Language Contact and the Development of Modern Hebrew 900430200X, 9789004302006

584 32 2MB

English Pages 356 [373] Year 2015

Report DMCA / Copyright


Polecaj historie

Language Contact and the Development of Modern Hebrew
 900430200X, 9789004302006

Citation preview

Language Contact and the Development of Modern Hebrew

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


The titles published in this series are listed at

Language Contact and the Development of Modern Hebrew Edited by

Edit Doron In cooperation with Ofra Tirosh-Becker and Sarah Bunin Benor


Originally published as Volume 3, Nos. 1–2 pp. 5–348 of Brill’s journal The Journal of Jewish Languages. In Cooperation with the Scholion Library The Institute of Jewish Studies The Hebrew University of Jerusalem Library of Congress Control Number: 2015955688

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 issn 0081-8461 isbn 978-90-04-30200-6 (hardback) isbn 978-90-04-31089-6 (e-book) Copyright 2016 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 List of Contributors ix List of Abbreviations xIV Introduction 1 Edit Doron

Clausal Predicates The Usual Suspects: Slavic, Yiddish, and the Accusative Existentials and Possessives in Modern Hebrew 25 Moshe Taube Predicate Nominal Sentences with the Hebrew ze and Its Russian Counterpart eto 36 Olga Kagan Bleached Verbs as Aspectual Auxiliaries in Colloquial Modern Hebrew and Arabic Dialects 49 Ophira Gamliel and Abed al-Rahman Mar’i Verbal Predicate Fronting in Modern Hebrew and Yiddish 63 Isaac L. Bleaman Circumstantial versus Depictive Secondary Predicates in Literary Hebrew—The Influence of Yiddish and Russian 76 Keren Dubnov

Clausal Periphery Modern Hebrew še- and Judeo-Spanish ke- (que-) in Independent Modal Constructions 89 Ora (Rodrigue) Schwarzwald and Sigal Shlomo Modern-Hebrew lama-še Interrogatives and Their Judeo-Spanish Origins 101 Itamar Francez



Colloquial Modern Hebrew Doubly-marked Interrogatives and Contact with Arabic and Neo-Aramaic Dialects 112 Samir Khalaily and Edit Doron The Right Periphery in Colloquial Hebrew: Modality and Language Contact Driven Effects 128 Yael Ziv Patterns of Dislocation: Judeo-Arabic Syntactic Influence on Modern Hebrew 146 Yehudit Henshke

Negation Superfluous Negation in Modern Hebrew and Its Origins 163 Aynat Rubinstein, Ivy Sichel, and Avigail Tsirkin-Sadan From Negative Polarity to Negative Concord—Slavic Footprints in the Diachronic Change of Hebrew meʔuma, klum, and šum davar 180 Einat-Haya Keren

Lexical Values The Sudden Disappearance of Nitpael and the Rise of Hitpael in Modern Hebrew, and the Role of Yiddish in the Process 197 Shira Wigderson Substrate Sources and Internal Evolution of Prescriptively Unwarranted Comitative Complements in Modern Hebrew 204 Yishai Neuman Inheritance and Slavic Contact in the Polysemy of bixlal 215 Avigail Tsirkin-Sadan The Expression of Material Constitution in Revival Hebrew 228 Chanan Ariel



Noun-Phrase Structure What Is New in the NP-Strategy for Expressing Reciprocity in Modern Hebrew and What Are Its Origins? 245 Elitzur A. Bar-Asher Siegal The Evolution of the Structure of Free Relative Clauses in Modern Hebrew: Internal Development and Contact Language Influence 258 Miri Bar-Ziv Levy and Vera Agranovsky The Impact of Contact Languages on the Grammaticalization of the Modern Hebrew Superlative 269 Yael Reshef The Impact of Contact Languages on the Degrammaticalization of the Hebrew Definite Article 281 Edit Doron and Irit Meir

Internal Development The Nature and Diachrony of Hebrew Quality Pseudo-Partitives: Are They a Calque from the Contact Languages? 301 Nimrod Shatil Reconsidering the Emergence of Non-core Dative Constructions in Modern Hebrew 309 Elitzur A. Bar-Asher Siegal and Nora Boneh

Borrowing A Constructional Idiom in Modern Hebrew: The Influence of English on a Native Hebrew Collocation 325 Malka Rappaport Hovav When the Construction Is Axla, Everything Is Axla: A Case of Combined Lexical and Structural Borrowing from Arabic to Hebrew 337 Roey J. Gafter and Uri Horesh Index 349

List of Contributors Vera Agranovsky is an MA student at the Hebrew University of Jerusalem in the department of Hebrew and Jewish Languages. Vera is responsible for teaching Hebrew to adults from the former Soviet Union at the Russian-speaking Jewry Unit of the Jewish Agency for Israel. Abed al-Rahman Mar’i is a senior lecturer of Hebrew literature at Beit Berl College. He has published several articles concerning the contact between spoken Hebrew and Palestinian Arabic. His book Walla Bseder: A Linguistic Profile of the IsraeliArabs ( Jerusalem: Keter, 2013) has attracted considerable media coverage in Israel. Chanan Ariel is a PhD candidate in the Hebrew Language department, and a member of the Research Group on the Emergence of Modern Hebrew at the Mandel Scholion Interdisciplinary Research Center in the Humanities and Jewish Studies, The Hebrew University of Jerusalem. He has worked at the Historical Dictionary Project of the Academy of the Hebrew Language. His dissertation focuses on the language of Maimonides. Elitzur A. Bar-Asher Siegal is a senior lecturer in the department of Hebrew Language and a fellow at the Language Logic Cognition Center at the Hebrew University of Jerusalem. His research concentrates on the Semitic languages and their history; in addition, he is interested in topics in semantics. His recent studies are on reciprocal constructions, datival expressions, and negation. Miri Bar-Ziv Levy is a PhD student at the Hebrew University in the department of Hebrew and Jewish Languages. She is a member of the Research Group on the Emergence of Modern Hebrew at the Mandel Scholion Interdisciplinary Research Center in the Humanities and Jewish Studies, The Hebrew University of Jerusalem. Isaac L. Bleaman is a doctoral student in the Department of Linguistics at New York University. His current research focuses on sociolinguistic variation in contemporary spoken Yiddish.


list of contributors

Nora Boneh is a lecturer of Linguistics at the Hebrew University of Jerusalem and a fellow at the Language Logic and Cognition Center at the Hebrew University of Jerusalem. Her research topics include the syntax and semantics of core and non-core datives, the interaction between tense, aspect, and modality, in particular in the expression of habituality, and the syntax of clausal possession. Edit Doron is Professor of Linguistics and a member of the LLCC (Language, Logic, and Cognition Center) at the Hebrew University of Jerusalem. Her research interests include syntax and semantics of Hebrew, Arabic, Neo-Aramaic, and French. She is a co-editor of Brill’s Journal of Afroasiatic Languages and Linguistics and an associate editor of Theoretical Linguistics. Keren Dubnov (MA in theoretical and applied linguistics, Moscow State University, PhD in Hebrew language, the Hebrew University of Jerusalem) is a scientific secretary at the Academy of the Hebrew Language and a lecturer at the David Yellin Academic College of Education in Jerusalem. Itamar Francez (BA 1999, The Hebrew University of Jerusalem; PhD 2007, Stanford University) is an Assistant Professor of Linguistics at the University of Chicago. His main research areas are in semantics, pragmatics, and the syntax-semantics interface. Ophira Gamliel is a teaching fellow at Bar-Ilan University and a language instructor at the Hebrew University of Jerusalem. She has published several articles on Jewish Malayalam and is currently working on an Introduction to Malayalam Grammar. Roey J. Gafter holds a PhD in linguistics from Stanford University, and is currently a postdoctoral fellow at the Tel Aviv University Department of Linguistics. His research interests are in the field of sociolinguistics, focusing mainly on issues of language and ethnicity, and how the two interact in Modern Hebrew. Yehudit Henshke is a senior lecturer in the Department of Hebrew Language at the University of Haifa. Her fields of interest include Mishnaic Hebrew, Jewish languages,

list of contributors


Judeo-Arabic, and Modern Hebrew. She has recently been awarded a grant from the Israel Science Foundation to pursue a project that will study the influence of Judeo-Arabic on contemporary Hebrew. Uri Horesh is Assistant Professor of Instruction and the language coordinator for the Program in Middle East and North African Studies at Northwestern University. He holds a PhD in Sociolinguistics from the University of Essex. His research interests are in language contact among Semitic languages, and language variation and change in varieties of Arabic. Olga Kagan completed her PhD studies at the Hebrew University of Jerusalem in 2008, with the dissertation “On the Semantics of Structural Case.” She has worked at the Hebrew University of Jerusalem, University of California, Santa Cruz, and Bar-Ilan University. Currently she is a lecturer at Ben-Gurion University of the Negev. Kagan is the author of Semantics of Genitive Objects in Russian: A Study of Genitive of Negation and Intensional Genitive Case (2013) and Verbal Prefixes in Slavic (forthcoming). Einat-Haya Keren is a PhD student in generative linguistics at the Hebrew University of Jerusalem. She is a member of The Emergence of Modern Hebrew group at the Mandel Scholion Center. Her main interests are Negation and the Creole and Koiné languages. She is especially interested in the link between those languages and Modern Hebrew. Samir Khalaily (BA 1986; MA 1992, The Hebrew University of Jerusalem; PhD 1997, Leiden University) teaches English syntax, morphology, and pragmatics at Al-Qasemi College, and English for academic purposes at Zefat Academic College. He is interested in natural language syntax, and in the semantics of negation, metaphor, proper names, and Sufi discourse. Irit Meir is a Professor of Linguistics and an Associate Director of the Sign Language Lab at the University of Haifa. Her research interests include sign language linguistics, focusing on morpho-syntax and argument structure, modality effects on linguistic structure, and the emergence of sign languages. She also studies morphology, syntax, and language change in Modern Hebrew.


list of contributors

Yishai Neuman is a research Assistant Professor in the Department of Hebrew Language at Achva Academic College in Israel. His particular fields of research are phonology, contrastive syntax, language evolution, Hebrew linguistics, contact linguistics, historical linguistics, language planning, and writing-induced language change. His publications include articles on various aspects of general and Hebrew linguistics in European, North-American, and Israeli academic journals. Malka Rappaport Hovav is Henya Sharef Professor of Humanities and a member of the Linguistics department and the Language, Logic, and Cognition Center at the Hebrew University of Jerusalem. Her research focuses on lexical semantics and its interface with syntax, morphology, and conceptual categories. She is an author, with Beth Levin, of Unaccusativity (MIT Press) and Argument Realization (CUP). Yael Reshef studies the formation processes of Modern Hebrew, and is a member of the research group on the Emergence of Modern Hebrew at the Mandel Scholion Interdisciplinary Research Center in the Humanities and Jewish Studies, The Hebrew University of Jerusalem. Aynat Rubinstein is a postdoctoral fellow at the Mandel Scholion Interdisciplinary Research Center in the Humanities and Jewish Studies at the Hebrew University of Jerusalem. Her research concerns the context dependency and gradability of modals and expressions of attitude across languages. She is currently studying modality in Modern Hebrew and its emergence during the revival period. Ora (Rodrigue) Schwartzwald is Professor Emerita at Bar-Ilan University in Israel. She is a member of the Academy of the Hebrew Language. Her main research interests are Hebrew phonology and morphology, as well as Ladino ( Judeo-Spanish), and she has written many books and articles on these subjects. Sigal Shlomo wrote her dissertation on generic sentences in biblical wisdom literature (2005). She teaches academic writing and editing in the Department of Hebrew and Semitic Languages at Bar-Ilan University, and also teaches Hebrew as a second language. Her research interests are syntax and style in the Hebrew language throughout the ages.

list of contributors


Nimrod Shatil is the head of the Hebrew department at Zefat Academic College. His primary research interests are the accommodation of inherited Hebrew linguistic material to contemporary Hebrew, mainly in morphology, syntax, and lexicon. His special interest is in the reorganization of nominal patterns. Ivy Sichel teaches in the Linguistics Department at the Hebrew University of Jerusalem. She has worked on the syntax-semantics of pronouns, nominalizations, and negative indefinites, and more recently on the sociolinguistics of the revival of Hebrew. Moshe Taube is a Professor in the Department of Linguistics and in the Department of German, Russian, and East-European Studies at the Hebrew University of Jerusalem, and holder of the Tamara and Saveli Grinberg Chair in Russian Studies. His publications embrace the fields of Modern Yiddish grammar, mainly syntax, and Medieval East Slavic philology, mainly Ruthenian and Russian translations from Hebrew. Avigail Tsirkin-Sadan is a PhD candidate at the department of linguistics and member of the Research Group on the Emergence of Modern Hebrew at the Mandel Scholion Interdisciplinary Research Center in the Humanities and Jewish Studies, The Hebrew University of Jerusalem. Her dissertation focuses on the semantics of repetitive expressions in Modern Hebrew. Shira Wigderson is a graduate student in linguistics at the Hebrew University of Jerusalem. Her research interests include Hebrew language and linguistics, word formation, morphology, lexicology, and semantics. Yael Ziv Professor Ziv has taught at the Hebrew University of Jerusalem since 1976 in the departments of Linguistics and English. Her primary research area is discourse and pragmatics, with specific interests in Relevance Theory, information structure, centering theory, discourse structure and discourse markers, and the syntax-discourse interface.

List of Abbreviations * Marking of ungrammaticality a / adj Adjective abs Infinitive absolute acc Accusative adv Adverb ap Adjective Phrase aux Auxiliary BIRP Bar-Ilan Responsa Project BYP Ben-Yehuda Project caus Causative ccp Comment Clause Parenthetical cnj Conjunction cl Clitic cond Conditional cop Copula CoSIH Corpus of Spoken Israeli Hebrew cp Complementizer Phrase cs       Construct State D Determiner dat Dative def Definite dem Demonstrative dm Discourse Marker dmi Doubly Marked Interrogative dp Determiner Phrase du Dual dur Durative emph Emphatic expl Expletive f Feminine fr Free Relative clause fut Future gen Genitive GIH General Israeli Hebrew H Hebrew borrowing into Jewish language HJP Historical Jewish Press

list of abbreviations

imp Imperfect impr Imperative inf Infinitive inst Instrumental inch Inchoative int Interrogative ip Inflection Phrase IPH Israeli Periphery Hebrew L1 First language L2 Second language ld Left Dislocation loc Locative m Masculine mh Modern Hebrew MSA Modern Standard Arabic mv Main Verb N Noun N-word Negative expression neg Negation nci Negative Concord Item nom Nominative np Noun Phrase npi Negative Polarity Item obj Object p / pl Plural poss Possessive pp Preposition Phrase prf Perfect pron Pronominal copula pronH Personal-pronoun copula pronZ Demonstrative-pronoun copula prp Preposition prs / pres Present prt Particle pst Past ptcp Participle qud Question Under Discussion rd Right Dislocation recp Reciprocal


xvi refl Reflexive rel Relative complementizer s / sg Singular subj Subjunctive Super Neg Superfluous negation V Verb vp Verb Phrase

list of abbreviations

Introduction Edit Doron What is the relation between Hebrew in the pre-modern period and its contemporary form Modern Hebrew? Hebrew ceased to be spoken by native speakers at the beginning of the 3rd century CE but has been used as a liturgical, scholarly, legal, and cultural language of Jewish communities world-wide. Modern Hebrew (MH), since the early 20th century, has been the native language of the Jewish community in Palestine and later Israel. Given these historical facts, the question is often asked whether Modern Hebrew could possibly be, from a linguistic perspective, a stage in the development of Hebrew, transmitted from earlier stages. In the view of its speakers and traditional scholars, it undoubtedly is.1 According to one radical linguistic hypothesis, however, it is not; rather it is a creole2 based on a substrate of contact languages (particularly Yiddish, the native language of many of the Jewish immigrants to Palestine), and a Hebrew superstrate, serving solely as a lexifier.3 The present collection is a first attempt to present detailed analysis of the syntax of Modern Hebrew (MH), which goes beyond the controversy between the traditional and the radical views. More specifically, it aims at laying the grounds for a theoretically sound approach to the genesis of MH. Before introducing these studies, a brief historical overview might be necessary.4

* The project presented in this volume started out with the work and findings of the members of the 2013–2016 Research Group on the emergence of Modern Hebrew at the Mandel Scholion Interdisciplinary Research Center in the Humanities and Jewish Studies at the Hebrew University of Jerusalem, who have all contributed chapters to this volume. I would like to thank the Mandel Scholion center for providing generous support to our group, and the JJL editors, Sarah Bunin Benor and Ofra Tirosh-Becker, for hosting a special thematic volume and for the immense editorial work they have invested in its publication. I am grateful to Chanan Ariel, Miri Bar-Ziv Levy, Malka Rappaport Hovav, and Yael Reshef for their help with writing this introduction. 1  Reshef, forthcoming. 2  Wexler 1990, following Holm 1988; Izre’el 2001. 3  For the relexification view of the nature of creoles see Lefebvre 1998. 4  For more detailed information about the various stages of Hebrew, I refer the reader to Harshav 1999 and to the relevant articles in the Encyclopedia of Hebrew Language and Linguistics (Khan et al. 2013).

© koninklijke brill nv, leiden, ���6 | doi ��.��63/9789004310896_002



Historical Overview

Hebrew ceased to be a spoken language at the end of the 2nd century CE. Between this period and the time of the establishment of MH as a spoken language in Palestine, Hebrew consisted of a large body of writings containing a core of scripture, liturgical and traditional legal works, and an extensive range of scholarly and literary works. The central religious works were read and studied and used in worship over the centuries in Jewish communities in Europe, the Middle East, and North Africa, but also as far as Central Asia, India, and China. The language of all the writings contained elements of the earliest stages of written Hebrew that emerged when Hebrew was still a language with native speakers, and also elements of the written language from subsequent periods when it was no longer spoken.5 The written texts of Hebrew that emerged when it was still a spoken language are classified into two main corpora. These are the Old Testament, composed before the Common Era (Biblical Hebrew) and the Mishnah and related texts (Mishnaic Hebrew). Mishnaic texts, although composed in a period when Hebrew was a language with native speakers, are traditionally classified as part of Rabbinic Hebrew, together with the Talmud and other texts which were composed after Hebrew ceased to be spoken. This joint classification is due to both corpora together forming the legal (halakhic) basis of orthodox Jewish life and to their close linguistic affinity, clearly distinguished from Biblical Hebrew. Both corpora were committed to writing in Hebrew in Palestine and Babylon, by people whose native language was Aramaic, a language that greatly influenced the Rabbinic stage of Hebrew. The following period, Medieval Hebrew, starts around the 7th century with the Arab conquest, when Arabic replaced Aramaic as the native language in many Jewish communities. Vast religious, philosophical, and scientific corpora, and a lot of poetry and liturgical hymns were written during this period in lands under Islamic rule in the Middle East, North Africa, and Europe, and in adjacent lands under Christian rule. The corpus created from the 15th century onward is sometimes called Late Rabbinic Hebrew. It consists of halakhic writings, but also administrative documents relating to the life of communities in Europe and some literary work. In the 18th century in Eastern Europe scholars identify Hasidic Hebrew as a distinct corpus, which was rich in hagiography and homiletics typical of the Hasidic branch of Judaism.

5  There is, however, evidence of very rare circumstances where Hebrew was sometimes spoken, e.g., Tirosh-Becker 2015.



Modern publications in Hebrew start appearing in the middle of the 18th century in Central and Eastern Europe, the manifestation of a new cultural movement expressing Jewish secular aspirations for emancipation and integration into the surrounding culture. The language of these publications is known as the Hebrew of the Haskalah (Enlightenment), also called Maskilic Hebrew. Under the influence of the neo-classical trend of the times, some of the notable literary oeuvres were written in Biblical Hebrew. However, most writings (political, scientific, newspaper articles),6 and even some literary work, kept up the tradition of utilizing a mixture of all the historical stages of Hebrew. This conscious linguistic mixture of different historical periods turned into the mainstream style of written Hebrew during the second half of the 19th century. At the end of the century, in the aftermath of widespread pogroms (massacres of Jews in the Russian empire), Jewish refugees and others who supported the ideas of Jewish nationalism emigrated to Palestine, bringing along their Hebrew (sometimes called Revival Hebrew) as part of the foundation for a projected autonomous political entity. Palestine at the end of the 19th century was a province of the Ottoman Empire, with a population of about 500,000 Muslims, 5,000 Christians (mostly German), and 40,000 Jews.7 The Jewish inhabitants, who formed what is known as the Old Yishuv, were the descendants of Jews who had immigrated over the previous millennium for religious reasons. They lived in Jerusalem, Safed, and Tiberias, where they constituted the majority of the population, and in other ancient towns such as Hebron, Peqiʿin, and Shefarʿam. The oldest communities of the Old Yishuv spoke Arabic, and others were divided into communities speaking different languages according to their land of origin: Judeo-Spanish in the case of Sephardic communities, and Yiddish (also Hungarian, Rumanian, etc.) in the case of Ashkenazi communities. Members of the Old Yishuv used Hebrew as a lingua-franca when they interacted among themselves. They spoke it according to the traditional phonology in which the Sephardic communities, which were the majority, conducted their traditional readings of the Bible and other religious rituals.8 The newcomers from Europe adapted this pronunciation in some ways to Ashkenazi phonology9 and became the first community of Modern Hebrew speakers. The subsequent large waves of Jewish immigration eventually adopted this language. 6  Newspapers in Hebrew were published from the mid-1800s in various cities, e.g., Berlin, Warsaw, Cracow, St. Petersburg. 7  Lewis 1954, Ben-Aryeh & Bartal 1983. 8 Fellman 1973. 9  Ofer 2007.



They called it Hebrew and consciously aspired to be faithful to its historical legacy.10 A full educational system was instituted, run entirely in Hebrew, from kindergarten11 all the way to ­university.12 In 1922, Hebrew was granted the status of official language, alongside English and Arabic, of the British Mandate for Palestine. In 1948, Hebrew was proclaimed the official language of the state of Israel together with Arabic.

Overview of Contributions

The chapters in the present volume document syntactic structures novel to Modern Hebrew (MH) and investigate the origins of the innovation, looking for parallel structures in the contact languages and also for possible internal precursors of these structures. The results of the study have direct bearing on the controversy regarding the nature of MH. If the genesis of MH followed a creolization process, where Hebrew was but a lexifier, the novel syntactic structures of MH should not be developments of previous Hebrew structures, but should be accounted for solely on the basis of the syntax of the contact languages. The converse does not hold if the genesis of MH was based on transmission from previous stages of Hebrew. That is, assuming that MH results from transmission, there is still room for the influence of contact languages. Even if MH was transmitted from previous stages of Hebrew, this could not be the familiar transmission by native speakers in a normal historical development. The first generation of MH speakers spoke it as a second language (L2). This is clearly atypical. Language is typically transmitted by native speakers for whom the transmitted language is a first language (L1). In the case of Hebrew, transmission was mediated by L2 speakers, who were the original speakers of MH. The linguistic knowledge of L2 speakers is imperfect and may contain 10  Tutored by the Committee of the Hebrew Language (predecessor of the present-day Academy of the Hebrew Language) founded in Jerusalem in 1889 by the lexicographer Eliezer Ben-Yehuda, who had been the pioneer ideologue of the transformation of Hebrew into a spoken language. 11  The first kindergarten in Palestine run in Hebrew was established in 1898 by Esther Shapira in the town of Rishon LeZion, founded in 1882 and considered the first town of the New Yishuv. 12  The Technion (Institute of Technology) was founded in Haifa in 1912. In 1914 it committed to a curriculum taught in Hebrew, after a year-long struggle known as the “War of Languages” fought against the prior decision of its board of trustees to conduct classes in German. The Hebrew University of Jerusalem was founded in 1918 and has taught in Hebrew ever since it opened its gates in 1925.



inconsistencies and disruptions which inhibit linguistic transmission of the language. This imperfect transmission characterizes the process of language shift, when a group of speakers shifts to a target language by acquiring it as an L2, thus failing to learn it perfectly.13 L2-mediated transmission is bound to give rise to change due to the influence of the L2 speakers’ native languages. Moreover, the next generation, which was the first generation of L1 speakers of MH, underwent a process of deficient learning based solely on L2 speakers. Normally, when language is transmitted from L1 speakers to L1 speakers, there are language-specific innate principles which guide learning and secure smooth transmission. The transmission-based hypothesis—that MH was indeed transmitted from previous stages of Hebrew (though deficiently through L2 speakers)—can be supported by comparing the patterns of changes found in MH with those found in other known examples of deficient transmission. The studies in this volume support the transmission-based hypothesis by describing changes which are indeed similar to what is known about deficient transmission via L2 speakers in other documented cases of language contact, such as Kroch 2001 and Meisel 2011. There have been in the past arguments adduced against the creole-based genesis, notably Goldenberg 1996, Blau 2002, Zeldes 2013. These scholars have argued that MH includes many constructions already found in historical stages of Hebrew. The present studies show something more dramatic: that many novel constructions of MH, including those triggered by contact, are actually based in one way or another on original Hebrew structures. In other words, in many cases the novel constructions seem to encode change applied to the previous stages of Hebrew, whether internally motivated or contact-induced. The transmission-based hypothesis is best discussed in the framework of the view found in the generative literature since Chomsky (1981), where language variation is conceived as following from parametric differences. Grammars are taken to differ from each other along the lines of different parametric choices. In other words, grammars make different choices in selecting the values of underspecified formal features found in the characterization of particular parameters. For example, Rizzi (1982) has suggested that a grammar which allows question formation from within structures otherwise thought to be syntactic islands differs from one which doesn’t in selecting CP (Complementizer

13  As known in general about L2 acquisition, Gass & Selinker (2008:89–155).



Phrase) rather than IP (Inflection Phrase) as the value of the bounding node parameter, at least for constructing questions.14 Since Lightfoot (1991, 1999), Roberts (1993, 1999), Roberts & Roussou (2003), syntactic change has been conceived as diachronic modification in the values of parameters. In particular, two types of parameter modification have been described in previous studies of contact: (a) value transfer, where there is direct transfer into the language of the L2 speakers of the value settings of the parameters in their native language (the contact language), and (b) value reset, where L1 speakers change the value of parameters that originates in a previous generation. L1 speakers acquire the language at an early stage, when they still have access to the mental mechanism which enables them to reset the values of parameters. They do so in the face of the ambiguous and conflicting triggers they are exposed to in the speech of the preceding generation of L2 speakers. The present studies are not formulated in terms of universal parameters but rather language-specific constructions. What these studies show is that novel MH constructions follow the two modification types formulated above: value transfer from the contact language, and value reset. Novel constructions are shown in the present studies to be related to particular constructions within previous stages of Hebrew. The difference between the novel and original versions of a construction are shown to result from value change (either value transfer or value reset) of a property of the original construction. When inspecting a particular novel construction, one can tell whether it is the outcome of value transfer or value reset. Value transfer results in properties of the construction not originally found in Hebrew but identical to the ones found in the parallel construction within the contact language. Value reset results in properties of the construction which are neither originally apparent in Hebrew nor identical to properties of parallel constructions in the contact languages. Value reset is different from Rosén’s 1995 account of the change within the language of L1 speakers, whereby native speakers are consciously taught the historical values of the Hebrew parameters at school, or make use of general cognitive mechanisms such as analogy and back-formation which lead them to simplify and neutralize the grammar of the previous generation of L2 speakers. Beyond the similarities between the transmission of Hebrew in its earliest stages and other cases of L2 transmission, the present studies show changes that are particular to the transmission of Hebrew.

14  The example cannot be further elaborated in this short introduction. I refer the reader to Rizzi 1982:49–76 and Lightfoot 2006:142–144 where such an analysis has been suggested.







In the case of Hebrew, the L2-stage differs from more familiar examples of L2 acquisition, as it took place when there was no community of native speakers of the L2. It may be akin to language acquisition in a scenario of heritage language, where there is reduced access to primary linguistic data, and the data available to the learners is inconclusive. Such a scenario has been studied cross-linguistically (Montrul 2004, Polinsky & Kagan 2007) and has been shown to result in incomplete acquisition and grammatical reanalysis heavily influenced by the first language of the learners, leading to language change. In Hebrew, there was some value resetting at the stage of L2 speakers, maybe because they often acquired L2 at the same age as L1. Since it is early acquisition which allows resetting, young L2 learners might have overcome conflicting triggers they were exposed to in their two languages by resetting relevant values of MH constructions. An example is discussed in the chapter by Taube. In cases where the original Hebrew value of a given parameter/construction remains in MH alongside the value transferred from the contact language, the two values serve to distinguish registers. As is to be expected, the original Hebrew value tends to be used in prescriptive MH, whereas the value transferred from the native language of the L2 speakers is reserved for colloquial MH. This is amply illustrated in the chapters by Kagan, Ziv, Schwarzwald & Shlomo, Francez, Bar-Asher Siegal, Neuman, and Doron & Meir. The contact languages which seem to have influenced MH most were the languages spoken by the Jewish communities whose members were the first L2 speakers of MH. Some are Jewish languages which have evolved over centuries remote from their non-Jewish correlates (Yiddish and Judeo-Spanish). Others are Jewish dialects of the surrounding languages (Judeo-Arabic, Jewish Neo-Aramaic). In addition, there was influence of the languages surrounding the communities, both in Eastern Europe and in Palestine (Slavic, Arabic). The present studies also demonstrate that language contact is not unique to MH, but has been pervasive in the development of Hebrew over the ages. There is repeated mention of the influence of Aramaic during the period of antiquity and the influence of Arabic and Romance languages on Medieval Hebrew.

Another fact brought to light by the present studies is that contrary to claims that Yiddish might be a Slavic language (Wexler 1990), there are actually crucial differences between Yiddish and Slavic, and it is possible to demonstrate



in many cases whether value transfer originated from Yiddish or Slavic. This is demonstrated in the chapters by Kagan, Keren, Tsirkin-Sadan, and Ariel. The volume is organized along two dimensions. One dimension is concerned with the type of transmission, and separates constructions with properties clearly affected by language contact resulting in language shift (first four sections) from those with properties which might be the result of internal development or borrowing (last two sections respectively). Another factor plays a role within this dimension, and distinguishes two types of valuechange in the properties of constructions whose transmission was mediated by language contact: L2 transfer (the vast majority of the cases) vs. L1 reset (which can be detected in the chapters by Widgerson, Bleaman, and Ariel). The second dimension separates constituents according to their structural complexity: phrasal vs. lexical, and their grammatical role: clause central vs. clause peripheral. Clausal Predicates The phenomenon described in the first chapter, by Taube, seems to involve L2 value reset (rather than mere transfer). The chapter accounts for the possessive/existential construction, a thetic construction that goes back to Biblical Hebrew. The same construction is found in MH, but unlike previous stages of Hebrew, MH extensively marks the possessum as accusative—a novel marking that has never been properly understood (the possessum is originally nominative in Hebrew). Taube makes the discovery that in Slavic, the marking is genitive and crucially only occurs in negative clauses. This means that originally, the special marking is not at all a characteristic of the possessive/existential construction, but of negation—the realization of the well-known Slavic case value known as the genitive of negation. Taube shows how the Slavic value was first transferred by speakers of Yiddish to their own negative existential construction, utilizing accusative morphology (Yiddish does not have special forms for the genitive case). Later, learners of Hebrew reset the case value within their L2 Hebrew. The change must have taken place at the L2-stage, in view of evidence that it was already in effect among speakers prior to 1911, when there were practically no L1-speakers of MH. It appears that L2 speakers, faced with the case conflict between positive and negative existential constructions, and between the possessive and existential constructions, reset the MH case value to the accusative in all possessive/existential constructions, irrespective of negation. The next chapters contain examples where at the L2 stage, only value transfer is attested, as expected, not reset. Kagan discusses another MH construction which originates in Biblical Hebrew, one in which a pronoun functions as copula. In Biblical Hebrew, only personal pronouns can be copulas, and



this has been analyzed as a case of grammaticalization which took place in Biblical Hebrew.15 In MH, a different process is attested, whereby demonstrative pronouns start to function as copulas directly under contact with Slavic (here there is no Yiddish mediation). The novelty of this second grammaticalization is documented by the vast difference between the properties of personal vs. demonstrative copulas. The former are grammaticalized as syntactic heads (Doron 1983, 1986), while the latter, both in Slavic (Van Gelderen 2011) and MH (Spector Shirz 2014) are specifiers. The latter process can be viewed as the result of value-transfer by L2 speakers of the Slavic value, and is considered colloquial. It seems that there was never subsequent value reset by L1 speakers to unify personal and demonstrative pronouns.16 Gamliel & Mar’i also describe a construction of MH that originates in Biblical Hebrew. This is an example whereby lexical verbs are grammaticalized as aspectual auxiliaries conjoined to the main verb.17 This grammaticalization is attested in Biblical Hebrew for the auxiliary hālaḵ ‘go,’ which denotes imperfectivity, and šāḇ ‘return’ which denotes repetitivity (ḥāzar in Mishnaic Hebrew). In MH, the process expands. The relevant contact language this time is Arabic. Under the influence of Arabic dialects, the grammaticalization is extended in MH to additional verbs in the same conjunctive construction: ba ‘come’ (ingressive aspect) and yašav ‘sit’ (durative aspect); Lucas 2015:528 shows the same grammaticalization of the latter verb in Maltese English, also under contact with Arabic. This transfer is facilitated by the fact that go is a grammaticalized auxiliary also in Arabic. Bleaman discusses another MH construction that originates in Biblical Hebrew, one where an infinitival form of the verb within a clausal predicate is used to reduplicate its finite form. In MH, a construction involving reduplication is found as well. Unlike previous views, which conclude that this is the same construction (Goldenberg 1971), Bleaman carefully traces vast differences between the two: morphological, syntactic and pragmatic. In particular, the MH construction is phrasal, unlike the lexical reduplication in Biblical Hebrew. The MH phrase always appears to the left of the entire clause, unlike the Biblical infinitive which immediately precedes or even follows the verb. Again, we have here a case where the basic construction is due to a Biblical Hebrew parameter (the separation of the lexical verb from the inflection). But the construction in MH is different in many ways from the basic Biblical 15  This is part of the “Copula Cycle” discussed by Van Gelderen 2011:Ch. 4. 16  Data discussed in Spector Shirtz 2014:Ch. 2 fn 4 perhaps show that at a certain point there might have been beginnings of such reset. 17  This falls under the “Aspect Cycle” discussed by Van Gelderen 2011:Ch. 7.6.



c­ onstruction. Rather it has the properties of a parallel Yiddish construction and is thus probably due to value-transfer by L2 MH speakers whose native language was Yiddish. There seems to have been subsequent reset by L1 speakers which allowed the new value to be adopted in contemporary MH, and the Biblical value was relegated to frozen examples within archaic stylized writing. Dubnov too discusses a predicational construction that originates in Biblical Hebrew, the secondary predicate. She shows that in the Old Testament, only depictive secondary predicates are used. These are secondary predicates which are typically adjectives, sometimes participles, they are participant related, they follow and form a prosodic unit with the main predicate, and they are obligatorily within the scope of sentential negation. Dubnov shows that in early MH literature a new type of secondary predicate enters the language— the circumstantial secondary predicate. As in the construction discussed by Bleaman, here too the new version has very different properties from the Biblical version. Circumstantial secondary predicates are typically participles rather than adjectives, they further describe the event rather than one of its participants, they can precede the main predicate and form a separate constituent, and finally they can escape the scope of sentential negation. Dubnov argues that this cluster of properties characterizes a special part of speech found in Slavic and Yiddish, the adverbial participle, whose only function is expressing a circumstantial secondary predicate. It thus seems that the value determining the choice of secondary predicates in these languages was transferred by the speakers to their L2 MH, allowing the inclusion of ordinary participles as secondary predicates in MH. Clausal Periphery Schwarzwald & Shlomo describe a construction which consists of a clause introduced by the complementizer ‘that’ (še in Hebrew). Originally in Hebrew, as is typically the case cross-linguistically, the complementizer is only found in embedded clauses. In the future tense, such clauses have a modal interpretation, and in MH are found in a construction consisting of main clauses expressing modality, e.g., requests and wishes. The authors argue that the novel construction evolved in MH under the influence of Judeo-Spanish, where the same construction is attested for parallel modal clauses with a complementizer (ke in Judeo-Spanish). This raises an interesting puzzle, since Judeo-Spanish is a minor contact language restricted to a very small subcommunity of the early MH L2 speakers. How did the new value main-clause for the distribution of this construction catch up with the entire community of MH speakers?



The answer is, as argued by the authors, that the new value main-clause had actually previously been introduced into Hebrew through medieval contact with Spanish and other Romance languages (with the same complementizer que). It made its way into Late Rabbinic texts, and was adopted by early users of MH, independently of transfer from Judeo-Spanish. Thus, this value became generally available. A particular variant of the same construction is described by Francez, which he argues likewise originates in Judeo-Spanish. Here the future-tense main-clause introduced by še ‘that’ is embedded under the question word lama ‘why.’ This particular variant has very special semantics-pragmatics, identical to those of the parallel construction in Judeo-Spanish, which substantiates a common origin. In particular, when the construction includes negation, it is interpreted both in MH and in Judeo-Spanish as a suggestion with a positive polarity (Francez calls the construction under such an interpretation “suggesterogative”). The converse is also true in both languages: when the construction does not include negation, it cannot be interpreted as suggesterogative. Francez explains the special contribution of negation in both languages as follows: the suggesterogative interpretation depends on negation taking scope outside both the that-clause and the modal that operator, i.e., it is a high negation operative at the level of the speech-act; this explanation is substantiated by the ungrammaticality of negative concord items within suggesterogatives in both languages, which is due to the fact that such items are not licensed within clauses not including the negation. The complex composition of the suggesterogative is probably not independent in the two languages, suggesting that this might be a case of L2 transfer. Another construction unique to the clause periphery, also embedded under lama ‘why,’ is described by Khalaily & Doron, interestingly originating in other contact languages. The construction entered MH through contact with Arabic dialects and Neo-Aramaic dialects, probably through L2 transfer, but is attested as a general Semitic construction also found in Rabbinic Hebrew. The ancient origin may have facilitated its entering MH. In this construction the clause embedded under why is not a that-clause but a question. This double question has its own special semantics-pragmatics, and is constructed as a rhetorical question which rejects a presupposition present in the discourse. The same double question with the same rejection function is also found in the construction within the contact languages, which argues for a common origin. The chapter by Ziv is concerned with the non-canonical word order found in colloquial Hebrew, typically in the spoken modality, where certain elements are placed in the right periphery of the clause (i.e., clause final). Examples



include parentheticals, tags and discourse markers, focus markers, certain conjunctions, and dislocated constituents. Ziv shows that these elements were found in Biblical Hebrew, but their function was limited, and they were typically clause initial rather than clause final. It is only in MH that these elements start to appear in the right periphery, particularly in the spoken modality, and acquire special functions determined by information structure considerations. Both the placement of these elements in a position which follows the constituents over which they scope, and the special functions they acquire in this position, were probably introduced by L2 MH learners by transferring the relevant value from Slavic and Yiddish, where the same word order with the same functions is attested as well. Henshke’s chapter is unique in that it shows the mechanism of value transfer at work in real time. Henshke studies Israeli Periphery Hebrew (IPH), spoken in Israeli communities that mostly consist of descendants of native speakers of North African Judeo-Arabic dialects. She argues that in IPH, the variety of values for the dislocation construction is strikingly richer than in colloquial MH in general and is the result of transfer from Judeo-Arabic. She shows many very subtle and detailed differences between IPH and general colloquial MH and argues that in all these cases the IPH values conform to Judeo-Arabic, demonstrating that contact language influence does not consist in diffuse language similarity but in a very precise transfer of particular values. For example, dislocated constructions in IPH serve a richer array of functions than in MH, such as the presentation of deictic relations, including social or locative relations anchored to the speaker, similarly to Judeo-Arabic. Moreover, the variety of elements which can appear dislocated in the (right or left) periphery of the clause is richer in IPH than in MH, and so is the type of clause-internal pronominals which resume the dislocated element. In MH, referential noun phrases can be dislocated and related to resumptive personal pronouns (including the pronominal clitics attached to prepositions). IPH freely allows, in addition, the dislocation of personal pronouns, which may be resumed not only by personal pronouns but also by verbal subject inflection. Moreover, IPH allows the dislocation of demonstrative pronouns. In such cases, the resumptive pronoun is demonstrative. Neither subject inflection nor demonstrative pronouns can function as resumptive pronouns in colloquial MH in general. IPH often makes use of the echo construction, where the resumptive pronoun is exactly of the same type as the dislocated element (both are personal pronouns, or both are demonstrative pronouns) and also the striking complex echo construction with simultaneous left and right dislocation of the same pronominal. Studies like Henshke’s real-time documentation are extremely valuable. For example, the documentation of the echo construction in IPH will prove crucial in a scenario



where the echo construction spreads to general MH, as it will demonstrate that it originates from contact with Judeo-Arabic (since the spread started in IPH) rather than Yiddish, which also has the same construction. Negation Rubinstein, Sichel, & Tsirkin-Sadan discuss superfluous negation (Super Neg)—negation which does not reverse the truth conditions of the clause. Constructions with such negation include free relatives, exclamative rhetorical questions, clausal complements of ‘until,’ ‘without,’ and ‘before,’ clausal complements of ‘fear’-type verbs, complements of negated ‘surprise,’ and the complement of ‘almost.’ These values have been transferred from Slavic and Yiddish. Haspelmath & König (1998) establish the areal nature of the phenomenon among certain eastern European languages. They speculate that Yiddish borrowed the construction from Russian, Polish, or Ukrainian. Yet Super Neg with slightly different values is attested in previous periods of Hebrew. Complements of ‘fear’ and other verbs of this class were introduced by both ‘lest’ and še-lo ‘that-neg.’ Super Neg uses of še-lo in this construction are attested in early Rabbinic texts. Super Neg in the complement of kimʕat ‘almost’ is attested in small numbers in Medieval Hebrew and later. Constructions with Super Neg in MH thus do not all share the same path of development. Several constructions disappeared (kimʕat še-lo ‘almost thatneg’) while others lived on to become part of MH grammar. Language contact may have reinforced existing patterns of Hebrew (‘fear’ verbs), led to reanalysis of others (ʕad še-lo ‘until that-neg’), and introduced altogether new forms into the language (such as in free relatives). Keren’s chapter is concerned with the change in interpretation and distribution of particular negative items. She shows that in Biblical and Rabbinic Hebrew, the negative items meʔuma, klum, and šum davar, best translated to English as anything, have the distribution of Negative Polarity Items (NPIs). These are lexical items which cannot freely appear in clauses of the language, but are restricted to clauses with the semantic property of being downward entailing (first formulated by Ladusaw 1979). Clauses which have this property include negative and interrogative clauses, protases of conditionals, clauses with particular quantifiers, and others. When examining Revival Hebrew texts, Keren finds a change in the distribution of meʔuma, klum, and šum davar. Their distribution is now more restricted, they only appear in negative clauses but not in other types of downward entailing clauses. Keren suggests that they are now better translated to English as nothing. Negative elements restricted to negative clauses have been called Negative Concord Items (Ladusaw 1992). Why would the NPIs meʔuma, klum, and šum davar turn into NCIs? Keren



examines their counterparts in Polish, Russian, and Yiddish, and shows that in Polish and Russian these items are strictly NCIs, but that is less clear for Yiddish. Thus the change is probably due to value transfer from Slavic. Keren’s chapter is unique in that it presents a case of language change which simply cannot be detected by comparative textual work. Without the theoretical distinction between NPIs and NCIs, it would be natural, in view of their Biblical and Rabbinic distribution, to consider insignificant the fact that meʔuma, klum, and šum davar are missing from Revival interrogatives, conditional protases, etc., especially if textual work is not buttressed by asking for speakers’ judgments. Without querying speakers, one would not know that MH speakers consider ungrammatical interrogatives and protases containing these items, unless negation is present as well. On the other hand, researchers who are native speakers might unconsciously rely on their own native-speaker judgments. They would thus be biased in favor of the modern distribution and conclude from the texts that meʔuma, klum, and šum davar are restricted to negative clauses throughout the history of Hebrew, and that the Biblical and Rabbinic interrogatives and protases are an unexplained quirk. Whichever distribution the researchers assume, the change would not be detected. Lexical Choices The chapter by Wigderson is concerned with L1 reset. In the L2 generation, speakers favored a verbal system with two different exponents of the middle intensive template, the Biblical form hitpaʕel and the Rabbinic nitpaʕel, which served to distinguish nuances of agentivity vs. passivity. Wigderson suggests that the distinction is due to an L2 transfer from Yiddish. What is surprising is the almost complete loss of the distinction in the speech of the succeeding L1 generation. Wigderson attributes it to the subset principle (Berwick 1985): children always make the most conservative hypothesis when engaged in value resetting. In the present example, children do not get enough evidence for using nitpaʕel, and they conclude that their language permits one exponent only, hitpaʕel. Only if children systematically heard adults using nitpaʕel would they conclude that their language permits it too. Assuming that the L2 speakers used nitpaʕel mostly in writing, but not much in their spoken language, especially with children, this may explain the value reset by the L1 generation. Neuman presents a case with L2 transfer of the values of lexical exponents from contact languages, eventually partly overruled by the original Hebrew values. He shows that the comitative reversible preposition with is syncretic in



many languages with the non-reversible instrumental preposition. This is the case in Yiddish and Arabic, and the syncretic value of the parameter was transferred to the speech of L2 learners of MH. But in writing, L2 users preserved the original Hebrew value, where the preposition is reversible only, giving rise to the prescriptive use where ʕim ‘with’ is comitative only and is distinct from the non-reversible be- ‘by means of.’ In the next generation, both values are used by L1 speakers, depending on the context of use. Even educated speakers start allowing non-reversible uses of ʕim ‘with’ (this might be partly due to borrowing from non-reversible, possessive, uses of the English with). An example of L2 transfer from Slavic (rather than Yiddish) is discussed by Tsirkin-Sadan. It involves two related but separate Rabbinic collocations which are eventually reinterpreted as a single adverb in the speech of the first L2 learners of MH in accordance with Slavic. One Rabbinic collocation is the combination of the preposition be- ‘in’ and the noun klal ‘totality,’ grammaticalized in Medieval Hebrew as the adverb biḵlal ‘in general.’18 The other is the Rabbinic negative polarity item klal ‘at all.’ In Slavic, but not in Yiddish, a single adverb carries both meanings, and it contains the locative preposition ‘in.’ The transfer of this complex meaning to MH promoted the creation of the single adverb bixlal with a blanket use of the locative preposition. The new adverb further acquires in MH the additional function of discourse marker that it has in Slavic. An additional example of the crucial role of Slavic is provided in the chapter by Ariel. The chapter follows the changes within the early stages of MH in the choice of the preposition which heads phrases expressing the material constitution of entities, e.g. ‘of glass.’ Ariel shows that historically, two options are found in the Hebrew sources: mostly the Rabbinic genitive preposition šel ‘of,’ but also the locative (source) preposition min ‘from,’ a vestige from Biblical Hebrew also used in Medieval Hebrew. Ariel shows that in the earliest phases of Revival Hebrew, writers originally favored the Rabbinic šel ‘of,’ gradually switching over the years to min ‘from’ as MH started to be spoken. This probably represents value transfer not from Yiddish, which uses the preposition fun, which has both possessive and locative interpretations, but from Russian, which uses the locative iz, which does not have a possessive meaning. L2 ­learners transferred the Slavic value and used the preposition min ‘from.’ For L1 speakers this becomes the only value, similarly to the process described in the chapters by Wigderson and Bleaman. 18  This is part of the “Case Cycle” discussed by Van Gelderen 2011:Ch 5.1.3.



Noun-phrase Structure Bar-Asher Siegal discusses two reciprocal constructions, the demonstrative construction originating in Mishnaic Hebrew, and the numeral construction calqued from a European construction, which came to be favored in (colloquial) MH probably because of L2 transfer from the contact languages. Similarly to other cases discussed in the volume, the original Hebrew construction was relegated to formal writing. Despite the semantic and syntactic resemblance between the new and the old constructions, they remain side by side as two independent constructions with different grammatical properties and semantic nuances. First, only a component of the numeral construction may be floated from complement to specifier position; this is impossible in the demonstrative construction, probably since a floated demonstrative is interpreted as referring to a different argument than a separate demonstrative in complement position. Second, only the components of the demonstrative construction can encode the singular/plural contrast and therefore distinguish between cases where reciprocity is held between individuals or between groups; plurality encoding is not attested in the numeral construction. Bar-Ziv Levy & Agranovsky show that the complementizer introducing free relatives had the Biblical value in the original stages of MH. The Mishnaic value was present as well, but it gained popularity once the use of MH became heavily influenced by the original languages of the users, and the Mishnaic value was favored since it matched the value transferred by speakers of the Yiddish/ Slavic value. In addition, the authors show that an innovative binary distinction in definiteness originates in transfer from the same contact languages (in particular Yiddish and Polish); this distinction is adopted and extensively made use of in the subsequent L1 generation. Reshef discusses a periphrastic superlative construction that was used in early Modern Hebrew texts and argues that it originates in Medieval contact with Arabic and Latin. This superlative consists of the definite determiner appended to a periphrastic comparative adjective of the form more + Adj. Reshef attributes the adoption of this definite periphrastic superlative to its affinity to the Hebrew definite form of the adjective which originally served as superlative but was less useful since it required explicit mention of the comparison class. In later texts of Modern Hebrew, the medieval periphrastic superlative is gradually lost and replaced by the original Hebrew periphrastic elative construction Adj + to-a-high-degree. Reshef attributes the loss of the previous construction more + Adj to its problematic word order, whereby the modifier more precedes the head Adj, which is problematic since Hebrew requires the opposite word order. This nevertheless does not determine the choice of the elative to replace the ungrammatically ordered comparative. Reshef therefore



concludes that the value-transfer from contact languages is merely functional rather than structural. But it is also possible to entertain a different hypothesis. In Slavic, the elative to a high extent/degree is used for the superlative. It is therefore possible that the adoption of the elative structure took place under the influence of the Slavic construction. Since Slavic was indeed a contact language in the period when the change took place, the 1920s, the change might after all be a case of structural L2 value-transfer. Doron & Meir discuss value changes of the Hebrew determiner (D). In Biblical Hebrew, as in Semitic in general, D is an inflectional feature of the category state (distinguishing the absolute/construct/emphatic states). Doron & Meir show a change in the morphosyntax of D in Medieval Hebrew, first internally within Rabbinic Hebrew texts, and later strengthened by contact first with Arabic and then with other European languages. The change in morphosyntax brought about a change in the values available to D and its reinterpretation as marking definiteness. In MH as spoken today D expresses values of definiteness, whereas Revival Hebrew preferred to stick to the state values of the category (still reflected in prescriptive Hebrew today). Under L2 transfer, the former value for D won in colloquial MH, due to the influence of Yiddish, the native language of many of the first L2 speakers of MH. Internal Change In some cases, the role of contact languages in influencing a construction is less clear. Shatil describes a Hebrew possessive structure of the form N1 of N2 which, as in other languages, came to have a particular evaluative interpretation—the quality pseudopartivite (also known as a binominal noun phrase). Shatil discusses the syntactic, semantic, prosodic and sociolinguistic features of the MH quality pseudopartitive. The structure is basically possessive, e.g., reax šel ʔorez, ‘smell of rice,’ but like in many languages the possessive structure also has pseudopartitive uses for particular types of N1, e.g., the quantity pseudopartitive kos šel ʔorez, ‘a cup of rice,’ including quality pseudopartitive uses where N1 is an evaluative noun, either positively or negatively, sometimes the nominalization of an adjective, such as yófi šel ʔorez, ‘good rice,’ literally ‘a beauty of rice.’ Pseudopartitives are special in that they are right-headed, unlike the original possessive which is left-headed. Shatil recognizes that the quality pseudopartivite construction might have originated in an inner process, but contact may still have played a role in the construction’s wide distribution, since it can be found in most of the languages of contact. The chapter by Bar-Asher Siegal & Boneh discusses a parameter which has remained stable and resisted change. Such parameters have been shown (by Biberauer & Roberts, forthcoming) to be the macro-parameters (so-called



by Baker 1996). They are unlike the micro-parameters, including language-specific constructions illustrated so far, which are sensitive to change as they are characterized by the value of a single property. Macro-parameters are clusters of properties and, as such, resist change. They often consist of particular categories like nouns, verbs, etc. Bar-Asher Siegal & Boneh discuss prepositions, in particular dative prepositions. They argue that the functions of dative prepositions are stable throughout Hebrew history, including MH. They point to what might be a counterexample, a MH use which is novel, the so-called discursive dative. Yet they argue that this use has probably existed all along, but has not been previously documented in Hebrew since it is restricted to the spoken modality. As such, it could not have been transmitted from previous stages of Hebrew to the L2 speakers of MH. Rather, it must have been transferred from their native languages, many of which included this use. Borrowing The chapter by Rappaport Hovav presents the one example in the volume of construction borrowing, which is a different type of contact-induced change from the shift discussed in all the other studies (the two types were distinguished by Thomason & Kaufman 1988). Borrowing happens in the converse direction: it happens when speakers reset a value within their language, not as part of the process of acquiring L2, but as part of L2 influence. There is no shift to L2, but rather an effect on L1 due to a relatively long period of contact with a prestige L2 (English in the present example), which is not adopted by the community but affects its L1. Recent changes within MH tend to be of the borrowing type, and the construction documented by Rappaport Hovav is very recent in MH. The construction involves the single collocation ʔibed ʔet ʕacmo l-a-daʕat ‘obliterate oneself consciously.’ Originally this construction was compositionally interpreted as ‘commit suicide,’ where the adverb consciously transparently contributed the conscious dimension of the act. This collocation froze in MH, together with the archaic prefixal form l- within the adverb consciously. In MH the transparency was lost, since adverbial prefixes were reinterpreted as prepositions. The preposition l- ‘to’ is systematically interpreted as goal in MH, including in resultative adjuncts. In order to preserve transparency, it made sense to reinterpret l- as a goal preposition introducing the result. This required reversing the sense of consciousness to denote the result of suicide, i.e., something akin to lack of consciousness, metaphorically death (reversal of the literal meaning of items in fixed collocations is often attested in Hebrew). This freed the verb within the collocation from carrying the resultative meaning of death and allowed its interpretation as a manner verb. Other manner verbs started being used as



well. The next step described by Rappaport Hovav involved the reinterpretation of the reflexive element as a functional element rather than necessarily as an argument, maybe akin to the prefixal reflexive morphology in the lexical version of obliterate oneself, i.e., hitʔabed. What was borrowed from English is the use of such a functional reflexive element to allow the application of resultatives to intransitive verbs. As a result, the range of possible verbs in the construction was extended to intransitives, and eventually to transitives with unspecified objects, very similarly to English, giving rise to a new productive construction of MH, a collocation with a variable verbal slot. When MH started being spoken in Palestine, lexical borrowing from Arabic was highly regarded, as it symbolized the metamorphosis of the Jewish new immigrants into authentic local inhabitants. Some of the borrowed items were later replaced, e.g., buǧaras ‘headache,’ čílba ‘enemy,’ čizbat ‘story,’ finǧan ‘coffeepot,’ ǧábal ‘mountain,’ ǧamáʔa ‘friends,’ ǧára ‘jar,’ ǧóra ‘sewage,’ sáxbak ‘friend,’ rásmi ‘formal,’ but many others remain integrated within MH, such as the adjective ʔáħla ‘more beautiful/most beautiful,’ borrowed as áxla. This lexical item is carefully studied in the chapter by Gafter & Horesh, and it allows them to unveil interesting characteristics of lexical borrowing. In Arabic, ʔáħla is clearly an adjective, as it exhibits the two morphosyntactic characteristics typical of comparative/superlative adjectives in Arabic. It is derived in the special template ʔaCCaC, and it can precede the head noun N in the attributive construction, unlike all other adjectives, which must follow N. áxla is borrowed together with its morphosyntactic characteristics. Yet these morphosyntactic characteristics do not characterize adjectives of any sort in Hebrew. The ʔaCCaC template is used to derive nouns rather than adjectives, and preceding the head N is only allowed for evaluative nouns in the attributive construction, never adjectives. Accordingly, áxla seems to have the form and d­ istribution of a noun in MH, though it definitely has an evaluative interpretation, something like ‘a good thing.’ Indeed, like other evaluative nouns, it is found together with the possessive preposition šel ‘of’ e.g., áxla (šel) órez ‘good rice’ within the Hebrew quality pseudopartitive construction discussed in Shatil’s chapter. Lexical borrowing thus seems to preserve morphosyntax and part of the semantics but surprisingly not necessarily the lexical category. In sum, the genesis of MH as studied in the present volume provides examples of various types of language change: reset, transfer, internal change, and borrowing. The various types of change illustrated in the volume collectively disprove the hypothesis that MH has a creole character and is based on a specifically Yiddish substrate with a Hebrew superstrate that serves solely as a lexifier. In particular, the examples of change as a result of contact appear to support the transmission hypothesis for MH. Though they involve constructions



influenced by contact with a range of languages, in many of the cases these constructions are modifications of constructions that existed in some form in earlier stages of Hebrew. References Baker, Mark. 1996. The Polysynthesis Parameter. New York: Oxford University Press. Ben-Aryeh, Yehoshua & Israel Bartal, eds. 1983. The History of Eretz Israel: The Last Phase of Ottoman Rule (1799–1917). Jerusalem: Keter (in Hebrew). Berwick, Robert. 1985. The Acquisition of Syntactic Knowledge. Cambridge, MA: MIT Press. Biberauer, Theresa & Ian Roberts. Forthcoming. “Parameter Setting.” In The Cambridge Handbook of Historical Syntax, eds. A. Ledgeway & I. Roberts. Cambridge: Cambridge University Press. Blau, Yehoshua. 2002. “Reflections on the Revival of Hebrew.” Leshonenu 65: 315–324 (in Hebrew). Chomsky, Noam. 1981. Lectures on Government and Binding. Dordrecht: Foris. Doron, Edit. 1983. Verbless Predicates in Hebrew. Ph.D. Dissertation, The University of Texas at Austin. ———. “The Pronominal Copula as Agreement Clitic.” In Syntax and Semantics 19: The Syntax of Pronominal Clitics, ed. H. Borer. New York: Academic Press, 313–332. Fellman, Jack. 1973. The Revival of a Classical Tongue: Eliezer Ben Yehuda and the Modern Hebrew Language. Berlin: Walter de Gruyter. Gass, Susan M. & Larry Selinker. 2008. Second Language Acquisition: An Introductory Course. 3rd edition. New York: Routledge. Goldenberg, Gideon. 1971. “Tautological Infinitive.” Israel Oriental Studies 1: 36–85. ———. 1996. “Hebrew as a Living Semitic Language.” In Evolution and Renewal— Trends in the Development of the Hebrew Language: Lectures Commemorating the 100th Anniversary of the Establishment of the Hebrew Language Council. Jerusalem: Publications of the Israel Academy of the Sciences and Humanities, 148–190 (in Hebrew). Harshav, Benjamin. 1999. Language in Time of Revolution. Stanford: Stanford University Press. Haspelmath, Martin & Ekkehard König. 1998. “Concessive Conditionals in the Languages of Europe.” In Adverbial Constructions in the Languages of Europe, ed. Johan van der Auwera. Berlin: Mouton de Gruyter, 563–640. Holm, John A. 1988. Pidgins and Creoles. Cambridge: Cambridge University Press. Izre’el, Shlomo. 2001. “On Processes in the Formation of the Hebrew Spoken in Israel.” In Speaking Hebrew: Research of Spoken Language and Linguistic Variation in Israel, Te’uda 18, ed. S. Izre’el. Tel Aviv: TAU Press, 217–238 (in Hebrew).



Khan, Geoffrey et al., eds. 2013. Encyclopedia of Hebrew Language and Linguistics. Leiden: Brill. Kroch, Anthony S. 2001. “Syntactic Change.” In The Handbook of Contemporary Syntactic Theory, eds. Mark. Baltin & Christopher Collins. Oxford: Blackwell, 699–729. Ladusaw, William A. 1979. Polarity Sensitivity as Inherent Scope Relations. Ph.D. Dis­ sertation, University of Texas, Austin. Also published, New York: Garland Press, 1980. ———. 1992. “Expressing Negation.” In Semantics and Linguistic Theory (SALT) II. Ithaca, NY: Cornell University Press. Lefebvre, Claire. 1998. Creole Genesis and the Acquisition of Grammar: The Case of Haitian Creole. Cambridge: Cambridge University Press. Lightfoot, David. 1991. How to Set Parameters. Arguments from Language Change. Cambridge, MA: The MIT Press. ———. 1999. The Development of Language. Oxford: Blackwell. ———. 2006. How New Languages Emerge. Cambridge: Cambridge University Press. Lewis, Bernard. 1954. “Studies in the Ottoman Archives—I.” Bulletin of the School of Oriental and African Studies 16.3: 469–501. Lucas, Christopher. 2015. “Contact-induced Language Change.” In The Routledge Handbook of Historical Linguistics, eds. Claire Bowern & Bethwyn Evans. New York: Routledge. Meisel, Jürgen M. 2011. “Bilingual Language Acquisition and Theories of Diachronic Change: Bilingualism as Cause and Effect of Grammatical Change.” Bilingualism: Language and Cognition 14.2: 121–145. Montrul, Silvina. 2004. “Subject and Object Expression in Spanish Heritage Speakers: A Case of Morphosyntactic Convergence.” Bilingualism: Language and Cognition 7.2: 125–142. Ofer, Yosef. 2007. “The Origin of Israeli Pronunciation.” In Sha’arei Lashon: Studies in Hebrew, Aramaic and Jewish Languages Presented to Moshe Bar-Asher, vol. III, eds. A. Maman, S.E. Fassberg, & Y. Breuer. Jerusalem: Bialik Institute, 166–172 (in Hebrew). Polinsky, Maria & Olga Kagan. 2007. “Heritage Languages: In the ‘Wild’ and in the Classroom.” Language and Linguistics Compass 1.5: 368–395. Pintzuk, Susan. 2008. “Variationist Approaches to Syntactic Change.” In The Handbook of Historical Linguistics, eds. Brian D. Joseph & Richard D. Janda. New York: Wiley. 509–528. Reshef, Yael. 2013. “Revival of Hebrew: Sociolinguistic Dimension.” In Encyclopedia of Hebrew Language and Linguistics, vol. 3, eds. G. Khan et al. Leiden: Brill, 408–415. ———. Forthcoming. Hebrew in the Mandate Period. Jerusalem: The Academy of the Hebrew Language (in Hebrew). Rizzi, Luigi. 1982. Issues in Italian Syntax. Dordrecht: Foris. Roberts, Ian. 1993. Verbs and Diachronic Syntax. A Comparative History of English and French. Dordrecht: Kluwer.



———. 1999. “Verb Movement and Markedness.” In Language Creation and Language Change: Creolization, Diachrony, and Development, ed. M. DeGraff. Cambridge, MA: The MIT Press, 287–327. ——— & Anna Roussou. 2003. Syntactic Change: A Minimalist Approach to Grammaticalization. Cambridge: Cambridge University Press. Rosén, Haiim B. 1995. Hebrew at the Crossroad of Cultures. Leuven and Paris: Peeters. Spector Shirtz, Ilona. 2014. The Syntax of Non-verbal Predication in Modern Hebrew: Predicate Nominals, Pseudoclefts and Clefts. Ph.D. Dissertation, The Hebrew University of Jerusalem. Thomason, Sarah G. & Terrence Kaufman. 1988. Language Contact, Creolization, and Genetic Linguistics. Berkeley, CA: UC Press. Tirosh-Becker, Ofra. 2015. “Eliʿezer Ben-Yehuda and Algerian Jews: Relationship and Language.” In Arabic and Semitic Linguistics Contextualized. A Festschrift for Jan Retsö, ed. L. Edzard. Wiesbaden: Harrassowitz Verlag, 432–433. Van Gelderen, Elly. 2011. The Linguistic Cycle: Language Change and the Language Faculty. Oxford: Oxford University Press. Wexler, Paul. 1990. The Schizoid Nature of Modern Hebrew: A Slavic Language in Search of a Semitic Past. Wiesbaden: Otto Harrassowitz. Zeldes, Amir. 2013. “Is Modern Hebrew Standard Average European? The View from European.” Linguistic Typology 17: 439–470.

Clausal Predicates

The Usual Suspects: Slavic, Yiddish, and the Accusative Existentials and Possessives in Modern Hebrew Moshe Taube

The Hebrew University of Jerusalem, Israel [email protected]

Abstract Existential and possessive constructions with a definite NP marked as object appear in both spoken and written Modern Hebrew. This paper ascribes their appearance to negative existential and possessive constructions with genitive accusative in Slavic languages (Polish, Russian, Ukrainian). These were reinterpreted in Yiddish as accusative and subsequently calqued by bilingual speakers of Modern Hebrew in the first generations of its emergence as a spoken language.

Keywords structural borrowing – calques – existential constructions – possessive constructions

Spoken Modern Hebrew, at least from the 1910s, and increasingly also some varieties of written Modern Hebrew, make use of existential and possessive constructions with the existential marker yeš ‘there is,’ or its negated counterpart ʔeyn ‘there is not,’ and the object marker ʔet followed by a definite NP. In the possessive variety of these existential constructions, the possessor is marked by the preposition le- ‘to.’ These constructions have attracted the attention of many linguists (e.g., Blanc 1989 [collected reprint of columns written in the early 1950s]:16; Rosén 1957:30–31; Ziv 1976, 1982a, 1982b; Berman 1980; Glinert 1990; Henkin 1994; Givón 2001, I:135; Kuzar 2002; Boneh 2003; Matras & Schiff 2005; Reshef 2008; Goldenberg 2013:300–311).

* I wish to thank Lea Sawicki and Dana Taube for their critical remarks.

© koninklijke brill nv, leiden, ���6 | doi ��.��63/9789004310896_003



This interest is apparently due to the “discrepancy tolerated within them between the basic assertion of existence and the object position of the existing entity” (Goldenberg 2013:310). This discrepancy stems from the fact that the existential construction “is a thetic sentence, ‘there is X,’ of which no part represents given information, and which is consequently not construable as a categorical judgment built of a subject and a predicate. It comprises an existing entity and the expression of existence but nothing that is thematic” (Goldenberg 2013:300). Hence, “there is constitutionally in such cases a discrepancy between the ‘logical contents’ on the information level and the usual /S-῀-P/ structure of predicative statements. This often causes instability, hesitation, and shifts in the structure of such sentences” (ibid.). Goldenberg goes on to show that such shifts occur in several Semitic languages, both ancient and modern, including Modern Hebrew and Palestinian Arabic. Long before they drew the attention of descriptive and theoretical linguists, these constructions, primarily of the possessive variety, were being repeatedly banished by custodians of the purity of the Hebrew language (for a review of these, see Reshef 2008). The earliest example of such banishment unearthed by Reshef (2008:226) is from 1911: (1)

ʔal tomar: ‘Don’t say:’ ha-yeš lexa INT-EXIST ‘Do you have this thing?’

ʔet ACC

ha-davar DEF-thing

ha-ze? DEF-this

ʔeyn lo ʔet NEG.EXIST to.him ACC ‘He does not have this thing.’

ha-davar DEF-thing

ha-ze DEF-this

ʔemor: ‘Say:’ ha-yeš lexa INT-EXIST ‘Do you have this thing?’

ha-davar DEF-thing

ha-ze DEF-this

ʔeyn lo ha-davar NEG.EXIST to.him DEF-thing ‘He does not have this thing.’

ha-ze DEF-this

The Usual Suspects


The earliest example of such construction in written Hebrew, cited by Reshef (2008:231), is from a satirical column by Amos Keinan which he published between 1950 and 1952 in the Haaretz newspaper: (2) ha-zaken haya mabsut nora ve-ʔamar la-cair: ʔeyze mazal še-yeš li ʔotxa which luck that-EXIST you.ACC ‘The old man was very glad and said to the young one: What luck that I have got you!’ As early as 1929, the use in Hebrew of the accusative to mark the possessum in the existential-possessive constructions, instead of the “correct” nominative, is condemned by the purists as reflecting the unwelcome influence of a foreign language, namely Yiddish, which marks the possessum by the accusative in the transitive construction with hobn ‘have’ (I. Avineri, quoted in Reshef 2008:229). H. Blanc and H.B. Rosén, who in the 1950s, first in magazine columns and radio talks and then in collections and monographs, launched the first non-normativist (indeed anti-normativist) descriptions of Modern Spoken Hebrew, while acknowledging the influence of Yiddish, pointed out nevertheless (Blanc 1989:16; Rosén 1957:30–31) that (a) these constructions are current in the mouths of native speakers of Hebrew, and that (b) in forms such as our example (2), where the possessum is pronominal, the condemned accusative is in fact the only form possible. On the final page of his last book, finished just a few months before his death, Gideon Goldenberg cautions: “The search for foreign influences on the expansion of such incoherent constructions would not by itself provide their linguistic analysis, and may produce simplistic explanations” (2013:311). Fully concurring with this caveat, I propose, without choosing between the various explanations offered so far or attempting an analysis of my own, to confine myself in the present paper to pointing out the Yiddish and Slavic equivalents which most likely enhanced, if they did not actually trigger, the expansion of these constructions in Modern Spoken Hebrew. Yiddish makes use of a negated existential marker nito / ništo ‘there is not’ (cognate of German nicht da ‘not here’) followed by the accusative form of the (definite) NP, whether nominal or pronominal: (3) ništo Berlen NEG.EXIST Berl.ACC ‘Berl is gone’ (= disappeared or dead) / ‘Berl is not here.’



ništo dem Rebn NEG.EXIST the.ACC.M.SG Rebbe.ACC.SG ‘The Rebbe is gone’ (= disappeared or dead) / ‘The Rebbe is not here.’ ništo mikh NEG.EXIST me.ACC.SG ‘I’m gone’ / ‘I’m out of here.’ ništo NEG.EXIST ‘You’re gone.’

dikh you.ACC.SG

ništo im NEG.EXIST him.ACC.SG ‘He’s gone’ / ‘He’s not here.’ ništo zi NEG. EXIST her.ACC.SG ‘She’s gone’ / ‘She’s is not here.’ These constructions are amply attested in Yiddish literature of the 19th and 20th centuries. (4) haynt ober ništo im dem malekh today but NEG.EXIST him.ACC DEF.ACC.SG angel ‘But today he is not there, the angel.’ (Moyshe Prager, Antologye fun religyeze lider un dertseylungen, 1955:139) (5) Vider tseveynt zi zikh shtil un ruft: again weep.INCH she REFL quietly and calls nito im, Leyblen, mer! NEG.EXIST him.ACC.SG Leybl.ACC more ‘She again starts weeping quietly, calling out: He is no more, Leybl!’ (Leyzer Vilenkin, Gerangl, 1973:8) The accusative forms have not, however, totally put out of use the forms of nominative, as evidenced by the next example:


The Usual Suspects

(6) ikh zukh khaven, vu iz I seek Khave.ACC where is —nishto khave! . . .  NEG.EXIST Khave.NOM ‘I seek Khave, where is Khave? No Khave.’ (Sholem Aleykhem, Tevye der milkhiker, 1927:129)

khave? Khave.NOM

The accusative forms in the negative turn seem to be limited to definite NPs, as no example with an indefinite NP has been found. It should be noted, however, that only proper names, as well as a very small group of common names, mark case distinctions on the noun itself (in the singular only), and that there is no case marked on the negated indefinite article keyn. With pronouns, only the masculine distinguishes the nominative from the accusative in the singular, whereas in the plural the cases are not distinguished in any of the three genders. Hence in most cases there is no way of telling the accusative from the nominative for indefinite NPs. These accusative forms in Yiddish calque similar constructions in Slavic. The Russian linguist Yevgeniy Kagarov listed, among the few instances of indisputable Slavic influence on Yiddish syntax that he was willing to recognize as such, “the use of the genitive in negation, e.g. ništo im” (1926:426). The reason for this apparently mistaken identification of Yiddish accusative forms like im as genitive is the fact that the Slavic forms of the genitive singular masculine (e.g., the anaphoric pronouns for the third person, Russian его [ yevo], Polish jego/go, Ukrainian його [yoho] ‘him.GEN/ACC.SG.M’) are historical genitives which had gradually expanded into the domain of the accusative, among them the function of marking the direct object. There are some differences, e.g., with feminine pronouns, between Polish on the one hand (keeping the original accusative) and Ukrainian and Russian on the other (switching to genitive). In any case, after negation, the genitive form is obligatory in all three genders in the three languages (on the genitive of negation see Partee & Borshchev 2002, 2004, 2006; for a review of genitive/accusative syncretism in Slavic, see Comrie 1978). Thus, the end of Matthew 28:6 οἶδα γὰρ ὅτι Ἰησοῦν τὸν ἐσταυρωμένον ζητεῖτε. οὐκ ἔστιν ὧδε ‘I know that you are looking for Jesus, who was crucified. He is not here’ has the following renderings in Slavic, all of them with the anaphoric pronouns in the genitive: (7) Polish: Nie ma NEG.EXIST

Go him.GEN.SG

tu here



Russian: Его yevo him.GEN.SG

нет nyet NEG.EXIST

здесь zdes’ here

Ukrainian: Нема nema NEG.EXIST

Його yoho him.GEN.SG

тут tut here

Yiddish, however, has lost the genitive case, a process well under way also in many German dialects (to wit, the hugely popular 2004 book by Bastian Sick, Der Dativ ist dem Genitiv sein Tod). It is therefore quite understandable that speakers of Yiddish, who calqued the Slavic genitive/accusative forms of the negative existential constructions, interpreted them as accusative and rendered them in Yiddish with the corresponding accusative forms ništo im / ništo zi / ništo Berlen / ništo dem Rebn, etc. Speakers of Modern Hebrew, in turn, calqued these Yiddish turns of phrase with the accusative (direct calquing of the Slavic genitive/accusative forms seems less likely), yielding constructions such as the line from the 1984 song od nipageš performed by Arik Einstein (the lyrics are his): (8) yeš kan hakol ʔaval ʔeyn EXIST here all but NEG.EXIST ‘There’s everything here but there ain’t you.’

ʔotax you.ACC.FSG

The expansion of the existential construction with the definite-object-marker to positive assertions of existence, such as those cited in Goldenberg 2013:310 ( yēš ʔεt ʔēlle še-yḵōlīm la-ʕăśōt ha-kol mi-klūm ‘there are objthose who can do all from nothing’; hī ṣrīḵā ʕεzrā, ʔaz yēš ʔεtḵεm ‘she needs help, so there are objyou’), seems to be (though there is no independent evidence, such as later attestation, to prove it) an internal development within Modern Hebrew, probably inspired by analogy with the negative construction and predicted by the instability of the construction. In any case, neither Yiddish nor Slavic have positive constructions of that form. As regards the possessive variety, consisting of the aforementioned existential constructions expanded by a PP with le- denoting the possessor, a similar equivalence obtains, in that negative assertions of possession in Slavic languages mark the possessed with the genitive. Although they are languages with a “have”-verb (Polish mieć, Russian иметь, Ukrainian мати) requiring the

The Usual Suspects


accusative (but the genitive in negation), the Slavic languages also make use of existential-locative constructions for predicating possession, with Russian having gone furthest in making the existential construction with the preposition u ‘at, by’ the default, while basically confining the “have”-forms to collocations with abstract nouns like “have the honor,” “have the right,” etc. Thus we frequently find Russian sentences like the following one: (9) Художественное образование?! ?! Нет, ero y меня xudožestvvennoye obrazovanye nyet yevo u menya Artistic education no it.GEN.SG at me.GEN.SG нет. nyet NEG.EXIST ‘Art education?! No, I do not have one.’ (, accessed August 23, 2014) The same is true for Ukrainian: (10) Ви знаєте, що в нього немає vy znayete ščo v nyoho nemaye you know that at him.GEN.SG NEG.EXIST паспорта? passport.GEN.SG ‘Do you know that he does not have a passport?’ (, accessed August 23, 2014) (11) (Телевізор—ящик з токсичним спамом.) televizor—yaščyk z toksyčnym spamom! У мене його немає і супер! u mene yoho nemaye i super! by me.GEN him.GEN.SG NEG.EXIST and super ‘TV—a box with toxic spam. I do not have one and super!’ (Телевізор-ящик-з-токсичнимспамом-У-мене-його-немає-і-супер/103620483025663?sk=info, accessed August 23, 2014) Polish, which uses the “have” verb for predicating possession, does employ existential constructions with the same PP u and a noun or pronoun in the



genitive, however not to express possession, but any kind of involvement, association or relevance: (12) Nie ma u mnie tego problemu NEG EXIST by me.GEN this.GEN.SG problem.GEN.SG ‘I do not experience this problem (sc. on my computer).’ (,972/jaka-wersja-oprogramowania?m=25892, accessed August 23, 2014) Note that the idiomatic rendering in English of such utterances would still be ‘I don’t have that problem.’ (13) U mnie nie ma zmiłuj by me.GEN NEG EXIST have.mercy ‘For/With me there’s no “have mercy.” ’ (, accessed August 24, 2014) (14) A u mnie nie ma paczkomatu. and by me.GEN NEG EXIST packmachine.GEN.SG ‘In my town, however, there’s no packmachine.’ (, accessed August 24, 2014) Yiddish, which also has a “have”-verb, has expanded, under the influence of Slavic (especially Russian and Ukrainian), the use of existential-locative constructions with the preposition bay ‘at, by’ for predicating possession as well as all these additional meanings (see Taube 1984), but we have not found examples with definite nouns or with pronouns in the accusative. As with the existential constructions (without PP) treated above, in the case of existential constructions with a PP and object marking, too, only the negated ones have pretty close equivalents in Slavic and Yiddish, whereas those positively asserting possession or involvement of any kind should be looked upon as the result of internal development. This, despite the fact that Palestinian dialects of Arabic do have quite similar constructions such as the following (see Goldenberg 2013:311, citing Rosenhouse 2004 and Sadka 2009; see also Henkin 1994:53, fn 27):

The Usual Suspects


ʕind-o iyyā-h (or ʕind-o-yyā-h) [by-him1 OBJ-3SG.M 2]‘he1 has it2’ ʕind-kum iyyā-hā [by-you.PL OBJ-3SG.F] ‘you have her,’ mā ʕind-i-š iyyā-h [NEGby-me-NEG OBJ-3SG.M]‘I don’t have it.’ The absence in Yiddish of negative possessive constructions with accusative leads me therefore to assume that the Modern Hebrew possessive constructions with the accusative are an internal development, an expansion of the existential construction without PP, in which the accusative (in the negative turns) owes its appearance to Yiddish. My reluctance to include Spoken Arabic among the usual suspects named as sources of influence on the emerging Spoken Hebrew stems from the fact that the overwhelming majority of the Jewish population in Palestine during the crucial years of the emergence of Spoken Hebrew (ca. 1900–1920) were native speakers of Yiddish, with knowledge of at least one Slavic language, whereas knowledge of Arabic was much less prevalent among them (for details, see Bachi 1956). Conclusion The Modern Hebrew possessive/existential constructions with object marking of the possessum undoubtedly owe their expansion in Modern Hebrew to the main contact languages spoken by the first generations of Modern Hebrew speakers, namely Yiddish, Polish, Russian, and Ukrainian, in which similar constructions are attested. The “discrepancy between their theticity on the information level and the dimorphemic structure of normal predicative statements” (Goldenberg 2013:309) may have resulted in the same shift, to which such constructions are prone, but the presence of very similar constructions in languages that were the native tongues to the majority of speakers of Modern Hebrew in the crucial period of its emergence surely was a decisive catalyzer. References Bachi, Roberto. 1956. “A Statistical Analysis of the Revival of Hebrew in Israel.” Scripta Hierosolymitana 3: 179–247. Berman, Ruth A. 1980. “The Case of an (S)VO Language: Subjectless Constructions in Modern Hebrew.” Language 56: 759–776.



Blanc, Haim. 1989. lešon bney-adam (Human Language). Jerusalem: Bialik Institute (in Hebrew). Boneh, Nora. 2003. “Modern Hebrew Possessive yeš Constructions.” In Research in Afroasiatic Grammar II: Selected Papers from the Fifth Conference on Afroasiatic Languages, Paris 2000, ed. Jacqueline Lecarme. Amsterdam: John Benjamins, 63–77. Comrie, Bernard. 1978. “Genitive-Accusatives in Slavic: The Rules and Their Motivation.” International Review of Slavic Linguistics 3: 27–42. Givón, Talmy. 2001. Syntax: An Introduction. 2 vols. Amsterdam: John Benjamins. Glinert, Eliezer. 1990. “ ‘yeš . . .’ and ‘yeš lo . . .’ in Formal Contemporary Hebrew— Subject or Object? Stylistic Study in Spoken Hebrew.” Balšanut ʕivrit 28–30: 207–211 (in Hebrew). Goldenberg, Gideon. 2013. Semitic Languages: Features, Structures, Relations, Processes. Oxford: Oxford University Press. Henkin, Roni. 1994. “ ‘yeš gam et ze’.” Balšanut ʕivrit 38: 41–54 (in Hebrew). Kagarov, Yevgeniy. 1926. “di grund=stixye fun yidišn sintaksis.” Filologiše Šriftn fun YIVO 1:425–428. Kuzar, Ron. 2002. “The Simple Impersonal Construction in Language Style presented as Colloquial.” In Te‘uda, ed. Shlomo Izre’el. Tel Aviv: Tel-Aviv University, 18: 329–352 (in Hebrew). Matras, Yaron & Leora Schiff. 2005. “Spoken Israeli Hebrew Revisited: Structures and Variation.” In Studia Semitica: The Journal of Semitic Studies Jubilee Volume, eds. Philip S. Alexander et al. Oxford: Oxford University Press, 145–191. Partee, Barbara H. & Vladimir Borschev. 2002. “Genitive of Negation and Scope of Negation in Russian Existential Sentences.” In Annual Workshop on Formal Approaches to Slavic Linguistics: The Second Ann Arbor Meeting 2001 (FASL 10),” ed. Jindrich Toman. Ann Arbor, MI: Michigan Slavic Publications, 181–200. ―――. 2004. “The Semantics of Russian Genitive of Negation: The Nature and Role of Perspectival Structure.” In Proceedings of Semantics and Linguistic Theory (SALT) 14, eds. Kazuha Watanabe & Robert B. Young. Ithaca, NY: CLC Publications, 212–234. ―――. 2006. “Information Structure, Perspectival Structure, Diathesis Alternation, and the Russian Genitive of Negation.” In Proceedings of Ninth Symposium on Logic and Language (LoLa 9), Besenyőtelek, Hungary, August 24–26, 2006, eds. Beáta Gyuris, László Kálmán, Chris Piñón, & Károly Varasdi. Budapest: Research Institute for Linguistics, Hungarian Academy of Sciences, Theoretical Linguistics Programme, Eötvös Loránd University, 120–129. Reshef, Yael. 2008. “On the History of Attestation of the Possessive Construction ‘yeš lo et’ in the Early Days of Spoken Hebrew.” Lešonenu Laʕam 56: 226–233 (in Hebrew). Rosén, Haiim B. 1957. ʕivrit tova (Good Hebrew). Jerusalem: Qiryat Sefer. (2nd ed. 1966, 3rd ed. 1977) (in Hebrew).

The Usual Suspects


Rosenhouse, Judith. 2004. “’ilo/‘indo/ma‘o as Used to Express Possession in Colloquial Arabic in Israel.” Mediterranean Language Review 15: 129–153. Sadka, Isaac. 2009. “From ʕindo ktāb (‘by him a book’) to ʕindoyyāh (by him OBJit/him’).” In Mas’at Aharon: Linguistic Studies Presented to Aharon Dotan, eds. Haim E. Kohen & Moshe Bar-Asher. Jerusalem: Bialik Institute, 611–618 (in Hebrew). Taube, Moshe. 1984. “Langues-être, langues-avoir et le yiddish.” Orbis 33: 222–235. Ziv, Yael. 1976. “On the Reanalysis of Grammatical Terms in Hebrew Possessive Constructions.” In Studies in Modern Hebrew Syntax and Semantics: The Transformational-Generative Approach, ed. Peter Cole. Amsterdam: North-Holland Publishing Compan, 129–152. ―――. 1982a. “On So-called ‘Existentials’: A Typological Problem.” Lingua 56: 261–281. ―――. 1982b. “Another Look at Definites in Existentials.” Linguistics 18: 73–88.

Predicate Nominal Sentences with the Hebrew ze and Its Russian Counterpart eto Olga Kagan

Department of Foreign Literatures and Linguistics, Ben-Gurion University of the Negev, Beer-Sheva, Israel [email protected]; [email protected]

Abstract The article is devoted to Hebrew predicate nominal sentences in which the lexical item ze fulfills a copula-like function. A hypothesis is put forward according to which the demonstrative ze has acquired its new function under the influence of Slavic contact languages.

Keywords predicate nominal sentences – Hebrew – Slavic – copula

Introduction In this article, I discuss the lexical item ze ‘this’ in Modern Hebrew, especially its occurrence in predicate nominal sentences. I argue that the use of ze as a copula-like element in predicate nominals, as well as, plausibly, its occurrence in pseudoclefts and clefts, is a result of contact with Slavic languages.

Demonstrative and Copular ze in Hebrew

The distribution of the lexical item ze is extremely varied in Modern Hebrew. First, under what plausibly constitutes its most basic use, ze functions as a demonstrative (1a), which also has an attributive use illustrated in (1b) and (1c): * Acknowledgements: I wish to thank Edit Doron, Ilona Spector Shirtz, and Yael Greenberg for their discussion of the issues raised in this article.

© koninklijke brill nv, leiden, ���6 | doi ��.��63/9789004310896_004

Predicate Nominal Sentences


(1) a. .‫תן לי את זה‬ ten li ʔet ze give me acc this ‘Give this to me.’ ‫האיש הזה‬ b. ha-ʔiš ha-ze the-man the-this.m sg ‘this man’

c. ‫האישה הזאת‬ ha-ʔiša ha-zot the-woman the-this.f sg ‘this woman’ Under the attributive use, the demonstrative agrees with the modified noun in gender and number. Ze represents the masculine singular form. An additional, strongly related, use is one whereby ze appears as a “pointer” subject (Geach 1968): (2)

.‫זה דני‬

ze dani this Dani ‘This is Dani.’

The fact that the same item is used as both a demonstrative and a pointer is not surprising; in fact, in many languages, including English, both constructions contain the same word (e.g., this). Both the demonstrative and the pointer uses of ze can be traced back to Biblical Hebrew. However, ze is characterized by additional uses, which have entered the language in its Revival period. One of these is the copula-like function that this item receives in predicate nominal sentences such as (3): (3)

.‫דני זה חבר טוב שלי‬

dani ze ħaver tov šeli. Dani this friend good my ‘Dani is my good friend.’

This construction has received considerable attention in the literature on the syntax and semantics of Hebrew copular sentences (e.g. Sichel 1997; Heller



1999, 2001; Greenberg 2009; Spector Shirtz 2014; Danon 2012, in press). Ze, which under this use is often referred to as pronZ following Doron’s (1983) terminology, seems to fulfill the function of a copula (but it has also been claimed to constitute the subject, cf. Spector Shirtz 2014). Indeed, in example (3), it can be replaced with another pronominal copula, hu (pronH). (Both versions are provided in Greenberg [2009].) (4)

.‫דני הוא חבר טוב שלי‬

dani hu ħaver tov šeli. Dani he friend good my ‘Dani is my good friend.’

The use of pronZ in predicate nominal sentences is characterized by the following properties. First, pronZ does not agree “to the left,” i.e., with the subject: (5) .‫צלילה זה \ *זאת דבר מסוכן‬ clila ze/ *zot davar diving.f sg this.m sg this.f sg thing.m sg ‘Diving is a dangerous thing.’ (based on Greenberg’s [2009] example)

mesukan. dangerous.m sg

In (5), the subject is feminine singular, whereas the nominal predicate is masculine singular. PronZ cannot be inflected in the feminine form. Secondly, pronZ can agree “to the right,” with the nominal predicate, although such agreement is not obligatory (Sichel 1997). (6) .‫עישון זה \ זאת פעילות מסוכנת‬ ʕišun ze/ zot peʕilut mesukenet. smoking.m sg this.m sg this.f sg activity.f sg dangerous.f sg ‘Smoking is a dangerous activity.’ (based on Greenberg’s [2009] example) In (6), the subject is masculine and the predicate, feminine; here, pronZ can appear in the feminine singular form zot. Thirdly, while pronZ is compatible with nominal predicates, it is often unacceptable when the predicate is realized as an adjective phrase (AP):


Predicate Nominal Sentences


.‫דני הוא \ *זה גבוה‬ Dani hu/ *ze gavoha. Dani he this tall ‘Dani is tall.’ (Greenberg 2009:20)

In fact, there does exist a group of sentences in which pronZ seems to combine with AP predicates, as illustrated in (8). Note, however, that these sentences are special in that (i) the AP does not agree with the subject (8b) and (ii) the property denoted by the predicate need not be attributed directly to the referent of the subject. As argued by Greenberg (2009), in (8), the property of being interesting is attributed not to the students themselves but rather to situations/eventualities that involve young students, e.g., supervising a young student, etc. (8) a. .‫סטודנט צעיר זה מעניין‬ student caʕir ze student.m sg young.m sg this.m sg ‘A young student—that’s interesting.’ (Greenberg 2009:16)

meʕanyen. interesting.m sg

b. .‫סטודנטית צעירה זה מעניין \ *מעניינת‬ studentit ceʕira ze meʕanyen/ student.f sg young.f sg this.m sg interesting.m sg/ *meʕanyenet. interesting.f sg ‘A female young student—that’s interesting.’ I will have more to say about the construction illustrated in (8) in the third section. It is also worth noting that pronZ is compatible with the verbal copula (9). (9)

.‫אף פרויקט שלו זה לא היה הצלחה‬

ʔaf proyekt šelo ze lo haya haclaħa. no project his this neg was success ‘No project of his was ever successful.’ (Spector Shirtz 2014:68)



Semantically, predicate nominal sentences with pronZ have been linked in the literature to equative/identity (as opposed to predicative) meanings, i.e., the ones whereby a semantic identity between two expressions is asserted (cf. Heller 1999, 2001; Greenberg 2009; Danon 2012, in press; Spector Shirtz 2014). The distinction is inspired by Higgins’ (1973) classification. The most prototypical example of an equative sentence with ze is provided in (10): (10)

.‫זורו זה דון דיאגו דה לה וגה‬

zorro ze don diego de la vega. Zorro this don Diego de la Vega ‘Zorro is Don Diego de la Vega.’

At the same time, it is easy to find ze-sentences that, at least superficially, seem to belong to the predicational type, such as (3) or (8). See Greenberg (2009) for a proposal of how to reconcile such sentences with the equation hypothesis. To the best of my knowledge, the copular use of ze is not attested in Biblical Hebrew. Further, it does not have a counterpart in modern Arabic. It is quite widespread in Hebrew today, however, and is attested in early Modern Hebrew: (11) a. .‫חופש המצפון הרי זה דבר שאינו מחייב כלום‬ ħofeš ha-macpun harey ze davar freedom the-consciousness indeed this a.thing še-ʔeyno meħayev klum. that-doesn’t oblige anything ‘After all, freedom of speech is a thing that does not oblige.’ b. !‫חבר הלאומים זהו אנגליה‬ ħever ha-leʔumim zehu ʔangliya! United Nations this.he England ‘United Nations—that’s England!’ Haim Arlozorov (, accessed December 16, 2014) In (11b), ze appears in combination with the copula hu (pronH), as a result of which we get the item zehu, a combination that is found in today’s Modern Hebrew, too.

Predicate Nominal Sentences


Slavic Influence: ze versus eto

Following Dubnov (2013), I propose that the pronZ use of ze has entered Modern Hebrew as a result of the influence of Slavic contact languages, such as Russian, Ukrainian, and Polish. All these languages have an item that shares the functions of a demonstrative, pointer, and pronZ. In Russian, this is the word eto, in Ukrainian, ce, and in Polish, to. In (12)–(14), the demonstrative and copular use of each word is illustrated. Russian: (12) a. это решение eto rešenie this.n sg solution.n sg ‘this solution’ b. эта девочка eta devočka this.f sg girl.f sg ‘this girl’ c. Миша—это мой лучший друг. Miša eto moj lučšij drug Misha this my best friend ‘Misha is my best friend.’ Polish (Citko 2008:262-263): (13) a. to dziecko this.n sg child.n sg ‘this child’ b. Jan to mój najlepszy przyjaciel. Jan this my best friend ‘Jan is my best friend.’



Ukrainian: (14) a. це рішення ce rišennja this.n sg decision.n sg ‘this decision’ b. Київ—це велике місто. Kiyiv ce velyke misto. Kiev this big city ‘Kiev is a big city.’ In what follows, I concentrate mainly on Russian, although most of the facts hold equally for the other two languages. (In fact, as pointed out by Dubnov [2013], the phonetic similarity between the Ukrainian ce and the Hebrew ze could catalyze the discussed extension.) The similarity between the Russian copula eto and the Hebrew copula ze is quite striking. First, as indicated above, in both cases we deal with a demonstrative whose range of functions is extended to that of a pronominal copula in predicate nominal sentences. Second, similarly to ze, eto is compatible with nominal predicates (12a) but as a rule not with adjectival ones (e.g., Markman [2008] and Dubnov [2013] also point to the unacceptability of sentences such as (15), as well as their Hebrew counterparts): (15)

Дима (*это) высокий. Dima (*eto) vysokij. Dima this tall ‘Dima is tall.’1

At the same time, Russian does have counterparts of ze-sentences of the type illustrated in (8) above (Spector Shirtz 2014).

1  As pointed out by an anonymous reviewer, both “Dima—eto vysokij” in Russian and “Dani ze ha-gavoah” in Hebrew are acceptable in the sense of choosing between several individuals: ‘Dima/Dani is the tall one.’ I believe, however, that in these cases the predicate is, in fact, a nominal phrase with a phonologically empty noun. This is suggested by the interpretation as well as by the fact that in Hebrew, the adjective combines with the definiteness marker ha-.


Predicate Nominal Sentences

(16) Молодой студент—это интересно. Molodoj student eto interesno. young.m sg student.m sg this interesting2 ‘A young student that’s interesting.’ Just like (8), (16) contains a determiner phrase (DP) subject followed by a pronominal element, which, in turn, is followed by the predicate. The predicate does not agree with the subject. Also, what it modifies is not the subject itself but rather events/situations that involve the subject, e.g., ones of supervising a young student. According to Markman (2008), in Russian sentences such as (16), the predicate is realized as a short3 adjective in the default, neuter singular form (but see discussion in the third section). Third, similarly to ze, the copula eto does not agree to the left, with the subject: (17)

Медитация—это / *эта отличный отдых. Meditacija eto/ *eta otličnij otdyx. meditation this.n sg this.f sg great.m sg rest.m sg ‘Meditation is a wonderful way to rest.’

The grammatical features of eto are neuter singular (this is the default form in Russian, just as masculine singular is the default in Hebrew). When the subject is feminine singular as in (17), we still get the form eto. In fact, eto never exhibits agreement under its copula-like use; in other words, it does not agree with the predicate either (18). In this respect, it differs from ze. (18)

Медитация—это / *этот отличный отдых. Meditacija eto/ *etot otličnij meditation this.n sg this.m sg great.m sg ‘Meditation is a wonderful way to rest.’

otdyx. rest.m sg

Similarly to ze, eto is compatible with the verbal copula (cf. e.g., Markman 2008:366, Spector Shirtz 2014:221). 2  The word interesno can be analyzed as either an adjective in a neuter singular form or as an adverb. See discussion in the third section. 3  Russian adjectives appear in long and short (case-lacking) forms, cf. e.g., Bailyn (1994).



(19) Царевич Пётр—это был новый самозванец . . . Carevič Pёtr eto byl novyj samozvanec . . . prince Piotr this.n sg was.m sg new.m sg impostor.m sg ‘Prince Piotr was a new impostor.’ (, accessed December 16, 2014) Finally, eto-sentences are associated with equative readings (Geist 2007), in more or less the same way as ze-sentences in Hebrew. The tendency is present, although potential counterexamples can be found, too. In fact, in the most typical equatives, which contain two DPs that are asserted to have the same referent, eto is not only possible but also obligatory (Geist 2007).4 (20)

Зорро—*(это) дон Диего де ла Вега. Zorro *(eto) don Diego de la Vega. Zorro this don Diego de la Vega ‘Zorro is Don Diego de la Vega.’

To sum up thus far, ze and eto share a whole range of properties. Both items constitute demonstratives that can also fulfill the functions of pointer subjects and copula-like elements in predicate nominal sentences. (In fact, an overlap in their distribution is even wider, as will be mentioned below.) Both are compatible with nominal predicates but not with agreeing APs. At the same time, both are found in sentences with (arguably) non-agreeing AP predicates. Neither ze nor eto agrees with the subject. Finally, both items tend to appear in equative sentences. I propose that this striking similarity is non-accidental and results from the fact that the copula-like use of ze emerged in Modern Hebrew by the influence of Slavic languages.

Copular Sentences with AP(?) Predicates

The generalization that copula-like ze and eto are incompatible with AP predicates is contradicted by such examples as (8) and (16) above. I wish, however, to challenge this counterevidence by pointing out that in both languages, default form adjectives (masculine singular in Hebrew and short neuter 4  It should be noted that Russian does not have a counterpart of the Hebrew PronH (Dubnov 2013).

Predicate Nominal Sentences


singular in Russian) are typically homonymous with adverbs. There is evidence that sentences such as (16) in Russian, in which eto seems to combine with a non-agreeing adjective, in fact, contain adverbs. This view is supported by the fact that eto can be followed in such sentences by a different class of adverbs, ones that are not morpho-phonologically identical to adjectives. They are formed from adjectives by the attachment of the prefix po- and the suffixes -emu, -omu, or -i (depending on the adjective), e.g., po-russki ‘(in) Russian,’ po-našemu ‘our way,’ po-bystromu ‘quickly.’ (21) Аншлаг—это по-нашему. Anšlag eto po-našemu. this along-our ‘Full house—that’s our way.’ (, accessed December 16, 2014) If the sentences in question indeed contain adverbs, this explains the fact that they ascribe properties to events associated with the subject, rather than the subject itself. Evidence that they may contain adverbs in Russian raises further questions regarding the grammatical category of the corresponding phrases in the Hebrew construction.

Clefts and Pseudoclefts

It is important to point out that ze in Modern Hebrew has additional uses that are plausibly related to its role in predicate nominal sentences. Specifically, ze appears in clefts (22) and pseudoclefts (23). A detailed analysis of these constructions can be found in Heller (1999, 2001) and Spector Shirtz (2014). (22) (23)

.‫זה דני ששבר את החלון‬

ze dani še-šavar ʔet ha-ħalon. this Dani that-broke acc the-window ‘It is Dani who broke the window.’ .‫מי שאוהב לשתות קולה מפחית זה אביב‬

mi še-ʔohev lištot kola mi-paħit ze ʔaviv. who that-loves to.drink cola from-can this Aviv ‘The one who likes drinking cola from a can is Aviv.’ (Spector Shirtz 2014:2)



Once again, eto in Russian has analogous uses. This is particularly obvious for pseudoclefts, as illustrated in (24): (24)

Кем Дима был, так это великим доктором! Kem Dima byl tak eto velikim doctorom! who Dima was emph this great doctor ‘What Dima was is a great doctor.’ (Markman 2008:368)

The state of affairs with clefts is somewhat less obvious, but the Russian counterpart of (22) above is provided in (25) and does contain eto: (25)

Это Дима разбил окно. Eto Dima razbil okno. this Dima broke window ‘It was Dima who broke the window.’

The question of whether sentences such as (25) are, in fact, clefts is not trivial. They have similar pragmatic, focus-related properties, but lack the bi-clausal structure that generally characterizes clefts (or at least so it seems superficially). Several authors, including Markman (2008) and Spector Shirtz (2014), treat such sentences as clefts. It is possible that the use of ze in cleft and pseudocleft sentences results from the contact of Modern Hebrew with Slavic languages, such as Russian. This view is supported by the fact that pseudoclefts exhibit the copula-like use of ze. Further, ze-clefts are not found in Arabic (Ouhalla 1999) or Biblical Hebrew. This issue, however, requires further investigation. Conclusion Ze in Hebrew and eto in Russian share a strikingly wide range of syntactic and semantic properties, especially in their distribution and copula-like pronZ use. I have proposed that the range of uses of ze has been extended as a result of contact with Slavic languages.

Predicate Nominal Sentences


References Arlozorov, Haim. 1926. Yoman Ženeva. Ben-Yehuda Project Website, http://benyehuda. org/arlosoroff/004.html (accessed December 16, 2014). Bailyn, John. 1994. “The Syntax and Semantics of Russian Long and Short Adjectives: An X’- Theoretic Account.” In Formal Approaches to Slavic Linguistics: The Ann Arbor Meeting, ed. J. Toman. Ann Arbor, MI: Michigan Slavic Publications, 1–30. Citko, Barbara. 2008. “Small Clauses Reconsidered: Not So Small and Not All Alike.” Lingua 118: 261–295. Danon, Gabi. 2012. “Nothing to Agree On: Non-agreeing Subjects of Copular Clauses in Hebrew.” Acta Linguistica Hungarica 59.1–2: 85–108. ———. In press. “Ma ze ze? nituaħ taħbiri šel ha-ʔoged ‘ze’ ba-ʕivrit ha-modernit.” Hebrew Linguistics (in Hebrew). Doron, Edit. 1983. Verbless Predicates in Hebrew. PhD diss., Austin, TX: University of Texas, Austin. Dubnov, Keren. 2013. “Russian and Slavic Influence on Modern Hebrew.” In The Encyclopedia of Hebrew Language and Linguistics, vol. 3, eds. Geoffrey Khan et al. Leiden: Brill, 576–578. Geach, P. T. 1968. Reference and Generality. Ithaca, NY: Cornell University Press. Geist, Ljudmila. 2007. “Predication and Equation in Copular Sentences: Russian vs. English.” In Existence: Semantics and Syntax, eds. I. Comorovski & K. von Heusinger. Berlin: Springer. Greenberg, Yael. 2009. “Predication and Identity in Hebrew Nonpseudocleft Copular Sentences.” In Generative Approaches to Hebrew Linguistics, eds. S. Armon-Lotem, G. Danon, & S. Rothstein. Amsterdam: John Benjamins. Heller, Daphna. 1999. The Syntax and Semantics of Specificational Pseudoclefts in Hebrew. MA thesis, Tel-Aviv University, Tel Aviv. ———. 2001. “On the Relation of Connectivity and Specificational Pseudoclefts.” Natural Language Semantics 10: 243–284. Higgins, F. R. 1973. The Pseudocleft Construction in English. PhD diss., Boston, MA: MIT Press. Markman, Vita G. 2008. “Pronominal Copula Constructions Are What? Reduced Specificational Pseudo-Clefts!” In Proceedings of the 26th West Coast Conference on Formal Linguistics, eds. Charles B. Chang & Hannah J. Haynie. Somerville, MA: Cascadilla Proceedings Project. Ouhalla, J. (1999). “Focus and Arabic Clefts.” In The Grammar of Focus, eds. G. Rebuschi & L. Tuller. Amsterdam: John Benjamins. Sichel, Ivy. 1997. “Two Pronominal Copulas and the Syntax of Hebrew Nonverbal Sentences.” In Texas Linguistic Forum 38: The Syntax and Semantics of Predication,



eds. R. Blight & M. Moosally. Austin, TX: University of Texas Department of Linguistics. Spector Shirtz, Ilona. 2014. The Syntax of Non-verbal Predication in Modern Hebrew: Predicate Nominals, Pseudoclefts, and Clefts. PhD Dissertation, The Hebrew University of Jerusalem.

Bleached Verbs as Aspectual Auxiliaries in Colloquial Modern Hebrew and Arabic Dialects Ophira Gamliel

Bar-Ilan University, Israel [email protected]

Abed al-Rahman Mar’i

Beit Berl College, Kfar Saba, Israel [email protected]

Abstract Verbal inflections in Classical Hebrew and Arabic encode aspectual information such as perfective and imperfective. In modern Arabic dialects, an aspectual system has evolved through the auxiliary usage of bleached verbs, replacing the older system of aspectual inflections. Arguably, a similar process in which bleached verbs acquire aspectual use is now evolving in Colloquial Modern Hebrew. The article discusses the functions of the bleached verbs ‘sit’ and ‘come’ in Colloquial Modern Hebrew and Arabic.

Keywords Colloquial Modern Hebrew – Arabic dialects – serialized verb – bleached verbs – aspects

Introduction Aspect in Modern Hebrew is typically not marked by verb inflection, but rather depends on semantic, lexical, and syntactic contexts (Rosén 1977:179–88, 193–5; Boneh & Doron 2008:323–4). Contrarily, modern Arabic dialects have evolved * Our article benefitted from discussions with Edit Doron, Nora Boneh, Guy Ron Gilboa, and Eliran Levi, and we are grateful for their scholarly remarks and suggestions. All errors are due to our own shortcomings. We are also grateful to friends who willingly engaged with us in casual discussions over the matter: Enad Azbarga, May Arow, Victor Manevich, Aref Nammari, Samer Azaizy, Amir Aharoni, Sameer Kadan, and Amal Sharar.

© koninklijke brill nv, leiden, ���6 | doi ��.��63/9789004310896_005


Gamliel and al-Rahman mar ’ i

diverse means of marking aspects such as punctual, durative, ingressive, and progressive (Benmamoun 2000:28–34; Dahlgren 2008:241–5). Such aspect markers result from processes of semantic bleaching that lead to grammaticalization (Esseesy 2008). In the least grammaticalized constructions, aspect is expressed in Arabic by serialized verbs (Hussein 1990:349–351), of which one or more bleached verbs serve as auxiliaries (AUX) modifying the main verb (MV). For example, (1) has the verb ‘sit’ modifying the MV: (1) ḫilāl yomīn qaʕado wa-ḥallo ʔil-moškila during day-du sit-pst.3pl cnj-solve-pst.3pl def-problem ‘They were disentangling the problem for two days.’1 The event in (1) is ‘solve,’ with qaʕado marking it as a process that takes some time to complete. This article discusses a particular subtype of verb serialization in Hebrew, where a bleached verb functions as AUX preceding and modifying the MV in a complex predicate. Such complex predicates are based on coordinated verb construction, usually with a conjunctive particle (CNJ).2 Arguably, several bleached verbs in Colloquial Modern Hebrew are comparable in their functions to parallel verbs in Arabic dialects, especially in contemporary Palestinian Arabic. Unlike Hebrew, Arabic dialects incorporate bleached verbs in a variety of constructions, ranging from verb serialization to periphrastic formations, as we shall see below; the verbs subject to semantic bleaching and their respective functions as AUX are strikingly similar. In Modern Hebrew, the verb ‘return’ (ħazar) denotes iteration in its AUX function (Schwarzwald 2001:64–5; Boneh 2013). Similarly, in Arabic dialects, the verb ‘return’ (as in: rijaʕ saʔal-nī, ‘he asked me once more’) denotes iteration in asyndetic coordination (Elihai 2004:22; Brustad 2000:193). This similarity between Hebrew and Arabic in the functional usage of the verb ‘return’ also applies to other bleached verbs, such as ‘sit’ and ‘come,’ typical of colloquial 1  Provided by a native speaker of the central Palestinian rural dialect. 2  Coordinated verb constructions with bleached verbs differ from two other subtypes of verb coordination: (i) coordination denoting a sequence of events with a series of two or more MVs and (ii) coordination denoting a single event expressed by synonymous verbs, conventionally termed “hendiadys” (Lillas-Schuil 2006). All three subtypes of coordinated verb formations strikingly resemble serial verb constructions common in many Asian and African languages. See Hussein 1990; Delancey 1991; Lord 1993:1–7; Jayaseelan 2004; Aikhenvald & Dixon 2006:1–68, 338–350. Cf. Lord 1993:1–2; Kuzar 1996:122; Lillas-Schuil 2006:90. Moshe Taube (forthcoming) refers to such complex predicates as “verbal hendiadys,” a term that we propose reserving for synonymous verbs as in type (ii).

Bleached Verbs as Aspectual Auxiliaries


registers in Modern Hebrew. An example is in (2), which is comparable to the Arabic example in (1) above: (2) ‫ הוא ישב ותכנן חודשים על שנים‬ hu yašav ve-tixnen 3sgm sit-pst.3sgm cnj-plan-pst.3sgm ħodašim ʕal šanim months on years ‘He was planning for months and years.’ (accessed February 2, 2013)3 The event in (2), ‘planned,’ is modified by ‘sit’ for marking the event as one that spans over a duration of time. Thus, ‘sit’ in both examples (1) and (2) functions to indicate a durative aspect. In what follows, section 2 presents the durative ‘sit,’ and section 3 discusses the ingressive usage of the verb ‘come’ in Colloquial Modern Hebrew and in Arabic dialects. Section 4 examines similar aspectual functions of these two verbs in late nineteenth-century Hebrew texts. Section 5 summarizes our preliminary observations.

The Durative ‘Sit’

The durative function of the verb ‘sit’ juxtaposes lexical and grammatical categories. On the one hand, it functions grammatically as an aspectual auxiliary; it situates the event in a temporal relation of inclusion vis-à-vis the reference time (Smith 1991:135–54). On the other hand, its grammatical function relies on its inherent lexical durative property, which is retained in all its occurrences in Colloquial Modern Hebrew. The durative property of ‘sit’ is retained even when annulled in contexts where an actual postural event is impossible. In (2) above, ‘sit’ co-occurs with temporal adverbials depictive of its durative and inherent lexical ­property, while, at the same time, it functions as grammatical AUX modifying the MV ‘plan.’ In Arabic dialects, ‘sit’ marks the durative in various syntactic constructions ranging from serialized verbs to periphrastic formations; the latter, and the more common, is constituted by a participle (PTCP) with an inflected 3  See D7%90-%D7%90%D7%AA-%D7%94%D7%A6%D7%91%D7%90/.


Gamliel and al-Rahman mar ’ i

i­mperfect form (IMP) (Piamenta 1991:406; Watson 1993:158–159; Brustad 2000:195; Dahlgren 2008:242).4 Example (1) above is of a coordinated verb construction in contemporary Palestinian Arabic, identical in structure and in the function of the AUX ‘sit’ to the Hebrew example in (2). In both (1) and (2), in proximity with the temporal adverbial ‘for two days,’ ‘sit’ is bleached of its postural meaning to function grammatically as an AUX, while at the same time it retains its durative lexical meaning, modifying the MV ‘solve.’5 Additionally, the durative modification stands in contrast to punctual adverbs in both Colloquial Modern Hebrew and Palestinian Arabic, as shown in (3) and (4) respectively: (3) ‫ ישבתי וחשבתי ופתאום נזכרתי‬ yašavti ve-ħašavti ve-pitʔom nizkarti sit-pst.1sg cnj-think-pst.1sg cnj-suddenly recall-pst.1sg ‘I was thinking when I suddenly recalled.’ (accessed March 24, 2008)6 (4) qaʕdat wa-katbat wa-faǧʔā simʕat sit-pst-1sg cnj-write-pst-1sg cnj-suddenly pst-hear-1sg infiǧār ʕind ʔil-ǧīrān blast at def-neighbors ‘I was writing when I suddenly heard a blast over at the neighbors’.’7

4  To the best of our knowledge, an exhaustive list of dialectical forms where ‘sit’ functions as an AUX still awaits compilation. Kerstin Eksell (1995:68) gives a comparative survey of periphrastic formations with AUX, adding their respective regions and characterizations based on grammar books. The list includes the following formations with ‘sit’: qāʕed + IMP, Iraq, actual present/progressive; qaʕad + IMP/PRF, Syria et al., ingressive/durative/perfective. The list includes also grammaticalized formations with qa-/da- + IMP. According to Otto Jastrow (2014:211), “in Iraqi Arabic the actual present markers derived from *qāʕid are invariable qad-, ~ qa- in the qəltu dialects and invariable gāʕid ~ da- in the gələt dialects.” Two formations in Palestinian Arabic are excluded from Eksell’s survey, and we have not found their attestation in the secondary literature available to us; these are qāʕed + b- + IMP for durative present and qaʕad + CNJ + PRF for durative past. 5  The durative ‘sit’ in both (1) and (2) modifies the aktionsart of the MV ḥall ‘solve’ into an activity, which requires using a different verb altogether in the English translation. 6  See 7  Provided by a native speaker of the central Palestinian rural dialect.

Bleached Verbs as Aspectual Auxiliaries


The durative/continuous/progressive function of postural verbs such as ‘sit,’ ‘stand,’ and ‘lie down’ is a phenomenon common in many languages, likely resulting from their stative aktionsart.8 Arguably, the function of ‘sit’ in Colloquial Modern Hebrew results from contact with Arabic dialects, since its durative function is widespread over dialects, and, furthermore, it manifests in diverse constructions as mentioned above. Example (5) demonstrates the more common use of ‘sit’ as a bleached verb, where it appears in its PTCP form modifying an inflected IMP form. It also shows how far the durative function of ‘sit’ can go in Colloquial Arabic, whereby even speakers wonder about the usage of ‘sit’ with ‘walk,’ obviously annulling the lexical properties of ‘sit’ as a postural verb in this case. In (5), Person A asks in an online forum how ‘sit’ is compatible with ‘walk,’ and Person B answers jokingly with another example for the durative ‘sit’: (5) A: leš baʕeḍ ʔil-nās yiqūlu qāʕed why some people say-imp.3pl sit-PTCP.sG.m ʔimši ʔau jāles ʔimši walk-imp.1sg or sit-PTCP.sG.m walk-imp.1sg B: ʔanā ʔal-ħīn qāʕed ʔafkar 1sg now sit-PTCP.sG.m think-imp.1sg b-il-jawāb hahaha prp-def-answer ha ha ha A: ‘Why do some people say: “I sit walking” or “I sit down walking”?’ B: ‘I now sit thinking of the answer . . . ha ha ha.’ (accessed March 28, 2014)9 While in Hebrew the durative ‘sit’ is incompatible with motion verbs such as ‘walk,’ in Arabic, the option of periphrastic construction of PTCP + IMP enables this combination, suggesting that the processes of semantic bleaching and grammaticalization are by far more advanced in Arabic than in Hebrew. Arguably, the phenomenon of semantic bleaching in Colloquial Modern 8  The durative/progressive function of ‘sit’ occurs also in North-Germanic languages and in Colloquial Bulgarian in syndetic constructions comparable to coordinated verb constructions in Hebrew (Ebert 2000:607; Kuteva 1999; 2004:43–74). 9  See user%3Fuserid%3D12058081480105288680.


Gamliel and al-Rahman mar ’ i

Hebrew is compatible with Colloquial Arabic on a scale broader than the ­isolated case of ‘sit’ as an aspectual auxiliary. Colloquial Modern Hebrew and Arabic dialects both use several other bleached verbs.10 One example is the verb ‘come’ for the ingressive aspect, to which we turn next.

The Ingressive ‘Come’

Ron Kuzar (1996:137) defines the AUX ‘come’ as a modal verb with two different functions: increasing the agency of the MV on the one hand and, on the other hand, being used pragmatically to convey the speaker’s position regarding the event. Based on functions of ‘come’ comparable with Arabic dialects, as shown below, we speculate that the grammatical function of ‘come’ in AUX positions is ingressive. See, for example, (6): (6) ‫ ואז היא באה ו'העיפה' את הערפל בנוגע לזה‬,‫ זה היה שמועה‬,‫ לפני כן‬ li-fne-xen ze haya šmuʕa previously it be.pst.3sgm rumor ve-ʔaz hi baʔa ve-heʕifa cnj-then she come.pst.3sgf cnj-fly.caus.pst.3sgm ʔet ha-ʕarafel be-nogeʕa le-ze acc def-fog regarding prp-it Previously, it was a rumor. Then the [company] diffused the fog regarding it. (accessed May 15, 2011)11 While Kuzar compares this usage of ‘come’ to ‘go ahead and’ in English, we argue that it is more closely comparable with the use of ʔağā, ‘come,’ in Arabic dialects. The auxiliary usage of ‘come’ in Arabic dialects is discussed in several studies (Hussein 1990:349; Palva 1991:59; Brustad 2000:194; Henkin-Roitfarb 2010:152– 3), though mostly not in coordinated formations. Heikki Palva (1991:59) convincingly demonstrates that it functions as ingressive, as in his example ʔağā

10  Other bleached verbs in Colloquial Hebrew such as ‘get up’ (qam) and ‘go’ (halax) are comparable to Arabic dialects in this respect. In Classical Hebrew too, such usages are documented, as with the bleached verbs ‘rise’ (qūm), ‘go’ (hālax), and ‘return’ (šūḇ) marking the ingressive, progressive, and iterative aspects respectively (Agustinus 2013). An extensive study, though a desideratum, is beyond the scope of the present article. 11  See

Bleached Verbs as Aspectual Auxiliaries


katab, ‘then he wrote.’ The ingressive function is often used with speech verbs,12 as in (7): (7) ‫ היא באה ואמרה לי שהיא התאהבה בידיד שלה היא שברה אותי‬ hi baʔa ve-ʔamra 3sgf come-pst.3sgf cnj-tell-pst.3sgf l-i še-hi hitʔahava dat.1sg rel-3sgf fall in love-pst.3sgf be-yadid šel-a hi prp-friend.m gen-3sgf sub.3sgf šavra ʔot-i break-pst.3sgf acc-1sg ‘She then told me that she fell in love with her friend. She broke me.’ (accessed February 6, 2014)13 This is similarly the case in Palestinian Arabic, as in (8): (8) ʔağā wa-ʔāl-li come-pst.3sgm cnj-tell-pst.3sgm-1sg ṭīr min hūn fly-imp.sgm from here ‘He dared telling me to piss off!’14 Palva (1991:63) states that the ingressive ‘come’ often co-occurs with the ethical dative.15 Roni Henkin (2010:153) comments that the ethical dative variants “contribute to portraying an action as prompt, directed, and resolute.” Also, in Colloquial Modern Hebrew, the ethical dative co-occurs with the ingressive ‘come,’ reinforcing the action as daring and provoking: (9) ‫ ואז בא ושוכב לי על הרצפה כמו ילד מפגר‬ ve-ʔaz ba ve-šoxev then come-prs.sgm cnj-lie-prs.sgm 12  Interestingly, the Israeli author and translator Ioram Melcer wrote an op-ed regarding the use of ‘come’ with speech verbs in the legal language register (“baʔim ve-ʔomrim,” Haaretz, August 8, 2007). 13  See 14  Provided by a native speaker from Lydda (urban dialect). 15  For the ethical dative in Modern Hebrew and Syrian Arabic, see Al-Zahre and Boneh, 2010:249–251. See also Boneh & Bar-Asher Siegal 2014; and Bar-Asher Siegal & Boneh 2015.


Gamliel and al-Rahman mar ’ i

l-i ʕal ha-ricpa dat.1sg on def-floor kmo yeled mefager like child retarded ‘Then he dares me lying down on the floor like a retarded child.’ (December 25, 2012)16 Like the durative function of ‘sit,’ the ingressive function of ‘come’ juxtaposes lexical and grammatical categories. Ostensibly, it is this juxtaposition that generates the pragmatic nuances of daring and provocation previously observed by Kuzar. Pragmatic connotations generated by bleached verbs are dealt with in studies on Southeast Asian languages, where they are often viewed as semilexical auxiliaries (Butt & Gueder 2001; cf. Kimmig 2014).

The Durative and the Ingressive in Early Modern Hebrew

Late nineteenth-century texts contain only a few examples of the durative and the ingressive functions of ‘sit’ and ‘come.’ Examples (10) and (12) are from Masaʕot Ħabšuš, “The Travels of Habshush,” a travelogue written in 1894 partly in Hebrew and partly in Judeo-Arabic by R. Hayyim Habshush, a Yemenite Jew, whose native tongue was Judeo-Yemeni Arabic. Example (11) is from a travelogue written by R. Jacob Sapir, who migrated from Romania to Safed when he was eleven years old. He was conversant in Colloquial Arabic, as evident by his narration, which is spiced with phrases in Arabic transcribed into Hebrew (as in 1866, I, 3) and with meta-linguistic comments on Arabic (as in 1874, II, 1). (10) ‫ כשלושת ימים אתה יושב ובטל‬ ki-šlošet yamim ʔata 16  See A9%D7%AA%D7%99-%D7%91%D7%98%D7%A2%D7%95%D7%AA%D7%91%D7%99%D7%9C%D7%93-%D7%9E%D7%94%D7%9B%D7% 99%D7%AA%D7%94-%D7%A9%D7%9C%D7%99-%D7%90%D7%96%D7%94%D7%95%D7%90-%D7%91%D7%90-%D7%90%D7%9C%D7%99%D7%9E%D7%94-%D7%90%D7%AA-%D7%9E%D7%AA%D7%A0%D7%92 %D7%A9%D7%AA-%D7%91%D7%99-%D7%A8%D7%95%D7%A6%D7%94%D7%9E%D7%9B%D7%95%D7%AA-%D7%90%D7%96%D7%90%D7%9E%D7%A8%D7%AA%D7%99-%D7%9C%D7%95%D7%91%D7%95%D7%90.

Bleached Verbs as Aspectual Auxiliaries


like-three days you yošev u-ḇatel sit-prs.sgm cnj-being idle-prs.sgm ‘You have been idle for some three days.’ (Goitein 1983:12) In (10), the function of ‘sit’ as a durative modifying the MV ‘being idle’ is further clarified by the adverbial phrase ‘for three days.’ The use of ‘come’ for marking the ingressive and coloring the MV with a provocative connotation is demonstrated in (11) and (12): (11)

‫ ואם גם אחרי המודעה הזאת יבוא מישהו ויאמר ולא היא‬

ve-ʔim gam ʔaħarey ha-modaʕa and-if also after def-notice ha-zot yavo mišehu def-this come-fut.3sgm someone ve-yomar ve-lo hi cnj-say-fut.3sgm cnj-not that ‘Even if, regardless of this notice, someone will dare say, “not so!” . . .’ (Sapir, “Introduction,” 1866:2)

Example (12) is in Judeo-Yemeni,17 further substantiating our assumption that bleached verbs in Colloquial Modern Hebrew and their functions result from contact with Arabic dialects. Moreover, in (12), the AUX construction is asyndetic and, thus, adheres more closely to the pattern of bleached verb auxiliation in Arabic dialects:18 (12) ‫לׁש ָּבת‬ ַ ‫שח ֯יתּונֵ א ֵצ ַירך ַא‬ ִ ִ‫נכאף מן אלּגֵ יִ ם לאנהם ֵיָּבאּו י‬ ֯ ‫ ואחנא‬ wa-ʔiħna nixāf min cnj-1pl fear-prs.1pl from ʔal-geyimH li-ʔan.hum yabeʔuH def-gentiles that-they come-fut.3pl yišħituneH ṣeyraxH ʔal-šabātH defile-fut.3pl commodities def-Saturday ‘We fear the gentiles will dare to defile our Saturday commodities.’ (Goitein 1983:259) 17  Superscript H signifies a Hebrew loanword. 18  Note that Palva (1991:59) attributes the IMP + IMP construction to the Arabic dialects of Syria et al.


Gamliel and al-Rahman mar ’ i

Conclusion The colloquial Modern Hebrew examples above and in the appendix below suggest the emergence of aspectual marking comparable to aspectual functions of ‘sit’ and ‘come’ in Arabic dialects. The genetic affinity between Hebrew and Arabic requires no elaboration; neither does the fact that these two languages were in contact all through medieval times and up to the modern period (Henkin-Roitfarb 2011; Mar’i 2013:119–162). Colloquial Modern Hebrew has been in ongoing contact with Palestinian Arabic and other Arabic dialects since its earliest phases of revival.19 The incorporation of bleached verbs and, in particular, ‘sit’ and ‘come’ in syndetic coordination in Palestinian Arabic requires further investigation; bleached verbs in Arabic dialects normally function as such in periphrastic constructions (see footnote 4 above). The verbal syndeton attested in Palestinian Arabic is possibly a recent development.20 Furthermore, the durative ‘sit’ in syndetic coordination is attested in North-Germanic and Bulgarian as well, possibly reflecting a cross-linguistic functional property of ‘sit,’21 rather than cross-linguistic influence on Colloquial Modern Hebrew. We have not delved into the theoretical questions that stem from our initial investigation regarding semantic bleaching, grammaticalization, serialized verbs, semi-auxiliation and, above all, the nature of the interface between grammatical and lexical categories. In this article, we aimed at presenting the data as evidence for syntactic affinities between Colloquial Modern Hebrew and Arabic dialects in Israel.

19  Interestingly, Yiddish, too, was in contact with Palestinian Arabic, enough to justify the publication of a tutorial in Arabic for Yiddish speakers in 1918 in New York by Getzl (George) Zelikovitz (Simon 2009). 20  At this stage of our research, we can only speculate that the syndetic verbal coordination in Palestinian Arabic may have emerged either through contact with Modern Hebrew or, alternately, through processes of “shifting from a synthetic language to an analytic one whose syntactic relationships are expressed through strings of discrete morphemes” (Brustad 2000:70). 21  See footnote 8. In the secondary literature regarding aspects and bleached verbs, we did not find evidence for the durative ‘sit’ in syndetic coordination in Russian. However, one Russian speaker did point out that this is possible in Colloquial Modern Russian as well.


Bleached Verbs as Aspectual Auxiliaries

Appendix: More Examples for ‘Sit’ and ‘Come’ as Semi-auxiliaries

Examples (13)–(16), below, show the durative function of ‘sit’ with animate and inanimate subjects. Examples (17)–(24) demonstrate the ingressive and provocative ­functions of ‘come,’ with (20)–(23) featuring inanimate subjects. Example (19) has two separate complex predicates: two with the durative ‘sit’ and one with provocative ‘come.’ Examples (24)–(25) demonstrate asyndetic formation with the durative ‘sit’ and the ingressive ‘come,’ respectively. ‫אני כתבתי ספר (ועוד אחד בשלבי כתיבה) וחודשים ארוכים אני יושבת ועושה תיקונים‬ .‫ולא מעזה לשלוח אותו להוצאות לאור‬

)accessed February 9, 2010(22

.‫מוחות גדולים ישבו והקדישו את כל חייהם לתיאוריה הזאת‬

)accessed March 3, 2011(23 .‫הרעיון ישב והתבשל בראשו של בורשטיין מספר חודשים‬ )accessed August 29, 2011(24

,‫ לא משנה עד כמה אני מרגיש ש"זה" יושב ומבעבע בפנים‬.‫אני אף פעם לא מצליח ליצור‬ .‫שזה רוצה להתפרץ ולצאת החוצה‬

)accessed April 2, 2004(25

????‫איך פתאום באים וממציאים לנו משהו נורא כמו מלפפונים "חמוצים" מתוקים‬ )accessed May 30, 2007(26 !!!‫בושה להם שהם באו ואכלו באמצע הצום‬ )accessed August 9, 2009(27 ‫[ בעיות‬sic] ‫[ אותי כי יש אנשים אם‬sic] ‫והוא יושב ומסתלבט על כולם זה מטריח‬ ‫[ לו‬sic] ‫אמיתיות והוא בא וצוחק על חשבון כולם אשכרה אנשים ישבו והגדישו‬ ?‫זמן ובשביל מה‬ )accessed May 21, 2014(28


)14( )15( )16(


)18( )19(

22  See message_id=656592. 23  See 24  See,7340,L-4113639,00.html. 25  See 26  See 27  See 1000004801000aRCRD&page=1&vgnextoid=8b6b8682356c2210VgnVCM100000290c10acR CRD&. 28  See


Gamliel and al-Rahman mar ’ i .‫ואז המדינה באה ומחריבה כל סיכוי להגיע להסכם פשרה‬


.‫ וחבל שהפלאש בא וקלקל אותם‬,‫צבעי הרקע טובים מאוד‬

)21( )22( )23(

)accessed June 20, 2013(29

)accessed August 5, 2005(30

.‫שמשהו יבוא ויגמור את הסיפור הזה בשבילי‬

)accessed May 16, 2007(31

‫עד שיש לנו שחקן צעיר שיכול להתפתח ולהוביל ת`נבחרת שלנו בעתיד אז הצבא בא‬

!. . .‫והורס לו את הקריירה‬

)accessed October 4, 2009(32

.‫ אלא ישב דיבר עם כרושצ'וב‬,‫קנדי כלל לא טרח לפנות לקסטרו‬

)accessed August 10, 2003(33

. . . ‫ כאילו הם באמת מבינים משהו‬,‫אנשים באים מטיפים לי‬

(accessed November 28, 2008)34

)24( )25(

Primary Sources Goitein, Shlomo Dov (ed. and tr.). 1983. Travels in Yemen: Joseph Halevy’s Journey to Najran in the Year 1870 as Related by Hayyim Habshush. Jerusalem: Ben-Zvi Institute (in Hebrew and Judeo-Yemeni Arabic). Sapir, Jacob, 1866. Even Sapir I. Lyck: L. Silbermann (in Hebrew). ———. 1874. Even Sapir II. Mainz: Y. Brill (in Hebrew).

References Agustinus, Gianto. 2013. “Semantic Bleaching.” In Encyclopedia of Hebrew Language and Linguistics, eds. G. Khan et al. Leiden: Brill. Aikhenvald, Alexandra Y. & R.M.W. Dixon. 2006. Serial Verb Constructions: A Crosslinguistic Typology. New York: Oxford University Press. al-Rahman Marʾi, Abed. 2013. Walla Bseder: A Linguistic Profile of the Israeli-Arabs. Jerusalem: Keter. 29  See 8y9AjhmpAPFPFM&v=jTJWJRiuQxw. 30  See 31  See 32  See 33  See 34  See

Bleached Verbs as Aspectual Auxiliaries


Al-Zahre, Nisrine & Nora Boneh. 2010. “Coreferential Dative Constructions in Syrian Arabic and Modern Hebrew.” Brill’s Annual of Afroasiatic Languages and Linguistics 2: 248–282. Bar-Asher Siegel, Elitzur A. & Nora Boneh. 2015. “On the Origin of Modern Hebrew Non-selected Datives.” Journal of Jewish Languages 3.1–2: 309-322. Benmamoun, Elabbas. 2000. The Feature Structure of Functional Categories: A Comparative Study of Arabic Dialects. New York: Oxford University Press. Boneh, Nora. 2013. “Aspect: Modern Hebrew.” In Encyclopedia of Hebrew Language and Linguistics, eds. G. Khan et al. Leiden: Brill. Boneh, Nora & Elitzur Bar-Asher Siegal. 2014. “Hebrew Non-core Datives in Context.” Lešonenu 76.4: 1–34. Boneh, Nora & Edit Doron. 2008. “Habituality and the Habitual Aspect.” In Theoretical and Crosslinguistic Approaches to the Semantics of Aspect, ed. Susan Rothstein. Amsterdam: John Benjamins, 321–348. Brustad, Kirsten. 2000. The Syntax of Spoken Arabic: A Comprehensive Study of Moroccan, Egyptian, Syrian, and Kuwaiti Dialects. Washington, D.C.: Georgetown University Press. Butt, Miriam & Wilhelm Geuder. 2001. “On the (Semi)lexical Status of Light Verbs.” In Semi-lexical Categories: The Function of Content Words and the Content of Function Words, eds. Norbert Cover & Henk van Rijmsdijk. Berlin: Mouton de Gruyter, 323–370. Dahlgren, Sven-Olof. 2008. “The Relevance of Tense and Aspect in Semitic Languages: The Case of Hebrew and Arabic.” In Interdependence of Diachronic and Synchronic Analyses, eds. Folke Josephson & Ingmar Söhrman. Amsterdam: John Benjamins, 221–247. Delancey, Scott. 1991. “The Origins of Verb Serialization in Modern Tibetan.” Studies in Language 15.1: 1–23. Ebert, Karen H. 2000. “Progressive Markers in Germanic Languages.” In Tense and Aspect in the Languages of Europe, ed. Östen Dahl. Berlin: Mouton De Gruyter, 605–654. Elihai, Yohanan. 2004. To Speak Arabic, vol. 2. Jerusalem: Minerva Instruction and Consultation Group (in Hebrew). Eksell, Kerstin. 1995. “Complexity of Linguistic Change as Reflected in Arabic Dialects.” Studia Orientalia 75: 63–74. Esseesy, Mohssen. 2008. “Semantic Bleaching.” In Encyclopedia of Arabic Language and Linguistics, vol. 4, eds. Edzard Lutz & Rudolf de Jong. Leiden: Brill, 160–164. Henkin-Roitfarb, Roni. 2010. Negev Arabic: Dialectical, Sociolinguistic, and Stylistic Variation. Wiesbaden: Harrassowitz Verlag. ———. 2011. “Hebrew and Arabic in Asymmetric Contact in Israel.” Lodz Papers in Pragmatics 7.1: 61–100.


Gamliel and al-Rahman mar ’ i

Hussein, Lutfi. 1990. “Serial Verbs in Colloquial Arabic.” In When Verbs Collide: Papers from the Ohio State Mini-Conference on Serial Verbs, eds. Brian D. Joseph & Arnold Zwicky. (Columbus, Ohio, May 26–27, 1990) Working Papers in Linguistics 39: 340–354. Jastrow, Otto. 2014. “Dialect Differences in Uzbekistan Arabic and Their Historical Implications.” In Alf lahga wa lahga: Proceedings of the Ninth Aida Conference, eds. Olivier Durand, Angela Daiana Langone, & Giuliano Mion. Vienna: LIT Verlag, 205–212. Jayaseelan, K.A. 2004. “The Serial Verb Construction in Malayalam.” In Clause Structure in South Asian Languages, eds. V. Dayal & A. Mahajan. Dordrecht: Kluwer Academic Publishers, 67–91. Kimmig, Rainer. 2014. “Verb-Verb Sequences in Old and Middle Indo-Aryan: A Preliminary Survey.” Paper presented at the Thirtieth South Asian Languages Analysis Roundtable (SALA), Centre for Applied Linguistics and Translation Studies, University of Hyderabad. Kuteva, Tania A. 1999. “On Sit/Stand/Lie Auxiliation.” Linguistics 32.7: 191–213. ———. 2004. Auxiliation: An Enquiry into the Nature of Grammaticalization. Oxford: Oxford University Press. Kuzar, Ron, 1996. “Tavnit ha-peʕalim ha-recifim ha-modaliyyim be-ʕivrit Yisraʔelit.” Lešonenu 68: 119–138 (in Hebrew). Lillas-Schuil, Rosmari. 2006. “A Survey of Syntagms in the Hebrew Bible Classified as Hendiadys.” In Current Issues in the Analysis of Semitic Grammar and Lexicon II, eds. Edzard Lutz & Jan Retsö. Wiesbaden: Harrassowitz Verlag, 79–100. Lord, Carol. 1993. Historical Change in Serial Verb Constructions. Amsterdam: John Benjamins. Palva, Heikki. 1991. “The Form jāk in Bedouin Narrative Style.” Studio Orientalia 67: 55–64. Piamenta, Moshe. 1991. Dictionary of Post-Classical Yemeni Arabic, vol. 2. Leiden: Brill. Rosén, B. Haiim. 1977. Contemporary Hebrew. The Hague: Mouton. Schwarzwald, Ora R. 2001. Modern Hebrew. Munich: Lincom Europa. Simon, Rachel. 2009. “Teach Yourself Arabic—In Yiddish!” MELA Notes 82: 1–15. Smith, Carlota. 1991. The Parameter of Aspect. Dordrecht: Kluwer Academic Publications. Taube, Moshe. Forthcoming. “Verbal Hendiadys in Yiddish.” In Yiddish and Typology, eds. N. Jacobs, H.I. Aronson, & T. Shannon. Watson, Janet C.E. 1993. A Syntax of Ṣanʿānī Arabic. Wiesbaden: Harrassowitz Verlag.

Verbal Predicate Fronting in Modern Hebrew and Yiddish Isaac L. Bleaman

Dept. of Linguistics, New York University, New York, NY, USA [email protected]

Abstract This article discusses the development of verbal predicate fronting (“predicate cleft”) in Modern Hebrew by comparing its properties with those of analogous constructions in Classical Hebrew and Yiddish, a critical contact language. The evidence, largely syntactic, lends support for contact-induced change as a plausible source of verbal predicate fronting in the contemporary spoken variety.

Keywords predicate cleft – syntactic borrowing – substrate effects – Hebrew – Yiddish

Introduction Modern Hebrew allows for verb phrase fronting of a topic or focus.1 In such a construction, one copy of the verbal predicate surfaces with infinitival morphology in the left periphery, while a second, tensed copy is obligatorily * I would like to thank Chris Collins, Edit Doron, Itamar Kastner, Daniel Kaufman, and my classmates at NYU for comments on this and related work. For their grammaticality judgments, I would like to acknowledge Edit Doron and Itamar Kastner (Hebrew), and Binyumen Schaechter and Gitl Schaechter-Viswanath (Yiddish). This article is based upon work supported by the National Science Foundation Graduate Research Fellowship under Grant No. DGE-1342536. 1  Many linguists (e.g., Cozier 2006) have described similar fronting constructions for topic/ focus as instances of “clefting.” Following Landau’s (2006) analysis of Hebrew, and Aboh’s (2006) analysis of Gungbe and other West African languages, I will refer to this construction as “VP-fronting” or “verbal predicate fronting” rather than as predicate cleft. Abbreviations used in this article: 1, 2, 3 = person, abs = infinitive absolute, acc = accusative, expl = expletive, inf = infinitive, prt = particle, sg = singular.

© koninklijke brill nv, leiden, ���6 | doi ��.��63/9789004310896_006



pronounced in its canonical position lower in the clause. This is illustrated in the following example: (1) .‫ היא לא קוראת בכלל‬,‫ לקרוא‬ li-kro hi lo koret bi-xlal to-read she not read at-all ‘As for reading, she does not read at all.’ (cf. Ziv 1997, ex. 2) Although VP-fronting (with doubling) in Modern Hebrew has historical precedents, ultimately I will argue in favor of contact as the more probable source of the construction. The following discussion will first describe the syntactic properties of VP-fronting in Modern Hebrew, and then offer a comparative analysis of parallel constructions in both Classical Hebrew and Yiddish. Finally, the syntactic evidence will be supplemented with historical data from the early decades of revitalized Hebrew.

Verbal Predicate Fronting in Modern Hebrew

As Landau (2006:37) shows, Modern Hebrew permits both “phrasal-infinitive fronting” (PI-fronting), where the verb is fronted along with its internal argument, and “bare-infinitive fronting” (BI-fronting), where the verb alone undergoes fronting. In both cases, a tensed copy of the verb is also pronounced in its canonical position lower in the sentence: (2) .‫ היא קנתה‬,‫לקנות את הפרחים‬ li-knot ʔet ha-praħim, hi kanta (PI-fronting) to-buy acc the-flowers she bought ‘As for buying the flowers, she bought them.’ (Landau 2006, ex. 8a) (3) .‫ היא קנתה את הפרחים‬,‫ לקנות‬ li-knot, hi kanta ʔet ha-praħim (BI-fronting) to-buy she bought acc the-flowers ‘As for buying, she bought the flowers.’ (Landau 2006, ex. 8b) The constituent located at the left edge of the sentence is a bare VP, allowing only for a verb and its arguments and right-adjoined adjuncts:

Verbal Predicate Fronting in Modern Hebrew and Yiddish


(4) .‫ גיל לא תמיד מוריד‬,)‫ להוריד את המים (במהירות‬ le-horid ʔet ha-mayim (bi-mhirut), Gil lo tamid morid to-flush acc the-water (in-speed) Gil not always flushes ‘As for flushing the toilet (quickly), Gil doesn’t always do it.’ (cf. Landau 2006, ex. 10a and footnote 9) Consequently, negation (‫ לא‬lo ‘not’) and sentential adverbs (e.g., ‫ תמיד‬tamid ‘always’), which normally precede the verb (as in sentence [4]), cannot be fronted: (5) .‫ גיל מוריד‬,‫ )*לא) (*תמיד) להוריד את המים‬ (*lo) (*tamid) le-horid ʔet ha-mayim, Gil morid (*not) (*always) to-flush acc the-water Gil flushes ‘As for (*not) (*always) flushing the toilet, Gil does it.’ (Landau 2006, ex. 10b) As the translations suggest, the construction is employed typically (but not exclusively) in contrastive contexts. With these general characteristics of VP-fronting in mind, let us turn to the question of the origins of the construction: is VP-fronting attested in older varieties of Hebrew, and if so, how does Modern Hebrew compare with such precedents?

Comparison with Classical Hebrew

Verbal predicate fronting and doubling is available in Classical Hebrew, although its form and syntactic behavior differ in important ways from VP-fronting in Modern Hebrew. Under the heading “the infinitive absolute,” Gesenius’s Classical Hebrew grammar (1910/1976:342) notes that when this non-finite form appears immediately before the verb, it typically “emphasize[s] . . . either the certainty (especially in the case of threats) or the forcibleness and completeness of an occurrence,” as in Genesis 2:17 ‫ ֥מֹות ָּת ֽמּות‬mōṯ tāmūṯ ‘thou shalt surely die.’ The usage that most closely mirrors the pragmatics of VP-fronting in Modern Hebrew, however, is when the non-finite form “give[s] emphasis to ֽ ֵ ֹ ‫ ָא ֥כֹל ּת‬ʾāḵōl an antithesis” (Gesenius 1910/1976:343), as in Genesis 2:16 ‫אכל‬ tōʾḵēl ‘thou mayest freely eat, but . . . [see previously cited verse].’ Despite this similarity, the infinitive absolute is morphologically distinct from the canonical l-prefixed infinitive of both Classical and Modern Hebrew. Moreover, the



fact that the non-finite verb can surface after the finite copy—often in narratives, wh-questions, imperatives, and injunctives (Harbour 2007:229–234), as in Job 37:2 ‫ ִׁש ְמ ֤עּו ָׁש ֣מ ַֹוע‬šimʿū šāmōʿa ‘hear ye attentively’—is incompatible with the facts of Modern Hebrew and suggests that the form may in fact be a cognate object.2 Harbour (1999) presents additional facts about Classical Hebrew that contrast sharply with the properties of VP-fronting in Modern Hebrew. First, he claims that the Classical Hebrew construction involves head movement (movement of the verb alone), since there is no evidence that complements can be pied-piped along with it. This is different from Modern Hebrew, where phrasal movement is attested (as in sentence [2]). A second point of contrast concerns the possibility of mismatched verbal templates, or binyanim: in Classical Hebrew, the two copies of the verb must be matched for their binyan, unless the copy in the left periphery is in binyan paʿal, in which case it can co-occur with any other binyan later in the sentence—“without any evident difference in meaning” between the two alternants (Harbour 1999:165). The availability of mismatched binyanim is shown in the biblical verses below (6). Modern Hebrew, by contrast, permits no mismatched combinations of binyanim (Landau 2006:49). (6) a. .‫נָ ׂ֥ש ֹא ֶא ָ ּׂ֖שא ָל ֶ ֽהם‬ nāśōʾ ʾɛśśā‌ʾ lā-hɛm carry.abs (paʿal) I.will.carry (paʿal) to-them ‘I will utterly take them away.’ (Hosea 1:6) b. .‫ נָ ׂ֥שֹוא יִ ּנָ ׂ֖שּוא‬ nāśōʾ yīnnāśūʾ carry.abs (paʿal) (nifʿal) ‘They must be carried.’ (Jeremiah 10:5) Since Classical Hebrew is a VSO language, perhaps it is not surprising that arguments are not permitted to intervene between the infinitive absolute and the tensed verb. Negation (‫ לא‬loʾ ‘not’), however, is permitted to intervene (Harbour 2   On the infinitive absolute as the accusative of the internal object, see also Joüon (1923/1996:421–422). Harbour (1999:164–165) argues against the view that the non-finite form is a cognate noun.

Verbal Predicate Fronting in Modern Hebrew and Yiddish


1999:163). This is a point of overlap between the two varieties of Hebrew, since in both cases negation occurs to the left of the tensed copy of the verb, but never before the fronted copy (compare sentences [7] below with [4] and [5] above), leading to the conclusion that in both cases the fronting operation targets a VP constituent not including negation. (7) a. .‫ישֹו‬ ֽ ‫הֹור‬ ִ ‫הֹוריׁש ֥ל ֹא‬ ֖ ֵ ְ‫ ו‬ wə-hōrēš lōʾ hōrīš-ō and-drove.out.abs not drove.out-him ‘And [Israel] did not utterly drive them out.’ (Judges 1:28; Gesenius 1910/1976:342; Harbour 1999, ex. 8b) b. .‫ישֹו‬ ֽ ‫הֹור‬ ִ ‫הֹוריׁש‬ ֵ ‫( *וְ לֹא‬no attested examples of this type)3 *wə-lōʾ hōrēš hōrīš-ō While some languages (such as Gungbe; Aboh 2006:39) disallow sentential negation in instances of verbal predicate fronting and doubling, others (including Yiddish and Russian) display the very same grammaticality facts as the Classical Hebrew data in (7): that is, they permit negation before the tensed copy of the verb but not before the fronted infinitive. Consequently, this point of similarity between Classical and Modern Hebrew cannot be taken as definitive support for verbal predicate fronting as a purely language-internal development, as opposed to a contact-induced change. The next section will highlight some of the similarities between Modern Hebrew and Yiddish in terms of the movement operations involved, lending support to one such alternative contact-based account.

Comparison with Yiddish

Both syntactically and pragmatically, verbal predicate fronting in Yiddish—a Germanic verb-second language—displays many of the same properties that have been outlined for Modern Hebrew. Consider the following alternation:

3   This prediction is assumed from Harbour’s (1999) discussion. However, Gesenius (1910/1976:344) lists three apparent exceptions with negation preceding both copies of the verb.



(8) .‫מאקס עסט ֿפיש‬ ַ Maks es-t fish Max eat-3.sg4 fish ‘Max eats fish.’ ַ ‫ עסן עסט‬ (9) .‫מאקס ֿפיש‬ es-n es-t Maks fish eat-inf Max fish ‘As for eating, Max eats fish.’ (Cable 2004, ex. 1a)

(10) .‫מאקס‬ ַ ‫ עסן ֿפיש עסט‬ es-n fish es-t Maks eat-inf fish Max ‘As for eating fish, Max eats them.’ (Cable 2004, ex. 1b) Like Modern Hebrew, but unlike Classical Hebrew, Yiddish allows for both BI-fronting (sentence [9]) and PI-fronting (sentence [10]). Yiddish, like Modern and Classical Hebrew, also bars the fronting of negation: (11) .‫לאכט זי נישט ָאּפ‬ ַ ‫ּפלאכן ֿפון זיך ַאלײן‬ ַ ‫ אבער (*נישט) ָא‬ ָ ober (*nisht) op-lakh-n fun zikh aleyn lakh-t but (*not) off.prt-laugh-inf from self alone zi nisht op she not off.prt ‘But as for laughing at herself—she won’t.’ (translation of Landau 2006, ex. 16b) As in Modern Hebrew, the Yiddish construction is typically employed in contrastive contexts, though a simple topic reading is also possible, as in the following exchange:

4  The glosses for Yiddish will be more explicitly morpheme-by-morpheme than was the case for Hebrew. This is because agreement and other features of Yiddish verbs are marked by affixes, whereas Hebrew displays the Semitic pattern of changes within the stem, and because person and number features will be shown to undergo fronting (see example [17]).

Verbal Predicate Fronting in Modern Hebrew and Yiddish


(12) Q: ?‫ֿפארסטו ַאהײם‬ ָ ‫ װען‬ ven for-st-u aheym? when homeward ‘When are you going home?’ ַ ‫ֿפאר איך‬ ָ ‫ֿפארן‬ ָ A: .‫נײן ַא זײגער‬ for-n for ikh nayn a zeyger travel-inf I nine o’clock ‘I’m leaving at nine o’clock.’ Finally, the analogous constructions in Yiddish and Modern Hebrew suggest a movement analysis rather than base-generation in the left periphery. The earliest such analysis of the Yiddish pattern comes from Davis & Prince (1986:92–93), who show that “Yiddish verb-topicalization,” as they call it, is constrained by subjacency: (13) a. Extraction from complement clause (crosses single bounding node, S): .‫געזאגט ַאז ער װײסט ַא סך‬ ָ ‫האסטו מיר‬ ָ ‫װײסן‬ veys-n ho-st-u mir gezogt [S az er veys-t a sakh] know-n said that he a lot ‘As for knowing, you told me that he knows a lot.’ b. Extraction from relative clause (crosses two bounding nodes): .‫װאס װײסט ַא סך‬ ָ ‫האב איך געזען דעם ייִ דן‬ ָ ‫*װײסן‬ a sakh]] *veys-n hob ikh gezen [NP dem yid-n [S vos veys-t know-n have I seen the Jew-acc that a lot ‘As for knowing, I saw the man that knows a lot.’ c. Extraction from indirect question (crosses two bounding nodes): .‫געזאגט װער עס װײסט ַא סך‬ ָ ‫האסטו מיר‬ ָ ‫ *װײסן‬ *veys-n hostu mir gezogt [S ver [S es know-n have-2sg-you said who expl veys-t a sakh] ] know-3sg a lot ‘As for knowing, you told me who knows a lot.’ Landau (2006:42–44) uses a number of diagnostics to show that A’-movement gives rise to the syntactic relation between the two copies of the predicate in Modern Hebrew, as well. First, the left copy can cross a finite clause boundary



(assuming successive cyclic movement; sentence [14]), but is also sensitive to the number of bounding nodes (e.g., a complex NP island; sentence [15]).5 (14)

)‫ גיל זכר שהוא הציע שהוא יעזור (לה‬,‫לעזור לרינה‬%

%la-ʕazor le-Rina, Gil zaxar [S še-hu hiciʕa to-help to-Rina Gil remembered that-he offered še-hu yaʕazor (la)] that-he (to.her) ‘As for helping Rina, Gil remembered that he’d offered to help her.’ (Itamar Kastner, p.c.; based on Landau 2006, ex. 21)

(15) .‫ גיל זכר את ההצעה שלו שהוא יעזור‬,‫ ??לעזור לרינה‬ ??la-ʕazor le-Rina, Gil zaxar [NP ʔet ha-hacaʕa šelo to-help to-Rina Gil remembered acc the-offer his [S še-hu yaʕazor] ] that-he ‘As for helping Rina, Gil remembered his offer to help her.’ (Kastner, p.c.) Space constraints prevent me from explicating other overlapping syntactic, prosodic, and discourse-pragmatic properties, including the possibility of fronted copular verbs but not fronted auxiliary verbs (Davis & Prince 1986:95; Landau 2006:41) and the marginal acceptability of certain pairs of non-identical lexical verbs across the sentence (Cable 2004:9; Landau 2006:47–48). However, a discussion of VP-fronting in Yiddish would be incomplete without mentioning a noteworthy difference from the Hebrew. The fronted copy of the predicate, imprecisely glossed as ‘INF’ in most of these examples, was instead glossed as ‘stem-n’ in (13). In most cases, the fronted copy coincides with the canonical citation form of the infinitive (as in ‫ עסן‬esn ‘to eat’ in sentence [9]); however, when an irregular tensed verb is fronted, what appears is neither the true infinitive nor an exact copy of the clause-internal verb, but rather a copy of the stem with the infinitival -n suffix. For this reason, earlier analyses have referred to the fronted constituent as a “pseudoinfinitive” (term from Mark 1978; later in Waletzky 1980):

5  The percent symbol indicates that informants differ on their acceptability judgments. The double question marks indicate that informants uniformly consider the example to be degraded.

Verbal Predicate Fronting in Modern Hebrew and Yiddish


(16) .‫גארנישט‬ ָ ‫*װײסט װײסט ער‬/‫*װיסן‬/‫ װײסן‬ veys-n/*visn/*veys-t veys-t er gornisht know-n/*know.inf/* he nothing ‘As for knowing, he knows nothing.’ (Davis & Prince 1986, ex. 3b) Indeed, some of the fronted “pseudoinfinitives” are lexical items found nowhere else in Yiddish, e.g., binen and izn in the following examples, whose stems bear the person and number features of the tensed verb: (17) a. .‫*זײן בין איך אין ַאמעריקע‬/‫בינען‬ ַ bin-en/*zayn bin ikh in Amerike am-n/*be.inf I in America ‘As for being, I am in America.’ (cf. Davis & Prince 1986, ex. 17b) ַ ‫*זײן איז ער ַא‬/‫איזן‬ ַ b. .‫טשודאק‬ iz-n/*zayn iz er a tshudak is-n/*be.inf he a oddball ‘As for being, he is an oddball.’

The Yiddish “pseudoinfinitive,” which is a morphological puzzle in its own right,6 can be contrasted with the simple infinitive-fronting strategy employed in Modern Hebrew. The difference between the two languages may derive from how they instantiate the spell-out of default (non-finite) tense features (one as stem+n, the other as infinitive). In any case, given the other overlapping properties, this point does not diminish the possibility that Yiddish provided a critical model upon which VP-fronting in Modern Hebrew was based.

Attestation in Early Modern Hebrew

The discussion presented thus far has highlighted the syntactic parallels between VP-fronting in Modern Hebrew and Yiddish, as well as the differences between Modern and Classical Hebrew. However, given that the “revernacularization” of Modern Hebrew was initiated by a learned Jewish elite (Spolsky 2014), a natural question that arises is why the so-called “infinitive absolute” 6  Why the Yiddish predicate is fronted along with the phi-features of the lower copy (as in example [17]) is beyond the scope of this survey article; see Cable 2003 for one proposal.



was not successfully revived. In fact, early fictional texts demonstrate that the older non-finite form was dominant in written narratives (sentence [18]), as well as in the contrastive contexts where VP-fronting might be used today (sentence [19]):7 (18) .‫ וארא שם אשה מכוסה בצעיף‬,‫ הלוך הלכתי לפנות ערב בין הבכאים‬ halox halaxti li-fnot ʕerev beyn ha-bxaim, walk.abs I.walked at-turn evening between the-mulberries ve-ʔere šam ʔiša mexusa be-caʕif and-saw there woman covered in-shawl ‘I walked in the late evening among the mulberry trees, and there I saw a woman covered in a shawl.’ (Mapu 1853, ch. 16) (19)

.‫הּפולנים ִּכ ְּבדו אותו מאד‬ ַ ‫ כי‬,‫ ורק שמוע נשמע‬,‫ ואין איש יודע דבר על–אודותיו‬

ve-ʔeyn ʔish yodeʕa davar ʕal-ʔodotav, ve-rak šamoʕa and-there.isn’t man knows thing about-him and-only hear.abs nišmaʕ, ki ha-polanim kibdu ʔoto meʔod was.heard that the-Poles respected him very ‘And there isn’t anybody who knows anything about him. It was only heard that the Poles respected him greatly.’ (Frishman 1881)

The precise moment when the fronting of VPs (headed by infinitives) began to supplant the Classical V-fronting paradigm is difficult to reconstruct. Considering the dearth of examples in any early written prose, it is likely that VP-fronting entered the language through speech, and was thus less susceptible to conscious correction. It is clear that by 1929, when Avigdor Hameiri penned ‫ בין שיני האדם‬beyn shiney haadam [Between the Teeth of Man], the modern pattern of infinitive-fronting had already entered Hebrew prose and varied alongside the Classical pattern: (20) .‫ ועד היום הזה אינו אוכל כלום‬,‫ אך לאכול לא אכל כלום‬,‫ נתתי לו לשתות עוד ועוד‬ natati lo li-štot ʕod va-ʕod, ʔax le-ʔexol lo ʔaxal I.gave to.him to-drink more and-more yet to-eat not he.ate klum, ve-ʕad ha-yom ha-ze ʔeyno ʔoxel klum nothing and-until the-day the-this he.not eats nothing 7  See Ziv (1997) for a more in-depth treatment of the discourse functions of the Modern Hebrew construction.

Verbal Predicate Fronting in Modern Hebrew and Yiddish


‘I gave him more and more to drink, but as for eating, he didn’t eat anything, and even to this day he still eats nothing.’ (Hameiri 1929, ch. 26) The fact that the infinitive absolute forms are morphologically distinct and used in limited, often highly stylized or liturgical contexts (Krohn 2011:229) may have accelerated their decline. Today, the change has proceeded to completion, such that it is relatively easy to find examples of VP-fronting in colloquial Hebrew (sentence [21]; note the number of arguments) but difficult to find any idiomatic use of the infinitive absolute. (21)

.‫ זה לא יחבב‬,‫הרי כל בר דעת מבין שלחבב את היהדות האורתודוקסית על הציבור הרחב‬

harey kol bar-daʕat mevin še-le-ħabev ʔet indeed every one.with-sense understands that-to-make.fond acc ha-yahadut ha-ʔortodoksit ʕal ha-cibur ha-raħav, ze the-Judaism the-Orthodox on the-public the-wide, it lo yeħabev no will.make.fond ‘Why, anyone with any sense understands that the last thing this would do is make the general public fond of Orthodox Judaism.’ (Persico 2011) Conclusion

It would be premature to identify syntactic borrowing from Yiddish as the lone source of verbal predicate fronting in contemporary vernacular Hebrew. First, it is important to mention that a number of other contact languages (including Russian) also allow for the fronting of infinitives and the spell-out of multiple copies; I chose Yiddish as the focus of this article because of its singular status as the mother tongue of most Hebrew revitalizers and Jewish immigrants from a variety of Eastern European locales (Spolsky 2014). Contact with other VP-fronting languages could plausibly have reinforced the use of the construction in Modern Hebrew. Second, because VP-fronting is attested in a number of languages that clearly played no role in the development of Modern Hebrew (e.g., Brazilian Portuguese and Gungbe), it is impossible to rule out an explanation rooted in linguistic universals: e.g., that the first generation of native Hebrew speakers needed some way of topicalizing or focalizing verbal predicates, and this cross-linguistically common strategy emerged organically.



Again, this could have reinforced the reliance on models from critical contact languages such as Yiddish. Ultimately, the syntactic evidence presented above, combined with the socio-historical reality of Yiddish-Hebrew contact, substantiates an account that invokes syntactic borrowing. References Aboh, Enoch O. 2006. “When Verbal Predicates Go Fronting.” ZAS Papers in Linguistics 46: 21–48. Cable, Seth. 2003. “The Remarkable Yiddish Pseudo-Infinitive: Evidence for Phases?” (Unpublished manuscript, MIT) (accessed March 1, 2014). ———. 2004. “Predicate Clefts and Base-Generation: Evidence from Yiddish and Brazilian Portuguese.” (Unpublished manuscript, MIT) scable/papers/Yiddish-Predicate-Clefts.pdf (accessed March 1, 2014). Cozier, Franz K. 2006. “The Co-occurrence of Predicate Clefting and wh-Questions in Trinidad Dialectal English.” Natural Language & Linguistic Theory 24: 655–688. Davis, Lori J. & Ellen F. Prince. 1986. “Yiddish Verb-Topicalization and the Notion ‘Lexical Integrity.’” In Papers from the 22nd Regional Meeting, Chicago Linguistic Society, eds. Anne M. Farley, Peter T. Farley, & Karl-Eric McCullough, 90–97. Frishman, David. 1881. “be-yom ha-kipurim” [On the Day of Atonement] (in Hebrew). (accessed September 12, 2014). Gesenius, Wilhelm. 1910/1976. Gesenius’ Hebrew Grammar, ed. Emil Kautzsch, trans. A. E. Cowley, 2nd English edition (1910), 13th impression. Oxford: Oxford University Press. Hameiri, Avigdor. 1929. Beyn shiney ha-adam [Between the Teeth of Man] (in Hebrew). (accessed September 12, 2014). Harbour, Daniel. 1999. “The Two Types of Predicate Clefts: Classical Hebrew and Beyond.” MIT Working Papers in Linguistics 34: 159–175. ———. 2007. “Against PersonP.” Syntax 10: 223–242. Joüon, Paul. 1923/1996. A Grammar of Biblical Hebrew, trans. T. Muraoka. Second reprint with corrections (1991). Rome: Editrice Pontificio Istituto Biblico. Krohn, Nitza. 2011. Reading Academic Hebrew: An Advanced Learner’s Handbook. Leiden: Brill. Landau, Idan. 2006. “Chain Resolution in Hebrew V(P)-Fronting.” Syntax 9: 32–66. Mapu, Avraham. 1853. Avahat tsiyon [The Love of Zion] (in Hebrew). http://benyehuda. org/mapu/ahavat_zion_complete.html (accessed September 12, 2014).

Verbal Predicate Fronting in Modern Hebrew and Yiddish


Mark, Yudl. 1978. gramatik fun der yidisher klal-shprakh (A Grammar of Standard Yiddish). CYCO: New York (in Yiddish). Persico, Tomer. 2011. “lama ha-feminizm kol kax mesukan la-yahadut ha-ortodoksit” [“Why Feminism Is So Dangerous to Orthodox Judaism”] (in Hebrew). http://tomer (accessed September 17, 2014). Spolsky, Bernard. 2014. “The Return to Zion and Hebrew.” The Languages of the Jews: A Sociolinguistic History. Cambridge: Cambridge University Press, 249–268. Waletzky, Joshua. 1980. “Topicalization in Yiddish.” In The Field of Yiddish: Studies in Language, Folklore, and Literature, vol. 4, eds. Marvin I. Herzog, Barbara KirshenblattGimblett, Dan Miron, & Ruth Wisse, Philadelphia: Institute for the Study of Human Issues, 237–315. Ziv, Yael. 1997. “Infinitivals Initially: Theme/Topic/Focus.” In Discourse and Pragmatics in Functional Grammar, eds. John H. Connolly, Roel M. Vismans, Christopher S. Butler, & Richard A. Gatward. Berlin: Mouton de Gruyter, 163–175.

Circumstantial versus Depictive Secondary Predicates in Literary Hebrew—The Influence of Yiddish and Russian Keren Dubnov

The Academy of the Hebrew Language and David Yellin College of Education, Jerusalem, Israel [email protected]

Abstract This article discusses the development of circumstantial secondary predicates in late Enlightenment and early Modern literary Hebrew under the influence of adverbial participle constructions in substrate languages, primarily Yiddish and Russian. Circumstantial secondary predicates differ from depictive secondary predicates, which are well-attested in Classical Hebrew.

Keywords participle – adjunct – circumstantial – depictive – secondary predicate – adverbial participle

Introduction Adjunct participial clauses are frequently found in contemporary Hebrew belles-lettres, both in postposition and in preposition to the main clause, as illustrated in (1) and (2) respectively. (1)

]. . .[ ‫ נמנעת מלהביט בעיניו‬,‫חייכתי אליו בנימוס‬ ħiyax-ti ʔel-av be-nimus, nimnaʕat mi-le-habit smiled-1s to-him in-politeness avoid.ptcp.fs from-to-look be-ʕen-av in-eyes-his

* The author is grateful to Edit Doron for her enormous help and advice.

© koninklijke brill nv, leiden, ���6 | doi ��.��63/9789004310896_007

The Influence of Yiddish and Russian


‘I smiled at him politely, avoiding looking in his eyes [. . .].’ (Katzir, Hine ani matħila, 2003) (2) .‫ חשה בלה באושר הקיים בה כממשות‬,‫סובבת בגן כבתוך חלום‬ sovevet b-a-gan ki-vtox ħalom, ħaša wander.ptcp.fs in-the-garden as-inside dream experienced.3.fs bela b-a-ʔošer ha-kayam b-a ke-mamašut Bela in-the-happiness rel-exists in-her as-reality ‘Wandering in the garden like in a dream, Bella experienced the happiness within her as reality.’ (Liebrecht, Mesibat ha-ʔerusin šel ħayuta, 1986). In this kind of sentence structure, the participial secondary predicate further specifies the circumstances of the event described by the predicate of the main clause. It differs in its semantic contribution from the secondary predicates found in the Bible, where the secondary predicate functions as a depictive predicate denoting an attribute of one of the events’ participants: (3) ‫וַ יֵ ֵצא ָה ָמן ַבּיֹום ַההּוא ָ ׂש ֵמחַ וְ טוֹ ב לֵ ב‬ way-yēṣē hāmān b-ay-yom ha-hu śāmēaḥ and-came out Haman in-the-day the-that joyful wə-ṭoḇ lēḇ and-good.cs heart ‘Then went Haman forth that day joyful and glad of heart.’ (Esther 5:9)1 Jackendoff (1990:200–209) syntactically distinguishes in English between adjuncts like the latter, which he calls depictive secondary predicates (typically adjectival phrases), and the former, which he calls gerundive secondary predicates (since in English they are headed by gerunds), but which I call circumstantial secondary predicates on the basis of their semantic function. Recent works in linguistic typology (Schultze-Berndt & Himmelmann 2004; Fabricius-Hansen & Haug 2012) have crosslinguistially substantiated this distinction between depictive and circumstantial secondary predicates. Depictives are often adjectives or adjective phrases (and more rarely other parts of speech, as shown in footnote 2 below), while circumstantials can only be deverbal. Some languages formally distinguish these two constructions, whereas other languages make use of the

1  All of the English translations of Biblical verses are from the JPS, 1917 edition.



same form of the participle in the two constructions. In the latter languages, the differences exist mainly at the semantic and prosodic levels. Depictives are participant-oriented, and form part of the same prosodic unit as the main predicate, whereas circumstantials are event-oriented and typically form a prosodic unit of their own (Schultze-Berndt & Himmelmann 2004:78–79). A useful test for distinguishing between these two types of secondary predicates is based on adding negation to the main clause. A depictive must be within the scope of main clause negation, whereas a circumstantial may escape the scope of negation (Schultze-Berndt & Himmelmann 2004:68– 69). Thus, sentential negation can only be interpreted as negating the depictive happy rather than the main verb in Bill didn’t enter the classroom happy, but can negate the main verb rather than the circumstantial in Bill didn’t enter the classroom, waiting for the teacher. In this article, I claim that unlike in Classical Hebrew, circumstantial secondary predicates have been developed in Modern literary Hebrew under the influence of adverbial participle constructions in Russian and in Yiddish. I focus on the early examples of circumstantial participles in Hebrew literature. I present citations from authentic Hebrew prose as well as translations from Yiddish and Russian where the original texts use adverbial participles (Rappaport:1984) in the circumstantial function.

Depictive Constructions in Classical Hebrew

In the Bible, depictive secondary predicates are usually adjectives, and they follow the predicate of the main clause: ָ ‫וַ יִ גְ וַ ע וַ יָ ָמת ַא ְב ָר ָהם ְב ֵש ָיבה‬ (4) ‫טֹובה זָ ֵקן וְ ָ ׂשבֵ ַע‬ way-ygwaʿ way-yāmot ʾaḇrāhām bə-śēḇā and-expired and-died Abraham in-old age ṭoḇā zāqēn wə-śāḇēaʿ good.fs ‘And Abraham expired, and died in a good old age, an old man, and full of years.’ (Genesis 25:8)

The appearance of a depictive secondary predicate before the predicate is very rare and only documented in poetic speech:

The Influence of Yiddish and Russian


(5) ‘‫ֲאנִ י ְמלֵ אָ ה ָה ַל ְכ ִתי וְ ֵר ָיקם ֱה ִש ַיבנִ י ה‬ ʾăni məlēʾā hālaḵ-tī wə-rēqām hɛ̆šiḇa-nī ʾădonāy I full.fs went-1s and-emptily returned.3sm-me Lord ‘I went out full, and the Lord hath brought me back home empty.’ (Ruth 1:21)2 Depictive secondary predicates may also be participles, typically of intransitive verbs: (6) ‫אתם ִמן ַה ִמ ְצ ָפה הֹלֵ ְך הָ ל ְֹך ּובֹכֶ ה‬ ָ ‫וַ יֵ ֵצא יִ ְש ָמ ֵעאל ֶבן נְ ַתנְ יָ ה ִל ְק ָר‬ way-yēṣē yišmāʿʾel bɛn nətanyā liqrāt-ām and-exited.3ms Ishmael son.cs Nethaniah towards-them min ha-miṣpa holēḵ hāloḵ u-ḇoḵɛ from the-Mizpah go.abs ‘And Ishmael the son of Nethaniah went forth from Mizpah to meet them, weeping all along as he went.’ (Jeremiah 41:6) I have only found one biblical example of a secondary predicate expressed by an active participle of a transitive verb with an obligatory direct object, which may be said to have circumstantial interpretation: ִ ‫יהם ַמ ְכ ִע‬ ֶ ‫וְ זֵ ָרם ֵמ ֵע ֶבר ַלנָ ָהר יַ ַען ֲא ֶשר ָעׂשּו ֶאת ֲא ֵש ֵר‬ (7) ‘‫יסים ֶאת ה‬ wə-zēr-ām mēʿēbɛr lan-nāhār yaʿan ʾăšɛr and-scatter-them beyond to-the-river religion REL ʿāśu ʾɛt ʾăšērē-hɛm maxʿīsīm ʾɛt-ʾadonāy made.3pl acc Asherim-their acc-Lord ‘. . . and will scatter them [people the kingdom of Israel] beyond the River; because they have made their Asherim, provoking the LORD.’ (1 Kings 14:15)

However, it is uncertain that the participle maxʿīsīm predicates people of Israel; it can also be a reduced relative clause modifying the object—Asherim.3 2  In the second part of this example, an adverb appears in a depictive function. Sadka (1997:311) mentions a Mishnaic example with a noun in the same function: ?‫הלומד ילד למה הוא דומה‬ ‘One who learns Torah as a child (in his childhood), what is this comparable to?’ (Avot 4:20). 3  In the following example, the participle has its own subject; therefore, it cannot be analyzed as a circumstantial secondary predicate, but as the main predicate in a circumstantial clause ‫ֹלהים ַהכ ֲֹהנִ ים ז ְֹר ִקים ֶאת ַה ָדם ִמיַ ד ַה ְלוִ יִ ם‬ ִ ‫תֹורת מ ֶֹשה ִאיׁש ָה ֱא‬ ַ ‫וַ יַ ַע ְמדּו ַעל ָע ְמ ָדם ְכ ִמ ְש ָפ ָטם ְכ‬



I conclude that secondary predicates only fulfill the depictive function, and that circumstantial secondary predicates are not attested in the Bible. Circumstantial adjuncts in the Bible are typically clausal, and start with wə- + subject, e.g.: (8) ‫וַ יֵ ָרא ֵא ָליו ה' ְב ֵאֹלנֵ י ַמ ְמ ֵרא וְ הוּא י ׁ ֵֹשב ֶפ ַתח ָהא ֶֹהל ְכחֹם ַהּיֹום‬ way-yērā ʾēl-āw Lord bə-ʾēlonē mamrē and-appeared.3ms to-him ʾădonāy in-terebinths.cs Mamre pɛtaḥ ha-ʾohɛl kə-ḥom hay-yom wə-hū yošēḇ and-he entrance.cs the-tent as-heat.cs the-day ‘And Lord appeared unto him by the terebinths of Mamre, as he sat in the tent door in the heat of the day.’ (Genesis 18:1) Rabbinic Hebrew is not characterized by new types of secondary predicates (except for the one illustrated in footnote 2), but rather by frequent usage of dependent adverbial clauses.4 Arabicized Hebrew in the Middle Ages Participles of some intransitive verbs are used as adjuncts following the main predicate in Medieval translations from Arabic calquing, ḥāl (‫ )حال‬construction in the original text. Here is an example from Maimonides’ More nevoxim (The Guide for the Perplexed) translated from Arabic into Hebrew by Shemuel ibn Tibbon (1224):5

‘And they stood in their place after their order, according to the law of Moses the man of G-d; the priests dashed the blood, which they received of the hand of the Levites’ (2 Chronicles 30:16). 4  For example: ‫ ולא עוד אלא אפילו נכנס כשהוא קורא‬.‫לא ייכנס אדם במבואות המטונפות ויקרא את שמע‬ ‫הרי זה מפסיק עד שיצא מרשות כל אותו מקום ויקרא‬ ‘A person should not enter into dirty lanes and read the Shema. Moreover, even if he already entered it while he was reading, he should stop until he exits from that place and then he should read’ (Tosefta, Berachot 2:19). The expression ʿamad tameah (stood amazed) in late Midrashim, such as Tanḥuma [[wa-yaqhel. 1]], probably developed under the influence of Aramaic. It is depictive and has much in common with ʾamar mitnaṣṣel, which is briefly discussed in the next section. 5  I am thankful to Chanan Ariel for drawing my attention to this example.


The Influence of Yiddish and Russian

(9) ‫ אשר החל ראש הפילוסופים לחקור ולעשות מופתים בעניינים עמוקים מאד אמר‬. . . ‫מתנצל דבר‬

ʾašer heḥel roš ha- p̄ ilosop̄ im laḥqor rel started.3ms head.cs the-philosophers examine.inf we-laʿasot mop̄ etim be-ʿinyanim ʿammuqim meʾod ʾamar and-do.inf evidences in-matters very said.3ms mitnaṣṣel davar thing ‘When the chief of philosophers (Aristotle) was about to inquire into some very profound subjects, and to establish his theory by proofs, he commenced his treatise with an apology.’ (trans. M. Friedlander, 1904)

The term ʾamar mitnaṣṣel translates the Arabic qāl muʕtadiran, where qāl=said.3MS; muʕatdiran=apologize.part.MS.ACC according to the Tibbonian word for ‘word method of translation.’ ʔamar mitnacel had been preserved in Modern Hebrew literature. In this kind of construction, as well as in parallel Arabic ones, the participles seem to be depictive, as they are in classical Hebrew, although they require a more detailed analysis based on additional examples.6 Enlightenment Hebrew The Enlightenment Hebrew prose in the middle of the 19th century provides many examples of secondary predicates with adjectives or participles, but unlike in Biblical Hebrew, they may precede the main clause, as in the following examples: (10) ‫מחריש ומשתאה עמדתי הכן על שפת הנחל‬ maħariš u-mištaʔe ʕamad-ti haxen be and-be amazed. stood-1s ready.abs ʕal sfat ha-naħal on bank.cs the-river ‘Silent and amazed I stood ready on the riverside.’ (Erter, Tašlix, 1840)

6  Some similar examples from Medieval translations from Arabic are quoted in GoshenGottstein (2006:121).



(11) ‫ עזבתי את המקום הזה‬,‫מר בחמת רוחי‬ mar ba-ħamat ruħ-i ʕazav-ti bitter in-anger.cs spirit-my left.1s ha-ze ‘With bitter anger I have left this place.’ (Rumsh, Kur Oni, 1862)

et acc

ha-makom the-place

The earliest writer who used secondary predicates with obligatory complements was, to the best of our knowledge, S.Y. Abramovich (Mendele Moykher-Sforim). Some examples may be interpreted with a circumstantial reading already in his early Hebrew novel Ha-avot ve-ha-banim—Fathers and Sons (1868:152): (12) ‫ מצפה לראות ולשמוע מה אחרית כל‬,‫שעות אחדות ישב דויד לבדו בחדר‬ ‫המדבר בו‬

šaʕot ʔaħadot yašav david levado mecape lirʔot hours several.fp sat.3ms David alone see.inf ve-lišmoaʕ ma aħarit kol ha-midbar b-o and-hear.inf what end.cs all the-conversation about-him ‘David sat in his room alone for several hours, expecting to see and hear what will be the end of all that conversation about him [. . .]’

Applying the negation test to this example shows that sentential negation negates the main verb yašav rather than the secondary predicate mecape. Moreover, the secondary predicate mecape with its complement is not necessarily included in the scope of the negation. Thus, the secondary predicate of this sentence is circumstantial, not depictive. It is particularly interesting to compare Mendele’s later Hebrew prose, which is self-translated from Yiddish, with the original Yiddish text. In Yiddish, as well as in Russian and in some other Slavic languages, there is a special part of speech—an adverbial participle (see for Russian, Rappaport 1984), whose only function is expressing a circumstantial secondary predicate. The adverbial participle is a non-conjugated deverbal part of speech.7 It is typically coreferent with the subject of the main clause. Both in Yiddish and in Russian, there are adverbial participles of present and past tense, but only those of present tense, with the ending –dik (variation: –dig) in Yiddish, –a/ja in Russian, have affected the syntax of the written Modern Hebrew (whereas past a­ dverbial 7  In Slavic languages, it originates in a fossilized participle.

The Influence of Yiddish and Russian


­ articiples are typically paraphrased into subordinate clauses in Hebrew p translations from these languages). Indeed, Hebrew circumstantial participle clauses in Mendele’s Hebrew prose frequently correspond to adverbial participles in its Yiddish parallels. Here is an example from his novel The Mare, which is called Di klyatshe (1873) in Yiddish, and Susati (1909) in Hebrew.8 (13) In Yiddish:

‫זֹוינס?—זאג איְך מיט גרֹויס רחמנֹות אֹויף דער‬ ָ ‫וואס טּוהט זיְך ָדא ַא‬ ָ ,‫מאמע‬ ַ ?‫מאמע‬ ַ ,‫וויינט—וואס וויינסטּו‬ ָ ‫ ווי זי‬,‫ זעהענדיג זי‬,‫מאמען‬ ַ

Mame vos tut zix do azoyns zog ix mit mother what do.2s refl here thus say I with groys raxmones oyf der mamen zeendig zi great pity on the.dat mother.dat see.adv.ptcp her vi zi veint vos veinstu mame how she cries what cry.2s mother ‘Mother, what are you doing?—I say with great pity to my mother seeing her cry—why are you crying, mother?’ In Hebrew:

‫ רואה את אמי בוכיה ומתמלא עליה רחמים—למה‬,‫ מה זאת?—תוהה אני‬,‫אמי‬ ?‫ אמא‬,‫תבכי‬

ʔim-i ma zot? tohe ʔani roʔe mother-my what this I ʔet ʔimi boxiya u-mitmale ʕale-ha acc mother-my cry.part.fs and-become-full on-her raħamim lama tivki ʔima? mercy why cry.fut.2fs mother ‘What is it, mother, I wonder seeing my mother cry and filling with pity toward her—Why do you cry, mother?’

In the Yiddish version, there is one adverbial participle, zeendig (=zeendik) ‘seeing,’ whereas in the Hebrew version, there are two participles in an adverbial function, roʔe ‘seeing’ and mitmale (raħamim) ‘becoming full (of mercy).’ In this example, a depictive interpretation of both Hebrew participles is impossible, and they are clear circumstantials. The same is true for all the examples in the next section. 8  Stiff (1930) noted that adverbial participles with the suffix –endik were not common in Yiddish prose before Mendele. Perhaps it is possible to assume that the frequent usage of adverbial participle phrases in Russian prose affected Mendele’s both in Yiddish and in Hebrew.



Early Modern Hebrew Literature Circumstantial participles are not rare in Bialik’s prose, e.g.: (14) ‫ מעקם פרצופו ומתעטש‬,‫אחי האמצעי רסק מרור במגררת‬ ʔaħ-i ha-ʔemcaʕi risek maror be-migreret brother-my the-middle scraped horseradish in-grater meʕakem parcuf-o u-mitʕateš face-his ‘My middle brother was scraping horseradish on a grater, twisting his face and sneezing.’ (Ha-ħacocra nitbayša—The Trumpet Was Ashamed, 1917) Both in Yiddish and in Russian, adverbial participles may be both postpositive and prepositive to the predicate. However, neither Mendele nor Bialik used circumstantial participle phrases preceding the main clause of the sentence. On the other hand, the authors of the later Hebrew prose, such as Hazaz and Fogel, favored this construction also in sentence-initial position: (15) ‫—סלע סל�ע‬: ‫הופתכת פניה כיוונה לסלע מכורתה וקראה ברחישת געגועים ועצב‬

!‫ שדה הנחושת‬,‫סלע‬

hofexet pane-ha kivna le-sela mexorat-a turn.ptcp.fs face-her aimed.3fs to-rock.cs homeland-her ve-karʔa bi-rħišat bi-rħišat we-ʕecev and-cried.3fs in-feeling.cs longing and-sadness selaʕ selaʕ selaʕ sde ha-neħošet rock rock rock field.cs the-copper ‘Turning away, she addressed her homeland and cried out with sadness and longing: Rock, rock, rock, a field of copper!’ (Hazaz, Bridegroom of Blood, 1929)

In many translations from Russian written in the first third of the 20th century, participle phrases consistently correspond to all the adverbial participle phrases in the original text, both in preposition and in postposition to the main clause. An example from Mordekhay Harizman’s Hebrew version of the Chekhov story “The Steppe” is shown below:9

9  For this text, the same English translation can be suitable both for the original text and for its Hebrew version.

The Influence of Yiddish and Russian


(16) Дрожа в воздухе, как насекомое, играя своей пестротой, стрепет поднялся высоко вверх по прямой линии. Droža v vozduxe kak nasekomoe tremble.adv.ptcp in air.loc like insect igraja svoej pestrotoj strepet play.adv.ptcp its-own.inst colorfulness.inst bustard podn’als’a vverx vysoko po pr’amoj linii rose up high.adv along straight line . . . ‫ הוָ ָתא התרומם למעלה בקו ישר‬,‫רוטט באוויר כרמש ומשחק בשלל גווניו‬ rotet b-a-avir ke-remes u-msaħek in-the-air like-insect bi-šlal gvan-av ha-vata hitromem in-abundance.cs hues-his the-bustard rose le-maʕla be-kav yašar to-upward in-line straight ‘Trembling in the air like an insect and playing with its variegation, the bustard flew up in a straight line.’ Conclusion Circumstantial participle constructions became standard in Modern Hebrew literature during the late 19th and early 20th centuries under the influence of adverbial participles in Yiddish and in Russian (and probably other Slavic languages that have this part of speech). Within the tense system of Modern Hebrew, the classical active participles have been reinterpreted as verbal forms expressing present tense; therefore, their acquiring a non-adjectival function is not a surprise. In spite of the resemblance between depictive and circumstantial secondary predicates in Hebrew, the two can be differentiated from one another. The former are found in the Bible and in Classical Hebrew in general, and it is possible that their superficial similarity inspired the emergence of the l­atter. Circumstantial secondary predicates often precede the main predicate in Modern Hebrew prose.



References Goshen-Gottstein, Moshe Henry. 2006. Syntax and Vocabulary of Mediaeval Hebrew as Influenced by Arabic. Revised by Shraga Assif and Uri Melamed. Jerusalem, Ben-Zvi Institute, Yad Izhak Ben-Zvi, and the Hebrew University of Jerusalem (in Hebrew). Fabricius-Hansen, Cathrine & Dag Haug, eds. 2012. Big Events, Small Clauses. The Grammar of Elaboration. Berlin: Walter de Gruyter. Jackendoff, Ray. 1990. Semantic Structures. Cambridge, MA: MIT Press. Jewish Publication Society. 1917. The Holy Scriptures according to the Masoretic Text. Philadelphia, PA: The Jewish Publication Society of America (JPS). Rappaport, Gilbert C. 1984. Grammatical Function and Syntactic Structure: The Adverbial Participle of Russian. Columbus, OH: Slavica Publishers. Sadka, Isaac. 1997. Studies in Syntax and Semantics. Beer-Sheva: Ben-Gurion University of the Negev Press (in Hebrew). Schultze-Berndt, Eva & Nikolaus P. Himmelmann. 2004. “Depictive Secondary Predicates in Cross-linguistic Perspective.” Linguistic Typology 8.1: 59–131. Stiff, Nahum. 1930. Yidishe Stilistic. Moscow: Kharkov and Minsk, 153–169 (in Yiddish). The Academy of the Hebrew Language. Ma’agarim: The Historical Dictionary of the Hebrew Language. Jerusalem (in Hebrew).

Clausal Periphery

Modern Hebrew še- and Judeo-Spanish ke- (que-) in Independent Modal Constructions Ora (Rodrigue) Schwarzwald and Sigal Shlomo

Department of Hebrew and Semitic Languages, Bar-Ilan University, Ramat Gan, Israel [email protected]; [email protected]

Abstract Modality in Modern Hebrew is expressed in different ways. This article concentrates on one special construction consisting of an independent clause introduced by the particle še followed by a future tense verb, which expresses a variety of modal meanings: desires, wishes, prohibitions, volitions, curses, commands, etc. This means of expressing modality is very common in spoken Modern Hebrew, and can be found in various literary genres. As for its origins, although several suggestions have been proposed, we argue that spoken Judeo-Spanish (the substrate language of the first users of spoken Modern Hebrew in Israel), rather than Yiddish or Russian, is the initial contributor to this widely used construction.

Keywords modality – Judeo-Spanish – Modern Hebrew – subjunctive – language contact – substrate

Introduction: Modal Expressions in Modern Hebrew

Modality in Modern Hebrew is expressed through numerous linguistic devices, lexical as well as grammatical (Boneh 2013). The focus of this article is a particular syntactic construction for the expression of modality: the prefixed particle še followed by a future tense verb (henceforth [še+FUT]) in an independent clause, as in (1). (1)

!‫ֶׁשּיֵ ֵלְך ְּכ ָבר‬

še-yelex kvar! ŠE-go.FUT.3M.SG already! ‘Let him be gone already!’

© koninklijke brill nv, leiden, ���6 | doi ��.��63/9789004310896_008


Schwarzwald and Shlomo

The particle še is commonly used in Hebrew to introduce subordinate clauses, both complement clauses, as in ˀamarti lo še-yelex ‘I told him to go (lit. said.I to.him that-go.FUT.3M.SG),’ and adjunct clauses: ha-sefer še-karati haya mešaˁamem ‘the book that I read was boring.’ When še introduces a subordinate clause, it is commonly translated as that. The modal use of [še+FUT] is unique in introducing a main, rather than subordinate, clause (Bar-Adon 1966). In the present article the uses of the modal construction will be described, and its origins in Modern Hebrew will be examined based on previous Hebrew sources and other languages. The claim will be made that Judeo-Spanish, rather than Yiddish or Russian, is the source for this widely used Modern Hebrew construction.

Modern Hebrew Use of [še+FUT]

The construction [še+FUT] in independent clauses is commonly used both in spoken and in various literary genres. The modal meanings refer to commands (1, 2), admonitions (3a, b, 10), oaths (4), tentative cohortatives (5), negative commands or prohibitions (6), blessings (7, 8), instructions, permission or encouragement (9), etc.1 Because še functions in independent clauses, we do not translate it as ‘that’ but rather leave it untranslated as the form ŠE. (2) !‫ ֶׁשּלֹא יֵ ֵלְך ְל ִאּבּוד‬,‫ֶׁש ַּת ֲחזִ יק אֹותֹו טֹוב‬ še-taħazik ˀoto tov, še-lo ŠE-hold.FUT.2M.SG it.ACC good, that-not yelex le-ˀibud! go.FUT.3M.SG to-lost! ‘Hold it well, lest it be lost!’ (Uri Asaf in the song caˁacuˁeha šel ˀosnat) (3) a. !‫ֶׁש ִּת ְת ַּבּיֵ ׁש ְלָך‬ a. še-titbayeš a. ŠE-shame.FUT.2M.SG a. ‘Shame on you!’


1  Shenhar (1957) lists several Modern Hebrew curses and blessings with allusions to classical sources, many of which start with še. Ben-Amotz (1962:20–22) lists a collection of humorous curses, most of which start with še, e.g., še-teradem ba-Kastel ve-titˁorer be-Hadasa ‘May you fall asleep in the Kastel (a hill on the way to Jerusalem), and wake up in Hadassa (a famous hospital in Jerusalem)!’

Modern Hebrew Še- and Judeo-Spanish Ke- ( Que- )

b. !‫ֶׁשּיִ ְת ַּבּיֵ ׁש לֹו‬ b. še-yitbayeš lo! b. ŠE-shame.FUT.3M.SG to.him! b. ‘He should be ashamed!’ (4) !‫ֶׁש ָּכ ָכה יִ ְהיֶ ה ִלי טֹוב‬ še-kaxa yihye li tov! ŠE-so be.FUT.3M.SG good! ‘Blimey!; Unbelievable!; Trust me!’ (5) ?‫ֶׁשּנַ ְת ִחיל ָל ִׁשיר‬ še-natħil la-šir? ŠE-begin.FUT.1PL to-sing? ‘Should we start singing?’ (6) !‫ֶׂשּלֹא ַּתּגִ יד ִלי ַמה ַל ֲעׂשֹות‬ še-lo tagid li ŠE-not say.FUT.2M.SG ‘Don’t tell me what to do!’ (7) .‫יֹום טֹוב ֶׁשּיִ ְהיֶ ה ְלָך‬ yom tov še-yihye day.M good.M ŠE-be.FUT.F ‘Have a good day.’

ma what

la-ˁasot! to-do

le-xa. to-you.

(8) !‫נּוח ְּב ָׁשלֹום ַעל ִמ ְׁש ָּכ ָבּה‬ ַ ‫ֶׁש ָּת‬ še-tanuaħ be-šalom ˁal ŠE-rest.FUT.3F.SG in-peace on ‘May she rest in peace in her grave!’

miškav-ah! bed-her!

(9) .‫ִמי ֶׁשּטֹוב לֹו וְ ָׂש ֵמ ַח ֶׁשּיִ ְר ַקע‬ mi še-tov lo ve-sameaħ še-yirkaˁ. who that-good to.him and-glad ŠE-stomp.FUT.3M.SG. ‘Whoever is well and glad should stomp (his feet).’ (Datia Ben-Dor in the song mi šetov lo vesameaħ) (10) .‫ לֹא ֲאנִ י‬,‫ֶׁשהּוא יִ ְת ַּבּיֵ ׁש‬ še-hu yitbayeš, lo ŠE-he shame.FUT.3M.SG, not ‘He should be ashamed, not me.’

ˀani. I.



Schwarzwald and Shlomo

The construction [še+FUT] is flexible in two senses: a. it may be preceded by other sentential material, as in (7, 9); b. the prefix še can be separated from the future tense verb by complements or by overt subjects, as in (4, 6, 10). As can be seen in the examples above, the use of this construction in the future tense is not restricted to the second person. It can be used in any other person for cohortative (self-encouragement) or jussive (wishes for other parties) purposes as well. In (1, 3b, 4, 7, 9, 10) it is used in the masculine third person singular, in (5) in the first person plural, and in (8) in the feminine third person singular. The subject of these sentences does not need to be human, as in (7).

Compared Constructions in Former Layers of Hebrew

The use of [še+FUT] constructions in ancient Hebrew texts is rare. In one of Bar Kosba’s letters, the following sentence occurs: (11) ‫שלום שתשלח תבי חמשת כורין חטין‬ šalom še-tišlaħ tavi(?) ħamešet peace ŠE-send bring(?) five korin ħitin kors (=dry measures) wheat ‘Peace, send (to) bring five measures wheat’ (Kutscher 1962:11) This sentence states a request and is built in the same way as the Modern Hebrew expression. In rabbinic literature, še occurs only in embedded clauses, but has a modal flavor which is not found in the Bible: (12) .‫זה בנין בית המקדש שיבנה במהרה בימינו אמן‬ ze binyan bet ha-mikdaš this building.of house.of the-sanctuary še-yibbane bi-mhera be-yame-nu ˀamen. that-be.built.FUT.3M.SG soon in-days-our amen. ‘This is the building of the Temple that (we wish it) will be built soon in our times. Amen.’ (Talmud Yerushalmi, Taanit 4:7) Rabbi Shlomo Itsħaki (Rashi; 1040?–1105), from France, uses the sentence given below in his interpretation of the Talmud (13). Rashi adds še to the future form

Modern Hebrew Še- and Judeo-Spanish Ke- ( Que- )


in an independent clause, whereas late Ashkenazi rabbis omit this from the same sentence. Rashi knew Old French, and because the [que+PRES/SUBJ] construction can express modality in French,2 it is clear that he was influenced by the French syntactic structure. (13) .‫יֹוד ֶיעּנָ ה‬ ִ ‫בּואה ֶׁש‬ ָ ‫ּיֹוד ַע ְּת‬ ֵ ‫ָּכל ִמי ֶׁש‬ kol mi še-yodeaˁ tevuˀa every who that-knows cereal.F.SG še-yodiˁ-ena. ŠE-announce.FUT.3M.SG-ACC.F.SG. ‘Whoever recognizes cereal should announce it.’ (Rashi, Babylonian Talmud, Menaħot 64b) Rabbi David Kimħi (RaDaK; 1160–1235), a medieval Provençal grammarian and commentator of Spanish ancestry, also uses [še+FUT] in an independent clause in his biblical commentary: (14) .‫ ֶׁש ֵּת ַדע ִּכי יָ ַד ְע ִּתי ַהּכֹל‬,‫ָּכל ַמה ֶּׁש ָע ִׂש ָית‬ kol ma še-ˁasita, še-tedaˁ all what that-do.PAST.2M.SG ŠE-know.FUT.2M.SG ki yadaˁti hakol. that know.PAST.1SG everything. ‘Whatever you did, you may know that I knew everything.’ (RaDaK, Ps. 50:21) Avineri (1964: 533, 1979:1229) calls the particle še in (13)3 ‫שי"ן טפלה בלשון ציווי‬ šin tfela bilšon civuy “a secondary unimportant šin (še) which functions as an imperative,” and claims that this same še can be found in Midrashic literature and in medieval interpretations, as in (15): ָ ְ‫עֹולם ֵאל ּגָ דֹול ו‬ ָ ‫ֶׁשּיִ ּוָ ַדע ָלְך ִא ִּמי ֶׁשּיֵ ׁש ָּב‬ (15) .‫נֹורא‬ še-yivadaˁ lax ˀimm-i ŠE-acknowledge.FUT.3M.SG mother-my še-yeš ba-ˁolam ˀel gadol ve-nora. in.the-world God big and-terrifying 2  The command yǝhī ˀōr ‘let there be light’ (Gen. 1:3) is translated into French as Que la lumière soit! 3  Avineri erroneously includes še-tedaˁ ‘so that you will know’ in Rashi’s interpretation of 1 Samuel 1:15 as belonging to this construction.


Schwarzwald and Shlomo

‘You should acknowledge, my mother, that there is in the world a big and terrifying God.’ (Otsar midrashim [Eisenstadt], Avraham Avinu 3) According to Even-Shoshan (2003:548), the expression še-yiħye ‘may he live’ (as well as še-tiħye ‘may she live’) in independent clauses originated in the Middle Ages, but the examples he cites are from Modern Hebrew literature. We found in the Reponsa Project (2014) that the addition of še-yiħye to names or titles of people was used mostly by rabbis who spoke Romance languages—Spanish, Judeo-Spanish, French, or Italian. The use of še-yiħye by Ashkenazi rabbis is much rarer. The addition of parenthetical še-tiħye is analogous to the use of še-yiħye. Early examples using [še+FUT] in Modern Hebrew can be found in Brener’s book šxol vexišalon (Bereavement and Failure, 1910), written after his immigration to Israel (16), and in his story nedudim “Wandering” (17), written before he immigrated. (16) . . . ‫בכל אופן שיימשך כך לעולם‬ be-xol ˀofen, še-yimašex In-every way, ŠE-continue.FUT.3M.SG ‘Anyway, let it last this way forever . . . ’ (šxol vexišalon, part one, chapter one)

kax so

le-ˁolam . . .  forever . . . 

(17)  . . . ‫ לא כיום‬,‫ שיהיה בית הכנסת מכובד היטב‬,‫מחר בבוקר כבואי הלום‬ maħar ba-boker ke-voˀ-i halom, Tomorrow in.the-morning as-come-I here, še-yihye bet-ha-kneset mexubad hetev, ŠE-be.FUT.3M.SG the-synagogue cleaned properly, lo ka-yom . . .  not as-today ‘tomorrow morning as I come here, let the synagogue be cleaned, not like today . . . ’ (nedudim) The use of [še+FUT] is also attested to in a book published in 1932 by a school teacher who collected data in the 1920s (Fectorit 1932:123):4 4  Yael Reshef found this example for us. Keren Dubnov supplied the examples from Brener.

Modern Hebrew Še- and Judeo-Spanish Ke- ( Que- )


(18) !‫הספר! שתהיי כבר בריאה‬-‫שתבואי מחר לבית‬ še-tavoˀi maħar le-vet-hasefer! ŠE-come.FUT.2F.SG tomorrow to-school! še-tihyi kvar briˀa! ŠE-be.FUT.2F.SG already healthy! ‘Come to school tomorrow! Be well!’ These examples prove that this construction was well known at the beginning of the 20th century.

Compared Constructions in Other Languages5

Kutscher claims: “It is believed that ‫ ֶׁש‬+ imperfect denoting the jussive, e.g., ‫ ֶׁשּיֵ ֵלְך‬, ‘let him go’ is a Yiddish calque, but this requires clear-cut proof” (1982:215). Avineri (1964:533) and Netzer (2007:229) assert that Yiddish is the source for ָ (German sollen) ‘must, necessary’ this construction, based on the verbs ‫זאלען‬ and ‫לאזען‬ ָ (German lassen) ‘let happen, ought.’ These verbs are translated into ָ ‫[ מע‬me zol nit visn Hebrew by [še+FUT] forms, e.g., ‫זאל ניט וויסען פון ַאזוינע צרות‬ fun ˀazoyne tsores] > ‫[ ֶׁשּלֹא נֵ ַדע ִמ ָּצרֹות ָּכ ֵא ֶּלה‬še-lo nedaˁ mi-tsarot ka-ˀele] ‘We shouldn’t know about such troubles’ (Matisoff 2000:47). The use of these auxiliary verbs in Yiddish is followed by an infinitive form, far from the [še+FUT] construction. The same applies to Russian. The particle Пусть pusṯi or puskay, which originates from the verb pustiṯi (or puskaṯi), meaning ‘to free, allow, let,’ is the standard way in which to express a wish. The structure chtob/chtoby ‘in order to, what’ followed by a subjunctive particle and a verb in the past tense is also used to express a wish, especially for curses or commands. This structure also occurs in subordinate clauses indicating wishes and intentions. In a high register, the particle da, derived from the verb meaning ‘to give,’ can be followed by a verb in third person to express modality. Thus, Yiddish and Russian may carry the same meanings expressed in examples (1–10), but their syntax is very different from that of the Hebrew expressions.

5  We are grateful to Keren Dubnov and Michael Ryzhik for clarifying the facts about Russian.


Schwarzwald and Shlomo

Judeo-Spanish Parallel Constructions

As in Hebrew, modality is expressed in Judeo-Spanish (and in Spanish) in a variety of ways, lexical as well as grammatical. The use of subjunctives and imperatives in the verbal system enables the modal expression in addition to the syntactic means (Bunis 1999:312–392; Seco 1992; Batchelor & San José 1910). The particle ke (Sp. que) ‘that’ is used in subordinate clauses in the same ways as še in Hebrew, but, in addition, it is used with the present subjunctive form to express modality in independent clauses. Grammar books tend to ignore this practice, but Spanish dictionaries cite these usages (DRAE, que, meanings 17–19; Nehama 1977:283). Here are a few examples: (19) ke vos kante: kwando KE DAT.2PL sing.PRES.SUBJ.1SG when el rey Nimrod? the king Nimrod? ‘Should I sing for you (the song beginning with) “When the King Nimrod”?’ (cf. 5) (20) anyáða bwéna ke tengáš! year good KE have.PRES.SUBJ.2PL! ‘May you have a good year!’ (=‘Happy New Year’) (cf. 7) (21) ke se mwéra! KE REFL die.PRES.SUBJ.3SG! ‘May he die!’ (=‘I wish him dead’) (cf. 3b) (22) el ke va vinír, he who go.PRES.3SG come.INF, ke vénga! KE come.PRES.SUBJ.3SG! ‘Whoever intends to come, let him come!’ (cf. 9) (23) azlaxá ke te méte successH KE DAT.2SG put.PRES.SUBJ.3SG el dyo! the God! ‘May God grant you success!’ (cf. 10)

Modern Hebrew Še- and Judeo-Spanish Ke- ( Que- )


(24) ke no dígas náða KE not say.PRES.SUBJ.2SG anything enfrénte de el! in.front of him! ‘Don’t say anything in front of him!’ (cf. 6) (25) gan éðen bwéno ke Garden.of EdenH good KE ‘May he/she rest in peace!’ (cf. 8)

ténga! have.PRES.SUBJ.SG!

(26) ken se axárva kon sus mános, who REFL hit with his hands, ke no yore. KE not cry.PRES.SUBJ.3SG. ‘Whoever hits himself with his own hands should not cry.’ (proverb) (cf. 9) (27) ke vénga el Mašíax, KE come.PRES.SUBJ.3M.SG the messiahH, ke séa en mwéstros días. KE be.PRES.SUBJ.3M.SG in our days. ‘May the messiah come, may it be in our days.’ (cf. 10) (28) bóka de león ke te kóma mouth of lion KE REFL eat.PRES.SUBJ.3SG i no óžo de benaðám! and not eye of personH! ‘It is better to be eaten by a lion than by people’s eyes (beware of jealousy)!’ (proverb) (cf. 10) As can be seen in these few examples, the structure is parallel to Hebrew: ke is followed by a subjunctive form in independent clauses [ke+PRES.SUBJ]. This structure can occur at the beginning of an utterance, as in (19, 21, 24, 27), or it can be preceded by other sentential material, as in (20, 22, 23, 25, 26, 28). It is also used in negative commands, as in (24). All these usages are similar to the uses in (1–10, 16–18), except that in Judeo-Spanish the subjunctive form is used instead of the Hebrew future form.


Schwarzwald and Shlomo

The “Internal Development Hypothesis”

Another possible explanation could be given to the occurrence of modal [še+FUT] independent constructions. Peretz (1942:297, 1962:73–74) claims that in spoken languages, speakers tend to shorten their utterances; hence, instead of ˀani mecavxa laševet ‘I order you to sit,’ speakers omit the preceding words and say just laševet ‘sit!’ The same applies to the sentence še-yelex beˁacmo ‘let him go by himself,’ which has changed from a subordinate clause into an independent clause in speech. The explanation in favor of the omission fits Zadka’s viewpoint within transformational grammar regarding deep performative verbs that are omitted from the surface structure, e.g., (ˀani mekalel ˀotxa) šetelex laˁazazel ‘(I curse you) that you (will) go to hell’ (1978:34, 1981:36–45). According to Netzer (2007:229–230), this omission occurs especially in emotional expressions. The process of change from a subordinate clause to an independent clause by the omission of performative verbs is possible, but this still leaves one question unsolved: why were these constructions absent in biblical and early rabbinic literature until they started being used sporadically in the Middle Ages, especially in Romance-speaking areas, and began to flourish in spoken and written Modern Hebrew as early as the beginning of the twentieth century? Conclusion We contend that Judeo-Spanish, the substrate language of most Hebrew speakers in Palestine at the end of the 19th century and the beginning of the 20th century, influenced Modern Hebrew in a variety of ways (Schwarzwald 1993, 2013), including the syntactic structure under discussion. The first adopters of Modern Hebrew as a spoken language were Judeo-Spanish speakers who lived in the big cities of Palestine: Jerusalem, Safed, Hebron, and Tiberias. They used Hebrew for communication, but their substrate language was Judeo-Spanish, and the syntax of this structure was naturally transferred into Hebrew. The resemblance between the Judeo-Spanish use of [ke+PRES.SUBJ] and Hebrew [še+FUT] in independent modal clauses is beyond accident. Future tense forms in Hebrew carry modal meanings, and they are parallel in meaning to the subjunctive verbal forms in Judeo-Spanish. The use of the Judeo-Spanish ke as the first constituent of both subordinate and independent modal clauses was used analogously in Hebrew; thus, the use of še in subordinate clauses was expanded to independent modal clauses in Hebrew as well. Without the Judeo-Spanish substratum and the close contact of its speakers with other

Modern Hebrew Še- and Judeo-Spanish Ke- ( Que- )


Modern Hebrew speakers, we doubt if this construction could have become so widely used in Modern Hebrew. References Avineri, Isaac. 1964. Yad ha-lašon (The Language Monument). Tel Aviv: Izre’el (in Hebrew). ———. 1979. Hexal rashi (Rashi’s Palace). Vol. 1. Jerusalem: Mossad Harav Kook (in Hebrew). Bar-Adon, Aharon. 1966. “New Imperative and Jussive Formation in Contemporary Hebrew.” Journal of the American Oriental Society 86: 410–413. Batchelor, Ronald A. & Miguel Ángel San José. 1910. A Reference Grammar of Spanish. Cambridge: Cambridge University Press. Ben-Amotz, Dahn. 1962. ˀex la-ˁasot ma? (How to Do What?). Tel Aviv: Amikam (in Hebrew). Boneh, Nora. 2013. “Mood and Modality: Modern Hebrew.” In Encyclopedia of Hebrew Language and Linguistics, vol. 2, eds. Geoffrey Khan et al. Leiden: Brill, 693–703. Bunis, David M. 1999. Judezmo: An Introduction to the Language of the Sephardic Jews of the Ottoman Empire. Jerusalem: Magnes. DRAE. Diccionario de la Lengua Española—Vigésima segunda edición, Real Academia Española, (accessed January 10, 2014). Even-Shoshan, Abraham. 2003. Ha-milon he-ħadaš (The New Dictionary). Tel Aviv: Am Oved, Kineret Zmora-Bitan, Dvir and Yediot Aharonot (in Hebrew). Fectorit, A. 1932. Mi-pinati: rešimot mi-tox ha-ˁavoda be-xita ˀalef (From My Corner: Notes from Work in First Grade). Tel Aviv: Psiot (in Hebrew). Kutscher, Eduard Y. 1962. “lešonan šel ha-ˀigrot ha-ˁivriyot ve-ha-ˀaramiyot šel bar kosba (The Language of the Hebrew and Aramaic Letters of Bar Kosba),” Lešonénu 26: 7–23 (in Hebrew). ———. 1982. A History of the Hebrew Language. Jerusalem & Leiden: Magnes & Brill. Matisoff, James A. 2000. Blessings, Curses, Hopes, and Fears: Psycho-ostensive Expressions in Yiddish. Stanford: Stanford University Press. Nehama, Joseph. 1977. Dictionnaire du Judeo-Espagnol. Madrid: CSIC. Netzer, Nissan. 2007. Hebrew in Jeans: The Image of Hebrew Slang. Beer-Sheva: BenGurion University Press (in Hebrew). Peretz, Yitzhak. 1942. “lilšon hayladim vehaˁam (Children’s and People’s Language).” Lešonénu 11: 296–300 (in Hebrew). ———. 1962. ˁivrit ka-halaxa (Hebrew as Required). Tel Aviv: Shreberk (in Hebrew). Seco, Rafael. 1992. Manual de Gramática Española. Madrid: Aguilar. Schwarzwald (Rodrigue), Ora. 1993. “Remnants of Judeo-Spanish in Modern Hebrew.” Pe’amim 56: 33–49 (in Hebrew).


Schwarzwald and Shlomo

———. 2013. “Between Hebrew and Judeo-Spanish.” hed ha-ulpan he-ħadash 101: 17–28 (in Hebrew). Shenhar, Isaac. 1957. “kilelat hedyot; ve-zot ha-braxa (Layman’s Curse; And This Is the Blessing).” Lešonenu Laˁam 8: 157–169, 253–260 (in Hebrew). Zadka, Yitzhak. 1978. Taħbir ha-mišpat (Sentence Syntax). Jerusalem: Akademon (in Hebrew). ———. 1981. Taħbir ha-ivrit be-yamenu (Hebrew Syntax in Our Days). Jerusalem: Kiryat Sefer (in Hebrew).

Modern-Hebrew lama-še Interrogatives and Their Judeo-Spanish Origins Itamar Francez

University of Chicago, Chicago, IL, USA [email protected]

Abstract This article describes a Modern Hebrew interrogative construction, not found in earlier varieties of the language, in which a wh-word is followed by a clause headed by the complementizer še ‘that.’ When that clause contains negation, the resulting sentence has the illocutionary force of a suggestion, with the opposite polarity to that of the complementizer clause. In this case, negation fails to license negative concord and negative polarity items. The main properties of the construction are described, an analysis is sketched, and evidence is given indicating Judeo-Spanish as the probable source for the construction.

Keywords interrogative – mood and modality – polarity – rhetorical question – presupposition – speech act – Judeo-Spanish

Introduction Modern Hebrew (MH) shows a contrast between two kinds of interrogative structures, exemplified in (1), the title of a newspaper article, and its constructed variant (2). The two sentences form a minimal pair, differing only in the occurrence of the complementizer še.

* I am grateful to Edit Doron for illuminating criticism and discussion of the issues discussed in this chapter. I have also benefited from discussion with and suggestions from Karlos Arregi, Jerry Sadock, Larry Horn, and Barbara Partee. I thank Aldina Quintana for her help with the Judeo-Spanish data. This chapter is dedicated to the memory of my paternal grandparents, Shlomo Francez and Rachel Mordechai, whom I never knew, but who would have been excellent consultants on Balkan Judeo-Spanish.

© koninklijke brill nv, leiden, ���6 | doi ��.��63/9789004310896_009




lama ha-zahav tamid yihiye ʔatraktivi? why the-gold always be.3msg.fut attractive ‘Why will gold always be attractive?’ (calcalist,, accessed July 24, 2008) (2) lama še ha-zahav tamid yihiye ʔatraktivi? why that the-gold always be.3msg.fut attractive ‘Why should gold always be attractive?’

For convenience, I refer to (1) as a lama-interrogative and to (2) as a lama-še interrogative. The two constructions differ in interpretation. The sentence in (1) expresses a question asking for reasons for the truth of what is taken to be an established fact (in (1), that gold will always be attractive). The sentence in (2) expresses a (often rhetorical) question, asking for motivations for a particular way of resolving an issue that is, crucially, considered to be unsettled at the time of utterance (in (2), whether or not gold will always be attractive). The focus of this article is negative lama-še interrogatives, such as (3) and (4), which express suggestions. I term such interrogatives suggesterogatives, alluding to Sadock’s (1971) coinage queclaratives for interrogatives that have declarative force. (3) lama še lo tagiš ʔet ʔexad why that neg submit.2msg.fut acc one šelxa ke-teza? yours as-thesis ‘Why don’t you submit one of your books as a thesis?’ (title of a blog post by journalist Raviv Druker,, accessed May 1, 2015)

ha-sfarim the-books

(4) ʔaz ʔim ʔanaxnu kvar kan, ʔaz lama še lo then if we already here, then why that neg nexayex ve-nište ʔeyze te vradim ʕim nana. smile.1pl.fut and-drink.1pl.fut which tea.cs roses with mint ‘So if we’re here, then why don’t we smile and drink some sort of rose-tea with mint.’ (found in Tal Linzen’s Israblog corpus,, accessed May 1, 2015) Strikingly, the suggestion made in (3) and (4) is of the opposite polarity to that of the sentence in the še-clause, i.e., despite the presence of negation, these sentences express a positive suggestion to submit a book or drink tea. Suggesterogative readings of lama-še interrogatives are diagnosable by the fact that beseder ‘ok’ is a natural response to them. This diagnostic seems to


modern hebrew lama-Še interrogatives

show that neither lama-interrogatives nor positive lama-še interrogatives have suggesterogative readings. Thus, neither (1) nor (2) can be followed up with beseder, whereas this is a perfectly natural response to (3) or (4). Negative lama-še interrogatives on their suggesterrogative interpretation have a parallel in English, in the form of why-don’t questions. (5) a) Why don’t you sit down a minute? b) Why don’t the children sit over there at the smaller table? In the next section, I list the main descriptive properties characterizing positive and negative lama-še interrogatives, distinguishing them from lama interrogatives. While še can occur with other wh-words in Hebrew, in which case the resulting structure shares some of the basic properties of the lama-še construction, none of these cases give rise to suggesterrogative readings.

Properties of lama-še Interrogatives

As mentioned, the two constructions differ in the kind of context they can be uttered in. lama- interrogatives call for a context in which some proposition is assumed (by both interlocutors) to be a fact. In contrast, lama-še interrogatives call for a context in which the truth of the relevant proposition is assumed to be unsettled. In such a context, asking for reasons why the issue of whether the proposition is true or not should be settled in one way rather than another leads to the implication that such reasons do not exist. Under certain circumstances, elaborated below, this implication gives rise to the illocutionary force of a suggestion. The lama-še interrogatives, unlike lama-interrogatives, are restricted to the future tense, as shown by the contrast between (6a)1 and (6b). This is unsurprising given that lama-še interrogatives ask about an unsettled issue. The past and present are by nature already settled. (6) a) *lama še hu why that he

lo neg

yašan / slept.3ms

yašen sleeps

ba-salon. in.the-living room

b) lama hu lo yašan / yašen ba-salon? why he neg slept.3ms sleeps in.the-living room ‘Why didn’t he sleep / isn’t he sleeping in the living room?’ 1  The * symbol in front of (6a) is the convention used in contemporary theoretical linguistics to indicate ungrammaticality. See examples (3) and (4) above for exemplification of lama-še interrogatives with the future tense.



Negation in lama-še interrogatives on a suggesterogative reading cannot license negative concord elements such as klum ‘nothing’ or negative polarity items such as bixlal ‘at all.’ For the difference between negative concord and negative polarity, see Keren (2015). Thus, the sentences in (7) cannot be interpreted as suggestions, but only as rhetorical questions, in response to an assertion denying the complement of the complementizer.2 (7) a)


lama še lo toxal klum? why that neg eat.fut.2msg nothing ‘Why should you not eat anything?’ NOT: ‘Why don’t you eat?’ lama še lo toxal bixlal? Why that neg eat.fut.2msg at.all ‘Why should you not eat anything at all?’ NOT: ‘Why don’t you eat?’

This property of lama-še suggesterogatives is shared with their English counterparts. English suggesterogatives also cannot feature negative polarity items, as shown in (8), where (8b) is not interpretable as a suggestion.3 (8) a) Why don’t you eat something? b) #Why don’t you eat anything? Whether a negated lama-še interrogative receives a suggesterogative interpretation depends on properties of the context. In particular, in a context in which an assertion has been made about how a metaphysically unsettled issue will be resolved, a lama-še interrogative can be used to take issue with that assertion. This is exemplified in the dialogue in (9): 2  An example of a natural context for, e.g., (7a) is as a response to (i): (i)

hi ʔamra še ʔani lo ʔoxal klum. she said.3fs that I neg eat.fut.1sg nothing ‘She said that I will not eat anything?’

3  This pattern has been known in the literature, going back at least to example (ii), attributed to Barbara Partee by Laurence Horn (Partee 2005). (ii) a) Why don’t you love me some more? (invitation) b) Why don’t you love me anymore? (question)

modern hebrew lama-Še interrogatives


(9) A: lo nirʔa li še ha-kelev yišan neg seems that the-dog sleep.fut.3msg ba-salon. in.the-living room ‘I don’t think the dog is going to sleep in the living room.’ B: lama še hu lo yišan ba-salon? why that he neg sleep.fut.3msg in.the-living room ‘Why shouldn’t he sleep in the living room?’ B’s response in (9) is not interpreted as a suggestion that the dog sleep in the living room (such a suggestion would be an odd one to make to someone who has just asserted the opposite), but instead as a rhetorical question, implying that the dog might well sleep in the living room. In a context in which A has not made her assertion and in which the issue at hand (or Question under Discussion, see, e.g., Ginzburg 1994; Roberts 1996) is where the dog should sleep, B’s assertion is interpreted as a suggesterogative, suggesting that the dog sleep in the living room. These two interpretations are distinguished by intonation. In both cases, the context of utterance is one in which it is unsettled where the dog will sleep. The suggesterogative reading of lama-še interrogatives arises only for negative sentences. Thus, (10), the title of an article in an online journal, cannot be interpreted as a suggestion, and, if it were uttered, could not be followed up by beseder, even though it does gives rise to the implication that we should not care whether or not Obama is reading our emails. (10) lama še yihye lanu ʔixpat ʔim ʔobama why that be.3msg care if Obama kore ʔet ha-meylim šelanu? reads acc the-mails of.1pl ‘Why should we care if Obama is reading our mails?’ (alaxson,, accessed June 13, 2013) Similarly, (11), which was written as a response to a question about the compatibility of a device with a certain package, has only declarative force and cannot be followed up with beseder. The suggestion that the addressee not buy the package is a perlocutionary, rather than illocutionary, effect.



(11) ʔaval lama la-ʕazʔazel  še tikne ʔet but why to.the-hell  that buy.2ms. fut acc ha-maʔaraz haze? hu ʕole faking ʔelef šekel! the-package this? it costs fucking thousand NIS ‘But why the hell would you buy this package? It costs a fucking thousand dollars!’ (, accessed May 1, 2015) The declarative force of positive lama-še interrogatives is further exemplified by the sentence in (12), a classic Sadock queclarative. The particle naxon ‘right’ generally embeds declarative sentences to form, roughly, a yes-no question. (12) ʔamarti lexa lo likro ʔaval lama told.1sg neg read.inf but why še takšiv li naxon? that listen.3ms.fut right ‘I told you not to read, but why listen to me, right?’ (Tal Linzen’s Israblog corpus,, accessed May 1, 2015) The lama-še interrogatives on their rhetorical, non-suggesterogative reading license free choice items that lama interrogatives do not, as shown in (13): (13) a. lama še ʔey-paʕam tišan ba-salon? why that ever sleep.fut.3msg in.the-living room ‘Why should you ever sleep in the living room?’ b. *lama ʔey-paʕam tišan ba-salon? why ever sleep.fut.3msg in.the-living room ‘*Why will you ever sleep in the living room?’ This is unsurprising, given the unsettledness of the truth of the propositions asked about in a lama-še interrogative in contrast with the settledness of the proposition asked about in a lama interrogative. When the subject of the sentential complement of lama-še is first or second person, the complementizer can be dropped. This is not possible with thirdperson subjects. (14) lama (še) lo nišan / tišan ba-salon? why that neg sleep.fut.1pl sleep.fut.2msg in.the-living room ‘Why don’t you / we sleep in the living room?’

modern hebrew lama-Še interrogatives


The Origins of lama-še Interrogatives

lama-še interrogatives are just lama interrogatives that embed a še-clause. This še-clause is precisely the structure that Schwarzwald & Shlomo (2015) discuss as the [še+fut] construction. As they show, the [še+fut] clause can appear as a matrix clause and has a range of irrealis readings involving the notions of hortative and optative modality. Schwarzwald & Shlomo demonstrate that [še+fut] entered Modern Hebrew through Judeo-Spanish. The Judeo-Spanish correspondent of [še+fut] is [ke+pres.subj], and it too can stand alone as a matrix clause with the relevant set of meanings. In fact, similar constructions are to be found in a whole range of Balkan and southern European languages (see Ammann & Van der Auwera 2004). Since [ke+pres.subj] can function as a matrix clause, it is expected that the language should be able to embed such a clause under a wh-word meaning ‘why’ and achieve the structure of a lama-še interrogative. This expectation is fulfilled. The following Judeo-Spanish example is from Luria (1930), reported by Kahane & Soporta (1951):4 (15) Pur ke k’afirmi? why that that-sign.1s.fut ‘Why should I sign?’ Here, the wh-word is clearly followed by a clause headed by the complementizer ke, and the structure is both syntactically and semantically identical to the lama-še interrogative. Furthermore, Judeo-Spanish also features the suggesterogative interpretation for negative [ke+pres.subj] interrogatives. Example (16) was reported to me by Aldina Quintana. (16) de ké no te asentas? why that neg cl.2s sit.2s.fut ‘Why don’t you sit?’

4  The datum in (15) comes from oral materials collected by Luria in Monastir (present-day Bitola). Aldina Quintana (p.c.) points out that the use of pur ke (and por ke) instead of the more common de ke is probably a feature of the Monastir dialect. The Jews of Monastir were, together with the rest of Macedonia’s Jews, transported to Treblinka and murdered there. Consequently, very little is known about this dialect.



Finally, Judeo-Spanish allows other wh-words to embed ke-clauses, as shown in (17), also from Kahane & Soporta. (17) ande ké la meta? where that 3f put.1s.fut ‘Where should I put it?’ Exactly the same structure is found in Modern Hebrew, as mentioned briefly in the last sentence of the introduction. (18) ʔefo še ʔani ʔasim ʔet ze? where that I put.fut.1sg acc it ‘Where should I put it?’ While more research is required to draw definitive conclusions, it seems highly probable that lama-še interrogatives entered Hebrew through Judeo-Spanish, together with the matrix [ke+pres.subj] construction.

The Suggesterogative Reading

In this section, I sketch an analysis of the interpretation of lama-še interrogatives, and the suggesterogative reading associated with negative ones. A key component of any satisfactory analysis is an explanation for the failure of negative lama-še interrogatives to license negative concord and negative polarity items. I propose that lama-še interrogatives are just lama interrogatives, i.e., ‘why’ questions, which ask for a reason for the truth of the complement of the wh- element. In this case, that complement is a [še+fut] structure, which expresses, roughly, weak bouletic necessity.5 To exemplify the analysis, I use the simple examples in (19), based on (14) above.

5  Bouletic modality has to do with what is possible or necessary given the desires, preferences, or wishes of a discourse participant. Weak bouletic necessity (Kratzer 1991) is truth in the worlds that best fit the relevant wishes, desires, or preferences. As noted above and discussed by Schwarzwald & Shlomo, the Hebrew [še+fut] construction on its own can express more specific notions of hortative and optative mood (e.g., wishes and curses). A semantic account of the modality of the [še+fut] construction is far beyond the scope of a short paper. See Boneh (2013) for a recent discussion of the expression of various types of mood and modality in Modern Hebrew.

modern hebrew lama-Še interrogatives


(19) a. lama še lo tišan ba-salon. why that neg sleep.2ms in.the-living room ‘Why don’t you sleep in the living room.’ b. lama še tišan ba-salon? why that sleep.2ms. in.the-living room ‘Why should you sleep in the living room?’ First consider (19a). My proposal is that the suggesterrogative interpretation of this sentence arises when the negation lo scopes outside the [še+fut] clause, as part of a why not? question asking for reasons against the truth of the [še+fut] clause, paraphrasable roughly as “you should sleep in the living room.”6 The proposed logical form for (19a) is (20). (20) Why-not [SHOULD (you sleep in the living room)] Given this logical form, the failure to license negative concord and negative polarity items is unremarkable. Negative concord items require the presence of negation in the same clause, and negative polarity items, on standard accounts following Ladusaw (1979), require a downward entailing environment.7 Neither of these is provided by the [še+fut] clause in (20). The suggesterogative reading arises as an indirect speech act through a pragmatic convention. When the issue at hand is where, among several salient possibilities, the addressee should sleep, a speaker uttering the question schematized in (20) indicates that she does not know, and cannot think of, any reasons against the truth of “you should sleep in the living room,” and therefore that she believes “you should sleep in the living room.” In situations in which it would be unreasonable for the hearer to assume that the speaker is genuinely asking her a question (for example, if it is up to the speaker to decide where the hearer should or shouldn’t sleep), the speaker is therefore interpreted as asserting what she believes, namely “you should sleep in the living room,” and thus, through a meta-linguistic convention of usage, indirectly making a suggestion.8 6  Recall that should here is to be read as weak necessity, i.e., as equivalent to “it is recommended / preferred that.” 7  A downward entailing environment is, very roughly, one which licenses inferences from general cases to stronger, more specific ones. For example, Eve didn’t eat a fruit entails Eve didn’t eat an apple. 8  A full explication of this account requires a theory of indirect speech acts. See Gordon & Lakoff (1971), Sadock (1974), Searle (1975), and, for the view of indirect speech acts as metalinguistic conventions of language use, Morgan (1977).



Recall that, as shown in (14), the suggesterogative reading is not obligatory and not conventionally encoded by (19a), which can also be uttered to ask a question. This happens in contexts in which it has already been suggested that the addressee should not sleep in the living room. The speaker can then genuinely ask the addressee for the motivations behind that suggestion, implicating conversationally that there are no good motivations. In this case, I propose, negation is interpreted within the [še+fut] clause, yielding the logical form in (21). (21) Why [SHOULD (you not sleep in the living room)] As discussed above for (7) and (8), in genuine why questions, polarity and concord items are licensed, as expected if negation is interpreted inside the [še+fut] clause. The implication that arises from an utterance of (19a) with the meaning in (21) is that the speaker does not endorse the truth of “you should not sleep in the living room,” and therefore, that as far as she is concerned, you are free to sleep there. This implication does not amount to a suggestion to sleep there, only to a relief from an obligation not to. This is why, on this reading, (19a), like (19b), cannot be followed by beseder, a response that indicates the endorsement of a suggestion. Conclusion This article has laid out the properties of the lama-še interrogative in Modern Hebrew. Such interrogatives differ from regular lama interrogatives mainly in that (a) they ask about motivations for a particular resolution of a yet-unsettled issue, rather than reasons for the truth of an established fact, and (b) they express suggestions rather than questions in their negative form. It was argued that the interpretation of lama-še interrogatives, including their suggesterogative reading, follows from the meaning of the še-clause they embed. Finally, strongly suggestive, if inconclusive, evidence was presented that lama-še interrogatives entered Hebrew through Judeo-Spanish. References Ammann, Andreas & Johan van der Auwera. 2004. “Complementizer-headed Main Clauses for Volitional Moods in the Languages of South-Eastern Europe: A Balkanism?” In Balkan Syntax and Semantics, ed. Olga Mišeska Tomić. Amsterdam: John Benjamins, 293–314.

modern hebrew lama-Še interrogatives


Boneh, Nora. 2013. “Mood and Modality.” In The Encyclopedia of Hebrew Language and Linguistics, vol. 2, eds. Geoffrey Khan et al. Leiden: Brill, 693–703. Ginzburg, Jonathan. 1994. “An Update Semantics for Dialogue.” In Proceedings of the 1st International Workshop on Computational Semantics, ed. Harry Bunt. Tilburg: ITK, Tilburg University. Gordon, David & George Lakoff. 1971. “Conversational Postulates.” In Papers from the Seventh Regional Meeting of the Chicago Linguistics Society, eds. Douglas Adams, Mary Ann Campbell, Victor Cohen, Julie Lovins, Edward Maxwell, Carolyn Nygren, & John Reighard. Chicago IL: Chicago Linguistic Society, 63–84. Kahane, Henry R. & Sol Saporta. 1953. “The Verbal Categories of Judeo-Spanish.” Hispanic Review 21.3: 193–214. Keren, Einat-Haya. 2015. “From Negative Polarity to Negative Concord—Slavic Footprints in the Diachronic Change of the Hebrew Items meʔuma, klum and šum davar.” Journal of Jewish Languages 3.1: 183–198. Kratzer, Angelika. 1991. “Modality.” In Semantics: An International Handbook of Contemporary Research, eds. Armin von Stechow & Dieter Wunderlich. Berlin: De Gruyter, 639–650. Ladusaw, William A. 1979.  Polarity Sensitivity as Inherent Scope Relations. PhD Diss., University of Texas at Austin. Luria, Max A. 1930. “A Study of the Monastir Dialect of Judeo-Spanish Based on Oral Material Collected in Monastir, Yugoslavia.” Revue Hispanique 79: 323–583. Morgan, Jerry L. 1977. “Two Types of Convention in Indirect Speech Acts.” In Syntax and Semantics, 9: Pragmatics, ed. Peter Cole. New York: Academic Press, 261–280. Partee, Barbara. 2005. “Reflections of a Formal Semanticist as of Feb 2005.” Unpublished ms, online at Roberts, Craig. 1996. “Information Structure in Discourse: Toward an Integrated Formal Theory of Pragmatics.” In OSU Working Papers in Linguistics, Vol 49: Papers in Semantics, eds. Jae H. Yoon & Andreas Kathol. Columbus, OH: The Ohio State University Department of Linguistics, 91–136. Searle, John. 1974. “Indirect Speech Acts.” In Syntax and Semantics (Vol. 3): Speech Acts, eds. Peter Cole & Jerry S. Morgan. New York: Academic Press, 59–82. Sadock, Jerrold M. 1971. “Queclaratives.” In Papers from the Seventh Regional Meeting of the Chicago Linguistics Society, eds. Douglas Adams, Mary Ann Campbell, Victor Cohen, Julie Lovins, Edward Maxwell, Carolyn Nygren, & John Reighard. Chicago IL: Chicago Linguistic Society, 223–232. ———. 1974. Towards a Linguistic Theory of Speech Acts. New York: Academic Press. Schwarzwald, Ora & Sigal Shlomo. 2015. “Modern Hebrew še- and Judeo-Spanish ke(que-) in Independent Modal Constructions.” Journal of Jewish Languages 3.1: 91–103.

Colloquial Modern Hebrew Doubly-marked Interrogatives and Contact with Arabic and Neo-Aramaic Dialects Samir Khalaily

Al-Qasemi College and Zefat Academic College, Israel [email protected]

Edit Doron

The Hebrew University of Jerusalem, Israel [email protected]

Abstract This article describes the innovative DMI construction—doubly-marked interrogative—of colloquial Modern Hebrew, in which a question is doubly marked as interrogative. A DMI consists of two parts: (i) an ordinary question, which we call the content question, and (ii) an additional wh-phrase, the attitude marker, which embeds the content question, and whose function is to assign it additional illocutionary force, typically that of rejecting a presupposition salient in the discourse. The article suggests that the DMI was (re-)innovated in Modern Hebrew as a result of contact with Modern Arabic and Neo-Aramaic dialects. It may have been previously innovated in an earlier stage of Hebrew due to its contact with Aramaic.

Keywords doubly-marked interrogative (DMI) – attitude marker – rhetorical question – presupposition – question under discussion (QUD) – Hebrew – Arabic – Neo-Aramaic * The Palestinian Arabic examples in the article are part of a study presented by the first author in conferences at the Al-Qasemi College (June 10, 2012) and the Zefat College (February 4, 2013). He would like to express his gratitude to the audiences (which included his Baqa alGharbiyya and Umm il-Fahim students) for their questions and suggestions, and to thank his informants Mustafa Sa’di, Khalid Ghanayyim, Ra’if Khalaily, Aida Zbedat, and Khalid Zbedat, all from Sakhnin. The second author gratefully acknowledges the fellowship from the Mandel Scholion Interdisciplinary Research Center in the Humanities and Jewish Studies at the Hebrew University of Jerusalem.

© koninklijke brill nv, leiden, ���6 | doi ��.��63/9789004310896_010

Colloquial Modern Hebrew Doubly-Marked Interrogatives


Introduction Modern Hebrew has been in contact with Modern Arabic dialects since the early stages of its revival, first Palestinian Arabic and later the Jewish Arabic dialects spoken by immigrants to Israel (see Mar’i 2013:119–162 and Henshke 2013 respectively, and references cited therein). Modern Hebrew has also been in contact with Neo-Aramaic dialects of the immigrants to Israel from the Kurdish areas of northern Mesopotamia (Khan 2011; Mutzafi 2014). Common to the Modern Arabic and Neo-Aramaic dialects, and now also to Modern Hebrew, is a doubly-marked interrogative construction (DMI), not previously discussed in the linguistic literature, and not mentioned in the grammars of Classical Arabic, Syriac, or Classical Hebrew. The DMI construction is interesting both in its special syntax and in the intricate relation between its semantics and its distinct pragmatic function. The article is structured as follows. We first present Hebrew examples and explain the function of DMI. Second, we discuss possible sources of the DMI in Modern Hebrew. We argue that the Modern Hebrew DMI emerged due to contact with Arabic and Aramaic dialects. However, Modern Hebrew also contains frozen vestiges of an older DMI construction originating from Aramaic. We conjecture that at some point, the DMI might have existed as a common feature of the Classical Central Semitic languages, preserved mostly in the Arabic and Aramaic dialects. The DMI in Colloquial Modern Hebrew Like many languages of the world, Modern Hebrew forms a constituent question both by rising intonation and by placing a wh-phrase in front of the clause. Yes-no questions, on the other hand, are typically formed by rising intonation only, and do not have a special syntax distinguishing them from declarative sentences. In recent years, a new construction—the DMI—has emerged in colloquial Modern Hebrew, expressing a novel type of complex question. The construction consists of a wh-phrase (which we call the attitude marker) embedding an ordinary question (which we call the content question), the latter either a constituent question or a yes-no question. The attitude marker and the content question form an amalgamated interrogative clause—the DMI, pronounced with the intonation contour of a single question. Thus, though a DMI is often introduced by two wh-phrases, it forms a single interrogative clause. In the DMI in (1a), for example, the content question what happened and the ­attitude


Khalaily and Doron

marker why form together a single interrogative clause, pronounced with continuous rising intonation into a single peak. This is very different from the intonation contour of the corresponding sequence of two separate interrogative clauses shown in (1b). The latter consists of two separate stretches of rising intonation into two high intonation peaks: (1) a. DMI ?‫למה מה קרה‬ lama ma kara? ‘Why what happened?’ b. Sequence of interrogatives ?‫למה? מה קרה‬

lama? ma kara? ‘Why? What happened?’ Semantically, too, the DMI is very different from a sequence of interrogatives. We illustrate this by examining both (1a) and (1b) in a context in which they are preceded by the echo question, “She should clear the table?”: (2a) DMI

?‫שתפנה מהשולחן? למה מה קרה‬

še-tefane me-ha-šulħan? lama ma  kara? ‘that-will.clear.3fs from-the-table? why  what happened? ‘She should clear the table? What happened? (and why assume that what happened would be reason enough to induce her to clear the table?)’ (, accessed December 10, 2014) (2b) Sequence of interrogatives

?‫שתפנה מהשולחן? למה? מה קרה‬

še-tefane me-ha-šulħan? lama? ma  kara? ‘that-will.clear.3fs from-the-table? why?  what happened? ‘She should clear the table? Why? What happened?’ In the DMI in (2a), the attitude marker why introduces a rhetorical question that expresses rejection of the obligation to clear the table. In (2b), the independent question Why? simply queries the reason for the obligation. It is true that asking for a reason often conversationally implies rejection, but in (2a) the rejection is conventionalized, having become part of the conventional meaning of the construction. The DMI is mostly found in colloquial oral speech, including informal web chats and blogs. It has very recently also found its way into journalistic writing, and

Colloquial Modern Hebrew Doubly-Marked Interrogatives


even into literary works—though still typically confined to direct speech in these contexts (the earliest printed examples that we have seen are from the 1990s). We informally sketch the semantics/pragmatics of the DMI as follows: – The content question is either a genuine quest for information or a rhetorical question. – Irrespective of whether the content question is originally genuine or rhetorical, the attitude marker assigns it the (additional) function of a rhetorical question. Similarly to rhetorical questions in general, the DMI has a strong speaker-oriented force and typically denotes a sense of negation (Sadock 1971, 1974; Krifka 1995; Han 1998, 2002). The disapproving function of some wh-phrases has also been noted for Chinese by Yang (2007). – In many examples, the attitude marker is why. Its function is to endow the content question with additional illocutionary force, that of rejecting either a salient presupposition that had been added to the common ground by the addressee, or the QUD (question under discussion) currently in the discourse. It does so by asking the addressee the rhetorical question, “Why assume the presupposition/QUD”?1 The following is an example from a blogger’s discussion of a driver’s rude behavior. The content question Who are you? attributed to the driver is rhetorical, and implies that the addressee (a pedestrian trying to cross the street at a crosswalk) is not a noteworthy individual. The attitude marker why endows the content question with the additional function of rejecting the implicit presupposition that she should stop at the crosswalk for a mere pedestrian. (3) ‫ משקפי שמש‬,‫ ברך מורמת על המושב‬,‫נהגת הג׳יפ שלא עוצרת במעבר חצייה‬ naheget driver.f.cs ħaciya crossing miškefey glasses.cs

ha-jip the-Jeep berex knee šemeš sun

. . . ‫ למה מי אתה‬,‫וסיגריה‬

še-lo ʕoceret be-maʕavar that-neg stops at-passage.cs muremet ʕal ha-mošav raised on the-seat ve-sigarya lama mi ʔata . . . and-cigarette why who you.m . . .

1  On the notion of QUD (question under discussion), see Ginzburg (2012). For an additional type of speech act performed by the use of a rhetorical why in Modern Hebrew see Francez (2015).


Khalaily and Doron

‘the Jeep driver who doesn’t stop for you at a crosswalk, her knee raised on the seat, with sun glasses and cigarette. Who are you? (and why do you assume the driver would stop for you?)’ ( E%D7%A4%D7%97%D7%99%D7%93/, accessed October 17, 2014) Another example is from the writer Sayed Kashua’s weekly column in the Israeli daily newspaper Haaretz: (4) ?‫למה מתי הוא ידע להנות מההצלחה זה‬ lama matay hu yadaʕ le-henot why when he knew to-enjoy me-ha-haclaħa ze from-the-success ‘When did this one know how to enjoy success? (and why expect he would this time?)’ (Haaretz, October 5, 2014) Kashua is a novelist, a bilingual speaker of Palestinian Arabic and Hebrew. In this particular example, he reports the (fictional) words of his mother, a speaker of Palestinian Arabic. As in the previous example, the content question is a rhetorical question. It implies that ‘this one’ (her son Kashua) never knew how to enjoy success. The attitude marker implies that there is no reason why the addressee, Kashua’s father, should presuppose that Kashua would enjoy his success on the occasion at hand. The next example is from the novel Dead Fish in Jaffa, in which the writer Dan-Benaya Seri reports a dialogue with a woman of the “Old Yishuv,” the Jewish community in Palestine, which lived in close contact with speakers of Palestinian Arabic in the days of the Ottoman Empire. In this example, the QUD Where? is explicitly put forward in the discourse, and the DMI rejects its being a valid issue by turning it into a rhetorical question with an obvious answer. (5) ?‫ לאן? למה איפה יקחו את החמור הזה‬.‫לקחו את המנוול שלי מהבית והלכו‬ laqħu ʔet took.3mpl acc ve-halxu and-left.3mpl ʔet ha-ħamor acc the-ass

ha-menuval šeli the-bastard mine leʔan lama ʔeyfo where why where ha-ze bet-mešugaʕim the-this house-lunatics


me-ha-bayit from-the-house yikħu would.take.3mpl

Colloquial Modern Hebrew Doubly-Marked Interrogatives


‘They took that bastard of mine from the house, and left. / Where? / Where would they take this ass? To the loony bin (and why assume there might be another option?)’ (Dan-Benaya Seri, Dead Fish in Jaffa, 2003:87) What is interesting about the next example is that the content question is not a rhetorical question but an ordinary informative where-question, querying about the whereabouts of the addressee (both the author and the addressee are schoolchildren); indeed, the next move in the dialogue is the addressee’s answer. Only when embedded under the attitude marker why does the question acquire a rhetorical dimension; it expresses rejection of the criticism implied in the previous question. (6) ‫ בכיתה ומתכתב? שלחתי עם סמיילי מזועזע‬.‫ החזיר‬,‫ בכיתה‬.‫אז איפה אתה? שאלתי‬

‫ אז שבי בשקט‬. . .‫למה איפה את? בכיתה‬

ʔaz ʔeyfo ʔata šaʔalti b-a-kita then where you.msg asked.1sg in-the-class heħzir b-a-kita ve-mitkatev šalaħti replied.3msg in-the-class and-corresponding sent.1sg ʕim smayli mezuʕazaʕ lama ʔeyfo ʔat with smiley shocked why where you.fsg b-a-kita ʔaz švi be.šeket in-the-class then sit.2fsg quietly ‘So where are you? I asked./ In class, he responded./ You are corresponding while in class? I sent with a shocked smiley./ Where are you? (and why assume that your location is more appropriate for SMSing than mine?)/ In class . . . / Then shut up.’ ( ‫‘ סיפורי אהבה‬Love stories,’ accessed September 17, 2014)2

So far, we have seen that the content question can be headed by a variety of wh-words such as what, who, when, and where (which and how are attested as well). It can also be a yes-no question: (7) ‫ למה נראה לך‬.‫בברזיל הקים בית חרושת לטקסטיל "וגם הייתי מבריח סחורות‬ be-brazil in Brazil “ve-gam

heki:m established.3msg hayiti mavriyaħ

bet.xarošet factory sħorot

"?‫שמעבודה אפשר לחיות‬

le-tekstil for-textile lama nirʔa

2  Unfortunately, as of November 2014, the text is no longer available at this site.


Khalaily and Doron

and-also was.1sg smuggling goods why seems l-ax še-me-ʕavoda ʔefšar li-ħyot?” to-you that-from-work possible to-live ‘In Brazil, he established a textile factory “and I also used to smuggle goods. Does it seem to you that it is possible to make a living by working?” (and why assume that smuggling goods is not an acceptable way to make a living?)’ (Haaretz, September 5, 2014, p. 20 of the Galeria section) This example is interesting for two reasons: 1.


The why-marker is ambiguous. One reading, the one given above, is the DMI reading with a yes-no content question. But since a yes-no question has the same syntax as a declarative clause, there is an additional reading, one in which why is understood as embedding a declarative rather than an interrogative clause. According to the latter reading, the whyquestion is an ordinary question querying why the addressee believes that it is possible to make a living by working. The latter reading is disfavored in this particular example. The author of the article switches mid-sentence from reported to direct speech (as witnessed by the switch from third to first person). This facilitates the use of the colloquial DMI construction. The colloquial nature of DMI is also attested by the impossibility of replacing the colloquial whyword lama with maduaʕ, which is the formal-register why-word.

There are also literary examples in which the DMI is not within direct speech but is part of the writer’s prose. In such examples, the construction is used sarcastically. The rejection expressed by the attitude marker is facetious: The writer actually shares the presupposition / the QUD, and only pretends not to. This is interesting from a sociolinguistic perspective, since it is the colloquial nature of the construction that indicates to the reader that the writer’s words should actually not be taken at face value. One sarcastic example is from a restaurant review criticizing a particular restaurant for its Khraime (traditional fish dish in a rich tomato sauce). The content question What did you think?, addressed to the readers, challenges, when embedded under the attitude marker, the readers’ assumption that Khraime would be a rich sauce rather than a mere cumin-spiced tomato paste. The use of the construction is clearly facetious, as the critic obviously shares the readers’ assumptions about Khraime.

Colloquial Modern Hebrew Doubly-Marked Interrogatives



‫ מקבלים רסק עגבניות‬.‫החריימה הבהירה מה קורה כשמשדכים רסק עגבניות לכמון‬ ?‫ למה מה חשבתם‬.‫עם כמון‬

ha- ħrayme hivhira ma kore kše-mešadxim the-Khraime demonstrated what happens when-combine.3mpl resek  ʕagvaniyot le-kamun mekablim  resek  paste tomatoes to-cumin get.3mpl paste ʕagvaniyot ʕim  kamun  lama ma   ħašavtem? tomatoes with cumin why what think.2mpl? ‘The Khraime demonstrated what happens when one combines tomato paste with cumin. One gets tomato paste with cumin. What did you think? (and why expect anything else?)’  (Haaretz, October 10, 2014, Restaurant Review, p. 18 of the Galeria section)

An additional sarcastic example is found in the title of a Haaretz article by the novelist David Grossman: (9) ?‫למה מי מת‬ lama mi met? why who died ‘Who died? (and why assume this is of consequence?)’ (Haaretz, February 24, 2012) The rhetorical content question implies that nobody (of consequence) died. The attitude marker rejects the concern that some of the readers were bound to be having at the time about the death a few days earlier of a Palestinian detainee, Omar Abu Jariban, while in Israeli custody. Grossman’s use of the construction is clearly facetious, as the whole point of his article is to enhance the public concern.3 Since the DMI affects the speech-act performed by the content question, it is typically found in main clauses, where it can directly relate to the speech situation and access the discourse presuppositions and QUDs. In this respect, it differs both from multiple questions and from conjoined questions, which are easily embeddable. Another difference is that the various wh-phrases in multiple and conjoined questions are all part of the content question. A 3  The article was translated to Arabic: in March 3, 2012 (accessed July 26, 2014). Although, naturally, the translation is to Modern Standard Arabic, the title is translated to Palestinian Arabic le:š mi:n ma:t ‘Why who died?’, since DMI is a construction found in Arabic dialects, but not in the standard language.


Khalaily and Doron

third difference is that in a multiple question, the wh-phrases are not stacked at the beginning of the clause (cf. [10a] below); and in a conjoined question (cf. [10b]), though the wh-phrases are all clause-initial, they are conjoined rather than stacked.4 (10a) ‫האתר הזה יאפשר לנו לדעת מי אוכל מה איפה‬ ha-ʔatar ha-ze yeʔafšer la-nu the-site the-this will-make-possible to-us la-daʕat mi ʔoxel ma ʔeyfo to-know who eats what where ‘This site will permit us to know who eats what where.’ (, accessed October 17, 2014) (10b) ?‫איך ולמה זה קורה להם‬ ʔex ve-lama ze kore la-hem how and-why this happens to-them ‘How and why does this happen to them?’ (Haaretz, theater review, October 10, 2014, p. 1 of the Galeria section) The DMI in Modern Arabic Dialects The DMI is a general feature of Modern Arabic dialects: (11) Palestinian Arabic le:š šu: štare:t why what bought.2ms ‘What did you buy? (and why assume it was nothing?)’ ʔami:ṣ blu:ze w-banṭalo:n shirt blouse and-trousers ‘A shirt, a blouse, and trousers.’ (, accessed July 26, 2014)

4  English has multiple questions and conjoined questions, but not the DMI construction. Superficially, one may find two non-conjoined clause-initial wh-phrases in English, too, but the second one only scopes over part of the clause, e.g., in the title of Dov Seidman’s book Why How We Do Anything Means Everything.

Colloquial Modern Hebrew Doubly-Marked Interrogatives


(12) Syrian Arabic A: ana ra:yiħ baddak ši: I going want.2ms thing ‘I am going. Do you want anything?’ B: šu: we:n ra:yiħ what where going ‘Where are you going?’ A: le:š inti šu: daxal-ek why you what concern-your ‘How does this concern you? (and why bother?)’ (, accessed January 12, 2012) (13) Lebanese Arabic le:š mi:n ʔa:l γe:r he:k why who said other.than that ‘Who said anything different than that? (and why do you assume such an option?)’ (, accessed April 11, 2010) (14) Egyptian Arabic ana le: mi:n fi d-dunya baʕd ħabi:b-i I why who in the-world after lover-my ‘Who do I have in the world other than my lover? (and why assume I would?)’ ( html, accessed July 29, 2009) Of the various dialects, Palestinian Arabic is the most accessible to us. It is the native language of the first author, who still remembers the words of his deceased mother when she would hear one of her children complaining: (15) Palestinian Arabic le:š šu: sa:yer ʕal-e:k? why what is.happening on-you ‘What is happening to you? (and why assume that complaining would help?)’ We also rely on the testimony of native speakers of Jewish Arabic dialects. There are scarcely any recordings of these dialects, and none that are a­ vailable


Khalaily and Doron

to us. Nevertheless, it is possible to elicit the DMI construction in Jewish Arabic dialects. Native speakers who were given the context of example (8) above were happy to produce the following DMIs in their dialects: (16) Moroccan Arabic, Casablanca Jewish dialect ʕəlas ʔas dhələkom why what ‘What did you think? (and why expect anything else?)’5 (17) Tunisian Arabic, Jewish dialect las sa fi-bal-kom why what ‘What did you think? (and why expect anything else?)’6 Since DMIs are not known as such in previous stages of Hebrew, nor in its European contact languages, we conclude that Arabic is probably a source of the Modern Hebrew DMI. Speakers of Hebrew whose native language was a dialect of Arabic may have been instrumental in introducing the DMI to Hebrew. Indeed, we find that Arabic allows a wider range of DMI types than Hebrew, both with respect to lexical options and structure. For example, Palestinian Arabic allows how as an attitude marker, which is not attested in Modern Hebrew:7 (18) ki:f ʔe:š rašid how what Rashid ‘What did Rashid buy?’8

ištara bought

Structurally, we find content questions in Palestinian Arabic DMIs which are multiple questions, a construction not attested in Hebrew: 5  We are grateful to Eli Ohayon for this example. 6  We are grateful to Yehudit Henshke for this example. 7  Palestinian Arabic also allows what as an attitude marker, and so does Modern Hebrew: (i) ?‫מה ממתי את אוהדת הפועל‬ ma mi-matay ʔat ʔohedet ha-poʕel what from-when Ha-Po’el ‘Since when are you an Ha-Po’el fan?’ ( .php?f=5&t=25035&start=25, accessed December 10, 2014). 8  This example was constructed by the first author.

Colloquial Modern Hebrew Doubly-Marked Interrogatives

(19) A: B:


ka:n fi: mufa:jaʔa:t ha:y l-le:li was there surprises this the-night ‘There were surprises tonight.’ le:š mi:n γalab mi:n why who defeated who ‘Who defeated whom? (and why assume that there is anything unexpected here?)’9

We also find imperative clauses replacing content questions within DMIs in Palestinian Arabic, for which Hebrew has no counterpart:10 (20) le:š ʔu:m why ‘Stand up, come hit me!’11

taʕa:l come.imp.2msg


We thus see that the distribution of DMI is wider in Arabic, a fact compatible with the direction of transfer from Arabic to Hebrew: imposition rather than borrowing. The existing recorded data does not allow us to determine whether it was Palestinian Arabic or the Jewish Arabic dialects that were most influential. It is possible that all of them contributed to some extent. Had we found attested Hebrew DMI examples prior to the 1950s, we would have concluded that their origin must be Palestinian Arabic. Since we have not found such examples, but they may well exist, we leave it to future research to produce conclusive evidence relevant to this issue. In the next section, we consider an additional possible source for the DMI in Modern Hebrew. The DMI in Neo-Aramaic The DMI is also found in Neo-Aramaic dialects, both in Christian communities that immigrated to Europe, North America, and Australia (e.g., Christian Barwar) and in communities that immigrated to Israel (e.g., the Jewish dialect of Zakho):

9  This example was constructed by the first author. 10  This example also illustrates the serial verb construction, discussed in Mar’i & Gamliel (2015). For further analysis of serial verb construction in Palestinian Arabic, see Khalaily (1997:238–242). 11  This example was constructed by the first author.


Khalaily and Doron


Christian Barwar nə́mu ʔáyya bə́rke d-ɛni-laʔáti sxáya gàwa why this pool of-who-cop that-you swim ‘Whose is this pool in which you are swimming? (= surely it is mine not yours)’ (Khan 2008:906)


nə́mu la-t-ðàʔə-lli why no-that-know.2ms-me ‘Don’t you know me anymore? (= surely you remember me)’ (Khan 2008, 1596:65)


Jewish Zakho12 qay-   mà why what ‘What happened?!’


qaymà   ʾuz-li         why what do.prf-1s ʾóṭo         k-ṣárx-ət ʾə̀ll-i?! like.that ind-shout.imp-2sm on-1s ‘What did I do that you are shouting at me like that?!’


qay-   kmà       k-táql-an     why how-much ind-weigh.imp-1fs g-ə́mr-et                     šamə̀ntawan?! ind-say.imp-2ms fat.fs cop.fs ‘How much do I weigh that you are saying I am fat?!’

bré-le?! happen.prf-3ms


We therefore conclude that the Jewish Neo-Aramaic dialects are a possible source for the DMI in Modern Hebrew.13 As pointed out to us by Geoffrey Khan, the DMI can actually be reconstructed as a general syntactic trait of the Semitic languages of the area. As such, it may in fact have had its origins in the special distribution of the Classical Aramaic

12  We are grateful to Oz Aloni for these examples, which he kindly transcribed from his recordings of the speech of Zakho native speakers. 13  Neither Barwar nor Zakho had been in contact with Arabic, therefore the DMI in NeoAramaic does not itself originate in the Modern Arabic dialects.

Colloquial Modern Hebrew Doubly-Marked Interrogatives


particle lema, originally le-ma ‘for what,’ etymologically related to the Classical Hebrew lamma ‘why.’ Classical Syriac uses a construction in which lema embeds a question. Our analysis of the DMI may be applicable to this Classical Syriac construction. Perhaps one could view the question embedded by lema as a content question, and lema itself as an attitude marker that denies the presupposition of doubt in the content question, e.g., lema emma w-atta ṭavan li men alaha ‘Are mother and wife better to me than God? (And why assume this may be true, of course God is better),’ quoted by Nöldeke (1904, §373).14 In Biblical Aramaic and other dialects of pre-modern Aramaic, the phrase d-lema/di-lema is used in the sense of ‘lest,’ e.g., Ezra 7:23 ‫די למה להוא קצף על‬ ‫‘ מלכות מלכא‬Lest there be wrath against the kingdom of the king.’ Perhaps the background of such constructions is: ‘For there will be wrath (But why assume this is inevitable?).’ Rabbinic Hebrew has the particle ‫שמה‬/‫ שמא‬šemma used in purpose clauses with the meaning of ‘lest.’ This particle is structurally equivalent to the Biblical Aramaic di-lema, since šemma < še-l-ma. The particle ‫שמה‬/‫ שמא‬in Rabbinic Hebrew is, in fact, also used in yes-no questions, apparently as an attitude marker, to deny the presupposition of the question, as in the following example from Mishnah Eduyyot 5.7: ‫ שמה עוולה מצאתה בי‬šemma ʕawla maṣata bi ‘Did you find wrong in me?’ (= surely not), quoted in Stadel (2014:314). If the reconstruction of the ancient why+question as a DMI is on the right track, then it may be that Modern Hebrew is regaining a lost Semitic construction through contact with Modern Arabic and Neo-Aramaic dialects, and that this development parallels the contact between ancient stages of Hebrew and Aramaic. On the one hand, the ancient etymology may shed light on the restriction we have found in Modern Hebrew of the attitude marker to lama ‘why’ (and perhaps ma ‘what’). On the other hand, the modern development may cast light on the earlier constructions, since we can observe their development in embryonic stages and get direct access to their pragmatic function.

14  The translation is actually not Nöldeke’s, but is the re-adaptation suggested by Geoffrey Khan in view of our analysis of the DMI. All the translations of the classical examples in the present section have likewise been re-adapted. Our re-adaptation assigns why in the relevant examples its ordinary lexical meaning, and, crucially, the special DMI function of an attitude marker conventionalizing the rhetorical question interpretation of ordinary why. This is different from the received translations of these classical examples, which postulate a special ad-hoc lexical meaning of why.


Khalaily and Doron

Conclusion The DMI construction of colloquial Modern Hebrew is a complex interrogative construction consisting of an extra wh-phrase (usually why) that embeds an ordinary question typically introduced by its own wh-phrase. Though the latter wh-phrase may be a genuine quest for information, the former wh-phrase endows the question with the very distinctive illocutionary force of rejecting a salient presupposition present in the discourse. The DMI is also found both in dialects of Modern Arabic and in dialects of Neo-Aramaic, including those with which Modern Hebrew has been in contact. Accordingly, we conjecture that the Hebrew DMI was imposed by contact with these dialects. It is very improbable that such a marked construction would emerge in Hebrew independently of its contact languages. Yet the imposition may have been facilitated by the historical vestiges of an ancient DMI construction that had been borrowed by Rabbinic Hebrew from its contemporary Aramaic. References al-Rahman Mar’i, Abed. 2013. Walla Bseder: A Linguistic Profile of the Israeli-Arabs. Jerusalem: Keter. ――― & Ophira Gamliel. 2015. “Bleached Verbs as Aspectual Auxiliaries in Colloquial Modern Hebrew and Arabic dialects.” Journal of Jewish Languages 3.1–2: 51–65. Francez, Itamar. 2015. “Modern-Hebrew lama-še Interrogatives and Their JudeoSpanish Origins.” Journal of Jewish Languages 3.1–2: 104–115. Ginzburg, Jonathan. 2012. The Interactive Stance: Meaning for Conversation. Oxford: Oxford University Press. Han, Chung-hye. 1996. “Deriving the Interpretation of Rhetorical Questions.” Proceeding of WCCFL 16: 1–17. ―――. 2002. “Interpreting Interrogatives as Rhetorical Questions.” Lingua 112: 201–229. Henshke, Yehudit. 2013. “On the Mizraḥi Sociolect in Israel: A Sociolexical Consideration of the Hebrew of Israelis of North African Origin.” Journal of Jewish Languages 1: 207–227. Khalaily, Samir. 1997. One Syntax for ALL Categories: Merging Nominal Atoms in Multiple Adjunction Structures. Doctoral dissertation. Leiden: Holland Institute of Generative Grammar (HIL). Khan, Geoffrey. 2008. The Neo-Aramaic Dialect of Barwar, Vol. 1. Leiden: Brill. ―――. 2011. “North-Eastern Neo-Aramaic.” In The Semitic Languages, ed. Stefan Weninger. Berlin: De Gruyter Mouton, 708–724.

Colloquial Modern Hebrew Doubly-Marked Interrogatives


Krifka, Manfred. 1995. “The Semantics and Pragmatics of Polarity Items.” Linguistic Analysis 25: 1–49. Mutzafi, Hezy. 2014. “Jewish Neo-Aramaic.” In Encyclopedia of Jews in the Islamic World, ed. Norman A. Stillman. Brill Online, encyclopedia-of-jews-in-the-islamic-world/jewish-neo-aramaic-COM_000701. Nöldeke, Theodor. 1904. Compendious Syriac Grammar. London: Williams and Norgate. [Translation by James A. Crichton of T. Nöldeke, Kurzgefasste syrische Grammatik, 2nd edition, 1889, Leipzig.] Sadock, J. M. 1971. “Queclaratives.” Papers from the Seventh Regional Meeting of the Chicago Linguistic Society 7, 223–232. ―――. 1974. Towards a Linguistic Theory of Speech Acts. New York: Academic Press. Stadel, Christian. 2014. “Interrogative: Rabbinic Hebrew.” In Encyclopedia of Hebrew Language and Linguistics, Vol. 2, eds. Geoffrey Khan et al. Leiden: Brill, 313–316. Yang, Barry Chung-Yu. 2007. “Rhetoric/Disapproving wh and Intervention Effect.” unpublished manuscript. Harvard. and%20Intervention%20Effect%202.0.pdf.

The Right Periphery in Colloquial Hebrew: Modality and Language Contact Driven Effects Yael Ziv

The Hebrew University of Jerusalem, Jerusalem, Israel [email protected]

Abstract The article discusses structural and functional properties of non-canonical constituents in the right periphery in Modern Hebrew, by comparison with their left periphery correlates. It is argued that the left periphery constructions are evident in both written and spoken modalities and that the right periphery constructions characterize the spoken modality. The discourse functions of both constructions are shown to follow from their respective modality constraints. It is suggested that the late emergence of the right periphery structures in Modern Hebrew might be due to their existence in the languages with which Hebrew was in contact.

Keywords written modality – spoken modality – left periphery – non-canonical right periphery constructions

Introduction A considerable amount of research has been devoted to the study of constituents in the so-called left periphery (initial position) of the sentence in the spoken modality of Modern Hebrew, but relatively little has been said about

* I should like to thank Emanuel Allon, Isaac Bleaman, Edit Doron, Keren Dubnov, Ronnie Henkin-Royt, Yehudit Henshke, Olga Kapeliuk, Yael Maschler, Asya Pereltsvaig, Chana Sagi, Lea Sawicki, Moshe Taube, Chava Turniansky, and Ghil’ad Zuckermann for providing relevant data and clarifying questions.

© koninklijke brill nv, leiden, ���6 | doi ��.��63/9789004310896_011

The Right Periphery in Colloquial Hebrew


constituents in the right periphery (final position). This study is an attempt to shed some light on the latter constructions, particularly non-canonical ones. An investigation into the nature of some left and right periphery constructions indicates that, despite certain similarities, the two are not mere mirror image structures, but rather differ in various syntactic and discourse-functional properties. Several of these properties will be examined here. In addition, the two constructions do not demonstrate identical modality characteristics. The relevant left periphery constructions appear in both the written and the spoken modalities, whereas the right periphery structures show up typically in the spoken mode. Since the spoken modality of Modern Hebrew has been significantly affected by languages with which it was in contact (e.g., Yiddish, Polish, Russian, and some dialects of Arabic), the right periphery constructions reflect parallels in these languages. It is true that the right periphery constructions appear to serve discourse purposes which are driven by modality considerations, so it might not be surprising to find similar constructions in languages with which Hebrew was not in direct contact, but the fact that in the languages with which Hebrew was in contact these constructions are used regularly might have constituted a driving force in their emergence in Hebrew.

Non-canonical Right Periphery Constructions

Colloquial Hebrew is considered to be an SVO language; hence, constituents which are not part of the complement and which occur in the right periphery are non-canonical. Among these constituents are: (a) comment clause parentheticals, (b) tags and discourse markers, (c) focus markers, (d) certain conjunctions, and (e) the explicit noun phrases (NPs) in sentences displaying right dislocation (RD). These are exemplified in the following, respectively, with the relevant constituent underlined: (1) (2)

‫ אני חושבת‬,‫הגנב נכנס מכאן‬

ha-ganav nixnas mi-kan ʔani ħoševet the-thief from-here I think ‘The thief got in from here, I think.’ ‫ נכון‬,‫הגנב נכנס מכאן‬

ha-ganav nixnas mi-kan naxon? the-thief from-here true? ‘The thief got in from here, right?’

130 (3)

Ziv ‫אני לא חושב על זה אפילו‬

ʔani lo ħošev ʕal ze ʔafilu I not think about it even ‘I don’t think about it even.’

(4) A: ‫ את תראי‬.‫אני אקבע איתך בתאריך שאני באה לתאילנד‬ ʔani ʔeqbaʕ ʔitax ba-taxʔarix še-ʔani baʔa le-tayland I will.fix in.the-date that- I come to-Thailand ʔat tirʔi you will.see ‘I’ll set (a meeting) with you when I come to Thailand, you’ll see.’ B: . . .‫אני אגיע לפניך אבל‬1 ʔani ʔagiaʕ lefanayix ʔaval I will.arrive, but ‘I will arrive before you, but . . .’ (5)

‫ הגנב‬,‫הוא נכנס מכאן‬

hu nixnas mi-kan ha-ganav he from-here the-thief ‘He got in from here, the thief.’

What seems to characterize the underlined non-canonical right periphery constituents above is that they do not contribute to the so-called “at issue” propositional content of the sentence, even when they are coreferential with an entity in the sentence which does contribute to the propositional content (i.e., RD, which prototypically reintroduces an entity previously mentioned in the discourse or occurring in the extra-linguistic context).2 Not surprisingly, these have left dislocated correlates, as in:3

1  The example is taken from the Corpus of Spoken Israeli Hebrew (CoSIH) (http://humanties, accessed July 23, 2014). It should be noted that there are several examples of this type in this corpus and it is clearly not a performance error. 2  Such a characterization rules out instances of extraposition and extraposed relative clauses, in which the right periphery constituents do contribute to the propositional content. 3  Note that so-called medial position correlates are only mentioned here briefly when they demonstrate unique properties.

The Right Periphery in Colloquial Hebrew


4‫אני חושבת (ש)הגנב נכנס מכאן‬ ʔani ħoševet še-ha-ganav nixnas mi-kan I think that-the-thief from-here ‘I think that the thief got in from here.’


‫נכון הגנב נכנס מכאן‬

naxon ha-ganav nixnas mi-kan? true the-thief got. in from-here ‘Is it true that the thief got in from here?’


5‫אפילו אני לא חושב על זה‬ ʔafilu ʔani lo ħošev even I not think ‘I don’t even think about it.’


‫אבל אני אגיע לפניך‬



ʕal ze about it

ʔaval ʔani ʔagiʕ lefanayix but I will.arrive ‘But I will get there before you.’ ‫ הוא נכנס מכאן‬,‫הגנב‬

ha-ganav hu nixnas mi-kan the-thief he got. in from-here ‘The thief, he got in from here.’ Left vs. Right Periphery Constructions

The two periphery constructions display somewhat different syntactic and discourse-functional properties, in addition to modality characteristics. The syntactic distinctions pertain to the so-called subjacency properties that constrain rightward but not leftward movement (to the end of the clause in which they occur, Ross 1967:185). Comment clause parentheticals as in (1) are often regarded as involving rightward movement. Final focus constructions as in 4  In the spoken modality of Colloquial Hebrew there are instances where the subordinator -‫ ש‬še- ‘that’ does not occur (cf. in this context Ziv, in press). In (2’) (following) the subordinator -‫ ש‬is often omitted as demonstrated here. 5  The position of the focus marker ‫ אפילו‬ʔaf’ilu ‘even’ is naturally determined by scope considerations and by intonation. Here there is an additional option following the first constituent.



(3) and structures displaying final conjunctions as in (4) might also be analyzed this way. The other periphery structures under consideration, however, do not involve movement. As already mentioned above, unlike the left periphery constructions, which occur in both spoken and written modalities (fulfilling similar discourse function), the right periphery structures characterize the spoken modality (cf. Beeching & Detges [2014] for cross-linguistic support).6 The different constructions display distinct functional properties in conformity with their different modes. Since the written modality is more likely than the spoken to represent pre-planned speech, it is expectable that the left periphery, occurring initially, should be associated with an explicit statement of parameters such as illocutionary force, stance, and epistemic status, whereas the right periphery, prototypically a spoken modality property, should demonstrate a retroactive, corrective role, clarifying the status of the prejacent. It is noteworthy that despite their origin, different right periphery constructions have become grammaticalized.

Comment Clause Parentheticals

Comment clause parentheticals (CCP) have long been analyzed as occurring in a variety of syntactic positions in the sentence, the initial position (as in 1’) usually not being regarded as parenthetical material. Lately, however, analyses have been proposed in which these, too, are regarded as parenthetical. This is particularly suggestive in cases where the subordinating conjunction is eliminated in the spoken modality (cf. Kaltenboeck 2011, for English, and Ziv, in press, for Hebrew). Tracing the distribution of CCP in Hebrew indicates that Biblical Hebrew displays numerous examples of left periphery CCP specifying the speech act type )e.g., ask, reply, swear), as in: (6)

‫מּואל וְ ָדוִ ד‬ ֵ ‫אמר ֵאיפֹה ְׁש‬ ֶ ֹ ‫וַ ִ ּי ְׁשאַ ל וַ ּי‬

way-yišʾal way-yōmɛr ʾēp̄ ō šəmuʾɛl wə-dāwid and-asked and-said where Shmuel and-David ‘And he asked and said, Where are Samuel and David?’ (1 Samuel 19:22)7

6  When they do occur in the written modality, they usually constitute renditions of the spoken mode or are driven by poetic considerations. 7  The biblical translations are from the King James Version.

The Right Periphery in Colloquial Hebrew


However, there are only a handful of right periphery CCP in the later books (mostly in poetic style) as in (7): (7)

'‫נִ ַחם ה' ַעל זֹאת לֹא ִת ְהיֶ ה ָא ַמר ה‬

niḥam H ʿal zōṯ lō ṯihəye ʾāmar H repented God on this not will be said God ‘The Lord repented for this. It shall not be, Saith the Lord.’ (Amos 7:3)

In Post-Biblical Hebrew, in particular Mishnaic and Early Rabbinic Hebrew, there is also a multiplicity of left periphery CCP and only a few right periphery counterparts. In cases where the two co-occur, the left periphery CCP clearly outnumber the instances of right periphery. In the Hebrew texts written during the Haskala (Jewish Enlightenment) period (e.g., Mapu’s ‫ אהבת ציון‬ʔahavat Zion ‘The Love of Zion’), we find an abundance of left periphery CCP, echoing the biblical style. In Y.L. Gordon (who writes toward the end of that period), we find CCP also in the right periphery. (e.g., in ‫שני ימים ולילה אחד בבית מלון אורחים‬ šney yamim velayla exad bebet melon orxim ‘Two Days and One Night in a Motel’). During the Early Modern Hebrew period, we find a fair amount of CCP in the right periphery, as, for example, in Brener’s ‫‘ זעיר שם‬zʕer šam ‘A Bit There’: (8)

‫ — רטן הוא‬. . . ‫ יש לי זמן‬. . . ‫יש לי זמן‬

yeš li zman . . . yeš li zman . . . ratan hu time time grumbled he ‘I’ve got (free) time, grumbled he.’

Currently, the spoken modality is rich with instances of both left and right periphery CCP (cf. example 1 above, and Ziv, in press).

Tags and Discourse Markers

Tags and discourse markers (DM) demonstrate a range of forms and functions, but the common denominator appears to be their non truth-conditional nature. Thus, their function is to provide instructions for interpretation.8 The instruction which the marker ‫ נכון‬naxon ‘true’ (in 2 and 2’) provides concerns the assessment of propositional givenness (accessibility), and, in line with its 8  Cf. the concept of “procedural meaning,” as formulated in Blakemore (1987), within Relevance Theory.



relative position (left or right periphery), the instruction is either an explicit request for confirmation (left periphery), or a retroactive attempt to verify the status of the prejacent (right periphery) (cf. Mittwoch 1979; Ziv 2006, in press). It is noteworthy that although written varieties of Hebrew prototypically lack the types of tags and discourse markers discussed here, it seems that organizational DM occur in the left periphery already in Biblical Hebrew. Thus, ‫ ועתה‬wǝ-ʿattā ‘and now’ functions as a left periphery segmentational DM as in Genesis 30:30: (9)

‫יתי‬ ִ ‫וְ ַע ָּתה ָמ ַתי ֶא ֱע ֶׂשה גַ ם ָאנ ִֹכי ְל ֵב‬

wǝ-ʿat-tā mātay ʾɛʿɛ̆śɛ ḡam ʾānōḵī lǝ-ḇēṯī and-now when also I to-my-house ‘and now when shall I provide for mine own house also’

‫ ועתה‬wǝ-ʿattā ‘and now’ clearly does not contribute its conceptual meaning of temporality, since its co-occurrence with the interrogative ‫ מתי‬matay ‘when’ would make the sentence incongruous in that the temporal indication is supplied (i.e., ‘now’) and queried (i.e., ‘when’) simultaneously. Rather, it functions procedurally along with the predictable features of DM (cf. Ziv 2001). In an interesting discussion of the use of DM during the Enlightenment and Early Modern Hebrew periods as reflected in Brener’s translation of Dostoyevsky’s Crime and Punishment (presumably around 1911) and in his own writings, Dubnov (in press) points out the abundance of DM such as ‫באמת‬ beʔemet ‘really,’ ‫ פשוט‬pašut ‘simply,’ and ‫ בכלל‬bixlal ‘at all.’ Significantly, these DMs occur quite extensively in colloquial Hebrew as well (cf. Ziv 2001, 2006, 2007, 2012). Dubnov indicates that Brener’s translation of the DM reflects the Russian correlates found in the original. She thus proposes, convincingly, that rather than reflecting the spoken mode of Hebrew of those times, the occurrence of the DM in question constitutes evidence for their borrowing from the languages with which Hebrew was in contact, in this particular case, Russian. Dubnov is not concerned with the position of the DM in the sentence. Colloquial Hebrew displays the types of DM under consideration in both the left and the right peripheries (cf., e.g., Ariel 1988 with respect to the givenness marker ‫הרי‬, harey ‘after all,’ Ziv 2001 with respect to ‫ פשוט‬pašut ‘simply,’ Ziv 2006, 2007 with respect to ‫ נכון‬naxon ‘true’ as in sentences (2) and (2’), and Ziv 2012 with respect to ‫ בכלל‬bixlal ‘at all’).9

9  On bixlal cf. Tsirkin-Sadan 2015.

The Right Periphery in Colloquial Hebrew


Focus Markers

Focus Markers (e.g., ‫ אפילו‬ʔafilu ‘even’ as in 3 and 3’) are known to be scopesensitive, and their distribution is mostly determined by scopal considerations. In the spoken modality they are evident in various positions in the sentence, while the written variety is more constrained with preferences for left periphery. Biblical Hebrew shows ‫ אף כי‬ʾap̄ ki ‘even’ (presumably the counterpart of ‫ אפילו‬ʔafilu) only in the left periphery, with ‫ כי‬ki functioning as a subordinating conjunction, as in: (10) ‫אכלּו ִמכֹל ֵעץ ַהגָ ן‬ ְ ֹ ‫ֹלהים לֹא ת‬ ִ ‫אמר ֶאל ָה ִא ָשה אַ ף ִּכי ָא ַמר ֱא‬ ֶ ֹ ‫וַ י‬ way-yōmɛr ʾɛl hā-ʾiš-šā ʾap̄ kī ʾāmar ʾɛ̆lōhīm lō tōḵlū and-said to the-woman even said God not (you) mik-kōl ʿēṣ hag-gān from.every tree the-garden

‘And he said unto the woman: “Yea, hath God said: Ye shall not eat of any tree of the garden?” ’ (Genesis 3:1)

‫ אפלו‬ʔap̄ illu ‘even’ occurs in the left periphery in post-biblical Hebrew with

both clausal and phrasal scope, as is evident in the following, respectively:

(11a) ‫אפלו חרב חדה מונחת על צוארו של אדם‬ ʾăp̄ illū ḥɛrɛḇ ḥadā mūnāḥat ʿal ṣāvarō šēl ʾādām even sword sharp lies on neck of man ‘Even (if) a sharp sword lies against a person’s neck.’ (Mishnah, Brachot 1:10)10 (11b) ‫אפלו כמלא הנימה‬ ʾăp̄ illū ki-mǝlō han-nimā even as full the (piece of) hair ‘even a tiny amount’ (Mishnah, Yoma 38:2) We find instances of ‫ אף כי‬ʔaf ki in Y.L. Gordon’s ‫שני ימים ולילה אחד בבית מלון‬ ‫ אורחים‬Šney yamim velayla exad bebet melon orxim, ‘Two Days and One Night in a Motel’ (Enlightment period) with similar left periphery distribution as in Biblical Hebrew, but no instances of ‫ אפלו‬ʔafilu. ‫ אפילו‬ʔafilu resurfaces in the Early Modern Hebrew period, e.g., in Mendele’s ‫ האבות והבנים‬Haʔavot 10  Author’s own translation, after examination of several interpretations.



­vehabanim ‘The Fathers and the Sons’ in the left periphery. In the spoken modality in colloquial Hebrew, we find ‫ אפילו‬ʔafilu with clausal and phrasal scope in the right periphery as well (cf. (3) above for an example with clausal sentential scope). Conjunctions Conjunctions prototypically occur in the left periphery and have only recently, due to careful examinations of the spoken modality, been observed to occur systematically also in the right periphery with clear functional characteristics. Thus, while the prototypical left periphery conjunction ‫ אבל‬ʔaval ‘but’ introduces information which denies expectations raised by preceding material, its right periphery counterpart introduces accessible, implied or presupposed information.11 In addition, final but may function to signal potential end of turn. Example (4) above in its original context (from the CoSIH) displays both discourse functions. The turn-yielding role of ‫ אבל‬ʔaval ‘but’ is exemplified in the sequence uttered by Speaker B, which is followed in the original context by an utterance by speaker A, and its so-called content function is demonstrated by the fact that the utterance by Speaker B pertains to the date of their meeting in Thailand (mentioned earlier in the discourse). Final ‫ אבל‬ʔaval structures have only emerged fairly recently in the spoken modality in Hebrew and there is no trace of them in earlier periods in the written modality. Discussions of final but in English demonstrate that these are systematically used and shift from the left periphery turn-continuing connective to a turn-yielding discourse particle, in a way that is consistent with what has been described in the grammaticization literature (Mulder & Thompson 2006). Since this right periphery structure has only emerged fairly recently, it might be the result of an independent development in the spoken modality of Hebrew. But the fact that these occur also in the languages with which Hebrew has been in contact might have constituted a driving force in this developmental path.

Right Dislocations

Right dislocations (RD) and left dislocations (LD) (as in (5) and (5’) above) both have organizational functions in the spoken modality, yet these differ in detail 11  In this function it could be regarded as a “backward looking” device, to adopt Grosz & Sidner’s (1986) distinction between forward and backward looking centers.

The Right Periphery in Colloquial Hebrew


(cf., e.g., Ziv 1994). RDs function to identify and introduce into the discourse model situational referents, which are naturally accessible to the addressee.12 LDs, in comparison, may introduce “brand new” entities (cf. Prince 1981).13 The difference then lies in the accessibility status of the referents in question to the addressee (cf. Ariel 1990), i.e., whether they are accessible to memory or brand new entities. RDs involve the first type of referents, but LDs involve both types. Both, however, may re-introduce textual referents that are judged not to be in the foreground of the listener’s consciousness (to use Chafe’s 1976 conception). Poetic forms such as those evident in Biblical Hebrew display few instances of LD and practically no instances of RD. Following is an example of a Biblical LD: (12)

‫רּוׁש ַליִ ֤ם ָה ִרים ָס ִביב ָלּה‬ ָ ְ‫י‬

yǝrūšālayim hārīm sāḇīḇ lā Jerusalem mountains around her ‘As the mountains are round about Jerusalem.’ Psalm 125:2

In addition, we can find several Post-Biblical examples of a similar type. Very few examples of what appear to be instances of RD are evident in Mishnaic Hebrew, as in the following example: (13)

‫ מתנה טובה יש לי בבית גְ נָ זַ י ושבת ְש ָמּה‬:‫אמר לו הקדוש ברוך הוא למשה‬

ʾāmar lō haq-qādōš bārūx hū lǝ-mōšɛ: said to.him the-holy blessed he to-Moshe: matānā tova yēš li bǝ-ḇ ēt present good house-in gǝnāzay wǝ-šabāt šǝmāh and-Shabbat name.her ‘And said to him God to Moses, I have a present in my archives and its name is Shabbat.’ (Mishnah, Shabbat 10:72)

Mishnaic writing is characterized by commentaries in the body of the text and in a case such as (13) above, the explicit NP ‫ משה‬moshe ‘Moses’ clarifies the identity of the referent introduced by the cataphoric pronominal ‫ לו‬lo ‘to him.’ It seems that this type of construction was codified later (and assumed a 12  In fact, RDs can also function as appositions providing a predicate characterizing the entity in question (cf. Doron 1994 for an analysis of appositions as predicates.) 13  I have only referred here to the discourse function, rather than the syntactic-therapeutic properties of LDs (cf. Prince 1992).



distinct discourse function in the spoken mode). Indeed, we find a few additional such instances in the Mishnah. Most display an adjacency between the cataphoric pronoun and the explicit co-referential NP, as in: (14)

‫ מפני מה אתה מצוי בין הגדרות‬:‫אמרו לו לנחש‬ āmrū lō lan-nāḥāš: mipnē ma said to him to-the-snake why bēn hag-gdērōt between the-fences

ʾattā you

māṣūy located

‘They said to him to the snake: Why are you between the fences?’ (Midrash Raba EMOR, 26:2)

This corroborates the idea that the initial function of the construction was clarificatory and the status of the referent could be “brand new,” unlike its status in the spoken modality currently. The cases where there is no adjacency, as in (13), appear to suggest codification. Interestingly, examples from the Enlightenment and Early Modern Hebrew literature also display the more frequent version with the adjacency between the cataphoric pronoun and the explicit NP. Singular examples displaying non-adjacency between the two coreferential NPs appear to echo the Mishnaic style. The spoken modality in Modern Hebrew shows a range of LD and RD constructions. The two structures appear to display a subject/non-subject asymmetry, such that in LDs, cases where the criterial coreferential NP is non-subject ((15), below), clearly outnumber those where it is the subject ((5’), above): (15)

‫ אני אף פעם לא אבין אותה באמת‬,‫רותי‬

ruti ʔani ʔaf paʕam lo ʔavin ʔota beʔemet Ruthie I even once not understand her in.truth ‘Ruthie, I will never really understand her.’

With RD, cases where the criterial coreferential NP is the subject ((5) above) clearly outnumber those where it is not ((16) below).14

14  Interestingly, in Mishnaic Hebrew and the later examples that echo this style, RD, like LD, shows preference for the criterial coreferential NP to be non-subject. The difference in statistical preference appears to be related to the discourse function in the spoken modality.

The Right Periphery in Colloquial Hebrew



‫ראיתי אותו בטלויזיה את אחיך‬

raʔiti ʔoto ba-televizya ʔet ʔaħix (I)saw him on-television acc ‘I saw him on television, your brother.’

It stands to reason that these distributional asymmetries are related to the discourse function of the construction in question in the relevant modality.

Contact Languages and Related Structures

As already intimated, non-canonical constituents in the right periphery occur in the spoken modality considerably more often than in the written modality. This is not the case with the corresponding constituents in the left periphery, which appear in both modalities in similar frequencies. Since the spoken modality of Colloquial Hebrew has been significantly affected by languages with which it was in contact (e.g., Yiddish, Polish, Russian) the right periphery constructions attested in Hebrew reflect parallels in these languages. Yiddish Yiddish has had a considerable effect on the word order in Colloquial Hebrew. Thus, despite problems involved in the analysis of Yiddish word order within topographical frameworks (e.g., Fox 2008) and information structure orientations (e.g., Hinterhölzl 2010), and questions pertaining to the unmarked order VO (e.g., Diesing 1997) or OV (e.g., Vinker 2001), the status of the right periphery structures under consideration is not affected by these analyses. Their occurrence in the spoken modality in Yiddish, as in the following, might have served as a driving force for their occurrence in Colloquial Hebrew. 15 Note that the numbers of the examples in this section correspond to the original Hebrew examples displaying the Right Periphery phenomena. (1a) der ganev iz fun danet arayn, meyn ikh the thief is from here inside think I ‘The thief got in from here, I think.’

15  These examples, and those from Polish are taken from field work with informants.



(2a) der ganev iz fun danet arayn, rikhtik the thief is from here inside right ‘The thief got in from here, right?’ (3a) ikh ken nisht trakhtn vegn dem afile I can not think about it even ‘I cannot think about it even.’ (4a) ikh farshtey nisht ober16 I understand not but ‘I cannot understand but/however.’ (5a) er iz fun danet arayn, He is from here inside ‘He got in from here, the thief.’

der ganev the thief

As for the RD construction (5), intuitions differ with respect to the relevance of adjacency between the two coreferential NPs, as well as that of the subject/non-subject distinction to grammaticality. Some speakers attribute significance to the distinctions, judging the sentences with the subject as criterial coreferential NP as better than those with the non-subject counterparts, and others regard the distinctions as insignificant, assigning a similar degree of grammaticality to both. The adjacency issue, too, is considered as criterial by some and immaterial by others. What is significant for our point here is that such distributional distinctions may be relevant for grammaticality judgment. Polish Polish is considered a free word order language, with SVO as the unmarked order and the full range of options determined informationally. Here, too, the relevant right periphery entities which characterize the spoken modality might have had an effect on the corresponding structures evident in the spoken modality of Hebrew. Except for the correlates of (4) with ‫ אבל‬ʔaval ‘but,’ all the relevant sentences occur in the spoken modality of Polish. (1b) (Ten) Złodziej wszedł tȩdy, myślę this thief (I) think ‘The thief got in from here, I think.’ 16  Some speakers indicated that this may be restricted to clause rather than sentence final position. The crucial factor in the current context is that ober is not restricted to clause initial position.

The Right Periphery in Colloquial Hebrew


(2b) (Ten) Złodziej wszedł tȩdy, prawda this thief true ‘This thief got in from here, true?’ (3b) nie myślę o tym nawet17 No (I) think about this even ‘I don’t think about it even.’ (4b) There is no Polish counterpart. (5b) On wszedł tȩdy, ten złodziej he this thief ‘He got in from here, the thief.’ Example (5b) above demonstrates the two coreferential NPs in the subject role. It is noteworthy that speakers of Polish who originally rejected RD examples where the criterial coreferential NP was non-subject, later came up with such examples where the coreferential NPs were adjacent, as in: (5b’) nie widziałem go tego sterownika not (I) saw him that controller ‘I haven’t seen him, that controller.’ (, accessed August 24, 2014) So parameters such as the subject/non-subject status of the coreferential NPs in the RD construction, as well as adjacency versus non-adjacency of the criterial NPs, appear to be relevant in the grammaticality judgment in Polish as well. Russian Like Polish, Russian, too, appears to display a range of word order combinations, which are determined by information structure considerations. The unmarked word order here as well is supposedly SVO. As in the case of Yiddish and Polish, the right periphery constituents under investigation appear to occur mostly in the spoken modality, and it is likely that they have affected the emergence of their correlates in the spoken modality in Hebrew:

17  Some native speakers rejected it, while others accepted it. The latter judgment appears to reflect its multiple occurrences in the Internet in this position.



(1c) vor prolez otsjuda, ya dumayu thief entered I think ‘The thief got in from here, I think.’ (2c) vor prolez otsjuda, verno thief entered true ‘The thief got in from here, true?’ (3c) ya ne dumayu ob etom dazhe I not think about this even ‘I don’t think about this, even.’ (4c) ya ne dumayu ob etom odnako18 I not think about this however ‘I don’t think about this however.’ (5c) on prolez otsjuda, vor he entered thief ‘He got in from here, the thief.’ With respect to the RD structure, the subject/non-subject distinction of the criterial coreferential NP as well as its adjacency/non-adjacency parameter appear to affect the acceptability judgment among speakers. Subjects rate better than non-subjects, and adjacency is worse when the relevant NP is subject than when it is not. Bedouin Arabic in the Negev An interesting observation was provided to me by Ronnie Henkin (p.c.) with respect to the RD-like constructions in the Arabic spoken by Bedouins in the Negev. Henkin reports that the most natural narrative counterpart of the RD sentence is the pronoun-less sentence with the subject occurring in final position (either as an afterthought or as an integral part of the sentence), as in the following example:

18  It has been pointed out to me that the colloquial “no” (but) cannot occur in final position, and a tentative prosodic rationale has been proposed (Keren Dubnov). As an alternative, the more formal “odnako” (‘though,’ ‘however’) was suggested.

The Right Periphery in Colloquial Hebrew


(5d) fāt min ihniyy al-ḥaṛāmiy entered from here the thief ‘ϕ got in from here, the thief.’ She suggests that the RD version (with the pronoun), as in (5e below) constitutes a special pattern where the subject is the discourse topic (of discussion), which presumably has already occurred in the discourse segment: (5e) hū fāt min ihniyy al-ḥaṛāmiy he entered from here the thief ‘He got in from here, the thief.’ As we have seen this is in line with the discourse function of RD. Bearing this observation in mind the influence of the contact with Arabic could also serve as a source for the emergence of the RDs in the spoken mode of Colloquial Hebrew. This suggestion could also be corroborated given that VS structures in Hebrew are mostly restricted to indefinite subjects (unlike the RDs under investigation) and hence could not possibly serve as the source of the Hebrew RD constructions. It is thus plausible to assume that the occurrence of RDs in Hebrew is the effect of language contact of some sort. Conclusion Having provided the correlates of the non-canonical right periphery constructions in Yiddish, Polish, and Russian, a caveat is in place. The non-canonical right periphery structures are characteristic of the spoken modality, and they are evident in languages other than Yiddish, Polish, and Russian (e.g., French). Hence, their occurrence in Modern Hebrew is clearly a development that is due to its having become a spoken tongue, and the fact that they occur in the languages with which Hebrew has been in contact since its revival, might have constituted a licensing factor and a driving force for this development. References Ariel, Mira. 1988. “Retrieving Propositions from Context: Why and How.” Journal of Pragmatics 12.5–6: 567–600. ———. 1990. Accessing Noun Phrase Antecedents. London: Routledge.



Beeching, Kate & Ulrich Detges, eds. 2014. Discourse Functions at the Left and Right Periphery: Crosslinguistic Investigations of Language Use and Language Change, Studies in Pragmatics 12. Brill: Leiden. Blakemore, Diane. 1987. Semantic Constraints on Relevance. Oxford: Blackwell. Chafe, Wallace L. 1976. “Givenness, Contrastiveness, Definiteness, Subjects, Topics, and Point of View.” In Subject and Topic, ed. Charles N. Li. New York: Academic Press, 25–55. Corpus of Spoken Israeli Hebrew (CoSIH). Accessed July 27, 2014. Diesing, Molly. 1997. “Yiddish VP Order and the Typology of Object Movement in Germanic.” Natural Language and Linguistic Theory 15: 369–427. Doron, Edit. 1994. “The Discourse Function of Appositives.” In The Proceedings of the 9th Annual Conference of Israel Association of Theoretical Linguistics and the Workshop on Discourse, eds. Rhonna Buchalla & Anita Mittwoch. Jerusalem: A.I. Weidenfeld Ltd., 53–65. Dubnov, K. In press. “The Buds of Modern Hebrew in Brener’s Translation of Crime and Punishment—Discourse Markers.” Haivrit 63 (in Hebrew). Fox, Gwendoline. 2008. “Describing Yiddish Word Order Using Topological Fields.” In Proceedings of LingO, ed. Miltiadis Kokkonidis. Oxford: Faculty of Linguistics, Philology, and Phonetics, University of Oxford. Grosz, Barbara J. & Candace L. Sidner. 1986. “Attention, Intentions, and the Structure of Discourse.” Journal of Computational Linguistics 12.3: 175–204. Hinterhölzl, Ronald. 2010. “Information Structure and Unmarked Word Order in (Older) Germanic.” In Information Structure from Different Perspectives, eds. C. Féry & M. Zimmermann. Oxford: Oxford University Press, 282–304. Kaltenboeck, Gunter. 2011. “Explaining Diverging Evidence: The Case of Clause-initial I Think.” In Converging Evidence: Methodological and Theoretical Issues for Linguistic Research, ed. D. Schönefeld. Amsterdam: John Benjamins Publishing Company, 81–112. Mittwoch, Anita. 1979. “Final Parentheticals with English Questions—Their Illo­ cutionary Function and Grammar.” Journal of Pragmatics 3: 401–412. Mulder, Jean & Sandra A. Thompson. 2006. “The Grammaticization of But as a Final Particle in English Conversation.” In Selected Papers from the 2005 Conference of the Australian Linguistic Society, ed. Keith Allan., 1–18. Accessed June 20, 2014. Prince, Ellen. F. 1981. “Toward a Taxonomy of Given/New Information.” In Radical Pragmatics, ed. Peter Cole. New York: Academic Press, 223–255. ———. 1992. “Left Dislocation in English Discourse: One Form Three Functions.” Paper presented at the Israel Association of Theoretical Linguistics annual meeting.

The Right Periphery in Colloquial Hebrew


Ross, John R. 1967. Constraints on Variables in Syntax. PhD dissertation, Pittsburgh, PA: MIT. Vikner, Sten. 2001. “Verb Movement Variation in Germanic and Optimality Theory.” Habilitation Thesis, University of Tübingen. Ziv, Yael. 1994. “Left and Right Dislocations: Discourse Functions and Anaphora.” Journal of Pragmatics 22: 629–645. ———. 2001. “It Ain’t Simple (‫ – פשוט‬pashut): Discourse Markers in Colloquial Hebrew.” Hebrew Linguistics 48: 17–29 (in Hebrew). ———. 2006. “Hebrew nachon (True) and ma (what): Codification of Givenness and Surprise as Tools in Information Processing.” In Studies in Language 10, eds. Aharon Maman & Steven Fassberg, 65–73 (in Hebrew). ———. 2007. “The Discourse Markers naxon and lo: Linguistic and Rhetorical Characterization.” Hebrew Linguistics 54: 21–27 (in Hebrew). ———. In press. “Selected Comment Expressions from CoSIH—Distribution, Functions, and Analytical Challenges.” In Studying Spoken Hebrew, ed. Einat Gonen. Teuda. Tel Aviv: Tel Aviv University (in Hebrew).

Patterns of Dislocation: Judeo-Arabic Syntactic Influence on Modern Hebrew Yehudit Henshke

Hebrew Language Department, University of Haifa, Israel [email protected]

Abstract This article deals with a phenomenon of Modern Hebrew that exhibits the influence of Judeo-Arabic: the phenomenon of dislocation as found in the Hebrew sociolect of the Israeli periphery, among descendants of Middle Eastern and North African Jews. I call this sociolect Israeli Periphery Hebrew (IPH). The article examines the widespread use of dislocation constructions in IPH—specifically pronominal dislocation, as well as echo and anchoring constructions—and their unique features. Even though dislocation is typical of spoken language in general, it is argued here that its frequency and unique constructions in IPH reflect the influence of the Judeo-Arabic substrate. The article attempts to illuminate the sources and roots of these dislocated constructions.

Keywords dislocation – pronominal dislocation – echo constructions – anchoring constructions – Modern Hebrew – Hebrew of the Israeli periphery – Judeo-Arabic

Introduction Although some of its features have, over time, trickled into General Israeli Hebrew (GIH),1 the influence of the Judeo-Arabic substrate on Modern Hebrew is most visible in the Hebrew of the Israeli periphery (IPH),2 namely, 1  This term was coined by Haim Blanc (1957, 1964) to describe the spoken Israeli Hebrew that is grounded mainly in the language of speakers of European origin. 2  The term “periphery” refers here to geographical and social periphery alike. The term IPH denotes the language of Israel-born native Hebrew speakers who are descendants of immi-

© koninklijke brill nv, leiden, ���6 | doi ��.��63/9789004310896_012

Patterns of Dislocation


in the speech of the descendants of speakers of Judeo-Arabic (Henshke 2013). Examined here is one feature of Modern Hebrew that can be attributed to the influence of Judeo-Arabic: the phenomenon of dislocation as exemplified in the Hebrew sociolect of the periphery. The data presented here are based on a broad field survey that I conducted among second and third-generation descendants of North African immigrants to Israel and on dialogues culled from written literature. Considering that dislocation constructions are very frequent in the JudeoArabic of North African Jews (Moshe Bar-Asher [personal communication]; Akun 2015:9–10; Caubet 1993:1:227–228, 2:5, among others), it is only natural that such constructions entered the Hebrew of Israelis whose parents or grandparents spoke Judeo-Arabic. Indeed, my examination of their Hebrew showed widespread use of means of dislocation in IPH, compared to GIH, and, moreover, unique dislocation constructions that are otherwise rare in Modern Hebrew. Several prominent constructions are examined here: widely occurring constructions involving the dislocation of personal and demonstrative pronouns, as well as echo and anchoring constructions. I claim that even though dislocation typifies spoken language in general, its multiplicity and unique constructions in the Hebrew of the periphery display the influence of the Judeo-Arabic substrate.

Personal Pronoun Dislocation

As noted, in the periphery some dislocated clauses are distinguished from GIH only by the frequency of their use (see Bar 1997 for a detailed discussion of the phenomenon). If tensed clauses with a pronominal subject are not widespread in GIH (Bar 1997:313–15), in the Hebrew of the periphery they constitute an active, dominant category with several syntactic variations, as outlined below. I analyze the pronominal subject as being a dislocated element resumed by various types of resumptive elements: a) verbal person-number-gender inflection; b) inflected prepositions.

grants from the Middle East and North Africa, many of whom live in development towns, small towns, and moshavim (communal agricultural settlements) in the north and south of Israel, or in distinctive urban neighborhoods.

148 (1)


Verbal Clitic as Resumptive Element .‫אני עכשיו הגעתי לבית שלך‬

ani ʕaxšav hegaʕti la-bait šel-ax3 I now arrive.pst.1sg to.the-house of-you ‘Me, now I have reached your house.’ (R., Jerusalem)4

All strata of Hebrew, including GIH, refrain from the use of a redundant pronominal first or second person subject without special cause (Cohen 1990:53; Bolozky 1984:126). Various aspects of the sporadic presence of this construction in verbal clauses in GIH have been noted and analyzed in the literature. Some note its distribution in terms of person and tense (Bolozky 1984); others regard it as a means of enhancing discourse accessibility, noting pragmatic motivations for its use. Among the latter, the most frequently noted motivation is topic shift. A tangential motivation for its use is to focus on the person performing the action (Ariel 1990:120; Polak-Yitzhaki 2004:33–34, 69);5 others observe the expressive nature (complaint, warning, insult) of the use of the extra pronominal subject (Bar 1997:314). Nonetheless, all agree that this structure is infrequent in spoken GIH and even rarer in written Hebrew.6 In IPH, however, I found widespread, explicit use of the personal pronoun preceding verbs in the past (and future) tense, whose source I attribute to Judeo-Arabic, as in the following Judeo-Arabic constructions, for example: (2) nti šǝfti-h waqt li- ža? you time ‘You, did you see him when he came?’ (3) ana ma mšit-ši l-l-ʕarš I neg go.pst.1sg-neg to-the-wedding ‘Me, I did not go to the wedding.’

3  The transcription of the examples is faithful to the phonology of IPH, rather than to GIH or to Hebrew spelling. 4  Informants are referred to in this article by their initial and town. 5  In Modern Hebrew this is the sole motivation for this redundancy. See Cohen 1990. 6  On the limited scope of this redundancy in spoken Hebrew, see Polak-Yitzhaki 2004:34.


Patterns of Dislocation

In IPH, this construction occurs in both first-person and second-person pronouns, but more frequently with the former.7 Regarding the rationale for this redundancy in IPH, only in a minority of cases can we discern the pragmatic or emotional motivation that is mentioned in the literature. Most of the occurrences serve to direct attention to the subject of the predicate, but this effect is not necessarily very strong; indeed, it is often almost imperceptible, as in the following examples: (4) (5)

.‫אני תמיד הייתי עובדת‬

ani tamid haiti I always ‘Me, I have always worked.’ (Motzafi-Haller 2012:39)


.‫אני כשהייתי בהצגה עם ציפי לא ידעתי שהיא מרוקאית‬

ani kše-hayiti ba-hacaga ʕim cipi I when-was.1sg in.the-play with Tsipi lo yadaʕti še-hi marokait neg know.pst.1sg that-she Moroccan.f ‘Me, when I was at the play with Tsipi, I didn’t know that she was Moroccan.’ (H., Yeruham)

These utterances are typical of the IPH sociolect. In GIH the equivalent utterances would be devoid of the initial pronoun. Inflected Prepositions as Resumptive Elements Dislocation of personal pronouns is also evident in verbal clauses in which the resumptive element is not the verbal inflection, but is an object or possessive pronoun; namely, the left dislocation is of a nominative form which is resumed by a pronoun in the direct/indirect object or possessor position. For example: (6) .‫הוא עוד מדברים עליו שנשאר מלך‬ hu ʕod medabrim ʕal-av še-nišar

melex he still of-him king

‘Him, they still say of him that he remained a king.’ (Shilo 2005:16)

7  A similar tendency with future verbs is also found in GIH. See Bolozky 1984:128–129. I did not provide examples of the third-person construction, because they are widespread in Hebrew (Bolozky 1984).



(7) .‫ תעזוב אותי פה רגע עם איציק‬,‫אני‬ ani taʕazov ot-i po regaʕ ʕim icik I acc-me here minute with Itzik ‘Me, you leave me here for a minute with Itzik.’ (Shilo 2005:55) (8) 8.‫אני הבנים שלי למדו בחינוך העצמאי‬ ʔani ha-banim šəl-i lamdu ba-ħinnux I of-me in.the-education ha-ʕacmaʔi the-independent ‘Me, my sons studied in the independent education system.’ (9) 9.‫אני המשפחה שלי עלתה ממרוקו ומתורכיה‬ ani ha-mišpaħa šel-i ʕalta I of- me mi-maroko ve-mi-turkiya and-from-Turkey from-Morocco ‘Me, my family immigrated [to Israel] from Morocco and Turkey.’ (10) .‫אני הסבתא שלי לא ידעה קרוא וכתוב‬ ani ha-safta šel-i lo yadʕa I the-grandmother of-me neg kro u-xtov read.inf and-write.inf ‘Me, my grandmother did not know how to read and write.’ (11) . . . ‫אני כואב לי כשאומרים‬ ani koev l-i kše-omrim I to-me ‘Me, it hurts me when they say . . .’ The examples above are verbal clauses; however, dislocation of nominative pronouns is also found in nominal possessive constructions ( yesh li ‘I have’/ 8   Rabbi Ovadia Yosef in an interview, (accessed October 6, 2014). 9  Examples (9)–(11) are cited from Schwarzwald (1986:332). Interestingly, Schwarzwald’s main informant for examples of this sort is a speaker whose family came from an Arabic-speaking land.

Patterns of Dislocation


ʔen li ‘I don’t have’). This dislocated construction is not found in GIH and is marked as typical of IPH. For example: (12)

‫אני יש לי חור בשכל‬

ani yeš l-i xoʁ ba-sexel I is to-me hole in.the-brain ‘Me, I have a hole in my brain.’ (R., Jerusalem)

(13) .‫היא אין לה את הבעיה עם השם שלי‬ hi en l-a et ha-beʕaya ʕim ha-šem šel-i she neg to-her acc the-problem with the-name of-me ‘Her, she doesn’t have a problem with my name.’ (Shilo 2005:178) (14) 10. . . ‫אני הייתה לי‬ ani hayta li I me ‘Me, I had . . .’ (Busi 2000) (15)

11.‫אני לא היה לי שעון‬ ani lo haya l-i šaʕon I neg was to-me watch ‘Me, I didn’t have a watch.’

The latter type of examples (12)–(15) has direct parallels in Judeo-Arabic: (16) ʔana ʕnd-i tlat bnat I by-me three ‘Me, I have three daughters.’

10  Citation from Muchnik 2004:10–11. She terms these “substandard expressions from spoken language . . . and expressions . . . typical of the uneducated.” 11  Rabbi Ovadia Yosef in an interview, (accessed October 6, 2014).



Demonstrative Pronoun Dislocation

The use of dislocated demonstrative pronouns is very frequent in IPH. Unlike the semantically empty expletive pronouns found in GIH sentences such as ze raq ba-ʿerev she-hi kazot, ‘it is only in the evening that she is like that’ (Halevy 2006; Bar 1997:310–11),12 dislocated demonstratives in IPH point to persons and generally have a specific referent known to the speaker and hearer. In Bar’s study of dislocation in Israeli Hebrew (Bar 1997:310–11), this pattern is not mentioned. Here, too, the influence of Arabic is readily apparent (Caubet 1:168–69). The following examples are taken from spoken Judeo-Arabic, the first of right dislocation and the second of left dislocation: (17) li tquli-lu yaʕməl-ək ada that do. -(for) you ‘Whatever you tell him he will do for you, that one.’ (18) hadi li ʕnd-a taʕṭ-ek that by-her -(to) you ‘That one, whatever she had she will give to you.’ In IPH as well, these dislocated elements are found in both left and right dislocation.

Right Dislocation

(19) ?‫איך הוא אמר זה‬ ex hu amar ze how he ‘How did he say, that one?’ (P., Yeruham) (20) .‫היא עדינה זאתי‬ hi ʕadina zoti she delicate ‘She’s delicate, that one.’ (Shilo 2005:69) 12  In Bar-Aba’s (2010:186–209) examples, most instances of ‫ זה‬ze ‘it,’ are not interchangeable with proper nouns. The few examples of this sort that she does present (Bar-Aba 2010:195–196) are not associated with speakers of a particular social background.

Patterns of Dislocation


.‫היא יכולה לבכות זאתי‬ hi yxola livkot zoti she cry.inf ‘She can cry, that one.’ (Shilo 2005:73)


.‫היא מחורפנת לגמרי זאתי‬ hi mexurfenet legamre zoti she completely ‘She’s completely screwed up, that one.’ (Adaf 2008:206)



Left Dislocation .‫ אשתו לובשת סינר‬,‫ מתי שהוא מגיע למטבח‬,‫זה‬ ze matay še-hu magiyaʕ la-mitbax when that-he to.the-kitchen išt-o lovešet sinar wife-his apron ‘That one, when he gets to the kitchen his wife wears an apron.’ (P., Yeruham)

(24) .‫ זה לא נכון‬,‫ כל מה שהוא אומר לך‬,‫זה‬ ze kol ma še-hu omer l-ax everything what that-he tell.ptcp.m.s to-you ze lo naxon this(it)13 neg correct ‘That one, everything he tells you, it’s not correct.’ (G., Yeruham) (25) .‫ הכל אצלה הצגה‬,‫זאתי‬ zoti ha-kol ecl-a hacaga the-everything by-her act ‘That one, everything is a show for her.’ (Shilo 2005:171)

13  This ze is a semantically empty expletive parallel to the expletive common in GIH, and not the demonstrative pronoun discussed in this section.



(26) .‫ ציפרניים של זאב יש לה‬-‫וזאתי ימח שמה‬ ve-zoti yimmaħ šem-a cipornayim šel name-her of

zeʾev yeš l-a wolf have to-her ‘And that one, may her name be cursed: claws of a wolf she has.’ (Adaf 2008:167)

The Echo Construction

A particularly striking construction is the echo construction (Taube 1997),14 also called the “sandwich construction” (Azar 1992:96), as in (27): (27) ?‫אתה רעב אתה‬ ata raʕev ata? hungry you you ‘You, are you hungry?’ (Oz 1983:40) On the face of it, example (27) might seem to be a case of right dislocation that just happens to duplicate the resumptive pronoun. However, though historically these sentences may indeed be analyzed as simple cases of dislocation, I maintain that in IPH they constitute a distinct construction: a fixed structure involving an echo, i.e., an exact duplication of the pronoun (whether personal or demonstrative). The essential feature of this structure is the full echo, which produces an intensifying effect. The echo construction, which occurs with personal and demonstrative pronouns, is typical of and very frequent in IPH, and appears in both declarative and interrogative sentences.15 Some of these clauses still retain their expressive function;16 in others, however, the expressive function has been eroded, leaving only the repetitive construction. 14  Taube (1997), who discusses the echo construction in Yiddish, notes that in Hebrew these repetitive constructions do not derive from Yiddish but rather from Arabic dialects (Taube 1997:418). The speakers who use it share a Judeo-Arabic substrate. 15  Contra Azar (1992:96), who argues that most occurrences of this construction are interrogative. 16  Bar-Aba (2010:171–72) maintains that this utterance expresses denigration or exhortation. This is not the case in IPH.


Patterns of Dislocation

Even though this construction exists in the Bible,17 its prevalence in IPH is clearly due to the influence of Judeo-Arabic. This construction is common in North African Arabic, e.g.: (28) hada mliḥ hada good ‘That’s good, that!’ (29) hadi ṣla hadi synagogue ‘That’s a synagogue, that.’ (30) ana mǝžnuna ana? I crazy I ‘I’m crazy, am I?’ In the Hebrew of the periphery this construction can take two forms, simple and complex: Simple Echo Construction This construction is composed of a clause with the addition of a personal pronoun duplicated at the beginning or the end, as in (27) above, and in the following examples: (31)

?‫למה את באה אתי את‬ lama at baʾa it-i at why with-me ‘Why are you coming with me, you?’ (A., Jerusalem)

(32) .‫את צריכה להיזהר את‬ at crixa le-hizaher you. need to-careful.inf ‘You have to be careful, you.’ (Adaf 2008:163)


17  Psalm 76:8: ‫ אתה נורא אתה‬ˀatta nora ˀatta ‘Oh You! You are awesome!’

156 (33)


.‫ גרוש הוא לא היה שם בצד‬,‫הוא‬ hu gruš hu lo haya sam ba-cad he penny he neg was put in.the-side ‘Him, a penny he didn’t set aside.’ (Shilo 2005:21)

Complex Echo Construction This construction is composed of a base clause with the addition of two identical echo pronouns at both ends. Again, these sentences may have developed through a process of dislocation. For example, the simple clause ma ani yaʕase, ‘what I’ in (34) became ani ma ani yaʕase, ‘me what I do,’ through left dislocation (or, alternatively, ma ani yaʕase ani through right dislocation), and then an additional process of dislocation yielded a second pronoun at the other end, forming an echo structure (ani ma ani yaʕase ani). However, I believe that today these sentences are instances of the independent echo construction as well. (34) ?‫אני מה אני יעשה אני‬ ani ma ani yaʕase ani? I what I I ‘Me, what will I do -, me?’ (G., Jerusalem) (35)

‫אני לא חסר לי שכל אני‬

ani lo ḥaser l-i sexel ani I neg lack to-me intelligence I ‘Me, I’m not lacking in intelligence, me.’ (Motzafi-Haller 2012:158)

Echo constructions (both simple and complex) can also feature demonstrative pronouns: Simple (36)

.‫זה מוסיקה טובה זה‬ ze muzika tova ze music good ‘That’s good music that.’ (M., Jerusalem)

Patterns of Dislocation



.‫זה סיפור מהחיים זה‬ ze sipur me-ha-ħaim ze story from-the-life ‘That’s a story from life that.’ (Motzafi-Haller 2012:35)

Complex (38) .‫זה הכלכלה שלו יקר זה‬ ze ha-kalkala še-lo yakar ze dem.m.sg18 the- economy of-him expensive ‘That one, supporting him is expensive, that one.’ (A., Yeruham)

The Anchoring Construction

Another dislocation construction that singles out one part of the sentence is the anchoring construction, involving a chain of social relationships. For example: (39) .‫ שכנים שלה לומדים אצלנו‬,‫ החברה שלו‬,‫שכן שלי‬ šaxen šel-i ha-xavera šel-o šxenim šel-a neighbor of-me the-friend of-him of-her

lomdim ecl-enu by-us ‘The neighbors of the friend of my neighbor study with us.’ Literally: ‘My neighbor, his friend, her neighbors study with us.’ (Ziv 2010:43)

This construction has received partial coverage in the literature (Ziv 2010), but no attention has been paid to the overall syntactic picture. Treating it here as part of the phenomenon of dislocation that typifies IPH, I examine its significations, uses and sources. Anchoring is a discourse construction that seeks to create an anchor for connectivity and for enhanced accessibility of discourse entities (Ziv 2010: 18  In its two appearances in the sentence ze refers to a person.



43–45). It is frequent in Judeo-Arabic and in the Hebrew of Israelis with origins in Arabic-speaking lands, and has even found its way into spoken GIH. Its advantages in terms of processing are readily apparent. Nonetheless, note that this is not an innovation, but rather a borrowing of a widespread Judeo-Arabic construction. Moreover, in Hebrew this construction serves the same purpose as in Judeo-Arabic, especially in the description of relationships in real or imagined space. It always opens with a coordinate known to the interlocutors and progresses two or three stages in order to map relationships. An additional example appears below: (40) .‫ פתחה מספרה‬,‫ גיסתה‬,‫ אחותו‬,‫אילן דדון‬ ilan dadon axot-o gisat-a patxa Ilan Dadon sister-his sister-in-law-her maspeʁa ‘The sister-in-law of Ilan Dadon’s sister opened a hairdressing salon.’ Literally: ‘Ilan Dadon—his sister—her sister-in-law; she opened a hairdressing salon.’ (H., Netivot) Thus, as we saw above, in discourse speakers use these dislocated utterances, which progress from one entity to the next, in order to guide the addressee to the destination by clarifying the link between each two entities in the chain.

A Presentative Construction

An additional type of anchoring construction involves locative phrases. Speakers of IPH who have to map out a route for their interlocutor note the main coordinates in nominal form, sometimes preceded by a presentative word. They sometimes use a combination of Hebrew and Arabic, as in (41), or only Hebrew, as in (42): (41) .‫דאר‬-‫תחנה א‬-‫א‬ ha-t-taḥana ha-d-dar here-the-station here-the-house ‘Here’s the station—here’s the house.’ (H., Yeruham)

Patterns of Dislocation



.‫הנה התחנה הנה הבית‬ hine ha-taḥana hine ha-bayit here the-station here the-house ‘Here’s the station—here’s the house.’ (S. Regev)19

Conclusion In sum, this article has examined a syntactic phenomenon in Modern Hebrew: the frequency of dislocated elements in IPH and the unique forms found only in this sociolect. Attributing this phenomenon to Judeo-Arabic sources has shed new light on the roots of the use of dislocation in IPH. The article has also reconsidered some elements sporadically addressed in the literature, such as echo and anchoring constructions, viewing them as part of a general trend of dislocation that draws on the Judeo-Arabic substrate. References Adaf, Shimon. 2008. Sunburnt Faces. Tel Aviv: Am Oved (in Hebrew). Akun, Natali. 2015. “Ha-aravit ha-mishtaqefet ba-ivrit shel Marocco.” Studies in the Culture of North African Jewry, Edited and Annotated Texts, eds. M. Bar-Asher & S. Fraade. Jerusalem: The Program in Judaic Studies, Yale University, New Haven and The Center for Jewish Languages and Literatures, The Hebrew University, Jerusalem, 29–52 (in Hebrew). Ariel, Mira. 1990. Accessing Noun-Phrase Antecedents. London: Routledge. Azar, Moshe. 1992. “Liqrat havanat mivneh ha-mishpat ha-memuqad be-ivrit bat zemanenu.” Hebrew: A Living Language 1: 87–99 (in Hebrew). Bar, Tali. 1997. “Extraposition in Contemporary Hebrew.” Lĕšonénu 60: 297–328 (in Hebrew). Bar-Aba, Esther Borochovsky. 2010. Issues in Colloquial Hebrew. Jerusalem: Bialik Institute (in Hebrew). Blanc, Haim. 1957. “Qetaʿ shel dibbur ‘ivri yisra’eli,” Lĕšonénu 21: 33–39 (in Hebrew). ———. 1964. “Israeli Hebrew Texts.” In Studies in Egyptology and Linguistics: In Honor of H. J. Polotsky, ed. H. B. Rosen. Jerusalem: Israel Exploration Society, 123–152.

19  See (accessed October 3, 2010).



Bolozky, Shmuel. 1984. “Subject Pronouns in Colloquial Hebrew.” Hebrew Studies 25: 126–130. Caubet, Dominique. 1993. L’arabe marocain, tombe I: Phonologie et morphosyntaxe; tombe II 2: Syntaxe et catégories grammaticales, Textes. Paris-Louvain: Peeters. Cohen, Chaim E. 1990. “The Independent Pronoun as the Subject of a Definite Verb in Tannaitic Hebrew.” Proceedings of the Tenth World Congress of Jewish Studies, Division D, vol. 1. Jerusalem: World Union of Jewish Studies, 53–59 (in Hebrew). Halevy, Rivka. 2006. “The Function of Nonlexical Zeh in Contemporary Hebrew.” Lešonénu 68: 283–307 (in Hebrew). Henshke, Yehudit. 2013. “Peripheral Hebrew: Between Error and Loan.” Language Studies, 14–15: 175–186 (in Hebrew). Motzafi-Haller, Pnina. 2012. In the Cement Boxes: Mizrahi Women in the Israeli Periphery. Jerusalem: Magnes (in Hebrew). Muchnik, Malka. 2004. “Tel Aviv ‘Southerner’ Dialect in The Moon Goes Green in the Wadi.” Helkat Lashon 35: 5–19 (in Hebrew). Oz, Amos. 1983. A Journey in Israel Autumn 1982. Tel-Aviv: Am Oved (in Hebrew). Polak-Yitzhaki, Hilla. 2004. “The Functions of Subject Personal Pronouns.” M. A. thesis, University of Haifa (in Hebrew). Schwarzwald, Ora. 1976. “Acceptability and Formation of Topicalized Sentences in Hebrew.” Bar Ilan 13: 321–340 (in Hebrew). Shilo, Sarah. 2005. No Gnomes Will Appear. Tel Aviv: Am Oved (in Hebrew). Taube, Moshe, 1997. “Echo-construction in Yiddish.” Massorot 9–11 (Gideon Goldenberg Festschrift): 397–420 (in Hebrew). Ziv, Yael. 2010. “Anchoring in Discourse Model.” Hebrew Linguistics 64: 37–47 (in Hebrew).


Superfluous Negation in Modern Hebrew and Its Origins Aynat Rubinstein,a Ivy Sichel,b and Avigail Tsirkin-Sadana

aMandel Scholion Interdisciplinary Research Center in the Humanities and Jewish Studies, The Hebrew University of Jerusalem, Jerusalem, Israel bDepartment of Linguistics and Program in Cognitive Science, The Hebrew University of Jerusalem, Jerusalem, Israel [email protected]; [email protected]; [email protected]

Abstract In this article, we survey a variety of constructions in contemporary Modern Hebrew that include seemingly superfluous instances of negation. These include free relatives, exclamative rhetorical questions, clausal complements of ‘until,’ ‘without,’ and ‘before,’ clausal complements of ‘fear’-type verbs, after negated ‘surprise,’ and the complement of ‘almost’ (a construction by now obsolete). We identify possible sources for these constructions in pre-modern varieties of Hebrew. When an earlier source cannot be found, we examine earliest attestations of the constructions in modern-era corpora and consider the role of contact (primarily with Yiddish and Slavic) in their development.

Keywords negation – superfluous negation – expletive negation – Modern Hebrew – language contact

* We thank Chanan Ariel, Edit Doron, Aviad Eilam, Yehudit Henshke, Samir Khalaily, Abed Al-Rahman Mar’i, Moshe Taube, and two anonymous reviewers for their input during the development of this article. Support from the Mandel Scholion Interdisciplinary Research Center in the Humanities and Jewish Studies at the Hebrew University of Jerusalem is gratefully acknowledged.

© koninklijke brill nv, leiden, ���6 | doi ��.��63/9789004310896_013


Rubinstein, Sichel and Tsirkin-Sadan

Introduction Superfluous negation (henceforth Super Neg) is the general term we will use for an instance of negation that appears not to have the usual reversal effect on the truth conditions of the containing sentence.1 While we believe there are reasons to suspect that this is not a unitary phenomenon (i.e., that, synchronically, not all the kinds of Super Neg that we have identified have the same underlying analysis), we will not attempt an analysis of the various constructions here. Our goal in this short contribution is much more modest. In the following section, we survey the constructions in which Super Neg is observed in contemporary Modern Hebrew. We then ask whether these constructions (and an additional construction, which is by now obsolete) existed in earlier stages of Hebrew and, if not, when they entered the modern language. We identify possible contact-induced sources for Super Neg, focusing primarily on Yiddish and Slavic. Survey Contemporary Modern Hebrew exhibits Super Neg with the negative morpheme ‫ לא‬lo in a variety of constructions.2 The negative morpheme generally resists stress when it is “superfluous” (Avinery 1964:242, 253; Eilam 2009).3 Free Relative Clauses Super Neg is observed with relativization from subject, object, and adjunct positions, and with a variety of interrogative pronouns (e.g., ‫ מה‬ma ‘what,’ ‫ מי‬mi

1  The phenomenon is variably referred to in the literature as expletive, pleonastic, redundant, supplementary, or paratactic negation. The broad definition given above may very well include instances of negation that do have the regular semantic contribution, though masked by other factors. An anonymous reviewer correctly points out, for example, that negation in Y/N questions could be considered superfluous by this broad criterion, even though it still probably has its regular semantic contribution. We leave for future study the proper analysis of the instances of Super Neg identified below, along with the question whether in all or any of them negation truly sheds its normal semantic contribution. 2  In some of these constructions, the negative marker can also have its usual contribution. We set such uses aside. Other negative morphemes in Modern Hebrew do not support Super Neg (see below). 3  The order of presentation of Super Neg constructions roughly represents the amount of attention that the various constructions received in the literature.

Superfluous Negation in Modern Hebrew and Its Origins


‘who,’ ‫ איזה‬ʔeyze ‘which,’ ‫ איפה‬ʔeyfo ‘where,’ ‫ מתי‬matay ‘when,’ ‫ איך‬ʔeyx ‘how’).4 With negation, the sentence conveys that the claim does not depend on the precise identity of the free relative’s referent (Eilam [2009] notes, employing terminology from the literature on free relatives crosslinguistically, that in Hebrew the inference is primarily one of indifference rather than ignorance).5 (1)

.‫מה שדני לא כתב התפרסם בעיתון‬ ma še-dani lo katav hitparsem ba-ʕiton what that-Danny neg wrote was.published in.the-newspaper ‘Whatever Danny wrote was published in the newspaper.’ (Eilam 2009)

Rhetorical Questions Expressing Surprise or Noteworthiness These interrogatives are used as exclamatives and convey a universal implication (in (2), he was blamed for everything).6 (2)

!?‫במה לא האשימו אותו‬ be-ma lo heʔešimu ʔoto in-what neg blamed.3mpl him ‘The things he’s been blamed for!’

Clausal Complements of ‫ עד‬ʕad ‘until’ The presence of negation contributes the sense that there is a non-accidental connection between the ‘until’-clause event and the matrix event (Eilam 2009; Avinery 1964) such that the ‘until’-clause describes a necessary condition for a change in the main-clause event. In (3), the sentence conveys that the team

4  Why-free relatives are ungrammatical in Hebrew, as in many languages (see Citko 2010:222 on Polish; Larson 1987), hence ‫ למה‬lama ‘why’ is not included in the list. See Francez (2015) on negative lama interrogatives and their interpretation as positive suggestions. 5  A note about glossing: since we do not commit ourselves to a semantic account of the various uses of superfluous negation in this paper, we gloss negation simply as neg and rely on the English translations to reflect the fact that in these constructions it does not have the standard effect. 6  A theoretical question we set aside is whether rhetorical questions and exclamatives are grammatically similar or distinct crosslinguistically. In relying for classificatory purposes on formal properties rather than function or use, we follow Eilam’s (2009) classification of this construction as a negative rhetorical question. We translate the rhetorical question as an idiomatic English exclamative only because English lacks negative rhetorical questions of this sort. See also Tzivoni (1993:320–321).


Rubinstein, Sichel and Tsirkin-Sadan

is in such bad shape that it would take a Russian billionaire to put it back on track. Without negation, the ‘until’-clause receives its usual interpretation. (3) .‫ ניוקאסל תתקשה‬,‫עד שלא יגיע מיליארדר רוסי‬ ʕad še-lo yagiaʕ milyarder rusi nukasel until that-neg will.arrive billionaire Russian Newcastle titkaše will.have.trouble ‘Newcastle will be in trouble until a Russian billionaire comes along.’ (, accessed August 26, 2014) Clausal Complements of ‫ לפני‬lifney ‘Before’, ‫ בלי‬bli ‘Without’ There is more variation among speakers regarding the acceptability of these examples. For those who accept them, the negation contributes the sense of a necessary condition noted above for ‘until’ (in (4), leaving without an answer is not possible). (4) .‫אני לא רוצה שמישהו ייצא מפה בלי שהוא לא קיבל מענה על השאלות שלו‬ ʔani lo roce še-mišehu yece mi-po bli I neg want that-someone will.leave from-here without še-hu lo kibel maʕane ʕal ha-šeʔelot šelo that-he neg received response on the-questions his ‘I don’t want anyone to leave here without having gotten answers to his questions.’ (Protocol of the Tel Aviv-Yafo local Design and Building Committee meeting of Aug. 24, 2011, accessed August 31, 2014)7 Embedded under Negated ‘Surprise’ 8 Negation in the embedded clause is optional and is naturally used when the speaker takes issue with an opposing expectation in the discourse. In (5), for example, the expectation that ‘he’ may be behind the incident is considered by the speaker to be at odds with the prevailing view. Super Neg is restricted to sentences with future tense morphology in the matrix clause, which, notably, involve the complementizer ‘if,’ raising the possibility that the clause under ‘if’ is a conditional adjunct clause. In the past tense, ‘surprise’ takes an ordinary ‘that’-complement and Super Neg is not licensed. 7  See the website at‫מליאת פרוטוקול‬ ‫ סטנוגרמה‬- )‫הועדה המקומית (תוכנית בניין עיר‬.pdf. 8  The pattern may extend to other expressions of expectation, e.g., ‫ שאני אמות אם‬. . . še-ʔani ʔamut ʔim . . . ‘I’ll be damned (lit. dead) if. . . .’

Superfluous Negation in Modern Hebrew and Its Origins


(5) .‫אני לא אתפלא אם הוא לא יזם את כל התקרית המכוערת הזו‬ ʔani lo ʔetpale ʔim hu lo yazam ʔet kol I neg if he neg initiated acc all ha-takrit ha- mexoʕeret ha-zo the-incident the-ugly the-this ‘I wouldn’t be surprised if he is the one behind this ugly incident.’ (Haim Kadman, Škiʕa ʔafrikanit ‘African Sunset,’ 2010, http://cafe.the, accessed September 17, 2014) Clausal Complements of ‘Fear’-type Predicates Another somewhat restricted manifestation of Super Neg occurs in the complements of verbs like ‫ פחד‬paħad ‘fear,’ ‫ דאג‬daʔag ‘worry,’ and ‫ חשש‬ħašaš ‘worry’ (and derived nouns), mainly in colloquial language (Sagi 2000:95). (6) .‫פחדתי שלא ינדו אותי בגלל שאני ערביה‬ paħadti še-lo yenadu ʔoti feared.1sg that-neg will.ostracize.3mpl me biglal še-ʔani ʕarviya because that-I Arab.fsg ‘I was afraid I would be ostracized because I was Arab.’ (, accessed August 30, 2014) Clausal Complement of ‫ כמעט‬kimʕat ‘Almost’ This is an obsolete construction that was short-lived in early Modern Hebrew. It was used to describe near-disastrous events (Avinery 1964:253; Sagi 1997, 2000; Farstey 2006; in (7), the revival of a blood libel). (7) .‫] כמעט שלא נתחדשה לפני ימים אחדים עלילת הדם‬. . .[ ‫בעיר פרערוי‬ ba-ʕir freroy kimʕat še-lo nitħadša lifney yamim in.the-city Freroy almost that-neg was.renewed before days aħadim ʕalilat ha-dam ones libel.cs the-blood ‘The blood libel was almost/all but revived in the town of Freroy a few days ago.’ (Ha-melic, February 26, 1886)9

9  The town referred to is probably Kremsier in Mähren, today Kroměříž in the Czech Republic.


Rubinstein, Sichel and Tsirkin-Sadan

Origins of Super Neg Constructions: First Attestations and Contact

We begin with a brief overview of other cases of so-called ‘redundant’ or ‘repetitive’ negation that have been identified in Biblical and Mishnaic Hebrew. We then turn to the contemporary constructions given above. Although in a number of cases there exist sporadic pre-modern examples that resemble the contemporary uses, we suggest that these constructions were not inherited with superfluous negation from previous layers of Hebrew. We date the earliest attestation of the modern uses in our corpora (Historical Jewish Press [HJP] and the Ben-Yehuda Project [BYP]) and consider the plausible role of contact in their emergence. Biblical Hebrew Two types of redundant negation in Biblical Hebrew are discussed by grammarians (Gesenius 1910:483; Joüon & Muraoka 2006:573). The first is exemplified by ‫ ְּב ֶט ֶרם לֹא‬bə-ṭɛrɛm lō ‘before neg’ used to express temporal precedence.10 Although it resembles Modern Hebrew lifney še-lo, the lexical choice of preposition is different. Another, more productive, type is mi-blī/mē-ʾēn. It is analyzed as containing two negative morphemes, ‘miwithout-neg,’ that give rise to a single, emphatic, negative interpretation (e.g., mi-blī yōšēḇ meaning ‘(land) without inhabitants’). In our survey above, the closest counterparts of these two types are classified together (lifney/bli še-lo ‘before/without that-neg’). Note that while negation is realized as blī/ʾēn in the Biblical Hebrew construction, in Modern Hebrew it is restricted to lo (*bli (še-)ʔeyn). Mishnaic Hebrew Both Ben David (1967) and Azar (1995) mention the same phenomenon under the heading ‘repetitive negation,’ where negation is marked on each element of a conjunction in addition to matrix negation (e.g., ʾeyn meḇarḵin lo ʿal ha-ner ve-lo ʿal ha-beśamim šel noḵrim, ‘It is not allowed to recite a blessing neg over candles and neg over fragrances of foreigners’; Beraḵot 8:6). This is an interesting construction that exists in Modern Hebrew too, but we do not consider it an example of Super Neg. The repetition of negation in the two conjuncts conveys emphasis, on a par with English neither . . . nor, and the additional repetition of negation in the matrix clause may be an instance of negative concord, obligatory in Modern Hebrew in the context of N-words.11 10  Zephaniah 2:2. 11  N-words are Negative expressions such as nobody in English ‫ אף אחד‬or ʔaf ʔexad ‘nobody’ in Hebrew. Since Hebrew has negative concord, an N-word is necessarily accompanied by

Superfluous Negation in Modern Hebrew and Its Origins


First Attestations and Possible Contact Free Relative Clauses Super Neg in free relative clauses (FRs) is not entirely absent from the language of the Talmud (Avinery 1964:289) and is also attested in responsa of the early modern period (Sagi 1997, 1999).12,13 The construction is well attested in literary writing of the revival period, with early examples in Modern Hebrew found around the turn of the 20th century (several instances in Mendele’s Be-ʕemek Ha-baxa [1896–1908] and Susati [1909], Eliyahu Maidanik’s 1900 publications, and a 1902 letter by Yosef Vitkin). We observe a clear quantitative preference (43 out of 53 examples in BYP) for Super Neg in ‫ כמה‬kama ‘how many/much’ FRs over three other representative types (‘what,’ ‘who,’ and ‘which.msg . . .’). This may be noteworthy in light of Sagi’s (1999) finding that kama accounted for all examples of Super Neg FRs in the 16th–19th c. responsa, and kama together with ‫ איך‬ʔeyx ‘how’ accounted for the majority of relevant examples in the 20th c. responsa. It may also represent a preference for Super Neg in adjunct over argument FRs, a possibility that merits further investigation. Despite the existence of attested examples in the pre-revival era, grammarians of Modern Hebrew tend to view Super Neg in FRs as resulting from Yiddish or Slavic influence (Garbell 1930; Haspelmath & König 1998; see also Blanc 1956, 1965; Altbauer 1964; Sagi 2000; Eilam 2008, 2009). Haspelmath and König (1998) establish the areal nature of the phenomenon among certain eastern European languages. They speculate that Yiddish borrowed the construction from Russian, Polish, or Ukranian (pp. 615–616). Two points should be noted regarding the proposed borrowing from Yiddish into Hebrew. First, Yiddish has two types of FRs that Hebrew could have potentially borrowed, one expressed with expletive negation and one with the focus particle nor ‘only’ (Haspelmath sentential negation. We remain agnostic as to whether the negation that is interpreted in Negative Concord is the actual negative marker or some other, abstract, negative operator (Zeijlstra 2011). 12  It is not entirely clear that negation in the example cited by Avinery is indeed an instance of Super Neg (‫אמר להן המלך יגע זה לשתי שעות יותר ממה שלא יגעתם אתם כל היום‬ ‫כולו‬, ‘The king said to them: he worked in two hours more than you neg worked all day long,’ Jerusalem Talmud, Beraḵot 2:8; two other versions of this text lack the comparative yoter mi- ‘more than,’ and negation does not seem superfluous: ‘. . . he worked in two hours what you did neg work all day long’; (Šir ha-Širim Raba 6:2, Qohelet Raba 5:11). (Ch. Ariel, p.c.). 13  Sagi mentions four occurrences in 16th–19th c. responsa but does not cite specific examples. A cursory search in the current version of the Responsa Project revealed many more examples. Notably, ‫ כמה שלא‬kama še-lo is preceded by ‫ כל‬kol, ‘every’ in many of them.


Rubinstein, Sichel and Tsirkin-Sadan

& König 1998:613). Only the first type is attested in Modern Hebrew.14 Second, while the Yiddish constructions typically involve subjunctive marking on the verb (e.g., Vos er zol ništ zogn, gleybt zi im ništ ‘Whatever he would tell her, she doesn’t believe him,’ Schaechter 1986:321), Hebrew Super Neg FRs are found with a variety of tense-aspect markings from early on. Sentence (8) is a 1904 example of Super Neg in a past tense free relative. (8)

‫[ וכמה שלא התגעגעתי להיות פעם אחת בקונגרס ולראות את יוצרו‬. . .]

ve-xama še-lo hitgaʕgaʕti lihiyot paʕam aħat and-how.much that-neg longed.1sg time one ba-kongres ve-lirʔot ʔet yocro in.the-congress and-to.see acc his.creator ‘And however much I wanted to attend the [Zionist] Congress once and see its creator [. . .].’ (I.L. Peretz, Be-ʕolam ha-ʔotiyot ha-maħkimot, 1904)

Rhetorical Questions Expressing Surprise or Noteworthiness Non-questioning uses of interrogatives are well attested in Biblical, Rabbinic, and Medieval Hebrew (Moshavi 2013, 2014; Stadel 2013; Gryczan 2013), as is the specific use of interrogatives to express exclamation (e.g., with ‫ מה‬ma ‘what’ in Biblical Hebrew; Moshavi 2013). These examples do not contain superfluous negation, however, and therefore the Modern Hebrew construction seems not to have been inherited from these earlier varieties.15 Our searches reveal many examples in Hebrew literature already in the 19th century, with ‫ מי לא‬mi lo ‘who 14  An anonymous reviewer correctly points out that Modern Hebrew does have occurrences of ‘only’ FRs of the sort found in Yiddish: i. ʔani ʔeten lax ma še-rak tirci I will.give you what that-only you.will.want ‘I will give you whatever you want.’ This variety is restricted in Hebrew to particular verbs, and especially want, as in (i). Note that with other verbs, such as ask below, ‘only’ FRs are degraded in Modern Hebrew: ii. *ʔani ʔaʕane lax ʕal ma še-rak tišʔali I will.answer you on what that-only you.will.ask Intended: ‘I will answer whatever you ask.’ 15  One issue under debate in the literature is whether the Biblical Hebrew particle ‫ ֲהלֹא‬hălō, which had a non-negative presentative function in rhetorical questions, should be analyzed as a combination of a polar interrogative hă- and negation. See Driver (1973) for an early discussion and Gzella (2013), Moshavi (2013) for a recent evaluation and additional references.

Superfluous Negation in Modern Hebrew and Its Origins


neg’ attested as early as Judah Leib Gordon’s ʔahavat David U-Mixal (1856) and ‫ מה לא‬ma lo ‘what neg’ in Avraham Mapu’s ʔašmat Šomron (1865). (9)

!?‫נער אדמוני הוא ויפה עינים—דוד הרעה—מי לא יכירנו‬ naʕar ʔadmoni hu vi-yfe ʕeynayim david lad redheaded he and-beautiful.cs eyes David ha-roʕe mi lo yakirenu the-shepherd who neg will.know.him ‘He is a red headed lad with beautiful eyes, David the shepherd, who doesn’t know him?!’ (Judah Leib Gordon, ʔahavat David u-Mixal, 1856)

A distinct pattern of emergence is observed in comparison to FRs: the majority of examples are found with ‫ לא‬. . . ‫מה‬/‫ מי‬mi/ma . . . lo ‘who/what . . . neg’ (103/24 respectively in a sample of BYP) while examples based on ‫ כמה‬kama ‘how many/much’ are not attested. Despite the superficial similarity between the two constructions, this type of Super Neg also seems to have emerged somewhat earlier than the FR type. Eilam (2008, 2009) proposes that while negative rhetorical questions are common crosslinguistically and may have developed independently in Modern Hebrew, it is plausible that the construction was in fact calqued from Yiddish. Clausal Complement of ‫ עד‬ʕad ‘until’ Examples of ʕad še with a negated clause are attested since at least Mishnaic Hebrew (Braverman 1995:172–173; Morgenstern 2013; possibly from Aramaic; Rosén 1956:64), but with a temporal precedence meaning as in (10). Avinery (1964:443) argues that ʕad is a variant of ʕod ‘while’ in these cases, such that ʕad še-lo contributes a ‘while not’ or ‘before’ meaning.16 In contemporary Modern Hebrew, ʕad means ‘until’ and no longer has the ‘while’ meaning. Early Modern Hebrew inherited the rabbinic ʕad še-lo (Eilam 2008, 2009), with examples attested in our corpora from the 1860s (11).

16  According to BDB (p. 725), ‘ad in the sense of ‘while’ is also found in Biblical Hebrew (rare). There are three instances of ‫‘ ַעד ֲא ֶׁשר לֹא‬ad ʾăšɛr lō in the sense of ‘while not’ in the Bible (all in Ecclesiastes 12:1,2,6), and another occurrence with no complementizer (‫ַעד לֹא‬ ‘ad lō ‘while’ in Proverbs 8:26).

172 (10)

Rubinstein, Sichel and Tsirkin-Sadan

]. . .[ ‫עד שלא יתחילו במלאכה צא ואמור להם‬ ʿad še-lo yatħilu ba-melaḵa ṣe while that-neg will.start.3mpl in.the-work go.out.2msg ve-ʾemor lahem and-tell them ‘Before they start working, go out and tell them . . .’ (Original rabbinic use; Babylonian Talmud, Bava Meciʕa 83A)

(11) ‫חובה עלינו לתת תודתנו לאלה החכמים אשר קדמו לעזור לנו עד שלא קראנו אליהם‬ [. . .] ħova ʕaleynu latet todatenu le-ʔele ha-ħaxamim ʔašer duty to.give our.thanks to-those the-wise who kadmu laʕazor lanu ʕad še-lo karanu ʔeleyhem were.early while that-neg we.called to.them ‘We are obliged to the wise who were early to help us before we asked them.’ (Rabbinic type; Ha-karmel, May 1, 1868) A random sample of examples in the Mishnah and Babylonian Talmud revealed a preference for verbs with past tense morphology in the adjunct. This preference seems to have been maintained in early Modern Hebrew (e.g., (11)), still with the ‘before’ meaning. The current Super Neg use is observable in the BYP and HJP from the 1880s: (12) ‫ ע"כ נצטוה [הנציב] לבלי צאת מסטשוואן עד שלא יוחלט הדבר בנוגע לתשלומי‬. . . .‫נזקי המיסיונרים‬

ʕal-ken nictava [ha-naciv] li-vli cet therefore was.ordered.3msg [the-commissioner.3msg] to-neg exit mi-setšuan ʕad še-lo yuħlat ha-davar be-nogeaʕ from-Sichuan until that-neg the-issue regarding le-tašlumey nizkey ha-misyonerim to-payments.cs damages.cs the-missionaries ‘Therefore the commissioner was ordered not to leave Sichuan until compensation is settled for the damage done by the missionaries.’ (New type; Ha-melic, August 1, 1895) Some of the modern examples utilize the complementizer ʔašer (of Biblical origin). Note the non-past morphology in the ‘until’-clause in (12) and the sense of a non-accidental connection between the events mentioned (recall

Superfluous Negation in Modern Hebrew and Its Origins


(3) above). As is also typical of the contemporary Super Neg use, negation can be omitted in (12) with only a slight change in meaning. Both the Rabbinic type and the new type of ʕad še-lo coexisted for a while in early Modern Hebrew, but by 1920 the new type had become prominent, accounting for over eighty percent of occurrences in our sample. As the new type continued to expand in Modern Hebrew, the Rabbinic type diminished and became rare and archaic (though examples are still found in 1937 and even later, e.g., in the Zionist Orthodox newspaper Ha-cofe, probably attributable to its readership’s familiarity with the Mishnah and Talmud). While the contemporary, new ʕad še-lo could in principle be seen as a language-internal development, i.e., as a semantic narrowing of an old Hebrew form, it is notable that Yiddish (Schaechter 1986:321) and Russian (Timberlake 2004:464; Abels 2005; Wade 2011:501–502) both have similar Super Neg constructions. Yiddish has been suggested as the source of the Modern Hebrew calque (Eilam 2008, 2009), and seems the more likely source of influence, since negation in the ‘until’-clause is optional in Yiddish, as in Modern Hebrew, but obligatory in Russian.17 Besides the main semantic shift, the greater variety of tense marking in the ʕad še-lo adjunct could be a reflection of the same freedom in Yiddish and Russian ‘until’-clauses (Schaechter 1986; Abels 2005; Wade 2011:502). Clausal Complements of ‫ לפני‬lifney ‘Before’, ‫ בלי‬bli ‘Without’ These forms were not found in BYP or HJP and seem to be more recent. As far as we can tell, they are not mentioned in the literature on Yiddish and Slavic, but both are attested, for example, in German (Krifka 2010) and in French. Interestingly, in French as in Modern Hebrew, there is some disagreement between grammarians as to their acceptability (van der Wouden 1994; Sanchez Valencia et al. 1994). It is possible that these Super Neg uses are an extension of the ‘until’ construction discussed above and not a direct result of contact. Embedded under Negated ‘Surprise’ The expression of expectation using superfluous negation in ‘if’-clauses following negated ‘surprise’ appears to be a recent development of Modern Hebrew. While ‫ לא יִ ָּפלא‬lo yipale ‘neg will.surprise’ occurs quite frequently in BYP, negation in the ‘if’-clause has its usual truth reversal effect:18 17  We refer specifically to poka . . . ne clauses in Russian, in which poka ‘while, by the time’ is obligatorily followed by a negative morpheme to give the meaning of ‘until.’ 18  Verbal patterns searched for included the roots ‫א‬.‫ל‬.‫ פ‬p.l.ʔ, ‫מ‬.‫מ‬.‫ ׁש‬š.m.m, and ‫י‬.‫א‬.‫ ׁש‬š.ʔ.y.


Rubinstein, Sichel and Tsirkin-Sadan

(13) ‫ועל כן לא יפלא כי רוב העם לא ידעו מה שכתוב במגלות ההן‬ ve-ʕal ken lo yipale ki rov ha-ʕam lo and-therefore neg that most the-people neg yedʕu ma še-katuv ba-megilot ha-hen will.know what that-written in.the-scrolls the-those ‘Therefore it is not surprising that most of the people do not know what is written in those scrolls.’ (Ordinary negation; Ephraim Deinard, Ha-yaʕar be-ʔeyn Dov, 1929) A similar construction exists in English and German (as in I won’t be surprised if he isn’t given a hard time),19 but, according to our informants, seems not to exist in Russian or in Yiddish. Clausal Complements of ‘Fear’-type Predicates In Rabbinic and Medieval Hebrew, complements of ‫א‬.‫ר‬.‫ י‬y.r.ʔ ‘fear’ and other verbs of this class were introduced by both ‫ שמא‬šema ‘lest’ (and other complementizers, e.g., ‫ פן‬pen) and ‫ שלא‬še-lo ‘that-neg’ (Avinery 1964:241–242; Sagi 2000:92). Super Neg uses of še-lo in this construction are attested in the early rabbinic texts (Avinery 1964:241ff.), in Medieval Hebrew (Goshen-Gottstein 2006:141–142), in the pre-Haskalah literature (e.g., Luzzatto’s Mesilat Yešarim, 1740), and from the mid-19th century throughout the revival literature (e.g., (14)). Our search in BYP retrieved thirteen relevant literary occurrences with še-lo (clearly the less common construction, as there were hundreds of examples with the specialized complementizers: 422 with šema and 703 with pen).20 šema and pen complementation represents a phenomenon distinct from Super Neg and is today formal and archaic. (14) ‫ כאילו מתוך יראה‬,‫ לראות עולמן בחייהן‬,‫ ממהרות להתענג‬,‫נשים טסות והולכות‬ . . . ‫שעה קלה לבטלה‬ ‫שלא לאבד‬ našim tasot ve-holxot, memaharot lehitʕaneg, lirʔot women fly and-walk, hurry to.enjoy, to.see ʕolaman be-ħayeyhen, keʔilu mi-tox yirʔa še-lo, as.if out.of fear that-neg leʔabed šaʕa kala le-vatala to.lose hour light to-idleness ‘Women rush by, hurrying to enjoy as much as they can in their lifetime, as if afraid to spend a single hour in vain.’ (“Ħulša,” Ha-šiloaħ, Eliyahu Maidanik, May 1904) 19  See, accessed September 17, 2014. 20  The roots searched for were ‫ד‬.‫ח‬.‫ פ‬p.ħ.d, ‫ג‬.‫א‬.‫ ד‬d.ʔ.g, ‫ש‬.‫ש‬.‫ ח‬ħ.š.š, and ‫א‬.‫ר‬.‫ י‬y.r.ʔ.

Superfluous Negation in Modern Hebrew and Its Origins


Super Neg in the complement of ‘fear’-type verbs is also widespread crosslinguistically (found in Yiddish, Russian, French, Shakespeare English, 17th-century Dutch, and Latin; Weinreich 1958; Schaechter 1986; Timberlake 2004; van der Wouden 1994, among others). While the construction is quite old in Hebrew, external influences may have facilitated its preservation over the ages. GoshenGottstein (2006:141–142) suggests the influence of Arabic on Medieval Hebrew, and the existence of corresponding constructions in Yiddish and Slavic may have played a role more recently. Clausal Complements of ‫ כמעט‬kimʕat ‘Almost’ Super Neg in the complement of kimʕat is attested in small numbers in Medieval Hebrew (Goshen-Gottstein 2006:131; Sagi 2000) as well as in responsa of the 16th–19th centuries (Sagi 2000). Our searches show that the construction existed in the late 19th century (with examples attested from 1872) and reached its peak around 1900–1902. A sharp decrease in use is observed around 1905, followed by a gradual decline. Today, Super Neg kimʕat še-lo sounds odd to native Hebrew speakers. (15) .‫זהירות של שוטר אנגלי‬-‫שלשום כמעט שלא נקטפו חיי אדם מחמת אי‬ šilšom kimʕat še-lo niktefu ħayey almost that-neg were.plucked life.cs ʔadam meħamat ʔi-zehirut šel šoter ʔangli human because.of lack-caution of policeman English ‘Two days ago a British officer’s carelessness almost cost a man’s life.’ (Doʔar ha-yom, November 19, 1931, Jerusalem) Goshen-Gottstein (2006:131) attributes two occurrences in medieval texts to Arabic influence, and Avinery (1964) attributes the occurrence in Rashi’s writing in the 11th–12th c. to French influence. The same construction exists in Yiddish (Schaechter 1986:322; Sadan 1971:121ff.) and in Russian (Wade 2011:113,295; Kagan & Wolf to appear), and was proposed to be another instance of Yiddish influence on Hebrew (perhaps from the 16th century; Sagi 1997, 2000; Farstey 2006). Hebrew prescriptivists shared this view and denounced the use of kimʕat še-lo (Lešonenu la-ʕam, December 14, 1934).21 Before concluding, we note another construction that is obsolete in contemporary Modern Hebrew and can potentially be analyzed as an instance of Super Neg: -‫ ל‬le ‘to/for’ and a special negative form (le-val, li-vli, le-vilti) in the complement of ‫ אסר‬ʔasar ‘prohibit, bond.’ Examples are attested in earlier 21  This may be one reason for the disappearance of the construction in Modern Hebrew, an issue we must leave for future research.


Rubinstein, Sichel and Tsirkin-Sadan

varieties of Hebrew and in the late 19th-century literature in the BYP and HJP. While Super Neg in complements of prohibition predicates is attested crosslinguistically (van der Wouden 1994:109 mentions ‘forbid’), the ambiguity of Hebrew ʔasar as meaning both ‘prohibit’ and ‘bond’ is compatible in principle with an analysis of the negation as non-superfluous in these examples.22 Conclusion Constructions with superfluous negation in Modern Hebrew do not all share the same path of development. While several constructions were denounced as “vulgar Russianisms or Polishisms”23 over the years, some disappeared (kimʕat še-lo) while others lived on to become part of the Modern Hebrew grammar. Language contact may have reinforced existing patterns of Hebrew (‘fear’ verbs), led to reanalysis of others (ʕad še-lo), and introduced altogether new forms into the language (FRs). A better understanding of the semantic contribution of negation in the different constructions may shed further light on these diverse paths of development. References Abels, Klaus. 2005. “ ‘Expletive Negation’ in Russian: A Conspiracy Theory.” Journal of Slavic Linguistics 13.1: 5–74.

22  The scope of our survey is limited to contemporary Hebrew and does not cover every historical case of Super Neg. We enriched the discussion by including obsolete kimʕat še-lo, and we suspect there may be similar cases of short-lived Super Neg constructions. One possible instance of Super Neg in Rabbinic Hebrew which did not survive to be part of Modern Hebrew is counterfactual ‘if’ ‫ אלמלא‬ʔilmale or ‫ אילולא‬ʔilule (M. Taube, p.c.). ʔilmale is used with a negated clause in the following sentence: .‫ אלמלי המסבות והמאורעות של אותו הזמן לא היו מפיחין בו‬,‫] ניצוץ זה היה עומם ונעלם‬. . .[ nicoc ze haya ʕomem ve-neʕelam, ʔilmale ha-mesibot spark this was dim and-disappear if.counterfactual the-circumstances ve-ha-meʔoraʕot šel ʔoto ha-zman lo hayu mefiħin bo and-the-events of same the-time neg were in.him ‘This spark would have dimmed and vanished, had the events of that time not brought it back to life.’ (Mendele Moxer-Sfarim’s Masʕot Binyamin ha-šliši, 2nd ed. (Tel Aviv: Dvir, 1949–1950), p. 12). 23  Shapiro (1909/1938).

Superfluous Negation in Modern Hebrew and Its Origins


Altbauer, Moshe. 1964. “New Negation Construction in Modern Hebrew.” In For Max Weinreich on His Seventieth Birthday: Studies in Jewish Languages, Literature, and Society, ed. Lucy S. Dawidowicz. The Hague: Mouton, 1–5. Avinery, Isaac. 1964. Yad ha-lašon. Tel Aviv: Yizra’el (in Hebrew). Azar, Moshe. 1995. The Syntax of Mishnaic Hebrew. Jerusalem: The Academy of the Hebrew Language (in Hebrew). Ben-David, Aba. 1967. Lešon mikra u-lešon ħaxamim. Tel Aviv: Dvir (in Hebrew). Blanc, Haim. 1956. Review of Haʕivrit šelanu: Dmuta be-ʔor šitot ha-balšanut [Our Hebrew, as Seen by the Methods of Linguistics] by Haiim Rosén. Language 32.4: 794–802. ———. 1965. “Some Yiddish Influences in Israeli Hebrew.” In The Field of Yiddish: Studies in Language, Folklore, and Literature, second collection, ed. U. Weinreich. The Hague: Mouton, 185–201. Braverman, Natan. 1995. Particles and Adverbs in Tannaitic Hebrew (Mishnah and Tosefta): A Syntactic-Semantic Analysis. PhD dissertation, The Hebrew University of Jerusalem (in Hebrew). Brown, Francis, Samuel Rolles Driver, & Charles Augustus Briggs. 1906. A Hebrew and English Lexicon of the Old Testament. Boston: Houghton-Mifflin. [BDB] Citko, B. 2010. “On the Distribution of -Kolwiek ‘Ever’ in Polish Free Relative Clauses.” Journal of Slavic Linguistics 18.2: 221–258. Driver, Godfrey R. 1973. “Affirmation by Exclamatory Negation.” Journal of Ancient Near Eastern Society of Columbia University 5: 107–114. Eilam, Aviad. 2008. “Modern Hebrew: Yiddish Patterns, Hebrew Forms.” Unpublished manuscript. Eilam, Aviad. 2009. “The Crosslinguistic Realization of -Ever: Evidence from Modern Hebrew.” In Proceedings of the 43rd Annual Meeting of the Chicago Linguistic Society (CLS), vol. 2, eds. Malcolm Elliott et al. Chicago: Chicago Linguistic Society, 39–53. Farstey, Hava. 2006. Various Phenomena in Written Modern Hebrew and Their Affinity to Yiddish. PhD dissertation, The Hebrew University of Jerusalem (in Hebrew). Francez, Itamar. 2015. “Modern Hebrew lama-še Interrogatives and Their JudeoSpanish Origins.” Journal of Jewish Languages 3.1: 104–115. Garbell (Chanoch), Irene. 1930. Fremdsprachliche Einflüsse im modernen Hebräisch. PhD dissertation, University of Berlin (in German). Global Jewish Database: The Bar Ilan Responsa Project on CD-ROM, version 21. Gesenius, Wilhelm. 1910. Gesenius’ Hebrew Grammar as Edited and Enlarged by E. Kautzsch, 2nd English edition, rev. in accordance with the 28th German ed. (1909) by A.E. Cowley. Oxford: Clarendon Press. Goshen-Gottstein, Moshe. 2006. Syntax and Vocabulary of Mediaeval Hebrew as Influenced by Arabic, revised by Shraga Assif and Uri Melammed. Jerusalem: Ben Zvi Institute (in Hebrew).


Rubinstein, Sichel and Tsirkin-Sadan

Gryczan, Barbara. 2013. “Verbal System: Medieval Hebrew Poetry.” In Encyclopedia of Hebrew Language and Linguistics, eds. Geoffrey Khan et al. Brill Online. http:// Gzella, Holger. 2013. “Presentatives.” In Encyclopedia of Hebrew Language and Linguistics, eds. Geoffrey Khan et al. Brill Online. encyclopedia-of-hebrew-language-and-linguistics/presentatives-EHLL_COM_ 00000022. Haspelmath, Martin & Ekkehard König. 1998. “Concessive Conditionals in the Languages of Europe.” In Adverbial Constructions in the Languages of Europe, ed. Johan van der Auwera. Berlin: Mouton de Gruyter, 563–640. Historical Jewish Press: (accessed August 2014). Joüon, Paul & T. Muraoka. 2006. Grammar of Biblical Hebrew. Rome: Editrice Pontificio Instituto Biblico. Kagan, Olga & Lavi Wolf. To appear. “Gradability versus Counterfactuality: Almost in English and Russian.” In Proceedings of the 30th Annual Conference of the Israel Association for Theoretical Linguistics, 2014. MIT Working Papers in Linguistics, ed. Nurit Melnik. Cambridge, MA. Krifka, Manfred. 2010. “How to Interpret ‘Expletive’ Negation under Bevor in German.” In Studia Grammatica 72 (Language and Logos: Festschrift for Peter Staudacher on his 70th Birthday), eds. Thomas Hanneforth & Gilbert Fanselow. Berlin: Akademie Verlag, 214–236.‫‏‬ Larson, Richard. 1987. “Missing Prepositions and the Analysis of English Free Relative Clauses.” Linguistic Inquiry 18: 239–266. Morgenstern, Matthew. 2013. “Temporal Clause: Rabbinic Hebrew.” In Encyclopedia of Hebrew Language and Linguistics, eds. Geoffrey Khan et al. Brill Online. http:// Moshavi, Adina. 2013. “Interrogative: Biblical Hebrew.” In Encyclopedia of Hebrew Language and Linguistics, eds. Geoffrey Khan et al. Brill Online. http://reference interrogative-biblical-hebrew-EHLL_COM_00000570. ———. 2014. “What Can I Say? Implications and Communicative Functions of Rhetorical ‘WH’ Questions in Classical Biblical Hebrew Prose.” Vetus Testamentum 64: 93–108. Rosén, Haim. 1956. Our Hebrew. Tel Aviv: ʕam ʕoved (in Hebrew). ———. 1977. Contemporary Hebrew. The Hague: Mouton. Sadan, Dov. 1971. “Šir, šier, šiur.” In Idiomatic Expressions of the Yiddish Language, ed. Dov Sadan. Buenos Aires: Asociación Pro-Cultura Judía, vol. 1: 121 (in Yiddish).

Superfluous Negation in Modern Hebrew and Its Origins


Sagi, Hanna. 1997. Selected Morpho-Syntactic Changes in Literary Translations of Sholom-Aleichem from Yiddish to Hebrew: A Study of the Influence of Yiddish on the Structure of Modern Hebrew. PhD dissertation, Bar-Ilan University (in Hebrew). ———. 1999. “ ‘Ma še-lo yikre, nišaer yedidim’: ‘Lo’ ke-šolel ʕodef.” Ħelkat Lašon 28: 7–21 (in Hebrew). ———. 2000. “ ‘Kimʕat še-lo nafalti’: ‘Lo’ ke-šolel ʕodef ba-ceruf ‘kimʕat še-lo’.” Ħelkat Lašon 29–32 (Maya Fruchtman Book): 86–96 (in Hebrew). Schaechter, Mordkhe. 1986. Yiddish II: A Textbook for Intermediate Courses. Philadelphia: Institute for the Study of Human Issues (in Yiddish). Shapiro, Aharon Y. (1909/1938). More nevoxey ha-lašon. Warsaw: Starovolsky (in Hebrew). Stadel, Christian. 2013. “Interrogative: Rabbinic Hebrew.” In Encyclopedia of Hebrew Language and Linguistics, eds. Geoffrey Khan et al. Brill Online. interrogative-rabbinic-hebrew-EHLL_COM_00000550. Timberlake, Alan. 2004. A Reference Grammar of Russian. Cambridge: Cambridge University Press. The Ben-Yehuda Project: (accessed February 2014). Tzivoni, Lea. 1993. Negation in Israeli Hebrew. PhD dissertation, The Hebrew University of Jerusalem (in Hebrew). Wade, Terence. 2011. A Comprehensive Russian Grammar (third edition), revised and updated by David Gillespie. Malden: Wiley-Blackwell. Weinreich, Uriel. 1958. “Yiddish and Colonial German in Eastern Europe.” In American Contributions to the Fourth International Congress of Slavicists, Moscow, September, 1958, eds. International Congress of Slavists, Moscow. The Hague: Mouton, 369–419. Van der Wouden, Ton. 1994. Negative Contexts. Doctoral dissertation, Gröningen University. Zeijlstra, Hedde. 2011. “On the Syntactically Complex Status of Negative Indefinites.” Journal of Comparative Germanic Linguistics 14: 111–138.

From Negative Polarity to Negative Concord— Slavic Footprints in the Diachronic Change of Hebrew meʔuma, klum, and šum davar Einat-Haya Keren

The Hebrew University of Jerusalem, Israel [email protected]

Abstract The article is concerned with a special kind of negative items that changed their distribution when Hebrew became a spoken language again, as an impact of the native languages of its first users. The main claim is that the items meʔuma, klum, and šum davar, which function as Negative Polarity Items (NPIs) in Biblical and Rabbinic texts, and are therefore translated as ‘anything,’ have changed their function into Negative Concord Items (NCIs) in the course of Hebrew’s so-called revival, and are now better translated as ‘nothing.’ Though both classes are often used with negation, there are contexts in which only NCIs or NPIs are allowed. Showing the difference in distribution between Modern and Classical Hebrew, the article compares meʔuma, klum, and šum davar to parallel NCIs in Polish, Russian, and Yiddish. It concludes that the shift in distribution and meaning of these items is mostly due to influence of Slavic languages.

Keywords Negative Concord Items (NCIs) – Negative Polarity Items (NPIs) – Modern Hebrew – Slavic – Yiddish – Diachronic Change

Introduction NCIs Negative Concord Items (NCIs)1 are elements with a certain puzzling nature. On the one hand, they must be used in structures containing sentential 1  Such items are usually referred to as N-words, after Laka (1990:106). Negative quantifiers not necessitating further negation are viewed as N-words by some authors, but not by others

© koninklijke brill nv, leiden, ���6 | doi ��.��63/9789004310896_014

From Negative Polarity to Negative Concord


­ egation; but on the other hand, the reading resulting from the combinan tion of these two negations yields just one logical negation. One can see that Modern Hebrew meʔuma, klum, and šum-davar qualify as NCIs by the following example: (1)

lo raʔ’iti meʔuma/ klum/ šum davar2 neg see.past.1s nothing ‘I didn’t see anything.’/ ‘I saw nothing.’

In (1), although the quantifiers meʔuma, klum, and šum davar are combined with the sentential negation lo, the interpretation of the sentence is ‘I saw nothing,’ with single negation. The sentence is not interpreted with double negation: ‘I didn’t see nothing’ (= ‘I saw something’). Moreover, this negation is necessarily present, and without it the sentence is ungrammatical. Strict versus Non-strict Use of NCIs There are different ways in which languages use NCIs. In some languages, the verb has to be negated, regardless of the function of the NCI (subject, object, or adjunct) or its position (before the verb or after the verb). Modern Hebrew is such a language: (2) ki b-a-tmuna lo raʔiti meʔuma. because in-the-picture neg saw.1s nothing ‘Because I didn’t see anything in the picture.’ (“Tapuz Anashim,” orumid=485&messageid=172190518#172190518, accessed December 4, 2014) (3) meʔuma lo yešane ʔet daʕatam nothing neg acc opinion.gen ʕal ha-ħisun about the- vaccination ‘. . . Nothing can change their opinion about the vaccination.’ (y-net, health+(Briut),,7340,L-4457560,00. html, accessed December 4, 2014)

(Willis, Lucas, & Breitbarth, 2013:32). For simplicity reasons, this article will only use the term “Negative Concord Items.” 2  Unless mentioned otherwise, all examples of Modern Hebrew were provided by myself as a native speaker.



Other languages are different. In many Romance languages, an NCI appearing post-verbally necessitates negation, while an NCI appearing pre-verbally either does not require it, or is, in fact, incompatible with it (Zeijlstra 2004:3). This was termed “Non-Strict Negative Concord” by Giannakidou (1997). A key feature of those languages is that an NCI appearing pre-verbally will function as the negator for an NCI appearing post-verbally.3 In Modern Hebrew, the preverbal NCI cannot function as negation for the post-verbal NCI, as can be seen in (4): The quantifier ʔaf ʔeħad, which is also an NCI, cannot give license to the NCIs meʔuma, klum, and šum davar; negation is obligatory here just as it is in all previous examples. This again shows that Modern Hebrew is a Strict Negative Concord language, as opposed to a Non-Strict one. (4)

ʔaf ʔeħad *(lo) ʔamar nobody *(neg) said.3ms ‘Nobody said anything.’

meʔuma/ klum/ šum davar nothing

meʔuma, klum, and šum-davar Viewed Diachronically

We have seen that meʔuma, klum, and šum-davar are used as NCIs in Modern Hebrew. By contrast, their distribution in Classical Hebrew was different. It seems to have matched that of a different category of elements known as Negative Polarity Items (NPIs). Negative Polarity Items Languages across the world are known to have classes of items that are sensitive to polarity. This means that they need to be licensed by certain semantic, and syntactic, environments. Otherwise, they will be ill-formed. There are Positive and Negative Polarity Items; this article will only deal with Negative ones (NPIs). NPIs are excluded from positive epistemic sentences (Giannakidou 2011:1661). The most common environment for them is within the scope of negation. However, NPIs can be grammatical in many contexts that do not include negation, for example, questions and conditional sentences (see (6), (9), and (11) below). There are differences in the semantics of NPIs and NCIs as well. Both of them deal with classes of objects. NCIs, however, negate the existence of even a single member of a class, while NPIs typically describe the existence of at least 3  This pattern is called “Negative Spread.” According to Giannakidou (2000:460), the term was coined by Den Besten (1986).


From Negative Polarity to Negative Concord

one member of that class. Thus, while the NCI šum davar negates the existence of any entity, the NPI davar va-ħaci davar describes the existence of very small things: (5)

ʔein šam neg.existential there ‘There’s nothing there.’

šum davar nothing

(6) ʔim timca šam davar va-ħaci davar, tagid if find.fut.2sm there thing and-half thing tell ‘If you find anything there, tell me.’

li dat.1s

To demonstrate the differences in distribution, we will contrast the NCI ʔaf paʕam ‘never’ with the NPI ʔei paʕam ‘ever.’ Both NCIs and NPIs are grammatical with negation: (7) lo hayiti šam ʔaf paʕam neg be.past.1s there neg time ‘I’ve never been there.’ (8) lo hayiti šam ʔei paʕam neg be.past.1s there NPI time ‘I haven’t ever been there.’ However, there are many contexts in which NPIs are grammatical and NCIs, by contrast, are ungrammatical. The classical ones are questions (see (9)–(10)), and antecedents of conditionals (see (11)–(12)).4 (9) hayita šam ʔei paʕam? be.past. 2SM there NPI time ‘Have you ever been there?’

4  The Ladusaw-Fauconnier hypothesis (Fauconnnier 1975; Ladusaw 1979) states that all the environments that can license NPIs share a property called “Downward Entailment”—­ reversal of the entailment direction of logical relations in their scope (Willis, Lucas, & Breitbarth 2013:28–29). Some authors claim that this property is not enough to capture all the environments capable of licensing NPIs (ibid.: 29). For reasons of space and simplicity, I will not discuss this issue here, and will use only questions and antecedents of conditionals as NPI licensing contexts.



(10) *hayita šam ʔaf paʕam? be.past.2SM there NEG time ‘Have you never been there?’ (11) ʔim ʔei paʕam tagiaʕ le- šam if NPI time arrive.fut.2sM to-there li mazkeret me souvenir.f ‘If you ever get there, bring me a souvenir.’ (12) *ʔim ʔaf paʕam tagiaʕ le-šam, if NEG time arrive.fut.2sm to-there ‘If you never get there, it will be a shame.’

tavi bring.fut.2sM

ze yihiye ħaval it be.fut.3sm shame

There are also contexts that allow NCIs but disallow NPIs:5 (13) Fragmented Answers: kama peʕamim hayita be-mongolia? how.many be.past.2sm in-mongolia ‘How many times have you been to Mongolia?’ a. ʔaf paʕam neg time ‘never’ b. *ʔei paʕam npi time 5  Another context that seems to allow N-Words in Modern Hebrew is the emphatic klum and šum-davar: ʔetmol halaxti mi merkaz ha-ʕir ʕad Bakʕa biglal yesterday walked.1.s from-center.m.gen the-city.f to Bakʕa because švitat ha-ʔotobusim. . . strike.f.gen ‘Yesterday I walked from the city center to Bak’a, because of the buses’ strike . . .’ ze klum! ʔani halaxti me-har ha-cofim Deictic nothing I walked.1.s from-mountain.m.gen ve-ʕad Ramot! and-to Ramot ‘That’s nothing! I walked from Mt. Scopus to Ramot!’ Krifka (1994:204–5) claims that the emphatic assertion, which marks that the stressed proposition is less likely than its alternatives, can license Strong NPIs. Those NPIs are only licensed by negation rather than downward entailing contexts, but are nevertheless ungrammatical in fragmented answers. Perhaps this feature is common to NCIs and Strong NPIs. This would

From Negative Polarity to Negative Concord


(14) Modification by almost: a. ʔani lo kotevet po kimʕat ʔaf paʕam I neg write.pres.sf here almost neg time ‘I hardly ever write here . . .’ (Taken from the forum “Kippah,” show/9974329, accessed July 12, 2015) b. *ʔani lo kotevet po kimʕat ʔei paʕam I neg write.pres.sf here almost npi time Table 1 summarizes the information presented in this section: while questions and conditionals allow NPIs but disallow NCIs, the pattern is reversed in fragmented answers and after almost. table 1

The available environments for Negative Polarity Items (NPIs) and Negative Concord Items (NCIs)



N-words (NCIs)

Negation Questions & conditionals Fragmented Answers After “almost”

yes yes no no

yes no yes yes

Data from Classical Hebrew When looking at the distribution of meʔuma, klum, and šum-davar in Classical Hebrew texts, one can see a difference: Unlike their distribution in Modern Hebrew, in classical texts their distribution seemed to have overlapped with that of NPIs. The following examples establish this claim. meʔuma məʔūmā in Biblical Hebrew appears not only with negation (15), but also in questions (16) and conditionals ( (17)–(18).6 explain why meʔuma, klum, and šum-davar are licensed in emphatic contexts without negation. I thank Avigail Tsirkin-Sadan for her help with this matter. 6  In 2 Samuel 13:2 məʾūmā appears in the complement of verb was difficult (wa-yippālε), which is a Downward Entailment environment (see footnote 3). There are also other places where məʾūmā or məʾūm appear in questions and antecedents of conditionals. For reasons of space, these were not included here.



(15) ‫אּומה‬ ָ ‫אמר יַ ֲעקֹב לֹא ִת ֶתן ִלי ְמ‬ ֶ ֹ ‫וַ י‬ va-yyomer yaʿakoḇ lo titten li mǝʾūmā and-say.3sm Jacob neg give.FUT.2sm dat.1s nothing/anything ‘And Jacob said, “Pay me nothing!” ’ (Genesis 30:31)7 ָ ‫אּוכל ַדבר ְמ‬ ַ ‫אתי ֵא ֶליָך ַע ָתה ֲהיָכֹול‬ ִ ‫ִהנֵ ה ָב‬ (16) ‫אּומה‬ hinnē bātī ʾēlεyḵā ʿattā hă-yāḵōl ʾūḵal here come.PAST.1s now Q-can.INF can.FUT.1s dabbēr məʾūmā talk.inf anything ‘And now that I have come to you, have I the power to speak freely?’ (Numbers 22:38)

(17) ‫אּומה לֹא ָתבֹא ֶאל ֵביתֹו ַל ֲעבֹט ֲעבֹטֹו‬ ָ ‫ִכי ַת ֶשה ְב ֵר ֲעָך ַמ ַשאת ְמ‬ kī ṯaššε ḇə-rēʿăḵā maššat məʾūmā lō if lend.FUT.2sm in-friend.gen.2sm thing anything neg ṯāḇō ʾεl- bēyṯō laʿăḇoṭ ʿaḇoṭō come.FUT.2sm to house.gen.3sm take.INF pledge.gen.3sm ‘When you make a loan of any sort to your countryman, you must not enter his house to seize his pledge.’ (Deuteronomy 24:10) ָ ‫ֹלהים וְ כֹה י ִֹסיף ִכי ִאם ִל ְפנֵ י בֹוא ַה ֶש ֶמׁש ֶא ְט ַעם ֶל ֶחם אֹו ָכל ְמ‬ ִ ‫כֹה יַ ֲע ֶשה ִלי ֱא‬ (18) ‫אּומה‬ kō yaʿăsε llī ʾɛ̆lōhīm wə-ḵō yōsip̄ kī ʾimso do.FUT.3sM dat.1s god and-so add.FUT.3sM if lip̄ nēy ḇō ha-ššεmεš ʾεṭʿam lεḥεm ʾō before come.inf the-sun taste.fut.1s bread or ḵol məʾūmā any anything ‘May God do thus to me and more if I eat bread or anything else before sundown.’ (2 Samuel 3:35)

7  All English translations of Biblical Hebrew are taken from the Jewish Publication Society. 1985. Tanakh—A New Translation of The Holy Scriptures according to the Traditional Hebrew Text. The Jewish Publication Society: Philadelphia, PA.

From Negative Polarity to Negative Concord


Klum Klum, a word which first appeared in early Rabbinic Hebrew (in the Mishnah), is mostly used with negation in Rabbinic texts (19). However, it too can be found in questions (20) and in conditionals (21): ֶ ‫ָּכל זְ ַמן ֶש ַה ֶּט ֶבל ְמ‬ (19) ‫רּוּבה לֹא ִהיפסיד ְּכלּום‬ kol zman šε-ha-ṭṭεḇεl mərūbbe lō hīp̄ sīḏ all time.M that-the-untithed great.m neg lost.PAST.3sm kəlūm nothing/anything ‘[W]hen the greater part is untithed produce naught is lost.’ (Mishnah, Zeraʿim, Demai 7:7, MS Kaufmann A 50)

(20) ,‫רֹואים ַא ֶתם ְּכלוּם ְּב ֶא ָחד ִמן ֶה ָה ִרים ַה ָּללוּ‬ ִ rōʾīm ʾaṯεm kəlūm bə-͗εḥāḏ min you.pLm anything in-one.M of hε-hārīm hallālū? the-mountains those? ‘Do ye see anything upon one of those mountains?’ (Pirke De-Rabbi Eliezer, chapter 31; taken from the 2005 edition, Zichron Aharon: Jerusalem) (21) ‫אם יאמר לך יתרו כלום מן השבועה‬ ʾim yōmar ləḵā Yiṯrō kəlūm min ha-ššḇūʿā if say.fut.3sm dat.2s Jethro anything from the-oath.F ‘Should Jethro at all remind you of your oath . . .’ (Exodus Rabbah 4:4; Hebrew: Friedlander 1981; English: Lehrman 1961) šum-davar Both šum and davar are Biblical words.8 The combination šum davar is considered either Rabbinic or Medieval.9 Example (22) is taken from the medieval text Mishneh Torah by Maimonides, which was relatively well known to the

8  šūm and davar appear in the Old Testament. šūm appears mostly in its earlier phonetic variant šēm (‫) ֵשם‬, except for the Aramaic texts in Ezra and Daniel, in which we already find šum in the orthography šum (‫) ֻשם‬. I thank Chanan Ariel for bringing this matter to my attention. 9  The combination first appears in Mishnah, Ketubot. It has several appearances in Mishnaic and Rabbinic Hebrew texts. In Avraham Even-Shoshan’s dictionary, however, it is categorized as Medieval. I am grateful to Elitzur Bar-Asher Siegal for his help with this matter.



first users of Modern Hebrew. One can see that in this quote šum davar is used in the antecedent of a conditional:10 (22)  ‫ ֲ ה ֵרי זֹו‬. . . ‫עֹולם‬ ָ ‫ אֹו ְלׁשּום ָּדבר ָּב‬. . . ‫אכה אֹו‬ ָ ‫אֹותּה ַה ְּמ ָל‬ ָ ‫ ְ ל‬. . . ‫ׁשֹואל ֶאת ַה ְּב ַע ִלים‬ ֵ ‫וְ ֶאחד ַה‬

‫ּופטּור‬ ָ ‫ְש ִא ָילה ַב ְב ָע ִלים‬

w-ʾεḥad ha-šōʾēl ʾεt ha-bəʿālim . . . lə-ʾōtah and-one.M acc the-owners . . . for-acc.sf ha-mməlāḵā ʾō. . .ʾō lə- šum davar b-ā-ʿōlām . . . the-work.f or or for-anything in-the-world . . . hărei zō šəʾīla ḇ-a-bəʿalīm u-p̄ āṭūr emphatic it.f borrowing.f in-the-owners and-quit.pass.3sm  ‘whether the commodatary borrowed the services of the owner or . . . whether he borrowed the services for the same work, or . . . or for anything in the world . . . it is a case of borrowing with the owner and the commodatary is quit.’ (Borrowing and Depositing, 2 1(6), The Book of Civil Laws, Mishneh-Torah [The Code of Maimonides]; Hebrew: Kook 1987; English: Rabinowitz 1949) To summarize, it seems that meʔuma, klum, and šum davar were NPIs in Classical texts.11 In Modern Hebrew, however, speakers will most likely judge meʔuma, klum, and šum davar as ungrammatical in contexts of questions or antecedents of conditionals not containing negation:12

10  Earlier examples for usage of šum davar as an NPI can be found in Baniel (1938 [‫]תרצ"ח‬: 317–319). I am grateful to Chanan Ariel for bringing this matter to my attention. 11  The distribution of meʔuma, klum, and šum davar is also compatible with simple nouns, of course, rather than NCIs or NPIs. However, no examples of these items in non-NPI contexts have been documented, to the best of my knowledge (in case of meʔuma, all of its occurrences in the Old Testament are in NPI contexts [Keren 2012]). Also, in the relevant texts’ translations, quantifiers are usually chosen for meʔuma, klum, and šum davar. 12  Perhapes native speakers of Hebrew who have been taught about the initial distribution of meʔuma, klum, or šum davar will accept them in such contexts. One of my informants in Keren (2012), for example, accepted the following sentence, claiming she had been taught in school that such sentences are acceptable: ʔata carix latet li klum you.MS need.PRES.SM give.INF dat.1s nothing ‘You don’t have to give me anything.’ However, native speakers who have not been taught anything in particular about meʔuma, klum, or šum davar are very unlikely to provide such judgments. None of my other informants, altogether more than twenty native speakers, have given such judgments concerning this sentence or any other sentence containing meʔuma, klum, šum davar, or the NCI for person af ʔeħad, without negation.

From Negative Polarity to Negative Concord


(23) * raʔita šam meʔuma / klum/ šum davar? see.past.2mS there nothing ‘Did you see anything there?’ (24) *’im tirʔe šama if see.fut.2mS there ‘If you see anything there . . .’

meʔuma/ klum/ šum davar . . . nothing

An Explanation of the Diachronic Change—Modern Hebrew’s Substrate Languages

What is the source of this shift in the distribution of meʔuma, klum, and šum davar? When looking at data from Polish, Russian, and Yiddish, a natural explanation presents itself.13 In all these languages, one can find an NCI equivalent to ‘nothing’ in its semantics, and a distribution that appears similar to that of Modern Hebrew’s meʔuma, klum, and šum davar: (25) Polish: Jan *(nie) chciał niczego kupować John not wanted nothing-gen buy-inf “John didn’t want to buy anything.” (Przepiórkowski & Kupść 1998:250) (26) Russian: Ivan ne znaet ničego. Ivan not knows n.what ‘Ivan does not know anything.’ (Fitzgibbons 2008:51) 13  One could ask if this isn’t a diachronic change. Similar changes from NPIs to NCIs have been documented in many Romance languages, for example, French (Roberts & Roussou 2003). However, one could ask if a language without native speakers could be expected to experience such natural processes. Also, these processes usually affect one lexical item at a time. In contrast, during Modern Hebrew’s period of revival, three lexical items corresponding to nothing were affected at the same time in which a lexical item corresponding to no one was formed. The Jespersen Cycle (Jespersen 1917)—a well-known phenomenon in which a pre-verbal marker of sentential negation gradually weakens, while a postverbal adverb, which was optional at first, overtakes its role—is not relevant here, since this article discusses meʔuma, klum, and šum davar rather than sentential negation. A detailed discussion of diachronic change is beyond the scope of the present article and is addressed more thoroughly in Sichel & Keren (forthcoming).



(27) Yiddish: Hot men gegrobn un gegrobn, un hot gornisht has one dug and dug and has nothing nit gefunen not found ‘One continued digging and one found nothing.’ (Olsvanger 1947: 110, as quoted in Van der Auwera & Gybels 2014:211) The use of nic (niczego), ničego, and gornisht with negation results in one logical negation. Their existence could be the reason that meʔuma, klum, and šum davar14 are used as NCIs in Modern Hebrew.15 But are niczego, ničego, and gornisht NCIs or NPIs? Przepiórkowski & Kupść (1998:2) show that Polish niczego is not licensed in NPI contexts and is therefore an NCI. Similarly, Russian ničego is an NCI rather than an NPI (Pereltsvaig 2006:155). Van der Auwera & Gybels (2014) do not address the question directly. However, they talk about Haspelmath’s “indirect negation”: contexts of implicitly negative expressions, such as the verbs deny and refuse, and superordinate negation. The authors (ibid.: 219–221) mention that Yiddish negative indefinites do not normally appear in such contexts.16 Haspelmath (1997:33) claims that indirect negation is one of the contexts for NPIs (1997:36). Therefore we may conclude that gornisht is not an NPI either. In short, niczego, ničego, and gornisht all seem to be NCIs. As meʔuma, klum, and šum davar usually appear in negated sentences, 14  One can ask if Aramaic lexical items could also have influenced the first Hebrew speakers when they mistook meʔuma, klum, and šum davar for NCIs. According to Elitzur BarAsher Siegal (PC, 2014), there don’t seem to be any examples of the parallel Aramaic item in downward entailing contexts. If this item is found in fragmented answers, this will prove that it had been an NCI. Related to that notion, one could also mention lo ḵlum as another Rabbinic NCI. However, regardless of the syntactic category of lo ḵlum and the Aramaic items and their impact on meʔuma, klum, and šum davar, we still have to account for the latter items’ changing distribution. 15  Altbauer (1964) and Ben-Nun (1965) have already proposed that the existence of Negative Concord in Modern Hebrew is due to the influence of European languages, while Rosén (1977:228–229) makes a statement particularly about Russian and Yiddish in this matter. However, they did not test the locality constraints on Negative Concord in Modern Hebrew or in Russian, Polish, or Yiddish. That work was done in Keren (2012), the dissertation that this article is based on. 16  Van der Auwera & Gybels (2014:219) mention that gornisht can appear in the context of without, which Haspelmath treats as indirect negation. However, Giannakidou claims that without can license NCIs in strict Negative Concord languages. This was shown for Greek (Giannakidou 2000:10) and Haitian Creole (Depréz 1999:418), among others.


From Negative Polarity to Negative Concord

a person whose native language uses NCIs could easily have overlooked their appearances in NPI contexts, and might mistakenly believe that they are NCIs rather than NPIs.17 Examples 25–27 show us that Polish, Russian, and Yiddish do exhibit Negative Concord. If these languages are Strict Negative Concord languages, like Modern Hebrew, we will be able to explain the distribution of meʔuma, klum, and šum davar in Modern Hebrew as a parallel of that of Polish niczego, Russian ničego, and Yiddish gornisht. Indeed, Russian and Polish are Strict Negative Concord languages (Zeijlstra 2004:122, among others).18 The Yiddish NCI gornisht, however, can be used without negation in regular sentences,19 as can be seen in (28). Therefore, Yiddish uses this NCI in a Non-strict way. (28) Hot Yankev gezogt ‘Zolst mir gornisht gebn.’ has Yankev said should me nothing give ‘Yankev said: “You should give me nothing.” ’ (Yehoyesh, Genesis, as quoted in Van der Auwera & Gybels 2014:211) table 2 The distribution of gornisht compared to NPIs, NCIs, and Negative Items in DN languages Environment

Negative Polarity Items

Negative Concord Items

Negative Items in Double Negation Languages


Negation Questions & conditionals Fragmented Answers After “almost” Positive Declarative

yes yes

yes no

yes yes

yes yes





no no

yes no

yes yes

yes yes

17  I thank Elitzur Bar-Asher Siegal for his help formulizing this point. 18  This was confirmed by native speakers of those languages whom I have consulted. 19  Not fragmented answers. This has already been noted by Birnbaum 1979:302. However, in Uriel Weinreich’s College Yiddish (1976, Hebrew edition), I found no reference to this matter.



Table 2 summarizes the distribution of gornisht, showing it is different from NCIs (resembling that of N-Words such as English nothing, which can appear without negation). From this, it doesn’t seem likely that meʔuma, klum, and šum davar were influenced by gornisht. This is surprising, as Yiddish syntax has been suggested as a major influence over the syntax of Modern Hebrew (for example, Wexler 1990). However, we can find expressions in Yiddish with a syntactic distribution that does parallel that of meʔuma, klum, and šum davar. According to Van der Auwera & Gybels (2014:205), the expression keiner is an NCI for person-­negation that cannot appear without negation. Yiddish also has NCIs for time-negation, keyn-mol, and place-negation, in ergets. They too necessitate negation, according to the intuitions of Moshe Taube, a native speaker of Yiddish (personal communication). An anonymous reviewer has suggested the following example in order to show the contrast between gornisht and negative determiner phrases of the type keyn-N: (29) un du host mir gornisht ibergelozt afile and you have me nothing left even keyn vort host du mir nisht ibergelozt NEG.DET word have you me not left ‘And you have left me nothing—not even a single word.’ (Sholem Ash, Meri, 1917:152) This suggests that the combination of keyn with nouns also necessitates negation. So it seems that gornisht is the only Yiddish NCI that can appear without negation. On top of that, we must remember that gornisht can appear with sentential negation and form one semantic negation with it (unlike English nothing, which in combination with negation yields a sentence with positive interpretation). So, the distribution of gornisht does not contrast with that of meʔuma, klum, and šum davar in Modern Hebrew, even if it does not parallel it. Conclusion The distribution of meʔuma, klum, and šum davar is very similar to that of Polish niczego and Russian ničego. Therefore, one can posit an explanation according to which native speakers of Polish and Russian misinterpreted meʔuma, klum, and šum davar and viewed them as NCIs, when they were in fact NPIs, and that this is the reason why their distribution in Modern Hebrew

From Negative Polarity to Negative Concord


is different from their distribution in Classical Hebrew. As for Yiddish, the ­distribution of ­gornisht is different from that of meʔuma, klum, and šum davar. However, gornisht can be used with sentential negation just like meʔuma, klum, and šum davar, and Yiddish does have other NCIs with a syntactic distribution seemingly parallel to that of meʔuma, klum, and šum davar. Therefore, altogether, there is not a major contrast between Yiddish syntax and the syntax of meʔuma, klum, and šum davar. Keeping in mind that the other NCIs in Yiddish are used in a strict way, this too could have contributed to the change in distribution of meʔuma, klum, and šum davar toward NCI use in Modern Hebrew. References Altbauer, Moshe. 1964. “New Negation Constructions in Modern Hebrew.” In For Max Weinreich on His Seventieth Birthday: Studies in Jewish Language, Literature, and Society, ed. Lucy S. Dawidowicz. London: Mouton, 1–5. Baniel (Berliner), Yaakov. 1938. “Heʕarot vehacaʕot.” Leshonenu Laʕam 9: 317–321 (in Hebrew). Bin-Nun, Yehiel. 1965. “The Negation Pronoun.” Leshonenu La-ʕam 16/4: 251–268 (in Hebrew). Birnbaum, Salomo A. B. 1979. Yiddish, a Survey and a Grammar. Toronto: University of Toronto Press. Danby, Herbert. 1950. The Mishnah. Oxford University Press: London. Déprez, Viviane. 1999. “The Roots of Negative Concord in French and French Based Creoles.” In Language Creation and Language Change, ed. M. DeGraff. Boston, MA: MIT Press, 375–427. Fitzgibbons, Natalia V. 2008. “Freestanding Negative Concord Items in Russian.” Nanzan Linguistics 3.2: 51–63, special issue. Friedlander, Gerald. 1981. Midrash pirḳê de Rabbi Eliezer (The Chapters of Rabbi Eliezer the Great). Sepher-Hermon Press: New York. Giannakidou, Anastasia. 1997. The Landscape of Polarity Items, PhD diss., University of Groningen. ———. 2000. “Negative . . . Concord?” Natural Language & Linguistic Theory 18.3: 457–523. ———. 2011. “Negative and Positive Polarity Items.” In Semantics: An International Handbook of Natural Language Meaning, eds. Claudia Maienborn, Klaus von Heusinger, & Paul Portner. De Gruyter: Berlin, 1660–1712. Jewish Publication Society. 1985. Tanakh: A New Translation of the Holy Scriptures according to the Traditional Hebrew Text. Philadelphia, PA: Jewish Publication Society.



Keren, Einat. 2012. Negative Concord in Modern Hebrew and the Attempt to Track Down Its Origins. MA Thesis, Hebrew University of Jerusalem. Krifka, Manfred. 1994. “The Semantics and Pragmatics of Weak and Strong Polarity Items in Assertions.” In SALT IV, eds. Mandy Harvey & Lynn Santelmann. Ithaca, NY: Cornell University Press, 195–219. Laka, Itziar M. 1990. Negation in Syntax: On the Nature of Functional Categories and Projections. PhD diss., MIT. Lehrman, S. M. 1961. “Exodus.” In Midrash Rabbah, eds. H Freedman & Maurice Simon. The Soncino Press: London. Pereltsvaig, Asya. 2006. “Negative Polarity Items in Russian and the ‘Bagel Problem.’ ” In Negation in Slavic, eds. Sue Brown & Adam Przepiórkowski. Bloomington, IN: Slavica Publishers, 153–178. Przepiórkowski, Adam & Anna Kupść. 1998. “Verbal Negation and Complex Predicate Formation in Polish.” In Procedings of the 1997 Texas Linguistics Forum Symposium on the Syntax and Semantics of Predication, Texas Linguistic Forum, vol. 38, eds. Ralph C. Blight & Michelle J. Moosally. Austin, TX, 247–262, The Linguistics http://bach.ipipan (accessed September 3, 2014). Rosén, Haiim B. 1977. “Contemporary Hebrew.” In Trends in Linguistics, State-of-the Art Reports (11), ed. Werner Winter. The Hague: Mouton, 225–229. Van der Auwera, Johan & Paul Gybels. 2010. “On Negation, Indefinites, and Negative Indefinites in Yiddish.” In Yiddish Language Structures, eds. Marion Aptroot & Bjorn Hansen. Berlin: Walter de Gruytner, 185. Weinreich, Uriel. 1977. College Yiddish: An Introduction to the Yiddish Language and to Jewish Life and Culture. Jerusalem: The Magnes Press, The Hebrew University. Wexler, Paul. 1990. The Schizoid Nature of Hebrew: A Slavic Language in Search of a Semitic Past. Wiesbaden: Otto Harrassowitz. Willis, David, Christopher Lucas, & Anne Breitbarth. 2013. “Comparing Diachronies of Negation.” In The Development of Negation in the Languages of Europe. Volume 1: Case Studies, eds. David Willis, Christopher Lucas, & Anne Breitbarth. Oxford: Oxford University Press, 1–50.

Lexical Values

The Sudden Disappearance of Nitpael and the Rise of Hitpael in Modern Hebrew, and the Role of Yiddish in the Process Shira Wigderson

The Hebrew University of Jerusalem, Israel [email protected]

Abstract In this article I will present some data about the uses and the distribution of Hitpael and Nitpael during the revival years of Modern Hebrew. While examining this new data, I will try to answer two questions: a) What caused the authors from 1870 to 1920 to prefer the Nitpael form over the Hitpael form; and b) What happened in later years when Hitpael took Nitpael’s place, and became much more frequently used in written and spoken Modern Hebrew? I suggest that Yiddish had influenced speakers to use the Nitpael form in earlier days, but that since its influence on speakers had diminished, and because the meanings of Nitpael are a sub-group of Hitpael’s meanings, Hitpael became the more frequent form of use.

Keywords Hitpael – Nitpael – Modern Hebrew – Yiddish

Introduction Hitpael and Nitpael are two binyanim in Modern Hebrew that originated in Biblical Hebrew and Mishnaic Hebrew, respectively. When Hebrew came to use anew in the 19th century, both forms were incorporated into accepted patterns of speech. Nowadays, the Nitpael form, though not common in use, has not disappeared from the language completely. Though it is very uncommon in spoken colloquial Hebrew, one can find it in written Hebrew or in formal speech. For example:1 1  In the glosses I use the abbreviations H = Hitpael and N = Nitpael.

© koninklijke brill nv, leiden, ���6 | doi ��.��63/9789004310896_015



a. ‫לחם הבננה שלכבודו נתכנסנו כאן היום‬ leħem ha-banana še-li-xvod-o bread.CS the-banana that-to-honor-its nitkanasnu kan hayom gathered.N.1PL here today ‘the banana bread, for which we have gathered here today’ (, accessed July 16, 2014) b. .‫לא נתברכתי ביכולת הבנה‬ lo nitbaraxti be-yexolet havana not blessed.N.1SG with-ability.CS understanding ‘I was not blessed with the capability of understanding.’ ( id=173552587, accessed July 19, 2014) Since the two forms showed no functional difference at the time each one was in use (Biblical and Mishnaic), we would expect to see, approximately, a 50–50 distribution in the texts, according to the authors’ preferences. Instead, we can see that during the early years of the Hebrew revival, the Nitpael has been much more widely used than the Hitpael. This becomes even more intriguing when we take into account the fact that today, the tables have turned, and Hitpael is the more frequently used binyan. Thus, one question that emerges is, what made the authors of the revival days prefer to use the Nitpael? Another significant question is, what caused the later change in the distribution of the two forms, such that Hitpael is the dominant binyan in use today? Only after defining the differences between the two forms and their distribution can we try to understand the reasons for the change that occurred in Modern Hebrew. Below, I will briefly review the theories concerning these two binyanim and will then present the data regarding the distribution in use of the two binyanim as well as a possible analysis of the data.

Theoretical Background

Doron (2008:57–88) and Siloni (2008:111–138) give a semantic-syntactic analysis of Hitpael. Doron classifies Hitpael verbs as middle voice verbs with a single argument, the logical object. This logical object is syntactically realized as a subject, and may for some roots be assigned the agent role in addition to the patient role. Siloni lists the possible meanings of Hitpael as reflexive, r­ eciprocal

Hitpael and Nitpael in Modern Hebrew


and inherent-reciprocal, unaccusative, passive, experiencer, and middle. To the Nitpael form she attributes meanings such as the passive, unaccusative, and the inherent-reciprocal. Shatil (2007:105–127) and Bolozky (2010:277–289) examine this issue from a different point of view and describe the differences between Hitpael and Nitpael. Shatil suggests that Nitpael preserves the characteristics of its Mishnaic origin as a reflexive, but also as the passive form of Piel. He points to another difference at the level of agentivity: the Nitpael form encodes very little agentivity, whereas Hitpael encodes both high and low agentivity. According to Bolozky, there is no real semantic difference between the two, and any root can be conjugated in both. But he also notes that Modern Hebrew speakers, when they use Nitpael, tend to use it for the inchoative and passive meanings, while Hitpael is used for the agentive meanings as well. This reinforces Shatil’s proposal. According to all the linguists mentioned above, the possible meanings expressed by Nitpael are limited to low-agentivity meanings, whereas Hitpael can express agentive ones also. Examples for some of the uses of Hitpael: a. Reciprocal: . . . ‫ אשר התפרדו בקול רעש והמולה‬. . . ‫ לאגודות‬. . . ‫ שם התלכדו‬ šam hitlakdu . . .  le-ʔagudot . . .  ʔašer there combined.H.3PL into-bundles that hitpardu be-kol raʕaš ve-hamula . . .  separated with-sound.CS noise and-hustle ‘There they bundled . . . into groups . . . which loudly disbanded . . .’ (Mendele Mocher Sforim, The Book of the History of Nature, 1872:71) b. Reflexive:  . . . ‫וכאשר התפשטה עד מעילה‬ ve-kaʔašer hitpašta ʕad meʕil-a . . .  and-when undressed.H.3FS to coat-her ‘and when she removed her coat . . .’ (Peretz Smolenskin, A Donkey Burial, 1873:527) c. Experiencer:  . . . ‫התאויתי פתאום תאוה‬ hitʔaveti pitʔom taʔava . . .  desired.H.1SG suddenly a.desire ‘I had a sudden craving . . .’ (Uri Nissan Gnessin, The Cross, 1913:38)



Data and Preliminary Analysis

In order to find the distribution of Hitpael and Nitpael over different contexts and during the revival years, I used the Historical Hebrew Dictionary by The Academy of the Hebrew Language. I chose this corpus because it contains about 120 million tokens, and because it is morphologically analyzed, therefore providing more precise results relative to other corpora. In addition, the possibility of seeing the context in which the verb appears aids in the determination of the correct interpretation of each verb. This corpus includes only some of the Hebrew publications during that time and excludes others, such as newspapers and letters. Due to the partial representation of Hebrew texts in the corpus, it may only indicate trends. I searched for occurrences of Hitpael and Nitpael in the past tense2 in all the works written from 1870 to 1932. I found 57 different3 roots and a total of 992 examples. The distribution of the two binyanim is 266 in the Hitpael form and 726 in the Nitpael form—i.e., 2.7 times more Nitpael occurrences than Hitpael. Fig. 1 shows the distribution of use of the two binyanim during each year for which data was available. The figure shows that, for the most part, Nitpael was in much wider use than Hitpael until the 1920s, when it began to appear much less frequently. Reshef (2009:143–176) also found that the greater use of Nitpael lessened during the mid-1920s. What is not clearly visible here is the shift in the use of the two binyanim that occurs slightly later on. This is due to the absence of works that were written later than the 1930s in the corpus of the Historical Hebrew Dictionary. Another thing to look at in the results is whether there is a difference in agentivity between Hitpael and Nitpael verbs. For the ten most frequent roots (those that appeared in the texts 29 times and more), the vast majority of subjects are inanimate for the Nitpael form. For example: a.  . . . ‫אך דעתי לא נתקררה עוד‬ ʔax daʕat-I lo nitkarera ʕod . . .  but mind-my not cooled.N.3FS yet ‘But my mind hadn’t settled yet . . .’ (Moshe Leib Lilienblum, The Nether World, 1873: 608) 2  I considered only past tense verbs, since the Nitpael present or future forms are the same as those of Hitpael. 3  I considered only roots which appeared in the Nitpael form more than three times in the texts.


100 90 80 70 60 50 40 30 20 10 0

hitpael nitpael

1870 1875 1879 1886 1890 1893 1896 1899 1902 1905 1909 1912 1916 1919 1923 1931

Number of Occurences

Hitpael and Nitpael in Modern Hebrew

Year Written figure 1

The distribution of Hitpael and Nitpael, 1870–1932

b.  . . . ‫ומענין לענין נתגלגלה השיחה‬ u-me-ʕinyan le-ʕinyan nitgalgela ha-siħa . . .  and-from-matter to-matter rolled.N.3FS the-conversation ‘And the conversation moved from one thing to another.’ (Mendele Mocher Sforim, Shlomo’s Life, 1910:14) Two other frequent roots ‫( כון‬to mean) and ‫( בלבל‬to confuse), though they take animate subjects, describe events in which these subjects are not agents but experiencers, for example: c. .‫אני לדבר אחר נתכונתי‬ ʔani le-davar ʔaxer nitkavanti I to-thing different meant.N.1S ‘I meant something else.’ (Mendele Mocher Sforim, The Book of Baggers, 1909:15) d. .‫ואולם המלצר העלוב נתבלבל עוד יותר‬ ve-ʔulam ha-melcar ha-ʕaluv nitbalbel and-yet the-waiter the-poor confused.N.3MS ‘And yet the poor waiter became even more confused.’ (Hayim Nahman Bialik, The Man of the Deck, 1931:104) This fact confirms the notion that Nitpael marks low agentivity.

ʕod still

yoter more



I now turn to the question of whether the distribution of Nitpael and Hitpael was influenced by Jewish languages spoken by the first Modern Hebrew speakers. In the Comprehensive Yiddish–English Dictionary (2013), one can find several Nitpael forms: ‫( נשתומם ווערן‬be astonished), ‫נשתנה ווערן‬ (be transformed), ‫( נתגלה ווערן‬be revealed), ‫( נתעשר ווערן‬become rich), and so on. The structure of this form—the Nitpael form followed by the verb ‫ווערן‬ (to be)—indicates that the Nitpael form was not perceived as a verb (Harshav 2006:32–46), but as a passive participle. It seems, then, that the Nitpael form in Yiddish took on a low agentivity meaning, or even a passive meaning. This might have influenced the early writers and speakers of Modern Hebrew and caused them to make a distinction in which Nitpael was lower in agentivity than Hitpael. Conclusion It is possible that Nitpael was much more frequently used during the early days of the Hebrew revival since it was already familiar to speakers of Yiddish. Moreover, the passive-like use of the Nitpael in Yiddish might have influenced its use in Modern Hebrew. Furthermore, I suggest that Nitpael became less frequent in Modern Hebrew because Nitpael’s possible meanings (passive, middle, and inchoative) are a subgroup of the possible meanings of Hitpael (reflexive, reciprocal, inchoative, unaccusative, passive, experiencer, and middle). In other words, since it is possible to express every meaning of Nitpael using Hitpael, but not the other way around, it was possible for speakers to retain a single binyan without sacrificing any nuance of meaning. This change may also be attributed to different preferences in style made during the revival years—in the beginning, the preference for Rabbinic (Mishnaic) Hebrew (and even rejection of Biblical Hebrew) led to a wider use of Nitpael; and later, a preference for the Biblical Hebrew enhanced the use of Hitpael. Nitpael’s present and future forms are indistinguishable from those of Hitpael. In addition, the past Nitpael form resembles, and is sometimes even identical to, the future form of the two binyanim. Therefore, it seems plausible that Modern Hebrew speakers abandoned the Nitpael form due to the phonological resemblance between the two binyanim. It thus seems that Nitpael hasn’t completely disappeared from Modern Hebrew, and that speakers of Modern Hebrew know when it is and is not possible to use it. Nitpael is still used as the low agentivity marked form, which makes it possible for us to emphasize a more passive meaning if we wish to do

Hitpael and Nitpael in Modern Hebrew


so. It is also a form used in a higher register than that of Hitpael and therefore used in formal contexts. The relative strength of these claims can be further tested through additional research on the appearance of Nitpael and Hitpael forms in a wider range of texts and authors than the data I used. References Beinfeld, Solon & Harry Bochner. 2013. Comprehensive Yiddish–English Dictionary. Bloomington, IN: Indiana University Press. Bolozky, Shmuel. 2010. “Nitpael and Hitpael in Israeli Hebrew.” In Mishnaic Hebrew and Related Fields: Studies in Honor of Shimon Sharvit, eds. Efraim Hazan & Zohar Livnat. Ramat-Gan: Bar-Ilan University Press, 277–289 (in Hebrew). Doron, Edit. 2008. “The Contribution of the Template to Verb Meaning.” In Modern Linguistics of Hebrew, ed. G. Hatav. Jerusalem: Magnes Press, 57–88 (in Hebrew). Harshav, Benjamin. 2006. Yiddish and Jewish Discourse, vol. 4. Jerusalem: Carmel, 32–46 (in Hebrew). Reshef, Yael. 2009. “Continuity vs. Change in the Emergence of Standard Modern Hebrew: The Verbal System in the Early Mandate Period.” In Modern Hebrew: Two Hundred and Fifty Years, ed. Chaim E. Cohen. Jerusalem: Hebrew Language Academy, 143–176 (in Hebrew). Shatil, Nimrod. 2007. “The Synchronic Status of Nitpa’el.” Divrei ha-ḥug ha-yisr’eli levalshanut 16: 105–127 (in Hebrew). Siloni, Tal. 2008. “About the Hitpael Binyan.” In Modern Linguistics of Hebrew, ed. G. Hatav. Jerusalem: Magnes Press, 111–138 (in Hebrew).

Substrate Sources and Internal Evolution of Prescriptively Unwarranted Comitative Complements in Modern Hebrew Yishai Neuman

Achva Academic College, Arugot, Israel [email protected]

Abstract Modern Hebrew usage of the preposition ‫ ִעם‬ʿim ‘with’ displays syntactic and semantic features which are either non-existent or extremely rare in Classical Hebrew, among which are the instrumental and the possessive. These new features correspond to universal tendencies in language evolution and seem to have prevailed in several substrate languages of Modern Hebrew, including Yiddish and Judeo-Spanish. Until recently, they had characterized mostly colloquial Hebrew, whereas written Hebrew had conformed to Classical Hebrew. In recent years, however, social change has brought about the integration of colloquial Hebrew and written Hebrew, which has led to a decrease in the latter’s resemblance to Classical Hebrew. Normative linguists have been speaking out against the phenomenon of “misplaced ʿim.” This article is a brief account of some Hebrew ‘with’ structures, considered typologically through diachronic analysis and synchronic description of the socially-stirred interplay between colloquial and standard written Hebrew.

Keywords comitative – instrumental – possessive – substrate influence on Modern Hebrew

The Research Question

It is widely accepted that the preposition ‫ ִעם‬ʿim ‘with’ conveys meanings and uses in Modern Hebrew that it did not convey in Classical Hebrew (Avinery 1964:450ff.; Netzer 1996:133–135, 2008:157; Allon 1998; Sovran 2008, 2008–2009). One such meaning is the instrumental,1 as in “eat with a fork,” which Classical 1  Fillmore (1968:24) defines instrumental as “the case of the inanimate force or object causally involved in the action of state identified by the verb.” © koninklijke brill nv, leiden, ���6 | doi ��.��63/9789004310896_016

Substrate Sources and Internal Evolution


Hebrew conveys using the preposition -‫ ב‬be- ‘in, at, by.’ Another meaning is possessive, ranging from inalienable possession, including the part-whole relation, such as “the girl with the blue eyes” (i.e., “the blue-eyed girl”) to various types of alienable possession, including attachment, such as “he always walks with a hat” (i.e., “he wears a hat all the time”) (Croft 1991; Stolz 2001). Other uses include syntactic rection, displaying ‘with’ in verbs and adjectives which in Classical Hebrew exhibit other prepositions, usually -‫ ב‬be- ‘in, at, by,’ e.g., ‘use with,’ ‘begin with,’ ‘proceed with,’ ‘end with,’ ‘full with.’ The semantic scope of ʿim is wider in Modern Hebrew, and its level of precision is therefore lower. The key claim, which is a normative one, is that since the instrumental and the possessive—as well as other meanings and uses—are not attested in Classical Hebrew, they fall within the category of “misplaced ʿim” and are thus unwelcome in Modern Hebrew. To remedy this “syndrome,” as some call it, works of normative Hebrew provide lists of do’s and don’ts (Amotz 1932:184; Vardi 1973; Bahat & Ron 2003; Tzivoni 2014). Sovran (2008, 2008–2009) lays out the most comprehensive account so far and is balanced between the linguistic-descriptive and the cultural-prescriptive approaches. In the following brief outline we shall analyze the diachronic factors of substrate sources and internal evolution and look into the social dynamics that command the current linguistic variability.

The State of Affairs in Classical Hebrew

Biblical Hebrew ‫ עם‬ʿim ‘with’ “expresses a variety of comitative relations,” and “[the] most common sense involves accompaniment . . . or addition.”2 Previous accounts dwell on the types of relations but not on the semantics of the syntactic arguments, both governing and governed. A quick overview3 of the semantic features of the NPs involved in the Biblical Hebrew “NP1 ʿim NP2” syntagm reveals that they are primarily [+human] (including collective nouns and anthropomorphized divinities), e.g.,4 “So Saul did eat with Samuel that day” (1 Sam. 9:24) and secondarily [+animate, -human], e.g., “And the wolf shall dwell with the lamb” (Isa. 11:6). What is manifestly characteristic of those types is their reversibility: “Samuel did eat with Saul” and “the lamb shall dwell with

2  Waltke & O’Connor (1990:219). Similarly, Brockelmann (1956:111–112). 3  Cf. previous relational classifications: HALOT, Kaddari 2006. 4  The source for the English translation of Biblical verses is the 1917 Edition of the JPS Tanakh retrieved from



the wolf”; in other words, they “did eat” and “shall dwell” together. The core usage of Biblical Hebrew ‫ עם‬ʿim is therefore a “reversal with.”5 But Biblical Hebrew is no monolithic block, and upon closer observation, one may note how the argument NPs of ʿim evolve into NPs that are neither [+human] nor [+animate]. Since reversibility does not apply for these NPs, a new type of Biblical Hebrew ʿim emerges: a “non-reversal with.” Such evolved usage conveys usually the locative and rarely also the possessive, and is extant more in figurative language than in prose and more in Late Biblical Hebrew than in Classical Biblical Hebrew. Thus, the possessive use, which is generally thought to have arisen in Modern Hebrew, is rare in Biblical Hebrew, but not absent, e.g., “Thine is an arm with might; strong is Thy hand” (Ps. 89:14), where strong equals with might, as well as “I even set the people after their families with their swords, their spears, and their bows” (Neh. 4:7). The following case may illustrate a possible tendency. The Biblical ָ ‫ְב‬ Hebrew passive adjective of .‫ל‬.‫ל‬.‫ ב‬BLL in qal ‘to mix,’ that is m. ‫ ָבלּול‬, f. ‫לּולה‬ ‘mixed,’ introduces its complements by -‫ ב‬be- ‘in, at, by’ throughout its entire 37 occurrences. Mishnaic Hebrew displays one occurrence using ‘with’: ‫‘ ברזל טמא שבללו עם ברזל טהור‬impure iron which they mixed with pure iron’ (Kelim 11:4) (Azar 1995:148), and later in ‫ דקדוקי הטעמים‬Diqduqe ha-Ṭeʿamim (Baer & Hermann 1879), the major treatise of 10th-century chief Masorete (grammarian) Aaron ben Moshe ben Asher, one reads: ‫שלא תהיה זו עם זו בלולה‬ (p. 27) ‘for it not to be mixed with each other.’ Such differences in rection require further study.

The Cognitive Factor of Semantic Change: Metonymy → Polysemy

It is assumed that comitative ‘with’ conveys primarily the sameness of an action with respect to two [+human] agents, where both togetherness and reversibility/reciprocity apply, as in “Harry went to the movies with Sally.” Metonymy may apply the pivotal component of sameness to more than just action and may thus allow arguments to go beyond the constraint [+human]. The semantic criterion “sameness of action” may thus evolve into other types of sameness, such as place and time, and the semantic constraint [+human] may be dropped, which explains the loss of reversibility (Cadiot 1997, chapters 5–6; Stroh 1998; Stolz 2001). Sameness of place implies locative relations (Croft 1991; Stolz 2001) and is indeed characteristic of any object and its possessum/part, neither of which is necessarily [+human]; this may be illustrated in u ­ tterances 5  On the reversibility and “together”-compatibility criteria, cf. Schlesinger (1979).

Substrate Sources and Internal Evolution


such as “the house with the red roof,” where “the house” is the object and “the roof” is a part. But sameness of place is also characteristic of an animate agent and an instrument, such as “fishermen/seagulls catch fish with their net/beak.” By maintaining the core semantic component of sameness while modifying other semantic constraints, ‘with’ is liable to evolve into conveying other meanings, including instrumental and possessive. The predisposition of ‘with’ to evolve in such a causal chain may take its course to various degrees in different languages at different periods of time. In Classical Hebrew, such evolution may have been responsible for much unattested personal variation, but it has most likely not attained the degree whereby ‘with’ conveys instrumental, and evidence for possessive is extremely rare and in no case systematic.

A Language Universal Particularly Salient in the European Linguistic Area

The fact that change from “reversal with” to “non-reversal with” is cognitively reasonable does not mean it necessarily happens everywhere. But it sometimes happens in specific language groups. Genetic affiliation, however, is not the only criterion for the constitution of language groups. Geographic adjacency is an important factor, and languages which share a common geographic sphere come into contact. Such contact is prone to trigger the transfer of linguistic traits, both lexicon and structure, and the emergence of a linguistic area (←Sprachbund). Europe as a linguistic area exhibits some common linguistic traits, of which comitative-instrumental syncretism is a salient feature: Languages may belong to one of three classes: They use two distinct markers to encode comitative and instrumental separately (asyncretic); they do not distinguish formally between the two categories (syncretic); or they employ two markers, of which one exclusively encodes either comitative or instrumental and the other one covers both functions (= mixed). Europe stands out from the rest of the world, as it is the only continent where the syncretic type is statistically the strongest (everywhere else the asyncretic type dominates by far). Thirty-five of the European sample languages (= 69%) are syncretic, 11 are asyncretic, and five are mixed. (STOLZ 2006:289)

The distribution of the comitative-instrumental syncretism as a variant on the linguistic map of Europe (Stolz 2006:289) shows that German, Yiddish



(Birnbaum 1979:296), and (Judeo-) Spanish are among the 69 percent exhibiting full syncretism. Thus, Weinreich’s Hebrew version of College Yiddish (1977:33) relates ‫ מיט‬mit ‘with’ to both ‫ עם‬ʿim ‘with’ and -‫ ב‬be- ‘by,’ and a following language exercise asks ‫וואס שרײַבט דער לערער‬ ָ ‫ מיט‬mit vos ʃrajbt der lerer ‘with what [does] the teacher write?’ What this means is that a critical number of Jews who shifted to Hebrew spoke a substrate language in which ‘with’ also conveyed the instrumental. Other types of ‘non-reversal with’ characterize other substrate languages, e.g. colloquial Arabic ‘tea with sugar.’ The cumulative typological evidence for this tendency is of major significance for the uses of ‫ ִעם‬ʿim ‘with’ in both (1) post-classical pre-vernacular written Hebrew and (2) colloquial Hebrew at its inception.

Post-classical Non-vernacular Written Hebrew ‘With’

Medieval Hebrew is non-vernacular by definition. So is Central- and EasternEuropean Maskilic Hebrew, from 1755 on. The writers of non-vernacular Hebrew were native speakers of languages that displayed various types of “nonreversal with.” Among their sources for composing in non-vernacular written Hebrew were (1) previously written Hebrew, which they mastered to various degrees, and (2) their own vernacular. Thus, Medieval and Maskilic Hebrew exhibit instances of “non-reversal with,” which can be ascribed to structural transfer from the vernacular. Avinery (1964:450) shows that against the Biblical Hebrew instrumental ‫ ִּכי ֶּב ָענָ ן ֵא ָר ֶאה‬kī bɛʿ:ån:ån ʾērʾ:åɛ ‘for I appear (over the ark cover) in a cloud’ ‘through a cloud’ (Lev. 16:2), Rashi, a native speaker of Old French, interprets ‫ כי תמיד אני נראה שם עם עמוד ענני‬ki tamid ʾani nirʾe šam ʿim ʿamud ʿanani ‘For I continuously appear there with My pillar of cloud’ (France, 11th c.).6 Following this path, a Responsa search ‫ עם מקל‬ʿim maqel ‘with a staff’ has retrieved this: against Biblical Hebrew ‫א ַלי ַּב ַּמ ְקלֹות‬-‫א‬ ֵ ‫א ָּתה ָב‬-‫י‬ ַ ‫ ִּכ‬kī ʾatt:å ḇ:å ʾēlay bammaqqǝlōṯ ‘thou comest to me with [←‘by’] staves’ (1 Sam. 17:43), Malbim (Eastern Europe, 19th c.) comments ‫‘ ללחום עם מקל ואבנים‬fight with a staff and stones’ (17:44). Further research will certainly find more cases that are similar. If the Yiddish common expression ‫ לא מיט אן אלף‬lóy mit an álef ‘absolutely not,’ literally, ‘no with an aleph’ represents a ‘with’ structure that is foreign to Hebrew (cf. the standard Modern Hebrew counterpart ‫ לא באל"ף רבתי‬lo be-ʾalef rabati ‘no with a capital N’), then one should deem 6  The source for the translation of Rashi’s commentary is aid/63255/jewish/The-Bible-with-Rashi.htm.

Substrate Sources and Internal Evolution


foreign Medieval Hebrew altogether, e.g., grammarian David Qimḥi’s phrasing ‫‘ והחמישית בת חמש אותיות עם התי"ו והמ"ם‬and the five-letter one with the tav and the mem’ (Provence, 12–13th c.) as well as grammarian Elie Levita’s phrasing: ‫כי שני מיני חירק הם אחד עם היוד והוא תנועה גדולה והשני בלא יוד והוא תנועה קטנה‬

‘There are two types of ḥireq, one with a yod which is a long vowel and the other without a yod which is . . .’ (Italy, 1518). In late Maskilic Hebrew one also finds possessive ‘with’: ‫ כפי שאנחנו מכירים אותה‬,‫‘ האשה עם החטם הארך‬the woman with the long nose, as we know her’ (Frischmann 1905), where the author’s comment ‘as we know her’ might allude to a level of familiarity, which favors colloquial language. Both natural tendencies in human language and the specific vernaculars of the scribes seem to have affected the Hebrew grapholect whose production stretches between the end of Hebrew speech around 200 CE and the beginning of Modern Hebrew speech around 1900. Those also may have left an impact on spoken Modern Hebrew at its inception. The methodological foundations laid by Kaddari (1984) in dealing with three specific cases of Jewish vernacular impact on both written Medieval Hebrew and colloquial Modern Hebrew may serve as a point of departure for further research on the question at hand.

Modern Hebrew: Written and Spoken, Contact and Convergence

Berman (1978:121) and Glinert (1989:233) provide simple statements with respect to Modern Hebrew instrumentals: Although bə- is the normative case marker for instrumentality, in colloquial usage the comitative marker im ‘with’ is generally used instead (in a way unacceptable to purists). (BERMAN)

The means by which something is done is usually expressed by the preposition -‫ ב‬be-, or in casual usage by ‫ עם‬ʿim (condemned by purists). (GLINERT)

Both depart from normative Hebrew, and Berman’s description includes the concessive relation, as if normative Hebrew had much say in colloquial Hebrew. Indeed, when observed with a higher resolution, Modern Hebrew displays one typology in writing and another in speech, with much autonomy for both. This is due to the lineage each follows. On the one hand, written



Hebrew follows the lines of the uninterrupted practice of writing various types of Hebrew while speaking different background vernaculars; the knowledge of written Hebrew has always depended on education. On the other hand, the newly spoken Hebrew, at its inception, followed two lines: (1) vernacularized written Hebrew and (2) the previous vernaculars of the new speakers of Hebrew, none of whom was native. With respect to the use of ‘with,’ colloquial Hebrew draws on ‘non-reversal with’ languages, mostly European, and written Hebrew draws on ‘reversal with’ in previous manifestations of written Hebrew. This is the initial state of affairs. Both idioms, however, have been in contact, similar to the situation involving written Hebrew and the previous vernaculars. Indeed, given that speakers of Modern Hebrew enjoy full literacy, spoken Hebrew and written Hebrew are two idioms in contact, both social and cognitive. Such contact situations usually stimulate reciprocal transfer. So, just as written Hebrew left a significant impact on the Jewish vernaculars, which, in turn, affected written Hebrew, both idioms in Israel, the spoken and the written, are in contact, simply because users of spoken Hebrew are most often also users of written Hebrew. Conditions are thus ripe for reciprocal structural transfer to take place. As a result of the popularization of writing, typically spoken features enter written production. This trend is responsible for the increasing instances of ‘nonreversal with’ in written Hebrew of recent years. Colloquial tendencies in journalese and even in prose may be explained as the authors’ wish to imitate their audience’s supposed spoken language (Shlesinger 2000:74). With respect to literary prose, Tzivoni (2014:281–282) draws up a list of novelists who use “misplaced ʿim,” among which she cites Eshkol Nevo, Yehonathan Geffen, David Grossman, and Michal Shalev. Undoubtedly, the quest for authenticity in realistic writing favors colloquial features in writing. In the opposite direction, education and the resulting impact of the written language on the spoken language are responsible for the fact that syntagms which in early spoken Hebrew displayed ‫ עם‬ʿim ‘with’ have gradually come to exhibit -‫ ב‬be-, e.g., ‫‘ משתמש עם‬to use with’ ‘to use’ → -‫‘ משתמש ב‬to use by’ ‘to use.’7 From a crossgenerational point of view, the normative -‫‘ משתמש ב‬to use by’ appears both in standard writing and in speech by educated speakers whose less educated parents use ‫‘ משתמש עם‬to use with’; and ‫‘ משתמש עם‬to use with’ appears in unedited writing by semi-educated language users, who can read and write to some degree and whose even less-educated parents had generally not had the skill of writing. Transfer is therefore extant, on the one hand, from colloquial Hebrew 7  A Google search has retrieved 3,040 and 5,590 hits for ‫ עם מה משתמשים‬and ‫במה משתמשים‬ (literally “use with” and “use by”) respectively (accessed August 31, 2014).

Substrate Sources and Internal Evolution


into semi-educated Hebrew writing, as in some Internet sources, and, on the other hand, from standard written Hebrew into educated spoken Hebrew. Given that (1) the uneducated can technically write but ignore much of the language of writing and that (2) the educated can imitate writing features in speech, transfer of features is reciprocal, but the streams are contradictory, as the educated and the uneducated are disjoint sets of language users. To illustrate this, we shall consider Google hits for “arrive by bus,” “arrive by taxi,” and “by public transportation.” Educated spoken Hebrew would use ‫ באוטובוס‬be-ʾoṭobus ‘by bus’ whereas spoken Hebrew, untouched by typologically different written Hebrew, maintains ‫ עם אוטובוס‬ʿim ʾoṭobus ‘with a bus.’ And just as educated speech displays transfer from written Hebrew, unedited Internet writing displays the impact of spoken Hebrew. Thus, a Google search of ‫‘ להגיע באוטובוס‬arrive by bus’ and ‫‘ להגיע עם אוטובוס‬arrive with a bus’ provides 14,600 and 64,100 hits (=18.5% vs. 81.5%) respectively; but since the more expensive means assumes a better socio-economic situation and concomitantly better education, the search ‫‘ להגיע במונית‬arrive by taxi’ and ‫‘ להגיע עם מונית‬arrive with a taxi,’ yielding 12,800 and 24,900 hits (=34% vs. 66%), shows a higher rate of the educated syntagm for taxi than for bus (34% vs. 18.5%). Finally, the formal term ‫‘ תחבורה ציבורית‬public transportation’ assumes an even higher level of education, and the results are 204,000 hits for ‫ עם‬ʿim and 275,000 for -‫ ב‬be- (=42.5% vs. 57.5%).8 The gradually increasing level of education associated with bus, taxi, and public transportation is matched by respective be- hit rates of 18.5%, 34%, and 57.5% at the expense of ʿim. As education spreads, the use of spoken features in writing falls and the use of written Hebrew features in speech rises. The writing of the uneducated and the speech of the educated enhance the convergence of both idioms, so both evolve in parallel. Last but not least, one should also consider the presence of English and its possible influence. Since the knowledge of English correlates with higher education and therefore with social prestige, ‘with’ structures in colloquial Hebrew whose English counterparts exhibit with, such as ‫‘ תה עם סוכר‬tea with sugar,’ may lose the stigma of being only colloquial. This means that English filter-influence may in effect account for the resistance of originally popular structures and expressions throughout society, regardless of the level of education. Further corpus-based quantitative research is needed to ascertain and quantify these impressions.

8  Results retrieved on August 31, 2014.



Conclusion The study of ‫ עם‬ʿim ‘with’ throughout the history of Hebrew shows how cognitive-internal, contact-external, and socio-cultural factors stimulate both convergent and divergent evolution in written Hebrew and spoken Hebrew. Given that the written Hebrew and the spoken Hebrew of our time stem from partly different sources, with substrate vernaculars determining much of spoken Hebrew but far less of written Hebrew, both idioms initially exhibit partly different typologies. At the same time, the spread of literacy among some and the increase in access to higher education among others encourage overall synchronic convergence between both idioms, but cross-generational changes due to the rise of education lead, in fact, to diachronic divergence within more homogeneous social groups. Proponents of normative Hebrew sound the alarm to make writers aware of how written Hebrew has been and therefore should remain, and their efforts would probably not surpass the barrier of education. This being said, the extant spread of higher education to more layers of Israeli society and the consequent generational change in language use between uneducated parents and educated children may foretell some level of success of the endeavor. From the purely typological point of view, the case of ‫ עם‬ʿim ‘with’ is yet another example of the intricate interplay of the spoken, the written, and their relations in an ever-changing social eco-system. References Allon, Emmanuel. 1998. “The Preposition ‫ ִעם‬ʿim ‘With’ in Israeli Hebrew.” Israeli Linguistic Society 10: 5–10 (in Hebrew). Amotz. 1932. “Comments and Suggestions.” Lešonénu 4: 184–185 (in Hebrew). Avinery, Isaac. 1964. Yad ha-lashon. Tel-Aviv: Izre’el (in Hebrew). Azar, Moshe. 1995. The Syntax of Mishnaic Hebrew: Jerusalem: The Academy of the Hebrew Language (in Hebrew). Baer, Seligmann & Hermann L. Strack. 1879. Die Dikduke ha-teamim des Ahron ben Moscheh ben Ascher und andere alte grammatisch-massorethische Lehrstücke. Leipzig: Fernau. Bahat, Yaakov & Mordechai Ron. 2003. Ve-dayek: Hebrew Language Guide for Grammar and Style, 28th edition. Tel Aviv: Hakibbotz Hameuchad (in Hebrew). Berman, Ruth. 1978. Modern Hebrew Structure. Tel-Aviv: University Publishing Projects. Birnbaum, Solomon A. 1979. Yiddish: A Survey and a Grammar. Toronto: University of Toronto Press. Brockelmann, Carl. 1956. Hebräische Syntax. Neukirchen: Kreis Moers.

Substrate Sources and Internal Evolution


Cadiot, Pierre. 1997. Les prépositions abstraites en français. Paris: Armand Colin. Croft, William. 1991. Syntactic Categories and Grammatical Relations: The Cognitive Organization of Information. Chicago: University of Chicago Press. Fillmore, Charles J. 1968. “The Case for Case.” In Universals in Linguistic Theory, eds. Emmon W. Bach & Robert T. Harms. New York: Holt, Rinehart, and Winston, 1–88. Frischmann, David. 1905. “Herbert Spencer and George Eliot.” (in Hebrew). Glinert, Lewis. 1989. The Grammar of Modern Hebrew. Cambridge: Cambridge University Press. HALOT = Köhler, Ludwig & Walter Baumgartner. 1994–2000. The Hebrew and Aramaic Lexicon of the Old Testament. Leiden: E. J. Brill. Kaddari, Menaḥem Z. 1984. “Hebrew mikol-makom, mamash, and nitpael in Responsa Literature (with Reference to Probable or Possible Influence on Hebrew).” Jewish Language Review 4: 30–33. ———. 2006. A Dictionary of Biblical Hebrew (Alef-Taw). Ramat-Gan: Bar-Ilan University Press (in Hebrew). Levita, Elia. 1527. Pirqe Eliyahu. Capitula cantici, specierum, proprietatum, & officiorum . . . / autore Elia Levita aeditum, & per Sebastianum Munsterum . . . Latine iuxta Hebraismum versum. Basel. Originally composed in 1518 (in Hebrew). Netzer, Nissan. 1995–1996. “Hebrew Prepositions: The Influence of Foreign Languages.” Leshonenu La’am 47: 132–138. ———. 2008. Hebrew in Jeans: The Image of Hebrew Slang. Beer-Sheva: Ben-Gurion University Press (in Hebrew). Schlesinger, I. M. 1979. “Cognitive Structures and Semantic Deep Structures: The Case of the Instrumental.” Journal of Linguistics 15: 307–324. Shlesinger, Yizhak. 2000. Journalistic Hebrew: Stylistic Aspects of Israeli Newspaper Sections. Beer-Sheva: Ben-Gurion University Press (in Hebrew). Sovran, Tamar. 2008. “The Logic of Addition: Changes in the Meaning of the Hebrew Preposition ‘im (With).” In Adpositions: Pragmatic, Semantic, and Syntactic Perspectives, eds. Denis Kurzon & Silvia Adler. Amsterdam: John Benjamins, 257–271. ———. 2008–2009. “‘Im (with) Syndrome’—Modes of Adding and Connecting in Israeli Hebrew.” Ha-ʿivrit ve-ʾaḥyoteha 8–9: 245–270 (in Hebrew). Stolz, Thomas. 2001. “To Be With X Is to Have X: Comitatives, Instrumentals, Locative, and Predicative Possession.” Linguistics 39.2: 321–350. ———. 2006. “Europe as a Linguistic Area.” In Encyclopedia of Language and Linguistics, 2nd edition, eds. Anne Anderson et al. Boston: Elsevier, 278–295. Stroh, Cornelia. 1998. “Die Geschichte der Mit-Relationen im Französischen: Komitativ-Instrumental-Synkretismus mit Hindernissen.” Sprachtypologie und Universalienforschung 51, no. 2: 131–156.



Tzivoni, Lea. 2014. Le-takanat ha-lashon: Normative Hebrew. Tel-Aviv: Tzivonim (in Hebrew). Vardi, Tzvi. 1973. “On the Use of the Preposition ‘im (With).” Leshonenu la-‘am 24, no. 5: 128–132 (in Hebrew). Waltke, Bruce K. & Michael P. O’Connor. 1990. An Introduction to Biblical Hebrew Syntax. Winona Lake, IN: Eisenbrauns. Weinreich, Uriel. 1977. College Yiddish: An Introduction to the Yiddish Language and to Jewish Life and Culture. Hebrew translation ‫ יידיש לאוניברסיטה‬by Shoshana Bahat. Jerusalem: Magnes Press, from the 5th English edition, 1974. New York: YIVO (in Hebrew).

Inheritance and Slavic Contact in the Polysemy of bixlal Avigail Tsirkin-Sadan

Mandel Scholion Interdisciplinary Research Center, The Hebrew University of Jerusalem, Jerusalem, Israel [email protected]; [email protected]

Abstract This article provides a brief diachronic view of the way the Modern Hebrew adverb ‫ בכלל‬bixlal came to acquire multiple meanings and functions, through a combination of inheritance from Classical Hebrew uses and language contact with the Slavic languages during the period of Hebrew Revival. The path of this development corresponds to regular changes documented in other languages.

Keywords ‫ בכלל‬bixlal‒ prepositional phrase (PP) ‒ discourse marker (DM)‒ diachronic develop-

ment of adverbs ‒ language contact ‒ Modern Hebrew

Introduction The Modern Hebrew adverb ‫ בכלל‬bixlal (‘in general’ / ‘at all’ / ‘actually’ / ‘generally’ / ‘anyway’) is puzzlingly versatile from a synchronic point of view, and has therefore been the focus of several papers analyzing its semantics (Migron 2003; Ziv 2012; Greenberg & Khrizman 2012; Kadmon & Sevi 2014; Greenberg

* I wish to thank Miri Bar-Ziv Levy, Hanan Bordin, Yael Greenberg, Yael Reshef, Moshe Taube, Hanna Volovici, Yael Ziv, two anonymous reviewers, and all the kind native speakers who have taken the time to answer my questionnaire, with special thanks to Edit Doron and Chanan Ariel for their invaluable help during the writing of this article; any errors or inaccuracies are mine. I gratefully acknowledge a PhD scholarship from the Mandel Scholion Interdisciplinary Research Center in the Humanities and Jewish Studies at the Hebrew University of Jerusalem.

© koninklijke brill nv, leiden, ���6 | doi 10.1163/9789004310896_017



2014). The different uses of bixlal sketch a diachronic path of functional development. It was first used in a prepositional phrase (PP), later became an adverb, and eventually acquired discourse-marker function. A similar path of adverbial development is reported for other languages (e.g., English and Japanese in Traugott & Dasher 2001, German in Abraham 1991:373, cited in Traugott & Dasher 2001:153–154). The following sections describe this process one use at a time, in the order of their emergence; the last section briefly discusses the role of contact and the status of the new developments as grammaticalization. Noun klal within the Prepositional Phrase bi-[ḵlal . . .]1 (Rabbinic Hebrew: Mishnah) The noun ‫ כלל‬klal ‘a totality’/‘a generalization/rule’ is derived in Mishnaic Hebrew from the root ‫ כלל‬k.l.l., which is also at the core of the Biblical Hebrew quantifier ‫ ּכֹל‬kōl ‘all, every’ and the Biblical Hebrew verb ‫ ָּכ ַלל‬kālal ‘to perfect,’ which acquired two more meanings in early Rabbinic Hebrew: ‘to include’ and ‘to generalize’ (Ben Yehuda 1980:2401). In Rabbinic Hebrew from the Mishnah on, the noun klal was often used within a PP headed by be-‘in:’ bi-[ḵlal.cs NP], where klal is in the construct state.2 The annex NP conventionally denotes a category to which a relevant halachic law applies, and the phrase as a whole conveys a relation of being included in that category. (1) ‫ שהדלועין בכלל ירק וקטנית אינן בכלל ירק‬. . . še-had-deluʿin bi-ḵlal yaraq ve-qiṭnit that-the-gourds in-totality.cs vegetable and-legume ʾeynan bi-ḵlal yaraq neg in-totality.cs vegetable ‘Yet gourds are included among vegetables, and pulse is not included among vegetables.’ (Mishnah, Nedarim 7a; Danby 1933:272) 1  To the extent that there might have been one authentic pronunciation of ‫בכלל‬, it is possible that it had a definite variant ba-klal ‘in.the-totality’ (such as inclusive until ‫עד ועד בכלל‬ ʕad ve-ʕad bi-ḵlal/ba-klal), which was not preserved. Such pronunciation is preserved, for example, in the Yemenite tradition, as can be seen in Yosef ʕamar’s rendition of the Babylonian Talmud (e.g., Temura 13b, Šabat 80b). However, the ancient tradition represented in Torat kohanim (Vatican 66 ms.) features the bi-ḵlal variant in similar cases (e.g., definite klal followed by a demonstrative, e.g. in Kedošim A1 or Nedava A1). Special thanks to Chanan Ariel for his help in exploring this question. 2  Construct state is abbreviated as cs in the glosses.

inheritance and slavic contact in the polysemy of bixlal


The noun klal also gained a specific scholastic flavor with the Talmudic hermeneutics of klal u- p̄ raṭ ‘general and particular’ used to label the extent of a Halakhic law and its specifics.3 klal and praṭ appear in prepositional phrases such as bi-ḵlal and bi-ḵlal u-p̄ raṭ (u-ḵlal). Used liberally in the Talmud, these PPs serve to apply the Talmudic hermeneutics klal u-p̄ raṭ, often as predicates in clauses where the subject is a biblical verse or term (2). (2) ‫ ומה קללה בכלל ופרט אף ברכה בכלל ופרט‬ u-ma qəlala bi-[ḵlal u-p̄ raṭ] ʾaf and-what curse in-[‘general and-particular’] also bəraḵa bi-[ḵlal u-p̄ raṭ] blessing in-[‘general and-particular’] ‘Curses are interpreted using the hermeneutics of “general and particular,” so blessings must be likewise interpreted using the hermeneutics of “general and particular.”’ (Babylonian Talmud, Sota 37b, BIRP) Revival Hebrew inherited the use of klal within a PP headed by the preposition be- (3), and it is still used in present-day Hebrew, mostly in the inclusion sense. The lexical status of the noun klal can be seen in its neutral pronunciation by Modern Hebrew speakers as be-klal, in contrast with the pronunciation bixlal (Neudecker 2013), the grammaticalized adverb. (3) ‫ואל יפלא לומר שקיסרי על שפת הים היא בכלל גליל העליון‬ ve-ʔal yipale lomar še-keysari ʕal sfat ha-yam and-not surprising to.say that- Caesarea on lip.cs the-sea hi bi-xlal galil ha-ʕelyon she in-totality.cs Galilee the-upper ‘And one should not be surprised when told that Caesarea on the shore of the Mediterranean is included in the Upper Galilee.’ (Ha-melic, February 21, 1861, HJP)

Negative Polarity klal ‘at All’ (Rabbinic Hebrew: Babylonian Talmud)

The Negative-Polarity Item (NPI) klal (3) cannot be dated as early as the Mishnah. The only reported occurrences of NPI klal in the Mishnah appear in 3  For more information on the hermeneutics, see “‫ פרט וכלל‬,‫ ”כלל ופרט‬in the Talmudic Encyclopedia, vol. 31, columns ‫רצ‬-‫( רסט‬269–290).



three successive sentences in Taʕanit 2:6 (Kosovsky 1967:966). However, klal is absent from the same verses in the Kauffman and Parma manuscripts,4 and must be a later addition. Originally, the Talmudic NPI use of klal could be related to the Biblical Hebrew quantifier kōl ‘all,’ which derives from the same root and also has an NPI function.5 It is much more prominent in the Aramaic portions of the Babylonian Talmud, in which its distribution is not limited to negative sentences but also includes if-clauses. This suggests that the source of NPI klal in Rabbinic Hebrew may have been contact with Aramaic. (4) .‫אין שואלין בשלום אשה כלל‬ ʾeyn šoʾalin bi-šlom ʾišša klal neg ask.3pl in-peace.cs woman at.all ‘One must not ask about a woman’s well-being at all.’ (Babylonian Talmud, Kidušin 70b; BIRP) In Medieval Hebrew the NPI acquired two binomial versions: ‫ כלל וכלל‬klal u-ḵlal ‘at all’ and ‫ כלל ועיקר‬klal ve-ʿiqar ‘at all’ (Ben Yehuda 1980:2406–7), added to a similar Talmudic use with the double form ‫ כלל כלל‬klal klal (Kosovsky 1954:441). Since Revival Hebrew, the NPI function has split into two co-existing forms: Inherited klal (5) and Modern bixlal (6). The traditional form klal is limited in Modern Hebrew to a formal register (Glinert 2013), despite its prescriptive status as the “correct” form (e.g., Barak & Gadiš 2008–2009:37, item 19). The two items also differ in distribution. Modern Hebrew bixlal appears in various syntactic positions in diverse downward-entailing environments such as negative clauses, interrogatives, and if-clauses (as in (6) below; Migron 2003, Greenberg & Khrizman 2012:140, Glinert 2013). klal, on the other hand, seems to prefer negative clauses in Modern Hebrew, where its distribution is relatively flexible, in comparison with other environments (such as if-clauses) where, to the extent that it can occur, klal is restricted to clause-final position.6

4  The manuscripts were browsed digitally at the Israel National Library ( dl/talmud/mishna, accessed July 15, 2014). 5  For a formal semantic treatment of Modern Hebrew kol in positive and negative environ­ ments, see Bar-Lev & Margulis 2013. 6  For more on diachronic processes of NPIs in Hebrew see Keren 2015.

inheritance and slavic contact in the polysemy of bixlal


(5) ‫] שבפסח העבר לא נזהר כלל בחמץ‬. . .[ še-be-pesaħ he-ʕavar lo nizhar klal be-ħamec that-in-Passover the-past not was.careful at.all in-leavened.dough ‘That last Passover he did not mind the ħamec (leavened dough banned during Passover) at all.’ (Yosef Haim Brenner, “xidušim,” c. 1904–1908, BYP) (6) ‫] והנני עוסק עתה בבירור‬. . .[ ‫מצאתי פה את ענינַ י סבוכים והכל בערבוביא‬

.‫החשבונות—אם בכלל אפשר יהיה לבררם‬

macati po ʔet ʕinyanay svuxim ve-ha-kol be-ʕirbuviya I.found here acc messy and-the-all in-disorder ve-hineni ʕosek ʕata be-verur ha-xešbonot— dealing now in-clarifying the-calculations ʔim bixlal ʔefšar yihiye levareram if at.all possible to.clarify.them ‘(Upon my return) I found my business in shambles, and I am now busy sorting out the accounting—if it can be sorted out at all.’ (Ahad Haʿam’s letter to Mordechai Ben Hillel Ha-Kohen, Warsaw, March 19, 1897, BYP)

Adverb ‘in General’ biḵlal (Medieval Hebrew)

The first stage of grammaticalization seems to have taken place in Medieval Hebrew, from the PP bi-ḵlal into the adverb biḵlal ‘in general’ (and, in parallel, into the matching ‫ בפרט‬bip̄ raṭ ‘in particular’). This use was acquired under the influence of a similar distinction in contemporaneous Arabic philosophical and scientific texts (Goshen-Gottstein 2006:130, item 122(3)b and fn. 29). biḵlal is a PP-derived adverb, and, as a preliminary observation which requires further research, tended in its early use to modify phrases (as in (7) below) rather than full clauses. (7) ‫ והוא כלי להגיע החיות אל כל האברים בכלל ולמוח‬,‫] הלב הוא עקר קיום הבעל חי‬. . .[ ‫בפרט‬

ha-leḇ hu ʿiqar qiyum ha-baʿal-ħay ve-hu the-heart he core.cs existence.cs the-living-organism and-he kəli lehagiaʿ ha-ħiyut ʿel kol ha-ʿavarim vessel to.make.arrive the-life to all biḵlal ve-la-moaħ bip̄ raṭ in-general and-to.the-brain in-particular



‘The heart is the core of existence of the live being, and it is a vessel to drive the force of life into all parts of the body in general and the brain in particular.’ (Rabbi Yosef Albo, Sefer ha-ʿiqarim, A:6, Spain, 15th century, BIRP) The adverbs bixlal and bifrat were inherited by Revival Hebrew from Medieval Hebrew (as in [8]).7 (8) .‫כי מרבה להטיב הוא עם כל יושבי ארצו הגדולה בכלל וגם עם היהודים שמה בפרט‬ ki marbe leheytiv hu ʕim kol yošvey ʔarco for does.much to.benefit he with all inhabitants.cs ha-gdola bixlal ve-gam ʕim ha-yehudim the-big in.general and-also with the-Jews šama bifrat there in.particular ‘For he constantly acts in the best interests of all the inhabitants of his great land in general and the local Jews in particular.’ (Ha-magid, December 25, 1862, HJP)

Adverb ‘Actually’/’Rather’ bixlal (Modern Hebrew)

This new adverbial use of bixlal is very different in meaning from the ‘in general’ adverb mentioned above. In the semantics literature it has been paraphrased as ‘actually’ or ‘in fact’ (Migron 2003, 2005a, 2005b; Greenberg & Khrizman 2012:140). It conveys that the truth of its clause is less expected than that of alternatives in the context, usually challenging another speaker’s statement or presupposition.8 bixlal ‘actually’ is widely used in present-day Hebrew­

7  This particular use in Modern Hebrew was recently analyzed by Sevi & Kadmon 2014 as establishing a category-specimen relation, especially when adjoined to the subject of a nominal sentence. 8  Migron describes two readings of unstressed bixlal in non-negative environments: ‘actually’ and ‘even,’ distinguished by the scalarity of the latter. Here, both are conflated into one, for methodological reasons. Two of Migron’s examples for ‘even’ and ‘actually’ are the following (2003:4, examples 14, 15 respectively): (14) A: dani ve-ruti makirim? ‘Are Dani and Ruti acquainted?’ B: betax, hem bixlal ħaverim ‘Of course. They are even friends.’ (15) A: ruti lo belgit, naxon? ‘Ruti isn’t Belgian, is she?’ B: lo, hi bixlal carfatiya ‘No, she’s actually French.’

inheritance and slavic contact in the polysemy of bixlal


conversation, though to a lesser extent in writing, and it is also manifest in Revival Hebrew, though less frequently. Sentence (9), for example, conveys that the Liberals’ joining the Catholics contradicts what may be expected from their differing views. (9) ‫ אשר בכלל‬,‫] אל הקאטאליקים‬. . .[ ‫] נתחברו הליברלים‬. . .[ ‫לרגלי סיבות שונות‬ ‫נפרדים הם עוד בדעותיהם‬

le-ragley sibot šonot nitħabru ha-liberalim ʔel for reasons various joined.3pl the-Liberals to ha-katolikim ʔašer bixlal nifradim hem ʕod the-Catholics that actually separate they still be-deʕoteyhem in-their.opinions ‘For various reasons the Liberals joined the Catholics, who are of rather different political views.’ (Ha-magid, January 1, 1885, HJP) Unlike the other Modern Hebrew uses of bixlal, which were inherited (with some variation) from older layers of Hebrew, the adverb bixlal ‘actually’ is new.

Discourse Marker ‘Generally’ bixlal (Modern Hebrew)

This use of bixlal is a continuation of the cline of ‘in general’ from full PP to adverb, and from adverb to discourse marker (DM).9 DM bixlal ‘generally’ differs from its adverbial counterpart in scope and in affecting the discourse: it is a sentential adverb with a discourse function, marking its clause as a generalization of an issue in the conversational context, whose specifics are expressed in utterance(s) in the discourse rather than phrases within the sentence. For example, (10) refers to the tendency of Hungarian Jews to openly confront antiSemites. This generalization follows a discourse chunk in which a specific case was discussed in detail.

9  Discourse markers (DMs) are defined as “expressions which signal a relationship across rather than within utterances, and contribute to the coherence of the discourse” (Fraser & Malamud-Makowski 1996:864, cited in Traugott & Dasher 2001:155). Indeed, sentential adverbs are used as DMs in various languages (e.g., Traugott & Dasher 2001).



(10) ‫בכלל זו רעה חולה מצד רבים מאחב"י באונגארן שנלחמים עם הצוררים ולא ישימו‬ ]. . .[ ‫] יגדילו עוד השנאה לישראל‬. . .[ ‫על לבם כי עי"ז‬ bixlal zo raʕa ħola mi-cad rabim me-ʔaħeynu generally this bad sick from-side many of-our.brothers bney-yisraʔel be-ʔungaren še-nilħamim ʕim ha-corerim Jews in-Hungary with the-persecutors ve-lo yasimu ʕal libam ki ʕal-yedey-ze yagdilu and-neg on their.heart that thus ʕod ha-sinʔa le-yisraʔel more the-hate to-Jews ‘Generally, many of our Jewish brethren in Hungary have a reprehensible tendency to fight the anti-Semites, without realizing that by so doing they deepen the hatred towards the Jews even more.’ (Ha-magid, February 5, 1885, HJP) Sentences with DM bixlal also serve to conclude one discourse topic and lead to another, contributing to the segmentation of the discourse. Traugott & Dasher (2001) claim that over time DMs derived from sentential adverbs lose much of their original semantics, and are eventually left with a purely discourse-procedural function. Some evidence that this applies to bixlal can be seen in the contrast between (10) above and (11) below: in (10) bixlal marks its clause as a summary of the discussion while retaining the ‘generally’ meaning, facilitating a change of topic; bixlal in (11), on the other hand, is used in a meta-discursive utterance that changes the topic without mention of previous content at all. Thus, DM bixlal can be used, as in (11), to affect the structure of the discourse without marking a generalization.10 (11) .‫ בזה הלא כבר דיברנו לא פעם‬:‫ הבה ונדבר בעניין אחר‬,‫ יווה איסאקֹובנה‬,‫ובכלל‬ u-vixlal, yeva isakovna, hava u-nedaber be-ʕinyan and-anyway Yeva Issakovna let’s in-topic ʔaħer: be-ze halo kvar dibarnu lo paʕam other: in-this prtcl already we.talked not once ‘Anyway, Eva Issacovna, let’s talk about a different topic: we have discussed this more than once, haven’t we.’ (Yossef Haim Brenner, Misaviv la-nekuda, chapter 12, 1904, BYP)11

10  Ziv (2012:268–269) also identifies a slightly different DM use of bixlal as expressing the speaker’s disapproval.

inheritance and slavic contact in the polysemy of bixlal


The Modern Uses and Contact with Slavic Languages Table 1 is a listing of the uses divided into inherited and new. The three new uses were introduced at once during the revernacularization of Hebrew. During this process various layers of Hebrew came into contact with Yiddish and Slavic languages spoken by the revivalists. The contribution of contact to the meaning changes of bixlal is supported by the existence of a single adverb, Russian вообще voobŝe and Polish w ogóle vogule, which covers the entire range of adverbial meanings (excluding the PP-inclusion meaning). Yiddish, in contrast, does not cover the full range of meanings with a single item. Further evidence of this asymmetry is found in the results of a questionnaire answered by native Yiddish, Polish, and Russian speakers, whose main results are summarized in Table 2.12 The results show that voobŝe/vogule are table 1

Inherited and new uses of bixlal

Inherited Uses

PP inclusion/generalization bi-ḵlal NPI ‘at all’ klal Adverb ‘in general’ biḵlal

Rabbinic Hebrew: Mishnah Rabbinic Hebrew: Babylonian Talmud Medieval Hebrew

New Uses Adverb ‘actually’ bixlal NPI ‘at all’ bixlal DM ‘generally’ bixlal

Modern Hebrew (since Revival Hebrew)

11  Yossef Haim Brenner, Brenner’s Collected Works, vol. 1. Tel Aviv: Ha-kibuc ha-meʔuxad, p. 497. This novel was published in installments in Ha-meʕorer, edited by Brenner in London, and can be dated to 1904. bixlal here is pronounced vixlal due to the phonological rule of spirantization (Rafi Tsirkin-Sadan, P.C.). 12   The questionnaires consisted of a translation task of the same twelve sentences in Hebrew with different uses of bixlal in random order. Most sentences were based on unambiguous examples from the semantic literature on bixlal (e.g., Migron 2003, Ziv 2012, Greenberg & Khrizman 2012), and some on attested examples. The participants were asked to translate the sentences from Hebrew into their mother tongues. Five Russian speakers, eight Polish speakers, and four Yiddish speakers answered the questionnaire in June-July 2014. The results are partially supported by relevant entries in Russian, Polish, and Yiddish dictionaries, limited to those uses which are listed in dictionaries. The order of the terms as presented in the table does not imply frequency. Rarer terms which occurred only for one speaker on one item of the questionnaire were usually left out.

224 table 2

tsirkin-sadan Questionnaire results Yiddish



NPI ‘at all’

in gancn; legamre; iberhoypt; bixlal

voobŝe; sovsem; vovce; covrešenno

Adverb ‘in general’

in algemeyn

vogule; v cale; pžeciež v cale; yakikolviek vogule; ogolnie; v ogolnosci

Adverb ‘actually’

dox; iberhoypt; in algemeyn

DM ‘generally’

say vi; iberhoypt; in algemeyn; fundestvegn

vogule; pžeciež; navet; v cale; calkiem; cupełnie; tak napravde vogule

voobŝe; v celom; v chastnosti voobŝe; daže; sobstvenno; na samom dele voobŝe, voobŝe govorya

consistently available for four types of bixlal (though in certain cases synonyms are preferred),13 while Yiddish speakers resort to different items in different contexts. Modern Hebrew bixlal emerged out of inherited uses of bi-xlal and klal on the one hand and Slavic voobŝe/vogule and their denotations on the other. Morphologically, voobŝe and vogule, too, are adverbs derived from PPs headed by the locative preposition v- (and allomorph vo-), which resembles Hebrew b- in both sound and function.14 The process should be described as reanalysis rather than calquing/borrowing, since bixlal already had a rich historical basis within the language, namely, bi-xlal and klal. Reanalyzing the traditional bi-xlal after the model of Slavic voobŝe/vogule resulted in widening the scope 13  Some Russian speakers prefer sovsem or vovce over voobŝe in a nominal sentence with a negated predicate, both of which resemble voobŝe in form and content; most Polish speakers who answered the questionnaire preferred pžeciež in one of the ‘actually’ uses. 14   Maschler (2009:215) mentions “a productive process in Hebrew grammar of deriving adverbs from nouns via prefixation with the preposition b(e)- (‘in’) . . .” One wonders whether this pattern of adverbial formation in general could be due to Slavic influence. In reference to Yiddish, Katz (1987:185) mentions prepositions in the Hebrew/Aramaic components of Yiddish that act as “adverb formers,” including be-.

inheritance and slavic contact in the polysemy of bixlal


of the existing ‘in general’ adverb to discourse level ‘generally,’ adding the ‘actually’ use, and subsuming under bixlal the NPI function previously associated with traditional klal. Turning away from the influence of language contact, this semantic process can be seen as a natural step in the development of the adverb within the language. The direction of its development over the ages is comparable with regular patterns of change of PP-derived adverbs into DMs in various languages (Traugott & Dasher 2001, a.o.), and also in Hebrew (Maschler 2009). Two of the changes along this path could be analyzed as grammaticalization: the early shift from PP to adverb in Medieval Hebrew; and the later contact-induced appearance of the DM use. The term ‘grammaticalization’ here includes category-shifts (PP into adverb) and the shift “from contentful to procedural” (Traugott & Dasher 2001:189) in the adverb-DM step of the process. To the extent that external influence is viewed as a legitimate trigger of grammaticalization, it is possible to conclude that bixlal in Modern Hebrew successfully simulates the result of a natural semantic development, via language contact. References Bar Ilan Responsa Project (BIRP). 2013. CD edition version 21. Bar-Lev, Moshe E. & Daniel Margulis. 2013. “Hebrew kol: A Universal Quantifier as an Undercover Existential.” In Proceedings of Sinn und Bedeutung 18 in the Basque Country, eds. Urtzi Etxeberria, A. Fălăuş, A. Irurtzun, & B. Leferman. Leioa: University of the Basque Country Press, 60–76. Ben-Yehuda, Eliezer. 1980. “‫( ”כלל‬klal). In Milon ha-lašon ha-ʕivrit ha-yešana ve-haxadaša (Dictionary of the Hebrew Language, Old and New, vol. 5). Jerusalem: Makor Publishing, 2402–2407 (in Hebrew). Ben-Yehuda Project (BYP). (accessed February 3, 2014). Barak, Smadar & Ronit Gadiš. 2008–2009. Safa Qama: Selections from the Leshonenu laʕam Column, 1932–1944. Jerusalem: The Academy of the Hebrew Language, 37 (Item 19) Coref, Efraim, “Tikunim: bixlal, me-ʔeyfo, pašat ve-lavaš, nasaʕ” (Corrections: bixlal, etc.) Lešonenu la-ʕam column 4, Ha-arec, December 2, 1932, cited in: S. Barak & R. Gadiš. 2008–2009. Safa Qama: Selections from the Leshonenu laʕam Column, 1932– 1944. Jerusalem: The Academy of the Hebrew Language, 37 (Item 19). Glinert, Lewis. 2013. “Negation: Modern Hebrew.” In Encyclopedia of Hebrew Language and Linguistics, eds. Geoffrey Khan et al. Brill Online. http://referenceworks.brill



Goldberg, Zalman N. et al., eds. 2013. “Klal u-frat, prat-u-xlal” (“General and Particular, Particular and General”). In Encyclopedia Talmudit (Talmudic Encyclopedia), vol. 31, columns ‫רצ‬-‫רסט‬. Jerusalem: Yad Ha-Rav Hertzog, 269–290 (in Hebrew). Goshen-Gottstein, M. H. 2006. Syntax and Vocabulary of Medieval Hebrew as Influenced by Arabic. Revised by Shraga Assif and Uri Melamed. Jerusalem: Ben-Zvi Institute and the Hebrew University of Jerusalem. Greenberg, Yael & Keren Khrizman. 2012. “Bixlal—A General Strengthening Operator.” In Proceedings of IATL 27, ed. Evan Cohen. Pittsburg: MIT Working Papers in Linguistics 65, 139–161. Historical Jewish Press (HJP). (accessed August 22, 2014). Kadmon, Nirit & Aldo Sevi. 2014. “Nuclear Accent and Lexical Meaning: The Case of bixlal.” Presentation at the Workshop on Focus Sensitive Expressions from a Cross Linguistic Perspective, Bar Ilan University, Ramat Gan. Katz, Dovid. 1987. Grammar of the Yiddish Language. London: Duckworth. Keren, Einat-Haya. 2015. “From Negative Polarity to Negative Concord—Slavic Footprints in the Diachronic Change of Hebrew meʔuma, klum and šum davar.” Journal of Jewish Languages 3.1-2: 183–198. Kosovsky, Haim Yehoshua. 1954. ʔocar lešon ha-talmud (Concordance of the Talmud) vol. 19, co-ed. Benjamin Kosovsky. Jerusalem: Israeli Ministry of Education in cooperation with the American Rabbinical Midraša (in Hebrew). ———. 1967. ʔocar lešon ha-mišna (Concordance of the Mishnah) vol. 3. Tel Aviv: Massada (in Hebrew). Ma’agarim Database (MD), The Historical Dictionary Project of the Academy of the Hebrew Language. (accessed September 4, 2014). Maschler, Yael. 2009. Metalanguage in Interaction: Hebrew Discourse Markers. Amsterdam: John Benjamins. Migron, Hagit. 2003. “Ma kore po bixlal? On the Polarity Sensitivity of Hebrew Bixlal.” Talk given at the departmental seminar of the English Department, The Hebrew University of Jerusalem. MILA HaAretz Corpus; Itai, Alon & Shuly Wintner. 2008. “Language Resources for Hebrew.” Language Resources and Evaluation 42.1: 75–98. Neudecker, Hannah. 2013. “Vocalization of Modern Hebrew and Colloquial Pro­ nunciation.” In Encyclopedia of Hebrew Language and Linguistics, eds. Geoffrey Khan et al. Brill Online. Traugott, Elizabeth C. 1995. “The Role of the Development of Discourse Markers in a Theory of Grammaticalization.” Paper presented at the International Conference on Historical Linguistics (ICHL) XII, Manchester 1995, version of 11/1997. http://, accessed July 15, 2014, 1–23.

inheritance and slavic contact in the polysemy of bixlal


——— & Richard B. Dasher. 2002. “The Development of Adverbials with Discourse Marker Function.” In Regularity in Semantic Change, eds. Elizabeth Closs Traugott & Richard B. Dasher. Cambridge: Cambridge University Press, 152–189. Ziv, Yael. 2012. “Kama ‘bixlal’ yeš bixlal?” (“How many bixlals are there, anyway?”). In On the Verso of T. P.: Studies in Modern Hebrew and Jewish Languages: Presented to Ora (Rodrigue) Schwartzwald, eds. Malka Muchnik & Tsvi Sadan. Jerusalem: Karmel, 259–271.

The Expression of Material Constitution in Revival Hebrew Chanan Ariel

The Hebrew University of Jerusalem, Israel [email protected]

Abstract This article deals with several constructions used in Hebrew to express material constitution. In Revival Hebrew, there are three different constructions in parallel use, all of which are also documented in earlier strata of Hebrew. These structures will be examined and compared to the former stages of the language and to their equivalents in contact languages. The article discusses the internal influences on these constructions, examines the effect of contact languages and register on the choice of structure, and also briefly compares the situation in Revival Hebrew with the situation in Contemporary Hebrew.

Keywords contact languages – revival of Hebrew – material constitution – šel-phrase – minphrase – construct state – influence of Judeo-Arabic – influence of Yiddish

Introduction This article is concerned expressions describing material constitution. Typi­ cally, such expressions involve two parts: (a) a part describing the material and (b) a part describing the item constituted of the material, henceforth called the “artifact.”

* I thank the members of the Emergence of Modern Hebrew group in the Mandel Scholion Interdisciplinary Research Center, especially Edit Doron, for their input. I also wish to thank Keren Dubnov for her helpful comments. Support from the Mandel Scholion Interdisciplinary Research Center in the Humanities and Jewish Studies at the Hebrew University of Jerusalem is gratefully acknowledged.

© koninklijke brill nv, leiden, ���6 | doi ��.��63/9789004310896_018

The Expression Of Material Constitution In Revival Hebrew


The Constructions in Revival Hebrew

In order to examine the structures used for the expression of material constitution in Revival Hebrew, I drew data from Hebrew press articles published in Palestine from 1882—when the pioneers of the first Aliyah (immigration) came to the country—through 1948, the eve of the establishment of the State of Israel. The corpus chosen for the examination consisted of phrases involving a specific material, namely, glass (‫ זכוכית‬zexuxit in Hebrew).1 The Hebrew of this period uses three types of phrases to express material constitution, as detailed in Table 1. The first construction, shown in row (a), is a bare Noun Phrase (NP). The other two constructions are Prepositional Phrases (PPs); the one in row (b) contains the preposition šel (‘of’), whereas the one in row (c) contains the preposition min (‘from’, spelled and pronounced mi- when cliticized onto the following noun). The three expressions of the material phrase have a similar distribution. All three can function attributively (as illustrated in column 1), as well as predicatively, as either primary predicates (column 2) or secondary predicates (column 3). In Revival Hebrew, these three constructions are used side by side (see Appendix). The first construction, the bare NP, is the most common in the corpus. An examination of the data for 1935, for example, reveals 76 occurrences of the bare NP construction2 versus 14 occurrences of the šel-PP and 16 of the min-PP. That is, the first construction is about five times more prevalent than each of the others. Fig. 1 presents data on the two less frequent constructions.34 The data show that the šel-PP was in use throughout the Revival Hebrew period, and that, until the late 1920s, it was significantly more common than the min-PP. The latter appears fewer than 10 times in the pre-1925 corpus (i.e., about four times less frequently than the šel-PP). However, after 1930, it became more common, eventually surpassing the šel-PP toward the end of the period. 1 The data was sourced from the Historical Jewish Press site (see References). If the construction appeared several times in the same text, it was counted once. If the same text (an article, article title, or advertisement) appeared several times in different newspapers, it was counted once. I also counted an article title which was cited by other authors as a single occurrence. It should be noted that the website’s search engine relies on OCR technology, which has its deficiencies, but seems to be sufficiently reliable to produce a general picture. All the examples in this article are attested to in the investigated corpus unless explicitly noted otherwise. 2 Only undetermined phrases were counted. 3 Both the attributive and predicative constructions were counted.

230 table 1

Ariel Constructions expressing material constitution in Revival Hebrew (1) attributive function

(2) predicative function (primary predicate)

(3) predicative function (secondary predicate)

(a) bare NP ‫סנדלי זכוכית‬ sandeley sandals.CS zexuxit glass ‘glass sandals’

‫התריסים העשויים זכוכית בית חרושת שקירותיו זכוכית‬

(b) šel-PP

,‫ איזו עין שלי היא של זכוכית‬The construction was not ‫ עין תותבת‬found in the corpus.4 ʔezo ʕain šeli which eye of.mine šel zexuxit hi PRON of glass ʕain totevet artificial eye ‘which eye of mine is of glass, an artificial eye’

‫מגבעת של‬ ‫זכוכית‬

migbaʕat šel hat of zexuxit glass ‘a glass hat’

(c) min-PP ‫רעפים מזכוכית‬ reʕafim shingles mi-zexuxit from-glass ‘glass shingles’

bet-ħarošet factory še-kirotav that-its.walls zexuxit glass ‘a factory whose walls are (made of) glass’

‫גם הגג הוא מזכוכית‬ gam ha-gag hu also the.roof PRON

mi-zexuxit from-glass ‘the roof is also made of glass’

ha-terisim the-shutters ha-ʕasuyim which-are.made zexuxit glass ‘the shutters which are made (of) glass’

‫חלונות האוטו נעשו מזכוכית‬ .‫חסונה ואטומה‬ ħalonot ha-ʔoto windows.CS the-car mi-zexuxit naʕasu were.made from-glass ħasuna va-ʔatuma and-solid bulletproof

‘The windows of the car were made of bulletproof and opaque glass.’

4 This construction is found in Mishnaic Hebrew; see example (13) below.


The Expression Of Material Constitution In Revival Hebrew 50




40 35







20 15 10 5





3 4




5 3



30 21

šel min



18 8

2‒ 1 18 884 85 ‒1 18 889 90 ‒1 18 894 95 ‒1 19 899 00 ‒1 19 904 05 ‒1 19 909 10 ‒1 19 914 15 ‒1 19 919 20 ‒1 19 924 25 ‒1 19 929 30 ‒1 19 934 35 ‒1 19 939 40 ‒1 19 944 45 ‒1 94 8








Frequency of šel-PP and min-PP in Revival Hebrew (1882–1948)

An examination of the context in which the constructions appear yields interesting findings. Register Thirty-eight instances of the min-PP occur in advertisements (as in example 1), versus only two instances of the šel-PP. (1) ‫שולחן מעץ צבע אגוז מכסה מזכוכית‬ šulħan me-ʕec ceva ʔegoz, mixse mi-zexuxit table from-wood color walnut cover from-glass ‘a table of wood, walnut colored, [with] glass cover’ (ad for execution sale; Ha-tzofeh, September 6, 1939) Since advertisements generally use more colloquial language than press articles, one may conclude that the min-PP is characteristic of the spoken register, while the šel-PP is characteristic of a higher register.



Syntactic Factors

Forty-two instances of the šel-PP and 96 instances of the min-PP occur in syntactic structures that compelled the writers to avoid the use of a bare NP, or at least encouraged them to use alternative structures. The following are the main syntactic constraints: i.

Predicative use: Revival Hebrew avoids expressing the material as a bare NP secondary predicate (a second direct object), unlike Biblical Hebrew (see example 7 below). For example: (2) .‫ גם בגדים עושים כבר מזכוכית‬ gam begadim ʕosim kevar mi-zexuxit also clothes they.make already from-glass ‘Even clothes are made of glass these days.’ (Davar, June 4, 1936)


Attributive use: An attributive bare NP can only be the complement of a construct-state (CS) noun. When the artifact whose material constitution is at stake is expressed by a phrase, it cannot function as the head of a construct, and consequently the material must be expressed by a PP. This happens in three types of examples: a) When the artifact is itself expressed by a construct, e.g., batey ħarošet (‘factories’) in the following example: (3) ‫חרושת‬-‫ בתי‬,‫יורק ובפרבריה יתנוססו דירות מזכוכית‬-‫עוד מעט ובחוצות ניו‬ ʕod meʕat u-v-ħucot nu york soon and-in-streets.CS New York u-v-farvareha yitnosesu dirot and-in-its.suburbs apartments mi-zexuxit, batey ħarošet mi-zexuxit from-glass factories from-glass ‘Soon the streets of New York and its suburbs will sport apartments of glass, factories of glass [. . .].’ (Davar, May 11, 1931)


b) When the artifact is expressed by a conjunctive phrase, e.g., kelim va-ħafacim (‘vessels and items’) in the example below:

The Expression Of Material Constitution In Revival Hebrew


(4) ‫כן נמצאו כלים וחפצים יפים מזכוכית ומברונזה‬ ken nimceʔu kelim va-ħafacim yafim also were.found vessels and-items beautiful mi-zexuxit u-mi-bronza from-glass and-from-bronze ‘Also found were beautiful vessels and items of glass and of bronze.’ (Moriah, June 5, 1912) c) When the noun expressing the artifact has a modifier expressed directly after the head, e.g., tavla ketana (‘small tablet’) in example (5):5 (5) ‫וראיתי לנגד עיני בקצה המראה הגדולה שלפני הנהג טבלא קטנה של זכוכית‬ ‫אטומה‬ ve-raʔiti leneged ʕenay bikce ha-marʔa and-I.saw before my.eyes at.the.edge the-mirror ha-gedola še-lifney ha-nahag tavla ketana the-big that-before the-driver tablet small šel zexuxit ʔatuma of glass opaque ‘And I saw before me, at the edge of the large mirror in front of the driver, a small tablet of opaque glass.’ (Do’ar ha-yom, January 20, 1933)

Previous Strata of Hebrew and the Influence of Contact Languages6

Biblical Hebrew In predicate nominal clauses, the material appears as a bare NP primary predicate without a preposition or a definite article, as in: (6)

‫הור‬ ֹ ‫יה זָ הָ ב ָט‬ ָ ‫ּומ ְחּת ֶֹת‬ ַ ‫יה‬ ָ ‫ּומ ְל ָק ֶח‬ ַ

ū-malqāḥɛhā ū-maḥtōtɛhā zāhāḇ ṭāhōr and-its.tongs gold pure ‘and its tongs and fire pans [are] of pure gold’ (Exod. 25:38, NJPS)

5 It should be mentioned that multiple modifiers are rare in Biblical Hebrew and Mishnaic Hebrew. 6 For lack of space, the survey presented here is very brief.



In verbal clauses, the material appears as a bare NP secondary predicate (a second direct object), as in: (7) ‫הור‬ ֹ ‫וַ ּיַ ַעׂש ֶאת ַה ְּמנ ָֹרה זָ הָ ב ָט‬ way-yaʿaś ʾɛṯ ham-mənōrā zāhāḇ ṭāhōr and-he.made ACC the.lampstand gold pure ‘He made the lampstand [of] pure gold.’ (Exod. 37:17, NJPS) According to Joüon & Muraoka (2009: 423, §125v) and others, the verbal clause in (7) is derived from the nominal clause (ham-mənōrā zāhāḇ ṭāhōr, ‘the lampstand—pure gold’)7 by adding an action verb (way-yaʿaś ‘he made’). The predicate in the nominal clause becomes a secondary predicate (a second object) in the verbal clause. When the material is expressed by a definite phrase, it cannot be expressed as a bare NP secondary predicate but only as a PP (min-PP), as in: (8)

‫וַ יַ ֲעׂשּו ָל ֶהם ַמ ֵּס ָכה ִמ ּ ַכ ְס ּ ָפם‬

way-yăʿaśū lāhɛm massēḵā mik-kaspām and-they.made for.them image from-their.silver ‘They have made them molten images [. . .] from their silver.’ (Hos. 13:2, NJPS)

A definite material may be expressed in a bare NP only when it is fronted:8 (9) ‫ּ ַכ ְס ּ ָפם וּזְ הָ בָ ם ָעׂשּו ָל ֶהם ֲע ַצ ִּבים‬ kaspām ū-zhāḇām ʿāśū lāhɛm ʿăṣabbīm their.silver they.made for.them images ‘Of their silver and gold they have made themselves images.’ (Hos. 8:4, NJPS) When the material functions attributively, it is only expressed as a bare NP, as in:

7 The example is unattested. 8 For an additional example, compare Exodus 38:28 (exhibiting the same construction as example 8) with Exodus 39:1 (exhibiting the same construction as example 9).

The Expression Of Material Constitution In Revival Hebrew


(10) ‫הור‬ ֹ ‫ית) ְמנ ַֹרת זָ הָ ב ָט‬ ָ ‫)וְ ָע ִׂש‬ (wə-ʿāśīṯā) mənōraṯ zāhāḇ ṭāhōr (and-you.shall.make) lampstand.CS gold pure ‘(You shall make) a lampstand of pure gold.’ (Exod. 21:31, NJPS) According to Segal (1927:189), there are only two exceptions to this rule in Biblical Hebrew (Lev. 13:48 and Ezra 1:11). Mishnaic Hebrew Mishnaic Hebrew continues to express the material phrase as a bare NP, similarly to Biblical Hebrew, but prefers the šel-PP construction,9 as in: (11) ‫חֹותם ֶׁש ָּלּה ׁ ֶש ּ ַל ּ ַמ ּ ֶתכֶ ת‬ ָ ְ‫ַט ַּב ַעת ׁ ֶש ּ ָלאַ ְלמוּג ו‬ ṭabbaʿat šel-la-ʾalmug we-ḥotam šellah šel-lam-matteḵet ring of-coral10 and-seal her of-metal ‘a ring is of coral and its seal of metal’ (Mishnah, Kelim 13:6, Danby 1933:623) The šel-PP is also used when the material is expressed as a secondary predicate in a clause, as in: (12) ‫] של עופרת וצריך לו—עושהו‬. . .[ .‫שאם היה שלזהב וצריך לו—עושהו שלכסף‬ ‫של עץ‬ še-ʾim haya šel-le-zāhāḇ we-ṣariḵ lo ʿosehu if it.was of-gold and-he.needs it he.makes.him šel-le-ḵesep̄ [. . .] šel ʿop̄ eret we-ṣariḵ lo ʿosehu šel ʿeṣ of-silver [. . .] of lead and-he.needs it of wood ‘If it (the original idol) was made of gold, and one needs it, one makes (a copy of the idol) out of silver [. . .] and if (it was originally made) of lead, and one needs it, one makes (a copy) out of wood.’ (Sifre on Deuteronomy, passage 43, Hammer 1986:93)

9 See Segal 1927:189. On morphological conditioning of the construct and šel-phrase, see Sarfatti 1997:89–97. For syntactic constrains, see Azar 1995:205–206. According to Bendavid 1967:128, the usage of šel is a calque of Aramaic di. 10 It should be noted that the apparent definite form of the determiners in example (10) actually reflects a phonological phenomenon. See Birnbaum 2009.



When the material is definite, it is expressed by a min-PP (as in Biblical Hebrew, see example 8): (13) )‫ ג‬,‫] מאוזני כלים מן השחולת ומן הגרודת (משנה כלים יא‬. . .[ ‫העושה כלים מן העשת‬ ha-ʿose kelim min ha-ʿešet me-ʾozne ḵelim the.maker vessels from the-iron.ore from-ears.CS vessels min haš-šeḥolet u-min hag-gerodet from the-metal.chippings and-from the-filings ‘Articles made from iron ore [. . .] or handles of other vessels, or metal chippings or filings.’ (Mishnah, Kelim 11:3, Danby 1933:620)

Medieval Hebrew as Influenced by Arabic

Judeo-Arabic expresses the material phrase both as a bare NP, which has an equivalent in Classical Arabic, and as a min-PP (even when the artifact is indefinite, as opposed to Classical Hebrew). The existence of the latter influenced the Hebrew that was in contact with Arabic. Below is a comparison between the original commentary on the Mishnah, written in Judeo–Arabic (Cairo, 1168), and the Hebrew translation of R. Yosef Ben Yiṣḥaq Ben Alfavell (Huesca, Aragon, 1297). Arabic Source (R. Qafiḥ’s edition)

(14) ‫ אלה—דרקה מן עוד‬.‫תריס—אלתראס‬ .‫[ וכלאהמא מן עוד‬. . .] ‘tăris’—at-tirās. ‘ʾalla’—daraqatun ‘tăris’ the-shield ‘ʾalla’ shield min ʿūdin [. . .] wa-kilāhumā min ʿūdin from wood [. . .] and-both from wood (Mishnah, Shabbat 6:4)

Hebrew Translation (Napoli press, 1492)

]. . .[ ‫ ואלה צנה מעץ‬.‫תריס הוא המגן‬

‫ושניהם מעץ‬

‘teris’ hu ham-magen. ‘teris’ he the-shield. we-‘alla’ ṣinna me-ʿeṣ [. . .] and-‘ʾalla’ shield from-wood [. . .] u-šnehem me-ʿeṣ and-both from-wood

‘Teris is “shield.” And ʾalla is “wooden shield” [. . .] and both are of wood.’

The Expression Of Material Constitution In Revival Hebrew


According to Blau (1980:8, 158), the replacement of the synthetic construct by a PP reflects the tendency of Judeo-Arabic to prefer analytical structures.11

Rabbinic Hebrew in 19th-Century Europe

Tifʼeret Yiśraʼel (Hanover 1878) is a Hebrew commentary on the Mishnah by R. Israel Lipschutz. In the left column of the table below is a segment from this book dealing with a mishnah (Kelim 9:2) about a clay jar with a metal siphon inside it. In the right column is a Yiddish translation of the same mishnah by R. Joseph Meir Jawitz: R. Israel Lipschutz

(15) ‫ ומינֶ קת‬.‫חַ ִּבית—של חרס של ע"ה‬ ‫בתוכה—העבער ממתכת מונח‬ .‫בחבית‬ xabis—šel ḥeres šel am hoʾorec. jar of clay of ʕam haʔarec u-meynekes besocoh—heber ‘u-meynekes besocoh’ siphon mi-matexes munox ba-xavis from-metal is.laid in-the.jar

R. Joseph Meir Jawitz

‫]) אַ ֶע ְר ֶדענֶע פֿאס פוּן אֵ יין ַעם‬. . .[ ‫( ָחבית‬ ‫ּתֹוכה) אּון ִאין ִדיא‬ ָ ‫ּומנִ ָיקת ְּב‬ ְ ( ]. . .[ ‫הָ אָ ֶרץ‬ ְ ‫עבער פֿוּן ְּב‬ ‫לעך‬ ּ ֶ‫ַפאס ִליגְ ט ֵאיין ה‬

(xovis [. . .]) aerdene fas fun eyn a.clay.ADJ jar from a am hoʾorec [. . .] (u-mnikos betoxo) ʕam haʔarec (and.siphon un in di fas ligt eyn heber fun blex and in the jar is.laid a siphon from metal

‘‫—חבית‬a jar of clay belonging to an ʕam haʔarec; ‫—ומניקת בתוכה‬a siphon of metal is laid inside the jar.’

Tifʼeret Yiśraʼel uses the šel-PP and the bare NP more often than the min-PP, but the last is also well attested. The appearance of the Yiddish word heber in the PP heber mi-matexes strongly suggests that the usage of the min-PP may be a result of Yiddish influence. 11 Other languages that influenced Medieval Hebrew, such as Provençal (as in the case of Rabbi David Qimḥi’s and Rabbi Levi Ben Gershom’s writings), Judeo–Italian (see Ryzhik 2009:446), and Judezmo (cf. Bunis 1999:150), will not be discussed here for lack of space. I thank Prof. Ryzhik and Prof. Bunis for these references.



The Influence of Yiddish and Russian

In Yiddish and Russian, the material may be expressed as a denonimal adjective (see, for instance, a-erd-ene fas—‘a jar of clay’—in example (15) above; erdene is an adjective derived from the noun erd ‘earth’), or by means of prepositions: fun in Yiddish (see eyn heber fun blex ‘a siphon of metal’ in the example above) and из ‘iz’ in Russian. In addition to marking the NP expressing the material, the Yiddish preposition fun has two other functions translated by two different prepositions in Hebrew: (1) šel, expressing possession, as in aerdene fas fun eyn am hoʾorec in example (15) above (‘a jar of clay belonging to an ʕam haʔarec’) and (2) min ‘from’, as in the following example (from Katz 1987: 253): (16) ‫טראז‬ ַ ‫אלקא‬ ַ ‫נטלאפֿן פֿון‬ ָ ‫איך בין ַא‬ ix bin antlofn fun alktaraz I PAST run.away from Alcatraz ‘I escaped from Alcatraz.’ As for the Russian preposition iz, it has only the meaning ‘from’ (and not the possessive meaning). As was demonstrated by Fig. 1 above, in the beginning of the Revival Hebrew period, the šel-PP (common in Mishnaic Hebrew) was more prevalent than the min-PP as a means of expressing material constitution. This may be due to the fact that, during this period, the majority of Hebrew writers were not native speakers of the language, and therefore they adopted the form familiar to them from Mishnaic texts. But as Modern Hebrew became more established as a spoken language, use of the šel-PP declined, whereas the min-PP (which was also found in Medieval Hebrew, and in Classical Hebrew when the material is determined) became increasingly common. This may be due to the influence of Yiddish fun. Perhaps ‘from’ was perceived as the basic meaning of this Yiddish preposition, leading Hebrew speakers to prefer the min-PP over the šel-PP as a translation. The fact that Russian iz has only the meaning ‘from’ (and not the possessive meaning) may have also played a role in this preference. In addition, the distribution of the šel-PP may have been affected by various changes that occurred in the general usage of the preposition šel in Modern Hebrew,12 a discussion of which is beyond the scope of this article. It is of interest to note that the Revival Hebrew corpus yielded no examples of zexuxit (‘glass’) expressed as a denominal adjective. Presumably this 12 For a discussion of changes in the functions of šel in Modern Hebrew, see Rosén 1977: 127–193; Ravid & Shlesinger 1995.

The Expression Of Material Constitution In Revival Hebrew


c­ onstruction is absent from Revival Hebrew because it did not exist in the earlier strata of Hebrew.13

Contemporary Hebrew

A search of NITE-Wiki 2013, conducted in the same manner as the search of the historical Jewish press corpus, shows that trends identified in Revival Hebrew continue and intensify in Contemporary Hebrew. As in Revival Hebrew, the material phrase is most commonly expressed as a bare NP. The šel-PP, whose prevalence began dropping during the Revival Hebrew period, is attested only rarely and seems to be confined to high-register expressions and to collocations such as ‫ לב של אבן‬lev šel ʔeven ‘heart of stone’ or ‫ כפפות של משי‬kefafot šel meši ‘kid gloves,’ while the min-PP occurs far more often (about 200 occurrences). Conclusions It was demonstrated that the dominant form used in Revival Hebrew to express the material phrase for the description of material constitution is a bare NP, also commonly found in Biblical Hebrew. Revival Hebrew uses alternative constructions when syntactic constraints require this, specifically when the material is in secondary predicate function, and when the artifact whose material constitution is at stake is expressed by a phrase rather than a single word. The šel-PP, which is the dominant form in Mishnaic Hebrew, was in use at the beginning of the period, but from 1930 onward its prevalence dropped, and it was gradually replaced by the min-PP. The latter is found in Biblical and Mishnaic Hebrew when the material is determined and in Medieval Hebrew as an alternative to the construct state. It seems that the increased use of the min-PP at the expense of the šel-PP may be due to the influence of the contact languages. Further examination is necessary to determine whether this was also influenced by changes in the usage of šel in Revival Hebrew and in Modern Hebrew. Instances of the material expressed as a denominal adjective (as in Yiddish and Russian) were not found, most likely because this construction did not exist in the earlier strata of Hebrew.

13 Zagorodsky 1939:14 used this construction to distinguish between ‫ אדמת חול‬ʔadmat ħol ‘silicious earth’ and ‫ אדמה חולית‬ʔadama ħolit ‘sandy soil.’



Appendix (1)

‫ אך המענין הוא כי כבר נבנה בית חרושת‬.‫כי אפשר לבנות בתי זכוכית אין כל חידוש בדבר‬ ‫ ולא רק בתים אפשר לבנות‬.‫שקירותיו זכוכית והוא עלה פחות מאשר בנין מחמרים רגילים‬ ‫ הגרה‬,‫ ויש מנבאים שבקרוב תהיה עקרת הבית‬.‫מזכוכית כי אם גם בגדים עושים כבר מזכוכית‬ ‫ והיא‬,‫ ואף המרבדים יהיו זכוכית‬,‫ נעולה סנדלי זכוכית‬,‫בבית זכוכית מלובשה בבגדי זכוכית‬ .‫תבשל בקדרות זכוכית‬ (1936 ‫ ביוני‬4 ,‫)דבר‬ (2) ‫ נעולה‬,‫ומעשה בכלה אחת שנכנסה לחופה כשהיא לבושה בשמלת כלולות עשוייה זכוכית‬ .‫ לראשה מגבעת של זכוכית ובידה ארנק של זכוכית‬,‫זכוכית‬-‫נעלי‬ (1945 ‫ במרץ‬9 ,‫)על המשמר‬

References Azar, Moshe. 1995. The Syntax of Mishnaic Hebrew. Jerusalem: The Academy of the Hebrew Language (in Hebrew). Bendavid, Abba. 1967. Biblical Hebrew and Mishnaic Hebrew. Tel-Aviv: Dvir (in Hebrew). Birnbaum, Gabriel. 2009. “The Vocalization of ‫ ל‬in the Particle -‫ של‬in MS Kaufmann of the Mishna.” Mas’at Aharon: Linguistic Studies Presented to Aron Dotan, eds. Moshe Bar-Asher & Chaim E. Cohen, 63–74. Jerusalem: The Bialik Institute. Blau, Joshua. 1980. A Grammar of Mediaeval Judaeo-Arabic, 2nd edition. Jerusalem: The Magnes Press (in Hebrew). Bunis, David M. 1999. Judezmo: An Introduction to the Language of the Sephardic Jews of the Ottoman Empire. Jerusalem: The Magnes Press (in Hebrew). Danby, Herbert. 1933. The Mishnah: Translated from the Hebrew with Introduction and Brief Explanatory Notes. Oxford: Clarendon. Hammer, Reuven. 1986. Sifre:‎A Tannaitic Commentary on the Book of Deuteronomy, Translated from the Hebrew with Introduction and Notes. New Haven: Yale University Press. HJP Historical Jewish Press. aspx (The search was conducted in the following newspapers: Al Hamishmar, Davar, Do’ar Hayom, Habazeleth, Hapoel Hatzair, Hashkafa, Moria, Hed Hamizrah, HaZvi, HaTzofeh, January 1, 1882–May 14, 1948) (accessed October 1, 2014). HLP Project, National Institute for Testing and Evaluation (NITE), 2013. NITE-Wiki: A Corpus of Articles from Hebrew Wikipedia, .aspx; (accessed October 1, 2014). Jawitz, Joseph Meir. 1878. Mišnayot seder teharot . . . ʕim peruš bi-lšon ʔaškenaz (Mishnah: The Order of Purities . . . with a Yiddish Commentary). Warsaw: Alapin Press (in Hebrew).

The Expression Of Material Constitution In Revival Hebrew


Jewish Publication Society. 1985. Tanakh: the Holy Scriptures: The New Translation according to the Traditional Hebrew Text (NJPS). Philadelphia: Jewish Publication Society. Joüon, Paul & Takamitsu Muraoka. 2009. A Grammar of Biblical Hebrew. 2nd edition. Rome: Pontifical Biblical Institute Press. Katz, Dovid. 1987. Grammar of the Yiddish Language. London: Duckworth. Lipschütz, R. Israel ben Gedaliah. 1830. Mišnayot ʕim peruš . . . tif ʔeret yisraʔel. (Mishnah with Commentary . . . Tif ʔeret Yisraʔel). Hannover: E. A. Telgener (in Hebrew). Qafiħ, R. Yosef. 1963. Mishnah ʕim peruš rabenu moše ben maimon: makor ve-targum (Maimonides’ Commentary on the Mishnah: The Arabic Text with a Hebrew Translation), vol. 2: Seder Moʕed. Jerusalem: Mossad Harav Kook (in Hebrew). Ravid, Dorit & Yitzhak Shlesinger. 1995. “Factors in the Selection of Compound-type in Spoken and Written Hebrew.” Language Sciences 17: 147–179. Rosén, Haim B. 1977. ʕivrit tova (Good Hebrew), 3rd edition. Jerusalem: Kiryat Sefer (in Hebrew). Ryzhik, Michael. 2009. “Fifteenth-century Hebrew Letters from Tuscany,” Mas’at Aharon: Linguistic Studies Presented to Aron Dotan, eds. Moshe Bar-Asher & Chaim E. Cohen. Jerusalem: The Bialik Institute, 442–452 (in Hebrew). Sarfatti, Gad B. 1997. In the Language of My People: Essays on Hebrew. Jerusalem: The Academy of the Hebrew Language (in Hebrew). Segal, Moshe Hirsh. 1927. A Grammar of Mishnaic Hebrew. Oxford: Clarendon. Zagorodsky, Melech. 1939–1941. Milon kol-bo le-ħaklaʔut (Comprehensive Dictionary of Agriculture), 2 vols. Tel-Aviv: Self-published with the assistance of Vaʕad ha-lašon ha-ʕivrit (in Hebrew).

Noun-phrase Structure

What Is New in the NP-Strategy for Expressing Reciprocity in Modern Hebrew and What Are Its Origins? Elitzur A. Bar-Asher Siegal

The Hebrew University of Jerusalem, Israel [email protected]

Abstract This essay focuses on various aspects of the NP-strategies for expressing reciprocity in Modern Hebrew and inquires about their origin. It attempts to determine the exact type of relationship that exists between the contemporary constructions and their equivalents in older periods. It describes a situation in which Modern Hebrew added a new NP-construction to express reciprocity, due to a calque of a construction existing in Indo-European languages. This is an interesting example of the way Modern Hebrew grows richer by incorporating external influence. The new construction did not replace the older one, an inheritance from Mishnaic Hebrew. Instead, it provided a means to distinguish between registers. Despite the semantic and the syntactic resemblance between the new and the old constructions, they remained independent, and they differ in their sociolinguistic distribution, grammatical properties, and semantic nuances.

Keywords reciprocal constructions – Modern Hebrew – Mishnaic Hebrew – calque

* I wish to thank Nora Boneh, Edit Doron, and Yael Reshef for reading and commenting on an early version of this article. This work is supported by European Union grant IRG 030-2227. Since the point of the Modern Hebrew examples in the article is to show the schematic distribution of the reciprocity elements, I am using very simple uncontroversial examples that I have constructed on the basis of my capabilities as a native Modern Hebrew speaker.

© koninklijke brill nv, leiden, ���6 | doi ��.��63/9789004310896_019


Bar-Asher Siegal

Introduction This short essay focuses on various aspects of the NP-strategies for expressing reciprocity in Modern Hebrew and inquires about their origin.1 It relies on my previous work on the constructions of the Semitic NP-strategy for expressing reciprocity and their history (Bar-Asher Siegal 2011, 2012, 2014a, and 2014b; for a definition of these strategies, see 2014b). Modern Hebrew has two major constructions and both consist of two pronominal expressions: The numeral-construction (1): the regular components of this construction are the cardinal number exad, ‘one,’ and a definite form of the ordinal number šeni, ‘second.’ Prima facie, this is an odd combination, consisting of a pair of cardinal and ordinal numbers. The demonstrative-construction (2): this construction consists of a repetition of demonstrative pronouns. (1)

Numeral construction:

‫הילדים שיחקו אחד עם השני‬

ha-yelad-im sixku exad im ha-šeni def-child-pl one.m with def-second.m ‘The kids played with each other.’ (, accessed December 5, 2010) (2) Demonstrative construction: ‫הילדים שיחקו זה עם זה‬

ha-yelad-im sixku ze im ze def-child-pl with dem-sg.m ‘The kids played with each other.’ ( _%E0%E7%F8%E9_%E4%F6%E4%F8%E9%ED, accessed February 16, 2003) These two constructions are used in different registers. The numeralconstruction is restricted in use to the informal, mostly spoken language. Occasionally, it appears in the written language, mostly in the context of 1  In the NP-strategy the encoding of the reciprocity is non-verbal, i.e., verbs in the relevant constructions are transitive (unlike verbal encoding of reciprocity), and it is not morpho­ logically encoded on the verb.

What Is New in the NP-Strategy for Expressing Reciprocity


informal venues, such as the Internet. In contrast, the demonstrativeconstruction is marked for the written language and other contexts where higher register is expected.2 This distribution is probably related to the fact that normativists approve only the demonstrative-construction, since only this construction has a precedent in Mishnaic Hebrew (and a few attestations in the Bible; see Bar-Asher Siegal 2012). The numeral-construction is considered as an innovation of Modern Hebrew,3 and in the following section I will discuss its origin and to what extent it is indeed an innovation. The driving question of this article is, therefore, the following: what can the similarities, both between modern and classical versions of Hebrew and between Modern Hebrew and other languages of the modern period, teach us about the historical connections between Modern Hebrew and these languages regarding the specific phenomenon?

The Origin of the Modern Hebrew Constructions

As is often the case with syntactic constructions, Modern Hebrew inherited the components of the reciprocal demonstrative-construction from Mishnaic Hebrew and not from Biblical Hebrew. This is to be expected, as the demonstrative-construction was the standard NP-strategy to express reciprocity in all the literature written in Hebrew in the Middle Period, when Hebrew ceased to be spoken. As for the peculiarity with the numeral-construction, that its first component is a cardinal number and the second an ordinal number, the origin of this construction seems to be a calque of a construction common in many of the Indo-European languages. The components of the relevant constructions in the Indo-European languages are “one” and “another” (inter alia, einander in German, l’un l’altro in Italian, and yek and din in Kurdish). This calque could have happened as a result of the fact that in various Jewish languages the relevant NP-constructions have these components as well (from ‫נאנדער‬ ַ ‫ ַא‬, an’ander in Yiddish, to xa, “one” and xit, “another” in the Neo-Aramaic of Zakho).

2  The phenomenon that a language with more than one reciprocal construction demonstrates a sociolinguistic specialization in their use is known from other languages as well. Similarly, Kjellmer (1982) and Biber et al. (1999) propose a sociolinguistic distinction between registers in the case of English. 3  Bahat & Ron (1980:177–178);, accessed 6 October 2014.


Bar-Asher Siegal

But how did the European “another” turn into “second” in Hebrew? This is probably due to the fact that the ordinary use of “another/other” is expressed in Hebrew as “second.” For example, if we consider the following sentence, “I met two people, one was tall and the other was short,” then the equivalent sentence in Hebrew, from all periods, would most likely be expressed with “one” and “second.” This can be shown in the following sentences and their English translations: (3) ‫ֵׁשם ָה ֶא ָחד ַּב ֲענָ ה וְ ֵׁשם ַה ֵּׁשנִ י ֵר ָכב‬ šem ha-ʾeḥād baʿăna we- šem ha-šēni name.of def-one.m Baanah and-name def-second.m ‘One was named Baanah and the other Rekab.’ (2 Samuel 4:2)

rēḵāb Rekab

(4) ‫ ומשאמר האחד דעתו אם אמר השני כמותו‬ u-miše-ʾāmar ha-ʾeḥād daʿt-o ʾīm def-one.m cond ʾāmar ha-šēni kmot-o def-second.m ‘As one of them expressed his opinion, if the other one expressed a similar opinion, . . .’ (Geonic Responas, Šaʿarey Ṣedeq, 4:36, from around the 8th century CE) (5) ‫ קנה השני דינרי כסף‬,‫מכי משך האחד דינרי זהב‬ mǝkī māšaḵ ha-ʾeḥād dīnār- ē zāhāb As def-one.m dinar-pl.of gold qānā ha-šēni dīnār-ē kesep̄ def.second.m dinar-pl of silver ‘As soon as the one pulled golden dinars, the other purchased silver dinars.’ (Obadiah ben Abraham of Bertinoro 1440–1510, Mishnah, B. Meṣiʿa 4: 1) Alternatively, or maybe in addition, it is possible to suggest that this construction developed in the vicinity of speakers of Lithuanian, in which one of the relevant NP-constructions consists of the components vienasantrạ.4 In Lithuanian, the word for “other” is antrạ. This word in Lithuanian 4  The historical connection between the numeral-construction and the Indo-European construction consisting of the elements “one-another” was proposed in the past (Baraḳ & Gadish 2008:192). As far as I know, the stronger connection to the Lithuanian elements was not mentioned.

What Is New in the NP-Strategy for Expressing Reciprocity


developed into the expression for the ordinal number “second.” Thus, the construction that consists of vienas-antrạ could have been translated literally either as vienas=one (cardinal number), antrạ=another or as vienas=one (cardinal number), antrạ=second (ordinal number). The calque in Modern Hebrew, accordingly, could be the result of the latter. This hypothesis is supported by the fact that in the database of the Responsa Project, which includes texts of all the rabbinic literature from the 2nd century CE to the contemporary era, the first appearances of the numeral-construction are in citations from rabbis of the eastern part of what today is Belarus, the area which historically belongs to Lithuania. Here are two examples: (6) ‫ביטל אחד את השני‬ bittēl eḥād et-ha-šēni one.m acc-def-second.M ‘They cancelled each other.’ (Responsa Divrey Malkiel 1:84, Malkiel Tenenbaum, lived in Gardinas at the end of the 19th century) (7) ‫כששני נביאים מכחישים בנבואתם אחד את השני‬ kəše-ššney nəbīʾ-īm makḥīš-īm when-two prophet-pl b-nəbuʾat-ām eḥād et-ha-šēni one.m acc-def-second.m ‘When two prophets contradict each other in their prophecy . . .’ (Hiddušey hagriz 103, Isaak Zeev Soloveitchic, who grew up in Valozhyn at the end of the 19th century)

The Two-unit Construction in Modern Hebrew

An additional dimension is relevant for the classification of constructions within the NP-strategy. Each of the two constructions mentioned above is further subdivided according to whether the subcomponents of the reciprocal expression form one unit within the clause (as was the case in all reciprocal examples given so far), or whether they are distributed over two separate units (Bar-Asher Siegal 2014b): I.

One-unit constructions: constructions with a one-unit expression, which co-refers with another plural NP in the clause and never occupies the non-embedded subject but all other positions, as required by the predicate.

250 II.

Bar-Asher Siegal

Two-unit constructions: constructions with two components, each filling a different argument position of the predicate.5

Akkadian, for example, has both types. The two-unit construction consists of a repetition of ah̬ um ‘brother,’ while the one-unit type contains variants of ah̬ āmiš/ah̬ āiš ‘each other.’ The former was predominant in the earlier dialects (8a), while the latter developed only in Middle Babylonian and Middle Assyrian (8b) (Bar-Asher Siegal 2011). (8) a. Old Akkadian: urkatam ah̬ -um ana ah̬ -im lā afterwards brother-nom to brother-gen neg inappuš ‘Afterwards one will not make a claim against the other.’ (TCL 19 63:45) b. Late Babylonian: ah̬ āmeš Ippalū recp ‘They will compensate each other.’ (Dar 321:29) The components of the numeral-construction, seen (1) above in the one-unit construction, appear also in the two-unit construction (9). While the numeralconstruction in general is affiliated with the lower register, in its two-unit appearance it has the flavor of a higher style, usually of writing. This variant almost never emerges in earlier periods of Hebrew (Bar-Asher Siegal 2012), but it is well attested among other Semitic languages (Bar-Asher Siegal 2011, 2014b).

5  There are, in addition, hybrid constructions, which have some of the characterizations of the one-unit constructions but are still two-unit constructions in their essence. The three types of constructions also represent stages of a diachronic development of such constructions (two-unit construction > hybrid construction > one-unit construction). From a typological point of view, the two Modern Hebrew constructions, as they appear in (1)–(2), belong to the same type of construction. Despite their appearance of fitting into the hybrid construction, Bar-Asher Siegal (2014b) demonstrates that, grammatically speaking, they are in fact a oneunit construction.

What Is New in the NP-Strategy for Expressing Reciprocity

(9) ‫ האחד שיחק עם השני‬- ‫יוסי ודני‬ yosi ve-dani ha-exad Yosi(m) and-Danny(M) def-one.m ha-šeni def-second.m ‘Yosi and Danny played with each other.’


sixek im with

On the other hand, the demonstrative-construction is incompatible with the two-unit construction. I could not locate anywhere in its long history a reciprocal use of the demonstrative-construction in a two-unit type prior to Modern Hebrew. In Modern Hebrew too such examples are extremely rare and sound marginal in acceptability.6 For our purposes it is worth mentioning that in other languages, when both the two-unit and one-unit construction are grammatically possible, it is demonstrable that in an earlier stage of the history of that specific language only the two-unit construction existed. This is the case, for example, in Italian (Vezzosi 2010, Bar-Asher Siegal 2014b), Akkadian (Bar-Asher Siegal 2011), and Arabic (Bar-Asher Siegal 2014b). Clearly, this is not the case in Modern Hebrew, for the simple reason that the components eḥād-šēni ‘one-second’ were never parts of a reciprocal construction in the past. In fact, in the classical texts of the Biblical and Mishnaic corpora, tokens of the two-unit constructions (in other constructions, which are not the demonstrative-construction) appear only rarely. Therefore, it is likely that once the numeral-construction was in use in Modern Hebrew, the two variants were easily accepted by speakers, since, as noted, in various languages that have a similar construction to those of Modern Hebrew, the two variants may occur. This, for example, is the case in the JudeoArabic dialect of Tafilalt:7 (10) a. yaʿku Jacob

u-musi and-Moses

si someone


b. yaʿku u-musi wkkelaw si Jacob and-Moses someone ‘Jacob and Moses fed each other.’

si someone

l-si def-someone

6  It is beyond the scope of the current article to discuss the conditions for when and why it is grammatical and when it is only marginally grammatical. 7  I wish to thank my consultant, Moshe Bar-Asher, for the data.


Bar-Asher Siegal

Thus, as long as the two elements of the reciprocal construction are phonologically separated (for example, in Modern Hebrew it is transparent that originally there were two components eḥād and ha-šēni),8 both the one-unit and two-unit constructions may co-exist. Once speakers have the two constructions in their native language, they could develop in Modern Hebrew a real two-unit construction as well, as a result of an analogy to a similar situation in their native languages. Why did this analogy not take place with the demonstrative-constructions? In the current context we may only speculate, and we can think of two reasons: 1) history/register: as noted, the demonstrative-constructions appear in high register, and this register imitates what is found in the old literature. In this literature, the two-unit version does not appear, hence it appears in Modern Hebrew only in the one-unit construction; 2) it might be related to the fact that the components of this construction still function in Modern Hebrew as demonstratives. When they are separated, they function as deictic expressions and therefore are interpreted in a way that each one of the demonstratives refers only to one individual, and not in a reciprocal manner.

Two Additional Notes Concerning the Demonstrative-construction

In the one-unit version of both constructions of Modern Hebrew, the numeral one and the demonstrative one, the pronominal expressions agree in gender with the subject (11a, 12a). When the participants of the reciprocal relation are not of the same gender, then both constructions have two variants (11b, 12b); either the two pronominal expressions are masculine, or one is masculine and the other is feminine: (11) a. ‫יעל ורבקה שיחקו אחת עם השנייה‬ yael ve-rivka sixku axat im ha-šniya Yael(f) and-Rivka(f) one.f with def-second.f ‘Yael and Rivka played with each other.’

8  As opposed to a situation such as, for example, that in the Bavarian dialect of German, where the bipartiteness of a(rà)nand(à), which originated from ein ‘one’ and ander ‘another,’ is not transparent (Plank 2008).

What Is New in the NP-Strategy for Expressing Reciprocity


b. )‫ יוסי ויעל שיחקו אחד עם השני(ה‬ yosi ve-yael sixku exad im Yosi(m) and-Yael(f) one.m with ha-šeni/ha-šniya def-second.m/f ‘Yosi and Yael played with each other.’ (12) a. ‫יעל ורבקה שיחקו זו עם זו‬ yael ve-rivka sixku zo im zo Yael(f) and-Rivka(f) with ‘Yael and Rivka played with each other.’ b. ‫זו‬/‫יוסי ויעל שיחקו זה עם זה‬ yosi ve-yael sixku ze im Yosi(m) and-Yael(f) with ze/zo ‘Yosi and Yael played with each other.’ The reason behind this variation is that, on the one hand, logically speaking, if strong reciprocity is indeed held between the participants (i.e., if each of the participants holds the relation described by the verb with each of the other participants), it would not follow to represent either the masculine or feminine demonstrative alone in a certain position, as each should appear in both positions. On the other hand, the attempt to have agreement causes the change of gender in the pronouns in order to match both members of the set represented by the subject. Normativists consider the change of gender a hyper-correction.9 This conclusion is due to the fact that allegedly, in similar cases in Mishnaic Hebrew, both demonstratives were masculine. In fact, however, this is not entirely true. For, throughout the history of Hebrew literature, we find similar “mistakes,” although it must be admitted that it is impossible to know whether they reflect the original languages or textual corruptions in their transmissions. Let us then consider the following examples:

9  Avinery (1964:161); see also Avshalom Kor, Beofen Miluli: 2011/01/14/%E2%80%9D%D7%96%D7%94-%D7%90%D7%AA-%D7%96%D7%94% E2%80%9D-%D7%90%D7%95-%E2%80%9D%D7%96%D7%94-%D7%90%D7% AA-%D7%96%D7%95%E2%80%9D/, accessed June 23, 2010.


Bar-Asher Siegal

(13) 10‫מזווגן זה לזו‬ mezawwəg-ān ze lā-zo ‘He couples them (a male and a female) with each other.’ (Leviticus Rabbah, Paraša 8, Piska 1) (14) ‫הזוג ישא זה את זו‬ ha-zug yissā ze def-couple ‘The couple will get married.’ (Responsa Harama, Poland 1525)

ʾet acc


Interestingly, perhaps in order to avoid the asymmetry of each option, various authors, writing centuries apart, repeat both directions: (15) ‫כתבו זה לזו וזו לזה‬ kātḇū ze lā-zo ve-zo lā-ze ‘They wrote to each other.’ (Maharam Mintz, Germany 1415) Similarly, the Israeli songwriter Neomi Shemer wrote in the song “‫”בלילה שכזה‬: (16) ‫ בלילה שכזה אהבנו זה את זו וזו את זה‬,‫בלילה שכזה‬ bǝ-laylā še-kkā-ze bǝ-laylā še-kkā-ze in-night in-night ʾāhavnū ze ʾet-zo ve-zo ʾet -ze ‘On an evening such as that, on an evening such as that, we loved each other.’ The second phenomenon is the semantic agreement in the case of plural subjects (17). Only in the case of the demonstrative-construction can the pronominal expression be plural (compare (17a) with (17b)):

10  This is the form in MS London (British Museum 340). MSS Oxford 147 and 2335, however, have two masculine forms of the demonstrative. 

What Is New in the NP-Strategy for Expressing Reciprocity


(17) a. ‫הילדים שיחקו אחד עם השני‬ ha-ylad-im sixku exad im ha-šeni def-boys-pl one.m with def-second.m ‘The boys played with each other.’

b. ‫אלו עם אלו‬/‫הילדים שיחקו זה עם זה‬ ha-ylad-im sixku ze im ze/ def-boys-pl with elu im elu with ‘The boys played with each other.’ The agreement, however, is not morphological but semantic. The target of the agreement is controlled by the actual number of members within each set that participate in the reciprocal relation. With singular demonstratives, example (17b) signifies that among the children various individuals played with each other; and with plural demonstratives, then example (17b) indicates that the children were divided into groups, each group consisting of at least more than one child, and that these groups played with each other (Glinert 1989:69; see Heine & Miyashita 2008:169–170 on the peculiarity of this phenomenon crosslinguistically). This is a regular phenomenon in Modern Hebrew but can also be recognized in Mishnaic Hebrew (Bar-Asher Siegal 2012, 2014b). Should we then consider this similarity between Mishnaic Hebrew and Modern Hebrew to be an example of a successful imitation in Modern Hebrew of the style of Mishnaic Hebrew? And, if so, then how exactly did this imitation occur? If we suppose such an imitation, then it could have happened in one of two ways: 1) the first speakers of Modern Hebrew were exposed to so many examples from Mishnaic Hebrew that they eventually internalized the grammatical rules of the classical form of their language; 2) they learned the rules from grammar books and forced an implementation of the relevant rules. Option 1 is quite unlikely due to the “poverty of stimulus;” there are very few relevant examples in the rabbinic literature, and, even in this corpus, there is a lack of consistency (Bar-Asher Siegal 2012). Option 2 is irrelevant, for the publication of good grammars of Mishnaic Hebrew did not happen before the middle of the 20th century, at which point the numeral-construction had already been in use for several decades. Moreover, even these grammars did not notice the relevant semantic distinction (cf. Segal 1946:63). It is, therefore, more likely that this similarity is not due to historical relation (inheritance or imitation), but rather in both cases the result of the same natural development. When the subject of the reciprocal construction is a plurality, languages usually do not


Bar-Asher Siegal

make the fine-grained distinction as to whether the reciprocal relation holds between singular individuals or between sub-pluralities. This lack of distinction is due to a lack of nominal declension in the relevant reciprocal expressions (consider for example the English “one”-“another”). In fact, very rarely do languages use existing pronouns such as demonstratives to express reciprocity. In Hebrew, on the other hand, the use of demonstrative pronouns allows number declension, and consequently the relevant semantic distinction can be developed. Thus, it is quite possible that the common morphology of Mishnaic Hebrew and Modern Hebrew allowed the independent development of the same distinction, without assuming that the former influenced the latter. Similarly, I believe this is also the case when reciprocity is held between pairs of different gender. It is possible that the tension existing today between the two options (illustrated in (11b) and (12b)), was also present for speakers of previous periods of Hebrew. It therefore seems unnecessary to argue for a historical relation or seek a different explanation.


This article described a situation in which Modern Hebrew added a new construction to express reciprocity, due to a calque of a construction existing in Indo-European languages. The emergence of a new construction, the numeralconstruction, a calque of a common Indo-European construction, did not replace the older construction, the demonstrative-construction, an inheritance from Mishnaic Hebrew. It only provided a means to distinguish between registers. Despite the semantic and the syntactic resemblance between the new and the old constructions, they remained two independent constructions with different grammatical properties and semantic nuances: 1) Only the components of the numeral-construction appear regularly as a two-unit construction. Two demonstratives in a two-unit construction most often are not interpreted as a reciprocal construction; 2) In the case of the demonstrative-construction it is possible to distinguish between cases where reciprocity is held between individuals or between groups denoted by the subject. The numeral-construction, as is the case with most constructions of the NP-strategy, cannot mark such a semantic distinction.

What Is New in the NP-Strategy for Expressing Reciprocity


References Avinery, Isaac. 1964. Yad ha-lashon: otsar leshoni be-seder alef-bet shel ha-nośʾim (Yad ha-lashon: Language Treasures Arranged in Alphabetical Order). Tel Aviv: Yizreʻel (in Hebrew). Bahat, Yaʻacov & Mordekhai Ron. 1980. Ṿe-dayeḳ: tiḳune lashon ve-shipur ha-signon (Be Precise: Language Amendments and Style Improvement. Mahad. 2, meḥuseshet. ed. (in Hebrew). Bar-Asher Siegal, Elitzur A. 2011. “Notes on Reciprocal Constructions in Akkadian in Light of Typological and Historical Considerations.” Semitica et Classica 4: 23–42. ———. 2012. “Diachronic Syntactic Studies in Hebrew Pronominal Reciprocal Constructions.” In Diachrony in Biblical Hebrew, eds. Cynthia Miller & Ziony Zevit. Winona Lake: Eisenbrauns, 209–244. ———. 2014a. “Notes on the History of Reciprocal NP-strategies in Semitic Languages in a Typological Perspective.” Diachronica 31: 337–378. ———. 2014b. “Reciprocal NP-Strategies in Jewish Dialects of Near Eastern NeoAramaic in Light of Parallel Semitic Constructions.” Journal of Jewish Languagues 2(1): 49–77, doi:10.1163/22134638-12340018. Baraḳ, Semadar & Ronit Gadish. 2008. Safah ḳamah: leḳeṭ mi-tokh ha-mador Leshonenu la-ʻam ʻiton ha-Arets 1932–1944 ‫( שפה קמה‬A Rising Language: Selections from the Leshonenu La-am column in Haretz, 1932–1944). Vol. 7, Asupot u-mevoʼot ba-lashon. Jerusalem: ha-Aḳademyah la-lashon ha-ʻIvrit (in Hebrew). Biber, Douglas, Stig Johansson, Geoffrey Leech, Susan Conradand, & Edward Finegan. 1999. Longman Grammar of Spoken and Written English. London: Longman. Glinert, Lewis. 1989. The Grammar of Modern Hebrew. Cambridge: Cambridge University Press. Heine, Bernd & Hiroyuki Miyashita. 2008. “The Intersection between Reflexives and Reciprocals: A Grammaticalization Perspective.” In Reciprocals and Reflexives: Theoretical and Cross-linguistic Explorations, eds. Ekkehard König & Volker Gast. Berlin: Mouton de Gruyter, 169–223. Kjellmer, Goran. 1982. “Each Other and One Another: On the Use of the English Reciprocal Pronouns.” English Studies 63: 231–254. Plank, Frans. 2008. “Thoughts on the Origin, Progress, and Pronominal Status of Reciprocal Forms in Germanic, Occasioned by Those of Bavarian.” In Reciprocals and Reflexives: Theoretical and Cross-linguistic Explorations, eds. Ekkehard König & Volker Gast. Berlin: Mouton de Gruyter, 347–373. Vezzosi, Letizia. 2010. “Micro-Process of Grammaticalization: The Case of Italian l’un l’altro.” In Grammaticalization: Current Views and Issues, eds. Katerina Stathi, Elke Gehweiler, & Ekkehard König, 343–372. Amsterdam: John Benjamins.

The Evolution of the Structure of Free Relative Clauses in Modern Hebrew: Internal Development and Contact Language Influence Miri Bar-Ziv Levy

The Hebrew University of Jerusalem, Israel [email protected]

Vera Agranovsky

The Hebrew University of Jerusalem, Israel [email protected]

Abstract The article discusses the evolution of the syntax of Free Relative clauses (FRs) in Modern Hebrew, from the beginning of the Revival period in the 1880s until the 1980s. Two different FR constructions are used during this period, one originating in Biblical Hebrew, and the other in Mishnaic Hebrew. The article points to two processes that affected these constructions and that have likely been influenced by the languages with which Modern Hebrew was in contact (Yiddish, Slavic). First, the Mishnaic construction gradually replaced the Biblical one. A factor favoring this process was the affinity of the Mishnaic construction to the structure of FRs in Yiddish and in Slavic. Second, the case marking of the Mishnaic construction (at least in direct object position) underwent a process of differentiation that encoded the semantic distinction between definite and universal FR interpretations. The same semantic distinction is also structurally encoded in the Yiddish/Polish FR construction.

* The authors wish to thank Scholion, the interdisciplinary research center in Humanities and Jewish Studies in which the article was written. We also express our deepest gratitude to Edit Doron for all her help, and to the members of the Research Group on the Emergence of Modern Hebrew, especially to Moshe Taube, Chanan Ariel, and Avigail Tsirkin-Sadan. Our thanks go also to Ivy Sichel, Elitzur Bar-Asher Siegal, Nora Boneh, Aynat Rubinstein, and Velvl Chernin for their contributions.

© koninklijke brill nv, leiden, ���6 | doi ��.��63/9789004310896_020

The Evolution of the Structure of Free Relative Clauses


Keywords Modern Hebrew – free relative clause (FR) – complementizer – Revival period – ­contact language

Introduction This article discusses the evolution of the structure of free relative clauses (FRs) in Modern Hebrew, from the beginning of the Revival period in the 1880s until the 1980s. We chose to focus on FRs in direct object position, which underwent some important processes that will be addressed in this article. In all periods of Hebrew, we find two main complementizers in FRs: ʔašer and še- (both: ‘that’). Below are two examples of sentences with these complementizers from newspapers of the Revival period. (1) .‫ הוא מהר לתקן את אשר עות‬. . . hu miher letaken ʔet ʔašer ʕivet he hurried to.fix acc that he.distorted ‘He was quick to fix what he had distorted.’ (Ha-zfira, August 23, 1914) (2) .‫ הממשלה תעשה מה שתחפוץ מבלי לשאול עוד את פי העם‬. . . ha-memšala taʕase ma še-taħpoc the-government what that-will.want mibli lišʔol ʕod ʔet pi ha-ʕam without to.ask more acc mouth.cs the-people ‘The government will do what it pleases without asking the people.’ (Ha-magid, January 18, 1882) One process that Modern Hebrew FRs underwent during the period under investigation is the gradual replacement of the Biblical complementizer ʔašer with the Mishnaic complementizer še- preceded by the interrogative ma ‘what.’ The second process is the increasing use of the accusative marker ʔet with FRs in direct object position. Below are examples of such FRs with and without ʔet. (3) .‫ לתומו ספר מה ששמע‬. . . le-tumo siper ma še-šamaʕ to-his.naivete he.told what that-he.heard


Bar-Ziv Levy and Agranovsky

‘He naively told what he had heard.’ (Ha-zman, January 13, 1904) (4) .‫ למחרת סיפר את מה שראה למדריך המתנדבים בקיבוץ‬ lemoħorat siper ʔet ma še-raʔa le-madrix he.told acc what that-he.saw to-guide.CS ha-mitnadvim ba-kibuc the-volunteers in.the-Kibbutz ‘The next day, he told what he had seen to the guide of the volunteers in the Kibbutz.’ (Maariv, August 8, 1972) After describing these two processes, we will attempt to identify their internal and external causes and also to characterize the difference between the alternatives in present-day Modern Hebrew. The data for this study was compiled by scanning press articles, sourced from the Historical Jewish Press website, for occurrences of ʔašer and še- in object position FRs. The corpus includes newspapers published between the year 1882, which is considered the beginning of the Hebrew Revival period, and the year 1980. Out of this vast corpus, we sampled all the newspapers published in January, at twenty-year intervals.

The Gradual Replacement of ʔašer with še-

An analysis of the data from the given period reveals that, throughout the development of Modern Hebrew, there has been a significant trend of replacing ʔašer with še-. One of the most prominent differences between ʔašer and še-in FRs is that še- is typically preceded by an interrogative pronoun occupying the specifier position.1 Here we shall only consider the interrogative pronoun ma ‘what,’2 leaving aside the interrogative mi ‘who,’ which can likewise be used in direct object FRs.

1  We assume that FRs headed by še- have the structure argued for by Caponigro (2002), with the wh-phrase in the specifier position of the FR. 2  The specifier ma very rarely occurs with ʔašer, in all periods of Hebrew. It does not appear in the Bible, but it appears in the Dead Sea Scrolls (1Q27(1QMyst)1i4). In Modern Hebrew literature, ma+ʔašer is not used (Rubinstein 1971:65).

The Evolution of the Structure of Free Relative Clauses


Another important difference between ʔašer and še- in direct object FRs is that, in all historical stages of Hebrew, ʔašer is usually preceded by the accusative marker ʔet. Conversely, FRs with še- and the interrogative ma (henceforth ma+še- FRs) almost always appeared without ʔet; this changed in the Modern Hebrew period. In Biblical Hebrew, FRs generally appear with the complementizer ʔašer (Peretz 1967:127). When in object position, they are usually marked with the accusative marker ʔet, as in (5) (though the Bible also features FRs with ʔašer that are not preceded by ʔet; see Peretz 1967:166). (5)

. . .‫ ְׁש ָמר ְלָך אֵ ת א ׁ ֲֶשר ָאנ ִֹכי ְמ ַצּוְ ָך ַהּיֹום‬

šəmor ləḵā ʾēṯ ʾăšɛr ʾānōḵī məṣawwəḵā ha-yyōm keep acc that I today ‘Mark well what I command you this day . . .’ (Exodus 34:11, NJPS)

FRs with ma+še- occur in the Bible, but only in Ecclesiastes, and only twice in direct object position, e.g.:3 (6)

. . .‫אינֶ ּנּו י ֵֹד ַע ַמה ּ ׁשֶּיִ ְהיֶ ה‬-‫י‬ ֵ ‫ ּכ‬ ִ

kī ʾēnɛnnū yōḏēʿa ma šɛ-yihyɛ for he.does.not know what ‘Indeed, he does not know what is to happen . . .’ (Ecclesiastes 8:7, NJPS)

In Mishnaic Hebrew, ma+še- FRs replace ʔašer:4 (7) . . .‫ אף אתה עשה עימנו מה שהבטחתנו‬,‫עשינו מה שגזרת עלינו‬. . . ʿāśīnū ma šɛ-gāzartā ʿālēnū we.did what that-you.decreed ʾap̄ ʾattā ʿăśē ma šɛ-hiḇṭaḥtānū also you do what ‘. . . we have done what thou hast decreed concerning us: do thou also what thou hast promised to us . . .’ (Mishnah, Maaser Sheni 5:13, Danby 1933)

3  In the Bible, there is also a rare use of še- without ma, preceded by ʔet (ʔet še-) (Peretz 1967:166). 4  The Mishnah likewise features a few cases of direct object FRs introduced by ʔet še-, without ma.


Bar-Ziv Levy and Agranovsky

Medieval Hebrew has direct object FRs with both complementizers. For example, Rashi, in his commentary on the Tanakh, uses ʔašer in some cases and ma+še in most cases, e.g.: (8) ‫ ממני תראו את אשר אני עושה ראו ועשו כן‬ mimmɛnnī ṯirʾū- ʾēṯ ʾăšɛr ʾnī ʿōśɛ rəʾū wa-ʿăśū ḵēn you.will.see acc that I do see so ‘Watch me (Judg. 7:17)—You shall look at me and do likewise.’ (9) ‫ אפו היום‬,‫ את אשר תאפו אפו מה שאתם רוצים לאפות בתנור‬ ʾēṯ ʾăšɛr tōʾp̄ ū ʾēp̄ ū- ma šɛ-ʾatɛm acc that you.will.bake bake- what that-you rōṣīm lɛʾɛ̆p̄ōṯ bā-tannūr ʾēp̄ ū ha-yyōm want to-bake in-the-oven bake today ‘Bake what you would bake (Exod. 16:23)—Bake today whatever you wish to bake in the oven.’ Late Rabbinic Hebrew uses both complementizers as well, though there is a preference for še-. As mentioned above, at the beginning of the Revival period a clear preference for ʔašer over še- is observed, but subsequently there is a gradual decline in the use of ʔašer, and še- becomes dominant in the clauses under investigation. In the 1882 data, 87 percent of the FRs contain ʔašer and only 13 percent contain ma+še-. Conversely, in the 1980 data, še- is used in 73 percent of the cases and ʔašer in only 27 percent. This gradual change is shown in Fig. 1, which also indicates that še- superseded ʔašer as the dominant complementizer around 1920.

Figure 1

ʔašer vs. ma+še- in FRs from 1882 to 19805

5  623 occurrences in total.

The Evolution of the Structure of Free Relative Clauses


The preference for the Biblical complementizer in the Revival period is probably due to the particular reverence felt toward Biblical Hebrew during this period and the tendency to adopt certain features of it (Bendavid 1971:253). The tendency to later minimize the use of ʔašer and expand the use of še- in FRs is compatible with the general tendency to replace ʔašer with še- throughout the development of Modern Hebrew. Since in the Classical Hebrew sources the complementizer še- is preceded by ma in FRs, it is not surprising that it appears with ma in Modern Hebrew as well. Another possible explanation for the preference of ma+še- over ʔašer is the influence of the main languages Modern Hebrew was in contact with. The period in which the transition occurred was characterized by the relatively strong influence of Yiddish and Russian. These two languages employ the interrogative pronoun ‘what’ (što in Russian and vos in Yiddish) in FRs. Here is an example in Russian followed by an equivalent example in Yiddish: (10) Ya sdyelal to, što on skazal. I did that what he said ‘I did what he said.’ (11) Ikh hob geton (dos), vos er hot gezogt. I have done this what he had said ‘I did what he said.’ Moving on from the Revival period to more recent Modern Hebrew, we find, based on the data culled from the 1980 newspapers, that both ʔašer and še- are in use; the dominant complementizer is še-, whereas ʔašer is largely confined to higher-register texts—in accordance with its general status as one of the markers of high register in more recent Modern Hebrew.

The Growing Trend of Using ʔet before ma+še- FRs

Alongside the gradual substitution of še- for ʔašer in FRs, another process occurs. In the Revival period, ma+še- FRs in object position were rarely preceded by the accusative marker ʔet, but over the years, ʔet became increasingly common in these clauses (contrast [3] and [4] above). In the Mishnah, ma+še- FRs nearly always appear without ʔet.6 For example:

6  Biblical Hebrew disallows ʔet before ma in general (Joüon & Muraoka 2008:415, 501). In the Mishnah, however, ʔet is commonly used before ma (Azar 1995:61), but not in FRs: out of

264 (12)

Bar-Ziv Levy and Agranovsky

‫ הוי שקוד ללמוד תורה ודע מה שתשיב לאפיקורוס‬

hewey šaqud lilmod tora we-ḏa be diligent torah and-know ma še-tašiḇ lā-ʾapiqoros what that-you.will.respond to-heretic ‘Be alert to study the law and know how to make answer to an unbeliever.’ (Mishnah, Avot 2:14, Danby 1933)

In Hebrew of the post-Mishnaic period, ma+še- FRs without ʔet continue to be dominant, although the variant with ʔet occurs as well, in a minority of cases.7 In Modern Hebrew object ma+še- FRs appear first without ʔet, and over the years a gradual increase in its use is clearly observed. In our data, the variant with ʔet is initially absent: our pre-1920 corpus contains no examples of this variant at all. In our 1920 data, the variant with ʔet accounts for 10 percent of the occurrences, while the variant without ʔet accounts for 90 percent. By 1980, ma+še- FRs with ʔet constitute 50 percent of the cases. This gradual change is shown in Fig. 2. Since the increase in the use of ʔet is consistent, it may be assumed that the trend continued after 1980, and that in current Hebrew the variant with ʔet is even more common.

Figure 2

ma+še- vs. ʔet+ma+še- in FRs in object position from 1882 to 1980

dozens of ma+še- FRs, the Mishnah includes only two examples of ma+še- FRs preceded by ʔet, both of them in the same sentence (Negaim 13). 7  A general survey of ma+še- FRs in Medieval and late Rabbinic Hebrew, through the BarIlan Responsa Project, reveals that, although some FRs are preceded by ʔet, most appear without it.

The Evolution of the Structure of Free Relative Clauses


This data gives rise to two questions. First, why are object ma+še- FRs with ʔet so rare in the Revival period? It seems that during the Revival period the influence of Classical sources was very significant, so these FRs appear without ʔet, just as they do in the Mishnah. Moreover, during the Revival period, there was a general tendency to downplay ʔet, which may have also contributed to this phenomenon (Reshef 2004:181).8 The second question is what caused the later insertion of ʔet in ma+še- FRs. Perhaps the insertion of ʔet stems from the need to mark these FRs for accusative case, as part of a general tendency of Modern Hebrew to become more consistent in marking definite direct objects with ʔet (Reshef 2004:180). But the fact remains that many ma+še- FRs in object position still appear without ʔet in present-day Modern Hebrew. One possible explanation is that the process whereby ʔet is becoming an obligatory marker of ma+še- FRs is not yet complete. However, there seems to be a subtle distinction between ma+še- clauses with and without ʔet. It appears that the clauses without ʔet can be interpreted in two ways: as definite or universal, whereas clauses with ʔet tend to be interpreted as definite. For example: (13) .‫ עשיתי מה שהוא אמר‬ ʕasiti ma še-hu ʔamar I.did what that-he said ‘I did what/whatever he said.’ (14) .‫ עשיתי את מה שהוא אמר‬ ʕasiti ʔet ma še-hu ʔamar I.did ACC what that-he said ‘I did what he said.’ The sentence without ʔet (13) allows two readings: a. I did whatever he said (universal); b. I did a particular thing he said (definite). The choice between the readings can depend on the context in which they appear (Larson 1987) and on the interpretation of ma.9 In the sentence with ʔet (14), the definite reading is much more prominent than the universal one. We can compare

8  Conversely, ʔašer in the Revival period is nearly always preceded by ʔet, though examples without ʔet are found as well. 9  ma can be interpreted as referring to a presupposed referent, in which case it is definite. Alternatively, it can be interpreted as not involving a referent, in which case it is not definite. Concerning the connection between presupposition and definiteness, see Strawson (1950).


Bar-Ziv Levy and Agranovsky

ma+še- FRs to questions introduced by ma ‘what,’ which display a similar distinction (Danon 2008:262).10 The distinction between FRs with and without ʔet may be due to the influence of Yiddish and perhaps also other contact languages, such as Polish. Examples (13) and (14) in Yiddish would be: (15) Ikh hob geton, vos er hot gezogt. I have done what he had said ‘I did what/whatever he said.’ (16) Ikh hob geton dos, vos er hot gezogt. I have done this what he had said ‘I did what he said.’ The pronoun dos is optional. Without it, both readings (definite and universal) are possible, but when it is present, the prominent reading is the definite one. In other words, the reading of the sentence with dos is similar to the reading of the sentence with ʔet, and the readings of the sentence without dos are similar to the readings without ʔet.11 This distinction does not seem to be found in Russian, but it is observed in Polish (Citko 2004:103): (17) Jan špiewa, co Maria špiewa. Jan sings what Maria sings John sings what/whatever Mary sings. (18) Jan špiewa to, co Maria špiewa. Jan sings this what Maria sings John sings/is singing what Mary sings/is singing. 10  ?‫(את) מה החלטת לקנות‬ (ʔet) ma heħlatta liknot (acc) what to buy What did you decide to buy? The question without ʔet is ambiguous: it may ask about the kind of thing the speaker wants to buy (e.g., a car as opposed to a motorcycle) or the particular referent among a given set of alternatives (e.g., the green car with the broken mirror as opposed to the blue car with the dented fender). The question with ʔet expects an answer of the latter type (a particular referent), whereas an answer of the former type (referring to a kind) is less expected. 11  Moshe Taube, Avraham Novershtern, David Roskies, p.c.

The Evolution of the Structure of Free Relative Clauses


A more precise characterization of the distinction between the readings in Modern Hebrew requires further research. Conclusion This article has dealt with two processes that FRs underwent in Modern Hebrew. The first is the gradual replacement of the Biblical complementizer ʔašer, which was dominant in the Revival period, with the Mishnaic complementizer še-, which ultimately became the more prevalent form. The second is a process whereby ma+še- FRs in direct object position increasingly appear with the accusative marker ʔet. We pointed out that the initial preference for the Biblical complementizer ʔašer over the Mishnaic complementizer še- during the Revival period may be due to the great influence of Biblical Hebrew in that period, in contrast to the lesser influence of Mishnaic Hebrew. As for the gradual replacement of ʔašer with še-, we suggested two possible causes. First, it may be seen as a part of the general process of expanding the use of še- at the expense of ʔašer in Modern Hebrew. Since in the Classical Hebrew sources the complementizer še- is preceded by ma in FRs, the same happens in Modern Hebrew as well. Another possible factor is the influence of Russian and Yiddish, which are the main languages Modern Hebrew came in contact with, and in which FRs appear with an interrogative pronoun meaning ‘what.’ Since Hebrew ma comes only with še- and not with ʔašer, this may have contributed to the replacement of the latter with the former. In present-day Hebrew, both ʔašer and še- are in use, with še- as the dominant complementizer and ʔašer as a more marginal variant characteristic of high-register texts. The absence of ʔet before ma+še- FRs in the Revival period may be due to the tendency to emulate the language of the Classical Hebrew sources, in this case Mishnaic Hebrew, which typically did not insert the accusative marker ʔet before FRs in object position. The later insertion of ʔet in these clauses may be part of a general tendency of Modern Hebrew to consistently mark definite direct objects with ʔet. In present-day Hebrew, ma+še- FRs in direct object position can appear either with or without ʔet. We assume that in present-day Hebrew there is a semantic-pragmatic distinction between the two alternatives. The distinction is subtle, but it seems that the clause without ʔet allows two readings—definite and universal—whereas in the clause with ʔet the definite reading is more prominent. The distinction may have emerged under the influence of the contact languages Yiddish and Polish, which exhibit a similar phenomenon.


Bar-Ziv Levy and Agranovsky

References Azar, Moshe. 1995. The Syntax of Mishnaic Hebrew. Jerusalem: The Academy of the Hebrew Language / Haifa University (in Hebrew). Bar-Ilan Responsa Project. (accessed Jan­uary 3, 2015). Bendavid, Abba. 1971. Biblical Hebrew and Mishnaic Hebrew. 2 vols. Tel-Aviv: Dvir (in Hebrew). Caponigro, Ivano. 2002. “Free Relatives as DPs with a Silent D and a CP Complement.” In Proceedings of WECOL 2000, ed. Vida Samiian. Fresno, CA: California State University, 140–150. Citko, Barbara. 2004. “On Headed, Headless, and Light-Headed Relatives.” Natural Language and Linguistic Theory 22.1: 95–126. Danby, Herbert. 1933. The Mishnah: Translated from the Hebrew with Introduction and Brief Explanatory Notes. Oxford: Clarendon. Danon, Gabi. 2008. “The Semantic Content of ‘ʔet’.” In Theoretical Hebrew Linguistics, ed. Galia Hatav. Jerusalem: Magnes, 253–277 (in Hebrew). Historical Jewish Press. (accessed January 3, 2015). Historical Dictionary Project of the Hebrew Language. http://maagarim.hebrew-acad (accessed January 3, 2015). Joüon, Paul & Takamitsu Muraoka. 2008. A Grammar of Biblical Hebrew. 2nd edition. Rome: Pontifical Biblical Institute. Larson, Richard. 1987. “Missing Prepositions and the Analysis of English Free Relative Clauses.” Linguistic Inquiry 19: 239–266. NJPS Tanakh. 1985. The New JPS Translation of the Holy Scriptures According to the Traditional Hebrew Text. Philadelphia: Jewish Publication Society. Peretz, Yitzak. 1967. The Relative Clause. Tel-Aviv: Dvir (in Hebrew). Reshef, Yael. 2004. The Early Hebrew Folksong: A Chapter in the History of Modern Hebrew. Jerusalem: Bialik Institute (in Hebrew). Rubinstein, Eliezer. 1971. The Verb Phrase. Tel-Aviv: Hakibbutz Hameuchad (in Hebrew). Segal, Moshe H. 1927. A Grammar of Mishnaic Hebrew. Oxford: Clarendon. Strawson, Peter F. 1950. “On Referring.” Mind 59.235: 320–344.

The Impact of Contact Languages on the Grammaticalization of the Modern Hebrew Superlative Yael Reshef

The Hebrew University of Jerusalem, Israel [email protected]

Abstract Modern Hebrew grammatical constructions include a tripartite paradigm of degree comparison consisting of the positive adjective, the comparative, and the superlative. Such a paradigm did not exist in classical Hebrew, and the expression of the superlative in both Biblical Hebrew and Rabbinic Hebrew required reference to a comparison class by means of a noun. Based on an examination of textual evidence from the initial phases of the formation of Modern Hebrew, this article traces the emergence of the modern superlative constructions and evaluates the role of contact languages in the process.

Keywords superlative – adjective degree comparison – revival of Hebrew – language contact – grammaticalization

Modern Hebrew possesses two common paradigms to express the degree comparison of adjectives: (1) (2) e.g.,

Positive ADJ ADJ gadol big ‘big’

Comparative yoter (‘more’)+ADJ ADJ+yoter (‘more’) yoter gadol/gadol yoter more big/big more ‘bigger’

Superlative haxi (‘most’)+ADJ ha-ADJ(the-ADJ)+beyoter (‘most’) haxi gadol/ha-gadol beyoter most big/the-big most ‘biggest’

© koninklijke brill nv, leiden, ���6 | doi ��.��63/9789004310896_021



Paradigm (1) is the unmarked option in speech and in casual writing, and paradigm (2) is used in more formal style (Shatil 2014:283–289, 291–292; Glinert 1989:213, 2013:799). From the structural viewpoint, haxi may be used with both adjectives and adverbs (e.g., haxi maher ‘most quickly’), whereas beyoter is restricted to adjectives. Alongside these common constructions, classical superlative constructions such as the Biblical ‫ ְק ֥טֹן ָּב ָנֽיו‬qəṭōn bānāw (small.CS sons.his) ‘his youngest son’ (2 Chr. 21:17) or the Rabbinic ‫ קטנה שבבנות‬qeṭanna šeb-ba-banot (small.F.SG that in-the-girls) ‘the smallest of the girls’ (m.Shabbat 8:4) are occasionally used as well. However, their distribution is sporadic and is restricted to literary and highly formal style (Glinert 1989:214, 2013:799). Syntactically, the classical constructions are less flexible than the modern ones, as they do not allow the degree comparison of an adjective independently of a noun, but require explicit reference to the comparison class. The use of a degree word to express the superlative first emerged in Hebrew in the medieval period due to the contact with Arabic on the one hand and with dialects of Latin origin on the other hand. A vast corpus of translations from Arabic gave rise to an Arabicized Hebrew style, imbued with Arabic features, particularly in the fields of syntax, phraseology, and the lexicon (Hopkins 2013). Since in Arabic the superlative is a distinct category, translators’ attempts to render it in Hebrew gave rise to a variety of non-classical constructions (Goshen-Gottstein 2006:95–96). One of these solutions was the common cross-linguistic strategy of deriving the superlative from the comparative through the use of the definite article (Bobaljik 2012:52–54), namely, ADJ—yoter+ADJ—ha-yoter+ADJ (e.g., mešubaħ ‘of high quality’—yoter mešubaħ ‘of higher quality’—ha-yoter mešubaħ ‘of highest quality’) (GoshenGottstein 2006:96; Simon Hopkins, personal communication). This particular strategy became popular among medieval Hebrew writers in Christian lands as well, and became the main strategy to express the superlative in written Hebrew in the following generations. The popularity of the ha-yoter+ADJ construction among medieval Hebrew writers is best attributed to the approximate correspondence with the parallel paradigm in the contact languages. In Arabic, the use of the definite article is one of several strategies to form the superlative, and in dialects of Latin origin this strategy is the default option (e.g., French grand—plus grand—le plus grand) (compare Shatil 2014:294–296). The emergence of adjective degree words in medieval Hebrew may therefore be explained as contact-induced in two respects: (a) the contact languages raised the need for a specific, fixed construction to express the superlative; and (b) their employment of the definite

the Impact of Contact Languages


article to that end contributed to writers’ preference to use in Hebrew the hayoter+ADJ construction over its alternatives for that end. The modern expression of the superlative with the modifiers beyoter and haxi is relatively new. In the first phases of the formation of Modern Hebrew, the dominant construction was ha-yoter+ADJ (the-more+ADJ), inherited from medieval Hebrew, e.g.: ‫ הלא הוא גירושם הנורא מארץ ספרד‬,‫גירוש היהודים היותר גדול בדברי הימים‬

geruš ha-yehudim ha-yoter gadol be-divrey ha-yamim, expulsion.CS the-Jews the-more big in-history halo hu gerušam ha-nora me-ʔerec sefarad surely is their.explusion the-terrible from-land.CS Spain ‘the biggest expulsion of Jews in history, namely their tremendous expulsion from Spain’ [Sokolov 1882:209] ‫אבל כאן חורבן עם במובן היותר פשוט של המושג הזה‬

ʔaval kan ħurban ʕam ba-muvan but here destruction.CS people in.the-sense ha-yoter pašut šel ha-musag ha-ze the-more simple of the-concept the-this ‘But here is a destruction of a people in the simplest sense of this concept.’ (Brenner, Between Water and Water, 1910:51) Despite its initial dominance, the medieval construction was replaced in the early decades of the 20th century by ha-ADJ+beyoter and haxi+ADJ.1 This diachronic change is briefly examined below.

The Textual Data

The forms haxi (originally: ‘indeed’) and beyoter (originally: ‘very, a lot’) were both inherited into Modern Hebrew from previous linguistic layers. Based on their use as reinforcing elements denoting “high degree,” they were reinterpreted in Modern Hebrew as superlatives denoting “highest degree.” This

1  The construction ha-yoter+adj. did not completely disappear, but it is used in contemporary Hebrew rather marginally as a definite comparative. However, at present no analyses exist of its exact functions and relationship with its alternatives but see Doron & Meir 2015.



transformation is relatively recent. The form beyoter acquired the function of superlative marker in a slow, gradual process from the late 18th century on, and the form haxi emerged in this function quite abruptly about a century later. In the early 20th century it was extensively used in written Hebrew for a short period, but due to prescriptive objection it acquired a colloquial tint. This development paved the way for beyoter to turn into the main superlativeforming device in more formal usage.


The form beyoter originates in Rabbinic Hebrew, where it is used as a degree word (‘a lot’), e.g., ‫ בניך יפין ביותר‬banaix yafin beyoter (sons.yours handsome very) ‘your sons are very handsome’ (Babylonian Talmud, Nedarim 20:2). In subsequent generations it was extensively used in Hebrew texts in that function, in both indefinite and definite constructions (Kaddari 1991:516–517, 541), e.g.: ‫ ועל ההרים הרמים ביותר שם האויר דק‬,‫בעמקים האויר הוא עב ביותר ולא טוב לנשימה‬ ‫מאד‬

ba-ʕamakim ha-ʔavir hu ʕav beyoter ve-lo tov in.the-valleys the-air is dense very and-not good li-nešima, ve-ʕal he-harim ha-ramim beyoter for-breathing and-on the-mountains the-high very šam ha-ʔavir dak meʔod there the-air thin very ‘In the valleys the air is very dense and it is not good for breathing, and on the very high mountains—there the air is very thin.’ (Pinhas Eliyahu Hurvits, The Book of the Covenant, 1797:31)

The acquisition of a superlative meaning by the definite construction is quite natural (cf. Bobaljik 2012:54), and in fact many contexts in which it occurs are ambiguous and tolerate both an interpretation of ‘high degree’ and ‘highest degree’ (Kaddari 1991:542). Cases in which the definite construction fills an unambiguous superlative function start to appear only in late 18th century, at first only sporadically, e.g.:

the Impact of Contact Languages


‫הפשטן ביותר מכל תופשי התורה הוא רש"י ז"ל‬

ha-paštan beyoter mi-kol tofsey ha-tora the literalist-most from-all keepers.CS the-Tora hu raši z.l. is Rashi his memory-blessed ‘The most inclined toward pešaṭ among the interpreters of the Torah is Rashi, blessed be his memory.’ (Pinhas Eliyahu Hurvits, The Book of the Covenant, 1797:18) Throughout the 19th century, occasional examples for unambiguous superlative meaning may be located in the texts, alongside a much wider distribution of the medieval construction ha-yoter+ADJ. At the same time, the traditional reinforcing function of beyoter (‘very’) remained extant in the texts in various syntactic constructions, including the definite construction used for the superlative as well, e.g.: ‫ ושלא לשבת בחדר‬,]. . .[ ‫] דרושה נקיות העור‬. . .[ ‫להשמר מהתקררות הגוף‬ ‫ ועקר הדבר תנועת הגוף באויר צח‬,‫ שלא לישן הרבה‬,‫החם ביותר‬

le-hišamer me-hitkarerut ha-guf [. . .] deruša to-self.guard from-catching.cold.CS the-body needed nekiyut ha-ʕor [. . .], ve-še-lo la-ševet ba-ħeder cleanliness.CS the-skin and-that-not to-sit in.the-room ha-ħam beyoter, še-lo lišon harbe, ve-ʕikar the-warm very that-not to-sleep a.lot and-main.CS ha-davar tenuʕat ha-guf ba-ʔavir caħ the-thing movement.CS the-body in-air clean ‘In order to safeguard the body from catching cold [. . .] the purity of the skin is needed [. . .], and not to sit in a very warm room, not to sleep a lot, and most of all—the movement of the body in fresh air.’ (David Gordon, Ways of Medicine, 1870:114)

The ambiguous nature of the construction seems to stand behind its slow expansion in the language. At the onset of the revival period it is already attested fairly regularly, alongside the medieval construction, e.g.:



(3) ‫המדרש‬-‫ הקרוב ביותר אל בית‬,‫ המובחר‬,‫היותר טוב‬ ha-yoter tov, ha-muvħar,2 ha-karov beyoter the-most good the select the-close most ʔel beyt ha-midraš to house.CS the-Midrash ‘the best, the choicest, the closest to the place of Torah study’ (Yosef Haim Brenner, In the Winter, 1903:408) (4) ‫ שחוקו המר והנמהר ביותר‬,‫הלא זהו שירו היותר חריף של הינה‬ halo zehu širo ha-yoter ħarif šel indeed his.poem the-most sharp of haine, sħoko ha-mar ve-ha-nimhar beyoter Heine the-bitter and-the-hasty most ‘This is indeed Heine’s sharpest poem, his most bitter song.’ (Haim Nahman Bialik, The Hebrew Book, 1913:422) Nevertheless, throughout the first decades of the 20th century the distribution of ha-+ADJ+beyoter as a superlative construction remained restricted, and the medieval construction ha-yoter+ADJ was statistically much more frequent (Reshef forthcoming2). The transformation of ha-+ADJ+beyoter into a major superlative-forming device occurred only from the 1930s on, after the rise and fall of a third construction, haxi+ADJ.


The form haxi originates in Biblical Hebrew. Primarily an interrogative element, in one context its meaning is vague, and it may not be interpreted as an interrogative: ‫( מן השלשה הכי נכבד‬2 Sam. 23:19; literally, of the three haxi important).3 Based on a parallel verse in Chronicles, which uses instead the form ‫[‘ הנו‬he is] indeed’, it is traditionally interpreted in this context as a reinforcing element. This reinforcing function was occasionally employed by writers (alongside the more common interrogative function) in both medieval and early modern texts (Berggrün 1981:16–18), e.g.:

2  This word appears without a degree word because it in itself has a semantic meaning of a superlative. 3  For a detailed discussion of the verse and its meaning see Berggrün 1981:21–26.

the Impact of Contact Languages


(5) ‫ מפאת הנשמה‬,‫ כי הכי נכבד הוא מכל גרם השמיימי‬,‫והכל ברא לכבודו של האדם‬

‫אשר באפו‬

ve-ha-kol bara li-xvodo šel he-ʔadam, and-the-all he.created for.sake-his of the-man ki haxi nixbad hu mi-kol gerem because more important he than-all body ha-šemeymi, mipeʔat ha-nešama ʔašer be-ʔapo the-astronomical due-to the-soul that in-him ‘And he created everything for the sake of man, since he is indeed more important than all astronomical bodies, due to the soul within him.’ (Pinhas Eliyahu Hurvits, The Book of the Covenant, 1797:24)

The employment of haxi as a superlative, first attested in the work of Mendele Mokher Sefarim (1836–1917), was initially exclusive to the quotation of the Biblical expression, e.g.: (6) ‫ ומהם הכי נכבד ספרו הגדול‬,‫] חבר ספרים הרבה בחכמת הטבע‬. . .[ ‫ר' מרדכי‬ "‫"תולדות הדגים‬

rabi mordexay [. . .] ħiber sefarim harbe Rabbi Mordekhai composed books many be-ħoxmat ha-tevaʕ, u-me-hem in-science.CS the-nature and-of-them haxi nixbad sifro ha-gadol ‘toledot most important book.his the-big ‘history.CS ha-dagim’ the-fish.’ ‘Rabbi Mordekhai [. . .] composed many biology books, and the most important of them is his big book The History of Fish.’ (Mendele Mokher Sefarim, A General View on Fish, 1877:491) (7) ‫]הכי נכבדים הם שני ההרגשים האלה‬. . .[ ‫בין ההרגשים הנטועים עמוק עמוק בהנפש‬ beyn ha-hergešim ha-netuʕim ʕamok among the-sentiments the-buried deep ʕamok be-ha-nefeš [. . .] haxi nixbadim hem deep in-the-soul most important are šeney ha-hergešim ha-ʔele two the-sentiments the-those ‘Among the sentiments that are buried deep in the soul [. . .] the two most important sentiments are the following.’ (Naħum Sokolov, Eternal Hatred to the Eternal People, 1882:7)



In isolated examples, this usage was extended to other adjectives, primarily to synonyms of the Biblical adjective, such as ‫ חשוב‬ħašuv ‘important’, ‫נבחר‬ nivħar ‘selected’, or ‫ נעלה‬naʕale ‘supreme,’ e.g.: (8) ‫בין הסבות השונות בדרכי החיים המשחיתות את תואר האדם הכי חשובות הן‬ ‫המחלות‬

beyn ha-sibot ha-šonot be-darxey among the-reasons the-various in-roads.CS ha-ħayim ha-mašħitot ʔet toʔar he-ʔadam the-life that-corrupt ACC appearance.CS the-man haxi ħašuvot hen ha-maħalot most important are the-diseases ‘Among the various factors in the course of life that corrupt the appearance of man, the most important are the diseases.’ (Yisrael Frankel, The Guardian of Health, 1890:20) However, the widespread use of haxi with a variety of adjectives is recorded only from the first decade of the 20th century on, alongside the other superlative constructions, e.g.: (9) ‫פחותה‬-‫על פי החוק של ההתנגדות הכי‬ ʕal pi ha-ħok šel ha-hitnagdut the-law of the-resistance haxi peħuta most meager ‘According to the law of the least strong resistance’ (Yosef Haim Brenner, Beyond the Borders, 1907:113) (10) ‫הכי קרובים לאבא הם המה היותר רחוקים ממנו‬ haxi kerovim le-ʔaba hem hema ha-yoter most close to-father are they the-more reħokim mimenu far from.him ‘The closest to father are those who are the furthest away from him.’ (Yosef Haim Brenner, Michael Kramer, 1911:4) As opposed to the slow expansion of beyoter, haxi was immediately accepted in general use. By the 1920s it turned into the most common construction in written Hebrew, rendering the formerly popular ha-yoter+ADJ obsolete

the Impact of Contact Languages


(Reshef forthcoming1, ch. 3; Reshef forthcoming2). However, its domination was short-lived. Due to its dubious Biblical origin, normative objection to its use as a superlative was repeatedly stated from the early 1920s on (see in detail Ben-Asher 1969:39–41; Perets 1965:166–168). As a result, it started to be viewed as unsuitable for formal usage, and the beyoter construction expanded at its expense in language-conscious usage. However, as colloquial usage was not affected by the normative demands and remained intact, the split between the colloquial and formal constructions found in contemporary usage was created.

Discussion and Conclusions

In a diachronic study of comparative and superlative constructions, Shatil attributes the emergence of the colloquial degree comparison paradigm of Modern Hebrew to the influence of two Romance languages on the linguistic habits of the first generation of speakers: Judeo-Spanish, spoken in the old Yishuv (as the Jewish community in Palestine was called), and French, taught in the Alliance schools (Shatil 2014:298–299, 307–308, 309). According to his thesis, the Biblical form haxi was interpreted as a composite form, headed by the definite article ha-, thus providing speakers with a close equivalent to the superlative-forming syntax used in the Romance languages (e.g., French grand—plus grand—le plus grand) (308). However, this explanation raises serious difficulties. The element ha- in the Biblical form haxi is an interrogative element, and it is doubtful if it was ever interpreted as a definite article. Moreover, the medieval paradigm which was dominant in the initial phases of the revival (namely, ADJ—yoter+ADJ— ha-yoter+ADJ) is in fact structurally closer to the paradigms used in JudeoSpanish and French, and it already includes a definite article. The influence of those languages on Modern Hebrew was relatively marginal, and may not be assumed without solid evidence. In the case of the superlative, the textual evidence does not support this assumption but points in the opposite direction, as haxi was adopted as the main means of expression (at least in writing) in the early 20th century, when the influence of those languages was evidently in decline. Hence, assuming that the contact with Judeo-Spanish and French triggered the emergence of the modern superlative constructions in Hebrew seems dubious. Nevertheless, explicit remarks by revival generation writers hint that the contact situation had an impact in the process. These writers r­epeatedly



­ ention the need felt for a straightforward Hebrew equivalent to the m superlative constructions known from other languages, a challenge that the classical constructions could not meet as they did not offer a modification of the adjective independently of the comparison class (e.g., Schneider 1930:189; Avinery 1964:128). An examination of the contact languages in the initial phases of the revival (Arabic, Judeo-Spanish, Yiddish, German, Russian) indicates, however, that they diverge in the formal structures used to that end. As none of them presents a close morphosyntactic equivalent to the Modern Hebrew constructions, singling out a specific language as the direct source of influence seems unfounded. At the same time, as all contact languages share the existence of the category of the superlative, though in different forms, it seems reasonable to consider the possibility of their functional impact. Such an explanation is compatible with the model proposed by Heine & Kuteva (2005) in a cross-linguistic study of language contact and grammatical change, who claim that “transfer of grammatical structures from one language to another without involving any linguistic form is perhaps more widespread than has previously been thought” (260–261). According to their findings, “speakers [. . .] do not simply imitate grammatical categories” (37), but “tend to develop new structures of grammatical expression, manipulating the linguistic resources available in ways that are most beneficial to them in the situation concerned” (35). In the case of the superlative, such a resource was located within Hebrew in two classical forms, the Biblical haxi and the Rabbinic beyoter. Based on sporadic ambiguous examples that could tolerate a superlative interpretation, these usages were gradually extended to further contexts, and their growing association with this new grammatical function resulted in their eventual grammaticalization as superlative markers (cf. Heine & Kuteva 2005:44–45). The preference for these constructions over the formerly dominant medieval construction may be attributed to the general tendency of Modern Hebrew towards the maintenance of external affinity with classical Hebrew, whereas changes of meaning of the inherited constructions are more easily tolerated (Ben Hayyim 1992; Goldenberg 1996). The medieval superlative constructions, in which the complement (yoter) precedes the head, evidently deviated from the formal inventory of classical usage, in which the opposite word order (namely, head+complement) is consistently employed in all syntactic contexts. By contrast, the transformation of the classical constructions haxi+adj. and ha-ADJ+beyoter into superlative-forming devices provided an apparent external continuity with the classical models.

the Impact of Contact Languages


In the interplay between form and meaning in the formation of the modern superlative constructions, it therefore seems that only the latter should be attributed to language contact. The forms themselves reflect the exhaustion of the latent potential of the inherited Hebrew inventory through grammaticalization processes that imbued existing elements with new functions. References Avineri, Isaac. 1964. yad hallaschon. Tel Aviv: Izre’el (in Hebrew). Ben-Asher, Mordechai. 1969. hitgabšut ha-diqduq ha-normativi (The Formation of Normative Grammar). Tel Aviv: Hakibbutz hameuchad (in Hebrew). Ben-Hayyim, Zeʾev. 1992. “lashon ʿatiqa bi-meṣiʾut x̱ adaša.” In The Struggle for a Language. Jerusalem: The Academy of the Hebrew Language, 36–83 (in Hebrew). Berggrün, Nissan. 1981. “berure lašon.” Lešonenu Laʿam 32: 297–305 (in Hebrew). Bobaljik, Jonathan David. 2012. Universals in Comparative Morphology: Suppletion, Superlatives, and the Structure of Words. Cambridge: The MIT Press. Doron, Edit & Irit Meir (2015). “The Impact of Contact Languages on the Degrammaticalization of the Hebrew Definite Article.” Journal of Jewish Languages 3, 1–2: 283–300. Goshen-Gottstein, Moshe. 2006. Syntax and Vocabulary of Medieval Hebrew. Jerusalem: Ben-Zvi Institute (in Hebrew). Glinert, Lewis. 1989. The Grammar of Modern Hebrew. Cambridge: Cambridge University Press. ———. 2013. “Elative: Modern Hebrew.” In Encyclopedia of Hebrew Language and Linguistics, eds. Geoffrey Khan et al., vol. 1. Leiden: Brill, 798–800. Goldenberg, Gideon. 1996. “ha-ʿivrit ke-lašon šemit x̱ aya.” In Evolution and Revival: Trends in the Development of the Hebrew Language. Jerusalem: Israel Academy of Sciences and Humanities, 148–168 (in Hebrew). Heine, Bernd & Tania Kuteva. 2005. Language Contact and Grammatical Change. Cambridge: Cambridge University Press. Hopkins, Simon. 2013. “Arabic Influence: Medieval Period.” In Encyclopedia of Hebrew Language and Linguistics, eds. Geoffrey Khan et al., vol. 1. Leiden: Brill, 137–143. Kaddari, Menahem Zevi. 1991. Post-Biblical Hebrew Syntax and Semantics. 2 vols. Ramat Gan: Bar-Ilan University (in Hebrew). Peretz, Yizhak. 1965. ivrit ka-halakha. Tel Aviv: Yosef Shreberk. Reshef, Yael. Forthcoming1. Hebrew in the Mandate Period. Jerusalem: The Academy of the Hebrew Language (in Hebrew).



———. Forthcoming2. “Written Hebrew of the Revival Generation as a Distinct Phase in the Evolution of Modern Hebrew.” Journal of Semitic Studies. Shatil, Nimrod. 2014. Developments in Contemporary Hebrew. Jerusalem: The Academy of the Hebrew Language (in Hebrew). Schneider, M. B. 1930. “berur temunot ha-lašon u-keviʿat šimušan.” Lešonenu 2: 184–190 (in Hebrew). Werner, Fritz. 2013. “Adjective.” In Encyclopedia of Hebrew Language and Linguistics, eds. Geoffrey Khan et al., vol. 1. Leiden: Brill, 35–44.

The Impact of Contact Languages on the Degrammaticalization of the Hebrew Definite Article Edit Doron

The Hebrew University of Jerusalem, Israel [email protected]

Irit Meir

The University of Haifa, Israel [email protected]

Abstract The Hebrew article ha- is apparently undergoing a process of degrammaticalization within Modern Hebrew. Its distribution has been changing in a particular direction that is unexpected from the point of view of historical linguistics. Whereas in Classical Hebrew it was found with a limited number of lexical items, it now attaches to a variety of phrases. This change is indicative of a change in its morpho-syntactic category: it is becoming more a clitic than an affix. The morpho-syntactic change is accompanied by a semantic change; its function is to mark the definiteness of the phrase it attaches to, rather than being part of the Classical Hebrew state system. We propose that the change has its roots in a language-internal change that affected the periphrastic genitive construction of Mishnaic Hebrew and was enhanced through several phases of language contact such as the contact of Medieval Hebrew with Arabic and the contact of nineteenth-century Hasidic Hebrew with Yiddish.

Keywords definiteness – emphatic state – construct state – degrammaticalization – language contact

* The first author acknowledges the support of the Israel Science Foundation grant #1157/10, and the fellowship from the Mandel Scholion Interdisciplinary Research Center in the Humanities and Jewish Studies of the Hebrew University.

© koninklijke brill nv, leiden, ���6 | doi ��.��63/9789004310896_022


Doron and meir

Introduction In Classical Hebrew (including Biblical and Rabbinic Hebrew), the article hais an inflectional affix marking the emphatic state of nouns and adjectives. As an inflectional affix, it attaches to these two types of lexical items, not to phrases or even compounds. In colloquial Modern Hebrew, the distribution and semantics of the article is dramatically changed. Though it is still a bound item, it is no longer an affix but is becoming a clitic, with increased autonomy and an expanded distribution. It has changed from a morphological marker of state inflection, which only indirectly contributes to definiteness, into a clitic, which contributes the meaning of definiteness. Details of the change are given in Meir & Doron (2013). The present paper traces the beginnings of this change, which we attribute to a combination of internal change and language contact.

The Origins of Definiteness within the Semitic State System

Nouns and adjectives in the Semitic languages are historically inflected for the category of state, in addition to more familiar inflectional categories such as gender and number. Three different states are distinguished in the Classical Hebrew inflectional system, as described by traditional Hebraists as early as the Renaissance (e.g., Reuchlin 1506; Buxtorf 1651): the absolute state, the construct state (cs), and the emphatic state.1 (1)

Classical Hebrew a. absolute state e.g. śimla ‘’ b. construct state e.g. śimlat- ‘’ c. emphatic state e.g. ha-śśimla ‘’

1  The morphological term emphatic is a Semiticist’s term marking a particular value of the inflectional state of a noun and is unrelated both to the phonological term emphatic in the sense of stressed and to the phonetic term emphatic in the sense of pharyngealized. The term emphatic state is commonly used with respect to the Aramaic -a suffix but for some reason has not been used for the Hebrew ha- prefix in the philological literature. However, the function of both affixes is parallel in the two classical languages, and in general in the Central Semitic languages. We explain below why the Classical Hebrew ha- is best treated as a wordlevel prefix marking state inflection rather than definiteness, which is a phrase-level category. Our approach favors the morphological origin of these Central Semitic affixes as the ProtoSemitic presentative adnominal affix hā/han/hal (Hasselbach 2007; Pat-El 2009), but is also compatible with the view that these are original demonstrative pronouns that underwent a process of grammaticalization (Rubin 2005:65–90 and references cited therein).

The Impact of Contact Languages


A noun inflected in the construct state is a bound form and must be attached to another constituent called the annex (possessor). The construction consisting of the construct-state head and the annex is known as the construct. Semantically, it expresses a variety of relations, typically possession (cf. Doron & Meir 2013 for a partial summary of the vast literature on this topic). The emphatic state is marked by prefixation with the article ha-;2 it forms the basis for definiteness, yet it is not strictly speaking interpreted as definite, since definiteness is a value assigned to phrases, not to lexical items. The tripartite state system survived intact throughout Classical Hebrew. In Modern Hebrew, we find the state system converting into a binary absolute vs. construct opposition, with the article ha- reanalyzed as a phrasal clitic divorced from the state system and marking definiteness. The change is prevalent in colloquial Modern Hebrew, and is spreading through Modern Hebrew as a whole, gradually invading the more formal registers.

The Change in the Distribution of the Article ha-

Consider the following contrasts in the distribution of the article ha-. Classical Hebrew Modern Hebrew3 (colloquial) ‫אבן‬-‫הלוחות‬ (2) a. )‫ יב‬,‫ ֻל ֣חֹת ָה ֶ֗א ֶבן (שמות כד‬b. luḥōṯ hā-ʾɛḇɛn ha-luħot even tablets.cs the-stone the-tablets.cs stone ‘the stone tablets’(Exodus 24:12) ‘the stone tablets’ (3) a. )‫ א‬,‫המינין (ירושלמי חלה א‬-‫ חמשת‬b. ‫החמש מינים‬ ḥamešet ha-mminin ha- ħameš minim five.cs the-kinds the-five kinds ‘the five kinds’ ‘the five kinds’ (Palestinian Talmud, Ḥalla 1:1) (4) a. )‫ יד‬,‫ּטֹובים (יהושע כג‬ ִ֗ ‫ ַה ְּד ָב ִ ֣רים ַה‬b. ‫הדברים טובים‬ ha-ddəḇārim ha-ṭṭōḇīm ha-dvarim tovim the-things the-good the-things good ‘the good things’ (Joshua 23:14) ‘the candies’ 2  We uniformly transcribe the Hebrew article as ha-, which is accurate for Modern Hebrew, though in Classical Hebrew the article includes a consonant that typically assimilates to the following consonant; e.g., ha-śśimla in (1c) would be strictly represented as haś-śimla. 3  All examples in Modern Hebrew are attested examples from the internet.


Doron and meir

(5) a. ‫ משנתו שאינה סדורה‬b. ‫משנתו הלא סדורה‬ )‫ ב‬,‫(בבלי תענית ז‬ mišnat-o še-ʾena sedura mišnat-o ha-lo sdura study-his that-neg orderly doctrine-his the-neg coherent ‘his confused study’ ‘his incoherent doctrine’ (Babylonian Talmud, Taʿanit 7.2) In Classical Hebrew, the article ha- inflects lexical items, e.g. the nouns stone in (2a) and kinds in (3a), and the adjective good in (4a). It does not attach to the full constructs in (2a) and (3a), the attributive construction in (4a), or the phrasal modifier in (5a). Moreover, a noun that it inflects is not necessarily interpreted as definite. In (2a), stone is a predicate that denotes material constitution or the type of objects counted. In (4a), things is in no way definite, it is good things that is definite. The emphatic marking of nouns does not make them definite, but it makes the noun phrase as a whole definite. The corresponding (b) examples are all prevalent in colloquial Modern Hebrew. The article ha- attaches to compounds in (2b) and (4b), to a noun phrase consisting of a noun specified by a numeral in (3b), and to an adjective phrase consisting of a negated adjective in (5b). Semantically, the article marks definiteness (or agreement in definiteness, for adjective phrases). Another aspect of the change in the distribution of the article is its association with a variety of word classes that were not historically inflected for state, such as prepositions (6), adverbs (7), various degree words (8), and infinitival verbs (9). (6) ‫הקבוצה המקומית מהמקום הלפני אחרון בטבלה‬ ha-kvuca ha-mekomit me-ha-makom ha-lifne the-group the-local from-the-place the-before.cs ʔaħaron b-a-tavla last in-the-table ‘the local team from the one-but-last place in the league table’ (, accessed March 29, 2015) (7) ‫הפינה הלפעמים שבועית של בן שש וחצי‬ ha-pina ha-lifʕamim švuʕit šel ben the-sometimes weekly of šeš va-ħeci six and-half ‘the sometimes weekly radio show of a six and a half year old’ ( &id=122441177783528, accessed March 26, 2015)

The Impact of Contact Languages


(8) ?‫מה זה הדבר הממש טעים הזה שיש במטבח‬ ma ze ha-davar ha-mamaš taʕim ha-ze še-yeš what this the-thing the-really tasty the-this that-is b-a-mitbaħ in-the-kitchen ‘What is this really tasty something in the kitchen?’ (H. Tzur, age 18, private conversation, July 25, 2003) (9) .‫ את הלקחת כלום ולעשות ממנו משהו‬,‫ את הייצור‬. . . .‫אני אוהב את זה‬ ʔani ʔohev ʔet ze . . . ʔet ha-yicur, ʔet I love acc this . . . acc the-production acc ha-la-kaħat klum ve-la-ʕasot mi-menu mašehu the-to-take nothing and-to-make from-it something  ‘I love this . . . the production, taking nothing and making it into something.’ (Avirama Golan, The Ravens, 2004:31)

The Direction of Change from Affix to Clitic: Degrammaticalization?

The changes in the distribution of the article ha- affected its morpho-syntactic status. While in Classical Hebrew it was an inflectional affix, in Modern Hebrew it exhibits more clitic properties than affix properties (cf. Zwicky & Pullum 1983): it attaches to phrases, often only cliticizing to the first element of the phrase, and is less choosy regarding the lexical category of its host. Furthermore, it has more systematic semantic interpretation, i.e. definiteness, and in some cases it does not participate in agreement processes, as in example (4b) above (Meir & Doron 2013). Such a direction of change runs counter to the much more widespread process of language change, namely grammaticalization, a term coined by Meillet (1912), which refers to a change from a less grammatical to a more grammatical element. The change in the status of the article goes in the opposite direction: from more grammatical (an affix) to less grammatical (a clitic). The latter type of change has been referred to as de-grammaticalization. Based on criteria developed by Norde (2009, 2010) to identify de-grammaticalization processes, we have argued (Meir & Doron 2013) that the change in the status of ha- in Modern Hebrew is an instance of de-grammaticalization. The article has become less bound to its host. It is no longer part of the category of state, which originally distinguished between the absolute, construct, and emphatic states. In Modern Hebrew, the original system is no longer operative, as is evidenced by the fact that the article can


Doron and meir

attach to a noun in the construct state, as in (2b) above. In the original system, there was no way to doubly inflect the same noun in both the construct state and the emphatic state.

Tracing Back the Origins of the Change

The change from an inflectional prefix to a phrasal clitic thus dramatically modifies the morpho-syntax of Hebrew noun phrases, and their semantic interface. When and how did this change take place? We suggest that the change has its roots in a language-internal change that affected the periphrastic genitive construction of Mishnaic Hebrew and was enhanced through the contact of Medieval Hebrew and Arabic, and the contact of 19th-century Hasidic Hebrew and Yiddish. The change was initiated in particular constructions. One is the periphrastic genitive construction, and another is the construct, including compounds and numeric specifiers; yet another one is the superlative phrase consisting of yoter ‘more’ together with an adjective. These constructions are independent of each other and different in nature. The first two relate to the structure of the noun phrase, and the third to the structure of the adjective phrase. Yet, all three have a similar effect on the distribution of the article. All result in the loosening of its attachment to lexical items and its reanalysis as a proclitic attached to a phrasal constituent. The combined effect of the change within the three constructions gave rise to a much wider change in the morpho-syntactic status of the article in Modern Hebrew, as it spread to other constructions as well. Crucially, though the change originates in a language-internal development within the periphrastic genitive construction, it was facilitated by the change in the superlative construction and in the construct under the influence of contact languages. The Rise of the Periphrastic Genitive Construction In the Biblical genitive construct shown in (2a) above, the head of the construction is in the construct state and must be adjacent to the annex. The definiteness of this construction is determined by the attachment of the article to the annex. Mishnaic Hebrew saw the rise of the periphrastic genitive construction (called ‫ סמיכות פרודה‬smixut pruda in Hebrew), where the head is in the absolute state, and is separated from the annex by the genitive preposition ‫של‬ šel ‘of’ as in (10) below. In the periphrastic genitive construction, the definiteness of the phrase is determined by the attachment of the article to the head, not the annex, unlike in the construct.

The Impact of Contact Languages



)‫ ז‬,‫הזנב של לטאה שהיא מפרכסת (משנה אוהלות א‬ ha-zzanaḇ šel leṭa⁠ʾa še-hi mep̄ arkeset the-tail of lizard that-she twitches ‘the tail of a twitching lizard’ (Mishnah, Ohalot 1:7)

Yet originally, the article was prefixed both to the head and the annex in the definite periphrastic genitive. The annex was originally introduced by the dative prefixal preposition le- ‘to,’ subordinated to the head of the structure by the prefixal complementizer še-. When these two prefixes were attached to an annex already prefixed with the article ha-, a portmanteau prefix was formed: še-ll(e)-a ‘that-to-the.’ The portmanteau prefix šella- ‘’ was spelled as part of the annex, as can still be seen in the Kaufmann manuscript (10th/11th century). In later non-vocalized editions, the prefix šella- lost both its final vowel and its status as a bound morpheme. Thus was born the free preposition šel ‘of,’ and, concomitantly, the article was lost.4 The example in (11a) shows the spelling found in the Kaufmann manuscript, and (11b), the spelling in contemporary printed editions of the Mishnah. The only article remaining in (11b) within the entire genitive structure is the one originally attached to the head noun (‫ הקיתון‬ha-kkiton, ‘the ewer’), now interpreted as marking the definiteness of the entire phrase.5 4  As pointed out to us by Chanan Ariel, in Judean Desert documents of the second century CE, šel already occurs as a free form. We speculate that, unlike the Galilean dialect that underlies Mishnaic Hebrew (Rensdburg 1992), the Judean Hebrew dialect lacked the portmanteau prefix šella- ‘,’ perhaps because its speakers actually pronounced the h- onset of the article ha- and therefore only used a portmanteau prefix šelle- ‘’ that did not include the article. This speculation is supported by the finding (Mor, in press, §2.10.3 and references cited therein) that in Judean Hebrew, unlike in the Mishnah, šel may precede an annex marked by the article ha-. Since the Judean dialect did not develop into subsequent stages of Hebrew, we do not discuss it further. 5  Though examples of the prefix šella- abound in the Kaufmann manuscript (Birnbaum 2008), examples of the prefix šelle- can be found as well, attached, as expected, to an annex lacking the article ha-, whether indefinite (i) or definite (ii): (i) )‫ א‬,‫ ורובו ֶש ְּלאחד (משנה חולין ב‬ (ii) )‫ ח‬,‫מסמר ֶש ְּלאבן השעות (משנה עדויות ג‬ we-rubb-o šelle-ʾeḥad masmer šelle-ʾeḇen ha-ššaʿot and-most-poss.3msg nail the-hours ‘and most of one’ (Mishnah, Hullin 2:1) ‘the style of the sundial’ (Mishnah, Eduyyot 3:8) Both šelle- (in Mishnah, Shekalim 6:1) and šella- (in Mishnah, Kelim 12:5) are found in the Kaufmann manuscript separated from the annex by a line break, which shows that these prefixes were considered all along a separate morpheme.


Doron and meir

(11) a. )‫ ה‬,‫לעולם כהן גדול מקדש ידיו ורגליו מן הקיתון ֶש ַּלּזהב (משנה יומא ד‬ leʿolam kohen gadol meqaddeš yad-av ve-ragl-av always priest high sanctifies hands-his and-feet-his min ha-kkiton šella-zzahaḇ from the-ewer ‘The high priest always sanctifies his hands and feet from the ewer [made] of gold.’ (Mishnah, Yoma 4:5) ‫הקיתון של זהב‬ b. ha-kkiton šel zahaḇ the-ewer of gold ‘the ewer of gold’ The interpretation of the phrase-initial article as marking the definiteness of the phrase as a whole prevailed in subsequent stages of Hebrew. It can be illustrated by the following example from Mendele Mokher Sforim (Maskilic Hebrew). The first occurrence of the noun phrase containing šel in this example is indefinite, with both head and annex indefinite. The second occurrence of the same noun phrase is definite, as is to be expected. Yet definiteness is expressed by a single article, attached to the phrase as a whole: (12)  ‫ ומה גבורתו של היהודי ושל העז‬. . . . ‫ִעזו של יהודי קפצה ואכלה גג של תבן חדש‬ ?‫והגג של תבן‬ iz-o šel yehudi qafca ve-ʔaxla gag šel goat-his of man jumped and-ate roof of teven ħadaš . . . . u-ma gvurat-o šel ha-yehudi straw new and-what bravery-his of the-man ve- šel ha-ʕez ve-ha-gag šel teven and-of the-goat and-the-roof of straw ‘A man’s goat jumped and ate a new roof of straw . . . . And what is the bravery of the man and the goat and the roof of straw?’ (Mendele Mokher Sforim, Travels of Benjamin the Third, chapter 1, 1878) The Change within the Superlative Construction As shown in Reshef (2015), the attachment of the article to a phrase is found in the comparative/ superlative construction of Medieval Hebrew, under the influence of Arabic. Since adjectives in Hebrew do not have a special comparative form, and since the need for such a form was probably felt because it existed in Arabic, a phrase came to be used in Medieval Hebrew for the comparative, where the

The Impact of Contact Languages


adjective is modified by yoter ‘more,’ for example ‫ יותר גדול‬yoter gadol ‘more big’ (Goshen-Gottstein 2006:95–96). Moreover, for the purpose of superlative formation, Hebrew attaches the article to the comparative adjective, and this was extended to the phrasal comparative as well, yielding the superlative phrase ‫ היותר גדול‬ha-yoter gadol ‘the more big,’ interpreted as ‘the biggest.’ The Medieval Hebrew examples in (13) below are taken from two different Hebrew translations of the same phrase from Maimonides’s Treatise on the Art of Logic, ca. 1158, published by Israel Efros in 1938.6 (13a) is the original Arabic phrase, spelled by Maimonides in Arabicized Hebrew script. The comparative adjectives are in the construct state and are interpreted as superlative because of the article/pronoun in the annex. The Hebrew translation in (b) is by Moshe ben Shmuel Ibn Tibbon of Provence, written ca. 1250, and the translation in (c) is by Ahituv ben Isaac of Palermo, known as “Ahituv the physician,” written ca. 1280. Each translator uses both the innovative phrasal and the traditional lexical comparative (though not for the same adjective). Similarly to Arabic, the comparative is interpreted as superlative because of the presence of the article. ֧ ׄ ‫אכמל אלשיאין‬ (13) a. ‫אפצ ׄלהמא‬ ʔakmal ʔaš-šayʔ-ayni ʔafḍal-humā more.perfect.cs the-thing-dual.gen more.honored.cs-3m.dual (Chapter 12, p. 33 of the Efros edition, 1938, New York: American Academy for Jewish Research)  ‫היותר שלם משני הדברים והנכבד משניהם‬ b. ha-yyoter šalem mi-ššene ha-ddeḇarim the-more perfect from-two.cs the-things we-ha-nniḵbad mi-ššene-hem and-the-honored from-both-them (p. 54 of the Efros edition, 1938, New York: American Academy for Jewish Research) c. ‫השלם שבשני הדברים והיותר משובח‬ ha-ššalem še-bbi-šene ha-ddeḇarim we-ha-yyoter the-perfect that-in-two.cs the-things and-the-more mešubbaḥ praised (p. 91 of the Efros edition, 1938, New York: American Academy for Jewish Research) all three: ‘the best and most honored of the two things’ 6  We are grateful to Chanan Ariel for the Medieval Hebrew examples.


Doron and meir

The phrasal attachment of the article was carried on into the Haskalah literature of Early Modern Hebrew, where this construction was the most prevalent way of expressing the superlative (additional examples in Reshef 2015): (14) .‫הצווי הוא המילה היותר קצרה בין מילות הפועל‬ ha-civuy hu ha-mila ha-yoter kcara the-imperative pron the-word the-more short ben milot ha-poʕal among words.cs the-verb ‘The imperative is the shortest form of the verb.’ (Samuel David Luzzatto, introduction to the 1855 edition of Sefer harikma by Jonah Ibn Janaħ) The superlative in Modern Hebrew is phrasal as well. Yet its form has changed. The Modern Hebrew superlative consists of the adjective either preceded or followed by an adverb meaning the most: ‫ הכי‬haxi or ‫ ביותר‬beyoter (Reshef 2015). The construction [ha-yoter adj] is now interpreted compositionally, i.e., as the definite form of the comparative: (15) )16 ‫כמה שאלות למשתמשים היותר צעירים בינינו (עד גיל‬ kama šeʔelot l-a-mištamšim ha-yoter ceʕirim ben-enu a.few question to-the-users the-more young among-us (ʕad gil 16) (till age 16) ‘a few questions to the younger users among us (up to the age of sixteen)’ ( c&t=9028, accessed March 26, 2015) It is possible that the change in the superlative construction paved the way to a broader change in the distribution of the article, namely, the possibility of attaching it to additional adverbials and degree words in the initial position of an adjective phrase, as in (8) above.

The Change within the Construct (Both Compounds and Numeric Constructs) An additional construction in which the change in the distribution of the article was initiated was the construct. Originally, the article was prefixed to the annex of the construct. The change consisted in the attachment of the article to the noun phrase as a whole rather than to the annex, and it occurred both in compounds and in phrasal constructs with numeric specifiers.

The Impact of Contact Languages


As in the case of the superlative, these changes are found in the Medieval Hebrew translations from Arabic (Goshen-Gottstein 2006:88–90; 107–109). The following examples too are from the Ahituv translation of the Treatise on the Art of Logic: (16)

.‫הבעל חיים קודם לאדם בטבע‬ ha-bbaʿal ḥayyim qodem l-a-ʾadam b-a-ṭṭeḇaʿ the-possessor.cs life precedes to-the-man in-the-nature ‘The animal is prior to Man in nature.’ (Chapter 12, p. 91 of the Efros edition)


‫השני הפכים שאין ביניהם אמצעי‬

ha-ššene hap̄aḵim še-ʾen bene-hem ʾemṣaʿi the-two.cs contraries that-neg between-them intermediate ‘the two contraries with no intermediate’ (Chapter 11, p. 91 of the Efros edition)

In Arabic, animal is monomorphemic, rather than compound as in Hebrew, and was therefore preceded in the Arabic original of (16) by the definite article: ʔal-ħayawān. As for noun phrases with numeric specifiers, such as (17), Wright (1896, book 3, §107[d]) mentions the construction in Classical Arabic corresponding to the Hebrew ha-ħamiša kfarim ‘the-five villages’ alongside the construction that corresponds to the Classical Hebrew ḥamešet ha-kefarim ‘five.cs the-villages’ as in (3a). In some cases, where the counted noun is singular, the former construction is obligatory, as it is in Hebrew: ha-tišʕim ʔiš/ * tišʕim ha-ʔiš ‘the-ninety man’ (= the ninety men).7 The attachment of the article to the noun phrase in Medieval Hebrew is not restricted to Arabic translations, but can be found in the Hebrew writings of Abraham Ibn Ezra, David Kimhi (“Radak”), Maimonides, and others: (18) ‫כי החמש אותיות הכפולות הם לסוף המילה‬ ki he-ḥameš ʾotiyyot ha-kkefulot hem le-sop̄ ha-mmilla for the-five letters the-double pron to-end the-word ‘For the five double letters are because of the end of the word.’ (Ibn Ezra, Yesod Diqduq, ca. 1145, 1984 edition by N. Aloni, p. 90)

7  In Arabic ʔal-xamsu qurān and xamsu l-qurā ‘the five villages’; ʔat-tisʕūna rağulān vs. *tisʕūna r-rağulā ‘the ninety man’ (Wright, ibid.)


Doron and meir

(19) ‫ואלו הן הארבעה שומרין‬ we- ʾellu hen ha-ʾarbaʿa šomerin and-these pron the-four bailees ‘and these are the four bailees’ (Maimonides, Mishneh Torah, ca. 1180, Hilkhot Skhirut 1:1) The attachment of the article to the noun phrase is already found in the writings of the Geʾonim (Early Medieval Hebrew), probably influenced by Arabic,8 but is found also in Rashi’s commentary on the Talmud.9 As pointed out by Avineri regarding Rashi’s writings (1985:92), most of these examples are of compounds (what he calls cerufim qevuʕim ‘permanent collocations’), which are probably expressed by single words in French. It thus seems that the change in the distribution of the article goes back at least to the 11th century. It was enhanced in Eastern Europe several centuries later, through language contact with Yiddish. Kahn (2013a, b) notices Hebrew compounds borrowed into Yiddish in Hasidic writings, and then modified for definiteness with the Yiddish definite article attached to the compound (20a). This construction is reflected in the parallel 19th-century Hasidic Hebrew construction (20b), from Kahn (2013b:175). (20) a. ‫( דער ראש ישיבה‬Yiddish) der roš yešive ‫( הראש ישיבה‬Hasidic Hebrew) b. ha-roš yešiḇa the-head.cs Yeshiva both: ‘the head of the Yeshiva’ It is very possible that this construction found its way into early Modern Hebrew.10 Such constructions are cited in descriptions of Modern Hebrew as early as the 1930s (Garbell 1930; Rosén 1957; Berman 1978; and many 8  For example, in a letter written by Elhanan ben Shemarya from the early 11th century, we find ha-ššne triyyim ‘the-two Triyyim (a specific coin)’ (p. 122, Cambridge, University Library, T-S Collection, 13J 16, 11). 9  Rashi lived in Provence in the 11th century, where there was probably no Arabic influence. We thank Ora Schwarzwald for bringing the Rashi examples to our attention. 10  Examples which also survive in Modern Hebrew are Hasidic Hebrew examples in which the article attaches to both the annex and the head, such as ‫ הבעל הבית‬ha-baʕal ha-bayit ‘the-owner.cs the-house’ (the house owner, Yiddish der balabos), ‫ הבית הכנסת‬ha-bet hakneset ‘the-house.cs the-gathering’ (the synagogue), cf. Kahn (2015:62).

The Impact of Contact Languages


others since). It is also possible that lexicalized compounds were perceived by the speakers as a single lexeme (as suggested by Berman 1978:250), perhaps by analogy to blends such as ‫ הכדורגל‬ha-kaduregel ‘the football’ (ball+foot) or ‫ המחזמר‬ha-maħazemer ‘the musical’ (play+music). Moreover, Kahn (2015:140) cites many examples from Hasidic tales from the 1900s of phrasal constructions with numeric specifiers in which the article precedes the noun phrase as a whole rather than inflecting the annex (cf. 17 above). These too reflect Yiddish structure:11 (21) a. ‫( די אכט טעג‬Yiddish) di acht teg ‫( השמונה ימים‬Hasidic Hebrew) b. ha-ššmona yamim the-eight days both: ‘the eight days’ (22) a. ‫( די זיבן מענטשן‬Yiddish) di zibn mentchn b. ‫( השבעה אנשים‬Hasidic Hebrew) ha-ššiḇʕa ʔanašim the-seven men both: ‘the seven people’ It thus appears that contact both with Arabic and with Yiddish triggered similar changes in the Hebrew construct, and that the similar effects from the two languages enhanced the entrenchment of the change in Hebrew in its revival stage. In Modern Hebrew, the construct became a very productive device for creating compounds (Nir 1993; Ornan 2003; Schwarzwald 2001), while the periphrastic genitive construction became the main construction for expressing possessive relations (Rosén 1957; Berman 1978; Schlesinger & Ravid 1998). The attachment of the article to both types of structures continued into Modern Hebrew. It seems, then, that there are (at least) three possible factors contributing to the change of the position of the article: the increased use of the periphrastic genitive for expressing possession, the use of the construct for compounding, and the influence of Arabic and Yiddish. We hypothesize that the reanalysis of the Mishnaic Hebrew periphrastic genitive created a 11  We are grateful to Dov Faust for the Yiddish translations.


Doron and meir

structure in which the article was interpreted as a phrasal clitic, paving the road to the constructions influenced by Yiddish and Arabic. The Change within Noun+Adjective Compounds In compounds consisting of N+A, attachment of the article to the compound as a whole replaces its traditional attachment to both the head and the adjective.12 This was illustrated in (4b) above, and again in (23) here: (23) .‫הכנתי טוסט לאחי ונגמרה הגבינה צהובה‬ hexanti tost le-ʔaħ-i ve-nigmera I-prepared toast to-brother-mine and-got-finished ha-gvina cehuba the-cheese yellow ‘I made a toast for my brother and we ran out of yellow cheese.’ (, accessed March 26, 2015) This is an example of a change that might have been facilitated by the lack of strict agreement in the emphatic marking of nouns and adjectives in Mishnaic Hebrew, but was strongly influenced and enhanced by contact with Yiddish many centuries later, which paved its way into Modern Hebrew. The examples in (24a-b) are from Kahn (2013b:175). (24) a. ‫( דּוא חיות רעות‬Yiddish) di xayes roes ‫( החיות רעות‬Hasidic Hebrew) b. ha-ħayot raʕot the-animals bad both: ‘the wild animals.’

12  Compounds consisting of N+A differ from phrases consisting of N+A in being far less compositional. The meaning of the compound is typically not compositionally constructed from the combination of the attributive meaning of the adjective with that of the noun. The adjective in a compound usually categorizes the head rather than modifying it; for example, ‫‘( גבינה צהובה‬yellow cheese’ in (23)) is not necessarily a cheese that is yellow but rather a particular type of cheese.

The Impact of Contact Languages


Conclusion We have argued that the de-grammaticalization of the Hebrew article was initiated both by internal developments within Hebrew and by constructions that were introduced into Hebrew through contact with other languages, first with Arabic, and then with Yiddish. These constructions include periphrastic genitives, superlatives, compounds, and noun phrases with numeric specifiers. The cliticization of the article to whole phrases within these constructions was present in the language in its revival stage, and it expanded the morphosyntactic environments in which the article could occur. The new environments contributed to the loosening of the bond between the article and its nominal hosts, and to the increase of the article’s independence, characteristic of degrammaticalization processes. The change did not stop in these constructions and spread to other phrasal constituents. The change in the status of the article constitutes an instance of a de-grammaticalization change that was possibly triggered or enhanced by language contact, providing us with the opportunity to study the contribution of language contact to changed grammaticalization. References Avineri, Yitzhak. 1985. Hekhal Rashi, B. Jerusalem: Mosad Harav Kook (in Hebrew). Berman, Ruth A. 1978. Modern Hebrew Structure. Tel Aviv: University Publishing Projects. Birnbaum, Gabriel. 2008. “Vocalization of l within šel in the Mishnah Kaufmann Manuscript.” In Mas’at Aharon: Linguistic Research Presented to Aharon Dotan, eds. M. Bar-Asher & H.A. Cohen. Tel Aviv: Mosad Bialik, 63–74 (in Hebrew). Buxtorf, Johann. 1651. Thesaurus Grammaticus Linguae Sanctae Hebraeae. [reprint forthcoming. Piscataway, NJ: Gorgias Press.] Doron, Edit & Irit Meir. 2013. “Construct State: Modern Hebrew.” In The Encyclopedia of Hebrew Language and Linguistics, Vol. 1, eds. G. Khan et al. Leiden: Brill, 581–589. ———. 2014. “Amount Definites.” Recherches Linguistiques de Vincennes 42: 139–165. Garbell, Irene. 1930. Fremdsprachliche Einflüsse im modernen Hebräisch. PhD Diss., University of Berlin. Goshen-Gottstein, Moshe H. 2006. Syntax and Vocabulary of Medieval Hebrew: Under the Influence of Arabic. Jerusalem: Ben-Zvi Institute, edited and published by Shraga Assif & Uri Melammed on the basis of the unpublished 1951 PhD diss. (in Hebrew). Hasselbach, Rebecca. 2007. “Demonstratives in Semitic.” Journal of the American Oriental Society 127.1: 1–27.


Doron and meir

Kahn, Lily Okalani. 2013a. “Nominal Possessive Constructions in the Early Modern Hasidic Hebrew Tale.” Bulletin of the School of Oriental and African Studies 76: 271–287. ———. 2013b. “Hasidic Hebrew.” In Encyclopedia of Hebrew Language and Linguistics, Vol. 2, eds. Geoffrey Khan et al. Leiden: Brill, 173–176. Meillet, Antoine. 1912. “L’évolution des formes grammaticales.” Scientia (Rivista di Scienza) 12: 384–400. Reprinted 1926 in Antoine Meillet, Linguistique historique et linguistique générale. Paris: Champion. Reprinted 1951, Klincksieck, Paris, 13–148. Meir, Irit & Edit Doron. 2013. “mi-curan netiya le-rexiv taħbiri: ha-šinuy ha-lešoni šel tavit ha-yiduaʕ ba-ʕivrit bat-yamenu” [Degrammaticalization: The Linguistic Change of the Definite Article in Modern Hebrew]. Lešonenu 75.2–3: 317–358 (in Hebrew). Mor, Uri. In press. Judean Hebrew: The Language of the Hebrew Documents from Judea between the First and the Second Revolts. Jerusalem: The Academy of the Hebrew Language (in Hebrew). Norde, Muriel. 2009. Degrammaticalization. Oxford: Oxford University Press. ———. 2010. “Degrammaticalization: Three Common Controversies.” In Gramma­ ticalization: Current Views and Issues, eds. Katerina Stathi, Elke Gehweiler, & Ekkehard König. Amsterdam: John Benjamins, 123–150. Nir, Raphael. 1993. darxey ha-yecira ha-milonit be-ʕivrit bat zmanenu [Word-Formation in Modern Hebrew]. Raanana: Open University Press (in Hebrew). Okalani Kahn, Lily. 2015. A Grammar of the Eastern European Hasidic Hebrew. Leiden: Brill. Ornan, Uzzi. 2003. ha-mila ha-ʔaħarona [The Last Word]. Haifa: Haifa University Press (in Hebrew). Pat-El, Na’ama. 2009. “The Development of the Semitic Definite Article: A Syntactic Approach.” Journal of Semitic Studies 54.1: 19–50. Rendsburg, Gary A. 1992. “The Galilean Background of Mishnaic Hebrew.” In The Galilee in Late Antiquity, ed. Lee I. Levine. New York: The Jewish Theological Seminary of America, 225–240. Reshef, Yael. 2015. “The Impact of Contact Languages on the Degrammaticalization of the Hebrew Definite Article.” Journal of Jewish Languages 3.1–2: 271–282. Reuchlin, Johannes. 1506. De rudimentis hebraicis. Pforzheim. Rosén, Haiim. 1957. ʕivrit tova: ʕiyunim be-taħbir [Good Hebrew: Studies in Syntax]. Jerusalem: Kiryat Sefer (in Hebrew). Rubin, Aaron D. 2005. Studies in Semitic Grammaticalization. Harvard Semitic Series. Winona Lake, IN: Eisenbrauns. Schwarzwald, Ora. 2001. “darxey tecura ve-ħidušey milim ba-ʕivrit be-hebet kamuti” [A Quantitative Study of Word Formation and Innovation in Hebrew]. In Sefer BenZion Fischler, eds. O. Schwarzwald & R. Nir. Tel-Aviv: Rexes, 265–276 (in Hebrew).

The Impact of Contact Languages


Shlesinger, Yitzhak & Dorit Ravid. 1998. “ha-smixut ha-kfula ba-ʕivrit ha-ħadaša: ʕodfut ʔo kiyum ʕacmaʔi?” [Clitic Doubling in the Modern Hebrew Construct: Superfluous or Productive?] Balšanut ʕivrit 43: 85–97 (in Hebrew). Wright, W. 1896. A Grammar of the Arabic Language, 3rd ed., Cambridge: Cambridge University Press. Zwicky, Arnold M. & Geoff K. Pullum. 1983. “Cliticization vs. Inflection: The Case of English n’t.” Language 59.3: 502–513.

Internal Development

The Nature and Diachrony of Hebrew Quality Pseudo-Partitives: Are They a Calque from the Contact Languages? Nimrod Shatil

Zefat Academic College, Zefat, Israel [email protected]

Abstract The article discusses the syntactic, semantic, prosodic, and sociolinguistic features of the contemporary Hebrew construction of the type ‘a beauty of a girl,’ in general N1 of N2, known as quality pseudo-partitive (also as binominal noun-phrase). In this construction, N1 is a nominalized adjective and N2 is the head. Semantically the syntagm is evaluative, either positively or negatively. The article examines the claim that the construction, first documented in 1928, emerged as an internally caused change, and concludes from the evidence that the construction was calqued from contact languages (English, French, German, Yiddish, and Judeo-Spanish).

Keywords quality pseudo-partitives – evaluative impact – agreement – indefiniteness – word order – contact languages – calque

Introduction Phrases like ‫יופי של סרט‬, yófi šel seret, ‘a beauty of a film,’ present an unusual use, called the quality pseudo-partitive, of the Hebrew construction N1 šel N2 ‘N1 of N2.’ The ordinary use of the construction is possessive, e.g., ‫הבית של יעל‬, ha-bayit šel yaʕel, ‘the house of Yael’/‘Yael’s house,’ but it is also known to have other pseudo-partitive uses, e.g., ‫כוס של אורז‬, kos šel ʔorez, ‘cup of rice.’ I will first discuss the properties of the quality pseudo-partitive and then the way it has evolved. All examples that represent actual usage were retrieved from the Google search engine (January–October 2014) unless specified otherwise.

© koninklijke brill nv, leiden, ���6 | doi ��.��63/9789004310896_023



Grammatical Properties

In the quality pseudo-partitive N1 of N2, N2 is the head and N1 the modifier. Accordingly, N2 determines the syntagm’s agreement features: (1)

!‫ הייתה יופי של אספה‬,‫הי‬

hey, hayta yófi šel ʔasefa hey was.3fs beauty.m of gathering.f ‘Hey that was a beauty of a gathering.’

The verb is marked feminine by agreement with the N2 ʔasefa ‘gathering,’ which is feminine, rather than with the N1 yofi ‘beauty,’ which is masculine. Additional arguments for viewing N2 as the head of the construction are given for English by Aarts (1998). In Hebrew, the indefiniteness of the construction is essential. Once a phrase like ‫יופי של רעיון‬, yófi šel raʕayon, ‘a beauty of an idea’ is marked as definite, it loses its quality pseudo-partitive interpretation and is only interpreted as an ordinary possessive phrase, as in (2b). (2) a ‫יופי של רעיון‬ yófi šel raʕayon beauty of idea ‘a beauty of an idea’ b ‫היופי של הרעיון‬ ha-yófi šel ha-raʕayon the-beauty of the-idea ‘the beauty of the idea’ The syntagm under discussion is evaluative, either positively or negatively. The first noun of these phrases exemplifies the two poles. For the positive pole we have items like: ‫יופי של ירקות‬, yófi šel yerakot, ‘a beauty of vegetables;’ ‫חומד של‬ ‫ילד‬, ħómed šel yeled, ‘a charm of a boy,’ etc. For the negative pole we have ‫גועל‬ ‫נפש של התנהגות‬, góʕal nefeš šel hitnahagut, ‘a disgust of a behavior;’ ‫חרא של‬ ‫דירוג‬, ħára šel derug, ‘a shit of a rating;’ ‫פיכס של מזג אוויר‬, fixs šel mezeg ʔavir, ‘a filth of a weather;’ ‫פויה של בן אדם‬, fuya šel ben ʔadam, ‘a disgust of a person’ (compare German pfui). Exclamations such as fixs and fuya serve as heads also in the parallel German phrases, e.g.: ein Hui von Felsenspitzen, ‘a wonder of rock outcrops.’

The Nature and Diachrony of Hebrew Quality Pseudo-Partitives


Prosodically, many examples are stress initial. N1 may have penultimate stress: , ħómed šel baħur, ‘a sweetness of a lad;’ ‫גועל של אנשים‬, góʕal šel ʔanašim, ‘a disgust of people;’ bómba šel saxkan, ‘a bomb of a (football) player.’ Others have monosyllabic N1: ‫כיף של חופשה‬, kef šel ħufša, ‘a delight of vacation;’ and ‫צחוק של הרפתקה‬, cħok šel harpatka, ‘a joke of an adventure.’ The penultimate accent of the noun búba in the phrase ‫ּבּוּבה של קונצרט‬, búba šel concert, ‘a charm of a concert,’ makes it evaluative, whereas the ultimate-stress form bubá or ‛doll’ is denotative. Nevertheless, the construction allows non-penultimate nouns as well, such as in: ‫פצצה של בדיחה‬, pcacá šel bdiħa, ‘a ‘bomb’ of a joke;’ ‫שיגעון של מסיבה‬, šigaʕón šel mesiba, ‘a mania of a party;’ ‫חוויה של טעם‬, ħavayá šel taʕam, ‘a good experience of taste;’ and ‫חלום של חופשה‬, ħalóm šel ħufša, ‘a dream of a vacation;’ ‫ אגדה של‬. . ., ʔagadá šel . . ., ‘a legend of a . . .’ as in commercial names (of country hotel rooms, kindergartens, etc.). Sapan (1966; 1972) treated these constructions as slang expressions. His successors, the lexicographers of Hebrew slang (Ben-Amots & Ben-Yehuda 1972; Rosenthal 2005), did not include any of these constructions in their dictionaries. This shows that at the time of the composition of these latter works, the form was felt to be a part of the linguistic standard and not sub-standard. As Halevy (2001) notes, in the last three decades the construction has become widespread. Entertainment producers, public relations persons, and commercial copywriters are powerful distribution agents of the construction. They are aware of the evaluative impact of the construction and use it in order to entice the public. The construction supplies the title of many shows, TV series, shops, etc., for example: ‫קסם של מתנה‬, késem šel matana, ‘an enchantment of a present;’ and ‫קסם של ניקיון‬, késem šel nikayon, ‘a charm of a cleanliness.’ Further evidence is found in the title of a popular linguistics book with prescriptive leaning, ‫יופי של עברית‬, yófi šel ʕivrit, ‘a beauty of Hebrew.’

The Diachronic Facet

Our aim is not to point out the exact point of birth of the construction within Hebrew, but to discuss the diachronic circumstances that led to its emergence. Concomitantly, we can try to detect its first manifestations. The first documentation that I could find is an advertising announcement of a car in the newspaper Doʾar ha-yom, on July 7, 1928:1 ‫יופי של צורה וצבע‬, yófi šel cura 1  I thank Ofir Zussman for the example.



ve-ceva, ‘a beauty of a form and a color.’ So one can guess that the form was well established in the nineteen twenties. A later example is in the same newspaper from July 17, 1935: ‫זרימה‬-‫יופי של קוי‬, yófi šel kave zrima, ‘a beauty of a streamlines.’ Another documented example appears in A. Shlonsky’s ‘The Adventures of Miki-Mahu:’2 ‫מין ֶח ְרּבֹון של מחזה‬, min ħerbón šel maħaze, ‘such a shit of a drama.’3 Shlonsky’s work was published as a serial in the children’s newspaper Davar li-Yladim from 1933 on. In his 1946 book, Avinery notes childish expressions of nouns “that substitute for adjectives” such as: !‫יופי‬, ‫יופי לי‬, yófi li, yófi, ‛lovely!’ In his 1964 book (written between 1926 and 1963), he notes phrases like: ‫יופי של שמלה‬, yófi šel simla, ‘a beauty of a dress.’ Halevy (2001) concluded from these tokens that the construction emerged over the period 1946–1956 as an internal development, but this is contradicted by the existence of a documented example from 1928. Five years later, Netzer opposed Halevy’s view: “Presumably, this pattern was introduced initially as a literal translation of certain Yiddish expressions and eventually was established in vernacular Hebrew (without relation to Yiddish)” (Netzer 2008:229). Germanic languages abound with examples of the quality pseudo-partitive construction: English: a hell of a woman; a devil of a temper; a charm of a place; a beauty of an actress; a horror of a man; a disgust of an organization. German: Ein Traum von Musik, ‘a dream of music;’ ein Engel von einem Pferd, ‘an angel of a horse;’ eine Schönheit von Frau, ‘a beauty of a woman;’ eine Scheußlichkeit von blauem Kleid, ‘a horror of a blue dress;’ eine Scheußlichkeit von Schiff, ‘a filth of a ship.’ Yiddish: , ‫א מיאוסקייט פון א יינגעלע‬, a miskayt fun a yingele, ‘a filth of a boy;’ ‫א שיינקייט פון פארברענג‬, a šeynkayt fun farbreng, ‘a charm of pleasure;’ ‫ א זיסקייט פון א קאץ‬, a ziskayt fun a kats, ‘a sweetness of a cat;’ ‫א שוידער און‬ ‫א שרעק פון א מעשה‬, a šoyder un a šrek fun a mayse, ‘a shudder and horror of a story.’ 2  The title is certainly a pun on Disney’s Mickey-Mouse. The phrase itself is outdated. Nonetheless, the phrase ħerbón šel maħaze is outdated. An internet search using the Google search engine (performed on Jan. 15, 2015) found 26 examples for the phrase ħára šel hacaga (the same meaning)(22 with the spelling ‫ חרא‬and 4 with the spelling ‫)חרה‬. The reason is presumably the preference for a penultimate head. 3  On page 99 of the work.

The Nature and Diachrony of Hebrew Quality Pseudo-Partitives


I conclude from these examples that contemporary Germanic languages have the quality pseudo-partitive construction in their non-formal registers at least. The same holds of the Romance languages: French: Une guilde, un enfer de management, ‘A guild, a hell of a management;’ une horreur de monde, ‘a horror of a crowd.’ Spanish: un infierno de hotel, ‘a hell of a hotel;’ un asco de vida, ‘a filth of life.’ These examples show that there is no dichotomy between these two groups of European languages but rather a common typology.

The Contribution of Romance Languages to the Formation Process of Contemporary Hebrew

In her article, Halevy (2001) claims that there is no living contact between contemporary Hebrew and the Romance languages, and consequently these languages have taken no part in the formation of the construction under discussion. Her claim is accurate for the general contours of the present linguistic situation but not for the situation existing during the first decades of the revival process, as shown clearly in Schwarzwald (1993).4 The first students learning Hebrew by the direct method were the students of Nissim Bachar and Eliezer Ben-Yehuda in the Ha-tora ve-ha-Mlaxa School that was founded in Jerusalem in 1882. The children of the Sephardic community of Jerusalem attended this school, which was a part of the Alliance Française school network. Many of the newcomers from the Balkan and the Near East countries were graduates of these schools. This system is responsible for the intrusion of French as a cultural language among the Sephardim. Moreover, French had some status in the new settlements that were financed by the Baron de Rothschild’s French administration for 20 years. The first Hebrew school in Rishon Lezion, the Habib School, is situated next to his administration building. We can interpret this fact as a hint to the inhabitants: “Try to learn Hebrew, but remember that French is rewarding.”5 Ornan (2013) claims that students of these schools were the generators of the “War of Languages.”6 To sum up, the facts show that the first generation of speakers 4  See also Shatil 2009. 5  See also Valden 2013. 6  In 1913, the German Jewish Ezra (aid) agency declared that the official language in the first technical academy established in Palestine would be German. This sparked a public



of Hebrew as a dominant communicative language were exposed to Romance languages, both French and Judeo-Spanish. When considering the impact of the Romance languages on contemporary Hebrew, we have to consider some other data as well. According to BenRafael & Ben-Rafael (2011), the estimated number of Israeli citizens with some access to French is 1,500,000, which is 20% of the population and the number of Francophones is 250,000–350,000. These numbers are substantial. Besides, there are still speakers of other Romance languages spoken by newcomers from Latin America, Europe, and Turkey. Going back to our construction, we can add that Judeo-Spanish itself possesses quality pseudo-partitives, for example: ‫מֹוזּורה ֵדי ִאיזָ 'ה‬ ָ ‫ ֵקי ֵא ְיר‬, ke ermozura ְ ‫ ֵקי ֵפ'יאֹור ֵדי‬, ke feor de fostan, ‘what uglide ija, ‘what a beauty of a girl;’ ‫פ'ֹוס ָטאן‬ ness of a dress.’7 Conclusion Reshef (2012; 2013) points out that a great deal of the characteristic traits of spoken Hebrew were born at the very beginning of the so-called “Revival of Hebrew,” even though they were not documented. It is possible that the quality pseudo-partitive began to incubate in the first decades of the revival, but there is no way of proving it. On the other hand, the construction is documented in the writing of graduates of high schools founded at the beginning of the 20 century (e.g., Shlonsky 1947; Yzhar 1958) and in a newspaper edited by Itamar Ben-Avi. We thus have two contradictory opinions of the emergence of the construction. It is either an internal process or a construction influenced by Germanic or Romance contact languages. I argue that even if the quality pseudo-partitive form was born as an inner process, it owes its wide distribution to its presence in most of the languages of contact.

c­ ontroversy, known as the “War of Languages,” between those who supported the use of Hebrew and those who supported the use of German. 7  I thank Prof. Ora Schwarzwald for her examples and explanations.

The Nature and Diachrony of Hebrew Quality Pseudo-Partitives


References Aarts, B. 1998. “Binominal Noun Phrases in English.” Transactions of the Philological Society 96.1: 117–158. Avinery, I. 1946. The Achievements of Modern Hebrew. Tel Aviv: Sifriat Hapoalim (in Hebrew). ———. 1964. Yad hallashon: Lexicon of Linguistic Problems in the Hebrew Language. Tel Aviv: Izreel (in Hebrew). Ben Amots, D. & N. Ben Yehuda. 1972. Universal Dictionary of Colloquial Hebrew. Jerusalem: Levin-Epstein (in Hebrew). Ben-Rafael, E. & M. Ben-Rafael. 2011. “Francophonie in the Plural: The Case of Israel.” Israel Studies in Language and Society 4, 1:39–72. Blank, H. 1956. “Some Yiddish Influences in Israeli Hebrew.” In The Field of Yiddish: Studies in Yiddish Language, Folklore, and Literature, 2, ed. U. Weinreich. New York: Linguistic Circle of New York, Columbia University, 185–201. Halevy, R. 2001. “Functional Changes of šel Phrases in Contemporary Hebrew.” Leshonenu 63, 1–2: 61–80 (in Hebrew). Netzer, N. 2007. Hebrew in Jeans: the Image of Hebrew Slang, Beer-Sheva: Ben-Gurion University Press (in Hebrew). Ornan, U. 2013. In the Beginning Was the Language. Jerusalem: The Academy of the Hebrew Language (in Hebrew). Ravid, D. & A. Baron. 2013. “Semantic Aspects of the Acquisition of the Constructs.” In Studies in Modern Hebrew and the Jewish Languages in Honor of Ora Schwarzwald, eds. M. Muchnik & Z. Sadan. Jerusalem: Carmel, 272–288 (in Hebrew). Reshef, Y. 2013. “The Language that Follows Speech Will Not Be the Same as the One that Preceded It: Spoken Hebrew in the Pre-State Period.” Journal of Jewish Studies 64,1: 157–186. ———. 2012. “How Was Hebrew Spoken? Some Linguistic Traits of the Spoken Hebrew in Its Early Period of Existence.” In The Speech Machine as Language Teacher. “One Speaks Hebrew Here:” Hebrew Voices from Nazi Germany as Evidence of the Living Language and the Life in Palestine during the British Mandate, ed. S. Izreel. Tel Aviv: University Press, 188–211 (in Hebrew). Rosenthal, R. 2005. Dictionary of Israeli Slang. Jerusalem: Keter (in Hebrew). Sapan, R. 1966. Dictionary of Hebrew Slang. Jerusalem: Kiryat Sepher, 2nd edition (in Hebrew). ———. 1971. The Mechanisms of Hebrew Slang. Jerusalem: Kiryat Sepher, 2nd edition (in Hebrew). Schwarzwald (Rodrigue), O. 1993. “Remnants of Judeo-Spanish in Modern Hebrew.” Peʿamim 56: 33–49 (in Hebrew).



Shatil, N. 2009. “The Evolution of Spoken Hebrew Comparative and Superlative Forms.” Ha-ʿivrit we-aḥyoteha, 8–9: 271–290 (in Hebrew). Shlonsky, A. 1947. Mickey Who? Tel Aviv: Sifriyat Hapoʕalim (in Hebrew). Walden, Z. 2013. “On the Place of Kindergarten in the Nursing of Hebrew—Then and Today.” In Language as Culture: New Perspectives on Hebrew, ed. Y. Benziman. Jerusalem: Van Leer; Tel-Aviv: Hakibbutz Hameuchad, 114–144 (in Hebrew). Yizhar, S. 1958. Days of Ziklag. Tel Aviv: ʕAm ʕOved (in Hebrew).

Reconsidering the Emergence of Non-core Dative Constructions in Modern Hebrew Elitzur A. Bar-Asher Siegal and Nora Boneh

The Hebrew University of Jerusalem, Israel [email protected]; [email protected]

Abstract This article critically scrutinizes the perceived view that the emergence of non-core dative constructions in Modern Hebrew is due to a Slavic-Yiddish influence. It studies the Biblical and Mishnaic sources, showing that these language strata contain highly similar constructions to the ones in Modern Hebrew. It additionally shows that parallel constructions existed in languages spoken in the Jewish communities at the time of the revival, revealing that this linguistic phenomenon is typologically widely attested. We therefore claim that this could be an example of an internalization of the old grammar in the new spoken language, enhanced by the fact that similar constructions are reflected in the non-Hebrew native languages of the revival era speakers. These speakers, at the same time, imported into their colloquial Hebrew a sub-type of non-core dative—the discursive dative—to which they could not have been exposed through the ancient written texts, since this type of dative construction occurs only in the spoken language.

Keywords non-core dative – affected dative – ethical dative – discursive dative – co-referential dative – Biblical Hebrew – Mishnaic Hebrew – Modern Hebrew

* We wish to thank Edit Doron, an anonymous reviewer, and the audience at the Departmental Seminar, Department of Hebrew Language, The Hebrew University of Jerusalem. This research is supported by an ISF grant #1366/14 to Nora Boneh and the European Union grant IRG 030-2227 to Elitzur Bar-Asher Siegal. The translation into English of the Biblical Hebrew examples is from The translation into English of the Rabbinic Hebrew examples is E. Bar-Asher Siegal’s.

© koninklijke brill nv, leiden, ���6 | doi ��.��63/9789004310896_024


Bar-Asher Siegal and Boneh

Introduction This article examines the emergence of constructions containing a non-core dative preceded by the dative preposition le- (henceforth, datives, dative constructions), which are frequently used in Modern Hebrew. It focuses on constructions in which a nominal phrase (henceforth DP) that is dative-marked is interpreted as an affected participant, positively or negatively affected by the underlying occurrence (i.e., as being a beneficiary or maleficiary participant or as being endowed possession or deprived of it—possessor dative), as a discursive dative, or as the reflexive co-referential dative, in which the dative-marked DP shares agreement features with the subject of the clause (for discussion of these classifications see Berman 1982; Halevy 2004, 2007, 2013; Dattner 2014; Bar-Asher Siegal & Boneh 2014, 2015a, 2015b).1 An investigation concerning the evolution or disappearance of other types of constructions featuring a dative marked DP are beyond the scope of this short paper; thus, we leave out of the discussion dative marked DPs that are clearly arguments selected by the predicate covering the semantic roles: recipient/goal, experiencer and the agentive dative (cf. Berman 1982) in alternations such as ‫הספר אבד לי‬/‫איבדתי את הספר‬ ʔibadeti ʔet ha-sefer/ʔavad li ha-sefer ‘I lost the book/I got the book lost’2 as they are known to be found in old strata of Hebrew (Gesenius 1910, §119s).3,4 Any inquiry about the origin of a given phenomenon in Modern Hebrew has to answer the following two questions: 1) 2)

Is the phenomenon under discussion attested in earlier Hebrew corpora, and if it is, are its distributional and interpretative properties similar to that of Modern Hebrew? Is it possible to identify a language, or several languages in the same geographical area, spoken by Jewish communities prior to the revival of

1  Previous literature also features the category of the ethical dative. Some of the examples from the literature that fall within this category are included as cases of affected dative (when there is a psychological effect), and some are characterized as discursive dative in the current typology (see Bar-Asher Siegal & Boneh 2015b for the relationship between the various classifications of the non-core datives). 2 (i) a. ʔibadeti ʔet ha-sefer Lost acc the-book ‘I lost the book.’ b. ha-sefer ʔavad l-i The-book lost ‘I got the book lost.’ 3  See 1 Samuel 9:20, Minor Tractates, Semachot 6:11, and Babylonian Talmud 8a. 4  See also Breuer 1987:132–134.

Reconsidering the Emergence of Non-core Dative Constructions


Modern Hebrew that is unique in instantiating the phenomenon under consideration, and which at the same time is not attested in Jewish languages that are unrelated geographically and genetically? When answering the first question negatively and the second question positively, it is reasonable to assume, from a diachronic point of view, that the language which is mentioned in the answer to the second question is the origin of the phenomenon under discussion in Modern Hebrew. Answering the first question positively, however, does not necessarily reject the possibility that the existence of a similar phenomenon in other contemporary languages contributed to the re-emergence of that phenomenon in Modern Hebrew. It is still possible, and perhaps even likely, that the internalization of the old grammar in the new spoken language has been enhanced by the fact that it was reflected in the non-Hebrew native language of the revival era speakers as well. We will show here that, for the most part, the first question should be answered positively and the second negatively. We will claim, however, that this is still not merely an internalization of the old grammar, as some development did take place in Modern Hebrew, although we cannot point to a specific origin, as there could be many. As we will show, beneficiary/maleficiary and so-called possessor datives, which we will refer to here as “affected datives,” are already attested in Biblical and Mishnaic Hebrew, in a way highly similar to Modern Hebrew, both distributionally and interpretatively. These constructions are also attested in languages that were spoken by speakers of different origins present at the time of the revival and, in general, by members of the Jewish communities prior to the time of revival in different regions of the Jewish Diaspora. 5 Importantly, this linguistic phenomenon is not limited only to these languages, as typologically non-core datives are a widely attested phenomenon (inter alia Payne & Barshi 1999; Zúñiga & Kittilä 2010; Horn 2008). Only one type of non-core dative, the discursive dative, which is a feature of spoken language, cannot be considered a re-emergence of an old phenomenon; since it does not appear in the ancient Hebrew texts, its appearance is probably purely due to the modern languages spoken at the time of revival. Affected and discursive datives are widely available in both European and Semitic languages spoken in the Jewish communities, so it is not at all clear to which languages the re-emergence of these constructions in Modern Hebrew should be attributed. Previously, it has been claimed that it is entirely due to a Slavic-Yiddish influence, and we wish to demonstrate that this is not a tenable claim as such 5  It is worth noting that not only Jews of East-European origin lived in pre-1948 Palestine and were part of the time of the revival. See for instance Vilnay (1963:183–190).


Bar-Asher Siegal and Boneh

and that it should be considerably nuanced. Consequently, the paper wishes to reject the following two claims, repeatedly made in the literature: Claim 1: Non-core dative constructions have a much wider range of uses than in earlier stages of the Hebrew language (this claim is mentioned in Berman 1982 and endorsed by others, such as Halevy 2007, 2013; Linzen 2014; Zeldes 2013).6 Claim 2: The new dative constructions of colloquial Modern Hebrew are due to a Slavic-Yiddish influence (Evan-Zohar 1986; Halevy 2007, 2013; Zeldes 2013). With respect to Claim 1, we will demonstrate that only one of its uses, the discursive dative, is unattested in the classical texts, whereas, with respect to Claim 2, we will argue that it is impossible to point to a specific language or group of languages as the source of the re-emergence of these constructions in colloquial Modern Hebrew.

Non-core Datives in Modern Hebrew

In order to examine the origin of the phenomenon under discussion, we begin with a brief semantic description of the relevant constructions in Modern Hebrew. Such a description will allows us to seek parallels in previous periods and in other languages. We discuss here three groups of non-core datives: affected datives, which cover beneficiary/maleficiary and possessive datives (and under some classifications also ethical datives) exemplified in (1); discursive datives in (2); and co-referential datives/reflexive datives in (3). The following are all attested examples: (1)

a. ‫ והשאיר אותי חרדה ולחוצה‬,‫כך הוא טס לי לשנה למזרח‬-‫אחר‬ ʔaħarkax hu tas l-i le-šana la-mizraħ ve-hišʔir then he flew to-year to.the-east and-left ʔoti ħareda ve-leħuca me anxious and stressed

6  It is likely that a similar assumption stands behind the prescriptive opposition to the use of some of these constructions (Sivan 1976); See also "‫אמור‬-‫רשימת "אל תאמר‬.

Reconsidering the Emergence of Non-core Dative Constructions


‘Then he flew on me to the Far East for a year, and left me anxious and stressed.’ ( D7%90%D7%99%D7%AA%D7%99-%D7%A2%D7%A9%D7%94%D7%A4%D7%A1%D7%99%D7%9B%D7%95%D7%9E%D7 %98%D7%A8%D7%99-%D7%99%D7%A6%D7%90%D7%94%D7%9C%D7%99-%D7%94%D7%A0%D7%A9%D7%9E%D7%9, May 19, 2007, accessed November 23, 2013) b. ‫מפתיע איך אף אחד לא שבר להם את הבית עד עכשיו‬ maftiyaˁ eyx af eħad lo šavar la-hem ʔet ha-bayit surprising how no-one not broke acc the house ʕad ʕaxšav till now ‘It’s surprising how come no one broke their house on them yet.’ (,1 1382,L-4527341,00.html, June 5, 2014, accessed November 23, 2013) (2) a. !?‫איזה טס למשחק? באשקלון הוא לא הולך פתאום טס לי לשוויץ‬ ʔeyze tas la-misħak? be-ʔaškelon hu lo holex which fly to.the-game in-Ashkelon he not go pitʔom tas l-i le-švayc?! suddenly flying to-Switzerland?! ‘What do you mean fly to the (soccer) game, even to Ashkelon he won’t go, suddenly he’s flying to Switzerland?! Unbelievable!’ (, June 9, 2005, accessed November 23, 2013 ) b. ‫המשגה הרביעי הוא החשיבה שמסחר לטווח קצר מספק הזדמנויות להרוויח‬

‫ וכאן שוברים לך שלושה חוקים מרכזיים של החיים‬,‫הרבה מהר ובקלות‬

ha-mišge ha-reviʕi hu ha-ħašiva the-mistake the-fourth the-thinking še-misħar li-tvaħ kacar mesapek hizdamnuyot that-commerce to.range short provides opportunity leharviaħ harbe maher u-vekalut ve-kan šovrim to.earn a lot quick and-easily and-here break le-xa šloša ħukim merkaziyim šel ha-ħayim three rules central of the-life


Bar-Asher Siegal and Boneh

‘The fourth mistake is thinking that commerce in the short run provides an opportunity to earn a lot quickly and easily, and here, contrary to your (=one’s) expectations, three central rules of life are broken.’ ( (January 31, 2007, accessed November 23, 2013) (3) a. ‫לא מזמן טסנו לנו לחצי שבוע בחו"ל‬ lo mizman tasnu l-anu le-ħaci šavuʕa be-ħul not long.time.ago flew to-half week in-abroad ‘Not long ago we flew for half a week abroad.’ ( messageid=106943816, October 13, 2007, accessed November 23, 2013) b. ‫נשבר לו עוד קיץ על המרפסת‬ nišbar l-o ʕod kayic ʕal ha-mirpeset broke another summer on the-balcony ‘Another summer came to an end on the balcony.’ ( kid=2235, (February, 17, 2008, accessed November 23, 2013) Affected datives contribute to truth conditions in that the dative marked DP denotes an event participant that is seen as affected by the occurrence expressed in the sentence either materially or psychologically (Bar-Asher Siegal & Boneh 2014, 2015b). In contrast, discursive datives and co-referential datives, which are only pronominal, do not have such a contribution to truth-conditions. The truth conditional meaning of the sentence is unaltered whether the dative pronoun is there or not. The discursive dative, restricted to first and second person pronouns, indicating the speech event participants, is felicitous in contexts where the proposition asserted constitutes an exception to a generalization available to the speech event participant in the conversational background (Bar-Asher Siegal & Boneh 2015a). Finally, the co-referential dative, easily identifiable as a subcategory of non-core datives because its inflectional features match those of the subject DP, is non-truth conditional, like the discursive dative (cf. Gesenius 1910, §119s). An adequate formulation of its interpretative contribution has not been achieved; see AlZahre & Boneh (2010) and Halevy (2004, 2007, 2013) for recent attempts. For studies on the various non-core dative constructions see Berman (1982), Borer & Grodzinsky (1986), Landau (1999), Halevy (2004, 2007, 2013), Dattner (2014), Gafter (2014), Linzen (2014), and Ariel et al. (2015).

Reconsidering the Emergence of Non-core Dative Constructions


Non-core Datives in Biblical and Mishnaic Hebrew

Speakers of Hebrew at the time of its revival were likely to be exposed to affected (4) and co-referential datives (5) found in Biblical and Classical Rabbinic texts (as for these datives in the Biblical corpus, see Muraoka 1978; Dan 2013): (4) a. ‫ְׁשמֹר ְל ַע ְב ְּדָך ָדוִ ד ָא ִבי ֵאת ֲא ֶׁשר ִּד ַּב ְר ָּת ּלֹו‬ šəmor lə-ʿaḇdə-ḵā Dāvid ʾāḇ-ī David ʾēt ʾăšer dibbartā l-ō acc rel ‘Keep for your servant David my father the promises you made to him.’ (1 Kings 8:25) b. ‫השוכר את הפועל לשמר לו את הפרה ולשמר לו את התינוק‬ has-soḵer ʾet hap-poʿel lišmor l-o et acc the-worker watch.inf to-3.m.sG ACC ʾet hat-tinoq hap-para we- lišmor l-o the-cow and-watch.inf to-3.M.SG acc the-child ‘When someone hires a worker in order to watch his cow or his child.’ (Tosefta, Šabb. 17:26) ָ ‫ִמי יֵ ֶל‬ c. ‫לנּו‬-‫ְך‬ mī yēleḵ=l-ānū who ‘Who will go for us?’ (Isaiah 6:8)

(5) a. ֹ‫ַהּגֶ ֶׁשם ָח ַלף ָה ַלְך לו‬ hag-gešem ḥālap̄ hālaḵ the-rain ‘The rains are over and gone.’ (Song of Songs 2:11)


b. ‫הלך לו למדינת הים‬ halak l-o li-mdinat hay-yam to-land.of the.sea ‘He is gone beyond seas.’ (Mishnah, Ketubbot 9:8) In the appendix we provide more examples of these constructions. As one can easily discern, among the examples mentioned in the appendix, there are


Bar-Asher Siegal and Boneh

various examples of affected non-core datives that are classified in the literature as possessive datives (4b, iii, iv, v, xii), beneficiary datives (4a, i, ii, ix, x, xi), or maleficiary datives (vi, vii, xiii), alongside non-truth conditional non-core datives, which are co-referential datives (5a–b, xiv). In addition, among the affected datives, the datival expressions are attached to various kinds of VPs, such as unaccusative (vi, ix, x) and unergative (4c), and also to some stative predicates, VP (vii, viii, xi) & AP (xiii). Finally, it must be noted that the use of the dative is not restricted to material effects but may equally describe psychological effects (4a, ii, vii, viii, xi, xiii). As for the referents of the datives, in most of the examples they are represented with pronouns of all persons, but examples of full DP are found too (4a, xii, xiii). In most cases, the referents themselves are animates, but examples of inanimates are documented as well (xii). These are highly similar in their interpretation and distributional properties to the Modern Hebrew affected datives discussed by Bar-Asher Siegal & Boneh (2014, 2015b). There are of course many Biblical and Mishnaic examples of co-referential datives (for recent studies concerning this construction in these corpora, see Shemesh 2010 and Dan 2013). It is sufficient to mention only a few examples (5, xiv) since there is an overall consensus (inter alia Berman 1982) that this is an original Hebrew phenomenon. Finally, thinking about the corpora to which speakers at the time of the revival of Modern Hebrew were exposed, one must remember that in the rabbinic style of the Middle Period of Hebrew, Mishnaic Hebrew and Babylonian Aramaic were mixed together, and due to the fact that Hebrew and this dialect of Aramaic are so close, both in vocabulary and in grammar, their grammar often converged. Therefore, it is worth noting that in Babylonian Aramaic affected and reflexive datives are well attested, and could affect those who had used Rabbinic Hebrew as well (Bar-Asher 2007; Bar-Asher Siegal 2014). Among the Modern Hebrew non-core datives, only the discursive dative group is unattested in the classical corpora. This is not surprising, since even in Modern Hebrew it is very rarely represented in the written language. As Ullendorf (1992:1) noted, this dative “belongs to the sphere of the colloquial, slightly sub-standard, and certainly to popular rather than high-flown speech or writing.” Thus, even if it were spoken in the ancient times when Hebrew was still a spoken language, its appearance in Modern Hebrew is most likely due to the foreign languages spoken by the revivers, but because it was available in all the major linguistic communities at the time of the revival, as we shortly show, it is impossible to pinpoint its origin to an exclusive Slavic-Yiddish origin. It is most likely that its origin is related to the foreign substrata of Modern Hebrew, as it could not have been an imitation of the classical layers of Hebrew. It is

Reconsidering the Emergence of Non-core Dative Constructions


reasonable to suggest that this is a calque at the level of a construction. Thus, speakers who had the three types of non-core datives in their native language, upon reviving the co-referential dative and the affected dative, imported the third type of the datival expression—the discursive one—into their colloquial Hebrew as well.

Non-core Datives in the Contact Languages

Now that we have established that there is equivalence between constructions containing non-core datives in Modern Hebrew and in the earlier Hebrew corpora, we turn to tackle the second question to be examined when seeking the origin of a construction: which languages could support its existence during the period of the revival of Modern Hebrew? Truth conditional and non-truth conditional non-core datives are attested in numerous Indo-European languages (Wierzbicka 1988; Zúñiga & Kittilä 2010; Bosse, Bruening, & Yamada 2012; Cuervo 2003, 2010; Horn 2008; Roberge & Troberg 2009 inter alia), including the immediate contact languages Yiddish (Mark 1978), Slavic (Evan-Zohar 1986), and Ladino (A. Quintana, p.c.), and can be traced back to Greek (Smyth 1920),7 Latin (Van Hoeke 1996), and Old Persian (Kent 1953:80–81, Haig 2008:55–58). Due to limited space, here we provide illustrations of discursive datives only from Hungarian (6a) and French (6b). In both these examples, the dative is found in an environment where the underlying proposition counters expectations of the speaker and/or addressee: (6) a. Hihetetlen! Képes és tényleg megnyeri unbelievable capable and indeed wins nekem a verseny-t. dat.1sg the race-acc ‘It is unbelievable! He can and will bloody well win the race.’ (Rákosi 2008:ex. 3d)

b. Au mont St Michel la mer te monte à une de ces vitesses ! At Mont St. Michel, the sea rises in one of these speeds ‘You won’t believe how quickly the sea raises at the Mont St. Michel!’ (Leclère 1976)

7  Bendavid (1967:146–147) attributes some of the features of the co-referential datives in Mishnaic Hebrew to Greek influence.


Bar-Asher Siegal and Boneh

Similarly, these constructions appear in many Semitic languages (Ullendorf 1992), ancient (Contini 1988, Joosten 1989) and modern, as dialectal Arabic also instantiates this array of constructions. Feghali (1928), Al-Zahre (2003), Al-Zahre & Boneh (2010), and Haddad (2014), for example, documented such constructions in the Levant, in the Syrian and Lebanese dialects. Yoda (2013) provides examples of these dative constructions in various Syro-Palestinian dialects, and Brustad (2000) discusses the function of these constructions in the Egyptian and Moroccan dialects. There is even evidence that some of these constructions were already in use in medieval Judeo-Arabic (Blau 2005). The following are examples of these datives in Judeo-Arabic dialect of Tafilalt.8 Example (7a) illustrates an instance of an affected dative, where the speaker is emotionally affected by the occurrence described in the clause “wanting to go to the land of Israel”; example (7b) is an instance of the discursive dative, where “his standing in the middle of the road” counters the expectations of the speaker. According to our consultant, this example does not convey that s/he is in some way affected by this occurrence. ḥəbbət təmsi=l-i l-īris (7) a. ʿla why ‘Why would you do that to me, to go to the land of Israel?’ b. hada wakf=l-i/na f-ṭ-ṭriq in-the-road ‘Look at him, he’s standing in the middle of the road!’ It seems then that this phenomenon is widely attested in many languages that were in direct contact with Hebrew at the time of the revival, but is not limited to those. Note, for instance, that discursive datives that clearly resemble the one attested in Modern Hebrew are available in languages that had relatively limited contact with the Hebrew of the revival, as the examples from Hungarian and French above illustrate. Conclusions In light of these data, it is evident that it is impossible to attribute the appearance of these constructions in Modern Hebrew to a single source. This seems 8  We wish to thank our consultant Moshe Bar-Asher for the examples in this dialect.

Reconsidering the Emergence of Non-core Dative Constructions


to be a case in which the re-emergence of a linguistic phenomenon in Modern Hebrew could have been influenced by parallel constructions in the native spoken languages of the speakers of the revival period. Therefore, not only were the old constructions revived, but an additional construction sharing the same morphology appeared, or at least has been attested and documented for the first time in the history of Hebrew. At the same time, this is a linguistic phenomenon that is typologically widely attested and might be a general feature of natural language. It is difficult, however, to determine to what extent language contact may have influenced conventions of use of non-core datives and issues pertaining to register. Previous studies that argued for a Slavic-Yiddish origin seem not to have seriously considered the older corpora, and have failed to provide any evidence that supports the claim that initially these constructions were used only among speakers of Slavic languages and not among Jews from other places in the Diaspora, especially in the absence of information pinpointing the emergence of the discursive dative. Their claim might therefore be taken to be sociologically biased at the expense of the linguistic facts. Appendix )‫יג‬,‫לי ָא ִחי הּוא (בראשית כ‬-‫י‬ ִ ‫ ִא ְמ ִר‬i .i )‫ ו‬,‫ה ֶח ֶסד ַהּגָ דֹול ַהּזֶ ה (מל"א ג‬-‫ת‬ ַ ‫לֹו ֶא‬-‫וַ ִּת ְׁש ָמר‬iiii )‫ ז‬,‫הּיָ ֶלד (שמות ב‬-‫ת‬ ַ ‫וְ ֵתינִ ק ָלְך ֶא‬iiiii )‫ ט י‬,‫ה ֲא ָד ָמה (שמו"ב‬-‫ת‬ ָ ‫וְ ָע ַב ְד ָּת ּלֹו ֶא‬iviv )‫ יג‬,‫שוברין לו דלתות הבית (תוספתא שבת טו‬v .v )‫ שבת קיז ע"ב‬,‫נשברה לו חבית בראש גגו (בבלי‬vivi )‫ כב‬,‫ּפ ָׁש ָעיו ֲא ֶׁשר ָע ָׂשה לֹא יִ ּזָ ְכרּו לֹו (יחזקאל יח‬-‫ל‬ ְ ‫ ָּכ‬vivii )‫ ויחי פרשה צט‬,‫והיה מתירא יהודה שלא יזכור לו מעשה תמר (בראשית רבה‬viviii ‫לֹו נַ ֲח ָלה ָל ֶׁש ֶבת ִּכי לֹא נָ ְפ ָלה ּלֹו ַעד ַהּיֹום ַההּוא‬-‫ ֵׁש ֶבט ַה ָּדנִ י ְמ ַב ֶּקׁש‬,‫ּובּיָ ִמים ָה ֵהם‬ ַ ixix )‫ א‬,‫ ְּבנַ ֲח ָלה (שופטים יח‬,‫ְּבתֹוְך ִׁש ְב ֵטי יִ ְׂש ָר ֵאל‬ )‫ ט‬,‫מי שהיו לו פירות שביעית שנפלו לו לירושה (משנה שביעית ט‬x .x - ‫כל זכיותיו של משה לא עמדו לו בשעת דחקו (מכילתא דרבי ישמעאל יתרו‬xixi )‫מסכתא דעמלק פרשה א‬ )‫ טז‬,‫צ ַֹהר ַּת ֲע ֶׂשה ַל ֵּת ָבה (בראשית ו‬xixii )‫יג‬,‫וְ ַאנְ ֵׁשי ְסד ֹם ָר ִעים וְ ַח ָּט ִאים ַליהוָ ה (בראשית יג‬xixiii )‫ כא‬,‫מ ְצ ַריִ ם (מל״ב יח‬-‫ל‬ ִ ‫ ַע‬,‫   ַע ָּתה ִהּנֵ ה ָב ַט ְח ָּת ְּלָך ַעל ִמ ְׁש ֶענֶ ת ַה ָּקנֶ ה ָה ָרצּוץ ַהּזֶ ה‬.xiv


Bar-Asher Siegal and Boneh


Al-Zahre, Nisrine. 2003. La Structure du Groupe Verbal en Arabe: Trois arguments empiriques en faveur d’une vision syntaxique de la structure argumentale. Université Paris 8, Saint-Denis. ——— & Nora Boneh. 2010. “Coreferential Dative Constructions in Syrian Arabic and Modern Hebrew.” Brill’s Annual of Afroasiatic Languages and Linguistics 2: 248–282. Ariel, Mira, Elitzur Dattner, John W. Du Bois, & Tal Linzen. 2015. “Pronominal Datives: The Royal Road to Argument Status.” Studies in Language 39.2: 257–321. Bar-Asher, Elitzur A. 2007. “The Origin and the Typology of the Pattern qtīl lî in Syriac and Babylonian Aramaic.” In Sha’arey Lashon: Studies in Hebrew, Aramaic, and Jewish Languages in Honor of Moshe Bar-Asher, eds. A. Mamman et al. Jerusalem: Mosad Bialik (in Hebrew). Bar-Asher Siegal, Elitzur A. 2014. “From a Non-Argument-Dative to an ArgumentDative: The Character and Origin of the qṭīl lī Construction in Syriac and Jewish Babylonian Aramaic.” Folia Orientalia 51: 59–101. ——— & Nora Boneh. 2014. “Modern Hebrew Non-core Datives in Their Context.” Lǝšonénu 74: 461–495 (in Hebrew). ——— & Nora Boneh. In press, to appear in 2015. “Decomposing Affectedness: Truth-Conditional Non-core Datives in Modern Hebrew.” MIT Working Papers in Linguistics, Proceedings of IATL30, ed. Nurit Melnik. Beit Hasefer Hareali. 1952. ‫אמור"מתוך "ידיעות למורים" של בית הספר הריאלי‬-‫"אל רשימת‬. ‫ י"ג באייר תשי"ב‬,‫בחיפה‬

Bendavid, Aba. 1967. Leshon miḳra u-leshon ḥakhamim (Biblical Hebrew and Mishnaic Hebrew). Tel-Aviv: Dvir. N.P. Berman, Ruth A. 1982. “Dative Marking of the Affectee Role: Data from Modern Hebrew.” Hebrew Annual Review 6: 35–59. Blau, Joshua. 2005. “The Dativus Ethicus in Medieval Judaeo-Arabic.” In Semitic Studies in Honour of Edward Ullendorff, ed. Geoffrey Khan. Leiden: Brill, 109–114. Borer, Hagit & Yosef Grodzinsky. 1986. “Syntactic Cliticization and Lexical Cliticization: The Case of Hebrew Dative Clitics.” In Syntax and Semantics, ed. Hagit Borer. New York: Academic Press. 175–217. Bosse, Solveig, Benjamin Bruening, & Masahiro Yamada. 2012. “Affected Experiencers.” Natural Language & Linguistic Theory 30: 1185–1230. Breuer, Yoḥanan. 1987. “On the Hebrew Dialect of the Amoraim in the Babylonian Talmud.” Language Studies 2–3: 127–153. Brustad, Kristen. 2000. The Syntax of Spoken Arabic : A Comparative Study of Moroccan, Egyptian, Syrian, and Kuwaiti Dialects. Washington, DC: Georgetown University Press. Contini, Riccardo. 1998. “Considerazioni sul presunto ‘dativo etico’ in aramaico Pre-Cristiano.” In Études sémitiques et samaritaines offertes à Jean Margain, eds.

Reconsidering the Emergence of Non-core Dative Constructions


Christian-Bernard Amphoux, Albert Frey, & Ursula Schattner-Rieser. Lausanne: Editions du Zèbre, 83–94. Cuervo, Maria Cristina. 2003. Datives at Large. PhD diss., MIT. ———. 2010. “Against Ditransitivity.” Probus 22: 151–180. Dan, Barak. 2013. “The Dativus Ethicus in Biblical Hebrew and Early Translation.” Language Studies 14–15: 1–49. Dattner, Elitzur. 2014. Mapping Hebrew Dative Constructions. PhD diss., Tel-Aviv University. Even Zohar, Itamar. 1986. “The Dialogues of Gnessin and the Question of Russian Models.” In Uri Nissan Gnessin: Research and Documents, eds. Dan Miron & Dan Laor. Jerusalem: Mosad Bialik, 11–41 (in Hebrew). Feghali, Michel T. 1928. Syntaxe des parlers arabes actuels du Liban. Vol. t. 9, Bibliothèque de l’École des langues orientales vivantes. Paris: Impr. nationale. Gafter, Roey. 2014. “The Distribution of the Hebrew Possessive Dative Construction: Guided by Unaccusativity or Prominence?” Linguistic Inquiry 45.3: 482–500. Gesenius, Wilhelm & E. Kautzsch. 1910. Gesenius’ Hebrew Grammar, second English ed. Oxford: Clarendon Press. Haddad, Youssef A. 2014. “Attitude Datives in Lebanese Arabic and the Interplay of Syntax and Pragmatics.” Lingua 145: 65–103, doi: .lingua.2014.03.006. Haig, Geoffrey L. J. 2008. Alignment Change in Iranian Languages: A Construction Grammar Approach, Empirical Approaches to Language Typology, vol.  37. Berlin: Mouton de Gruyter. Halevy, Rivka. 2003–2004. “Šoteq lo ha-’adon’: The Function of the Construction ‘Verb + l + Personal Reflexive Pronoun in Contemporary Hebrew.” Lǝšonénu 65: 113–143 (in Hebrew). ———. 2007. “The Subject Co-referential ‘l-‘ Pronoun in Hebrew.” In Studies in Semitic and General Linguistics in Honor of Gideon Goldenberg, eds. T. Bar & E. Cohen. Münster: Ugarit-Verlag, 299–321. ———. 2013. “The Dative in Modern Hebrew.” In Encyclopedia for Hebrew Language and Linguistics, eds. Geoffrey Khan et al. Leiden: Brill Online, dative-modern-hebrew-EHLL_COM_00000339. Horn, Laurence R. 2008. “I Love Me Some Him: The Landscape of Non-argument Datives.” In Empirical Issues in Syntax and Semantics, eds. O. Bonami & Cabredo Hofherr. Paris: CSSP, available online only,, 169–192. Joosten, Jan. 1989. “The Function of the So-called Dativus Ethicus in Classical Syriac.” Orientalia 58: 473–492. Kent, Roland G. 1953. Old Persian: Grammar, Texts, Lexicon, American Oriental Series. New Haven, CT: American Oriental Society.


Bar-Asher Siegal and Boneh

Landau, Idan. 1999. “Possessor Raising and the Structure of VP.” Lingua 107: 1–37. Leclère, Christian. 1976. “Datifs syntaxiques et datif éthique.” In Méthodes en grammaire française, eds. Maurice Gross & Jean-Claude Chevalier. Paris: Klincksieck, 73–96. Linzen, Tal. 2014. “Parallels between Cross-linguistic and Language-internal Variation in Hebrew Possessive Constructions.” Linguistics 52: 759–792. Mark, Yudel. 1978. A Grammar of Standard Yiddish. New York: Publications of the Congress for Jewish Culture, Wladon Press, Inc. Muraoka, T. 1978. “On the So-called Dativus Ethicus in Hebrew.” The Journal of Theological Studies 29.2: 495–498, doi: 10.1093/jts/XXIX.2.495. Payne, Doris L. & Immanuel Barshi, eds. 1999. External Possession, Typological Studies in Language, vol. 39. Amsterdam: John Benjamins. Rákosi, György. 2008. “Some Remarks on Hungarian Ethical Datives.” In When Grammar Minds Language and Literature. Festchrift for Prof. Béla Korponay on the Occasion of his 80th Birthday, eds. Béla Hollósy, József Andor, Tibor Laczkó, & Péter Pelyvás. Debrecen: Institute of English and American Studies, University of Debrecen, 413–422. Roberge, Yves & Michelle Troberg. 2009. “The High Applicative Syntax of the Dativus Commondi/Incommodi in Romance.” Probus 21: 249–289. Shemesh, Rivka. 2010. “The Usage of the Co-agentive Dative Pronoun in the Mishnah and Tosefta.” Folia Orientalia 47: 141–179. Smyth, Herbert W. 1920. Greek Grammar for Colleges. New York: American Book Company. Ullendorf, Edward. 1992. “Some Observations on the Dativus Ethicus in Semitics and Elsewhere.” Jerusalem Studies in Arabic and Islam 15: 1–9. Van Hoecke, Willi. 1996. “The Latin Dative.” In The Dative, eds. W. Van Belle & W. Van Langendonck. Amsterdam: John Benjamins, 3–38. Vilnay, Zev. 1963. “The Maghreb? Jews as Pioneers of the Yishuv in the Land of Israel.” Ozar Yehudey Sefarad le-Xeker Toldot Yehudey Sefarad ve-Tarbutam 6: 183–190 (in Hebrew). Wierzbicka, Anna. 1988. The Semantics of Grammar. Amsterdam: John Benjamins. Yoda, Sumikazu. 2013. “On the So-called Ethical Dative in Syro-Palestinian Arabic dialects.” In Nicht Nur Mit Engelszungen: Beiträge zur semitischen Dialektologie. Festschrift für Werner Arnold zum 60. Geburtstag, eds. Renaud Kuty, Urlich Seeger, & Shabo Talay. Wiesbaden: Otto Harrassowitz, 391–398. Zeldes, Amir. 2013. “Is Modern Hebrew Standard Average European? The View from European.” Linguistic Typology 17: 439–470. Zúñiga, Fernando & Seppo Kittilä, eds. 2010. Benefactives and Malefactives: Benefactives and Malefactives. Amsterdam-Philadelphia: John Benjamins.


A Constructional Idiom in Modern Hebrew: The Influence of English on a Native Hebrew Collocation Malka Rappaport Hovav

The Hebrew University of Jerusalem, Israel [email protected]

Abstract This article discusses a constructional idiom that has developed in recent Modern Hebrew, based on the Classical Hebrew collocation expressing the committing of suicide, with the verb replaced by an open position that can be filled by a wide range of verbs. It is argued that the development of the idiom involved a syntactic re-analysis of the original collocation whereby the PP is reanalyzed as a result PP and the reflexive is analyzed as a non-subcategorized NP. It is suggested that the idiom developed under the influence of a similar productive construction in English. The interpretation of the constructional idiom is briefly explored and comparison is made with another constructional idiom based on a group of native collocations under the influence of English.

Keywords constructional idiom – resultative construction – reflexive – fake reflexive – change of state – collocation – productivity

* This article was written during my stay at the Mandel Scholion Interdisciplinary Research Center in the Humanities and Jewish Studies, whose support I gratefully acknowledge. Thanks to the members of the Research Group on the Emergence of Modern Hebrew for helpful discussion.

© koninklijke brill nv, leiden, ���6 | doi ��.��63/9789004310896_025


Rappaport Hovav

Introduction: A Constructional Idiom in Modern Hebrew

In 2007, Prime Minister Ehud Olmert expressed his opposition to the depletion of government resources by excessive fortification of buildings in southern Israel thus: (1)

.‫לא נמגן את עצמנו לדעת‬ lo nemagen ʔet neg fortify acc

ʕacmenu ourselves

la-daʕat to.def-mind

I have intentionally not included the translation at this point. The sentence contains a constructional idiom—a construction with an open position, in which the fixed components contribute a meaning which is not fully compositional. Here, the fixed components are the possessed form of ‫ עצם‬ʕecem ‘self’ and ‫ לדעת‬la-daʕat ‘to.def-mind.’ The position of the verb is open.1 The following illustrate the idiom with other verbs, still without translation: (2) a. .‫בוגי יעלון מפטפט את עצמו לדעת‬ bogi yaʕalon mefatpet ʔet ʕacmo Bogi Yaalon blabbers acc himself b. .‫השמאל משקר את עצמו לדעת‬ hasmol mešaker ʔet ʕacmo def.left lies acc itself c. .‫הארץ ממחזר את עצמו לדעת‬ haʔarec memaħzer Haaretz (newspaper) recycles

la-daʕat to.def-mind

la-daʕat to.def-mind ʔet ʕacmo la-daʕat acc itself to.def-mind

I argue that the idiom results from the grafting of an English construction onto a Hebrew collocation, via a reanalysis of the syntax of the original collocation. The collocation that serves as the basis of this idiom is ‫ איבד את עצמו לדעת‬ʔibed 1  In more traditional idioms, such as ‫ גמר אומר‬gamar ʔomer in Hebrew (‘made a decision;’ literally, ‘finished saying’), or “spilled the beans” in English, all components are fixed. For more on constructional idioms, see Goldberg (1995), Jackendoff (1997; 2002), and Espinal & Mateu (2010), among others. An insightful discussion of the constructional idiom under discussion appears in Borochovsky & Sovran (2003). They, too, point to the fact that the construction involves a shift in the syntax of the original Hebrew phrase ‫איבד את עצמו לדעת‬ ʔibed ʔet ʕacmo la-daʕat ‘committed suicide.’

A Constructional Idiom in Modern Hebrew


ʔet ʕacmo la-daʕat, which translates, leaving aside tense and agreement, as ‘to commit suicide.’ The English construction under whose influence the idiom seems to have developed is illustrated in English translations of Olmert’s statement in the electronic press: “Olmert: ‘We Won’t Fortify Ourselves to Death;’ ”2 “We can’t fortify ourselves senseless, says Olmert.”3 The English translations instantiate subtypes of the resultative construction, which is extremely productive in English but is only marginally instantiated in Modern Hebrew and not productive. I argue that the process of replacing the verb ‫ לאבד‬le-ʔabed ‘to obliterate’ with an open position in which many verbs may appear4 has effectively introduced a productive resultative construction into Modern Hebrew, under the influence of a specific subtype of the English resultative construction. In the following sections, I first review the properties of the Hebrew collocation. I then provide a short overview of the range of resultative constructions in English, including the reflexive resultative. Next, I show how the replacement of the verb in the native Hebrew collocation with an open position produced a constructional idiom with a meaning similar to the English reflexive resultative. Finally, I show that there is another constructional idiom developing in Modern Hebrew under the influence of another subtype of the English resultative construction—the way construction.

Committing Suicide in Hebrew

The collocation ‫ איבד את עצמו לדעת‬ʔibed et ʕacmo la-daʕat ‘to commit suicide’ appears in Mishnaic Hebrew. (3) )‫אלא שאין דרכו של אדם לאבד את עצמו לדעת (בראשית רבה פב‬ ʼella še-ʼen dark-o šel ʼadam le-ʼabbed but that-neg way-POSS.3MSG of man to-lose ʼet ʻacmo la-daʻat acc himself to.def-mind ‘It is not the way of a person to commit suicide.’ (Genesis Rabbah 82) 2  See Accessed on June 15, 2015. 3  See,7340,L-3418913,00.html. Accessed June 14, 2015. 4  It remains to be determined which verbs in fact appear in this construction. Languages that have resultative constructions vary in the range of verbs allowed to appear in them. See, for example, Washio (1997); Croft et al. (2010).


Rappaport Hovav

Biblical Hebrew ‫ לאבד‬lə-ʔabbed instantiates the root ‫ אבד‬in the Piʕel template. This root is loosely associated with the sense of ‘loss,’ and in Piʕel it means ‘to obliterate,’ where this can be taken to mean ‘cause loss of existence.’ The verb is always used with an agentive subject; therefore, the phrase ‫לאבד את עצמו‬ le-ʼabed ʼet ʻacmo means ‘to obliterate oneself’ or, more specifically, to ‘kill oneself.’ The ‫ לדעת‬la-daʻat ‘to.def-mind’ here is used adverbially, meaning ‘knowingly, intentionally.’ ‫ דעת‬daʻat ‘knowledge’ with this meaning appears frequently in Rabbinic Hebrew with the preposition le- ‘to,’ but also with the preposition me- ‘from’: ‫ יאוש שלא מדעת‬yeʼuš šello mi-ddaʻat (unconscious resignation ‫ בבא מציעא‬Bava Meziʼa), and sometimes with be- ‘in’: ‫פרט למכה שלא‬ ‫ בדעת‬pǝrat lǝ-make šello be-daʻat, ‘except for one who smites unknowingly’ (‫ ספרי מ"א‬Sifre 41). While the entire phrase is interpreted compositionally, it is a collocation because ‫ איבד‬is the only verb that regularly appears in this frame. We do not find, for example, ‫ הרג את עצמו לדעת‬harag ʼet ʻacmo la-daʻat with the verb ‫להרוג‬ la-harog ‘to kill,’ or ‫ השמיד את עצמו לדעת‬hišmid ʼet ʻacmo la-daʻat with the verb ‫ להשמיד‬le-hašmid ‘to destroy’ to indicate suicide. Moreover, in this particular collocation we find only the preposition le- ‘to’ but not be- or me-.5 The collocation continues to appear in all strata of Rabbinic Hebrew, including the Responsa literature of all ages.6 In (non-Rabbinic) Modern Hebrew, the verb ‫ לאבד‬le-ʔabed is almost exclusively used in the sense of loss of possession— and the verb ‫ להתאבד‬le-hitʔabed ‘commit suicide’ (derived from the same root and template but in the middle voice) has almost completely replaced the collocation. Based on Google N-gram Viewer, the use of ‫איבד את עצמו לדעת‬ ʔibed et ʕacmo la-daʕat drops precipitously in the 1930s, when the use of ‫התאבד‬ hitʔabed begins to increase significantly. By the year 2000, the latter has almost completely replaced the former.

5  The sole exception is the phrase as it appears a number of times in Masexet Semaḥot, for example: ‫ אין מתעסקין עמו בכל דבר‬- ‫ המאבד את עצמו בדעת‬ha-mmeʼabbed ʼet ʻacmo be-daʻat ʼen mitʻasqin ʻimmo be-kol davar ‘He who commits suicide, nothing is done for him.’ 6  Only one other verb regularly appearing with ʼet ʻacmo la-daʻat but only in halakhic texts: hikša ‘to harden oneself,’ where the interpretation is sexual. But this is only in halakhic texts and is certainly less common than ‫מאבד את עצמו לדעת‬.

A Constructional Idiom in Modern Hebrew


Resultative Constructions in English and Hebrew

We return to the English translations of (1) above representing two subtypes of the resultative construction.7 Resultative constructions describe the causation of a change of state (or location) by some specified activity. (4a–c) include transitive verbs: the activity described by the verb causes the referent of the direct object to enter into the state specified by the adjective phrase (AP) or prepositional phrase (PP). (4) a. John wiped the table clean. b. My son smashed the vase to pieces. c. They stabbed him to death. In contrast to the sentences in (4), in (5a, b) the post-verbal NP is not an argument of the verb: without the result AP or PP the verb cannot appear with a direct object (*The audience laughed the actor; *The dog barked the neighbor). Levin & Rappaport Hovav (1995) call this the “nonsubcategorized NP resultative.” In (5c), the post-verbal NP does not correspond to the verb’s normal object, and is based on the independently available use of the verb without its object: (John drove (his car)). (5) a. The audience laughed the actor off the stage. b. The dog barked the neighbor awake. c. Drive your engine clean (gasoline advertisement). (6a, b) are a special case of a non-subcategorized NP resultative in which the post-verbal NP is a reflexive, often called a fake reflexive. In contrast, in (6c), the reflexive is an argument of the verb. (6) a. Ari sang himself hoarse. b. He worked himself to death. c. John frightened himself silly. The result XPs (that is, either APs or PPs) with subcategorized NPs typically express a result that comes about as a natural consequence of the action denoted by the verb; therefore, the verb and the result XP mutually constrain one another (wipe the table clean, but not *wipe the table empty). The semantic 7  For analyses of this construction see Simpson (1983), Hoekstra (1988), Levin & Rappaport Hovav (1995, 2001), Goldberg (1995), and Boas (2003).


Rappaport Hovav

dependencies between the verb and the result XP with non-subcategorized NP resultatives are much looser than with subcategorized resultatives (Wechsler 2005); as a consequence, new combinations of verbs, APs, and PPs are found all the time. The non-subcategorized NP resultatives, then, provide English with a very productive means of expressing the causation of a result state via a specified activity. These are extensively illustrated in works cited above. Hebrew of all strata has no AP resultative constructions.8 Modern Hebrew does, however, have resultatives in which the result state is expressed as a PP.9 Some examples appear below: (7) a. .‫דני ניפץ את האגרטל לרסיסים‬ dani nipec ʔet haʔagartal li-rsisim Dani smashed acc def.vase to-smithereens ‘Dani smashed the vase to smithereens.’ b. 10.‫ חתכתי את העוגה לשמונה חתיכות‬ ħataxti ʔet haʕuga li-šmone ħatixot acc def.cake to-eight slices ‘I cut the cake into eight slices.’ c. .‫ דן דימם למוות‬ dan dimem la-mavet Dan bled to.def-death ‘Dan bled to death.’

8   Hebrew, like other languages without AP resultative constructions, has V NP AP sequences where the verb requires both NP and AP: ‫ עשו אותו שחצן‬ʕasu ʔoto šaħcan ‘they made him conceited’ (see Napoli (1992) on Italian, and Green (1973) on French). 9   Earlier stages of Hebrew, including Biblical Hebrew, have examples such as ‫וְ ֶאת ָה ַאיִ ל נִ ַתח‬ ‫‘( ִלנְ ָת ָחיו‬and he cut the ram into slices,’ Lev. 8:20). Most examples involve actions such as cutting and slicing with the direct object something of a cognate object, sharing a root with the verb. An exception is ‫ למוות‬la-mmaweṯ ‘to death.’ I have found one such example in the Bible (‫ ֶה ֱע ָרה ַל ָמוֶ ת נַ ְפׁשֹו‬, Isa. 53:12), and it is not clear whether it is representative. 10  Rapoport (1999) suggests that the slicing examples are not true resultatives. Broccias (2008) shows that in Old English, the resultative construction was also very limited and began with a small class of verbs, including verbs of cutting. Modern Hebrew has examples such as ‫ טחנתי את הקפה דק‬taħanti ʔet ha-kafe dak ‘I ground the coffee thin.’ Other languages purportedly lacking AP resultatives have parallel examples as well (e.g., Italian; Napoli 1992). Levinson (2010) and Geuder (2000) argue that these are not true resultatives.

A Constructional Idiom in Modern Hebrew


d. .‫ דן שתה לשוכרה‬ dan šata le-šoxra Dan drank to-insobriety ‘Dan drank himself drunk.’ The preposition is always le- ‘to.’ This is because the result state can be seen as a kind of goal, and le- is a typical marker of goal in Hebrew. More generally, one finds across languages that markers of goals are found in resultative constructions (for example, to death in English, a morte in Italian). The range of verbs that can appear with these PPs in Hebrew is restricted, and the verbs dictate a narrow range of NPs in the result PP. The PP specifies a natural end result of the activity denoted by the verb. Actions such as cutting and slicing naturally lead to the formation of pieces and slices; drinking leads to insobriety, and bleeding leads to death. While the same generalization holds for subcategorized NP resultatives in English, the range of result PPs in Hebrew is more restricted. For example, one cannot say ‫ הוא טאטא את החדר לניקיון‬hu tite ʔet haħeder le-nikayon ‘He swept the floor to cleanliness.’11 In general, Hebrew does not have resultatives predicated of nonsubcategorized NPs; there are no word-for-word translations of (5a–c) and (6a, b). With its restriction of the result phrase to a PP and the absence of nonsubcategorized NP resultatives, Hebrew hasn’t been able to use the resultative construction as a productive strategy for the expression of the causation of a change of state by a specified activity. That is, until the development of the constructional idiom under discussion.

V ʔet ʕacmo la-daʕat as a Type of Resultative

The constructional idiom V et ʕacmo la-daʕat has provided Hebrew with a productive strategy for expressing in a single clause the causing of a change of state by an activity specified by the verb. This has been accomplished by two types of re-analysis of ʔibed ʔet ʕacmo la-daʕat. First, the PP la-daʕat has been reinterpreted as a result PP. Second, the reflexive was given the possibility of being interpreted as a non-subcategorized NP, thus allowing both transitive and intransitive verbs to fill the open position of the verb. I expound on each of these re-analyses in turn. 11  Of course, in English, this would be ‘He swept the floor clean.’ The point is that the corresponding PP resultative is not acceptable in Hebrew.


Rappaport Hovav

I suggest that the PP la-daʕat has been reanalyzed as a resultative PP along the lines of the Modern Hebrew result PP la-mavet (‘to death’). In the original collocation, the preposition does not have a clear goal function. Though the adverbial could appear with the prepositions be- or me-, eventually the collocation stabilized exclusively with le-. Since all resultative PPs in Modern Hebrew contain the preposition le- ‘to,’ this adverbial PP could be easily reinterpreted as a goal-marked result phrase. What is the meaning associated with the result PP? The collocation ‫ איבד את עצמו לדעת‬indicates suicide, where the result state is clearly death. Indeed, ‫ למוות‬la-mavet ‘to death’ is one of the more frequent resultatives in Modern Hebrew. However, there are only few examples of the idiom in which ‫ לדעת‬is literally interpreted as death, even with verbs expressing an activity that normally leads to death.12 More often, it is some unspecified negative effect, the nature of the effect being contextually determined. For example, a newspaper article from 1985 describes a commercial that was prohibited from being aired: it depicts a woman being seduced by a man, suddenly putting him off, realizing he reeks of cigarette smoke. The narrator says: ‫ אל תעשן את עצמך לדעת‬ʔal teʕašen ʔet ʕacmexa la-daʕat (don’t smoke yourself la-daʕat). This might be translated into English as: ‘Don’t smoke yourself out of a good catch!’ In the following, the negative effect is made explicit: (8) .‫ הולך ומצטרד‬,‫רון שר את עצמו לדעת‬ ron šar ʔet ʕacmo la-daʕat holex Ron sang acc himself to.def-mind going u-mictared and-becoming.hoarse ‘Ron sang himself ladaʔat, slowly becoming hoarse.’ = ‘He sang himself hoarse.’ (,7340,L-1722–288–10366789,00.html, accessed June 14, 2015) The negative effect is often loss of senses (cf. the second of the two English translations of Olmert’s declaration above):

12  One does find cases of nonsubcategorized reflexives with the result phrase ‫ למוות‬lamavet ‘to death.’ This, however, is much less common and is more transparently a translation from English: ‫ האיש שאכל את עצמו למוות‬ha.ʔiš še-ʔaxal ʔet ʕacmo la-mavet ‘the man who ate himself to death,’, accessed June 14, 2015.

A Constructional Idiom in Modern Hebrew


(9) .‫את עצמם לדעת בסן פרנסיסקו‬ ‫כשילדי הפרחים סיממו‬ keše yaldei hapraħim simemu ʔet ʕacmam when child.cs drugged acc themselves la-daʕat be-san fransisko to.def-mind in-San Francisco ‘When the flower children drugged themselves senseless in San Francisco.’ (, accessed June 14, 2015) This is rather ironic, since it seems to reverse the literal meaning of the PP, which would be to arrive at ones sense’s (cf. the preposition le- ‘to’). I mentioned above that the non-subcategorized NP resultatives lend the true productivity to the construction in English. The reflexive in ʔibed ʔet ʕacmo la-daʕat is a true subcategorized NP; the verb ‫ איבד‬ʔibed selects a direct object. However, in ‫ הוא פטפט את עצמו לדעת‬hu pitpet ʔet ʕacmo la-daʕat ‘He blabbered himself senseless,’13 the verb is intransitive (there is no ‫ * הוא פטפט את עצמו‬hu pitpet ʔet ʕacmo ‘he blabbered himself’), and so the object is a fake reflexive. I suggest that this re-analysis of the reflexive as a possible fake reflexive came about under the influence of English.14 The fake reflexive can also appear with a transitive verb whose normal object is omitted as in (10) and (11). (10) She drank herself senseless. (11) a. .‫מגזר הסלולר תובע את עצמו לדעת‬ migzar haselular toveaʕ ʔet ʕacmo la-daʕat sector def.cellular sues acc itself to.def-mind ‘The cellular sector is suing itself out of business.’

13  Since ‫ לדעת‬la-daʕat is a fixed expression and the result phrase is contextually interpreted, I take some freedom in translating the expressions into English. 14  The relatively late entry of this construction into Modern Hebrew suggests that it did not come about under the influence of the contact languages of Revival Hebrew. I have found no examples of the construction before 1980, and the construction has become common only since 2000. Of the possible contact languages, only German has a similar productive construction. Russian, Yiddish, and Arabic do not. Even today, one finds it mainly in the press and in blogs. Interestingly, when the idiom is used with a nonsubcategorized NP, it is almost exclusively a reflexive and hardly ever a full NP, in line with the original collocation.


Rappaport Hovav

b. .‫המדינה מפריטה את עצמה לדעת‬ hamedina mafrita ʔet ʕacma la-daʕat def.state privatize acc itself to.def-mind ‘The state is privatizing itself to death.’ (, accessed June 14, 2015) It can also be interpreted as a true object for verbs that allow the subject and the object to be co-referential, as in: (12) .‫ בית המשפט העליון מבזה את עצמו לדעת‬ betha.mišpat ha.ʕelyon mevaze ʔet ʕacmo la-daʕat court supreme contempt acc itself to.def-mind ‘The Supreme Court is “contempting” itself to death.’ (, accessed June 14, 2015)

Another Constructional Idiom

More recently, a constructional idiom has arisen out of a syntactic re-analysis of another native (family of) collocations involving [V ʔet dark-o ʔel NP] ([V acc way.poss to NP)], as in the examples below:15 (13) a. .‫אינטל רוכשת את דרכה אל דור העתיד של ייצור השבבים‬ intel roxešet ʔet dark-a ʔel dor Intel acquiring acc way-poss.3sg to generation.cs heʕatid šel yicur hašvavim def.future of production.cs def.chips ‘Intel is acquiring its way to the new generation of chip production.’ b. ‫מסעו של טבח המבשל את דרכו אל החופש הפיננסי‬ masaʕ-o šel tabaħ hamevašel ʔet trek-poss3msg of chef acc dark-o ʔel ha-ħofeš ha-finansi way-poss3msg to def.freedom ‘the trek of a chef who cooks his way to financial freedom’ 15  Borochovksy & Sovran (2003) remark that Hebrew has no equivalent of the English way construction. However, it is amply exemplified in the press. It seems to have entered Modern Hebrew a bit later than the la-daʕat idiom.

A Constructional Idiom in Modern Hebrew


These examples are reminiscent of the way construction in English, described by Goldberg (1995) and Goldberg & Jackendoff (2004), illustrated in the translations of these examples. The way construction also allows for the expression of the achievement of a state by a specified activity in a single clause. The result in this case is depicted as the achievement of a goal by the traversal of path, where the noun “way” designates the path. These examples are also based on (a family of) Classical Hebrew collocations, involving a number of verbs that take the possessed form of ‫ דרך‬derex ‘way’ as their direct object followed by a goal phrase, as in ‫פרץ את דרכו אל‬/‫פילס‬/‫שרך‬/‫ עשה‬ʕasa/sarax/piles/parac ʔet darko ʔel ‘made/plod/straighten/break his way.’16 In these examples, the direct object is an argument of the verb. With the verbs ‫ עשה‬and ‫פרץ‬, the expression is similar to the English ‘to make one’s way.’ The other verbs indicate creation of the way. In contrast, (13a, b) above involve transitive verbs without their normal direct object; in these cases, then, the possessed form of ‫ דרך‬derex ‘way’ has become a nonsubcategorized NP. Here, too, then, we have a constructional idiom, based on a native Hebrew collocation, in which a set (group of) verb(s) is replaced by an open position that can be filled by a wide range of verbs. This provides Modern Hebrew with another productive strategy for describing the attainment of a goal by some specified activity.


Boas, Hans C. 2003. A Constructional Approach to Resultatives (Stanford Monograph in Linguistics). Stanford, CA: Center for the Study of Language and Information; Amsterdam: John Benjamins. Booij, Geert. 2002. “Constructional Idioms, Morphology, and the Dutch Lexicon.” Journal of Germanic Linguistics 14.4: 301–327. Borochovsky, Esther & Tamar Sovran. 2003. “Hebrew Construction Grammar,” in Hebrew—A Living Language III, eds. R. Ben-Shahar & G. Toury. Tel Aviv: The Porter Institute, Tel Aviv University, and Hakibbutz Hame’uxad, 31–50 (in Hebrew).  Broccias, Cristiano. 2008. “Towards a History of English Resultative Constructions: The Case of Adjectival Resultative Constructions.” English Language and Linguistics 12.1: 1–28. Croft, William, Jóhanna Barðdal, Willem Hollmann, Violeta Sotirova, & Chiaki Taoka. 2010. “Revising Talmy’s Classification of Complex Event Constructions.” In Contrastive Studies in Construction Grammar, ed. Hans Boas. Amsterdam: John Benjamins, 201–236. 16  It is actually not entirely clear in all these cases whether the PP headed by ʔel ‘to’ is a complement of the noun or a complement of the verb. This deserves further study.


Rappaport Hovav

Espinal, M. Theresa & J. Mateu. 2010. “On Classes of Idioms and Their Interpretation.” Journal of Pragmatics 42: 1397–1411.  Geuder, Wilhelm. 2000. Oriented Adverbs: Issues in the Lexical Semantics of Event Adverbs. PhD Diss., Universität Tübingen. Goldberg, Adele E. 1995. Constructions: A Construction Grammar Approach to Argument Structure. Chicago: University of Chicago Press. ——— & Ray Jackendoff. 2004.  “The English Resultative as a Family of Constructions.” Language 80: 532–568. Green, Georgia M. 1973. “A Syntactic Syncretism in English and French.” In Issues in Linguistics: Papers in Honor of Henry and Reneé Kahane, ed. Braj B. Kachru. Champaign, IL: University of Illinois Press, 257–278. Hoekstra, Teun. 1988. “Small Clause Results.” Lingua 74: 101–139. Jackendoff, Ray S., 1997. “Twisting the Night Away.” Language 73: 534–559. ———. 2002. Foundations of Language: Brain, Meaning, Grammar, Evolution. Oxford: Oxford University Press, 477. Levin, B. & M. Rappaport Hovav. 1995. Unaccusativity: At the Syntax-lexical Semantics Interface. Cambridge, MA: MIT Press. Levinson, Lisa. 2010. “Arguments for Pseudo-resultative Predicates.” Natural Language and Linguistic Theory 28.1: 135–182. Napoli, D. J. 1992. “Secondary Resultative Predicates in Italian.” Journal of Linguistics 28: 53–90. Rapoport, T. R. 1999. “Structure, Aspect, and the Predicate.” Language 75: 653–77. Simpson, Jane. 1983. “Resultatives.” In Papers in Lexical-functional Grammar, eds. Lori Levin, Malka Rappaport, & Annie Zaenen. Bloomington, IN: Indiana University Linguistics Club, 143–157. Washio, Ryuichi. 1997. “Resultatives, Compositionality, and Language Variation.” Journal of East Asian Linguistics 6: 1–49. Wechsler, Stephen. 2005. “Resultatives under the Event-Argument Homomorphism Model of Telicity.” In The Syntax of Aspect-Deriving Thematic and Aspectual Interpretation, eds. Nomi Erteschik-Shir & Tova Rapoport. Oxford: Oxford University Press, 255–273.

 hen the Construction Is Axla, Everything Is Axla: W A Case of Combined Lexical and Structural Borrowing from Arabic to Hebrew Roey J. Gafter

Tel Aviv University, Tel Aviv, Israel [email protected]

Uri Horesh

Northwestern University, Evanston, IL, USA [email protected]

Abstract This article examines a borrowing from Arabic into Hebrew, which is a combination of a lexical borrowing and a structural one. The Arabic superlative aħla ‘sweetest, most beautiful,’ pronounced by most Modern Hebrew speakers [axla], has shifted semantically to mean ‘great, awesome.’ Yet, as our corpus-based study illustrates, it was borrowed into Hebrew—for the most part—with a very particular syntactic structure that, in Arabic, denotes the superlative. In Arabic itself, aħla may also denote a comparative adjective, though in different syntactic structures. We discuss the significance of this borrowing and the manner in which it is borrowed both to the specific contact situation between Arabic and Hebrew and to the theory of language contact in general.

Keywords Arabic – Hebrew – language contact – structural borrowing – lexical borrowing – ­construct state – superlatives – elative

Introduction When the history of Modern Hebrew is explored, language contact surfaces on every level. Hebrew had fallen into disuse as a vernacular language centuries ago, until a famously successful revitalization project brought it back to the status of a mother tongue for many speakers. But despite conscious attempts to reconstruct the language as it was historically spoken, the Modern Hebrew

© koninklijke brill nv, leiden, ���6 | doi ��.��63/9789004310896_026


gafter and horesh

that emerged is, of course, not the same as Biblical Hebrew, or any other historical form of Hebrew, and has been greatly influenced both by the first languages of the early revitalizers and by the surrounding languages (Zuckermann 2003). Of particular interest is the contact between Hebrew and Arabic. Before the Jewish immigration, Palestinian Arabic was the majority language of the area that was to become the State of Israel, and many Jewish immigrants from Arab countries were native speakers of other varieties of Arabic as well. Unsurprisingly, many Arabic loanwords became very common in Hebrew, and it is hard to imagine current Israeli slang without some of its Arabic staples, yala ‘come on’ (< Arabic yaɫɫa), wala ‘whatever, really?!’ (< Arabic waɫɫaːh), basa ‘darn, bummer’ (< Arabic baʔs), and a range of others.1 In this article, we focus on a particular loanword in Modern Hebrew: axla (< Arabic (ʔ)aħla), which means ‘sweeter/sweetest’ in Standard Arabic, but has a far more general meaning in Hebrew—more similar to ‘great, awesome.’ This word is considered somewhat non-standard, yet is extremely common in Modern Hebrew. Perhaps the most memorable example of its use is in the name of the ubiquitous hummus brand Hummus Axla, and in its popular slogan shown in (1): (1) kše- ha- xumus axla hakol axla when the hummus great everything great ‘When the hummus is great/axla, everything is great.’

Beyond Lexical Borrowing

There is more to axla, however, than simply a commonly used borrowed lexical item. In the example in (1), axla serves as a clausal predicate, and as such behaves like a normal Hebrew adjective. However, in cases where it modifies a noun, its behavior is more peculiar; whereas adjectival modifiers in Hebrew follow the noun they modify, axla can appear either before or after the noun, as shown in (2) and (3): (2) davar axla thing great ‘a great thing’

1  The glosses provided are for the loanwords as they are interpreted in Modern Hebrew.

when the construction is axla, everything is axla


(3) axla davar great thing ‘a great thing’ The word order shown in (2) is the normal order for modifiers in Hebrew. But as corpus analysis will show, the order in (3) is actually the preferred one for this word. The unusual word order in (3) is the one in Arabic, suggesting that axla was not simply borrowed as a lexical item, but as a construction, that is, along with its particular structure. Before going into the distribution of axla in Modern Hebrew, however, let us explore the Arabic origins of the word and the construction it takes part in. In Arabic, as in Hebrew, adjectives generally follow the noun they modify. The word order associated with axla is thus not the default order in Arabic either, but rather, the result of a general and productive process. Palestinian Arabic, like many other varieties of Arabic, has a morphological template (ʔ)aC1C2aC3, which denotes either a comparative or superlative nominal.2 This is usually an adjective, but since Arabic—like many other Semitic languages— lacks a morphological distinction between adjectives and substantive nouns, the latter interpretation is quite often implied, especially when a definite article is present. This comparative/superlative form in Arabic is also known jointly as the “elative” form (see e.g., Girod 2007; Shahin 2008:532; Cowell 1964:310–315). The derivation of the elative form is as follows. Palestinian Arabic adjectives come in various templates, yet their elative (i.e., comparative/superlative) will always take the form aC1C2aC3, as follows: (4) C1C2iːC3 (5) C1iC2C3 (6) C1aːC2iC3

kbiːr ‘big’ ṣiʕb ‘difficult’ baːrid ‘cold’

→ → →

akbar ‘bigger, biggest’ aṣʕab ‘more/most difficult’ abrad ‘colder, coldest’

Like other Arabic adjectives, the word ħilw/ħilu ‘sweet, beautiful’ derives its elative form in the aC1C2aC3 template, resulting in aħla ‘sweeter, sweetest, beautiful, more beautiful.’ This item was borrowed into Hebrew, but due to the fact that most Hebrew speakers merge /ħ/ and /x/ (Blanc 1968; Yaeger-Dror 1988; Gafter 2014, among others), in Hebrew this elative is usually pronounced axla. The meaning has changed as well—whereas in Arabic the original meaning denotes sweetness,3 and the more general meaning of ‘good’ is an extension, in Hebrew axla has lost its original meaning of sweetness and only maintains 2  The glottal stop is presented here in parentheses, as it is often deleted in Palestinian Arabic as part of a consonantal chain shift. 3  In some Arabic dialects, ħilw/ħilu also carries a secondary meaning of beauty.


gafter and horesh

a meaning of good quality. Furthermore, although it was the elative that was borrowed, Hebrew axla does not denote a comparative or superlative, but rather, its meaning is closer to that of a simple adjective—‘great.’ Of particular interest in this discussion is the syntax of superlatives in Arabic.4 Ryding (2005) provides a detailed breakdown of the syntactic constructions in use for superlatives in Modern Standard Arabic (MSA). Many of these constructions are shared with Levantine Arabic, the group of dialects to which Palestinian Arabic belongs. What follows is a summary of the available constructions in MSA that are shared by Syrian and/or Palestinian Arabic, based on the authors’ knowledge of these dialects, and cross-referenced with Cowell (1964) for Syrian Arabic, and Seeger (2013) for Palestinian Arabic. The examples have been changed to include the superlative adjective aħla. In the examples below, this term is translated in its original MSA meaning, ‘the sweetest.’ As the examples show, Arabic aħla can follow the noun, like most Arabic adjectives, and as such it agrees in gender (as shown in (7) and (8)). However, there is another construction, traditionally understood to coincide with the construct state,5 in which aħla precedes the noun and is invariant (shown in (9)–(10)). (7) al- ʕasal al- aħla the honey (MASC) the sweetest-MASC ‘the sweetest honey’ (8) al- baqlaːwa al- ħulwa the baklava (FEM) the sweetest-FEM ‘the sweetest baklava’ (9) aħla ʕasal sweetest-MASC honey (MASC) ‘the sweetest honey’

4  The syntax of comparatives is not discussed here, as it is not relevant to the use of axla in Modern Hebrew. 5  Wright’s (1964:218, 226) analysis of superlatives is couched within his chapter on “the status constructus and the genitive,” and he is unequivocal about superlatives such as those in the aC1C2aC3 template being substantive nouns, not adjectives, as they “govern” the genitive nouns that follow them.

when the construction is axla, everything is axla


(10) aħla baqlaːwa sweetest-MASC baklava’ (FEM) ‘the sweetest baklava’ Note that in examples (9) and (10), there is no overt definite article (al- in Arabic), but the phrases in both examples are definite by virtue of their inclusion in a construction known in Arabic as ʔiḍaːfa (literally, ‘addition’), which is a synthetic possessive phrase similar to the Hebrew construct state (smixut). The idiomatic translation, therefore, encapsulates this definiteness with the English article the. As illustrated in example (1), aħla/axla in both languages can serve as a clausal predicate. In Arabic we find such examples as (11)–(12): (11) al- sukkar aħla min al- milħ the sugar sweeter from the salt ‘Sugar is sweeter than salt.’ (12) al- ʕasal aħla ši bi- l- dunya the honey sweetest thing in the world ‘Honey is the sweetest thing on earth.’ We shall return to this issue and its significance when we discuss parallel Hebrew data.

Corpus Analysis

We argue that it is the construction in (9)–(10), with an invariant aħla preceding the noun, which was borrowed into Hebrew, resulting in forms that feature a borrowing not only of a word but also of the word order of a particular construction—such as axla davar in (3). However, while forms like the one in (3) are clearly attested, the extent of their current use is not empirically validated. In order to assess the distribution of axla, we searched the Israblog Corpus (Linzen 2010), which was extracted in September 2008 from the Israeli bloghosting site (accessed September 15, 2014). The corpus comprises blogs written between 2005 and 2008. This corpus is particularly suitable for our purposes, since its source materials are blogs, and therefore it features a variety of registers but, importantly, has ample representation of a less formal writing style, which is where we expect to find axla.


gafter and horesh

We searched for the first 500 tokens of the word axla6 to determine that speakers indeed use the axla NOUN order, and not just the NOUN axla order (which, again, would be the unmarked order for a noun and adjective in Hebrew). The results show that they undoubtedly do. Out of these tokens, 63 percent (313 tokens) are in the axla NOUN order (exemplified in (13)–(14)), and only 2 percent (8 tokens) of the hits are in the NOUN axla order (exemplified in (15)–(16)). (13) nekave še- tihye axla mesiba we’ll.hope that great party ‘We’ll hope that it’ll be a great party’ (14) ve- gam ha- nof yafe šam, axla nof and also the view pretty there great view ‘and the view there is pretty too, a great view’ (15) kodem kol, nose axla first all topic great ‘first of all, great topic’ (16) xaxa hu benadam maze axla haha he person so great ‘Haha, he’s such a great person’7 The data show that for these writers, axla NOUN is not only common, but the preferred order: axla is not reanalyzed as a Hebrew adjective, but rather, is used as part of the entire borrowed construction. Thus, the Hebrew axla maintains an interesting combination—the lexical item was borrowed through the aħla NOUN construction, and tends to maintain that order, although it is not the basic one for adjectival modification in Hebrew, nor is it a generalized construction. It is important to point out, however, that while the word order is superficially similar to that of Arabic, the resulting construction in Hebrew is different from the Arabic one. In the Arabic construction aħla NOUN (which again, in Arabic, is general to adjectives of this template, not restricted to a 6  In Hebrew orthography, the word axla (‫ )אחלה‬can be ambiguous with various verbal forms such as exle (‘I will get sick’) and ex(a)la/ix(a)la (‘she wished’). We did not consider such tokens. 7  This example also shows the adjectival behavior of axla, in that it is able to get an adverbial modifier.

when the construction is axla, everything is axla


particular lexical item), the noun phrase is definite as well, and the meaning is technically ‘the sweetest thing,’ not ‘a sweet thing’ (as shown in (9)–(10)). In Hebrew this meaning is lost; it has neither a comparative nor a superlative meaning, nor is it definite. In addition to its use as a modifier, axla is also frequently used as a predicate. In our corpus, 25 percent of the hits were predicative uses, in which axla does not modify a noun at all, as exemplified in (17)–(18).8 As mentioned before, the Arabic aħla can be a predicate as well, but it is used differently. In Arabic, it is a comparative in its indefinite form and a superlative in its definite form, but once again, this meaning is lost in Hebrew. (17) gam ha- xaverim šela me- ha- becefer ha- kodem also the friends hers from the school the previous hegiu ve- hem mamaš axla arrived and they really great ‘her friends from her previous school also came and they were really axla’ (18) haya nexmad, ha- hofaot hayu axla was nice the shows were great ‘it was nice, the shows were axla’ Examples (17)–(18) show another important way in which axla shows that it does not behave like a normal Hebrew adjective. Whereas Hebrew adjectives agree in gender and number, axla does not (the subject in both cases is plural, requiring the agreement suffix -im/-ot). This is not an exception, but the rule: in our corpus, axla never showed agreement, regardless of syntactic position. It is invariant as a predicate, in the axla NOUN order, and in the NOUN axla order. This is different from the pattern in Arabic shown in (7)–(8), in which aħla agrees when used as an adjectival modifier. This supports our proposal that axla was borrowed into Hebrew along with the axla NOUN construction, and suggests that it was nativized as a single, invariable, lexical item, such that even when the more typical Hebrew word order is used (NOUN axla), it does not behave like a regular Hebrew adjective. We have argued that the greater use of axla NOUN compared to NOUN axla suggests that for most speakers, axla was not reanalyzed as an adjective, and 8  The use of axla as a predicate is also very common in short answers to questions. In fact, the first definition given to axla in the popular online Hebrew dictionary Milog is: “A positive answer when someone is asked how they are doing.” %D7%90%D7%97%D7%9C%D7%94 (accessed February 5, 2015).


gafter and horesh

rather, the construction was borrowed wholesale from Arabic. It is interesting to note that axla, which was borrowed along with the Arabic word order, is not showing signs of losing its special distribution. The Israblog corpus represents a variety of speaker/writer ages, but due to the nature of its source material, it is highly skewed towards younger speakers—almost all are younger than 40, and many are in their teens—reflecting the demographics of speakers likely to be writing blogs. Thus, the sample over-represents younger speakers, most of whom are likely to have little to no knowledge of Arabic, let alone the complexities of its syntax. If this sample shows such a preference for axla NOUN, it is safe to assume that it is not the case that early borrowers from Arabic borrowed the construction and younger generations are just using the lexical item based on Hebrew syntax. Rather, the borrowed structure appears to be here to stay.

A Note on šel Constructions

Finally, of the 500 tokens from the Israblog corpus, 10 percent (51 instances) are neither axla NOUN nor NOUN axla. Rather, they use the general possession/ genitive marker šel between axla and the noun, as exemplified in (19)–(20): (19) naxal tsafit, axla šel maslul la- yom ha- rišon river tsafit great of route for.the day the first ‘Tsafit river, an axla route for the first day’ (20) pašut axla šel xofeš simply great of vacation ‘simply a great vacation’ This use of šel is a Hebrew innovation and does not come from the Arabic. Recall that in Arabic the aħla NOUN construction is a construct state; Arabic also has an analytic possession which can be used instead of the construct state forms, such as tabaʕ/šeːt in Palestinian Arabic,9 but constructions such as aħla tabaʕ/šeːt NOUN (parallel to Hebrew axla šel NOUN) have not, to our knowledge, been documented. Rather, axla here seems to be following a

9  These forms are increasing in their distribution in speech communities that are in close contact with Hebrew, such as Jaffa (Horesh 2014).

when the construction is axla, everything is axla


Hebrew pattern of using a noun as a modifier.10 This is shown in (21)–(22) with the nouns kesem (‘charm’) and motek (‘sweetness’): (21) kesem šel yalda charm of girl ‘a charming girl’ (22) motek šel xatul sweetness of cat ‘a sweet cat’ (23) xara šel seret shit of movie ‘a shitty movie’ The common use of axla šel NOUN gives further evidence that axla was not reanalyzed as a Hebrew adjective, as this strategy is used for nouns, not adjectives. This raises the question: since axla does not show agreement, and can participate in this clearly nominal construction, was it simply reanalyzed as a noun? That may very well be the case, but it is important to point out that axla does not exhibit the patterns of a typical Hebrew noun either. First, in our corpus data, the two Hebrew modification strategies, NOUN axla and axla šel NOUN, are far less common than the borrowed construction, axla NOUN. Furthermore, in examples such as (19)–(20), the šel can be freely dropped, resulting in the very common axla NOUN form, whereas in examples (21)–(22) the šel cannot be dropped (*motek yalda, *kesem xatul),11 and these nouns cannot have structure equivalent to axla NOUN. Interestingly, in (23), the šel can be dropped, resulting in the perfectly grammatical xara seret, but it is our impression that examples of nouns such as xara, which can allow for such a construction, are quite rare, whereas (21)–(22) appears to be the general p ­ attern.12 Thus, while axla does arguably appear more nominal than adjectival in its syntactic properties, it also appears not to have the typical distribution of any major lexical category of Hebrew.

10  For more on the origins of this use of šel in Hebrew, see Shatil 2015. 11  For a more complex and somewhat modified analysis of Hebrew šel constructions based on prosody, see Botwinik & Albert 2012. 12  It is interesting to point out that xara is also a borrowing from Arabic, whereas the other two examples are native Hebrew words.


gafter and horesh

Conclusion In conclusion, the data presented here clearly show that axla is not a simple lexical borrowing, as it was borrowed from Arabic to Hebrew with a good deal of syntactic structure intact. This makes axla an interesting data point in light of the broader theoretical discussion regarding different types of borrowings and language contact (Thomason 2001). With respect to Arabic, Thomason (2007) makes a distinction between the contact Arabic may have with other languages and its interference with these languages’ phonologies, as opposed to “hints of more extensive structural interference,” which she argues only occur in Arabophone areas and areas that border on Arabophone areas. Modern Hebrew falls squarely into the latter case.13 It clearly fits the geographical classification, and, as shown here, axla exhibits structural interference, since the borrowing includes a rather complex syntactic structure in Arabic, which it does not share with Hebrew: a genitive construct, which does not denote a semantic genitive, but rather a superlative. Of course, axla is just one example, but its syntactic behavior is indicative of the extent of contact between Hebrew and Arabic, with structural influence that goes well beyond the occasional borrowing of a lexical item.14 References Blanc, Haim. 1968. “The Israeli Koine as an Emergent National Standard.” In Language Problems of Developing Nations, eds. J. Fishman, C.A. Ferguson, & J. Das Gupta. New York: J. Wiley & Sons, 237–251. Botwinik, Irena & Aviad Albert. 2012. “The Occurrence of šel in Hebrew Binominal Constructions: A Prosodic Account.” Paper presented at the 28th Annual Meeting of the Israeli Association of Theoretical Linguistics, Tel Aviv University. Cowell, Mark W. 1964. A Reference Grammar of Syrian Arabic. Washington, DC: Georgetown University Press.

13  In fact, unlike some of the cases Thomason describes in which Arabic phonology interferes with that of the languages with which it comes in contact (e.g., Wolof, Hausa, Turkish), in the case of Modern Hebrew, there is statistical evidence (Horesh 2014) of Hebrew interference in Arabic phonology. 14  For a detailed account of the contact between these two languages, see Henkin-Roitfarb 2011.

when the construction is axla, everything is axla


Gafter, Roey Jaakov. 2014. The Most Beautiful and Correct Hebrew: Authenticity, Ethnic Identity and Linguistic Variation in the Greater Tel Aviv Area. PhD dissertation, Stanford University. Girod, Alain. 2007. “Elative.” In Encyclopedia of Arabic Language and Linguistics, Vol. 2, eds. Kees Versteegh et al. Leiden: Brill, 12–16. Henkin-Roitfarb, Roni. 2011. “Hebrew and Arabic in Asymmetric Contact in Israel.” Lodz Papers in Pragmatics 7.1: 61–100. DOI: 10.2478/v10016-011-0004-7. Horesh, Uri. 2014. Phonological Outcomes of Language Contact in the Palestinian Arabic Dialect of Jaffa. PhD thesis, University of Essex. Linzen, Tal. 2010. Hebrew Statistical Linguistics Using a Morphologically Analyzed Blog Corpus. Paper presented at the Israeli Seminar on Computational Linguistics, Tel Aviv, Israel, June 2010. Ryding, Karin C. 2005. A Reference Grammar of Modern Standard Arabic. Cambridge: Cambridge University Press. Seeger, Ulrich. 2013. Lehrbuch des Palästinensischen Arabisch: Der Dialekt der Städter. Wiesbaden: Harrassowitz Verlag. Shahin, Kimary N. 2008. “Palestinian Arabic.” In Encyclopedia of Arabic Language and Linguistics, Vol. 2, eds. Kees Versteegh et al. Leiden: Brill, 526–538. Shatil, Nimrod. 2015. “The Nature and Diachrony of Hebrew Quality Pseudo-Partitives: Are They a Calque from the Contact Languages?” Journal of Jewish Languages 3.1–2: 301–308. Thomason, Sarah G. 2001. Language Contact. Edinburgh: Edinburgh University Press. ———. 2007. “Language Contact.” In Encyclopedia of Arabic Language and Linguistics, Vol. 2, eds. Kees Versteegh et al. Leiden: Brill, 664–674. Wright, William. 1964 (reprint). Grammar of the Arabic Language; translated from the German of Caspari and edited with numerous additions and corrections. Vol. 2. London: Cambridge University Press. Yaeger-Dror, Malcah. 1988. “The Influence of Changing Group Vitality on Convergence toward a Dominant Linguistic Norm: An Israeli Example.” Language & Communication 8.3: 285–305. Zuckermann, Ghil’ad. 2003. Language Contact and Lexical Enrichment in Israeli Hebrew. New York: Palgrave Macmillan.

Index A’-movement 69 Absolute  17, 282–283, 285–286 addressee 105, 109–110, 115–118, 137, 158, 317 adjacency 138, 140–142, 207 adjective degree comparison  269 adjunct depictive 10, 51, 76–83, 85–86 resultative 18–19, 325, 327, 329–332 adverb 10, 15, 18, 20, 43, 45, 51–52, 57, 65, 76, 78–80, 82–86, 117, 178, 189, 215–216, 217, 219–225, 227, 270, 284, 290, 328, 332, 336, 342 affix 68, 281–282, 285 agreement 38, 43, 68, 173, 253–255, 284–285, 294, 301–302, 310, 327, 343, 345 Akkadian 250–251, 257 Aktionsart 52–53 ambiguity 176 analysis 1, 6–7, 13, 27, 33–34, 45, 62–64, 69, 81, 101, 108, 123, 125, 127, 137, 139, 164, 176–178, 198, 200, 204, 224, 260, 268, 286, 293, 325–326, 331, 333–334, 339–341, 345 n. 11 anchoring 146–147, 157–160 anaphoric 29 Arabic Arabic dialect 9, 11–12, 49–51, 53–54, 57–58, 61, 113, 119–124, 126, 154, 251, 318, 322, 339 n. 3, 347 Judeo-Arabic 7, 12–13, 56, 146–148, 151–152, 154–155, 158–159, 228, 236–237, 251, 318 Palestinian Arabic 26, 50, 52, 55, 58, 112 n. 1, 113, 116, 119 n. 3, 120–123, 322, 338–340, 344, 347 Standard Arabic 119, 338, 340, 347 Aramaic 2, 7, 80 n. 4, 112–113, 124–126, 171, 187 n. 8, 190 n. 14, 218, 235 n. 9, 282 n. 1, 316 article definite article 29, 233, 270, 277, 279, 281, 291–292, 296, 339, 341 indefinite article 29 Ashkenazi 3, 93, 94 aspect 9, 49–51, 54, 58, 170 assertion 26, 30, 104–105, 184, 194

Assyrian 250 attitude marker 112–119, 122, 125 attributive 19, 36–37, 229–230, 232, 234, 284, 294 auxiliary 9, 49, 51, 54, 70, 95 auxiliation 57, 58 Babylon 2, 93, 172, 216–218, 223, 250, 272, 284, 310, 316, 320 Babylonian 93, 172, 216–218, 223, 250, 272, 284, 310, 316, 320 Belarus 249 binyan 66, 92, 197–198, 200, 202–203 borrowing 8, 15, 18, 19, 25, 122, 134, 158, 169, 224, 337, 341, 345, 346 bounding 6, 69–70 Brazilian Portuguese 73–74 calque 16, 25, 29–30, 95, 171, 173, 235, 245, 247, 249, 256, 301, 317, 347 canonical non-canonical 11, 128–130, 139, 143 case accusative 8, 25, 27–34, 63, 66, 199, 202, 259, 261, 263, 265, 267, 316 dative 18, 55, 61, 287, 309, 310–312, 314–322 genitive 8, 15, 25, 29–32, 34, 281, 286–287, 293, 295, 340, 344, 346 genitive of negation 8, 29, 34 nominative 8, 27–29, 149–150 vocative 57, 59 change of state 325, 329, 331 Christian 2–3, 123–124, 127, 179, 270, 321–322 clause clausal predicate 8–9, 23, 338, 341 embedded clause 10, 92, 166 free relative 13, 16, 163–165, 169–170, 177–178, 258–259, 268 interrogative clause 13–114, 118 main clause 10–11, 76–78, 81–82, 84, 110, 119, 165 matrix clause 107, 166, 168 clause periphery 11

350 clause (cont.) relative clause 69, 79, 130, 164, 169, 177–178, 258–259, 268 subordinate clause 83, 90, 95–96, 98 cleft pseudo-cleft 47 collocation 239, 292, 325–328, 332–335 comitative 14–15, 204–207, 209, 213 comparative 14, 16, 19, 21, 52, 61, 64, 169, 179, 269–271, 277, 279, 288–290, 308, 320, 337, 339–340, 343 comparison 16, 65, 67, 128, 137, 171, 218, 236, 269–270, 277–278, 325 complement 5, 10–11, 13, 16, 66, 69, 82, 90, 92, 101, 104, 106–108, 110, 129, 163, 165–167, 171–176, 185, 204, 206, 232, 259, 261–263, 267–268, 278, 287, 335 complementizer 5, 10–11, 16, 101, 104, 106–107, 110, 166, 171 n. 16, 172, 174, 259, 261–263, 267, 287 composition 11, 18, 290, 294, 303, 326, 328, 336 compositional 18, 290, 294 n. 12, 326, 336 compositionality 336 conditional 14, 133, 166, 182, 188, 314, 316–317 conjunction 12, 129, 132, 135, 136, 168 construction 5–13, 16–20, 25–35, 37, 39, 45, 47, 50–53, 57–58, 60–66, 68–69, 72–73, 76–78, 80–81, 84–85, 89–90, 92–93, 95–96, 98–99, 101–103, 107–108, 111–114, 118–120, 122–123, 125–126, 128–129, 131–132, 137–143, 146–151, 154–160, 163–165, 167–171, 173–178, 193, 228–231, 234–235, 239, 245–252, 254–258, 269–274, 276–279, 281, 283–284, 286, 288, 290–296, 301–306, 309–322, 325–327, 329–331, 333–337, 339–346 constructional idiom 325–327, 329, 331, 333–335 construct state 216, 228, 232, 239, 281–283, 286, 289, 295, 337, 340–341, 344 contact language 1, 4, 6–7, 9–12, 14, 16–17, 33, 36, 41, 63, 73–74, 122, 126, 139, 228, 233, 239, 258–259, 266–267, 269–270, 278–279, 281, 286, 296, 301, 306, 317, 333, 347 context 15, 22, 49, 51, 61, 65, 68, 72–73, 103–105, 110, 112, 114–115, 130–131, 136,

Index 140, 143, 168, 179–180, 182–185, 188, 190–191, 200, 203, 220–221, 224, 231, 246–247, 252, 265, 272, 274, 278, 314, 320, 332–333 copula 8–9, 20, 36–44, 46–47, 70 copy 63–64, 66–67, 69–71, 235 corpus 2, 102, 106, 130, 144, 200, 211, 226, 229–230, 238–240, 255, 260, 264, 270, 315, 337, 339, 341, 343, 344, 345, 347 creole 1, 5, 19–21, 190, 193 creolization  4, 22 data 7, 9 n. 16, 58, 64, 67, 94, 123, 128 n. *, 147, 185, 189, 197–198, 200, 203, 229, 251 n. 7, 260, 262–265, 271, 306, 318, 320, 341, 342, 345–346 dative affected dative 309, 310 n. 1, 312, 314, 316–318 co-referential dative 309–310, 312, 314, 315–317 discursive dative 18, 309–312, 314, 316–319 ethical dative 55, 309–310 n. 1, 312, 322 non-core dative 61, 309–312, 314–317, 319–320 definiteness 16–17, 42 n. 1, 144, 265 n. 9, 281–288, 292, 301–302, 341 degrammaticalization 279, 281, 285, 295–296 degree 16–17, 140, 207–208, 210, 269–272, 274 n. 2, 277, 284, 290 demonstrative 9, 12, 16, 36–37, 41–42, 44, 147, 152, 153 n. 13, 154, 156, 216 n. 1, 246–247, 251–256, 282 n. 1, 295 depictive 10, 51, 76–83, 85–86 derivation 339 descriptive 26, 103, 205 determiner 16–17, 43, 192, 235 n. 10 diachronic diachronic change 21, 111, 180, 189, 226, 271 diachronic development 215, 250 n. 5 diachrony 22, 257, 301, 347 dialect 7, 9, 11–12, 30, 32, 49–52 nn. 4, 7, 53–55 n. 14, 57–58, 61–62, 107 n. 4, 111–113, 119 n. 3, 126, 129, 154 n. 14, 160,

Index 250–252 n. 8, 257, 270, 287 n. 4, 316, 318, 320, 322, 339 n. 3, 340, 347 discourse discourse marker, DM 12, 15, 129, 133–134, 144–145, 215–216, 221–227 dislocated constituent 12 dislocation left dislocation 136, 144, 149, 152–153, 156 right dislocation 12, 129, 136, 145, 152, 154, 156 distribution 10, 13–14, 17, 19, 36, 44, 46, 123–124, 132, 135, 145, 148, 177, 180–183, 185, 188 nn. 11, 12, 189, 190 n. 14, 191–193, 197–198, 200–202, 207, 218, 229, 238, 245, 247, 270, 273–274, 281, 282–286, 290, 292, 303, 306, 321, 339, 341, 344–345 doubly-marked interrogative DMI 112–116, 118–120, 122–126 downward entailment 183 n. 4, 185 n. 6 Eastern Europe 2–3, 7, 13, 73, 110, 169, 179, 208, 292, 296 echo 12–13, 114, 138, 146–147, 154–156, 159–160 elative 16–17, 279, 337, 339, 340, 347 emphatic state 17, 281–283, 285–286 English 4, 9, 13, 15, 18–19, 21, 32, 37, 47, 52 n. 5, 54, 74, 77, 84 n. 9, 103–104, 120 n. 4, 132, 136, 144, 165 n. 5, 6, 168, 174–175, 177–178, 186 n. 7, 187–188, 192, 202–203, 205 n. 4, 211, 214, 216, 226, 247 n. 2, 248, 256–257, 268, 297, 301–302, 304, 307, 309 n. *, 321–322, 325–327, 329–334 n. 15, 335–336, 341 entailment 183 n. 4, 185 n. 6 environment 109, 182–183 n. 4, 185, 191, 218, 220 n. 8, 295, 317 entailment downward entailment 183 n. 4, 185 n. 6 European 13, 16–17, 73, 107, 122, 146 n. 1, 169, 190 n. 15, 207–208, 210, 248, 305, 311 evaluative impact 301, 303 exclamative 13, 163, 165 existential 8, 25–27, 30–35, 183, 225 exponent 14 extraction 69 feature feminine 29, 38, 43, 92, 252, 253, 302

351 gender 29, 37, 147, 252–253, 256, 282, 340, 343 masculine 29, 37–38, 43–44, 92, 252–254 n. 10, 302 neuter 43–44 number 13, 37, 68 n. 4, 71, 73, 139, 147, 168, 175, 186, 201, 208, 246–247, 249, 255–256, 281–282, 306, 328 n. 5, 335, 343 plural 6, 29, 92, 249, 254–256, 307, 343 singular 16, 29, 37, 38, 43–45, 63 n. 1, 73, 92, 138, 255–256, 291 float 16 focus 12, 46–47, 63, 73, 75, 78, 89, 102, 129, 131, 135, 148, 169, 215, 226, 259, 338 force illocutionary 101, 103, 105, 112, 115, 126, 132, 144 perlocutionary 105 form logical 26, 109–110, 181, 183 n. 4, 190, 198 phonological 202, 223 n .11, 235 n. 10, 282 n. 1, 347 formation 5–6, 20, 59, 99, 160, 194, 224 n. 14, 269, 271, 279, 289, 296, 305, 331 free relative, FR 13, 16, 163–165, 169–171, 177–178, 258–259, 260 n. 1, 268, 321 French  21, 93–94, 143, 173, 175, 189 n. 13, 193, 208, 220 n. 8, 270, 277, 292, 301, 305–306, 317–318, 330 n. 8, 336 fronting 63–68, 70–74 function 8–12, 15, 18, 29, 36–38, 41–42, 44, 49–59, 61, 72 n. 7, 75, 77–79 n. 2, 80, 82–83, 85–86, 90, 93, 107, 112–113, 115, 125, 128–129, 131–134, 136–139, 143–145, 154, 160, 165 n. 6, 170 n. 15, 178, 180–182, 194, 198, 207, 215–216, 218, 221–222, 224–225, 227, 229, 230, 232, 234, 238–239, 252, 271 n. 1, 272–274, 278–279, 281–282 n. 1, 318, 321, 332, 336 functional 17, 19, 50, 58, 61, 75, 128–129, 131–132, 136, 194, 198, 216, 278, 307, 336 grammar 5–6, 13, 21, 34, 47, 52 n. 4, 62, 65, 74–75, 86, 93, 95–96, 99, 113, 176, 224 n. 13, 225, 309, 311, 316 grammatical 7–9, 16, 43–45, 51, 54, 56, 58, 89, 96, 182–183, 245, 251 n. 6, 255–256, 269, 278, 285, 302, 345

352 grammaticalization 9, 50, 53, 58, 216, 219, 225, 269, 278–279, 282 n. 1, 285, 295 Gungbe 63, 67, 73 Halakhic 2, 217, 328 n. 6 Hasidic 2, 281, 286, 292–294 Haskala 3, 113, 174, 290 hearer 109, 152 Hebrew Biblical Hebrew 2–3, 8–10, 12, 15, 17, 37, 40, 46, 81, 132–135, 137, 168, 170–171 n. 16, 185–186n7, 197, 202, 205–206, 208, 216, 218, 232–233, 235–236, 239, 247, 258, 261, 263, 267, 269, 274, 309, 328, 330 n. 9, 338 Classical Hebrew 49, 54 n. 10, 63–68, 71, 76, 78, 81, 85, 113, 125, 180, 182, 185, 193, 204–205, 207, 215, 236, 238, 263, 267, 269, 278, 281–285, 291, 325, 335 Colloquial Modern Hebrew 49–55, 57–58, 112–113, 126, 209, 282,–284, 312 Hasidic Hebrew 2, 281, 286, 292–294 Late Rabbinic Hebrew 2, 262, 264 n. 7 Maskilic Hebrew 3, 208–209, 288 Medieval Hebrew 2, 7, 13, 15, 17, 170, 174–175, 208–209, 218–219, 220, 223, 225, 236–237 n. 11, 238–239, 262, 270–271, 281, 286, 288–289, 291–292 Mishnaic Hebrew 2, 9, 16, 137–138 n. 14, 168, 171, 197, 206, 216, 230 n. 4, 233 n. 5, 235, 238–239, 245–247, 253, 255–256, 258, 261, 267, 281, 286–287 n. 4, 293–294, 309, 311, 315–317 n. 17, 327 Rabbinic Hebrew 2, 11, 13, 17, 80, 125–126, 133, 176 n. 22, 187 n. 9, 216–218, 223, 237, 262, 264 n. 7, 269, 272, 282, 309 316, 328 Revival Hebrew 3, 13, 15, 17, 217–218, 220–223, 228–232, 238–239, 333 n. 14 Hitpael 197–203 identity 40, 137, 165 imposition 123, 126 indefiniteness 301–302 infinitive absolute 63 n. 1, 65–66, 71, 73 instrumental 15, 122, 204–205, 207–209 internal development 8, 30, 32–33, 67, 98, 173, 258, 286, 295, 304

Index interrogative 13–14, 101–108, 110, 112–114, 118, 126, 134, 154, 164–165, 170, 218, 259, 260–261, 263, 267, 274, 277 intonation 105, 113–114, 131 n. 5 Islam 2 island 5, 70 Israel 1, 4, 25, 36, 49, 55 n. 12, 58, 67, 76, 79, 89, 94, 112–113, 123, 128, 146, 174, 150, 163, 180, 197, 204, 210, 215, 218 n. 4, 228–229, 237, 245, 254, 258, 269, 281, 289, 301, 309, 318, 325–338 Israeli periphery 12, 146 Jewish 1–7, 19, 21, 33, 61, 71, 73, 86, 113, 116, 121–124, 126–127, 133, 159, 160, 163, 168, 177–178, 186 n. 7, 193–194, 202–203, 208 n. 6, 209–210, 222, 229 n. 1, 239, 240–241, 247, 260, 268, 277, 279, 289, 296, 305 n. 6, 307, 309–311, 320, 322, 338, 347 Jewish Neo-Aramaic 7, 124 Judaism 2, 73, 75 Judeo-Arabic 7, 12–13, 56, 146–148, 151–152, 154 n. 14, 155, 158–159, 228, 236, 237, 251, 318 Judeo-Spanish 3, 7, 10–11, 89, 90, 94, 96–101, 107–108, 110–111, 126, 177, 204, 277–278, 301, 306–307 Kurdish 113, 247 language language acquisition 7, 20–21 colloquial language 167, 209, 231 contact language 1, 4, 6, 7, 9–12, 14, 16–17, 33, 36, 41, 63, 73–74, 122, 126, 139, 228, 233, 239, 258–259, 266–267, 269–270, 278–279, 281, 286, 296, 301, 306, 317, 333 n. 14, 347 First Language L1 4, 7, 338 heritage language 7 language internal development 67, 173, 286 native language 1, 2, 5–7, 10, 17–18, 121–122, 180–191, 252, 309, 311, 317 Second Language L2 4 language shift 5, 8 standard language 119 n. 3 target language 5 war of languages 4 n. 12, 305



lexical lexical borrowing 19, 337–338, 346 lexical item 13, 19, 36, 71, 189 n. 13, 190 n. 14, 281–284, 286, 338–339, 342–344, 346 lexicon 62, 177, 207, 213, 270, 307, 321, 335 lexifier 1, 4, 19 Lithuanian 248 loanword 57 n. 17, 338

negative concord 11, 13, 101, 104, 108–111, 168–169, 180–182, 185, 190 n. 15, 191, 193–194, 226 polarity 13, 15, 21, 101–102, 104, 108–111, 127, 180, 182, 185, 191–194, 217, 226 Neo-Aramaic 7, 11, 112–113, 123–127, 247, 257 Nitpael 197–203, 213 North Africa 2, 12, 126, 146, 147, 155, 159

marker focus marker 12, 129, 131, 135 material constitution 15, 228–230, 232, 238–239, 284 metaphoric 18 metonymy 206 Middle East 2, 146–147 Mishnah 2, 125, 135, 137–138, 172–173, 177, 187, 193, 216–217, 223, 226, 235–237, 240–241, 248, 261, 263–265, 268, 287–288, 295, 315, 322 modality epistemic 132, 182 metaphysical 104 spoken modality 11–12, 18, 128–129, 131–133, 135–136, 138–141, 143 written modality 128, 132, 136, 139 modifier 9, 16, 233, 271, 284, 302, 338, 339, 342 n. 7, 343, 345 mood  finite 65–66, 69, 71–72 infinitive 9, 20, 63–65, 66–67, 70–74, 95 non-finite 65–66, 71–72 pseudo-infinitive 74 Morocco 150, 287 morpheme 58, 68, 164, 168, 173, 287 morphology 8, 19, 63, 166, 172, 256, 279, 319, 335

object 19, 21, 25–26, 29–30, 32, 33–34, 66, 79, 144, 149, 164, 181–182, 198, 204 n. 1, 232, 234, 258–265, 267, 272, 329, 330, 333–335 obsolete 163–164, 167, 175–176, 276 Old Testament  2, 10, 177, 187, 188 n. 11, 213

negation expletive negation 163, 169, 176, 178 paratactic negation 164 redundant negation 168 sentential negation 10, 67, 78, 82, 169, 181, 189, 192–193 superfluous negation, Super Neg 13, 163–165, 168, 170, 173, 176

Palestine 1–4, 7, 19, 33, 98, 116, 229, 277, 305 n. 6, 311 n. 5 parameter 5–7, 15, 17, 141–142 macro-parameter 17–18 micro-parameter 18 parenthetical 12, 94, 129, 131–132 participle adverbial participle 10, 76–78, 82–86 past participle 202 partitive pseudo-partitive 301–307 quality pseudo-partitive 301–307 path of development 13, 176 periphery left periphery 12, 63, 66, 69, 128–129, 132–136, 139 right periphery 11–12, 128–137, 139–141, 143–145 periphrastic 16, 50–53, 58, 281, 286–287, 293, 295 phonology 3, 148 n. 3, 346 n. 13 phrase adjective phrase, AP 38, 77, 284, 286, 290, 329 adverbial phrase, AdvP 57 noun phrase, NP 16, 159, 243, 301 determiner phrase, DP 43–44, 192, 310, 314, 316 prepositional phrase, PP 30–33, 215–217, 219, 221, 223, 225, 229, 329 verb phrase, VP 63–67, 70–72, 316

354 polarity negative 11, 13, 15, 101, 104, 108–109, 111, 180–183, 185, 187, 189, 191, 193–194, 217 opposite 101–102 positive 11 Polish 13–14, 16, 25, 29–31, 33, 41, 129, 139–141, 143, 165, 169, 176–177, 180, 189–192, 194, 223–224, 258, 266–267 polysemy 206, 215 possessive 8, 15, 17, 19, 25–27, 30, 33–35, 149–150, 204–207, 209, 238, 293, 296, 301–302, 312, 316, 341 pseudo-cleft 36, 45–48 pragmatics 11, 65, 115 predicate complex predicate 50, 59 predicate fronting 63–65, 67, 69, 71, 73 main predicate 10, 78, 79 n. 3, 80, 85 nominal predicate 38, 42, 44 predicate nominal 36–40, 42, 44–45, 233 nominal sentence 36–40, 42, 44–45, 220 n. 7, 224 n. 13 secondary predicate 10, 76–82, 85–86, 229–230, 232, 234–235, 239 verbal predicate 63–65, 67, 73 prefix 18, 45, 65, 89, 92, 224 n. 14, 282 n. 1, 283, 286, 287, 290 preposition comitative preposition 14–15 dative 18, 310 goal preposition 18 instrumental 15 locative 15, 224 non-reversible 15 possessive preposition 19 reversible 14–15 source preposition 15 prescriptive 7, 15, 17, 204–205, 218, 272, 303, 312 n. 6 prescriptivist 175 presupposition 11, 101, 112, 115, 118, 119, 125–126, 220, 265 n. 9 productivity 325, 333 pronoun anaphoric 29 cataphoric 137–138 deictic 12, 184 n. 5, 252

Index demonstrative 9, 12, 16, 36–37, 41–42, 44, 147, 152–154, 156, 216 n. 1, 246–247, 251–256, 282 n. 1 interrogative 13–14, 101–108, 110–115, 117–119, 126, 134, 154, 164–165, 170, 177–179, 218, 259–261, 263, 267, 274, 277 personal 8–9, 12, 34, 147–149, 154–155, 192, 207, 270 pointer 37, 41, 44 pron 230, 290–292 pronH 38, 40, 44 n. 4 pronZ 38–41, 46 reciprocal 16, 198–199, 202, 210–211, 245, 247, 249, 251–252, 255–256 reflexive 19, 198–199, 202, 257, 310, 312, 316, 321, 325, 327, 329, 331–333 resumptive 12, 147–149, 154 pronominal 12, 27, 38, 42, 43, 137, 146–148, 246, 252, 254, 314 property 39 prosody 345 n. 11 pseudo-partitive 301–302, 304–306 pseudo-infinitive 70–71 quality pseudo-partitive 301–302, 304–306 quantifier 13, 180 n. 1, 181–182, 188 n. 11, 216, 218 queclarative 102, 106 question content question 112–113, 115–119, 122, 123, 125 question intonation 105, 113–114, 131 n. 5 rhetorical question 11, 13, 101–102, 104–105, 112, 114–117, 125, 163, 165, 170–171 wh-question 66 yes-no question 106, 113, 117–118, 125 question under discussion, QUD 105, 112, 115–116, 118–119 reanalysis 7, 13, 176, 224, 286, 293, 326 reciprocal 16, 198–199, 202, 210–211, 245, 247, 249, 251, 252, 255–256 reference 51, 113, 170 n. 15, 191 n. 19, 224 n. 14, 229 n. 1, 237 n. 11, 269–270, 282 n. 1, 287 n. 4, 304 n. 2 referent 39, 44, 137–138, 152, 165, 265 n. 9, 266 n. 10, 316, 329

Index reflexive fake reflexive 325, 329, 333 register 7, 51, 245–247, 256, 283, 305, 341 relative free relative 13, 16, 163, 164, 165, 169, 170, 258, 259 relative clause 69, 79, 130 n. 2, 164, 169, 258, 259 relexification 1 n. 3 reset 6–10, 14, 18–19 resultative 18, 19, 325, 327, 329–333 revival of Hebrew 228, 269, 306 Revival period 37, 169, 258–260, 262–263, 265, 267, 273, 319 rhetorical question 11, 13, 101–102, 104–105, 112, 114–117, 125–126, 163, 165, 170–171 Russian 3, 13–15, 25, 29–33, 36, 41–46, 58 n. 21, 67, 73, 76, 78, 82–85, 89–90, 95, 129, 134, 139, 141, 143, 166, 169, 173–176, 180, 189–192, 223–224, 238–239, 263, 266–267, 278 scope 10–12, 78, 82, 108–109, 120 n. 4, 131 n. 5, 135–136, 182–183 semantic bleaching 50, 53, 58 semantics 11, 19, 37, 113, 115, 182, 189, 205, 215, 220, 222, 282 Semitic 11, 17, 26, 68 n. 4, 89, 113, 124–125, 246, 250, 282, 311, 318, 339 semi-lexical auxiliary 56 Sephardic 3, 305 serialized 49–51, 58 slang 303, 338 Slavic 7–10, 12–17, 25, 27, 29–33, 36, 41, 44, 46, 82, 85, 163, 164, 169, 173, 175, 180, 215, 223–224, 258, 309, 311–312, 316–317, 319 socio-linguistic 17, 21, 61, 75, 118, 245, 247, 301 Spanish Judeo-Spanish 3, 7, 10–11, 89, 90–91, 93–98, 101, 107–108, 110, 204, 277–278, 301, 306 speaker 1–10, 12, 14–18, 21, 25, 27, 30, 33, 50, 52–55, 58, 73, 98–99, 109–110, 115–116, 121–122, 124, 136, 140–142, 146–147, 150, 152, 154 n. 14, 158, 166, 175, 181 n. 2, 188–189 n. 13, 190 n. 14, 191 n. 18, 192, 197, 199, 202, 208, 210, 215, 217, 220, 222–224, 238, 245, 248, 251–252, 255–256,

355 277–278, 287 n. 4, 293, 305–306, 309, 311, 315–319, 337–339, 342–344 speech act 11, 101, 109, 115 n. 1, 119, 132 spellout 71, 73 Sprachbund 207 structural structural borrowing 25, 337 structure 4–5, 16–17, 26, 33–34, 46, 52, 77, 93, 95, 97–98, 101, 103, 107–108, 122, 128–129, 132, 136, 138–143, 148, 154, 156, 180, 202, 204, 207–208, 211, 222, 228–229, 232, 237, 258–260, 278–279, 286–287, 293–294, 337, 339, 344–346 information structure 12, 139, 141 subjacency 69, 131 subject 12, 26, 37–39, 43–45, 50, 59, 79–82, 92, 106, 138, 140–143, 147–149, 164, 181, 198, 200–201, 217, 220 n. 7, 249, 252–256, 310, 314, 328, 334, 343 subjunctive 89, 95–98, 170 subset principle 14 substrate 19, 63, 76, 89, 98, 146–147, 154, 159, 189, 204–205, 208, 212 suffix 45, 70, 83 n. 8, 282 n. 1, 343 suggesterogative 11, 102, 104–110 superlative 16–17, 19, 269–279, 286, 288–291, 295, 337, 339–340, 343, 346 superstrate 1, 19 synchrony 164, 204, 212, 215 syncretic 14–15, 207 syntactic syntactic borrowing 63, 73–74 syntax 1, 4, 20–22, 29, 37, 82, 95, 98, 113, 118, 192–193, 258, 270, 277, 286, 326, 340, 344 šel-phrase 228, 235 n. 9 Talmud 2, 92, 93, 169, 172–173, 216 n. 1, 217–218, 223, 272, 283, 284, 292, 310 n. 3 template, binyan active 79, 85, 101–102, 147 hitpael 197–203 intensive 14 middle 14, 198–199, 202, 328 passive 199, 202, 206 tense present tense 82, 85 past tense 82, 95, 166, 170, 172, 200 future tense 10–11, 89, 92, 98, 103, 166

356 text 2, 11, 13, 16–17, 51, 56, 72, 78, 80, 82, 84, 92, 117 n. 2, 113, 137, 169 n. 12, 174–175, 180, 185–186 n. 7, 187–188, 198, 200, 203, 219, 229 n. 1, 238, 249, 251, 263, 267, 272–274, 309, 311–312, 315, 328 n. 6 thematic role agent 198, 206–207, 303 cause 26, 148, 253, 260, 267, 328–329 goal 18, 164, 310, 331–332, 335 tongue 33, 56, 73, 143, 223 n. 12, 337 topic 63, 68, 143, 148, 222, 283, 342 trace 9, 136, 269, 282 transfer 6, 8–12, 14–17, 19, 123, 207, 208, 210–211, 278 transformation 4 n. 10, 98, 272, 274, 278 transmission hypothesis 1, 5, 14, 17, 19, 36, 40, 98, 193 n. 4, 249 imperfect transmission 5 mediated transmission 5 truth condition 13, 133, 164, 314, 316, 317 Ukrainian 13, 25, 29–30, 71, 32–33, 41–2 universal 6, 72, 165, 204, 207, 258, 265, 266–267 usage colloquial 7, 9, 11–12, 16–17, 49–58, 73, 112–114, 118, 126, 128,–129, 131 n. 4, 134, 136, 139, 142 n. 18–143, 167, 197, 204, 208–211, 231, 272, 277, 282–284, 309, 312, 316–317 formal 272, 277

Index value 5–18 verb bleached 49–54, 56–58 durative 9, 50–53, 56–59 ingressive 9, 50–52 n. 4, 54–57, 59 postural 51–53 serialized 49–51, 58 verbal syndeton 58 vocabulary 316 word order SVO 129, 140–141 VSO 66 Yiddish 1, 3, 7–10, 12–17, 19, 25, 27–30, 32–33, 35, 58 n. 19, 62–64, 67–71, 73–76, 78, 82–86, 89, 90, 95, 99, 129, 139, 141–44, 154 n. 14, 160, 163–164, 169–171, 173–175, 177–180, 189–194, 197, 202–204, 207–208, 212, 214, 223–224, 226, 228, 237–241, 247, 258, 263, 266–67, 278, 281, 286, 292–295, 301, 304, 307, 309, 311–312, 316–317, 319, 322–333 n. 14 Yishuv New Yishuv 4 n. 11 Old Yishuv 3, 116, 277 Zakho 123–124, 247