Semitic Languages in Contact 9004300155, 9789004300156

Citation preview

Studies in Semitic Languages and Linguistics Editorial Board A. D. Rubin and C. H. M. Versteegh

VOLUME 82

The titles published in this series are listed at brill.com/ssl

Semitic Languages in Contact Edited by

Aaron Michael Butts

LEIDEN | BOSTON

Library of Congress Cataloging-in-Publication Data Semitic languages in contact / edited by Aaron Michael Butts.   pages cm. — (Studies in Semitic languages and linguistics; Volume 82)  Includes bibliographical references and indexes.  ISBN 978-90-04-30014-9 (hardback : alk. paper) — ISBN 978-90-04-30015-6 (e-book) 1. Semitic languages—Grammar, Comparative. 2. Semitic languages—Grammar. 3. Languages in contact. I. Butts, Aaron Michael, editor. PJ3021.S48 2015 492’.045—dc23 2015024968

This publication has been typeset in the multilingual “Brill” typeface. With over 5,100 characters covering Latin, ipa, Greek, and Cyrillic, this typeface is especially suitable for use in the humanities. For more information, please see www.brill.com/brill-typeface. issn 0081-8461 isbn 978-90-04-30014-9 (hardback) isbn 978-90-04-30015-6 (e-book) Copyright 2015 by Koninklijke Brill nv, Leiden, The Netherlands. Koninklijke Brill NV incorporates the imprints Brill, Brill Hes & De Graaf, Brill Nijhoff, Brill Rodopi and Hotei Publishing. All rights reserved. No part of this publication may be reproduced, translated, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without prior written permission from the publisher. Authorization to photocopy items for internal or personal use is granted by Koninklijke Brill nv provided that the appropriate fees are paid directly to The Copyright Clearance Center, 222 Rosewood Drive, Suite 910, Danvers, ma 01923, usa. Fees are subject to change. This book is printed on acid-free paper.

Contents Preface vII List of Figures xi Abbreviations xii Contributors xx A Thamudic B Abecedary in the South Semitic Letter Order 1 Ahmad Al-Jallad and Ali Al-Manaser Ethiopian Semitic and Cushitic. Ancient Contact Features in Ge‘ez and Amharic 16 David Appleyard Hebrew Adverbialization, Aramaic Language Contact, and mpny ʾšr in Exodus 19:18 33 Samuel Boyd and Humphrey Hardy The Distribution of Declined Participles in Aramaic-Hebrew and Hebrew-Aramaic Translations 52 Yochanan Breuer The Proto-Semitic “Asseverative *la-” and the Innovative 1sg Prefixes in South Ethio-Semitic Languages 68 Maria Bulakh Egyptianizing Features in Phoenician and Punic Inscriptions from Egypt 97 David Calabro Head-Marking in Neo-Aramaic Genitive Constructions and the ezafe Construction in Kurdish 114 Eran Cohen Notes on Foreign Words in Hatran Aramaic 126 Riccardo Contini and Paola Pagano Language, Writing, and Ideologies in Contact: Sumerian and Akkadian in the Early Second Millennium bce 158 C. Jay Crisostomo

vi

contents

Inner-Semitic Loans and Lexical Doublets vs. Genetically Related Cognates 181 Lutz Edzard Structural Change in Urban Palestinian Arabic Induced by Contact with Modern Hebrew 198 Uri Horesh Language Contact as Reflected in the Consonant System of Ṭuroyo 234 Otto Jastrow Lexical Borrowings in the Eastern European Hasidic Hebrew Tale 251 Lily Kahn Possible Ugaritic Influences on the Hurrian of Ras Shamra-Ugarit in Alphabetic Script 267 Joseph Lam The Lexical Component in the Aramaic Substrate of Palestinian Arabic 280 Mila Neishtadt The Classification of Hobyot 311 Aaron D. Rubin Expression of Attributive Possession in Tunisian Arabic: The Role of Language Contact 333 Lotfi Sayahi Aramaic Loanwords in Gǝʿǝz 348 Jürgen Tubach Language Contact between Akkadian and Northwest Semitic Languages in Syria-Palestine in the Late Bronze Age 375 Juan-Pablo Vita Semitic Languages in Contact—Syntactic Changes in the Verbal System and in Verbal Complementation 405 Tamar Zewi and Mikhal Oren Index 423

Preface Over the past several decades, contact linguistics has grown into a discrete field within linguistics. There were certainly earlier publications that laid the foundation for this, probably none more important than U. Weinreich Languages in Contact. Findings and Problems (New York, 1953).1 The late 1980s, however, witnessed the publication of several studies that would ultimately lead to the development of contact linguistics into a field of research in its own right. Undoubtedly the most important of these was S. Thomason and T. Kaufman’s Language Contact, Creolization, and Genetic Linguistics (Berkeley, 1988). By analyzing a broad range of contact situations, Thomason and Kaufman were able to develop a general theory of language contact that remains influential today.2 It is no exaggeration to say that the study by Thomason and Kaufman served as one of the main catalysts for the explosion of publications on language contact witnessed over the past several decades.3 This growing body of literature, as well as the scholars responsible for producing it, have helped to transform what was once a subfield within historical linguistics into a new field of research: contact linguistics. The establishment of this field can be illustrated by the creation of new journals, such as Bilingualism: Language and Cognition (since 1998) and The Journal of Language Contact (since 2007/2008), and new book series, such as Studies in Language Contact (Oxford University Press, since 1990) and Approaches to Language Contact (Cambridge University Press, since 2011), as well as by the publication of The Handbook of Language Contact (Hickey 2010a). Following this development with contact linguistics, Semitic linguistics has in recent years started to pay more attention to language contact. S. Weninger’s recent handbook on the Semitic languages (2011), for instance, contains individual chapters on a number of contact situations involving Semitic languages, including Akkadian and Sumerian, Akkadian and Aramaic, Aramaic and Iranian, Aramaic and Arabic, Arabic and Berber, Arabic and Persian, Arabic and Modern European Languages, and Ethiopian Semitic and Cushitic. This 1 This is not to minimize earlier studies, such as two important articles by Haugen (1950a, 1950b), or the even earlier works by Whitney (1881), Schuchardt (1884), and Windisch (1897). 2 The year of 1988 also saw the publication of F. Van Coetsem’s Loan Phonology and the Two Transfer Types in Language Contact (Dordrecht, 1988). Though overshadowed by Thomason and Kaufman’s classic, this study is now gaining traction as presenting a more robust theory of language contact, at least to some, including the editor of this volume (see Butts Forthcoming: Ch. 2) as well as, for instance, Winford (2005; 2007). 3 For a recent survey of this research, see Hickey 2010b.

viii

preface

can be contrasted with earlier handbooks, such as the one edited by Hetzron (1997), which does not include substantial discussions of language contact in its chapters on particular languages much less present individual chapters dedicated specifically to language contact. Or, to take another example, a recent issue of The Journal of Language Contact (6.2 [2013]), dedicated to contact among genetically related languages, contains several articles dealing with Semitic languages (Al-Jallad 2013; Beaulieu 2013; Pat-El 2013), and two of the three guest editors of the issue are Semitists, John Huehnergard and Na’ama Pat-El (the third editor, Patience Epps, specializes in Amazonian languages). As one final example, over the last couple of decades, a number of monographs have appeared dealing with language contact involving Semitic languages, including studies on contact between Arabic and Berber (Sayahi 2014), Egyptian and Northwest Semitic (Muchiki 1999), Akkadian and Biblical Hebrew (Mankowski 2000), Akkadian and West Semitic (Pentiuc 2001), Syriac and Iranian (Ciancaglini 2008), and Syriac and Greek (Butts Forthcoming), to name only a few. Despite this growing body of literature, there has not until now been a collection of studies devoted to language contact involving Semitic languages.4 The present volume contains twenty individual case studies on Semitic language contact. The languages treated span from ancient Semitic languages, such as Akkadian, Aramaic, Gəʿəz (Classical Ethiopic), Hebrew, Phoenician, and Ugaritic, to modern ones, including languages/dialects belonging to the Modern Arabic, Modern South Arabian, Neo-Aramaic, and Neo-Ethiopian branches of the Semitic family. The topics discussed involve various aspects of language, including writing systems, phonology, morphology, syntax, and lexicon. The approaches range from what might be termed traditional philology to more theoretically-driven linguistics. Despite this diversity, the studies in this volume are united by the theme of language contact. This volume is not intended to be comprehensive or exhaustive. Every reader will undoubtedly want studies on additional topics or contact situations and employing different methodologies and theories. Even the editor would have loved to have included so many more studies, whether on diglossia in Arabic, the influence of Aramaic on Akkadian, code-switching in Modern Hebrew, or contact features in Mehri due to Arabic. The limitations of space do not, however, permit the inclusion of studies on every aspect of language contact involving a Semitic language within a single volume. Though this volume does not aim to be comprehensive or exhaustive, it is meant to be representative of 4 See, however, Crass and Meyer 2009 for a collection of studies on contact-induced change in Ethiopian Semitic.

preface

ix

the range of contact situations involving Semitic languages that are attested as well as of the types of approaches by which these contact situations are currently being analyzed. In this way, the volume seeks to provide the status quaestionis of the study of language contact among the Semitic languages. It is hoped that this snapshot of the current state of the field will encourage new research on Semitic language contact. Aaron Michael Butts Bibliography Al-Jallad, A. 2013. “Arabia and Areal Hybridity,” Journal of Language Contact 6: 220–242. Beaulieu, P.-A. 2013. “Aspects of Aramaic and Babylonian Linguistic Interaction in First Millennium BC Iraq,” Journal of Language Contact 6: 358–378. Butts, A. M. Forthcoming. Language Change in the Wake of Empire: Syriac in its GrecoRoman Context (LSAWS 11). Eisenbrauns. Ciancaglini, C. A. 2008. Iranian Loanwords in Syriac. Wiesbaden. Crass, J. and R. Meyer. 2009. Language Contact and Language Change in Ethiopia. Köln. Haugen, E. 1950a. “Problems of bilingualism,” Lingua 2: 271–290. (= Haugen 1972: 59–78) ———. 1950b. “The Analysis of Linguistic Borrowing,” Language 26: 210–231. (= Haugen 1972: 79–109) ——— (selected and introduced by A. S. Dil). 1972. The Ecology of Language. Essays by Einar Haugen. Stanford. Hetzron, R. 1997. The Semitic Languages. London. Hickey, R. 2010a. The Handbook of Language Contact. Chichester, West Sussex. ———. 2010b. “Language Contact: Reconsideration and Reassessment,” in Hickey 2010a: 1–28. Mankowski, P. V. 2000. Akkadian loanwords in Biblical Hebrew (HSS 47). Winona Lake. Muchiki, Y. 1999. Egyptian Proper Names and Loanwords in North-West Semitic. Atlanta. Pat-El, N. 2013. “Contact or Inheritance? Criteria for distinguishing internal and external change in genetically related languages,” Journal of Language Contact 6: 313–328. Pentiuc, E. J. 2001. West Semitic Vocabulary in the Akkadian Texts from Emar (HSS 49). Winona Lake. Sayahi, L. 2014. Diglossia and Language Contact. Language Variation and Change in North Africa. Cambridge. Schuchardt, H. 1884. Slawo-Deutsches und Slawo-Italienisches. Graz.

x

preface

Thomason, S. G. and T. Kaufman. 1988. Language Contact, Creolization, and Genetic Linguistics. Berkeley. Van Coetsem, F. 1988. Loan Phonology and the Two Transfer Types in Language Contact (Publications in Language Sciences 27). Dordrecht. Weinreich, U. 1953. Languages in Contact. Findings and Problems. New York. Weninger, S. (in collaboration with G. Khan, M. P. Streck, and J. C. E. Watson). 2011. The Semitic Languages. An International Handbook. Berlin. Whitney, W. D. 1881. “On Mixture in Language,” in Transactions of the American Philological Association (1869–1896) 12: 5–26. (Reprinted in M. Silverstein [ed.]. 1971. Whitney on Language. Selected Writings of William Dwight Whitney. Cambridge, MA. 170–191.) Windisch, E. 1897. “Zur Theorie der Mischsprachen und Lehnwörter,” Berichte über die Verhandlungen der Königlich Sächsischen Gesellschaft der Wissenschaften zu Leipzig. Philologisch-historische Classe 49: 101–126. Winford, D. 2005. “Contact-induced changes. Classification and processes,” Diachronica 22.2: 373–427. ———. 2007. “Some issues in the study of language contact,” Journal of Language Contact THEMA 1 (2007): 22–39.

List of Figures 1 2 3 4 5 6 7 8

Photograph of the basalt slab bearing three Thamudic B inscriptions 15 Hand copy of the three Thamudic B inscriptions 15 Hand copies of Phoenician graffiti from the temple of Seti I at Abydos 102 Variations in the oath formula 172 Cross-tabulation of (ʕ) by speaker, community, and 2 types of lenition 208 Example of emphatic misspelled as non-emphatic by native speaker 215 Carthamus glaucus 301 Centaurea pallescens 305

Abbreviations General Akk. Akkadian Amh. Amharic ANA Ancient North Arabian Ar. Arabic Arg. Argobba Arm. Armenian ASA Ancient South Arabian BA Biblical Aramaic Bab. Babylonian BH Biblical Hebrew Čah. Čaha CA Classical Arabic CPA Christian Palestinian Aramaic Dob. Dobbi ENA Eastern Neo-Aramaic End. Endegañ Enm. Ennemor ES Ethiopian Semitic Gaf. Gafat Gr. Greek HA Hatran Aramaic Har. Harari Ḥars. Ḥarsusi Has. Hasaitic Heb. Hebrew Hob. Hobyot HP Sefer Halachot Pesuqot HR Halaḵot Pesuqot ʾo Hilḵot Reʾu JBA Jewish Babylonian Aramaic Jib. Jibbali JPA Jewish Palestinian Aramaic JudA Judean Aramaic Kst. Kestane Lat. Latin LB Late Babylonian LBH Late Biblical Hebrew Lih. Lihyanite

abbreviations

MA Middle Assyrian Md. Mandaic Med. Median MIr. Middle Iranian MSA Modern South Arabian ms(s). manuscript(s) Msq. Masqan Muh. Muher NA Neo-Aramaic / Neo-Assyrian Nab. Nabatean NB Neo Babylonian NENA North-Eastern Neo-Aramaic NP Neo-Persian OA Old Assyrian OAkk. Old Akkadian OAram. Old Aramaic OAr. Old Arabic OB Old Babylonian OfA Official Aramaic OIr. Old Iranian OM Omani Mehri OP Old Persian OSyr. Old Syriac PA Palestinian Arabic Palm. Palmyrene Parth. Parthian PS Proto-Semitic QA Qumran Aramaic QH Qumran Hebrew SA Samaritan Aramaic Sab. Sabaic Saf. Safaitic SBH Standard Biblical Hebrew Sel. Selṭi SES South Ethio-Semitic SP Sumerian Proverbs Sum. Sumerian Syr. Syriac WNA Western Neo-Aramaic Wol. Wolane YM Yemeni Mehri

xiii

xiv

abbreviations

Bibliographic AELAC ÄF AfO AHw AION ALASP ANES AO AOAT AoF APAL AS AuOr BICs BiOr BK BASOR BSOAS CAD CHANE CGC CRAIBL CSCO DJA DJBA DJPA DNWSI DSA

Association pour l’étude de la littérature apocryphe chrétienne Äthiopistische Forschungen Archiv für Orientforschung Soden, W. von 1965–81. Akkadisches Handwörterbuch, Vol. I–III. Wiesbaden. ΑΙΩΝ. Annali del Dipartimento di Studi del Mondo Classico e del Mediterraneo Antico Abhandlungen zur Literatur Alt-Syrien-Palästinas Ancient Near Eastern Studies Aula Orientalis Alter Orient und Altes Testament Altorientalische Forschungen Annual Publication in African Linguistics Assyriological Studies Aula Orientalis Bulletin of the Institute of Classical Studies Bibliotheca Orientalis Biberstein Kazimirski, A. De. 1860. Dictionnaire arabe-français, Vol. I–II. Paris. Bulletin of the American Schools of Oriental Research Bulletin of the School of Oriental and African Studies Gelb, I. J. et al. 1956–2010. The Assyrian Dictionary of the Oriental Institute of Chicago. Chicago. Culture and History of the Ancient Near East Catalogue général des antiquités égyptiennes du Musée du Caire Comptes rendus (des séances) de l’Académie des Inscriptions et BellesLettres / Académie des Inscriptions et Belles-Lettres. Comptes rendus. Corpus Scriptorum Christianorum Orientalium Sokoloff, M. 2003. A Dictionary of Judean Aramaic. Ramat-Gan. Sokoloff, M. 2002. A Dictionary of Jewish Babylonian Aramaic of the Talmudic & Geonic periods. Ramat-Gan. Sokoloff, M. 2002. A Dictionary of Jewish Palestinian Aramaic of the Byzantine period (2nd ed.). Ramat-Gan. Hoftijzer, J. and Jongeling, K. 1995. Dictionary of North-West Semitic inscriptions (HdO I/21), Vol. 1–2. Leiden. Tal, A. 2000. A dictionary of Samaritan Aramaic, Vol. I–II (HdO I/50). Leiden.

abbreviations

DULAT

xv

del Olmo Lete, G. and J. Sanmartín. 2003. A Dictionary of the Ugaritic Language in the Alphabetic Tradition. Leiden – Boston. EA Études assyriologiques EALL Versteegh, Kees. 2005–2009. Encylopedia of Arabic Language and Linguistics (5 vol.). Leiden. EHLL Khan, Geoffrey (ed.), with Shmuel Bolozky, Steven Fassberg, Gary A. Rendsburg, Aaron D. Rubin, Ora R. Schwarzwald, Tamar Zewi. 2013. Encyclopedia of Hebrew Language and Linguistics (4 vol.). Leiden. EI2 Encyclopaedia of Islam (2nd ed.) ELO Elementa Linguarum Orientis GEDSH Brock, S. P., A. M. Butts, G. A. Kiraz, L. Van Rompay (eds.). 2011. Gorgias Encyclopedic Dictionary of the Syriac Heritage. Piscataway. GLECS Groupe Linguistique d’Études Chamito-Sémitique HALOT Köhler, L., W. Baumgartner, J. J. Stamm, et al. (trans. M. E. J. Richardson). 2001. The Hebrew and Aramaic Lexicon of the Old Testament. Study Edition. Leiden. HdO Handbuch der Orientalistik HSK Handbücher zur Sprach- und Kommunikationswissenschaft HSS Harvard Semitic Studies Hugoye Hugoye: Journal of Syriac Studies IEJ Israel Exploration Journal IF Indogermanische Forschungen IOS Israel Oriental Studies IOS Israel Oriental Studies JA Journal asiatique JAOS Journal of the American Oriental Society JCS Journal of Cuneiform Studies JANER Journal of Ancient Near Eastern Religions JES Journal of Ethiopian Studies JJS Journal of Jewish Studies JNES Journal of Near Eastern Studies JNSL Journal of Northwest Semitic Languages JQR The Jewish Quarterly Review JSAI Jerusalem Studies in Arabic and Islam JSOT Journal for the Study of the Old Testament JSS Journal of Semitic Studies KBL2 Koehler, Ludwig. 1958. Lexicon in Veteris Testamenti libros (2nd ed.). Leiden. KBL3 Koehler, Ludwig and Walter Baumgartner. 1967–1995. Hebräisches und aramäisches Lexikon zum Alten Testament. Leiden.

xvi KTU

abbreviations

Dietrich, Manfried, Oswald Loretz, and Joaquín Sanmartín. 2013. Die keilalphabetischen Texte aus Ugarit, Ras Ibn Hani und anderen Orten (3rd ed.; AOAT 360/1). Münster. LSAWS Linguistic Studies in Ancient West Semitic LiCCOSEC Lingua-Culture Contextual Studies in Ethnic Conflicts of the World LS Brockelmann, C. 1928. Lexicon Syriacum (2nd ed.). Halis Saxonum. LSP Schulthess, F. 1903. Lexicon syropalaestinum. Berlin. MD Drower, E. S. and R. Macuch. 1963. A Mandaic dictionary. Oxford. MDOG Mitteilungen der Deutschen Orient-Gesellschaft zu Berlin MLR Mediterranean Language Review MRLLA Magical and Religious Literature of Late Antiquity NABU Nouvelles Assyriologiques Brèves et Utilitaires NESE Neue Ephemeris für semitische Epigraphik OC Oriens Christianus. OLA Orientalia Lovaniensia Analecta OLZ Orientalistische Literaturzeitung Or Orientalia PAT Hillers, D. R. and Cussini, E. 1996. Palmyrene Aramaic texts. Baltimore. PLO Porta linguarum orientalium (neue Serie) RA Revue d’assyriologie RIÉ Bernand, É., A. J. Drewes, and R. Schneider (eds.). 1991–2000. Recueil des inscriptions de l’Éthiopie des périodes pré-axoumite et axoumite, Vol. 1–3. Paris. RLA Reallexikon der Assyriologie SAAS State Archives of Assyria Studies SAOC Studies in Ancient Oriental Civilization SBT Studien zu den Boğazköy-Texten SCCNH Studies on the civilization and culture of Nuzi and the Hurrians SEL Studi Epigrafici e Linguistici SL Sokoloff, M. 2009. A Syriac Lexicon. A Translation from the Latin, Correction, Expansion, and Update of C. Brockelmann’s Lexicon Syriacum. Winona Lake – Piscataway. StOr Studia Orientalia TS Payne Smith, R. 1879–1901. Thesaurus Syriacus, Vol. 1–2. Oxford. UF Ugarit Forschungen WO Welt des Orients ZA Zeitschrift für Assyriologie und verwandte Gebiete

abbreviations

ZAL ZDMG ZDPV ZS

xvii

Zeitschrift für arabische Linguistik Zeitschrift der Deutschen Morgenländischen Gesellschaft Zeitschrift des Deutschen Palästina-Vereins Zeitschrift für Semitistik und verwandte Gebiete Abbreviations in Linguistic Glosses and Paradigms

Several of the papers in this volume make use of linguistic glosses. The glossing system of each author has been been retained without any attempt to standarize, since there is not a unified glossing system in Semitic linguistics. In general, however, most of the authors follow the Leipzig Glossing Rules with occasional deviations. A list of abbreviations occurring in linguistic glosses is as follows: 1 first person 2 second person 3 third person ABS absolutive / status absolutus ACC accusative ADVZ adverbializer AUX auxiliary C common gender CON status constructus CONJ conjunction D dual DAT dative DECL declarative DEF definite article DEM demonstrative pronoun DN Divine Name DOM direct object marker EMP emphatic EMPH status emphaticus ERG ergative EX existential EZ Persian ezāfé construction F / FEM feminine GEN genitive

xviii

abbreviations

IMPF imperfect / imperfective IND indicative LPC long prefix conjugation m / masc masculine gender MVM main verb marker N noun NEG negative / negation NTRL neutral (focus) NOM nominative NP noun phrase NUC nucleus OBL oblique PC prefix conjugation PERF perfect / perfective PL plural PN personal name POSS genitive exponent PP preposition phrase PREP preposition PRES present PRO pronoun PRT preterite REL relative RN Royal Name S sentence SG singular SPC short prefix conjugation SC suffix conjugation SN Settlement Name SUBJ subjunctive

Abbreviations of Biblical Books

Following the Society of Biblical Literature (SBL), Biblical books are abbreviated as follows: Gen Genesis Exod Exodus Lev Leviticus

abbreviations

Num Numbers Deut Deuteronomy Josh Joshua Judg Judges Ruth Ruth 1Sam 1 Samuel 2Sam 2 Samuel 1Kgs 1 Kings 2Kgs 2 Kings 1Chron 1 Chronicles 2Chron 2 Chronicles Ezra Ezra Neh Nehemiah Esth Esther Job Job Ps Psalms Prov Proverbs Eccl Ecclesiastes Song Song of Solomon Is Isaiah Jer Jeremiah Lam Lamentations Ezek Ezekiel Dan Daniel Hos Hosea Joel Joel Amos Amos Obad Obadiah Jon Jonah Mic Micah Nah Nahum Hab Habakkuk Zeph Zephaniah Hag Haggai Zech Zechariah Mal Malachi

xix

Contributors Ahmad Al-Jallad (PhD Harvard University) is Assistant Professor of Arabic Linguistics and Ancient Arabia at Leiden University in the Netherlands. His research focuses on the epigraphy of Ancient Arabia, with special focus on North Arabia. More broadly, he publishes on issues in the comparative grammar of the Semitic languages, language contact, and the historical grammar of Arabic. Some of his key works include Graeco-Arabica I (2015) and an ancient Arabian zodiac. The constellations in the Safaitic inscriptions, part I (2014). He is author of the book An Outline of the Grammar of the Safaitic Inscriptions (Brill, 2015). Ali Al-Manaser (PhD Free University of Berlin) is a researcher on the project Online Corpus of the Inscriptions of Ancient North Arabia (OCIANA). His research focuses on the cultural context of the Safaitic inscriptions. He is the author of Ein Korpus neuer safaitischer Inschriften aus Jordanien (2008) and is currently preparing several editions of unpublished Safaitic inscriptions for publication. David Appleyard is Professor Emeritus of the Languages of the Horn of Africa in the University of London, having retired from the Department of the Languages and Cultures of Africa, SOAS, in 2006. His research focusses on the descriptive and comparative linguistics of Ethiopian Semitic and Cushitic languages, in the latter instance especially of the Central Cushitic or Agaw languages. More recently he has turned in particular to the study of contact-induced influence between Ethiopian Semitic and Cushitic. Among his major publications are Colloquial Amharic (Routledge) and A Comparative Dictionary of the Agaw Languages (Helmut Buske Verlag). He has also been one of the English-language editors of the Encyclopaedia Aethiopica. In 2014 he was awarded The British Academy Edward Ullendorff Medal for Semitic Languages and Ethiopian Studies in recognition of his scholarly achievements in both fields. Samuel L. Boyd (PhD University of Chicago) is Assistant Professor of the Bible and the Ancient Near East in the Department of Religious Studies and in the Program in Jewish Studies at the University of Colorado, Boulder. His research combines biblical studies, reception history, and sociolinguistics. He is particularly interested in the intersection between linguistics and historical-critical approaches to the

contributors

xxi

Hebrew Bible. He has published on the book of Jonah and Semitic epigraphy, and he is currently editing his dissertation for publication. Aaron Michael Butts (PhD University of Chicago) is Assistant Professor of Semitic Languages and Literatures in the Department of Semitic and Egyptian Languages and Literatures at The Catholic University of America. His research focuses on the dialectology and reconstruction of the Semitic language family. Among the Semitic languages, he specializes in Aramaic (including Syriac) as well as (Christian) Arabic and Classical Ethiopic. He is translator of Jacob of Sarug’s Homily on the Tower of Babel (Gorgias Press, 2009) and a co-editor of the Gorgias Encyclopedic Dictionary of the Syriac Heritage (Gorgias Press, 2011). His forthcoming monograph Language Change in the Wake of Empire: Syriac in its Greco-Roman Context (Eisenbrauns) analyzes contact-induced changes in Syriac due to Greek. Yochanan Breuer (PhD The Hebrew University of Jerusalem) is a professor of The Hebrew Language and Aramaic in the Department of The Hebrew Language and Jewish Languages at The Hebrew University of Jerusalem. His research focuses mainly on Mishnaic Hebrew and Babylonian Aramaic, but he also deals with other layers of the Hebrew Language. He is the author of The Hebrew in the Babylonian Talmud according to the Manuscripts of Tractate Pesahim (2002). Maria Bulakh studied linguistics at the Russian State University for the Humanities (Moscow) and did her Ph.D. on the etymology of color terms in Semitic. Since 2008 she has been a lecturer in the Institute for Oriental and Classical studies, Russian State University for the Humanities. Her main areas of research are comparative Semitic studies (with focus on the Ethio-Semitic languages), typology of semantic shifts, and language description. She was a member of the editorial team of the Encyclopaedia Aethiopica from 2004 to 2008 and a member of the editorial board of the volume Languages of the World. Semitic languages. Ethio-Semitic languages (in Russian) published in 2013. Since 2011 she has been engaged in a project on the description of the Soqotri language and recording of Soqotran folklore. David Michael Calabro (PhD University of Chicago) is a researcher at the Neal A. Maxwell Institute for Religious Scholarship at Brigham Young University. A specialist in the

xxii

contributors

intersection of language and culture in the ancient Near East, his published articles include studies on the Egyptian hieratic tradition in preexilic Judah, the semiotics of Northwest Semitic ritual gestures, and Egypto-Semitic etymology. Eran Cohen is Associate Professor in Semitic Linguistics in the Department of Linguistic in the Hebrew University of Jerusalem. The center of attention in his work mostly has to do with the syntax and discourse of various Semitic languages— Akkadian, Biblical Hebrew, Modern Hebrew, and Neo-Aramaic. Riccardo Contini is Professor of Semitics at the University of Naples “L’Orientale” and currently the coordinator of its PhD program in Ancient Near Eastern Studies. His research interests include Semitic comparative linguistics and interlinguistics, Aramaic philology (he edited with the late Cristiano Grottanelli Il saggio Ahiqar: fortuna e trasformazioni di uno scritto sapienziale [Brescia, 2005]), Syriac historiography, Arabic dialectology, and the history of Semitic studies. C. Jay Crisostomo (PhD University of California, Berkeley) is Assistant Professor of Assyriology at the University of Michigan. He has worked extensively on digital editions and analysis of cuneiform texts, especially for the Digital Corpus of Cuneiform Lexical Texts (oracc.org/dcclt). His research explores language, scribal practices, and social history in Babylonia in the third through first millennia BCE. His work on language focuses on multilingualism, translation, and sociolinguistic variation with particular emphasis on Sumerian and Akkadian. His dissertation, the recipient of the International Association for Assyriology’s 2015 De Gruyter award, focused on questions of multilingualism and scribal knowledge in Old Babylonian schools and is currently being revised for publication. Lutz Edzard (PhD University of California Berkeley, Habilitation University of Bonn) is Professor of Semitic linguistics at the University of Oslo and Professor of Arabic and Semitic linguistics at the University of Erlangen-Nürnberg. His research interests include comparative Semitic and Afroasiatic linguistics with a focus on phonology, Arabic and Hebrew linguistics, and the history of diplomatic documents in Semitic languages. He is the editor, with Rudolf de Jong, of the Encyclopedia of Arabic Language and Linguistics online at Brill, and, with Stephan Guth, the editor of the online Journal of Arabic and Islamic Studies.

contributors

xxiii

He serves as the Semitics editor of Zeitschrift der Deutschen Morgenländischen Gesellschaft and, with Stephan Guth, as the editor of the series Porta Linguarum Orientalium at Harrassowitz. Humphrey Hardy (PhD University of Chicago) is Assistant Professor of Old Testament and Semitic Languages at Southeastern Seminary in Wake Forest, North Carolina. His research concentrates on Ancient Hebrew, Comparative Semitic Philology, and Historical Linguistics particularly focused on grammaticalization and diachronic issues within the corpus of the Hebrew Bible. He has published articles on Aramaic inscriptions, Hebrew lexicography, and grammaticalization of particles in peer-review journals such as the Journal of Semitic Studies, Vetus Testamentum, and Bulletin of the American Schools of Oriental Research. Uri Horesh (PhD University of Essex) is Assistant Professor of Instruction in Arabic and Language Coordinator in the Program in Middle East and North African Studies at Northwestern University. His research deals with outcomes of contact between Arabic and Hebrew in historic Palestine, as well as Arabic sociolinguistics in general. He has taught at various institutions, including Georgetown University, the University of Texas at Austin, and Franklin & Marshall College. He is currently co-editing (with Enam Al-Wer) The Routledge Handbook of Arabic Sociolinguistics, which is in preparation for future publication. Otto Jastrow (PhD University of Saarbrücken) has served as professor ordinarius of Semitic Philology in the universities of Heidelberg and Erlangen-Nürnberg. He is currently holding the office of professor of Arabic in the University of Tallinn (Estonia) where he has inaugurated the first program of Arabic studies in Estonia. His research interests are centered on linguistics and spoken language, with special emphasis on Arabic dialects and Neo-Aramaic languages and dialects. His publications count 12 books and around 80 articles, equally divided between the two areas mentioned. Among his monographs are Lehrbuch der Turoyo-Sprache (1992), Der neuaramäische Dialekt von Mlahso (1994), Die mesopotamisch-arabischen qəltu-Dialekte (2 vols.; 1978, 1981), Die arabischen Dialekte der Juden von Arbil und Aqra (1990). He is currently working on a series of monographs devoted to individual qəltu dialects. Otto Jastrow is co-editor (together with Werner Arnold) of Zeitschrift für Arabische Linguistik / Journal

xxiv

contributors

of Arabic Linguistics (ZAL) and editor of the monograph series Semitica Viva and its companion series Semitica Viva—Series Didactica (published by Otto Harrassowitz). Lily Okalani Kahn (PhD University College London) is Lecturer in Hebrew at that university. Her main research area is Hebrew in Eastern Europe. She is also interested in Yiddish, comparative Semitics, Jewish languages, and global Shakespeare. Her publications include The Verbal System in Late Enlightenment Hebrew (Brill, 2009), Colloquial Yiddish (Routledge, 2012), The Routledge Introductory Course in Biblical Hebrew (Routledge, 2014), and A Grammar of the Eastern European Hasidic Hebrew Tale (Brill, 2015). She is currently working on a bilingual edition of the earliest Shakespeare plays translated into Hebrew entitled The First Hebrew Shakespeare Translations (UCL Press). Joseph Lam (PhD University of Chicago) is an Assistant Professor in the Department of Religious Studies at the University of North Carolina at Chapel Hill, where he teaches broadly in the areas of ancient Near Eastern languages, history, and culture. His research focuses on the written traditions of the Levant in the 2nd and 1st millennia BCE, including the texts from Ugarit (modern Ras Shamra) and the Hebrew Bible. He is particularly interested in examining ancient ritual practices and religious conceptualizations as they are represented in the textual material. He is currently preparing his first book, Patterns of Sin in the Hebrew Bible: Metaphor, Culture, and the Making of a Religious Concept, for publication with Oxford University Press. Mila Neishtadt completed her BA and MA studies in the Department of Arabic and Islamic Studies at Tel Aviv University. She later received the Polonsky Scholarship for Outstanding Doctoral Candidates and is currently a PhD candidate in the Department of Arabic Language and Literature at the Hebrew University of Jerusalem. She is interested in Semitic Linguistics, while focusing on Arabic linguistics and language contact. In her dissertation, “Studies in the Aramaic Substrate of Palestinian Arabic,” she is analyzing the Aramaic lexical remnants in Palestinian Arabic from phonological, morphological and semantic aspects and is compiling an indexed lexicon of substrate elements in Palestinian Arabic. Within this study, special attention is paid to Aramaic substrate lexica in Palestinian Arabic which can be traced back to Hebrew.

contributors

xxv

Mikhal Oren completed her PhD at the University of Haifa in 2012. Her fields of interest include ancient Semitic languages and texts and modern Hebrew lexicography. Paola Pagano is a graduate student in Ancient Near Eastern Studies at the University of Naples “L’Orientale” specializing in Assyriology, Aramaic philology, and Arabian archaeology. Her research interests focus on the Akkadian lexical heritage in Eastern Aramaic. Aaron D. Rubin (PhD Harvard University) is Malvin E. and Lea P. Bank Professor of Jewish Studies, Classics & Ancient Mediterranean Studies, and Linguistics at Penn State University. His research interests include the whole of the Semitic language family, ancient and modern, with special attention to Hebrew and Modern South Arabian. He is the author of five monographs, the most recent of which is The Jibbali Language of Oman (Grammar and Texts) (2014). He is also co-editor of Brill’s Encyclopedia of Hebrew Language and Linguistics (2013) and Brill’s forthcoming Handbook of Jewish Languages. Lotfi Sayahi is Associate Professor of Linguistics in the Department of Languages, Literatures, and Cultures at the University at Albany, State University of New York. His research focuses on language contact and language change, particularly in North Africa and Spain. His recent publications include the book Diglossia and Language Contact: Language Variation and Change in North Africa (Cambridge University Press, 2014). He has also published articles that have appeared in The Journal of Language Contact, The International Journal of the Sociology of Language, Revista Internacional de Lingüística Iberoamericana, The Journal of Language Sciences, and The Journal of Sociolinguistics, among others. Johann Jürgen Tubach was born in 1947 in Bruchsal (district of Karlsruhe). In 1966/67–1972, he completed theological studies at the University of Heidelberg, with his first theological examination in 1972. Starting in the winter term of 1972/73, he studied Oriental studies, Greek and Roman history, and history of old oriental religions at the Universities of Heidelberg and Hamburg and then Christian Oriental studies, history of religions, and Semitic languages at the University of Bonn.

xxvi

contributors

He earned a Dr. phil. in 1982 from the the University of Bonn (Im Schatten des Sonnengottes [Wiesbaden, 1986]) and a Dr. habil. in 1989 from the same institution (Die Herkunft der spätantiken Sonnenreligion im Licht christlichorientalischer Quellen [unpublished; pp. v+706). He completed a second theological examination (Badische Landeskirche, Karlsruhe) in 1984. Since 1995, he has been Prof. for Christian Oriental Studies at the Martin-LutherUniversity in Halle (upon Saale). His interests include the old non-Christian religions in the Near East in Hellenistic-Roman times, Ethiopian history till the 17th century, church history of Nubia, Manichaeism, history of the Nestorian church, mission of the Nestorians along the Silk Road, and the history of Christianity in South India till the 17th century. Juan-Pablo Vita (PhD University of Murcia) is Tenure Scientist in the Instituto de Lenguas y Culturas del Mediterráneo y Oriente Próximo (ILC) of the Consejo Superior de Investigaciones Científicas (CSIC) in Madrid. His main lines of research focus on the study of the languages and the socio-economic history of SyriaPalestine in the second millennium BC. Within this area, he has carried out an important part of his work on the text archives of Ugarit, Emar, and El-Amarna. He is the author of El ejército de Ugarit (CSIC, 1995), A Concordance of Ugaritic Words (Gorgias Press, 2003; co-authored with J.-L. Cunchillos and J.-Á. Zamora), Manual de lengua acadia (IEIOP Zaragoza, 2005 and 2006; co-authored with F. Malbran-Labat), Das Kanaano-Akkadische der Amarnazeit (Ugarit-Verlag, 2010; co-authored with J. Tropper), and Canaanite Scribes in the Amarna Letters (Ugarit-Verlag, 2015). Tamar Zewi (PhD the Hebrew University of Jerusalem, 1993) is Professor of Hebrew and Semitic Linguistics at the University of Haifa. She is the author of A Syntactical Study of Verbal Forms Affixed by -n(n) Endings in Classical Arabic, Biblical Hebrew, El-Amarna Akkadian and Ugaritic (Ugarit-Verlag, 1999), Parenthesis in Biblical Hebrew (Brill, 2007), and The Samaritan Version of Saadya Gaon’s Translation of the Pentateuch (Brill, forthcoming) and is one of the editors of the Encyclopedia of Hebrew Language and Linguistics (Brill, 2013). Her fields of interest include Hebrew and Semitic linguistics (especially syntax), ancient Semitic languages (Biblical Hebrew, Ugaritic, El-Amarna Akkadian, Ge’ez), Classical Arabic and Mediaeval Judaeo-Arabic, Tafsīr Saadya Gaon and Bible translations into Semitic languages.

A Thamudic B Abecedary in the South Semitic Letter Order Ahmad Al-Jallad and Ali Al-Manaser 1 Introduction The rock under consideration was discovered by Dr. Sabri Abbadi in 2004 during a survey in Wādī al-Ḥašād about 45 km to the northeast of aṣ-Ṣafawī, near the north-eastern border of Jordan.1 Dr. Ali al-Manaser kindly provided the photograph, but the exact location of the rock remains unknown. The obeliskshaped slab of basalt, measuring 133 centemeters, from the base to the apex, by 34 centemeters, bears two Thamudic B inscriptions on a single face. While it is not unusual to encounter the occasional Thamudic B inscription in northeastern Jordan and southern Syria, they are far more common in what is today Saudi Arabia. The significance of this particular find, however, lies in its contents. The inscriptions attest several unique letter shapes not known from elsewhere, unusual inscriptional formulae, and the first attestation of a Thamudic B abecedary, which follows the South Semitic order. Before examining the texts themselves, a few remarks on the language inscribed in the Thamudic B script is in order. As is well known, Thamudic is a conventional label applied to the Ancient North Arabian (ANA) scripts which have not been precisely classified in any of the established categories;2 the term, therefore, does not refer to a single script, nor does it suggest that the scripts labeled Thamudic are more closely related to each other as opposed to other varieties of the South Semitic script. Our knowledge of the languages these scripts express is even more fragmentary. While it has been conventional to consider all of the varieties attested in the ANA scripts reflexes of a single language, there is no a priori reason to assume so. The different North Arabian epigraphic corpora are separated by great expanses of space and time, and to date no shared linguistic innovations to support the existence of an ANA subgrouping have been identified.3 1  We thank Dr. Sabri Abbadi for giving us permission to publish this photograph. 2  The established categories include: Taymanitic, Dadanitic, Hismaic, and Safaitic (Macdonald 2000). 3  On this issue, see Al-Jallad 2015, especially n.12. We differ here from the recent assertion of Hayajneh (2011: 760) that Ancient North Arabian constitutes as single “linguistic area.” There

© koninklijke brill nv, leiden, ���5 | doi ��.��63/9789004300156_002

2

Al-Jallad and Al-Manaser

Most Thamudic B inscriptions contain only a few words, primarily onomastica and short invocations to various deities. But even these brief texts can testify to considerable internal linguistic diversity, which could signal both regional and chronological variation. For example, word final triphthongs/ diphthongs are often preserved, but can occasionally collapse to a vowel: for instance, the divine name */ruḍaw-/ is spelled as both rḍ but more commonly as rḍw. One of the unique features common in the Thamudic B inscriptions is the particle nm, which is no doubt a reflex of lm, the dative preposition plus an enclitic mem, with regressive assimilation.4 Nevertheless, the basic form l is sometimes attested. Beside the definite article h-, ʾl is very rarely attested (see, for example, WTI-Thamudic 25). Other notable grammatical features include the consonantal reflex of the first person singular clitic pronoun -y (Macdonald 2004: 507), the ʾ- causative prefix (see 1a below), and the leveling of the -t suffixes to the first person of the suffix conjugation, rfʾt ‘I am healed’ (Hayajneh 2011: 770). One of the most obvious gaps in our knowledge concerns the lexicon. Many of the bizarre translations found in the text editions suggest that the reliance on the Classical Arabic lexicons as the sole point of reference is misguided. As will be seen, two important terms in 1a found better semantic matches in Ancient South Arabian (ASA) and Hebrew rather than Classical Arabic. Clearly, the process of linguistic decipherment is still underway, and much more work is needed to determine the linguistic character and affiliation of the varieties attested in the Thamudic scripts. 2

The Texts

2.1 Text 1 The first text has two components: (a) a building and prayer text and (b) a partial abecedary in the South Semitic order. The text curves around the edge of the face starting at the bottom right, with each of its components sharing roughly an equal portion of the stone. is significant linguistic variation not only between different script categories, but within them as well. So large a claim cannot be asserted in such a cavalier manner, but must instead be closely argued with the support of a large amount of linguistic evidence. It would seem highly unlikely that the authors of the Taymanitic inscriptions had anything to do with the authors of the Safaitic inscriptions, who not only wrote hundreds of miles away, but centuries later. As one might expect, the languages attested in both corpora exhibit considerable differences. 4  Knauf (2009: 481) has suggested that this particle functions as a direct object marker. In some cases, however, it seems to have an introductory or presentative function as well.

A Thamudic B Abecedary in the South Semitic Letter Order

3

1.a Transliteration h nṣb ḥrt w nhy ʾtmn ʾqṭr w ws²{ʿ } l -h mn fnw [-k] Translation Verily, Ḥrt has erected a cult-stone, so, Nhy, restore ʾqṭr to health and may he be strengthened by means of [your] countenance. 2.2 Commentary h nṣb: There are two ways of interpreting the relationship between the first three words, both dependent upon the way we understand the word nṣb. Macdonald (2012: 291) has recently suggested that the nṣb stone acted as a commemorative symbol of a divine being, or even a representation of the being itself. Indeed, such a hypothesis seems likely as the nṣb is mentioned in connection with the names of divinities in the Safaitic inscriptions, e.g., w nṣb ʾṯʿ ‘and he erected the cult-stone for ʾṯʿ (a well-known divinity in the Safaitic inscriptions, our insertion)’ (Macdonald et al. 1996: 456). Given that the third word, Ḥrt, is not a known divine name, it is probably best to take nṣb as a verb with the intended deity, nhy, being mentioned in the following clause. In light of this understanding, it is tempting to see the intial h as a causative morpheme, as in Sabaic and Dadanitic; however, no such forms have been identified in Thamudic B, nor are they known from the inscriptions carved in the Safaitic, Hismaic, Taymanitic, and early Arabic scripts. It would therefore seem better to take the h as a presentative particle, cognate with Arabic ʾin(na) and Hebrew hēn and hinnē. If the Thamudic form was hn, then it is possible that the two word-boundary n’s were treated as a geminate consonant and consequently written with a single n, hnṣb = */hinnaṣab/. On the other hand, if we chose to understand nṣb more generically as ‘a monument’ or a ‘votive stele’, as found in Phoenician and Punic (Hoftijzer and Jongeling 1995: 750), then the noun probably stands in construct with the personal name Ḥrt. The intial h, then, should be understood as a demonstrative particle, rending ‘this is the monument of Ḥrt’. Indeed, a simple h- demonstrative seems to be known from other ANA corpora (Macdonald 2004: 518). Since this sense is not attested in the Safaitic or Hismaic inscriptions and seems to be at odds with the function of the nṣb in North Arabia, we favor the former interpretation. However, both possibilities should at least be considered on account of the fact that this is the first attestation of the word in Thamudic B, and so its exact meaning in this epigraphic context cannot be securely established. Ḥrt: This name is probably a reflex of the common Arabian name ḥāreṯat-, which is attested widely in the North Arabian epigraphy, in various forms, and in Nabataean. Its particular form here suggests that it was drawn from a

4

Al-Jallad and Al-Manaser

language that had merged ṯ with t, a sound change typical of Aramaic. I have suggested elsewhere5 that this particular form reflects an Aramaic calque of Arabic ʾal-ḥāreṯ, namely, ḥārtā. For a possible alternative, Thamudic B-internal explanation, see below § 3. w nhy: The deity nhy, probably *nuhay based on the neo-Assyrian transcription dnu-ḫa-a-a in the Esarhaddon prism, is widely attested in Thamudic and is also occasionally found in Safaitic and Hismaic, although considerably less frequently. ʾtmn: Winnett (1987, 240) understood this verb to mean ‘to make complete or whole physically’, that is, ‘to restore to health’. This seems much more convincing than the sexual interpretation given to it by Jamme (1968).6 Winnett, following Littmann, interpreted the final n, which sometimes appears on this verb, as a reflex of the energic ending common on imperative forms. Thus, the form ʾtmn would reflect an imperative of the C-stem (form IV/causative) of √tmm ‘to complete’, with an energic ending, meaning ‘restore to health!’. ʾqṭr: The formula h- divine name, ʾtm(n) + personal name is rather common in the Thamudic B inscriptions (again, see Winnett 1987), and so this fact motivates us to understand ʾqṭr as an anthroponym rather than a common noun. Compare:7 Hu 104 h yġ ʾtm (ṣ)ʾm wdd -h O Yġ, restore Ṣʾm, his friend, to health Hu 128 h rḍw ʾtmn ddy O Rḍw restore my uncle [or the PN Ddy, our insertion] to health 5  See Al-Jallad (2015, §3.2 n. 68) on the absence of spirantization in Nabataean Aramaic and on the different realizations of this name. 6  The sexual sense of this word for Safaitic also seems unjustified. Winnett took the phrase ʾtm zf in C 285 as meaning ‘he ripped [the genitalia] of Zawf’ based on the presence of nyk ‘to copulate’ in the same pericope, and a misidentification of qd with Arabic qāda ‘to act as pimp’. The inscription should read and translate as: l bṯ w nyk bgrmh f ʾtm zf f qd ʾbṣ ‘by Bṯ and he copulated with Bgrmh , for he had celebrated the sending of the bride and was satisfied’. The qd of this inscription should in fact be compared with the preverbal marker in Classical Arabic, qad. Recently, Hayajneh (2011) has given the translation ‘to complete the pleasure’ to the Thamudic verb, but does not justify it. 7  The readings and interpretations follow Winnett 1987; see that publication for alternative sigla as well.

A Thamudic B Abecedary in the South Semitic Letter Order

5

Hu 753 h nhy ʾtm tmʿzz O Nhy, restore Taimʿazīz to health This previously unattested name seems to reflect an elative pattern or a C-stem suffix conjugation of the root qṭr (Lane 1863–1893: 2542). It is impossible to determine its meaning with any certainty, but a connection with the sense ‘to drop, or flow’ would seem to be suitable, especially if referring to water or rain (Lane 1863–1893: 2542b–2543b), comparable to names like mṭr ‘rain’ (Harding 1971: 551). w ws2{ʿ} l -h: The rectangular shape of the w with a single diagonal line is sporadically attested in Thamudic B, but the circular w is more common. The rectangular w, however, is attested in two out of the three known Dumaitic inscriptions.8 The ʿ is left partially open, which causes it to resemble an r or an f, but neither of these options provides a suitable root. Moreover, the f occurs later in the inscription and is distinct from the glyph under consideration. While ʿ in the alphabetic portion of this inscription is closed, the circle of the q and ḍ remain partially open, suggesting that this was perhaps a peculiarity of the author’s hand. On the other hand, its particular shape may be due to a writing error. The word ws2ʿ should be connected with Sabaic ws2ʿ, which is attested in the C-stem hws2ʿ, meaning ‘to bestow, grant; give aid’, said particularly of a god (Biella 1982, 152; Beeston et al. 1982: 163). This root is not known in Classical Arabic; however, a similar meaning is associated with √wsʿ, which can mean ‘to grant generously’ (Lane 1863–1893: 3052c). The correspondence between Arabic s and Thamudic and Sabaic s2 is problematic. The true cognate of the Classical Arabic root seems to be Sabaic ws3ʿ, meaning ‘to decree, ordain’, and in the C-stem ‘to give sufficient water’ (Beeston et al. 1982: 163; Biella 1982: 151). While the sense of the root seems suitable, its morphological identity is more difficult to determine. The form ws2ʿ may reflect an imperative or optative use of the suffix conjugation, perhaps in the D-stem, which can sometimes be identical in meaning to the C. But there are a few problems with this interpretation. The benefactive is treated as a direct object in Sabaic, whereas here it would seem to be introduced by the preposition phrase, l- or, less likely [ʿ ] l, with a word-boundary spelling. There is of course no reason to assume an identical syntax between Thamudic and Sabaic, but we encounter yet another 8  See WTI-Dumaitic 21–23.

6

Al-Jallad and Al-Manaser

discrepancy between the two usages: the Sabaic verb is ditransitive, and it is always accompanied by its theme, or direct object, which is introduced by b-. No direct object is mentioned in this inscription. One could suggest that ws2ʿ without a direct object meant simply to ‘bestow divine favor’. This would suit nicely the reconstruction of the preposition phrase as [ʿ ]l -h ‘upon him’. Attractive as this solution may seem, such a usage is unattested elsewhere. We think it is better to take ws2ʿ as a nominal form meaning ‘strength’ or ‘aid’, similar to Sabaic ws2ʿn ‘idem.’ (Biella 1982: 152). This would render the first part an existential sentence, which must be taken pragmatically with an optative force, meaning ‘may he be strengthened’ (lit. ‘may there be strength for him’). mn fnw [-k]: The prepositional phrase introduced by mn clearly modifies the verb ws2ʿ. We read the second word as fnw. While the f is not carved in the typical Thamudic B manner, it is identical to the glyph which occupies the position of f in the abecedary that follows. While the rock is broken at the top, the final word can be none other than fnw ‘face’, cognate with Hebrew pānâ, panîm and Sabaic fnw (Beeston et al. 1982: 45; Biella 1982: 405–6). In Sabaic, its meaning has shifted to features of the land and/or canals and can be used adverbially to indicate direction (ibid.). The root has undergone a semantic shift in Arabic as well, where it has come to mean ‘to pass away’, ‘to be extinguished’, probably a development from an earlier meaning, ‘to leave’, paralleling the development of the root √hlk from original ‘to walk’ to ‘to perish’. One could take fnw in this inscription to mean something like ‘death’, which would lead to the translation ‘may he be strengthened against death’. How­ ever, we would expect the preposition ʿl rather than mn to express ‘against’ in such a context. On the other hand, the word’s original meaning ‘face’ is quite suitable. The use of ‘face’ to signify divine countenance is common in Near Eastern religious phraseology. Consider, for instance, Ps 89:16: ʾašrê hāʿām yōḏʿê tərûg̠â YHWH bəʾôr-pānek̠ā yəhallēk̠ûn ‘Happy is the people who know the joyful shout; they walk, YHWH, in the light of thy countenance’. The Qurʾān uses waǧh- in a similar way. The phrase waǧhu llāh ‘the countenance of Allāh’ and waǧhu rabbihim ‘the countenance of their lord’ are frequently attested (for example, Q 13:22; 30:39; 76:9). It seems best to understand fnw, in conjunction with mn, as meaning ‘by means of your (i.e., Nhy) countenance’. For this reason, too, we restore a final k, the 2nd person clitic pronoun, in the portion that has been broken off. 1.b The second portion of the inscription consists of a partial alphabet in the South Arabian order. The letters themselves agree mostly with Thamudic B

A Thamudic B Abecedary in the South Semitic Letter Order

7

shapes, but a few forms are unique and require discussion. The implications of this will be discussed in the historical commentary below; only the forms of the letters will concern us at the moment. h l ḥ m q w s2 r b t s1 k n ḫ ṣ ṯ f ʾ ʿ ḍ 2.3 Commentary m: The m in South Arabian and Thamudic B are essentially identical, except for their orientations. The two triangles of the m in 1a faced upwards and downwards, while the triangles of the m in the alphabet portion face towards the end of the inscription, typical of the ASA form. w: The circular w in the alphabet portion is closer to the ASA w, and indeed the w typically used in Thamudic B, than the form used in the first half of the inscription. s2: The s2 has been damaged, but the glyph clearly reflects the typical Thamudic B shape, namely, a vertical serriform line with ears on one or both ends. t: The cross-shaped t is commonly found in Thamudic B against the X-shaped ASA glyph. n: The n is decidedly Thamudic B. ḫ: The ḫ attested here seems to be a compromise between the South Arabian 𐩭 and the X-shaped glyph common in many ANA scripts, including Thamudic B. This causes it to resemble the alif of Safaitic and Hismaic. The present shape could reflect an archaic form, ancestral to the X-shaped ḫ of other ANA scripts. A similar simplification process seems to have operated in Safaitic and Hismaic where the alif has a by-form which is essentially an X. ṣ: It seems that the author first inscribed a glyph virtually identical to the preceding ḫ, but then closed the furcular opening at the top to create the ṣ. This form agrees with the ANA and middle and late Sabaic shapes. A possible explanation for its strange shape may be that the author mistakenly carved a second ḫ, and then attempted to correct it rather than discarding the entire inscription. ṯ: The author has inscribed the ṯ glyph, which is identical in Thamudic B and ASA, in place of ASA s3. f: The f of this inscription has a unique shape not encountered elsewhere. It consists of two bow shaped strokes opening towards each other but with their ends not touching. This could be a variant of the diamond-shaped ASA glyph. The closest ANA shape is found in Dadanitic or Taymanitic, where f consists of an open circle, resembling the ʿ of 1.a, or open diamond shape. The typical Thamudic B glyph bears no resemblance to the form given here.

8

Al-Jallad and Al-Manaser

ʾ: Thamudic B has two forms of the alif, one which resembles the ASA glyph and the other, with two ears, resembling the shape found in the Oasis North Arabian scripts. The present glyph reflects the latter; the furcular opening facing upward is typical, but the trunk is unusually elongated. ḍ: This glyph is decidedly Thamudic B. 2.4 Text 29 The second text is situated in the middle of the stone and consists of two lines written from left to right. The first line contains only two words and is centered, while the second spans the length of the stone and curves around the text of the first line. Transliteration (1) l drr (2) {w} h nhy s¹ʿd [d]rr ʿl wdd-h {y}ʾṭ {b}b-yd [d]rr Translation (1) By Drr (2) {and}, O Nhy, help Drr in the matter of his love, {Yʾṭb}—by the hand of Drr 2.5 Commentary 2.5.1 Script and Formulae The script of this inscription fits well within the standard variation of Thamudic B and does not exhibit the rectangular w’s of 1a–b. In both the introduction and narrative component, the author’s d’s face towards the beginning of the inscription while his r’s open facing the end, but this alignment is reversed in the signature component, byd[d]rr. Playfulness of this sort with the script is typical of ANA graffiti. The text exhibits a unique mix of formulae found in both Thamudic B and Nabataean graffiti. The expression h nhy s1ʿd + PN ʿl X is well known; a virtually identical expression is found in Dumaitic: WTI-Dumatic 23: h rḍw w nhy w ʿtrs1m s1ʿdn ʿl wddy ‘O Rḍw and Nhy and ʿtrs1m help me in the matter of my love’ (Winnett and Reed 1970: 80). The signature ‘by the hand of PN’, on the other hand, is attested here for the first time in Thamudic B. To our knowledge, it has not been attested in any other ANA corpus either, but resembles the common Nabataean signature formula found in graffiti and papyri, ktb yd PN ‘the writing of the hand of PN’. Another unusual feature for Thamudic B is the use of l as an introductory particle instead of nm. Perhaps 9  We thank M. C. A. Macdonald who read and interpreted the signature component of this inscription.

A Thamudic B Abecedary in the South Semitic Letter Order

9

it is significant that the former is typical of the Safaitic inscriptions, which are usually carved in this region. 2.5.2 Word-boundary Sandhi and a Ligature Two clear cases of word-boundary spellings are present in the inscription, both with d. The first d must be interpreted as representing the last d of the imperative verb s1ʿd ‘aid, help’ and the first letter of the authors name drr. The same applies to the signature, bydrr, which must be interpreted as b- yd [d]rr ‘by the hand of Drr’. This phenomenon minimally indicates that the dialect of this inscription lost word-final [i]. The two small, and rather crudely carved, b’s share a single horizontal upperline. It seems that the author originally carved a disproportionately large b, perhaps indicating that he originally intended it to be the final character of the inscription. Then, when he decided to sign the inscription, he added two vertical strokes to the interior of the b, effectively splitting it into two b’s. The rightmost line of the first b was extended to the edge of the rock, and a similar line was extended from the s1 downwards as well. 2.5.3 Onomastica Drr: This name has been previously attested in Safaitic (Harding 1971: 238); Harding connects it with the Classical Arabic darra ‘to flow, shine brightly’. {y}ʾṭ{b}: We have identified this word as a personal name in apposition with the noun wdd -h ‘his beloved’. A small line extends from the h to the first letter of this word, which is a circle. The patina is identical to the letters of the rest of the inscription, suggesting the line was intentional. Thus, the first letter is probably a malformed y, slanting to the right. The author may have carved the circle first and then upon realizing that he did not have space for a sufficiently long line downwards, on account of the previously inscribed k of 1b, carved it diagonally instead. The same explanation can account for the diagonal placement of the alif, which is carved considerably above the following ṭ. The name yʾṭb is unknown, but would seem to derive from the root √ṭyb ‘to be good, pleasant’ (Lane 1863–1893: 1900). The present form would be derived from a C-stem verb from a language which preserved the causative ʾ- in the prefix conjugation. A possible example of such a name is found in Taymanitic, e.g., yʾrs2l (JSTham 542 = Esk 72), perhaps ultimately derived from yuʾawriṯʾil ‘God grants inheritance’. 3

Historical Commentary

The upper and lower chronological limits of the Thamudic B script are, at our current state of knowledge, unclear. We know that Thamudic B was in use in

10

Al-Jallad and Al-Manaser

the mid-sixth century BCE, as an inscription carved in this script mentions the Babylonian king Nabonidus, but no other inscriptions with absolute chronological information have been discovered.10 Of the Thamudic scripts, B is the most widespread, with specimens found as far south as Yemen and as far north as Syria, and nearly everywhere in between. Alphabetic texts have been discovered for Hismaic, Safaitic, and Dadanitic, but only the Dadanitic abecedary arranges the letters in the hlḥm letter order. The single Hismaic abecedary is essentially in the NWS order, with a few irregularities, and each of the several Safaitic alphabetic texts is arranged according to a unique letter order.11 The present inscription contains the first evidence for the use of the hlḥm letter order in a desert ANA script.12 It seems unlikely that the author was trying to reproduce the ASA alphabet, as most of the letter forms agree with the Thamudic B shapes. The f and the ḫ are exceptions, but they still do not match the forms encountered in the ASA script. South Arabia is not the only source for the hlḥm order. Two abecedaries in alphabetic cuneiform in which the letters are arranged according to this order have been discovered (KTU 5.25 = 8.1 [Bēt Šemeš]; KTU 9.426 = RS 88.2215 [RasŠamra]).13 Both are very difficult to interpret, but they nevertheless differ in important ways from the canonical ASA order: KTU 5.24 h l ḥ m q w ṯ r t š k n h̠ [ṣ s p ʾ] ʿ ẓ g d ġ ṭ [z] [ḏ] y RS 88.2215 h l ḥ m q w ṯ r b t ḏ š k n ẖ ṣ s p ʾ ʿ ẓ g {d} ġ ṭ z y ASA h l ḥ m q w s2 r b t s1 k n ẖ ṣ s3 f ʾ ʿ ḍ g d ġ ṭ z ḏ y ṯ ẓ KTU 5.24 is fragmentary, and it is therefore difficult to determine the degree of congruence between the arrangement of its letters and the ASA order. The b is missing, but this may be a writing error, and ṯ occupies the position of s² in the ASA order. RS 88.2215, on the other hand, is basically complete and exhibits three points of difference vis-à-vis the ASA order. As with KTU 5.24, a glyph which most closely resembles a ṯ occupies the position of ASA s2.14 The 10  See Macdonald (2010: 16 n. 27) on this inscription. 11  For an excellent analysis of these, see Macdonald 1986; the most recent collection of the Safaitic alphabets is found in Rawan 2013: 84–86. 12  On the divisions of the ANA scripts, see Macdonald 2000; 2004. 13  These are discussed most recently in Bordreuil and Pardee 1995; Hayajneh and Tropper 1997; Sass 2005. The readings provided here are taken from Hayajneh and Tropper 1997: 27. 14  The reason for this is unknown, but it is possible that the writer was arranging the cuneiform letters according to the order of the ASA script, which kept the three voiceless sibilants apart graphically. Consequently, he judged the interdental ṯ the closest match to the lateral (=*s2), which perhaps did not exist in the language the alphabet cuneiform script normally expressed.

A Thamudic B Abecedary in the South Semitic Letter Order

11

interdental ḏ is inserted between t and s1 rather than following z, as in the ASA order. Finally, cuneiform ẓ occupies the position of ASA ḍ. While the Ugaritic language does not maintain a distinct reflex of ḍ, the sound merged with *ṣ rather than *ẓ, and therefore its position cannot simply be attributed to the loss of the phoneme. Similarly, the Thamudic B abecedary also differs from the ASA order with respect to ṯ, yet unlike the two cuneiform abecedaries, the glyph occupies the position of s3. There are several ways to interpret this, but first a short discussion on the phonetic values of the sibilants is in order. In former times, it was thought that the Proto-Semitic voiceless sibilants series was essentially identical to Hebrew, where *s1 was realized as [ʃ], *s2 as a lateral [ɬ], and *s3 as a plain sibilant [s]. However, since the thorough work of Faber (1984; 1985; 1992) the phonetic values of two of these phonemes have come to be understood differently. *S1 was probably realized as a plain voiceless sibilant [s], a realization Arabic preserved. However, one may also admit the possibility of an apical allophone, similar to the voiceless sibilant in Modern Greek or Dutch. The value of *s2 remains a voiceless lateral fricative, but s3 was probably the plain counterpart to the ejective affricate *ṣ [ts’], meaning that it was realized as [ts]. Its merger, then, with *s1 in many of the languages of North Arabia was the result of an areal phenomenon of deaffrication. If the hlḥm order was the inherited letter sequence of the Thamudic B script, one could suggest that the special placement of ṯ was simply the consequence of the loss of s3 in the dialect(s) this script represented. Such an explanation, however, cannot account for why ṯ was singled out and moved from the penultimate position in the canonical order to replace s3. Moreover, Thamudic B was a script employed almost exclusively by nomads for informal purposes. The absence of schools or an academic tradition means that the alphabet was passed on from user to user, and that there would have been no tradition to preserve a single arrangement of letter forms (Macdonald 1986: 115–116). For this reason, the known ANA abecedaries reflect either random orderings of the letters or a borrowed sequence from sedentary scripts. We would argue that the Thamudic B abecedary falls into the latter category and reflects a borrowing from sedentary people employing the ASA alphabet or even a sedentary ANA script that used the South Semitic order, such as Dadanitic. In this light, the placement of the ṯ in the position of s3 is reminiscent of the confusion of these two phonemes in Ḥaḍramitic and in the Amiritic dialect of Sabaic. In Ḥaḍramitic, the confusion of the two sounds seems to be a part of a larger phenomenon involving the merger of the interdentals with the affricates/sibilants, suggesting the sound changes [θ] > [ts] or [s]. If the author had learned the alphabet order from diction, then the sound occupying this position would have been [s] or [ts] and not [θ], and so a Ḥaḍramitic source seems unlikely. On the other hand, Amiritic merged s3 and s1, such that the reflex of

12

Al-Jallad and Al-Manaser

etymological *s3 was consistently written with s1, suggesting the sound change ts > s. Following this, the reflex of *ṯ was written with the s3 glyph (Stein 2004: 228f.). This could suggest two things: (1) the interdental had become an affricate, [θ] > [ts] or (2) the old affricate sign was reassigned to the interdental. Given that there is no evidence for the confusion of ḏ and z in Amiritic, the latter explanation would seem more likely. Thus, if one learned the alphabet order from diction from an Amiritic source, the sound occupying 16th position of the alphabet would have been [θ], even if it were written with the glyph that originally signified [ts]. If our author was arranging the Thamudic B letters in an ASA, and in particular Amiritic, order, the Thamudic B glyph corresponding to this sound would have been ṯ. While this is the most likely scenario, one could also consider the possibility, following Knauf (2009: 479), that *ṯ and *s³ merged to an affricate in this particular variety of Thamudic B, as in Taymanitic. If this were the case, then Thamudic B ṯ would most closely resemble the original value of *s³ [ts] against [θ], and would therefore have been the most suitable glyph to fill the position of the plain affricate in the ASA letter order. Such a scenario could indirectly be supported by the anthroponym ḥrt. If this name was not drawn from an Aramaic source, but instead represents a Thamudic-internal attempt at approximating Arabic ḥāreṯ, then it would indicate that Thamudic B t was the closest approximation of Arabic [θ], in turn suggesting that the glyph ṯ had some other value, perhaps [ts]. Attractive as this might seem, such a scenario seems far less likely since, unlike Taymanitic, the varieties carved in Thamudic B do not merge *ḏ and *z, and it seems that in other circumstances the reflex of *s³  is written with the s¹ glyph.15 An Amiritic source is especially appealing, as the Amiritic inscriptions come mostly from the Jawf and reflect the northern most dialect of ASA. Several linguistic peculiarities point towards a non-ASA admixture/substratum, including the use of t rather than k suffixes on the suffix conjugation, negation with lm, and the preposition mn ‘from’ in place of bn (see Stein 2004). These have led some to posit an Arabic substratum, but this would stretch the evidence too far. These features are typical of any non-ASA Arabian language and cannot, in and of themselves, identify the source of contact. If, however, the particular ordering of the letters in this inscription is indicative of an Amiritic provenance, then it may suggest that the linguistic influence on that particular 15  For example, the word s¹mk occurs in Thamudic B which seems to be a reference to the season smāk (mid-Febuary–mid-April), Eut 561 b rḍw ʿdd s¹mk ‘by Rḍw comes the abundance of Smāk’. The word is surely derived from the root √s³mk (see Beeston et al. 1982: 138).

A Thamudic B Abecedary in the South Semitic Letter Order

13

dialect of Sabaic is connected to a variety inscribed in the Thamudic B script, and not Arabic at all. A connection between the two must remain tentative at the current moment, but this inscription sheds a small ray of light on the diverse ways the languages of the nomads and sedentaries can entwine. Unlike the texts discussed in Macdonald (1986), the Thamudic B abecedary does not appear to be a practice exercise or the consequence of a writing error (ibid. 102). That it occupies exactly one half of the face of the rock suggests that it was an intentional component of Text 1. It is possible that the author inscribed the letters in a foreign order to serve some magical purpose, but with no parallels in other ANA inscriptions, any further remarks would amount to little more than speculation. It is hoped that future finds will shed light on the religious use of writing among the inhabitants of the desert. 4 Sigla Esk HU

Taymanitic inscriptions in Eskoubi 1999. Taymanitic, Hismaic, and Thamudic B, C, and D inscriptions copied by C. Huber and renumbered in van den Branden 1950. JSTham Taymanitic, Hismaic, and Thamudic B, C, and D inscriptions in Jaussen and Savignac 1909–1922. WTI Dumaitic, Hismaic, and Thamudic B, C, and D inscriptions in Winnett and Reed 1970. Bibliography Al-Jallad, Ahmad. 2015. “Graeco-Arabica I: The Southern Levant,” in F. BriquelChatonnet, M. Debié, and L. Nehmé (eds.), Le context de naissance de l’écriture arabe. Écrit et écritures araméennes et arabes au 1er millénaire après J.-C. (OLA) Louvain. Beeston, Alfred F. L., Mahmoud A. Ghul, Walter W. Müller, and Jacques Ryckmans. 1982. Sabaic Dictionary (English-French-Arabic). Louvain-la-Neuve. Biella, Joan Copeland. 1982. Dictionary of Old South Arabic: Sabaean Dialect (HSS 25). Chico, CA. Bordreuil, Pierre and Dennis Pardee. 1995. “Un Abécédaire du type sud-sémitique découvert en 1988 dans les fouilles archéologiques françaises de Ras ShamraOugarit,” CRAIBL: 855–60. Eskoubi, Khaled. 1999. An Analytical and Comparative Study of Inscriptions from “Rum” Region, South West of Tayma. Riyadh. (in Arabic)

14

Al-Jallad and Al-Manaser

Faber, Alice. 1984. “Semitic sibilants in an Afro-Asiatic context,” JSS 29: 189–224. ———. 1985. “Akkadian Evidence for Proto-Semitic Affricates,” JCS 37: 101–107. ———. 1992. “Second Harvest: šibbōliθ Revisited (Yet Again),” JSS 37: 1–10. Harding, G. Lankester. 1971. An Index and Concordance of Pre-Islamic Arabian Names and Inscriptions (Near and Middle East 8). Toronto. Hayajneh, Hani. 2011. “Ancient North Arabian,” in Stephan Weninger with Geoffrey Khan, Michael Streck, and Janet Watson (eds.), The Semitic Languages: An International Handbook (Handbücher Zur Sprach- Und Kommunikationswis­ senschaft 36). Boston-Berlin. 756–782. Hayajneh, Hani and Josef Tropper. 1997. “De genese des altsudarabische Alphabets,” UF 29: 183–98. Hoftijzer, Jacob and Karel Jongeling. 1995. Dictionary of the North-West Semitic Inscriptions (Handbuch der Orientalistik, Erste Abteilung: Der Nahe un Mittlere Osten, 21). Leiden. Jamme, Albert. 1968. “The Thamudic Verb ‘Tm,” JAOS 88: 290–94. Knauf, E. Axel. 2009. “Thamudic,” EALL 1.477–483. Lane, E. W. 1863–1893. An Arabic-English Lexicon. London. Macdonald, Michael C. A. 1986. “ABCs and Letter Order in Ancient North Arabian,” Proceedings of the Seminar for Arabian Studies 16: 101–70. ———. 2000. “Thamudic,” EI2. ———. 2004. “Ancient North Arabian”, Roger Woodard (ed.), The Cambridge Encyclopedia of the World’s Ancient Languages. Cambridge. 488–533. ———. 2010. “Ancient Arabia and the Written Word,” in Michael C. A Macdonald (ed.), The Development of Arabic as a Written Language (Supplement to the Proceedings of the Seminar for Arabian Studies 40). Oxford. 5–27. ———. 2012. “Goddesses, dancing girls or cheerleaders? Perceptions of the divine and the female form in the rock art of pre-Islamic North Arabia,” in Isabelle Sachet and Christian Robin (eds.) Dieux et déesses d’Arabie Images et representations: Actes de la table ronde tenue au Collège de France (Paris) les 1er et 2 octobre 2007. Paris. 261–297. Macdonald, Michael C. A., Muna al-Mu’azzin, and Laïla Nehmé. 1996. “Les inscriptions safaïtiques de syrie, cent quarante ans après leur découverte,” CRAIBL: 435–494. Rawan, Schirin Faisal. 2013. Neue Safaitische Inschriften aus Süd-Syrien (Semitica et Semitohamitica Berolinensia 16). Aachen. Sass, Benjamin. 2005. The Alphabet at the Turn of the Millennium. The West Semitic Alphabet ca. 1150–850 BCE. The Antiquity of the Arabian, Greek and Phrygian Alphabets (Journal of the Institute of Archaeology of Tel Aviv University Occasional Publication 4). Tel Aviv. Stein, Peter. 2004. “Zur Dialektgeographie des Sabäischen,” JSS 49: 225–45. Winnett, Fidelity V. 1987. “Studies in Ancient North Arabian,” JAOS 107: 239–44. Winnett, Frederick Victor and William LaForest Reed. 1970. Ancient Records from North Arabia (Near and Middle East 6). Toronto.

A Thamudic B Abecedary in the South Semitic Letter Order

FIGURE 1  Photograph of the basalt slab bearing three Thamudic B inscriptions, courtesy of Sabri Abbadi.

FIGURE 2  Hand copy of the three Thamudic B inscriptions, by Ahmad Al-Jallad.

15

Ethiopian Semitic and Cushitic. Ancient Contact Features in Ge‘ez and Amharic David Appleyard It is probably impossible to find an Ethiopian language not influenced by [an]other language or languages. (Zaborski 1991: 123)

The Semitic languages of Ethiopia have evolved, influenced, and been influenced by a range of Cushitic languages over an extensive period of two and a half millennia and probably longer. This obvious fact has been recognized for almost as long as Ethiopian Semitic languages have formed the object of scholarly study, and already in the nineteenth century the influence was noted of the Cushitic languages of the region on Ge‘ez and other languages, notably Amharic and Tigrinya. In two articles Praetorius (1889, 1893) listed some twenty items of Ge‘ez vocabulary to which he ascribed a Cushitic origin, as well as two morphological formatives. Earlier in the century, others had spoken of Amharic as “degenerate” from the supposed Semitic “type” or as “sub-Semitic,” based somewhat impressionistically on syntactic patterns, phonology, and presumably lexicon, too, implicitly seeing the influence of Cushitic languages at work. In the above remarks it may seem that there has been the assumption that the influences, the contact features, have passed from Cushitic to Ethiopian Semitic (hereafter ES). However, features may of course move in both directions between languages in contact. Whilst it might be relatively easy to identify a lexical item, for instance, Ge‘ez dorho ‘chicken’, as a loanword that has come from a Cushitic source because no cognates occur in Semitic languages outside Ethiopia, whereas there are similar terms in various Cushitic languages, it is less easy always to be so certain about the direction of borrowing of many non-lexical features. For instance, all ES languages, Ge‘ez as well as the modern languages aside from Tigre, have a special converb form, traditionally called the gerund or gerundive, to denote a sequentially subordinate action or event to the main verb. The exact formation of the converb is, however, not the same across Ethiopian Semitic, and the various formations can be explained wholly within the terms of ES morphology. The category of converb does not, however, exist elsewhere in Semitic, but is present in the verbal systems of many Cushitic languages. The assumption is that ES devel-

© koninklijke brill nv, leiden, ���5 | doi ��.��63/9789004300156_003

Ethiopian Semitic and Cushitic

17

oped a converb “under Cushitic influence” (Hetzron 1972: 99), and though it is likely, it is not so readily provable. More will be said about the converb below. Equally difficult to ascribe to a particular origin are semantic parallels between ES and Cushitic, for example, as exemplified by the widespread use across the region of a causative derivative of the verb ‘want’ in the sense of ‘be necessary’, e.g. Amh. asfällägä ‘it is necessary’ from fällägä ‘he wanted’ and Oromo barbaachisa ‘it is necessary’ from barbaada ‘he wants’. It cannot be said that one is explicitly a borrowing from or rather a calque on the other, that is in which direction the feature is copied. This kind of correspondence leads naturally to a consideration of the question of an Ethiopian language area, in which within a given region languages, which are not very closely related, show parallels, implicitly due to contact. The idea of an Ethiopian language area was originally expounded by Ferguson in two slightly different versions (1970, 1976) and was further refined by Zaborski (1991) into several subareas. The language area concept does not of course necessarily and simply propose to identify a given feature as borrowed from language X into language Y, but encompasses a geographical area in which languages (i.e. their speakers) are in contact and have influenced one another, including the possibility of parallel developments beyond the notion of direct, conscious borrowings. The Ethiopian region, which can be extended beyond the confines of the Ethiopian and Eritrean political entities, is one of the world’s more complex linguistic regions in terms of both the number of languages and the different language families involved. Alongside the ES languages, which number up to 23 varieties1 grouped into three principal branches, there are over 30 Cushitic languages, properly speaking, and more if varieties are taken into account, stretching from the far south of Egypt to Tanzania. Not all of these are relevant to the issue of contact with ES, of course. In addition, there are also Omotic and a number of Nilo-Saharan languages within the “Ethiopian” language area, which again do not seem to be relevant to the issues of language contact with ES that is the subject of the present study. However, when it comes to looking for those languages that have been in contact with ES over a long period of time and which demonstrate contact features, beyond isolated lexical borrowing, the number is greatly reduced. The time factor is important here as well as the nature of the contact features, because whilst it can be demonstrated that Amharic, for instance, has borrowed lexicon from a range of Ethiopian languages and that a 1  I use the neutral term ‘variety’ so as not to prejudge what are languages or dialects. Some of the separate varieties subsumed under the name Gurage are probably better described as dialects. The figure of 23 ES varieties is drawn from Hetzron’s 1972 study. For Cushitic, see inter alia Appleyard 2012: 198–201.

18

Appleyard

possibly even larger number of languages has borrowed lexicon from Amharic because of the language’s dominant position over recent centuries, the languages with which Amharic has historically been in contact and where that contact appears to have given rise to features other than new vocabulary are even fewer. We do not know for sure what Cushitic languages were spoken where in the region in which Amharic developed, or for that matter what languages were spoken where Ge‘ez and its geographical successor Tigrinya developed, but the most obvious and generally agreed candidate is an earlier form or forms of the Central Cushitic or Agaw languages, which today constitute small islands mostly within the Amharic-speaking area.2 By the 14th century it has been conjectured that Agaw languages formed a block between the early Tigrinya and Amharic speaking regions (Haberland 1965: 14–16), and the name Agaw, in the form Athagaus (Αθαγαους), already occurred on the 3rd-century Monumentum Adulitanum and as Atagäw in one of the 4th-century Axumite inscriptions of ‘Ezana (Appleyard 2006: 2), although the exact location of these Atagäw is uncertain. There are also some possible traces of contact, in the form of lexical borrowing, with an East Cushitic language or languages already in Ge‘ez, whose sole representatives today in Eritrea and northern Ethiopia are the closely related Saho and ‘Afar. Amharic and its closest relatives also show some lexical borrowing from an undefined Highland East Cushitic language indicative of an old, pre-modern contact. There are contact features between ES and Cushitic languages at every level of linguistic analysis, not just in the lexicon, but also in phonetics and phonology, syntax, semantics, and morphology, both inflectional and derivational. I do not intend simply to provide an itemization of such features, however. For one thing, if the list were exhaustive it would be too long for this short study, as well as being not especially illuminating if we are to try and make an estimation of the mechanisms involved. Firstly, in what follows I shall focus chiefly on those features in Amharic and its oldest contact neighbours, namely the Agaw languages. At the same time, because I believe it is more interesting to look at the phenomenon of contact over as long a time span as possible, I shall also look at Ge‘ez, as the single ancient ES language of which we have evidence, referring also at times to Tigrinya, as the northern neighbour of Amharic, as well as to the closest relatives of Amharic, the other members of what has been called the Transversal South Ethiopic subgroup (Hetzron 1972: 36ff). Secondly,

2  One Agaw language, Bilin, is today spoken in Eritrea in contact with Tigre to the north and Tigrinya to the south. Tradition, however, ascribes the presence of the Bilin there to migrations from further south first in the 9th or 10th centuries.

Ethiopian Semitic and Cushitic

19

I shall focus on a number of morphological features in Amharic and these other languages that show signs of being contact features. 1 Lexicon Before proceeding to morphology, however, it might be instructive to look at a few instances of probable contact features in other areas in order to appreciate such issues as the “direction” of influence, the difficulty of identifying the precise “origin” of some features, and if possible the time depth or dating of contact influences. Since lexical borrowing is typically used as the first evidence of language contact, I shall begin with considering a few items firstly in Ge‘ez and then in Amharic that are fairly obviously loans, but for which the identification of a precise source is for various reasons not that easy, if at all possible. For Ge‘ez, a somewhat longer list is discussed in Appleyard 1978. There are also the comments in Leslau’s 1987 dictionary, though for some items without obvious Semitic cognates outside ES a list of Cushitic forms may be given with the note ‘in Cushitic’, or ‘from Cushitic’, or ‘passed into Cushitic’ without further discussion. For Amharic, however, there is still no etymological dictionary. My own studies, published in condensed form in Appleyard 1977, only go a little way towards the preparation of a true etymological dictionary. For other ES languages there is Leslau’s dictionary of Harari (1963) and of course his monumental dictionary of the Gurage languages (1979). Ge‘ez: śǝga ‘flesh, meat’: this is usually seen as a loan from Agaw (Proto-Agaw *sǝx-a)3 with the same meaning, though the g:x correspondence needs explaining. The situation is further complicated, on the one hand, by the fact that possible cognates of the Agaw term elsewhere in Cushitic show a medial ʔ (Proto-East Cushitic *soʔ-), and on the other hand, by the existence of a similar East Cushitic form *šaʕ- ‘cow’ (cf. also Beja šʔa ‘cow’ and ša ‘meat’). Yet another factor that seems relevant is the ‘Afar/Saho term saga ‘cow’, which shows the same medial as the Ge‘ez term, though of course has the meaning ‘cow’ and not ‘meat’. This puts one in mind of the discussions whether Ge‘ez lahm ‘cow, ox’ and its ES cognates are cognate with the Semitic root laḥm- with a general sense of ‘food’ despite the 3  I use established reconstructed proto-forms to avoid giving potentially long lists of cognate forms.

20

Appleyard

different medials. Incidentally, the common Semitic root *bśr ‘meat’ occurs in ES languages other than Ge‘ez, Tigre, Tigrinya and Amharic: Harari bäsär, etc., whilst Ge‘ez basor, baśor ‘flesh’ is a loan from Hebrew. ṣaʕda ‘white’: this looks like a loan from Agaw, though it is difficult to reconstruct a common proto-form from the extant forms (Bilin č’aʕǝd-, Xamtanga s’ar-, Kemant šay-) not the least because the pharyngeal ʕ in Bilin is usually confined to loanwords from ES except in a couple of problematic items (Appleyard 2006: 18). Even supposing a Proto- or “Preproto-Agaw” form the resulting reconstruction *ʦaʕd- cannot be easily connected with other Cushitic items, though it looks as if it ought to be related to Lowland East Cushitic ʕad- ‘white’, though leaving the initial consonant unexplained. ʕoqä ‘know’: this is likely to be from Proto-Agaw *-aq, which occurs only in Kemant ax- and in Awngi, as a relic prefix-inflecting verb which in turn has a sole cognate in the rest of Cushitic in Somali -aq, also a relic prefix-inflecting verb. The other Agaw languages use a different root whose Cushitic cognates have the sense of ‘see’. The Agaw root and its Somali cognate are monoconsonantal and show no evidence of an initial ʕ, even in Somali where the pharyngeal occurs as a phoneme. The Ge‘ez root has “created” a medial radical w to conform with the triradical nature of verbal roots, and this appears in modern ES cognates such as Amharic awwäqä, whilst in Argobba ōnqa and Harari āqa the w surfaces only in derived stems. The Ge‘ez root ʕwq cannot therefore be related directly to any extant Agaw or Cushitic root, but there can be little doubt, everything considered, that it is of Agaw origin. śǝrnay ‘wheat’: despite early attempts to link this with the common Semitic root *śʕr,4 this item and its ES cognates (Tigrinya sǝrnay, Amharic sǝnde, Muher sǝrre, Masqan sǝnne, etc.) is probably of Cushitic, and specifically East Cushitic origin. It has no cognates in Agaw other than obvious recent loans from ES. The precise origin of the Ge‘ez item is, however, unclear: Saho sinraa, ‘Afar sirray, Somali sareen, and even an unconfirmed Beja seraam. Of course, crop names readily come under the heading of Wanderwörter and the Cushitic words are quite likely of ES origin anyway. dorho ‘chicken’: this is probably of Lowland East Cushitic origin, with the closest cognates in ‘Afar dorrahi and Saho dorho, two languages which are of course spoken close to the region where Ge‘ez developed. Beja, 4  Dillmann in his 1865 dictionary of Ge‘ez. However, the loss of medial ʕ and the suffix –nay are difficult to explain.

Ethiopian Semitic and Cushitic

21

which appears to have entered the Eritrean region during Aksumite times, has a slightly different form, ʔandiirho. The Agaw languages have an obviously cognate form, which is however less close to Ge‘ez: ProtoAgaw *dirw-a. The front vowel of this and the Beja form is, however, reminiscent of Tigre derho. Amharic has doro, with a back vowel like Ge‘ez. In attempting to be more precise about the origin of the word in ES there remains the consideration that, like the preceding item, this word is a prime candidate to be a Wanderwort. Therefore, beyond saying that it is a culture word of the north-central Ethiopian and Eritrean region, greater precision as to its origin is impossible. Amharic: wǝha ‘water’: Amharic shares this item with most South Ethiopic languages (Argobba ähʷa, Chaha ǝxa, Inor ǝhǝ, etc.), though the original Semitic root is retained in Harari and the East Gurage languages as well as in North Ethiopic (Ge‘ez may, etc.) In the light of forms like Argobba ähʷa it is possible that the Amharic form is metathesized from *ǝhwa, thus approaching more closely the probable Agaw origin in such as Kemant axʷ from Proto-Agaw *ʔaqʷ. However, it is perhaps coincidental, but Highland East Cushitic languages have a term with initial w, traceable to a common proto-form *waʔa, cognate with an East Cushitic root *waʕ-/wiʕ- meaning ‘flowing water’. čǝggǝr ‘difficulty, trouble, want’, also čǝggar ‘famine, deprivation’ and the associated verbal root čgr ‘be difficult’: this has no Semitic cognates and few ES cognates, which suggests it might be a very localized borrowing. In Agaw there is a range of forms meaning ‘be hungry’ which are similar (Bilin tǝgǝrt-, Xamtanga gǝdǝrz-, Kemant dǝkǝrt-, etc., all derived stems) but not identical to the Amharic. There the origin of the palatalization can be explained as originating in the B-type verb stem *teggär- > čäggär- (cf. Argobba čeggär-), which suggests a fairly old loan, further supported if time is allowed for the various metatheses and voicing/ devoicing developments seen in the Agaw forms when compared with the more archaic Bilin. gulbät, gʷǝlbät ‘knee’: cognates of this item occur throughout South Ethiopic but not in North Ethiopic aside from Tigrinya gʷǝlbät ‘strength’ which is an obvious borrowing from Amharic as its meaning indicates (‘strength’ is a secondary meaning in Amharic). All the ES forms, aside from Harari gǝlib, which has a different vocalization, can be derived from a common *gʷǝlbät. Cognates also occur throughout Cushitic: East

22

Appleyard

Cushitic: *gilb-/gulb-: Oromo jilba, Somali jilib, Konso kilpa, Saho and ‘Afar gulub, Sidaama gulube, Hadiyya gurubbo, Burji gilba, and so on; in Agaw, however, the common form can derived from *gǝrb- without any rounding, and in Beja there is gumba, with rounding but with a nasal instead of l or r. So, the closest of the Cushitic roots to the ES terms is East Cushitic *gulb-, but no more precise origin can be specified and the borrowing may therefore be ancient. It is interesting, by the way, that the Semitic ending -ät looks to have been added as a substitute for Cushitic -a. The inherited Semitic root is retained in North Ethiopic as the regular term for ‘knee’: Ge‘ez bǝrk, etc., but Harari bǝrxi/bǝxri ‘the unit between two joints’ (in a finger, sugar cane, etc.), bǝrk in East Gurage, may or may not be a retention with specialized meaning or a borrowing; Amharic does have bǝrk, used only and rarely in literary language in the sense of ‘knee’, but meaning more commonly ‘trembling, hysteria, loss of strength due to fright’, which in turn relates to the derived verb täbräkäkkä ‘tremble, shake with fear, give way (legs due to fright)’. Further complicating the picture in Amharic is the verb bärräkä ‘kneel, bend’, which again could be a retention of the root *brk, or an amplification of the root borrowed from Ge‘ez. These few examples give an idea of the complexities involved in tracing the histories of some lexical borrowings. Before leaving the lexicon, it might be instructive to mention briefly a couple of short surveys, one dealing with Amharic (Appleyard 1979) and one looking at Ethiopian Semitic as a whole (Bender 1986). The former took as its corpus firstly one of the major dictionaries of Amharic then available and secondly a range of texts. In the first instance it appeared that almost 73% of the root entries in the dictionary whose origins could be safely identified were inherited Semitic, and only 12% could be assigned a Cushitic origin. However, the percentage of roots whose origin could reasonably be identified was as little as 55%. The percentage of inherited Semitic roots in the texts, furthermore, rose to 84% of those whose origin could be identified and almost 60% of the total. Amharic, and especially the written language, has been expanding its lexicon for many years using Ge‘ez as a resource in much the same way as the classical languages have been used in building English lexicon. Bender’s survey, which operated with only a 100-item word list, perhaps unsurprisingly, concluded that Tigre and especially Tigrinya shared a higher percentage (77% and 80%, respectively) of items in common with Ge‘ez than did Amharic (72%) or the various so-called Gurage languages (60%). Nonetheless, these figures do show that Modern ES does retain a higher proportion of Semitic lexicon than initial impressions may suppose.

Ethiopian Semitic and Cushitic

23

2 Phonology When we turn to the fields of phonetics and phonology there are fewer probable contact features to discuss. The consonant system of Ge‘ez shows only a few changes from a putative Proto-Semitic, all essentially mergers of earlier separate phonemes. It is true that the consonant system of Amharic, for instance, is much further removed from the classical Semitic type, but many of the processes involved when viewed alongside Ge‘ez, for instance, can be explained by common developments seen in other languages, such as loss of laryngals, palatalization, and weakening of stops through fricativization (k > h, b > w, m > w) that do not need to be attributed to contact. In addition, it is something of a truism but it is worth reminding that the presence of glottalized phonemes in ES has nothing to do with Cushitic. It is true that most but not all Cushitic languages have some glottalized phonemes (Beja and Somali are exceptions) typically forming triads with plain voiceless and voiced stops and sometimes affricates. However, glottalized phonemes occur in Semitic outside the ES region in Modern South Arabian. There are, however, two features of the common ES system, one consonantal and one vocalic, that may be explained as due to contact with Cushitic, and specifically Agaw. ES developed a series of labialized velars or labiovelars (in Ge‘ez kʷ, gʷ, qʷ, ḫʷ), deriving from a plain velar in contact with an original (long) rounded vowel (Gragg 1997: 245). This development need not necessarily be a contact feature, but it is interesting that the Agaw languages also have a set of labiovelars and labio-uvulars (Proto-Agaw *kʷ, *gʷ, *ŋʷ, *xʷ, *qʷ, *ɣʷ).5 The Agaw labialized phonemes also appear to have arisen in the environment of a rounded vowel, though the paucity of certain lexical cognates with other Cushitic languages makes it difficult to give precise details. The other phonemic feature of ES that may be due to contact again with Agaw is the seven-term vowel system: ä, ǝ, a, i, u, e, o. When compared with Proto-Semitic, the process must have begun with the merger of the two short high vowels i and u into a high centralized ǝ, and the (partial ?) centralization of a, leaving phonemic vowel length redundant. So, in the traditional pronunciation of 5  I have retained the familiar transcription q for the ES glottalized velar stop [k’], whereas there is good reason to believe that Proto-Agaw *q was a voiceless uvular stop, as still in Awngi and apparently in late 19th-cent. Bilin. The voiced pair, transcribed as ɣ, i.e. [G], occurs only in Awngi. An interesting example of an ES to Agaw contact feature in connection with the former is that the modern pronunciation of the Bilin phoneme /q/ is as a glottalized [k’], evidently due to the influence of Tigrinya and Tigre, in one of which most Bilin speakers are bilingual.

24

Appleyard

Ge‘ez vowel length is not phonemic. However, Tigre preserves phonemic length in the contrast between a ([ɐ] and [a]), corresponding to Ge‘ez ä, and ā ([a:]), corresponding to Ge‘ez a.6 The same seven-term system occurs in most Agaw languages and can be reconstructed for Proto-Agaw (Awngi shows some further developments). Again, comparison with the rest of Cushitic suggests a similar development to that seen in ES. It is, however, less clear how the mid vowels e and o arose in Agaw, and they are comparatively rare. In Ge‘ez, of course, they arose from the monophthongization of the diphthongs ai and au, respectively. 3 Syntax As regards contact features in the area of syntax I shall have little to say here, other than the obvious statement that the predominant SOV, Head final structures of modern ES languages is closely paralleled throughout Cushitic, leading to the often stated and natural conclusion that such structures in the former are due to the influence of the latter. Ge‘ez, of course, retains a “typically Semitic” VSO, mostly Head initial syntax. One small point that is of interest here, however, is that whilst the common typological concomitant feature of an SOV, Head final system is the use of postpositions over prepositions, Amharic and most modern ES languages retain a small number of prepositions. These often combine with a larger number of postpositions, but in spoken Amharic there is a tendency to drop the preposition leaving just the postposition: bä-betu wǝsṭ or ǝ-betu wǝsṭ ‘in[side] the house’ becoming just betu wǝsṭ. In Harari the process of shifting to postpositions has been carried one step further insofar as original simple prepositions have become postpositions: mišti-zo-le ‘to his wife’, gār-zina-be ‘in our house’. This development may not necessarily be due to contact in itself, but is a natural progression of the SOV structure type. It is at the same time true that Cushitic languages all use postpositions and/or case suffixes. Another syntactic feature that may be due to Cushitic influence is the prevalence in Amharic and some other ES languages for focus marking using cleft-clause constructions (Appleyard 1989). Many Cushitic languages have a range of focus and topic marking constructions, the former often deriving from the same kind of cleft constructions seen in ES involving the same or similar forms of the verb as used in relative constructions and an overt or covert copula.

6  Phonemic vowel length, apparently secondary in origin, also occurs in Silt’i and in Peripheral Western Gurage (Gyeto, Inor, Endegen, and Ener).

Ethiopian Semitic and Cushitic

25

4 Morphology Turning now to morphology, contact features in ES concern primarily patterns and structures and not actual morphemes. The only instance of the latter, which was recognized already by Praetorius (1889: 318–21), comes from derivational morphology and is an abstract noun formative in -na. This exists throughout ES, including in Ge‘ez, where -nna is a frequent abstract noun suffix: qǝddǝsǝnna ‘holiness’, mǝlkǝnna ‘rule, lordship’, ṭäbibǝnna ‘philosophy, intellectuality’, and so on. Similar formatives occur across ES, often extended by -t, as in the Amharic abstract noun suffix -nnät: kǝfunnät ‘badness’, tǝhutǝnnät ‘humility’, alämawinnät ‘worldliness”, and so on. This morpheme probably has its origins in the common Agaw verbal noun formative -na (so in Bilin and Kemant, -änäw in Xamtanga and -äna as the instrument noun suffix). On the other hand, the Amharic suffix -nnät seems to have been “re-borrowed” back into Kemant, for instance, in the form -näy (showing the regular t > y shift in postvocalic position) as a formative for some abstract nouns and language names. A morphological development in Amharic that is shared with the other Transversal South Ethiopic languages is the loss of gender distinction in the plural in pronouns and in the inflection of verbs. The same occurs in the now extinct Gafat, an Outer South Ethiopic that early came under the influence of Amharic. All the other ES languages maintain the distinction between masculine and feminine in the plural. This development in Amharic and its closest relatives may be due to contact with Cushitic languages, though we cannot be specific about which as no Cushitic language makes a gender distinction in the plural. However, as the North Ethiopic languages and the Outer South Ethiopic languages other than Gafat retain gender distinction in the plural, the loss of the latter in Amharic et al. could simply be a random development. On more secure ground is the question of the converb or gerund(ive) in ES. This is a special form of the verb that is used for chaining clauses describing simultaneous or sequential events, as well as for various kinds of general adverbial subordination.7 Such converbs are found throughout ES, with the exception of Tigre, though the strategies and forms employed differ from one group of languages to another. Hetzron (1972: 99) says that the existence of converbs, etc., in ES is “clearly under Cushitic influence”, though here, too, the strategies employed differ. What we may call the archetypal converb in the Ethiopian context, a single distinct form fully marked for person and 7  A fuller discussion of the ES situation than is given here can be found in Hetzron 1972: 98–115, and the question of converbs in East Cushitic, where they are sometimes called Participles, is treated in Banti 2010.

26

Appleyard

number but unmarked for tense, occurs only in Ge‘ez, Tigrinya, Amharic, and Argobba. In origin, most clearly visible in Ge‘ez, this converb is formally the accusative case of a verbal noun in C1äC2iC3- to which possessive pronominal suffixes are added. To such a form the Amharic converb can be traced directly (with reduction of the stem to C1äC2C3-), whilst in Tigrinya the old accusative case vowel -ä- is dropped, and in Argobba an element -d- is further added to the stem C1äC2C3- before the personal endings which retain traces of the accusative marker. So, for example, corresponding to Ge‘ez säbiräkä ‘you having broken’ are Tigrinya säbirka, Amharic säbräh, and Argobba säbrǝdah. The “intrusive” -d- in Argobba is probably of the same origin as the -t- added as a substitute C3 in Amharic in verb stems whose third radical, typically an original laryngeal or semivowel, has been lost. This same -t- seems to occur also in the converb in Central and Peripheral Western Gurage languages,8 though here the personal endings are those of the perfect. Other South Ethiopic languages have quite a different form in the function of a converb, with a range of different tense and mood bases (so marking tense-mood unlike the Ge‘ez converb and its relatives), to which an ending -ma, -mä, -m, or -ani, -ane is added according to language. The situation is more complex than I can present here and Hetzron’s discussion (op. cit.) should be consulted for details. What is clear is that the old type of converb seen in Ge‘ez has been lost in a number of South Ethiopic languages and has been replaced by a more analytical form employing a common ES connective. This latter situation is reminiscent of the converb strategies in many East Cushitic languages. The closest parallels, however, in Cushitic to the archetypal ES converb are found in the Highland East Cushitic and Agaw languages, where there are distinct, fully inflecting paradigms that are not derived from main verb tenses. Within both language groups, however, there is no agreement of the actual forms but a range of different vocalic elements (in bold) occur after the personal markers: Sidaama ros-t-e ‘you/she having learnt’, Hadiyya it-t-a ‘you having eaten’, it-t-aʔa ‘she having eaten’, Kambaata fan-t-i ‘you/she having opened’, Bilin gäb-r-o ‘you/she having refused’, Xamtanga k’äb-ǝr9 ‘you/she having cut’, Awngi des-t-ata ‘you/she having studied’. A further parallel is that in Agaw, just as in Ge‘ez, Amharic, etc., the negative equivalent of the converb is quite disparate and is not formed simply by adding a negative morpheme. Therefore, I think whilst agreeing with Hetzron that the ES converbs were developed under Cushitic influence, we may reasonably safely conclude that the archetypal converb of Ge‘ez, Amharic, etc., emerged specifically under the influence 8  But see Hetzron’s reservations (1972: 103–4 and fn. 76.) 9  The Xamtanga form in fact has zero ending after the person marker, which is here -[ǝ]r.

Ethiopian Semitic and Cushitic

27

of Agaw, and because this converb exists already in Ge‘ez, the influence of Agaw here must be ancient. The last piece of morphological evidence for contact between ES, and in this instance specifically Amharic and its closest relatives (Argobba, Harari, and East Gurage) and Cushitic, again specifically Agaw, that I want to examine here concerns the creation of some distinct affirmative verb paradigms in main verb position. Cushitic languages, by and large, make a clear distinction between main indicative forms in the imperfect (or non-past) tense and subordinate ones. There is good reason to believe, however, that this development in Amharic, etc., is of much more recent date than others of the contact features that I have examined here. The evidence also makes it less easy to specify in which direction the influence has operated, whether the Amharic feature is due to Agaw influence or vice versa. So what are the details? In Amharic an imperfect form like yǝsäbǝr can only occur either as part of a subordinate verb complex: yämm-isäbǝr ‘who breaks’, b-isäbǝr ‘if he breaks’, or in conjunction with negative affixes: a-ysäbr-ǝm10 ‘he doesn’t break’, yämm-aysäbǝr, ‘who does not break’, b-a-ysäbǝr ‘if he does not break’, etc. In affirmative main verb position this must be combined with an auxiliary element which transparently derives from the non-copular or existential verb ‘to be’: yǝsäbrall ‘he breaks’. On the other hand, the perfect säbbärä may occur in all positions, main and subordinate, affirmative and negative. When we look at other modern ES languages with respect to a distinction between main and subordinate verb inflection, we find that there are broadly three types: (1) languages such as Tigre and Tigrinya that make no formal distinction between main and subordinate verbs, and of course the same applies to Ge‘ez; (2) languages like Amharic that only make a distinction in the affirmative imperfect or non-past, employing a main verb marker (MVM) deriving from the non-copular verb ‘to be’; lastly (3) languages which make a distinction in both the perfective (past) and the imperfective (non-past), employing main verb markers which are of different origin, probably derived from a copula. The situation is, however, a little more complex, as Modern Harari, Silt’i, Zay, and Wolane, which are all classified as Transversal South Ethiopic languages along with Amharic and Argobba, use the main verb imperfective in adnominal (relative) constructions unlike Amharic and Argobba. Additionally, Zay also distinguishes perfective main verbs from subordinates, like many Outer South Ethiopic languages. Details can be found in Appleyard 2002, but the following table will illustrate the range of forms: 10  Amharic subordinate negatives differ from main verb negatives only in the absence of -m[m].

28

Appleyard Past Main verb

Non-past Subordinate

Tigrinya Amharic Argobba Harari

säbärä säbbärä säbbära sabara

Silt’i

sabara

Wolane

säbärä

Zay

säbärä-nu

säbär[ä]

Main verb

Subordinate

yǝsäbbǝr yǝsäbr-all yǝsäbǝr yǝsäbr-äl yǝsäbǝr yisabr-aal yisabri-z-aal 11 yisabri isabr-aan isabr-aan isabr yǝsäbr-an yǝsäbr-an isäbr yǝsäbǝr-äl ~ yǝsäbǝr-äl yǝsäbr-ǝna yǝsäbǝr

As mentioned above, the main verb marker (MVM) derives from the noncopular verb ‘to be’, transparently so in Amharic and Argobba where the forms of the marker are identical to the independent verb except in the 3.M.SG where, for instance, Amharic -all occurs instead of allä. The suffixes in Harari and the East Gurage languages are obvious reductions: e.g. Wolane tǝsäbr-at ‘she breaks’ but ʔalät ‘she is’; tǝsäbrw-ahum ‘you [pl.] break’ but alähum ‘you [PL] are’; the suffix -an of the 3.M.SG and 3.PL seems to show a shift l > n, which is well known from other Gurage languages. The MVMs in Zay -nu and -na are of different origin, like those in the Outer South Ethiopic languages, where they involve the same elements n, t, and u as the suffixed copula in those languages, in which their origin may most likely be found.12 We are fortunate that for Amharic and Harari, at least, there are records of earlier stages of both languages, in both cases dating back to between the 16th and 18th centuries. Here we see a somewhat more fluid situation in which the unaugmented affirmative imperfective or non-past could also be used in main verb position. Lastly, before looking at the situation in Agaw, it should be said that compounds of 11  Adnominal (relative) where -z- is the relative marker. In Silt’i, Wolane and Zay there is zero relative marker. 12  I shall not examine the Outer South Ethiopic forms here, nor possible similarities of pattern in Highland East Cushitic. More details can, however, be found in Appleyard 2002: 18–19.

Ethiopian Semitic and Cushitic

29

the imperfective and the existential verb ‘to be’ also occur in Tigre and Tigrinya, as well as a handful of rare instances in Ge‘ez, but these do not have the syntactic function of marking main verbs, but are aspectual (durative, progressive, habitual, or immediate), as Tigre nǝtfarrar hallena ‘we are going out’, lǝmayto hallaw ‘they (regularly) die’, and whilst Tigrinya ʔǝxäyyǝd ʔalloxu is the exact formal equivalent of Amharic main verb ǝhedallähu ‘I go, am going, shall go’, it has the function of marking immediacy, ‘I go now, am going now’. Moving to Cushitic, all the Agaw languages make a clear formal distinction between verb forms, negative as well as affirmative, that are used in main and in subordinate clauses. The situation in Agaw is, however, rather more complex than what we find in ES. All Agaw languages also have a very highly developed system of subordinate verb paradigms, with two separate sets of adnominal subordinate paradigms according to whether or not the head noun is co-referential with the subject of the subordinate clause or not—i.e. Subjectival and Oblique (or Complemental in Hetzron’s terminology) Relatives—and in addition an extensive series of adverbial subordinate paradigms which incorporate the marker of the type of subordination such as temporal, causal, final, terminative, conditional, etc. Whilst the underlying system is essentially the same in all Agaw languages, there are a few significant differences, particularly between Awngi, on the one hand, and the remaining languages, on the other hand. The most important of these concern the range and form of tenses used in main verb position and the range of differing types of negative marking. Especially relevant to the present discussion is that the indicative affirmative main verb paradigms are constructed with a different suffix, marking tenseaspect and person, from verbs used in subordinate functions, and this main verb suffix is clearly cognate with one of the Agaw verb roots meaning ‘to be’: e.g. Bilin ʔak- ~ ʔäx- (imperfect and perfect aspect, respectively). Unlike ES, a distinction is also made in the negative, where a different negative marker occurs in main and subordinate verbs, without any ‘to be’ element. So for example, in Bilin we have k’ʷal-äkʷǝn ‘I see’ and k’ʷal-ǝxʷǝn ‘I saw’, k’ʷal-äli ‘I do not see’ and k’ʷal-li ‘I did not see’ (all main verbs), but as examples of subordinate verbs: k’ʷal-ä-n ‘if I see’ or k’ʷal-ä-xǝr ‘I who see/saw’ and k’ʷaläg-ä-n ‘if I do not see’, k’ʷal-äg-ä-xǝr ‘I who do not see’ and k’ʷal-g-ä-xǝr “I who did not see’. The situation in Xamtanga and Kemant is much the same as in Bilin, but in Awngi there is a more complicated, less clear-cut differentiation, though it is still true to say, following Hetzron, that one can distinguish between main verbal forms capable of forming an independent predicate, and subordinate verbal forms that depend on a main form. Unlike Bilin, Awngi has four basic affirmative main verb tenses (using Hetzron’s terminology): Perfect Definite, Perfect Indefinite, Imperfect Definite, and Imperfect Indefinite.

30

Appleyard

The difference between ‘definite’ and ‘indefinite’ revolves around whether the speaker vouches for the certainty of the event. What is especially interesting, however, is that the ‘definite’ tenses are marked by a suffix deriving from the cognate of the ‘to be’ verb that appears in the affirmative main verb forms in the other languages, whilst the ‘indefinite’ tenses have a simple, ostensibly older suffix that is not dissimilar to tense/aspect and person marker of subordinate forms. There are also differences in the negative verb forms in Awngi which need not concern us here. The fairly close parallels between Amharic with its closest relatives and Agaw in the use of a non-copular verb ‘to be’ to mark affirmative main verb forms must surely be a contact feature. Compounds with ‘be’ do occur in other Cushitic languages but as in Tigre and Tigrinya these have an aspectual function. In this connection, it is worthwhile mentioning that Amharic, for instance, has a compound form comprising the converb and an often reduced form of the same verb ‘to be’, but this, though indeed confined to main verb position, has an aspectual, perfective meaning: nägrän + all > nägränall ‘we have spoken’, like the equivalent Tigrinya compound form: räxibna ʔallona ‘we have found’. Interestingly, identical structures (converb + ‘be’) occur in Xamtanga, wašǝŋ-kʷuŋ, and Kemant wasǝn-wanäkʷǝn ‘they have heard’, perhaps the two Agaw languages most influenced by Amharic. In sum, therefore, although there is not an exact match, it would seem obvious that the ES and Agaw structures to mark affirmative main verbs are a contact feature, though it is perhaps impossible to say for sure in which direction the influence passed. I would, however, suggest that the influence came from Agaw, for two reasons. Firstly, in Agaw the ‘be’ constituent is added directly to the verb stem and there is no trace of verbal inflection on the stem having been ‘eroded’, and the construction may therefore be quite old, whilst in ES both elements are inflected and if there is any ‘erosion’ it is on the ‘be’ auxiliary. Secondly, the evidence of North Ethiopic suggests that the compound imperfect + ‘be’ was already in existence with an aspectual function, and it would be more economical to envisage a shift of function of an already existing structure in Amharic, etc., under the influence of a contact language or languages, i.e. Agaw. 5

Concluding Remarks

In conclusion, we have seen that ES languages show signs of contact with Cushitic languages at all levels, from phonology to syntax, from lexicon to morphology, and influence that has been continuous from the early times of Semitic presence in the Horn of Africa. There are without doubt many overlapping

Ethiopian Semitic and Cushitic

31

layers of influence and many centers of contact with some features being closely regionally defined and others of a wider currency. The linguistic politics of the region have obviously also played a part at various times, with Ge‘ez first being the dominant language of the Aksumite State and then for many centuries the language of the Ethiopian Church, and then with Amharic later assuming the same role in the Ethiopian kingdom. Cushitic languages have influenced ES languages and vice-versa, and amongst ES languages Ge‘ez has influenced Amharic as a language of learning and Amharic has influenced other ES languages as the language of the rulers. On the side of Cushitic, since its appearance in the Ethiopian region in the 16th century Oromo has achieved the role of lingua franca in certain areas of the country, in the center and south, and as such has influenced other Cushitic as well as ES languages, especially in the field of lexicon. Saying this, we may return to the quote which headed this brief study: “it is probably impossible to find an Ethiopian language not influenced by [an]other language or languages” (Zaborski 1991: 123). Bibliography Appleyard, David L. 1977. “A comparative approach to the Amharic lexicon,” Afroasiatic Linguistics 5: 43–109. ———. 1978. “Linguistic evidence of non-Semitic influence in the history of Ethiopian Semitic: lexical borrowing in Ge‘ez and other Ethiopian Semitic languages,” Abbay: documents pour servir à l’histoire de la civilization éthiopienne 9: 49–56. ———. 1979. “A statistical survey of the Amharic lexicon,” JSS 24: 71–97. ———. 1989. “The relative verb in focus constructions: an Ethiopian areal feature,” JSS 34: 291–305. ———. 2002. “The Morphology of Main and Subordinate Verb Form Marking, with Special Reference to Ethiopian Semitic and Agaw,” Afrikanistische Arbeitspapiere [AAP] 71: 9–31. ———. 2006. A Comparative Dictionary of the Agaw Languages (Kuschitische Sprachstudien 24). Köln. ———. 2012. “Cushitic,” in Lutz Edzard (ed.), Semitic and Afroasiatic: Challenges and Opportunities (PLO 24). Wiesbaden. 199–295. Banti, Giorgio. 2010. “Remarks on the typology of converbs and their functional equivalents in East Cushitic,” in Sascha Völlmin, Azeb Amha, Christian Rapold, Silvia Zaugg-Coretti (eds.), Frankfurter Afrikanistische Blätter 20. Köln. 1–51. Bender, M. Lionel. 1986. “Lexical retention in Ethio-Semitic: checking up on a myth,” in Joshua A. Fishman et al. (eds.), The Fergusonian Impact. New York and Amsterdam. 1.291–299.

32

Appleyard

Ferguson, Charles A. 1970. “The Ethiopian language area,” JES 8: 67–80. ———. 1976. “The Ethiopian language area,” in M. L. Bender, J. D. Bowen, R. L. Cooper, and C. A. Ferguson (eds.), Language in Ethiopia. London. 63–76. Gragg, Gene. 1997. “Ge‘ez (Ethiopic),” in Robert Hetzron (ed.), The Semitic Languages. London – New York. 242–260. Haberland, Eike. 1965. Untersuchungen zum äthiopischen Königtum (Studien zur Kulturkunde 18). Wiesbaden. Hetzron, Robert. 1972. Ethiopian Semitic. Studies in Classification (JSS Monograph 2). Manchester. Leslau, Wolf. 1963. Etymological Dictionary of Harari (Near Eastern Studies 1). Berkeley and Los Angeles. ———. 1979. Etymological Dictionary of Gurage (Ethiopic) (3 vols.). Wiesbaden. ———. 1987. Comparative Dictionary of Ge‘ez. Wiesbaden. Praetorius, Franz. 1889. “Hamitische Bestandtheile im Aethiopischen,” ZDMG 43: 317–326. ———. 1893. “Kuschitische Bestandtheile im Aethiopischen,” ZDMG 47: 385–394. Zaborski, Andrzej. 1991. “Ethiopian language subareas,” in Stanisław Piłaszewicz and Eugeniusz Rzewuski (eds.), Unwritten Testimonies of the African Past. Proceedings of the International Symposium held in Ojrzanów n. Warsaw on 07–08 November 1989. Warsaw. 123–134. ———. 2010. “Language subareas in Ethiopia reconsidered,” Lingua Posnaniensis 52: 99–110.

Hebrew Adverbialization, Aramaic Language Contact, and mpny ʾšr in Exodus 19:18 Samuel Boyd and Humphrey Hardy Adverbializers are a subclass of subordinators, or subordinating conjunctions, that mark an intra-clausal, adverbial relation.1 Hebrew expresses adverbialization in a variety of ways. In this article, two syntactic patterns of adverbialization with prepositions are surveyed from the perspective of historical linguistics. Both of these arrangements are found in the earliest strata of Hebrew. Later evidence demonstrates a shift in the usage of one pattern in favor of another. We propose a historic process through which these patterns emerged in Biblical Hebrew (BH), track the diachronic distribution in Hebrew, demonstrate the role of language contact on the changes, and apply this frame work to a particularly difficult interpretive problem in the Hebrew Bible. 1

The Development of Adverbializers

The primary means of adverbialization in BH is a function of prepositions. Two productive structures consist of prepositions governing a following clause or an embedded relative clause: (1) PREP+S and (2) PREP+REL+S. (1) ‫ַא ֲח ֵרי ֵה ַסּבּו אֹתֹו‬ ʾaḥăre hesabbu ʾoṯo AFTER-PREP lead.around-SC.3.M.PL DOM+it ‘After they led it around’ (1Sam 5:9)

* Abbreviations: 1 = first person; 3 = third person; advz = adverbializer; BH = Biblical Hebrew; c = common gender; dom = direct object marker; LBH = Late Biblical Hebrew; m = masculine gender; N = noun; np = noun phrase; pc = prefix conjugation; pl = plural; pp = preposition phrase; prep = preposition; rel = relative; s = sentence; SBH = Standard Biblical Hebrew; sc = suffix conjugation; sg = singular. 1  The use of the term ‘adverbializer’ to indicate an element that converts a constituent into an adverb (Gensler 2000) is not in view here. Rather, the standard description of this term from language typology is employed designating a type of subordinating conjunction which forms an adverbial clause (Thompson and Longacre 1985).

© koninklijke brill nv, leiden, ���5 | doi ��.��63/9789004300156_004

34

Boyd and Hardy

(2) ‫ַעד ַׁש ַּק ְמ ִּתי‬ ʿaḏ šaq-qamti UNTIL-PREP REL+arise-SC.1.C.SG ‘Until I arose’ (Judg 5:7) The syntax of these patterns finds its origin in the Semitic noun phrase (NP), particularly, the genitive construction. Through examining these modifying structures, the development of adverbializers in Hebrew is detectable. In cross-linguistic comparison, the modification of nouns may be classified into three constructions: genitives, adjectives, and relatives.2 These may be exemplified in BH with the following: (3) a genitive (i.e. the Semitic construct phrase), (4) an adjective phrase, and (5) a relative construction. The first two expressions are types of NPs, and the last consists of a relative clause. (3) ‫ֹלהים‬ ִ ‫ְּבנֵ י ָה ֱא‬ bne hɔ-ʔɛ̆lohim sons.of the-gods ‘the gods’ sons’ (4) ‫ֶה ָה ִרים ַהגְ ב ִֹהים‬ hɛ-hɔrim hag-gḇohim the-mountains-M the-high-M.PL ‘the high mountains’ (5) ‫ּטֹובה ֲא ֶשר ָע ָשה ְל ָדוִ ד‬ ָ ‫ ַה‬ haṭ-ṭoḇɔ ʔăšɛr ʕɔśɔ lḏɔwiḏ the-good REL do-SC.3.M.SG FOR+David ‘the good that he did for David’ A fourth construction, though less frequent, is known in BH. The modification strategy in Example (6) evidences a NP (kɔl-yme ‘all of the days of’) in construct with a clause (hiṯhallaḵnu ʔittɔm ‘we walked with them’).3 It is weakly differentiated from the genitive construction. As such, the construction may demonstrate a mediating position between the syntax of genitive and relative constructions.

2  Gil 2013. 3  This expression has been understood alternatively as having a covert relativizer, that is, as a null-headed relative (Holmstedt 2001; 2006).

35

Hebrew Adverbialization, Aramaic Language Contact

(6) ‫ל־יְמי ִה ְת ַה ַל ְכנּו ִא ָתם‬ ֵ ‫ָכ‬ kɔl-yme hiṯhallaḵnu ʔittɔm all.of+days.of walk.about-1.C.PL WITH+them ‘all of the days we walked with them’ Goldenberg classes these kinds of constructions as examples of “endocentric modification” or simply “attributive.”4 He demonstrates the correspondence between the genitive and relative structures throughout the Semitic languages. Examples from Syriac and Akkadian are provided in Chart 1 with (a) a nominal head, (b) a pronominal head, and (c) a nominal head and an appositional coreferential genitive construction. In this paradigm, the first two constructions are parallel to Example (3) above and the third is parallel to Example (5). CHART 1

Attributive Noun Phrases Apposition

Syriac

Akkadian

(a) (b) (c) (a) (b) (c)

deḥlṯɔ

bītum

Head

Attributive

Gloss

deḥlaṯ ddbīt ša ša

alɔhɔ alɔhɔ alɔhɔ ilim ilim ilim

‘fear of god’ ‘N of god’ ‘fear of god’ ‘house of god’ ‘N of god’ ‘house of god’

Regarding phrases of the relative type as in Example (5), Goldenberg claims that these patterns are “shown to be syntactically equivalent to genitive nominal forms.”5 The constructions with a verbal attributive (Chart 2), then, express the same basic relation as those with a nominal attributive. It is also important to note that relatives are derivative in each of the Semitic languages.6 That is to say, the constructions with relatives are secondary and developed independently in the individual daughter languages. As a result, these common NP types need to account for the construction similarities.

4  Goldenberg 1998. 5  Goldenberg 1998: 56–57. 6  Huehnergard 2006.

36

Boyd and Hardy

CHART 2  Attributive Relative Clauses Apposition Akkadian (a) (b) (c) elippum

Head

Attributive

Gloss

kasap ša ša-

išqulu izzazzu uṭebbû

‘the money (that) he paid’ ‘N that stands’ ‘the boat that he sank’

Following Goldenberg, the forms of NP modification in BH may be collapsed into two construction types. Chart 3 presents these two BH varieties using the previous examples. Type 1 is the genitive construction with either a nominal or verbal attributive that may be headed by either a preposition or a relative. Type 2 is the attributive adjective construction—hɛ-hɔrim hag-gḇohim ‘the high mountains’. CHART 3  BH Modification of NPs Apposition Type 1

(a) (b) (c)

Type 2

haṭ-ṭoḇɔ

Head

Attributive

Gloss

bne kɔl-yme ʔăšɛr

hɔ-ʔɛ̆lohim hiṯhallaḵnu ʔittɔm ʕɔśɔ lḏɔwiḏ

hɛ-hɔrim

hag-gḇohim

‘the gods’ sons’ ‘all of the days we walked with them’ ‘the good that he did for David’ ‘the high mountains’

Classical grammatical and historical linguistic approaches have established that most BH prepositions originate from NPs.7 Aspects of the originating constructions are preserved even as these lexemes acquire new statuses and 7  Nearly every BH grammar presents the general word-class change of Noun to Preposition (see, e.g., Blau 2010: 283). Using grammaticalization theory, Hardy (2014) provides more details for the origin and development of BH prepositions that may be etymologically related to nominals.

Hebrew Adverbialization, Aramaic Language Contact

37

are incorporated into the linguistic system as function words. In particular, prepositions—and by extension adverbializers—emerge from these nounmodification constructions displaying the syntactic arrangement and morphology of their originating forms. The emergence of adverbialization in BH is situated at the nexus of genitive constructions and the development of prepositions. Combining these two components, a plausible development of adverbialization may be outlined. The syntax of BH prepositions patterns on the genitive constructions of the first type (Chart 4), corresponding to the previously discussed genitive of NPs. These preposition constructions consist of two varieties where the complement may be a NP (a) or a sentence either without a relative (b1) or with a relative (b2). In the case of the relative, the construction follows the genitive type of Chart 3 (b) kɔl-yme hiṯhallaḵnu ʔittɔm ‘all of the days we walked with them’, where the relative is the attributive in the larger construct phrase. The result is two productive BH patterns, PREP+S and PREP+REL+S, whereby adverbialization could be expressed. CHART 4  BH Prepositions PREP

Complement

Gloss

ʔaḥăre ʔaḥăre ʔaḥăre

bne hɔ-ʔɛ̆lohim hiṯhallaḵnu ʔittɔm ʔăšɛr ʕɔśɔ lḏɔwiḏ

‘after the gods’ sons’ ‘after we walked with them’ ‘after what he did for David’

Type 1

(a) (b1) (b2)

2

Linguistic Variation in Hebrew Adverbializers

The structure suggested in the previous section is based on the reconstruction of the genitive construction and the emergence of prepositions functioning as adverbializers. Before turning our attention to outlining the proposed contactinduced change, we shall review more closely the linguistic variation in BH and later Hebrew. What will be demonstrated is that the adverbializer constructions followed a clear development trajectory of incremental change. At each step, a lone pattern emerged from polysemy, and in turn the grammatical function was reshaped as a result of internal and external influences. BH evidences both patterns of adverbialization discussed previously. However, the patterns are not evenly distributed. The PREP+S pattern is found

38

Boyd and Hardy

fifty-nine times.8 Of the nine different prepositions (ʔɛl, ʔaḥar, ʔaḥăre, bilʕăle, kmo, kʕal, min, ʕad, and ʕal), the most frequent semantic relation is the temporal function—either AFTER (ʔaḥar/ʔaḥăre) or UNTIL (ʕad). Three-fourths of the adverbializers consist of this latter temporal element. The only “inseparable” preposition is the expanded independent form kmo ‘like, as’. The distribution of this PREP+S pattern within the BH corpus is diverse. There are nineteen examples in the Pentateuch and Deuteronomistic History, fifteen in the prophets, nineteen in the writings, and six examples in Chronicles and Ezra-Nehemiah. Isaiah (First and Second), Psalms, and Job witness the largest number of instances. The PREP+REL+S pattern is attested more widely within BH, both pertaining to types and tokens. Twelve different prepositions are evidenced amounting to more than seven hundred examples. The inseparable variety include: b- ‘in’, k- ‘like, as’, l- ‘to, for’, and me- ‘from’ (< min-). Eight separable prepositions are evidenced (ʔɛl, ʔɛt, ʔaḥar, ʔaḥăre, min, ʕad, ʕal, and taḥat). The overwhelming majority of these examples (515 instances) consists of a lone type, kaʔašɛr ‘just as’. The distribution of the PREP+REL+S pattern throughout the BH corpus is not altogether different from that of the previous construction. Comparing the two patterns, one finds considerable overlap in semantics, usage, and dispersion within the corpus. Seven of these constructions use the same preposition element (Chart 5). Six distinct prepositions are evidenced uniquely with PREP+S (bilʕăle, kʕal) or PREP+REL+S (ʔɛt, b-, l-, taḥat). Two lexemes demonstrate formal variances relating to the contrast of the independent form as compared to the dependent: kmo and kaʔašɛr; min and meʔašɛr. Both forms are found in early texts, so-called Standard Biblical Hebrew (SBH), and the later corpus, Late Biblical Hebrew (LBH). The only readily seen differences between the two constructions are semantic. This variance is based on the distribution of the specific preposition element where the PREP+S pattern is typically temporal and the PREP+REL+S pattern is comparative.

8  Gen 38:11; Exod 15:16 (2x); Lev 14:43; 25:48; Num 23:24; Deut 33:11; Josh 2:22; 10:13; 1Sam 1:22; 5:9; 14:19; 2Sam 10:5; 21:10; 1Kgs 11:16; 15:29; 2Kgs 3:25; 7:3; 10:17; 1Chron 15:12; 2Chron 21:15; 29:34; 36:21; Ezra 3:11; 8:29; Job 8:21; 14:6; 27:5; 32:11; 34:32; 42:7; Ps 57:2; 58:8; 71:18; 73:17; 94:13; 110:1; 132:5; 141:10; Prov 7:23; 12:19; 23:7; Is 22:14; 26:18, 20 (2x); 32:15; 42:4; 59:18; 62:1, 7 (2x); Jer 41:16; Lam 3:50; Ezek 28:15; 39:15; Dan 11:36; Hos 10:12; Zech 10:8.

Hebrew Adverbialization, Aramaic Language Contact

39

CHART 5  Prepositions found in Adverbializer Constructions PREP+S

PREP+REL+S

ʔɛl

ʔɛl ʔɛt ʔaḥar ʔaḥăre b-

ʔaḥar ʔaḥăre bilʕăle kmo kʕal min ʕad ʕal

klmeʕad ʕal taḥat

The linguistic variation seen with these patterns is not static. And adverbialization did not cease developing but continued to evolve in later strata of Hebrew. The evidence from the Hebrew of the Dead Sea Scrolls (3rd century BCE– 1st century CE) indicates a preference for the PREP+REL+S pattern to the exclusion of the other. These constructions include many of the same prepositions as in BH (ʔḥr ʔšr, ʔl ʔšr, ʔt ʔšr, bʔšr, bʕbwr ʔšr, lʔšr, mʔšr, mš, kʔšr, and ʕd ʔšr). The semantics of the most common construction kʔšr, however, have expanded from the BH usage as a comparative to include the temporal relation (e.g. CD-A 15: 12 wkʔšr yqym ʔwtw . . . ‘now when he sets himself up . . .’). In contrast, no example of the well-attested BH construction ʕad+S is found in these documents. Instead one finds ʕad ʔšr ‘until’ as its equivalent at Qumran.9 It is reasonable to conclude that the general pattern with the relative has expanded to be the sole adverbialization strategy in these documents. This change appears to be the result of an internal shift toward pattern regularizing in the linguistic strata subsequent to BH. With the early Rabbinic period (70 CE–3rd century CE), the situation had evolved further. The adverbializer pattern remained PREP+REL+S, but the 9  Five examples are extant in the Temple Scroll (11Q19 45:8, 17, 18: 50:3; 58:20).

40

Boyd and Hardy

lexeme functioning as the relative shifted from ʔšr to the monosyllabic proclitic š-.10 The former lexeme is known from biblical quotations and a handful of what appear to be static idioms, such as kʔšr—which itself is found only about a dozen times. In Rabbinic Hebrew, the latter pattern PREP+š(REL)+S is the productive means of marking adverbialization (e.g., ʔḥr š-, byn š-, kmw š-, kš-, lpy š-, mš-, ʕl š-, etc.).11 For instance, only three examples of the adverbializer ʕd ʔšr are evidenced in the Tosefta, whereas several hundred examples consist of ʕd š- ‘until’ throughout the corpus of Rabbinic Hebrew. In light of the origin of adverbialization and the distribution of the constructions, a plausible picture of the development may be sketched. The BH polymorphous adverbializer patterns, PREP+REL+S and PREP+S, were reduced to the single construction with the relative ʔšr by the early centuries BCE. This form appears to be a conservative feature of the Hebrew known from the Dead Sea Scrolls.12 Rabbinic Hebrew perpetuated the PREP+REL+S adverbialization pattern but exchanged the relative š- in place of ʔšr.13 3

Language Contact and Adverbializers in BH

In the previous sections, we explored the emergence of the two syntactic patterns of adverbializers in BH and examined the changes to these patterns in later Hebrew. In the following, we analyze the role of language contact of Aramaic in the evolution of Hebrew adverbialization. Doing so allows for a linguistic basis in approaching a particularly difficult issue in Exod 19:18, which will be examined in the final section. As Pat-El argues extensively, Aramaic lost “independent subordinators” early in its existence, and, more specifically, “never developed independent adverbial subordinators.”14 Instead, various dialects of Aramaic formed adverbial subordination through extensive use of the relative pronoun. According to Pat-El, this role for the relative particle increased over time in the language and any form of subordination required the use of the relative. 10  For the two primary derivational options for these relatives, see Heuhnergard 2006 and Holmstedt 2007. The replacement of ʔšr with š- may be seen as a broader change in Rabbinic Hebrew (Fernandez 1999: 50). 11  See further Segal 1927: 148. 12  Qimron 1986: 82. 13  Fernandez notes that while the origin of the relative/genitive particle ‫ ֶׁשל‬may be found in Hebrew, “the extent of its use has been influenced by the Aramaic relative ‫( ” ִּדי‬1999: 6). This connection should perhaps be expanded further to include the proclitic relative šɛ-. 14  Pat-El 2008: 59.

Hebrew Adverbialization, Aramaic Language Contact

41

With this mapping of the historical development of adverbial subordination in Aramaic established, Pat-El proposes a contact-induced change in BH causal subordination due to Aramaic influence.15 This change in BH diverges from the standard formation of this type of subordination that marks relationships between clauses. Both languages shared kî as a causal conjunction; however, Aramaic lost this conjunction early in its history.16 In order to compensate and formulate causal subordination, Aramaic developed a variety of expressions using the relative pronoun.17 Pat-El identifies in LBH two Aramaic loans on the basis of constructions with the relative pronoun. The first is Example (7) where Aramaic bdyl ‘because of’ corresponds to BH ba-ʾăšɛr l- ‘on account of’: (7) Aramaic: bdylh b-dy-l-h PREP+REL+PREP+him ‘because of him’18

BH: ‫ ַב ֲא ֶשר ְל‬ ba-ʾăšɛr lPREP+REL PREP ‘on account of . . .’ (Jon 1:8)

The second example relates Aramaic bdyl d ‘because’ to BH b-šɛ-l ʾăšɛr ‘though’: (8) Aramaic: bdyl dy hwʾ ydʿ b-dy-l dy hwʾ ydʿ PREP+REL+PREP REL he VERB ‘because he knows’19

15  Pat-El 2012. 16  We would like to thank A. M. Butts for drawing our attention to Syriac kay, which may be a remnant of this particle, though in Syriac it does not function as a clausal conjunction. 17  According to Pat-El (2012), while BH contains some instances of causal subordination through the adverbialization pattern PREP+REL+S, this usage of the construction is rare. For example, Gen 39:9 contains b-ʾšr used causally (‘He has not kept back anything from me, except you, because [b-ʾšr] you are his wife’), an adverbializer that can also be used spatially (Judg 17:8). 18  This particular construction in Aramaic tends to occur with pronominal suffixes. Pat-El (2012) cites Targum Onkelos Gen 3:17 (bdylk ‘because of you’) and 11QTgJob 38:3 (bdylh ‘because of him’). 19  4Q197 frg. 4 ii 4, as cited in Pat El 2012.

42

Boyd and Hardy

BH: ‫ ְב ֶשל ֲא ֶשר‬ b-šɛ-l ʾăšɛr PREP+REL+PREP REL ‘because (though) . . .’ (Eccl 8:17)

Pat-El’s study of Aramaic’s influence on causal subordination in BH serves as a basis for further theorization of issues of language contact and for the development of adverbializers in BH. Pat-El makes the case for language contact in the development of causal subordination in BH and places the issue of syntactic borrowing within a larger framework of contact linguistics discussed in the work of Thomason and Kaufman.20 Yet the very notion of syntactic borrowing has become a controversial area in the study of language contact. Linguists such as Winford have challenged the assertion that structure can be independently borrowed from lexicon, suggesting that syntactic borrowing is instead the product of extensive calquing.21 In this fashion, the syntax from the source language (SL) is borrowed into the recipient language (RL) as the agents of change are the speakers of the RL modelling expressions through loan translations. Additionally, Winford has argued many of the cases claimed to be syntactic borrowings are actually examples of SL agentivity;22 if so, then the label “borrowing,” which usually denotes RL agentivity, is problematic in these examples between Aramaic and BH. Pat-El notes that “syntactic interference happens in a situation when the source language is considered dominant over the receiving language for any number of reasons.”23 In this situation, Winford would demure the label “borrowing” and rather call such syntactic contact-induced change the result of shift or imposition. In defense of syntactic borrowing, Heine and Kuteva have examined situations in which syntax can be borrowed without language shift, showing that 20  This framework has been further developed in Thomason 2001. In more recent writings, Thomason has acknowledged the similarities between her typology developed with Kaufman and that derived independently in the work of Frans van Coetsem, a similarity that van Coetsem had also already acknowledged (Thomason 2003: 692; see also the brief comments in Thomason 2001: 95; van Coetsem 1990: 60–62). As Thomason and Kaufman note, they were not the first to notice the linguistic processes under examination, but rather they sought to refine the analysis of these linguistic outcomes through the accurate labeling and assessment of the linguistic processes underneath contact-induced phenomena (Thomason and Kaufman 1988: 42). 21  Winford 2003: 97; 2010: 175–77. 22  Winford 2010: 179–81. 23  Pat-El 2012: 251.

Hebrew Adverbialization, Aramaic Language Contact

43

structural linguistic elements can be borrowed.24 While Pat-El is, therefore, on safe ground in her analysis of syntactic borrowing, another topic in the study of language contact is relevant for her discussion, particularly as it relates Aramaic influence on Hebrew. That topic is the matter (MAT) and pattern (PAT) distinction. Sakel and Matras define MAT and PAT borrowing in the following manner: MAT loans refer to the process in which any morphological element and “its phonological shape” is transferred from the SL to the RL; PAT loans involve the borrowing of patterns and shapes of constructions from the SL to the RL without transferring the morphemes themselves.25 (“Morphemes” in this definition include lexemes.) MAT loans do not require a high level of bilingualism whereas PAT loans typically involve a higher degree of proficiency in both the SL and RL in the speech community. This feature of PAT loans is logical inasmuch as speakers of the RL have to be bilingual enough in the SL to be able to identify the pattern and then produce it sensibly in the RL but using the morphemes and linguistic features of the RL. The MAT/PAT distinction helps make sense of some of Pat-El’s observations. For example, in her discussion of Aramaic influence on causal subordination in Jon 1:8 and Qoh 8:17 she states that “all the elements of the preposition and conjunction shown above in Jonah and Qoheleth are original Hebrew features, but their new structural configuration as well as the causal function are the 24  For this counterargument to Winford, see Heine and Kuteva 2005: 157–59. They use an example from Malcom Ross’ study of contact between Takia and Waskia. It is certain that Proto-Western Asiatic, the language family of which Takia is a part, had a determiner (‘the’) that preceded the noun; however, Takia developed a postpositioned determiner, based on the model of Waskia. The process through which this change occurred in Takia involved patterning the Takia near deictic morpheme, which follows the noun when used attributively, on the SL (or, “model language” to use Heine and Kuteva’s terminology), namely Waskia. According to Heine and Kuteva, structural changes such as this example show that syntactic influence (in this case, through grammatical replication, though Takia also borrowed the SOV syntax from Waskia as an independent development) can occur without language shift attested as being a broader part of the contact situation. Since this situation does not involve shift, the example of the postpositioned determiner in Takia shows how structural borrowing can occur even without a (or, as a separate process from) change in word order: the Proto-Western Asiatic determiner was not simply placed behind the noun in Takia; rather, another lexeme that already fit into this postpositioned syntactic slot (the near deictic pronoun) was desemanticized through grammaticalization and modeled on the definite article in Waskia. This example shows, then, a clear case in which syntax is borrowed without language shift. For further arguments concerning the borrowing of structural linguistic elements, see Butts Forthcoming. 25  Matras and Sakel 2007: 15.

44

Boyd and Hardy

result of language contact.”26 This statement conforms to the definition of the PAT loans mentioned above. Finally, a sociolinguistic facet of the MAT/PAT distinction also undergirds and supports Pat-El’s thesis about language contact and causal subordination in Aramaic and Hebrew. In Matras and Sakel’s study, they catalogued over twenty languages and their various contact situations, noting the sociological backgrounds of each.27 In doing so, they found that the predominant sociolinguistic context for PAT loans involved situations in which the SL was not only a socially dominant language generally, but more specifically functioned as a lingua franca. This aspect of PAT loans, and the fact that causal subordination in a few examples in BH can be explained due to this contact-induced change with Aramaic, is consistent with a larger argument in Pat-El’s article. She claims that this feature in BH does not occur in SBH, but only in LBH. While her point was to display the fact that diachrony in BH is discernable, the relevance of her diachronic argument here is that the period when such PAT loans appear in BH aligns with the status of Aramaic as a lingua franca in the Persian period.28 In this section, we examined the role of contact-induced change from Aramaic causal subordination into BH. We affirmed the insights of previous studies that contact-induced change is present; however, we proposed another process was at play in the concept of a PAT loan. Understanding this contact scenario as a PAT loan has the advantage of explaining why the grammatical elements remained Hebrew even as Aramaic influence is evident as well as elucidating the sociolinguistic background of Aramaic as a lingua franca. 4

Causal Subordination, PAT Loans, and a Literary Problem in the Pentateuch

The history of the two adverbializing patterns (1) PREP+S and (2) PREP+REL+S in Hebrew and the contact-induced change in causal subordination from 26  Pat El 2012: 258. 27  Matras and Sakel 2007: 1–13 (especially 4–13). 28  While Aramaic was certainly a lingua franca in the Assyrian heartland in the eighth century BCE (and well attested in the seventh century BCE), it does not follow that it functioned similarly in the southern Levant in Judah at this time as well. For arguments concerning Assyrian, not Aramaic, influence in the speech of the Rav Shaqeh, see Oren 2013: 7–10. Moreover, the earliest epigraphic evidence of Aramaic being used administratively in Judah is the Aḥiab seal, dated to the late sixth century BCE (the early Persian/ Achaemenid era) (Vanderhooft 2011: 538).

Hebrew Adverbialization, Aramaic Language Contact

45

Aramaic is significant for more than purely linguistic reasons. Identifying patterns in BH that show evidence of internal development or Aramaic influence can also have benefits for the interpretation of the Pentateuch, particularly from a documentary perspective. In this section, a pattern that matches the second adverbializer construction and that may also show evidence of contact-induced change from Aramaic is presented. Its appearance in Exod 19:18 is analyzed, revealing that the adverbializer and the clause it governs cannot be easily aligned to any of the traditional sources. Finally, we will offer a suggestion concerning the nature of the causally subordinated clause in Exod 19:18. The second class of adverbializers, PREP+REL+S, as has been shown above, has a long history in Hebrew. As also discussed above, Aramaic contact-induced change in causal subordination in BH manifests itself in LBH. This causal subordination in Aramaic can have a similar pattern as BH PREP+REL+S. In Aramaic, the relative particle /d/, according to Wertheimer, nominalizes the following clause, in effect translating the clause into a nominal constituent.29 The addition of the preposition in front of the relative particle then adverbializes the relative phrase, which can denote any number of adverbial expressions. In Syriac, two constructions in this pattern are used to form adverbial expressions: ʿl(PREP)+d(REL)+S; and mṭl(PREP)+d(REL)+S.30 This particular pattern, being common in Aramaic and BH, cannot be ascribed to language contact. Moreover, the construction ʿl(PREP)+ REL +S appears in both BH (in a wide variety of literary strata) and Aramaic, and also cannot be attributed to contact-induced change. The shift toward standardization in later Hebrew may well have been influenced by the Aramaic pattern and certainly the eventual shift to the mono-syllabic šɛ- evinces a likely PAT borrowing situation. The occurrence of the adverbial marker mpny ʾšr (marking reason) in Exod 19:18 and Jer 44:23 is peculiar in BH. This fact alone is not sufficient to propose language contact. Given the literary problems in both verses, however, in which the adverbial clauses are likely later additions to each text, the uniqueness of the construction mpny ʾšr in BH merits further examination. The exceptional nature of Jer 44:23 has been discussed in detail. The content of the verse governed by mpny ʾšr is an explanation of Jer 44:22, and Jer 44:23 is peculiar inasmuch as it is the only instance in BH in which the causal subordinator mpny begins a sentence in which the verb is a suffix-conjugation (SC).31 This 29  Wertheimer 2001: 289. 30  Wertheimer 2001: 284. 31  Goldstein states (2013: 100): ‘‫בלׁשון המקרא איננו מוצאים בדרך כלל ׁשימוׁש במילית ‘מפני‬ ‫ בלׁשון המקרא נפתחת פסוקית סיבה על ידי “יען״ או‬,‫ יתר על כן‬.‫בפתיחת מׁשפט מׁשועבר‬

46

Boyd and Hardy

construction is unique in BH syntax, and could be explained as a later gloss on the meaning of Jer 44:22. In Aramaic, the adverbializer b-dyl d- expresses a causal meaning when governing a SC and a resultative meaning when governing a PC. As Pat-El observes, the constructions b-ʾšr l- in Jon 1:8 and b-šl ʾšr in Qoh 8:17 are loans from the Aramaic causal subordinator b-dyl d-; however, in these verses in the Hebrew Bible the causal subordinators govern prefixconjugation (PC) verbs, the reverse of the usage in Aramaic.32 Pat-El correctly notes that the RL does not have to have the “same distribution” of borrowed features as the SL. If, however, the adverbializer mpny ʾšr is patterned from a similar construction in Aramaic (see below), then the oddity that Jer 44:23 is the only sentence in the Hebrew Bible that begins with a causal conjunction and governs a clause with the SC could find an explanation as a contactinduced change. Even more peculiar syntactically, then, is Exod 19:18, in which the adverbial construction mpny ʾšr functions within the sentence as an explanation for the reason why Sinai was wrapped in smoke. The adverbializer mpny ʾšr marks an explanation of the first part of the verse, which has been identified as the J source; however, in J the deity does not descend on the mountain until Exod 19:20, which contradicts the content of the clause governed by mpny ʾšr.33 In the E source, however, the deity states that he will imminently descend in a cloud (Exod 19:9a), as per Exod 19:16–17. As such, the content of the mpny ʾšr clause matches the narrative of E, and explains why, according to Exod 19:17, the people stand at the foot of the mountain (since the deity had descended on top of the mountain).34 Such an explanation in Exod 19:18, however, is hardly ,‫ כפותח מׁשפט סיבה מופיע רק במקום אחר נוסף במקרא‬,“‫ והצירוף “מפני אׁשר‬,“‫“יען כי‬ .‫ יח‬,‫במ‘ יט‬ 32  Pat-El 2012: 258. 33  For a source discussion of J and E in Exod 19, see Baden 2012: 77 (J), 117–18 (E). See also page 128, in which he observes the fact that Deut 4:11 reflects only the narrative of Exod 19:16–17, the only E material in the otherwise exclusively J block of Exod 19:10–18, is consistent with how D uses E and J separately. Baden here builds on the work of August Dillmann, who states, “Aus dem Dt. kann ein Beweis für die Existenz eines BC vor dem ABC auch nicht gezogen werden. Denn zwar ist wohl richtig, dass D in seiner Geschichtsdarstellung sich an B u. C anschliesst; aber schon S. 609f. ist gezeigt, dass D den B u. C noch als besondere Schriften vor sich hatte, während ein sicherer Beweis dafür, dass er BC als Ganzes benützte, nicht geführt werden kann” (Dillmann 1886, 679; for Dillmann’s exegesis leading to this conclusion, see 609–11). For more of a discussion of Dillmann in recent theories of the development of the Pentateuch, see Baden 2009: 37–40. For a more extended analysis of D’s use of E and J, see Baden 2011: 327–44. 34  The trembling of the people within the camp in Exod 19:16b is connected with the trembling of the mountain in Exod 19:18b. An argument could be made that Exod

Hebrew Adverbialization, Aramaic Language Contact

47

necessary in E as the narrative connection between Exod 19:9a and Exod 19:16 already makes it clear that the deity has come down on top of Sinai.35 From a source critical perspective, the clause governed by mpny ʾšr in Exod 19:18 belongs neither to the J nor to the E source and can be plausibly identified as a later harmonistic insertion. Moreover, the marking of the phrase in Exod 19:18 by mpny ʾšr is unparalleled as a causal subordinator in BH, excepting Jer 44:23, prompting further exploration as a basis for explanation of the adverbial marking and the literary nature of the verses in their context. Two explanations can be offered for the origin of this construction in BH. First, given the similarities between mpny ʾšr and later mpny š in Tannaitic Hebrew, one could argue that it is a back-formation from this idiom into BH. The scribe who made the insertion could have recognized the dialectal distinction and translated the later Hebrew form into the biblical form with ʾšr for the insertion. The lack of other clear examples of this process, however, may argue against this solution. Second, the presence of mpny ʾšr in Exod 19:18 and Jer 44:23 could be explained as a result of a PAT loan from Aramaic. Wertheimer has explored a linguistic process in Syriac in which a relative pronoun substantivizes the linguistic elements it governs (a process called “translation,” borrowed from Tesnière).36 A further (or, second order) translation of PREP+REL converts the construction into an adverbial clause in a complex sentence, forming a subordinating conjunction marking reason and other adverbial relationships. For 19:18b should be changed to ‘all the people trembled very much’, following the reading in the Septuagint. A similar confusion between people and the mountain occurs in the Samaritan Pentateuch in Exod 19:12a. In this understanding of Exod 19:18b, the people are increasingly in awe and tremble as the theophany happens before their eyes, which explains their response in Exod 20:18–20. On the other hand, reading with the MT allows for another literary explanation. The people tremble in the camp in Exod 19:16b, and after Moses brings them out of the camp for the encounter with the deity he then stations them at the foot of the mountain. To communicate the sheer immensity of the divine encounter, the narrator claims that the mountain, the location of the theophany, had an even more severe response than the people, trembling very much. 35  An argument might be made that a preposition governing a SC can locate the event in the non-simple past, envisioning a completed event in the future leading to the harmonistic translation in Exod 19:18 ‘because God will descend. . .’. In this understanding, Exod 19:18 could belong to J. For example, Exod 10:3 begins with the temporal marker ʿd mty, which sets the SC mʾnt as a completed future event, ‘How long will you refuse . . .?’ (Waltke and O’Connor 1990: §30.4.c). Against this understanding of Exod 19:18, this use of SC is marked with temporal adverbs, not causal. The SC in Exod 19:18 is therefore better translated ‘because God had descended’. 36  Wertheimer 2001: 261–63.

48

Boyd and Hardy

example, the NP ‘the words of the king’ (or, ‘the king’s words’) can be converted into an adverbial phrase with the addition of mṭl as in ‫ ܡܛܠ ̈ܡܠܝ ܡܠܟܐ‬mṭl mly mlkʾ ‘because of the words of the king’. The PREP+REL forms a subordinate conjunction, “marking the beginning of a clause in a complex sentence”: ‫ ܡܛܠ ܕܡܠܟܐ ܡܠܠ‬mṭl d-mlkʾ mll ‘because the king spoke’.37 In this case, the PREP+REL (mṭl d-) functions as an adverbial conjunction denoting reason. The relative in this pattern PREP+REL is a translatif, or a morphological marking of the translation or transfer of a word from one grammatical category to another, according to Tesnière’s framework. In complex sentences, a double translation occurs in which the REL marks a first stage in which a verbal phrase is translated to a nominal element in the sentence and the PREP turns the new nominal element into an adverb. Wertheimer’s overall argument is to show that the REL in Aramaic in all its forms is a nominalizing particle. Regarding BH, the REL ʾšr nominalizes the following verbal clause, and the use of mpny turns the entire phrase into an adverbial element modifying the first part of Exod 19:18. One could argue that mpny ʾšr in Exod 19:18 is patterning after this construction in Aramaic given the literary and linguistic issues that mark the clause governed by mpny ʾšr as a later, post-redactional addition and given the time period (Persian era when Aramaic was a lingua franca) that such an addition would occur. Against this explanation is the fact that Wertheimer’s study was based on Syriac, at a much later period than the biblical text, and the types of scribal additions proposed for Exod 19:18. Moreover, an internal explanation within Hebrew may lie in the possibility that a similar process may be behind the development of the preposition ʿl ‘upon’, which can be causal on its own,38 but also governs the REL ʿl ʾšr meaning ‘because’.39 In a similar manner, Hebrew mpny can have a causal meaning on its own,40 and mpny ʾšr could be understood as a more explicit marking of an adverbial translation, converting the 37  Wertheimer 2001: 284. 38  Ps 119:136 39  Num 20:24 40  It should be noted that to the best of our knowledge the causal use of mpny by itself never governs a verb. A further subordinator is added when a verbal element is needed, as in Jer 51:64: ‫ מפני הרעה אׁשר אנכי מביא עליה‬mpny hrʿh ʾšr ʾnky mbyʾ ʿlyh ‘because of the evil which I am bringing upon her’, where ʾšr functions as a further subordinator that governs a verbal element. We are aware of the following causal uses of mpny by itself: Gen 6:13; 41:31; Exod 3:7; 8:20; Deut 28:20; Judg 6:6; 1Kgs 8:11; Job 37:19; Is 7:2; Jer 9:6; 23:10; 51:64; Ezek 14:15; Lam 5:10; Amos 2:9; Hos 10:15; Mic 1:4; Mal 3:14. None of these cases consists of mpny governing a verb. In this manner, the use of ʾšr in mpny ʾšr may allow for such verbal governance.

Hebrew Adverbialization, Aramaic Language Contact

49

following verbal phrase into a substantive element by use of the REL and then transforming the whole construction into an adverb by use of the PREP.41 In this section, a peculiar adverbial construction in Exod 19:18 and Jer 44:23 was examined. The clauses governed by these adverbializers are literarily problematic and in the case of Exod 19:18 the clause has been suggested to be a later scribal harmonization between two sources. Given the time period for such scribal additions (Persian era or later when Aramaic pattern loans occur elsewhere in BH), Aramaic influence is one possible explanation for this unique construction. Since an internal suggestion can tentatively also be made, however, the suggestion that mpny ʾšr is a pattern loan is an admittedly hypothetical, albeit intriguing, possibility to a problematic construction that deserves further thought. Bibliography Baden, Joel. 2009. J, E, and the Redaction of the Pentateuch (Forschungen zum Alten Testament 68). Tübingen. ———. 2011. “Deuteronomic Evidence for the Documentary Theory,” in Thomas B. Dozeman, Konrad Schmid, and Baruch J. Schwartz (eds.), The Pentateuch: International Perspectives on Current Research (Forschungen zum Alten Testament 78). Tübingen. 327–44. ———. 2012. The Composition of the Pentateuch: Renewing the Documentary Hypothesis (Anchor Bible Reference Library). New Haven. Blau, Joshua (ed. by Michael Patrick O’Connor and Cynthia L. Miller). 2010. Phonology and Morphology of Biblical Hebrew. An Introduction (LSAWS 2). Winona Lake. Butts, A. M. Forthcoming. Language Change in the Wake of Empire: Syriac in its GrecoRoman Context (LSAWS 11). Winona Lake. Coetsem, Frans van. 1990. “Review: Language Contact, Creolization, and Genetic Linguistics by Sarah Thomason; Terrence Kaufman; Lectures on Language Contact

41  The preposition mn plus ʾšr is typically non-causal in BH. The following are instances in which a non-causal use of mn plus ʾšr is evident: Gen 31:1; Exod 5:11; 29:27 (x2); Lev 14:30; Josh 10:11; Judg 16:30; Ruth 2:9; 2Kgs 6:16; Esth 4:11; Eccl 3:22; Is 47:13; Jer 40:7. The causal use occurs twice for this construction: Num 6:11; Is 43:4. The second is clearly exilic or postexilic, and the Septuagint did not render the construction in Num 6:11 as causal, but rather ‘concerning the things (in which) he sinned’. The Aramaic construction mn dy is clearly causal in Dan 3:22. The form mzy in Sefire 1:25 is adverbial, representing mn and zy, and is likely temporal; a causal translation (‘because Ashur reigns!’) is theoretically possible, but does not make sense in the context of the curse formulae.

50

Boyd and Hardy

by Ilse Lehiste; Languages in Competition, Dominance, Diversity, and Decline by Ronald Wardhaugh,” Language in Society 19: 260–62. Dillmann, August. 1886. Die Bücher Numeri, Deuteronomium, und Joshua (Kurzgefasstes exegetisches Handbuch zum Alten Testament 13). Leipzig. Fernández, Miguel P (trans. John Elwolde). 1999. An Introductory Grammar of Rabbinic Hebrew. Leiden. Gensler, Orin D. 2000. “Why Semitic Adverbializers (Akkadian -iš, Syriac -āʾīṯ) should not be Derived from Existential *ʾīṯ,” JSS 45: 233–265. Gil, David. 2013. “Genitives, Adjectives and Relative Clauses,” in Matthew S. Dryer and Martin Haspelmath (eds.), The World Atlas of Language Structures Online. Max Planck Institute for Evolutionary Anthropology. (Available online at http://wals .info/chapter/60, accessed on 2014–08–26.) Goldenberg, Gideon. 1998. “Attribution in Semitic Languages,” in his Studies in Semitic Linguistics. Jerusalem. 1–20. Goldstein, Ronnie. 2013. The Life of Jeremiah: Traditions about the Prophet and their Evolution in Biblical Times. Jerusalem. (Hebrew) Hardy, Humphrey. 2014. Diachronic Change in Biblical Hebrew Prepositions: A Case Study in Grammaticalization. Ph.D. Diss., The University of Chicago. Heine, Bernd and Tania Kuteva. 2005. Language Contact and Grammatical Change (Cambridge Approaches to Language Contact). Cambridge. Holmstedt, Robert. 2001. “Headlessness and Extraposition: Another Look at the Syntax of ‫אׁשר‬,” Journal of Northwest Semitic Languages 27: 1–16. ———. 2006. “The Story of Ancient Hebrew ʾăšer,” Ancient Near Eastern Studies 43: 7–26. ———. 2007. “The Etymologies of Hebrew ʾăšer and šeC-,” JNES 66: 177–191. Huehnergard, John. 2006. “On the Etymology of the Hebrew Relative šɛ-,” in S. E. Fassberg and A. Hurvitz (eds.), Biblical Hebrew in Its Norwest Semitic Setting. Typological and Historical Perspectives. Jerusalem – Winona Lake. 103–125. Matras, Yaron, and Jeanette Sakel. 2007. Grammatical Borrowing in Cross-linguistic Perspective (Empirical Approaches to Language Typology 38). New York. Oren, Mikhal. 2013. “Interference in Ancient Languages as Evidenced by Governed Prepositions,” JSS 58: 1–11. Pat-El, Na’ama. 2008. “Historical Syntax of Aramaic: A Note on Subordination,” in H. Gzella and M. L. Folmer (eds.), Aramaic in Its Historical and Linguistic Setting. Wiesbaden. 55–76. ———. 2012. “Syntactic Aramaisms as a Tool for the Internal Chronology of Biblical Hebrew,” in Cynthia L. Miller-Naudé and Ziony Zevit (eds.), Diachrony in Biblical Hebrew (LSAWS 8). Winona Lake. 245–63. Qimron, Elisha. 1986. The Hebrew of the Dead Sea Scrolls (HSS 29). Winona Lake.

Hebrew Adverbialization, Aramaic Language Contact

51

Segal, M. H. 1927. A Grammar of Mishnaic Hebrew. Oxford. Thomason, Sarah. 2001. Language Contact: An Introduction. Washington, D.C. ———. 2003. “Contact as a Source for Language Change,” in Brian D. Joseph and Richard D. Janda (eds.), The Handbook of Historical Linguistics. Malden, MA. 687–712. Thomason, Sarah and Terence Kaufman. 1988. Language Contact, Creolization, and Genetic Linguistics. Berkeley. Thompson, Sandra and Robert E. Longacre. 1985. “Adverbial Clauses,” in Timothy Shopen (ed.), Language Typology and Syntactic Description. Cambridge. 171–234. Vanderhooft, David. 2011. “ʾel-mĕdînâ ûmĕdînâ kiktābāh: Scribes and Scripts in Yehud and in Achaemenid Transeuphratene,” in Oded Lipschits, Gary N. Knoppers, and Manfred Oeming (eds.), Judah and the Judeans in the Achaemenid Period: Negotiating Identity in an International Context. Winona Lake. 529–44. Waltke, B. K. and M. O’Connor. 1990. An Introduction to Biblical Hebrew Syntax. Winona Lake. Wertheimer, Ada. 2001. “The Functions of the Syriac Particle d-,” Le Muséon 114: 259–89. Winford, Donald. 2003. An Introduction to Contact Linguistics (Language in Society 33). Malden, MA. ———. 2010. “Contact and Borrowing,” in Raymond Hickey (ed.), The Handbook of Language Contact (Blackwell Handbooks in Linguistics). Malden, MA. 170–187.

The Distribution of Declined Participles in Aramaic-Hebrew and Hebrew-Aramaic Translations Yochanan Breuer 1 Introduction Hebrew and Aramaic are cognate languages, which became even closer due to a long period of contact. However, certain features, which evolved in Eastern Aramaic, made it quite distinct from other Aramaic dialects as well as from Hebrew. In this article, I will describe how a feature of this kind is treated in Aramaic-Hebrew and Hebrew-Aramaic translations. The description is based on three sources: (1) Hilḵot Reʾu (HR), a Hebrew translation of Halaḵot Pesuqot (HP). The original was composed in Babylonian Aramaic during the Geonic period. The time and place of the translation is unknown, but it seems that it was composed in Palestine during the Geonic period. Its style resembles that of the Mishna according to its Palestinian tradition.1 (2) Targum Onqelos. (3) The Babylonian tradition of Mishnaic Hebrew, i.e., Mishnaic Hebrew as transmitted in the Babylonian Talmud. In Hebrew and in most Aramaic dialects, while the perfect and imperfect decline according to person, the participle does not have a person marker. In Eastern Aramaic, on the other hand, also the participle declines according to person. The following is the declension of the singular participle in Babylonian Aramaic with the equivalent Hebrew forms (the feminine and plural decline in a similar way): Aramaic Hebrew 1st ‫ ָּכ ֵת ְבנָ א‬katevna ‫ אני כותב‬ʾani kotev I write 2nd ‫ ָּכ ְת ַב ְּת‬katvat ‫ אתה כותב‬ʾatta kotev you write 3rd ‫ ָּכ ֵתב‬ katev ‫ הוא כותב‬hu kotev he writes * A Hebrew version of this article will appear in the Ilan Eldar Festschrift. 1  I will not elaborate here on this point, since I intend to devote a comprehensive description to this translation. The numbers following the quotations from HP refer to Sefer Halachot Pesuqot, edited by S. Sasoon (Jerusalem, 1950); in HR, to Halaḵot Pesuqot ʾo Hilḵot Reʾu, edited by A. L. Schlossberg (Versailles, 1886). The quotations from HP, HR and all the rabbinic sources follow the readings in Maagarim, the Archive of the Historical Dictionary Project of the Academy of the Hebrew Language, presented at . The conclusions are based on 66% of the parallel material: HP pp. 1–126 and HR pp. 1–97. © koninklijke brill nv, leiden, ���5 | doi ��.��63/9789004300156_005

aramaic-hebrew and hebrew-aramaic translations

53

In the 1st and 2nd persons, only declined forms are used, which were created by the suffixation of the independent pronoun to the participle: ‫ ֲאנָ א‬+ ‫ ָּכ ֵת ְבנָ א > ָּכ ֵתב‬katev + ʾana > katevna, ‫ ַא ְּת‬+ ‫ ָּכ ְת ַב ְּת > ָּכ ֵתב‬katev + ʾat > katvat. When the participle is used, an independent pronoun does not appear, since the subject is already expressed by the participle form.2 This is clear in the 1st and 2nd persons, but it is also true for the 3rd person: the base form with zero morpheme, by contrast with the other declined forms, clearly expresses the 3rd person, hence it also does not require an independent pronoun. In Hebrew, on the other hand, the base form of the participle is not limited to the 3rd person and requires the appearance of an independent pronoun to mark the person (1st, 2nd, or 3rd, SG or PL). Consequently, translation of the Aramaic participle into Hebrew requires the addition of an independent pronoun, as in the table above. Indeed, this is the case in many examples: HP HR 1st ‫ חזינא ליה לרב כהנא‬ ‫אני רואה רב כהנא‬ ḥazena leh le-Rav Kahana (29) ʾani roʾe Rav Kahana (26) ‘I see R. Kahana.’ 2nd ‫ מאי בעית בהאיי ביתא‬ ‫מה אתה מבקש בבית זה‬ may baʿet be-haye beta (88) ma ʾatta mevaqqeš be-vayit ze (68) ‘What do you want in this house?’ 3rd ‫ כשמע להי אסדר ברכות שמע להי‬ ‫כשהוא שומען על סדר ברכות הוא שומען‬ ki-šamaʿ le-hi ʾa-sseder beraḵot ke-she-hu šomʿan ʿal seder beraḵot šamaʿ lehi (20) hu šomʿan (19) ‘When he hears them, it is according to the order of the blessings that he hears them.’ However, in most cases, the participle is translated in a different way, and a distinction should be made between the 1st and 2nd persons and the 3rd person.

2  It does appear for the purpose of focalization, e.g., ‫אנא אמינא לך רבן יוחנן בן זכאי ואת אמרת‬ ‫ לי רב ושמואל‬ʾana ʾamina laḵ Rabban Yoḥanan Ben Zakkay we-ʾat ʾamrat li Rav u- Šmuʾel ‘I say to you Rabban Yoḥanan Ben Zakkay and you say to me Rav and Šemuʾel’ (B. Talmud Beṣa 5a), exactly as in the Hebrew declined forms, e.g., ‫ אני אמרתי להשפילן ואתה אמרת לגאותן‬ʾani ʾamarti le-hašpilan we-ʾatta ʾamarta le-ga‌ʾotan ‘I wanted to humiliate them and you wanted to glorify them’ (P. Talmud Taʿanit 67:1).

54 2

Breuer

1st-2nd Persons: Change of Tense

2.1 Hilḵot Reʾu In the 1st and 2nd persons, the tense is usually changed and the participle is translated by a perfect or an imperfect,3 e.g.: 1st person singular: perfect



‫ כל לא ידענא היכא אתנחתינהו‬ ‫פשיעה היא‬

‫כל לא ידעתי היכן הנחתי פשיעה היא‬

kol la yadaʿna heḵa kol lo yadaʿti heḵan hinnaḥti ʾatnaḥtinhu pešiʿa hi (82) pešiʿa hi (64) ‘Whenever (he says) I do not know where I put it, it is a violation.’

imperfect ‫אבקש [חוטרא] לידי ומר[א] לקבורה בעיאנא חוטרא ליד ומרא לקבורה‬ baʿyana ḥuṭra le-yad u-mara ʾavaqqeš ḥuṭra le-yadi u-mara li-qvura (113) li-qvura (87) ‘I want a can for my hand and a mattock for burial.’ 2nd person singular: perfect ‫ אי כתיב יקנו לבסוף כי דקא אמרת‬ ‫אם כתב יקנו לבסוף כשאמרתה‬ ʾi ketiv yiqnu li-vsof ki de-qa ʾim katav yiqnu le-va-ssof ke-še ʾamrat (64) ʾamarta (51) ‘If yiqnu was written at the end, it would have been as you say.’ imperfect ‫ ושבקת רבנן ועבדת כר' אליעזר‬ ‫ותניח חכמ' ותעש' כר' אליעזר‬ we-šavqat rabbanan we-ʿavdat we-tanniaḥ ḥaḵamim we-taʿase ke-Rabbi ʾEliʿezer (27) ke-Rabbi ʾEliʿezer (25) ‘Will you leave the sages and do as R. Eliezer?’ 1st person plural: Perfect ‫ ושמעינן מינה דיש לבכור בשבח‬ we-šamʿinan minnah de-yesh li-vḵor bi-švaḥ (59)

‫ושמענו ממנה שיש לבכור בשבח‬

we-šamaʿnu mimmenna še-yyesh li-vḵor bi-švaḥ (47)

3  The sum of the participle forms found in the corpus is 2,293, out of which 1,018 were deducted according to the considerations enumerated in Appendix 1. The conclusions are thus based on an analysis of the translation of 1,275 participle forms.

aramaic-hebrew and hebrew-aramaic translations

55

‘And we conclude from it that the first-born takes part of the profit.’ imperfect ‫ אי איכא פנים חדשות מברכינן כלהי‬ ‫אם יש פנים חדשות נברך את כולן‬ ʾi ʾikka panim ḥadašot ʾim yeš panim ḥadašot nevareḵ mevarḵinan kullehi (109–110) ʾet kullan (84) ‘If there are new participants, we say all the blessings.’ 2nd person plural: imperfect ‫ כי כתביתו אקניאתא‬ ‫כשתכתבו הקנאה‬ ki katvitu ʾaqnayata (69) ke-še-tiḵtevu haqna‌ʾa (55) ‘When you write deeds of transfer.’ Expressions created by this process deviate from Mishnaic Hebrew, where the participle appears as in Aramaic. See, e.g., – ‫ לא ידעתי‬lo yadaʿti ‘I do not know’: ‫ כל לא ידענא היכא אתנחתינהו פשיעה היא‬

‫כל לא ידעתי היכן הנחתי פשיעה היא‬

kol la yadaʿna heḵa ʾatnaḥtinhu kol lo yadaʿti heḵan hinnaḥti pešiʿa hi (82) pešiʿa hi (64) ‘Whenever (he says) I do not know where I put it, it is a violation.’

In Mishnaic Hebrew, the expression ‫ לא ידעתי‬lo yadaʿti does not appear, while the participial expression ‫ איני יודע‬ʾeni yodeaʿ appears more than 70 times. –

‫אמר‬ ַ ֹ ‫ מניין נ‬minnayin nomar ‘From where do I deduce it?’: ‫ מנא אמינא לה דתניא רבי אומר והיתה‬ ‫ לאהרן ולבניו‬

‫מניין נומר דתני ר' אומ' והיה לאהרן‬ ‫ולבניו‬

mena ʾamina lah de-tanya Rabbi minnayin nomar di-tne Rabbi ʾomer we-hayta le-ʾAharon ʾomer we-haya le-ʾAharon u-l-vanaw (118) ʾu-l-vanaw (90) ‘From where do I deduce it? (From) that is said, Rabbi says, and it will be for Aharon and for his sons.’ The expression ‫ מניין נאמר‬minnayin nomar does not appear in Mishnaic Hebrew. Instead, we find 14 times a similar expression in the participle ‫ מניין אתה אומר‬minnayin atta ʾomer ‘from where do you deduce it’, e.g., ‫ומניין‬

56

Breuer

‫ אתה אומר שלא ירדו אבותינו למצרים אלא כדי שיעשה ניסין וגבורה‬u-minnayin ʾatta ʾomer še-llo yardu ʾavotenu le-Miṣrayim ʾella kede še-yyaʿase nissin u-gvura ‘From where do you deduce that the reason why our forefathers went down to Egypt was only so that He will do miracles and might?’4

–

‫ אבקש‬ʾavaqqeš ‘I want’: ‫ מישום דקא בעינא אינסובי לרב זביד‬

‫בשביל שאבקש להנשא לרב זביד‬

mi-ššum de-qa baʿena ʾinnesove bišvil še-ʾavaqqeš le-hinnase le-Rav le-Rav Zevid (118) Zevid (91) Since I want to get married to R. Zevid.

‘I want’ is expressed in Mishnaic Hebrew by the participle, e.g., ‫ אני מבקש לילך לאלכסנדריא‬ʾani mevaqqeš leleḵ le-ʾAleksandria ‘I want to go to Alexandria’.5 –

‫ אם תתיר‬ʾim tattir ‘if you permit’:

‫ דאי שרית ליה למיכל בחמש‬

‫שאם תתיר לו לאכל בחמש‬

de-ʾi šaret leh le-meḵal be-ḥameš (8) še-ʾim tattir lo loḵal be-ḥameš (12) ‘Since if you permit him to eat at five.’ In Mishnaic Hebrew, the participle is used: ‫אם אתה מתיר לו מקצת התר לו את הכל‬ ʾim ʾatta mattir lo miqṣat hatter lo ʾet hakkol ‘If you permit him part, permit him all’ (Tosefta Sheviʿit 1:5). –

General obligation:

When a general obligation is expressed in HP by the 1st person participle plural, it is frequently changed in HR into a plural participle without a subject, e.g.,

‫ האיי אפטורפא דמפסיד מסלקינן ליה‬

‫איפטרופוס שמפסיד מסלקין אותו‬

haye ʾafṭurfa de-mafsid mesalleqinan ʾepiṭropos še-mmafsid mesalleqin leh (56) ʾoto (45) ‘A custodian who causes losses we dismiss him.’

However, when the 1st person plural is maintained, it is changed into the imperfect: 4  Siphre ad Deuteronomium, edited by L. Finkelstein (Berlin, 1939), 342. 5  Mechilta d’Rabbi Ismael, edited by H. S. Horovitz and I. A. Rabin (2nd ed.; Jerusalem, 1970), 137.

aramaic-hebrew and hebrew-aramaic translations

‫ תפסינן אילא תפסינן‬

tafsinan ʾila tafsinan (90) ‘Do we take it or not?’

57

‫נתפוש או לא נתפוש‬

nitpos ʾo lo nitpos (69)

‫ לאכראגא למזוני ולקבורה מזבנינן‬ ‫למס למזונות ולקבורה נמכור‬ ‫ בלא אכרזתא‬ ‫בלא הכרזה‬

le-ʾaḵraga li-mzone we-li-qvura le-mas li-mzonot we-liqvura mezabbeninan be-la ʾaḵrazta (56) nimkor be-lo haḵraza (44) ‘For tax, for food and for burial we sell without a proclamation.’



‫אי גברא דאמיד למישתא‬ ‫ בכסא דכספא הוא עליה אתו ואילא‬ ‫ אכסא דכספא אתו‬

‫נראה אם אדם שהוא רגיל לשתות‬ ‫בכוס שלכסף עליו באו ליסטים ואם‬ ‫לאו על כוס באו‬

ḥazenan ʾi gavra da-ʾamid le-mište nirʾe ʾim ʾadam še-hu ragil lištot be-ḵasa de-ḵaspa hu ʿaleh ato we-ʾi-la be-ḵos še-lle-ḵesef ʿalaw ba‌ʾu lisṭim ʾa-kkasa de-ḵaspa ʾato (85) we-ʾim law ʿal kos ba‌ʾu (66) ‘We see, if he is a person who is used to drink with a silver cup, they came for him, and if not, they came for the silver cup.’

In Mishnaic Hebrew, a general obligation is expressed by the participle; compare, e.g., ‫ רואין אם במיפה כחו כותבין ואם לאו אין כותבין‬roʾin ʾim bi-myappe koḥo kotvin we-ʾim law ʾen kotvin ‘they see, if it is to give him a prerogative, they write, and if not, they do not write’ (B. Talmud Bava Batra 135b). The tendency to change the tense requires an explanation. At first glance, it would seem to reflect a difference in the tense system: if the scope of the participle in Babylonian Aramaic is wider than in Hebrew, it could be an adaptation to the Hebrew system.6 Alternatively, it could be the influence of Biblical Hebrew or Arabic, where the participle is not part of the tense system, and the perfect or imperfect are used instead. The possible influence of these languages cannot be excluded in the Geonic period. Although these factors may have also played some role, they cannot account for the wide scope of this process: the difference between Aramaic and Hebrew is not so great as to cause a change in most of the forms,7 as can be proven by the comparison with 6  For the use of the participle to denote the past in Babylonian Aramaic, see, e.g., S. Morag, “Ha-ʾaramit Ha-bbavlit be-masoret Teman: Ha-ppoʿal Ha-ššalem” in S. Lieberman, S. Abramson, E. Y. Kutscher and S. Esh (eds.), Henoch Yalon Jubilee Volume (Jerusalem, 1963), 193. For the use of the participle in Babylonian Aramaic in general, see M. Schlesinger, Satzlehre der Aramäischen Sprache des Babylonischen Talmuds (Leipzig, 1928), 40–42. 7  It is true that in Babylonian Aramaic the participle may denote the past and the future (see previous note), but this is true for Mishnaic Hebrew as well. For the use of the participle in Mishnaic Hebrew, see M. Mishor, The Tense System in Tannaitic Hebrew (Ph.D. Diss.,

58

Breuer

Mishnaic Hebrew, where similar expressions are phrased in the participle, as in Aramaic. As for Biblical Hebrew and Arabic, I could not trace an influence of these languages in other domains. Usually, the language is quite similar to Mishnaic Hebrew. It seems to me that the reason for this phenomenon stems from the adherence of this translation to the style of the original. There are no elaborations, and even technical Talmudic terms are translated literally, word for word, with no expansions or explanations. In translating the Aramaic participle, using the Hebrew participle would require splitting the word into two by adding an independent pronoun. In order to keep the original rhythm and to introduce a word for word translation (in other words: in order to translate a declined Aramaic form by a declined Hebrew one), the Hebrew declined forms are used, which are the perfect or the imperfect. This explanation is supported by the difference in translation between the 1st and 2nd persons and the 3rd person:

1st–2nd 3rd

Participle

Perfect and imperfect

52 (20%) 770 (76%)

205 (80%) 248 (24%)

In the 1st and 2nd persons, tense is changed in 80% of the forms,8 while in the 3rd person, only in 24%.9 If the change resulted from a difference in the tense The Hebrew University of Jerusalem, 1983), 162–313 (in Hebrew). For its use to denote the future, see ibid., 253–255, and to denote the past see ibid., 255–258. It is thus difficult to decide whether there is a difference between Babylonian Aramaic and Mishnaic Hebrew without a comprehensive investigation. I assume that the difference is not so great as to force the translator, who naturally tends to imitate the original, to change the tense in the overwhelming majority of forms. 8 In most forms, tense can be determined according to the orthography. In two forms the spelling is ambiguous: 2nd person ‫ כתבת‬ketavt/katvat and 3rd person ‫ כתב‬ketav/katev may be either a perfect or a participle. In these forms, the tense must be decided according to context, and there may be a difference between my analysis and that of the translator. Therefore, I also calculated statistically without these forms; this yields even a more clear distinction: in the 1st and 2nd persons, 29 forms are translated by the participle (14%) as against 181 by perfect or imperfect, while in the 3rd person, 259 forms are translated by the participle (78%) as against 71 by the perfect or the imperfect. 9 The following example may demonstrate the difference: in the root .‫י‬.‫ע‬.‫ ב‬b.ʿ.y., in the 1st person, a translation with the imperfect ‫ אבקש‬ʾavaqqeš or ‫ ארצה‬ʾerṣe appears 10 times, and

aramaic-hebrew and hebrew-aramaic translations

59

system, there would have been no reason for a difference between the persons. On the other hand, a desire to introduce declined forms for declined forms naturally leads to such a difference: the Aramaic declined forms of the 1st and 2nd persons are translated by Hebrew declined forms (perfect or imperfect), while the Aramaic base form of the 3rd person is translated by the Hebrew base form.10 2.2 Targum Onkelos The Aramaic of Targum Onqelos is basically Palestinian Aramaic of the 2nd century CE, but it also manifests features of Eastern Aramaic, and it has often been discussed whether it should be classified as Eastern or Western.11 One of the confusing features is the participle form: in the 1st and 2nd persons, base forms (as in Classical and Western Aramaic) and declined forms (as in Eastern Aramaic) are used side by side. I have found no attempt in the scientific literature to arrange the forms according to any rule.12 Indeed, it seems that no internal principle may explain the distribution between the forms. However, if we consider the Hebrew forms of the original, and take into account the influence of the original on the translation, we find a very clear rule:13 a Hebrew participle is always translated by the base form, e.g.:14 only twice with the participle; in the 3rd person, on the other hand, it is translated 19 times with the participle and never with the imperfect, e.g., ‫דדילמא משכח לה איניש‬ ‫ ואזיל כתיב מלעיל מאי דבאיי‬de-dilma maškaḥ lah ʾinneš we-ʾazel katev mi-lleʿel may devaye (67) ‘because someone may find it and he may go and write above whatever he wants’—‫ שמא ימצאה אדם וילך ויכתוב מלמעלה משהוא רוצה‬šemma yimṣa‌ʾah ʾadam we-yeleḵ we-yiḵtov mi-llemaʿla ma-šše-hu roṣe (53). 10  In fact, in Aramaic, the base form, too, should be viewed as a declined form, in which the person is expressed by a zero ending, so even in the 3rd person a pronoun should be added in Hebrew (this consideration may have caused the change of tense in 24% of the cases). However, the Hebrew base form is similar to the Aramaic base form, and this caused a failure to add a pronoun; see more on this in §3. 11  See M. Z. Kaddari, “Research on Onqelos Today,” in B. Uffenheimer (ed.), Bible and Jewish History (Tel-Aviv, 1971), 370–374 (in Hebrew). The readings from Targum Onqelos are according to A. Sperber, The Bible in Aramaic, vol. 1. (Leiden, 1959). 12  A. Dodi (The Grammar of Targum Onqelos [Ph.D. Diss., Bar-Ilan University, 1981], 420–421 [in Hebrew]) cites examples for the participle with a pronominal suffix expressing the agent, but does not specify any rule for its appearance. The same is true for Dalman (see the following note). 13  I collected all the participle forms in Genesis, with the addition of all the declined forms from Targum Onqelos quoted in G. Dalman, Grammatik des jüdisch-palästinischen Aramäisch (2nd ed.; Leipzig, 1905), 289–291. All the forms are enumerated in Appendix 2. 14  I counted 44 Hebrew participle forms translated by the base form of the participle. Only once it is translated by a declined participle: ‫ כי ארד אל בני אבל שאלה‬ki ʾered ʾel beni ʾavel šeʾola ‘for I will go down into the grave unto my son mourning’ (37:35)—

60

Breuer



‫אנא קאים אנכי נצב‬

ʾanoḵi niṣṣav ʾ ana qa‌ʾem ‘I stand’ (Gen 24:13) However, when the Aramaic participle translates other tenses, the declined participle may be used, e.g.:15

‫גניבנא גֻ נבתי‬

gunnavti genivna ‘I was stolen’ (Gen 40:15) ‫יכילנא אּוכל‬

ʾuḵal yeḵilna ‘I can’ (Num 22:38) ‫ תשתרר‬ ‫מתררבת‬

tistarer mitrarvat ‘you make yourself a prince’ (Num 16:13)

‫הידעיתון הידעתם‬

ha-ydaʿtem ha-yadʿitun ‘do you know’ (Gen 29:5) Here again the Hebrew declined forms (perfect and imperfect) are considered equivalent to the Aramaic declined participle. In the same way that in the Aramaic-Hebrew translation the Aramaic declined participle is translated by other tenses, in the Hebrew-Aramaic translation other tenses are translated by the Aramaic declined participle.16 ‫ ארי איחות לות ברי כיד אבילנא לשאול‬ʾare ʾeḥut lewat beri kid ʾavelna li-šʾol. It seems to me that this was done in order to avoid a confusion involving ‫‘ אבל‬mourning’: does it relate to ‫‘ ארד‬I will go down’ or to ‫‘ בני‬my son’. 15  In Targum Neophiti, written in Palestinian Aramaic, only the base form of the participle is used, as against the declined participle in Onqelos, e.g. (the forms here are from Genesis), ‫( הידעתם‬29:5)—‫( ידעיתון‬Onqelos)—‫( חכימין אתון‬Neophiti); ‫אנכי אערבנו‬ )43:9(—‫( אנא מערבנא ביה‬Onqelos)—‫( אנה ערב יתיה‬Neophiti); ‫( ידעתי בני ידעתי‬48:19)— ‫( ידענא ברי ידענא‬Onqelos)—‫( ידע אנא ברי ידע אנה‬Neophiti); or forms with the same tense as in the original, e.g., ‫( גֻ נבתי‬40:15)—‫( גניבנא‬Onqelos)—‫( אתגנבת‬Neophiti). 16  This discussion hardly contributes to the question of whether the Aramaic of Targum Onqelos should be regarded as Eastern or Western, since the appearance of both forms, the declined along the base form of the participle, points in contradicting directions. Considering the influence of the original on the translation, it may be claimed that only

aramaic-hebrew and hebrew-aramaic translations

3

61

3rd Person: Addition of the Pronoun

3.1 Hilḵot Reʾu While the Babylonian Aramaic 3rd person participle ‫ ָּכ ֵתב‬katev (without a person marker) contains a built-in expression of the 3rd person, there is only one Hebrew participle form for all persons, so the person is not expressed. Hence, in translating the Aramaic 3rd person participle into Hebrew, the independent pronoun should be added, as is done in the following examples:

‫ דקא אכיל ארעא דיתמי‬ ‫שהוא אוכל שדה יתומים‬

de-qa ʾaḵel ʾarʿa de-yatme (121) še-hu ʾoḵel sede yetomim (93) ‘That he enjoys the usufruct of a field of orphans.’

‫ בעידניה יתיב‬

‫במסכתו הוא יושב‬

be-ʿiddaneh yatev (122) be-masaḵto hu yošev (94) ‘He is engaged in his learning.’

‫ יכלא למימרא ליה‬ ‫יכולה היא לומר‬

yaḵla le-memar leh (107) yeḵola hi lomar (82) ‘She can say to him.’ ‫ בההיא הנאה דקא ציתי הדדי‬ ‫באותה הנאה שהין שומעין זה לזה‬

be-hahi hana‌ʾa de-qa ṣayte be-ʾotah hana‌ʾa še-hen šomʿin ze hadade (96) laze (73) ‘In the benefit that they listen to each other.’ '‫מה הן אומ‬ may ʾamre (18) ma hen ʾomrin (18) ‘What do they say?’ ‫ מאי אמרי‬

However, the independent pronoun is added only in the minority (14%) of examples, while in most cases it is not added:17 the declined form may prove an independent formulation according to Eastern Aramaic, while the base form may be attributed to the influence of the original. However, the base form also appears as a translation of other tenses, where no influence of the original may be assumed. 17  In my statistical calculations I did not take into account cases in which the subject already appears in the original, e.g., ‫ אינהו קא אמרי‬ʾinnehu qa ʾamri ‘they say’ (126)—'‫ הם אומ‬hem ʾomrim (97); ‫ בתראה קני‬batra‌ʾa qane ‘the last one has the purchase’ (119)—‫אחרון קונה‬ ʾaḥaron qone (91).

62

Breuer

‫ דמשכינו ליה דקלי וקא אכיל פיריהו‬

‫שהירהין לו דקלים ואוכל פירותיהם‬

de-maškinu leh diqle we-qa ʾaḵel še-hirhin lo deqalim we-ʾoḵel perayhu (48) perotehem (38) ‘That they pledged to him date palms and he eats their fruit.’

‫ וכד מתנח מזוזה מבריך‬

‫וכשמניח מזוזה מברך‬

we-ḵad matnaḥ mezuza u-ḵ-še-manniaḥ mezuza mevareḵ (36) mevareḵ (31) ‘And when he puts a mezuza he says.’

‫ אם בעיא יתבא גביה‬ ‫אם רוצה יושבת אצלו‬

ʾim baʿya yatva gabbeh (100) ʾim roṣa yoševet ʾeṣlo (77) ‘If she wants, she may stay with him.’ ‫ ברשותיה קימי וכדידיה דאמו‬

‫ברשותו עומדין וכשלו דומין‬

bi-ršuteh qayme u-ḵ-dideh bi-ršuto ʿomdin u-ḵ-šello damu (102) domin (77) ‘They are in his possession and are considered his.’ ‫ אישתכח דקא עיילין לבי דינא‬

‫לא נמצא שכשנכנסין לבית דין‬

ʾišteḵaḥ de-qa ʿaylin le-ve dina (123) lo nimṣa še-kke-še-niḵnasin le-vet din (94) ‘It turns out that they enter the court-room.’ Translating the participle into Hebrew without adding an independent pronoun is in fact an omission of the subject. The reason for this is obvious: the Aramaic 3rd person participle does not bear any visible person marker, and the 3rd person is expressed only by contrast with the other declined forms. Consequently, the base Hebrew participle, which also lacks a person marker, seems as if it is equivalent. 3.2 The Hebrew of the Babylonian Talmud One of the differences between Babylonian and Palestinian Hebrew is the insertion of a pronoun between the relative pronoun and the participle, which is typical of Palestinian Hebrew but is missing in Babylonian Hebrew, e.g., Palestinian Babylonian

‫ כאדם שהוא קורא בתורה‬ ‫כאדם שקורא בתורה‬

ke-ʾadam še-hu qore ba-ttora ke-ʾadam še-qqore ba-ttora (Mishna Beraḵot 1: 2; P. Talmud (B. Talmud Beraḵot 10b x3) Beraḵot 3a; 3c x2; 7b; Shabbat 3b x2) ‘As a man who reads in the Tora.’

aramaic-hebrew and hebrew-aramaic translations

63

Here again it seems to be an influence of Babylonian Aramaic on Babylonian Hebrew: since the Babylonian Aramaic 3rd person participle does not need an independent pronoun, the pronoun is also omitted in Hebrew, although in Hebrew this means an omission of the subject altogether.18 3.3 Conditioning Factor for the Appearance of the Pronoun As said above, the pronoun is added in HR only in a minority of cases. Still, it seems that its appearance is not altogether arbitrary and depends to some extent on the Babylonian Aramaic prefix qa: when the pronoun appears in Hebrew, in most cases qa appears in Aramaic, and when the pronoun does not appear in Hebrew, in most cases qa does not appear in Aramaic:19

18  See M. Bar-Asher, “The different traditions of Mishnaic Hebrew,” in D. M. Golomb (ed.), “Working with no data”, Semitic and Egyptian Studies presented to Thomas O. Lambdin (Winona Lake, 1987), 27, 33. According to Bar-Asher, it is not necessarily a Babylonian Aramaic influence, since it may also be a continuation of a Late Biblical Hebrew feature, as in ‫ שנאחזים‬še-nneʾeḥazim ‘that are taken’ (Eccl 9:12); ‫ ֶׁשּי ָֹצא‬še-yyoṣa ‘which proceeds’ (Eccl 10:5); ‫ ֶׁשּי ֵֹרד‬še-yyored ‘that goes down’ (Ps 133:2, 3). However, in the Bible, even in its late layer, there are interchanges, e.g., ‫ שהוא ָע ֵמל‬še-hu ʿamel ‘wherein he labors’ (Eccl 2:22); ‫ אשר הוא גר‬ʾašer hu gar ‘where he sojourns’ (Ezra 1:4); ‫ אשר הם בונים‬ʾašer hem bonim ‘which they build’ (Neh 3:35); ‫ אשר הם חיים‬ʾašer hem ḥayyim ‘which they live’ (2Chron 6:31); ‫ אשר המה חיים‬ʾašer hemma ḥayyim ‘which they live’ (Eccl 4:2). As a matter of fact, even in the early books there are interchanges, e.g., ‫ אשר יושב‬ʾašer yošev ‘which dwells’ (Num 21:34; Deut 1:4 [2x]; 3:2; 4:46); ‫ אשר י ֵֹצא‬ʾašer yoṣe ‘that springs’ (1Kgs 5:13); ‫ אשר מבקשים‬ʾašer mevaqšim ‘that seek’ (Jer 38:16), as against ‫ אשר הוא יושב‬ʾašer hu yošev ‘that they dwell’ (Num 13:19); ‫ אשר הוא עֹשה‬ʾašer hu ʿose ‘that he does’ (Gen 39:3); ‫אֹוכל‬ ֵ ‫ אשר הוא‬ʾašer hu ʾoḵel ‘which he eats’ (Gen 39:6); ‫ אשר הוא חֹנה‬ʾašer hu ḥone ‘where he encamps’ (Exod 18:5); ‫ אשר הוא מושיע‬ʾašer hu mošiaʿ ‘who saves’ (1Sam 10:19). In the Babylonian Talmud, on the other hand, this is the rule. Accordingly, although the option to omit the pronoun may have resulted from Biblical Hebrew, it is the uniformity in the Babylonian Talmud that requires an explanation, and in my view, it should be attributed to Aramaic influence. 19  The reason for the difference in the number of occurrences of the 3rd person participle between here and above (2.1) is that here, sentences with an explicit subject in the original were omitted (see above, n. 17). Such cases are significant for the tense but not for the pronoun.

64

Breuer

Hebrew

with a pronoun without a pronoun

Aramaic with qa

without qa

45 (76%) 26 (7%)

14 (24%) 344 (93%)

It is difficult to guess how this prefix was interpreted by the translator, but it is improbable that he regarded it as a pronoun.20 It seems that this, too, reflects a desire to resemble the original rhythm. This was done by exploiting two differences between the languages in the 3rd person participle: in Aramaic, there is a prefix qa but no pronoun; in Hebrew, there is no prefix qa but a pronoun is needed. Consequently, the pronoun tends to appear when qa appears in the original. 4 Conclusion This article focuses on the ways in which the Eastern Aramaic declined participle is translated into Hebrew. An accurate translation should have produced a Hebrew participle with the addition of an independent pronoun. This was done only in a minority of examples. In most cases, it is translated by only one word: the 1st and 2nd person participles are translated by the Hebrew perfect and imperfect; in the 3rd person, an independent pronoun is not added. Both phenomena have parallels in other sources: in the 1st and 2nd persons in Targum Onqelos, too, the Aramaic declined participle parallels the Hebrew perfect and imperfect; in the 3rd person, in Babylonian Hebrew, too, a pronoun is missing in the 3rd person participle. In both phenomena, resemblance of form is preferred over precision in function: change of tense means introducing a similar form (declined) in spite of the flaw in the tense system; and failing to add a pronoun results in the omission of the subject in order to introduce a similar form. This description also shows the difference between domains of the language in accepting a foreign influence. In all the cases, only pure Hebrew forms are used. While trying to imitate the Aramaic declined participle, no Hebrew 20  For a discussion on the function of this prefix, see Y. Breuer, “The Function of the Particle Qa in the Aramaic of the Babylonian Talmud,” Leshonenu 60 (1997): 73–94. (in Hebrew).

65

aramaic-hebrew and hebrew-aramaic translations

declined participle was created. Even the Hebrew declined participle of the 1st person (as in ‫ סבורני‬savurni ‘I think’), known from other sources, is not used here.21 It is only the syntax which is affected, whether by change of tense or by omission of the subject. This shows how far the syntax is ready to accept a foreign influence in comparison to morphology.

Appendix 1: Forms not Included in the Statistical Calculations

For calculating the relationship between the forms, the following types are not included: (1–4) forms which must be translated by the participle; I took into account only forms in which the translator could choose between the tenses; (5) sentences phrased in a different way; (6) forms whose tense is unknown. The following are the types of these cases with an example for each. 1.

A compound tense in Aramaic:

‫ והא אית לי שהדי דכל שתא הוה‬ ‫ אתית תלתין יומי‬

‫והלא יש לי עדים שכל שנה הייתה בא‬ ‫לכן שלשים יום‬

we-ha ʾit li sahade de-ḵol šatta wa-halo yeš li ʿedim še-kkol šana hawa ʾatet telatin yome (88) hayita ba le-ḵan šelošim yom (68) ‘But I have witnesses that every year you would come for thirty days.’ 2.

A compound tense in Hebrew:

‫ כי אתו קמיה דראבא בר רב‬ ‫ חונא אמ' להו‬

‫כשהיו באין לפני רבא בר רב חונה‬ ‫היה אומ' להן‬

ki ʾatu qameh de-Rava bar Rav ke-še-hayu ba‌ʾin lifne Rava bar Rav ḥuna ʾamar le-hu (123) ḥuna haya ʾomer lahen (94) ‘When they would come before Rava bar Rav Huna he would say to them.’ 3. State:

‫ אנא חזיאנא‬

ʾana ḥazyana (113) ‘I am entitled.’

‫ ניכסאי להאיי דמעברת‬

‫אני ראויה‬

ʾani reʾuya (87) ‫נכסיי לזה שאת מעוברת‬

niḵsay le-haye di-mʿabberat (118) neḵasay laze še-ʾat meʿubberet (90) ‘My property will go to the one with which you are pregnant.’ 21  On the rareness of a foreign influence in morphology, especially in declination, see, e.g., U. Weinreich, Languages in contact (The Hague,‎1963), 29, 31.

66

Breuer

4.

Participle plural expressing an obligation: the Aramaic 1st person plural participle, when expressing an obligation, is usually translated by a Hebrew plural participle without a subject. Since this is a fixed formulation, change of tense is not possible:



‫ מאי טעמא מברכינן על נטילת לולב‬ ‫ ואהנך אחריני לא מברכינן‬

‫מה טעם מברכין על נטילת לולב‬ ‫ועל הללו אין מברכין‬

may ṭaʿma mevarḵinan ‘al neṭilat ma ṭaʿam mevarḵin ‘al neṭilat lulav we-ʾa-hanaḵ ʾaḥrine la lulav we-‘al hallalu ʾen mevarḵinan (35) mevarḵin (30) ‘Why do we say the blessing on the ‘Why do (they) say the blessing on holding of the lulav but we do not the holding of the lulav but (they) say the blessing on the other do not say the blessing on the other species?’ species?’ 5.

Change of the formulation:

‫פרע לי מאה זוז שיש לי עמך פירען שית מאה זוזי דמסיקנא בך‬

pirʿan šit meʾa zuze de-massiqna peraʿ li meʾa zuz še-yyeš li bax (126) ʿimmax (97) ‘Pay me six hundred zuz that you owe me.’ 6.

The tense of the Hebrew form is unknown:

‫ אי מיתניס אי מיגניב‬

‫אם נאנס או נגנב‬

ʾi mitnis ʾi miggeniv (49) ʾim neʾenas ʾo nignav (39) ‘If it is seized or stolen.’

Appendix 2: The Participle in Targum Onqelos

1 The Base Form The base form usually occurs when there is a Hebrew participle, but it also is found when the Hebrew has other tenses. Here are the forms:22 Perfect: ‫( לית אנא ידע—לא ידעתי‬27:2); ‫( ואת מפקד—ואתה ֻצויתה‬45:19). Participle: ‫( ליט את—ארור אתה‬3:14; 4:11); ‫( הא אנא מיתי—הנני מביא‬6:17); ‫( אנא מחית מטרא—אנכי ממטיר‬7:4); ‫( הא אנא מקים—הנני מקים‬9:9); ‫—ואנכי הולך‬ ‫( ואנא אזיל‬15:2); ‫( דאין אנא—דן אנכי‬15:14); ‫( אנא ערקא—אנכי ברחת‬16:8); ‫( את מעדיא—הנך הרה‬16:11); ‫( דאת חזי—אשר אתה ראה‬13:15); ‫המכסי—המכסה אני‬ ִ ‫מחבלין‬ ‫( אנא‬18:17); ‫( דאנא עביד—אשר אני עשה‬18:17); ‫אנחנא—משחתים אנחנו‬ 22  Most of the forms are from Genesis. In quotations from other books, the book is listed.

‫‪67‬‬

‫‪aramaic-hebrew and hebrew-aramaic translations‬‬

‫—אשר אתה עשה ;)‪ (20:7‬ליתך מתיב—אינך משיב ;)‪ (20:3‬את מאית—הנך מת ;)‪(19:13‬‬ ‫;)‪ (24:13‬אנא קאים—אנכי נצב ;)‪ (24:3‬דאנא יתיב—אשר אנכי יושב ;)‪ (21:22‬דאת עביד‬ ‫—אנכי נצב ;)‪ (24:42‬דאנא אזיל—אשר אנכי הלך ;)‪ (24:37‬דאנא יתיב—אשר אנכי ישב‬ ‫משלהי אנא—עיף אנכי ;)‪ (24:49‬אם איתיכון עבדין—אם ישכם עשים ;)‪ (24:43‬אנא קאים‬ ‫אשר אתה ;)‪ (27:8‬דאנא מפקדא—אני מצוָ ה ;)‪ (25:32‬אנא אזיל—אנכי הולך ;)‪(25:30‬‬ ‫חזי אנא—רֹאה אנכי ;)‪ (28:20‬דאנא אזיל—אשר אנכי הולך ;)‪ (28:13‬דאת שכיב—שכב‬ ‫—אנכי מבקש ;)‪ (32:12‬דחיל אנא—ירא אנכי ;)‪ (31:43‬דאת חזי—אשר אתה רֹאה ;)‪(31:5‬‬ ‫—הנני עֹמד ;)‪ (41:9‬אנא מדכר—אני מזכיר ;)‪ (37:30‬אנא אתי—אני בא ;)‪ (37:16‬אנא בעי‬ ‫;)‪ (43:4‬אם איתך משלח—אם ישך משלח ;)‪ (42:18‬אנא דחיל—אני ירא ;)‪ (41:17‬האנא קאים‬ ‫;)‪ (43:18‬אנחנא מיתעלין—אנחנו מובאים ;)‪ (43:5‬ואם ליתך משלח—ואם אינך משלח‬ ‫—הנני מפרך והרביתך ;)‪ (45:12‬ארי בלישנכון אנא ממליל עמכון—כי פי המדבר אליכם‬ ‫אנא—אני נאסף ;)‪ (48:21‬האנא מאית—הנה אנכי מת ;)‪ (48:4‬האנא מפיש לך ומסגי לך‬ ‫‪ (50:24).‬אנא מאית—אנכי מת ;)‪ (50:5‬הא אנא מאית—הנה אנכי מת ;)‪ (49:29‬מתכניש‬ ‫—תעמד ;)‪ (19:19‬לית אנא יכיל—לא אּוכל ;)‪ (16:8‬את אזלא—ואנה תלכי ‪Imperfect:‬‬ ‫—תבקש ;)‪ (32:18‬את אזיל—תלך ;)‪ (31:39‬את בעי לה—תבקשנה ;)‪ (24:31‬את קאים‬ ‫;)‪ (41:15‬דאת שמע חלמא ומפשר ליה—תשמע חלום לפתר אתו ;)‪ (37:15‬את בעי‬ ‫‪ (46:30).‬אילו אנא מאית זמנא הדא מנחם אנא—אמותה הפעם‬ ‫ ‪2‬‬ ‫‪The Declined Participle‬‬ ‫‪The declined participle parallels only the Hebrew perfect or imperfect (with‬‬ ‫‪one exception).‬‬ ‫;)‪ (12:11‬ידענא—ידעתי ;)‪ (9:15‬ודכירנא—וזכרתי ;)‪ (4:9‬לא ידענא—לא ידעתי ‪Perfect:‬‬ ‫דענָ א—ידענו ;)‪ (29:5‬הידעיתון—הידעתם‬ ‫אתה ;)‪ (30:26‬את יָ ַדעת—אתה ידעת ;)‪ (29:5‬יָ ִ‬ ‫ידעתי בני ;)‪ (44:15‬ידעיתון—ידעתם ;)‪ (40:15‬גניבנא—גֻ נבתי ;)‪ (30:29‬את יָ ַדעת—ידעת‬ ‫מהודענא—והודעתי ;)‪ (Ex 18:16‬דאיננא—ושפטתי ;)‪ (48:19‬ידענא ברי ידענא—ידעתי‬ ‫ודכרנא—וזכרתי ;)‪ (Ex 23:9‬ידעיתון—ידעתם ;)‪ (Ex 21:5‬רחימנא—אהבתי ;)‪(Ex 18:16‬‬ ‫—רכבת ;)‪ (Num 22:30‬אליפנא—הסכנתי ;)‪ (Num 22:6‬ידענא—ידעתי ;)‪(Lev 26:45‬‬ ‫‪ (Num 23:11).‬מבריכת—ברכת ;)‪ (Num 22:30‬רכיבת‬ ‫—אדבר ;)‪ (Ex 16:7‬מתרעמתון—תלינו ;)‪ (43:9‬מערבנא ביה—אערבנו ‪Imperfect:‬‬ ‫‪ (Num 22:38).‬יכילנא—אּוכל ;)‪ (Num 16:13‬מתררבת—תשתרר ;)‪ (Num 12:8‬מלילנא‬ ‫כיד—א ֵבל ‪Only once the declined participle parallels the Hebrew participle:‬‬ ‫ָ‬ ‫‪ (37:35).‬אבילנא‬

The Proto-Semitic “Asseverative *la-” and the Innovative 1SG Prefixes in South Ethio-Semitic Languages Maria Bulakh 1 Introduction The verbal paradigms of the prefix conjugation in many Ethio-Semitic (ES) languages have subject indexes (in traditional terminology, subject agreement markers) which coincide in shape—all having the proto-form *l(V)—but not necessarily in meaning and distribution. At least some of them go back to the common Semitic asseverative particle *la-. This particle has been reconstructed in Huehnergard 1983, where the functions of its reflexes across Semitic are described. According to Huehnergard (1983: 579–580, 592), the “asseverative *la-” is the source of the “injunctive” proclitic la- in Geʿez and of a number of subject indexes in modern ES languages. Huehnergard’s reconstruction implies the existence of a proto-ES “injunctive” particle *la-, compatible only with jussive verb forms. Geʿez is then the only ES language which preserves this element intact as far as its shape, functions, and distribution are concerned. In Tigrinya, no reflexes of the “injunctive” *la- have been found. In Tigre and South Ethio-Semitic (SES), the reflexes are no more clitics: they have developed into subject indexes of the verbal para­digms. Such a development is not an exclusive feature of ES (similar developments came about in Akkadian, Aramaic, and MSA). The innovative prefixes in SES are mostly limited to 1SG. Huehnergard (1983) aims at reconstructing several formally and functionally similar elements of proto-Semitic. In his brief, albeit very accurate, analysis *  The investigation was carried out in the framework of the project #2992 supported by the Ministry of Science and Education of Russia. I am deeply grateful to Dr. S. Loesov, who carefully read the draft version and amended the text considerably. Needless to say, all the errors are mine. Note the following abbreviations: Amh. = Amharic; Arg. = Argobba; Čah. = Čaha; DECL = declarative; Dob. = Dobbi; End. = Endegañ; Enm. = Ennemor; ES = Ethio-Semitic; Gaf. = Gafat; Har. = Harari; Kst. = Kestane; LPC = long prefix conjugation; MSA = Modern SouthArabian; Msq. = Masqan; Muh. = Muher; IMPF = imperfective; NEG = negative; PC = prefix conjugation; PS = Proto-Semitic; SPC = short prefix conjugation; Sel. = Selṭi; SES = South Ethio-Semitic; Wol. = Wolane.

© koninklijke brill nv, leiden, ���5 | doi ��.��63/9789004300156_006

Innovative 1SG Prefixes in South Ethio-Semitic Languages

69

of the ES data he discusses only the jussive subject indexes and makes no claim concerning the indicative paradigms. Yet certain innovative subject indexes of the indicative are traceable to the *la- of the ES jussive. In Bulakh 2013 I reconstructed the way in which a jussive prefix lə- had penetrated into the indicative paradigm of Tigre. A closer investigation of ES material reveals that the subject indexing prefixes containing the element *l (henceforth referred to as *l-prefixes) do not necessarily all go back to the same entity. For instance, the negative indicative paradigms in Zay and Wolane are analyzed by R. Meyer (2005: 191–192; 2006: 107, 110) as containing a negative prefix ʔa- and a 1sg prefix -l-: ʔa-l-šäwwər-u NEG-1SG-return.IMPF-DECL ‘I do not return’. Yet historically -l- of the negative paradigm is not a subject index but a part of the negative marker ʔal(cf. Section 4.1). Some other cases are less transparent and require special consideration. The present investigation explores the innovations in the 1SG prefixes in various paradigms of the prefix conjugation in SES and attempts to establish their origins, various strategies of readjustments of the subject indexes, as well as various factors that may have influenced these readjustments. 2

Shared Innovations, Parallel Development and Areal Diffusion

The analysis of the innovations in the 1SG of the prefix conjugation allows us to establish several isoglosses. Some of them are shared innovations in R. Hetzron’s (1976) sense, that is, they are inherited from a common protostage. Others are parallel developments (common drifts). Parallel developments may take place in languages that have no contact with each other as well as in neighboring languages. In the latter case, one usually surmises areal diffusion rather than truly independent changes in individual languages (see Aikhenvald and Dixon 2006: 4). Therefore, geographic proximity and/or intensive contact between separate languages within SES should not be overlooked in the present investigation. As is well known, ES languages have always existed in a situation of manifold contact, and many features characteristic of ES languages have areal rather than genealogical origin. Special studies on language contact in Ethiopia mostly focus on the interaction between Cushitic and Semitic (e.g., Crass and Meyer 2011). Contacts between separate Ethio-Semitic languages have received less attention. Still, it is an important factor which cannot be ignored in EthioSemitic studies. Adstratal influence of Amharic as a lingua franca and the language of instruction is frequently mentioned in modern descriptions of other

70

Bulakh

SES languages (e.g., Meyer 2005: 16–17 for Zay, Meyer 2006: 17–18 for Wolane). In the case of Argobba, direct contact with Amharic is due to the geographic proximity (see Wetter 2010: 18, 22–23). Likewise, contact between various SES languages of the Gurage zone is well known and frequently mentioned (see, e.g., Meyer 2006: 18 for Wolane, Meyer 2007: 1–2 for Muher). Besides, languages which are no more neighbors might have been in contact for a long time in the past. Such is probably the case of Argobba and Harari. The southern variety of Argobba was spoken in the vicinity of Harar till the first half of the 20th cent. (see Wetter 2010: 24 with further references). It stands to reason that Argobba was more widespread in medieval Ethiopia (see Wetter 2010: 25) and that the presence of Argobba speakers around and perhaps even within Harar led to intensive language contact. The East Gurage Selṭi (Braukämper 2004: 65 with further references) and Wolane (Meyer 2001) were likely settled in the Harar region before migration towards the Gurage area. This reconstruction is supported by their linguistic closeness to Harari language. Thus, they might have been geographically close both to Harari and Argobba. There is no special link to Harar in the history of Zay, the third East Gurage language, yet it is not improbable that for some time after the split from the Selṭi and Wolane it remained their neighbor and thus could have been in contact with Harari and Argobba as well. Besides, all three East Gurage languages could have been in contact with each other before they migrated westwards. The geographical closeness and contact between separate SES languages during various periods of their history likely facilitated the development of common features not inherited by them from proto-SES. 3

The Innovative *l-prefix in the Proto-SES Prefix Conjugation

As is well-known, the ES languages possess two types of prefix conjugation, which are distinguished through two bases: the “long” (*-С1aC2(C2ə)C3) and the “short” (*-C1C2a/əC3) base. The “long” prefix conjugation (LPC) is employed to express NON-PAST INDICATIVE, whereas the “short” prefix conjugation (SPC) is the exponent of JUSSIVE. In traditional grammars, they are referred to as the Imperfect and the Jussive, respectively. In modern descriptions, “the Imperfective” has replaced “the Imperfect”. The subject indexes are either prefixes or combinations of prefixes and suffixes (circumfixes; throughout the article, the terms “prefixes” and “suffixes” will be applied both to independent morphemes and parts of circumfixes). In Geʿez, as well as in Tigrinya, LPC and SPC have the same set of subject indexes

Innovative 1SG Prefixes in South Ethio-Semitic Languages

71

(marking person, gender and number), which is apparently very close to, or even identical with, the proto-ES set of subject indexes (see Table 1). TABLE 1  The subject indexes of the prefix conjugation in proto-ES

1 2 MASC 2 FEM 3 MASC 3 FEM

SG

PL

*ʔə*tə*tə- . . . -i *yə*tə-

*nə*tə- . . . -u *tə- . . . -ā *yə- . . . -u *yə- . . . -ā

In SES and in Tigre, the prefixes in LPC and SPC are not fully identical. The innovations in the 3 person prefixes in Tigre are discussed in Bulakh 2013. As for SES, the most conspicuous difference between the subject indexes of LPC and SPC is in the 1SG: LPC preserves the archaic *ʔə- in all the SES languages at least as one of the allomorphs, whereas SPC is characterized by the ubiquitous innovative 1SG prefix *lə-, going back to the “injunctive” *la-.1 None of the innovative 1SG prefixes of LPC is traceable to proto-SES, and thus, the distribution between the 1SG prefixes *ʔə- (LPC) and *lə- (SPC) can be safely reconstructed for proto-SES. Another specifically SPC morpheme appearing in most SES languages is the prefix yä- in 3SG.MASC/PL (opposed to the archaic yə- in the corresponding slots of LPC). According to E. Wagner (1968, see also Hetzron 1977: 79–80), this morpheme is a result of a merger of the “injunctive” *la- and the subject index *yə-. Wagner dates the merger as early as the proto-SES stage and explains the absence of the innovative yä- in Amharic and Argobba by secondary paradigmatic levelling.2 If Wagner’s hypothesis is accepted, the subject indexes for LPC and SPC of proto-SES should be reconstructed as in Table 2. 1  For the pan-SES status of this latter feature, see Bulakh and Kogan 2010: 279. 2  The proto-SES status of this feature is disputable: Amharic and Argobba may have never possessed this feature and hence it may have emerged after proto-SES split into daughter languages. Still, the advantage of Wagner’s reconstruction is that it presupposes a simultaneous fusion of the clitic with the subject index in the 1st and 3rd person. Within Wagner’s hypothesis, the early proto-SES picture is quite similar to that of Soqotri, where l-, the reflex of the

72

Bulakh

TABLE 2  The subject indexes of the prefix conjugation in proto-SES LPC

1 2 MASC 2 FEM 3 MASC 3 FEM

SPC

SG

PL

SG

PL

*ʔə*tə*tə- . . . -i *yə*tə-

*nə*tə- . . . -u *tə- . . . -a3 *yə- . . . -u *yə- . . . -a

*lə- (< *la-ʔə-) *tə*tə- . . . -i *yä- (< *la-yə-) *tə-

*nə*tə- . . . -u *tə- . . . -a *yä- . . . -u (< *la-yə- . . . -u) *yä- . . . -a (< *la-yə- . . . -ā)

Thus, on the proto-SES level the integration of the element *la- into the system of subject indexes was restricted to SPC. This original distribution is well compatible with the distribution of the Geʿez “injunctive” clitic la- and speaks in favour of Huehnergard’s reconstruction. The emergence of *l-prefixes in LPC must be dated to a more recent period. Are they directly taken over from SPC? In the subsequent sections, I will try to answer this question. 4

Subtypes of the LPC in SES

LPC in SES develops a number of subtypes, and the innovative 1SG prefixes are often restricted to some of these subtypes. Firstly, all of the languages of this branch oppose affirmative and negative paradigms of LPC (Muh. ä-säbr-u ‘I break’ vs. annə-säbər ‘I do not break’). Secondly, in Argobba of Ṭollaha, East Gurage, Gafat and a number of Gunnän-Gurage languages the word-initial personal prefixes of LPC are not identical to the personal markers preceded by the conjunction prefixes: Zay yə-näḳəl-āhw ‘I take’ vs. tī-l-näḳəl-u ‘while I take’. For brevity’s sake, I will refer “injunctive” clitic, appears in the 1SG, 3SG.MASC, 3PL.MASC/FEM (and 1 and 3DU, absent in ES; see Simeone-Senelle 1997: 405) of jussive, thus, before historical prefixes *ʔə- and *yə-, but not elsewhere. 3  The distinction between masculine and feminine plural is lost in Amharic, Argobba, Harari and East Gurage. In Gunnän-Gurage, the markers are innovative. The simplest solution is to assume that the archaic morphemes were still preserved in proto-SES.

Innovative 1SG Prefixes in South Ethio-Semitic Languages

73

to this dichotomy as the opposition between Anlaut and Inlaut paradigms. Occasionally I will also speak of Anlaut und Inlaut forms in Amharic and Argobba of Aliyu Amba, alluding to verbal forms without and with conjunction prefixes, albeit these languages employ the same set of personal markers in both cases. In Harari (both Ancient and Modern), Inlaut forms of LPC are virtually absent since conjunctions are mostly postpositional in this language. Finally, all modern Transversal SES languages4 distinguish between the simple and compound LPC (simple and compound imperfect in most grammatical descriptions). The simple paradigm is restricted to certain syntactic contexts (subordinate clauses, analytic constructions), whereas the compound LPC is primarily used to express the non-past tense of a verb in the main clause. The paradigm of the simple LPC (Amh. yəsäbər) is the direct reflex of the proto-ES paradigm of LPC; the paradigm of the compound LPC (Amh. yəsäbrall) goes back to the analytic construction LPC + auxiliary *hallawa (in Ancient Harari, they apparently still function as analytic constructions,5 whereas the simple forms are commonly used both in main and dependent clauses). In most of the languages, the simple and compound LPC employ the same set of subject indexes. Only Argobba of Aliyu Amba has two different subsets of subject indexes (see Section 8.3.2). 4.1 Negative Paradigm in Proto-SES Among the paradigmatic oppositions outlined above, only the affirmative vs. negative dichotomy existed as early as proto-SES. Cumulative morphemes which express subject indexing and negation are found in most SES languages (see Bulakh 2012: 393–398) and thus can be reconstructed for proto-SES. In these morphemes, the opposition between the 1SG and the rest of the paradigm is especially prominent: the element -l- surfaces in NEG.1SG, whereas elsewhere, the final consonant of the negative morpheme is assimilated to or absorbed by the initial consonant of the subject index. As a result, the element -l- of the 1SG becomes opposed to the other consonants of the cumulative morphemes and is easily reinterpreted as the subject index. This reinterpretation 4  The term “Transversal SES” will be understood here as a cover term for Amharic, Argobba, Harari and East Gurage (Wolane, Selṭi, Zay) without implying a genealogical unity (see Bulakh and Kogan 2013: 604–606). 5  I was unable to find examples of the 1SG in such analytic constructions of Ancient Harari. Therefore, strictly speaking, it remains unclear whether they employed the same subject indexes as the simple LPC. It is, however, highly improbable that two subsets of subject indexes existed at this early stage, before the analytic forms were grammaticalized into the compound LPC.

74

Bulakh

is facilitated by the presence of a subject index l- in SPC of proto-SES (see Table 3). TABLE 3  The subject indexes of SPC in the proto-SES compared to those of the negative LPC SPC

1 2 MASC 2 FEM 3 MASC 3 FEM

Negative LPC

SG

PL

SG

PL

*lə- (< *la-ʔə-) *tə*tə- . . . -i *yä- (< *la-yə-) *tə-

*nə*tə- . . . -u *tə- . . . -a *yä- . . . -u (< *la-yə- . . . -u) *yä- . . . -a (< *la-yə- . . . -ā)

*ʔal*ʔat(tə)*ʔat(tə)- . . . -i *ʔay*ʔat(tə)-

*ʔan(nə)*ʔat(tə)- . . . -u *ʔat(tə)- . . . -a *ʔay- . . . -u *ʔay- . . . -a

The -l- of the negative paradigm can hardly be related to the “asseverative *la-” (for the criticism of such comparison, see Bulakh 2012: 393–394). The formal identity of the element -l- in the subject index of SPC with the element -l- of the negative LPC is purely accidental.6 However, the structural similarity between subject index sets of SPC and the negative LPC is not a coincidence. Rather, it is a good example of a parallel development triggered by one and the same factor. In both cases original proclitics (the injunctive and negative markers) were fused with the subject indexing prefixes and subsequently entered the system of subject indexes. The injunctive *la- has persisted in the 1SG slot only. The negative marker *ʔal- appears as such in the 1SG slot only, with -l- opposed to the consonants of the subject indexing prefixes elsewhere;7 such a distribution made it possible to reanalyze the cumulative morpheme and to treat the element -l- as a 1SG marker (with ʔa- as a negative marker; such interpretations have been offered 6  The similarity between SPC and the negative LPC was restricted to the subject indexes: the bases of SPC and the negative LPC were different in proto-SES. As for the employment of the SPC base in the negative LPC, it occurs in a few daughter languages only and is obviously a more recent phenomenon caused by Cushitic influence (see Wagner 1968; Bulakh 2012: 387–388). 7  In Bulakh 2012: 407–409, the SES negative marker is thought to go back to *ʔay-, with -ystrengthened to -l- in the 1SG.

Innovative 1SG Prefixes in South Ethio-Semitic Languages

75

for Zay and Wolane, see Meyer 2005: 191, Meyer 2006: 107). In both cases, the result of the innovation is the pronounced opposition between the 1SG prefix and the rest of the subject indexes. As it turns out, the fusion of the 1SG prefix and the preceding morpheme both in SPC and in the negative LPC paradigms is due to acoustic similarity between 1SG and 3SG.MASC prefixes, which will be discussed below (see Section 7.1). 5

Phonological Shape of the 1SG Prefixes of LPC in SES

The 1SG markers found in various subtypes of the prefix conjugation in SES (see Table 4) are either archaic (that is, traceable to the proto-ES *ʔə-) or innovative. TABLE 4  1SG markers in the prefix conjugation of SES8 Affirmative LPC (Simple) SPC Simple Compound Inlaut Negative LPC (Anlaut) (Anlaut) (NEG.1SG)9 Amh. Arg. of Aliyu Amba Arg. of Ṭollaha Ancient Har.

əəl-/ä∅u-, i-

əə∅u-, i-

-ə-(ə)l-l-/-ll—

alalaC-10 al-

lələləli-/la- (rarely n-)

8  Tables 4–6 contain the data from those SES languages for which I was able to find the relevant forms (dashes indicate the absence of the subtype in the given language, question marks indicate the absence of reliable data). The sources are Leslau 1995 for Amharic, Leslau 1997 for Argobba of Aliyu Amba, Wetter 2010 for Argobba of Ṭollaha, Leslau 1958 for Modern Harari, Cerulli 1936 and Wagner 1983 for Ancient Harari, Gutt 1997 for Selṭi, Meyer 2005 for Zay, Meyer 2006 for Wolane, Leslau 1956 for Gafat, Leslau 1968 for Kestane, Leslau 1981 for Muher, Hetzron 1977 for Dobbi, Leslau 2004 for Masqan, Leslau 1983 and Hetzron 1977 for Čaha, Leslau 1983 for Ennemor, Leslau 1971 (repr. in 1992) and Eyassu Nega 2003 for Endegañ. For Inlaut paradigms, see also fn. 13. 9  The prefixes are given as they occur in simple LPC forms. In most Transversal SES languages, the same morphemes occur in the compound LPC; only in Harari, Selṭi and Wolane, the compound forms attach the negative morpheme in a different way and lack the cumulative morpheme comparable with that of the simple LPC; for details, see Bulakh 2012, Table 1, with fn. 46. 10  C stands for the gemination of the first radical.

76

Bulakh

TABLE 4  1SG markers in the prefix conjugation of SES (cont.)

Simple (Anlaut) Modern Har. Sel. Zay Wol. Gaf. Kst. Dob. Muh. Msq. Čah. Enm. End.

Affirmative LPC (Simple) SPC Compound Inlaut Negative LPC (Anlaut) (NEG.1SG)

iiyəyəəääääää? ə- or ä-11

ii*yəyə— — — — — — — —

— -l-l-l-l-ä-ä-ä? -n-n?

analʔalʔalʔalannəannəannəannəanãan-

nala*lälälənänənənənənənä-

5.1 Reflexes of the Archaic *ʔəAmharic, Argobba, Harari, Selṭi and Gafat display reflexes of the archaic prefix *ʔə-: ə- in Amharic, Argobba and Gafat, i- in Harari, phonologically conditioned i-/u- in Ancient Harari. Almost all Gunnän-Gurage languages (see fn. 11 on Endegañ) employ the 1SG prefix ä- in the Anlaut. This is a result of the word-initial shift ə- > ä- that took place in all the peripheral SES languages (see Hetzron 1977: 34).12 However, in some Gunnän-Gurage languages (Kestane, Dobbi, Muher) the same prefix appears in the Inlaut. The word-internal shift ə- > ä- cannot be explained in terms of phonologically conditioned changes in Gunnän-Gurage. Yet it is likely that the vocalic opposition between the 1SG and the rest of the LPC pre11  It is ə- according to Leslau 1971 (see Leslau 1992: 468–469, 473; note, however, that Leslau doubts the correctness of his own data), but ä- according to Eyassu Nega 2003: 83. 12  It is hardly possible to regard this marker as a continuation of the combination *la-ʔə(similar to *la-yə-, the predecessor of the SPC 3rd person marker yä-, reconstructed by E. Wagner, see Section 3). Firstly, the form *la-ʔə- lacks the palatalyzing factor -y-. Secondly, unlike the 3rd person SPC marker yä-, the 1SG prefix with ä-vocalism never manifests the initial y-, which could be traced back to l.

Innovative 1SG Prefixes in South Ethio-Semitic Languages

77

fixes, originating in the word-initial shift, was reinterpreted as a morphological opposition and as such taken over into the Inlaut paradigm. In Argobba of Aliyu Amba, the prefix ä- (a variant of əl- in the word-initial position of the simple LPC) can also be traced back to the archaic *ʔə-. The word-initial phonological shift ə- > ä- in Argobba is optional (see Leslau 1997: 9). The distribution of the 1SG marker ä- in Argobba is narrower than that of its counterpart in Gunnän-Gurage: according to Leslau, the compound LPC employs the prefix ə- (see Section 8.3.2). The 1SG marker in the LPC paradigm of Endegañ is transcribed as either əor ä- (see fn. 11). In both cases it is comparable to other SES morphemes going back to the archaic *ʔə-. 5.2 Innovative Morphemes The affirmative LPC paradigms of SES contain the following innovative 1SG markers: 5.2.1 The prefixes that go back to *-l-: a) -l-: Argobba of Aliyu Amba, Argobba of Ṭollaha (Inlaut paradigm), East Gurage (Inlaut paradigm), Gafat (Inlaut paradigm), b) -n-: West Gurage languages13 (Inlaut paradigm), 5.2.2 əl-: Argobba of Aliyu Amba (simple Anlaut paradigm),14 5.2.3

yə-: Zay and Wolane (Anlaut paradigm),

5.2.4

Zero marker in Argobba of Ṭollaha (Anlaut paradigm).

The subsequent sections will deal with the innovative prefixes and the reasons for their emergence. 13  In all the involved languages, -n- is the regular reflex of the preconsonantal -l- (Hetzron 1977: 40–41; Leslau 1979: xlviii). There are no data on the Inlaut 1SG prefix (whether -n- or otherwise) for all Gunnän-Gurage languages. Hetzron (1977: 78) makes a general statement, “ä- is replaced by -n- after a subordinating prefix,” but he adduces examples only from Čaha and Eža (both belong to West Gurage). It appears that this -n- is not ubiquitous in Gunnän-Gurage: note such forms as Kst. t-äḳärs ‘I do not begin’ (Leslau 1992: 167, Goldenberg 1968: 95), Dob. t-äfärd ‘when I judge’ (Hetzron 1977: 164), Muh. b-änäfa ‘while I shout’ (Leslau 1981: 120). 14  It is treated here separately from the -l- of the Inlaut paradigm. For a detailed discussion, see Section 8.3.2.

78 6

Bulakh

Syncretism Involving 1SG Prefixes in SES

In SES, the 1SG prefixes sometimes coincide with the 1PL prefixes15 (see Table 5) or with the 3SG.MASC prefixes (see Table 6). TABLE 5  1SG vs. 1PL in LPC (simple Anlaut, simple Inlaut and negative paradigms) and in SPC16 Simple Anlaut LPC

Amh. Arg. of Aliyu  Amba

Inlaut

SPC

(Simple) Negative LPC

1 SG

1 PL

1 SG 1 PL

1 SG

1 PL

1 SG

1 PL

əəlä-

ən(nə)əl- . . . -n

-ə-l-

-nn-l- . . . -n

lələ-

ənnələ- . . . -n

alal-

anal- . . . -na-

ənnəna-

aCal-

annəan-

anal-

anal- . . . -na

ʔalʔal-

ʔal- . . . -n ʔal- . . . -nä

Arg. of Ṭollaha ∅Ancient Har.

u-, i-

ənnənu-, ni-

-ll—

-nn—

Modern Har.

ii-

nii- . . . -na

— -l-

— ? -l- . . . -na

ləli-/la(rarely n-) nala-

Wol.

*yəyə-

*yə- . . . -na *yə- . . . -na-

-l-l-

-l- . . . -nä -l- . . . -nä

*lälä-

nala- . . . -na ya- . . . -na *lä- . . . -nälä- . . . -nä

Gaf.

ə-

ənnə-

-l-

-nn-

lə-

?

ʔal-

annə-

Kst.

nənə- . . . -nänə- . . . -nä-

Msq.

ä-

ä- . . . -nä

änə- . . . -nä änə- . . . -nä ? ə- or ä-17 nə- . . . -nä

-nn? -nn- . . . -nä ? ? ? ?

nə-/ənnə-

Čah.

-ä-ä-ä? -n-n?

nä-

Muh.

äää-

nənənənənənä-

nə- . . . -nä nə- . . . -nä nə- . . . -nä nə- . . . -nä nə- . . . -nä nä- . . . -nä

annəannəannəannəanãan-

annəannə- . . . -nä annə- . . . -nä annə- . . . -nä an- . . . -nä ã- . . . -näan- . . . -nä-

Sel. Zay

Dob.

Enm. End.

15  This is a well-known instance of a parallel development within Semitic, for which cf. Goldenberg 1977: 481–482. 16  For the sources, see fn. 8. 17  See fn. 11.

79

Innovative 1SG Prefixes in South Ethio-Semitic Languages TABLE 6  1SG vs. 3SG.MASC in LPC compound, simple Anlaut and simple Inlaut paradigms18 Compound (Anlaut)

Amh. Аrg. of Aliyu Amba Arg. of Ṭollaha Modern Har. Ancient Har. Sel. Zay Wol. Gaf. Kst. Dob. Muh. Msq. Čah. Enm. End.

Simple (Anlaut)

Simple (Inlaut)

1SG

3SG.MASC

1SG

3SG.MASC

1SG

3SG.MASC

əə-

yəyə-

əəl-/ä-

yəyə-

-ə-l-

-i-i-

∅i? iyəyə-

∅yiyiiyəyə-

∅iu-, iiyəyə-

∅yiyiiyəyə-

-l— — -l-l-l-

-i— — -∅-∅-i-

— — — — — — — —

— — — — — — — —

əääääää? ə- or ä-19

yəyəyəyəyəyəyə? ə- or yə-20

-l-ä-ä-ä? -n-n?

-i-i-i-i? -i-i-i-ī-

As Table 5 shows, most languages that have the same prefix in the 1SG and PL forms preserve the number opposition, which is usually expressed by the innovative 1PL suffix *-na: Zay läsbär ‘I shall break’ vs. läsbär-nä ‘we shall break’. Thus, we are dealing with the formal identity of prefixes rather than morphemes (block syncretism rather than whole-word syncretism in terms of Stump 2001: 217–218). Whole-word syncretism is observed only rarely: Har. SPC

18  For the sources, see fn. 8. 19  See fn. 11. 20  ə- according to Leslau 1971, see also Leslau 1992: 468; yə- according to Eyassu Nega 2003: 82.

80

Bulakh

nasbar ‘let me/us break’. An idiosyncratic development is observed in Kestane: only the consonants of the 1SG and 1PL prefixes coincide, whereas the vowels successfully preserve their opposition. Thus, no syncretism is observed at all: nä-sfər ‘let me measure’ vs. nə-sfär/ənnə-sfär ‘let us measure’. The 3SG.MASC prefix is identical with the 3PL prefix throughout SES. Therefore, the languages with syncretism between 1SG and 3SG.MASC are also characterized by the homophony of the 1SG and 3PL prefixes. This formal identity of prefixes does not lead to the syncretism of morphemes. As shown in Table 1, the subject indexing in proto-ES was expressed by combinations of prefixal and suffixal elements. The suffixal elements (sometimes developed into suprasegmental morphemes—labialization or palatalization applied to the whole word) are preserved in the daughter languages. The presence of these elements rules out the possibility of the formal coincidence between 1SG and 3PL: Wol. yəsäbr ‘I break’ vs. yəsäbru ‘they break’ (Meyer 2006: 110), Zay yənäḳlən ‘I take’ vs. yunoḳlun ‘they take’ (Meyer 2005: 96). On the contrary, there is no suffixal element in the 3SG.MASC of proto-ES or its daughter languages. As a result, lack of distinction between the prefixes *ʔəand *yə- leads to the whole-word syncretism between the 1SG and 3SG.MASC: Zay yənäḳlən ‘I take’/‘he takes’. 7

The Conditions that Favour the Emergence of the Innovative LPC Prefixes

7.1 The Formal Similarity between the Prefixes *ʔə- and *yəIn a study focusing on the evolution of the prefix conjugation in Eastern Aramaic languages Rubin (2007) posits a common factor that has triggered parallel processes in individual languages of this group. This factor is the acoustic similarity between the 1SG prefix *ʔV- and the 3SG.MASC/PL prefix *yV- and hence the danger of loss of their distinction. According to Rubin, one of these morphemes has to be replaced by an innovative one in order to prevent their merger. Both *ʔV- and *yV- are safely reconstructed for the proto-languages of all branches of Semitic and for proto-Semitic. Thus, the potential trigger of the innovation was present in proto-Semitic and in the ancestor languages of all the subbranches. Moreover, the phenomena which enhance the danger of the loss of distinction between ʔV- and yV- are not restricted to the Aramaic language group. Rubin reconstructs a shift *yə-(yi-) > i- as the first step toward the potential merger between the two prefixes. Similar processes (loss of word-initial y-,

Innovative 1SG Prefixes in South Ethio-Semitic Languages

81

“de-jotation”), as well as opposite developments (insertion of prothetic y-— that is, “jotation”—or ʔ- to avoid the vocalic Anlaut) are attested in various Semitic languages, including ES (see Tropper 1995: 30–35). In theory, the most straightforward result of the weakening of the formal distinction between the reflexes of *ʔV- and *yV- would be the complete loss of distinction between the two prefixes. Thus, one expects to find in Semitic results of two alternative paths of development: either the invention of a new device to reinforce the morphological opposition (as in Eastern Aramaic) or the syncretism of 1SG and 3SG.MASC prefixes. For the brevity’s sake, I will refer to the former process as “paradigmatic dissimilation,” and to the latter one as “paradigmatic merger.”21 Both developments are actually attested in Semitic languages. Yet, the “paradigmatic merger” is less common than the “paradigmatic dissimilation” and is mostly restricted to certain subtypes of prefix conjugation. The conditions which favor the “paradigmatic merger” of the two prefixes will be discussed below (Section 7.2). The “paradigmatic dissimilation” can be observed in various subtypes of prefix conjugation in different branches of Semitic (see Rubin 2007, Loesov 2012: 442–452 for Eastern Aramaic; Rubin 2007: 17–19, Bulakh 2013 for Tigre; Bulakh 2012: 397, fn. 16, for tentative examples in the negative PC of ES; Bulakh 2013: 386–387 for further examples). The same type of development can be seen in the proto-SES paradigms of SPC and the negative LPC (see Section 4.1). The innovative morphemes are mostly restricted to some subtypes of PC. Sometimes, the restricted compatibility of an innovative morpheme can be explained by the properties of its source. Thus, the innovative subject index going back to the “injunctive” enclitic *la- is reconstructed only for SPC in protoSES (see Section 3). The morphemes that result from interaction between the subject indexes and the negative marker can of course occur in the negative paradigms only (see Section 4.1). 21  These terms will be used to refer both to morphological readjustments and to the changes which are the direct results of phonological processes. Incidentally, none of the SES para­ digms considered in this paper demonstrates syncretism between 1SG and 3SG.MASC which can be attributed to regular phonological rules. As for the “paradigmatic dissimilation”, in Anlaut paradigms of Gunnän-Gurage we do observe a blind sound change which leads to a more pronounced opposition between the reflexes of *ʔə- and *yə- (see Sections 5.1, 7.2). The same process in Argobba of Aliyu Amba and in the Inlaut paradigms of Gunnän-Gurage should be interpreted as a morphological readjustment, in the absence of an obligatory rule (see 8.3 and below in this section). Thus, the phonological and morphological processes are interrelated and both should be considered in this study.

82

Bulakh

The emergence of innovative morphemes in the Inlaut paradigms (and the preservation of the archaic prefixes in the Anlaut paradigms) has a natural explanation in terms of phonology and supplies still another argument in favour of Rubin’s hypothesis. The word-internal position is especially vulnerable as far as the potential merger between *ʔə- and *yə- is concerned. Both intervocalic ʔ and y are easily elided. As a result, the formal distinction between the morphemes *ʔə- and *yə- becomes realized as a vocalic opposition (as in Amh. s-ə-säbər ‘while I break’ vs. s-i-säbər ‘while he breaks’), less pronounced than the consonantal one. Moreover, the distinction is inevitably lost in the languages in which i functions as an allophone of ə, such as Harari (see Garad and Wagner 1998: 168) or Zay (Meyer 2005: 46). It is thus to be expected that the Inlaut paradigms will more readily participate in innovations that involve the prefixes *ʔə- and *yə-. The empiric data show that the “paradigmatic merger” seldom takes place word-internally in ES (for an explanation see Section 7.2).22 On the contrary, “paradigmatic dissimilation” restricted to Inlaut paradigms is quite well attested in SES. In fact, it is this process that brings about the split into Inlaut and Anlaut paradigms in a number of SES languages: East Gurage, Gafat, Čaha, Ennemor (see Table 4; on Argobba of Aliyu Amba see Sections 7.2, 8.3.2). In these languages, the Inlaut paradigm is characterized by the innovative 1SG morpheme, *-l-, which is absent from (or, in Argobba of Aliyu Amba, only secondarily introduced into) the Anlaut paradigms. Another indication of the tendency towards “paradigmatic dissimilation” in word-internal positions can be observed in a number of Gunnän-Gurage languages (see Sections 5.1, 8.1), in which the vowel -ä-, a regular reflex of *ʔə- in the Anlaut paradigm, is taken over into the Inlaut paradigm. This process was likely triggered by the tendency toward “paradigmatic dissimilation” between the 1SG and 3SG.MASC in the Inlaut forms. The appearance of -l- in the 1SG of the negative LPC, in opposition to the other slots of the paradigm, can also be manifestation of the tendency to reinforce the opposition between 1SG and 3SG.MASC word-internally (see Section 4.1).

22  The only reliable example of this kind comes from outside SES: in Tigrinya, where i and ə tend to become allophones of one phoneme, the relative forms zə-ḳättəl ‘he, who kills’ and zə-ḳättəl ‘I, who kill’ coincide in pronunciation (Leslau 1941: 62–63), although in the written language the opposition can be kept up (Voigt 2011: 1160). In SES, no such merger is registered even in those languages which, similar to Tigrinya, tend to lose the phonological opposition between i and ə.

Innovative 1SG Prefixes in South Ethio-Semitic Languages

83

7.2 Absence of Multiple Exponence As shown in 5.2.1, the innovative 1SG marker *-l- is a distinctive feature of Inlaut paradigms in quite a few SES languages. The respective Anlaut paradigms predominantly lack the innovative markers. As a rule, they preserve the old opposition and are only affected by the blind sound changes. Thus, the initial *ʔ- is dropped or loses its phonemic status throughout SES, and the initial ə- is changed to ä- in some languages. The latter shift does serve to reinforce the opposition between 1SG and 3SG.MASC (*ʔə- vs. *yə- > *ə- vs. *yə- > ä- vs. yə-), but this is obviously an accidental effect of a phonological development. Yet, the opposition between *ʔə- and *yə- is not omnipresent in the Anlaut paradigms of SES. As one can see in Table 6, the prefixes of 1SG and 3SG.MASC are not distinguished in Argobba of Ṭollaha and East Gurage languages. Neither the omission of initial *ʔə- and *yə- in Argobba nor the replacement of ʔ- by y- in East Gurage can be viewed as a result of a phonological rule. The languages share the general pattern—block syncretism of 1SG and 3SG.MASC/PL—but employ different strategies (spread of the 3SG.MASC/PL shape to 1SG or vice versa in East Gurage vs. omission of both prefixes in Argobba of Ṭollaha), which points toward a parallel development in East Gurage and in Argobba of Ṭollaha. The acoustic similarity between *ʔə- and *yə- (see 7.1) has apparently favored this development. The same languages do distinguish between 1SG and 3SG.MASC prefixes in the Inlaut paradigms, where the innovative *l-prefix emerges. What explanation can be offered for this distribution: “paradigmatic merger” in the Anlaut paradigm vs. “paradigmatic dissimilation” in the Inlaut paradigm? And why is the “paradigmatic merger” restricted to Argobba of Ṭollaha and East Gurage and does not occur elsewhere? It turns out that another important factor influences the evolution of the subject indexes in SES: “paradigmatic merger” and “paradigmatic dissimilation” correlate with the presence/absence of multiple exponence. “Multiple exponence” (or extended exponence) is a term used in contemporary linguistics to describe the employment of two or more morphemes within one word to express the same grammatical meaning (see, e.g., Caballero and Harris 2012 with further references). In Transversal SES, multiple exponence is especially conspicuous in the compound LPC (an Anlaut paradigm, which originates in an analytic construction with the locative verb *hallawa as an auxiliary): Amh. tə-säbr-iy-all-äš ‘youfem break’ (both -iy and -äš are the suffixes of 2SG.FEM). The association of the simple Anlaut forms in Transversal SES with the multiple exponence is less obvious, yet it plays an important role in the development of this subtype of LPC. Indeed, the compound LPC ousts the bare simple forms in the main clauses, whereas in subordinate clauses, the

84

Bulakh

verbal form of the simple LPC is usually preceded by a conjunction prefix.23 As a result, the simple Anlaut forms are usually restricted to analytic constructions. The construction with the auxiliary *nabara ‘to sit; to live’ is most widely used. Historically, this construction is a past-tense counterpart of the abovementioned construction with the auxiliary *hallawa. Originally, both analytic constructions (from now on referred to as *nabara- and *hallawa-constructions) were characterized by a double set of subject indexes—a feature which is sometimes preserved in the modern languages and which can be conveniently labeled as multiple exponence. The evolution of *nabara- and *hallawa-constructions in Transversal SES is analysed in a special investigation (Bulakh 2014). Here, it suffices to outline how the factors of multiple exponence and of acoustic similarity between *ʔəand *yə- interact in Argobba of Ṭollaha and in East Gurage. In these languages, the presence of multiple exponence in *nabara- and *hallawa-constructions obviously favours the “paradigmatic merger.” The block syncretism of the 1SG and 3SG.MASC prefixes of the main verb does not lead to homophony, since the two forms are distinguished by means of the subject indexes of the auxiliary elements: Zay yə-näḳəl-āhw ‘I take’ vs. yə-näḳəl-äl ‘he takes’; yə-näḳəl nārux ‘I used to take/I was taking’ vs. yə-näḳəl nār ‘he used to take/he was taking’.24 The tendency to avoid multiple exponence is not sufficient to account for the merger between *ʔə- and *yə-. One has to explain why only some of the prefixes grew syncretic, whereas the rest of the paradigm kept the redundant morphemes unchanged. This selective syncretism can be conveniently 23  An important exception is Harari, where the Inlaut paradigm is virtually absent, as observed above. 24  In Argobba of Ṭollaha, the “paradigmatic merger” is achieved by dropping both 1SG and 3SG.MASC/PL prefixes. The resulting “zero” elements are opposed to the overt prefixes of the rest of the paradigm. A similar phenomenon, namely, omission of *ʔə- and *yə(as well as 1PL *nə-), is observed in southern dialects of Tigrinya (for Tigrinya of MayČ�̣ äw, see Voigt 2004: 346; 2006; for Tigrinya of Wajerat, see Tsehaye Kiros Mengesha 2009: 51). The material presented in this section suggests that the merger/omission of *ʔə- and *yə- is usually favoured by double marking. Unfortunately, due to lack of comprehensive descriptions of the dialects in question, it remains unclear to what extent they employ the bare forms of the simple LPC. It appears that at least in Tigrinya of Wajerat, such forms are not common. They are certainly not used to express the present tense; in fact, in Tsehay Kiros Mengesha’s description, only the compound LPC is mentioned (as the exponent of present progressive; see Tsehay Kiros Mengesha 2009: 53). In Tigrinya of MayČ�̣ äw, the simple LPC as the exponent of present seems to be replaced by the compound LPC (see Voigt 2006: 894–895). Thus, it is not improbable that both dialects share with Argobba of Ṭollaha the features that favour the loss of the prefixes.

Innovative 1SG Prefixes in South Ethio-Semitic Languages

85

ascribed to the interaction between two factors: the multiple exponence and the acoustic similarity of the prefixes *ʔə- and *yə-. The combination of these two factors can lead towards “paradigmatic merger” between *ʔə- and *yə- and, simultaneously, to the constraint on the deletion of subject indexes of the auxiliaries. The latter constraint is best manifested in Selṭi, where the subject indexes of the auxiliary *nabara are consistently deleted except for the 1SG: iḳaba nār ‘he was painting’, iḳābu nār ‘they were painting’, but iḳaba nārku ‘I was painting’ (Gutt 1997: 922). Obviously, the merger between *ʔə- and *yə- preceded the deletion of subject indexing suffixes of auxiliary in this lect. In Amharic and, probably, in Argobba of Aliyu Amba (on which see below), in the absence of “paradigmatic merger” between *ʔə- and *yə-, the deletion of the subject indexes of the auxiliary *nabara is not constrained by any additional factor and affects all the slots: Amh. ə-säbər näbbär ‘I was breaking/I used to break’ vs. yə-säbər näbbär ‘he was breaking/he used to break’. The optional status of the forms with the deleted subject indexes in Amharic suggests that this innovation is a recent one. At the earlier stage, the multiple exponence must have been present both in *hallawa- and *nabara-constructions. Thus, the multiple exponence is a necessary, but not a sufficient condition for the “paradigmatic merger” between *ʔə- and *yə-. Absence of multiple exponence can lead to “paradigmatic dissimilation” in SES: the innovative *l-prefix appears in the LPC sub-paradigms which do not display multiple exponence. As one can see from Table 4, it is usually restricted to the Inlaut forms, which typically are not combined with auxiliaries. It is only in Argobba of Aliyu Amba that the Anlaut paradigm also employs the innovative 1SG marker dissimilar to the 3SG.MASC. In the latter variety, the innovative 1SG marker əl- appears in the Anlaut simple LPC, alongside the more archaic ä-. Now, in Argobba of Aliyu Amba multiple exponence in all probability is not typical of the constructions which employ the simple Anlaut LPC forms. A complete paradigm of the *nabara-construction is missing from Leslau 1997, but it is likely that multiple exponence is abandoned here in the way similar to that of Amharic, that is, by means of deletion of the subject indexes of the auxiliary (this is supported by the shortened form of the auxiliary in the construction yəhed əmbär ‘he used to go’ quoted in Leslau 1997: 49, instead of əmbära, with the overt subject index -a, see Leslau 1997: 36). The optional preservation of the more archaic ä- in the Anlaut simple LPC paradigm of Argobba of Aliyu Amba suggests that the innovative 1SG əl- in this sub-paradigm is more recent than in the Inlaut paradigm. In fact, it may have been triggered by the deletion of the subject indexing suffixes of the auxiliary əmbära (this deletion is likely a recent development, as in Amharic, see above).

86

Bulakh

Why does the absence of multiple exponence block “paradigmatic merger”? In the absence of multiple exponence, the prefixes *ʔə- and *yə- are the only markers of 1SG and 3SG.MASC, respectively, and the potential merger between them would lead to whole-word syncretism. It is therefore hardly surprizing that in the paradigms which lack multiple exponence (inter alia, the Inlaut paradigms),25 we are mostly confronted with “morphological dissimilation”: Zay *bi-l-hān > bīlhān ‘if I were’ vs. *bi-y-hān > bī-hān ‘if he were’. However, the “paradigmatic dissimilation” in the absence of multiple exponence is by no means ubiquitous in SES. Thus, no innovative prefixes were developed in Harari. This fact can be explained by the lack of Inlaut verbal forms in this language: word-initially, the prefixes *ʔə- and *yə- are better distinguished and hence, the “paradigmatic dissimilation” is not necessary (see Section 7.1). Yet, even in the Inlaut paradigm, absence of multiple exponence does not always lead to “paradigmatic dissimilation.” An obvious example is Amharic: *bə-əhon > bəhon ‘if I were’ vs. *bə-yə-hon > bihon ‘if he were’. Thus, absence of multiple exponence is a necessary, but not a sufficient condition for the “paradigmatic dissimilation” between 1SG and 3SG.MASC. 8

Innovative 1SG Markers of the LPC in SES and their Origins

In the present section, I will survey the material shapes and the sources of the innovative 1SG prefixes in LPC of SES. I will pay special attention to the problem of distinguishing between the innovations inherited from a common proto language and those to be attributed to parallel development (and perhaps areal diffusion). 8.1 Parallel Development or Genealogical Inheritance? As the survey in the Section 5.2 shows, the most widespread innovative feature is the *l-prefix in the Inlaut LPC paradigm. Shall we reconstruct the marker -l- (as the Inlaut allomorph of *ʔə-) for proto-SES and assume its eventual 25  At least, this is the most typical employment of the Inlaut paradigms. Of course, the Inlaut forms can in turn participate in a number of analytic constructions, which results in the double subject indexing: Zay *ti-l-bälə-n ʔalä-hʷ-u > tī-l-bälən ʔalo-hu ‘I am eating’ (Meyer 2005: 186). The constructions that exhibit such redundancy are more recent and less integrated into the verbal systems of the respective languages. Hence, preservation of the double marking in such cases does not contradict my claim. It can be expected that further grammaticalization of these constructions will lead to elimination of some of the redundant subject indexes.

Innovative 1SG Prefixes in South Ethio-Semitic Languages

87

disappearance in Amharic, Harari, and some of Gunnän-Gurage languages? Or shall we regard it as a parallel development in several languages? A common origin of the 1SG -l- of LPC is unviable: why should the assumed innovative prefixes of the affirmative LPC have been replaced by more archaic markers in quite a few languages, while they have been preserved in SPC and in the negative LPC? There are no reasons to suspect areal diffusion, either. The languages sharing this innovation—Argobba, East Gurage, Gafat, West Gurage—constitute no language continuum today, and there is not any evidence for the existence of such a continuum in the past, before the migrations that led to the present location of these lects. More likely, we are dealing with common drift, triggered by a combination of the factors discussed in the Section 7 (acoustic similarity between the 1SG and 3SG.MASC prefixes in the absence of multiple exponence). Yet, after several *l-prefixes have come into existence in subbranches of SES, genealogical inheritance or language contact could have played some role in their further spread. Firstly, it is unlikely that the innovative prefix emerged independently in all three languages of East Gurage: rather it is to be reconstructed for proto-East Gurage. The same consideration could be valid for West Gurage, but lack of information forces one to be more cautious: we only know that this feature occurs in Čaha and Ennemor. Secondly, language contact may have furthered the spread of the *l-prefix. On the modern language map, the geographic proximity between Selṭi and West Gurage within Gurage language area cannot be overlooked. Historically, language contact between Argobba and East Gurage languages (or even protoEast Gurage) is not improbable (see Section 2). We do not possess any information on the linguistic surroundings of Gafat, but in principle, language contact between Gafat and some of the West Gurage languages is not to be excluded. Thus, some of the languages in question may have borrowed the innovative feature from their neighbours. In Zay, Wolane and Argobba of Ṭollaha, the emergence of an innovative 1SG marker in the Inlaut paradigm is concomittant with the emergence of another innovative 1SG marker (yə- in Zay and Wolane and “zero” morpheme in Argobba of Ṭollaha) in the Anlaut paradigm, syncretic with the 3SG.MASC morpheme. East Gurage and Argobba of Ṭollaha share the same innovative pattern rather than the shape of the innovative elements. Hence, once again we are faced with a parallel development, probably under the influence of the same factors (acoustic similarity of the 1SG and 3SG.MASC prefixes + multiple exponence in the relevant forms).

88

Bulakh

Could language contact between Argobba and East Gurage have played some role in the spread of this feature? Hardly so. Such a suggestion would imply that either the hypothetical contact between Argobba and East Gurage involved the Argobba of Ṭollaha variety only or that Argobba of Aliyu Amba secondarily, under Amharic influence, reintroduced the opposition between 1SG and 3SG.MASC into the Anlaut paradigm. Neither of these hypotheses look convincing. Wolane and Zay share both the pattern and the actual shape of the innovative prefix yə- in their Anlaut paradigms. The third East Gurage language, Selṭi, has i- in the 1SG and 3SG.MASC/PL of the Anlaut paradigm. In Selṭi, i is the regular reflex of *ə, hence the 1SG prefix i- must be archaic. Its merger with the 3SG.MASC/PL prefix must have taken place after the split of protoEast Gurage, but cannot be fully independent of a similar merger (yet, with the reverse direction, from 3SG.MASC to 1SG) in Wolane and Zay. It is likely a common drift in closely related languages, a development in the Anlaut LPC paradigms inherited from proto-East Gurage. Areal diffusion could have been an additional factor at work: the three languages could have been in contact for some period of time after their split. The innovative 1SG prefix əl- of Argobba of Aliyu Amba is unique both in its shape (with the initial ə-) and in distribution (Anlaut paradigm) and thus is to be treated as an independent development having no exact parallels elsewhere in SES (or beyond). Finally, -ä- instead of -ə- in the 1SG forms of the Inlaut paradigms is registered in Kestane, Dobbi, and Muher. In these forms, the shift from ə to ä is not phonologically-conditioned and rather is to be treated as another instance of “paradigmatic dissimilation” (see Sections 5.1, 7.1). Since these languages constitute an areal language group rather than a genealogical unit (see Hetzron 1977: 22), the change can be an example of areal diffusion (for some other features shared by these three languages, see ibid.). However, it remains uncertain whether this feature is absent from other Gunnän-Gurage languages (see fn. 13). Innovative 1SG Markers as a Result of “Paradigmatic Merger” between 1SG and 3SG.MASC The “paradigmatic merger” is attested in Argobba of Ṭollaha and in East Gurage. In both cases, it involves the Anlaut paradigms only. What are the sources of the innovative 1SG morphemes? In Zay and Wolane, the directionality of the syncretism is obvious: the 3SG. MASC/PL prefix yə- is taken over into the 1SG forms. In Selṭi, the direction is apparently reverse: the 1SG prefix (*ʔə- > i-) has spread into the 3SG.MASC/PL. 8.2

Innovative 1SG Prefixes in South Ethio-Semitic Languages

89

In Argobba of Ṭollaha, we are dealing with non-directional syncretism: the omission of the prefixes *ʔə- and *yə- leads to the syncretism of morphologically unmarked forms (see Baerman, Brown, and Corbett 2005: 25–26). Innovative 1SG Markers as a Result of “Paradigmatic Dissimilation” between 1SG and 3SG.MASC The “paradigmatic dissimilation” in SES always changes the 1SG prefix (rather than 3SG.MASC/PL, as, for instance, is the case in Tigre). However, the shape of the new morpheme varies from language to language. What are the material sources of these divergent morphemes? In some cases, the origin is the archaic *ʔə-. Thus, (-)ä- in the Inlaut paradigms of some of Gunnän-Gurage languages (Kestane, Dobbi, Muher) and in the simple Anlaut LPC of Argobba of Aliyu Amba ultimatedly go back to *ʔə- (see Section 5.1). In both cases, the phonological sound change was reinterpreted as a morphological innovation, aiming to reinforce the opposition between 1SG and 3SG.MASC. The origin of the innovative 1SG *-l- in the Inlaut paradigms of a number of SES languages and of the 1SG əl- in the Anlaut paradigm of Argobba of Aliyu Amba deserves special discussion. 8.3

8.3.1 The Origin of the *l-prefixes in the Inlaut Paradigms of LPC As shown in 8.1, the *l-prefixes of LPC emerged independently in several subgroups of SES. How could it happen that a parallel development resulted in several formally identical innovative morphemes? A plausible explanation is that the languages involved shared not only the structural “inner dynamic” but the material source of the innovative morpheme as well. However, it was only after the split of proto-SES, and only in some of the daughter languages, that this potential was realized. What is this common source for the innovative *l-prefixes in the LPC of SES? Were they taken over from the SPC (as is claimed in Goldenberg 1977: 482), in other words, do they ultimately go back to the archaic “injunctive” *la-? At first glance, this hypothesis is supported by the seemingly parallel developments in Tigre and Eastern Aramaic (on which see Rubin 2007, Loesov 2012, Bulakh 2013). In these languages, the path of development may be reconstructed in the following way: “injunctive” *la- > subject index of the jussive paradigm (=SPC in ES) > subject index of the jussive and indicative non-past paradigms (=SPC and LPC in ES). However, as argued in Bulakh 2013, both in Tigre (in the derived stems) and Eastern Aramaic this expansion was favoured by the identity of bases of the

90

Bulakh

jussive and indicative non-past paradigms (thus, the innovation led towards total merger between the two paradigms). In SES, there are no such favourable conditions: the LPC innovative prefix appears in all stems, although in most of them the LPC bases are distinct from the SPC bases. Besides, in the proto-SES, an element *l in the 1SG appeared not only in SPC, but in the negative LPC as well (see Section 4.1). As argued in 4.1, the *l in SPC and in the negative LPC have different origins and their formal identity is a pure accident. Still, this formal identity poses a serious problem for our investigation. Indeed, one has to choose between two potential sources for the LPC 1SG prefix *lə-: the jussive prefix *lə- and the element *l of the negative paradigm. Admittebly, borrowing of the LPC prefix from the negative paradigm is an unprecedented development in Semitic. Yet, this hypothesis finds support in the fact that the innovative *l-prefixes in the affirmative LPC are mostly restricted to Inlaut paradigms. Indeed, the reanalysis of the cumulative *ʔalNEG.1SG into *ʔa-l- NEG-1SG mentioned in the Section 4.1 makes it possible to treat the negative paradigm as a subtype of Inlaut paradigms. Within this approach, one does not need to explain the expansion of a subject index from a negative into an affirmative paradigm. Rather, we can postulate the expansion of a subject index from one Inlaut context into all Inlaut contexts—a generalization which led to the emergence of the Inlaut paradigm as a morphologically distinct subtype of LPC. In fact, a third possibility is not to be excluded, namely, that the *l-prefixes in the affirmative LPC are not directly borrowed from any other subtype of PC, but rather are due to the influence both from SPC and the negative paradigm. Since both paradigms had the formally identical elements *l in the 1SG, the salience of these elements increased, which facilitated their expansion into another subtype of PC. 8.3.2 Argobba of Aliyu Amba: 1SG Prefix əl- in the Anlaut Paradigm In Argobba of Aliyu Amba the Anlaut simple paradigm has an innovative 1SG marker əl- (at free variation with the archaic ä-): əl-säkər/ä-säkər ‘I get drunk’ (simple LPC) vs. ə-sädb-älluh ‘I offend’ (compound LPC). This 1SG prefix, with the initial vowel, does not find immediate parallels elsewhere in ES. Moreover, it is the only l-prefix of LPC in SES which appears in the Anlaut paradigm. Two circumstances must be taken into account. Firstly, apart from əl- in the Anlaut paradigm, this lect is also characterized by the innovative 1SG -l- in the Inlaut paradigm, and it can be argued that the emergence of the Inlaut -l- preceded the emergence of əl- (see Section 7.2).

Innovative 1SG Prefixes in South Ethio-Semitic Languages

91

Secondly, the prefix əl- is not restricted to the 1SG. In Argobba of Aliyu Amba, most subtypes of PC are characterized by the syncretism of the 1SG and 1PL prefixes (see Section 6). In SPC and in the Inlaut LPC of Argobba of Aliyu Amba, the direction of the syncretism is from 1SG to 1PL (the 1SG prefix -l- is extended to 1PL): SPC lə-sdäb ‘let me offend’ vs. lə-sdäb-ən ‘let us offend’, Inlaut LPC lə-l-säkər ‘so that I get drunk’ vs. lə-l-säkr-ən ‘so that we get drunk’. The same pattern is found in the compound LPC forms, with the prefix ə- taken over into the 1PL forms: ə-sädbənällən ‘we offend’. The syncretic paradigm of the compound LPC coexists with the archaic one, which employs ə- (< *ʔə-) for the 1SG and əl-/ən- (< *ʔən-) for the 1PL: ə-sädb-älluh ‘I offend’ vs. əl-sädb-ällən/ən-sädb-ällən. Alternation between l and n is found in Argobba (see Leslau 1997: 1) as well as elsewhere in SES (see Podolsky 1991: 48 for Amharic). In Argobba, it is not a regular phonological process. In the 1PL of the compound LPC, the alternation could have been triggered by the existence of *l-prefixes in the 1PL forms of SPC and the Inlaut LPC (where they were taken over from the 1SG forms). The 1SG/1PL prefix əl- of the simple Anlaut LPC must be related to the 1PL prefix əl-/ən-. This means that the direction of the syncretism in this sub-paradigm is from 1PL to 1SG. The development in Argobba of Aliyu Amba is then to be reconstructed as follows. At first, the syncretic 1SG/PL *l-prefixes emerged in SPC and the Inlaut LPC (their emergence must have been more or less contemporary to the introduction of the innovative 1PL *-na): SPC *lə-sdäb ‘let me offend’ vs. *nə-sdäb ‘let us offend’ > lə-sdäb vs. lə-sdäb-ən, Inlaut LPC *lə-l-säkər ‘so that I get drunk’ vs. *lə-n-säkər ‘so that we get drunk’ > lə-l-säkər vs. lə-l-säkr-ən. The second step is the shift from ən- to əl- in the 1PL of the Anlaut LPC, where the original ən- came to be perceived as the phonological variant of 1PL -l(ə)of the rest of the PC paradigms: simple Anlaut LPC *ən-sädəb > *əl-sädəb/*ən-sädəb, compound LPC *ən-sädb-ällən > əl-sädb-ällən/ən-sädb-ällən ‘we offend’. Finally, the 1PL əl- is extended into the 1SG of the simple Anlaut LPC (simultaneously, the 1PL suffix is introduced):

92

Bulakh

*ə-sädəb ‘I offend’ vs. *əl-sädəb/*ən-sädəb ‘we offend’ > əl-sädəb vs. əl-sädb-ən. Of the two coexisting allomorphs ən-/əl-, only the l-form was chosen, undoubtedly due to the influence of l-prefixes of the Inlaut LPC and of SPC. Why did the innovation affect the simple LPC only, while the compound LPC continued to employ the archaic prefix? The answer is that it was triggered by the elimination of multiple exponence in the analytic constructions which involve simple Anlaut LPC. The innovation in this position is thus a manifestation of the same process as that of the Inlaut paradigms of Argobba of Aliyu Amba and other SES languages (“paradigmatic dissimilation” in the absence of multiple exponence, see Section 7.2). The shape of the innovative prefix was suggested by the syncretism of 1SG and 1PL in other sub-paradigms of PC: the innovative prefix was taken over from the 1PL. 9 Conclusions In this article, I have surveyed various innovative 1SG prefixes in the PC of SES, have described two basic patterns of evolution leading to their emergence, have considered various factors that favour this or that type of development and have reconstructed the origin of each of the innovative 1SG prefixes. Within SES, two major types of “common drifts” leading to innovations in 1SG of PC can be distinguished. One of them, labeled here as “paradigmatic merger”, is the introduction of a 1SG prefix homophonous with that of 3SG.MASC/PL. The other, labeled as “paradigmatic dissimilation”, is manifested in the replacement of 1SG marker *ʔə- by an innovative prefix, conspicuously different from the 3SG.MASC/PL prefix *yə-. The impulse for both developments is to be sought in the acoustic similarity between the archaic 1SG prefix *ʔə- and the 3SG.MASC/PL prefix *yə-. This or that type of evolution is chosen in accordance with the morphological distinction principle, on the one hand, and the economy principle, on the other hand (see Kibrik 1992: 188; for an English summary see Testelets 2008: 318): the “paradigmatic merger” is favoured in the situation of multiple exponence, whereas the “paradigmatic dissimilation” usually takes place if the prefixes involved are the only exponents of subject indexing. The role of language contact in the spread of these developments is apparently not very high. The languages in question exhibit a great deal of diversity as far as the innovative morphemes are concerned, and the established isoglosses in most cases include languages remote from each other both

Innovative 1SG Prefixes in South Ethio-Semitic Languages

93

genetically and geographically. Yet, in some cases, areal diffusion is not to be excluded (see Section 8.1). The investigation confirms Huehnergard’s claim that the “injunctive *la-” of proto-ES (going back to “asseverative *la-” of proto-Semitic) is the origin of the innovative 1SG prefix of SPC in Amharic as, indeed, elsewhere in SES and in proto-SES. However, the situation is less clear in respect of the innovative 1SG prefixes in various subtypes of LPC. There are no solid arguments in favor of their etymological link to the corresponding SPC markers and thus, to the “injunctive *la-”. An alternative and equally plausible reconstruction implies that they were shaped under the influence of two subtypes of PC, namely, of SPC and of the negative LPC. In this case, the emergence of the innovative 1SG *l-prefixes in LPC of SES is not parallel to the expansion of the jussive *l-prefixes into the indicative paradigm in Eastern Aramaic and in Tigre. Bibliography Aikhenvald, Alexandra Y. and Robert M. W. Dixon. 2006. “Introduction,” in Alexandra Y. Aikhenvald and Robert M. W. Dixon (eds.), Areal diffusion and genetic inheritance: problems in comparative linguistics. Oxford. 1–26. Baerman, Matthew, Dunstan Brown, and Greville G. Corbett. 2005. The Syntaxmorphology Interface: A Study Of Syncretism. Cambridge. Braukämper, Ulrich. 2004. Islamic History and Culture in Southern Ethiopia. Collected Essays. Münster. Bulakh, Maria. 2012. “Negative markers *ʔay-, *ʔi- and *ʔal- in Ethio-Semitic,” Babel und Bibel 6: 385–420. ———. 2013. “Origin and Evolution of the Personal Marker l(ə)- in the Prefix Conjugation of Tigre,” Babel und Bibel 7: 375–388. ———. 2014. “Multiple exponence in the long prefix conjugation paradigms of Transversal South Ethio-Semitic languages,” in Ronny Meyer, Yvonne Treis, and Azeb Amha (eds.), Explorations in Ethiopian Linguistics: Complex Predicates, Finiteness and Interrogativity (Abhandlungen für die Kunde des Morgenlandes). Wiesbaden. 147–176. Bulakh, Maria and Leonid Kogan. 2010. “The Genealogical Position of Tigre and the Problem of North Ethio-Semitic Unity,” ZDMG 160: 273–302. ———. 2013. “More on Genealogical Classification of Ethiopian Semitic,” Babel und Bibel 7: 599–608. Caballero, Gabriela and Alice C. Harris. 2012. “A working typology of Multiple Exponence,” in Ferenc Kiefer, Mária Ladányi, and Péter Siptár (eds.), Current issues in morphological theory: (ir)regularity, analogy and frequency: selected papers from

94

Bulakh

the 14th International Morphology Meeting, Budapest, 13–16 May 2010. Amsterdam – Philadelphia. 163–188. Cerulli, Enrico. 1936. Studi Etiopici, Vol. 1. La lingua e la storia di Harar. Rome. Crass, Joachim and Ronny Meyer. 2011. “Ethiosemitic-Cushitic language contact,” in Stefan Weninger, in collaboration with Geoffrey Khan, Michael P. Streck and Janet C. E. Watson (ed.), Semitic Languages: An International Handbook. Berlin. 1266–1275. Eyassu Nega. 2003. The Morphology of Endegañ, M.A. thesis, Addis Ababa University. Garad, Abdurahman and Ewald Wagner. 1998. Harari-Studien: Texte mit Übersetzung, grammatischen Skizzen und Glossar. Wiesbaden. Goldenberg, Gideon. 1977. “The Semitic Languages of Ethiopia,” BSOAS 40: 461–507. ———. 1968. “Kəstanəňňa: Studies in a Northern Gurage Language of Christians,” Orientalia Suecana 17: 61–102. Gutt, Ernst-August. 1997. “Concise Grammar of Silt’e,” in Eeva H. M. Gutt and Hussein Mohammed, Silt’e-Amharic-English Dictionary (with Concise Grammar by ErnstAugust Gutt). Addis Ababa. 895–957. Hetzron, Robert. 1968. “Main Verb-Markers in Northern Gurage,” Africa 38: 156–72. ———. 1972. Ethiopian Semitic. Studies in classification. Manchester. ———. 1976. “Two principles of genetic reconstruction,” Lingua 38: 89–108. ———. 1977. The Gunnän-Gurage Languages. Naples. Huehnergard, John. 1983. “Asseverative *la and hypothetical *lu/law in Semitic,” JAOS 103: 569–593. Kibrik, Aleksandr. 1992. “Podlezhaschee i problema universal’noy modeli yazyka,” Aleksandr Kibrik, Ocherki po obschim i prikladnym voprosam yazykoznanija. Moscow. 179–197. [Kibrik, Aleksandr, “The subject and the problem of the universal model of language,” in Alexandr Kibrik, The essays on the general and applied issues of language studies. Moscow. 179–197.] Leslau, Wolf. 1941. Documents tigrigna (ethiopien septentrional). Grammaire et textes. Paris. ———. 1956. Étude descriptive et comparative du Gafat (Éthiopien méridional). Paris. ———. 1958. The Verb in Harari (South Ethiopic). Berkeley – Los Angeles. ———. 1968. Ethiopians Speak. Studies in Cultural Background, Vol. 3. Soddo. Berkeley – Los Angeles. ———. 1971. “The verb forms of the Gurage dialect of Endegeñ,” in Afrikanische Sprachen und Kulturen—Ein Querschnitt (Hamburger Beiträge zur Afrika-Kunde 14). Hamburg. 180–187. (repr. in Wolf Leslau. 1992. Gurage studies: Collected Articles. Wiesbaden. 466–473.) ———. 1979. Etymological Dictionary of Gurage (Ethiopic), Vol. 3. Etymological section. Wiesbaden.

Innovative 1SG Prefixes in South Ethio-Semitic Languages

95

———. 1981. Ethiopians speak: Studies in Cultural Background, Part IV. Muher. Wiesbaden. ———. 1983. Ethiopians Speak: Studies in Cultural Background, Part V. Chaha – Ennemor. Wiesbaden. ———. 1992. Gurage studies: Collected Articles. Wiesbaden. ———. 1995. Reference Grammar of Amharic. Wiesbaden. ———. 1997. Ethiopic Documents: Argobba. Grammar and Dictionary. Wiesbaden. ———. 2004. The Verb in Mäsqan as Compared with Other Gurage Dialects. Wiesbaden. Loesov, Sergey. 2012. “A New Attempt at Reconstructing Proto-Aramaic. I,” Babel und Bibel 6: 421–456. Meyer, Ronny. 2001 [2002]. “Wäy läne, Wäy lähagäre! Oral Traditions of the Wolane’s Past,” Hristianskiy Vostok 3 (IX): 457–472. ———. 2005. Das Zay. Deskriptive Grammatik einer Ostguragesprache (Äthiosemitisch). Köln. ———. 2006. Wolane. Descriptive Grammar of an East Gurage Language (Ethiosemitic). Köln. ———. 2007. “Self-agentive motion-verbs in Muher,” APAL 5: 1–18. Podolsky, Baruch. 1991. Historical Phonetics of Amharic. Tel-Aviv. Rubin, Aaron D. 2007. “On the Third Person Preformative l-/n- in Aramaic, and an Ethiopic Parallel,” ANES 44: 1–28. Simeone-Senelle, Marie-Claude. 1997. “The Modern South Arabian Languages,” in Robert Hetzron (ed.), The Semitic Languages. London. 378–423. Stump, Gregory Thomas. 2001. Inflectional Morphology: A Theory of Paradigm Structure. Cambridge. Testelets, Yakov G. 2008. “Russian works on linguistic typology in the 1960–1990s,” in Martin Haspelmath, (ed.), Language typology and language universals: an international handbook, Vol. 1. Berlin – New York. 306–323. Tropper, Josef. 1995. “Die Phönizisch-Punischen Kausativbildungen im Lichte von Präjotierung und Dejotierung im Semitischen,” ZDMG 145: 28–37. Tsehaye Kiros Mengesha. 2009. A comparison of Wajerat Tigrigna vs. Standard Tigrigna. M.A. Thesis, Addis Abeba University. Voigt, Rainer M. 2004. “Der Wegfall der Personalelemente in den Präfixkonjugationen des Semitischen,” in Verena Böll, Denis Nosnitsin, Thomas Rave, Wolbert Smidt, and Evgenia Sokolinskaia (eds.), Studia Aethiopica In Honour of Siegbert Uhlig on the Occasion of his 65th Birthday. Wiesbaden. 345–354. ———. 2006. “Südtigrinische Dialekte: Das einfache und zusammengesetzte Präsens im Dialekt von May-Č�̣ äw (Tigray),” in Siegbert Uhlig (ed.), Proceedings of the XVth International Conference of Ethiopian Studies, Hamburg July 20–25 2003. Wiesbaden. 893–898.

96

Bulakh

———. 2011. “Tigrinya,” in Stefan Weninger, in collaboration with Geoffrey Khan, Michael P. Streck and Janet C. E. Watson (ed.), Semitic Languages: An International Handbook. Berlin. 1153–1169. Wagner, Ewald. 1968. “Drei Miszellen zum südostsemitischen Verbum,” Wissen­ schaftliche Zeitschrift Universität Halle 1: 207–215. ———. 1983. Harari-Texte in Arabischer Schrift mit Übersetzung und Kommentar. Wiesbaden. Wetter, Andreas. 2010. Das Argobba: Eine deskriptive Grammatik der Varietät von Shonke und T’ollaha (Zentraläthiopien). Köln.

Egyptianizing Features in Phoenician and Punic Inscriptions from Egypt David Calabro From the Persian period through the Ptolemaic period, as had been the case in earlier eras, a great many Semitic-speaking mercenaries, merchants, pilgrims, and refugees settled in Egypt or at least traversed its soil (see, inter alia, Bresciani 1987; Chiera 1987; Vittmann 2003: 44–83). Among the marks left by these Semitic-speaking people are inscriptions. The two largest contingents of the Semitic-speaking immigrant population during these periods were speakers of Aramaic and speakers of Phoenician. Because of the large quantity of recovered Aramaic texts from Egypt, we know a great deal about contact between Aramaic and Egyptian from the Persian to the Ptolemaic periods. However, the Phoenician-speaking contingent has yet to be fully integrated into our picture of language contact during this time. My purpose in the present study is to point out some aspects of the Phoenician corpus from Egypt that show apparent influence from Egyptian and that seem distinct, both from Phoenician as attested elsewhere and from Egyptian Aramaic. I will also be concerned with explaining these features in terms of the depth of interaction with Egyptian, or at least assessing whether such an explanation is possible. Phoenician and Punic inscriptions from Egypt that are currently available for study include a collection of about twenty Phoenician ostraca (mixed with five Aramaic ostraca) from north Saqqara; a couple of letters on papyrus, one from Saqqara and the other unprovenanced; about seventy Phoenician graffiti (mixed with sixteen Aramaic graffiti) on the walls of the temple of Seti I at Abydos; eleven urn inscriptions from Thebes; about sixty Phoenician jar inscriptions (mixed with thirteen Aramaic jar inscriptions) from the Persianperiod fortress at Elephantine; eight graffiti on the two colossi at the entrance to the temple at Abu Simbel; assorted ostraca from Tell el-Maskhuta and Abusir; and various inscribed statuettes and other objects from all over Egypt, mostly without a definite provenance.1 Based on this corpus, I discuss here six features that are indicative of language contact with Egyptian: (1) transliteration of Egyptian consonants, (2) sound shifts similar to those that take place in Egyptian, (3) “loanwords” 1  References to the publications of these inscriptions are found in the discussion below, where some of the inscriptions are quoted.

© koninklijke brill nv, leiden, ���5 | doi ��.��63/9789004300156_007

98

Calabro

(broadly conceived as Egyptian words other than proper nouns appearing in Phoenician script), (4) a possible calque, (5) Egyptian-like personal names, and (6) verbal formulae resembling those used in Egyptian. Where possible, I draw conclusions about what these features mean for our understanding of contact between Phoenician, Egyptian, and Aramaic communities in Egypt during the first millennium BCE. 1

Conventions for Transliterating Egyptian Consonants

The name ṣḥpmw occurs four times in jar inscriptions from the Persian-period fortress at Elephantine (Lidzbarski 1912), a corpus which also includes many Aramaic inscriptions. 1.1. lwʾ br ṣḥpmw LWʾ son2 of Ṣaḥpimaw 1.2. ṣḥpmw / bn ḥr Ṣaḥpimaw / son of Ḥori 1.3. ṣḥpmw Ṣaḥpimaw 1.4. lṣḥpmw bn ṣḥʾ (Belonging) to Ṣaḥpimaw son of ṢḤʾ This name is to be explained as the Egyptian ṯꜣy-ḥp-ı͗m=w ‘May the Apis take them’ (Ranke 1935: 388, no. 2; Spiegelberg in Lidzbarski 1912: 7 n. 2).3 Likewise, in the same corpus, the name ṣknsmw ‘May Khonsu take them’ (Ranke 1935: 388, no. 6) has the same verbal element: 1.5. ṣknsmw bn pṭʾs Ṣakonsimaw son of Peṭʾisi

2  The word br 'son' is Aramaic, but the ductus of the inscription is Phoenician. 3  Ranke cites Akkadian syllabic evidence for the vocalization.

Egyptianizing Features in Phoenician and Punic Inscriptions

99

The spelling with ṣ for the Egyptian palatal plosive ṯ = /c/ (Loprieno 1995: 33)4 is characteristically Phoenician, as Spiegelberg points out, the Aramaic spelling having š instead (Spiegelberg in Lidzbarski 1912: 7 n. 2). The spelling ḥrpkrṭ is the Phoenician transliteration of the Egyptian divine name ḥr-pꜣ-ẖrd (‘Horus the child’, Harpocrates) and occurs on two unprovenanced (but presumably Egyptian-originating) statuettes of this god (Gibson 1982: 142–43; Donner and Röllig 2002: no. 52). The first of these, which is less complicated, runs as follows: 1.6. ḥrpkrṭ ytn ḥym l ʿbdy lʿbdʾšmn bn ʿštrtytn bn mgn bn ḥnts bn pṭ bn ṭṭ bn pšm[. .]y bšt [. . .] May Harpocrates give life to his servant, to ʿAbdiʾešmun son of ʿAštartyaton son of Magon son of ḤNTS  son of PṬ son of ṬṬ son of PŠM[. .]Y BŠT [. . .] The pronunciation of the consonant ẖ in earlier Egyptian is believed to have been either a palatalized voiceless velar fricative /xj/ (Allen 2000: 16) or a voiceless palatal fricative /ç/ (Loprieno 1995: 33, 35). By the Demotic period, this consonant seems to have merged with /x/ in the dialects known to us (Maspero 1917: 46–50; Vycichl 1983: 282). Aramaic transliterations of the name Harpocrates differ from the Phoenician in rendering this consonant by a ḥ (Degen 1970: 220), as with the use of ṣ vs. š in the names ṣḥpmw and ṣknsmw in the Elephantine jar inscriptions. The difference in Aramaic and Phoenician transliterations of these names is summarized in Table 1.

4  A different account of this Egyptian letter’s ancient pronunciation, namely as /ɵ/, is given by Maspero (1917: 18–22), but this is not generally accepted. In many words, ṯ merged with t by the New Kingdom, but in other cases, including ṯꜣı͗ ‘take’, the pronunciation as a palatal plosive was retained through the Coptic period (see Maspero 1917: 18–22; Vycichl 1983: 322–335).

100

Calabro

TABLE 1  Phoenician vs. Aramaic Transliteration of Egyptian Names Egyptian

Phoenician

Aramaic

Translation

ṯꜣy-ḥp-ı͗m=w

ṣḥpmw

šḥpymw

ṯꜣy-ḫnsw-ı͗m=w

ṣknsmw

---

ḥr-pꜣ-ẖrd

ḥrpkrṭ

ḥrpḥrṭ (in PN pṭḥrpḥrṭ)

‘May the Apis take them’, a PN ‘Khonsu has taken them’, a PN ‘Horus the child’, a DN

These examples show that Phoenician and Aramaic scribes consistently followed two separate orthographic conventions for transliterating Egyptian consonants (Table 2). It might be suggested that in the case of the transliterations of ṯ, the different conventions are due to differences in the phonetic inventories of Phoenician and Aramaic. Specifically, Egyptian Aramaic may have lost the affricate pronunciation of ṣ as /tsʾ/ by this period, while Phoenician may have retained it; thus an Aramaic speaker would have had to resort to š as the closest approximation to Egyptian ṯ in his native phonetic inventory. However, it is also possible that there was no phonetic difference in ṣ in the two languages, but that the different scribal conventions simply reflect different aspects of the target phoneme: some members of the Phoenician-speaking community heard Egyptian ṯ as similar to ṣ in the plosive release of the consonant (which sounds similar to affrication), while some members of the Aramaic-speaking community heard a similarity to š in the aspect of the palatal place of articulation. Steiner (1982: 57–59), based on other assorted evidence from transliterations of and into Aramaic, is of the opinion that Aramaic speakers in Persian-period (and later) Egypt pronounced ṣ as an affricate. TABLE 2  Conventions of Transliteration Egyptian

Phoenician

Aramaic

ṯ = /c/ ẖ = /ç/ or /xj/

ṣ = /tsʾ/ k

š = /š/ or /s/ ḥ

Egyptianizing Features in Phoenician and Punic Inscriptions

2

101

Sound Shifts

Ten graffiti from the temple of Seti I at Abydos (Derenbourg 1885; ClermontGanneau 1905; Lidzbarski 1915: 93–116, pls. 7–11; Donner and Röllig 2002: no. 49) have /l/ for /n/ or vice versa. The four examples showing the shift of /l/ to /n/ are the following: 2.1. ʾnk pʿlʾbst bn ṣdytn bn grṣd hṣry yšb [ʿ/d]ky bʾn mṣrm bpṭrt ʿbdmnqrt hʾn[y] ʾnk p



I am Paʿala‌ʾubast son of Ṣadyaton son of Gerṣad the Tyrian, inhabi­tant of [. . .] in Heliopolis of Egypt, in the freedom of ʿAbdimelqart the Heliopolitan. I [. . .]

2.2. ʾnk bʿlyḥn bn mnqrtḥny I am Baʿlyaḥon son of Melqartḥanniya. 2.3. ʾnpqd / nʿʿ ym ʾElpaqod / the diurnal temple musician 2.4. ngrb(ʿl) For Gerba(ʿl) Those substituting in the opposite direction (/l/ for original /n/) are six in number. In addition to the examples from the temple of Seti I at Abydos, a pilgrim graffito on a colossus at the entrance to the temple at Abu Simbel (Friedrich 1964) shows the same phenomenon (2.11). 2.5. ʾlk yḥlbʿl bn bʿlpls I am Yiḥannibaʿl son of Baʿlpilles. 2.6. ʾlk ʿbdṣpl bl ʿbdmlkt bl mlqrtʿm- / s I am ʿAbdiṣapon son of ʿAbdimalkat son of Melqartʿamo- / s 2.7. ʾlk bnb[ʿl] I am Banoba[ʿl].

102

Calabro

2.8. ʾl! ʿbdšmš I am ʿAbdišamaš. 2.9. ʾbʿl ʾbl mtl ʾAbbaʿl son of Matton 2.10. ʾl! hl ʾlm bd ǀǀ [.]t ʾyt gm ǀǀ [. . .] I was here: [. . .]. 2.11. ʾlk kšy I am Kušiya. Since the Phoenician graphemes for l and n are distinguished only by the length of the descender, one may think that this is an error in the reading of the inscription, or at least a purely graphic phenomenon. However, there is no empirical basis for either of these explanations. Due to the deterioration of the stone, Lidzbarski’s hand copies are the best images of these graffiti. According to him, the reading as n rather than l is “ganz sicher” (Lidzbarski 1915: 99). Some of the graffiti show an n and a clearly distinct l (with a much shorter descender) in the same inscription. The environments in which the variants occur are fairly consistent: /l/ > /n/ in the divine name Melqart, and /n/ > /l/ in the pronoun ʾnk ‘I’ and in variants of the word bn ‘son’. The /l/ > /n/ shift also occurs once with the preposition l ‘to, for’, which is interesting, since the Egyptian preposition having the same meaning of ‘to, for’ is n. Given these facts, it seems likely to me that this is a phonetic and not just a graphic phenomenon. Lidzbarski has attempted to place this two-sided phenomenon in a broader Phoenician context (Lidzbarski 1915: 99–100). However, the attempt is largely

FIGURE 3

Hand copies of Phoenician graffiti from the temple of Seti I at Abydos (see text citations 2.1 and 2.2), taken from Lidzbarski 1915: plate 10.

Egyptianizing Features in Phoenician and Punic Inscriptions

103

unsuccessful, and it only highlights how rare this phenomenon is outside of the Phoenician inscriptions in Egypt. In terms of relative frequency, it is entirely unique compared to the rest of Phoenician as far as I am aware. Therefore, we have here a feature that distinguishes “Egyptian Phoenician” from Phoenician as attested in inscriptions elsewhere. I would suggest that what we have here, in the case of the /l/ > /n/ shift, is phonetic influence from Egyptian. Some (though not all) dialects of ancient Egyptian lost the phoneme /l/, which merged with /n/, /r/, or /ʾ/. Further, the fact that a few Phoenician inscriptions with no obvious tie to Egypt also exhibit a shift between /l/ and /n/, as cited by Lidzbarski, could mean that a native (though rare) phenomenon facilitated the shift for Phoenician speakers in Egypt. These speakers could have selected an option known from their native repertoire and regularized it in certain contexts (like the divine name mnqrt) so that the phenomenon appears much more frequently here than elsewhere. For the opposite shift, that of /n/ > /l/, we may perhaps invoke hypercorrection. Phoenician speakers, realizing that /menqart/ was not strictly correct for the divine name Melqart, began pronouncing not only this name, but also other words (particularly the pronoun ʾnk), with /l/. The evidence allows this interpretation: 2.1 and 2.2 have ʾnk and mnqrt, while 2.6 has ʾlk and mlqrtʿm; none of the inscriptions have ʾlk and mnqrt. In the graffiti on the colossi at Abu Simbel, there are three occurrences of m for *n in the word bn ‘son’ (2.12, 2.13, and 2.14). A similar phenomenon may be at work in the sequence ʾmk (perhaps for ʾnk) in a short tomb graffito (CIS, no. 98) near Tell el-Amarna (2.15). 2.12. grhkl bm ḥlm / ʾš ʿl š[d] kš [r]lḥms Gerhekal son of Ḥalom / who is over the field of Kush (?) [. . .] 2.13. kšy bm ʿbdpʿm / ʾš ʿl šd kš rlḥms Kušiya son of ʿAbdipaʿm / who is over the field of Kush (?) [. . .] 2.14. kšy / bm ʿbdpʿm Kušiya / son of ʿAbdipaʿm 2.15. [. . .] ʾmk [. . .] [. . .] I [. . .] In Demotic, and even beginning in Late Egyptian, /n/ often appears for etymological */m/ (Spiegelberg 1925: 7, 123; Erman 1933: 23). Synchronically, this

104

Calabro

might have been understood as arbitrary variation between the two phonemes in some morphological environments.5 Therefore, this could, like the interchange of n and l, point to phonological influence from Egyptian. Again, while shifts between m and n are common generally in the Semitic languages, such a shift in the environments of the pronoun ʾnk and the word bn is rare and distinguishes these inscriptions in the “Egyptian Phoenician” corpus from the rest of Phoenician. 3

Loanwords or Egyptian Words in Alphabetic Script

Detecting loanwords in an ancient text is a notoriously problematic enterprise. Several sequences of letters in the Phoenician corpus from Egypt could be loanwords, since, if we are reading the letters correctly, they do not readily seem to make sense in terms of what we know of Phoenician. However, there is a lot of room for uncertainty, given that the readings of the letters are sometimes in doubt, the contexts are fragmentary or nonexistent, and the overall Phoenician corpus is small. Nevertheless, there are a couple of cases that are higher on the continuum of probability. A jar inscription from Elephantine (Lidzbarski 1912: 17, no. 50) has an inscription which does not readily make sense in Phoenician: 3.1. npr q Good: (One) q(ab) (?) Lidzbarski analyzes this sequence only as “anscheinend ägyptisch.” Along these lines, one interpretation is that it is the Egyptian word nfr ‘good’ followed by an abbreviation for the amount contained in the jar: one qab (about 2 liters). An alternative interpretation could be to parse it as n pr Q ‘(Belonging) to the estate of Q(. . .)’. In either case, this provides an example of an Egyptian word other than a proper name being written in alphabetic script. In an unprovenanced Phoenician letter on papyrus (Aimé-Giron 1939; Donner and Röllig 2002: no. 51), the sequence hrnpʾrt or hrnkʾrt occurs.

5  “Sehr häufig ist altes m im Demotischen (wie ja vielfach schon im Neuäg.) in n übergegangen, vor allem in der Praeposition n” (Spiegelberg 1925: 7). The original /m/ is retained with pronoun suffixes and in compound prepositions: Middle Egyptian ı͗m=f > Demotic n-ı͗m=f > Coptic emmōf; Demotic m-ḳdy, m-bꜣḥ, m-šs, m-sꜣ, m-ḏr. As some of these examples show, the shift is not always toward assimilation; compare Demotic n pt ‘in heaven’. The phenomenon also occurs sporadically in cases other than the preposition (Spiegelberg 1925; Erman 1933).

Egyptianizing Features in Phoenician and Punic Inscriptions

105

Aimé-Giron and Donner and Röllig read hrnkʾrt, but the fourth letter could be k or p (see Aimé-Giron 1939: 6). As this group of letters is nonsensical if taken as Phoenician, I think this may be a phrase borrowed from Egyptian, possibly hr n pꜣ rt ‘what is pleasing of the harvest’; compare the extant phrase ḫrp n pꜣ rt ‘first of the harvest’ (Chicago Demotic Dictionary, under rt ‘harvest’). The context is as follows: 3.2.

[mn] ʿbdmlkt lbdbʿl lrb [. . . . . . . . . . .]ty wbrb šl[m w-] [. . . . .]m[. .]ḥtn l[.]n šlš 1 wqllm w[. . . .]wbnšmm ʿlm 2 wbṣl rʾšy[. .]wbṣl plnʿ wpʾl hrnpʾrt [. .]brktmlqrt šlmm 3 wzyt mšqn 25 wšqdm wkmn wlʿrty ʿʿm wššmn wʿqlm 3 wl[. . .] 5 klʾ l yrḥm[.]ʿ [. . . . . . .]wḥ[t]m wḥtmm bṭbʿt w[. .] whn š[. . . . . .]



[From] ʿAbdimalkat to Bodbaʿl the supervisor. [. . .] and in the abundance of well-be[ing] [. . .] three 1, jars and [. . .]; and “sons of the sky”:6 ʿLM-crops:7 2, and RʾŠY onions [. . .] and PLNʿ onions and beans, what is pleasing of the harvest [. . .] Barkatmelqart. Šalamim offerings: 3; olives from [. . .]: 25; almonds, cummin, [. . .] and sesame seeds; wicker baskets: 3; and [. . .]: 5; all for the months [. . . . . .] and a seal, and seal them with a signet ring; and [. . .]

This letter appears to contain a list of requested commodities, each commodity being followed by a number indicating the quantity requested. The proposal that hrnpʾrt is from Egyptian hr n pꜣ rt raises the problem of the status of the Egyptian Aleph in the article pꜣ and its transmission into Phoenician. We have seen that the Egyptian Aleph does not transfer into the Phoenician 6  My reading of bnšmm as ‘sons of the sky’, i.e. crops, goods that grow by means of sun and rain, is new. Aimé-Giron and Donner and Röllig leave this sequence untranslated. 7  ʿlm is most likely the plural form of a word for some fruit or grain; compare Arabic ǵallatun ‘produce, crops, grain, fruit’.

106

Calabro

writing of the divine name ḥrpkrt (from Egyptian ḥr-pꜣ-ẖrd). Names with the compound element pṭ (from Egyptian pꜣ-dı͗), discussed in section 5 below, show the same lack of transmission of this phoneme. The status of the phoneme in Egyptian of the Late Period is not known for certain; by the Coptic period, the phoneme had elided, as shown in spellings like prōme ‘the man’ (from earlier Egyptian pꜣ rmṯ). However, none of these data rule out the proposed rendering of hrnpʾrt as being borrowed from Egyptian. For example, the loanword could have entered the variety of Phoenician spoken by the community in Egypt at an earlier period when the Aleph was heard more distinctly (the many periods of intensive contact between Egypt and Phoenicia, from prehistory through the Late Period, allow this possibility). It is also possible that this is a learned loanword, taking into account the Egyptian writing of the article with Aleph. Thus, although the proposed reading of this sequence is somewhat problematic and not entirely satisfactory, it is a viable possibility, far better than having no reading at all. 4

Possible Calque

Loanwords such as the ones I have just discussed demonstrate some intelligent interaction with Egyptian, although the kind of interaction with the language indicated by loanwords does not have to be very deep. However, the situation is different with calques, or literalistic translations of idioms from one language into another. The existence of a calque implies that speakers of the borrowing language have interacted with the lending language sufficiently to parse the idiom and translate each part. One possible (and, I think, probable) example of a calque has been suggested in the Phoenician corpus from Egypt. This occurs in an urn inscription from Thebes (Lidzbarski 1915: 126): 4.1. šd ʾlnm / ʿbdḥmn bn ʾbṣ[d] Necropolis: / ʿAbdiḥammon son of ʾAbiṣa[d]8 Lidzbarski reports a suggestion by Möller that šd ʾlnm corresponds to Egyptian ẖr.t-nṯr ‘necropolis’, literally ‘what is under the god’. The phrase indicates where the remains of the deceased are placed, coinciding with the purpose of the

8  This patronym has multiple possible readings, depending on the reconstruction of the last letter(s); Lidzbarski deems ʾbṣ[d] the most likely.

Egyptianizing Features in Phoenician and Punic Inscriptions

107

urn. This, then, would be a calque, with šd ‘field’ corresponding to Egyptian ẖr.t and ʾlnm ‘gods’ corresponding to Egyptian nṯr. 5

Egyptianizing Personal Names

Several Egyptian names occur in the Phoenician Egyptian corpus. Those which are fully Egyptian, including no Semitic elements, all come from the jar inscriptions from Elephantine and from unprovenanced inscriptions. We have already seen examples of the names ṣḥpmw, ṣknsmw, and pṭʾs (1.1 through 1.5). The inscriptions containing the remaining names are quoted here. These consist of three Elephantine jar inscriptions (5.1 to 5.3) and an unprovenanced bronze statuette of the deified minister Imhotep (5.4). The three jar inscriptions are published by Lidzbarski (1912), and the statuette is published by Aimé-Giron (1924). 5.1. lʿptḥ bn klblʾ (Belonging) to ʿAptaḥ son of KLBLʾ 5.2. lmlk ṭ / ʿptḥ (Belonging) to the king, Ṭ / ʿAptaḥ 5.3. pṭkns / bn msk[.] Peṭkons / son of MSK[.] 5.4. lwḥprʿ bn ʾšmnytn (Belonging) to Waḥipreʿ son of ʾEšmunyaton Lidzbarski proposes that ʿptḥ in 5.1 and 5.2 is ʿApptaḥ < ʿAbptaḥ < ʿAbdptaḥ (Lidzbarski 1912: 15, no. 42). However, in other inscriptions from the same corpus, we have the full writing ʿbdptḥ (Lidzbarski 1912). It is likely that a vowel was inserted between the two morphological elements to prevent a very difficult consonant cluster (four consonants in a row), as reflected in my vocalization /ʿAbdiptaḥ/. As an alternative to Lidzbarski’s analysis, I would suggest that the name ʿptḥ could be parsed as ꜥꜣ-ptḥ ‘Ptah is great’. Ranke shows this name as being attested in the Middle Kingdom, but the similar name ꜥꜣ-ptḥ-mn-nfr ‘Ptah of Memphis is great’ is attested in the Late Period, though in a context in which the name is partially obscured (Ranke 1935: 57, nos. 15 and 18). The names pṭʾs (1.5) and pṭkns (5.3) are theophoric names of the type pṭDN, a type which is well-known from the Egyptian Aramaic corpus and is also found in the Hebrew Bible. This type of name comes from Egyptian pꜣ-dı͗-DN

108

Calabro

‘the one whom DN gave’. In 1.6 (line 3), we encountered the name pṭ, apparently a hypocoristicon of the type pṭ-DN ‘DN has given’. A name of the same form appears on a Moabite seal (see Calabro 2011: 107). It is interesting that in the pilgrim graffiti on the walls of the temple of Seti I at Abydos, the only three instances of this type of name are clearly Aramaic, as indicated by the script: pṭmwn, pṭnʾsy, and pṭsry (Lidzbarski 1915: 107, 113–14). In addition, the three instances of this type of name in the Phoenician corpus from Egypt are in contexts in which it could be argued that the people named are not necessarily Phoenician. On the basis of this, one could argue that this type of name was characteristically Aramaic and not Phoenician in some locales in Egypt; however, the evidence is equivocal. There is at least one other example in the broader Phoenician corpus outside of Egypt (Degen 1970: 219; Donner and Röllig 2002: no. 29), but the fact that there is only one is perhaps telling. It could be that, at least in some Egyptian locales, pṭ-names were rare or nonexistent among the Phoenician population and were instead indicative of Aramaic or Hebrew linguistic affiliation. In any case, pṭ-names do not seem to have been as prevalent among the Phoenicians as among the Aramaic speakers living in the same locales. The name wḥprʿ in 5.4 is from the Egyptian name wꜣḥ-ı͗b-rꜥ ‘the heart of Re endures’ (Ranke 1935: 72, no. 28). The names I have just discussed are all theophoric names in which both the verbal or construct element and the divine name are Egyptian (or hypocoristic names in which only an Egyptian verbal or construct element remains). These names are gathered in table 3. Another, much larger class of Phoenician names attested in the inscriptions includes a Semitic verbal or construct element combined with an Egyptian divine name. I have not discussed examples of this type here, since the type is very widespread in the Phoenician world in general, and there is nothing remarkable in its occurrence in the corpus from Egypt. Neither of these types is necessarily indicative of deep interaction with Egyptian language. The first type may be indicative of intermarriage between Phoenicians and Egyptians or may simply indicate an affinity for Egyptian names, and the second type is indicative of the adoption of Egyptian gods into the Phoenician pantheon, but in neither case does a Phoenician person have to be able to speak Egyptian in order to give the name to a child. Thus both types of names are indicative of shallow cultural interaction, on the order of broadly-defined “cultural influence.” What would be indicative of deeper interaction would be names in which the verbal or construct element is Egyptian but the divine name is Semitic, since this would imply knowledge of the meanings of Egyptian onomastic elements, though not necessarily by the one bestowing the name in a particular instance; there are actually names of this type attested in Hebrew inscriptions, like pṭyhw (Calabro 2011: 107–8),

Egyptianizing Features in Phoenician and Punic Inscriptions

109

but I have not encountered any certain examples of this type in the Phoenician corpus from Egypt. TABLE 3  Egyptianizing Personal Names Phoenician

Egyptian

Reference

ṣḥpmw ṣknsmw pṭʾs pṭkns pṭ ʿptḥ wḥprʿ

ṯꜣy-ḥp-ı͗m=w ṯꜣy-ḫnsw-ı͗m=w pꜣ-dı͗-ꜣst pꜣ-dı͗-ḫnsw pꜣ-dı͗ ꜥꜣ-ptḥ wꜣḥ-ı͗b-rꜥ

1.1 to 1.4 1.5 1.5 5.3 1.6 5.1, 5.2 5.4

6

Distinctive Formulae

In the large group of pilgrim graffiti from Abydos, the typical formula used in the Phoenician graffiti is distinct from that used in the Aramaic graffiti. Concerning this distinction and the origins/antecedents of the two formulae, Lidzbarski (1915: 95) writes: Die Sprache der semitischen Graffiti ist phönizisch und aramäisch. Beide Arten haben typische Formen. Die phönizischen lauten: “Ich bin x” mit oder ohne nähere Bestimmungen, während die aramäischen eine Benediktion enthalten: “Gesegnet sei x vor Osiris”. Beide Arten hatten sicherlich ihre Vorbilder außerhalb Ägyptens. So finden sich unter den sinaitischen Inschriften Graffiti ähnlicher Form, wie die der aramäischen in Abydos. In his reference to “sinaitischen Inschriften,” Lidzbarski likely has reference not to the famous early alphabetic inscriptions from Sinai, nor to Egyptian inscriptions from that area, but to Nabatean graffiti from Sinai (Cooke 1903: 258–62). In at least one of these Nabatean graffiti, there is expressed a desire to be in remembrance before the god Baal, similar to the Aramaic graffiti from Abydos (Cooke 1903: no. 105). As for the Phoenician formula, it is true that the authors of Egyptian autobiographical steles frequently identify themselves with the formula: (list of titles) + (PN) + ‘who says, I am a man who is’ + (list of attributes)—that is, a formula

110

Calabro

using the first person singular pronoun and the author’s personal name and having the purpose of perpetuating remembrance of the author. However, it is probable that the Phoenician graffito formula has a more appropriate parallel in Phoenician tomb inscriptions, as in the following sampling:9 KAI 10 ʾnk yḥwmlk mlk gbl bn yḥrbʿl bn bn ʾrmlk mlk gbl . . .  I am Yaḥawwimilk, king of Byblos, son of Yiḥarbaʿl, grandson of ʾUrimilk king of Byblos . . . KAI 13 ʾnk tbnt khn ʿštrt mlk ṣdnm bn ʾšmnʿzr khn ʿštrt mlk ṣdnm . . .  I am Tabnit, the priest of Astarte, king of the Sidons, son of Eshmunazor the priest of Astarte, king of the Sidons . . . KAI 24 ʾnk klmw br ḥy[ʾ] mlk gbr ʿl yʾdy . . .  I am Kilamuwa son of Ḥayy[a], mighty king over Ya‌ʾudiya . . . Therefore, rather than seeing the formulae used in these graffiti as derived from Egyptian sources, we should view them as native Phoenician and Aramaic formulae; the Egyptian influence can be observed not in the origin of the formulae, but in the fact of their being addressed to the Egyptian deity Osiris. However, there is another Phoenician formula that may show closer links to Egyptian. This is the formula ‘may DN give life to PN’. We find this formula in Phoenician inscriptions on the bases of two statuettes of the god Harpokrates. The first of these was quoted above (1.6). The second reads as follows: 6.1.

dı͗ ṯꜣ(w) ḥr-p(ꜣ)-ẖrd n ws-ꜥnḫ sꜣ pthy [. . .] ḥrpkrṭ ytn ḥyym lʿms bn ʾšmnytn bn ʿzrmlk wḥq nn bʿl bnʾ



May Harpocrates give breath to Wasʿanḫ son of PTHY [. . .] May Harpocrates give life to ʿAmos son of ʾEšmunyaton son of ʿIzrimilk and Ḥeqanun master of [. . .].

In Egyptian, the giving of life or breath to mortals by the god is a prevalent theme and occurs in many formulaic contexts. In inscriptions from the Phoenician homeland, such as that on the sarcophagus of Eshmunazor, deities are entreated to provide blessings and favor, but not usually life or breath. 9  In these quotes, the abbreviation KAI stands for Donner and Röllig 2002.

Egyptianizing Features in Phoenician and Punic Inscriptions

111

It is also interesting that the direct object in the hieroglyphic portion of 6.1 comes directly after the verb but before the subject, which violates standard Egyptian word order. An inscribed statuette of the god Imhotep, which bears a hieroglyphic and a Phoenician inscription like the second Harpocrates statuette discussed above, is interesting for the way in which it seems to combine its two inscriptions. 6.2. ı͗ı-͗ m-ḥtp sꜣ ptḥ dı͗ ꜥnḫ lwḥprʿ bn ʾšmnytn

Imhotep son of Ptah, given life (or: may he give life) to Waḥipreʿ son of ʾEšmunyaton

As noted by Aimé-Giron (1924: 2–5), these inscriptions can be understood either separately (‘Imhotep son of Ptah, given life’ and ‘[Belonging] to Wahipre son of Eshmunyaton’) or in combination as a single message (‘May Imhotep son of Ptah give life to Wahipre son of Eshmunyaton’). The latter reading would align this example with 1.6 and 6.1 above. Vittmann calls the two inscriptions on this object “quasi-komplementar” (Vittmann 2003: pl. 8, caption). However, as with the hieroglyphic portion of 6.1, the word order in this combined reading is ungrammatical according to standard Egyptian, which would normally have a VSO word order in this formula. The apparent affinities with Egyptian language and formulaic usage that we see in this small group of inscriptions, together with the divergences from standard Egyptian word order, raise many questions. It is obvious that they provide evidence of interaction with Egyptian, but what kind of interaction, and what variety of Egyptian? We find in these inscriptions neither homeland Phoenician nor standard Egyptian, but something in between. This evidence is tantalizing; however, the evidence is not adequate to answer the many questions it raises. With regard to these three statuettes, we can at least say that they demonstrate a level of interaction that goes beyond the shallow influence noted earlier in the case of proper names. 7 Conclusions There were at least two distinct Northwest Semitic cultures forming communities in Egypt during the Iron Age. The first, which is the focus of this study, is marked by Phoenician language, the ‘I am PN’ graffito formula (which seems to have its origin in native Phoenician tradition), and personal names compounded with the divine names Baʿl and Rašap. The second culture is

112

Calabro

marked by Aramaic language, the ‘May PN be blessed before DN’ graffito formula, and personal names of the form pṭ-DN (from Egyptian pꜣ-dı͗-DN). In addition, the two cultures had distinct scribal traditions, each marked by a different ductus and different conventions for transliterating Egyptian consonants. The two communities lived side-by-side in various Egyptian locales and had business dealings with one another, and both communities assimilated elements of Egyptian culture to a large degree. However, they retained their differences and assimilated Egyptian cultural elements in different ways. One could attribute this to their different contact trajectories (the Phoenician communities were tied into maritime routes along the Mediterranean sea, whereas the Aramaic communities were tied into land-based caravan routes), and possibly to different modes of contact with Egypt. There are some indications that Phoenician communities in Egypt developed a distinct speech variety that was influenced by Egyptian language. The inscriptions in the present study show examples of Egyptian borrowed words (e.g. npr in 3.1, hrnpʾrt in 3.2) and one possible example of a calque (šd ʾlnm in 4.1). In addition, several inscriptions show evidence of phonological shifts, namely the interchange between /l/ and /n/ and the occasional shift of */n/ > /m/. These shifts possibly indicate Egyptian influence on the phonology of Phoenician speakers, in that certain options in the Phoenician repertoire were frequently selected by some speakers, forming a contact variety of the language based on compromise. Other speakers hypercorrected to avoid Egyptian phonological patterns, and in doing so further differentiated their speech from standard Phoenician. In my opinion, therefore, we can see in the Phoenician inscriptions from Egypt a greater and deeper degree of influence from Egyptian than what we see in Phoenician and Punic inscriptions from the Phoenician homeland, the Levant, and the western Mediterranean. Bibliography Aimé-Giron, N. 1924. “Glanures de mythologie syro-égyptienne,” Bulletin de l’Institut français d’archéologie orientale 23: 3–25. ———. 1939. “Adversaria semitica,” Bulletin de l’Institut français d’archéologie orientale 38: 1–18 with plate 1. Allen, James P. 2000. Middle Egyptian: An Introduction to the Language and Culture of Hieroglyphs. Cambridge. Bresciani, Edda. 1987. “Fenici in Egitto,” Egitto e Vicino Oriente 10: 69–78. Calabro, David. 2011. “Personal Names with Egyptian Elements in Pre-exilic Hebrew Inscriptions,” in Aaron Demsky (ed.), These Are the Names: Studies in Jewish Onomastics, Vol. 5. Ramat Gan. 95–118.

Egyptianizing Features in Phoenician and Punic Inscriptions

113

Chiera, G. 1987. “Fenici e cartaginesi a Menfi,” Rivista di studi fenici 15: 127–31. Clermont-Ganneau, C. 1905. “Proscynèmes phéniciens et araméens d’Abydos,” in Recueil d’archéologie orientale, Vol. 6. Paris. 391–400. Cooke, G. A. 1903. A Text-Book of North-Semitic Inscriptions. Oxford. Daressy, M. G. 1903. Textes et dessins magiques (CGC 9 [Nos. 9401–9449]). Cairo. Degen, Rainer. 1970. “Der Name Harpokrates in phönizischer und aramäischer Umschreibung,” WO 5: 218–21. Derenbourg, Joseph, and Hartwig Derenbourg. 1885. “Les inscriptions phéniciennes du temple de Seti à Abydos, publiées et traduites d’après une copie inédite de M. Sayce,” RA 1: 81–101 with plates 1–4. Donner, Herbert, and Wolfgang Röllig. 1964. Kanaanäische und aramäische Inschriften, Vol. 2. Kommentar. Wiesbaden. ———. 2002. Kanaanäische und aramäische Inschriften, Vol. 1. Erweiterte und überarbeitete Auflage. Wiesbaden. Erman, Adolf. 1933. Neuaegyptische Grammatik (2nd ed.). Leipzig. Friedrich, Johannes. 1964. “Ein kurzer phönizischer Graffito aus Abu Simbel,” ZDMG 114: 226–227. Friedrich, Johannes, and Wolfgang Röllig (edited by Maria Giulia Amadasi Guzzo). 1999. Phönizisch-punische Grammatik (3rd ed.). Rome. Gibson, John C. L. 1982. Textbook of Syrian Semitic Inscriptions, Vol. 3. Phoenician Inscriptions. Oxford. Lidzbarski, Mark. 1898. Handbuch der nordsemitischen Epigraphik (2 vol.). Weimar. ———. 1907. Kanaanäische Inschriften. Giessen. ———. 1912. “Phönizische und aramäische Krugaufschriften aus Elephantine,” Abhandlungen der königlich preussischen Akademie der Wissenschaften (Philosophisch-historische Classe). ———. 1915. Ephemeris für semitische Epigraphik, Vol. 3. Giessen. Loprieno, Antonio. 1995. Ancient Egyptian: A Linguistic Introduction. Cambridge. Mariette, Auguste. 1869. Abydos: Description des fouilles, Vol. 1. Paris. Maspero, G. 1917. Introduction a l’étude de la phonétique égyptienne. Paris. Möller, G. 1936. Hieratische Paläographie (3 vol.). Osnabruck. Ranke, H. 1935. Die ägyptischen Personennamen, Vol. 1. Glückstadt. Segal, J. B. 1983. Aramaic Texts from North Saqqâra with Some Fragments in Phoenician. London. Spiegelberg, Wilhelm. 1904. Die demotischen Inschriften (CGC 16). Leipzig. ———. 1925. Demotische Grammatik. Heidelberg. Steiner, Richard. 1982. Affricated Ṣade in the Semitic Languages. New York. Vittmann, Günter. 2003. Ägypten und die Fremden im ersten vorchristlichen Jahrtausend. Mainz am Rhein. Vycichl, Werner. 1983. Dictionnaire étymologique de la langue copte. Leuven.

Head-Marking in Neo-Aramaic Genitive Constructions and the ezafe Construction in Kurdish Eran Cohen 1 Preamble Genitive constructions (Hebrew smīḵūṯ, Arabic ʾiḍāfa) are an important syntactic feature of ancient Semitic. Semitic genitive constructions have an extensive range having nouns, pronouns, and clauses as attributes (Goldenberg 1995; Izre’el and Cohen 2004: 75–78; Cohen 2008: 36, n. 16). The range of application as well as the extent of explicit marking of genitive constructions seems to have shrunk over time in many Semitic languages, especially in Aramaic. Consequently, it was quite a surprise to find that one Neo-Aramaic dialect, Jewish Zakho, has an elaborate and highly explicit system of genitive constructions (mentioned and discussed first in Goldenberg 1993: 296–298), which in many respects reminds us of ancient Semitic. However, it is quite clear that this situation in Jewish Zakho is not a retention from ancient Semitic, since Aramaic has never shown the high level of complexity of genitive constructions compared with Akkadian, Arabic, or Ethiopic. Accordingly, there is probably some other explanation for this elaborate system in Jewish Zakho. The most likely candidate, which explains some of the puzzle, is language contact between Neo-Aramaic (henceforth NA) and Northern Kurdish, or Kurmanjî. In this paper, I will attempt first to provide a short survey of the Semitic genitive constructions, especially their status and extent in ancient Semi­tic syntax; second, to demonstrate their full revival in one NA dialect; and third, to offer a partial explanation for at least some of these features in NA within the framework of language contact. 2

Ancient Semitic Genitive Constructions

Genitive constructions are an ancient Semitic feature; unlike such constructions in ancient Indo-European for instance, the Semitic constructions were *  I would like to thank my colleague Eitan Grossman for his great help in several stages in the preparation of this paper.

© koninklijke brill nv, leiden, ���5 | doi ��.��63/9789004300156_008

115

Head-marking in Neo-Aramaic Genitive Constructions

originally marked twice—once on the head (that is, the entity in construct state, which survives to this day in many languages) and once on the attribute (which was marked by the genitive case in ancient Semitic). Table 1 provides an illustration of the functional extent of these constructions as well as the marking: TABLE 1  Genitive constructions in ancient Semitic (Old Babylonian Akkadian, ca. 1800 BCE) Scheme Head

Attribute

Example Head1

Attribute

Translation

Remarks

‘your house’

bīt house.NUC

-ka GEN.2.M.SG ab-ī-ni father-GEN-GEN.1.PL ēpuš-u 1.sg.do.PRT-SUBJ

(suffix pronoun)

ša pron.NUC

kâta OBL.2.SG ab-ī-ni father-GEN-GEN.1.PL ēpuš-u 1.Sg.do.PRT-SUBJ

‘that (of) yours’ ‘that (of) our father’ ‘that (which) I built’

(oblique pronoun)

‘like you’

preposition

kīma as.NUC

kâta OBL.2.SG ab-ī-ni father-GEN-GEN.1.PL ēpuš-u 1.Sg.do.PRT-SUBJ

‘like our father’ ‘as/that I built’

conjunction

pronoun substantive

substantive clause

pronoun pronoun

substantive clause

pronoun adverb

substantive clause

1  Glossed throughout by NUC (for ‘nucleus’).

‘house of our father’ ‘the/a house I built’

‘relative clause’

116

Cohen

In Akkadian (as representative of the ancient Semitic languages), the mechanism is found with all prepositional phrases, conjunctional subordination, relative clauses, etc.: the syntactic head could be a noun (including adjectives2 and infinitives), a pronoun or an adverb, and be expanded by substantives, pronouns, and clauses. Two points are fundamental: both head and attribute are initially marked: the substantival head, by the apocopate form (generally; not with the plural masculine of substantives in Akkadian), whereas the attribute is marked as such by the genitive case in the singular and the oblique in the plural. In addition, in Akkadian only, verbal forms that function as attributes are marked by the subjunctive form. This form mostly occurs in this genitive function, namely, following any head in construct state (see the third line in each section of table 1 above). 3

Aramaic Genitive Constructions

Syriac is an intermediate phase between the ancient Semitic model and NA. The genitive construction with apocopate marking still exists (table 2, line 1: bar ʾalāhā) but the construction with d- predominates (on table 2, line 2 and 3: brā d-alāhā as well as br-ēh d-alāhā, that is, head noun with an attributive pronoun, then a head pronoun [i.e., d-] with an attribute. For the construction see also Hopkins 1997). d- is a pronoun, which is clearly capable of being anaphoric (e.g., table 2, line 5, where the two cases of the pronoun d- clearly represent a person and a thing respectively). As a mechanism to express definiteness, one finds in Syriac ʿamm-ēh d-ḡaḇrā (line 4; also Arayathinal 1959, 1.89). However, when there is no antecedent nearby, the d- is often strengthened by another correlative pronoun (Nöldeke 1904: §236 hāw d-). Marking genitive constructions in Syriac is limited to the head, which is fairly consistently marked as such. Table 2 summarizes the different options found in Syriac:

2  For which see Cohen 2008: 48–50.

117

Head-marking in Neo-Aramaic Genitive Constructions TABLE 2  The genitive construction and related patterns in Syriac3 Head

Attribute

br son-NUC ʿamm with daPRON.NUC

bar son-NUC brā dson PRON.NUC -ēh dGEN.3.M.SG PRON.NUC -ēh dGEN.3.M.SG PRON.NUC ḥṭap̄ drob-PST.3.M.SG PRON.NUC

1 2 3 4 5

4

Head

Attribute

Translation

ʾalāhā God ʾalāhā God ʾalāhā God ḡaḇrā man ḥaḇr-ēh friend-GEN.3MS

‘God’s son’ ‘(the) son of God’ ‘the son of God’ ‘with the man’ ‘the one (who) robbed that (which is) his friend(’s)’

Source

Nöldeke 1904: §205

Ibid. §222 Ibid. §2094

Genitive Constructions in Neo-Aramaic

The Jewish dialect of Zakho consistently marks both head and attribute (Cohen 2010; 2012: 91–139). The data are summarized in Table 3a and 3b. The head may, as in ancient Semitic, consist of substantives, pronouns, adverbs, infinitives, and adjectives. It is marked either by the apocopate form (the old Semitic construct state, found occasionally in other dialects, e.g., Khan 2008b: 174–176) or by the ending -ıd (or -ıt). This morpheme is very common in many NA dialects (Khan 2008a: §§10.16 and 14.5 for the dialect of Barwar, 2008b: 174–176 for Jewish Urmi). In early Jewish NA (Sabar 1976), most prepositions have this

3  The syntactic analysis of the Syriac material in the table follows the one conducted in Goldenberg 2013: 236–238. For another view about the nature of d- in Syriac, see Wertheimer 2001. 4  See also nessaḇ mennan da-y(h)aḇ lan ‘he will take away from us that which he has given us’ (Goldenberg 1995: 14, my emphasis).

118

Cohen

marking too: ʾımm-ıd ‘with-NUC’, ʾıll-ıd ‘to-NUC’, b-ıd ‘in-NUC’, etc. In Jewish Zakho the form -ıd is best analyzed synchronically as equivalent to the apocopate marking of the head in construct state (and not, for instance, as equivalent to the pronoun d-, as stated by Goldenberg 1993),5 most conspicuously because it occurs after prepositions (b-ıd ‘in-NUC’, ımm-ıd ‘with-NUC’). In this position, there is no motivation for the pronoun d- to occur, since there is no sense in pronominally representing the preposition (as there is, e.g., between two nouns, where d- perfectly represents the first noun). This morpheme occurs also with adverbials (qarwāwe ‘around’, qarwāw-ıt ‘close to’), pronouns (ku-d, mani-d ‘whoever’; mā-d ‘whatever’, etc.) and adjectives (ʾaw xšīw-ıt kullu ‘the (most) important of them all’), whereas the pronouns d- or dīd (which are almost free variants of one another) do not.6 In other dialects, this -ıd is analyzed in several ways, often as syntactically equivalent to d- (e.g., Khan 2008A: §§10.16 and 14.5). The genitive attribute7 (on table 3b) is regularly marked by special allomorphs of the determiners and pronouns which are found only in this function, in a manner similar to the ancient Semitic genitive marking. Moreover, like Akkadian, which marks its attributive clauses after a construct state with the subjunctive form of the verb, in Jewish Zakho there is a special allomorph for the present copula which is found only as attribute. Table 3a

Genitive constructions in Jewish Zakho: head

Forms

• substantive (any) • pronoun (many) • adverb (=preposition/ conjunction) • infinitive • adjective (special cases)

Marking

1. apocopate forms (like Akkadian) 2. -ıd

5  Goldenberg (2000: 79) speaks of alternation between the apocopate form of the nucleus and the forms with -ıd. However, in an earlier work (Goldenberg 1993: 296–297), he shows the correspondence between, e.g., bēs ʾīlāha (‘God’s house’) and bēsı-d ʾīlāha (‘the house of God’), but analyzes each of them differently, explaining that this -d “should be regarded as basically embodying the head, and in apposition to the explicit noun which it represents, so that bēs ʾīlāha would structurally parallel d īlāha” (296). 6  This system is elaborately described in Cohen 2010: 81–87; 2012: 92–105. 7  For the attribute, see Cohen 2010: 87–93; 2012: 105–123.

119

Head-marking in Neo-Aramaic Genitive Constructions Table 3b

Genitive constructions in Jewish Zakho: attribute

forms

marking

remarks

substantive pronoun adverb(ials)

genitive determiners genitive pronouns none verbal forms: none attributive copula

used only for genitive attributes

clause

main subordination strategy

The origin of the morpheme -ıd is generally agreed to be the Aramaic pronoun d-, in one of two historical structures. One is found on table 4, line 1 (baytā d-ḡaḇrā), where the pronoun d- was supposedly re-analyzed as an ending (e.g., Tsereteli 1965: 229); the other is what we have in table 4, line 2 (bayt-ēh d-ḡaḇrā). The advantage of the second structure is that it can also explain the ending -ıd on a preposition (see line 3 on the same table). TABLE 4  Historical source of -ıd Syriac

1 baytā house 2 bayt-ēh house-GEN.3.M.Sg 3 ʿamm-ēh with-GEN.3.M.Sg

Jewish Zakho

dPRON.NUC dPRON.NUC dPRON.NUC

ḡaḇrā8 man ḡaḇrā man ḡaḇrā (see explanation in man Mengozzi 2005 for -ıd on prepositions)

bēṯ-ıd house-NUC bēṯ-ıd house-NUC ʾımm-ıd with-NUC

gōra man gōra man gōra man

But to what do we owe this change, that is, that the historical Aramaic pronoun d- has become the morpheme -ıd which signals the head as such? The claim is that it is in fact a pattern replication of a similar construction found in Kurmanjî, which shows some basic similarities (although it is by no means identical) to the Jewish Zakho genitive constructions. 8  The equivalent syntagm is still attested in Jewish Zakho: bēsa dīd daw šıhāra ‘the house of the blind (man)’ (341).

120 5

Cohen

EZAFE Constructions in Kurmanjî

Kurmanjî, or Northern Kurdish, is a language which has been in contact with the dialect of Jewish Zakho and many other NA dialects in North Iraq. It is an Indo-European Iranian language, which, like other Iranian languages, has a construction termed in Persian ezāfé (a term originating in the Arabic term ʾiḍāfa, that is, annexation, or genitive construction).9 The head of the ezafe construction in Kurmanjî is characteristically marked by an ending (table 5a has a summary of the data), it inflects for gender, number, and definiteness, basically appended to nouns which are further expanded by substantives, pronouns, prepositional phrases, adjectives, and clauses (table 5b as well as ex. [3]). Table 5a

Genitive constructions in Kurmanjî: head

Forms

Marking

• substantives • (denominal) adverb­ial conjunctions • infinitives pronoun (that is, the demonstrative ezafe, the unappended ending)

Table 5b

def indef

ezafe endings m.sg f.sg -(y)ê -(y)a -î -e

yê

ya

c.pl -(y)ên (-e)

yên

Genitive constructions in Kurmanjî: attribute

Forms

Marking

Remarks

substantive pronoun prepositional phrase adjective clause

oblique case none10

oblique case is also used for objects and after prepositions

none

9  For this construction in Kurmanjî see MacKenzie 1961, Schroeder 1999, and Haig 2011. 10  The adjective and the clause attributes are not marked as such.

Head-marking in Neo-Aramaic Genitive Constructions

121

The ezafe ending has a peculiarity. The “demonstrative ezafe” (MacKenzie 1961: §266, table 5a) occurs when the construction does not immediately follow its referential noun, and in that case, the ending is found separated from its head noun, and is followed by the attribute (substantives, pronouns, prepositional phrases, etc.), as shown in yê pola in ex. [1] below. 6

A Tentative Explanation

We know that in all probability the -ıd ending originates in the pronoun d-. It is suggested here that what triggered this was the ezafe morpheme in Kurmanjî. Matras and Sakel (2007) and Sakel (2007) discuss pattern replication, namely, when what is borrowed is rather a function with its structural features, while the actual phonological form is not borrowed (which type is termed matter replication). The most basic step in the process of pattern replication is pivot matching: spotting a pivot structure in the model language, and matching it with a structure in the replica language, assigning a similar function to the new structure, termed replica construction. The model pivot in our case is the demonstrative ezafe morpheme in Kurmanjî, and the matched pivot in the replica language, namely, NA, is the pronoun d-. In comparing exx. [1] and [2] it is easy to see why they are matched: [1] dil-ê wî yê pola (N . . .) ez+x heart-EZ.M OBL.3.M.SG EZ.M steel ‘his heart of steel’ (Kurmanjî, Thackston 2006: §5) [2] lıbb-ıd Parʿo u d- mıṣrāy-e (N . . . ) d+x heart-NUC CONJ PRON.NUC egyptian-PL ‘the heart of Pharaoh and that (of the) egyptians’ (Early NA, Sabar 1976: 51) Both the ezafe exponent and the pronoun d- may stand alone as heads but at the same time they are referential to some substantive in the background (in both examples it is ‘heart’). In addition, both are further expanded by several possible attributes. Stage 1 in Table 6 lists the situation after matching the pivots, but before the change. We should assume that NA had at that time the original Aramaic genitive construction (#Nconst x#), which is retained in several dialects, including

122

Cohen

Jewish Zakho.11 Nevertheless, there is no symmetry, and one more step is needed, namely, to create an extension so that the existing pronoun d- can be fully used as the ezafe, namely, appended to the end of the substantive (stage 2 in table 6): TABLE 6 

2nd stage

Kurdish Neo-Aramaic

With an immediately preceding substantive

(N . . .) EZ+x (ex. [1]) (N) d-+x (ex. [2])

N-EZ x (ex. [3]) ?

⇐

1st stage

Without an immediately preceding substantive

N-d x (new construct state, ex. [4])

These changes are reflected in exx. 3 and 4: [3] germî-ya dil-ên kurd-ên qenc N-ez x warmth-EZ.F heart-EZ.PL Kurd-EZ.PL good ‘The warmth of the hearts of the good Kurds’ (Kurmanjî, Thackston 2006: 180) [4] ʿazız-t-ıt lıbb-ıt ḥakōma N-d x dear-FS-NUC heart-NUC king ‘dearestF.SG of the king’s heart’ (Jewish Zakho) The pronoun d- has grammaticalized into a suffix that marks the preceding entity as the syntactic head and, by default, the entity that immediately follows it as the attribute. Kurmanjî temporal conjunctions—dema, gava, çaxê, wexta (Thackston 2006: §29)—seem in fact to stem originally from nouns denoting time (‘time’, ‘moment’, etc.) with an ezafe ending, which are able to join a clause either directly or by means of the subordinator ku. Prepositions in Kurmanjî are not 11  For instance, bēsNUC gōra ‘house [of a/the] man’)

Head-marking in Neo-Aramaic Genitive Constructions

123

generally marked by the ezafe, which seems to be essentially limited to nouns (including infinitives and these denominal temporal expressions). However, Jewish Zakho is synchronically much less limited: the morpheme -ıd marks as head almost anything except a finite form of the verb. Jewish Zakho also has a rich formation of the original construct state/nucleus marking, meaning that the two strategies often compete with the same entity. The suggested pattern replication may be helpful in explaining one critical issue: the compatibility of clauses with the construct state (or the head) in Jewish Zakho. This feature of ancient Semitic (Akkadian, Hebrew, Ethiopic, Classical Arabic) is entirely missing in the earlier stages of Aramaic, but exists nevertheless in Jewish Zakho, as is shown in exx. [5] and [6]: [5]

ʾıl xabır {mxē-la baxt-e to word-NUC hit-PRET-3.F.SG wife-3.M.SG at bıt ʾarya} (unbound form: xabra) at lion ‘to the word(s) {his wife said about the lion}’ (Jewish Zakho)

[6] bıt xabr-ıt {mír-rē-la} in word-NUC say-PRET-3.M.SG-DAT.3.F.SG ‘about the word(s) {he told her}’ (Jewish Zakho) It may have been triggered by the ability of the ezafe to occur with a clause,12 or, alternatively, by the fact that d-, the original form, had been compatible with a clause all along. The difficulty with the latter idea is that it does not account for cases such as the one in ex. [5], in which the head occurs with a clause when marked by the apocopate form, that is, where the pronoun d- has never been involved. 7 Conclusions Along with the solution offered as regards the inception of the head marker -ıd in NA, some questions remain open and cannot be explained based on language contact with Kurmanjî:

12  aw k’as-ē awwilī b-ē-t DEM person-EZ.M.SG first SUBJ-come-PRES.3.SG ‘that person who comes first’ (McKenzie 1961: §307, glossed apud Haig 2011: 366)

124

Cohen

1. How come not only substantives (or denominal entities, as is the situation in Kurmanjî) in NA may occur as heads in genitive constructions, but also pronouns, adjectives, adverbs (which turn in the process into prepositions and conjunctions), in line with the ancient Semitic system? In other words, what could explain the almost full-scale re-establishment of the ancient Semitic system of genitive constructions? 2. What triggered the development of a specialized marking for the genitive attribute in these constructions in NA, especially in Jewish Zakho? Kurmanjî marks the genitive attribute very imperfectly (table 5b), and this function is not marked by a unique signal (since the oblique case, found with the substantive and pronominal attribute, is also found with verbal and prepositional complements), whereas Jewish Zakho marks the attribute whenever it occurs with a definite determiner, a pronoun, or when it contains a copula (table 3b and n. 6). Bibliography Arayathinal, Th. 1959. Aramaic grammar (2 vol.). Mannanam, Kerala State, S. India. Cohen, E. 2008. “Adjectival ša Syntagms and Adjectives in Old Babylonian,” BSOAS 71: 25–52. ———. 2010. “Marking nucleus and attribute in North-Eastern Neo-Aramaic,” in Proceeding of the VIII Afro-Asiatic Congress (September 2008, Naples), Studi Maghrebini (Nuova Serie), VI, 79–94. ———. 2012. The Syntax of Neo-Aramaic: The Jewish Dialect of Zakho (Neo-Aramaic Studies). Piscataway. Goldenberg, G. 1993. Review Article of Jastrow 1988. JSS 38: 295–308. ———. 1995. “Attribution in Semitic Languages,” Langues Orientales Anciennes: Philologie et Linguistique 7: 1–20. (= Goldenberg 1998: 46–65) ———. 1998. Studies in Semitic Linguistics. Jerusalem. ———. 2000. “Early Neo-Aramaic and Present-Day Dialectal Diversity,” JSS 45: 69–89. ———. 2013. Semitic languages: Features, structures, relations, processes. Oxford. Haig, G. 2011. “Linker, relativizer, nominalizer, tense-particle: On the Ezafe in West Iranian,” in Foong Ha Yap, Karen Grunow-Hårsta, Janick Wrona (eds.), Nominalization in Asian Languages: Diachronic and Typological perspectives. Amsterdam. 363–390. Hopkins, S. 1997. “On the construction šmēh l-gaḇrā (the name of the man) in Aramaic,” JSS 42: 23–32. Izre’el, Shlomo and Eran Cohen. 2004. Literary Old Babylonian (Lincom Europa Languages of the World/Materials 81). Munich. Jastrow, Otto. 1988. Der neuaramäische Dialekt von Hertevin (Provinz Siirt). Wiesbaden.

Head-marking in Neo-Aramaic Genitive Constructions

125

Khan, G. 2007. “North Eastern Neo-Aramaic,” in Y. Matras and J. Sakel (eds.), Grammatical Borrowing in Cross-Linguistic Perspective. Berlin. 197–214. ———. 2008a. The Neo-Aramaic Dialect of Barwar, Vol. 1. Grammar. Leiden. ———. 2008b. The Jewish Neo-Aramaic Dialect of Urmi (Neo-Aramaic Studies 2). Piscataway. Krotkoff, G. 1982. A Neo-Aramaic Dialect of Kurdistan. New Haven. Mackenzie, D. N. 1961. Kurdish Dialect Studies, Vol. 1. London. Matras, Y. and Sakel J. 2007. “Investigating the mechanisms of pattern replication in language convergence,” Studies in Language 31: 829–865. Mengozzi, A. 2005. “Extended Prepositions in Neo-Aramaic, Kurdish and Italian,” in Alessandro Mengozzi (ed.), Studi Afroasiatici. IX Incontro Italiano di Linguistica Camitosemitica. Milan. 371–90. Nöldeke, Th. 1904. Compendious Syriac Grammar. Translated from the second and improved German edition by James A. Crichton. London. Sabar, Y. 1976. Pǝšaṭ Wayǝhī Bǝšallaḥ. A Neo-Aramaic Midrash on Beshallaḥ (Exodus). Introduction, phonetic transcription, translation, notes and glossary. Wiesbaden. Sakel, J. 2007. “Types of loan: Matter and pattern,” in Yaron Matras and Jeanette Sakel (eds.), Grammatical borrowing in cross-linguistic perspective. Berlin – New York. 15–29. Schroeder, Ch. 1999. “Attribution in Kurmanji (Nordkurdisch),” in K. H. Wagner and W. Wildgen (eds.), Studien zur Phonologie, Grammatik, Sprachphilosophie and Semiotik. Bremen. 43–63. Thackston, W. M. 2006. Kurmanji Kurdish—A Reference Grammar with Selected Readings. Unpublished manuscript. (available at http://www.fas.harvard.edu/ ~iranian/Kurmanji/) Tsereteli, K. 1965. “A Type of Nominal Syntagm in Modern Aramaic Dialects,” BSOAS 28: 227–232. Wertheimer, Ada. 2001. “The Functions of the Syriac Particle d-,” Le Muséon 114: 259–289.

Notes on Foreign Words in Hatran Aramaic Riccardo Contini and Paola Pagano 1

Hatran Aramaic and its Language Contacts1

Among the “Middle Aramaic”2 or “Later Imperial Aramaic”3 varieties, the Aramaic of Hatra and Assur has been hitherto rather less studied, particularly regarding the foreign components in its vocabulary. In this perspective, not only Judean and other Palestinian Aramaic varieties, but also Nabataean, Palmyrene, and even Old Syriac have been repeatedly investigated, particularly in the last 25 years.4 In sharp contrast, Hatran Aramaic vocabulary is of course 1  R C is responsible for the non-Akkadian lexical data, PP for the Akkadian ones, whereas we are jointly answerable for the general structure of this paper and for the etymological solutions presented here. Materially, RC wrote §§1, 2 and the first half of §4; PP wrote §3 and the second half of §4. Note the following abbreviations: Akk. = Akkadian; ANA = Ancient North Arabian; Arm. = Armenian; BA = Biblical Aramaic; Bab. = Babylonian; CA = Classical Arabic; CPA = Christian Palestinian Aramaic; Gr. = Greek; HA = Hatran Aramaic; Has. = Hasaitic; JBA = Jewish Babylonian Aramaic; JPA = Jewish Palestinian Aramaic; JudA = Judean Aramaic; Lat. = Latin; LB = Late Babylonian; LBH = Late Biblical Hebrew; Lih. = Lihyanite; MA = Middle Assyrian; Md. = Mandaic; Med. = Median; MIr. = Middle Iranian; NA = Neo Assyrian; Nab. = Nabatean; NB = Neo Babylonian; NP = Neo-Persian; OA = Old Assyrian; OAkk. = Old Akkadian; OAr. = Old Arabic; OAram. = Old Aramaic; OB = Old Babylonian; OfA = Official Aramaic; OIr. = Old Iranian; OP = Old Persian; OSyr. = Old Syriac; Palm. = Palmyrene; Parth. = Parthian; QA = Qumran Aramaic; QH = Qumran Hebrew; SA = Samaritan Aramaic; Sab. = Sabaic; Saf. = Safaitic; Sum. = Sumerian; Syr. = Syriac. 2  We shall conventionally employ this expression in the chronological meaning introduced in 1969 by Joseph A. Fitzmyer to denote the phase of Aramaic roughly from 200 BC to 200 AD; for other understandings of “Middle Aramaic” in the history of Aramaic studies, see Moriggi 2012, with essential bibliography. 3  This alternative terminology has lately been advocated by Gzella 2008: 108ff and 2011: 604ff, who labels “Eastern Mesopotamian Aramaic” (in conformity with Beyer’s suggested designation “Ostmesopotamisch” [1984: 46 and 1998: 7]) the variety dominating the area between Hatra and the ancient city of Assur, which we conventionally name “Hatran Aramaic” in the present paper (“Hatran” is here the linguistic, “Hatrene” the non-linguistic adjective related to Hatra; cf. Kaizer 2006: 139 n. 2 for a similar solution). 4  All of these Aramaic varieties are covered in Hoftijzer and Jongeling 1995. No less than four lexicographical tools are devoted to the Dead Sea material, all with concise etymological information: Beyer 1984: 499–741; 1994: 301–432; 2004: 339–506; Sokoloff 2003; to these may be added the glossary for Judean Aramaic in Yadin et al. 2002: 398–404. Nabataean, albeit certainly in need of an updated dictionary complementing Cantineau 1932: 53–173 and the

© koninklijke brill nv, leiden, ���5 | doi ��.��63/9789004300156_009

Notes on foreign words in Hatran Aramaic

127

recorded in concise glossaries appended to the main text collections,5 as well as in Hoftijzer and Jongeling’s dictionary of epigraphic North-West Semitic (1995), but no specific study has been devoted so far to a discussion of foreign (loan)words in the language, with the exception of Harnack’s diffuse but today largely outdated essay (1970) on Iranian titles in the inscriptions of Hatra: this gap in Aramaic lexicology has prompted us to offer here a first assessment of the data, with some tentative etymological indications. Our starting point was of course Beyer’s excellent glossary (1998, integrated in 2013), offering some sparing etymological information, mostly implicit, and partly inferable from the hypothetical vocalization and translation provided, according to the style of presentation adopted for his masterly glossaries of Dead Sea Aramaic texts. Hatran Aramaic (HA) is documented in ca. 600 lapidary inscriptions and graffiti, a part of which are dated between 44 BC and 238 AD, belonging to different epigraphic genres, each characterized by its distinctive formulary and specialized lexis: building, votive, funerary, memorial, and legal texts, to which may be added the sub-genre of blessings and curses.6 No texts on soft media, such as are extant for Judean Aramaic, Palmyrene, and Old Syriac, are so far known for HA.7 The remarkable linguistic and orthographic uniformity of the HA corpus has brought Holger Gzella to advance the quite plausible hypothesis of the existence of a royal chancellery, whose activity imposed some degree of standardization throughout the whole area of diffusion of this Aramaic variety

somewhat disappointing Jobling 1995, has been repeatedly investigated regarding its lexical borrowings from Arabic (O’Connor 1986, Greenfield 1992, and the glossary for Nabataean in Yadin et al. 2002: 405–410) and from Greek and Latin (Healey 1995, in comparison with Old Syriac, and Monferrer Sala 2013). Palmyrene has an excellent glossary in PAT 333–421, while attention has been devoted both to its Greek and Latin (Brock 2005, in comparison with Syriac; Gzella 2006: 26–31, in a broader sociolinguistic frame) and to its Arabic loanwords (Maraqten 1998). Old Syriac vocabulary is accounted for in an essential word list in Drijvers and Healey 1999: 267–275, while its foreign loanwords are discussed by Healey 1995, in comparison with Nabataean. 5  Essentially Vattioni 1981: 111–119 (bare lexical index, including proper names without distinction); Aggoula 1991: 185–190; Beyer 1998: 169–185 with Beyer 2013: 56–61; del Río Sánchez 2006: 81–84; Healey 2009: 345–348 (lexical index and concordance). Bertolino’s treatment of foreign words in Hatran Aramaic takes just 5 lines (2008: 52), whereas Ciancaglini’s first-class monograph on Iranian loanwords in Syriac (2008) also includes a discussion of 6 Hatran words (bḥṭšʾ, dḥšpṭʾ, nḥšrpṭʾ, pdḥšʾ, pšgrybʾ, pzgrybʾ). 6  Representative annotated anthologies of HA texts have recently been contributed by del Río Sánchez 2006: 51–86, Bertolino 2008: 67–82, and, most usefully, Healey 2009: 276–310. 7  Healey 2009: 23, 27.

128

Contini and Pagano

in Parthian and early Sasanian times.8 In comparison with the varieties at the Western end and in the middle of the Middle Aramaic linguistic continuum, such as Nabataean and Palmyrene, HA displays a much more innovative linguistic profile, sharing a few features with Old Syriac, but also showing some relevant differences with this (as well as with other Eastern Aramaic local or regional varieties),9 which also affect its vocabulary, including the variable incidence of Greek and Iranian loanwords in both dialects.10 Hatra’s foreign connections were shaped chiefly by its political subordination to the Arsacid Parthian Empire, by its strategic role in international relations and trade,11 by its hosting a center of the worship of the Mesopotamian sun-god Šamaš (possibly vocalized as Šmeš in HA)12 much visited by pilgrims from the entire Eastern Jazirah,13 and lastly by having a social organization patterned according to a dimorphic model, which saw the steppe surrounding the town inhabited by “Arabs” (ʿrb), presumably speaking one or more Arabian varieties. Hatrene rulers, on the other hand, bore personal names of various linguistic affiliation: Iranian, Arabian, and Aramaic.14 Only HA seems to have been in public use in Hatra and the surrounding region in the Arsacid times. Whereas the large number of bilingual Greek and Aramaic inscriptions found at Palmyra has given rise to some discussion concerning the respective status and areas of employment of the two languages in the epigraphic habit,15 the single short HA—Greek bilingual inscription so far documented (D4), originating from Dura-Europos,16 is rather more informative for the correspondence between the Hatrene and Greek solar cults and divine names than for the sociolinguistic relationship between the two languages.17 Iranian languages, Greek, Latin, and Arabian varieties are thus the sources to be expected for words of foreign origin in HA, with the significant addition 8  Cf. Gzella 2008: 109 and n. 10 for the possibility that the operation of this chancellery may largely antedate the earliest surviving HA material. 9  Gzella 2008: 109f., with pertinent exemplification. 10  A systematic comparison between Old Syriac and HA is, however, still a desideratum. 11  Hatra would, however, not meet Millar 1998’s strict requirements for being characterized as a “caravan city.” 12  Beyer 1998: 150f. 13  Dirven 2006–2007. 14  Details in Beyer 1998: 153–167. 15  Taylor 2002; Healey 2009: 20. 16  On Hatran inscriptions found in Dura-Europos, see Bertolino 2007: 48–54; on different manifestations of Aramaic discovered there, see Gzella 2011: 605; on the coexistence of sojourners of various ethnic origins, religious persuasion, and linguistic affiliation at Dura, see now Van Rompay 2011, with essential bibliography. 17  Milik 1972: 334–337, and particularly Kaizer 2009, with copious references.

Notes on foreign words in Hatran Aramaic

129

of the lexical legacy of pre-classic Mesopotamia, essentially Akkadian, preserved in its vocabulary. 2

Loanwords from Greek, Latin, Iranian, and Arabian

Among several dozens of HA words for which a foreign origin could be argued, the following were chosen, as endowed with a higher degree of probability:18 1) ʾHL ‘family’ (Beyer 2013: 56: ʾahl ARAB‘Familie’) ← OAr. *ʾahl, cf. CA ʾahl ‘family, home’ (BK 65f.); Inscr. CPA ʾhl (Beyer 1984: 506: ‘Leute’ < Arab). 2) ʾS ‘copper coin’, also abbreviated as {ʾ} (Beyer 1998: 169: ʾas ‘eine Münze’; Hoftjzer and Jongeling 1995: 86f. ʾs1: ‘certain weight [14 gr.] / certain amount of money’) = Lat ās, assis ‘a Roman copper coin, penny, worth about one sixteenth of a denarius’, but cf. Gr. ἀσσάριον ← Lat. assarius (nummus); Palm. ʾsr ‘assarius (coin, unit of money)’ (PAT 341b; Brock 2005: 12f.); Syr. ʾassārā ‘id.’ (LS 38a); Jewish Aramaic ʾysr ‘Ass, eine römische Kupfermünze’ (Krauss 1899: 37f). In the other Aramaic dialects the source of the loan is the derivate assarius via Gr. ἀσσάριον: a direct borrowing from Latin ās seems unlikely, possibily ʾs and {ʾ} are both abbreviations for *ʾsr. 3) ʾWṬRNʾ SG.EMPH ‘veteran’ (Beyer 2013: 56: ʾuweṭrānā ‘Veteran’) ← Lat. veteranus via Gr. οὐετeρανός; cf. Palm. wṭrn (Rosenthal 1936: 92; PAT 361a; Brock 2005: 19f.); Syr. ʾwṭrnʾ (LS 8a–b), wṭrnʾ (Brock 2005: 20). This spelling with prothetic ʾ- seems preferable to Beyer’s previous reading [wṭr]nʾ in T3,5 (Beyer 1998: 174), possibly to be emended to [ʾwṭr]nʾ, while the reconstructed reading [?]ṭrnʾ in H324 could likewise—by a not overly reckless application of Occam’s razor—be emended to [ʾw]ṭrnʾ, thus obliterating the only conjectural occurrence in HA of a further loanword from Greek τίρων ‘recruit’ (Beyer 1998: 176: ṭrōnā ‘Rekrut’).19

18  Other words were excluded either because of their uncertain readings (e.g. [bry]qrʾ Beyer 1998: 171: ‘Künstler’; brmrkʾ Beyer 1998: 171: ‘Angestellter’; ḥṣʾ Beyer 1998: 176: ‘Oberfläche’) or because their being foreign borrowings appeared debatable (e.g. glp / glwpʾ ‘to carve’ / ‘sculptor’, occurring also in Palmyrene and Old Syriac, whose derivation from Gr γλύφω seemed doubtful also to Brock 2005: 11 n. 4, among others, on the grounds of the great rarity of direct verbal borrowings from Greek into Aramaic; mlʾ ‘defensive ramp, artificial terrace’ [discussed in Steiner’s masterly essay (1989) in Wörter und Sachen philology] Beyer 1998: 179: ‘Rampe’). 19  This word could however easily have been borrowed into HA, as parallels in Aramaic dialects show: see JPA ṭyrwn (Krauss 1899: 265a); Syr. ṭrwnʾ (Payne Smith 1879–1901: 1517).

130

Contini and Pagano

4) ʾMNYK ‘necklace’ (Beyer 1998: 169: ʾamnīḵ [for hamnīḵ] ‘Halskette’) possibly ← a MIr. reflex of OP *hamyānaka ‘necklace’ (cf. Ciancaglini 2008: 163f); cf. BA k. hmwnkʾ, q. hamnîḵâ ‘necklace’ (Dan 5:7) (Ellenbogen 1962: 70; HALOT 1860f; Beyer 1984: 563); Syr. hamniḵā ‘necklace’ (LS 177b); JBA məniḵā ‘id.’ (DJBA 687b); JPA mnyk ‘necklace’ (Krauss 1899: 343; DJPA 318b). This hapax legomenon (H292,7) in HA shows initial ʾ-, as against h- in Biblical Aramaic and Syriac and an absence of the initial laryngal in the Jewish Palestinian and Jewish Babylonian Aramaic forms. The h- > ʾ- shift has been described by Beyer (1998: 127) as an instance of weakening of h in HA, a phenomenon attested only twice in the corpus. The matter is further complicated by the controversial Iranian etymon(s) of the different Aramaic forms: the first set (Biblical Aramaic, Syriac) has most recently been derived by Ciancaglini from OP *hamyānaka ‘necklace’ (Hinz 1975: 114), which may ultimately share the same etymon as the other Iranian loanword in Syriac hmynʾ ‘belt, girdle’, only augmented by the diminutive suffix -ka. The Jewish Palestinian and Jewish Babylonian Aramaic forms, as well as Greek μανιάκης ‘gold necklace’, should be traced back instead to a different Old Iranian etymon, i.e. *manyaka- ‘(gold) necklace’ (Hinz 1975: 160).20 5) ʾSṬWʾ SG.EMPH ‘portico’ (Beyer 1998: 169, Beyer 2013: 56: ʾesṭwā ‘Säulenhalle’) ← Gr. στοά ‘roofed colonnade, portico’; see Palm. ʾsṭwʾ (Rosenthal 1936: 92; PAT 341a; Brock 2005: 21); Syr. ʾesṭwā (LS 32b; Schall 1960: 49); JPA sṭw (DJPA 372b), sṭyw (Krauss 1899: 379), ʾsṭbʾ (Krauss 1899: 117f); CPA *ʾsṭwʾ, pl. ʾsṭwn (LSP 15a). 6) ʾSPPṬʾ SG.EMPH ‘head of cavalry’ (Beyer 1998: 169; Beyer 2013: 56: ʾaspapṭā ‘Kavallerieoffizier’; Aggoula 1991: 185b: ‘préfet’) ← MP Inscr. ʾspptʾ = Parth. ʾsppty ‘chef de la cavalerie’ (Gignoux 1972: 17a, 47a). This loanword, which occurs twice in HA, is hitherto not documented in any other Aramaic dialect; pace Aggoula, it has nothing to do with an OP *spaḏapati (actually a ghostword, cf. Sima 1997: 321a), nor with MP spʾhptʾ /spāhbed/ ‘general; commander’, which is the etymon of Syr. ʾsphbyd, sphbyd ‘general’.21 7) ʾSTR[Y]N PL ‘staters’ (Beyer 1998: 169: ʾestērīn ‘Stater’ eine Münze; alternative reading: ʾstrʾ ‘statère’ Aggoula 1991: 185b, Hoftjzer and Jongeling 1995: 92) ← Gr στατήρ ‘stater; standard coin’; see OfA sttr, sttry ‘id.’ (Hoftjzer and Jongeling 1995: 805); sttry (Muraoka and Porten 1998: 377); Syr ʾestirā, ʾestrā (LS 38b); JBA *ʾstyr, ʾstyrʾ, ʾsṭrʾ (Krauss 1899: 98). 20  Ciancaglini 2008: 163f., with detailed history of research and a convincing refutation of Walter Belardi’s explanation of Syr. hamniḵā from OP *manyaka- by addition of an alleged “inorganic h-”, in fact unattested in Syriac. 21  Sima 1997: 321a; Ciancaglini 2008: 110f., with full bibliography.

Notes on foreign words in Hatran Aramaic

131

The adaptation of this Greek borrowing into HA follows the same lines as into Syriac and Jewish Babylonian Aramaic, differently from the forms without prothetic ʾ- and with double t in 5th cent. BC Official Aramaic. 8) ʾQLWTʾ, ʾQLTʾ SG.EMPH ‘servant, follower, acolyte; association of servants, of acolytes’ (Beyer 1998: 169: ʾaqlūṯā ‘Diener(schaft)’; Aggoula 1991: 133: ‘association’, 185b: ‘acolyte’; Hoftjzer and Jongeling 1995: 99 ʾqlt “exact meaning unknown”) prob. ← Gr. ἀκόλουθος ‘following, follower’; cf. Palm. [ʾ]qlwtʾ (Rosenthal 1936: 92; reading not accepted in Hoftjzer and Jongeling 1995 or Hillers and Cussini 1996); later Syr. ʾqlwtyʾ (ἀκολουθία) ‘ordo’ (LS 45a); JPA ʾqwlyws ‘in good order’ (← ακολούθως) (DJPA 72b). The two forms documented in HA seem to denote an individual acolyte of a princess (ʾqlwtʾ, H112,7) and an association of followers of the god Barmaren (ʾqltʾ, H280,1),22 respectively; while the semantics of this loanword is still unclear, a borrowing from Greek seems assured. 9) *ʾŠPZKN SG.CON + 3.SG suffix ʾšpzknh ‘(his/her) steward, majordomo, chamberlain’ (Beyer 1998: 170: ʾašpazkāne/ah ‘[sein/ihr] Hausverwalter’; Aggoula 1991: 136: ‘son hôtelier [intendant]’, 185b: ‘son majordome’) prob. ← Early MP *aspinjakān ‘host, inn-keeper’ (Ciancaglini 2008: 118f); see also Syr. ʾšpzknh, ʾšpyzknh ‘major-domo, host’ (LS 53b; Ciancaglini 2008: 119); JBA ʾušpizkānā, ʾwšpyzkʾnʾ, ʾyšpyzknʾ ‘landlord, host’ (DJBA 99a). This loanword, occurring twice in HA, was already ascribed to the Middle Iranian Nebenüberlieferung on the foundation of its presence in the late Syriac native lexicographers and in Jewish Babylonian Aramaic.23 10) BṬḤŠʾ, PDḤŠʾ SG.EMPH ‘high official, governor’ (Beyer 1998: 170, 182: bṭaḥšā, pḏaḥšā ‘hoher Beamter’; Aggoula 1991: 77: ‘pitiax’, 185b: ‘maître des gardes’; Hoftjzer and Jongeling 1995: 150 btḥš, 902 pdḥš ‘designation of high [Parthian] official’) ← MP and Parth. bidaxš (Ciancaglini 2008: 230f); see Later OfA Armazi btḥš (Hoftjzer and Jongeling: 150); Syr. pṭaḥšā ‘praefectus, satrapa provinciae magnae regni Sasanidarum’ (LS 564a); MP bythš /bidaxš/ ‘viceroy’ (Mackenzie 1971: 18); Inscr. MP btḥšy = Parth. bytḥš ‘vice-roi’ (Gignoux 1972: 20b, 50b); Gr. πιτάξης, βιτάξης; Lat. vitaxa (Amm. Marc. 23, 6, 14). Attested as a loanword in a wide range of languages and transcriptions as well as in multiple allotropes (as the HA doublets also witness), the Middle Iranian title bidaxš is problematic as regards both its etymon and its semantics:

22  Sima 1997: 320b even considers the possibility that the whole phrase bny ʾqltʾ dy br mryn may be a tribal name. 23  See Hinz 1975: 47 and above all Ciancaglini 2008: 118f. for a full survey of the evidence.

132

Contini and Pagano

in the Arsacid period it probably denoted ‘a governor, a plenipotentiary representative of the king of kings in the great districts of the kingdom’.24 11) GZBRʾ SG.EMPH ‘treasurer’ (Beyer 1998: 172: gazzaḇrā ‘Schatzmeister’) ← OIr. *ganzabara- ‘treasurer’ through an OfA forerunner gzbrʾ, gyzbrʾ (Ciancaglini 2008: 27; Zadok 2010: 756); see OIr. (Med.) *ganzabara- (Hinz 1975: 102; Tavernier 2007: 422); Parth. gnzbr, gznbr = Inscr. MP gnzbwr ‘trésorier’ (Gignoux 1972: 51, 22b); LB ganzabāru ‘Schatzmeister’ (AHw 281a); OfA gnzbrʾ (Harnack 1970: 547f); BA gizzaḇrayyâ ‘the treasurers’ (Ezra 7:21) (Ellenbogen 1962: 55; Beyer 1984: 544; 1994: 339; HALOT 1843f); Syr. gzbrʾ, gyzbrʾ gezzaḇrā, gnzwrʾ ganzurā (= reborrowing, Ciancaglini 2008: 142) (LS 111a, 125a); JBA gizbərā (DJBA 273b); Md. ganzibra ‘treasurer; the ecclesiastical rank above that of tarmīda (priest)’ (MD 77a–b).25 12) GNS ‘clan’ (Beyer 1998: 172: gnès ‘Sippe’) ← Gr γένος ‘race, stock, kin, kind, clan, family’; cf. Palm. gns ‘kind, way’ (Rosenthal 1936: 92; Hoftjzer and Jongeling 1995: 230 [gns], PAT 354a; Brock 2005: 13); Syr. gensā ‘stock, kind, family’ (LS 125b; Schall 1960: 89f and 119); JBA ginsā ‘type, sort’ (DJBA 297a); etc. 13) *DḤŠʾ PL.CON + 3.M.SG suffix dḥšyhy ‘(his) guards’ (Beyer 1998: 172: daḥšēh ‘seiner Wächter’; Aggoula 1991: 186b: ‘gardien, soldat’) ← Parth. *daxš‘guard, executioner’ (Ciancaglini 2008: 150); cf. Inscr. MP dḥyc ‘bourreau’ (Gignoux 1972: 21a); Syr. daḥšā ‘guard, executioner’ (LS 149b). The Parthian etymon of this HA loanword, occurring only twice, is reconstructed on the basis of the following. 14) DḤŠPṬʾ SG.EMPH ‘chief of the guardsmen, executioner’ (Beyer 1998: 172: daḥšapṭā ‘Ordnungshüter’; Aggoula 1991: 186b: ‘chef [maître] des gardes’) ← Parth. *daxšpat ‘chief of the guard’ (Greenfield 1987: 258); cf. Arm. dahčapet ‘chief guard, executioner’. In HA this Parthian title (a compound consisting of *daxš- ‘guard’ + pat ‘lord, master’)26 is attributed to Nergal, god of the underworld, when it is supposed to mean ‘executioner’.27

24  See Hinz 1975: 47 and above all Ciancaglini 2008: 118f. for a full survey of the evidence. 25  A very old Iranian loanword in Semitic; see the extensive Iranian and Aramaic evidence in Ciancaglini 2008: 59, 142f. 26  Greenfield 1987: 258; Kaizer 2006: 143, 147; Ciancaglini 2008: 150; this interpretation seems semantically preferable to the alternatives put forward by Harnack 1970: 546f. (‘Herr der Diener’) and Pennacchietti 1996: 63f. (‘δεκαδάρχης, commander of a decuria, decurion’, whence in HA a semantic extension to ‘Capitano delle Milizie’ or some such title in the Parthian army or court hierarchy). 27  Greenfield 1989, with full discussion.

Notes on foreign words in Hatran Aramaic

133

15) *DNR PL.EMPH DNRʾ ‘denarius, gold denar’ (Beyer 1998: 174: dēnārē ‘Denar’) ← Lat. denarius via Gr. δηνάριον; cf. Palm. dnr, dynr, SG.EMPH dnrʾ, PL.ABS dnryn (Rosenthal 1936: 91; PAT 356b; Brock 2005: 14); JudA dynr, PL.ABS dnryn (Hoftjzer and Jongeling 1995: 256 [dnr]); OSyr. EMPH dynrʾ (Healey 1995: 81); Syr. dinārā (LS 160b; Schall 1960: 41 and 98); JPA dynr (Dalman 1897–1901: 101f; Beyer 1984: 557; DJPA 197f; Beyer 1994: 335); CPA dynrʾ (LSP 45b); JBA dynʾrʾ (DJBA 334a–b); Md. dynʾrʾ (MD 108b); also Aramaic → CA (Fraenkel 1886: 191f); etc.28 16) DRYKNʾ PL.EMPH ‘darics, Persian gold coins’ (Beyer 1998: 174: dārēḵānē ‘Golddareikos’; Aggoula 1991: 143 ‘deniers or [dariques]’; Hoftjzer and Jongeling 1995: 261: drykn ‘Persian gold coin, exact type and value in Hatra unknown’) ← OP *dārayaka- ‘gold coin introduced by Darius’ (Hinz 1975: 83), possibly via Gr δαρεικός, δαρικός; cf. LBH PL ʾăḏarkōnîm ‘darics’ (Ezra 8:29) (Ellenbogen 1962: 17f, HALOT 17a); Syr. driḵonā ‘daricus’ (LS 167a). The HA (a hapax legomenon) and Syriac forms, as well as the Hebrew one (with prothetic ʾ-)29 seem to be adaptations of the Iranian borrowing in Greek δαρεικός, δαρικός. The -ōn- segment in Hebrew and Syriac is usually explained as derived from the Greek GEN.PL ending -ων,30 though an alternative hypothesis of the morphological integration of the loanword could consider the Greek case-ending replaced by the North West Semitic derivational suffix -ōn/-ān, which would account also for Beyer’s reconstructed suffix *-ān in the HA form. 17) HDRPṬʾ SG.EMPH ‘chiliarch, commander’ or ‘fire priest’ (Beyer 1998: 174: hḏārpṭā ‘Oberst’; Harnack 1970: 496ff: hērpat ‘Feuerpriester’; Aggoula 1985: 18: ‘maître du feu [?]’; DNWSI 270f: hdrpṭ prob. ‘fire priest’); 1. see OP *hazārapati(Hinz 1975: 120); MP hzʾlpt /hazārbed/ ‘chiliarch’, Inscr. MP hzʾlwpt, Parth. hzrwpt (Ciancaglini 2008: 161f); Syr. hazārpāṯ ‘chiliarch, lord of a thousand’ (LS 174a; Ciancaglini 2008: 161f); 2. see MIr. Nisa 280,1 ʾtwršpty ← OIr. *atropati- ‘lord of fire’ (interpreted as compounded of OIr. *ātar ‘fire’ + *pati- ‘lord’) (Harnack 1970: 496ff). By his translation Beyer seems implicitly to advocate a connection with Syr. hazārpāṯ ‘chiliarch’, which has a consistent Iranian pedigree.31 The very short HA text containing this hapax legomenon (H83,2),32 though, is inscribed on the base of a fire altar, which gives some contextual plausibility to the alternative 28  See also De Romanis 2004. 29  See Schwyzer 1931: 15. 30  First suggested by Schwyzer 1931: 14ff., accepted, e.g., by HALOT 17a. 31  Albeit there is a fierce controversy on its precise etymological connection within Iranian; full survey of the evidence and critical discussion in Ciancaglini 2008: 161–162. 32  A second occurrence in H362,2 is highly uncertain.

134

Contini and Pagano

interpretation, already put forward by the inscription’s editor princeps Fuad Safar and argued in Iranian philology by Harnack,33 which is however rather weakened by the unusual initial h- in the HA form.34 The presence of -d- in the HA form is still unexplained according to either etymological solution. 18) ZNPṬʾ SG.EMPH ‘master of arms’ (Beyer 1998: 175: zēnapṭā ‘Waffenmeister’; Aggoula 1991: 187a: ‘maître d’armes’) ← OIr. (Med.) *zainapati(Hinz 1975: 276) (cf. Gr. Σινιπάτης); Inscr. MP zynpt = Inscr. Parth. zynpty ‘chef de l’armement’ (Gignoux 1972: 39b, 68b); the first element of this compound was borrowed in Aramaic at an early date: cf. OP (Med.) *zaina- (Hinz 1975: 276; Ciancaglini 2008: 174); OfA zynʾ ‘weapon’ (DNWSI 319 [zyn2]); Syr. zaynā ‘id.’ (LS 195a; Ciancaglini 2008: 174); JBA zênā, zaynā ‘id.’; Md. zaina ‘weapon, armour, equipment; fetter(s)’ (MD 158a); JPA zyyn (DJPA 175b); CPA zyn (LSP 56a). 19) ḤṬMʾ PL.EMPH ‘camel-muzzles’ (Beyer 1998: 175: ḫṭāmē pl. ‘Kamelhalfter’; Healey 2009: 300f: ‘camel-halter’; Aggoula 1977: 182: ‘masse qui sert à casser les blocs de pierre’; Degen 1978: 68f: ‘Schutzdach’; Aggoula 1991: 187a: ‘masse’; DNWSI 364: ḥṭm poss. ‘sledge-hammer’) ← OAr. *ḫiṭām, cf. CA ḫiṭām ‘muzzle’ (BK 596b); Sbaʿa Bedouin Arabic ḫṭāme ‘camel’s leather muzzle’ (de Boucheman 1934: 42); Najdi Arabic ḫṭām ‘noseband, halter; the entire halter used for riding a camel’ (Kurpershoek 2005: 77). In the context of H281, a decree against the theft of objects “from the work of Barmaren,” this noun has been interpreted mainly by deriving it from the Semitic roots—both well documented in Arabic—ḤṬM ‘to break, to shatter, to wreck’ and ḪṬM ‘nose’.35 The second root, corresponding to Beyer’s and Healey’s choice, is indeed documented in Aramaic (cf. Targumic Aramaic ḥuṭmā ‘nose, nostril, snout’; Jastrow 1903: 431b), but not with this technical meaning, consistent with Bedouin material culture, which would as a working hypothesis qualify this hapax legomenon in HA as borrowed from Old Arabic, rather than as a cognate of the Arabic forms listed above. As Healey pertinently remarks, in H281,4 this translation fits the preceding word ‘tents’ better than the following list of tools, but its intermediate position between the two kinds of material referents does not make these conclusive grounds for rejecting this interpretation. Alternatively, however, it would be possible to connect the word with the √ḥṭm and reconstruct a substantive *ḥaṭṭām ‘sledgehammer, mallet’, thus going back to Aggoula’s and DNWSI’s interpretation as denoting a 33  Harnack 1970: 496–508; see also DNWSI 270f. 34  Greenfield 1989: 136 and n. 10, with further Iranological references. 35  The exception being Degen’s proposal, which is however not supported by a convincing etymological connection.

Notes on foreign words in Hatran Aramaic

135

building tool: this would still be compatible with classifying this noun as borrowed from Old Arabic, though an Aramaic cognate could not be excluded. 20) ḤYNʾ SG.EMPH ‘time’ (Beyer 1998: 175: ḥīnā ‘Zeitpunkt’) ← OAr. ḥīn, cf. CA ḥīn ‘time, hour, epoch, moment, season’ (BK 528b); cf. also JudA (Babatha Archive) ḥynh ‘(seine) Zeit’ (Beyer 2004: 396). Segal (1986: 71)’s classification of this word as an Arabic loanword seems justified and was indeed accepted by Beyer (1998: 111); this is also attested among the recently discovered Arabic loanwords in Judaean Aramaic documents of the Babatha archive. 21) *ṬWHM SG.CON + 3.M.SG suffix ṭwhmh ‘(his) clan’ (Beyer 2013: 58: ṭohmeh ‘(seine) Sippe’) ← OP *tauma- (= tauhmā) ‘race, stock, family’ (Ciancaglini 2008: 182f, cf. Hinz 1975: 235); cf. Arm. loanword. tohm; MP twhm, twm /tōhm, tōm/; Inscr. Parth twḥm ‘race, semence’ (Gignoux 1972: 65b);36 Syr. ṭwhmʾ ṭohmā ‘race, stock, family, descent, origin’ (LS 268b; Ciancaglini 2008: 192f.; Zadok 2010: 756); Md. ṭuhma ‘race, lineage, stock, origin, seed, kin, etc.’ (MD 177a). This is one of the Old Persian loanwords that reached Syriac (and doubtless also HA) through (so far unattested) Official Aramaic according to Ciancaglini (2008: 27). 22) KLYLYʾ SG.EMPH ‘tax collector’ (Beyer 1998: 78: klīlāyā ‘Eintreiber der Kronensteuer’); alternative interpretations listed in DNWSI 512: 1. ‘the man of Klyl’, 2. ‘the one of the crown / garland / turban’, 3. ‘the one of the arch (i.e. its sculptor)’; cf. Syr. klilā ‘crown, diadem’ (LS 327a–b); JudA and JPA SG.EMPH klylʾ ‘crown’ (DJPA 260a), PL.EMPH nisba adj. klylyʾ (Beyer 1984: 607: ‘Eintreiber der Kronensteuer’); JBA klilā 1. ‘crown’, 2. ‘decorated headband’, 3. ‘wreath, garland’, 4. ‘coronation tax’ (DJBA 582f); etc. Beyer’s translation of this hapax legomenon in a dedication (H253,2) seems to presuppose—as in the Jewish Aramaic occurrence quoted in Beyer 1984: 607—a substantivated nisba-adjective derived from a meaning of HA klīlā calqued on the Latin (aurum) coronarium,37 as is recorded above for Jewish Babylonian Aramaic (dmy klylʾ ‘payment of the coronation tax’, lit. ‘payment of the crown’): though not a borrowing, this would show crucial semantic influence from a foreign word. 23) KSYDʾ PL.EMPH ‘stone-cutting tool’ (Beyer 1998: 177: ksīḏē “Stemmeisen”; DNWSI 523 ksyd subst. indicating certain tools: ‘stone-mason’s axe, hatchet, paring-chisel’; Kaizer 2006: 141f.: ‘pickaxe’) prob. ← Gr. ξοίς / -ίδος ‘chisel’, cf. Syr. ʾaḵsīḏā ‘culter lapicidae’ (LS 18b). 36  Further Iranian cognates, with abundant bibliography, in Ciancaglini 2008: 182f. 37  On this irregular form of taxation see Millar 1977: 140ff.

136

Contini and Pagano

24) KRNYTʾ sg.emph ‘general’ (Beyer 1998: 178: kārānyāṯā ‘Heerführer’) possibly ← a compound of OP *kāra- ‘army’ (Hinz 1975: 147ff; Tavernier 2007: 557) + a cognate of OIr. *naitā, NOM.SG of *naitār = Vedic netār- ‘leader, guide’ (St. Zimmer, apud Sima 2000: 165). So far, the only etymological proposal concerning this Iranian loanword in HA has been advanced by St. Zimmer, in Sima 2000: 165; it may well have been imported already into Official Aramaic. 25) MKLʾ PL.EMPH ‘bars, bolts’ (Beyer 1998: 179: moḵlē ‘Riegel’; DNWSI 624: mkl2 ‘bar, lever’; Kaizer 2006: 141f with n. 19: ‘bars’) ← Gr. μοχλός ‘lever, crowbar, bolt’; cf. Syr. moḵlā ‘bar’ (LS 385a: ‘obex, sera’; Schall 1960: 50 and 103; Healey 1995: 80; Healey 2009: 301); Jewish Aramaic *mwgls ‘Riegel’ (Krauss 1899: 324); also Aramaic → CA (Fraenkel 1886: 17). The occurrence of this hapax legomenon in the list of building tools in H281,6 fits better this meaning, documented in contemporary literary Syriac, than the alternative interpretation as some kind of ‘measure’ suggested by Degen 1978: 68ff, probably on the analogy of Palm. mkl (DNWSI 624 mklⁱ), which however would seem to denote a liquid measure.38 26) NḤŠRPṬ[ʾ] SG.EMPH ‘master of the hunt’ (Beyer 1998: 180; 2013: 52, 54: naḥšīrpṭā ‘Jagdherr’; Greenfield 1987: 258: ‘in charge of the hunt’; Aggoula 1991: 188b: ‘maître des guerriers’; DNWSI 726: nḥšrpṭ ‘master of the hunt’) ← Inscr. Parth. nḥšyrpty (= Inscr. MP nḥcyrpt) ‘maître de chasse’ (Gignoux 1972: 59 and 30; Sima 1997: 318); for the first element of the compound, see OP *naxačarya(?) (Hinz 1975: 172); MP naḥčīr (Mackenzie 1971: 58; Nyberg 1974: 136a); QH nḥšyr ‘Blutbad, Gemetzel’ (Beyer 1984: 636; 1994: 381); Syr. naḥširā ‘hunting; wild beasts; hunter’ (LS 424a–b; Ciancaglini 2008: 213); Md. našira ‘chase, quarry’ (MD 287a). The right interpretation of this Iranian loanword in HA had already been put forward by Safar and Caquot and its etymology sketched by Harnack, before being better defined in light of a more precise assessment of the Iranian evidence.39 27) SMʾ SG.ABS, SMYʾ SG.EMPH, SMYTʾ PL.EMPH ‘(the) cultic standard(s)’ (Beyer 1998: 150; 2004: 59: sēmē, sēmyā ‘(die) Götterstandarte, Kultstandarte’; Aggoula 1991: 189a ‘remède(s)’; DNWSI 790f: s.v. smʾ ‘standard’) ← Gr. σημεῖον or σημεία ‘military standard’; cf. Palm. smy SG.ABS, smytʾ < σημεῖον ‘figure, image’ (Rosenthal 1936: 91; PAT 392b; Brock 2005: 20); Syr. SG simiyon, PL simyʾā

38  PAT 381a: ‘measure of wine’; Kaizer 2002: 191 and n. 96. 39  Harnack 1970: 544f.; Greenfield 1987: 258; DNWSI 726 (nḥšrpṭ); Ciancaglini 2008: 213, with full survey of the Iranian data.

Notes on foreign words in Hatran Aramaic

137

‘signum, milliarium’, etc. (LS 480a); JPA smh, PL.EMPH smhwtʾ ‘Feldzeichen’ (Beyer 1984: 647).40 Rather than Aggoula’s proposal of a connection with Aramaic samm ‘medicine, poison’, a plausible alternative to an Aramaic borrowing from Greek would be the influence of Greek words like σῆμa, σημεῖον etc. on an original Semitic word derived from the √swm (/sym) ‘to set up’, as suggested in DNWSI 791. 28) PZGRYBʾ, PŠGRYBʾ SG.EMPH ‘crown prince’ (Beyer 1998: 182f: pzaḡrīḇā, pšaḡrīḇā ‘Kronprinz’; DNWSI 904 [pzgryb], 946 [pšgrb]; Aggoula 1991: 189b: ‘lieutenant, locum tenens’) ← MIr. *pačāγrīw or *pašāγrīw (Ciancaglini 2008: 235f); cf. OSyr. pṣgrybʾ ‘crown prince’ (Healey 1993: 81); Syr. pṣgrybʾ, bṣgrybʾ ‘magistratus persicus’ (LS 586b and 86a; Ciancaglini 2008: 235f). Before being found also in HA, this Parthian title was attested in Old Syriac (2nd cent. AD) inscriptions and, in two different allotropes, very early in Classical Syriac (e.g., in the Book of the Laws of Countries of the school of Bardesanes and in the Hymn of the Pearl).41 Its precise Middle Iranian etymon is still controversial,42 while its political meaning has been lately specified by Tommaso Gnoli as ‘heir to the throne, crown-prince’, rather than ‘representative of the king’ or ‘viceroy’.43 29) PLṬYʾ PL.EMPH ‘palace servants’ (Beyer 1998: 182: plāṭāyē ‘Bediensteter des Schlosses, palatium’; DNWSI 915: plṭy³ PL.EMPH plṭyʾ: substantivated adj. ‘traveller’ [Aggoula]) ← Gr. παλάτιον + adjectival suffix + PL.EMPH ending; cf. Syr. palāṭin ‘palace’ = παλάτιον palatium (LS 574b); JPA palāṭīn ‘palace’ (< παλάτιον, Lat. palatium) (Krauss 1899: 457; DJPA 435b). This hapax legomenon in G2,4 has been convincingly interpreted by Pennacchietti as a substantivized nisba adjective derived from Lat. palatium via Gr. παλάτιον ‘palace’,44 a solution also accepted by Beyer: this adapted loanword would have a parallel in a form like epigraphic Byzantine Greek παλατῖνος 40  Further Aramaic forms are discussed by Tubach 1986: 190–191 and Dirven 2005. 41  Poirier 1981: 212–223 (with abundant bibliography); Tubach 1986: 248 n. 60; Greenfield 1987: 258. 42  MIr. *pačāγrīw or *pašāγrīw < OIr. *pašă- or *pasčā-grīwa- lit. ‘behind / instead of the self’ (i.e., of the king), according to Ciancaglini 2008: 235f., who also gives extensive bibliographical data on the discussion of the phonological problems regarding the Middle Iranian form(s) that was the model of the Aramaic loanwords. 43  Gnoli 2002; the same interpretation was advocated, on philological grounds, by Harnack 1970: 516–519, who also suggested that the Middle Iranian compound may have been calqued on Greek διάδοχος. Aggoula’s translation still reflects the older interpretation: cf. Sima 1997: 318a. 44  Pennacchietti 1988: 143: *pallāṭāyē ‘gli addetti al palazzo, il personale, la servitù del palazzo’.

138

Contini and Pagano

‘palace official’, recorded in Glare and Thompson 1996: 236a. Aggoula’s alternative interpretation ‘traveller’ seems semantically rather less likely, being apparently founded on the √plṭ meaning ‘to escape, to flee’ in Aramaic, and ‘to come unexpectedly upon someone’ in Classical Arabic (IV. form). 30) PRSDʾ PL.EMPH ‘anteroom, colonnade’ (Beyer 1998: 103, 183: prosdē ‘Vorhalle[n]’ [προστάς]; DNWSI 941: prsd: SG.EMPH prsdʾ ‘column’ [< Gr. παραστάς ‘id.’ rather than < προστάς ‘Vorhalle’]) prob. ← Gr. προστάς ‘vestibule, porch, portico’ or παραστάς ‘pilaster, doorpost, space enclosed between the παραστάδες, vestibule or entrance of a temple or house’; cf. Syr. presṭē ‘peristylia’ (LS 601a) (< παραστάς Healey 2009: 281) and prwsṭdʾ < προστάδα ‘Vorzimmer’ (Krauss 1899: 484a), ‘porticus’ (LS 601a; Schall 1960: 57). The precise Greek etymon of this loanword, a hapax legomenon in HA, is still unclear: actually, the definitions of προστάς and παραστάς provided by Liddell and Scott, quoted above, allow sufficient semantic overlap to be both compatible with the architectonic context of the HA occurrence. 31) PRPYṬʾ PL.EMPH ‘wanderers’ (?) (Beyer 1998: 13, 182: perpīṭē ‘die Wandelnden [περιπατοῦντες, eine religiös-philosophische Gruppe?]’) possibly ← Gr. περιπατοῦντες ‘wanderers’; cf. Later Syr. prpṭn, prypṭwn etc. περίπατος ‘peripatus; peripateticus’ (LS 604a). The context of this hapax legomenon in A11b,2 (read differently by Aggoula 1985: 32) is difficult: the first sentence is ingeniously translated by Beyer (1998: 13) as follows: kpʾ dy ʾḥtw prpyṭʾ ‘Der Stein (Material), welche die Wandelnden hergebracht haben’. Though it is hardly conceivable that the adherents of a religious-philosophical community would transport stones to build a temple in Assur, no other alternative interpretation springs to the mind, despite the high number of Greek nouns prefixed by παρα-, περι- or προ- followed by a π and a τ to be found in the Greek lexica; neither do the rich lists of Greek words borrowed into Syriac seem to be enlightening in this respect. 32) *TG tgh ‘crown’ + 3.F.SG suffix (Beyer 1998: 61: tāḡah ‘ihre Krone’) ← MIr. *tāγ ‘crown’ (Ciancaglini); cf. Palm. SG.EMPH tgʾ ‘crown’ (PAT 418b; DNWSI 1203 [tg¹]); Syr. tāḡā ‘id.’ (LS 815b); JBA tāḡā ‘id.’ (DJBA 1193f); JPA tgʾ ‘id.’ (Beyer 1984: 721); OAr. tg, CA tāǧ ‘crown, diadem’ ← Aramaic (Fraenkel 1886: 62); MP tāg ‘id.’ (Nyberg 1974: 189a); NP tāǧ ‘id.’ (Ciancaglini 2008: 266f). A hapax legomenon in HA, this Iranian loanword was already known in Palmyrene, Syriac, and Eastern Aramaic dialects, as well as in Arabic, since its earliest attestation in the Nemāra inscription.45

45  As rightly pointed out by Siddiqi 1919: 84; cf. also Asbaghi 1988: 67.

Notes on foreign words in Hatran Aramaic

3

139

Words of Akkadian Origin

Several HA words could be credited with an Akkadian etymon. We have chosen the following as more likely instances:46 1) ʾBWLʾ SG.EMPH ‘city gate’ (Beyer 1998: 169; 2013: 56: ʾəḇollā ‘Tor’) ← Akk. abullu ‘city gate’ (from OA, OB on; CAD A/1 82ff; AHw 8f; Zimmern 1917; 14; Kaufman 1974: 32; Marrassini 1971: 126f; DNWSI 5 [ʾbl3]), cf. Palm. ʾblʾ ‘doorway (of a tomb)’ (Rosenthal 1936: 90; PAT 334a); Syr. ʾbwlʾ ‘portico’ (SL 3a); JBA ʾbwlʾ1 ‘city gate’ (DJBA 74a); Md. ʾbwlʾ ‘gate, gateway’ (MD 3b). 2) ʾPKLʾ SG.EMPH ‘high priest’ (Aggoula 1991: 354: ‘juge’; Beyer 1998: 169: ʾpkallā rabbā ‘Oberpriester’) prob. ← Sum. via Akk. apkallu ‘wise man’ (Bab., NA; CAD A/2 171ff; AHw 58a–b; Zimmern 1917: 29; Kaufman 1974: 34; DNWSI 95f. [ʾpkl]), cf. Palm. ʾpklʾ (Rosenthal 1936: 90; PAT 342b);47 Nab. ʾpklʾ (Cantineau 1935: 66; Healey 1993: 154; Sima 2000);48 ANA (Lih., Has.) (Sima 1997: 318–319; Sima 2000: 449); Sab. ʾfklt hapax legomenon RES 3945/16 ‘diviner’ (Beeston 1994: 40). The only instance in Ancient South Arabian is the plural or collective ʾfklt of Sabaic, pace Healey (1993: H 16,8) who quotes Hommel’s Chrestomathie for a Minaic example: today this is classified as Hasaitic, one of the Eastern dialects of Ancient North Arabian (Sima 1999: 449). Maraqten (2000: 263f), following Cantineau (1935: 150) and Kaufman (1974: 34), considers this word as 46  We excluded two words as uncertain on semantic and / or phonological grounds: 1. KṢR PL.EMPH ‘straw-knots’ (?) (Beyer 1998: 178: keṣrē ‘Strohknoten’) hypothetically ← Akk. kiṣrum; translated ‘straws’ by Ibrahim and Segal (1982: 113, 1986: 74), who see it as an equivalent to Syr. qṣrʾ with shift q > k; according to Aggoula (1983: 33) it would be the collective noun *ksārā, not attested in Syr. with the meaning of ‘pebbles’, and attributable to CA kusāra ‘pebbles, fragments’, rather cavalierly disregarding the shift from Arabic s to HA ṣ; questionable—because of the considerable phonological and orthographic difficulties—seems also the etymological hypothesis of Pennacchietti (1988: 52), who translates the noun ‘beams’ connecting it to Syr. kāšurā, kšurā, kāššurā ← Akk. gušūru; HA kṣr undoubtedly denotes some kind of building material, at present impossible to identify. 2. ṬPʾ PL.EMPH ‘?’ hapax legomenon H 290,4 (Beyer 1998: 176: ṭappē ‘Konsole’) ← hypothetically Akk. t/ṭuppum (?). Aggoula (1991: 140) translates the lexeme ‘platform’, Beyer as ‘Konsolen’, both well suited to the architectural context of the inscription; the solution suggested by Sima (2000: 166)—“inscription in whose furrows lead was poured”—appears to be less consistent with the context and with the immediately preceding mention of arches: however, inscriptions carved on a plaque in limestone intended to be subsequently filled with lead are well attested in Hatra. 47  For the attestations of this priestly title in Palmyra, see Kaizer 2002: 237. 48  For the evidences of ʾpkl/ ʾpklt in Hegran Nabatean, see Alpass 2013: 130.

140

Contini and Pagano

a very early loan from Akkadian into Arabian and through this in Palmyrene. The Akkadian loans in West Semitic are sometimes easy to recognize: ʾpklʾ, for example, displays a quadrisyllabic structure which is distinctively not Aramaic. Scholarly opinions vary regarding the semantic difference between HA ʾpklʾ and the more frequent kmrʾ, whose root is common Semitic: ʾpklʾ is apparently tied to the cult of Šamaš (ʾpklʾ rbʾ dšmš ʾlhʾ) and therefore reserved for the royal house, denoting a high priest, maybe the foremost religious authority of the nation, while kmrʾ might have been a commoner title familiar to the Arabs of the city (Segal 1986: 66); Teixidor (1966: 91–93) sees in ʾpklʾ the evidence of a theocratic power (on the model of South Arabian mukarrib) preceding the beginning of the monarchy with Sanatruq I; Aggoula (1986: 353–54) takes the word as meaning ‘judge’, but a shift from the semantic sphere to which the lexeme belongs in the various Aramaic and Arabic dialects considered above does not seem necessary. In Nabatean and Safaitic it is attested in personal names (Nab. ʾfklw: Winnett and Reed 1970, 148 n. 34; Saf. ʾfkl and ʾfklt: see Harding 1971: 59). 3) ʾRDKLʾ SG.EMPH ‘mason’ (Beyer 1998: 170; Beyer 2013: 56: ʾardēklā ‘Baumeister’) ← Akk. arad-ekalli ‘a designation of a specific profession’ (NB only; CAD A/2 210f. 2; AHw 1465 B4; Zimmern 1917: 26; Kaufman 1974: 35; DNWSI 103f. [ʾrdkl]), cf. OfA ʾrdyklʾ ‘builder’ (Muraoka and Porten 1998: 33 and 376); Syr. ʾrdyklʾ ‘builder’ (SL 96a); JBA ʾrdyklʾ, ʾrdkl ‘builder, stone-mason’ (DJBA 164b); Md. ʾrdklʾ ‘architect, master-mason, builder, compiler’ (MD 36f). 4) ʾŠYTʾ SG.EMPH ‘(city) wall’ (Beyer 1998: 170: ʾāšīṯāʾ, PL + 3.F.SG suffix ʾāšyātah ‘Mauer’) ← Akk. asītu, asa‌ʾittu ‘wall’ (MA, NA; CAD A/2 332f.; AHw 74a–b; Zimmern 1917: 14; Kaufman 1974: 37; DNWSI 123 [ʾšyt], cf. Syr. ʾšytʾ ‘column; pillar’ (SL 106f); JBA ʾšytʾ1 ‘wall’ (DJBA 172f); Md. ʾšytʾ ‘wall’ (MD 40b); Aramaic → CA ʾasiyah ‘column’ (Fraenkel 1886: 11). There is no phonological evidence to prove borrowing from Akkadian, but the linguistic distribution of the word points in that direction. Variation between sibilants is well attested in Neo-Assyrian, where written {s} was pronounced as š and written {š} as s. The evidence for this alternation comes predominantly from foreign loanwords in Neo-Assyrian and, conversely, from Neo-Assyrian words in other languages (Luukko 2004: 74). 5) ʾŠKPʾ SG.EMPH ‘leatherworker’ hapax legomenon H212,2 (Beyer 1998: 170: ʾaškāpāʾ ‘Lederarbeiter’) ← Sum. via Akk. aškāpu ‘leatherworker’ (from OA, OB on; CAD A/2 442f; AHw 81a; Zimmern 1917: 28; Kaufman 1974: 39; DNWSI 123 [ʾškp]), cf. Syr. ʾškpʾ ‘shoemaker’ (SL 107a); JBA ʾwškpʾ ‘shoemaker’ (DJBA 98a–b); Aramaic → CA ʾiskāf ‘shoemaker’ (Fraenkel 1886: 256). Aggoula (1991: 105) proposes a different interpretation for ʾškpʾ, connecting it to the Syriac verb škaf  ‘to coat; daub’. According to this view, the word should

Notes on foreign words in Hatran Aramaic

141

denote the craftsman charged with coating the walls with mortar. This second hypothesis is, however, rather weakened by the absence of nomina professionis prefixed with ʾ- elsewhere in Aramaic. 6) ʾŠLʾ PL.EMPH ‘tow-rope, measuring string’ hapax legomenon H462,3 (Beyer 2013: 56: ʾašlā ‘Meßschnüre’) ← Akk. ašlu ‘rope’ (from OA, OB on; CAD A/2 447ff. A; AHw 81b; Zimmern 1917: 35; Kaufman 1974: 39; DNWSI 124 [ʾšl1], cf. OfA ʾšlw ‘measuring rope’ (Muraoka and Porten 1998: 376); Syr. ʾyšlʾ ‘rope’ (SL 39a); JBA ʾšlʾ1 ‘rope, a measure of length’ (DJBA 173a–b); Md. ʾšlʾ ‘draw string, rope for pulling, tow rope, cable; path (?)’ (MD 41a); Aramaic → CA ʾašl ‘rope’ (Fraenkel 1886: 93). Though the word may actually be a cognate in Aramaic and Akkadian, the latter could have influenced its semantics in Aramaic (appui, in Y. Malkiel’s terminology). 7) ʾŠPRʾ SG.EMPH ‘weaver’ hapax legomenon H283,2 (Beyer 1998: 170: ʾešprā ‘Weber’) ← Akk. išparu ‘weaver’ (from OB on; CAD I/J 255f; AHw 397a; Zimmern 1917: 27; Kaufman 1974: 59; DNWSI 125 [ʾšpr]), cf. Syr. ʾšprʾ ‘tailor’ (SL 108a); JBA ʾšprʾ ‘mender of garments’ (DJBA 174a). The identity of meaning between the HA and Akkadian forms resulting from Beyer’s translation rather militates in favor of an Akkadian loanword, though the alternative of cognate words cannot be excluded. Less probable are Safar’s different translation “blanchisseur, celui qui nettoie les habits” (1971: 69) and Degen’s reading the word as a personal name (1978: 74). 8) BBYʾ SG.EMPH ‘gate-keeper’ hapax legomenon H413c,3 (Beyer 1998: 170: bāḇāyā ‘Türhüter’) nisba adjective ← Akk. bābu ‘opening, doorway, gate’ (OA on; CAD B 14ff; AHw 95a–b; Zimmern 1917: 30; Kaufman 1974: 40f.; DNWSI 142f. [bby]), cf. OfA bb, bbʾ, bbh ‘gate; (accounting) entry’ (Muraoka and Porten 1998: 376); Palm. bbʾ ‘gate, doorway’ (PAT 345a), Magical Syr. bbh ‘door’ (Naveh and Shaked 1993: 265; Moriggi 2014: 225); Md. bbʾ 1. ‘gate, door, entrance, portal’, 2. ‘sect’, 3. ‘versicle, paragraph, verse, portion, part’ (MD 45a–b). Substantivized nisba adjective, denoting either the origin of the person designated in the epigraph ‘from Bābā’ (there are many known place names containing the element bāb ‘door’: in the Tabula Peutingeriana -Section XIHatra is connected with two main itineraries or routes, one of which leading to various towns among which appears the name Baba),49 or his craft: ‘gatekeeper’, according to Beyer’s interpretation, or ‘banker’ according to Aggoula’s interpretation (1990: 417). This lexeme for ‘door’ is maintained only in Eastern dialects of Aramaic (including eastern varieties of Official Aramaic represented in Elephantine), whereas it is absent from Classical Syriac where tarʿā is 49  Miller 1916: 771.

142

Contini and Pagano

found.50 See also H202i, k: tarrāʿē/tārōʿē ‘gate-keeper’. In this view, bbʾ might be an Eastern Aramaic lexical feature inherited from the Akkadian substratum. 9) BYRTʾ SG.EMPH ‘castle’ (Beyer 1998: 170: bīrtā ‘Burg’) ← Akk. birtu ‘citadel, castle, fort’ (from OB on; CAD B 261ff; AHw 129 I; Zimmern 1917: 14; Kaufman 1974: 44; Marrassini 1971: 105f; DNWSI 155f [byrh]), cf. OfA byrt, b(y)rtʾ ‘fortress’ (Muraoka and Porten 1998: 376); Syr. byrtʾ ‘palace, fortress’ (SL 143b); JBA byrtʾ1 ‘citadel’ (DJBA 206a–b); QA byrh ‘fortress’ (Beyer 1984: 532); JPA byrh ‘castle, fortress’ (DJPA 102a). The emphatic state of the Aramaic word might have been modeled directly on the Akkadian form implying the feminine ending -t. Less probable seems the alternative advanced by Lemaire and Lozachmeur (1987: 161f.) that the Akkadian word was derived from North West Semitic. On the other hand, the NB/LB plural birānātu stems from Aramaic, witness the presence of the -ān suffix in the plural form (AHw 129). 10) GṢʾ SG.EMPH ‘mortar’ (Beyer 1998: 172: gaṣṣā ‘Gips’) ← Akk. gaṣṣu ‘gypsum, whitewash’ (from OB on; CAD G 54f, AHw 282f III; Zimmern 1917: 60; DNWSI 231[gṣ]), cf. Syr. gṣʾ ‘chalk, lime’ (SL 254b); JBA gṣʾ ‘lime’ (DJBA 297b); Aramaic → CA ǧiṣṣ ‘mortar’ (Fraenkel 1886: 10). Both AHw and CAD consider this lexeme as a loanword in Aramaic and from there into Arabic, whereas Kaufman does not accept this as an Akkadian loanword.51 11) HYKLʾ SG.EMPH ‘temple’ (Beyer 1998: 174: hēḵlā ‘Tempel’) ← Sum. via Akk. ekallu ‘palace’ (from OAkk. on; CAD E 52ff, AHw 191ff.; Zimmern 1917: 8; Kaufman 1974: 27; DNWSI 278 [hykl], cf. OfA hyklʾ ‘palace’ (Aḥiqar 9, 17, 23, 44); JudA hykl, hkl ‘temple’ (DJA 43f); Palm. hyklʾ ‘temple’ (PAT 359b); Syr. hyklʾ 1. ‘palace’, 2. ‘temple, holy place, shrine in a temple, nave, church’, 3. ‘name of a star’ (SL 340f); JBA hyklʾ ‘the Temple, temple; homestead’ (DJBA 377b); Md. hyklʾ 1. ‘hall, temple’, 2. ‘palace, homestead’, 3. ‘structure, stature’ (MD 143b); Aramaic of Targum Onqelos hyklʾ ‘temple’ (Cook 2008: 68); JudA hyklʾ ‘palace, temple’ (Beyer 1984: 562; Beyer 1994: 338); JPA hyklʾ ‘temple’ (DJPA 163a); Aramaic → CA haykal ‘temple’ (Fraenkel 1886: 274). There are many doubts about the nature of this loanword. Apparently we are faced with a noun borrowed very early into North West Semitic (as evidenced

50  For the Eastern character of bābā and its attestation in Manichaean Syriac, see Contini 1995: 78f., 89f., n. 134, quoting Lidzbarski 1918–1920: XIII and 1927: 916f., who recognized in a Middle Persian poem from Turfan the pun “Babel—door (of truth)” which presupposes the contrast between bʾbyl and bʾbʾ dkwšṭ. 51  Kaizer 2006: 145, nn. 5 and 11 for an exhaustive list of the interpretations proposed.

Notes on foreign words in Hatran Aramaic

143

by the presence of Ugaritic hkl). The preservation of h in all attested Aramaic forms shows the duration of the loan as such and that it was not re-borrowed. 12) ZPʾ SG.EMPH ‘slander’ hapax legomenon H74,7 (Beyer 1998: 175: zēpeh ‘Verleumdung’) ← Akk. ziʾpu, zīpu ‘mold’ (OB, SB, LB; CAD Z 86f; AHw 1529f. III; Zimmern 1917: 27; Kaufman 1974: 113; DNWSI 338f), cf. OfA zptʾ (Aḥiqar 130, 131); Syr. zʾpʾ ‘lie, falsehood, fraud, deception; dross, slag’ (SL 361a–b); JBA zypʾ2 ‘falsity’ (DJBA 410b); Md. zypʾ ‘falsity, falseness, deceit, fraud’ (MD 167a); SA zyp ‘deceit, lie’ (DSA 229a–b); Aramaic of Targum Onqelos zypʾ ‘mold’ (Cook 2008: 74); Aramaic → CA zayf ‘fake money’ (Fraenkel 1886: 195f). For the semantic extension from the concrete sense ‘mold’ to the metaphorical one ‘falsehood’, see Kaufman (1974: 113 n. 410). 13) ḤṢNʾ PL.EMPH ‘axes’ hapax legomenon H281,7 (Beyer 1998: 176: ḥaṣṣīnē ‘Beile’) ← Akk. ḫaṣṣinnu ‘axe’ (from OAkk. on; CAD Ḫ 133f, AHw 332a–b; Zimmern 1917: 12; Kaufman 1974: 54; DNWSI 400 [ḥṣn2]), cf. Syr. hṣynʾ ‘pickaxe, axe’ (SL 483b); JBA hṣynʾ ‘ax’ (DJBA 479a–b); Targumic Aramaic hṣynʾ; Aramaic → CA ḫṣyn ‘axe’ (Fraenkel 1886: 86f). According to Kaufman “there is no reason to suppose that this old culture word necessarily entered Aramaic from Akkadian”; so Degen 1978: 68 and Kaizer 2006: 142 n. 7. For the opposite view see now Cohen et al. 1970–: 910a. 14) *MḤZH SG.CON + 3.M.SG suffix ‘city’ hapax legomenon H1039,7 (Beyer 1998: 178: māḥōzeh ‘(seine) Stadt’) ← Akk. māḫāzu ‘town, sanctuary, harbor’ (from OAkk. on; possibly via Canaanite maḫōz [Beyer 1994: 372]; CAD M/1 85ff, AHw 582a–b; Zimmern 1917: 9; Kaufman 1974: 68; Marrassini 1971: 79f; DNWSI 611 [mḥz1]), cf. OfA *mḥwzʾ hypothetical original of MP MḤWZʾ ‘region’ (Gignoux 1972: 57a); Nab. mḥwzʾ (= mḥwz ʿgltyn) ‘harbor’ (Yardeni 2000: 155b); Palm. mḥwz ‘city’ (PAT 379b); Syr. mḥwzʾ ‘walled city’ (SL 737a); JBA mḥwzʾ ‘walled city, town’ (DJBA 654b); Md. mʾhwzʾ ‘town, small walled city or village’ (MD 240a); JudA mḥwz ‘port’ (presumably derived from a lost OfA *mḥwzʾ forerunner of mḥwz) (DJA 62a); Aramaic of Targum Onqelos mḥwz ‘harbor’ (Cook 2008: 149); JPA mḥwz ‘harbor; district’ (DJPA 299b).52 Contrary to the hypothesis of an Akkadian loanword in Aramaic, see Amadasi-Guzzo (1982) who derives all West Semitic forms from an hypothetical root *mḫd meaning ‘harbor’, different from Akkadian māḫāzu ‘place from which one can take’ (←√*ʾḫḏ). As a third possibility, Beyer (1994: 372) suggests a Canaanite intermediation maḫōz ‘harbor’ between the Hebrew and Aramaic forms and Akkadian māḫāzu ‘Markt – und Kultstadt’. 52  The original vowel quality is usually preserved in Aramaic loanwords from Akkadian, except for Akkadian ā > ō in this and few other words. The ō could also be due to Canaanite influence.

144

Contini and Pagano

15) MNDLTʾ SG.EMPH ‘relief-portrait’ hapax legomenon H1019,1 (Beyer 1998: 179: mandaltā ‘Personenrelief’) ← Akk. manzaltu ‘stand for an object or image’ (from OAkk. on; < mazzaz/štu, manzaz / štu: -št- > -lt-; CAD M/1 228ff. A; AHw 638a–b; Zimmern 1917: 62; Kaufman 1974: 69f; DNWSI 609 [mzl]), cf. Syr. mwzlʾ ‘astronomical sphere’ (SL 723a); JBA mzlʾ ‘zodiacal station, planet; fortune, guardian angel’ (DJBA 653f); Md mʾnzʾlʾ ‘astronomical and astrological term opposed to zaqfa and ziqpa I (q. v.), setting, constellation; star of destiny, stellar influence, horoscope’ (MD 248a–b); JPA mzl ‘constellation; lot, luck’ (DJPA 298b). Two different semantic shifts are presupposed by Beyer’s conjectural translation and suggested Akkadian derivation: 1. from a metaphorical to a concrete sense, 2. a semantic extension by metonymic change from the stand to the object which it supports. Beyond the semantic difficulty, however, there is the phonological correspondence of Aramaic d to Akkadian z, which would be more consistent with a cognate relationship than with a loan. 16) MTʾ SG.EMPH ‘country’ (Beyer 1998: 180: māṯeh; Beyer 2013: 59: māṯā ‘Land’) ← Akk. mātu ‘country’ (from OAkk. on; CAD M/1 414ff, AHw 633f. I; Zimmern 1917: 9; Kaufman 1974: 71; DNWSI 706f [mt1]), cf. OAram. mt (Tell Fekheriye; ed. Abou-Assaf et al. 1982: 75); OfA mtʾ ‘land’ (Muraoka and Porten 1998: 377); Syr. mtʾ ‘region, home-land’ (SL 858a); JBA mtʾ, mʾtʾ ‘town, place; built up area; first element in GN’s’ (DJBA 718a–b); Md. mtʾ ‘territory, (small) town, village’ (MD 256b); QA and JudA mt ‘region, district’ (Beyer 1984: 632; Beyer 1994: 378). The distribution and semantic development of this lexeme clearly indicate that we are in the presence of a loan. It is attested earlier in the Old Aramaic of Tell Fekheriye, later on in Official Aramaic which preserves the original meaning of ‘country’; subsequently this word is found only in Eastern Aramaic, but with a semantic narrowing: in HA, Syriac, Jewish Babylonian Aramaic, and Mandaic, it denotes a region or town. 17) PḤRʾ SG.EMPH ‘meeting hall’ hapax legomenon H283,1 (Beyer 1998: 182: poḥrā ‘Versammlungsraum’)53 ← Akk. puhru ‘assembly’ (from OA, OB on; CAD P 485ff, AHw 876f; Zimmern 1917: 46; Kaufman 1974: 83; DNWSI 907f [pḥr1]), cf. Syr. pwhrʾ ‘banquet; bands, crowds’ (SL 1161b); Md. pwhrʾ1 ‘feast, banquet’ (MD 367b). While this noun is found also in Ugaritic (DULAT 669–670 [pḫr I ← Akk.]), the cognate verb paḫāru is known only in Akkadian. 18) PRYKʾ SG.EMPH ‘incense-altar’ (Beyer 1998: 182: prekkā ‘Räucheraltar’) ← Sum. via Akk. parakkum ‘dais, pedestal, socle’ (from OB, MA on; CAD P 145ff A; AHw 827f; Zimmern 1917: 68; Kaufman 1974: 80 and 150; DNWSI 938 53  For HA pḥrʾ ‘meeting hall’ rather than ‘banquet’, see also Contini 2012: 334 n. 9.

Notes on foreign words in Hatran Aramaic

145

[prk2]), cf. Syr. prkʾ ‘altar’ (SL 1241a); Md. prykyʾ ‘pagan shrines and their spirits’ (MD 379a–b). In Aramaic this noun is epigraphically attested only in HA; the word survives in Syriac and Mandaic with semantically comparable meanings. 19) PTRʾ SG.EMPH ‘altar-table’ (Beyer 1998: 183: pāṯūrā ‘Tisch’) ← Sum. via Akk. paššūrum ‘offering table’ (from OAkk. on; CAD P 259ff 2; AHw 845f; Zimmern 1917: 33; Kaufman 1974: 81f; DNWSI 953f [ptr1]), cf. OfA ptwr ‘table’ (Degen 1978: 31); Syr. ptwrʾ 1. ‘table’, 2. ‘course, dish, plate for host’, 3. ‘banquet’, 4. ‘contribution, gift’, 5. ‘moneychanger’s table’ (SL 1264b); JBA ptwrʾ ‘table; platter’ (DJBA 945b); Md. pʾtwrʾ ‘platter, dish; try-table (or try for ritual food)’ (MD 366a); QA ptwr ‘table’ (Beyer 1984: 673, Beyer 1994: 401); Aramaic of Targum Onqelos ptwrʾ ‘table’ (Cook 2008: 231); JPA ptwr1 ‘table’ (DJPA 454b); Aramaic → CA fāṯūr ‘table (mostly of marble)’ (Fraenkel 1886: 83). The correspondence of Aramaic {t} to Akkadian -šš- may be explained by the spirantization of post-vocalic /bgdkpt/ since the 5th.–4th. cent. BC, combined with the influence of a contiguous /r/ (Lipiński 1978: 207). 20) ŠDNʾ SG.EMPH ‘possessed person’ hapax legomenon A33f,2 (Beyer 1998: 185: šēḏānā ‘Besessener’) ← Akk. šēdu (šēddu, šīdu) ‘a spirit or demon representing the individual’s vital force’ (from OB on; CAD Š/2 256ff A, AHw 1208a–b I; Zimmern 1917: 69; Kaufman 1974: 101; DNWSI 1111 [šdn]), cf. OfA *šdyʾ hypothetical original of MP ŠDYʾn, ŠYDʾn ‘demons’ (Gignoux 1972: 34f); Palm. šdyʾ, šedayyā ‘les démons’ (Cantineau 1935: 152; PAT 413a [‘Shadya (name of a deity); or, demon (?)’]); Syr. šʾdʾ ‘demon, one possessed by a demon’ (SL 1496a); JBA šydʾ ‘demon’ (DJBA 1132b); Md. šydʾ1 ‘demon’ (MD 460a); Aramaic of Targum Onqelos šd2 ‘demon’ (Cook 2008: 275); JPA (Galilaean) and JudA šyd ‘bad demon’ (Beyer 1984: 706; Beyer 1994: 420); JPA šd ‘demon’ (DJPA 538a). Akkadian loanword with Aramaic -ān suffix appended to the nominal Akkadian base šdy. Peter Jensen (Andrae – Jensen 1920: 19 n. 2) interprets šdn as a specific term for a religious role which he relates to the Arabic sâdin ‘overseer’, a suggestion followed by Milik (1972: 338). 4

Provisional Conclusions

Our sample of 32 foreign words—many of them hapax legomena—in HA includes 14 words with an Iranian, 11 with a Greek, 3 with a Latin, and 3 with an Arabian etymon, and 1 possible calque on Latin. The number of Greek loanwords (14, including 3 Latin) is vastly inferior to the 75 (including 19 Latin) counted by Brock (2005: 11) in Palmyrene: this is only natural, in consideration not only of the larger textual corpus

146

Contini and Pagano

available for Palmyrene, but above all of the longer and more pervasive presence of Hellenistic culture and Roman administration in the Palmyrene area compared with the Hatrene. Only for one Latin loanword, ʾs ← ās, assis, a direct loan might be considered, the word being hitherto unattested in this form in Greek: however, Latin loanwords are normally inducted by way of Greek in Nabataean and Palmyrene, almost always in Syriac.54 Most Greek loanwords in HA are also attested in either Palmyrene (possibly ʾqlwtʾ, ʾqltʾ ‘servant, follower’) or Syriac (ʾstr[y]n / ʾstrʾ; dryknʾ; ksydʾ; mklʾ; prsdʾ), or both (ʾwṭrnʾ; ʾsṭwʾ; gns; dnrʾ; smyʾ); only 2 Greek words in HA have at present no Palmyrene or Syriac counterpart: plṭyʾ ‘palace personnel’; prpyṭʾ ‘wanderers / peripatetic philosophers’. A few remarks shall follow concerning the manner in which Greek words have been taken over into HA.55 a) Orthography: as in Palmyrene and Syriac, there is a regular equivalence for Gr. consonants, normally maintaining a distinction between k = χ, q = κ, ṭ = τ, t = θ: the only exceptions seem to be ʾqltʾ and ʾstrʾ, which show /t/ in correspondence of τ.56 Again as in Palmyrene and Syriac, Greek initial consonant clusters beginning with a sibilant are provided with a prothetic ʾ-, e.g., ʾstr[y]n or ʾstrʾ ← Gr. στατήρ. b) Treatment of Greek case-endings: /-os/ is only represented once by -s (gns ← Gr. γένος), more frequently by the Aramaic SG.EMPH ending -ʾ (e.g., ʾwṭrnʾ ← Lat. veteranus via Gr. οὐετeρανός; ʾql(w)tʾ prob. ← Gr. ακόλουθος), a clear token of morphological adaptation.57 Integrated Greek loanwords also show the adjunction of an Aramaic either ABS or EMPH PL ending, e.g., ’str[y]n (si vera lectio) ← Gr. στατήρ; dnrʾ ← Lat. denarius via Gr. δηνάριον. c) Pronominal suffixes: the adjunction of an Aramaic possessive pronominal suffix to a Greek word, considered by Brock a sure sign of its full integration in Aramaic, is so far unattested in the HA corpus. d) Derivational suffixes: in contrast with Palmyrene and Syriac, but similarly to Nabataean, the Aramaic abstract suffix /-ūṯ/ is never attached to words of Greek origin in HA.58 On the other hand, a derivational suffix /-ān/, followed 54  Brock 2005: 23. Butts (forthcoming) also reaches the same conclusion. 55  For the sake of comparison with the Palmyrene and Syriac evidence, we follow here as far as possible the frame of Brock 2005’s presentation. 56  See the discussion of an exception to the correspondence τ = /ṭ/ in Palmyrene in Gzella 2007: 721. 57  Compare the assessment in Brock 2005: 24 on the increasing morphological adaptation of Greek loanwords into Palmyrene and Syriac evident in the sequence of reflexes of Gr. /-os/: 1) -0, 2) -ws, 3) -s, 4) -sʾ, 5) -ʾ. 58  Compare Brock 2005: 25 for Palmyrene and Syriac.

Notes on foreign words in Hatran Aramaic

147

by the EMPH PL ending -ʾ, added to the stem of Gr. δαρεικός, δαρικός (in its turn borrowed from Old Persian) may be traceable in dryknʾ ‘darics’. One substantivized adjective shows the affixation of the Aramaic nisba suffix -āy to a Greek loanword: plṭyʾ ‘palace personnel’ ← παλάτιον, palatium. e) Evidence of borrowings taken over from Greek forms different from the nominative case, e.g. ksydʾ ← ξοίς / -ίδος, the presence of -d- implying derivation from either the genitive or the accusative; the same phenomenon may be posited (with the loss of -t/ṭ- before -d-) for prsdʾ ← Gr. προστάς / -άδος.59 Latin loanwords are generally imported into HA—as into Palmyrene and Old Syriac—through Greek, with the only possible exception of ʾs ←ās, assis, so far unattested in Greek except as borrowed through the derived form assarius → ἀσσάριον, borrowed in its turn into several Aramaic dialects. A semantic shift of HA klylʾ ‘crown’ → ‘coronation tax’, calqued on Latin (aurum) coronarium (as attested in Jewish Babylonian Aramaic), may be reconstructed at the origin of the substantivated nisba-adjective klylyʾ ‘tax collector’. Words with a recognizable Iranian etymon in HA have been traced by Iranian philologists to different phases and dialects: though most of them were probably borrowed directly from a Middle Iranian dialect, either Middle Persian or Parthian, a minority of words of Iranian origin seem to be the continuators of Old Iranian (Old Persian and less frequently the variety labeled as “Median” by Walther Hinz) borrowings into Official Aramaic:60 gzbrʾ, possibly znpṭʾ, *ṭwhm, and krnytʾ seem to belong to this class. For lack of specific competence, we shall leave to Iranian philologists the individuation of the punctual dialectal affiliation of each Iranian loanword in HA, as well as the discussion of the phonological and morphological features of their adaptation to Aramaic, which are pertinent for this classification. We may simply remark that the index of high morphological integration recognized by Brock in the affixation of Aramaic pronominal suffixes to foreign loanwords, while not documented for Greek or Latin borrowings, is however attested for words of Iranian origin in HA, all of them bearing a possessive 3.M.SG suffix, e.g., ʾšpzknh ← Early MP *aspinjakān ‘host, inn-keeper’; ṭwhmh ← OP *tauma- (= tauhmā) ‘race, stock, family’; tgh ← MIr. *tāγ ‘crown’. Only three Arabian loanwords,61 which can with some confidence be assigned to Old Arabic, are so far documented in HA: ʾhl ← OAr. *ʾahl ‘clan, tribe’; 59  Compare Brock 2005: 25 for Palmyrene and Syriac. 60  For this category in Syriac, see Ciancaglini 2008: 25–28. 61  A fourth item would be ʾpkl, which we chose to treat here among the words of Akkadian origin, despite its very probable importation into HA through an Arabian intermediate ʾfkl, witness its specific cultic meaning.

148

Contini and Pagano

ḥṭmʾ prob. ← OAr. *ḫiṭām ‘muzzle’ (the modern terminology for Bedouin material culture confirms this meaning: in Sbaʿa Arabic ḫṭāme denotes the camel’s muzzle, while the halter attached to it is named ǧirīr and the complex formed by muzzle + halter is resan; this makes evident the semantic link of the OAr. and HA word with the Semitic √ḫṭm meaning ‘nose, snout’);62 ḥynʾ ← OAr. *ḥīn ‘time, moment’. None of these words is actually attested in the very meagre corpus of Old Arabic inscriptions, either in Ancient South Arabian, Ancient North Arabian, or Nabataean heterography63 or in the Old Arabic script,64 but several dozens of words are today known in the Nebenüberlieferung of the language in Nabataean, Judean Aramaic, Palmyrene, and Old Syriac,65 so that the addition of HA to this list of witnesses should cause no surprise. The words in which an Akkadian lexical legacy can be recognized are twenty, more than half of them hapax legomena. The Akkadian lexical entries making their appearance in the HA vocabulary ought to be considered inherited from the Official Aramaic lexicon. Except for the professional cultic noun ʾpkl, attested in Palmyrene, Nabatean, and various Ancient North Arabian dialects and Sabaic, all the other lexical items are also attested in Late Eastern Aramaic dialects (Syriac, Jewish Babylonian Aramaic, Mandaic). Five of these words are ultimately of Sumerian origin (ʾpklʾ, ʾškpʾ, hyklʾ, prykʾ, ptwrʾ), but they of course entered Aramaic through Akkadian. Sometimes words thought to be of Akkadian origin were part of the West Semitic lexicon for a long time, so the validity of the loan should be discussed (e.g., hēklā < ekallu). In his monograph on HA inscriptions Beyer explicitly posits an Akkadian origin only for two items: zēpeh and mandaltā: while the first is a clear and widely recognized instance of the Akkadian lexical legacy, the second displays several semantic and phonological difficulties that incline us to refute the proposed Akkadian derivation. An assessment of the HA vocabulary, greatly increased in recent decades, has allowed us to enlarge the data offered by Kaufman (1974: 161) about the Akkadian loanwords in Aramaic “Monumental Dialects.”66 62  See Cohen et al. 1970–: 970f., with copious references. 63  Hopkins 2008 has recently made a case for the possibility of reading at least part of the pre-Islamic Jewish inscriptions in the square Aramaic script from Ḥijāz as heterographic Judaeo-Arabic texts. 64  See Mascitelli 2006 and Macdonald 2008, both with abundant bibliography. 65  See n. 4 above for the essential literature. One rather misses a reference to this material, as well as a lexical index, in Mascitelli’s very useful monograph on Old Arabic (2006). 66  Only three words were recognized as influenced by Akkadian: parakku, aškapu, and the unusual word ʾpkl (the first two attested only in HA).

Notes on foreign words in Hatran Aramaic

149

As can be seen from our list of proposed words of Akkadian origin in HA, it is not difficult to identify a link between these entries and specific spheres of Hatrene life which seem to show a continuity with the Mesopotamian legacy and which are generally more conservative: 1. building: ʾabwlʾ, ʾšlʾ, ʾštʾ, byrtʾ, gṣʾ, hyklʾ, hṣynʾ, lbynʾ, mndltʾ, prykʾ, ptwrʾ 2. topographical / administrative features: mḥzh, mtʾ, phrʾ 3. professions / crafts: ʾpklʾ, ʾrdklʾ, ʾškpʾ, ʾšprʾ, possibly bbyʾ. To fully assess the weight of the Akkadian component in the HA lexis, we must consider the location of Hatra and the surrounding region in a formerly Assyrian area, whose cultural legacy it still shows. Hatrene culture provides an exceptional testimony of the entire range of the Assyro-Babylonian civilization, progressively subjected to Hellenistic, Parthian, Roman, and Arab influence. Hence many cultural elements in Hatra bear Mesopotamian imprints and are still denoted by words of Akkadian origin. The Akkadian cultural influence on Hatra is particularly noticeable in the religious sphere. Architectonically, the persistence of Mesopotamian models is evident in the organization of the spaces shaping the temple complex dedicated to Šamaš (Temple H), which echoes the genuine Mesopotamian type of “temple in antis.”67 “The Hatra of the Parthians—the last refuge of the Assyrian gods—,” according to Leo Oppenheim’s famous definition,68 is still reflected in the divine names and in the theophoric elements of HA personal names. The evidence for the worship of gods with Akkadian names in Hatra shows a clear continuity with the pre-classical Mesopotamian tradition: however, the general religious physiognomy of Hatrene religion is markedly syncretistic, so that one should not overstate the actual survival of ancient Mesopotamian deities and theologoumena.69 Both extensive archaeological remains and the legends of Hatrene coins witness to the presence in Hatra of an important solar cult. The sumptuous temple complex,70 built in honor of the Mesopotamian sun god Šamaš is sometimes called in HA Saggīl hēklā rabbā,71 echoing the name of Babylon’s famous Esagila, the dwelling-place of the city-god Marduk.

67  Sommer 2003: 63–70. 68  Oppenheim 1964: 134. 69  For the adaptation of Babylonian and Assyrian elements to this syncretistic culture, one thinks of the “model of an additive extension of an open system where choices are made on the principle of extending, rather than replacing, a cultural system” (Bendlin 1997: 53). 70  Ibrahim 1986: 123–127 and pl. 59. 71  H 107,6; 191,2; 202,q; 225,1; 240,2; 244,2; 245,2; 246,3.

150

Contini and Pagano

Other deities of Mesopotamian origin are Nabū, known in the inscriptions with the name of sāprā dmāran ‘Nabū the scribe of Māran’ (H389); Nergal, a chthonic Babylonian deity connected with forest fire, plague, and other disasters, identified at Hatra with the Greco-Roman Heracles.72 Linked to the cult of Šamaš is the goddess Nanaya—Ištar, attested however only in the Hatrene onomasticon,73 which shows a high proportion of theophoric personal names including Mesopotamian theophoric components: Šamaš-names are the most popular,74 but other Mesopotamian deities such as Assur, Nabū, and Nergal are both named in the inscriptions and represented in HA theophoric names.75 All in all, this rather cursory and provisional76 survey of words of foreign origin in HA, be they borrowed from languages in contact with it or ascribable to the Akkadian lexical legacy, rather confirms what could a priori be expected on the strength of Hatra’s known political and cultural history. While HA vocabulary reflects Arsacid northern Mesopotamia’s debt to cuneiform culture and the importance of the cult of the solar god Šamaš / Šmeš at Hatra, the origin of the great majority of foreign loanwords in HA is consistent with the Hellenistic and Parthian influences which are quite visible in Hatra’s art and architecture.77 Iranian accounts for fourteen out of our selection of thirty-one foreign loanwords (+ 1 calque) in HA, most of them being represented by Parthian administrative, civil, or military titles (ʾsppṭʾ, *ʾšpzkn, bṭḥšʾ / pdḥšʾ, gzbrʾ, dḥšʾ, dḥšpṭʾ, hdrpṭʾ, znpṭʾ, krnytʾ, nḥšrpṭ[ʾ], pzgrybʾ / pšgrbʾ), with the addition of a term denoting an object in some way connected to the administrative sphere (ʾmnykʾ). Greek and Latin loanwords cover a slightly broader semantic scope: the economic and fiscal (ʾs, ʾstr[y]n / ʾstrʾ, dnrʾ, dryknʾ, klylyʾ) or military sphere (ʾwṭrnʾ), architecture (ʾsṭwʾ, prsdʾ), building implements (ksydʾ, mklʾ), cultic objects (smyʾ), palace or temple personnel (ʾqlwtʾ, plṭyʾ), and terms difficult to classify because they are hapax legomena whose very fragmentary context does not afford a precise understanding of their semantic range (e.g., gns, apparently meaning ‘clan’, and prpyṭʾ, conjecturally interpreted by Beyer as ‘the wanderers’). Arabian influence in Hatra is mainly visible in its 72  For the syncretistic phenomena of Nergal’s cult at Hatra, see Kaizer 2000, with full references. 73  See Beyer 1998: 152. 74  Abbadi 1983: xxii–xxiii. 75  For a comprehensive discussion of the Mesopotamian profile of the HA onomasticon in the Assur materials, see Livingstone 2009. 76  Admittedly, our presentation is occasionally not innocent of some almost inevitable concessions to speculative reasoning. 77  See Tubach 1986: 213–335, Sommer 2003; 2005: 355–380; Healey 2009: 16ff., all with abundant references to earlier literature.

Notes on foreign words in Hatran Aramaic

151

onomastics (beside several other ethnic and linguistic components),78 but the three Old Arabic loanwords we have discussed above show at least some degree of exposure of HA to a lexis consistent with Arab tribal organization or Bedouin material culture. All of the foreign loanwords in HA are nouns, in conformity with the general cross-linguistic prominence of this category among borrowed lexical classes.79 As is already the case for the Nabataean, Palmyrene, and Old Syriac corpora, the foreign connections of the lexical evidence from Hatra should therefore also be considered in their own right in any account of its historical and cultural foreign connections, besides being integrated into the history of languages in contact with Aramaic through the ages.80 Bibliography Abbadi, S. 1983. Die Personennamen der Inschriften aus Hatra (Texte und Studien zur Orientalistik 1). Hildesheim. Abou-Assaf, A., P. Bordreuil, and A. Millard. 1982. La statue de Tell Fekherye et son inscription bilingue assyro-araméenne (EA 7). Paris. Aggoula, B. 1975. “Remarques sur les inscriptions hatréennes III,” Syria 52: 181–195. ———. 1983. “Remarques sur les inscriptions hatréennes (VII),” AuOr 1: 31–38. ———. 1985. Inscriptions et graffites araméens d’Assour (AION Suppl. 45). Naples. ———. 1986. “Remarques sur les inscriptions hatréennes (XII),” Syria 63: 353–374. ———. 1990. “Remarques sur les inscriptions hatréennes (XVI),” Syria 67: 397–421. ———. 1991. Inventaire des inscriptions hatréennes (Bibliothèque archéologique et historique 139). Paris. Alpass, P. 2013. The religious life of Nabataea. Leiden. Amadasi Guzzo, M. G. 1982. “Il vocabolo mʾḫd / mḥz in ugaritico e fenicio,” in S. Ribichini et al. (eds.), Materiali lessicali ed epigrafici, vol. 1. Rome. 31–36. Andrae, W. and P. Jensen. 1920. “Aramaïsche Inschriften aus Assur und Hatra aus der Partherzeit,” MDOG 60: 1–54. Asbaghi, A. 1988. Persische Lehnwörter. Wiesbaden.

78  Abbadi 1983 rather urgently needs to be updated, but for the time being Beyer’s onomastic indexes (1998: 144–168; 2013: 50–55) offer at least a tentative compendium and interpretation of the evidence. 79  Matras 2009: 167ff. 80  As canvassed selectively, e.g., by Khan 2007.

152

Contini and Pagano

Beeston, A. F. L. 1994. “Foreign loanwords in Sabaic,” in N. Nebes et al. (eds.), Arabia Felix. Beiträge zur Sprache und Kultur des vorislamischen Arabien. Festschrift Walter W. Müller zum 60. Geburtstag. Wiesbaden. 39–45. Bendlin, A. 1997. “Peripheral centres—central peripheries: religious communication in the Roman Empire,” in H. Cancik – J. Rüpke (eds.), Römische Reichsreligion und Provinzialreligion. Tübingen. 35–68. Benveniste, E. 1966. Titres et noms propres en iranien ancien. Paris. Bertolino, R. 2007. Corpus des inscriptions sémitiques de Doura-Europos (AION Suppl. 94). Naples. ———. 2008. Manuel d’épigraphie hatréenne. Paris. Beyer, K. 1984. Die aramäischen Texte vom Toten Meer. Göttingen. ———. 1986. The Aramaic language. Its distribution and subdivisions. Göttingen. ———. 1994. Die aramäischen Texte vom Toten Meer, Ergänzungsband. Göttingen. ———. 1998. Aramäische Inschriften aus Assur, Hatra und dem übrigen Ostmesopotamien. Göttingen. ———. 2004. Die aramäischen Texte vom Toten Meer, Vol. 2. Göttingen. ———. 2013. “Die aramäischen Inschriften aus Assur, Hatra und dem übrigen Ostmesopotamien,” WO 43: 25–62. Biberstein Kazimirski, A. De. 1860. Dictionnaire arabe-français, Vol. I–II. Paris. Boucheman, A. de 1934. Matériel de la vie bédouine recueilli dans le désert de Syrie (tribu des Arabes Sbaca) (Documents d’études orientales III). Damascus. Brock, S. P. 2005. “Greek and Latin words in Palmyrene inscriptions: a comparison with Syriac,” in Cussini 2005: 11–25. Brockelmann, C. 1928. Lexicon Syriacum (2nd ed.). Halis Saxonum. Butts, A. M. Forthcoming. “Latin Words in Classical Syriac,” Hugoye: Journal of Syriac Studies. Cantineau, J. 1930–32. Le nabatéeen, I–II. Paris. ———. 1935. Grammaire du palmyrénien épigraphique. Cairo. Ciancaglini, C. A. 2008. Iranian loanwords in Syriac (Beiträge zur Iranistik 28). Wiesbaden. Cohen, D. et al. 1970–. Dictionnaire des racines sémitiques. Paris – Louvain. Contini, R. 1995. “Hypothèses sur l’araméen manichéen,” Annali di Ca’ Foscari 34/3: 65–107. ———. 2012. “Banchetti e pasti comunitari nell’Arabia antica: appunti sull’evidenza epigrafica e letteraria,” in L. Milano (ed.), Mangiare divinamente. Pratiche e simbologie alimentari nell’antico Oriente (Eothen 20). Firenze. 331–353. Cook, E. M. 2008. A glossary of Targum Onkelos. Leiden. Cussini, E. (ed.). 2005. A journey to Palmyra. Collected essays to remember Delbert R. Hillers (CHANE 22). Leiden.

Notes on foreign words in Hatran Aramaic

153

Dalman, G. 1897–1901. Aramäisch-neuhebräisches Wörterbuch zu Targum, Talmud und Midrasch, Vol. I–II. Frankfurt. Degen, R. 1978. “Weitere Inschriften aus Hatra (Nr. 281–335),” NESE 3: 67–111. De Romanis, F. 2004. “Sul prestigio della moneta aurea romano-bizantina in Oriente (secc. I–VI d.C.),” in La Persia e Bisanzio. Atti del Convegno Internazionale dell’Accademia Nazionale dei Lincei (Roma, 14–18 ottobre 2002). Rome. 291–308. Dirven, L. 2005. “Σημειον, SMYʾ, signum. A note on the Romanization of the Semitic cultic standard,” Parthica 7: 119–136. ———. 2006–2007. “Hatra: a ‘Pre-Islamic Mecca’ in the Eastern Jazirah,” Aram 18–19: 363–380. Drijvers, H. J. W. and Healey, J. F. 1999. The Old Syriac inscriptions of Edessa and Osrhoene. Leiden. Drower, E. S. and R. Macuch. 1963. A Mandaic dictionary. Oxford. Ellenbogen, M. 1962. Foreign words in the Old Testament. Their origin and etymology. London. Fraenkel, S. 1886. Die aramäischen Fremdwörter im Arabischen. Leiden. Gelb, I. J. et al. 1956–2010. The Assyrian Dictionary of the Oriental Institute of Chicago. Chicago. Gignoux, Ph. 1972. Glossaire des inscriptions pehlevies et parthes (Corpus Inscriptionum Iranicarum, Suppl. Series 3). London. Glare, P. G. W. and Thompson, A. A. 1996. Greek-English Lexicon Revised Supplement. Oxford. Gnoli, T. 2002. “Paṣgribā at Hatra and Edessa,” in A. Panaino and G. Pettinato (eds.), Ideologies as intercultural phenomena. Proceedings of the Third Annual Symposium of the Assyrian and Babylonian Intellectual Heritage Project Held in Chicago, Usa, October 27–31, 2000 (MELAMMU Symposia III). Milan. 79–87. Greenfield, J. C. 1987. “Aramaic II: Iranian loanwords in Early Aramaic,” in E. Yarshater (ed.), Encyclopaedia Iranica 2.256–261. ———. 1989. “Nergol DḤŠPṬ’,” in W. Sundermann et al. (eds.), A green leaf: Papers in honour of Professor Jes P. Asmussen. Leiden. 135–143. (= Paul et al. 2001: 418–426) ———. 1992. “Some Arabic loanwords in the Aramaic and Nabatean texts from Naḥal Ḥever,” JSAI 15: 10–21. (= Paul et al. 2001: 497–508) Gzella, H. 2006. “Das Aramäische in dem römischen Ostprovinzen: Sprachsituationen in Arabien, Syrien und Mesopotamien zur Kaiserzeit,” BiOr 63: 15–39. ———. 2007. Rev. of Cussini 2005. BiOr 64: 720–725. ———. 2008. “Aramaic in the Parthian period: The Arsacid inscriptions,” in H. Gzella and M. Folmer (eds.), Aramaic in its historical and linguistic setting (Akademie der Wissenschaften und der Literaturen · Mainz / Veröffentlichungen d. Orientalischen Kommission 50). Wiesbaden. 107–130.

154

Contini and Pagano

———. 2011. “Late Imperial Aramaic,” in S. Weninger et al. (eds.), Semitic languages: An international handbook (HSK 36). Berlin. 598–609. Harding G. L. 1971. An index and concordance of pre-Islamic Arabian names and inscriptions (Near and Middle East Series 8). Toronto. Harnack, D. 1970. “Parthische Titel, vornehmlich in den Inschriften von Hatra,” in F. Altheim and R. Stiehl (eds.), Geschichte Mittelasiens im Altertum. Berlin. 492–549. Healey, J. F. 1986. “The Nabataeans and Madāʾin Ṣāliḥ,” Atlal 10: 108–116. ———. 1993. The Nabataean tomb inscriptions of Mada’in Salih. Oxford. ———. 1995. “Lexical loans in Early Syriac: A comparison with Nabataean Aramaic,” SEL 12: 75–84. (= Healey 2011: no. XIV + pp. 8–9) ———. 2009. Textbook of Syrian Semitic inscriptions, Vol. 4. Aramaic inscriptions and documents of the Roman period. Oxford. ———. 2011. Law and religion between Petra and Edessa. Studies in Aramaic epigraphy on the Roman frontier (Variorum Collected Studies Series). Farnham – Burlington. Hillers, D. R. and Cussini, E. 1996. Palmyrene Aramaic texts. Baltimore. Hinz, W. 1975. Altiranisches Sprachgut der Nebenüberlieferungen. Wiesbaden. Hoftijzer, J. and Jongeling, K. 1995. Dictionary of North-West Semitic inscriptions (HdO I/21), Vol. 1–2. Leiden. Hommel, F. 1893. Süd-arabische Chrestomathie: minäo-sabäische Grammatik, Bibliographie, minäische Inschriften nebst Glossar. Munich. Hopkins, S. 2008. “The earliest texts in Judaeo-Middle Arabic,” in J. Lentin and J. Grand’Henry (eds.), Moyen arabe et variétés mixtes de l’arabe à travers l’histoire. Louvain-la-Neuve. 231–249. Ibrahim, J. Kh. 1986. Pre-Islamic settlement in Jazirah. Baghdad. Jastrow, M. 1903. A dictionary of the Targumim, the Talmud Babli and Yerushalmi, and the Midrashic literature, Vol. I–II. Philadelphia. Jobling, W. J. 1995. Nabataean-Aramaic: a provisional lexicon (Nablex). Kensington, MD. Kaizer, T. 2000. “The ‘Heracles Figure’ at Hatra and Palmyra: Problems of interpretation,” Iraq 62: 219–232. ———. 2002. The religious life of Palmyra. A study of the social patterns of worship in the Roman period (Oriens et Occidens 4). Stuttgart. ———. 2006. “Capital punishment at Hatra: Gods, magistrates and laws in the RomanParthian period,” Iraq 68: 139–153. ———. 2009. “Religion and language in Dura-Europos,” in H. Cotton et al. (eds.), From Hellenism to Islam. Cultural and linguistic change in the Roman Near East. Cambridge. 235–253. Kaufman, S. A. 1974. The Akkadian influences on Aramaic (AS 19). Chicago. Khan, G. 2004. “Aramaic and the impact of languages in contact with it through the ages,” in P. Bádenas de la Peña et al. (eds.), Lenguas en contacto: el testimonio escrito. Madrid. 87–108.

Notes on foreign words in Hatran Aramaic

155

Köhler, L., W. Baumgartner, J. J. Stamm, et al. (trans. M. E. J. Richardson). 2001. The Hebrew and Aramaic Lexicon of the Old Testament. Study Edition. Leiden. Krauss, S. 1898–99. Griechische und lateinische Lehnwörter im Talmud, Midrasch und Targum, Vol. I–II. Berlin. Kurpershoek, P. M. 2005. Oral poetry and narratives from Central Arabia, Vol. V. Voices from the desert: glossary, indices, & list of recordings (Studies in Arabic Literature 17/V). Leiden. Lemaire, A. and H. Lozachmeur. 1987. “Bīrāh/birtā en araméen,” Syria 64: 261–266. Lipiński, E. 1978. “La correspondance des sibilantes dans les textes araméens et les textes cunéiformes néo-assyriens,” in P. Fronzaroli (ed.), Atti del secondo congresso internazionale di linguistica camito-semitica Firenze, 16–19 aprile 1974 (QS 5). Firenze. 201–210. Livingstone, A. 2009. “Remembrance at Assur: The case of the dated Aramaic memorials,” in M. Luukko et al. (eds.), Of god(s), trees, kings, and scholars. Neo-Assyrian and related studies in honour of Simo Parpola (StOr 106). Helsinki. 151–157. Luukko, M. 2004. Grammatical variation in Neo-Assyrian (SAAS 16). Helsinki. Macdonald, M. C. A. 2008. “Old Arabic (epigraphic),” in EALL 3.464–477. MacKenzie, D. N. 1971. A concise Pahlavi dictionary. London. Maraqten, M. 1998. “The Arabic words in Palmyrene Inscriptions,” Aram 7: 89–108. ———. 2000. “Der Afkal/Apkallu im arabischen Bereich: eine epigraphische Untersuchung,” in J. Marzahn and H. Neumann (eds.), Assyriologica et Semitica— Festschrift J. Oelsner (AOAT 252). Münster. 263–283. Marrassini, P. 1971. Formazione del lessico dell’edilizia militare nel semitico di Siria (QS 1). Firenze. Mascitelli, D. 2006. L’arabo in epoca preislamica: formazione di una lingua (Arabia Antica 4). Rome. Matras, Y. 2009. Language contact (Cambridge Textbooks in Linguistics). Cambridge. Milik, J. T. 1972. Recherches d’épigraphie proche-orientale I. Dédicaces faites par des dieux (Palmyre, Hatra, Tyr) et des thiases sémitiques à l’époque romaine (Institut Français d’Archéologie de Beyrouth. Bibliothèque archéologique et historique 92). Paris. Millar, F. 1977. The Emperor in the Roman world (31 BC–AD 337). London. ———. 1998. “Caravan cities: the Roman Near East and long-distance trade by land,” in M. Austin et al. (eds.), Modus operandi: essays in honour of Geoffrey Rickman (BICS Supp. 71). London. 119–137. (= Millar 2006: 275–299) ——— (ed. H. M. Cotton and G. M. Rogers). 2006. Rome, the Greek world, and the East, Vol. 3. The Greek World, the Jews, and the East. Chapel Hill, NC – London. Miller, K. 1916. Itineraria romana: römische Reisewege an der Hand der Tabula Peutingeriana. Stuttgart. Monferrer Sala, J. P. 2013. “Greek administrative loanwords in Nabataean inscriptions,” MLR 20: 97–115.

156

Contini and Pagano

Moriggi, M. 2004. La lingua delle coppe magiche siriache (QS 21). Firenze. ———. 2012. “Middle Aramaic: outlines for a definition,” in L. Bettini and P. La Spisa (eds.), Au-delà de l’arabe standard. Moyen-arabe et arabe mixte dans les sources médiévales, modernes et contemporaines (QS 28). Firenze. 279–289. ———. 2014. A corpus of Syriac incantation bowls. Syriac magical texts from LateAntique Mesopotamia (MRLLA 3). Leiden. Muraoka, T. and Porten, B. 1998. A grammar of Egyptian Aramaic (HdO I /32). Leiden. Naveh, J. and Sh. Shaked. 1993. Magic spells and formulae. Aramaic incantations of Late Antiquity. Jerusalem. Nyberg, H. S. 1974. A manual of Pahlavi, Part 2. Glossary. Wiesbaden. O’Connor, M. 1986. “The Arabic loanwords in Nabatean Aramaic,” JNES 45: 213–229. Oppenheim, A. L. 1964. Ancient Mesopotamia. Londra. Paul, Sh. M., M. E. Stone, and A. Pinnick. 2001. ‘Al Kanfei Yonah: Collected studies of Jonas C. Greenfield on Semitic philology, Vol. 1–2. Leiden – Jerusalem. Payne Smith, R. 1879–1901. Thesaurus Syriacus, Vol. 1–2. Oxford. Pennacchietti, F. A. 1988. “Tre note di epigrafia hatrena,” Mesopotamia 23: 43–64 + 1 tav. ———. 1996. “Il medio aramaico orientale alla luce di testi epigrafici inediti di Hatra (Irak),” Incontri Linguistici 19: 59–68. Poirier, P.-H. 1981. L’Hymne de la Perle des Actes de Thomas. Introduction, texte— traduction, commentaire (Homo religiosus 8). Louvain-la-Neuve. Río Sanchez, Francisco del 2006. Textos epigráficos en arameo palmireno, hatreo y nabateo (Estudios de Filología Semítica 6). Barcelona. Rosenthal, F. 1936. Die Sprache der palmyrenischen Inschriften und ihre Stellung innerhalb des Aramäischen. Leipzig. Safar, F. 1971. “Kitābāt al-Ḥaḍr,” Sumer 27: 3–14. Schall, A. 1960. Studien über griechische Fremdwörter im Syrischen. Darmstadt. Schulthess, F. 1903. Lexicon syropalaestinum. Berlin. Schwyzer, E. 1931. “Awest. aspərənō und byzantin. ’άσπρον. Beiträge zur griechischorientalischen Münznamenforschung,” IF 49: 1–45. Segal, J. B. 1982. “Aramaic legal texts from Hatra,” JJS 33: 109–115. ———. 1986. “Arabs at Hatra and the vicinity: Marginalia on New Aramaic texts,” JSS 31: 57–80. Siddiqi, A. 1919. Studien über die Persischen Fremdwörter im klassischen Arabisch. Göttingen. Sima, A. 1997. Review of Aggoula 1991. AfO 42–43: 316–322. ———. 1999. Review of Healey 1993. AfO 44–45: 446–450. ———. 2000. Review of Beyer 1998. OLZ 95: 162–167. Soden, W. von 1965–81. Akkadisches Handwörterbuch, Vol. I–III. Wiesbaden.

Notes on foreign words in Hatran Aramaic

157

Sokoloff, M. 2002a. A dictionary of Jewish Palestinian Aramaic of the Byzantine period (2nd ed.). Ramat-Gan. ———. 2002b. A dictionary of Jewish Babylonian Aramaic of the Talmudic & Geonic periods. Ramat-Gan. ———. 2003. A dictionary of Judean Aramaic. Ramat-Gan. ———. 2009. A Syriac lexicon. A translation from the Latin, correction, expansion, and update of C. Brockelmann’s Lexicon Syriacum. Winona Lake – Piscataway. Sommer, M. 2003. Hatra, Geschichte und Kultur einer Karawanenstadt im römischparthischen Mesopotamien. Mainz. ———. 2005. Roms orientalische Steppengrenze. Palmyra – Edessa – Dura Europos – Hatra. Eine Kulturgeschichte von Pompeius bis Diokletian (Oriens et Occidens 9). Stuttgart. Steiner, R. C. 1989. “New light on the Biblical millo from Hatran inscriptions,” BASOR 276: 15–23. Tal, A. 2000. A dictionary of Samaritan Aramaic, Vol. I–II (HdO I/50). Leiden. Tavernier, J. 1997. Iranica in the Achaemenid period (ca. 550–330 BC): Lexicon of Old Iranian proper names and loanwords, attested in non-Iranian texts (OLA 158). Leuven. Taylor, D. G. 2002. “Bilingualism and diglossia in Syria and Mesopotamia,” in J. N. Adams et al. (eds.), Bilingualism in ancient society. Language contact and the written text. Oxford. 298–331. Teixidor, J. 1966. “Notes hatréennes,” Syria 43: 91–97. Tubach, J. 1986. Im Schatten des Sonnengottes. Der Sonnenkult in Edessa, Ḥarrān und Ḥaṭrā am Vorabend der christlichen Mission. Wiesbaden. Van Rompay, L. 2011. “Dura Europos,” in GEDSH, 132–133. Vattioni, F. 1981. Le iscrizioni di Hatra (AION Suppl. 41). Naples. ———. 1994. Hatra (AION Suppl. 81). Naples. Winnett, F. V. and W. L. Reed. 1970. Ancient records from North Arabia. Toronto. Yadin, Y., J. C. Greenfield, A. Yardeni, and B. A. Levine. 2002. The documents from the Bar Kokhbah period in the Cave of the Letters: Hebrew, Aramaic and Nabatean-Aramaic papyri. Jerusalem. Yardeni, A. 2000. Textbook of Aramaic, Hebrew and Nabataean documentary texts from the Judaean Desert and related material. Jerusalem. Zadok, R. 2010. Review of Ciancaglini 2008. OLZ 105: 754–761. Zimmern, H. 1917. Akkadische Fremdwörter als Beweis fur babylonischen Kultureinfluss (2nd ed.). Leipzig.

Language, Writing, and Ideologies in Contact: Sumerian and Akkadian in the Early Second Millennium BCE C. Jay Crisostomo 1 Introduction Sumerian and Akkadian textual bilingualism represents the earliest documented case of language contact. The present contribution examines variations in writing practices resulting from language ideologies and attitudes in the early second millennium BCE. In this study, I identify prevailing language ideologies associated with the use of Sumerian and Akkadian, the two most prominent languages in the Mesopotamian linguistic area. I argue that, despite the disappearance of Sumerian from the vernacular, metalinguistic ideologies valorizing Sumerian expressed primarily in literature correspond with language use in context. These ideologies are represented in writing variations used in various social contexts and textual genres including, as I show here, legal contracts. This brief contribution represents a preliminary investigation into the sociolinguistic parameters of Mesopotamian societal multilingualism reflected in writing practices. This initial introduction presents the methodologies informing the study. I first discuss the linguistic situation in Mesopotamia at the beginning of the second millennium BCE. I then present relevant native metalinguistic data. Next I compare the metalinguistic ideologies with language use “in action” by examining writing variations in the phrase mu—pad (Sum.) corresponding to nīš—itma (Akk.) ‘(s)he swore an oath’1 as used in legal real estate documents 1  Sumerian words are indicated in bold-faced font and glossed immediately afterwards (e.g., mu ‘name’); words and their morphophonemics are connected with hyphens (-), which separate the individual signs. Akkadian words are given in linguistically normalized form (as opposed to transliterated form wherein individual signs are represented) and indicated in bold-faced, italicized font and glossed immediately afterwards (e.g., šumu ‘name’). Em-dashes (—) separate parts of idiomatic phrases. Where relevant, I provide linguistic analysis of morphology. The following abbreviations are used: 3 = 3rd person; ABS = Absolutive; ACC = Accusative; Akk. = Akkadian; CONJ = Conjunction; DN = Divine Name; ERG = Ergative; IMPF = Imperfective; NTRL = neutral (focus); NOM = Nominative; PN = Personal Name; RN = Royal Name; SG = Singular; SN = Settlement Name; Sum. = Sumerian.

© koninklijke brill nv, leiden, ���5 | doi ��.��63/9789004300156_010

sumerian and akkadian in the early second millenium bce

159

from the city of Sippar, a genre removed from the social context that produced much of the metalinguistic data. 1.1 Languages in Contact Several studies have analyzed the effects of language contact on the linguistic structures of Sumerian and Akkadian, including phonology, morphology, syntax, and the lexicon.2 The majority of these language contact studies examine how language structures change as a result of contact. My analysis follows Y. Matras (2009, 2010) in taking a functional approach to language contact, based on the assumption that multilingual speakers (or writers) “have a complex repertoire of linguistic structures at their disposal” that does not conform to rigid language systems. Such language users select contextually appropriate structures, including levels of mixing, according to habitually ingrained practices of language socialization (Matras 2009: 4). The present contribution focuses on variation in a multilingual society rather than on structural changes resulting from contact. This study takes writing practices as the focus of investigation. Very few studies on language contact recognize the importance and complications of the written medium, focusing instead on the effects of language contact on speech practices.3 The effects of Sumerian-Akkadian contact are often obfuscated by the writing system, most notably via alloglottography (see especially Rubio 2006b). Both Sumerian and Akkadian are written in cuneiform, a polyvalent

2  Studies on the structural effects of contact on Akkadian and Sumerian include Oberhuber (1981), Haayer (1986), Pedersén (1989), Karahashi (2008) and Zólyomi (2011). Loanwords and lexical borrowing between the two languages has been the object of considerable study. D. O. Edzard (2003, 178) estimates that 7% of Akkadian vocabulary are Sumerian loans. As P. Michalowski notes, such a small percentage does not indicate mass lexical interference (Michalowski 2006: 169). The opposite phenomenon, Semitic/Akkadian loans in Sumerian, has received extended attention in recent years (see Krebernik 1998: 265; Sommerfeld 2006; Civil 2007). These studies, especially Civil (2007), demonstrate the pervasive extent of contact. One suspects that careful attention would demonstrate further Semitic loans in Sumerian that have undergone alterations and are therefore less easily recognized. For example, Sumerian mud ‘blood’ may be a loan from Akkadian dāmu—‘blood’, with phonological metathesis. Moreover, some Semitic loans are represented in Sumerian as a noun—verb lexically compounded phrase, such as Sumerian šu—ḫu-uz ‘to burn’ from Akkadian šūḫuzu ‘to flare up’ or Sumerian saŋ—rig₇ ‘to give’ from Akkadian šarāku ‘to give’ (on such loans, see provisionally Karahashi 2000; Edzard 2003: 175–76; Karahashi 2004). 3  On the issue of the written medium as object of historical linguistic study, see further Romaine (1982: 14–21) and, more recently, Coulmas (2013).

160

Crisostomo

and polysemic writing system.4 Sumerian, however, tends to use signs to represent individual words; Akkadian tends to write words syllabically. As a result of contact, Akkadian words may be represented by the sign used for a semantically or phonologically corresponding Sumerian word. Thus, the sign MU (𒈬) may be used to write Sumerian mu ‘name’ as well as the semantically corresponding Akkadian šumu ‘name’. This micro-alloglottography is common in almost all Akkadian texts, beginning in the latter half of the third millennium. Some texts extend alloglottography over entire clauses or even a whole text. In a recent contribution on contact languages, L. Johanson examines “ways of writing in which the genre requires some kind of intertwining” (Johanson 2013: 273). Johanson categorizes Sumerian and Akkadian alloglottography as his Type E, which he describes as “Elements of a higher-ranking code are used in texts to represent a lower-ranking code” (Johanson 2013: 274). That is, the higher code graphically represents the lower code, which is read aloud. As Johanson notes, “This technique manifests a loyalty to a high-ranking code and its script regardless of the acts of reading. . . . The higher-ranking codes used for writing . . . offer cultivated formal ways of expression” (Johanson 2013: 314). In Johanson’s rankings, Type E reflects the category in which the lower-ranking code demonstrates highest degree of passivity. That is, the higher-ranking code is expressed representing the lower-ranking code; the lower-ranking code is completely obscured. Johanson’s discussion provides a typological background under which to organize and analyze the written expressions of Sumerian and Akkadian contact. 1.2 Language Ideologies P. Kroskrity discussed language ideologies in a recent overview as follows: [B]eliefs, feelings, and conceptions about language structure and use which often index the political economic interests of individual speakers, ethnic and other interest groups, and nation states. These conceptions, whether explicitly articulated or embodied in communicative practice, represent incomplete, or ‘partly successful’, attempts to rationalize language usage; such rationalizations are typically multiple, context-bound, and necessarily constructed from the sociocultural experience of the speaker. (Kroskrity 2010: 192)

4  For brief introductions to the cuneiform writing system, see Michalowski, Cooper, and Gragg (1996), Woods (2010), or Finkel and Taylor (2015).

sumerian and akkadian in the early second millenium bce

161

Kroskrity’s characterization provides a working definition of this analytical tool.5 Moreover, language attitudes are significant social predictors in contactinduced change (Thomason 2001: 77–85). As I show in this study, such ideologies are embedded in metadiscourse and usage practices most apparently with reference to Sumerian. Ideologies for Akkadian are evident in regular use, especially in contrast with the use of Sumerian. 2

Linguistic Context

The Mesopotamian linguistic area had a long tradition of multilingualism (Vanstiphout 1999). As early as the first half of the third millennium BCE, textual evidence indicates that Semitic and Sumerian language communities coexisted in roughly the same geographic location. At the site of Tell Abū Ṣalābīḫ (approximately 130 km SE of Baghdad), texts were composed in Sumerian by scribes with linguistically Semitic names (Biggs 1967). SumerianSemitic language contact almost certainly predates the inscribed texts from the third millennium. M. Civil (2007), in particular, shows that Sumerian borrowed from Semitic early in its history, arguing that the number of Semitic loans into Sumerian is “significant and consonant with the coexistence of Sumerian- and Semitic-speaking populations that must have taken place from the dawn of history” (Civil 2007: 17), echoing earlier observations by J. Cooper (1973: 239–40). Moreover, early multilingualism, despite the trends in the textual record, cannot be restricted to bilingualism. As G. Rubio notes, “The picture of the linguistic situation of Mesopotamia of early periods should be that of fluidity, of words traveling together with the objects and techniques they designate, of different languages and their dialects (most of which have left no traces or just a few . . . in surviving language), all of them sharing the same space and perhaps even sometimes the same speakers” (Rubio 1999: 11). Sumerian, an isolate, is the first attested language in the region. In the latter half of the third millennium, Akkadian emerged in administrative writing as the language of a new dynasty.6 The Akkadian languages, including both 5  For discussion of approaches to language ideologies and attitudes in historical linguistics, see Milroy (2012) with literature. 6  Assyriologists are often lax in their descriptions of this language group. It is collectively known in the literature as Akkadian, but there are numerous variations (dialectal? geographic?). The most prominent are what are known as Babylonian, the dialect (for lack of a better term) generally used in the southern part of Mesopotamia beginning in the late third millennium, and Assyrian, the dialect generally used in the northern part of Mesopotamia.

162

Crisostomo

Assyrian and Babylonian forms, constitute a branch of the Semitic family. Sometime in the last centuries of the third millennium, Sumerian likely died out as a vernacular (Cooper 1973; Michalowski 2006; Rubio 2006a), yet continued to be written in certain social contexts.7 Multiple languages were in use in southern Mesopotamia during the early second millennium BCE, the period known as the Old Babylonian period (henceforth OB). Four are attested in writing: Sumerian, Akkadian, Hurrian, and Elamite. Another, Amorite, is attested primarily in personal names, but not in contextual usage. Other languages beyond these five were almost certainly used in society, but have not been preserved in writing and only sporadically in personal names. During the OB period, the true extent of multilingualism is obfuscated by the shifting statuses of languages. Three different languages clearly operated in various levels of society. Akkadian was presumably the predominantly spoken language, largely used to record personal letters, administrative and legal texts in addition to some literary works, mathematical practice texts, and divinatory treatises. Sumerian, though no longer a widely spoken vernacular—if spoken at all—rose to prominence as a literary and cultural language. Sumerian was confined to education, magical incantations, liturgy, and specific administrative documents. Amorite, a West Semitic language (as opposed to the East Semitic branch to which Akkadian belongs) was the ancestral language of the ruling family in Babylon; however, no textual material has been preserved. Some form of Amorite was apparently spoken by non-sedentary populations in the Mari region, along the bend of the Euphrates. It is not clear how much Amorite was used in urban areas in southern Mesopotamia. The exact character of the contact relationship among these three languages is difficult to quantify. Amorite and Akkadian are typologically similar (both fall within the

Despite problems with the nomenclature, I adopt the terminology of the field in the present study. 7  The death of Sumerian as a vernacular is a contentious topic in Assyriology. Until the late 1990s, the clear majority position was that Sumerian was dead or dying in the last century of the third millennium. Contrary opinions have arisen that at least challenge that consensus (Sallaberger 2004; Edzard 2005; Woods 2006), bringing the death of Sumerian as a spoken language into the first few centuries of the second millennium. Woods’ (2006) suggestion that Sumerian was spoken in some parts, albeit not widely, as late as the eighteenth century cannot be dismissed out of hand. The extant data, however, does not deal with any possible outliers.

sumerian and akkadian in the early second millenium bce

163

Semitic language family), whereas Sumerian differs drastically from either.8 Elamite and Hurrian, unrelated to the previous three, appear in incantations and personal names.9 Other languages almost certainly could have been heard on some streets and social situations. The multilingual character of the OB period thus presents a case for the analysis of ideology and differentiation (Gal and Irvine 1995; Irvine and Gal 2000), but the textual record preserves only the ideological distinction between Akkadian and Sumerian. 3

Metalinguistic Data and Language Ideology

In this section, I discuss the metalinguistic data from the OB period, which indicates the social position of Amorite, Akkadian, and Sumerian. I show that, at least for the scribal elite, Sumerian was viewed as a language of culture and education. 3.1 Amorite Little is known about Amorite linguistically (Durand 2012). A recent publication (Ziegler and Charpin 2007) elucidates some prevailing linguistic attitudes about the language. Shamshi-Addu, the ruler of the northern Mesopotamian kingdom Ekallātum, in a letter written in Akkadian to his son, Yasmakh-Addu, ruler of the city Mari to the west of the Mesopotamian heartland, berates his son for asking for someone to presumably teach him Sumerian, instead of learning the more practical language, Amorite: ‘Regarding (the request) you sent me to send you one man who (can) read Sumerian, why do you want a man who (can) read Sumerian? Learn to speak Amorite!’ (M.7930+M.8157).

8  Since little is known about Amorite, my attention focuses primarily on the difference between Akkadian and Sumerian including the following generalities: Sumerian is considered a split-ERGATIVE language, whereas Akkadian is a NOMINATIVE-ACCUSATIVE language; Sumerian is agglutinative, Akkadian is inflectional; Akkadian nominals are marked with cases; Akkadian utilizes prepositions, Sumerian uses post-positions marked on noun phrases; both languages differ in phonemic inventory, with Sumerian’s less understood and the differences complicated by the shared writing system. Some, such as J. Høyrup (1992), have argued for Sumerian as a mixed language, but such discussions have not been widely accepted. For brief introductions to grammar, see Huehnergard and Woods (2004) for Akkadian and Michalowski (2004) for Sumerian. 9  For grammatical introductions, see Khačikjan (1998) or Stolper (2004) for Elamite and Wilhelm (2004) for Hurrian.

164

Crisostomo

In another letter, this time from Yasmakh-Addu to Shamshi-Addu, Yasmakh-Addu addresses his father’s complaint: ‘The tablet of my lord arrived here saying . . . “You are incapable of speaking Amorite with them (the nomads)!” . . . Without exaggeration, I will learn to speak Amorite!’ (A.3823). From these two letters, we see the ideological contrast between the three languages. Yasmakh-Addu, an Akkadian speaker, wishes to learn Sumerian, but should be learning Amorite. Moreover, from Shamshi-Addu’s viewpoint, Amorite would be more useful to Yasmakh-Addu in governing his kingdom. We see then a continuum of Amorite, the language of Yasmakh-Addu’s subjects, to Akkadian, the language of correspondence for the two rulers, to Sumerian, a specialists’ language with enough esteem that Yasmakh-Addu wanted to learn it.10 3.2 Sumerian Most metalinguistic data from this period concerns Sumerian. All of it emerges from a particular social field—scribal education. N. Veldhuis has described scribal education as a process wherein students, children of elite members of society, were enculturated into an imagined Sumerian cultural heritage (Veldhuis 2004). Scribal education was not merely training in literacy, but the routinization of habits focused on the Sumerian language—a language no longer useful in everyday practices—and the development of a bureaucratic “fraternity” (Michalowski 1987; 2012), an interpretive community shaped by shared experiences and knowledge. In short, scribal education laid the ideological ground for a Sumerian identity formation that indexed power, scholarly literacy, and tradition. It comes as no surprise, then, that the metalanguage derived from the literature of scribal education (and written almost exclusively in Sumerian) extols Sumerian as the premier language and reflexively emphasizes the necessity of Sumerian for belonging to the scribal community. The preeminence of Sumerian over all other languages is presented in a small section in the narrative Enmerkar and the Lord of Aratta. In this story, Enmerkar, the king of Uruk (the prototypical Sumerian city) challenges the ruler of Aratta, a mythical Shangri-la representing the wealth and splendor of the eastern lands. The point of the story is Uruk’s superiority over Aratta. In a much discussed, although relatively small section, the story implicitly aggrandizes the Sumerian language.

10  Very little metalinguistic data exists for Akkadian in this period, except in contrast to the other languages.

sumerian and akkadian in the early second millenium bce

165

At that time, there was no snake, there was no scorpion, there was no hyena, there was no lion, there was no dog or wolf, there was no fear or trembling—man had no rival. At that time, the lands of Shubur and Hamazi, the distinctly tongued; Sumer, the great mountain, the essence of nobility; Akkad, the land possessing the befitting, and the land of Martu (the Amorite land), lying in safety—the totality of heaven and earth, the well-guarded people, indeed they proclaimed (the great god) Enlil in a single language. At that time, the contest lord, the contest prince, the contest king, (the god) Enki, the contest lord, the contest prince, the contest king, the contest lord, the contest prince, the contest king, Enki, the lord of abundance and true word, the lord of wisdom who watches over the land, the expert of all the gods, the chosen of wisdom, the lord of Eridu, (Enki) set up the complete and total alteration of the language in their mouths. The speech of humanity is one. (ELA 136–55)11 According to the story, the god Enki made all people temporarily speak one language in order to facilitate the contest between the two rulers.12 Since the narrative is written in Sumerian and because the point of the story is the superiority of Uruk—representative of all Sumer—that one language is presumably Sumerian. Thus, Sumerian is foremost over all the languages of the region, including Akkadian and Amorite. Other metalinguistic examples explicitly regard Sumerian as a requisite for true scholarship. Sumerian proverbs (SP) not only taught scribes Sumerian grammar (Veldhuis 2000), but also provided an opportunity to reinforce ideology and particular social perspectives.13 11  The most recent edition of this story is Mittermayer (2009). See also the Electronic Text Corpus of Sumerian Literature 1.8.2.3 (etcsl.orinst.ox.ac.uk/) and Vanstiphout (2003). The translation here is my own. 12  The interpretation of this passage has been much debated in Assyriological literature. The interpretation presented here is detailed in Crisostomo (forthcoming). The two most prominent explanations are polar opposites: either Enki, in the past, divided all the languages (and thus a parallel to the biblical story of the Tower of Babel) or he, in the future, will make all people speak one language. No matter the interpretation, the mythical association of Mesopotamia with multilingualism stands and Sumerian emerges as the most prominent of those languages. For discussion, see Mittermayer (2009) and Keetman (2010) with previous literature. 13  See also Alster and Oshima (2006). Editions of the SP may be found in Alster (1997) and the Electronic Text Corpus of Sumerian Literature 6.1.02.

166

Crisostomo

A scribe who does not know Sumerian, what (kind of) scribe is he? (SP 2.47) A scribe who does not know Sumerian, how will he get a translation? (SP 2.49) In a group of satirical literary texts known as the Eduba texts (Eduba is the Sumerian word for school) which purportedly depict life in the school, Sumerian language ideology permeates the narratives and, consequently, the practices of the scribes depicted in the stories (and those copying the texts). This group of texts, although not indicative of reality, nevertheless articulate the ideology of the school (Volk 1996) and the principle subject in the school, Sumerian.14 Do you speak Sumerian as well as I do? (D3: 66) (Supervisor): Young man, are you a student? (Student): I am a student. (Supervisor): If you are a student, do you know Sumerian? (Student): I can speak (OR: translate?) Sumerian (Supervisor): You are young; how do you speak it (so well)? (Student): I have listened carefully to my teacher’s instructions (EdD: 1–6)15 (Student): I know all my scribal arts; I am not restricted by anything; my teacher shows me one wedge (and) I can add one or two from memory. I have sat (in school) for the complete duration (so) I wrestle with Sumerian, the scribal arts, interpretation, and accounts. I speak (OR: translate?) Sumerian. (EdD: 32–38) In the second and third examples above, both from the composition Eduba D, the two participants engage in a series of boasts about their competence in scholarly activities. Here, and similarly in the first example, ability in Sumerian 14  Some of the compositions cited still await publication in critical editions. Lines cited may be found primarily in Sjöberg (1976) or Volk (1996). Exceptions are noted. This category of Eduba texts comprises Eduba A (EdA): Schooldays; Eduba B (EdB): The young scribe and his father; Eduba C (EdC): A supervisor’s advice to a young scribe; Eduba D (EdD): Scribal activities; Eduba R (EdR): Eduba regulations; Dialogue 1 (D1): Dialogue between two scribes; Dialogue 2 (D2): Dialogue between Enkihegal and Enkitalu; Dialogue 3 (D3): Dialogue between Enkimansum and Giriniisag. 15  For an edition of the first seventy-one lines of this composition, see Civil (1985).

sumerian and akkadian in the early second millenium bce

167

is the first issue addressed, indicating its basic importance to the scribal arts. Excellent knowledge of Sumerian represented the basic prerequisite for being a scribe. Similar compositions depict the two interlocutors debating each others’ merits as a scribe. One of the basic insults is to question one’s capacity to understand Sumerian. (Student): I speak (OR: translate?) Sumerian (Supervisor): Indeed, yet Sumerian is obscure to you! (EdD: 38–39) Your tongue does not produce Sumerian correctly! (D1: 56) He is “heavy” for Sumerian; His tongue does not move correctly (D2: 97) (He is) deaf about the scribal arts, deaf about Sumerian (D3: 10) Direct opposition between Sumerian and Akkadian is provided in the composition Eduba A. This satirical composition traces a student’s day in school, during which he is constantly accused of doing something incorrectly by the staff/faculty and punished. Among these offenses: The one (in charge of) Sumerian (said): “He spoke Akkadian!” Then he caned me. (EdA: 40) The teacher admonishes the student for speaking Akkadian, implying that doing so in school was improper and needed correction (see Kroskrity 1998: 107–9). Similarly, from a literary letter also composed within the field of scribal education: Do not neglect Sumerian! Once more, I am responding to you in correct language. (Sumerian Epistolary Miscellany 22: 3–4)16 Here, the letter writer seems to imply that a previous correspondence from the addressee was written in a language other than Sumerian. The writer corrects the addressee and composes his letter properly in Sumerian. 16  This humorous, artificial letter concerns students attempting to fool the teacher into letting them leave school with falsified parental notes (see Kleinerman 2011: 181–84). M. Civil (2000: 108–09) interprets ‘correct language’ (eme si-sa₂) here as grammatically correct, rather than ethically or socially correct.

168

Crisostomo

3.3 Conclusions from the Metalinguistic Data Many other references could be brought forward as evidence, but the data from Enmerkar and the Lord of Aratta, Sumerian Proverbs, and the set of Eduba texts provide typical examples of metadiscourse on the ideology of Sumerian. In the corpus of Sumerian literature which existed in the social life of scribal education, competence in Sumerian served as the most basic aptitude required for admission into the scribal community. As such, the metalinguistic data portray Sumerian as distinct among languages. Moreover, the use of any language other than Sumerian in school was deemed inappropriate. The limits of metalinguistic discourses are well-established (see Silverstein 1981; Lucy 1993; Silverstein 1993, 1998).17 The metadiscourses presented here, particularly those focused on Sumerian, offer idealistic portrayals stemming from a social field attempting to reflexively define its uniqueness and viability. The ideologies represented with such bias should be regarded as belonging strictly to the field of scribal education. Any assertion that such ideologies which posture Sumerian as the preeminent language represent language ideology in the broader multilingual society should be regarded with skepticism. Such metadiscourse represents only one aspect in our attempt to analyze (and reconstruct) language ideology. The application of sociolinguistic hierarchical values of prestige or higher versus lower codes does not always apply during the OB, nor are the values consistent among all discourse genres. Personal letters and divinatory compendia, for example, are written exclusively in Akkadian implying that Akkadian was appropriate linguistic code for those genres. The metalinguistic data from scribal education, however, explicates Sumerian as a higher ranked language due to its association with traditional scribal culture. The following section tests whether the cultural prestige of Sumerian may be detected beyond the realms of a single social field. 4

Language in Use

The native metalinguistic data present a picture that glorifies Sumerian— which by the time of the OB period was no longer productive—over the languages of everyday use, most notably Akkadian. In this section, I contrast 17  Following Silverstein’s (1998) distinctions, I have provided a group-level representation of metadiscourse—ideologies intended to differentiate social groups—and an example of nomic calibration in the mythical example above, a rationalized basis for the language ideology.

sumerian and akkadian in the early second millenium bce

169

tokens of language use in texts with the ideological picture given in literary texts produced by a social field which had reason to reinforce the importance of Sumerian. I consider a genre removed from the world of scribal education, legal documents, focusing on variation in the phrase (Sum.) mu—pad = (Akk.) nīš—itma ‘(s)he swore an oath’. I also provide a few examples of changes in language structure represented in or resulting from the phrase under examination. In this analysis, I show that socially-conditioned “ways of speaking” documented in language use in context reinforces literary ideological portrayals of Sumerian. 4.1 Writing Variation in Oath Taking Apart from letters and administrative documents, legal contracts provide the closest available data for accessing regular social life. In contracts, a principle ratified certain stipulations by taking an oath, swearing on the king, city, and/ or gods. In Sumerian, this oath was formulated via the base phrase mu—pad ‘to name the name’ (literally): mu dutu dmarduk hammurapi u irisipparki mu utu marduk hammurapi u sippar.(ak).Ø name DN DN RN CONJ SN.GEN.ABS in-pad3 i.n.pad.Ø NTRL.ERG.swear[-IMPF].ABS ‘He swore on the gods Utu (and) Marduk, king Hammurapi, and the city Sippar.’ The corresponding Akkadian phrase was nīš—itma ‘swear on the life’ (lit.): nīš dutu dmarduk hammurapi u irisipparki life DN DN RN CONJ SN.GEN.ACC itma swear[3;SG, -IMPF] ‘He swore on the gods Utu (and) Marduk, king Hammurapi, and the city Sippar.’ In this period, a type of orthographic code-switching can occur: mu dutu dmarduk hammurapi u irisipparki itma ‘He swore on the gods Utu (and) Marduk, king Hammurapi, and the city Sippar.’

170

Crisostomo

Moreover, by analogy with Akkadian, which had a suffixed plural morpheme on the verb, and a misunderstanding (or, perhaps, adaptation) of the Sumerian plural copula me.eš, the phrase with plural actors could be rendered in a morphologically variant contrived Sumerian form in equal distribution with the regular plural form:18 mu dutu dmarduk hammurapi u irisipparki mu utu marduk hammurapi u sippar.(ak).Ø name DN DN RN CONJ SN.GEN.ABS in-pad3-de3-meš i.n.pad.e.meš NTRL.ERG.swear[-IMPF].PL ‘They swore on the gods Utu (and) Marduk, king Hammurapi, and the city Sippar.’ For this study, I surveyed 844 real estate documents from the city of Sippar (30 km SW of Baghdad) from ca. 1900–1600 published by L. Dekiere (1994). Dekiere published a corpus of texts now located in the British Museum and is, thus, a representative sample. I follow Dekiere in separating the texts into four distinct chronological groups: Pre-Hammurapi (Pre-Ha; ca. 1900–1792), Hammurapi (Ha; 1791–1750), Samsu-iluna (Si; 1749–1712), and Post-Samsu-iluna (Post-Si; 1711–1595). I counted as evidence only those instances in which both parts of the verbal phrase are clearly present and obvious as to their form. That is, I omitted any examples that, due to damage to the material object, required reconstruction of either part of the phrase. Moreover, I counted envelopes

18  Literary Sumerian and Sumerian contracts from earlier periods would render the verb:  in-pad3-de3-eš  i.n.pad.eš Here the combination of pre-verbal affix [n] plus the post-verbal affix [eš] indicates the plural AGENT. The Sumerian form me.eš is the verb ‘to be’ with a plural AGENT marker. Beginning in the early second millennium, the form meš is used, mostly in Akkadian contexts (especially when using Sumerian glyphs to represent Akkadian words), to indicate plurality. Early and literary Sumerian typically did not mark plurality or, if marked, plurality would be indicated by reduplication (diŋir diŋir ‘gods’) or a suffixed morpheme for animates [ene] (diŋir.ene ‘gods’). Speakers of Akkadian, which marked plurality inflectionally within its morphology, were apparently dissatisfied with the lack of a plural marker for certain noun classes, e.g., cities, and began using the plural form of the copular verb as a plural marker (iri.meš ‘cities’) and eventually extended its use to all noun classes (diŋir.meš ‘gods’) and, as in the above examples, some conventional verbal expressions.

sumerian and akkadian in the early second millenium bce

171

as separate attestations.19 Even though the envelopes routinely duplicated the tablet, there were at least a dozen examples of alternations in this phrase between tablet and envelope. I further sub-divided the texts into genres such as sales, litigation, inheritance, and adoption; these categorizations were not immediately productive. I found 234 admissible tokens of the oath clause. The texts are unevenly distributed diachronically. Moreover, the sample is too small from which to draw statistically reliable conclusions, but the trends represented are nevertheless suggestive. Based on a brief survey of further textual data from the site, further study is likely to present similar data.20 Other cities offer different writing practices, but only the Sippar material allows for the diachronic analysis adopted here. The corpus of investigation presents several difficulties for linguistic analysis. One issue is the language under examination in a given text and the linguistic phenomena in operation. During the OB period, there is evidence that contracts seemingly written in Sumerian represent alloglottography and were read in Akkadian (Schorr 1909). Differences between some tablets and their envelopes—which ostensibly provide the same data—further confirm this understanding. For example, one tablet reads: mu dutu mu dmarduk zabium u irisipparki teš₂-bi in-pad₃-de₃-eš They altogether swore on the name of Utu, the name of Marduk, king Zabium and Sippar The envelope to this document reads: mu dutu dmarduk zabium u irisipparki itmû They swore on the name of Utu, Marduk, king Zabium, and Sippar Where the tablet writes the verb in Sumerian, the envelope writes the Akkadian equivalent. The choice to present stock phrases orthographically as Sumerian implies a stylistic variation, which may be attributed to an 19  Some documents were covered with a thin layer of clay on which the content of the internal document was duplicated, summarized, or excerpted. Envelopes served as a means of both legal and physical security. For further on envelopes, see Taylor (2011: 19–21). 20  For example, in the corpus examined in the sixth volume of Dekiere (1994), a slightly different corpus not included in my principal analysis, the percentages of variations follow. Pre-Ha (n = 1): variant (5) = 100%; Ha (n = 8): (3) = 37.5%, (1) = 37.5%, (4) = 25%; Si (n = 18): (3) = 27.8%, (4) = 72.2%; Post-Si (n = 8): (1) = 12.5%, (3) = 12.5%, (4) = 75%. See below for typologies of variations and analysis of the primary corpus.

172

Crisostomo

ideological stance. The practice of representing language in Sumerian further affirms the higher ranked cultural status of Sumerian during the OB, explicated in the literary texts attributed to the field of scribal education. In Johanson’s typology discussed above, Sumerian as the higher ranked linguistic code is contrasted with Akkadian, which is not overtly conveyed. The writing practices and the variations in representation further illuminate ideological expression. At issue, then, is not language or supposed speech practice, but rather writing practice. Given the above information, one can conceive six basic variations of the oath clause: (1) mu—pad. (2) nīš—itma. (3) mu—itma. (4) mu—i.n.pad.e.meš. (5) nīš—pad. (6) nīš—i.n.pad.e.meš.21 Of these possible variations, however, (5) and (6) were not found at all in this corpus. Additionally, variation (2), the unilingual Akkadian formation of the formula, is only found Pre-Ha. In fact, Pre-Ha shows the most variation of orthographies and conjugations.22 The final three periods show a movement towards variations (3) and (4), as shown in Figure 4. 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0%

Variant (4) Variant (3) Variant (2) Variant (1)

Pre-Ha. (ca. 1900–1792) n = 110

Ha (1791–1750) n = 59

Si (1749–1712) n = 47

Post-Si (1711–1595) n = 18

FIGURE 4  Variations in the oath formula.23 21  One regular variation is mu—Ø where the verbal element is elided. I did not admit this variation in my data. Another possible variation, the “native” Akkadian phrase Ø—itma, the verb as intransitive, is never used among the documents surveyed. 22  For example, the standard formula is aspectually PERFECTIVE. In Pre-Ha, some Sumerian versions of the formula are IMPERFECTIVE:  i.n.pad.e.Ø  N TRL.ACC.swear.IMPF.NOM[3;SG] One other variant in Pre-Ha is the nominalized form i.n.pad.a. 23  Alternative variations are omitted from the percentages in the chart. Only Pre-Ha used variations deviating from the norm, discussed in the preceding note.

sumerian and akkadian in the early second millenium bce

173

The greater degree of variation in Pre-Ha indicates that the formula was more flexible and not yet standardized. As time progresses, variant (1) is absorbed into variant (4), regardless of the number of actors. As shown in Figure 4, variants (3) and (4) are the preferred instantiations of the formula, especially towards the end of the period investigated. Variant (2) disappears entirely after Pre-Ha. Moreover, no forms with nīš are used after Pre-Ha. As mentioned earlier, neither variations (5) or (6) are found; that is, there are no combinations of Akkadian nīš with Sumerian pad.24 At this time, given the scope of this study, I cannot determine the pragmatics of variation nor can I detect patterns for ordered differentiation.25 The data suggest that writers of contracts at Sippar preferred more “Sumerian-ized” versions, particularly later in the OB. Thus, even in a genre far removed from the Sumerian-centric field of scribal education, orthographic conventions promote an ideology of Sumerian as an eminent cultural language. 4.2 Structral Changes The oath clause variations discussed above also indicate changes in linguistic structure. I briefly survey three such changes, including morphological, lexical, and graphic changes resulting from the contact of Sumerian and Akkadian. The form i.n.pad.e.meš (written in-pad₃-de₃-meš; see footnote 18) represents either a morphological change reflected in writing resulting from a nonnative misunderstanding or adaptation (see Thomason and Kaufman 1988; Thomason 2001: 66–76) of the Sumerian form me.eš—a form of the plural copula or verb ‘to be’ plus plural agent morpheme—or, less likely, a historical grammaticalization in written Sumerian. The writing i.n.pad.e.meš does not occur in Sumerian literary texts produced during the OB period nor in any genres from earlier periods. In such documents, the typical verbal form expressing a plural agent is written i.n.pad.eš. Beginning in the OB period, meš is regularly used to graphically mark plurality when a logogram is used to express an Akkadian word. Post OB, meš is also extended to mark plurality in Sumerian contexts. Although most often used on nouns, meš may also be used on verbs, as in the case of the oath phrase. As shown above, meš in the oath phrase is applied to all forms of the verb, regardless of the actual number 24  The formula zi—pad ‘to call on the life’ (literally), a more precise Sumerian equivalent of Akkadian nīš—itma, is used in the OB, but was not found in the corpus under investigation. Conversely, I find no instances of šum—itma ‘to swear on the name’, the more precise Akkadian equivalent to Sumerian mu—pad. 25  Not enough data exists to determine variation along traditional sociolinguistic markers such as sex or gender, age, or class.

174

Crisostomo

of agents. That is, the form i.n.pad.e.meš occurs with both singular and plural agents. The use of meš in legal expressions seems to have been taught in schools. The Sippar Legal Phrasebook, a list of business expressions which student scribes copied as part of their education, gives verbal forms such as in-kur₂-e-me-eš ‘they will deny’, using meš to mark plurality.26 Literary Sumerian during the OB period would have written in-kur₂-re-eš or similar. The use of meš in a text from scribal education thus corresponds to the use of meš in legal contracts.27 The stock phrase nīš—itma ‘he swore on the life’ in these contracts is a pseudo-calque on the Sumerian phrase mu—pad. The Akkadian verb tamû semantically entails an act of oath-taking and does not require an object upon which to swear an oath. The influence of Sumerian mu—pad, an idiomatic phrase which typically requires both components, dictated that the equivalent Akkadian phrase also be represented with both a nominal and verbal element. Thus, in legal contracts, the full phrase nīš—itma is utilized, even if one or both parts is represented as the corresponding Sumerian mu—pad. As a result of the calque, the sign form MU (𒈬)—which signifies Sumerian mu ‘name’ and, in the oath phrase, Akkadian nīšu ‘life’—adopts further signifieds in post-OB texts. For example, a mid-first millennium prayer uses the esoteric writing: ina . . . MU šU-MU to represent Akkadian ina . . . nīš qātiya ‘in . . . the raising of my hands (i.e., prayer)’. In this phrase, the first iteration of MU represents the Akkadian participle nīšu ‘raising’. The equivalency MU is nīšu ‘raising’ is based on the homophony of nīšu ‘raising’ and nīšu ‘life’ as found in the oath clause. Thus, the calque of mu—pad, brought into Akkadian as nīš—itma, allows the sign MU to represent the syllable /niš/ and is thus adapted to a different word nīšu ‘raising’ and new context. The graphic adaptation is an example of a contact-induced change in the semiotic system.

26  On the legal phrasebooks used in education, see Veldhuis (2014: 188–94). The tablet cited here for the Sippar Legal Phrasebook is HS 1702 (to be published in a forthcoming volume of the Texte Materialen Hilprecht-Collection). For a preliminary edition, see the Digital Corpus of Cuneiform Lexical Texts the Sippar Legal Phrasebook (oracc.org/dcclt/ Q000260). 27  Similar phrases are found in the localized Nippur Legal Phrasebook (oracc.org/dcclt/ Q000045), such as in-na-an-la₂-me-eš ‘they paid him’. Notably, to my knowledge, neither legal contracts nor model contracts—used in schools to practice writing legal contracts (oracc.org/obmc)—from the city of Nippur use meš to mark plurality as discussed above. In other words, grammatical forms taught in the lists in schools at Nippur were not utilized in writing contracts at the same city.

sumerian and akkadian in the early second millenium bce

175

4.3 Conclusions from Language in Use The analysis of real estate contracts from Sippar presented above suggests that writing practices of the oath clause favored Sumerian forms over Akkadian forms. Moreover, the data further imply movement toward increased use of Sumerian forms over time. In addition to variation, the above discussion of the oath clause demonstrates some examples of contact-induced structural change. 5 Conclusions The study of contact between Sumerian and Akkadian presents complex challenges. Among these complexities are the obscurities caused by the writing system and modern imperfect understanding of Sumerian linguistic structure, a problem exacerbated by the Akkadian lens through which we understand Sumerian. Despite such complications, sociolinguistic research on the effects and outcomes of contact on these two languages remains a viable pursuit. The present study has attempted initial research on the topic by demonstrating that linguistic ideologies prominent in literary texts are reflected also in the writing practices used in non-literary texts, particularly legal contracts. The ideology of Sumerian as an esteemed cultural language pervaded society and affected writing practices beyond the social field which produced explicit metadiscourse on the subject. Future analyses of variation may further elucidate and qualify the nature of multilingualism in the Mesopotamian linguistic area and present a clearer picture of the earliest documented case of language contact. Bibliography Alster, Bendt. 1997. Proverbs of Ancient Sumer: The World’s Earliest Proverb Collections (2 vol.) Bethesda. Alster, Bendt and Takayoshi Oshima. 2006. “A Sumerian Proverb Tablet in Geneva, with Some Thoughts on Sumerian Proverb Collections,” Orientalia 75: 31–72. Biggs, Robert D. 1967. “Semitic Names in the Fara Period,” Orientalia 36: 55–66. Civil, Miguel. 1985. “Sur les ‘Livres d’écolier’ à l’époque Paléo-Babylonienne,” in JeanMarie Durand and Jean-Robert Kupper (ed.), Miscellanea Babylonica: Mélanges offerts a Maurice Birot. Paris. 67–78. ———. 2000. “From the Epistolary of the Edubba,” in A. R. George and I. L. Finkel (ed.), Wisdom, Gods and Literature: Studies in Assyriology in Honour of W. G. Lambert. Winona Lake. 105–18.

176

Crisostomo

———. 2007. “Early Semitic Loanwords in Sumerian,” in Martha T. Roth, Walter Farber, Matthew W. Stolper, and Paula von Bechtolsheim (eds.), Studies Presented to Robert D. Biggs June 4, 2004 (Assyriological Studies 27). Chicago. 11–33. Cooper, Jerrold S. 1973. “Sumerian and Akkadian in Sumer and Akkad,” Orientalia 42: 239–46. Coulmas, Florian. 2013. Writing and Society: An Introduction (Key Topics in Socio­ linguistics). Cambridge. Crisostomo, C. Jay. forthcoming. “ ‘Recount for Him the Spell of Nudimmud’ . . . yet again: Reassessing the Babylonian Babel,” JAOS. Dekiere, Luc. 1994. Old Babylonian Real Estate Documents (6 vol.; Mesopotamian History and Environment: Texts). Ghent. Durand, Jean-Marie. 2012. “Réflexions sur un fantôme linguistique,” in Catherine Mittermayer and Sabine Ecklin (ed.), Altorientalische Studien zu Ehren von Pascal Attinger: mu-ni u₄ ul-li₂-a-aš ĝa₂-ĝa₂-de₃ (Orbis Biblicus et Orientalis 256). Fribourg. 165–91. Edzard, Dietz Otto. 2003. Sumerian Grammar (Handbuch der Orientalistik: Der Nahe und der Mittlere Osten). Leiden. ———. 2005. “Wann ist Sumerisch als gesprochene Sprache ausgestorben?,” Acta Sumerologica 22: 53–70. Finkel, Irving and Jonathan Taylor. 2015. Cuneiform. London. Gal, Susan and Judith Irvine. 1995. “The Boundaries of Languages and Disciplines: How Ideologies Construct Difference,” Social Research 62: 967–1001. Haayer, G. 1986. “Languages in Contact: The Case of Akkadian and Sumerian,” in H. L. J. Vanstiphout, K. Jongeling, F. Leemhuis and G. J. Reinink (eds.), Scripta, Signa, Vocis: Studies about Scripts, Scriptures, Scribes and Languages in the Near East, presented to J. H. Hospers by his pupils, colleagues and friends. Groningen. 77–84. Høyrup, Jens. 1992. “Sumerian: The Descendant of a Proto-Historical Creole? An alternative approach to the ‘Sumerian problem’,” AION 14: 21–72. Huehnergard, John and Christopher Woods. 2004. “Akkadian and Eblaite,” in Roger D. Woodard (ed.), The Cambridge Encyclopedia of the World’s Ancient Languages. Cambridge. 218–87. Irvine, Judith and Susan Gal. 2000. “Language ideology and linguistic differentiation,” in Paul V. Kroskrity (ed.), Regimes of Language: Ideologies, Polities, and Identities. Santa Fe. 35–83. Johanson, Lars. 2013. “Written Language Intertwining,” in Peter Bakker and Yaron Matras (eds.), Contact Languages: A Comprehensive Guide (Language Contact and Bilingualism 6). Berlin. 273–31. Karahashi, Fumi. 2000. Sumerian Compound Verbs with Body-Part Terms. Ph.D. Diss., University of Chicago, Chicago.

sumerian and akkadian in the early second millenium bce

177

———. 2004. “Some Observations on Sumerian Compound Verbs,” ORIENT 39: 96–110. ———. 2008. “Sumerian Enclitic -àm and Akkadian Enclitic -ma: From Copula to Focus Marker,” in Robert D. Biggs, Jennie Myers, and Martha T. Roth (eds.), Proceedings of the 51st Rencontre Assyriologique Internationale Held at the Oriental Institute of the University of Chicago July 18–22, 2005 (Studies in Ancient Oriental Civilization 62). Chicago. 85–91. Keetman, Jan. 2010. “Enmerkar und Šulge als sumerische Muttersprachler nach literarischen Quellen,” ZA 100: 15–31. Khačikjan, M. 1998. The Elamite Language (Documenta Asiana 4). Rome. Kleinerman, Alexandra. 2011. Education in Early 2nd Millennium BC Babylonia: The Sumerian Epistolary Miscellany (Cuneiform Monographs 42). Leiden. Krebernik, Manfred. 1998. “Die Texte aus Fāra und Tell Abū Ṣalābīḫ,” in Pascal Attinger and Markus Wäfler (eds.), Mesopotamien: Späturuk-Zeit und Frühdynastische Zeit (Orbis Biblicus et Orientalis 160/1). Fribourg. 235–427. Kroskrity, Paul V. 1998. “Arizona Tewa Kiva Speech as a Manifestation of a Dominant Language Ideology,” in Bambi B. Schieffelin, Kathryn A. Woolard, and Paul Kroskrity (eds.), Language Ideologies: Practice and Theory. Oxford. 103–22. ———. 2010. “Language Ideologies—Evolving Perspectives,” in Jürgen Jaspers, Jan-Ola Östman, and Jef Verschueren (eds.), Society and Language Use. Amsterdam – Philadelphia. 192–210. Lucy, John A. 1993. “Reflexive Language and the Human Disciplines,” in John A. Lucy (ed.), Reflextive Language: Reported Speech and Metapragmatics. Cambridge. 9–32. Matras, Yaron. 2009. Language Contact (Cambridge Textbooks in Linguistics). Cambridge. ———. 2010. “Contact, Convergence, and Typology,” in Raymond Hickey (ed.), The Handbook of Language Contact. West Sussex. 66–85. Michalowski, Piotr. 1987. “Charisma and Control: On Continuity and Change in Early Mesopotamian Bureaucratic Systems,” in McGuire Gibson and Robert D. Biggs (eds.), The Organization of Power: Aspects of Bureaucracy in the Ancient Near East (Studies in Ancient Oriental Civilization 46). Chicago. 45–57. ———. 2004. “Sumerian,” in Roger D. Woodard (ed.), The Cambridge Encyclopedia of the World’s Ancient Languages. Cambridge. 19–59. ———. 2006. “The Lives of the Sumerian Language,” in Seth L. Sanders (ed.), Margins of Writing, Origins of Cultures (Oriental Institute Seminars 2). Chicago. 159–84. ———. 2012. “Literacy, Schooling and the Transmission of Knowledge in Early Mesopotamian Culture,” in W. S. van Egmond and W. H. van Soldt (eds.), Theory and Practice of Knowledge Transfer. Studies in School Education in the Ancient Near East and Beyond. Papers Read at a Symposium in Leiden, 17–19 December 2008. Leiden. 39–57.

178

Crisostomo

Michalowski, Piotr, Jerrold Cooper, and Gene Gragg. 1996. “Mesopotamian Cuneiform,” in Peter T. Daniels and William Bright (eds.), The World’s Writing Systems. Oxford. 33–72. Milroy, James. 2012. “Sociolinguistics and Ideologies in Language History,” in Juan Manuel Hernández-Campoy and Juan Camilo Conde-Silvestre (eds.), The Handbook of Historical Sociolinguistics (Blackwell Handbooks in Linguistics). West Sussex. 571–84. Mittermayer, Catherine. 2009. Enmerkara und der Herr von Arata: Ein ungleicher Wettstreit (Orbis Biblicus et Orientalis 239). Fribourg. Oberhuber, Karl. 1981. “Kontaktwirkungen der Symbiose Sumerisch-Akkadisch: Bemerkungen zum akkadischen Lehngut im Sumerischen,” in Al-Hudhud: Festschrift Maria Höfner zum 80. Geburtstag. Graz. 257–61. Pedersén, Olof. 1989. “Some Morphological Aspects of Sumerian Akkadian Linguistic Areas,” in Hermann Behrens, Darlene Loding, and Martha T. Roth (eds.), Dumu-E2DUB-BA-A. Studies in Honor of Åke W. Sjöberg. Philadelphia. 429–38. Romaine, Suzanne. 1982. Socio-Historical Linguistics (Cambridge Studies in Linguistics 34). Cambridge. Rubio, Gonzalo. 1999. “On the Alleged ‘Pre-Sumerian Substratum’,” JCS 51: 1–16. ———. 2006a. “Šulgi and the Death of Sumerian,” in Piotr Michalowski and Niek Veldhuis (eds.), Approaches to Sumerian Literature. Studies in Honour of Stip (H. L. J. Vanstiphout) (Cuneiform Monographs 35). Leiden. 167–79. ———. 2006b. “Writing in Another Tongue: Alloglottography in the Ancient Near East,” in Seth L. Sanders (ed.), Margins of Writing, Origins of Cultures (Oriental Institute Seminars 2). Chicago. 33–70. Sallaberger, Walther. 2004. “Das Ende des Sumerischen: Tod und Nachleben einer altmesopotamischen Sprache,” in P. Schrijver and P.-A. Mumm (eds.), Sprachtod und Sprachgeburt. Bremen. 108–40. Schorr, M. 1909. “Das Sumerische in den Rechturkunden der Ḫammurabi-Periode,” in Hilprecht Anniversary Volume. Studies in Assyriology and Archaeology. Leipzig. 20–32. Silverstein, Michael. 1981. The Limits of Awareness. Sociolinguistic Working Paper, number 84. Austin: Southwest Educational Development Laboratory. ———. 1993. “Metapragmatic Discourse and Metapragmatic Function,” in John A. Lucy (ed.), Reflexive Language: Reported Speech and Metapragmatics. Cambridge. 33–58. ———. 1998. “The Uses and Utility of Ideology: A Commentary,” in Bambi B. Schieffelin, Kathryn A. Woolard, and Paul V. Kroskrity (eds.), Language Ideologies: Practice and Theory. Oxford. 123–45.

sumerian and akkadian in the early second millenium bce

179

Sjöberg, Åke W. 1976. “The Old Babylonian Eduba,” in Stephen J. Lieberman (ed.), Sumerological Studies in Honor of Thorkild Jacobsen on His Seventieth Birthday, June 7, 1974 (Assyriological Studies 20). Chicago. 159–79. Sommerfeld, Walter. 2006. “Die ältesten semitischen Sprachzeugnisse—eine kritische Bestandsaufnahme,” in G. Deutscher and N. J. C. Kouwenberg (eds.), The Akkadian Language in its Semitic Context: Studies in the Akkadian of the Third and Second Millennium BC (PIHANS 106). Leiden. 30–75. Stolper, Matthew W. 2004. “Elamite,” in Roger D. Woodard (ed.), The Cambridge Encyclopedia of the World’s Ancient Languages. Cambridge. 60–94. Taylor, Jon. 2011. “Tablets as Artefacts, Scribes as Artisans,” in Karen Radner and Eleanor Robson (eds.), The Oxford Handbook of Cuneiform Culture. Oxford. 5–31. Thomason, Sarah G. 2001. Language Contact: An Introduction. Edinburgh. Thomason, Sarah Grey and Terrence Kaufman. 1988. Language Contact, Creolization, and Genetic Linguistics. Berkeley. Vanstiphout, H. L. J. 1999. “The Twin Tongues. Theory, technique, and practice of bilingualism in Ancient Mesopotamia,” in H. L. J. Vanstiphout (ed.), All Those Nations . . . Cultural Encounters withing and with the Near East. Studies presented to Han Drijvers at the occasion of his sixty-fifth birthday by colleagues and students. Groningen. 141–59. ———. 2003. Epics of Sumerian Kings: The Matter of Aratta (Writings from the Ancient World 20). Atlanta. Veldhuis, Niek. 2000. “Sumerian Proverbs in their Curricular Context,” JAOS 120: 383–99. ———. 2004. Religion, Literature, and Scholarship: The Sumerian Composition Nanše and the Birds. With a Catalogue of Sumerian Bird Names (Cuneiform Monographs 22). Leiden. ———. 2014. History of the Cuneiform Lexical Tradition (Guides to the Mesopotamian Textual Record 6). Münster. Volk, Konrad. 1996. “Methoden altmesopotamischer Erziehung nach Quellen der altbabylonischen Zeit,” Saeculum 47: 178–216. Wilhelm, Gernot. 2004. “Hurrian,” in Roger D. Woodard (ed.), The Cambridge Encyclopedia of the World’s Ancient Languages. Cambridge. 95–118. Woods, Christopher. 2006. “Bilingualism, Scribal Learning, and the Death of Sumerian,” in Seth L. Sanders (ed.), Margins of Writing, Origins of Cultures (Oriental Institute Seminars 2). Chicago. 91–120. ———. 2010. “The Earliest Mesopotamian Writing,” in Christopher Woods (ed.), Visible Language: Inventions of Writing in the Ancient Middle East and Beyond (Oriental Institute Museum Publications 32). Chicago. 33–50.

180

Crisostomo

Ziegler, Nele and Dominique Charpin. 2007. “Amurritisch Lernen,” in Festschrift für Hermann Hunger zum 65. Geburtstag gewidmet von seinen Freunden, Kollegen und Schülern (Wiener Zeitschrift für die Kunde des Morgenlandes 97). Wien. 55–77. Zólyomi, Gábor. 2011. “Akkadian and Sumerian Language Contact,” in Stefan Weninger (ed.), Semitic Languages. An International Handbook (Handbücher zur Sprach- und Kommunikationswissenschaft 36). Berlin. 387–93.

Inner-Semitic Loans and Lexical Doublets vs. Genetically Related Cognates Lutz Edzard 1 Introduction A recent quotation pinpointing the intricate issue of inner-Semitic loans as opposed to mere cognates is the following: “The problem with Semitic loans is to distinguish them from cognates. And in respect of non-Semitic loans, if they cannot be tracked, it is not always possible to determine whether they are Kulturwörter, Wanderwörter, cross-cultural loanwords or even isoglosses” (Watson 2005: 194). The phenomenon of lexical doublets in the service of semantic differentiation or “fine-tuning” may complicate or even put into question traditional genetic subgrouping, at least as long as such subgrouping essentially rests on phonological criteria. In order to approach this subject, I shall first present a brief overview of typical loans, focussing on Akkadian, Hebrew, Arabic, and Amharic examples, loans from non-Semitic source languages as well as other Semitic source languages (i.e., inner-Semitic loans). In a further step, I shall look more closely at the issue of lexical doublets emerging in the process of inner-Semitic borrowing. Finally, I shall briefly consider whether the existence of such doublets can affect traditional models of subgrouping in Semitic. 2

Borrowing in Individual Semitic Languages

2.1 Akkadian Akkadian draws on a large stock of loaned vocabulary, mostly from Sumerian (see Lieberman 1977), but also other source languages. Here are a few examples: (1) Loanwords in Akkadian Akkadian Source word in the language of origin dû ‘hill’ Sumerian du ekallu ‘palace’ Sumerian é.gal ‘big house’ erēqu ‘flee’ Aramaic √ʿRQ

© koninklijke brill nv, leiden, ���5 | doi ��.��63/9789004300156_011

182

edzard

gadalû isinnu kasulatḫu kisallu laqāḫu paraššannu purkullu qaṭālu ṭupšarru uriḫullu

‘dressed in linen’ Sumerian gada lá ‘festival’ Sumerian ezen ‘a device of copper’ Hurrian kasulatḫ‘court’ Sumerian kisal ‘take’ Ugaritic √LQḤ ‘part of horse riding gear’ Hurrian paraššann‘stone cutter’ Sumerian bur-gul ‘kill’ Aramaic (Amurrite?) √QṬL ‘scribe’ Sumerian dub.sar ‘scribe’ ‘conventional penalty’ Hurrian uriḫull-

An especially interesting case represents the following lemma, in which the borrowing process takes a return path: (2) Re-borrowing Akkadian – Sumerian – Akkadian Akkadian šakkanakku

Twisted path ‘military governor’ Akkadian > Sumerian > Akkadian

2.2 Hebrew 2.2.1 Common Northwest-Semitic Stock but also Canaanite-Aramaic Isoglosses While there clearly exists a shared Northwest-Semitic lexical stock, one also finds Canaanite-Aramaic isoglosses, such as the following (see McCarter 2008: 78): (3) Canaanite-Aramaic isoglosses Canaanite √ ʿ–l–y √ b-w-ʾ √ d–b–r

vs. vs. vs.

Aramaic √ s–l–q √ ʿ–l–l √ m–l–l

‘to ascend’ ‘to enter’ ‘to speak’

2.2.2 Loanwords in Biblical Hebrew Biblical Hebrew also features a certain deal of loanwords, notably from Akkadian (or ultimately Sumerian) (see Mankowski 2000: 167–170) or from Egyptian (see McCarter 2008: 79):

inner-semitic loans and lexical doublets

183

(4) Akkadian and Egyptian loanwords in Biblical Hebrew Biblical Hebrew Source word in the language of origin hēḵå�l ‘temple, palace’ Akkadian ekallu (Sum. é.gal ‘big house’) ṭap̄ så�r ‘military officer’ Akkadian ṭupšarru ‘scribe’ (Sum. dub.sar ‘scribe’) ʾiggɛ́rɛṯ ‘letter’ Akkadian egirtu miskēn ‘pauper’ Akkadian muškēnu šēš ‘linen’ Egyptian šś ṭabbáʿaṯ ‘sealing ring’ Egyptian ḏbʿwt 2.2.3 Loanwords in Medieval Hebrew Notable additions to the Hebrew lexicon in the post-Classical (Rabbinical) and medieval periods originated in Greek and Aramaic. The following examples are notable: (5) Greek and Aramaic loanwords in Medieval Hebrew from Greek (several via Aramaic): ʾawīr ‘air’ ʾambaṭiya ‘bathroom’ ʾaḵsanya ‘inn’ ʾasimon ‘telephone token’ basis ‘base’ zug ‘pair’ parṣup̄ ‘face’ pinqas ‘notebook’ tīq ‘file’

from Aramaic: le-sayyeaʿ ‘to help’ le-targem ‘to translate’ girsa ‘version’ darga ‘rank, step’ ʿuḇda ‘fact’ zabban ‘salesman’ sappar ‘hairdresser’ maškanta ‘mortgage’ harpatqa ‘adventure’

2.2.4 Loanwords in Modern Hebrew Modern Hebrew draws from a wide range of languages, both Semitic and nonSemitic. As is well known, Eliezer Ben-Yehuda’s strategy always was to give priority to other Semitic languages like Arabic over European languages when need arose to create new vocabulary. Here are three examples:

184

edzard

(6) Arabic loanwords in Modern Hebrew Arabic Hebrew ta‌ʾrīḫ tarix ‘date (calender)’ ʾadīb adiv ‘polite’ muhāǧir mehager ‘emigrant/immigrant’ Often, formal Hebrew neologisms, proposed by the Hebrew Language Academy, have not really caught on in the colloquial language, as opposed to European-based loanwords, e.g., formal šmarṭaf (‘who watches baby’) vs. colloquial bebisiṭer ‘baby sitter’, or formal (Academy-proposed) paʿaman (‘time(s) instrument’) vs. colloquial metronom ‘metronome’. Hebrew also draws on older Semitic languages. An interesting example is the neologism balšan ‘linguist’. According to Kutscher (1982: 234), the historical source of this term is found in Ezra 2:2 mårdəḵay bilšān (< Marduk bēl-šunu ‘Marduk is their Lord’), a term which engendered the following Midrashic exegesis: bilšan ‘he knew [seventy] languages’, from which the nouns balšan ‘linguist’ and balšanut ‘linguistics’ were derived. 2.3 Borrowing in Arabic 2.3.1 Classical Arabic Scientific Terminology An attempt at a typology of loan formations as relating to scientific Classical Arabic was made in Edzard 1998: 36f., based on Endreß 1992: 14–23. Interestingly, loans in this semantic category come about either directly from a non-Semitic source (typically from Greek or Persian) or via other Semitic (typically Syriac Aramaic) intermediary stages. Here are few relevant examples: (7) Loanwords in scholarly Arabic (7a) Direct loans from Greek via Syriac Greek Syriac Arabic γένος gensā jins ‘kind’ στοιχεῖον ʾesṭuḵsā ʾusṭuquss ‘element’ (7b) Direct loans from Syriac Syriac Arabic kyānā kiyān ‘element’ ʿellṯā ʿilla ‘cause’ (‘defect’)

(see Greek φύσις) (see Greek α̉ίτιον)

185

inner-semitic loans and lexical doublets

(7c) Direct loans from Persian Persian Arabic goh(ar) jawhar ‘substance’ (‘jewel’) (see Greek οὐσία) purmahi burmāhi ‘full moon’ (7d) Direct loans from Greek in quasi-transcription Greek Arabic κατηγορίαι qāṭāġūriyās σοφιστική sūfistiqā

‘categories’ ‘sophistics’

(7e) Loan translations (“calques”) Greek Arabic λογική manṭiq α̉rητή faḍīla

‘logic’ ‘high quality’ (‘best’)

(7f) Secondary creation after original borrowing Greek Arabic-1 στοιχεῖον ʾusṭuquss ‘element’ α̉ίτιον ʿilla ‘cause’ (‘defect’)

Arabic-2 ʿunṣur ‘element’ sabab ‘cause’

(7g) Abstract neologisms Greek Arabic τὸ πρός τι ʾiḍāfa τὸ τί ε̉στι māhīya

‘relation’ ‘essence’

2.3.2 Aramaic Loanwords in Arabic The Aramaic stock in or the Aramaic substrate of Arabic has, ever since Fraenkel 1886 and Jeffery 1938, been recognized in its importance. In a recent acute analysis, Retsö (2006) highlights the following complications that complicate a straightforward phonological derivation of the Arabic terms. (a) the rather complicated rules for the postvocalic shift of /b, g, d, k, p, t/ (spirantization) in Aramaic; (b) the reduction of short vowels in unstressed open syllables to šwā or zero;

186

edzard

(c) the shift ā > o (ō) in northern-central Syria and Mesopotamia (but not in other Aramaic-speaking areas); (d) the circumstance that common Semitic p always appears as f in Arabic; (e) the rather complicated distribution of sibilants; (f) the fact that both Aramaic and Arabic feature t / ṯ and d / ḏ, but in different distribution (originally allophonic in Aramaic, at times also phonemic; regularly phonemic in Arabic). Commonly accepted examples of Aramaic loans in Arabic include the following lexical items: (8) Aramaic loanwords in Arabic Aramaic Arabic ašlem ʾaslama ‘to submit [to the new religion]’ bāḇā bāb ‘gate’ biʿṯā bīʿa ‘church’ šḇīlā sabīl ‘path’ tḇar ‘break’ tatbīr ‘destruction’ ʿālmā ʿālam ‘world’ ṣlōṯā ṣalāt ‘prayer’ purqānā furqān ‘redemption’ The status of tatbīr ‘destruction’ as an inner-Semitic loan is evidenced by the first root consonant t, instead of a lautgesetzlich expected ṯ (see Hebrew šāḇar ‘to break’). Retsö (2006) identifies, among others, the following semantic doublets in Arabic that emerged through borrowing from Aramaic parallel to the original Arabic terms (no phonological differentiation in these examples, though): (9) Semantic doublets in Arabic Aramaic Arabic kataba ‘to write’ (from Aramaic/ vs. ‘to put/sew together’ Hebrew) (Arabic) darasa ‘study’ (from Aramaic draš) vs. ‘to wipe out’ (Arabic) dīn ‘judgment’ (from Aramaic vs. ‘debt’ (Arabic) dīnā)

inner-semitic loans and lexical doublets

187

2.3.3 Ethiopic Loanwords in Arabic Inner-Semitic loans are also attested from Ethio-Semitic to Arabic (see Weninger 2007). The following are plausible examples of loans from Gəʿəz: (10) Ethio-Semitic loanwords in Arabic Ethiopic Arabic waqf  ‘bracelet’ waqf  ‘bracelet’ kabaro ‘drum, timbrel’ kabar ‘kettle-)drum’ maṣḥaf  ‘(any kind of) book’ muṣḥaf  ‘book (especially copy of the Qurʾān)’ manbar ‘chair, throne’ minbar ‘pulpit’ (no root √ n–b–r ‘to sit’ in Arabic) 2.3.4 European Loanwords in Arabic European loans in modern Arabic, in both the standard language and the dialects, are well documented (see, e.g., Weninger 2001; Edzard 2012a). Words can either be fully morphologically assimilated, e.g., as in raskala ‘recycling’, or be borrowed in an unassimilated forms (save for means of linear derivation, typically nisba endings), e.g, as in bībliōgrāfiyā ‘bibliography’ or tiknōqrāṭ ‘technocrat’ (see Badawi et al. 2004: 742). Native forms preferred by language academies may coexist with borrowed forms, a notable example pair being hātif ‘invisible voice’ and tilifōn ‘telephone’. Regional factors play a role as well; thus, one finds French-based ʾotél ‘hotel’ and Turkish-based lokanda besides the native Arabic terms for ‘hotel’, funduq and nazl (see Badawi et al. 2004: 743). (For the reverse scenario, i.e., Arabic loanwords in Germanic, see Vennemann 2003, 2011, 2012.) 2.4 Borrowing in Ethio-Semitic 2.4.1 Cushitic Substratum There are ubiquitous traces of Cushitic substratum in Ethio-Semitic (see Appleyard 1977; 1978). Appleyard (1977) identifies Cushitic loans in the following semantic fields (see also Edzard 2012b: 47): (11) Cushitic loanwords in Amharic (a) “man”, comprising general terms, kinship terms, and parts of the body, e.g., Amharic aggot ‘uncle’, cf. Bilin (Central Cushitic) ʾäg; (b) “the domestic environment”, comprising agricultural activities and implements, crops, domestic animals, food and its preparation, and the

188

edzard

[realm of the] house, e.g., Amharic doro ‘chicken’, cf. Saho-Afar (EastCushitic) dorho; (c) “the natural environment”, comprising natural phenomena, flora, and fauna, e.g., däga ‘highlands’, cf. Bilin (Central Cushitic) dag ‘summit, above’; (d) “social organization”, comprising law and government, economy, warfare, and religion, e.g., aṭe ‘emperor’, cf. Kemant (Central Cushitic) ašena; (e) “grammatical items”, comprising pronouns, numerals, and particles, e.g., ši(h) ‘thousand’, cf. Bilin (Central Cushitic) šix. 2.4.2 European Loanwords in Ethio-Semitic European, notably Italian loans also play an important role in Ethio-Semitic (see, e.g., Abraham Demoz 1963, Edzard 2003). The following items are extrapolated from Leslau 1976; all transcriptions follow the Ethiopic syllabary: (12) Italian loanwords in Amharic (a) persons, professions: ʾəmamma ‘mom’, mammo ‘male baby’, gutto ‘little stout person’ (Ital. gatto ‘tomcat’), listro ‘shoe shine boy’ (Ital. lustro ‘shoe crème’); (b) vehicles, technique, building material: fabrika ‘factory’, mäkina ‘car’, gomma ‘rubber, tire’, bukko ‘hole in tire’ (Ital. buco), bonda ‘iron fixation’, targa ‘name tag’ (Ital. targa ‘tag’), tubbo ‘[lead] pipe’ (Ital. tubo), ʾantena ‘antenna’, siminto ‘concrete’ (Ital. cemento); (c) appliances, furniture, tools, instruments, further items: banko ‘bar table’, bañño ‘bathtub’, ṭrumba ‘trumpet’, pakko ‘packet [of cigarettes]’, pippa ‘pipe’, kandella ‘cigarette lighter’, samuna ‘soap’ (Ital. sapone), karta ‘map’, gazeṭa ‘newspaper’. Comparable loanwords are also attested in the Cushitic languages Somali (see, e.g., Agostini et al. 1985) and Oromo (see Gragg 1982). 2.4.3 Diachronic Borrowing from Gəʿəz (loss of gutturals) in Amharic Phonological, but not necessarily semantic doublets have emerged in modern Amharic through borrowing from an earlier stratum of the same branch, Gəʿəz. While the older gutturals have largely vanished in the modern verbal paradigms, nominal patterns continue to retain these gutturals. Here are a few examples:

189

inner-semitic loans and lexical doublets

(13) Phonological doublets in modern Amharic guttural retained təmhərt mäṣḥaf däḫna səʿəl

ትምህርት መጽሐፍ ደኅና ሥዕል

3

guttural lost ‘lesson’ vs. ተማረ tämari ‘book’ vs. ጻፈ ṣafä ‘well’ vs. ዳነ danä ‘picture’ vs. ሣለ salä

‘student’ ‘he wrote’ ‘he recovered’ ‘he painted’

Inner-Semitic Loans / Lexical Doublets

3.1 Typology of Minimal Pairs Of special interest is the emergence of lexical doublets in Semitic that constitute phonological and semantic minimal pairs. In this context, the differentiating factor can be phonological stress, as in the following modern Hebrew example (see Kubcher 1982: 229f.): (14) Differentiation by stress in Modern Hebrew mašqé ‘something to drink’ vs. mášqe ‘alcoholic drink’ Here, the word with penultimate stress was re-borrowed from Yiddish into modern Hebrew. Stress can, of course, simply be used to distinguish semantic shades, e.g., German ′übersetzen ‘to cross a river’ (preterite er setzte über ‘he crossed a river’) vs. übersétzen ‘to translate’ (preterite er übersetzte ‘he translated’) (see also Vennemann 1986: 30f.). Of special interest is the phonematization of the /b, g, d, k, p, t/ early on in the history of Aramaic (see Edzard 2001). The differentiation may affect the word semantics, the category of a word, or even the person, as in the following examples: (15) Differentiation by spirantization of /b, g, d, k, p, t/: re-phonematization in Syriac Aramaic semantics: ḥaḏūṯā ‘joy’ vs. ḥadūṯā ‘cave for wheat and barley’ category: garḇā ‘scabies’ vs. garbā ‘scabious’ person: ḥḏīṯ ‘I rejoiced’ vs. ḥḏīt ‘you (m.s.) rejoiced’

190

edzard

Vowel quality can also serve to support semantic fine-tuning. In the following example, the different pronunciation of the qamaṣ in the first closed syllable distinguishes the word ‘patience’ from the word ‘tolerance’: (16) Differentiation by vowel quality in Modern Hebrew savlanut ‘patience’ vs. sovlanut ‘tolerance’ In other cases, parallel forms may coexist without any overt semantic differentiation involved, as in Hebrew verbal nouns of the hifʿīl binyan, some of which also, or even predominantly, are attested in a parallel Aramaic pattern. The weakening of word-initial h in non-normative modern Hebrew blurs the picture even more, as the Aramaic form sounds exactly like the non-normative pronunciation of the Hebrew form. Here are two examples found in the modern Hebrew dictionaries: (17) Parallel Hebrew and Aramaic forms in Modern Hebrew haflaya/ʾaflaya ‘discrimination’ havṭala/ʾavṭala ‘unemployment, laying-off’ Next to straightforward lautgesetzlich correspondences one also finds freer variation in the composition of roots, as in the Semitic etymon for ‘to laugh’: (18) Multiple roots: the etymon ‘to laugh’ in Semitic √ ś–ḥ–q (Hebr.) √ ś–ḥ–q (Eth.) √ ḍ–ḥ–k (Arab.) ‘to laugh’ √ ṣ–ḥ–q (Ugar.) √ ṣ–ḥ–q (Hebr.) ‘to laugh’ √ g–ḥ–k (Syr.) √ d–ḥ–k (Pal.) ‘to laugh’ Modern Hebrew draws upon several of these roots. While the positively connotated participle/adjective maṣḥīq ‘amusing’ is derived from the root √ ṣ–ḥ–q, the negatively connotated participle/adjective məguḥaḵ ‘ridiculous’ is derived from the root √ g–ḥ–k. In turn, the noun daḥka ‘prank’ appears to be derived from the Arabic root √ ḍ–ḥ–k, even though √ d–ḥ–k is already attested in Palestinian Aramaic. Interesting minimal pairs can arise in the course of secondary formations, in which original affixes are integrated into the root (see Ussishkin 1999; Edzard 2011). In Arabic, one finds the following minimal pair:

inner-semitic loans and lexical doublets

191

(19) Minimal pairs as a result of root integration ʾaslama, yuslimu ‘to be come a Muslim’ vs. ʾaslama, yuʾaslimu ‘to islamize’ While the former represents the regular form IV in the Arabic binyan system, the latter represents a quadriliteral form in which the prosthetic ʾalif is fully integrated into the root, comparable to the ʾalif  in ʾamraka, yuʾamriku ‘to americanize’; thus one can also easily derive the reflexive ta‌ʾaslama, yata‌ʾaslamu ‘to be islamized’, just as ta‌ʾamraka, yata‌ʾamraku ‘to be americanized’ is derived. Lexical doublets can also be the result of multiple borrowing at different stages in language history. The following in an instructive example, loaned in English twice from French (ultimately from Latin caput ‘head’), at different times and in different meanings (see Kutscher 1982: 49): (20) Different forms as the result from borrowing at different historical stages chief  ‘head [of an organization]’ vs. chef  ‘cook’ In modern German, comparable examples exist as well. Service, for instance, refers to ‘service’ in an English pronunciation, but to a fancy set of porcelain in a French pronunciation of the same word. Needless to say, the meaning of a word can take interesting spins in a borrowing process. The English adjective handy, for instance, has adopted the meaning of ‘cellular phone’ in contemporary German. Finally, lexical doublets can emerge in the process of re-borrowing (a linguistic “pizza effect”). In modern Hebrew, one finds an opposition between the regular participle of the root √ m–s–r and the same form re-borrowed from Yiddish, in a specialized negatively tainted meaning (see Edzard 2014). Here, the vowel quality in the re-borrowed form is crucial, as opposed to accent above in (14): (21) Different forms as the result from re-borrowing mōser ‘transmittor’ vs. moiser ‘traitor within the Jewish community’ (via Yiddish) East Semitic vs. West Semitic: Problems with the Distribution of Gutturals Two small case studies shall conclude this typological overview. The first concerns Akkadian. Unexpected parallel forms in Akkadian, which involve the uneven or unexpected distribution of gutturals, have attracted the attention of 3.2

192

edzard

Tropper (1995), Huehnergard (2003), Kogan (2012: 109–114), and Edzard (2015), among others. The following set of parallel forms (alphabetically ordered) represents crucial examples in this context: (22) Lexical doublets/variants in Akkadian (Edzard to appear) ḫabālu ‘bind’ vs. ebēlu ‘to catch with a net’ ḫadāru vs. adāru ‘to be dark, sinister’ ḫazzatu(m) vs. azzatu(m) ‘goat’ ḫullu(m) vs. ullu ‘neckring’ Many of the variants with ḫ appear to be Neo-Assyrian or jungbabylonisch, others are of Mari provenance. Michael Streck (personal communication) kindly informs me that these examples can be categorized as follows: (a) one etymon is loaned from West-Semitic, whereas the other one is genuinely East Semitic Akkadian, e.g., ḫabālu and ebēlu; (b) both etyma are loaned, and the alternative ḫ ≈ ʾ is just of an orthographical nature, e.g., ḫazzatu(m) and azzatu(m), both of which represent loaned ʿazzatu(m); the genuine Akkadian counterpart is enzu; (c) both etyma are genuinely Akkadian, and Ḫ simply renders /ʾ/, e.g., ḫadāru and adāru; (d) both etyma are genuinely Akkadian, whereby Ḫ represents a guttural phoneme still extant in older Akkadian (e.g., /ġ/), whereas the younger orthography represents the disappearance of the guttural, e.g., ḫullu(m) and ullu (see also Kogan 2012: 112). Such doublets are not only found in the realm of gutturals. Turning to sibilants, one finds the well-known but still unexplained pair salāmu(m) ‘to be in peace with each other’ and šalāmu(m) ‘to be in good shape’; the latter variant represents the lautgesetzlich “expected” form in East-Semitic. 3.3 Phonological Doublets in Ṭūrōyo The second case study concerns the Central neo-Aramaic language Ṭūrōyo and refers to the research of Sina Tezel (2011). Again, the point of departure are phonological doublets that arose in the course of inner-Semitic borrowing (for questions of language contact with modern Aramaic, see also Kapeliuk 2002). The Arabic verb θāba ‘to return’ is paralleled by the etymologically identical verb tāba ‘to repent’, which constitutes a loan from Aramaic (see Hebrew šāḇ ‘he returned’). Similarly, one finds the Aramaic causative verb šēṣī ‘to accomplish’, which in its phonological structure is to be explained as a loan from

inner-semitic loans and lexical doublets

193

Akkadian, in contrast to the Aramaic basic form iʿā ‘to grow’ (see Brockelmann 1908: 119). In the following schematic representation, the relevant root consonants are marked in bold: (23) Arabic and Aramaic doublets as a result of inner-Semitic borrowing Arabic θāba ‘to return’ Aramaic tāḇ ‘he repented’ > Arabic tāba ‘¨to repent’ Aramaic iʿā ‘to grew’ Akkadian ušēṣī ‘he sent out’ > Aramaic šēṣī ‘to accomplish’ In Ṭūrōyo, the focus of Tezel 2011, comparable doublets have emerged as a result of borrowing from Arabic. In the following examples, the first line represents the expected form in Aramaic, whereas the second line represents the parallel form borrowed from Arabic (in the last example, an additional form loaned via Kurdish is present as well). Again, the consonants relevant for the issue at hand are marked in bold: (24) Ṭūrōyo doublets as a result of borrowing from Arabic Ṭūrōyo √dyq, e.g., mādəqle ‘to look out’ Arabic √ðyq > Ṭūrōyo √ðyq, e.g., ðəqle ‘to taste’ Ṭūrōyo √nṭr, e.g., nṭərle ‘to wait for’ Arabic √nð̣r > Ṭūrōyo √nð̣ r, e.g., nað̣ra ‘look, supervision’ Ṭūrōyo √tny, e.g., lō kō mtāne ‘he doesn’t oppose’ Arabic √θny > Ṭūrōyo √θny, e.g., θnēle ‘to plough twice’ Ṭūrōyo √shð/ḏ (< √shd), e.g., shəð/ḏle ‘to witness’ Arabic √šhd > Ṭūrōyo √šhd, e.g., mšāhadle ‘to acknowledge Islam’ Ṭūrōyo √byʿ, e.g., bēʿe ‘eggs’ Arabic √byð̣ /ẓ > Ṭūrōyo √byð̣ /ẓ, e.g., mbāyað̣ /ẓle ‘to tin’ Ṭūrōyo √sgd, e.g., sɣ/ḡəð/ḏle ‘to adore, worship’ Arabic √sǧ/jd- > Ṭūrōyo √sǧ/jd, e.g. səǧ/jǧ/jāde ‘carpet, rug’; but > Ṭūrōyo məzgafte (via Kurdish mizgeft) The complete inventory of doublets in Tezel’s (2011) corpus, which draws both on written sources (notably Ritter 1967–1990) and her own native intuition, involve the following consonants and sound correspondences (the consonants in parentheses represent /bgdkpt/ allophones that should not be confused with the corresponding Arabic fricatives and therefore do not undermine the picture):

194

edzard

(25) Sound correspondences between Common Semitic, Arabic, and Aramaic Common Semitic Arabic g ≈ ǧ / j ð ≈ ð x / ḫ ≈ x / ḫ θ̣ ≈ ð̣ / ẓ ś ≈ š ɣ / ġ ≈ ɣ / ġ ð̣̣ / ṣ́ ≈ ḍ š ≈ s θ ≈ θ

Syriac/Ṭūrōyo ≈ g (ɣ / ḡ) ≈ d (ð / ḏ) ≈ ḥ ≈ ṭ ≈ s ≈ ʿ ≈ ʿ ≈ š ≈ t (θ / ṯ)

4 Summary In sum, the described inner-Semitic loans, even though statistically not significant within the overall vocabulary of the respective languages, work to blur our traditional understanding of lautgesetzlich sound correspondences. With respect to the East Semitic vs. West Semitic scenario in Akkadian gutturals, Tropper (1995: 64) allows for borrowing in individual cases, while Militarev and Kogan (2000: lxxiv–lxxv) point to both borrowing and different geographical distribution. Huehnergard (2003: 111) argues for (inner-Semitic) inter-language loan as well as semantic contamination. Edzard (2014a, to appear), in view of both the Akkadian and the Ṭūrōyo scenario, highlights the potential for phonemic splitting in the course of inner-Semitic borrowing, in line with thoughts about linguistic development, expressed in Trubetzkoy 1939, Pisani 1975, or Edzard 1998. While it may seem to be relatively easy to identify inner-Semitic loanwords in a contemporary Semitic language like Ṭūrōyo, this is not necessarily the case in older stages of Semitic, as could be observed with the WestSemitic loans in Akkadian and the Aramaic loans in Arabic: the greater the temporal distance the more inner-Semitic loans appear to be integrated in the phonological system of the language in question, and the harder it is to establish criteria for the identification of inner-Semitic loans.

inner-semitic loans and lexical doublets

195

Bibliography Abraham Demoz. 1963. “European loanwords in an Amharic daily newspaper,” in John Spencer (ed.), Language in Africa. London. 116–122. Agostini, Francesco, Annarita Puglielli, and Ciise Moxamed Siyaad. 1985. Dizionario somalo-italiano. Rome. Appleyard, David L. 1977. “A comparative approach to the Amharic lexicon,” Afroasiatic Linguistics 5: 1–67. ———. 1978. “Linguistic evidence of non-Semitic influence in the history of Ethiopian Semitic,” Abbay 9: 49–56. Badawi, Elsaid, Michael G. Carter, and Adrian Gully. 2004. Modern Written Arabic. A Comprehensive Grammar. London – New York. Brockelmann, Carl. 1908–1913. Grundriss der vergleichenden Grammatik der semitischen Sprachen (2 vol.). Berlin. Edzard, Lutz. 1998. Polygenesis, Convergence, and Entropy. An Alternative Model of Linguistic Evolution Applied to Semitic Linguistics. Wiesbaden. ———. 2001. “Problems with post-vocalic spirantization in Syriac: Cyclic rule ordering vs. ‘early phonemization with paradigmatic leveling’,” JSS 46: 77–95. ———. 2003. “Externe Sprachgeschichte des Italienischen in Libyen und Ostafrika,” in G. Ernst (ed.), Romanische Sprachgeschichte, Vol. 1. Berlin. 966–972. ———. 2011. “Die SIG7.ALAN = Nabnītu-Liste und das Konzept der semitischen Wurzel,” ZDMG 161: 17–37. ———. 2012a. “Language contact between Arabic and modern European languages,” in Stefan Weninger (ed.), Semitic Languages. An International Handbook. Berlin. 1022–1032. ———. 2012b. “Introduction: Semitic and Afroasiatic,” in Lutz Edzard (ed.), Semitic and Afroasiatic. Challenges and Opportunities. Wiesbaden. 23–58. ———. 2014. “Hebrew and Hebrew–Yiddish terms and expressions in contemporary German: Some (socio-)linguistic observations,” in Lutz Edzard and John Huehnergard (eds.), Proceedings of the Oslo–Austin Workshop in Semitic Linguistics, Oslo, May 23 and 24, 2013. Wiesbaden. 127–143. ———. 2015. “Zum Verhältnis von Lehnbildung und genetischer Verwandtschaft im Semitischen,” in Viktor Golinets, Hanna Jenni, Hans-Peter Mathys, and Samuel Sarasin (eds.), Neue Beiträge zur Semitistik. Fünftes Treffen der Arbeitsgemeinschaft Semitistik in der Deutschen Morgenländischen Gesellschaft vom 15.–17. Februar 2012 an der Universität Basel. Münster. 19–33. Endreß, Gerhard. 1992. “Die wissenschaftliche Literatur,” in Wolfdietrich Fischer (ed.), Grundriß der arabischen Philologie, vol. 3. Wiesbaden. 3–152. Fraenkel. Sigmund. 1886. Die aramäischen Fremdwörter im Arabischen. Leiden. Gragg, Gene. 1982. Oromo Dictionary. East Lansing.

196

edzard

Huehnergard, John. 2003. “Akkadian ḫ and West Semitic *ḥ,” in Leonid Kogan (ed.), Studia Semitica. Moscow. 102–119. Jastrow, Otto. 1997. “The neo-Aramaic languages,” in Robert Hetzron (ed.), The Semitic Languages. London. 334–377. Jeffery, Arthur. 1938. The Foreign Vocabulary in the Qurʾān. Baroda. Kapeliuk, Olga. 2002. “Languages in contact: The contemporary world,” in Shlomo Izre’el (ed.), Semitic Linguistics: The State of the Art at the Turn of the 21st Century (IOS 20). Winona Lake. 307–340. Kogan, Leonid. 2012. “Proto-Semitic Phonetics and Phonology,” in Stefan Weninger et al. (eds.). The Semitic Languages. An International Handbook. Berlin. 54–151. Kutscher, Eduard Yechezkel. 1982. A History of the Hebrew Language. Jerusalem. Leslau, Wolf. 1958. Ethiopic and South Arabic Contributions to the Hebrew Lexicon. Wiesbaden. ———. 1945. “The influence of Cushitic on the Semitic languages of Ethiopia. A problem of substratum,” Word 1: 59–82. ———. 1952. “The influence of Sidamo on the Ethiopic languages of Gurage,” Language 28: 63–81. ———. 1969. Hebrew Cognates in Amharic. Wiesbaden. ———. 1976. Concise Amharic Dictionary. Wiesbaden. Lieberman, Stephen J. 1977. The Sumerian Loanwords in Old Babylonian Akkadian, Vol. 1. Prolegomena and Evidence (HSS 22). Missoula. Mankowski, Paul V. 2000. Akkadian Loanwords in Biblical Hebrew (HSS 47). Winona Lake. McCarter, P. Kyle. 2008. “Hebrew,” in Roger D. Woodard (ed.), The Ancient Languages of Syria-Palestine and Arabia. Cambridge. 36–81. McCarthy, John J. 1991. “Semitic gutturals and distinctive feature theory,” in Bernard Comrie and Mushira Eid (eds.), Perspectives on Arabic Linguistics III. Papers from the Annual Symposium on Arabic Linguistics. Amsterdam. 63–92. Militarev, Alexander and Leonid Kogan. 2000. Semitic Etymological Dictionary, Vol. 1. Anatomy of man and animals. Münster. Pisani, Vittorio. 1975. Introduzione alla linguistica indoeuropea. Torino. Puglielli, Annarita and Biancamaria Tedeschini Lalli. 1981. Ricerce sull’insegnamento dell’ italiano in Somalia. Rome – Mogadishu. Retsö, Jan. 2006. “Aramaic/Syriac loanwords,” in Kees Versteegh (ed.), Encyclopedia of Arabic Language and Linguistics, Vol. 1. A–Ed. Leiden. 178–182. Ritter, Hellmut. 1967–1990. Ṭūrōyo: Die Volkssprache der syrischen Christen des Ṭūr ʿAbdīn (5 vols). Beirut – Stuttgart. Tezel, Sina. 2011. Arabic Borrowings in Ṣūrayt/Ṭūrōyo within the Framework of Phonological Correspondences: in comparison with other Semitic languages. Uppsala.

inner-semitic loans and lexical doublets

197

Tropper, Josef. 1995. “Akkadisch nuḫḫutu und die Repräsentation des Phonems /ḥ/ im Akkadischen,” ZA 85: 58–66. Trubetzkoy, Nikolaus. 1939. “Gedanken über das Indogermanenproblem,” Acta Linguistica 1: 81–89. Ussishkin, Adam. 1999. “The inadequacy of the consonantal root: Modern Hebrew denominal verbs and output-output correspondence,” Phonology 16: 401–442. Vennemann, Theo. 1986. Neuere Entwicklungen in der Phonologie. Berlin. ——— (ed. Patrizia Noel Aziz Hanna). 2003. Europa Vasconica – Europa Semitica. Berlin – New York. ———. 2011 “Arabic loanwords in German(ic),” in Lutz Edzard and Rudolf de Jong (eds.). Encyclopedia of Arabic Language and Linguistics Online. Leiden. ——— (ed. Patrizia Noel Aziz Hanna). 2012. Germania Semitica. Berlin – New York. Wagner, Esther-Miriam. 2013. “Germanic languages, Hebrew loanwords in,” in Geoffrey Khan (ed.), Encyclopedia of Hebrew Language and Linguistics, Vol. 2. G–O. Leiden. 57–60. Watson, Wilfred G. E. 2005. “Loanwords in Semitic,” AO 23: 191–198. Weninger, Stefan. 2001. “Zur Wiedergabe englischen Sprachmaterials im modernen Hocharabisch,” in Dörte Borchers, Frank Kammerzell, and Stefan Weninger (eds.), Hieroglyphen, Alphabete, Schriftreformen: Studien zu Multiliteralismus, Schriftwechsel und Orthographieneuregelungen. Göttingen. 175–191. ———. 2007. “Ethiopic loanwords,” in Kees Versteegh (ed.), Encyclopedia of Arabic Language and Linguistics, Vol. 2. Eg-Lan. Leiden. 56–57.

Structural Change in Urban Palestinian Arabic Induced by Contact with Modern Hebrew Uri Horesh 1 General In her overview of over a century of contact between Arabic and Hebrew in Palestine, Henkin-Roitfarb “stress[es] that the contact is asymmetric” (2011: 61). This qualification has to do mostly with the imbalance of dual language proficiency across the Jewish-Israeli community, where Hebrew is used as the primary language of communication, and the Palestinian community, whose native language is Palestinian Arabic. While the former includes native speakers of various dialects of Arabic, though typically not Palestinian Arabic (more common are North African and Mesopotamian dialects), as well as a small percentage of non-native speakers who have mastered Arabic as a second language, they do not compare to the overwhelming numbers of Palestinians who have studied within the Israeli educational system and acquired Hebrew to a high level of proficiency. The goal of the current paper is to focus on a number of structural changes in a contemporary dialect of Palestinian Arabic and contextualize them as contactinduced changes. Because Arabic and Hebrew share—or had shared—certain features throughout history, and because some of the changes described may not be unique to Arabic varieties that are in contact with Hebrew, there will be an almost forensic nature to some of the argumentation herein. Using a combination of quantitative methods, general principles of diachronic linguistics and language contact and—first and foremost—basing the analysis on empirical data recorded in Palestine between 1999 and 2005, I hope to establish that much of the deviation exhibited in the dialect at hand from neighboring dialects that are not in contact with Hebrew is indeed a result of this very contact. Thomason and Kaufman (1988: 67) argue that “long-term contact with widespread bilingualism among borrowing-language speakers is a prerequisite for extensive structural borrowing.” Nagy (1996) provides us with a body of work which has taken this notion and incorporated it within the theoretical thinking and methodology of variationist sociolinguistics. Nagy (1996: 41) lists three groups of intensity of contact factors: amount of contact, cultural identity and linguistic factors.

© koninklijke brill nv, leiden, ���5 | doi ��.��63/9789004300156_012

structural change in urban palestinian arabic

199

Palestinian citizens of Israel attend, for the most part, public schools whose main language of instruction is Arabic. Officially, the variety of Arabic used in the school system is Modern Standard Arabic. Pupils study Hebrew starting in the third grade of elementary school (Amara 2001: 160).1 However, scholars in educational linguistics raise serious doubt regarding the degree of proficiency these students typically achieve by the end of high school. In Jaffa, where the main bulk of data for the current study were collected, this is further complicated by the fact that many families prefer to send their children to schools in which Hebrew, not Arabic, is the main language of instruction. Other children attend private parochial schools, where other languages (e.g., French) are also added into the mix. According to the most recent report by the Israeli Central Bureau of Statistics, “Arabs” (i.e., Palestinians) constitute 20.69% of Israel’s population of 8,114,000.2 Talmon (2000) reports that while most (ca. 65%) of the speakers of Palestinian Arabic within Israel are concentrated in the Galilee and in Haifa, i.e., in the northern part of the country, some 100,000 live in the southern Nagab region, and over 200,000 live in the central region, within the greater metropolitan area whose core is in Tel Aviv-Jaffa. The Palestinians living in the muθallaθ ‘Triangle’ area north and northeast of Tel Aviv-Jaffa are not quite a part of the cluster of suburbs and semi-industrial towns of the metropolis. Those living in the mixed (i.e., Jewish-Arab) towns of Lydd and Ramle are closer to that status. Those living in Jaffa (Arabic jaːfa; Hebrew ˈjafo, in more formal registers jaˈfo),3 formerly an autonomous municipal entity and since shortly after the formation of the State of Israel part of the city of Tel Aviv–Jaffa, are in many ways full participants in the urban experience, culturally and financially. There appears to be some controversy surrounding the number of Palestinians currently living in Jaffa. According to the demographic section of the 2012 statistical bulletin for the City of Tel Aviv—Jaffa,4 out of 404,800 people living in the city as a whole (Tel Aviv and Jaffa combined), at the end of 2011, 388,100 1  I have been informed (inter alii by Elana Shohamy of Tel Aviv University School of Education, p.c., Feb. 2007) that this is changing and that due to pressure from parents and other community factors, Hebrew is now entering the Arab schools as early as the second and first grades in many instances. 2  This adds up to 1,678,400 people. This is based on the Central Bureau of Statistics’ Monthly Bulletin of Statistics—December 2013, published 7 January 2014: http://www.cbs.gov.il/ publications14/yarhon1213/pdf/b1.pdf (accessed 3 February 2014). 3  Unless otherwise noted, transcriptions in this paper are based on IPA notation. 4  http://www.tel-aviv.gov.il/TheCity/Documents/‫שנתון‬%202012/‫מבנה‬%20‫דמוגרפי‬/‫הסבר‬%20 ‫מילולי‬.pdf (accessed 3 February 2014).

200

horesh

(95.87%) were “Jewish and other non-Arabs” and the remaining 16,700 (4.13%) were “Arabs” (i.e., Palestinians). However, only about 14,000 of these (83.83%) lived in Jaffa. The numbers reported by the League of the Arabs of Jaffa, a local group that describes its goal as “to preserve the Arab presence in Jaffa and to protect the rights of the Palestinians in Jaffa as an Arab Palestinian Minority [sic],”5 are higher. The League devotes part of its web site6 to “Historical Background”— about Jaffa and its Palestinian identity, not about the League itself—and opens with the following: The number of Arabs in Jaffa today reaches about 23,000 inhabitants. Before Jaffa fell in Israeli hands in 1948, the Arab inhabitants of Jaffa counted more than 120,000. Most of them were forced to leave their city. Only 3900 Arabs were able to stay in Jaffa. Today Jaffa is one of the six so called Palestinian mixed cities targeted by the Israeli authorities. Jewish new comers live in the homes of the Palestinians who were forced to leave in 1948. All six cities are targeted by the Israeli authorities to turn them into marginal insignificant minorities. It is worth noting that this web page has a “© 2007” notation at its bottom, implying that the information in it may not have been updated in seven years. This actually brings the population figures more in line with the official ones from the municipal authorities (which, in turn, are drawn from the Central Bureau of Statistics). The municipal report (written in Hebrew), indicates that the non-Jewish population of Tel Aviv—Jaffa had peaked in 2003 at 28,200, having doubled from what it had been in 1991. But the current (i.e., 2011) statistic represents a fall, following a decrease in annual natural growth, currently at -0.3%. 2

The Envelope of Variation

2.1 Overview The study summarized in this paper comprises two main sociolinguistic variables. As is customary in variationist sociolinguistics, these variables will be referred to using regular parentheses ( ). In addition to the main variables, 5  http://www.arabyaffa.org/en/DynamicPage.aspx?PageID=98&CheckedOrdinal=1&Men uID=1 (accessed 3 February 2014). 6  http://www.arabyaffa.org/en/DynamicPage.aspx?PageID=99&CheckedOrdinal=1&Men uID=1 (accessed 3 February 2014).

structural change in urban palestinian arabic

201

I will discuss a number of additional variables, for which only preliminary data are currently available. However, the preliminary data for the additional variables, alongside the abundant data and careful analysis for the two main variables, paint what I believe to be a solid picture of a contemporary language contact situation between two Semitic languages. The two main variables are (ʕ) and (EMPH). The former refers to the voiced pharyngeal fricative in Arabic, which in Palestinian Arabic is often depharyngealized in a variety of phonetic realizations. The latter is in fact a cluster of variables: (dˤ), (sˤ) and (tˤ)—the three so-called emphatic consonantal phonemes of urban Palestinian Arabic.7 While Arabic and Hebrew are both Semitic languages of the Central Semitic branch,8 the phonemic inventory of Hebrew is impoverished in comparison with Arabic. Old Hebrew had already lost several Proto-Semitic (PS) consonantal phonemes. Modern Hebrew has since lost several more. Classical Arabic (CA), with its 28 consonantal phonemes, has all but one of the PS consonants. The 29th, a voiceless lateral fricative /ɬ/ (conventionally marked as /ś/ by most Semitists), is found in Old Hebrew but has merged in Modern Hebrew with the voiceless alveolar fricative /s/. CA has a set of emphatic (CA mufaxxam) consonants, which are pharyngealized or velarized (or, according to Shahin 1995, 1996, uvularized)9 counterparts of non-emphatic consonants:

7  Note that urban Palestinian Arabic, the variety under investigation, has merged the two voiced pharyngealized alveolars: ðˤ>dˤ. For a detailed account of this merger, and the question of whether it is indeed a merger per se, see Al-Wer 2004. 8  Arabic used to be classified as a South Semitic language, alongside the Ethiopian and South Arabian languages, but a newer classification had been proposed by Hetzron in 1972, and Rubin’s (2008) family tree is something of a refinement of that tree. Rubin means for it to represent the “subgrouping of the Semitic family as it is best understood given the facts available to date” (Rubin 2008: 61). Note that in a subsequent analysis, Huehnergard and Rubin (2011) consider various models of “phyla,” represented graphically by trees, but also entertain the thought that language contact may have played such a significant role in antiquity as to render the genetic model of language classification inadequate. They rely in part on a statement by Labov (though of course, they could have cited many others) and conclude: “Only by integrating the two models of language change, the family tree model and the wave model, can we explain the relationships among the Semitic language [sic]” (Huehnergard and Rubin 2011: 267). 9  I wish not to enter the debate on the precise phonetic nature of “emphasis” in Arabic. In Shahin 1996 it is strongly argued that pharyngealization and uvularization are two discrete processes. McCarthy 1994 is also of the view that: “The so-called pharyngealized consonants of Arabic should really be called uvularized.” On the other hand, both traditional groupings of Arabic consonants and modern acoustic accounts find that the emphatics share features with the pharyngeal and uvular consonants alike and that the emphatics do have

202

horesh

/ḍ/, /ṣ/, /ṭ/, /ð̣ /.10 Old Hebrew merged the first and fourth of these with the second, and Modern Hebrew merged the third with /t/, and the merged Old Hebrew /ṣ/ is pronounced as an affricate /c/ [ʦ] (see Steiner 1982: 11–44 for a discussion of the origins of the diachrony of the affricated reflexes of Hebrew /ṣ/ in the various locales in which it was spoken). In Palestinian Arabic, as in virtually every contemporary vernacular of Arabic, Classical Arabic */ḍ/ and */ð̣ / are merged (or appear to be merged; see Al-Wer 2004) either as a stop or a fricative, depending on whether the dialect in general has retained the pronunciation of interdental fricatives. Dialects that merged /θ/ and /ð/ with /t/ and /d/, respectively, typically only have a voiced emphatic alveolar stop as a reflex of both /ḍ/ and /ð̣ /. Dialects that have retained the non-emphatic interdentals have a voiced emphatic interdental fricative as the merged (or nearly-merged) fricative. In some dialects, a new variant, a voiced emphatic alveolar fricative /ẓ/ has emerged, usually by means of lexical diffusion and borrowing from CA or MSA into the vernacular. In the Jaffa dialect, a typical urban Mediterranean variety, all historical interdental fricatives have alveolar plosive reflexes. Another difference between contemporary Hebrew and Arabic is that Arabic has distinctions of quantity: consonant gemination and vowel length. Biblical Hebrew (as far as the Tiberian “pointing” system for indication of vocalization can tell us) was beginning to lose some of the length distinctions for certain vowels, in some cases substituting different vowel qualities for a PS long vowel. Also in Biblical Hebrew, certain “guttural” consonants (pharyngeals, laryngeals and the liquid /r/) were not geminated, often with compensatory lengthening of a preceding vowel. Modern Hebrew is much more categorical: gemination and long vowels do not exist. Finally, most speakers of Modern Hebrew do not have the PS (and Old Hebrew) pharyngeal fricatives that most varieties of Arabic have retained. In Modern Hebrew, /ħ/ has merged with /x/, and /ʕ/ has merged with the glottal stop /ʔ/ (both of which, as well as /h/, are often realized as a phonetic zero).



constriction in the pharynx. I will therefore use the term pharyngealized for the emphatics and transcribe them with a superscript ˤ accordingly. 10  In this section only, I am transcribing the emphatics with an underdot, as is customary among Semitists. This is mostly because it is unclear what the exact nature of PS emphatics was. It is common to think that they were historically ejective. See Bergsträßer (1983: 4): “The oldest pronunciation of the emphatics was probably with following release of the glottal stop, as is still the case in modern Ethiopic; this is widely replaced by a weakened pronunciation with velarization—broader contact between tongue and palate, particularly the soft palate.” Elsewhere, as stated above, I use a more IPA-compliant transcription.

structural change in urban palestinian arabic

203

2.2 Lenition in Palestinian Arabic The processes of sound change that I am grouping together as “lenition” include the following: 1. 2. 3.

Shortening of long vowels V1ː→V111 Degemination of consonants C1C1→C1 Depharyngealization of the voiced pharyngeal ʕ→ʔ~∅ fricative 4. Depharyngealization of secondary pharyngeal  dˤ    d articulation of emphatic alveolar stops and fricative  –  sˤ – → –  s – tˤ    t I am using “lenition” as a categorization of both types of features (those involving loss of pharyngeal articulation and those involving loss of length distinction), mainly because the end result of each of these processes is a less complex system, insofar as it includes fewer features from which the speaker needs to choose, and the features that are taking over are in a sense of simpler articulatory nature. This is in line with Campbell’s definition: “Lenition is a reasonably loose notion applied to a variety of kinds of changes in which the resulting sound after the change is conceived of as somehow weaker in articulation than the original sounds” (1998: 41). 3

The Fieldwork

3.1 The Sample The data for this study were extracted from sociolinguistic interviews conducted in Palestine by the author with native speakers of Palestinian Arabic, mostly in 2004 and 2005 (two interviews from a pilot study in 1999 were also used for parts of the analysis). For the purpose of controlling for and analyzing the role of contact between Arabic and Hebrew, a two-pronged approach was employed. First, the sample included not only speakers for whom it was known that contact was a part of their linguistic repertoire, but also a subsample of speakers (referred to hereafter as the “control group”) for whom it was assumed a priori that contact with Hebrew was nonexistent or negligible 11  In the formal representation of these processes I am using the synchronic arrow (→) rather than the diachronic angled bracket (>), as for now I am only treating these processes as variable rules, which do not necessarily affect the underlying phonological value of the features involved.

204

horesh

at best. The former group of speakers consisted of natives of Jaffa, a Palestinian city that was ethnically cleansed in 1948 and became a mixed Arab-Jewish town, and later a borough of the larger, predominantly Jewish, City of Tel AvivJaffa. The latter was interviewed in Jerusalem and Ramallah and comprised residents of these two West Bank cities and their environs. Secondly, the speakers within the Jaffa sample are stratified (albeit not evenly) to represent various levels of language contact within the town itself. For instance, speakers of different age groups are assumed to have had varying degrees of exposure to Hebrew during their formative years and throughout their primary and secondary education. Additional social factors, including language of primary and secondary education (schooling in Jaffa is offered in Arabic, Hebrew, and French, depending on the type of school), proficiency in Hebrew as assessed by the researcher, and others (see Horesh 2014 for details) were coded for and included in a careful quantitative analysis of the data. TABLE 1  Number of speakers sampled by age and socioeconomic status Jaffa

West Bank

Age | Sex

F

M

F

M

14–35 36–60 61+ Total Grand total

9 3 3 15

7 1 1 9

0 5 0 5

3 4 0 7

24

12

3.2 The Interviews In addition to some of the standard urban topics of discussion that sociolinguists use to elicit vernacular forms (danger of death, premonitions, childhood games, etc.), I saw the need to construct a number of modules that would address questions of language contact and language attitude. Examples of questions of this sort can be found in the interview excerpts in a study of Anglophones in Quebec, by Nagy, Moisset, and Sankoff (1996). In the Jaffa case, similar modules were adapted to fit the local setting. Part of my strategy was to conduct the interview with a short Hebrew component,

structural change in urban palestinian arabic

205

leading to a longer portion in Arabic. It had been my experience that as a non-Arab who happens to speak Arabic, I am often identified as an “other” (more specifically, a Jewish Israeli, regardless of my own personal views of my identity). Oftentimes, when I initiate a dialogue in Arabic with Palestinians, my interlocutors reply in Hebrew and impose a switching of the language of the interaction. Since I wanted to gather some information not only about the speakers’ own assessment of their Hebrew and their level of contact with Hebrew speakers, but also about their actual level of proficiency in Hebrew and the degree to which their Hebrew resembled that of native speakers, it seemed like a good idea to commence each interview with the Hebrew component, including, inter alia, an explicit language-centered module of questions and a short reading passage, and then introducing Arabic through an abrupt shift on my part in the form of “okay, now in Arabic!” (uttered in Arabic). 4

Analysis and Discussion: Variable 1—(ʕ)

4.1 Description of the Variable 4.1.1 A Complex Variable Upon embarking on the fine-grained work of coding the corpus for this variable, it became clear that it is a more complex variable than I had originally envisaged it to be. I had suspected that there would be more to this variable than simply a pharyngeal realization—the traditional [ʕ]—and a mere deletion of the segment. What I had thought would be an intermediate variant, based on casual auditory observation of the data I had been collecting, was a sort of a glottal stop [ʔ]. 4.1.2 Two Types of Vocalization However, when I began coding, I discovered the existence of at least two types of vocalic variants of (ʕ) in the Palestinian dialect. One of them I am calling ‘compensatory lengthening’, as it involves simply the lengthening of the preceding otherwise short vowel, e.g., baʕdeːn → baːdeːn ‘later.’ The second type of vocalic variant is ‘syllabic vocalization’. It consists not only of the addition of a vocalic mora, but also of resyllabification of the word. It often occurs at the final word boundary, though it is not limited to this position. A prime, recurring example of this phenomenon is found in the word us.buːʕ → us.bu.a ‘week.’ What I find intriguing about this particular variant in the context of the Arabic-Hebrew language contact situation—and this specific lexical item happens to illustrate this quite neatly—is that there is a similar phenomenon in

206

horesh

Hebrew, traditionally known as pataħ gənuva such as in the Hebrew equivalent of us.buːʕ ~ us.bu.a—ʃa.vu.a(ʕ)—the [ʕ] only pronounced nowadays by Hebrew speakers whose dialects are influenced by Arabic (usually because of ancestry within the Middle East or in North Africa). 4.1.3 Creaky Voice It is worth noting here that an additional type of variant was quite common in the data analyzed for this study. I had originally considered coding this variant, often referred to in the literature as ‘creaky voice’, as separate from the traditional pharyngeal [ʕ], and in a subsequent study may in fact return to these tokens and study them in their own right. However, I have chosen, for the sake of both convenience and a certain degree of logic, which I shall defend henceforth, to code the creaky voiced tokens as ‘regular’ pharyngeals. The convenience argument has to do with the lack of certainty in distinguishing creaky tokens from pharyngeal ones in a good deal of the cases. On the one hand, many profoundly creaky tokens do show up on spectrographic images in quite recognizable fashions. On the other hand, I decidedly only used the freeware Praat and its spectrographic feature as an aid in determining the quality (and in the case of compensatory lengthening—quantity) of variants, secondary to my own auditory impressions of the phonations. For this reason, in those cases in which there was a discrepancy between Praat showing some degree of creakiness, but the auditory impression being that of an actual pharyngeal, I preferred to code the tokens as the latter. 4.1.4 A Hierarchy of Lenition In terms of logic, my rational was as follows. The idea behind coding for multiple types of variants was that there was a hierarchy of lenition, possibly related to contact between Arabic and Hebrew, and that this hierarchy was gradient. This gradient nature of the variants would subsequently allow for a multivariate analysis (e.g., using Rbrul),12 treating the variables on a continuous scale. It was fairly clear where on this scale the four variants I eventually coded for would fall. Adding ‘creaky voice’ into the mix would have potentially jeopardized the analysis, as it would have been virtually impossible to assess where this variant belongs on the continuous scale. It is worth noting here that ‘creaky voice’ (or ‘creaky phonation’ as a more technical, phonetic term) has been widely observed in a variety of dialects 12  Rbrul is the current standard software among many variationist sociolinguists for conducting multivariate statistics, replacing the hitherto ubiquitous Varbrul/Goldvarb program (see Johnson 2009).

structural change in urban palestinian arabic

207

(see, e.g., Heselwood 2007: 6, 13, 17). One variety in which it is common to see this feature is Maltese, which is of course tempting for drawing analogies with Palestinian Arabic because of the contact situations both varieties have been subject to. What is interesting about creaky voice as a reflex for /ʕ/ in Maltese, however, is that the literature seems to mostly attest to this feature being prevalent in the dialects of Gozo, the smaller inhabited island of Malta, which is considered to have a more conservative dialect, one that preserves pharyngeal consonants (see, e.g., Hume et al. 2009: 15, fn 1). Agius informs us that “[a] ‘creaky voice’ to describe the alternation in the pronunciation of the voiced pharyngeal fricative [. . .] occurs practically in all Gozitan dialects” (1992: 130). He further asserts that (similarly to the Palestinian case) “at times it is lost except for a compensatory lengthening of adjacent vowels.” Compensatory lengthening for historical /ʕ/ is also the focus of Hume et al.’s 2009 study involving two speakers from the island of Malta itself (not Gozo). Their findings, however (e.g., p. 42) indicate that unlike the Palestinian data, in Maltese, vowels adjacent to a historical voiced pharyngeal are typically not lengthened to the same extent as phonemically long vowels. In summation, the five variants eventually coded for with respect to the (ʕ) variable are as follows, in ascending order of lenition: 1. 2. 3. 4. 5.

Pharyngeal Glottal Compensatory lengthening Syllabic vocalization Deletion

4.2 Quantitative Analysis 4.2.1 Cross-tabulation Figure 5 is a graphic representation of a cross-tabulation of the data, in which for each speaker, two values are reported. The left-hand side bar for each speaker represents a value for the application of a variable rule that assumes deletion of the pharyngeal only, ignoring all intermediate variants. The bar on the right for each speaker aggregates all of the lenited variants: deletion (i.e., ∅), compensatory lengthening, and syllabic vocalization, and treats them collectively as the application value, in opposition to the fully pharyngeal [ʕ] variant. The purpose of this cross-tabulation and its graphic representation is mostly for getting a general feel of the data, prior to submitting it to more rigorous “number crunching.” What we see here is a rather convincing confirmation

208 0.9

horesh West Bank speakers

Jaffa speakers

0.8 0.7 0.6 0.5

App=∅ (vs. [ʕ])

0.4

App=any lenition

0.3 0.2

0

ʕuθmaːn ħusni faːtˤme inaːs nuha raːwi ʕali amal bijaŋka ʒamiːl ʒamiːla marjam muːsa muniːra nabiːl neviːn saːlim saːmi sawsan tˤaːriq umm-xaliːl umm-jaziːd widaːd xaːlid jaziːd zahra total

0.1

FIGURE 5

Cross-tabulation of (ʕ) by speaker, community and 2 types of lenition.

of the general hypothesis at hand that correlates lenition of pharyngeal articulation with contact a speaker has with Hebrew. This is true across the sample. In other words, while the distinction between the bilingual Jaffa speakers and the non-bilingual West Bank speakers is helpful to maintain a perspective regarding speakers’ general proficiency with the superstrate language and its potential to induce change on a daily basis, high variability is apparent within the bilingual group as well. Several Jaffa speakers have relatively low values for deletion (the blue bars in Figure 5), but when considering all forms of lenition (the red bars), their values surge. Consider such speakers as ‘tˤaːriq’,13 who was an 18 year-old high school student at the time of the interview (born 1987). His value for ∅ realization of (ʕ) is 0.109, but his value for lenition in general is 0.562. ‘umm xaliːl’ is an even more striking case. Not only is she much older (born 1928), but her life trajectory is much more complex. She had left Jaffa in 1946 to marry a man in the village of abu kiʃk, where the dialect spoken is of a Bedouin type (I have met her 13  All speakers are referred to by pseudonyms, hence the single quotes.

structural change in urban palestinian arabic

209

husband, and he still speaks a Bedouin dialect), and after being expelled from their village (which was eventually destroyed) in 1948 to Gaza and to northern Sinai in Egypt, they returned to Palestine, but to a different village altogether, ʒalʒuːlje, where I ended up interviewing her in 2004. Similarly to ‘tˤaːriq’, ‘umm xaliːl’s’ speech exhibits a very low rate of full deletion of the voiced pharyngeal fricative (0.111). Her value for lenition at-large is seven-fold at 0.752. I have no clear-cut explanation for this pattern, only a few educated guesses. We can begin with the gender hypothesis, which is known from general sociolinguistics. As a woman, albeit the oldest woman in the sample, it is not unreasonable for us to expect somewhat more advanced tokens from her than from her male counterparts. Also recall that at the time of the interview, ‘umm xaliːl’ had been displaced from the original environment of her native dialect for six decades. She had acquired neither the rural, fallaːħi, dialect of her “new” domicile nor any of the dialects of the places in which she lived during her transition there (including her husband’s Bedouin dialect). She reports still visiting Jaffa from time to time (a mere 35-km drive from her current village). While her Hebrew proficiency is the lowest of all the Jaffa speakers in the sample, and her level of contact with Hebrew speakers the lowest as well, she is still exposed to Hebrew through mass media, shopping trips to Jewish cities, and other sources. All of the above amount to a rather nebulous set of factors, which admittedly could have led to either a more conservative linguistic behavior or, as indeed we see in this case, a more advanced one. I tend to think that her rather special position in her family and community have led her to assert her urban origins in some ways that she is aware of (e.g., the non-pronunciation of interdentals and the glottal realization of historical /q/, see Shahin 2008: 527) and in others—such as the lenition in the case of the (ʕ) variable—that are below her level of consciousness. 4.2.2 Multivariate Analysis Elsewhere (Horesh 2014) I explain in greater detail the rationale behind the quantitative methodology I have used to attain the results I report in this section. I also report the statistics much more exhaustively there than I do here. What I wish to provide below in terms of the quantitative component of the linguistic analysis of the variation is the conclusion arising from the statistical modeling, with only a representative set of numerical corroboration. In Table 2 below, “Log-odds” are numbers between -∞ and ∞, such that any positive numbered log-odd indicates favoring application of the variable rule, and negative numbered log-odds indicate disfavoring. A log-odd of 0 indicates neutrality (Johnson 2009: 361).

210 TABLE 2

horesh Rbrul results for Jaffa & West Bank (binary: all lenition variants)

R2=0.212

Age group (p > > >

ḥ ʕ t d ṭ ʕ

5  Inscriptions found in North-Eastern Syria and adjoining areas of Anatolia and northern Mesopotamia, see Kaufman 1997: 114.

238

Jastrow

The third sibilant of Proto-Semitic, *ś, is still distinguished in Biblical Aramaic but in Syriac, the predecessor of modern Ṭuroyo,6 it is merged with the phoneme s. Thus, with the end of the Old Aramaic period and the emergence of Middle Aramaic varieties such as Syriac, Aramic had suffered a massive reduction of its phoneme inventory by at least seven consonant phonemes. Had this tendency continued, modern Neo-Aramaic languages would now be counted among those languages of the world which have an extremely reduced consonant inventory. But already in the Middle Aramaic period there was a new sound shift underway which eventually would restore to Aramaic some of the lost phonemes. This sound shift which is conventionally referred to by the term of “Begadkefat rule” consists in the splitting up of six stops (p, b, t, d, k, g) into two allophones each, one a stop and one a fricative, as shown in table (5): (5) Middle Aramaic: Begadkefat rule (appearance of fricative allophones) Labial

/p/ > /b/ >

Dental

[p] [f] [b] [v]

/t/ > /d/ >

Velar

[t] [ṯ] [d] [ḏ]

/k/ > /g/ >

[k] [x] [g] [ġ]

By this process a number of consonants which had belonged to the older phoneme inventory before they were eliminated by sound shifts now reappeared as allophones, namely the dental fricatives [ṯ] and [ḏ] and the velar fricatives [x] and [ġ]; in addition, two new consonants were introduced, namely the labial fricatives [f] and [v]. The original distribution of Begadkefat allophones follows a simple rule: stops are retained after a preceding consonant and word initially, fricative allophones appear after a preceding vowel. This distribution pattern, however, was 6  It is common lore that Syriac (in its West Syriac form) cannot be regarded as the actual predecessor of modern Ṭuroyo but is still close enough to serve as basis for historical deductions, pretty much as Latin is used to illustrate the earlier form of the Romance languages because the presumed “Vulgar Latin” is only sparcely attested.

239

the Consonant System of Ṭuroyo

gradually weakened already in Syriac, mainly due to vowel elision, as shown by the example in table (6). (6) Syriac: Beginning phonemicization of allophones

a) b) c)

‘Leprosy’

‘Leprous’

*garbā *garbā garbā

*garibā *garivā garvā

The fricative allophone [v] in *garivā was caused by the preceding vowel i, however, after the elision of the vowel, it appears in a position which would call for a stop. As a result Syriac garbā: garvā emerge as a minimal pair, thereby constituting a phonemic opposition between /b/ and /v/. In Syriac, the phonemic opposition between stops and fricatives resulting from the Begadkfat rule is still marginal, in the majority of cases they rather appear according to their allophonic distribution, however, the beginning of a gradual phonemicization can be observed. In Neo-Aramaic this development has reached its final stage, resulting in the addition of six new phonemes (i.e., the former fricative allophones) to the phoneme inventory, as shown in figure (7) for Ṭuroyo: (7) Ṭuroyo: Phonemicization completed Labial stops fricatives

/p/ /f/

Dental

/b/ /w/

/t/ /ṯ/

Velar

/d/ /ḏ/

/k/ /x/

/g/ /ġ/

As demonstrated in (6) above the process of phonemicization was often brought about by changes in the syllable structure (e.g., vowel elison). Another important factor was analogical restructuring of root consonants. This can be shown by the following Ṭuroyo examples in (8):

240

Jastrow

(8) Ṭuroyo: verbs ‘to come’ and ‘to sit down’ Root

Present tense

Past tense

*ʾty *ytb

oṯe < *āṯē ‘he comes’ yotu < *yāṯeḇ ‘he sits’

āṯi < *attī 7 ‘he came’ yātu < *yattīḇ ‘he sat’

Obviously, both present tense forms, harking back to the old participle *pāʿel, should have a fricative /ṯ/, while both past tense forms, harking back to an old participial form *paʿʿīl, should have the stop /t/. In actual fact, however, the fricative has been generalized in the verb ‘to come’, while the stop has been generalized in the verb ‘to sit down’. Thus oṯe/yotu and aṯi/yatu in Ṭuroyo are contrastive pairs demonstrating that /t/ and /ṯ/ can appear in the same phonetic environment. Here are some further minimal or contrastive pairs demonstrating the phoneme status of the former allophones in Ṭuroyo (9): (9) Minimal and contrastive pairs in Ṭuroyo /t/:/ṯ/ /d/:/ḏ/ /k/:/x/ /g/:/ġ/

kǝtyo kǝṯyo admo aḏno kǝkwo kǝxlo mdaglo raġlo

‘there is’ ‘she comes’ ‘blood’ ‘ear’ ‘star’ ‘she eats’ ‘she lies (tells lies)’ ‘foot’

7  The past tense pattern paʿǝl < *paʿʿīl in Ṭuroyo exists only in the first stem (old Peal) and is lexically restricted to a number of mainly intransitive verbs, e.g., aṯi ‘he came’, yatu ‘he sat down’, damǝx ‘he slept; lay down’, kali ‘he stopped’, qayǝm ‘he stood up’, etc.

241

the Consonant System of Ṭuroyo

To sum up the preceding discussion: The fricative phonemes *ṯ, *ḏ, *x, and *ġ which were lost during language evolution from Proto-Semitic to Old Aramaic or Late Old Aramaic, reappear as new phonemes in Neo-Aramaic after going through an allophonic stage in Middle Aramaic. Ṭuroyo is, however, a very conservative Neo-Aramaic language which preserved *ṯ, *ḏ, *x, and *ġ unchanged;8 in most NENA dialects they were subject to further sound changes by which some of them were again eliminated. This is shown for /ṯ/ and /ḏ/ in the following table (10): (10) Reflexes of Middle Aramaic *ṯ and *ḏ in some NENA dialects9 Middle Aramaic Iraqi Kurdistan – Christians, Jews Iraqi Kurdistan – Jews (Zaxo) Iranian Azerbaijan – Christians Iranian Kurdistan – Jews

bayṯā ‘house’ beṯa besa beta belá

īḏā ‘hand’ iḏa10 iza ida ilá

Whereas in the dialect of the Jews of Zaxo the former interdentals had shifted to sibilants, in Jewish dialects further to the south, e.g., in Dehok province, ṯ and ḏ were preserved (see Mutzafi 2007: 15); this may likewise be due to the closer proximity of the Arabic-speaking area (Mossul). III The system of consonant phonemes of present-day Ṭuroyo is shown in the following table (11).

8   The same is true, mutatis mutandis, for Western Neo-Aramaic. 9   See Hopkins 1999: 322. 10  In NENA vowel length is no longer phonemic, except for some marginal cases. It is therefore not indicated in transcription although stressed vowels in open syllables are usually pronounced long, e.g., [bēṯa], [īḏa].

242

Jastrow

(11) Consonant phonemes of present-day Ṭuroyo Bilabial/ labiodental

Dental/ interdental

stop

pb

affricate fricative

td ṭ (ḍ)

fv

ṯḏ (ḏ̣) sz ṣ n (ṇ) l (ḷ) r (ṛ)

sibilant nasal m lateral trill approximant w

Palatal

čǧ

Velar

Uvular Pharyngal Glottal

kg

q

xġ

(ʔ)

ḥʕ

h

šž

y

Symbols in parentheses designate phonemes of limited distribution. ḍ and ḏ# occur in loans from Arabic and may be replaced by d and z, respectively. ʔ is distinctive only word-internally where it is rare. ṇ, ḷ, and ṛ occur in verb inflection as a result of the assimilation of stem-final -r to suffix-initial n, l, or k, e.g., *komar-no → komaṇṇo ‘I say’, *mīr-le → mǝḷḷe ‘he said’, *mīr-Ke → mǝṛṛe ‘they said’. The fricative consonants resulting from the so-called bgdkpt split have all been preserved, as can be seen in the above table: f, w, ṯ, ḏ, x, ġ. Note that the reflex of *ḇ is w, e.g., lowəš ‘he puts on’, not v; v is a rare consonant occurring only in loanwords (see V below). Middle Aramaic *p/*p̄ have both yielded f in Ṭuroyo, e.g., ftəḥle ‘he opened’, foṯo ‘face’ < *pāṯā, however, the shift is not yet completed as some Ṭuroyo dialects (not Midyat) preserve a sporadic p, e.g., poṯo. The consonant phoneme inventory of present-day Ṭuroyo thus is not only considerably larger than that of late Old Aramaic and Middle Aramaic, it even exceeds the reconstructed phoneme system of Proto-Semitic. The reason for this wealth of phonemes lies, in the view of the present author, in the influence of the contact languages which have been active in two ways: in encouraging or furthering the preservation of some inherited phonemes and in introducing new phonemes via loanwords.

243

the Consonant System of Ṭuroyo

IV The contact language Arabic (Anatolian or qəltu Arabic, here the subgroup of the Mḥallami dialects) was spoken fluently by part of the Ṭuroyo population. Thus, for example, the inhabitants of Midyat also spoke the Mḥallami dialect of the neighbouring Muslim village of Astal. Although the Arabic-Aramaic bilingualism did not extend to the whole Ṭuroyo area it must have exerted a considerable influence on Ṭuroyo, both by linguistic practice and by the large number of Arabic loanwords. In the view of the present author, these two factors contributed considerably to the preservation of a number of consonants which did not survive in the NENA dialects to the east which were not in contact with Arabic. a) The most striking case of preservation is that of the two pharyngeal fricatives, voiceless /ḥ/ and voiced /ʕ/. Due to the contact of spoken Arabic and the large amount of Arabic loanwords, these two phonemes are still currently part of the Ṭuroyo phoneme inventory. A few examples are given in (12) and (13): (12) /ḥ/ in Ṭuroyo

word initial word internal word final

Inherited words

Loanwords

ḥreno ‘other’, ḥoṯo ‘sister’ aḥna ‘we’, yarḥo ‘month’, naḥit ‘he descended’ fotiḥ ‘he opens’

ḥarb ‘war’, ḥŭdūd ‘border’ mḥalle ‘quarter’

(13) /ʕ/ in Ṭuroyo

word initial word initial word final

Inherited words

Loanwords

ʕisri ‘twenty’, ʕamo ‘people’, ʕayiš ‘he lived’ naʕimo ‘small’, aʕme ‘with him’, greʕo ‘servant’ abiʕ ‘he wanted’

ʕāde ‘costum’ sāʕa ‘hour/watch’, ṣinʕa ‘profession’ rabiʕ ‘spring’

244

Jastrow

In the NENA dialects which are not in direct contact with Arabic ʕ was shifted to a glottal stop ʔ and, in the more progressive dialects, to a glide (w or y) or Ø, as illustrated by figure (14) below. Betanure (Mutzafi 2007) is a conservative Jewish dialect from central northern Iraq, Čāl is a more progressive Christian dialect of the same area, but situated in Turkey right across the Turkish-Iraqi border.11 (14) Reflexes of Middle Aramaic *ʕ in some NENA dialects Ṭuroyo

NENA (Betanure)

NENA (Čāl)

Gloss

šawʕo beʕe bǝʕto tarʕo

šōʔa beʔe bǝʔta tarʔa

šawwa beye beta taṛṛa12

‘seven’ ‘eggs’ ‘egg’ ‘door’

b) Similarly, the velar fricatives /x/ and /ġ/ in Ṭuroyo most likely owe their preservation to the fact that they are likewise part of the phonemic inventory of Anatolian Arabic. Since /x/ and /ġ/ in Ṭuroyo result from the Begadkefat rule (see II above) they occur in native words only in word internal and word final position; in word initial position they are almost always loan words from Arabic, see (15), (17): (15) /x/ in Ṭuroyo Inherited words word initial word internal

gboxe ‘he weeps’, hawxa ‘so’, frixle ‘he rubbed’

Loanwords

xabro ‘word’, xalyo ‘empty’, xayifo ‘quick’ šaxṣ ‘person’

11  See Talay 2008: 70 ff. 12  In this form, *rʔ assimilated to ṛṛ; there are dialects which have tarra without velarization and others in which ʔ was simply elided, causing a short ă vowel in open syllable: tăra, see Talay l.c.

245

the Consonant System of Ṭuroyo

word final

Inherited words

Loanwords

aʕmux ‘with you (m.)’, baytax ‘your (f.) house’, kimhalix ‘he walks’

šēx ‘shaikh’, wăsax ‘dirt’

Whereas Ṭuroyo preserved /ḥ/ and /x/ as independent phonemes they merged to *ḥ in NENA. *ḥ in turn shifted to /x/ in the large majority of NENA, only a small group of dialects in eastern Turkey, the best-known of which is that of Hertevin,13 continued the older pronunciation *ḥ. This is illustrated by figure (16) below. (16) Middle Aramaic *x and *ḥ Middle Aram.

Ṭuroyo

NENA (Hert.)

NENA

Gloss

māḥē ḥamšā bāxē

moḥe ḥamšo boxe

maḥe ḥamša baḥe

maxe xamša baxe

‘he hits’ ‘five’ ‘he weeps’

(17) /ġ/ in Ṭuroyo Inherited words word initial word internal word final

raġlo ‘foot’, syoġo ‘wall’, lawġul ‘inside’ mašíġ! ‘wash!’

Loanwords

ġálăbe ‘very’, ġēr ‘except’ baġlo ‘mule’, šuġlo ‘work’ ṣāyiġ ‘jeweler’

In NENA ġ, a rare consonant, has not left any traces, e.g. šula ‘work’ < arab. šuġl, cf. Ṭuroyo šuġlo. 13  See Jastrow 1988.

246

Jastrow

c) Even the interdental fricatives /ṯ/ and /ḏ/ in Ṭuroyo might owe their preservation to the fact that they are likewise part of the phonemic inventory of Anatolian Arabic. Since /ṯ/ and /ḏ/ in Ṭuroyo result from the Begadkefat rule (see II above) they occur in native words mostly in word internal and word final position; in word initial position they are rare and almost always loanwords from Arabic, see (18), (19) (18) /ṯ/ in Ṭuroyo Inherited words

Loanwords

word initial

ṯnīḥ14 ‘he rested’

word internal

tloṯo ‘three’, qriṯo ‘village’, saṯwo ‘winter’, šabṯo ‘week’ māyiṯ ‘he died’, māqíṯ ‘light!’ (e.g., the fire)

ṯabuto ‘firm, stable’, ṯēni ‘second’ maṯlo ‘proverb’

word final

(19) /ḏ/ in Ṭuroyo Inherited words word initial word internal word final

iḏo ‘hand’, ḥḏōḏe ‘one another’ gmawqiḏ ‘he burns (sth.)’, Məḏyaḏ ‘Midyat’ (town)

Loanwords

ḏanbo ‘crime’

In Neo-Aramaic areas which were not in contact with Arabic, /ṯ/ and /ḏ were not preserved but underwent different sound changes as shown for /t/ in (10) above.

14  This is a back formation from məṯnəḥ < * mettĕnīḥ ‘he rests’.

247

the Consonant System of Ṭuroyo

V Another group of consonant phonemes in Ṭuroyo were not inherited (and replenished by borrowings) but introduced exclusively via borrowings, see (20). (20) Introduction of new consonant phonemes into Ṭuroyo via borrowing Phoneme

Loanword

Source

Gloss

/v/ /č/

tavda čāra čīrōke ǧāmūdo žinníke ṭāži

Kurdish Turkish Kurdish Arabic Kurdish Kurdish

‘together’ ‘remedy’ ‘popular tale’ ‘cold’ ‘woman’ ‘greyhound’

/ǧ/ /ž/

VI In concluding, I would like to draw attention to the “Sprachbund” phenomenon which was shortly mentioned in the beginning. As pointed out in footnote 1, the convergencies resulting from close language contact over many centuries are not restricted to phonology but manifest themselves also in morphology (e.g., verb inflection) and in the lexicon, however, even a segment of the phoneme inventory of the languages in question can illustrate the extent of these convergencies. Table (21) shows the present-day inventories of labial, alveolar, and velar stops/affricates and corresponding fricatives in the three languages in question. (21) Convergence of consonant systems

Kurdish (Kurmancî) Arabic (Mḥallami)

Bilabial

Alveolar

Velar

pb fv pb fv

čǧ šž čǧ šž

kg xġ kg xġ

248

Jastrow

(cont.)

Aramaic (Ṭuroyo)

Bilabial

Alveolar

Velar

pb

čǧ

kg

fv

šž

xġ

Synchronically these systems are identical; in a historical perspective, however, in the inventory of each language some consonants were lacking. The consonants marked in bold were not originally part of the system of the respective language but were introduced via borrowings. Thus, for instance, Arabic /g/ was introduced via loanwords from Kurdish and Turkish, while /č/, /ǧ/, and /ž/ in Ṭuroyo were introduced via loans from Arabic, Kurdish, and Turkish, etc. The South-east Anatolian “Sprachbund” was made possible by the centuryold, not always peaceful but on the whole stable, coexistence of several ethnic groups. Now that the latter has become a thing of the past, the “Sprachbund” itself will soon become a subject of historical linguistics. Bibliography Beyer, Klaus (trans. John F. Healey). 1986. The Aramaic Language. Its Distribution and Subdivisions. Göttingen. Bedir Khan, Emir Djeladet and Roger Lescot. 1970. Grammaire kurde (dialecte Kurmandji). Paris. Blau, Joyce. 1965. Dictionnaire Kurde-Français-Anglais. Brussels. Fassberg, Steven E. 2010. The Jewish Neo-Aramaic dialect of Challa. Leiden. Grigore, George. 2007. L’arabe parlé à Mardin—monographie d’un parler arabe périphérique. Bucharest. Hopkins, Simon. 1999. “The Neo-Aramaic Dialects of Iran,” in Shaul Shaked and Amnon Netzer (eds.), Irano-Judaica. Studies Relating to Jewish Contacts with Persian Culture throughout the Ages, Vol. IV. Jerusalem. 311–327. Jastrow, Otto. 1967. Laut- und Formenlehre des neuaramäischen Dialekts von Midǝn im Ṭur ʿAbdin. Ph.D. Diss. Saarbrücken. ———. 1977. “Zur Phonologie des Kurdischen in der Türkei,” Studien zur Indologie und Iranistik 3: 84–106.

the Consonant System of Ṭuroyo

249

———. 1978. Die mesopotamisch-arabischen qǝltu-Dialekte, Vol. 1. Phonologie und Morphologie (Abhandlungen für die Kunde des Morgenlandes 43.4). Wiesbaden. ———. 1981. Die mesopotamisch-arabischen qǝltu-Dialekte, Vol. 2. Volkskundliche Texte in elf Dialekten (Abhandlungen für die Kunde des Morgenlandes 46.1). Wiesbaden. ———. 1992. Lehrbuch der Ṭuroyo-Sprache (Semitica Viva, Series Didactica 2). Wiesbaden. ———. 1993. Laut- und Formenlehre des neuaramäischen Dialekts von Midǝn im Ṭur ʿAbdin (Semitica Viva 9; 4th ed.). Wiesbaden. ———. 1994a. Der neuaramäische Dialekt von Mlaḥsô (Semitica Viva 14). Wiesbaden. ———. 1994b. “Neuentdeckte aramäische Dialekte in der Türkei,” in Cornelia Wunsch (ed.), XXV. Deutscher Orientalistentag in München, 1991. Ausgewählte Vorträge (ZDMG, Supp. 10). Stuttgart. 69–74. ———. 1997. “The Neo-Aramaic Languages,” in Robert Hetzron (ed.), The Semitic Languages. London. 334–377. ———. 1997. “Zum neuaramäischen Dialekt von Hassane (Provinz Şirnak),” in Asma Afsaruddin and A. H. Mathias Zahniser (eds.), Humanism, Culture and Language in the Near East. Studies in Honor of Georg Krotkoff. Winona Lake. 275–281. ———. 2001. “Aramäische Lehnwörter in den arabischen Dialekten der Südost-Türkei,” in Stefan Wild und Hartmut Schild (eds.), Akten des 27. Deutschen Orientalistentages. Würzburg. 615–621. ———. 2003. Arabische Texte aus Kinderib (Semitica Viva 30). Wiesbaden. ———. 2004. “Aramäische Lehnwörter im arabischen Dialekt von Kinderib (SüdostTürkei),” in Estudios de dialectología norteafricana y andalusí 8. Homenaje a Peter Behnstedt. 99–103. ———. 2005. Glossar zu Kinderib (Anatolisches Arabisch) (Semitica Viva 36). Wiesbaden. ———. 2006a. “Anatolian Arabic,” in EALL 1.86–96. ———. 2006b. “Arabic dialects in Turkey—towards a comparative typology,” Türk Dilleri Arastırmaları, 16. Istanbul. 153–164. ———. 2006–2007. “Where do we stand in the research on the Anatolian qǝltu dialects?,” in Nadia Anghelescu (ed.), Peripheral Arabic Dialects (= Romano-Arabica VI–VII). Bucharest. 63–69. Kaufman, Steven A. 1997. “Aramaic,” in Robert Hetzron (ed.), The Semitic Languages. London. 114–130. Mutzafi, Hezy. 2008. The Jewish Neo-Aramaic dialect of Betanure (province of Dihok) (Semitica Viva 43). Wiesbaden. Omar, Feryad Fazil. 1992. Kurdisch-Deutsches Wörterbuch (Kurmancî). Berlin. Ritter, Hellmut. 1968–69. “Kurmānci-Texte aus dem Ṭūrʿabdīn,” Oriens 21–22: 1–135.

250

Jastrow

Talay, Shabo. 2001. “Der arabische Dialekt von Hasköy (Dēr Khāṣ), Ostanatolien. I. Grammatische Skizze,” ZAL 40: 71–89. ———. 2002. “Der arabische Dialekt von Hasköy (Dēr Khāṣ), Ostanatolien. II. Texte und Glossar,” ZAL 41: 46–86. ———. 2008. Die neuaramäischen Dialekte der Khabur-Assyrer in Nordostsyrien (Semitica Viva 40). Wiesbaden. ———. 2009. Neuaramäische Texte in den Dialekten der Khabur-Assyrer in Nordostsyrien (Semitica Viva 41). Wiesbaden. Wittrich, Michaela. 2001. Der arabische Dialekt von Āzǝx (Semitica Viva 25). Wiesbaden.

Lexical Borrowings in the Eastern European Hasidic Hebrew Tale Lily Kahn 1 Introduction The Eastern European Hasidic Hebrew tale corpus is a large collection of hagiographic stories composed during the late nineteenth and early twentieth centuries by followers of the Hasidic spiritual movement in a region spread chiefly over parts of present-day Poland, Ukraine, and Russia. The tales, which focus on the lives and works of the rebbes, or Hasidic spiritual leaders, provide an unparalleled linguistic insight into the historical development of the Hebrew language. Firstly, they offer a unique perspective on the role of Hebrew in traditional Ashkenazi society in the pre-modern period because they constitute the sole extensive record of narrative and discursive language use from this setting. Secondly, the language of the tales is one of the two chief forms of Hebrew, along with that employed by the Maskilim (adherents of the Jewish Enlightenment), that immediately preceded the revernacularization project in Palestine beginning in the 1880s. Therefore, a thorough understanding of Hasidic Hebrew can help pinpoint ways in which Hasidic literature influenced contemporaneous and subsequent forms of the Hebrew language. However, despite their significance the tales have not previously been the subject of detailed linguistic analysis. One of the many fascinating and hitherto unexamined features of this idiom is its utilization of lexical borrowings from other languages. The Hasidic Hebrew tales are the product of a very distinctive type of multilingual environment. Their authors were all native Yiddish speakers, but they were schooled from an early age in written and recited varieties of Biblical and post-Biblical Hebrew and, from a slightly later age, trained in a range of Aramaic-language texts ranging from the Aramaic portions of the Bible (Daniel and Ezra) to the Palestinian and Babylonian Talmuds to the kabbalistic writings contained in the Zohar and other medieval mystical literature. In addition to their everyday spoken use of Yiddish and their familiarity with written Hebrew and Aramaic,

* The author gratefully acknowledges the British Academy for their generosity in funding the research project of which this paper forms a part.

© koninklijke brill nv, leiden, ���5 | doi ��.��63/9789004300156_014

252

Kahn

the authors lived in a Slavic-speaking (most typically Ukrainian, Polish, and Russian) environment. It is therefore informative to examine the ways in which these different linguistic influences, the Indo-European Yiddish and Slavic vernaculars on the one hand, and the fellow Semitic written language Aramaic on the other, are reflected in the lexis of their Hebrew writings, as well as the question of whether any other languages played a role in the lexical composition of these texts. I shall present and analyse the characteristic Yiddish, Aramaic, Slavic, and other borrowings appearing in the Hasidic Hebrew corpus in turn in order to illustrate the key patterns that they exemplify. 2 Yiddish Because the authors of the Hasidic Hebrew hagiographic tales were all native Yiddish speakers, and because many of the tales are thought to have initially been transmitted orally in Yiddish (Dvir-Goldberg 2003: 19), it is unsurprising that they turned to that language in order to supplement the Hebrew vocabulary of their compositions. However, the specific ways in which they utilized Yiddish are instructive, as they reveal striking patterns about the relationship between the authors’ vernacular and their written Hebrew. Yiddish loanwords are an extremely common element of Hasidic Hebrew. They are comprised almost entirely of nouns, with other content or function words attested only very infrequently (see Kahn 2015: 378–380 for details of these rare cases). These Yiddish nouns are a deeply entrenched component of Hasidic Hebrew: for example, they are regularly incorporated into Hebrew grammatical patterns, taking Hebrew prefixes such as the definite article and inseparable prepositions. These trends conform to cross-linguistic tendencies regarding loanwords in multilingual environments, whereby nouns are typically the most common type of content word to be borrowed and such nouns are generally subject to the inherited morphosyntactic processes of the borrowing language (Matras 2009: 167, 173). The sentence shown in (1) illustrates this typical Hasidic Hebrew treatment of Yiddish loanwords: the Yiddish term ‫ אונטער שלאק‬untershlak ‘lining’ is prefixed by the Hebrew definite article and integrated into the surrounding text without any explicit orthographic indication (such as vocalization, quotation marks, or brackets) of its borrowed nature. 1.

‫תחת האונטער שלאק תפור קמיע‬

taḥat ha-untershlak tap̄ ur qameʿa Under the lining a talisman is sewn (Rodkinsohn 1865: 23)

lexical borrowings in hasidic hebrew

253

The Yiddish lexical component of Hasidic Hebrew literature is dispersed fairly consistently throughout the corpus and is represented in a very wide range of semantic fields. However, despite this diversity the Yiddish loanwords almost always label concrete objects with practical, everyday uses and strong roots in contemporary Ashkenazi culture (as opposed to abstract concepts, philosophical and theological terms, etc.). These patterns are in keeping with the role of Yiddish as the authors’ vernacular language. Much of this Yiddish vocabulary denotes items and concepts referring to various aspects of contemporary Jewish life in Eastern Europe for which there were no precise Hebrew equivalents. This same phenomenon is also found in Ashkenazi Responsa literature (Betzer 2001: 102) and may therefore constitute a shared feature of a more widespread Eastern European Hebrew idiom. Yiddish lexical items appearing in the Hasidic Hebrew tales include terms relating to the fields of administration, buildings, business, Christian religion, clothes, currency, Eastern European Jewish customs, education, food, furniture and other household objects, ethnic labels, institutions, law enforcement, materials, measurements, medicine, military, occupations, transport, and others. Due to space constraints only a representative sample of these categories will be presented in the present paper; see Kahn (2015: 367–378) for the complete list. Note that the spelling of Yiddish words in the Hasidic Hebrew tales often differs from the standardized Yiddish orthography established in subsequent decades; moreover, the same Yiddish word may be spelt in various ways within the Hasidic Hebrew corpus and even within the same text. In this paper Yiddish words are presented as they appear in the texts, but standardized spelling following the YIVO system is provided in transliteration following each example. 2.1 Clothes Yiddish words are often used to denote specific items of clothing. In most cases the authors’ motivation for their selection is relatively straightforward, as many of these did not have well-established Hebrew equivalents at the time of writing. However, in some cases, there is a Hebrew equivalent: for example, (11), ‫ פאטשיילע‬fatsheyle, can be equated with the Hebrew ‫ מטפחת‬miṭpaḥat, which appears in the Bible (in Ruth 3:15) and Mishna (e.g. Kilayim, Ḥullin) with similar meaning. In these cases the Yiddish loanword may have been selected because the authors (perhaps subconsciously) strongly associated the garment in question with the contemporary Ashkenazi cultural setting embodied in the

254

Kahn

vernacular label, in contrast to the Hebrew term, which may have felt somewhat more removed from this context. This suggestion is supported by the fact that the loanwords presented here all refer to clothes that were very commonly worn by Eastern European Jews at the time of writing and are prominent features of contemporaneous Yiddish literature. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11.

‫ ארבל‬arbl ‘sleeve’ (Teomim Fraenkel 1911: 38) ‫]ברילין‬. . .[ [. . .]briln ‘eyeglasses’ (Kaidaner 1875: 15b) ‫ הויזן‬hoyzn ‘trousers’ (Sobelman 1909/10 pt. 3: 11) ‫]יארמילקע‬. . .[ [. . .]yarmlke ‘yarmulke (skullcap)’ (Lieberson 1913: 46) ‫ לייביל‬laybl ‘short jacket’ (Ehrmann 1903: 48a; Breitstein 1914: 6) ‫ פאנטעפל‬pantofl ‘slippers’ (Chikernik 1903a: 21) ‫]אונטער שלאק‬. . .[ [. . .]untershlak ‘lining’ (Rodkinsohn 1865: 23) ‫ שטריימיל‬shtrayml ‘shtreimel’1 (Bromberg 1899: 12) ‫ שטערינטוכיל‬shterntikhl ‘headscarf’ (Munk 1898: 50) ‫ פאטשיילע‬fatsheyle ‘shawl/kerchief’ (Bodek 1865: 21); cf. Hebrew ‫מטפחת‬

miṭpaḥat

2.2 Food and Drink Yiddish terms are often used to denote specific types of food and drink. As in the case of the clothes vocabulary discussed above, in most cases these terms refer to items lacking established Hebrew equivalents because they did not exist until the medieval or early modern period or because they were not common fare outside of Eastern Europe. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21.

‫ איינגימאחטש‬ayngemakhts ‘preserves’ (Rodkinsohn 1864: 6) ‫ אינדיק‬indik ‘turkey’ (Berger 1910: 125; Michelsohn 1912: 37) ‫]בארשט‬. . .[ [. . .]borsht ‘borscht’ (J. Duner 1899: 82) ‫]בייגיל‬. . .[ [. . .]beygl ‘bagel’ (Greenwald 1899: 54a) ‫]טשאלענט‬. . .[ [. . .]tsholnt ‘cholent’2 (Gemen 1914: 63) ‫ קאטלעטין‬kotletn ‘cutlets’ (Walden 1914: 122) ‫ קאשע‬kashe ‘porridge’ (Chikernik 1903a: 7) ‫ קוגע''ל‬kugl ‘kugel’3 (Bromberg 1899: 42) ‫ טייא‬tey ‘tea’ (Landau 1892: 34) ‫ קאווע‬kave ‘coffee’ (Kaidaner 1875: 21b)

1  Fur hat worn by Hasidic men on the Sabbath and Jewish festivals. 2  A slow-cooked stew of meat, potatoes, and beans eaten at the Sabbath lunchtime meal. 3  Noodle or potato casserole.

lexical borrowings in hasidic hebrew

255

2.3 Furniture and Household Objects Yiddish vocabulary is also commonly used to designate furniture and household objects that were commonplace in the authors’ Eastern European setting. As above, these terms generally refer to items for which no widely accepted Hebrew terms existed at the time of writing. However, in certain cases the authors employ a Yiddish loanword despite the existence of a well-known Hebrew counterpart. Such cases are shown in (29) and (30). This phenomenon may be attributable to the fact that the Hebrew terms in question are themselves used in Yiddish and have very specific elevated connotations in that language (‘[God’s] throne’ and ‘seven-branched Temple candelabra’ or ‘oil lamp used in the Hanukkah festival’, respectively); the authors’ use of their much more prosaic Yiddish equivalents may reflect a desire (whether conscious or subconscious) to avoid these inappropriate connotations in the settings concerned. 22. 23. 24. 25. 26. 27. 28. 29. 30.

‫ קאלדרא‬koldre ‘bedspread’ (Landau 1892: 21) ‫ קלאמק''ע‬klyamke ‘doorknob’ (Bromberg 1899: 42) ‫]הענגלאחטער‬. . .[ [. . .]henglaykhter ‘chandelier’ (Lieberson 1913: 46) ‫ שיפלעדיל‬shufledl ‘little drawer’ (Sofer 1904: 26) ‫]שאנק‬. . .[ [. . .]shank ‘wardrobe’ (Rapaport 1909: 22) ‫ שפיגעל‬shpigl ‘mirror’ (Munk 1898: 74) ‫ שראנק‬shrank ‘closet’; ‘cupboard’ (Brandwein 1912: 23) ‫ לאמפין‬lompn ‘lamps’ (N. Duner 1912: 37); cf. Hebrew ‫ מנורות‬mənorot ‫ בענקיל‬benkl ‘chair’ (Sobelman 1909/10 pts. 1–2: 15); cf. Hebrew ‫ כסא‬kisseʾ

2.4 Law Enforcement While Hebrew terms are typically employed in reference to specifically Jewish legal concepts, e.g. ‫ בית דין‬bet din ‘(Jewish) law court’, Yiddish vocabulary is typically employed with reference to non-Jewish forms of policing and law enforcement. In these cases the motivation for selecting Yiddish terms is straightforward: the concepts in question derive from and describe aspects of non-Jewish society, and as such the authors most likely chose them because the closest Hebrew equivalents seemed inappropriate or irrelevant. 31. 32. 33. 34. 35. 36. 37.

‫ אדוויקאט‬advokat ‘lawyer’ (Munk 1898: 21) ‫ זונדארין‬zhundarin ‘policemen’ (Munk 1898: 30) ‫ פאליצייא‬politsey ‘police’ (Rodkinsohn 1864: 18) ‫]פראצעס‬. . .[ [. . .]protses ‘court case’ (Michelsohn 1912: 102) ‫ ריכטער‬rikhter ‘judge’ (Sofer 1904: 26) ‫ קינע‬kune ‘pillory’ (Kaidaner 1875: 42a) ‫]קלוצוויט‬. . .[ [. . .]klyutshvoyt ‘rural police commissioner’ (Kaidaner 1875: 16b)

256

Kahn

2.5 Materials Yiddish terms are often used to designate various types of materials, ranging from different sorts of cloth to ceramics and metals. Like the law enforcement vocabulary discussed above, these Yiddish borrowings were most likely selected because the items in question lacked well-known Hebrew designations. 38. 39. 40. 41. 42. 43. 44. 45. 46.

‫ בערשטינן‬burshtinen ‘amber’ (Sobelman 1909/10 pts. 1–2: 34) ‫]בירלאנטין‬. . .[ [. . .]brilyantn ‘diamonds’ (Sofer 1904: 30) ‫]גיוואנט‬. . .[ [. . .]gevant ‘fabric’ (Gemen 1914: 60) ‫ ַװאט‬vat(e) ‘cotton wool’ (Landau 1892: 21) ‫]זילבער גלוייט‬. . .[ [. . .]zilber geloyt ‘silver solder’ (Kaidaner 1875: 41a) ‫ לייווינט‬layvnt ‘canvas’ (Sofer 1904: 12) ‫ )מאגנעט =( מאגנעס‬magnet ‘magnet’ (Munk 1898: 77) ‫ קאחיל‬kakhil ‘tile’ (Rodkinsohn 1864: 9) ‫ קרייד‬krayd ‘chalk’ (J. Duner 1899: 48)

2.6 Measurements The Hasidic Hebrew authors typically employ Yiddish loanwords to indicate measurements, most commonly of distance, as in (47)–(52), but sometimes of time, as in (53). As in many of the categories discussed above, the use of Yiddish in this type of context seems to be the result of the lack of Hebrew terms for practical, everyday concepts at the time of composition. 47. 48. 49. 50. 51. 52. 53.

‫וויארסט‬/‫ ווירסט‬virst/vyorst ‘verst’4 (Heilmann 1903: 99) ‫ לויט‬loyt ‘half-ounce’ (Kaidaner 1875: 41a) ‫ ליטרא‬liter ‘litre’ (Walden 1912: 79) ‫ מיל‬mil ‘mile’ (Bodek 1866: 5) ‫ קווארט‬kvort ‘quart’ (Rakats 1912 pt. 1: 48) ‫ קלאפטער‬klafter ‘fathom’ (Ehrmann 1911: 10b) ‫ מינוט‬minut ‘minute’ (Kaidaner 1875: 40b)

2.7 Medicine Yiddish lexical items are often used to convey everyday medical concepts. As in the case of many other semantic categories discussed previously, this tendency is typically due to a lack of standard Hebrew equivalents. However, in some cases the authors employ a Yiddish term despite the existence of a Hebrew counterpart. This is illustrated in (57), in which the Yiddish ‫דאקטער‬ dokter appears in addition to the Hebrew ‫ רופא‬rop̄ eʾ. Unusually, in contrast to 4  Archaic Russian measure of distance equivalent to 1.06 kilometres.

lexical borrowings in hasidic hebrew

257

the cases discussed above these Yiddish and Hebrew terms are used in free variation and the authors seem to have regarded them as synonyms without different semantic nuances. 54. 55. 56. 57. 58. 59. 60. 61.

‫]אפטייק‬. . .[ [. . .]apteyk ‘chemist, pharmacy’ (Hirsch 1900: 46) ‫ אפעראציע‬operatsye ‘operation’ (Landau 1892: 56; Michelsohn 1910: 190) ‫ באנקיס‬bankes ‘cupping’ (Zak 1912: 21) ‫ דאקטער‬dokter ‘doctor’ (Sofer 1904: 12); cf. ‫]רופא‬. . .[ [. . .] rop̄ eʾ ‘doctor’

(Lieberson 1913: 39) ‫]פיאווקיס‬. . .[ [. . .]pyavkes ‘leeches’ (Zak 1912: 21) ‫]קליניק‬. . .[[. . .]klinik ‘clinic’ (Ehrmann 1903: 6b) ‫ רעצעפט‬retsept ‘perscription’ (Berger 1910: 20) ‫]שפיטאל‬. . .[ [. . .]shpitol ‘hospital’ (Munk 1898: 36)

2.8 Occupations Yiddish terms are often used to designate occupations and professions common in the tales’ settings. As in many of the categories discussed above, most of these terms lacked obvious Hebrew labels at the time of writing. However, some of them, e.g., that shown in (68), do have equivalents in well-known earlier Hebrew texts. As in many cases discussed above, it is possible that the authors selected the Yiddish terms because they (perhaps subconsciously) identified them strongly with the Eastern European context of the tales, in contrast to their Hebrew counterparts which may have lacked such associations. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71.

‫ בוכהאלטער‬bukhhalter ‘bookkeeper’ (Ehrmann 1905: 36) ‫ אפטייקער‬apteyker ‘pharmacist’ (Hirsch 1900: 81) ‫ וואנער‬vayner ‘wine merchant’ (Munk 1898: 40) ‫ זייגער מאכער‬zeygermakher ‘watchmaker’ (Rakats 1912 pt. 2: 20) ‫ טרעגער‬treger ‘porter’ (Lieberson 1913: 43) ‫]מילנער‬. . .[[. . .]milner ‘miller’ (Berger 1910: 50) ‫]קעך‬. . .[[. . .]kekh ‘cook’ (Singer 1900: 16); cf. Hebrew ‫ טבח‬ṭabbaḥ ‫ קרעטשמער‬kretshmer ‘innkeeper’ (Ehrmann 1903: 7b) ‫ שווייגער‬shvayger ‘milkman’ (Bromberg 1899: 32) ‫ שיפסקאפיטאן‬shifskapitan ‘ship’s captain’ (Sofer 1904: 29)

2.9 Eastern European Jewish Religion and Culture The Hasidic Hebrew authors refer to certain Ashkenazi cultural and religious practices and concepts by their Yiddish designations. In this case, the use of Yiddish is most likely rooted in the fact that many of the practices concerned are folk customs which would have been most familiar to the authors and

258

Kahn

audience in these forms (even if in some cases Hebrew equivalents may technically have existed). This area is one of the few deviations from the pattern suggested above whereby Yiddish terms serve to denote concrete objects rather than abstract concepts: the loanwords shown below include abstract nouns, e.g. (79), and adjectives, e.g. (75) and (81). 72. 73. 74. 75. 76. 77. 78. 79.

‫ אונטער פירר‬unterfirer ‘wedding usher’ (Sobelman 1909/10 pt. 3: 16) ‫ באדעקונס‬badekns ‘veiling of the bride’ (Munk 1898: 36) ‫ גוטער יוד‬guter yid ‘Hasidic rebbe’ (J. Duner 1899: 39) ‫ גלאט כשר‬glat kosher ‘glatt kosher’5 (Sofer 1904: 24; Rapaport 1909: 29) ‫דריידיל‬ ֵ dreydl ‘dreidel’6 (Landau 1892: 35) ‫ וואכנאכט‬vakhnakht ‘watch night’7 (Berger 1910: 137) ‫ יארצייט‬yortsayt ‘anniversary of a death’ (Gemen 1914: 87) ‫ יודישקייט‬yidishkayt ‘Jewishness’; ‘Jewish religious culture’ (Rakats 1912

pt. 1: 31) 80. ‫ פאר=שפיל‬forshpil ‘pre-wedding dance party’ (Zak 1912: 136) 81. ‫ פרום‬frum ‘religiously observant’ (Bromberg 1899: 8; Heilmann 1902: 87)

As this analysis has illustrated, Yiddish loanwords are a prominent feature of Hasidic Hebrew lexis and are not restricted to a small range of semantic domains but rather are distributed through a wide variety of categories. Nevertheless, they most commonly serve to designate practical, everyday concepts, particularly concrete objects. In many cases this is due to the lack of established Hebrew terms for such concepts and objects; by contrast, in certain instances the authors select Yiddish borrowings despite the existence of a Hebrew equivalent, often because of the culturally specific or quotidian resonances of the Yiddish terms in question. 3 Aramaic Like their Yiddish counterparts, Aramaic loanwords appearing in the Hasidic Hebrew tale corpus consist primarily of nouns rather than other parts of speech and are fully incorporated into Hebrew grammatical patterns. However, Aramaic loanwords differ significantly from their Yiddish equivalents in 5  Conforming strictly to the Jewish dietary laws. 6  Spinning top played at the festival of Hanukkah. 7  The night before the circumcision ceremony when a vigil is kept over the baby.

lexical borrowings in hasidic hebrew

259

three key ways. Firstly, they are drawn from a much more limited—and very dissimilar—semantic range, being typically restricted to Jewish religious, legal, philosophical, and mystical concepts. Secondly, they usually designate abstract notions rather than physical objects. Thirdly, they most commonly consist of noun phrases rather than individual nouns, typically possessive constructions, two nouns linked by the conjunction waw, and noun-adjective syntagms. These Aramaic borrowings are almost always traceable to sources such as the Talmud that would have been very familiar to the tale authors. Many of these terms were also used independently in medieval forms of Hebrew and in Yiddish. This suggests that Aramaic was not an extremely well-integrated or productive element of the authors’ linguistic repertoire, but rather that they inserted certain common words and set phrases into their work based on their familiarity with these elements from their frequent appearance in Aramaic and Hebrew texts and from their own vernacular. Indeed, the Hasidic Hebrew authors’ use of Aramaic does not seem to have been disproportionately high when compared with Yiddish or with well-known written forms of Hebrew. Moreover, given the intimate perceived relationship between Hebrew and Aramaic in the authors’ native Yiddish (wherein both languages are referred to by the single designation ‫קודש‬-‫ לשון‬loshn-koydesh ‘the holy tongue’) and the close links between study of written Hebrew and Aramaic in Talmudic academies, it is possible that they did not really regard these Aramaisms as loanwords at all. Finally, it is significant that the majority of these nouns and phrases were subsequently absorbed into the lexis of Modern Hebrew, particularly in the higher registers (see Even-Shoshan 2003, which lists many of the Aramaisms to be discussed below). Aramaic thus appears to constitute a relatively superficial component of the Hasidic Hebrew corpus in comparison to the deeply entrenched and pervasive Yiddish. The Aramaic nouns and noun phrases appearing in the tale corpus can be divided into the few semantic categories outlined below. 3.1 Jewish Religious Culture A large proportion of the Aramaic lexical items appearing in the tale corpus belong to the specific domain of Jewish religious culture, including texts, rabbinic institutions, education, and legal concepts. In some cases, as in Yiddish, the Aramaic form is employed due to the lack of a Hebrew equivalent, e.g. (82) ‫ טפחא‬ṭip̄ ḥā⁠ ‘tipha (a cantillation note)’. In most others, the Aramaic terms are widespread and familiar features of rabbinic literature and therefore the authors’ selection of them is not surprising despite the existence of Hebrew alternatives.

260 82. 83. 84. 85. 86. 87. 88. 89. 90. 91.

Kahn

‫ טפחא‬ṭip̄ ḥā ‘tipḥa’ (Rodkinsohn 1864: 22) ‫ אורייתא‬ʾoraytā⁠ ‘Torah’ (Kaidaner 1875: 37b) ‫ ספיקא‬səfeqā⁠ ‘suspicion (in Jewish law)’ (Bodek 1866: 56) ‫ יומא דהלולא‬yomā⁠ dəhillulā⁠ ‘holiday’ (Shenkel 1903: 3) ‫ מרא דאתרא‬mārā⁠ dəʾatrā⁠ ‘local rabbinic authority’ (Lieberson 1913: 44) ‫ דינא‬dinā⁠ ‘law’ (Sobelman 1909/10 pts. 1–2: 15) ‫]איסורא‬. . .[[. . .] ʾissurā⁠ ‘prohibition (in Jewish law)’ (Ehrmann 1903: 15b) ‫]פלפולא דאורייתא‬. . .[ [. . .] pilpulā⁠ dəʾoraytā⁠ ‘Torah debate’ (Zak 1912: 13) ‫ חומרא‬ḥumrā⁠ ‘strictness (in Jewish law)’ (Landau 1892: 15) ‫ ריש דוכנא‬reš duḵnā⁠ ‘teaching assistant’8 (Chikernik 1903b: 4)

Within the broad category of Jewish religious culture, Aramaisms are often used with reference to Jewish eschatological and messianic themes, as in the following examples. This is most likely due to the fact that such topics are frequently discussed in the Babylonian Talmud, which was composed largely in Aramaic. 92. ‫ במתיבתא דרקיע‬bimətiḇtā⁠ dirəqiaʿ ‘in the heavenly academy’ (Bodek 1866: 28) 93. ‫ עלמה דשקרא‬ʿālmā dəšiqrā⁠ ‘this world (lit: the world of lies)’ (Ehrmann 1903: 22b) 94. ‫]עלמא דקשוט‬. . .[ [. . .]ʿālmā⁠ diqəšoṭ ‘the world to come (lit: the world of truth)’ (Stamm 1905: 35) 95. ‫ עקבתא דמשיחא‬ʿiqḇātā⁠ diməšiḥā⁠ ‘sign (lit: footsteps) of the Messiah’ (Walden 1912: 46) 96. ‫ שכינתא בגלותא‬šəḵintā baḡluṯā ‘the Divine Presence in exile’ (Breitstein 1914: 51) Similarly, many Aramaic borrowings belong specifically to the domains of Jewish mysticism and Hasidism, as shown below. This is logical given that the chief work of Jewish mysticism, the Zohar, which has a prominent place in Hasidic philosophy, is written in Aramaic. 97. 98. 99. 100.

‫ גושפנקא דמלכה‬gušpanqā dəmalkā ‘the seal of the king’9 (Munk 1898: 6) ‫ סטרא דקדושה‬siṭrā diqədušā ‘the side of holiness’10 (J. Duner 1899: 63) ‫ מלכא דשידי‬malkā dəšede ‘king of demons’ (Shenkel 1903: 11) ‫ חסידא קדישא‬ḥasidā qaddišā ‘holy Hasid’ (Kaidaner 1875: 47a)

8   In the ḥeder, traditional Jewish school. 9   A concept in Jewish numerology. 10  A concept in Jewish mysticism.

lexical borrowings in hasidic hebrew

261

101. ‫ חסידא ופרישא‬ḥasidā up̄ arišā ‘Hasid and excellent one’ (Berger 1906: 22) 102. ‫ מילי דחסידותא‬mille deḥasidutā ‘words of Hasidism’ (Gemen 1914: 57) 3.2 Uncategorized While most of the Aramaic lexical borrowings appearing in the Hasidic Hebrew corpus clearly belong to the specialized semantic fields discussed above, occasionally uncategorized nouns and noun phrases do appear. However, in keeping with the rest of the Aramaisms attested in the tales, such borrowings typically denote abstract concepts rather than concrete items. They are typically derived from rabbinic literature in Aramaic, particularly the Babylonian Talmud. The Hasidic Hebrew authors do not seem to have had a clear motivation for selecting these Aramaic forms, as they generally have Hebrew equivalents and are not used in particular semantic or syntactic contexts. As suggested above, it is possible that the authors did not perceive them as Aramaisms as such but rather regarded the nouns as straightforward synonyms of their Hebrew counterparts and the noun phrases simply as fixed expressions. Semantically uncategorized nouns and noun phrases include the following: 103. 104. 105. 106. 107. 108.

‫ יומא דשוקא‬yomā dəšuqā ‘market day’ (Greenwald 1899: 51a) ‫ גופא דעובדא‬gup̄ ā dəʿuḇdā ‘the heart of the matter’ (Bodek 1866: 5) ‫ בעינא פקיחא‬bəʿenā pəqiḥā ‘with a sharp eye’ (Teomim Fraenkel 1911: 6) ‫ בצוותא חדא‬bəṣawtā ḥadā ‘together’ (Kaidaner 1875: 28b) ‫]מלתא דבדיחותא‬. . .[ [. . .]millətā diḇədiḥutā ‘joke’ (Walden 1913 pt. 2: 29) ‫ צפרא טבא‬ṣap̄ rā ṭāḇā ‘good morning’ (Singer 1900: 11)

The above examination has illustrated that the Aramaic lexical borrowings appearing in the Hasidic Hebrew tale corpus are much more restricted in scope and number than their Yiddish counterparts, typically deriving from only a few semantic domains most often relating to Jewish religious, legal, and mystical abstract concepts. This is in keeping with the most prominent sources of these Aramaic loanwords, namely the Talmuds, the Zohar, and other mystical writings. 4

Slavic Languages

While the Hasidic Hebrew tales contain a significant proportion of Yiddish vocabulary, and a not insignificant amount of Aramaic, they are almost entirely devoid of loanwords deriving from the Slavic languages in whose territory their authors lived. The corpus contains only the smallest handful of borrowings stemming directly from a Slavic language (rather than via the

262

Kahn

Slavic component of Yiddish). Slavic borrowings in the tales are invariably written in the Hebrew alphabet; however, in contrast to Yiddish and Aramaic loanwords they are usually overtly marked as foreign elements through labels and orthographic techniques, as will be seen below. The example in (109) (marked in bold) is a gloss that is explicitly designated by the author as a Russian term and appears with vocalization. This is unusual because vocalization is not usually employed in Hasidic Hebrew narrative literature, and as such its use here serves to highlight the word as a foreign term that may not otherwise be immediately recognizable to readers. The gloss is presented as synonymous with the Yiddish word ‫ פורטקאלין‬protokoln ‘protocols, record’ but derives from the Russian noun вопрос vopros ‘question, enquiry’. Interestingly, it appears with the Yiddish plural suffix ‫ין‬- -in (‫ן‬- -n in Standard Yiddish orthography) instead of with the Russian plural suffix –ы -y, despite the fact that this word is not employed in Yiddish. This may point to the author’s unfamiliarity with Russian grammatical patterns. The example in (110) contains two Ukrainian or Polish pronouns (again, marked in bold) that are used in combination with Hebrew nouns to form a rhyming phrase. As in (109), the fact that they are vocalized serves to highlight their foreign status. Interestingly, this case deviates from the tendency seen in the Yiddish and Aramaic sections above for the authors to borrow nouns and noun phrases rather than function words. 109. ‫אסין‬ ִ ‫אפ ָר‬ ְ ‫ואח"כ קבלו בכל פעם פורטקאלין הניקרא בלשון רוסיʹ ַװ‬ və-ʾaḥ”k (=ʾaḥar kaḵ) qibbəlu bəḵol paʿam protokaln han-niqra⁠ʾ biləšon rusi’ vaprosin And afterwards every time they would receive records, which are called in the Russian language voprosn (Rodkinsohn 1864: 7) Russian вопрос vopros ‘question’; ‘enquiry’ 110. ‫ טאַקי ליל שמורים‬,‫יאַקי פורים‬ yaqi purim, ṭaqi lel šəmurim As Purim is, so the first night of Passover will be (Michelsohn 1912: 139) Ukrainian який jakij; Polish jaki ‘which’; ‘what kind of’ Ukrainian такий takij, Polish taki ‘such a’ 5

Other Languages

Like Slavic loanwords, lexical borrowings from more distant languages are almost nonexistent in the Hasidic Hebrew tales. On a very few occasions an isolated German word or phrase appears in quotation marks within the

lexical borrowings in hasidic hebrew

263

Hebrew text, as in (111) (in bold). More strikingly, one Romance gloss (of uncertain provenance; possibly Italian, Spanish, or Judeo-Spanish) is attested in the corpus, shown in (112) (in bold). This is introduced by the designation ‫בלע"ז‬ bəlaʿaz ‘in Romance/a foreign tongue’. Again like the Slavic lexical items shown above, and in contrast to the numerous Yiddish and Aramaic borrowings, these rare items are explicitly marked out as foreignisms through the use of quotation marks and explanatory labels. The authors’ motivation for including these particular glosses is unclear: the immediate context does not logically call for them; they do not seem to reflect the actual words of a speaker; and they do not appear to aid understanding. 111. ‫לבשל אפילו מעט תפוחי אדמה „איין ביסכען ערדעפפֿעלʺ להחיות את נפשך‬ ləḇaššel ʾap̄ illu məʿat tappuḥe ʾadama “ein bißchen Erdäpfel” ləhaḥayot ʾet nap̄ šəḵa ‘[. . .] To cook even a bit of potatoes “ein bißchen Erdäpfel” to revive you [. . .]’ (Ehrmann 1903: 17b) German ein bißchen Erdäpfel ‘a bit of potatoes’ 112. ‫ולבש עצמו בלבוש מכוסה במראות (שפיגעל) אספיגו בלע"ז‬ velaḇaš ʿaṣmo biləḇuš məḵusse bəmarʾot (shpigl) espego bəlaʿaz And he dressed himself in a garment covered in mirrors (shpigl), espego in La’az (Munk 1898: 74) Cf. Italian specchio and Spanish or Judeo-Spanish espejo ‘mirror’ 6 Conclusion This paper has demonstrated that the lexical borrowings attested in the Hasidic Hebrew corpus exhibit three key patterns. Firstly, Yiddish loanwords are extremely widespread and most frequently constitute nouns denoting concrete objects; these are distributed over an extensive range of semantic fields associated with everyday life. Secondly, Aramaic loanwords are comparatively restricted in number and in semantic range and, in contrast to their Yiddish counterparts, typically constitute nouns and noun phrases designating abstract concepts from the theological, legal, and mystical domains. Thirdly, the corpus is almost entirely devoid of Slavic and other loanwords; moreover, on the rare occasions when these do appear they are generally presented explicitly as foreign elements, in contrast to the Yiddish and Aramaic borrowings. These findings illustrate three significant broader points about the lexical trends and sociolinguistic conditions that contributed to the tales’ compo­ sition. Firstly, the authors’ non-Semitic vernacular seems to have played

264

Kahn

a much more prominent role in the enrichment of their Hebrew lexis than their linguistically more similar Aramaic textual repertoire did. Secondly, the authors associated Yiddish primarily with the tangible and practical while they linked Aramaic chiefly with the conceptual and theological. These tendencies are themselves in keeping with the two languages’ wider roles within Eastern European Jewish society. Finally, the authors appear to have had a striking lack of direct lexical engagement with their Slavic-speaking host societies, and an even more marginal relationship with languages from further afield. References

Primary

Berger, Israel. 1906. ‫[ עשר קדושות‬Ten holinesses]. Piotrkow. Repr. in ‫ספרים קדושים‬ ‫[ מתלמידי בעל שם טוב הק׳ זצלה“ה‬Holy books from the students of the holy Ba’al Shem Tov of everlasting memory], vol. 97, New York, 1996. ———. 1910. ‫[ עשר צחצחות‬Ten radiances]. Piotrkow. Repr. in ‫ספרים קדושים‬ ‫[ מתלמידי בעל שם טוב הק׳ זצלה“ה‬Holy books from the students of the holy Ba’al Shem Tov of everlasting memory], vol. 97, New York, 1996. Bodek, Menahem Mendel. 1865. ‫[ קהל קדושים‬The assembly of the holy]. Lemberg. Repr. in ‫[ ספרים קדושים מתלמידי בעל שם טוב הק׳ זצלה“ה‬Holy books from the students of the holy Ba’al Shem Tov of everlasting memory], vol. 27, New York, 1985. ———. 1866. ‫[ מפעלות הצדיקים‬The works of the righteous]. Lemberg. Brandwein, Eliezer. 1912. ‫[ דגל מחנה יהודה‬The standard of the camp of Judah]. Lemberg. Repr. in ‫[ ספרים קדושים מתלמידי בעל שם טוב הק׳ זצלה“ה‬Holy books from the students of the holy Ba’al Shem Tov of everlasting memory], vol. 31, New York, 1985. Breitstein, Solomon Zalman. 1914. ‫[ שיחות חיים‬Talks of life]. Piotrkow. Bromberg, Israel Moses. 1899. ‫[ תולדות הנפלאות‬The chronicle of wonders]. Warsaw. Repr. in ‫[ ספרים קדושים מתלמידי בעל שם טוב הק׳ זצלה“ה‬Holy books from the students of the holy Ba’al Shem Tov of everlasting memory], vol. 3, New York, 1984. Chikernik, Isaiah Wolf. 1903a. ‫[ ספורים ומאמרים יקרים‬Precious stories and writings]. Warsaw. ———. 1903b. ‫[ ספורים נחמדים‬Pleasant stories]. Zhitomir. Duner, Jacob Shalom HaKohen. 1899. ‫[ דרך האמונה ומעשה רב‬The way of faith and story of the Rebbe]. Warsaw. Duner, Nathan Neta. 1912. ‫[ בוצינא קדישא‬The holy enlightened one]. Piotrkow. Ehrmann, Dov Baer. 1903. ‫[ דברים ערבים‬Pleasant words]. Part 1. Munkacs. Repr. in ‫[ ספרים קדושים מתלמידי בעל שם טוב הק׳ זצלה“ה‬Holy books from the students of the holy Ba’al Shem Tov of everlasting memory], vol. 8, New York, 1983. ———. 1905. ‫[ דברים ערבים‬Pleasant words]. Part 2. Munkacs.

lexical borrowings in hasidic hebrew

265

Gemen, Eliezer Dov. 1914. ‫[ ספרן של צדיקים‬The book of the righteous]. Warsaw. Repr. in ‫[ ספרים קדושים מתלמידי בעל שם טוב הק׳ זצלה“ה‬Holy books from the students of the holy Ba’al Shem Tov of everlasting memory], vol. 68, New York, 1988. Greenwald, Yehiel Moses. 1899. ‫[ ליקוטים חדשים‬New selections]. Warsaw. Heilmann, Hayim Meir. 1902. ‫[ בית רבי‬The house of the Rebbe]. Berdichev. Hirsch, Isaac Dov. 1900. ‫[ אמונת צדיקים‬Faith of the righteous]. Warsaw. Repr. in ‫[ ספרים קדושים מתלמידי בעל שם טוב הק׳ זצלה“ה‬Holy books from the students of the holy Ba’al Shem Tov of everlasting memory], vol. 42, New York, 1985. Kaidaner, Jacob. 1875. ‫[ סיפורים נוראים‬Wondrous tales]. Lemberg. Repr. in ‫[ ספרים קדושים מתלמידי בעל שם טוב הק׳ זצלה“ה‬Holy books from the students of the holy Ba’al Shem Tov of everlasting memory], vol. 3, New York, 1981. Landau, Isaac. 1892. ‫[ זכרון טוב‬A good memory]. Piotrkow. Repr. in ‫ספרים‬ ‫[ קדושים מתלמידי בעל שם טוב הק׳ זצלה“ה‬Holy books from the students of the holy Ba’al Shem Tov of everlasting memory], vol. 3, New York, 1984. Lieberson, Hayim. 1913. ‫[ צרור החיים‬Eternal life]. Bilgoray. Repr. in ‫ספרים‬ ‫[ קדושים מתלמידי בעל שם טוב הק׳ זצלה“ה‬Holy books from the students of the holy Ba’al Shem Tov of everlasting memory], vol. 7, New York, 1983. Michelsohn, Abraham Hayim Simhah Bunem. 1910. ‫[ דובר שלום‬Speaker of peace]. Przemysl. Repr. in ‫[ ספרים קדושים מתלמידי בעל שם טוב הק׳ זצלה“ה‬Holy books from the students of the holy Ba’al Shem Tov of everlasting memory], vol. 30, New York, 1985. ———. 1912. ‫[ מקור חיים‬The source of life]. Bilgoray. Repr. in ‫ספרים קדושים‬ ‫[ מתלמידי בעל שם טוב הק׳ זצלה“ה‬Holy books from the students of the holy Ba’al Shem Tov of everlasting memory], vol. 30, New York, 1985. Munk, Faivel. 1898. ‫[ שיחות צדיקים‬Talks of righteous ones]. Warsaw. Repr. in ‫[ ספרים קדושים מתלמידי בעל שם טוב הק׳ זצלה“ה‬Holy books from the students of the holy Ba’al Shem Tov of everlasting memory], vol. 27, New York, 1985. Rakats, Yo’ets Kim Kadish. 1912. ‫[ תפארת היהודי‬The glory of the Jew]. 2 parts. Piotrkow. Repr. in ‫[ ספרים קדושים מתלמידי בעל שם טוב הק׳ זצלה“ה‬Holy books from the students of the holy Ba’al Shem Tov of everlasting memory], vol. 3, New York, 1984. Rapaport, Solomon Zalman. 1909. ‫[ ויקהל שלמה‬And Solomon gathered]. Piotrkow. Rodkinsohn, Michael Levi Frumkin. 1864. ‫[ שבחי הרב‬In praise of the Rebbe]. Lemberg. Repr. in ‫[ ספרים קדושים מתלמידי בעל שם טוב הק׳ זצלה“ה‬Holy books from the students of the holy Ba’al Shem Tov of everlasting memory], vol. 27, New York, 1985. ———. 1865. ‫[ עדת צדיקים‬The congregation of the righteous]. Lemberg. Singer, Isaac. 1900. ‫[ שבע רצון‬Satisfaction]. Podgorze. Repr. in ‫ספרים קדושים‬ ‫[ מתלמידי בעל שם טוב הק׳ זצלה“ה‬Holy books from the students of the holy Ba’al Shem Tov of everlasting memory], vol. 31, New York, 1985. Sobelman, Abraham Isaac. 1909/10. ‫[ סיפורי צדיקים החדש‬New stories of righteous men]. 3 parts. Piotrkow. Repr. in ‫ספרים קדושים מתלמידי בעל שם טוב הק׳ זצלה“ה‬

266

Kahn

[Holy books from the students of the holy Ba’al Shem Tov of everlasting memory], vol. 4, New York, 1982. Sofer, Jacob. 1904. ‫[ ספורי יעקב‬Stories of Jacob]. Husyatin. Repr. in ‫ספרים קדושים‬ ‫[ מתלמידי בעל שם טוב הק׳ זצלה“ה‬Holy books from the students of the holy Ba’al Shem Tov of everlasting memory], vol. 17, New York, 1984. Stamm, Shalom Elijah. 1905. ‫[ זכר צדיק‬The memory of a righteous man]. Vilna. Repr. in ‫[ ספרים קדושים מתלמידי בעל שם טוב הק׳ זצלה“ה‬Holy books from the students of the holy Ba’al Shem Tov of everlasting memory], vol. 35, New York, 1986. Teomim Fraenkel, Judah Aryeh. 1911. ‫[ אהלי שם‬Tents of renown]. Bilgoray. Repr. in ‫[ ספרים קדושים מתלמידי בעל שם טוב הק׳ זצלה“ה‬Holy books from the students of the holy Ba’al Shem Tov of everlasting memory], vol. 17, New York, 1984. Walden, Moses Menahem. 1914. ‫[ אהל יצחק‬Isaac’s tent]. Piotrkow. Repr. in ‫[ ספרים קדושים מתלמידי בעל שם טוב הק׳ זצלה“ה‬Holy books from the students of the holy Ba’al Shem Tov of everlasting memory], vol. 11, New York, 1984. Zak, Reuben. 1912. ‫[ בית ישראל‬The house of Israel]. Piotrkow. Repr. in ‫ספרים קדושים‬ ‫[ מתלמידי בעל שם טוב הק׳ זצלה“ה‬Holy books from the students of the holy Ba’al Shem Tov of everlasting memory], vol. 5, New York, 1983.

Secondary

Betzer, Tzvi. 2001. ,‫ החטיבה הביניימית‬:‫ החטיבה השנייה‬,‫פרקים בתולדות הלשון העברית‬ ‫הביניים‬-‫ הלשון הרבנית של ימי‬:7 ‫[ יחידה‬History of the Hebrew language: The medieval division, unit 7: Rabbinic Hebrew]. Tel Aviv. Dvir-Goldberg, Rivka. 2003. ‫ עיון בסיפורי מעשיות מפי צדיקים‬:‫הצדיק החסידי וארמון הלויתן‬ [The zaddik and the palace of Leviathan: A study of Hassidic tales told by zaddikim]. Tel Aviv. Even-Shoshan, Abraham (ed. Moshe Azar, Ilana Shamir, and Ya’el Yannai). 2003. ‫ מחודש ומעודכן לשנות האלפיים‬:‫שושן‬-‫[ מילון אבן‬The Even-Shoshan dictionary: Revised and updated for the 21st century], 6 vols. Israel. Kahn, Lily. 2015. A grammar of the Eastern European Hasidic Hebrew tale. Leiden. Matras, Yaron. 2009. Language contact. Cambridge.

Possible Ugaritic Influences on the Hurrian of Ras Shamra-Ugarit in Alphabetic Script Joseph Lam The interaction between Ugaritic and Hurrian in the Late Bronze Age city of Ugarit (modern Ras Shamra in Syria) represents a fascinating example of language contact from the ancient Near East. Ugaritic and Akkadian are the two languages that dominate the textual record at the site, with Ugaritic (approximately 2000 texts) being the primary spoken vernacular of the ancient city and Akkadian (approximately 2500 texts) functioning as the diplomatic lingua franca of the period (Bordreuil and Pardee 2009: 8). However, among the remaining six languages that are known from the Late Bronze Age at Ras Shamra (Sumerian, Hittite, Hurrian, Luwian, Cypro-Minoan, and Egyptian), Hurrian stands out as the one most likely to have represented a spoken language for a significant minority segment of Ugaritic society.1 A number of factors point in this direction. First, Hurrian is the only language other than Ugaritic (and Akkadian) to have been inscribed in the locally-invented alphabetic cuneiform script.2 In fact, unlike Ugaritic (mostly alphabetic cuneiform) and Akkadian (mostly logosyllabic cuneiform), Hurrian is more evenly attested in the two major scripts used at the site, with texts in each of the scripts numbering in the dozens; their distribution reflects the particular importance of Hurrian in hymnic and ritual contexts.3 In addition, a significant number 1  For a useful survey of languages and scripts attested at Ugarit, see Malbran-Labat (1999). 2  As Juan-Pablo Vita (1999: 457) observes, “the only successful adaptation of the Ugaritic alphabet to another language, in a significant way, was to Hurrian,” notwithstanding the several isolated examples of Akkadian written in alphabetic script. On the invention of the Ugaritic alphabetic cuneiform system, see Pardee (2007). 3  The syllabic material includes the multilingual vocabulary (Sa) texts that contain a Hurrian column (Huehnergard 1987: 22–23; van Soldt 1990: 728–729; André-Salvini and Salvini 1998; 1999), one Akkado-Hurrian “wisdom” text (RS 15.010), two letters (RS 11.853, RS 23.031 [unpublished]), and about twenty examples (counting the numerous fragments of RS 19.164 as a single text) of the so-called “musical” texts, which contain religious lyrics accompanied by notations for musical performance. The alphabetic Hurrian material consists of between one and two dozen texts of a ritual nature (the numerical uncertainty deriving from the difficulty of language identification for small fragments), plus five other ritual texts in mixed Ugaritic and Hurrian (see discussion to follow). For overviews of the Hurrian material from Ugarit, see Dietrich and Mayer (1999: 58–61) and Vita (2009: 219–220).

© koninklijke brill nv, leiden, ���5 | doi ��.��63/9789004300156_015

268

Lam

of Hurrian proper names4 and loanwords5 are known in the Ugaritic texts. These facts, along with evidence of Hurrian linguistic influence at nearby sites such as Alalakh, Emar, and Qatna, reflect the existence of a Hurrian culturallinguistic element in northern Syria in the latter part of the 2nd millennium BCE (Salvini 1995; Richter 2005). Last but not least, among the alphabetic material are five bilingual Ugaritic and Hurrian tablets,6 all of them ritual in nature, attesting to priest-scribes that were at least functionally bilingual and capable of inscribing Hurrian when a specific part of a ritual necessitated it. However, the contact situation between Ugaritic and Hurrian in Late Bronze Ras Shamra has been explored only to a limited extent. On the one hand, since the influence of Hurrian on Ugaritic seems restricted entirely to lexical borrowings, the most detailed study to date is the series of four articles by Wilfred Watson on “Non-Semitic Words in the Ugaritic Lexicon” (1995; 1996; 1998; 1999) in which he identifies approximately 80 possible Hurrian loanwords in Ugaritic. On the other hand, the nature of Ugaritic influence on Hurrian at Ras Shamra remains difficult to characterize, because of the general difficulty and paucity of the Hurrian texts from Ras Shamra.7 Limited observations have been made on the topic, though invariably as part of broader grammatical overviews, such as those of Bush (1964), Laroche (1968), and Dietrich and Mayer (1999). Most notably, Laroche (1968: 533) gave a brief paragraph describing the “style” of the 4  According to Hess (1999: 509), about 23% of the personal names from Ugarit (both syllabic and alphabetic) are Hurrian. Similarly, van Soldt (2003: 702), in a more focused study of select groups of persons at Ugarit, observed that Hurrian names comprise between 5 and 30% of the various lists. While acknowledging that the language of one’s name does not necessarily correspond to the language that one speaks, the numbers nonetheless confirm the importance of the Hurrian linguistic-cultural element at Late Bronze Ugarit. Earlier studies analyzing the Hurrian names at Ugarit include Gröndahl (1967) and Kinlaw (1967). 5  See Watson (1995; 1996; 1998; 1999) and discussion to follow. 6  R S 24.254, RS 24.255, RS 24.261, RS 24.291, and RS 24.643 (Ugaritic with a five-line Hurrian section); see Pardee 1996; Lam 2006. 7  Notwithstanding the considerable progress made in the study of Hurrian in the last two or three decades, especially after the discovery of the Hurro-Hittite bilingual materials at Boğazköy in the 1980’s, it remains one of the more poorly-understood languages of the ancient Near East. The Hurrian texts inscribed in the alphabetic cuneiform script pose unique challenges for interpretation. While consonantal writing is well suited to syllabic structure and the derivational patterns of a Semitic language (based on triliteral roots), it seems rather maladapted for Hurrian, with its agglutinative structure and great number of morphemes that differ only by a vowel or by the gemination of a consonant. The result is that much of the alphabetic Hurrian corpus remains difficult to interpret if not impenetrable to modern scholars.

Ugaritic Influences on the Hurrian of Ras Shamra-Ugarit

269

alphabetic Hurrian texts relative to the language of the Mittani Letter, while Dietrich and Mayer (1999: 68–69) offered a list of 24 possible loanwords from Semitic into Hurrian. Even if one does not accept all of their proposals, the extensiveness of the lexical borrowing is clear, given the paucity of textual evidence from which the examples are drawn. Against this background, the goal of this paper is modest: to highlight three possible cases of linguistic influence of Ugaritic on the Hurrian of Ras Shamra (focusing on the alphabetic material), and to consider their implications for the Ugaritic-Hurrian contact situation at Ras Shamra. The examples will be presented in decreasing order of plausibility; while the first two represent fairly clear examples of changes resulting from language contact, the third is more uncertain, illustrating the potential pitfalls involved in the identification of linguistic interference. Then, I will offer brief comments on the place of Hurrian in ancient Ugaritian society in light of this evidence. 1

Assimilation of /n/

The first example comes from the realm of phonology. Familiar to students of Semitic is the phenomenon of regressive (anticipatory) assimilation of /n/ when in contact with a following consonant, a conditioned sound change that is found in a number of Semitic languages including Hebrew and Akkadian. For instance, in Ugaritic: (1) {ytt} /yatattu/ < /*yatantu/ ‘I gave’ {a̓t} /’atta/ (written at-ta in syllabic transcription) < /*’anta/ ‘you’ In (standard) Hurrian, /n/ does not show regressive assimlation; the regular rule is that of progressive (lag) assimilation of /n/, particularly in two situations:

‧ When an enclitic =nna follows the ergative ending /ž/, the /nn/ shows par‧

tial assimilation to the voiceless counterpart /š/, i.e., -ž + =nna > -šša (Farber 1971) Between two instances of /n/, /r/, /l/, or the dentals, after the syncopation of a vowel between them, the second consonant assimilates (totally) to the first, e.g., *avari=ne > avarre (Wilhelm 2004: 100; Wegner 2007: 60)

In the alphabetic Hurrian of Ras Shamra, however, there are more than a few instances of regressive assimilation of /n/ as typical of Ugaritic phonology. Examples of this include:

270

Lam

‧ Comparing the examples {tzġ . ảrm . ttb . ṯu̓ ṯk . ḫnz˹r˺} (RS

24.643 [KTU 1.148] 17) and {tzġ . ảrm . ttp . [. . .] /[ṯ]ủṯk ḫzr} (RS 24.644 [KTU 1.149] 11), the latter spelling of {ḫzr} reflects the assimilation of an /n/, perhaps based on a form with the syllable structure /*ḫVnzVr/; it is uncertain whether the writing of {ḫnz˹r˺} implies phonological or merely orthographic variation (see Lam 2006: 409–410, esp. n. 40). The spelling of {pdgl} (RS 1.004 [KTU 1.42] 48) for pendigalli, i.e., dPé-en-tikal-li in KUB 27.13 I 20 (Laroche 1980: 200; 1968: 529). The form {mḏgd} (RS 1.004 [KTU 1.42] 39) for mužungi=da (‘order, arrangement’); compare mu-šu-un-ki (ChS I/9 119 I 7ʹ; ChS I/9 120 I 18) and mu-šuun-ki-ya (ChS I/9 4 III 10ʹ, 26ʹ; ChS I/9 5 III 11ʹ; ChS I/9 53 II 19) (see Richter 2012: 256, 532; Laroche 1980: 173). The form {i̓ḏr}, which occurs 19 times in RS 1.004 (KTU 1.42), for enžari (Lam 2011: 153–154): etymologically, this word consists of the noun eni (“god”) with the derivational morpheme =žari which produces collectives (Wegner 2007: 57). In the syllabic texts, this word is always found with evidence of the /n/ preserved (Laroche 1980: 80–82)—e.g., e-in!-ša-a-ri in VBoT 14, 10’ and e-enza-a-ri in KBo 32.14 Vs. I 36 (see Neu 1996: 139). The spelling with {ḏ} is also unexpected from a Hurrian phonological perspective: normally in Hurrian, in the case of a “doubled” letter, one expects the voiceless allophone /š/ to be represented (written Vš- šV in syllabic texts, and with {ṯ} in the alphabetic script). Thus, the form {i̓ḏr} looks like an original form *en=žari adapted to the phonological patterns of Ugaritic (i.e., /ežžari/ < /*enžari/).

‧ ‧ ‧

Given the clear contrast between the behavior of /n/ in standard Hurrian and in Ugaritic, and the way in which these examples conform to the Ugaritic pattern, they likely represent instances of phonological interference from Ugaritic on Hurrian. Moreover, in light of the paucity of the alphabetic Hurrian material at Ras Shamra, even the few examples cited here are not insignificant. In fact, there are probably additional instances of this phenomenon that have yet to be identified due to uncertainties in interpretation. 2

Genitival Phrases

The second example relates to both morphology and syntax. In Hurrian, genitival constructions involve the marking of the dependent noun with the postposition =ve; the dependent noun can precede or follow the head noun:

Ugaritic Influences on the Hurrian of Ras Shamra-Ugarit

271

(2) šēn(a) = iffu = ve ašti brother+ poss1SG+GEN wife+ØABS ‘the wife of my brother’ (Mitt. III 21) Moreover, when the noun phrase is in a case other than the absolutive (which has zero case ending), the phenomenon of Suffixaufnahme occurs in which the case markers on the head noun are reproduced on the dependent noun, attached to the relator =ne: (3) šēn(a) = iffu = ve = ne = va ašt(i)= ī =va brother+poss1SG+GEN+REL+DAT wife+poss3SG+DAT ‘for the wife of my brother’ (Mitt. II 6) This pattern is observable in the alphabetic Hurrian of Ras Shamra: (4) i̓[n] a̓mrw i̓n u̓ grtw eni amurri=ve eni ugaritta=ve god+ØABS Amurru+GEN god+ØABS Ugarit+GEN ‘mṯtmrw ‘ammiṯtamra=ve Ammistamru+GEN ‘the god of Amurru, the god of Ugarit, of Ammistamru’ (RS 24.274 [KTU 1.125] 6–7) But in a number of other instances, one finds a genitival phrase (in an oblique case) constructed without any genitive marker at all (see Dietrich and Mayer 1999: 67; Tropper and Vita 2003): (5) i̓n ṯlnd eni šelli=ne=da god (+?) household+REL+DIR (instead of en[i]=ne=da šelli[=ni]=ve=ne=da) ‘for “the-god-of-the-household” ’ (RS 24.254 [KTU 1.110] 1) (6) i̓n a̓tnd eni attan(i)=ne=da god (+?) father+REL+DIR ‘for “the-god-of-the-father” ’ (RS 24.254 [KTU 1.110] 2)

272

Lam

(7) i̓n a̓rdnd eni arde=ne=da god (+?) city+REL+DIR ‘for “the-god-of-the-city” ’ (RS 24.254 [KTU 1.110] 2) It appears that these examples represent a borrowing of the pattern of the Semitic construct phrase, with the head noun preceding the dependent noun, and understood as bound together without any explicit genitival marker (Lam 2006: 404 n. 21). Two points of elaboration are in order. First, in the Ugaritic construct phrase, the dependent noun would have been marked by a genitive case ending /i/ or /īma/, while the head noun would have retained the case ending necessitated by its syntactic function in the sentence. Unfortunately, the consonantal orthography does not give us any indication of whether these elements would have been present. Second, note that the interference here from Ugaritic is structural, affecting both morphology and syntax. Both words of the construct phrase are Hurrian, with the phrase representing a sort of loan-translation of a Ugaritic divine epithet; when this phrase is in an oblique case, the relator and postposition are placed after the entire phrase, not after the head noun.8 This suggests the construal of the entire syntactically-Semiticized divine epithet as a compound lexical unit. At the same time, the fact that the frozen element is the Hurrian rendering of a Ugaritic phrase as opposed to simply a borrowed Ugaritic epithet still suggests a deeper structural influence from Semitic on Hurrian grammar. 3

The Question of Enclitic -m(a)

The third example concerns the question of enclitic -m(a) and has actually been suggested sporadically in the literature as an explanation for individual forms in these texts. But this has usually been done in passing and in isolation, without a full explication of the dimensions of language contact that are potentially involved. Hurrian possesses a relatively rich set of enclitic particles and pronouns that can be attached to various forms. These include (Wilhelm 2004: 155; Wegner 2007: 117): 8  The addition of the relator (=ne) appears to be characteristic of all divine names in Hurrian that originated as epithets (e.g., Šimige=ne=, Kumarbi=ne=, etc. [Khachikyan 2005: 189]). Examples of this at Ras Shamra include the spellings {kmrwnd} (RS 24.254 [KTU 1.110] 5) and {ṯmgnd} (RS 24.254 [KTU 1.110] 7).

Ugaritic Influences on the Hurrian of Ras Shamra-Ugarit

273

‧ sentence particles (Mittani Letter): =an, =mān, =man, =mmaman, =nīn ‧ the two forms of the 2nd person sg. pronoun: =mma/=m ‧ the “Old Hurrian” connective =ma Ugaritic also has a number of enclitic particles, the most important of which are {-y}, {-n}, and {-m} (Tropper 2012: 823–835)—all three of these are attachable to any part of speech. The enclitic {-m}, in particular, has an extremely wide array of functions.9 In the alphabetic Hurrian corpus of Ras Shamra, the more complex forms of the Mittani Letter sentence particles (=mān, =man, =mmaman, =nīn) appear to be quite rare. Instead, one finds quite a few words spelled with just final {m} or {n}.10 For the forms with final {m}, the question arises as to whether these are 2.sg personal pronouns, examples of an Old Hurrian connective, or imitations of the Ugaritic enclitic particle {m} (representing /-ma/). In such cases, the determination of the function of the {m} can be difficult given the wideranging and overlapping uses of the relevant particles in the two languages. In particular, to establish definitively that a given {m} is a borrowing of the Ugaritic form, one would need to identify a contextual function that is encompassed by Ugaritic -ma but not characteristic of (or replaceable in context by) any of the corresponding Hurrian particles spelled with {m}. Dietrich and Mayer (1999: 67) have offered three suggestions for possible instances of enclitic -ma in the alphabetic Hurrian of Ras Shamra. Although they appear ambivalent as to linguistic origin,11 their suggestions are worth considering as possible instances of distinctly “Ugaritic-style” -ma. The first is the phrase {kby kbny mdm kt[y]} in RS 24.285 (KTU 1.131) 3–4, which they 9   Tropper (2012: 825–832) identifies the following nine major categories of functions: (i) to mark an adverbial; (ii) to mark a vocative; (iii) to emphasize various nominal expressions; (iv) attachment to a finite verbal form (precise function uncertain); (v) as a stylistic marker in poetry; (vi) to mark the second of a two-member nominal expression; (vii) with the meaning ‘also, yet, again, even’ on the second (or subsequent) member of a list; (viii) to mark the ellipsis of an expected word; (ix) on pronouns and particles. 10  As an anecdotal (but nonetheless useful) measure, in the longest of the alphabetic Hurrian texts, RS 1.004 (KTU 1.42), one finds the following 19 forms in 62 lines of text: pġdm (l. 3), a̓ wrnm (l. 4), a̓ rdnm (l. 4), ti̓nm (l. 8), i̓ym (l. 15), ṯnm (l. 16), ti̓nm (l. 28), [. . .]nm (l. 30), a̓ rdnm (l. 31), pddm (l. 43), ṯnm (l. 46), [. . . ?]m (l. 49), ttm (l. 52), psm (l. 53), [. . .]nm (l. 55), a̓ ṯthnm (l. 56), hbtm (l. 56), ṯnm (l. 59), ṯnm (l. 62). 19 forms in 62 lines comes out to roughly one form every three lines, in a text that has 3–4 words per line. 11  In particular, in their description of the functions of -ma as “irrespective of whether its origin is Hurrian or Ugaritic” (Dietrich and Mayer 1999: 67), they manage to sidestep the issues addressed in the present discussion.

274

Lam

translate as the parallelistic couplet ‘who plans her destiny, // who utters wisdom’. Here they analyze {mdm} as madi=ma (from the noun madi, ‘wisdom’), with an ostensibly Ugaritic-style enclitic -ma to indicate “the inversion of predicate and object in the second member” of the couplet. While this is an intriguing suggestion and would correspond to the “poetic” function of enclitic -ma in Ugaritic (Tropper 2012: 830), their interpretation is rendered dubious by the uncertainty of their lexical identifications. In particular, while it is plausible to take {kt[y]} as a form of the verb kad- (‘to speak’), their translation of the first phrase as ‘who plans her destiny’ depends on two questionable analyses: that of {kby} as a form of a verb kup- meaning ‘to plan, conspire’ and based on a Hittite etymology; and that of {kbny} as a substantivized adjectival form (from a verbal root kib-, ‘to set, place’) with apparently a resultative sense, i.e., ‘that which is set’ > ‘destiny’ (see Dietrich and Mayer 1999: 65; 1995: 26; 1994: 99).12 Even if one accepts their lexical identifications, the question is potentially rendered moot by the use of =ma in Old Hurrian as a clause connective, which could equally explain the usage here. Their second suggestion, in which they analyze the form {ḫlm} in RS 24.285 (KTU 1.131) 13 as an imperative meaning ‘speak up!’ (Dietrich and Mayer 1999: 67), is more promising as a possibility, though still not entirely unambiguous. Certainly, Ugaritic enclitic -ma is known to occur on certain finite verbal forms (Tropper 2012: 829–830; Bordreuil and Pardee 2009: 62), and its potential focusing function here would make good contextual sense. Alternatively, the {m} can be interpreted as indicating an Old Hurrian clause connective, though the obscurity of the broader context makes it difficult to divide up the clauses definitively and (in turn) to evaluate that possibility. Their third suggestion is questionable, not due to the ambiguity related to overlapping functions of Hurrian and Ugaritic morphemes, but because of uncertainty as to whether the {m} represents an enclitic particle at all. The Hurrian form {a̓ṯḫlm} appears to denote a ritual/sacrificial category (perhaps roughly synonymous with the Ugaritic noun dabḥu) and, in the alphabetic Hurrian corpus, always introduces a list of sacrifices (Dietrich and Mayer 1999: 67).13 If the {m} at the end of the form is in fact an enclitic particle, then it certainly would fall under the range of uses of the Ugaritic enclitic -ma (but 12  On the form {kbny}, although the Hurrian morpheme sequence =i=nni does create adjectives/substantives from verbal roots, the form generally denotes the doer of the action, a kind of nomen professionis (see examples in Wegner 2007: 55). 13  RS 24.254 (KTU 1.110) 1; RS 24.255 (KTU 1.111) 3!, 8; RS 24.261 (KTU 1.116) 3, 9, 10; RS 24.274 (KTU 1.125) 1.

Ugaritic Influences on the Hurrian of Ras Shamra-Ugarit

275

not the Hurrian connective =ma)—that is, marking a fixed nominal phrase (Bordreuil and Pardee 2009: 62; see also Tropper 2012: 827). However, this analysis assumes the existence of a noun spelled alphabetically {a̓ṯḫl} but not otherwise attested in that form, even though a verb ašḫ- (‘to offer, to sacrifice’) is known in syllabic Hurrian (Richter 2012: 54–56; see also Laroche 1980: 59–61). In fact, an alternate analysis is possible in which the {m} represents the infinitive morpheme =umme (Laroche 1980: 60; 1968: 501) and the {l} is a verbal derivational morpheme =ol possibly denoting a sense of motion (Wegner 2007: 88). The resultant form, ašḫ=ol=umme, would be a verbal noun meaning ‘sacrifice, offering up’. At our current state of knowledge of Hurrian lexicography, this analysis must be considered more likely. Having raised questions regarding the possible instances of enclitic -ma offered by Dietrich and Mayer, allow me to mention, for the sake of completeness, two other groups of possibilities. The first group consists of the two forms {ti̓nm} and {ṯnm}, which occur repeatedly near the ends of paragraphs in RS 1.004 (KTU 1.42) (Lam 2011: 161–162). Although the interpretation of these phrases is uncertain, the fact and locations of their recurrence would make an adverbial function for each of them at least plausible, reflecting a typical usage of Ugaritic enclitic -ma. With {ṯnm}, there is the possibility of an adverbial meaning ‘twice’ based on the Ugaritic numeral {ṯn} ‘two’, though in that case one would have to consider the entire adverbial form {ṯnm} as a lexical borrowing from Ugaritic. The other group of possibilities concerns the instances of enclitic {m} found in the Hurrian sections of the mixed Ugaritic-Hurrian text RS 24.261 (KTU 1.116): specifically, the forms/phrases {a̓gndym} in line 4, {i̓nṯt ṯlnnṯtm} in line 11, {i̓nṯt a̓tnṯtm} in line 12, and {nntdm} in line 22. The syntax of these occurrences of {m} are particularly interesting yet perplexing. Laroche (1968: 503) suggested the interpretation of these as cases of enclitic -ma that were attached to the first of pairs (or greater numbers) of elements, in order to emphasize the parallel linking of the respective elements. For instance, he takes {a̓gndym ṯdndy} (line 4) to mean ‘in the ‘AGND and in the ṮDND’ (Laroche 1968: 500), with enclitic -ma denoting ‘and’ (which is then followed by two other phrases similarly spelled with {y} to mark the locative -e but without enclitic {m}). Similarly, the phrase {nntdm kltd} in lines 22–23 (found in the middle of a long list of divine recipients of offerings each marked with directive =da) he would render as ‘for Ninatta-(and-)Kulitta’ (Laroche 1968: 500–501), identifying a pair of goddesses known as female attendants to Ishtar (e.g., Güterbock 1983: 156). However, this syntactic pattern for the {m} goes contrary to what we would expect for both the (Old) Hurrian connective and the Ugaritic enclitic

276

Lam

respectively.14 In fact, the position of the tokens of {m}, if anything, reflect the usage of Akkadian enclitic -ma. Alternatively, one could interpret these as Ugaritic enclitic -ma but with uncertain function, though such a solution merely represents an avoidance of the problem. It is notable that, in these cases in which the {m} in question is most likely to be a identified as a Ugaritic-style enclitic -ma from a strictly morphological point of view, the syntactic function remains obscure. In sum, the existence of Ugaritic enclitic -ma as a linguistic borrowing in the alphabetic Hurrian texts of Ras Shamra cannot be established with certainty, though it remains a possibility in a number of cases. 4 Conclusion In light of the evidence just reviewed, what can be said regarding the place of Hurrian in ancient Ugaritian society? Notwithstanding the uncertainties that exist in the identification of individual elements of linguistic influence, I believe that enough evidence exists to support the contention that Ugaritic influence on the Hurrian of Ras-Shamra was structural and extensive, spanning the areas of lexicon, phonology, morphology, and syntax, and thus reflecting a fairly intense contact situation.15 By contrast, Hurrian influence on Ugaritic appears to consist almost entirely of lexical borrowings.16 In typological terms, this assymetry of language contact is consistent with the dominant role of Ugaritic as the local vernacular vis-à-vis Hurrian, a language that might have been spoken by a larger proportion of society in earlier periods, but towards the end of the Ugaritic kingdom (the time of the texts in question) was limited to a small group of speakers—who perhaps no longer acquired it as a first language, and who used it almost exclusively in cultic contexts. Accordingly, 14  In Hurrian, when =ma (or another enclitic) functions as a clause connective, it is invariably found on the first word of the second clause, not the final word of the first. Accordingly, when it connects smaller constituents of clauses, it appears to attach to the second and subsequent elements of the list (see the examples furnished by Speiser [1941: 178], which Laroche [1968: 503 n. 2] cites in support of his interpretation). In Ugaritic, when enclitic -ma is found in lists, it is similarly used to emphasize the second and subsequent elements (Bordreuil and Pardee 2009: 62; Tropper 2012: 831–832). 15  In my view, it would correspond to the “more intense contact” level in the scale given by Thomason (2001: 70–71), though I must emphasize that this assessment is approximate given the dearth of linguistic data at our disposal. 16  As would be expected, these are almost all nouns, mostly of a technical and/or ritual nature (Watson 1995; 1996; Thomason 2001: 70).

Ugaritic Influences on the Hurrian of Ras Shamra-Ugarit

277

the influence of Hurrian on Ugaritic is minimal, given the relatively small proportion of ancient Ugaritians who knew Hurrian, whereas the influence of Ugaritic on Hurrian is more extensive, reflecting the fact that the Hurrian at Ras Shamra would have been produced by bilingual Hurrian-Ugaritic speakers. At the same time, contrary to how the situation is sometimes characterized, the scribes who produced the alphabetic Hurrian texts of Ras Shamra really did know Hurrian, however imperfect this knowledge might have been. In other words, the extent of the Hurrian preserved at Ras Shamra, and the knowledge of Hurrian grammatical structures the texts imply, precludes the possibility of the material in question representing merely “frozen ritual phrases” that were recited and inscribed but not understood. For instance, it is difficult to imagine a 62-line Hurrian ritual text like RS 1.004 (KTU 1.42) being inscribed (in the predominantly vowelless alphabetic script) by anyone other than a real Hurrian speaker. The size of the Hurrian-speaking “community” at Ugarit may have been small by the end of the Late Bronze Age—perhaps being restricted mostly to certain cultic functionaries. But the Hurrian texts from Ras Shamra in alphabetic script should be understood as preserving a linguistic tradition that still found spoken expression toward the end of the kingdom of Ugarit.17 Bibliography André-Salvini, Béatrice and Mirjo Salvini. 1998. “Un nouveau vocabulaire trilingue sumérien-akkadien-hourrite de Ras Shamra,” in David I. Owen and Gernot Wilhelm (eds.), Studies on the Civilization and Culture of Nuzi and the Hurrians, Vol. 9. General Studies and Excavations at Nuzi 10/2. Bethesda. 3–40. ———. 1999. “Additions and Corrections to SCCNH 9 (1998) 3–40,” in David I. Owen and Gernot Wilhelm (eds.), Studies on the Civilization and Culture of Nuzi and the Hurrians, Vol. 10. Nuzi at Seventy-Five. Bethesda. 434–435. Bordreuil, Pierre and Dennis Pardee. 2009. A Manual of Ugaritic (LSAW 3). Winona Lake. Bush, Frederic William. 1964. A Grammar of the Hurrian Language. Ph.D. Diss., Brandeis University. Dietrich, M., and W. Mayer. 1994. “Hurritische Weihrauch-Beschwörungen in ugaritischer Alphabetschrift,” UF 26: 76–112. ———. 1995. “Sprache und Kultur der Hurriter in Ugarit,” in M. Dietrich and O. Loretz (ed.), Ugarit: Ein ostmediterranes Kulturzentum im Alten Orient. Ergebnisse und

17  For another recent discussion, with full bibliography, see Vita 2009.

278

Lam

Perspektiven der Forschung, Vol. 1. Ugarit und seine altorientalische Umwelt (ALASP 7). Münster. 7–42. ———. 1999. “The Hurrian and Hittite Texts,” in W. G. E. Watson and N. Wyatt (ed.), Handbook of Ugaritic Studies (HdO I/39). Leiden. 58–75. Farber, Walter. 1971. “Zu einigen Enklitika im Hurrischen,” Or 40: 29–66. Gröndahl, Frauke. 1967. Die Personennamen der Texte aus Ugarit (Studia Pohl 1). Rome. Güterbock, Hans G. 1983. “A Hurro-Hittite Hymn to Ishtar,” JAOS 103: 155–164. Hess, Richard. 1999. “The Onomastics of Ugarit,” in W. G. E. Watson and N. Wyatt (ed.), Handbook of Ugaritic Studies (HdO I/39). Leiden. 499–528. Huehnergard, John. 1987. Ugaritic Vocabulary in Syllabic Transcription (HSS 32). Atlanta. Khachikyan, Margarit L. 2005. “On the Origin and Evolution of the Particle -ne in Hurrian,” in David I. Owen and Gernot Wilhelm (eds.), Studies on the Civilization and Culture of Nuzi and the Hurrians, Vol. 15. General Studies and Excavations at Nuzi 11/1. Bethesda. 187–191. Kinlaw, Dennis Franklin. 1967. A Study of the Personal Names in the Akkadian Texts from Ugarit. Ph.D. Diss., Brandeis University. Lam, Joseph. 2006. “The Hurrian Section of the Ugaritic Ritual Text RS 24.643 (KTU 1.148),” UF 38: 399–413. ———. 2011. “A Reassessment of the Alphabetic Hurrian Text RS 1.004 (KTU 1.42): A Ritual Anointing of Deities?,” JANER 11: 148–169. Laroche, Emmanuel. 1968. “Documents en langue hourrite provenant de Ras Shamra,” in J.-C. Courtois (ed.), Ugaritica V (Mission de Ras Shamra 16). Paris. 447–544. ———. 1980. Glossaire de la langue hourrite. Paris. Malbran-Labat, Florence. 1999. “Langues et écritures à Ougarit,” Semitica 49: 65–101. Neu, Erich. 1996. Das hurritische Epos der Freilassung I: Untersuchungen zu einem hurritisch-hethitischen Textensemble aus Ḫattuša (SBT 32). Wiesbaden. Pardee, Dennis. 1996. “L’ougaritique et le hourrite dans les textes rituels de Ras ShamraOugarit,” in Françoise Briquel-Chatonnet (ed.), Mosaïque de langues, mosaïque culturelle: Le bilinguisme dans le Proche-Orient ancient (Antiquités Sémitiques 1). Paris. 63–80. ———. 2002. Ritual and Cult at Ugarit (Writings from the Ancient World 10). Atlanta. ———. 2007. “The Ugaritic Alphabetic Cuneiform Writing System in the Context of Other Alphabetic Systems,” in Cynthia L. Miller (ed.), Studies in Semitic and Afroasiatic Linguistics Presented to Gene B. Gragg (SAOC 60). Chicago. 181–200. Richter, Thomas. 2005. “Hurriter und Hurritisch im Bronzezeitlichen Syrien,” in Doris Prechel (ed.), Motivation und Mechanismen des Kulturkontaktes in der späten Bronzezeit (Eothen 13). Florence. 145–178. ———. 2012. Bibliographisches Glossar des Hurritischen. Wiesbaden. Salvini, Mirjo. 1995. “Ougarit et les Hourrites,” in M. Yon, M. Sznycer, and P. Bordreuil (eds.), Le pays d’Ougarit autour de 1200 av. J.-C.: Actes du Colloque International Paris, 28 juin–1er juillet 1993 (Ras Shamra-Ougarit XI). Paris. 89–97.

Ugaritic Influences on the Hurrian of Ras Shamra-Ugarit

279

van Soldt, Wilfred. 1990. Review of Huehnergard 1987. BiOr 5/6: 728–736. ———. 2003. “The Use of Hurrian Names at Ugarit,” UF 35: 681–707. Speiser, E. A. 1941. Introduction to Hurrian (Annual of the American Schools of Oriental Research 20). New Haven. Thomason, Sarah G. 2001. Language Contact: An Introduction. Washington. Tropper, Josef. 2012. Ugaritische Grammatik (2nd ed.; AOAT 273). Münster. Tropper, Josef and Juan-Pablo Vita. 2003. “Der Gott des Hauses,” UF 35: 673–680. Virolleaud, Charles. 1929. “Les inscriptions cunéiformes de Ras Shamra,” Syria 10: 304–310. Vita, Juan-Pablo. 1999. “The Society of Ugarit,” in W. G. E. Watson and N. Wyatt (ed.), Handbook of Ugaritic Studies (HdO I/39). Leiden. 455–498. ———. 2009. “Hurrian as a Living Language in Ugaritic Society,” in D. A. Barreyra Fracaroli and G. del Olmo Lete (eds.), Reconstruyendo el Pasado Remoto: Estudios sobre el Próximo Oriente Antiguo en homenaje a Jorge R. Silva Castillo (Aula Orientalis Supplementa 25). Sabadell. 219–231. Watson, W. G. E. 1995. “Non-Semitic Words in the Ugaritic Lexicon,” UF 27: 533–558. ———. 1996. “Non-Semitic Words in the Ugaritic Lexicon (2),” UF 28: 701–719. ———. 1998. “Non-Semitic Words in the Ugaritic Lexicon (3),” UF 30: 753–760. ———. 1999. “Non-Semitic Words in the Ugaritic Lexicon (4),” UF 31: 785–799. Wegner, Ilse. 2007. Einführung in die hurritische Sprache (2nd ed.). Wiesbaden. Wilhelm, Gernot. 2004. “Hurrian,” in R. D. Woodard (ed.), The Cambridge Encyclopedia of the World’s Ancient Languages. Cambridge. 95–118.

The Lexical Component in the Aramaic Substrate of Palestinian Arabic Mila Neishtadt 1

A Brief Background to Palestinian Arabic

The term ‘Palestinian Arabic’ (hereafter PA) designates several socially heterogeneous dialect groups. Scholarly research traditionally divides PA into three such groups, spoken by town-dwellers, villagers, and Bedouin.1 Each of these dialect groups displays religious and/or geographical variations. PA is spoken by a variety of religious groups: Muslims, Christians, Druze, and Samaritans. Gender and age are sometimes important factors in language variation as well. The political situation also has an impact: the strong influence of Modern Hebrew (evident especially in the lexicon) can be immediately noticed in the Arabic spoken by Palestinians/Arabs with Israeli citizenship.2 Consequently, as with every ‘language’, there is much more than one variety of PA. It would be misleading to understand the term PA in an exclusively geographical sense: Arabic dialects spoken by the Bedouin population in the Negev and in Galilee are both different from each other and different from PA as a whole.3 In order to avoid confusion, the term PA is usually applied to a single linguistic type: the sedentary Arabic of the Syro-Palestinian dialectal group spoken in Palestine.4 It is widely accepted that Hebrew ceased to be spoken in Palestine around the 2nd century CE and was gradually supplanted by Aramaic, several varieties of which were spoken by Christians, Jews, Samaritans, and pagans. During the Roman and the Byzantine periods, before the Arab conquests, these varieties * I would like to thank Simon Hopkins for introducing me to this fascinating subject and guiding me through my first steps. I am grateful to Yishai Neuman for his most helpful remarks. Any mistakes are mine alone. 1  Heikki Palva, “A General Classification for the Arabic Dialects Spoken in Palestine and Transjordan,” StOr 55 (1984): 371–372. 2  Muhammad Hasan Amara, “Hebrew and English Borrowings in Palestinian Arabic in Israel: A Sociolinguistic Study in Lexical Integration and Diffusion,” in Yasir Suleiman (ed.), Language and Society in the Middle East and North Africa (Richmond, 1999), 81–103. 3  Palva, “Classification,” 371–372; Rony Henkin, Negev Arabic: Dialectal, Sociolinguistic, and Stylistic Variation (Wiesbaden, 2010), 48–49. 4  Simon Hopkins, “Notes on the History of the Arabic Language in Palestine,” LiCCOSEC 20 (2011): 51. © koninklijke brill nv, leiden, ���5 | doi ��.��63/9789004300156_016

the aramaic substrate of palestinian arabic

281

of Aramaic, influenced by Greek, were the dominant spoken language of Palestine.5 Like other cases of language shift, the shift from Aramaic to Arabic in Palestine must not be understood as a sharp replacement of one spoken language by another accomplished within a generation or two, but rather as a gradual and lengthy process, probably with a significant phase of AramaicArabic bilingualism.6 Moreover, the contact between Aramaic and Arabic in this area was not limited to the period following the Arab conquests: Arabicspeaking communities existed on the outskirts of Palestine before the time of the conquests;7 in fact, most of the few known Arabic inscriptions which predate Islam are from Syria and Palestine.8 On the other hand, some Aramaicspeaking communities maintained Aramaic as their spoken language well into the Muslim period. One example is the Samaritans, who became a monolingual Arabic-speaking community apparently only from the 11th century CE onwards.9 The Aramaic > Arabic shift in the Levant has not yet been entirely completed: until very recently, three Western Neo-Aramaic-speaking communities populated the Syrian villages of Maʽlūla, Baxʽa, and Jubbʽadīn. During the course of the Syrian Civil War, Maʽlūla and Baxʽa were severely hit and depopulated.10 As in other cases of language shift, the supplanting language (Arabic) was not left untouched by the supplanted language (Aramaic) and the existence of an Aramaic substrate in Syro-Palestinian colloquial Arabic has been widely 5   Klaus Beyer (trans. J. F. Healey), The Aramaic Language: Its Distribution and Subdivisions (Göttingen, 1986), 34–43, 46–53. 6   On the history of Arabic-Aramaic language contact, see Stefan Weninger, “AramaicArabic Language Contact,” in idem (ed.), The Semitic Languages (Berlin, 2011), 747–755. For an example of the linguistic status in the monastic communities of Palestine in the Byzantine and the early Islamic periods, see Sidney Griffith, “From Aramaic to Arabic: The Languages of the Monasteries of Palestine in the Byzantine and Early Islamic Periods,” Dumbarton Oaks Papers 51 (1997): 11–31. 7   Isaac Hasson, “The Penetration of Arab Tribes in Eretz Israel during the First Century of the Hegira,” Cathedra 32 (1984): 60–62. (in Hebrew) 8    A brief summary may be found in Ahmad Al-Jallad, Ancient Levantine Arabic: A Reconstruction Based on the Earliest Sources and the Modern Dialects (Ph.D. Diss., Harvard University, 2012), 10–24. 9   Haseeb Shehadeh, “When Did Arabic Replace Samaritan Aramaic?,” in Moshe Bar-Asher et al. (eds.), Hebrew Language Studies Presented to Professor Zeev Ben-Ḥayyim (Jerusalem, 1983), 515–528. (in Hebrew) 10  I owe this information to a personal correspondence with Werner Arnold. See also: http://www.telegraph.co.uk/news/worldnews/middleeast/syria/10768900/Syria-warMaaloulas-monastery-destroyed-after-Assad-forces-drive-rebels-out.html—retrieved on 26–05–2014.

282

Neishtadt

accepted.11 The influence of the Aramaic substrate is especially evident in many Palestinian place names12 and in the vocabulary of traditional life and industries: agriculture, flora, fauna, food, tools, utensils, etc. Some elements of the substrate lexicon go even farther back, to Canaanite/ Biblical Hebrew, and in some cases may indeed be recognized as belonging to that group. This article aims to demonstrate the basic methodology which will help us to identify and analyze Canaanite and/or Aramaic lexical elements which survived in PA, the supplanting language of Palestinian Aramaic dialects. 2

General Theoretical Premises

In a diachronic study like this, which examines the influence of the substrate language on the lexicon of the supplanting language, it is important to draw a clear distinction between loanwords and substrate words.13 Both terms refer to words having migrated from one language to another. However, they convey different diachronic realities. Loanwords enter an already existing target language from a source language, whereby both languages are not mutually involved in language shift, i.e., when the source language is not being abandoned by its speakers in favour of the target language, for example, Modern Hebrew ‫ מחסום‬max̱som14 ‘check-point, barrier’ was adapted in PA as maḥs(ū/ō)m ‘check-point’, whose form corresponds to the Arabic noun pattern mafʽūl, whence pl. mafāʽīl > maḥāsīm. Substrate words, on the contrary, are words that remain in use in the course of language shift, a process whereby a speech community ceases to speak one language and acquires another one instead. Features of the abandoned language (in our case Aramaic) which are still extant in the supplanting language (in our case Arabic) after the language shift may be regarded as substrate influences. Substrate words are one example of such features. In the process of language shift speakers may continue to use part of the lexicon of the abandoned language, thus defining their own version of the supplanting language. Although such words may go through several phonological and/or semantic changes and adaptations, a substrate 11  Werner Diem, “Studien zur Frage des Substrats im Arabischen,” Der Islam 56 (1979): 43. 12  For an important study of this subject, see Y. Elizur, Ancient Place Names in the Holy Land (Jerusalem, 2004). (Revised Hebrew version: Jerusalem, 2009). 13  For this basic distinction, see Sarah Grey Thomason and Terrence Kaufman, Language Contact, Creolization and Genetic Linguistics (Berkeley, 1988), 38–41. 14  I have followed the Hebew transcription/romanization conventions of EHLL, vol. 1, xiii–xiv.

the aramaic substrate of palestinian arabic

283

word is not a new loanword but rather an already existing word which participates in the formation of the newly acquired language. As an example we may consider the Coptic substrate in spoken Egyptian Arabic, e.g., baqrūr ‘frog’, libīs ‘carp-like Nile fish’, and mirīsi ‘south wind (in the jargon of Nile boatmen)’.15 Once it has been established that the goal is to determine which of the lexical items of PA are substrate words, we shall define terms such as candidate word and supportive evidence. A candidate word is a word of PA which either (1) has been mentioned in previous scholarly research as a substrate word, or (2) based on supportive evidence (see below), we deem a possible substrate word whose history we are attempting to elucidate. To find out whether a candidate word is a genuine substrate word or not, we shall employ various methods of investigation, involving different types of supportive evidence. Such evidence may be linguistic or circumstantial. 3

Review of Previous Studies

3.1 Fraenkel (1886) One of the pioneer studies of the Aramaic vocabulary in Classical Arabic (CA) is Fraenkel’s Die aramäischen Fremdwörter im Arabischen,16 which contains more than 1,500 entries in CA and in late medieval Arabic literature, classified according to several semantic fields. Each entry can be relatively easily traced due to the book’s index, which is arranged in alphabetical order. In his introduction Fraenkel suggests a few guidelines for recognizing Aramaic loanwords,17 among them the following: (1) consonant correspondence; (2) noun pattern; (3) lack of standardization and variability (vocalization, gender, plural forms). But despite its broad scope and reliability, and although some colloquial Arabic words attested in late medieval Arabic literature are mentioned in Fraenkel’s study, the question of the Aramaic substrate lexicon

15  Bishai lists 109 words considered to be genuine Coptic substrate words, see Wilson B. Bishai, “Coptic Lexical Influence on Egyptian Arabic,” JNES 23 (1964): 40, 43. For a detailed discussion of several wrong identifications in Bishai’s study, see Federico Corriente, “Coptic Loanwords of Egyptian Arabic in Comparison with the Parallel Case of Romance Loanwords in Andalusi Arabic, with the True Egyptian Etymon of Al-Andalus,” Collectanea Christiana Orientalia 5 (2008): 59–123. 16  Siegmund Fraenkel, Die aramäischen Fremdwörter im Arabischen (Leiden, 1886). 17  Ibid., xiv–xvii. As we shall see, many of these criteria apply equally to the identification of substrate lexical elements in PA.

284

Neishtadt

in Arabic dialects is not the main focus of that work, and thus remained largely unanswered. 3.2 Féghali (1918) One of the earliest comprehensive and systematic studies of Aramaic loanwords and substrate words in Lebanese Arabic was undertaken by Féghali.18 This work follows an earlier work of Hobeïka19 and mentions over two hundred words. Féghali’s study is important mainly because he was a pioneer who collected material from spoken Arabic, his native language. Féghali did consider the question of identification and analysis of Aramaic loanwords/substrate words by using sound shifts and consonant correspondence as his main criterion. There are, however, several problems with his summary and with the analysis of his findings. In his review of this work, Brockelmann rightly mentions that Féghali should have treated sound shifts with much more care and attention. For example, Féghali’s analysis of sound shifts ignores chronology (old vs. late Aramaisms).20 Yet a bigger methodological flaw lies in the fact that he reached general conclusions by analysing all the sound shifts in the attested lexicon as a whole, rather than checking and analysing each case in its own right. As a result, Féghali’s basic categorization of the attested lexicon in three different categories21 is incoherent. Another methodological problem with Féghali’s work is that no distinction is made between loanwords and substrate words. The existence of a certain word in Syriac thus serves as a starting point to establish Aramaic origin in both of these two distinctive cases. Except for the ‘church loans’, learned

18  Michel T. Féghali, Étude sur les emprunts syriaques dans les parlers arabes du Liban (Paris, 1918). 19  J. Hobeïka, Al-dawāṯir: baḥṯ fī baqāyā al-luġa al-suryāniyya fī al-luġa al-ʽarabiyya al-ʽāmmiyya fī Lubnān wa Sūriyya: Étymologie arabo-syriaque: Mots et locutions syriaques dans l’idiome vulgaire du Liban et de la Syrie (2 vols.; Basconta, 1902 and 1904). This book was partially published in alphabetical order in al-Machriq 37 (1939): 289–412. This 1939 edition was most recently reprinted: Youssef Hobeica, The Influences of Syriac on the Lebanese and Syrian Dialects (Piscataway, 2011). Unless indicated otherwise, the latest edition, published in 2011, is cited for the purposes of this paper. 20  C. Brockelmann, “Review of Féghali, M. F. É tude sur les emprunts syriaques dans les parlers arabes du Liban. Paris: H. Champion 1918,” OLZ 29 (1926): 197. 21  Féghali, Emprunts, 22–24.

the aramaic substrate of palestinian arabic

285

Syriac clerical loanwords, it would have been methodically more appropriate to search for the origin of the words used in daily life in Western Aramaic dialects and not in Syriac.22 3.3 Diem (1979) Diem’s “Studien zur Frage des Substrats im Arabischen” was a milestone in the study of the substrate question in Arabic.23 Diem favours a strict minimalist approach, according to which, for the assumption of ‘substrate influence’ not to remain merely a hypothesis, clear and strict criteria should be set. Based on conditions formulated by Schmitt24 Diem outlined two general conditions for claiming substrate influence: 1.

2.

We must have some knowledge of the substrate language. The suspected substrate elements have to be recognizable. If we lack sufficient attestation of the substrate language, then at least data from living daughter languages should be available. The phenomenon under investigation should not be attested outside the specific geographical area of the hypothesized substrate influence, unless the distribution of the phenomenon can be explained by migration or wave motion.25

According to Diem, if the first condition is not satisfied, the substrate hypothesis ought to be rejected. If only the first condition appears to be met, the substrate influence hypothesis is questionable at the very least, since even if it cannot be entirely disproven that a certain phenomenon may spontaneously arise in one place and independently exist in another place in same language region due to substrate influence, in such a case this influence cannot necessarily be regarded as the cause. The second condition challenges researchers with the ‘collection of data’: Arabic dialectology and its rich literature allow

22  Brockelmann, “Review of Féghali,” 198. 23  Diem, “Studien,” 12–80. 24  Alfred Schmitt, “Die Germanische Lautverschiebung und ihr Wert für die Frage nach der Heimat der Indogermanen,” in Helmut Arntz (ed.), Germanen und Indogermanen: Volkstum, Sprache, Heimat, Kultur: Festschrift für Herman Hirt (2 vols.; Heidelberg, 1936), vol. 2, 354–357. 25  Diem, “Studien,” 16.

286

Neishtadt

us only a glimpse. Thus, there is a constant danger that a certain phenomenon will be declared to be unusual and explained in terms of substrate theory, when in fact the same phenomenon prevails in various Arabic-speaking communities with a different substrate.26 Therefore, the first line of investigation should always be to look for an internal explanation of the phenomenon, based on the development of the dialect itself. Even if both conditions are fulfilled there are other uncertainty factors involved and it is only rarely that substrate influence may serve as an explanation in the full sense of the word.27 Diem’s highly minimalist approach raises important points concerning the detection and identification of substrate elements. It seems reasonable to adopt a more sympathetic yet cautious approach. When Diem’s first condition is not met the substrate hypothesis may still be true, just not provable. It is true that the available data on Arabic dialectology are far from exhaustive. This is certainly the case with Aramaic, too. But it is possible that a combination of various data may lead us to the conclusion that a specific feature could most plausibly be attributed to substrate influence. In other cases we might be less certain and reach the conclusion that based on available data a certain feature is perhaps only likely, or possibly, a substrate-feature. The degree of probability may vary, thus rather than either accepting or rejecting a certain phenomenon as being a substrate-feature we can talk of a scale of probability. 3.4 Other Sources Several other works refer to the Aramaic substrate in dialectal Arabic. Among these one can mention the works of Rosenthal,28 Frayha,29 Seidel,30

26  For several such examples from Syro-Palestinian Arabic dialects, see Diem, “Studien,” 44. 27  Ibid., 16–17. 28  Franz Rosenthal, “Aramäische Überreste in den heutigen arabischen Dialekten Syriens,” in his Die aramäistische Forschung seit Th. Nöldeke’s Veröffentlichungen (Leiden, 1939), 169–172. 29  Anis Frayha, Muʽğam al-alfāẓ al-ʽāmmiyya: A Dictionary of Non-Classical Vocables in the Spoken Arabic of Lebanon (Beirut, 1973; originally published in 1947). 30  Ulrich Seidel, “Studien zum Vokabular der Landwirtschaft im Syrischen,” AoF 15 (1988): 133–173; idem, 16 (1989): 89–139.

the aramaic substrate of palestinian arabic

287

Arnold and Behnstedt,31 Contini,32 Mubaraka,33 Jastrow,34 Retsö,35 and Río Sánchez.36 Much data on the Aramaic substrate in Cypriot and Bedouin Arabic can be found in several studies published by Borg.37 As for the Canaanite and/or Aramaic substrate specifically in PA, several articles were dedicated to particular lexical items: ḥazīṭ,38 sifl,39 ṣarār,40 and 31   Werner Arnold and Peter Behnstedt, Arabisch-aramäische Sprachbeziehungen im Qalamūn (Syrien) (Wiesbaden, 1993), 62–95. 32  Riccardo Contini, “Le substrat araméen en néo-arabe libanais: Préliminaires à une enquête systématique,” in Marcello Lamberti and Livia Tonelli (eds.), Afroasiatica Tergestina: Papers from the 9th Italian Meeting of Afro-Asiatic (Hamito-Semitic) Linguistics, Trieste 23–24 April 1998 (Padova, 1999), 101–128. 33  Fadel M. Mubaraka, Baqāyā al-ārāmiyya fī luġat ahl Ṣadad al-maḥkiyya: The Aramaic Influence in the Vernacular of Sadad (Damascus, 1999). 34  Otto Jastrow, “Aramäische Lehnwörter in den arabischen Dialekten der Südost-Türkei,” in Stefan Wild and Hartmut Schild (eds.), Akten des 27. Deutschen Orientalistentages (Bonn – 28. September bis 2. Oktober 1998): Norm und Abweichung (Würzburg, 2001), 615–621; idem, “Aramäische Lehnwörter im arabischen Dialekt von Kinderib (Südost-Türkei),” in Manfred Woidich and Jordi Aguadé (eds.), Estudios de dialectología norteafricana y andalusí (EDNA) 8. Homenaje a Peter Behnstedt en su 60 aniversario / Festschrift für Peter Behnstedt zum 60. Geburtstag (Zaragoza, 2004), 99–103. 35  Jan Retsö, “Aramaic/Syriac Loanwords,” in EALL, vol. 1, 178–182. 36  Francisco del Río Sánchez, “Influences of Aramaic on Dialectal Arabic,” in Juan Pedro Monferrer-Sala and Wilfred G. E. Watson (eds.), Archaism and Innovation in the Semitic Languages (Cordoba, 2013), 129–136. 37  For Cypriot Arabic, see Alexander Borg, “Aramaic Substrata in Eastern Arabic with Special Reference to Cypriot Arabic,” in Daniel Sivan and Pablo-Itshak Halevy-Kirtchuk (eds.), Yaakov Bentolila Jubilee Volume: Research Papers in Hebrew Linguistics, Hebrew Literature and Jewish Languages (Occasional Publications in Jewish Studies 8; Beer-Sheva, 2003), 1*–18*. For more data, see idem, Comparative Glossary of Cypriot Maronite Arabic (Arabic – English) (Leiden, 2004). For Bedouin Arabic, see idem, “Lexical Aramaisms in Nomadic Arabic Vernaculars,” in Stephan Procházka and Veronika Ritt-Benmimoun (eds.), Between the Atlantic and Indian Oceans: Studies on Contemporary Arabic Dialects: Proceedings of the 7th AIDA Conference, held in Vienna from 5–9 September 2006 (Vol. 4) (Münster, 2008), 89–112. 38   M. Piamenta, “Ḥazīṭ ‘Unfortunate’—A Decaying Element in Colloquial Arabic,” in J. Mansour (ed.), Arabic and Islamic Studies (Bar-Ilan Departmental Researches 1; RamatGan, 1973), xlvi–xlviii and 260. (p. 260 in Hebrew) 39  Simon Hopkins, “Ṣarār ‘pebbles’—A Canaanite Substrate Word in Palestinian Arabic,” ZAL 30 (1995): 41–43. 40  Ibid., 43–49.

288

Neishtadt

bandūq.41 A number of recent papers with detailed information on several substrate words in PA were published by Halayqa42 and Bassal.43 A more detailed study dedicated to Aramaic substrate in PA is a dissertation by Bassal.44 With over three hundred and fifty entries altogether, it serves as an important source of Aramaic substrate words. However, it does suffer from several methodological and practical problems. From the diachronic point of view it is true that Christians might have retained Aramaic substrate words longer than Muslims. The Western Neo-Aramaic-speaking communities in Syria bear witness to this: the population of Maʽlūla was mostly Christian until the recent events in Syria; apparently Jubbʽadīn still had a completely Christian population until the middle of 16th century CE.45 Yet some formerly Aramaic-speaking Christian communities became Arabic-speaking already a century after the Arab conquests without undergoing any religious change. For example, the Palestinian Melkite community, the former speakers of Christian Palestinian Aramaic (CPA), is known to have shifted to Arabic in both speech and writing as early as the 8th century CE while retaining their Christian identity.46 Whether today the spoken Arabic of the Christian inhabitants of 41  Simon Hopkins, “On the Etymology of Arabic bandūq ‘bastard’,” in Renaud Kuty, Ulrich Seeger, and Shabo Talay (eds.), Beiträge zur semitischen Dialektologie: Festschrift für Werner Arnold zum 60. Geburtstag (Wiesbaden, 2013), 145–150. 42  Issam K. H. Halayqa, “Canaanite and Aramaic Linguistic Remains in the Names of the Tools and Objects in the Palestinian Vernacular,” UF 40 (2008): 421–428; idem, “The Names of the Traditional Establishments for Animal Breeding,” MLR 20 (2013): 117–131; idem, “The Names of the Traditional Water Facilities in the Palestinian Colloquial,” in Renaud Kuty, Ulrich Seeger, and Shabo Talay (eds.), Beiträge zur semitischen Dialektologie: Festschrift für Werner Arnold zum 60. Geburtstag (Wiesbaden, 2013), 125–134; idem, “The Traditional Names for Land Markers and of Protective Establishments of Land Produce in the Palestinian Colloquial,” ZAL 58 (2013): 66–88. 43  Ibrahim Bassal, “Ancient Hebrew and Aramaic Words of Agriculture Remaining in Spoken Galilean Arabic,” Haivrit Weaḥyoteha 6–7 (2006–2007): 89–114 (in Hebrew); idem, “Strata of Foreign Languages in Palestinian Spoken Arabic in Israel,” LiCCOSEC 15 (2010): 3–18; idem, “Hebrew and Aramaic Substrata in Spoken Palestinian Arabic,” MLR 19 (2012): 85–104. 44  Ibrahim Bassal, Hebrew and Aramaic Elements in the Vernacular Christian Arabic in Israel and in the Written Christian Arabic in the Holy Land, Syria and Lebanon (Ph.D. Diss., Haifa University, 2004). (in Hebrew) 45  Muhammad Adnan Bakhit, “The Christian Population of the Province of Damascus in the Sixteenth Century,” in Benjamin Braude and Bernard Lewis (eds.), Christians and Jews in the Ottoman Empire: The Functioning of a Plural Society (2 vols.; New York, 1982), vol. 2, 23. 46  Milka Rubin, “Arabization versus Islamization in the Palestinian Melkite Community during the Early Muslim Period,” in Arieh Kofsky and Guy G. Stroumsa (eds.), Sharing

the aramaic substrate of palestinian arabic

289

Kufr Yāsīf and other villages in Galilee should be necessary defined as ‘Christian Arabic’ as formulated by Bassal requires further investigation. In his study Bassal distinguishes between Aramaisms in ‘spoken Christian Arabic’, ‘written Christian Arabic’, and ‘both spoken and written Christian Arabic’. However, the ‘written’ vs. ‘spoken’ distinction is a side issue. The more basic distinction which deserves to have been used is between two kinds of Aramaisms: (a) ‘Aramaic loanwords’—learned clerical borrowings from Syriac and (b) ‘Aramaic substrate words’—colloquial elements which remained in the speech of speakers of Aramaic who over centuries shifted to Arabic and which survived through oral transmission in PA to this day.47 Two recent studies are those of Basis.48 Although sometimes lacking proper argumentation, the importance of these works lies primarily in the fact that they bring to light some unique substrate words.49 Another important written source for Aramaic substrate words in PA are ethnographical and folklore studies attesting to daily life of traditional Palestinian society. Although most of these studies are not linguistically oriented, they still provide rich and sometimes unique relevant vocabulary. First and foremost of these sources is Dalman’s Arbeit und Sitte in Palästina.50 In the seven volumes of this monumental work we find a comprehensive overview of traditional daily life in Palestine in the first third of the twentieth century. The importance of this study lies not only in its scope, but also in the fact that it was written by a skilled Semitic linguist, whose transcriptions and linguistic remarks throughout the book are of major importance. The value of this exemplary work is considerably enhanced by indices at the end of each volume, listing the Hebrew, Aramaic, and PA words mentioned in it as well as by Dalman’s personal photographs and illustrations. As we have seen, considerable material on the Aramaic substrate in PA is already available. While some identifications are no more than guesses, others are less disputable and a good number may be regarded as certain. Some scholars are maximalists, others minimalists. So far, two points have been the Sacred: Religious Contacts and Conflicts in the Holy Land: First-Fifteenth Centuries CE (Jerusalem, 1998), 156–158. 47  For this distinction, see also Borg, “Substrata,” 12. 48  Rabia Basis, The Aramaic and Hebrew Elements in the Israeli Druzes Dialect (MA Thesis, Haifa University, 2000) (in Hebrew); eadem, Aramaic and Hebrew Elements in the Spoken and Written Arabic of the Druze in Israel (Ph.D. Diss., Haifa University, 2009). (in Hebrew) 49  See, e.g., Hopkins, “bandūq,” 145–150. 50  Gustav Dalman, Arbeit und Sitte in Palästina (7 vols. in 8; Gütersloh, 1928–1942; vol. 8; Berlin 2001).

290

Neishtadt

made clear: (1) An overt distinction should be made between loanwords and substrate words; (2) Substrate words should be identified and classified by rigorous criteria. 4

Basic Methodology

In this article I would like to provide a few general guidelines to help decide whether a certain candidate word is indeed a substrate word. As a preliminary step, a cumulative list of candidate words should be drawn from previous scholarly studies, be they linguistic or not; then, using supportive evidence, we can consider whether each candidate word is an Aramaic substrate word or not, and to what degree of likelihood. The list has to be sifted, and rigorous criteria are needed for determining whether a word should be in the list or not. I believe that a sympathetic minimalist approach is advisable, and many candidate words will remain in the ‘possible’ category. As a rule of thumb, each candidate word will be expected to be absent from CA and thus not be attested in canonical dictionaries of CA, such as Lisān al-ʽarab. But even if a certain word is attested in CA dictionaries, it may still be a genuine substrate word in PA, especially when it is (a) qualified in CA dictionaries as typical to the spoken language within the Syro-Palestinian Arabic dialect region and/or (b) ascribed a meaning that differs from its meaning in PA, provided the latter is identical to the meaning in Aramaic (see durdār below). Sometimes it is this kind of attestation in CA dictionaries that may lead to the suggestion that a specific usage of a certain word in PA is not of Arabic origin. The criteria which should be applied may be summarized in five points, as follows. 4.1 Phonology: Irregular Consonant Correspondence Checking the root of a candidate word in light of consonant correspondence in the Semitic languages can often help us to determine the likelihood of an Aramaic and sometimes even Canaanite/Hebrew provenance. For example, the root brḫ is found in two PA denominative verbs in everyday use, baraḫ and barraḫ, meaning ‘to kneel, sit down to relax, sit on the ground (animal)’ and ‘to make kneel, sit down to relax, sit on the ground (animal)’, respectively.51 The verb barraḫ is cited in several CA dictionaries; e.g., Lisān 51  Attestation of these verbs seems to be restricted to PA sources, see Charles ClermontGanneau, Archaeological Researches in Palestine during the Years 1873–1874 (2 vols.; London, 1896), vol. 2, 481; Leonhard Bauer, Das palästinensische Arabisch: Die Dialekte des Städters und des Fellachen: Grammatik, Übungen und Chrestomathie (4th ed.; Leipzig, 1926), 186;

291

the aramaic substrate of palestinian arabic

al-ʽarab cites the following line of poetry: wa-law yuqālū barriḫū la-barraḫū ‘and if it was said (to them) “kneel” they knelt’. The imperative barriḫū is explained by the Arabic imperative barrikū ‘kneel!’ and the meaning of the verb barraḫū is given as ḏallū wa-ḫaḍaʽū ‘were subdued and humbled’; furthermore, the verb is said to be of Aramaic (nabaṭiyya) origin.52 However, other than in a number of CA dictionaries, this verb is not attested in CA literature. The common CA root is brk and the forms barak and barrak occur with the same meanings in several Arabic dialects,53 including PA.54 Whereas in Arabic the phoneme /k/ displays no environmentally-conditioned plosive/spirant alternation, in both Hebrew and Aramaic this phoneme is realized in two conditioned allophonic variants traditionally transcribed as [k] (voiceless velar plosive) and [ḵ] (voiceless velar spirant), as part of the begadkefat spirantization phenomenon: Proto-Semitic CA

PA

/k/

/k/ (> /č/)

brk

/k/

Aramaic

Hebrew

/k/

/k/

[k] [ḵ] barak ‘to kneel’ baraḫ ‘to kneel, sit down Syriac Peshitta: to relax, sit on the ground (animal)’ barrak ‘to make barraḫ ‘to make kneel, w-’ab̠ reḵ gamlaw(hy) kneel’ sit down to relax, sit on ‘He made his camels the ground (animal)’ kneel down’ (Gen 24:11)

[k] [ḵ] Biblical Hebrew:

way-yab̠ rēḵ hag-gəmallīm ‘He made the camels kneel down’ (Gen 24:11)

Georg Kampffmeyer, Glossar zu den 5000 arabischen Sprichwörtern aus Palästina (Berlin, 1936), 4; ‛Abd al-Laṭīf Maḥmūd al-Barġūṯī, Al-qāmūs al-‛arabī al-ša‛bī al-Filasṭīnī: al-lahğa al-Filasṭīniyya al-dāriğa (Rāmallah, 2001), 135; Ulrich Seeger, Der arabische Dialekt der Dörfer um Ramallah (3 vols.; Wiesbaden, 2009–2013), vol. 2, 21. 52  See Ibn Manẓūr, Muḥammad b. Mukarram, Lisān al-‛arab (6 vols.; Cairo, 1981), vol. 1, 247 (s.v. brḫ). 53   For Syro-Palestinian Arabic, see Adrien Barthélemy (vols. 4–5 ed. by H. Fleisch), Dictionnaire arabe-français: Dialectes de Syrie: Alep, Damas, Liban, Jérusalem (5 vols.; Paris, 1935–1954), vol. 1, 39. For Egyptian Arabic, see Martin Hinds and El-Said Badawi, A Dictionary of Egyptian Arabic (Beirut, 1986), 69. 54  See Leonhard Bauer (with assistance of Anton Spitaler), Deutsch-arabisches Wörterbuch der Umgangssprache in Palästina und im Libanon (2nd ed.; Wiesbaden, 1957), 176 (s.v. ‘Knie’ – barrak ‘knien lassen’).

292

Neishtadt

The root brk is indeed attested in Aramaic55 and in Biblical Hebrew,56 and realized with a velar spirant [ḵ] in most forms. Under no phonological condition is CA velar plosive /k/ known to have merged with PA velar spirant /ḫ/.57 When we thus add the attestation in Aramaic and Hebrew to the attestation and distribution in Arabic, the best explanation for the situation in PA would be substrate influence. From a diachronic point of view, the roots brk and brḫ in PA are cognate, i.e., they stem from a single Proto-Semitic etymon: brk. They are thus doublets, like English chamber and camera. Synchronically, PA exhibits a clear phonological distinction between brk and brḫ: the former is the etymologically expected Arabic form, and the latter goes back to the Aramaic pronunciation of the same Proto-Semitic root, thus helping us to establish substrate influence. In PA, both Arabic brk and Aramaic brḫ possess practically the same meaning58 and some native speakers of PA seem to use both forms more-or-less interchangeably.59 One may thus wonder why the root brḫ survived in PA and was not completely substituted by the root brk. One possible answer may lie in Clermont-Ganneau’s attestation of the imperative form obroḫ ‘sit down!’ as a common interjection used by camel-drivers.60 This use as a common interjection may have had a positive impact on preserving productive usage of this verb throughout the paradigm. 55  For example, in Syriac: peʽal ‘to kneel down’ and afʽel ‘to order s.o. to kneel down; to cast down, cause s.o. to prostrate’; see Carl Brockelmann, Lexicon Syriacum (2nd ed.; Halle, 1928), 96; Michael Sokoloff, A Syriac Lexicon: A Translation from the Latin, Correction, Expansion, and Update of C. Brockelmann’s Lexicon Syriacum (Winona Lake – Piscataway, 2009), 189–190; or in Samaritan Aramaic: afʽel ‘to cause to kneel (camels)’; see Abraham Tal, A Dictionary of Samaritan Aramaic (2 vols.; Leiden, 2000), vol. 1, 116. 56  In pāʽal, ‘to kneel down’ (Ps 95:6; 2Chron 6:13) and in hifʽīl: ‘to cause to kneel (camels)’ (Gen 24:11). 57  Federico Corriente, “A Survey of Spirantization in Semitic and Arabic Phonetics,” JQR 60 (1969): 147–171; Kimary N. Shahin, “Palestinian Arabic,” in EALL, vol. 3, 526–538. 58  A slight semantic difference is mentioned by Seeger: baraḫ / yubruḫ ‘sich setzen um sich zu erholen, zu Boden gehen, aus Müdigkeit am Boden sitzen/liegen (von Mensch und Tier gesagt)’ compare barak / yubruk ‘sich auf den Boden legen, auf dem Boden sitzen/ liegen (nur von Tieren gesagt)’; see Ulrich Seeger, Wörterbuch Palästinensisch – Deutsch, in Zusammenarbeit mit Laṭīfe Abu l-ʿAsal, Taḥsīn ʿAlāwnih, Rāmi il-ʿArabi—work in progress, Status: April 2014. http://www.uni-heidelberg.de/fakultaeten/philosophie/ori/semi tistik/seeger_dictionary.html. 59  I wish to thank Ulrich Seeger for pointing this out to me in a private correspondence. 60  Clermont-Ganneau, Archaeological Researches, vol. 2, 481.

the aramaic substrate of palestinian arabic

293

It is worth noting that the root brḫ in Syro-Palestinian colloquial Arabic is attested not only as a substrate word as shown above, but also as a Syriac learned clerical loanword in the standard phrase b(ā/a)reḫ m(ō/o)r meaning ‘Hello, Father’ lit. ‘Bless (me), my Father’ from Syriac ‫ ܒܪܟ ܡܪܝ‬barreḵ mār(y).61 4.2 Morphology: Noun Patterns, Affixes, Etc. The morphology of a candidate word may lead to corroborative evidence for a certain candidate word to qualify as a substrate word. One possible example is the suffix -ōn, a common diminutive suffix in Aramaic which is also attested in Hebrew, although not always as a diminutive.62 This suffix, realized as -(ū/ō)n, may be found in several colloquial Arabic dialects, including PA.63 One such ‫ن‬ example is the word �‫< ج��م�لو‬ğmlwn> ‘gable; ridged roof’, which although absent in CA Lisān al-ʽarab dictionary is widely attested in late medieval Arabic texts.64 It is also attested, as ‫< גמלון‬ğmlwn>, in an early Judaeo-Arabic glossary dating around the 10th century CE,65 as a translation of the Biblical Hebrew ‫’ ַא ִּתיק‬attīq in Ezek 42:3 and in an explanation of the Biblical Hebrew hapax legomenon ‫יטנּו‬ ֵ ‫ ַר ִה‬rahīṭēnū (Song 1:17) in Tanḥum Ha-Yerushalmi’s commentary from the 13th century CE.66 In Syro-Palestinian dialects this word is attested in various

61  Hobeica, Influences, 10; Féghali, Emprunts, 36; Barthélemy, Dictionnaire, vol. 1, 34. 62  J. Barth, Nominalbildung in den semitischen Sprachen (2nd ed.; Leipzig, 1894), 349; compare Ben-Zion Gross, The Nominal Patterns ‫ פעלון‬and ‫ פעלן‬in Biblical and Mishnaic Hebrew (Jerusalem, 1993), 278–280. (in Hebrew) 63  A similar Arabic suffix -ūn is attested in some Arabic names of endearment, e.g., ḥamdūn, ḫaldūn, etc., but this seems to be an internal development. For a detailed discussion of the etymology of this suffix in different names of this type, see Georg Kampffmeyer, “Südarabisches,” ZDMG 54 (1900): 621–660. See also Carl Brockelmann, Grundriss der vergleichenden Grammatik der semitischen Sprachen (2 vols.; Berlin, 1908–1913), vol. 1, 394; Henri Fleisch, Traité de philologie arabe, 1: Préliminaires, phonétique, morphologie nominale (Beirut, 1961), 453–454. 64  On the attestation of this word in Arabic, see Haggai Ben-Shammai, “On Meira Poliak and Sasson Somekh ‘Two Biblical Judaeo-Arabic Glossaries from Cairo Geniza’, Peʽamim 83 (2000): 15–47,” Peʽamim 88 (2001): 125–130 (in Hebrew); See also Reinhart Pieter Anne Dozy, Supplément aux dictionnaires arabes (2nd ed.; 2 vols.; Leiden, 1927), vol. 1, 219; Fraenkel, Fremdwörter, 29. 65  Meira Poliak and Sasson Somekh, “Two Biblical Judaeo-Arabic Glossaries from Cairo Geniza,” Peʽamim 83 (2000): 34. (in Hebrew) 66  Ms. Poc. 320, fol. 19a, see Aryeh Zoref, Tanchum ha-Yerushalmi’s Commentary on Canticles: Studies in Its Tendencies and Its Jewish, Sufi-Islamic and Christian Sources, with a Critical Edition (Ph.D. Diss., The Hebrew University of Jerusalem, 2012), vol. 2, 21. (in Hebrew)

294

Neishtadt

forms: ğamal(ū/ō)n,67 ğaml(ū/ō)n,68 ğabalūn,69 ğumlūn,70 and ğumlūl 71 with both broken and sound plural forms (ğamālīn72 and -āt,73 respectively) and is also common in the spoken Arabic of Egypt74 and Iraq.75 The phonological makeup of this word is neutral and the expected diminutive parallel *gamlōnā is not attested in Aramaic; the fact that this word occurs in Egyptian Arabic, a dialect with a different substrate, is yet another point against the substrate influence theory in PA. Nonetheless, I believe that we are still on firm ground when treating this non-classical word as a likely Aramaic lexical remnant. The word ‫ ܓܡܐܠ‬gamlā without the diminutive suffix -onā is attested in Syriac in both meanings (‘gable’ and ‘ridged roof’) in Bar Bahlul’s dictionary (10th century CE).76 Further evidence comes from the North Eastern Neo-Aramaic (NENA) dialect of Qaraqosh, where the word gumla means ‘vaulted roof of a church’.77 Identifying colloquial Arabic ğamal(ū/ō)n 67  Buṭrus al-Bustānī, Muḥīṭ al-muḥīṭ: qāmūs muṭawwal li-l-luġa al-‛arabiyya (Beirut, 1987; originally published in 1867–1870), 135; Dozy, Supplément, vol. 1, 219; J. G. Hava, Al-farâ’id Arabic-English Dictionary (Beirut, 1915), 99; Barthélemy, Dictionnaire, vol. 1, 122; Dalman, Arbeit, vol. 7, 152; Frayha, Dictionary, 29; Basis, MA Thesis, 61–62; Bassal, PhD Dissertation, 154–155; Basis, PhD Dissertation, 178–180. Canaan transcribes this word as djamālōn and djamâlōn, see T. Canaan, The Palestinian Arab House: Its Architecture and Folklore (Jerusalem, 1933), 23, 45–46, 74. In his dictionary of PA al-Barġūṯī mentions the adjective ğamalūnī, see al-Barġūṯī, Qāmūs, 290. 68  Hobeica, Influences, 17; Basis, MA Thesis, 61–62; Bassal, PhD Dissertation, 154–155; Basis, PhD Dissertation, 178–180. 69  Transcribed žabaloone, see Karl Stowasser and Moukhtar Ani, A Dictionary of Syrian Arabic: English-Arabic (Washington D.C., 1964), 99. 70  Al-Bustānī, Muḥīṭ, 135. 71  Ibid. 72  Hava, Al-farâ’id, 99; Dozy, Supplément, I, 219. 73  Dozy, Supplément, vol. 1, 219; Frayha, Dictionary, 29; Stowasser and Ani, Dictionary, 99; Basis, MA Thesis, 61–62. 74  S. Spiro, Arabic-English Dictionary of the Modern Arabic of Egypt (2nd ed.; Cairo, 1923), 120; Hinds and Badawi, Dictionary, 173. 75  D. R. Woodhead and W. Been, A Dictionary of Iraqi Arabic, Arabic-English (Washington D.C., 1967), 77. There the expression saquf ğamalūn (transcribed saquf jamaloon) is translated ‘pitched roof’. 76  Rubens Duval, Lexicon Syriacum auctore Hassano bar Bahlule (3 vols.; Paris, 1888–1901), vol. 1, 500. A possible etymological connection between the Arabic ğamalūn and the Syriac gamlā was already suggested in Robert Payne Smith, Thesaurus Syriacus (2 vols., Oxford, 1879–1901), vol. 1, 736. See also Fraenkel, Fremdwörter, 29. 77  Geoffrey Khan, The Neo-Aramaic Dialect of Qaraqosh (Leiden, 2002), 733.

the aramaic substrate of palestinian arabic

295

as a likely Aramaic substrate word contributes to Aramaic lexicography, as this word in colloquial Arabic may be regarded as a living evidence that this word was used with a diminutive suffix in Aramaic. 4.3 Attestation and Distribution in Aramaic There are numerous dialectal words that do not occur in CA, nor can they be found in Aramaic in a related meaning and a similar formal shape. Any attempt to attribute such words to an Aramaic substrate is more than doubtful. One such example is the spoken Arabic word iḫtyār / ḫityār ‘old man’, which has been repeatedly and erroneously attributed to the Syriac verb ‫ ܟܬܪ‬kattar ‘to stay, continue, persist, remain’ and explained by the semantic shift ‘stay, remain > grow old > old man’.78 Although the attestation of this word in colloquial Arabic is unique to the Levant region,79 a former Aramaic speaking area, an Aramaic origin seems highly unlikely in every other respect. The presence ‫خ‬ of the word ihtiyar in Turkish (‫ ا �ت��ي���ا ر‬in Ottoman-Turkish) denoting ‘aged, old; old man’80 suggests that iḫtyār / ḫityār in PA is rather an Ottoman-Turkish loanword of Arabic origin, i.e., ‘chosen leader’ from CA ’iḫtiyār ‘choice’,81 compare PA muḫtār ‘village chief, mayor of the village’. In order to claim an Aramaic substrate in PA we ought to go back to Diem’s first condition, i.e., ‘knowledge of the substrate’. If we lack compelling evidence for the existence of a certain candidate word in the substrate language (in our case Aramaic) it will be difficult to establish substrate influence. Such candidate words may well be Aramaic (which is far from being fully documented!), but until we can make a plausible case for attributing such a word to the substrate, the matter will remain open until new data become available. Attestation in the Western Aramaic group, to which the Aramaic substrate of PA belongs, is of particular importance. If a candidate substrate word in

78  Hobeïka, Dawāṯir, vol. 2, 28–29 (not mentioned in later editions); Basis, MA Thesis, 57–58. This was also suggested as a possible etymology by al-Mawṣilī and by Frayha (Dāwud al-Mawṣilī, Al-āṯār al-ārāmiyya fī luġat Mawṣil al-ʽāmmiyya [Mosul, 1935], 36–37; Frayha, Dictionary, 42). 79  Peter Behnstedt and Manfred Woidich, Wortatlas der arabischen Dialekte, I: Mensch, Natur, Fauna, Flora (Leiden, 2011), 29. 80  J. W. Redhouse, Redhouse yeni Türkçe-İngilizce sözlük: New Redhouse Turkish-English Dictionary (Istanbul, 1979), 522. 81  Dozy, Supplément, vol. 1, 416; Barthélemy, Dictionnaire, vol. 1, 209–210; Behnstedt and Woidich, Wortatlas, 29.

296

Neishtadt

PA is present in the Western Neo-Aramaic of Maʽlūla and vicinity,82 the living descendant of the old Western Aramaic dialects, the chances of this PA word being an Aramaic substrate word considerably increase (see mašāṭīḥ ‘wide/big feet’ below). Although the Neo-Aramaic of Maʽlūla has borrowed many words from Arabic, often recognizable as such, the distribution is still very suggestive. An interesting example of the importance of the Western Neo-Aramaic dialect of Maʽlūla for establishing substrate influence in PA is the word m(i/u)šṭāḥ, pl. mašāṭīḥ. In PA as in other Syro-Palestinian Arabic dialects this word is attested mostly in two different meanings: (a) ‘place on which one dries fruit (figs, dates, grapes)’83 and (b) ‘a kind of flat loaf of bread’.84 Yet a third meaning (c) ‘wide/big feet’, is attested in the plural form mašāṭīḥ by Basis, e.g., mašāṭīḥu kbār ‘his feet are wide/big’; (a)bū mašāṭīḥ ‘a wide/big footed man’.85 The morphology of m(i/u)šṭāḥ is neutral and does not seem necessarily indicative. The Arabic noun pattern mifʽāl, pl. mafāʽīl does of course exist in

82  A short glossary by Bergsträsser has been the source thus far, see Gotthelf Bergsträsser, Glossar des neuaramäischen Dialekts von Maʽlūla (Leipzig, 1921). Materials published later by, e.g., Spitaler and Arnold are another source: Anton Spitaler, “Neue Materialen zum aramäischen Dialekt von Maʽlūla,” ZDMG 107 (1957): 299–339; idem, “The Aramaic dialect of Maʽlūla,” in Franz Rosenthal (ed.), An Aramaic Handbook (2 vols., Wiesbaden, 1967), vol. 2/1, 62–68; idem, vol. 2/2, 82–96; Werner Arnold, Das Neuwestaramäische (5 vols.; Wiesbaden, 1989–1991). A forthcoming dictionary of Western Neo-Aramaic by Arnold will give us a much richer viewpoint. 83  Barthélemy, Dictionnaire, vol. 2, 391; Basis, MA Thesis, 90–91; Bassal, PhD Dissertation, 250; Basis, PhD Dissertation, 142–143. 84  Hobeica, Influences, 46; Féghali, Emprunts, 31; Barthélemy, Dictionnaire, vol. 2, 391; Frayha, Dictionary, 95; Basis, MA Thesis, 90–91; Basal, PhD Dissertation, 250; Basis, PhD Dissertation, 142. Féghali described this kind of bread as a long piece of puff pastry sprinkled with water, impressed with four fingerprints and put into the tannūr (oven), see Michel T. Féghali and Joseph Féghali, Contes, légendes et coutumes populaires du Liban (2nd ed.; 2 vols.; Louvain-la-Neuve, 1976), vol. 1, 36, 42, 50. (Originally published under the title: Contes, légendes et coutumes populaires du Liban et de la Syrie [Paris, 1935]). Dalman attested mušṭāḥ in a popular song as ‘eine Art Brot’ (Gustaf Dalman, Palästinischer Diwan: Als Beitrag zur Volkskunde Palästinas [Leipzig, 1901], 18); in his Arbeit und Sitte in Palästina he translates ‘ein schlecht gebackener Brotkuchen’ (vol. 3, 102) and ‘ein schlecht ausgefallenes Brot’ (vol. 4, 84). 85  Basis, MA Thesis, 90–91; eadem, PhD Dissertation, 142. Basis mentions ‘wide/big feet’ as a unique meaning with no equivalents in other Arabic dialects, yet mišṭāḥ meaning ‘big foot’ in Lebanese colloquial Arabic is mentioned by Frayha (Dictionary, 95).

the aramaic substrate of palestinian arabic

297

CA, mostly for words denoting tools (e.g., CA miftāḥ ‘key’).86 However, the phonological makeup of PA m(i/u)šṭāḥ in the light of consonant correspondence serves as solid evidence for substrate influence, as in Arabic we would have expected /s/ instead of the Canaanite and Aramaic /š/. The occurrence of the Proto-Semitic root šṭḥ (Akkadian šeṭȗ ‘to spread out, lay out’ [CAD Š 343]) in spoken Arabic dialects and its absence in CA provide further evidence. The CA counterpart is the root sṭḥ (see below). Yet several words deriving from the non-classical Arabic root šṭḥ are well attested in Arabic dialects of the Levant, e.g., the verbs šaṭāḥ ‘to extend, lie down; to fell, slip (on the ground); to make a big step; to move away, depart; to go out, go for a walk, have a picnic, wander’;87 šaṭṭāḥ ‘to make someone lie down, make someone extend’;88 tšaṭṭāḥ ‘to lay oneself down, extend on the ground’;89 and the nouns šaṭḥa ‘walk, outing, picnic, trip, excursion’;90 šṭīḥ ‘flat plate usually made of sheepskin (which is used to cover the bread when kneaded;91 on which one puts the dough and then cuts it into small pieces)’,92 and šaṭṭīḥ ‘tubular, finger bar twice as long as the finger that is slipped on the finger for better gripping of the grain (straw)’.93 The root šṭḥ with the same basic meaning of ‘to spread out, lay out’ is also well attested in the Judaeo-Arabic of Iraq.94 The word mišṭāḥ itself is not used in CA; corresponding expected CA forms of this word, m(a/i)sṭaḥ and misṭāḥ

86  Barth, Nominalbildung, 250; William Wright, A Grammar of the Arabic Language (3rd ed.; 2 vols.; Cambridge, 1896–1898), vol. 1, 130. 87  Al-Bustānī, Muḥīṭ, 465; Hobeica, Influences, 46; Hava, Al-farâ’id, 365; Féghali, Emprunts, 31; Dalman, Arbeit, vol. 1.2, 427, 431, 439–440; Barthélemy, Dictionnaire, vol. 2, 391; Halim Dammous, Qāmūs al-ʽawāmm: Dictionnaire populaire (Damascus, 1923), 154; Frayha, Dictionary, 95; Basis, MA Thesis, 90–91; al-Barġūṯī, Qāmūs, 678; J. Elihay, The Olive Tree Dictionary: A Transliterated Dictionary of Conversational Eastern Arabic (Palestinian) (Jerusalem, 2005), 519–520. 88  Hobeica, Influences, 46; Barthélemy, Dictionnaire, vol. 2, 391. 89  Barthélemy, Dictionnaire, vol. 2, 391; Frayha, Dictionary, 95. 90  Hobeica, Influences, 46; Dalman, Arbeit, vol. 1.2, 427, 431, 588; Barthélemy, Dictionnaire, vol. 2, 391; Frayha, Dictionary, 95; al-Barġūṯī, Qāmūs, 678; Elihay, Dictionary, 520. 91  Féghali, Emprunts, 31; Barthélemy, Dictionnaire, vol. 2, 391. 92  Hobeica, Influences, 46; Frayha, Dictionary, 95. 93  Kampffmeyer, Glossar, 37; also cited in Claude Denizeau, Dictionnaire des parlers arabes de Syrie, Liban et Palestine: Supplément au dictionnaire arabe-français de A. Barthélemy (Paris, 1960), 280. 94  Yitzhak Avishur, A Dictionary of the New Judeo-Arabic Written and Spoken in Iraq (1600– 2000) (3 vols.; Tel-Aviv, 2010), vol. 3, 525–526. (in Hebrew)

298

Neishtadt

(with /s/), are attested in the meaning of ‘a place for drying fruit’.95 In fact, m(i/u)sṭāḥ with /s/ is also attested in PA96 alongside m(i/u)šṭāḥ, thus providing an example of a colloquial equivalent of the CA /s/ variant and its Aramaic substrate /š/ doublet (see below). A semantically close enough equivalent of PA m(i/u)šṭāḥ may be found in Hebrew. In Biblical Hebrew (Ezek 26:5; 26:14) we find *mišṭå̅ḥ in the expression ‫ ִמ ְש ַטח ֲח ָר ִמים‬mišṭaḥ ḥărå̅mīm meaning ‘a place for spreading nets to catch fish’, perhaps an aramaicized version of the synonymous Hebrew term ‫טֹוח ַל ֲח ָר ִמים‬ ַ ‫ִמ ְש‬ mišṭōaḥ la-ḥărå̅mīm in Ezek 47:10.97 It is worth noting that mišṭaḥ ḥărå̅mīm / mišṭōaḥ la-ḥărå̅mīm is translated into Judaeo-Arabic as ‫< משטאח ללציאדין‬mšṭ’ḥ l-l-ṣy’dyn> in Bodleian ms. Hunt 206, an early manuscript copied at the end of 12th century CE.98 In the Rabbinic Hebrew of the Babylonian Talmud the root šṭḥ occurs several times with precisely the same meaning of ‘a place on which one dries fruit’, e.g., ‫ שוטח שם פירות‬šoṭeaḥ šam perot ‘spreading out the fruit (to dry) there’.99 Derivatives of šṭḥ with this and closely related meanings are attested in several Aramaic dialects, old and new. In Syriac there are two close equivalents: (1) ‫ ܡܫܛܚܐ‬mašṭāḥā in the phrase ‫ ܒܝܬ ܡܫܛܚܐ‬bēṯ mašṭāḥā in the Peshitta translation of the aforementioned Biblical Hebrew expressions mišṭaḥ ḥărå̅mīm (Ezek 26:5; 26:14) and mišṭōaḥ la-ḥărå̅mīm (Ezek. 47:10) and in the ̈ mašṭāḥē in the Peshitta translation of Biblical Hebrew plural form ‫ܡܫܛܚܐ‬ ‫ ָשטֹוח‬šå̅ṭōaḥ (Num 11:32) ‘things spread out to dry; expanse’; (2) the plural form ̈ mašṭuḥē meaning ‘things spread out (perhaps nets)’, mentioned in ‫ܡܫܛܘܚܐ‬

95  See Ibn Manẓūr, Lisān, vol. 3, 2006; Muḥammad b. Muḥammad Al-Murtaḍā al-Zabīdī, Tāğ al-‛arūs min ğawāhir al-qāmūs (25 vol.; Beirut, 1994), vol. 4, 89–90; Edward William Lane, Arabic English Lexicon (8 vols.; London, 1863–1893), vol. 4, 1357. The CA word m(a/i) sṭaḥ in meaning (a) was mentioned by al-Ğawālīqī (12th century CE) in his Muʽarrab as a Persian loanword (Mawhūb b. Aḥmad al-Ğawālīqī, Al-muʽarrab min kalām al-aʽğamī ʽalā ḥurūf al-muʽğam [Damascus, 1990], 596). Fraenkel considered it a plausible Aramaic loan (Fraenkel, Fremdwörter, 136). 96  Dalman, Arbeit, vol. 1.2, 478, 559; vol. 4, 339, 350; al-Barġūṯī, Qāmūs, 677; also attested by Basis as pronounced with emphatic [ṣ] (Basis, PhD Dissertation, 142). 97  Fraenkel, Fremdwörter, 136. 98  Yitzhak Avishur, A Medieval Translation of the Latter Prophets into Iraqi and Syrian JudaeoArabic: Ezekiel and the Minor Prophets: The Text of Bodleian Manuscript Hunt. 206 with Introduction and Notes (Tel Aviv, 2000), 43, 72, 165. (in Hebrew) 99  ʿAboda Zara 70b.

the aramaic substrate of palestinian arabic

299

Bar Hebraeus’ Chronicon.100 In his Syriac dictionary Audo (1853–1918) mentions mašṭāḥā as having the meaning, among others, of ‘the place in the vineyards on which one spreads out grapes’, yet he does not mention the exact Syriac source from which this meaning was taken, possibly because he took this meaning from the spoken language and not from classical Syriac written sources.101 In Jewish Babylonian Aramaic (JBA) the verb ‫ שטח‬šṭaḥ occurs in the same context, e.g., ‫< דחזיא למישטחא בה פירי‬d-ḥzy’ l-myšṭḥ’ bh pyry> ‘(the Sabbatical year) that is suitable for spreading out fruit to dry’.102 This meaning is attested in several modern Neo-Aramaic dialects, e.g., in the Western NeoAramaic of Jubbʽadīn as mušṭōḥa;103 in Ṭūrōyo as mašṭūḥo104 and mǝštāġa105 and in NENA dialects—in the Barwar dialect as mǝštaxa, pl. mǝštaxe106 and in the Jewish dialect of Zakho as mištāxa.107 Scholars thus rightly agree that the colloquial Arabic word m(i/u)šṭāḥ in meaning (a) is of Aramaic origin.108 As we have seen, the phonological makeup of the word, accompanied by secure attestations in both Arabic and Aramaic, is indeed highly indicative. This word in meaning (a) ought therefore to be considered as a certain instance of substrate influence in PA. Some scholars suggest that m(i/u)šṭāḥ in meaning (b) is of Aramaic origin as well,109 although we do not have evidence from Aramaic for this exact meaning. In my opinion, 100  Jean-Baptiste Abbeloos and Thomas Joseph Lamy, Gregorii Bar-Hebraei chronicon syriacum (3 vols.; Louvain, 1872–1877), vol. 3, 73, as cited in Payne Smith, Thesaurus, vol. 2, 4134. See also Brockelmann, Lexicon, 772; Sokoloff, Syriac Lexicon, 844–845. 101  Thomas (Toma) Audo, Simtā d-leššānā suryāyā (2 vols.; Mosul, 1897), vol. 2, 566. This dictionary has been reprinted several times, most recently under the title Treasure of the Syriac Language: A Dictionary of Classical Syriac (Piscataway, 2008). 102  B M 106a; see also BB 29b (fruit); 68a (sesame). See Michael Sokoloff, A Dictionary of Jewish Babylonian Aramaic of the Talmudic and Geonic Periods (Ramat-Gan, 2002), 1129. 103  Arnold, Das Neuwestaramäische, vol. 2, 88–89. 104  Helmut Ritter, Ṭūrōyo: Die Volkssprache der syrischen Christen des Ṭūr ʽAbdȋn, B: Wörterbuch (Beirut, 1979), 333; Jastrow, “Aramäische Lehnwörter in den arabischen Dialekten der Südost-Türkei,” 618; Aziz Tezel, Comparative Etymological Studies in the Western NeoSyriac (Ṭūrōyo) Lexicon (Uppsala, 2003), 175. 105  Tezel, Etymological Studies, 175. 106  Geoffrey Khan, The Neo-Aramaic Dialect of Barwar (3 vols.; Leiden, 2008), vol. 2, 1336. 107  Yona Sabar, A Jewish Neo-Aramaic Dictionary (Wiesbaden, 2002), 226. 108  Fraenkel, Fremdwörter, 136; Barthélemy, Dictionnaire, vol. 2, 391; Basis, MA Thesis, 90–91; Bassal, PhD Dissertation, 250; Basis, PhD Dissertation, 142. This word is also listed in the lexicon of Aramaic substrate words in the Arabic of Qalamūn (Arnold and Behnstedt, Sprachbeziehungen, 85). 109  Hobeica, Influences, 46; Féghali, Emprunts, 31; Frayha, Dictionary, 95.

300

Neishtadt

meaning (b) in all probability belongs to the substrate as well. New evidence from Aramaic dialects, old and new, may provide the missing link. What about meaning (c) ‘wide/big feet’? Basis suggested this word might be of Aramaic origin, yet no exact or semantically close attestation in Aramaic was provided.110 I did not find any evidence of an Aramaic cognate of this SyroPalestinian word, with the exception of the Neo-Aramaic dialect of Maʽlūla, in which the basic meaning ‘to spread, flatten’ of the root šṭḥ is also applied to feet: mušṭōḥa ‘a (bigger than normal) foot’.111 This unique attestation in the Western Neo-Aramaic of Maʽlūla, accompanied by phonological and distributional evidence in Arabic and Aramaic, qualifies the Syro-Palestinian word mašāṭīḥ with the meaning of ‘wide/big feet’ as certainly belonging to the substrate. It is important to stress that the PA substrate word m(i/u)šṭāḥ in all its three meanings is not to be confused with the loanword mištāḥ, pl. mašātīḥ ‘a wooden pallet on which units of stones, building material etc. are stacked and moved from one place to another’ from Modern Hebrew ‫ משטח‬mištax̱ ‘surface’.112 This distinction is further supported by the non-emphatic /t/ in PA, as is pronounced by native speakers of Modern Hebrew. The pharyngeal /ḥ/ in this loanword is no exception. Although the Hebrew grapheme ח‬corresponds to the voiceless velar spirant [x] in the speech of an overwhelming majority of Jewish speakers of Modern Hebrew, it frequently corresponds to the voiceless pharyngeal spirant /ḥ/ in PA loanwords from Modern Hebrew, e.g., maḥs(ū/ō)m (also maḫsūm) ‘checkpoint’ (Hebr. ‫ מחסום‬max̱som ‘checkpoint, barrier’); maḥfēr ‘excavator’ (Hebr. ‫ מחפר‬max̱ per ‘excavator’); ḥam(a/e)mōt ‘greenhouse’ (Hebr. pl. ‫ חממות‬x̱amamot ‘greenhouses’).113 One 110  Basis, MA Thesis, 90–91; eadem, PhD Dissertation, 142. 111  I am grateful to Werner Arnold for this information. 112  Compare Basis, MA Thesis, 90. See also Basal, PhD Dissertation, 250; Nancy Hawker, Palestinian-Israeli Contact and Linguistic Practices (Oxford, 2013), 43, 153; Seeger, Wörterbuch—work in progress. 113  Nancy Hawker, Palestinian-Israeli Contact and Linguistic Practices (Oxford, 2013), 5 and passim (maḥs[ū/ō]m); 38, 43 (ḥamemōt); Abed al-Rahman Marʽi, Walla bseder: A Linguistic Profile of the Israeli-Arabs (Jerusalem, 2013), 40–41, 101, 201, 216 (in Hebrew); Seeger, Wörterbuch—work in progress. The facts are rather complex, since there is some inconsistency and variation in usage; consider, e.g., biddi aḥōfeš (verbal form) min-i š-šuġl ‘I want to take a day-off (from work)’ (Hebr. ‫ חופש‬x̱ofeš ‘day-off, vacation’), laḫmanyōt ‘buns’ (Hebr. sg. ‫ לחמניה‬lax̱ maniya ‘bun’, pl. ‫ לחמניות‬lax̱ maniyot), see Marʽi, Walla bseder, 33, 199; for (ḫ/ḥ)ōf(e/i)š ‘day-off, vacation’; laḥmaniya ‘bun’ attested in the West-Bank, see Hawker, Palestinian-Israeli Contact, 39, 55; Seeger, Wörterbuch—work in progress. Further investigation and research of the phenomenon in several varieties of PA (among speak-

the aramaic substrate of palestinian arabic

FIGURE 7

301

Carthamus glaucus. Photo: Avinoam Danin.

possible explanation of the pharyngeal /ḥ/ in PA mištāḥ as in other similar cases is graphophonemic assignment: Hebrew ח‬PA /ḥ/.114 4.4 Attestation and Distribution in Arabic We should, of course, also consider the occurrence of the candidate word in Arabic. We have already mentioned the non-attestation of a candidate word in CA as important supportive evidence for the claim that it is a substrate word. Distribution in spoken Arabic is important as well. As a rule of thumb, the more geographically unique to Palestine a word under investigation is, the greater are the chances that it can qualify as a substrate component. So if a unique PA word also exists in Aramaic, especially in Western Aramaic, it is very likely that this word is an Aramaic residue.

ers of different levels of knowledge and literacy in both spoken and written Hebrew) is required. 114  On graphophonemic assignment, see Yishai Neuman, “Graphophonemic Assignment,” in EHLL, vol. 2, 135–145.

302

Neishtadt

An interesting example of this kind is the word q(ō/ū)ṣ. In PA it is one of the names for the plant commonly called ‘distaff thistle’ (Carthamus glaucus). Nothing in this word’s morphological pattern suggests a foreign origin (PA qōṣ is similar in structure to, e.g., PA tōm CA ṯawm ‘garlic’). Nevertheless it is not attested in CA. Comparative Semitic etymology points to the Biblical Hebrew lexeme ‫ קֹוץ‬qōṣ ‘thorn’ (12 occurrences). The possible connection between PA q(ō/ū)ṣ and Biblical Hebrew qōṣ was far from unnoticed by earlier scholars, yet they did not provide arguments for a connection between the two words other than the observable semantic and phonological match.115 As we shall see, we can be rather confident about the direct connection between them. The distribution of q(ō/ū)ṣ ‘thorn’ in Arabic shows clearly that it is a local PA word. Several regional and chronological variants of this word are mentioned in ethnographic sources documenting Palestinian flora and traditional life.116 Al-Barġūṯī also mentions the denominative verb qawwaṣ ‘to uproot, weed (thorns)’, qawwāṣa ‘a Y-shaped stick with which one collects thorns and then cuts them using a qālūše’,117 and qawwāṣ ‘person who uses the qawwāṣa’.118‫ق‬ This word is still in use: in a written website attestation from 2009 we find ‫�و�ص‬ in a list of common plants near ʽAqraba village located in the district of Nāblus.119 This word is also attested in Nehmé’s two dictionaries of Lebanese

115  Leonhard Bauer, Volksleben im Lande der Bibel (Leipzig, 1903), 182, f. 5; Ephraim Rubinovitch, Ḥidušim ba-šemot ha-ʽaraviyim šel ṣimḥe Ereṣ Yisra’el u-ve-qešer ’itam ba-šemot ha-ʽivrim (Innovation of Arabic Names of Plants in the Palestinian Flora in Relation to Their Hebrew Cognates) (Tel Aviv, 1923), 2–3 (in Hebrew); Immanuel Löw, Die Flora der Juden (4 vols., Vienna, 1924–1934), vol. 1, 394–396; vol. 4, 36; Dalman, Arbeit, vol. 1.1, 51; vol. 2, 315–316. 116  See above-mentioned sources. See also E. von Mülinen, “Beiträge zur Kenntnis des Karmels,” ZDPV 30 (1907): 136; Grace M. Crowfoot and Louise Baldensperger, From Cedar to Hyssop (London, 1932), 61–62. Bauer mentions both qūṣ and qūs (Bauer, Wörterbuch, 77 [s.v. ‘Distel’]). 117  The word qālūše, pl. qawālīš and -āt of the pattern fāʽūl(a) means ‘a small sickle’ and is itself a fascinating case study of a most likely substrate word which clearly fulfills most of the above-mentioned criteria. As already suggested by Dalman, it is reminiscent of the Biblical Hebrew hapax legomenon ‫ ִק ְלׁשֹון‬qilləšōn (1Sam 13:21) (Dalman, Arbeit, vol. 3, 25). See also Arnold and Behnstedt, Sprachbeziehungen, 89. 118  Al-Barġūṯī, Qāmūs, 1042. 119  See http://www.palestineremembered.com/GeoPoints/_Aqraba_556/Article_13586.html —retrieved on 27.05.2014.

the aramaic substrate of palestinian arabic

303

and Syrian flora. There qūs (realized with non-emphatic [s]) appears among several names of Carthamus tenuis, another species of Carthamus.120 This word also appears as part of a pair ‫< ואלקוץ ואלדרדרי‬w-’l-qwṣ w-’l-drdry> in an early Judaeo-Arabic commentary (ms. BL 5562A.1) to Gen 3:18 ‫וְ קֹוץ וְ ַד ְר ַדר‬ wə-qōṣ wə-ḏardar, which was published by Tobi.121 One could argue that here both and are embedded Hebrew borrowings, i.e., loanwords and not substrate words. However, the modern usage of qōṣ and—as we shall see—durdār in PA suggests that this text could well be Palestinian and the specific usage in Arabic could, in fact, help us determine the geographic origin of this Judaeo-Arabic commentary. Now let us discuss the attestation and distribution in Aramaic. An equivalent of PA q(ō/ū)ṣ is documented in several Aramaic sources from the relevant region. One piece of evidence comes from Samaritan Aramaic. In the Samaritan translation of Gen 3:18 the Biblical Hebrew pair wə-qōṣ wə-ḏardar is translated ‫< וקוצן ודרדר‬w-qwṣn w-drdr>.122 The Biblical Hebrew plural form ‫ ק ִֹצים‬qōṣīm in Exod 22:5 is translated ‫< קוצנים‬qwṣnym> (ms. British Museum Or. 7562) with the variant ‫< קוצים‬qwṣym> (ms. Vatican Cod. Barberini Or 1).123 It is interesting to note that Arabic translations of the Samaritan translation to the‫ق‬ ‫ن‬ Pentateuch (dating to the beginning of the 13th century) also suggest � ‫�و�ص�ا‬ in a translation of this verse.124 Evidence from Samaritan Aramaic as well as from Arabic is found also in Ha-meliṣ, a Samaritan lexicographic glossary of the Pentateuch which consists of three parallel columns: Hebrew, Samaritan Aramaic, and Arabic. The Aramaic part is dated to the end of the 10th century CE. As Aramaic gradually ceased to be spoken and the understanding of Aramaic among the Samaritans diminished, an Arabic column was added by a later hand, probably between 120  Moustapha Nehmé, Dictionnaire étymologique de la flore du Liban: Noms scientifiques et leur étymologie: noms français, anglais et arabes: Noms arabes translittérés (Beirut, 2000), 75; idem, Etymological Dictionary of Syrian Flora: Scientific Names and Their Etymology: Arabic, English and French Names (Beirut, 2008), 83–84. Dalman also mentions qōs and qūṣ as PA words for Carthamus tenuis (Dalman, Arbeit, vol. 2, 312). 121  Yosef Tobi, “Early Judaeo-Arabic Translations of the Pentateuch,” Haivrit Weaḥyoteha 4–5 (2004–2005), 128. (in Hebrew) 122  Abraham Tal, The Samaritan Targum of the Pentateuch: A Critical Edition, Vol. 1. Genesis, Exodus (Tel-Aviv, 1980), 12; idem, Dictionary, vol. 1, 193. 123  Tal, Samaritan Targum, 312–313; idem, Dictionary, vol. 2, 769. 124  Haseeb Shehadeh, The Arabic Translation of the Samaritan Pentateuch, Vol. 1. Genesis— Exodus (Jerusalem, 1989), 360. In PA qōṣān is attested as one of the names for the ‘globe thistle’ (Echinops viscosus), see Dalman, Arbeit, vol. 2, 312.

304

Neishtadt

the early 11th to the 14th century CE. In this glossary we find this word in the Hebrew column both in the plural and in the singular. The Hebrew plural form qōṣīm in Exod 22:5 is translated as ‫< קיצין‬qyṣyn> in Aramaic and ‫< שויך‬šwyk> ‘small thorns’ in Arabic, and the Hebrew singular form qōṣ in Gen 3:18 is translated as ‫< כבין‬kbyn> ‘thorns’ in Aramaic and as ‫< קוץ‬qwṣ> in Arabic. This seems to be one of the earliest attestations of q(ō/ū)ṣ in Arabic.125 Other evidence comes from the Jewish Palestinian Aramaic (JPA) of the Palestinian Targums. In the Fragment Targum (formerly known as Targum Yerushalmi II) in ms. Paris one finds ‫< קוצין‬qwṣyn> (Gen 3:18)126 and in Cambridge University Library ms. T-S 20.155r one finds ‫< קוצנין‬qwṣnyn> (Exod 22:5).127 In Targum Neofiti 1 we find ‫< קציין‬qṣyyn> as a marginal gloss in Gen 3:18128 and ‫< קוצצין‬qwṣṣyn> as a marginal gloss in Exod 22:5.129 It seems that this word is absent from Eastern Aramaic and is not attested in Syriac, Mandaic, or modern Eastern Aramaic dialects. The plural form ‫קוצי‬ does occur three times in the Babylonian Talmud.130 It is, however, most likely that the plural form in the Aramaic of the Babylonian Talmud is a literary borrowing from Hebrew with Aramaic declension rather than a genuine Aramaic word. The Hebrew lexeme ‫< קוץ‬qwṣ> frequently occurs in the Talmud not only in Hebrew, but also as an embedded Hebrew loan in Aramaic, e.g., ‫< והא נמי חזיא למשקלא בה קוץ‬w-h’ nmy hzy’ l-mšql’ bh qwṣ> ‘and this too is suitable for removing a splinter’.131 Furthermore, two of the three occurrences of appear in the expression ‫< דאית בה קוצי‬d-’yt bh qwṣy> ‘that has thorns in it’ (B.B. 19b; 20a) which is found in a slightly different version ‫< דאית בה קוצים‬d-’yt bh qwṣym> in Beṣ 30a, where the plural form ‫קוצים‬ is a clear embedded Hebrew loan. In light of all the corroborative evidence presented thus far, namely the lack of attestation in CA and the fact that its occurrence is limited to Palestine in 125  Z. Ben-Ḥayyim, The Literary and Oral Tradition of Hebrew and Aramaic amongst the Samaritans (5 vols.; Jerusalem, 1957–1977), vol. 2, 580. (in Hebrew). On this commentary and its dating, see idem, vol. 1, lxv–lxxiii. 126  Michael Klein, The Fragment-Targums of the Pentateuch according to Their Extant Sources (2 vols.; Rome, 1980), vol. 1, 46. 127  Michael Klein, Genizah Manuscripts of Palestinian Targum to the Pentateuch (2 vols.; Cincinnati, 1986), vol. 1, 289. 128  Alejandro Díez Macho, Neophyti 1: Targum Palestinense ms de la Biblioteca Vaticana (6 vols.; Madrid, 1968–1979), vol. 1, 17. 129  Ibid., vol. 2, 143. 130  Ḥul. 43b; B.B. 19b; 20a. 131  Šab. 123a.

the aramaic substrate of palestinian arabic

305

FIGURE 8  Centaurea pallescens Photo: Avinoam Danin

Arabic, Aramaic, and finally—Hebrew, we can conclude that this word should be added to the relatively short list of examples of the Canaanite > Aramaic substrate in PA. Put differently, this is a ‘remnant word’ which survived in local speech from the Biblical period to the present time. 4.5 Semantics From a semantic perspective, we would like to examine to what extent a certain Syro-Palestinian candidate word diverges from its Arabic meaning and/or is attested with a related meaning in Aramaic. The closer to Western Aramaic its meaning is, the higher its chance to qualify as a substrate word. An interesting case study from the realm of flora is that of pa durdār ‘yellow star thistle’ (Centaurea pallescens), also known in PA as murrēr and murrār.132 132  F. Petermann, Reisen im Orient (2nd ed.; 2 vols.; Leipzig, 1865), vol. 1, 74; Rubinovitch, Ḥidušim, 5; Löw, Flora, vol. 1, 405; Dalman, Arbeit, vol. 1.1, 51; vol. 2, 316–317; Crowfoot

306

Neishtadt

In this case, however, it is primarily the meaning that serves as crucial evidence that it is an Aramaic substrate word. There are a number of references to durdār or a similar word in the spoken Arabic of the Syro-Palestinian region.133 Ibn al-Bayṭār (13th century CE) men‫ة‬ tions that the form ��‫< د ر د ر�ي‬drdryh> is commonly used by people of Diyār Bakr, today’s Diyarbakır (Turkey), a former Aramaic-speaking region, to denote a thorny plant also called murrār and ‫< �مر�ر‬mryr>.134 In Judaeo-Arabic a close ‫ي‬ relative of this word also appears in Saadia Gaon’s commentary on Gen. 3:18 in the singular ‫< דרדר‬drdr>135 and according to Yemenite tradition in the plural ‫< דראדר‬dr’dr>.136 It also appears in the pair ‫< ואלקוץ ואלדרדרי‬w-’l-qwṣ w-’ldrdry> in the previously mentioned Judaeo-Arabic commentary to Gen 3:18 published by Tobi.137 The CA equivalent of PA durdār is dardār. This word’s formal shape consisting of two identical repeated pairs of consonants does not provide us with evidence for substrate influence, since words of this pattern are not rare in CA, e.g., balbāl ‘excitement’, ṣalṣāl ‘clay, loam’, ḫalḫāl ‘anklet’.138 We now turn to the semantics of this word. In Lisān al-ʽarab we find dardār meaning ‘a type of well-known tree’.139 The Tāğ al-ʽarūs dictionary suggests it is šağarat al-baqq ‘a baqq tree’, a term explained as follows, as translated by Lane: ‘There come forth from it various ’aqmā‛ (app. excrescences of the nature of gall-nuts), like pomegranates, in which is a humour that becomes baqq and Baldensperger, From Cedar to Hyssop, 40. In his dictionary, s.v. ‘Distel’, Bauer notes that durdār is used in PA in the North and murrēr and murrār in the South (Bauer, Wörterbuch, 77). 133  For PA, see above-mentioned sources. For Lebanese and Syrian Arabic, see al-Bustānī, Muḥīṭ, 275; Dozy, Supplément, vol. 1, 432; Hobeica, Influences, 25; Féghali, Emprunts, 59. The term dardār is mentioned as the name of Centaurea calcitrapa, another species of Centaurea, see Nehmé, Flore du Liban, 76; idem, Dictionary of Syrian Flora, 86. 134  Ḍiyā’ al-Dīn ʽAbdallāh b. Aḥmad Ibn al-Bayṭār, Al-ğāmiʽ li-mufradāt al-adwiya wa-laġḏiya (4 vols.; Cairo, 1874), vol. 4, 148. See also Dozy, Supplément, vol. 1, 432; Hobeica, Influences, 25. 135  According to ms. Saint-Petersburg Eвр. II C and J. Derenbourg (ed.), Oeuvres complètes de R. Saadia Ben Iosef al-Fayyoûmî: Volume premier: Version arabe du pentateuque (Paris, 1893), 9. 136  Yehuda Ratzabi, A Dictionary of Judaeo-Arabic in R. Saadya’s Tafsir (Ramat-Gan, 1985), 64. (in Hebrew) 137  See 4.4 above; Tobi, “Early Translations,” 128. 138  Barth, Nominalbildung, 205–206. 139  Ibn Manẓūr, Lisān, vol. 2, 1359 (s.v. dardār).

the aramaic substrate of palestinian arabic

307

(i.e. bugs or gnats, for both are signified by this word) and when they burst open, the baqq come forth: its leaves are eaten, in their fresh state, like herbs, or leguminous plants.’140 In Modern Standard Arabic (MSA) the term dardār usually applies to an elm (Ulmus) or an ash (Fraxinus).141 The identification of dardār with an elm tree was already mentioned by Ibn al-Bayṭār: būqīṣā huwa šağarat-u al-dardār-i al-maʽrūfat-u bi-l-šām-i wa-l-ʽirāq-i bi-šağarat-i al-baqq-i ‘An elm-tree is a dardār tree, known in Syria and in Iraq as a baqq tree’.142 Although both of these words (CA dardār and PA durdār) belong to the same semantic field of flora, the meanings are rather different: a type of a tree as opposed to a short dry thistle bush. The semantic resemblance between durdār in colloquial Syro-Palestinian Arabic and Biblical Hebrew and/or Aramaic equivalents was mentioned in earlier studies.143 Dalman briefly notes that the names qōṣ and durdār are reminiscent of the Biblical pair qōṣ wə-ḏardar (Exod 3:18; Hos 10:8) and that the meaning of ‘thorn’, as he puts it, ‘does not suit an ash tree’.144 But it was Basis who not only noted the semantic difference between CA dardār and spoken PA durdār, but rightly pointed out the importance of this difference for determining that this PA word can be traced back to its Aramaic substrate. Basis suggests that this word is a Hebrew remnant which survived via Aramaic in PA, but takes no further steps to provide actual evidence, for example by examining the distribution of this word in Arabic and Aramaic.145 The Canaano-Aramaic equivalents are indeed a much better semantic match. The Biblical Hebrew lexeme ‫ ַּד ְר ַּדר‬dardar occurs twice in the pair qōṣ wə-ḏardar (Exod 3:18; Hos 10:8) commonly translated as ‘thorns and thistles’. As for attestation and distribution in Aramaic, the same ‘thistle’ meaning may be found in Western Aramaic. In the Samaritan translations of the Pentateuch the Biblical Hebrew wə-qōṣ wə-ḏardar (Gen 3:18) is translated ‫וקוצן ודרדר‬

140  Murtaḍā al-Zabīdī, Tāğ al-‛arūs, vol. 6, 400 (s.v. dardār). For translation, see Lane, Lexicon, vol. 3, 864–865. 141  Hans Wehr (trans. and ed. J. Milton Cowan), A Dictionary of Modern Written Arabic (3rd ed.; Ithaca, 1976), 277; Nehmé, Flore du Liban, 122, 267; idem, Dictionary of Syrian Flora, 143–144, 313. 142  Ibn al-Bayṭār, Al-ğāmiʽ, vol. 1, 127. 143  Petermann, Reisen, vol. 1, 74; Hobeica, Influences, 25; Féghali, Emprunts, 59; Rubinovitch, Ḥidušim, 5; Löw, Flora, vol. 1, 407; Dalman, Arbeit, vol. 1.1, 51; vol. 2, 316–317. 144  Dalman, Arbeit, vol. 1.1, 51; vol. 2, 316–317. 145  Basis, MA Thesis, 32–33. Basis did not develop this suggestion in her later PhD dissertation.

308

Neishtadt

.146 In CPA the plural form ‫< ܕܪܕܪܝܢ‬drdryn> and in JPA ‫דרדרין‬ and ‫< דרדורין‬drdwryn> appear in the same meaning as well.147 Equivalents of CA dardār and PA durdār are well attested in Eastern Aramaic. The word ‫ דרדרא‬dardrā is mentioned in the Babylonian Talmud and explained there as ‫< מוריקא דחוחי‬mwryq’ d-ḥwḥy> ‘prickly safflower’.148 In Syriac we find two semantically distinctive lexemes.149 One of these, ‫ ܕܕܪܐ‬daddārā,150 is attested with the meaning of ‘elm tree’ and is repeatedly explained in medieval Syriac-Arabic dictionaries by the aforementioned CA term šağarat al-baqq or by the Arabic term ‫ د د ا ر‬dVddār, closely resembling the CA form dardār in both form and meaning.151 As for the other lexeme, the plural form ‫ܕ̈ܪܕܪܐ‬ dardrē is widely attested with the meaning of ‘thistles’: e.g., in the Peshitta of the Old Testament dardrē occurs four times, as the translation of different Biblical Hebrew words for ‘thistles’: dardar (Exod 3:18; Hos 10:8), ‫ ִס ִירים‬sīrīm (Hos 2:8), and ‫חֹוח‬ ַ ḥōaḥ (Is 34:13).152 In Neo-Aramaic dialects we find both meanings. In Maʽlūla (Western NeoAramaic) ḏartarīṯa153 and tartōra (collective)154 are attested as ‘Ulme, Esche’ (‘elm, ash’). In the NENA dialect of Hertevin we find darderra as a general term for construction materials of a roof, translated by Jastrow as ‘Bauholz’

146  Tal, Samaritan Targum, 12. 147  For CPA, see Fridericus Schulthess, Lexicon Syropalaestinum (Berlin, 1903), 48. For JPA, see Díez Macho, Neophyti 1, vol. 1, 17; Klein, Fragment-Targums, vol. 1, 46; Michael Sokoloff, A Dictionary of Jewish Palestinian Aramaic of the Byzantine Period (2nd ed.; Ramat-Gan, 2002), 213. 148  Giṭ 70a; translated by Löw as ‘stachlige Saflor’ (Löw, Flora, vol. 1, 400). 149  These lexemes are discussed by Löw, yet he does not present the whole complexity of this case study (Immanuel Löw, Aramäische Pflanzennamen [Leipzig, 1881], 98–101). 150  The lexeme ‫< ܕܪܕܐ‬drd’> is probably a misspelling of ‫< ܕܕܪܐ‬ddr’> daddārā, the only difference between the Syriac ‫< ܕ‬d> and ‫< ܪ‬r> being a diacritic point (see Payne Smith, Thesaurus, vol. 1, 946; Löw, Pflanzennamen, 99). 151  Payne Smith, Thesaurus, vol. 1, 824–825, 946–947. On the etymology of daddārā and its connection to Persian ‫ د ي��د ا ر‬dīdār / ‫ د ر د ا ر‬dardār, Biblical Hebrew ‫ ִת ְד ָהר‬tiḏhå̅r (Is 41:19; 60:13) and CA dardār, see Löw, Pflanzennamen, 99; Paul de Lagarde, Uebersicht über die im Aramäischen, Arabischen und Hebräischen übliche Bildung der Nomina (Göttingen, 1889), 130; Brockelmann, Lexicon, 142. 152  Payne Smith, Thesaurus, vol. 1, 946–947. 153  Bergsträsser, Glossar, 22. According to information kindly provided by Werner Arnold, this word is no longer used. The current corresponding word is tartarīṯa. 154  Arnold, Das Neuwestaramäische, vol. 4, 334. It could be that tartarīṯa and tartōra are Arabic loanwords, since in genuine Aramaic words Old Aramaic /d/ at the beginning of a word should appear as /ḏ/ in the Western Neo-Aramaic dialects.

the aramaic substrate of palestinian arabic

309

(‘timber’).155 In the NENA dialect of Telkepe near Mosul daṛədṛa is ‘a kind of thistle with edible seeds’.156 In the NENA dialect of Qaraqosh the word darədra, pl. darədre means ‘tick (on animals)’.157 The cumulative data on the PA word durdār (Centaurea pallescens) allow us to ascribe it to the Aramaic substrate with a high degree of probability. The Canaanite-origin hypothesis is indeed appealing and quite possible. We may therefore conclude with considerable confidence that the PA word durdār goes back at least to an Aramaic, if not to an earlier Hebrew origin, since this word has a meaning which is not attested in CA, but is found in both Eastern and Western Aramaic and farther back in Biblical Hebrew. The distribution of the word in spoken Arabic in former-Aramaic speaking regions (Syria-LebanonPalestine) provides further evidence for this claim. 5

Summary of Findings

The existence of Aramaic substrate elements in Palestinian Arabic, the supplanting language of Palestinian Aramaic dialects, is widely accepted and is especially evident in the lexical component. Yet detection, identification, and—ultimately—analysis of those elements are still far from being complete. This article aims to provide some general methodological guidelines for this purpose and to demonstrate them by a few detailed case studies. A cumulative list of possible substrate words should be drawn from previous studies and additional sources; using supportive evidence, we can consider whether each of these candidate words is an Aramaic substrate word or not, and to what degree of likelihood. As a rule of thumb, candidate words will be expected to be absent from CA. Among the linguistic criteria applied are: (a) phonology; (b) morphology; (c) attestation and distribution in Aramaic; (d) attestation and distribution in Arabic; (e) semantics. It is true that even if candidate words are examined in light of all the proposed criteria, it will not always be possible to arrive at unequivocal conclusions, and in many cases it may be difficult to determine whether a certain word should indeed be 155  Otto Jastrow, Der neuaramäische Dialekt von Hertevin (Provinz Siirt) (Wiesbaden, 1988), 115, 185. 156  I am grateful to Hezy Mutzafi for this information. 157  Khan, Qaraqosh, 730. Providing no similar word for ‘tick’ exists in the region, it would be difficult to think of a different etymology in this case. One possible explanation may be elm tree galls, abnormal outgrowths of plant tissues that can be caused by various parasites and very closely resemble the shape of a tick on an animal’s skin.

310

Neishtadt

considered as belonging to the Aramaic substrate, even though it might be quite possible. Hopefully, cumulative evidence will help us resolve many cases. Some words will seem to qualify on all or almost all counts (e.g., baraḫ / barraḫ); in others it might be primarily the meaning which will reveal their origin (e.g., durdār). For yet others attestation (or non-attestation) and distribution in Arabic and Aramaic will be highly indicative (e.g., q(ō/ū)ṣ; mašāṭīḥ meaning ‘wide/big feet’). It is important to note the different levels of certainty involved in identifying substrate words. Many PA words which the existing literature sometimes considers as Aramaic substrate words turn out to be completely unrelated to Aramaic (e.g., iḫtyār / ḫityār). A considerably large number of PA words which previous studies considered as Aramaisms might possibly or probably be substrate words, although further supportive evidence is desirable (e.g., m(i/u)šṭāḥ meaning ‘kind of flat loaf of bread’). Other candidate words some of which have not been mentioned in the existing literature seem almost or absolutely certain to belong to the substrate. In other words, rather than determining whether a certain candidate word qualifies as a substrate item or not, we ought to map each potential case study on a ‘scale of probability’, with a rough division into the following four levels of certainty: (a) certain; (b) likely; (c) possible, and (d) negative (but claimed to be Aramaic substrate influence in the existing literature). 6

Expected Contribution

To conclude this paper, I would like to point out the importance of identifying and classifying the Aramaic component of PA. Such a study is expected to contribute to the field of Semitic linguistics and lexicography (both in Arabic and Aramaic), and perhaps also shed light on Canaanite/Hebrew, which preceded Arabic and Aramaic in Palestine. Linguistics, however, is not the only realm of interest for such a study. In effect, this study may be valuable to other disciplines such as archaeology and the material culture of the region, Biblical studies and Talmud scholarship as well as the historical study of Palestine.

The Classification of Hobyot Aaron D. Rubin 1 Introduction There are six languages that make up the so-called Modern South Arabian (MSA) group of languages, a subgroup of the western branch of the Semitic family.1 Three of the languages, Mehri, Soqoṭri, and Jibbali, were first brought to the attention of Europeans in the 1830s and 40s. These also happen to be the three languages with the greatest number of speakers at present. Two more languages, Ḥarsusi and Baṭḥari, were first mentioned in the work of Bertram Thomas (1937), who was also the first European to cross the Empty Quarter. The sixth language, Hobyot, was discovered only in the late 1970s, by the British scholar T. M. Johnstone, and first mentioned in print in 1981.2 For more than thirty years after this first mention, Hobyot data remained extremely scant. Although Johnstone mentions Hobyot in a couple of his works, he did not actually publish any data on the language. And in the thirty-plus years since this first mention, very little data had been published on the language until 2013. The first data on Hobyot, a dense two pages worth, was published in a 1985 article by Antoine Lonnet; a bit more appeared in subsequent publications of Lonnet, and also a few of Marie-Claude Simeone-Senelle, both of whom did fieldwork in Yemen in the 1980s. A short article was published by Werner Arnold in 1993, based on some meetings he had with a native Hobyot speaker in Syria.3 In the summer of 2013, a lexicon was published posthumously by the Japanese scholar Aki’o Nakano, giving us a relative wealth of new data to consider. In the last couple of years two teams of scholars have been doing some fieldwork on Hobyot, one group from the UK, and one from France, and so no doubt much more data is forthcoming. But with the publication of Nakano’s dictionary, we can make a preliminary sketch of the grammar, and with this we 1  My sincere thanks to Julien Dufour and Antoine Lonnet for their invaluable comments and corrections to this chapter, especially, but by no means only, for the insight they provided on Soqoṭri. Thanks also to Leonid Kogan and Maria Bulakh for sharing their own Soqoṭri data before publication. 2  The brief mention in Johnstone (1991) was actually sent for publication prior to the note in Johnstone (1981), but obviously not published until much later. 3  Much data was collected by Miranda Morris during this time, but was not published.

© koninklijke brill nv, leiden, ���5 | doi ��.��63/9789004300156_017

312

Rubin

can at least ascertain where Hobyot fits in within the Modern South Arabian family tree. Specifically, we can determine if it is more closely related to Mehri or Jibbali. This paper will attempt exactly this. Nearly all of the Hobyot data in the following presentation comes from Nakano’s lexicon, and that lexicon may or may not, of course, be completely reliable. So forthcoming data from other sources may make us reconsider some of the conclusions in this study. Even so, the comparative data is still interesting, and sets up a framework for comparative study of the Modern South Arabian languages. Hobyot is spoken along the Yemeni-Omani border, on both sides. It has only perhaps one or two thousand speakers left. It has less prestige than Mehri, Jibbali, and Soqoṭri, and it is not clear how well children are learning Hobyot. Most speakers are also speakers of Mehri and/or Jibbali, as well as Arabic. All speakers are bi- or tri-lingual. And so there are two big problems with Hobyot data. The first is that there appears to be a lot of variation among speakers, which is actually not unusual for a language with so few speakers. The second is that there is a lot of interference from Mehri and/or Jibbali, and because those languages are so similar, it is not always clear what is pure Hobyot and what is interference. One fieldworker related to me that one native speaker “seemed to feel that he had the right to pick up any lexeme he wanted from Jibbali or Mehri when speaking Hobyot.”4 Even if we can identify what is “real” Hobyot, the geographical overlap between Hobyot, Mehri, and Jibbali also causes some problems for classification, since there are clearly features of these languages that are the result of areal phenomena, and not genetic relationships. For example, Omani Mehri (OM), Jibbali, and Ḥarsusi all have a prefixed definite article, while further west, Yemeni Mehri (YM) dialects and Hobyot have no article. The innovation of the article in the more easterly languages—or perhaps the loss of the article in the more westerly languages—appears to be an areal phenomenon, and does not speak to the genetic classification of the languages. Some other examples of areal phenomena are:

‧ the loss of ʿ (> ʾ or ∅) in Hobyot, Omani Mehri, and Ḥarsusi, but retention in Yemeni Mehri and Jibbali. the ‧ loss of l (in certain environments) in Omani Mehri and Jibbali but retention in Yemeni Mehri, Hobyot, and Ḥarsusi.

4  Julien Dufour, p.c., Feb. 1, 2014.

313

The Classification of Hobyot

‧ the pronunciation of /g/ as [dʒ] in Western Jibbali, Hobyot, and Yemeni Mehri; as g or gy in Omani Mehri, Central/Eastern Jibbali, and Ḥarsusi. ‧ the particle wot (vars. wut, wat) ‘when/if’ in Sharqīyah Yemeni Mehri and Hobyot, not present in more westerly Yemeni Mehri varieties or in Omani Mehri.

In the sections that follow, I will show that Hobyot is more closely connected with Mehri, Ḥarsusi, and Baṭḥari than with Jibbali, based on phonological, morphological, and lexical evidence. Specifically, I will argue, based on current available data, for a family tree that looks as follows: Proto-MSA Western MSA Mehri, Ḥarsusi, Baṭḥari Hobyot

Eastern MSA Jibbali

Soqoṭri

2 Phonology The few sound changes briefly mentioned in the previous section all appear to be areal phenomena. However, there are also sound changes that appear to reflect genetic relationships. The first of these pertains to the outcome of the Proto-Semitic phoneme *s, the phoneme that is more traditionally reconstructed š.5 In Mehri, Ḥarsusi, Baṭḥari, and Hobyot, this phoneme has become h, while in Jibbali it has become š. In Soqoṭri, things are a bit complicated, as the outcome can be š, h, or a third phoneme that I transcribe here yh. In this case, Hobyot patterns clearly with Mehri and its dialects. Some examples are:

5  See Faber (1981; 1985) and Kogan (2011: 61–70) for discussions on the reconstruction of this sibilant.

314

Rubin Mehri6

Ḥars.

Hob.

Jib.

Soqoṭri

Cognates/Notes

ham hōba(ʿ) hēxǝr

hǝm hōba xāhǝr

hum hōbǝʾ hāxǝr7

šum šōʿ šáxǝr

Heb. šēm, Ar. ismHeb. šēḇaʿ, Ar. sabʿ(metathesis in Ḥarsusi & Baṭḥari)

ḥǝ-rōh hīma/ hūma ‘he cooked’ hǝbhōl

ḥǝ-rīh hōma

ḥā-rēh hīmǝʾ

rɛš šĩʿ

šɛm yhóbiʿ yhɛ́ḥar8 ‘man’ (pl. šḥer) ri(h) hímaʿ

abhōl

abhōl

ɛbšél

ébhǝl

Heb. hiḇšīl, Ge‘ez ʾabsala

‘name’ ‘seven’ ‘old man’

‘head’ ‘he heard’

Heb. rōš, Ar. raʾsHeb. šāmaʿ, Ar. samiʿa

Another sound change that reflects a shared development is a bit more complicated. Put simply, there is a whole set of words that historically had an initial glottal stop or y that show up in some of the languages with an initial pharyngeal ḥ-. Examples are:

‘father’ ‘hand’ ‘ear’

Mehri

Ḥarsusi9

Hobyot

Jibbali

Soqoṭri

ḥayb ḥayd ḥayðēn

ḥayb ḥayd ḥayðēn

ḥīb ḥīd ḥayðēn

ī (< *ʾabī) id iðɛ́n

íifǝ(ʾǝ)d ídhɛn

In the languages that have a definite article, namely Omani Mehri and Ḥarsusi, this initial ḥ- shows up sometimes only as the definite form. 6  Cf. also Baṭḥari hum ‘name’, haba ‘seven (f.)’, and xāhǝr ‘old man’. 7  Nakano (2013) has hāḥǝr (pl. hīḥar) in the main entry in his lexicon (p. 101), but hāxǝr (pl. hīxār) elsewhere (e.g., pp. 131, 280). The forms with ḥ are either variants or, more likely, errors in transcription. 8  Soqoṭri x > ḥ is a regular sound change in Eastern Soqoṭri. Soqoṭri forms cited herein are Eastern, unless otherwise noted. 9  Cf. Baṭḥari ḥayb ‘father’.

315

The Classification of Hobyot

‘(the) road’ ‘(the) moon’ ‘(the) sun’ ‘(the) women’

Mehri10

Ḥarsusi11

Hobyot

Jibbali

Soqoṭri

ḥōrǝm ḥārīt ḥǝyáwm ḥaynīt

ḥōrǝm ḥārēt ḥǝyōm ḥānīt

ḥōrǝm ḥārēt ḥǝyūm ḥaynāt

ɔ́ rǝm ɛrét ɛyūm ínɛ́t

óʾorǝm12 é(ʾ)re [šɔm] [ǝghétǝn]

The prefixed ḥ- of these Mehri and Ḥarsusi definite forms stands in contrast to the normal definite article, which is a prefixed a (corresponding to Jibbali ɛ-). Hobyot has no definite article, but still these nouns, and many others like them, have the prefixed ḥ. The only difference is that the ḥ- has become part of the base, as with Mehri nouns like ḥayb and ḥayðēn, noted above. This prefix, which was treated in an important article by Alexander Sima (2002), is a feature again common to Mehri and its dialects and Hobyot, in contrast to Jibbali and Soqoṭri. It should be added that etymological initial ḥ- is unchanged in all the languages, as is illustrated in the following table:

‘cold(ness)’ ‘male in-law’ ‘he dug’ ‘he milked’

Mehri

Ḥarsusi

Hobyot

Jibbali

Soqoṭri

ḥǝbūr ḥaym ḥfūr ḥǝlūb

ḥǝbūr ḥaym ḥfōr ḥǝlōb

ḥǝbōr ḥīm ḥfōr ḥǝlōb

ḥōr ḥim ḥfɔr ḥɔlɔ́ b

ḥébhor ḥam ḥfor ḥélob

10  The Mehri forms listed in this table are all definite. The indefinite forms are wōrǝm (root ʾrm), rīt, yawm, and yǝnīṯ. 11  The Ḥarsusi forms listed in this table are all definite. The indefinite forms correspond to those of Mehri, with minor differences. Cf. Baṭḥari ḥārīt ‘the moon’ and ḥǝyōm ‘the sun’. 12  Western Soqoṭri has hórǝm ‘road’ and hére ‘moon’ (Julien Dufour, p.c.), but this appears to reflect a different sound change than the one found in Mehri, Ḥarsusi, and Hobyot.

316

Rubin

3 Pronouns Following are the forms of the independent pronouns in the six MSA languages:13 Mehri

Ḥarsusi

Baṭḥari

Hobyot

Jibbali

Soqoṭri

hō hēt hēt (OM/WYM)

hō hēt hēt

hō hēt hīt

hō hēt hīt

he hɛt hit

hó(hon) ɛ́(hen)/hɛt í(hin)/hit

3 M SG 3 F SG 1C D 2C D 3C D 1 C PL 2 M PL

hē sē kay/kīh ətay/tīh hay/hīh nḥā ətēm/tām

hē sē ǝtī ǝtī hī nḥā ǝtōm

hē sē ? ? ? nḥā ǝtō

hē sē tīh tīh hīh nḥā tom

šɛ sɛ (ə)s̃i (ə)ti ši nḥa(n) tum

2 F PL 3 M PL 3 F PL

ətēn/tān hēm/hān sēn/sān

ətēn hōm sēn

ətēn hō sēn

tan hom san

tɛn šum sɛn

yhé(hǝn) sé(hǝn) kí(hǝn) tí(hǝn) yhí(hǝn) ḥán(hǝn) tim/tin/ tǝ́n(hǝn) tǝ́n(hǝn) yhǝ́n(hǝn) sǝn

1 C SG 2 M SG 2 F SG

/hīt (EYM)

There are no special innovations among the independent pronouns that speak to any connection Hobyot might have to the other languages, other than the shift of *s to h found in most third person forms. There are some interesting changes in individual languages, like the loss of gender distinction in the second person singular in most Mehri dialects, the levelling of the vowels in the Mehri second and third person plural forms, the vowel change in the Jibbali first person singular pronoun, and the longer forms of Soqoṭri, but these reflect changes internal to each language. The demonstrative pronouns are as follows:14

13  Forms separated by slashes reflect dialectal variants. 14  Slashes indicate variant forms.

317

The Classification of Hobyot Mehri

Ḥarsusi

Hobyot15

Jibbali16

Near M SG

ðōməh / ðɛh

ðānǝh (ðānuh)

ðɛ́nu

Near F SG

ðīməh / ðīh

ðīnǝh (ðínuh)

ðínu

Near C PL

əlyōməh

lānǝh (lanúh)

iźɛ́nu

Far M SG Far F SG Far C PL Remote M SG

ðɛn / ðánǝmǝh / ðɛh ðǝ́nnǝh / ðǝ́nǝmǝh / ðī ǝlǝ́nǝmǝh / lǝ́lǝmǝh — — — ðákǝmǝh / ðǝk

— — — ðákəməh / ðēk / ðǝk ðə́kəməh / ðayk / ðə́kəməh / ðīk ðīk əlyákəməh / əlyēk ǝ́lǝmǝh / lǝk

ðēk ðīk lōk ðówwǝh (ðōh)

ðɔ́ hun ðúhun/ðíhun iźɔ́ hun ðɔ́ kun / ðǝk

ðúwwǝh (ðūh)

ðíkun / ðúkun(u)/ ðik iźɔ́ kun / iźɔ́ k

Remote F SG Remote C PL

lówwǝh (lōh)

In the abovementioned article by Werner Arnold (1993), he compared the Hobyot demonstratives with Mehri and Jibbali, and based on just those three languages Hobyot does indeed look a lot more like Jibbali than Mehri. In Mehri the near demonstratives have m in the singular (ðōmǝh/ðīmǝh), while Jibbali and Hobyot have n.17 However, if we look outside of Mehri proper, at Ḥarsusi and Baṭḥari,18 we see that these languages also have n. So we can assume that Mehri has innovated, and that the correspondence between Hobyot and the remaining four MSA languages is a shared retention. A more interesting similarity between Hobyot and Jibbali is the existence of a three-way series of demonstratives. If Nakano is to be trusted, then the Hobyot far demonstratives are cognate with the Jibbali remote series, while the Hobyot remote series is cognate with the Jibbali far series. I am a little suspicious. Regardless, the fact that both languages have a three-way series sets them apart. But it is not clear if this should be considered a shared innovation, a retention, or an areal feature. Moving to the interrogatives (pronouns and adverbs), we see some variation in the forms. The full set of forms is: 15  Hobyot forms in parentheses come from Arnold (1993). 16  Jibbali ź, a voiced lateral fricative, is an allophone of l. 17  Cf. also Soqoṭri dɛ́nʿa ‘this (ms)’ (Leslau 1938: 131; but Naumkin et al. 2014b: 523 ‘that’). 18  Cf. Baṭḥari ðɛn ‘this’ (Thomas 1937: 97).

318

who? what? where? how? how many?

Rubin Mehri

Ḥarsusi

Baṭḥari

Hobyot

Jibbali

Soqoṭri

mōn hɛ̄śǝn ḥõ / hõ hībō kǝm

mōn háśǝn ḥōnǝh hǝbō kǝm

mān (h)ínɛ ḥānǝh hābō ?

mōn iníh b-hóh / wōr / hóh-ṭoh hǝbō kam

mun mɔn ínɛ́ inɛ́m hun / húṭun ó(ʾo) yɔl / yɔh19 ifú(l) mśé kam

If we just compare Mehri, Jibbali, and Hobyot, the word for ‘what?’ clearly groups Hobyot with Jibbali and Soqoṭri. However, when we include Ḥarsusi and Baṭḥari, we can see that there is not a clear connection between Hobyot and Jibbali and Soqoṭri. The only thing that links Hobyot with Jibbali is the longer form of the interrogative ‘where?’ with the element ṭoh or ṭun, which is probably an areal feature. 4 Numerals The cardinal numerals are very similar among all the languages and do not add to any discussion on classification. The ordinals are more interesting. Compare the following forms:

1st 2nd 3rd 4th

Mehri (m)20

Ḥarsusi (m)

Hobyot (m)

Jibbali (m)

Soqoṭri21

ḥāwǝláy/ ḥāwīl mǝšēġǝr/ tōni/mǝtálli śōlǝt rōbaʾ

ḥāwīl

ḥawlí

ɛnfí

néśhǝr22

məs̃áġər

máṣif

śólət (=cardinal)

? ?

mǝšēġǝr mǝtálli /tēni/mǝtálli śēlǝt/śēlǝś ? rēbaʾ rōbǝʾ

19  Jibbali yɔ(l) and Soqoṭri ifú(l) are cognate with hībō, etc. 20  Only some dialectal variants have been indicated. Cf. also Baṭḥari ḥāwīl ‘first’. 21  Ordinals beyond ‘2nd’ in Soqoṭri are apparently formed by the genitive exponent di- + cardinal, though further investigation is needed. The forms for ‘1st’ and ‘2nd’ come from Leslau (1938), but have been confirmed by Leonid Kogan (p.c.). 22  This is derived from the verb néśor ‘get ahead’ (Leslau 1938: 278).

319

The Classification of Hobyot

5th 6th 7th 8th 9th 10th

Mehri (m)

Ḥarsusi (m)

Hobyot (m)

Jibbali (m)

Soqoṭri

xōmǝs/ xōmah sōdǝs sōbaʾ/hōbaʿ tōmǝn tōsaʾ ōśǝr/áyśǝr

xāmǝh

xōmǝh

(=cardinal)

?

hēt hēbaʾ tēmǝn tēsaʾ āśǝr

? ? ? ? ?

(=cardinal) (=cardinal) (=cardinal) (=cardinal) (=cardinal)

? ? ? ? ?

Mehri and Ḥarsusi have ordinal numerals from one through ten, and, based on the limited forms found in Nakano’s lexicon, Hobyot seems to as well. Mehri, Ḥarsusi, and Hobyot also share the same root used for ‘first’, and the Hobyot word for ‘second’ has cognates in Mehri and Ḥarsusi. Jibbali and Soqoṭri, on the other hand, use unique roots for ‘first’, and Jibbali (probably also Soqoṭri) has special ordinal forms only for ‘first’, ‘second’, and ‘third’. So in the ordinal system, Hobyot aligns lexically and structurally with Mehri, against Jibbali and Soqoṭri. 5 Verbs Verbal morphology is an area where we see some of the clearest data relevant to classification. First of all, the innovation of a future tense is a feature that can be used in classifying the languages. Compare the following future-tense constructions: Mehri sǝyūr ‘he went’, sīrōna ‘he will go’ (future ← active participle in *-ān) Hobyot sǝyōr ‘he went’, mádǝh yīsǝr ‘he will go’ (future ← particle mad- + subjunctive) Jibbali aġád ‘he went’, (d)ḥa-yġád ‘he will go’ (future ← particle (d)ḥa- + subjunctive) In Mehri, as also in Ḥarsusi, and Baṭḥari, the future tense is expressed with a form that derives from an earlier participial form, a nominal form with a suffix *-ān(a).23 Forms like sīrōna, given above, behave like nouns in that they decline only for number and (usually) gender, but not for person. In Hobyot, the 23  See Rubin (2007) for further discussion on the development of this form in Mehri.

320

Rubin

future is expressed with a particle mad- plus a dependent verb, like madǝh yīsǝr ‘he will go’. In Jibbali also, the future is formed with a particle plus a dependent verb. Soqoṭri has no explicit future. Mehri, Ḥarsusi, and Baṭḥari are linked by a shared innovation, while Hobyot and Jibbali exhibit their own independent innovations. The Jibbali particle (d)ḥa- likely derives from a verb meaning ‘want’,24 and it has been credibly suggested that the Hobyot particle mad comes from a word meaning ‘wish’ or ‘intention’.25 In fact, the Mehri verb ḥōm ‘want’ can be used to indicate a future (alongside the other future construction) in some Yemeni dialects. So it is possible that a ‘want’-future developed as an areal phenomenon, though the fact that such developments are so common cross-linguistically certainly does not make this necessary. Even if so, the morphological and lexical differences between the languages would still make clear that Hobyot has developed independently of all the other languages with regard to this future tense construction. While a comparison of the future tense constructions make clear that Hobyot is neither Mehri nor Jibbali, some other features of verbal morphology link Hobyot with Mehri. Consider the following conjugations of a G-Stem (basic stem) verb in Mehri, Jibbali, and Soqoṭri in the perfect, subjunctive, and imperative.26 G-Stem Perfect

1 C SG 2 M SG 2 F SG 3 M SG 3 F SG 1 C PL 2 M PL 2 F PL 3 M PL 3 F PL

Mehri

Jibbali

Soqoṭri

ḳədə́rk ḳədə́rk ḳədə́rš ḳədūr ḳədərūt ḳədūrən ḳədə́rkəm ḳədə́rkən ḳədáwr (OM) / ḳǝdūrǝm (YM) ḳədūr

ḳɔdɔ́ rk ḳɔdɔ́ rk ḳɔdɔ́ rs̃ ḳɔdɔ́ r ḳɔdɔrɔ́ t ḳɔdɔ́ rən ḳɔdɔ́ rkum ḳɔdɔ́ rkən ḳɔdɔ́ r ḳɔdɔ́ r

ḳédork ḳédork ḳédorš ḳédor ḳedɛ́ro ḳedórən ḳedórkən ḳedórkən ḳédər ḳédor

24  See further in Rubin (2012). 25  Lonnet (1994: 228–30). 26  The verb means ‘be able’ in Mehri and Jibbali, but ‘cook’ in Soqoṭri.

321

The Classification of Hobyot G-Stem Subjunctive

1 C SG 2 M SG 2 F SG 3 M SG 3 F SG 1 C PL 2 M PL 2 F PL 3 M PL 3 F PL

G-Stem Imperative27

Mehri

Jibbali

Soqoṭri

ləḳdēr təḳdēr təḳdēri yəḳdēr təḳdēr nəḳdēr təḳdērəm təḳdērən yəḳdērəm təḳdērən

lɔ́ ḳdər tɔ́ ḳdər tíḳdir yɔ́ ḳdər tɔ́ ḳdər nəḳdɛ́r təḳdɔ́ r təḳdɛ́rən yəḳdɔ́ r təḳdɛ́rən

ləḳdɛ́r təḳdɛ́r təḳdír liḳdɛ́r təḳdɛ́r nəḳdɛ́r təḳdə́r təḳdɛ́rǝn liḳdə́r təḳdɛ́rǝn

Mehri

Jibbali

Soqoṭri

ḳədēr ḳədēri

ḳədɛ́r ḳədír

ǝḳdɛ́r ǝḳdír

ḳədērəm ḳədērən

ḳədɔ́ r ḳədɛ́rən

ǝḳdǝ́r ǝḳdérǝn

Perhaps the most important difference among the paradigms is that, in Mehri, there is a suffix -m used in various forms, namely, the 3.M.PL perfect, 2/3.M.PL subjunctive (likewise in the imperfect, not shown here), and in the M.PL imperative. In Yemeni Mehri, the 3.M.PL perfect always has the suffix -m, while in Omani Mehri, the suffix is found only with certain conjugations and root types. In the imperfect, subjunctive, and imperative, the dialects all have -m in the abovementioned forms. Where Mehri has this -m, Jibbali and Soqoṭri have ablaut or nothing. And Hobyot behaves just like Mehri, specifically like Yemeni Mehri, as will be seen below.28 When it comes to the peculiarities associated with the conjugation of verbs with weak root consonants, Hobyot usually patterns with Mehri. For example, consider the following forms of a final-weak (III-w/y) verb, especially the 3.M.SG forms:

27  Despite contrary claims by some scholars (most recently, Simeone-Senelle 2011: 1096), an imperative does exist in Soqoṭri. See Naumkin et al. (2014a: 46–47). 28  Ḥarsusi also behaves like Mehri. Baṭḥari data from Thomas (1937) suggests that it actually does not, but Thomas’s data can be unreliable at times.

322

Rubin

G-Stem root ḳrV ‘hide’:

Omani Mehri Yem. Mehri/  Hobyot Ḥarsusi Jibbali Soqoṭri

3 M SG PERF

3 M PL PERF

1 C SG PERF

3 M SG IMPF

3 M SG SUBJ

ḳǝrō

ḳǝrīw

ḳōrǝk

yǝḳáyr

yǝḳrē

ḳǝrū

ḳǝrīyǝm

ḳōrǝk

yǝḳīr

yǝḳrē

ḳǝrō ḳéré ḳárǝ

? ḳéré ḳárǝ

ḳārǝk ḳórǝk ḳárǝk

yǝḳáyr yǝḳɔ́ r yíḳar

yǝḳrē yéḳǝr liḳrɛ́

An initial root consonant w has an effect on all verb forms in Jibbali and Soqoṭri, but behaves mostly as a strong consonant in Mehri and Hobyot. Compare, for example, the following table, where the root consonant w is lost in all Jibbali and Soqoṭri forms, but only the G-Stem subjunctive in Mehri, Hobyot, and Ḥarsusi: G-Stem root wḳf ‘be silent’:

Omani Mehri29 Hobyot Ḥarsusi Jibbali Soqoṭri

3 M SG PERF

3 M PL PERF

1 C SG PERF

3 M SG IMPF

3 M SG SUBJ

wǝḳáwf wǝḳōf 30 wǝḳōf (e)ḳɔ́ f ʾö́ḳaf

wǝḳáwf wǝḳōfǝm wǝḳōfǝm (e)ḳɔ́ f ʾéḳǝf

wǝḳǝ́fk wǝḳōfk wǝḳōfk (e)ḳɔ́ fk ʾö́ḳafk

yǝwḳōf yǝwḳōf yǝwḳōf yéḳɔ́ f yóuḳaf

yāḳāf yǝḳāf yāḳāf yǝḳɛ́f liḳáf

29  Among the Mehri dialects, certain types of I-w verbs show some variation in the loss or retention of w. 30  Whether or not Hobyot really has long vowels in the verb forms presented in this and the following tables remains to be conclusively determined. The issue has no bearing on the arguments of this paper.

323

The Classification of Hobyot

When w is the second root consonant, it also often behaves as a strong consonant in Mehri and Hobyot, as one can see in the derived Š2-Stem verb and in the doubly-weak G-Stem verb in the following two tables. Š2-Stem root nwḥ ‘quarrel’:31

Mehri Hobyot Ḥarsusi Jibbali Soqoṭri

3 M SG PERF

3 M PL PERF

1 C SG PERF

3 M SG IMPF

3 M SG SUBJ

šǝnēwǝḥ šǝnāwǝḥ šǝnēwǝḥ šenīḥ šéneḥ

šǝnáwḥǝm šǝnáwḥǝm

šǝnáwḥǝk

yǝšnáwḥǝn yǝšnáwḥǝn

yǝšnēwǝḥ yǝšnāwǝḥ

yǝšeníḥǝn yǝšnɛ́yḥǝn

yǝs̃enáḥ lišénaḥ

šenīḥ

G-Stem root twy ‘eat’:

Omani Mehri Hobyot Jibbali Soqoṭri

3 M SG PERF

3 M PL PERF

1 C SG PERF

3 M SG IMPF

3 M SG SUBJ

tǝwō tǝwū tē tǝ

tǝwīw tǝwīyǝm tē tǝ

tūwǝk tūk/túwǝk tɛ̄k toyk

yǝtáyw yǝtū yǝté yítǝ

yǝtē yǝtē yit or yǝtɛ́ litɛ́

The preservation of w in Mehri and Hobyot is most likely a shared retention, rather than a shared innovation, but at least it is evidence for the subgrouping of Jibbali and Soqoṭri, to the exclusion of Hobyot. In the next table, we see another link between Jibbali and Soqoṭri, namely the irregular loss of r in the verb ‘to say’:

31  The meaning varies slightly in the languages.

324

Rubin

G-Stem root ʿmr ‘say’:

Omani Mehri Sharqīyah  Yemeni Mehri Hobyot Ḥarsusi Jibbali Soqoṭri

3 M S PERF 3 M PL PERF

1 C SG PERF

3 M SG IMPF

3 M SG SUBJ

āmōr32

āmáwr

āmǝ́rk

yōmǝr

yaʾmēr

ʿamōr

ʿamōrǝm

ʿamǝ́rk

yǝʿōmǝr

yaʿmēr

ʾǝmōr ʾamōr ʿõr ʿámor

ʾǝmōrǝm ʾamōrǝm ʿõr ʿémǝr

ʾǝmōrǝk ʾamōrǝk ʿõk33 ʿámok34

yǝʾōmǝr yōmǝr yǝʿõr yaʿámǝr

yǝʾmēr yaʾmēr yáʿmɛr laʿmɛ́r

In the causative stem, there seems to be another connection that links Jibbali and Soqoṭri, namely the vowel alternation in the base of the perfect, and the full vowel after the first root consonant in the imperfect: H-Stem (various roots): 3 M SG PERF

3 M PL PERF

1 C SG PERF

3 M SG IMPF

3 M SG SUBJ

Omani Mehri Yemeni Mehri

hǝrkūb hǝrkūb

hǝrkǝ́bk hǝrkǝ́bk

yǝhǝrkūb yǝhǝrkūb

yǝhárkǝb yǝhárkǝb

Hobyot Ḥarsusi Jibbali Soqoṭri

axxǝrōj arkōb (ɛ)flét éḳdǝm

hǝrkīb hǝrkūbǝm / hǝrkībǝm axxǝrōjǝm

axxǝrōjǝk

(ɛ)flét éḳdǝm

(ɛ)flɔ́ tk éḳdomk

yaxxǝrōj yarkōb yǝffélɔ́ t yǝḳádom

yáxrǝj yǝhárkǝb yɛ́flǝt láḳdɛm

There are, however, cases in which Hobyot patterns with Jibbali and Soqoṭri, against Mehri and Ḥarsusi. In the Gb-Stem, the equivalent of the Arabic faʿila type, the imperfect and subjunctive have fallen together in Mehri and Ḥarsusi, 32  ʿ > ʾ or ∅ is regular in Omani Mehri, Hobyot, and Ḥarsusi. 33  Irregular loss of r is normal in all 1/2 singular forms, sometimes in 1/2 dual and 2 plural. Loss of intervocalic m is regular in Jibbali. 34  Irregular loss of r is normal in all 1/2 forms, except 1.C.PL.

325

The Classification of Hobyot

but remain distinct in Hobyot, Jibbali, and Soqoṭri. This is a shared innovation of Mehri and Ḥarsusi, but a shared retention among the others. An example is: Gb-Stem root ġẓ́n ‘feel compassion for; love’:

Mehri Hobyot Ḥarsusi Jibbali Soqoṭri

3 M SG PERF

3 M PL PERF

1 C SG PERF

3 M SG IMPF

3 M SG SUBJ

ġáyẓ́ǝn ġīẓ́ǝn ġáyẓ́ǝn ġéẓ́ǝn ʿéẓ́an/ ʿö́ẓ́an

ġǝ́ẓ́ǝnǝm ġīẓ́ǝnǝm

ġǝ́ẓ́nǝk

ġéẓ́ǝn ʿéẓ́ǝn

ġéẓ́ǝnk ʿóẓ́ank

yǝġẓ́ōn yǝġaẓ́ōn yǝġẓ́ōn yǝġaẓ́ún yiʿáẓ́on

yǝġẓ́ōn yǝġẓ́ōn yǝġẓ́ōn yǝġẓ́ún laʿẓ́ón

Each language, of course, has its own irregular and anomalous verbs, but there are some interesting examples where a verb is anomalous in one language, but totally regular in the other. Such are:

‧ Jibbali s̃éf ‘he slept’ < *s̃sé̃ f < *s̃ǝs̃éf < *s̃ǝws̃éf < Š1-Stem *s̃ǝwkéf (compare Mehri and Hobyot Š1 šǝwkūf ) ‧ Mehri šfūḳ ‘they (F) married’ < Š1-Stem *šǝhfūḳ < *šǝsfūḳ (compare Jibbali G šfɔḳ and Š1 s̃ǝšféḳ; Hobyot Š1 šǝhfōḳ) ‧ Mehri (t)tǝḳ(ḳ) ‘drink’, Jibbali šús̃ị (< *šúts̃i < *šútḳi), both T-Stems of the root *sḳy (> Mehri *hḳy; Jibbali šḳy). Compare Hobyot hútḳi ‘drink’.

6 Auxiliaries Another important similarity between Hobyot and Mehri can be seen in the forms of two common auxiliaries. These are ber, which, among other things, can mean ‘already’, and ʿad, which, among other things, can mean ‘still’ or ‘yet’.35 These two auxiliaries are conjugated like verbs in Jibbali and Soqoṭri, but decline like nouns in Mehri and Hobyot. As can be seen in the table below, the first person singular forms in Jibbali and Soqoṭri have the suffix k, like verbs, while those of Mehri and Hobyot have -i, like a noun with a possessive suffix. Jibbali and Soqoṭri also share some irregularities, namely the loss of r in the first and second person singular forms of ber, and the loss of d in all first and 35  On ber, see Rubin (2010: 248–51; 2014: 164–68). On ʿad, see Rubin (2010: 241–43; 2014: 168–71) and Watson (2012: 126–27).

326

Rubin

second person forms of ʿad except the 1.C.PL. So, again, Hobyot patterns clearly with Mehri, in this case with a probable shared innovation, while other shared innovations pair Jibbali and Soqoṭri.

1 C SG 2 M SG 2 F SG 3 M SG 3 F SG 1/2 C D 3M D 3F D 1 C PL 2 M PL 2 F PL 3 M PL 3 F PL

Jibbali

Soqoṭri36 Mehri37

Hobyot

Jibbali

Soqoṭri

Mehri38

Hobyot

bek bek bis̃ ber berɔ́ t bérs̃i berɔ́ bertɔ́ bérǝn bérkum bérkǝn ber ber

bǝ(r)k bǝ(r)k bǝš ber bíroh bíki bíroh birɛ́toh bérǝn bé(r)kǝn bé(r)kǝn ber ber

béri bérǝk bérǝš bérǝh bérǝs ? ? ? ? ? ? bérhǝm bérsǝn

ʿɔk ʿɔk ʿɔs̃ ʿɔd ʿɔt ʿɔ́ sĩ ʿɔdɔ́ ʿɔdtɔ́ ʿɔ́ dǝn ʿɔ́ kum ʿɔ́ kǝn ʿɔd ʿɔd

ʿak ʿak ʿaš ʿad ʿédoh ʿak ʿédoh ʿedɛ́toh ʿádǝn ʿákǝn ʿákǝn ʿad ʿad

ʾādi ʾādǝk ʾādǝš ʾādǝh ʾādǝs ? ? ? ʾādǝn ʾādkǝm ʾādkǝn ʾādhǝm ʾādsǝn

ʾādi ʾādǝk ʾādǝš ʾādǝh ʾādǝs ʾadki ? ? ʾādǝn ? ? ʾādhǝm ʾādsǝn

bǝ́ri bǝrk bǝrš bǝ́rǝh bǝrs bǝ́rki bǝ́rhi bǝ́rhi bǝ́rǝn bǝ́rkǝ bǝ́rkǝn bǝ́rhǝ bǝ́rsǝn

7 Lexicon Lexicon is not the most reliable source for determining historical relationships between languages, but it can nevertheless be reflective of such. In comparing a list of basic nouns, Hobyot patterns with Mehri in the majority of cases, though there are some cases where it patterns with Jibbali, or even has a unique word. In short, the morphological evidence suggests a relationship between Hobyot and Mehri, and the lexical evidence supports this idea. Compare the following list of common nouns:

36  Soqoṭri forms in this table were taken both from the glossary of Naumkin et al. (2014b) and from the unpublished field data of Antoine Lonnet. The two sources differ slightly in the transcription of some vowels, but these differences are not relevant for the arguments herein. 37  Ḥarsusi behaves like Mehri. 38  Ḥarsusi behaves like Mehri.

327

The Classification of Hobyot Mehri

Ḥarsusi

Hobyot

Jibbali

Soqoṭri

ʿouyɛ́ghǝn / mǝ́brǝhe (pl. ouyǝ́ghon/ ǝmbórye) śíbib śḥaf hélob

‘boy’

ġiggēn (pl. ġǝggēn (pl. ǝmbǝráwtǝn) ǝmbǝrǝ́ttǝn)

ġajjēn (pl. ḳalyūn)

ǝmbérɛʾ (pl. ɛrśɔ́ t)

‘old woman’ ‘milk’ ‘buttermilk/ whey’ ‘town’

āgáwz śxōf ātǝrēt

āgǝz39 śxōf ātǝrēt

ʾajūz śxōf ḥamẓ́

šxarét nuśǝb ḥalɔ́ b

rǝḥbēt

rǝḥbēt

rǝḥbēt

‘donkey’ ‘snake’

ḥayr rēśīt

ḥayr rǝśēt

‘cow’ ‘chest’ ‘stone, rock’

bǝḳǝrēt gawf ṣāwǝr

bǝḳǝrēt gawf ṣǝ́wwǝr/ kēbǝt

ḥallɛ́t or (ṣ̌irét?) ṣ̌irɛ́t ḥayr ḳéraḥ šánḳahal reśēt (gen.)/ hɔ̄ t (gen.) bǝkéle hábyǝt (big, black) lē léʾ ǝ́llǝhe jōf géhɛ gɛ́hɛ ṣō(wǝ)r fúdún/dɔf fǝ́dhon/ ó(ʾo)bǝn

A bit more instructive than common nouns are functional words, like the words associated with time in the following table: Mehri

Ḥarsusi

Baṭḥari

Hobyot

Jibbali

‘(in) morning’ k-sōbǝḥ ‘(at) night’ bǝ-ḥǝlláy

k-sōbǝḥ bǝ-ḥǝlláy

? ?

k-sábḥa bǝ-ḥallē

k-ḥáṣaf ġasré

yǝllīlǝh gēhǝmǝh yǝllōh sǝnēt / ḥawl l-āyūmǝn

ǝllīlǝh gēhǝmǝh yǝllōh sǝnēt / ḥawl l-ōm

? gēhǝmǝh ? sǝnēt/ ḥawl fǝnnɛ́-l-ōm

allēlǝh jáhma allóh sǝnēt / ḥōl l-ʾōm

‘tonight’ ‘tomorrow’ ‘last night’ ‘year’ ‘last year’

39  Cf. also Baṭḥari ʿagūz.

Soqoṭri

ṣá(b)ḥ bǝ-lǝ́lhe/ ḥte ǝl-ʿéni ? ḳǝrérɛ ḳarére mənhínəm lílin ʿónut / ʿéno ḥabl berhón birǝhan

328

Rubin

(cont.) Mehri

Ḥarsusi

Baṭḥari

Hobyot

‘today’

yǝmō(h)

yǝmō(h)

ḥor

‘day’

nǝhōr nǝhōr / (pl. yūm) ḥǝyōmǝt (pl. yīm)

ḥor / axxór šḥór / šḥér ḥer (< hxór < sḥr?) ḥǝyūm yum yom / šɔm (pl. ḥǝyēm) (pl. ɛ̄m) (pl. énhor)

yōm / yáwmǝt

Jibbali

Soqoṭri

Here Hobyot again usually patterns with Mehri, Ḥarsusi, and Hobyot, and where it doesn’t pattern with Mehri, like in the words for ‘today’ and ‘day’, it still patterns with Baṭḥari. A comparison of basic verbs show that again Hobyot clearly patterns with Mehri and Ḥarsusi in nearly all of the most common verbs, including even the verb ‘to be’: Mehri

Ḥarsusi

Baṭḥari

Hobyot

Jibbali

Soqoṭri

‘be’ ‘go’ ‘sit, stay’ ‘come’

wīḳa sǝyūr śxǝwlūl nūka

wēḳa sǝyōr śxǝwlōl nōka

? sǝyɛ̄r śxāwal nōka

wīḳǝʾ sǝyōr śxōlūl41 nūkǝʾ

kɔn ʿod/ṭáhɛr ézʿǝm42 gédaḥ/éraḥ

‘do’

áymǝl

amōl

?

ʾīmǝl

kun40 aġád skɔf zǝḥám/ níkaʿ s̃érék

šégɛ/ ʿémor43

40  The G-Stem of kwn is, in fact, used in Mehri, Ḥarsusi, and Hobyot, but in the imperfect only, and with limited functions. 41  Other researchers have apparently recorded skof. 42  Julien Dufour has recorded Western Soqoṭri tśxɔl ‘sit down!’, a relic of the root found in Mehri, Ḥarsusi, Baṭḥari, and Hobyot. 43  The basic meaning of the verb ʿémor (root ʿmr) is ‘say’, and is not connected with the root ʿml ‘do’ found in Mehri, Ḥarsusi, and Hobyot (which also have the root ʿmr ‘say’). The Jibbali verb ʿõr ‘say’ can also mean ‘do’, though it is not common; see the comment to text 49: 18 in Rubin (2014).

329

The Classification of Hobyot Mehri

Ḥarsusi

Baṭḥari

Hobyot

Jibbali

Soqoṭri

‘slaughter’ sǝḥāṭ ‘pass the day’hǝḳṣáwm ‘stay the hātūm night’ ‘die’ mōt ‘carry, lift’ śǝl

sǝḥāṭ aḳṣáwm ātōm

? ? ?

sǝḥāṭ aḳṣōm aʾtōm

ḥez ḳǝhɛ́b aġsǝré

ṣélob/ḥez ḳóhob ?

mōt śǝl

? ?

mōt śǝl

xarɔ́ g ḥõl

‘be hungry’ gáwya ‘run’ bǝḳōẓ́,

géwya bǝḳōẓ́, abárḳa yǝxōm wǝkōb ġátri

?44 baḳāẓ́

jíyǝʾ bǝḳōẓ́

telf šaʿé

ṣámǝ ʿédol/zégod/ ḥémol ṣéṭaʿ šéʿe

ḥām ? ġátri

yǝxxōm wǝkōb ġútri

ʿágǝb45 égaḥ herɔ́ g

? ? ?

axxǝṣōb rḥū ṣǝxāb

ʿer rḥē ṣǝxáb

?

kǝmūz, ðǝlōf

kũz, dɔlɔ́ f, ðɔlɔ́ f

‘want’ ‘enter’ ‘speak’

abárḳa yǝḥōm wǝkūb ġátri

‘send’ ‘lick’ ‘have pain’

xǝṣáwb lāt bǝxāṣ

‘jump’

xǝṣáwb lāt ṣǝxáb/ bǝxāṣ āfōd, dǝlūf afōd, ðǝlōf

ʿégǝb ékob/égaḥ šémtǝl/ ʿotérǝ (ʿ < ġ) bélog/éṭbaḳ réḥe/ḳáḥaṣ éẓ́aḥ ḳélom, ḳéroy, sábaḥ, yhálaḥ

8 Prepositions Each language, and some dialects, have slightly different system of prepositions, both in terms of form and usage. A comparison of those systems show some additional evidence of the relationship of Hobyot with Mehri and Jibbali with Soqoṭri. For example, Mehri, Ḥarsusi, and Hobyot share the word bǝ(r)k ‘inside’, which is not present in Jibbali or Soqoṭri. Jibbali uses the innovative preposition ʿaḳ (grammaticalized from ʿamḳ ‘middle’), while Soqoṭri uses the inherited Semitic bǝ-. 44  Compare Baṭḥari giʾan ‘hungry’, from the same root. 45  Jibbali ʿágǝb and Soqoṭri ʿégǝb also (and originally) mean ‘like, love’; compare Mehri and Ḥarsusi áygǝb, Hobyot ʾījǝb ‘like, love’.

330

Rubin

The syntax of the prepositions can differ as well. In comparison of adjectives, Mehri, Ḥarsusi, and Hobyot all use the preposition mǝn ‘from’ as the equivalent of English ‘than’, while Jibbali and Soqoṭri use ʿan,46 even though mǝn exists in both of those languages. Moreover, this preposition ʿan is common only to Jibbali and Soqoṭri, and is unknown in the other languages. Mehri, Ḥarsusi, Hobyot, and Jibbali all share the word *(ǝl-)his ‘like’. This word does not behave like a true preposition, since a pronominal “object” is indicated not with a pronominal suffix, but rather with a following direct object pronoun t- in Mehri, Ḥarsusi, and Hobyot, and with a following independent pronoun in Jibbali. Soqoṭri shares the same structure as Mehri, Ḥarsusi, and Hobyot, but uses a different lexeme for ‘like’. So in this way of expressing comparison, Mehri, Ḥarsusi, and Hobyot again form a group. These prepositional data are summarized in the following table: Mehri

Ḥarsusi

Hobyot

Jibbali

Soqoṭri

‘in(side)’

bǝrk47

bǝrk

bak (bark-)

ʿaḳ

‘better than’ ‘like’+PRN OBJECT ‘like me’

axáyr mǝn (ǝl-)hīs t-

axáyr mǝn hǝs t-

axáyr mǝn hīs t-

axér ʿar (ʿan-) (ǝl-)hés PRN

bǝ- (or compounds) ʿan ‘than’ tóʾo t-

(ǝl-)hīs tī

hǝs tǝ́ni

hīs tī

(ǝl-)hés hé

tóʾo tho48

9 Semantics Shared semantic development can also provide evidence of relationship, an example of which is the verb ‘to want’ (shown in the table in section 7), which in both Jibbali and Soqoṭri comes originally from the verb that originally means ‘love, like, or desire’ (root ʿgb). Another interesting example of semantic development can be seen in the Jibbali and Soqoṭri adjectives for ‘big’, which 46  In Jibbali, the independent preposition has become ʿar, but the suffixed forms retain the base ʿan-. 47  Yemeni Mehri uses bak or brak for the bare form, but bark- with suffixes (Jahn 1905: 123; Watson 2012: 114). 48  Note also that the Jibbali counterpart of Soqoṭri tho ‘me (direct object)’ is tɔ, also with a back vowel, while Mehri and Hobyot have the suffix -i with a front vowel.

The Classification of Hobyot

331

derive from the inherited Semitic roots for ‘father’ and ‘mother’. Here Hobyot again patterns with Mehri and the others, against Jibbali and Soqoṭri. Soqoṭri: ǝ(ʾǝ)b (m. only; pl. ǝbhétǝn), SHARED STRUCTURE & LEXEME am (f. only, pl. ǝmhétǝn) Jibbali: eb (m. only; pl. ētǝ), um (f. only; pl. emíti or ĩti) SHARED STRUCTURE (AREAL) OM and Ḥarsusi: śōx (m. only; pl. śīyǝx), nōb(ǝt) (f. only; pl. nǝyōb) YM: śōx (pl. śīyāx), f. śaxt (pl. śīyaxtǝn) SHARED LEXEME Hobyot: śōx (pl. śiyāx), f. śoxt (pl. śiyáxtǝ) (Hob. ~ Mehri; Baṭḥari: nōb (pl. niyāb), f. nōbiyǝt (pl. nōbiyāt) Baṭ. ~ Mehri) Alongside the shared semantic innovation of Jibbali and Soqoṭri, we find that the lexeme for ‘big’ used in Hobyot is the same as found in Mehri and Ḥarsusi. More interestingly, Omani Mehri and Ḥarsusi pattern with Jibbali and Soqoṭri in their structure, in that they all use distinct masculine and feminine words for ‘big’. This is an areal development shared by Omani Mehri, Ḥarsusi, and Jibbali, which is separate from the shared semantic innovation that links Jibbali and Soqoṭri. Hobyot patterns with Yemeni Mehri in this case, both lexically and structurally. 10 Conclusion There are a variety of phonological, morphological, and lexical features that suggest a relationship between Hobyot, Mehri, Ḥarsusi, and Baṭḥari, as against Jibbali and Soqoṭri. At the same time, there are a number of features, only some of which are mentioned above, that suggest a relationship between Jibbali and Soqoṭri, against the other languages. So based on the limited Hobyot data that is available at present, the family tree that was suggested above seems likely. Bibliography Arnold, Werner. 1993. “Zur Position des Hóbyót in den neusüdarabischen Sprachen,” Zeitschrift für arabische Linguistik 25: 17–24. Faber, Alice. 1981. “Phonetic Reconstruction,” Glossa 15: 233–262. ———. 1985. “Akkadian Evidence for Proto-Semitic Affricates,” Journal of Cuneiform Studies 37: 101–107.

332

Rubin

Jahn, Alfred. 1905. “Grammatik der Mehri-Sprache in Südarabien,” Sitzungsberichte der Kaiserlichen Akademie der Wissenschaften in Wien, Philosophisch-Historische Klasse 150/6. Johnstone, T. M. 1977. Ḥarsūsi Lexicon and English-Ḥarsūsi Word-List. London. ———. 1981. Jibbāli Lexicon. London. ———. 1987. Mehri Lexicon and English-Mehri Word-List. London. Kogan, Leonid. 2011. “Proto-Semitic Phonetics and Phonology,” in Stefan Weninger et al. (eds.), The Semitic Languages: An International Handbook. Berlin. 54–151. Leslau, Wolf. 1938. Lexique Soqoṭri (sudarabique moderne). Paris. Lonnet, Antoine. 1985. “The Modern South Arabian Languages in the P.D.R. of Yemen,” Proceedings of the Seminar for Arabian Studies 15: 49–55. ———. 1994. “Le verbe sudarabique moderne: hypothèses sur des tendances,” Matériaux arabes et sudarabiques NS 6: 213–55. Nakano, Aki’o (ed. Robert Ratcliffe). 2013. Hōbyot (Oman) Vocabulary: With Example Texts. Tokyo. Naumkin, Vitaly et al. 2014a. “Studies in the Verbal Morphology of Soqotri I/1: Strong Triconsonantal Roots in the Basic Stem (the Analysis),” Zeitschrift für arabische Linguistik 59: 25–56. ———. 2014b. Corpus of Soqotri Oral Literature, Vol. 1. Leiden. Rubin, Aaron D. 2007. “The Mehri Participle: Form, Function, and Evolution,” Journal of the Royal Asiatic Society (Series 3) 17: 381–88. ———. 2010. The Mehri Language of Oman. Leiden. ———. 2012. “Grammaticalization and the Jibbali Future,” in Domenyk Eades (ed.), Grammaticalization in Semitic. Oxford. 193–203. ———. 2014. The Jibbali (Shaḥri) Language of Oman: Grammar and Texts. Leiden. Sima, Alexander. 2002. “Der bestimmte Artikel im Mehri,” in Werner Arnold and Hartmut Bobzin (eds.),“Sprich doch mit deinen Knechten aramäisch, wir verstehen es!” 60 Beiträge zur Semitistik: Festschrift für Otto Jastrow zum 60. Geburtstag. Wiesbaden. 647–68. Simeone-Senelle, Marie-Claude. 2011. “Modern South Arabian,” in Stefan Weninger et al. (eds.), The Semitic Languages: An International Handbook. Berlin. 1073–1113. Thomas, Bertram. 1937. “Four Strange Tongues from Central South Arabia—The Hadara Group,” Proceedings of the British Academy 23: 231–331. Watson, Janet C. E. 2012. The Structure of Mehri. Wiesbaden.

Expression of Attributive Possession in Tunisian Arabic: The Role of Language Contact Lotfi Sayahi 1 Introduction Possession is a universal concept that describes a particular link between two entities. It can denote a set of relationships that range from strict ownership to more loose connections that are open to interpretations. At the semantic level, a major distinction has been established between alienable possession and inalienable possession. Alienable possession refers to relationships that are temporary or which can be ended freely such as ownership of material objects. In the case of inalienable possession the possessed entity is intrinsically linked to the possessor, as in the case of part-whole relations (e.g., body parts) or kinship relations. Languages, however, vary when it comes to what counts as alienable vs. inalienable possession (Heine 1997; Payne and Barshi 1999; Herslund and Baron 2001). At the structural level, two major types of possessive structures exist. The first type, which is not the object of the current study, is predicative possession which uses verb forms to express the possessive link. The second type is attributive possession, also referred to as adnominal possession, which marks the possession in the noun phrase through different morphosyntactic structures. Despite being such a common category, only a few studies have described change in the expression of possession in cases of language contact. Weinreich (1963) in his seminal work mentions the case of Estonian, Amharic, and Modern Hebrew which developed or increased the use of analytic possession constructions as a result of contact with other languages (Weinreich 1963: 41–42). Hickey (2010: 16) makes an argument for the role of Celtic languages in the change in Middle English towards the use of possessive pronouns with inalienable nouns instead of the old Germanic form that marked this type of possession through use of the personal dative. In the United States, a few studies have shown variation in the expression of possession in the speech of Hispanic speakers as a result of contact between Spanish and English (Wolford 2006; Montoya 2011; Orozco 2009). Montoya, in her study of second generation Hispanic immigrants to New York State, found that they extend the use of the possessive adjective to contexts where the Spanish definite article is the unmarked option in non-contact Spanish varieties. She attributes such © koninklijke brill nv, leiden, ���5 | doi ��.��63/9789004300156_018

334

sayahi

extension in the use of the possessive adjective to influence from English. In this paper the objective, then, is to look into the expression of attributive possession in Tunisian Arabic and the possible role that contact with French plays in the additional spread of the analytic form. In Standard Arabic, synthetic constructions are formed by means of the construct state (ʔiḍa:fa): two nouns are juxtaposed, with the second noun being the possessor (N+N) as in (1), or through a suffixed pronominal possessor instead of a noun (N+PRO) as in (2). As mentioned by Eksell Harning (1980: 10), prepositions can also be used to express possession in Standard Arabic but the focus here is on attributive possession within the noun phrase and not on possession in other classes of phrases. (1) kita:bu Salma book Salma ‘Salma’s book’ (2) kita:bu-ha: book-GEN.3.SG.F ‘her book’ Analytic constructions, understood here as constructions that are formed with the genitive exponent as opposed to constructions formed with prepositions, on the other hand, are non-existent in Standard Arabic, although they appear in the majority of the dialects (Eksell Harning 1980; Taine-Cheikh 2010; among others). In Maghrebi Arabic, in addition to the synthetic constructions mentioned for Standard Arabic, there are analytic constructions that are formed through the use of the genitive exponent mtɛ:ʕ, originally a noun meaning ‘possession’, or similar particles such as dyal in Moroccan Arabic, followed either by a lexical or a pronominal possessor: (3) l-ktɛ:b mtɛ:ʕ Salma DEF-book of Salma ‘Salma’s book’ (4) l-ktɛ:b mtɛ:ʕ-ha DEF-book of-3.SG.F ‘her book’ Much importance has been given to the value of the genitive exponent in showcasing the analytic nature of the dialects (Eksell Harning 1980), but also

Attributive possession in tunisian arabic

335

in variation and change across regions and social groups. The argument that will be advanced in this study is that the presence of a considerable number of French loanwords and code-switched noun phrases in the Tunisian dialect, as evidenced in previous work (Sayahi 2011, 2014), contributes to the spread of the analytic form. This represents a case of indirect transfer as opposed to direct morphosyntactic interference given that the analytic from whose usage increases already exists in vernacular Arabic (Silva-Corvalán 2008; Poplack and Levey 2010). It is true that the analytic form has been spreading in Arabic vernaculars in general, but the position that will be adopted here is that lexical insertions from other languages have historically contributed to the acceleration of such a change in the affected dialects. I will start by showing that such a process, an increase in the frequency of the usage of the genitive exponent as a result of lexical borrowing and code-switching, has been documented in previous studies of attributive possession in Arabic in contact with other languages. 2

Attributive Possession in Arabic in Contact

Among the Arabic varieties, Maltese stands out as one that has gone through intense contact with other languages for centuries. Contact with Romance languages, principally Sicilian and Italian, and more recently English, has led to hundred of loanwords in Maltese and to significant structural change (Aquilina 1958, 1959; Borg and Azzopardi-Alexander 1997; Misfud 1995; among others). Estimates vary as to the exact number of loanwords in Maltese. In a work in progress, Comrie (2011) estimated that about 61.2% of Maltese vocabulary is Arabic while the rest is from other sources. Mori (2009: 295), on the other hand, suggested that about 57% of the Maltese lexicon is in fact nonSemitic. The degree of impact that other languages have had on Maltese led Misfud (1995: 33–34) to distinguish between a Semitic Maltese morphology and a non-Semitic Maltese morphology. With regard to possession, it is notable that Maltese uses the synthetic form mainly with inalienable nouns (Borg 1994: 33; Fabri 1996: 230; KoptjevskajaTamm 1996: 264), while it uses the analytic construction, formed with genitive exponent ta’, as the less marked option to express attributive possession (Borg and Azzopardi-Alexander 1997: 112–113). Analyzing historic data, Mori (2009: 299) argues that, increasingly, the synthetic form has been reduced to use with inalienable nouns but that was not always the case. Given the number of loanwords in Maltese, which do not affect in a significant manner the basic vocabulary, it could be a possible player in the contraction of the synthetic form as posited by Koptjevskaja-Tamm: “One of the possible factors promoting

336

sayahi

the Maltese pattern may be a strong influence of Italian which extensively uses analytical genitives” (1996: 271). In a corpus of Nigerian Arabic that contains lexical insertions from English, Owens (2005) argued that, when the possessor is a pronoun, overwhelmingly the analytic form is favored when the possessed noun is in English (83/89) as opposed to Arabic (6/89). In total, analytic constructions, which predominantly occur with an English possessed noun, represented only 12.7% of the total possessive constructions. On the other hand, synthetic constructions represented 87.3% of all possessive construction in the corpus and happened overwhelmingly with an Arabic noun in the role of the possessed noun 99.97%. English nouns were used only marginally in synthetic constructions: 21 out of 610 (.03%). In his study of the use of the particle hana as part of the analytic construction in Nigerian Arabic, Owens reaches the conclusion that “Englishinserted items overproportionally use the hana possessive and underutilize the idafa” (2002: 188). For Moroccan Arabic, a study of children living in the Netherlands by Boumans showed that lexical insertions also favor the analytic from: “When Standard Arabic or French words for kinship terms or body parts are used as possessed forms in MA [Moroccan Arabic], they occur in the analytic construction” (Boumans 2006: 220). The same was true for an older member of the western Arabic dialect group: Andalusi Arabic. This variety was in intense contact with Romance leading to bidirectional code-switching and lexical borrowing (Thomas and Sayahi 2012; Sayahi 2014). With regard to the role of contact with Romance on the expression of possession in Hispanic Arabic, Ferrando argued that: It seems that in many cases, especially those that include Romance nouns, but not only these, the disappearance of the iḍāfa and its replacement by a preposition represents an attempt to translate too literally the Romance version. This Romance syntactic influence must have had an important contribution to the development of the analytical genitive in our materials, particularly related to the Mozarabic and Christian culture and institutions. (1995: 74; my translation from Spanish) All these cases point towards the fact that lexical insertions into the vernacular varieties from other languages favor the use of the analytic form over the synthetic form. I will now show how this obtains in the case of Tunisian Arabic.

Attributive possession in tunisian arabic

3

337

Attributive Possession in Tunisian Arabic

3.1 Current Study The linguistic situation in Tunisia bears many similarities to Algeria and Morocco with regard to the existence of classical diglossia between Arabic and the dialects and also societal bilingualism between Arabic vernaculars and French. The presence of Berber in Tunisia is miniscule when compared to the robust, although dwindling, numbers of Berber speakers in the other two countries. French introduced during the Protectorate (1881–1956) was not wide-spread during the colonial period among the Tunisian population although France managed to form a francophone elite that kept French language as part of the educational system after independence. Interestingly, with the democratization of education, the French language started to spread further among the educated Tunisian population which has been among the fastest growing in the Arab world. According to Foster, cited in Payne (1983: 264), after independence, the French educated elite “had succeeded in achieving what the French had failed to do, that is make almost the whole of Tunisian education Francophone.” Today education is compulsory in Tunisia, and all students have access to French starting from elementary school. While it is true that some school subjects have been Arabized, the majority of the science and technology subjects continue to be taught in French. More significantly, at the higher education level, many fields of specialization such as the hard sciences, medicine, or engineering are taught entirely in French. This strong presence of the French language in the educational system carries over to the professional fields with varying degrees. Although public administration has been officially Arabized, for practical purposes, code-switching between Arabic and French is common in many professions particularly those that have to do with science and technology. In addition, mass media has been open to the use of French since independence and continue to play a major role in its maintenance as a viable code of communication. The sociopolitical changes that started with the Tunisian Revolution of 2011 have led to the emergence of private mass media outlets that are more open to the use of French as an embedded language in Tunisian Arabic discourse. While competence in the French language varies across a wide spectrum that ranges from advanced competence to passive familiarity, as is the case in any situation of educational bilingualism, access to French is available to any Tunisian with some educational background. This access to French has allowed its usage in the different domains mentioned above and often leads to frequent code-switching between Tunisian Arabic and French (Belazi 1992; Lawson and Sachdev 2000;

338

sayahi

Sayahi 2011; among others). Tunisians, however, remain dominant speakers of Arabic making Tunisian Arabic almost always the base language from which they may code-switch to French. The nature of code-switching between the two languages more often than not takes the form of single insertions, in the majority of the cases nouns. Many terms that refer to modern life, science, and technology are used in French in an otherwise Arabic discourse. This was also observed by Bentahila and Davies (1995) who studied code-switching between Moroccan Arabic and French among the younger generations in Morocco. The proliferation of single-word code-switches provides a good context for the study of the expression of possession within a context of language contact. When French nouns are inserted in an Arabic discourse speakers have the choice between the synthetic and the analytic forms described above. In her comparative study, Eksell Harning (1980: 102) looked at the expression of attributive possession in the Tunisian and Libyan Arabic dialects. Her main findings show that in the texts she examined, extracted from previously published works by other scholars, there are overall many more instances of the synthetic form than the analytic form. She estimated that, in urban Tunisian dialects, there was one occurrence of the analytic form for every three occurrences of synthetic constructions. In particular, an important remark that she made is the almost categorical rule for the use of the analytic form with lexical insertions from other languages to the degree that “in some cases, it may even be difficult to imagine the possibility of a synthetic construction” (Eksell Harning 1980: 107). In order to test Eksell Harning’s claim and evaluate the role insertions from French, be they cases of nominal code-switching or established loanwords, play in increasing the use of the analytical form, this paper examines the use of possession in twelve sociolinguistic interviews with Tunisian speakers. The interviews were recorded with participants from the Greater Tunis region. For the purpose of this study, I use an average of fifteen minutes from each interview for a total of three hours analyzed. The participants include a variety of speakers that are divided in three main groups according to gender and level of education: 4 university-educated male speakers, 4 university-educated female speakers, and 4 high school-educated male speakers. Lack of availability of high school-educated female speakers at the time of the interviews did not allow the possibility for a more balanced set of groups of speakers that would have included high school-educated female speakers. The interviews covered different topics but the major focus was on education, family, and work. The speech sample analyzed is representative of the Tunisian dialect,

Attributive possession in tunisian arabic

339

including frequent insertion of French items especially by university-educated speakers (Sayahi 2011, 2014). In addition to analyzing the overall frequency for each type of constructions per group of speakers, I use a set of linguistic and extra-linguistic variables, based on Rosenbach (2002) and Wolford (2006), to determine what drives the choice for one form over the other. The linguistic variables considered include: language of the possessed and the possessor (Arabic vs. French), gender of the possessed noun, number of the possessed/ possessor, final sound in the possessed noun (vowel vs. consonant), class of the possessor (lexical vs. pronominal), use of determiner with the possessed and/ or possessor, use of adjectives with the possessed and/or the possessor, animacy of the possessor/ possessed, multiple annexation (when multiple layers of possession exist in the same phrase, as in: the book of the principal of the school), coordinate possessed and or possessor nouns (as in: the book and the pen of the principal or the books of the students and the teacher), and nature of the possession relation (real possessives vs. non-real possession). At the extra-linguistic level, I consider gender of the speaker and level of education. Age was not a factor as speakers’ ages ranged between 25 and 45 at the time of the interviews. I recognize that the use of the analytic or the synthetic forms, as expected, can be conditioned by pragmatic functions as well. In (5), the speaker talking about his profession as a barber describes how standing as part of this job is too much and is tiring. First, he uses waqfa with the synthetic form, then he reiterates what he said by emphasizing that standing is tiring but this time he uses waqfa with the genitive exponent and the possessive pronoun. This type of variation at the pragmatic level is not, however, the object of this study. In addition, given the similar nature of the interviews, no stylistic factors were considered in this analysis. For a discussion of the role of stylistic factors in the choice between synthetic and analytic constructions of possession, see Brustad (2000). (5) a-waqfit-ha barcha, l-waqfa mtɛ:ʕha ttaʕʕib standing-GEN.3.SG.F a lot DEF-standing of-GEN.3.SG.F tires s-saqqi:n the legs ‘[Its] standing is a lot, [its] standing tires the legs’ 3.2 Results In terms of absolute frequency, the data analyzed shows that there are 472 tokens of attributive possession in the sample, 68.64% of the cases are synthetic

340

sayahi

constructions (N = 324) while 31.36% are analytic constructions (N = 148). The fact that, overall, tokens of synthetic possessive constructions are more frequent than tokens of analytic constructions is not surprising. This is what would have been expected given previous studies on possession in the Arabic dialects and the conclusions reached by Eksell Harning (1980: 102) that in Tunisian Arabic there is much more usage of the synthetic form. In fact, she argued that, in her analysis mentioned above, synthetic forms are three times more frequent than analytic forms, while the data here shows that it is closer to being only two times more. One possible explanation is that the analytic form is continuing to spread in Tunisian Arabic. A major factor is the fact that, in my sample, the participants are all educated with 8 out of 12 possessing a higher education. Access to higher education is closely tied to competence in French, and, by consequence, it leads to more frequent insertions from that language into the speech of these speakers. As will be argued now, French nouns predominantly favor the analytic form. With regard to the language of the possessed noun, there were 385 cases where the possessed noun was in Arabic (81.57%), and in 87 cases it was in French. In the case of the Arabic possessed nouns, 81.6% (N = 314) were in synthetic constructions, and only 18.4% (N = 71) were in analytic construction. On the other hand, in cases where the possessed noun was in French, 11.5% (N = 10) of the cases were synthetic constructions, and 88.5% (N = 77) were analytic constructions. There is a statistically significant correlation between the language of the possessed noun and the type of the construction confirmed through a Pearson Chi-Square test (P = .00). In (6), the same speaker uses the word ‘pronunciation’ in Arabic and French but with the Arabic word she uses the synthetic construction while with the French equivalent she uses the analytic form. In some cases, both the possessed noun and the possessor noun are in French but linked with the Arabic genitive exponent as in (7). French possessed nouns that are part of a noun phrase are always in the analytic form as in (8) (6a) ma naʕrifiʃ yizrib wa illa nuṭqu xa:yib [I] don’t know [he] goes fast or pronunciation-GEN.3.SG.M is bad ‘I don’t know, he speaks too fast or his pronunciation is bad’ (6b) hiyya déjà l-prononciation mtɛ:ʕha ndra kifɛ:ʃ she already DEF-pronunciation of-GEN.3.SG.F who knows how ‘Already her pronunciation is strange’

Attributive possession in tunisian arabic

341

(7a) naqra fi cours mtɛ:ʕ traduction [I] study in course of translation ‘I am taking a course in translation’ (7b) ingénieur mtɛ:ʕ informatique engineer of computer science ‘a computer science engineer’ (8a) les certificats medicaux mtɛ:ʕu the certificates medical of-GEN.3.SG.M ‘his medical certificates’ (8b) c’est le rythme mtɛ:ʕ l- ħayɛ:t l- ʕa:di it is the rhythm of the-life the-normal ‘It is the daily rhythm of life’ Use of synthetic form is more common with inalienable nouns, especially those that are of higher frequency such as ru:ħ in (9), where ru:ħ is used to mean ‘self’ as opposed to Standard Arabic ‘spirit’. This particular word was used in 37 cases significantly contributing to the overall higher number of synthetic forms. While an argument can me made about the grammaticalization of this particular word, its contribution to a higher number of synthetic constructions is also true for other inalienable nouns that are much less commonly used with the genitive exponent (e.g., umm ‘mother’, bu ‘father’, ṣa:ħib ‘friend’, damm ‘blood’, etc.). This doesn’t mean the impossibility of the occurrence of the analytic construction with these and other similar inalienable words but that it is much less frequent and pragmatically more neutral. (9a) tħis ru:ħik fi ɣa:ba [you] feel self-GEN.2.SG.M in jungle ‘You feel yourself in a jungle’ (9b) hu:ma déjà yaʕtabru arwɛ:hum n-nuxba they already consider selves- GEN.3.PL.M the-elite ‘They already consider themselves the elite’ There are cases that are in-between code-switching and borrowing as in (10). The fact that these nouns are used with the synthetic form is an additional argument for their consideration as borrowings given their behavior as Arabic

342

sayahi

nouns and not French nouns, as argued by Poplack and Meechan (1998). In (10b), the word moyenne is used with the plural marker -ɛ:t which facilitates its treatment as an Arabic word to which the possessive pronoun can be attached (see Owens 2002: 192 for a discussion of the same phenomenon in Nigerian Arabic). (10a)

ħatta kɛ:n moyenne-ik even if average-GEN.2.SG.M ‘Even if your grade average is bad’

xa:yib bad

(10b) dima moyenne-ɛ:ti ça va always average-GEN.1.SG.M fine ‘My grade averages are always fine’ (10c) salaire-hum ʃay aʕjabb salary-GEN.3.PL.M something extraordinary ‘Their salary is extraordinary’ It has to be noted that established loans from French, such as blaṣa < place ‘place’, are here considered Arabic words. The fact that these words are known and used by even monolingual speakers and that they do not get reanalyzed by educated speakers with access to French is a strong argument in favor of considering them established loans behaving like other native vocabulary. It is easier for these words to be used with the synthetic construction as in the case of the word blaṣa in (11). (11a) tawwa fi blayiṣ-hum now in place-GEN.3.PL.M ‘now in their places’ (11b) ma hiʃi blaṣ-tik mʕa:na not it[is] place-GEN.2.SG.M with-us ‘Your place is not with us’ In the case of the language of the possessor, the same tendency is observed with slightly different rates. In total, there were 432 cases where the possessor was in Arabic and only 40 cases where the possessor was in French. This is due mainly to the fact that the pronominal possessors were all in Arabic and the nature of the conversation, which evolved around the speakers’ education, profession, and family and didn’t allow for more French possessors. The

Attributive possession in tunisian arabic

343

speakers more often than not were speaking about themselves and entities that are linked to them. In additions, these conversations, although containing frequent code-switching to French, remain predominantly in Tunisian Arabic, and the majority of the code-switches are at the single word level. In 73.6% of the cases where the possessor was in Arabic, the construction was a synthetic one (N = 318), and in 26.4% (N = 114) it was an analytic construction. In the case of the French possessors, 15.0% (N = 6) were part of a synthetic construction, and 85.0% (N = 34) were in analytic constructions. Despite the smaller number of tokens, the language of the possessor clearly plays a role in determining the type of the construction (P = .00), as it does in the case of the possessed noun. On the other hand, pronominal possessors strongly favor synthetic constructions in 83.28% (N = 284) of the cases, while lexical possessors favor analytic possession in 69.47% (N = 91). Pronominal possessors are easier to attach to the possessed nouns while with lexical possessors there is a need for the construct state to be used, a form that is not much different than the constructions that use the genitive exponent. This also reflects the fact already mentioned above that all pronominal possessors are in Arabic. The Chi-Square test confirms that there is a statistically significant correlation between the class of the possessor, i.e., pronominal vs. lexical, and the type of construction (P = .00), with lexical possessors favoring analytic possession. Another factor, animacy, also favors synthetic forms which correlates with inalienable possession: when both the possessor and the possessed are animate, 90.91% (N = 80) of the cases were used in a synthetic construction. This refers specifically to kinship relationships. When it comes to the nature of the possessive relation itself, real possession or prototypical possession as defined in the literature (Rosenbach 2002: 121; Hammarberg and KoptjevskajaTamm 2003: 126), strongly favors the synthetic form (90.12%, N = 292), while other types of possession slightly favor the analytic form at 54.06% (N = 80). This is also tied up with the fact that inalienable possession, including family members and body parts, favors the synthetic form. Finally, both first person and second person strongly favor synthetic forms (first person 88.05% [N = 140]; second person 92.86% [N = 26]), while third person shows preference for the synthetic form but with a significant use of the analytic at 44.5% (N = 126). While first person and second person frequently referred to both alienable and inalienable possession, many cases where third person possessors were used referred to inanimate entities in alienable possessive relations. The results also show that the gender of the possessed noun doesn’t affect the choice between synthetic and analytic possessive constructions and neither does the number of the possessed or the possessor. Other factors that did

344

sayahi

not prove significant are multiple annexation and coordination as there were very few cases of these two. In sum, at the linguistic level, the factors that favor the use of the synthetic form as opposed to the analytic form are: the use of Arabic nouns, pronominal possessors, animacy of the entities involved, and first and second persons. Particularly significant for this study is the fact that the language of the possessed noun shows that French nouns heavily favor the usage of the analytic construction. With regard to the social factors, both groups of males and females produced overall more synthetic forms that analytic ones. Males used a synthetic form in 63.88% (N = 168) of the cases and females in 74.64% (N = 156) of the cases. With regard to education, comparing university-educated male speakers to high school-educated male speakers shows that there is a correlation between level of education and use of either forms of possession (P = .02). While both groups produced similar numbers of tokens, high school-educated participants produced 137 cases while university educated produced 126, high school-educated male speakers used synthetic possession in 70.07% (N = 96) of the cases while university-educated speakers used it in 57.14% (N = 72) of the cases. This could be explained by the fact that university-educated participants used more possessed nouns in French (27.87%, N = 35) than their high school educated peers (16.06%, N = 22). This difference has to do with the use speakers made of French nouns when talking about their professions and education. As mentioned above, more educated Tunisians tend to code-switch to French frequently in conversations that deal with these two topics. (12) s-service mtɛ:ʕ-na l-mortalité mtɛ:ʕ-u service of-GEN.1.PL.M the-mortality of- GEN.1.PL.M vingt-sept pour cent twenty seven per cent ‘The mortality rate of our unit is twenty seven percent’ Example (12) shows how with French nouns the tendency is to use the genitive exponent as opposed to the suffixed pronoun. Although different forms of the word service can be produced in Tunisian Arabic, including a more established loan where the first vowel is raised and the labiodental sound is devoiced, here the speaker uses the genitive exponent. In the case of the word mortalité, the use of the exponent appears to be more necessary as the use of a French feminine noun with a suffixed genitive possessor would require the obligatory insertion of the feminine marker [-it] which would further modify the morphophonological structure of the word. Owens raised this issue when he

Attributive possession in tunisian arabic

345

discussed English insertions in Nigerian Arabic. He argued that in the case of the need to add the feminine marker “The use of the hana possessor, however, obviates the need to make morphophonological adjustments to the Englishmixed word” (2002: 191). French possessed nouns that end in vowels represent 52.87% of all French possessed nouns, and in their majority they favored analytic possession (93.48%) at the same time 82.93% of possessed French nouns that end in consonants favor analytic possession as well. This is not a big different especially given the small size of the sample. 4 Conclusion This study shows that in the Tunisian dialect the rate of usage of the analytic possessive construction is a little less than half that of the synthetic form. This rate is higher than what was found in the study of Eksell Harning (1980). While the internal change towards a more analytic structure should be the main reason, an argument can be made based on the results discussed in the previous section and the results reported in the literature, that contact with French is accelerating the change by increasing the use of the analytic form with French nouns that favor it. In addition to established borrowings, French bare nouns are commonly used by the population that has access to French and often tend to preserve the morphological shape of the inserted noun and “maintain its discrete integrity” (Owens 2002: 190) by not using the synthetic form. The fact that the spread of education increased the use of French insertions by a growing sector of the Tunisian population should not be underestimated as a factor in accelerating a change that has been in progress for a long time. References Aquilina, Joseph. 1958. “Maltese as a mixed language,” JSS 3: 58–79. ———. 1959. The structure of Maltese: A study in mixed grammar and vocabulary. Valletta. Baron, Irène, Michael Herslund, and Finn Sørensen. 2001. Dimensions of possession. Amsterdam. Belazi, Hedi M. 1992. Multilingualism in Tunisia and French/Arabic code switching among educated Tunisian bilinguals. Ithaca. Bentahila, Abdelali and Eirlys E. Davies. 1995. “Patterns of code-switching and patterns of language contact,” Lingua 96: 75–93.

346

sayahi

Borg, Alexander. 1994. “Observations on some evolutionary parallels and divergences in Cypriot Arabic and Maltese,” in Jordi Aguadé, Federico Corriente, and Marina Marugán (eds.), Actas del Congreso Internacional sobre Interferencias Lingüísticas Arabo-Romances y Paralelos Extra-Iberos. Zaragoza. 21–40. Borg, Albert and Marie Azzopardi-Alexander. 1997. Maltese. London. Boumans, Louis. 2006. “The attributive possessive in Moroccan Arabic spoken by young bilinguals in the Netherlands and their peers in Morocco,” Bilingualism: Language and Cognition 9: 213–231. ———. 2005. “Lexical variation in Moroccan Arabic,” Estudios de dialectología norteafricana y andalusí 9: 125–137. Brustad, Kristen. 2000. The syntax of spoken Arabic: a comparative study of Moroccan, Egyptian, Syrian, and Kuwaiti dialects. Washington, DC. Comrie, Bernard. 2009. “Maltese and the World Atlas of Language Structures,” in Bernard Comrie, Ray Fabri, Elizabeth Hume, Manwel Mifsud, Thomas Stolz, and Martine Vanhove (eds.), Introducing Maltese linguistics. Amsterdam. 3–11. ———. 2011. Maltese Loanword typology: a preliminary assessment. Paper presented at GĦILM 3rd Conference on Maltese Linguistics University of Malta, 2011 April 8–10. Eksell Harning, Kerstin. 1980. The analytic genitive in the modern Arabic dialects. Göteberg. Fabri, Ray. 1996. “The construct state and the pseudo-construct state in Maltese,” Revista di Linguistica 8: 229–244. Ferrando, Ignacio. 1995. El dialecto andalusí de la Marca Media: los documentos mozárabes toledanos de los siglos XII y XIII. Zaragoza. Hammarberg, Bjorn and Maria Koptjevskaja-Tamm. 2003. “Adnominal possession: combining typological and second language perspectives,” in Giacalone Ramat (ed.), Typology and Second Language Acquisition. Berlin. 125–180. Heine, Bernd. 1997. Possession cognitive sources, forces, and grammaticalization. Cambridge. Hickey, Raymond. 2010. “Language Contact: Reassessment and reconsideration,” in Raymond Hickey (ed.), The Handbook of Language Contact. Malden. 1–28. Koptjevskaja-Tamm, Maria. 1996. “Possessive noun phrases in Maltese: alienability, iconicity, and grammaticalization,” Rivista di Linguistica 8: 245–274. Lawson, Sarah and Itesh Sachdev. 2000. “Codeswitching in Tunisia: attitudinal and behavioural dimensions,” Journal of Pragmatics 32: 1343–1361. Mifsud, Manwel. 1995. Loan Verbs in Maltese: A Descriptive and Comparative Study. Leiden. Montoya, Maria Cristina. 2011. Expression of Possession in Spanish in Contact with English: A Sociolinguistic Study Across Two Generations in the Greater New York Metropolitan Area. Albany.

Attributive possession in tunisian arabic

347

Mori, Laura. 2009. “The shaping of Maltese throughout the centuries,” in Bernard Comrie, Ray Fabri, Elizabeth Hume, Manwel Mifsud, Thomas Stolz, and Martine Vanhove (eds.), Introducing Maltese linguistics. Amsterdam. 291–307. Orozco, Rafael. 2009. “La influencia de factores sociales en la expresión del posesivo,” Lingüística 22: 25–60. Owens, Jonathan. 2002. “Processing the world piece by piece: iconicity, lexical insertion, and possessives in Nigerian Arabic codeswitching,” Language Variation and Change 14: 173–210. ———. 2005. “Bare forms and lexical insertions in codeswitching: A processing-based account,” Bilingualism: Language and Cognition 8: 23–39. Payne, R. M. 1983. Language in Tunisia. Tunis. Payne, Doris L., and Immanuel Barshi. 1999. External possession. Amsterdam. Poplack, Shana and Marjory Meechan. 1998. “How languages fit together in code-mixing,” International Journal of Bilingualism 2: 127–138. Poplack, Shana and Stephen Levey. 2010. “Contact-induced grammatical change,” in Peter Auer and Jürgen Erich Schmidt (eds.), Language and Space—An international handbook of linguistic variation, Vol. 1. Berlin. 391–419. Rosenbach, Anette. 2002. Genitive variation in English: conceptual factors in synchronic and diachronic studies. Berlin. Sayahi, Lotfi. 2011. “Code-switching and language change in Tunisia,” International Journal of the Sociology of Language 211. 113–133. ———. 2014. Diglossia and language contact: Language variation and change in North Africa. Cambridge. Silva-Corvalán, Carmen. 2008. “The limits of convergence in language contact,” Journal of Language Contact 2: 213–224. Taine-Cheikh, Catherine. 2000. “A propos de l’opposition ‘type synthétique’ vs ‘type analytique’ en arabe, in Aspects of the Dialects of Arabic Today,” in Abderrahim Youssi, Fouzia Benjelloun, and Mohamed Dahbi (eds.), Proceedings of the 4th Conference of the International Arabic Dialectology Association (AIDA). Rabat. 234–244. Thomas, Juan Antonio and Lotfi Sayahi. 2012. “A quantitative analysis of codeswitching in the Arabic-Romance kharjas,” Journal of Language Contact 5: 1–17. Weinreich, Uriel. 1963. Languages in Contact: Findings and Problems. The Hague. Wolford, Tonya. 2006. “Variation in the expression of possession by Latino children,” Language Variation and Change 18: 1–13.

Aramaic Loanwords in Gǝʿǝz Jürgen Tubach 1 Introduction The Semitic languages of Ethiopia belong to the southwest Semitic branch of the Semitic group and altogether to the Afro-Asiatic family of languages. The Semitic languages of Ethiopia were put into writing in modern times, especially at the beginning of the last century and at the end of the 19th century. Only Gǝʿǝz, the Old-Ethiopic language, and Gē sinān (Ḥararī) have an old tradition. The latter is the language of Ḥarar (ሐረር) and plays a special role, because it does not belong to the other southwest Semitic languages of Ethiopia in a strict sense. Originally Gǝʿǝz was the language of Axum and became the administrative language of the Axumite empire. When Ezana (ʿĒzānā) converted to Christianity in the first half of 4th century, Gǝʿǝz received an additional function. It became the language of the growing church of the empire. The script and language was brought to the northern highland of Ethiopia by Sabaean colonists in the first half of the 1st millennium BC. It is still clearly recognizable that the forms of the single characters of the Ethiopian script have South Arabian ancestors. The older inscriptions are still Sabaean, but from the first century AD on the language of the inscriptions can be called Gǝʿǝz. The longest texts are trilingual inscriptions of the 4th century. The language underwent some modifications which happened in the time of Ezana or earlier. In contrary to all other Semitic languages (except Akkadian) Gǝʿǝz now writes the vowels and changed the direction of the script which runs from left to right. The system of vowel signs attached to a letter comes from India. The basic form of a letter is pronounced with the vowel a as in many scripts of India. When the first settlers came from South Arabia they did not find a deserted land. The northern part was inhabited by Cushitic people with central Cushitic languages. The Agaw people may have been the original inhabitants of the Ethiopian highlands. For this reason the language of the Sabaean colonists adopted many Cushitic loan words. The Christianisation and the subsequent translation of biblical texts and theological literature lead to the reception of Greek words1 including some Latin loan words in the Greek language. Probably there were many educated Axumites (including the court) who understood 1  Weninger 2005: 469–471, 481–483.

© koninklijke brill nv, leiden, ���5 | doi ��.��63/9789004300156_019

Aramaic Loanwords in Gǝ ʿ ǝz

349

Greek. Ezana (and one of his predecessors) left inscriptions in Greek,2 and one can assume that he had learned Greek in his youth (from his teachers Frumentios and Aidesios). Zoskales (1st cent. AD), the first known Axumite king is described by the anonymous author of the Periplus maris Erythraei as ‘a fine person and well versed in reading and writing Greek’.3 The coins of the Axumite kings4 bears legends in Greek (3rd–7th cent. AD).5 Gǝʿǝz is used for silver and copper coins in the 5th and 7th century.6 For this reason some of the Greek words must be older than the translation of the bible. Furthermore Gǝʿǝz possesses Arabic loan words7 (including Greek loanwords in Arabic transcription),8 some Coptic terms,9 and very few words borrowed from India.10 While sōkar (or sōkkar, ሶከር)11 is borrowed via the Arabic sukkarun from a Middle Indian language (Pali sakkharā < Sanskrit śarkarā) and not a direct loan word from the Indian subcontinent.12 There is at least one case where the word was adopted from India without detour. Nagayāt (ነገያት, nagēyāt ነጌያት, nagǝyāt ነግያት, SG nagē ነጌ),13 one of the names for the elephants,14 2   RIÉ I 359–387 (no. 269–286A); RIÉ IIIA. 3  Casson 1989: 52 (ln. 20–22), 53; Huntingford 1980: 21 (ch. 5). It is not quite clear whether Zoskales was the nǝgūša nagašt residing in Axum or the nǝgūš/nagaši of the littoral coast at the Red Sea with Adulis as center (Huntingford 1980: 147f), who was more or less independent from the king of kings (for the various possibilities, see Casson 1989: 109f; cf. Bowersock 2013: 26, 31–44, 43, 53; Seland 2010: 38). 4  Munro-Hay 1986; 1999; Munro-Hay and Juel-Jensen 1995; Pedroni 1997; Godet 2004; Hahn 2000: 281–311; West 2006. 5  Hahn 1994; 2010. Cf. Bausi 2003: 171ff. 6  West 2006. Cf. Bausi 2003: 166ff. 7  Weninger 2005: 473–475. 8  Voigt 1991. 9  Weninger 2005: 469–471, 475–482. 10  Weninger 2005: 479. 11  Dillmann 1865: 380; 1866: 199. 12  Leslau 1987: 497; Littmann 1926: 411.417. Cf. Weninger 2005: 479. 13  Dillmann 1865: 685; 1866: 228; Grébaut 1952: 432; Leslau 1987: 390; 1989: 127. The Gǝʿǝz word nagōt means ‘trunk of an elephant’ and is derived from nagē (Leslau 1987: 393). The word is preserved in Tǝgrē/ Tǝgrǝñña as nǝgōt ‘trunk (of the elephant)’, while ʾabūnǝgōt (‘father of the trunk’) means the elephant (Littmann and Höfner 1956–1962: 342; Leslau 1982: 55; cf. Kane 2000: 1.1373). 14  The usual name for the elephant in Gǝʿǝz and Tǝgrē/ Tǝgrǝñña is ḥarmaz / ḥarmāz / ḥārmāz (Leslau 1987: 243; Leslau 1989: 17; Littmann and Höfner 1956–1962: 66; Kane 2000: 1.185; cf. Dillmann 1865: 83). It already occurs in Axumite inscriptions with a picture of an elephant (RIÉ I 340 [no.255]). Another word for elephant is falfal which can mean ‘water buffalo’ as well (Leslau 1987: 159; cf. boves Lucas = elephant [Pliny, Naturalis

350

tubach

comes from India15 and apparently means the Indian elephant (elephas maximus indicus) in contrast to the African one.16 The Axumite army used war elephants,17 and the kings must have imported elephants from India trained

historia, VIII.6 §16]). Sometimes the latter is the only meaning given by the dictionaries (Leslau 1989: 250; cf. Dillmann 1865: 1347, 84). The word has an intricate history. It must be derived from the Arabic fīlun ‘elephant’, which is borrowed from the Aramaic/Syriac pīl/ pīlā (Beyer 1984: 668; Dalman 1938: 332; Brockelmann 1928: 566) or the Middle Persian pīl (Nyberg 1974: 161; Boyce 1977: 76; MacKenzie 1990: 68; Durkin-Meisterernst 2004: 289). The word is not attested in Biblical Hebrew, but it later occurs in Hebrew (pīl). The orthography in Akkadian—it can be pīru or pīlu (cf. Armenian pił)—suggests that the word is not of Semitic origin, although there existed elephants in Syria and Mesopotamia till the first Millenium BC (AHw 3.867). It may be that the Hamitic word eḷu was combined with the Egyptian definite article p and adopted as pīl/pīr in Syria/Mesopotamia (cf. Frisk 1960: 493). Curiously the Oriental fil is attested in Old Norse and Old Danish as fill/fil (elephant, cf. Buck 1949: 189, 190). 15  Dillmann 1865: 685; Leslau 1987: 390. The loan word is not mentioned in Schneider’s comprehensive work (2004). 16  Halévy 1896: 262. 17  One of the sarāwīt/sǝraw/sǝrāw (troops, legions, detachments, military units) of the Axumite army consisted of elephants apparently mounted with turrets for archers (Altheim and Stiehl 1964–1969: Vol. V.2, 1230f; 1971: 1.423f; cf. Sergew 1972: 95). The sarwē dākēn (ሰርዌ ዳኬን) or sarwē dākuēn (ሰርዌ ዳኳን) is mentioned in two inscriptions of Ezana found in Aksum (Littmann 1913: 24f, 27 [no. 9, ln. 21.24], 28f, 3 [no. 10, ln. 10]; 1950: 110 [Nr.10]; RIÉ I 260 [no.188, ln. 10], 256 [no.187, ln. 21.24]). Already Dillmann (1879: 213 esp. fn. 2, cf. Pereira 1898: 6–8; Halévy 1896: 262) recognized that dākuēn/dākēn—he read it as dakūn or dūkūn—is a loanword from a Cushitic language meaning elephant (with reservation, Littmann 1913: 31; 1926: 413; 1950: 113). He thought that it is borrowed from Saho, and he quotes Reinisch (1878: 423: dakani) and Munzinger (1863: 142) including the Amharic cognates zähōn/zāhōn/zǝhōn (Isenberg 1841: 155; Leslau 1976: 178). The word dakāno (PL dakūn) means elephant (Reinisch 1890: 105, 408; Vergari and Veragi 2003; Munzinger 1864: 142 fn. 1: dekeni = elephant). Cognates exist in several Cushitic languages (e.g., in ʿAfar dakano/dakanu, cf. Reinisch 1890: 105; Morin 2004: 193; 2012: 269; dakkani: Isenberg 1840: 3, 13; dakkěnï: Isenberg 1841: 155). The term occurs as loan word in some of the Semitic languages of Ethiopia (Leslau 1963: 55 [sub doxon, read ž instead of z exept Amharic, Argobba, Aymellel] and esp. Leslau 1979: 2.216f; 3.721; Altheim and Stiehl 1964–1969: vol. V.2, 230f), but not in Gǝʿǝz. The Gǝʿǝz term is probably derived from the plural. The Saho are agro-pastoral nomads living in Eritrea and Northern Ethiopia, a region, which corresponds to the old Axumite empire. The language of the ʿAfar (Arabic Danākil, SG Dankalī) is closely related to Saho. The ʿAfar people live in Eastern Eritrea, North-Eastern Ethiopia, and Djibouti.

Aramaic Loanwords in Gǝ ʿ ǝz

351

for combats,18 if it was not possible to domesticate Ethiopian elephants19 with Indian help.20 Close contacts to the Indian subcontinent existed at that time, as a Brahmi seal from Adulis,21 Axumite coins found in South India,22 and the Kushana coin hoard23 from Dabra Dāmō show. African bush or forest elephants (Loxodonta Africana and Loxodonta cyclotis)24 can normally not be tamed and afterwards trained for battles. In medieval times, Gǝʿǝz adopted several Amharic words, which is not astonishing because the language of the imperial court and the language of the people in the central highlands was Amharic. Words borrowed from Arabic in Axumite and post-Axumite times are a difficult stratum of the vocabulary. Furthermore the inventory of the language contains several loanwords from Hebrew and Aramaic which are an integral part of the language. The most interesting words of this group of borrowings are several Termini Technici of the religious world, either the Jewish or the Christian religion. These words can be divided into three categories: (a) Some words belong to the so-called cultural borrowings and come from the Sumerian or Akkadian language. They were adopted by Hebrew and Imperial Aramaic (or already by Old Aramaic) in the first millennium BC. 18  Zebu bulls (bos indicus) were imported in Axumite times (Ziegert and Wendowski 2003: 217). It is therefore not improbable that elephants were brought to Axum as well. 19  Nonnosos, the leader of a Byzantine embassy, was sent by Justinian (527–565) to the court of Caleb. On his journey from Adulis to Aksum, he saw a huge herd of elephants near Aue (Αὔη) feeding in the plain (probably the fertile Ḥazamo/Ḥazumo plain). He estimated that there were nearly 5,000 animals (Muller 1851: 4.180a; Henry 1959: 1.6; Sergew 1972: 139; Altheim and Stiehl 1964–1969: vol. V.2, 230; Schwarz 1894: 360). The Ptolemies caught elephants in the region of Adulis and brought them from Gabaz, the seaport of Adulis, to Egypt on special transport ships called Elephantagoi (Jackson 2002: 86; Gates 2005: 44; Altheim and Stiehl 1964–1969: vol. I, 67f, 130; Bard 2008: 293; Bowersock 2013: 34ff). These elephants belonged to an extinct species, the so-called North African Forest elephant (Loxodonta africana pharaoensis), which was much smaller than his Indian cousins. 20  It was often thought in antiquity that elephants can only understand the ‘Indian’ language (cf. Hofmann 1975: 62–65). 21  RIÉ I 399f, np. 291. 22  Nawartmal and Nawartmal 1998; Nawartmal 1999; Juel-Jensen 1999; Krishnamurthy 1999; 2000; Metlich 2006; Day 2011. 23  104 gold coins of Vima Kadphises, Kaniška, Huviška, and Vasudeva, cf. Mordini 1960; 1967; Göbl 1970; Berzina 1984. 24  Krumbiegel 1943: 28ff, 35–49.

352

tubach

(b) There are several words which belong to a typical Jewish sphere and are useless in a Christian context. (c) Some words can have a Jewish or Christian origin. None of these words was directly borrowed from Hebrew or Aramaic. They were used in the Jewish or Christian communities in South Arabia25 and came to Ethiopia in early Axumite times as we will see later. All these words are not perceived as a kind of foreign vocabulary; they are typical Ethiopian words like the German Fenster which comes from the Latin fenestra while the English word window is a real Germanic word which means an unglazed hole in the roof, a blowhole. 2

Cultural Loanwords

A cultural loanword is haykal (ሀይከል)26 with a long history of borrowing. The long walk of this word starts in the 3rd millennium BC. Literally the Sumerian word e-gal means ‘big (gal) house (e)’ and designates the palace of the king or the temple of a god. The Sumerian word was adopted by Akaddian as ēkallu in the same meaning.27 A palace or a temple for one of the great gods had a lot of rooms, because both king and god had a royal or heavenly household with many servants etc. Minor gods belonged to the suite of a great god and all needed rooms in a e-gal, in a ‘big house’. The word spread from Mesopotamia to Syria and Phoenicia.28 In all Canaanite languages it means both palace and temple, e.g., Hebrew hēkāl (‫) ֵה ָיכל‬.29 At the same time the word came into the old Aramaic language (or languages) and afterwards into Imperial Aramaic.30 It was fully integrated into the inventory of the vocabulary. Later it appears in all Aramaic dialects of the post-Achaemenid time.31 In Hebrew the word 25  Many of the Hebrew, Aramaic, or Syriac loanwords appear in the middle and late Sabaic period of the language. Several words belong to the secular sphere and have nothing to do with the religious world (Beeston 1994: 39–45 esp. 42f; Gajda 2010: 109–113). 26  Dillmann 1865: 15; 1866: 162; Grébaut 1952: 11; Leslau 1987: 221; 1989: 3; Nöldeke 1910: 32. 27  AHw 1.191–193. 28  Kaufman 1974: 27. 29  KBL2 230f; KBL3 234f; HALOT 1.244f; Gesenius 1987–2010: 274. 30  Beyer 1984: 562; Jean and Hoftijzer 1965: 64; DNWSI 278; Kaufman 1970: 40f; cf. Tropper 2008: 40; DULAT 336; Aistleitner 1963: 86f (nr. 827 [hkl = palace]); Krahmalkov 2000: 157 (hkl as deity). 31  Jean and Hoftijzer 1965: 64; DNWSI 278; Beyer 1984: 174; Dalman 1938: 112; Jastrow 1886– 1903: 345; Levy 1924: 1.464.

Aramaic Loanwords in Gǝ ʿ ǝz

353

can take a special meaning, it is used as designation for the room of the temple between ʾŪlām (or ʾĒlām) and the adytum which is called dǝbīr or qodäš qŏdāšīm (with the ark of the covenant).32 The Hēkāl bears a second name. This is qodäš ‘holy one’. The room had an altar. When the Bible was translated into Syriac, hayklā (SG.ABS haykal) could still be used for the palace of the king, but in non-biblical texts it is normally the name for the church building or a part of it. The special meaning ‘nave’ in East Syriac designates that ‘part of the church which is for the people’33 and is evolved from the Hebrew or at least inspired by Hebrew. It corresponds to the Qodäš or Hēkāl of the temple, but it was not allowed for the normal believers to enter this room, it was only reserved for the priests. The Akkadian word ēkallu lost gradually its secular meaning when it came into Aramaic and later to the Arabian Peninsula where it is attested in middle Sabaic (as hykl / ḥykl ‘temple, palace’).34 In contrast to Aramaic, that is, Syriac, the Arabic35 and old Ethiopic haykal developed the additional meaning ‘altar’. The reason for this is that the Hēkāl of the Temple in Jerusalem had an altar. 3

Jewish Loanwords

Several words in Gǝʿǝz have a Hebrew or Aramaic origin. All belong to the sphere of religion and must have been familiar to the people of Jewish and later Christian communities. The following nouns or verbs are part of this category: sanbat ‘Sabbath, Sunday, week’, mǝṣwāt ‘alms, almsgiving, charity’, ṣalōt ‘prayer’, ṭāʿōt ‘idol, false god’, haymānōt ‘belief’, ʾōrīt ‘Torah, law’, sīʾōl ‘netherworld, She’ol’, gahannam ‘hell’, fesḥ ‘Passach, Easter’, ʿarb ‘Friday’, nabīy ‘prophet’, masīḥ ‘anointed, Christ’, talmīd ‘disciple’, aṭhara ‘to clean, purify, lustrate’ (esp. de lustrationibus Judaeorum), tābōt ‘ark (esp. Noah’s ark), ark of the covenant’, kāhǝn ‘priest’, and malak ‘angel’.

32  KBL2 231; KBL3 235; HALOT 1.245; Gesenius 1987–2010: 274; Dalman 1938: 112; Jastrow 1886– 1903: 345; Levy 1924: 1.464. 33  Payne Smith 1903: 103; Payne Smith 1879–1901: 1.1004; cf. LS 174; SL 340f. 34  Beeston 1994: 41; Biella 1982: 108f; for other meanings of the lexeme, see Beeston, Ghul, Müller, Ryckmans 1982: 58, 74. 35  Wehr 1985: 1368; Wahrmund 1898: vol. I.2, 1142.

354

tubach

The Hebrew šabbāt36 (‫ > ַׁש ָּבת‬Greek σάββατον37 > Coptic ⲥⲁⲃⲃⲁⲧⲟⲛ38) came into Aramaic as šabbā (ABS).39 The last consonant was misunderstood as feminine ending (Syr. šabbṯā, ABS šabbā, PL.ABS šabbīn). The noun received the additional meaning ‘week’ (like šabbāt in post-biblical Hebrew40 and the Hebrew loanword šambat in Iran41). When the planetary week was introduced in Syria, the Jewish and Christian communities refused to accept this innovation and counted the week days with numbers, only for Friday and Saturday nouns were applied. For Saturday the name Sabbath was used. The doubling of the middle consonant was dissolved in some Aramaic dialects by insertion of the consonant nun. This occurs in East Aramaic in Mesopotamia (sometimes in East Syriac, Hatrene Aramaic, Mandaic, Jewish Aramaic). From Mesopotamia, especially from Babylonia, the word Sabbath spread to to Persia42 and Arabia, probably to South Arabia and from this region to Ethiopia. The vocalisation (as ሰንበት sanbat)43 in Gǝʿǝz shows that the archetype was not the Aramaic but the Hebrew word which was probably used as citation loan word in Arabia in Jewish circles. It must have come from Babylonia to Eastern Arabia and South Arabia because of the substitution of the lengthening of the ٌ َْ middle consonant.44 The Qurʾanic sabtun (�‫��س�ب�� ت‬, 2,65, etc.) is used for the Jewish Sabbath,45 but it is probably not derived from the Jewish Aramaic or Syriac šabbṯā. It belongs to the nominal form qatl and is an arabizised adaption of the Hebrew word. As already mentioned the week days have no names, but are counted with numbers in Syriac (and Hebrew). Sunday is ‘the first day of the week’ (Syr. ḥaḏbšabbā ‘[day] one in the week’, cf. Mark 16:2 par.; Jewish Aramaic ḥaḏbšabbā/ḥaḏbšubbā;46 Mandaic habbšabbā; Middle Parthian 36  KBL2 947f; KBL3 1310–1312; HALOT 4.1409–1411; Gesenius 1987–2010: 1321f; Dalman 1938: 414; Jastrow 1886–1903: 1520; Levy 1924: 4.506f. 37  Bauer 1988: 1479f. 38  Böhlig 1958: 10, 396. 39  Beyer 1984: 699; Dalman 1938: 414; Jastrow 1886–1903: 1520; Levy 1924: 4.507. 40  Dalman 1938: 414. 41  See footnotes 45–47 below. The meaning ‘week’ can be concluded from the system of counting the days. 42  Middle Parthian šambat (Boyce 1977: 85; Durkin-Meisterernst 2004: 318) and Sogdian šambed/δ (Sims-Williams and Durkin-Meisterernst 2012: 48.184); cf. Middle Persian šambat/šambad (Durkin-Meisterernst 2004: 101). 43  Dillmann 1865: 369; 1866: 198; Leslau 1987: 505f; Leslau 1989: 72f; Nöldeke 1910: 37. 44  Spitaler 1954: 257–266 = 1998: 3–13. 45  Jeffery 1938: 160f. 46  Margolis 1910: 169*; Dalman 1938: 137; Jastrow 1886–1903: 425; Levy 1924: 2.15; 4.493.

Aramaic Loanwords in Gǝ ʿ ǝz

355

ēwšambat ‘[day] one of the week’47 = Middle Persian yekšambat/yekšambad48 = Sogdian ēwšambed/δ49). The same system was applied in Ethiopia. Sunday bears the name ʾǝḥūd (እሑድ Gǝʿǝz/Amharic).50 When the day is called Sanbat as well it is connected with the medieval controversy whether there are one or two sabbatical days a week. Emperor Zarʿa Yāʿqōb (1434–1468, *1399) decided at a synod held in 1450 in the monastery of Dabra Mētmāq that both days are official holidays. The custom to celebrate both days was very popular among the northern monks from the center of the old Axumite empire. If one wanted to avoid any ambiguity one could speak of Sanbat krǝstyān and Sanbat ʾAyhud. If the Saturday is called qadām (ቀዳም)51 and the Sunday ʾǝḥūd the ambiguity is not neutralized, because both words have the same meaning. The designation Sabbath as name of Saturday including the counting of the weekdays is Jewish in origin and was used by Jewish converts to Christianity. This was the case in Palestine/Syria/ Mesopotamia, and it can be assumed that the same process first happened in South Arabia and afterwards in Ethiopia. More striking is the name for Friday. It makes sense in a Jewish community, but in a Christian environment it is useless and meaningless. The Gǝʿǝz word ʿarb (ዐርብ,52 Amharic ዓርብ,53 Tǝgrǝñña ዓርቢ54) has close parallels in Syriac, Armenian, and Coptic. The Syriac ʿrūḇtā55 and the Arabic ʿurūbatun/ ʿarūbatun, used in Pre-Islamic times,56 are loanwords from Jewish Aramaic in nearly the same vocalisation (ʿărūḇtā)57 and mean the preparation day for the

47  Boyce 1977: 25; Durkin-Meisterernst 2004: 101 (Sunday [without further explanation]). 48  Durkin-Meisterernst 2004: 374 (Sunday [without further explanation]). According to Durkin-Meisterernst (2014: 193 [§402]), the Middle Persian (and Parthian) word is a translation of the corresponding Syriac expression, but Mani did not use classical Syriac for his canon. He used an East Aramaic dialect (from Babylonia) for his holy books (Beyer 1994: 26). 49  Sims-Williams and Durkin-Meisterernst 2012: 48, cf. 184. 50  Dillmann 1865: 723; Leslau 1987: 12; 1989: 133. 51  Dillmann 1865: 463; Leslau 1987: 421; 1989: 90; cf. Leslau 1961: 62f. 52  Dillmann 1865: 964; 1866: 252; Grébaut 1952: 288; Leslau 1987: 69; 1989: 172; cf. Leslau 1961: 65.68; Chaîne 1907: 94 (§171). 53  Leslau 1976: 126. 54  Kane 2000: 2.1847. 55  Payne Smith 1879–1901: 2.2984; Payne Smith 1903: 427f; LS 546; SL 1136. 56  Wahrmund 1898: Vol. I.2, 246. 57  Dalman 1938: 322; Jastrow 1886–1903: 1114; Levy 1924: 3.691f.

356

tubach

Sabbath.58 The same word is used in Mandaic for Friday (ʿurūḇṯā/ʿurūp̄ ṯā).59 In Egypt the day is called ⲡⲁⲣⲁⲥⲕⲉⲩⲏ,60 but this is borrowed from the JewishHellenistic synagogue (παρασκευή).61 The meaning is the same as in Syriac or the Jewish-Aramaic. The Copts still have two typical Christian names for this day. It is either ‘the sixth [day]’ (ⲡⲥⲟⲟⲩ) or ‘the big fasting, the big day of abstinence’ (ⲧⲛⲟⳓ ⲛ̄ⲛⲏⲥⲧⲓⲁ, ⲧⲛⲏⲥⲧⲓⲁ ⲱ)62 as remembrance of the crucifixion. The Armenian ourpat (ուրբաթ) is an adaption the corresponding Syriac or Jewish-Aramaic word. The Gǝʿǝz ʿarb is etymologically related to ʿrūḇtā / ʿărūḇtā, but not borrowed from these Aramaic idioms. It is derived from the middle Hebrew name for the day. This is äräb šabbat or äräḇ šabbaṯ63 (with the double pronounciation of the begadkefat from the 1st BC on),64 the preparation day for the Sabbath. When the name was borrowed, it was shortenend and the second part of the genitive was omitted.65 In a Christian parish such a day has lost its meaning. A preparation for the day of the Lord can not take place on Friday. The only explanation for such a borrowing is due to the fact that numerous members of the Jewish community or “God-fearers” as sympathizers to Judaism changed their religion, but not the vocabulary for certain Termini technici. Some of the former words for religious matters were kept up and subsequently maintained. Although the Christians of the Nile valley invented two other names for Friday which were conform to the new religion, nobody in Ethiopia had the impression that the name of the Friday should be changed similar to Egyptian tradition and be called ‘the big fast day’, although Christians in Ethiopia and Egypt observe two fasting days every week (Wednesday and Friday) which is a Jewish practice (Tuesday and Thursday). Probably the name for Friday was understood as fasting day, because the original meaning was more or less forgotten.66 Sometimes another word is used for Friday which corresponds to the usual counting of the weekdays with numbers.

58  Another name for Friday is maʿălē/meʿālē šabbṯā, which literally means ‘entrance, commencement of the Sabbath’ (Margolis 1910: 147*, 169*; Dalman 1938: 246; Jastrow 1886– 1903: 817; Levy 1924: 3.193). 59  Shooshtary 2012: 104; MD 346. 60  Böhlig 1958: 393. 61  Bauer 1988: 1257. 62  Till 1961: 89 (§180); Westendorf 1965–1977: 203.138; Crum 1939: 368. 63  Dalman 1938: 322; Jastrow 1886–1903: 1111; Levy 1924: 3.691b. 64  Beyer 1984: 126–128. 65  The root √ʿrb is attested in Geʿez, but there is no word meaning ‘evening’ or like that. 66  Cf. arbä räbuʿ (= fasting days), the name for Wednesday and Friday in Tǝgrǝñña (Kane 2000: 2.1847).

Aramaic Loanwords in Gǝ ʿ ǝz

357

This is sadūs (ሰዱስ)67 which can mean either ‘the sixth day, the sixth day of the week (= Friday)’ or ‘six days’, because the nominal form qatūl is used for the ordinal and cardinal numbers. The verb aṭhara (አጥሀረ II.1) which means ‘to clean, purify, lustrate’ in a religious sense,68 has a strong Jewish background and is influenced by Jewish practice. Purity plays an important role in Judaism and is associated with many lustration rites. The same customs occur in Ethiopian Christianity. According to the law of the Pentateuch there exist clean and unclean animals. These prescripts and the laws for cultic purity are sometimes observed very strictly although it is stressed in many texts that the ‘new law’ of the New Testament abrogates the old law. The word ṭāʿōt (ጣዖት) ‘idol, false god, ungodliness’69 is borrowed from the Hebrew ṭāʿūt70 or the Jewish-Aramaic ṭāʿūṯā ‘error, idol’.71 The corresponding word in Syriac adds the letter yōḏ and can be excluded for this reason. If the Syriac ṭāʿyūṯā72 would have been the linguistic model, it should be ṭāʿyōt in Gǝʿǝz. The Syriac ṭāʿyūṯā normally only means ‘error, ignorance, forgetting’ and not ‘idol’ as in Hebrew. The word mǝṣwāt (ምጽዋት) ‘almsgiving, alms, charity’73 is derived from the Hebrew miṣwā ‘order, commandment, obligation’74 in an Aramaic form (miṣwǝṯā).75 It is normally explained as plural of the Jewish Aramaic miṣwǝṯā which got the additional meaning ‘almsgiving, alms’ when it came into Aramaic. The plural miṣwāṯā (Hebrew miṣwōt) is the archetyp, the model of the Gǝʿǝz form. But it cannot completely be excluded that mǝṣwāt is a shortening of an Aramaic miṣwāt ʾalāhā (< Hebrew miṣwōt ʾälōhīm/YHWH) in the sense that almsgiving is a divine obligation. The word has no counterpart in Syriac and must have found its way into Gǝʿǝz from a Jewish environment or more specifically from the synagogue. 67  Chaîne 1907: 94 (§171); Leslau 1987: 486; 1989: 76. Dillmann (1865: 395) and Leslau (1961) do not mention this name for Friday. 68  Dillmann 1865: 1213 (purgare, purificare; lustrare, de lustrationibus Judaeorum sacris); cf. Leslau 1987: 589; 1989: 214; Grébaut 1952: 341; Nöldeke 1910: 36; cf. Dillmann 1866: 274. 69  Dillmann 1865: 1243; 1866: 277; Leslau 1987: 484; 1989: 220; Nöldeke 1910: 35. 70  Dalman 1938: 172; Jastrow 1886–1903: 542; Levy 1924: 2.170f. 71  Dalman 1938: 172; Jastrow 1886–1903: 542; cf. Levy 1924: 2.171. 72  Payne Smith 1879–1901: 1.1494; Payne Smith 1903: 178; LS 282; SL 542. 73  Dillmann 1865: 228; 1866: 182; Grébaut 1952: 118; Leslau 1987: 371; 1989: 47; Nöldeke 1910: 36; Polotsky 1964: 5f; Altheim and Stiehl 1964–1969: 3.30. 74  KBL2 556f; KBL3 588f; HALOT 2.622; Gesenius 1987–2010: 724; Dalman 1938: 248; Jastrow 1886–1903: 823f; Levy 1924: 3.207f. 75  Dalman 1938: 248; Jastrow 1886–1903: 824; Levy 1924: 3.208.

358

tubach

Gǝʿǝz knows several words for ‘priest’. ʾAklīrōs (አክሊሮስ76 < Arabic ʾiklīrūs77) and qlīrīqōs/ qǝlīrīqōs (ቅሊሪቆስ)78 are the Greek words klēros (κλῆρος, Itacistic pronunciation [kliros], Syr. qlērōs) and klērikos (Syr. qlērīqā),79 while šawāʿī (ሠዋዒ) ‘priest, sacrificer’ is a genuine Gǝʿǝz word.80 The common word for ‘priest’ in Syriac is qaššīšā (and qaššā).81 It originated from the adjective ‘old’. The Gǝʿǝz word qassīs (ቀሲስ) and qas (ቀስ)82 has the same meaning including ‘presbyter’ and ‘elder’ as in Syriac. The original meaning of the adjective (= old) was dropped. Another very common word for priest is kāhǝn (ካህን)83 in Gǝʿǝz. It has several derivates and is fully integrated into the phonetic vocabulary. The latter word was taken over from Hebrew but not directly. Gǝʿǝz kāhǝn goes back to the Hebrew kōhen,84 but it was imparted by Jewish-Aramaic (kāhănā)85 or Syriac (kāhnā, ABS kāhen).86 The normal word for priest is in Aramaic komar87 and in Syriac kumrā88 (Hebrew komär ‘pagan priest’89 > Jewish Aramaic and Mandaic: kumrā ‘idol-priest’90). The Aramaic qaššīš91 (> Sabaic qassīs or qissīs92 76  Dillmann 1865: 785; Leslau 1987: 15. 77  Wehr 1985: 34. The borrowing is late and comes from the Middle Ages. The Greek letter κ does not correspond to the Semitic k in ancient Greek transcriptions of Semitic names. The elimination of the double consonance is a feature of classical Arabic, but occurs in other Semitic languages as well. 78  Leslau 1987: 430. 79  Payne Smith 1879–1901: 2.3633; Payne Smith 1903: 507; LS 671; SL 1371. The Greek κ is transliterated as q in Antiquity; only before the aspiration of the Semitic letter k (earlier than 250 BC), it could express the Greek kappa (Beyer 1984: 125f). 80  Dillmann 1865: 256; 1866: 185; Grébaut 1952: 133; Leslau 1987: 538; 1989: 53. 81  Payne Smith 1879–1901: 2.3765.3766f; Payne Smith 1903: 522; LS 702; SL 1418, 1419. 82  Dillmann 1865: 431, 433; Grébaut 1952: 175; Leslau 1987: 447; 1989: 85; Nöldeke 1910: 37; cf. Dillmann 1866: 205. 83  Dillmann 1865: 813f; 1866: 239; Grébaut 1952: 255; Leslau 1987: 278; 1989: 148; Nöldeke 1910: 36. 84  KBL2 424f; KBL3 440f; HALOT 2.461f; Gesenius 1987–2010: 529f; Dalman 1938: 193; Jastrow 1886–1903: 615; Levy 1924: 2.299; cf. DNWSI 490–492. 85  Dalman 1938: 193; Jastrow 1886–1903: 615; Levy 1924: 2.300; cf. Beyer 1984: p. 603. 86  Payne Smith 1879–1901: 1.1683; Payne Smith 1903: 206; LS 319; SL 601. 87  Jean and Hoftijzer 1965: 122; DNWSI 515f; Beyer 1994: 364. 88  Payne Smith 1879–1901: 1.1757f; Payne Smith 1903: 209; LS 332; SL 608. 89  KBL2 442; KBL3 459; Gesenius 1987–2010: 552f; Dalman 1938: 194; Jastrow 1886–1903: 621; Levy 1924: 2.346; Phoenician: Jean and Hoftijzer 1965: 122; DNWSI 515f. 90  Dalman 1938: 201; Jastrow 1886–1903: 621; Levy 1924: 2.346; MD 207. In Jewish Aramaic there exists the vocalisation kummārā. 91  Beyer 1984: 688; Smith 1879–1901: 1.3766; Payne Smith 1903: 522; LS 702; SL 1419; Dalman 1938: 394; Jastrow 1886–1903: 1431. 92  Beeston 1994: 42; Gajda 2010: 111.

Aramaic Loanwords in Gǝ ʿ ǝz

359

and Arabic qissīs93) or the Greek presbyteros (πρεσβύτερος)94 is one of the ‘elders’ of the Jewish and the early Christian community. The office of the Christian presbyteros/qaššīš or ‘elder’ was specialized and the word became a synonym for the ‘priest’ of a Christian parish. In this meaning it is attested in late Sabaic. The English or German word priest or Priester originated from presbyteros. Probably the Gǝʿǝz word kāhǝn comes from the Jewish community because it is not so common in Syriac Christianity. The Ethiopian word for hell gahannam (or gahānnam, ገሀነም, ገሃነም)95 is derived from the Hebrew Gē Hinnōm,96 a valley south of Jerusalem. Because children were burnt as sacrifice for the god Moloch (= Greek; Hebrew Moläk / Moläḵ)97 in that valley (2Kgs. 23:10) the metaphoric meaning hell arose in the pseudepigraphic literature in post-Old-Testamental times. Already in the Gospels γέεννα, the Greek transcription, is a common word for hell.98 The form of the word supposes that the last consonant of the Jewish-Aramaic Gē Hinnām99 was dropped. The Semitic gutturals cannot be expressed in script and are often omitted as in the Greek loanword used in the Gospels. Γέεννα is derived from a word which must be identical with the Syriac gēhannā (> West Syriac gīhanō, gīnō)100 with the absence of the final mēm. The Ethiopic gahannam and the Arabic ǧahannamu101 are borrowed more or less directly from the Jewish-Aramaic Gē Hinnām (< Hebr. Gē Hinnōm). 4

Jewish or Christian Loanwords

Many loanwords of the religious realm can be domiciled either in a Jewish or Christian community, because the converters did not invent new theological expressions, but used the old and familiar one in their new religious community. The main difference was the restricted validity of the law and the coming 93  Kawerau 1976: 118; Wehr 1985: 1023. 94  Bauer 1988: 1402f. 95  Dillmann 1865: 1137; 1866: 267; Grébaut 1952: 327; Leslau 1987: 186; 1989: 200; Nöldeke 1910: 34f. 96  KBL2 179; KBL3 181a; HALOT 1.188b; Gesenius 1987–2010: 212; Dalman 1938: 76; Jastrow 1886–1903: 236; Levy 1924: 1.323. 97  KBL2 560; HALOT 2.592; Gesenius 1987–2010: 686; Levy 1924: 3.131; Heider 1985; Day 1989; Moscati 1991. 98  Bauer 1988: 306. 99  Dalman 1938: 76; Jastrow 1886–1903: 236; Levy 1924: 1.323. 100  Payne Smith 1879–1901: 1.707; Payne Smith 1903: 68; LS 106f; SL 229. 101  Jeffery 1938: 105f.

360

tubach

of the Messiah. Most of the expressions with a theological or religious connotation could easily be kept. Several of these words have Hebrew ancestors. A new word appearing in the vocabulary is wangēl ‘gospel’ (ወንጌል).102 Originally it is taken from the Greek εὐαγγέλιον [evangelion] (in the pronunciation of Greek at the time of Christ)103 and shortened at the beginning and the end (= vangel). Probably it was taken over by Christians in South Arabia from Aramaic/Syriac (ʾewangelīyōn),104 before it came to the Axumite empire. Otherwise one must assume that Christian merchants conveyed the word to their Ethiopian stewards. In classic Arabic it appears as ʾinǧīl105 in the nominal form qittīl which is typical for many loanwords from the religious sphere (e.g., Ǧibrīl, ʾIdrīs, qissīs). Other words like nabīy (ነቢይ),106 ṣalōt (ጸሎት),107 and masīḥ (መሲሕ)108 are borrowed from Hebrew or Aramaic before the 3rd century AD. These words are used by the Jewish or Christian community. They are important, but not specific words of both groups which characterize one of these communities without any doubts. The borrowing of these words must have taken place before the first half of the 3rd century AD, when the unstressed short vowels in open syllables were dropped in Aramaic, resp. Syriac.109 lf they were taken over after this date, they should be written nǝbīy, ṣǝlōt, and mǝsīḥ in Gǝʿǝz according to the Syriac nḇīyā, ṣlōṯā, mšīḥā, but in the status absolutus or constructus. The Gǝʿǝz language can express consonants without a vowel or with a so called murmured vowel in the orthography. The list with words of Hebrew or Aramaic origin is not exhaustive.110 They especially turn up in the religious vocabulary, but there are several examples from the secular field as well. Furthermore Christianity in Ethiopia possesses many striking characteristics which can be labelled as Jewish in a general sense.111 The possibility that all these elements are taken from the Old Testament can be excluded because of linguistic reasons. The mentioned 102  Dillmann 1865: 918; 1866: 248; Leslau 1987: 615; 1989: 164. 103  Bauer 1988: 643f. 104  Payne Smith 1879–1901: 1.71; Payne Smith 1903: 6; LS 8; SL 17. 105  Jeffery 1938: 71f. 106  Dillmann 1865: 658; 1866: 226; Grébaut 1952: 219; Leslau 1987: 385; 1989: 123; 1958: 32; Nöldeke 1910: 34. 107  Dillmann 1865: 1263; 1866: 278; Grébaut 1952: 355; Leslau 1987: 556, 557; 1989: 224; Nöldeke 1910: 34. The term occurs in Late Sabaic (Beeston 1994: 42; Gajda 2010: 110). 108  Dillmann 1865: 177; 1866: 177; Leslau 1987: 363; 1989: 34; 1958: 32; Nöldeke 1910: 34. The word is attested in Late Sabaic as well (Beeston 1994: 42; Gajda 2010: 111). 109  Beyer 1984: 128–136. 110  Nöldeke 1910: 32–46; Polotsky 1964; Altheim and Stiehl 1964–1969: 3.29–32; cf. Witakowski 1989–90: 191f, who wants to declare some words as Syriac (similar Heyer 1971: 218). 111  Ullendorff 1956; Hammerschmidt 1963.

Aramaic Loanwords in Gǝ ʿ ǝz

361

loanwords contradict such an assumption. They are inherited from the Jewish community in Ethiopia, when either full members of Judaism or the so called god-fearers112 changed their religion and turned to Christianity. They kept a great part of their old and familiar religious vocabulary, even some words which make no sense in the new Christian environment. These words can not have their origin in the missionary activity of Frumentios113 and the Nine Roman Saints.114 Otherwise they would have brought a vocabulary to the Axumite empire which is typical for a Jewish community and which did not exist in the homeland of these saints in this manner as a whole. The Hebrew and Aramaic words with a special Jewish connotation—except other Jewish elements—require Jewish communities in the Axumite empire. lf this were not the case, the number of these words is not explainable. It should not exceed the number of such loan words in Greek, English, or another European language. The postulated Jewish background of Ethiopian Christianity corresponds to the Axumite tradition as heirs of the Old Testament legacy. The result which can be drawn from the loanwords, is a confirmation of the Tradition of the Kǝbra Nagašt115 (and other texts), that the majority of the Ethiopians were adherents of the Old Testamental belief before the introduction of Christianity and not pagans. In the national legend of Ethiopia Solomon’s son Menelik and his mother, the queen of Sheba, the queen of the South, introduced the belief of the Old Testament resp. Judaism in Ethiopia. The Ethiopians are the true Israelites which did not later reject Christ and his message. 5 Conclusion Can such a reconstruction of the Ethiopian past claim to be true? An exact counterpart and parallel is South Arabia with a similar development. In the motherland of the Axumites both religions, Judaism and Christianity, are well attested. The religious vocabulary is not known in detail,116 but it must be the 112  Hegermann 1990: 307–314. 113  Dihle 1965; Dinkler 1976; 1977; Heyer 1971: 217ff; Sergew 1972: 97–113; Kaplan 1982; MunroHay 1980; 1988; 1999; Dombrowski and Dombrowski 1984; Rodinson 2001; Brakmann 1994: 51–77, 117–121; 1999: 408–413 (=20–25); 2001: 745–751, 766–769; Murav’yov 2009; Haas 2008; Hatke 2013: 89, 91, 93, 96, 164; Phillipson 2012: 91ff; Bowersock 2013: 63ff. 114  Sergew 1972: 115–121; Brakmann 1994: 127–132; 1999: 426–429 (=38–40); 2001: 772–777; Marrassini 1999; Brita 2007. 115  Edited in Bezold 1909; translations in Budge 1932; Strelcyn 1956; Brooks 1996; Colin 2002; Mahler 2007; Beylot 2008; Raineri 2008. 116  Cf. fn. 25.

362

tubach

same as in Ethiopia or partly in the Qurʾan with the same Jewish and Christian background. The Jewish communities in Axum must have their origin in the Sabaean and Himyarite kingdoms.117 One further close parallel exists between Himyar and Axum: in both cases the ‘Lord of heaven’ (ʾǝgzīʾa samāy/mrʾ smyn) is invoked in inscriptions.118 This is a neutral phrase, acceptable for Jews and Christians (Dan 5:23, cf. Gen 24:3, 7; Dan 2:18f, 37:44; Jon 1:9, etc.). The political and cultural connections between Axum and the motherland were very close and sometimes very difficult and coined by tensions, especially when Axurn’s sovereigns tried to conquer South Arabia. In 525 Caleb (Kālēb, throne name: Ǝlla Aṣbǝḥa/ Aṣbǝḥā, ca. 510–540)119 succeeded in doing this. He defeated Yūsuf Asʾar Yaṯʾar (517 or 522–525), the last Jewish king of Himyar.120 Sometimes the Axumites were in possession of parts of Arabia and stationed military garrisons there. The Axumite army fought at least in the beginning of the 3rd century in South Arabia.121 It can be assumed that Axumites in South Arabia came into contact with Jews and converted or that Jewish merchants brought their religion to northern Ethiopia. From the 4th century on the Himyarite kings were either full members or sympathizers of Judaism.122 In any case they fostered the Jewish religion and no longer the old pagan belief. Judaism and Christianity came to South Arabia very early. Christians existed there at least in the first half of the 4th century, if not earlier. Judaism was extant in South Arabia much earlier, the Jewish religion found its way to Sabaean kingdom probably already in pre-Christian times. The Jewish faith is the dominant religion in the 4th century in South 117  Scholz (1988: 317, 320) thinks that Jewish missionary activities and immigrants had their origin in the Nile valley (Elephantine, Meroe) and South Arabia (cf. Altheim and Stiehl 1964–1969: 3.19f, 21f, 32). 118  R IÉ I 263–265 (no. 189, ln. 1, 5, 38f, 40f, 45, 45f, 52); Littmann 1913: 32–35 (no. 11); 1950: 115, 125f (no. 10); Ryckmans 1946; 1958; see now esp. Gajda 2009; 2010. 119  Hatke 2011; Hahn 2000; 2006; Fiaccadori 2003; Brakmann 2001: 753–762; 1999: 413–424 (=25–36); 1994: 81–109; Harmatta 1974; Letsios 1991; Drewes 1978; 1987; Christides 1972; Sergew 1972: 123–158; Altheim and Stiehl 1964–1969: Vol. V.2, 199–202 et passim; 1971: 432ff, 443ff, 447f. 120  Power 2012: 68–74; Hatke 2011; Robin 2008; Altheim 1969; Altheim and Stiehl 1964–1969: Vol. V.1, 305–391; 1971: 440–447; Engelhardt 1974: 171ff; Beeston 1989: 1–6; 1975: 124–126; cf. Nebes 2008; 2010; Al-Asali 1968; Bowersock 2013: 92ff. 121  The Axumite king Gadarat (or Gadara) and his successor ʿAḏaba controlled parts of South Arabia, especially the region of Nagrān at the so-called incense road. For a short time Gadarat was in possession of the Himyarite capital Ẓafār and the royal palace of Raydān (Brakmann 1994: 12–18; 2001: 722–724; Sima 2005; Robin 1989; cf. Yule 2013; 2014). Later the Axumite kings did not give up the claim to be rulers of South Arabia. 122  Garbini 1996; Lecker 1995; Robin 2003; 2004; cf. Lundin 1999; Frantsouzoff 1999; Bowersock 2010; 2013: 78ff; Rubin 1995; 2000; Gajda 2009; 2010; Stein 2010.

Aramaic Loanwords in Gǝ ʿ ǝz

363

Arabia. At that time the circumstances in Axurn changed. The king and his family became Christians. Many people especially the god-fearers adopted the new religion. But not all wanted to change their religion, a rest remained and adhered to their belief faithfully. They did not want to convert and retreated to the Southern highland of Ethiopia.123 These were the Bēta ʾƎsrāʾēl (ቤተ እስራኤል) or Falashas (Amharic ፈላሻ),124 as they are called by their Christians neighbours. Their retreat into the southern highland of Ethiopia separated them from the contact to the Jewish world.125 Bibliography Abbink, Jon. 1990. “The Enigma of Beta Esra’el Ethnogenesis: An Anthro-Historical Study,” Cahiers d’études africaines 30: 397–449. Aistleitner, Joseph. 1963. Wörterbuch der ugaritischen Sprache. Berlin. Altheim, Franz. 1969. “Übersehenes zur Geschichte Ḏū Nuwās,” in his Geschichte der Hunnen, Vol. 2. Die Hephthaliten in Iran (2nd ed.). Berlin. 40–46. Altheim, Franz and Ruth Stiehl. 1964–1969. Die Araber in der alten Welt. Berlin. ———. 1971. Christentum am Roten Meer. Berlin. Al-Asali, Khalid Salih. 1968. South Arabia in the 5th and 6th centuries CE, with reference to relations with Central Arabia. Ph.D. Diss., University of St. Andrews. 123  Messing 1982: 13. 124  Leslau 1976: 244. The word is either a synonym for gōlā/gālūt (‫ּגו ָֺלה‬/‫ )ּגָ לּות‬or means προσήλυτος, if it can be derived from Ge‛ez falās/ falāsī (Dillmann 1865: 1341f). 125  Cf. Jones-Monroe 1935: 39–41. It is often assumed that the Falashas arose in the Middle Ages and have no (real) connection with the normally postulated Jewish communities in Axumite times (e.g., Brakmann 1994: 43–50, 172–165, esp. 49; 2001: 743–745, 794–799; Kaplan 1992: 5–10, 13–32, esp. 9, 32, 55, 157; 1993; Pankhurst 1992: 582; Shelemay 1986: 17–38, esp. 2, 20ff, 205ff; Quirin 2010; Abbink 1990; Krempel; complete refusal of this theory: Teferi 2005). Sometimes the existence of Jewish communities in Axumite times is declared as unprovable (Abbink 1990: 405: “If they were there, they have not left any traces in language, writing and architecture . . .”). In this case the Falashas are a kind of Christian sect, Judaized Christians with a special reverence of the Old Testament retaining all the Hebrew-Jewish elements of their former faith, the Ethiopic Christianity. But only Christians who first were members of the synagogue would have the idea to maintain customs of their former faith, especially when such things do not contradict the new views. Christianity was for them a completion of the old faith, the prophecies of the Old Testament were fulfilled. It was not necessary to wait for the coming of the Messiah. If the connection between the Falashas and the postulated Jewish communities in Axumite times is cut, it is difficult to explain why Christians became Jews (cf. Teferi 2005: 179–182). Kessler (1996: 42–49, 63f) assumes that the Jewish communities in South Arabia and Axum have an independent origin. The precursors of the later Falashas came from Meroe/Egypt.

364

tubach

Bard, Kathryn A. 2008. An Introduction to the Archaeology of Ancient Egypt. Malden. Bauer, Walter (ed. Kurt Aland und Barbara Aland). 1988. Griechisch-deutsches Wörterbuch zu den Schriften des Neuen Testaments und der frühchristlichen Literatur (6th ed.). Berlin. Bausi, Alessandro. 2003. “Numismatica Aksumita, Linguistica e Filologia,” Annali dell’Istituto Italiano di Numismatica 50: 157–175. Beeston, A. L. F. 1975. “The Realm of King Yusuf (Dhu Nuwas),” BSOAS 38: 124–126. ———. 1983. “The Chain of al-Mandab,” in Ulla Ehrensvärd and Christopher Toll (eds.), On both Sides of al-Mandab. Ethiopian, South-Arabic and Islamic Studies presented to Oscar Löfgren on his ninetieth birthday 13 Mai 1988 by colleagues and friends. Istanbul – Stockholm. 1–6. ———. “Foreign Loanwords in Sabaic,” in Norbert Nebes (ed.), Arabia Felix. Beiträge zur Sprache und Kultur des vorislamischen Arabien. Festschrift Walter W. Müller zum 60. Geburtstag. Wiesbaden. 39–45. Beeston, A. F. L., M. A. Ghul, W. W. Müller, J. Ryckmans. 1982. Sabaic Dictionary (EnglishFrench-Arabic) / Dictionnaire sabéen (anglais – français – arabe). Louvain-la-Neuve – Beirut. Bernand, É., A. J. Drewes, and R. Schneider (eds.). 1991–2000. Recueil des inscriptions de l’Éthiopie des périodes pré-axoumite et axoumite, Vol. 1–3. Paris. Berzina, S. Ja. 1984. “Kushana coins in Axum,” International Association for the Study of the Cultures of Central Asia. Information Bulletin 7: 57–64. Beyer, Klaus. 1984. Die aramäischen Texte vom Toten Meer. Göttingen. ———. 1994. Die aramäischen Texte vom Toten Meer, Ergänzungsband. Göttingen. ———. 1998. Aramäische Inschriften aus Assur, Hatra und dem übrigen Ostmeso­ potamien. Göttingen. Beylot, Robert. 2008. La Gloire des Rois ou l’Histoire de Salomon et de la reine de Saba (Apocryphes. Collection de poche de l’AELAC 12). Brepols. Bezold, Carl. 1909. Kebra Nagast. Die Herrlichkeit der Könige. Nach den Handschriften in Berlin, London, Oxford und Paris zum ersten Mal im äthiopischen Urtext herausgegeben und mit deutscher Übersetzung versehen (Abhandlungen der philosophischphilologischen Klasse der Königlich Bayerischen Akademie der Wissenschaften, 23. Band, in der Reihe der Denkschriften der LXXVII. Band. I. Abteilung). Munich. Biella, Joan Copeland. 1982. Dictionary of Old South Arabic, Sabaean Dialect (HSS 25). Chico. Böhlig, Alexander. 1958. Die griechischen Lehnwörter im sahidischen und bohairischen Neuen Testament (2nd ed.) (Studien zur Erforschung des christlichen Aegyptens 2). Munich. Bowersock, G. W. 2010. “The new Greek inscription from South Yemen,” in Jean-François Salles (ed.), Qāni’. Le port antique du Ḥaḍramawt entre la Méditerranée, l’Afrique et l’Inde. Fouilles russes 1972, 1985–1989, 1991, 1993–1994. Brepols. 393–396.

Aramaic Loanwords in Gǝ ʿ ǝz

365

———. 2013. Throne of Adulis. Red Sea wars on the eve of Islam. Oxford. Boyce, Mary. 1977. A Word-list of Manichaean Middle Persian and Parthian with a Reverse Index. Leiden. Brakmann, Heinzgerd. 1994. ΤΟ ΠΑΡΑ ΤΟΙC ΒΑΡΒΑΡΟΙC ΕΡΓΟΝ ΘΕΙΟΝ. Die Einwurzelung der Kirche im spätantiken Reich von Aksum. Bonn. ———. 1999. “Religionsgeschichte Aksums in der Spätantike,” in Piotr O. Scholz (ed.), Äthiopien gestern und heute. Akten der 1. Tagung der Orbis Æthiopicus Gesellschaft zur Erhaltung und Förderung der äthiopischen Kultur (Nubica et Æthiopica IV/V). Warszaw. 401–430. (= 13–42) ———. 2001. “Axomis (Aksum),” in Reallexikon für Antike und Christentum. Sachwörterbuch zur Auseinandersetzung des Christentums mit der Antiken Welt, Supplement-Band I. Aaron-Biographie II. Stuttgart. 718–810. Brita, Antonella. 2007. “Nine Saints,” in Siegbert Uhlig (ed.), Encyclopaedia Aethiopica, Vol. 3. Wiesbasen. 1188–1191. ———. 2010. I racconti tradizionali sulla « seconda cristianizzazione » dell’Etiopia. Il ciclo agiografico dei nove santi (Studi Africanistici Serie Etiopica 7). Naples. Brockelmann, Carl. 1928. Lexicon Syriacum (2nd ed.). Halis Saxonum. Brooks, Miguel F. 1996. A modern translation of the Kebra nagast (The glory of kings). The history of the departure of God and His Ark of the Covenant from Jerusalem to Ethiopia, and the establishment of the religion of the Hebrews and the Solomonic line of kings in that previously pagan country. Compiled, edited and translated. Lawrenceville. Buck, Carl Darling. 1949. A Dictionary of Selected Synonyms in the Principal IndoEuropean Languages. A Contribution to the History of Ideas. Chicago. Budge, E. A. Wallis. 1932. The Queen of Sheba and her only Son Menyelek (I) being the “Book of the Glory of Kings” (Kebra Nagast), a work which is alike the traditional history of the establishment of the religion of the Hebrews in Ethiopia, and the patent of sovereignty which is now universally accepted in Abyssinia as the symbol of the divine authority to rule which the kings of the Solomonic line claimed to have received through their descent from the house of David. Translated from the Ethiopic (2nd ed.). Oxford. Casson, Lionel. 1989. The Periplus Maris Erythraei. Text with Introduction, Translation, and Commentary. Princeton. Chaîne, Marius. 1907. Grammaire éthiopienne. Beirut. Christides, Vassilios. 1972. “The Himyarite-Ethiopian War and the Ethiopian Occupation of South Arabia in the Acts of Gregentius (ca. 530 AD),” Annales d’Éthiopie. Revue d’archéologie, de philosophie et d’histoire 9: 115–146. Colin, Gerárd. 2002. La Gloire des rois (Kebra nagast), épopée nationale de l’Éthiopie, traduction française intégrale (Cahiers d’orientalisme 23). Geneva. Crum, W. E. 1939. A Coptic Dictionary. Oxford.

366

tubach

Dalman, Gustaf, H. 1938. Aramäisch-neuhebräisches Handwörterbuch zu Targum, Talmud und Midrasch (3rd ed.). Göttingen. Day, John. 1989. Molech. A God of Human Sacrifice in the Old Testament (University of Cambridge Oriental Publications 41). Cambridge. Day, Rebecca. 2011. “Imitation in Aksumite coinage and Indian imitations of Aksumite coins,” Rosetta. Papers of the Institute of Archaeology and Antiquity, University of Birmingham 9.5: 16–22. Dihle, Albrecht. 1965. “Frumentios und Ezana,” in his Umstrittene Daten. Untersuchungen zum Auftreten der Griechen am Roten Meer (Wissenschaftliche Abhandlungen der Arbeitsgemeinschaft für Forschung des Landes Nordrhein-Westfalen 32). Köln – Opladen. 36–64. Dillmann, A. 1865. Lexicon linguae aethiopicae cum indice latino. Leipzig. ———. 1866. Chrestomathia Aethiopica. Leipzig. Dinkler, Erich. 1976. “New Questions Concerning King Ezana of Axum,” Etudes et Travaux. Centre d’Archéologie Méditerranéenne de l’Académie Polonaise des Science / Studia i prace Zakład Archeologii Śródziemnomorskiej Polskiej Akademii Nauk 9: 6–15. ———. 1977. “König Ezana und das Christentum. Ein Randproblem der Geschichte Nubiens,” in Erika Endesfelder, Karl-Heinz Priese, Walter-Friedrich Reineke, and Steffen Wenig (eds.), Ägypten und Kusch. Sammelband Fritz Hintze zum 60. Geburstag gewidmet (Schriften zur Geschichte und Kultur des Alten Orients 13). Berlin. 121–132. Drewes, A. J. 1978. “Kaleb and Himyar. Another reference to ḤYWN?,” Raydān. Journal of ancient Yemeni antiquities and epigraphy 1: 27–30. ———. 1987. “Ethiopië en Zuid-Arabië tussen 450 voor en 550 na het begin van onze jaartelling,” Phoenix. Bulletin uitgegeven door het Vooraziatisch-Egyptisch Genootschap “Ex Oriente Lux” 33: 55–71. Dombrowski, B. W. W. and Franz Amadeus Dombrowski. 1984. “Frumentius/Abbā Salāmā. Zu den Nachrichten über die Anfänge des Christentums in Äthiopien,” Oriens Christianus 68: 114–169. Drower, E. S. and R. Macuch. 1963. A Mandaic dictionary. Oxford. Durkin-Meisterernst, Desmond. 2004. Dictionary of Manichaean Texts, Vol. 3. Texts from Central Asia and China, Part 1. Dictionary of Manichaean Middle Persian and Parthian (Corpus Fontium Manichaeorum. Subsidia). Brepols. ———. 2014. Grammatik des Westmitteliranischen (Parthisch und Mittelpersisch). Wien. Engelhardt, Isrun. 1974. Mission und Politik in Byzanz. Ein Beitrag zur Strukturanalyse byzantinischer Mission zur Zeit Justins und Justinians (Miscellanea Byzantina Monacensia 19). Munich. Fiaccadori, Gianfranco. 2003. “Kaleb,” in S. Uhlig (ed.), Encyclopaedia Aethiopica, Vol. 1. Wiesbaden. 329–332.

Aramaic Loanwords in Gǝ ʿ ǝz

367

Frantsouzoff, Serguei A. 1999. “Judaism in Hadramaut on the eve of Islam,” in Ephraim Isaac and Yosef Tobi (eds.), Judaeo-Yemenite Studies. Proceedings of the Second International Congress. Princeton. 27–40. Frisk, Hjalmar. 1960. Griechisches etymologisches Wörterbuch, Vol. 1. Heidelberg. Gajda, Iwona. 2009. Le royaume de Ḥimyar à l’époque monothéiste. L’histoire de l’Arabie du Sud ancienne, de la fin du IVe siècle de l’ère chrétienne jusuq’à l’avènement de l’Islam. (Mémoires de l’Académie des inscriptions et belles-lettres, nouvelle série 40). Paris. ———. 2010. “Quel monothéisme en Arabie du Sud ancienne?,” in Joëlle Beaucamp, Françoise Briquel-Chatonnet, and Christian Julien Robin (eds.), Juifs et chrétiens en Arabie aux Ve et VIe siècles. Regards croisés sur les sources. Paris. 107–122. Garbini, Giovanni. 1996. “La dinastia di Malkikarib Yuha’min, il primo re di Saba giudeo,” Atti délia Accademia nazionale dei Lincei, Serie IX, Anno CCCXCIII, Classe di Scienze morali, storiche e filologiche. Rendiconti, Vol. 7, fasc. 2, Rome. 237–242. Gates, Jennifer Erin. 2005. Traveling the Desert Edge. The Ptolemaic Roadways and Regional Economy of Egypt’s Eastern Desert in the Fourth through First Centuries BCE. Ph.D. Diss., University of Michigan, Ann Arbor. Gesenius, Wilhelm (ed. Rudolf Meyer and Herbert Donner). 1987–2010. Hebräisches und Aramäisches Handwörterbuch über das Alte Testament (18th ed.). Berlin – New York. Godet, Eric. 2004. Le monnayage de l’Éthiopie ancienne ( fin du IIIe – milieu du VIIe siècle après Jésus-Christ). Catalogue raisonné des monnaies axoumites du Musée national d’Addis-Abeba. Ph.D. Diss, Université Pantheon-Sorbonne, Paris. Göbl, Robert. 1970. “Der kusanische Goldmünzschatz von Debra Damo (Äthiopien) 1940. (Vima Kadphises bis Vasudeva I.),” Central Asiatic Journal 14: pp. 241–252. Grébaut, Sylvain. 1952. Supplément au Lexicon linguae aethiopicae de August Dillmann (1865) et édition du lexique de Juste d’Urbin (1850–1855). Paris. Haas, Christopher. 2008. “Mountain Constantines: The Christianization of Aksum and Iberia,” Journal of Late Antiquity 1: 101–126. Hahn, Wolfgang. 1994. “Déclinaison et orthographe des légendes Grecques sur les monnaies d'Axoum,” Bulletin de la Société Française de Numismatique 49 no. 9: 944–948. ———. 2000. “Aksumite Numismatics—a critical survey of recent Research,” Revue Numismatique 155: 281–311. ———. 2000. “Der Heilige Kaleb Ella Asbeha—König des Abessinierlandes und seine Münzen,” moneytrend. Das Münzenmagazin 32.3: 60–67. ———. 2000. “Nachtrag zu Kaleb Ella Asbeha, König des Abessinierlandes,” moneytrend. Das Münzenmagazin 32.4: 50–51. ———. 2006. “Ezanas and Caleb, the Pair of Saintly Kings,” in Siegbert Uhlig (ed.), Proceedings of the XVth International Conference of Ethiopian Studies, Hamburg, July 20–25, 2003 (ÄF 65). Wiesbaden. 260–265.

368

tubach

———. 2010. “BĀXCACA. Eine neuentdeckte Legendenvariante auf einer aksumitischen Silbermünze des anonymen Typs bekräftigt die Lesung des Titels König der Abessinier,” Mitteilungsblatt des Instituts für Numismatik und Geldgeschichte der Universität Wien 40: 8–9. Halévy, J. 1896. “Traces d’influence indo-parsie en Abyssinie,” Revue sémitique d’épigraphie et d’histoire ancienne 4: 258–265. Hammerschmidt, Ernst. 1963. Stellung und Bedeutung des Sabbats in Äthiopien (Studia Delitzschiana. Abhandlungen und Texte aus dem Institutum Judaicum Delitzschianum, Münster [Westfalen] 7). Stuttgart. Harmatta, János. 1974. “The Struggle for the Possession of South Arabia between Aksūm and the Sāsānians,” in IV Congresso Internazionale di Studi Etiopici (Roma, 10–15 aprile 1972), Vol. 1. Rome. 95–106. Hatke, George. 2011. Africans in Arabia Felix. Aksumite Relations with Ḥimyar in the Sixth Century CE. Ph.D. Diss., Princeton University. ———. 2013. Aksum and Nubia. Warfare, commerce, and political fictions in ancient Northeast Africa. New York. Hegermann, Harald. 1990. “Das hellenistische Judentum,” in Johannes Leipoldt und Walter Grundmann (eds.), Umwelt des Urchristentums, Vol. 1. Darstellung des neu­ testamentlichen Zeitalters (8th ed.). Leipzig. 292–345. Heider, George C. 1985. The Cult of Molek. A Reassessment (JSOT Supp. 43). Sheffield. Henry, René. 1959. Photius, Bibliothèque. Paris. Heyer, Friedrich. 1971. Die Kirche Äthiopiens. Eine Bestandsaufnahme (Theologische Bibliothek Töpelmann 22). Berlin. Hofmann, Inge. 1975. Wege und Möglichkeiten eines indischen Einflusses auf die meroitische Kultur (Studia Instituti Anthropos 23). Bonn. Hoftijzer, J. and Jongeling, K. 1995. Dictionary of the North-West Semitic inscriptions (HdO I/21), Vol. 1–2. Leiden. Huntingford, G. W. B. 1980. The Periplus of the Erythraean sea by an unknown author. With some Extracts from Agatharkhidēs ‘On the Erythraean Sea’. London. Isenberg, C. W. 1840. A Small Vocabulary of the Dankali Language. London. ———. 1841. Dictionary of the Amharic language. London. Jackson, Robert B. 2002. At Empire’s Edge. Exploring Rome’s Egyptian Frontier. New Haven. Jastrow, Marcus. 1886–1903. Dictionary of the Targumim, the Talmud Babli and Yerushalmi, and the Midrashic Literature. New York. Jean, Charles-F. and Jacob Hoftijzer. 1965. Dictionnaire des inscriptions sémitiques de l’Ouest. Leiden.0Jeffery, Arthur. 1938. The Foreign Vocabulary of the Qur’ān. Baroda. Jones, A. H. M. and Elizabeth Monroe. 1935. A History of Abyssinia. Oxford. Juel-Jensen, Bent. 1999. “More Gold Coins out of India of King Ousanas of Aksum,” Spink Numismatic Circular 107 no.6: 176. Kane, Thomas Leiper. 2000. Tigrinya-English Dictionary. Springfield.

Aramaic Loanwords in Gǝ ʿ ǝz

369

Kaplan, Steven. 1982. “Ezana’s Conversion Reconsidered,” Journal of Religion in Africa 13: 101–109 (repr. in Alessandro Bausi [ed.]. 2012. Languages and Cultures of Eastern Christianity: Ethiopian. Farnham–Burlington. 27–34. [no. 2]). ———. 1992. The Beta Israel (Falasha) in Ethiopia. From Earliest Times to the Twentieth Century. New York. ———. 1993. “The Invention of Ethiopian Jews: Three Models,” Cahiers d’études africaines 33: 645–658. Kaufman, Stephen A. 1974. The Akkadian influences on Aramaic (AS 19). Chicago. Kawerau, Peter. 1976. Christlich-arabische Chrestomathie aus historischen Schriftstellern des Mittelalters, Vol. 1, Part 2. Glossar (CSCO 374). Louvain. Kessler, David. 1982. The Falashas. The forgotten Jews of Ethiopia. London. Koehler, Ludwig. 1958. Lexicon in Veteris Testamenti libros (2nd ed.). Leiden. Koehler, Ludwig and Walter Baumgartner. 1967–1995. Hebräisches und aramäisches Lexikon zum Alten Testament. Leiden. Köhler, L., W. Baumgartner, J. J. Stamm, et al. (trans. M. E. J. Richardson). 2001. The Hebrew and Aramaic Lexicon of the Old Testament. Study Edition. Leiden. Krahmalkov, Charles R. 2000. Phoenician-Punic Dictionary (OLA 90). Leuven. Krempel, Veronika. 1973. Die soziale und wirtschaftliche Stellung der Falascha in der christlich-amharischen Gesellschaft von Nordwest-Äthiopien. Ph.D. Diss., Freie Univ. Berlin. Krishnamurthy, Ramasubbu. 1999. “Aksumite Coins of Ethiopia from Karur, Tamil Nadu,” Studies in South Indian Numismatics 8: 58–64. ———. 2000. Non-Roman Ancient Foreign Coins from Karur in India. Chennai. Krumbiegel, Ingo. 1943. Der afrikanische Elefant (Monographien der Wildsäugetiere 9). Leipzig. Lecker, Michael. 1995. “The conversion of Ḥimyar to Judaism and the Jewish Banū Hadl of Medina,” WO 26: 129–136. (reprinted in his Jews and Arabs in Pre- and Early Islamic Arabia. Ashgate. no. XIII) Leslau, Wolf. 1958. Ethiopic and South Arabic Contributions to the Hebrew Lexicon (University of California Publications in Semitic Philology 20). Berkeley – Los Angeles. ———. 1961. “The Names of the Weekdays in Ethiopic,” JSS 6: 62–70. (in idem. 1992. Gurage studies. Collected articles. Wiesbaden. 686–694.) ———. 1963. Etymological Dictionary of Harari (University of California Publications Near Eastern Studies 1). Berkeley – Los Angeles. ———. 1976. Concise Amharic Dictionary. Amharic-English, English-Amharic. Wiesbaden. ———. 1979. Etymological Dictionary of Gurage (Ethiopic), Vol. 1–3. Wiesbaden. ———. 1982. North Ethiopic and Amharic cognates in Tigre. Naples. ———. 1987. Comparative dictionary of Geʿez (Classical Ethiopic). Wiesbaden. ———. 1989. Concise dictionary of Geʿez. Wiesbaden.

370

tubach

Letsios, Dimitrios G. 1991. “Die äthiopisch-himyaritischen Kriege des 6. Jahrhunderts und die Christianisierung Äthiopiens,” Jahrbuch der österreichischen Byzantinistik 41: 25–41. Levy, Jacob. 1924. Wörterbuch über die Talmudim und Midraschim nebst Beiträgen von Heinrich Leberecht Fleischer (2nd ed.). Berlin. Littmann, Enno. 1913. Deutsche Aksum-Expedition herausgegeben von der General­ verwaltung der Königl[ichen] Museen zu Berlin, Vol. 4. Sabaische, griechische und altabessinische Inschriften. Berlin. ———. 1926. “Indien und Abessinien,” in Willibald Kirtel (ed.), Beiträge zur Literaturwissenschaft und Geistesgeschichte Indiens. Festgabe Hermann [Georg] Jacobi zum 75. Geburtstag ‹11. Febr. 1925›. Bonn. 406–417. ———. 1950. “Äthiopische Inschriften,” in Miscellanea Academica Berolinensia. Gesammelte Abhandlungen zur Feier des 250jährigen Bestehens der Deutschen Akademie der Wissenschaften zu Berlin, Vol. II.2. Berlin. 97–127. Littmann, Enno and Maria Höfner. 1956–1962. Wörterbuch der Tigrē-Sprache (Akademie der Wissenschaften und der Literatur. Veröffentlichungen der orientalischen Kommission 11). Wiesbaden. Lundin, Abraham G. 1999. “The Jewish Communities in Yemen in the 4th–6th Centuries (according to epigraphic material),” in Ephraim Isaac and Yosef Tobi (eds.), JudaeoYemenite Studies. Proceedings of the Second International Congress. New Jersey. 17–25. MacKenzie, D. N. 1971. A concise Pahlavi dictionary. London. Mahler, Samuel. 2007. Kebra Negast. La Gloire des Rois d’Éthiopie traduit du guèze. Noisy le Grand. Margolis, Max L. 1910. Lehrbuch der aramäischen Sprache des babylonischen Talmuds (Clavis linguarum semiticarum 3). Munich. Marrassini, Paolo. 1999. “Ancora sul problema degli influssi siriaci in età aksumita,” in Luigi Cagni (ed.), Biblica et Semitica. Studi in memoria di Francesco Vattioni (Istituto Universitario Orientale. Dipartimento di Studi Asiatici. Series Minor 59). Naples. 325–337. (English translation in Alessandro Bausi [ed.]. 2012. Languages and Cultures of Eastern Christianity: Ethiopian. Farnham–Burlington. 209–219. [no. 13]) ———. 1990. “Some Considerations on the Problem of the ‛Syriac Influences’ on Aksumite Ethiopia,” Journal of Ethiopian Studies 23: 35–46. Messing, Simon D. 1982. The Story of the Falashas. “Black Jews” of Ethiopia. Brooklyn. Metlich, Michael Andreas. 2006. “Aksumite Gold Coins and their Relation to the Roman-Indian Trade,” in Federico De Romanis and Sara Sorda (eds.), Dal denarius al dinar. L’Oriente a la Moneta romana (Studi e materiali 12). Rome. 99–103. Mordini, A. 1960. “Gli aurei Kushana del convento di Dabra Dammò. Un’indizio sui rapporti commerciali fra l’India e l’Etiopia nei primi secoli dell’era volgare,” in Atti del convegno internazionale di Studi Etiopici (Roma 2–4 aprile 1959). Rome. 249–254.

Aramaic Loanwords in Gǝ ʿ ǝz

371

———. 1967. “Gold Kushana coins in the Convent of Dabra Dammo,” Journal of the Numismatic Society of India 29.2: 19–25. Morin, Didier. 2004. Dictionnaire historique afar (1288–1982). Paris. ———. 2012. Dictionnaire afar-français (Djibouti, Érythrée, Éthiopie). Paris. Moscati, Sabatino. 1991. Gli adoratori di Moloch. Indagine su un celebre rito cartaginese (Arte, storia, archeologia 4). Milan. Müller, Karl. 1851. Fragmenta Historicorum Graecorum, Vol. 4. Insunt Praxagorae Athenensis, Bemarchii Caesariensis, Eustathii Cappadocis, Eutychiani Cappadocis, Magni Carrheni, Eunapii Sardiani, Olympiodori Thebaei, Prisci Panitae, Malchi Philadelphensis, Capitonis Lycii, Candidi Isauri, Eustathii Epiphaniensis, Hesychii Milesii, Nonnosi, Petri Patricii, Anonymi, Menandri Protectoris, Theophanis Byzantii, Joannis Epiphaniensis, Joannis Antiocheni, aliorum auctorum supra quingentos numero quorum aetas est incerta. Accedunt addenda et indices locupletissimi auctorum titulorum, rerum nominumque. Paris. Munro-Hay, S. C. 1980. “ʿĒzānā (Ēzana/Ēzanas). Some Numismatic Comments,” Azania. Archaeological Research in Africa 15.1: 109–119. ———. 1986. The Munro-Hay Collection of Aksumite Coins. Naples. ———. 1988. “The Dating of Ezana and Frumentius,” Rassegna di Studi Etiopici 32: 111–127. (reprinted in Alessandro Bausi [ed.]. 2012. Languages and Cultures of Eastern Christianity: Ethiopian. Farnham–Burlington. 57–74. [no. 4]) ———. 1999. Catalogue of the Aksumite Coins in the British Museum. London. ———. 1999. “Ethiopia and Alexandria. Chronology of the Conversion of Ethiopia”, in Piotr O. Scholz (ed.), Äthiopien gestern und heute. Akten der 1. Tagung der Orbis Æthiopicus Gesellschaft zur Erhaltung und Förderung der äthiopischen Kultur (Nubica et Æthiopica IV/V). Warszaw. 455–464. (= 67–76) Munro-Hay, S. C. and Bent E. Juel-Jensen. 1984. Aksumite Coinage. London. Munzinger, Werner. 1883. Ostafrikanische Studien (2nd ed.). Basel. Murav’yov, Alexei. 2009. “Nachalo vtoroi volny christianizatsii Aksuma,” Vestnik Drevnei Istorii 181–97. Nawartmal, Hanuman and Lakshmi Nawartmal. 1998. “Spätantikes Handelsgold in Südindien (aus dem Englischen von Andrea Luegmeyer),” Money Trend 30 nr. 11: 52–57. Nawartmal, Hanuman. 1999. “Aksumite Coins in India—Some New Evidence,” Spink Numismatic Circular 1971: 1–2. Nebes, Norbert. 2008. “Die Märtyrer von Nagran und das Ende der Himyar. Zur politischen Geschichte Südarabiens im frühen sechsten Jahrhundert,” Aethiopica 11: 7–40. ———. 2010. “The Martyrs of Najran and the End of the Himyar: On the Political History of South Arabia in the Early Sixth Century,” in Angelika Neuwirth, Nicolai Sinai, Michael Marx (eds.), The Qurʾān in Context. Historical and Literary

372

tubach

Investigations into the Qurʾānic Milieu (Texts and Studies on the Qurʾān 6). Leiden. 27–59. Nöldeke, Theodor. 1910. “Lehnwörter in und aus dem Äthiopischen,” in Neue Beiträge zur semitischen Sprachwissenschaft. Straßburg. 31–66. (reprinted in idem. 1982. Beiträge und Neue Beiträge zur semitischen Sprachwissenschaft. Amsterdam. 31–66) Nyberg, Henrik Samuel. 1974. A manual of Pahlavi, Part 2. Glossary. Wiesbaden. del Olmo Lete, G. and J. Sanmartín. 2003. A Dictionary of the Ugaritic Language in the Alphabetic Tradition. Leiden – Boston. Pankhurst, Richard. 1992. “The Falashas, or Judaic Ethiopians, in Their Christian Ethiopian Setting,” African Affairs 91: 567–582. Payne Smith, J. (Mrs. Margoliouth). 1903. A Compendious Syriac Dictionary. Oxford. Payne Smith, R. 1879–1901. Thesaurus Syriacus. Oxford. Pedroni, Luigi. 1997. “Una Collezione di Monete Aksumite. Catalogo. Analisi microchimiche di Guido Devoto,” Bolletino di Numismatica ser. I, anno 15, no. 28–29: 7–147. Pereira, Francisco Maria Esteves. 1898. O elephante em Ethiopia. Lisbon. Phillipson, David W. 2012. Foundations of an African Civilisation. Aksum & the Northern Horn 1000 BC–AD 1300. Woodbridge. Polotsky, H. J. 1964. “Aramaic, Syriac and Geez,” JSS: 1–10. (reprinted in idem. 1971. Collected Papers. Jerusalem. 1971. 8–17 and in Alessandro Bausi [ed.]. 2012. Languages and Cultures of Eastern Christianity: Ethiopian. Farnham–Burlington. 187–196. [no. 11]) Power, Timothy. 2012. The Red Sea from Byzantium to the Caliphate AD 500–1000. Cairo. Quirin, James. 2010. The Evolution of the Ethiopian Jews. A History of the Beta Israel (Falasha) to 1920. Philadelphia. ———. 1998. “Caste and Class in Historical North-West Ethiopia: The Beta Israel (Falasha) and Kemant, 1300–1900,” Journal of African History 39.2: 195–220. ———. 1979. “The Process of Caste Formation in Ethiopia: a Study of the Beta Israel (Felasha), 1270–1868,” International Journal of African Historical Studies 12.2: 235–258. Raineri, Osvaldo. 2008. La Gloria dei Re. Salomone e la regina di Saba nell’epopea etiopica tra testo e pittura traduzione dal testo etiopico (ge‘ez). Rome. Reinisch, Leo. 1878. “Die Sahosprache,” ZDMG: 415–464. ———. 1890. Die Saho-Sprache, Vol. 2. Wörterbuch der Saho-Sprache. Wien. Robin, Christian. 1989. “La première intervention abyssine en Arabie méridionale (de 200 á 270) de l’ère chrétienne,” in Taddese Beyene (ed.), Proceedings of the Eighth International Conference of Ethiopian Studies, Univesity of Addis Ababa, 1984, Vol. 2. Addis Ababa. 147–162. ———. 2003. “Le judaïsme de Himyar,” Arabia. Revue de Sabéologie 1: 97–172. ———. 2004. “Himyar et Israël,” CRAIBL 148.2: 831–908. ———. 2008. “Joseph, le dernier roi de Ḥimyar (de 522 à 525) ou une des années suivantes,” JSAI 34: 1–124.

Aramaic Loanwords in Gǝ ʿ ǝz

373

Rodinson, Maxime. 2001. “La conversion de l’Éthiopie,” Raydān. Journal of Ancient Yemeni Antiquities and Epigraphy 7: 225–262. Rubin, Zeev. 1995. “ ‘The Martyrdom of Azqir’ and the Struggle between Judaism and Christianity in South Arabia during the 5th Century CE,” in Aryeh Kasher and Aharon Oppenheimer (eds.), Dor Le-dor: From the End of Biblical Times Up to the Redaction of the Talmud. Studies in Honour of Joshua Efron. Tel-Aviv. 251–284. ———. “Judaism and Raḥmanite Monotheism in the Ḥimyarite Kingdom in the Fifth Century,” in Tudor Parfitt (ed.), Israel and Ishmael. Studies in Muslim-Jewish Relations. Richmond – New York. 32–51. Ryckmans, G. 1946. “Les inscriptions monothéistes sabéennes,” in Miscellanea historica in honorem Alberti De Meyer, Universitatis catholicae in oppido Lovaniensi iam annos XXV professoris, Vol. I. Louvain. 194–205. ———. 1958. “Heaven and Earth in South Arabian Inscriptions,” JSS 3: 225–236. Schneider, Pierre. 2004. L’Éthiopie et l’Inde. Interférences et confusions aux extrémités du monde antique (VIIIe siècle avant J.-C.–VIe siècle après J.-C.) (Collection de l’École Française de Rome 335). Rome. Schwarz, Wilhelm. 1894. “Aethiopien,” Rheinisches Museum für Philologie NF 49: 353–361. Scholz, Peter Otto. 1988. Frühchristliche Spuren im Lande des ἀνὴρ Αἰθίοψ. Historischarchäologische Betrachtungen zur Apostelgeschichte 8: 26–40. Ph.D. Diss., Bonn. Seland, Eivind Heldaas. 2010. Ports and political power in the Periplus. Complex Societies and Maritime Trade on the Indian Ocean in the First Century AD (BAR International series 2102/ Society for Arabian Studies Monographs 9). Oxford. Sergew, Hable Sellassie. 1972. Ancient and Medieval Ethiopian History to 1270. Addis Ababa. Shelemay, Kay Kaufman. 1986. Music, Ritual, and Falasha History (Ethiopian series monograph 17). East Lansing. Shooshtary, Dakhil. 2012. Mandaic Dictionary: English Mandaic. Bloomington. Sima, Alexander. 2005. “GDR(T),” in S. Uhlig (ed.), Encyclopaedia Aethiopica, Vol. 2. Wiesbaden. 719. Sims-Williams, Nicholas and Desmond Durkin-Meisterernst. 2012. Dictionary of Manichaean Texts, Vol. 3. Texts from Central Asia and China, Part 2. Dictionary of Manichaean Sogdian and Bactrian (Corpus Fontium Manichaeorum. Subsidia). Brepols. Soden, W. von 1965–81. Akkadisches Handwörterbuch, Vol. I–III. Wiesbaden. Sokoloff, M. 2009. A Syriac Lexicon. A Translation from the Latin, Correction, Expansion, and Update of C. Brockelmann’s Lexicon Syriacum. Winona Lake – Piscataway. Stein, Peter. 2010. “Himyar und der Eine Gott: Südarabien in den letzten zwei Jahrhunderten vor dem Islam. Besprechungsartikel zu I. Gajda, Le royaume de Himyar à l’époque monothéiste: L’histoire de l’Arabie du Sud ancienne de la fin du

374

tubach

IVe siècle de l’ère chrétienne jusqu’à l’avènement de l’Islam, Paris 2009,” Orientalia 79: 558–566. Strelcyn, Stefan. 1956. Kebra Nagast czyli Chwala królów Abisynii. Fragmenty. Z języka abisyńskiego przelożyl oraz wstępem zaopatrzyl. Warsaw. Teferi, Amaleletch. 2005. “About the Jewish Identity of the Beta Israel,” in Tudor Parfitt and Emanuela Trevisan Semi (eds.), The birth of an elite. London. 173–192. Till, Walter C. 1961. Koptische Grammatik (Saïdischer Dialekt) mit Bibliographie, Lesestücken und Wörterverzeichnissen (2nd ed.) (Lehrbücher für das Studium der orientalischen und afrikanischen Sprachen 1). Leipzig. Tropper, Josef. 2008. Kleines Wörterbuch des Ugaritischen (ELO 4). Wiesbaden. Ullendorff, E. 1956. “Hebraic-Jewish Elements in Abyssinian (monophysite) Christianity,” JSS 1: 216–256. (reprinted in idem. 1987. Studia Aethiopica et Semitica [ÄF 24]. Wiesbaden. 2–42) Vergari, Moreno and Roberta Vergari. 2003. A Basic Saho-English-Italian Dictionary, with an English-Saho Index and Grammatical Notes. Asmara. Voigt, R. 1991. “Greek loan-words in Gə‛əz (Classical Ethiopic): The Role of Arabic,” Graeco-Arabica 4: 65–272. Wahrmund, Adolf. 1898. Handwörterbuch der neu-arabischen und deutschen Sprache (3rd ed.). Giessen. Wehr, Hans. 1985. Arabisches Wörterbuch für die Schriftsprache der Gegenwart. ArabischDeutsch (5th ed.). Wiesbaden. Weninger, Stefan. 2005. “Der Wortschatz des klassischen Äthiopisch,” in Bogdan Burtea, Joseph Tropper, and Helen Younansardaroud (eds.), Studia Semitica et Semitohamitica. Festschrift für Rainer Voigt anläßlich seines 60. Geburtstages am 17. Januar 2004 (AOAT 317). Münster. 465–488. West, Vincent. 1999. “Ge‛ez Legends on Aksumite Coins,” Oriental Numismatic Society Newsletter 159: 5–6. ———. 2006. “Recent Research on Aksumite Coinage,” London Numismatic Club. Newsletter. The Journal of the London Numismatic Club 8.9: 15–24. Westendorf, Wolfhart. 1965–1977. Koptisches Handwörterbuch. Heidelberg. Witakowski, Witold. 1989–1990. “Syrian Influences in Ethiopic Culture,” Orientalia Suecana 38/39: 191–202. (repr. in Alessandro Bausi [ed.] 2012. Languages and Cultures of Eastern Christianity: Ethiopian. Farnham–Burlington. 197–208. [no. 12]) Yule, Paul. 2013. Late antique Arabia. Ẓafār, Capital of Ḥimyar. Rehabilitation of a “Decadent” Society. Excavations of the Ruprecht-Karls-Universität Heidelberg 1998– 2010 in the Highland of the Yemen (Abhandlungen der Deutschen Orient-Gesellschaft 29). Wiesbaden. ———. 2014. Himyar: Die Spätantike im Jemen / Late Antique Yemen (2nd ed.). Aichwald. Ziegert, Helmut and Marlies Wendowski. 2003. “Aksum at the transition to Christianity,” Annales d’Ethiopie 19: 215–230.

Language Contact between Akkadian and Northwest Semitic Languages in Syria-Palestine in the Late Bronze Age Juan-Pablo Vita 1 Introduction Akkadian, a native language of Mesopotamia, “is by far the best attested Semitic language in the ancient world prior to the appearance of Arabic in the historical record” (Kouwenberg 2011: 330).1 The oldest traces of Akkadian can be dated to the middle of the third millennium BC, and the last text is datable to the first century AD (Kouwenberg 2011: 330–332). During the second millennium BC this language was used as a lingua franca in the ancient Near East by a large number of States, reaching its peak as the language of diplomacy in the Middle Babylonian period (ca. 1500–1000).2 As Akkadian expanded beyond the geographical area where it was spoken, it came into contact with both Semitic (e.g., Amorite, Ugaritic, Canaanite) and non-Semitic languages (e.g., Hittite, Hurrian). These languages were often spoken by societies who had not previously developed a script system for their native languages. Given the new situation, extra-Mesopotamian scribes adopted at times Mesopotamian cuneiform script to write texts only in Akkadian. On other occasions, they used the Mesopotamian script to write texts both in Akkadian and in the local language. In both cases Akkadian became imbued, to various degrees, with the diverse local languages it came 1  This chapter is one result of the research project ‘Native Languages, linguae francae, and Graphics Traditions in Late Bronze Age Syria and Palestine: Three Case Studies (Canaan, Ugarit, Emar)’ (FFI2011-25065), funded by the Spanish Ministry for Economic Affairs and Competitiveness within the National Plan for Scientific Research, Development and Technological Innovation (I+D+I). A previous version of this work was presented on 17 July 2014 as a lecture at the Seminar für Sprachen und Kulturen des Vorderen Orients—Assyriologie of the University of Heidelberg titled “Akkadisch und einheimische semitische Sprachen im spätbronzezeitlichen Syrien-Palästina: Sprachkontakt und Schreiberschulen.” I wish to thank Professor Dr. Stefan Maul for kindly inviting me to give this lecture and to also thank those who attended for their valuable observations and remarks. 2  van Soldt 2011: 405; Streck 2011: 376. See also Márquez Rowe 2006: 140–166. In some cases scribes used Mesopotamian cuneiform script and Akkadian both for international exchanges and for the purposes of local administration.

© koninklijke brill nv, leiden, ���5 | doi ��.��63/9789004300156_020

376

Vita

into direct contact with.3 A significant result of this process was that, quite often, Akkadian acted as a sort of “contrast medium,” bringing to light languages (Semitic and non-Semitic, such as Amorite and Kassite respectively) which had not been recorded in writing up until that moment and which would otherwise have probably gone totally unnoticed. Within the aforementioned framework, this chapter shall exclusively focus on the contact between Akkadian and three other Semitic languages. This contact therefore involves genetically related languages, a linguistic situation with peculiar issues.4 The three languages in question are Emarite, Ugaritic, and Canaanite, attested in Syria-Palestine in the Late Bronze Age. Unlike Akkadian, a language within the East Semitic branch of Semitic languages, the aforementioned three languages are part of the Northwest Semitic sub-family, corresponding to clearly differentiated geographical, social, and political environments.5 No texts are preserved in Emarite or in Canaanite from this time period; our knowledge of these languages is limited to the linguistic elements preserved in Akkadian texts written in the kingdom of Emar and in various localities in Canaan. These languages can be considered “Trümmersprachen” or partially documented languages, whose knowledge is consequently partial and defective.6 Ugaritic, in turn, is attested to not only in Akkadian texts but above all in hundreds of texts written entirely in that language using a specific alphabet. In the following lines I will describe firstly the type of linguistic contact established between Akkadian and Emarite (§2), Akkadian and Ugaritic (§3), and Akkadian and Canaanite (§4). I will next try to elucidate what the attitude of scribes was towards the registering in writing their mother tongue and how they perceived their own language in connection with the Akkadian used in written texts, a goal of a rather sociolinguistic nature. I will try to reach, in conclusion, an understanding of what kind of relationship existed between these three vernacular languages and Akkadian as a lingua franca. 3  van Soldt 2011; Streck 2011: 366–367, 376–378. From this viewpoint, Akkadian has also been the subject of study in linguistic analyses focused on contact languages; see for instance Thomason 2001: 6, 89; Márquez Rowe 2004; Johanson 2013. 4  As expounded by Epps, Huehnergard, and Pat-El 2013. 5   Regarding the Northwest Semitic languages family, see especially Huehnergard 2005; Hasselbach and Huehnergard 2007; Gzella 2011. 6  A category where Amorite, for instance, could also be included. Regarding the problems posed by this type of partially documented languages see Hackett 2002. See also Gzella 2011: 426–428.

Language Contact between Akkadian and NwS Languages

2

377

Akkadian and Emarite

2.1 Texts Found in Emar The archaeological excavations carried out since 1972 in Meskene Qadime (Arab Republic of Syria), ancient Emar, have brought to light about 1,200 texts and fragments written mainly in Akkadian.7 The Akkadian language used in these documents has been appropriately studied by Seminara (1988)8 while the onomastics was studied by Pruzsinszky (2003). These texts can be dated to the 14th and 12th centuries BC, that is, to the Syrian Late Bronze Age, and they enable us to study the political and social organization as well as the languages used in a Syrian kingdom already referred to by the Eblaite sources from the 3rd millennium BC.9 The Akkadian texts are of different genres (legal and clerical documents, literary, religious, and lexical texts), and they have an added ingredient of interest: they also attest to the local language of Emar, currently known as “Emarite.” 2.2 Emarite The classification of Emarite as a Semitic language has been the subject of some debate, but in general it may be classified as Northwest Semitic (NWS).10 The main difficulty in studying this language is posed by the fact that no text in Emarite has ever been preserved: all elements of Emarite are found inserted, here and there, within syllabic texts in Akkadian written by local scribes. For example:

7   A small number of documents were also in Hittite (published) and in Hurrian (unpublished; cf. Giorgieri 2013: 166). 8   See also Faist and Vita 2008. 9   A full and updated bibliography on all the aspects of the studies of Emar can be found in Faist, Justel, and Vita 2003, 2005, 2007, 2009. See also “Bibliografía de los estudios de Emar, 2006. Página en español, inglés y alemán,” at http://www.propylaeum.de/index. php?id=655. To date, the archive of Emar has been the subject of two specific collective works: Chavalas 1996 and d’Alfonso, Cohen, and Sürenhagen 2008. Regarding the chronology of the Emar archives, see d’Alfonso and Cohen 2008; Pruzsinszky 2009; Cohen 2013. 10  Pruzsinszky 2003: 41–47, provides a brief history of the investigation. Linguistic links between Emarite and Ugaritic have been explored by Watson 1992; 1993. More recently, del Olmo 2012 has analysed the links between Emarite, Ugaritic, and Old South Arabian.

378

Vita

– lú.meš gal lú.meškà-ma-ri ì.kú ‘The nobles (and) priests eat’.11 – dInanna ṣu-pá-ra-ti ‘Ishtar of the goats’.12 – dkaskal.kur.ra.meš ša ḫi-iṭ-ṭi ‘(god) Balih of the wheat’.13 The Emarite linguistic material was compiled and studied by Pentiuc (2001). His work serves as a reference for any analysis of Emarite, though it must currently be complemented with and amended by other studies, especially those by Sjöberg (1998), Ikeda (2003),14 and Cohen (2002; 2009; 2010a; 2010b). A glossary comprised exclusively of NWS linguistic elements found in the syllabic texts of Emar would yield approximately 145 lemmata which, according to Ikeda (2003, 268), may be classified within the following semantic fields:15 1) topographic features,16 2) architectural structures, 3) professional and social status, 4) divine names and epithets, 5) festival and month names, 6) agricultural products, 7) animals, 8) food, 9) vessels and tools, 10) units of measurement. Nevertheless, not all of these items can be directly considered part of the Emarite language, as Ikeda (2003, 268) reckons that only half of these items “may safely be regarded as Emarite bases”; the remainder could be classified within the following four main groups: 1) West Semitic bases which are not necessarily Emarite, 2) West Semitic loan-words into Akkadian, 3) Akkadianized Emarite roots, and 4) Akkadian lexemes with Emarite interference (Ikeda 2003, 268–269). 2.3 Emarite: Genres, Chronology, and Statistic Data According to Pentiuc (2001: 12), “The largest concentration of non-normative Akkadian forms (native and foreign) is found in religious, legal and economic documents. In second place come the lexical and literary texts, and letters provide only one possible example.” The following lines—exclusively focused on the Northwest Semitic elements—aim to qualify and explain Pentiuc’s statements also taking into consideration the diachronic dimension of the various text genres.

11  Emar 446: 38. Pentiuc 2001: 95: “/kamarū/ and /kumarū/ WS n.m.pl. ‘priests’ (KMR).” 12  Emar 379: 6. Pentiuc 2001: 164: “/ṣuparātu/ WS n.f.pl. ‘goats (ṢPR).” 13  Emar 378: 43’. Pentiuc 2001: 70: “/ḥiṭṭu/ WS n.m.s. ‘wheat (ḤNṬ).” 14  Ikeda 2003 provides a very good assessment of the work of Pentiuc 2001 and also takes into consideration texts which are not reviewed by Pentiuc. 15  Pentiuc’s glossary (2001, 19–198) also compiles Hittite and Hurrian terms. 16  See also Reculeau 2010.

Language Contact between Akkadian and NwS Languages

379

Approximately 700 texts found in Emar have been published to date. Two thirds of these consist of legal documents regarding private law.17 Only 133 of these texts contain a NWS term. Legal texts from Emar correspond, in turn, to two different scribal traditions, known as ‘Syrian’ and ‘Syro-Hittite’, each with their own palaeographic and grammatical characteristics.18 The ratio between both traditions is approximately one to one.19 In terms of absolute chronology, the documents of Syrian tradition span from the early 14th century to the mid 13th century, while the Syro-Hittite tradition starts ca. 1275 and ends with the fall of the city around 1175 BC.20 Of the 133 legal texts containing NWS terms 95 are of Syrian tradition, and the remainder are of Syro-Hittite tradition. The corpus of clerical texts from Emar consists of approximately 175 documents, chronologically corresponding to the final stages of the history of the city.21 These documents also reflect the aforementioned Syrian and SyroHittite traditions.22 Only 33 clerical texts contain NWS linguistic items, 25 of which are of Syro-Hittite tradition. As for other genres in the text corpus of Emar, the situation can be summed up as follows: only 8 out of 70 lexical texts, 61 of ca. 320 literary and ritual texts,23 and just one out of ca. 20 preserved letters contain NWS linguistic elements. The only preserved medical text contains just one NWS term. The following conclusions can be drawn from the aforementioned data: 1) the number of texts from Emar containing NWS linguistic elements is quite low both from the point of view of the entire text corpus from this locality and from a genre standpoint; 2) as legal and clerical texts reveal, both Syrian and Syro-Hittite traditions attest to NWS linguistic elements, neither of which reveal a higher prevalence than the other of these items if both genres are taken into consideration; 3) the textual corpus from Emar attests to NWS linguistic elements throughout its entire chronology, spanning from the early 14th century to ca. 1175 BC. 17  Cf. Démare-Lafont 2010: 43. Westbrook (2003: 657) reckons that Emar and sites in the vicinity have yielded approximately 500 legal documents. 18  Regarding these two scribal traditions, see the relevant clarifications produced by Fleming and Démare-Lafont 2009, with preceding bibliography. See also recently Cohen 2012. 19  Cf. Ikeda 1999: 166; Pentiuc 2001: 10. 20  Cohen and d’Alfonso 2008: 24; Cohen 2013: 289. 21  Faist 2008: 201: “roughly between the end of the 13th and the beginning of the 12th century BC.” 22  Faist 2008. 23  Approximately 20 literary and 300 ritual texts are preserved. The following observation by Ikeda (2003: 273) should also be taken into consideration: “The religious texts were composed in Emar, while the literary texts were not.”

380

Vita

In short, little more than half (379) of the 700 published texts of the Emar corpus contain a NWS linguistic element. The aforementioned figure of 145 NWS lemmata spans over two centuries of history. 2.4 The Attitude of Emarite Scribes towards their Mother Tongue Bearing all these data in mind, what may have been the attitude of Emarite scribes towards their own mother tongue as a written language? An illustrative example in this respect is provided by the lexical lists known as Urra = ḫubullu, lú = ša, and izi = išātu found in Emar. These texts produce a column for the Sumerian entries, a second column in Akkadian (with the occasional insertion of terms in the local language or simply non-Akkadian terms) as a translation of the terms listed in the first column, and a third column with the occasional presence of terms which could be considered as “second glosses.” Here are some examples where the secondary glosses may be interpreted as WS terms, quite probably belonging to the local language: – (giš) apin á-kár – (giš) sag-du – az

u-nu-tu ‘utensil’ a-sú-ú ‘loom part’ a-su ‘bear’

: a-nu ‘utensil’24 : na-ba-lu ‘harp/hanging’25 : za-ba-ú ‘animal-type’26

The “second glosses” have been studied in detail by Cohen (2010a), who reaches the following conclusions regarding their function: “Because of the high percentage of Akkadian words, it is clear that the inclusion of the second glosses was not primarily intended to supply translations of Akkadian words into the local language. Rather the intention was to supply synonyms, related terms, or explanations to the first gloss, without a specific regard of the target language . . . The purpose of the second gloss . . . was not merely to translate, but rather to amass additional data to the lexical lists” (Cohen 2010a: 834–835).27 From what we have seen so far it may be deduced that, generation after generation, Emarite scribes were not particularly keen on recording in writing their own mother tongue.28 The presence of NWS elements in general and of Emarite elements in particular in Akkadian texts from Emar may be 24  Emar 545 D: 69’, 136’. Cohen 2010a: 814. 25  Emar 545 H: 72, 243’. Cohen 2010a: 824. 26  Emar 551 A: 115, 37’. Cohen 2010a: 831. 27  See also Cohen 2010b: 29 and 40. 28  Cf. also Cohen 2009: 242–243: “it is worth asking why the Emar lexical lists were not supplied with a column of the local language as in other peripheral sites, like Ugarit and Ḫattuša.”

Language Contact between Akkadian and NwS Languages

381

considered sporadic and limited to the semantic fields pointed out by Ikeda (cf. §2.2), which were actually the only cases where the use of terms written in the local language were required. As a general outcome of this attitude of scribes, the Emar Akkadian texts contain hardly half a dozen verbal forms reflecting West Semitic Patterns;29 significantly enough, none of these verbal forms are to be found in letters. As Cohen sums up, in Emar the local language “was never committed to writing, reflecting the situation in other scribal centres along the Middle to Upper Euphrates valley during the Old Babylonian and Middle Babylonian periods where the local language or languages were never written” (Cohen 2009: 14). Overall, an effective sort of barrier seems to be in place in Emar blocking the permeation of linguistic elements from the local language into written Akkadian, save for the minimum requisite items. What could this barrier consist of? Cohen has already pondered this matter, and he points out several possible reasons: it is worth asking . . . why Emar was resistant to writing its religious or other literature in its own tongue. We can only speculate for now that the city’s geographical position in the Middle Euphrates region contributed to a low language barrier between the local spoken language, Emarite, and the written Akkadian, just as it did in Old Babylonian Mari, or in other contemporary sites, such as Ekalte or Tell Hadidi, unlike the high (or higher) language barrier between Akkadian and the ‘vernaculars’ of the Lebanon coast, Canaan, and Anatolia. Which may have prompted the writing of one’s own language in cuneiform or in indigenous scripts (as Hittite Hieroglyphic or the alphabet). This factor, combined with a certain degree of conservatism, may have contributed to the lack of translations of the scholarly materials into the local tongue and the absence of literature written in Emarite. (Cohen 2009: 242–243) Alongside these general possible causes, I believe that the matter ought to be focused, from a more specific and explicit standpoint, on Emarite scribes who were the actual authors of the texts we have recovered. In this sense, I believe that the conclusion reached by Ikeda (2010) in his study on diglossia in Emar ought to be taken into consideration:30 “Akkadian was not only written but also

29  Cf. Pentiuc 2001: 247–248. 30  On this matter, see also previously Ikeda 2003: 276.

382

Vita

spoken in Emar . . . at least in the royal scribal circle.”31 This fact could indeed have acted as an efficient barrier against the inclusion of Emarite linguistic elements in the Akkadian texts from Emar, despite the clear tension between both languages which can be occasionally noted in some Akkadian texts.32 But it may not have been the only one, as pointed out later sub §3.4. 3

Akkadian and Ugaritic

3.1 The Kingdom of Ugarit In terms of geographic area, political power, and military capacity, Ugarit (also located in the modern Syrian Arab Republic) was an average sized kingdom by the standards of Syria in the second millennium BC, yet a leading one in economical terms. It vanished (like Emar) from history in the early twelfth century BC, swept away by the wave of the so-called “Sea Peoples,” a time which witnessed the collapse of the main Late Bronze Age Syrian centers and the disappearance of the kingdom of Ḫatti.33 The kingdom of Ugarit was contemporary with Emar, with whom direct and close trading relations were maintained.34 It is the only kingdom in Late Bronze Age Syria to have yielded abundant literature written in the local language (Ugaritic). 3.2 Languages and Scripts The language and text situation in Ugarit was very different from that of Emar. Ugarit has yielded instances of various scripts (such as Egyptian hieroglyphs or Cypro-Minoan script)35 and languages (such as Hurrian or Hittite),36 but mostly texts in Ugaritic (the local language) and in Akkadian (non-local language).37 It is probable that, as in the case of Emar (§2.4), Akkadian may

31  Previously, Seminara 1998: 56 concluded that Akkadian in Emar was born and developed as a written language, deprived of substantial contributions from the actual use of the language (an opinion recorded by Pentiuc 2001: 12). Later, however, Seminara (1998: 601) did admit the possibility that Akkadian was spoken in Emar, at least in certain circles and under certain circumstances. 32  See, for instance, the case of text E 17; cf. Seminara 1998: 604–606; Cohen 2009: 86. 33  On the history of Ugarit, see Singer 1999 (= 2011: 19–146); Freu 2006. 34  Cf. Cohen and Singer 2006; Malbran-Labat and Roche 2007. 35  For the former, see Lagarce-Othman 2013 (with previous bibliography). For the latter, see most recently Ferrara 2012: 132–145; 2013: 108–113; Egetmeyer 2013. 36  Dietrich and Mayer 1999; Vita 2009; 2013; Giorgieri 2013. 37  See for example van Soldt 1995a: 206.

Language Contact between Akkadian and NwS Languages

383

have been spoken in Ugarit to a certain extent in the circles of Ugarit scribes.38 Akkadian texts found in Ras Shamra-Ugarit comprise a wide range of genres (school, literary, religious texts, letters, legal, clerical texts, etc.) using, as could be expected, the cuneiform logo-syllabic script of Mesopotamian origin.39 They span from the 14th century BC to the destruction of the kingdom in the early 12th century BC. The texts in Ugaritic were written mostly using a cuneiform alphabet of 30 signs. The moment when this alphabet might have been created continues to be debated though it may be agreed that most of the preserved texts from Ugarit are dated to the 13th and early 12th centuries BC.40 The most recent edition of Ugaritic texts (Dietrich, Loretz, and Sanmartín 2013) includes around 1,570 texts and fragments of various genres (myths, rituals, letters, legal, and clerical documents, etc.). Ugaritic is, therefore, a well-attested Semitic language with its own peculiar script.41 3.3 Interactions between Akkadian and Ugaritic Mesopotamian script tradition had a direct impact on several aspects of the alphabetical Ugaritic texts. Scribes from Ugarit used preferably clay tablets to write alphabetic texts in Ugaritic.42 As a result, alphabet signs acquired a cuneiform appearance. As in Akkadian, texts ran from left to right.43 These traits derive from a centuries-old tradition of use of the Mesopotamian cuneiform script and of the Akkadian language in Ugarit. Huehnergard concludes about the Akkadian of Ugarit that, “While the underlying matrix of the grammar is essentially Middle Babylonian, there are relic Old Babylonian features, Assyrian features, and many features that do not conform to the normative grammar of any Mesopotamian dialect” (1989: 271). Van Soldt adds that “the Akkadian at Ugarit was basically Babylonian, but with a number of influences from other languages: Hurrian, Assyrian and Ugaritic” (1999: 42). These linguistic influences would be directly linked to the political situation at each 38  On this matter, see van Soldt 1995: 205 and 206; 1995b: 186; Márquez Rowe 2006: 149. 39  These documents were partly produced in the kingdom of Ugarit itself and partly in other kingdoms of contemporary Syria such as Ḫatti, Cyprus, or Egypt, or in coastal Syro-Palestine towns. On the Akkadian of Ugarit, see Huehnergard 1989; van Soldt 1991; Malbran-Labat 1995. 40  See discussion in Pardee 2012a: 11–12, 45–46. 41  See Bordreuil and Pardee 2009; Huehnergard 2012; Tropper 2012. On the distribution of Akkadian and Ugaritic in the text corpus of Ugarit, see Malbran-Labat 1996 (esp. 56–61); 1999. 42  For Ugaritic texts outside clay tablets, see Zamora 2006; 2007. 43  Except for a few instances of texts written with the so-called “short alphabet”; see Dietrich and Loretz 1988: 145–275.

384

Vita

historical moment. Thus, Hurrian influence was particularly strong in the first period of the archives of Ugarit, when Mittani was a great power. After the defeat of Mittani, the rise of Assyria brought about a higher influence both of Assyrian and Babylonian on the Akkadian at Ugarit (van Soldt 1991: 522– 523; 1999: 45). As could only be expected, Akkadian also left an imprint in the Ugaritic lexicon. According to Watson’s analysis, “Of the approximately 2325 words in the Ugaritic lexicon about 175 have been borrowed from Akkadian (directly or indirectly), which is roughly 7%,” hence, generally speaking, “only a small number of words were borrowed from Akkadian whereas in the hippiatric texts and the birth omens the percentage is considerably higher” (Watson 2007: 117).44 Conversely, what influence did Ugaritic have on the Akkadian at Ugarit? From a chronological viewpoint, “Ugaritic influence, already strong at the beginning of the historic period, becomes stronger as time passes . . . In the older texts the Ugaritic influence is mainly noticeable in syntax, but gradually it also becomes stronger in morphology” (van Soldt 1999: 45 and 44). Huehnergard (1989: 280–281) summarizes as follows the traits of Akkadian in Ugarit which may have a Ugaritic foundation: a) the syntax of expressions involving measures, b) the frequency of bound forms for genitive relationships, c) atypical theme-vowels in some G verbs and a number of unique or atypical D forms, d) some of the instances of ištu for expected ina, e) the frequent use of non-coordinating -ma and some of the unexplained instances of enclitic -mi, f) a number of cases of incorrect agreement may reflect the gender of an underlying Ugaritic noun, and g) some examples of the predicative verbal adjective and a few transitive paris forms probably reflect the Ugaritic suffix-conjugation. As in Ugaritic texts (cf. Tropper 2012: 447–452), syllabic texts also reflect the application of the Barth-Ginsberg law in verbal prefixes (van Soldt 1991: 431). In terms of syntax, the word order in clauses “is much freer than in native Akkadian and the verb is seldom placed at the end” (van Soldt 1991: 518). Instances where syntax follows the order subject – verb – object – indirect object (instead of the habitual Babylonian order subject – object – indirect object – verb) may generally be explained by the direct influence of Ugaritic.45 As for the lexicon, what impact did Ugaritic have on the Akkadian texts? The glossary of Ugaritic words in syllabic texts completed by Huehnergard

44  See, in more detail, Watson 2007: 65–118. 45  Cf. van Soldt 1999: 44. See also in detail Huehnergard 1989: 217–224; van Soldt 1991: 475– 490, 518. See also, more briefly, Streck 2011: 377.

385

Language Contact between Akkadian and NwS Languages

(2008: 103–189) produces approximately 290 entries.46 Of this amount, 130 originate exclusively in the group of lexical texts known as “polyglot Syllable Vocabulary A” (Sª Voc.).47 For example:48 Sign

Akkadian

Hurrian

NU KAR NU

a-bu šu-zu-bu la-a

at-ta-ni a-bu-uš-ku-me ma-nu-ku

Ugaritic

a-da-nu pu-la-ṭu la-a

/ʾadānu/ /pullaṭu/ /lā/

‘father’ ‘to save’ ‘not’

About 160 entries come from non-lexical texts throughout a period of little more than a century.49 In general, these lexemes are found either “(a) as parts of the names of local geographical features or plots of land or, much more often, (b) by chance, essentially lapsus calami in which the scribe either forgot the appropriate Akkadian word and substituted a Ugaritic form or thought that the (Ugaritic) form he was writing was proper Akkadian” (Huehnergard

46  This figure includes the entries which must be eliminated and included in Huehnergard’s glossary according to the “Appendix” in Huehnergard 2008: 189–194 (cf. p. 404 and https:// www.eisenbrauns.com/assets/errata/Huehnergard-AddendaCorrigenda.pdf). See also Huehnergard 1999: 134: “Embebbed within the syllabic cuneiform texts written by scribes at Ugarit are over three hundred Ugaritic lexical items. These Ugaritic forms appear in all genres of Akkadian texts.” 47  “The Sª Voc. was a Mesopotamian lexical series in which columns of individual cuneiform signs, in a fixed order, were equated with one or more Akkadian words in a second column . . . The Ugarit exemplars . . . are unusual in that they . . . add either one additional column giving lexical equivalents in Hurrian or, more often, two additional columns with equivalents in both Hurrian and Ugaritic . . . Nearly all parts of speech are attested among these forms” (Huehnergard 1999: 134). Regarding this, see the analysis carried out by Huehnergard 2008. Cf. also van Soldt 2008; Veldhuis 2014: 298–299. 48  Cf. Huehnergard 1999: 135. 49  See also detailed statistics by Huehnergard 2008: 12–13. It may be noted that, from a merely quantitative viewpoint, the figure of approximately 160 Ugaritic words in Akkadian texts is not too far-off the 145 Northwest Semitic words preserved in the Akkadian texts from Emar (cf. §§2.2 and 2.3). It should also be noted that scholars distinguish approximately 90 Amorite entities which are loanwords into Akkadian and Sumerian (Streck 2000: 82–128, 135; 2010: 39; 2011: 453; see also Knudsen 2004).

386

Vita

1999: 134).50 The impact of Ugaritic is particularly noticeable in daily life texts, especially in legal and economic ones. In legal texts, the local linguistic background is remarkably noticeable in the use of specialized vocabulary— fed by the legal Syro-Babylonian tradition and by local technical terms— and in the verbal morphology, using mixed verbal forms.51 The influence of Ugaritic is even stronger, at all levels, in syllabic clerical texts,52 to the extent that it may be claimed that in most of these documents and despite the script used, the language transcribed was not Akkadian but Ugaritic.53 3.4 The Attitude of Ugaritic Scribes towards their Mother Tongue From the previous section it may be concluded that Ugaritic had a noticeable impact on Akkadian texts written in Ugarit, particularly in certain genres, a circumstance which, to a certain extent, may have been related to the level of training of the scribes.54 In the daily performance of their job, Ugaritic scribes demonstrated that it was technically possible to write in their mother tongue using the logo-syllabic cuneiform script. Yet, the Ugaritic column of the Sª Voc. (§3.3) and, to a certain extent, the syllabic administrative texts, reveal that furthermore (and conversely to what probably happened in Emar) scribes did this being fully aware and deliberate in their action.55 Consequently, scribes 50  “About a fourth of the Akkadian texts contain one or more such Ugaritic words . . . Most of the syllabically-written Ugaritic words are also attested in alphabetic texts. Over one-fifth of the forms, however, are thus far unknown in alphabetic form” (Huehnergard 1999: 136, 138). Huehnergard (2008: 12) reaches his conclusions on the basis of a total of 565 syllabic texts. This figure, however, will soon be notably increased by the edition of the syllabic texts found in the “House of Urtēnu”; see, regarding this, the catalogue made by MalbranLabat 2008. 51  On this matter, see detailed analysis by Malbran-Labat 1999: 78–87. See also Márquez Rowe 2006: 167–174. 52  Approximately a total of 144 texts, according to calculation by Roche 2010: 119. 53  See Malbran-Labat 1999: 96: “l’examen du lexique et des formes verbales semble conduire à la conclusion que la langue des documents administratifs [syllabiques] n’était pas l’akkadien, mais un jargon dans lequel il est difficile de voir un dialecte akkadien tant il est imprégné d’ougaritique . . . Ainsi l’opposition n’est-elle pas entre ‘textes ougaritiques alphabétiques’ et ‘textes akkadiens syllabiques’ mais entre ‘textes en cunéiformes alphabétiques’ et ‘textes en cunéiformes suméro-akkadiens’: plus qu’un bilinguisme, c’est un digraphisme qui était pratiqué par l’administration ougaritaine.” Roche has delved into this idea and concluded that “if a small number of the logo-syllabic administrative texts were written and intended to be ‘read’ in Akkadian, the majority (perhaps even a large majority) appear to have been ‘read’ in the local language, Ugaritic” (Roche 2010: 119). See also van Soldt 1995: 211–212; 2010: 351–355; Márquez Rowe 2006: 152–154. 54  See in this respect van Soldt 1995a: 206; Márquez Rowe 2006: 153. 55  Thus already also Huehnergard 1999: 134, for the Ugaritic column of Sª Voc.

Language Contact between Akkadian and NwS Languages

387

might have opted to go a step beyond, that is, to write full texts in Ugaritic using the logo-syllabic script of Mesopotamian origin. Nonetheless, save for one sole exception which may have been mere experimentation,56 they did not do so. Ugaritic scribes did not have the interest or the will to exploit that option.57 Why did scribes not take a step which was feasible and might seem (from our perspective) even reasonable? One initial reason may be similar to that pointed out for the case of Emarite (§2.4), that is, that Akkadian in Ugarit was a language spoken to a certain extent in scribal circles (cf. §3.3). One second possible reason, which would support the first, could be the rigorous Mesopotamian school tradition in Ugarit.58 That tradition may have made Ugaritic scribes conceive that Mesopotamian cuneiform script was not appropriate, merely from a conceptual viewpoint, to transcribe the local Semitic language.59 This belief was probably supported by the possibility of channelling the local language through the cuneiform alphabet. It was a script of a much shorter life and extent, free from the burden of a long and prestigious foreign scribal tradition. Consequently, it was conceptually apt to transcribe the local language while it might also act as an instrument which could be used for experimenting with other languages,60 as well as for its joint use, in one single text, alongside logo-syllabic script.61 56  It consists of text RS 20.163 (= KTU 10.1), regarding which see Huehnergard 1987: 11–12, and KTU 739. 57  Therefore, should the alphabetic script not have caught on in Ugarit, our current knowledge of Ugaritic vocabulary and grammar would be as fragmented as it is for Emarite. 58  On the weight of Mesopotamian tradition on the Ugarit scribal teaching system as well as on the presence in Ugarit of Babylonian and Assyrian scribes, see Malbran-Labat 1999: 76, 100; van Soldt 1995b; 2001; 2002; 2012 (cf. in this matter Roche 2008a: 207 n. 12); Hawley 2008a: 2008b. 59  This attitude could be due, to a certain extent, to the “impermeability rule” described by Nougayrol 1962: 31 and 32; an impermeability which could have been a necessity rather than a rule for the scribes. 60  In line with Mesopotamian tradition (cf. Malbran-Labat 1999: 76), Ugaritic scribes had already proved their capacity to exploit the possibilities rendered by the Mesopotamian logo-syllabic script to develop, e.g., complex graphic sets; see in this respect Roche 2008a; van Soldt 2010. This capacity for exploration continued, for instance, in the use of the alphabetic cuneiform script to write texts in Akkadian and Hurrian; see in this respect especially the work of scribe Ṯab’ilu, studied by Pardee 2010; 2012b; Lam and Pardee 2012: 421–422; Bordreuil, Hawley, and Pardee 2013; Hawley, Pardee, and Roche-Hawley 2013; Roche-Hawley and Hawley 2013. On the adaptation of the cuneiform alphabet to Hurrian language see Vita 2013. 61  The use of alphabetic and logo-syllabic scripts in one single text in Ugarit is particularly frequent in texts written in Ugaritic (cf. Malbran-Labat 1999; Roche 2008b), but not on

388 4

Vita

Akkadian and Canaanite

4.1 Amarna Letters The so-called ‘El-Amarna Letters’ were found around 1887 in the homonymous Egyptian locality, the ancient capital of Egypt under pharaoh Amenophis IV. This correspondence, a total of 350 letters and fragments, constitutes the main documentation on international relations in the ancient Near East in the 14th century BC. It also constitutes the main documentation on the history and language of Syria-Palestine in the 14th century BC.62 4.2 Canaanite and Canaano-Akkadian The Syro-Palestinian Amarna letters are linguistically interesting from various points of view. They were written on clay tablets, and the language used is Akkadian on the basis of an ancient Babylonian dialect using the Mesopotamian logo-syllabic script.63 But they were written by autochthonous Syro-Palestinian scribes whose mother tongue was not Akkadian but a Northwest Semitic language which is quite often reflected in the Akkadian they use, morphologically, syntactically, or lexically. The Akkadian used is, therefore, imbued with traits from other linguistic systems, the Northwest Semitic local languages and dialects.64 Consequently, Rainey (2010: 852) describes the language used in these letters as a hybrid made up of three layers: 1) the basic Old Babylonian, 2) the local modifications, and 3) the purely Northwest Semitic elements. As well as Old Babylonian, the letters also use (sometimes jointly) the Middle Babylonian and Middle Assyrian dialects (Rainey 2010: 853). The “local modifications” would be expressions that are neither Babylonian nor Canaanite, but reflect Canaano-Akkadian mixed forms. The purely the contrary, as stated by Malbran-Labat 1999: 96: “Il semble que ce soit un caractère fundamental de l’écriture d’Ougarit que de ne pas mêler de caracteres alphabétiques au sein d’un document noté en cuneiformes mésopotamiens.” The Ugaritic text RS 94.2411, recently edited (Bordreuil, Pardee, and Hawley 2012, 11), reveals that the mixture of scripts could occur not only in one single text but even in one single line within the text (cf. lines 8 and 9 of the text). 62  See regarding this the work by Cohen and Westbrook 2000. The main editions of the texts are Knudtzon 1907–1915 (complemented by Rainey 1978) and Rainey 2015. See also the useful transcription of the Canaanite letters provided by Sh. Izre’el in http://www.tau. ac.il/humanities/semitic/amarna.html. The main translations correspond to Moran 1992, Liverani 1998 and 1999, Cochavi-Rainey 2005, Rainey 2015. See also Mynářová 2007; 2014. 63  Rainey 1996a: 17–32. 64  This morphosyntactic phenomenon seems to go back to the letters of Taanak, from the mid-fifteenth century BC; see Rainey 1996a: 31–32.

Language Contact between Akkadian and NwS Languages

389

Northwest Semitic features present in the letters reflect the oldest linguistic characteristics of the Canaanite family.65 According to Rainey (2010: 855), this hybrid language “was a real dialect alongside Middle Babylonian and Middle Assyrian”. Thus, Northwest Semitic languages or local dialects appear in three main forms in this hybrid language: 1) Canaanite lexemes may appear in the middle of a sentence either independently or acting as glosses, translating a preceding word or phrase:66 – ìr-ka a-nu-ki I am your sevant’67 – ù sag.du-nu: ru-šu-nu i-na qa-te-ka ‘and our head is in your hand’68 – iA-ia-ab: ḫe-eḫ-bé-e ‘He has hidden Ayyab’69 – [ù] li-iḫ-šu-uš-mi: ia-az-ku-ur-mi ilugal-ri en-ia mi-im-ma ša in4-né-pu-uš-mi ugu uruḪa-ṣú-raki ‘[So] may the king, my lord, take thought for everything that has been done against Hazor’70 – ù iz-zi-iz-mì egir-šu: aḫ-ru-un-ú ‘and I took my place behind him’71 2) As in Ugarit, terms in the local language can also be found in lexical texts. In Canaan, unlike in Ugarit, just a small amount of these types of texts has been found though they are quite significant.72 The most relevant in this sense is a fragment of a lexical text found in Aphek (Aphek No 5837/1).73 The five preserved lines (none of them complete) reveal that the text was structured in three columns: the first for the Sumerian term, the second for the Akkadian

65  Cf. Izre’el 1995: 103; Steiner 1997: 146–147; Moran 2003: 343; Edzard 2011: 481. 66  This glosses could incorporate—or not—a so-called “gloss-wedge” or “Glossenkeil.” On the phenomenon of codeswitching in the Canaanite glosses existing in these letters, see Izre’el 1995: 101–122; Gianto 1995; Liverani 1998: 24–27; Tropper and Vita 2010: 26. 67  EA 287: 66. WS /ʾanōki/. 68  EA 264: 18. WS /rōšu-nū/. 69  EA 256: 7. WS /hiḥbê’/. Rainey 1996a, 12: 122; Tropper and Vita 2010: 101. 70  EA 228: 18–23. yazkur: a jussive form, cf. Rainey 1996a: 245. 71  EA 245: 10. Moran 1992, 299. WS /’aḫrōn-hu/ (Can. ’aḫr- + nun energicum -ōn- (< ān) + -hu; cf. Rainey 1996b: 122; Tropper and Vita 2010: 114). 72  I consider the idea of Tadmor (1977: 101–102) appropriate in the sense that it must be accepted “that centres of cuneiform literacy existed in each of the city-states of Canaan from the Middle Bronze to the close of the Late Bronze Ages.” It should be supposed that each chancery could have been, at least potentially, a school, even if it was a modest one. For a different viewpoint see recently van Soldt 2013. 73  Rainey 1976; Horowitz, Oshima, and Sanders 2006: 31–32.

390

Vita

version, and the third for the version in the local language. The first two lines read: 1’  . . . a].meš : ma-wu : 2’ gešti]n.meš : ka-ra-nu :

mu-mi74 ‘water’ ye-nu ‘wine’

It is quite probable that a second fragment of a lexical text also found in Aphek, as well as another found in Ashkelon, may also have had a column for Canaanite terms.75 The latter (part of the series Urra = ḫubullu) “follows the Emar version exactly” (Huehnergard and van Soldt 1999: 191), whereas the texts from Aphek show more freedom compared to the Mesopotamian models: “Neither of the Aphek lexical texts conforms to any known Mesopotamian series” (Rainey 1976: 139).76 All three fragments may be dated to the 13th century BC.77 So far, Canaan has not yielded any instances of lexical texts datable to the 14th century BC. Nonetheless, the Amarna letters do seem to reveal indirect evidence of the existence of lexical texts in Canaanite scribal centres during that period. Indications of this are provided by a number of glosses that appear to be unnecessary from a functional point of view which, in my opinion, could be elements that scribes extracted from lexical lists that they had used in their training with the purpose of hinting at their professional background, at their training as writing professionals, and at their expertise as scribes. Consequently, these glosses could somehow directly reflect elements from the lexicographical education of the scribes and may be of assistance in the identification of underlying lexical texts.78 The following case, identified by Shlomo Izre’el in a letter from Meggido, should be added to the instances reviewed in Vita (2012):79 ù la-a-mi ni-le-ú ZÚ.SI.GA ba-qa-ni: qà-˹ṣí ˺-ra ù la-a-mì ni-le-ú a-ṣí KÀ a-bu-ul-lí: ša-aḫ-ri iš-tu pa-ni ila-ab-a-ya ‘And we cannot do the plucking, and we cannot go out of the gate because of Lab’aya’ (EA 244: 13–17; Izre’el 2003: 91) 74  Concerning mu-mi (mū/ōmi), see Izre’el 2003: 77. 75  Aphek No 5837/1; cf. Rainey 1975; Horowitz, Oshima, and Sanders 2006: 29–30. Fragment from Ashkelon: Huehnergard and van Soldt 1999; Horowitz, Oshima and Sanders 2006: 42: “When complete, most likely Sumerian, Akkadian and West-Semitic entries.” 76  In this sense also Horowitz, Oshima, and Sanders 2006: 30. 77  Horowitz, Oshima, and Sanders 2006: 29, 32, 42; Huehnergard and van Soldt 1999: 184. 78  On this type of glosses, see in detail Vita 2012. 79  This example, overlooked in Vita 2012, was kindly pointed out to me by Prof. Shlomo Izre’el via email. See also the case of EA 244: 30–33 discussed in Izre’el 2003: 100.

Language Contact between Akkadian and NwS Languages

391

Izre’el remarks: “The scribe of this letter interestingly uses tripartite gloss sets. It reminds us of the trilingual vocabularies used in cuneiform schools, where each lexeme is represented by a Sumerian logogram, its Akkadian counterpart, and finally the local word, the latter two in syllabic writing . . . The employment of such tripartite gloss sets is understandable in the context of cuneiform scribal education” (2003: 91). 3) More frequently, however, the local language is apparent through forms which hybridise Canaanite and Akkadian elements, particularly frequent in the verbal system. Generally, scribes took a verbal base attested in Akkadian and combined it with Canaanite prefixes and suffixes. In the resulting form, Canaanite affixes predominated in terms of expressing time and aspect, the Akkadian verbal base being of little relevance in that sense:80 da-ag-la-ti7 ki-ia-am ù da-ag-la-ti7 ki-ia-am ù la-a na-mi-ir ‘I looked this way, and I looked that way, and there was no light’81 a-na-ku aq-bu [šum-ma UD.K]ÁM.MEŠ yi-iš-mu lugal-ru [ù UD].KÁM. MEŠ yi-ìl-te9-qú-šu-nu ‘I have been saying: “[If one da]y the king should hear, [then in one da]y he could seize them” ’82 a-na mi-ni7 yi-iš-tap-ru iRi-ib-dIŠKUR ki-na-an-na-ma ṭup-pa a-na É.GAL ‘Why does Rib-Haddi continually send a tablet in this manner to the palace?’83 ù ir-ka-ab-mi it-ti iYa-aš-da-ta ‘So I mounted up with Yashdata’84

80  See in detail Rainey 1996a; Izre’el 2005. Also, amongst others, von Dassow 2004: 644–647; Tropper and Vita 2010. 81  EA 292: 8–10. Moran 1992: 335. Akk. dagālu ‘to look’; qatl-ātī (qatl+ā+tī), cf. Rainey 1996a: 285; Tropper and Vita 2010: 70–71. 82  EA 109: 15–17. Rainey 1996a: 100. yi-iš-mu: a form using the Akkadian preterite theme (Rainey 1996a: 60–61); yi-ìl-te9-qú: a form based on the Akkadian -t- preterite (Rainey 1996a: 54). 83  EA 106: 14. Based on G preterite with infixed -t- or on Gt preterite, cf. Rainey 1996a: 102; Tropper and Vita 2010: 64 (yi-QTL-u). 84  Regarding the Akkadian term rakābu ‘to ride, to mount’, probably to represent Canaanite *’irkab as opposed to the Akkadian form arkab, cf. Rainey 1996a: 71.

392

Vita

This language, currently known as Canaano-Akkadian,85 is the product of the coexistence and interaction of two Semitic languages and does not constitute a cohesive linguistic system.86 It is an example of the situation where two different linguistic systems interacted, producing a novel grammatical organization. This situation is referred to as a language contact phenomenon and typically involves linguistic objects such as pidgins, creoles, koinés, and mixedlanguages. Consequently, Canaano-Akkadian is typically analysed within a language-contact framework.87 Yet, the exact linguistic nature of this language, and its possible oral use has been the subject of very varied proposals and continues to be debated.88 In our opinion,89 although Canaano-Akkadian emerged from the situation of diglossia and had its roots in two linguistic systems (Akkadian and Canaanite), the real state of affairs at the time of composing the Amarna letters corresponded rather to triglossia. That is to say, three languages coexisted at the scribal centres: Akkadian, Canaanite, and the outcome of their “blending,” Canaano-Akkadian.90 The Akkadian was a high variety and superstrate: it was typically written but possibly also spoken, although very infrequently, for instance during the Akkadian lessons. Canaanite was a low variety and substrate: it was used by the vast part of the society; it was typically spoken and only infrequently written,91 e.g., glosses. Canaano-Akkadian was a high variety “mixture” of the superstrate and substrate: it was typically written although, in agreement with Izre’el (2012), we suggests that Canaano-Akkadian was a 85  A denomination used first by Shlomo Izre’el 2005 (first edition 1998). 86  Izre’el 2005: 3, appropriately describes the situation as follows: “The CanAkk texts are characterized by inherent variation. Variation may be geographically dependent. It may depend on the scribal tradition of different cities, but also on imported traditions or variant local ones . . . Geographic variation is dependent upon scribal traditions and scribal education. By and large, there is correlation between the provenance of a letter and its linguistic structure. The farther south one travels in Canaan the more remote becomes a CanAkk text from Akkadian, and the closer it becomes to the Canaanite vernacular of that region.” 87  See recently Johanson 2013: 281–282. 88  Regarding the various proposals of linguistic classification of Canaano-Akkadian and relevant bibliography, see in detail Tropper and Vita 2010: 20–24; Andrason and Vita 2014. More significant studies: Gianto 2000; von Dassow 2004; Sanders 2009; Rainey 2010; Izre’el 2012. 89  On the subject expounded next, see in more detail Andrason and Vita 2014. 90  These three tongues match the three layers of the language of the letters as posited by Rainey, see above. 91  At least by means of the Mesopotamian logosyllabic writing.

Language Contact between Akkadian and NwS Languages

393

type of a mixed-language that was genuinely spoken at the scribal centres, constituting a professional jargon. It is possible that conforming to the typical behaviour of jargons, scribes may have been using Canaano-Akkadian as their in-group code for all types of conversations, not only the formal ones. Thus, to an extent, Canaano-Akkadian would be a middle stage between its two sources, having emerged from the previously mentioned diglossial circumstances. All of this suggests that, rather than a single static definition, one may understand Canaano-Akkadian as a dynamic fuzzy (yet synchronic) object where properties and situations typical for pidgins (more specifically, their sub-type, koinés), mixed-languages, and jargons intervene.92 On the whole, it is a unique linguistic organization shaped by its exceptional sociolinguistic environment.93 4.3 The Attitude of Canaanite Scribes towards their Mother Tongue As we have seen, in the Amarna corpus the pressure of the local language over Akkadian is notably stronger than in Ugarit. The use of hybrid CanaanoAkkadian forms, for instance, is more intense and systematic than in Akkadian texts from Ugarit, resulting in the language known as “Cannaano-Akkadian” (§4.2). A more likely atomized training in Akkadian in Canaan, scattered about in smaller and more or less independent centers,94 may have contributed to this outcome in the lack of a strong and centralized scribal center, which in all likelihood existed in Ugarit. The Canaanite terms and grammar elements present in the Amarna letters (cf. §4.2) jointly demonstrate that, as well as Ugarit scribes (cf. §3.4), Canaanite scribes were able to use the Mesopotamian logo-syllabic writing system in order to render their mother tongue. Nevertheless, both the lack of texts written in the local language95 and the instability of writing of Canaanite forms96 suggest that the scribes—again, similar to their colleagues from Ugarit—were uninterested in developing an orthographic convention with which they could render their Northwest Semitic mother tongues or dialects by employing this type of writing system. Along with the influence of the Mesopotamian school 92  On the contrary, the idiom does not fulfil the necessary condition of a creole—it was never nativized. 93  The “dynamic” classification of Canaano-Akkadian formulated above, albeit distinct from other views, should not been understood as invalidating the linguistic tradition of the Canaano-Akkadian studies. Quite the reverse, we see our proposal (Andrason and Vita 2014) as profoundly inclusive. 94  For a more negative opinion on scribal schools in Canaan, see von Dassow 2004: 667. 95  A possible exception may be RS 94.2615, edited by Arnaud 2006. See, however, the view presented by von Dassow 2010: 901 n. 13. 96  Regarding this, see Andrason and Vita 2014.

394

Vita

tradition, and as in the case of Emar and Ugarit, one possible reason for this attitude may have been the fact that Canaano-Akkadian was a language spoken at the scribal centers and may have constituted a professional jargon (§4.2). Hence, the writing of the local West Semitic languages or dialects had to be achieved via different manners, most probably by means of the alphabet (as in Ugarit) and using mostly perishable materials. In this sense, we ought to remember that in this geographical area there seems to be alphabetic evidence going back to even before the 14th century BC,97 and that Canaanite scribes in general developed their own strong personality boasting great creativity and an experimental disposition.98 In this manner, Canaanite would offer a situation similar to what can be observed in Ugarit, where the scribes chose to develop (quite probably earlier than in Ugarit, §3.4) an alphabetic cuneiform system in order to note their mother tongue. 5

Final Comments and Conclusions

Of all the Northwest Semitic languages considered, Emarite and Canaanite are only partially documented. No full texts in these languages are preserved, just linguistic elements interspersed, to varying degrees, in the Akkadian language used by local scribes. Ugaritic is a remarkable exception, particularly due to the large number of texts in this language written by means of an alphabetic script. Epps, Huehnergard, and Pat-El (2013) have pointed out the specific problems attached to the study of contacts between genetically related languages. All three cases dealt with in this chapter could fit into this category of linguistic contact, though two particular features can be noted. On the one hand, Akkadian was not a language spoken in society in Emar, Ugarit, or Canaan (at least not significantly). In those places, therefore, prolonged and significant contact between two linguistic communities interacting in the same area did 97  According to Hachmann, Kāmid el-Lōz could have yielded lineal alphabetic inscriptions archaeologically dating from the 14th century BC (Hachmann, 1993: 228) plus one in cuneiform alphabet dating from the 14th or 15th century BC (Hachmann 1980: 106; against this, see Sass 1988: 165–166). The Proto-Canaanite Inscriptions are also to be borne in mind (Sass 1988; Finkelstein and Sass 2013). 98  As we have seen (§4.2), some of the school texts found in Canaan are adaptations of Mesopotamian models for local needs or circumstances. This creative and adaptive ability goes back to the third millennium, as demonstrated by the syllabary found in Byblos (Dossin 1969: 247; Edzard 1982: 58; Cavigneaux 1980–1983: 618) and by the so-called “pseudo-hieroglyphic script”, also from Byblos.

Language Contact between Akkadian and NwS Languages

395

not exist.99 Contact between Akkadian and the aforementioned Northwest Semitic languages occurred within communities of scribes, that is, in the bosom of small communities of (in principle) highly specialized professionals. The mother tongue of those scribes was Northwest Semitic, but their language of study, work, and prestige was Akkadian.100 The prestige of Akkadian among these communities of scribes produced a second peculiarity in this type of linguistic contact concerning writing: scribes in Syria-Palestine in the Late Bronze Age did not seem to conceive of the use of Mesopotamian logo-syllabic script for writing local Northwest Semitic languages. This fact may concern, in the ancient Near East of the second millennium BC, only this type of languages and does not affect, however, non-Semitic languages such as Hurrian or Hittite. At least in the cases of Ugarit and Canaan, scribes demonstrated that this way was feasible though it was not exploited. Inhibiting elements must therefore have been in play which prevented Northwest Semitic languages from being written in Mesopotamian cuneiform script. One reason could have been the aforementioned prestige of Akkadian in the Syro-Palestinian scribal centers, where the direct presence of Babylonian and Assyrian scribes can also be detected;101 a prestige Akkadian maintained throughout the first millennium BC.102 This prestige could also have caused Akkadian to be a language (to a greater or lesser extent) spoken in the circles of Syro-Palestinian scribes. These could probably have produced professional jargons where Akkadian was interspersed, to varying degrees, with local linguistic traits, a situation undoubtedly propitiated by the genetic relationship

99  This historical and linguistic situation would substantially differ from that described, for instance, by Beaulieu for the case of contacts between Akkadian and Aramaic in the first millennium BC, where both languages “were in intimate contact for centuries, being spoken and written side by side in the same society” (Beaulieu 2013: 358). 100  As previously seen (§3.4), the writing of Ugaritic was also taught at schools largely following the Mesopotamian curricular model. 101  The presence of Mesopotamian scribes in Ugarit has already been pointed out above (§3.4). Regarding Emar, see Cohen 2009: 183. The existence of foreign scribes in Canaan, highly plausible, remains yet to be studied in detail. 102  In the case of contact between Akkadian and Aramaic, Beaulieu 2013: 375 states: “Babylonian resisted the influence of Aramaic to a considerable degree. This could be attributed to some extent to a systematic and self-conscious purist endeavour on the part of scribes and the learned classes . . . Babylonian enjoyed a more prestigious status as the language favoured by the monarchy and the urban elites . . . while Aramaic seems to have been confined largely to an administrative and juridical role, albeit a significant one in those areas.”

396

Vita

between the languages involved.103 In that sense, Canaano-Akkadian seems to set, in Syria-Palestine in the Late Bronze Age, the limit reached (in the written and spoken registers) by linguistic contact between Semitic languages.104 The scribes’ perception that Akkadian and Mesopotamian logo-syllabic scripts were barely dissociable must have also had an impact.105 Consequently, the logo-syllabic script may not have been considered adequate for writing in other Semitic languages; a situation where, once more, the genetic relationship between the languages involved may have played a part (yet to be defined). Some of those Northwest Semitic languages, conversely, were written using alphabetic scripts and partly perishable materials, at least in the Levantine strip of Ugarit and Canaan.106 These scripts were free of the weight of centennial scribal traditions and therefore conceptually appropriate to transcribe local languages of this geographical area. Bibliography Andrason, A. and Vita, J.-P. 2014. “From Glosses to the Linguistic Nature of CanaanoAkkadian,” Folia Orientalia 51 (2014). In press. Arnaud, D. 2006. “Un fragment de lettre en canaanéen: RS 94.2615, provenant sans doute de Tyr,” AuOr 24: 7–15. Beaulieu, P.-A. 2013. “Aspects of Aramaic and Babylonian Linguistic Interaction in First Millennium BC Iraq,” Journal of Language Contact 6: 358–378. Bordreuil, P. and D. Pardee. 2009. A Manual of Ugaritic. Winona Lake. Bordreuil, P., R. Hawley, and D. Pardee. 2013. “Données nouvelles sur le déchiffrement de l’alphabet et sur les scribes d’Ougarit,” in P. Bordreuil, F. Ernst-Pradal, M. G. Masetti-Rouault, H. Rouillard-Bonraisin, M. Zink (eds.), Les écritures mises au jour sur le site antique d’Ougarit (Syrie) et leur déchiffrement. Paris. 319–332. Cavigneaux, A. 1980–1983. “Lexikalische Listen,” RLA 6: 609–641. 103  See Beaulieu 2013 regarding linguistic transfers between Aramaic and Akkadian propitiated precisely by the close genetic relation between both languages. 104  Canaano-Akkadian will also need to be studied within a broader framework including the linguistic phenomena resulting from contemporary contacts between Akkadian and Hurrian and between Ugaritic and Hurrian. Therefore including the archives of Alalaḫ IV (slightly earlier in time; von Dassow 2012) and the mixed Akkadian-Hurrian language of Qaṭna (Richter, in Richter and Lange 2012: 29–42, with previous bibliography); see also von Dassow 2010: 905. Regarding Hurrian in Ugarit, see recently Vita 2009; 2013; Giorgieri 2013: 167–178. 105  See already in this respect the relevant remarks by Nougayrol 1962: 32. 106  Emar, conversely, was far from the area of influence of alphabetic scripts; see also comments by Millard 2007: 86–87.

Language Contact between Akkadian and NwS Languages

397

Chavalas, M. W. (ed.). 1996. Emar: the History, Religion and Culture of a Syrian Town in the Late Bronze Age. Bethesda. Cochavi-Rainey, Z. 2005. To the King my Lord. Letters from El-Amarna, Kumidu, Taanach and Other Letters from the Fourteenth Century BCE. Jerusalem. Cohen, Y. 2002. “The West Semitic/Peripheral Akkadian Term for ‘Lung’,” JAOS 122: 824–827. ———. 2009. The Scribes and Scholars of the City of Emar in the Late Bronze Age. Winona Lake. ———. 2010a. “The ‘Second Glosses’ in the Lexical Lists from Emar: West Semitic or Akkadian?,” in L. Kogan et al. (eds.), Proceedings of the 53e Rencontre Assyriologique Internationale, Vol. 1, Part 2. Language in the Ancient Near East (= Babel und Bibel 4/2). Eisenbrauns. 813–839. ———. 2010b. “Rara Avis—A Study of the HU Section of the Sa Vocabulary,” in. D. Baker, E. Robson, and G. Zolyomi (eds.), Your Praise is Sweet: A Memorial Volume for Jeremy A. Black from Students, Colleagues and Friends. Oxford. 29–40. ———. 2012. “An Overwiew on the Scripts of Late Bronze Age Emar,” in E. Devecchi (ed.), Palaeography and Scribal Practices in Syro-Palestine and Anatolia in the Late Bronze Age. Leiden. 33–45. ———. 2013. “Problems in the History and Chronology of Emar,” KASKAL 10: 281–294. Cohen, Y. and I. Singer. 2006, “A Late Synchronism between Ugarit and Emar,” in Y. Amit et al. (eds.), Essays on Ancient Israel in Its Near Eastern Context. A Tribute to Nadav Na’aman. Winona Lake. 123–139. Cohen, R. and R. Westbrook. 2000. Amarna Diplomacy. The Beginnings of International Relations. Baltimore. d’Alfonso, L. and Cohen, Y. 2008. “The Duration of the Emar Archives and the Relative and Absolute Chronology of the City,” in d’Alfonso, Cohen, and Sürenhagen 2008: 3–25. d’Alfonso, L., Y. Cohen, and D. Sürenhagen. 2008. Control, Interaction, and Inter­ dependence. The City of Emar among the Late Bronze Age Empires (14th–12th centuries BC) (AOAT 349). Münster. 2008. von Dassow, E. 2004. “Canaanite in Cuneiform,” JAOS 124: 641–674. ———. 2010, “Peripheral Akkadian Dialects, or Akkadography of Local Languages?,” in L. Kogan et al. (eds), Proceedings of the 53e Rencontre Assyriologique Internationale, Vol. 1. Part 2. Language in the Ancient Near East (= Babel und Bibel 4/2). Winona Lake. 895–924. ———. 2012. “Gloss Marking and the Language of the Alalaḫ IV Texts,” in E. Devecchi (ed.), Palaeography and Scribal Practices in Syro-Palestine and Anatolia in the Late Bronze Age. Leiden. 201–216. del Olmo Lete, G. 2012. “Ugaritic and Old(-South)-Arabic: Two WS Dialects?,” in F. Corriente, G. del Olmo Lete, Á. Vicente, J.-P. Vita (eds.), Dialectology of the Semitic

398

Vita

Languages. Proceedings of the IV Meeting on Comparative Semitics, Zaragoza 11/6– 9/2010. Sabadell. 5–23. Démare-Lafont, S. 2010. “Éléments pour une diplomatique juridique des textes d’Émar,” in S. Démare-Lafot and A. Lemaire (eds.), Trois millénaires de formulaires juridiques. Geneva. 43–84. Dietrich, M. and O. Loretz. 1988. Die Keilalphabete. Die phönizisch-kanaanäischen und altarabischen Alphabete in Ugarit. Münster. Dietrich, M., O. Loretz, and J. Sanmartín. 2013. The Cuneiform Alphabetic Texts from Ugarit, Ras Ibn Hani and Other Places (Third, Enlarged Edition). Münster. Dietrich, M. and W. Mayer. 1999. “The Hurrian and Hittite Texts,” in W. G. E. Watson and N. Wyatt (eds.), Handbook of Ugaritic Studies. Leiden. 58–75. Dossin, G. 1969. “Trois inscriptions cunéiformes de Byblos,” Mélanges de l’Université Saint-Joseph 45: 243–255. Edzard, D. O. 1982. “Der Aufbau des Syllabars ‘Proto-Ea’,” in Society and Languages of the Ancient Near East. Studies in Honour of I. M. Diakonoff. Warminster. 42–61. Edzard, L. 2011. “Biblical Hebrew,” in S. Weninger (ed.), The Semitic Languages. An International Handbook. Berlin. 480–514. Egetmeyer, M. 2013. “Ougarit et le déchiffrement de ses inscriptions en syllabaire chypro-minoen,” in P. Bordreuil, F. Ernst-Pradal, M. G. Masetti-Rouault, H. RouillardBonraisin, M. Zink (eds.), Les écritures mises au jour sur le site antique d’Ougarit (Syrie) et leur déchiffrement. Paris. 133–155. Epps, P., J. Huehnergard, and N. Pat-El. 2013. “Introduction: Contact Among Genetically Related Languages,” Journal of Language Contact 6: 209–219. Faist, B. I. 2008. “Scribal Traditions and Administration at Emar,” in d’Alfonso, Cohen, and Sürenhagen 2008: 195–205. Faist, B. I., J.-J. Justel, J.-P. Vita. 2003. “Bibliografía de los estudios de Emar,” UF 35: 191–230. ———. 2005. “Bibliografía de los estudios de Emar (2),” UF 37: 329–340. ———. 2007. “Bibliografía de los estudios de Emar (3),” UF 39: 141–160. ———. 2009. “Bibliografía de los estudios de Emar (4),” UF 41: 469–504. Faist, B. I. and J.-P. Vita. 2008. “Der Gebrauch von ashar in den akkadischen Texten aus Emar,” WO 38: 53–60. Ferrara, S. 2012. Cypro-Minoan Inscriptions, Vol. 1. Analysis. Oxford. ———. 2013. Cypro-Minoan Inscriptions, Vol. 2. The Corpus. Oxford. Finkelstein, I. and B. Sass. 2013. “The West Semitic Alphabetic Inscriptions, Late Bronze II to Iron IIA: Archeological Context, Distribution and Chronology,” Hebrew Bible and Ancient Israel 2: 149–220. Fleming, D. and S. Démare-Lafont. 2009. “Tablet Terminology at Emar: ‘Conventional’ and ‘Free Format’,” AuOr 27: 19–26. Freu, J. 2006. Histoire politique du royaume d’Ugarit. Paris.

Language Contact between Akkadian and NwS Languages

399

Gianto, A. 1995. “Amarna Lexicography: The Glosses in the Byblos Letters,” SEL 12. 65–73. ———. 2000. “Amarna Akkadian as a Contact Language,” in K. van Lerberghe and G. Voet (eds.), Languages and Cultures in Contact. Leuven. 123–132. Giorgieri, M. 2013. “Diffusion et caractéristiques de la culture écrite d’origine hourrite dans le Proche-Orient asiatique et à Ougarit,” in P. Bordreuil, F. Ernst-Pradal, M. G. Masetti-Rouault, H. Rouillard-Bonraisin, M. Zink (eds.), Les écritures mises au jour sur le site antique d’Ougarit (Syrie) et leur déchiffrement. Paris. 157–185. Gzella, H. 2011. “The Semitic Languages and Dialects III: North-West Semitic,” in S. Weninger (ed.), The Semitic Languages. An International Handbook. Berlin. 425–451. Hachmann, R. 1980. “Zur Stratigraphie des Gefässfragments mit alphabetischer Keilinschrift,” in R. Hachmann (ed.), Kāmid el-Lōz 1968–70. Bonn. 103–109. ———. 1993. “Ostraka mit eingeritzten und eingestempelten Zeichen aus Kāmid el-Lōz (Libanon),” in M. J. Mellink, E. Porada, and T. Özgüç (eds.), Aspects of Art and Iconography: Anatolia and Its Neighbours. Studies in Honor of N. Özgüç. Ankara. 227–272. Hackett, J. A. 2002. “The Study of Partially Documented Languages,” in Sh. Izre’el (ed.), Semitic Linguistics: The State of the Art at the Turn of the Twenty-First Century. Winona Lake. 57–75. Hasselbach, R. and J. Huehnergard. 2007. “Northwest Semitic Languages,” in EALL 3.408–422. Hawley, R. 2008a. “Apprendre à écrire à Ougarit: une typologie des abécédaires,” in C. Roche (ed.), D’Ougarit à Jérusalem. Recueil d’études épigraphiques et archéologiques offert à Pierre Bordreuil. Paris. 215–232. ———. 2008b. “On the Alphabetic Scribal Curriculum at Ugarit,” in R. D. Biggs, J. Myers, and M. T. Roth (eds.), Proceedings of the 51st Reencontre Assyriologique Internationale Held at The Oriental Institute of the University of Chicago. July 18–22, 2005. Chicago. 57–67. Hawley, R., D. Pardee, and C. Roche-Hawley. 2013. “À propos des textes akkadiens alphabétiques conservés au Musée d’Alep: notes épigraphiques,” in V. Matoïan and M. Al-Maqdissi (eds.), Études ougaritiques III. Leuven. 395–401. Horowitz, W., T. Oshima, T., and S. Sanders. 2006. Cuneiform in Canaan. Cuneiform Sources from the Land of Israel in Ancient Times. Jerusalem. Huehnergard, J. 1989. The Akkadian of Ugarit. Atlanta. ———. 1999. “Ugaritic Words in Syllabic Texts,” in W. G. E. Watson and N. Wyatt (eds.), Handbook of Ugaritic Studies. Leiden. 134–139. ———. 2005. “Features of Central Semitic,” in A. Gianto (ed.), Biblical and Oriental Essays in Memory of William L. Moran. Rome. 155–203. ———. 2008. Ugaritic Vocabulary in Syllabic Transcription (2nd ed.). Atlanta.

400

Vita

———. 2012. An Introduction to Ugaritic. Peabody. Huehnergard, J. and W. H. van Soldt. 1999. “A Cuneiform Lexical Text from Ashkelon with a Canaanite Column,” IOS 49: 184–192. Ikeda, J. 1999. “Scribes in Emar”, in K. Watanabe (ed.), Priests and Officials in the Ancient Near East. Heidelberg. 163–185. ———. 2003. “A New Contribution to Northwest Semitic Lexicography,” BiOr 60: 263–279. ———. 2010. “Was Akkadian Spoken in Emar? Diglossia in Emar,” in L. Kogan et al., (eds), Proceedings of the 53e Rencontre Assyriologique Internationale, Vol. 1. Part 2. Language in the Ancient Near East (= Babel und Bibel 4/2). Winona Lake. 841–850. Izre’el, Sh. 1995. “The Amarna Glosses: Who Wrote What for Whom?,” IOS 15: 101–122. ———. 2003. “Canaanite Varieties in the Second Millennium BC: Can We Dispense with Anacronism?,” Orient 38: 66–104. ———. 2005. Canaano-Akkadian (2nd ed.). München. ———. 2012. “Canaano-Akkadian: Linguistics and Sociolinguistics,” in R. Hasselbach and N. Pat-El (eds.), Language and Nature. Papers Presented to John Huehnergard on the Occasion of His 60th Birthday. Chicago. 171–218. Johanson, L. 2013. “Written Language intertwining,” in P. Bakker and Y. Matras (eds.), Contact Languages. A Comprehensive Guide. Berlin. 273–331. Knudsen, E. E. 2004. “Amorite Vocabulary. A Comparative Statemen,” in J. G. Dercksen (ed.), Assyria and Beyond. Studies Presented to Mogens Trolle Larsen. Leiden. 319–331. Knudtzon, A. 1907–1915. Die El-Amarna-Tafeln. Leipzig. Kouwenberg, B. 2011. “Akkadian in General,” in S. Weninger (ed.), The Semitic Languages. An International Handbook. Berlin. 330–340. Lagarce-Othman, B. 2013. “Un nouveau base inédit d’Horemheb,” in V. Matoïan and M. Al-Maqdissi (eds.), Études ougaritiques III. Leuven. 347–364. Lam, J. and D. Pardee. 2012, “Diachrony in Ugaritic,” in C. L. Miller-Naudé and Z. Zevit (eds.), Diachrony in Biblical Hebrew. Winona Lake. 407–431. Liverani, M. 1998. Le lettere di el-Amarna. 1. Le lettere dei “Piccoli Re”. Brescia. Malbran-Labat, F. 1995. “Éléments de la structure du discours dans l’akkadien d’Ugarit,” GLECS 31: 87–85. ———. 1996. “Akkadien, bilingues et bilinguisme en Élam et à Ougarit,” in F. BriquelChatonnet (ed.), Mosaïque de langues, mosaïque culturelle. Le bilinguisme dans le Proche-Orient ancien. Paris. 33–61. ———. 1999. “Langues et écritures à Ugarit,” Semitica 49: 65–101. ———. 2008. “Catalogue raisonné des textes akkadiens de la ‘Maison d’Urtēnu’,” in C. Roche (ed.), D’Ougarit à Jérusalem. Recueil d’études épigraphiques et archéologiques offert à Pierre Bordreuil. Paris. 21–38.

Language Contact between Akkadian and NwS Languages

401

Malbran-Labat, F. and C. Roche. 2007. “Urtēnu Ur-Tešub,” in J.-M. Michaud (ed.), Le royaume d’Ougarit de la Crète à l’Éuphrate. Nouveaux axes de recherche. Sherbrooke. 63–104. Márquez Rowe, I. 2004. “Reflexiones sobre el acadio como lengua de contacto en el Antiguo Oriente,” in P. Badenas, S. Torallas, E. R. Luján, and M. Á. Gallego (eds.), Lenguas en contacto: el testimonio escrito. Madrid. 17–34. ———. 2006. The Royal Deeds of Ugarit. A Study of Ancient Near Eastern Diplomatics. Münster. Millard, A. 2007. “Alphabetic Writing, Cuneiform and Linear, Reconsidered,” Maarav 14: 83–93. Moran, W. L. 1992. The Amarna Letters. Baltimore. ———. (ed. J. Huehnergard and Sh. Izre’el). 2003. Amarna Studies. Collected Writings. Winona Lake. Mynářová, J. 2007. Language of Amarna—Language of Diplomacy. Perspectives on the Amarna Letters. Prague. ———. 2014. “Egyptian State Correspondence of the New Kingdom: The Letters of the Levantine Client Kings in the Amarna Correspondence and Contemporary Evidence,” in K. Radner (ed.), State Correspondence in the Ancient World. From New Kingdom Egypt to the Roman Empire. London. 10–31. Nougayrol, J. 1962. “L’influence babylonienne à Ugarit, d’après les textes en cuneiformes classiques,” Syria 39: 28–35. Pardee, D. 2010. “Un chant nuptial ougaritique (RS 5.194 [CTA 24]). Nouvelle étude épigraphique suivie de remarques philologiques et littéraires,” Semitica et Classica 3: 13–46. ———. 2012a. The Ugaritic Texts and the Origins of West-Semitic Literary Composition. Oxford. ———. 2012b. “RS 5.229: Restitution d’une nouvelle signature du Scribe Ṯab’ilu,” in C. Roche-Hawley and R. Hawley (eds.), Scribes et érudits dans l’orbite de Babylone. Paris. 31–49. Pentiuc, E. J. 2001. West Semitic Vocabulary in the Akkadian Texts from Emar. Winona Lake. Pruzsinszky, R. 2003. Die Personennamen der Texte aus Emar. Bethesda. ———. 2009. “Notes on the Chronology and Kings of Emar,” SCCNH 18: 421–430. Rainey, A. F. 1975. “Two Cunieform Fragments from Tel Aphek,” Tel Aviv 2: 125–129. ———. 1976. “A Tri-Lingual Cuneiform Fragment from Tel Aphek,” Tel Aviv 3: 137–140. ———. 1978. El Amarna Tablets 359–379. Supplement to J. A. Knudtzon Die El-AmarnaTafeln (2nd ed.). Neukirchen-Vluyn. ———. 1996a. Canaanite in the Amarna Tablets, vol. 2. Leiden. ———. 1996b. Canaanite in the Amarna Tablets, vol. 3, Leiden.

402

Vita

———. 2010. “The Hybrid Language Written by Canaanite Scribes in the 14th Century BCE,” in L. Kogan et al. (eds), Proceedings of the 53e Rencontre Assyriologique Internationale, Vol. 1. Part 2. Language in the Ancient Near East (= Babel und Bibel 4/2). Winona Lake. 851–861. ______. 2015. The El-Amarna Correspondence. A New Edition of the Cuneiform Letters from the Site of El-Amarna based on Collations of all Extant Tablets. Leiden. Reculeau, H. 2010. “Périphérique ou local? Le vocabulaire des paysages de la Vallée de l’Euphrate au IIe millénaire av. n. è.,” in L. Kogan et al. (eds), Proceedings of the 53e Rencontre Assyriologique Internationale, Vol. 1. Part 1. Language in the Ancient Near East (= Babel und Bibel 4/1). 505–520. Richter, T. and S. Lange. 2012. Das Archiv des Idadda. Die Keilschrifttexte aus den deutsch-syrischen Ausgrabungen 2001–2003 im Königspalast von Qatna. Wiesbaden. Roche, C. 2008a. “Jeux de mots, jeux de signes en Ougarit ou de l’influence des textes lexicaux sur les scribes de périphérie,” in C. Roche (ed.), D’Ougarit à Jérusalem. Recueil d’études épigraphiques et archéologiques offert à Pierre Bordreuil. Paris. 205–214. ———. 2008b. “Classification de l’utilisation du cunéiforme mésopotamien dans les textes ougaritiques,” in R. D. Biggs, J. Myers, and M. T. Roth (eds.), Proceedings of the 51st Reencontre Assyriologique Internationale Held at The Oriental Institute of the University of Chicago. July 18–22, 2005. Chicago. 155–170. ———. “Language and Script in the Akkadian Economic Texts from Ras Shamra,” in W. H. van Soldt (ed.), Society and Administration in Ancient Ugarit. Leiden. 107–122. Roche-Hawley, C. and R. Hawley. 2013. “An Essay on Scribal Families, Tradition, and Innovation in Thirteenth-Century Ugarit,” in B. J. Collins and P. Michalowski (eds.), Beyond Hatti. A Tribute to Gary Beckman. Atlanta. 242–264. Sanders, S. L. 2009. The Invention of Hebrew. Urbana. Sass, B. 1988. The Genesis of the Alphabet and Its Development in the Second Millennium BC. Wiesbaden. Seminara, S. 1998. L’accadico di Emar. Rome. Singer, I. 1999. “A Political History of Ugarit,” in W. G. E. Watson and N. Wyatt (eds.), Handbook of Ugaritic Studies. Leiden. 603–733. ———. 2011. The Calm before the Storm. Selected Writings of Itamar Singer on the End of the Late Bronze Age in Anatolia and the Levant. Atlanta. 19–146. Sjöberg, Å. W. 1998. “Studies in the Emar Sa Vocabulary,” ZA 88: 240–283. van Soldt, W. H. 1991. Studies in the Akkadian of Ugarit. Dating and Grammar. Neukirchen-Vluyn. ———. 1995a. “The Akkadian of Ugarit: Lexicographical Aspects,” SEL 12: 205–215. ———. 1995b. “Babylonian Lexical, Religious and Literary Texts and Scribal Education at Ugarit and its implications for the alphabetic literary texts,” in M. Dietrich and O. Loretz (eds.), Ugarit. Ein ostmediterranes Kulturzentrum im Alten Orient.

Language Contact between Akkadian and NwS Languages

403

Ergebnisse und Perspektiven der Forschung, Vol. 1. Ugarit und seine altorientalische Umwelt. Münster. 171–212. ———. 1999. “The Society of Ugarit,” in W. G. E. Watson and N. Wyatt (eds.), Handbook of Ugaritic Studies. Leiden. 28–45. ———. 2001. “Naḫiš-šalmu, an Assyrian Scribe Working in the ‘Southern Palace’ at Ugarit,” in W. H. van Soldt, J. G. Dercksen, N. J. C. Kouwenberg, and Th. J. H. Krispijn (eds.), Studies Presented to Klaas R. Veenhof on the Occasion of His Sixty-Fifth Birthday. Leiden. 429–444. ———. 2002. “The Orthography of Ugaritic Words in Texts Written by the Assyrian Scribe Naḫiš-šalmu,” in O. Loretz, K. A. Metzler, H. Schaudig (eds.), Ex Mesopotamia et Syria Lux. Festschrift für Manfried Dietrich zu seinem 65. Geburtstag. Münster. 685–697. ———. 2008. “The Ugarit Version of Syllabary A,” in R. J. van der Spek (ed.), Studies in Ancient Near Eastern World View and Society presented to Marten Stol on the Occasion of His 65th Birthday, 10 November 2005, and His Retirement from the Vrije Universiteit Amsterdam. Bethesda. 255–275. ———. 2010. “The Interaction of Syllabic and Alphabetic Cuneiform Writing in Ugarit,” in A. de Voogt and I. Finkel (eds.), The Idea of Writing: Play and Complexity. Leiden. 345–355. ———. 2011. “Akkadian as a Diplomatic Language,” in S. Weninger (ed.), The Semitic Languages. An International Handbook. Berlin. 405–415. ———. 2012. “The Palaeography of two Ugarit Archives,” in E. Devecchi (ed.), Palaeography and Scribal Practices in Syro-Palestine and Anatolia in the Late Bronze Age. Leiden. 171–183. ———. 2013. “The Extent of Literacy in Syria and Palestine during the Second Millennium BCE,” in L. Feliu, J. Llop, A. Millet, J. Sanmartín (eds.), Time and History in the Ancient Near East. Proceedings of the 56th Rencontre Assyriologique Internationale at Barcelona. 26–30 July 2010. Winona Lake. 19–31. Steiner, R. C. 1997. “Ancient Hebrew,” in R. Hetzron (ed.), The Semitic Languages. London. 145–173. Streck, M. P. 2000. Das amurritische Onomastikon der altbabylonischen Zeit, Vol. 1. Die Amurriter. Die onomastische Forschung. Orthographie und Phonologie. Nominal­ morphologie. Münster. ———. 2010. “Großes Fach Altorientalistik: Der Umfang des keilschriftlichen Textkorpus,” MDOG 142: 35–58. ———. 2011. “Babylonian and Assyrian,” in S. Weninger (ed.), The Semitic Languages. An International Handbook. Berlin. 359–396. Tadmor, H. 1977. “A Lexicographical Text from Hazor,” IEJ 27: 98–102. Thomason, S. G. 2001. Language Contact. An Introduction. Edinburgh. Tropper, J. 2012. Ugaritische Grammatik (2nd ed.). Münster.

404

Vita

Tropper, J. and Vita, J.-P. 2010. Das Kanaano-Akkadische der Amarnazeit. Münster. Veldhuis, N. 2014. History of the Cuneiform Lexical Tradition. Münster. Vita, J.-P. 2009. “Hurrian as a living language in Ugaritic society,” in D. A. Barreyra Fracaroli and G. del Olmo Lete (eds.), Reconstructing a Distant Past. Ancient Near Eastern Essays in Tribute to Jorge R. Silva Castillo. Sabadell-Barcelona. 219–231. ———. 2012. “On the Lexical Background of the Amarna Glosses,” AoF 39: 278–286. ———. 2013. “Alphabet ougaritique et langue hourrite: interactions et adaptations,” in P. Bordreuil, F. Ernst-Pradal, M. G. Masetti-Rouault, H. Rouillard-Bonraisin, M. Zink (eds.), Les écritures mises au jour sur le site antique d’Ougarit (Syrie) et leur déchiffrement. Paris. 203–217. Watson, W. G. E. 2002. “Emar and Ugaritic,” NABU 2002: 10–11 (no. 9). ———. 2003. Review of Pentiuc 2001. JSS 48: 355–357. ———. 2007. Lexical Studies in Ugaritic. Sabadell – Barcelona. Westbrook, R. 2003. “Emar and Vicinity,” in R. Westbrook (ed.), A History of Ancient Near Eastern Law, vol. 1. Leiden. 657–691. Zamora, J.-Á. 2006. “Les utilisations de l’alphabet lors de IIe millénaire av. J.-C. et le développement de l’épigraphie alphabétique: une aproche à travers la documentation ougaritique en dehors des tablettes (II),” in G. del Olmo, Ll. Feliu, and A. Mollet (eds.), Šapal tibnim mû illakū. Studies Presented to Joaquín Sanmartín on the Ocasión of His 65th Birthday. Sabadell-Barcelona. 491–528. ———. 2007. “Les utilisations de l’alphabet lors de IIe millénaire av. J.-C. et le développement de l’épigraphie alphabétique: une aproche à travers la documentation ougaritique en dehors des tablettes (I),” in J.-M. Michaud (ed.), Le royaume d’Ougarit de la Crète à l’Éuphrate. Nouveaux axes de recherche. Sherbrooke. 9–47.

Semitic Languages in Contact—Syntactic Changes in the Verbal System and in Verbal Complementation Tamar Zewi and Mikhal Oren 1 Introduction Morphology, as regards both derivation and inflection, is perceived as reflecting the backbone of any individual language. To a great extent, derivation patterns and inflection paradigms are the foundation supporting the structure of the language. Morphology is therefore the most conservative domain in the language, even though it too is susceptible to internal changes and external influences.1 As opposed to morphology, syntax is not perceived by the speakers as reflecting the fundamental essence of their language and is open to many influences. Contact between languages therefore leaves its mark more easily on the syntax, and foreign influences penetrate the syntax to a large extent unheeded. The different periods or stages in the history of a given tongue, as well as the influences it was subjected to due to its contact with other languages, can thus be clearly identified on the syntactic level. As is well known, contact between languages may occur in a variety of situations and for a variety of reasons. Languages are put in contact, for example, in the event of emigration of individuals or communities; in the case that the use of a language other than the mother tongue is necessary for communication purposes; in the course of a translator’s work; and so forth. The contact between languages may lead to the adoption of syntactic principles and patterns from a foreign language into the speakers’ native tongue, or, conversely, to the transference of the native language’s syntax into the second language. Influence frequently occurs in both directions simultaneously.

* Sections 1, 2.1, 3.1, and 4 in this article are essentially the revised version of an article published by T. Zewi in Hebrew (Zewi 2002–2003); section 2.2 and part of section 3.2 are based on material taken from M. Oren’s doctoral dissertation, written under the supervision of T. Zewi (Oren 2012). Work on this dissertation was made possible by the award of a Rotenstreich Fellowship for Outstanding Doctoral Students in the Humanities from the Israeli Council for Higher Education, as well as a grant from the Memorial Foundation for Jewish Culture. 1  See, for instance, Weinreich 1967: 31–37, Goldenberg 1996: 158–165.

© koninklijke brill nv, leiden, ���5 | doi ��.��63/9789004300156_021

406

Zewi and Oren

In the following article we shall focus on the latter tendency, i.e., the influence of the native language’s syntactic principles and patterns on the syntax of a second language. We shall discuss two situations in which the speakers of one Semitic language wrote, for certain purposes, in another Semitic language; in both cases the writers’ first language (or main language) had a deep impact on the language in which they wrote, the second language. We shall examine the effects of this influence both on the verb itself, in the verb’s inflection and the use of verbal tenses and modes, and on the verb’s complementation in the verb phrase—two domains that are among the most easily accessible to foreign influence. The first case we shall investigate (section 2) is that of the Akkadian used in the letters written by Canaanite vassal rulers to the Egyptian Pharaoh in the 14th century BCE, which were found in the archive at Tell el-Amarna. The second case (section 3) concerns so-called ‘Arabicised’ Hebrew, namely the Hebrew used in medieval times in Arabic-speaking milieux for writing both original compositions and translations from Arabic. 2

Amarna Akkadian

The influence of the native dialect(s) of the 14th-century Canaanite scribes on the Akkadian they used as lingua franca in official international correspondence can be discerned on various levels. Section 2.1 presents the Canaanite features that may be identified in the verb itself, in both its morphology (the verb’s inflected forms) and its syntax (the temporal, aspectual, and modal values associated with these forms). Section 2.2 deals with changes affecting verb+preposition combinations. 2.1 Changes in the Verbal System Significant alterations of the Akkadian verbal system in the Amarna texts originating from Canaan and its surroundings were first noted by Ebeling.2 His findings were later confirmed and supplemented by other researchers, chief among whom was Moran, who wrote his 1950 dissertation on the syntax of the Akkadian dialect found in the Amarna letters from Byblos (in Phoenicia, present-day Lebanon).3 Moran was the first to show how the Canaanite suffix conjugation (in his terms the ‘perfect’, i.e., the qatala/qatila/qatula forms of the Qal stem and their equivalents in other stems) and prefix conjugation 2  Ebeling 1910. 3  Moran 1950 (= Moran 2003: 1–130).

changes in the verbal system & in verbal complementation

407

(the ‘imperfect’, i.e., the yaqtulu/yiqtalu forms in Qal and their equivalents in other stems) were employed with various temporal and aspectual values reflecting these forms’ uses in Canaanite,4 and displacing in part the Akkadian iprus, iptarras, and iparras conjugations.5 Most of Moran’s conclusions were corroborated by Rainey, who demonstrated they applied not only to the Byblos correspondence but to letters from other Canaanite sources as well; Rainey also corrected several mistakes in Moran’s work.6 Our discussion of the modifications of the Akkadian verbal system in the Amarna letters will be divided into three parts, dealing with the changes in the suffix conjugation (section 2.1.1), with the changes in the prefix conjugation (section 2.1.2) and with the energic, a morpheme that is affixed to verbal forms (mostly belonging to the prefix conjugation) and that has a special syntactic role (section 2.1.3). 2.1.1 Suffix Conjugation Akkadian possesses a suffix conjugation, the so-called ‘stative’ or ‘permansive’ forms, very similar in morphology to the Canaanite suffix conjugation forms: the main difference is the suffix for the 1st person singular, which is -āku in Akkadian and -ti in Canaanite. Clear evidence of the Canaanite influence on the suffix conjugation forms in the Amarna letters may thus be seen in the fact, noted by Rainey, that out of the approximately 120 occurrences of 1st person singular suffixed forms in the Amarna letters from Canaan, only twenty-eight use the Akkadian suffix -āku, while the rest use Canaanite -ti.7 Further to be noted is the widespread use, for transitive verbs, of the base qatal (with two a vowels), which is common in Canaanite but rare in Akkadian, where paris is preferred for transitive verbs and paras is reserved mainly for verbs denoting a state and for the passive.8 Apart from these morphological particularities, the suffix conjugation in the Amarna letters also displays unusual syntax. Whereas the stative in Akkadian functions essentially as a nominal system denoting states, in the Amarna letters the suffix conjugation is used to express a variety of temporal, aspectual,

4  Moran 1950: 24–56 (= 2003: 24–53), 1960 (= 2003: 179–195); cf. also Rainey 1996: 2.5–8. 5  Whereas the root q-t-l is used for denoting morphological patterns in most Semitic languages, for the patterns of Akkadian the root p-r-s is used. 6  Rainey 1975, 1986, 1990, 1991–1993, 1996: 2.195–264, 347–366. 7  Rainey 1996: 2.283–287. 8  Rainey 1996: 2.295–305.

408

Zewi and Oren

and modal values, many of which resemble those of the suffix conjugation in Biblical Hebrew.9 Suffix conjugation forms in the Amarna letters are used primarily with a past meaning (narrative past), mirroring one of the main uses of the equivalent forms in Biblical Hebrew.10 Thus, for instance, in example (1) it is clear that the form ša-mi-ti refers to the past, since in parallels in other letters we find verbs of the iprus or iptaras conjugations, denoting respectively the preterite and perfect in Akkadian, e.g., iš-te-me in EA 141: 8–9. This use of the suffix conjugation can be compared to its use in Biblical Hebrew in examples such as (2). (1) EA 362: 5–6 (Byblos) a-nu-ma ša-mi-ti7 a-wa-teMEŠ LUGAL EN-ia ‘Now I have heard the words of the king, my lord’11 (2) Deut 5:28 ‫שׁמ ְע ִתּ י ֶאת־קֹול ִדּ ְב ֵרי ָה ָעם ַהזֶּ ה ֲא ֶשׁר ִדּ ְבּרוּ ֵא ֶליָך‬ ַ ָ  ‫אמר יְ הֺוָ ה ֵא ַלי‬ ֶ ֹ ‫וַ יּ‬ ‘And the Lord said to me, “I have heard the plea that this people made to you”.’12 Suffix conjugation forms in the Amarna letters from Canaan are also used with a present import. Since this usage is found especially with verbs describing a state or denoting the passive, it is likely that it reflects the uses of the Akkadian stative.13 Suffix conjugation forms may also refer to the future, usually in the protases and apodoses of conditional clauses,14 and occasionally also in optative expressions.15 An example of the use of the suffix conjugation in the apodosis of a conditional clause may be found in (3), both in the Akkadian verb ḫal-qa-at and in its Canaanite gloss a-ba-da-at (an instance of the Canaanite scribes’ practice of occasionally adding words in their own language beside Akkadian words to clarify their meaning). This conditional use of the suffix 9   For the substance of Rainey’s discussion of the syntax of the suffix conjugation, see Rainey 1973; 1996: 2.347–366. 10  Rainey 1996: 2.348–352. 11  Taken from Rainey 1996: 2.348. 12  Translations of the biblical text follow that of the JPS (Jewish Publication Society), with minor changes. 13  Rainey 1996: 2.352–355. 14  Rainey 1996: 2.355–364. 15  Rainey 1996: 2.364–365.

changes in the verbal system & in verbal complementation

409

conjugation is paralleled in Biblical Hebrew, for instance in (4), where suffix conjugation forms are used in both protasis and apodosis. (3) EA 288: 51–53 (Jerusalem) [ù] šum-ma ia-a-nu-mi ERÌN.MEŠ pi-ṭa-tu4 i-na MU an-ni-ti ḫal-qa-at a-ba-da-at \ gáb-bi KUR.ḪÁ šàr-ri EN-ia ‘But if there are no regular troops this year, all the lands of the king, my lord, are lost’16 (4) Josh 23:16 ‫ֹלהים ֲא ֵח ִרים‬ ִ ‫יכם ֲא ֶשׁר ִצוָּ ה ֶא ְת ֶכם וַ הֲלַ ְכ ֶ תּם וַ עֲ בַ ְד ֶ תּם ֱא‬ ֶ ‫ֹלה‬ ֵ ‫ת־בּ ִרית יְ הוָֹ ה ֱא‬ ְ ‫ְבּ ָע ְב ְר ֶכם ֶא‬

‫טֹּובה ֲא ֶשׁר נָ ַתן ָל ֶכם‬ ָ ‫יתם ָל ֶהם וְ ָח ָרה ַאף־יְ הוָֹ ה ָבּ ֶכם וַ אֲבַ ְד ֶ תּם ְמ ֵה ָרה ֵמ ַעל ָה ָא ֶרץ ַה‬ ֶ ִ‫וְ ִה  ְ שׁ  ַתּחֲו‬

‘If you break the covenant that the Lord your God enjoined upon you, and go and serve other gods and bow down to them, then the Lord’s anger will burn against you, and you shall quickly perish from the good land that He has given you.’

2.1.2 Prefix Conjugation As with the suffix conjugation, the effects of Canaanite influence on the Akkadian prefix conjugation can be discerned in both morphology and syntax. Morphologically, for the 3rd person masculine singular the Akkadian prefix i- is replaced by the Canaanite ya- (yaqtul(u) vs. iprus)17 and for the 3rd person masculine plural the prefix i- and suffix -ū are replaced by t- and -ū or -ūna (taqtulū/taqtulūna vs. iprusū), as is also found in Ugaritic.18 Syntactically, Canaanite influence is clearly seen in the temporal, aspectual, and modal values accorded to the prefix conjugation.19 In West Semitic languages, the prefix conjugation presents three variants distinguished by their suffixes: yaqtulu denotes the imperfect in the indicative mood, yaqtula the volitive, and yaqtul (with a Ø-morpheme suffix) the jussive as well as the indicative preterite. The three variants are found in Biblical Hebrew and in 16  Taken from Rainey 1996: 2.358. 17  Rainey 1996: 2.34–37. 18  Moran 1951 (= 2003: 159–164), Rainey 1996: 2.43–45. The Canaanite prefix conjugation, which denotes the imperfect, is compared here to the Akkadian iprus conjugation marking the preterite, rather than to the iparras conjugation marking the present and future, as it is morphologically closer to the former. Regarding the forms in Ugaritic, see, e.g., Sivan 1997: 111–112, 115–120. 19  On the syntax of the prefix conjugation, see Rainey 1975; 1986; 1990; 1991–1993; 1996: 2.195–264; 2003.

410

Zewi and Oren

Ugaritic, and the Amarna letters provide evidence of the existence of at least yaqtulu and yaqtul forms, and possibly of yaqtula forms, in Canaanite.20 The yaqtulu (indicative imperfect) forms found in the Amarna letters may refer to actions in the present or future, as for example in (5), in which context they replace the iparras forms normally used in Akkadian for the expression of the present/future. They may also refer to habitual or continuous actions in the past, as in (6). (5) EA 142: 16–17 (Beirut) i-na-ṣa-ru-šu a-di yi-im-lu-ku LUGAL ‘I am guarding him until the king takes counsel.’21 (6) EA 85: 36–37 (Byblos) ša-a yu-da-nu pa-na-nu i-na URUṢu-mu-ra [y]u-da-na7 i-na-na i-na URUGub-la ‘What used to be sold formerly in Ṣumur, let it be sold now in Byblos.’22 Similar uses of the prefix conjugation are found in Biblical Hebrew. In example (7), just as in (5) above, the prefix conjugation verb form, referring to a future situation, follows the conjunction ‘until’ (Akkadian adi, Hebrew ʿaḏ). In example (8), as in (6), the prefix conjugation verb form refers to a habitual action in the past, indicated by the adverb ‘formerly’ (Hebrew ləp̄ ānīm, Akkadian pānānu). (7) Gen 38:11



‫ית־א ִביְך ַעד־יִ גְ  ַדּל ֵשׁ ָלה ְבנִ י‬ ָ ‫ְשׁ ִבי ַא ְל ָמנָ ה ֵב‬

‘Stay as a widow in your father’s house until my son Shelah grows up.’

20  The suffixes of the West Semitic yaqtulu and yaqtula forms being formally identical to the Akkadian subjunctive suffix -u and ventive suffix -a(m), the distinction between them is not always clear. The possibility of interpreting -a as the West Semitic volitive rather than as the Akkadian ventive was first suggested by Moran (1960 = 2003: 179–195); however, Rainey, who originally supported this hypothesis, later showed that most of the -a suffixes in the Amarna letters are best interpreted as Akkadian ventives. Cf. Rainey 1991–1993, 1996: 2.202–211, 254–263. 21  Taken from Rainey 1996: 2.231. 22  Taken from Rainey 1996: 2.233.

changes in the verbal system & in verbal complementation

(8) 1Sam 9:9



411

‫ִכּי ַלנָּ ִביא ַהיֹּום יִ  ָקּ ֵרא ְל ָפנִ ים ָהר ֶֹאה‬

‘For the prophet of today was formerly called a seer.’

An example of the use of a yaqtul form for the jussive in the Amarna letters is given in (9), where the Canaanite jussive ia-az-ku-ur-mi parallels the Akkadian precative li-iḫ-šu-uš-mi to which it serves as a gloss. A comparable example from Biblical Hebrew is given in (10). (9) EA 228: 18–25 (Hazor) [ù] li-iḫ-šu-uš-mi \ ia-az-ku-ur-mi Išar-ri EN-ia mi-im-ma ša in4-né-puuš-mi UGU URUḪa-ṣú-raKI URU.KI-ka ù UGU ÌR-ka ‘So may the king, my lord, take thought for everything that has been done against Hazor, your city, and against your servant.’23 (10) 2Sam 14:11 ‫ֹלהיָך‬ ֶ ‫יִ זְ   ָכּר־נָ א ַה ֶמּ ֶלְך ֶאת־יְ הֺוָ ה ֱא‬ ‘Let Your Majesty be mindful of the Lord your God.’ 2.1.3 The Energic In the Amarna letters from Canaan a suffix -una or possibly -unna (with geminate n) is sometimes added to verbs, mainly of the prefix conjugation. This suffix has been recognised as an energic morpheme, probably meant to emphasise certain modal values of the verb. This can be seen from the syntactic contexts in which the suffix is found: mostly in questions (some of them rhetorical), but also in promises, admonishments, wishes, threats, protases of conditional clauses, and so on. In all these contexts the indicative is also frequently used without this suffix, the use of which appears to be optional.24 Energic -un(n)a is unknown in Akkadian outside El-Amarna, but formally similar morphemes are used for like purposes in Biblical Hebrew, Ugaritic, and Classical Arabic, and possibly also in Aramaic.25 One may therefore surmise that the use of the energic in the Amarna letters originates from West Semitic influence. An example of a question featuring a verb with the energic in an Amarna letter is given in (11). In Biblical Hebrew, the energic is found in the n of certain accusative suffixes, as for instance in -nnâ in example (12). 23  Taken from Rainey 1996: 2.245. 24  On the energic in the Amarna letters, see Moran 1950: 53–56 (= 2003: 50–53), Rainey 1996: 2.234–244 and Zewi 1999: 157–170. 25  Zewi 1999. Regarding Aramaic, see ibid. 7–8 n. 7.

412

Zewi and Oren

(11) EA 74: 63–64 (Byblos) mi-na i-pu-šu-na a-na-ku i-na [i]-de-ni-ia ‘What can I do by myself?’26 (12) Is 14:27 ‫וּמי יְ  ִ שׁיבֶ נָּה‬ ִ ‫וּמי יָ ֵפר וְ יָ דֹו ַהנְּ טוּיָ ה‬ ִ ‫ִכּי־יְ הֺוָ ה ְצ ָבאֹות יָ ָעץ‬ ‘For the Lord of Hosts has planned, Who then can foil it? It is His arm that is poised, And who can stay it?’ 2.2 Changes in Verbal Complementation Beyond the verb itself, the influence of one language on another can also be found in the verb’s complementation, i.e., in the elements with which the verb combines. A verb with a given meaning may necessitate certain complements and often imposes requirements on their grammatical form: some verbs govern a noun or nominal phrase in the accusative (direct object); others a prepositional phrase with a particular preposition; others yet a content clause. Certain verb+preposition combinations found in the Amarna letters are rarely, if ever, found in other Akkadian texts, but have equivalents in West Semitic languages. Such combinations thus bear witness to Canaanite (or more generally West Semitic) influence on Amarna Akkadian. An example of this, discussed elsewhere in detail,27 is the use of the phrase eli X ašāšu in two letters from Byblos (EA 83: 35, 122: 38–39). The verb ašāšu usually means ‘to be worried, distressed’ and is not normally used with a complement in Akkadian; but since this verb was glossed, in two other letters from Byblos (EA 82: 50–51, 93: 4–5), by naqṣupu, derived from the same West Semitic root as the Hebrew verb qāṣap̄ ‘to be angry’, it is likely that the scribe used ašāšu with the latter meaning and combined it with the preposition eli as a calque on the Canaanite equivalent of Hebrew qāṣap̄ ʿal ‘to be angry with’. Several similar cases will be examined below. damāqu ‘to be good’. Most of the occurrences of this verb in the Amarna letters are in the phrases ana/ana pāni/ina pāni X damāqu ‘to seem good, be agreeable to (someone)’, as in example (13); the verbs lamānu ‘to be bad’ and tarāṣu ‘to be correct’ are also found in such phrases. Since this type of phrases is not common in Akkadian, their use in the Amarna letters likely originates in West Semitic influence.28 They have a close parallel in Biblical Hebrew, where ṭōḇ ‘good’, raʿ ‘bad’, and yāšār ‘right’ are frequently combined with bəʿênê ‘in the 26  Taken from Rainey 1996: 2.236. 27  Oren 2013: 3–5. 28  Moran 1950: 17 (= 2003: 17); Rainey 1996: 3.50–51.

changes in the verbal system & in verbal complementation

413

eyes of’; a double example is given in (14). Incidentally, in late Biblical Hebrew the preposition bəʿênê ‘in the eyes of’ in these phrases is partially displaced by lip̄ nê ‘in front of’ (equivalent, word for word, to ina pāni!) and ʿal ‘on’, as may be seen in example (15). This is once again the result of the influence of the commonly spoken language on a language used only or mainly in writing—in this case, Aramaic influence on Hebrew in post-exilic times.29 (13) EA 64: 10–14 (South Canaan) ù yi-da-mi-iq i-na pa-ni LUGAL-ri EN-ia yu-wa-ši-ra 1 LÚ.GAL a-na na-ṣa-ri-ia ‘and may it seem good to the king, my lord, to send an official to protect me’ (14) Jer 40:4 ָ ‫ וְ ִאם־ ַרע  ְ בּ ֵעינ‬. . . ‫ֶיך ָלבֹוא ִא ִתּי ָב ֶבל בֹּא‬ ָ ‫ִאם־טוֹ ב  ְ בּ ֵעינ‬ ‫בֹוא־א ִתּי ָב ֶבל ֲח ָדל‬ ִ ‫ֶיך ָל‬ ‘If it seems good to you to go with me to Babylon, come . . . And if it seems bad to you to come with me to Babylon, you need not’ (15) Neh 2:5–6 ָ ‫יטב ַע ְב ְדָּך ְלפָ נ‬ ‫הוּדה‬ ָ ְ‫ֶיך ֲא ֶשׁר ִתּ ְשׁ ָל ֵחנִ י ֶאל־י‬ ַ ִ‫וָ א ַֹמר ַל ֶמּ ֶלְך ִאם־ ַעל־הַ  ֶמּלֶ ְך טוֹ ב וְ ִאם־י‬ ‫ִיטב לִ ְפנֵי־הַ  ֶמּלֶ ְך וַ יִּ ְשׁ ָל ֵחנִ י‬ ַ ‫ וַ יּ‬. . . ‘If it please the king, and if your servant has found favour with you, send me to Judah . . . So it was agreeable to the king to send me.’ palāḫu ‘to fear’. In ‘central’ (as opposed to peripheral) Akkadian, combinations of palāḫu ‘to fear’ with the preposition ištu pāni ‘from’ appear only in the first half of the first millennium BCE (Neo-Assyrian and Neo-Babylonian dialects).30 The presence of the phrase ištu pāni X palāḫu in the Amarna letters, several centuries earlier, therefore suggests West Semitic influence. Indeed, the phrase has an exact equivalent in Biblical Hebrew yārēʾ mippənê. The use of the phrase in Amarna Akkadian and Biblical Hebrew may be compared in examples (16) and (17) below. The preposition ištu pāni further appears in Amarna Akkadian with

29  The replacement of Hebrew by Aramaic as the language of everyday use in post-exilic Jewish communities provides another example of Semitic languages in contact, which we shall not discuss further here. On the Aramaic influence on Hebrew, see among others Tur-Sinai (Torczyner) 1950, Kutscher 1964, Hurvitz 1972, Oren 2013: 5–7. 30  Rainey 1996: 3.52–53 in fact attributes the emergence of the phrase ištu pāni X palāḫu in Neo-Assyrian and Neo-Babylonian to Aramaic influence.

414

Zewi and Oren

the verb na‌‌ʾbutu (N-stem form of abātu) ‘to flee’, mirroring Biblical Hebrew bāraḥ mippənê.31 (16) EA 153: 6–8 (Tyre) pal-ḫa-at gáb-bi KUR-ti iš-tu pa-ni ERÌN.MEŠ LUGAL EN-lí-ia ‘The entire land is afraid of the troops of the king, my lord.’ (17) 2Kgs 25:26 ‫שׂ ִ דּים‬  ְ  ַ‫ל־ה ָעם ִמ ָקּטֹן וְ ַעד־גָּ דֹול וְ ָשׂ ֵרי ַה ֲחיָ ִלים וַ יָּ בֹאוּ ִמ ְצ ָריִם ִכּי י ְָראוּ ִמ ְ פּנֵי כ‬ ָ ‫וַ יָּ ֻקמוּ ָכ‬ ‘And all the people, young and old, and the officers of the troops set out and went to Egypt because they were afraid of the Chaldeans.’ šemû ‘to hear’. In the Amarna letters this verb is relatively frequently accompanied by the preposition ana ‘to’;32 an example is given in (18). Since in Akkadian šemû almost always governs a direct object, the phrase ana X šemû in the Amarna letters most likely reflects the West Semitic usage evidenced in Biblical Hebrew šāmaʿ ʾel and šāmaʿ lə(qōl) (besides šāmaʿ bəqōl), in Old Aramaic šmʿ lql and in Ugaritic šmʿ l(gh) ‘to listen to, heed, obey (the voice of)’.33 An example from Biblical Hebrew is given in (19). (18) EA 290: 19–20 (Jerusalem) li-iš-me LUGAL-ri a-na IÌR-ḫi-ba ÌR-ka ù lu-ma-šar ERÌN.MEŠ pi-ṭa-ti ‘May the king listen to ʿAbdi-Ḫeba your servant and send troops.’ (19) 1Kgs 12:15 ‫יְתה ִס ָבּה ֵמ ִעם יְ הוָֹ ה‬ ָ ‫י־ה‬ ָ ‫וְ לֹא־  ָשׁ ַמע ַה ֶמּ ֶלְך אֶ ל־הָ ָעם ִכּ‬ ‘The king did not listen to the people; for the Lord had brought it about.’ 3

Arabicised Hebrew

The discussion that follows is concerned with linguistic features present in the Hebrew written by Arabic-speakers during the Middle Ages. Medieval Hebrew prose was often a translation of Arabic works, but also includes some original Hebrew texts. The Arabic influence on these writings, discernible in vari31  Rainey 1996: 3.52–53. 32  The verb šemu appears some 140 times in the Amarna letters, of which some 25 times in combination with ana. 33  The West Semitic origin of ana X šemû has been noted by Moran 1950: 17 (= 2003: 18) and Rainey 1996: 3.14.

changes in the verbal system & in verbal complementation

415

ous linguistic levels, gave rise to the term ‘Arabicised Hebrew’. As regards the verbal system, one innovation of Arabicised Hebrew is the introduction of a complete system of compound tenses based on those of Arabic.34 Clear Arabic influence on the Hebrew verbal system is further apparent in the use of prefix conjugation forms to express the future or present/future. Both these matters will be examined below in section 3.1. With regard to verb complementation, Arabicised Hebrew displays certain trends in the use of accusative complements that suggest Arabic influence. It also introduces new verb+preposition combinations calqued on Arabic. These issues will be described in section 3.2. 3.1 Changes in the Verbal System 3.1.1 Compound Tenses Arabicised Hebrew presents many instances of the use of a system of compound tenses as found in Arabic. This system relies primarily on the auxiliary verb hāyâ ‘to be’ and in some cases on the adverb kəḇār ‘already’.35 Examples of the full variety of compound tenses in Arabicised Hebrew from different sources were presented by Goshen-Gottstein in his 1951 doctoral dissertation.36 According to Goshen-Gottstein, the compound tenses include combinations of hāyâ or yihye (i.e., suffix or prefix conjugation forms of the verb ‘to be’) with prefix conjugation verb forms, with suffix conjugation verbs forms,37 or with the particles yēš asserting existence and ʾên negating it. These combinations serve to convey temporal and aspectual nuances, presenting actions as continuous or as completed at a past or future time, much like kāna faʿala, kāna yafʿulu, yakūnu faʿala, and yakūnu yafʿulu in Arabic. Additional compound tenses are created, according to Goshen-Gottstein, by means of the adverb kəḇār ‘already’: kəḇār is combined with suffix conjugation verb forms to denote perfected actions, in the same manner as Arabic qad faʿala. A Hebrew combination of the type hāyâ kəḇār ʿāśâ thus expresses the past perfect (‘had done’), mirroring Arabic kāna qad faʿala, while Hebrew yihye kəḇār ʿāśâ, like Arabic yakūnu qad faʿala, expresses the future perfect (‘will have done’). Hebrew combinations of the type kəḇār yaʿăśe or hāyâ kəḇār yaʿăśe, matching Arabic qad yafʿulu or kāna qad yafʿulu, do not indicate a point in time but rather convey modality: ‘may do’.

34  Goshen-Gottstein 2006: 59–64 (§§1–14). 35  Goshen-Gottstein 2006: 59 (§1). 36  A revised version of this dissertation has now been published as Goshen-Gottstein 2006. All references to Goshen-Gottstein’s dissertation are here based on this edition. 37  Goshen-Gottstein refers to the prefix and suffix conjugations as ‘future’ and ‘past’, respectively.

416

Zewi and Oren

In the excerpts of Sefer ha-Riqma, Yehuda Ibn Tibbon’s translation of Yonah Ibn Janāḥ’s Kitāb al-Lumaʿ, that we examined with the purpose of adding to Goshen-Gottstein’s material, the only compound forms we found were combinations of kəḇār with suffix conjugation verbs. A few examples are shown below:38 (20) Sefer ha-Riqma, p. 11 (author’s introduction) ‫אבל אנשי לשוננו בדורנו זה כבר השליכו החכמה הזאת אחרי גויהם‬ ‘But the men of our generation have thrown this wisdom behind their backs.’ (21) Sefer ha-Riqma, p. 17 (author’s introduction) ‫וכבר ראיתי ראש הישיבה רבינו סעדיה ז"ל נוהג המנהג הזה ברוב פירושיו‬ ‘And I have seen our master Rabbi Saadia do so in most of his interpretations.’ (22) Sefer ha-Riqma, p. 20 (author’s introduction) ‫ בשער יחדנוהו לו בחלק הזה הראשון‬,‫ מורה על זולתו‬,‫וכבר זכרנו אנחנו ממנו כלל‬

‫מחבורנו זה‬

‘And we ourselves have provided a record of (instances of) this, indicative of others, in the section we devoted to it in the first part of our work.’

(23) Sefer ha-Riqma, p. 35 ‫וכבר הוסיפו אנשים מבעלי הדקדוק על הענינים האלה ענינים אחרים‬ ‘And some among the grammarians have added other matters to these.’ 3.1.2 Use of the Prefix Conjugation with Future and Present Values The regular forms of the prefix conjugation (i.e., the indicative imperfect forms, as opposed to the shortened forms of the jussive and the lengthened forms of the cohortative) had several uses in Arabicised Hebrew. GoshenGottstein briefly notes the use of these forms in general, atemporal assertions, as well as their appearance in conditional clauses.39 It appears, however, that these forms frequently served also to indicate the future and present/future. Examples are given below: while yṣṭrk in (24), yrdwp and ystwm in (25) and yzhr in (26) have a general, atemporal value, ywbn in (25) refers to the future, and ʾmṣʾ, ʾmnʿ, and ʾzhr in (26) to the present/future.

38  Page references follow Wilensky’s edition (Wilensky 1928–1930). 39  Goshen-Gottstein 2006: 64 (§14), 243 (§287).

changes in the verbal system & in verbal complementation

417

(24) Sefer ha-Riqma, p. 5 (translator’s introduction) ‫ בעבור הדברים אשר זכרתים כבר‬,‫ואני ידעתי כי העבודה כבדה והמלאכה קשה‬

‫בפתיחת העתקת ספר חובות הלבבות אשר יצטרך אליהם המעתיק‬

‘And I knew that the work was burdensome and the task difficult, because of the things that I already mentioned in the introduction to the translation of Sefer Ḥovot ha-Levavot that the translator requires.’

(25) Sefer ha-Riqma, p. 5 (translator’s introduction) ‫ ולרשת‬,‫כי קוצר הלשון הנמצא בידנו דוחק את המעתיק להרחיב מעט את גבולו‬

‫ כי ראיתי אשר טוב למעתיק‬,‫ על אפניו ותכונתו‬,‫ כדי שיובן ענין מליצתו‬,‫משכנות לא לו‬ ‫ מאשר ירדוף אחרי המלות ויסתום העניינים‬,‫לדחוק עצמו מעט בבנינים‬

‘For the insufficiency of the language at our disposal forces the translator to slightly broaden his domain and take possession of foreign dwellings, so that the meaning of his expressions, in its particulars and its essence, be understood, for I have seen that it is better for the translator to force his way somewhat with constructions than to stick to the words and make matters unintelligible.’

(26) Sefer ha-Riqma, p. 17 (author’s introduction) ‫ומה שלא אמצא עליו עד ממה שזכרתיו ואמצא עליו העד מן הלשון הערבי לא אמנע‬

‫ כאשר יזהר מזה מי‬,‫להביא עד מן הגלוי בו ולא אזהר מקחת ראיה מהנראה ממנו‬ ‫שדעתו חלושה והכרתו מעטה מאנשי דורנו‬

‘And for the things I find no evidence for in the aforementioned and find evidence for in the Arabic tongue, I shall not refrain from providing evidence from what is apparent nor beware of taking a proof from what can be seen, as do those whose mind is weak and knowledge slight among the men of our time’.

Use of the prefix conjugation to indicate the future or present is, of course, already found in Biblical Hebrew, where the indicative imperfect had a number of uses combining temporal and aspectual values, as well as modal uses.40 However, in Mishnaic Hebrew these forms ceased to be associated with time and aspect and were used only with modal values.41 It is therefore likely that the future and present/future uses of the prefix conjugation in Arabicised Hebrew were not carried over from Biblical Hebrew but rather reflect the uses of the corresponding forms in Arabic.

40  See for example Kautzsch 1910: 313–319 (§107); Joüon and Muraoka 2006: 337–345 (§113). 41  Sharvit 1980: 110, 114.

418

Zewi and Oren

3.2 Changes in Verbal Complementation In his study of Arabicised Hebrew, Goshen-Gottstein devotes a chapter to the particularities related to the use of verbs with accusative complements. He mentions several phenomena which were not new in Hebrew in themselves, but whose increased occurrence in Arabicised Hebrew attests to Arabic influence: namely, the absence of the accusative marker ʾēṯ for direct objects and the use of predicative accusatives and of internal object accusatives.42 GoshenGottstein further touches upon the use of the preposition lə- as an accusative marker. While this phenomenon is known to be fairly frequent in Late Biblical Hebrew and in Mishnaic Hebrew, where it reflects the influence of Aramaic,43 the particular cases where accusative lə- is combined with participial verb forms should be attributed, according to Goshen-Gottstein, to Arabic rather than Aramaic influence.44 More importantly, Goshen-Gottstein’s study comprises a lexicon of the verbs which he found to reflect Arabic influence in their meaning, the forms of their complements, or both.45 Among his gleanings one may find, for instance, the Hebrew phrase šāmaʿ min, calqued on Arabic samaʿa min, replacing Classical Hebrew šāmaʿ ʾel ‘to heed, obey’ that we mentioned earlier as the model for ana X šemû in Amarna Akkadian (compare examples (18) and (19) in section 2.2 above); Hebrew qāraḇ min, mirroring Arabic qaruba min, replacing Classical Hebrew qāraḇ ʾel ‘to approach, be close to’; or Hebrew dāmâ bəreplacing Classical Hebrew dāmâ lə- and dāmâ ʾel ‘to resemble, be like’ as the translation of Arabic šabbaha bi-.46 In some cases, inter-linguistic influence may be favoured by intralinguistic factors. Goshen-Gottstein mentions the phrase ʾānap̄ ʿal (as opposed to Classical Hebrew ʾānap̄ bə-) employed in one text as the translation of Arabic g1 aḍiba ʿalā ‘to be angry with’.47 In combining the verb ʾānap̄ with the preposition ʿal, the writer may have been influenced not only by the Arabic phrase 42  Goshen-Gottstein 2006: 111–116 (§§96–101). On the corresponding structures in Biblical Hebrew, see, e.g., Kautzsch 1910: 362–376 (§§117–118), Joüon and Muraoka 2006: 410–430 (§§125–126). 43  The use of lə- as the accusative marker in Late Biblical Hebrew is discussed in greater detail in Oren 2013: 5–7. More generally concerning the Aramaic influence on Hebrew, see the other references given in note 29 above. 44  Goshen-Gottstein 2006: 188 (§202b). Note however that according to Kautzsch 1910: 366 (§117n), the use of lə- as the accusative marker in Biblical Hebrew is more common with participial verb forms. 45  Goshen-Gottstein 2006: 255–436. 46  For these particular examples, see Goshen-Gottstein 2006: 283, 400, 422. 47  Goshen-Gottstein 2006: 262.

changes in the verbal system & in verbal complementation

419

he was translating but also by Hebrew phrases such as kāʿas ʿal or qāṣap̄ ʿal, where the verb, though identical in meaning to ʾānap̄ , governs the preposition ʿal rather than bə-. An enlightening comparison can be made to ḥārâ ʾappô ‘his wrath was kindled’, a locution related to ʾānap̄ in both meaning and etymology: in Biblical Hebrew alone, it is found over forty times combined with the preposition bə- but also three times with ʿal.48 Here no foreign language influence is at play, but only the analogy to such phrases as qāṣap̄ ʿal ‘to be angry with’ or šāp̄ aḵ ḥămāṯô ʿal ‘to pour one’s wrath on’ or, more broadly, the general adversative nuance often carried by the preposition ʿal.49 4 Conclusion The discussion above was concerned with changes occurring in the verbal system and in verbal complementation as a result of contact between languages. The case of the Akkadian of the Amarna letters from Canaan (section 2) and that of Arabicised Hebrew (section 3) are far removed from one another in time and place, and the changes evidenced in these languages are also very different. What these cases have in common is the facility with which linguistic practices from the speakers’ (or in these cases, the writers’) main language penetrate the second language they must use for certain purposes. In the use of verbal tenses and modes and in the complementation of the verb, elements that are characteristic of the first language or mother tongue spread almost unheeded into the second language. Verbal complementation is a syntactic issue, as is, at least in traditional grammar, the system of verbal tenses and modes. The ease with which significant changes take place in these domains demonstrates the ease with which syntax as a whole may be affected through contact between languages. References Ebeling, E. 1910. “Das Verbum der El-Amarna Briefe,” Beiträge zur Assyriologie und semitischen Sprachwissenschaft 8: 39–79.

48  Zech 10:3 and, with minor variations, Num 24:10 (ḥārâ ʾappô ʾel: on ʾel used for ʿal, see, e.g., Joüon and Muraoka 2006: 456 [§133b]) and Job 19:11 (heḥĕrâ ʾappô ʿal with a Hip̄ ʿīl verb form). 49  Cf. Kautzsch 1910: 383–384 (§119dd), Joüon and Muraoka 2006: 460–461 (§133f).

420

Zewi and Oren

Goldenberg, G. 1996. “Hebrew as a Living Semitic Language,” in J. Blau (ed.), Evolution and Renewal: Trends in the Development of the Hebrew Language. Jerusalem. 148–190. (in Hebrew) Goshen-Gottstein, M. H. (revised by Sh. Assif and U. Melammed). 2006. Syntax and Vocabulary of Mediaeval Hebrew as Influenced by Arabic. Jerusalem. Hurvitz, A. 1972. The Transition Period in Biblical Hebrew. Jerusalem. (in Hebrew) Joüon, P. and T. Muraoka. 2006. A Grammar of Biblical Hebrew (2nd ed.). Rome. Kautzsch, E. (tran. A. E. Cowley). 1910. Gesenius’ Hebrew Grammar (2nd ed.). Oxford. Kutscher, E. Y. 1964. “Aramaic Calque in Hebrew,” Tarbiz 33: 118–30. (in Hebrew) Moran, W. L. 1950. A Syntactical Study of the Dialect of Byblos as Reflected in the Amarna Tablets. Ph.D. Diss., John Hopkins University. (= Moran 2003: 1–130) ———. 1951. “New Evidence on Canaanite taqtulū(na),” JCS 5: 33–35. (= Moran 2003: 159–164) ———. 1960. “Early Canaanite yaqtula,” Orientalia 29: 1–9. (= Moran 2003: 179–195) ———. (ed. J. Huehnergard and Sh. Izre’el). 2003. Amarna Studies: Collected Writings. Winona Lake. Oren, M. 2012. Verbal Rection in the Northwest Semitic Languages in the Second Half of the Second Millennium and the First Half of the First Millennium BCE. Ph.D. Diss., University of Haifa. (in Hebrew) ———. 2013. “Interference in Ancient Languages as Evidenced by Governed Prepositions,” JSS 58: 1–11. Rainey, A. F. 1973. “Reflections on the Suffix Conjugation in West Semitized Amarna Tablets,” UF 5: 236–262. ———. 1975. “Morphology and the Prefix Tenses of West Semitized El-ʿAmarna Tablets,” UF 7: 395–426. ———. 1986. “The Ancient Hebrew prefix Conjugation in the Light of Amarna Canaanite,” Hebrew Studies 27: 4–19. ———. 1990. “The Prefix Conjugation Patterns of Early Northwest Semitic,” in Tz. Abusch et al. (eds.), Lingering over Words: Studies in Ancient Near Eastern Literature in Honor of William L. Moran. Atlanta. 407–420. ———. 1991–1993. “Is There Really a yaqtula Conjugation Pattern in the Canaanite Amarna Tablets?,” JCS 43–45: 107–117. ———. 1996. Canaanite in the Amarna Tablets, I–IV. Leiden. ———. 2003. “The Yaqtul Preterite in Northwest Semitic,” in M. F. J. Baasten and W. Th. Van Peursen (eds.), Hamlet on a Hill: Semitic and Greek Studies Presented to Professor T. Muraoka on the Occasion of his Sixty-Fifth Birthday (OLA 118). Leuven. 395–407. Sharvit, Sh. 1980. “The ‘Tense’ System of Mishnaic Hebrew,” in G. B. Sarfatti et al. (eds.), Studies in Hebrew and Semitic Languages. Ramat-Gan. 110–125. (in Hebrew) Sivan, D. 1997. A Grammar of the Ugaritic Language. Leiden.

changes in the verbal system & in verbal complementation

421

Tur-Sinai (Torczyner), N. H. 1950. “The influence of Aramaic on Hebrew,” in Encyclo­ paedia Biblica. Jerusalem. 1.593–5. (in Hebrew) Weinreich, U. 1967. Languages in Contact. The Hague. Wilensky, M. 1928–1930. Sefer ha-Riqma (Kitāb al-Lumaʿ) ab Jona ibn Ğanāḥ, Hebraice vertit Jehuda ibn Tibbon, I–II. Berlin. Zewi, T. 1999. A Syntactical Study of Verbal Forms Affixed by -n(n) Endings in Classical Arabic, Biblical Hebrew, El-Amarna Akkadian and Ugaritic (AOAT 260). Münster. ———. 2002–2003. “Semitic Languages in Contact—Syntactic Changes in the Verbal System,” Haivrit Weaḥyoteha 2–3: 243–253. (in Hebrew)

Index adaptation 57, 131, 133, 146–147, 170, 173–174, 270, 282, 354, 356 adverbializer 33–49 ʿAfar 18–20, 22, 188, 350n17 Agaw 18–30, 348 Aikhenvald, Alexandra 69 Akkadian 35–36, 68, 98n3, 114–116, 118, 123, 129, 139–151, 158–175, 181–183, 191–194, 267, 269, 276, 297, 348–353, 375–396, 406–414, 418, 419 alloglottography 159–160, 171 Amarna Akkadian See Canaanite Amharic See Ethiopian Semitic Amiritic 11–12 Amorite 162–165, 182, 375–376, 385n49 Ancient North Arabian 1–13, 15, 139, 148. See also Dadantic, Dumaitic, Hasaitic, Hismaic, Liḥyanite, Safaitic, and Taymanitic Ancient South Arabian See Old South Arabian apocopate 116–118, 123 Arabic 2–6, 9, 12–13, 57–58, 105n7, 114, 120, 123, 126–127, 134–135, 138–140, 142, 145, 147–148, 151, 181, 183–187, 190–194, 280–310, 312, 324, 334, 349–351, 353, 355, 358–360, 375, 406, 411, 414–419. See also Maltese Anatolian 234–248 Andalusi 336 Cypriot 224n17, 287 Egyptian 283, 291n53, 294 Hispanic 336 Iraqi 294, 297–298 Jordanian 212, 221 Lebanese 220n16, 284, 296n85, 306n133 Libyan 338 Maghrebi 334 Middle Judaeo 294, 298, 303, 306 Moroccan 221, 225, 227, 334–336, 338 Najdi 134 Nigerian 336, 342, 344–345 Palestinian 198–229, 280–310 Qalamūn 299n108 Samaritan 303

Sbaʿa Bedouin 134, 148 Tunisian 225, 333–345 Aramaic 4, 12, 40–49, 52–68, 80–82, 89, 93, 97–100, 107–112, 114, 116–117, 119, 121, 123, 126–151, 181–186, 190, 192–194, 213, 234, 241–242, 244, 251–252, 258–264, 280–310, 348–363, 395–396, 411, 413, 418. See also Neo-Aramaic Achaemenid (Imperial / Official) 44, 126, 130–132, 134–136, 140–145, 147–148, 234, 351–352 Biblical 49n41, 130, 234, 238, 251 Christian Palestinian Aramaic 129–130, 133–134, 288, 307 Hatran 126–151, 354 Jewish Babylonian Aramaic 54–59, 61–64, 130–135, 138–145, 147–148, 234, 251, 299, 308 Jewish Palestinian Aramaic 60n15, 129–131, 135, 137–138, 142–145, 190, 251, 304–305, 307 Judean Aramaic 126–127, 133, 135, 142–145, 148 Mandaic 132–136, 139–145, 148, 235, 304, 354, 356, 358 Nabataean 3–4, 8, 109, 126–128, 139–140, 143, 146, 148, 151 Old 49n41, 144, 237–238, 241, 351–352, 414 Palmyrene 126–129, 131–133, 136, 138–143, 145–148, 151 Qumran Aramaic 41, 126n4, 127, 142, 144–145  Samaritan 143, 292n55, 303–304, 307 Syriac 35, 41n16, 45, 47–48, 116–117, 119, 126–148, 151, 184, 189–190, 194, 234, 238–239, 284–285, 289, 291–293, 295, 298–299, 304, 308, 249–360 Targum Onkelos 41, 59–60, 66–67, 142–143, 145 Argobba See Ethiopian Semitic Armenian 132, 135, 236n4, 349–350, 355–356 assimilation 2, 73, 104n5, 242, 244n12, 269–270 Assyrian See Akkadian Awngi 20, 23–24, 26, 29–30

424 Aymellel See Kestane under Ethiopian Semitic Babylonian See Akkadian Barth-Ginsberg Law 384 Barwar See Neo-Aramaic Baṭḥari 311, 313–321, 327–329, 331 Baxʿa See Neo-Aramaic begadkepat See spirantization Beja 19–23 Betanure See Neo-Aramaic Bilin 18n2, 20–21, 23n5, 25–26, 29, 187–188 bilingualism 23n5, 43, 128, 158, 161, 198, 208, 212, 215, 225, 228, 243, 268, 277, 281, 312, 337 Burji 22 Čāl See Neo-Aramaic calque 4, 17, 42, 98, 106–107, 112, 135, 137n43, 145, 147, 150, 174, 185, 225, 412, 415, 418 Campbell, Lyle 203 Canaanite 143, 182, 282, 287, 290, 297, 304, 309–310, 352, 375–376, 388–394, 406–414, 418. See also Hebrew, Phoenician, Punic Canaano-Akkadian See Canaanite Chaha (Čaha) See Ethiopian Semitic Christian Palestinian Aramaic See Aramaic cleft 24 code-switching 169, 226–228, 335–338, 341, 343–344, 389n66 compensatory lengthening 202, 205–207 converb 16, 25–30 Coptic 99n4, 104n5, 106, 283, 349, 354–356 copula 24, 27–28, 118–119, 124, 170, 173 creole 392–393 Cushitic 16–31, 69, 74n6, 187–188, 348, 350n17. See also ʿAfar, Agaw, Awngi, Beja, Bilin, Burji, Hadiyya, Kemant, Konso, Oromo, Saho, Sidaama, Somali, Xamtanga Cypro-Minoan 267, 382 Dadantic 1n2, 3, 7, 10–11 de-gemination 202–203, 224 diffusion 69–70, 86–88, 93, 127, 202 diglossia 337, 381, 392–393

Index Dixon, Robert M. W. 69 Dobbi See Ethiopian Semitic doublet, lexical 131, 181–194, 292, 298 drift 69, 87–88, 92. See also parallel development Dumaitic 5, 8 Eblaite 377 Egyptian 97–112, 182–183, 267, 349–350, 382. See also Coptic Elamite 162–163 elative 5 Emarite 376–382, 387, 394 enclitic m 2, 272–276 Endegen (Endegañ) See Ethiopian Semitic Ener See Ethiopian Semitic energic ending 4, 389n71, 407, 411–412 Ennemor See Ethiopian Semitic ergative 163n8, 269 Ethiopian Semitic 16–31, 68–93, 114, 123, 187–189, 201n8, 202n10, 348–363 Amharic 16–31, 69–73, 75–76, 78–79, 82–83, 85–88, 91, 93, 181, 187–189, 333, 350–351, 355, 363 Argobba 20–21, 26–28, 70–73, 75–79, 81–85, 87–92, 350n17 Chaha (Čaha) 21, 76–79, 82, 87 Dobbi 76–79, 88, 89 Endegen (Endegañ) 24n6, 76–79 Ener 24n6 Ennemor 76, 78–79, 82, 87 Gafat 25, 72, 76–79, 82, 87 Gǝʿǝz 16–31, 68, 70, 72, 187–188, 348–363 Gurage 17, 19, 21–22, 24n6, 26–28, 70, 72–73, 76–77, 81, 82–84, 87–89 Gyeto 24n6 Harari 19–22, 24, 27–28, 70, 72–73, 75–76, 78–80, 82, 84n23, 86–87, 348 Inor 21, 24n6 Kestane 76–80, 88–89, 350n17 Masqan 20, 76, 78–79 Muher 20, 70, 72, 76–79, 88–89 Selṭi 24n6, 27–28, 70, 73, 75–76, 78–79, 85, 87–88 Tigre (Tǝgrē) 16, 18n2, 20–25, 27, 29–30, 68–69, 71, 81, 89, 93, 349n13–14

Index Tigrinya (Tǝgrǝñña) 16, 18, 20–23, 26–30, 68, 70, 82n22, 84n24, 349n13–14, 355–356 Wolane 27–28, 69–70, 73n4, 75–77, 79–80, 87–88 Zay 27–28, 69–70, 72–73, 75–80, 82, 84, 86–88 Ethiopic, Classical see Gǝʿǝz under Ethiopian Semitic ezafe 120–121 fricativization 23. See also spirantization Gafat See Ethiopian Semitic Gǝʿǝz See Ethiopian Semitic Ǧenēt See Neo-Aramaic genitive See possession, attributive gerund See converb gerundive See converb GoldVarb 206n12 grammaticalization 36n7, 43n24, 86n25, 173, 341 Greek 126–138, 145–147, 183–185, 281, 348–349, 354, 358–361 Ǧubbʿadīn See Neo-Aramaic Gurage See Ethiopian Semitic Gyeto See Ethiopian Semitic Hadiyya 22, 26 Ḥaḍramitic 11 Hakkari See Neo-Aramaic Harari See Ethiopian Semitic Harbol See Neo-Aramaic Ḥarsusi 311–331 Hasaitic 139 Hassana See Neo-Aramaic Hatran See Aramaic Hebrew 2–3, 6, 20, 33–49, 52–67, 108–109, 114, 123, 133, 143, 186, 190, 192, 201–202, 219–221, 223–224, 251, 269, 290–291, 293, 303–304, 310, 351–361 Biblical 33–49, 57–58, 63n18, 107, 182–183, 214–215, 219–221, 224, 282, 291–292, 294, 298, 302–303, 307–309, 349–350, 352, 408–414 Early Modern 251–264 Medieval 183, 414–419

425 Modern 183–184, 189–191, 198–229, 251, 280, 282, 300, 333 Post-Biblical 39–40, 45, 47, 52–67, 280, 298 Heine, Bernd 42–43, 333 Hertevin See Neo-Aramaic Herzog, Marvin I. 214  Hickey, Raymond vii, 333 Hismaic 1n2, 3–4, 7, 10 Hittite 267–268, 274, 375, 377–379, 382, 395 Hobyot 311–331 Hurrian 162–163, 182, 267–277, 375, 377–378, 382–385, 387n60, 395–396 hypercorrection 103, 112 Indo-European 114, 120, 216, 252. See also Armenian, Greek, Hittite, Iranian, Kurdish, Latin, Luwian, Pali, Polish, Russian, Sanskrit, Ukranian, Yiddish inheritance 86–88 innovation 1, 69–70, 75, 80, 82, 85–87, 89–90, 92, 312, 316–317, 319–320, 323, 325–326, 331, 415 Inor See Ethiopian Semitic Iranian 120, 127–138, 145–147, 150, 184. See also Median, Middle Persian, Old Persian, Parthian, Sogdian Jewish Babylonian Aramaic See Aramaic Jewish Palestinian Aramaic See Aramaic Jewish Zakho See Neo-Aramaic Jibbali 311–331 Johanson, Lars 160, 172, 376n3, 392n87 Jubb‘adīn See Neo-Aramaic Kassite 376 Kaufman, Terrence vii, 42, 173, 198, 222–223, 282n13 Kemant 20–21, 25, 29–30, 188 Kestane See Ethiopian Semitic koiné 392–393 Konso 22 Kroskrity, Paul 160–161, 167 Kulturwort 181 Kurdish 114–124, 193, 234–248 Kurmanjî See Kurdish Kuteva, Tania 42–43

426 Labov, William 201n8, 213–214, 216–217, 228 language area (linguistic area / Sprachbund) 1n3, 17, 87, 158, 161, 175,  234n1, 247–248 Latin 126–129, 131, 133, 135, 137, 145–147, 150, 191, 238n6, 348, 352 Levey, Stephen 335 lexical borrowing See loanword Liḥyanite 139 lingua franca 31, 44, 48, 69, 267, 375–376, 406 linguistic area See language area loan-translation 272 loanword (lexical borrowing) 16–22, 104–106, 112, 126–151, 159n2, 181–194, 234, 242–248, 251–264, 268–269, 275–276, 282–285, 289–290, 293, 295, 300, 303, 308n154, 335–336, 338, 341–342, 345, 348–363, 385n49 Luwian 267, 381 Maltese 207, 225–226, 229, 335–336 Maʿlūla See Neo-Aramaic  Mandaic See Aramaic Masqan See Ethiopian Semitic Matras, Yaron 43–44, 121, 151n79, 159, 252 Matrix Language Frame Model See Myers-Scotton, Carol Median 132, 134, 147 Meechan, Marjory 342 Mehri 311–331 merger 4, 11–12, 23, 26, 71, 80–81, 99, 103, 201–202, 214–219, 221–226, 228, 237–239, 245, 292. See also paradigmatic merger metathesis 159n2, 314 metonymy 144 Middle Persian 142n50, 147, 349–350, 355 Midyat See Neo-Aramaic Milroy, James 161n5, 214 minimal pair 189–191, 228, 239 mixed-language 392–393 Modern South Arabian 311–331. See also Baṭḥari, Ḥarsusi, Hobyot, Jibbali, Mehri, Soqoṭri Muher See Ethiopian Semitic multiple exponence 83–93 Myers-Scotton, Carol 227

Index Nabataean See Aramaic Nagy, Naomi Gail 198, 204 Neo-Aramaic 114–124, 234–248 Barwar 117, 299 Baxʿa 234, 281 Betanure 244 Čāl 244 Ǧenēt 235 Ǧubbʿadīn (Jubb‘adīn) 234–235, 281, 288, 299 Hakkari 235 Harbol 235 Hassana 235 Hertevin 235, 245, 308 Iranian Azerbaijan (Christian) 241 Iranian Kurdistan (Jewish) 241 Iraqi Kurdistan (Christian) 241 Iraqi Kurdistan (Jewish) 241 Maʿlūla 234–235, 281, 288, 296, 300, 308 Midyat 235, 242–243 Neo-Mandaic 235 Qaraqosh 295, 308 Telkepe 308 Ṭuroyo 192–194, 234–248, 299 Umra 235 Zakho (Jewish) 114–124, 241, 299 Neo-Mandaic See Neo-Aramaic Nevalainen, Terttu 214 Nilo-Saharan languages 17 nisba adjective 135, 137, 141, 147, 187 Ohala, John 222 Old Persian 135, 147 Old South Arabian (Ancient South Arabian) 2, 139, 148, 377n10. See also Amiritic, Ḥaḍramitic, and Sabaic Omotic 17 Oromo 17, 22, 31, 188 Pali 349 Palmyrene See Aramaic paradigmatic dissimilation 81–93 paradigmatic merger 81–93 parallel development 17, 69–70, 74, 78n15, 83, 86–89. See also drift Parthian 128, 131–137, 147, 149–150, 354–355 pattern replication 43–44, 119, 121, 123

427

Index Persian See Iranian Phoenician 3, 97–112 pidgin 392–393 pivot 121 Polish 252, 262 Poplack, Shana 226–227, 335, 342 possession, attributive 33–37, 116, 224–226, 270–272, 333–345 Praat 206 Punic 3, 97–112 Qaraqosh See Neo-Aramaic Rbrul 206, 210 retention 22, 114, 312, 317, 322–323, 325 Romaine, Suzanne 159n3, 214 Russian 217, 252, 256n4, 261–262 Sabaic 3, 5–7, 11, 13, 139, 148, 348, 352–353, 358–360, 362–363 Safaitic 1–4, 7, 9–10, 140 Saho 18–20, 22, 88, 350n17 Sakel, Jeanette 43–44, 121 Sankoff, Gillian 204 Sanskrit 349 Schuchardt, Hugo vii n1 Selṭi See Ethiopian Semitic shift, language 42–43, 281–283 Sidaama 22, 26 Siltʾi See Selṭi under Ethiopian Semitic Silva-Corvalán, Carmen 335 Sogdian 355 Somali 20, 22–23, 188 Soqoṭri 71n2, 311–331 spirantization 4n5, 145, 185, 189, 219, 238–239, 242, 244, 246, 291–292, 356. See also fricativization Sprachbund See language area

substrate 185, 280–310, 290, 392 Sumerian 140, 142, 144–145, 148, 158–175, 181–182, 267, 351–352, 380, 385n49, 389–391 superstrate 208, 212, 392 Syriac See Aramaic Taymanitic 1–3 , 7, 9, 12 Telkepe See Neo-Aramaic Thamudic B 1–13, 15 Thomason, Sarah vii, 42, 161, 173, 198, 222–223, 276n15, 282n13, 376n3 Tigre (Tǝgrē) See Ethiopian Semitic Tigrinya (Tǝgrǝñña) See Ethiopian Semitic Trudgill, Peter 214 Turkish 234, 234, 236–237, 247–248, 295 Ṭuroyo See Neo-Aramaic Ugaritic 10–11, 143–144, 182, 267–277, 375–377, 382–387, 393–396, 409–411, 414 Ukranian 252, 262 Vabrul 206n12 Van Coetsem, Franz vii n2, 42n20 Wanderwort 20–21, 181 Weinreich, Uriel vii, 65n21, 214, 333, 405n1 Whitney, Dwight vii n1 Windisch, E. vii n1 Winford, Don vii n2, 42–43 Wolane  See Ethiopian Semitic Xamtanga 20–21, 25–26, 29–30 Yiddish 189, 191, 216, 251–264 Zakho See Neo-Aramaic Zay See Ethiopian Semitic