ʾaw-wārīḏaṯṯe nāfīqi larwal. In some senses, it differs from wārīḏo, as is the case with its sense ‘foundation’, as in ʾūkök dǝ-Bsōrīno laǧ-ǧǝddōnayḏan māḥatte ‘our forefathers laid the foundations of Bsōrīno’. So also in the sense ‘origin, descent’. Examples: ʾū-kök d-Sabri mē Mīdǝnyo ‘in origin, Ṣabri is from Mīdǝn’, ʾū-kök daččalkōye suryōyo ‘in origin, the Yezidis are Syrians’ (examples from my informants). It is not permissible to say ʾū-wārīḏo dǝ-Bsōrīno but rather ʾū-rǝkǝn, which was the word used in Ṣūrayt/Ṭūrōyo before the introduction of Turkish kök in the sense ‘foundation’. 62

84

AZIZ TEZEL

mer (ʾaṣǝl) is probably a direct borrowing from Arabic, while the latter (ʿasǝl) has been taken over from its borrowed form in Kurdish.63 The normal word for ‘perhaps’ among the speakers in question nowadays is bälki, from Turkish belki, a hybrid from Arabic bel + Persian ki, according to Türkçe Sözlük.64 There are no traces in Ṣūrayt/Ṭūrōyo of Syriac kəḇar, with the same meaning, and I do not know of any serious attempt to introduce it into Ṣūrayt/Ṭūrōyo. In its sense ‘it may happen; maybe’ the word ‘perhaps’ may in Ṣūrayt/Ṭūrōyo be expressed by the native expression kō-we d-hōwe. hīč ‘not at all, never; without’ has got a foothold in the language. It is borrowed from Turkish hiç, which is of Persian origin. Examples of its usage in Ṣūrayt/Ṭūrōyo are: b-mar Gabrīyel šarḥ hīč latwo ‘There was in Mar Gabriel no translation (of Syriac into Ṣūrayt/Ṭūrōyo) at all’; hīč ḥarbūṯo lō saymīwayna ‘We never did mischief’. Sometimes this word also means ‘unimportant’, for example, hāṯe hīč-yo ‘this does not matter; this is quite unimportant’ as a synonym of hāṯe or hāno latyo mēde. In meaning, it normally corresponds to Western Syriac sōḵ, status absolutus of sōḵō, of which there are no traces in Ṣūrayt/Ṭūrōyo as a genuine case. The same hīč in combination with ʾolmaṣṣa, thus hīč ʾolmaṣṣa (Turkish olmasa), has the meaning ‘at least’, which frequently is used by Ṣūrayt/Ṭūrōyo-speakers in Istanbul. The same is true of its synonym ʾän ʾāzǝndan (Turkish en azından). In meaning, the former corresponds to Ṣūrayt/Ṭūrōyo mǝqa dlōwe, and the latter to būnōqūṣo, in which nōqūṣo is of Arabic origin. qōlay qōlay; by repeating the Turkish word kolay ‘easy’, which is of frequent occurrence also as an adjective among the Ṣūrayt/Ṭūrōyo-speakers in or from Turkey, the word is given an adverbial function, for example, qōlay qōlay ʾī-ʿītāṯe lō kūyo ḥḏō ‘so easily can this church not be one’; cf. b-hasāni, with the same meaning, which is an older Ṣūrayt/Ṭūrōyo word borrowed from Kurdish.65 In meaning, it corresponds to dialectal Arabic b-suhūle, used by Ṣūrayt/Ṭūrōyo-speakers in or from The word rǝkǝn, which is from Arabic rukn, came into Ṣūrayt/Ṭūrōyo either from Arabic or via Kurdish. It is still in use, especially among the older generation. The word ʾaṣǝl occurs in a frequent phrase: ʾaṣǝl hātat (mīdǝn ʿasǝl hātǝt), which means ‘it is up to you; it is for you to decide’. 63 The form ʿasǝl occurs at least in the dialect of Mīdǝn. 64 See Türkçe Sözlük (2005: 240b). 65 The Kurdish form is hêsanî. It is also found in the form asan, which is to be compared to Persian āsān, āsānī; hence asan, asanî in Ottoman Turkish. The initial laryngeal h in the Kurdish hêsanî and our hasāni seems to be of a secondary origin (for a comparison of the Kurdish forms to the Persian āsān, see Chyet (2003: 9b); for the Ottoman Turkish forms, see Redhouse (1999: 79b).

THE TURKISH LEXICAL INFLUENCE ON ṢŪRAYT/ṬŪRŌYO

85

Arabic-speaking countries. A possible importation from Syriac could be fšīqūṯo, which with the Ṣūrayt/Ṭūrōyo prep. bǝ- thus: bǝ-fšīqūṯo would have the same meaning. näysä (< Turkish ne ise) ‘anyway; briefly’, for example, näysa, qāyǝm ʾazze, sǝmme qontrol, ḥzalle kāle mū-tansiyon ‘Briefly, he went to (the physician), they examined him (and) saw that (the cause) was the blood-pressure’. The older Ṣūrayt/Ṭūrōyo word with, more or less, the same meaning is w-alḥāṣǝl, which is of Arabic origin and is still in use. yal(ǝ)nǝz (Turkish yalnız) ‘only; but’, for example, yalǝnǝz ʾaḥna kǝzzān lī-gīhāno ‘Only we shall go to Hell’? It functions also as a conjunction, for example, gdōbēnōlux darmōno, yalnǝz kōlōzǝm saymǝt dǝqqa ʾīnaqla dmōḥǝtle l-ʿaynux ‘I will give you medicine, but you must be careful when you spread it on your eyes’. As the older word with the same meanings is Ṣūrayt/Ṭūrōyo bäs, on which see below. ḥaqqätan (< Turkish hakıkaten, of Arabic origin) ‘really; indeed’, for example, ḥaqqatan hawxāwa, bǝšrōro ‘Indeed, so was the case, indeed’. The same meaning may be expressed by the native bǝ-šrōlo; the form šrōro in bǝ-šrōro instead of the dissimilated form šrōlo, occurs among the purists. It reflects Western Syriac šrōrō. dāyīma ‘always’ (< Turkish daima < Arabic dāʾīman), as in būbaytōste dāyīma mṣālīnāwa ‘Also at home we always used to pray’; dāyīma suryōyo mǝǧġōlīwayna ʿam ḥḏōḏe ‘We always used to speak Syriac’. The same meaning may, more or less, be expressed by kul naqla, which literally means ‘every time’. From the same Arabic root is derived the base form däwäm- of däwämli ‘all the time’ (< Turkish devamlı; Turkish devam- < Arabic dawām), as in däwämli ʾaḥna bayn ʾaṭṭāye qāyīmīna ‘All the time we grew up among the Muslims’. This borrowing has thus been taken over from Turkish, as is evident from the form of the word in question. ṭābi (Turkish tabiî < Arabic ṭabīʿī) ‘naturally, of course’ (occasionally only as a filler), as in ṭābi bīdayrōste ḥzēlan ġālābe mēde däyīšǝk66 ‘Naturally, also in the monastery we saw a very different situation’. This is a doublet of ṭābīʿi, which is nowadays mostly used as an adjective, especially among Ṣūrayt/Ṭūrōyo-speakers in or from Arabic-speaking countries. nāṣǝlṣa (Turkish nasılsa) ‘in any case, under all events; somehow or other’, e.g. ʾama hayna dkǝtne dtāmo lō ḥārūwe mǝnne, yāni67 kǝmmi nāṣǝlṣa tahlīka layt härke ‘But

66

The adjective däyīšǝk, too, is a Turkish loanword, from Turkish deǧişik: cf. the borrowed Ṣūrayt/Ṭūrōyo verbal root d-g-š ‘to change’ below. 67 The form of the word yāni ‘that is, namely’, which ultimately comes from Arabic yaʿnī (root ʿ-n-y), is suggestive of a borrowing from the Turkish form yani, the same.

86

AZIZ TEZEL

they do not pay attention to those who are there, namely they say: in any case, there is no danger here’. An equivalent Ṣūrayt/Ṭūrōyo expression is ʾīdärbo dhōwo. gȫya (< Turkish gûya)68 ‘according to reports, from information received’, for example, göya ʾad-dāmāre d-lēbe mqapṭi69 ‘From information received, the arteries of his heart are closed’. A native phrase ʿal ʾū-mamro has the same meaning, which is still in use. This is one of the finest Ṣūrayt/Ṭūrōyo phrases whose use should not be neglected. ʾolāraq (Turkish olarak) ‘as being’, for example, ‘dīsa ʿlaymo olāraq ʾāṯi… läšän dhōwe dayrōyo ‘Again, as a young he came to become monk’. A native word with the same meaning is xud (< *ʾax d). gerči (= Turkish gerçi, of Persian origin) ‘to be sure, it is true that’, as in gerči zlām ṭāwōyo, ʾama ʾīṭābīʿayḏe latyo bāsǝmto ‘To be sure, he is a good man, but his nature is not gentle’. ʿaǧaba (interrogative particle, from Turkish acaba), as in ʿaǧaba gdōṯe ġālābe nōše l-ärke? ‘Will many people come here? The Turkish acaba itself is ultimately of Arabic origin, from the Arabic form ʿaǧaban, or via the Arabic loanword ʿaǧabāʾ in Persian.70 The form and the usage of the Ṣūrayt/Ṭūrōyo word in question point to a borrowing from Turkish. The older Ṣūrayt/Ṭūrōyo word, with the same function, known to me, is gälo, a borrowing from Kurdish, as has already been stated by Ritter. gāyät ‘quite’ (< Turkish gayet < Arabic ġāya(t)), as in hāṯe gāyät normalyo ‘This is quite normal’. ʾama ‘but’ is of frequent occurrence in the language. It corresponds in meaning to another Ṣūrayt/Ṭūrōyo word, namely bäs. The latter is also found in Kurdish (bes), which is perhaps the immediate source of our bäs. However, this bäs, which

68

The change û > ö in this case is interesting, as such a change in Ṣūrayt/Ṭūrōyo is unusual, as has been stated by Jan Retsö, see Tezel (2003: 28). However, it is not clear whether the change in question took place in Ṣūrayt/Ṭūrōyo, as the Turkish word concerned, which is of Persian origin, also occurs in Anatolian Arabic in the forms gūya, gōya. 69 Note that three words in the sentence gȫya ʾad-dāmāre d-lēbe mqapṭi are borrowings from Turkish, which represents an exceptional case. The same meaning may be expressed thus ʿal ʾū-mamro ʾaw-wārīḏe dlēbe ṣxīri, in which all words are genuine Ṣūrayt/Ṭūrōyo words. 70 Türkçe Sözlük (2005: 5a) indicates an Arabic form ʿacebā (our transcription = ʿaǧabā) as the source, but I have failed to find such an independent Arabic form ending in -ā in my Arabic sources. In Persian, on the other hand, ʿaǧabā is indicated alongside ʿaǧaban, see Steingass (2006: 837a).

THE TURKISH LEXICAL INFLUENCE ON ṢŪRAYT/ṬŪRŌYO

87

probably is a homonym both in Kurdish and our language,71 has other meanings and functions, such as ‘enough, only’, as in bäs! ‘That is enough!’; hāṯe bäs? ‘This only’? Thus, the use of ʾama, with a single and clearly-defined meaning, came to be favoured. It corresponds in meaning to Syriac bram, of which there are no traces in Ṣūrayt/Ṭūrōyo. It should be noted that the Turkish ama itself is a borrowing from Arabic ʾammā. demäk or demäkki ‘that means, that means to say’, as in kǝte ʾapartmāne, demek latne kāfīne ‘they have blocks of flats, that means that they are not hungry’, demäk lānīyo ‘that means it is for these’. This is a synonym of ʾī-maʿnayḏa72, which also occurs frequently, especially among Ṣūrayt/Ṭūrōyo-speakers in or from Arabicspeaking countries. čünkü (Turkish çünkü, of Persian origin) ‘because, therefore’. This word is of frequent occurrence. In Arabic-speaking countries, Ṣūrayt/Ṭūrōyo-speakers use lī-ʾan (of Arabic origin), with the same meaning. They correspond in meaning to Syriac ʿal d and meṭṭul. The former also occurs in Ṣūrayt/Ṭūrōyo, where ʿal d and mīḏe d (< *men ʾīḏeh d)73 were the normal words for ‘because, therefore’, before the Turkish word in question was borrowed. Whether the latter, meṭṭul, is a living word in Ṣūrayt/Ṭūrōyo is not quite clear. I have not witnessed its occurrence.74 gēč (Turkish geç) ‘late’ is of very frequent occurrence both as an adverb and an adjective. Examples: qāyǝmno gēč ‘I got up late’; fāyəšno geč ‘I am late’. Before this Turkish borrowing gēč took place, Kurdish däräng, with the same meaning, was used by several dialects in Ṣūrayt/Ṭūrōyo. From Arabic we have ʿ-w-q, underlying mʿāwǝ/aq ‘to be late’. The purists occasionally make use of mšawḥǝ/ar (from Syriac šawḥar, root ʾ-ḥ-r), with the same meaning as that of mʿāwǝ/aq.

71

The word bäs with the meaning ‘but’ seems to go back ultimately to Persian pas, which, among other things means ‘but, moreover, however; then’. This is also to be found in Ottoman Turkish in the form pes as a borrowing, see Steingass (2006: 249a); Redhouse (1999: 929a). Persian has also bas, which, among other things, means ‘enough, sufficient’, see Steingass (2006: 184b), hence Ottoman Turkish bes, see Redhouse (1977: 161b). With the meanings ‘enough, only’, our bäs and Kurdish bes are, thus, to be connected with this word. For the Kurdish bes, see Chyet (2003: 46a), comparing it to Persian bas, Sorani bes. 72 The word maʿna ultimately comes from Arabic maʿnan ‘meaning, sense’ (root ʿ-n-y), found also in Turkish in the form mana. 73 Literally mīḏe d (< *men ʾīḏeh d) means ‘from his/its hand that’; cf. Syriac b-yaḏ d ‘therefore, because’, and men yaḏ ‘on account of’, in which yaḏ is the absolute state of ʾīḏō ‘hand’. 74 Michael Gabriel (Bsōrīno) told me that he has heard the form mǝṭlōṯi ‘for my sake, on my account’ used by certain speakers of the dialect of Bsōrīno.

88

AZIZ TEZEL

ẓātan (Turkish zaten) ‘as a matter of fact’, for example, ẓātan ʾaġlab daqǝryāwōṯe hāwǝn qurūğīye ‘As a matter of fact, the majority of the villages became rural guards’. Mostly, it is used for emphasis, as is the case in Turkish. This word is of Arabic origin, from Arabic ḏātan ‘personally’. Before borrowing this word from Turkish, another loanword, namely Kurdish (ʾǝ)šxa, was used by several dialects in Ṣūrayt/Ṭūrōyo. ṣanki (Turkish sanki) ‘as if, as though’ is of frequent occurrence among Ṣūrayt/Ṭūrōyo-speakers in Istanbul. A Ṣūrayt/Ṭūrōyo phrase dǝmmǝt qay has the same meaning. yoqsa (< Turkish yoksa) ‘otherwise’, as in kǝtwayli šuġlo ġālābe, yoqsa gdōṭēwayno ‘I had much to do, otherwise I would had come’. A Kurdish borrowing nāxa in some dialects has the same function. The same meaning may be expressed by Ṣūrayt/Ṭūrōyo lō lō-ste, repeating the negative particle lō and adding the suffixste ‘also, too’. qarši (< Turkish karşı) ‘against’, as in hūli ʾōy qarši dīḏe ‘I have voted against him’; qāyǝm qarši dīḏe ‘He contradicted him’. In the case of the latter, one may very well use a native Ṣūrayt/Ṭūrōyo word qūm-, as in qāyǝm qūme. Otherwise, the Turkish loanword qarši in meaning chiefly corresponds to another loanword, in this case an Arabic, namely mqābǝl; cf. Syriac luqəḇal, which in Ṣūrayt/Ṭūrōyo has changed into lqul ‘before, in front of, in presence of’ (root q-b-l).

9. SOME VERBS

Verbs borrowed from Turkish into Ṣūrayt/Ṭūrōyo before the introduction of the Turkish school system seem to be rare. Such a verb is dmǝġ-le, which means ‘to mark; to mark with a stamp’. It is likely to be a denominative of damġa, ultimately from Turkish damga ‘instrument for stamping; stamp’. It is possible that this word found its way into Ṣūrayt/Ṭūrōyo through an Arabic dialect, in which the Turkish loanword damga was pronounced damġa.75 A debatable verbal root known to the older generation is t-r-ṣ, underlying tōrǝṣ (stem I, present). It is used in combination with qōlǝb in the phrase kō-qōlǝb w-kōtōrǝṣ with the meaning ‘to toss and turn’ (in bed). The same root underlies matraṣ- in matraṣ-le (stem III), also used in combination with the root q-l-b, as in maqlable wmatraṣle ʾū-qumaš ‘He turned the fabric over (from the side with the seam to right

75

A denominative verb from the same Turkish word is also to be found in Arabic, see, among others, Wehr (1976: 293a).

THE TURKISH LEXICAL INFLUENCE ON ṢŪRAYT/ṬŪRŌYO

89

side)’. Ritter connects this root with Turkish ters, 76 which, among other things, means ‘back’ or ‘reverse’ (of a thing). I have carefully suggested the possibility that it could reflect Syriac t-r-ṣ having meanings such as ‘to do rightly, set right, direct’, which in Ṣūrayt/Ṭūrōyo has survived dialectally in the form t-r-s ‘to send’, underlying mtāras- in mtārasle. If this is the case, we are then dealing with a doublet in the language.77 If it is not a reflex of the Syriac t-r-ṣ, Ritterʼs statement is valid. A denominative verb in Ṣūrayt/Ṭūrōyo, formed either from Kurdish päyda or Turkish peyda (of Persian origin) is mpāyad-le/mfāyad-le with the meanings ‘to procure, obtain, provide’ (as a synonym of mqāfē-le), as already stated by Ritter.78 Both Kurdish peyda kirin and Turkish peyda etmek have the same or similar meanings. Most other Turkish verbs seem to have been borrowed into Ṣūrayt/Ṭūrōyo after the introduction of the Turkish school system. Among these the verb mdāgaš- in mdāgaš-le, with the meaning ‘to change’, has an exceptional position in Ṣūrayt/Ṭūrōyo. Almost all the Ṣūrayt/Ṭūrōyo-speakers living in Turkey whom I have met use this verb, which is formed from the Turkish verb değişmek. This applies also to most Ṣūrayt/Ṭūrōyo-speakers coming from Turkey and now living in the diaspora. The use of this verb was probably spread after the emigration of the Ṣūrayt/Ṭūrōyospeakers to Istanbul. Its use, however, is not confined to the young generation. It occurs among the older generation as well, as has already been indicated. Interestingly enough, Ṣūrayt/Ṭūrōyo possesses other verbs with the same meanings. The native word for ‘to change’ in Ṣūrayt/Ṭūrōyo is mḥālaf- in mḥālafle, which is a synonym of mġāyar- in mġāyar-le and mbādal- in mbādē-le, with the same basic meaning. Both the verbal roots ġ-y-r and b-d-l are of Arabic origin. They are of frequent occurrence in the dialect of Mǝḏyaḏ. The same is true of their use among Ṣūrayt/Ṭūrōyospeakers in or from Arabic-speaking countries. It should be noted that the use of the native verb mḥālaf-le has decreased greatly, being more and more replaced by the foreign verbs mentioned above. Interestingly enough, when the purists want to use a native word for signifying the same action, today they rather make use of its šafʿel form, namely they say mšaḥlaf-le instead of mḥālaf-le. The form mšaḥlaf-le in Ṭūr ʿAbdīn was of rare occurrence. The genuine Ṣūrayt/Ṭūrōyo word mḥālaf-le covers all senses expressed by the other foreign verbs, which means that there is actually no need for using the foreign verbs instead of the native one. This is one of the important native verbs whose use should not be neglected.

76

See Ritter (1990: 100). For details, see Tezel (2003: 199–200). 78 See Ritter (1990: 494). 77

90

AZIZ TEZEL

The same is true of a native verbal root ṣ-x-r (< *s-x-r/s-k-r) ‘to close’, underlying ṣxər- in ṣxər-le, whose use by the Ṣūrayt/Ṭūrōyo-speakers in or from Turkey has decreased considerably, giving way to a Turkish borrowing mqāpaṭ- in mqāpaṭ-le, from Turkish kapatmak, which does not differ in meaning from the native root ṣ-x-r. In this connection we note that the Turkish verb kapatmak also occurs in Anatolian Arabic. While the native word mḥālaf-le came to be used together with its synonyms mġāyar-le and mbādē-le even among the Ṣūrayt/Ṭūrōyo-speakers in or from Arabicspeaking countries, I do not know of any Arabic word that is used as a synonym of the native ṣxǝr-le among these speakers. This may be due to the fact that several Arabic dialects, at least in Syria and Lebanon, having Ṣūrayt/Ṭūrōyo-speaking communities, make use of Arabic sakkar to mean ‘to close’, which is the etymological equivalent of the Ṣūrayt/Ṭūrōyo ṣ-x-r. The identity in meaning and the close similarity in phonological shape between the Ṣūrayt/Ṭūrōyo ṣ-x-r and Arabic s-k-r was perhaps a hindrance to a borrowing from the Arabic root s-k-r, with the meaning ‘to close’. To the list of Turkish verbs, which have got a foothold in Ṣūrayt/Ṭūrōyo, belongs mčaylaš- in mčaylaš-le, used in some phrases, such as mčaylaš-le qūme ‘to fight for someone’; mčaylašle bad-därsayḏe ‘to work on one’s homework’. I have noted mčālaš-li in the expression hāwīlīyo dart mčālašli ‘It became a worry to me (so) I endeavoured’. It is borrowed from Turkish çalışmak, which is the normal Turkish word for ‘to work’. The diphthong ay in the form mčaylaš-le is formed by analogy with ay in other quadriradical verbal formations having ay. We note that the normal word for ‘to work’ in Ṣūrayt/Ṭūrōyo is ʿwǝd-le in some dialects and šāġǝl in others. The former is native, while the latter is ultimately of Arabic origin, found also in Kurdish with the same meaning. As has already been indicated, mčaylaš-le or mčālaš-le is mainly used in senses other than those ʿwǝd-le and šāġǝl, so there are well-founded reasons for their borrowing. mqāraš-le. Among the verbs of Turkish origin one should note mqāraš- in mqāraš-le ‘to interfere, meddle with’, used mainly along with the negative lō, as in lō mqaršǝt, lat yō šuġlux ‘do not interfere; it is not your own business’. In this case Ṣūrayt/Ṭūrōyo has no native verb for denoting the same meaning. The word for ‘to be destroyed’, ḥāru (root ḥ-r-w), as in hat lō ḥāru lux, with emphasis, has, to be sure, a similar sense, but this expression means, without emphasis, ‘do not worry, do not take notice’, and the like. Other phrases such as lō čaykǝt rūḥux or nḥīrux; lō māʿabrǝt rūḥux have a similar sense. An Arabic loanword mdāxal- as in lō mdax-lǝt, has an equivalent sense. Further, we note the word for ‘to win, earn, acquire, gain; win over’, mqayzanin mqayzan-le, which is formed from Turkish kazanmak. Older Ṣūrayt/Ṭūrōyo words, with the same or similar meanings, are mḥayṣal- in mḥayṣē-le and rābǝḥ, both being

THE TURKISH LEXICAL INFLUENCE ON ṢŪRAYT/ṬŪRŌYO

91

of Arabic origin and still in use. The diphthong ay in mqayzan-le is the result of an analogical formation; cf. ay in mḥayṣal (< *mḥāsal) and in mčaylaš mentioned above. There are in Ṣūrayt/Ṭūrōyo no traces of any verbal forms from the Syriac root q-n-y, with the same meanings. I know only of one derivative of this root in Ṣūrayt/Ṭūrōyo, namely qanyōne (pl.) ‘cattle’. I do not know of any attempt to introduce any verbal form of the root in question into Ṣūrayt/Ṭūrōyo. An introduction of q-n-y into Ṣūrayt/Ṭūrōyo should take the verbal form qnē-le. To the frequent verbs borrowed from Turkish belongs the word for ‘deceive’, mqandar- in mqandar-le, whose base form qandar- comes from Turkish kandırmak. An older Ṣūrayt/Ṭūrōyo verb, with the same or similar meaning, is mxāpē-le, coming from Kurdish xapandin.79 Ṣūrayt/Ṭūrōyo has no a proper verb denoting ‘to rub, wipe, clean’ (a writing on the blackboard or any handwriting). The Ṣūrayt/Ṭūrōyo verbal root ḥ-r-w, with a basic meaning ‘to destroy’, has in fact the sense ‘to rub’, as in maḥraw ʾǝšmi ‘rub out my name!’, whose use with the sense in question has decreased considerably. Among the younger generation, this sense of the verb maḥraw-le is almost forgotten. Ṣūrayt/Ṭūrōyo-speakers in or from Turkey nowadays make use of the verbal forms slǝg-le and sī-le, with the same meaning as that of a Turkish verb silmek. As we can see, the base of the Ṣūrayt/Ṭūrōyo verb slǝg- in slǝg-le contains a g, which is not the case with the base of the Turkish verb sil- in silmek. The Ṣūrayt/Ṭūrōyo verb slǝg- in slǝg-le can then be best explained as a denominative from sǝlgīye ‘rubber, eraser’, from the Turkish silgi, and thus not a direct formation from the Turkish verb silmek, from which, on the other hand, the form *səl- in sī-le (< *səl-le) could have been borrowed directly.80 The purists nowadays make use of lḥē-le (root l-ḥ-y), with the same meaning, which recently has been introduced into Ṣūrayt/Ṭūrōyo from Syriac.81 The same is true of lāḥayto ‘a rubber, eraser’, formed as a neologism from the same root. The same action is also signified by expressions such as mbēle mūdarbo and čīk taḥte ʿāqıl dīḏe. The former literally means ‘he led him from the way’, and the latter ‘he permeated through his mind’. Sometimes the expression čīk taḥte ʿāqǝl dīḏe serves as ingressive to the expression mbēle mūdarbo, thus: čīk taḥte ʿāqǝl dīḏe w-mbēle mūdarbo. 80 The form sī-le is to be found in Ritter (1990: 506). The most frequent form is, however, slǝg-le. The form sīle, which is mainly used in the dialect Mǝḏyaḏ, seems to be formed from the Turkish verb silmek, as is evident from its form. It is thus not a denominative from sǝlgīye. 81 The root l-ḥ-y, underlying lḥē- in lḥē-le, occurs at least once in Ritter (1971: III, 402), where we read: ʾī-däwla dī-Osmanlīye glōḥēno ʾǝšma mīyarʿo, “ich (komme über dich und) vertilge den namen des Osmanischen Reiches vom erdboden”, Ritter (1971: III, 403). The occurrence of l-ḥ-y here may reflect a genuine Ṣūrayt/Ṭūrōyo case, and thus not a borrowing from 79

92

AZIZ TEZEL

Among the foreign verbs belongs māpǝs-le ‘to run down someone’. This is a denominative from pīs ‘dirty, foul’, found both in Turkish and Kurdish. It is possible that the word pīs found its way into Ṣūrayt/Ṭūrōyo via Kurdish and not directly from Turkish. According to Redhouse Sözlüğü, this word is of Persian origin.82 Such a statement is not to be found in Türkçe Sözlük. To the list of foreign verbal roots also belongs q-y-d ‘to register’, underlying mqāyad- in mqāyad-le. Although the verbal root in question is ultimately of Arabic origin, it is possible that in its present usage it reflects a borrowing from Turkish kaydetmek, with the same meaning. To be sure Arabic qayyada has, among other things, the same meaning, but the normal Modern Arabic word for ‘to register’ is saǧǧala, which in the form saǧǧal is also the word for ‘to register’ in many Arabic dialects. Hence Ṣūrayt/Ṭūrōyo-speakers in or from Arabic-speaking countries use msāǧē-le (< *msāǧal-le) for signifying the same act. A native verbal root k-ṯ-w, with a basic meaning ‘to write’, has also the sense ‘to register’, when used alongside ʾǝšmo ‘name’, or, rūḥo ‘self’, both being in this case used with pronominal suffixes, thus: kṯūle ʾǝšme, rūẖe. z-n-g-n ‘to enrich’, as in ʾaq-qǝštōṯāni mzanganne ʾī-ʿīto ‘these arches have enriched the church’. This is a formation either from the Turkish verb zenginleştirmek or a denominative of Turkish zengin ‘rich’,83 which today is of frequent occurrence in Ṣūrayt/Ṭūrōyo, replacing two older loanwords maldār and dōlāmand, which probably came into the language via Kurdish maldar and dewlemend.84 The Western Syriac word for ‘a rich’, ʿattīrō (in Ṣūrayt/Ṭūrōyo ʿātīro), is mainly known to the Ṣūrayt/Ṭūrōyo-speakers from its biblical usage to denote the rich man (Dives) in the parable in Luke 16: 19-31. However, the purists nowadays make use of ʿātīro in its basic sense ‘rich’ for replacing the foreign words mentioned above. No verbal forms of the Syriac root ʿ-t-r ‘to be rich’ are familiar to me in Ṣūrayt/Ṭūrōyo. The same meaning is expressed by the verb hāwi ‘to become’ in combination with one of the words maldār, dōlāmand, zängīn, ʿātīro and ġani, the last occurring mostly among Ṣūrayt/Ṭūrōyo-speakers in or from Arabic-speaking countries. Among the Syriac. I thank Michael Gabriel, Bsōrīno, who informed me about the occurrence of l-ḥ-y in this passage. 82 Cf. Persian pes ‘leprosy; vile, ignoble’? 83 The Turkish zengin is a borrowing from Persian sengīn, according to Türkçe Sözlük (2005: 2230a). 84 Both the words have their origin in Persian, having māldār and dawlatmand. Both the words are hybrids, the former from Arabic māl + Persian dār, and the latter from Arabic dawlat + Persian mand. They were also borrowed from Persian into Ottoman Turkish in the forms maldar and devletmend.

THE TURKISH LEXICAL INFLUENCE ON ṢŪRAYT/ṬŪRŌYO

93

Ṣūrayt/Ṭūrōyo-speakers in or from Turkey hāwi zängīn is nowadays of frequent occurrence. From a morphological point of a view there is no hindrance for introducing two verbal forms from the Syriac root ʿ-t-r, namely ʿātǝr ‘to be or become rich’, and maʿtar-le ‘to enrich’. A quadriradical verbal formation s-k-t-r ‘to drive a person away’ (by using coarse language), underlying msaktar- in msaktar-le, is a denominative of sǝkter (< Turkish siktir) ‘hop it, scram!’. Another denominative quadriradical verbal formation is s-r-g-n ‘to banish’, underlying msargan- in msargan-le (usually in the pl. msarganne), from Turkish sürgün ‘banishment; place of exile’, which is also the normal word among the Ṣūrayt/Ṭūrōyo-speakers in Turkey. The verbal root b-ḥ-s ‘to mention’, underlying mbāḥas- in mbāḥas-le, is also likely to be a denominative formation. It is either an internal Ṣūrayt/Ṭūrōyo formation from baḥs, or has been formed under the influence of Turkish bahs etmek or Kurdish baḥs kırın, both having the same meaning. The word baḥs, bahs in these languages reflects a borrowing from Arabic baḥṯ ‘research; discussion’, which is used by Ṣūrayt/Ṭūrōyo-speakers in or from Arabic-speaking countries. This means that the word in question occurs as a doublet in Ṣūrayt/Ṭūrōyo. To the list of denominative verbal roots also belongs k-ġ-t ‘to paper’, underlying mkāġat- in mkāġat-le,85 from Turkish kâğıt ‘paper’, found also in Ṣūrayt/Ṭūrōyo both as kāġat and kāġaz.86 The latter has been taken over from Kurdish. Another denominative verbal root is l-k-h ‘to spot, stain’, underlying mlākah- in mlākah-le, the h being an augmented element. The Persian word laka, lakka was borrowed into several languages in the region, its form in Turkish being leke and in Ṣūrayt/Ṭūrōyo läkke; cf. also its Iraqi Arabic form lakka. It is not clear whether the word was spread in the region through Persian or Ottoman Turkish. In the case of Ṭūrōyo, it is of a later date, the older word, used with the same meaning, is däqqe. Denominative verbal forms from this word are found even in other languages; cf. Turkish lekelemek and Iraqi Arabic l-k-l-k.87 The word leke is also to be found in Kurdish.

85

It is to be found in Ritter (1990: 162, 226). Ritter (op. cit.: 226) considers it a Kurdish word. However, the immediate source of kāġat and mkāġat-le is the Turkish kâǧıt, which comes from Persian kāġaẕ, kāġiẕ, according to Türkçe Sözlük (2005: 1033b). The Kurdish form used in Ṭūr ʿAbdīn is kāġaz. 86 The word for ‘paper’ is also termed warqo (< Arabic waraq) among the Ṣūrayt/Ṭūrōyospeakers in the diaspora. 87 For the Iraqi Arabic lakka and l-k-l-k, see Woodhead and Beene (2003: 426b).

94

AZIZ TEZEL

To verbal roots of Turkish origin belongs also y-s-q ‘to forbid, prohibit’, underlying myāsaq- in myāsaq-le, which is formed either from the Turkish yasak etmek, with the same meaning, or only as a denominative from yasak ‘prohibition; forbidden, prohibited’. The Turkish word yasak nowadays is one of the frequent loanwords in Ṣūrayt/Ṭūrōyo; cf. its use in the following example: b-mar Gabriyel ʾū-Ṭūrōyo yāsaq-wa ‘in Mar Gabriel it was forbidden to speak Ṭūrōyo’ (in order to encourage the use of Syriac among the students). The corresponding term among Ṣūrayt/Ṭūrōyo-speakers in or from Arabic-speaking countries is mamnuʿ, with the verbal form mamnaʿ-le.

10. SOME TURKISH LOANWORDS CONSTRUCTED WITH SƏMLE

Ṣūrayt/Ṭūrōyo-speakers in or from Turkey have borrowed several Turkish words in the form of substantives and adjectives into their language, where the words in question are used in phrases with verbs sǝm-le ‘he did, made’, hāwi ‘he became’, and maḥwē-le ‘he showed’. Here follows a selection of examples with sǝm-le. The English meanings are given in the infinitive, although səm-le is in the preterite. sǝm-le baṣqi (with the preposition ʿal) ‘to oppress’. The older word for signifying the same or a similar meaning is ẓlǝm-le, or with sǝm-le, as in sǝm-le ẓǝlmīye ʾaʿle, the base form ẓlǝm- being ultimately of Arabic origin. A native verbal root d-y-š, with a basic meaning ‘to tread’, has nowadays such a sense. Thus, one may say ʾī-ḥukūme kō-dayšo ʾū-ʿāmēḏa ‘the government oppresses its people’, instead of ʾī-ḥukūme kōsaymo baṣqi ʿal ʾū-ʿāmēḏa. This d-y-š corresponds in meaning to the Turkish verb basmak, from which baskı is derived. sǝm-le ʾäläštri (Turkish eleştiri) ‘to criticize’, which corresponds in meaning to nqǝd-le, in which nqǝd- is an Arabic borrowing, used mainly by Ṣūrayt/Ṭūrōyospeakers in or from Arabic-speaking countries. Ṣūrayt/Ṭūrōyo does not possess a proper native word for signifying the same meaning. The word for ‘to strike’, mḥēle, along with the preposition b, thus mḥēle b-, as in mḥēle bī-ḥukūme ‘he criticized the government’, has in fact an equivalent meaning.88 88

I do not know of any proper verbal form in Syriac with the specific sense ‘to criticize’. The roots s-t-r, ʿ-d-l and r-š-y seem to be the most proper ones for signifying such a meaning. There are no traces of the roots in question in Ṣūrayt/Ṭūrōyo. In the event of introducing a verbal form from Syriac into Ṣūrayt/Ṭūrōyo for replacing the foreign words concerned, the Syriac root ʿ-d-l/ʿ-ḏ-l must be avoided, as it may lead to confusion with Ṣūrayt/Ṭūrōyo ʿ-d-l ‘to correct’, which is of Arabic origin. Some purists already use sturyo, with the sense ‘criticism’. This is a derivative of the Syriac root s-t-r with meanings such as ‘to break down; to refute, confute, to ruin; to speak against’; cf. Hebrew ś-t-r and Arabic š-t-r. This Syriac s-t-r is to be

THE TURKISH LEXICAL INFLUENCE ON ṢŪRAYT/ṬŪRŌYO

95

sǝm-le ʾätki (Turkish etki) ‘to effect; to have influence upon’ is a synonym of sǝm-le tāsīr; cf. Turkish tesir etmek (Turkish tesir < Arabic taʾṯīr, root ʾ-ṯ-r). The corresponding term used by the Ṣūrayt/Ṭūrōyo-speakers, who do not use the Turkish form etki, is mʾāṯar-le, from Arabic ʾ-ṯ-r. As an abstract noun for denoting ‘effect, influence’, the purists nowadays make use of māʿǝbḏōnūṯo (root: Syriac ʿ-b-d/ʿ-ḇ-ḏ = Ṣūrayt/Ṭūrōyo ʿ-w-d), being recently introduced from Syriac into Ṣūrayt/Ṭūrōyo. Its verbal form should be maʿbaḏ-le, which has not made itself felt among the speakers. The expression sǝm-le ʾätki ʾaʿle, in which only ʾätki is of Turkish origin, may be replaced by sǝm-le māʿǝbḏōnūṯo ʾaʿle. sǝm-le toplantīye ‘to hold a meeting, a gathering’, which corresponds in meaning to sǝm-le ʾǝǧtīmāʿ, used by Ṣūrayt/Ṭūrōyo-speakers in or from Arabic-speaking countries. Today, the purists use the phrase sǝm-le k/xnušyo, in which k/xnušyo has been introduced into the language from Syriac. A verb from the same word would be mǝṯkānašše (pl.). The phrase lātīmi or lāyīmi lǝḥḏōḏe has an equivalent sense.89 sǝm-le qōnušma (Turkish konuşma) ‘to make a speech; to give a lecture’. The corresponding expression among the purists nowadays is mǝr-le or mqādam-le mǝlṯo, in which mǝlṯo has been introduced from Syriac into Ṣūrayt/Ṭūrōyo; cf. Arabic ʾalqā kalima(t), in which kalima(t) corresponds in meaning to Western Syriac melləṯō. The Syriac root m-l-l, with a basic meaning ‘to speak’, from which is derived melləṯō, does not seem to have survived in Ṣūrayt/Ṭūrōyo. sǝm-le ʾanlašma (Turkish anlaşma) ‘to reach an agreement’. Ṣūrayt/Ṭūrōyo has no a proper native word for denoting an ‘agreement’. A word with the same or a similar meaning is qawl, used in the phrase sǝm-le qawl, in which qawl is ultimately of Arabic origin. Otherwise, a native phrase ʾāṯǝn lǝḥḏōḏe is used to mean ‘to come to an agreement’. The proper word for an ‘agreement’ in Syriac is ʾawyūṯō (Western Syriac pronunciation),90 which nowadays has been introduced into Ṣūrayt/Ṭūrōyo. It is used by the purists also with the sense ‘solidarity’.

kept separate from another Syriac s-t-r, having meanings such as ‘to conceal, cover; to protect’; cf. Hebrew and Arabic s-t-r, see Brockelmann (1982: 502b–503a–b). In the sense ‘to protect’, this root is retained in Ṣūrayt/Ṭūrōyo, where it occurs both as ṣ-t-r and s-t-r, the form ṣ-t-r being more common. Some words with the form s-t-r are of Arabic origin, for example, yā-star (yā sǝttār) ‘God preserve!’, from Arabic yā sattār. 89 As to the root l-y-m, underlying lāyīmi, compare Syriac l-m-m, with the sense ‘to gather’, with Arabic l-m-m as the etymological equivalent. The passage of mediae geminatae roots into II-y roots is regular in Ṣūrayt/Ṭūrōyo. As for the root l-t-m, underlying lātīmi, compare Syriac ltm in the sense ‘to pile up, bring together’. 90 It is derived from the Syriac root ʾ-w-ʾ.

96

AZIZ TEZEL

sǝm-le dänämä (Turkish deneme) ‘to make an attempt; to have a try’, as a synonym of mğārab-le, in which ǧ-r-b is an Arabic borrowing. The Syriac root n-s-y, with the same meaning, has not survived in any verbal form in Ṣūrayt/Ṭūrōyo, to my knowledge. An introduction of a verbal form from this root into Ṣūrayt/Ṭūrōyo should take the verbal form mnāsē-le; cf. the Ṣūrayt/Ṭūrōyo noun nǝsyūno, ‘temptation’, reflecting Western Syriac nesyūnō. sǝm-le qullanmīš (Turkish kullanmış) ‘to make use of’, for example, kō-saymīna qullanmīš xabre tǝrki ‘We use Turkish words’. In meaning, it corresponds to kōmǝstaʿmlīna, from Arabic istaʿmala (root ʿ-m-l). The normal word that expresses this in Syriac is ʾeṯḥaššaḥ (with the preposition b), the ʾeṯpaʿʿal form of the root ḥ-š-ḥ, of which there are no traces in Ṣūrayt/Ṭūrōyo. No serious attempts to introduce this verbal stem into Ṣūrayt/Ṭūrōyo are familiar to me. From a morphological point of view, there are no obstacles to introduce it into Ṣūrayt/Ṭūrōyo, where one very well could say kō-mǝṯḥašḥīna instead of the foreign words. sǝm-le (or hū-le) ʾātāxīye,91 (with the preposition ʾaʿl -), as in sǝmme (huwwe) ʾātāxīye ʾaʿle ‘They informed against him’. The word ʾātāxīye is a corrupted form of Turkish yataklık, ‘bedstead’, which along with etmek, thus yataklık etmek means ‘to harbour (a criminal), conceal (stolen goods)’. Turkish words that have lost their literal meanings tend to undergo sweeping changes; cf. ʾaprāxe, ‘stuffed vine-leaves’, whose base form ultimately comes from Turkish yaprak.

11. SUMMARY

In this article it has been shown that Ṣūrayt/Ṭūrōyo-speakers in or from Turkey use a considerable number of Turkish elements in their speech. They have entered Ṣūrayt/Ṭūrōyo at different periods, mostly from the period after the introduction of Turkish school system into Ṭūr ʿAbdīn in the 1940s, and after the Ṣūrayt/Ṭūrōyospeakers began to migrate from Ṭūr ʿAbdīn to Istanbul in the beginning of 1950s and onwards. The majority of the borrowed particles and verbs stem from the latter pe-

91

It is not clear whether ʾātāxīye is a direct borrowing from Turkish or it came into the language via Kurdish. According to some informants, the Kurdish spoken in Ṭūr ʿAbdīn has a similar form. In Chyet (2003: 5b) I could find only altax ‘informer, slanderer’, and altaxî ‘slander, informing on’, the former compared by the author to Turkish alçak ‘low; base; mean’. It is probable that the Kurdish forms quoted have nothing to do with Turkish alçak, but rather with Turkish yataklık, with the l a secondary addition. If that is the case, our form ʾātāxīye would be formed from a Kurdish *ataxi < *yataxi.

THE TURKISH LEXICAL INFLUENCE ON ṢŪRAYT/ṬŪRŌYO

97

riod. Some technical terms and words with cultural significations pertaining to currency, fabrics, shoes, are of an earlier date. It is likely that not all the borrowed elements mentioned above have entered the language directly from Turkish, some of them being taken over from Kurdish and Arabic dialects, where the same Turkish elements were/are in use. For example, the form of čärxi ‘a coin of five piaster’ points to a borrowing from Kurdish çerxî, and not from Turkish çeyrek, and the meaning of qämäri ‘a piaster, penny’ can be identified in Levantine Arabic. The immediate sources of some borrowed elements, especially those of the terms källä ‘money’ and šūqo ‘a kind of linen fabric’, remain uncertain. The meaning of the former and the form of the latter (with š- instead of an expected č- or ǧ-) seem to be typical for Ṣūrayt/Ṭūrōyo, where, moreover, šūqo is homonymous, coinciding with another Ṣūrayt/Ṭūrōyo-word, šūqo ‘a street; the centre of the city’. Two other homonyms discussed above, namely čīt ‘chintz’, čīt ‘fence’, and bäs ‘but’, bäs ‘enough’, have arisen in other languages in the region. The study shows that some of the borrowed elements denote objects and social activities for which Ṣūrayt/Ṭūrōyo has no words of its own, while some others came to replace or are in the process of replacing native words and phrases, for example, gȫya ‘according to reports’, instead of the native ʿal ʾū-mamro, čünkü ‘because, therefore’, instead of the native ʿal d and mīḏe d. This is especially true of two borrowed Turkish verbs, namely mdāgaš- ‘to change’, and mqāpat- ‘to close’, instead of the native mḥālaf- and ṣxər-, respectively. Therefore we have given an account for the Ṣūrayt/Ṭūrōyo lexical elements which were in use, before the Turkish elements were borrowed, partly in order to demonstrate some important changes that took place within the Ṣūrayt/Ṭūrōyo lexicon, and partly also to encourage their use so that they hopefully will not disappear. The article also provides some suggestions for neologisms and draws attention to some neologisms introduced into the language from Syriac.

REFERENCES

al-Munǧid. 1975. = al-Munǧid fī al-luġa wal-ʾaʿlām. Beirut: Dār al-mašriq. Baalbaki, Rohi. 1991. al-Mawrid, A Modern Arabic-English Dictionary. Beirut: Dar elilm lilmalayin. Barthélemy, Adrien. 1935-1969. Dictionnaire arabe-français; Dialects de Syrie: Alep, Damas, Liban, Jérusalem, Paris: Geuthner. Beṯ-Ṣawoʿe, Jan. 1997. ǝno mǝrli ʿammo Išoʿ Qašo Malke madʿarle. Södertälje: BeṯFroso Nsibin. Brockelmann, Carl. 1928. Lexicon Syriacum. Editio secunda aucta et emendata. Halis Saxonum, Niemayer: Tübingen.

98

AZIZ TEZEL

Bustānī, Butrus. 1870 (reprint 1983). Muḥīṭ al-muḥīṭ, qāmūs muṭawwal lil-luġa alʿarabiyya, Beirut: Maktabat Lubnān. Chyet, Michael Lewisohn. 2003. Kurdish-English Dictionary. New Haven, CT: Yale University Press. Collins. 1991. Collins English Dictionary (3rd ed.). Glasgow: Harper Collins. Derleme Sözlüǧü. 1963–1977. Türk Dil Kurumu Yayınları. Ankara: Türk Tarih Kurumu Basımevi. Fraenkel, Siegmund. 1886. Die aramäischen Fremdwörter im Arabischen. E.J. Brill: Leiden. Hony, Henry Charles. 1957. A Turkish-English Dictionary (2nd ed.). Oxford: Clarendon Press. İzoli D. 1992. Ferheng: Kurdi-Turki, Türkçe-Kürtçe. Istanbul: Deng Yayınları. Kazimirski de Biberstein, Albin. 1860 (Réimpr. 1960). Dictionnaire arabe-français, 12. Paris: Maisonneuve. Lane, Edward William. 2003. Arabic-English Lexicon, 8 vols., London 1863–93. Reprinted by the Islamic Texts Society (UK), 2 vols. MacLean, Arthur John. 1901. A Dictionary of the Dialects of Vernacular Syriac as Spoken by the Eastern Syrians of Kurdistan, North-West Persia, and the Plain of Mosul. With Illustrations from the Dialects of the Jews of Zakhu and Azerbaijan, and of the Western Syrians of Tur ‘Abdin and Ma‘lula, Oxford: Clarendon. Payne Smith, Jessie. 1903. A Compendious Syriac Dictionary, Oxford: Clarendon. Payne Smith, Robert, ed. 1981. Thesaurus Syriacus, I–II, Oxford 1879–1901. Nachdruck der Ausgabe Oxford 1879–1901, Hildesheim, New York: G. Olms. Redhouse, James. 1999. A Turkish/Ottoman-English Dictionary. Istanbul: SEV Matbaacılık ve Yayıncılık. Ritter, Helmut. Ṭūrōyo. Die Volkssprache der syrischen Christen des Ṭūr ‫ﻉ‬Abdîn. 5 vols. Stuttgart: Steiner. —. 1971. Ṭūrōyo. Die Volkssprache der syrischen Christen des Ṭūr ‫ﻉ‬Abdîn. Texte Band III. Beirut: Steiner (in Kommission). —. 1979. Ṭūrōyo. Die Volkssprache der syrischen Christen des Ṭūr ‫ﻉ‬Abdîn. Wörterbuch. Beirut: Steiner (in Kommission). —. 1990. Ṭūrōyo. Die Volkssprache der syrischen Christen des Ṭūr ‫ﻉ‬Abdîn. Grammatik: Pronomen, „sein, vorhanden sein“, Zahlwort, Verbum. Stuttgart: Steiner. Sabar, Yona. 2002. A Jewish Neo-Aramaic Dictionary: Dialects of Amidya, Dihok, Nerwa and Zakho, northwestern Iraq, Wiesbaden: Harrassowitz. von Soden, Wolfram. 1965–1981. Akkadisches Handwörterbuch, I–III, unter Benutzung des lexikalischen Nachlasses von Bruno Meissner (1868–1947). Wiesbaden: Harrassowitz. Steingass, Francis. 2006. Persian-English Dictionary. New Delhi: Manohar Publishers.

THE TURKISH LEXICAL INFLUENCE ON ṢŪRAYT/ṬŪRŌYO

99

Tarama Sözlüǧü. 1974. Türk Dil Kurumu Yayınları. Ankara: Türk Tarih Kurumu Basımevi. Tezel, Aziz. 2003. Comparative Etymological Studies in the Western Neo-Syriac (Ṭūrōyo) Lexicon. With Special Reference to Homonyms, Related Words and Borrowings with Cultural Siginification, Studia Semitica Upsaliensia 18. Elanders Gotab: Stockholm. Thesaurus Syriacus; see Payne Smith, Robert. Türkçe Sözlük. 2005. Türk Dil Kurumu Yayınları. Ankara: Akşam Sanat Okulu Matbaası.‫‏‬ Vocke, Sibylle, and Waldner, Wolfram. 1982. Der Wortschatz des anatolischen Arabisch. Magister-Arbeit, Universität: Erlangen-Nürnberg. Wehr, Hans. 1976. A Dictionary of Modern Written Arabic; edited by J.M. Cowan (3rd ed.). Ithaca, NY: Spoken Language Services, Inc. Woodhead, Daniel R., and Beene, Wayne (eds.). 2003. A Dictionary of Iraqi Arabic: Arabic-English. Washington, DC: Georgetown University Press.

NEOLOGISMS IN ṢŪRAYT/ṬŪRŌYO

1

SINA TEZEL A situation of diglossia similar to that between Modern Standard Arabic and Arabic dialects exists between Western Syriac and Ṣūrayt/Ṭūrōyo. The Ṣūrayt/Ṭūrōyospeakers in general view Ṣūrayt/Ṭūrōyo as the everyday language of communication, while they consider Western Syriac as the cultural and ecclesiastical language with higher prestige. This is where we can find answers to why Ṣūrayt/Ṭūrōyo almost solely forms its neologisms from linguistic material from Western Syriac, or takes over pre-formed Syriac words which are given new meanings without considering them as borrowings into Ṣūrayt/Ṭūrōyo. Today, a great number of such neologisms exist in Ṣūrayt/Ṭūrōyo, most of which probably did not exist before the 1950s, judging from their absence in Ritter’s Ṭūrōyo collection. The phenomenon appears to have first begun spreading at the beginning of the 1970s, when Ṣūrayt/Ṭūrōyospeakers began to emigrate to Sweden and other Western European countries. The formation of their own clubs and associations in these countries, and the publication of their own newsletters and magazines have played an important role. They did not have the freedom to engage in such communal activities to the same extent in their home countries. The exact number of neologisms and their spread among the Ṣūrayt/Ṭūrōyo-speakers is not known. At any rate, it is clear that the neologisms are used by the purists, in clubs and associations, in television programmes and in newspapers. They are disseminated through these means. The neologisms in Ṣūrayt/Ṭūrōyo are formed for two reasons, namely to replace foreign words and to describe new phenomena in society.2

1

This article is an expanded version of a lecture I gave at the International Conference of Neo-Aramaic Dialectology, June 26–27, 2013, at The Hebrew University of Jerusalem. 2 See Tezel (2011: 18).

NEOLOGISMS IN ṢŪRAYT/ṬŪRŌYO

101

1. DEFINITION AND TYPES OF NEOLOGISMS

Trask defines the term ‘neologism’ as follows: “A recently coined word, especially one constructed consciously and deliberately by one person or by an official body” (Trask 2000: 228). There are different types of neologisms. They can be divided into three different categories. The first one includes only neologisms that have not existed up to a certain period of time and cannot be found in the texts written before a given moment, for example, the Ṣūrayt/Ṭūrōyo-neologisms that are not found in Ritter’s Ṭūrōyo collection. The second one includes the neologisms that have changed their meaning but retained their old form, with their old meaning lost, or relegated to secondary importance. Applying this to Ṣūrayt/Ṭūrōyo, we can mention the example fulḥōno, which today in the diaspora denotes mostly ‘activities’ in an association or ‘political activities’ in general, while in Ṭūr ʿAbdīn the same word denoted ‘an arable land’. The third category contains neologisms that have only added one or more new meanings without losing the significance of the old ones. 3 Applying this to Ṣūrayt/Ṭūrōyo, we can adduce the example qōlo ‘voice; church song’, which in the diaspora has acquired a new meaning, namely ‘vote’, without losing its original meanings. The main focus is in this article will be on the first type of neologism mentioned above, with a particular focus on neologisms relating to the organisation life in the diaspora.

2. WORDS RELATING TO THE ORGANISATION LIFE

As has been mentioned above, many of the neologisms formed in Ṣūrayt/Ṭūrōyo relate to the activities of the associations. The word for ‘an association or a club’, which is termed ḥuḏ̣ro, is itself a neologism. It is the Western Syriac ḥūḏrō ‘circle, course, a company’, which is derived from the Syriac root ḥ-d-r/ḥ-ḏ-r ‘to go around’. The meaning ‘an association or a club’ is thus new. The word ḥuḏ̣ro has replaced two foreign words, namely Arabic nādi and Turkish dernek. The word is related to the Ṣūrayt/Ṭūrōyo preposition ḥēḏ̣ər/ḥḏ̣ōr, ‘around’, which is the only genuine Ṣūrayt/Ṭūrōyo word derived from the root in question; cf. Western Syriac ḥəḏōr ‘around’. Words derived from this root are to be kept separate from the Arabic borrowing in Ṣūrayt/Ṭūrōyo ḥ-ḏ̣-r ‘to be present’, whose root in this case is Arabic ḥ-ḍ-r. The pronunciation of these two etymologically different roots has, thus, merged in Ṣūrayt/Ṭūrōyo. 3

For the different types of neologisms, see Levchenko (2010: 14–15).

102

SINA TEZEL

After the word ḥuḏ̣ro there was also a need for a word denoting ‘a board, committee’, which came to be termed səʿto. This is from the Western Syriac sīʿtō ‘succour, troupe, band, company’, from which is derived the denominative Syriac verb sayyaʿ ‘to uphold, support; to help, assist, succour, relieve’,4 a verb that does not seem to occur in Ṣūrayt/Ṭūrōyo. Not only the word səʿto, but also its meaning is new in the language. Note the change ī > ə, which has taken place in the closed syllable of səʿto. The word is used alongside two Arabic loanwords, namely, hayye and ləǧne. The use of səʾto is fairly widespread, but not familiar to everyone. The word is also used in compounds, such as səʿto dan-nīše ‘womenʼs committee’, səʿto dī-ʿīto ‘church board’, səʿto dī-ʿlaymūṯo ‘youth committee’. The abstract noun ʿlaymūṯo ‘youth’ and the word for ‘young person’, ʿlaymo (m.), ʿlaymṯo (f.), are all neologisms imported from Western Syriac, having the same forms and meanings. Before introducing these words from Western Syriac into Ṣūrayt/Ṭūrōyo, the normal word used for ‘youth, young a person’ was the Kurdish loanword xort (m.), xorte (f.).5 An inherited word grēʿūno ‘youth, a young person’, also occurs in Ṭūr ʿAbdīn and is still in use. There was also need for a word denoting ‘a member’, hādōmo,6 which is directly imported from Western Syriac haddōmō. This is a Persian loanword that came into the language at an early period, but Ṣūrayt/Ṭūrōyo-speakers generally do not seem to be aware of its origin and consider it to be a native word. Previously in Ṭūr ʿAbdīn, the sense of ‘a member’ was expressed by two words, viz. ʿāẓa and ʿəḏ̣u (both derived from Arabic ʿuḍw), depending on whether the Ṣūrayt/Ṭūrōyo-speakers are from Turkey or Arabic-speaking countries. The board and its members need ‘a chairman’, which is termed rīšo/rīšōno (m.), rīšōnīṯo (f.), with rīšūṯo/rīšōnūṯo ‘chairmanship’ as the abstract. In these cases the meanings are new, as the word forms are not imported from Western Syriac. The Ṣūrayt/Ṭūrōyo word rīšo, which has the common meaning ‘head’, is used in the sense of ‘leader’ in phrases such as rīšo dī-qrīṯo means ‘the leader of the village’. The board and its members also need to arrange ‘a meeting’, which came to be termed knušyo. This is pronounced by some Ṣūrayt/Ṭūrōyo-speakers xnušyo, which, with regard to the fricativization of k in word-initial position, is one of three words demonstrating the fricativization of bǝġaḏkǝfaṯ in such a position. This is a new phenomenon in the diaspora, for there is no evidence of fricativization of bǝġaḏkǝfaṯ in initial position in Ṭūr ʿAbdīn. The word knušyo has been introduced directly from

4

See Sokoloff (2009: 1006). See Ritter (1979: 565). 6 The word hādōmo is attested in Ritter (1979: 190), with the meaning ‘Knochen’, without referring to its occurrence in any text but to an informant. 5

NEOLOGISMS IN ṢŪRAYT/ṬŪRŌYO

103

Western Syriac, reflecting Western Syriac knūšyō, derived from k-n-š ‘to gather; to sweep’; cf. Ṣūrayt/Ṭūrōyo k-n-š, ‘to sweep’. From the same root is derived the word ‘broom’ maknəšto, which is a genuine Ṣūrayt/Ṭūrōyo word. That the word is not termed *maxnəšto, which should be according to the bǝġaḏkǝfaṯ-rule, is due to the maintenance of the semantic connection between it and the root k-n-š. Another neologism derived from the root k-n-š is kūnōšo, kūnōše (pl.) ‘result’, which is not as widespread as the other neologism knušyo. These two neologisms, namely knušyo and kūnōšo, are historically derived from two different verbal stems, the former from kənaš and the latter from kanneš. A meeting, in turn, needs ‘a programme’, which came to be termed taḥrazto. This is a new formation from the Syriac root ḥ-r-z ‘to perforate for stringing together; to place, arrange’ or ‘set in order’, especially with mellē ‘words’, i.e. ‘to string words together’, ‘to compose (a discourse)’, for which no genuine verbal root is attested in Ṣūrayt/Ṭūrōyo. The neologism taḥrazto is an abstract noun, formed with the prefix ta-. There are some other neologisms in Ṣūrayt/Ṭūrōyo formed in this way (see taglīṯo below). The root ḥ-r-z, underlying taḥrazto, is cognate with Arabic x-r-z ‘to pierce, bore (something)’. This was borrowed into Ṣūrayt/Ṭūrōyo, where xrəz- in xrəzle means ‘to thread, string’.7 An association has also to organize ‘an election’, which is termed gūbōyo/ġūbōyo, introduced into the language from Western Syriac, where gūbbōyō means ‘selection; the cleaning, winnowing of corn’. The form ġūbōyo is a second word manifesting the fricativization of bǝġaḏkǝfaṯ in word-initial position. This neologism is related to a genuine Ṣūrayt/Ṭūrōyo root g-w-y ‘to beg’, underlying gwē- in gwēle, for Ṣūrayt/Ṭūrōyo g-w-y reflects a root such as Syriac g-b-y, with the shift b  ḇ  w. The semantic differentiation between Ṣūrayt/Ṭūrōyo g-w-y and the neologism gūbōyo/ġūbōyo is due to the fact that the former is derived from the basic verbal stem gəḇō ‘to choose, approve, appoint; to prove, purge; to collect, exact (tribute)’, while the latter is derived from the paʿʿel form gabbī ‘to choose, select, collect; to clean corn’. In this connection we also note a third Ṣūrayt/Ṭūrōyo word mǝǧbīye, which is an Arabic loanword, whose root is ǧ-b-y, which is cognate with the Syriac root g-b-y.8 In ‘an election’ one has also to cast ‘a vote’, which is termed qōlo, a genuine Ṣūrayt/Ṭūrōyo form whose original pre-dominant meanings are ‘voice’ and ‘church song’. The sense ‘vote’ is thus new. The root underlying this word is q-w-l.

7 8

For taḥrazto and the cognate root x-r-z with x instead of ḥ, see S. Tezel (2011: 104). See Tezel (2011: 51–52).

104

SINA TEZEL

There was also a need for the associations to form ‘a federation’, which came to be termed ḥūyōḏo, imported from Western Syriac ḥūyyōḏō ‘a uniting; union’, the abstract noun from ḥayyeḏ ‘to make one, unite’, a denominative verb from ḥaḏ ‘one’.9 ‘A congress’ held by a federation is termed lūmōḏo, which is a new coinage, formed from Syriac lammeḏ ‘to bring together’. The Syriac verbal root l-m-d/l-m-ḏ does not seem to occur in any verbal form in Ṣūrayt/Ṭūrōyo. The Ṣūrayt/Ṭūrōyo word talmīḏo ‘disciple’ is from the Western Syriac talmīḏō, which in the lexicons of Syriac is, to be sure, indicated sub the root l-m-d/l-m-ḏ, but nowadays some suggest that this word is a borrowing from Akkadian talmīdu.10 An association also organizes activities for its members. The terms used for ‘activities’ themselves are šūšōṭe and ḥfīṭwōṯo, but more common today is the term fulḥōno (pl. fulḥōne), whose meaning, but not form, is new, for Ṣūrayt/Ṭūrōyo has already fulḥōno with the meaning ‘arable land’. Both the terms šūšōṭe and ḥfīṭwōṯo are directly imported from Western Syriac. However, even in these cases, their Ṣūrayt/Ṭūrōyo meanings are new, for Western Syriac ḥfīṭūṯō (sg.) has the meanings ‘exhortation, diligence, earnest care’, a derivative of the root ḥ-f-ṭ. Western Syriac šūšōṭō (sg.) has the meanings ‘advance, growth, progress, course’. In the case of ḥfīṭwōṯo, no related words are attested in Ṣūrayt/Ṭūrōyo. The word šūšōṭe is closely related to Syriac šawšeṭ ‘to stretch out’ (šafʿel of the root y-š-ṭ), with which one may compare Ṣūrayt/Ṭūrōyo mžd ‘to stretch out’, whose historical root likewise is y-š-ṭ.11 Among the regular activities organized by an association we can mention ḥāgo ‘a feast’, which in Ṭūr ʿAbdīn seems to have been of rare occurrence,12 but today in the diaspora is a frequent word. The form ḥāgo is derived from a form such as Western Syriac ḥaggō ‘feast, festal day, feasting; a festal assembly (hence) a fair, any assembly; a company, dance’, by loss of gemination and then compensatory lengthening of the preceding vowel. To an association’s social activities in the diaspora also belong the institution of a ‘social gathering’ šahro, which is usually arranged in the evenings of the holidays. The occurrence of the word šahro in Ṭūr ʿAbdīn seems to have been restricted to the festivities maintained to the memory of ‘Saintʼs Day’.13 A verbal form from the same root, šāhīri ‘to pass the night awake; to attend a gathering in the evening or at night’ 9

See Payne Smith (1903: 139); Brockelmann (1928: 219). See von Soden (1965–1981: 1311); Leslau (1987: 575); Sokoloff (2009: 1650). Thanks to Sergey Loesov for drawing my attention to the Akkadian form. 11 See Tezel (2003: 87); Tezel (2011: 165-166); Ritter (1990: 664–665). 12 See Ritter (1979: 212). 13 See Ritter (1979: 483), indicating šahro in the example adyawma ʾū-šahro d-mōr Hōbil-yo, ‘heute ist der feiertag von dem Heiligen Abel’. 10

NEOLOGISMS IN ṢŪRAYT/ṬŪRŌYO

105

is in the process of replacing an Arabic borrowing s/ṣāhīri. The Aramaic root š-h-r and the Arabic root s-h-r are cognates with phonological correspondences.14 To the activities of an association in the diaspora also belong the arrangement of ‘a youth camp’ mašrīṯo, which has been directly imported from Western Syriac mašrīṯō ‘a camp; a troop; an army’. This is derived from the root š-r-y ‘to loosen; untie’, which is also preserved in Ṣūrayt/Ṭūrōyo with the same meanings. The paʿʿel stem of the same root šarrī in Syriac means ‘to begin’, which in the form mšārēle is used by the purists to replace a common Arabic loanword bdē-le in Ṣūrayt/Ṭūrōyo. The establishment of cultural and national associations also gave rise to the establishment of magazines. The foreign words used to denote ‘a newspaper or magazine’ are gazeta, ǧarīde and maǧalle, depending on whether the Ṣūrayt/Ṭūrōyospeakers are from Turkey or Arabic-speaking countries. These words have in some Ṣūrayt/Ṭūrōyo-speaking groups been replaced by one imported from Western Syriac, namely mġalṯo, which in Syriac actually denotes ‘a scroll’. The meaning ‘newspaper or magazine’ is thus new and is associated with the meaning of the Arabic word maǧalla/e. The root of mġalṯo is g-l-l and its original form in Syriac is məġalləṯā. Even international words such as ‘television’ and ‘radio’ are expressed by their neologisms. Thus the word for ‘television’ is termed frōs ḥezwo and that for ‘radio’ is termed frōs qōlo. Even if the roots f-r-s and ḥ-z-y, from which these neologisms are formed, are found in Ṣūrayt/Ṭūrōyo, such constructions, where the first part is in status constructus, have the form of Western Syriac. Not only status constructus, which is unusual in Ṣūrayt/Ṭūrōyo, but also other signs show that the neologisms in question are not genuine Ṣūrayt/Ṭūrōyo forms. The long -ō- in a closed syllable is unusual. Also -e- in a closed syllable does not exist in genuine Ṭūrōyo words, but rather it undergoes changes in such a position. Finally, Ṭūrōyo has no genuine abstract nouns from the root ḥ-z-y with -w-.15 With the neologisms for ‘television’ and ‘radio’ also came the neologism ṭēbe (pl.) ‘news’, from Western Syriac ṭebbē ‘a message, news’, whose root is ṭ-b-b. Also an Arabic loanword xabre (pl.) is used with the same meaning. The associations arrange also ‘a demonstration’, which is expressed by the neologism taglīṯo. This is formed from the root g-l-y, preserved in Ṣūrayt/Ṭūrōyo in all three stems, namely glē-le, mgālē-le and maglē-le, all with the meaning ‘to take the cover off; to expose, uncover’. Both the form and meaning of taglīṯo, which belongs to the neologisms formed with the prefix ta-, are new. In this connection we should mention also the neologism gōlūṯo, which denotes ‘a diaspora’. It is based on a form 14 15

See Tezel (2011: 179–180). See Tezel (2011: 206–207).

106

SINA TEZEL

such as Western Syriac gōlūṯō ‘captivity; captives, exiles’, which is considered a derivative of the same root g-l-y. In connection with the activities of the associations are formed also football teams. This has given rise to new words with new meanings that have replaced the corresponding foreign words. The word for ‘football’ is termed ʾesfirreġlo, which is formed from Western Syriac ʾesfīr ‘sphere, globe, ball’ + reġlō ‘foot’. The first part ʾesfīr is of Greek origin. The Ṣūrayt/Ṭūrōyo word ṣfēro ‘omlette’ actually reflects the same Syriac word in its emphatic state esfīrō, thus esfīrō > Ṣūrayt/Ṭūrōyo sfēro.16 As for reġlo ‘foot’, the genuine Ṣūrayt/Ṭūrōyo form is raġlo, exhibiting the change e > a in a closed syllable. The word for ‘football team’ is termed gūdo. It is a new meaning compared to the original meaning of Ṣūrayt/Ṭūrōyo gūdo ‘church choir’, which reflects a form such as Western Syriac gūdō ‘musical chorus’. The word for ‘a football player’ is termed məštaʿyōno, whose meaning, but not form, is new in the language. The word in Ṣūrayt/Ṭūrōyo means ‘player’, which is derived from the Ṣūrayt/Ṭūrōyo verbal stem məštāʿē-le ‘he played’; cf. Syriac ʾəštaʿʿī ‘to play; to tell’. The root in this case is š-ʿ-y. In this context we may note another derivative of the root š-ʿ-y, namely tašʿīṯo ‘tale, story’, which has been introduced from Western Syriac into the language. The word for ‘trainer, coach’ is mdaršōno and that for ‘training’ is dūrōšo, reflecting Western Syriac mdarrəšōnō ‘schoolmaster, teacher’,17 and dūrrōšō ‘exercise; teaching’, respectively.18 Both are formed from the paʿʿel stem of Syriac root d-r-š ‘to tread, find out or prepare (a path, metaph.); to practise, train, instruct; to dispute, to debate’, which does not seem to have survived in Ṣūrayt/Ṭūrōyo.

3. WORDS RELATING TO EDUCATION

From the root d-r-š is formed also madrašto ‘school’, which has come to replace, or is used alongside two foreign words, maktab and madrase,19 in Ṣūrayt/Ṭūrōyo. The word is directly ḥ-ḏ-r taken from Western Syriac maḏreštō/maḏraštō20 ‘school’, which is assumed to be modelled on Arabic madrasa(t)Error! Bookmark not defined..21

16

See Tezel (2003: 148). See Payne Smith (1903: 254). 18 See Payne Smith (1903: 88). 19 Both the foreign words maktab and madrase are ultimately of foreign origin, but they are or were used as loanwords in other languages in the region. 20 Some lexicons of Syriac have the vocalization maḏreštō (thus in Brockelmann 1928: 168), and in Payne Smith (1903: 254), others have the vocalization maḏraštō (thus in Manna 17

NEOLOGISMS IN ṢŪRAYT/ṬŪRŌYO

107

The word for ‘university’ is termed bēṯ ṣawbo, which is directly imported from Western Syriac bēṯ ṣawbō ‘place of meeting’. Compounds with the form bēṯ are not genuine Ṣūrayt/Ṭūrōyo constructions, for the genuine Ṣūrayt/Ṭūrōyo equivalent is bē, in which the ṯ is elided, for example bē qāšo ‘the house of the priest’. The word for ‘a student’ is termed yōlūfo, which is derived from the root y-l-f, the normal Ṣūrayt/Ṭūrōyo root for ‘to learn’. The form yōlūfo, however, has been introduced from Western Syriac into Ṣūrayt/Ṭūrōyo. Another neologism related to education is fərʿo ‘branch, study programme’, from Western Syriac ferʿō ‘a shoot, sucker; ramification of shoot’. The new meaning of Ṣūrayt/Ṭūrōyo fərʿo ‘branch, study programme’ is associated with the metaphorical meaning of ḥ-ḏ-r, Arabic farʿ.

4. MEANS OF CONVEYANCE

Among the words denoting means of conveyance, the word for ‘a car’ rāḏ̣ayto is a neologism formed from a Syriac root r-d-y/r-ḏ-y ‘to journey, go forward, proceed on the way’. It is modelled after the Arabic word sayyāra(t), with the same meaning, which is also used by Ṣūrayt/Ṭūrōyo-speakers in or from Arabic-speaking countries. Ṣūrayt/Ṭūrōyo-speakers in or from Turkey often use the word ʿaraba. The Syriac root r-d-y/r-ḏ-y, from which rāḏ̣ayto is formed, does not seem to occur in any verbal form in Ṣūrayt/Ṭūrōyo. From this is to be kept separate an Arabic borrowing r-ḏ̣-y (< Arabic r-ḍ-y) ‘to be content’; cf. Syriac r-ʿ-y, with the same meaning.22 Another word in this field is ṭāyosto/ṭāyəsto ‘an airplane’, which is formed from a Syriac root ṭ-w-s/ṭ-y-s ‘to fly on high’. From the same Syriac root is formed the second part of another neologism bēṯ ṭawso ‘an airport’. The Syriac root ṭ-w-s/ṭ-y-s does not seem to occur in any verbal form in Ṣūrayt/Ṭūrōyo, which uses the verbal root f-y-r for ‘to fly’.23 Another word in this field is qṭōro ‘a train’, whose meaning is borrowed from Arabic qiṭār, but its form is the same as Ṣūrayt/Ṭūrōyo qṭōro ‘a ceiling’, which in this case is a genuine Ṣūrayt/Ṭūrōyo word, derived from the root q-ṭ-r ‘to ceil; to congeal, curdle’, the same root as Syriac q-ṭ-r. The Arabic word qiṭār in older Arabic

1900: 164 and in Audo 1896: vol. 1, 218). Note that the West-Syrians in their actual reading say madraštō, with d and thus not with ḏ, hence also Ṣūrayt/Ṭūrōyo madrašto, with d. 21 See Sokoloff (2009: 718). 22 For the Arabic borrowing r-ḏ̣-y, see Tezel (2011: 157–158). 23 See Tezel (2003: 68).

108

SINA TEZEL

sources has the meaning ‘a train of camels’, which in this meaning was borrowed into Syriac.24

5. UTENSILS

Among the neologisms in the field of utensils we may mention the word for ‘refrigerator’ maqrōno, and the word for ‘freezer’ māgəlḏōno/magləḏōno, which are formed from the roots q-r-r and g-l-d/g-l-ḏ, both occurring also in Ṣūrayt/Ṭūrōyo, although in the form q-y-r ‘to cool’ and g-l-ḏ ‘to freeze’. They have replaced, or are used alongside, the Turkish loanword buzdōlabi and the Arabic loanword barrād. Another neologism worth mentioning in this field is mašəlyūno ‘fork’, which is a diminutive formation of Western Syriac mašləyō, denoting ‘pitchfork’, which is a derivative of the Syriac root š-l-y ‘to draw out’.25 By the diminutive form mašəlyūno a new word with a partly new meaning has been created. It has replaced or is used alongside two foreign words, namely Turkish çatal and Arabic šawke.

6. COURTESY PHRASES

Among the neologisms in the field of the courtesy phrases, we should mention the word for ‘apology; excuse’, šubqōno, and the word for ‘thanks’, tawdi, which have been introduced from Western Syriac šūbqōnō and tawdī, respectively. They are derivatives of the roots š-b-q, which among other things, means ‘to forgive’, and y-d-y ‘to acknowledge: to thank’. These roots do not seem to have survived in genuine Ṣūrayt/Ṭūrōyo words. The form tawdī is a shortened form (status absolutus) of tawdīṯō ‘thanksgiving; religion’. The form tawdīṯo, with the meaning ‘thanksgiving’, is attested in Ritter once.26 These two courtesy phrases have become common and replaced almost all foreign phrases with the same meanings. A third neologism in this field is fqūḏ ‘please; you are welcome’. It is an imperative formed from the Syriac root p-q-ḏ ‘to visit, enquire; to command, order’. The word fuqdōno27 ‘order’, a derivative of the same root, is used by the purists. Among the courtesy phrases we should also mention lalyo ṭōwo or alternatively lalyo ṭōbo ‘good night’, and brīx safro/safro ṭōbo ‘good morning’. The form ṭōbo has been introduced from Syriac, as the genuine Ṣūrayt/Ṭūrōyo form is ṭōwo/ṭawo. An-

24

See Tezel (2003: 112); Sokoloff (2009: 1358). To be sure, Ṣūrayt/Ṭūrōyo has a root š-l-y, ‘to become quite, to calm down’, but this reflects the same root as the Syriac š-l-y, ‘to be silent’, for which see Tezel (2011: 183–184). 26 See Ritter (1979: 518). 27 The word fuqdōno is also indicated in Ritter (1979: 165). 25

NEOLOGISMS IN ṢŪRAYT/ṬŪRŌYO

109

other phrase for ‘good night’ is lalyo danyōḥo. The abstract noun nyōḥo as a genuine Ṣūrayt/Ṭūrōyo word actually means ‘recovering’, from nayəḥle ‘he recovered’, but in the phrase lalyo danyōḥo, the form nyōḥo has been introduced from Syriac into the language. Finally, in this field I shall mention the word kūšōro ‘good luck’, imported directly from Western Syriac kūššōrō ‘success, good fortunate’.

REFERENCES

Audo, Thomas. 1896. Sīmtō d-lešōnō sūryōyō, I–II. Urmia. (Reprinted by the Assyrian Federation in Sweden, Stockholm 1979, with the English title: Dictionary of the Assyrian Language). Brockelmann, Carl. 1928. Lexicon Syriacum. Editio secunda aucta et emendata. Halis Saxonum, Niemayer: Tübingen. Leslau, Wolf. 1987. Comparative Dictionary of Geez (Classical Ethiopic): GeezEnglish/English-Geez. Wiesbaden: Harrassowitz. Levchenko, Yaroslav. 2010. Neologism in the Lexical System of Modern English. On the Mass Media Material. Master’s Dissertation. Munich: GRIN Verlag. Manna, Jacques Eugène. 1900. Leksīqōn kaldōyō-ʾarabōyō. Qāmūs kaldānī-ʿarabī. Chaldean-Arabic Dictionary. Mosul. (Reprinted with a new appendix by Dr. Raphael J. Bidawid, Chaldean bishop of Beirut, 1975. Babel Center Publications: Beirut). Payne Smith, Jessie. 1903. A Compendious Syriac Dictionary. Oxford: Clarendon Press. Ritter, Helmut. 1979. Ṭūrōyo. Die Volkssprache der syrischem Christen des Ṭūr ‫ﻉ‬Abdîn. Wörterbuch. Beirut: Steiner (in Kommission). —. 1990. Grammatik: Pronomen, “sein, vorhanden sein”, Zahlwort, Verbum. Stuttgart: Steiner. Tezel, Aziz. 2003. Comparative Etymological Studies in the Western Neo-Syriac (Ṭūrōyo) Lexicon. With Special Reference to Homonyms, Related Words and Borrowings with Cultural Siginification. Uppsala: Uppsala Universitet. Tezel, Sina. 2011. Arabic Borrowings in Ṣūrayt/Ṭūrōyo within the Framework of Phonological Correspondences: In Comparison with other Semitic Languages. Uppsala: Uppasala Universitet. Trask, Robert Lawrence. 2000. The Dictionary of Historical and Comparative Linguistics. Edinburgh: Edinburgh University Press. von Soden, Wolfram. 1965–1981. Akkadisches Handwörterbuch, I–III. Wiesbaden: Harrassowitz.

THE THREE /R/S OF BAṚTƏḶḶA KRISTINE MOLE 1. INTRODUCTORY REMARKS

Within the group of dialects called North-Eastern Neo-Aramaic (NENA) it is common to encounter more than one kind of /R/ sound,1 e.g. emphatic [ṛ], retroflex [ɽ]/[ɻ] and the tap/trill [r]. These occur frequently as allophones of /R/, sometimes as independent phonemes, e.g. /r/ and /ɽ/ or /r/ and /ṛ/, but in most cases there are only two independent /R/ phonemes. The only published research into a dialect with three independent /R/ phonemes is that of Werner Arnold (2001) on the dialect of Baṛṭəḷḷa. There is also unpublished research by Dr Roberta Borghero describing three independent /R/ phonemes in the dialect of Dere, which she presented at the NeoAramaic seminar in Cambridge. I am grateful to her for letting me refer to some of her data. The article below explores the ternary (three-way) system of three independent /R/ phonemes in the Baṛṭəḷḷa Neo-Aramaic (BNA) dialect. It discusses possible conditioning phonetic factors for their distribution and compares this to Werner Arnold’s previous research on the subject. Finally, it compares data about the three /R/s in BNA to the three /R/s of Dere, and to the binary system of the emphatic /ṛ/ and the tap /r/ in the dialect of Barwar.

1.1. Baṛṭəḷḷa (Bəɽəṭle)

Baṛṭəḷḷa is an Aramaic-speaking village of about 10,000 people, situated just off the road between Mosul and Erbil. It is 20 minutes by car from Mosul to the west, to the 1

Capital /R/ is used in this article as a shorthand form of referring to a grouping of /R/ phonemes, in this case a catch-all term for emphatic /ṛ/, retroflex /ɽ/ and the tap /r/ = the three /R/s.

THE THREE /R/S OF BAṚTƏḶḶA

111

south it is 10 minutes from Karemlesh and 15 minutes from Qaraqosh (Baghdede). Dialects of NENA are spoken in both Karemlesh and Qaraqosh and there are active links between the villages. In spite of this, the occurrence of three independent /R/ phonemes is a peculiar phenomenon to the dialect of Baṛṭəḷḷa only, as the dialects of Karemlesh and Qaraqosh do not feature more than one /R/. Arabic is spoken in the village as well as Baṛṭəḷḷa Neo-Aramaic (BNA). All teaching at schools and any official correspondence is conducted in Arabic. Television and entertainment tends to be in Arabic. There has until very recently been very strong links with Arabic-oriented Mosul, though this has been compromized by recent unrest in Mosul. Kurdish is not spoken in the village, although it is clear from the lexicon of BNA that there has been very strong influence from Kurdish in the past. Some Shabaks live in the village, about 1500–2000 people, but there is minimal contact between the two groups.2

2. THE THREE /R/S

In the NENA dialect of Baṛṭəḷḷa there are three independent /R/ phonemes: an emphatic /ṛ/, a retroflex /ɽ/ and the tap /r/. There is a minimal ‘triplet’ which illustrates this: páɽe < *pārā páṛe < *parre páre < *pāre

‘bran, husks left after sieving’ ‘lambs’ ‘money’ (< Kurd. pāra)

The minimal pairs below also testify to the existence of three independent /R/ phonemes.

2.1. Minimal Pairs /ɽ/ : /r/

2

ɽoq roq

‘flee’ (IMP.SG) ‘spit’ (IMP.SG)

ʾáɽya ʾárya

‘may she hold’ ‘lion’

Shabaks are an ethnic and religious minority group, with small communities in the Zinjar and Nineveh region of northern Iraq. Their language, Shabaki, is a north-western Iranian language close to Gorani Kurdish. They follow an independent religion, related to but distinct from orthodox Islam and Christianity.

112

KRISTINE MOLE

2.2. Minimal Pairs /ṛ/ : /ɽ/ góṛa góɽa

‘man’ ‘may she marry’

ʾámṛa ʾámɽa

‘wool’ ‘may she say’

qṭíṛa qṭíɽa

‘knotted, tied’ ‘a handful’

maṛéṯa maɽéṯa

‘rennet’ ‘pain’

2.3. Minimal Pairs /ṛ/: /r/

There is only one minimal pair attested between the emphatic /ṛ/ and the tap /r/ (apart from the minimal triplet above). daṛóye daróye

‘to look after’ ‘to winnow, to throw something up in the air’

It is unusual in phonology to get oppositions that are three-way (ternary) rather than two-way (binary). This ternary system appears to include limited conditioning factors for the tap /r/ (i.e. only at the beginning of a word in words with no other conditioning factors, or in loanwords). One might expect that the retroflex /ɽ/ and emphatic /ṛ/ would occur only in vastly different phonetic environments. To some extent this is true, but there is also some overlapping in relation to environments with labials.

3. DISTRIBUTION OF THE THREE /R/S

Faced with three different /R/s in BNA, it is necessary to study their distribution, to compare the phonetic environments in which they occur, and to attempt to define what factors condition their presence. Werner Arnold’s research into the three /R/s of Baṛṭəḷḷa concludes that there are strong correlations between the other phonemes in a word and the distribution of /R/. According to Werner Arnold (WA), the distribution of the three /R/s in the dialect is as follows:

THE THREE /R/S OF BAṚTƏḶḶA

113

3.1. Summary of WA’s Phonetic Conditioning Factors Emphatic /ṛ/

Retroflex /ɽ/ (after these sounds)

Labials Emphatic /ṭ/

Glottal /ʾ/ Glottal /h/ Back consonants /k/, /g/, /x/, /q/ Back vowels /u/ and /o/

Tap /r/ Retained after and palatals

dentals

WA concludes his summary with a list of words that cannot be classified, and concedes that among these there are many words with labials that have retroflex /ɽ/, not emphatic /ṛ/, as one would expect within the conditioning factors of WA.

3.2. Summary of this Article’s Phonetic Conditioning Factors

My own research into conditioning phonetic factors for the three /R/s coincide with some of those of WA. The main differences are the addition of labials as a conditioning factor for retroflex /ɽ/s and more specific categories for the tap /r/. To facilitate easy comparison, the table below has a summary of my phonetic conditioning categories. The differences between these and WA’s categories are marked in bold.

Emphatic /ṛ/

Retroflex sounds)

Labials Emphatic /ṭ/3

Labials Glottal /ʾ/

In loanwords, especially from Arabic

Glottal /h/

3

/ɽ/

(after

these

Tap /r/ In word-initial position Where no other phonological factors are present In loanwords, especially from Kurdish

Back consonants /k/, /g/, /x/, /q/ Back vowels /u/ and /o/

Arguably, this may also be conditioned by other ‘emphatic elements’, not frequent enough to warrant a separate category, e.g. ʾamṛa ‘wool’ (*ʿamṛa) and ṣehṛa ‘moon’ and arguably qaṛiṛa ‘cold’.

114

KRISTINE MOLE

3.3. A Note on the Data

The data for this article have been collected from speakers of BNA in the UK and from fieldwork in Baṛṭəḷḷa in September 2010. Only words of Aramaic origin have been included as evidence for the categories proposed in the study below. This is based on the assumption that, like in other NENA dialects, the /R/s of BNA have developed from an earlier form of Aramaic which contained only one /R/. To date there is no evidence to the contrary. After studying and categorizing words with the three /R/s of the Aramaic corpus, loanwords have been traced where possible and added to the classification, to provide insight into their behaviour within the framework of the /R/s. It is worth noting that words from minimal pairs are excluded from any classification, as developments in minimal pairs are often irregular in order to avoid homonyms.

4. EMPHATIC /Ṛ/

The emphatic /ṛ/ or pharyngealized [rˁ] often affects the vowels in a word. The /a/ vowel especially tends to change quality from [a] or [ӕ] to [ɑ]. In fast speech, this change in vowel quality is more easily recognisable to the listener than the emphatic /ṛ/ itself. Within the corpus of loanwords, the emphatic /ṛ/ appears to be especially prominent in words of Arabic origin. The emphatic /ṛ/ may be affected by preceding or succeeding sounds (unlike the retroflex /ɽ/, which is only affected by preceding sounds).

4.1. Emphatic /ṛ/ with Labials

As WA described in his article, there is strong evidence to suggest that the emphatic /ṛ/ is conditioned by the presence of preceding or succeeding labials, e.g.: ʾámṛa ʾaṛmónta báθəṛ baṛáye baxóṛe búṛga búṛma máṛa măṛe maṛíṛa mdúṛa

‘wool’ ‘pomegranate’ ‘after, following’ ‘outside’ ‘incense’ ‘hole’ ‘cooking pot’ ‘spade’ ‘mirror’ ‘bitter’ ‘round’

THE THREE /R/S OF BAṚTƏḶḶA məṭṛa pəṣṛa qaṛíwa qóṛma ṣápṛa

‘rain’ ‘meat’ ‘relative; best man’ ‘tree-trunk; shot-gun’ ‘morning prayer; tomorrow’

Verbs: b-ṛ-x ṛ-w-y

II ‘to bless’ I ‘to grow up; be tipsy but not drunk’

115

4.2. Emphatic /ṛ/ with Emphatic /ṭ/ Pharyngealisation in emphatic consonants tends to spread to other segments, and this is illustrated in words with emphatic /ṭ/, which tends to condition the emphatic /ṛ/. maṭṛán məṭṛa qṭáṛta ṭaṛníθa ṭáṛpa ṭ-ṛ-p n-ṭ-ṛ

‘Metropolitan’ ‘rain’ ‘ceiling, dome’ ‘rack for bread’ ‘leaf’ II ‘to clap’ I ‘watch, wait’

4.3. Emphatic /ṛ/ in Loanwords Loanwords are a common feature in the dialect of Baṛṭəḷḷa, especially from Arabic but there are a couple from Kurmanji Kurdish. ʿəṛaq dəwáṛa fṭáṛta qánṭaṛa qhíṛa ṛəẓẓa ṣádṛa sáqṛa šáṛba

‘Iraq’ (Ar.) ‘circle’ (Ar.) ‘breakfast’ (Ar.) ‘old style domed roof’ (Ar. qanṭara ‘arch; vault’) ‘to be angry, be in mourning’ (Ar.) ‘rice’ (Ar.) ‘chest’ (Ar.) ‘falcon’ (Ar.) ‘pottery jar for one person’ (Ar.)

116

KRISTINE MOLE Verbs: ṛ-m-y ṛ-ʿ-y s-k-ṛ

I ‘to lay an egg’ (Ar.) I ‘to shepherd’ (Ar.)4 I ‘to be drunk’ (Ar.)

There is also one loanword from Kurdish with emphatic /ṛ/, bəsmáṛa ‘nail; corn on foot’ (Kurd. bizmar, Ar. mismār). It is possible that the pronunciation of this has been influenced by its cognate in Arabic. It may also be that there has been a secondary emphatic development later due to the number of labials in the word. In BNA there is a strong tendency for loanwords from Arabic to have emphatic /ṛ/, whereas loanwords from Kurdish tend to have the tap /r/, with a few having retroflex /ɽ/.

4.4. Emphatic /ṛ/ in other Words

There are a handful of words where there is no obvious explanation for the presence of emphatic /ṛ/. deṛa géṛa šéṛa nəšṛa qaṛiṛa ṛiṛe š-d-ṛ xóṛa

‘monastery’ ‘rolling pin’ ‘festival’ ‘eagle’ ‘cold’ ‘dribble’ II ‘to send’ ‘friend’

Of the words above, the three first in the list do appear at first glance to share a similar structure, but their original historical forms are very dissimilar, e.g. deṛa ‘monastery’ < *dayrā, géṛa ‘rolling pin’ < *gerrā, šéṛa ‘festival’ < *šahṛā. It is possible that the emphatic /ṛ/ in these words signals an elided segment, though there is not a lot of other evidence to support this. This particular word structure Céṛa is not exclusive to words with emphatic /ṛ/, as the following words illustrate: béɽa ‘well’, néɽa ‘yoke’, štéɽa ‘she descended’. It is unclear why nəšṛa ‘eagle’ has emphatic /ṛ/. There is no conditioning factor in ṛiṛe ‘dribble’ for an emphatic /ṛ/, in earlier Aramaic it is *ryrʾ, without any historical conditioning factors, e.g. /ʿ/ for the emphatic /ṛ/. The dialect of Dere also has ṛiṛe as emphatic.

4

This form exists in parallel with the earlier Aramaic form ṛ-ʾ-y.

THE THREE /R/S OF BAṚTƏḶḶA

117

The words ṣéhṛa ‘moon’ and qaṛíṛa ‘cold’ also have emphatic /ṛ/, but also the emphatic /ṣ/ and /q/. In BNA, /q/ sometimes behaves like an emphatic version of /k/, just like in some dialects of Arabic.5

5. RETROFLEX /Ɽ/

The retroflex /ɽ/ is never word-initial, and appears, therefore, to be conditioned by preceding sounds only. In the few cases where it appears to be word-initial it is clear that the conditioning segment has been elided. This is generally a glottal /ʾ/, e.g. ɽuta ‘Friday’ < *ʿaruḇtā and ɽəqle ‘he fled’ *mawtā> *gawzā> *ʿaḇdā>

poxa *mōtā> mɔta jöza ʾöda do.3SG.F

‘wind’ ‘death’ ‘walnut’ ‘(that) she does’

It has to be recognized that fronting of the back rounded vowels in CDZ is not a productive process, thus the words like poxa and bašule do not shift to *pöxa or *bašüle. We can, therefore, observe a fairly regular vowel chain shift, but the origin of the fronting process itself should probably be sought in the languages in contact (see above). It can be inferred that the process of vowel fronting occurred prior to the spread of emphasis over a word. Taking as an example the word ṱüṛa ‘mountain’, it is hardly probable that the vowel /ü/ developed between the emphatic *ṭ (of which the unaspirated /ṱ/ is here a reflex) and the emphatic /ṛ/ with its potential for emphasis spread (recall the plural form [ˈtʏ.rʕɑ:nɛ] ‘mountains’). Rather, it can be posited that first the emphasis in *ṭ was lost, leaving the lack of aspiration, and later the *u vowel shifted to /ü/. The emergence of the emphatic /ṛ/ can be construed as an independent development, detached from historical emphasis. The appearance of this new emphatic phoneme, with a potential for the spreading of emphasis would, therefore, be the most recent process operative in the dialect. The development of the words containing a historical emphatic and fronted vowels may be summarized as follows: Earlier Aramaic *ṭūrā> *ṱūrā>

*ṱüra >

CDZ ṱüṛa

CDZ ṱüṛane (back [ɑ])

We can, then, posit that historical emphasis was lost, leaving some traces limited to particular segments. After the stage of vowel fronting, a ‘new wave’ of expanding emphasis took place, which is what we now find in the synchronic state of the dialect. The following paragraphs outline more closely the chronological sequencing of these processes.

A DIACHRONIC PERSPECTIVE ON EMPHASIS IN CHRISTIAN DIYANA-ZARIWAW

141

3.2.4. Emphasis vs. fronted vowel distribution

Other dialects, in which both the fronted vowels and emphasis are present, further support the complementary distribution of fronting and backing. JU has generally no fronted vowels in the back words, or emphasis in words containing /ö/ and /ü/. In Haṣṣan from eastern Turkey the historical emphasis is present, but its spread is blocked by the fronted vowel, e.g. ṭüra (aspirated /ṭ/). Furthermore, some instances of non-fronted /u/ instead of expected /ü/, can be found in words with historical emphasis, like in xamuṣa ‘sour’ (Damsma 2009). The table below illustrates this complementary occurrence of emphasis and the fronted vowels. Note that where in Haṣṣan the original emphasis is maintained, it is lost in CDZ, and in the Urmi dialects it has developed into secondary emphasis, encompassing the whole word. The history of emphasis in CDZ could, thus, be reconstructed as a development from historical segmental emphasis, as in Haṣṣan, which was blocked by the presence of fronted vowels, and then subsequently weakened, followed by a second period of emphasis in which consonants became pharyngealized with consequential suprasegmental spreading. Earlier Aramaic *ṭūrā *ḥāmūṣā *gawrā *ḥlūlā

Haṣṣan ṭüra xamuṣa -

CDZ

JU

Christian Urmi + + ṱüṛa tura ṱuyra + xamüsa xamusa +xamuysa + güṛa jura + xḷüḷa xlüla xluyla

gloss ‘mountain’ ‘sour’ (m.) ‘big’ (m.) ‘wedding’

Having said that, it might be of interest to recall the remark which Garbell made about the emphasis spread in JU. Namely, she noted that the elderly female speakers tend to pronounce all the words, even the otherwise “plain” ones, in an emphatic manner, unless there is an /ö/ or /ü/ vowel in them (Garbell 1965: 33). This might further support the view that in dialects like JU and CDZ the process of emphasis spread is operative on the synchronic level, unless checked by the fronted qualities.

3.2.5. CDZ and languages in contact

The phonetic processes outlined for CDZ can be linked to the historical location of the CDZ speaking community and the issue of language contact. At an earlier stage, the impact of Turkish on the CDZ community would have been stronger, inducing the development of the fronted vowels. Later on the migration towards the south would have exposed the speakers of the dialect to a more intense contact with Kurdish on the one hand, and with the other NENA dialects on the other. These NENA dialects from the area of the lake of Urmi are characterized by the presence of su-

142

LIDIA NAPIORKOWSKA

prasegmental emphasis spread, which is also reported to exist to some extent in Kurdish (Garbell 1965; Hoberman 1985; Kurdoev 1978). This contact could have resulted in a tendency toward intensification of emphasis and its spreading beyond the unit of a syllable. Having mentioned the main aspects of emphasis in some NENA dialects and their geographical distribution, an attempt of classification with regard to emphasis could be advanced. Drawn with very rough strokes, the picture would look in the following way: North Syria emphatic segments emphasis spread Turkey (Khabur) emphatic segments (Jilu, Čalla) fronted vowels (Haṣṣan) Iran loss of emphasis, back vowels (Jewish Salamas) Iraq West East segmental emphasis emphatic segments synharmonism emphasis spread emphasis spread (the Urmi dialects) (Alqosh, JKS fronted vowels Barwar, Arhadin) (CDZ, the Gargarnaye cluster, other dialects?)

South

4. SUMMARY

loss of historical emphasis secondary emphatic segments (Jewish Sanandaj)

To conclude, it can be argued that the dialect of CDZ presents a case of transition with respect to emphasis. The process of diachronic development involved the loss of the historical emphatic consonantal segments, which is marked by the lack of aspiration in stops and affricates, and the later development secondary emphasis in liquids with consequential suprasgemental spread. The loss of the historical emphatic consonants in CDZ resulted in the innovative development of back vowels in otherwise non-emphatic words, which are the vestiges of the emphasis spread at an

A DIACHRONIC PERSPECTIVE ON EMPHASIS IN CHRISTIAN DIYANA-ZARIWAW

143

earlier period. Finally, after the development of secondary emphasis in liquids fronted round vowels are found to be flanked by emphatic consonants. For establishing the stage of emphasis that CDZ displays a comparison with other NENA dialects has been sketched above. It has also been argued that the disappearance of earlier suprasegmental emphasis or its weakening to back vowels is the only possible reconstruction. It can be claimed that the emphasis spread is on the rise in CDZ. This suggestion is based on the fact that vowel fronting must be chronologically anterior to the synchronic emphasis spread of liquids. Assuming this, the weakening or loss of original segmental emphasis in consonants and fronting of the vowels would precede the emergence of the new emphatic consonants, and the spread of emphasis. Therefore, the situation in CDZ would present a case of a shift from segmental to suprasegmental emphasis. Moreover, such an interpretation is further borne out by the overall picture of geographical distribution of emphasis across the NENA speaking area. The exposure to different languages at different stages of the dialect’s development is likely to account for such a scenario of the development of emphasis.

REFERENCES

Bohas, Georges, and Ghazali, Salem. 1995. “Le prétendu synharmonisme dans les parlers araméens modernes.” Langues orientales anciennes. Philologie et linguistique 5/6: 153–165. Damsma, Alinda. 2009. “The Christian Dialect of Haṣṣan in SE Turkey.” Materials from ARAM 26th International Conference in Oxford 2009 (unpublished). Garbell, Irene. 1964. ““Flat” Words and Syllables in Jewish East New Aramaic of Persian Azerbaijan and the Contiguous Districts (A Problem of Multilingualism).” In Studies in Egyptology and Linguistics in Honor of Hans Jakob Polotsky, edited by Haim B. Rosén, 86–103. Jerusalem: The Israeli Exploration Society. —. 1965. The Jewish Neo-Aramaic Dialect of Persian Azerbaijan. Linguistic Analysis and Folkloric Texts. The Hague: Mouton&Co. Hetzron, Robert. 1969. “The Morphology of the Verb in Modern Syriac (Christian Colloquial of Urmi).” Journal of the American Oriental Society 89: 112–127. Hoberman, Robert. 1985. “The Phonology of Pharyngeals and Pharyngealization in Pre-Modern Aramaic.” Journal of the American Oriental Society 105 (2): 221– 231. —. 1988. “Emphasis Harmony in a Modern Aramaic Dialect.” Language 64 (1): 1–26. —. 1989. “Parameters of Emphasis.” Journal of Afroasiatic Languages 2 (1): 73–97. Jakobi, Heidi. 1973. Grammatik des thumischen Neuaramäisch (Nordostsyrien). Wiesbaden: Deutsche Morgenländische Gesellschaft.

144

LIDIA NAPIORKOWSKA

Jakobson, Roman. 1962. “Mufaxxama.” In Selected Writings, vol. 1, Phonological Studies, 510–522. The Hague: Mouton&Co. Jušmanov, Nikolaj. 1938. “Syngarmonizm urmijskogo narechija.” In Pamjaty akademika N. Ja. Marra, 295–314. Izdatielstwo Akademij Nauk SSSR: Moscow. Khan, Geoffrey. 2008. The Jewish Neo-Aramaic Dialect of Urmi. Piscataway, NJ: Gorgias Press. —. 2009. The Jewish Neo-Aramaic Dialect of Sanandaj. Piscataway, NJ: Gorgias Press. —. 2013. “Phonological Emphasis in North-Eastern Neo-Aramaic.” In Base Articulatoire Arrière. Backing and Backness, edited by Jean Léo Léonard and Samia Naïm, 111–132. Munich: Lincon Europa. —. 2014. “Phonological Emphasis in North-Eastern Neo-Aramaic.” Edited by Yehudit Henshke, Aharon Maman, and Tamar Zewi. Carmillim: For the Study of Hebrew and Related Languages 10: [15]–[27]. Kurdoev, Kanat. 1978. Gramatika kurdskogo iazyka: na materiale dialektov kurmandzhi i sorani. Moscow: Nauka. Ladefoged, Peter. 1971. Preliminaries to Linguistic Phonetics. Chicago and London: The University of Chicago Press. —. 1993. A Course In Phonetics (3rd ed.). Fort Worth: Harcourt Brace College Publisher. Lagdefoged, Peter, and Maddieson, Ian. 1997. The Sounds of the World’s Languages. Oxford: Blackwell. Mutzafi, Hezy. 2004. The Jewish Neo-Aramaic Dialect of Koy Sanjaq (Iraqi Kurdistan). Wiesbaden: Harrassowitz. Odden, David. 1991. “Vowel Geometry.” Phonology 8: 261–289. Odisho, Edward Y. 1988. The Sound System of Modern Assyrian (Neo-Aramaic). Wiesbaden: Harrassowitz. Payne Smith, Jessie. 1999. A Compendious Syriac Dictionary: Founded upon the Thesaurus Syriacus of R. Payne Smith. Eugene: Wipf and Stock Publishers. Pennacchietti, Fabrizzio, and Tosco, Mauro. 1991. Testi neo-aramaici dell’Unione Sovietica raccolti da Enrico Cerulli. Napoli: Instituto Univeristario Orientale. Soane, Ely Bannister. 1913. Grammar of the Kurmanji or Kurdish Language. London: Luzac&co. Talay, Shabo. 2008. Die neuaramäischen Dialekte der Khabur-Assyrer in Nordostsyrien: Einführung, Phonologie und Morphologie. Wiessbaden: Harrassowitz. Younansardaroud, Helen. 2001. Der neuostaramäische Dialekt von Särdä:rïd. Wiesbaden: Harrassowitz.

DOMAINS OF EMPHASIS, SYLLABLE STRUCTURE AND

MORPHOLOGICAL BOUNDARIES IN THE CHRISTIAN URMI DIALECT OF NEO-ARAMAIC

GEOFFREY KHAN 1. THE PHONOLOGICAL DOMAIN OF EMPHASIS

In the North-Eastern Neo-Aramaic (NENA) dialect spoken by the Assyrian Christians of Urmi (henceforth C. Urmi) all words have one of two suprasegmental settings, viz. emphatic (i.e. pharyngealized involving retraction of the tongue root to the upper pharynx) or plain (i.e. lacking pharyngealization). Historically emphasis was a feature of emphatic consonantal phoneme segments, such as ṭ or ṣ, which had a tendency to spread to adjacent segments in the phonetic realization of a word. This is the case in, for example, the NENA dialects of Iraq. In C. Urmi this phonetic spread of emphasis has undergone a reanalysis in the historical evolution of the dialect and emphasis now has the status of a suprasegmental feature of the entire word, rather than a feature rooted in a one phoneme segment.1 The domain of emphasis in C. Urmi also includes enclitic grammatical elements of no lexical content that are attached to the end of a word. In such cases the emphasis setting of the main word spreads rightwards to the enclitic. These enclitic elements are marked as attached to the word in the transcription used here by the . Such enclitics include the following grammatical elements: (i) the copula 1

For a description of this process see Khan (2013). There have been various treatments of emphasis in the C. Urmi dialect in previous scholarship. For an overview of the literature see Younansardaroud (2001: 19–63), who herself adopts the approach of Tsereteli (1972; 1976; 1978: 28).

146

GEOFFREY KHAN

and (ii) the inclusive particles da and žə. In the following examples suprasegmental emphasis is marked by the symbol + at the beginning of the emphasis domain and the domain boundaries are marked by /, e.g.: /+ʾúllul/ /+ʾúllul ə/ /+ʾúllul

‘above’ ‘he is above’ ‘also above’

These enclitics may be combined, in the order copula + inclusive particle. In such cases both enclitics are included within the domain of emphasis of the base word, e.g.: /+ʾúllul ə da/

‘he is also above’

When an emphatic word is combined with a preceding word or particle in a single stress group, the preceding element is generally not within the domain of emphasis and so is not pronounced emphatic if, when standing independently, it is plain. These preceding elements are marked as combined to the following word by a single hyphen in the transcription, e.g.: /mən-/+ʾúllul/ /ʾa-/+jóra/ /lá-/+xəllə/

‘from above’ ‘this man’ ‘he did not eat’

Conversely, if the first element in a stress group is emphatic when standing independently and is followed by a word that is plain, the emphasis does to extend to the following word, e.g. /+ʾal/-/ʾídi/

‘on my hand’

In certain cases, however, when the first element is closely bonded grammatically with the following word, the first element is within the domain of emphasis of the second word, e.g. in compound nouns such as: /+bra sup̂p̂a/ /

‘thimble’ ‘glove’

In the first example, the bra element harmonizes with the emphasis of +sup̂ p̂a, whereas in the second it harmonizes with the plainness of ʾida. This is represented by the double hyphen. The form bra is derived historically from the old construct state of brata ‘daughter’ (bra < *brat). Although brata is plain, the relationship of the form bra to this lexical item is now completely opaque, especially since it is referring to an inanimate object, and it is devoid of lexical content. Conversely, tightly-knit phrases that contain elements whose original lexical content is still identifiable maintain their own autonomous domains of emphasis and each component is

DOMAINS OF EMPHASIS, SYLLABLE STRUCTURE AND MORPHOLOGICAL BOUNDARIES 147 treated like a separate word. This applies also to old construct forms and contracted words, e.g.: /+bar/-náša/ ‘son-man’ = ‘human being’ (+bar, cf. bruna) + / xvār-/də́k̭na/ ‘white-beard’ = ‘old man’ /+ʾarp̂ ó/-šiba/ ‘four in-week’ (< +ʾarp̂a-b-šiba) = ‘Wednesday’ /bət-/+k̭úra/ ‘house-grave’ = ‘cemetery’ (bət < betət) + /bax-/ mámu/ ‘wife-paternal uncle’ = ‘the wife of a paternal uncle’ (bax < baxtət) In all these examples the lexical content of the two components is still identifiable and they have autonomous domains of emphasis. The placement of the stress on the numeral component of names of days of the week is a further sign that the two components of such phrases are treated as separate words.2 An apparently exceptional case is that of the tens and hundreds paradigms of numerals. The word for ‘hundred’ is plain (ʾəmma), but it harmonizes with the first decade cardinal that is attached to it, e.g.: ʾə́mma trəmma + ṱləmma ~ +ṱlamma ʾarbə́mma, +ʾarp̂ ámma

‘one hundred’ ‘two hundred’ ‘three hundred’ ‘four hundred’

The explanation seems to be that these are analysed as unitary lexical items and not compounds, and the final element -mma is treated like a derivative morpheme. The same applies to the tens paradigm, the forms of which are treated as unitary lexical items with a derivative morpheme -(s)sar, e.g.: taryəssar ṱalṱásar

+

‘twelve’ ‘thirteen’

Prepositions before nouns have their own autonomous domain of emphasis: /+ʾal/-/ʾídi/ /mən/-/+ṱuyra/

‘on my hand’ ‘from the mountain’

The domain of emphasis is a phonological phenomenon. The boundaries of the domain are observed in slow careful speech. In fast speech, however, phonetic coarticulation sometimes breaches the boundaries. In such cases the setting of the main lexical word affects that of the first element. In the stress group +ʾal-ʾidi, for exam-

2

In the Jewish Urmi dialect several corresponding phrases have a single domain of emphasis, e.g. +xwar-dəqna ‘old man’, +bel-šultana ‘house of the king’ (Hoberman 1989: 85).

148

GEOFFREY KHAN

ple, the plain setting of the lexical word ʾidi may be extended to the first element, resulting in the realization [ʔal-ˈʔiːdiˑ]. Likewise in the stress groups mən-+ʾúllul, lá+ xəllə, the emphatic setting of the second word can spread to the first: [m̴ə̴n-̴ ˈʔ̴ul̴lu ̴ l̴], [ˈl̴ɑ-̴ x̴ə̴l̴lə̴ ̴]. Since this is a phonetic coarticulatory phenomenon, it is not marked in the transcription.

2. SYLLABLE STRUCTURE

At the beginning of a word the onset of a syllable may consist of a single consonant C or a cluster CC, e.g.: pá.təx ptə́x.lə

‘he opens’ ‘he opened’

A consonant cluster at the beginning of a word is generally not split by an epenthetic vowel. Some examples of epenthetics between the consonants, however, are found. This occurs, for example, after a tense dorsal /k̭/, e.g.: k̭dála [ḵədaːla]

‘neck’

This would have the following syllable structure: k̭da.la. Initial clusters may be of decreasing sonority, e.g.: npə́l.lə rxə́ṱ-lə

‘he fell’ ‘he ran’

Clusters at the onset of a syllable are expected to exhibit rising sonority onsets (Ewen and Hulst 2001: 136–141; 147–150). The fact that the clusters can be of decreasing sonority, beginning with sonorants, suggests that in such cases the first consonant of a cluster should be analysed as extra-syllabic, thus: n.pə́l.lə

‘he fell’

Word-initial strings with decreasing sonority such as npəllə are often pronounced with epenthetic prosthetic vowels, which is evidence for this analysis of the syllable structure, e.g.: nšə́k̭lə + nṱúrrə rjə́dlə + rxə́ṱlə spadíta švik̭áli

[ɪnˈʃɪḵlɪ] [u̴n̴ˈt̴u̴rr̴ ə̴ ]̴ [ɪɾˈʤɪdlɪ] [ə̴rˈ̴ x̴ət̴ ̴lə̴ ]̴ [ɪspaˈdiːtʰa] [ɪʃviːˈḵaːliç]

‘he kissed him’ ‘guard it’ ‘it trembled’ ‘he ran’ ‘pillow’ ‘I left her’

DOMAINS OF EMPHASIS, SYLLABLE STRUCTURE AND MORPHOLOGICAL BOUNDARIES 149 The prefixing of a prosthetic vowel is a permanent feature of several words that are historically monosyllabic, in which the originally epenthetic vowel has been reanalysed as a syllable nucleus. The prosthesis is exploited to augment the word to two syllables. These prosthetic vowels, which bear the stress, are represented in the transcription. Note that in all cases the cluster is of falling sonority, which disfavours syllable onset formation, e.g.: ʾə́rxə ʾə́šta ʾə́ltəx

‘water-mill’ ‘six’ (m.) ‘below’

< *rxe < *raḥyā < *štā < *l-tēx

When a prefixed element consisting of one consonant is attached to a word beginning with a cluster CC, a cluster consisting of three consonants is created in the onset CCC. An epenthetic is obligatorily placed between the particle and the remaining consonants, e.g.: dətré [dɪˈtreː] + dəṱlá [d̴əˈ̴ t̴l̴ɑ̴ː]

‘the second’ ‘the third’

In a case such as the ones cited, the stress position demonstrates that the vowel is an epenthetic, since the canonical position of word stress is on the penultimate syllable. The monosyllabic word tre takes final stress although a vowel exists after the /d/. If the vowel after the initial /d/ were a syllable nucleus, one would expect the stress placement to take it into account and the stress would be placed on this syllable due to its penultimate position, the canonical place for stress. It should be interpreted, therefore, as an epenthetic. Since, however, it is obligatory in such positions, it is represented in the transcription. Enclitic elements that are attached to the end of a host word may have a zero syllabic onset, e.g.: ʾúl.lul.i.lə (= +ʾúllul ə)

+

‘He is above’

This is permitted since the copula ilə does not occur in word-initial position. Zero onsets are not tolerated at the beginning of a word in the phonetic form of the word, although at an underlying level words sometimes have a zero onset. Such words acquire an initial laryngeal /ʾ/, which fills the zero onset slot, e.g.: /∅ilə/ > ʾilə /+∅axəl/ > +ʾaxəl

‘he is’ ‘he eats’

The rhyme of a syllable may consist of a vowel nucleus only or may consist of a vowel nucleus and a following consonant. The phonetic duration of vowels varies according to the position in a word. There is reason to believe, however, that on a phonological level syllable rhymes should be analysed as VV or VC, i.e. they have the weight of two moras.

150

GEOFFREY KHAN

In an open stressed syllable a vowel is long. A syllable with a short vowel nucleus in the same conditions has to have the same weight of rhyme, i.e. VC: pá.təx CVV.CVC

‘he opens’

šám-max CVC-CVC

‘we hear’

+

It can be assumed, therefore, that open and closed syllables are of the same phonological moraic weight also in other conditions, although the phonetic duration of the vowel or of a geminated consonant may be reduced due to the prosody of the utterance. Although a word-final open syllable is phonologically long (bimoraic), the second vowel mora tends to lose audibility due to glottalization. This applies in particular to stressed word-final syllables in stress groups such as xá-yuma ‘one day’. In such cases the glottalization fills the timing of the second mora (marked here with the symbol ˀ.3 The symbol # is used to mark a word boundary: [ˈxaˀ-juːmaː] xá.-yu.ma CVV.#CVV.CVV#

In their citation form and in pause at the end of the intonation group, wordfinal vowels are generally long, and within intonation groups their duration is reduced. This can be regarded as a prosodic phenomenon independent of the phonological length. The symbol | is used to mark an intonation group boundary: pát.xa| CVC.CVV#

[ˈpʰatʰ.xaː]

‘she opens’

pát.xa CVC.CVV#

[ˈpʰatʰ.xa]

‘she opens’

The bimoricity of such word-final vowels is demonstrated by the fact that they are clearly phonetically long when they take the stress in word-internal position after the attachment of a suffix, e.g.: pat.xá.va [pʰatʰ.ˈxaː.va] ‘she used to open’ CVV.CVV.CVV#

3

For a similar phonological interpretation of word-final glottalization in Italian see Vayra (1994).

DOMAINS OF EMPHASIS, SYLLABLE STRUCTURE AND MORPHOLOGICAL BOUNDARIES 151 Other evidence for their bimoricity is the fact that the second mora of final -i surfaces as the palatal [ç] in pause before an intonation group boundary due to glottalization: bé.ti| CVV.CVV#

[ˈbeː.tʰiç]

‘my house’

The vowel of penultimate post-stress syllables also reduces in phonetic duration when not in pause at the end of the intonation group, e.g.: bé.tə.lə [ˈbeː.tʰɪ.lɪ] CVV.CVV.CVV#

‘it is a house’

The rhyme VC occurs also in final closed syllables, e.g.:

p.túx C.CVC#

[p.tʰux]

‘open!’

Before intonation group boundaries the vowel of a word-final CVC syllable is often extended phonetically, but this need not affect the analysis of its phonological mora weight, since it is likely to be a prosodic phenomenon independent of the syllable’s phonological weight. Its bimoricity is demonstrated by the fact that when it is put in word-internal position by the addition of a suffix, it is always phonetically short: p.túx.lə C.CVC.CVV# Certain words contain a long vowel in closed syllables even when a suffix is added. These forms have developed by contraction of two syllables into one, e.g. + ́ ṱāl ‘he plays’, +ṱālta ‘game’. The syllable structure of these forms can be analysed thus: +

ṱā.l CVV.C#

‘he plays’

ṱā.l.ta CVV.C.CVV#

‘game’

+

Here the consonant following the long vowel is extrasyllabic and does not affect the bimoraic weight of the syllable. Evidence for this analysis is that the contracted syllable can be optionally restored, e.g. + ́ ṱā.l ~ +ṱá.vəl ‘he plays’ CVV.C#

CVV.CVC#

152

e.g.:

GEOFFREY KHAN A CC cluster may occur in the onset of a syllable that occurs word-internally, +

pur.ṱk̭al.ta CVC.CCVC.CVV#

‘orange’

xər.ṱman.ta CVC.CCVC.CVV#

‘chickpea’

+

If the second consonant of the word-internal sequence CCC is more sonorant than the third, the syllabification C.CC is not favoured, since the onset would have falling rather than rising sonority. In such cases one finds forms such as mắdənxa ‘east’. These exhibit several apparently irregular properties, a short stressed vowel in an open syllable and the incidence of stress two syllables back from the end of the word instead of in the canonical position on the penultimate syllable. The explanation is that the /ə/ is an epenthetic vowel and at an underlying level the initial syllable is closed: mắdənxa /ˈmad.n.xa/ CVC.C.CVV

The vowel length and the stress placement are determined by the underlying syllable structure. The fact that the epenthetic occurs before the /n/ is evidence that this consonant is not in the onset of the final syllable. As indicated, the /n/ would be extra-syllabic, just as it is in a form such as n.pə́l.lə ‘he fell’, in which the marked sonority sequencing is resolved in the same way. Short vowels that were originally epenthetics have often become phonologized as syllable nuclei in C. Urmi. This applies, for example, to words such as čalə́bta ‘bitch’ cf. šarə́xta ‘heifer’ cf. səppə́rta ‘sparrow (f.)’ cf.

čálba ‘dog’ šárxa ‘calf’ sə́pra ‘sparrow’

In the related C. Salamas dialect these words have the form cəlba ‘dog’, cilə́bta ‘bitch’; šərxa ‘calf’, širə́xta ‘heifer’. In the history of the C. Salamas dialect there was a shift of short /a/ > /ə/ in closed syllables. It can be assumed that originally a word such as C. Urmi čaləbta had a syllable structure such as that of mắdənxa, i.e. */čal.b.ta/, and C. Salamas ciləbta would have originally had the syllable structure */cəl.b.ta/. These would have been realized as *čắləbta and cə́ləbta respectively, since the epenthetic before the /b/ would have been ignored and the first syllable would have been closed at an underlying level and the vowel short. Subsequently the epenthetic became analysed as a syllable nucleus. As a consequence, the stress

DOMAINS OF EMPHASIS, SYLLABLE STRUCTURE AND MORPHOLOGICAL BOUNDARIES 153 shifted onto it and the initial syllable was opened, resulting in the lengthening of the vowel: C. Urmi čaləbta [ʧaːlɪptʰa], C. Salamas kiləbta [cʰiːlɪptʰa]. In the causative verbal pattern III and quadriliteral verbs an original epenthetic between the two final radicals has likewise become phonologized, with the result that it takes the stress and is treated as a syllable nucleus with canonical rhyme length achieved by geminating the following consonant, e.g. *ham.zma > ham.zə́m.ma

‘she speaks’

Some loanwords end in an consonantal cluster, e.g. +dost ‘friend’, +xiyavand ‘road’. These clusters only occur in word-final position and the last consonant can be regarded as extrasyllabic, so the syllable is bimoraic: dós.t xi.yá.van.d

+ +

‘friend’ ‘road’

Following the analysis by Kiparsky (2003) of Arabic syllable structure, we may say that such unsyllabifiable consonants in forms such as n.šuk̭ ‘kiss’, +ṱā.l ‘he plays’, mădənxa ‘east’ /mad.n.xa/ and +dos.t are licensed by moras adjoined to the higher node of the prosodic word rather than the syllable node. Kiparsky refers to these consonants as ‘semisyllables’. In the following trees ω = word node, σ = syllable node, μ = mora. In the phonological transcriptions of the trees in slanted brackets syllables are enclosed in rounded brackets and consonants left outside of the rounded brackets are extrasyllabic:

154

GEOFFREY KHAN

μ n

š

ω

ω

σ

σ

μ

μ

u

k̭

ṱ

/n.(šúk̭)#/

μ

μ

a

a

l

/(ṱáa).l#/

ω

ω

σ

d

μ

σ

μ

μ

μ

o

s

t

/(dós).t#/

m

σ

μ

μ

μ

a

d

n

μμ x

aa

/(mád).n.(xaa)#/ [ˈmadənxa]

Loanwords that remain unadapted to the phonological and morphological system of the NENA dialect exhibit the syllable structure of the source language. In loanwords that are adapted to the phonology of the dialect short open syllables are generally changed to bimoraic CVV syllables, e.g.: pári [ˈpaːɾiˑ] ‘fairy’ < Pers. părī ́ CVV.CVV

háva [ˈh̴ɑ̴ːv̴ɑ̴ˑ] CVV.CVV +

‘air, weather’ < Pers. (< Ar.) hăvɑ̄́

Short back rounded vowels in open syllables in the source language may be lengthened and be realized as the bimoraic diphthong /uy/, e.g. + k̭úymar [ˈḵujm̴ɑr̴ ̴] ‘gambling’ < Pers. (< Ar.) qŏmɑ̄ŕ CVV.CVC

When the prefixed particle b, which consists of a single consonant, creates an initial CC cluster, a vowel often splits the cluster to ensure that the particle is kept distinct from the following word. The inserted vowel may be short [ɪ] or [ə], in

DOMAINS OF EMPHASIS, SYLLABLE STRUCTURE AND MORPHOLOGICAL BOUNDARIES 155 which case the following consonant is generally geminated, or the vowel may be long /i/, e.g.: b-síla [ˈpsiːla] b-síla [bɪsˈsiːla] b-síla [biːˈsiːla]

‘with sand’

The syllabification of the forms with the inserted vowel can be interpreted as the phonologization of an epenthetic as a vowel nucleus. As a result of this the syllable rhyme has to have the canonical bimoraic weight, which is achieved by geminating the following consonant or lengthening the vowel. The realization with the lengthened vowel, such [biːˈsiːla], can be regarded as reflecting a looser connection of the preposition with the noun than is the case with the other realizations. Since various syllabifications are attested, including the clustering of the b with the following consonant, it is preferable to normalize the transcription to b- in all contexts. This includes where the b- is followed by two consonants and is regularly followed by an epenthetic, e.g. b-prəzla [bɪpʰrɪzla]

‘with iron’

An epenthetic keeps the particle b distinct before labial consonants, e.g.:

b-báxtu [bɪbˈbaxtu] b-momíta [bɪmmoːˈmiːtʰa]

‘on his wife’ ‘with an oath’

The preposition b can be separated from a noun by a hesitation, in which case it is followed by a long vowel, e.g. bí ... k̭ésə [biː …. ˈḵeːsɪː]

‘with wood’

The phonotactics of b are different before progressive verbal forms, which are derived historically from the combination of the preposition b- with a following infinitive form. If the b is followed by a single non-labial consonant, this consonant is regularly geminated and the realization with a lengthened vowel, reflecting a looser connection, is not permitted. Nor can the b be separated from the verb by a hesitation. The epenthetic has been phonologized as a short vowel nucleus and has, indeed, also been grammaticalized, in that it is treated as an integral part of the morphological base of the progressive verbal form rather than an affix. The gemination and epenthetic, therefore, should be represented in the transcription, and hyphens should not be used. The b is always in the domain of emphasis of the verb, which is not regularly the case with b before nouns, e.g.: bərrák̭ələ + bəddáyələ

‘he is running’ (< b + rak̭ələ) ‘he knows’ (< b + +daya)

156

GEOFFREY KHAN

Likewise, the epenthetic after the particle b before CC clusters in such progressive verbal forms should be represented in the transcription, e.g. +

bəṱlabələ

‘he is requesting’

In such verbal constructions, however, the b is generally not kept distinct before labials and it is, in most cases, elided, e.g.: bə́xyələ vádələ páyələ márələ prak̭ələ

‘he ‘he ‘he ‘he ‘he

is weeping’ (< b + bəxyələ) is doing’ (< b + vadələ) is baking’ (< b + payələ) is saying’ (< b + marələ) is finishing’ (< b + prak̭ələ)

This reflects the higher functional load of the particle b before nouns than in these verbal constructions before the infinitive stem. It also demonstrates that at some stage of the derivation the particle b must have been in contact with an initial labial without an intervening vowel. Since there are grounds for arguing that the epenthetic in forms such as bərrakələ and +bəṱlabələ has now been phonologized as a vowel nucleus and is a component of the morphological pattern of the verbal base, the contact in question is best considered to have existed at some earlier historical period rather than to exist synchronically at an underlying level of derivation. The foregoing behaviour of b before a labial does not apply to the combination of b with a noun, in that it is always kept distinct before a noun beginning with a labial, reflecting its greater functional load, e.g.: b-+mxáṱa [b̴əm ̴ ̴ˈx̴ɑ̴ːt̴ɑ̴] b-+p̂ láša [b̴əˈ̴ p̴l̴ɑ̴ːʃ̴ɑ̴]

‘with a needle’ ‘by war’

This applies also to verbal nouns, which behave like nouns in this respect rather than verbs, e.g.: b-prák̭ta [bɪpɾaḵtʰa] prák̭ələ

‘at the end’ (noun) ‘he is finishing’ (verb)

The future preverbal particle sometimes has the form of a single labial segment b or p, which are shortened forms of bət. These behave differently from the progressive preverbal particle b, in that there is no restriction on the clustering of the particle with an initial labial, e.g.: b-bášəl [ˈbbaːʃɪl] p-párək̭ [ppʰaːˈrɪḵ]

‘he will cook’ ‘he will finish’

In the last example the particle is devoiced before the voiceless /p/. This clustering of the particle with an initial labial can be interpreted as reflecting that, unlike in progressive constructions, there is a word division between the particle and

DOMAINS OF EMPHASIS, SYLLABLE STRUCTURE AND MORPHOLOGICAL BOUNDARIES 157 the verb, and that at an underlying level it does not cluster in the onset of the first syllable of the verb. One could posit an underlying form of the particle bət which reduces phonetically to a single segment late in the derivation, viz. /bət#bašəl/ > [b#baːʃɪl] /bət#parək̭/ > [p#pʰaːrɪḵ] This analysis is supported by the fact that the particle bət can be separated from the verb by a hesitation, e.g. bə́t … +ʾáxəl

‘he will eat’

3. MORPHOLOGICAL BOUNDARIES

A ‘word’ in an informal sense consists of a number of different phonological domains, the boundaries of which do not necessarily coincide. The domain of suprasegemental emphasis, as we have seen, includes following enclitic elements. Other phonological domains, however, recognize a boundary between the main word and clitics. This is the case with the domain of a word-final devoicing rule and a wordstress placement rule. Word-final voiced stops, fricatives and affricates consonants are devoiced. Voiced stops are devoiced to aspirated unvoiced stops. Examples: máġġəb ʾávəd bárəz ȷâ vəȷ ̂

[ˈmaɣɣɪpʰ] [ˈʔavɪtʰ] [ˈbarɪs] [ˈʣavɪʦʰ]

‘he loves’ ‘he does’ ‘it dries’ ‘it moves’

This applies also to loanwords, e.g.: xiyávand [x̴iˈjɑ̴ːv̴ɑ̴n̴t̴ʰ] k̭ássab [ˈḵɑ̴s̴sɑ ̴ ̴p̴ʰ] + časəb [ˈʧ̴ɑ̴sə̴ p ̴ ʰ] + sárbaz [ˈs̴ɑ̴r̴bɑ ̴ ̴s̴] dūz [duːs] + ʾótaġ [ˈʔ̴o̴ːt̴ɑ̴x̴] + +

‘road’ (Pers.) ‘butcher’ (Pers. < Ar.) ‘poor’ (Pers. < Ar.) ‘soldier’ (Pers.) ‘true’ (Azer.) ‘room’ (Pers. < Turk.)

When inflectional or derivative affixes are attached to the end of phonologically integrated loanwords the consonants retain their voicing, e.g.: ʾavə́dva [ʔaˈvɪdva] barə́zva [baˈrɪzva] + xiyavándə [x̴iˈjɑ̴ːv̴ɑ̴nd ̴ ə] + k̭assábə [ḵɑ̴sˈ̴ s̴ɑ̴ːb̴ə]̴ + ʾotáġə [ʔ̴o̴ːˈt̴ɑ̴ːɣə] + sarbázan [s̴ɑr̴ ̴ˈb̴ɑ̴ːz̴a̴n̴]

‘he used to do’ ‘it used to dry’ ‘roads’ ‘butchers’ ‘rooms’ ‘our solider’

158

GEOFFREY KHAN duzúyta [duːˈzujtʰa]

‘truth’

A phonological transcription of the final consonant in such forms should represent it as voiced in all contexts, as can be seen from the examples above. When enclitics are attached to the end of such words, the final consonant remains unvoiced, e.g.: + +

sárbaz ə sárbaz

[ˈs̴ɑ̴r̴bɑ ̴ ̴sə̴ ̴lə̴ ̴] ‘he is a soldier’ [ˈs̴ɑ̴r̴bɑ ̴ ̴sd ̴ ɑ ̴ ̴] ‘also a soldier’

Notice also that the stress does not shift, as it does in forms with inflectional and deriviational affixes, but remains in the position it has when there is no enclitic, i.e. on the penultimate syllable. The domains of word-final devoicing and stress are, therefore, the main word excluding enclitic elements. This differs, therefore, from the domain of emphasis, which includes the enclitics. Differences in phonotactic boundaries within similar looking constructions can result in a different phonological form. Consider the pair: ṱ-ilə

‘which is’

do

‘of that’

In both forms the initial consonant is historically the voiced annexation particle *d. In the first form ṱ-ilə the particle has become a unvoiced unaspirated /ṱ/. This is the result of a coalescence of *d with an following laryngeal: d + ʾilə > ṱ-ilə. The copula has this initial laryngeal when standing independently: ʾilə ‘he is’. The fact that the particle in do has not been devoiced shows that it was combined with the following demonstrative without an intervening laryngeal: d + o, despite the fact that the demonstrative has an initial laryngeal when standing independently: ʾo ‘that’. This can be explained by the hypothesis that the initial laryngeal is not part of the word at some underlying level of the derivation but is added by a phonotactic rule that requires word-initial vowels to be preceded by minimally by a laryngeal onset, thus: #∅ilə > #ʾilə According to this rule, therefore, there is no word boundary after the initial particle in do but rather the form is treated like a single word and the d is an integral element of this word, hence its transcription without a hyphen. The form ṱ-ilə, however, is treated by this rule as two words combined in a single stress group, since the word-initial laryngeal placement rule operated on the ʾilə before it was combined with the particle d. We can summarize this as follows using the symbol # to indicate the morphological boundary: d + #∅ilə > ṱ-ilə

#d + o > do

DOMAINS OF EMPHASIS, SYLLABLE STRUCTURE AND MORPHOLOGICAL BOUNDARIES 159 The phonetic rule that adds a laryngeal /ʾ/ to the front of words that have a zero onset vowel does not apply to a word that is attached to a preceding word as an enclitic, e.g.: ʾ ə sárbaz ə

‘He is above’ ‘He is a soldier’

+ +

As we have seen, such enclitics are included within the domain of emphasis of the word to which they are attached but they are not integrated into the word with regard to the features of stress placement and word-final devoicing. The syllable structure of the host word, furthermore, ignores their presence. This is seen by the fact that in a form such as +ʾúllul ə, the final syllable of ʾullul continues to be treated as a closed syllable with a VC rhyme after the addition of the enclitic beginning with a vowel (+ʾul.lul.i.lə). The syllable structure is not adjusted by lengthening the vowel or geminating the final consonant of this syllable. Another issue relating to morphological boundaries is illustrated by forms such as the following: b-+xalta

‘with eating’

+

bixálələ

‘he is eating’

The first form consists of a preposition b and a verbal noun. Such a phrase is realized phonetically in a variety of ways: [b̴x̴ɑ̴lt̴ ɑ ̴ ̴], [b̴ə̴x̴xɑ ̴ ̴l̴tɑ ̴ ̴], [biːx̴ɑl̴ ̴tɑ ̴ ̴]. When the b clusters with the initial consonant of +xalta or is syllabified with the initial consonant through the gemination of this consonant, the b is emphatic. When the b is more independent from the following word in the realization [biːx̴ ɑ̴lt̴ ̴ɑ]̴ , it tends not to be pronounced emphatic. The long [iː] in this form is a lengthened epenthetic. Due to these various realizations, the transcription of the preposition in such contexts before nouns has been normalized to b- in all contexts. In the similar looking form +bixálələ, on the other hand, which developed historically from the combination of the preposition b with the infinitive, the initial /b/ is always followed by long /i/. The explanation is that the verbal root in question at some stage of historical development was treated as initial /y/: y-x-l and the infinitive was originally *ixala. The /y/ of the verb was elided in word initial position before a consonant, e.g. +xala ‘to eat’, +xalta ‘eating, food’, +xə́llə ‘he ate’, and the root came to be interpreted as ∅-x-l.4 The /y/ was preserved, however, when not in word-initial position, e.g. in the noun +mixúlta ‘food’. This would apply to +bixálələ, in that the word boundary is interpreted as falling before the /b/ whereas the word-boundary of b-+xalta is after the /b/, since, as we have seen earlier, the /b/ in the progressive form +bixálələ is

4

In C. Urmi slots in verbal roots can be filled by a zero radical ∅ (see Khan forthcoming).

160

GEOFFREY KHAN

now analysed as a component of the morphological pattern of this form rather than as an affix: b # +xalta

#+b ixalələ

The domain of emphasis regularly includes the b in the verbal form +bixalələ, as is the case with +mixulta. One word may be annexed to the following word by the annexation clitic -ət, e.g. ʾídət málča

‘the hand (ʾida) of the king (malča)’

On some occasions the annexation particle is omitted and the first word is combined more tightly with the second: ʾīd-málča

In this phrase, the vowel of the form ʾīd remains long, as if it were still in an open syllable. It is transcribed by a macron, since vowels in closed syllables would normally be expected to be short. The syllable structure of ʾidət and ʾīd would be, therefore: ʾi.dət CVV.CVC

ʾī.d CVV.C

The final consonant is extrasyllabic, or a ‘semi-syllable’ to use Kiparsky’s (2003) terminology. One can compare this to the syllabification of forms such as +ṱāl (CVV.C), which is a contraction of +ṱavəl (CVV.CVC). One may say that the underlying form of ʾīd is ʾidət. When after the loss of the annexation particle the first word ends in a voiced stop, it does not devoice, e.g. ʾīd-málča [ʾiːdˈmalʧa]

‘the hand of the king’

< ʾídət málča

One can interpret this as reflecting that the rule of word-final devoicing operates on the underlying form ʾidət, in which the /d/ is not word-final, and so does not devoice.

REFERENCES

Ewen, Colin J., and Harry van der Hulst. 2001. The Phonological Structure of Words. Cambridge: Cambridge University Press. Hoberman, Robert. 1989. “Parameters of Emphasis: Autosegmental Analyses of Pharyngealization in Four Languages.” Journal of Afroasiatic Languages 2 (1): 73–98. Khan, Geoffrey. 2013. “Phonological Emphasis in North-Eastern Neo-Aramaic.” In Base Articulatoire Arrière. Backing and Backness, edited by Jean Léo Léonard and Samia Naïm, 111–132. Munich: Lincon Europa.

DOMAINS OF EMPHASIS, SYLLABLE STRUCTURE AND MORPHOLOGICAL BOUNDARIES 161 —. forthcoming. The Neo-Aramaic Dialect of the Assyrian Christians of Urmi. Kiparsky, Paul. 2003. “Syllables and Moras in Arabic.” In The Syllable in Optimality Theory, edited by Caroline Féry and Ruben Florentius Hendricus Eduardus van de Vijver, 147–182. Cambridge: Cambridge University Press. Tsereteli, Konstantin. 1972. “Neuaramäisch.” In Von Nag Hammadi Bis Zypern, edited by Peter Nagel, 47–64. Berlin: Akademie-Verlag. —. 1976. “Obrazcy Sovremennoj Assirijskoj Reči (salamasskie Teksty).” In Semitiskie Jazyki, edited by Grigorii Shamilevich Sharbatov, vol. 3, 208–219. Moscow: Nauka. —. 1978. The Modern Assyrian Language. Moscow: Nauka. Vayra, Mario. 1994. “Phonetic Explanations in Phonology: Laryngealization as the Case for Glottal Stops in Italian Word-Final Stressed Syllables.” In Phonologica 1992: Proceedings of the 7th International Phonology Meeting, edited by Wolfgang U. Dressler, Martin Prinzhorn and John R. Rennison, 275–293. Torino: Rosenberg&Sellier. Younansardaroud, Helen. 2001. Der Neuostaramäische Dialekt Von Särdä:rïd. Semitica Viva Bd. 26. Wiesbaden: Harrassowitz.

THE HISTORY OF THE FUTURE SAMUEL ETHAN FOX We are accustomed to thinking of the future as essentially comparable to present and the past, as merely the right-hand portion of the time spectrum that begins in the past and centers on the present. We understand that languages divide this spectrum differently, but we see it as essentially an unbroken unity. Krotkoff’s classic diagram (Krotkoff 1982: 31) of the tenses of Aradhin Neo-Aramaic reflects such a view, with the past on the left, the present as a central pale, and the future on the right. In this diagram, the future tense differs from the general imperfect and the general present only in its position on the spectrum. The relevant lines of the diagram are reproduced below: Time Continuum Psychological Present ------------------------------------------------------------------------------------------------ General Present -- -- -- -- -- - --|-- -- -- -- --|-- --  i-palxin IND-work.1SG General Imperfect -- -- | i-palxin-wa IND-work.1SG-CONV

|

|

Future bet-palxin FUT-work.1SG

|

|

|-- -- -- 

In fact, this is a natural way of looking at the future, and time in general, one which embodies the metaphor described by Lakoff and Johnson as “Time is stationary, and we move through it” (Lakoff and Johnson 1980: 43). This metaphorical view of the passage of time is manifest in such expressions as ‘As we go through the

THE HISTORY OF THE FUTURE

163

years’, and this rather breathless sentence from the diary of Queen Victoria “This day I go out of my teens and become twenty!”1 However, a contrary view of the future, according to which it differs from the present and past in almost every way, is also widely held, and with reason. The future, in this view, is different in meaning, form, and development, and perhaps should even be banished from the category of tense altogether. First of all, the future is unreal, since events in its realm have not yet occurred. While the present and past can be regarded as simply stating the facts, it is impossible to say anything about the future without an element of the speaker’s attitude: the future is a prediction, and expresses the speaker’s expectation. The future “necessarily involves an element of prediction or some related modalization” (Fleischman 1982: 24). Thus, it is closely allied to modal forms that express the speaker’s assessment of the likelihood of an event or his wish that it may occur. The future also typically differs from other tenses in form. It may be indicated by a particle or auxiliary where the present and past are marked by affixes or stem changes. Further, in some languages the future is not marked at all. In these cases context or adverbial elements tell the listener when future reference is intended. Even where there is a formal marker of the future tense, it may be optional or restricted to some cases. Consider the English ‘We’re leaving for Brecon in the morning’. where the future is unmarked, alongside ‘We will leave for Brecon in the morning’. where the future is explicit. Further, distinctions that are made elsewhere in the verbal system of a language are often attenuated in the future. For example, the past and present may express aspectual distinctions which are neutralized in the future, or the future may not have a separate negative, this being replaced by some other tense form. Turning now to Aramaic, we will review the future as it appears in the various forms of Middle Aramaic, and then in some of the dialects of Neo-Aramaic. We will discuss the developments which connect the present of the future with its past. Finally, we will consider these developments in the light of the broad investigation of the process of grammaticalization which is presented in an important work on this process, The Evolution of Grammar (Bybee, Perkins and Pagliuca 1994). Jewish Babylonian Aramaic (JBA) is one of the three attested forms of Eastern Middle Aramaic, the others being Syriac and Mandaic. JBA is the language of the Aramaic portions of the Babylonian Talmud and the somewhat later Gaonic documents. As regards tense usage, JBA is in the midst of a transition. Earlier forms of Aramaic used the imperfect to express the future, and while this is still common in 1

Quoted in Sampson and Sampson (1988: 50).

164

SAMUEL ETHAN FOX

JBA, the future can now also be expressed by the present participle, sometimes with the prefix qa-, a form which most often expresses the present. Alongside the imperfect and the present participle a third construction has also emerged: the present participle of the very common verb bāʿē ‘to want’ with a following infinitive. This construction is used for the imminent future, that is for an event which is about to occur. For example: (1) ḥazy-eh see.PRF-3SG.M bāʿē want.PTCP.3SG.M

də-havā REL-be.PRF.3SG.M lǝ-mihăḏar

PRP-return.INF

qā

PRP/PRET

lǝ-mizkē

PRP-acquire.INF

bǝ-hu in-3PL

‘He saw that he was about to acquire them again’. (Baba Meṣi’a 30b) (2) kǝ-də-vāʿēnan as-REL-want.PTCP.1PL

lǝ-memar PRP-say.INF

lǝ-qamān PRP-below

‘as we are about to say below’ (Baba Meṣi’a 21b) Syriac uses both the imperfect and the active participle for the present and the future. According to Nöldeke (1904: 208), the imperfect is still commonly used for the future and for cases where “there is the slightest modal colouring”, as well as for “a wish, request, summons, or command.” but rarely for the present. The active participle has almost entirely replaced the imperfect as an expression of the present and is also often used for the future. (Nöldeke 1904: 211). In Mandaic, the third attested dialect of Middle Aramaic, the Imperfect is used for the future as well as the present (Macuch 1965: 430). After the preposition ad- or alma ḏ- ‘until’ the active participle is used with future meaning. So outside of JBA, the future is expressed either by the fading imperfect, or by the rising active participle, and there is no special marker which distinguishes the future from the present. Let us turn now to the modern dialects, beginning with NENA. In NENA the future is expressed by the descendent of the Middle Aramaic active participle which I will refer to, following Hetzron (1969), as the J stem, with a prefix of the form b- or bǝt-. These prefixes have been held since Nöldeke to be descended from a form like bāʿē d- or bʾē d-, the imminent future that we saw in JBA. This contrasts with the situation in JBA, where bāʿē is followed by the infinitive, a form which no longer exists in NENA. Middle Aramaic constructions of a finite verb followed by an infinitive have been replaced in NENA by two finite verbs linked by the prefixed subordi-

THE HISTORY OF THE FUTURE

165

nator d-. With the loss of the infinitive, bāʿē, folowed by the subordinator d- evidently came to be preposed to the active participle, and the sequence bāʿē d- was then reduced to the attested forms. In many NENA dialects the future is sometimes also expressed by the J stem with no prefix, or with the prefix for the present tense, and in most cases when the future is negated the prefix is lost. These two phenomena are instances of the unconditioned and conditioned loss of the future marker in some cases, which we noted also occurs in English and in many other languages. In the following table, the forms of the prefixes in a selection of dialects are given in the first column. The second column shows other prefixes, or the absence of a prefix, which can be used with the J stem to express the future. The third column shows the prefix, or absence of prefix, which is used when the future is negated.

Dialect Jewish Urmi

2

Jewish Betanure

3

Future Prefix

Alternate

Negative

b-, be-

∅

∅/b-

b-

∅

(k-)

gǝbe

immediate future

Jewish Barzani Aradhin5

4

bǝd ~ b-

∅

y-/g-

bed-, bet-, bd-,

?

∅

ptJilu

b(t)-,

∅

∅

Qaraqosh 7

bǝt-, bd-, b-, d-

k-

k-

Bespin

b-

-

?

6

8

Hertevin Bohtan 10

2

9

bǝd ~ bbǝt/bǝd, b-

Garbell (1965: 65–68). Mutzafi (2008: 54, 84). 4 Mutzafi (2002: 59–60). 5 Krotkoff (1982: 33). 6 Fox (1997: 32–33). 7 Khan (2002: 98, 100). 8 Sinha (2000: 104). 9 Jastrow (1988: 54–55). 10 Fox (2009: 56). 3

∅ ∅

∅

166

SAMUEL ETHAN FOX

To give one example of a dialect, in Bohtan the future tense is formed from the J stem, preceded by the future particle bǝt/bǝd or its short form, the prefix b-. The subject is marked by the A set suffixes and the object is marked by the L set suffixes. It is used to indicate future action. (3) sǝttiya lady

b-ačča FUT-come.3SG.F

palg-ǝd half-REL

yoma. day

‘The lady will come at noon’. (Fox 2009: 56) (4) bǝt-howe FUT-be.3SG.M

toma. ADV.there

‘It will be there’. (5) oná-se I-too

bǝd-zoli FUT-go.1SG

hawdax this.INDF.PRON

b-awdena. FUT-do.1SG

‘I will also go and will do that’. (6) ból-ahan b-howe face-POSS.3PL FUT-be.3SG.M

l-áġdode. PRP-together

‘Their faces will be together’. The present, which in Bohtan has no prefix except where the verbal stem begins with a vowel, can also be used with future meaning, particularly when there is modal colouring: (7) b-layle in-night

palṭǝx go_out.1PL

yarqǝxna run_away.1PL

‘In the night we will go out, we will run away’ (Fox 2009: 126) In the negative future the particle bǝt/b- is deleted: (8) le neg

ote come.3SG.M

aryona. rain

‘It will not rain (lit. the rain will not come)’. (Fox 2009: 59) An interesting further development has taken place in Hertevin and some other NENA dialects, where the future prefix is also used for habitual action: (9) daḥ how

bet-ḥalwi-le? FUT-milk.3PL-them

‘How do they milk them?’ (Jastrow 1988: 108)

THE HISTORY OF THE FUTURE

167

In all of the above dialects of NENA, we find a future marker of the form b(ǝd)-, evidently descended from a form similar to the JBA bāʿē. It is significant that the coterritorial dialects of Kurdish possess a future marker (Blau 1975: 78–79). For example, in the Kurdish of Amadiya we have the second position particle dê: (10) ez I.DIR

dê

hêm come.1SG

dê

kî who

PTC.FUT

‘I will come’ (11) ma

INTRJ

PTC.FUT

be-t take-3SG

‘But who will take her?’ This contrasts with a prefix, t- in Amadiya, which marks the present indicative (Blau 1975: 78): (12) min I.OBL

t-vȇt IND-want.2SG

sindoq-ek cashbox-ART

‘I want a cashbox’. (13) tu you.DIR

t-bȇjî IND-say.2SG

‘You say’ On the other hand, in Sorani Kurdish, the future is not formally distinguished from the present (Blau 1980: 52). This is reflected in the fact that three NENA dialects from the area where Sorani dialects are spoken also lack a formal marker of the future. In Jewish Sulemaniyya (Khan 2004: 285) the indicative qăṭil kill.3SG.M form may be used with future tense reference. There is no verbal form in the dialect that specifically expresses the future. In Christian Koy Sanjaq the category k-pātəx (Mutzafi 2004a: 254) is used for both the present and the future: k-pātǝx IND-open.3SG.M ‘he opens, will open’. There is an additional present progressive which is formed with the additional prefix lā: lāk-pātǝx PROG-IND-open.3SG.M ‘he is opening’. In Jewish Koy Sanjaq (Mutzafi 2004b: 111), the imperfective expresses an indicative mood and an imperfective aspect referring to events in the present or future. There is only one form which has a distinct future meaning: the form of the verb h-w-y ‘to be’ k-awe IND-be.3SG.M ‘he will be’ (Mutzafi 2004b: 90). I would suggest that these dialects have lost a distinct future that they earlier possessed as a result of the influence of the coterritorial dialects of Kurdish.

168

SAMUEL ETHAN FOX

Now let us turn to Neo-Mandaic and Ṭuroyo, dialects of Neo-Aramaic which lie outside the realm of NENA. In Neo-Mandaic, the active participle with qa- prefix is used for both the present and the future (Macuch 1965: 433). Here are two examples from Häberl (2009: 243): (14) min from

hǝnā here

tum until

ārbin 40

yum-ā day-AUG

qǝ=maṯin=ḵon IND=come.IMPRF.1SG=2PL

‘Forty days from now, I will bring you’. (15) ǝmal-∅=l=i say.PRF-3SG.M=OBJ=3SG.M

perš tomorrow

q-aṯinā IND-come.IMPRF.1SG

‘He said to him, “I will come tomorrow.”’ The lack of an explicit future in Neo-Mandaic may be principally attributed to the fact, noted above, that there was no future marker in Classical Mandaic. It may also be relevant that colloquial Farsi uses the present tense forms for the future. A periphrastic future with ‘want’ is now only used in writing. However, Arabic was probably the dominant language in the areas where Mandaic was spoken over most of the history of the language, so the influence of Farsi was most likely not a factor. In Ṭuroyo, the present is marked with the prefix ko-/k-, probably cognate with similar present prefixes in NENA and in Mandaic. The future has a distinct prefix gǝd-/gd- before a vowel, g- before a single consonant, and gǝ- before a consonant cluster. Talay (2009: 173) suggests that both the present and future prefixes derive from the Syriac particle kad ‘when’. In the case of the present, the combination kad hu ‘when he’ was contracted to ko-. In the case of the future, kad was followed by the particle d- ‘that’ and was contracted to kad-, and with regressive assimilation of voicing and vowel reduction to gǝd-. However, comparison of Ṭuroyo with the neighboring NENA dialects of Bohtan and Hertevin suggests another explanation. The future marker gǝd- seems to be cognate with the preverbal particles ked of Hertevin and qǝt of Bohtan which mean ‘can, is able’. These particles are a reduced form the the Arabic qadir. In Ṭuroyo the meaning of this particle has evolved from ability to a future marker. Development of the meaning of a grammatical marker from possibility to intention to future is also seen in Cantonese, where there are two particles expressing the future that derive from verbs that mean ‘be able to’ (Bybee, Perkins and Pagliuca 1994: 265–266). This path of development is not as frequent as that from volition to future marker, but its occurrence in Cantonese demonstrates that it is a possibility. Explicit future tense forms exist in many, though not all, languages. Where they exist, they usually can be shown to have developed from a fairly restricted set of

THE HISTORY OF THE FUTURE

169

earlier forms. The major sources of future markers that Bybee, Perkins and Pagliuca (1994: 252–253) identify in their broad cross-linguistic study are: A.Movement Verbs a.Come

B.Desire

b.Go

C.Ability

D.Other Verbal Sources a.be, become

b.Other Verbs (try to, need, owe, get, have, do, make, look for)

E.Temporal Adverbs

Three languages (Danish, Inuit, and Tok Pisin) in their sample of ninty-four have futures that are clearly derived from forms with the meaning of ‘desire’. In addition, four other languages in their sample (Buli, Nimboran, Bongu, and Dakota) had future tense markers that also could be used to express ‘desire’, but in these cases there is no evidence to show which meaning came first. Of course, one could easily find other languages, not included in their sample, which exhibit such a development, English and Greek being among them. One of the hypotheses of the study is that there are certain common paths of development in grammaticization that are followed independently by many languages. Specifically, they suggest that forms indicating desire such as Middle Aramaic bāʿē move to future meaning along this path: Desire > Willingness > Intention > Prediction11 The development of the NENA future follows this very frequent path. In conclusion:

 The development of the future in NENA follows a recognized path which can be found in many other languages.  The use of the present for the future and the neutralization of the future in the negative are also common phenomena in other languages.  The future prefix of Ṭuroyo is a development of the prefix expressing ability which is found in Hertevin and Bohtan.

11

Bybee, Perkins, and Pagliuca (1994: 256)

170

SAMUEL ETHAN FOX  The future prefix is lacking in several dialects of NENA due to influence of the coterritorial Kurdish.

REFERENCES

Blau, Joshua. 1975. Le Kurde de cAmādiya et de Djabal Sindjār. Paris: Klincksieck. Blau, Joyce. 1980. Manuel de Kurde, Dialect Sorani. Paris: Klincksieck. Bybee, Joan, Perkins Revere, and Pagliuca, William. 1994. The Evolution of Grammar. Chicago: University of Chicago Press. Fleischman, Suzanne. 1982. The Future in Thought and Language. Cambridge: Cambridge University Press. Fox, Samuel Ethan. 1997. The Neo-Aramaic Dialect of Jilu. Wiesbaden: Harrassowitz. —. 2009. The Neo-Aramaic Dialect of Bohtan. Piscataway, NJ: Gorgias Press. Garbell, Irene. 1965. The Jewish Neo-Aramaic Dialect of Persian Azerbaijan. The Hague: Mouton. Häberl, Charles. 2009. The Neo-Mandaic Dialect of Khorramshahr. Wiesbaden: Harrassowitz. Hetzron, Robert. 1969. ‘The Morphology of the Verb in Modern Syriac (Christian Colloquial of Urmi).” Journal of the American Oriental Society 89: 112–27. Jastrow, Otto. 1988. Der neuaramäische Dialekt von Hertevin (Provinz Siirt). Wiesbaden: Harrassowitz. Khan, Geoffrey. 2004. The Jewish Neo-Aramaic Dialect of Sulemaniyya and Ḥalabja. Leiden: Brill. Krotkoff, Georg. A Neo-Aramaic Dialect of Kurdistan: Texts, Grammar, and Vocabulary. New Haven: American Oriental Society. Lakoff, George, and Johnson, Mark. 1980. Metaphors We Live By. Chicago: University of Chicago Press. Macuch, Rudolf. 1965. Handbook of Classical and Modern Mandaic. Berlin: De Gruyter. Mutzafi, Hezy. 2004a. “Features of the Verbal System in the Christian Neo-Aramaic Dialect of Koy Sanjaq and Their Areal Parallels.” Journal of the American Oriental Society 124: 249–262. —. 2004b. The Jewish Neo-Aramaic Dialect of Koy Sanjaq (Iraqi Kurdistan). Wiesbaden: Harrassowitz. —. 2008. The Jewish Neo-Aramaic Dialect of Betanure (Province of Dihok). Weisbaden: Harrassowitz. Nöldeke, Theodor. 1904. Compendious Syriac Grammar. London: Williams and Norgate. Sampson, Anthony, and Sampson, Sally. 1988. The Oxford Book of Ages. Oxford: Oxford University Press.

THE HISTORY OF THE FUTURE

171

Sinha, Jasmin. 2000. Der neuostaramäische Dialekt von Bēṣpən (Provinz Mardin, Südosttürkei). Eine grammatische Darstellung. Wiesbaden: Harrassowitz. Talay, Shabo. 2009. Die neuaramäischen Dialekte der Khabur-Assyrer in Nordostsyrien. Semitica Viva 40. Wiesbaden: Harrassowitz.

THE ORIGIN OF THE PERIPHRASTIC PRETERITE

KƏM/QAM-QĀṬƏLLE IN NORTH-EASTERN NEO-

ARAMAIC

STEVEN E. FASSBERG 1. INTRODUCTION

The reconstruction of Proto-Neo-Aramaic in general and Proto-NENA (North-Eastern Neo-Aramaic) in particular is challenging because the transition period from Late Aramaic to Neo-Aramaic is virtually undocumented. The oldest substantial written evidence of Neo-Aramaic, both Jewish and Christian, is dated to the 17th century, and the serious academic study of the oral language began only in the 19 th century. Nonetheless, several pioneering works of reconstruction have been attempted, and noteworthy treatments, to name just a few, include those of R. Hoberman on pharyngealization (1985) and pronouns (1988; 1989), G. Khan (2001 and 2011) on the copula, S. E. Fox (2008) on the relationship of NENA in general to Late Aramaic dialects, and H. Mutzafi on etymologies.1 One specific feature of reconstruction that has long aroused interest is the periphrastic Preterite kəm/qam-qāṭə́lle, i.e., the preverbal particle kəm/qam followed by the Subjunctive base with L-suffixes. It occurs primarily with transitive verbs that have suffixed direct objects (Hoberman 1989: 52–55).2 Several variants of qam are attested: qəm/kam/kəm/ḳam/ḳəm/tem-. Unlike the frequent NENA prefix k- found on the Subjunctive base and its variants (g- before voiced consonants, q before an adja-

E.g. the glossaries in Mutzafi (2004 and 2008b) as well as the discussions in Mutzafi (2007 and 2008a). 2 For occurrences without objects, see the references in Pennacchietti (1997: 479, ft. 19). 1

THE ORIGIN OF THE PERIPHRASTIC PRETERITE KƏM/QAM-QĀṬƏLLE

173

cent q, č-, and rarely x and h),3 and which is derived by almost all from the proclitic )‫ ק(א‬qā and freestanding ‫ קא‬qā/‫ קאי‬qāʾe attested in Jewish Babylonian Aramaic, and the proclitic ‫קא‬/‫ קי‬in Classical Mandaic (Brockelmann 1921: 1913),4 kəm/qam- and its variants have been explained in different ways. No scholarly consensus has yet been reached. I would like to review the etymologies that have been offered since Th. Nöldeke (1868) first discussed the form. The most comprehensive and recent discussions appear in F. A. Pennacchietti (1994 and 1997) and A. D. Rubin (2005). In my remarks I shall relate to the extent of the phenomenon as well as the merits and weaknesses of each suggestion. I shall also bring additional data to bear on the discussion. The construction kəm/qam-qāṭəlle is a salient feature of many NENA dialects. It is attested in most but not all Jewish lishana deni dialects (it is neither in Jewish Nerwa and Amedia texts from the 17 th century nor in modern Jewish Challa)5 as well as Christian dialects from the plain of Mosul (the so-called Felliḥi or Alqosh dialects) up into the mountainous regions of south-eastern Turkey as far east as the Iranian border.6 It is not found in the periphery of NENA, which leads one to conclude that the phenomenon must have originated on and around the plain of Mosul: it did not spread to the north-west (e.g. Heretvin and the Bohtan cluster) and western (Ṭuroyo cluster) NENA areas, nor to Trans-Zab Iraq (with the exception of Senaya) and western Iran. As for north-western Iran, kəm/qam-qāṭə́lle is found in the literary Urmi Aramaic of the 19th and early 20th centuries described by D. T. Stoddard (1855: 104– 105), Nöldeke (1868: 296–297), A. J. Maclean (1895: 82, 141), H. J. Polotsky (1961: 21), Q. I. Marogulov (1976: 60–61), K. Tsereteli (1978: 93), 7 and H. L. Murre-van den Berg (1999: 210–211); O. Kapeliuk (2006: 386) adds that it is rare in the translations of Russian works in the ‘New Alphabet’. It is absent, however, in the descriptions of modern Jewish Urmi by I. Garbell (1965) and Khan (2008b), and of modern Christian Urmi as described by R. Hetzron (1969).

Fox (1994: 158); Jastrow (1997: 365); Heinrichs (2002: 244–247). A shift of initial q > k in non-verbs is attested in Hertevin as described by Jastrow (1988: 7). See also Contini (1996: 157–158). 5 Sabar (1976); Fassberg (2010). 6 For early works mentioning the syntagm, see Stoddard (1855: 104); Nöldeke (1868: 296– 297); Sachau (1895: 46–47); Maclean (1895: 82, 141 and 1901: 270) and Rhetoré (1912: 225). 7 His grammar also relates to the Christian vernacular. 3 4

174

STEVEN E. FASSBERG

2. SUGGESTED ETYMOLOGIES

Nöldeke (1868: 296) seems to have been the first to derive the preverbal particle from the general Aramaic preposition qŏḏām, which he translated, however, as an adverb ‘früher, vorher’. He has been followed by several scholars, including J. Rhétoré (1912: 225, ft. 1), R. Macuch and E. Panoussi (1974: 117), D. Cohen (1984: 520), and S. Talay (2001: 18).8 A second proposed etymology derives kəm/qam- and its congeners from a verbal form, either the Classical Aramaic G perfect qḏam or the D qaddem. This verbal derivation was first suggested by Maclean (1895: 82; 1901: 270), who opted for the D verb, and has been adopted by many, among them Rubin (2005: 34) and Khan (2008a: 177). Tsereteli compared qam with the Arabic particle qad, which is commonly derived from the D verb *qad(d)ama (Brockelmann 1913: 507; Pennacchietti 1997: 477). The most recently proposed etymology has been suggested by Pennacchietti (1994; 1997). In a general discussion of the expression of the preterite with objects, which was followed by a shorter treatment that focused on the etymology of kəm/qam-, Pennacchietti derived the form from the Neo-Aramaic Subjunctive of the Aramaic verb q-w-m ‘stand up, arise’, which, in Semitic, generally marks ingression when functioning in hendiadys.9 Pennacchietti believed that kəm/qam- was a grammaticalized, phonetically reduced form of qāʾim. The impetus for his proposal comes from the frequent grammaticalization of verbs of motions (e.g. Arabic raḥ- < rāyiḥ before imperfects to mark futurity, or qa(d) < qāʿid before imperfects to mark the present) and the use in NENA of the Subjunctive base qāʾim in hendiadys with another Subjunctive in certain dialects. He pointed out that qāʾim as an auxiliary verb preceding a Subjunctive expressed an imminent future in the dialects of Upper Tiari and Ashita (Maclean 1895: 82), but in the dialects of the plain of Mosul it related an action in the past that had happened immediately following a Preterite. His data were drawn from the Grammaire de la langue soureth by Rhétoré (1912: 225–226), in which qāʾim was presented as “une forme énergique” and translated “aussitôt”. Pennacchietti (1997: XLI) explained this as “a new periphrastic construction in some old NENA dialects”; he believed that the construction was employed as a stylistic device aimed at “enlivening the narrative”, which in time replaced the Preterite. He cited Though in his grammar of the Khabur dialects, Talay (2008: 309, ft. 361) appears to have changed his mind. 9 For examples in Western Neo-Aramaic and Neo-Mandaic, see Correll (1978: 79–80) and Häberl (2009: 234). For examples in some other Aramaic dialects and in Semitic in general, see Dobbs-Allsopp (1995) and Contini (1996: 154–156). 8

THE ORIGIN OF THE PERIPHRASTIC PRETERITE KƏM/QAM-QĀṬƏLLE

175

parallels from Catalan, medieval French, and Guardia Piemontese, in which a periphrastic tense made up of a present develops into a past. According to Pennacchietti, the clincher for the grammaticalization of qāʾim is the existence of the construction qəm-ʾāte+ Subjunctive, e.g. (1) kāʾ one

nāšā man

səw-le rābā grow_old.PST-3SG.M very

qəm-ʾātē QƏM-come.3SG.M

wə and

brūn-ēh on-his

dīy-ēh PRON-3SG.M

ṭāʾən-nēh carry.3SG.M-OBJ.3SG.M

‘A man grew very old and his son carried him (on his shoulders)’. His proposal has been embraced by Contini (1997: 153).

3. STRENGTHS AND WEAKNESSES OF PROPOSED ETYMOLOGIES

Each of the three proposed etymologies has its strengths and weaknesses. With regard to the preposition qŏḏam, though the general semantics of the word fit perfectly, the form in earlier Aramaic functioned as a preposition and not as an adverb, as noted by Pennacchietti (1997: 476) and Rubin (2005: 34). See e.g. Biblical Aramaic qŏḏām, Jewish Babylonian Aramaic qamme, Syriac qḏām, and Mandaic qudam. In Neo-Aramaic qam still serves as a preposition. The assimilation of d does not pose a problem in the light of Jewish Babylonian qamme, which is a reflex of a base *qadm-. qŏḏam did function as part of a conjunction ‘before’ in Late Eastern Aramaic dialects, but only with the addition of the relative particle d-, whose extensive presence in many NENA dialects would have led one to expect a trace of this d-, if this indeed is the origin of the preverbal qam. If, however, a related adverb expressing the past is sought, a better choice, and one that is etymologically related to qŏḏām, would be a clipped form of the Neo-Aramaic qamāye ‘formerly, originally, a long time ago’, which may have developed from the older Aramaic ordinal (sg.m.) qaḏmāyā ‘first’, (pl.m.) qaḏmāye, or, as suggested by Khan (2008a: 430), from an adverbial *qaḏmāyāṯ, which is attested in Syriac. The derivation of kəm/qam- from a verb, be it the G stem Perfect qḏam or the D stem qaddem, has much to offer since there is widespread hendiadys use of the verb in Late Eastern Aramaic dialects. Rubin (2005: 34, ft. 95) stressed the extent of the phenomenon in the Peshitta. It should be added that there is also significant evidence for the hendiadys construction in Jewish Babylonian Aramaic (Sokoloff 2002: 984b–985a in the G and C stems) and Mandaic (Drower and Macuch 1963: 399, 405 in the G and C stems). Rubin finds additional support in the Senaya form tem

176

STEVEN E. FASSBERG

(Panoussi 1990: 118, 129 and Heinrichs 2002: 243), which he takes to be a reflex of *dVm < qdam/qaddem.10 Pennacchietti (1997: 477) objected to a G or D stem Perfect verbal origin of qam since Syriac qaddem in a hendiadys construction does not occur before participles as does qam in NENA. This objection is not valid, however in the case of Jewish Babylonian Aramaic, where one does find participial forms of the verb q-d-m followed by other participles. Another reservation in deriving kəm/qam- from a Perfect verb involves the elision of the consonant d, which is widely preserved throughout NENA either as d, ḏ, or t11 and is retained in forms of the verbal root q-d-m found in different NENA dialects.12 Though processes of grammaticalization frequently result in substantial phonetic reductions, one might have expected the apocope of final m in the NENA form, especially in the light of its frequent elision in Jewish Babylonian Aramaic,13 compare also Arabic qad, if it really is a reflex of a verb qad(d)ama. Pennacchietti’s view of qāʾim as the source of kəm/qam- is phonetically the most attractive of the etymologies offered so far. It is also fetching because of the Aramaic and general Semitic hendiadys use of the verb q-w-m, though the common meaning of the adverbial use of q-w-m, as already noted, is to mark ingression and not to mark the past. Pennacchietti’s examples of qāʾim followed by the Subjunctive in a past context are taken from Rhétoré’s grammar and text samples. One may also add examples of this phenomenon from the grammar of Christian Aradhin by Krotkoff (1982: 56) or the grammar of Neo-Mandaic by Häberl (2009: 234) as well as many examples of this construction in the past and other contexts from the Jewish Neo-Aramaic dialect of Challa, e.g. (2) zəl-le go.PST-3SG.M

mjāməʿ gather.3SG.M

denānəd debtors.REL

gyān-e. self-3SG.M

Could the t of tem possibly have been influenced at an earlier stage by ‫‘ תמה‬there’ (in modern Senaya ṭōma)? Note Samaritan Aramaic ‫‘ > תמן‬then’ (Tal 1980: 955a). 11 Less commonly d > ḏ > ∅ (Nöldeke 1868: 44) and in Jewish Iranian dialects d > ḏ > l (Kapeliuk 1997 and Khan 2009: 8, 23). But as noted above, the assimilation of d is attested in qamme < *qadm already in Jewish Babylonian Aramaic. 12 Cf. Mutzafi (2008a: 424–425), who derives Trans-Zab bqatta ‘morning’ < *bqadamta < *b-qaddamtā. 13 Particularly in forms of the verb q-w-m. See Boyarin (1976) and Morgenstern (2011: 142, 174). 10

THE ORIGIN OF THE PERIPHRASTIC PRETERITE KƏM/QAM-QĀṬƏLLE qemi get_up.PL

177

qaṭli-le kill.3PL-OBJ.3SG.M

‘He went to gather together his debtors. They up (and) kill him’. The existence of the two competing constructions qāʾim + Subjunctive and kəm/qam + Subjunctive in the material gathered by Rhétoré and in Christian Aradhin as described by Krotkoff, in addition to the inflection of qāʾim in all persons, shows, to my mind, that the Subjunctive of qāʾim is far from being grammaticalized and therefore kəm/qam has a different origin.14 With regard to examples of the construction qəm-ʾāte + Subjunctive found in Ms. Sachau 337 (Lidzbarski 1896, I: 80188) and which Pennacchietti considers the linch-pin of his argument, it should be noted that Mutzafi (2007: 353) considers the manuscript to be “a concoction of various Christian NENA dialectal words and forms and is replete with fictitious and hyper-corrected forms, although quite a few other words and forms indeed reflect Tyare.” Rather than assuming, as does Pennacchietti, that the construction qāʾim + Subjunctive has developed into a substitute expression for the past, I think it preferable to interpret the construction as being merely contextually determined, i.e., it is the previous Preterite or the general context that determines the past meaning. Fluctuation of verbal forms in past context are common in conversational discourse in NENA (Khan 2009), and the use of the ingressive qāʾim before a Subjunctive, as opposed to the use of the Subjunctive alone, should be viewed as the mark of a particularly free and colloquial register. Yet another objection to Pennacchietti’s reconstruction of the origin of kəm/qam from qāʾim comes from Rubin (2005: 34), who observes that the latter is the source of the k- present tense particle. I think Ruben’s observation points in the right direction.

4. A NEW PROPOSAL

There is a widespread phenomenon in NENA that may be directly related to the origin of kəm/qam-, namely, the k- prefix found on the Subjunctive base. It is almost universally acknowledged that this prefix in NENA as well as the q- prefix found in Neo-Mandaic, developed directly from the Late Aramaic form qa-/qāʾ/qāʾe (
ky > y, then the existence of the present indicative prefix is even more extensive in Neo-Aramaic than previously thought, and, if one also relates Ṭuroyo ko- to it, then it is everywhere in Neo-Aramaic except in the West (Maʿlula, Baxʿa, Jubbʿadin). Breuer (1997) examined closely the function of qā- in Jewish Babylonian Aramaic, and demonstrated for the first time in a systematic manner that the primary meaning of the prefix qā- before a Participle in Jewish Babylonian Aramaic was to mark continuous action.16 Surprisingly, the expression of continuous action is not the only use of qā- in Jewish Babylonian Aramaic, nor, in fact, is it even the most frequent. Breuer showed that qā also marks a past action previously mentioned that is still relevant, or in other words, qā + Participle may also function as a Perfect.17 This important insight explains why Amoraic Hebrew translates Aramaic qā + Participle with Perfect forms and also why qa- is prefixed in the Yemenite oral tradition of the Babylonian Talmud to Perfect forms, e.g. (3)

‫ ַק ְס ַבר‬qa-svar QA-think.PRF.3SG.M

‘and he thought and still thinks’, and not only Participles. I think this hitherto unidentified use of qa- has direct relevance to the origin of kəm/qam-. Talay (2008: 306, ft. 354) plays with the possibility that the Syriac conjunction kaḏ is the source of k/g- and Ṭuroyo ko-. Contini (1996: 159–160) also entertains the idea but eventually dismisses kaḏ as the origin of the Ṭuroyo future preverbal particle g/k/gə/gd/d/kt/t-. Macuch and Panoussi (1974: 69) hesitantly mention the Syriac particle kay, which emphasizes expressions of doubt, desire or interrogation, as the source of k/g-. Lipiński (2001: 348–349) derives the prefix k/g- from a form of the verb k-w-n. Heinrichs (2002: 246–247) surveys the forms and suggested etymologies, noting also the particle kā mentioned by Barhebraeus. See also Tezel (2003: 35, ft. 52). 16 For bibliography on the subject see Breuer (1997: 73–74) and Heinrichs (2002: 251– 255). 17 Furthermore, Wajsberg (2011) has recently argued that the apparent free interchange between the Participle ‫ ָּפ ֵע ל‬and the Preterite ‫ ְפ ַע ל‬in Jewish Babylonian Aramaic historical accounts (Nöldeke 1875: 375, ft. 1), especially in introductory formulae, is conditioned: ‫ָּפ ֵעל‬ functions as a Perfect and ‫ ְפ ַע ל‬as a Preterite. He notes (Wajsberg 2011: 141) that, according to Breuer, ‫ קא‬is optional before Participles. 15

THE ORIGIN OF THE PERIPHRASTIC PRETERITE KƏM/QAM-QĀṬƏLLE

179

The pervasiveness of the preverbal k/g– (to which y/i- may be related) before a Subjunctive as well as the dominant use in Jewish Babylonian Aramaic of qa + Participle to express a Perfect leads me to propose that the preverbal k/g- + Subjunctive functioning in past contexts as a historical present may be the source of the preverbal kəm/qam- before a Subjunctive. For example, (4) nāše people ́ wīša dry

sē-lu go_down.PST-3PL b-ıd in-REL

dúmūt image

g-bahti. G-be_astonished.3PL

l-ṣḷōṣa. to-prayer

d-īd

PRON-REL

g-mēnxi, G-look.3PL

́ ʿinsān human

xa one

wē-la COP.PST-3SG.F

kḗpa stone l-tāma. to-there

g-ımri… G-say.3PL

‘People came to prayer. They look. A solid stone in human form was there. They are astonished. They say…’ (Jewish Zakho; Meehan and Alon 1978: 184) If Talay is correct and y/i- is a palatalized reflex of k-, then it is striking that in all dialects in which kəm/qam- occurs, one also finds one or more of the allomorphs of k-, though not all dialects that have an allomorph of k-, have kəm/qəm-. If kəm- derives from k-, what is the origin of the m of kəm/qam-? I would like to suggest that kəm/qam- may have been created in Proto-NENA from a metanalysis of the prefix kə- and the prefix m- of D (stem II) and C (stem III) Subjunctive forms of the verb, e.g. (5) kə + mšādər-re > (stem D)

kəm-šādər-re KƏM-send.3SG.M-OBJ.3SG.M (stem D)

‘he sent him’ (and with junctural doubling kəm-mšādərre); (6) kə + maqṭəl-le > (stem C)

kəm-aqtəlle KƏM-kill.3SG.M-OBJ.3SG.M (stem C)

‘he had him killed’ (with junctural doubling kəm-maqṭəlle). The derivation of kəm/qam from kə- and the m- of the D or C stems is strengthened by the fact that the direct objects suffixed to the Subjunctive in the construc-

180

STEVEN E. FASSBERG

tion are to be expected precisely with D and C stems verbs, which are transitive and require a direct object.18 At a later stage the particle kəm may have been transferred to G (stem I) verbs: (7) kəm-qāṭəl-le KƏM-kill.3SG.M-OBJ.3SG.M (stem G) ‘he killed him’ There are manuscripts like Codex Sachau 223, which was published by Li-

dzbarski (1896, I: 346–386),19 in which the scribe prefixed kəm- to the following sub-

junctive, e.g., before a G stem verb ‫ ܟܡܨܠܘܝܠܐ‬kəmṣalwile ‘they crucified him’. In the

same manuscript only one mem is usually written in the case of D and C stem verbs, e.g.

‫ ܟܡܚܠܨܠܐ‬kəmxāləṣla ‘he saved her’ (Mengozzi, 2002, I: 147, line 99d), ‫ܟܡܩܡܠܐ‬

kəmāqəmle ‘he raised him’ (Mengozzi 2002, II: 69 line 51a), though infrequently one finds a double writing of mem, e.g. 2002, I: 63 line 37a),

‫ ܟܡܡܠܩܠܐ‬kəmmalqāle ‘they met him’ (Mengozzi

‫ ܟܡܡܠܒܫܠܐ‬kəmmalwəšle ‘and he dressed him’ (Mengozzi 2002,

I: 135 line 50c). Fluctuation in the writing of kəm- is by no means unusual in texts from the 19th and early 20th centuries. For example, see also manuscripts containing the poem on the plague of Pioz: ‫ܫܒܩܠܢ‬

‫ ܟܡ‬kem šāwəqlan ‘he abandoned us’ (Poizat 1990: 175 line 116c) vs. ‫( ܟܡܫܩܠܢ‬line 117a); ‫ ܟܡܢܦܩܠܢ‬kəmnāpəqlan ‘he brought us out’ (Codex Sachau 117b) vs. ‫ ܟܡ ܡܦܩ ܠܢ‬printed edition)20; or in the texts presented in The ubiquity of k/g-emər, k/g-əmri as a historical present may have also played a role in the feeling that the sequence kəm- was a preverbal particle. This assumes that there was no shift of k > g/-m in early NENA as is the case in many dialects today. k- is written before mem in Christian Aramaic manuscripts and printed additions. This is not the case in the 17 th century Jewish Nerwa and Amedia texts. kəm/qam + āmərre (ʾ-m-r ‘say’) is relatively rare, the verb ‘say’ usually takes an indirect object. For example, in the hundreds of pages of texts in Barwar published by Khan (2008, III), I have found only two examples of qəm + āmər: mə́re bár qəmamrə́tla ʾanna xabráne ʾə́lli ‘He said, “Since you have said these things to me”’ (III, 1552, line 50) and ʾá-dana qəm-amə̀rra ‘Then he said to them’ (III, 1874, line 3), as opposed to scores and scores of (y)-amər, (y)-amri. 19 See Mengozzi (2002, II: 49–57) on the salient features of the scribe of Codex Sachau 223. On the scribe’s use of kəm-, see 40–41. 20 Poizat (1993: 249). 18

THE ORIGIN OF THE PERIPHRASTIC PRETERITE KƏM/QAM-QĀṬƏLLE

181

Guidi (1883: 304 line 54) qem wağib vs. kemwáğib (313 line 1). Rhétoré (1912: 225, ft. 1) notes that kəm is written qam as a separate particle in Persia and in the moun-

tains.

The piling up of particles and their subsequent metanalysis is a well-known and old phenomenon: see, e.g. Biblical Aramaic: (8) ‫ָּכל‐ ֳק ֵבל‬ k

PRP.like

+l + qŏbel > + PRP.to +oposite

kol-qŏḇel

‘opposite’ The metanalysis of accumulated preverbal particles in NENA can be clearly seen in the Jewish Nerwa and Amedia texts from the 17th century, where the orthography of the negative la preceding the present tense prefix g- is written by scribes together with g- as a separate particle lag: (9) ‫גל נאציה ולג קאטל‬ lag- nāṣe NEG.PRS.PTCP-fight.3SG.M

w-lag-qāṭel CONJ-NEG.PRS.PTCP-kill.3SG.M

‘He does not fight and he does not kill’. (Sabar 1976: 213, line 9), or the common particle mā + d- > mād INTER+ REL> PRON ‘whatever’ attested in many dialects today. If my proposed reconstruction is correct, then the velar k and the shewa vowel of kəm- might be the early NENA form, whereas the uvular q and a-vowel of qam possibly arose later in NENA by a semantic association with the preposition qam ‘before’ and the adverb qamāye ‘formerly’. Another reconstruction, however, is also conceivable: that qam with uvular q is a direct reflex of the Late Aramaic particle qāand that q> k. Theoretically, one could even posit a split in Proto-NENA in which the uvular reflex was exploited for the expression of the perfect, like the use of qādescribed by Breuer for Jewish Babylonian Aramaic, whereas a derived velar k- was used by speakers to express continuous action and the present. In short, I believe that scholarly reservations about the previously suggested etymologies of the preverbal particle kəm/qam in the construction kəm/qam-qāṭə́lle are

182

STEVEN E. FASSBERG

not unwarranted. I suggest relating the particle to the widespread k-, which clearly derives from the preverbal qā or qāʾe in Late Aramaic and which not only marks continuity but also functions already as a perfect. kəm/qam-qāṭə́lle would be appear to be a relatively recent innovative feature that originated in the plain of Mosul. The absence of the construction from the 17th century Jewish Nerwa and Amedia texts (vs. its frequency in most modern lishana deni dialects) and its rarity in the Christian Neo-Aramaic poems from northern Iraq in the 17th and 18th centuries (vs. its regularity in poems from the 19th and 20th centuries21) point to the diffusion of the phenomenon sometime during the 18th and 19th centuries.

REFERENCES

Avineri, Iddo. 1988. The Aramaic Dialect of the Jews of Zākhō. Jerusalem: Israel Academy of Sciences and Humanities. [in Hebrew] Boyarin, Daniel. 1976. “The Loss of Final Consonants in Babylonian Jewish Aramaic.” Afroasiatic Linguistics 3: 103–107. Brockelmann, Carl. 1913. Grundriss der vergleichenden Grammatik der semitischen Sprachen, II: Syntax. Berlin: Reuther & Reichard. Breuer, Yochanan. 1997. “The Function of the Particle Qā- in the Aramaic of the Babylonian Talmud.” Leshonenu 60: 73–94. [in Hebrew] Contini, Roberto. 1997. “Alcuni casi di grammaticalizzazione (e degrammaticalizzazione) in ṭūrōyō.” In Afroasiatica Neapolitana. Contributi presentati all’80 Incontro di Linguistica Afroasiatica (Camito- Semitica) Napoli, 25–26 Gennaio 1996, edited by Alessandro Bausi and Mauro Tosco, 151–167. Napoli: Istituto Universitario Orientale. Correll, Christoph. 1978. Untersuchungen zur Syntax der neuwestaramäischen Dialekte des Antilibanon (Maʿlūla, Baḫʿa, Ğubb ʿAdīn) mit besonderer Berücksichtigung der auswirkungen Arabischen Adstrateinflussess. Abhandlungen für die Kunde des Morgenlandes 44.4. Wiesbaden: Deutsche Morgenländische Gesellschaft. Dobbs-Allsopp, F. W. 1995. “Ingressive qwm in Biblical Hebrew.” Zeitschrift für Althebraistik 8: 31–53. Drower, Ethel S., and Macuch, Rudolf. 1963. A Mandaic Dictionary. Oxford: Clarendon. Fassberg, Steven E. 2010. The Jewish Neo-Aramaic Dialect of Challa. Studies in Semitic Languages and Linguistics 54. Leiden: Brill. Fox, Samuel Ethan. 1994. “The Relationships of the Eastern Neo-Aramaic Dialects.” Journal of the American Oriental Society 114: 154–162. 21

See Mengozzi (2012).

THE ORIGIN OF THE PERIPHRASTIC PRETERITE KƏM/QAM-QĀṬƏLLE

183

—. 2008. “North-Eastern Neo-Aramaic and the Middle Aramaic Dialects.” In NeoAramaic Dialect Studies: Proceedings of a Workshop on Neo-Aramaic held in Cambridge 2005, edited by Geoffrey Khan, 1–17. Piscataway, NJ: Gorgias Press. Garbell, Irene. 1965. The Jewish Neo-Aramaic Dialect of Persian Azerbaijan. The Hague: Mouton. Guidi, Ignazio. 1883. “Beiträge zur Kenntnis des neuaramäischen Fellihi-Dialektes.” Zeitschrift der Deutschen Morgenländischen Gesellschaft 37: 293–318. Heinrichs, Wolfhart. 2002. “Pecularities of the Verbal System of Senāya within the Framework of North Eastern Neo-Aramaic (NENA).” In Sprich doch mit deinen Knechten aramäische, wir verstehen es! 60 Beiträge zur Semitistik: Festschrift für Otto Jastrow zum 60.Geburtstag, edited by Werner Arnold and Hartmut Bobzin, 237–268. Wiesbaden: Harrassowitz. Hetzron, Robert. 1969. “The Morphology of the Verb in Modern Syriac (Christian Colloquial of Urmi).” Journal of the American Oriental Society 89: 112–127. Hoberman, Robert D. 1985. “The Phonology of Pharyngeals and Pharyngealization in Pre-Modern Aramaic.” Journal of the American Oriental Society 105: 221–231. —. 1988. “The History of the Modern Aramaic Pronouns and Pronominal Suffixes.” Journal of the American Oriental Society 108: 557–576. —. 1989. The Syntax and Semantics of Verb Morphology in Modern Aramaic: A Jewish Dialect of Iraqi Kurdistan. American Oriental Society 69. New Haven, Connecticut: American Oriental Society. Jastrow, Otto. 1988. Der neuaramäische Dialekt von Hertevin (Provinz Siirt). Semitica Viva 3. Wiesbaden: Harrassowitz. —. 1997. “The Neo-Aramaic Languages.” In The Semitic Languages, edited by Robert Hetzron, 324–377. London: Routledge. Kapeliuk, Olga. 1997. “Spirantization of ṯ and ḏ in Neo-Aramaic.” Massorot 9–11: 527–544. [in Hebrew] —. 2006. “The Neo-Aramaic Tense System in the Light of Translations from Russian.” In Loquentes Linguis: Studi Lingusitici e Orientali in Onore Fabrizio A. Pennacchietti, edited by Gorgio Borbone, Alessandro Mengozzi and Mauro Tosco, 381–390. Wiesbaden: Harrassowitz. Khan, Geoffrey. 2001. “Quelques aspects de l’expression d’être” en néo-araméen.” In Langues de Diaspora: Langues in Contact, edited by Anaid Donabéian, 139–148. Faits de Langue Revue de Linguistique 18. Paris. —. 2005. “Remarks on the Function of the Preterite and the Perfect in North-Eastern Neo-Aramaic.” In Neo-Aramaic Dialect Studies: Proceedings of a Workshop on NeoAramaic held in Cambridge 2005, edited by Geoffrey Khan, 105–130. Gorgias Neo-Aramaic Studies 1. Piscataway, NJ: Gorgias Press. —. 2007. “The North-Eastern Neo-Aramaic Dialects.” Journal of Semitic Studies 52: 1–20.

184

STEVEN E. FASSBERG

—. 2008a. The Neo-Aramaic Dialect of Barwar. 3 vols. Handbook of Oriental Studies 96. Leiden: Brill. —. 2008b. The Jewish Neo-Aramaic Dialect of Urmi. Gorgias Neo-Aramaic Studies 2. Piscataway, NJ: Gorgias Press. —. 2009. The Jewish Neo-Aramaic Dialect of Sanandaj. Gorgias Neo-Aramaic Studies 10. Piscataway, NJ: Gorgias Press. —. 2011. “Remarks on Constructions with the Copula in North-Eastern Neo-Aramaic Dialects.” Aula Orientalis 29: 105–119. Krotkoff, Georg A. 1982. Neo-Aramaic Dialect of Kurdistan: Texts, Grammar, and Vocabulary. American Oriental Series 64. New Haven, Connecticut: American Oriental Society. Lidzbarski, Mark. 1896. Die neu-aramäischen Handschriften der Kgl. Bibliothek zu Berlin. 2 vols. Wiemar: Emil Felber. Lipiński, Edward. 2001. Semitic Languages: Outline of a Comparative Grammar (2nd ed.). Orientalia Lovaniensia Analecta 80. Leuven: Peeters. Maclean, Arthur John. 1895. Grammar of the Dialects of Vernacular Syriac. Cambridge: Cambridge University Press. Macuch, Rudolf, and Panoussi, Estiphan. 1974. Neusyrische Chrestomathie. Wiesbaden: Harrassowitz. Marogulov, Q. I. 1976. Grammaire néo-syriaque pour écoles d’adultes (dialecte d’Urmia), translated by O. Kapeliuk (from the Russian edition, Moscow 1935). Paris: Geuthner. Meehan, Charles, and Alon, Jacqueline. 1979. “The Boy Whose Tunic Stuck to Him: A Folktale in the Jewish Neo-Aramaic Dialect of Zakho (Iraqi Kurdistan).” Israel Oriental Studies 9: 174–203. Mengozzi, Alessandro. 2002. Israel of Alqosh and Joseph of Telkepe: A Story in a Truthful Language, Religious Poems in Vernacular Syriac (North Iraq, 17th Century), edited with introduction and translation. Corpus Scriptorum Christianorum Orientalium 590–591 (Scriptores Syri 230–231). Louvain: Peeters. —. 2012. “The Contribution of Early Christian Vernacular Poetry from Northern Iraq to Neo-Aramaic Dialectology: Preliminary Remarks on the Verbal System.” ARAM 24: 25–40. Morgenstern, Matthew. 2011. Studies in Jewish Babylonian Aramaic Based upon Early Eastern Manuscripts. Harvard Semitic Studies 62. Winona Lake, Indiana: Eisenbrauns. Murre-van den Berg, Hendrika Lena. 1999. From a Spoken to a Written Language: The Introduction and Development. of Literary Urmia Aramaic in the Nineteenth Century. Leiden: Nederlands Instituut voor het Nabije Oosten. Mutzafi, Hezy. 2004. The Jewish Neo-Aramaic Dialect of Koy Sanjaq (Iraqi Kurdistan). Semitica Viva 32. Harrassowitz: Wiesbaden.

THE ORIGIN OF THE PERIPHRASTIC PRETERITE KƏM/QAM-QĀṬƏLLE

185

—. 2007. “The Sound Change *t > š in Ṭyare Neo-Aramaic.” Le Muséon 120: 351– 364. —. 2008a. “Trans-Zab Jewish Neo-Aramaic.” Bulletin of the School of Oriental and African Studies 71: 409–431. —. 2008b. The Jewish Neo-Aramaic Dialect of Betanure (Province of Dihok). Semitica Viva 43. Harrassowitz: Wiesbaden. —. 2014. Comparative Lexical Studies in Neo-Mandaic. Studies in Semitic Languages and Linguistics 73. Leiden: Brill. Nöldeke, Theodor. 1868. Grammatik der neusyrischen Sprache. Leipzig: Weigel. —. 1875. Mandäische Grammatik. Halle: Buchhandlung des Waisenhauses. Panoussi, Estiphan. 1990. “On the Senaya Dialect.” In Studies in Neo-Aramaic, edited by Wolfhart Heinrichs, 107–129. Harvard Semitic Studies 36. Atlanta, Georgia: Scholars Press. Pennacchietti, Fabrizzio A. 1994. “Il preterito neoaramaico con pronome oggetto.” Zeitschrift der Deutschen Morgenländischen Gesellschaft 144: 259–283. —. 1997. “On the Etymology of the Neo-Aramaic Particle qam/kim-.” Massorot 9–11: 475–482. [in Hebrew] Poizat, Bruno. 1990. “La Complainte sur la peste de Pioz.” In Studies in Neo-Aramaic, edited by Wolfhart Heinrichs, 161–179. Harvard Semitic Studies 36. Atlanta, Georgia: Scholars Press. —. 1993. “La peste de Pioz. Suite et fin.” In Semitica: Serta philological Constantino Tsereteli dicata, edited by Roberto Contini, Fabrizzio A. Pennacchietti and Mauro Tosco, 227–272. Torino: Silvio Zamorani. Polotsky, Hans Jakob. 1961. “Studies in Modern Syriac.” Journal of Semitic Studies 6: 1–32. —. 1979. “Verbs with Two Objects in Modern Syriac (Urmi).” Israel Oriental Studies 9: 204–227. Rhétoré, Jacques. 1912. Grammaire de langue soureth ou chaldéen vulgaire selon le dialecte de la plaine de Mossoul et des pays adjacents. Mossoul: Imprimerie des Pères Dominicains. Ruben, Aaron D. 2005. Studies in Semitic Grammaticalization. Harvard Semitic Studies 57. Winona Lake, Indiana: Eisenbrauns. Sabar, Yona. 1984. Homilies in the Neo-Aramaic of the Kurdistani Jews on the Parashot Wayḥi, Beshallaḥ and Yitro: Edition, Hebrew Translation and Introduction. Jerusalem: Israel Academy of Sciences and Humanities. [in Hebrew] Sachau, Eduard. 1895. Skizze des Fellichi-Dialekts von Mosul. Berlin: Verlag der Königl. Akademie der Wissenschaften. Sara, Solomon I. 1974. A Description of Modern Chaldean. Mouton: The Hague/Paris. Sokoloff, Michael. 2002. A Dictionary of Jewish Babylonian Aramaic. Ramat-Gan: Bar Ilan University Press/Baltimore, Maryland: Johns Hopkins University Press.

186

STEVEN E. FASSBERG

Stoddard, David Tappan. 1855. A Grammar of the Modern Syriac Language as Spoken in the Oroomiah, Persia, and in Koordistan. London: Trubner&Co. Tal, Abraham. 2000. A Dictionary of Samaritan Aramaic. 2 vols. Handbuch der Orientalistik 50. Leiden: Brill. Talay, Shabo. 2001. “Grammatikalische Anmerkungen und Texte zum neuaramäischen Dialekt von Nerwa (Nordirak).” Mediterranean Language Review 13: 1–37. —. 2008. Die neuaramäischen Dialekte der Khabur-Assyrer in Nordostsyrien. Semitica Viva 40. Wiesbaden: Harrassowitz. Tezel, Aziz. 2003. Comparative Etymological Studies in the Western Neo-Syriac (Ṭūrōyo) Lexicon: With Special Reference to Homonyms, Related Words and Borrowings with Cultural Signification. Studia Semitica Upsaliensia 18. Uppsala: Uppsala Universitet. Tsereteli, Konstantin G. 1978. Der modernen assyrischen Sprache (Neuostaramäisch), translated by P. Nagel. Leipzig: VEB Verlag Enzykopädie Leipzig (translated from the Russian edition, Moscow, 1964). Wajsberg, Ejakim. 2011. “The Babylonian Aramaic Tense System: ‫ ָּפ ֵעל‬and ‫ ְּפ ַעל‬in Historical Accounts.” Leshonenu 73: 139–166. [in Hebrew]

THE PRESENT CONTINUOUS IN THE NEO-ARAMAIC

DIALECT OF ʿANKAWA AND ITS AREAL AND TYPOLOGICAL

PARALLELS

1

ROBERTA BORGHERO 1. INTRODUCTION

The dialect of ʿAnkawa belongs to the North-Eastern Neo-Aramaic (NENA) subgroup. The town of ʿAnkawa, located in the province of Arbil, in north-eastern Iraq, was and still is inhabited by Catholic Christians (Chaldeans). In comparison with other Chaldean villages of Iraq, ʿAnkawa has a relatively large population in the present day, as the number of people still living in the town is in the order of thousands. During the past decades, however, many of the original inhabitants of ʿAnkawa have left their homeland and moved to various locations around the world, mainly Sweden, Australia, Canada and the United States. This paper is the result of the fieldwork that I have conducted among speakers of the dialect of ʿAnkawa recently emigrated to Sweden, the United Kingdom, the United States and Australia. It will focus on the present continuous construction in ʿAnkawa. This will be compared with present continuous forms documented in other NENA dialects and in Kurdish (Kurmanji) dialects spoken in the area of ʿAnkawa. Comparisons with constructions attested in other Semitic and non-Semitic languages will also be made in order to show typological parallels between ʿAnkawa and other languages of the world. In the transcription system used here for ʿAnkawa and other NENA dialects, stress is penultimate unless indicated otherwise. The vowels are generally short in 1

The idea of developing this subject grew out of my reading of Pennacchietti (2007).

188

ROBERTA BORGHERO

closed syllables and open unstressed final ones, and long otherwise. They are indicated by a macron or a breve only where the occurrence of a long or a short vowel goes against the general tendencies of length distribution. In historical forms, however, long vowels are always indicated by a macron.

2. THE PRESENT CONTINUOUS CONSTRUCTION IN ʿANKAWA: UNINFLECTED PARTICLE DƏ- + PRESENT INDICATIVE

The present continuous is generally expressed in ʿAnkawa by means of the uninflected preverbal particle də- followed by the present indicative of the verb, e.g. də-kšaqəl ‘he is taking’. The full paradigm is as follows:

3

ms. fs. pl.

də-kšaqəl də-kšaqla də-kšaqli

‘he is taking’ ‘she is taking’ ‘they are taking’

2

ms. fs. pl.

də-kšaqlət də-kšaqlat də-kšaqlitun

‘you (ms.) are taking’ ‘you (fs.) are taking’ ‘you (pl.) are taking’

1

ms. fs. pl.

də-kšaqlən də-kšaqlan də-kšaqlux

‘I (m.) am taking’ ‘I (f.) am taking’ ‘we are taking’

The use of an uninflected preverbal particle followed by a present base verb to express the present continuous is documented in other NENA dialects spoken in the area of Arbil.2 Examples of this kind are C. Koy Sanjaq lā-kpaθəx, C. Bedyal ma-kpaθəx, J. Arbel lā-palə́x ‘he is opening’.3 The same construction is also found outside the province of Arbil, in the contiguous region of ʿAqra (in the Mosul province), e.g. C. ʿAqra, Shōsh-u-Sharmən and Nargəzine-Xarjawa nə-kqaṭəl, Khirpa ho-kqaṭəl ‘he is killing’.4 The preverbal particles used in the NENA dialects of the area appear to be derived from the present copula, whose original form (*ʾila, *hole, etc.) varies in the different dialects.5 The hypothesis of a copular origin of the preverbal particle finds 2

See Mutzafi (2004: 260–262). Mutzafi (2004: 260); Khan (1999: 111). 4 Mutzafi (2004: 261); Coghill (2008: 103). 5 For the copular origin of the preverbal particle in the mentioned dialects see Khan (1999: 111–114); Mutzafi (2004: 260–261); Coghill (2008: 103–104). 3

THE PRESENT CONTINUOUS IN THE NEO-ARAMAIC DIALECT OF ʿANKAWA

189

support in the comparison with the present continuous construction attested in Qaraqosh and Karamlesh (NW Iraq), e.g. kilə kšaqəl ‘he is taking’, which consists of the deictic present copula (kilə) followed by the present indicative of the verb (kšaqəl).6 Although the evidence from other NENA dialects of the area seems to suggest a possible copular derivation for the preverbal particle, in the case of ʿAnkawa it is difficult to identify the original copula from which the particle də- would have derived.7 Alternatively, one could suggest that the particle də- in the ʿAnkawa present continuous construction də-kšaqəl is related to the deontic particle that is used with the imperative in ʿAnkawa and other NENA dialects to give immediacy to orders, e.g. ʿAnkawa də-mšari ‘begin!’, J. Arbel da-holli ‘give me, please!’,8 Qaraqosh də-rxoš qalla ‘walk quickly!’,9 Ashitha (SE Turkey) de-pluṭ m-axxa ‘go out of here!’,10 Barwar (NW Iraq) də-murri ‘tell me!’.11 In NENA, broadly speaking, the particle is found not only with the imperative, but also with other verbal forms and is generally used to attract attention to a salient point of the narration, e.g. Barwar ʾana də-ṱ-amrə́nnox ʾó-mdi taxrə̀nne ‘I shall tell you what I remember’, de-píšla ʾap-ay báxta d-àwwa dewə̀rrəš ‘Now, she became the wife of that vagabond’.12 Geoffrey Khan (2008) hypothesizes an etymological relation of this ‘particle of immediacy’ with the word for ‘now’ (diya in Barwar) and believes that this is the same particle used in the ʿAnkawa present continuous construction.13 Constructions containing an analogous particle are attested also in Kurdish and in some Arabic dialects. Forms consisting of a deontic particle de or da followed by the imperative are documented in Kurdish (both Kurmanji and Sorani), e.g. Kurmanji de were ‘come on!’,14 and in some Arabic dialects of Mesopotamia, in particular in 6

Khan (2002: 331 ff.); Borghero (2008: 82). It might have resulted from the contraction of an original copula of the type *dule, which is however not attested in the area, or, less plausibly, from the contraction of the deictic copula ʾāylə, attested in ʿAnkawa. One could reconstruct a shift *ʾāylə > *ʾāymə> ʾāynə, parallel to the shift *hole>*home> hone hypothesized by Coghill (2008: 103) for the particle nə- in the ʿAqra dialects. However a further shift *nə > də with the denasalization of /n/ (exceptional in NENA) is not very convincing. 8 Khan (1999: 282–283). 9 Khan (2002: 350). 10 Borghero (2005a: 126). 11 Khan (2008: 742). 12 Khan (2008: 1590). 13 Khan (2008: 576, 742). The word for ‘now’ in ʿAnkawa is dăha. 14 Thackston (2006a: 207). For a parallel form in Sorani Kurdish, see Thackston (2006b: 7

190

ROBERTA BORGHERO

the qəltu dialects of the Mardin area, e.g. Arbel dē dōṛu ʿaláyu ‘look for him!’,15 and in the Muslim, Christian and Jewish dialects of Baghdad, e.g. denām ‘sleep!’.16 Moreover, it should be noted that a particle di-, de-, da- is attested in Kurdish (Kurmanji and Sorani, in the standard form of the language and in the dialects) for various types of verbal forms, including the present indicative and the present continuous, e.g. Kurmanji ez di-girîm ‘I cry’, ez di-birrime ‘I am cutting’,17 and the habitual past and the past continuous, e.g. ez di-girîyam ‘I used to cry’, di-parastîye ‘he was defending’.18 A seemingly interesting parallel with ʿAnkawa can be found also in the Muslim dialect of Baghdad, where the form da + imperfect can be used to express the present continuous, e.g. damši ~ daʾamši ~ dayemši ‘I am walking’.19 However, the particle da in this form appears to be derived from gāʿed, which is attested, albeit less frequently, in the same construction with the same function.20 Besides the possibilities discussed above for the derivation of the preverbal də in the ʿAnkawa construction də-kšaqəl, in my opinion a further hypothesis should be taken into consideration, that is that the particle was in origin the relative d (~ də), attested throughout NENA. This will be discussed in the following paragraph.

3. DƏ-KŠAQƏL ‘HE IS TAKING’: A PSEUDO-RELATIVE CONSTRUCTION?

Let us consider the hypothesis that the preverbal particle də in the construction də-kšaqəl is derived from the relative d (~ də).

39).

15

Jastrow (1978: 310). Blanc (1964: 117). 17 Blau and Barak (2003: 48, 51). For other examples in Kurmanji Kurdish, see Thackston (2006a: 33–35), MacKenzie (1961: 89). For examples in Sorani Kurdish, see Thackston (2006b: 26–29), Blau (2011: 51–54). 18 Blau-Barak (2003: 71, 73). For other examples in Kurmanji Kurdish, see Thackston (2006a: 47–48), MacKenzie (1961: 96). For examples in Sorani Kurdish, see Thackston (2006b: 40–45), Blau (2011: 68–69). 19 Blanc (1964: 115–116). 20 Blanc (1964: 116). In Jewish and Christian Baghdad the present continuous can be expressed by means of the particles qad and qa respectively, both derived from *qāʿid ‘sitting’, e.g. J. qadamši, C. qa(ʾ)amši ‘I am walking’. Referring to Van Wagoner (1944: 56–57), Blanc states that a ‘present marker’ da ~ jāʿed exists also in rural lower Iraq. I have not been able to check Van Wagoner (1944) in person. 16

THE PRESENT CONTINUOUS IN THE NEO-ARAMAIC DIALECT OF ʿANKAWA

191

As shown in the table below, besides the present continuous construction də-kšaqəl, there is also a very marginal one, where the form is preceded by the deictic copula (e.g. 3ms. ʾāylə), for instance ʾāylə də-kšaqəl ‘he is taking’. The full paradigm of the construction is given below.

3

ms. fs. pl.

ʾāylə də-kšaqəl ʾāyla də-kšaqla ʾāylu də-kšaqli

‘he is taking’ ‘she is taking’ ‘they are taking’

2

ms. fs. pl.

ʾayyət də-kšaqlət ʾayyat də-kšaqlat ʾáyyitun də-kšaqlitun

‘you (ms.) are taking’ ‘you (fs.) are taking’ ‘you (pl.) are taking’

1

ms. fs. pl.

ʾayyən də-kšaqlən ʾayyan də-kšaqlan ʾayyux də-kšaqlux

‘I (m.) am taking’ ‘I (f.) am taking’ ‘we are taking’

The deictic copula ʾāylə is itself extremely rare in ʿAnkawa, and the enclitic copula -ilə is generally preferred to it. The deictic copula appears to be used only to indicate a contingent situation and not a permanent property of the subject. Being presentative by nature, it expresses more immediacy than the enclitic copula, e.g. enclitic copula b-béθelə ‘he is in the house’, deictic copula ʾāylə b-beθa ‘he is [as we speak] in the house’. In a similar way, the present continuous construction with the deictic copula is used with more immediacy and salience than the form without the copula. In particular, the construction with the deictic copula appears to be used to draw attention to a process that is taking place in the presence of the speaker, e.g. ʾāylə də-kšaqəl ‘he is right here taking’/‘he is taking right now’, while the form də-kšaqəl alone is used with the more general meaning ‘he is taking’. The construction ʾāylə də-kšaqəl could be analysed as follows: (1) ʾāylə COP.PRS.3SG.M he_is

də-kšaqəl REL-take.3SG.M that-takes

Therefore ʾāylə də-kšaqəl could be literarily translated ‘he is that takes’, or, considering the deictic force of the copula ʾāylə, ‘here he is that takes’. A construction typologically parallel to ʾāylə də-kšaqəl ‘[here] he is that takes’ can be found in French and Italian. It is normally documented in locative and existential contexts, after deictics (such as Italian ecco and French voilà ‘here is’) or after verbs of perception. Some examples of the construction are given below.

192

ROBERTA BORGHERO

In locative contexts Italian (2) Mario Mario

ѐ be.3SG

qui here

che REL

aspetta wait.3SG

‘Mario is here waiting’. French (3) elle 3SG.F

est be.3SG

là there

qui

REL.NOM

mange eat.3SG

‘She is there eating’. After deictics Italian (4) ecco here_is

Giovanni Giovanni

che REL

torna come_back.3SG

‘Here is Giovanni coming back’. French (5) voilà Antoine here_is Antoine

qui

REL.NOM

arrive arrive.3SG

‘Here is Antoine arriving’. After verbs of perception Italian (6) vedo see.1SG

un

INDEF.ART.SG.M

uomo man

che

professeur professor

qui

REL

corre run.3SG

‘I see a man running’. French (7) je 1SG

vois see.1SG

le

DEF.ART

‘I see the professor smoking’.

REL.NOM

fume smoke.3SG

THE PRESENT CONTINUOUS IN THE NEO-ARAMAIC DIALECT OF ʿANKAWA

193

This particular type of relative construction has been named in various ways, two examples among several being ‘deictic relative’ and ‘situational relative’.21 The most popular term used to define it is ‘pseudo-relative’, introduced by Radford (1975).22 Pseudo-relative constructions are generally found in contexts like the ones mentioned above and are normally used to indicate or draw attention to a process taking place in the current speech situation, or concomitant with the act of perceiving it. The subject of the pseudo-relative is caught in the act of performing an action, and is somehow defined by the action itself. Going back to the Neo-Aramaic dialect of ʿAnkawa, we could hypothesize an original pseudo-relative construction ʾāylə də-kšaqəl ‘[here] he is that takes’. This construction, used to draw attention to a particular process happening in the presence of the speaker, would have developed into a verbal form expressing concomitance and consequently assumed the function of the present continuous ‘he is taking’. At a later stage, the form də-kšaqəl (lit. ‘that takes’), originally in a construction with the deictic copula ʾāylə, would have been reanalysed as a full predicate, becoming the primary way of expressing the present continuous. As a result of this process, the original relative particle də in də-kšaqəl would have been reinterpreted as a preverbal particle expressing the function of the present continuous. This process of grammaticalization might have been favoured due to the tendency of the relative particle to become a clitic.23 An original relative form də-kšaqəl ‘that takes’ (relative particle + present indicative) used as a full predicate to express the present continuous ‘he is taking’ has a very interesting parallel in a construction attested in some northern Kurdish (Kurmanji) dialects of North Iraq.24 Here the present continuous can be expressed by the ezafe followed by the present indicative of the verb. Among its various functions, the ezafe can be used, like d in NENA, as a nota genitivi or to introduce a relative

21

‘rélatives déictiques’ (Cadiot 1976; Benzakour 1984) and ‘relative situazionali’ (Strudsholm 1999; 2007). 22 For a thorough analysis of the pseudo-relative constructions and an outline of the history of studies on the subject, see Scarano (2002). 23 In relative clauses the particle d ~ də has a tendency to cliticize, being often either suffixed to the end of the head nominal or prefixed to the first word of the relative clause. The passage through a clitic stage is considered one of the main steps of the ‘grammaticalization cline’: lexical item > clitic > affix (Hopper and Traugott 1993: 132). 24 The type of construction analysed in this paragraph has been attested for the dialects belonging to group II in MacKenzie (1961). These are all west of ʿAnkawa, but include the neighbouring region of ʿAqra. Also Blau (1975) and Haig (2011) document the construction.

194

ROBERTA BORGHERO

clause.25 Unlike its Persian cognate, the northern Kurdish ezafe is inflected for gender and number, e.g. sg.m. (y)ē ~ (w)ē, sg.f. (y)ā ~ (w)ā, pl. (y)ēt ~ (w)ēt. It is frequently reduced to an enclitic vowel, e.g. sg.m. -ē, sg.f. -ā.26 Examples of present continuous constructions consisting of the ezafe followed by the present indicative of the verb are given below. They are taken from different Kurmanji dialects of North Iraq.27 Jabal Sinjār (8) ez 1SG

ê

dibêjim28 say.1SG

yê

xwarinê meal.OBL

REL.SG.M

‘I am saying’ Zakho (9) ez 1SG

REL.SG.M

çêdikim29 make.1SG

‘I am making/preparing a meal [right now]’. ʿAqra (10) dast-ē hand-GEN

min-ē 1SG.OBL-REL.SG.M

tēšīt30 ache.3SG

‘My hand is aching’.

25

For the ezafe in Northern Kurdish, see Haig (2011). The multiple morphosyntactical functions of d in NENA are typical of its origin as a determinative pronoun. For a thorough study of Semitic determinative pronouns, see Pennacchietti (1968). 26 MacKenzie (1961: 162); Haig (2011: 365). 27 The quoted examples come from various sources, indicated in the notes. I am following Pennacchietti (2007) in the interpretation of the ezafe as a relative particle. Haig (2011: 370– 375) is inclined to read the ezafe in this type of constructions (‘tense ezafe’) as a demonstrative/anaphoric rather than relative. Note that in the examples I have kept the transcription systems of the different authors, and this has resulted in some slight inconsistencies. 28 Blau (1975: 40). 29 Haig (2011: 372). 30 Mackenzie (1961: 205).

THE PRESENT CONTINUOUS IN THE NEO-ARAMAIC DIALECT OF ʿANKAWA

195

ʿAqra (11) Maryam-ā Maryam-REL.SG.F

txôt31 eat.3SG

‘Maryam is eating’. According to Fabrizio Pennacchietti, in Kurdish present continuous constructions like the ones above the syntagma ezafe + present indicative is to be considered as the predicate of a nominal sentence with zero copula, e.g. in the last example, ‘Maryam [is] that (f.) eats’.32 Pennacchietti’s hypothesis of a zero copula in the Kurmanji forms finds an indirect confirmation in the Neo-Aramaic dialect of ʿAnkawa, where, as explained above, the present continuous construction is attested also with the copula, e.g. ʾāylə də-kšaqəl, lit. ‘he is that takes’. On the other hand, Pennacchietti’s interpretation of the Kurmanji forms could provide evidence for the derivation of the present continuous də-kšaqəl from ʾāylə də-kšaqəl in ʿAnkawa. The passage from ʾāylə də-kšaqəl to də-kšaqəl could be explained as an internal development of the NENA dialect of ʿAnkawa, but it could have also been facilitated by the presence of forms such as ā txôt and ē tēšīt in the Kurdish dialects of the neighbouring area. The present continuous construction də-kšaqəl finds parallels also in Semitic. In Modern South Arabian the present continuous can be expressed by means of the preverbal particle d ~ ð followed by the prefix conjugation (‘imperfect’) of the verb. It is a widespread opinion among scholars that the particle d ~ ð preceding the imperfect is in origin the relative pronoun.33 Below are some examples of the construction in Mehri: (12) ḥəbūr cold

ð-yəzyūd34 REL-increase.IMPRF.3SG.M

‘The cold is increasing’. (13) b=xayr well

hē 3SG.M

‘He is well and he is working’. 31

wə and

ð-yəxawdəm35 REL-work.IMPRF.3SG.M

MacKenzie (1961: 205) discussed in Pennacchietti (2007). Pennacchietti (2007: 140). 33 See Wagner (1953), Pennacchietti (2007) and Rubin (2010). Simeone-Senelle (2003: 248) stresses rather the demonstrative/deictic nature of the particle in such constructions. 34 Rubin (2010: 145). 35 Rubin (2010: 145). 32

196 (14) hu 1SG

ROBERTA BORGHERO d-esiûr REL-go.IMPRF.1SG

bi-sebîl LOC-path

de-bâlī 36 GEN-God

‘I am following God’s path’. Pennacchietti interprets Modern South Arabian constructions of the type above, similarly to the Kurdish forms, as pseudo-relative constructions where d + prefix conjugation is the predicate of a nominal sentence with zero copula, e.g. in the last example, ‘I [am] that follow God’s path’.37 The constructions attested in the NENA dialect of ʿAnkawa (də-kšaqəl), Kurmanji Kurdish (ē tēšīt) and Modern South Arabian (ð-yəzyūd) appear to have parallel forms and functions. In all three cases the relative followed by a finite verb (present indicative/prefix conjugation) stands as a full predicate and is used to express the present continuous. The case of ʿAnkawa is particularly interesting, as the form ʾāylə də-kšaqəl might document an earlier stage of evolution of the pseudo-relative, where the syntagma də + present indicative was found in construction with the copula.38

4. THE PRESENT CONTINUOUS IN ʿANKAWA AND ITS RELATION TO PRESENT CONTINUOUS CONSTRUCTIONS IN OTHER NENA DIALECTS

If we assume that in ʿAnkawa the preverbal də of the present continuous də-kšaqəl is the relative particle, what remains to be explained is the connection with the present continuous constructions of other NENA dialects of the area, e.g. C. Koy Sanjaq lā-kpaθəx, C. Bedyal ma-kpaθəx, J. Arbel lā-palə́x ‘he is opening’, where, as explained above, the preverbal particle appears to be derived from the present copula. As already pointed out, an indication of the copular derivation of the preverbal particle in these dialects can be found in the comparison with present continuous constructions attested in Karamlesh and Qaraqosh, e.g. kilə kšaqəl ‘he is taking’, which consist of the deictic present copula (kilə) followed by the present indicative of the verb (kšaqəl). In my opinion, the forms from Karamlesh and Qaraqosh could be interpreted as asyndetic pseudo-relative constructions governed by the verb ‘to be’, i.e. kilə 36

Bittner (1914: 48 ft. 24; 82, ft. 16) discussed in Pennacchietti (2007). Pennacchietti (2007: 137; 141). 38 A construction parallel to ʾāylə də-kšaqəl is attested also in Modern Ethiopian Semitic. Here the copula combined with a pseudo-relative form of the type relative + prefix conjugation can be used as a full predicate, albeit not with the function of a present continuous (Kapeliuk 1988; forthcoming). I am very grateful to Olga Kapeliuk for giving me a copy of her as yet unpublished paper. 37

THE PRESENT CONTINUOUS IN THE NEO-ARAMAIC DIALECT OF ʿANKAWA

197

kšaqəl ‘he is [that] takes’.39 A similar interpretation could apply to the present continuous forms documented in the dialects of the region around ʿAnkawa, which appear to have originated from a construction parallel to the one attested in Karamlesh and Qaraqosh.40 The connection between the present continuous construction in ʿAnkawa and the one documented in Karamlesh and Qaraqosh (and possibly also the constructions from which the present continuous of the dialects spoken in the area of ʿAnkawa originated) would be that both are expressed by a pseudo-relative form, which is syndetic in ʿAnkawa and asyndetic in the other dialects.

5. PSEUDO-RELATIVES AND PARTICIPLES

The evolution of an original relative də-kšaqəl (‘that takes’) into a full predicate (‘he is taking’) could be related to the trend of participial forms to acquire verbal properties.41 The form də-kšaqəl (‘that takes’) actually has a function similar to the active participle (‘taker’) and could be considered somehow equivalent to it. The development of də-kšaqəl could be compared to the process of verbalization undergone by the earlier Aramaic active participle *šāqil, which in NENA has acquired full verbal force and has become the present base of the verb, e.g. Ashitha (SE Turkey) šaqəl ‘he takes’, ʿAnkawa kpaθəx ‘he opens’. Interestingly, in ʿAnkawa the active participle šaqala appears to be undergoing a similar process of verbalization. The pattern CaCaCa (~ CăCaCa) is the reflex of

39

The same remarks made for the present continuous can apply to past continuous constructions, such as Qaraqosh kiwa kšaqəl ‘he was taking’, kiwa kqaṭəl ‘he was killing’ (deictic past copula + present indicative). These could be considered parallel to Classical Arabic past continuous forms like kāna yaqtulu ‘he was killing’ (past of ‘to be’ + prefix conjugation of the verb), which have been interpreted by Pennacchietti (2007: 142–143) as asyndetic pseudo-relative constructions literary meaning ‘he was [that] kills’. 40 This should apply at least to the majority of the dialects in the region of ʿAnkawa, which appear to have a present continuous construction where the preverbal particle is followed by the present indicative of the verb (see Mutzafi 2004: 260). Note, however, that, unlike the present continuous in Karamlesh and Qaraqosh, where the copula is inflected (e.g. Qaraqosh kilə kšaqəl ‘he is taking’, kiyət kšaqlət ‘you (sg.m.) are taking’), in the dialects of region of ʿAnkawa the copula has become fossilized and remains the same throughout the paradigm, e.g. C. Koy Sanjaq lā-kpaθəx ‘he is opening’, lā-kpaθxət ‘you (sg.m.) are opening’ (Mutzafi 2004: 256). 41 For a thorough account of the tendency of nominal forms to acquire verbal force in NENA, see Kapeliuk (2008). For more examples in this sense, with special regard to constructions with the resultative participle and the infinitive, see Borghero (2005b; 2006); Khan (2006).

198

ROBERTA BORGHERO

earlier Aramaic *CaCCāCa.42 In NENA it normally indicates a nomen professionalis or the performer of a habitual activity, e.g. qaṭala ‘killer’, zamara ‘singer’, baxaya ‘weeper’. In ʿAnkawa the pattern appears to be formed productively from any verbal root and to be used without lexical restriction as an active participle of all stem I verbs. The active participle of the derived stems is formed by means of the suffix -ana, e.g. mšadrana ‘sender’ (š-d-r II), maxlana ‘feeder’ (ʾ-x-l III). The active participle may be combined with the present and the past copula and used as an independent verbal form. In combination with the present copula it functions as an immediate future, normally translated with English ‘going to ...’, e.g. pasáʾelə ‘he is going to walk’, paláṭelə? ‘is he leaving?’, paθáxelə ṭăra ‘he is going to open the door’: (15) paθáxelə (paθaxa-ilə) open.ACT.PTCP.SG.M-COP.PRS.3SG.M

ṭăra door

‘He is going to open the door’. The active participle combined with the past copula has normally the function of a ‘future in the past’ (‘he was going to ...’), often in a counterfactual sense, i.e. ‘he would have ... (but …)’, ‘he was supposed to... (but …)’, e.g. šaqálta-yawa kθawah, bas là-kəm-šaqlalə ‘she was supposed to take her book, but she did not take it’, mbašlaníθa-yawa, bas là-mbušəlla ‘she was supposed to cook, but she did not cook’: (16) mbašlaníθa-yawa cook.ACT.PTCP.SG.F-COP.PST.3SG.F

bas but

là not

mbušəl-la cook.PST-3SG.F

‘She was supposed to cook, but she did not cook’. It is interesting to point out that in ʿAnkawa the constructions active participle + present copula and active participle + past copula are used productively with any verb, be it transitive or intransitive, and can govern an explicit object.43

42

Nöldeke (1904: 72). The same constructions are attested for any type of verb also in Tisqopa (NW Iraq), e.g. ʾiwən šaqalaḥ ~ šaqálaḥ-iwən ‘I (m.) am going to take it (f.)’, wətwa šaqalaḥ ~ šaqálaḥ-wətwa ‘you (m.) were supposed to take it (f.)’ (examples according to my informants). For similar forms in other dialects see Borghero (2008) and Kapeliuk (2008). 43

THE PRESENT CONTINUOUS IN THE NEO-ARAMAIC DIALECT OF ʿANKAWA

199

6. PSEUDO-RELATIVES, PARTICIPLES AND GERUNDS IN EXPRESSIONS OF CONCOMITANCE

Further indications of a parallel between the pseudo-relative də-kšaqəl and the active participle can be found. A link between a pseudo-relative construction of this type and the participle has been suggested for other Semitic languages, in particular for Modern South Arabian and Modern Ethiopian Semitic. As regards Modern South Arabian, Ewald Wagner and Fabrizio Pennacchietti consider such a construction as a substitute for the participle.44 As for Modern Ethiopian Semitic, Olga Kapeliuk notes the attributive nature of the ‘relative verb’ (mostly static, unlike the regular verb, mostly dynamic) and sees this as an indication of it being a possible replacement of the participle.45 A further sign of the analogy between the pseudo-relative də-kšaqəl and the active participle can be seen in the comparison between the present continuous in ʿAnkawa and a parallel construction attested in Qaraqosh for the verbs ‘to come’ and ‘to go’. For these verbs in Qaraqosh the present continuous may be expressed by combining the deictic copula with the active participle, e.g. for the verb ‘to come’ (ʾ-θ-y): (17) kilə

COP.PRS.3SG.M

ʾaθoya46 come.ACT.PTCP.SG.M

‘He is coming’. As explained above, in ʿAnkawa, besides the form də-kšaqəl, we also find a marginal present continuous construction ʾāylə də-kšaqəl, where the pseudo-relative is preceded by the deictic copula. For instance, still for the verb ‘to come’, we have: (18) ʾāylə COP.PRS.3SG.M

də-kaθə REL-come.IND.3SG.M

‘He is coming’. As shown by the two examples, the present continuous constructions in ʿAnkawa and Qaraqosh are typologically parallel, and the pseudo-relative (də-kaθə) in ʿAnkawa has the same function as the active participle (ʾaθoya) in Qaraqosh.47 44

Wagner (1953: 120–121); Pennacchietti (2007: 145). Kapeliuk (forthcoming). The type of construction defined here and in Pennacchietti (2007) as ‘pseudo-relative’ is called by Kapeliuk and others ‘relative verb’. 46 Khan (2002: 349). 47 Remember, however, that this type of present continuous construction is documented in 45

200

ROBERTA BORGHERO

In the expression of the present continuous an analogy can be noted also between the two constructions above and another, very widespread in NENA, of the type hole b-šqala ~ hole šqala ‘he is taking’, e.g. Barwar hole b-xpara ~ hole xpara ‘he is digging’.48 The form hole b-šqala is generally considered the original one.49 In this construction the copula hole is followed by the preposition b- combined with the infinitive šqala. The infinitive following the preposition is actually a deverbal noun, and the form b-šqala can be considered as a single morphological unit and defined as a gerund.50 Keeping the verb ‘to come’ for the sake of comparison, such forms can be analysed as follows: (19) hole COP.PRS.3SG.M

b=θaya come.GER

‘He is coming’. In the three examples of present continuous constructions above, the active participle, the pseudo-relative and the gerund have the same function and can be translated with English ‘coming’.51 An analogy among pseudo-relatives, participles and gerunds has been pointed out for Latin and Italian, where the three forms appear in parallel constructions to express the idea of concomitance.52 In Latin, the present participle had verbal force and was used in particular constructions, notably after verbs of perceptions, to indicate a concomitant action: 53 (20) video see.1SG

hominem man.ACC

currentem54 run.PRS.PTCP.SG.ACC

‘I see a man running’.

Qaraqosh only for the verbs ‘to come’ and ‘to go’. 48 Khan (2008: 191–192). 49 Goldenberg (2000: 84); Khan (2006: 102). 50 This is discussed in Polotsky (1991: 270–272; 1996: 19–20); Kapeliuk (1996); Goldenberg (2000: 82–84). 51 It is interesting to note that the English present continuos ‘he is coming’ could be considered the result of the confluence of two different constructions: a) present of ‘to be’ + present participle of ‘to come’, b) present of ‘to be’ + preposition + gerund of ‘to come’ (on this point, see Elsness 1994: 6–8). 52 Scarano devotes a whole chapter of her book to this subject (Scarano 2002: 111–125). 53 ‘participio di simultaneità’ (Tekavčić 1972 II: 384–386; Scarano 2002: 114–115). 54 This and the following Latin and Italian examples are taken from Tekavčić (1972 II: 277) and discussed in Scarano (2002: 115–116).

THE PRESENT CONTINUOUS IN THE NEO-ARAMAIC DIALECT OF ʿANKAWA

201

In the example above the present participle (currentem) has the function of a predicate of the object (hominem) and is used to describe an action that is concomitant with the act of perceiving it. In such contexts, pseudo-relative forms were alternative constructions to the participle: (21) video see.1SG

hominem man.ACC

qui

REL.NOM.M

currit run.3SG

‘I see a man running’. In Italian the present participle gradually lost its verbal force, and its function became restricted to one of an adjective. In older sources it is still possible to find participial constructions like the Latin ones: (22) vedo see.1SG

un

INDEF.ART.SG.M

uomo man

corrente55 run.PRS.PTCP.SG

‘I see a man running’. However, these forms gradually disappeared and were replaced by constructions with the gerund instead of the present participle:56 (23) vedo see.1SG

un

INDEF.ART.SG.M

uomo man

correndo57 run.GER

‘I see a man running’. Yet such constructions were of ambiguous interpretation, as the gerund could be seen both as a predicate of the object and as a predicate of the subject.58 Because of their ambiguity, constructions of this type have almost disappeared in Modern

55

Constructions of this type are documented, for instance, in Boccaccio, e.g. Ed Espero già si poteva vedere infra li tiepidi raggi di Febo cercante l’occaso (quoted in Scarano 2002: 119). 56 Actually the use of a gerund with the function of a present participle goes back to Late Latin (Scarano 2002: 114). 57 This type of construction can be found, for instance, in Dante and Boccaccio, e.g. Dante Come occhio segue suo falcon volando; Quando la madre da Chirone a Sciro trafugò lui dormendo (quoted in Scarano 2002: 112; 113); Boccaccio Quivi trovarono i giovani giuocando; Ma Florio lasciò la madre piangendo (quoted in Scarano 2002: 113; 122). 58 Therefore the example could be interpreted both as ‘I saw a man running’ and ‘I saw a man while I was running’.

202

ROBERTA BORGHERO

Standard Italian.59 Pseudo-relative constructions parallel to the ones attested in Latin, instead, continue to be used in Italian:60 (24) vedo see.1SG

un

ART.INDEF.SG.M

uomo man

che REL

corre run.3SG

‘I see a man running’. In the examples above, as in the examples from NENA, the active participle, the pseudo-relative and the gerund have the same function. They appear to indicate a process taking place in front of the speaker, or a situation concomitant with the act of perceiving it. Similarly to the Latin and Italian constructions, where the participle, the pseudo-relative and the gerund are the predicate of the object of the verb of perception, in the present continuous constructions described for NENA (kilə ʾaθoya, ʾāylə də-kaθə, hole b-θaya) the three forms can be seen as the predicate of the subject. In actual fact, there is a more relevant analogy between NENA and Italian. This is the case when the participle, the pseudo-relative and the gerund are found after ecco, e.g. participle: Ecco Boezio in quello De Consolatione dicente: [...];61 gerund: Ed ecco il veglio onesto gridando; 62 pseudo-relative: Ecco il treno che arriva.63 These examples are noteworthy, as ecco in Italian has a presentative, attention drawing function analogous to the one of the deictic copula in NENA.64 59

There are very rare instances of this type of constructions in the 19th and 20th centuries authors, e.g. Verga Lodovico scorse Giovanni e Maria in piedi ciarlando affabilmente (Scarano 2002: 123); Bernari Sicuramente quando tornerò la troverò piangendo (Scarano 2002: 112). The construction is still in use in some Italian dialects, e.g. Calabrian U dassai mangiandu; A trovai durmennu (Scarano 2002: 113). 60 Scarano (2002: 115–116); Tekavčić (1972 II: 277). In order to express a concomitant action after verbs of perception, Latin and Italian use also a type of construction with the infinitive, e.g. Latin Video hominem currere, Italian Vedo un uomo correre ‘I see a man running’. According to Scarano (2002: 95–110; 115) this construction is not exactly equivalent to the others. Expressions of concomitance containing an infinitive could be compared to NENA present continuous forms like hole θaya (copula + infinitive) ‘he is coming’. As explained above, however, these are likely to be derived from constructions, such as hole b-θaya, where the infinitive was preceded by the preposition b-. 61 Lit. ‘Here is Boethius saying in De Consolatione: [...]’ (Dante, Convivio, IV, xii; Scarano 2002: 121). 62 Lit. ‘And here is the honest old man crying’ (Dante, Purgatorio, II, vv. 119–120; Scarano 2002: 113). 63 Lit. ‘Here is the train arriving’ (Modern Standard Italian). 64 Incidentally, there is a clear relationship between deictics and verbs of perception, as

THE PRESENT CONTINUOUS IN THE NEO-ARAMAIC DIALECT OF ʿANKAWA

203

As a form equivalent to a participle, the pseudo-relative də-kšaqəl has gone through a similar development. Like the earlier Aramaic active participle *šāqil, which has specialized as the present of the verb, and the NENA active participle šaqala, which in combination with the copula has taken the functions of immediate future and ‘future in the past’, də-kšaqəl (originally ‘that takes’) has found its own place in the verbal system of ʿAnkawa and has specialized as a present continuous (‘he is taking’). The process of grammaticalization undergone by də-kšaqəl is remarkable. It is not attested in other Neo-Aramaic dialects, nor, to my knowledge, in earlier Aramaic, but has parallels in South Semitic (Modern South Arabian and Modern Ethiopian Semitic) and in Kurdish.65 This phenomenon deserves to be included in the linguistic literature on grammaticalization.

REFERENCES

Benzakour, Fouzia. 1984. “Les Relatives déictiques.” In Recherches en Pragma-Semantique. Etudes. Recherches Linguistiques 10, edited by Georges Kleiber, 75–106. Paris: Klincksiek. Bittner, Maximilian. 1914. Studien zur Laut- und Formenlehre der Mehri-Sprache in Südarabien. Wien: Sitzungsberichte der Kais. Akademie der Wissenschaften in Wien, Philologisch-Historische Klasse 176.1. Blanc, Haim. 1964. Communal Dialects in Baghdad. Cambridge (Mass.): Harvard University Press. Blau, Joyce. 1975. Le Kurde de ʿAmâdiya et de Djabal Sindjâr. Paris: Klincksiek. —. 2011. Manuel de Kurde Sorani. Paris: L’Harmattan. Blau, Joyce, and Barak, Veysi. 2003. Manuel de Kurde Kurmancî. Paris: L’Harmattan. Borghero, Roberta. 2005a. The Neo-Aramaic Dialect of Ashitha. Unpublished PhD Dissertation. University of Cambridge.

they both imply an element of presentation, and it is not surprising to find the same type of constructions in the two contexts. On this point, Strudsholm (2007) talks about a continuum between ‘explicit’ and ‘implicit’ perception. 65 The existence of pseudo-relative constructions used as fully independent verbal forms in Neo-Aramaic could be significant, as this phenomenon has been hypothesized to be an areal feature or an isogloss of other sub-groups of Semitic (Pennacchietti 1993; 2007; Kapeliuk forthcoming), that is Modern South Arabian and Ethio-Semitic (Kapeliuk) or the broader area of South Semitic and Arabic (Pennacchietti, who includes in his analysis also Arabic constructions interpreted as asyndetic relatives).

204

ROBERTA BORGHERO

—. 2005b. “The Evolution of the Verbal System in the North Eastern Neo-Aramaic Dialect of Ashitha.” In Afro-Asiatic Studies, 11th Italian Meeting of Afro-Asiatic Linguistic, edited by Alessandro Mengozzi, 325–336. Milano: Franco Angeli. —. 2006. “Some Features of the North Eastern Neo-Aramaic Dialect of Ashitha.” In Loquentes Linguis. Linguistic and Oriental Studies in Honour of Fabrizio A. Pennacchietti, edited by Giorgio Borbone, Alessandro Mengozzi and Mauro Tosco, 111–122. Wiesbaden: Harrassowitz. —. 2008. “The Verbal System of the Neo-Aramaic Dialect of Karamlesh.” In Neo-Aramaic Dialect Studies, edited by Geoffey Khan, 75–90. Piscataway, NJ: Gorgias Press. Cadiot, Pierre. 1976. “Les relatives et infinitives ‘deictiques’ en français.” Documentation et Recherche en Linguistique Allemande Contemporaine 13: 1–64. Coghill, Eleanor. 2008. “Some Notable Features in North-Eastern Neo-Aramaic Dialects of Iraq.” In Neo-Aramaic Dialect Studies, edited by Geoffrey Khan, 91–104. Piscataway, NJ: Gorgias Press. Elsness, Johan. 1994. “On the Progression of the Progressive in Early Modern English.” International Computer Archive of Modern and Medieval English 18: 5–25. Goldenberg, Gideon. 2000. “Early Neo-Aramaic and Present-day Dialectal Diversity.” Journal of Semitic Studies 46: 69–89. Haig, Geoff L. J. 2011. “Linker, Relativizer, Nominalizer, Tense-Particle: On the Ezafe in West Iranian.” In Nominalization in Asian Languages: Diachronic and Typological Perspectives (Typological Studies in Languages 96), edited by Foong H. Yap, Karen Grunow-Hårsta and Janick Wrona, 361–390. Amsterdam: John Benjamins Publishing Company. Hopper, Paul J., and Traugott, Elizabeth C. 1993. Grammaticalization. Cambridge: Cambridge University Press. Jastrow, Otto. 1978. Die mesopotamisch-arabischen qeltu-Dialekte. Band I: Phonologie und Morphologie. Wiesbaden: Steiner. Kapeliuk, Olga. 1988. “Nominalization in Amharic.” Aethiopische Forschungen 23. Wiesbaden: Steiner. —. 1996. “The Gerund and Gerundial Participle in Eastern Neo-Aramaic.” Sprachtypologieund Universalien Forschung 51 (3): 276–288. —. 2008. “Between Nouns and Verbs in Neo-Aramaic.” In Neo-Aramaic Dialect Studies, edited by Geoffrey Khan, 131–147. Piscataway, NJ: Gorgias Press. —. forthcoming. “The Relative Verb: a Common Isogloss of Modern South Arabian and Ethio-Semitic?” In Afro-Asiatic Studies, 14th Italian Meeting of Afro-Asiatic Linguistic, edited by Alessandro Mengozzi. Khan, Geoffrey. 1999. A Grammar of Neo-Aramaic. The Dialect of the Jews of Arbel. Leiden: Brill. —. 2002. The Neo-Aramaic Dialect of Qaraqosh. Leiden: Brill.

THE PRESENT CONTINUOUS IN THE NEO-ARAMAIC DIALECT OF ʿANKAWA

205

—. 2006. “Remarks on Compound Verbal Forms in North Eastern Neo-Aramaic.” In Current Issues in the Analysis of Semitic Grammar and Lexicon II, vol. 2, edited by Lutz Edzard and Jan Retsö, 101–114. Wiesbaden: Harrassowitz. —. 2008. The Neo-Aramaic Dialect of Barwar. Leiden: Brill. MacKenzie, David Neil. 1961. Kurdish Dialect Studies I. London: Oxford University Press. Mutzafi, Hezy. 2004. “Features of the Verbal System in the Christian Neo-Aramaic Dialect of Koy Sanjaq and Their Areal Parallels.” Journal of the American Oriental Society 124 (2): 249–264. Nöldeke, Theodore. 1904. Compendious Syriac Grammar. Eugene, Oregon: Wipf and Stock Publishers. Pennacchietti, Fabrizio A. 1968. Studi sui pronomi determinativi semitici. Napoli: Istituto Orientale di Napoli. —. 1993. “Le forme verbali pseudorelative: isoglossa strutturale del semitico sudoccidentale.” In Serta Philologica Constantino Tsereteli Dicata, edited by Ricardo Contini, Fabrizio A. Pennacchietti and Mauro Tosco, 213–225. Torino: Silvio Zamorani Editore. —. 2007. “L’impiego di frasi pseudorelative come verbi finiti.” In Relative e Pseudorelative tra Grammatica e Testo, edited by Federica Venier, 133–148. Alessandria: Edizioni dell’Orso. Polotsky, Hans J. 1991. “Modern Syriac Conjugation.” Journal of Semitic Studies 36: 263–277. —. 1996. “Notes on a Neo-Syriac Grammar.” Israel Oriental Studies 16: 11–48. Radford, Andrew. 1975. “Pseudo-Relatives and the Unity of Subject Raising.” Archvium Linguisticum 6: 32–64. Rubin, Aaron. 2010. The Mehri Language of Oman. Leiden: Brill. Scarano, Antonietta. 2002. Frasi relative e pseudo-relative in italiano. Roma: Bulzoni. Simeone-Senelle, Marie-Claude. 2003. “De quelques fonctions de đ- dans les langues suda- abiques modernes.” In Perspectives synchroniques sur la grammaticalisation. Polysémie, transcatégorialité et échelles syntaxiques, edited by Stéphane Robert, 239–252. Louvain-Paris: Peeters. Strudsholm, Erling. 1999. Relative situazionali in italiano moderno. Münster: LIT. —. 2007. “La ‘relativa situazionale’ tra testo e contesto. Una reinterpretazione della cosiddetta pseudo relative.” In Relative e Pseudorelative tra Grammatica e Testo, edited by Federica Venier, 117–132. Alessandria: Edizioni dell’Orso. Tekavčić, Pavao. 1972. Grammatica storica dell’italiano. Bologna: Il Mulino. Thackston, Wheeler. M. 2006a. Kurmanji Kurdish: A Selected Grammar with Selected Readings. Lecture Notes. Iranian Studies, Harvard University. —. 2006b. Sorani Kurdish: A Reference Grammar with Selected Readings. Lecture Notes. Iranian Studies, Harvard University.

206

ROBERTA BORGHERO

Van Wagoner, Merrill Y. 1944. A Grammar of Iraqi Arabic. Unpublished PhD Dissertation. Yale University. Wagner, Ewald. 1953. Syntax der Mehr-Sprache. Berlin: Akademie Verlag.

CAUSATIVE-INCHOATIVE ALTERNATION IN NORTHEASTERN NEO-ARAMAIC

KATHRIN GÖRANSSON 1. INTRODUCTION (1) Assyrian Kirkuk a. kteta bšǝlla gu-bašlanta. ‘The chicken cooked in the kitchen’. b. bašlana bušəlle gu-bašlanta. ‘The cook cooked in the kitchen’. What is the difference between these two clauses? At first sight, their structures appear to be identical: both have a subject, a verb and a prepositional phrase. They differ, however, in terms of transitivity and grammatical voice. The verb in the first clause is an intransitive and semantically middle verb and will be referred to as inchoative1, whereas the verb in the second clause is an active transitive verb that functions as a causative (a direct object, e.g. kteta ‘chicken’, being implied). It should be noted that from a semantic point of view both transitivity and voice are relative and not absolute categories; there is a continuum between transitive and intransitive clauses as well as between active, middle and passive voice (Hopper and Thompson 1980; Kemmer 1993).

1

This term will henceforth be used to denote an event that happens spontaneously without implying a cause. It should not be confused with the aspectual category bearing the same name. It has also been labelled ‘anticausative’, which in this paper refers to a particular morphological type of inchoative (see footnote 14 below).

208

KATHRIN GÖRANSSON

The grammatical subject of a causative verb is a cause (i.e. a causer or a causing event), which brings about some kind of change in the direct object (the patient). An inchoative verb, on the other hand, presents an action as happening spontaneously. Its subject is a patient, and thus has the same semantic role as the direct object of the corresponding causative. Thus, in North-Eastern Neo-Aramaic (henceforth NENA) and in many other languages, the same event can be expressed either as an inchoative (if the cause is unknown, irrelevant or lacking) or as a causative (containing additional information about the cause): (2) Assyrian Kirkuk a. kteta chicken(F) patient

bšǝl-la cook.PST-ERG.3SG.F2 inchoative (stem I)

gu-bašlanta. in-kitchen(F)

‘The chicken cooked in the kitchen’. b. bašlana

bušl-a-le

agent

causative (stem II)

cook(M)

cook.PST-ABS.3SG.F-ERG.3SG.M

kteta

chicken(F) patient

gu-bašlanta.

in-kitchen(F)

‘The cook cooked the chicken in the kitchen’. In this paper, the character of causative-inchoative alternations will first be explored from a typological perspective and the difference between the inchoative and the passive discussed. Second, a concise overview will be given of the different formal causative-inchoative alternation types that occur cross-linguistically. Next, the way causative-inchoative alternation manifests itself in NENA will be examined by studying the dialects of Assyrian Kirkuk, Christian Barwar, Jewish Sanandaj and Jewish Urmi. In addition, a type of causative-inchoative alternation that was used in earlier Aramaic and that has left traces in NENA will be discussed.3 Finally, some 2

NENA has morphological ergativity (cf. Khan and Doron 2012). In transitive clauses in the past base (denoting perfective aspect), ergative suffixes (also known as L-suffixes) encode the subject, whereas absolutive suffixes (also called S-suffixes) mark the direct object, which alternatively may be replaced by a prepositional phrase or be added in the form of another Lsuffix after the subject. The subject of intransitive clauses is, depending on the dialect and the type of verb, marked by an absolutive or by an ergative suffix. In the present base, which inflects according to a nominative-accusative pattern, this is reversed: absolutive suffixes mark the subject and L-suffixes the direct object. 3 The data of Assyrian Kirkuk have been obtained through my own fieldwork, while those

CAUSATIVE-INCHOATIVE ALTERNATION IN NORTH-EASTERN NEO-ARAMAIC

209

examples of Assyrian Kirkuk verbs will be provided to illustrate how causativeinchoative alternation interacts with other verbal stems and periphrastic constructions.

2. CAUSATIVE AND INCHOATIVE

Causative-inchoative alternation is a transitivity alternation, which is also known as ‘causative alternation’, ‘inchoative/causative alternation’ or ‘causative-anticausative alternation’. Haspelmath (1993), whose terminology will be employed here, uses the terms ‘causative alternation’ and ‘anticausative alternation’ for particular morphological types of causative-inchoative alternation (see section 3). This is why the alternation in general is referred to here as ‘causative-inchoative alternation’ rather than ‘causative alternation’ or ‘causative-anticausative alternation. While the causative is relatively easy to identify, this is not always the case with the inchoative, which is closely related to the passive voice.4 The inchoative is part of a cluster of semantically related categories, which are sometimes referred to as middle voice and centre around the direct reflexive.5 According to Kemmer 1993, the middle consists of various lexical classes such as grooming verbs, verbs of motion or cognitive verbs that share certain semantic properties. The inchoative alternant of the causative-inchoative alternation constitutes one of these lexical classes:6 (3) Assyrian Kirkuk šmaṭa ptaxa myata

break (intr.) open (intr.) die

of C. Barwar come from Geoffrey Khan’s work (2008b). The data of J. Urmi and J. Sanandaj are based partly on my fieldwork and partly on Khan 2008a and 2009. I would like to thank Nineb Lamassu, Danny Avrahami, Sarah Avrahami, and Aviv Nisan for their time and patience in answering my questions; Geoffrey Khan for countless helpful and illuminating discussions; and Holger Gzella, Leonid Kulikov, as well as the participants of the conference on ‘NeoAramaic Dialectology: Jews, Christians and Mandaeans’ in Jerusalem in June 2013, where an earlier version of this paper was presented, for their insightful comments. 4 See Kulikov 2001 for an overview on causatives. The causative should not be considered a voice, because it can be combined with voice markers in one form (whereas true voices cannot; Kulikov 2001: 889). 5 Cf. Kemmer’s typological study (1993), which provides a comprehensive account of the middle voice, and its relations to the passive, the reflexive and the reciprocal. 6 Referred to as ‘spontaneous events middle’ by Kemmer (1993).

210

KATHRIN GÖRANSSON pšara mlaya pqaya

melt (intr.) fill (intr.) explode (intr.)

The inchoative is closely related to the passive: in many languages, inchoatives such as ‘break (intr.)’, ‘open (intr.)’ or ‘fill (intr.)’ are not formally distinguished from the passives ‘be broken’, ‘be opened’ or ‘be filled’. This is in contrast to the corresponding causatives, which, being active transitive, take a different form: (4) Swedish Inchoative/passive

Causative

bryta-s

break (intr.)/be broken

bryta

break (tr.)

öppna-s

open (intr.)/be opened

öppna

open (tr.)

fylla-s

fill (intr.)/be filled

fylla

fill (tr.)

Although the inchoative and the passive have similar semantics and can take one and the same form, they differ in one crucial aspect. The inchoative has only one argument (a patient) and is therefore both semantically and syntactically relatively intransitive. The passive voice, in contrast, is syntactically intransitive but semantically transitive. Syntactically, it appears as a low-transitivity event due to the demotion of the agent, whereas semantically it corresponds to an active event that has two arguments (an agent and a patient) and is therefore relatively high in transitivity. Thus, although it is impossible to formally distinguish the inchoative from the passive in some languages, there is a crucial semantic difference between the two: the inchoative presents an event as occurring spontaneously, whereas the passive at least implies an agent, and sometimes makes it explicit by means of an oblique prepositional phrase: (5) a. Inchoative7 Assyrian Kirkuk bzi-la pierce.PST-ERG.3SG.F

gutta. ball(F)

‘The ball got pierced’. 7

J. Sanandaj has a similar construction. However, the past base is used with S-suffixes (glossed as absolutive) instead of the L-suffixes (glossed as ergative) that are employed in dialects such as Assyrian Kirkuk or C. Barwar.

CAUSATIVE-INCHOATIVE ALTERNATION IN NORTH-EASTERN NEO-ARAMAIC

211

b. Passive (periphrastic) in Christian dialects8 Assyrian Kirkuk gutta ball(F)

pǝš-la become.PST-ERG.3SG.F

bzita pierce.RES.SG.F

(b-šušan). by-Susan

‘The ball was pierced (by Susan)’. c. Passive (periphrastic) in Jewish dialects9 J. Sanandaj (Khan 2009: 312) mam-i paternal.uncle-my

qṭila kill.RES.SG.M

xar. become[ABS.3SG.M]10

‘My uncle will be killed (by somebody)’. In NENA the passive is expressed by a periphrastic construction consisting of the verb pyš (Christian dialects) or xdr (Jewish dialects), both meaning ‘become’, together with the resultative participle. However, in addition to the fact that a large number of NENA dialects have several passive constructions, the inchoative, too, is more complex than it might appear at first sight. In both Assyrian Kirkuk and J. Sanandaj, the inchoative can occur with a prepositional phrase denoting a cause. In NENA, as in languages such as Swedish, the border between inchoative and passive therefore seems to be less than clear-cut.11 This seems to perpetuate an old tradition, since earlier Aramaic had one form, the t-stems, to express the inchoative and the passive, as well as the reflexive and reciprocal.

8

This construction is apparently restricted to Christian dialects and is attested, apart from in Assyrian Kirkuk, at least in the Christian dialects of Barwar, Qaraqosh, Khabur and Bohtan (cf. Khan 2008b; 2002; Talay 2008; Fox 2009). 9 This construction is used with telic verbs only (Khan 2009: 309) and seems to be restricted to Jewish dialects; it has been found in the Jewish dialects of Sanandaj, Urmi and Koy Sanjaq (cf. Khan 2009; 2008a; Mutzafi 2004). 10 Since the present base has many different functions, it will not be given a particular label. 11 However, this is still a tentative finding and part of my ongoing PhD research at the University of Cambridge. For a cross-linguistic discussion of the smooth transition between inchoatives (i.e. spontaneous events middles) and passives see Kemmer 1993.

212

KATHRIN GÖRANSSON

3. FORMAL TYPES OF CAUSATIVE-INCHOATIVE ALTERNATION

There are different ways of expressing causative-inchoative alternation crosslinguistically. In addition to English labile verbs, which mark neither the inchoative nor the causative alternant (i.e. verbs such as break, melt, open, which are both intransitive and transitive), there is a range of formal devices that languages employ: morphological markers (for either or both the causative and the inchoative alternant), auxiliary verbs and different lexemes.12 Haspelmath (1993) has suggested three main types of causative-inchoative alternation,13 according to the way they mark the inchoative and/or causative alternant, viz. causative, anticausative14, and non-directed alternation. Every expression type can be marked in three different ways: by stem modification, affixation, or the use of an auxiliary verb. The three main expression types are distinguished by the element(s) that is (are) marked in this way: the causative alternant, the inchoative alternant, or both or neither the causative and the inchoative alternant. Causative alternation refers to an expression type in which the inchoative alternant is formally basic or less marked and the causative alternant is derived or more marked. This type of alternation occurs, for instance, in French: (6) French fondre faire fondre

melt (intr.) melt (tr.)

inchoative causative

In anticausative alternation the causative alternant is basic or less marked, whereas the inchoative alternant is formally derived or more marked. Modern Hebrew is an example of a language with anticausative alternation:

12

Cf. Comrie 2006 and Nichols, Peterson and Barnes 2004 for typological studies of causative-inchoative alternation and of transitivization respectively. 13 Henceforth referred to as ‘expression type’ or ‘alternation type’. 14 With ‘inchoative’ and ‘anticausative’ there are two different terms available that help to distinguish between the semantic and the morphological aspect of the intransitive alternant respectively. Unfortunately, there is only a single term for the transitive alternant: no matter whether it is used in a semantic (the counterpart to ‘inchoative’) or a morphological sense (referring to forms with a causative marker, its counterpart being ‘anticausative’), it is always referred to as ‘causative’ (Haspelmath 1993: 108).

CAUSATIVE-INCHOATIVE ALTERNATION IN NORTH-EASTERN NEO-ARAMAIC

213

(7) Modern Hebrew15 patax ni-ftax

open (tr.) open (intr.)

causative inchoative

Non-directed alternation constitutes, next to causative and anticausative alternation, the third main type of causative-inchoative alternation. In non-directed alternation, neither the causative nor the inchoative alternant is formally derived from the other. Haspelmath (1993: 91) distinguishes three categories of non-directed alternation, which go back to Nedjalkov (1969): labile, equipollent and suppletive alternation.16 Two of these subtypes, labile and suppletive alternation, are attested in NENA. Many languages have verbs in which the difference between the causative and the inchoative alternant is not apparent because they share the same form. Such verbs are commonly referred to as labile verbs. They are very common in English: (8) The vase broke. John broke the vase.

inchoative causative

The term ‘form’, which Haspelmath and other scholars use, is a highly ambiguous term. In NENA, causative-inchoative alternation is not expressed by means of vowel changes within stems or different suffixes, and alternation pairs that differ in inflection (cf. example 17 below) are therefore still considered labile verbs. In this article, a labile verb is defined as a verb whose inchoative and causative alternants are not formally distinguished from each other by means of derivational or syntactic processes, i.e. by using different stems or auxiliary verbs. In equipollent alternation, both verbs are derived from the same stem by two different markers, none of which seems to be more basic or less marked. (9) Japanese atum-aru atum-eru

15

gather (intr.) gather (tr.)

Note that Modern Hebrew employs a wide range of alternation types. In addition to anticausative alternation, it makes frequent use of causative alternation and to a lesser extent non-directed alternation. See Haspelmath 1993 for two examples of equipollent verbs as well as one labile verb (the latter of which, however, seems questionable). On the other hand, it appears that the Hifʿil, which corresponds to NENA stem III and the C-stem of earlier Aramaic, has a number of labile verbs such as hilbin ‘whiten (intr. and tr.)’, similarly to stem III alternation in NENA (see section 5.1 below). 16 However, Nedjalkov used different terms for these three subtypes.

214

KATHRIN GÖRANSSON

Suppletive alternation denotes the use of two different lexemes for the inchoative and causative alternant respectively: (10) Russian goret’ žeč

burn (intr.) burn (tr.)

The choice of expression type for a particular verb pair gives an indication of how the speakers of a language conceptualize the relationship between the inchoative and the causative alternant of the verb (Haspelmath 1993; Comrie 2006; Haspelmath et al. 2014). Cross-linguistically, if the inchoative alternant of a particular verb is the default form, i.e. the one that is most widely used by the speakers of a language, that verb tends to occur in causative alternation. In many of the languages that Haspelmath, Comrie and Haspelmath et al. surveyed, this is the common pattern for verbs such as boil, freeze or dry; these verbs most naturally seem to happen spontaneously. Conversely, if the causative is the more commonly used alternant, the verb is more likely to occur in anticausative alternation. In a large number of languages this pattern prevails with verbs such as split, close, or break, where the causer often seems to play an important role. While many languages have grammatical markers for both the causative and the inchoative, not all of them do. These tendencies apply across languages rather than to individual languages. Finally, there are alternation pairs where no alternant is more marked than the other. In NENA (and in other languages), non-directed alternation in the form of labile alternation resulted from changes in the verbal system rather than being induced by the semantics of a verb.17 Apart from the fact that semantics has an impact on the expression type in which individual verbs are used, languages have the tendency to make disproportionate use of one or two expression types (cf. Haspelmath 1993). Causative alternation is the most widespread and productive expression type in NENA, followed by labile alternation, which is also relatively common. Suppletive alternation is extremely rare.

17

However, since labile alternation seems to occur mainly in verbs that previously had anticausative alternation (see section 6), it may be possible that in NENA many labile verbs have semantics that are typical of anticausative verbs.

CAUSATIVE-INCHOATIVE ALTERNATION IN NORTH-EASTERN NEO-ARAMAIC

215

4. CAUSATIVE ALTERNATION

In NENA the causative alternant is either marked by stem modification (most frequently) or by the use of a causative auxiliary such as šwq ‘allow, cause, leave’. The latter is, however, rarely used as causative alternant of an inchoative verb and mainly has other, related functions instead. Causative alternation by means of stem modification has a long history in Aramaic. Already in Old Aramaic, the C(ausative) stem, the predecessor of what in most dialects corresponds to stem III in NENA, was used to derive causative from inchoative (as well as from other, more transitive) verbs. Today stem I/III causative alternation18 is the most widespread and productive type of causative-inchoative alternation and seems to be attested in most, if not all, NENA dialects:19 (11) Stem I/III, Assyrian Kirkuk a. Inchoative, stem I č̭mi-le extinguish.PST-ERG.3SG.M +

nura. fire(M)

‘The fire went out’. b. Causative, stem III muč̭mi-la extinguish.PST[ABS.3SG.M]-ERG.3SG.F +

‘She extinguished the fire’.

nura. fire(M)

Another type of causative alternation in NENA is stem I/II alternation (see also example 2 above): (12) Stem I/II, Assyrian Kirkuk a. Inchoative, stem I bdǝr-run spill.PST-ERG.3PL

‘The water spilled’. 18

miya. water(PL)

Hereafter referred to as ‘stem I/III alternation’; given that two stems are involved it is obvious that this is a type of causative rather than labile alternation. Similarly, e.g. ‘stem I labile alternation’ will be called ‘stem I alternation’. 19 However, in some dialects east of the Great Zab river such as J. Urmi, which have lost one stem, it is stem II which functions as causative of stem I (cf. Khan 2008a: 66).

216

KATHRIN GÖRANSSON b. Causative, stem II budr-e-le spill.PST-ABS.3PL-ERG.SG.M

‘He spilled the water’.

miya. water(PL)

Although stem I/II alternation occurs in a large number of dialects, it is used much less frequently than stem I/III alternation. Apart from C. Barwar, Assyrian Kirkuk and J. Sanandaj, it is attested in C. Qaraqosh, C. Sardarid, J. Challa, J. Amədya, and presumably other dialects.20 On the other hand, there are many dialects, mainly situated east of the Great Zab river, in which the D-stem of earlier forms of Aramaic (which in the modern dialects that preserved all three active stems survives in the form of stem II) merged into the modern stem III and/or stem I (Mutzafi 2004: 12). This appears to be an areal feature, since it is attested in both Jewish and Christian dialects, such as J. Urmi, J. Barzan, J. Koy Sanjaq, C. Koy Sanjaq, C. Sulemaniyya, and C. Sanandaj. Due to the loss of the former D-stem, these dialects do not have an equivalent to the causative I/II alternation that is found in the dialects that have preserved all three active stems of pre-modern Aramaic. The meanings of stem II and III verbs seem to overlap. However, generally speaking stem II has a more idiosyncratic meaning and is less likely to be the causative counterpart of a stem I inchoative. Stem III, on the other hand is—and already was in earlier Aramaic—the most productive way of forming causatives (cf. Khan 2008b: 264). A third and rare type of causative alternation involves the causative auxiliary verb šwq. It is used in Assyrian Kirkuk as the causative counterpart of mainly inchoative (but also other semantically middle) stem III forms: (13) Stem III/šwq, Assyrian Kirkuk a. Inchoative, stem III nisan d-wərre spring REL-past

wardə flowers

‘Last spring the flowers sprouted’.

mujyən-nun. sprout.PST-ERG.3PL

b. Causative, aux. verb šwq + stem III šwəq-li cause.PST-ERG.1SG

‘I made them sprout’. 20

d-majyənn-i. REL-sprout-ABS.3PL

Cf. Khan (2002); Younansardaroud (2001); Fassberg (2010) and Greenblatt (2011).

CAUSATIVE-INCHOATIVE ALTERNATION IN NORTH-EASTERN NEO-ARAMAIC

217

These stem III forms are characterized by their semantics, which often denote spontaneous events without an obvious cause. Despite the fact that they occur in stem III, their interpretation as inchoatives is therefore most natural. For situations in which speakers nevertheless wish to express a cause, the option with šwq is available. A much more widespread function of šwq is to denote a permissive or indirect causative meaning. While either interpretation is possible, at least in Assyrian Kirkuk this construction is more often used to express the permissive: (14) Permissive/indirect causative with šwq + stem II, Assyrian Kirkuk21 šwǝq-le cause.PST-ERG.3SG.M

d-badr-a-lun

REL-spill-ABS.3SG.F-OBJ.3PL

miya. water(PL)

‘He allowed/caused her to spill the water’. Compared with a direct causative with an agentive causer and a patientive causee (see examples 11 and 12 above), an indirect causative has one additional argument: an agentive causer causes an agentive causee 1 to do something to a patientive causee2. Thus, there is an intermediary causee and as a result the causer “does not get physically involved in the execution of the caused event” (Shibatani and Pardeshi 2002: 140). The permissive has the same argument structure and only differs from the indirect causative in that it is a weaker, less controlling form of causation. The expression of the indirect causative and the permissive differs in the various dialects.22 Stem I/III, and to a much lesser extent stem I/II, alternation is not restricted to causative-inchoative alternation: non-inchoative, both intransitive and transitive, verbs can also be causativized.23 If derived from a transitive verb, the resulting causative may retain the original object of the transitive verb as its object and so has a passive sense:24

21

This example involves a stem II form of a stem I/II alternation verb, but šwq can also be used as an indirect causative or permissive with any other stem (see section 7). 22 Regarding the situation in C. Urmi see Khan (forthcoming). 23 As Nichols et al. (2004) and Haspelmath (1993) demonstrated, this is a common phenomenon cross-linguistically. 24 For a similar construction (direct causatives turning into passives) in various languages cf. Kulikov 2011: 394f. For this phenomenon in NENA, see Khan (forthcoming).

218

KATHRIN GÖRANSSON

(15) Causativisation of transitive verbs, C. Barwar (Khan 2008b) qṭala maqṭole

kill cause to kill/ cause to be killed

stem I stem III

transitive causative

bnaya mabnoye

build cause to build/ cause to be built

stem I stem III

transitive causative

Also intransitive and middle verbs (as understood by Kemmer 1993) can be causativized: (16) Causativization of intransitive and middle verbs, Assyrian Kirkuk dmaxa madmuxe

sleep cause to sleep

stem I stem III

pšama mapšume

become sad; feel sorry stem I sadden; cause to feel sorry stem III

intransitive causative emotion middle causative

5. NON-DIRECTED ALTERNATION

In addition to causative alternation, non-directed alternation in the form of labile alternation is common in NENA. In the dialects surveyed, there is also at least one instance of suppletive alternation.

5.1. Labile alternation NENA labile verbs are attested in all four stems. However, they occur mostly in stem I, and to a lesser extent in quadriliteral stems (Q). Labile stem III verbs are quite rare and stem II verbs very rare. The origins of these labile alternations differ, and so do the reasons for the emergence of labile alternation in the four stems, some of which will be explored below. Stem I alternation is common in all four dialects: (17) Stem I, J. Urmi a. Inchoative top ball(F)

+

‘The ball burst’.

pqe-la. burst.PST-ERG.3SG.F

CAUSATIVE-INCHOATIVE ALTERNATION IN NORTH-EASTERN NEO-ARAMAIC

219

b. Causative yale children

ǝl-+top OBJ-ball(F)

pəqy-a-lu. burst.PST-ABS.3SG.F-ERG.3PL

‘The children burst the ball’. J. Urmi seems to have a higher proportion of labile stem I verbs than the other three dialects. This might be due to the fact (already mentioned above) that the D-stem has been lost in this dialect. As a result, such verbs have merged with stem I verbs, and stem I/II alternation became stem I alternation verbs. This is exactly the opposite of what appears to have happened to the former t-stems, where the inchoative t-stem alternant may have joined the causative stem I alternant in stem I (see section 6). It can be safely assumed that stem I alternation is the oldest type of labile alternation in Aramaic, since stem I has always been undetermined for transitivity;25 there might even be examples of labile stem I (i.e. G-stem) verbs in earlier Aramaic.26 Moreover, stem I alternation is by far the most common type of labile alternation, followed by labile stem Q verbs, which seem to be a more recent development. Although stem Q alternation occurs less frequently than stem I alternation, it is well represented in all four dialects: (18) Stem Q, C. Barwar (Khan 2008b) a. Inchoative našuθa humanity(F)

la-talq-a NEG-perish-ABS.3SG.F

mxaθxəθ-a. renew-ABS.3SG.F

‘Humanity should not disappear but should be renewed’. b. Causative qəm-mxaθxəθ-la PST-renew[ABS.3SG.M]-OBJ.3SG.F

xəzmayuθa. kinship.relationship

‘He renewed the family relationship’.

25

This is demonstrated by the fact that pre-modern Aramaic had both causative and anticausative alternation, in which stem I forms, being unmarked, functioned as inchoatives (in causative alternation) and causatives (in anticausative alternation) respectively. 26 This finding is still tentative.

220

KATHRIN GÖRANSSON

The reason for the relatively large number of stem Q verbs (compared to stem III or stem II alternation) may be that these verbs cannot usually be causativized. This is due either to their large number of root consonants, which by adding an mprefix would increase even more, or because they already have an m-prefix in both alternants, as for instance in C. Barwar (cf. Khan 2008b: 163). There is no m-prefix in stem Q in J. Sanandaj and it seems to be very rare in Assyrian Kirkuk. In J. Urmi, where it is also generally lacking (cf. Khan 2008a: 66f.), the m-prefix can occasionally be used to distinguish between the inchoative and the causative. However, in some cases, it may be employed also with the inchoative alternant, whereas in others it can only be used with the causative alternant if the causer is an agent. In J. Urmi, too, the alternation involving stem Q is therefore labile rather than causative. Both stem Q and stem II (see also below) include denominative verbs. Stem II verbs have traditionally been relatively high in transitivity. Consequently, stem II would not have been the obvious means for the formation of denominative verbs, especially in cases in which the intransitive is the more commonly used alternant. Stem Q, in contrast, is not marked for transitivity, which resulted in the emergence of a relatively large number of stem Q alternation verbs. Since such verbs ended up in labile alternation regardless of their semantics, this might to some degree distort the otherwise strong tendency (and possibly even rule) of labile verbs being restricted to non-agentive and telic verbs (Haspelmath et al. 2014: 5). This contrasts with causative alternation, which is less restrictive (as is demonstrated in examples 15 and 16 above). Another reason for the preferred use of stem Q for denominative verbs might also be that it is phonologically more accommodating of nouns and adjectives than stem II due to its larger number of root consonants. Stem III alternation features in all four dialects, but it is not very common: (19) Stem III, Assyrian Kirkuk a. Inchoative ’awo that

bara mubrəq-le light shine.PST-ERG.3SG.M

‘That light shone’. b. Causative ’awo that

bara light(M)

mubrəq-a-le shine.PST-ABS.3SG.F-ERG.3SG.M

pǝnjara window(F)

‘That light made the window shine’. In J. Urmi (as in those dialects that have lost stem II, see section 4 above), stem III alternation corresponds to stem II alternation in the other three dialects:

CAUSATIVE-INCHOATIVE ALTERNATION IN NORTH-EASTERN NEO-ARAMAIC

221

(20) Stem II, J. Urmi a. Inchoative pǝsra meat(M)

mǝgdǝl-le. freeze.PST-ERG.3SG.M

‘The meat froze’. b. Causative tǝmmal yesterday

pǝsra meat(M)

mǝgdǝl-li-le. freeze.PST-ERG.1SG-OBJ.3SG.M

‘Yesterday I froze some meat’. A large number of labile stem III verbs are weak verbs. This is of course partly due to the fact that an estimated half of NENA verbs are weak. However, another reason may be that weak inchoative stem I verbs tend to be ‘upgraded’ to stem III (which has an m-prefix) because of their lack of phonological body. This would explain the shift from stem I/III causative alternation to stem III labile alternation that seems to have occurred in some verbs, thereby decreasing phonological ambiguity and morpho-syntactic complexity while increasing semantic ambiguity. Similarly, causative alternants of stem I alternation verbs seem to migrate to stem III, which may be observed in a number of verbs that appear in the process of moving from stem I alternation to stem I/III alternation, thereby decreasing semantic and phonological ambiguity.27 In a further step, a few Assyrian Kirkuk verbs that may originally have been used in stem III alternation (which still appears to be the case in C. Barwar in these verbs) are now used in the causative stem III/šwq type of alternation, in which stem III is used only for the intransitive alternant, whereas the causative is expressed by the auxiliary šwq (as discussed in section 5). While it is difficult to estimate how many (or few) verbs are affected by this shift, this seems to be a recent development in Assyrian Kirkuk.28 Thus, there seems to be a tendency in this dialect to resolve the ambiguity that characterizes labile alternation by marking the causative alternant

27

These verbs, too, might eventually become stem III labile verbs, if the inchoative follows the causative alternant to stem III. 28 However, I have found one verb in J. Sanandaj (whose intransitive alternant is not inchoative) that seems to show stem III/šwq alternation, and there might be other dialects that have it as well.

222

KATHRIN GÖRANSSON

with šwq, thereby employing the same ancient device (causative alternation) in a new way. Finally, labile stem II verbs are found in Assyrian Kirkuk, C. Barwar and J. Sanandaj:29 (21) Stem II, J. Sanandaj (Khan 2009) a. Inchoative šəm-ef name(M)-his

maḥe! erase[ABS.3SG.M]

‘May his name get erased!’ b. Causative šəm-ef name(M)-his

maḥe-le. erase[ABS.3SG.M]-OBJ.3SG.M

‘He will erase his name’. Some labile stem II verbs might be derived from adjectives or nouns, borrowed from other languages (mainly Arabic) or be derived from former t-stem verbs, like other non-labile stem II verbs (cf. Khan 2008b: 162, 263). However, only extremely few stem II verbs are clearly labile, and a few more are possibly labile. Among my data, not a single verb is attested in stem II alternation in all or even in two of the three dialects (excluding J. Urmi, which lacks stem II). The main reason why stem II alternation is so rare may be that stem II is neither a prototypical inchoative nor causative alternant: many stem II verbs have historically been relatively transitive with agentive subjects30 (apart from those that go back to t-stems) and therefore did not typically have inchoative alternants, which are low in transitivity. The causative alternants, on the other hand, have throughout the attested history of Aramaic been expressed by stem III, which, as mentioned above, is still the most productive way of deriving causatives in NENA today. Therefore, both alternants—and not only one, as in stem I and stem III—would somehow have needed to end up in stem II. However, judging from the evidence, this happened only very rarely. Moreover, compared to stem I and stem III, stem II often has a more specific, i.e. less widely used, and idiosyncratic meaning, which may exclude an inchoative or causative alternant. 29

As already mentioned, J. Urmi lost that stem. J. Urmi stem II alternation seems to correspond to stem III alternation in other dialects. 30 For relative degrees of transitivity see Hopper and Thompson (1980).

CAUSATIVE-INCHOATIVE ALTERNATION IN NORTH-EASTERN NEO-ARAMAIC

223

To sum up, stem I, III and Q alternations are attested in all four dialects, whereas stem II alternation occurs sporadically in all dialects except for J. Urmi, which has lost the equivalent of stem II. Labile alternation is therefore a common means of expressing causative-inchoative alternation in the dialects surveyed, and stem I and Q exhibit this alternation type more frequently than stem III and II.

5.2. Suppletive alternation So far I have encountered only one instance of suppletive alternation: in all four dialects, the causative-inchoative alternation pair ‘die/kill’ is expressed by the stem I verbs myl31 and qṭl32. In Assyrian Kirkuk, C. Barwar and J. Urmi a stem III form of myl exists. However, in Assyrian Kirkuk it is used in the sense of ‘pretend to be dead’ and ‘kill’ is expressed by qṭl. In C. Barwar, too, stem III of myθ seems to be used mainly or even exclusively in the sense of ‘pretend to be dead’, whereas in the many examples involving killing that I found in the text corpus only qṭl was used. Similarly, my J. Urmi informant used +qtl to express the causative ‘kill’ right after having given an intransitive example with stem I of myl. Regarding J. Sanandaj, I am not aware of a stem III of myl; qṭl is used to denote the causative alternant for the inchoative stem I myl also in this dialect. The myl/qṭl ‘die/kill’ alternation is therefore clearly suppletive in all four dialects. It is probably no coincidence that this verb pair also occurs in suppletive alternation in many other languages (cf. Haspelmath 1993). For some reason, 16 out of 21 languages in Haspelmath’s sample have suppletive forms for ‘die’ and ‘kill’, which therefore generally seem to be perceived as having quite different meanings. Suppletive alternation is thus certainly not a common way in which causativeinchoative alternation manifests itself in NENA. This may be a universal tendency. In Haspelmath’s study, only very few alternation pairs occur in the form of suppletive alternation, and there is no language in his sample that has more than 3 suppletive verb pairs (out of a total of 31).

6. THE TRANSFORMATION FROM ANTICAUSATIVE INTO LABILE ALTERNATION VERBS

In NENA, anticausative alternation is not in use anymore. In earlier Aramaic, however, this alternation type was common:

31

In J. Urmi and J. Sanandaj; myt in Assyrian Kirkuk and myθ in C. Barwar. qtl in J. Urmi.

32 +

224

KATHRIN GÖRANSSON

(22) Jewish Babylonian Aramaic (Sokoloff 2002: 946) a. Inchoative, Gt-stem ʾtptx ‘open (intr.), be opened’ (lost in NENA) ytptxwn open.FUT.3PL

ly šʿry ḥkmh for.1SG gates wisdom

w-ʾhgy and-meditate.1SG

bhw in.3PL

‘The gates of wisdom will open for me and I shall meditate in them’. b. Causative, G-stem ptx ‘open (tr.)’ (corresponds to NENA stem I) ptx open.PST[3SG.M]

yʿqb Jacob

ʿyny-h eyes-his

w-ʾxyk and-laugh.PST[3SG.M]

‘Jacob opened his eyes and laughed’. In the course of time the anticausative marker (the t-infix33) was lost and some, and possibly a large proportion, of former anticausative alternation pairs seem to have come to be expressed by labile alternation in NENA:34 (23) Stem I, Assyrian Kirkuk a. Inchoative pǝnjara window(F)

ptǝx-la. open.PST-ERG.3SG.F

‘The window opened’. b. Causative hawa wind(M)

ptix-a-le open.PST-ABS.3SG.F-ERG.3SG.M

pǝnjara. window(F)

‘The wind opened the window’. This mostly seems to have occurred in stem I alternation, probably because either the -t- of the anticausative verb assimilated, or the anticausative merged with stem I for some other reason. However, the history of t-stem verbs (and especially of the Dt- and the Ct-stem) needs to be investigated further. A major problem is that verbs are attested in multiple stems and in different varieties of pre-modern Aramaic, none of which is the precursor of NENA. This makes it difficult to identify the older alternation type(s) of a verb. The situation is further complicated by the fact 33 34

Or historical prefix, depending on which theory one adheres to. This finding is still tentative.

CAUSATIVE-INCHOATIVE ALTERNATION IN NORTH-EASTERN NEO-ARAMAIC

225

that the extant material of early forms of Aramaic is limited and in some parts quite scarce.

7. CAUSATIVE-INCHOATIVE ALTERNATION WITHIN THE VERBAL SYSTEM

A few Assyrian Kirkuk verbs will be used to illustrate how the most common alternation types (stem I/III, I/II, I, III and Q alternation) interact with other verbal stems and with periphrastic constructions, including those employing the auxiliary verbs šwq and pyš. While the following examples are expected to be relatively representative, it goes without saying that there are other verbs that diverge from the examples given. Furthermore, since there are indications that some verbs have switched alternation types in the past or are in the process of doing so, more variation exists that cannot be presented here.

7.1. Stem I/III alternation qyada/maqude ‘burn’ Stem I

burn (intr.)

inchoative

Stem III

burn (tr.)/cause to get burnt causative

šwq + stem I

allow/cause to burn (intr.)

permissive/indirect causative

šwq + stem III allow/cause to burn (tr.)

permissive/indirect causative

pyš + stem III

passive

be burnt

It is even possible (though not very common) to combine šwq and pyš: (24) Periphrastic causative, šwq + pyš + stem III b-šoq-ǝn-ne FUT-cause-ABS.1SG-OBJ.3SG.M muqd-a burn.RES-3SG.M

naxir-e nose-his

d-payǝš REL-become[3SG.M]

(b-ide d-gane). by-hand-his REL-REFL

‘I will cause his nose to get burnt (by his own hand)’. The šwq + stem I construction may best be translated as ‘allow/cause to get burnt’. It should be noted that the meaning of this construction is inchoative rather than passive, since it is impossible to add an agent phrase. This is in contrast to the causative-passive meaning of stem III, where the expression of the agent is allowed:

226

KATHRIN GÖRANSSON

(25) a. Indirect causative, šwq + stem I šwǝq-li cause.PST-ERG.1SG

d-qeda REL-burn[ABS.3SG.F]

id-ux hand(F)-your.SG.M

‘I caused your hand to get burnt (*by someone35)’. b. Causative, Stem III muqd-a-li burn.PST-ABS.3SG.F-OBJ.1SG

id-ux (b-brun-i). hand(F)-your.SG.M by-son-my

‘I caused your hand to get burnt (by my son)’. The crucial difference between the indirect causative šwq + stem I construction and the causative of stem III with a passive sense is, therefore, that in šwq + stem I the stem I form has only inchoative (and not passive) meaning, whereas the causative stem III may, in addition to its active causative meaning ‘burn (tr.)’, be conceived of as a causative with a passive sense, meaning ‘cause to get burnt (by someone or something)’.

7.2. Stem I/II alternation bdara/badure ‘scatter; spill’ Stem I

scatter (intr.)

inchoative

Stem II

scatter (tr.)

causative

Stem III36

—

—

šwq + stem I allow/cause to scatter (intr.)

permissive/indirect causative

šwq + stem II allow/cause to scatter (tr.)

permissive/indirect causative

pyš + stem II be scattered

passive

For some verbs two different expression types are attested, which may at least partly be attributed to the legacy of the pre-modern Aramaic verbal system. For in35

Note, however, that expressing a non-agentive cause is allowed: b-šoqǝn qeda idux b-nura. ‘I will cause your hand to get burnt by fire’. 36 Stem III is attested in some stem I/II alternation verbs. With some verbs (e.g. pšr ‘melt’) stem III has the same or a similar meaning as stem II, whereas with others (e.g. tlx ‘collapse; destroy’) it functions as a causative of stem II.

CAUSATIVE-INCHOATIVE ALTERNATION IN NORTH-EASTERN NEO-ARAMAIC

227

stance, the reason that ṭpy ‘kindle’ occurs in stem I/II and I/III alternation in Assyrian Kirkuk and C. Barwar might be found in the verb’s history: stem I/II alternation of this verb has possibly developed from the Dt/D-stem alternation of earlier forms of Aramaic, with the D-stem corresponding to the modern stem II and the Dt-stem merging with the G-stem into what became stem I. Similarly and somewhat more straightforwardly, stem I/III alternation might go back to the G/C-stems of earlier Aramaic. A slightly different example of a verb that seems to have recently switched from one alternation type to another is the verb tlx ‘collapse; destroy’. It is used in stem I/II alternation in Assyrian Kirkuk but shows traces of labile stem I alternation. Interestingly, the causative of this verb is expressed by both stem I and stem II also in Barwar (Khan 2008b: 259). Given that usually only one option is available for each alternant, this indicates that tlx is moving either towards stem I or stem I/II alternation.

7.3. Stem I alternation ptaxa ‘open’ Stem I

open (intr.) open (tr.)

inchoative causative

Stem III

cause to open (intr.) cause to be opened

indirect causative causative

šwq + stem I allow/cause to open (intr.) allow/cause to open (tr.)

permissive/indirect causative permissive/indirect causative

pyš + stem II be opened

passive

Similarly to stem I/III alternation above, the stem III form of stem I alternation verbs has two functions; to convey indirect causative meaning as well as a causative with a passive sense. Stem I in the šwq construction can be used either as an inchoative or a causative: (26) Stem I a. Inchoative šwǝq-le cause.PST-ERG.3SG.M

d-patx-a REL-open-ABS.3SG.F

‘He allowed/caused the window to open’.

pǝnjara. window(F)

228

KATHRIN GÖRANSSON b. Causative šwǝq-le cause.PST-ERG.3SG.M

brun-e son-his

d-patǝx-la pǝnjara. REL-open[ABS.3SG.M]-OBJ.3SG.F window(F)

‘He caused his son to open the window’. Apart from the (optional) use of a noun or a pronominal phrase denoting the direct object, the inchoative and causative alternants can be distinguished based on differences in agreement. The causative form of ptx carries a direct object suffix denoting the patient (the patientive causee), which is expressed by an absolutive suffix in the inchoative. The agentive causee, in turn, which is lacking in the clause with the inchoative form, is marked as subject in the causative form.37

7.4. Stem Q alternation šaxlupe ‘change’ Stem Q

change (intr.) change (tr.)

inchoative causative

šwq + stem Q allow/cause to change (intr.) permissive/indirect causative allow/cause to change (tr.) permissive/indirect causative pyš + Q

be changed

passive

In Assyrian Kirkuk stem Q alternation is practically always labile, i.e. stem Q verbs virtually never take a m-prefix. Since stem Q alternation has only one form at its disposal, the šwq construction, too, is, like in stem I alternation, labile .

7.5. Stem III alternation mamyule ‘become blue; bruise’ Stem III

become blue bruise

inchoative causative

šwq + stem III

allow/cause to bruise

permissive/indirect causative

pyš + stem III

be bruised

passive

37

For a discussion of the syntax of causative and inchoative verbs in NENA compared to that in other languages, see Khan (forthcoming).

CAUSATIVE-INCHOATIVE ALTERNATION IN NORTH-EASTERN NEO-ARAMAIC

229

Similarly to the situation in stem I alternation verbs, šwq + stem III can be used to express the permissive and the indirect causative of both the inchoative and the causative alternant. This verb is somewhat special in that the šwq construction is not used to express ‘allow/cause to become blue’. However, with other stem III verbs, such as gdl ‘freeze’, both meanings occur: (27) šwq + stem III a. Indirect causative šwǝq-li cause.PST-ERG.1SG

d-magdǝl-la-le REL-freeze-ERG.3SG.F-OBJ.3SG.M

bǝsra meat(M)

‘I caused/got her to freeze the meat’. b. Permissive šwǝq-li cause.PST-ERG.1SG

d-magdǝl-la REL-freeze-ERG.3SG.F

‘I let her freeze (to death)’.

8. CONCLUSION

Causative-inchoative alternation is a type of transitivity alternation that is attested in NENA as well as in many other languages. According to Haspelmath (1993), there are three types of causative-inchoative alternation, based on the way in which the inchoative and/or causative alternants are marked: causative alternation, anticausative alternation and non-directed alternation. The latter can be subdivided into labile, equipollent and suppletive alternation. NENA mainly has causative and labile alternation. Causative alternation is productive, whereas at least part of the verbs undergoing labile alternation appear to have developed historically from an original anticausative alternation, which was common in pre-modern Aramaic. Due to the profound transformation of the verbal system in Aramaic, anticausative alternation has been lost. As a result of the old anticausative stems merging into the modern stem I, labile alternation evolved (or became at least greatly enhanced). NENA stem I alternation thus seems to have replaced the anticausative alternation of earlier Aramaic. The other types of labile alternation seem to have developed for other reasons. Some or even most stem Q and (the very rare) stem II alternation verbs have emerged from denominatives. Moreover, in stem Q it may have been impossible for phonological reasons to use stem III to express the causative alternant. Stem III alternation appears to have resulted from inchoative (and especially from weak) verbs ending up in that stem, combined with—as may have been the case with stem Q—

230

KATHRIN GÖRANSSON

the lack of a suitable causative marker. Nevertheless, it has been shown that in the Assyrian Kirkuk dialect in my corpus stem III causatives may be distinguished from their inchoative counterparts by employing the auxiliary verb šwq, which is otherwise used mainly to express indirect causative and permissive meaning. Causative alternation, on the other hand, has not changed much since premodern Aramaic times. As in labile alternation there are several grammatical means of expressing causative alternation in NENA (stem I/III, I/II and III/šwq alternations), but stem I/III alternation is by far the most productive and the most common. In addition to causative-inchoative alternation, stem I/III alternation is also used to causativise intransitive as well as transitive verbs.

REFERENCES

Comrie, Bernard. 2006. “Transitivity Pairs, Markedness, and Diachronic Stability.” Linguistics 44 (2): 303–318 (Special issue Operations on Argument Structure, edited by Daniel Hole and Peter Siemund). Berlin: Mouton de Gruyter. Fassberg, Steven Ellis. 2010. The Jewish Neo-Aramaic Dialect of Challa. Studies in Semitic Languages and Linguistics 54. Leiden: Brill. Fox, Samuel. 2009. The Neo-Aramaic Dialect of Bohtan. Neo-Aramaic studies 9. Piscataway, NJ: Gorgias Press. Greenblatt, Jared R. 2011. The Jewish Neo-Aramaic Dialect of Amadiya. Studies in Semitic Languages and Linguistics 61. Leiden: Brill. Haspelmath, Martin, Calude, Andreea, Spagnol, Michael, Narrog, Heiko, and Bamyacı, Elif. 2014. “Coding Causal-Noncausal Verb Alternations: A Form-Frequency Correspondence Explanation.” Journal of Linguistics, 1–39. Haspelmath, Martin. 1993. “More on the Typology of Inchoative/Causative Verb Alternations.” In Causatives and Transitivity, edited by Bernard Comrie and Maria Polinsky, 87–120. Amsterdam: John Benjamins. Hopper, Paul J., and Thompson, Sandra E. 1980. “Transitivity in Grammar and Discourse.” Language 56: 251–99. Kemmer, Suzanne. 1993. The Middle Voice. Typological Studies in Language 23. Amsterdam: John Benjamins. Khan, Geoffrey, and Doron, Edit. 2012. “The Typology of Morphological Ergativity in Neo-Aramaic.” Lingua 122: 225–240. Khan, Geoffrey. 2002. The Neo-Aramaic Dialect of Qaraqosh. Studies in Semitic Languages and Linguistics 36. Boston, MA: Brill. —. 2008a. The Jewish Neo-Aramaic Dialect of Urmi. Gorgias Neo-Aramaic Studies 2. Piscataway, NJ: Gorgias Press. —. 2008b. The Neo-Aramaic Dialect of Barwar. Handbook of Oriental Studies 96. Leiden: Brill.

CAUSATIVE-INCHOATIVE ALTERNATION IN NORTH-EASTERN NEO-ARAMAIC

231

—. 2009. The Jewish Neo-Aramaic Dialect of Sanandaj. Neo-Aramaic studies 10. Piscataway, NJ: Gorgias Press. —. forthcoming. “Causative constructions in Neo-Aramaic (Christian Urmi Dialect).” In Arabic and Semitic Linguistics Contextualized: a Festschrift for Jan Retsö, edited by Lutz Edzard. Wiesbaden: Harrassowitz. Kulikov, Leonid. 2001. “Causatives.” In Language Typology and Language Universals. An International Handbook, edited by Martin Haspelmath et al., 886–898. Vol. 2. Berlin: De Gruyter. —. 2011. “Voice Typology.” In The Oxford Handbook of Linguistic Typology, edited by Jae J. Song, 368–398. Oxford: Oxford University Press. Mutzafi, Hezy. 2004. The Jewish Neo-Aramaic Dialect of Koy Sanjaq (Iraqi Kurdistan). Semitica Viva 32. Wiesbaden: Harrassowitz. Nedjalkov, Vladimir P. 1969. “Nekotorye verojatnostnye universalii v glagol’nom slovoobrazovanii.” In Jazykovye universalii i lingvisticheskaja tipologija, edited by I.F. Vardul, 106–114. Moscow: Nauka. Nichols, Johanna, Peterson, David A., and Barnes, Jonathan. 2004. “Transitivizing and Detransitivizing Languages.” Linguistic Typology 8 (2): 149–211. Shibatani, Masayoshi, and Pardeshi, Prashant. 2002. “The Causative Continuum.” Typological Studies in Language 48: 85–126. Sokoloff, Michael. 2002. Dictionary of Jewish Babylonian Aramaic of the Talmudic and Geonic periods. Ramat-Gan, Israel: Bar Ilan University Press; Baltimore and London: Johns Hopkins University Press. Talay, Shabo. 2008. Die Neuaramäischen Dialekte der Khabur-Assyrer in Nordostsyrien: Einführung, Phonologie und Morphologie. Semitica Viva 40. Wiesbaden: Harrassowitz. Younansardaroud, Helen. 2001. Der Neuostaramäische Dialekt von Särdä:rïd. Semitica Viva 26. Wiesbaden: Harrassowitz.

SPLIT ERGATIVITY IN THE NENA DIALECTS ALESSANDRA BAROTTO Different types of split ergativity are attested in North-Eastern Neo-Aramaic (NENA) dialects, a group of over a hundred varieties spoken by Jewish and Christian communities in south-eastern Turkey, northern Iraq and north-western Iran. Although historically Aramaic is nominative-accusative, some ergative phenomena developed early through the influence of Iranian languages, especially Kurdish, which are ergative or have been ergative at some stage of their history. Nevertheless, the NENA dialects show some interesting autonomous developments, such as the extension of the ergative suffix to all intransitive verbs. The different types of split ergativity exhibited by the NENA dialects concern the expression of tense and aspect of the verbal phrase (ergativity developed only in the perfective aspect), the gradual extension of the ergative suffix to intransitive verbs, the expression of the pronominal object and the creation of new accusative markers and, finally, the expression of the pronominal subject. All these split-ergativity phenomena create a rather complex alignment system which can provide useful insights for future linguistic studies on ergativity.

1. TYPES OF SPLIT ERGATIVITY IN THE NENA DIALECTS

According to the collected data,1 the NENA dialects show four types of split ergativity, which affect the verbal system on several levels.

1

The dialects I have compared for the present studies are the Jewish dialect of Sanandaj (Khan 2009), Sulemaniyya and Ḥalabja (Khan 2004), Urmi (Khan 2008b), Challa (Fassberg 2010), Betanure (Mutzafi 2008), Amadiya (Hoberman 1989), Arbel (Khan 1999), and the Christian dialects of Barwar (Khan 2008a), Hertevin (Jastrow 1988), Qaraqosh (Khan 2002), Bohtan (Fox 2009) and Koy Sanjaq (Mutzafi 2004).

SPLIT ERGATIVITY IN THE NENA DIALECTS

233

The first type concerns the expression of tense and aspect of the verbal phrase. While the present base of verbs keeps the classical Aramaic accusative alignment, ergativity developed in the perfective aspect. Because of its uniformity and lack of significant variation, this type is typologically uninteresting and it will not be discussed any further. The second type concerns the extension of the ergative suffix to intransitive verbs. In NENA dialects, some intransitive verbs behave like the transitive ones, taking the ergative marking. This type of split ergativity is not uniform. Dialects exhibit varying degrees of ergativity because of the differences in the distribution of the ergative suffix among intransitive verbs. We can divide the NENA dialects into three types according to this parameter, following the categorization made by Doron and Khan (2012: 225). We call Split-S those dialects in which the ergative marker is extended to transitive and unergative verbs. (1) Jewish Sanandaj (Doron and Khan 2012: 230) Transitive: barux-awal-i friend-PL-my

brat-i daughter-my

gərš-a-lu pull.PRF-ABS.3SG.F-ERG.3PL

‘My friends pulled my daughter’. Intransitive unergative: kalba dog

nwəx-le bark.PRF-ERG.3SG.M

‘The dog barked’. Intransitive unaccusative: brat-i qim-a daughter-my rise.PRF-ABS.3SG.F ‘My daughter rose’. In the second type, the Dynamic-Stative dialects, the ergative marker is extended to all intransitive verbs and the absolutive marking of unaccusative verbs survives as a resultative stative. (2) Jewish Urmi (Doron and Khan 2012: 233) Transitive (perfective): baruxaw-i friend.PL-my +

brat-i daughter-my

‘My friends pulled my daughter’.

gərš-a-lu pull.PRF-ABS.3SG.F-ERG.3PL

234

ALESSANDRA BARPTTO Intransitive unergative (perfective): ayné they

rqǝ̀l-lu dance.PRF-ERG.3PL

‘They danced’. Intransitive unaccusative (perfective): brat-i qəm-la daughter-my rise.PRF-ERG.3SG.F ‘My daughter rose’. Intransitive unaccusative (resultative stative): brat-i qim-a daughter-my rise.PRF-NOM.3SG.F ‘My daughter has risen’. Finally, we call Extended Ergative those dialects in which an ergative marker is extended to all intransitive verbs and the absolutive marker is no longer used. Although the marking system in these dialects appears more uniform (all intransitive verbs take the same marking), their theoretical status is ambiguous. Even if the markedness proportions are those of a classical ergative system (with the ergative “case” more marked than the absolutive one), the case alignment follows the nominative system, where S (the subject of intransitive verbs) shares the same morphological coding of A (the agent of transitive verbs). (3) Christian Barwar (Doron and Khan 2012: 231) Transitive: xawr-awaθ-i friend-PL-my

brat-i daughter-my

‘My friends pulled my daughter’. Intransitive unergative: kalba dog ‘The dog barked’.

nwix-le bark.PRF-ERG.3SG.M

griš-a-la pull.PRF-ABS.3SG.F-ERG.3PL

SPLIT ERGATIVITY IN THE NENA DIALECTS

235

Intransitive unaccusative: brat-i daughter-my

qim-la rise.PRF-ERG.3SG.F

‘My daughter rose’. This alignment type—which is apparently rare across languages, but rather common or, we could even say dominant in NENA—is labeled by Dixon either as “marked nominative’ or “extended ergative”.2 The linguistic status of this alignment is problematic and, in this regard, the NENA dialects exhibit peculiarities which are not found elsewhere. This topic is worthy of being investigated further from the point of view of linguistic typology. However, it would fall beyond the scope of this paper.

The third type of split ergativity is characterized by the use of new accusative markers. These accusative markers do not replace radically the absolutive ones to cross-reference the object, but became the only acceptable way to express the 1st and 2nd person objects in some dialects. The absolutive marking of the 3 rd person object is still rarely used in certain constructions, however, the accusative marking is preferred and used more frequently. There is a fourth type of split ergativity which is confined to a small group of NENA dialects, the so-called “Hertevin paradigm”. It concerns the expression of the pronominal subject. The oblique series of preterite endings in the 1 st and 2nd person used to cross-reference the subject is replaced by the direct endings preceded by lwhenever the pronominal object is explicit. This change does not take place with 3 rd person subject pronouns. These last two types of split ergativity will be the central focus of this paper, especially as regards the possible correlation between different degrees of ergativity and strategies to mark pronominal objects.

2. SPLIT ERGATIVITY CONDITIONED BY THE SEMANTIC NATURE OF THE OBJECT IN THE NENA DIALECTS

According to the data, in most cases, ergative alignment occurs only with 3rd person object pronouns. With 1st and 2nd person objects, dialects must pick different strategies, which cause a reanalysis of the entire alignment system.

2

See Dixon (1994: 63–69).

236

ALESSANDRA BARPTTO

(4) Jewish Sanandaj (Khan 2009) barux-awal-i friend-PL-my

bron-i son-my

gərš-∅-lu pull.PRF-ABS.3SG.-ERG.3PL

‘My friends pulled my son’. barux-awal-i friend-PL-my

brat-i daughter-my

gərš-a-lu pull.PRF-ABS.3SG.F-ERG.3PL

‘My friends pulled my daughter’. barux-awal-i friend-PL-my

bron-awal-i son-PL-my

gərš-i-lu pull.PRF-ABS.3PL-ERG.3PL

‘My friends pulled my sons’. Still, in a small group of dialects, incorporated pronominal objects (objects cross-referenced by absolutive marking, also known as S-suffixes) are allowed for all persons. These dialects are the only NENA examples in which we find a perfect ergative alignment, at least in the past transitive paradigm with pronominal objects. An example is provided by the Jewish dialect of Challa: n-š-q ‘to kiss’ (Fassberg 2010: 124) nšiqətte (nšiq+ ət + le)

‘he kissed you (sg.m.)’

nšiqatte (nšiq+ at + le)

‘he kissed you (sg.f.)’

nšiqétūle (nšiq+ etun + le)

‘he kissed you (pl.)’

nšiqənne (nšiq+ ən + le)

‘he kissed me (m.)’

nšiqanne (nšiq+ an + le)

‘he kissed me (f.)’

nšiqaxle (nšiq+ ax + le)

‘he kissed us’

The occurrence of this rare ergative paradigm in conservative dialects provides evidence that the restriction of incorporated objects to the 3rd person is a further step towards the loss of ergativity. In other words, we can imagine a historical development, in which the NENA dialects gradually stopped using the S-suffixes to cross-reference the object for all person, and, at the same time, introduced new accusative markers to recreate a nominative-accusative alignment. There is evidence that the NENA dialects are developing towards this process of decay of ergativity and not the other way around. First of all, thanks to texts and manuscripts, we know that until the 17th century incorporated object was a common strategy to express the direct object in Jewish and Christian Neo-Aramaic literary

SPLIT ERGATIVITY IN THE NENA DIALECTS

237

texts, and it was available for all persons. The manuscript transmission of Christian texts shows that since the 19th century, incorporated objects have been confined to the 3rd person. Secondly, even in dialects where intra-conjugational representation of objects is attested, although confined to the 3rd person, it is never the only strategy available and it is rarely used. Finally, the 1st and 2nd person incorporated objects are attested only in conservative and peripheral dialects, where, probably, innovations struggle to settle.3 Considering these pieces of evidence, the loss of incorporated objects can be considered one of the NENA “antidotes” to ergativity. The intra-conjugational representation of objects confined to the 3rd person is an intermediate stage: the S–suffixes do not cease immediately to cross-reference the patient in the preterite, but they are confined to the 3rd person. This causes a split in the verbal system. While the 3rd person maintains the ergative-absolutive alignment, the 1st and 2nd persons take accusative markers, creating a nominative-accusative system.

Ergativeabsolutive alignment Intra-conjugational representation of pronominal objects for all persons by absolutive S-suffixes (i.e. Jewish dialect of Challa, Zahko, Amadiya…)

Decay >

> Intra-conjugational representation of pronominal objects by absolutive Ssuffixes confined to the 3rd person

Nominativeaccusative alignment

All pronominal objects are represented by accusative suffixes

This system follows what Dixon labels as “split ergativity conditioned by the semantic nature of the SN”.4 In this particular split, if pronouns and nouns have different systems of case inflection, noun phrases which are lowest in the animacy hierarchy will follow the ergative system, while the ones which are higher will follow the accusative system. The NENA examples are partially different from those analyzed by Dixon, but they reflect the same theoretical assumption. In the NENA examples, this phenomenon does not concern the agent, but only the expression of pronominal objects. The pronominal subjects do not change according to the animacy hierarchy, but they are indeed affected by the consequences of split ergativity. In fact, even if transitive sub3 4

See Pennacchietti (1994: 264–280). See Dixon (1994: 83–97).

238

ALESSANDRA BARPTTO

jects are always marked by L-suffixes, the markedness proportion changes according to what marker is used to cross-referenced the object. An interesting peculiarity is the status of the 3rd person singular masculine suffix. There are some dialects (i.e. Urmi, Hertevin, Arbel and Qaraqosh), which treat the 3rd person singular masculine suffix as the 1st and 2nd person suffixes. In other words, they cannot be incorporated into the past base by S-suffixes. We have to be careful in dealing with these verbal forms in order to identify this phenomenon. Since the 3rd person singular masculine suffix is ∅, we can assume that forms like + qtəlle (in the Jewish dialect of Urmi, +qtəl + ∅ + le) express the 3rd person singular masculine direct object and translate ‘he killed him’.5 For some dialects this is correct, for others such a form does not express an object (and thus they must be interpreted ‘he killed’). (5) Christian Barwar (Khan 2008a) qṭíl-∅-le kill.PRF-ABS.3SG.M-ERG.3SG.M ‘He killed him’. (rare) qəm-qaṭəl-∅-le PAST-kill-NOM.3SG.M-ACC.3SG.M ‘He killed him’. (most common) Christian Qaraqosh (Khan 2002) kəm-naqəš-∅-lə PAST-kill-NOM.3SG.M-ACC.3SG.M ‘He hit him’. (only way to express the object) The different treatment of the 3 rd person singular masculine suffix shows some consequences also regarding the development from an ergative system to an accusative one. If there is an intermediate step between the intra-conjugational representation of objects confined to the 3rd person and the complete loss of this strategy, this stage will involve the different treatment of the 3rd person singular masculine. The different treatment of the singular and plural is certainly less problematic: the implicit definiteness of a single entity provides more animacy, compared to a group of participants.

5

See Khan (2008b: 139).

SPLIT ERGATIVITY IN THE NENA DIALECTS

239

According to this, one might expect a different treatment of 3rd person singular (masculine and feminine) and plural suffixes. However, the split operates also between the two singular pronouns (masculine vs. feminine). The explanation for this splitting may be provided in terms of markedness: generally speaking, if a language has genders, the feminine (as well as the plural) will be morphologically more marked than the masculine. In this case we cannot speak of animacy, but there are some linguistic theoretical parameters which explain why the masculine can be put on a different level, compared to the feminine and the plural. Consequently, they explain also the differences in syntactical treatment.

3. ALTERNATIVE STRATEGIES TO THE INTRA-CONJUGATIONAL REPRESENTATION OF OBJECTS

The NENA dialects have three alternative strategies to intra-conjugational representation of objects. These strategies show various degrees of pervasiveness and distribution among the dialects. The first one involves the creation of a new accusative marker. The object is cross-referenced by a preposition such as ʾill- with pronominal suffixes. The second one involves a complete reshaping of the past tense paradigm. The past base of the verb is replaced by the present base, prefixed by the preverb which has the form kəm- or slight variants thereof. The case inflection follows the nominative-accusative alignment found in the present base (S-suffixes cross-reference the subject, while L-suffixes cross-reference the object). The third involves the use of L-suffixes as a dedicated accusative marker, symmetrical to the present base. It is a peculiar alternative because it creates a system in which, in the preterite, subject and object are cross-referenced by the same marking. The distribution of these strategies varies among the NENA dialects. The use of a new accusative prepositional marker such as ʾill would appear to be the most common scheme in the dialects I have compared for the present study. The accusative marker ʾill is commonly used in Sulemaniyya, Sanandaj, Koy Sanjaq, Arbel, Barwar and Challa. It does not occur in the dialects of Urmi, Bohtan, Hertevin, Qaraqosh, Betanure and Amadiya. In the Extended Ergative dialects such as the Christian dialects of Barwar and Qaraqosh, and in the Jewish dialects of Betanure, Amadiya and Zahko, the strategy with a preverbal particle (kəm- or one of its variants) seems to be dominant. Compared to the kəm- strategy, the use of ʾill is less invasive. The past base is maintained, including the traces of ergativity: the use of L-suffixes to cross-reference the agent and, in some circumstances, the use of S-suffixes to mark intransitive subjects.

240

ALESSANDRA BARPTTO

Although the usage of ʾill- does not eliminate the ergative construction completely, it creates new markedness proportions between the markers used to crossreference the subject and the object. The latter is more marked than the former: a type of proportion which is much closer to the one found in a canonical nominativeaccusative system. Interestingly, in a dialect like Jewish Arbel, the preposition ʾill is used, however rarely, as an accusative marker also with the present base: (6) Jewish Arbel (Khan 1999: 118) lā-daʿ-èt PROG-revile-NOM.2SG.M

ʾillú ACC.3PL

‘You are reviling them’. The second strategy, i.e. the kəm-prefixed paradigm, appears in the Extended Ergative dialects: Christian Barwar and Qaraqosh and in the Jewish dialects of Betanure, Amadiya and Zahko. This correlation between the Extended Ergative dialects and the preterite with kəm- leads to the assumption that, enlarging the sample, the number of dialects using the kəm- strategy would increase as well. Pennacchietti (1994: 269) described the geographical distribution of the kəmprefixed preterite. He notes that wherever the kəm- model is the dominant strategy, it excludes the use of L-suffixes as accusative marker on past base verbs and allows only the optional use of incorporated objects in the past base of verbs. Compared to ʾill, the preterite with kəm- is a much more pervasive strategy. It eliminates completely the past base and all traces of ergativity. The present base used with kəm- preserves its nominative-accusative alignment. (7) Christian Qaraqosh (Khan 2002) Present tense: k-naqəš-∅-la IND-hit-NOM.3SG.M-ACC.3SG.F ‘He hit her’. Past tense: kəm-naqəš-∅-la PAST-hit-NOM.3SG.M-ACC.3SG.F ‘He hit her’. However, the preterite with kəm- does not cause a complete reshaping of the past tense paradigm, since it is used only when a pronominal object is explicitly ex-

SPLIT ERGATIVITY IN THE NENA DIALECTS

241

pressed. If the object is not explicitly expressed, the verb keeps its past base with the ergative marking (L-suffixes) to cross-reference the subject. (8) Christian Qaraqosh (Khan 2002) Without an explicit object: nqəš-lə hit.PRF-ERG.3SG.M ‘He hit’. With an explicit object: kəm-naqəš-∅-la PAST-hit.IMPRF-NOM.3SG.M-ACC.3SG.F ‘He hit her’. If the preterite with kəm- was to become the only way to express the preterite regardless of the explicit expression of the object, ergativity—even extended ergativity—would disappear. This development does not occur in any dialect. Nevertheless, there are some isolated examples where kəm- is used with the verb ʾ-m-r ‘to say’, followed by an indirect object: (9) Christian Qaraqosh (Khan 2002: 317) kəm-amər-∅ PAST-say.IMPRF-NOM.3SG.M

dà-mm-əḥ to-mother-his

‘He said to his mother’. The construction with ʾill- and the one with kəm- do not seem to be mutually exclusive. In our sample, there is at least one dialect, in which both are used, i.e. the Christian dialect of Barwar. Nevertheless, kəm- seems to be most commonly used. (10) Christian Barwar (Khan 2008a) Intra-conjugational representation of object (S-suffixes): qṭil-á-le kill.PRF-ABS.3SG.F-ERG.3SG.M ‘He killed her’. *ʾill qṭíl-le kill.PRF-ERG.3SG.M

ʾəlla ACC.3SG.F

242

ALESSANDRA BARPTTO

‘He killed her’. Present base prefixed by the preverb *kəm-: qəm-qaṭəl-∅-la PAST-kill.IMPRF-NOM.3SG.M-ACC.3SG.F ‘He killed her’. Although the presence of both constructions in a dialect is rare and uneconomical, it shows that NENA varieties are still experimenting with different strategies to mark objects differently and confine the ergative phenomenon.

4. THE L-SUFFIXES AS DEDICATED ACCUSATIVE MARKER

Curiously, in the Dynamic-Stative dialects (i.e., dialects in which the resultative stative of unaccusative verbs is expressed ergatively; in our sample they are represented by Christian Hertevin and Bohtan, and the Jewish dialect of Urmi) the two strategies described above are rather marginal or, in the case of kəm-, entirely unattested. Doron and Khan (2012: 228) demonstrate the nature of L-suffixes as clitics, in contrast to the inflectional nature of S-suffixes. For example, L-suffixes (as other inflected prepositions) follow the anteriority auxiliary -wa, while S-suffixes precede the auxiliary. Furthermore, two L-suffixes can be combined in the same verb, the first as an ergative marking and the second as an accusative one. However, the majority of NENA dialects do not follow this strategy, developing dedicated accusative markers instead of L-suffixes. In contrast to this trend, most of the Dynamic-Stative dialects use L-suffixes to cross-reference the object also in the ergative preterite. The accusative use of L-suffixes became the only alternative available to mark pronominal objects. (11) Jewish Urmi (Khan 2008b) qtəl-la-le kill.PRF-ERG.3SG.F-ACC.3SG.M

+

‘She killed him’. (12) Christian Bohtan (Fox 2009) ptǝ́x-li-le open.PRF-ERG.1SG-ACC.3SG.M ‘I opened him’.

SPLIT ERGATIVITY IN THE NENA DIALECTS

243

(13) Christian Hertevin (Jastrow 1988) ḥzé-le- la see.PRF-ERG.3SG.M-ACC.3SG.F ‘He saw her’. It should be noted that using L-suffixes to cross-reference the object is the easiest solution in order to create a parallelism with the present base of the verb, and thus an accusative alignment. Since L-suffixes are also used as the ergative marker, this solution is potentially problematic on a communicative level, despite being accessible and simple; this is a likely explanation for its rarity. The second L-suffix should be considered as an accusative marker, used in the past tense by analogy with the present tense: (14) Jewish Urmi (Khan 2008b) xze-le-le see.PRF-ERG.3SG.M-ACC.3SG.M ‘He saw him’. The accusative use of L-suffixes also tends towards a nominative-accusative alignment. However, compared to constructions with ʾill- and kəm-, it is less effective. Apart from using the same marker to cross-reference subject and object, L-suffixes as accusative markers do not change the markedness proportions between subject and object, since such constructions still have a marked ergative subject. All these reasons possibly explain why it is a relatively rare strategy, restricted to few dialects, despite being the easiest solution available. Although the Dynamic-Stative dialects seem to accept the consequences of having two L-suffixes combined on the same verb, there is at least one case, in which an attempt was made to further modify the paradigm to get closer to the present base alignment: the so-called Hertevin paradigm.

5. THE HERTEVIN PARADIGM

The so-called Hertevin paradigm can be seen as a further “antidote” against ergativity, attested for centuries in other NENA dialects, though in very few occurrences, as a marginal form. In the dialect of Hertevin, the oblique series of preterite endings used to crossreference the subject is replaced in the 1 st and 2nd person by the direct endings pre-

244

ALESSANDRA BARPTTO

ceded by l- whenever the pronominal object is explicit. This change does not take place with 3rd person subject pronouns. (15) Christian Hertevin (Jastrow 1988) 3rd person singular masculine subject: ḥze-le see.PRF-ERG.3SG.M ‘He saw’. ḥzé-le-le see.PRF-ERG.3SG.M-ACC.3SG.M ‘He saw him’. 1st person singular subject: ḥze-li see.PRF-ERG.1SG ‘I saw’. ḥzé-len-ne6 see.PRF-ERG.NOM.3SG.M7-ACC.3SG.M ‘I saw him’. As for the expression of the object, the split caused by the Hertevin paradigm can be explained by the animacy hierarchy. The 3rd person shows the higher degree of ergativity: the ergative marker is still used to cross-reference the subjects. By contrast, the 1st and 2nd person seem to shift towards an accusative system, using a hybrid set of suffixes which is halfway between the S-suffixes, the canonical nominative marker, and L-suffixes, the preterite endings that cross-reference the subject. Outside the dialect of Hertevin, this paradigm is no longer used in the verbal systems of the modern spoken NENA dialects. As noted by Pennacchietti (1991) and 6

This is caused by the assimilation of the /l/ of the L-suffixes: ḥzélenne> *ḥzélen + le. It was not an easy task to indicate the Hertevin paradigm in the gloss. Finally, I decided to treat -lén- as a grammatical element that consist of two parts: the /l/ can be seen as the ergative preterite ending which cross-references the subject, while -en is the canonical nominative marker. Though puzzling, ERG-NOM seems the best way to indicate the hybrid set of suffixes. 7

SPLIT ERGATIVITY IN THE NENA DIALECTS

245

Goldenberg (1993), the Hertevin paradigm does occur as a marginal form in early Christian Neo-Aramaic texts. In the dialect of Umra, we find a particular variant of the Hertevin paradigm, available only for the 2nd person: qṭǝ́llǝtta instead of *qṭǝ́lluḥla ‘you (sg.m.) killed her’.8 Interestingly, the Hertevin paradigm occurs only in dialects which use L-suffixes to cross-reference the object in the preterite. Contrary to the majority of the NENA dialects, which recreate the accusative alignment by changing the object marker, the Hertevin paradigm acts on the subject marker, but not randomly. As Pennacchietti explained, the hybrid set of suffixes can be seen as an analogical extension of the direct endings of the present tense to the preterite. (16) Hertevin Present tense

VB.IMPRF – S-suffix – L-suffix

Preterite tense

VB.PRF – l- + S-suffix – L-suffix

The attestation of these hybrid forms testifies other attempts to limit ergativity ever since the 17th–18th centuries. Even the poor success of the Hertevin paradigm as an alternative to incorporated object exhibits some theoretical interest. Without a standard, dialects act with a greater degree of freedom. The large number of splitergative phenomena in the NENA dialects is good evidence for that. Data show that these dialects have experimented with different solutions, not all of them equally successful, in order to neutralize ergativity.

6. CORRELATIONS BETWEEN TYPES OF SPLIT ERGATIVITY AND ALTERNATIVE STRATEGIES FOR THE EXPRESSION OF THE PRONOMINAL OBJECT

The decline of incorporated objects is attested in written texts since the 19 th century. At the same time, NENA dialects developed alternative solutions, already available in the linguistic system although rarely used. Some of these solutions proved to be more successful, becoming potentially the only strategies available, while others have completely decayed. The reasons behind the success or the failure of these strategies depend on several factors. While the paradigm of Hertevin creates a new hybrid set of suffixes, the other two strategies use the material already available in the language in a more effective way. The construction with kəm- is perhaps the more effective from this

8

See Hobrack (2000: 90).

246

ALESSANDRA BARPTTO

point of view, since it recreates the best analogy with the present tense paradigm, at least under certain conditions. By contrast, the creation of a new set of suffixes is likely to create confusion and to be redundant in the language.

6.1. Split-S Dialects

Split-S dialects taken into account in this essay are the Jewish dialect of Sanandaj (Khan 2009) and of Sulemaniyya and Ḥalabja (Khan 2004). The intra-conjugational representation of object is limited to the 3rd person. The only alternative strategy to express the object is the accusative marker ʾill-, available for all person and obligatory for the 1st and the 2nd person. The construction with kəm- is not attested. The NENA dialects exhibiting the Split-S type are relatively rare. They are all part of the western Iranian sub-group of Jewish Neo-Aramaic dialects.

6.2. Dynamic-Stative Dialects

The Dynamic-Stative dialects analyzed, i.e. Urmi (Khan 2008b), Hertevin (Jastrow 1988), Bohtan (Fox 2009) and Koy Sanjaq (Mutzafi 2004), use the intraconjugational representation of object only with the 3rd person feminine singular and 3rd person plural. The use of the L-suffixes as an accusative marker seems to be the favourite alternative. The only exception is the Jewish dialect of Koy Sanjaq which uses the accusative marker ʾill-. In Koy Sanjaq, the L-suffixes as accusative marker are used only in the present tense. Constructions with kəm- are not attested. The Christian dialect of Hertevin shows yet another type of split. The Hertevin paradigm is not attested elsewhere, except in the dialect of Umra, where it is found only for the 2nd person.

6.3. Extended Ergative Dialects

Extended Ergative dialects taken into account in this essay are the Jewish dialect of Arbel (Khan 1999), Challa (Fassberg 2010), Betanure (Mutzafi 2008) and Amadiya (Hoberman 2000) and the Christian dialects of Barwar (Khan 2008a) and Qaraqosh (Khan 2002). Amongst all the dialects analyzed, the Extended Ergative dialects exhibit the higher degree of variation in the strategies adopted as alternatives to incorporated objects. The intra-conjugational representation of the object is optional in all the Extended Ergative dialects in varying degrees: while Challa uses it for all persons, the majority of the Extended Ergative dialects have limited it to 3rd person.

SPLIT ERGATIVITY IN THE NENA DIALECTS

247

The accusative marker ʾill- is attested in all the Extended Ergative dialects, with the exception of Qaraqosh. The construction kəm- is attested in all dialects, with the exception of Arbel. In Challa, as noted by Fassberg (2010: 125), there are some rare examples of this construction, which are probably a result of the influence of surrounding dialects such as Zakho.

7. CONCLUSIONS

The most interesting point is that the distribution of alternative strategies to incorporated objects shows a parallel with the degree of ergativity that characterizes the various NENA dialects. Split-S dialects exhibit the highest degree of ergativity: the ergative marker is restricted to transitive and unergative verbs, and is not found with unaccusative verbs. Whenever the ergative scheme is not available (with 1st and 2nd person objects), dialects opt for the less pervasive alternative. They simply replace the absolutive marker with a dedicated accusative one, the pronominal suffixes combined with the preposition ʾill-. Extended ergative dialects exhibit the lowest degree of ergativity, to the point that this alignment could be interpreted as a hybrid halfway between an ergative and an accusative system. This is the only type of dialects in which the present base prefixed by the preverb kəm- is the predominant strategy. Although the construction with kəm does not occur in all the Extended Ergative dialects, if ʾill- and kəm- are both available in a dialect, kəm- seems to be the favourite strategy and used more frequently. According to the collected data, the only Extended Ergative dialect which does not use the construction with kəm- is the Jewish dialect of Arbel. In order to explain this exception, we can formulate some hypotheses. Following the geographical distribution provided by Pennacchietti (1988: 276), the construction with kəm- is predominant especially among Christian dialects, with some small exceptions (such as the Jewish dialects of Amadiya and Zahko). The combination of the peripheral position of Arbel and the influence of other Jewish dialects could be the reasons behind the lack of kəm-. Likewise we can consider the Jewish dialect of Challa. Its peripheral position makes it less innovative in many respects than other dialects. The preference for the construction with kəm- displayed by the extended ergative dialects can be seen as further evidence that ergativity is perceived by NENA dialects as an unstable and extraneous phenomenon. While the Split-S dialects just opt for the introduction of a dedicated accusative marker to cross-reference the object, the Extended Ergative dialects prefer a strategy

248

ALESSANDRA BARPTTO

that completely changes the past paradigm and makes it symmetrical with the present tense. Finally, the last interesting point concerns Dynamic-Stative dialects. In fact, this type of dialect prefers using L-suffixes as an accusative marker to cross-reference the past object. Pennacchietti (1994: 272) observes that this model is now restricted to two peripheral areas (the South-East of Turkey and the Iranian Kurdistan), separated by a zone in which the construction with kəm- is dominant. Pennacchietti assumed that the use of L-suffixes as an accusative marker is an innovation compared to intraconjugational representation of objects. However, it was subsequently replaced by the construction with kəm-. The kəm- paradigm can be considered the most innovative strategy adopted by NENA dialects, whose expansion has restricted the use of Lsuffixes as an accusative marker. Since we have found at least one case (the Jewish dialect of Koy Sanjaq) in which a Dynamic-Stative dialect cross-references the object by the construction with ʾill- and not with the expected L-suffixes, the correlation between the dynamicstative split and the use of L-suffixes for direct objects may be simply explained as an areal feature, with no connection with ergativity.

REFERENCES

Dixon, Robert M. W. 1994. Ergativity. Cambridge: Cambridge University Press. Doron, Edith, and Khan, Geoffrey. 2012. ‘The Typology of Morphological Ergativity in Neo-Aramaic.” Lingua 122: 225–240. Fassberg, Steven E. 2010. The Jewish Neo-Aramaic Dialect of Challa. Leiden: Brill. Fox, Samuel Ethan. 2009. The Neo-Aramaic Dialect of Bohtan. Piscataway, NJ: Gorgias Press. Goldenberg, Gideon. 1993. “Otto Jastrow, Der neuostaramäische Dialekt von Hertevin – A Review Article.” Journal of Semitic Studies 38: 295–308. Hoberman, Robert D. 1989. The Syntax and Semantics of Verb Morphology in Modern Aramaic. A Jewish Dialect of Iraqi Kurdistan. New Haven: American Oriental Society. Hobrack, Sebastian. 2000. Der neuaramäische Dialekt von Umra (Dereköyü): Laut und Formenlehre, Texte, Glossar. Unpublished Master’s Dissertation. Friedrich Alexander Universität, Erlangen – Nürnberg. Jastrow, Otto. 1988. Der neuostaramäische Dialekt von Hertevin (Provinz Siirt). Wiesbaden: Harrassowitz. Khan, Geoffrey. 1999. A Grammar of Neo-Aramaic: The Dialect of the Jews of Arbel. Leiden: Brill. —. 2002. The Neo-Aramaic Dialect of Qaraqosh. Leiden: Brill. —. 2004. The Jewish Neo-Aramaic Dialect of Sulemaniyya and Ḥalabja. Leiden: Brill.

SPLIT ERGATIVITY IN THE NENA DIALECTS

249

—. 2008a. The Neo-Aramaic Dialect of Barwar. Leiden: Brill. —. 2008b. The Jewish Neo-Aramaic Dialect of Urmi. Piscataway, NJ: Gorgias Press. —. 2009. The Jewish Neo-Aramaic Dialect of Sanandaj. Piscataway, NJ: Gorgias Press. Mutzafi, Hezy. 2004. The Jewish Neo-Aramaic Dialect of Koy Sanjaq (Iraqi Kurdistan). Wiesbaden: Harrassowitz. —. 2008. The Jewish Neo-Aramaic Dialect of Betanure (Province of Dihok). Wiesbaden: Harrassowitz. Nöldeke, Theodor. 1898. Kurzgefasste syrische Grammatik. Leipzig: T.O. Weigel. Pennacchietti, Fabrizzio. 1988. “Verbo neo-aramaico e verbo neo-iranico.” In Tipologie della convergenza linguistica. Atti del Convegno della Società Italiana di Glottologia. Bergamo 17–19 dicembre 1987, edited by Vincenzo Orioles, 93–110. Pisa: Giardini Editori e Stampatori. —. 1991. “Gli allomorfi della flessione preteritale del dialetto neoaramaico orientale di Hertevin (Turchia) in prospettiva storica.” In Semitic Studies in Honor of Wolf Leslau on the Occasion of His Eighty-Fifth Birthday November 14th, 1991, vol. II, edited by Alan. S. Kaye, 1197–1202. Wiesbaden: Harrassowitz. —. 1994. “Il preterito neoaramaico con pronome oggetto.” Zeitschrift der Deutschen Morgenländischen Gesellschaft 144: 259–283.

THE PARTICLE WAL IN THE JEWISH NEO-ARAMAIC DIALECT OF ZAKHO

ERAN COHEN 1. INTRODUCTION

The particle wal in the Jewish Neo-Aramaic dialect of Zakho (henceforth JZ) occurs mainly in conjunction with verbal forms (and only rarely with other predicative expressions such as ıtli ‘I have’). The particle itself is defined in Sabar (2002: 154b) as an emphatic particle—‘surely, indeed’ and as having a relation with walox, essentially an inflecting vocative exponent. However, the lexical translations which seem to fit are not consistent throughout and seem to work only in individual cases. The need naturally arises to introduce some order and to try to see what this particle has in common in all its uses, namely, what function it has in the system. The verbal complexes with the particle wal are discussed below and compared with analytic verbal forms which are based on the deictic, or presentative copula.1 The particle seems to have a significant and definable function in the verbal system of JZ, presenting further distinctions in the system.

* I would like to thank my teacher and colleague Gideon Goldenberg and my student and colleague Michal Marmorstein for their ideas, comments and corrections. 1 The deictic copula (as it is sometimes called in NENA circles) occurs throughout NENA, and is composed of the original copula preceded by various deictic elements, see Goldenberg (1993: 298–302), where it is termed “presentative”.

THE PARTICLE WAL IN THE JEWISH NEO-ARAMAIC DIALECT OF ZAKHO

251

2. THE FORMS

The etymology of the particle wal is not entirely clear. The closest option might be an inflecting vocative particle walox (2sg.m.), walax (2sg.f.), etc. In addition, it may be related to another element, (-)wā, probably making up the presentative copula (wēle< wa+īle) and possibly found in the distal demonstrative ʾwā(ha) ‘that yonder’. The particle wal is compatible only with the forms šqıl-le/qam šāqıl-le and k-šāqıl (and not, for instance, with p-šāqıl, as is clear from table 1). There are only a few examples with ıtl- ‘x has’ and īb- ‘there is’. The particle never occurs with any forms of the verb h.w.y. ‘be’, or with any form of the copula, which may point to the mutual exclusion of these forms. The level of juncture of the particle with the verbal form is low: wal precedes the verbal form, as well as its negation (see table 1). It seems that one particle is basically enough to mark several consecutive verbal forms.2 This means that it is only at the beginning of a grammaticalization process. Nevertheless, the forms with wal turn out to be central to the verbal system: Table 1: the internal structure of verbal forms 3

2

1

pwal la ∅

∅

kqam—

resulting (maximal) form

p-šāqıl šāqıl wal-lak-šāqıl wal-la-qam-šāqıl-le

šqılle wal-la-šqıl-le

As far as paradigmatic relations inside the syntagm are concerned, wal shows mutual exclusion with ∅ and with the backshifting morpheme -wa with the forms šqıl-le/qam šāqıl-le and k-šāqıl, see table 2 below. There are, however, difficulties, since -wa- may occur with other forms, namely, p-šāqıl (‘future’) and ∅-šāqıl (subjunctive) as well, whereas wal never does. This fact may seem almost trivial, but a similar situation obtains in Ṭuroyo (Hemmauer and Waltisberg 2006: 27, ft. 15), where the prefix ko- and the suffix -wa

2

This is in some analogy with qad or la- in Classical Arabic, and shows that the chain of verbal forms is treated as a unit (Michal Marmorstein, personal comunication).

252

ERAN COHEN

seem to exclude each other. This mutual exclusion may have some general consequence in Neo-Aramaic (henceforth NA).3 These affixes in complementary distribution have in common the function of marking, inter alia, various tense and aspect values, as is clear looking at the functions in table 2. These different functions are discussed and explained below. form affix simple (+∅) function +walfunction +-wa(-) function

Table 2: the paradigmatic relations of wal šqıl-le

qam šāqıl-le

šqıl-le

qam šāqıl-le preterite wal šqıl-le wal qam šāqıl-le present perfect šqıl-wā-le qam šāqıl-wā-le pluperfect

k-šāqıl k-šāqıl aorist wal k-šāqıl actual present k-šāqıl-wa imperfect

3. A SHORT LITERATURE REVIEW

Several ideas for a solution are found in other dialects. Pre- (or post-) verbal particles are not new to NA: One prominent case which immediately arises is the particle lā- in the Jewish dialect of Arbel (Khan 1999: 111–114 [discussing form and crossdialectal relations]; 265–275 [describing the syntax], as well as Khan 2000). This particle is central to the Arbel verbal system, it is involved, as “a marker of prominence and intensity” (Khan 2000: 330) in the formation of perfects and perhaps also actual presents. The problem is that the resolution is not very high, probably due to movement away from an earlier phase where this particle, now thought of as a frozen copula, used to have basically these functions. The functions it fulfils in the synchronic state of the dialect are sometimes contradictory. Another case is the particle (käl)ko- in Ṭuroyo. Ritter (1990: 50–54) mentions the particle (käl)ko- as a part of the formation of the perfect, as well as of what he terms “imperfect”, aiming, it seems, at the present (without actually insisting on a specific type of present). Jastrow (1993: 146–147) tells us that in order to insist on the “actual present” value, the particle käl precedes k(o)-. There is some connection with perception as well, as this Verstärkung is mentioned in Ritter (1990: 97) as occurring following verbs of perception. This function is resumed further below. 3

The opposition between the affixes may be explained as pointing to the relative distance from the deictic center -wal (or in Ṭuroyo, ko-) brings us closer, while -wa- removes us away from it (Michal Marmorstein, personal communication).

THE PARTICLE WAL IN THE JEWISH NEO-ARAMAIC DIALECT OF ZAKHO

253

Heinrichs (2002), in a programmatic paper, ties together the various formatives, or strategies, of present perfects and actual presents in Neo-Aramaic, especially those forms which are based on a particle ~ copula and a finite verbal form. This strategy is represented in the left hand side of table 3. The right hand side of the table shows the other strategy, namely, an auxiliary with a gerund or perfect participle. Note that the functions are deemed similar by Heinrichs: Table 3: the formation of present perfects and actual presents

1. finite verb strategy +

šāqıl

šqıl-le

particle ~ copula

function

actual present

present perfect

2. auxiliary strategy

copula

+

gerund

perfect participle

Heinrich begins to explain the relationship between the present perfect and the actual present as having to do with the term presentative (which may be done more extensively) and concludes that there is a cyclic movement from the function of actual present to the function of general present, this movement generally being analogous to the cyclic change from perfect function to past (a case in point is NA šqıl-le, which probably started as a perfect). The perfect has been discussed in Hopkins (2002, formation and levels of grammaticalization in Kerend), Khan (2008a, formation and function) and Kapeliuk (2008, function). The momentary or actual present is briefly mentioned in Kapeliuk (2006), but is otherwise given very little attention in the general linguistic literature.

4. PARTIAL SOLUTION

Compared with, for example, Jewish Arbel, the situation in JZ is different. To begin with, the basic formal distinction between the verbal forms is somewhat more detailed (table 4):

254

ERAN COHEN Table 4: the subjunctive base and its tenses preformative

1

∅

2

p-

3

k-

4

base šāqıl

qam-

5

—

šqıl-

form

basic function

p-šāqıl

future/epistemic

šāqıl

subjunctive

k-šāqıl

aorist4

qam-šāqıl-le

preterite

šqıl-le

Since the basic distinction between the various verbal functions already exists at this level, wal would be expected to expand these categories further, and indeed it does. The discussion starts with the dialogue: a basic premise is that when the primary linguistic material at one’s disposal mostly consists of narrative texts—folktales etc.—it is desirable to keep narrative and dialogue apart and to describe each of them independently. The beginning of the solution for the function of wal forms is hinted at in the selection of Frei’s “deux-milles phrases”, translated into JZ in Avinery 1988. The material itself is problematic, heavily elicited and devoid of context. Nevertheless, we find several correspondences, as given by the informants, which may point to a possible solution. The translation was given originally in Modern Hebrew, which had neither perfect nor progressive tenses at the time. Examples (1) and (2) each bring together wal plus preterite with the analytic, usual form of the perfect: the deictic or presentative copula plus the participle in example (1) and future/epistemic expression of h-w-y ‘be’ in example (2): (1) hawwa la good PTC

kull-u all-3PL

wal wal

šmeʾ-lu hear.PST-3PL

ʾıbb-a in-3SG.F

‘After all, everyone (has) heard about it’. hawwa la good PTC

xabra word

kull-a all-3SG.F

māsa village

wēl-e COP-3SG.M

zvīra spread.PTCP.SG.M

go in

‘After all, the word (has) spread in the entire village’. (2000/894) 4

The “aorist” is used as it is used in Turkish (Lewis 1967: 115–122), that is, the as an unbounded tense. This term is equally used by Polotsky (1991: 273).

THE PARTICLE WAL IN THE JEWISH NEO-ARAMAIC DIALECT OF ZAKHO (2) A) ʾatta la k-āw-ıt now NEG PRS-be-2SG.M

255

xīla (?) eat.PTCP.SG.M

You probably haven’t eaten yet (?) [Hebrew: ‫]עוד לא אכלת‬ B) bale wınne wal but indeed wal

xıl-li mqabıl eat.PST-1SG.C before

sē-li. come.PST-1SG.C

‘But indeed I ate’. (=have eaten 2000/1021) [Hebrew: ‫]אכלתי‬ In example (1), the two parts are given as semantic parallels. In example (2) the question in A uses an auxiliary form of ‘to be’ which denotes epistemic meaning (“probably”).5 The answer is produced in a classical situation for the use of the perfect. Both examples illustrate the relationship between the two different strategies (see table 3 below). Examples (3) to (5) show us that the speakers feel there is some correspondence between the progressive present (copula with the gerund) and the form composed of wal and the aorist/present form: (3) wal wal

g-yatw-ın PRS-sit-1SG.M

(=w-ın (COP-1SG.M

bītāwa) ... sit.GER)

xazır near

dūk-ıt place-NUC

k-ḥaml-i ṭrombēl-e PRS-stand-3PL bus-PL ‘I (am) sit(ting) ... near the place (where) the buses stand’. (2000/392) (4) wal wal

k-kasw-ın

PRS-write-1SG.M

xakma DET.INDF.PL

(=w-ın bıksāwa) (COP-1SG.M write.GER)

ksāw-e letter-PL

‘I write/am writing letters’. (2000/1960) (5) lē-w-an NEG-COP-1SG.F wal wal

la

NEG

bıġzāya see.GER

k-xazy-an

PRS-see-1SG.F

‘I do not see anything’. (2000/10)

čı

NEG.DET

čı

NEG.DET

mındi thing mındi thing

Example (6) shows the same correspondence, but in the 3 rd person: 5

The form la k-āw-ıt, being also the negative of p-āw-ıt ‘you will/may be’, occasionally has this epistemic value. See Cohen (2008: 52–53).

256

ERAN COHEN

(6) wēl-e COP-3SG.M wal wal

bılyāpa study.GER

mušpāṭ-e. law-PL

g-lāyıp PRS-study.3SG.M

maḥkam-at. law-PL

‘He studies/he is studying law’. (2000/732) However, the only example which actually hints at the function is example (7): (7) ma what

wal wal

bıʾwāza do.GER

g-ō-t ʾatta? PRS-do-2SG.M now

ma what

w-ıt COP-2SG.M

ʾatta? now

‘What are you doing now?’ (2000/1006) The compatibility with the adverb ʾatta is another step towards a solution. It seems that here too, as in Heinrichs (2002), the natural comparison of wal forms is with the analytic verbal forms: Table 5: the present-related analytic verbal forms structure 3rd person 1st ~ 2nd person function copula+participle wēle šqīla wın ~ wıt šqīla present perfect copula+gerund

wēle bıšqāla

wın ~ wıt bıšqāla

progressive present

Note that compound tense expressions occur with a special distribution of copula type, depending upon the person: the present copula (wın ~ wıt) occurs with the 1st and 2nd person, whereas the presentative copula (wēle) only occurs with the 3rd person. This distribution is valid in verbal complexes only, where the present copula (īle) rarely occurs.6 Starting with the better known notions, first the analytic expression of the perfect in JZ is compared with the construction in question in context:7 (8) mtaslım

yān surrender-IMP.SG or

6

p-qaṭl-ın-nox FUT-kill-1SG.M-ACC.2SG.M

ʾaxxa. here

For the copular system of JZ see Cohen (2008), especially pp. 61–62 for the presentative copula. 7 Most of the examples from here onwards are taken from the Polotsky Zakho text collection, the main corpus at the base this enquiry, and their location in the text is given by original page. Other sources are duly indicated.

THE PARTICLE WAL IN THE JEWISH NEO-ARAMAIC DIALECT OF ZAKHO w-ın COP-1SG.M

musya ʾımm-i ṭallasma gūre bring.PTCP.SG.M with-1SG.C 300 men

257

naṣāy-e fighter-PL

‘Surrender or I will kill you here. 
I have brought with me three hundred fighting men’. (954) (9) yā

VOC

ʾıstāz-an lord-1PL

ḥakōma, 
 king

maxwax-lox show.1PL-DAT.2SG.M

wal wal

ʾanya

DET-PL

musē-lan bring.PST-1PL qazān-at pot-PL

‘O our lord the king, 
we have brought these pots to show you’. (469) In examples (8) and (9), both perfects show current relevance—in the first it is the rationale for the threat, and in the second it explains the reason why the speakers came to the king. The following pair occurs in a situation typical of perfects as well: (10) ay

kma

DEM QUANT

waʿda īs ınnu time EXIST COMP

le-w-en NEG-COP-1SG.F

mpīqa go_out.PTCP.SG.M

‘It has been some time that I have not gone outside’. (106) (11) ay

ʾī-b-a

DEM EXIST-in-3SG.F

wal wal

qarīwıt close-NUC

mpıq-le go_out.PST-3SG.M

pāpor ship

+

šāta year

palge 
 half

tıjjār-e merchant-PL

‘It has been (lit. there is) almost a year and a half 
(since) a merchant ship has left’. (424) Example (10) covers a span of time during which some action has not been performed. Example (11) occurs in a context which makes it easy to analyze—the boat has left some time ago and there are implications: the protagonist now needs to wait for several more years for its return. The semantic link with current relevance is transparent in example (12) as well: (12) wal wal

ʾuz-li make.PST-1SG.C

tafq-ım-b-u find-1SG.M-in-3PL

nīšanq-e sign-PL

b-urxa 
 in-road

māṭo how

hayya quickly

‘I have made signs on the way 
how to find them quickly’. (915)

258

ERAN COHEN

Making the signs in the examples means that it is now possible to find the way to the treasure.8 The same principle applies to example (13): (13) tımmal láswā-le xa zwaʾta ʾāxıl yesterday NEG-EXIST.PST-DAT.3SG.M INDF.DET piece eat.3SG.M muġ nāše ū like people and

ʾıdyo today

wal wal

pıš-le become.PST-3SG.M

mare bēsa… owner house

‘Yesterday he did not have a loaf of bread to eat 
like human beings, 
but today he has become a house-owner’. (926) The negative form simply follows the particle wal (example 14): (14) yā

VOC

flankaso, husband

ʾıbb-ox in-2SG.M

hīl ʾatta till now

wal wal

la

NEG

mqōrıš-li 
 interfere.PST-1SG.C

p-či-mındi in-NEG.DET-thing

‘O husband, until now I have not interfered 
with you at all’. (927) When the particle wal occurs with the aorist form k-šāqıl, it seems that we have some relationship with the progressive, or more specifically, one of its functions, namely, expressing the actual present, referring to the specific moment in question: (15) mā

what

ū and

wal wal

g-ōz-ēt, PRS-do-2SG.M

yā

VOC

gōra man

jındāya, fine

āhet

PRON.2SG.M

ʾanya čōl-e? DEM.3PL desert-PL

‘What are you doing, o fine man, 
in (lit. you and) this wilderness?’ (290) (16) dı-mar-ri PTC-tell.IMP.SG-DAT.1SG.C šūl-ox occupation-2SG.M

ū and

yā

nāša, man

VOC

mā what

ṣaneʾta craft

ma-yl-e 
 what-COP-3SG.M

g-ōz-et? PRS-do-2SG.M

‘Tell me, o man, what is your occupation 
and what craft do you practice?’ 8

wal with the preterite seems to have some kind of logical connection with causality, not the local circumstantial type, but rather the textual type that expressed by the particle ʾinna in Classical Arabic (Michal Marmorstein, personal communication). This bears further investigation.

THE PARTICLE WAL IN THE JEWISH NEO-ARAMAIC DIALECT OF ZAKHO

259

(512) The difference between examples (15) and (16) brings out the distinction between a particular activity and a generally practiced activity: what the addressee is doing at a specific place and time and what the addressee does for his living. The same distinction is made between examples (17) and (18): (17) sāl-ax come.IMP-2SG.F wal wal

ū and

xā one

baxta woman

kısl-ēni with-1PL

žān-e birth_pain-PL

k-ēsē-l-a PRS-come.3PL-DAT-3SG.F

‘Come, one of our women is having birth pains’. (297–298) (18) rāba sqīla ū jındāya-le 
 ū very pretty and fine-COP-3SG.M and ta to

ḥukum rule

dīd

gyān-e. self-3SG.M

PRON.NUC

k-ēse PRS-come.3SG.M

mın from

k-šākıl

PRS-be_worthy.3SG.M

bale but

xapča a bit

rīxa smell

pumm-e mouth-3SG.M

‘He is very pretty and fine 
and he is worthy of his rule. 
But a little smell comes out of his mouth’. (678) Birth pains are sudden and occur in a given, bounded, specific time. The bad smell seems to be considered characteristic (or, at any rate, unmarked in this respect). An analogous relationship is found between examples (19) and (20): (19) yā

VOC

brōn-i, son-1SG.C

mēka k-ēs-ıt bıd whence PRS-come-2MS in

mír-rī-l-a: say.PST-1SG-DAT-3SG.F ṭāʾın-n-i look_for-1SG-DAT.1SG

w-ın

COP-1SG.M

xa

INDF.DET

dō

DEM.SG.M

bızvāra wander.GER dūk-ıt place-NUC

lēle? night

ta to

dmāxa sleep.INF

‘My son, where do you come from tonight? I said: 
I am wandering in order to look for a place for me to sleep’. (408–409) (20) yāla wēl-e rıš pımm-e. 
 ʾay dammıd g-zāvır, child COP-3SG.M on mouth-3SG.M whenever PRS-turn.3SG.M g-māreʾ-lax PRS-hurt.3SG.M-DAT.2SG.F

260

ERAN COHEN

‘The child is [stuck] on his mouth. 
Whenever he turns around, you feel pain’. (SAG 1.14) Note, however, that in example (19), the actual present is expressed by the analytic form, namely, by the progressive present. The following pair shows this distinction in forms in the 2nd person: (22) ṭamā ́

why bıd in

w-ētın COP-2SG.M

bīzāla go.GER

b-urxa in-way

ū and

mēnōxe look.GER

ʾarʾa? ground

‘Why are you walking the road and looking at the ground?’ (520) (21) wal-ox āhet mani VOC-2SG.M you.SG.M who

w-ētın, COP-2SG.M

dīd

COMP

kud lēle each night

k-šaql-ıt xā līra mın dē dıkkāna ū g-ēz-ıt? PRS-take-2SG.M one lira from DET.SG.F store and PRS-go-2SG.M ‘You, who are you that every night you take 
one coin from this store and go (away)?’ (4) To sum up, both wal k-šāqıl and wēle ~ wın bıšqāla as actual presents are opposed to the aorist or “general present” expressed by the form k-šāqıl. This difference is already formulated for Literary Urmi by Polotsky (1991: 273), who makes the distinction between the form ki pǝtix, which he succinctly describes as an aorist (see footnote 5 above), and the form biptǝxili which he regards as the “true present”. Olga Kapeliuk uses the term “actual present” in the same context: When combined with the copula ijli in Neo-Aramaic and with the copula näw in Amharic this present gerund forms the actual present. The biptəxili present in Urmi Neo-Aramaic is opposed to the general present ki pətix as general versus concrete action with a specific or definite object, thus providing a kind of definiteness in a language which uses no definite article. (Kapeliuk 2010: 129)

5. BETWEEN PROGRESSIVE AND ACTUAL PRESENT

These two functions, “present progressive” and “actual present”, are occasionally considered to be one and the same in the literature, the terms “progressive” and “actual” being used interchangeably, for instance, in Heinrichs (2002: 240)—“actual (progressive) present tense”. The progressive, however, may be associated with another direction of grammaticalization, i.e., that of imperfectivity, for instance, in Khan (2002), where the form lā qaṭil is discussed and linked with “present progres-

THE PARTICLE WAL IN THE JEWISH NEO-ARAMAIC DIALECT OF ZAKHO

261

sive or habitual aspect”. Progressives basically grammaticalize into presents, as well as into imperfectives (already Bybee et alii 1994: 140–149), as in fact is shown by Heinrichs (2002). For instance, the following pair of examples feature the same progressive forms, with somewhat different functions: (23) k-īʾ-en PRS-know-1SG

wēl-u

COP-3PL

ʾatta now

bıṇṭāra guard.GER

bāb-i-yımm-i (ı)l-dáy father-1SG.C-mother-1SG.C to-DET.SG.F

daqīq-ıt minute-NUC

yaʾl-ın enter-1SG.M

b- tarʾa in-door ‘I know: my parents are now waiting 
for the minute I come in through the door’. (425) (24) 
ʾaxnan PRON.1.PL dē

DET.SG.F

kull-a šāta all-3SG.F year

w-ax COP-1PL

bınṭāra guard.GER

ıl to

dalīsa vine

‘All year long we have been guarding this vine’. (655) Whereas the progressive in example (23) denotes the actual present, referring to the very moment of speech, the expression in example (24) has a year’s span (the present-perfect-progressive translation into English may indicate the difference). In the following example, the span is a day: (25) yā

VOC

w-axni COP-1PL

ıstāz-i lord-1SG.C

ḥakōma, king

bıṭʾāya look_for.GER

axni

PRON.1PL

kull-e all-3SG.M

yōma day

ıll-ox to-2SG.M

‘Our Lord the King, 
all day long we have been looking for you’. (48–49) In short, the progressive is a bounded, on-going action or state. It forms a scale from focalized to durative (Bertinetto et alii, 2000),9 which are in effect two oppo9

Bertinetto et alii (2000) define the two poles of the progressive as follows: “(i.) “Focalized” progressive constructions (henceforth Foc-PROG), i.e. those expressing the notion of an event viewed as going on at a single point in time, here called “focalization point”.

262

ERAN COHEN

sites. The focalized present progressive yields, so it seems, a focalized present. The durative one does not. Note that a durative progressive may grammaticalize into a general imperfective, that is, the other end of the scale. Judging from our examples, the form wal k-šāqıl is not progressive: unlike the latter, it is never put between two temporal boundaries (the adverbials kulle yōma ‘all day (long)’, kulla šāta ‘all year (long)’, etc.), and all it does is to convey focalized “nowness” in the dialogue.

6. THE PRESENTATIVE FUNCTION: PERCEPTION CONTENT

The term “presentative” is used throughout NA mainly with reference to particles which “draw attention to referents or situations” (Khan 2009: 243). However, in the Barwar dialect (Khan 2008b §§15.3.2; 15.5.5; 17.6.3 and 18.1.5), the presentative is shown to be a form with a textual function, most prominently as a complement of verbs of perception. A similar function is found in JZ, in which the so called deictic, or presentative copula follows a verb of perception as one of several complementation strategies, which is mostly found in narrative: (26) wēl-a 
 OP-3SG.F C

tūta sit.PTCP.SG.F

marōḥe ventilate.GER yōm-āsa. day-PL bıdyāqa knock.GER

dīd-e OBJ-3SG.M

xá-bēna one-moment

xá-yōma one-day ū and

xa-rēš-e near-head-3SG.M

bıbxāya weep.GER

mux like

kull-u 
 all-3PL

k-šamʾ-a xáʾ wēl-e PRS-hear-3SG.F someone COP-3SG.M

tarʾa
 door


‘She’s sitting one day near him, 
fanning him and weeping as all days.
 Suddenly The focalization point may be overtly expressed in the sentence, or else it may be r ecovered through the context, being the object of a presupposition. Needless to say, the focalization point does not exhaustively localize the event; it simply indicates a point in time overlapping the progressive event, while the actual duration of the latter remains indeterminate. (my e mphasis) (ii.) “Durative” progressive constructions (henceforth Dur-PROG), i.e. those that are evaluated relative to a larger interval of time. Here again, however, the actual duration of the event remains indeterminate. Even when a durative temporal adverbial is present, this does not deli mit the event but merely yields a vantage point from which the situation is observed.” (Bertinetto et alii 2000: 527)

THE PARTICLE WAL IN THE JEWISH NEO-ARAMAIC DIALECT OF ZAKHO

263

she hears: someone is knocking on the door’. (687) This strategy is used in order to reflect a switch of perspective, namely, to that of the character in question,10 or in other words, his point of view as displayed from his own hic et nunc. Accordingly, it is attested only in the 3 rd person. The clause is not subordinate and hence considered a strategy on a different (macro-)syntactic level. The presentative is comparable in this function to direct speech, but within the domain of perception. In addition to the presentative copula, wal forms too are found in this function (example 27):

(27) pıš-la become.PST-3SG.F xakma INDF.PRON.PL

wal wal

bazıl mbínōke, dawn

ʾōqat so_much

g-dēq-i PRS-knock-3PL

ıl to

ġzē-le 
 see.PST-3SG.M

tarʾa door

‘It was dawn, no sooner did he look: 
some (people) are knocking on the door’. (716–717) Examples (26) and (27) reflect what is actually seen (or heard) by the character in question. Both formal options are found mixed in this function: (28) wal k-xāze mā wal PRS-see.3SG.M what wēl-e COP-3SG.M

bīsāya mın raḥūqa come.GER from afar

wēl-e rıš COP-3SG.M on šıkıl sort

wal wal

ū and

xá

kapōra giant

INDF.DET

dāra, tree

kāp-e wēl-e dwīqa shoulder-3SG.M COP-3SG.M hold.PTCP.SG.M

ḥēwān-ıt dunye animal-NUC world

wēl-e COP-3SG.M ū and

k-xāze xa PRS-see.3SG.M INDF.DET

rūvīk-e fox-PL

mtultya dōhun hang.PTCP.SG.M OBJ.3PL k-ēse PRS-come.3SG.M

mın from

pıḷıng-e tiger-PL

ʾaryı-wāsa, lion-PL

bıd daw in DET.SG.M

kull-u all-3PL ū and

dāra tree

nēčir hunt

‘He suddenly sees—what does he see?—a giant is coming from afar, a tree is on his shoulder, he has caught all kinds of animals: foxes, tigers and li10

One finds a similar idea in Biblical Hebrew with the exponent wǝ-hinne ‘lo’, see Berlin (1983: 62–63).

264

ERAN COHEN ons, and he has hung them on the tree and he is coming back from the hunt’ (534)

In example (28), the analytic forms wēle bīsāya ‘he is coming’, wēle dwīqa ‘he has caught’ and wēle mtultya dōhun ‘he has hung them’ are together in the same construction as wal kēse ‘he is coming (back)’. The first verbal form (wēle bīsāya) and the last (wal kēse) are almost synonymous.11 Inside the narrative, the verbal forms do not exhibit a deictic function; consequently, these forms do not indicate tense as they might do in dialogue. Accordingly, this is no “actual present”, but rather what is captured in the eye of the character—either things that are happening right that moment (or, in the case of the perfect—things that have occurred before, but left some indication for what has taken place). In this presentative strategy that is all we get—a current, on-going action or a perfect kind of action, of which merely the resulting state may be witnessed. Only rarely are wal forms attested in perfect presentative function: (29) qım-le, psıx-le stand.PST-3SG.M open.PST-3SG.M sāfīna dīd ship PRON.NUC gıvanān-e bank-PL

ʾāla side

dō

DET.SG.M

mīsa dead

šıbāka, window wal wal

g-mēnıx, 
 PRS-look.3SG.M mṭē-la arrive.PST-3SG.F

ıl to

xēta other

‘So he opened the window, looking, 
the ship of the dead (person) 
 has reached the banks of the other side’. (MA 18.4) The boat is currently at the other bank, hence it is concluded that it has reached the bank. It is important to note the difference between normal narrative, where events follow one another, where the (present) perfect function is uncalled for, and this special function, whose deictic coordinates (that is, the hic et nunc) belong to the seeing character, and therefore can exist within the narrative.

7. TWO TENSES: PRESENT PERFECT AND ACTUAL PRESENT

The question arises why the particle wal interacts only with these verbal forms, and whether there is something in common between the resulting tenses. Saying that 11

It is also possible that the first form, wēle bīsāya, functions as a circumstantial (although this does not tend to happen after a verb of perception), while the final form wal kēse function as the main verb, actual present of the beholder.

THE PARTICLE WAL IN THE JEWISH NEO-ARAMAIC DIALECT OF ZAKHO

265

both are present tenses (one concrete and the other resultative) is true, but not enough: the aorist is a kind of present as well, and so is the present progressive. The present perfect and actual present seem at first not to be structurally opposed to one another at all, since they may be compatible with one another (and thus not in the same paradigm), as in ‘I have been doing’, which is present perfect progressive. They are, however, basically opposed in NA (literary Urmi included,12 despite similar compatibility), and not only in the present: any expression of being, in any tense, mood, or aspect may show the opposition between a predicative gerund and a perfect participle. Back to their expression in the present—the dimension we have been discussing—they are both related to a strong deictic centre, the holy trinity of I-here-now, namely, the one related to the spatio-temporal point of utterance—the speaker, in the case of the dialogue. Other tenses (e.g. the future and the past) are removed, and have no direct link with the moment of utterance and consequently bear less significantly for the moment. Actually, even the aoristic present (the form k-šāqıl), often covering a generally-valid-fact or habit, is not particularly related to the moment of utterance, it merely crosses it.

8. PRESENTATIVE FUNCTION AND NOW-RELATED TENSES

We have seen that both sets, the wal forms as well as the analytic verbal forms occur in: A. presentative function following verbs of perception. This occurs in narrative as well as elsewhere (that is, in dialogue as well, as long as the forms follow a verb of perception). B. In dialogue, as present perfect and actual present. The similarity of the two apriori different environments, namely, dialogue vs. the complement slot of perception verbs, requires an explanation. The forms in question are closely related, in dialogue, to the hic-et-nunc of the speaker. In the presentative function, on the other hand, they are closely related to the hic-et-nunc of 12

Polotsky 1991: 265 adduces the following possibilities: 1 bivǝjǝGER [i]jli biptǝxǝGER/ ptijxǝPTCP lit. ‘he is being opening/having opened’ 2 vijjǝPTCP [i]jli biptǝxǝGER/ptijxǝPTCP lit. ‘he has been opening/had opened’. There is syntagmatic compatibility in that the gerund and the participle may co-occur in the same compound verbal form. Since the auxiliary is always some expression of being (either verb or copula), the participle and the gerund, which may occur in the same slot, can in addition be regarded as constituting together a paradigm and thus as being in paradigmatic opposition. In some lexemes, e.g. postural ones, no such opposition exists: wın tīwa ‘I have sat, am sitting, am seated’.

266

ERAN COHEN

the character, namely, the one who sees. But the principle that unifies the nature of the tenses in both environments is actually the field of vision. The latter as a component of the deictic nature of tense is used, for instance, in Janssen’s (1993)13 conception of tense. For our purposes a considerably less-abstract concept of vision is needed: what can be physically observed is only 1. what is actually happening or 2. what has happened but left a trace clear enough for us to reconstruct. These two perspectives coincide with the actual present and the present perfect. The future, the past or the pluperfect are not observable, hence not in the field of vision of either the speaker or the viewer. So it is this relatively simple principle of observable realities which constitutes the explanation for both functions of the forms. It is worth noting that in Biblical Hebrew, where the presentative wə-hinne is used, it too occurs with participles, non-verbals or with the perfect qāṭal, but nothing else. The wal construction occasionally occurs in subordination. In this case, its function is analogous to the one it fulfils in dialogue, but the reference point depends on the location of the subordinate clause: for instance, if it occurs in narrative, that would be its point of reference (examples 30 and 31): (30) mṭē-le arrive.PST-3SG.M wal wal

ıl to

gı-ḥākım PRS-rule.3SG.M

day

DET

bāžer 
 dīd city PRON.NUC

axōn-e brother-3SG.M

rēš-a on-3SG.F

‘He arrived at that town 
which his brother was (lit. is) (currently) ruling’. (149) (31) bale but

lá

NEG

rʾıš-le feel.PST-3SG.M

d-bēs-e COMP-house-3SG.M

wal wal

l-gyān-e to-self-3SG.M xrū-le be_ruined.PST-3SG.M

‘But he did not notice that his home had (lit. has) been destroyed’. (826)

13

Janssen (1993: 88–92) explains tenses and demonstratives as having in common 1. the speaker’s vantage point; 2. a region in the speaker’s mental field of vision; and 3. the entity. Discussing only preterite and present, the difference between which is found in the speaker’s mental field of vision, more specifically, “the region of (dis)focal referential concern”, where the present is focal and the past is not. It seems that Janssen’s speaker’s vantage point could be transferred into narrative, where what would be in focus is rather the “preterite”, which is the main narrative form.

THE PARTICLE WAL IN THE JEWISH NEO-ARAMAIC DIALECT OF ZAKHO

267

Both examples provide us with a relative actual tense (example 30) and a relative perfect (example 31). It should be noted that only the verbal forms with wal may occur in subordination.

9. CONCLUSIONS

In discussing the particle wal, several issues have been covered: A. the forms with which it occurs; B. the function it has in dialogue (two tenses: present perfect and actual present); C. the relationship between progressive and actual present; D. the presentative function, which is one narrative strategy of presenting the content of perception; E. the relationship between the two tenses—present perfect and actual present— which are exclusively now-related tenses; and F. the relationship between the presentative function and now-related tenses with regard to the actual field of vision. The data and peripheral issues discussed are put in table 6: Table 6 is a summary of the comparison conducted among the forms discussed up to this point. Table 6: summary functions/ environments 1 presentative

wal šqıl-le

wal k-šāqıl-le w-ın ~ wē-l-e šqīla wın ~ wē-l-e bıšqāla

rare

+

+

++

both functions ⇕ occur in 3rd pers. only 2

circumstantial

functions in 3a dialogue

3b

remarks for dialogue

4 subordination 5

syntactic dependency

no perfect this form denotes only presentperfect

+

actual present perfect ~ passive this form is not progressive (stative lexemes are compatible)

this form could denote resultative and passive in addition to present perfect

progressive: focalized ~ durative 1. progressive ranges from durative to focalized 2. the progressive is by definition given boundaries

yes

no

no

yes

268

6 other remarks

ERAN COHEN 1. wal forms show relatively low juncture (preceding the negative particle, etc.) and are incompatible with h.w.y. as well as with -wa-.

2. Several element orders exist: a subject nominal may occur first, second or third. But, one condition is always kept: that wēle precedes the participle or gerund.

Row 1. All forms are found in presentative function, following a perception verb. Row 2. The compound tenses with the presentative copula are found in circumstantial functions (for instance at beginning of example 26: wēla tūta ‘she is sitting’), but not the verbal forms with wal. Incidentally, in both presentative and circumstantial function only the 3 rd person is attested. Row 3. Whereas the compound forms with the presentative copula may be ambiguous (perfect or passive), or polysemic (focalized or durative progressive), wal forms in dialogue designate precise present perfect and actual present. Here, incidentally, all persons occur. That the form is not progressive is indicated by the fact that stative lexemes are compatible with the actual present. Row 4. Of the two types, only wal forms may occur in subordination. Row 5. While wal forms may occur independently, the compound forms with the presentative copula do not occur by themselves and need some formal support (very much like the Biblical Hebrew wǝ-hinne). Row 6.1. wal forms show weaker juncture (table 1), but are in strong interaction with the affix -wa. In addition, wal is incompatible with h-w-y in general.
 Row 6.2. The presentative copula always precedes the participle or gerund; a subject nominal, when present, may occur anywhere: first, in the middle, or at the end, namely: gōra wē-l-e ʾısya ~ wē-l-e gōra ʾısya ~ wē-l-e ʾısya gōra.

REFERENCES

Avinery, Iddo. 1988. The Aramaic Dialect of the Jews of Zakho. Jerusalem: Israel Academy of Sciences and Humanities. [in Hebrew] Berlin, Adele. 1983. Poetics and Interpretation of Biblical Narrative. Sheffield: Almond. Bertinetto, Pier Marco et alii. 2000. “The progressive in Europe.” In Tense and Aspect in the Languages of Europe, edited by Östen Dahl, 517–558. Belin: Mouton de Gruyter. Bybee, Joan et alii. 1994. The Evolution of Grammar—Tense, Aspect and Modality in the Languages of the World. Chicago: University of Chicago Press. Cohen, Eran. 2008. “The Copular Sentence in Jewish Zakho Neo-Aramaic.” Journal of Semitic Studies 53: 43–68. Goldenberg, Gideon. 1993. “Review Article of Jastrow, Otto, Der neuaramäische Dialekt von Hertevin (Provinz Siirt).” Journal of Semitic Studies 38: 295–308.

THE PARTICLE WAL IN THE JEWISH NEO-ARAMAIC DIALECT OF ZAKHO

269

Heinrichs, Wolfhart. 2002. ‘‘Peculiarities of the Verbal System of Senāya within the Framework of North Eastern Neo-Aramaic (NENA).” In Sprich doch mit deinen Knechten aramäisch, wir verstehen es!: 60 Beiträge zur Semitistik: Festschrift für Otto Jastrow zum 60.Geburtstag, edited by Werner Arnold and Hartmut Bobzin, 237–268. Wiesbaden: Harrassowitz. Hemmauer, Roland, and Waltisberg, Michael. 2006. “Zum relationalen Verhalten der Verbalflexion im Ṭurojo.” Folia Linguistica Historica 27 (1-2): 19–59. Hopkins, Simon. 2002. “Preterite and Perfect in the Jewish Neo-Aramaic of Kerend (Southern Iranian Kurdistan).” In Sprich doch mit deinen Knechten aramäisch, wir verstehen es!. 60 Beiträge zur Semitistik: Festschrift für Otto Jastrow zum 60.Geburtstag, edited by Werner Arnold and Hartmut Bobzin, 281-298. Wiesbaden: Harrassowitz. Janssen, Theo. 1996. “Deictic and Anaphoric Referencing of Tenses.” In Anaphores temporelles et (in-)cohérence, edited by Walter de Mulder, Liliane Tasmowski and Carl Vetters, 79–107. Amsterdam: Rodopi. Jastrow, Otto. 1993. Laut- und Formenlehre des neuaramäischen Dialekts von Mīdin im Ṭur-‘Abdīn. Wiesbaden: Harrassowitz. Kapeliuk, Olga. 2006. “The Neo-Aramaic Tense System in the Light of Translations from Russian.” In Loquentes linguis: studi linguistici e orientali in onore di Fabrizio A. Pennacchietti, edited by Gorgio Borbone et alii, 381–390. Wiesbaden: Harrassowitz. —. 2008. ‘The Perfect Tenses in Urmi Neo-Aramaic.” In Aramaic in its Historical and Linguistic Setting, edited by Holger Gzella and Margaretha Folmer, 313–334. Wiesbaden: Harrassowitz. —. 2010. “Some Common Innovations in Neo-Semitic.” In Proceedings of the 13th Italian Meeting of Afro-Asiatic Linguistics Held in Udine, May 21st‒24th, 2007, edited by Frederick Mario Fales and Giulia Francesca Grassi, 123–131. Padova: S.A.R.G.O.N. Khan, Geoffrey. 1999. A Grammar of Neo-Aramaic: The Dialect of the Jews of Arbel. Leiden: Brill. —. 2000. “The Verbal System of the Jewish Neo-Aramaic Dialect of Arbel.” Journal of the American Oriental Society 120 (3): 321–332. —. 2002. “The Neo-Aramaic Dialect of the Jews of Rustaqa.” In Sprich doch mit deinen Knechten aramäisch, wir verstehen es!. 60 Beiträge zur Semitistik: Festschrift für Otto Jastrow zum 60.Geburtstag, eited by Werner Arnold and Hartmut Bobzin, 395–410. Wiesbaden: Harrassowitz. —. 2008a. “Remarks on the Function of The Preterite and the Perfect in NorthEastern Neo-Aramaic.” In Neo-Aramaic Dialect Studies, edited by Geoffrey Khan, 104–130. Piscataway, NJ: Gorgias Press. —. 2008b. The Neo-Aramaic Dialect of Barwar. Leiden: Brill.

270

ERAN COHEN

—. 2009. The Jewish Neo-Aramaic Dialect of Sanandaj. Piscataway, NJ: Gorgias Press. —. 2012 “The Evidential Function of the Perfect in North-Eastern Neo-Aramaic Dialects.” In Language and Nature. Papers presented to John Huehnergard on the Occasion of his 60th birthday. Studies in Ancient Oriental Civilization 67, edited by Rebecca Hasselbach and Naʾama Pat-El, 219–227. Chicago: The Oriental Institute of the University of Chicago. Lewis, Geoffrey. 1967. Turkish Grammar. Oxford: Oxford University Press. (Reprinted with corrections 1986). MA—see below. Meehan, Charles, and Alon, Jacqueline. 1979. “The Boy whose Tunic Stuck to him: A Folktale in the Jewish Neo-Aramaic Dialect of Zakho.” Israel Oriental Studies 9: 174–203. Polotsky, Hans Jakob. 1991. “Modern Syriac Conjugation.” Journal of Semitic Studies 36: 263–277. Ramat, Anna G. 1997. “Progressive Periphrases, Markedness, and Second-Language Data.” In Language and its Ecology: Essays in Memory of Einar Haugen, edited by Stig Eliasson and Ernst Håkon Jahr, 261–285. Berlin: Mouton de Gruyter. Ritter, Helmut. 1990. Ṭūrōyo. Die Volkssprache der syrischen Christen des Ṭūr ‫ﻉ‬Abdîn. C: Grammatik. Pronomen, „sein, vorhanden sein“, Zahlwort, Verbum. Stuttgart: Steiner. Sabar, Yona. 2002. A Jewish Neo-Aramaic Dictionary. Dialects of Amidya, Dihok, Nerwa and Zakho, Northwestern Iraq. Wiesbaden: Harrassowitz. —. 2007. “Agonies of Childbearing and Child Rearing in Iraqi Kurdistan: A Narrative in Jewish Neo-Aramaic and its English Translation.” In Studies in Semitic and General Linguistics [AOAT 334], edited by Tali Bar and Eran Cohen, 107–145. Münster: Ugarit Verlag. SAG—Sabar 2007. 2000—Avinery 1988: 76–208 (by number).

MORE FROM H. J. POLOTSKY’S NACHLAẞ ON THE VERB IN URMI

OLGA KAPELIUK 1. INTRODUCTORY REMARKS

Upon his death in August 1991, Professor H. J. Polotsky left most valuable material on Neo-Aramaic which he had collected and on which he had been working during several decades. His Nachlaß may be divided into two parts: on the one hand, a huge quantity of texts in the Jewish dialect of Zakho in his handwriting, collected by him under dictation from native speakers and a draft of a grammar of this dialect, which, however, seems to have been lost and, on the other hand, his material on literary Urmi, which was entrusted to me. The latter includes, among others, three thick files of what might be considered as a very first draft of a book on the Urmi verb. It is clear that Polotsky planned to write a book in English under the title “Modern Syriac Conjugation”. The three files hold 550 pages in all, mainly typewritten, but also partly covered with his handwriting. Not all the pages were fully used; some contain only a few lines, others are fully covered with text. Most of the material is composed of non-translated examples arranged according to subjects and classified in a tentative table of contents. The latter is placed, together with two bibliographical lists— one of the sources and the other of grammatical and lexical works—at the beginning of the second file. The sources mostly used are the works written by Father Paul Bedjan1 and literary works translated from Russian in the 1930’s and written in the

1

Born in Khosrova (1838–1920) near Salamas, to the north of lake Urmi, P. Bedjan studied at the French Lazarist Seminary in his native town and later in Paris, where he was ordained priest. After serving for 20 years as a missionary in Khosrova and Urmi he finally settled in Europe, first in Paris, then in Belgium, and finally in Germany, where most of his works in Old Syriac and in Neo-Aramaic were printed (Murre-van der Berg 1994; id. 1999: 111–115; Vosté 1945 and the bibliography adduced there; Kapeliuk 2002: 361–363 [id 2009: 526–528] and Polotsky 1962: 276 note 2 [id. 1971: 620 note 2]).

272

OLGA KAPELIUK

Soviet Latin script.2 The Soviet script, known as the New Alphabet, was adopted by Polotsky in his publications and manuscripts as a phonological transcription for all Urmi examples, including those from P. Bedjan which were originally printed in the Nestorian characters. Also examples from other sources, using their proper phonetic transcription, such as texts published by A. Kalašev, K. Tsereteli and R. Duval, were retranscribed in the New Alphabet. There is also a fourth file with parts of the same material as in files 1-3 but written in French under the title “Conjugasion néosyriaque”. Most of the examples in this file come from P. Bedjan’s writings and some of them are accompanied by their source in French or Latin. I have the impression that this file preceded the English ones. H. J. Polotsky had already used many examples from his files in the various articles published by him,3 but large parts of the material remain still unpublished. Besides the examples, there is very little written text formulated by Polotsky himself, such as would deal with the characteristics of the examples, the grammatical rules governing them and any linguistic statement in general that would offer us guidelines for reconstructing what he intended to write in his description of the Urmi verb. When such statements are found, they give the impression of being sketchy early drafts rather than well formulated grammatical rules. Only a few subjects received a somewhat more detailed treatment pointing to a more advanced stage of writing, but these paragraphs are still very fragmentary and tentatively phrased, as may be deduced from the numerous erasures, correction, additions and repeated attempts at formulating. The more detailed sections contain the basic verbal parts, a subject taken up by Polotsky in most of his articles (1961; 1979; 1984-86 and 1996); the system of stems (dəsti according to the terminology of the native grammarians) in Urmi with a comparison with Zakho, which I published as the first paper of this series under the title: “From H. J. Polotsky’s Nachlaß on the Verb in Urmi” (Kapeliuk 2005) and, towards the end of the third file, a text that seems to be a final summary of the whole tense system, which will be copied almost in its entirety in § 2.2 below. I will also add in §§2-3 some material on more minute points of interest from Polotsky’s files. All this is accompanied, in the spirit of this volume devoted to language contacts, by some comparisons with Kurdish, which are all my own suggestions (see also Kapeliuk forthcoming a).

2

For this material see Polotsky (1961: 3–10 [id. 1971: 587–595]), Kapeliuk (2006: 381– 383 [id. 2009: 553–555]). In the Soviet script ə represents a front a, and ь a central i. 3 A complete list of H. J. Polotsky’s articles on Urmi Neo-Aramaic is included in the Bibliography at the end of this article.

MORE FROM H. J. POLOTSKY’S NACHLAß ON THE VERB IN URMI

273

2. MODERN SYRIAC TENSES

Towards the end of the third file of Polotsky’s material, i.e. near the end of the whole collection of examples on the Urmi verb, there are a few attempts at drawing tables representing the tense system in its entirety. Table I and Table II reproduced below seem final and there is also a reproduction of them on a stencil, proving that Polotsky presented it to his students in one of his classes, or to some other audience as was his custom, which may suggest that he found these tables satisfactory. Besides these two tables, Polotsky had drawn in his publications two different Tables of Tenses. The first one was included in his path-breaking article “Studies in Modern Syriac” (1961: 22 [id. 1971: 606]),4 and the second, disavowing the first one, in his paper “Neusyrische Konjugation” (1984–1986: 324). The two tables from the file are followed by a few pages of text, which will be reproduced here almost verbatim. At the end of this paragraph I will add, on my proper initiative, some points of comparison with the Kurmanji Kurdish tense system, compiled from the following sources: Baran Rizgar (1993: Appendix 386–399), Joyce Blau and Veysi Barak (1999) and Wheeler M. Thackston (2006). To distinguish between Polotsky’s text and mine, I am putting his between quotation marks.

2.1. Tables TABLE I Modern Syriac Tenses SIMPLE COMPOUND & BIFURCATE

with h-v-j ptixl-

vili+biptǝxǝ/ptijxə

qəm pətixl

[-------]

bit pətix

bit hǝvi

ki pətix

ki hǝvi

4

DOUBLY COMPOUND & BIFURCATE

with the copula

with h-v-j & copula

biptǝx-ili5

bivǝjili biptǝxǝ/ptijxǝ

ptijx-ili

vijli biptǝxǝ/ptijxǝ

This article was hailed as the opening of a new era in the scientific study of Neo-Aramaic, after a standstill of half a century, due to the two World Wars (Poizat, 1973–1979: 392). 5 In subordination and in negation the copula precedes the lexical verb form: ijli biptǝxǝ/ptijxǝ and leli biptǝxǝ/ptijxǝ.

274

OLGA KAPELIUK

TABLE II Tenses with vǝ “in Praeterito” SIMPLE SIMPLE +vǝ

COMPOUND & BIFURCATE with h-v-j + vǝ

ptixli

ptix-vǝ-li

[--------]

qǝm pǝtixl-

qǝm pǝtix-vǝ-l

[-------]

bit pǝtix

bit pǝtix-vǝ

bit hǝvi-vǝ + biptǝxǝ/ptijxǝ

ki pǝtix

ki pǝtix-vǝ

ki hǝvi-vǝ

ijli

ijvǝ

bivǝji-vǝ/viji-vǝ

2.2. Explanation of the Tables

All of Polotsky’s remarks that follow refer to Table I: “The tenses fall into four groups. The preceding table presents them in what I believe to be their mutual relationship. It is less easy to give them appropriate names. In order to avoid unnecessary terminological quibbles I shall try to choose names which are formally descriptive, in preference to abstract ones. “The 1st group can be called “Simple Tenses” in the sense that their component parts have no independent existence in the language.6 “Simple” is thus meant as the opposite of “compound”. The other tenses are “compounded” with auxiliary elements. They have an additional characteristic in the bifurcation of the stem into two branches, viz. biptəxə and ptijxə. They can therefore with equal justification be called compound and bifurcate tenses, according to whether we wish to stress either the presence of the auxiliary elements or the bifurcation of the verb stem. They are further subdivided according to the auxiliary element. In the second group the auxiliary element is one of the simple tenses of the verb h-v-j;7 the simple tense qəm pətixl- has

6

Here Polotsky chose to ignore the independent use of the subjunctive pǝtix because Tables I and II are centered on the indicative forms of the verb and the subjunctive and imperative moods are not included in them, although the conditional bit pǝtixvǝ is included in Table II (O. K.). 7 For h-v-j as a full verb of existence and becoming see Polotsky (1996: 39–42). On the orthography of the copula in the New Alphabet see Kapeliuk (forthcoming b).

MORE FROM H. J. POLOTSKY’S NACHLAß ON THE VERB IN URMI

275

no counterpart in this group, because h-v-j, as an intransitive verb, takes no object suffixes.8 “The four (2x2) groups of tenses can be looked upon as a progressive expansion, which at the same time implies the bifurcation of the verb-stem into two branches, viz. biptəxə and ptijxə. The first step of expansion is brought about by one of two auxiliaries, either h-v-j or the Copula, the next step by both auxiliaries together. The fourth group stands in the same relation to the third as the second does to the first, inasmuch as both are expanded with the help of h-v-j; just as the second group is compounded with the simple tenses of h-v-j, i.e. the forms of the first group, so the fourth group is compounded with bivəjə and vijə inflected by the Copula, i.e. the two tenses of the third group. “The difference between the first group and the second, is to a large extent stylistic, i.e. it depends on the choice of the writer, not on the nature of the event described: ptixli simply reports the event, while vili biptəxə/ptijxə is used when the writer wishes to dwell on details, or to indicate, by implication or by an explicit adverb, a certain length of time. The same event is narrated in Hist. 3109 by drilun qǝm qǝmcij “they flogged him” and in Manuel 448/11–12 by vilun dirju qǝm qǝmcij.10 This example shows that with certain transitive verbs ptijxǝ is required rather than biptǝxǝ. Such is the case with verbs of markedly ingressive meaning. In this respect such transitive verbs are closely related to verbs of motion and posture; the contrast of ingressive vs. stative seems to be the decisive factor. “In the third group the auxiliary is the copula, forming an equivalent of the Present Continuous (biptǝxǝ) and the Present Perfect (ptijxǝ) respectively. This form contrasts with ki hǝvi; so far from being merely a stylistic variant of ki pǝtix, depending on the choice of the speaker, it has the distinctive function of referring the statement to the speaker, i.e. the time of speaking and the speaker’s sphere of interest (in statements which can be qualified by adverbs like “now”). Its natural field of employment is therefore direct speech. “The last group is doubly compound (“temps surcomposés”) or doubly bifurcate inasmuch as the bifurcation occurs in the auxiliary as well as in the verb stem. Of 8

qǝm hǝvili exists only as a full verb with the meaning ‘to have’ and ‘to be born to’ (on this verb see Polotsky 1979: 209–210). The use of qǝm is otherwise restricted to transitive verbs with definite objects and necessitates the presence of the object pronoun -li (O. K.). 9 Examples from Polotsky’s files are quoted by an abbreviated or by a full form of the first noun of their title. The references of the examples could not be checked because Polotsky’s library was sold and is no longer accessible. For the same reason the bibliographical indications about his sources are partly incomplete. 10 For more about this opposition see Polotsky (1984–1986: 329).

276

OLGA KAPELIUK

the four possible combinations only vijili biptǝxǝ is fairly common: it expresses an action which has been going on for a stated length of time which extended shortly before the moment of speaking, e.g. Hist. 183: ţla şinni vitivǝn mamusux ‘For three years, I have been giving you suck’, Hist. 237: hǝr min d ijlǝ vьrta, hǝr vitilǝ binşǝqe ǝqlij ‘Ever since she has come in, she has been kissing my feet’. The fourth group stands in the same relation to the third as the second to the first, inasmuch as both are expanded with the help of h-v-j: just as the second group is compounded with hv-j in the forms of the first group, so the fourth group is compounded with h-v-j in the two varieties of the third group.”

2.3. Comparison with Kurmanji Kurdish The tense system of Urmi and of Neo-Aramaic in general is much more complicated and richer than that of Kurmanji Kurdish, but the impact of the Iranian languages on the basic structure of the Neo-Aramaic system is undeniable. I have already drawn attention to parallels between Urmi and Modern Persian tenses (Kapeliuk 1996a: 6268 [id. 2009: 499–504]) to which the Kurdish tense system is quite similar. The first basic common principle in Kurdish and in Neo-Aramaic is the split of the verb into a present and a past stem, e.g. Kurmanji: from ‘to fall’ ketin: present stem: kev, past stem: ket, which are paralleled by Urmi: ‘to open’ pǝtix for the present and ptix-l for the past. The second common principle is the presence of modal markers, preposed to both stems, di- for the present and for the progressive/habitual aspect in the past, bi- for the subjunctive, the imperative and the conditional moods, and ê bi- for the future in Kurmanji, and the tense markers ki for the present, bit for the future and qǝm for the past of transitive verbs as a variant of ptix-l in Urmi. The third common principle is the suppletion in the auxiliaries: the copula e/ye ‘he is’ and the verb bûn ‘to be, to exist, to become’ in Kurmanji and ijli as the copula and the root h-v-j for the rest in Urmi. The fourth common principle consists in the formation of compound tensed with the auxiliaries and the passive/perfect participle ketî in Kurmanji and the participle ptijxǝ/ptixtǝ/ptijxi in Urmi. But the most extraordinary common feature, a real anomaly for a Semitic language, is the presence of an ergative construction in Neo-Aramaic, parallel to Kurdish, even though in Urmi it is limited to the preterite ptix-lij ‘I opened’ alone, whereas in Kurmanji it includes all the forms derived from the past stem, namely just for the indicative: the preterite ez ket-im ‘I fell’, the durative/progressive past ez di-ket-im ‘I was falling’, the present perfect ez keti-me ‘I have fallen’,11 and the pluperfect ez keti-bûm ‘I had fallen’. 11

Blau and Barak (1999: 71) include, in addition, a form of the present perfect preceded by the durative/progressive marker di which they call “imparfait narrative”. This form is not

MORE FROM H. J. POLOTSKY’S NACHLAß ON THE VERB IN URMI

277

There are, however, two major differences between these two tense systems, which also explain the relative simplicity of the Kurdish system as compared with Urmi. In Kurmanji the auxiliaries are used only in their simple form, at least in the indicative, with which we are dealing here, e.g. keti-ye ‘he has fallen’ and keti-bû ‘he had fallen’. In Neo-Aramaic, on the other hand, the auxiliaries not only may have, beside their simple form, an enlarged form in praeterito, i.e. with the past converter vǝ: ki hǝvi-vǝ and bit hǝvi-vǝ, but also the stem of the auxiliary h-v-j undergoes a bifurcation into the present gerund bivǝjilǝ and the participle vijǝ/vitǝ/viji producing the compound forms of the auxiliary: bivǝj-ili, vij-ili, bivǝj-ivǝ and vij-ivǝ.12 The second major difference concerns the nature of the verb form that is combined with an auxiliary in the compound tenses. Whereas in Kurmanji it has only one form, the past participle, in Urmi13 it is bifurcated in two: the present gerund consisting of the preposition bi and the infinitive on the one hand, and the participle on the other hand. This bifurcate form of the verb that is combined with the auxiliary allows the Urmi verbal system to render the opposition between a durative and a stative aspect and it literally doubles the number of existing compound tenses as compared with Kurmanji.14 In the following paragraphs, some characteristics of the two bifurcate forms are discussed.

3. VERBAL NOUNS

The two most important verbal nouns: the participle ptijxǝ/ptixtǝ/ptijxi and the infinitive ptǝxǝ may be both predicative and nominal. ptijxǝ/ptixtǝ/ptijxi is predicative in its bare form, ptǝxǝ basically has to be preceded by the preposition bi- ‘in/by’.

3.1. Verbal Nouns with Predicative Function

The infinitive preceded by the preposition bi- and the participle in its bare form are made fully predicative and finite with the help of the auxiliaries formed from the copula or from the verb h-v-j, and as such they are fully incorporated into the Urmi mentioned by Thackston, nor by Rizgar, but it is similar to the Modern Persian “continuative perfect” ‫( ميخريده است‬see example (13) below). 12 Modern Persian also has a compound auxiliary for the past preterite: ‫( خريده بوده است‬see Kapeliuk 1996a: 63–65 [id. 2009: 500–502]). 13 Not all the Neo-Aramaic dialects use the infinitive + bi. 14 I hope to be able to investigate in the near future the function of the diverse tenses in both languages since I have recently acquired, thanks to the help of our Russian colleague, Alexey Lyavdansky, a long text in standardized Urmi, of which I also have a parallel Kurmanji version.

278

OLGA KAPELIUK

tense system as shown above.15 In the absence of an auxiliary, the two verbal nouns still remain predicative though not finite. biptǝxǝ corresponds to a present gerund, as in examples (1) (2). The participle ptijxǝ/ptixtǝ/ptijxi corresponds to a past gerund (Kapeliuk 1996b [id. 2009: 483–495]; id. 2008: 133–145 [id. 2009: 565–577]), and it is active if derived from an intransitive verb or from a transitive verb accompanied by an object pronoun, as in examples (3) (4), and passive if derived from a transitive verb without an object, as in example (5): (1) U: xzi-lun xǝ see.PST-3PL INDF.ART dimm-u blood-his

paxra body

d of

barnǝşǝ man

bь-slaja in/by-descend.INF

‘They saw a body of a man, his blood flowing’. (Vies 301/21) (2) U: jimm-ij qim-la min mother-my get_up.PST-SUB.3SG.F from lǝ

NEG

bi-ljǝzǝ griş-lǝ in/by-hurry.INF pull.PST-SUB.3SG.F

qǝm before

mis u table and

l-kis pǝnçǝrǝ towards window

‘My mother got up from the table and without hurrying (lit. not hurrying) withdrew towards the window’. (MXS 10/8)16 (3) U: qim-lǝ b pǝlg-id get_up.PST-SUB.3SG.F in half-of brun-ij son-my

leli, night

ǝnǝ I

ţlьta, sleep.PTCP.SG.F

şqil-lǝ take.PST-SUB.3SG.F

‘She got up at midnight, I [being] asleep, she took my son’. (Hist. 112/16) (4) U: xzi-li Polus … brijkǝ briş birkǝk-u, riş-u see.PST-SUB.1SG Polus kneel.PTCP.SG.M on knees-his head-his tqijl-u u ijd-u murm-e swing.PTCP-OBJ.PRON.3SG.M and hands-his raise.PTCP-OBJ.PRON3PL l-şmǝjjǝ to-heaven 15

See also Polotsky (1961: 20–23 [id. 1971: 604–607]; id. 1984–1986: 327–329; id. 1996: 12–31.) 16 Examples collected by me are quoted by the initials of the author and of the title.

MORE FROM H. J. POLOTSKY’S NACHLAß ON THE VERB IN URMI

279

‘He saw Paulus … kneeling (lit. kneeled) on his knees having lifted his head and having raised his hands to heaven’. (Vies 148/20) (5) U: dranan-e arms-their qǝm

PAST

sьjrь ǝnij bind.PTCP.PL they

dǝrij-lun put.PL-OBJ.PRON.3PL

qǝm before

am with

ǝbunǝ priest kǝliskǝ carriage

d of

hǝkim governor

‘Their arms bound, they with the priest, they put them before the carriage of the governor’. (Vies 118/8) When it forms the gerund the preposition bi- ‘in/by’ is written in the Soviet script directly before the infinitive, whereas when used as a regular preposition ‘by’ with the infinitive serving as a noun, it is written separately, as any other preposition in this script and it has a long ij as its vowel, as may be seen in example (6). In example (7), the preposition bij before the infinitive is used in its full form because it indicates here the instrument and has no gerundial function.17 The special spelling of bi- before the infinitive points to a full grammaticalization of the gerundial form biptǝxǝ and it was rightly felt as such by those who devised the Soviet script. The gerundial bi- becomes zero in the first stem before roots starting with a labial consonant and in all the heavy stems on account of the original initial consonant m-, which is still present in other dialects but not any more in Urmi. But even in the first stem the preposition bi is not always found where it is expected. If the infinitive is repeated expressively in adverbial expressions marking a continuous concomitant action or state, the bi- is dropped as in examples (8) (9) and (10). This construction is also characterized by a special stress pattern with a single stress on the last syllable of the first infinitive, another sign of grammaticalization; in the Soviet script there is a hyphen between the two infinitives. Also the participle may be repeated for expressive means, as in example (11). bi- may also be missing sometimes with the negation lǝ, as in example (12). It is interesting to note that in contemporary literary texts, such as those which are currently published in the Journal of Assyrian Academic Studies, the use of bi is very irregular and it is often missing without any morphological or syntactical justification: (6) U: alma bi-xzǝj-e-vǝ people in/by-see.INF-OBJ.PRON.3PL-COP.PST.AUX

17

See also Polotsky (1984–1986: 328 [id. 1996: 18–20]).

bijj-u in-him

kul-le all-3PL

280

OLGA KAPELIUK ǝn mjatratь d those virtues which bij in

ijvǝ

COP.PST.AUX

xizj-e see.PTCP-OBJ.PRON.3PL

Mar M. Mar M.

‘The people used to see in him all those virtues which they had seen in Mar M’. (Vies 195/13) (7) U: [A man is asked how it is possible to attain holiness. The answer is] bij sjama soma u bij by fast.INF fast.SUBS and by

salujь sluta pray.INF prayer.SUBS

‘By fasting and praying’ (Vies 229/16) (8)

U: zil-lun go.PST-SUB.3PL

ǝnni the

dervijş-i dervish-PL

zmǝrǝ-zmǝrǝ sing.INF-sing.INF

‘The dervishes went singing’. (Merx 10/10) (9) U: rxaţà-rxaţa xdьr-e-li run.INF-run.INF circle.PST-OBJ.PRON.3PL-SUB d arpa of four

ǝrzijbǝni surroundings

vagunь train_cars

‘Running he circled the surroundings of four train cars’. (MXS 55/27) (10) U: b in

xǝdutǝ u joy and

zdajà-zdaja gxik-li fear.INF-fear.INF laugh.PST-SUB.3SG.M

‘He laughed in joy and apprehension’. (Gor. 75/3)

(11) U: ǝtxǝ up ǝt lazьm … thus too you it_is_necessary ţalb-ьt-la demand-SUB.2SG.M-OBJ.PRON.3SG.F

qijdǝ-qijdǝ burn.PTCP-burn.PTCP komǝg help

‘So you too need to ask most arduously (lit. burned-burned) God’s help’. (Imit. 152/12) (12) U: lǝ

NEG

dard-ux pain-your

jimm-ux mother-your

lǝ

NEG

jaqurta heavy xzǝjǝ! see.INF

bit

FUT

hojǝ, be.3SG.F

in if

met-it die-2SG.M

‘Won’t your pain be heavy if you die without seeing your mother!’ (GSI 21/32)

MORE FROM H. J. POLOTSKY’S NACHLAß ON THE VERB IN URMI

281

It is interesting to point out that the two Neo-Aramaic forms with gerundial function, namely the participle and the infinitive preceded by the preposition bi-, have their parallels in Kurmanji literary texts written in the Hawar (Latin) script. The use of the participle ket-î is quite frequent and here too as in the case of Urmi, the participle of intransitive verbs is active as in example (13) and passive with transitive verbs as in example (14). This is a purely Indo-European phenomenon of active/passive diathesis not encountered in the other Semitic languages,18 but frequent also in Modern Persian, and its presence in Neo-Aramaic may be considered a clear case of Iranian interference. In Kurmanji the use of the infinitive or of an abstract deverbal noun with the preposition bi ‘with/by’ as circumstantial complement also exists. In his book “Nivîsa min – Mon livre” Bedir Khan (1965: 97) writes: “En kurde il n’existe pas de participe présent. On traduit à l'aide des prépositions “dans” bi, di … de etc. et par le substantif ou l’infinitif correspondant pris substantivement: di xewê de=en dormant (dans le sommeil); di xwendînê de=en lisant (dans la lecture19); bi girînê de=en pleurant (dans les pleurs20)”. The gerundial use in Kurmanji of an infinitive with the preposition bi is further illustrated in examples (15)21 and (16). The latter may be compared with the circumstantial complement in the NeoAramaic translation, as in (17), of the same Russian source quoted in example (18): (13) K: çend hev ji vexwarin-ê mest bûyî … some together from drink.INF-OBL drunk be/become.PTCP xir-xir xir-xir

di-kiri-ne PROG-make.PTCP-COP.AUX.PL

‘Some of them altogether, having become drunk, … were snoring’. (EŞŞ 250/46) (14) K: li paş wan xulam-ên at behind they.OBL servant-CNST.PL

18

wan … bi xencer, they.OBL with dagger

Occasional use of a passive participle with active meaning in Post Biblical and Israeli Hebrew, e.g. IH: savur-ni/svur-ani ‘I reckon’ is not a real case of active/passive diathesis because Hebrew uses besides an active participle, which is not the case in Neo-Aramaic nor Kurmanji form (Polotsky 1996: 25–28; Kapeliuk 2008: 137–145 [id. 2009: 569–577]). 19 Literally ‘in to read’. 20 Literally ‘with/by to cry’. 21 I copy the original French translation (Blau and Barak 1999: 220).

282

OLGA KAPELIUK şûr, debençe û tifing çekkirî, sword revolver and gun arm_make.PTCP

di-skinî-n PROG-stand.PST-SUB.3PL

‘Behind them stood their servants armed with daggers, swords, revolvers and guns’. (EŞŞ 240/11–13) (15) K: dayik-a wî bi dirûn-ê mother-CNST he-OBL with/by sew.INF-OBL xwe her_own

xwedî owner

zarûk-ên child-CNST.PL

kir-in do.PST-OBL.3PL

‘Sa mère subvenait aux besoins de ses enfants en faisant de la couture (lit. with/by to sew)’. (Blau and Barak 1999, 109/12) (16) K: diya min mother my û and

hanî, brought

bav-ê father-CNST

çû, go.PST.3SG bi with

şinasname identity_card

min… my

derxist take_out.PST.3SG

destlerizîn da hand_shake.INF give.PST.3SG

min my

‘My mother went, took out my identity card … and brought [it], her hand shaking she gave it to my father’. (PKK 15/18–19) (17) U: jimm-ij mucx-ǝ-lǝ mother-my find.PST-OBJ.3SG.F-SUB.3SG.F

pasport-ьj… identity_card-my

b-ijd-o rǝgud-tǝ ǝj juv-ǝ-lǝ in-hand-her shake.ADJ-F she give.PST-OBJ.3SG.F-SUB.3SG.F qǝ to

bǝb-ij father-my

‘My mother found my identity card … her hand shaking she gave it to my father’. (PBQ 9/21–23) (18) R: матушка отыскала мой mother found my

паспорт… и вручила его identity_card and handed it

батюшк-е дрожаще-ю рук-у father-DAT shake.PTCP.INS hand-INS ‘My mother found my identity card … and handed it with shaking hand to my father’. (PKD 112)

MORE FROM H. J. POLOTSKY’S NACHLAß ON THE VERB IN URMI

283

3.2. Deverbal Nouns with Nominal Function

Besides their predicative function as gerunds, the two deverbal nouns may act also as plain nominals—the infinitive as a noun and the participle as an adjective. An interesting aspect of the infinitive is that sometimes it is used in the feminine as may be seen from the resumptive pronouns in examples (19) and (20), although its morphological form corresponds to a Neo-Aramaic masculine. This anomaly is a loan from Kurmanji, which uses the infinitive in the feminine. Many Urmi nouns, either original Aramaic or borrowed from another Semitic language have adopted the feminine gender from Kurmanji, contrary to their original masculine morphological form, e.g. Urmi mьţra ‘rain’ is feminine (Maclean 1901: 172) according to baran (f.) ‘rain’ in Kurmanji, but contrary to meţrā (m.) in Old Syriac: (19) U: hərgis le never NEG.FUT d lə that NEG

həvit nəşə ruxanaja…, həl become.2SG.M person spiritual… until

tərgit-lə leave.2SG.M-OBJ.PRON.3SG.F

həmzum … speak.INF

‘You will never become a spiritual person if you don’t stop speaking…’. (Imit. 59/15) (20) U: qraja22 read.INF dənni those

qarax-la kuljom read.1PL-OBJ-3SG.F every_day

xə one

min from

təxmən-jəti thought-PL.F

‘To read, let us read every day one of those thoughts’. (Man. 189/2–3) The infinitive of transitive verbs may carry a suffixed pronoun which marks the object. While in the other Semitic languages a pronoun suffixed to the infinitive may render either the subject or the object, in Neo-Aramaic it normally has the function of an object pronoun, as in example (21). It is only extremely rarely that the suffixed pronoun marks the subject as in example (22), where the couple of intransitive verbs form a kind of compound, as may be inferred from the absence of a coordinating conjunction between the two. The usual way of marking by a suffix pronoun the subject with an abstract deverbal noun is by using instead of the infinitive the noun of action ptəxtə which can carry a pronoun both as subject and object (Kapeliuk

22

The infinitive, which functions here as an internal object or ‫ مفعول مطلق‬is definite.

284

OLGA KAPELIUK

1992: 69–72 [id. 2009: 564–568]). In example (23) we find an infinitive with a suffix pronoun as subject by analogy to the two nouns of action preceding it: (21) U: kull-e şul-ь baţьjl-ь-na şvuq all-PL matter-PL vain-PL-COP-3PL except maxxub-u love.INF-OBJ.PRON.3SG.M

marja Lord

min from

ələhə God

‘All the matters are vain except the love (lit. to love Him) of God the Lord’. (Imit. 2/11) (22) U: qjəm-u stand_up.INF-his d of

tjəv-u lə həvi sit.INF-his NEG be-3SG.M

b urxə by way

almajь laymen

‘His behaviour (lit. his to stand his to sit) should not be by the way of laymen’. (Imit. 217/20) (23) U: hivij hope

d of

plaţt-ij get_out.NOM.ACT-SUB.PRON.1SG

m-əxxə from-here

qţь-la, pjəşt-ij b urxə u cut.PST-SUB.PRON.3SG.F stay.NOM.ACT-SUB.PRON.1SG in road and tləq-ij be_lost.INF-SUB.PRON.1SG

b-itəj-inə in-come.INF-COP.3PL

qəm ajn-ь before eye.PL-my

‘The hope of getting out of here was cut, my staying on the road and my being lost are coming before my eyes’. (MPX 65/14–15) With the participle the suffixed pronouns always mark the object and it is according to the presence or absence of the object pronoun that the active/passive diathesis comes to light, as in examples (24) and (25). Occasionally an object pronoun is found with the participle of an intransitive verb, as in example (26), where it resumes an internal object. It is a well known fact that when the object is definite it has to be resumed by a pronoun attached to the verb and that with the infinitive and participle this pronoun is identical to the adnominal possessive pronoun. The function of this pronoun, whether it marks the direct object or the recipient, is determined by rules of definiteness.23 If the direct object is indefinite, then the pronoun 23

For a detailed discussion on the marking of the object see Polotsky (1979) and (1994).

MORE FROM H. J. POLOTSKY’S NACHLAß ON THE VERB IN URMI

285

attached to the verbal base marks the recipient, as in example (27). If the direct object is definite the pronoun takes the place of the recipient which has to be expressed by a preposition, as in example (28): (24) U: lə

NEG

dvur-ri all-ьj tarra d shut.IMP.SG-OBJ.PRON.3SG.M on-me door of

d ijli that COP.AUX.3SG.M

mar-an lord-our

purqənə salvation

ptijx-u open.PTCP-OBJ.PRON.3SG.M

‘Do not shut on me the door of salvation which our Lord has opened (it)’. (Vies 192/20) (25) U: qə do buş for DEM more d of

purqənə salvation

basura d small of

rajatt-ьj flock-my

tarra door

ptijx-ili open.PTCP-COP.3SG.M

‘For the smallest one of my flock the door of salvation is open(ed)’. (id. 191/11) (26) U: kull-ə e all-SG.F DEM

xidjutə joy

d that

ijvə

COP.PST.AUX.3SG

xdijt-o rejoice.PTCP-OBJ.PRON.3SG.F ‘All this joy that he had rejoiced (it)’. (Marie 310/9–10)

(27) U: b-ijəv-e-və in/by-give.INF-OBJ.PRON.3PL-COP.PST.AUX

libbə heart

‘He was encouraging them (lit. giving them heart)’. (Vie 249/13) (28) U: huqi-lun… relate.PST-SUB.PRON.3PL tarra d hemənutə qə door of faith to

d that dən those

ijvə

COP.PST.AUX

millət-ti nation-PL

ptijx-u open.PTCP-OBJ.PRON.3SG.M

xənp-i heathen-PL

‘It was related … that he opened the door of faith to those heathen nations’. (Hist. 355/7–9) With compound verbs, extremely frequent in Urmi, as well as in Kurmanji (Kapeliuk 2002 [id. 2009: 526–542]), there is another possibility of indicating the pronominal recipient, which seems to have escaped Polotsky’s attention. In this alternative construction the recipient is rendered by a possessive pronoun suffixed to

286

OLGA KAPELIUK

the nominal component of the verb as in examples (29) and (30), paralleled by Kurmanji24 in examples (31) and (32): (29) U: çuvvab-e hər answer-their at_all

levə COP.NEG.PST.AUX

maddur-u return.INF-OBJ.PRON.3SG.M

‘He was not answering them at all (lit. their answer he was not returning it)’. (GSI 6/30) (30) U: şurь-la lom-u vəd-u start.PST-SUB.3SG.F reproach-his make.INF-OBJ.PRON.3SG.M ‘She started to blame him (lit. his blame to make it)’. (Hist. 37/2)

(31) K: ji bo from to min my

peydakirin-a find.INF-CNST.F

gur-an wolf-OBL.PL

alîkarî-ya help-CNST.SG.F

bike make.IMP

‘Help me to find the wolves (lit. my help make)’. (Blau and Baran 1999: 109) (32) K: diy-a min pesn-ê mother-CNST.F my praise-CNST.M

min my

di-da PROG-give.PST.3SG

‘My mother was praising me (lit. my praise was giving)’. (EŞŞ 12/15)

REFERENCES Sources

EŞŞ=Ereb, Şemo 1989. Şivanê kurd – Le berger kurde Paris: Institut Kurde. [in Kurmanji and French] Gor.=Gorqij, M. 1936. Həqjətti. Moscow. GSI=Gorqij, M. 1934. Sotə Izergil. Moscow: Basmaxana d Məlkutə. Hist.=Bedjan, Paul. 1888. Histoire sainte. Paris. Imit.=Bedjan, Paul. 1885. Imitatio Christi. Paris. Man.=Bedjan, Paul. 1893. Manuel de piété. Paris. Marie=Bedjan. Paul. 1904. Mois de Marie. Paris. Merx=Merx, Adalbert. 1873. Neusyrisches Lesebuch. Tübingen: Tübingen Univärsitet.

24

For a similar construction in Sorani Kurdish see Traidia (2008: 91).

MORE FROM H. J. POLOTSKY’S NACHLAß ON THE VERB IN URMI

287

MPX=Marogulov, Q. and D. Petrosov 1935. Xrestomatija d saprajuta qə mədrəsə d şuraja. Moscow: Basmaxana d Məlkutə. MXS=Marogulov, Q. 1935. Xrestomatija d saprajuta qə mədrəsə mьsseta. Moscow: Basmaxana d Məlkutə. PBQ=Puşqin, Aleksander. 1937. Brətə d qapitan, translated by U. Bedrojev. Moscow: Basmaxana d Məlkutə. PKD=Pushkin, Aleksander S. 1978. Kapitanska dochka. Moscow: Chudozhestvennaja Literatura. [in Russian] PKK=Pûşkîn, Aleksander S. 1988. Keça kapîtan [In Kurmanji], translated by Hesenê Metê. Stockholm: Weşanên Welat. Vies=Bedjan, Paul. 1912. Vie des saints. Paris.

Bibliography

Bedir Khan, N. 1965. Nivîsa min – Mon livre: Cours pratique de la langue kurde. Paris: [École Nationale des Langues Orientales Vivantes]. Blau, Joyce, and Barak, Veyci. 1999. Manuel de kurde kurmanji. Paris: l’Harmattan. Kapeliuk, Olga. 1992. ‟Miscellanea Neo-Syriaca.” Jerusalem Studies in Arabic and Islam 15: 60–73. —. 1996a. “Is Modern Hebrew the only “Indo-Europeanized” Semitic Language? And what about Neo-Aramaic?” Israel Oriental Studies 16: 59–70. —. 1996b. “The Gerund and Gerundial Participle in Neo-Aramaic.” Sprachtypologie und Universalien Forschung 51 (3): 276–88. —. 2002. “Compound Verbs in Neo-Aramaic.” In Sprich doch mit deinen Knechten aramäisch, wir verstehen es! Festschrift für Otto Jastrow zum 60. Geburtstag, edited by Werner Arnold and Hartmut Bobzin, 361–377. Wiesbaden: Harrassowitz. —. 2005. “From H. J. Polotsky’s Nachlaß on the Verb in Urmi.” In Afro-Asiatic Studies – 11th Italian Meeting of Afro-Asiatic Linguistics, Bergamo 2003, edited by Alessandro Mengozzi, 349–358. Milano: Francoangeli. —. 2006. “The Neo-Aramaic Tense System in the Light of Translations from Russian.” In Loquentes Linguis – Studi Linguistici e Orientali in Onore di Fabrizio A. Pennacchietti, edited by Gorgio Borbone, Alessandro Mengozzi and Mauro Tosco, 381–390. Wiesbaden: Harrassowitz. —. 2008. “Between Nouns and Verbs in Neo-Aramaic.” In Neo-Aramaic Dialect Studies, edited by Geoffrey Khan, 131–147. Piscataway, NJ: Gorgias Press. —. 2009. Selected Papers in Ethio-Semitic and Neo-Aramaic Linguistics. Jerusalem: Magnes Press. —. 2012. “The Copula in Suppletion wih the Verb h-v-j in Standardized Urmi.” ARAM 24: 73–86.

288

OLGA KAPELIUK

—. forthcoming. “Language Contact between Aramaic Dialects and Iranian.” In The Semitic Languages and Dialects, edited by Stefan Weninger. Berlin – New York: de Gruyter. Maclain, Arthur John. 1901. Dictionary of the Dialects of Vernacular Syriac. Cambridge: Clarendon Press (reprinted in 1972). Murre-van den Berg, Hendrika Lena. 1994. “Paul Bedjan (1838–1920) and his NeoSyriac Writings.” In Symposium Syriacum VI, edited by René Lavenant, 381–382. Rome: Pontifico Istituto Orientale. —. 1999. From a Spoken to Written Language. Leiden: Instituut voor het Nabije Oosten. Poizat, Bruno. 1973–79. “Une bibliographie commentée pour le néo-araméen.” Groupe Linguistique d'Études Chamito-Sémitiques 18–23: 347–411. Polotsky, Hans Jakob. 1961. “Studies in Modern Syriac.” Journal of Semitic Studies 6: 1–32. —. 1962. “Review of J. Friedrich Zwei russische Novellen in neusyrischer Übersetzung und Lateinschrift.” Orientalia 31: 273–283. —. 1971. Collected Papers. Jerusalem: Magnes Press. —. 1979. “Verbs with two Objects in Modern Syriac.” Israel Oriental Studies 9: 204– 27. —. 1984–1986. “Neusyrishe Konjugation.” Orientalia Suecana 33–35: 324–32. —. 1994. “Incorporation in Modern Syriac.” In Semitic and Cushitic Studies, edited by Gideon Goldenberg and Shlomo Raz, 90–102. Wiesbaden: Harrassowitz. —. 1996. “Notes on Modern Syriac Grammar.” Israel Oriental Studies 16: 11–48. Rizgar, Baran. 1993. Kurdish-English, English-Kurdish Dictionary. London: Lithosphere. Thackston, Wheeler M. 2006. Kurmanji Kurdish – A Reference Grammar with Selected Readings. Iranian Studies, Harvard University. Traidia, Sandrine. 2008. “Les verbes composés (nom-verbe) en kurde (sorani).” Études Kurdes 9: 83–100. Vosté, J.-M. 1945. “Paul Bedjan, le lazariste persan.” Orientalia Christiana Periodica 11: 45–102.

CHRISTIAN SALAMAS AND JEWISH SALMAS: TWO SEPARATE TYPES OF NEO-ARAMAIC1

HEZY MUTZAFI 1. INTRODUCTION

It is common knowledge in the field of Neo-Aramaic that the dialectal diversity of NENA (North-Eastern Neo-Aramaic) is not only based on geographical distribution, but is also very much related to communal, ethnolinguistic factors, namely to the confessional affiliation of NENA speakers as Jews or Christians. 2 In no location throughout the entire NENA-speaking area is there a case of Jews and Christians speaking one and the same dialect. Their co-territorial dialects are rather always distinct from one another on the levels of phonology, morphology and lexicon, and in some regions also in syntax. Such confession-based dialectal differences are far more substantial in the eastern and south-eastern peripheries of NENA, viz. in areas where Trans-Zab Jewish dialects3 overlap (or rather overlapped) with Christian NEResearch on the Jewish dialect of Salmas was supported by the Israel Science Foundation (Grant No. 339/11) and is partly derived from my work on an etymological dictionary of the Jewish Neo-Aramaic (JNA) dialects. The data in this paper are primarily based on fieldwork on various NENA dialects and Neo-Mandaic (NM). Fieldwork was carried out on Christian Salamas (C.Sal.) in the Republic of Georgia and on Jewish Salmas (J.Sal.) in Israel. All my informants were born in Tbilisi, except one Salamas-born C.Sal. speaker. Western Neo-Aramaic (NA) data are extracted from Prof. Werner Arnold, personal communication. 2 Already a century ago Rhétoré suggested that “les divers parlers soureth peuvent se diviser d’abord en parlers chrétiens et en parlers juif; puis les uns et les autres peuvent se distinguer en parlers de la plaine de Mossoul, parlers de la montagne de Hakkari, et parlers de la Perse.” (Rhétoré 1912: ix). For recent treatments of this point see, inter alia, Hopkins (1999: 321–322) and Jastrow (2002: 368). 3 For Trans-Zab Jewish Neo-Aramaic see Mutzafi (2008). 1

290

HEZY MUTZAFI

NA ones, as in the district of Salmas (salmāś )4—or Salamas, as it is called by the Assyrian Christians—in Iranian Azerbaijan.5 The aim of this paper is to demonstrate a radical case of a Jewish-Christian NENA dialectal split by highlighting some of the major phonological, morphological and lexical differences between the two NENA ethnolects which were both spoken in the district of Salmas until early in the 20 th century. Even a cursory perusal of Duval’s texts of these two dialects6 is sufficient for the reader to realize that they are two extremely different types of NENA, so different that it is difficult to imagine that speakers of the two varieties could freely communicate in Aramaic with one another,7 unless they had been intensely exposed to the spoken idiom of the other ethnic group. Therefore, we can safely regard these dialects as representing two separate NENA languages. J.Sal. is part of the Jewish Azerbaijan dialect-cluster or language,8 which is itself part of the much larger Jewish Trans-Zab continuum. On the other hand, C.Sal. is akin, in quite a few features, mostly to the NENA varieties of the areas of Van and northern Hakkari across the Turkish border, such as Ṣara, Timur, Gawar, Jilu and Dez, as well as to the Christian Urmi dialect cluster spoken to the south of Salmas. Both NENA dialects of Salmas are characteristically innovative in comparison with many other NENA dialects, in particular those in the western parts of Hakkari (e.g. Ṭyare, Tkhuma), Cudi Dağ, northernmost Iraq and the environs of Mosul. Yet J.Sal. appears to be even more innovative, being one of the most progressive dialects in the entire spectrum of Neo-Aramaic varieties. Until 1918 Jewish and Christian NENA-speakers lived side by side in the region of Salmas. The Jews lived in the small town of Kuhneh Shahr (J.Sal. name: salmāś , presently Tazeh Shahr), a few miles west of Dilman (presently named Salmas), the capital town of the district; whereas the local Christians lived mostly in the large village of Khosrava (official name: Khosro Abad), south-east of Dilman, as well as in Dilman itself and in a number of other villages nearby.9 Notes on transcription: Stress is penultimate unless otherwise indicated. The sign ̭ marks an unaspirated consonant and uppercase + marks word-emphasis. Apart from the short vocalic phoneme ə, the vowels i, e, a, o, ɔ and u are generally short in closed syllables and open unstressed ones, and long otherwise. Long vowels in positions other than these are indicated by a macron. 5 Nowadays Salmas is a county in the Iranian province of Western Azerbaijan. 6 See Duval (1883). 7 As already stated in Hopkins (1999: 321). 8 The best known dialect of this group is J.Urmi, see Khan (2008). 9 See Wilmshurst (2000: 325–328) for historical and demographic details on the Assyrians 4

CHRISTIAN SALAMAS AND JEWISH SALAMAS

291

In 1918 the district of Salmas was ravaged by events related to the First World War, and by the end of that year it had been completely depopulated of its Aramaicspeaking inhabitants. The assassination of the Assyrian Patriarch by the local Kurdish chieftain Simko, which took place in Kuhneh Shahr in 1918, stirred heavy Assyrian-Kurdish fighting and the frenzied killing of numerous non-Christians, including dozens of Jews, by vengeful Assyrians and Armenians. Soon afterwards, an invasion of overwhelmingly powerful Ottoman forces uprooted the Assyrian population, which fled southward to Urmi.10 Many of the surviving Jews fled northward and settled in Tbilisi, where some of the Assyrians had found refuge earlier during the war. According to my informants, a small number of Assyrian and Jewish families repatriated to Khosrava and Kuhneh Shahr, respectively, after the war. In 1930, however, the tiny Jewish community of Kuhneh Shahr left for Urmi in the aftermath of a devastating earthquake, and about 20 years later joined the emigration of most of the Jews of Urmi to Israel. The Salmasi Jewish families who had settled in Tbilisi and had not repatriated were deported to Kazakhstan in 1950 and established their own community in Alma-Ata (Almaty), where some J.Sal. speakers are still to be found to this day. As for the Assyrians, according to my informants, some still lived in Khosrava at least as late as 2005, but I have been unable to ascertain whether these were descendants of native Khosrava Assyrians, and whether the unique Christian Neo-Aramaic dialect of the region is still spoken today on its native soil. Both the Jewish and Christian dialects of Salmas are presently rare in the extreme and are highly endangered. It is now hardly possible to find reliable speakers of J.Sal. in Israel, and it was likewise difficult to find good speakers of C.Sal. in Tbilisi when I visited there for fieldwork in 2005. In what follows, I shall discuss some salient differences between C.Sal. and J.Sal. Cognate forms in other Neo-Aramaic varieties are adduced so as to represent antecedents of the current forms in the dialects of Salmas.

of the Salmas district until 1913 (the town ‘Salmas’ in chart No. 36 refers to Dilman), as well as the map at http://aina.org/maps/urmiamap.htm (where “Khoskawa” should be read “Khosrava”). 10 For a description of these events from an Assyrian perspective see Werda (1924: 120– 134, 149–155 and 159–162). The destruction of the Jewish community of Kuhneh Shahr in 1918 is described in the diary of Senyor Demirel: see http://sinyor.demirel.co.il (in Hebrew).

292

HEZY MUTZAFI

2. PHONOLOGY 2.1. Reflexes of *ḏ

The reflex of *ḏ is d in C.Sal., whereas in J.Sal., as a Trans-Zab dialect, reflex l, as in: Ṭyare nəqḏa rquḏ

C.Sal. nəqda rqud

J.Sal. nəqlá rqül

*ḏ has the

‘thin’ ‘dance (sg.)!’

A marginal reflex in J.Sal. is d, which occurs in a few words, such as: Ṭyare ʾiḏa ʾawəḏ

C.Sal. ida avəd

J.Sal. idá ȫd

‘hand’ ‘he might do’

Additional cases of *ḏ > d in J.Sal. are found in hudaá ‘a Jew’ and dəxde (< *d-ʾəxḏaḏe) ‘one another’.

2.2. Reflexes of *ṯ

Whereas in J.Sal., being a Trans-Zab dialect, the regular reflex of *ṯ is l, C.Sal. has lost the interdental phoneme ṯ, presumably through intermediate h (*ṯ > *h > ∅). The postulated shift *ṯ > h occurs in closely related dialects such as Baz, Jilu, Ṣara and Dez.11 Betanure maṯa kusiṯa sahḏuṯa

C.Sal. mā kusiya sāduvva

J.Sal. malá ksilá sahlül(ġ)á

‘village’ ‘hat’ ‘testimony’

In C.Sal. The off-glides y and *w > *ww appeared after i and u, respectively, and at a later stage *ww shifted to vv. The odd-looking ending -ülġá beside expected -ülá in J.Sal. will be explained in § 3.1. below. Apart from the regular reflexes of *ṯ, both dialects evince marginal irregular reflexes. These are ç (Ich-Laut) and t in C.Sal.12 and d in J.Sal.,13 as in the following examples: Cf., e.g. Baz *maṯa > măha ‘village’, Ṣara *ṯele > hilə ‘he came’. The aberrant reflex t was possibly induced by contact with the neighbouring Urmi dialect-cluster. 13 Presumably via the the process *ṯ > *ḏ > d. 11 12

CHRISTIAN SALAMAS AND JEWISH SALAMAS Betanure ʾaṯe ṯele moṯele

C. Sal. ati çili muli

J. Sal. adé ədyele mdele

293

‘he might come’ ‘he came’ ‘he brought’

C.Sal. ç and t both occur in the verb ∅.ç.y. ~ ∅.t.y. ‘to come’, where t is found in the 3sg.m. and pl. inflections of the present base, viz. 3sg.m. ati, 1pl. atax, 2pl. ātitux (!), 3pl. ate, and ç occurs in all other forms of this verb, e.g. ki-açya ‘she comes’, ə́ççəli ‘he has come’, çítəla ‘she has come’ (see further forms in § 3.9. below). Additional cases of C.Sal. *ṯ > t are p.t.x. ‘to open’, k.t.v. ‘to write’, ktava ‘book, letter’, m.y.t. ‘to die’, mota ‘death’ and +kimətra ‘pear’. In J.Sal. *ṯ > d is found in the verbs ∅.d.y. ‘to come’ (see § 3.9. below) and m.d.y. ‘to bring’, as well as in šəndá ‘sleep’ (noun) and arməldá ~ almərdá ‘widow’.

2.3. Suprasegmental Emphasis

J.Sal. has lost the trait of word-emphasis (suprasegmental velarization), which is still preserved in its closest congener J.Urmi and also occurs in C.Sal., as in: C.Sal. amra + bəzza + susiya +

J.Urmi + amrá + bəzzá + susultá

J.Sal. amrá bəzzá susəltá

‘wool’ ‘hole’ ‘plait, pigtail’

J.Sal. appears to be the only Neo-Aramaic dialect that has lost emphasis altogether.14 This loss does not seem to be very old, since the texts published by Duval in 1883 clearly attest to preservation of emphasis.15 The loss of emphasis in contemporary J.Sal. can plausibly be attributed to the impact of Georgian, which would explain the absence of this unique J.Sal. trait in Garbell’s description of Jewish Azerbaijan Neo-Aramaic, based on informants from Iran (Garbell 1965), but I would

This is already evident from the J.Sal. text recorded in 1963 in Alma Ata by Gabrielova (1969), e.g. p. 50/2 šultoná ‘king’, p. 50/44 salmá ‘face’. 15 Duval’s transcription marks emphasis not only for etymological ṣ and ṭ, but also for nonetymological emphasis in certain words. His notation of emphasis in these words is corroborated by cognates in other Tras-Zab dialects, e.g. in ṭlāʾ́ a ‘three’ (107/16, 135/21), corresponding to Arbel ṭḷahá, J.Urmi +tahá; in ṭáḫya (107/16) ‘lower’, corresponding to Rustaqa ṭəxyá; and the ṭ in ṭrṓṣa ‘true’ (118/16 et passim), corresponding to J.Urmi +trosá. These correspondences exclude the a priori unlikely possibility that Duval marked ṣ and ṭ according to etymology even though emphasis had already been lost in J.Sal. 14

294

HEZY MUTZAFI

not rule out the possibility that emphasis has already been obsolescent due to contact with Azerbaijani sometime between 1883 and 1918.16 Two J.Sal. features, however, are vestiges of an earlier phase where suprasegmental emphasis was still retained: The first vestige is a shift of long a to o in formerly emphatic words, as in the following examples: J.Urmi raba [ˠˈrɑ:bɑ] + xyatá [ˠxjɑ:ˈtɑ] +

Early J.Sal. *+roba *+xyotá

J.Sal. roba xyotá

‘many, very’ ‘to sew’

This feature does not apply to the 3sg.m. present inflection of neo-pəʿal (stem I) such as palə́t ‘he goes out’, tašé ‘he hides, keeps’ (cf. J.Urmi +palət, +taše). These forms preserve the vowel by analogy with the rest of the paradigm, which includes forms like 3sg.f. paltá, 1pl. paltéx. The second vestigial feature involves the occurrence of the back vowels o, u in formerly emphatic words as against their front equivalents ö [ø], ü [y] in historically non-emphatic words, as is shown in the following minimal pairs: J.Urmi17 kösá : +kosá küpá : +kupá

J.Sal. kösá : kosá küpá : kupá

‘beardless’ : ‘bun of hair’ ‘hunchbacked’ : ‘clay vat’

There are also cases of ö vs. o in J.Sal. in which o stems from long a in emphatic words, as in: J.Urmi J.Sal. yömá : +yamá yömá : yomá höl : +hal ~ hāl höl : hol

‘day’ ‘give (sg.)!’

: ‘sea, lake’ : ‘until’

Garbell (1965) lumped J.Sal. together with other closely related dialects like J.Urmi where emphasis is not lost. It is unlikely that she would have failed to detect lack of emphasis had she encountered such a feature, yet it might be that Garbell was insufficiently acquainted with the dialect, having hardly worked with J.Sal. speakers, hence no name of a J.Sal. infor mant is mentioned on p. 16 and no J.Sal. text or specific grammatical trait is furnished. Another possibility is that Garbell’s J.Sal. informant, being most probably from Urmi (like all, or almost all, Salmasi Jews in Israel in the 1950s and 1960s), exhibited a highly ‘U rmized’ idiolect of J.Sal. 17 The following J.Urmi words kösá and küpá are represented with a narrow transcription (a broader, phonemic transcription would be kosá and kupá). 16

CHRISTIAN SALAMAS AND JEWISH SALAMAS

295

By having discarded the feature of suprasegmental emphasis, J.Sal. has practically adopted Turkic-like back-front vowel contrasts as regards ö ‒ o and ü ‒ u.

2.4. Reflexes of *ū and *əw < *iw A well-known hallmark of C.Sal. is the epenthetic fricative velar consonants x or ġ after u < *ū, including *ū which stems from *əw < *iw. The nature of the intrusive consonant as x or ġ is dependent on the following consonant’s being unvoiced or voiced, respectively.18 J.Sal. is much more conservative in this respect, retaining separate reflexes of Proto-NENA *ū and *iw, as in the following examples: Ṭyare suse duša19 nura tiwta xliwle

J. Sal. süsé düšá nürá ytəvtá xləvle

C. Sal. suxsa duxša nuġra tuxta20 + xluġli

‘horse’ ‘honey’ ‘fire’ ‘seated (sg.f.)’ ‘he milked’

The origin of this feature appears to be a fortition of y which was part of a former diphthong *uy. This diphthong is typical of C.Urmi, as in suysə ‘horse’, nuyra ‘fire’. Exceptional cases in C.Sal. whereby *ū yields u occur when u precedes q in šuqa ‘market’ and imuqa ‘deep’, and another exceptional case is duka ‘place’.21 The fricative velar intrusive consonant occurs in loanwords as well, e.g. in tuġman ‘tuman (toman), Persian monitary unit’, butuġli ‘bottles’, pl. of butul. Note that u which is a regular reflex of *o (see § 2.5. below) is never followed by ġ/x, hence, e.g. *mbaṭṭole >+buṱuli ‘to annul, cancel’22 vs. butuġli ‘bottles’. An irregular case is *bnawši > *bnəwši > *bnuši > bnuxši ‘he himself, he alone’ (and likewise bnuxš- with other pronominal suffixes).

18

For further treatment of C.Sal. ux/uġ feature see Polotsky (1961: 11–14). The diphthong *əw in forms like *dəwša ‘honey’ was contracted to u already in Proto(or early) NENA. 20 A homophone of tuxta ( u discussed in §2.5. below. 22 For further examples see ibid. 19

296

HEZY MUTZAFI

2.5. Reflexes of *e, *o

Erstwhile *e and *o were raised to i and u, respectively, in C.Sal., whereas J.Sal. has preserved these vowels (except for the fronting of o in accordance with §2.3. above), as in: Ṭyare dewa mnela xabuše

J.Sal. devá mnela xabüšé

C.Sal. diva mnila xābuxši

‘wolf’ ‘she counted’ ‘apples’

Ko (Ṭyare) smoqa sʾora ṛəmmone

J.Sal. smöqá zörá armoné

C.Sal. smuqa sura + rumuni23

‘red’ ‘small, young’ ‘pomgranates’

2.6. *aC: > *eC > iC

Unlike J.Sal. where *a is preserved in all positions, C.Sal. evinces a raising of a before a geminated consonant prior to de-gemination. This raising was initially to e, and remains so in closely related dialects that did not undergo the shift *e > i, such as Sat, Jilu and Gawar, where *aC: yielded eC. In C.Sal. and other closely related dialects, such as Timur, Dez and Nudez, e which stemmed from *aC: was further raised to i as part of a general shift *e > i discussed in the previous paragraph, thus *aC: > *eC > iC. Consider the following comparison: Neo-Mandaic kakkɔ bərattɔ mšadder

J.Sal. kaká bratá šadə́r

Sat keka breta šedər

C.Sal. kika brita šidər

‘tooth (NM: molar tooth)’ ‘daughter’ ‘he might send’

In one attested case e which stems from *aC: was not raised to i, as is shown below: Western NA marra

J.Sal. mará

Sat + mera

C.Sal. + mera24

‘spade, hoe’

Furthermore, a has remained unraised in C.Sal. in the following attested cases: 23

Based on an informant, whereas Duval 14/9 rammounyáiḣ stands for the form rammunyāyç or +rammunyāyç. 24 Apart from eliciting this word from an informant, it is attested in Duval (1883: 16/1) as mḗra.

CHRISTIAN SALAMAS AND JEWISH SALAMAS

297

1. Before *rʿ > rr in a few nouns, as in: Western NA ʾarʿa ṯarʿa

J.Sal. ará tará

Sat + ʾera + tera

C.Sal. + arra + tarra

‘land, ground’ ‘door’

Additional such nouns in C.Sal. are +karra ( qatulá ‘cat’. The restriction of the form -ülġa to the abstract ending renders the assumption that the origin of the J.Sal. intrusive ġ is the Azerbaijani equivalent suffix all the more compelling.

3.2. 3SG Independent Pronouns Ṭyare dialects ʾahu, ʾāw ʾahi, ʾāy

C.Sal. āv ~ avən āy ~ ayən

J.Sal. ö ö

‘he’ ‘she’

J.Sal., as a Trans-Zab dialect, has lost gender distinction between the 3sg.m. and 3sg.f. independent pronouns, since the former was generalized as a 3sg.c. pronoun. Other Neo-Aramaic varieties, including C.Sal., retain this distinction, except the dialect of Challa, which evinces the 3sg.c. forms ʾaya ~ ʾā. The short C.Sal. alternants are clearly derived from ʾahu and ʾahi, whereas the long alternants include a final syllable ən which is of uncertain etymology. J.Sal. ö is a contraction of āw.

3.3. Third Person Pronominal Suffixes

The 3sg. person pronominal suffixes are reflexes of m. -ew and f. -aw in both dialects, yet C.Sal. (as also C.Urmi) contracted these to -u and -o, respectively, whilst J.Sal. exhibits the uncontracted forms -év and -áv. The 3pl. forms represent two different precursors: J.Sal. -ú goes back to the older 3pl.m. form *-un (as still in C.Urmi) or *-ohun (as marginally in Lishana Deni alongside -u), whereas C.Sal. -é goes back to the erstwhile 3pl.f. form *-ayhen, which still exists in various Christian NENA dialects, such as Marga (Hakkari): Marga tawrew tawraw tawrayhen

Arbel toréw toráw torú

J.Sal. törév töráv törǘ

C.Sal. toru toro toré

‘his ox’ ‘her ox’ ‘their ox’

CHRISTIAN SALAMAS AND JEWISH SALAMAS

299

3.4. Present Copula

C.Sal. retains an older paradigm of the present copula closer to the postulated ProtoNENA dichotomy of w-allomorph in the 1st and 2nd persons and l-allomorph in the 3rd person, whereas in J.Sal., as a Jewish Azerbaijan dialect, the whole paradigm was levelled in favour of the 3rd person l-allomorph: C. Sal. ivən ivan ivax ivət ivat itux ili ila ile ~ ina

J. Sal. ilen(a) ilan(a) ilex(a) ilet(a) ilat(a) iletün ile ila ilü

‘I (m.) am’ ‘I (f.) am’ ‘we are’ ‘you (sg.m.) are’ ‘you (sg.f.) are’ ‘you (pl.) are’ ‘he is’ ‘she is’ ‘they are’

3.5. Reflexes of Neo-paʿʿel (Stem II) Forms Betanure mqoləple mqaləp mqulpa mqaləp mqalope

C.Sal. quləpli qiləp qulpa qiləp qulupi

J.Sal. qləple qalə́p qlipá qlüp qalöpé

‘he peeled’ ‘he might peel’ ‘peeled’ ‘peel (sg.)!’ ‘to peel’

Both dialects evince an elision of the initial m- of neo-paʿʿel, yet each dialect also exhibits its own innovations. The main ones in C.Sal. are the rule of raising *aC: >*eC > iC in the present base and imperative, and vowel assimilation in the infinitive, as in *mqallope > *qilupi > *qulupi. J.Sal., being a Jewish Azerbaijan dialect, has innovative past (qləple), perfect (qlipá) and imperative (qlüp) forms which were spawned by full analogy with their counterparts in neo-pəʿal (stem I).

3.6. 3PL Subject Ending of the Present Base Ṭuroyo nəšqi ḥozən

Ṭāl našqi xazay

C.Sal. našqi(ç) xaze

J.Sal. našqí ~ našqeni xazeni

Neo-Mandaic nɔšqen hɔzen

The modern 3pl. endings of the present base are derived from pre-modern Aramaic participles such as nāšqīn ‘kissing (m.pl.)’ and, in III-y verbs, *ḥāzayn ‘seing (m.pl.)’. Closely related forms still survive in various Neo-Aramaic varieties, e.g.

300

HEZY MUTZAFI

Ṭuroyo nəšqi ‘they kiss’, ḥozən ‘they might see’. C.Sal. and J.Sal. evince markedly different forms of these endings. In C.Sal. alongside našqi we find an alternant with final ç, našqiç, which reflects a general tendency for fortition of final *ī (cf., e.g. kəlbi ~ kəlbiç ‘my dog’), apparently via diphthongization to *iy. This final ç appears to be more common in pause. The C.Sal. III-y ending -e, as in xaze, reflects a contraction of -ay (cf. Ṭāl xazay). In J.Sal. the III-y form is a combination of -en (< *ayn) and the i of the rest of verb classes, hence, e.g. xazeni ‘they see’. This final -eni in III-y verbs (typical of many Trans-Zab dialects), has spread in J.Sal. to all other verbal classes as a unique free variant of erstwhile -i (< īn) in verbs such as našqí ~ našqeni ‘they kiss’. Note that a similar analogy and levelling occurred in NeoMandaic, where the III-y ending, as in hɔzen ‘they might see’, has spread to all other verbs, hence, e.g. nɔšqen ‘they might kiss’.

3.7. Positive and Negated Present and Future C.Sal. ki-nášəq → b(əd)-nášəq →

lé-našəq lé-našəq

J.Sal. našə́q b-našə́q

→ →

lá-našəq b-lá-našəq ~ lá-našəq

C.Sal. has indicative and future prefixes, the former being ki- < *k + *y, and the latter is b(əd)-. The negation of both C.Sal. present and future conjugations evinces a common form with the negator le. The latter has emerged from la followed by the element y, which is combined in the indicative prefix ki-. In J.Sal. the indicative prefix k- has generally been lost,27 whereas the future prefix b- has been retained. The negator is la, and the negated future mostly involves a curious insertion of the negator between the future prefix and the verb, as in b-lánašəq ‘he will not kiss’. The negated future without b- is quite rare in my data.

3.8. Habitual Past and Continuous Past C. Sal. ki-+raxṱáva + bərxáṱəva28

J. Sal. raxotá-vela raxotá-vela

‘she used to run’ ‘she was running’

C.Sal. has two separate conjugations to express habitual vs. continuous past, the former with the present base converted to past by the suffix -va and the latter

27

eats’.

28

It survives only with a handful of weak first radical verbs, such as ∅.x.l. ‘eat’, kxəl ‘he Also +brə́xṱəva, with the gerund based on the form +rəxṱa.

CHRISTIAN SALAMAS AND JEWISH SALAMAS

301

involves the gerund and past copula (< *bərxaṭa-iwawa). In J.Sal., on the other hand, habitual past, inherited from Proto-NENA, succumbed to the late NENA construction of gerund with past copula which was originally restricted to expressing the continuous past.

3.9. Selected Irregular Verbs Two selected verbs, denoting ‘to know’ and ‘to come’, have been chosen to demonstrate the fact that more often than not cognate verbal forms stemming from the same Aramaic etyma have undergone strikingly distinct developments in each of the two dialects in question. For instance, both C.Sal.+dili ‘he knew’ and J.Sal. yəlle ‘he knew’ evolved from *yḏiʿle, but, among other changes, whereas in C.Sal. the first radical y has been elided, in J.Sal. it has been retained, and, moreover, the J.Sal. form has undergone metathesis, viz. *yḏiʿle > *yʿiḏle > *yʿille (cf. Jewish Koy Sanjaq yʾille) > yəlle ‘he knew’. Another example is the distinct imperative forms of the verb ‘to come’, namely ça in C.Sal. and idá in J.Sal. Both these forms stem from *iṯa (cf. Jewish Dohok yṯa). In C.Sal. *ṯ possibly first shifted to h, yielding *iha, and then h was partially assimilated to i before the latter was elided. Unlike C.Sal., J.Sal. keeps the initial i and also displays an aberrant shift of *ṯ to d, quite possibly through the process *ṯ > *ḏ > d. ‘to know’

Past Resultative Irrealis Present Imperative Infinitive Pres. Prog. ‘to come’ Past Resultative Irrealis

29

C.Sal. dili + diya + yāy + ki-yāy + di ~ +dīç + daya + bidāyli

J.Sal. yəlle yilá ayyə́l kyəl yül ayölé ayölé

C.Sal. çili əçça29 ati

J.Sal. ədyele ədyá adé

+

< əçya, attested in Duval (1883: 6/15).

302

HEZY MUTZAFI Present Imperative Infinitive Pres. Prog.

ki-ati ça çā30 biçāyli

gde idá idaá idāyle

4. LEXICON 4.1. Selected Lexical Differences C.Sal. dəxya šipira pā + ruxša qdala nāxira bázzuġni zvana šqala

J.Sal. qlivá sqilá salmá kpaná pqará pöqá šətqel šaqölé antöé

‘clean’ ‘beautiful’ ‘face’ ‘shoulder’ ‘neck’ ‘nose’ ‘last year’ ‘to buy’ ‘to take’

The list above contains only a small selection of inherited Aramaic words out of many lexical differences between the two dialects.

4.2. Selected Cognate Nouns C. Sal. + suppa nāviya xā təlgəssar tinuġra misayya

J. Sal. spotá nholtá xalüntá tresár tanürá msalá

‘finger’ ‘ear’ ‘sister’ ‘twelve’ ‘oven’ ‘scales (for weighing)’

Many C.Sal.–J.Sal. cognates look worlds apart, as in the case of the cognate forms denoting ‘ear’ in these dialects, both ultimately stemming from the plural form

30

çyā.

Thus according to one of my informants, but in Duval 1883: 12/16, 25/9 the form is

CHRISTIAN SALAMAS AND JEWISH SALAMAS

303

*ʾeḏnāhāṯā > 17th century NENA texts nhaṯa.31 The latter is the precursor of J.Sal., whereas the C.Sal. from derives from the intermediate antecedent *naṯiṯa.32 As for the cognate words denoting ‘twelve’, J.Sal. has a transparent form tresár, whereas C.Sal. displays a bewildering and unique oddity təlgəssar, according to my informants, and təlyisar or təlyesar according to Duval.33 These C.Sal. forms await satisfactory elucidation as to the processes which led to their emergence.34

4.3. False Friends

Finally, the following are a few false friends (with their Aramaic etyma in parentheses) that further demonstrate the enormous gaps between the two ethnolects: p-l-x š-l-y + r-v-y, r-v-y xaé šita, šitá tina, tiná bəzza, bəzzá

C.Sal. to work (*p-l-ḥ) to calm down (*š-l-y) to be drunk (*r-w-y) their sister (*ḥāṯayhen) year (*šattā) fig (*tenā) breast (*bezzā)

J.Sal. ‘to open (*p-ṯ-ḥ)’ ‘to card wool (*š-ḏ-y)’ ‘to grow up (*r-b-y)’ ‘life (*ḥayyē)’ ‘plastered.SG.F (*šiʿtā)’ ‘mud (*ṭinā)’ ‘hole (*bezʿā)’

REFERENCES

Duval, Rubens. 1883. Les dialectes néo-araméens de Salamas. Paris: F. Vieweg. Gabrielova, A. 1969. “A Sample of the Aramaic Speech of the Jews of Salamas.” Восточная Филологиа (Philologia Orientalis) 1: 47–58. [In Georgian] Garbell, Irene. 1965. The Jewish Neo-Aramaic Dialect of Persian Azerbaijan. The Hague: Mouton. Gottheil, Richard. 1893. “The Judaeo-Aramaean Dialect of Salamas.” Journal of the Americal Oriental Society 15: 297–310. Hopkins, Simon. 1999. “The Neo-Aramaic Dialects of Iran.” In Irano-Judaica IV, edited by Shaul Shaked and Amnon Netzer, 311–327. Jerusalem: Ben-Zvi Institute.

See Mutzafi 2005. For *nhaṯa and *naṯiṯa see ibid., 235–236 and 237–238, respectively. 33 Duval 1883: 27/3 tilyisár, 56/5 tilyesár. 34 The form təlgəssar may well have emerged by analogy with the structure of +təltasar ‘thirteen’, but this suggestion falls short of explaining the enigmatic epenthetic g (Duval y) and the geminated s. A similar case with an epenthetic velar consonant in C.Sal. is mdikta ‘town, city’. 31 32

304

HEZY MUTZAFI

Jastrow, Otto. 2002. “Neo-Aramaic Dialectology: The State of the Art.” Israel Oriental Studies 20: 365–377. Khan, Geoffrey. 2008. The Jewish Neo-Aramaic Dialect of Urmi. Piscataway, NJ: Georgias Press. Mutzafi, Hezy. 2005. ‘‘The Reflexes of the Word ‫‘( אדנא‬ear’) in Eastern Neo-Aramaic: Etymology, Diversification and Innovation.” In Samaritan, Hebrew and Aramaic Studies Presented to Professor Abraham Tal, edited by Moshe Bar-Asher and Moshe Florentin, 229–242. Jerusalem: Bialik Institute. [In Hebrew] —. 2008. “Trans-Zab Jewish Neo-Aramaic.” Bulletin of the School of Oriental and African Studies 71 (3): 409–431. Polotsky, Hans Jakob. 1961. “Studies in Modern Syriac.” Journal of Semitic Studies 6: 1–32. Rhétoré, Jacque. 1912. Grammaire de la langue Sureth ou Chaldéen vulgaire. Mossoul: Imprimerie de pères Dominicains. Werda, Joel E. 1924. The Flickering Light of Asia, or the Assyrian Nation and Church (Reprinted in Chicago: Assyrian Language and Culture Classes Incorporated 1990). Wilmshurst, Davis. 2000. The Ecclesiastical Organisation of the Church of the East, 1318–1913. Leuven: Peeters.

SOME FEATURES OF THE GAZNAX DIALECT (SOUTHEAST TURKEY)

1

ARIEL GUTMAN 1. INTRODUCTION

Gaznax2 is the native name of a Chaldean village located in the Şırnak province in south-east Turkey. The name is derived, according to a traditional etymology, from ܽ ܳ ܰ Gazzā d-Noḥ ‘Noah’s Treasure’, in accordance with the belief that Syriac ‫ܓܙܐ ‏ܕܢܘܚ‬ Noah’s ark landed on the nearby Mount Judi (Joseph Alichoran, personal communication). It was one of 8 Chaldean villages in this region.3 The inhabitants of these villages all spoke similar dialects of Neo-Aramaic, and were thus grouped together by Sinha (2000) as the Mount Judi (Cudi Dagı ̆ in Turkish) dialects. A first-hand de-

I am grateful to members of the Yaramis family in France for sharing with me their knowledge of their dialect: Ciko, Herman, Isa, Joseph, Nouri, Paul and Zackarie Yaramis, as well as the late Memo Yaramis. Thanks are also due to Isa Hamdo and his family, for providing me with data on the Harbole dialect. Particular thanks go to Joseph Alichoran, lecturer of Soureth in the INALCO (Paris) and a native speaker of Neo-Aramaic, for the invaluable information he provided me with, and for his immense help which made the transcription of the interview with Memo Yaramis possible. I am also grateful for the comments of Dr. Eleanor Coghill and Prof. Eran Cohen on drafts of this paper. Finally, I would like to thank the participants of the Neo-Aramaic Dialectology conference in Jerusalem, where this paper was first presented, for their comments, and especially Dr. Hezy Mutzafi for pointing out some errors. The research was conducted and funded in the scope of the DFG project ‘Neo-Aramaic morphosyntax in its areal-linguistic context’ led by Dr. Eleanor Coghill. 2 This is the linguistic transcription of the name, the last segment being a velar fricative. Common ways of writing the name include Gaznakh or Geznakh. The official Turkish name of the village Cevizaği (Sinha 2000: XV). 3 Sinha (2000: 5) provides a map with the location of the villages. An online map, created by the author, is available at http://tinyurl.com/kaqv7a2. 1

306

ARIEL GUTMAN

scription of these villages is given by Poizat (1986).4 Yaramis (2010), a native of Gaznax, gives an account of its life and folklore. Due to the armed conflict between the Turkish government and the Kurdistan Workers’ Party (the PKK), the Chaldean inhabitants had to leave their villages by the beginning of the 1990’s. Many of them moved first to Istanbul, and then left Turkey. The fate of Gaznax was not different: according to Yaramis (2010: 87), about 10,000 of the descendants of Gaznax live today in Iraq, Russia, the United States and Australia. In Europe, about 1,500 of them live in Belgium, France, Germany, Denmark and Sweden. Quite uniquely, however, some of the inhabitants of Gaznax have been able to return to their village since 2004, as reported by Thiry (2007). The current paper is based on field research which the author carried out with speakers of the Gaznax dialect from the Yaramis family, who live today in two suburbs north of Paris, namely Sarcelles and VilliersLe Bel. The number of Gaznax speakers in this area is estimated to be around 400 (Yaramis 2010: 87), out of 10,000 Christian Aramaic speakers in the Parisian suburbs, most of which come from the Judi villages (Alichoran and Sibile 2013: 873). As mentioned, the dialect of Gaznax is quite similar to other Judi dialects, and in particular to the Bēṣpən dialect described in detail by Sinha (2000).5 While Sinha does give some information regarding the variation among the other village dialects, she has little information on the dialect of Gaznax (Sinha 2000: 9ff.), since she had no recordings of it (personal communication). Nonetheless, this dialect merits attention due to its peripheral location amongst the Judi dialects, standing between them and the former Hakkari dialects. Indeed, the speakers of Gaznax clearly perceive it as distinct. In the diaspora situation, however, the speakers who were interviewed live in close proximity to speakers of other Judi dialects. The influence of the other dialects on their speech has clearly grown, leading to a blurring of the dialectal differences, especially in the younger generation, which was not born in Gaznax. The aim of the current paper is, thus, to elucidate some of the features of the Gaznax dialect, and, where possible, to contrast them with features of other Judi dialects.

Poizat included in his survey also some other villages, notably Arivan, or Hertevin, whose quite different dialect is described by Jastrow (1971 and 1988). Poizat also mentions Deran, Djenet, and Birinji, on which I have no further information. 5 The supra-dialect taught in Poizat (2008) is in fact also quite similar to the Judi dialects, as it is based partly on the speech of Joseph Alichoran, who is fluent in the Harbole dialect (Poizat, personal communication). 4

SOME PECULIARITIES OF THE GAZNAX DIALECT

307

2. PHONOLOGY

Figure 1: The vowel system of Gaznax Phoneme i u e o a ă

Tense [i(ː)] [u(ː)] [e(ː)] [o(ː)] [ɑ(ː)] -

Lax

[ɪ] [ʊ] [ɘ] [o] [æ] [æ]

Table 1: Vowels of Gaznax and their realisation The phonological system of Gaznax is similar, if not identical, to that of Bēṣpən. Thus, the consonants are those given by Sinha (2000: 48 ff.). It is worth noting that Gaznax is an /x/ dialect, as the former Aramaic /ḥ/ phoneme has shifted to /x/, fusing thus with the [ḵ] allophone of /k/, of which the allophony is no more productive (cf. Sinha 2000: 51). As for the vowel system, we need to clearly distinguish between the phonetic and the phonological inventories, contrasted in Figure 1. As we can see, Gaznax has only 6 vocalic phonemes, but different phonetic realisations of these. For clarity, the phonemes are repeated in Table 1 together with their main realisations. Generally speaking, vowels in an open syllable are realized as a tense and long allophone, while in a closed syllable they appear as a more centralized allophone of short duration.6 These distinctions are in general not phonemic, except marginally in 6

The two variants are termed “long” and “short” by Sinha (2000: 35ff.) and elsewhere, but

308

ARIEL GUTMAN

the opposition /a/~/ă/. The latter short vowel can appear in stressed open syllables, especially in verbal forms of weak roots. We find oppositions such as măra ‘it hurts’ vs. mari ‘my Lord’. The phone [ɛ] only appears as a reduced variant of the diphthong /ay/, which is regularly realized as [ɛj]~[ɛɪ]~[ɛ] or sometimes as [aj] (usually near an emphatic consonant). It has no distinct phonemic status. The phones [y] and [ø] appear only in a handful of Turkish loanwords (such as the country name Türkiye ‘Turkey’ or the word öğretmen ‘teacher’, for which a native word malpana exists as well), and are not part of the native system.

3. DISTINCTIVE LEXICAL ITEMS

Some lexical items of Gaznax differ from those in the other Judi dialects. Table 2 shows some of these differences, contrasting the Gaznax dialect with the Bēṣpən and Harbole dialects.7 Gaznax ‘man’ zalame ‘brother’ xona ‘brothers’ xunwata ‘frog’ peqa ‘sheep’ p̣arạ ‘shepherd’ raya ‘red’ smoka ‘all’ kun ‘want’, pres. base bay‘go’ r-x-š

Bēṣpən/ Harbole gawra/ gura (Har.) axona axawata pŭqa barxa šwana (< Kurdish) smoqa kul(e) kibbʾ-z-l

Table 2: Distinctive Lexical Items of Gaznax

at least in the context of Gaznax such names would be somewhat misleading, since the tense allophones are generally realized without the length feature [ː] in an unstressed syllable. Note also that closed mono-syllabic words (not clitics) always have a tense and long allophone as their vocalic nucleus. 7 The data on Bēṣpən come from Sinha (2000). The data from Harbole come the author’s fieldwork. Some lexical differences were pointed out to me by the Gaznax speakers. In addition to these differences in form, there are also identical lexemes which differ only in grammatical gender.

SOME PECULIARITIES OF THE GAZNAX DIALECT

309

4. PRONOMINAL SYSTEM

The pronominal system of Gaznax is typical of the Judi dialects. Table 3 presents the independent pronouns, as well as the S-, L- and possessive pronominal suffixes. Note the 2sg. and 3pl. independent forms /ate/ and /an(e)/ respectively, pace Sinha (2000: 69): Ind. 1SG.M 1SG.F 2SG.M 2SG.F 3SG.M 3SG.F 1PL 2PL 3PL

-S

ana ate

-L

-Poss.

-in

-li

-i

-an -it

-li -lux

-i -ux

-lax -le [lɘ] -la -lan -lawxun -lay [lɛ(ɪ)]

-ax -u -aw -an -awxun -ay [ɛ(ɪ)]

-at awa -∅ aya -a axnan -ux axnutin -utin an(e) -i

Table 3: Pronouns and pronominal suffixes of Gaznax

5. COPULAS 1SG.M 1SG.F 2SG.M 2SG.F 3SG.M 3SG.F 1PL 2PL 3PL

Independent Enc.

Past

Neg.

Neg.P.

ʾoli ʾoli ʾolux ʾolax ʾol(e) ʾola ʾolan ʾolawxun ʾolay

inwa inwa itwa itwa iwa iwa ixwa itunwa iwa

lewin lewan lewit lewat lele lela lewux lewutin lele

lenwa lenwa letwa letwa lewa lewa lexwa letinwa lewa

iwin iwan iwit iwat ile ila iyux itun ilay

Table 4: Copulas of Gaznax The copular conjugation resembles in general the other Judi dialect forms, though there are some differences. Table 4 presents the independent, enclitic, past,

310

ARIEL GUTMAN

negative, and past negative copular forms. Those which differ significantly from the Bēṣpən copular forms (Sinha 2000: 144–148) are marked by italics. In contrast to the Bēṣpən forms, the past affirmative and negative copulas do not have a /-wV-/ segment, as for example iwənwa (1sg.m.) or iwanwa (1sg.f.) in Bēṣpən. Consequently, gender distinction disappears for these forms. As for the enclitic plural copulas, these are iwux (1pl.) and iwútən (2pl.) in Bēṣpən.

5.1. Emergence of a Deictic Copula

Gaznax Bēspən

Deictic

General

ho-

ʾo-

ho-

Table 5: Split of the independent copula in Gaznax The independent copula is used in the Judi dialects when talking about the immediate present (or in a narrative present). As such it also participates in the formation of the present progressive verbal tense paradigm (see below). Syntactically, it is the independent counterpart of the enclitic copula. In most Judi dialects, the independent copula is formed by using the base /ho/ + L-suffix. In the Gaznax dialect, however, the corresponding base form is /ʾo/, which is regularly used in the present progressive paradigm, as well as in nominal clauses in the present tense. Nonetheless, the base /ho/ is retained (or possibly innovated by dialect contact) for true deictic uses: (1) ha behold

ho-la COP-3SG.F

‘It is here!’, ‘Voici!’ (2) ho-la COP-3SG.F

ʾaxa here

‘Here she is!’ Thus, in the Gaznax dialect the independent copula has split into two differing forms, as summarized in Table 5.

SOME PECULIARITIES OF THE GAZNAX DIALECT

311

5.2. Sandhi of the Enclitic Copula

In the Judi dialects, the enclitic copula starts with an /i-/ segment.8 Following a singular predicate, this segment is regularly fused to a preceding /a/ vowel, resulting in the phonological diphthong /ay/ which is sometimes realized as the diphthong [ɛɪ], but more often in this context simply as the monophthong [ɛ] (cf. Sinha 2000: 147; Poizat 2008: 33). However, in the Gaznax dialect, the presence of an emphatic consonant near the /ay/ diphthong alters its realisation to [aj] (instead of the regular [ɛ(ɪ)]).9 The effect of the emphatic consonant is, however, blocked in the presence of the feminine suffix /-ta/. As a consequence, the sandhi behaviour of the clitic copula with nouns containing an emphatic consonant depends on the gender of the noun. Table 6 shows the different sandhi patterns with the example of the adjective ṭawa ‘good’.10 Gender Abstract form M F

/ṭawa=ile/

/ṭawta=ila/

Realisation

Sandhi realisation

[ṭɑwajle]

[aj] or [æj]

[ṭɑwtɛla]

[ɛ] or [ɛj]

Table 6: Sandhi patterns of enclitic copula with the adjective ‘good’ As described above, this phenomenon is simply a case of allophony of the /ay/ diphthong. Curiously, according to my observations, some speakers may have started to extend this pattern by analogy to other adjectives, which do not contain an emphatic consonant. In such a case, the allophonic pattern is becoming an allomorphic pattern. To ascertain this claim, however, more exact acoustic measurements are needed.

8

One may wonder whether this segment is simply part of the copular base, or is in fact the indicative prefix i- used in the verbal system, as has been suggested to me by Prof. Bruno Poizat. We note that the synchronic question is independent from the diachronic one, i.e. whether both have the same source. 9 Some variation as to the application of this rule is attested between speakers. 10 The sandhi pattern of the plural copula ilay with plural forms ending with /-e/ is another issue. The sequence /ey/ is sometimes realized as [e], sometimes as [ɛ], and sometimes as [i] (as if the /-e/ ending was simply elided).

312

ARIEL GUTMAN

6. VERBAL CONJUGATION

This section presents some features of the Gaznax verbal system, some of which are unique to this dialect, and some which are shared with other Judi dialects, but deserve special attention.

6.1. The Indicative Present The indicative present is formed using the present base (e.g. šaql- ‘take’) together with an S-pronominal suffix. Additionally, in most Judi dialects, a y- ~ i- prefix precedes the present base, similarly to the prefix k- in some other dialects (cf. Poizat 2008: 78). In Gaznax, however, the indicative prefix y- occurs only before vowelinitial present bases, which are derived from I /ʾ/ roots in the first stem formation.11 Contrast the following two examples: (3) y-atin IND-come.1SG ‘I come’ (4) našqin kiss.1SG ‘I kiss’ Notwithstanding this, the negation of present verbs remains constantly [lɛ], derived from /la/ + /y/, irrespective of whether /y-/ appears before the verb or not: (5)

lɛ

NEG

y-axl-in-ne IND-eat-A.1SG-P.3SG.M

awa DEM.SG.M

‘I do not eat this!’ (6) lɛ

NEG

šat-in-ne drink-A.1SG-P.3SG.M

awa DEM.MS

‘I do not drink this!’

11

In the Bēṣpən dialect, on the other hand, this prefix has a broader distribution. It can appear before any verb of the first stem formation, but not before verbs of other stem formations, which commence with an /m/ segment.

SOME PECULIARITIES OF THE GAZNAX DIALECT

313

6.2. The Present Progressive

Sinha lists 3 present progressive constructions in the Judi dialects (see Table 7). Gaznax speakers use frequently only the construction formed using the independent copula together with the infinitive (e.g. šqala ‘to take’). A second construction, formed using the present base, does appear from time to time but to a lesser degree, and may very well be the product of contact with other Judi dialects. Bēspən Construction

Occurrence in Gaznax

Ind. copula + (b) + infinitive

✓

Ind. copula + present base

rarer

(b) + infinitive + enclitic copula not found Table 7: Present progressive constructions in Judi dialects In contrast to other Judi dialects, the b- prefix (stemming from the preposition b- ‘in’), which precedes the infinitive in the present progressive formation, occurs only before vowel-initial present bases.12 Contrast: (7) ʾo-li COP-1SG

b-itaya PREFIX-come.INF

‘I am coming’. (8) ʾo-li COP-1SG

klaya wait.INF

‘I am waiting’. An interesting variant construction was presented by one speaker, who occasionally omitted the copula altogether. In the following example, note also the optionality of the object pronominal suffix on the infinitive: (9) awa 3SG.M

nšaqa{-w} kiss.INF-{3SG.F}

brat-i daughter-POSS.1SG

‘He is kissing my daughter’. A similar construction was noted in the Barwar dialect (Khan 2008: 726). 12

In the Bēṣpən dialect, the distribution of b- is similar to that of the y- indicative prefix: it appears only before infinitives of the first stem formation.

314

ARIEL GUTMAN

6.3. The Preterite

In Gaznax, as in all Judi dialects, there is no preverbal qam particle (or the like) for forming the preterite (cf. for example Cohen 2012: 458; Khan 2008: 176f.). Consequently, the preterite (i.e. perfective past) is always formed using the preterite base (e.g. šqil ‘took’). A suffixed L-suffix indexes the subject/agent13 of the verb (glossed A). As for the object/patient (glossed P), 3 possibilities exist: For the direct object, the S-suffix can be used for all persons. In the following examples, note the S-suffixes which are glossed as patient (P) markers: (10) nšiq-ux-le kissed-P.1PL-A.3SG.M ‘He kissed us’. (11) nšiq-at-wa-li kissed-P.2SG.F-CONV-A.1SG.F ‘I had kissed you (f.)’ This usage is not restricted to the Gaznax dialect, but occurs in other Judi dialects as well. The following examples, which appear in an exercise of Poizat (2008: 97), are in fact based on true conversation of speakers from the Judi region:14 (12) Yawsep, la J. NEG

xz-et-ti saw-P.2SG.M-A.1SG

tama there

‘Joseph, I did not see you there’. (13) la

NEG

xz-en-nox saw-P.1SG.M-A.2SG.M

men=sabab from=reason

la

NEG

te-li came-1SG

‘You did not see me because I did not come’. These forms, however, are not so common, and did not seem to be entirely intuitive to the speakers during the elicitation sessions. An apparently easier and more frequent possibility is to use the preposition b- (originally meaning ‘in, with’) with a pronominal suffix:

13

The notions of Agent and Patient are used here as semantic macro-roles, much like the Actor and Undergoer in Gutman (2008). 14 The transcription of these examples is normalized according to the system used here.

SOME PECULIARITIES OF THE GAZNAX DIALECT (14) nšiq-li kissed-1SG

315

biy-ux ACC-2SG.M

‘I kissed you (m.)’. (15) nšiq-wa-lax kissed-CONV-2SG.F

b-i

ACC-1SG

‘You had kissed me’. The usage of the b- preposition as an object marker is also attested in the Bēspən dialect: (16) […] b-gawr-at FUT-marry-2SG.F

b-i

ACC-1SG

[…]

‘... you (sg.f.) shall marry me ...’ (Sinha 2000: 212 (181)) The fact that the same argument which can be indexed by an S-suffix is marked by a preposition shows that the b- preposition has been grammaticalized as an accusative marker. As a preposition b- has either a locative, instrumental, or comitative (‘with’) meaning. None of these meanings has been reported as having been grammaticalized as a direct object marker by the World Lexicon of Grammaticalisation (Heine and Kuteva 2002). The path of the grammaticalisation may have gone through a different meaning, though. We note that in some cases b- adds a malefactive sense (MAL) to the argument: (17) galak many

mindyane things

(w)ud-ux-wa do-1PL-CONV

biy-ay MAL-3PL

‘We made many things against them’. Such a use of b- is also known from Amharic (Kane 1990: 853). It may be this sense which served as an intermediate stage on the way to become an accusative marker. We can contrast the use of b- with the preposition il(l)-. In some NENA dialects, the latter serves as a direct object marker (see, for example, Khan 2008: 808 (iv)). In Gaznax, however, it is reserved to mark a syntactically indirect (i.e. oblique) object, which can never be substituted by an S-suffix. Semantically, though, it can sometimes denote the patient of the action: (18) kle-li waited-1SG ‘I waited for him’.

ill-u OBL-3SG.M

316

ARIEL GUTMAN

(19) mxe-lox struck-2MS

ill-i

OBL-1SG

‘You struck me’.

7. GENITIVE CONSTRUCTIONS

In the Gaznax dialect, as in other Judi dialects, the classical Semitic annexation construction with the old construct state is hardly apparent. Instead, three other constructions are available to mark the genitive relationship.15 Morphologically, the head noun may be marked functionally as a construct state noun by the -id ~ -it suffix: (20) awa this

bayt-id house-CNST

bab-i father-POSS.1SG

‘this house of my father’s’ (21) tre two

čant-id sack-CNST

rizza rice

‘two sacks of rice’ In a separate, syntactic strategy, a linking particle /ʾad/ (glossed LNK) may intervene between the head NP and the genitive noun, especially when the head noun is modified additionally by an adjective (cf. Poizat 2008: 61): (22) šula work

zaḥme hard

ʾad

LNK

{d}-awa GEN-this

zalame man

‘the hard work of this man’16 (23) bayta house

raba big

ʾad

LNK

bab-i father-POSS.1SG

‘the big house of my father’

15

We deal here with the genitive construction involving two nouns. For genitive pronouns see Sinha 2000: 70ff. 16 The /d/ segment before awa is difficult to ascertain due to its position after the final /d/ of ʾad. If it is there, it can be seen as genitive case on the demonstrative pronoun, follo wing Cohen (2010).

SOME PECULIARITIES OF THE GAZNAX DIALECT (24) awa this

bayta house

ʾad

LNK

317

bab-i father-POSS.1SG

‘this house of my father’ Additionally, a -(y)e suffix (glossed EZ), probably borrowed from the Kurdish Ezafe, is used when the head and genitive nouns are both proper names, indicating affiliation relationships. (25) Yaqo-ye Y.-EZ

Musa M.

‘Yaqo son of Musa’

8. A TEXT SAMPLE

The following section presents a short text sample, transcribed from an interview made with the late Memo Yaramis, who lived most of his life in Gaznax, until his migration to Istanbul and subsequently to France.17 The facts covered in the text are similar to those related by Yaramis (2010), which is also based on the memories of Memo Yaramis. The transcription and translation of the text would not have been possible without the keen help of Joseph Alichoran. Time permitting, we hope to publish a longer extract of the text. Unless marked otherwise, the word accent is penultimate. Note that clitics (separated by an = symbol) do not normally change the accent position. Intonation group boundaries are simply marked by punctuation marks (comma or final point, according to the context). Note also that initial glottal stops /ʾ/ are systematically omitted in the transcription. The symbol of 3 dots (…) in the text indicates hesitation of the speaker. (1) ana 1SG

šimm-i name-POSS.1SG

Mamo=le, M.=COP.3SG.M

u=bab-i and=father-POSS.1SG

Xammo=le X.=COP.3SG.M ‘My name is Memo, and my father is Khammo’. 17

This interview was conducted by Antoine Yalap and Pierre Palais for a show of Ishtar TV, featuring interviews with elderly speakers from villages of Mardin and Hakkari regions. I am grateful to the copyright holder (© Archives de Samuel Yalap) for granting me the right to use this text for scholarly purposes.

318

ARIEL GUTMAN

(2) u=saw-i and=grandfather-POSS.1SG

Mamluk=ile. M.=COP.3SG.M

‘And my grandfather is Mamluk’.18 (3) Mamluk, šimm-i=ži Mamluk=ile M. name-POSS.1SG=also M.=COP.3SG.M y-imr-i-li IND-say-A.3PL-P.1SG

bas but

Mamo. M.

‘My name is also Mamluk, but I am called Memo’. (4) u=šimm-it and=name-CNST

tot-i grandmother-POSS.1SG

Basse=wa. B.=COP.PST.3

‘And the name of my grandmother was Basse’. (5) u=šimm-it and=name-CNST

saw-i grandfather-POSS.1SG

Mamluk=iwa. M.= COP.PST.3

‘And the name of my grandfather was Mamluk’. (6) bab-it father-CNST

bab-i, father-POSS.1SG

sawi. grandfather-POSS.1SG

‘The father of my father, my grandfather’. (7) u=ana and=1SG

ahl-(i)d people-CNST

Gaznax=iwin. G.=COP.1SG.M

‘And I am from Gaznax’.19 (8) axnan, 1PL

m=qam from=before

u=xamši and=fifty

anne DEM.PL

arb-emma, four-hundred

šinne years

‘We, before four hundred and fifty years,’20

18

Being a loan name (probably from Turkish), this name is pronounced [Mamlʉk]. The CONST suffix is realized as /-d/ in this sentence, probably due to the preceding sonorant /l/. 20 Yaramis (2010) situates this story in the beginning of the 14th century. 19

SOME PECULIARITIES OF THE GAZNAX DIALECT (9) maxket-id speech-CNST

bab-i father-POSS.1SG

Ṭiyare=ixwa T.=COP.PST.1PL

319

min from

tiye. come.RES.PL

‘According to my father, we came from Ṭiyare’. (10) min=Ṭiyare, m=…Ṭiyare from=T. from=T. Gippa G.

ʿelayta, upper,

m=Bnene from=sons

te-lan came-1PL

‘From Ṭiyare, (hesitating) Upper Ṭiyare. We came from the Gippa Clan’.21 (11) duk-id place-CNST

malka=iwa, ay king=COP.PST.3 DEM

duk-id place-CNST

malka=iwa. king=COP.PST.3

‘(From) where the king was’.22 (12) te-lay came-3PL

tḷata three

xunwata: Išo, u=Qaša, … Hawel brothers I. and=Q. H.

‘There came three brothers: Isho, Qasha and (hesitating) Hawel’. (13) ane

DEM.PL

tḷata three

xunwata brothers

te-lay. came-3PL

‘These three brothers came’. (14) madya Išu u=… Qaša, an DEM.PL regarding I. and=Q.

kle-lay stayed-3PL

b=Gaznax. in=G.

‘Regarding Isho and (hesitating) Qasha, they stayed in Gaznax’.23

21

Yaramis (2010: 13) links the Gippa Clan with Upper Tiyare as well. Nineb Lamassu (Cambridge University) pointed out to me, however, that the Gippa Clan was in fact in Lower Ṭiyare. 22 Yaramis (2010: 13) describes the king of Ṭiyare as the ‘Assyrian-Nestorian Duke’. Note that /malka=iwa/ is realized as [malkɛwa]. 23 The word madya ‘regarding’ may possibly be decomposed as ma diya ‘what of it (sg.f.)’.

320

ARIEL GUTMAN

(15) madya Hawel, bre-la da naxošuta l-u regarding H. happened-3SG.F INDF.SG.F disease to-3SG.M ‘Regarding Hawel, he got a disease’. (16) mxe-la struck-3SG.F ʾo-la COP-3SG.F

u=rxiš and=went.3SG.M qurb-id close-CNST

il=… il=dà to=to=INDF.SG.F

ṃata village

Oz Oz

‘It struck him and he went to (hesitating) a village near Oz’.24

REFERENCES

Alichoran, Joseph, and Sibille, Jean. 2013, “L’araméen.” In Histoire sociale des langues de France, edited by Georg Kremnitz, 869–875. Rennes: Presses universitaires de Rennes. Cohen, Eran. 2010. “Marking Nucleus and Attribute in North-Eastern Neo-Aramaic.” Proceeding of the VIII Afro-Asiatic Congress (September 2008, Naples), Studi Maghrebini (Nuova Serie) VII, edited by Sergio Baldi, 79–94. Naples: Università degli Studi di Napoli. —. 2012. The Syntax of Neo-Aramaic: The Jewish Dialect of Zakho. Gorgias NeoAramaic Studies 13, Piscataway, NJ: Gorgias Press. Gutman, Ariel. 2008. “Reexamination of the Bare Preterite Base in the Jewish NeoAramaic Dialect of Zakho.” Aramaic Studies 6 (1): 59–84. Heine, Bernd, and Kuteva, Tania. 2002. World Lexicon of Grammaticalization. Cambridge: Cambridge University Press. Jastrow, Otto. 1971. “Ein neuaramäischer Dialekt aus dem Vilayet Siirt (Ostanatolien).” Zeitschrift der Deutschen Morgenländischen Gesellschaft 121: 215–222. —. 1988. Der neuaramäische Dialekt von Hertevin (Provinz Siirt), vol. 3. Wiesbaden: Harrassowitz. 24

The village of Oz, or Hoz, is one of the Judi villages. The form rxiš ‘went’ is quite particular since it lacks the L-suffix -le indexing the agent. This may be a lapsus caused by the adjacent preposition il ‘to’, which resembles an L-suffix. Alternatively it may be a rare usage of a bare preterite, in which the agent is understood from the context. Such cases have been attested in the Jewish dialect of Zakho, though predominantly with transitive verbs (Gutman 2008). Notice that the use of S-suffixes as agent markers of intransitive preterite forms is unattested in the Judi dialects, unlike some Jewish dialects of Iran (see for example Khan 2009: 71ff.).

SOME PECULIARITIES OF THE GAZNAX DIALECT

321

Kane, Tomas Leiper. 1990. Amharic-English Dictionary. Wiesbaden: Harrassowitz. Khan, Geoffrey. 2008. The Neo-Aramaic Dialect of Barwar, 3 vols. Leiden: Brill. —. 2009. The Jewish Neo-Aramaic Dialect of Sanandaj. Gorgias Neo-Aramaic Studies 10, Piscataway, NJ: Gorgias Press. Poizat, Bruno. 1986. “The Sureth-Speaking Villages in Eastern Turkey.” Journal of the Assyrian Academic Society 1: 17–23. —. 2008. Manuel de Soureth: Initiation à l’Araméen d’Aujourd’hui, Parlé et Écrit, Avec la collaboration de Yawsep Alichoran et Yohanan Binouissa. Paris: Geuthner. Sinha, Jasmin. 2000. Der neuostaramäische Dialekt von Bēṣpən, (Provinz Mardin, Südosttürkei) Eine grammatische Darstellung. Wiesbaden: Harrassowitz. Thiry, August. 2007. Geznakh – Back to Hakkari. [online] http://www.shlama.be/shlama/content/view/11/12/ (accessed on 06/2012). Yaramis, Zackarie. 2010. Gaznakh: Nid des Civilisations. Millau: Transversales Editions.

HOW A NEO-ARAMAIC SPEAKER (AVIDANI OF

ʿAMIDYA) COPES WITH A BIBLICAL ARAMAIC TEXT (BOOK OF DANIEL): A SURVEY OF MISTRANSLATIONS OF

VERBAL AND NOMINAL FORMS

1

YONA SABAR American universities encourage professors to do “community service”. This is, of course, in addition to the primary tasks of research and teaching. So, from time to time I give a talk on the history of Aramaic to non-scholarly audiences, in synagogues, Jewish community centres, churches, book clubs, etc. On such occasions I am often asked by some curious members of the audience “If you were given a page of Talmud, would you understand it easily?” or “If Jesus were alive today could you talk to him and could you understand each other?” I have to disappoint them, saying that the answer is mostly negative, mutual understanding between speakers of old and new historical stages of a language are very limited, if any. Indeed in many cases it is so even in contemporary regional/religional dialects of the same language, as we know from the situation of Neo-Aramaic (NA) in our time. Therefore I thought it to be of interest to examine how a Neo-Aramaic speaker copes with a Biblical Aramaic text. I shall take as my case study the translation of the book of Daniel by Hakham Avidani of Amidya. I shall focus, in particular on the lexical and morpho-

1

Based on an expanded introduction to my book Daniel in Neo-Aramaic (forthcoming), and two lectures at HUJ, IAS, May–June, 2013. Note these abbreviations: BA = Biblical Aramaic of the Book of Daniel; NA = Neo-Aramaic.

HOW A NEO-ARAMAIC SPEAKER COPES WITH A BIBLICAL ARAMAIC TEXT

323

logical categories in which Avidani encounters difficulties and has a tendency to make translational errors. Old Jewish Aramaic was, of course, the second language of Judaism after Hebrew. Avidani is not completely ignorant of old Jewish Aramaic texts. He was certainly familiar with texts such as the Targums of Onqelos and Jonathan, the Zohar and Kabbala, and ritual texts such as Kaddish, Kol Nidre.2 The book of Daniel, however, was hardly used in any ritual. Whereas Hebrew texts of the Torah and Prophets were often recited every Shabbat and Holiday, the Aramaic texts of Daniel and Ezra were hardly used or known. 3 The familiarity of Avidani with the book of Daniel, therefore, is much more tenuous than with texts that were known from liturgical use. Moreover, Avidani was quite old in the 1960s when a recording was made of him performing an oral translation of the book of Daniel (in the 1960s), and this certainly affected his general linguistic ability, including his grasp of his native NA.

1. VERBS4

Of course there are many everyday BA verbal roots and forms that remain, I assume, transparent and familiar to a NA speaker like Hakham Avidani, e.g. š-p-r ‘to be pleasing’, š-r-y ‘to become lose’, m-ṭ-y ‘to reach’, x-z-y ‘to see’, b-ʾ-y ‘to want’, p-l-x ‘to worship, serve’, š-w-q ‘to leave’ (BA š-b-q), y-ʾ-l (BA ʿ-l-l), y-s-q (BA s-l-q) ‘to go up’, ʾ-z-l ‘to go’, ʾ-ṯ-y ‘to come’. Even roots of some obscure forms are identfiable: manṯırun (n-ṯ-r) (BA 4:11 ʾattārū) ‘to cause to shed off (leaves)’; BA tindaʿ ‘you will know’ = NA yāʾıt; BA lehěwōn = NA hāwe, etc. In spite of generally translating quite literally, Avidani deviates from a strict literal translation in certain BA verbal forms. It is not always clear if a (mis)translation is intentional or simply due to a misunderstandng. Here is my survey of the various kinds of deviations from literal translations, or mistranslations:5 1) He translates imperative forms with optative forms (probably intentionally): BA 2:4 malkā ləʿalmīn ḥĕyī ‘O king, live for ever!’ = NA ḥakōma lʾabad xāye ‘May the 2

Some of these texts, such Passover’s bi’ur ḥameṣ (‘burning the Ḥameṣ’) were almost identical with words of Avidani’s speech, e.g. OA kol ḥamīrā wa-ḥamīʿā də-ḥazīteh u-də-ľā ḥazītēh = NA kulle xmīra u-xmīʾa ʾōd xzēli u-ʾōd la xzēli; for details see Sabar (2000). 3 Except in Mišmara, a memorial reading ritual, see ʿAmedi (1967: 49). 4 For verbal forms in the Jewish Amedia dialect see Greenblatt (2011). 5 This survey is not exhaustive; for more complete details, see Introduction in Sabar (forthcoming).

324

YONA SABAR

king live for ever!’ Similarly, BA ʾemar (2:4) ‘Say!’ = NA ʾāmır ‘Let him say’; BA 2:6 ḥelmā u-pišrēh haḥăwōnī ‘tell me the dream and its interpretation!’ = NA xılma utafsir dīde mḥākıtūli ʾaxtun ‘You (yourselves) will tell me … ’. 2) Active participles expressing the historical present are often translated (mostly intentionally) by the past tense (preterite): BA 2:8 ʿānē wə-ʾāmar ‘answers and says’ = NA mōjıble u-mırre ‘answered and said’; BA 5:23 ḥamrā šātayin ‘(You are) drinking wine’ = xamra štēlōxun ‘You drank wine’; BA 4:10 mın šəmayyā nāḥiṯ ‘it descends from heaven’ = NA mın šımme kušle ‘descended’; BA 4:15 lā yāḵəlīn ‘(they are) not able’ = la ımṣēlu ‘they were not able’. 3) Even when active participles indicate the general present, they are often translated by past tense : BA 2:28; BA 2:40 kol-qŏḇel parzılā məhaddēq wə-ḥāšēl kollā ‘because iron grinds and pulverizes everything’ = NA dıqle u-ḥrıšle ‘grinded and pulverized’; BA 2:28 gālē rāzīn ‘revealer of secrets (= God)’ = NA glēle… g-gāle ‘revealed… reveals’ (showing hesitation). 4) The present continuous may also be translated by a past continuous: BA 5:5 u-malkā ḥāzē pas yəḏa dī kāṯəḇā ‘the king is seeing a hand that is writing’ = NA xzēle … ʾōd k-kaṯwāwa ‘saw … was writing’; or just Past: BA 2:27 rāzā dī malkā šāʾēl ‘the secret that the king is asking’ = NA mbōqırre = ‘asked’. 5) The past continuous may be translated by a past iterative: BA 4:26 ʿal hēḵal malkūṯā … məhallēk hăwā ‘he was walking on (the roof of) the royal palace’ = NA gēzılwa ‘he used to walk’ (instead of bizāla wēle?); BA 2:31 ḥāzē hăwayṯā ‘you were seeing’ = NA ġzēlox ‘you saw’ (instead of bıġzāya wēlox). 6) Even unacceptable types of compound forms are found: BA 4:7 ḥāzē hăwēṯ ‘I was seeing’ = NA xzēli wēli ‘I saw I was’ (instead of bıxzāya wēli); cf. BA 2:34 ḥāzē hăwayṯā ‘you were seeing’ = NA xzēlox wēlox ‘you saw you were’; BA 7:13 ʾāṯē hăwā ‘he was coming’ = NA ṯēle wēle ‘he came was’. 7) The Imperfect may also be translated by a past tense: BA 4:14 u-šəfal ʾănāšīm yəqīm ʿălah ‘He installs over it the lowliest of men’ = NA mōqımle ‘he installed’; BA 4:33 wə-lī…yəvaʿōn ‘they were seeking me’ = NA ṭlıblu ‘they sought’. 8) Nouns are mistranslated as verbs: BA 4:22 dī šallīṭ ʿillāyā ‘that the Supreme One is the ruler’ = NA ʾōd g-ḥākım ‘he is (the one) who rules’; BA 4:23 šallīṭīn ‘(are) rulers’ = NA g-ḥākmi ‘they rule’; BA 7:6 šolṭān ‘rule, kingship’ = NA g-ḥakmāwa ‘(she) was ruling’ (but after some hesitation corrected to šulṭanūṯa ‘kingship’); BA

HOW A NEO-ARAMAIC SPEAKER COPES WITH A BIBLICAL ARAMAIC TEXT

325

3:23 qāl qarnā mašrōqīṯa ‘sound of horn (and) pipe’ = NA qāl qāna bıṣfāra ‘sound of horn whistling’; BA 6:10 rəšam kəṯāvā we-ʾĕsārā ‘he sealed the letter and the ban’ = NA mhırre, kṯūle, muḥrımle ‘he sealed, wrote, and banned’ (impossible sequence!); BA 6:21 kə-miqrəḇēh ‘at his approaching’ = NA qrūle ‘he approached’. Occasionally he confuses lehĕwōn (‫ לֶ ֱהֹון‬6:2) ‘they will be’ with ləhōn (‫)להֹון‬ ְ ‘to them’. 9) Stative past tense forms are mistranslated: BA 4:8 rəvā ʾīlānā u-təqif ‘the tree grew up and became strong’ = NA ʾurwa … u-qūya ‘(is) great and strong’ (instead of rwēle u-qwēle). Note also double translations, probably a sign of hesitation: BA 6:24 ṭəʾēḇ ʿalōhī ‘he felt good’ = NA špırre, frıḥle, ʾılle; BA 7:10 dīnā yəṯīv we-sifrīn pəṯīḥū ‘the court sat and the books were opened’ = NA dīn ytīwa u-daftāre pṯīxe (both past participles, instead of the expected ytūle… pṯıxlu). 10) The passive voice is normally translated correctly with the auxiliary verb ʾ-ṯ-y ‘to come’,6 e.g. BA 2:19 rāzā gălī ‘the secret was revealed’ = NA ṯēle glāya (lit. ‘it came revealing’); BA 3:13 hēṯāyū ‘they were brought’ = NA ṯēlu maṯōye (lit. ‘they came bringing’); BA 4:13 yiṯyəhiv ‘(a heart) will be given (to him)’ = NA ʾāṯe yhāwa (lit. ‘will come giving’); BA 6: 3 lehĕwē nāziq ‘be harmed (active participle)’ = NA hāwe ʾāṯe mʾadōye = lit. ‘be come harming’; BA 7:9 rəmīw ‘(they) were cast’ = NA ṯēlu mandōye ‘came casting’. 11) The passive voice may also be correctly translated by an impersonal form: BA 6:18 hēṯāyiṯ ʾeven… sūmaṯ ‘a stone was brought … placed’ = NA muṯēlu … drēlu ‘(they) brought… placed’. 12) However, the passive is often misunderstood and/or mistranslated: BA 3:29 u-minnī sīm ṭəʿēm ‘an order was placed by me’ = mınni drēli = ‘from me I have placed …’ (instead of ṯēle drāya); BA 5:20 honḥat ‘was taken down’ = NA kušle ‘went down’ (instead of ṯēle makōše); BA 5:15 huʿallū ‘(wise men) were brought in (to me)’ = NA muyʾılli ‘I brought (them) in’; BA 5:13 huʿal ‘(Daniel) was brought in’ = NA yʾılli ‘I (Daniel) entered’; BA 4:33 hūsəp̄at ‘(it) was added’ = NA zıdla ‘added, increased’; BA 5:24 rəšīm ‘was inscribed (by the hand)’ = NA ršımla ‘(the hand) inscribed’ (r-š-m is a loan from Israeli Hebrew); BA 2:5 bāttēḵōn nəwālī yittəsāmūn ‘your homes will be turned into a dunghill’ = NA b-dārınnu qāṭōna ‘I (the king) will turn into a them dunghill’; BA 7:8 ʾeṯʿăqarā ‘(they) were uprooted’ = NA ʾqirīla ‘(it, f.) 6

The use of this auxiliary in expressions of the passive is due to the influence of Kurdish; cf. Rees (2008: 59–60 and 148–150).

326

YONA SABAR

uprooted them’; BA 6:23 zākū hištəkaḥat lī ‘a merit was found for me’ = NA l-záxut muġzēle ʾāli ‘for a merit he showed me (?)’ (= he saw me meritorious?); BA 2:35 dāqu ‘were ground’ = NA dıqlu = ‘(they) ground, pulverized’. 13) Occasionally, vice versa, active voice is mistranslated as passive voice or impersonal: BA 4:30 yēḵul ‘he eats’ = NA mōxıllu ʾāle ‘(they) fed him’ = ‘he was fed’; BA 6:9 tiršum kəṯāḇā ‘(you) will seal the letter’ = NA ʾāṯe mhāra kṯāwa ‘the letter shall be sealed’; BA 2:29 gālē rāzayyā ‘revealer of mysteries’ = NA ṯēle glāya ṣurr mın ‘mystery was revealed from (God)’; BA 7:12 includes both examples: heʿdīw šolṭānəhōn we-ʾarḵā … yəhīvat ləhōn ‘they (impersonal) removed their rule but extension was given to them’ = NA muxlifla … hūlu ‘(their rule) was removed … but they (impersonal) gave them (extension)’; BA 7:13 haqrəḇūhī ‘(they) brought him near’ = NA ṯēle maqrūwe ‘he was brought near’; BA 7:25 we-šulṭānēh yəhaʿdūn ‘they (impersonal) will take away his kingship’ = NA u-šulṭāne b-āṯe lı-mxalōfe ‘his kingship will be taken away’. 14) Causative verbs are confused with basic forms: BA 2:18 dī lā yəhōḇədūn daniʾel ve-ḥaḇrōhī ‘so that they will not annihilate Daniel and his friends’ = NA la ḍēʿī ‘(they) will not be lost’; BA 2:28 haškaḥat ‘I have found’ = NA muġzēli ‘I have shown’. Occasionally, vice versa: BA 5:7 ʾargəwānā yilbaš ‘(he) will wear purple’ = NA ʾarjawan malušınne ‘I’ll dress him in purple’; BA 7:20 nəp̄ alā ‘(they f.) fell’ = NA munpılla ‘she knocked (them) down’. 15) Pronominal subjects of verbal forms are confused: BA 2:36 pišrēh nēmar ‘we will say its interpretation’ = NA bamrın ‘I will say’; BA 4:15 ḥelmā ḥăzēt ‘I saw a dream’ = NA ġzēlox, ġzēli ‘you saw, (no,) I saw’; BA 2:23 bəʿēnā ‘we have requested’ = NA gıbın mınnox, ṭlıbli mınnox ‘I want … (no,) I requested’; BA 2:47 ʾĕlāhăḵōn … dī yəḵēltā ‘your God (is revealer of mysteries) … so you were able …’ = NA ımṣēle ‘he (God) was able’; BA 6:19 hanʿēl ‘(he) brought in’ = NA ṯēlu, yʾıllu ‘(they) came, entered’; BA ʾămar malkā we-hayṯīw ‘the king ordered and they brought’ = NA demēṯun ‘Bring (now)!’; BA 7:19 ṣəḇīt ‘I wished’ = NA ʾjıble ‘he wished’. However, BA 3:15 lə-ṣalmā dī ʿaḇḏēṯ ‘the statue (idol) that I have made’ = NA l-ṣanam d-ʾudle ʾo gōra ‘the idol that that man made’ is due to the desire to avoid saying ‘I’ in such a case, due to a theological taboo. 16) Semantic nuances are missed: BA 4:9 taṭlēl ‘finds shelter’ = NA gōda ṭılla ‘makes shade’ (not suitable). BA b-ʿ-y has a few nuances that are not in its NA cognate b-ʾ-y, e.g. BA 2:23 bəʿēnā = ‘we (Daniel and friends) requested,’ is mistakenly translated by the cognate gıbın ‘I want’, but after some hesitation Avidani corrected it to ṭlıbli ‘I have requested’, which fits the context much better (though to be exact it

HOW A NEO-ARAMAIC SPEAKER COPES WITH A BIBLICAL ARAMAIC TEXT

327

should have been ṭlıblan ‘we have requested’); BA 2:49 Daniyyēl bəʿā min malkā ‘Daniel requested from the king’ = NA bʾēle ‘wanted’; BA 6:13 kul ʾĕnāš dī yiḇʿē ‘every man who pleads’ = NA mṣāle, ṭālıb ‘prays, requests’; BA 6:7, 12, 16 hargišū = ‘(they) gathered (urgently)’ = NA jmıʿlu ‘assembled’ (7), rʾıšlu ‘noticed’ (12), mōlızlu ‘hurried’. 17) Mistranslations are sometimes caused by misleading similarities of sounds: BA 5:7 lə-heʿālā ‘to bring in’ = NA l-masōqe ‘to bring up’ (cf. Hebrew haʿălaʾa ‘raising’); BA 3:22 hassīqū ‘they brought up’ = NA mōqıdlu ‘they burned’ (cf. Hebrew hassaqa ‘heating’); BA 6:4 malkā ʿăšīṯ ‘the king thought’ = NA ʾudle, xšūle = ‘he made, (no,) he thought’; BA 6:8 yitrəmē = NA ʾāṯe ʾimāra, ʾāṯe mandōye ‘will be said, (no,) will be cast’.

2. NOUNS

7

1) Avidani has less difficulties with BA noun morphology than with the BA verbal system, since many forms being almost identical, e.g. NA rēši, rēše, rēša ‘my, his, her, head’ = BA rēšī, rēšēh, rēšah. There are also many nouns, prepositions, and verbal roots that remain without any, or very little, phonetic or semantic change. Here is a selected list of examples (from NA translation of the book of Daniel): nūra ‘fire’; ʾatūna ‘furnace’; tarʾa ‘gate’; ʾilāna ‘tree’; xamra ‘wine’; dehwa ‘gold’; bēṯa ‘house’; rabūṯa ‘greatness’ (BA rəḇūṯā); gūre ‘men’ (BA guḇrīn); qāna ‘horn’ (BA: qarnā); hımyanka ‘embellished belt’ (< Modern Persian/Kurdish) (BA hamnīḵā < OP); ʾasıqṯa ‘ring’ (BA 6:18 *ʿizqəṯā); ʾurwāne ‘dignitaries’ (BA raḇrəḇānayyā); BA 7:9 təlaḡ ḥiwwār ‘white snow’ = NA talga xwāra; BA 7:9 ʿămar nəqē ‘clean wool’ = NA ʾamra ṣıpya (not nıqya, since this denotes a good person). Prepositions: BA qŏḏāmay ‘before me’ = NA qāmi; BA bāṯar dənā ‘after that’ baṯır ʾadya. Some verbal roots: h-w-y ‘to be’; š-p-r ‘to be pleasing’; š-r-y ‘to become lose’; m-ṭ-y ‘to reach’; x-z-y ‘to see’; b-ʾ-y ‘to want’; p-l-x ‘to worship, serve’; š-w-q ‘to leave’ (BA š-b-q); y-ʾ-l (BA ʿ-l-l) ‘to enter’; y-s-q (BA s-l-q) ‘to go up’; ʾ-z-l ‘to go’; ʾ-ṯ-y ‘to come’; NA manṯırun (n-ṯ-r) = BA 4:11 ʾattārū ‘to cause to shed off (leaves)’; NA yāʾıt (y-ḏ-ʾ) = BA tindaʿ ‘you will know’. 2) There are some cases of ambiguity in the translations. In Avidani’s Amidya dialect, for example, mın = ‘from’ and ‘with’, which may cause ambiguity when translating Biblical Aramaic, which has two different preposi7

For details of nouns in the Jewish Amedia dialect see Greenblatt (2011), Sabar (2002).

328

YONA SABAR

tions to express these relations, e.g. BA 3:32 ʿăḇaḏ ʾelāhā ʿimmī ‘God has done with (for) me’ = NA ʾudle ʾilāha mınni. 3) Several opportunities to use cognates are missed by the translator: BA ḥarṣā (5:6) is translated by nafṭanga (< Kurdish) ‘loins’ instead of cognate xāṣa; BA ʿiddānā = majal (Ar.). Avidani probably did not even associate it with the cognate dāna (lesson, Jewish school, ‘stunde’); BA zimnā ‘time’ = waʿda (Ar.); cf. ʾiddānayyā wə-zimnayyā ‘seasons and times’ (2:21) = majāle u-waʿde. The word zimnā appears many times in BA, but it is never translated by the NA obsolete cognate zūna (< ziḇna/zimna), or zaman (Ar.). BA 6:20 bə-nāḡhā ‘at light’ = NA bıd kıxūṯıd badıl bınhe ‘at the star of morning shift’ (I am not sure if he saw the connection); BA 4:6 ḥezwā ‘vision’ = rang (< Persian ‘colour’) from the common root x-z-y ‘to see’. 4) Cognates are used even when their meaning has changed: BA 3:4 qārē bəḥayil ‘calling forcefully’ = NA. xēla ‘force’, although in NA it now has a very limited distribution, mostly in the idiom: nxılle xēle = ‘his (mental) energy has dissipated, he has become exasperated’. In other contexts ḥēlā may be translated differently: BA (4:32) ḥēl šəmayyā ‘host of heaven’ = ʿaskar [d]šımme; cf. also BA 3:20 guḇrīn gibbārē ḥayil dī bə-ḥaylēh ‘mighty men in his army’ = gūre gıbbārē d-xēla ʾōd b-ʿaskāre dīde; BA 7:6 bāṯar dənā ‘after this’ = NA bıd dūka ʾadya ‘in this place’ (but corrected after some hesitation to baṯır ʾadya ‘after this’); BA 2:41 ḥasaf-ṭīnā ‘clay material’ = NA xıspa u-ṭīna ‘(hard) clay and (!) (soft) mud (like after rain)’. 5) Mistranslations caused by false similarities of sounds: BA 4:27 ḥisnā ‘strength’ = NA ḥusın ‘beauty’ (< Ar.); BA 3:24 haddaḇrōhī ‘his counselors (?)’ = NA dabra-kare ‘(king’s) care-takers (?), i.e. ‘administrators’ (< Kurdish?); sārḵīn (6:3) ‘commanders’ = NA sardāre ‘chiefs’ (< Persian/Kurdish ‘heads’). 6) Persian or Greek loanwords are mistranslated in BA 2:5 haddamīn titʿabbəḏūn ‘you will be made into limbs (= dismembered)’ (haddamīn < Persian) = NA dımmāhe b-ōdınnu ‘I’ll make them bloods’ (= spill their blood?). BA milləṯā minnī ʾazdā ‘the matter has been made public by me’ (ʾazdā < Persian) = NA taneʾṯa mınni zılla, nıšyāli ‘the word has gone (away) from me, I have forgotten it’ (probably thinking about Nebuchadnzzar’s dream). 7) Anachronistic translations: BA 3:2 has a list of administrative functionaries siḡnayyā, paḥăwātā, etc. = NA names of Muslim functionaries: qādīye, muftiye, etc.; ʾăḥašdarpənayyā ‘satraps’ = NA ʾāġāye ‘Aghas’; BA 2:15 and 3:5 has a list of Greek loanwords for various musical instruments: qaṯros, pəsantērīn, sumpōnyā, etc. = NA

HOW A NEO-ARAMAIC SPEAKER COPES WITH A BIBLICAL ARAMAIC TEXT

329

contemporary instruments: kamanča, ẓurna; BA 2:44 malḵū dī lə-ʿālmīn lā titḥabbal ‘a kingdom that will never be harmed’ = NA adds a small interpolation: ʾādi-la madinat yisraʾel ‘It is the State of Israel’.8 8) Some noun morphology is analyzed incorrectly: BA 5:11 sāḵləṯānū ‘understanding’ (abstract noun) = NA sēxıl dīde ‘his mind’; BA 4:24 ḥăṭāʾāḵ … ʿăwāyāṯāk ‘your sins … your iniquities’ = NA xaṭa … plīmūṯa ‘sin … crookedness’; BA 3:28 malʾāḵeh ‘his angel’ = NA malʾāxe dīde ‘his angels’; BA 7:8 qarnayyā ‘horns’ = NA qāna ‘horn’, and in 7:11 qanāna (back formation); BA 7:7 taqqīfā yattīra ‘intensely aggressive’ = qwīṯa u-zodanta = ‘strong and additional’; BA 7:7 šinnayin də-p̄ arzel lah raḇrəḇān, ʾāxlā ‘it has teeth of iron, and eats ...’ = NA kāke … rāba ʾaxlāwa ‘teeth … it would eat a lot’. Occasionally he hesitates about the standard plural form: BA 6:11 kawwīn ‘windows’ = NA kāwe, kawāwe; the second form is the “correct” one; but BA 6:11 birḵōhī ‘his knees’ = bırke, instead of the expected bırkāke (dīde). BA 7:17 ʾıllēn ḥēwāṯā raḇrəḇāṯā ‘these four great beasts’ = NA ʾanna ḥaywāne, [la] ʾe ḥaywan rabṯa ‘these beats, no, this great beast’. However, some changes are intentional due to a theological taboo: BA 4:5 kə-šūm ʾelāhī ‘as the name of my god’ (Nebuchadnzzar’s words) = NA mux šımmıd ʾilāhıd do gōra ‘as the name of the god of that man’. 9) Mechanical translations are made of a word even when it has different or even opposite sense, e.g. BA ṭəʿēm/ṭaʿămā is translated rāyi ‘advice’ or by Hebrew ṭaʿam ‘taste, reason’ in almost every occurrence, including BA 6:14 D. lā sām ʿălāḵ ṭəʿēm ‘he did not obey you’ = NA la drēle ʾıllox rāyi ‘he did not give (‘put’) you advice’. 10) New or rare words: BA 7:19 dəḥīlā = NA ṣadūlta, zadoʾta ‘frightening, frightening’ (adds a second translation, feeling the first is not sufficiently known enough); BA 3:21 sarbālēhōn, paṭṭəšēhōn wə-karbəlāṯhōn ‘their mantles, their tunics and their hats’ = NA ʿasqallāṭe, kulke, bandane; pəneḵem zoʾăp̄ im (Dan. 1:10, Hebrew) = NA mṣurtıne < ṣurta ‘picture, face’; BA 2:35 ʿūr ‘chaff’ = NA barba = ? perhaps parta ‘sawdust’; BA 2:41 peḥār ‘potter’ raṣsāṣ/rassās?; BA 5:24 passā dī yəḏā ‘palm of hand’ = NA laqsıd/naqṣıd ʾīda (< laxtid ?; cf. Syriac laḥta, Ar. rāḥa); BA 7:10 ʾelep̄ ʾalp̄ īm ‘one thousand thousand’ = NA ʾalpá ʾalōpe (normally ʾalpe, ʾalpāye).

8

Cf. Rees (2008: 116, 167–68); Sabar (2006: 10 introduction); Sabar (1983: 35 introduction).

330

YONA SABAR

It is clear that the old Aramaic text, whose verbal system is considerably different from Avidani’s Neo-Aramaic dialect, presents him with more difficult and obscure forms than an average Biblical Hebrew text. We have to remember that Avidani never formally studied the grammar of Biblical Aramaic, and his knowledge of the text is based on content, context (at times ambiguous even to scholars), and traditional commentators, such as Rashi and Ibn Ezra. Nevertheless, if Daniel and Avidani were alive today, they would probably be able to converse with each other and make some sense of each other’s words, if Daniel did not use old Persian and Greek loanwords and Avidani did not use Kurdish, Turkish, and Arabic loanwords.

REFERENCES

ʿAmedi, Yitzhak. 1967. “Ritual Customs of Kurdistani Jews.” In Yalquṭ Minhagim, edited by Avraham Ben Yaʾaqov, 43–47. Jerusalem. [in Hebrew] Greenblatt, Jared R. 2011. The Jewish Neo-Aramaic Dialect of Amədya. Leiden: Brill. Rees, Margo. 2008. Lishan Didan, Targum Didan: Translation Language in a NeoAramaic Targum Translation. Piscataway, NJ: Gorgias Press. Sabar, Yona. 1983. Genesis in Jewish Neo-Aramaic. Jerusalem: Magnes Press. —. 2000 “Mah Nishtannah, Comparison of Two Haggadahs, of ʿAmidya and of Zakho.” Leshonenu 64: 73–94. [in Hebrew] —. 2002. A Jewish Neo Aramaic Dictionary, Dialects of Amidya, Dihok, Nerwa, and Zakho, Northwestern Iraq. Semitica Viva 28. Wiesbaden: Harrassowitz. —. 2006. Five Scrolls in Jewish Neo-Aramaic Translations. Jerusalem: The Hebrew University. —. forthcoming. The Book of Daniel in Neo-Aramaic Translation of the Jews of ‘Amidya, with Comparisons from the Jewish Dialects of Zakho, Syriac Peshiṭta, Christian Neo-Aramaic of Urmi, and Judeo-Arabic of Yemen.

FOLK-NARRATIVES IN THE JEWISH NEO-ARAMAIC DIALECT OF ZAKHO: THE CASE OF YOSEF VE-ʾEḤAV1

OZ ALONI In the following paper, I shall describe three areas of my current work on folknarratives in the Jewish Zakho dialect, which I am researching for my Ph.D. thesis: (1) folklore and approaches to its study; (2) narrative analysis; and (3) paremiology, i.e. the study of proverbs. I will conclude with (4) an sample from my corpus of Jewish Zakho narratives. Over the past three years, I have been recording native speakers of the dialect, most of whom were born in Zakho. I have recorded 27 informants, the majority in Jerusalem. Out of approximately 70 hours of recordings, I estimate that about 2/3 can be described as containing some genre of folk-narrative.

1. FOLKLORE AND APPROACHES TO ITS STUDY

I use the term folk-narrative in a wide, generic sense. Folk-narratives are not only folk-tales of various genres (fairy-tales, legends, moralistic parables, enriched biblical stories, etc.), but also jokes, riddles, folk-poetry, food recipes, and stories of mnemo-historic nature, i.e. the history of the community as remembered and recounted by the members of the community themselves. Thus the texts (i.e. the audio-texts) that I regard as in the scope of my current study can range from the elaborate or ‘enriched’ biblical story of Yosef ve-ʾEḥav ‘Joseph and his Brothers’, told by Samra Zaqen (discussed below), through the miraculous folktales of Ḥabuba Masusoni (the daughter of Mamo Yona Gabbay), the adventurous stories and sarcastic jokes of Yistḥaq Zizi, to the descriptions of household life and social structure by Zakiyya Ben Naḥum. 1

This paper was presented at the conference ‘Neo-Aramaic Dialectology: Jews, Christians and Mandeans’, Jerusalem, June 2013.

332

OZ ALONI

Is there anything in common between all these narratives, beside the fact that they are all told in the Jewish Zakho dialect? What justifies treating all of these varied and very different texts, which are classified under different genres of verbal performance, as items in a single category of folk-narratives? I would like to broaden or rephrase that question: Is there anything that signifies folk-narratives and distinguishes them from any other form of verbal expression? Is there anything that justifies the treatment of folk-narrative, or any other form of folk-creativity for that matter, as an independent category, worthy of its own research methodologies? The answer to that question, according to what is regarded by many as the founding programmatic essay of contemporary folklore study, the article ‘Folklore as a Special Form of Creation’ by Jakobson and Bogatyrev,2 is, of course, yes. Jakobson and Bogatyrev claim that folklore is indeed a special, unique, form of human creativity. According to them, the relation between a potential item of folklore, one which exists as knowledge common to many members of a community, and between its actual, concrete, individual performance, is parallel to the relation between the two Saussurean concepts of langue and parole. Pursuing the parallelism between the study of folklore and linguistics, the authors claim that the adaptation of an “item of folklore” by society is likewise parallel to processes of grammaticalization and other innovative transformations in language. They point out a fundamental difference between written literature, on the one hand, and oral, folkloristic literature, on the other. The former is an idiosyncratic product of one particular person, and is a permanent imprint of individual expression, whereas the latter retains its existence through transmission within a community across generations. Its preservation, but not necessarily its inception, is dependent on its acceptance by the society that hands it down. In their words: “... in folklore only those forms are retained which hold a functional value for the given community”.3

2. NARRATIVE ANALYSIS

In my work on the Jewish Zakho folk-narratives I approach the material using several theoretical frameworks. Roughly speaking, most of the various contemporary approaches to folk-narrative analysis can be classified as belonging to one of three main schools.4 The first is the Comparative Method—also known as the Geographical-Historical Method or the Finnish School. Its main concern is the classification, 2 3 4

Jakobson and Bogatyrev (1980 [1929]). Jakobson and Bogatyrev (1980 [1929]: 6). See Hasan-Rokem (1987).

FOLK NARRATIVES IN THE JEWISH NEO-ARAMAIC DIALECT OF ZAKHO

333

analysis, and cross-cultural comparison of folk-narratives, mainly of proper folktales. This is done on the basis of “types” and “motifs”. Simply put, motifs are abstractions of narrative events that recur cross-culturally, while types are larger sets of motifs that share features, as it were a semantic field of motifs. The most widely used catalogue of tale types is the Aarne-Thompson classification system.5 A second school, known as the Neo-Functionalist or Contextual School, scholars such as Malinowski and von Sydow, which criticized some aspects of the Comparative School, introduced the important term “theme” or Oikotype. This term refers to sets of motifs that undergo processes of change over periods of time in a communal or regional framework. A contextual thematological analysis of a folktale would try to identify its set of motifs, its theme, in other synchronically and diachronically attested versions. In the context of Neo-Aramaic, important sources for comparison with the Zakho material would be published folk-narratives in other Jewish and Christian dialects, larger corpora of general Middle Eastern folk-narratives and earlier Jewish literature. The goal of such an approach would be to characterize and explain the changes that occurred in the theme on both the historical axis and the cross-communal axes.6 A sample of a thematological analysis is presented below. The third approach, that of the Structuralist School, originates in the work of Vladimir Propp. Propp’s Morphology of the Folktale,7 indeed, pioneered the entire discipline of narratology, which, with further contributions of Levi-Strauss and Barthes, became a highly influential discipline in literary criticism, even more than in the study of folklore. Propp characterized folk-tales by constructing a reduced form, a skeleton, of the narrated events, which he terms functions. The aim of his analysis was to construct a narrative-grammar—not a grammar of the language of the narrative, but rather a grammar for the behaviour of the narrative itself. A different aspect of folk-narrative analysis is, of course, linguistic analysis and its interaction with narrative structure. An extensive chapter on this topic, narrative syntax, was recently published by Eran Cohen in his book on the syntax of the Jewish Zakho dialect.8 In my work, I try to identify the narrative significance of several selected linguistic features across a variety of folk-narrative genres, with comparison to other dialects. Such features include, for example: the distribution of direct speech versus narration; the markers of direct speech, i.e. the various forms of the verb ‘to say’; the distribution of various verbal forms, such as šqəlle, šqəlwale, qam-šaqəlle, within the

5

Aarne (1961). Elstein and Lipsker (2004: 33–49) refer to this as the “heterogeneous series”. 7 Propp (1968). 8 Cohen (2012). 6

334

OZ ALONI

text; the distribution of various demonstrative pronouns within the text; the distribution of personal versus pronominal character reference.

3. PROVERBS

Another field that I include in the framework of folk-narrative analysis is the study of proverbs, known as ‘paremiology’. Both narratives and conversations in Jewish Zakho, like in many other dialects, are embellished with an abundance of proverbs, some of them in rhyme. Two examples of this are the following:9 (1) šúl work[M]

ʾozí-le make.3PL-ACC.3MS

k-čáhe IND-become_tired.3SG.M

xuràs-i,ˈ friends-POSS.1SG

lə̀bb-iˈ heart-POSS.1SG

u-g-néx-I and-IND-rest-3PL

ʾizàs-i.ˈ hands.PL.F-POSS.1SG ‘Work that is done [for me] by my friends—[it is] my heart [that] gets tired, while my hands rest’.10 (2) be-kálo house.GEN-bride

šʾə̀š-lu,ˈ shake.PST-3PL

be-xə́tna house.GEN-bridegroom

la-rʾə̀š-lu.ˈ

NEG-feel.PST-3PL

‘[In] the house of the bride they are [already] rejoicing, [but in] the house of the groom they have not [yet] felt [anything]’. These interpretive remarks about states of affairs are sophisticated means of conveying complex, ironic messages, which may serve various functions in narrative and in discourse. Four collections of proverbs in the Jewish Zakho dialect have been published: Bar-Adon Rivlin (1945), Segal (1955), and Sabar (1978). Together they contain 9

In the transcription I follow the practice of the grammars of Geoffrey Khan and mark nuclear stress by a grave accent (v̀), non-nuclear word-stress by an acute accent (v́) and intonation group boundaries by a superscribed vertical line (ˈ). 10 See Bar-Adon (1930: 12) proverb no. 1; Segal (1955: 264) proverb no. 85; Sabar (1978: 229) proverb no. 127.

FOLK NARRATIVES IN THE JEWISH NEO-ARAMAIC DIALECT OF ZAKHO

335

around 400 proverbs. These important collections, however, do not record the verbal environment in which the proverbs were used and do not reveal their social and discursive contexts. The central question, from the standpoint of narrative and discourse analysis, is formulated by paremiologist Peter Seitel as follows: “By application of what set of rules does [one utter proverbs] in a culturally appropriate manner and by what criteria does [one] judge correctness of another’s usage?”11 Proverbs can be regarded as ‘out-of-context’ speech acts that are intended to further some end within the discourse. 12 Various aspects of their function can be analysed. Firstly, proverbs can be categorized as either deictic or anaphoric, i.e. some proverbs refer to extra-linguistic events or situations, while others refer to the content of the discourse. Secondly, one can observe in proverbs three sets of relations, which have a complex correlation between them: the circumstantial relation between the addressee of the proverb and its addresser; the relations between the elements of the social, or other, function that the proverb is intended to express; and the logical relations between the constituents of the proverb itself.13 The ‘internal’ features of a proverb, its syntactic structure, lexical characteristics, prosody etc., in addition to being a fruitful subject of study in their own right, are relevant to the proverb’s linguistic and extra-linguistic contexts, and play an important role in the perception of proverbiality, i.e. the interpretation of an utterance as proverbial by speakers in a discourse.14 An expression of the interface between linguistics and folklore studies is HasanRokem’s application of a set of two Saussurean concepts to the phenomenon of the proverb.15 A proverb once used creates a collocation, a link, between the situation at hand and a chain of past situations that the same proverb may apply to. HasanRokem terms this “the paradigmatic aspect of proverb usage”. When proverbs are used within a narrative, it is this aspect that creates intertextuality, a relationship with other narratives and situations in which the same proverbs appear. By referring a situation to the community’s values and transferring it to a conceptual level, the proverb restores equilibrium to the situation. The aforementioned ability of the individual to use a proverb in acceptable, “correct” contexts is termed “the syntagmatic aspect”.

11

Seitel (1969: 144). This ability of using proverbs appropriately is termed by HasanRokem (1982: 11) ‘proverb competence’. ‘Competence’ is here used in the Chomskian sense. 12 Seitel (1969: 145). 13 Seitel (1969: 157). 14 See Arora (1994). 15 Hasan-Rokem (1982: 11).

336

OZ ALONI

4. A SAMPLE FROM A JEWISH ZAKHO NARRATIVE

In what follows, a short sample of a thematological analysis of a particular Zakho narrative will be given.16 The study of thematology of Jewish literature, i.e. the study of recurrence and development of themes over long periods of history, has progressed considerably in the last decade, with the publication of a series of thorough thematological studies by scholars such as Edelstein, Lipsker, and Kushelevsky; and with the on-going publication of Sippur ʿOqev Sippur, an encyclopaedia of the Jewish story.17 The particular Zakho text I shall discuss is the story of Yosef ve-ʾeḥav, ‘Joseph and his Brothers’, told by Samra Zaqen. This story belongs to a very popular genre of folk-narratives amongst the Jews of Zakho, the genre of enriched biblical narratives. Samra’s story is to a large extent a part of the tradition of the tafsirim, the re-telling of usually heroic or dramatic biblical tales, with many additions from a variety of sources. Sometimes the tafsirim are in the form of epic poems, performed by singing.18 They are described by Yona Sabar as “the foremost literary product of the Ḥakhamim of Kurdistan”.19 This ‘literary product’, which at some point was a product of the literary creativity of a single author, was transmitted orally, and therefore became part of the folk-repertoire, and developed many different versions. Samra is an exceptionally talented storyteller. She captures the listener’s attention by her fascinating way of unfolding the story. She accompanies her stories with rapidly changing intonations and elaborate body language. She was born in Zakho, and arrived in Jerusalem together with the first group of immigrants from Zakho in 1951. She was in her teens then and already married. Samra is fluent in four languages: Neo-Aramaic, Kurdish, Arabic and Modern Hebrew, but she reads or writes none of them. All of her many, very long, and exceptionally detailed stories are purely oral. Samra was recorded telling the story of Joseph and his Brothers in September 2011. The framework of the story is the biblical narrative about Joseph and his brothers in Genesis 37–47. Numerous additions, originating from various sources, are, however, incorporated into the storyline. The biblical story and the multiple added embellishments are amalgamated into one, cohesive narrative. The final

16

For a more detailed thematological analysis of this particular narrative, and for its full transcription and translation, see Aloni (2014). 17 Elstein, Lipsker, and Kushelevsky (2004). 18 For a parallel Christian Neo-Aramaic tradition, see Mengozzi (2009). 19 Sabar (1982: xxxvi).

FOLK NARRATIVES IN THE JEWISH NEO-ARAMAIC DIALECT OF ZAKHO

337

product as told by Samra gives no sign of its composite nature. It has become a homogenous narrative. Of the many themes appearing in Samra’s story that are not derived directly from the biblical text I have selected four examples. In Samra’s story, some years after Joseph has already been sold to the Egyptians, and famine has struck the land, the children of Jacob in Canaan ask him to identify a country in which wheat can still be purchased. He does so by using his sense of smell. In Samra’s words: (3)

yaʿákov Jacob

H H

ʾaví-nuH father-POSS.1PLH

g-əmr-í-le IND-say-3PL-DAT.3SG.M

tú-le sit.PST-3SG.M

ʾàba,ˈ father

móx smell.IMP.SG.M

b-əd ʾéma bážer ʾíz in-GEN which.INTER city[F] there_is d-əd of-GEN

xə̀ṭṭe.ˈ wheat[F]

g-er IND-say-3SG.M rìxa,ˈ smell.M

mə́x-le smell.PST-3SG.M

ríxa smell.M mə́x-le smell.PST-3SG.M

mən-məṣráyim from-Egypt sá-un go.IMP-PL

k-ése IND-come.3SG.M

l-məṣràyim.ˈ to-Egypt

‘Jacob our Father was sitting, his sons asked him: “Father, smell [impr] in which city there is the smell of wheat.” He smelled [and] smelled, and said: “The smell comes from Egypt, go to Egypt.”’ I have not found any older source for this aggadah. It seems to be an original creative addition of the Jews of Kurdistan, or an instance of preservation of an older motif, otherwise not attested. Jacob’s association with the sense of smell may originate from the blessing on him by his father, Isaac, in Genesis 27:27: ‘...[He] blessed him, and said: See, the smell of my son is as the smell of a field which the Lord has blessed’.20 This aggadah does appear, in a different formulation, in a recorded performance of an epic poem that recounts the same story and shares many features with Samra’s story. This is the performance by the cantor Naʿim Shalom.21 It does

20 21

JPS 1917 translation (with slight modification). Shalom (CD).

338

OZ ALONI

not appear, however, in the parallel poem in a published collection of epic poems of the Jews of Zakho by Rivlin,22 which differs considerably in various details from Naʿim Shalom’s rendering. Later in the story, Judah confronts his brother Joseph in Egypt, without knowing it is his brother. Joseph orders his sons, Menassehand Ephraim, to restrain Judah. Judah identifies the strength of their hands as strength typical of his father’s house. In Samra’s words: (4) g-émer IND-say.3SG.M dìd-e,ˈ of-POSS.3SG.M

ta to

menáše Menasse

u-fráyim and-Ephraim

g-er

xpùq-u-le,ˈ embrace.IMP-PL-ACC.3SG.M

IND-say.3SG.M

drá-un put-PL

ʾiz-óxun hand[SG.F]-POSS.3PL

kúl-lu all-GEN.3PL

wárize veins

g-émer IND-say.3SG.M ʾíz-[ə]d hand[F]-GEN

̀ wāy!ˈ oh!

[rə]š-kàp-e,ˈ on-shoulder[SG.M]-POSS.3SG.M

díd-e POSS-3SG.M

qam-xabqí-le PAST-embrace.3PL-ACC.3SG.M ʾánya these

bnón-e sons-POSS.3SG.M

b-yàtw-i.ˈ

FUT-sit-3PL

ʾúz-lu do.PST-3PL

hàdxa,ˈ thus

ʾizása hands

be-bàb-i=lu,ˈ house.GEN-father-POSS.1SG=COP.3PL

‘Joseph said to Menasseh and Ephraim his sons: “Embrace him, put your hands upon his shoulder, [so that] all of his blood vessels will settle down.” They embraced him, in this manner, and he said [in surprise]: “Oh! These hands are hands of my father’s house”’. This aggadah can be found in an older source, viz. in Midraš Tanḥuma. There, however, as well as in Rivlin’s collection23 and in Naʿim Shalom’s performance,24 it is

22

Rivlin (1959: 122–152). Rivlin (1959: 128) : ‫‘ זה אינה אלא של בית אבא‬This [blow] is none other than [a blow] of [my] father’s house’; Naʿim Shalom: ʾe dárbet bés be-bábi ‘This is the blow of my father’s house’. 24 Shalom (CD). 23

FOLK NARRATIVES IN THE JEWISH NEO-ARAMAIC DIALECT OF ZAKHO

339

Shimʿon who identifies Menasseh after the latter had struck him, before he, Shimʿon, was imprisoned by Joseph.25 Further in the story, Seraḥ, the daughter of Asher, tells Jacob her grandfather the good tidings about Joseph: he is alive and ruling Egypt. For being the bearer of such good news, she is rewarded with eternal life. This aggadah appears in several places in a relatively late layer of Rabbinic literature, for instance in Targum Pseudo-Jonathan, 26 the Alphabet of Ben Sirah, 27 Berešit Rabbati, 28 and Midrash Sekhel Ṭov.29 This addition, with the deliberately indirect manner in which Seraḥ tells Jacob the good news, by singing a lullaby containing the information about Joseph to her baby brother, appears also in an important source of the aggadic elements of Samra’s story, as well as of the epic song in Rivlin’s book and of the recorded performance by Naʿim Shalom, viz. the book Toqpo Šel Yosef by Rabbi Yosef Shabbetai Farḥi, published in 1846 in Livorno.30 This is a collection of elaborated aggadot and folk traditions, weaved artistically into a continuous narrative. Like the more popular book by the same author, ʿOse P̄ele,31 it is one of the first works belonging to a trend prominent in Jewish, and also non-Jewish, literature of the second half of the 19th century, namely anthologies of folk traditions. This book Toqpo Šel Yosef seems to have been very popular amongst the Ḥaḵamim of Kurdistan, and many of the stories in it were incorporated into the Jewish traditions of the region. Farḥi’s book, drawing from folk traditions, on the one hand, and inspiring other folk traditions, on the other, is a case that is an example of a phenomenon discussed in the aforementioned essay by Jakobson and Bogatyrev:32 the possibility for reciprocal relations between folklore and written literature, i.e. ‘recycling’ of folklore. Despite their separate functions in human culture and differences in their development, artistic literary works and folkloristic works may influence one another and may constitute the raw material for one another. The final example is the heart-rending dialogue between Joseph and his mother Rachel, which occurs earlier in the story. This section of the story is sung by Sam25

Midraš Tanḥuma, Parašat Va-Yyiggaš, Siman 4. Targum Yerushalmi, for Book of Numbers 26: 46. 27 In Eisenstein (1915: 50). 28 On Genesis 24: 34. 29 By Rabbi Menachem son of Rabbi Shlomo. On Genesis 45: 26. Sekhel Ṭov, Buber edition (1900: 288). 30 Farḥi (1846: 58a). 31 Farḥi (1870). 32 Jakobson and Bogatyrev (1980 [1929]: 13–15). 26

340

OZ ALONI

ra.33 It does not appear, however, in Naʿim Shalom’s recorded performance.34 After he has been sold and is on his way to Egypt with the convoy of merchants, Joseph visits his mother’s tomb. He then hears the voice of his mother and the two of them have this emotional dialogue. The theme appears, with some changes, in both Sefer Ha-Yašar, first printed in Venice 1625,35 as well as in Toqpo Šel Yosef.36 (5) wal V.EMPH

zonì-leˈ buy.3PL-ACC.3SG.M

g-er IND-say.3SG.M

bàxxad ʾiláha nábl-u-li kəz ràḥel mercy.GEN God take.IMP.PL-ACC.1SG to(chez) Rachel zerə̀n-na,| visit.1SG.M-ACC.3SG.F

ziyára visit[F].VERB.NOUN(=tomb)

ʾurx-òxun híle way[F]-POSS.2PL COP.3SG.M mən from H

go in

ʾazH soH

H

wèlu,ˈ COP.3PL

sàdeH H field.MH

H

dámmed when

bet-lèḥem.ˈ Bethlehem[M]

ṃə̣̀ḷyeˈ full.PL

mal-u-mə́ndi property[M.SG]-and-thing.M.SG mzàbni.ˈ sell.3PL

mən from

qam-nablí-le PAST-take.3PL-ACC.3SG.M

zə́l-le l-bet-lèḥem,ˈ go.PST-3SG.M to-Bethlehem

tə̀jar-eˈ merchant-PL

yı̀mm-i,ˈ mother-POSS.1SG

zə̀l-leˈ go.PST-3SG.M

b-nàbliˈ

rušalàyim,ˈ Jerusalem

bet-lèḥem.ˈ Bethlehem ʾánya these

pàreˈ money.PL FUT-carry.3PL

go inside

dìd-a.ˈ of-GEN.3SG.F

kùl-luˈ all-GEN.3PL

ṃə̣̀ḷyeˈ full.PL go in

ʾŭ̀r-reˈ enter.PST-3SG.M

mə́ṣer Egypt g-èr,ˈ IND-say.3SG.M

ṣrə́x-le xa qàla ʾəl yı̀mm-e,ˈ bə-d kùrdi,ˈ cry_out.PST-3SG.M a/one voice.M to mother-POSS.3SG.M in-GEN Kurdish

33

Recording reference: SZ110922T6 01:35–02:56. Shalom (CD). 35 Page 83a in that edition. 36 Farḥi (1846: 12a). 34

FOLK NARRATIVES IN THE JEWISH NEO-ARAMAIC DIALECT OF ZAKHO g-er [singing:] IND-say.3SG.M

yá oh.VOC

xzè-li,ˈ see/look.IMP.2F.SG-ACC.1SG

yımm-ìˈ mother-POSS.1SG

ha

zə́nj-ʾə́zə́ˈ [=zənjír-e], chain-PL

kápəkyàsa handcuff.PL

DEIC

wél-u PRES.COP-3PL

b-ʾaklàs-i,ˈ in-leg.PL-POSS.1SG

marawàs-i, owner.PL- POSS.1SG

wən COP.1SG.M

b-ʾákl-i in-leg-POSS.1SG

marawàsi,ˈ owner.PL-POSS.1SG

xa qála a/one voice[M]

wən COP.1SG.M

mzŭ̀ḇna sell.STAT.PTCP

sè-le, [declaiming:] náše come.PST-3SG.M man.PL yı́mm-e mother-POSS.3SG.M

qú-lax rise.IMP-2SG.F

b-ʾizàs-i,ˈ xòle in-hand.PL-POSS.1SG rope.PL

də́rye put.STAT.PTCP.PL

z(i)yár tomb[F].GEN

qú-lax rise.IMP-2SG.F

yá yımm-ìˈ oh.VOC mother-POSS.1SG

xzè-liˈ see.IMP.2SG.F-ACC.1SG

341

mzŭ̀bna sell.STAT.PTCP

ta to

ta to

mən yı́mm-e from mother-POSS.3SG.M kúl-lu all-GEN.3PL

šʾə̀š-la,ˈ shake.PST-3SG.F

mzurzèʿ-lu,ˈ shiver.PST-3PL kúl-lu all-GEN.3PL

nášeˈ rʾə̀l-lu.ˈ [singing:] xá qalàˈ m-yı́mm-e man.PL tremble.PST-3PL a/one voice.M from-mother-POSS.3SG.M sè-le,ˈ come.PST-3SG.M lá

NEG

zadʾèten,37ˈ fear.2SG.M

zadʾèten, fear.2SG.M

37

ṣrə́x-le cry.PST-3SG.M yá oh.VOC

yà oh.VOC

bron-í son-POSS.1SG

bron-í son-POSS.1SG yosèf,ˈ Joseph

[declaiming:] ṭḷahàgaˈ yá three_times oh.VOC

Long form of pronominal subject inflection.

yosèf,ˈ Joseph

lá

NEG

bron-í son-POSS.1SG

342

OZ ALONI yosèf,ˈ Joseph

lá

NEG

tàkleten,38ˈ trust.2SG.M

zadʾèten,ˈ [singing:] fear.2SG.MS dúkset place.F-GEN

b-ə́d in-GEN

ʾázətˈ go.2SG.M.

ḥàkom king.GEN

ṃə́ṣer Egypt

pešèten.40 FUT.become.2SG.M

ḥákom king.GEN

ṃə́ṣer Egypt

pešèten.ˈ FUT.become.2SG.M

ʾilàh-ox God-POSS.2SG.M b-najḥèten,39 FUT-win.3SG.M

‘Indeed [or: when] they bought him. He said: “For God’s sake, take me to Rachel my mother, so that I may visit her, her tomb. It is [on] your way [to] Bethlehem.” From Jerusalem, from the field, they took him to Bethlehem. So, when he went to Bethlehem—they were all merchants, full [of], full [of] money. They carry property and goods, selling [it] in Egypt. He went [and] entered [to the tomb], he says, he cried out [with] a [loud] voice to his mother, in Kurdish [=Aramaic], he says [=as follows]: [singing:] “Oh Mother! Rise and look at me [=see the poor condition I am in], Oh Mother! Rise and look at me! Here, on my leg there are chains, Handcuffs on my hands, Ropes are placed on my legs, I have been sold to my owners, I have been sold to my owners.” A voice came from his mother, all of the men shivered. The tomb of his mother shook. All of the people trembled. A voice came from his mother, 38

Long form of pronominal subject inflection. Long form of pronominal subject inflection. 40 Long form of pronominal subject inflection. 39

FOLK NARRATIVES IN THE JEWISH NEO-ARAMAIC DIALECT OF ZAKHO

343

He [the voice] cried out: “Oh my son Joseph, do not fear, Oh my son Joseph, do not fear. Three times: Oh my son Joseph, do not fear. Trust in your God, Wherever you go, you shall win [or: In the place that you go to, you will win]. You shall become the king of Egypt. You shall become the king of Egypt.”’

REFERENCES

Aarne, Amtii 1961. The Types of the Folktale: A Classification and Bibliography, translated and Enlarged by Stith Thompson (2nd ed.). Helsinki: Suomalainen Tiedeakatemia. Aloni, Oz. 2014. The Neo-Aramaic Speaking Jewish Community of Zakho, A Survey of the Oral Culture. MPhil Dissertation under the supervision of Geoffrey Khan, Faculty of Asian and Middle Eastern Studies, University of Cambridge. Cambridge: Lambert Academic Publishing. Arora, Shirley. 1994. “The Perception of Proverbiality.” In Wise Words—Essays on the Proverb, edited by Wolfgang Mieder, 3–29. New-York: Garland. (Reprinted from Proverbium 1, (1984)) Bar-Adon, Pesaḥ. 1930. “Me-Haʾaramit Ha-Mdubberet ʾEṣel Ha-Yhudim HaKurdiyyim.” Zion—Reports of the Land-of-Israel Society for History and Ethnography 1:12–13. [in Hebrew] Cohen, Eran. 2012. The Syntax of Neo-Aramaic: The Jewish Dialect of Zakho. Piscataway, NJ: Gorgias Press. Elstein, Yoav, Lipsker, Avidov, and Kushelevsky, Rella, eds. 2004. Sippur ʿOqev Sippur, Encyclopedia of the Jewish Story. Ramat Gan: Bar-Ilan University. [In Hebrew] Elstein, Yoav, and Lipsker, Avidov. 2004. “Qavey Yesod Ba-Tematologia Šel Sifrut ʿAm Yisrael.” In Sippur ʿOqev Sippur, Encyclopedia of the Jewish Story edited by Elstein, Yoav, Lipsker, Avidov and Kushelevsky Rella, 25–51. Ramat Gan: BarIlan University. [in Hebrew] Eisenstein, Yehudah David. 1915. ʾOṣar Midrašim. New York: Eisenstein. [In Hebrew] Farḥi, Yosef Shabetay. 1846. Toqpo Šel Yosef. Livorno: Rabbi Yisrael Qushta and Associates, Moshe Yeshuʿa Toviana Press. [In Hebrew] —. 1870. Sefer ʿOse P̄ele. Livorno: Rabbi Yisrael Qushta and Associates. [In Hebrew]

344

OZ ALONI

Hasan-Rokem, Galit. 1987. “The Snake at the Wedding—A Semiotic Reconsideration of Comparative Method of Folk Narrative Research.” ARV Scandinavian Yearbook of Folklore 43: 73–87. —. 1982. Proverbs in Israeli Folk Narratives: A Structural Semantic Analysis. Helsinki: Suomalainen Tiedeakatemia, Academia Scientiarum Fennica. Jakobson, Roman, and Bogatyrev, Petr. 1980 (1929). “Folklore As a Special Form of Creation.” Translated by: John M. O’Hara, Introduction by: Felix J. Oinas. Folklore Forum 13: 1–21. —. 1982 (1929). “Folklore As a Special Form of Creativity, Translated by: Manfred Jacobson.” In The Prague School—Selected Writings, 1929–1946, edited by Peter Steiner, 33–46. Austin: University of Texas Press. JPS. 1917. The Holy Scriptures According to the Masoretic Text: a New Translation with the Aid of Previous Versions and with Constant Consultation of Jewish Authorities. Philadelphia: Jewish Publication Society of America. Mengozzi, Alessandro. 2009. “Religious Poetry from Alqosh and Telkepe (North Iraq): Contacts between Sureth-Speaking Communities and Europe in the 19 th Century.” ARAM 21: 49–59. Propp, Vladimir. 1968. Morphology of the Folktale, Translated by, Laurence Scott, with an Introduction by Svatava Pirkova-Jakobson. Austin: University of Texas Press. Rivlin, Yosef Yoʾel. 1945, 1946. Pitgamim Bi-Lšon Ha-Ttargum. Reshumot (new series), two parts. Tel Aviv: Dvir. [in Hebrew] —. 1959. Širat Yehudey Ha-Ttargum. Jerusalem: Mosad Bialik. [in Hebrew] Sabar, Yona. 1982. The Folk Literature of the Kurdistani Jews: An Anthology. New Haven and London: Yale University Press. —. 1978. “Multilingual Proverbs in the Neo-Aramaic Speech of the Jews of Zakho, Iraqi Kurdistan.” International Journal of Middle East Studies 9 (2): 215–235 Segal, Judah Benzion. 1955. “Neo-Aramaic Proverbs of the Jews of Zakho.” Journal of Near Eastern Studies 14 (4): 251–270. Seitel, Peter. 1969. “Proverbs: A Social Use of Metaphor.” Genre 2 (2): 143–161. Sekhel Ṭov, Buber, ed. 1900. Midraš Sekhel Ṭov by Rabbi Menachem Son of Rabbi Shlomo, Italy 1139, Edited and Published with an Introduction and Comments by Shlomo Buber. Berlin: Zvi Hirsch son of Rabbi Yitzḥak Itzkovsky Print. Shalom, Naʿim. 1986. Naʿim Šalom Bi-Qriʾa Be-Targum Qurdi: Yosf Ve-ʾeḥav, ʿAqedat Yitsḥaq. [CD] Jerusalem: Naʿim Shalom. Catalogued in the National Library of Israel as: Širim ʾEpiyim Be-ʾAramit Ḥadaša, CD 07534, Naʿim Shalom, Singer; ʾAvraham Bero, Oud; Mərad Salman, violin.

SONGS OF THE ASSYRIAN MOUNTAINEERS NINEB LAMASSU Prior to the First World War and the tragic events that radically altered the demographical map of the region, a closely-knit community of Assyrians lived in the mountainous area of what is generally known as Hakkārī (for a treatment of these events see Travis 2006). The Hakkārī-Assyrians were divided into two classes— ‘freemen’ (ʿāšīratte) and ‘subject’ communities (rāʾyattē). They lived for the most part in the present-day province of Hakkārī in Turkey, their villages followed the northward course of the Upper Zāb River in Çölemerik (today Hakkārī) in the North and the districts of Barwār, Āmadīya in the South (Sanders 1997). The ʿāšīrattē consisted of five great clans: the largest among them were the Ṭyārē, accounting for nearly one half of the Hakkārī Assyrians. This being the largest tribe in Hakkārī, it was divided into two parts: (i) Upper Ṭyārē, whose civil centre was based in Čamba dMālek on the Upper Zāb (37.39872778 N 43.49264187 E), its spiritual centre being the celebrated church of Mār Sāwā in Rūmta (37.37059786 N 43.51611128 E), and (ii) Lower Ṭyārē whose centre was based in Lēzan (37.30837673 N 43.50560968 E). To the north-east of Ṭyārē was the smaller tribe of Dezīn (37.60021892 N 43.88774768 E), which was responsible for the protection of the Nestorian patriarch. The great tribe of Jīlū (37.295911 N 44.000225 E) had its spiritual centre in Mâ (village) d-Mār Zayyā, which contained the monastery of Mār (Saint) Zayyā, and its civil centre in Zerīnē. Adjacent to the Jīlu was the Bāz tribe, of which the civil centre was Māhē Xtayyā (37.27 N 43.53 E), with ʾArgab’s Mār Qeryāqūs monastery acting as its spiritual centre. The other primary tribe was Txūmā, the principal civil centre of which was Txūmā Gawāyā with Rabban Piṯyo’s monastery acting as its spiritual centre. The inhabitants of the districts of Tāl (37.43584603 N 43.75738503 E), Wālṭū (37.42683205 N 43.61300055 E), Ashīṯā (37.33321658 N 43.3928543 E) and Eshtāzīn were under the protection of Txūmā, Upper Ṭyārē, Lower Ṭyārē and the Jīlū tribes, respectively (Fiey 1964). It is worth noting that the HakkārīAssyrians were not totally isolated from other regions. The mountain ranges of Ṭur ʿAbdīn constituted their western border. They were connected to the North by the

346

NINEB LAMASSU

southern districts of Lake Van. Iranian Azerbaijan adjoins it to the East, and Ṭūrā dMātīna to the South (Aboona 2008: 2). There are various misconceptions regarding the ethnic identity of the inhabitants of Hakkārī. Asahel Grant, for example, advanced the notion of them being the remnants of the ten lost tribes of Israel (Grant 1844), while John Joseph, who is frequently cited as an authority on the identity of the modern Assyrians of the Middle East (Joseph 2000), postulated that the term “Assyrian” was only adopted after their encounter with the Anglican Church’s mission to the Middle East. More recent studies, however, have revealed evidence challenging these views, such as the works of a sixteenth century Kurdish historiographer Sharīf Xān ibn Shams Al-Dīn AlBadlīsī (Al-Badlisi 1962, vol. I: 91). Al-Badlīsī refers to the Hakkārī Christian mountaineers as Asūrī (Assyrians), long before the appearance of the English missionaries in the region.1 The Anglican missionaries, who lived among the tribes and studied them, advanced the theory that these Assyrian colonies were established during the 13th century following the Mongol, and particularly the Timurid, invasions of lower Mesopotamia. This was notably the opinion of the Rev. Percy Badger, author of the The Nestorians and their Rituals (Badger 2004). The documents from familial archives of some of the Assyrian clans of Hakkārī do indeed support this theory. These include the family of Mālek Čikkū Gīyū, who left a record of his family’s migration dating from their departure from Arbil (Modern Erbil) in 1310. Similarly, Qashā Daniel of Bāz traced his family records to Arbil. Moreover, the Bē Qellayta of Upper Ṭyārē and Mār Bīshūʾ took their name from the citadel of that city. There is evidence that some members of the Bāz tribe originally came from Tekrit and Samarra. Some families from Jīlū’s Mā d-Mār Zayyā originated from ʾAnkāwā (now almost part of modern Erbil), their ancestor Haj bar Haj having left ʾAnkāwā only twelve generations previously, i.e. some 300 years ago (Fiey 1964). In opposition to Badger’s theory, Fiey traces the Assyrian presence in the region to at least the first Christian century. He uses ecclesiastical and synodical records to reject the notion that the entire Assyrian populace of Hakkārī migrated northwards in the 13th century (ibid; see also Aboona 2008: 13). In terms of architecture evidence, Sir Charles Wilson, in his survey of the Ottoman Asiatic regions, stated that there existed, in Hakkārī, historical monuments going back to the period of the Assyrian empire (Aboona 2008: 21). Emhardt and Lamsa also share this view and adduce:

1

For a detailed study on this subject see Lamassu (2013).

SONGS OF THE ASSYRIAN MOUNTAINEERS

347

“… they built strong ‘kalas’ or fortresses in which they maintained ample army posts. Some of these fortifications and military barracks still remain. Other traces of Assyrian civilization in these highlands are found in the form of masonry, bricks, bridges, water pipes, spear heads, and many other relics of native art. Even some of the roads constructed by the Assyrians are still in use in these regions.” (Emhardt and Lamsa 1926: 19)

In their oral traditions, the Assyrians have always maintained that their churches in Hakkārī were built upon sites predating the advent of Christianity (Ashitha 2002: 29). Indeed, Grant is believed to have found cuneiform inscriptions in one of the churches within the environs of Qudshānūs (Grant 1984: 177). Whilst Maclean and Browne both observed that the timing of the feasts pertaining to a number of the Hakkārī saints, most notably Mār Sāwa, dated the traditions to the preChristian era (Maclean and Browne 1892: 298–299; 301). Until recently, these Assyrians had a lively oral literature, especially the Hakkārī-Assyrians who, as Pennacchietti remarked, “... have known how to keep their own folklore intact from Arabic, Kurdish, Turkish or Persian influences” (Pennacchietti 1985–1986: 40). Their fierce independence over several centuries was an important factor that helped them maintain a distinct culture (Aboona 1999: 250; Aboona 2008: 9). The Ottoman Christian genocide, however, destroyed many of their habitations and forced tens of thousands to migrate to Mesopotamia, Persia, Armenia, Georgia, Russia, Lebanon, Syria, Europe and the United States (Travis 2006). In such émigré communities this rich oral culture stood little chance of preservation. Knowledge of this important component of Modern Assyrian culture is, therefore, gradually declining among the current generations. This means, as scholars, “... we are obligated to devote all our resources to collecting information on the existing spoken dialects before they disappear” (Hoberman 1990: 79). I would add that preserving stories, songs and myths is as important as studying the grammatical or syntactical structure of the language. The risk of these texts falling into oblivion, due to displacement of the community and the passage of time, obliges us to devote ourselves not only to document as much of it as possible but also to subject this corpus to serious literary criticism and analysis. Another difficulty we face is that this rich tradition has not been documented in its original and genuine surroundings but in distant cities such as Baghdad and Mosul (Pennacchietti 1985-1986: 40), and more recently in the various countries of the West. This logistical fact runs the danger of disconnecting our study of the Assyrian oral tradition from the very environment that influenced its development and so losing the nuances of many terms and expressions. We are able to differentiate between some of the genres of this oral tradition through its internal terminology. Thus, for example, we know that sheddūlē are lull-

348

NINEB LAMASSU

abies sung by mothers for their babies; dīwanī, although difficult to translate, resembles something between songs of disputation and court poetry; lēlyānā are sung by female relatives of the groom as part of the wedding festivities (Youkhanna 1998), especially during what is generally known as the bathing of the groom (for a detailed study of this genre see Bet Benyamin 2009). Since this paper will mostly focus on zmīryāṯā d-rāwē, we will address this genre in more detail. The first to document this genre was Socin, but Pennacchietti was the first to attempt to etymologize the term rāwē, which, according to him, is derived from the Arabic rawī “the letter which remains the same throughout the entire poem and binds the verses together...” (Pennacchietti 1985–1986). It is proposed here, in agreement with Bet Benyamin (1998), Ashitha and Qasrayta (1998), and Youkhanna (1998), that rāwē is derived from the verbal root r-w-ʾ, which generally means: ‘to drink water until one is satiated’, or ‘to drink wine until one is inebriated’, and can also have the following nuances: ‘to passionately fall in love with someone’, or ‘to be in a state of rapture or ecstasy’ (cf. Arabic ṭariba). Since love and passion account for the majority of rāwē’s verses, it is indeed likely that rāwē is derived from this native verbal root. This etymology is further supported by the fact that the HakkārīAssyrians would spent the four months of winter indoors drinking wine and singing rāwē.2 As for the antiquity of rāwē, Pennacchietti sees it as “extremely old,” since it has already lost meaning among the Nestorians of Persia, and the Jacobites of Ṭūr ʿAbidīn (Pennacchietti 1985–1986). Through references made to certain known historical events in rāwē, one can safely trace some of its verses back more than two hundred years (Bet Benyamin 1998, Ashitha 1998), if not further. The origins of this oral heritage may be proved to extend even further back in antiquity, but this subject will not be explored here since it is outside the scope of this paper.3 The melody of rāwē is quite distinct, and it seems to be peculiar to the HakkārīAssyrians. Again, Pennacchietti gives us a remarkable observation, and a vivid description of it: “the melodic beat was repeated three times and embraced each line as well as the first accented word of the following line, which means that each word was pronounced twice. It had a modal tune, achieved by means of using chromatics at small intervals, executed with a surprising speed. You had the impres2

This information has been obtained from various informants whom I have recorded in northern Iraq, and Khabour, Syria. 3 Consult Lamassu (2009) for a discussion of the antiquity of the genre of rāwe.

SONGS OF THE ASSYRIAN MOUNTAINEERS

349

sion that the width between the highest note and the lowest never exceeded the interval our major sixth would make, even if the most important part of this tune appeared to be limited within the span of a major fourth. Once the stornello was finished, those present showed their appreciation as to the choice of the theme and its execution by singing out a series of stressed “o”s that finished up on a series of high and extremely sharp “i”s.” (Pennacchietti 1985-1986)

Among the Assyrians, it is generally believed that rāwē was only sung by the people of Ṭyārē, Txūma and Barwār (Bet Benyamin 1998). This is probably due to the fact that its survival, in modern times, is restricted to these three communities. As remarkeī earlier, it may have existed among the communities of Iran and Ṭūr ʿAbdīn, which have since lost it. Furthermore, we know that it was probably popular among the Jīlū Assyrians as well, since we know that Socin was first introduced to this genre, in Damascus, by an Assyrian from that tribe (Pennacchietti 1985–1986). I am familiar with a rāwē triplet composed in the Bāzīyānā dialect, which may serve as an indication of its popularity in the Bāz region: (1)

ܵ ‫ܲܚܕ ܫܸܡܫܝ ܲܣ‬ ܵ ‫ܗܪܐ‬ ‫ܨܦܝܝ‬ ܲ ‫ܐܝܕܘ ݂ܟ ܵܠܐ ܕܪܝ‬ ‫ܓܘ ܵܦܝܝ‬ 4 ܵ ̈ ܲ ܲ ‫ܵܩܛܠܝ ܠܘܟ ܐܚܘܢܘܝܝ‬ xá one

šǝ́mš-i sun-GEN

sára moon

ṣpáyi good

ʾíd-ux lá hand-2SG.M NEG

drí put_forward.IMP.SG.M

qáṭli-lux kill.3PL-OBJ.2SG.M

ʾaxónway-i brothers-1SG

gú at

páy-i face-1SG

It is a perfect full moon You should not touch my face My brothers shall (see us and) kill you 4

All quotations of original texts come from the corpus documented by me, unless indicated otherwise.

350

NINEB LAMASSU

Writing about the dūrǝkyāṯā, Mengozzi has the following to say about rāwē, which he loosely calls “erotic triplets”: “the erotic triplets that Socin asked to hear from the two boys in the Dominican Monastery of St. Jacob (Mar Yaʿqu) are not less ‘Chaldean’ than the religious poem dictated to him by a blind bard...” (Mengozzi 2002: 6). Whether or not this can be taken to be true, we are unable to verify it due to the paucity of our sources. This paucity could be attributed to the fact that most scholars had clergy as their informants, and people of the cloth had no interest in lay literature. For example, we learn that in a reply to a friend’s request from Belgium querying the existence of such folk literature in Shaqlāwā, Fr. Elias Sher wrote in 1937: “in your letter dated 14th of September, in which you asked me to document for you our oral literature, which we recite at our homes in cold winter nights, we unlike the Kurds don’t recite such literature” (Sher 2007: 97). I myself, however, have documented a rich oral literature among the Assyrians of Koy and Shaqlāwā. This attitude has changed since Fr. Sher wrote his letter in 1937, Fr. Ḥabib AlNawfali5 writing about his village Bēqūpā documents a few examples of the folkloric songs of his village. These songs seem to reflect the characteristics of our rāwē genre (Al Nawfali 2002: 62), at least in terms of lyrics: (2)

‫ܘ ܹܵܠܐ ܒܝ ܵܣ ܵܩܐ ܠܬ ܸܵܠܐ‬ ܲ ‫ܲܕ‬ ܵ ‫ܠܬ ݁ܗ ܓ‬ ܵ ‫ܪܦ‬ ‫ܸܩܠܐ ܘܙ ܸ ܵܠܐ‬ ܲ ‫ܩܪܘ ܵܢܨܪ‬ ܲ ‫ܕ‬ ‫ܡܫܠܠ ܸܵܠܐ‬ ܹ ܹ wéla COP.3SG.F

bi-sáqa PROG-go_up.INF

l-tǝ́lla to-tell

darpált-a hem-3SG.F

jǝ́q-la tear.PST-3SG.F

ʾu-zǝ́l-la and-go.PST-3SG.F

dé-qro PTC-call.IMP.PL

náṣǝr Nasir

mašallǝl-la weave.3SG.M-OBL.3SG.F

Now the Chaldean bishop of Basra, Iraq. In a private conversation with Al-Nāwfalī, I was informed that this genre of “erotic triplet” is not peculiar to the few triplets which he has documented in his book but it is more of widespread phenomena among the villagers of the Nineveh Plains. 5

SONGS OF THE ASSYRIAN MOUNTAINEERS

351

She is on her way walking up the tell The hem of her skirt tore and perished Call Nasir to come and sew it for her (3)

ܵ ‫ܸܦܠ ܵܛܐ ܡܢ ܥ‬ ܵ ‫ܘ ܹܵܠܐ ܒ‬ ‫ܹܕܬܐ‬ ܲ ‫ܲܡܚܟܹܗ ܡܸ ܲܢܗܝ‬ ‫ܒܨܢܝ ܵܬܐ‬ ܵ ‫ܠܗ ܕܦ‬ ܵ ‫ܕܢܫ‬ ݁ ‫ܘܫܒ ܲܘܩ‬ ܵ ‫ܹܝܬܐ‬ ݂ ‫ܸܩܠܐ‬ wéla COP.3SG.F

bǝ-pláṭa PROG-go_out.INF

máxke talk.3SG.M

mǝ́n-na with-3SG.F

d-našǝ́q-la‫‏‬ REL-kiss.3SG.M-OBJ-3SG.F

mǝ́n from

ʾéta church

b-ṣaníta in-skill ʾu-šúq-la and-let-OBJ.3SG.F

d-féta REL-pass.3SG.F

As she is coming out of the church He chats her up with skill Let him kiss her, then let her pass The mountainous habitat of the Hakkārī-Assyrians did not provide sufficient farming lands, and so they had to carry soil from distant places to fill the man-made terrace fields, called sheshalāṯā, on the mountain slopes (Aboona 2008: 8). Despite this, we are told by Shamasha Gewargis d-Ashitha that in times of droughts and famine, many Assyrians of the plains, from as far as Alqosh, Tel Kepe and even Mosul would go up into the mountains. Although farming was difficult, there was enough rainfall in Hakkārī, even in dry spells, to support agriculture and meet people’s needs (Shamasha Gewargis D’Ashitha 1982: 58). Other than farmers, we know some of these tribes contained also professional artisans, who earned their livelihood through their trade. For example, the Bāznāyē were both masons and blacksmiths, whose skills were renowned throughout the entire region (DeBaz I, DeBaz II, DeBaz II). Many rāwē verses touch on this feature of Assyrian society, for example: (4) ܵ ‫ܲܨ‬ ܵ ‫݂ܕܪ ݁ܗ ܲܕ‬ ܵ ܵ ‫ܪܓܐ‬ ‫ܕܙ ̈̈ܪܢܝ ܹܐ‬

ܵ ‫ܩ݂ܕܝ ܵܠܐ ܡܬܘܪ ܵܨܐ‬ ‫ܠܒ ܵܙܢܝ ̈ ܹܐ‬ ܵ ‫ܠܹܐ ܵܦܬ ܹܚ‬ ‫ܛܠܐ ܢܘ ݂ܟ̈ܪܝ ܹܐ‬

352

NINEB LAMASSU ṣáḏr-a‫‏‬ bosom-3SG.F qḏíla key lé

dárga gate

mtúrṣa craft.RES.SG.M paṯǝx open.3SG.M

NEG

d-zarnáye REL-Zarnaye

ṭlá to

‫‏‬

l-baznáye by-Baznaye nuxráye strangers

Her bosom is [like] the gate of Zārnāye The key has been crafted by Bāznāye It shall not open to strangers The huge and intricately decorated gates of the Zārnāye are used here figuratively, and their physical attributes—immense size and beauty—are likened to the bosoms of the beloved. We are told that strangers are denied access to her beautiful bosom (i.e. no one other than her lover). The use of such metaphorical language is a characteristic feature of such oral literature. Other subjects alluded to in rāwē are those of long distance trade, something for which the Jīlwē Assyrians were famous for (Fox 1997: 122). When visiting the church of Mār Zayyā in Jīlū, W.A. Wigram (Wigram and Wigram 1922: 67) found a curious collection of items, including pendulums and clocks, sacerdotal vestments of Russian manufacture spread out across the walls, ostrich eggs, coral from Malabar and porcelain bottles. A rāwē refers to such contact with the outside world in the following two triplets: (5)

ܲ ‫ܸܚܩܐ ܝܠܹܗ‬ ܵ ‫ܪ‬ ‫ܒܐ ݂ܬ ܵ̈ܪ ܵܘ ݂ ܵܬܐ‬ ܵ ‫ܵܦܦܹܐ ܛܘ̈ܪ ܹܐ‬ ‫ܘܝ ̈ܵܡ ݂ ܵܬܐ‬ ܵ ‫ܘ ܵܨ݂ܕܠܝ‬ ‫ܕܢܦܹܠ ܠܐ ̈ܝ ݂ ܵܬ ݂ ܵܬܐ‬

rǝ́xq=ele far=COP.3SG.M p-ápe across

b-ʾaṯrawáṯa in-lands

ṭúre mountains

ʾu-ṣáḏ-li and-fear-OBL.1SG

ʾu-yamáṯa and-oceans

d-nápǝl REL-fall.3SG.M

l-ʾiṯáṯa to-hands

SONGS OF THE ASSYRIAN MOUNTAINEERS

353

He is faraway in [foreign] lands Across [so many] mountains and oceans And I fear he will fall into trouble (lit. into hands) (6)

ܵ ‫ܚܸܡ̈ܪ ܹܐ‬ ‫ܕܩ݂ܕ ܵܠܐ ܬܪ ܹܝ ܟܝ ̈ ܹܣܐ‬ ܵ ‫ܲܡ ̈ܡܸ ܵܟ ݁ܗ ܣ‬ ‫ܹܐܡܐ ̈̈ܪܝܣܹܐ‬ ܲ ‫ܪܙ‬ ܲ ‫ܒܫܘ ܵܩܐ‬ ܲ ‫ܕܘ‬ ‫ܬ̈ܪܝܨܹܐ‬ ̈ ‫ܪܘܢ‬ xǝ́mre beads mammǝ́kk-a breast-3SG.F b-šúqa in-market

d-qḏál-a REL-neck-3SG.F séma silver d-warzarún REL-Erzurum

tré two

kíse pouches

ríse sprinkle.RES.PL tríṣe craft.RES.PL

The necklace [lit. beads] around her neck weighs two pouches Her breasts are covered with silver Crafted in the markets of Erzurum Animal husbandry was the primary source of livelihood of the HakkārīAssyrians, especially those of Ashīṯā; with some families owning up to one thousand sheep. This way of life created a close and intimate relationship between the people and their flocks, so much so that Maclean remarked: “it is strange, too, to us to find sheep called by their names…” (Maclean 1888). Other oral poetry of the Assyrians also deals with this special bond, such as the popular song of Xūshābī, in which the hero, Xūshābī, kills Ūsman, a Kurdish Aġa (an act which would often result in a long blood feud). Knowing how Xūshābī loved his goat, affectionately called Nūne, Ūsman ordered him to have it slaughtered and cooked for him. Xūshābī pleaded with the Aġa, offering as many of his sheep as his guest wished, to no avail. As a consequence, Xūshābī kills Ūsman. The following is a sample stanza from this popular song (Giwu):6 6

Also based on some recordings of a bard called Taoma D’Wila, which I recorded in Do-

354

NINEB LAMASSU (7)

݁ ‫ܲܘܝ ܠܝ ܠܥ ܸܙܝ‬ ‫ܕܗܝ ܢܘܢ ܹܐ‬ ܲ ‫ܠܟ‬ ܵ ‫ܕ ܸ ܵܡ ݁ܗ ܪܝ ܵܣܐ‬ ‫ܢܘ ̈ܢ ܹܐ‬

ܲ ‫ܐܘ‬ ܲ ‫ܣܡܢ ܡܘܟܪ ܹܐ ܠܹܗ ܲܙ ݂ܒܢ ܹܗ‬ wáy-li woe-OBL.1SG

l-ʾǝ́zz-i for-goat-1SG

dǝ́mm-a blood-3SG.F

rísa splash.RES.SG.M

ʾúsman Usman

dé

REL.that

núne Nune

l-kanúne to-hearths

múkr=ele shorten.RES.SG.M=COP.3SG.M

záwn-e time-3SG.M

Woe for that Nūne goat of mine Her blood now stains the hearth Thus, ʾUsman has shortened his life The importance of the sheep in these people’s lives is reflected by a letter written by a father to his son that has been published by Ashitha (2008). In this letter, the father asks after the health of his son first and informs him of the news of the family. Then he immediately gives an account of the welfare of his sheep, which are referred to by name: “if you are concerned about the sheep, three of them are barren, and two ewe have died including Ezzī the naked one”. As animal husbandry was of vital importance to these people, it also determined their movement. Despite the fact that the people of Ashīṯā lived in sedentary settlements, they often had to move around in order to accommodate their flocks. Here, once again, rāwē was employed to reflect such socio-economic circumstances, as is demonstrated below. In Ashīṯa, winter would start towards the end of October, and it would begin to snow by the end of November. From December to February, it would snow heavily, and the people who remained in their villages would not leave their homes except to attend the church. Many of the people—mostly young men and women—would travel down to the plains, usually to the areas between Alqosh and Mosul, and Tūr Alpap/Alfaf; although some would only travel as far as Sarsīng and Sapnā. Those huk, Iraq.

SONGS OF THE ASSYRIAN MOUNTAINEERS

355

who remained in their villages included the elderly, a number of young able men to protect the settlement and some women to care for the children and housekeeping. These aspects of rural life are reflected in rāwē such as the following way: (8)

ܲ ‫ܟܠ ܫܹܢܬ‬ ‫ܒܐܢ ܹܐ ܗ ܹܵܘ ̈ܢ ܹܐ‬ ܲ ‫ܪܒ ܵܗܪ ܵܨܠܝ‬ ܵ ‫ܲܒ‬ ‫ܡܘܥ ̈ܢ ܹܐ‬ ܲ ‫ܡܫ‬ ‫ܸܛܪܘ ܵܒܕܠܝ ܕܘ ܵܟ ̈ܢ ܹܐ‬ kúl every

šíta year

bar-báhar PREP-spring

b-ʾánne in-these

hewáne seasons

ṣáli m-wáne go_down.3PL with-sheep.PL

m-šǝ́ṭro bádli with-beautiful_one swap.3PL

dukáne places

Every year around this season With the approach of spring, they descend along with the sheep And swap places with the gorgeous one (9)

ܵ ‫ܒܡܪܝ ܲܡ ܲܬܝ‬ ܵ ‫ܵܝܡܝܵܢ‬ ‫ܒܓܢ ܹܗ‬ ܲ ‫ܘܒܣܸܬ̈ܪ ܹܐ‬ ‫ܕܡܕܒܚܹܗ ܡܝ ܵܠ ̈ܢ ܹܐ‬ ܵ ‫ܡܒܘܪ ܵܟܢ ܹܗ‬ ܲ ‫ܵܒܒܝ‬ ‫ܒܓܢ ܹܗ‬ yámyan promise.1SG.F

b-már máttay by-saint Mattay

b-ján-e in-REFL.PRON-3SG.M

ʾu-b-sǝ́tre and-in-curtains

d-madǝ́bx-e REL-altar-3SG.M

miláne blue.PL

báb-i father-1SG

mbúrx-an-ne marry.PST-OBJ.1SG.F-3SG.M

b-ján-e by-REFL.PRON-3SG.M

356

NINEB LAMASSU

I solemnly promise by St. Matay himself7 And by the blue curtains of its altar I was married off by my father himself (10)

ܵ ‫ܲܫ ܵܡ ܵܫܐ‬ ‫ܕܓܘ ܒܹܝܬ ܲܩܢ ̈ܟܹܐ‬ ‫ܲܥܠܹܗ ܚܘ ܵܪ ܵܪܐ ܕ ܲ̈ܪܢܓ ܹܐ‬

ܵ ‫ܵܩܪ ܹܐ‬ ‫ܘܥܓ ܵܒܐ ܠܝ ܲܕܢܓ ܹܗ‬ šamáša deacon

d-gú REL-inside

be-qánke house.GEN-sanctuary

ʾáll-e on-3SG.M

xuráre gridles

d-ránge REL-colours

qáre read.3SG.M

ʾu-ʾajbá-li and-please.3SG.F-OBL.1SG

dáng-e voice-3SG.M

[He is] a deacon [standing] in the sanctuary Dressed in a multi-coloured girdle He recites and his voice pleases me (11)

ܵ ‫ܝܘܢ ܗ ܵܘܐ ܠ‬ ܵ ‫ܐܙܝ‬ ܵ ‫ܠܬܐ‬ ‫ܸܩܪ ݂ܵܒܐ‬ ܵ ݂ ‫ܚܙܹܐ ܠܝ ܙܝ ܵܢܐ ܒܪ ܹܫ‬ ‫ܟܬ ݂ܵܒܐ‬

‫ܦܠܝܛ ܠܝ ܒܠܸܒܝ ܚܘ ܵܫ ݂ܵܒܐ‬ zílta=wanwa go.RES.SG.F=COP.PST.1SG.F xzé-li see.PST.OBJ.3SG.M-1SG

7

lǝ-qráwa to-receive_communion.INF

zína standing

b-réš at-on

kṯáwa book

Mar Matay is a famous monastery located on Tūr Alpap near Mosul.

SONGS OF THE ASSYRIAN MOUNTAINEERS plíṭ-li go_out.PST-1SG

b-lǝ́bb-i in-heart-1SG

357

xušáwa thought

I went to church to receive communion I saw him standing over the book [of Psalms] I came out, and my heart filled with a desire [for him] By March, the people would begin to return to their villages and start working the fields. In April and May they would start by planting the seeds, thus we often find a lover expressing the fear that the beloved might be over-worked in the fields: (12)

ܲ ‫ܘܙ ̈ܓ ܹܐ‬ ܵ ‫ܒܐ‬ ܵ ‫ܵܝܘܢܝ‬ ܵ ‫ܝܠ ݁ܗ‬ ‫ܩܠ ݁ܗ‬ ܵ ‫ܘܝܬܝ ݂ܒ ܵܬܐ ܠܕܘ ܵ̈ܪܢ ܹܐ ܕܚ‬ ‫ܸܩܠܐ‬

ܵ ‫ܕܝܠܝ‬ ‫ܕܒ ܲܘܩ ܵܠ ݁ܗ‬ ݂ ‫ܝܠ ݁ܗ ܲܚ ݂ܒܪܝ‬ yáwn-i=la pigeon-1SG=COP.3SG.F

ʾu-záge and-bells

tíwta sit.RES.SG.F

d-xǝqla REL-field

l-duráne on-edges

díyyi=la mine=COP.3SG.F

xáwr-i friend-1SG

b-ʾáql-a on-foot-3SG.F

dúq-la catch.IMP.SG-OBJ.3SG.F

She is my pigeon, with bells on her foot Sitting on the edging of the field She is mine, catch her, [least she falls], my friend. By June, they would start renovating their homes which had physically deteriorated over the course of the year: (13)

ܵ ‫ܠܒܢ‬ ܵ ‫ܕܡܚܹܐ ܵܪ‬ ܵ ‫ܵܛ‬ ‫ܘܕ ܵܢܐ‬ ܵ ‫ܕܬܠܚܝ ܓܘ ܵܕܐ‬ ܵ ܵ ‫ܘܐܦ‬ ‫ܣܘ ܵܢܐ‬ ‫ܲܨ ܵܢܥܐ ܵܕ ݂ܒ ܵܩܐ ܠܝ ܵܐ ܵܢܐ‬

ṭálban ask.1SG.F

d-máxe REL-hit.3SG.M

rawḏána earthquake

358

NINEB LAMASSU

d-tálxi REL-perish.3PL ṣánna artisan

gúda wall

ʾu-ʾáp and-also

dáwqa-li catch.3SG.F-OBJ.1SG

swána eaves

ʾána

PRON.1SG

I wish for an earthquake to strike May the wall collapse along with the eaves So that she hires me as a [her] artisan In July, when the weather became hot, a great number of the people would migrate to their summer pastures, where they would build mashǝknē or kwīnē, i.e. ‘tents’, or what was referred to as qūprāne (‘dwelling booths’), and ērzālē elevated ‘portable beds’, to avoid bugs and enjoy the summer breeze at night. Once again, some members of the community would remain in the village while the young able men would go and gather firewood and fodder for the sheep to be stored for winter. We know the people of Ashīṯā had to travel considerable distances to gather enough fodder for their large number of sheep. According to D’Ashitha, they had to travel at least two to five miles on mountainous terrain twice a day, carrying as much as 150 pounds on their backs (D’Ashitha 1982: 41): (14)

ܲ ‫ܙܘ ̈ ܹܡܐ ܠܹܐ‬ ܲ ‫ܝܠܝܗܝ ܲܒܣܝ ̈ ܹܡܐ‬ ܲ ‫ܬܪ ܹܝ ܩܘܪ̈ܪ ܹܐ ܠܹܐ‬ ‫ܝܠܝܗܝ ̈ܪܝܡܹ ܐ‬ ‫ܦܝ ̈ܫܹܐ ܒܚ ̈ ܹܸܡܐ ܲܚܡܝ ̈ ܹܡܐ‬ zóme summer_pastures d-tre REL-two píše remain.RES.PL

lélay basíme COP.NEG.3PL joyful.PL

k̭úrre lélay ríme lads COP.NEG.3PL go_up.RES.PL b-xǝ́mme in-heat.PL

xamíme hot.PL

The summer-pastures are bereft of joy For two lads have not ascended [to them] They are left behind in the scorching heat

SONGS OF THE ASSYRIAN MOUNTAINEERS

359

(15)

ܵ ‫ܩܘ‬ ܲ ‫ܦܪ ܵܢܐ ܕܹܐ ܐ‬ ‫ܣܘܪܘ ܵܠ ݁ܗ‬ ‫ܐܸܢ ܛܠܝ ݂ ܵܬܐ ܲܡܪܥܫܘ ܵܠ ݁ܗ‬

‫ܠܒܹ̈ܪܝܝ ܹܐ ܪܚܝܫ ܲܠܝܗܝ ܚܘ ܵ̈ܪ ݁ܗ‬ quprána booth ʾǝ́n if

dé

PART

sorú-la tie.IMP.PL-OBL.3SG.F

ṭlíṯa asleep.RES.SG.F

l-beríyye by-milkmaids

marrǝ́ššu-la wake.IMP.PL-OBL.3SG.F

xíš-lay go.PST-3PL

xúr-a friends-3SG.F

Hasten and build her a booth If she is asleep, do wake her up Her friends the milkmaids have already gone (16)

ܵ ‫ܘܠܐ ܪܥܝ‬ ܵ ‫ܝܘܢ ܗ ܵܘܐ‬ ܵ ‫ܛܠܝܥ ݂ ܵܬܐ‬ ‫ܫܬܐ‬ ܵ ‫ܸܪܙܠܝ ܲܠ ܵܗ ݁ܘ ܫܥܝ‬ ܵ ‫ܘܥ‬ ‫ܫܬܐ‬ ‫ܲܚ ݂ܒܪܝ ܵܡ ܵܪܐ ܕܓܸܫܢ ܸܫ ܵܬܐ‬ ṭlíṯa=wanwa asleep.RES.SG.F=COP.PST.1SG.F ʾu-ʾǝ́rzal-i and-bed -1SG xáwr-i friend-1SG

la-háw by-3SG.M mára owner

ʾu-lá and-NEG šíšta shaken.SG.F

d-gǝ́žnǝ́žta REL-earring

I was fully asleep and not awake It was he who shook my portable bed My lover, the one with an earring

ríšta awake.RES.SG.F

360

NINEB LAMASSU (17)

ܵ ‫ܵܦܥ ̈ ܹܠܐ ܕܟܝ ݂ܵܒ ܵܪܐ‬ ‫ܕܐܙܠܝ‬ ܲ ܵ ‫ܕܝ‬ ܵ ‫ܘܓܙܝ‬ ݂ ܵ ‫ܗܒܢ ܠܝܗܝ ܵܫܗܝ‬ ܵ ‫ܕܒ ܵܠܐ ܲܡܣܝܡܝ‬ ܵ ‫ܠܒܙܝ‬ pále labourer.PL

d-kiwára REL-thistle

d-yáhwan-lay FUT-give.1SG.F-OBL.3PL d-bála REL-attention

d-ʾázi REL-go.3PL šáhi šahi

masími pay.3PL

ʾu-ġázi and-ġazi l-báz-i to-lover-1SG

The labourers that are going to gather thistle I will give them a Šahi and a Ġazi8 So they would look after my lover (18)

ܲ ‫ܲܒܣܝ ̈ ܹܡܐ‬ ‫ܝܠܝܗܝ ܵܙܘܙܵ ̈ܢ ܹܐ‬ ܵ ܹ ‫ܚܒܝ ̈ܟܹܐ ܒ̈ܪ‬ ‫ܝܚܢ ܹܐ‬ ݂ ‫ܲܘ̈ܪ̈ܕܹܐ‬ ܵ ‫ܟܠ‬ ܵ ‫ܡܬ ܵܡܐ‬ ܵ ‫ܕܐܬ ܹܐ‬ ‫ܡܬܢ ܹܐ‬ basíme=lay nice.PL=COP.3PL wárde flowers

zozáne summer_pastures

xwíxe mixed.PL

kúl d-ʾáte every REL-come.3SG.M

8

Turkish and Persian currency.

b-rexáne with-thyme.PL m-táma from-there

mtáne tell.3SG.M

SONGS OF THE ASSYRIAN MOUNTAINEERS

361

The summer pastures are pleasing (nice) Mixed with flowers and thyme Whoever comes from there says so The people would remain in the summer pastures until September, at which time they would all descend to their villages, their sheep having fattened. They would remain in their villages until the beginning of winter in mid-October and then they would go down to kūzāwāṯā (‘winter pastures’) to repeat the cycle again. It was in September that the feast of Mār Sāwā was held in upper Ṭyārē. The people of Ashīṯa, like all Assyrians of Hakkārī, would gather there to celebrate this occasion: (19)

ܵ ‫ܵܙ ݂ܵܒܐ ܙܝ݂ܕ ܲܠܝܗܝ ܲܥ‬ ‫ܝܢ ̈ܬ ܹܗ‬ ܲ ܵ ‫ܘܬ‬ ‫ܕܒܝܩ ܠܹܗ ܐܘ̈ܪ̈ ܵܚܬ ܹܗ‬ ݂ ‫ܠܓܐ‬ ܵ ‫ܕܠܐ ܵܡܨܐ‬ ܵ ‫ܘܨ݂ܕܠܝ‬ ܲ ‫ܕܐܬ ܹܐ‬ ܹ záwa Zab

zíd-lay increase.PST-3PL

ʾu-tálga and-snow

ʾenáṯe springs

dwíq-le block.PST-3SG.M

ʾu-ṣáḏ-li d-lá and-fear-OBL.1SG REL-NEG

ʾurxáṯe ways.3SG.M

máṣe d-ʾáṯe be_able.3SG.M REL-come.3SG.M

The Zāb source-waters have risen And snow has blocked his passages I fear he may not be able to come across (20)

‫ܬ ܸܫܪܝܢ ܝܠܹܗ ݁ܗܘ ܲܩܕ ܵܡܝܵܐ‬ ܵ ܲ ܵ ‫ܡܪܐ ܒܣܝ ܵܡܐ ܠ‬ ‫ܸܫܬ ܵܝܐ‬ ‫ܚ‬ ܵ ܲ ‫ܲܚ ݂ܒܪܝ‬ ‫ܠܙܘ ̈ ܹܡܐ ܝܠܹܗ ܪܦܝܵܐ‬ tešíri=le Tešri=COP.3SG.M

ʾó that

qamáya first.M

362

NINEB LAMASSU xámra wine

bsíma become_pleasant.RES.SG.M

l-štáya to-drink.INF

xáwr-i friend-1SG

l-zóm=ele rpáya to-summer_pastures=COP.3SG.M dash.INF

It is Tešri the first9 The wine has matured [ready] for consumption My beloved is returning from the summer-pastures (21)

‫ܵܚ ܵܙ݂ܕܝ ܬܝ ܵܦܐ ܬܝ ܵܦܐ‬ ܵ ‫ܲܡ‬ ‫ܓܠܐ ܒܐܝܕܹܗ ܲܚܪܝ ܵܦܐ‬ ܲ ‫ܙܥܘ ܵܪܐ ܝܠܹܗ ܵܗܕܹܐ ܝܠܝ ܵܦܐ‬ xazáḏ-i reaper-1SG

típa bent.SG.M

mágla sickle

b-ʾíḏ-e in-hand-3SG.M

zór=ele young=COP.3SG.M

háde just

típa bent.SG.M xarípa sharp lípa learnt.SG.M

My reaper’s posture is bent [With] a sharp sickle in his hand He is young [and] only just a novice (22)

ܲ ܲ ‫ܠܐ ܲܢܝ ܲܝ̈ܪ̈ܚܹܐ‬ ‫ܕܐܘܪܝ ܗ ܵܘܘ‬ ܲ ‫ܘܫ ܵܒܬ̈ ܹܐ ܲܡܠܝܙܝ ܗ ܵܘܘ‬ ܵ ‫ܕܐܙܹܠܚ ܗ ܵܘܘ ܠܫܹܗ ܵܪܐ ܕܨܠܝ ݂ܵܒܐ‬

9

I.e. October.

SONGS OF THE ASSYRIAN MOUNTAINEERS l-ʾánne to-these

yárxe months

363

d-ʾóri-wa PTC-pass.3PL-CONV

ʾu-šabáte and-weeks

malyǝ́zzi-wa hurry.3PL-CONV

d-ʾázǝx-wa REL-go.1PL-CONV

l-šéra to-festival

d-ṣlíwa REL-cross

[If only] these month would pass And the weeks would speed on So we could go to the Festival of the Cross The rāwē songs also convey the Assyrian people’s sense of unity and national consciousness, often lamenting the lost heritage of their nation and expressing hopes for its revival: (23)

ܵ ‫ܡܚ‬ ܵ ‫ܵܛ‬ ܵ ‫ܠܒܢ ܕܛܝܵܪ ܹܐ‬ ‫ܝ݂ܕܐ ܗ ܵܘܬ‬ ܵ ܵ ‫ܕܒ‬ ܵ ‫ܘܐ ܵܢܝܘ ݂ ܵܬܐ‬ ‫ܛܠܐ ܗ ܵܘܬ‬ ܵ ‫ܘܡܐ‬ ܵ ‫ܵܩ‬ ܵ ‫ܕܐ ݂ܬ ܲܘܪ ܵܩ‬ ‫ܝܡܐ ܗ ܵܘܬ‬

ṭálban ask.1SG

d-ṭyáre REL-Ṭyare

ʾu-ʾanayúṯa and-selfishness qáwma maybe

mxáyḏa-wa unite.3SG.F-CONV

d-báṭla-wa REL-cease.3SG.F-CONV

d-ʾáṯur REL-Assyria

qáyma-wa rise.3SG.F-CONV

I wish Ṭyārē would unite And selfishness would be abolished Maybe then Assyria would rise (24)

ܲ ‫ܡܘ ܲܚ ܵܢܐ‬ ܲ ‫ܗܘܠܝ‬ ܲ ‫ܡܐ ݂ܬ ܵܪܐ‬ ܵ ܵ ‫ܘܨ‬ ܲ ‫ܵܙ ݂ܟ ܲܘ‬ ‫ܦܢܐ ܘܢܘ ܲܗ݂ܕܪܐ‬ ܵ ‫ܠܩܘܫ ܘܫܒ ܲܘ ܵܬ ݁ܗ ܵܒ ܲܥ‬ ܲ ܲ ‫ܡܐ‬ ‫݂ܕܪܐ‬ ݂ ݂

364

NINEB LAMASSU

hóli COP.1SG záxo Zaxo

muxánna long.RES.SG.M ʾu-ṣápna and-Ṣapna

m-ʾálqoš from-Alqosh

m-ʾáṯra from-country

ʾu-nuháḏra and-Nuhaḏra

ʾu-šwóṯ-a and-neighbour.F-3SG.F

baʾáḏra Baʿadra

I have longed for [my] country Zaxo, Ṣapna and Nuhadra (Dohuk) Alqoš, and its neighboring Baʿadrā (25)

‫ܵܐ ݂ܬ ܲܘܪ ܛܘ̈ܪ ܹܐ ܵܟܪܟܝ ܵܠ ݁ܗ‬ ̈ ‫ܟܸܟ ̈ܒܹܐ‬ ‫ܕܫ ܲܡܝܵܐ ܵܚܦܩܝ ܵܠ ݁ܗ‬ ܲ ‫ܲܘ̈ܪܕܐ ܘܒܝ ̈ܒܠܹܐ‬ ‫ܡܣܩܠܝ ܵܠ ݁ܗ‬ ܹ ʾáṯur Assyria

ṭúre mountains.PL

čárxi-la surround.3PL-OBJ.3SG.F

kǝ́xwe stars

d-šmáyya REL-sky

xápqi-la embrace.3PL-OBJ.3SG.F

wárde flower

ʾu-bíble and-camomiles

msáqli-la adorn.3PL-OBJ.3SG.F

Assyria is surrounded by mountains Embraced by the stars of the sky Adorned with flowers and camomiles This primary objective of this paper has been to offer an alternative approach to the study of modern Assyrian oral literature. The concern of many scholars has hitherto been to record this corpus solely for the documentation of the various dialects in order to subject them to philological and grammatical analysis, as they are fast disappearing. I hope I have shown here the value of studying the literary features of the texts and their reflection of the society in which they were composed.

SONGS OF THE ASSYRIAN MOUNTAINEERS

REFERENCES

365

Aboona, Hermis. 1999. Al-Ashuriyun Baʿda Suqut Naynawa, vol. 10, Al-ʿAshaʿer AlAshuriyya Al-Mustaqella Fi Tyari Wa Hakkari Wa Al-Aqqalim Al-Ashuriyya AlMuhita Biha. Chicago: Alpha Graphics. [in Arabic] —. 2008. Assyrians, Kurds, and Ottomans: Intercommunal Relations on the Periphery of the Ottoman Empire. New York: Cambria Press. Al-Badlisi, Sharif Khan ibn Shams Al-Din. 1962. Sharifnama, vol. 1. (Translated by Mohammad ʿAli ʿAuni), vol. 2. Baghdad: Dar Ahyaʿ Al Kutub Al Arabiya. [in Arabic] Al-Nawfali, Fr. Habib. 2002. Baqupa Habbat Kherdal. Baghdad: St. George Chaldean Church. [in Arabic] Ashitha, Odisho M. Gewargis. 1997. Hilqa De Leshana: Assyrian-Arabic Dictionary. Baghdad: Al-Maghreb. —. 2002. Sepra D’Ashitha. Baghdad: Al-Maghreb. [in Arabic] —. 2008. “Men Qanye Sapraye Atiqe.” Maʾalta 3–4 (2): 26–47. [in Suret] Ashitha, Odisho M. Gewargis, and Qasrayta, Susan Yawsep. 1998. Bahare DʾQinatha – Rawe. Baghdad: Al-Maghreb. [in Suret] Badger, George. P. 2004. The Nestorians and their Rituals: with the Narrative of a Mission to Mesopotamia and Coordistan in 1842–1844, and of a Late Visit to those Countries in 1850: also, Researches into the Present Condition of the Syrian Jacobites, Papal Syrians, and Chaldeans, and an Inquiry into the Religious Tenets of the Yezeedees, 2 vols. Kessinger Publishing Co. Bet Benjamin, Daniel Dawed. 2009. Zmiratha D’Lilyana. Arizona: self-published. [in Suret] —. 1998. Zmiratha D’Rawatha. Chicago: self-published. [in Suret] D’Ashitha, Shamasha Gewargis D’Bet Benjamin. 1982. Khayutha Gaw Tyare wu Lwar Men Tyare. Chigaco: Nineveh Press. [in Suret] DeBaz I, Rab Tremma Awdisho Bar Malik Natan. n.d. Mawdʿanuta Bud Atra D’Baz. Unpublished manuscript (personal collection). [in Suret] DeBaz II, Khammo Beqasha. 1899. Personal Diary of Khammo Beqasha DeBaz. Unpublished manuscript (personal collection). [in Suret] DeBaz III, Yawsip Khammo, n.d. A Brief History of Baz. Unpublished manuscript (personal collection). [in Suret] Emhardt, William C., and Lamsa, George M. 1926. The Oldest Christian People: a Brief Account of the History and Traditions of the Assyrian People and the Fateful History of the Nestorian Church. New York: Macmillan. Fiey, Jean M. 1964. “Proto-histoire chrétienne du Hakkari Turc.” L’Orient Syrien 9: 443–472. Fox, Samuel Ethan. 1997. The Neo-Aramaic Dialect of Jilu. Wiesbaden: Harrassowitz.

366

NINEB LAMASSU

Giwu, Shamasha Adham Sawa. n.d. Untitled Chronicle of Events and Documentation of Oral Literature. Unpublished manuscript (personal collection). [in Suret] Grant, Asahel. 1844. The Nestorians, or, The Lost Tribes: Containing Evidence of their Identity, their Manners, Customs, and Ceremonies: with Sketches of Travel in Ancient Assyria, Armenia, Media and Mesopotamia, and Illustrations of Scripture Prophecy. London: John Murray. Hoberman, Robert D. 1990. “Reconstructing Pre-Modern Aramaic Morphology: The Independent Pronouns.” In Studies in Neo-Aramaic, edited by Wolfhart Heinrichs, 79–88. Atlanta: Scholars Press. Joseph, John. 2000. The Modern Assyrians of the Middle East: Encounters with Western Christian Missions, Archaeologists, and Colonial Power. Leiden: Brill. Lamassu, Nineb. 2009. “The Female Voice in Rāwe: The Strive for Gender Equality.” Journal of Assyrian Academic Society 23 (2): 38–50. —. 2013. “Süryani Kadim’in Etnik Kimliği ve Milli Rönesansı.” Mardin ve Çevresi Toplumsal ve Ekonomik Tarihi Konferansı, edited by Emre Ayvas and Altuğ Yılmaz, 172–182. Istanbul: Mardin Tebligleri, Hrant Dink Vakfı Yayınları. [in Turkish] Maclean, Arthur J. 1888. Some Account of the Customs of the Eastern Syrian Christians, otherwise Variously Known as Assyrians, Chaldeans, or Nestorians. London: Office of the Archbishop’s Mission to the Assyrian Christians. Maclean, Arthur John, and Browne, William Henry. 1892. The Catholicos of the East and His People. London: SPCK. Mengozzi, Alessandro. 2002. Israel of Alqosh and Joseph of Telkepe: a Story in a Truthful Language: Religious Poems in Vernacular Syriac (North Iraq, 17th century). Lovanii: Peeters. Pennacchietti, Fabrizio. 1985–1986. “Zmiryata-D Rawe.” Journal of Assyrian Academic Studies 1: 39–44. Sher, Elias. 2007. Egratha D’Qashisha Elias Sher. Ankawa: Knushta D’Mardutha Kaldayta. [in Syriac] Travis, Hannibal. 2006. “Native Christian Massacred”: The Ottoman Genocide of the Assyrians during World War I.” Genocide Studies and Prevention 3 (1): 327–371. Wigram, William A., and Wigram, Edgar T. A. 1922. The Cradle of Mankind: Life in Eastern Kurdistan (2nd ed.). London: Black. Youkhanna, Fr. Emanuel. 1998. Zmiratha D’Rawe. Jönköping: Ashurbanipal Press. [in Suret]

NEO-MANDAIC IN MANDAEAN MANUSCRIPT SOURCES

1

MATTHEW MORGENSTERN 1. INTRODUCTION

Mandaic is a south-eastern variety of Aramaic that is closely related to the Jewish Babylonian Aramaic of the Talmud and post-Talmudic rabbinic literature. It is essentially a communal dialect, in that all attested sources demonstrating uniquely Mandaic features are written in the script that is employed for Mandaean religious documents.2 Furthermore, the overwhelming majority of the texts written in this lan1

This article is based upon a lecture presented at the conference Neo-Aramaic Dialectology: Jews, Christians, and Mandaeans, held at the Institute of Advanced Studies, Jerusalem on 26–27 June 2013. I wish to thank the organizers of the conference, Professors Simon Hopkins, Steven Fassberg and Hezy Mutzafi for inviting me to speak, and Professor Geoffrey Khan for accepting the paper in its present form for publication. I also wish to thank Dr. Tania Notarius, Maleen Schlüter, Elisheva Bard, Tom Alfia and Livnat Barkan for their assistance in preparing the materials discussed herein. Professors Shaul Shaked, Hezy Mutzafi, Charles Häberl and James Nathan Ford kindly shared with me their unpublished works. Citations from the Rbai Rafid Collection are reproduced by kind permission of the custodian of the collection. Mandaic written sources are provided in Mandaic script followed by a Latin transliteration in round brackets, e.g. #‫ר‬#‫( חיו‬hiuara) ‘white (m.sg.)’. The transliteration system follows that employed Rudolf Macuch, except for ‫ ע‬and ! (the Arabic pharyngeal ‫ ﻉ‬employed exclusively in loanwords and proper nouns), which are represented respectively by (ʿ) and (ʕ). Neo-Mandaic words from spoken sources are presented in italics, e.g. həwɔrɔ ‘white’ according to a standardized transcription, while for the sake of simplicity, in the references preference has been given to the glossaries that accompany these texts. This research was supported by the Israel Science Foundation grant no. 419/13. 2 Exceptions are citations of Mandaean literature found in Syriac, and some later manuscripts with accompanying transcriptions into Arabic letters to aid the study of the texts or to

368

MATTHEW MORGENSTERN

guage contain elements that reflect the Mandaean religion. This includes the evergrowing corpus of spells written on clay bowls and lead scrolls, which according to most estimations were written between the fourth to seventh centuries CE3 and provide the earliest material evidence for the language, even though the original date of the compositions contained in both this epigraphic corpus and the later manuscripts may be earlier and extend as far back as the second century CE. Most Mandaean texts are not preserved in early epigraphic sources but rather in much later manuscripts. The earliest Mandaean manuscript known to scholarship remains the Bodleian Library’s codex Marsh. 691, a small selection of prayers copied in Ḥuweiza in 936 AH (1529–1530 CE). Very few manuscripts survive from the 16 th century, and several works survive only in manuscripts that were copied as late as the 20th century. The lack of primary textual sources over a period of some 900 years presents a considerable impediment to the diachronic study of Mandaic. Yet while the fact that the available textual witnesses were copied so long after the redaction of the works that they contain is very frustrating for the study of pre-modern Mandaic, thanks to the researches of Jorunn Buckley, it has become clear that even late manuscripts may teach us much about the more recent history of the Mandaean community.4 During the course of our research projects related to the preparation of a new dictionary of the Mandaic language,5 considerable efforts have been made to expand the corpus. Thanks to the collaboration of several scholars and members of the Mandaean community around the world, these labours have proven successful, and we are now in possession of twice as many written sources than were previously available.6 These cast significant light not only on the textual transmission and the contents of Mandaean literature, enabling us to make many improvements to the interpretation of the Mandaean lexicon (the primary aim of the research projects), but also cast considerable light on the pre-history of contemporary Neo-Mandaic (NM).

render them readable to lay Mandaeans who need to recite them as part of the ritual. 3 On the archaeological evidence see, e.g., Hunter (1995). 4 See in particular Buckley (2010). 5 Two research projects have been funded by the Israel Science Foundation: ‘Materials for a New Dictionary of Mandaic’ (2010–2014) and ‘The Lexical Analysis of Mandaic’ (2013– 2016). 6 The largest single contribution has been the acquisition of digital images of the Rbai Rafid Collection. A full account of the current state of evidence for Mandaean literary works is presented in Morgenstern (in preparation).

NEO-MANDAIC IN MANDAEAN MANUSCRIPT SOURCES

369

As we shall see in what follows, the discovery of NM as a linguistic entity worthy of philological study was a gradual process. Macuch’s claims to have been the first scholar to discover that Mandaic was still spoken have now been conclusively disproven by Häberl,7 but his contributions to the study of NM were for many years the only significant source of information about the idiom. The recent publications by Häberl and Mutzafi represent valiant efforts to salvage from aging informants what remains of this critically endangered language, as well as to submit Macuch’s findings and conclusions to closer scrutiny.8 The current paper aims to provide an outline of the sources available to us for understanding the development of NM prior to the pioneering fieldwork of Rudolf Macuch in the 1950s. While some of these have been mentioned previously in the literature,9 others have not been given sufficient attention. Furthermore, our recent manuscript discoveries provide valuable new evidence for the earlier stages of NM. Within the scope of this article, it is not possible to discuss all these new findings in detail, and these will be presented in separate studies that consider the different types of sources in greater depth.

2. EINE SPRACHE OHNE FORSCHUNGSGESCHICHTE

In describing the state of scholarship in NM in his groundbreaking (if flawed) chrestomathy of NM texts, Rudolf Macuch observed: Obwohl man schon seit den ersten europäischen Kontaken mit den Mandäern durch katholische Missionare des 17. Jh. und später durch Reisende und Orientalisten, die zu diesem Völkchen durchgedrungen waren, wußte, daß die Mandäer ein der Sprache ihrer heiligen Bücher verwandtes Idiom sprachen, hat das gesprochene Mandäisch im Gegensatz zu den neusyrischen und westaramäischen Dialekten keine Forschungsgeschichte.10

Macuch ascribed this lack of recognition to the Mandaeans’ own suspicion of outsiders. However, it seems that Macuch’s statement is both an exaggeration and, at the same time, overlooks an important source for scholarly scepticism regarding NM.11 7

Häberl (2009: 25–26), and see further below. Häberl (2009, 2010; 2013); Mutzafi (2014) and Mutzafi (forthcoming). 9 See in particular Macuch (1989: 4–5): ‘Ältere und neuere Quellen der mandäischen Volkssprache’. 10 Macuch (1989: 5). 11 For an account of the exposure of western scholars to Mandaic prior to Macuch’s studies 8

370

MATTHEW MORGENSTERN

Theodore Nöldeke’s seminal Mandäische Grammatik of 1875 provided what remains the most important description of Classical Mandaic (CM) as it is found in manuscripts. (The epigraphic corpus was unknown at that time.) In summarizing the nature of his sources, Nöldeke ascribed a particularly late date to two of them: Die letzte Form der Sprache zeigen endlich die jüngsten Theile des Asfar Malwâše und die Berichte der Abschreiber über ihre Zeit (16. Jahrh. bis zur Gegenwart). Hätten wir in diesen Stücken wirklich einen modernen lebenden Dialect, so wären sie von grosser Wichtigkeit; aber sie bieten uns nur ein unerquickliches Gemisch von Formen der alten Sprache, welche man noch immer zu schreiben meint, und ganz jungen. Nicht bloss der Wortschatz, sondern auch die Grammatik ist von arabischen und persischen Elementen durchdrungen. Man sagt z. B. ‫„ ראבתאר‬grösser“ mit dem persischen Suffix tar, und gebraucht im aramäischen Text arabische Formen wie ‫يظهر = יידהאר‬. Ein Studium der lebenden Sprache, welche den Texten zu Grunde liegt, wäre natürlich von Interesse, aber diese wird hier eben wegen der Rücksicht, die man auf die alte Sprache und vielleicht auch auf fremde Schriftsprachen nimmt, durchaus nicht treu ausgedrückt.12

Nöldeke’s opinion was similarly negative regarding the 17th century Glossarium, now known as the Leiden Glossarium but at that time in Amsterdam: In diesen letzten Zeitraum fällt auch das von einem katholischen Missionär mit Hülfe eines Mandäers verfasste arabisch-mandäisch-lateinisch-persischtürkische Glossar, welches in einem Amsterdamer Codex enthalten ist. Dasselbe erweist sich bei äusserst behutsamem Gebrauch nützlich, kann aber den Unkundigen leicht sehr stark irre führen. Der Verfasser selbst hat sich oft genug geirrt und hatte keine Kenntniss von der Literatur.13

This is not to say that Nöldeke regarded Mandaic as a dead language, as he clarified shortly thereafter: Gern hätte ich die Entwicklung des Mandäischen bis auf unsere Zeit dargestellt, aber wir haben eben nur für die ältere Periode zuverlässige Quellen, da ja, wie gesagt, die jüngeren Schriften keineswegs die Sprache ihrer Gegenwart rein darstellen.14 see Häberl (2009: 16–26) and Mutzafi (2014: 5–9). 12 Nöldeke (1875: XXIV–XXV). 13 Nöldeke (1875: XXV). 14 Nöldeke (1875: XXV).

NEO-MANDAIC IN MANDAEAN MANUSCRIPT SOURCES

371

We must recall that Nöldeke had no direct contact with the Mandaeans, and was unable to determine the true nature of the spoken Mandaean language. By contrast, Ethel Stefana Drower, the most important western student of Mandaean culture,15 encountered the spoken Mandaic of Iran in the 1930s when she visited Khorramshahr with Sheikh Abdallah Khaffagi.16 Drower also recorded a version of the popular Mandaean legend The Bridge of Šuštar in a Latin transcription and partial translation. 17 This experience with spoken Mandaic seems to have influenced Drower’s view of the language of the Book of the Zodiac, as she writes: I venture to think that Noldeke (sic) is mistaken, and that the language is not artificially archaic, but represents a transitional period. In the later fragments, in which Arabic and Persian elements are, as he says, very evident, we get something very near the spoken Mandaean of today, hence, philologically, it is of importance.18

As we shall see below, both Drower and Nöldeke were correct to a degree. While the literary idiom of almost all late Mandaic texts is to a degree archaizing

15

It is interesting to note that the most accomplished translator of Mandaean texts, Mark Lidzbarski, whose publications provided the foundation for all subsequent research of Mandaean literature, had almost nothing to say about late Mandaic, beyond noting that Arabic loanwords indicate that some of the Mandaean wedding poems, drawn from folk songs are late. See Lidzbarski (1920: XI ft. 2), and the discussion of NM poetry below. 16 Buckley (2010: 115). An account of Drower’s visit, and her impressions upon seeing Mandaean children conversing in the Mandaic language, is provided by Nasser Sobbi in Häberl (2009: 276–277). 17 The text has been published with a translation and notes in Häberl (2013). Examination of Drower’s notes, kindly made available to me by Häberl, reveals a fair level of competence on Drower’s behalf, and a good ability to relate the words in her phonetic transcription to the related lexemes in literary Mandaic, Arabic and Persian. All of this belies Macuch’s claim (Macuch 1965: xlvi) that “The existence of a vernacular dialect spoken by Mandaean laymen in Khuzistan, which as a living language deserves, at least, as much attention as the traditional pronunciation of the literary tongue, remained completely unknown until my personal discovery at the occasion of my visit to the Mandaean community of Ahwāz in 1953”, as Häberl (2009: 25–26) justifiably noted. 18 Drower (1949: 2). Drower’s recognition of Neo-Mandaic as a language was tempered by a distinct distaste for it. Even in her last published statement on the matter, she described it as “a modernized and debased form of the tongue, incorporating a number of Persian and Arabic words” (Drower 1960: 1).

372

MATTHEW MORGENSTERN

and influenced by the literary language of earlier Mandaean works, the late texts are replete with neologisms and expressions drawn from the vernacular.

3. WHAT IS NEO-MANDAIC?

As we mentioned, there are no primary textual witnesses for the Mandaic language between the inscribed bowls and lead scrolls of the late Sassanid and early Islamic period and the manuscripts of the 16 th century. While it is certain that texts were being copied and composed during this period,19 they are only known today from later exemplars. The precise dating of almost all works of Mandaic literature is impossible, but it seems likely that the Ginza Rba was redacted in something close to its present form in the early Islamic period. However, Ginza Rba as it survives is unlikely to be linguistically identical to the work as it was redacted. Even a superficial comparison of the epigraphic texts with the earliest surviving manuscripts reveals differences, most notably in the orthography; while in the epigraphic texts, the use of matres lectionis to mark medial a/ā vowels (and to a lesser extent the other full vowels) is optional, in the Mandaic manuscripts it is obligatory. This implies that the copies of early works now in our hands have undergone some linguistic change.20 Owing to the large gap in primary sources, it is hard if not impossible to isolate when specific grammatical or lexical neologisms entered the Mandaic language. This is particularly true in the fields of phonology and morphology.21 Ostensibly late elements found in Mandaean works may result from the later copyist who consciously or unconsciously introduced vernacular elements into the written text, and not stem from the time that the composition was originally put down in writing. For example, #‫ר‬#‫( חיו‬hiuara) ‘white (m.sg.)’ and #‫רתי‬#‫( חיו‬hiuartia) ‘white (f.sg.)’ appear in Gy 9: 8 and 15 according to CS 1, the oldest manuscript of the Ginza Rba which was copied in 968 AH (c. 1560 CE). By contrast DC 22, which was copied in 1247 AH (c. 1831 CE) reads #‫ר‬#‫ו‬#‫( ח‬hauara) and #‫רתי‬#‫ו‬#‫( ח‬hauartia). #‫ר‬#‫( חיו‬hiuara) is the older form, and is found in the epigraphic corpus (e.g. BM 117880 [Segal 19

Buckley (2010: 231–273). The language of even the earliest manuscripts differs in many aspects from that of the epigraphic texts, most of which have not been published. A comprehensive study of the orthography, phonology and morphology of the epigraphic texts is now being prepared by Mr. Ohad Abudraham at Ben Gurion University, Be’er Sheva, Israel. See meanwhile Morgenstern (forthcoming) for a discussion. 21 Syntactic differences, such as verbal use, are probably less subject to scribal alteration. For an attempt to trace the development of syntactic patterns in Mandaic texts, see the study by Häberl in this volume. 20

NEO-MANDAIC IN MANDAEAN MANUSCRIPT SOURCES

373

081M]:16),22 while #‫ר‬#‫ו‬#‫( ח‬hauara) represents a spelling that reflects NM həwɔrɔ, wherein the shewa may be realized as a rounded vowel [ʊ].23 The digraph ‫ו‬# (au) represents this rounded vowel. The classical spelling, which is retained in early copies of the Ginza Rba, has been substituted in the later manuscript by the more phonetic spelling, which reflects the later developments of spoken language. The distribution of these apparent NM forms is not always according to earlier or later manuscripts. For example, one of the most salient features of NM is the shift of īṯ, ūṯ to ext, oxt.24 This process is already attested in the colophon and ritual instructions in the oldest known manuscript, Marsh. 691,25 and in the Leiden Glossarium in the form #‫( מכת‬mḵṯa) ‘death’, Lat.: mors, Ar.: ‫موت‬ َ (154:1). More examples are now forthcoming from early colophons, e.g. #‫( שופרוכת‬šuprukta) ‘kindness, favour’ (NM ešbəroxtɔ)26 in the colophon of RRC 1C from 1074 AH (1663–1664 CE). This phenomenon is not attested in the epigraphic sources discovered to date, but is attested in the form #‫( אליכת‬alikta) ‘fatty tail’ (NM əlexta)27 in CS 1 (Gy 234:2), which as we have seen is the earliest surviving manuscript of the Ginza. However, in CS 2, which was copied in 1042 AH (1632–1633 CE), i.e. over eighty years later, we find the more conservative form #‫( אלית‬alita). Given what we know about Mandaic in its earlier period, it is likely that #‫( אלית‬alita) is the original reading, since the Ginza is an early text. In DC 12:190, a copy of Pašar Haršia, a Mandaean amulet formula of unknown date copied in 1196 AH (1781–1782 CE), we find #‫( גוריכת‬gurikta) ‘bitch’, and the same reading is shared by RRC 1X which was copied in 1248 AH (1832-1833 CE); but in Ms Berlin or. 8° 3634 d, which was copied in 1231 AH (1816 CE), the reading for the same text is #‫( גורית‬gurita), representing the earlier form

22

242).

23

The correct reading was presented in Müller-Kessler (2001–2002: 131) and Ford (2002a:

See Häberl (2009: 85), where the forms həwārā ‘white’ and həwarānā are transcribed phonetically as [ˈhwɔː.rɔ] and [hʊ.wɛ.ˈrɔː.nɔ]. Note that Macuch (1993: 383) heard an o vowel quite distinctly. 24 Kim (2011: 325); Mutzafi (2014: 172 ft. 122). 25 Nöldeke (1875: 78) recognized it as a late form. 26 Macuch (1989: 245); Macuch (1993: 402). 27 Mutzafi (2014: 92–93).

374

MATTHEW MORGENSTERN

without the NM sound shift.28 Once again, we find the typologically earlier form in a later manuscript; but since the date of this composition is unclear, it is not possible to determine if the typologically earlier #‫( גורית‬gurita) is the original reading or a classicizing correction. Moreover, we must also consider the possibility that some of these “late” elements may have much earlier roots, and represent a hidden non-literary stratum that only rarely found expression in writing.29 For example, the “Neo-Mandaic” form for water, menā/mienā,30 already known from late manuscript sources,31 has now been identified in the Mandaic magic bowls from the late Sassanid or early Islamic period.32 We may also cite a grammatical example: while in CM and the majority of the pre-classical texts the 3m.pl. possessive morphemes (also employed with prepositions) are ‫( –חון‬-hun), ‫( –איחון‬-aihun), ‫( –ון‬-un) or ‫( –איחון‬-aiun) (in pre-classical texts written also with the defective spellings of ‫( –יחון‬-ihun) and ‫( –יון‬-iun)), in NM these forms have been replaced by -u, which is employed also as an object pronoun. Nöldeke already noted that sporadic examples of ‫( ו‬-u) are found attached to 28

Only a few lines later, all of these manuscripts read #‫( חוכצ‬hukṣa) ‘palm-frond’. Drower

and Macuch (1963: 135) derived this word from BA ‫( הוצא‬Sokoloff 2002: 373) and Syriac ‫ܚܘܨܐ‬ (Sokoloff 2009: 430), and indeed the direct cognate of these forms is now found in at least two

pre-classical Mandaic sources as #‫( חוצ‬huṣa): MS 2087/9:22, a lead scroll which I am currently preparing for publication, and JNF 40:19, a magic bowl which will be published in Ford

and Morgenstern (forthcoming). If #‫( חוכצ‬hukṣa) and #‫( חוצ‬huṣa) are related, then we have an exceptional example of the shift of ū to ox in different contexts, though see also NM

həwexṣɔ, həwexṣ ‘a dish made of rice flour, dates and sesame’, derived from həḇiṣā (Mutzafi 2014: 14.). It also remains unclear if Mandaic #‫( עוצ‬ʿuṣa, Drower and Macuch 1963: 334-335)

is related to this word, as Nöldeke (1875: 61) proposed. Mandaic #‫( חוצ‬huṣa)/#‫( חוכצ‬hukṣa)

and its cognates always refer to palm branches, which is not the case with #‫( עוצ‬ʿuṣa). 29 Several characteristic NM forms are already attested in early Jewish Babylonian Aramaic sources; see Morgenstern (2010). 30 Macuch (1989: 236); Macuch (1993: 414–415); Häberl (2009: 338). 31 Drower and Macuch (1963: 242) s.v. maina; 267, s.v. mina. 32

See the correct reading of BM 103365 [Segal 102M]:12, ‫ינה‬#‫( מ‬mainẖ) ‘his water’, pre-

sented in Ford 2002a: 259. The same form is now attested in defective spelling, ‫( מינה‬minẖ), in an unpublished bowl from the Martin Schøyen Collection, MS 2054/72:3.

NEO-MANDAIC IN MANDAEAN MANUSCRIPT SOURCES

375

prepositions in the corpus that he examined, and we now have many more examples of ‫( ו‬-u) and ‫( וה‬-uẖ) from the magic texts.33 A morpheme which might previously have been considered to represent NM influences on manuscript traditions has now been shown to have its roots in the earliest attested levels of the language.34

4. THE LEIDEN GLOSSARIUM

The provenance of the Leiden Glossarium has been convincingly established thanks to the work of Borghero.35 The evidence indicates that it was composed in Basra in 1651 by the Carmeline missionary Matteo de San Giuseppe. As we saw, Nöldeke’s opinion of this text was not favourable. Indeed it is true that the Glossarium contains many errors, and the author appears to have had difficulty in distinguishing between the phonemes s and ṣ (though some of the errors might reflect genuine interchanges between these phonemes).36 Nonetheless, precisely on account of the author’s lack of familiarity with the written language, which Nöldeke regarded as a shortcoming, the Glossarium is an invaluable source of contemporary Iraqi Mandaic. In addition, the author’s unique method of marking Mandaic vowels and plosive and fricative pronunciations of the consonants provides unparalleled information about the phonology of NM in the 17th century.37 Borghero has drawn attention to the fact that while in CM the form of the numeral ‘three’ is #‫ת‬#‫( תל‬tlata), the Glossarium employs #‫ת‬#ֺ ‫( כל‬klaṯa), a NM isogloss that is found in all of its forms and clearly distinguishes NM from Classical Mandaic, and has recalled Macuch’s reaction to this form:38 The author of the “Glossarium Sabico-Arabicum-Latinum-Turcum Persicum” gives some words as pronounced only as a proof of his ignorance of the regular spelling. Reading his work before my visit to the Mandaeans, I was especially struck by the word klata “three” written with an initial k as klata. I imagined that the author was partially deaf. My interviews with the Mandaeans,

33

For a detailed discussion see Morgenstern (forthcoming §4). For additional examples drawn from the field of verbal morphology, see Abudraham and Morgenstern (forthcoming). 35 Borghero (1999–2000). 36 Borghero (2004). 37 Ibid. 38 Ibid. (74–75). 34

376

MATTHEW MORGENSTERN however, proved to me that this dissimilated form represents the real pronunciation.39

However, much evidence previously existed for the authenticity of the Glossarium’s form. Already in 1900, Zwemer had recorded examples of the Mandaic language including the days of the week, amongst which was #‫ חובשאב‬#‫ת‬#‫( אכל‬aklata hubšaba) aklatha hofshaba ‘Tuesday’.40 Moreover, evidence was forthcoming from the colophons of manuscripts already in western collections for the existence of such a form. The colophon of CS 23 (fol. 96b) attests that the manuscript was copied #‫ר אלפ‬#‫ת‬#‫י אב‬#‫ אמ‬#‫ת‬#‫ כל‬... ‫ת‬#‫) לשנ‬lšnat…41 klata amai abatar alpa( ‘in the year of three hundred after a thousand’ i.e. 1300 AH (= 1882–1882 CE). Yet another example is found in DC 36 in an historical account added after the cholera epidemic of 1247 (1831 CE). In line 2725, the scribe corrects #‫ת‬#‫( כל‬klata) to #‫ת‬#‫( תל‬tlata), demonstrating his awareness of both the colloquial form and the literary standard. Borghero has also shown that although much material from the Glossarium was incorporated into Drower and Macuch’s Mandaic Dictionary (henceforth MD), many entries have been omitted.42 There is no apparent justification for the arbitrary selection, and it has led to useful information being omitted. For example, the Glossarium (31:1) records a form #‫( אנונ‬anuna) Lat.: aures, Ar.: ‫اذان‬, i.e. ‘ears’, which must be related to the plural form onɔnɔ ‘ears’ in the dialect of Ahvaz.43 This valuable evidence of an etymologically Aramaic lexeme is not recorded in MD, while numerous loanwords from Arabic found in the Glossarium are presented. Even when the material is recorded, the irregular spellings of the Glossarium’s author and the dictionary’s problematic method of recording entries has on occasion led to the Glossarium’s evidence being separated from the main lexeme it represents. For example, in two places the Glossarium presents a reduplicative plural for the noun sheep: Mand.: ‫י‬#‫ר‬#‫ אנבר‬. #‫( אנברונ‬anbruna, anbrarai), Ar.:‫‏حمالن ‏‬,‫‏حمل‬.

39

Macuch (1965: 1–2). Zwemer (1900: 287). I owe this reference to Charles Häberl. 41 Here an erroneous date has been corrected. 42 Borghero (2004: 64). 43 Macuch (1989: 246); (1993: 392). 40

NEO-MANDAIC IN MANDAEAN MANUSCRIPT SOURCES

377

Lat.: ovis, Turk.: 44 ‫ قويين‬Pers.: 45 ‫( كوسفند‬Gl. 68:13); Mand.: ‫רי‬#‫( אנבר‬anbrari), Ar.: ‫رعية‬, Lat.: grex, Turk.: ‫رعية‬, Pers.: ‫رعيت‬ (Gl. 85:9). In addition to these, we find once a non-reduplicative plural: Mand.: ‫י‬#‫ אנבר‬.#‫( אנבר‬anbra anbrai), Ar.: ‫خروف‬, Lat.: ovis, Turk.: ‫قويين‬, Pers.: ‫( كوسفند‬Gl. 85:9). This reduplicative plural is not mentioned in the MD’s entry ʿmbra ‘sheep, lamb’46 but instead is listed as a separate entry, ambra,47 a singular form that is not attested in the Glossarium at all. Furthermore, the reduplicative plural is also attested in the Mandaic corpus in the form ‫ריא‬#‫( עומבר‬ʿumbraria) in DC 46. 15:13 but is again listed separately in the MD under the lemma umbra.48 It is not surprising that the single attestation of the reduplicated form in the Mandaic literary corpus occurs in the instructions for preparation of an amulet in DC 46 since, as we shall see below, these instructions are replete with NM lexemes and forms. However, a reader of the MD who wished to examine the distribution of the plural forms of this noun in the various historical stages of Mandaic would be hard pressed to locate them in the dictionary. More work remains to be done on comparing the vocabulary and morphology of the Mandaic in the Glossarium to those of the surviving NM dialects.49 The Glossarium has been employed to excellent effect to establish the earlier history of several NM lexemes still used today,50 but further attention needs to be given to the lexemes that are no longer in use. A preliminary survey reveals that many of them are Arabic loanwords, and it is possible that some were local loanwords that were always specific to Basra.51 Some of these loanwords were so deeply entrenched in the 44

For standard Ottoman Turkish ‫قويون‬. For standard Persian ‫گوسپند‬. 46 Drower and Macuch (1963: 352). 47 Ibid. (22). 48 Ibid. (344). The initial ʿ is accidentally omitted in the dictionary, though the form is listed under ʿ. 49 Ms. Tom Alfia of the University of Haifa is currently working on an MA dissertation that will cover aspects of these topics. 50 Mutzafi (2014, passim). 51 The lack of evidence for the dialects of this period makes it difficult to establish the ex45

378

MATTHEW MORGENSTERN

Mandaic of the period that they were used to translate a different Arabic word, e.g. ‫‘ اتلف‬destroy’ (Lat.: perdere) in Gl. 21:15 is translated ‫( ֺאסיא‬aṡia),52 from the Arabic root ḍ-y-ʿ. Similarly, the extremely common root n-z-q ‘be close’ apparently a denominative derivative of Persian nazdiq, has almost entirely disappeared from the modern spoken languages, though nazdiq itself ostensibly survives in the dialect of Khorramshahr53 and naziq is employed in the colophon of DC 35.54 The Glossarium also indicates that many Aramaic roots and lexemes which have now been lost continued to function in 17 th century Iraq, e.g. #‫( חבר‬hbra) ‘friend’, Lat.: amicus, Ar.: ‫( صاحب‬Gl. 156:9) which has now been superseded by Persian dūs,55 or Arabic rafīq.56 The rootׁ‫אפך‬ ֺ (apḵ) and its secondary development ‫תפך‬ ֹ (tpḵ)57 are attested in a very wide range of meanings in the Glossarium: ׁ‫תפך‬ ֺ (tpḵ), ׁ‫גאמתפך‬ ֺ (gamtpḵ), Lat.: turbari, Ar.: ‫اضطرب‬, ‫( يضطرب‬Gl. 9:8–10) ׁ‫תפך‬ ֺ (mtpḵ),ׁ‫גאמתפך‬ ֺ (gamtpḵ), Lat.: venire in dissidium, Ar.: ‫اختلف‬, ‫( يختلف‬Gl. 14:7–8) ׁ‫אתפך‬ ֺ (atpḵ), Lat.: motus, commotio, Ar.: ‫( اضطراب‬Gl. 30:2) ‫אפכאיי‬ ֺ (apḵaii), Lat.: fluctus, Ar.: ‫( امواج‬Gl. 30:3) ‫אפכא‬ ֺ (apḵa), ‫גאפך‬ ֺ (gapḵ) Lat.: seduci, Ar.: ّ‫تضل‬, ّ‫( يتضل‬Gl. 52:3–4) ׁ‫אפך‬ ֺ (apḵ),ׁ‫גאפך‬ ֺ (gapḵ) Lat.: advolvere, revolvere, Ar.: ‫حرج‬, ‫( يحرج‬Gl. 62:1–2)

tent to which Iranian spoken Mandaic differed from that of Iraq during the 17 th century. 52 The strange use of Mandaic ‫( ס‬s) with the diacritical point seems to have arisen from the author’s confusion of this letter with the graphically similar Arabic ‫ص‬. See Borghero (2004: 69). 53 Häberl (2009: 341). 54 Macuch (1955: 362). It is striking, however, that neither nazdiq nor naziq are actually used in the published corpora of spoken texts. 55 Macuch (1993: 381). 56 Macuch (1989: 254). Hezy Muzafi informs me that rāhem (Macuch 1989: 254) is a classicism. 57 Compare Sokoloff (2002: 1225), s.v ‫תפך‬.

NEO-MANDAIC IN MANDAEAN MANUSCRIPT SOURCES

379

ׁ‫אפך‬ ֺ (apḵ),ׁ‫תפך‬ ֺ (tpḵ),ׁ‫גאמתאפך‬ ֺ (gamtapḵ) Lat.: evertere, subvertere, revolvere, Ar.: ‫قلب‬, ‫( يقلب‬Gl. 133:5–6) ׁ‫תפאך‬ ֺ (tpaḵ),ׁ‫גאמתאפך‬ ֺ (gamtapḵ) Lat.: adversari, Ar.: ‫قاوم‬, ‫( يقاوم‬Gl. 134:5–6) ׁ‫אתפך‬ ֺ (atpḵ),ׁ‫גאמתאפך‬ ֺ (gamtapḵ) Lat.: revolvere, Ar.: ‫دخرج‬, ‫( يدخرج‬Gl. 181:14) In spite of this widespread usage in the Glossarium, the root has been entirely superseded in the living Mandaic dialects by a different Mandaic root, k-m-r.58 The verbal root ‫בל‬#‫( ק‬qabl), imperfective ‫בל‬#‫מק‬#‫( ג‬gamqabl) ‘receive’ (Lat.: recipere, Ar.: ‫اقبل‬, ‫يقبل‬, Gl. 7:7–8, Lat.: recipere, Ar.: ‫قبل‬, ‫يقبل‬, Gl. 132:7–8 and Lat.: excipere, Ar.: ‫استقبل‬, ‫يستقبل‬, Gl. 19:11–12) is no longer employed, and survives in NM only in the loan-cognate qabul (< Persian < Arabic) ‘acceptance’.59 Semantically, its place is taken by the roots l-x-ṭ60 and d-r-y.61 #‫פ‬#‫( קל‬qlapa) ‘shell’ (Lat.: cortex, Ar.: ‫قشور‬, Gl. 138:6) has been superseded in NM by miškɔ,62 a lexical impoverishment that loses the original distinction between the softer skin and harder shell. #‫נ‬#‫ת‬#‫חור‬ (huratana) ‘elders’ (Lat.: senes, Ar.: ‫مشيخة‬, Gl. 155:12), and ‫חוראן‬, ‫( חוראתאני‬huran, huratani) ‘elder, elders’ (Lat.: senis, Ar.: ‫شيخ‬, ‫ مشيخة‬Gl. 96:15) are forms that are attested in neither literary nor in spoken Mandaic sources. In contemporary NM, deqen həvar or rīš həvar ‘elder’ is employed in the dialect of Ahvaz (literally ‘white beard’ and ‘white head’),63 attested in the plural as daqqen həvarānā ‘elders’ in the dialect of Khorramshahr.64 It is possible that the increased urbanization and secularization amongst the Iraqi Mandaeans, which brought them into greater contact with Arabic, may have brought about the decline of this type of spoken Mandaic in Iraq.65 It is also feasible that already by the mid-19th century, the Iraqi dialect had been supplanted by the 58

Macuch (1989: 229); Macuch (1993: 406); Häberl (2009: 331). Macuch (1993: 340); Häberl (2009: 345). 60 Macuch (1989: 232); Macuch (1993: 330); Häberl (2009: 334). 61 Macuch (1989: 313–314); Macuch (1993: 382–383); Häberl (2009: 310). 62 Mutzafi (2014: 129–130). 63 Macuch (1993: 385). 64 Häberl (2009: 311). 65 On Iraqi NM and its demise, see Mutzafi and Morgenstern (2012). As Buckley (2010: 6) notes, Petermann did not find Mandaeans in Baghdad in 1854, and M. N. Siouffi, who also came to Baghdad in 1873, did not mention it as a city where Mandaeans lived. By Drower’s time, it was one of the important Mandaean population centres. 59

380

MATTHEW MORGENSTERN

dialects from Iran. Certainly the colophons of Mandaic manuscripts attest to great movement over the whole southern Mesopotamian area during this period.

5. FROM BAGHDAD TO PARIS

Sustained texts written by Mandaeans in vernacular Mandaic—in contrast to late Mandaic literary texts containing strong influences of spoken Mandaic—are extremely rare. A selection of such texts was published in 1904 by J. de Morgan in his volume of Mandaic manuscript finds under the title “Histoires en mandaïte vulgaire”, but was almost entirely ignored by scholars until Macuch presented a study of them in 1989 as part of his Chrestomathie.66 The origin of de Morgan’s texts remained uncertain, and de Morgan did everything that he could to conceal their source, writing dramatically in the introduction to the volume: Je ne dirai pas comment je me suis procuré ces manuscrits, ni quels ont été mes intermédiaires: ce serait mettre en danger l’existence de ceux qui vivent parmi les Sabéens ou qui doivent encore traverser leur pays. Divulguer un livre est considéré chez ces gens comme une trahison méritant la mort, et chaque Sabéen est chargé de l’exécution de cet arrêt.67

As Macuch observed, the vernacular texts appeared to have been composed a short time prior the acquisition of the manuscripts, “etwa am Ende der siebziger oder Anfang der achtziger Jahre des vorigen Jahrhunderts”.68 In particular, Macuch noted that the text mistitled “Historiette racontée en langue mandaïte vulgaire”,69 in reality a letter, mentions ‫ס‬#‫סת‬#‫תריא אנ‬#‫( פ‬patria anastas), adding: “Sollte die Identifizierung eines der im mandäischen Brief … angesprochenen christlichen Priesters ... mit Père Anastase Marie ... stimmen, hätte dieser damals noch ziemlich jung (höchstens 30 Jahre alt) gewesen sein müssen. In diesem Fall wäre aber kaum eine frühere Datierung möglich.”70 Given that Père Anastase Marie de St. Elie (1866– 1947) of the Carmelite Mission in Baghdad was known for his contacts with Man-

66

Macuch (1989: 168–191). De Morgan (1904: VIII). 68 Macuch (1989: 12). 69 De Morgan (1904: 282). 70 Macuch (1989: 12). 67

NEO-MANDAIC IN MANDAEAN MANUSCRIPT SOURCES

381

daeans and his expertise in the Mandaic language,71 Macuch’s identification was quite justified, and was subsequently adopted by Häberl in his own study of the text at hand; as Häberl succinctly stated: “It is unlikely that there were multiple Mandaeophone Père Anastases running around Baghdad at the turn of the century”.72 A recent discovery now removes all doubt regarding the identification. Amongst the Mark Lidzbarski papers held by the Bibliothek der Deutschen Morgenländischen Gesellschaft is a letter from Père Anastase to Lidzbarski written on 20 November 1913. Although the letter deals primarily with questions relating to fishing terminology, which occupied Lidzbarski’s interest during his study of Draša ḏ-Yahia (Johannesbuch),73 it appears from the response that Lidzbarski asked Anastase other questions, including several relating to issues of pronunciation and the existence of other manuscripts. Anastase responded: J’ai reçu des leçons, il y a près de 10 ans d’un professeur mandéen qui me faisait prononcer comme je vous le dis. Ce professeur comprenait un peu sa langue, mais il savait bien le mandaïte vulgaire actuel. Je possède des lettres et des historiettes en ce dialecte que je payais par page (5 francs chacune). Leur langue a fameusement baissé. Tous les manuscrits que possèdent depuis 15 ans les bibliothèques de Londres, Paris, Berlin et autres ont été acquis par moi. Ceux qu’a publiés M. de Morgan ne font pas exception.74

71 72

ure.

73

Macuch (1989: 184–185) noted these contacts to lend support to the identification. Häberl (2010: 551–552), who presents additional information about this interesting fig-

The letter is the first of the two mentioned explicitly in Lidzbarski (1915: 141–142), and cited sporadically through the work. 74 The reference to Berlin is interesting, since the manuscript of “Iniani” published by de Morgan—now Codex Sab. 28 in the Bibliothèque nationale de France—is the twin of another copy found in the Berlin Staatsbibliotek codex Mq 1032. Codex Sab. contains a French description of the manuscript’s contents that seems to have served as a basis for the description of the “Iniani” published by de Morgan, but the writer’s identity and place of origin have been erased. The author might be Anastase, who appended similar descriptions to other manuscripts he supplied to western libraries (e.g. Bodleian Syr. f. 2 and Syr. g. 2.) but comparison of the handwriting with known samples of Anastase’s writing has not led me to any firm conclusion.

382

MATTHEW MORGENSTERN

Again, these manuscripts are exceptional, in that they are written entirely in the vernacular and represent genres—personal letters and folk literature—that are not commonly found in Mandaean libraries. I have found nothing similar in nature in any of the Mandaean collections that I have examined. From Anastase’s response it seems likely that they were written specifically for him. The teacher’s lack of familiarity with the literary idiom would also explain why these texts are so free of classicisms which characterize even the more colloquial writings and speech of learned Mandaeans, and why the spellings are so often corrupt. Both Macuch (1989: 11-12) and Häberl (2011: 560) have emphasized the great similarity between the language of the de Morgan (Anastase) texts and that of the contemporary NM dialects. This resemblance stands in contrast to other sources of NM, which show some variance from the currently spoken dialects.

6. MANDAEAN COLOPHONS

Mandaic literary texts are generally accompanied by a two-part colophon. The first part provides the name of the work, the name of the scribe, and recounts the textual history of the work, i.e. “I, PN son of PN, copied from a manuscript that was copied by PN son of PN, who copied from a manuscript that was copied by etc.”. The personal names are often accompanied by honorific titles. Following the textual history, the second part lists the date and circumstances in which the manuscript was copied. In several manuscripts, this is in turn followed by historical accounts of important events, e.g. persecutions or disasters, and the reconstruction of Mandaean communities following these calamities. Although the copying histories and dating formulae follow a set literary model, they occasionally integrate elements from the contemporary language of the copyists. This is all the more true of the historical accounts, wherein the copyists had to describe current events and society. It is in the colophons that we find some of the strongest influences of contemporary spoken Mandaic. Jorunn Buckley has used the material in the colophons to great effect to try to reconstruct the lost history of the Mandaeans.75 They are equally valuable as a source of NM. In discussing the orthography of Mandaic, Nöldeke already observed the colophons would frequently employ diacritical symbols to represent sounds in words borrowed from Arabic and Persian that did not exist in CM, namely ʿ, ḥ, j and č.76 However, the colophons available to Nöldeke rarely contained sustained NM passages, and the influences such as they are in those texts are limited primarily (though 75 76

Buckley (2010). Nöldeke (1875: 1–3).

NEO-MANDAIC IN MANDAEAN MANUSCRIPT SOURCES

383

not exclusively) to loanwords. It is therefore not entirely surprising that Nöldeke would have regarded these colophons as reflecting a literary idiom written in a classicising Mandaic and incorporating elements of the local spoken languages rather than representatives of a living language. Again, Macuch was the first scholar to explicitly declare materials from the colophons to be written in NM rather than simply containing late elements. In his review of Drower’s edition of DC 35, The Baptism of Hibil Ziwa, Macuch remarked “Der Schluß der Handschrift ist höchst interessant. Er wurde im Neumandäischen der zweiten Hälfte des vorigen Jahrhunderts vom Kopisten, welcher Petermanns mandäischer Lehrer war... zugefügt”.77 On the basis of his first-hand knowledge of NM, Macuch was able to improve the readings and interpretations of several passages, though it must be stated that Drower had done remarkably well in translating the colophon.78 As we have seen, there is now clear evidence that Drower had first-hand acquaintance with the raṭna, both the Mandaising Arabic speech of the Iraqi Mandaeans and the NM speech of Mandaeans of Iranian origins.79 Nevertheless, certain structures escaped her. For example, in two places I have found that Drower did not correctly interpret the possessive structure noun+d+pronoun, which is one of the salient features that distinguish NM from CM. In the same late historical colophon from DC 35 that Macuch mentioned, which relates an event following the cholera epidemic of 1247 AH (1831–1832 CE) up to 1265 AH (1848–1849 CE), we read #‫ב‬#‫דיון עור‬#‫ן עב‬#‫תד‬#‫ר‬#‫ר װ‬#‫שושת‬-$ #‫בעוחר‬ ‫י‬#‫ת‬#‫כס‬#‫י ות‬#‫( לױימיד מינד‬bʿuhra ḏ-šuštar g̤aratdan ʿbadiun ʿuraba lh̤ imid mindai utaksatai). Drower translated: “On the Šuštar road, Arabs, for greed, surrounded us and were covetous of my things and my clothes”, and wrote of the word ‫ן‬#‫תד‬#‫ר‬#‫װ‬ (g̤aratdan) “P[ersian] ‫‘ ڭردان‬a circuit’, ‘a turn’; hence here ‘made a circuit about us’, 77

Macuch (1955: 362). It was this review that drew Drower’s attention to Macuch and his work and ultimately led to their collaboration on the Mandaic Dictionary. See Macuch (2008: 13). Macuch (1989: 4–5) also called attention to the significance of the colophons as a source of Neo-Mandaic. 78 Contra Macuch (1989: 5), who wrote of Drower’s attempts “daß sie trotz ihres langjährigen Verkehrs mit den Mandäern im Irak, bei denen sich schon seit längerem das bei den Mandäern der südiranischen Provinz Chūzistān noch am Leben erhaltene neumandäische Idiom im Absterben befand, die größten Schwierigkeiten hatte und selten den richtigen Sinn traf”. 79 On the distinction between these two see Mutzafi and Morgenstern (2012: 160–162).

384

MATTHEW MORGENSTERN

‘surrounded us’. The sentence is obscurely phrased, but the meaning is obvious from the context”.80 However, the correct interpretation of the phrase would appear to be ġārat-dan əwadyon ‘(they) raided us’ wherein ġārat-dan is derived from Arabic/Persian ‫‘ غارة‬making a hostile incursion into an enemy’s country; a raid; plunder, pillage; havoc, devastation’81 construed as a phrasal verb on the model of Persian ġārat kardan. Similarly, in the colophon of BL. Or 6592, copied in Muḥammara in 1289 AH (1872–1873 CE) we read: #‫ כדיב חו‬#‫דת‬#‫כירדה ח‬#‫לדה ו‬#‫ שארױ או‬#‫אנסית מן חד‬$ #‫חו כתיב חו‬#‫וה במ‬#‫ כידבית על‬#‫ אנ‬#‫חו כדיב חו‬#‫ מ‬#‫ת‬#‫לה אס‬#‫ת‬#‫חו‬#‫( ל‬ḏ-ansit mn hda šarh̤ aualdẖ uakirdẖ hadta kdib hua lahuatalẖ asata mahu kdib hua ana kidbit ʿlauẖ bmahu ktib hua). Drower translated “I copied from a šaḥr (!) belonging to his father (?) (walidẖ) and its latter part was newly written. He had no copies. Whatever was written on it (in it) I wrote down, I wrote whatsoever was written”.82 It seems better to translate: ‘that I copied from a šarḥ, the beginning of which and end of which had been written anew. It didn’t have copying traditions. 83 Whatever was written I wrote as it was written’.‫כירדה‬#‫לדה ו‬#‫( או‬aualdẖ uakirdẖ) are NM forms awwaldi ‘its beginning’ and āxirdi ‘its end’. As Häberl has pointed out regarding the NM of Khorramshahr, “foreign nouns will only take possessive suffixes via an intermediary morpheme, -d-”.84 A full study of the language of these colophons has yet to be carried out.85 Nonetheless, even a cursory glance at MD reveals that several lexemes that are characteristic of NM are found only in the colophons. While some have survived into contemporary NM, e.g. the verbal root ‫( בגץ‬b-g-ṣ) ‘stay, wait’,86 others are appar-

80

Drower (1953: 88). Steingass (1892: 877). 82 Drower (1962: 36). 83 For this use of #‫ת‬#‫( אס‬asata) compare ‫ לה‬#‫חו‬#‫ יאתיר ל‬#‫ת‬#‫ס‬#‫( ו‬uasata iatir lahualẖ) ‘and it had no more copying traditions’ (DC 36:825). 84 Häberl (2009: 33). An exception to this rule is found twice in Macuch (1993: 212): merīšdẖ “von seinem Anfang” (l. 1181); ʾāx rīšdẖ “sein anderes Ende” (l. 1182). 85 I am currently preparing such a study for publication. 86 Drower and Macuch (1963: 52); it is already found in the Glossarium. 81

NEO-MANDAIC IN MANDAEAN MANUSCRIPT SOURCES

385

ently not employed in the surviving dialects, e.g. ‫( מוד‬mud)87 with its variant ‫אמוד‬ (amud) ‘according to, whatever’. 88 A similar word that escaped the attention of Drower and Macuch is ‫( מול‬mul), which from the context means ‘that (which)’. I have found it in two contexts, one of which (DC 27) was known to the authors of MD: ‫י חזאת מן גינזא אנסית‬#‫( וכול מול אינ‬ukul mul ainai hzat mn ginza ansit) ‘and all that my eye saw from the library89 I copied’ (DC 27:558, copied 1088 AH [16771678 CE]);90 and ‫ כול מול‬#‫שאמ‬#‫יכון ל‬#‫ן חזאת עודונ‬#‫זיא מן כול אינ‬#‫ ח‬#‫יכון ל‬#‫אינ‬ #‫ת מן בניא חשוכ‬#‫ן שימ‬#‫( עודנ‬ainaikun la hazia mn kul ainan hzat ʿudunaikun lašama kul mul ʿudnan šimat mn bnia hšuka) ‘may your eye not see any of that which our eye has seen and may your ear not hear all that which our ear has heard from the sons of darkness’ (RRC 2M, also copied 1086 AH [1675–1676 CE]). ‫ן‬#‫( אכתינ‬aktinan) ‘we are’ (DC 51:834), found in the colophon of an amulet formulary written in 1277 AH (1860-1861 AH), was correctly identified in MD as a NM form.91 Another example occurs in the same colophon: ‫( עכתינו‬ʿktinu) ‘they are’ (DC 51:816).92 It is worth noting that in both cases, the forms of the copula in DC 51 differ from those of the contemporary NM dialects, which are respectively extan ‘we are’ and extu ‘they are’.93 But while ‫ן‬#‫( אכתינ‬aktinan) ‘we are’ was accurately identified in MD, the wider context has not been correctly analysed. The full context (DC 51:835–836) reads: #‫ וגוצ‬#‫מ‬#‫ וג‬#‫מ‬#‫ן גוד ח‬#‫ אכתינ‬#‫מ‬#‫( וליליא ועומ‬ulilia uʿumama aktinan gud hama ugama uguṣa), and should be translated ‘and day and night we were in trouble and distress and worry’. MD emends ‫( גוד‬gud) to the graphically similar ‫( גוב‬gub) and interprets this as a phonetic spelling for guw ‘in’;94 however, since ‫( גוד‬gud) is now attested numerous times in several colophons in the Rbai

87

Ibid. (260). Ibid. (22). Perhaps a reflex of this lexeme is found in NM hemmed ‘whatever’. See Macuch (1989: 218; 1993: 387). 88

89

For this meaning of ‫( גינזא‬ginza) see Drower and Macuch (1963: 90). For a different interpretation see Burtea (2008: 118–189). 91 Drower and Macuch (1963: 18). 92 Drower and Macuch (1963: 349) s.v. ʿkt-. 93 Macuch (1965: 380); Macuch (1993: 92); Häberl (2009: 231). 94 Drower and Macuch (1963: 82). 90

386

MATTHEW MORGENSTERN

Rafid Collection95 there is clearly no need for emendation. Like ‫( מוד‬mud), ‫גוד‬ (gud) may be explained as a compound of gu and d-. #‫מ‬#‫( ח‬hama) is erroneously derived in MD from Arabic ‫ َحم‬,96 but ‫ هم ‏وغم‬is a fairly common word-pair in Arabic, and the NM Mandaic transcription should be interpreted accordingly. As with the Glossarium, the colophons employ loanwords which are now no longer in use, e.g. #‫( תוף וגוליל‬tup ugulila), apparently ‘cannon(s) and ball(s)’, from Turkish top ve gülle. The words are used to describe the weapons employed by the Turkish army against the Arab militias of the Muntafiq tribe (DC 43I:181), and thus represent the borrowing of terms for contemporary military technology which are now obsolete.

7. SFAR MALWAŠIA (THE BOOK OF THE ZODIAC)

Above we saw that this work, which comprises of assorted prognostications, was already known to Nöldeke,97 who expressed a negative opinion regarding its language and style, and was published by Drower, who regarded it as reflecting a transitional period between CM and NM.98 All copies of this book known today may be traced back to the work of a single scribe, Yahia Ram Zihrun br Mhatam,99 from whose work Drower’s manuscript was directly copied in Qurna in 1247 AH (18311832 CE). The work is clearly divided into several original collections, the first two of which have their own colophons (DC 31. 1–106; 106–256). The second colophon ultimately extends back to one Muʿalia br Anuš Bihdad, whom Buckley has tentatively placed around 1100 CE.100 The first colophon extends back a further two generations, but neither of the scribes mentioned is known from other texts. The contents of these sources may be in parts much older, and evidence has been adduced for their reliance upon Babylonian (Akkadian) models101 and for their similarity to

95

So far I have identified the form in RRC 1C, 2M, 2O and 4G. Drower and Macuch (1963: 122). 97 Nöldeke made use of MS Berlin Cod. Petermann 155, copied in Suq eš-Šuiūḵ in 1270 AH (1853–1854 CE). 98 Drower (1949). 99 This scribe was also active in Qurna, where he copied Oxf. Ms. Syr. G 2 (R) in 1231 AH (1816 CE) and Oxf. Ms. Asiat. Syr. C 13 in 1233 AH (1817–1818 CE); see Buckley (2010: 271). He later copied RRC 3F in 1238 AH (1822–1823 CE). 100 Buckley (2010: 241). 101 Rochberg (1999–2000). 96

NEO-MANDAIC IN MANDAEAN MANUSCRIPT SOURCES

387

Jewish materials found in the Cairo Geniza.102 The third collection (DC 31. 257– 289), which lacks a textual history, is quite clearly of much later origin, and is replete with Arabic loanwords and influences of NM. Let us consider a single passage: ‫לין‬#‫ מן ח‬#‫ניא ודר‬#‫כירמ‬$ ‫יניא איתיא תותיא‬#$ ‫בדית תותיא‬#$ ‫ישֵיא‬ ֵ #‫ו‬#ֵ‫לין ֵח‬#‫ח‬ ‫ך‬#‫רס‬#‫ ם ב‬#‫ריר ולילי‬#ֵֵ‫!ימ ונוחלה בח‬#‫ לתותיא ושוגה נ‬#‫תנ‬#‫ניא ו‬#‫רמ‬#‫ד‬ ‫יניא‬#-$ ‫ויא זיחוא‬#‫ר ויאק‬#‫ך ומינח‬#‫ינ‬#‫שאכבית מליא ב‬ (halin hauaiš̤ia ḏ-abdit tutia ḏ-ainia aitia tutia ḏ-kirmania udra mn halin darmania uatna ltutia ušugẖ naʕim unuhlẖ bh̤arir ulilia kḏ barsak šakbit mlia bainak uminhar uiaqauia zihua ḏ-ainia) These are the medicines that you may make a collyrium for the eyes; bring the collyrium (stone) of Kirman, and take of these medicines and set down the collyrium (stone) and mill it fine, and sift it with silk, and at night, when you lie in your bed, fill it in your eye and your eyesight will be enlightened and grow strong (DC 31. 286:38–41).

Most of the language here is late: ‫יײיא‬#‫ו‬#‫( ױ‬h̤ auaiš̤ia) is an Arabic loanword that appears here in its NM meaning of ‘medicine’;103 ‫איניא‬$ ‫( תותיא‬tutia ḏ-ainia) is a calque of Persian tūtiyāʾi dīda ‘A collyrium or medicine for the eyes’;104 ‫ושוגה‬ (ušugẖ) appears in the NM meaning of ‘to rub a special type of stone against a millstone so as to extract a substance used as eye medicine’;105‫!ימ‬#‫( נ‬naʕim) is an Arabic loanword as the pharyngeal ʕ indicates; similarly ‫ריר‬#‫( ױ‬h̤ arir), as indicated by the diacritical pointing that marks the pharyngeal ḥ; ‫ויא‬#‫( ויאק‬uiaqauia) appears to be based upon Arabic root q-w-y ‘be strong’, and employs the y- verbal prefix of Arabic; and ‫( זיחוא‬zihua)106 corresponds to NM zehwɔ ‘light’; while NM zehəw al-inɔ means ‘eyesight’. 107 ‫ניא‬#‫רמ‬#‫( ד‬darmania), though a loanword from Persian, is already found in more ancient Mandaic sources.108

102

Bohak and Geller (2013). Mutzafi (2014: 8). 104 Steingass (1892: 333). 105 Mutzafi (2014: 97–98). 106 In DC 31 the ‫( ח‬h) is written above the line as scribal correction, but the reading is supported by the other textual witnesses. 107 Mutzafi (2014: 174–176). 108 Pognon (1898: 123, text 4:1). 103

388

MATTHEW MORGENSTERN

What is significant about this passage is that within the space of a few lines, we find such a high concentration of modern vocables. True, we still find classicisms here, e.g. ‫לין‬#‫( ‏ח‬halin) ‘these’ for NM ahni(n) or hanni,109 and ‫( ם‬kḏ) ‘when’ for NM ke, or waxti ke,110 but that fact that so many of the late usages are shared with contemporary NM shows that the thin façade of the literary conservatism cannot conceal the spoken NM of the scribe who formulated this passage.

8. GRIMOIRES AND AMULETS

Beyond the religious works, a large proportion of Mandaic literature comprises amuletic materials, mostly for healing and protection. The style and content of many of the amulet formulae preserved in late manuscripts accord with those found in the much earlier epigraphic materials, demonstrating the great antiquity of the manuscripts’ formulae.111 In recent times, many of the shorter formulae were collected into grimoires (magical ‘recipe books’) which provided professional amulet writers with their raw materials.112 Numerous Mandaic amulets written in modern times are found in western libraries and private collections, of which some are actual objects written for real clients, while others are formularies which could serve as a model for the copying of amulets. Four substantial grimoires are known to me, all from western collections: DC 45, DC 46, CS 24 and CS 27. None of these are dated, but all appear to have been copied in the late 19th or early 20th centuries. Some additional fragments are found in libraries and private collections.113 Several lexemes from CS 27 entered MD from Lidzbarski’s card index of words, e.g. ‫יזא‬#‫( נ‬naiza) ‘[term] used of a reed’,114 while selections from DC 45 and DC 46 were published by Drower115 and most of their vocabulary was also recorded in MD.

109

516).

110

Macuch (1989: 52; 1993: 56); Häberl (2009: 162); and see Morgenstern (2010: 512–

Macuch (1989: 95); Häberl (2009: 332; 361). For previous discussions and literature, see Ford (2002b: 44–45) and Müller-Kessler (2010: 453–454). 112 A description of the structure of the grimoires appears in Morgenstern and Alfia (2013), but at the time of writing, only some of the manuscripts were known to us. 113 A full list of these manuscripts is found in Morgenstern (in preparation). 114 Drower and Macuch (1963: 283). Lidzbarski’s acquaintance with CD 27 is evident from his card index which is now held in the library of the Deutsche Morgenländische Gesellschaft. I have found no evidence that suggests that Drower herself saw either of the manuscripts from Paris. 111

NEO-MANDAIC IN MANDAEAN MANUSCRIPT SOURCES

389

In many modern manuscript amulets or formularies, and to an even greater extent in these grimoires, the protective formulae were furnished with instructions for the preparation of the amulet.116 Since these instructions had to be comprehensible to the modern copyist, and related to the realia of contemporary life (e.g. materials on which the amulet is to be written, the type of ink to be used, the location in which it is to be placed etc.), these instructions tend to be extremely rich in NM lexemes.117 In addition to the Mandaic formulae, a large number of formulae in the grimoires and occasionally in the amulet scrolls are in Arabic, usually in Mandaic transcription and often corrupt. The transcription shows clear signs of the Muslim gelet type dialects.118 Nonetheless, even when the formulae are in Arabic, the instructions for their use are generally presented in the same type of late literary Mandaic replete with NM influences that characterizes the other rubrics. Furthermore, not all of the Mandaic formulae are of the same ancient date. For example, a lengthy series of amulet formulae presented in narrative form, each of which begins #‫יו‬#‫ד‬$ #‫ב‬#‫( ב‬baba ḏ-daiua) ‘a spell for (lit. of) a dev’, is unmistakably of late origin and contains many NM lexemes.119 It seems that Drower was correct that the series, which is preserved in DC 46 and CS 24, was based upon an Arabic model.120 NM evidence was applied to good effect to interpret the late language of the baba ḏ-daiua texts, and to the examples already identified in MD we may now add one more. Mutzafi has recently drawn attention to the NM use of the root r-d-f in the meaning of ‘to shiver, tremble’.121 This usage of r-d-f may be used to interpret one of the baba ḏ-daiua texts, in which the demon states #‫ת‬#‫עוחר‬$ ‫י לרישא‬#‫דוכת‬ #‫דיפנ‬#‫שא מר‬#‫בר אנ‬$ ‫ויא ועדיא וליגריא‬#‫( ח‬duktai lriša d-ʿuhrata hauia uʿdia uligria

115

Drower (1943). Morgenstern and Alfia (2013: 156–157). 117 For examples see Mutzafi and Morgenstern (2012: 162–163). 118 Ibid. (163–164). 119 Morgenstern and Alfia (2013: 156). 120 Drower (1943: 155–1556); she was apparently unaware that a second copy existed in Paris. 121 Mutzafi (2014: 24–25). 116

390

MATTHEW MORGENSTERN

ḏ-br anaša mradipna) ‘my haunt is at the junction (?) and I cause the hands and legs of a man to tremble’ (DC 46. 63: 11–12).122 Not all of the late linguistic material has been presented accurately in MD or explicitly marked as post-classical. For example, Mutzafi has noted the use of derkɔ, derk in meaning ‘bank (of river, stream, canal)’ as characteristic of NM.123 The same definition is recorded in MD with a citation but no reference.124 Examination of the sources reveals that the citation, #‫חר‬#‫נ‬$ ‫( תרין דירכיא‬trin dirkia ḏ-nahra), is drawn from the instructions for writing on an amulet found in two parallel copies in DC 45. 70:16 and DC 46. 210:1. The phrase #‫חר‬#‫נ‬$ #‫( דירכ‬dirka ḏ-nahra) appears elsewhere in these manuscripts meaning ‘bank of the river’, but always within the context of such instructions. I have not found it elsewhere in Mandaic literature. Here we have a clear example wherein these instructions employ a NM usage which is not otherwise attested in Mandaic literature, but which is not unambiguously marked as a late usage in MD. MD’s somewhat haphazard system of marking post-classical lexemes and definitions has proven an additional hindrance to the diachronic study of Mandaic.125 Regarding loanwords in these texts, there are several indications these are drawn from the spoken language. For example, ‫ר‬#‫( ביסמ‬bismar) ‘nail’ is not simply a ‘coll[oquial] corruption of Ar[abic] ‫ ’ ِمسْمار‬as MD suggests,126 but rather reflects the spoken Arabic of Baghdad.127 The number of loanwords in the baba ḏ-daiua texts and the rubrics is considerable.

9. LATE MANDAIC POETRY

In his discussion of Mandaic poetry, Lidzbarski noted that while most of the wedding ritual is of a religious nature, some “volkstümliche Liedchen” are added, but notes “Bei ihrem volkstümlichen Charakter, der vom Inhalt des sonstigen man-

122

Reading and interpretation contra Drower and Macuch (1963: 425) s.v. RDP Part. pres. Mutzafi (2014: 101–102). 124 Drower and Macuch (1961: 109) s.v. dirka. 125 Morgenstern and Mutzafi (2012: 162–164); Morgenstern (2012: 184). 126 Drower and Macuch (1963: 62). 127 Woodhead and Beene (1967: 35). 123

NEO-MANDAIC IN MANDAEAN MANUSCRIPT SOURCES

391

däischen Schrifttums stark abweicht, ist das Verständnis sehr schwierig”.128 Although Lidzbarski seems to have regarded these poems of being early, he also remarked as an aside “Natürlich können im Einzelnen die Stücke jung sein und sind es auch, wie die arabischen Wörter zeigen”.129 I believe that Macuch was correct to regard that poems as evidence for early NM, and regard his assessment that they “bilden eine Zwischenstufe zwischen einer älteren und der heutigen Volkssprache” to be more accurate than Lidzbarski’s assessment.130 The earliest evidence for these wedding songs is found in CS 15, which was copied in Basra in 1086 CE (=1675–1676 CE).131 The texts were published by Drower, without reference to the older Paris manuscript, primarily on the basis of DC 38, which was copied in 1216 AH (1801–1802 CE).132 It is clear from the translation that Drower did not always understand their late language. Some of these errors were corrected in MD, though the correct interpretations are not always immediately apparent since they were not recorded under their lexical entry. For example, in one song we read #‫אנ‬$ ‫ן‬#‫ בורכ‬#‫ח‬#‫ת אל‬#‫ אנ‬#‫שט‬#‫י פ‬#‫ר‬#‫ למ‬#‫ ועד‬#‫ימ‬#‫ ק‬#‫אתותיא דיקל‬ ‫ לביחומ‬#‫נ‬#‫ל‬#‫( ק‬atutia diqla qaima uʿda lmarai pašṭa anat alaha burkan ḏ-ana qalana lbihum),133 which Drower translated: ‘Standing beneath the date-palm, And outstretching (my) hand to my lord, (I pray), “Thou, God bless me, So that I may speak with the stranger!”’.134 In MD, s.v. bihum, the latter part has been corrected to ‘thou, God, bless me that I may go to the stranger (?)’,135 wherein #‫נ‬#‫ל‬#‫( ק‬qalana) is cor-

128

Lidzbarski (1920: IX). Ibid.: XI. 130 Macuch (1989: 16). 131 Macuch (1989: 4) erroneously reported that the manuscript is CS 11 (this was the old numbering of the Paris manuscripts, still employed by Nöldeke) and that it was copied in 1529/30 (this is the date of Oxf. Marsh. 691, which does not contain these songs). 132 Drower (1950). 133 Published in Drower (1950: 19). Drower’s edition mistakenly omits part of the text, and reads: Atutia diqla anat alaha burkan ḏ ana qalana lbihum, thought the missing words are represented into the translation. 134 Ibid.: 69. 135 Drower and Macuch (1963: 59). In light of Mutzafi’s discovery of 1 f.s. participles in the pattern of našqɔ́ as well as more conservative našqɔnɔ we may translate the first part of this verse ‘I stand beneath the date-palm, and stretch out (my) hand to my lord’. The personal pronoun of ‫י‬#‫ר‬#‫( מ‬marai) ‘my lord’ requires such an interpretation. See Mutzafi (forthcoming). 129

392

MATTHEW MORGENSTERN

rectly identified as NM qalɔnɔ ‘I go’.136 Similarly, ‫חיך‬#‫זיא בעוחרה וג‬#‫( ט‬ṭazia137 bʿuhrẖ ugahik),138 which Drower translated ‘That those who mourn in the roads will laugh’ has been corrected in MD s.v. ṭ- to ‘who goes’.139 Nevertheless, some modern forms have been overlooked. In the continuation of the previous text, we read in DC 38: #‫ר‬#‫גול חיו‬-$ #‫ונ‬#‫ מיל‬#‫ שאדויא סקיל‬#‫נ‬#‫ב‬#‫( ק‬qabana šaduia sqila milauna ḏ-gul hiuara), which Drower translated ‘They clad him in a robe, an elegant one, Coloured rose and white’,140 deriving #‫נ‬#‫ב‬#‫( ק‬qabana) from Arabic ‫قباء‬.141 It seems, however, that #‫נ‬#‫ב‬#‫( ק‬qabana) is to be interpreted as ‘I (f.sg.) desire’, and that #‫נ‬#‫ב‬#‫( ל‬labana) in the same text means ‘I (f.sg.) don’t desire’.142 The remainder of the poem remains difficult to interpret, particularly since the manuscripts contain differences at several crucial points. Aside from the wedding songs, I know of only one other example of late Mandaic poetry. It is a polemical poem written against the consumption of meat that has not been ritually slaughtered. The poem appears at the end of RRC 4G, a copy of Draša ḏ-Yahia (the Mandaean Book of John) copied in Qurna in 1248 AH (1832– 1833 CE) by Yahia Bihram br Adam Yuhana, and appears to be his own composition.143 Although it is mostly formulated in a classicizing style, the author employs post-classical forms such as the demonstrative pronoun ‫ך‬#‫( ת‬tak) ‘that’144 for Classi-

136

32).

137

For the etymology of this form see Morgenstern (2010: 519–523) and Mutzafi (2014:

CS 15. 20b:2 reads ‫אזיא‬$ (ḏ-azia). On interchanges of ḏ and ṭ see now Ford (2012). Drower (1950: 18). 139 Drower and Macuch (1963: 171). See also Macuch (1965: 11) (where, however, ṭazia is erroneously identified with ḏ-masgia) and Macuch (1989: 1). 140 Drower (1950: 66). 141 Drower and Macuch (1963: 398 s.v.). 142 Hezy Mutzafi (personal communication) confirms that in the Neo-Aramaic dialect of Ahvaz, the form qəbɔnɔ is employed in this category. 143 On Yahia Bihram’s life and works see Buckley (2010: 133–147). 144 This form of the demonstrative pronoun is not recorded in MD or Macuch’s supplements (Macuch 1965: 527–543; Macuch 1976: 1–146) but is the far-deixis form of #‫( ת‬ta), for which see Drower and Macuch 1963: 477 s.v. It is already found in Drower’s collection in the expression #‫נ‬#‫ך עד‬#‫( בת‬btak ʿdana) ‘at that time’ DC 42: 828. The earliest example of ‫ך‬#‫ת‬ (tak) ‘that’ currently known to me is in the NM colophon of RRC 1C, copied in 1074 AH (1663–1664 CE). 138

NEO-MANDAIC IN MANDAEAN MANUSCRIPT SOURCES

393

cal Mandaic ‫תה‬#‫נ‬#‫( ח‬hanatẖ) or NM ax or axu.145 The text also employs occasional loanwords, e.g. #‫( ײאר‬š̤ara) ‘cure’ from Persian čāre.146 Though not a rich source of NM, the poem nonetheless demonstrates that even classicizing writings were not free of NM influences. As with the grimoires and magic texts, it seems that during the later stages of editorial redaction, some parts of the classical liturgy were provided with rubrics that show influences of post-classical Mandaic, e.g. ‫( חידוכתיא‬hiduktia) ‘wedding days’.147 These late usages in the rubrics do not necessarily indicate that the liturgical poems that they accompany are late.

10. SHEIKH NEJM’S GLOSSARY

This text has been discussed in detail in a recent article, wherein it was shown to contain several NM lexemes.148 Although not a rich source of material, it indicates the degree to which NM remained familiar to Iraqi Mandaeans in the 20 th century and was prone to affect their writings.

11. CONCLUSION

All surviving Mandaic manuscripts were copied from the 16th century onwards, i.e. some 900 years later than the epigraphic texts (the lead scrolls and the inscribed clay bowls). During this period, the earlier parts of the Classical Mandaic corpus were redacted, the orthography was to an extent standardized, and the later texts were probably composed. By the time the earliest surviving manuscripts were copied, a very different type of Mandaic was apparently being spoken, and this is reflected in the manuscripts in the ways outlined above. Evidence for this spoken language is unambiguously forthcoming from the 17th century onwards, particularly in the Leiden Glossarium and in some colophons. Fewer sources may be dated with certainty to the 18th century, but a considerable body of material is available from the 19th century. All of these sources demonstrate that the NM dialects spoken today 145

Macuch (1989: 52); Macuch (1993: 56); compare Morgenstern (2010: 515). Compare Häberl (2009: 309). In light of the diacritical marks in RRC 4G and the pronunciation in NM, Häberl (2010: 556-557) was correct to read ‫י‬#‫( שאר‬šarai) in the NM letter published by De Morgan as čāre (contra Macuch 1989: 184, 257, who read šāre). 147 Lidzbarski (1920: 239: 7, 245: 1; 250: 2). See already Nöldeke (1875: 78). 148 Mutzafi and Morgenstern (2012). 146

394

MATTHEW MORGENSTERN

represent only a small amount of the Mandaic linguistic riches and varieties that existed when NM was spoken by much larger circles of Mandaeans over a wider geographical area. The devastating effects of the cholera epidemic, persecutions and political upheavals that befell the Mandaean community in the 19 th century, and which are described so vividly (in NM) in the historical accounts appended to the colophons, greatly reduced the scope of NM use. Today, the last remaining vestiges of this language survive on a communal level only in Ahvaz. A preliminary survey of written evidence for NM prior to Macuch’s fieldwork indicates that they may serve as a valuable source for uncovering these last chapters in the lost history of the Mandaic language. It also makes clear that, following the title of Macuch’s seminal work of 1965, MD must be understood as a Dictionary of Classical and Modern Mandaic.

REFERENCES Manuscript abbreviations CS—Codex Sabéen, Bibliothèque nationale de France, Paris. DC—Drower Collection, Bodleian Library, Oxford. RRC—Rbai Rafid Collection.

Bibliography

Bohak, Gideon, and Mark Geller. 2013. “Babylonian Astrology in the Cairo Genizah.” In Envisioning Judaism: Studies in Honor of Peter Schafer on the Occasion of his Seventieth Birthday, edited by Raʿanan S. Boustan, Klaus Herrmann, Reimund Leicht, Annette Y. Reed and Giuseppe Veltri, 607–622. Tübingen: Mohr Siebeck. Borghero, Roberta. 1999–2000. “A 17th Century Glossary of Mandaic.” ARAM 11: 311–319. —. 2004. “Some Linguistic Features of a Mandaean Manuscript from the Seventeenth Century.” ARAM 16: 61–83. Buckley, Jorunn J. 2010. The Great Stem of Souls. Piscataway, NJ: Gorgias Press. Burtea, Bogdan. 2008. „Zihrun, das verborgene Geheimnis“: Eine mandäische priesterliche Rolle. Edition, Übersetzung und Kommentierung der Handschrift DC 27 Zihrun Raza Kasia. Wiesbaden: Harrassowitz. Morgan, Jacques de. 1904. Mission scientifique en Perse, tome V (études linguistiques), deuxième partie: Textes mandaïtes, histoires en Mandaïte vulgaire. Paris: Imprimerie Nationale. Drower, Ethel S. 1943. “A Mandæan Book of Black Magic.” Journal of the Royal Asiatic Society 1943: 149–181.

NEO-MANDAIC IN MANDAEAN MANUSCRIPT SOURCES

395

—. 1949. The Book of the Zodiac = Sfar malwašia: D. C. 31. London: Royal Asiatic Society. Drower, Ethel S. 1950. Šarh d-qabin d-Šišlam Rba (D.C. 38): Explanatory Commentary on the Marriage-ceremony of the Great Šišlam. Roma: Pontificio Istituto Biblico. —. 1953. The Haran Gawaita and the Baptism of Hibil-Ziwa. Vatican City: Biblioteca Apostolica Vaticana. —. 1960. Alf Trisar Šuialia, The Thousand and Twelve Questions. Berlin: AkademieVerlag. —. 1962. The Coronation of the Great Šišlam, being a Description of the Rite of the Coronation of a Mandaean Priest according to the Ancient Canon. Translation from Two Tanuscripts Entitled “The Coronation of Šišlam-rba”, DC 54 Bodleian Library, Oxford (1008 A.H.) and Or. 6592, British Museum (1298 A.H.) with Discussion of the “Words Written in the Dust.” Leiden: Brill. Drower Ethel S., and Rudolf Macuch. 1963. A Mandaic Dictionary. Oxford: The Clarendon Press. Ford, James N., and Matthew Morgenstern. 2002a. “Review of J.B. Segal, Catalogue of the Aramaic and Mandaic Incantation Bowls in the British Museum.” Jerusalem Studies in Arabic and Islam 26: 237–272. —. 2002b. “Another Look at the Mandaic Incantation Bowl BM 91715.” JANES 29: 31–47. —. 2012. “Phonetic Spelling of the Subordinating Particle d(y) in the Jewish Babylonian Aramaic Bowls.” Aramaic Studies 10: 215–247. —. in preparation. New Mandaic Incantation Bowls. Häberl, Charles. 2009. The Neo-Mandaic Dialect of Khorramshahr. Wiesbaden: Harrassowitz. —. 2010. “Neo-Mandaic in Fin de siècle Baghdad.” Journal of the American Oriental Society 130 (4): 551–560. Hunter, Erica. 1995. “Combat and Conflict in Incantation Bowls: Studies on Two Aramaic Specimens from Nippur.” In Studia Aramaica: New Sources and New Approaches. Journal of Semitic Studies Supplement Series 4, edited by Markham J. Geller, Jonas C. Greenfield and Michael P. Weitzman, 61–76. Oxford: University Press. Kim, Ronald. 2011. “Review of Häberl 2009.” Journal of the American Oriental Society 131: 323–327. Lidzbarski, Mark. 1915. Das Johannesbuch der Mandäer. Giessen: Töpelmann. —. 1920. Mandäische Liturgien, mitgeteilt, übersetzt und erklärt. Berlin: Weidmann. Macuch, Maria. 2008. “‘And Life is Victorious!’ Mandaean and Samaritan Literature—In Memory of Rudolf Macuch (1919–1993).” In Und das Leben ist siegreich, And Life is Victorious, edited by Rainer Voigt, 9–16. Wiesbaden: Harrasowitz.

396

MATTHEW MORGENSTERN

—. “Review of Drower 1953.” Zeitschrift der Deutschen Morgenländischen Gesellschaft 105: 357–363. —. 1965. Handbook of Classical and Modern Mandaic. Berlin: De Gruyter. —. 1976. Zur Sprache und Literatur der Mandäer. Berlin: De Gruyter. —. 1989. with Klaus Boekels. Neumandäische Chrestomathie mit grammatischer Skizze, kommentierter Übersetzung und Glossar. Wiesbaden: Harrassowitz. —. 1993. Neumandäische Texte im Dialekt von Ahwaz. Wiesbaden: Harrassowitz. Morgenstern, Matthew. 2010. “Diachronic Studies in Mandaic.” Orientalia 79: 505– 525. —. 2012. “Review of Häberl 2009.” Journal of Semitic Studies 57: 182–184. —. forthcoming. “Rare Forms in Eastern Aramaic.” Language Studies 16. [in Hebrew] —. in preparation. “Mandaic Literature: A Guide to Manuscript Sources.” Morgenstern, Matthew, and Tom Alfia. 2013. “Arabic Magic Texts in Mandaic Script: A Forgotten Chapter in Near-Eastern Magic.” In Durch Dein Wort ward jegliches Ding! / Through Thy Word All Things Were Made! – II Mandäistische und Samaritanistische Tagung , 163–179. Wiesbaden: Harrassowitz. Müller-Kessler, Christa. 2001–2002. “Die Zauberschalensammlung des British Museum.” Archiv für Orientforschung 48/49: 115–145. —. 2010. “A Mandaic Incantation against an Anonymous Dew Causing Fright: Drower Collection 20 and its Variant DC 43 E.” ARAM 22: 453–476. Mutzafi, Hezy. 2014. Comparative Lexical Studies in Neo-Mandaic. Leiden: Brill. —. forthcoming. “Verbal Conjugations in the Neo-Mandaic Dialect of Ahvaz.” Nöldeke, Theodor. 1875. Mandäische Grammatik. Halle an der Salle: Waisenhaus. Pognon, H. 1898. Inscriptions mandaïtes des coupes de Khouabir. Paris: Imprimerie Nationale. Rochberg, Francesca. 1999–2000. “The Babylonian Origins of the Mandaean Book of the Zodiac.” ARAM Periodical 11–12: 237–247. Reprinted in Rochberg, Francesca. In the Path of the Moon: Babylonian Celestial Divination and its Legacy. Ancient Magic and Divination 6, 223–235. Leiden: Brill, 2010. Sokoloff, Michael. 2002. A Dictionary of Jewish Babylonian Aramaic of the Talmudic and Geonic Periods. Ramat Gan: Bar Ilan University Press, and Baltimore and London: The John Hopkins University Press. Steingass, Francis J. 1892. A Comprehensive Persian-English Dictionary, 2 vols. London: Routledge & K. Paul. Woodhead, Daniel R., and Wayne Beene. 1967. A Dictionary of Iraqi Arabic: ArabicEnglish. Washington, D.C.: Georgetown University Press. Zwemer, Samuel M. 1900. Arabia: The Cradle of Islam. New York: Fleming H. Revell Company.

TENSE, ASPECT AND MOOD IN THE DOCTRINE OF JOHN

CHARLES G. HÄBERL 1. INTRODUCTORY REMARKS

As the only surviving Gnostic religion from Late Antiquity, Mandaeism is unique among the religions of the world, and the sole surviving inheritor to one of the world’s most widespread and influential religious traditions. Its sacred texts and liturgy are recorded in a dialect of Aramaic, and compose one of that language’s largest corpora. Its adherents have preserved both spoken and written forms of their language, as well as a complex body of rituals and a developed commentarial tradition. Despite all this, the history of scholarship on the Mandaeans is one of fits and starts, separated by long periods of virtual inactivity, and many important questions about their religion remain unanswered, including the precise nature of its relationship to Judaism, Christianity, and Islam. Buckley provides a useful synthesis of recent scholarly activity on this matter.1 Much has been made of the fact that no known copies of the Mandaean scriptures antedate the 16th century, the oldest being the Bodleian Library’s MS Marshall 691, copied in 1529,2 and that occasional references to Islam and the prophet Muhammad within them provide a date for their final redaction of some time after the mid-7th century. Nevertheless, a decisive consensus has ultimately emerged in favour of the antiquity of Mandaean traditions. Parallel traditions indicate that these same texts were sometimes adopted and circulated by the adherents of neighbouring religious traditions prior to the advent of Islam. Already in 1949, Torgny Säve1 2

Buckley (2010: 291–314). Buckley (2010: 197).

398

CHARLES G. HÄBERL

Söderberg discovered significant portions of the principal Mandaean scriptures, the Great Treasure, the Canonical Prayerbook, and the Doctrine of John, reproduced nearly word-for-word within the fourth-century Coptic Psalms of Thomas.3 Due in large part to Lidzbarski’s research, the prevailing thesis today is that the sacred books of the Mandaeans comprise a number of distinct texts which were composed at some point in Late Antiquity, redacted together, and acquired their present form shortly after the advent of Islam as a result of the impetus provided by the institution of the dhimma contract.4 This corresponds in part to the Mandaeans’ own traditions about their early encounter with Islam, as chronicled in one of their own texts, the Scroll of Inner Harran.5 Ultimately, the question of influence and the ultimate origins of their texts lies at the centre of the debate over the nature of their relationship to Judaism, Christianity, and Islam.

2. THE PROBLEM

Put simply, according to one school of thought, the later the date of the original composition of the Mandaean sacred texts (be they composed in written format or orally), the less relevant they may be to our understanding of the origins and formative years of these other traditions, a question that has preoccupied scholars of religion for at least a century and a half. On a foundation of earlier research using qualitative cues (e.g. references to the Islamic conquest or historical events that subsequently happened), I propose to develop a relative chronology of the texts that were eventually redacted into the Mandaean Doctrine of John (Draša ḏ-Iahia), one of the central texts of the Mandaean scriptural canon. Säve-Söderberg clearly demonstrated that some portions of the Doctrine of John predate Islam (and even Manichaeism). If it is true that other portions were composed after Islam, how can we distinguish between these portions and determine which are the oldest parts of the composition, and which are the youngest? Is it possible to divide the text into different portions and arrange them according to a relative chronology of composition? If so, assuming that the text assumed its present form shortly after the advent of Islam, as most scholars do, can we then extrapolate which portions were likely composed in the early Islamic period and which were likely composed beforehand? Lidzbarski and other translators have identified several qualitative features useful for dating these texts: Säve-Söderberg (1949: 156–158). Lidzbarski (1915: v–vi). 5 Drower (1953). 3 4

TENSE, ASPECT AND MOOD IN THE DOCTRINE OF JOHN

399

1) References to Islam or historical events that followed its advent; 2) Orthographic conventions, such as historical or phonetic spellings; 3) The presence of Arabic words, either in the form of loanwords or names; and finally, 4) “Modern” (Neo-)Mandaic forms such as the indicative particle -q(a) or the 3rd pl. personal morpheme -iun on the perfect These are all found to some degree or other in different chapters of the Doctrine of John. The problem with most of them is that they are subjective and open to interpretation. Historical references can be interpolated at a later date; manuscripts are copied and recopied constantly throughout the centuries, which presents abundant opportunities to revise their spelling; and transparently Arabic loanwords are rare, at least within this text, and the etymology of a given word is not always unimpeachable given the proximity of the languages. To these, I would add another qualitative distinction, and perhaps a quantitative one, related to the restructuring of the Aramaic verbal system. At some intermediary stage (or stages) between Late Aramaic and Neo-Aramaic, some time between the period of Late Antiquity and the present day, the Eastern Aramaic dialects underwent a major restructuring of their verbal system, as a result of which the ancestors of most surviving Aramaic dialects lost their perfect and imperfect conjugations, to be replaced by new periphrastic conjugations based upon the participles. These participial constructions were already present in earlier Aramaic dialects, in which they were used to supplement the inherited tenses.6 In most Neo-Aramaic dialects, the West Semitic perfect was replaced by a new periphrastic preterite conjugation (with a stem built upon the pattern *qṭil l-), which is conspicuously absent from Neo-Mandaic, but does appear occasionally in Classical Mandaic.7 Likewise, the old Semitic imperfect or “prefix conjugation” was completely replaced by a periphrastic present-future conjugation (built upon the pattern *qāṭil) in all Neo-Aramaic languages save for Western Neo-Aramaic, including NeoMandaic. At this point I would like to advance a hypothesis: the texts offer evidence for the progress of this restructuring, and this evidence can be used to establish a relative chronology of the texts. See Rosenthal (1963: 55) for Biblical Aramaic, Nöldeke (1875: 230–233; 373–386) for Classical Mandaic. 7 Nöldeke (1875: 379): “Dieses hat im Aramäisch eine ähnliche Bedeutungsverwandtschaft mit dem Perfekt wie das aktiv Partizip mit dem Imperfekt; das passiv Partizip verdrängt denn auch im Neusyrisch das Perfekt gänzlich.” 6

400

CHARLES G. HÄBERL

3. METHODOLOGY

As the 18th chapter of the Doctrine of John contains an indisputable late 7 th c. reference (to the Dome of the Rock, which was completed in 691 CE), it serves as a useful “anchor” for my relative chronology. The first step is to establish reasonable limits to the proportional analysis on either side of this anchor. Two texts will represent these limits: the Peshitta version Chapter 1 of the Gospel of Luke in Syriac, on the Annunciation to Zechariah, adopted at some point prior to the early 5 th century,8 and an extended Neo-Mandaic narrative text that I collected in 2003. I chose a Syriac text for the purposes of comparison not only on account of the genre and subject of the text, which are identical, but also the relationship between Classical Syriac and Classical Mandaic, which are closely related, and the absence of any Classical Mandaic text of comparable and demonstrated antiquity. The former contains 205 verbal forms, giving us a 6.9% margin of error with a 95% confidence level, and the latter contains 300 verbal forms, giving us a 5.7% margin of error. These verbal forms include:  Simple Perfect (the “Suffix Conjugation”)  Periphrastic Preterite  Simple Imperfect (the “Prefix Conjugation”)  Periphrastic Present-Future  Imperatives  Infinitives Having collected these forms, I proceeded to compare the proportion of inherited verbal forms to innovated forms in each text.

4. RESULTS

The most common forms in the Peshitta version of Luke, chapter 1, are the inherited suffix and prefix conjugations, and it is clear that these are also the most common ways to express action in the past and present/future, respectively. Table 1: Luke 1

Simple Perfect Periphrastic Preterite Simple Imperfect Periphrastic Present-Future Imperatives Infinitives 8

See Brock (2006: 17).

Absolute Number 103 16 57 25 0 4

Relative Proportion 86.6% 13.4% 69.5% 30.5%

TENSE, ASPECT AND MOOD IN THE DOCTRINE OF JOHN

401

On the other side of the spectrum is the Neo-Mandaic text, which has a considerably simplified verbal morphology relative to the Syriac text. Ignoring the narrative present, 100% of past tense action is expressed by the inherited suffix conjugation and 100% of the present and future tense action is expressed by what Macuch (1965) calls the “participial present-future tense.” The inherited prefix conjugation is conspicuous by its complete absence. Table 2: Neo-Mandaic Text IV

Simple Perfect Periphrastic Preterite Simple Imperfect Periphrastic Present-Future Imperatives Infinitives

Absolute Number 146 0 0 132 22 0

Relative Proportion 100% 0% 0% 100%

In our “anchor” text, the overwhelming majority of action occurring in the past is expressed by the inherited suffix conjugation, and the majority of action occurring in the present or future is expressed by the participial present-future conjugation. Prima facie it would seem that the imperfect or prefix conjugation still occurs with some frequency, about 12% of the time. Table 3: Chapter 18

Simple Perfect Periphrastic Preterite Simple Imperfect Periphrastic Present-Future Imperatives Infinitives

Absolute Number 145 14 13 95 7 5

Relative Proportion 91.2% 8.8% 12.0% 88.0%

On closer examination, however, the majority of these imperfect forms appear to consist of a quotative formula, ḏ-nimar-l- “that (which) he says to x,” which can be inflected for other persons. I have divided the examples into two groups, including this formula, which accounts for 10% of the inherited forms, leaving only two examples as the modest remainder: (1) azal-∅ go.PRF-3SG.M

luat by

liliuk {u}mn PN from

šinta sleep

ni-randid=ẖ 3SG.M-rouse.IMPRF=3SG

‘He went to Lilioch to rouse him from his slumber’. (Chapter 18, JB68.12)

402

CHARLES G. HÄBERL

(2) puq-∅ leave.IMP-SG.M

mn from

iahud PN

ḏ=la=ti-rmia REL=NEG=2SG-throw.IMPRF

b=eurašlam tigra in=PNquarrel ‘Leave Judaea, lest you start trouble in Jerusalem’. (Chapter 18, JB71.9) This suggests that the imperfect is severely restricted to certain contexts (in this case, both apparently purpose clauses) and once the formulaic examples are excluded, the distribution begins to look something very much like Neo-Mandaic: 97.9% innovated tenses to only 2.1% inherited. In Table 4 below, I summarize the results from all the chapters that I have subjected to this analysis thus far, showing the proportion of the inherited forms to the innovated forms for the present and future tenses. The data clearly show a progression from Syriac, in which the imperfect is least restricted in use and therefore accounts for the largest proportion of non-past tense verb forms, to Neo-Mandaic, in which the imperfect has been completely supplanted by the periphrastic presentfuture. Table 4

Luke 1 (Syriac) 55) “Abel's Queries” 3) “Yoshamin” 36) “The Soul Fisher” 11) “The Good Shepherd” 18) “Jack-John” 1) “Truth's Questions” Neo-Mandaic Text IV

inherited (*niqṭul) 69.5% 57.8% 14.5% 11.6% 8.3%

Innovated (*qāṭil) 30.5% 42.2% 85.5% 88.4% 91.7%

Margin of Error 6.8% 7.7% 6.9% 6.9% 9.1%

2.1% 2% 0%

97.9% 98% 100%

5.9% 10.6% 5.7%

Already, “clusters” of texts are beginning to emerge on the basis of the data. At one end of the continuum is Neo-Mandaic Text IV (with 100% periphrastic forms), followed immediately by two chapters from the Doctrine of John, our “anchor text” chapter 18 and chapter 1, both tied at roughly 98% periphrastic vs. 2% simple forms. On the other end of the continuum are those more conservative texts, for which the maximum (at nearly 70% simple against 30% periphrastic forms) is the Syriac text from Luke, followed by chapter 55 from the Doctrine of John (at 58% simple), in which the saviour spirit Radiant Abel (hibil ziua) laments the role he has been

TENSE, ASPECT AND MOOD IN THE DOCTRINE OF JOHN

403

forced to play, governing the material world, and describes the fate of righteous and unrighteous Nazoreans in the afterlife. The rest of the sample texts falls somewhere between these two extremes. Chapter 11 could cluster with chapters 18 and 1 (on the one hand) or chapters 36 and 3 (on the other) because the results are still within the margin of error, but for now it would appear that 3 and 36 form their own distinct cluster, as does 55, all off on its own. Intriguingly, two of these texts, chapters 55 and 1, belong to exactly the same genre, which consists of a dialogue between two “lightworld” beings, one posing questions and the other supplying responses. Despite belonging to the same genre and format, the language of the two could not be more different, with the former resembling Syriac in its Tense-Aspect-Mood system, and the latter resembling NeoMandaic. The last point bears elaboration. In all of the Classical Mandaic texts that I have sampled, including these two, the use of the imperfect is restricted to dependent clauses, such as examples 1 and 2 above, and non-indicative contexts, like the hortatory in example 3 or the conditional in example 4: (3) ana 1SG

e-mar=l=ak 1SG-tell.IMPRF=OBJ=2SG.M

ei-apriš=ak 1SG-explain.IMPRF=2SG.M

kḏ how

kušṭa truth

u=ana and=1SG

ḏ=hu-at REL=COP.PRF-3SG.F

‘Let me tell you the truth, and explain to you how it was’. (Chapter 1, JB4.8–9) (4) kulman whoever

ḏ=adria-∅ REL=carry.PTCP-3SG.M

engirta letter

ni-hlip 3SG.M-pass.IMPRF

‘Whoever carries a letter will pass’. (Chapter 55, JB206.13–207.1) In Chapter 55, the periphrastic conjugation is used for all declarative sentences (example 5), but the imperfect is regularly used for interrogative sentences (examples 6 and 7): (5) mihdia rejoice.INF

kma how

ḏ=ebd-it REL=do.PRF-1SG

hadi-na rejoice.PTCP-1SG

lbab-∅ heart-1SG

ebid-ata deeds-PL.F

b=gau=ẖ in=inside=3SG

ḏ=hazin GEN=this

alma world

‘How joyfully my heart rejoices (at) the things that I have done in this world!’ (Chapter 55, JB202.5–6)

404

CHARLES G. HÄBERL

(6) ana 1SG

ekma ei-anha how 1SG-calm.IMPRF

tir-ata conscience-PL.F

el

OBJ

lib=ai heart=1SG

u=kma and=how

e-šduk 1SG-settle.IMPRF

‘How shall I calm my heart and clear my conscience?’ (Chapter 55, JB202.11– 12) (7) kma how

∅-kbuš sahr-ia 1SG-suppress.IMPRF demon-PL

u=kma and=how

e-gṭal 1SG-kill.IMPRF

mard-ia rebel-PL ‘How shall I suppress demons and slay rebels?’ (Chapter 55, JB202.13–14) In this regard, Chapter 55 differs markedly from both Chapter 1 (example 8) and from our “anchor” text, Chapter 18 (example 9). In both the latter chapters, the periphrastic present-future conjugation is used much more extensively, not only for all declarative sentences but also for interrogative sentences, as demonstrated by examples 8 and 9: (8) amar-u=l=ia tell.IMP-PL.M=OBJ=1SG

arqa kma Earth how

haui-a COP.PTCP-3SG.F

sumk=ẖ width=3SG

‘Tell me, how wide is the Earth?’ (Chapter 1, JB2.7) (9) man who

ḏ=lagiṭ-∅ REL=take.PTCP-3SG.M

aspar book

ḏ=ampašiq-∅=l=kun REL=interpret.PTCP-3SG.M=OBJ=2PL.M

hilm-ia dream-PL hilm-ia dream-PL

ḏ=hz-aitun REL=see.PRF-2PL.M

‘Who is the one who has the book of dreams, which interprets for you the dreams that you saw?’ (Chapter 18, JB67.16–17)

6. CONCLUSIONS

The gradual shift that resulted in the complete replacement of the old Semitic imperfect (ipfv) by the periphrastic present-future conjugation (ptc) in all imperfective non-past contexts is reflected in the texts surveyed, and the language of the Doctrine of John reflects not one but several distinct stages of this shift:

TENSE, ASPECT AND MOOD IN THE DOCTRINE OF JOHN

Stage a b c d e

Indicative ipfv ptc ptc ptc ptc

Interrogative ipfv ipfv ptc ptc ptc

Conditional ipfv ipfv ipfv ptc ptc

405

Irrealis ipfv ipfv ipfv ipfv ptc

In the earliest stage, all imperfective contexts are regularly indicated by the imperfect and only optionally with the periphrastic conjugation. In subsequent stages, the periphrastic ceases to be optional and becomes obligatory for more and more of the primary functions of the imperfect, which becomes relegated to increasingly restricted functions, until it falls out of use entirely. With further research, it may be possible to identify additional stages within this process, and assign individual Classical Mandaic texts to one stage or another. While it will likely never be possible to derive an absolute chronology of these texts in this manner, contextual clues within some of the texts (such as the aforementioned reference to the Dome of the Rock) provide us with valuable data concerning the composition of each text and their relationships to one another. In the case of the Doctrine of John, all references to Islam and other such clues indicating a later composition occur only in those portions of the text that pertain to one of the more advanced stages of the process (d). Most other portions of the text reflect earlier stages of the process (a, b, and c). It is unlikely that all of these portions were composed by the same individual or even by a group individuals belonging to the same generation and living in close proximity to one another. The evidence therefore supports the hypothesis (first advanced by Lidzbarski and subsequently followed by most other scholars) that the Doctrine of John consists of multiple discrete texts, composed by different individuals separated from one another by several or more generations, and redacted together sometime in the early Islamic era. The remarkable absence of the periphrastic preterite from Neo-Mandaic makes it much more difficult to identify trends in its distribution over the same span of time. Nonetheless, it is attested in the Classical Mandaic texts, and further analysis of its distribution there might yield some insight into how and why it came to predominate in other Eastern Neo-Aramaic dialects but disappear completely from NeoMandaic. These trends may also provide useful information for a taxonomy of the Classical Mandaic texts.

406

REFERENCES

CHARLES G. HÄBERL

Brock, Sebastian. 2006. The Bible in the Syriac Tradition. Piscataway, NJ: Gorgias Press. Buckley, Jorunn J. 2010. The Great Stem of Souls. Piscataway, NJ: Gorgias Press. Drower, E. Stefana. 1953. The Haran Gawaita and the Baptism of Hibil-Ziwa. Città del Vaticano: Biblioteca Apostolica Vaticana. Häberl, Charles G. 2009. The Neo-Mandaic Dialect of Khorramshahr. Wiesbaden: Harrassowitz. Lidzbarski, Mark. 1915. Das Johannesbuch der Mandäer, vol. 2. Giessen: Töpelmann. Macuch, Rudolf. 1965. Handbook of Classical and Modern Mandaic. Berlin: Walter de Gruyter. Nöldeke, Theodor. 1875. Mandäische Grammatik. Halle: Waisenhaus. Rosenthal, Franz. 1963. A Grammar of Biblical Aramaic (2nd revised ed.). Wiesbaden: Harrassowitz. Säve-Söderberg, Torgny. 1949. Studies in the Coptic Manichaean Psalm-book. Stockholm: Almqvist&Wiksells.

VERB-GOAL (VG) WORD ORDER IN KURDISH AND

NEO-ARAMAIC: TYPOLOGICAL AND AREAL CONSIDERATIONS1

GEOFFREY HAIG 1. INTRODUCTION

In south-eastern Turkey and the adjacent regions of Iran, Iraq and Syria, the dialects of North-Eastern Neo-Aramaic (NENA) have co-existed with various varieties of Kurdish over many centuries. Unsurprisingly, the languages show a number of deep commonalities, perhaps most evident in the long lists of Kurdish loanwords compiled in the many available descriptions of the NENA dialects. Influence in morphology and syntax is also evident, though unlike lexical similarities, cases of structural similarity generally require much more careful assessment before they can be classified as contact-induced. In this paper, I will briefly examine some candidates for contact-induced shifts in syntax, thereby drawing on more general considerations of language typology. The focus is on a “signature” word order trait of the west Iranian languages, which I term post-predicate goals, or VG word order, and its (possible) relation to NENA. Within language typology, the position of Goal arguments relative to the predicate has not received anything like the same attention as, for example, the relative positioning of direct object and verb (OV vs. VO), and there is a corre-

1

I am very grateful to the participants at the Cambridge conference for valuable feedback on an earlier presentation of this subject matter. Over the last two years a good deal of additional data have become available, which has prompted me to modify my original presentation in several ways. I am very grateful to Eleanor Coghill, Diana Forker, Ergin Öpengin and Don Stilo for providing comments and data on a number of languages, and to the editors of the volume for additional corrections; the responsibility for all remaining errors is my own.

408

GEOFFREY HAIG

sponding lack of large-scale surveys. However, I believe within the general context of contact-induced word-order change, the evidence from such “minor” word order patterns is highly relevant, and has long been overlooked. In particular, the importance attached to the OV/VO parameter is not matched by its saliency in actual discourse: on closer inspection, clauses instantiating this word order parameter are seen to be quite a small minority. This article begins with a brief overview of word order in Kurdish from a typological perspective, then looks at VO/OV in NENA, before turning to a survey of VG word order in Kurdish and closely related languages. Given the breadth of the topic, and its comparative neglect up to now, it goes without saying that what is presented here is still very much provisional, but I think it is possible to identify some intriguing patterns and some key questions for future research.

2. WORD ORDER IN KURDISH: TYPOLOGICAL CONSIDERATIONS

Kurdish here includes varieties belonging to both northern Kurdish (Kurmanji) and Central Kurdish (Sorani); I will also be discussing closely related languages such as Gorani and Zazaki at various points. Like all west Iranian languages, Kurdish has OV word order. Direct objects may be fronted for pragmatic purposes, but are only exceedingly rarely positioned after the predicate.2 Turning now to the three mostwidely cited word-order parameters in typology, Kurdish patterns as follows: (1) OV (direct object-Verb); prepositional (bit not consistently, see below); NG (noun before genitive) This particular combination of word-order parameters is in fact extremely rare in the languages of the world. In the sample of 1142 languages in Dryer (2013), only 14 languages are listed that combine OV word order with prepositions. Among those 14, there are eight with NG word order, i.e. precisely the constellation given in (1); among those eight we find three West Iranian languages (Persian, Tajik and Central Kurdish), and one variety of NENA spoken under heavy Iranian influence (Arbel Jewish). In other words, the constellation given in (1) is not only rare on a global scale, but its occurrence is also disproportionately concentrated in a narrow genetic grouping (West Iranian), and among languages under contact influence of that group.

2

One transitive verb, rahištin (with regional variants) ‘pick up, remove, take away’ actually requires its object to be post-verbal; this appears to be an oddity linked to a single lexical item (thanks to Ergin Öpengin for initially drawing my attention to this).

VERB-GOAL (VG) WORD ORDER IN KURDISH AND NEO-ARAMAIC

409

One conclusion that could be drawn from the global facts is that this particular word order constellation arises only under rather specific conditions, hence its low frequency. In fact, a glance at the global distribution of the combination of O/V word order, adpositions, and NG/GN (cf. Features 95A and 86A in Dryer and Haspelmath 2013) already points to the special position of west Iranian: basically, the entire Asian land block eastwards of Anatolia, and including most of the Indian sub-continent, is dominated by languages with OV, postpositions, and GN word order (Indo-Aryan, Turkic, Uralic, Dravidian, Mongolian, Tungusic etc.). At its western fringe, this block overlaps with languages of North Africa, the Mediterranean, and western Europe, which generally differ on at least two of the three parameters mentioned. West Iranian, then, can be viewed as the westernmost outlier of an OV/postpositional/GN block, situated in the overlap with a neighbouring linguistic macro-area. Its unusual word-order profile can, thus, be considered a compromise profile, in the sense of Stilo (2005). Harris and Campbell (1995: 137) argue that in fact, certain word order constellations only arise in contact situations; one of the constellations they discuss is that given in (1) above.3 Note, however, that west Iranian languages spoken further to the North and the East, e.g. Māzanderāni in the Caspian region, are largely GN, and extensively postpositional. The constellation of word-order parameters shown in (1) is, thus, not valid for all west Iranian languages, but all those spoken in Anatolia and much of central and western Iran. Note further that even among the latter group of west Iranian languages, there are differences in the extent to which they can be considered “prepositional”. For Central Kurdish, and the dialects of Northern Kurdish spoken in southeast Anatolia, however, the dominance of prepositions is undeniable. And it can hardly be coincidence that in precisely these regions contact with prepositional Semitic languages has been most intense .

3. OV WORD ORDER IN NENA

In a similar vein, NENA can be considered the north-easterly outlier of Afro-Asiatic, and the shift to OV word order found in the trans-Zab Jewish dialects squarely reflects this fact: the languages on the overlap of two macro-areas exhibit the properties of both. The dialect of Bohtan (Fox 2009), though not from a Jewish speech

3

Harris and Campbell (1995) also include the relative positioning of Adjective and Noun (NA/AN). However, more recent work by Dryer (e.g. 2011) shows that NA/AN does not correlate significantly with the other parameters, hence it is not included here.

410

GEOFFREY HAIG

community, also shows predominant OV order, thus bringing it into line with the profile given in (1). The following examples illustrate OV in this dialect:4 (2) brota girl

axəst-aw ring-POSS.3SG.F

yawó-la give.PST-3SG.F

l-jambali to-Jambali

‘The girl gave her ring to Jambali’. (Fox 2009: 101) (3) danw-i tail-POSS.1SG

nəmmun why

qṭəlax-le cut_off.PST.2SG.F-OBJ

‘Why did you cut off my tail?’ However, OV word order in Bohtan is not the only option, and Fox provides several examples with post-predicate objects (VO): (4) abre-ni son-POSS.1PL

kebé-la love.3SG.M-OBJ

bratexun daughter.POSS.2PL

‘Our son loves your daughter’. (Fox 2009: 108) Thus the OV word order of the Bohtan dialect, unlike the OV order of neighbouring Kurdish dialects, is not a categorical rule, but rather a preference, i.e. “the most common order of constituents” (Fox 2009: 107). While it can scarcely be doubted that the shift to predominant OV order in Bohtan and other varieties of NENA is linked to contact influence from neighbouring OV languages (Azeri Turkish and Kurdish), the question arises as to just how such a shift proceeds. Silva-Corvalán (1998) argues at some length against the notion of wholesale “borrowing” of syntactic structure from a donor into a recipient language. Instead, she suggests that on closer inspection, most of the claimed examples are amenable to an analysis in terms of extension in the distribution of existing structures into novel semantic and pragmatic contexts. In other words, surface structure is not imported; rather, the distributional restrictions on existing surface structures are relaxed to bring the form-meaning pairings of the recipient language closer to those of the donor language. In the present case, we might expect to find OV word order as a—perhaps pragmatically-marked and low-frequency option—for transitive clauses in NENA. Khan (2008a) discusses different word orders for objects in the Barwar dialect of NENA, indicating that direct objects may occur pre-predicatively, 4

Examples from NENA have been supplied with a rudimentary morphological gloss, where only those grammatical features relevant to the discussion at hand are included, and object language forms are general not segmented. In some cases I have also minimally simplified the transcription in comparison to the original.

VERB-GOAL (VG) WORD ORDER IN KURDISH AND NEO-ARAMAIC

411

even though this is not the pragmatically neutral order. A corpus count of 104 transitive clauses with overt NP objects (from the first 864 clauses of the text Gozali and Nazali, Khan 2008b) revealed 10 cases of fronted NP objects from a total of 104 full NP objects, i.e. around 10% (cf. Schnell and Haig (forthcoming) for the background to this study). The following examples illustrate pre-predicate objects in the Barwar dialect: (5) ʾáyya this.PRON

gán-i self-POSS.1SG

fadən-na ransom.1SG-OBJ

‘Her (lit. ‘this.SG.F’), (I) myself will ransom her’ (6) ʾánna these

kúl-la θárwa all-3SG.F wealth

šáwp-i place-POSS.1SG

díya OBL.PRON.3SG.F

b-šaqlì-la=w FUT-take.3PL-OBJ=and

b-šaqlí-le=u FUT-take.3PL-OBJ=and

‘All this wealth (of the kingdom) they will take it, my place they will take it and …’. In the restricted number of other NENA sources available to me, pre-verbal direct objects are also quite easy to find, as in the following example from Ashitha: (7) dawarɛ́ha mule.PL

mròpyə-lla-wɛwa release.RES.PL-OBJ-COP.PST.3PL

‘They (the Kurds) had released their mules’ (Borghero 2005: 374, sentence 16) Now there are obviously structural factors in NENA that favour such constructions, in particular the use of the L-suffixes to index objects on the verb. Thus at least some of the cases of pre-predicate objects can meaningfully be analysed as fronted topics of the sort: (8)

[Xi, (Y) verbed-iti]

In fact, the ten examples of pre-predicate objects noted in Schnell and Haig (forthcoming) are all accompanied by an L-suffix on the verb indexing the object, suggesting that this feature licenses (or at least favours) the non-canonical positioning of the object. The same is true of other examples from Barwar not included in the text investigated by Schnell and Haig (forthcoming), as in the following:

412

GEOFFREY HAIG

(9) ʾána-ži PRON.1SG-ADD ʾá-kθɛθa the-chicken

m-kɛ́rb-i of-anger-POSS.1SG mútt-ən-na put.RES-1SG-OBJ

ʾǝ́θyən come.RES.M.1SG rəš-núra on-fire

ʾu-xìl-ən-na. and-eat.RES-1SG.M-OBJ

‘I, out of my anger, came and put the chicken on the fire and ate it’. (Khan 2008b, A2:11) Whatever the pragmatics behind such pre-predicate objects may be, the simple fact remains that pre-predicate objects are not entirely alien to the surface structure of NENA syntax, though we await a more representative survey of the actual extent of such constructions in usage. It therefore does not seem unreasonable to see the apparent “borrowing” of OV syntax from Kurdish in terms of a “de-pragmatization” of structures such as (8), leading to an expansion of this word-order variant into pragmatically neutral contexts. Thus the shift evident in NENA dialects such as Bohtan to a predominant OV structure may in fact be less of a radical shift than it appears at first sight. And conversely, the relative stability of OV-word order in West Iranian (no variety is known that has shifted to VO) could be linked to the lack (or exceedingly low frequency) of a pragmatically-determined VO word order variant in these languages. Now a difference between OV and VO may appear, at least from the perspective of language typology, like a fairly drastic word-order mis-match. But it is worth considering it from the perspective of actual discourse. Transitive clauses with overt NP objects are not a particularly frequent clause type: of the 864 clauses of NENA discourse examined in Schnell and Haig (forthcoming), only 104 (12%) had overt NP objects (and recall that among these, 10 were pre-verbal). A further 120 had clitic pronoun objects, but note that such pronominal objects in NENA attach to the verb, as do pronominal objects (generally) in Central Kurdish. Given that in both language groups all Goals are post-predicate (see next section), and most of NENA dialects have innovated clause-final copulas, bringing their copula clauses in line with those of Central Kurdish, it is evident that with the exception of clauses with overt NP objects, most of the clauses in actual connected discourse in both NENA and Central Kurdish exhibit the same word order. In other words, the fact that much of NENA still has VO, while Kurdish has OV, suggests a much higher degree of surface divergence in word order than is actually warranted. Once one looks at connected discourse, we observe that the quantitative impact of this difference is not high, affecting perhaps 10-20% of actual clauses. Outside of this clause type, in natural connected discourse NENA and Central Kurdish word orders have synchronized to yield a very high level of word-for-word isomorphism.

VERB-GOAL (VG) WORD ORDER IN KURDISH AND NEO-ARAMAIC

413

In the next section, we turn to a different word order pattern which has received comparatively little attention in typology, but which is, from a frequency perspective, scarcely less important than OV/VO.

4. VG WORD ORDER IN KURDISH

Although Kurdish is consistently OV, it is certainly not consistently “verb final”. In fact, all varieties of Kurdish systematically place certain elements after the verb; the most prominent kind of post-predicate argument is what I will term “Goals”. The term “G(oal)” is used here as a cover term to refer to: locational goals of verbs of motion, which for the sake of simplicity covers both human and non-human directional goals (though there are reasons to distinguish them, as will become apparent) recipients of verbs of transfer addressees of verbs of speech Thus, my term “Goal” covers what is traditionally termed “Indirect Object”, but also includes other semantic roles. The grouping together of these different object types is purely dictated by the grammars of the languages concerned, which tend to treat all Goals in this sense in a similar manner. Language typology has until very recently remarkably little to say on the order of Goals relative to verbs, and I am unaware of any large-scale systematic survey of the issue. The most well-known cases are the Mande languages of Liberia and the Republic of Guinea (Niger-Congo languages). In Kpelle, objects occur before the predicate (10), but prepositional phrases are regularly post-predicate (11):5 (10) galoŋ chief

a

AGR

pɛ́rɛ tɔɔ̂i house build

‘The chief is building a house’. (11) e

AGR

sɛŋ-kâu money

tèe sent

ḱâloŋ-pə chief-to

‘He sent money to the chief’.

Travis (1989: 269), citing Gay and Welmers (1971) and Givòn (1975) respectively, hence the minor differences in transcription; glosses follow Travis (1989). 5

414

GEOFFREY HAIG

Following an inquiry on the Linguistic Typology mailing list, a number of researchers on OV languages reported widespread use of post-predicate goals, notably often in languages in contact situations (e.g. Mono-Uruvuan, a small group of Oceanic languages which have shifted from inherited Oceanic VO to OV [Bill Palmer, personal communication]). For the Skou language of Papua New Guinea, Donohue (2004) notes that “... the only nominals that can appear in a postverbal position are goals and locations” (Donohue 2004: 131). Another case is the Oceanic language Saliba, spoken on Papua New Guinea, which has become OV presumably through contact with Papuan languages. In Saliba, locational Goals are generally expressed through bare NPs after the predicate—remarkably reminiscent of the situation in Kurdish: (12) ya-dobi maketi 1SG-go_down market ‘I go down to the market’. (Margetts 1999: 286) For Hinuq, a Nakh-Daghestanian OV language of Daghestan (Forker 2013), goals are apparently placed after the predicate in 20-30% of clauses (based on a count of 1000 clauses, Diana Forker, personal communication). The phenomenon of post-predicate Goals in OV languages appears, therefore, to be a solidly attested option in discourse, but the extent of such usage obviously varies from language to language, and has never been systematically examined. It is also unclear whether we are dealing with pragmatically-determined occasionalisms, or grammaticalized phrase-structure rules. Frequent post-predicate positioning of Goals would appear to make sense from a cognitivist perspective: Clause-final position of Goals could be seen as an iconic reflection of Goals as the natural endpoints of events. This, however, remains a question for future reference. But the conclusion that can reasonably be extracted from these preliminary findings is that OV languages generally seem to permit a certain amount of “leakage” rightwards of the predicate, and that Goals are among the most frequently post-posed constituents. This fact suggests that they are less strictly linearized, and hence inherently more susceptible to contact influence. VG word order occurs in all West Iranian languages known to me, though the frequencies of occurrence vary. Some examples are: (13) Persian raft-e bud-am go.PST.PTCP AUX.PST-1SG

jangal forest

‘I went to the forest’. (Roberts 2009: 483)

VERB-GOAL (VG) WORD ORDER IN KURDISH AND NEO-ARAMAIC (14) agarna pesar=e=rā otherwise boy=DEF=OBJ.MARK dast=e hand=EZ

415

mi-sepord-am IMPRF-turn_over.PRS-1SG

amnie-hā gendarmerie-PL

‘Otherwise I would have handed the boy over to the gendarmerie’. (Roberts 2009: 140) (15) Northern Kurdish (the Zakho-dialect) rābī get_up.PST(3SG)

dē so

ē-n bô come.SUBJ-PL to

māl-ē house-OBL

‘(They) set off to come home’. (MacKenzie 1962: 354, transcription modified)6 (16) kutilik-ā rissole-EZ t

ADP

dayk-ā mother-EZ

sēnīk-a tray-EZ

xwa

REFL

xāl-ē uncle-EZ

kir put.PST(3SG) xwa-dā REFL-ADP

‘(He) put his mother’s rissole onto his uncle’s tray’. (17) Vafsi bæ-væsd PUNCT-jumped(3SG)

man middle

aw-e water-OBL

‘He jumped into the water’. (Stilo 2005: 231, transcription and gloss slightly modified) (18) Gorani ya one

guɫ-e rose-EZ

bi-ya SUBJ-give.PRS

ī

DEM

bāxč-at-a garden=2SG=DEM.CLC

min 1SG

‘Give me a rose of this garden of yours!’ (Mahmoudveysi et al. 2012: ex. 4: 54)

6

MacKenzie (1962: 355) interprets rābī differently; the translation given here was suggested by Ergin Öpengin, and appears to be preferable.

416

GEOFFREY HAIG

(19) Dialect Of Sivand čāder-et-ā veil-2SG.CLC-OBJ.MARK

be-de SUBJ-give.PRS(2SG)

ba to

me me

‘Give me your veil!’ (Le Coq 1979: 89, Sentence 16) (20) DELVARI, SOUTHWEST IRAN sova tomorrow

va-mi-gard-om PREV-IMPRF-return-1SG

delvar Delvar

‘Tomorrow I will return to Dalvar’. (Bušehr, Haig and Nemati 2013) (21) Sistan Balochi wat-ī mardum-ān-ā REFL-GEN people-PL-OBJ mnī 1SG.GEN

dēm face

day give.IMP.SG

bagg-ay herd_of_camels-GEN

sarā to

‘... send your people to my herd of camels’. (Delforooz 2010: 331) Not all varieties of Iranian, however, use post-predicate goals to the same extent, and a closer look at texts reveals some interesting tendencies. In what follows I present some provisional findings from texts in a number of West Iranian languages, with the main focus on different varieties of Kurdish

4.1. Gorani and Central Kurdish

Consider first the Gorani dialects of the Gawrajū and Zarda, two villages of West Iran. The dialects represent pockets of what was once a much larger Gorani language region, which extended into North Iraq. There are obvious similarities to the chain of dialects spoken in modern North Iraq, almost as far as Mosul. The text collections published in Mahmoudveysi et al. (2012) and Mahmoudveysi and Bailey (2013) show that in both dialects, Goals as defined above, are overwhelmingly postpredicate. The only significant exception in Gawrajū are addressees, which seem to allow both orders, post verbal in (22) and preverbal in (23): (22) wat=iš=a say.PST=3SG.AGT=DIR

dayka [...] mother

‘(She) said to (her) mother ...’ [Gawrajū 8: 157]

VERB-GOAL (VG) WORD ORDER IN KURDISH AND NEO-ARAMAIC (23)

ī this

kurat-a young_man-DEM

dī wan=im see.SUBJ.2SG to=me

417

bwā say.IMP.SG

‘When you see this young man tell me’. [Gawrajū 8: 139] In Zarda, basically the same situation obtains, but prepositional phrases with pay ‘for, too’ are generally pre-predicate, when the complement is human, and particularly when it is pronominal: (24) labās=iš pay clothes=3SG.AGT for pay for

ī minal-ī-a this child-OBL-DEM

wē-m

REFL-POSS.1SG

hāwird, bring.PST

hāwird bring.PST

‘She brought clothes for me, she brought clothes for this child’. Otherwise, however, general expressions of direction or Goal are postpredicate. They may be with a preposition, as in (25), or simply a bare NP, as in (26):7 (25) duwāra min=iš bar(d) pay again 1SG=3SG.AGT take.PST to

ya a

bīmārisān-ī hospital-INDF

‘Again he took me to a hospital’. [Zarda 2-1: 50] (26) hāwird-im=šān ābād-ī wē=mān bring.PST-1PL=3PL.AGT village-EZ REFL=POSS.1PL ‘The brought us to our village’. [Zarda 1-1: 150] Turning to Central Kurdish, the dialect of Suleimaniye (MacKenzie 1962) exhibits essentially the same system for Goals as does Gorani. Non-human goals of verbs of motion are always post-predicate, either bare or with a preposition. As in Gorani, Central Kurdish Goals are also generally accompanied by a “directional” particle on the verb, most probably the relic of an old preposition which must originally have introduced the Goal argument (the directional is not attached to vowelfinal verb forms). The human complements of verbs of speech appear either pre- or post-predicate (there are too few overtly expressed addressees to discern a clear tendency; generally they are pronominalized and cliticize to the verb, or they are left unexpressed). In Central Kurdish, as in all other varieties of Kurdish known to me, 7

There is possibly a frequency effect here; highly frequent goals such as house/home, or village may be more acceptable without a preposition, but this is still speculative.

418

GEOFFREY HAIG

the recipients of ‘give’ are expressed post-verbally, unless they are pronominal, in which case they occur as verbal affixes or clitics hosted by a constituent of the VP (cf. Öpengin 2013 for details).

4.2. Northern Kurdish (Kurmanji)

Goals in the texts from Midyat (Ritter 1976), Amadiya (Blau 1975), and Muš (my own field work) were investigated. The first three are intended as representative of the Northern Kurdish dialects spoken in close proximity to NENA dialects, while the Muš variety is intended as representative of Central Anatolia, and more generally what Haig (forthcoming) refers to as the Caspian/Caucasian sphere in Anatolia. The general tendencies outlined above obtain in all varieties: non-human Goals are overwhelmingly post-predicate, recipients of ‘give’ are always post-predicate, including pronominals (there are no clitic pronouns in these dialects). However, there are two discernible differences between the varieties of southeast Anatolia and North Iraq, on the one hand, and the Muš variety, on the other. First, the position of Addressees. In the south-eastern, or Mesopotamian dialects, addressees are preferred in the post-predicate position, while in Muš, pre-predicate positioning is the rule. For example, in the text from Midyat, the four attested examples of addressees were all post-verbal (cf. 27–28), and whereas in the Muš texts, all six examples were preverbal, complements of the circumposition ji ... ra (29–30): (27) gōt-a Ūsufšá, xatûn-ē gō... say.PST-DIRC Usufsha, lady-OBL.F say.PST ‘(She) said to Usufsha, the lady said …’ (28) bêj-a say.IMP.SG-DIRC

wī

DEM.OBL.M

ṡuwâr-ī horseman-OBL.M

‘Say to this horseman’ (29) paşa Pasha

ro-k-î day-INDF-OBL

ji

ADP

wezîr-ê Vizier-EZ.M

xwe

REFL

ra

ADP

got say.PST

‘One day the Padishah said to his Vizier’ (30) Wezîr Vizier

ji

ADP

jin-a wife-EZ.F

xwe

REFL

ra

ADP

got-î-ye say.PST-PTCP-COP.3SG

‘The Vizier said to his wife’ Note also the use of the post-predicate ‘directional particle’ on the verbs in (2728), also a feature that is most commonly found in dialects further south (and of course in Central Kurdish).

VERB-GOAL (VG) WORD ORDER IN KURDISH AND NEO-ARAMAIC

419

The pre-predicate pattern is also the general one for Kurdish of Armenia, but the circumposition ji ... ra is reduced to just the postpositional element -ra: (31) mi 1SG.OBL

xewn-ek dream-INDF

xewn-a xu dream-EZ.F REFL

dît-î-ye see.PST-PTCP-COP.3SG te-ra 2SG.OBL-to

ez=ê 1SG=FUT

bêj-im say.PRS.SUBJ-1SG

‘I had a dream, I will tell you my dream’. (Djelil and Djelil 1978: 294) Exactly the same tendency holds for Zazaki, where adressees are prepredicative and marked by a clitic =rē: (32) šofird-ē-xo=rē driver-EZ-REFL=to

vano say.PRS.3SG

‘(He) says to his driver’ (Paul 1998: 218) (33) Mudir-ī=rē director-OBL-to

vano say.PRS.3SG

‘(He) says to the director’ (Paul 1998: 218) The second major split between the North Iraq/South-East Anatolian varieties of Northern Kurdish, and the central Anatolian varieties, is the use of prepositions with post-predicate Goals. In the south, there is a pronounced tendency for postpredicate Goals to be accompanied by a preposition or circumposition. The following examples are illustrative, as are (16) and (25) above: (34) waxtē when

čô go.PST.3SG

bô to

māl-ē house-OBL.F

‘when he went to the house’ (dialect of Gullī, near Zakho, MacKenzie 1962: 348, par. 752) (35) ... u and

kilîlk' kir-e key put.PST-DIRC

di

ADP

berîk-a pocket-EZ.F

xo

REFL

da

ADP

‘... and put the key into his pocket’ (Blau 1975: 106 par. 29) The same tolerance of post-predicate prepositions is also found in Mukri Kurdish (Öpengin 2013), where likewise the directional particle is also used (when the verb form ends in a consonant):

420

GEOFFREY HAIG

(36) were come.IMP

de-t-be-m-e IMPRF-2SG.P-bring-1SG-DRCT

bo to

māł-e house-EZ

xūšk=im sister=POSS.1SG ‘Come, I will take you to my sister’s house’. In the dialects of Central Anatolia, however, post-verbal Goals are not accompanied by prepositions, but are simply bare NPs.8 The only exception are human Goals, where a preposition is used, e.g. cem, as in: were cem min ‘come to me’ (Muš). Thus we find in the dialects of the south a comparatively high tolerance of clauses with the sequence V ADP NP, whereas in the Central Anatolian varieties, this combination is very restricted. To sum up this brief overview, Kurdish and closely related languages of East Anatolia and North Iraq share certain commonalities in the placement of Goals, but there are also areally-determined differences. The following two tables present differences and commonalities: Table 1: Goals in North Iraq, southeast Anatolia pre-/post-verbal

adposition/bare

Recipients of ‘give’

post-verbal

bare, but often with direction-

Addressees

post-verbal pre-

adp./bare with directional

Local Goals, human

post-verbal

with directional particle and

Local Goals, non-human

post-verbal

bare/may take adposition

8

ferred

al particle on verb particle

adposition

and/or directional particle

Though they may be accompanied by ‘local nouns', such as ser, lit. ‘head’, grammaticalized to a near-preposition ‘on’.

VERB-GOAL (VG) WORD ORDER IN KURDISH AND NEO-ARAMAIC

421

Table 2: Goals in Central Anatolia, including Zazaki pre-/post-verbal

adposition/bare

Recipients of ‘give’

post-verbal

bare

Addressees

pre-verbal preferred

with adposition

Local Goals, human

post-verbal

with adposition

Local Goals, non-human

post-verbal

bare

4.3. Interpreting the findings

Two broad features stand out as relatively consistent across all the varieties of Kurdish considered here: the post-predicate expression of recipients of the verb ‘give’, and of many non-human locational Goals. Both features are also well-attested in west Iranian languages outside the NENA area, as shown above, suggesting an inherited trait from the earliest predecessors of west Iranian. However, there are a number of finer-grained differences in the way these features are manifested, which point to some interesting areal effects. Perhaps the most general is the greater tolerance of post-predicate prepositional phrases in South-East Anatolia/North Iraq. If we consider the directional particle as the relic of an old preposition, then this tendency is further reinforced. Thus these dialects regularly instantiate clauses with the wordorder V ADP NP. This is a surface configuration that is completely typical of all varieties of NENA. The Kurdish dialects of Central Anatolia, and Zazaki, have only a very restricted use of such a word-order type. Either Goals are pre-predicate, and introduced by a circumposition or postposition, or they are post-predicate, but bare. It is worth asking how the historically consistently OV Iranian languages should have developed post-predicate Goals at all. Given the wide distribution of the feature, it seems reasonable to assume that it is an old trait of West Iranian that was inherited by the daughter languages, rather than a parallel independent development. Old Persian certainly allowed post-predicate local Goals, as in (37) and (38): (37) pasāva then

Gaubaruva Gobryas

hadā with

kārā army

ašiyava ūvjam marchedto_Elam

‘Then Gobryas marched to Elam with an army’. [DB V, 9–10, Kent 1953: 133] (38) pasāva hadā then with

kārā army

ašiyavam went.1SG

abiy to

sakām Scythia

‘Then I went with an army to Scythia’. [DB V, 21, Kent 1953: 133] However, VG in Old Persian is not as consistent as it is in modern Gorani, or Central Kurdish, though a thorough investigation of the phenomenon in Old Persian has yet to be carried out. With regard to recipients and benefactives in the Old Per-

422

GEOFFREY HAIG

sian corpus, a definitive statement is not possible. Most are pronominal, which might have differing word order properties from full NPs. Free pronoun recipients, however, regularly appeared in the pre-predicate position (cf. Haig 2008: 48–57 for examples). This would be extremely unusual in, for example, modern Gorani. Therefore, a preliminary conclusion would be that already in the Old Iranian period, VG existed as a possible variant, but the modern languages such as Gorani and Central Kurdish developed their consistent post-predicate Goals some time during the formative period of these languages from a presumed Middle Iranian ancestor. Utas (2013: 170–173) discusses the possible influences of Aramaic on Old and Middle Iranian, suggesting that, for example, the ezafe-construction that emerged in Parthian and is characteristic for many (but not all) modern West Iranian languages may have been influenced by an Aramaic model. In the same vein, the development of VG word order in the west Iranian languages, particularly those of northern Iraq and west Iran, could be considered part of the ancient imprint of Aramaic. We know from the facts of Turkmen of Iraq that VG word order is certainly borrowable into OV languages; Bulut (2007: 175) suggests that Iraqi Turkmen has borrowed the post-verbal positioning of Goal arguments from “Iranian languages”. Likewise, the Azeri texts in Kiral (2001) from the Tabriz region also show ample evidence of VG word order, almost certainly through Iranian influence (Kiral 2001: 142, sentence 6): (39) vä sähär-lär and morning-PL

tez early

dur-ar-di get_up-AOR-PST

ged-ar-di go-AOR-PST

ambar-a cellar-DAT

‘and in the mornings he would get up early and go into the cellar’ If we assume that Kurdish evolved from a group of west Iranian dialects spoken originally in the region of Hakkari/North Iraq, then I would assume that it was already characterized by consistent VG order. Dialects that remained in that region, such as Central Kurdish and some Northern Kurdish dialects, maintained close contact with NENA, and retained VG (or perhaps even extended it), basically leaving the word-order profiles of these languages almost identical (see section 3 above). The dialects of Kurdish that expanded westward and northward into Anatolia came under stronger contact influence from Armenian and Turkish. Possibly as a result of this, we find the shift towards (i) increasing use of pre-verbal addressees; (ii) decreasing tolerance of post-verbal prepositions. The dialects of Central Anatolia, and Zazaki, resemble in these respects much more closely Turkish and East Armenian.

5. CONCLUDING REMARKS

Thus, the scenario I am suggesting to account for the areal differences discernible in the VG word orders of different varieties of Kurdish would assume that an original “proto-Kurdish” was already characterized by consistent VG order, perhaps through

VERB-GOAL (VG) WORD ORDER IN KURDISH AND NEO-ARAMAIC

423

early Aramaic/Iranian contacts. This pattern has remained essentially unchanged in Central Kurdish and the southernmost dialects of Northern Kurdish which share the same areal distribution as NENA. The Central Anatolian dialects, on the other hand, have weakened VG to varying degrees as they progressively spread out of the NENA region. In NENA, I assume VG was the inherited norm anyway, and it has scarcely changed—even in those dialects which have shifted to OV. Thus Kurdish and neighbouring NENA dialects have locked into a basically identical word order profile, which is presumably of considerable antiquity. The fact that they (mostly) differ with regard to the ordering of direct objects is undoubtedly very significant, but viewed from the perspective of overall discourse patterning, is perhaps no more significant than the identical positioning of Goals in both groups. More generally, these findings confirm Silva-Corvalán’s (1996: 18) conviction that contact influence may be more readily detected “through differences in the frequency of use of a certain structure rather than its categorical occurrence or non-occurrence”.

REFERENCES Blau, Joyce. 1975. Le Kurde de ‘Amādiya et de Djabal Sindjār. Analyse Linguistique, textes folkloriques, glossaires. Paris: Klincksieck. Borghero, Roberta. 2005. The Neo-Aramaic dialect of Ashitha. Unpublished PhD Dissertation. University of Cambridge. Bulut, Christiane. 2007. “Iraqi Turkman.” In Languages of Iraq, Aancient and Modern, edited by John Postgate, 159–187. Cambridge: British School of Archaeology in Iraq. Delforooz, Behrooz Barjasteh. 2010. Discourse Features in Balochi of Sistan (Oral Narratives). Uppsala: Studia Iranica Upsaliensa. Djelil, Ordixane, and Djelil, Celile. 1978. Zargotina K’urda. Moscow. Donohue, Mark. 2004. A Grammar of the Skou Language of New Guinea, Online Draft Version. [online] http://www.papuaweb.org/dlib/tema/bahasa/skou/Skoufull1.pdf. Dryer, Matthew. 2011. “Noun-modifier Order in Africa.” In Geographical Typology and Linguistic Areas, with Special Reference to Africa, edited by Osamu Hieda, Christa König and Hiroshi Nakagawa, 287–311. Amsterdam: Benjamins. Dryer, Matthew S. 2013. “Relationship between the Order of Object and Verb and the Order of Adposition and Noun Phrase.” Max Planck Institute for Evolutionary Anthropology, Leipzig. [online] http://wals.info/chapter/95 Dryer, Matthew, and Haspelmath, Martin. 2013, (eds.). The World Atlas of Language Structures Online. Munich: Max Planck Digital Library.

424

GEOFFREY HAIG

Forker, Diana. 2012. A Grammar of Hinuq. Berlin: Mouton de Gruyter. Fox, Samuel Ethan. 2009. The Neo-Aramaic Dialect of Bohtan. Piscataway, NJ: Gorgias Press. Gay, John, and Welmers. William. 1971. Mathematics and Logic in the Kpelle Language. University of Ibadan, Institute of African Studies. Occasional publications no. 21. Givón, Talmy. 1975. “Serial Verbs and Syntactic Change: Niger-Congo.” In Word Order and Word Order Change, edited by Charles Li, 47–112. Austin: University of Texas Press. Haig, Geoffrey. 2008. Alignment Change in Iranian Languages. A Construction Grammar Approach. Berlin: Mouton de Gruyter. —. forthcoming. East Anatolia as a Linguistic Area? Conceptual and Empirical Issues. Haig, Geoffrey, and Nemati, Fatemeh. 2013. “Clitics at the Syntax-pragmatics Interface: The Case of Delvari Pronominal Enclitics.” Paper held at the Fifth International Conference on Iranian Linguistics (ICIL5). [online] http://bamling-research.de/data/2013-icil5/presentations/ Haig, Geoffrey, Schnell, Stefan, and Wegener, Claudia. 2011. “Comparing Corpora from Endangered Languages. Explorations in Language Typology Based on Original Texts.” In Documenting Endangered Languages: Achievements and Perspectives, edited by Geoffrey Haig et al., 55–86. Berlin: Mouton de Gruyter. Harris, Alice, and Campbell, Lyle. 1995. Historical Syntax in Cross-linguistic Perspective. Cambridge: Cambridge University Press. Kent, Roland. 1953. Old Persian (2nd revised ed.) American Oriental Society: New Haven, Connecticut. Khan, Geoffrey. 2008a. The Neo-Aramaic Dialect of Barwar, Vol. 1: Grammar. Brill: Leiden. —. 2008b. The Neo-Aramaic Dialect of Barwar, Vol. 3: Texts. Brill: Leiden. Kiral, Filiz. 2001. Das gesprochene Aserbaidschanisch von Iran. Eine Studie zu den syntaktischen Einflüssen des Persischen. Wiesbaden: Harrassowitz. Lecoq, Pierre. 1979. Le Dialect de Sivand. Reichert: Wiesbaden. MacKenzie, David Neil. 1962. Kurdish Dialect Studies, Vol. II. London: Oxford University Press. Mahmoudveysi, Parvin, Bailey, Denise, Paul, Ludwig, and Haig, Geoffrey. 2012. The Gorani language of Gawraju (Gawrajuyi), a Village of West Iran. Texts, Grammar and Lexicon. Reichert: Wiesbaden. Margetts, Anna. 1999. Valence and Transitivity in Saliba, an Oceanic Language of Papua New Guinea. Max Planck Institute for Psycholinguistics: Nijmegen. Öpengin, Ergin. 2013. Clitic/affix Interactions. A Corpus-based Study of Person Marking in the Mukri Variety of Central Kurdish. Unpublished PhD Dissertation. Paris Sorbonne/University of Bamberg.

VERB-GOAL (VG) WORD ORDER IN KURDISH AND NEO-ARAMAIC

425

Paul, Ludwig. 1998b. Zazaki. Grammatik und Versuch einer Dialektologie. Reichert: Wiesbaden. Ritter, Helmut. 1976. “Kurmânci-Texte aus dem Ṭûr’abdîn, Teil II.” Oriens 25-26: 1– 37. Roberts, John. 2009. A Study of Persian Discourse Structure. Uppsala: Acta Universitatis Upsaliensis. Schnell, Stefan, and Haig, Geoffrey. forthcoming. “Assessing the Relationship between Object Topicalisation and the Grammaticalisation of Object Agreement.” Proceedings of the Australian Linguistics Society, 2013, edited by Lauren Gawne and Jill Vaughan. Silva-Corvalán, Carmen. 1996. Language Contact and Change. Spanish in Los Angeles. Oxford: Clarendon Press. —. 1998. “On Borrowing as a Mechanism of Syntactic Change.” In Romance Linguistics: Theoretical Perspectives. Selected Papers from the 27th Linguistic Symposium on Romance Languages (LSRL XXVII), Irvine, 20–22 February, 1997, edited by Armin Schwegler, Bernard Tranel and Myriam Uribe-Etxebarria, 225–246. Amsterdam: John Benjamins. Stilo, Donald. L. 2005. “Iranian as Buffer Zone between the Universal Typologies of Turkic and Semitic.” In Linguistic Convergence and Areal Diffusion: Case Studies from Iranian, Semitic and Turkic, edited by Éva Ágnes Csató, Bo Isaksson and Carina Jahani, 35–63. London: Routledge. Travis, Lisa. 1989. “Parameters of Phrase Structure.” In Alternative Conceptions of Phrase Structure, edited by Baltin Mark L. and Anthony Kroch, 263–279. Chicago: University of Chicago Press. Utas. Bo. 2013. “The Grammatical Transition from Middle to New Persian.” In From Old to New Persian, edited by Bo Utas, Jahani, Carina and Mehrdad Fallahzadeh, 251–259. Beiträge zur Iranistik, Band 38. Wiesbaden: Reichert.

ON THE CONVERGENCE OF VERBAL SYSTEMS OF ARAMAIC AND ITS

NEIGHBOURS. PART I: PRESENT-

BASED PARADIGMS

PAUL M. NOORLANDER AND DONALD STILO

1

1. INTRODUCTION

Neo-Aramaic, specifically North-Eastern Neo-Aramaic (NENA), presents us with one of the most multi-faceted cases of language contact and areal phenomena among the world’s languages. In the transition from Old Aramaic to Neo-Aramaic, several profound morphosyntactic changes occurred attesting to various fascinating evolutionary cycles (cf. Jastrow 2008). The gradual loss of morphological case-marking through phonetic erosion and the merging of the Tense-Aspect-Mood (TAM) system in North-West-Semitic had a major impact on the restructuring of her Iron Age descendants, which are characterized by the concomitant emergence of a definite article and new dependent-marking strategies to indicate argument structure (Gzella 2013). Many developments in Aramaic are paralleled by those in Arabic and Hebrew (Rendsburg 1991). This drift continued in Neo-Aramaic, but assumed greater proportions in NENA. The most drastic innovation is its verbal system (cf. Pennacchietti 1988; Hoberman 1989; Polotsky 1996; Kapeliuk 1996; Khan 2007a: 12–14). Analytical constructions already on the verge of grammaticalization in Middle Aramaic gave rise to entirely new inflectional paradigms. The loss of the essential ingredients of the West Semitic 1

We would like to indicate that Paul M. Noorlander is responsible for those sections and paragraphs concerning the history of Aramaic and its modern dialects and Donald Stilo for the remaining part of the present article.

ON THE CONVERGENCE OF VERBAL SYSTEMS. PART I

427

verbal system, the suffix- and prefix-conjugations, and their replacement by originally non-finite constructions are unparalleled within modern Semitic (Hopkins 2005). Tsereteli (1979: 57–9) lists 14 TAM categories (plus a full set of revamped passives) for Christian Urmi Neo-Aramaic. For a Semitic language, this is not just a few new paradigms, but a virtual explosion of paradigms! The truly intriguing point—and the main goal of this article—in regard to the innovated verbal system of NENA comes from outside Aramaic: this hugely expanded system of TAM paradigms in NENA is closely mirrored in certain non-Semitic languages of the area that also underwent their own significant transformations and restructurings to arrive at these parallel modern inventories of TAM paradigms. In the core area we find 1) in addition to NENA, 2) aberrant Armenian dialects found from the Urmia area to the Araxes river and beyond in various unconnected pockets in Nakhichevan (Azerbaijan), Armenia, and Mountainous Karabagh, 3) Udi, now spoken in two villages in Azerbaijan and one in Georgia, the southernmost Daghestanian language and the descendent of Caucasian Albanian (CA), a major language of the 5th–8th century Christian East covering virtually all of eastern Transcaucasia (Bosworth 1987: 520), and 4) Northern Tati and some varieties of Talyshi, both members of the Tatic family. In an extended area beyond the core, we see that Standard Armenian, Azerbaijani, and Caucasian Tat (CT) also participate to a lesser degree in this convergence of verbal systems. Persian, the varieties of Kurdish, other Iranian languages, and Georgian also have marginal participation in only some of the stages presented below. We will refer to the area incorporating all the above-mentioned languages as the Araxes-Iran Linguistic Area (AILA), roughly including Georgia, Armenia, Azerbaijan, western and central Iran, northern Iraq and eastern Turkey. While there are clear areal subdivisions within the larger area, many phonological, grammatical, and lexical isoglosses unite the whole zone (Stilo: in progress). While the core group, often including CT, are at present found in very disparate areas (western Iran, northern Iraq, Daghestan), they would have had a different, and larger, distribution some 1000–1500 years ago. The various groups very likely underwent population movements and language shifts due to political upheavals, the Arab and Mongol invasions, and religious persecution (even by former co-religionists in cases of conversion of some communities to Islam). In brief, this article is really not about the evolution of Aramaic per se, but focuses rather on the shared features that we find throughout the whole area and especially the dialectal varieties whose verbal systems are in many respects now parallel to those of Aramaic. These parallel restructurings may have occurred at a historical point before the massive contact with Kurdish. We must also not exclude the potential for Aramaic to have been a major contributor, rather than just a receiver.

428

PAUL M. NOORLANDER AND DONALD STILO

We are only hoping to make a preliminary statement by offering some intriguing proposals that merit fuller investigation. Due to space limitations, only a brief survey is possible for now. The relevant South-West Iranian (SWI) languages include Persian and CT, a variety of Persian resettled to the south Caucasus from southern Iran ±1500 years ago. Spoken in the Republic of Azerbaijan, CT has Muslim and Jewish (Juhuri) varieties; Juhuri is also found across the border in Daghestan, Russia. The pertinent North-West Iranian (NWI) groups include Tatic (Tati-Talyshi), Caspian (Mazanderani, Gilaki), Central Dialects, (North, Central, South) Kurdish, Zazaki, and Hawrami. Tatic includes 1) Talyshi along the west Caspian littoral in Iran northward into Azerbaijan and 2) the Tati group found in sporadic pockets from Northern Tati along the Araxes in Iran and Nakhichevan to the Saveh area in Iran. The names Tati (NWI) and Caucasian Tat (SWI) may seem misleading at first as Tat is a Turkic word referring to various linguistic groups in the sense of ‘non-Turks’. Speakers of these languages now also use Tat/Tati themselves for lack of a better blanket term. Tatic and CT are very different, mutually unintelligible languages and the similarity in their names is only coincidental.

2. TYPOLOGICAL EVOLUTION OF THE VERBAL SYSTEMS

We would like to show how the languages of different families in the core group gradually underwent similar innovations, shifts, and obsolescence in their verbal systems, eventually developing strikingly analogous systems. The languages of the extended group participated to some degree as well. In Part I we will concentrate on the Present tense, the Present Progressive, the Subjunctive, and the Future tense. The past equivalents of these forms will be addressed in Part II, which follows below. We will start with the most basic Present tense formations and then gradually introduce other paradigms by moving through the stages that NENA, Iranian, Armenian, and CA underwent leading to the formation of new, usually more complex verbal paradigms. The languages of the core and extended areas originally exhibited quite different verbal systems, ranging from very simple to tense-mood inventories of varying sizes and richness. Then, as we propose, in various stages of convergent evolution due to diffusion, language shifts, and multilingualism in contact situations (often on multiple fronts), these languages over time acquired rather large, parallel inventories. The exact timing of these stages is not completely clear and may, in fact, be unknowable. The following presentation is organized according to typological rather than strictly diachronic criteria. Despite the exceptionally long continuously documented

ON THE CONVERGENCE OF VERBAL SYSTEMS. PART I

429

history of Aramaic, the direct ancestors of the Eastern Neo-Aramaic (ENA) dialects are thus far unattested. None of the earlier literary languages could be considered to fulfill this role but they may obviously shed some light on Proto-ENA, being closely related to it. We should not be surprised that spoken Aramaic varieties have a history of their own parallel to, though more or less independent of, their written counterparts. In other words, it is not known when the ENA dialects underwent the stages below to develop their own unique shape. Likewise, most Iranian languages (other than Middle Persian and Early New Persian), Armenian dialects, and Udi (except for the 7th century texts) are lacking any long-term diachronic data and are only known from the relatively recent past. Thus there is no secure way of dating the changes or the directions of the diffusion and expansion of isoglosses in the verbal systems of these languages.

2.2. The Original Areal Diversity of Verbal Systems Stage 1: The base stage In the base stage, the four families we are presently examining originally each formed the Present (or Imperfective) according to completely different structural patterns. The Subjunctive may or may not have had an independent typology, depending on the language.

Iranian

In Middle Persian (MP) and Parthian, the Present Indicative and Subjunctive had contrasting Person-Agreement Markers (PAMs) but no dedicated TAM markers. These paradigms were organized into a completely different system from any modern languages of the area: Parthian

Pres. Indicative

Pres. Subjunctive

2SG

kun-ā

1SG

kun-ām, -am

3SG

kun-ēd

kun-ē

Stem + IND-PAMs

kun-ān

kun-ā-(δ)

Stem + SUBJ-PAMs

‘to do/make’ (Skjaervø 2009: 219)

Classical Armenian (Grabar) The Classical Armenian Present Indicative had no TAM but the Present Subjunctive did. Both forms shared identical PAMs:

430

PAUL M. NOORLANDER AND DONALD STILO Present

Subjunctive

1SG

sir-em

sir-ec‘-em

3SG

sir-ē

sir-ec‘-ē

2SG

sir-es

sir-ec‘-es

‘I love’ (Gasparyan 2000: 42–43)

Aramaic Aramaic originally had yet another system different from both Iranian and Armenian. Older Aramaic, as other West-Semitic languages, mainly indicated the Present by means of what is traditionally known as the Imperfect or prefix conjugation (yaqṭul) vs. the Perfect or suffix-conjugation (qaṭal) generally conveying completion in the past. For an example of this prefix conjugation, see Table 1 below in Stage 2.

Udi, Caucasian Albanian, and Daghestanian

The reconstruction of the proto-NE Caucasian (Nakh-Daghestanian) verbal system is based on modern Nakh (Chechen-Ingush-Batsbi) and Daghestanian languages. Verbs typically lack person agreement, but coindex the class/gender of the head noun, which is the noun in the Absolutive case, i.e. Subjects of intransitives, Patients of transitives. NE Caucasian noun classes, depending on the language, range from two to eight (Gudava 1986: 739). Avar (Daghestanian), for example, has three classes, two of which by and large correspond to natural gender: class I = human masculine and Class II = human feminine (van den Berg 2005: 156). Thus the verb coindexes the gender but not the person of the head noun or pronoun: (1) Avar (Daghestanian) w-ač̣-ana CLI-come-PRET ‘I (m.)/ you (sg.m.)/ he came’. (2) emen father

w-ač̣-ana CLI-come-PRET

‘Father came’. (3) y-ač̣-ana CLII-come-PRET ‘I (f.)/ you (sg.f)/ she came’.

ON THE CONVERGENCE OF VERBAL SYSTEMS. PART I

431

(4) ebel y-ač̣-ana mother CLII-come-PRET ‘Mother came’. (5) ču horse

b-ač̣-ana CLIII-come-PRET

‘A horse came’. (examples (1), (3): van den Berg 2005: 166; (2), (4), (5): Gudava 1986: 739) The only old attestation we have for Daghestanian consists of a recent discovery of a 7th century palimpsest of major parts of the New Testament in CA, which clearly shows that this language is an older form of modern-day Udi (Gippert et al. 2008). This language had four different verbal stems, each with a thematic vowel. The thematic vowel of the Present was -a, but this stem was used for both the Present tense and the Imperative (Gippert et al. 2008 II: 44). It does not seem to be a dedicated TAM. Maysak (2008b: 208, ft. 176) calls it a marker of the Imperfective aspect, harking back to a proto-Lezgian ablaut that served as one of the means of encoding aspect. (6) Present

Subjunctive

owḳ-a=z say-TV=1SG

ba-al=anḳe=zow do.PRS-PRS.PTCP=CONJ=1SG

‘I say’

‘so that I may do’

(Gippert et al. 2008, II: 48)

(Gippert et al. 2008, II: 50)

Stage 2: Innovation I. AILA languages develop ∅-STEM-PAM

In Stage 2 the Present tense typologies of the various families as described in Stage 1 are brought closer to each other through two innovations: 1) the creation of a ∅-marked Present form, where not already existing, which often includes Subjunctive usages but may also contrast with the Subjunctive, and 2) the innovation of suffixed PAMs in early NENA and early CA.2 In a parallel fashion Early Judaeo-Persian (EJP) and Early New Persian (ENP) lose the distinction between Present and Subjunctive PAMs (but see Stage 4 below).

2

Since Caucasian Albanian person agreement marking was already in place in the 7th century manuscript, it would have appeared prior to that time.

432

PAUL M. NOORLANDER AND DONALD STILO

Thus the four families (or at least some varieties in each) ultimately all converge into a parallel typological Present (±Present Subjunctive) characterized by a lack of TAMs but with one set of PAMs used universally for all the Present-based tenses within each group.

Aramaic The original Aramaic system soon becomes areally consistent with Iranian, Armenian, and CA in Stage 2 by undergoing two well-known paradigmatic shifts plus the innovation of new PAMs. First, the typically West Semitic system in Aramaic discussed above presumably becomes gradually obsolete in early ENA, while a new structure emerges. Late Eastern Aramaic dialects start out with a continuation of the West Semitic Imperfect (yešqol) of Stage 1 with a distinctive n/l- in the 3rd person masculine prefix (cf. Table 1, columns 1 and 2). The functional range of the Imperfect, which used to include the general present, is gradually narrowed down by a new progressive, which we call here—for consistency with other families at this stage—a ∅-marked Present Indicative (Table 1, column 3). The Imperfect is now practically a Subjunctive (cf. Jewish Babylonian Aramaic [JBA] le-šqol ‘let him take, (that) he may take’), being the basic finite form in modal complements. It also generally expresses futurity, e.g. Classical Syriac (CS) ʾe-šqol ‘I will take’, a tense well-known to be cross-linguistically connected with deontic modality. The newly created Present Indicative (∅-šaqel=nā ‘I (sg) take, am taking’) grammaticalized on the basis of the active (or imperfective) participle šāqel- in the absolute state declined for gender and number like an adjective (respectively šāqel [sg.m.], šāql-ā [sg.f.], šāql-īn [pl.m.], šāql-ān [pl.f.]) combined with a personal enclitic (see column 3): Table 1. The outcome of Stage 2 in Classical Syriac

1SG 2SG.M 2SG.F 3SG.M 3SG.F

(Stage 1) > BASE > Present > ʾe-šqol > te-šqol > te-šql-in > ye-šqol > te-šqol, etc.>

(Stage 2) SHIFT Subjunctive ʾe-šqol te-šqol te-šql-in ne-šqol te-šqol, etc.

INNOVATION Present šāqel=nā šāql=att šāql=att (~ šaql-ā=att) šāqel šāql-ā, etc.

š.q.l. ‘to take’ In Aramaic, “the extension of the use of the active participle in the verbal system is accompanied by its acquisition of verbal properties, including, most conspicuously, its inflection for person and number” (Khan 2007b: 86). Indeed, this verbal

ON THE CONVERGENCE OF VERBAL SYSTEMS. PART I

433

adjective, šāqel-, employed predicatively as in non-verbal clauses, (see Nöldeke 1904: 45, §64) had become a major integral part penetrating the verbal system to express mainly the Indicative counterpart to the present Subjunctive ne-šqol (earlier *ye-šqol). The diachronic changes in the Neo-Aramaic past and non-past verbal paradigms mainly involved innovations organized around the development of analytic formations using infinitives and participles as well as a new set of PAMs out of formerly enclitic pronouns (see Table 2 below). In so doing, Late Aramaic now converges areally with Middle Iranian and the oldest forms of ENP and EJP, Classical Armenian, and CA, all of which already lacked a TAM in the Present.

2.3. New PAMs in Early NENA and Caucasian Albanian

In a process widely found cross-linguistically, both NENA and CA developed the new PAMs alluded to above from grammaticalized and phonetically reduced forms of full personal pronouns. This exact process is not directly attested for NENA but welldocumented data for other varieties of Late Eastern Aramaic (CS, JBA) serve as indirect evidence for the situation of the ancestor of Proto-NENA. The PAMs are already fully grammaticalized in NENA and are no longer enclitics but suffixes (see Stages 3, 4, 6). The actual phonetic shapes of the Syriac enclitics were formed by extracting different phonetic material from the full-form pronouns from that which was extracted in (proto-)NENA, e.g. CS ḥnan vs. NENA ʾaxnan ‘we (full form)’ > CS =nan vs NENA =ax ‘1st plural PAM’, as we see in Table 2: Table 2. Innovation of PAMs Caucasian Albanian

1SG 2SG 3SG 1PL 2PL 3PL

Full pron zu vown žan vʕan

Enclitic PAM =z(u) =nown =ne* =žan =nan =ne*

pres.ind ‘to search’ bes-a=z bes-a=nown bes-a=ne bes-a=žan bes-a=nan bes-a=ne

Pre-Proto-NENA

1SM 1SF 2SM 2SF 1PL 2PL

Full Enclitic pron PAM *ʾana *=na *ʾana *=na *ʾat *=at *ʾat *=at *ʾaxnan *=ax *ʾaxtun *=tun

C. Urmi Present ‘to open’ patx-in patx-a-n patx-it patx-a-t patx-ax patx-iy-tun

*(derived from a focus particle =ni, also spreading to the 2nd person SG/PL =nown/=nan) (Gippert et al. 2008, II: 54)

Note the following quotes about CA and early NENA: In contrast to modern Udi, “...agreement clitics are strongly bound to verbal stems in Caucasian Albanian, as long as they do not have the copular function.” (Gippert et al. 2008, II: 52–53) Likewise, “In NENA, the enclitic pronouns have become bound subject-markers.

434

PAUL M. NOORLANDER AND DONALD STILO

They cannot be separated from the verbal base and must be repeated for every verb.” (Coghill 1999). As for copular usage, at a later stage NENA innovated new enclitic PAMs to become the analytical copula as in other languages of the area. In certain paradigms not only do the new PAMs eventually develop into enclitics in their use as an AUX (Stage 5) but they are also mobile in these same paradigms (Part II, below). This mobility is closely mirrored in the other core group languages (Stilo 2008). The ∅-marking strategy for the Present of Stage 2 eventually either becomes obsolete in almost all these languages or shifts to another function. Very few modern languages of the Araxes area still use the ∅-marking typology in the Present tense: Table 3. ∅-marked Present tenses Mod. North-West Iranian

W. Gilaki Southern Koluri (inter alia) (Caspian) Talyshi (Central Tati) ∅-dān-ǝm ∅-zon-ǝm ∅-zān-ǝm ‘I know’ (contrasting with Subjunctive)

Mod. Eastern Aramaic

C. Ashitha, Ṭyare3 (Borghero 2006: 128) ∅-az-in ‘I go’ (Indic. = Subj.)

Thus with the creation of a ∅-marked Present in early NENA, and the development of one set of universal PAMs in ENP/EJP (presumably also undocumented NWI languages), CA, and NENA, we now find a fairly uniform typology of Present tense formation in the early stages of these languages in the extended AILA area. Once all four groups are typologically lined up here in Stage 2, they are then all ready to follow parallel paths in Stages 3–7 below. Stages 3 and 4, next, do not necessarily occur in a specific order but most likely appeared interactively and perhaps simultaneously.

Stage 3: Innovation II. The Present Indicative acquires a TAM

The next link in the chain of events gradually leading to typologically near-identical verbal paradigms in the modern languages is the innovation of Present TAMs. These TAMs may derive from verbal or adverbial (especially temporal) categories, but often their origin is simply unclear. In the process of grammaticalization and depending on the language, the new TAMs may appear before or after the verb stem or even alternate between both positions. The TAM still has this same variation in some of the modern AILA languages, as shown below.

3

In theory, however, the Ṭyare dialects could also have entirely lost the TAM of the next stage (see Stage 3).

ON THE CONVERGENCE OF VERBAL SYSTEMS. PART I

435

Aramaic

Unlike Classical Syriac, other varieties of Eastern Aramaic, such as Jewish Babylonian Aramaic and Classical Mandaic, expanded the same participial construction as sketched above with the preverbal particle qā and its variants (for example qāʾe and kā, Sokoloff 2002: 549a, 976–977), an eroded form of the active participle qāʾem ‘standing’ of the root q-w-m ‘to stand’. Postural verbs tend to grammaticalize into markers of continuous aspect or copulas cross-linguistically (Heine and Kuteva 2002: 280–282). Likewise, this durative preverb was semantically and structurally dependent upon the active participle, to which it may even be prefixed in the case of high frequency verbs, such as qā-sāḇar ‘he opines’, even contracted q-āmar ‘he says’ (Morgenstern 2011: 172–173). In all likelihood, qā either eroded or dissimilated (Gzella 2006: 184, ft. 3) to k- in Proto-NENA and was applied to all verbs. The above would be the most likely ultimate origin of the main Present Indicative TAMs found in ENA languages (NENA k(i)-, Ṭuroyo ko-, Neo-Mandaic qa-4), which would mean that, at least in this respect, Classical Syriac is more distantly related to them in lacking such an incipient TAM.5 Moreover, it cannot be excluded that Arabic influence at least played a partial role, where phonetically similar kū (< yakūnu or kūwe) exhibits similar behaviour (Jastrow and Fischer 1980: 154). (See also W. Armenian below for an analogous TAM derived from ‘stand’.) In many NENA varieties, this k(V)- prefix was gradually lost except for a small subset of about a dozen verb roots (either with initial /ʾ/ or /h/ or an irregular verb), e.g. Jewish Urmi Indicative g-de ‘they come’ vs. Subjunctive ∅-ade ‘let them come’ (see Khan 2008b). On the other hand, in some dialects, such as Christian Urmi, the ki- Present Indicative TAM marker is still consistent throughout: Indicative ki-xaz-in ‘I see (sg.m.)’ vs. Subjunctive ∅-xaz-in ‘(that) I (sg.m.) may see’.

Iranian

Except for a handful of languages (see Table 3), modern Iranian languages of the extended AILA area eventually developed TAMs of adverbial origin (Windfuhr 2009: 26), the most widespread of which are mV-, stressed in Modern Persian but unstressed in most other languages, and æ(t)-, both of which are actually durative

See Nöldeke (1868: 294, 1875: §261); Breuer (1997); Heinrichs (2002); Gzella (2006); Khan (2007b: 91 and elsewhere). 5 Talay (2008: 173) suggests an alternative source, however: the conjunction kaḏ ‘when, while’, which could also be combined with the participle in a circumstantial progressive or gerundive sense, but this would fail to explain Neo-Mandaic qa-. 4

436

PAUL M. NOORLANDER AND DONALD STILO

markers, as they are also used in the Imperfect. Examples in modern Iranian languages: (7) mVModern Persian mí-dunæm Talvela (Hawrami, Iraq) m-sanu Kafteji (C. Tati) me-gæneǽ

‘I know’ ‘I buy’ ‘she falls (3SG.F)’

(8) æ(t)Vafsi (S. Tati) Lori

æt-aróm i-δunóm

‘I bring’, æd-oróm ‘I eat’ ‘I know’

(9) æ(t)-? Lachin Kurmanji Zakho Kurmanji Suleymanie (C. Kurd.)

dæ-xwazɪm t-bôr-it a-xwā-t

‘I want’ (Bakaev 1965: 162) ‘it passes’ (MacKenzie 1962, 2:356) ‘s/he eats’ (McCarus 2009: 609)

While the seeds for the later TAMs (durative) mi ~ mē and (punctual) be- (see Stage 4, below) were already planted in Middle Iranian, they do not serve as dedicated TAMs at this stage nor even in the very earliest forms of New Persian: “In Middle Persian a pair of new pre-verbs were gradually introduced, probably in order to express resultativity and durativity, respectively: the former the previously mentioned adverb bē (BR’) ‘away’ and the latter the adverb hamē(v) ‘always’. ... The use of these pre-verbs was originally quite free but was gradually grammaticalized during later centuries...” (Utas 2013: 257)

The very earliest EJP texts, for example an 8 th century EJP letter discovered at Dandān Uiliq near Khotan in modern Xinjiang province, China, show us that TAMs are not used in the verbal domain. Paul (2013: 114) states that “The absence of (ha)mē from D[andān Uiliq]1/2 is in agreement with the early date of these texts, when (ha)mē was still far away from being grammaticalized.” In later EJP texts the particle hamē “is typically used to characterize an action or state as durative, iterative, or progressive (...) It is not yet fully grammaticalized, and although it usually precedes the verb immediately, its position in the sentence is variable in some texts (...) [and it] sometimes retain[s] its MP meaning “always” (Paul 2013: 115). In ENP, i.e. the earliest forms of New Persian written in the Arabic script but whose significant prose texts begin some two centuries later than the earliest EJP texts, hamē was still mobile and could also appear somewhat removed from the verb (Lazard 1963:274). Examples:

ON THE CONVERGENCE OF VERBAL SYSTEMS. PART I (10) buxār steam

bar up

āy-aδ come-3SG

437

hamē PTC

‘The steam rises’ (Lazard 1963: 275-276). (11) hamē ptc

amīrī emirate

i

EZ

Madīna Madina

dih-am give-1SG

‘I will give (them) the Emirate of Madina’. The particle hamē still carries the sense of durative, iterative, or progressive here as in EJP but seemingly becomes obligatory when the verb has a progressive sense (Lazard 1963: 297). In other types of duratives, however, it is optional. Additionally, as we see in (11), in these intervening two centuries hamē has acquired a slightly broader range of functions in ENP, even including immanent future (Lazard 1963: 285). Thus a step has clearly been made towards its grammaticalization as a TAM, even though it also “still behaves like a kind of adverb in some rather rare and special usages ...” (Lazard 1963: 297). In later stages of Modern Persian, of course, it became a fully grammaticalized prefix, mi- ~ mē-. The three positions (prefix, suffix, alternating) for the TAM *a(t)/e(t) are still found in different Central Dialects (NWI): (prefix) Ārāni æ-vin-õ ‘I see’; (suffix) Esfahani Jewish veroθ-un-e ‘I get up’; (alternating) Yazdi Zoroastrian e-væj-i ‘you say’ vs. næ-væj-i-e ‘you don’t say’ (Stilo 2007b: 106–108). (See also (13), Erzerum dialect of Armenian for another alternating type). In addition, Windfuhr (2009: 26) identifies dV-~ t- Present/Durative TAM of Northern Kurdish and the a(t)- of Central Kurdish with this affix. As in Aramaic, some NWI Iranian languages also base their Present Indicative on a present participle: Caspian dialects, some varieties of the Semnan group, and Zazaki/Dimli of eastern Turkey (a transplant from the Caspian area). The Old Iranian participial formant *-ant becomes phonetically reduced via a process of grammaticalization (-and > -nn > -n), yielding a true TAM, e.g., Zazaki gir-en-o (sg.m.) ‘he takes’, Babolsari (Mazanderani, Caspian) gir-n-ɛ, Lahijani (Gilaki, Caspian) gi-n-ə (root: -gir-), Aftari (Semnan area, Borjian 2008) gi-nn-e (root: -gir-). This suffix only occurs in the Present Indicative; the paradigms of the Past system (Imperfect, Past Progressive) are formed according to other strategies. Nothing is known about the concrete evolution or timing of this pattern.

Western Armenian

The Present TAM gə ~ ḳə of Western Armenian—actually a durative marker—is also found in most of its dialects and is one of the principal diagnostics for distinguishing Western from Eastern Armenian. In Standard Western Armenian it is an immobile prefix, gə-məna-m ‘I stay’ and now also contrasts with the ∅-marked Subjunctive. In

438

PAUL M. NOORLANDER AND DONALD STILO

some Western dialects gə~ḳə commonly follows the verb or, as in the Karin (Erzerum) dialect, freely alternates as preposed or postposed to the verb root, as in (13) below under Stage 4. In a process parallel to NENA qāʾem > kV- above, Armenian also derives this gə~ḳə Present marker from the postural verb ḳay- ‘stand’. The latter root grammaticalizes in two important directions: first, it is evolves into the Existence verb, ḳa; second, it becomes phonetically reduced to the W. Armenian durative marker gə~ḳə (Donabedian 2008). For its further development in Eastern Armenian, see Stage 7.

Udi

The Caucasian Albanian Present tense, as presented at the end of Stage 1 above, eventually shifts in Udi to a Subjunctive paradigm (see next stage) and a new Present is formed (see Stage 5).

Stage 4: Shift I. The original ∅-marked Present shifts to Subjunctive

As a result of the appearance of an overt Present/Durative TAM (Stage 3 or 5, below) and the use of the ∅-marked form for the Subjunctive paradigm (this stage), the Present Indicative and the Present Subjunctive forms now contrast in the languages that reach this stage. This is not so much a shift as it is a restricting or narrowing of the domain of the ∅-marked form. As TAMs were introduced to mark durative uses, the domain of the Present, which also included subjunctive senses, narrowed down to a new, independent Subjunctive. See also Matras (2009: 260) for a table of the TAM-marked Present (durative) vs. the ∅-marked Subjunctive in this area, although the languages included there and in our study do not completely overlap.

Aramaic

The ∅ TAM Present Indicative (∅-šaqəl) of Stage 2 developed into the new Subjunctive and the innovated equivalent with a preverbal TAM element (k-šaqəl) became the new Present Indicative. This provided a symmetrical distribution of the qā-based TAMs, its presence or absence denoting respectively indicative or non-indicative modality (cf. Khan 2007b: 86): Indicative

Non-indicative

Step 1 šāqel

ne-šqol

Step 2 šaqəl

šaqəl

Step 3 k-šaqəl

∅-šaqəl

ON THE CONVERGENCE OF VERBAL SYSTEMS. PART I

439

Iranian

In EJP and ENP the distinct Indicative vs. Subjunctive PAMs of Middle Iranian (Stage 1 above) are leveled out, except for the occasional use of a special 3rd singular form (Paul 2013: 122). In many Iranian languages the Subjunctive eventually acquires the overt prefix be in contrast to—and most likely along with or after the grammaticalization of—the durative marker (ha)mē, a tendency already apparent in Middle Iranian. “EJP be is grammaticalized to a lesser degree than EJP (ha)mē” (Paul 2013: 115). In the older stages, be mostly had a limitative or perfective sense (Lazard 1963: 298) and was used in the Subjunctive, Imperative, and Preterite, as is still the case in many Iranian languages. In EJP and ENP it occasionally even had durative/iterative (Lazard 1963: 299) or future (Paul 2013: 121) senses, hence its uneven process of grammaticalization in the early stages where we cannot yet detect any “important difference in the use of bi-...” (Lazard 1963: 299). While the morpheme bé- is always stressed, the Subjunctive be- and the initial stress should be seen as separate phenomena since the stress still moves leftward to the first syllable of the verb even in those situations where there is no be-. The only modern Iranian languages of the area still at the older stage of no overt Subjunctive TAM, at least for most verbs, are CT (both Muslim Tat and Juhuri), Razajerdi (a Tati language to the north-east of Qazvin), Keringāni and Harzani (N. Tati varieties along the Araxes river in Iran), where the Subjunctive is only marked by initial stress, e.g. Razajerdi (IND) me-ruš-ém > (subj.) ∅-rúš-em ‘I sell’ and Juhuri čü ∅-gúy-um? ‘What should I say?’ Even in Persian, this innovated bé- generally does not appear in many compound verbs or after adverbial preverbs, thus leaving some traces of the basic Stage 4 typology in these verbs as well: the Present is TAM-marked (mí-) and the Subjunctive is ∅-marked.

Armenian

The Classical Armenian Subjunctive becomes obsolete and the former ∅-marked Present Indicative shifts to take its place at some point in the post-classical period of the language, paralleling the same shift in other languages at this stage. The Western Armenian durative TAM gə ~ ḳə now contrasts with the nul-marked Subjunctive, ∅-hasn-i, e.g. in the Karin (Erzerum) dialect as mentioned above:

440

PAUL M. NOORLANDER AND DONALD STILO

(13) Karin dialect Indicative

Indicative

Subjunctive

hasni-ḳə arrive-PRS

ḳə-ləvas-v-i PRS-wash-REFL-3SG

thoγ ∅-hasn-i. let SUBJ-arrive-3SG

ləvas-v-i-ḳə wash-REFL-3SG-PRS ‘He arrives and washes himself’. ‘He washes himself’.

‘May he arrive’.

(Mkrtch‘yan 1968: 502, 505)

Udi

As mentioned, the CA Present tense, as presented at the end of Stage 1 above, eventually shifts to a Subjunctive paradigm and a new synthetic Present is formed, which most likely then shifted to a Future (see Stage 6). Maysak (2008b: 210) states that the Udi Subjunctive “...was historically most likely a present form of the indicative.” (14)

CA: Present

>

Modern Udi: Subjunctive

owk-a=z ‘I say’

>

uk:-a=z ‘(that) I say’

(Gippert et al. 2008, II: 48)

(based on Maysak 2008b: 163)

Maysak (2008b, 212) states that such a ‘restructuring most likely took place as a result of contact with those languages in which the finite strategy is quite common—particularly under Indo-European and Turkic influence”. The primary languages Maysak refers to are Persian, Armenian and Caucasian Tat, since “in all of them the older form of the present tense is now preserved in such restricted contexts as dependent [forms] in modal predicates, certain imperatives, or optative clauses, etc.” We cannot rule out the possibility that some aspects of the Udi Subjunctive “... could have appeared as a result of the calquing of the corresponding usages and the functionally similar forms in the other languages” (Maysak 2008b: 212). Any putative Turkic influence in this case, however, will depend on whether these shifts occurred prior to or after the 11 th century influx of the Oğuz/Seljuks. Two additional factors must be considered regarding convergence features in the subjunctive. Firstly, counter to the other languages mentioned, the Azeri Subjunctive paradigm is not ∅-marked, e.g. ged-æ-sæn/al-a-san ‘(that) you go/take’. Secondly and more importantly, the Subjunctive plays a much more prominent role in colloquial Azeri, even as spoken in the Republic of Azerbaijan, than we see in written styles or formal

ON THE CONVERGENCE OF VERBAL SYSTEMS. PART I

441

grammars of the language. This is clearly an influence of the substratal Iranian languages of Azerbaijan (see also Stilo in press and forthcoming).

Stage 5: Innovation III. A second Present tense, analytic formation Another Present tense, usually a progressive, is innovated periphrastically based on non-finite forms, particularly the infinitive,6 with a clitic or affixed locative or allative marker plus an enclitic copula. Such periphrastic progressives are well known cross-linguistically, e.g. colloquial Dutch ik ben aan ’t eten ‘I am eating’ and English I am going < I am a-going < I am at going (itself possibly a copy from Celtic). In some cases this paradigm is the sole Present tense instead of the more restricted Progressive (see Table 4 below). This formation is an areal isogloss of the northern AILA zone: (Progressive) Caspian group, colloquial Azerbaijani,7 colloquial Georgian, and a few varieties of NENA; (sole Present) N. and C. Talyshi, N. Tati, CT, and Udi. Eastern Armenian and many of its dialects have both a Present and a Present Progressive of this type. Other Iranian languages such as Kurmanji, C. Talyshi, and Persian innovate Progressive paradigms of various typologies not related to the analytic infinitival formation as shown here.

Aramaic and Iranian In several NENA dialects, including Christian Urmi, and in Asālemi (C. Talyshi), Harzani (N. Tati), and Lahij (Lahıc) dialect of (Muslim) Caucasian Tat the analytic paradigms consist of a Locative or Allative preposition + Infinitive + copula. (15) Christian Urmi Aramaic (Progressive) bi-qraya=vin

+

LOC-read.INF=COP.1SG

‘I am reading’

Asālemi C. Talyshi

Lahij (Lahıc) Caucasian Tat

(Simple Present) bæ-gæt-é=im* ALLT-take-INF=COP.1SG

bæ-xund-ǽn-um ALLT/LOC-read-INF-COP.1SG

‘I take’ *(pron: bægætím)

‘I am reading’ (Hüseynova 2002: 106)

As the Iranian infinitive is usually formed on the past root of the verb, e.g. Persian foruš/foruxt->foruxtæn ‘to sell (INF)’, these periphrastic forms may seem to be a past formation but in fact the infinitive is tenseless and tense is encoded in the AUX, lit. I to.sell-in=am/was ~ I to-to.sell=am/was, see (15)-(17). 7 This periphrastic progressive is possibly a feature of Azeri as spoken in the Republic of Azerbaijan. I (DS) have never really encountered it in Iranian Azeri or in collections of naturalistic spoken texts, e.g. Kıral (2001). 6

442

PAUL M. NOORLANDER AND DONALD STILO

N. Talyshi has a similar typology but with a Locative postposition, -dæ. In Gilaki the AUX used for this construction is actually a locational copula ‘to be in’, formed with the postposition *dar, no longer used in the language. Both -dæ and dar derive from older Iranian *andar ‘in’. (16) Leriki N. Talyshi

Rashti W. Gilaki

(sole Present) gæt-é-dæ=m take-INF-LOC=COP.1SG

Lahijani E. Gilaki

(Present Progressive) gift-ə́n=dər-əm take-INF=LOC.be-1SG

‘I take’

git-é=der-əm take-INF=LOC.be-1SG

‘I am taking’

In an unusual situation, Northern Talyshi and Harzani use both types of periphrastic constructions, but neither language now has a Stage 3-type Present built on the present stem. Northern Talyshi uses the Locative for the Present (16) and the Allative for the Future (Table 4) while Harzani uses them for the Present and Progressive, respectively: (17) Harzani (N. Tati) (Simple Present)

(Progressive)

bö-ot=en ALLT-say.PST=COP.1SG

öt-d-en say.PAST-LOC-COP.1SG

‘I sleep’

‘I am saying’ (Mortazavi 1962: 481)

In a third formation with similar typology, Jewish Urmi (as well as other NENA dialects) and Kalāsuri (N. Tati) delete the Locative/Allative preposition of this construction and reduce the structure to only infinitive + copula, which can probably be interpreted as one additional degree of grammaticalization. (18) Infinitive + copula in NENA and Tati (Present Progressive) Jewish Urmi Aramaic dúgl-e dagol-èt lie-PL lie.INF-COP.2SG.M ‘you are lying’ (Khan 2008b: 402, 35)

ON THE CONVERGENCE OF VERBAL SYSTEMS. PART I

443

Kalāsuri (N. Tati) olæt=em šušt-e clothes=COP.1SG wash-INF ‘I am washing clothes’ (Yarshater 2005: 278). (Present) Juhuri/Jewish Caucasian Tat jǽʕ fish

gürd-ɛ n ̀ -üm take-INF-COP.1SG

‘I catch fish’. The next Present tense type takes grammaticalization of the above structure yet one step further by adding the well-known Iranian TAM mi- into the equation: mV+ infinitive + copula. (19) Daγkušču (Dağkuşçu) Dialect, Muslim Caucasian Tat mæn I

raft-ǽn go-INF

mu-xast-ǽn-üm DUR-want-INF-COP.1SG

‘I want to go’. (Grjunberg 1963: 132, 234)

Udi The modern Udi Present tense marker -sa is of locative origin based on the fusion of two morphemes: the infinitive in the Dative case, e.g., CA owk-e-s-a = eatTV- INF-DAT > Udi uk-sa=ne ‘s/he eats (variant form)’ (see Gippert et al. 2008, II: 47 and Maysak 2008b: 168-171, respectively). The Udi Dative derives from a protoLezgian inessive and, among other usages, marks locatives and allatives in the modern language. Udi also has an additional periphrastic paradigm formed with -sax (an infinitive + an alternate Dative form) that is formally close to the Present marked by -sa and is used to encode progressives but is quite rare in speech (Maysak 2008b: 168).

Armenian The Eastern Armenian Present is formed with what is traditionally called the Present Participle (Gasparyan 2002: 53), formed from the present root of the verb and the suffix -um. At the same time, however, we cannot ignore the fact that the latter suffix is also the marker of the Locative in nominal paradigms as in (20). Given the Locative-type Present formations of AILA languages, especially in the core area to the north (Georgian), east (Azerbaijani, Udi), and south (Aramaic) of Armenian, we

444

PAUL M. NOORLANDER AND DONALD STILO

cannot discount the possibility that these two markers, Present -um and Locative -um, are in fact identical. This Present tense is an analytic paradigm using the enclitic forms of the copula to mark person: (20) gradaran-um library-LOC

ḳart‘-um=e read-PRES=COP.3SG

‘She reads in the library’. (eanc: OSD fav film 13) As another Stage 5 typology, Eastern Armenian also has an independent Progressive based on a ‘simultaneous participle’ (Gasparyan 2002: 52), also called ‘the processual participle’ (Dum-Tragut 2009: 198–199) with the marker -(e)lis (grelis=em ‘I am writing’). In a few Eastern Armenian dialects, this form has shifted to the sole Present tense, e.g. Artvin (NE Turkey) ert‘-lis=im ‘I go’ (Grigorian 1957: 443). In a parallel fashion to all Armenian dialects, the TAM-marked Present Indicative contrasts with a ∅-marked Subjunctive here as well.

Stage 6: Innovation IV. A Future tense is innovated The existence of a formal Future tense seems to be areally distributed across the northern AILA zone, except for a small isolated pocket in central Iran. The Future is more prevalent in the south Caucasus area (Georgian, Eastern Armenian, Azerbaijani, Talyshi, Northern Tati, Udi) and in the west (Turkish, Western Armenian, Kurmanji), while in the east (Gilaki, some C. Tati, Persian, Qohrudi), this tense is rather marginal.8 The Future has come into being along two paths: either an original Present shifts to a Future (see Stage 7) or a form is innovated with various types of Future TAMs added to the Subjunctive (ultimately also of Present origin) as discussed here. There are three Future paradigms in Standard Eastern Armenian and Udi and two in Azerbaijani. While these tenses are formed by typologically different strategies within each family, they correspond across genetic groups. The following main types are included here: A) an invariable, shortened form of ‘want’ (as in the Balkan Sprachbund) followed by a Subjunctive (see Noorlander forthcoming): Many NENA dialects, in addition to b-~ p- (< g-be-∅ ‘he wants’) use an alternate Future marker bd-~ bt-~ p̣t- “regarded as a phonetically attentuated form of the While languages with a fully integrated Future, e.g. E. Armenian, French, etc. could conceivably have a string of six or more Future-marked verbs in a clause, an equivalent in Persian, where the Future is a feature of the formal written language, would sound certainly questionable and distinctly bizarre. 8

ON THE CONVERGENCE OF VERBAL SYSTEMS. PART I

445

verbal construction *baʿe d- ‘he wants to’, or *bʿe d- ‘it is desired that’ (Nöldeke 1868: 294–296; Pennacchietti 1994a: 281; 1994b: 137; Cohen 1984: 520).” (Khan, 2008a: 175, ft. 3) (21) Christian Barwar Subjunctive

Future

∅-ʾaz-ət SUBJ-go-2SG.M

bṱ-az-a FUT-go-3SG.F

(Khan 2008a: 175) Christian Urmi Subjunctive

Future

∅-ʾaz-in SUBJ-go-1SG.M

p̂ ṱ-az-in FUT-go-1SG.M

Jewish Urmi Subjunctive

Future

∅-od-en SUBJ-do-1SG.M

b-od-en FUT-do-1SG.M

(Garbell 1965: 123) The Lahijani Future parallels the NENA typology, except that the Subjunctive has the overt marker bV- and is impersonal, that is, it lacks a PAM (see Stage 4 above): (22) Lahijani (E. Gilaki) Subjunctive

>

Future

< Present, ‘want’

bú-go-m SUBJ-say-1SG

xa-n bú-go-m want-PRS SUBJ-say-1SG

xa-n-ə want-PRS-3SG

‘(that) I say’

‘I shall say’

‘he wants’

B) in the reverse of type A, this type consists of a conjugated (and possibly shortened) form of ‘want’ followed by an infinitive or other non-finite form of the verb:

446

PAUL M. NOORLANDER AND DONALD STILO

(23) ‘want’-based Future auxiliary Rashti (W. Gilaki, Caspian)

Formal Written Persian

Qohrudi (Central Plateau)

xay-əm gift-ən want-1SG take-INF

xah-æm gereft want-1SG take.INF

kem-ande want-3PL

‘I will take’

gera take.INF

‘they will take’

Kurmanji’s formal Future is encoded with the particle dē plus the Present tense, while Lachin Kurmanji uses the Subjunctive plus the enclitic =æ, hosted only by an overt subject and is thus similar to those stages in which the TAM has not grammaticalized as an affix or clitic to the verb (Bakaev 1965: 71).

Stage 7: Shift II. The Present paradigm shifts to Future

In some languages, an original Present paradigm of Stages 3 and/or 5 shifts to a formal Future tense and an analytical Progressive of Stage 5 then assumes the role of the sole Present tense. Often these shifts contrast with the original situation in other related dialects or languages. While Northern Talyshi only has analytic paradigms for the Present and Future (see Stage 5), in Central Talyshi (Asalemi) the Allative formation remains as the only Present and no Locative Progressive develops. Table 4 shows the unshifted original state (•) in some languages or dialects vs. the shifted equivalent (–) in closely related varieties: Table 4. Shift of Present to Future •

W. Armenian

Present

gə-məna-m

– E. Armenian

>

•

bæ-gætí=m*

*Progressive? >

C. Talyshi

– N. Talyshi

*Progressive? >

>

ḳə-məna-m

mən-um=em

gæt-é-dæ=m

Future

>

bæ-gæté=m*

Meaning

‘I stay’

‘I will stay’ ‘I stay’

‘I take’

‘I will take’ ‘I take’

ON THE CONVERGENCE OF VERBAL SYSTEMS. PART I •

Persian

– Muslim Cauc. Tat

*Progressive9 >

mí-gir-am

>

mægüftǽn=üm

mægürǘm

447

‘I take’

‘I will take’ ‘I take’

(*Some details have been disregarded here for the sake of brevity)

Maisak (2008a: 123) also states that the Udi Future I (the basic Future) with the TAM suffix -al represents a former Present tense “that gradually lost the present sense and was completely displaced towards the domain of future”. In addition, the Future II (Future Potential) with the TAM suffix -o “apparently also arose as a current-ongoing form, which in time allowed it to begin to be used in the capacity of a narrative tense as well (...) [and] fairly quickly became the general present, expressing the full spectrum of present meanings ... but one way or another the “general present” in -o was gradually pushed out of the basic present contexts.” (Maysak 2008b: 186–187) Thus we see that three of the four families have variants affected by these shifts; Aramaic seems to be excluded from this shift as it forms the Future from a TAM of present modal origin (‘want’) prefixed to the Subjunctive, itself a former Present.

3. CONCLUSION

In examining the verbal systems of very different language families of the Araxes-Iran linguistic area, we see that many parallel typological processes can be identified in their evolution. New paradigms appear that are innovations within their respective language families and it is not yet clear which language(s) may have contributed in concrete terms, and at which stages, to this striking convergence. We understand many questions still remain open. We did not intend to, in fact, could not possibly show the exact diachrony or contact scenarios that would have been the conduits for this typological convergence. On the other hand, bringing these phenomena to light can ultimately assist in uncovering the processes and perhaps their sources. Given the complex history of the area and the lack of documentation of language shift due to population movements, the likely sources for these parallelisms may not be so apparent and could indeed come from an unexpected direction, given our present perspective. We are making the assumption here that the periphrastic Present tenses in Eastern Armenian, Northern Talyshi, and Caucasian Tat were originally Progressive forms because of their typology based on non-finite forms + AUX. This is only conjecture on our part and the true diachrony of this putative shift may never be known. 9

448

PAUL M. NOORLANDER AND DONALD STILO

It might at first seem most obvious to attribute the primary source of influence to Iranian languages. While such may have indeed been the case, there is also the likelihood that convergence may also have gone towards Iranian, given that the verbal systems of the core area Iranian languages do not exhibit typically Iranian structures. This points to the fact that these languages have undergone heavy contact phenomena and possibly even language shift from other local languages. The area along the southern banks of the Araxes in Iran is well known to have until recently had a substantial, heavily mixed population of speakers of Northern Tati dialects and local, quite unique Armenian dialects (now relocated to Armenia). The same area would certainly have had a previous Udi and/or Neo-Aramaic presence. Here we find an occasion for both language contact leading to diffusion and likely language shifts. What was the role of certain populations shifting to local Iranian languages along with probably not so uncommon concomitant conversions to Islam of some of these above-mentioned Armenian-, Udi- or Aramaic-speaking Christian and Jewish communities ultimately affecting the structures (and the locations) of these local Iranian languages? What do we actually know about the pre-Kurdish and pre-Turkic contacts of Aramaic with other ‘Christian languages’ (Armenian, Caucasian Albanian, Georgian, possibly local NWI languages) and ‘Jewish languages’ (Juhuri/Caucasian Tat, a myriad of NWI Central dialects, e.g. Esfahani Jewish, Georgian, and possibly Armenian or Armenian dialects)? Perhaps the contact situations for these languages were quite different from today and we may need to look more closely at the mutual interactions of NeoAramaic, Armenian, and Udi/Caucasian Albanian. Local Turkic verbal systems—for now a glaring omission from this discussion—and, given the heavy structural influence of the local non-Persian Iranian (and non-Iranian) substrata on Oğuz varieties west of the Iranian Kavir, the diachrony of Azerbaijani and Turkish verbal paradigms will especially need to be factored into the picture.

REFERENCES

Bakaev, Cherkes Kh. 1965. Yazyk azerbaijanskikh kurdov. Moscow: Nauka. Borghero, Roberta. 2006. The Neo-Aramaic Dialect of Ashitha. Unpublished PhD Dissertation. University of Cambridge. Borjian, Habib, and Maryam Borjian. 2008. “The Last Galesh Herdsman: EthnoLinguistic Materials from South Caspian Rainforests.” Iranian Studies 41 (3): 365–402. Bosworth, Clifford E. 1987. “Arran.” In Encyclopaedia Iranica, II edited by Ehsan Yarshater, 520–522. New York: Bibliotheca Persica Press. Breuer, Yoḥanan. 1997. “The Function of the Particle ‘qā’ in Babylonian Aramaic.” Leshonenu 60: 73–94. [in Hebrew]

ON THE CONVERGENCE OF VERBAL SYSTEMS. PART I

449

Coghill, Eleanor. 1999. The Verbal System of North-East Neo-Aramaic. Unpublished MPhil Dissertation. University of Cambridge. Donabédian-Demopulos, Anaïd and Ouzounian, Agnès. 2008. “Diachronic and Dialectological Variation of Verb Morphology in Armenian: Internal and/or Contact-Induced Changes?” Paper given at Morphological Variation and Change in Languages of the Caucasus, workshop at 13 th International Morphology Meeting, February 5–6, 2008, Vienna. Dum-Tragut, Jasmine 2009. Armenian: Modern Eastern Armenian. AmsterdamPhiladelphia: Benjamins. eanc: Eastern Armenian National Corpus, 2007. [online] http://www.eanc.net/corpus technologies. Garbell, Irene. 1965. The Jewish Neo-Aramaic Dialect of Persian Azerbaijan. The Hague: Mouton. Gasparyan, Gohar. 2000. Das armenische Tempus- und Modussystem: Synchrone und diachrone Analyse. Wiesbaden: Reichert. Gippert, Jost, Schulze, Wolfgang, Aleksidze, Zaza, and Mahé, Jean-Pierre. 2008. The Caucasian Albanian Palimpsests of Mt. Sinai, Vol. I, II. Turnhout (Belgium): Brepols. Grigorian, Aharon V. 1957. Hay barbaṙagitut‘yan dasnt‘ac‘ [A Course in Armenian Dialectology]. Yerevan: Yerevani Hamalsarani Hratarakch‘uty‘un. [in Armenian] Grjunberg, Aleksandr Lenovich. 1963. Jazyk Severoazerbajdžanskix Tatov. Moscow: Izdatel’stvo Akademii Nauk SSSR. Gudava, Togo Evstaf’evič. 1986. “North Caucasian languages.” In Encyplopædia Britannica 15th edition, 738–740. Gzella, Holger. 2004. Tempus, Aspekt und Modalität im Reichsaramäischen. Leiden: Brill. —. 2006. “Zu den Verlaufsformen für die Gegenwart im Aramäischen.” Orientalia 75: 184–188. —. 2013. “Convergence” In Encyclopedia of Hebrew Language and Linguistics, edited by Geoffrey Khan et al., 607–608. Leiden: Brill. Heine, Bernd, and Kuteva, Tania. 2002. World Lexicon of Grammaticalization. Cambridge: Cambridge University Press. Heinrichs, Wolfhart. 2002. “Peculiarities of the Verbal System of Senāya within the Framework of North Eastern Neo-Aramaic (NENA).” In “Sprich doch mit deinen Knechten aramäisch, wir verstehen es!” 60 Beiträge zur Semitistik: Festschrift für Otto Jastrow zum 60. Geburtstag, edited by Werner Arnold and Hartmut Bobzin, 237–268. Wiesbaden: Harrassowitz. Hüseynova, Gülsüm. 2002. Lahıc Tatlarının Dili. Baku: Nurlan.

450

PAUL M. NOORLANDER AND DONALD STILO

Hoberman, Robert. 1989. The Syntax and Semantics of Verb Morphology in Modern Aramaic: A Jewish Dialect of Iraqi Kurdistan. American Oriental Series 69. New Haven: American Oriental Society. Jastrow, Otto and Fischer, Wolfdietrich. 1980. Handbuch der arabische Dialekte. Wiesbaden: Harrassowitz. Jastrow, Otto. 2008. “Old Aramaic and Neo-Aramaic: Some Reflections on Language History.” In Aramaic in its Historical and Linguistic Setting, edited by Holger Gzella and Margaretha L. Folmer, 1–10. Wiesbaden: Harrassowitz. Kapeliuk, Olga. 1996. “Is Modern Hebrew the only ‘Indo-Europeanized’ Semitic Language? And what about Neo-Aramaic?” Israel Oriental Studies 16: 59–70. Khan, Geoffrey. 2007a. “The North Eastern Neo-Aramaic Dialects.” Journal of Semitic Studies 52: 1–20. —. 2007b. “Indicative Markers in North-Eastern Neo-Aramaic.” In XII Incontro Italiano di Linguistica Camito-semitica (Afroasiatica): Atti, edited by Marco Moriggi, 85–97. Soveria Manelli: Rubbettino. —. 2008a. The Neo-Aramaic Dialect of Barwar. Leiden: Brill. —. 2008b. The Jewish Neo-Aramaic Dialect of Urmi. Gorgias Neo-Aramaic Studies 2. Piscataway, NJ: Gorgias Press. Kıral, Filiz. 2001. Das gesprochene Aserbaidschanisch von Iran. Wiesbaden: Harrassowitz. Lazard, Gilbert. 1963. La langue des plus anciens monuments de la prose persane. Paris: Klincksieck. MacKenzie, David Neil. 1961–1962. Kurdish Dialect Studies, Vol 1, 2. London: Oxford University Press. Matras, Yaron. 2009. Language Contact. Cambridge: Cambridge University Press. Maysak, Teymur. 2008a. “Глагольная парадигма удинского языка (ниджский диалект).” In Удинский сборник: грамматика, лексика, история языка. Москва: Academia, 96–161. —. 2008b. “Семантика и происхождение глагольных форм настоящего и будущего времени в удинском языке.” In Удинский сборник: грамматика, лексика, история языка, Москва: Academia, 162–222. McCarus, Ernest N. 2009. “Kurdish.” In The Iranian Languages, edited by Gernot L. Windfuhr, 587–633. London-New York: Routledge. Mkrtch‘yan, M. et al., eds. 1968. Hay žołovrdakan hek‘iat‘ner [Armenian Folk Tales], Vol. IX. Yerevan: Haykakan SSH Gitut‘yunneri Akademiayi Hratarakch‘uty‘un. Morgenstern, Matthew. 2011. Studies in Jewish Babylonian Aramaic: Based upon Early Eastern Manuscripts. Winona Lake, Indiana: Eisenbrowns. Mortazavi, Manuchehr, 1341/1962. “Fe‘l dar zabān-e harzani.” NDATabriz, 14: 453– 488. [in Persian]

ON THE CONVERGENCE OF VERBAL SYSTEMS. PART I

451

Nöldeke, Theodor, 1868. Grammatik der neusyrischen Sprache am Urmia-See und in Kurdistan (reprinted in 1974, Hildesheim), T.O. Weigel: Leipzig. —. 1875. Mandäische Grammatik. Halle: Buchhandlung des Waisenhouses. —. 1904. Compendious Syriac Grammar (translated by J.A. Crichton). London. Noorlander, Paul M. 2013. “Prospective and Avertive Semantics in Eastern NeoAramaic Dialects in Light of Kurdish (and Arabic).” Presentation held at The International Symposium on the Perspective as a Grammatical Category: Evidence from Turkic, Iranian and Beyond, Goethe Universität Frankfurt am Main, 23–25 September 2013. —. forthcoming. “To the Future and Back Again: The Case of North Eastern NeoAramaic “wants”.” Presentation to held at The 42nd Annual Meeting of the North American Conference on Afroasiatic Linguistics (NACAL 42), Leiden University, 14–16 February 2014. Paul, Ludwig. 2013. A Grammar of Early Judaeo-Persian. Wiesbaden: Reichert. Pennacchietti, Fabrizzio A. 1988. “Verbo neo-aramaico e verbo neo-iranico.” In Tipologie della convergenza linguistica. Atti del Convegno della Società Italiana di Glottologia, edited by Vincenzo Orioles, 93–101. Pisa: Giardini. Polotsky, Hans Jakob, 1996, “Notes on a Neo-Syriac Grammar.” Israel Oriental Studies 16: 11–48. Rendsburg, Gary A. 1991. “Parallel Developments in Mishnaic Hebrew, Colloquial Arabic, and other Varieties of Spoken Semitic.” In Semitic Studies in Honor of Wolf Leslau, vol. II, edited by Alan S. Kaye, 1265–1277. Wiesbaden: Harrassowitz. Skjaervø, Prods Oktor. 2009. “Middle West Iranian.” In The Iranian Languages, edited by Gernot L Windfuhr, 196–278. London and New York: Routledge. Sokoloff, Michael. 2002. A Dictionary of Jewish Babylonian Aramaic of the Talmudic and Geonic Periods. Ramat-Gan: Bar Ilan University Press/Baltimore: The Johns Hopkins University Press. Stilo, Donald L. 2007a. “Isfahan, xix. Jewish Dialect.” In Encyclopaedia Iranica, Vol. XIV, edited by Ehsan Yarshater, 77–84. New York: Encyclopædia Iranica Foundation. —. 2007b. “Isfahan, xxi. Provincial Dialects.” In Encyclopaedia Iranica, Vol. XIV, edited by Ehsan Yarshater, 93–112. New York: Encyclopædia Iranica Foundation. —. 2008. “Two Sets of Mobile Verbal Person Agreement Markers in the Northern Talyshi Language.” In Aspects of Iranian Linguistics: Papers in Honor of Mohammad Reza Bateni, edited by Simin Karimi et al., 363–390. Newcastle-upon-Tyne: Cambridge Scholars Press. —. in press. “Further Notes on the Iranian Substratum of Azerbaijani Turkish.” —. forthcoming. “On the non-Persian Iranian Substratum of Azerbaijan.” —. in progress. Atlas of the Araxes-Iran Linguistic Area.

452

PAUL M. NOORLANDER AND DONALD STILO

Talay, Shabo. 2008. “Bridging the Tigris: Common Features in Ṭuroyo and NorthEastern Neo-Aramaic.” In Suryoye l-Suryoye: ausgewählte Beiträge zur aramäischen Sprache, Geschichte und Kultur, edited by Shabo Talay, Bibliotheca Nisibinensis 1. Piscataway, NJ: Gorgias Press. Tsereteli, Konstantin G. 1978. The Modern Assyrian Language. Moscow: Nauka. Utas, Bo. 2013. “The Grammatical Transition from Middle to New Persian.” In From Old to New Persian, edited by Bo Utas, Jahani, Carina and Mehrdad Fallahzadeh, 251–259. Beiträge zur Iranistik, Bd 38. Wiesbaden: Reichert. van den Berg, Helma. 2005. “The East Caucasian Language Family.” Lingua 115: 147–190. Windfuhr, Gernot L. 2009. “Dialectology and Topics.” In The Iranian Languages, edited by Gernot L. Windfuhr, 5–42. London-New York: Routledge. Yarshater, Ehsan. 2005. “Tāti Dialect of Kalāsur.” In Languages of Iran: Past and Present, edited by Dieter Weber, 269–284. Iranica 8. Wiesbaden: Harrassowitz. Any data not cited are derived from D. Stilo’s field notes.

ON THE CONVERGENCE OF VERBAL SYSTEMS OF ARAMAIC AND ITS ADIGMS

NEIGHBOURS. PART II: PAST PAR-

DERIVED FROM PRESENT EQUIVALENTS

DONALD STILO AND PAUL M. NOORLANDER

1

1. INTRODUCTION

In this article, we will examine the strategies for the Past derivation of various verbal paradigms by building on the processes presented in Part I. We will only focus on the languages of the core area2 of the Araxes-Iran Linguistic Area (AILA, see Part I): Northeast Neo-Aramaic (NENA), rather outstanding Armenian dialects of a small area in Iran, Armenia, and Azerbaijan, varieties of Tati and Talyshi (NWI), and Udi. For comparison purposes we will also marginally include some languages of the extended AILA area: Standard Eastern Armenian, Persian, Northern and Central Kurdish (Kurmanji and Sorani-Mukri respectively), and Caucasian Tat (CT, a SWI language). While we have already discussed the formation of the Present, Present Progressive, Subjunctive, and Future in Part I, here we will examine the Past equivalents of each of these paradigms: the Imperfect, Past Progressive, Past Subjunctive, and Condi-

1

I (Donald Stilo) would like to express my gratitude to Paul M. Noorlander, who, as joint author, has offered me invaluable assistance in writing those sections and paragraphs pertaining to Aramaic. 2 As discussed in Part I, our usage of the term ‘core area’ may be somewhat misleading: in discussing the core group and their extensive shared features, please keep in mind that although these languages are at present located in very disparate areas (north-central Iran, northern Iraq, Daghestan), they certainly had a different, and larger, distribution some 10001500 years ago (see Part I for additional discussion).

454

DONALD STILO AND PAUL M. NOORLANDER

tional, respectively. The latter set of paradigms is derived from the former simply by the addition of what is often called the ‘Past Converter’ (PC)3 in the representatives of the four families of the core group. We will also include four additional paradigms that use a PC. These were not discussed in Part I because they either are based on the Past tenses or are defective verbs. These paradigms include the Present and Past forms of 1) the copula, 2) the existence verb (or particle, depending on the language), 3) the periphrastic expression of Predicative Possession (‘have’), based on the existence verb, and 4) the Present Perfect and the conversion to Past Perfect. Beginning with the various Present system paradigms of Part I as bases, only the core AILA languages add a morpheme of ‘pastness’, the PC, to fully conjugated Present forms to derive the corresponding set of paradigms of the Past system. What is significant here is that this PC is usually an invariable form of the 3 rd SG past copula (COP.PST.3, here also glossed as AUX and occasionally ‘was’). This morpheme is added in a canonical agglutinative fashion.4 That is, rather than conflate the categories of person and tense together in a fusional strategy—e.g. the 1st singular markers in Modern Greek θel-o ‘I want’ > i-θel-a ‘I wanted’ or the change of the Present stem of the verb to its Past stem of Persian mi-xah-æm ‘I want’ > mi-xast-æm ‘I wanted’ or Kurdish (Zakho): āvēž- > āvēt- ‘throw’—the core AILA languages do not vary from Present to Past when the PC is added.

2. RELATIONSHIP OF THE COPULA TO THE PAST CONVERTER

The relationship of the PC to the past copula in the various languages of the core area can be reduced to a continuum of five typologies: 1) The PC is completely identical to the 3rd singular past copula in the given language (aberrant Araxes-area Armenian dialects, Southern Talyshi, Northern Tati dialects (Kalāsuri, Khoynarudi))

3

This is glossed in examples below as CONV, as elsewhere in this volume. Even Turkic languages are one small degree lower on the scale of agglutination by violating one of the key features of agglutination here, specifically the requirement that a given category should have one invariable form. The PAMs of local Turkic languages are suppletive according to tense in some persons, e.g. Turkish: gid-iyor-sun > gid-iyor-di-n, gidiyor-uz > gid-iyor-di-k, gid-iyor-sunuz > gid-iyor-di-niz ‘you (SG), we, you (PL) are going > were going’. 4

ON THE CONVERGENCE OF VERBAL SYSTEMS. PART II Urmia Armenian

Southern Talyshi

Kalāsuri (N. Tati)

Past Cop. Past 3rd SG Converter =er =er

Past Cop. Past 3rd SG Converter =a =a

Past Cop. Past rd 3 sg Converter =u =u5

455

2) The PC is almost identical to the 3rd sg. past copula differing from it only slightly as in many NENA varieties. This small change represents one small step along the path of grammaticalization. Chr. Urmia Aramaic Past Cop. Past rd 3 SG Conv. =ivæ =væ

Chr. Barwar Aramaic Past Cop. Past rd 3 SG Conv. =iwa =wa

Chr. Jilu Aramaic Past Cop. Past 3rd SG Conv. =iwa =wa

The Neo-Aramaic PC, wa is derived from earlier hwā ‘he was’, the 3rd person masculine singular of the suffix-conjugation of h-w-y ‘be, exist’, which functioned as the past copula. In conjunction with the participles, it forms the past counterpart to the present progressive or habitual (see Part I), compare Classical Syriac (CS) šāqel=nā ‘I take, am taking’ and šāqel=wēṯ ‘I used to take, was taking’. Note the elision of the initial h (*hw- > w-). The past copula gradually became frozen and uninflectable in Late Aramaic, as in (a), though agreement was not always altogether lost (b): Jewish Babylonian Aramaic (Sokoloff 2002: 372b) (a) hwa yaṯḇi=nnan was sit.IMPRF.PTCP.PL=1PL.C ‘We used to sit’

(b) hwu ʿaḇḏi=nnan they_were act.IMPF.PTCP.PL=1PL.C ‘We used to act’

As the examples above show, this increasingly grammaticalized auxiliary precedes the main verb it modifies in Jewish Babylonian Aramaic. This is also the situation we observe in Western Neo-Aramaic. There the particle wa precedes the predicate consisting of a nominal or participle (or existential īṯ); for example, wa lōyef ‘he was learning’ (Arnold 1990: 190). Interestingly, in NENA (and Məδyoyo Ṭuroyo), on the other hand, the PC wa directly follows the finite verb (before the oblique L-suffixes) as in CS.

5

The forms of S. Talyshi =a and N. Tati =u are to be considered as identical since there is a regular sound change of *ā > u throughout Kalāsuri, e.g. Māsulei a/ Kalāsuri u: dar/dur ‘tree’, pa/pu ‘foot’, adəm/ udæm ‘person’ (Yarshater 2005).

456

DONALD STILO AND PAUL M. NOORLANDER

(3) The PC is related to, but somewhat removed from, the past copula of the present language. Either only some of the original phonetic material of the diachronic past copula is still recognizable or the structure of the past copula differs from that of the PC. In several varieties of NENA, the past copula is the preterit of the verb hwy ‘be, exist’ (we- be.PRF < *hwē ‘been’). In the course of time, the present copula, itself most likely pronominal in origin (Khan 2001 and elsewhere), developed a suppletive relation with this verb: J. Urmia Aramaic (also Jewish Zakho, Jewish Arbel, etc.) Past Cop. Past 3rd SG Conv. =we-le (M), =we-la (F) =wa (4) The form of the PC in the contemporary language can only be traced historically to an older, often reconstructed past copula, as in Udi. Regarding the Caucasian Albanian and Udi PC =i/=y, Maysak (2008a: 130), partially citing other authors, states: “As the -y clitic attaches to already fully formed finite verb forms and its semantic function represents a shift of temporal reference to the past system (and with past tense forms [a shift] to the domain of “hyperpast”), it serves as a typical marker of a “retrospective shift”, in the terminology of V.A. Plungian. Analogous to the Udi =y, for example, is the suffix -y in Lezgi, which also builds the forms of the “past system” that attach to finite forms of the “present system”, cf. raxazva ‘speaks’ ~ imperfect raxazva-y ‘spoke [говорил]’, etc. ... Diachronically the Udi as well as the Lezgi morpheme goes back to a past copula [Alekseev 1985: 97-98]. In a number of Lezgic languages, the present and past tenses of the copula function equally, forming an equipollent opposition, cf., for example, Agul meHemed müʕelim e ‘Mohammed is a teacher’ vs. meHemed müʕelim iy ‘Mohammed was a teacher’...”

NENA: Jewish Sanandaj Aramaic is also type 4, as the Past Copula (=ye-le) and the PC (=wa) have no transparent connection in the present language. The element /y/ presumably goes back to an earlier */iw/, cf. C. Barwar =iwən ‘I am’ and C. Qaraqosh =iyən ‘I am’ (Khan 2011; 2012) and spread to the entire paradigm in J.Sanandaj, where it is inflected like hwy ‘be’. (5) The PC derives from ‘become’. In most AILA varieties ‘be’ and ‘become’ are forms of the same root or, if different, ‘become’ replaces ‘be’ in the paradigms that the latter lacks. As for the PC, some Armenian dialects use a grammaticalized form of lin- ‘to be/become’: Karchevan -le and Agulis -nel (Nakhichevan), Meghri (Armenia) -læ, Hadrut (Mountainous Karabagh) -læl.

ON THE CONVERGENCE OF VERBAL SYSTEMS. PART II

457

All other AILA varieties have a past morpheme but it differs from the PC variously, depending on the language: 1) it is not a copula, e.g. Georgian v-svam> vsvam-d-i ‘I drink/used to drink’, Azeri diy-ir-æm > diy-ir-di-m ‘I say/used to say’; 2) it is not encliticized to a fully formed present tense, e.g. S. Talyshi zon-ə́m > zon-í-m ‘I know/knew’ (but not *zon-ə́m + =i), cf. past stem zonəst (see also §3.3); 3) in some languages the Imperfect also requires a past root of the verb (see Persian mixah-æm > mi-xast-æm above).

3. PARADIGMS USING THE PAST CONVERTER 3.1. The Copula: Present > Past

The Present vs. Past copula paradigms show exact parallelisms among the core area languages in a process that, to our mind, is crosslinguistically very unusual: they form the Past copula by simply adding a PC (itself of past copular origin) to the Present copula: ‘I am-was’.

Aramaic In the modern dialects of NENA, there are different types of PCs in relation to the past copula according to the typologies established above (for the sake of brevity only masculine forms are given): Copula: ‘to be good; to be fine’ Typology 2

Typology 3

Christian Urmia

Jewish Urmia

Present.M + spay=vin + spay=vit + spay=ile

Present jwān=ilen jwān=ilet jwān=ile

Past + spay=vin=væ + spay=vit=væ + spay=ivæ

Typology 2

Past jwān=wéli jwān=wéli jwān=wele

Typology 4

Christian Barwar

Jewish Sanandaj

(Khan 2008a: 622, 180)

(based on Khan 2009: 85-87)

Present spay=iwən spay=iwət spay=ile

Past spay=iwən-wa spay=iwət-wa spay=wa-wa

Present ʿayza=yena ʿayza=yet ʿayza=y

Past ʿayza=yeli ʿayza=yelox ʿayza=yele

458

DONALD STILO AND PAUL M. NOORLANDER

Armenian

The Armenian dialects of the core Araxes area, but not Standard Armenian, have various AM+WAS typologies. The Urmia-Khoy dialect (type 1) adds an invariable 3rd sg Past copula to the Present copula. Since Meghri, Agulis, Hadrut, and Karchevan dialects use a grammaticalized form of ‘become’, this particle does not replace 3rd sg Present PAM as in Urmia Armenian, but adds the PC to it: Type 1

‘to be good; to be fine’

Type 5

Urmia Armenian

Meghri (Ałayan 1954: 233)

Present lav=em lav=es lav=a

Present lav=əm lav=əs lav=ə

Past lav=em=er lav=es=er lav=er

Past lav=əm=læ lav=əs=læ lav=ə=læ

The enclitic copula and PC in Agulis, however, are more flexible than in other AILA languages/dialects since they may occur in either order: Type 5, mobile Agulis Armenian (Sargseants 1883: 106) Karchevan (Muradyan 1960: 128) Present lav=əm lav=əs lav=a

Past lav=əm=nel lav=əs=nel lav=a=nel

Alternate lav=nel=əm lav=nel=əs lav=nel=a

PresentPast lav=im lav=im=le lav=is lav=is=le lav=a lav=a=le

While local Armenian dialects have AM+WAS typology, Standard East Armenian has the AM > WAS typology found in Persian (=æm > bud-æm), English, or Spanish (soy > era), without the use of a PC. The Armenian Past copula is also enclitic like the Present but its forms are similar to the Preterite verbal PAMs: Standard Eastern Armenian Present ≠ Past = Preterite PAMs (‘speak’) lav=em lav=ei xos-ec‘-i. lav=es lav=eir xos-ec‘-ir lav=e lav=er xos-ec‘-∅

Iranian (Talyshi)

The S. Talyshi Past copula with a PC by analogy has also affected the Present copula by encliticizing an invariable 3rd sg. Present copula to all persons of the Present copu-

ON THE CONVERGENCE OF VERBAL SYSTEMS. PART II

459

la, thus matching the AM+WAS pattern by an AM+IS pattern, an innovation not paralleled in other varieties of Talyshi or other AILA languages: Typology 1 (‘to be good, fine’) Urmia Armenian Pres. Past lav=em lav=em=er lav=es lav=es=er lav=a lav=er

Māsulei (S. Talyshi) Pres. Past xob=em=æ xob=em=a xob=eš=æ xob=eš=a xob=æ xob=a

Udi Due to the innovation of PAMs in Udi (see Part I), the original Present copula was lost and the Past equivalent is formed with the core area AILA pattern of adding the PC =y to the new Present copula: alim=zu ‘I am a teacher’ alim=zu=y ‘I was a teacher’ (Nij dialect, adapted from Maysak 2008a: 131)

jæyil=zu ‘I am young’ jæyil=zu=y ‘I was young’

3.2. Present Tense > Imperfect

In the four families of the core AILA area—NENA, Udi, all forms of Armenian, and all forms of Talyshi—the Imperfect is formed on the Present stem of the verb. In only three of these groups (excepting Talyshi), however, is the Imperfect formed with a fully conjugated Present tense + PC without any additional change. This section covers both synthetic and analytic paradigms for converting Present/Present Progressive > Imperfect/Past Progressive via the PC. We will also try to highlight the contrast between the Past stem vs. Past forms derived from the Present stem+PC with verbs whose Present vs. Past stems differ significantly or are suppletive. Present

Imperfect

Christian Urmia Aramaic

Present

Imperfect

Urmia-Khoy Armenian

ki-æz-in ki-æz-in-væ oṭ-es=em DUR-go-1SG.M DUR-go-1SG.M-CONV eat-DUR-COP.1 ‘I (M) go’

‘I (M) used to go

(Preterite: xiš-l-iy)

‘I eat’

oṭ-es=em=er eat-DUR=COP.1=CONV ‘I was eating’

́ (Preterite: ḳir-im Asatrian 1962: 135)

460

DONALD STILO AND PAUL M. NOORLANDER Jewish Urmia Aramaic

Meghri Armenian (Ałayan 1954: 232)

g-ez-en g-ez-en-wa əṭ-es=em əṭ-es=em=læ DUR-go-1SG.M DUR-go-1SG.M-CONV eat-DUR=COP.1 eat-DUR=COP.1=CONV ‘I (M) go’

‘I (M) used to go’

‘I eat’

‘I was eating’

(Preterite: zil-li)

(Preterite: ḳer-al=əm ‘I ate’)

Christian Barwar Aramaic

Agulis Armenian

y-azəl y-azəl-∅-wa naha-m=əm naha-m=əm=nel DUR-go-3SG.M DUR-go-3SG.M-CONV say-DUR=COP.1 say-DUR=COP.1=CONV ‘he goes’

‘he used to go’

‘I say’

‘I was saying’

(Preterite: zil-le ‘he went’)

(Preterite: aha-m ‘I said’)

(Khan 2008a: 598)

(Sargseants 1883: 114-115, 117)

Udi (adapted from Maysak 2008b: 163) e=z-sa e=z-sa=y [essa(y)] nex=e nex=e=y come-1SG-DUR come-1SG-DUR-CONV say.DUR=3SG say.DUR=3SG=CONV ‘I come’

‘I was coming’

(Preterite: har-i-zu ‘I came’)

‘s/he says’

‘s/he was saying’

(Preterite: p-i-ne ‘s/he said’)

3.3. Imperfect < Present stem, no PC As opposed to a PC-type formation of the Imperfect, many AILA languages beyond the core area base the Imperfect on a Past stem, e.g. Persian mí-foruš-æm ‘I sell’> Imperfect mí-foruxt-æm ‘I used to sell’. This typology is only mentioned in passing in this article. Some AILA varieties beyond the core area, however, have a transitional type of Imperfect formation. That is, Standard E. and W. Armenian (and less aberrant dialects), Georgian, Laz, all forms of Talyshi, and a few Central Tati dialects neighboring on S. Talyshi base the Imperfect on the Present stem but without a copuladerived PC. Instead, most replace the enclitic Present AUX/PAMs with the equivalent Past forms to form the Imperfect from the Present tense. Good examples of this type are Standard Eastern and Western Armenian in which both person and tense are included fusionally in one morpheme: -(e)m ‘1st sg., Present’ vs. -(e)i ‘1st sg., Past’:

ON THE CONVERGENCE OF VERBAL SYSTEMS. PART II

461

Standard Eastern Armenian Present

Imperfect

Preterite

dən-um=em put-DUR=COP.1SG

dən-um=ei put-DUR=COP.PST.1SG

dər-ec‘-i put.PST-PST.1SG

‘I put (PRS)’

‘I used to put’

‘I put’

Standard Western Armenian Present

Imperfect

Preterite

g-erta-m DUR-go-1SG

g-erta-i DUR-go-PST.1SG

ka-c-i go.PST-PST-PST.1SG

‘I go’

‘I used to go’

‘I went’

As mentioned, all forms of Talyshi base the Imperfect on the present stem of the verb. Compare all three Talyshi clusters with Rashti (Caspian), which also uses the -i- suffix in the Imperfect, and Kurmanji:

PRS PST 1SG 2SG 3SG

Augment + -i-

-i-, no Augment

General Durative

Leriki

Asālemi

Māsulei

Yerevan Kurmanji

N.Talyshi

C.Talyshi

S.Talyshi Caspian

N. Kurdish

‘hit’ žænžæ-

‘run’ virijvirit-

‘cook’ pepætt-

‘fall’* khævkhæt-

æ-žæn-i-m æ-žæn-i-š æ-žæn-i-∅

æ-vrij-i-m æ-vrij-i-š æ-vrij-i-∅

p-i-m p-i-š p-i-∅

Rashti ‘sleep’ xusxuft-

xúft-i-m xúft-i-(i) xúft-i-∅

də-khæt-əm də-khæt-i də-khæt-∅

*(Adapted from Kurdoev/Cukerman 1950: 46)

N. Talyshi and Hadrut Armenian (Mountainous Karabagh), geographically not very far from each other, both have only one Present tense but two Past duratives. 1) Hadrut has the Standard Armenian type Imperfect: Present AUX > Past AUX (no PC), while N. Talyshi has the Imperfect, as above; 2) an analytic Past Progressive formed from the Present in both languages: a PC strategy in Hadrut while N. Talyshi uses the Present AUX > Past AUX strategy (like Standard or Hadrut Armenian 1, this paragraph). The N. Talyshi paradigm now serves as a quite general Past durative, often overlapping with the Imperfect.

462

DONALD STILO AND PAUL M. NOORLANDER Hadrut Armenian (məna- ‘to stay’, Połosian 1965: 178) Present

Imperfect (point 1) Past Continuous (point 2)

məna-s=əm məna-s=əs məna-s=a

məna-s=i məna-s=ir məna-s=ar

məna-s=əm=læl məna-s=əs=læl məna-s=a=læl

stay-DUR=AUX stay-DUR=AUX.PAST stay-DUR=AUX=CONV Leriki (Northern Talyshi; žæn-/ž(æ)- ‘to hit’) ž-e-dæ=m æ-žæn-i-m hit-INF-LOC=AUX.1SG AUG-hit-DUR=1SG

ž-e-dæ=b-im (point 2) hit-INF-LOC=AUX.PST-1SG

Other languages in the immediate area also form the Imperfect on the present stem to which a past morpheme is added that is not quite the same as a PC, nor is it related to the copula: Georgian

Laz (Kartvelian, Turkey)

Present

Imperfect

Preterite

Present

Imperfect

v-amb-ob

v-amb-ob-di

v-t‘k‘v-i

b-zop’on

b-zop’on-t’i

‘I said’

‘I say’

‘I used to say’

Preterite v-t‘k‘v-i

1SG-say-DUR 1SG-say-DUR-PST 1SG-say.PST-PST SG-say-DUR 1SG-say-DUR-IMPRF

1SG-say.PST-PST

‘I say’

‘I said’

‘I used to say’

(Lacroix 2009: 724: 1803b) The Azeri and Turkish paradigms are also not exactly parallel to this situation since, as agglutinative languages, they do not exhibit Present vs. Past stems. While there is a morpheme of pastness, it is quite different from the PC. Also, the Turkic past morpheme is not simply added to the Present tense to make the Imperfect since the PAMs are not quite the same (see footnote 4 above).

3.4. Future > Conditional

The core area languages that have a formal Future also have a formal Conditional with a PC (except for N. Talyshi). Persian, Qohrudi, Gilaki, where the Future is marginal (see Part I), do not have a dedicated Conditional paradigm and usually replace it with the Imperfect. Future

Conditional

Future

Conditional

bit-æmr-in FUT-say-1SG.M

bit-æmr-in=væ FUT-say-1SG.M=CONV

b-amr-en FUT-do-1SG.M

b-amr-en=wa FUT-do-1SG.M=CONV

‘I’ll (m.) say’

‘I (m.) would say’

‘I’ll (m.) say’

‘I (m.) would say’

Christian Urmi Aramaic

Jewish Urmi Aramaic

ON THE CONVERGENCE OF VERBAL SYSTEMS. PART II Urmia-Khoy Armenian

Karchevan Armenian (Muradyan 1960: 149)

as-elü=yem as-elü=yem=er say-FUT=COP.1 say-FUT=COP.1=CONV

as-elac‘uḳ=em say-FUT=COP.1

as-elac‘uḳ=em-læ say-FUT=COP.1-CONV

‘I will say’

‘I will say’

‘I would say’

‘I would say’

Agulis Armenian

463

Hadrut Armenian

nah-il-əm nah-il=əm=nel ḳ-ərt‘-il-æḳan=əm ḳ-ərt‘-il-æḳan=əm=læl say-FUT-COP.1 FUT-go-FUT-FUT=COP.1 say-FUT=COP.1=CONV FUT-go-FUT-FUT=COP.1=CONV

‘I will say’

‘I would say’

(Sargseants 1883: 119, 123)

‘I will go’

‘I would go’

(Połosian 1965: 189–190)

Since Udi has three Future tenses, as does standard Armenian, it also has three dedicated Conditional paradigms formed with the PC but the true conditional senses of this paradigm seem to be only marginal (based Maysak 2008b: 173–182): Future I

Conditional I

Fut. Potential

Conditional Potential

uḳ-al=zu uḳ-al=zu=y say-FUT-1SG say-FUT-1SG=CONV

u=z=ḳ-o u=z=ḳ-o=y say=1SG=say-FUT say=1SG=say-FUT=CONV

‘I will say’

‘I will say’

‘I would say’

‘I would have said’

As Standard East Armenian (see the discussion of the Future in Part I) does not have a PC strategy, it forms its Conditional paradigms by replacing the Present copula of the Futures with the Past copula: Future I

Conditional I

Future II

Conditional II

ḳə-dən-em ḳə-dən-ei dən-elu=yem FUT-put-1SG FUT-put-PST.1SG put-FUTII-1COP

dən-elu=ei put-FUTII=1COP.PST

‘I will put’

‘I would put’

‘I would put’

3.5. Past Subjunctive

‘I will put’

The Past Subjunctive is formed with the Present Subjunctive plus the PC in the languages of the core area. Many of its usages are similar to those of the Present Subjunctive but in a past context (see excursus):

464

DONALD STILO AND PAUL M. NOORLANDER

Pres. Subj. Past Subjunctive Christian Urmi Aramaic ∅-æmr-in SUBJ-say-1SG.M

Pres. Subjunctive Past Subjunctive Urmia-Khoy Armenian

∅-æmr-in-væ ∅-as-em SUBJ-say-1SG.M-CONV SUBJ-say-1SG

‘(that) I (M) say: Pres > Past’ Udi

∅-as-em=er SUBJ-say-1SG= COP.PST.3

‘(that) I say: Pres > Past’ Agulis Armenian

b-a-z do-SUBJ-1SG

b-a-zu=y do-SUBJ-1SG=CONV

∅-ah-im SUBJ-say-1SG

∅-ah-im=nel SUBJ-say-1SG=CONV

‘(that) I do: Pres > Past’

‘(that) I say: Pres > Past’

(based on Maysak 2008b: 206–207)

(Sargseants 1883: 119, 123)

3.6. Past Subjunctive < Present stem, no PC

To form the Past Subjunctive, all varieties of Talyshi take the fully conjugated form of the Imperfect and prefix the usual Subjunctive marker bV- to it and also move the stress to the initial syllable: Leriki (N.Talyshi) Pres. Subj.

bə́-žæn-əm SUBJ-hit-1SG1

Imperfect

Past Subj.

Preterite

æ-žæn-í-m b-ǽ-žæn-i-m žæ=m-e AUG-hit-IMPRF-1SG1 SUBJ-AUG-hit-IMPRF-1SG1 hit.PST-1SG2-AUX

All forms of Armenian have a dedicated Past Subjunctive but, except for the dialects mentioned above, instead of the PC added to the Present PAMs, they have Past PAMs, as we saw under the Standard Armenian copula, the derived Imperfect, and the Conditional above: Standard Armenian

Hadrut Armenian (Połosian 1965: 180)

Pres. Subj.

Past Subjunctive

∅-as-em

∅-as-ei

∅-məna-m

∅-mən-i

∅-as-i SUBJ-say-PAM

∅-as-er SUBJ-say-PAM.PST

∅-məna-∅ SUBJ-say-PAM

∅-məna-r SUBJ-say-PAM.PST

∅-as-es

∅-as-eir

‘(that) I say > might say’

Pres. Subj. ∅-məna-s

Past Subjunctive ∅-mən-ir

‘(that) I stay > might stay’

ON THE CONVERGENCE OF VERBAL SYSTEMS. PART II

465

Excursus: Past Subjunctive

Since a dedicated Past Subjunctive has a restricted areal distribution—NENA, Udi, Armenian, and Talyshi6—we will discuss the uses of this paradigm at greater length with more examples than for the other paradigms. All varieties of Talyshi also use the Past Subjunctive in the equivalent sentences below but, as mentioned, its typology is not of the PC type. Persian, however, exclusively uses the Imperfect for (3, 4, 7) and the Present Subjunctive for (1, 2). The following are some typical usages of the Past Subjunctive among others: (i) Commonly after the Imperfect of the modals ‘want’ and ‘can’ and main verbs with control features over subordinate verbs in most AILA languages, the Past Subjunctive is more common in dependent clauses after Imperfect than Preterite modals, cf. Hoberman (1986: 68); Maysak (2008b: 207): (1) (a) Urmia Armenian, (b) C. Urmia Aramaic, (c) Leriki Present Modal + Present Subjunctive > (a)

ḳənṭə-nér-ə

uz-es=en

(c)

žen-on

pe-dæ=žon i

(b) bæxt-ætə

ki-+bay-i

me

xæ

xaṭ

xənʓor

∅-oṭ-en.>

gəlæ

sef

bə́-hæ-n.>

dænæ* xæbuyšæ

∅-+axl-i.>

Past Modal + Past Subjunctive (a)

ḳənṭə-nér-ə uz-es=en-er me xaṭ xənʓor woman-PL-DEF DUR-want-3PL-CONV one NC apple

(b) bæxt-ætə woman-PL (c)

∅-oṭ-en=er. SUBJ-eat-3PL=CONV

ki-+bay-i-va xæ dænæ xæbuyšæ ∅-+axl-i-va.* DUR-want-3PL-CONV one NC apple SUBJ-eat-3PL-CONV

žen-on pe-dæ=žon=be i gəlæ sef b-ǽ-hæ-y-n. woman-PL want-DUR-3PL2=COP.PST one NC apple SUBJ-AUG-eat-IMPRF-3PL

but Persian (Present Subjunctive): (d) zæn-a woman-PL

mí-xast-ænd DUR-want.PST-3PL2

ye one

dune sib bó-xor-ænd. NC apple SUBJ-eat-3PL

‘The women want > wanted to eat an apple’. Aside from Georgian and Laz, already mentioned, various Caspian dialects (Stilo 2001) also have formal Past Subjunctives but their formation is of a completely different typology not based on the present stem and are not included in this discussion. 6

466

DONALD STILO AND PAUL M. NOORLANDER

((a) Ani Beyt-Movsess, personal communication, others adapted by authors; *NC optional in CUA) (2) (a) Nij dialect of Udi, (b) Jewish Urmia Aramaic, (c) Urmia Armenian, (d) Leriki (N. Talyshi) (a) čur=yan-sa=y bazar-e want=1PL-DUR-CONV market-DAT

toy-d-a=yan-iy.* sell-LV-SUBJ-1PL-CONV

(b) g-b-ex-wa g=šuqa DUR-want-1PL-CONV in=market

∅-zabn-ex-wa-la. SUBJ-sell-1PL-CONV-OBJ.3SG.F

(c)

uz-es-enḱ‘=er bazar-i want-DUR-1PL=CONV market-DAT

(d) pe-dæ-mon=be want-DUR-1P2= COP.PL

meč‘ ∅-c̣ax-enḱ‘=er* in SUBJ-sell-1PL-CONV

vožor=ædæ market=in

b-ǽ-hvat-i-mon. SUBJ-AUG-sell-IMPF-1P1

‘We wanted to sell (it < the cow) at the market’. (based Maysak 2008b: 207; others adapted by the authors); *(Indexing of Objects in the verb is not licensed in Armenian and thus does not play a role in Differential Object Marking as in NENA.)

(ii) When encoding past necessity, especially in languages where the ‘must’ modal is invariable and does not include a means of expressing tense. Invariable ‘must’ Invariable ‘must’ (3) Jewish Urmia Aramaic, Urmia Armenian āt́

gắrag en-óx

là

ṭü

ṗiṭi

č‘ə-p‘aḳ-es-er NEG-close-2SG-CONV-(OBJ.3PL)*

ašk‘-eṭ

you must eye-your

čem-ət-wá-lu*.

+

‘You should not have closed your eyes’. (Khan 2008b: 408, 60; Urmia Armenian adapted by authors) (4) (a) Leriki, (b) Udi, (c) Urmia Armenian, (d) J. Urmia Aramaic (a) bæ-pe ïštæ jif=o b-ǽ-hæ-y-m. ALLT-must self pocket=from SUBJ-AUG-eat-IMPRF-1SG < (Past root: hard-) (b) gæræg ič must self

jub-n-axo pocket-OBL-ABL

uk-a-zu=y eat-SUBJ-1SG-CONV < (Past root: kæy-)

ON THE CONVERGENCE OF VERBAL SYSTEMS. PART II (c)

ṗiṭi must

im my

ḳərban-e-n pocket-ABL-DEF

∅-oṭ-em=er SUBJ-eat-1SG-CONV

< (Past root: ḳer-)

(d) gắrag +m=bəlwan-ət noš-i ∅-+axl-en-wa must from=pocket-LNK self-1SG2 SUBJ-eat-1SG.M-CONV< (Past root: +xil-) but Persian (Imperfect): bayæd æz jib-e must from pocket-EZ

xód-æm self-1SG2

mí-xord-æm. SUBJ-eat.PAST-1SG1

‘I had to eat out of my own pocket (i.e., at my own expenses)’. (Talyshi from field notes, others adapted by authors) Tense-marked ‘must’ (5) Jewish Sanandaj Aramaic g-bé-wa ∅-hez-í-wa zărá DUR-must-CONV SUBJ-go-3PL-CONV wheat ∅-hăm-èn-wa-le=o SUBJ-bring-3PL-CONV-OBJ.3SG.M=and

∅-šaql-ì, SUBJ-buy-3PL ∅-găb-èn-wa-le. SUBJ-buy-3PL-CONV-OBJ.3SG.M

‘They [hadPST to] goCONV and buynoCONV wheat, bringCONV it back, and sortCONV it’ (adapted from Khan 2009: 422, 58) No Past Subjunctive (6) Bayat Azerbaijani > Present Subjunctive göray=edi bu ev-lær-i gæz-æ-g gör-æ-k must=COP.PST.3SG this house-PL-ACC search-SUBJ-1PL see-SUBJ-1PL kim=edi. who=COP.PST.3SG ‘We should have searched these houses to see who it was’. (adapted from Bulut 2006: 244) (iii) in the protasis of counterfactual conditional clauses (7) Jewish Urmia Aramaic (adapted from Khan 2008b: 379), Urmia Armenian ăgar ana ∅-ay-en-wa

ād

bela=il-et,...

if

you

house=COP.2SG

t‘æ

yes ∅-gíd-im=er I

SUBJ-know-1SG=CONV

ṭü

‘If I had known that you were at home, ...’.

ṭon=es,...

467

468

DONALD STILO AND PAUL M. NOORLANDER

(iv) in purposive clauses in the past tense (8) (a) Standard Armenian, (b) Urmia Armenian; (c) J.Urmia Aramaic (a) ∅-p‘aḳ-eir

vor

∅-ḳara-yinḱ‘

∅-bac‘-einḱ‘.

(b) ∅-p‘aḳ-es=er

vor

∅-ḳar-inḱ‘=er

∅-ṗac‘-inḱ‘=er

(c) b-doq-ət-wa-le

ki

+

∅-məss-ex-wa

FUT-close-2SG.M-CONV-OBJ.3SG.M

SUBJ-can-1PL-CONV

∅-palx-əx-wa-le SUBJ-open-1PL-CONV-OBJ.3SG

SUBOR

‘You should have closed it for us to be able to open it’. (eanc: OSD Polylogue 131) (9) Standard Armenian, J. Urmia Aramaic zang-el=er vor ∅-as-er ring-PST.PTCP=CONV SUBOR SUBJ-say-PST.3SG telefon widéwa* phone do.PST.PTCP.COP.PST.3SG ∅-amər-∅-wa ‘la’. SUBJ-say-3SG.M-CONV no

‘č‘e’. no

ki

SUBOR

*(< wida-ywa)

‘He had called (in order) to say “no”’. (eanc: OSD polylogue 052) (v) Independent usages (see also (8) above and (20) below): (10)

Pres. Subjunctive

Past Subjunctive

Nij Udi:

hikæ uḳ-a-z? > what say-SUBJ-1SG

hikæ uḳ-a-zu=y? what say-SUBJ-1SG=CONV

Urmia Armenian:

inč‘ ∅-as-em > what SUBJ-say-1SG

inč‘ ∅-as-em=er? what subj-say-1SG=CONV

‘What should I say?’

‘What should I have said?’

(Maysak 2008b: 206)

(adapted from same)

ON THE CONVERGENCE OF VERBAL SYSTEMS. PART II

469

(11) Urmia Armenian æγællæk‘æm xənʓor-k‘-ér-ə at.least apple-PL-PL-DEF

∅-t‘æl-en=er

SUBJ-throw-3PL=CONV

c̣óv sea

c̣unæ-nér-ə fish-PL-DEF

> -oṭ-en-er. SUBJ-eat-3PL=CONV

‘They should have at least thrown the apples into the sea for the fish to eat’. (Ani Beyt-Movsess, personal communication)

3.7. Present Perfect > Past Perfect

The paradigms of the past were not covered in Part I, but since the Perfects exhibit yet another example of a Past form deriving from a fully conjugated Present form+PC,7 we felt it would be relevant and useful to introduce these paradigms from the Past system. Jewish Urmia Aramaic

Urmia-Khoy Armenian

plix-én (*plixá=ilen) open.PST.PTCP-1SG.M

as-ir=em say-PST.PTCP=1SG

‘I have opened’

‘I have said’

plix-én-wa open.PST.PTCP-1SG.M-CONV

as-ir=em=er say-PST.PTCP=1SG=CONV

‘I had opened’

‘I had said’ (Asatrian 1962: 105)

Nij Udi har-e=zu

Māsulei (S. Talyshi) har-e=zu=y

hardǽ=m-æ

hardǽ=m-a

‘I have eaten’

‘I had eaten’

come-PRF=1SG come-PRF=1SG=CONV come.PST.PTCP-1SG2-AUX COME.PST.PTCP-1SG2-CONV

‘I have come’ ‘I had come’ (based on Maysak 2008a: 113)

7

Note also the first quote above from Maysak in this article: ‘...(and with past tense forms [a shift] to the domain of “hyperpast”)’. (Maysak 2008a: 130).

470

DONALD STILO AND PAUL M. NOORLANDER

NENA and Udi have two Past Perfects: one adds the PC to the Present Perfect; the other adds it to the Preterite. While these forms in fact belong to two separate categories with subtle yet important differences in NENA, we have grouped them together on typological-comparative grounds. The Udi forms may indeed also have similar nuances. Present Perfect

Past Perfect I

Preterite

Past Perfect II

ptix-l-iy

ptix-væ-l-iy

‘I opened’

‘I had opened’

Christian Urmia Aramaic ptiyxe=vin ( PAST

The expression of existence at least in the Present affirmative in the AILA zone is generally encoded by a dedicated verb form or particle: ‘there is’

‘there was’

‘there is’

‘there was’

Christian Urmia Aramaic

Meghri Armenian (Ałayan 1954: 235)

it there_is

ḳa there_isthere.is

it-væ there_is-CONV

Māsulei (S. Talyshi)

Kalāsuri (No. Tati)

éss-æ éss=a there_ist-COP.3 there_is=CONV

hest there_is

Mukri (Öpengin 2013: 233)

ḳa=læ there_is=CONV

hest=u there_is=CONV

Yerevan Kurmanji (Bakaev 1965: 46, 48)

he-ye he-bū hæ-yæ hæ-bu there_is-COP.3 there_is=COP.PST.3 there_is-COP.3 there_is=COP.PST.3

ON THE CONVERGENCE OF VERBAL SYSTEMS. PART II Nij Udi (Maysak 2008a: 109-110)

471

Azerbaijani (Tabrizi)

bu=ne bu=ne=y var(=di) var=ïdi there_is=3SG there_is=3SG=CONV there_is-COP.3 there_is=COP.PST.3 As Table 1 shows, existence forms are found at least in the Present affirmative throughout but are deficient in certain other paradigms, depending on the language: Present negative, Past, and all other tenses are replaced by the verb(s) ‘be/become’ (> be(come)). In the negative, the existence verb may either have a dedicated form, it may be negated as other verbs (+ negator), or it may be replaced by the negative of ‘be(come)’.

Table 1: Paradigms of Existence Present

Negation

Past

dedi-

dedi-

+neg -

> be

+PC

Sub-

Past

junct

Perf

past

> be

+past

> be

> be

cated

cated

ator

(come)

PAM

(come)

‘be’

(come)

(come)

most NENA

+

–

+

+

+

–

–

–

+

+

Urmia

+

–

+

+

–

+

–

–

+

+

+

–

+

–

–

+

–

–

+

+

N. Talyshi

+

–

–

+

–

–

–

+

+

+

S. Talyshi

+

–

–

+

–

–

–

+

+

+

Udi

+

–

+

–

+

–

–

–

+

+

Azerbaijani

+

+

–

–

–

–

–

+

+

+

Turkish

+

+

–

–

–

–

–

–

2

+

+

Kurmanji

+

+

–

–

–

–

–

+4

+

+

Sorani/

+

–

–

+

–

–

–

+

?

?

Persian

+1

–

–

+

–

–

+

–

+

+

Muslim

+

–

–

+

–

–

+

–

+

+

+

–

–

+

+

–

–

–

+

+

Armenian St. E Armenian

1

1

3

4

Mukri

Cauc. Tat Kalāsuri (N. Tat) 1

Derived from ‘be’ and may still have some uses as ‘be’. 2 Formed with the usual morpheme of

pastness (not a PC).

3

Applies to Optative, Future, etc.

4

Kurmanji has a dedicated existence

verb, derived from ‘be’, in all paradigms. For Mukri (C. Kurdish) the affirmative Present and

472

DONALD STILO AND PAUL M. NOORLANDER

Past (and possibly other tenses) retain the same dedicated form derived from ‘be’ as Kurmanji but the negative forms here are replaced by ‘be’.

In languages in which ‘be’ replaces the above deficient paradigms, ‘be’ itself may also be replaced by ‘become’ in these paradigms (see (12) below), e.g. Azeri: ægær ol-sun (3rd sg., Optative) ‘if there is, if s/he is, if s/he becomes, if it is possible’. Persian only has a formal existence verb in the Present affirmative; all other forms are filled in by ‘be’, which is fully independent of ‘become’: Pres. hæst, Past bud, Subjunct. báš-e (= ‘be’) vs. bé-š-e ‘becomes (subj.)’. Note the following correlations in Vafsi: (1) copula, (2) affirmative existence, (3) negative existence/‘be’, (4) past existence/‘be’: (1) =e, (2) h-e, (3) n-e, (4) b-e. The Udi existence verb only occurs in the Present (bu-), Past (bu=y, +PC), and negated forms of these. In other tenses, bak- ‘become’ assumes the functions of copula and existence: pasčaγ=nu ‘you are king’ > pasčaγ bak-al=lu ‘you will be king’ (adapted from Harris 2002: 60-61); ?ḳæ=æ bu=y? ‘what was there (there)?’ vs. ḳæ=æ bak-e? ‘what happened (= became)?’ (adapted from Maysak 2008a: 150). (12) (a) Colloq. East Armenian; (b) Jewsih Urmia Aramaic; (c) Leriki (N. Talyshi); (d) Colloq. Persian; (e) Colloq. Azerbaijani if music NEG-SUBJ-become-3SG (a) et‘e yeražəšṭut‘yun č‘ə-∅-lin-i,... (b) ăgar musika la-∅-hawy-a,... (= ‘be’) (c) ægæm musiγi nḯ-bu-∅,... (= ‘be’) (d) æge musiqi nǽ-baš-e,... (= ‘be’) (e) ægær musigi ol-ma-sun,... (ol- ‘become’) ‘If there is no music, (I can’t drive the car.)’ (eanc: OPD 161-163) While the N. Talyshi Past copula does not use a PC, the Existence verb, which is conjugable,8 does:

8

For example, Armenian Pres/Past: ḳam/ḳai, ḳas/ḳair, ḳa/ḳar, etc.: uγγaḳi eṭ gorc̣-um č‘-ḳa-m (NEG-EXIST-1 SG) ‘I’m not directly involved in that business’ (eanc: Xač‘atryan Vahe 1997); Persian: bazi darim. hǽsti? ‘We’ve got a game going. Are you in?’; Azeri: Aff. varám, varsan, var(dir); Neg. yoxám, yoxsán, yox(dïr), e.g. górxma, mæn varám ‘Don’t be afraid, I’m here (for you), cf. Ital. ‘non aver paura, ci sono io’, or ilan? mæn yoxám! ‘Snakes? Count me out!’

ON THE CONVERGENCE OF VERBAL SYSTEMS. PART II Leriki: Copula Pres. =im =iš =e

Past b-im b-iš b-e

473

Existence verb Pres. Past hïst=im hïst=im=be (cf. Pers. *hæst-æm=bud) hïst=iš hïst=iš=be hïst=e hïst=e=be

5. PREDICATIVE POSSESSION: ‘HAVE’ > ‘HAD’

Except for Armenian, all other languages of the core area (and extending westward) express Predicative Possession periphrastically, cf. Latin mihi est, Russian u men’a yes’t’, Arabic ʿand-ī, etc. ‘I have’. The basis for these constructions in the core area languages is the existence verb/particle we just saw above. Note that the position of the PC in Aramaic differs from most of the other languages. In addition, -liy is a suffix (actually a double suffix: -l-iy) while in the other languages Set2 is a mobile enclitic: Christian Urmia Aramaic it-l-iy it-væ-l-iy

Azerbaijani (Tabrizi) var=ïm=di var=ïm=ïdi

‘I have’

‘I have’

there_is-L-1SG2 there_is-CONV-L-1SG2

Kalāsuri ní=m=æ

‘I had’

there_is=1SG2=COP.3

ní=m=u.

there_is=1SG2=COP.PST.3

‘I had’

Mukri (Central Kurdish) he=mān=e he=mān bū

be.NEG=1SG2=COP.3 be.NEG=1SG2=COP.PST.3 there_is-1P2=COP .3 there.is-1P2= COP.PST.3

‘I don’t have’ ‘I didn’t have’ (adapted from Yarshater 2005: 281)

‘we have’ ‘we had’ (adapted from Öpengin 2013: 233-234)

Of the three sets of verbal PAMs in Udi, the possessive set is used with the Existence verb in the have-construction, although the usual set of PAMs (Harris 2002: 29) may also be used (see 13a) below):

474

DONALD STILO AND PAUL M. NOORLANDER

‘have’ Pres.

‘have’ Past

‘have’ Pres.

Udi (adapted from Harris 2002: 28) bu=bez bu=bez=i bu=vi bu=vi=y bu=ṭa bu=ṭa=y there_is=PAM.POSS

there_is=SET2.POSS=CONV

‘have’ Past

Leriki (N. Talyshi) hḯst=ïm=e9 hḯst=ïm=be (hḯst=ï=e)10 hḯst=e=be hḯst=ïž=e hḯst=ïž=be

there_is=1SG2=COP.3SG

there_is=1SG2=COP.PST.3SG

5.1. The Possessor Clitics and their Placement The Set2 clitics of the above paradigms are secondary or oblique clitics (or clitic-like suffixes) that encode or coindex the possessor. They are called Set2 here since in many AILA languages they are polyfunctional as: A) possessors (‘my, etc’.), B) direct object PAMs encoded in transitive verbs, which may or may not be restricted to the Present system, and C) Agents of Past transitives in Tati-Talyshi, Sorani-Mukri, NENA (in some varieties intransitive subject in past as well). Kurmanji, Caspian languages, and Caucasian Tat completely lack these original Set2 clitics in all domains. Except for NENA, the Set2 clitics in have-constructions are mobile. They usually front to the possessum or occasionally to other hosts (see (17), Kalāsuri) in the clause, but not to the possessor: Udi (13) xib xinær=re bu=i three girl=3SG there_is=CONV

Leriki haft zuæ=šhest=be seven son=3SG2there_is=COP.PST.3SG

‘He had three daughters’. (Harris: 2002, 62) ‘He had seven sons’. (Miller 1930: 83) Tabriz Azeri (14) gonaγ-un var guest-2SG2 there_is ‘You have a guest’. (Kıral 2001: 148, 86)

9

Wārmāwa (Central Kurdish) sē kuṙ=ī bū three son=3SG2 be.PST.3SG ‘He had three sons’. (MacKenzie 1962: 252)

The form hḯst=ïm=e ‘I have’, based on the Existence verb, should not be confused with the equivalent person of the Existence verb hḯst-im ‘I exist’ (ï vs. i). 10 The form hḯst=ï=e ‘you had’ is avoided for phonological reasons, since it would yield hḯst=e which is the 3rd singular of the Existence verb alone, i.e., ‘there is’.

ON THE CONVERGENCE OF VERBAL SYSTEMS. PART II

475

In Turkish and C. Kurdish the Set2 clitic is obligatorily fronted. In other languages—and in Mukri when it is the sole host in the clause (Öpengin 2013: 233234)—it is optionally hosted by the existence word: Tabriz Azeri (15) pul ḱi

yox-umuz=di.

money EMPH NEG.EXIST-our=COP.3SG

‘We just don’t have money’. (Kıral 2001: 17)

Leriki hič tævæqqoní=m=e. none

expectation be.NEG=1SG2=COP.3SG

‘I don’t have any expectations’. (Miller 1930: 118)

In addition, the possessor may either be marked by a case form, with the polyfunctional =ra clitic (Caucasian Tat) well-known from Persian, or remain unmarked (NENA, Central Kurdish, etc.). Tabriz Azeri Leriki (16) tæbriz-in bi=dana böüx bazar-ï var. čïmï če hæqq hés-e? Tabriz-GEN one=NC big bazaar-3SG2 there_is I.GEN what right there_is-3SG ‘Tabriz has a big bazaar’. (Kıral 2001: 164-167)

‘What right do I have?’ (Miller 1930: 121)

Juhuri (J. Caucasian Tat) (17) mɛ́=rɛ híst-i čor ʿayíl. I=RA there_is-3SG four child

Kalāsuri (Yarshater 2005: 281) hæsæn-i here=š üæ hest-æ. PN-OBL three=3SG2 egg there_is-3SG1

‘I have four children’.

‘Hassan has three eggs’.

Since the existence verb/particle is deficient in certain paradigms, as mentioned, the forms in Predicative Possession based on them are also filled in by forms of the verb ‘be’, and, depending on the language, the latter may coincide with the forms of ‘become’: (18) Present Subjunctives > ‘be’ ~ ‘become’: (a) Jewish Sanandaj Aramaic; (b) Leriki (N. Talyshi); (c) Azeri Possessive enclitic unfronted Possessive enclitic fronted (a) puḷé mən=léka hăwé-wa-li*? money from=where be(come).SUBJ.3PL1-CONV-1SG2 (b) pül čï=kovræ bḯ-bu-∅=m pul=ïm čï=kovræ bḯ-bu-∅

money from=where SUBJ-be.SUBJ-3SG1=1 SG2 money=1SG2 from=where SUBJ-be-3 SG1

476

DONALD STILO AND PAUL M. NOORLANDER

(c)

pul=ïm har-dan ol-sun*? money=1SG2 where-ABL become-OPT.3SG

‘From where would I have the money?’ (Khan 2009: 89) *(the J. Sanandaj form is Past Subjunctive < Present Subjunctive+CONV; the Azeri form is 3rd sg. Optative of ‘become’ used as ‘be, exist, have’).

5.2. ‘Have Drift’

The various have-constructions above show varying levels of grammaticalization in a process called ‘have drift’ (Stassen 2005: 475). This progression ranges from 1) a fully analytic have-construction in which the possessive clitic never attaches to the Existence particle/verb, (Standard Turkish, Kurmanji) > 2) an analytic haveconstruction in which the possessive clitic sometimes attaches to the Existence element (colloquial Azerbaijani, Northern Talyshi) > 3) a fully grammaticalized, but still not transitive, have-(pseudo-)verb, as in NENA. The NENA form is the most advanced along this progression. All varieties of Armenian and Modern Persian have a transitive have-verb but it is an innovation in Persian. Early Judaeo-Persian still had a periphrastic construction (Paul 2013: 149) while a slightly later form of the language, Early New Persian, was in the process of acquiring a have-verb via a certain fluctuation (Lazard 1963: 376-377); (former: Possessor marked by -rā + ‘be/exist’) > Possessor marked by -rā + ‘have’> unmarked Possessor + ‘have’: Early Judaeo-Persian (19) dawlat man=rā hest. there_is wealth I=RA ‘I have wealth’.

Early New Persian vai=rā ša’n-ē ʿaẓīm dāšt s/he=RA stature-INDEF great have.PST ‘He had a great imminence’.

While an independent have-verb is an innovation in Iranian, it seems to be much older in Armenian. Early contacts with Armenian (and/or Arabic) may have played a role in this have-drift in Aramaic.

6. CLITIC FLOATING 6.1. Fronting (Leftward Mobility)

Fronting of PAMs is intimately bound with stress placement, which is in turn triggered by information structure. They are part and parcel of the same phenomenon, but not always, as we see with negation in the AILA zone. At least extending consistently from Georgian through Persian, the negator is generally automatically

ON THE CONVERGENCE OF VERBAL SYSTEMS. PART II

477

stressed irrespective of information structure.11 The negator usually takes highest precedence in the sentence stress placement hierarchy; stress will move away from it only when information structure places special emphasis on another element in the clause. Thus in those dialects (and those paradigms) that license AUX fronting, the negator hosts the fronted AUX: (20) (a) Udi, (b) C. Urmia Aramaic, (c) Urmia Armenian, (d) Asālemi (S. Talyshi), (e) Kalāsuri (No. Tati) Affirmative > Negative (a) ava-za=y > té= za ava=y hikæ b-a=zu=y. NEG=1SG.EXP know-CONV what do-SUBJ=1SG=CONV + (b) +bdaya=vin-va> lé= vin-væ bdaya mudi ∅-od-in-væ. NEG=1SG.M-CONV know.INF what do-SUBJ=1SG=CONV (c) gin-æs‘=em-er > č‘-ém-er gin-æli inč‘ ∅-æn-em-er. NEG-1SG-CONV know-PTCP what do-SUBJ-1SG-CONV (d) b-æ-znost=ím > nə́=m-æ-znost číči b-æ-kær-i-m. NEG-1SG-DUR-know.PST what SUBJ-DUR-do-IMPF-1SG1 (e) be-zunest-e=u-m > né=u-m be-zunest-e not available. NEG-CONV-1SG ALLT-know-INF ‘I didn’t know what to do’. (adapted from Maysak 2008a: 128) (Asālemi and Kalāsuri do not have PCs but Asālemi does have a Past Subjunctive form; all negative sentences above have initial stress, i.e., on the negator) In the larger picture of information structure, we see that the AUX/PAMs are more widely mobile in East Armenian (and its dialects), Udi, C. Urmia Aramaic, and within Talyshi only in two of the four dialect zones of Northern Talyshi (the Lerik and Astara clusters). In these cases, leftward mobility of the AUX is triggered by information structure.

11

One often hears this incorrect copy in English with speakers of these languages when the negator is stressed in a completely neutral situation: ‘What’s his name?’ ‘I don’t know’ vs. English: ‘I don’t know’.

478

DONALD STILO AND PAUL M. NOORLANDER

(21) Chr. Urmia Aramaic; (b) Udi; (c) St. E. Armenian; (d) Leriki (a) ænæ betæ b-ríxšæ=vin ænæ bétæ=vin b-rixšæ ǽnæ=vin betæ b-rixšæ I home LOC-go=1SG I home=1SG LOC-go I=1SG home LOC-go (b) zu kua ta=z-sa* zu ḳua=z tay-sa zú=zuḳua tay-sa I home go=1SG-LOC I home=1SG go-LOC I=1SGhome go-LOC (c) yes ṭun gn-um=em yés ṭun=em gn-um yes=em ṭungn-um I home go-LOC=1SG I home=1SG go-LOCI=1SG homego-LOC (d) az bæ kæ šé-dæ=m az bæ kǽ=m še-dæ az=im bæ kæše-dæ I to home go-LOC=1SG1 I to home=1SG1 go-LOC I=1SG1 to homego-LOC ‘I am going home’

‘I am going home’

‘I am going home’

*(ta=z-sa > tassa; (b) based on Harris, n.d. and Harris 2002: 55-56) In C. Urmia Aramaic and Urmia-Khoy Armenian, the PC stays with the AUX when the latter is fronted. In the latter dialect, however, as opposed to NENA, when the main verb is stranded by the fronting of the AUX or when the AUX is a ∅ (3rd sg.), the verb changes to a participial form with no TAM: kart‘-ás=em > č-em kart‘-alí ‘I read/don’t read’. (22) Christian Urmia Aramaic

Urmia-Khoy Armenian

bə-qraya=vin-væ (unfronted)

ḳart‘-es=em-er (unfronted)

+

> xæ ktævæ=vin-væ

+

bə-qraya >

one book=1SG-CONV LOC-read.INF ‘I was reading a book’.

me xat‘ girk‘=em-er kart‘-alí one

NC

book=1SG-CONV read.PTCP

‘I was reading a book’.

The situation in Talyshi differs according to dialect. S. Talyshi, as mentioned above, has a PC identical to a Past Copula but it is only used to convert the Present copula and the Existence verb to Past and the Present Perfect to Past Perfect. Another major difference is that the AUX and the PC are not mobile in this dialect. While the AUX is quite mobile in N. Talyshi, the past equivalents do not have a PC per se, but a fusional (and suppletive) Past of the Copula/AUX. The latter is also mobile and follows the same patterns as the Present AUX: (23) Leriki (N. Talyshi) Present hand-é-dæ=m (unfronted) > az kitob=ïm hand-é-dæ I book=1SG1 read-INF-LOC ‘I am reading a book’.

Past Continuous hand-é-dæ be-m (unfronted) > az kitob be-m hand-é-dæ I book AUX.PST-1SG1 read-INF-LOC ‘I was reading a book’.

ON THE CONVERGENCE OF VERBAL SYSTEMS. PART II

479

Udi is yet again different: while mobility of the PAMs is an outstanding feature of the Udi verbal system, the PC does not follow fronted PAMs and is generally rightmost in the verbal complex (Harris 2002: 133): (24) zu ḳua=z tay-sa I home.DAT=1SG go-LOC ‘I am going home’. (Harris, n.d.)

šæhær-e=yan tay-sa=y city-DAT=1PL go-LOC-CONV ‘We were going the city’. (Maysak 2008b: 167)

The AUX in the Perfect paradigms is also regularly fronted:

(25) (a) Urmia Armenian (Asatrian 1962: 179); (b) C. Urmia Aramaic (adapted from Tsereteli 1978: 68); (c) Leriki; (d) Udi (adapted from Harris 2002: 42, 109) (a) inč‘=es ænel? (b) mudi=vit viydæ? (c) čič=ï kardǽ? (d) eḳa=nu b-e?

inč‘=es=er ænel? mudi=vit-væ viydæ? čič=ï=be kardǽ? eḳa=nu b-e=y?

‘What have you done?’

‘What had you done?’

6.2. Non-Fronting

Another striking parallelism is found in the core area: as opposed to the analytic tenses in which the AUX may move leftward (see (20) above), in all these four families the Subjunctive only has a synthetic formation (as does the Past Subjunctive + PC). As an analytic form the Subjunctive PAMs are thus suffixes and are not mobile, which means that the negator or information structure cannot attract them leftwards: Chr. Urmia Urmia Armen. No. Talyshi l-əmr-in č‘-as-em mæ-vot-ïm

7. CONCLUSIONS

Udi (Maysak 2008a: 135) ma-uḳ-a-z

In examining the verbal systems of very different language families of the AILA zone, we see that many typological parallelisms can be identified in their development. These convergences are further reinforced by a multitude of shared features in other linguistic domains such as phonology, prosody, morphology beyond the verb (which has itself certainly not been exhausted, e.g. verbal negation), syntax, etc., all of which need to be examined both in more depth and breadth. The all-pervasive structural changes within all four language families of the core area that have led to strikingly convergent verbal systems, including some ra-

480

DONALD STILO AND PAUL M. NOORLANDER

ther unusual issues, esp. those discussed in Part II, all point to a picture of language shifts in the area, most likely continually in progress somewhere on the map at different periods. On the other hand, we may never discover concrete evidence for this convergence, especially in the cases of language shift. Fortunately, however, we now have a much clearer understanding of the typological effects of substrata and language shift on the phonology and grammar of target languages. These phenomena can ultimately assist in uncovering these processes and perhaps their sources. Given the complex history of the area and the paucity of documentation of language shift due to population movements, the likely sources for these parallelisms may not be so apparent and could indeed prove to be unexpected from our present perspective. It might at first seem most obvious to attribute the primary source of influence to Iranian languages. While such may have indeed been the case, there is also the likelihood that convergence may also have flowed towards Iranian, given that the verbal systems of the local Iranian languages do not exhibit typically Iranian structures. The areas both north and south of the Araxes-Kura would surely have had a previous Udi, Armenian and very likely Neo-Aramaic presence. Some of the Caucasian Albanian community is known to have converted to Islam12 and in the process most likely shifted to various local Iranian languages, the precursors of Northern Tati, Talyshi, and Caucasian Tat. At a later stage some of these speakers most likely then shifted to Azerbaijani, contributing to the genesis of this language (see Stilo, 1994; Stilo forthcoming; Stilo in press). Those that remained Christian may have in large part moved to Armenia and those that stayed in place may have left a significant trace in today’s aberrant dialects of Armenian in the Republic of Armenia and nearby, both north and south of the Araxes. Stilo’s current research is also preliminarily pointing to the possibility of a shift of some of these same Armenian dialects to the local Iranian languages. After a long break of close to a millennium with no information on the fate of the Caucasian Albanian language, in the 19th century, when ethnographic and linguistic research began, these speakers were reduced to only two villages of Udi speakers, Vartashen and Nij. Modern-day Udi is the direct descendant of (probably a subdialect of) Caucasian Albanian.

12

‘The Armenians, Georgians, and Albanians were vassals of the Arabs rather than an integral part of the caliphate. ... Conversions to Islam were numerous in Albania and the surviving Christian population appears to have taken refuge in the mountains of eastern Armenia. The last Mihrānid, Varaz-Tiridates II, was slain in 821, after which Albania ceased to exist as a Christian political entity’. (Shopkow: 125)

ON THE CONVERGENCE OF VERBAL SYSTEMS. PART II

481

At this stage, these proposals are all conjectural. Our main point here is to suggest possibilities, some more remote than others but not infeasible. The role of language shift needs to be considered more deeply. That is, while not disregarding current (or past) language contact through bi- and multilingualism, we must also look beyond it. At any rate, the picture was certainly quite complex, possibly even more than we can imagine given the tumultuous ethnic, religious, and linguistic history of the area. As for NENA, what do we actually know about the pre-Kurdish and pre-Turkic contacts of Aramaic with languages of other Christian communities (Armenian, Caucasian Albanian, Georgian, possibly local NWI languages) and Jewish communities (Juhuri/Caucasian Tat, Persian, a myriad of NWI Central Dialects such Esfahani Jewish, Kashani Jewish, etc.), Georgian, and possibly Armenian and/or Armenian dialects)? Perhaps the contact situations for these languages were quite different from today and we may need to look more closely at the mutual interaction of Neo-Aramaic, Armenian, and Udi/Caucasian Albanian. Local Turkic varieties and their own genesis must also be added to the picture, both as sources and recipients of convergence features.

REFERENCES

Ałayan, Ēdvard. 1954. Mełru Barbaṙ. Yerevan: Haykakan SSH Gitut‘yunneri Akademiayi Hratarakch‘uty‘un. [in Armenian] Arnold, Werner. 1990. Das Neuwestaramäische V: Grammatik. Wiesbaden: Harrassowitz. Asatrian, Manvel E. 1962. Urmiayi (Xoyi) Barbaṙë. Yerevan: Yerevani Petakan Hamalsarani Hratarakch‘uty‘un. [in Armenian] Bakaev, Cherkes X. 1965. Yazyk Azerbaidžanskix Kurdov. Moscow: Nauka. Bulut, Christiane. 2006. On Culture and Language in the Turkic Region of Bayâdîstân (Habilitationsschrift). Mainz: Johannes Gutenberg-Universität Mainz. eanc: EANC: Eastern Armenian National Corpus, 2007. [online] http://www.eanc.net/. Göksel, Aslı, and Celia Kerslake. 2005. Turkish: a Comprehensive Grammar. London and New York: Routledge. Grigorian, Aharon V. 1957. Hay barbaṙagitut‘yan dasnt‘ac‘ [A Course in Armenian Dialectology]. Yerevan: Yerevani Hamalsarani Hratarakch‘uty‘un. [in Armenian] Grjunberg, Aleksandr Lenovich. 1963. Jazyk Severoazerbajdžanskix Tatov. Moscow: Izdatel’stvo Akademii Nauk SSSR. Harris, Alice C. 2002. Endoclitics and the Origins of Udi Morhosyntax. Oxford: Oxford University Press.

482

DONALD STILO AND PAUL M. NOORLANDER

—. n.d. Recipes [Udi texts, glossed and translated]. Unpublished manuscript. University of Massachusetts Amherst. Hoberman, Robert D. 1989. The Syntax and Semantics of Verb Morphology in Modern Aramaic: A Jewish Dialect of Iraqi Kurdistan. New Haven: American Oriental Society. Kıral, Filiz. 2001. Das gesprochene Aserbaidschanisch von Iran. Wiesbaden: Harrassowitz. Khan, Geoffrey. 2001. “Quelques aspects de l’expression d’“être” en néo-araméen.” In Langues de Diaspora. Langues en Contact (Faits de Langues Revue de Linguistique XVIII), edited by Anaïd Donabédian, 139–148. Paris: Ophrys. —. 2008a. The Neo-Aramaic Dialect of Barwar I: Grammar, II: Lexicon, III: Texts. Leiden: Brill. —. 2008b. The Jewish Neo-Aramaic Dialect of Urmia. Gorgias Neo-Aramaic Studies 2. Piscataway, NJ: Gorgias Press. —. 2009. The Jewish Neo-Aramaic Dialect of Sanandaj. Gorgias Neo-Aramaic Studies 2. Piscataway, NJ: Gorgias Press. —. 2011. “Remarks on Constructions with the Copula in North-Eastern Neo-Aramaic Dialects.” Aula Orientalis 29: 105–119. —. 2012. “Remarks on the Historical Development of the Copula in Neo-Aramaic.” In Dialectology Of The Semitic Languages: Proceedings Of The IV Meeting on Comparative Semitics Zaragoza 11/6-9/2010, edited by Federico Corriente et al., 25– 31. Sabadel-Barcelona: Editorial Ausa. Kurdoev, Kanat K., and Isaac I. Cukerman. 1950. “Kurdskie teksty.” In Iranskie iaazyki II, edited by V. I. Abaev, 29–60. Moscow-Leningrad: Trudy instituta yazyka i myšleniya 6. Lacroix, René. 2009. Description du dialecte laze d’Arhavi (caucasique du sud, Turquie): Grammaire et textes. PhD Dissertation in Linguistics. Université Lumière Lyon 2. Lazard, Gilbert. 1963. La langue des plus anciens monuments de la prose persane. Paris: Klincksieck. MacKenzie, David Neil. 1962. Kurdish Dialect Studies, Vol 2. London: Oxford University Press. Maysak, Teymur, 2008a. “Глагольная парадигма удинского языка (ниджский диалект)”.’ In Удинский сборник: грамматика, лексика, история языка, 96– 161. Москва: Academia. —. 2008b. „Семантика и происхождение глагольных форм настоящего и будущего времени в удинском языке.” In Удинский сборник: грамматика, лексика, история языка, 162–222. Москва: Academia.

ON THE CONVERGENCE OF VERBAL SYSTEMS. PART II

483

Miller, Boris Vsevolodovič. 1930. Talyšskie teksty. Moscow: Naučno-issledovatel’skiy institut nacional’nyx i e’tničeskix kul’tur narodov vostoka SSSR. Mkrtch‘yan, M. et al., eds. 1968. Hay žołovrdakan hek‘iat‘ner [Armenian Folk Tales], Vol. IX. Yerevan: Haykakan SSH Gitut‘yunneri Akademiayi Hratarakch‘uty‘un. Muradyan, H. V. 1960. Karčevani Barbaṙë. Yerevan: Haykakan SSH Gitut‘yunneri Akademiayi Hratarakch‘uty‘un. [in Armenian] Öpengin, Ergin. 2013. Clitic/Affix Interactions: A Corpus-based Study of Person Marking in the Mukri Variety of Central Kurdish. PhD Dissertations. Université Sorbonne Nouvelle (LACITO) and Otto-Friedrich Universität Bamberg. Połosian, A. M. 1965. Hadrut‘i Barbaṙë. Yerevan: Haykakan SSH Gitut‘yunneri Akademiayi Hratarakch‘uty‘un. [in Armenian] Sargseants, Sargis. 1883. Aguleyay Barbaṙë (Zokeri Lezun), I. Moscow: O. Gerbek. [in Armenian] Schulze, Wolfgang. 2010. “Šükürbakala P’urio – The Grateful Dead: A Vartashen Udi Folktale (Dirr 1928:60–63).” [online], http://wschulze.userweb.mwn.de/ WS0910/ danktot1.pdf. Shopkow, Leah. 1982. “Albanian (Caucasian).” In Dictionary of the Middle Ages, Volume 1, edited by Joseph R. Strayer, 123–125. New York: Charles Scribner’s Sons. Stassen, Leon. 2005. “Predicative Possession.” In The World Atlas of Language Structures, edited by Martin Haspelmath, Matthew S. Dryer, David Gil and Bernard Comrie. Oxford: Oxford University Press. Stilo, Donald L. 2008. “Two Sets of Mobile Verbal Person Agreement Markers in the Northern Talyshi Language.” In Aspects of Iranian Linguistics: Papers in Honor of Mohammad Reza Bateni, edited by Simin Karimi et al., 363–390. Newcastleupon-Tyne: Cambridge Scholars Press. —. 2001. “Gilan, x. Languages.” In Encyclopaedia Iranica, Volume X, edited by Ehsan Yarshater, 660–668. New York: Bibliotheca Persica Press. —. 1994. “Phonological Systems in Contact in Iran and Transcaucasia.” In Persian Studies in North America: Studies in Honor of Mohammad Ali Jazayery, edited by Mehdi Marashi, 75–94. Bethesda, Md: Iranbooks. —. forthcoming. “On the non-Persian Iranian Substratum of Azerbaijan.” —. in press. “Further Notes on the Iranian Substratum of Azerbaijani Turkish.” —. in progress. Atlas of the Araxes-Iran Linguistic Area. Tsereteli, Konstantin G. 1978. The Modern Assyrian Language. Moscow: Nauka. Yarshater, Ehsan. 2005. “Tāti Dialect of Kalāsur.” In Languages of Iran: Past and Present, edited by Dieter Weber, 269–284. Iranica 8. Wiesbaden: Harrassowitz. Any data without citations are derived from Don Stilo’s field notes.

484

DONALD STILO AND PAUL M. NOORLANDER

Adapted examples that were added to an original example in other languages for comparative purposes were all constructed by Stilo to the best of his ability following the rules of the languages as found in the relevant grammars. Any errors in morphological forms are all his personal responsibility. Small errors, however, are unlikely to change the basic premise of the parallelisms or near-isomorphic correspondences of these examples across the languages of the four families introduced here.

INDEX General index active participle, 4, 164, 168, 197–200, 202–203, 281, 324–325, 432, 435 actual present, 47, 49–50, 252–253, 258, 260–261, 264–268 alignment, 11, 232–237, 239–240, 243, 245, 247 anticausative, 207, 209, 212–214, 219, 223–224, 229 Arabic loanword, 7, 19, 21–22, 70, 73, 76, 79–92, 94–95, 101–103, 105, 108, 114, 128, 330, 371, 377, 387, 399 areal features, 187, 203, 216, 246, 248, 421–422, 429, 441 aspect, 2–3, 47, 51, 60–61, 163, 207– 208, 403, 426, 431, 435 back vowel, 119, 124, 128, 131, 133, 135, 137–138, 142, 294 borrowing, 69–73, 75–80, 82–88, 90– 93, 96, 100, 104, 107, 386, 410, 412 causative, 153, 207–209, 212–223, 225–230 clitic, 3, 157, 160, 193, 242, 307–308, 311, 412, 418–419, 433, 441, 446, 456, 474–476 clitic floating, 476 colophon, 373, 376, 380, 382–386, 393 concomitance, 193, 200, 202 conditioning factors, 113, 123

convergence, 427, 440, 447–448, 480– 481 convergent verbal systems, 428, 479 denominative, 88–89, 91–94, 102, 104, 220, 378 devoicing, 157–160 diaspora, 30, 79–80, 89, 93, 101–102, 104–105, 306 discourse grammar, 53–54, 59–60, 67 Doctrine of John, 397–400, 402, 404– 405 emphasis, phonological, 90, 130–132, 134–146, 155, 157–160, 293, 293– 295, 297, 308, 311 emphatic /ṛ/, 110–117, 119–120, 122– 123, 125–128 future tense, 162, 401 genitive construction, 316 gerund, 200–202, 253, 255–256, 260, 265, 268, 277–279, 300–301 goals, 412–423 grammaticalization, 155, 163, 174, 176–177, 193, 203, 251, 253, 260– 262, 279, 315, 332, 414, 420, 426, 432–437, 439, 442–443, 446, 455– 456, 458, 476 have drift, 476 Hertevin paradigm, 235, 243–246

486 imperfect, 3, 162–164, 190, 195, 252, 324, 399–405, 430, 432, 436–437, 453, 456–457, 459–462, 464–465, 467 imperfective, 61, 167, 262, 379, 404– 405, 429, 431–432 inchoative, 207–217, 219–223, 225– 230 independent pronoun, 54, 60 infinitive, 2, 204, 279, 313, 433 intransitive, 1, 4, 8, 13, 18, 22–23, 198, 207–208, 210, 212, 217–218, 220– 221, 223, 230, 232–234, 239, 275, 278, 281, 283–284, 320, 474 junctural doubling, 179 kal copula, 38, 40, 45–46, 50 kal particle, 31–32, 38–39, 41, 48 kit copula, 32–35, 39, 50 Kurdish loanword, 7, 102, 120, 122, 128, 407 language shift, 23, 427–428, 447–448, 480–481 Leiden Glossarium, 370, 373–379, 384, 386, 393 metanalysis, 179, 181 middle voice, 209 minimal pair, 110–112 modal marker, 276 morphological boundary, 147, 151, 157–159 narrative techniques, 53, 60–61, 64, 174 native word, 80, 82, 86, 89–90, 94–95, 97, 102, 308 neologism, 77–79, 91, 100–103, 105, 107–108, 372 now-related tense, 265 oral literature, 2, 347, 350, 352, 364

INDEX participle, 2, 4, 47, 176, 178–179, 200, 204, 266, 287, 299, 324–325, 391, 399, 433, 443, 455 passive, 2, 4, 23, 26, 207–211, 217, 225–228, 267–268, 276, 278, 281, 284, 325–326, 427 past converter, 277, 454–465, 469–473, 478–479 past stem, 276, 454, 457, 459–460 past tense, 32, 243, 324 perfect, 22, 174–175, 178–179, 181– 183, 236, 252–256, 261, 264–270, 275–276, 299, 349, 399–401, 430, 454, 469–470, 478–479 perfective, 208, 232–234, 314, 439 periphrastic present-future, 399, 402 periphrastic preterite, 172, 405 Persian loanword, 70, 78, 81, 84, 86– 87, 89, 92, 328, 330, 347, 387 poetry, 331, 348, 353, 371, 390, 392 predicative possession, 454, 473, 475, 483 present continuous, 187–191, 193, 195– 197, 199–200, 202–203, 275, 324 present perfect, 252, 256, 265, 267– 268, 276 present progressive, 167, 260, 262, 265, 310, 313, 455 present stem, 276, 442, 461–462, 465 present tense, 21, 32, 60, 243, 245–246, 428, 434, 438, 441, 446 presentatives, 250–251, 254, 256, 262– 263, 266, 268 preterite, 2–5, 7, 9, 11, 13, 17, 21–22, 25–27, 172, 174, 177, 183, 245, 269, 314, 320, 400–401, 439, 458– 462, 464–465, 470 l-preterite, 16, 26 qatǝl-shaped preterite, 7–8, 18

INDEX preverbal particle, 156, 172, 174, 178, 180–181, 188, 190, 193, 195–197, 239, 435 progressive, 70, 155–156, 159, 254– 256, 258, 260–262, 265, 267–268, 275–276, 290, 313, 432, 435–437, 441 proverbs, 331, 334–335 pseudo-relative, 193, 196–197, 199– 203, 205 purists, 70, 76–79, 81, 83, 85, 89, 92, 95, 100, 105, 108 qatǝl, 1, 7–9, 11 qāṭil, 399, 402 qṭil l-, 399 qtīl, 6, 23 rāwē, 348–352, 354–355, 363 relative particle, 175, 193–194, 196 relative verb, 199 retraction of the tongue root, 132, 134, 145 retroflex /ɽ/, 110–114, 116–125, 127– 128 retroflex, 110–114, 116–117, 119–128 root, 2–7, 16, 21–22, 26, 70–74, 77–78, 80–82, 85, 87–92, 93–96, 101, 103– 108, 132, 134, 159, 176–177, 198, 220, 276, 279, 308, 312, 323, 327– 328, 348, 374, 378–379, 384, 387, 389, 435, 441, 443, 456, 457, 466– 467 split ergativity, 2, 8, 232–233, 235, 237 split-S, 233, 246–247 subjunctive, 172, 174, 176–177, 179, 428–435, 437–440, 444–447, 453, 463–468, 476–477, 479 substrate, 133, 441, 448, 451, 480, 483

487 syllable, 6, 74, 77–78, 102, 105–106, 132, 138, 142, 148–155, 157–160, 188, 279, 290, 298, 307–308, 439, 464 tense, 2–3, 22, 31, 47, 163, 165–169, 175, 177, 181, 194, 232–233, 239– 240, 245–246, 248, 252, 254, 260, 264–267, 273–274, 276–278, 307– 308, 310, 324–325, 401–402, 426, 428, 431–432, 434, 440–441, 443– 444, 446–447, 454, 456–457, 459– 462, 466, 468, 469 bifurcate, 274–275, 277 compound, 256 doubly compound, 275 simple, 252, 274–275 textual theme, 34, 36, 38–39, 44, 46– 48, 50 three /R/s, 110 translational errors, 323 Turkish loanword, 79, 81–83, 85, 88, 90, 96, 108 typology, 235, 407–408, 412–413, 429, 434, 439, 442, 444, 445, 447, 458, 460, 465 unaspirated stops, 130, 138 Verb Glossary, 1, 3 verbum sentiendi, 38–39, 42, 46–47, 49– 50 verbal particles, 252 voice, 4–5, 37–38, 101, 103, 207, 209– 210, 325–326, 340–343, 356 vowel harmony, 131, 139 word order, 43, 407–410, 412–414, 422–423 Yosef ve-ʾEḥav ‘Joseph and his Brothers’, 331

488

INDEX

Index of languages, personal and geographical names Amidya, 98, 270, 322, 327, 330 Anatolia, 409, 418, 420–422, 424 ʿAnkawa, 187–191, 193, 195–199, 203 Arabic, 2, 26, 69–82, 84–88, 90, 93, 95, 97–103, 105–109, 111, 114, 116, 121, 129, 131, 138, 153, 161, 168, 174, 176–177, 189, 197, 203, 206, 222, 231, 251, 258, 330, 336, 348, 365, 367, 371, 378–379, 383, 386– 387, 389–390, 395–396, 399, 426, 435–436, 451, 473, 476 Arabic dialects, 6–70, 72, 74, 90, 92, 97, 100, 117, 189 Armenian dialects, 427, 429, 444, 448, 453–454, 456, 458, 480–481 Assyrian Kirkuk, 207–211, 215–218, 220–225, 227–228, 230 Assyrian, 30, 109, 124, 144–145, 161, 207–211, 215–218, 220–225, 227– 228, 230, 279, 290–291, 304, 319, 321, 345–347, 349, 351, 363–366, 452, 483 Assyrians, 290–291, 345–353, 361, 365–366 Baṛṭəḷḷa, 110–112, 114–115, 117 Barwar, 68, 110, 126–129, 142, 180, 184, 189, 200, 205, 208–211, 216, 218–223, 227, 230, 232, 234, 238– 241, 246, 249, 262, 269, 313, 321, 410–411, 424, 445, 450, 455–457, 460, 482 Biblical Aramaic, 175, 181, 322–330, 374, 399, 406 Bohtan, 165–166, 168–170, 173, 211, 230, 232, 239, 242, 246, 248, 409– 410, 412, 424 Chaldean, 109, 185, 187, 305, 350, 365

Chaldeans, 187, 365–366 Christian Diyana-Zariwaw, 130–143 Christian Salmas, 289–293, 295–303 Christian Urmi, 30, 145, 173, 231, 290, 295, 298, 427, 435, 441, 445, 457, 459, 462, 464, 470, 473, 478 Classical Aramaic, 174, 233 Classical Mandaic, 168, 370, 374–375, 382–383, 386, 393, 399, 400, 405 Dere, 110, 116, 125–126, 128 Drower, Ethel S., 175, 182, 371, 374, 376–377, 380, 383–386, 388–392, 394–396, 398, 406 France, 305–306, 317, 320, 381, 394 Gaznax, 305–319 Hakkari, 289–290, 295, 298, 306, 317, 321, 345–348, 351, 361, 365, 422 Hebrew, 1, 11, 52, 82, 94, 100, 144, 178, 182–183, 185–186, 212–213, 254–255, 263, 266, 268, 281, 287, 291, 304, 323, 325, 327, 329–330, 336, 343–344, 396, 426, 448–451 Jewish Azerbaijan dialect, 290, 293, 299 Jewish Babylonian Aramaic, 22, 163– 164, 167, 432–433 Jewish dialect of Zakho, 271, 320 Jewish Salmas, 289–303 Jewish Sanandaj, 137, 142, 208, 233, 236, 456, 457, 467, 475 Jewish Urmi, 135–136, 139, 141, 147, 165, 173, 208, 233, 242–243, 290, 293–294, 435, 442, 445, 457, 460, 462 Jewish Zakho, 138, 179, 250, 253–254, 256, 262, 268, 331–333, 334, 336, 456

INDEX Judi, 305–306, 308–314, 316, 320 Kurdish, 7, 69–79, 81–82, 84, 86–93, 96–98, 111, 113, 115–116, 120–121, 125, 128, 133, 141, 144, 167, 170, 187, 189, 190, 193–196, 203, 205, 232, 272–273, 276–277, 286, 288, 291, 308, 317, 325, 327–328, 330, 336, 340, 342, 346–347, 353, 407– 410, 412–413, 414, 416–422, 424, 427–428, 437, 450–451, 453–454, 461, 471, 473–475, 481–483 Kurmanji, 115, 128, 144, 187, 189–190, 193, 195–196, 205, 273, 276–277, 281, 283, 285–288, 408, 418, 436, 441, 444, 446, 453, 461, 470–471, 474, 476 Late Aramaic, 172, 177, 181–182, 399, 433, 455 Macuch, Rudolf, 164, 168, 170, 174– 175, 178, 182, 184, 367, 369, 371, 373–386, 388, 390–395, 401, 406 Mandaic, 1, 163–164, 168, 170, 173– 177, 182, 185, 289, 296–297, 299– 300, 367–383, 386–390, 392–396, 399–403, 405–406, 435 Middle Aramaic, 163–164, 169, 183, 426 Morgan, Jacques de, 380–382, 393–394 Mosul, 98, 109–111, 173–174, 182, 185, 188, 290, 347, 351, 354–356, 416 NENA, 1, 2, 16, 32, 110–111, 114, 125, 128, 130–133, 134, 137–138, 141– 143, 145, 154, 164–170, 172–177, 179–181, 183, 187–190, 193–198, 200, 202–203, 208, 211, 213–215, 217–218, 221–224, 228–230, 232– 233, 235–237, 239, 242–248, 250, 269, 289–290, 295, 298–299, 301, 303, 315, 407–412, 418, 421–423,

489 426–428, 431, 433–435, 438, 441– 442, 444–445, 449, 453, 455–457, 459, 465–466, 470–471, 474–476, 478, 481 Neo-Aramaic, 1–2, 27, 29–31, 33, 47, 51–54, 68, 71, 74–75, 77, 98, 100, 110–111, 129–130, 143–145, 160– 163, 168, 170, 172, 174–178, 182– 185, 187, 193, 195, 203–205, 207– 209, 230–232, 236, 245–246, 248– 250, 252–253, 260, 268–273, 276– 277, 281, 283, 287, 289–291, 293, 298–299, 303–305, 320–322, 330– 331, 333, 336, 343–344, 365–367, 392, 399, 405, 407, 423–424, 426, 429, 433, 448–452, 455, 480–482 NENA ethnolects, 290 Neo-Mandaic, 168, 289, 296, 367–369, 371, 373–375, 377, 379, 382–394, 399, 405 Nineveh Plains, 350 Nöldeke, Theodore, 22, 28, 164, 170, 173–174, 176, 178, 185, 198, 205, 249, 370–371, 373–375, 382–383, 386, 391, 393, 396, 399, 406, 433, 435, 445, 451 Ṣūrayt/Ṭūrōyo, 69–97, 100–109 Sureth, 122, 304, 321, 344 Syriac, 3, 9, 22, 27–28, 70, 72–80, 82– 85, 87–89, 91–92, 94–99, 101–109, 123, 143–144, 163–164, 168, 170, 175–176, 178, 183–186, 205, 270– 271, 273, 283, 288, 305, 329–330, 366–367, 374, 400–403, 406, 432– 433, 435, 451, 455 Talyshi, 427–428, 434, 441–442, 444, 446–447, 451, 453–455, 457–462, 464–467, 469–472, 474–480, 483 Trans-Zab, 136, 173, 176, 185, 289– 290, 292, 298, 300, 304

490 Turkey, 30, 56, 72–73, 76–80, 82–84, 89–91, 93–94, 96, 102, 105, 107, 131, 141–143, 173, 189, 197, 232, 248, 305, 308, 321, 345, 407, 427, 437, 444, 462 Turkish, 7, 69–98, 101, 108, 131, 133, 139, 141, 254, 270, 290, 305, 308, 318, 330, 347, 360, 366, 377, 410, 422, 444, 448, 451, 454, 462, 471, 475–476, 481, 483 Ṭuroyo, 1–10, 22–23, 26–27, 29–33, 51, 53–54, 60, 62, 67–68, 168–169, 173, 177–178, 251–252, 299–300, 435, 452, 455 Udi, 427, 429–431, 433, 438, 440–441, 443–444, 447–448, 453, 456, 459– 460, 463–466, 468–474, 477–483 Werner, Arnold, 110, 112–114, 117, 119, 120, 129, 183, 269, 287, 289, 449, 481 Western Syriac, 78, 81, 83–84, 95–96, 100–104, 106, 108–109 Zakho, 52, 98, 184, 247, 250, 256, 268, 270, 272, 320, 330–331, 333, 336, 338, 343–344, 415, 419, 436, 454

INDEX