Diachrony in Biblical Hebrew 9781575066837

Diachrony in Biblical Hebrew is an indispensable publication for biblical scholars, whose interpretations of scriptures

211 39 8MB

English Pages 544 [536] Year 2012

Report DMCA / Copyright

DOWNLOAD PDF FILE

Recommend Papers

Diachrony in Biblical Hebrew
 9781575066837

  • 0 0 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview

Diachrony in Biblical Hebrew

Linguistic Studies in Ancient West Semitic edited by

Cynthia L. Miller-Naudé and Jacobus A. Naudé The series Linguistic Studies in Ancient West Semitic is devoted to the ancient West Semitic languages, including Hebrew, Aramaic, Ugaritic, and their near congeners. It includes monographs, collections of essays, and text editions informed by the approaches of linguistic science. The material studied will span from the earliest texts to the rise of Islam.  1. The Verbless Clause in Biblical Hebrew: Linguistic Approaches, edited by Cynthia L. Miller  2. Phonology and Morphology of Biblical Hebrew: An Introduction, by Joshua Blau  3. A Manual of Ugaritic, by Pierre Bordreuil and Dennis Pardee  4. Word Order in the Biblical Hebrew Finite Clause: A Syntactic and Pragmatic Analysis of Preposing, by Adina Moshavi  5. Oath Formulas in Biblical Hebrew, by Blane Conklin  6. Biblical Hebrew Grammar Visualized, by Francis I. Andersen and A. Dean Forbes  7. Time and the Biblical Hebrew Verb: The Expression of Tense, Aspect, and Modality in Biblical Hebrew, by John A. Cook  8. Diachrony in Biblical Hebrew, edited by Cynthia L. Miller-Naudé and Ziony Zevit

Diachrony in Biblical Hebrew

Edited by

C ynthia L. M iller -N audé and Z iony Z evit

Winona Lake, Indiana E isenbrauns 2012

Copyright © 2012 Eisenbrauns All rights reserved. Printed in the United States of America.

www.eisenbrauns.com

Library of Congress Cataloging-in-Publication Data Diachrony in biblical Hebrew / edited by Cynthia L. Miller-Naudé and Ziony Zevit.    p.  cm. — (Linguistic studies in ancient West Semitic ; 8) Includes bibliographical references and indexes. ISBN 978-1-57506-253-2 (hardback : alk. paper) 1.  Hebrew language—Grammar.  2.  Bible O.T.—Language, style.  3.  Linguistic change.  I.  Miller, Cynthia L. (Cynthia Lynn), 1957–  II.  Zevit, Ziony. PJ4556.D485 2012 492.4′5—dc23 2012028645

The paper used in this publication meets the minimum requirements of the American National Standard for Information Sciences—Permanence of Paper for Printed Library Materials. ANSI Z39.48-1984.♾™

To

Jacobus, Nelmarie, and Salomie and

Yotan, Hillel, Maʾayan, Yahav, Rei, Oriya, and Lia

Contents Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii

Part 1: Introduction Diachrony in Biblical Hebrew: Linguistic Perspectives on Change and Variation . . . . . . . . . . . . . . . . . . . . . . . . Cynthia L. Miller-Naudé

3

Part 2: Theoretical and Methodological Perspectives on Diachrony Methodological Issues in the Dating of Linguistic Forms: Considerations from the Perspective of Contemporary Linguistic Theory . . . . . . . . . . . . . . . . . . B. Elan Dresher Biblical Hebrew as a Diachronic Continuum . . . . . . . . . . . . . . . . T. Givón Diachrony in Biblical Hebrew and a Theory of Language Change and Diffusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Jacobus A. Naudé Detecting Development in Biblical Hebrew Using Diachronic Typology . . . . . . . . . . . . . . . . . . . . . . . . . John A. Cook Historical Linguistics and Biblical Hebrew . . . . . . . . . . . . . . . . Robert D. Holmstedt

19 39

61

83 97

Part 3: Examining Diachrony in Biblical Hebrew Orthographic Features Dwelling on Spelling . . . . . . . . . . . . . . . . . . . . . . . . . . . 127 A. Dean Forbes and Francis I. Andersen vii

viii

Contents

Morphological Features The Third-Person Masculine Plural Suffixed Pronoun -mw and Its Implications for the Dating of Biblical Hebrew Poetry . . . . . . Yigal Bloch The Kethiv/Qere ‫הוא‬, ִ Diachrony, and Dialectology . . . . . . . . . . . Steven E. Fassberg Discerning Diachronic Change in the Biblical Hebrew Verbal System . Martin Ehrensvärd The Archaic System of Verbal Tenses in “Archaic” Biblical Poetry . . . Tania Notarius

147 171 181 193

Syntactic Features Diachronic Syntactic Studies in Hebrew Pronominal Reciprocal Constructions . . . . . . . . . . . 209 Elitzur A. Bar-Asher Siegal Syntactic Aramaisms as a Tool for the Internal Chronology of Biblical Hebrew . . . . . . . . . . . . . . . . . . . . . . . . . 245 Naʿama Pat-El

Lexical Features The “Linguistic Dating of Biblical Texts”: Comments on Methodological Guidelines and Philological Procedures . . . . . 265 Avi Hurvitz The Evolution of Literary Hebrew in Biblical Times: The Evidence of Pseudo-classicisms . . . . . . . . . . . . . . . . 281 Jan Joosten Signs of Late Biblical Hebrew in Isaiah 40–66 . . . . . . . . . . . . . 293 Shalom M. Paul

Sociological and Dialectal Considerations Language Variation, Discourse Typology, and the Sociocultural Background of Biblical Narrative . . . . . . . . . . 301 Frank H. Polak Northern Hebrew through Time: From the Song of Deborah to the Mishnah . . . . . . . . . . . . 339 Gary A. Rendsburg

Contents

ix

Text-Critical Considerations Diachrony in Biblical Hebrew Lexicography and Its Ramifications for Textual Analysis . . . . . . . . . . . . . . . . . . . . . . . . 361 Chaim Cohen

Part 4: Comparative Semitic Perspectives on Diachrony Outline of Aramaic Diachrony . . . . . . . . . . . . . . . . . . . . . . 379 Michael Sokoloff Diachrony in Ugaritic . . . . . . . . . . . . . . . . . . . . . . . . . . . 407 Joseph Lam and Dennis Pardee Diachrony in Akkadian and the Dating of Literary Texts . . . . . . . . 433 N. J. C. Kouwenberg

Part 5: Afterword Not‑So‑Random Thoughts Concerning Linguistic Dating and Diachrony in Biblical Hebrew . . . . . . . . . . . . . . . . . . . 455 Ziony Zevit Indexes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 491 Index of Authors . . . . . . . . . . . . . . . . . . . . . . 491 Index of Ancient Sources . . . . . . . . . . . . . . . . . 499 Index of Scripture . . . . . . . . . . . . . . . . . . . . . 505 Index of Hebrew Words . . . . . . . . . . . . . . . . . . 519 Index of Subjects . . . . . . . . . . . . . . . . . . . . . . 520

Preface This volume is compiled of papers delivered at four sessions on “Diachrony in Biblical Hebrew” at the National Association of Professors of Hebrew in 2009 and 2010. The impetus in planning the sessions was the announcement in 2008 of the publication of the two-volume work Linguistic Dating of Biblical Texts, by Ian Young, Robert Rezetko, and Martin Ehrensvärd. It is clear that their research has raised important questions concerning linguistic evidence for diachronic change in Biblical Hebrew, the possibility (or impossibility) of using linguistic methods to date the linguistic artifacts found in the biblical text to a particular historical period, and the interrelation of language variation and language change. The significance of these issues and the fact that many biblicists and historians are unaware of developments in historical linguistics make these questions all the most important. Several additional papers are included that were not presented at the conference, namely articles by Steven E. Fassberg, Talmy Givón, N. J. C. Kouwenberg, Joseph Lam and Dennis Pardee, Gary A. Rendsburg, and Michael Sokoloff. We regret that Ian Young and Robert Rezetko decided not to publish their conference papers in this volume; Martin Ehrensvärd’s contribution thus represents the only contribution from the three authors whose work sparked intense interest in the questions addressed herein. We are grateful to Jim Eisenbraun for agreeing to publish the volume and to his staff for their care in producing it. We are also grateful for the peer review of all of the essays in the volume. The first editor thanks two of her graduate students for their assistance in preparing the documents: Jared Henson and Wendy Widder. The second editor thanks the first editor, and vice versa, for the cordial process of jointly editing the volume. It is our hope that this is only the beginning of consideration of the questions of language change and variation as reflected in the biblical text. The continuously evolving fields of historical linguistics and language variation have much to offer biblical scholars as well as historians of ancient Israel’s culture, literature, and beliefs. Conversely, analyses of change and variation with respect to ancient languages and texts such as those presented here have much to offer linguistics. Cynthia L. Miller-Naudé University of the Free State Bloemfontein, South Africa

and

xi

Ziony Zevit American Jewish University Los Angeles, California

Abbreviations General

A Afʿel ABH Archaic Biblical Hebrew ACC accusative ADV adverbial Akk. Akkadian ALL allative Ass Assyrian Atr. Atra-ḫasīs BA Biblical Aramaic Bab Babylonian BH Biblical Hebrew C common CBH Classical Biblical Hebrew* COMPL complex hypotaxis CPA Christian Palestinian Aramaic D D-stem DEF definite DU dual DUR durative EA Eastern Aramaic EA El Amarna letters EBH Early Biblical Hebrew* EE Enuma Eliš ELC explicit, lexicalized constituent EMPH emphatic F feminine frg. fragment G G-stem GAp Genesis Apocryphon GEN genitive Gilg. Gilgameš GN geographical name H Hafʿel IH Israelian Hebrew *  This term has different definitions in the various essays.

xiii

xiv

Abbreviations

IMP imperative IND indicative INDF indefinite INF infinitive IRR irrealis JBA Jewish Babylonian Aramaic JH Judahite Hebrew JPA Jewish Palestinian Aramaic kjv King James Version of the Bible LBH Late Biblical Hebrew LSLA Late Standard Literary Aramaic M masculine Ma Mandaic MA Middle Assyrian MB Middle Babylonian MH Mishnaic Hebrew MT Masoretic Text NAPH National Association of Professors of Hebrew NEG negative NG noun group njps New Jewish Publication Society version of the Bible nrsv New Revised Standard Version of the Bible NOM nominative NP noun phrase NWS Northwest Semitic OA Old Aramaic OA Old Assyrian (only in Kowenberg, pp. 433–451) OB Old Babylonian obl. oblique OfA Official Aramaic OSA Old South Arabian OT Old Testament P passive stem PAR participle PASS passive PERF perfect PL plural PN personal name POSS possessive PRET preterite PS Proto-Semitic PST past QA Qumran Aramaic RECP reciprocal RH Rabbinic Hebrew REL relative

Abbreviations

xv

rev. reverse RIS Ras Ibn Hani RS/RSO Ras Shamra / Ras Shamra–Ougarit S singular Sef. Sefire SA Samaritan Aramaic SB Standard Babylonian SBH Standard Biblical Hebrew* SBJV subjunctive SLA Standard Literary Aramaic SUB subordinator / subordination Syr Syriac t infix t TA Targumic Aramaic WA Western Aramaic

Reference Works AB AfO AHw

Anchor Bible Archiv für Orientforschung Von Soden, W. Akkadisches Handwörterbuch. 3 vols. Wiesbaden: Harrassowitz, 1965–81 ALASP Abhandlungen zur Literatur Alt-Syren-Palästinas und Mesopotamiens ANEP Pritchard, J. B., editor. The Ancient Near East in Pictures Relating to the Old Testament. 2nd ed. Princeton: Princeton University Press, 1969 ANES Ancient Near Eastern Studies AnOr Analecta Orientalia ARM Archives royales de Mari AOAT Alter Orient und Altes Testament AoF Altorientalische Forschungen AOS American Oriental Series AS Assyriological Studies BASOR Bulletin of the American Schools of Oriental Research BBB Bonner biblische Beiträge BDB Brown, F.; Driver, S. R.; and Briggs, C. A. Hebrew and English Lexicon of the Old Testament. Oxford: Clarendon, 1907 BE Babylonian Expedition of the University of Pennsylvania, Series A: Cuneiform Texts BETL Bibliotheca ephemeridum theologicarum lovaniensium BHS Elliger, K., and Rudolph, W., editors. Biblia Hebraica Stuttgartensia. Stuttgart: Deutsche Bibelgesellschaft, 1984 Bib Biblica BibOr Biblica et Orientalia BN Biblische Notizen BO Bibliotheca Orientalis *  This term has different definitions in the various essays.

xvi BSOAS BZAW CAD CahRB CAL CAT

Abbreviations Bulletin of the School of Oriental and African Studies Beihefte zur Zeitschrift für die alttestamentliche Wissenschaft Oppenheim, A. L., et al., editors. The Assyrian Dictionary of the Oriental Institute of the University of Chicago. 21 vols. (A–Z). Chicago: Oriental Institute, 1956–2011 Cahiers de la Revue Biblique Comprehensive Aramaic Lexicon. http://cal1.cn.huc.edu Dietrich, M.; Loretz, O.; and Sanmartín, J. The Cuneiform Alphabetic Texts from Ugarit, Ras Ibn Hani, and Other Places. 2nd ed. Münster: Ugarit-Verlag, 1995 Catholic Biblical Quarterly

CBQ CIS Corpus inscriptionum semiticarum CRAIBL Comptes rendus de l’Académie des inscriptions et belles-lettres CTA Herdner, A., editor. Corpus des tablettes en cunéiformes alphabétiques. Paris: Imprimerie Nationale, 1963 Dar Strassmaier, J. N. Inschriften von Darius, König von Babylon. Babylonische Texte 10–12. Leipzig: Pfeiffer, 1892–97 DJA Sokoloff, M. A Dictionary of Judean Aramaic. Ramat-Gan: Bar-Ilan University Press, 2003 DJBA Sokoloff, M. A Dictionary of Jewish Babylonian Aramaic of the Talmudic and Geonic Periods. Ramat-Gan: Bar-Ilan University Press / Baltimore: Johns Hopkins University Press, 2002 DJD Discoveries in the Judaean Desert DJPA Sokoloff, M. A Dictionary of Jewish Palestinian Aramaic of the Byzantine Period. 2nd ed. Ramat-Gan: Bar-Ilan University Press / Baltimore: Johns Hopkins University Press, 2002 DNWSI Hoftijzer, J., and Jongeling, K. Dictionary of the North-West Semitic Inscriptions. 2 vols. Leiden: Brill, 1995 DSA Tal, A. A Dictionary of Samaritan Aramaic. 2 vols. Leiden: Brill, 2000 DSD Dead Sea Discoveries ErIsr Eretz-Israel FAT Forschungen zum Alten Testament GAG Von Soden, W. Grundriss der akkadischen Grammatik. 2nd ed. Rome: Pontifical Biblical Institute, 1969 GKC Kautzsch, E., editor. Gesenius’ Hebrew Grammar. Translated by A. E. Cowley. 2nd ed. Oxford: Clarendon, 1910 HALOT Koehler, L.; Baumgartner, W.; and Stamm, J. J. The Hebrew and Aramaic Lexicon of the Old Testament. Translated and edited under the supervision of M. E. J. Richardson. 5 vols. Leiden: Brill, 1994–2000 HAR Hebrew Annual Review HO Handbuch der Orientalistik HS Hebrew Studies HSM Harvard Semitic Monographs HSS Harvard Semitic Studies HUCA Hebrew Union College Annual IEJ Israel Exploration Journal

Abbreviations IOS JANES JAOS JBL JCS JEOL

xvii

Israel Oriental Studies Journal of the Ancient Near Eastern Society Journal of the American Oriental Society Journal of Biblical Literature Journal of Cuneiform Studies Jaarbericht van het Vooraziatisch-Egyptisch Gezelschap (Genootschap) Ex Oriente Lux JHS Journal of Hebrew Scriptures. http://www.jhsonline.org JNES Journal of Near Eastern Studies JNSL Journal of Northwest Semitic Languages JQR Jewish Quarterly Review JSOTsup Journal for the Study of the Old Testament: Supplement Series JSS Journal of Semitic Studies KAI Donner, H., and Röllig, W. Kanaanäische und aramäische Inschriften. 2nd ed. Wiesbaden: Harrassowitz, 1966–69 KTU Dietrich, M.; Loretz, O.; and Sanmartín, J., editors. Die Keilalpha­ betischen Texte aus Ugarit. Alter Orient und Altes Testament 24. Kevelaer: Butzon & Bercker / Neukirchen-Vluyn: Neukirchener Verlag, 1976 Leš Lešonénu LSAWS Linguistic Studies in Ancient West Semitic MARI MARI: Annales de recherche interdisciplinaires MDOG Mitteilungen der Deutschen Orient-Gesellschaft MRS Mission de Ras Shamra MVAG Mitteilungen der Vorderasiatisch-ägyptischen Gesellschaft OBO Orbis Biblicus et Orientalis OLA Orientalia Lovaniensia Analecta Or Orientalia (n.s.) OtSt Oudtestamentische Studiën PEQ Palestine Exploration Quarterly PIHANS Publications de l’Institut historique-archéologique néerlandais de Stambul RB Revue biblique RBL Review of Biblical Literature RevQ Revue de Qumran SAA State Archives of Assyria SBLDS Society of Biblical Literature Dissertation Series SBLRBS Society of Biblical Literature Resources for Biblical Study SBLWAW Society of Biblical Literature Writings from the Ancient World SBS Stuttgarter Bibelstudien SJOT Scandinavian Journal of the Old Testament SMEA Studi Micenei ed Egeo-Anatolici SR Studies in Religion SSN Studia Semitica Neerlandica STDJ Studies on the Texts of the Desert of Judah SubBi Subsidia Biblica

xviii

Abbreviations

Sokoloff, M. A Syriac Lexicon. Winona Lake, IN: Eisenbrauns / Piscataway, NJ: Gorgias, 2009 TAD Porten, B., and Yardeni, A. Textbook of Aramaic Documents from Ancient Egypt. Vols. A–D. Jerusalem: Academon, 1986–99 TCL Textes Cunéiformes du Louvre TDT1 Yardeni, A. Textbook of Aramaic, Hebrew and Nabataean Documentary Texts from the Judaean Desert, A: The Documents. Jerusalem: Dinur Center for Research in Jewish History, 2000 [Hebrew] UF Ugarit-Forschungen VT Vetus Testamentum VTSup Vetus Testamentum Supplements WO Die Welt des Orients YOS Yale Oriental Series, Babylonian Texts ZA Zeitschrift für Assyriologie ZAH Zeitschrift für Althebraistik ZAW Zeitschrift für die alttestamentliche Wissenschaft ZDMG Zeitschrift der deutschen morgenländischen Gesellschaft SyrLex

Introduction

Diachrony in Biblical Hebrew: Linguistic Perspectives on Change and Variation Cynthia L. Miller-Naudé 1.  Diachrony and Synchrony To speak of diachrony is necessarily to place it in opposition to synchrony. Ferdinand de Saussure, who first characterized the study of language from these two perspectives, described synchrony as “the axis of simultaneities,” which considers the “relations of coexisting things and from which the intervention of time is excluded” (de Saussure 1959: 80). Diachrony, by contrast, is “the axis of successions,” in which “only one thing can be considered at a time but upon which are located all the things on the first axis together with their changes” (1959: 80). For Saussure, then, a diachronic description of language examines a single linguistic feature as it changes over time. From a diachronic point of view, the linguistic sign is mutable, the continuity of language in time results in inevitable change (1959: 74), whereas from a synchronic point of view the linguistic sign is immutable in that the individual speakers of a language cannot “willfully” change the system (1959: 69). The “synchronic truth” and the “diachronic truth” are both true—“one truth does not exclude the other” (1959: 96)—but for Saussure, the synchronic truth is the more important and forms the proper study of language. Saussure’s description of diachrony (as well as its differentiation from synchrony) highlights the centrality of the notions of change and variation for historical linguistics, but precisely how change and variation are understood and explained is, in many respects, open to debate. In this introductory essay, I consider some of the theoretical underpinnings, disputes, and methodologies of historical linguistics and highlight the ways in which the various essays in this volume relate to ongoing trends and debates within the field of historical linguistics. At the end of this volume, the essay by Zevit summarizes the history of the debate concerning diachrony in Biblical Hebrew and provides an in-depth analysis of Young, Rezetko, and Ehrensvärd 2008. 2.  What Is Change and How Does It Happen? 2.1.  Defining Change The term change is currently used in a variety of ways by historical linguists, some of them contradictory (see Janda and Joseph 2003: 12–14). Milroy, for 3

4

Cynthia L. Miller-Naudé

example, differentiates innovation, an act of an individual speaker, regardless of whether or not it later catches on in a speech community, from change, an innovation that has been widely adopted by members of the speech community (Milroy 1992: 219–22). Some generative linguists, by contrast, differentiate change within the grammar of an individual speaker, regardless of whether or not it is adopted by the speech community, from diffusion as the spread of the change throughout the speech community (Hale 2007: 33–47; Lightfoot 1979; 1991; 1999); other generative linguists use the terms actuation and implementation, respectively, for these concepts (see Fischer 2007: 4). These contrasting definitions of change highlight differing positions on the locus of language change—in the community of speakers or in the grammar of the individual. A third distinction is diachronic correspondence (Janda and Joseph 2003: 13). This last term is a means to differentiate the juxtaposition of two related items, an earlier item and a later one, that are not necessarily adjacent in time; in other words, item A does not necessarily “change” directly into item B, but they are historically related. In this volume, Cook, Holmstedt, and Naudé use the terminology of change and diffusion; the other authors do not differentiate among the various types of change but (apparently) use change either to refer to a diachronic correspondence or to the acceptance of innovation by speakers in a language community. 2.2.  How Do Languages Change? The divergent viewpoints concerning the meaning of change relate in large measure to differing theories concerning how languages change over time. 2.2.1.  Neogrammarian Approach The Neogrammarians, historical linguists from the end of the 19th century, were critically concerned with the reconstruction of proto-languages, especially Proto-Indo-European. They therefore focused on sound changes as observed in the descendant languages, which they viewed as “systematic”—that is, regular and purely phonetically conditioned (Hale 2003: 343–44). One of their central contributions to historical linguistic methodology includes the development of the comparative method, which has proved to be critical for Semitic philology (see the essays by Bar-Asher Siegal and Cohen in this volume for the use of comparative data in diachronic analyses of Biblical Hebrew). The comparative method is, so to speak, a “horizontal” comparison between related languages in order to reconstruct previous stages of language, as opposed to internal reconstruction, which involves a “vertical comparison” between different stages of the same language (Janda and Joseph 2003: 41; see also Fox 1995 and Campbell 2004).

Diachrony in Biblical Hebrew

5

2.2.2.  Formal Approaches Among modern historical linguists, approaches to language change have been fundamentally split between formal (generative) approaches and functional approaches (see Fischer 2007 for a comparison of these approaches). Generative approaches to language change see the locus for language change in child language acquisition. In this theory, the acquisition of the language of the child is not identical to that of the “input” (the “primary linguistic data” that the child receives from parents or other speakers in the environment) and thus the child’s mental grammar necessarily is different from the input; the difference is the change (Hale 2003: 345–46). The discontinuity of language transmission across generations (as parents transmit language to children), even in cases in which changes are not systemic, is the locus of language change. It is only when a number of speakers acquire the same structural feature that the change is said to be diffused throughout the speech community. Language change is thus instantaneous (in the acquisition of language), whereas diffusion is gradual. The generative approach to language change is thus connected to a theory of child language acquisition. Three essays in this volume take a formal approach to language change. Dresher presents examples of language change within the history of English from a generative point of view to illustrate how Biblical Hebrew should be approached diachronically. Holmstedt argues that the -h on first-person singular wayyiqtol forms represents an example of this kind of change. Naudé, employing a modified formal approach based on the insights of complex systems theory (see below), makes a similar claim concerning the change and diffusion of independent personal pronouns and the so-called “consecutive” verbal forms. 2.2.3.  Functional Approaches Functional approaches to language change, by contrast, view change as relating to the functionality of language for communication. An example of functional approaches to language change is found in the work of Harris and Campbell (see Pat-El’s essay in this volume appropriating the approach of Harris and Campbell 1995 to syntactic change), who see all change as communicationally based. They classify three types of language change: extension, reanalysis, and borrowing (discussed further below). Another major figure who has developed a functional approach to historical linguistics is Givón (see his essay in this volume); most recently, he has been concerned with an evolutionary approach to language change (Givón 2009; see also §2.2.6 below). Givón has been especially influential in developing the notion of grammaticalization as a means of language change and coined the dictum “today’s morphology is yesterday’s syntax” (Givón 1971: 413). Grammaticalization in general relates to the emergence and development of grammatical categories;

6

Cynthia L. Miller-Naudé

a linguistic expression is grammaticalized as a grammatical form, or a grammatical form becomes grammaticalized into a different grammatical form. For example, ‫ לֵאמֹר‬was originally a preposition followed by an infinitive construct; the form became grammaticalized as a complementizer, a grammatical function word used to introduce direct speech complements (Miller 1996: 200–212; for additional examples of grammaticalization in Semitic, see Rubin 2005). While the term grammaticalization can refer to the linguistic process, it is also used to refer to the theory that describes and explains grammaticalization phenomena (Heine 2003: 575; see also Heine, Claudi, and Hünnemeyer 1991; Traugott and Heine 1991; and Hopper and Traugott 1993). In this volume, Bar-Asher Siegal, Cook, Givón, and Naudé provide explanations of Biblical Hebrew features by means of grammaticalization. 2.2.4. Typology A way of looking at language change that is often correlated with grammaticalization involves language typology and especially diachronic typology. Language typology, in general, explores the range of variation present cross-linguistically among the languages of the world (see, for example, Greenberg 1963; Comrie 1989; Croft 1990; and especially Haspelmath et al. 2005). Although synchronic in its orientation, it provides an important way to use “the present for the illumination of the past” (Janda and Joseph 2003: 21) in that reconstructed language features or systems should be currently attested in some manner among the world’s languages. Diachronic typology explores the trajectories of change routinely followed by various linguistic features or constructions (especially those involving grammaticalization) by comparing cross-linguistic paths of change. For example, the classic work by Bybee, Perkins, and Pagliuca (1994) investigates the kinds of grammaticalization and change involving verbal forms and extrapolates from them various cross-linguistic generalities concerning pathways of change. These two kinds of typology are employed in the essays by Bar-Asher Siegal, Cook, and Givón. 2.2.5. Borrowing Borrowing, the appropriation of linguistic features or expressions across language boundaries, has its source in contact between speakers of different languages (see the general remarks concerning borrowing in the essays by Holmstedt and Pat-El). Interference across language boundaries may result in the loss of old features from the recipient language, the addition of new features, or the replacement of old features by new ones (Thomason 2001 and the updated analysis in Thomason 2003). One critical question involves whether or not the speakers who introduce the interference feature(s) fluently speak the language into which the features are introduced; if they do not, then imperfect learning of a second language plays an important role (Thomason 2003: 691–93). The first scenario, involving native speakers, may be referred to as borrowing proper; the second scenario, involving non-native speakers carrying

Diachrony in Biblical Hebrew

7

over features of their native language as they learn a new language, may be referred to as “shift-induced interference” (Thomason 2003: 691). In borrowing, “interference always begins with non-basic vocabulary unless languages A and B have mostly or entirely identical lexicons” (2003: 692); in shift-induced interference, phonological and syntactic interference predominate (2003: 693). The question of contact linguistics is especially important for the study of diachrony in Biblical Hebrew because of the problematic nature of the relationship between Hebrew and Aramaic, both linguistically (in terms of shared linguistic structures) and socially (in terms of intensive contact in the Persian period as well as across the “amorphous border” of Israel and Aram, as described by Rendsburg in this volume). In this volume, the question of diachronic change in Biblical Hebrew as a function of borrowing from Aramaic is addressed with respect to lexical items and expressions (in the essays by Holmstedt, Hurvitz, Paul, and Rendsburg), morphology (in the essay by Rendsburg), and syntax (in the essays by Pat-El and Rendsburg). 2.2.6.  Language Evolution and Complex Systems Analysis The most recent linguistic approaches to language change involve language evolution (Givón 2009; Mufwene 2008) and complex systems analysis (Larsen-Freeman and Cameron 2008; Ellis and Larsen-Freeman 2009; and Sampson, Gil, and Trudgill 2009). In this volume, Naudé describes these recent developments in his essay and synthesizes them to propose a methodology for examining language change in Biblical Hebrew from the viewpoint of language complexity. 3.  The Many Varieties of Variation While diachronic change necessarily involves variation, not all variation reflects past change or leads to future change. This section provides a brief overview of some of the kinds of variation attested in language generally and the role of variation in diachronic change. Linguistic variation has been described from the vantage points of multiple subdisciplines within linguistics; this brief survey cannot, therefore, be comprehensive. One way to differentiate types of variation is to contrast variation that takes place in the speech of individual speakers as opposed to variation that takes place across groups of speakers (Schilling-Estes 2002: 375). The variation of individual speakers may relate to dialect (including the possibility of shifting in and out of dialect varieties), register, or genre. Stylistic variation may involve a cluster of linguistic features, so that it is possible to speak of a “formal style,” a “conversational style,” and the like (Schilling-Estes 2002: 376). Individual style may function on a wide variety of levels to accomplish a wide range of sociological functions, such as to indicate group affinity and to achieve conversational purposes. Variation occurs at every level of language. Phonetically, the acoustic shape of every phoneme varies each time it is pronounced, even when it is

8

Cynthia L. Miller-Naudé

pronounced by the same speaker. (As Wolfram [2006] notes, the amazing thing is that speakers and hearers are able to generalize across the variation to communicate.) Phonologically and morphologically, language exhibits conditioned variants as a structural feature (namely, allophones and allomorphs, respectively). This kind of change does not result in linguistic change, though it may itself be the result of past linguistic change. Variation, especially involving optional, alternative syntactic constructions or lexical choice, may be employed for prosodic purposes or for information structure. Poetic language as described by Jakobson involves a focus on the message and using alternative variants for poetic effect by selection and combination. In Jakobson’s terms, the “poetic function projects the principle of equivalence from the axis of selection into the axis of combination” (Jakobson 1987: 71). In other words, poetic language uses variant resources within language for poetic effect. In a similar way, a speaker may use variants of language for information-processing concerns. That is, for cohesive, discourse, or discourse-pragmatic purposes, a speaker may use alternatives within language (Renkema 2009). This sort of variation—the use of the resources of language for special effects or communicative purposes—is motivated by particular speakers in particular contexts and is diachronically stable. Another kind of variation involves two distinct grammatical options “in areas of grammar that do not ordinarily permit optionality” (Pintzuk 2003: 510). In effect, the two “grammars” are in competition; the new grammar competes with the old grammar and eventually replaces it. The ratio of the old grammar to the new grammar follows an S-shaped curve through time, with the new grammar infrequently attested, then increasing rapidly, and finally (as it is fully adopted) slowing down. Significantly, however, the ratio of old grammar to new grammar remains the same in each grammatical context in which it is attested. What this means is that the grammatical change occurs on a “deep” structural level with the result that all of the “surface” manifestations occur at the same rate, a situation known as the Constant Rate Effect (Kroch 1989; Pintzuk refers to it as the Constant Rate Hypothesis [2003: 511]). The papers in this volume by Cook, Dresher, Holmstedt, and Naudé use this methodology. Yet another type of variation involves what sociolinguists refer to as the “linguistic variable,” which is “a structural unit that includes a set of fluctuating variants showing meaningful co-variation with an independent set of variables” (Wolfram 2006: 333). Wolfram describes the vast array of linguistic features and relationships that may constitute a linguistic variable (2006: 334): The linguistic variable may consist of lexical choice (for example, soda versus pop versus coca-cola in various dialects of American English). Structural linguistic variables may involve a grammatical category (for example, a tense form), a phoneme (for example, variation between -ing versus -in’ in American English), or a natural class of items in a specific linguistic context.

Diachrony in Biblical Hebrew

9

A linguistic variable may involve variation in the application of a syntactic process (for example, the variation in English between she has not versus she hasn’t) or in terms of syntactic relationships (for example, negative concord He didn’t do anything about any problem versus He didn’t do nothing about no problem). The linguistic variable may also relate to alternative expressions of a speech act (for example, the expression of an apology). In synchronic studies of language variation, phonological and morphological features have predominated because they are based largely on oral language; in diachronic studies of language variation, grammatical features are usually studied, because written texts usually do not display phonological or phonetic variation. The linguistic variable correlates in frequency with other independent variables. These are categorized as constraints on variability, where the term constraint indicates a factor that systematically correlates with a higher frequency of occurrence. Constraints are of two types: internal constraints, consisting of “structural linguistic factors related to the linguistic system itself” (Wolfram 2006: 335), and external constraints, involving “social or sociopsychological factors of various types that exist apart from the linguistic system” (2006: 335). Internal constraints thus relate to traditional linguistic analysis (such as phonetic environment, type of morpheme, lexical category). External constraints include “demographic variables (for example, age, social class, region), constructed social groupings and practices of various types (for example, communities of practice, social networks), interactional dynamics (for example, power relations, solidarity), and even personal presentation styles and registers (for example, performance, mimicking”; 2006: 335). In other words, external constraints are, in effect, variations based on ethnography. Both internal and external constraints are usually relevant to the synchronic description of the systematic variation of a particular linguistic variable. Furthermore, while language change implies co-variation (that is, linguistic variables), some variation is diachronically stable and does not participate in the dynamics of change. The question of whether the variations attested in Biblical Hebrew morphology, syntax, and lexicon indicate diachronic change, geographical dialect variants (Young 2003), or “style” (Young, Rezetko, and Ehrensvärd 2008) is critical. In this volume, Ehrensvärd’s essay considers several grammatical variables as indicative of alternative styles of the biblical writers rather than diachronic change. One variation of this sort involves the use of the infinitive absolute to express the meaning of a finite verbal form, which Ehrensvärd sees as solely stylistic. In contrast, Paul’s essay sees the same variation as indicative of diachronic change, whereas Rendsburg sees it as a matter of geographical dialect as well as linguistic change. The essay by Polak handles the question of the sociolinguistics of variation in a sophisticated way, by comparing the linguistic features of three corpora in the biblical text and correlating them with oral versus written style, the sociological factors of writing (scribal practice),

10

Cynthia L. Miller-Naudé

and the situation of language contact and status in the postexilic period. Rendsburg considers factors of dialect geography (the northern variety of Israelian Hebrew) as well as diachrony, and Holmstedt describes the variation between ‫ֲשׁר‬ ֶ ‫ א‬and ‫שׁ‬ ֶ as related to matters both of diachrony and of style. Finally, Fassberg identifies both dialect and diachrony at work in an early variant form of the third-person feminine singular pronoun, as preserved in the Kethiv of the Masoretic Text in the Pentateuch. The stability of linguistic features over time (that is, the relative resistance of linguistic features to change, loss, or borrowing) is also an important area for linguistic inquiry (see Nichols 2003). In this regard, the essay by Rendsburg is significant in its identification of stable grammatical features of Israelian Hebrew. 4.  Diachrony and Biblical Hebrew Linguistic Features The essays in this volume contribute in significant ways to the discussion concerning diachrony in Biblical Hebrew by examining a multitude of linguistic features on a wide range of linguistic levels. Before summarizing these contributions, however, I would like to reiterate the distinction made by a number of the authors between the text of the Bible as linguistic artefact and the linguistic system(s) of Biblical Hebrew (see the essays by Cook, Dresher, Holmstedt, and Naudé). On the purely phonological level, there is very little biblical evidence available for discerning sound changes within the span of Biblical Hebrew. However, Forbes and Andersen have presented evidence that some of the variations in the spelling conventions for the matres lectionis reflect change over time. Not all of these orthographic variations relate directly to phonological change. Many reflect only the visual representation of the phonology of the language. A number of studies treat morphological features. These include: the distribution of the third-person masculine-plural suffixed pronoun -mw (Bloch), evidence that the Kethiv ‫הוא‬ ִ representing the third-person feminine singular independent pronoun is, in fact, an early dialectal variant (Fassberg), the relative markers ‫ אשׁר‬and ‫( שׁ‬Holmstedt, Givón, Rendsburg), the increased use of ‫ הלך‬and ‫ פחד‬in the Piel (Paul), the reduplicated plural of geminate nouns (Rendsburg), feminine-singular nouns ending in -ût (Rendsburg), third-person feminine singular nouns of III-y verbs ending in -at (Rendsburg), the secondperson feminine independent pronoun ‫( אתי‬Rendsburg), and the femininesingular demonstrative ‫זו‬/‫( זה‬Rendsburg). Morphosyntactic features of Biblical Hebrew verbal forms are examined extensively including: the predicate split encoding of stative adjectives (Cook), the distribution of wetiqtol versus weqataltá /qĕtol (Ehrensvärd), the use of the infinitive absolute for a finite verb (Ehrensvärd, Paul, Rendsburg), the use of yiqtol as a present progressive in conversational contexts, and the use of qatal

Diachrony in Biblical Hebrew

11

in poetic narrative contexts within Archaic Biblical Hebrew (Notarius), the T-stem to express passive (Rendsburg), the use of the passive participle to express active voice (Rendsburg), the loss of consecutive verbal forms (Naudé), the correlation of tense-aspect to word order (Givón). Syntactic features include the following: pronominal reciprocal constructions (Bar-Asher Siegal), the complements of verbs of cognition, perception, and utterance (Givón), adverbial clauses (Givón), the reduced use of the definite object marker ʾet as a function of Aramaic influence (Pat-El) or time (Paul), causal subordination with šel (Pat-El), the periphrastic use of ‫ היה‬with the participle (Paul), the interrogative expression ‫ או‬. . . ‫( ה‬Rendsburg), and the passive infinitive construct with prepositional subject (Ehrensvärd). The Biblical Hebrew lexical inventory and facts about its distribution are examined by Dresher (‫ ממלכה‬versus ‫)מלכות‬, Hurvitz (‫מדרשׁ‬, ‫)אגרת‬, Paul (a large number of verbs, nouns, phrases, and expressions), and Rendsburg (the phrase ‫)זה היום‬. Joosten considers the use of various words and expressions as pseudo-classicisms, especially in cases in which a “poetic” word is used in a prosaic context or in which the writer attributed to words meanings that arose from biblical interpretation. The ramifications of lexical change for textcritical analysis are considered by Cohen. Finally, features of Biblical Hebrew reflecting sociolinguistic variation are examined in detail by Polak and Rendsburg. These include lexical, morphological, and syntactic features of language. 5.  Comparative Ancient Near Eastern Perspectives on Diachronic Variation The final three essays in the volume examine cognate languages and their written evidence with respect to the question of diachronic variation. In particular, these studies are included to provide a broader perspective from the ancient Near East concerning the possibilities and difficulties for discerning diachronic change in languages with longer or shorter linguistic histories, as well as various sorts of conventions concerning the copying and transmitting of texts. The essay by Sokoloff examines the long and dialectally fragmented history of Aramaic and illustrates the intertwined nature of dialectal and diachronic variation. The essay by Lam and Pardee looks at the relatively short history of Ugaritic texts to determine what kinds of language change can be detected in those texts, many of them literary. Finally, the essay by Kouwenberg describes the long and important history of Akkadian and its equally long history of textual transmission to determine how the language of Akkadian varied over time. Kouwenberg concludes that the nonliterary texts show clear variation with respect to dialect and diachrony. In contrast, texts written in the literary dialect of Standard Babylonian show little change over time. Interestingly,

12

Cynthia L. Miller-Naudé

Akkadian literary texts confirm the observations of Shi (1989) concerning linguistic change within Classical Chinese literary texts. Shi found that, because Classical Chinese texts tend to maintain conservative features, the rate of diffusion of a change occurred much more rapidly in the vernacular language than was reflected in the literary texts. 6.  Future Research The question of language variation and change within the biblical text is a critical one for linguists as well as for biblical scholars. This avenue of inquiry has recently been revitalized by the publication of Young, Rezetko, and Ehrensvärd (2008), but much work remains to be done, both from a theoretical and a methodological point of view. Theoretically, diachronic research on Biblical Hebrew will make even greater strides to the extent that it is informed by the most current (or at least, one of the current) theories of linguistic change, as illustrated by some of the essays in this volume; recent research on the mechanisms of linguistic change provides new means for grappling with long-standing problems. Furthermore, given the presence of variation within the biblical (and ancient Near Eastern) texts, the sophisticated models developed by linguists for identifying multiple sources of variation (including change over time) provide important resources for future research. Finally, Hebraists can contribute to linguistic theory by finding ways to incorporate text-critical considerations as well as the writtenoral nature of the biblical text into the complex theories currently available for the analysis of linguistic change and variation. Methodologically, the wealth of linguistic data found in the Hebrew Bible can be more thoroughly and insightfully analyzed in accordance with historical-linguistic and variationist techniques. Among the many possible avenues of research are the following. First, having identified a significant number of diachronic correspondences within texts, Hebrew linguistics can shift their focus to attempts to explain the path(s) of change and/or the impetus(es) for the innovation(s). Second, as Hebraists work to identify the multiple types of variation that are present in the biblical text—those that are related to linguistic change as well as those that are not—it is important also to consider that variation usually has multiple sources and thus multiple explanations. Third, the written, literary characteristics of the biblical text as well as the oral features that hint at the vernacular lying behind (portions of) the biblical text must be further identified together with an exploration of their implications for the identification of linguistic change and variation. Comparison with other oral and written literatures (such as pre-Islamic Arabic poetry, the medieval piyyuṭîm, or modern written records of oral literature from Africa) may provide helpful analogies for analysis. Finally, comparison of proposed analyses of linguistic change (including the various synchronic states) in ancient He-

Diachrony in Biblical Hebrew

13

brew with cross-linguistically (or, typologically) similar kinds of variation and change will provide additional confirmation that the analyses are linguistically feasible. References

Bybee, J. L.; Perkins, R. D.; and Pagliuca, W. 1994 The Evolution of Grammar: Tense, Aspect, and Modality in the Languages of the World. Chicago: University of Chicago Press. Campbell, L. 2004 Historical Linguistics: An Introduction. 2nd ed. Edinburgh: Edinburgh University Press. Comrie, B. 1989 Language Universals and Linguistic Typology: Syntax and Morphology. 2nd ed. Chicago: University of Chicago Press. Croft, R. 1990 Typology and Universals. Cambridge Textbooks in Linguistics. Cambridge: Cambridge University Press. Eckert, P., and Rickford, J. R. 2001 Style and Sociolinguistic Variation. Cambridge: Cambridge University Press. Ellis, N. C., and Larsen-Freeman, D., eds. 2009 Language as a Complex Adaptive System. Oxford: Blackwell. Fischer, O. 2007 Morphosyntactic Change: Functional and Formal Perspectives. Oxford Surveys in Syntax and Morphology 2. Oxford: Oxford University Press. Fox, A. 1995 Linguistic Reconstruction: An Introduction to Theory and Method. Oxford Textbooks in Linguistics. Oxford: Oxford University Press. Givón, T. 1971 Historical Syntax and Synchronic Morphology: An Archaeologist’s Field Trip. Chicago Linguistics Society 7: 394–415. 2009 The Genesis of Syntactic Complexity: Diachrony, Ontogeny, Neuro-cognition, Evolution. Philadelphia: Benjamins. Greenberg, J. H., ed. 1963 Universals of Language. 2nd ed. Cambridge, MA: MIT Press. Hale, M. 2007 Historical Linguistics: Theory and Method. Blackwell Textbooks in Linguistics 21. Malden, MA: Blackwell. Harris, A. C., and Campbell, L. 1995 Historical Syntax in Cross-Linguistic Perspective. Cambridge Studies in Linguistics 14. Cambridge: Cambridge University Press. Haspelmath, M.; Dryer, M. S.; Gil, D.; and Comrie, B., eds. 2005 The World Atlas of Language Structures. Oxford: Oxford University Press.

14

Cynthia L. Miller-Naudé

Heine, B. 2003 Grammaticalization. Pp. 575–601 in The Handbook of Historical Linguistics, ed. B. D. Joseph and R. D. Janda. Blackwell Handbooks in Linguistics. Oxford: Blackwell. Heine, B.; Claudi, U.; and Hünnemeyer, F. 1991 Grammaticalization: A Conceptual Framework. Chicago: University of Chicago Press. Hopper, P., and Traugott, E. C. 1993 Grammaticalization. Cambridge: Cambridge University Press. Janda, R. D., and Joseph, B. D. 2003 On Language, Change, and Language Change—Or, Of History, Linguistics, and Historical Linguistics. Pp. 3–180 in The Handbook of Historical Linguistics, ed. B. D. Joseph and R. D. Janda. Blackwell Handbooks in Linguistics. Oxford: Blackwell. Jakobson, R. 1987 Linguistics and Poetics. Repr., pp. 62–93 in R. Jakobson, Language in Literature, ed. K. Pomorska and S. Rudy. Cambridge, MA: Harvard University Press. [Reprinted from Style in Language, ed. T. A. Sebeok. Cambridge, MA: MIT Press, 1960.] Kiesling, S. C. 2011 Linguistic Variation and Change. Edinburgh Sociolinguistics. Edinburgh: Edinburgh University Press. Kroch, A. S. 1989 Reflexes of Grammar in Patterns of Language Change. Journal of Language Variation and Change 13: 199–244. Labov, W. 1994 Principles of Language Change: Internal Factors. Oxford: Blackwell. Larsen-Freeman, D., and Cameron, L. 2008 Complex Systems and Applied Linguistics. Oxford: Oxford University Press. Lightfoot, D. 1979 Principles of Diachronic Syntax. Cambridge: Cambridge University Press. 1991 How to Set Parameters: Arguments from Language Change. Cambridge, MA: MIT Press. 1999 The Development of Language: Acquisition, Change, and Evolution. Oxford: Blackwell. Miller, C. L. 1996 The Representation of Speech in Biblical Hebrew Narrative: A Linguistic Analysis. HSM 55. Atlanta: Scholars Press. Milroy, J. 1992 Linguistic Variation and Change. Oxford: Blackwell. Mufwene, S. S. 2008 Language Evolution: Contact, Competition and Change. London: Continuum. Nichols, J. 2003 Diversity and Stability in Language. Pp. 283–310 in The Handbook of Historical Linguistics, ed. B. D. Joseph and R. D. Janda. Blackwell Handbooks in Linguistics. Oxford: Blackwell.

Diachrony in Biblical Hebrew

15

Pintzuk, S. 2003 Variationist Approaches to Syntactic Change. Pp. 509–28 in The Handbook of Historical Linguistics, ed. B. D. Joseph and R. D. Janda. Blackwell Handbooks in Linguistics. Oxford: Blackwell. Renkema, J. 2009 The Texture of Discourse: Towards an Outline of Connectivity Theory. Amsterdam: Benjamins. Rubin, A. 2005 Studies in Semitic Grammaticalization. HSS 57. Winona Lake, IN: Eisenbrauns. Sampson, G.; Gil, D.; and Trudgill, P. 2009 Language Complexity as an Evolving Variable. Oxford: Oxford University Press. Saussure, F. de 1959 Course in General Linguistics, ed. C.  Bally and A.  Sechehaye. Trans. W. Baskin. New York: McGraw-Hill. Schilling-Estes, N. 2002 Investigating Stylistic Variation. Pp. 375–401 in The Handbook of Language Variation and Change, ed. J. K. Chambers, P. Trudgill, and N. Schilling-Estes. Blackwell Handbooks in Linguistics. Oxford: Blackwell. Shi, Ziqiang 1989 The Grammaticalization of the Particle le in Mandarin Chinese. Language Variation and Change 1: 99–114. Thomason, S. G. 2001 Language Contact: An Introduction. Washington, DC: Georgetown University Press. 2003 Contact as a Source of Language Change. Pp. 687–712 in The Handbook of Historical Linguistics, ed. B.  D. Joseph and R.  D. Janda. Blackwell Handbooks in Linguistics. Oxford: Blackwell. Traugott, E. C., and Heine, B., eds. 1991 Approaches to Grammaticalization. 2 vols. Amsterdam: Benjamins. Wolfram, Walt 2006 Variation and Language: Overview. Pp. 333–41 in Encyclopedia of Language and Linguistics, ed. K. Brown. 2nd ed. Oxford: Elsevier. Young, I., ed. 2003 Biblical Hebrew: Studies in Chronology and Typology. JSOTSup 369. New York: Continuum. Young, I.; and Rezetko, R.; with Ehrensvärd, M. 2008 Linguistic Dating of Biblical Texts. 2 vols. London: Equinox.

Methodological Issues in the Dating of Linguistic Forms Considerations from the Perspective of Contemporary Linguistic Theory B. Elan Dresher 1. Introduction The purpose of this essay is to present some methodological principles that underpin contemporary work in diachronic linguistics and sociolinguistics and see to what extent they can be brought to bear on current controversies concerning the dating of Biblical Hebrew. While recognizing that the dating of biblical texts poses unique challenges, one can nevertheless seek to evaluate hypotheses about the development of Biblical Hebrew in the light of what is known about language variation and change in general. In Linguistic Dating of Biblical Texts (henceforth LDBT), Young, Rezetko, and Ehrensvärd (2008) question the possibility of dating biblical books using linguistic criteria. Although some points they make about earlier claims may be valid, they go too far in my view in discounting the possibility of any sort of diachronic account of the variation found in the texts. This is because the methodology they use to argue against particular diachronic interpretations of variation is overly rigid and would, if applied to other languages, fail to identify even well-attested diachronic variation in texts. Further, their central arguments against diachronic accounts rest on flawed reasoning and unrealistic assumptions about dialects and language change. I begin by considering a major difference between working with the biblical texts and dating historical texts in Old English, tasks that have sometimes Author’s note:  I am grateful for comments and discussion to Vincent DeCaen, Robert Holmstedt, Cynthia L. Miller-Naudé, Tobie Strauss, participants in the sessions on Diachrony in Biblical Hebrew (National Association of Professors of Hebrew) at the 2009 annual meeting of the Society of Biblical Literature, and colleagues in the Department of Linguistics at the University of Toronto. All errors are my own. This research was supported in part by grant 410-08-2645 from the Social Sciences and Humanities Research Council of Canada.

19

20

B. Elan Dresher

been argued to be comparable. Following this, I consider how the study of Biblical Hebrew might have more in common with the study of the history of other languages than LDBT might allow. I then discuss some ways in which I disagree with the methodological principles adopted by LDBT and propose a different methodology for Biblical Hebrew linguistics. 2.  How Biblical Hebrew Is Different First, I agree with LDBT that the problems confronting a historical linguist dealing with the biblical texts are not the same as the problems one typically faces in Old English. In an example alluded to by LDBT (1.61), Kofoed (2006) suggests that we can follow the example of Amos (1980) in her monograph titled Linguistic Means of Determining the Dates of Old English Literary Texts. His point is that, if Amos can do this for Old English, similar methods should apply in the case of Biblical Hebrew. Consider Amos’s first test, which she concludes is not a reliable indicator for dating. The linguistic phenomenon is the deletion of the high vowels i and u when in the contexts stated in (1a). There is no doubt or controversy that pre–Old English forms such as *wordu ‘words’ or *fœ̅ti ‘feet’ (1b) changed to word and fœ̅ t, respectively (1c). In these words, -u and -i are in an open syllable following the heavy stems word (closed by two consonants) and fœ̅ t (containing a long vowel). In principle, then, if we find a text with undeleted i or u in words like these, we should be able to conclude that the language is very early. And if these vowels have been deleted, then it is later than the change. (1)  Old English high vowel deletion (HVD) as a test for dating a text (Amos 1980) a.  High vowels in an open syllable (a syllable that is not closed by a consonant) delete following a heavy stem, that is, a stem with a long vowel or closed by more than one consonant (Campbell 1959: 144–47; Hogg 1992a: 227–30). b.  Before HVD: *wordu ‘words’ *fœ̅ ti ‘feet’ c.  After HVD: word fœ̅ t d.  Questionable 1: ætgæ̅ ru gloss of Latin framea ‘spear’ e.  Questionable 2: ðweoru gloss of Latin prava ‘perverse’ So why does Amos conclude that this phenomenon is not very useful for dating? Because the change occurred very early, so much so that only a few possible examples exist of the pre-change forms, and these forms are dubious. Nouns in Old English belonged to a variety of declension classes. The form word is a member of the a-stems (named after an ancient but no-longer-occurring stem ending -a), a large and productive class, and fœ̅ t belongs to a minor declension class that has only a few nouns. These classes, and most of the “regular” Old English declension classes, undergo HVD in a straightforward way. But there exist other noun classes in which a suffixal -u was protected from deletion by

Methodological Issues in the Dating of Linguistic Forms

21

a preceding -i which underwent HVD in regular fashion. Such classes include the ja-stems and jō-stems, so-called because of the -i (sometimes designated -j) that intervened between the stem and the inflectional suffixes. For example, wītu ‘punishments’ is the post-HVD plural of a neuter ja-stem: starting from *wītiu, HVD applies to the -i, leaving behind the -u. Therefore, in deciding if HVD should have applied to a word, one must know to which declensional class it belongs. In the forms discussed by Amos, this is not always clear. For example, there is a form ætgæ̅ru (1d), found in early glossaries as the Old English gloss of Latin framea ‘spear’. This looks as though it could be a relevant example, because a final u appears after the heavy syllable gæ̅r ; we would expect it to delete after the introduction of HVD, suggesting that the form predates this rule. But the declension class, gender, and case of this form are all uncertain, making it difficult to draw a firm conclusion. The form might be the normal plural of a neuter ja-stem, like wītu. 1 Or it could be the singular of a feminine jō-stem, which also had an intervening -i that may have protected the final -u from deletion. 2 Somewhat less plausibly, it may not even be a genuine case of a final -u, but may represent a dative plural ætgæ̅ rum: the final m is sometimes indicated with a bar over the vowel, and this bar may have been omitted. Another dubious example is ðweoru, a nominative feminine-singular adjective that occurs in the Vespasian Hymns as a gloss for Latin prava ‘perverse’ (1e): one source of uncertainty here is whether the stressed vowel was long or short (vowel length was not marked in the text in question). If the vowel was short, no deletion is expected. However, even if the vowel was long, it is not clear that we should expect deletion of -u, because the long vowel would have been due to the loss of h between vowels from the early Old English form *ðweorh-u (compare the masculine nominative singular ðweorh; Campbell 1959: 264–65; Hogg 1992a: 173–75). This compensatory lengthening, if it indeed occurred, was most likely later than HVD and thus would not affect the final -u. In these cases, and all the others, the diachronic stages of the language are not in doubt. What is in doubt is the interpretation of individual forms and whether we can find useful reflections of these stages in the texts. In Biblical Hebrew, we are faced with a different situation. Here, it is the diachronic stages themselves that are in doubt. Therefore, we are not able simply to emulate Amos (1980). 1.  A plural suffix would not match the Latin, but the glossaries are not always precise in such matters. 2.  Amos (1980: 21) comments that scholars differ with regard to whether the -u would be expected to delete in jō-stem nouns. This class of nouns was subject to influences from other classes that obscure the expected phonological changes. For a brief survey of the Old English declension classes, see Hogg 1992b.

22

B. Elan Dresher

More generally, certain diagnostics that linguists rely on in dating Old English texts are not available to us in Biblical Hebrew. First, the biblical books appear to have been revised heavily, and the language of the original composition of the early books may have been updated. Conversely, there is also evidence that later authors sometimes attempted to make their compositions sound more ancient by using archaisms. Both tendencies make dating a tricky proposition. In Old English, by contrast, we often know the identity and general dates of authors of manuscripts. A second problem is that sound change, which is a mainstay of historical reconstruction, is not available to us. Hebrew writing was originally consonantal, with little or no indication of vowels. A system of representing some vowels by consonants (called matres lectionis) developed gradually from the tenth through the sixth centuries, and a more radical set of changes took place after 586 b.c.e. through the period of textual stabilization, ca. 200 b.c.e.–100 c.e. 3 The first texts we have with diacritic markings for vowels and other prosodic and phonological marks are the Masoretic codices that date from around 900 c.e.—that is, around 1,000 years after the fixing of the consonantal text. The phonology indicated by the Masoretic texts is largely uniform; therefore, phonological changes are generally not accessible to us through the biblical books. A third problem is that in the biblical period we do not have a lot of evidence for the state of the language outside the Bible. This is not to say there is none, but there is not a great deal. Therefore, the chief source of evidence for what is Early Biblical Hebrew and what is Late Biblical Hebrew comes from the Bible itself. Thus, features characteristic of early books are considered Early Biblical Hebrew, whereas features characteristic of later books are attributed to later Biblical Hebrew. One might detect some circularity here, and this is what is charged by LDBT: a feature is early because it occurs in an early book; and a book is early because its language has early features. To some extent, this critique is valid: our current models of the development of Hebrew are not especially sophisticated, and some particulars are open to the charge of circularity. 3.  How Biblical Hebrew Is the Same Nevertheless, this does not mean that Biblical Hebrew is impervious to linguistic investigation. Hebrew is a language like other languages, and therefore we may assume that basic assumptions about language in general apply to Hebrew as well. Some basic assumptions that most linguists agree on are listed in (2) and discussed below. 3.  For various views on the development of the matres lectionis, see Zevit 1980; Andersen and Forbes 1986. Gogel 1998 contains a detailed discussion of this issue and other aspects of Epigraphic Hebrew.

Methodological Issues in the Dating of Linguistic Forms

23

(2)  Some basic premises a.  All natural languages change. b.  All languages have dialects (regional, social, etc.). c.  Diachronic change begins with synchronic variation. d.  We must distinguish between a language and its reflection in texts. e.  To use linguistic criteria as an aid in dating texts, we must have a model of the history of the language—that is, of both diachronic and synchronic variation. All natural languages change. Although the rate of change is not necessarily constant, and the direction of changes may not be predictable, it appears to be part of the nature of things for languages to keep changing. The second assumption follows from the first. If languages are always changing, and if the directions of change are not predictable, it follows that a language will change in different ways in different subgroups of speakers, giving rise to dialects. These dialects may be regional, or social, or even age-based. The third premise, that diachronic change begins with synchronic variation, follows from the observation that many linguistic changes begin as variation within the grammars of individual speakers. Added to this is the fact that speakers (or writers, in our case) with different grammars exist at the same time, so learners (or philologists) may be receiving input from speakers with slightly different grammars, creating both intra- and inter-grammatical variation. A fourth premise appears to be obvious, but it is worth stating at the outset: we must distinguish between a language and its reflection in texts. Historical linguists are mainly interested in trying to reconstruct the history of a language. This is never a simple task, even in the case of languages that are well documented with texts whose authors and dates of composition are known. As William Labov has famously remarked (1994: 11), historical linguistics is “the art of making the best use of bad data.” In the case of Biblical Hebrew, we have to make the best use of very bad data. Nevertheless, we still aim to arrive at the most plausible scenario we can, using all the evidence available to us. In the case of Early Hebrew, this means relying heavily, though not exclusively, on the biblical texts. Given the doubts about the circumstances in which these texts were created, linguistic arguments have played, and will continue to play, an important role in establishing their provenance. But this puts us in the somewhat uncomfortable position, as LDBT reminds us, of using the language to date the texts, and then using the texts as evidence for the history of the language. In this situation, it follows that dating the texts cannot be our primary goal; rather, establishing a plausible history of the language is a prerequisite to dating texts. This is because dating a text using linguistic evidence is a more difficult problem than establishing a diachronic sequence for a language. If we

24

B. Elan Dresher

have some notion of the history of the language, we can say, for example, that a given form in a given text comes from an earlier or later stage of the language (or alternatively, from this or that synchronic dialect). But making this sort of determination still leaves many unanswered questions about how the text as a whole came to have this form in it. It could represent the date the text was composed, or it could be a later insertion into an older text, or a borrowing from another dialect, and so on. If the editorial history of a text is particularly complex, there may not be a well-defined answer to the question “To what date should this text be assigned?” Thus, LDBT may well be correct in asserting (2.100) that “the outward form of the biblical texts was in constant flux. In this context, the question of the ‘original date’ when a biblical book was composed is anachronistic and irrelevant.” My disagreements with LDBT concern its model of linguistic change in general and the history of Hebrew in particular. 4.  An Example of Variation: mamlākâ and malkût Let us take as an example the much discussed variation between several Biblical Hebrew forms for ‘kingdom’, in particular mamlākâ and malkût. According to LDBT (1.21 n. 21), the distribution of these forms is considered a “classic illustration” of a diachronic shift, with mamlākâ being the older form and malkût the newer form. A table showing the number of occurrences of each form in each book is given in (3). (3)  Number of Occurrences of mamlākâ and malkût in Biblical Texts Book

mamlākâ

malkût

% malkût

Book

mamlākâ

malkût

% malkût

Gen

2

0

0

Mic

1

0

0

Exod

1

0

0

Nah

1

0

0

Num

2

1

33

Zeph

1

0

0

Deut

7

0

0

Hag

2

0

0

Josh

2

0

0

Ps

6

6

50

1 Sam

6

1

14

Lam

1

0

0

2 Sam

6

0

0

Qoh

0

1

100

1 Kgs

12

1

8

Esth

0

26

100

2 Kgs

5

0

0

Dan

0

16

100

Isa

14

0

0

Ezra

1

6

86

Jer

17

3

15

Neh

1

2

67

Ezek

4

0

0

1 Chr

3

11

79

Amos

3

0

0

2 Chr

19

17

47

Total

117

91

44

Methodological Issues in the Dating of Linguistic Forms

25

If we sort the books in terms of rising percentage of malkût, we can display them on a chart as in (4): books with less than three examples in both columns have been omitted. The books on the left side of the chart are, for the most part, those that are conventionally considered to be early, and the books on the right side are considered to be late. This is a very rough first approximation: we are assuming that all the books can be treated as uniform wholes, which is not the case. (4)  mamlākâ and malkût (at least 3 of either form)

A diachronic interpretation of this distribution appears to be supported by extrabiblical attestation, as LDBT points out. In (5) is a listing of the distribution of these forms in Ben Sira, Qumran nonbiblical documents, and the Mishnah (numbers provided by Robert Holmstedt). They also fit in on the right side of the chart, as we might expect if they are late books. (5)  Extrabiblical Occurrences of mamlākâ and malkût Book

mamlākâ

malkût

% malkût

Ben Sira

3

2

40

Qumran

36

52

59

0

20

100

39

74

65

Mishna Total

5.  LDBT’s Central Argument against a Diachronic Interpretation of Variation LDBT does not accept the conventional diachronic interpretation of the distribution of these forms. The authors’ arguments with respect to this example

26

B. Elan Dresher

are fairly typical of their general position, so it is worth considering them in some detail. Their first argument runs as follows: if malkût is a late form, then its appearance in a text indicates that the text is late. This would make Numbers, Samuel, and Kings late books—a conclusion that would be unacceptable to almost all writers who support a distinction between Early Biblical Hebrew (EBH) and Late Biblical Hebrew (LBH). Anticipating the obvious reply to this line of reasoning, LDBT presents perhaps its central argument against the entire diachronic project (1.86; cf. 2.84–85): If against this is it argued that the LBH linguistic feature found in the EBH text is not actually ‘late’ but was also available in an early period, then its value for dating texts ‘late’ is negated. . . . Therefore, if EBH texts are early, and most LBH features are attested in EBH texts, then LBH features already existed in an early period, and were available to early authors, and thus their use is a matter of style, not chronology.

As stated, this argument is untenable. It is a well-attested fact in many languages that competing forms may coexist over a period of time, and thus a late form may occur sporadically in early texts, and an early form may survive in late texts. It is an empirical question, in any given case, whether the distribution of forms has a diachronic dimension or not; there is no basis for ruling out chronology as part of the story. Thus, it does not follow from the mere fact of coexistence that all the variation in the distribution of these forms must be stylistic and not diachronic. Let us consider more closely the claim that, if a late feature existed at an earlier period, it was therefore “available” to early authors. This notion of “availability” is contrary to findings in historical linguistics that much synchronic variation has a diachronic trajectory. Contrary to LDBT’s assertion that the coexistence of competing forms “negates” their value for dating texts, it can be shown that the proportion in which the forms occur has a characteristic signature in a given time and place and can have considerable predictive value in dating a text. 6.  The Rise of English Periphrastic Do In this connection I would like to look at the rise of periphrastic do in English. In present-day English, an auxiliary verb do must appear in a variety of contexts, as shown in (6). (6)  Present-day English contexts requiring periphrastic do 1.  Negative declarative sentences (a)  She does not deserve it. (b)  *She deserves it not.

Methodological Issues in the Dating of Linguistic Forms

27

2.  Negative imperatives (a)  Do not look at the answers. (b)  *Look not at the answers. 3.  Yes-no questions (a)  Do you know the answer? (b)  *Know you the answer? 4.  wh-adverbial questions (a)  Why does she deserve a reward? (b)  *Why deserves she a reward? In each type of sentence in (6), the (a) sentence with do is grammatical, and the (b) sentence, in which the main verb moves to a position before the negative marker or the subject, is ungrammatical (as indicated by the asterisk).  4 In Old and Middle English, the equivalents of the (b) sentences were all grammatical, and do was not used in these constructions. Beginning around the year 1400, verbs, with the exception of be, have, and modals (shall, will, may, can, etc.), began to lose the ability to move to the front of the sentence. Periphrastic do began to be used in sentence types that require the tensed verb to be before a subject or not. This change began slowly and took hundreds of years to complete. A graph showing the percentage of do in different types of sentences is shown in example (7) on p. 28. Building on Kroch (1989) and Han and Kroch (2000), Warner (2006) demonstrates that the changes in the percentage of do in the different sentence types advance in lockstep. The reason, according to these authors, is that a single basic change in the grammar affects all these sentence types, and in each period the old grammar and the new grammar coexist in a proportion that manifests itself in each type of sentence. The graph also illustrates another characteristic of language change— namely, the S-shaped curve of an innovation. Thus, periphrastic do advances relatively slowly at first until just before the year 1500, when it takes off and rises at an increasing rate (with some local dips) until it reaches about 90%, at which point the rate of change necessarily slows as the change moves to completion. Warner (2006) also argues for a stylistic influence on the development of do in negative declarative sentences (the dotted line in the middle of the graph). Beginning in about 1575, the percentage of periphrastic do in this type of sentence dipped and fell far behind the affirmative sentences (the solid line 4.  Some readers may find some of the (b) sentences to be high rhetorical or a bit archaic but not ungrammatical. These constructions can become familiar through exposure to Shakespeare or the King James Version of the Bible. However, for contemporary children learning English, or in the context of informal conversation, sentences of this sort would definitely be odd enough to be considered ungrammatical.

28

B. Elan Dresher

(7)  Percentage of Do in Different Types of Sentences (Ellegård 1953: 162; cited by Warner 2006: 48)

just above it). This deviation is an apparent counterexample to the claim that periphrastic do increased at a constant rate across sentence types. Warner argues that a more detailed analysis shows that the dip was not universal but occurred mainly in texts of what he calls “high lexical complexity”—that is, texts that use longer words and a greater variety of words. Such texts are more literary and sophisticated than texts with low lexical complexity, which tend to be more colloquial and closer to speech. Warner proposes that the drop in do not after 1575 in texts of higher lexical complexity was due to a stylistic avoidance of the sequence do not. This stylistic dispreference did not extend to texts of lower lexical complexity, meaning that, in the spoken language, do not continued to advance, and over the long run this stylistic tic did not significantly impede the rise of periphrastic do in negative declaratives. The conclusions I want to draw from this example are summed up in (8).

Methodological Issues in the Dating of Linguistic Forms

29

(8)  Conclusions from the rise of periphrastic do a.  Old and new forms can coexist over a long period. b.  This coexistence is not static but changes systematically over time. c.  The proportions of old and new forms are highly significant and can be used to estimate the date of texts. d.  There is a place for stylistic variation, but the stylistic influences are specific and occur in the context of ongoing diachronic change. When we apply these conclusions to the Hebrew example of the words for ‘kingdom’, it follows that the differing proportions of the two forms in different texts could well point to a diachronic difference in the texts. LDBT ignores differences in proportions and considers only presence versus absence of forms, a criterion that would miss the entire diachronic development of periphrastic do. Thus, they observe (1.88) that the EBH books Numbers, Samuel, and Kings have both mamlākâ and malkût, “but so do Jeremiah, Ezra, Nehemiah and Chronicles.” The suggestion is that these books are all the same in having both forms. The table in (9) reprises the numbers of forms in these books. (9)  Distribution of mamlākâ and malkût in Selected Books Book Numbers

mamlākâ

malkût

% malkût

2

1

33

Samuel

12

1

8

Kings

17

1

6

Jeremiah

17

3

15

1

6

86

Ezra Nehemiah

1

2

67

Chronicles

22

28

56

If we exclude Numbers and Nehemiah, which have only three forms each, it is apparent that the two groups of books (actually three, since Jeremiah occupies an intermediate position) are quite different with respect to the distribution of the two forms. In fact, they fit quite well the conventional division of books into periods reviewed by LDBT (1.11), as shown in example (10) on p. 30. 7.  External Attestation and Dialect Variation We observed above that the diachronic interpretation of this distribution is supported by extrabiblical attestations in Ben Sira, Qumran Hebrew, and the Mishnah. LDBT is not impressed by these facts, however, arguing that Mishnaic Hebrew (MH) is not in fact later than Biblical Hebrew. Again, the

30

B. Elan Dresher

(10)  Conventional Division into Periods

reasoning is flawed and rests on unrealistic assumptions about dialects and language change. LDBT (2.76) argues that the “nineteenth-century model of a steady development from EBH to LBH to MH is in conflict with the evidence.” This may be true but is of dubious relevance, because a diachronic account of differences between MH and BH does not depend on this model. Scholars such as Kutscher (1982) and Sáenz-Badillos (1993) also reject the nineteenth-century model but still accept that late biblical texts would be expected to show more MH elements than early texts. The general consensus is that MH developed from a vernacular dialect of Hebrew, whereas BH was a literary language that coexisted with vernacular dialects (Bar-Asher 1999). LDBT adopts a similar position but with a significant twist (2.77): “MH is an independent Hebrew dialect of great antiquity. Both ‘Aramaisms’ and ‘Mishnaisms’, far from being markers of a late date, were available in all periods of Hebrew.” The above quotation appears to suggest that there was no diachronic development in the MH dialect. It is one thing to say that Mishnaic Hebrew develops from vernacular dialects that can be traced back to preexilic times (see Bar-Asher 1999). It does not follow from this that preexilic Mishnaic

Methodological Issues in the Dating of Linguistic Forms

31

Hebrew forms all persisted unchanged for hundreds of years and could appear in any proportion in any text written in this period. If this were the case, then, as Delitzsch (1877: 190) remarked in a similar context, there is no history of the Hebrew language! 8.  Diachronic Discontinuities Maybe the authors of LDBT do not intend this radical interpretation. Perhaps they allow that MH changed over the course of hundreds of years. They could still argue that the fact that Biblical Hebrew did not become Mishnaic Hebrew poses problems for diachronic interpretations of variation, because “MH is simply a different dialect of Hebrew” (1.227). In this respect, Hebrew is not so different from other languages the history of which is better documented. The earliest attested examples of Old English, for example, tend to come from the Northumbrian dialect in the north. Beginning around 715, the Mercian kingdom in the Midlands became ascendant, and the Mercian dialect became the standard. In 825, the West Saxons in the south defeated the Mercians, and West Saxon became the standard until the end of the Old English period. Therefore, in studying the history of Old English, as one moves back in time, one also moves farther north. For example, a Mercian form from 700 is both older and from a different dialect than a West Saxon form from 1000. But even though early Mercian is not the ancestor of late West Saxon, for many purposes one can pretend that it is. The reason is that in many respects these dialects were similar and underwent many of the same diachronic changes. Thus the Mercian form might reveal to us the original vowels that appear in reduced form in later West Saxon. The main point is that dialect differences do not negate diachrony but must be considered together with diachrony. Toon (1983) discusses the problem of variation in Old English texts. He writes (1983: 106–7), “It is important to students of the language that variable data need not preclude, as it has for some, meaningful analysis.” One problem he considers is the spelling of the vowel in the Old English ancestor of the word ‘man’. We observe the distribution of spellings shown in (11). (11)  Spellings of the Vowel in Mercian Old English ‘Man’ Text

Date

a

o

Epinal Glossary

ca. 700

58

1

Erfurt Glossary

ca. 750

32

33

Corpus Glossary

ca. 800

38

95

Vespasian Psalter

ca. 830

none

all

Toon observes that the mixed spellings in the glossaries might lead one to suppose that they are the result of dialect mixture or of idiosyncratic stylistic

32

B. Elan Dresher

choice. But he argues that one can make sense of the variation in terms of diachrony. Like other Old English dialects, early Mercian originally had the vowel /a/ in the word mann ‘man’. A sound change then occurred in Mercian whereby /a/ became o before a nasal consonant. We can see the very beginning of this change in the Epinal Glossary. The later texts reflect later stages in which the change was either becoming more established in the spoken language or, alternatively, was becoming more acceptable as a written form. By the time of the Old English gloss of the Vespasian Psalter, o was the only option. It would be misleading and unproductive, in this case, to argue that both a and o spellings were “available” to Mercian scribes in the entire period 700–830, and that therefore the choice of one over the other was a matter of style, not chronology. Both spellings overlapped for a time, but they were not equally “available”; their distribution has a chronological as well as synchronic dimension. As Toon (1983) shows, there is also a political dimension to the variation in spellings. The table in (12) is a summary of spellings in Kentish charters. (12)  Spellings of the Vowel of Old English ‘Man’ in Kentish Charters Dates

a

o

Before Mercian influence

Period

679–741

5

0

Mercian ascendancy

803–824

0

64

End of Mercian influence

833–870

23

65

After Mercian exodus

859–868

13

3

958–1044

25

0

Late Kentish

In Toon’s interpretation, the Kentish vernacular dialect never underwent the change of a to o. The o spellings reflect the Mercian standard; once Mercian influence was gone, the a spellings return. This example shows that we cannot simply label a spellings as “early” and o spellings as “late.” This equation does hold within Mercian, but it is only part of the story. If we include other dialects, we see that it is also true that o spellings are “northern” and a spellings are “southern.” In the Kentish documents, where a forms are both early and late, it can be said that o spellings reflect the official standard spelling and a spellings the vernacular. All these dimensions play a role in fashioning a coherent account of the variation in spellings. It is also relevant to note that the dialects of the texts are not necessarily different stages of a single dialect. I have argued (Dresher 1985), following Kuhn (1939), that the dialect of the Corpus Glossary is not the direct descendant of the dialect of the Epinal Glossary but is more like a younger sister. Thus, it can be shown that two sound changes that were active in the same period reached these dialects in different orders, as displayed in (13). The fact that the Epinal

Methodological Issues in the Dating of Linguistic Forms

33

Glossary dialect coexisted with the Corpus Glossary dialect does not preclude us from assigning a diachronic dimension to the variation in these documents. (13)  Sound changes in different orders a.  Epinal Glossary: Second Fronting of a before Back Mutation Earlier forms fatu weras Second Fronting fætu Back Mutation fæatu weoras Gloss ‘vessel, nominative/ ‘man, nominative/ accusative plural’ accusative plural’ b.  Corpus Glossary: Back Mutation before Second Fronting of a Earlier forms fatu weras Back Mutation — weoras Second Fronting fæatu 9.  Accounting for Variation in the Biblical Texts Returning to Biblical Hebrew, the central empirical problem we are dealing with is: what is the best way to account for the variation in the texts? In the particular example we have been looking at, how can we account for the distribution of mamlākâ and malkût in the biblical texts? In (14), I summarize the two theories in front of us. (14)  Two theories of the variation of mamlākâ and malkût a.  The “chronological” theory (i)  Diachronic Mamlākâ is the earlier form, and malkût is a later form. Books with mixed forms show different stages in the rise of malkût. (ii)  Stylistic Chronicles and Ezra are both late, but the former was more concerned to imitate elements of the earlier grammar. b.  LDBT’s theory: “Multiple contemporary styles of literary Hebrew” (i)  Stylistic Mamlākâ predominates in books written in the “conservative” (EBH) style (= “moderate, cautious, avoiding extremes,” not older); malkût is preferred in the style (LBH) that is “more open to using a variety of linguistic forms.” In (14a) is the conventional theory, what LDBT calls the “chronological model” (2.85). It accounts for much of the distribution by diachrony but does not attribute all variation to chronology. The difference in (10) between Chronicles

34

B. Elan Dresher

and Ezra, for example, may be partly explained in terms of style, in that the former was more concerned to imitate certain elements of the earlier grammar. In other cases, dialect differences have also been invoked, as well as genre differences between prose and poetry. LDBT (2.96) proposes replacing this theory with (14b), “a model of multiple contemporary styles of literary Hebrew.” They designate EBH as a “conservative” style, whereas “LBH authors/editors/scribes are more open to using a variety of linguistic forms” (1.141). They hasten to stress (1.141 n. 91) that they “use ‘conservative’ here in the sense of ‘moderate, cautious, avoiding extremes’ rather than conservatism in the sense of favouring an older style . . . both the conservative and non-conservative styles co-existed throughout the period of the composition of the biblical literature.” Let us now consider the empirical status of the two theories, summed up in (15). (15)  Empirical status of the two theories a.  The “chronological” theory (i)  LDBT presents no compelling argument against this model. (ii)  The variation profile is entirely consistent with what we would expect to find and is in keeping with the English cases we have examined. b.  LDBT’s theory has no testable empirical consequences; therefore, it does not explain why mamlākâ and malkût occur in the attested proportions. Looking first at the chronological theory in (15a), I have argued that LDBT presents no compelling argument against this model. Moreover, the variation profile is entirely consistent with what we would expect to find and is in keeping with the English cases that we have looked at. Of course, this does not prove that the diachronic account is correct but only that it is plausible and consistent with the evidence that we have reviewed. Let us turn to LDBT’s alternative. Does this theory explain the variation in the forms mamlākâ and malkût as we find them in the texts? I do not see how it does. Why was mamlākâ considered a conservative form and malkût not? We can no longer say it is because mamlākâ was an older form or belonged to a more prestigious dialect. While rejecting these hypotheses, LDBT does not replace them with anything that can explain why EBH and LBH have the properties that they do. How do we account for the variation in books that contain both these forms, and why in the proportions that they do? The chronological hypothesis suggests an answer—perhaps a wrong answer, but something we can try to support further or disconfirm. But LDBT suggests in the end that all variation is due to “stylistic choices of authors and scribes” (2.95). Because

Methodological Issues in the Dating of Linguistic Forms

35

these choices are “unpredictable,” the proposed model has no testable empirical consequences. I mentioned at the outset that some model of how Hebrew developed, some notion of chronological stages and dialects, is a prerequisite to being able to date texts, because we need to have some sense of where the forms in the texts originate. By removing time and space from consideration, the authors of LDBT make it impossible to arrive at a coherent model of the history of Hebrew. 10.  An Analogy: Coexisting Achaemenid-Period Aramaic Styles in Elephantine What is lost is nicely illustrated by an analogy that LDBT (1.294; 2.99) draws between the authors’ own proposal and coexisting styles of Achaemenidperiod Aramaic in Elephantine as portrayed by Kutscher (1970: 362) and Folmer (1995: 709–10). According to these authors, there were two dialects of Aramaic, Eastern and Western, existing at the same time. Note here the introduction of geography. Writers in Elephantine, which is in the west, used the Western dialect in their ordinary writings, as we might expect. The Western dialect is in greater continuity with Old Aramaic than the Eastern; note here the introduction of a diachronic dimension. When writing legal documents, the Elephantine writers wrote, as we do, in a more conservative style that had older elements of the language (more diachrony). The Eastern dialect was more innovative and had more Persian loanwords (geography again). In letters directed to the Persian authorities, Elephantine scribes tried hard “to write in the official style of the royal chancelleries” (Folmer 1995: 727)—that is, in the Eastern dialect; here is a political dimension. (16)  Elements of the Kutscher-Folmer account (LDBT 1.294; 2.99) a.  Geographic (i)  Two coexisting dialects of Aramaic—Eastern and Western. Eastern Aramaic (in Persia) has more Persian and Akkadian loanwords. (ii)  Elephantine writers use their native Western dialect in private letters. b.  Diachronic (i)  The Western dialect is closer to Old Aramaic than the Eastern. (ii)  Elephantine legal documents are in a more conservative (= older) style. c.  Political In letters directed to the Persian authorities, Elephantine scribes try hard “to write in the official style of the royal chancelleries” (Folmer 1995: 727)—that is, in the Eastern dialect.

36

B. Elan Dresher

I think this is a very plausible and convincing analysis. Here is what LDBT (2.99) says about it: “It shows us that there is no need to posit chronological or geographical distance to explain the use of different styles of language.” But we have seen that both chronology and geography in addition to politics are crucial in explaining why the various styles are the way they are. LDBT has in mind that the same community in the same time and place could produce two different styles of writing; but without a diachronic and synchronic account of Aramaic, we would not be able to make sense of these two different styles. LDBT suggests that its account of EBH and LBH is very similar to the account of the two types of Aramaic produced in Elephantine. LBH writers, they propose, were trying to “distance this style of literature from literature produced in the EBH style. Rather than geographical or chronological distance, we would have intellectual or ideological distance.” However, without history or geography, or even a clear idea of who the two groups were, we have none of the elements that make the Elephantine analysis so compelling. Rather, juxtaposing the Kutscher-Folmer account with LDBT’s only serves to highlight the elements that LDBT is lacking. 11.  A Methodology for Biblical Hebrew Linguistics To return again to our example of mamlākâ and malkût: I have argued that LDBT does not provide a real alternative to the “chronological” model in (14a). This does not mean that this model is correct. There will always be a number of ways to account for the variation in any one feature studied in isolation. The real challenge is to arrive at a consistent model that can account for all the variation in the biblical texts, or as much of it as is feasible. This model should make use of any internal or external evidence available and should incorporate contemporary theories of linguistic change and typology. Thus, we can consider the chart in (4) to give us a profile of the variation between mamlākâ and malkût. We can similarly plot the profiles of other variable features. As DeCaen (2001) has argued, the traditional division into EBH and LBH is too simplistic: language change does not present us with only “early” features and “late” features. Rather, every linguistic change follows its own route. As we saw with the Mercian glosses, changes begin at different times in different places and move at different rates. Therefore, we do not expect every variable feature to give us the same profile as mamlākâ and malkût. The grid we need to construct is not one-dimensionally diachronic but multidimensional, including time and space as well as genre, politics, and style. This sort of project was proposed by DeCaen (2001: 23): “One form or one contrast yields precious little, but all possible variants statistically correlated should yield much.” Though I have taken issue with LDBT’s methodology and some of its conclusions, the authors’ detailed discussion and compilation of many variants will be a great assistance in carrying this project forward.

Methodological Issues in the Dating of Linguistic Forms

37

12. Old English Again I began by showing how the problem of dating Biblical Hebrew texts is different from the parallel problem in Old English. But in some instances, Old English presents similar difficulties. This is the case in trying to date the language of poems Beowulf, for example, survived in a manuscript from the end of the tenth century but was probably composed much earlier. The language shows a mixture of forms that suggest a complex history. Friedrich Klaeber, editor of the authoritative edition of Beowulf, had this to say about linguistic tests for dating Old English poems (1950: cviii–cix; footnotes omitted); I think it holds equally well for Biblical Hebrew: Investigations have been carried on with a view to ascertaining the relative dates of Old English poems by means of syntactical and phonetic-metrical tests. . . . It must be admitted that these criteria are liable to lead to untrustworthy results when applied in a one-sided and mechanical manner and without careful consideration of all the factors involved. Allowance should be made for individual and dialectal variations, archaizing tendencies, and . . . scribal alterations. . . .Yet it cannot be gainsaid that these tests, which are based on undoubted facts of linguistic development, hold good in a general way.

References

Amos, A. C. 1980 Linguistic Means of Determining the Dates of Old English Literary Texts. Cambridge, MA: Medieval Academy of America. Andersen, F. I., and Forbes, A. D. 1986 Spelling in the Hebrew Bible. Rome: Pontifical Biblical Institute. Bar-Asher, M. 1999 Mishnaic Hebrew: An Introductory Survey. HS 40: 115–51. Campbell, A. 1959. Old English Grammar. Oxford: Clarendon. DeCaën, V. 2001 Hebrew Linguistics and Biblical Criticism: A Minimalist Program. Journal of Hebrew Scriptures 3: article 6. http://www.purl.org/jhs. Delitzsch, F. 1877 Commentary on the Song of Songs and Ecclesiastes. Translated by M.  G. Easton. Edinburgh: T. & T. Clark. Dresher, B. E. 1985 Old English and the Theory of Phonology. New York: Garland. Ellegård, A. 1953 The Auxiliary ‘Do’: The Establishment and Regulation of Its Use in English. Stockholm: Almqvist & Wicksell. Folmer, M. L. 1995 The Aramaic Language in the Achaemenid Period: A Study in Linguistic Variation. Leuven: Peeters.

38

B. Elan Dresher

Gogel, S. L. 1998 A Grammar of Epigraphic Hebrew. Atlanta: Scholars Press. Han, C., and Kroch, A. 2000 The Rise of Do-Support in English: Implications for Clause Structure. Pp.  311–25 in Proceedings of the 30th Annual Meeting of the North East Linguistics Society, ed. Masako Hirotani et al. Amherst, MA: GLSA. Hogg, R. M. 1992a A Grammar of Old English, vol. 1: Phonology. Oxford: Blackwell. 1992b Phonology and Morphology. Pp. 67–167 in The Cambridge History of the English Language, vol. 1: The Beginnings to 1066, ed. R. M. Hogg. Cambridge: Cambridge University Press. Klaeber, F. 1950 Beowulf and the Fight at Finnsburg. 3rd ed. Boston, MA: Heath. Kofoed, J. B. 2006 Using Linguistic Difference in Relative Text Dating: Insights from other Historical Linguistic Case Studies. HS 47: 93–114. Kroch, A. 1989 Reflexes of Grammar in Patterns of Language Change. Language Variation and Change 1: 199–244. Kuhn, S. M. 1939 The Dialect of the Corpus Glossary. PMLA 54: 1–19. Kutscher, E. Y. 1970 Aramaic. Pp. 347–412 in Current Trends in Linguistics, vol. 6: Linguistics in South West Asia and North Africa, ed. T. A. Sebeok. The Hague: Mouton. 1982 A History of the Hebrew Language, ed. R. Kutscher. Jerusalem: Magnes. Labov, W. 1994 Principles of Linguistic Change: Internal Factors. Oxford: Blackwell. Sáenz-Badillos, A. 1993 A History of the Hebrew Language, trans. J. Elwolde. Cambridge: Cambridge University Press. Toon, T. E. 1983 The Politics of Early Old English Sound Change. New York: Academic Press. Warner, A. 2006 Variation and the Interpretation of Change in Periphrastic Do. Pp. 45–67 in The Handbook of the History of English, ed. A. van Kemenade and B. Los. Oxford: Blackwell. Young, I.; Rezetko, R.; and Ehrensvärd, M. 2008 Linguistic Dating of Biblical Texts. 2 vols. London: Equinox. [LDBT] Zevit, Z. 1980 Matres Lectionis in Ancient Hebrew Epigraphs. Cambridge, MA: American Schools of Oriental Research.

Biblical Hebrew as a Diachronic Continuum T. Givón 1. Introduction The rationale for interpreting grammatical variation across the Biblical Hebrew (BH) texts as a diachronic continuum is remarkably straightforward and may be presented based on the old adage: if it looks like a duck, walks like a duck, and quacks like a duck, then by golly it must be a duck. Translated into the current context: if the cross-textual grammatical variation in Biblical Hebrew lines up along known, universal, directional trends of diachronic change, then it must represent a diachronic continuum. The force of must here is not that of logical necessity but, rather, of plausible hypothesis. Like all useful hypotheses, ours makes testable predictions that can either be falsified or not.  1 In this particular case, the predictive force of our diachronic hypothesis may be rendered as follows: (1)  If ​ multiple grammatical variables turn out to line up the same books of the

Old Testament in the same temporal order, then the hypothesis about each individual variable is correspondingly strengthened.

Obviously, the methodology I propose to employ here depends on conceding the existence of universals of grammatical change. Under the current label of grammaticalization, the study of the diachrony of grammar harks back to the very roots of modern linguistics in the nineteenth century, with illustrious forebears such as Franz Bopp (1820) and Hermann Paul (1890), among others. More-modern renditions and elaborations may be found in Givón (1971; 1979), Heine et al. (1991), Traugott and Heine (1991), Hopper and Traugott (1993; 2003), Bybee et al. (1994), Givón (2000), Heine and Kuteva (2007), and Givón Author’s note: The materials presented here have been condensed and in some cases reanalyzed from two previous works (Givón 1977; 1991). I am indebted to Agustino Gianto, Eitan Grossman, and Bernd Heine for comments and suggestions. The imperfections that remain in the manuscript are entirely my own. 1.  As is the case for all explanatory hypotheses in science in general, the alternative of no hypothesis leaves vexing puzzles either unexplained or else explained in a less-systematic, more ad hoc fashion.

39

40

T. Givón

(2009), among others. For the purpose of this essay, the following universal directional trends in grammaticalization are invoked: directional trends in grammaticalization (2)  Universal ​ a.  Morphologization: Grammatical morphemes arise from lexical words. b.  Cliticization: When lexical words are cliticized as morphemes, they become phonologically reduced, shortened and destressed. c.  Irregularity: Morphophonemic irregularity and variability increases with the age of grammatical morphemes. d.  Syntacticization: Tightly-packed syntactic constructions arise from loosely-concatenated paratactic configurations. e.  Functional motivation: Grammatical change is communicatively motivated, so that functional extension always precedes structural readjustment. f.  Functional ambiguity: Consequently, the early stages of grammatical change are characterized by functional ambiguity, whereby the construction performs both its old and new functions. g.  Word-order change: The directionality of word-order change (without substratum contact) tends to be: SOV > free/pragmatic word-order > VSO/VOS > SVO A strong claim implicit in universals (2a–g) is that of unidirectionality; that is, grammatical change proceeds, overwhelmingly, in the directions noted in (2a–g) but not in the opposite directions. Many observed universal trends are specific to particular grammatical constructions or morphemes. To the extent that all these directional trends are valid, they allow us to make precise judgment about a grammatical construction or morpheme’s being older or younger along a diachronic continuum. In the case of Biblical Hebrew, this allows us to place at least some of the books of the Old Testament along this sort of continuum. 2.  Texts and Grammatical Variables The Old Testament as a whole is a complex mix of early oral traditions that were later pressed into written form, together with texts that were probably written originally. Natural diachronic change, however, takes place overwhelmingly in the spoken language. As long as a written text reflects, albeit with considerable time lag, the naturalness of a spoken dialect; and as long as the scribes were fluent native speakers of the language, one could assume that their grammar was coherent and natural. But to a grammarian, it is fairly clear that many late texts in the Old Testament, especially texts committed to writing or reedited after the Babylonian Exile, were written by people whose first

Biblical Hebrew as a Diachronic Continuum

41

language was not Hebrew. Consequently, vexing methodological issues plague the selection of Old Testament books to be investigated in a study such as this. In principle, all books can be subjected to the sorts of tests used here, because the grammatical variables chosen are of sufficient text-frequency. This assumption of general applicability raises relatively few problems with the Pentateuch and the Former Prophets. However, problems begin to crop up with some of the later texts. Three examples illustrate the point. On our proposed diachronic continuum, the book of Lamentations is problematic. Its language seems progressive (“later”) according to some grammatical variables but conservative (“earlier”) according to others. The book of Job is by all philological evidence a relatively late book, reflecting contact with Platonic Greek philosophy following the Alexandrian conquest. But its grammar is as conservative as that of Genesis and gives all appearances of a non-native-speaking scribe using older written texts as a stylistic template. Finally, Song of Songs (Canticles) may have been codified either early or late, but its grammar is the most progressive (“late”) on our proposed continuum. The reason probably has to do with sociolinguistics with regard to both the genre and the temporal provenance—this book is couched in the most earthy, folksy, oral language of the Old Testament, reflecting more closely the spoken language of its later era. Another issue in selecting BH texts for this type of study is potential genre differences with regard to the distribution of grammar and usage. Poetic texts represent a potential problem, because their language often harks back to an older layer of usage, as compared with the narrative within which they are embedded. As much as possible, I try to deal with comparable narrative texts.  2 It is beyond my competence and the scope of this essay to resolve all these issues. The texts studied here were selected by purely linguistic criteria: the presumed grammatical changes must line up along the same directional continuum for all variables studied and thus conform to our prediction (1). The strength of our proposed method rests on the assumption that grammatical change is coherent rather than chaotic. The following books seem to line up coherently along the continuum: (3)  The ​ Biblical Hebrew diachronic continuum (i)  Early Biblical Hebrew (EBH) (a)  Genesis (chaps. 1–12 or 1–20) (b)  2 Kings (chaps. 1–12 or 1–20)

2.  The three obvious but unavoidable exceptions are Lamentations, which is an exhortative prayer text; Qoheleth, an aphoristic text; and Song of Songs, a highly poetic text with many dialogic features. Of the three, only Lamentations proved problematic.

42

T. Givón (ii)  Late Biblical Hebrew (LBH) (a)  Esther (entire text) (b)  Lamentations (entire text) (c)  Qoheleth (entire text) (d)  Song of Songs (entire text) (iii)  Mishnaic Hebrew (MH) (a)  M. Berakot (entire text)

The fact that Misnhaic Hebrew (ca. second century a.d.) conforms in all grammatical variables studied to the Song of Songs serves to uphold the usefulness of our methodology. 3 The grammatical variables studied were variables (4)  Grammatical ​ a.  Word order b.  Tense-aspect c.  Relative clauses d.  Contraction of the subordinator ʾăšer vs. šae.  Verbal complements f.  Adverbial clauses These are all relatively frequent constructions or morphemes—items that presumably do not correlate with the contents of the texts. 3.  Grammatical Variation in Biblical Hebrew 3.1.  Word-Order and Tense-Aspect While technically distinct, word order and tense-aspect in fact show a considerable measure of correlation in Early Biblical Hebrew. This is so because one word order, VSO (V-first), is strongly associated with the preterite (perfective; the wayyiqtol form, erroneously labeled ‘imperfect’), the most frequent tense-aspect in narrative text; while another, SVO (or X-V [i.e., another constituent precedes the verb], or T-V-X [i.e., the topic precedes the verb and another constituent follows the verb]), is strongly associated with either the perfect (the qatal form) or with stative copular-predicate clauses. This association is motivated by the pragmatics of thematic continuity, a metafunctional dimension that underlies both aspect and word-order choice (Hopper 1982; Givón 1983; 2001: vol. 1). As an example of the strong association of thematic (or referential) continuity with both the VS word order and the preterite/perfective tense-aspect in EBH, consider: 3.  In addition, there are many grammatical innovations in Mishnaic Hebrew that are not yet attested in Song of Songs.

Biblical Hebrew as a Diachronic Continuum

43

ʾĕlōhîm “yĕhî ʾôr” (5)  ​ a.  way-yōʾmer and-3ms/say/PRET God 3ms/be/IRR light ‘And God said, “Let there be light”’ (Gen 1:3a) b.  wa-yĕhî ʾôr and-3ms/be/PRET light ‘and there was light’ (Gen 1:3b)5 c.  way-yarʾ ʾĕlōhîm ʾet-hā-ʾôr kî ṭôb and-3ms/see/PRET God ACC-the-light SUB-good ‘and God saw the light that it was good’ (Gen 1:4a) d.  way-yabdēl ʾĕlōhîm bên hā-ʾôr û-bên ha-ḥōšek and-3s/divide/PRET God between the-light and-between the-dark ‘and God separated the light from the dark’ (Gen 1:4b) e.  way-yiqrāʾ ʾĕlōhîm l-ā-ʾôr yôm and-3ms/call/PRET God to-the-light day ‘and God called the light day (Gen 1:5a) f.  wĕ-l-ā-ḥōšek qārāʾ lāylâ and-to-the-dark call/3ms/PERF night ‘and the dark he called night’ (Gen 1:5b) In all but the last verbal clauses in (5), the V-first (or VS) order is associated with the preterite tense-aspect; in all but (5f), the topic/theme is continuous. The last clause switches the topic—here the object—from ‘light’ to ‘dark’, using a contrastive fronting device (“Y-movement” or “inverse”). This thematic shift precipitates the O-V order and the perfect tense-aspect. Some of the complexity of using the double contrast of tense-aspect-withword-order to mark thematic/topical contrast of continuity vs. discontinuity is further illustrated in: ʾet-ḥawwâ ʾišt-ô (6)  ​ a.  wĕ-hā-ʾādām yādaʿ and-the-man know/3ms/PERF ACC-Eve wife-3ms ‘And Adam knew his wife Eve’ (Gen 4:1a) b.  wat-tahar wat-tēled ʾet-qayin and-3fs/conceive/PRET and-3fs/give.birth/PRET ACC-Cain ‘and she conceived and gave birth to Cain’ (Gen 4:1b) c.  wat-tōʾmer “qānîtî ʾîš ʾet-Yhwh” and-3fs/say/PRET bought/1s/PERF man ACC-Yhwh ‘and she said, “I bought a man with [from?] God”’ (Gen 4:1c) d.  wat-tōsep lā-ledet ʾet-ʾāḥî-w ʾet-hābel and-3fs/add/PRET to-give.birth ACC-brother-3ms ACC-Abel ‘and she went on to give birth to his brother Abel’ (Gen 4:2a)

44

T. Givón wa-yĕhî hebel rōʿēh ṣōʾn and-3ms/be/PRET Abel herder/of sheep ‘and Abel was a sheep-herder’ (Gen 4:2b) wĕ-qayin hāyâ ʿōbēd ʾădāmâ and-Cain be/3ms/PERF worker/of soil ‘and Cain was a tiller of the soil’ (Gen 4:2c)

There are three loci of thematic discontinuity in (6): (6a) is a chapter-initial clause, introducing a brand new subject/topic—Adam, thus employing the perfect-with-SV-order. In (6b), Eve takes over as the subject/topical, having first been introduced as the object in (6a) and thus deemed continuous. The narrative continues in the preterite using V-first word order through (6b–d). 4 In (6c), the direct-quoted verbal complement is treated as a thematic lacuna and is thus marked by the perfect aspect. The V-first word order is the predictable consequence of referential continuity (Givón 1977). The subject/topic shift in (6e) is not considered a serious thematic break, since Abel already appeared in the preceding clause (6d) as the topical object. So (6e) proceeds with the preterite aspect and V-S word order. Finally, in (6f), the switch from Abel to Cain as subject/topic precipitates the use of the perfect with SV word order. A consistent if less-than-perfect deployment of word-order-with-tenseaspect combination to code continuity vs. discontinuity may also be seen in:  5 bārāʾ ʾĕlōhîm ʾet-haš-šāmayim (7)  ​ a.  bĕ-rēʾšît in-beginning create/3ms/PERF God ACC-the-sky   wĕ-ʾet hā-ʾāreṣ  and-ACC-the-earth ‘In the beginning God created the heaven and the earth’ (Gen 1:1) b.  wĕ-hā-ʾāreṣ hāytâ tōhû wa-bōhû and-the-earth be/3fs/PERF chaos and-confusion ‘and the earth was chaos and confusion’ (Gen 1:2a) wĕ-hōšek ʿal-pĕnê tĕhôm and-dark on face/of precipice ‘and darkness over the precipice’ (Gen 1:2b) wĕ-rûaḥ ʾĕlōhîm mĕrahepet ʿal-pĕnê ham-māyim and-spirit/of God hover/fs/PAR upon face/of the-water ‘and the spirit of God was hovering over the water’ (Gen 1:2c) 4. Referential continuity (here, of Eve) makes a subject-NP superfluous, given the obligatory subject pronominal agreement of Hebrew. 5.  Clause (7a) is another example of the less-than-perfect pairing of the perfect aspect and S-V (or X-V; or T-V-X; Vennemann 1973) “inverse” word order (Givón 1994). Because of traditional misunderstanding of the role of the perfect in EBH as a marker of thematic discontinuity, (7a) is sometimes mistranslated as a nominalized temporal ADV-clause—that is, ‘In the beginning of God’s creation of the sky and the earth . . .’.

Biblical Hebrew as a Diachronic Continuum

45

c.  way-yōʾmer ʾĕlōhîm “yĕhî ʾôr” and-3ms/say/PRET God 3ms/be/IRR light ‘and God said, “Let there be light”’ (Gen 1:3a) wa-yĕhî ʾôr and-3ms/be/PRET light ‘and there became light’ (Gen 1:3b) d.  way-yarʾ ʾĕlōhîm ʾet-hā-ʾôr kî ṭôb and-3ms/see/PRET God ACC-the-light SUB-good/ms ‘and God saw the light that it was good’ (Gen 1:4) Clause (7a) is text-initial: by definition, a context of thematic discontinuity. It is marked with the perfect tense-aspect, but the novel subject/topic—God— appears in the VS word order, probably because of the fronted adverb (see further §3.4 below). The subject/topic switch in (7b) is marked by the predicted combination—the perfect tense-aspect with SVO word order. The subject/ topic switch in (7d) is marked predictably with the SVO word order, but the imperfective aspectuality is coded by the participle. Imperfective forms are distributed in language as thematic lacunas, strongly associated with thematic discontinuity (Hopper 1982; Givón 2001: vol. 1). In (7e–g), at long last, the action picks up with the thematic backbone of the narrative and its main topic/ subject—God—with the predictable combination of the preterite-with-V-first word order. The next example shows lacunas of thematic discontinuity in both an ADVclause (8d) and a V-complement (8f, g, i), again marked by the perfect (or other discontinuity aspects, such as imperfective or irrealis), with the V-first order again the consequence of referential continuity: rāʿāb b-ā-ʾāreṣ (8)  ​ a.  wa-yĕhî and-3ms/be/PRET famine in-the-land ‘And there was famine in the land’ (Gen 12:10a) way-yēred ʾabrām miṣraymâ lā-Gûr šām and-3s/descend/PRET Abram Egypt-ALL to-live there ‘and Abram went down to Egypt to live there’ (Gen 12:10b) kî-kābēd hā-rāʿāb b-ā-ʾāreṣ SUB-heavy the-famine in-the-land ‘because the famine was heavy in the land’ (Gen 12:10c) b.  wa-yĕhî ka-ʾăšer hiqrîb lā-bôʾ and-3ms/be/PRET SUB-REL near/3ms/PERF to-come  miṣrāymâ  Egypt-ALL ‘and so when he neared Egypt’ (Gen 12:11a) way-yōʾmer ʾel-śāray ʾišt-ô and-3ms/say/PRET to Sarah wife-3ms ‘he told his wife Sarah’ (Gen 12:11b)

46

T. Givón hinnēh-nāʾ yādaʿtî kî ʾiššâ yĕpat marʾeh ʾāt lo-EMPH know/3ms/PERF SUB-woman pretty/of visage you ‘Indeed I have known that you are a good-looking woman’ (Gen 12:11c) wĕ-hāyâ kî-yirʾû ʾōt-āk ham-miṣrîm and-be/3ms/PERF SUB-3mp/see/IRR ACC-2fs the-Egyptians  wĕ-ʾāmrû  and-say-3mp/PERF ‘so if it were that the Egyptians saw you and said,’ (Gen 12:12a) ʾišt-ô zōʾt wife/of-3ms this/fs ‘“she is his wife”’ (Gen 12:12b) c.  wĕ-hargû ʾōt-î and-kill/3mp/PERF ACC-1s ‘and they would kill me’ (Gen 12:12c) wĕ-ʾōt-āk yĕhayyû and-ACC-2fs 3mp/let.live/IRR ‘and will let you live’ (Gen 12:12d)

Relative clauses are likewise thematic lacunas, often pointing back to previous loci in the discourse, with a high frequency of perfect aspect, though invariably with V-S order. 6 Thus, consider: ʾĕlōhîm ʾet-kōl- ʾăšer ʿāśâ (9)  ​ a.  way-yarʾ and-3ms/see/PRET God ACC-all REL make/3ms/PERF ‘and God saw all that he had done’ (Gen 1:31) b.  wĕ-ʾet kōl nepeš ha-ḥayyâ hā-rōmeśet ʾăšer and-ACC-all soul/ of the-animal the-crawling REL  šārṣû        ham-mayim  spawn/3mp/PERF the-water ‘and all the crawling living souls that the water had spawned’ (Gen 1:21) Last, the use of the perfect-with-SV to mark discontinuity is most apparent in the series of subject/topic switches in:

6.  The use of the VS word order in object REL-clauses and ADV-clauses, often in strong association with the perfect aspect, is common in languages that retain some subject-position flexibility, such as in Spanish, Arabic, Swahili, or Spoken German (Vennemann 1973). Such subject post-posing is associated with the topicalization—by fronting—of a non-subject argument (Vennemann 1973; Givón 1976; 2001: vol. 1, chap. 5). I used to think it also reflected the putative syntactic conservatism of subordinate clauses, but I am no longer as sure about this.

47

Biblical Hebrew as a Diachronic Continuum Table 1.  Tense-Aspect and Word Order in EBH Genesis Preterite VS SV

Perfect VS SV

Partic. VS SV

Pred. VS SV

Irrealis VS SV

main clauses 177  —   2  21  —   4   5  72  12  12 fronted OBJ/ADV  —  —  13  —  —   1   2  15   1  — subord. clauses   2  —  25  —   2  —  10   1   2  — total (379) 179  —  40  21   2   5  17  88  15  12 percent    47%    16%    1.8%    27%    7%

yullad gam-hûʾ (10)  ​ a.  û-lĕ-šēm and-to-Shem PASS/bear/3ms/PERF also-he ‘and to Shem (to him too) was born’ (Gen 10:21) b.  wĕ-ʾarpakšad yālad ʾet-šālah and-Arpachshad bear/3ms/PERF ACC-Shelah ‘and Arpachshad sired Shelah’ (Gen 10:24a) wĕ-šelah yālad ʾet-ʿēber and-Shelah bear/3ms/PERF ACC-Ever ‘and Shelah sired Ever’ (Gen 10:24b) c.  û-lĕ-ʿēber yullad šĕnê bānîm and-to-Ever PASS/bear/3ms/PERF two sons ‘and to Ever were born two sons’ (Gen 10:25) The distribution of the SV and VS word orders in the various tense-aspects and clause types in EBH is provided in table 1. Table 1 suggests some interesting associations: a.  The preterite shows a 100% association with VS word order. b.  The preterite appears almost exclusively (177/179, or 98%) in main clauses. c.  The preterite is the most frequent tense-aspect in the text (179/379, or 47%) as compared with 16% for the perfect and 27% for the stative/ copular. d.  The perfect in main clauses shows a 21/23, or 91% association with SV word order. e.  The perfect occurs most frequently (38/61 instances, or 62%) in either subordinate clauses or topic-shifting (“inverse”) construction.

48

T. Givón f.  Subordinate clauses show a 33/34 (97%) association with the VS word order. g.  The overall text frequency of the VS word order is 253/379, or 66.7%.

The overall text-frequency of the preterite in table 1 (47%) is probably under­ represented due to two factors: (1) zero-subject clauses (that is, pronominal agreement is part of the verb) were not counted in table 1; and (2) the preterite is the main coding venue for high thematic and referential continuity, with the latter commonly associated with zero-subject. To estimate the magnitude of this underrepresentation: the overall occurrence of the preterite—with and without an overt subject—was counted in the first 12 chapters of Genesis. The number of all preterite-marked clauses was 379, and 200 of them appeared in zero-subject clauses, as against 179 with an overt subject as listed in table 1. The percent of zero-subject, preterite-marked clauses in the 12 chapters is thus 200/379 or 52.7%. The underrepresentation of the text-frequency of the preterite in table 1 is thus considerable. While other tense-aspects may have suffered similar underrepresentation, their lower text-frequency and their stronger association with thematic discontinuity suggest a lower impact. At the tail end of the BH diachronic continuum, one finds the Song of Songs. By this stage, the two grammatical variables tested here—tense-aspect and word order—had undergone profound restructuring. First, the EBH perfect had become the main preterite-perfective tense-aspect in LBH verbal clauses (excluding verbless predicate clauses). And second, the highly-marked SV word order, strongly associated in EBH with the perfect, had become the more frequent, unmarked word order in Late Biblical Hebrew. To illustrate briefly the strong association between the perfect, SV word order and thematic or referential continuity, consider: ʾet-kuttān-tî ʾêkākâ ʾelbāš-ennâ (11)  ​ a.  pāšaṭtî take-off/1s/PERF ACC-dress-1 show 1s/put.on/IRR-3fs ‘I took off my dress; how shall I put it on?’ (Song 5:3a) rāḥaṣtî ʾet-raglay ʾêkākâ ʾăṭannĕp-ēm wash/1s/PERF ACC-feet-1s how 1s/dirty/IRR-3mp ‘I washed my feet; how shall I dirty them?’ (Song 5:3b) b.  dôd-î šālaḥ yād-ô min-ha-ḥōr lover-1s send/3ms/PERF hand-3ms from-the-hole ‘my lover sent his hand through the opening’ (Song 5:4a) û-mēʿ-ay hāmû ʿāl-āyw and-guts-1s buzz/3mp/PERF for-3ms ‘and my guts called for him’ (Song 5:4b) c.  qamtî ʾănî li-ptōaḥ lĕ-dôd-î rise/1s/PERF 1s to-open to-lover-1s ‘I got up to open (the door) for my lover’ (Song 5:5a)

49

Preterite Table 2.  Tense-Aspect and Word Order in LBH Song of Songs

main clauses fronted OBJ/ADV subord. clauses total (151) percent

Preterite VS SV 4 5 1 — — —

Perfect VS SV 7 18 4 0 15 2

Partic. VS SV — 8 — — — —

Pred. VS SV 16 65 — — 6 4

Irrealis VS SV — — — — — —

5 5 6%

26 20 30%

— 8 5%

22 65 57%

— — 0%a

a.  Due to the largely dialogic nature of the Song of Songs text, irrealis clauses, with either first- or second-person subject, are coded with pronominal agreement on the verb, thus with no overt NP.

wĕ-yād-ay nāṭpû môr and-hands-1s drip/3mp/PERF myrrh ‘and my hands dripped with myrrh’ (Song 5:5b) The text distribution of the VS vs. SV word order in the various tense-aspect and clause types in Song of Songs is presented in table 2 above. The more salient features of the distributions in table 2 may be summed up as follows: a.  The old preterite now constitutes only 6% of the total sample. b.  The perfect—in its function as the new preterite—now constitutes 30% of the sample. c.  Verbless predicate clauses form 57% of the total sample, probably due to the poetic nature of the text. d.  In main clauses, the perfect outweighs the old preterite 25/9. It now constitutes 25/34, or 73% of the total sample of main clauses. e.  Of main clauses, 101/123, or 82% now show SV word order. For the sake of brevity, I will present in table 3 below only the summary of my two main variables: 7 (1) the distribution percentage of the SV word order in main clauses; and (2) the overall text-frequency of the perfect form in all verbal clauses (excluding verbless predicate clauses) in the texts selected to represent our proposed diachronic continuum. As noted above, Lamentations is the only problematic text in our sample, because of overuse of the perfect. Otherwise, both variables move in the same direction, placing the bend in the curve of diachronic change and the acceleration of both changes somewhere between Esther and Lamentations.  8 7.  The full tabulations may be found in Givón 1977. 8.  While diachronic-change curves tend to be L-shaped, like typical learning curves in psychology, the BH books studied here are not plotted on a monotonic temporal scale.

50

T. Givón Table 3.  Text-Frequency of SV and Perfect Verbal Form Genesis 2 Kings Esther Lamentations Qoheleth Song/Songs MH m. Berakot)

VS 196 215 103 45 43 27 9

SV Total %SV Perfect Total 109 305 35% 61 274 94 309 30% 125 399 56 159 35% 76 226 49 94 52% 90 119 120 163 73% 60 185 96 123 78% 46 64 115 124 92.8 n/a n/a

%Perfect 22% 31% 33% 75% 32% 71% n/a

3.2.  Relative Clauses The dominant pattern of relativization throughout the BH continuum is finite, overwhelmingly with V-first word order and with the subordinator ʾăšer/ ša-. This may be seen in the following EBH examples: (12)  ​ a.  Subject REL-clause min-hā-ʾădāmâ ʾăšer pāṣtâ ʾet-pî-hā from-the-earth REL open3fs/PERF ACC-mouth-3fs from the earth that has opened its (gaping) mouth (Gen 4:11) b.  Direct-object REL clause way-yāśem šām ʾet-hā-ʾ ādām ʾăšer yāṣār and-3ms/put-PRET there ACC-the-man REL make/3ms/PERF and he put there the man that he had made (Gen 2:8) c.  Prepositional-object kōl-māqôm ʾăšer tidrōk kap-raglĕ-kem b-ô all place REL 3fs/step/IRR sole/of foot-3mp on-3ms every place that the sole of your foot steps on it (Josh 1:3) d.  Manner REL-clause kĕ-kōl ʾăšer šāmaʿnû ʾel-mōšeh like-all REL listen/1p/PERF to-Moses just (like) the way we listened to Moses (Josh 1:17) e.  Predicate, locative šĕnayim šĕnayim mik-kōl hab-bāśār ʾăšer b-ô rûaḥ ḥayyîm two two from-all the-flesh REL in-3ms soul/of life two from each flesh that has a living soul in it (Gen 7:15) f.  Predicate, possessive wĕ-ʾēt kōl ʾăšer lā-hem and-ACC all REL to-3mp and everything that belongs to them (Josh 2:13)

Biblical Hebrew as a Diachronic Continuum

51

The same finite pattern also predominates in REL-clauses in the LBH texts, with the subordinator ʾăšer now truncated to ša-: (13)  ​ a.  Subject REL-clause kĕ-ʿēder hā-ʿizzîm šeg-gālšû mē-har  Gilʿād like-herd/of the-goats REL-slide/3mp/PERF from-mount  Gilead like the herd of goats that slid down from Mount Gilead (Song 4:1) b.  Direct-object REL-clause biqqaštî ʾēt še- ʾāhăbâ napš-î sought/1fs/PERF ACC-REL-love/2fs/PERF soul-1s I looked for the one my soul loved (Song 3:1) c.  Prepositional object REL-clause b-ā- ʿăṭārâ še- ʿiṭṭĕrâ l-ô ʾimm-ô at-the-crown REL-crown/3fs/PERF to-3ms mother-3ms at the crown that his mother made for him (Song 3:11) d.  Manner REL-clause šek-kākâ hišbaʿtānû REL-thus swear/2fs/PERF-1p that you have sworn us like this (Song 5:9) e.  Predicate, possessive karm-î šel-l-î lĕ-pān-āy vineyard-1s REL-to-1s to-faces-1s my vineyard is in front of me (Song 8:12) The second REL-clause pattern is specific to the participial/imperfective tense-aspect, and only for subjects. Examples from EBH are: (14)  ​ a.  Definite û-bĕ-kōl ḥayyâ hā-rōmeśet ʿal-hā-ʾāreṣ and-over-all animal the-crawl/fs/PAR on the-earth and over all the animals crawling upon the earth (Gen 1:28) b.  Indefinite: dešeʾ ʿēṣeb mazrîaʿ zeraʿ lĕ-mîn-ēhû turf grass seed/ms seed to-kind-3ms turf of grass seeding (its) seeds of all its kinds (Gen 1:12) The frequency distribution of ʾăšer vs. ša- REL-subordinators across the BH dialect continuum is given in table 4 (following Givón 1991). The general trend that lines up the books in the same order along the continuum persists.

52

T. Givón Table 4.  The Contraction of ăšer to ša- in REL Clauses ăšer Genesis Esther Lamentations Qoheleth Song of Songs MH (m. Berakot)

N 93 100 9 69 — 1

% 100% 100% 75% 52% — 1%

šaN — — 3 62 25 122

% — — 25% 48% 100% 99%

Total N % 93 100 100 100 12 100 131 100 25 100 123 100

3.3.  Verbal Complements In this section, I look only at indirect-quotation complements of cognitionperception-utterance verbs such as ‘know’, ‘see’, ‘hear’, or ‘say’/‘tell’. Two main finite patterns predominate in Biblical Hebrew. The first pattern—with either the preposition kî (‘as’, ‘like’) or the de-verbal wĕ-hinnēh (‘and-be’, ‘and-lo’) as subordinator—predominates in the EBH texts. Thus consider: the kî- subordinator (15)  With ​ a.  way-yarʾ ʾĕlōhîm kî ṭôb and-3ms/see/PRET God SUB good ‘and God saw that it was good’ (Gen 1:10) b.  wa-yĕhî ki-šmōaʿ ʾĕlîšāʿ ʾîš hā-ĕlōhîm and-3s/be/PRET SUB-hear/INF Elisha man/of the-God ‘and so it was that when Elisha the man of God heard kî qāraʿ melek yisrāʾēl ʾet-bĕgād-āyw SUB tear/3ms/PERF king/of Israel ACC-clothes-3ms that the king of Israel had torn his clothes’ (2 Kgs 5:8) With the wĕ-hinnēh subordinator c.  way-yarʾ wĕ-hinnēh hārbû pĕnê and-3ms/see/PRET and-be dry/3mp/PERF face/of  hā-ʾădāmâ  the-earth ‘and he saw that, lo, the face of the earth had dried up’ (Gen 8:13) The second pattern, with the REL-subordinator ʾăšer/ša-, predominates in LBH and is phased in gradually along the BH continuum. Some examples of its use from the latter portions of the continuum are: (16)  ​ a.  ṭôb ʾăšer lōʾ tiddōr good SUB NEG 2ms/swear/IRR it is better that you don’t swear an oath (Qoh 5:4)

Biblical Hebrew as a Diachronic Continuum

53

Table 5.  Distribution of V-Complement Subordinator Patterns

Genesis Esther Lamentations Qoheleth Song of Songs MH (m. Berakot)

kî/wĕhinnēh N % 24 100% 7 50% 7 87% 12 34% — — n/a n/a

ʾăšer/šaN % — — 7 50% 1 13% 23 66% 24 100% n/a n/a

Total N % 24 100 14 100 8 100 35 100 24 100 n/a n/a

b.  wĕ-rāʾîtî ʾānî šey-yēš yitrôn l-a-hokmâ and-see/1s/PERF I SUB-is advantage to-the-wisdom ‘and I saw that there was advantage in wisdom’ (Qoh 2:13) c.  mah-taggîdû l-ô še-hôlat ʾahăbâ ʾānî what 2mp/tell/IRR to-3ms SUB-sick/fs/of love 1s ‘what should you tell him? That I am sick with love’ (Song 5:8) wĕ-yādaʿtî gam-ʾānî šem-miqreh ʾehād and-know/1s/PERF too 1s SUB-fate one  yiqreh       ʾet-kull-ām   3ms/befall/IRR ACC-all-3mp ‘and I knew too that one fate will befall all of them’ (Qoh 2:14) In table 5, I present the relative distribution of the two main V-complement subordination patterns across the BH diachronic continuum (following Givón 1991). By and large, with Lamentation again the “odd man out,” the distributional data here conform to the general trend of our putative diachronic continuum. Table 6 shows the distribution of the subordinator ʾăšer vs. the contracted ša- in V-complement clauses across the BH diachronic continuum. This affords us a comparison with the REL-clause data given in table 4. Largely conforming to the frequencies shown in table 4, the data again tag Qoheleth as the midpoint of the continuum with 52% completion; and Song of Songs is again at the 100% completion point, with only ša-, just as in Mishnaic Hebrew. 9 3.4.  Adverbial clauses In this section we inspect the distribution of two major patterns of ADVclause formation across the BH diachronic continuum. Both are already attested in EBH. The first is a nominalized pattern whereby a preposition such 9.  The only occurrence of the unreduced ʾăšer in the Song of Songs is the first clause of the text: šîr haš-šîrîm ʾăšer li-šlōmōh—no doubt an instance of literary hypercorrection.

54

T. Givón Table 6.  The Contraction of ʾăšer to ša- in V-Complements

Genesis Esther Lamentations Qoheleth Song of Songs MH (m. Berakot)

ʾăšer N % — — 7 100% — — 11 48% — — — —

šaN — — — 12 25 24

% — — — 52% 100% 100%

Total N % — — 7 100 — — 23 100 25 100 24 100

as b- (‘at’), k- (‘like’) or l- (‘to’) is attached to the nominal/infinitive form of the verb, or to some locative/temporal noun that often partakes in this pattern. Either the subject or object argument in this pattern is marked as genitive. This pattern predominates in EBH texts, and there are reasons to believe that it is the older pattern, not only for ADV-clauses but also REL-clauses (Givón 1991; 2009: chap. 4). Some examples of this nominalized pattern from EBH texts are: Yhwh ʾĕlōhîm ʾereṣ wĕ-šāmāyim (17)  ​ a.  bĕ-yôm ʿăśôt at-day/of make/NOM/of Yhwh God earth and-sky ‘on the day when God made earth and sky’ (lit.: ‘upon the day of God’s making earth and sky’, Gen 2:4) bĕ-hibbārʾ-ām at-PASS/create/NOM-3mp ‘when they were created’ (lit.: ‘upon their creation’, Gen 2:4) b.  ʿad šûb-ĕkā ʾel hā-ʾădāmâ time return/NOM-2ms to the-earth ‘till you return to the ground’ (lit.: ‘the time of your return to the ground’, Gen 3:19) c.  ʾaḥărê hôlîd-ô ʾet mahălalʾēl following/of sire/NOM-3ms ACC-Mahalel after he sired Mahalel (lit.: ‘after his siring Mahalel’, Gen 5:13) d.  ki-šmōaʿ kōl-malkê hā-ʾĕmōrî like-hear/NOM/ of all/of kings/ of the-Amorite ‘when the Amorite kings heard’ (lit.: ‘upon the Amorite kings’ hearing’, Josh 5:1) e.  ki-lḥōk haš-šô ʾēt yereq haś-śādeh like-chew/NOM/ of the-ox ACC-grass/of the-field ‘the way the ox grazes the grass of the field’ (lit.: ‘like the ox’s grazing the grass of the field’, Num 22:4)

Biblical Hebrew as a Diachronic Continuum

55

The second ADV-clause pattern is finite, marked with the REL-subordinator ʾăšer/ša-, and often combining with a locative/temporal noun preceding it. This pattern is at the beginning of launching itself in EBH, and there are good grounds to believe that it is being extended from the REL-clause pattern. In LBH, in contrast, this is the predominant pattern. Some examples from EBH are: (18)  ​ a.  Temporal ka-ʾăšer hiqrîb lā-bôʾ micrāymâ like-REL near/3ms/PERF to-come Egypt-ALL ‘when he neared coming to Egypt’ (Gen 12:11) b.  Purpose lĕ-maʿan ʾăšer yĕṣawweh to-answer/NOM REL 3ms/order/IRR ‘so that he order’ (lit: ‘for the purpose that he order’, Gen 18:19) c.  Purpose wĕ-śamtî ʾet-zarʿ-ăkā ka-ʿăPar hā-ʾāreṣ and-put/1s/PERF ACC-seed-2ms like-dust/of the earth ‘and I shall scatter your seed like the dust of the earth’ ʾăšer ʾim yûkal ʾîš li-mnôt REL if 3ms/can/IRR man to-count ‘so that if a man could count’ (Gen 13:16) Examples of this finite ADV-clause pattern from LBH are: ʾet hā-ʾahăbâ ʿad šet-teḥpāṣ (19)  ​ a.  wĕ-ʾim tĕʿôrĕrû and-if 3mp/wake/IRR ACC-the-love time REL-3fs/desire/IRR ‘and if you will wake up love till it will desire’ (Song 2:7) b.  ʿad šey-yāpūaḥ hay-yôm time REL-3ms/blow.away/IRR the-day ‘till the day expires’ (Song 2:17) c.  ki-mʿat še-ʿābartî mē-hem like-little REL-pass/1s/PERF from-them ‘(when) I had almost passed them’ (Song 3:4) d.  mah naʿăśeh la-ʾăḥōt-ēnû b-ay-yôm what 1p/do / IRR to-sister-1p on-the-day   šey-yĕdubbar b-āh   REL-3ms/speak/PASS/IRR at-3fs ‘what shall we do with our sister on the day she is spoken about’ (Song 8:8) A few examples of the nominalized AV-clause pattern survive in LBH, as in, for example:

56

T. Givón

Table 7.  The Distribution of the Two ADV-Clause Patterns Nominalized

Finite

N

%

Genesis

41

95%

2

Esther

13

81%

3

2

50%

2

Song of Songs

2

18%

MH (m. Berakot)

7

12%

Lamentations

N

Total %

N

%

%

43

100

19%

16

100

50%

4

100

9

82%

11

100

47

88%

54

100

Qoheleth

ḥătunnāt-ô û-bĕ-yôm śimḥat (20)  bĕ-yôm ​ in-day/of wed/NOM-3ms and-in-day/of rejoice/NOM/of  lib-bô  heart-3ms ‘on the day of his wedding and on the day of his heart’s rejoicing’ (Song 3:11) The distribution of the two patterns for temporal ADV-clauses across the BH diachronic continuum is provided in table 7 (Givón 1991). Again, the general pattern follows the same trends that we saw above: the same BH books are lined up in the same relative order. 4. Discussion The diachronic grammatical changes discussed above and their frequency distribution across the BH continuum are largely independent of each other. Each one of these changes occurred in response to its own distinct motivating pressures and according to its own distinct dynamic mechanisms, in spite of some obvious relatedness in areas of grammar. The changes are: a.  Change of word order b.  Complex changes in the functional distribution of tense-aspects c.  The phonological attrition of a lexeme into a bound morpheme d.  The transfer of morphosyntactic patterns from REL-clauses to V-complements e.  The transfer of morphosyntactic patterns from REL-clauses to ADV-clauses Although the motivation and mechanisms driving these changes were not emphasized here, they are well understood and amply argued elsewhere (Givón

Biblical Hebrew as a Diachronic Continuum

57

1977; 1991). What is more, these changes conform to well-known universals of grammaticalization. To give just one example of a fairly transparent mechanism, consider the hybrid construction whereby a verb takes both a nominal object and a verbal complement. Such a construction constitutes an analogical bridge between REL-clauses and V-complements and is found in both EBH and LBH. With relatively little reanalysis, the V-complement clause following an NP can be interpreted as a REL-clause, given that REL-clauses follow a head noun, while V-complement follow a verb. Thus: ʾĕlōhîm ʾet-kōl-ʾăšer ʿāśâ (21)  ​ a.  way-yarʾ and-3ms/see/PRET God ACC-all REL make/3ms/PERF ‘. . . and God saw all that he had done’  wĕ-hinnēh ṭôb  mĕʾōd   and-be   good very   ‘and lo it was very good’ (Gen 1:31) b.  way-yarʾ ʾĕlōhîm ʾet-hā-ʾôr kî ṭôb and-3ms/see/PRET God ACC-the-light SUB good ‘and God saw that the light was good’ (lit.: ‘and God saw the light that it was good’, Gen 1:4) c.  šāmaʿnû ʾēt ʾăšer hôbîš Yhwh ʾet-mê hear/1p/PERF ACC-REL dry/3ms/PERF Yhwh ACC-water/of  yam  sûp   sea/of reed ‘we heard that God had dried up the waters of the Red Sea’ (Josh 2:10) d.  ʾal tirʾû-nîî še-ʾănî šĕ-ḥarḥōret NEG 3mp/see/IRR-1s REL-1s swarthy ‘don’t see that I am swarthy’ (lit.: ‘don’t see me that I am swarthy’, Song 1:6) e.  yādaʿtî šeg-gam-zeh hûʾ raʿyôn rûaḥ know/1s/PERF REL-too this 3ms folly/of spirit ‘I knew that it too was folly’ (Qoh 1:17) Exx. (21a, b) are the old hybrid constructions in EBH. Exx. (21c, d) are more advanced intermediates with the REL-subordinator ʾăšer invading the paradigm. And (21e) is the final product in LBH. Like all grammatical change, and like the everyday communication that gives rise to it, the changes studied here occur subconsciously and unobtrusively. This is what makes the pattern of distribution of morphosyntactic facts in the BH texts all the more remarkable. With relatively few perturbations, notably the book of Lamentations, these grammatical changes line up the very same books in the very same directional—and thus, I am suggesting, temporal—order.

58

T. Givón

Whether the diachronic interpretation of this order is supported by nonlinguistic evidence is a question I am not qualified to answer. But if the distributional patterns shown above are the result of fortuitous accidents, the accidents were surely governed by a perverse intelligence that somehow conspired to make Biblical Hebrew, with its incredibly lush beauty, look like a natural language, with natural-seeming distributions, variability, and change. For my money, still, if it looks like a duck, waddles like a duck, and quacks like a duck, it would be a bit shocking if it turned out not to be a duck. References

Bopp, F. 1820 Analytic Comparison of the Sanskrit, Greek, Latin and Teutonic Languages. Reprinted in Amsterdam Studies in the Theory and History of Linguistic Sciences, vol. 3, ed. E. F. K. Koerner. Amsterdam: Benjamins, 1974. Bybee, J.; Pagliuca, W.; and Perkins, R. 1994 The Evolution of Grammar: Tense, Aspect and Modality in Languages of the World. Chicago: University of Chicago Press. Givón, T. 1971 Historical Syntax and Synchronic Morphology: An Archaeologist’s Field Trip. Chicago Linguistics Society 7: 394–415. 1975 Serial Verbs and Syntactic Change: Niger Congo. Pp. 47–112 in Word Order and Word Order Change, ed. C. Li. Austin: University of Texas Press. 1976 On the VS Word-Order in Israeli Hebrew: Pragmatics and Typological Change. Pp.  153–81 in Studies in Modern Hebrew Syntax and Semantics, ed. P. Cole. Amsterdam: North-Holland. 1977 The Drift from VSO to SVO in Biblical Hebrew: The Pragmatics of TenseAspect. Pp. 181–254 in Mechanisms of Syntactic Change, ed. C. Li. Austin: University of Texas Press. 1979 On Understanding Grammar. New York: Academic Press. 1983 Topic Continuity in Discourse. Typological Studies in Language 3. Amsterdam: Benjamins. 1991 Serial Verbs and the Mental Reality of “Event.” Pp. 81–127 in vol. 2 of Approaches to Grammaticalization, ed. E. Traugott and B. Heine. 2 vols. Typological Studies in Language 19. Amsterdam: Benjamins. 1994 Voice and Inversion. Typological Studies in Language 23. Amsterdam: Benjamins. 2000 Internal Reconstruction: As Method, as Theory. In Reconstructing Grammar: Comparative Reconstruction and Grammaticalization, ed. S. Gildea. Typological Studies in Language 43. Amsterdam: Benjamins. 2001 Syntax: An Introduction. 2 vols. Amsterdam: Benjamins 2009 The Genesis of Syntactic Complexity: Diachrony, Ontogeny, Neuro-Cognition, Evolution. Amsterdam: Benjamins.

Biblical Hebrew as a Diachronic Continuum

59

Heine, B.; Claudi, U.; and Hünnemeyer, F. 1991 Grammaticalization: A Conceptual Framework. Chicago: University of Chicago Press. Heine, B., and Kuteva, R. 2007 The Genesis of Grammar. Oxford: Oxford University Press. Hopper, P., ed. 1982 Tense and Aspect: Between Semantics and Pragmatics. Typological Studies in Language 1. Amsterdam: Benjamins. Hopper, P., and Traugott, E. 1993 Grammaticalization. Cambridge: Cambridge University Press. 2003 Grammaticalization. Rev. ed. Cambridge: Cambridge University Press. Paul, H. 1890 Principles of the History of Language. London: Swan, Sonnenschein. Traugott, E., and Heine, B., eds. 1991 Approaches to Grammaticalization. 2 vols. Typological Studies in Language 19. Amsterdam: Benjamins. Vennemann, T. 1973 Explanations in Linguistics. Syntax and Semantics 2, ed. J.  Kimball. New York: Academic Press.

Diachrony in Biblical Hebrew and a Theory of Language Change and Diffusion Jacobus A. Naudé 1. Introduction The goal of this essay is to continue developing a theory of language change and diffusion (Naudé 2000a; 2000b; 2003; 2010) in the light of recent developments in contemporary linguistics. The study of diachrony and language change in Biblical Hebrew cannot be separated from new developments in contemporary diachronic linguistics (Fischer 2007), theories of language evolution (Mufwene 2008), language complexity (Larsen-Freeman and Cameron 2008; Ellis and Larsen-Freeman 2009; Givón 2009; Sampson, Gil, and Trudgill 2009), or studies on the vernacular use of Biblical Hebrew (inter alia, Sanders 2009; Schniedewind 2005). The focus in this essay is on the description and explanation of linguistic variation in Biblical Hebrew. The problems posed by Biblical Hebrew are similar to those posed by modern languages (compare, for example, Chambers et al. 2002). Recently published ideas on the concepts of diachrony and language change as well as the metalanguage for describing them have had a profound effect on this research. These new ideas contribute insights that promote viewing the diverse aspects of language change holistically (that is, understanding the causal dependencies and emergent processes among the elements that constitute the whole system) rather than viewing them partially and in isolation (see Clancey 2009: 11–34) as some scholars of Biblical Hebrew are doing now. This last problematic methodology characterizes two recent approaches (Hurvitz 2006 and Young et al. 2008) that are vying for legitimacy in recent discussions of diachrony in Biblical Hebrew. Both approaches are problematic because they are not psychologically true (Halle et al. 1978: xi–xv) but are counterintuitive. The recent developments on diachrony and language change are discussed first (§2). Second, best practices concerning the objectives, methods, and nature of evidence in diachronic study are discussed in order to move beyond the formal versus functional divide in historical linguistics (§3.1). Third, I show how a theory of language change and diffusion can be adjusted to recognize 61

62

Jacobus A. Naudé

the inevitable and irrefutable fact of language change and diffusion over time through recognition of four essential components (§3.2). This section discusses what should be used as the basis for comparison in identifying language change—the source and nature of the linguistic elements to be compared—as well as the role of internal and external factors in language change, especially the role played by the development of a written standard (see Sanders 2009). The final section (§4) illustrates the new approach to language change described in this study with examples from the Hebrew corpus. The core assumptions of this article are, first, the fact that a language inevitably changes and diffuses over time; and second, that a language inherently displays variation. The emphasis throughout, however, is on the theoretical issues because theory is the starting point for furthering the discussion on diachrony and language change productively. 2.  New developments on Diachrony and Language Change 2.1.  Language as a Complex Adaptive System 2.1.1.  Larsen-Freeman and Cameron (2008) Larsen-Freeman and Cameron (2008: 43–78) address the problem of change within complex systems theory. A system has elements that interact to form a connected whole. Complex systems are systems with many different types of processes, elements, and agents—usually in large numbers—that connect and interact in multiple ways. Furthermore, complex systems are systems that are heterogeneous concerning elements and agents, dynamic (that is, everything changes all the time), nonlinear, adaptive (that is, change in one area of the system leads to change in the system as a whole or to only part of it), and open. There is interconnectedness of both system and context. An open system cannot be independent of its context since there is a flow of energy or matter between system and environment; the context is part of the system and its complexity. Systems are continually adapting to contextual changes and may change internally as a result of adapting to external change. Adaptation, self-organization, and emergence are all types of change in complex systems. Internal self-organizing changes alter the structure of a system, while responding to energy and matter coming from outside leads to adaptive change that maintains order or stability. Emergence is the appearance in a complex system of a new state at a level of organization higher than the previous one. The emergent behavior or phenomenon has some recognizable “wholeness.” There is also a sense in which emergence produces simplicity from complexity. A painting, as a whole, is in some sense simpler than the multiple interactions of color and form from which it emerges. Higher level, emergent phenomena have an identity as wholes and can be labeled as such: for example, Monet’s Water Lilies (Larsen-Freeman and Cameron 2008: 36, 43, 58–60).

Diachrony and a Theory of Language Change and Diffusion

63

The phenomenon of change—interpreted as constant, dynamic, or on­ going—is a phenomenon related in the first instance, not to a physical understanding of things, but to a biological one: “an organism’s ongoing activity continuously changes its neural states, just as growth changes the physical dimensions of the body” (Larsen-Freeman and Cameron 2008: 17; see also pp. 29, 32, 72). In a complex systems approach, the emphasis is therefore on dynamics and thus requires “us to look for change and for processes that lead to change, rather than for static, unchanging entities” (Larsen-Freeman and Cameron 2008: 16; also p. 26). The flux or change that is the focus of the approach is interpreted in an organically dynamic way. Systems maintain stability even though they are perpetually dynamic—the stability gives the appearance of no change at all. This is the adiabatic principle (Larsen-Freeman and Cameron 2008: 88). A second important and related tenet concerns the linearity of the processes of change. Since all the various components of a complex system such as language (including what was previously sometimes inappropriately sidelined as “context”) are in continuous interaction, there is “reciprocal causality” (Larsen-Freeman and Cameron 2008: 7, 60), in which there is upward emergence of the patterns from individuals interacting that is nonetheless downwardly constrained due to both the historic trajectory of the system and its present-day sociocultural norms. (In this respect, complex systems theory is opposed to classical Darwinism and behaviorism, which viewed the environment as external to and independent of the organism.) Complex systems use spatial and topographical images to describe how a system changes over time. A complex dynamic system is visualized as wandering across a landscape, up hills and down through valleys, occasionally coming to a halt when a valley is too deep to get out of easily, but resuming its journey if it gathers enough energy to escape. The processes of change can therefore be described as a “movement in a trajectory across a ‘state space’ or ‘phase space’” (Larsen-Freeman and Cameron 2008: 20, 43, 45–49), which is a collection of all possible states of a system with dimensions relating to change over time. The change process, if drawn toward a sufficiently powerful “attractor,” which is a region of a system’s state space into which the system tends to move (Larsen-Freeman and Cameron 2008: 20, 43–78), can come to a provisional stability around, for example, the strong attractor of the notion of a standard language. Complexity theory has shown that three types of attractors can occur in the state space of complex dynamic systems: fixed-point attractors, cyclic attractors, and chaotic attractors (Larsen-Freeman and Cameron 2008: 56–57). Fixed-point attractors are the simplest type, representing a system moving into a stable, preferred state and remaining there. In a cyclic or closed-loop attractor, the system moves periodically between several different attractor states. A chaotic or strange attractor is a region of state space in which the system’s

64

Jacobus A. Naudé

behavior becomes quite wild and unstable, as even the smallest perturbation causes it to move from one state to another. Complex systems can change smoothly and continuously for periods of time but may then go through more-dramatic types of change, when they alter their nature radically, sometimes entering a period of turbulence, or chaos, when the system keeps on changing dramatically (Larsen-Freeman and Cameron 2008: 43–44). A system undergoes a phase shift or bifurcation when its behavior changes suddenly to a new and radically different mode (LarsenFreeman and Cameron 2008: 128). The states of the system before and after a phase shift are very different. From a complex systems perspective, language develops in a process of coadaptation (Larsen-Freeman and Cameron 2008: 67–68, 84, 126–28). Coadaptation is the interaction of two or more complex systems, each changing in response to the other. Since humans, as language-using agents, “soft-assemble” language from the resources at their disposal, the expectations that derive from previous experiences of cocreating and aligning oneself with other languageusing agents through discourse constitute an important resource. Adapting human resources sometimes means appropriating extant language-use patterns; at other times, it means innovating. When two individuals soft assemble using their language resources on a given occasion and then interact and adapt to each other, the state space of both their language resources changes as a result of coadaptation. On a longer time scale and within a speech community, these local interactions are what transform the state space of the language in an ongoing way. The passing on of change across a system is a sort of feedback, and the process in which a system adjusts itself in response to changes in its elements is a form of self-organization. As with other complex systems, language-using patterns are heterochronous; a language event on some local time scale may simultaneously be part of language change on longer time scales. Historical change on a long time scale arises from small local changes made possible by variability around current stabilities or the recurrent regularities. So, too, learning a language is seen as language development rather than as acquisition. In other words, language is a process of dynamic adaptation (Larsen-Freeman and Cameron 2008: 157) rather than something that, once learned, is “possessed” for all time. From a complexity point of view, language can never be in an entirely stable state, so it cannot be “acquired” once and for all. In this respect, complexity theory is opposed to a generative approach to language acquisition and change. A complexity perspective also provides a particular viewpoint on the composition of a written text. As the text is composed, it moves through different versions, changing and adapting in the process of composition. This dynamic process operates not just at the whole text level but with variability at all levels: the writer selects the best word, tries out several ways of writing the same

Diachrony and a Theory of Language Change and Diffusion

65

idea, adapts syntactic formulations of clauses and sentences, moves around paragraph content and sections. Eventually the text stabilizes into a form that stops changing. Various pressures ensure that most texts are finalized, even though potentially endless further variation would have been possible. Written texts can be viewed as co-constructed but asynchronically collaborative compositions in which writers imaginatively engage with, and thus interact with, prospective readers (Larsen-Freeman and Cameron 2008: 188). To summarize, Larsen-Freeman and Cameron (2008) have made the case for conceiving of language as a complex, adaptive, dynamic system that emerges and self-organizes from frequently occurring patterns of language use at different times and across a range of levels (from the individual to interacting pairs to entire speech communities), rather than as a fixed, autonomous, closed, and atemporal system. Accordingly, the evolution of language, language change, language diversity, language development, language learning, and language use are emergent from the dynamic processes of change that operate in all languages at all times. 2.1.2.  Ellis and Larsen-Freeman (2009) The “Five Graces Group” 1 demonstrates in a position paper in Ellis and Larsen-Freeman (2009: 1–26) that the processes of language acquisition, usage, and change are not independent of one another but are facets of the same complex adaptive system. The system consists of multiple agents (the speakers in the speech community) interacting with one another. The system is adaptive, that is, speakers’ behavior is based on their past interactions, while current and past interactions together feed forward into future behavior. A speaker’s behavior is thus the consequence of competing factors ranging from perceptual constraints to social motivations. Cognition, consciousness, experience, embodiment, brain, self, human interaction, society, culture, and history are all intertwined in rich, complex, and dynamic ways in language. The structures of language emerge from interrelated patterns of experience, social interactions, and cognitive mechanisms. They emerge as synchronic patterns of linguistic organization at numerous levels (phonology, lexis, syntax, semantics, pragmatics, discourse, genre, etc.), as dynamic patterns of usage, as diachronic patterns of language change (such as linguistic cycles of grammaticalization, pidgeonization, creolization, etc.), as ontogenic developmental patterns in child language acquisition, as global geo-political patterns of language growth and decline, dominance and loss, etc. (Ellis and Larsen-Freeman 2009: 1–3, 19). 1.  The “Five Graces Group” is a working group consisting of the following 10 scholars, sponsored by the Santa Fe Institute: Clay Beckner, Richard Blythe, Joan Bybee, Morten H. Christiansen, William Croft, Nick C. Ellis, John Holland, Jinyun Ke, Diane Larsen-Freeman, and Tom Schoenemann. The group is named after their special accommodations in Santa Fe, NM, during a conference there in 2007.

66

Jacobus A. Naudé

Language exists both in individuals (as idiolect) and in the community of users (as a communal language; Ellis and Larsen-Freeman 2009: 14–15). Language is emergent at these two distinctive, but interdependent levels: An idiolect is emergent from an individual’s language use through social interaction with other individuals in the communal language, whereas a communal language is emergent as a result of the interaction of the idiolects. Both communal language and idiolects are in constant change and reorganization. Languages are in constant flux, and language change is ubiquitous. At the individual level, every instance of language use changes an idiolect’s internal organization. The mechanisms of language change—such as production, economy, and frequency effects that result in phonetic reduction—may not be at work in every individual in the same way or at the same time. Moreover, functional or social mechanisms that lead to innovation in the early stages of language change need not be at work in later stages, because an individual may acquire innovations later, when they are established and broadly diffused in the communal language. For instance, language may change in the tug-of-war of conflicting interests between speakers and listeners. Speakers prefer productive economy, which encourages brevity and phonological reduction, whereas listeners want perceptual salience, explicitness, and clarity, which require elaboration (Ellis and Larsen-Freeman 2009: 16). Therefore, language change is a phenomenon observable at the communal level, but the actual process of language change is complicated and is woven from a myriad of factors. Language change is a cultural evolutionary process and takes place at two linked levels: replication and selection (Ellis and Larsen-Freeman 2009: 9). Replicators are units such as genes or words that are replicated with some chance for innovation and variation. Selection is a process by which individuals interacting with their environment cause replication of the replicators to be differential; that is, some replicators are replicated more than others, which in the extreme case leads to fixation of the former and extinction of the latter. Due in part to the indeterminacy of communication, the replication process produces variation. In complex systems, small quantitative differences in certain parameters often lead to phase transitions (that is, qualitative differences; Ellis and LarsenFreeman 2009: 16–17). Additionally, in a dynamic system, even when there is no parametric change, at a certain point behavior can change dramatically, going through a phase transition. Developmental lexical spurts often lead to rapid grammatical development. The S-curve shape of dynamics in language change is also a kind of phase transition (see also Larsen-Freeman and Cameron 2008: 85–88). Grammaticalization as a result of language use may be another consequence of such phase transitions, in which lexical items become grammatical items. For instance, in the history of English it may be observed that the main verb cunnan ‘to know’ underwent incremental changes that re-

Diachrony and a Theory of Language Change and Diffusion

67

sulted in a qualitative difference, so that can now functions as an auxiliary expressing possibility. Such changes accumulate gradually as the result of individual speaker utterances. Yet at a certain point in the shift, processes of change may seem to escalate: as the meaning of can became more abstract and general, it increased further in frequency, which further drove the development of grammatical status. 2.2.  Language Change as Language Evolution 2.2.1.  Mufwene (2008) Mufwene (2008: 16) applies the term “language evolution” to specific languages in a way similar to language change. However, for Mufwene (2008: 16) the term “evolution” covers more than the traditional term “change.” In addition to traditional concerns with structural and pragmatic changes, it also covers language speciation and language birth and death—processes to which the term “change” has not previously been applied in linguistics. For example, in Mufwene’s view, the emergence of creoles as new vernaculars (a kind of language birth) involved a concomitant loss of their creators’ heritage languages (a kind of language death; Mufwene 2008: 6). However, language contact between individuals rather than between populations plays an important role in Mufwene’s approach (Mufwene 2008: 17). Parallels to biological evolution are used by Mufwene (2008) in developing a general model for language evolution. The central idea in Mufwene’s approach is that languages are like biological species, and individual dialects or idiolects are analogous to individual organisms. To be more specific, languages are like parasitic species (the host is the biological human species in this case), in which individuals can be “infected” by multiple idiolects. This perspective naturally leads to an emphasis on variation within populations, with change happening through competition and selection at the level of features and idiolects. As in biology, “competition” refers to the unequal ways that the variants and coexistant language varieties are weighted by their speakers (the immediate external ecology through which all other ecological pressures are filtered), while “selection” describes the resolution of the competition, determining which particular variant or language variety is favored over which particular competitors (see Mufwene 2008: 115–32). Every idiolect is unique, and none replicates the idiolect of any other current speaker, because native speakers acquire their vernaculars naturalistically, through interactions with current speakers (Mufwene 2008: 3, 11–28). The acquisition of an idiolect proceeds piecemeal as an individual attempts to communicate, which leaves a lot of room for inaccurate inferences and therefore deviations in language that can evolve into divergences. Attempts to communicate force speakers to rely on partial analogies and to use both well-formed and ill-formed utterances, thereby improving their productions

68

Jacobus A. Naudé

and their interpretations of other speakers’ utterances in the process. Speakers often accommodate each other, thus keeping their idiolects and their languages in constant states of flux. As a result, languages are constantly being (re)shaped by their speakers, although not in noticeable ways. Linguistic structures may also not be as neatly integrated as traditionally claimed. Although much of the reshaping goes unnoticed, amounting to producing variants that are already current within the relevant community of speakers or variants that are rejected as abnormal, a few do contribute to structural changes. Much of this unpredictable behavior has to do with how the “invisible” hand works within a naturally heterogeneous population of speakers, producing new patterns of convergence among idiolects, driving some variants out, sometimes producing new variants, and generating new norms. As with biological species, the external ecology of a language works on variation within the language to drive its evolution (Mufwene 2008: 3). Languages are heterogeneous beyond the nonmonolithic architecture of idiolects, a property that follows from the fact that idiolects vary among themselves. A communal language exists only as a social construct, suggested by the ability of speakers in a particular setting to communicate successfully when they use similar sequences of spoken or signed gestures. Lumping idiolects into the same language is a matter of likeness, not of sameness or identity; idiolects are alike by “family resemblance.” Idiolects are thus like biological populations, in which each organism preserves its genotypic and phenotypic individuality, despite the many features it shares with its conspecifics. Languages evolve in nonuniform ways, with some idiolects, sociolects (social varieties of language), or dialects (regional varieties of language) being more engaged in some changes than others. Idiolects are complex adaptive systems just as are the communal languages that extrapolate from the idiolects. Because they exist by virtue of being spoken by individuals, communal languages are inherently variable (Mufwene 2008: 131–32). 2.2.2.  Hauser, Chomsky, and Fitch (2002) versus Everett (2005) A longstanding linguistic axiom is that all languages and language varieties are similar in complexity, and they have been so at all times in the past. This issue and the issue of language evolution surfaces in a joint paper by Hauser, Chomsky, and Fitch (2002), where they assert that recursivity is the critical feature separating human language from prehuman communication. Recursivity is the embedding of a clause inside another clause; that is, a node at a lower hierarchic level is dominated by a node of the same type at a higher level (for example, Women [who like smart men] are smart). Chomsky’s claim concerning the universal necessity of recursivity has been challenged by the evidence of at least one language (Piraha, an Amazonian isolate) that has no embedded relative clauses or verb complements. The communicative function of these constructions is performed when speakers use paratactic functional equivalents

Diachrony and a Theory of Language Change and Diffusion

69

(Everett 2005). Everett’s claim that Piraha has no recursivity hinges on the viewpoint that embedded clauses diachronically arise out of paratactic precursors (see Givón 2009: 41). 2.2.3.  Sampson, Gil, and Trudgill (2009) The book by Sampson et al. (2009) is the outcome of a workshop held at the Max Planck Institute for Evolutionary Anthropology, Leipzig, in April 2007. It challenges the widely held assumption that human languages are both similar and constant in their degree of complexity (see also §2.2.2) and poses language complexity as an evolving variable rather than a constant across languages (Sampson et al. 2009: 269). It also challenges an inevitable outcome of one of the central axioms of generative linguistic theory—namely, that the mental architecture of language is fixed and is therefore identical in all languages (Sampson et al. 2009: 270–71). A third challenge concerns the relevant linguistic data that should be investigated by linguists. Grammar (morphology and syntax), especially clause structure and clause subordination, is central to the assessment of language complexity (Sampson et al. 2009: 269–70). When language complexity was an issue in linguistics in the early twentieth century, syntax was not central to the discussion. 2.2.4.  Lee, Mikesell, Joaquin, Mates, and Schumann (2009) Lee et al. (2009) explore the evolution of language from the theoretical view that language could have emerged without a biologically instantiated universal grammar. A hominid group with a lexicon of about 600 words could combine these items to make larger meanings. Combinations that are successfully produced, comprehended, and learned become part of the “grammar” of the language. Any combination that is incompatible with human mental capacities is abandoned. Contrary to Pinker (1994), language structure emerged through interaction constrained by human psychology and physiology. In other words, language is not part of biology but a cultural artifact that emerged out of interactions among hominids. As a result, languages are not genetically but, rather, culturally transmitted. Language acquisition is based on an interactional instinct that emotionally bonds the infant to its caregivers. This relationship provides children with a motivational and attentional mechanism that ensures their acquisition of language. In adult second-language acquisition, the interactional instinct is no longer operating, but in some individuals with sufficient aptitude and motivation, successful second-language acquisition can be achieved. 2.2.5.  Givón (2009) Givón (2009) also treats change as an integral part of the evolutionary rise of human communication. For him, the issue is whether or not syntactic

70

Jacobus A. Naudé

complexity depends on a chance mutation in recent human evolution that introduced an entirely novel cognitive ability called recursion (see §2.2.2). According to Givón (2009: 41) there are three grand developmental trends of human language that fashioned the way language and languages are now: evolution, the development of the language capacity of the human species, as well as pidiginization, the development of new languages; diachrony, the communal enterprise directly responsible for fashioning synchronic morphosyntax and cross-language diversity—that is, the historical development of particular languages; and ontogenesis, the individual endeavor directly responsible for acquiring the competent use of grammar—that is, the emergence of language in children. The evolutionary relevance of language diachrony and language ontogeny is the organism’s adaptive on-line behavior—invention, learning, and skill acquisition—which constitutes the common thread running through all three developmental trends. Of these three, diachrony is a uniquely human phenomenon, a cumulative historical growth through cultural transmission. What is more, diachrony has the most direct causal bearing on the shape of any particular language and thus on the diversity of all human languages. (I return to this in §3.) Six general principles exert control on both language diachrony and biological evolution (Givón 2009: 42): (1)  ​(i)  graduality of change, (i)  adaptive-selectional motivation, (ii)  functional change and ambiguity before structural change and specialization, (iii)  terminal addition of new structures to older ones, (iv)  local causation (but global consequences), and (v)  unidirectionality of change. In what follows, I show that the insights provided by complexity theory and language evolution enrich a theory of language change and diffusion. 3.  Toward a Psychologically Feasible Methodology 3.1.  Moving beyond the Formal-Functional Divide A methodology for analyzing language change and diffusion must be “psychologically feasible” (Chomsky 1975: 166–67; Halle et al. 1978). This means that the linguistic description and explanation of language change and diffusion must reflect the reality of the phenomena investigated. As the basis of my proposed methodology, I begin with Fischer (2007). She argues that in order to further the study of diachrony the following two issues must be addressed. First, scholars must determine the best practices concerning the objectives, methods, and nature of evidence for diachronic study. This pertains to what data are used as the basis for comparison (that is, the source

Diachrony and a Theory of Language Change and Diffusion

71

and nature of the linguistic elements to be compared) as well as the role of internal and external factors in language change (for example, the role played by the development of a written standard). Second, scholars must use insights from the two main theoretical frameworks that offer explanations for linguistic change thus far—namely, the general formalist approach (Lightfoot 1979; 1991; 1999; 2002) and the functionalist approach (Traugott and Heine 1991; Labov 1994; Hopper and Traugott 2003). The formalist approach is interested in how change in language output is caused by change in the formal grammar system (the top-down approach), while the functionalist approach concentrates on how linguistic utterances are used in communication and how this leads to grammatical change (the bottom-up approach). Both form and function are equally important, and both approaches offer opportunities for a better understanding of linguistic structure and language change. A combination of the two approaches provides a fuller understanding of the causes and mechanism of language change and, ultimately, of the system underlying language. In light of the new developments described in §2 and Fischer’s methodology, I consider the following components essential for a theory of language change and diffusion: (a) the individual dimension: language change (the creation of an idiolect); (b) the sociological dimension: diffusion of language change; (c) the time dimension: diachronic/chronological development; and (d) a consideration of the nature of written language. I describe each of these components in §3.2. Although it must be acknowledged that Young et al. (2008) contribute much to demolish the simple linear model of Ancient Biblical Hebrew, Early Biblical Hebrew, Late Biblical Hebrew, and Mishnaic Hebrew by suggesting a more complex model of language development (namely, that the data that are used to distinguish preexilic from postexilic Hebrew are no more than manifestations of synchronic styles available to biblical authors). Unfortunately their suggested model demonstrates a long list of shortcomings. It is based exclusively on (author) style (= endless styles that account for all its differences), denying chronology and language variation based on idiolects and dialectology. Their data are based mainly on the distribution of lexical items, phonology, morphology, and morphosyntax. No instances of syntactic change and pragmatico-semantic change form part of the core of their analyses, nor do these linguistic features play a role in their argumentation. The interpretation of data is based on frequency and does not explore the developmental history of grammatical constructions. Until now the debate has excluded almost all theoretical developments in mainstream historical and diachronic linguistics concerning the nature of language change and the diffusion of changes through a language system. The same criticisms pertain to the (Neogrammarian) claims of Hurvitz (2006) for the traditional division of Biblical Hebrew into chronological periods by assuming a single, homogenous spoken and written Hebrew at any

72

Jacobus A. Naudé

one time with exceptionless and irreversible changes into another variety. The following quotations substantiate my characterization of his work: In other words, the theory which seeks to interpret the relationship between Standard Biblical Hebrew and Late Biblical Hebrew in non-chronological terms (diglossia or otherwise) can hardly be sustained. It reflects an attempt by modem scholars to impose on Biblical Hebrew theoretical models which are incompatible with the evidence of the ancient texts themselves. (Hurvitz 2006: 208) Instead of the uniform style which characterizes the earlier literature, here the personality and talent (or lack of talent) of the writer is revealed. The greater part of the late authors aspired to perpetuate the old biblical tradition, but not all of them met with equal success . . . the basic argument of the present discussion remains linguistically unchanged. (Hurvitz 2006: 209–10, quoting Abba Ben­ david).

For Hurvitz, the distinctive linguistic nature of the late elements makes them valid chronological markers that may be used for dating purposes in approaching biblical texts the historical age of which is debated. 3.2.  Components of a Theory of Language Change and Diffusion 3.2.1.  The Individual Dimension The individual dimension of language change involves the creation of an idiolect, the innate grammar of an individual speaker. This is the empirical object of change (Mufwene 2008). Similarly, I (Naudé 2000; 2003: 209) view changes as revisions and differences in the features of lexical items in the mental lexicon of the individual. Language change on the individual level occurs when the grammar of an individual differs from that of the input source (for example, child versus parents). Language is always emergent and may continue to change in adult life on the basis of interaction with more general cognitive learning principles. The individual dimension is psycholinguistic in nature and relates to the ontogenesis described by Givón (2009). 3.2.2.  The Sociological Dimension The sociological dimension relates to the diffusion or implementation of language change (that is, the creation of an idiolect) as one differing grammar becomes dominant and gains acceptance by the local speech community and later by society at large. Note here that the diffused language variety competes with similar diffusions of other innovative grammars and, at the same time, is shaped by the standard language. A group or network of people has more or less the same language because they learn and influence one another in all sorts of behavior including language; this is the adaptive-selectional motivation of Givón (2009). Because diffusion is naturally a continuous process, language variation always exists in a speech community, let alone a society; and a given language (viewed as a multilayered, complex system) is always in transition. Language diffusion follows certain observable patterns (see Pintzuk 2003). It

Diachrony and a Theory of Language Change and Diffusion

73

occurs gradually and in the shape of an S-shaped curve, with the new option beginning slowly, then accelerating, and finally leveling off once the competition is resolved. Kroch (1989) has described what he calls the Constant Rate Effect: the frequency of use of the innovative variant may differ across contexts, but the rate of change for each context is the same. This dimension reflects the functional change and ambiguity of Givón (2009) (diffusion elements) and is sociolinguistic in nature. 3.2.3.  The Chronological Dimension On the basis of Givón (2009: 42–43), I contend that a chronological dimension can be added to the theory of language change and diffusion. The chronological dimension relates to diachronic/chronological development. The everyday adaptive communicative experimentation of adult speakers accumulates in the linguistic structures with which they constantly tamper. Cultural transmission is the main instrument of passing on adaptively driven language change to future generations. Some fossilized linguistic structures, especially morphological features, may persist for millennia and thus can assist linguistics in reconstructing the earlier stages of syntactic structure. The attrition (that is, erosion, elimination, simplification, and loss) of phonological and morphological structures may proceed to its ultimate end, eroding first morphology and then syntactic constructions. This leads, in due course, to the eventual renovation of morphosyntactic structure with “deep” structured changes, which I will name the “diachronic cycle” with respect to the structural change. This renovation happens against the language variation of competitive culturally transmitted (fossil) structures (which I call “stylistic fossils”), especially morphological structures that are largely the result of functional change/ difference. These stylistic fossils are in competition—at certain stages they are dominant and at other stages they are dominated—and they may be present in the speech community for centuries. New structures are recruited to pick up the slack and the diachronic cycle starts over again. But the diachronic cycle is not a reversal of directionality, only the termination of one unidirectional process and the restarting of another in the same general direction. At the point of restarting, speakers may choose to pursue other structural options for performing the same communicative functions. This is where a language, or a functional domain within a language, may change its structural phenotype (see §2.2.1). 3.2.4.  Written Language as a Factor The nature of written language is the fourth factor in language change and diffusion. It relates to the differences between speech and writing. Writing is secondary to speech and employs special forms of language for its unique purposes (for example, the use of devices for the organization of discourse; see §2.1.1).

74

Jacobus A. Naudé

The problem with biblical texts is that they are so heavily mediated: they were transmitted through multiple editors and copyists, rather than archaeologically excavated straight from their original context. Sanders (2009) approached the Biblical Hebrew texts in the light of recent epigraphic discoveries on the extreme antiquity of the alphabet and its use as a deliberate and meaningful choice. For 2,000 years, Near Eastern kingdoms and empires shared cuneiform—a script, not a language. In the Late Bronze Age, the Levant was ruled by the Egyptian empire, which used Babylonian cuneiform as an administrative tool, not a form of expression. This cosmopolitan writing system’s lack of inherent connection with spoken language was an advantage for the outstretched empire and linguistically unrelated agents. School texts were used to train scribes in the non–living languages of Sumerian and Old Babylonian. Scribes used writing to list people and things that the empire owned and to write letters for officials who could not have understood each other in person. It was the only way everyone involved could communicate. Alphabetic texts of the Iron Age played a role that was very different from cuneiform texts. Beginning with traces of previously oral genres, the newly written discourses used the technology of alphabetic writing to address a West Semitic culture of kinship and political communication transformed by and transforming a historical context of imperialism. This West Semitic culture of politics is found in the narratives of ninth-century royal inscriptions (for example, the inscription of Mesha and the Tel Dan inscription) alongside the first standardized local written languages of the first millennium, of which at least four appeared simultaneously in writing (Ammonite, Aramaic, Moabite, and Hebrew), each script adapted from Phoenician (or a related Canaanite script tradition) with no previous trace of their written existence. Hebrew as the first widely written vernacular introduced this change. Local writers began to inscribe things they had not learned from empires: prophecies addressed to a people, letters of protest and complaint, monuments by workmen, graffiti from fugitives. History began to speak in a nonmonarchic voice reflecting the distinctive biblical discourses being configured—history, law, and prophecy. This was a history with a people and prophets (not just kings) as the main actors; a law that directly addressed a collective “you” (not third-person imperial subjects); and a prophecy, claiming that ultimate authority was mediated by God’s messengers to a public. Sanders (2009:102) postulated on the basis of Pollock (2006: 328) “a strong tendency, perhaps even a law: it is only in response to a superposed and prestigious form of preexistent literature that a new vernacular literature develops.” Like the core of the Bible, the inscriptions from Israel speak in a standardized vernacular language to the public. The Siloam inscription tells the story of a royal building project from the viewpoint of the people who actually built it, without mentioning the king. Israel’s form of language was the result of its unique historical and social situation. Two worlds, local and cosmopolitan, existed side by side with no interest in what

Diachrony and a Theory of Language Change and Diffusion

75

the other was writing. The Mesopotamian and Egyptian cultures left so many more documents because they were the seats of geographically large empires, richly productive economies that could support many scribes, and long continuous written traditions on the scale of 3,000 years; as opposed to the less than 300 years from which there are continuous Hebrew epigraphic texts. None of those characteristics was true of Israel or Judah. Since the Hebrew Bible gives clear signs of having emerged from a distinctively Iron Age culture of communication, Sanders (2009: 7) claims that it is possible to project centuries (500–800 years) back from the existing biblical manuscripts to reconstruct the hypothetical ancestors of the biblical texts in the times of kings Josiah, Hezekiah, and Solomon and then, on the basis of this reconstruction, to reconstruct the scribal culture that produced them. The most plausible context for a thriving scribal literary culture is the late Judean monarchy (late eight through early sixth centuries). There is a total disappearance of Hebrew from the epigraphic record between the Babylonian Exile and the Hellenistic period. Writing continues as before, but in Aramaic, with many of the same generic uses: letters, economic documents, and property markers. This shift finds its cause in the destruction of the Judean political infrastructure in the early sixth century by the Babylonians, combined with the economic devastation and mass deportation of craftspeople and artisans. At the same time as written Hebrew completely vanishes from the epigraphic record, the most extensive internal evidence of Hebrew texts’ being edited and rethought appears (see van der Toorn 2007). The production of extended texts in Hebrew flourished in literature, not in daily life. Epigraphy does not demonstrate biblical texts as Iron Age artifacts, but the inscriptions and biblical texts can be read together to help explain key biblical discourses as new ways of writing history, law, and prophecy. These are the components essential for a theory of language change and diffusion, of which some will be illustrated in §4 with data on the wawconsecutive and distribution of the independent personal pronouns in Biblical and Qumran Hebrew. 4.  Biblical and Qumran Hebrew in Light of a Theory of Language Change and Diffusion Diachronic cycles of structural changes are not studied in Young et al. (2008). Structural change begins with the eroding of morphological structures. The loss of morphological structures then impacts the syntactic structures, which leads in turn to the eventual renovation of morphosyntactic structure. This diachronic cycle is evident in structural changes that began in Biblical Hebrew, escalated in Qumran Hebrew, and were finished in Mishnaic Hebrew. The syntax of the finite verb provides a good example of this sort of syntactic change.

76

Jacobus A. Naudé

In BH, the consecutive waw with perfective verb has the same semantic nuance as a preceding imperfective verb. Similarly, the consecutive waw with imperfective verb has the same semantic nuance as a preceding perfective verb. Less frequently in BH, in place of consecutive verbal forms, the simple conjunctive waw plus verbal form appears. In certain grammars (that is, idiolects) of Qumran Hebrew, the consecutive verbal forms are replaced with the corresponding simple verbal forms with conjunctive waw. The consecutive verbal forms at Qumran have become dramatically less frequent than the conjunctive waw with finite verb and, in fact, the conjunctive forms are statistically more frequent in Qumran Hebrew than in Biblical Hebrew. By the time of the Mishnah, consecutive verbal forms are not used at all except in biblical quotations. These structural changes are illustrated with the following examples: (2a) Isa 8:21

‫ְו ָהיָה ִכי־י ְִרעַב ְו ִה ְתקַ צַּף‬ and it will turn out that when they are hungry, they will be enraged (perfect consecutive) (2b) 1QIsaa 8:15

‫והיה כי־ירעב יתקצף‬ and it will turn out that when they are hungry, they will be enraged (imperfect) In (2a), the Biblical Hebrew verse has the consecutive waw with the perfective verb form. In the same biblical verse at Qumran in (2b), 1QIsaa, the consecutive verbal form has been replaced by the simple conjunctive form with the imperfective. The same situation can be seen in the biblical verse from Deuteronomy in (3a) and its appropriation at Qumran in the Rule of the Community in (3b). The consecutive waw construction with the perfective verb form in Biblical Hebrew (‫)ו ִה ְתבָּרֵ ְך‬ ְ is replaced by an imperfective verb form in Qumran Hebrew (‫)יתברך‬. (3a) Deut 29:18

ְ ‫ֶת־דּ ְברֵי ָה ָאלָה הַזֹּאת ְו ִה ְתבּ‬ ‫ָרֵך ִבּ ְלבָבוֹ‬ ִ ‫שׁ ְמעוֹ א‬ ָ ‫ְו ָהיָה ְבּ‬

and it shall be when he hears the words of this curse, that he will bless himself (perfect consecutive) in his heart

(3b) 1QS 2:12–13

‫והיה בשׁמעו את־דברי הברית הזות יתברך בלבבו‬ and it shall be when he hears the words of this covenant he will bless (imperfect) in his heart

Diachrony and a Theory of Language Change and Diffusion

77

Occasionally, however, waw-consecutive forms are attested at Qumran outside biblical quotations, as in (4): (4) 1QM 14:5 ‫ויקרא כושלים ל[גבור]ות פלא וקהל גויים אסף לכלה אין שארית‬ He has called those who stumble unto wondrous [accomplishments] and he has gathered a congregation of nations for annihilation without remnant In this example, the first sentence uses the waw-consecutive imperfect form (‫ )ויקרא‬to indicate a narrative event, whereas the second sentence uses a perfective form (‫)אסף‬. The waw-consecutive form is a stylistic fossil, a remnant of the Biblical Hebrew usage. Mishnaic Hebrew uses only the coordinate waw construction. In the Mishnaic Hebrew example in (5), the coordinate waw construction with perfective verb forms is used instead of the consecutive waw construction with imperfective verb forms that is found in Biblical Hebrew. (5) m. Soṭah 7:5 ‫ ונטלו את‬. . . ‫הביאו את האבנים ובנו את המזבח וסדוהו בסיד וכתבו עליו‬ ‫האבנים ובאו ולנו במקומן‬ They brought the stones, and built the altar, and plastered it with plaster, and wrote on it . . . and took the stones and came, and lodged in their place. These facts about the change and diffusion of a structural change concerning the consecutive verbal forms are well known. 2 However, from the standpoint of a theory of language change and diffusion as presented above, they are highly significant. In the biblical period, the change occurred in one or more idiolects. The change then diffused so that, by the time of Qumran, there were fewer examples of the consecutive form and more examples of conjunctive waw. The examples of consecutive forms that remained were stylistic fosssils. 2.  Kesterson (1984) mentions a number of aspects of the verbal system in Qumran Hebrew that differ from Biblical Hebrew: (1) The perfective verb form has supplanted the imperfective verb form for the expression of past activity after the conjunction term (Kesterson 1984: 7–8). (2) The positive volitive construction (simple waw plus volitive after volitive, question, etc.) is in demise and is replaced by an infinitive (Kesterson 1984: 10). (3) In 1QS, the coordinate waw construction with the imperfective verb form often appears where Biblical Hebrew would employ the consecutive waw construction with the perfective form, which points to some breakdown in the use of the consecutive waw in Qumran Hebrew (Kesterson 1984: 10–11). (4) Introductory wayhî (‘and it happened’) is absent (Kesterson 1984: 11). (5) In contrast with Biblical Hebrew, the predicative participle without the copular verb hāyâ can manifest the modal nuance must/should (Kesterson 1984: 12). For the data concerning the decline and loss of the waw-consecutive, see Smith 1991a; 1991b; and 1991c.

78

Jacobus A. Naudé

By the time of the Mishnah, the cycle of change was complete: the consecutive forms occurred only in biblical quotations, and the vernacular language was structurally different from the language in the Bible. These facts about the distribution and frequency of the consecutive and conjunctive verbal forms illustrate the S-shaped curve of change and diffusion. Furthermore, the morphological change of the loss of waw-consecutive was a deep structural change that affected other aspects of the syntax of Hebrew— namely, word order and the independent personal pronouns. Since the consecutive waw construction involved subordination, the consecutive waw was followed immediately by the verb, and nothing could be inserted between the conjunction and the verb, as illustrated in (6a): (6a) 1 Sam 22:18

‫הנִים‬ ֲֹ ‫ִפגַּע־הוּא בַּכּ‬ ְ ‫ַויִּסֹּב דּוֹאֵג ָהאֲד ִֹמי ַויּ‬ Doeg the Edomite set out, and he fell upon the priests When the lexical features of the waw changed, the obligatory connection of the verb to the consecutive was lost, resulting in a change in word order that affected independent personal pronouns. At Qumran, in contrast to Biblical Hebrew, the independent personal pronouns do not appear after consecutive perfective and imperfective verbs (see Naudé 1996), as illustrated in (6b): (6b) 11QTa 63:7–8

‫ואתה תבער את דם נקי‬

And you will eradicate the innocent blood The independent personal pronouns in Qumran Hebrew are used as topics with perfective and imperfective verb forms. They appear in the topic position; that is, they occur before the verb. The loss of the features of the waw-consecutive construction thus led to a change in word order and to a syntactic change in the independent pronouns. These additional changes and subsequent diffusions occurred parallel with the loss of waw-consecutive. These facts illustrate the Constant Rate Effect. Young et al. (2008) provide lists of these stylistic fossils in Biblical Hebrew. Their contribution, which I describe as a “Stylistic Fossil Approach,” shows that these stylistic fossils function/exist not only in Late Biblical Hebrew but also in Early Biblical Hebrew—a fact that Hurvitz (2006) was not willing to demonstrate. However, the stylistic fossil data that Young et al. (2008) present are not further analyzed/tested/verified according to the S-shaped curve or the Constant Rate Effect methodologies of historical linguistics. Their claim that these data do not reflect language change is not substantiated.

Diachrony and a Theory of Language Change and Diffusion

79

5. Conclusion Recent developments in contemporary diachronic linguistics (Fischer 2007), in theories of language evolution (Mufwene 2008), and in language complexity (Larsen-Freeman and Cameron 2008; Ellis and Larsen-Freeman 2009; Givón 2009; Sampson, Gil, and Trudgill 2009) as well as studies on the vernacular use of Biblical Hebrew (inter alia, Sanders 2009; Schniedewind 2005) have implications for the study of diachrony and language change in Biblical Hebrew. In light of these recent developments in contemporary linguistics, a theory of language change and diffusion (Naudé 2000a, 2000b, 2003, 2010) was further developed to include the following components: (a) the individual dimension: language change (the creation of an idiolect); (b) the sociological dimension: diffusion of language change; (c) the time dimension: diachronic/ chronological development; and (d) the nature of the written language. More research and refinement are necessary to further the developmental picture of Hebrew. With the assistance of diachronic cycles, the diachronic development/history/broad parameters of a language can be outlined, although dating for any historical linguistic will be relative as opposed to absolute (see the implications in the introductory chapter of Joseph and Janda 2003). However, with the assistance of methods from historical linguistics (cf. Pintzuk 2003), namely the S-shaped curve and Constant Rate Effect, these parameters can be more fully represented. References

Chambers, J. K.; Trudgill, P.; and Schilling-Estes, N., eds. 2002 The Handbook of Language Variation and Change. Malden, MA: Blackwell. Chomsky, N. 1975 Questions of Form and Interpretation. Pp. 159–96 in The Scope of American Linguistics: Papers of the First Golden Anniversary of the Linguistic Society of America. Lisse, The Netherlands: de Ridder. Clancey, W. J. 2009 Scientific Antecedents of Situated Cognition. Pp. 11–34 in The Cambridge Handbook of Situated Cognition, ed. P. Robbins and M. Aydede. New York: Cambridge University Press. Ellis, N. C., and Larsen-Freeman, D., eds. 2009 Language as a Complex Adaptive System. Oxford: Blackwell. Everett, D. 2005 Cultural Constraints on Grammar and Cognition in Piraha: Another Look at the Design Features of Human Language. Current Anthropology 76: 621–46. Fischer, O. 2007 Morphosyntactic Change: Functional and Formal Perspectives. Oxford: Oxford University Press.

80

Jacobus A. Naudé

Givón, T. 2009 The Genesis of Syntactic Complexity Diachrony, Ontogeny, Neuro-cognition, Evolution. Philadelphia: Benjamins. Halle, M.; Bresnan, J. W.; and Miller, G. A. 1978 Linguistic Theory and Psychological Reality. Cambridge, MA: MIT Press. Hauser, M.; Chomsky, N.; and Fitch, W. T. 2002 The Faculty of Language: What Is It, Who Has It, How Did It Evolve? Science 298: 1569–79. Hopper, P. J., and Traugott, E. C. 2003 Grammaticalization. Cambridge: Cambridge University Press. Hurvitz, A. 2006 The Recent Debate on Late Biblical Hebrew: Solid Data, Experts’ Opinions, and Inconclusive Arguments. HS 47: 191–210. Joseph, B. D., and Janda, R. D. 2003 The Handbook of Historical Linguistics. Malden, MA: Blackwell. Kesterson, J. C. 1984 Tense Usage and Verbal Syntax in Selected Qumran Documents. Ph.D. diss., Catholic University of America. Kroch, A. 1989 Reflexes of Grammar in Patterns of Language Change. Journal of Language Variation and Change 1: 199–244. Labov, W. 1994 Principles of Linguistic Change: Internal Factors. Oxford: Blackwell. Larsen-Freeman, D., and Cameron, L. 2008 Complex Systems and Applied Linguistics. Oxford: Oxford University Press. Lee, Namhee; Mikesell, Lisa; Joaquin, Anna Dina L.; Mates, Andrea W.; Schumann, John H. 2009 The Interactional Instinct: The Evolution and Acquisition of Language. Oxford: Oxford University Press. Lightfoot, D. 1979 Principles of Diachronic Synatx. Cambridge: Cambridge University Press. 1991 How to Set Parameters: Arguments from Language Change. Cambridge, MA: MIT Press. 1999 The Development of Language: Acquisition, Change, and Evolution. Malden, MA: Blackwell. Lightfoot, D., ed. 2002 Syntactic Effects of Morphological Change. Oxford: Oxford University Press. Mufwene, S. S. 2008 Language Evolution: Contact, Competition and Change. London: Continuum. Naudé, J. A. 1996 Independent Personal Pronouns in Qumran Hebrew Syntax: A Minimalist Approach. D.Litt. diss., University of the Free State, Bloemfontein. 2000a Diachronic Syntax and Language Change: The Case of Qumran Hebrew. Southern African Linguistics and Applied Language Studies 18: 1–14. 2000b Qumran Hebrew Syntax in the Perspective of a Theory of Language Change and Diffusion. JNSL 26: 105–32.

Diachrony and a Theory of Language Change and Diffusion

81

2003 The Transitions of Biblical Hebrew in the Perspective of Language Change and Diffusion. Pp. 150–63 in Biblical Hebrew: Studies in Chronology and Typology, ed. I. Young. London: T. & T. Clark. 2010 Linguistic Dating of Biblical Hebrew Texts: The Chronology and Typology Debate. JNSL 36: 1–22. Pinker, S. 1994. The Language Instinct. New York: Harper Perennial. Pintzuk, S. 2003 Variationist Approaches to Syntactic Change. Pp. 509–28 in The Handbook of Historical Linguistics, ed. B.  D. Joseph and R.  D. Janda. Malden, MA: Blackwell. Pollock, S. 2006 The Language of the Gods in the World of Men: Sanskrit, Culture, and Power in Premodern India. Berkeley: University of California Press. Rendsburg, G. A. 1990 Diglossia in Ancient Hebrew. New Haven, CT: American Oriental Society. Sampson, G.; Gil, D.; and Trudgill, P. 2009 Language Complexity as an Evolving Variable. Oxford: Oxford University Press. Sanders, S. L. 2009 The Invention of Hebrew. Urbana: University of Illinois Press. Schniedewind, W. M. 2005 How the Bible Became a Book: The Textualization of Ancient Israel. Cambridge: Cambridge University Press. Smith, M. S. 1991a The Origins and Development of the Waw-Consecutive: Northwest Semitic Evidence from Ugarit to Qumran. Atlanta: Scholars Press. 1991b Converted and Unconverted Perfect and Imperfect Forms in the Literature of Qumran. BASOR 284: 1–16. 1991c The Waw-Consecutive at Qumran. ZAH 3: 161–64. Toorn, K. van der 2007 Scribal Culture and the Making of the Hebrew Bible. Cambridge: Harvard University Press. Traugott, E. C., and Heine, B., eds. 1991 Approaches to Grammaticalization. 2 vols. Amsterdam: Benjamins. Young, I.; Rezetko, R.; and Ehrensward, M. 2008 The Linguistic Dating of Biblical Texts. 2 vols. London: Equinox.

Detecting Development in Biblical Hebrew Using Diachronic Typology John A. Cook 1. Introduction Studies of Biblical Hebrew have long operated with the familiar three-stage diachronic model of Archaic Biblical Hebrew, Standard Biblical Hebrew, and Late Biblical Hebrew (for example, Sáenz-Badillos 1993). In recent years, there has been growing criticism that this tripartite model is too facile to account for the language variation found in Biblical Hebrew. Scholars have argued that at least some of the language variation in the biblical text may admit other explanations, such as diglossia (for example, Rendsburg 1990), or regional dialects (for example, Rendsburg 1999; 2000), or language contact (for example, Young 1993: 60–63). These explanations, in contrast to the traditional tripartite analysis, share in common the view that the observed language variations are synchronic rather than diachronic in character. Although there is no a priori reason why diachronic and synchronic sorts of explanations should be mutually exclusive, Young, Rezetko, and Ehrensvärd (2008) have argued that scribal activity has left the biblical text in such a state of disarray as to render impossible the drawing of diachronic conclusions about Biblical Hebrew from the text. As Young succinctly puts it, “We [Young, Rezetko, and Ehrensvärd] claim instead that the nature of the biblical texts is such that this chronology, however, is not visible in any way that makes linguistic dating of biblical texts possible” (Young 2010: 1). In this essay, I take the contrasting position that linguistics offers usable models and methods for discerning diachronic differences in the language of the biblical text. I begin with a clarification of these contrasting positions. 2.  Misunderstandings and Differences of Opinion Young, Rezetko, and Ehrensvärd (2008) have performed an inestimable service to the study of Biblical Hebrew not only by prompting a reevaluation of entrenched positions but by pointing out some inherent weaknesses of these positions. In particular, they rightly make the criticism that, in general, studies of the language of the Bible have failed to grapple with the state of the text of 83

84

John A. Cook

the Bible. This failure to take text criticism into account is evident in discussions that appear to treat individual biblical books as datable in toto, when in fact a critical analysis of the text demonstrates that there are various textual layers within the books (see Young, Rezetko, and Ehrensvärd 2008: 2.1–71). We are indebted to them for pointing out these blind spots, as well as for bringing together a wealth of data and for summarizing previous studies. However, at the same time the very character of the disagreement has been muddled through some basic misunderstandings. On the one hand, these misunderstandings are ours—we who have made it our task to study the language of the Bible. In particular, we have sometimes failed to distinguish clearly between the text itself and the language represented in the text, the latter of which is the proper object of linguistic study. Citing the conflation of these two as common among historical linguists generally, Hale (2007: 22) states the distinction and its importance trenchantly: If we are to keep this discussion coherent at all, we must carefully distinguish between the features of the text (established by philological methods) and features of the language of the text (which can be established by a linguistic analysis of the contents of the text). This is not an easy task, since . . . philological and linguistic analyses show a mutual interdependence and are frequently carried out by one and the same person. However, accepting that it is necessary to distinguish between the text itself and the linguistic structures hypothesized to be evidenced in that text, the linguistics/​philology contrast seems to correlate with this distinction quite well.

On the other hand, there is a misunderstanding of this same distinction on the part of Young, Rezetko, and Ehrensvärd that is a product of their coming to the discussion as text-critical scholars: they have stated multiple times their central argument—namely, that linguistics cannot be used to date biblical texts (Young 2005; 2010, cited above). In light of Hale’s statement (above), this claim misses the real point of contention: linguistics is not tasked with dating biblical texts; this is one of the chores of philology. Linguistics is concerned with language and philology with texts (see Barr 1969; Gleason 1974; Bodine 1987; Miller 2004; Holmstedt 2006). Once this is clarified, it becomes evident that the claim about linguistics’ being incapable of dating biblical texts is correct, though not for the reasons that Young, Rezetko, and Ehrensvärd claim, because their claim is based on the nature of the biblical text and not the character of the linguistic versus philological enterprises. Even if one concedes that the text is as skewed by scribal activities as Young, Rezetko, and Ehrensvärd argue, it cannot follow that the text is bereft of any usable linguistic data, only that it is more difficult to obtain these data than has sometimes been realized or admitted. Young, Rezetko, and Ehrensvärd (2008) concede this argument to a point: they themselves argue that there is language variation in the text and attribute it to scribal dialectal differences rather than

Detecting Development in Biblical Hebrew

85

the traditional diachronic explanation. But this sort of blanket explanation falls under the same censure that they have leveled at the traditional diachronic model. One cannot a priori rule out certain explanations of language variation, as Young, Rezetko, and Ehrensvärd have done with diachronic explanations; each case must be individually examined. Although Young, Rezetko, and Ehrensvärd have done the field of biblical studies—especially Biblical Hebrew linguistics—a yeoman’s service, it is not in the form of an alternative theory that explains the language variation in the biblical text, as the traditional chronological model tried to do; rather, they have given a clarion call to scrutinize the foundations and begin sifting through the data anew. If we are to proceed with such a reassessment of the data in good faith, we must not prejudge any language variation datum as precluding explanation by any of the possible accounts of language variation (and others) listed in the beginning of this essay—including diachronic models. To make this point, I illustrate below that some sorts of language variation attested in the biblical text are best explained as a consequence of diachronic (or chronological) change and diffusion. Given the nature of the debate, there are two desiderata to make my case. First, I want to find the sort(s) of language variation that is/are not susceptible to “imitation” (for example, archaizing) by scribes in order to exclude intentional scribal change as an explanation for the variation. Such variation must be of a sort that affects an entire part of the grammar system, making it increasingly impossible, because of diffusion of the new structure, to avoid ambiguity or confusion by the employment of the earlier grammatical construction. For example, English second-personsingular pronouns thee/thou/thy fell out of use and were replaced by the originally plural “you” form, which now serves for both singular and plural. As a result, employment of the archaic informal thee/thou/thy forms nowadays is generally mistaken for archaic formal speech. Second, I want some external means of validating any claim that the variation in question is more likely a diachronic change than some variety of synchronic variation. Diachronic typology serves as just such a means of external validation in that it provides a wealth of data regarding typical diachronic changes in human language (see Heine and Kuteva 2002). In the following section, I discuss in more detail diachronic typology, somewhat anticipating its application to the case of ‫ידע‬ ‘know’ in Biblical Hebrew. 3.  Diachronic Typology and Tense-Aspect-Mood (TAM) Systems Linguistic typology classifies languages in terms of a given linguistic structure and then develops generalizations regarding patterns of linguistic structures across languages (Croft 2003: 1). For example, languages may be classified based on whether or not they have a perfective verb conjugation. Based on discernible patterns of perfective verbs in TAM (tense-aspect-mood)

86

John A. Cook

systems across languages, linguists have made the following generalization: perfective verbs only develop in TAM systems that already have an imperfective verb, in opposition to which the new perfective stands (Bybee, Perkins, and Pagliuca 1994: 92). Hence, in analyzing Biblical Hebrew, if we conclude that it has a perfective verb conjugation, such as the Perfect (qatal), then based on this generalization we might reasonably identify the Imperfect (yiqtol) as an imperfective verb. Diachronic typology, to use Croft’s (2003: 233) term, represents a “dynamicization” of typology, whereby synchronic language states are reanalyzed as stages in the process of language change. Thus, the typological tasks of classification and generalization become applied in diachronic typology to the developmental axis of language structures. In other words, based on the classification of a given linguistic structure at one stage of a language, generalizations are developed regarding preceding and following stages in the development of that particular linguistic structure. An oft-cited example is that resultative constructions frequently give rise to perfect forms, as illustrated by a comparison of the kjv and the nrsv translations of the verb in Deut 13:13: are gone out (resultative, kjv) versus have gone out (perfect, nrsv). Several generalizations have emerged from a variety of studies over the past quarter century about tense-aspect-mood (TAM) systems that illustrate diachronic typology. One of the generalizations is described by Heine (2005: 594) as follows: “Verbal aspect categories give rise to tense categories, . . . while processes in the opposite direction are unlikely to happen.” A more specific example, consistent with this generalization, is that imperfective and present verb conjugations develop from progressive constructions (Bybee, Perkins, and Pagliuca 1994: 91; Heine 2005: 594). 4.  Stative Encoding and Aspect-Tense Shift One well-known example of linguistic variation in Biblical Hebrew is the predicate “split-encoding” of stative adjectives, such as ‫מלא‬, ‫כבד‬, and ‫זקן‬. When used predicatively, this small, closed class of adjectives may be encoded either as verbs or as nouns; in the latter case, they are copular complements (see Cook 2008). These alternatives are illustrated in (1). (1)  Stative encodings a.  Verbal encoding

‫ָמים‬ ִ ‫ָאתי ַּבּי‬ ִ ‫ְּתי ּב‬ ִ ‫אנִי זָקַ נ‬ ֲ

I am old/aged, advanced in days. (Josh 23:2) b.  Nominal encoding

‫ָמים‬ ִ ‫ָאים ַּבּי‬ ִ ‫ׂש ָרה ְז ֵקנִים ּב‬ ָ ‫ְוא ְַב ָרהָם ְו‬ Now Abraham and Sarah (were) old, advanced in days. (Gen 18:11)

Detecting Development in Biblical Hebrew

87

My characterization of the stative as a small, closed class of adjectives is justified by a comparison of the number and frequency of stative pattern adjectives versus other adjectival patterns in Biblical Hebrew and from the pattern’s lack of productivity in postbiblical Hebrew; that is, no new split-encoded stativepattern adjectives appear in postbiblical Hebrew. However, the question is whether the decline of the stative adjective is diachronically significant. In order to answer this question, I first need to construct a larger diachronictypological argument that will provide a framework within which to answer this question. 4.1.  The Diachronic-Typological Argument My diachronic-typological argument involves three interrelated generalizations. The first, which I mentioned above in illustrating diachronic typology (§3), is that TAM systems tend to develop from aspectual categories and oppositions toward tense systems rather than the reverse (Heine 2005: 594). The second generalization is that perfective and past conjugations are distinct in their interaction with stative predicates, so as to create a marked opposition (see Bybee, Perkins, and Pagliuca 1994: 95): the unmarked member of the opposition is the perfective with stative predicates, which may express either present or past states; the marked member of the opposition is the past with stative predicates, which may only express past states. This opposition characterizes the Perfect (qatal) and Past Narrative (wayyiqtol) conjugations in Biblical Hebrew, as illustrated in (2). (2)  The stative with Perfect and Past Narrative a.  Qal Perfect + stative = present state ‫א־רמּו עֵינַי‬ ָ ֹ ‫יְהוָה לֹא־ָגבַּה ִל ִּבי ְול‬ Yhwh, my heart is not high / my eyes are not elevated. (Ps 131:1). b.  Qal Perfect + stative = past state ‫ְקּיָהּו ִּכי ָגבַּה ִלּבֹו‬ ִ ‫ְחז‬ ִ ‫ֵׁשיב י‬ ִ ‫א־כ ְגמֻל ָעלָיו ה‬ ִ ֹ ‫ְול‬ But Hezekiah did not respond according to the benefit to him, because his heart was high. (2 Chr 32:25). c.  Qal Past Narrative + stative = past state ְ ‫ִתיַּצֵב ְּב‬ ‫ִג ַּבּה ִמּכָל־ ָהעָם‬ ְ ‫תֹוך ָהעָם ַוּי‬ ְ ‫ַוּי‬ And he stood in the midst of the people and he was taller than all the people. (1 Sam 10:28) The third generalization Stassen (1997: 347) terms the “tensedness universal of adjective encoding.” In his study of intransitive predication, Stassen shows a correlation between the choice of encoding strategy for adjectival predicates and whether a language’s TAM system is aspectual or tensed: languages that

88

John A. Cook

have predominately aspectual categories tend to encode adjectival predicates according to their verbal strategy, whereas languages with predominately tense categories tend to encode adjectival predicates according to one or more of their nominal strategies (see Stassen 1997: 347–57). These three generalizations are interrelated: as a language’s TAM system shifts from aspect toward tense, which may happen according to the first generalization, its aspectual categories decline in productivity or may shift in meaning. For example, a perfective conjugation may become past tense; such a shift could be discerned by its pattern of interaction with statives as outlined in the second generalization. As the shift occurs, the strategies for expressing present states also shift, based on the tensedness parameter, given as the third generalization. 4.2.  The Data The foregoing diachronic-typological argument points to the decline of the stative in Hebrew as diachronically significant. But while this decline is easily appreciated through a comparison of Biblical and postbiblical Hebrew, the question remains to what extent this diachronic shift is measurable within the Biblical Hebrew corpus. The obvious place to look would be the variation of the verbal and nominal encoding of stative predicates. Unfortunately, we face the obstacle of a dearth of data, which is compounded by having to exclude the ambiguous masculine-singular form (i.e., the masculine-singular nominal encoding and Qal third-person masculine-singular verbal encoding are morphologically identical—for example, ‫— ָּכבֵד‬and syntactically identical in the absence of an overt copula). Thus, among 60 of the most frequent stative verbs, there are only 12 that occur frequently enough to exhibit unambiguous split encoding. These 12 verbs provide 183 instances of unambiguous verbal encoding and 57 occurrences of unambiguous nominal-encoded predicates. The sparse data caution us against drawing hasty conclusions. However, the following observations are warranted with reference to the table in example (3) on p. 89. First, 7 books (Amos, Obadiah, Haggai, Song of Songs, Esther, Daniel, and Ezra) offer no data, and another 8 show just one example—either nominal or verbal. It may reflect the general decline in the use of the stative adjective patterns that 11 of these 14 books are philologically datable to the exilic or postexilic periods (Jonah, Chronicles, Ruth, Nehemiah, Qoheleth, Song of Songs, Obadiah, Ezra, Haggai, Esther, and Daniel). 1 Second, and by contrast, 1.  Young’s (personal communication) complaint that analyses of Biblical Hebrew fail to distinguish textual layers in the books may have some validity, but the convention of referencing data by book, chapter, and verse, as well as the fact that the books have long been recognized as distinct compositions from one another, some with clear termini post quem based on their content (for example, postexilic Ezra-Nehemiah and Chronicles), make continued reference to data by book convenient, at least at this level of preliminary analysis. Note

89

Detecting Development in Biblical Hebrew

Book Psalms Isa 1–39 Deut Jer Ezek Job Lev Gen Sam Isa 40–66 Judg Joel Josh Exod Hab

Verbal 24 18 18 17 14 12 11  9  8  7  6  4  4  4  3

Nominal 2 0 2 4 5 1 7 1 3 1 0 0 1 3 0

Book Zech Mic Kgs Hos Prov Mal Lam Zeph Jon Chron Ruth Neh Nah Num Qoh

Verbal 3 3 3 2 2 2 2 1 1 1 1 1 1 1 0

Nominal 0 1 8 0 0 0 2 0 0 0 0 0 0 0 1

(3)  Verbal versus nominal encodings of 12 stative adjectives 2 most of the books with a majority of verbal encoding contain material philologically datable to the preexilic period (14 of 21: Genesis, Exodus, Leviticus, Deuteronomy, Joshua, Judges, Samuel, Isaiah 1–39, Jeremiah, Hosea, Micah, Habakkuk, Psalms, and Proverbs).  2 A strategy to increase the database is to examine the general decline of verbally encoded statives in the Qal Perfect between the “Standard Biblical Hebrew” corpus of Genesis through 2 Kings and the corpus traditionally classified as “Late Biblical Hebrew” (Ezra, Nehemiah, Chronicles, Haggai, Zechariah, Malachi, Esther, Daniel, and Qoheleth). 3 The frequency of 60 statives in how my statements here illustrate the requisite give and take between linguistic analysis and philology in investigating these matters; neither discipline can nor should work in isolation. 2.  The 12 statives are ‫ ׂשמח‬,‫ קרב‬,‫ קלל‬,‫ מלא‬,‫ כבד‬,‫ ירא‬,‫ יבׁש‬,‫ טמא‬,‫ חפץ‬,‫ חדל‬,‫ זקן‬,‫דלל‬. 3.  I have placed these terms (Standard and Late Biblical Hebrew) in quotation marks because of their centrality to the dispute over whether the language represented in these texts is datable. I am convinced, however, that we should interact with the previous consensus (albeit no longer naïvely) rather than rejecting it wholesale. The rejection strategy has led Young, Rezetko, and Ehrensvärd (2008) to conclude that no diachronic observations about the text are possible. Certainly, if two texts differ chronologically on philological grounds, there is every possibility that differences in their language are diachronic. Here I do not intend to defend or validate the traditional diachronic scheme, but I am using it as a foil to demonstrate that the variations seen in the Hebrew stative occurrences and encodings are likely diachronic in character inasmuch as they do not entirely contradict conventional wisdom regarding the relative dating of these texts, as the graph in (6) below demonstrates.

90

John A. Cook

Qal Perfect as a percentage of all Qal Perfects in these books shows an almost equal amount between the two groups of books—about 22% of all perfects are stative verbs. 4 However, the data are skewed by the increase of ‫היה‬, which jumps from 8% to 12.5% frequency in the data. Given the shift of the Hebrew TAM from aspectual categories toward tense, this increase is fully expected given that the copula is a main strategy for overtly signaling tense.  5 If we set aside this single stative verb, the data show a decline in frequency from 14% in the SBH corpus to 9% in LBH. 6 Yet a third means of examining the data is to chart the decline of present state expression by statives verbs in the Qal Perfect conjugation. To do this, we must identify the “new” strategy or strategies employed for the expression of present states for each stative verb. Here I illustrate with ‫ ידע‬because of its relative frequency (over 300 occurrences), though I have found similar but less consistent results in analyses of other stative verbs. 7 The shift from present to past state is illustrated in (4a–b), while the examples in (4c–d) illustrate the main alternative strategies: an active participle encoding and an Imperfect verbal encoding. 8 (4)  Encoding strategies and the interpretation of ‫ידע‬ a.  Qal Perfect = present state ‫ָאתי‬ ִ ‫אנִי ָחט‬ ֲ ‫ִּכי יָדַ ע ע ְַב ְּדךָ ִּכי‬ For your servant knows that I have sinned. (2 Sam 19:21; see also 2 Sam 14:22) b.  Qal Perfect = past state

 ַ‫י־מ ִּל ְפנֵי יְהוָה הּוא בֹרֵ ח‬ ִ ‫ָׁשים ִּכ‬ ִ ‫אנ‬ ֲ ‫ָדעּו ָה‬ ְ ‫ִּכי־י‬ For the men knew that he was running away from Yhwh. (Jonah 1:10)

c.  Qal Active Participle = present state ‫ׁשל`ִּי ַה ַּסעַר ַהּגָדֹול ַהּזֶה עֲלֵיכֶם‬ ֶ ‫ִּכי יֹודֵ ַע ָאנִי ִּכי ְב‬ For I know that on my account this storm has (come) upon you. (Jonah 1:12) 4.  Verheij (1990: 32) has demonstrated that there is a general decline in the frequency of verbal predications between Samuel–Kings and Chronicles. Therefore, it is preferable to measure the data in terms of the frequency of the stative adjective in Qal Perfect as a percentage of all Qal Perfects in these books. 5.  This rise is further confirmed by the frequency of ‫ היה‬in the post-BH literature: 83 times in Ben Sira, 945 times in Qumran, and 1,738 times in the Mishnah, based on the databases available in Accordance software. 6.  This represents an actual decrease of 457 examples (564 versus 107 occurrences). 7.  I have analyzed ‫מלא‬, ‫חפץ‬, ‫אהב‬, and ‫ ׂשנא‬with less-consistent results, probably due to the size of the data. 8.  For simplicity’s sake, I am excluding lexical alternatives, which are in any case much less frequent than the strategies examined here; for example, the Hiphil of ‫( בין‬Job 28:23).

Detecting Development in Biblical Hebrew

91

d.  Qal Imperfect = present state

‫ֶג ִּדי ָת ִמיד‬ ְ ‫אתי נ‬ ִ ‫ַּט‬ ָ ‫אנִי א ֵָדע ְוח‬ ֲ ‫ׁשעַי‬ ָ ‫י־פ‬ ְ ‫ִּכ‬ For I know my transgressions / and my sin is continually before me. (Ps 51:5)

The data appear in (5) arranged in ascending frequency of new constructions as a percentage of all occurrences in the data. The list excludes the nine books that offer no relevant data. 9 (5)  Old forms and new forms of ‫ ידע‬expressing a present state (see p. 92) A source of widespread confusion in discussions of linguistic data and dating of biblical texts is the difference between linguistic change and linguistic diffusion. Responding to this confusion in historical linguistics generally, Hale (2007: 33) explains that change is strictly intergenerational, “when transmission is flawed with respect to some features.” Diffusion is the spread of change among speakers and is not instantaneous but exhibits a distinct pattern that has been described as an S-curve: the diffusion of the new construction is gradual in the beginning, accelerates in the middle stage, and slows again as it gains ascendancy at the close of the process (Pintzuk 2005: 512). Thus, we fully expect to find old and new constructions side-by-side in the data, but we also expect to find an increasing frequency of the new form in the shape of an S-curve. The graph in (6) charts the data from (5) against the plot of an ideal S-curve. I have excluded from this chart the six books that offer just one example as statistically less significant (i.e., it is difficult to draw any conclusions from these). (6)  Frequency of new construction expressing a present state for ‫ידע‬

9.  Leviticus (including 5 irrealis qatal forms of ‫)ידע‬, Obadiah, Habakkuk, Haggai, Malachi, Ruth, Lamentations, Ezra, and Nehemiah.

92

John A. Cook

Book Perf.a Partic.b Impf.c Freq. %

Book Perf. Partic. Impf. Freq. %

Song

1

0

0

0%

Jer

17

1

3

19.0%

Mic

1

0

0

0%

Sam

17

3

2

22.7%

Dan

1

0

0

0%

Isa

11

1

3

26.7%

Judg

3

0

0

0%

Job

20

0

8

28.6%

Ezek

3

0

0

0%

Ps

22

6

6

35.3%

Amos

3

0

0

0%

Prov

6

3

4

53.8%

Zech

3

0

0

0%

Chron

2

1

3

66.7%

Hos

4

0

0

0%

Jon

1

2

0

66.7%

Num

4

0

0

0%

Qoh

5

12

4

76.2%

Exod

11

0

0

0%

Est

0

2

0

100.0%

Deut

10

0

1

9.1%

Nah

0

1

0

100.0%

Gen

17

2

0

10.5%

Joel

0

1

0

100.0%

Kgs

22

1

3

15.4%

Zeph

0

1

0

100.0%

Josh

5

1

0

16.7%

Total 189

38

37

28.41%

a. Gen 4:9; 12:11; 18:19; 20:6; 21:26; 22:12; 27:2; 29:5; 30:26, 29; 31:6; 39:8; 43:22; 44:15, 27; 47:6; 48:19; Exod 3:7, 19; 4:14; 5:2; 9:30; 18:11; 23:9; 32:1, 22–23; 33:12; Num 10:31; 11:16; 20:14; 22:6; Deut 1:39; 3:19; 7:15; 9:2; 22:2; 28:33; 29:15; 31:21, 27; 34:6; Josh 2:4–5, 9; 14:6; 22:31; Judg 15:11; 17:13; 18:14; 1 Sam 17:28; 20:3, 30; 22:15; 24:21; 25:11; 28:9; 29:9; 2 Sam 1:5; 2:26; 3:25; 7:20; 14:22; 17:8; 19:7, 21, 23; 1 Kgs 1:11, 18; 2:5, 15, 44; 5:17, 20; 8:39; 17:24; 22:3; 2 Kgs 2:3, 5; 4:1, 9; 5:15; 7:12; 8:12; 9:11; 17:26; 19:27; Isa 1:3; 29:12; 37:28; 42:16; 44:18; 56:10–11; 59:8, 12; 63:16; Jer 1:6; 4:22; 5:4–5; 9:2; 10:23, 25; 11:19; 12:3; 14:18, 20; 15:15; 17:16; 18:23; 29:11; 33:3; 48:30; Ezek 11:5; 17:12; 37:3; Hos 5:3–4; 7:9; 8:2; Amos 3:2, 10; 5:12; Jonah 4:11; Mic 4:12; Zech 4:5, 13; 7:14; Ps 14:4; 20:7; 31:8; 35:11; 40:10; 41:12; 50:11; 53:5; 56:10; 69:6, 20; 71:15; 73:11; 79:6; 82:5; 91:14; 119:75; 135:5; 139:2, 4; 140:13; 142:4; Job 9:2, 28; 10:13; 12:9; 13:2, 18; 15:9; 19:25; 20:4; 21:27; 22:13; 23:10; 30:23; 32:22; 34:33; 35:15; 38:21, 33; 39:1; 42:2; Prov 4:19; 7:23; 9:13, 18; 23:35; 30:18; Song 6:12; Qoh 3:12, 14; 4:13; 7:22; 10:15; Dan 10:20; 2 Chr 2:7; 25:16 (two others from Chronicles are excluded since they are paralleled in the Samuel–Kings source text). b. Gen 3:5; 33:13; Josh 22:22; 1 Sam 23:17; 2 Sam 12:22; 17:10; 2 Kgs 17:26; Isa 29:15; Jer 29:23; Joel 2:14; Jonah 1:12; 3:9; Nah 1:7; Zeph 3:5; Ps 1:6; 37:18; 44:22; 90:11; 94:11; 139:14; Prov 12:10; 14:10; 24:22; Qoh 2:19; 3:21; 4:17; 6:12; 8:1, 7, 12; 9:1, 5; 11:5–6; Esth 4:11, 14; 2 Chr 2:7. c. Deut 20:20; 1 Sam 20:9; 2 Sam 3:38; 1 Kgs 3:7; 8:39; 18:12; Isa 40:21; 55:5; 58:3; Jer 5:15; 13:12; 40:14; Ps 35:8; 39:7; 51:5; 73:22; 92:7; 138:6; Job 8:9; 11:8; 15:9; 36:26; 37:5, 15–16; 38:5; Prov 24:12; 27:1; 28:22; 30:3; Qoh 9:12; 10:14; 11:2, 5; 2 Chr 6:30; 20:12; 32:13.

Notice that the frequency of new and old constructions in the books that exhibit both is fairly consistent with the philological dating of the various books: at the low end are, notably, a majority of the books from the Primary History (Genesis, Exodus, Numbers, Deuteronomy, Joshua, Samuel, and Kings) along with the prophetic books of Hosea, Jeremiah, and Isaiah; at the midpoint, we

Detecting Development in Biblical Hebrew

93

find the mixed poetic material of Job, Psalms, and Proverbs; and at the high end, we find Chronicles, Jonah, and Qoheleth. This analysis should, of course, be further refined by distinguishing among blocks of material within the different books and by including postiblical Hebrew data. 10 However, it is a promising beginning: promising inasmuch as it does not depart radically from the received wisdom of philological dating of these books; a beginning in that we need to begin to move forward with the data, plotting each individual phenomenon in a similar manner in order cautiously and securely to build up a picture of the diachronic development of BH. 11 5. Conclusion No one will deny that the linguistic data of BH are difficult. For this reason, not only must we be cautious in drawing diachronic conclusions from them, but at the same time we should not quickly give up the enterprise, as has been suggested. Other possibilities may present themselves for drawing meaningful conclusions from the data. Among these possibilities, I have examined how diachronic typology serves to set up expectations of the data that make the observed variations more meaningful. 10.  Some general statistics on ‫ ידע‬in the postbiblical materials are as follows: Ben Sira has 3 Perfects (present) versus 5 Participles and 5 Imperfects; the Qumran data are large enough to require separate treatment (122 Perfects; 72 Participles; 71 Imperfects); the Mishnah shows a clear preference for Participle (112 times) versus 51 Perfects and 16 Imperfects. 11.  With regard to absolute dating, one would want to rely on the relatively more-fixed dates of the epigraphic texts. Unfortunately, they provide little help with regard to the stative, which is almost wholly lacking in the preexilic texts. It is only somewhat confirmatory that there are three instances in the preexilic epigraphic texts of ‫ ידע‬conjugated (unambiguously) in the Perfect (Arad 40:9; Lachish 2:6; 3:8), two of which clearly express present states (Lachish 2:6; 3:8).

References

Barr, J. 1969 The Ancient Semitic Languages: The Conflict between Philology and Linguistics. Transactions of the Philological Society 1968: 37–55. Bodine, W. R. 1987 Linguistics and Philology in the Study of Ancient Near Eastern Languages. Pp. 39–54 in “Working with No Data”: Semitic and Egyptian Studies Presented to Thomas O. Lambdin, ed. D. M. Golomb. Winona Lake, IN: Eisenbrauns. Bybee, J.; Perkins, R.; and Pagliuca, W. 1994 The Evolution of Grammar: Tense, Aspect, and Modality in the Languages of the World. Chicago: University of Chicago Press. Cook, J. A. 2008 The Participle and Stative in Typological Perspective. JNSL 34/1: 1–19.

94

John A. Cook

Croft, W. 2003 Typology and Universals. 2nd ed. Cambridge Textbooks in Linguistics. Cambridge: Cambridge University Press. Gleason, H. A. 1974 Linguistics and Philology. Pp.  199–212 in On Language, Culture, and Religion: In Honor of Eugene A. Nida, ed. M. Black and W. A. Smalley. The Hague: Mouton. Hale, M. 2007 Historical Linguistics: Theory and Method. Blackwell Textbooks in Linguistics 21. Oxford: Blackwell. Heine, B. 2005 Grammaticalization. Pp. 575–601 in The Handbook of Historical Linguistics, ed. B. D. Joseph and R. D. Janda. Blackwell Handbooks in Linguistics. Oxford: Blackwell. Heine, B., and Kuteva, T. 2002 World Lexicon of Grammaticalization. Cambridge: Cambridge University Press. Holmstedt, R. D. 2006 Issues in the Linguistic Analysis of a Dead Language, with Particular Reference to Ancient Hebrew. JHS 6/11. http:// www​.jhsonline.org/. Miller, C. L. 2004 Methodological Issues in Reconstructing Language Systems from Epigraphic Fragments. Pp. 281–305 in The Future of Biblical Archaeology: Reassessing Methodologies and Assumptions, ed. J. K. Hoffmeier and A. Millard. Grand Rapids, MI: Eerdmans. Pintzuk, S. 2005 Variationist Approaches to Syntactic Change. Pp. 509–28 in The Handbook of Historical Linguistics, ed. B. D. Joseph and R. D. Janda. Blackwell Handbooks in Linguistics. Oxford: Blackwell. Rendsburg, G. A. 1990 Diglossia in Ancient Hebrew. New Haven, CT: American Oriental Society. 1999 Notes on Israelian Hebrew (I). Pp. 255–58 in Michael: Historical, Epigraphical and Biblical Studies in Honor of Prof. Michael Heltzer, ed. Y. Avishur and R. Deutsch. Tel-Aviv: Archaeological Center Publications. 2000 Notes on Israelian Hebrew (II). JNSL 26/1: 33–45. Sáenz -Badillos, A. 1993 A History of the Hebrew Language, trans. J. Elwolde. Cambridge: Cambridge University Press. Stassen, L. 1997 Intransitive Predication. Oxford Studies in Typology and Linguistic Theory. Oxford: Clarendon. Verheij, A. J. C. 1990 Verbs and Numbers: A Study of the Frequencies of the Hebrew Verbal Tense Forms in the Books of Samuel, Kings, and Chronicles. SSN 27a. Assen, The Netherlands: Van Gorcum.

Detecting Development in Biblical Hebrew

95

Young, I. 1993 Diversity in Pre-exilic Hebrew. FAT 5. Tübingen: Mohr. 2005 Biblical Texts Cannot be Dated Linguistically. HS 46: 341–51. 2010 Text Critical Observations on the (Im)Possibility of Linguistic Dating of Hebrew Biblical Texts. Paper read at National Association of Professors of Hebrew, Atlanta, GA. Young, I.; Rezetko, R.; and Ehrensvärd, M. 2008 Linguistic Dating of Biblical Texts. 2 vols. London: Equinox.

Historical Linguistics and Biblical Hebrew Robert D. Holmstedt 1. Introduction “Variety is the spice of life,” goes the old proverb, and life in Hebrew historical linguistics has become very spicy indeed. For generations of Hebraists who grappled with linguistic variation in the Hebrew Bible and other ancient Hebrew sources, situating a given feature diachronically was often the first and primary concern. Along the way, some features were given other, nonchronological explanations, such as northern-versus-southern dialects and literaryversus-colloquial registers, but these sorts of explanation were often proposed if and only if a chronological explanation was not apparent. Now, thanks to the tenacious work of certain “Young Turks,” there is an alternative explanation for the variation in the Hebrew Bible that is at its core nonchronological: LBH [is] merely one style of Hebrew in the Second Temple and quite possibly First Temple periods. Both EBH and LBH are styles with roots in preexilic Hebrew, which continue throughout the postexilic period. ‘Early’ BH and ‘Late’ BH, therefore, do not represent different chronological periods in the history of BH, but instead represent coexisting styles of literary Hebrew throughout the biblical period. (Young and Rezetko 2008: 2.96).

To switch from spice to the high seas, this is not the proverbial shot across the bow but a full broadside against diachronic Hebrew studies. The implications are enormous: in the process of arguing that it is impossible to date texts linguistically, these authors have effectively blocked access to diachronically meaningful data, and the entire history of Hebrew has been rewritten. That is, within their alternative explanation, there is no discernible linguistic history in the Hebrew Bible. In the face of this broadside, the question is, of course, will the “diachronic” ship sink? 2.  Methodological Issues One of the healthiest, most productive activities following a strong challenge to a consensus is a careful, agenda-free assessment of methodological principles and practices. We should put aside the Hebrew data and analyses for a long moment and revisit the linguistic literature that addresses diachronic 97

98

Robert D. Holmstedt

issues. We should be asking ourselves regularly, “How is this done in general?” and “How can this be applied to Hebrew, if at all?” Although there may be a few new sets of data to analyze, most of the issues are well known and certainly are not going to change (now). Thus, we should take this opportunity to reestablish an explicit and sound methodology for discussing variation and language change in ancient texts. 2.1.  Artifacts, Reconstructed Texts, and Grammars: What Is the Object of Study? A primary challenge to carrying out historical linguistic research on biblical texts is the nature of the texts themselves. Linguistic research on ancient Hebrew texts requires the researcher to be both philologist and linguist. We must not only deal with the challenges posed by paleography, orthography, textual lacunae, and—especially for the Bible—the complexities of composition and textual traditions, we must also navigate the theories and terminology of general linguistics. In concrete terms, this means, for example, that one cannot simply pick up a Hebrew Bible, turn to some datum in Isaiah 45, take it as evidence from the late eighth century b.c.e., and work it into a linguistic analysis. That is, we must be aware of the majority positions in biblical studies: namely, that Isaiah 45 is typically understood to have late-sixth-century origin, based on, among other clues, the mention of Cyrus (550–530 b.c.e.), the Persian ruler who took Babylon in 539 b.c.e. Given the complexity of Hebrew historical linguistics, how can we go about this task? Mark Hale, in his Historical Linguistics: Theory and Method (2007), presents what I consider to be one of the clearest discussions of philology as it deals with linguistic artifacts. 1 Hale attempts to clarify the real object of historical-linguistic study when we are working with ancient languages. We all begin with historical artifacts as the source of linguistic data, whether an inscription, a scroll from the Qumran caves, or a medieval Masoretic codex, but these are not in fact the primary object of linguistic study. Due to the accidents of history, the chronological sequence of artifacts may not reflect the chronological sequence of the linguistic data they contain. For example, most of us think that Masoretic codices, such as B19a (the Leningrad Codex), often preserve earlier data than the corresponding manuscripts from Qumran, written a millennium earlier. Thus, philologists establish “the attributes of a text, many of which may be relevant for subsequent linguistic analysis” (Hale 2007: 21). The product of philological analysis is localized and dated texts (dated relative to other sources, at least). Yet even this reconstructed philological text is not the object of historical-linguistic analysis. 1.  On the distinction between philology and linguistics, see also Holmstedt 2006. On the challenges in reconstructing ancient languages, see Miller 2004.

Historical Linguistics and Biblical Hebrew

99

Figure 1.  The relationship between grammars (Hale 2007: 25, fig. 2.6). Reproduced with permission.

Historical linguists are not interested in the “texts” but what the texts contain, specifically the relationships between linguistic structures contained within the different texts (Hale 2007: 22 n. 5). Hale asserts that “a rigorous linguistic analysis of the text which was philologically established from a given historical artifact would lead, in principle, to a hypothesized linguistic structure for the relevant aspects of that text” (2007: 23). This linguistic text presents us with a set of linguistic representations for the analyzed philological text. Again, it is not the texts (whether the artifact, the philologist’s reconstruction, or the linguist’s reconstruction) that are compared in historical linguistics but the grammatical features—and the grammars behind these features—represented within the texts (see 2007: 25). Figure 1 presents Hale’s visualization of the layers involved in getting to the real object of historical linguistic analysis. Figure 1 illustrates the complexity of the historical linguist’s task. In the top row are the artifacts, the concrete texts existing on some physical surface. Assuming that each artifact can be dated, even roughly, by some external factor (such as archaeological context, scribal note, historical reference on the manuscript), the next challenge is determining the status of the text represented by the artifact. The crossing dotted lines in Hale’s figure represent the not uncommon situation in which an artifact (for example, the physical manuscript on which text exists) is externally dated to a period later than the philologist dates the text on the basis of internal language clues or historical references. Perhaps

100

Robert D. Holmstedt

the artifact preserves a copy of the text that was composed much earlier and preserved via a conservative scribal process. Thus, a great part of the philologist’s task is to determine the relationship of the artifact to the text. Another part of the philological task is to reconstruct any parts of the text that have suffered from, for example, scribal errors, weather, or other physical damage. The result of this activity is the “philological text” in the second row of fig. 1. The linguist uses the philological text to reconstruct a “linguistic text,” represented in the third row of fig. 1. It is from this linguistic text that the linguist deduces the grammar (in the fourth row) that produced the language “output” represented in the linguistic text. The grammar itself is constrained by grammatical principles arising from theoretical and cross-linguistic research. Thus, the grammar can in turn influence the reconstruction of the linguistic text. This relationship is represented by the bidirectional arrows between the third and fourth rows in fig. 1. Once the grammar of each text is established with reasonable confidence, the grammars may be compared with each other and, using both external and internal information, situated relative to each other chronologically. Note that text in this discussion is not the same thing as composition, as the term is often used in biblical studies. For instance, if we took the book of Ezekiel in B19a and began reconstructing the philological text, the result would not necessarily be the prophet Ezekiel’s work. Rather, the philological text should be very much like the text-critical goal of the last redaction. 2 This suggests very strongly that text critics and linguists should work much more closely than is often the case. 3 The text-critical argument is sometimes set up as an obstacle to historical linguistics, in general, and to the dating of texts, in particular. Admittedly, the 2.  This accords with Tov’s approach (2001: 288): as a rule, this branch of textual criticism aims neither at the compositions written by the biblical authors, nor at the previous oral stages, if such existed, but only at that stage of the composition which is attested in the textual evidence. Textual analysis does not aim at oral or literary stages beyond this evidence. Moreover, earlier in the book, Tov makes it clear that, even when he discusses the “single original text,” he is not referring to “the most ancient form or earliest literary strand of a biblical book nor to the earliest attested textual form, but rather to the text or edition (or a number of consecutive literary editions) that contained the finished literary product and which stood at the beginning of the process of textual transmission” (Tov 2001: 171). 3.  A related point about the Masoretic tradition is critical. If the Masoretic textual tradition has preserved linguistic elements that are no longer part of whatever Hebrew the Masoretic scribes knew, then we have historical-linguistic evidence that can be used to build profiles of linguistic change. The result in most scribal traditions is a text in which it is possible to recognize the various linguistic layers that have accreted (see, for example, Ó Buachalla 1982). The linguist’s task is therefore not merely extracting linguistic data in a naïve way but, rather, dating (relatively) the discernible layers and establishing a (relative) linguistic chronology.

Historical Linguistics and Biblical Hebrew

101

reconstructive process is challenging, whether the goal is textual or philological; but the challenge should not be exaggerated. For example, while Young and Rezetko may be right to criticize the general disconnect between Hebrew historical study and text-critical study, it does not follow that “textual stability is essential to linguistic dating of texts. . . . If texts were modified to any serious degree, then we cannot use language as a criterion for talking about the original language of biblical books or the dates of original authors” (2008: 344). If the artifacts from which philological texts are reconstructed can be placed in sequential order, then the differences between them are prime evidence for language change. Moreover, even Young and Rezetko discuss numerous examples where differences among the textual evidence show “that the language was subject to constant revision at the hands of editors and scribes who passed down the biblical tradition through many generations” (2008: 359; also see p. 351). If they can identify textual differences that stand in an obvious chronological relationship, then they have engaged in the very reconstruction work they seem to disallow. It comes down to a simple principled position: either we are limited to the historical artifact, or we are allowed to access information, whether textual or linguistic, “behind” the historical artifact by means of reconstruction. 2.2.  Language Change Now that I have briefly discussed the object of study from a philological and linguistic perspective, let us turn to the issue of language change. The nature of language change has not been discussed with necessary clarity in ancient Hebrew studies. First, the categories of “Archaic BH,” “Standard/Classical/Early BH,” and “Late BH” are not only unhelpful, they have no empirical status. As Naudé reminds Hebraists in his 2003 study of the transitions of Biblical Hebrew, it is only the formal grammar represented in the output of an individual, the I-language represented by an idiolect, that is a discrete object of scientific study (Naudé 2003: 197; see also Hale 2007: 3–18). From an I-language perspective, “change results when transmission is flawed with respect to some features. When transmission is not flawed (with respect to some feature), there has been no change in the strict sense” (Hale 2007: 36). This feature, the product of imperfect transmission in the acquisition process, spreads or becomes diffuse, when it is accurately acquired by another speaker. This view of change and diffusion has a number of implications for how we may even talk about the history of Hebrew. First, as Naudé rightly points out, the notion of a “transitional” stage between SBH and LBH is not justifiable (2003: 202). Each I-language, represented by the idiolect that is itself represented in the language of the philologically reconstructed texts of the biblical books, is its own “stage,” as it were. The idea underlying this approach can be expressed by borrowing and modifying the dialectological dictum slightly:

102

Robert D. Holmstedt

every change and its resulting diffusion (if it becomes diffuse) has its own history. It is unlikely, therefore, that any two change-and-diffusion features will have the same origin. It is also unlikely that any two I-languages will reflect the same cluster of change-and-diffusion features, which implies that the exact order of texts may vary for each feature analyzed. But no single feature set can be determinative for a relative order, since the texts (or, the I-languages represented within the texts) do not stand in a two-dimensional line; rather, since each I-language is a unique constellation of features, some I-languages/ texts will stand to the “left” or “right” of any two-dimensional line of descent. (For further discussion of statistical analysis and relative dating with regard to Biblical Hebrew, see Andersen and Forbes 1986.) Two more features of change-and-diffusion are critical before a coherent picture emerges, from which we may derive an analytical framework for Hebrew. First, it necessarily follows from the acquisition-related change-and-diffusion framework that a new form representing the acquisition and diffusion of a given change will coexist with the older form within the speech community, perhaps for many generations, such as in the case of a dialect of an elderly or immigrant community coexisting with their children’s and grandchildren’s dialects. In other words, it may take generations for a single change to become diffuse throughout an entire speech community (and thus to be reflected in every text coming from that community). Indeed, this precise pattern has been observed many times over. As Wolfram and Schilling-Estes point out: Speakers do not suddenly adopt a new form as a categorical replacement for an older form, whether the form involves a gradual imperceptible change in the phonetic value of a vowel within a continuum of phonetic space or an abrupt, readily perceptible change involving the metathesis of consonants or the linear realignment of constituents within a syntactic phrase. Instead, there is a period of variation and coexistence between new and old forms in the process of change. This transitional period of fluctuation has often been ignored in historical linguistics under the assumption that language change cannot be directly observed. (Wolfram and Schilling-Estes 2003: 715–16)

Second, the historical linguist Charles Bailey observed that the diffusion of changes over time follows a Sigmoid, or “S”-shaped, curve: A given change begins quite gradually; after reaching a certain point (say, twenty per cent), it picks up momentum and proceeds at a much faster rate; and finally tails off slowly before reaching completion. The result is an S-curve: the statistical differences among isolects in the middle relative times of the change will be greater than the statistical differences among the early and late isolects. (Bailey 1973: 77; see also Kroch 1989; Pintzuk 2003)

An idealized example of such an S-curve is provided in fig. 2.

Historical Linguistics and Biblical Hebrew

103

Figure 2.  The S-curve of diffusion.

Although the S-curve has not been invoked in ancient Hebrew studies, it is quite likely that, if the features commonly cited were plotted along the dimensions of time and frequency, an S-curve such as the one above would emerge, and the relative order of books would match the general order achieved in more traditional analyses. Thus, the common refrain from the challengers that “the existence of a supposed late feature in a supposedly earlier text invalidates the entire approach” falls on its face for lack of linguistic awareness. Again, it is simply a fact that old and new forms do coexist, often for hundreds of years.  4 What is needed is a reevaluation of all previous results in which the features are each plotted separately and the results are overlaid, increasing with each additional layer the statistical probability of the accuracy of the synthesis.  5 Now, to return to the claim at the outset of this section: the reorientation required concerns the traditional categories of “Archaic BH,” “Standard/Classical/ Early BH,” and “Late BH.” The change-and-diffusion framework, with the S-curve describing the temporal path of diffusion, suggests that any categories such as “Standard” or “Late” are at best conveniences and at worst 4.  A commonly cited syntactic example is the development of ‘do’-support in Middle English, in which ‘do’ appears as an auxiliary verb (or better, as the finite verb carrying the bundle of inflectional features) in questions (‘do you want?’), clauses with an initial adverb (‘rarely did they want’), and other restricted environments. This development began in a restricted environment and then spread to other contexts. Moreover, non-‘do’-support clauses coexisted with the newer construction for over 300 years, until finally being replaced entirely by the ‘do’-support construction (see Lightfoot 1979; Kroch 1989; 2001; and Dresher in this volume). 5.  DeCaen makes a similar argument about the use of statistics on historical Hebrew analysis: “One form or one contrast yields precious little, but all possible variants statistically correlated should yield much” (DeCaen 2001: 23).

104

Robert D. Holmstedt

inaccurate and misleading. These labels represent generalizations that obscure the presence of numerous cases of change-and-diffusion within the corpus that the label supposedly covers. Changes, and thus the dialects within which they occur, can only be described as “earlier” or “later” relative to each other. Considering two categories of language change, with examples, will help to clarify the diachronic discussion: exogenous, or externally-motivated changes, and endogenous, or internally-motivated changes. Externally-motivated changes are due to language contact, such as borrowing or interference due to acquisition of a second language (see Thomason 2003). Internally-motivated changes are due to the nature of language acquisition itself: the “perfect” transmission of one parent’s grammar to a child is precluded by the reality of the output-input situation. 2.2.1.  Exogenous Change The most commonly studied type of exogenous change, at least for ancient languages, is borrowing. This is certainly true for Hebrew studies: there are numerous works devoted to the topic and any biblical commentary worth its salt discusses the issue at some point. But, while identifying borrowed words in biblical books and discussing the possible implications for the temporal, geographic, or social origin of the book in question is almost commonplace, the types of borrowing and its motivation are rarely discussed. Thus, it is worth considering the nature of linguistic borrowing and the closely related phenomenon of code-switching. Languages borrow words from other languages for two primary reasons: need and prestige (Campbell 2004: 64–65). 6 The word ‘coffee’ is a good modern example of need-based borrowing: European languages borrowed the word from Arabic through Turkish. 7 A likely example of need-based borrow6.  A third category of borrowing that relates to Hebrew is loan translations or “calques.” In contrast to borrowed words, for which something of the borrowed item’s phonetic shape and meaning continue into the borrowing language, loan translations use native words to translate a borrowed concept. Thus, Modern English ‘gospel’ is derived from Old English gód spel ‘good tidings’, which was a translation, through Latin evangelium, of the Greek εὐαγγέλιον ‘good tidings’. A possible Hebrew example is the use of ‫ ְּבצֵל‬in Qoh 7:12 (see also 6:12, 8:13; see Wise 1990 for the full argument). In this verse, instead of the normal Hebrew meaning ‘in the shadow (of)’, ‫ ְּבצֵל‬may be a loan translation of the cognate Aramaic ‫בטלל‬, which, unlike the Hebrew phrase, went through a series of semantic shifts: ‘in the shadow (of)’ > ‘with the help (of)’ > ‘because (of)’. Thus, in Qoheleth, the Hebrew phrase ‫ ְּבצֵל‬has neither the normal denotation of ‘in the shadow (of)’ nor the metaphorical meaning ‘in the protection (of)’ but is a loan translation of what the cognate Aramaic phrase had become, ‘because’. 7.  The etymological entry in the Oxford English Dictionary explains its background and notes its diffusion as a borrowed word: Arab. qahwah, in Turkish pronounced kahveh, the name of the infusion or beverage; said by Arab lexicographers to have originally meant ‘wine’ or some kind of

Historical Linguistics and Biblical Hebrew

105

ing in Hebrew is the word ‫‘ קֹוף‬monkey’. Since monkeys were not native to ancient Israel, it is understandable that Hebrew had no native word for the animal. However, when the need arose to mention this particular animal (1 Kgs 10:22 // 2 Chr 9:21), the word was borrowed into Hebrew, perhaps via Egyptian (although its origin may be Sanskrit kapi; see HALOT, s.v.). Few examples of need-based borrowing figure into the reconstruction of Hebrew diachrony. 8 It is prestige-based borrowing that figures prominently in ancient Hebrew studies. Prestige-based borrowing reflects a sociolinguistic situation in which a foreign language, whether closely related or not, is associated with higher social or political status or is simply a dominant linguistic cultural influence (for example, a lingua franca). For example, during the Norman French dominance in England (1066–1300), many French words were borrowed into English (for example, ‘pork’ > Fr. porc) even though English already had serviceable terms (for example, pig meat; Campbell 2004: 64). The reason was that at that time French was considered more prestigious than English. In Hebrew, prestige borrowing is often invoked to explain the increasing number of Aramaisms (for example, ‫‘ ְזמָן‬time’) as well as the few Persianisms (for example, ‫‘ ּפ ְַרּדֵ ס‬garden’, ‘royal enclosure’) found in some biblical texts. The prestige status for Aramaic came from its role as the administrative language of both the Neo-Babylonian and Persian empires; for Persian, the prestige status no doubt derived from the political dominance of the Persians from the sixth to fourth centuries b.c.e. Whether words are borrowed due to need or prestige, it is important to recognize that the borrowed item is normally adapted and accommodated to the borrowing language’s phonology and morphology (see Campbell 2004: 65–69). For instance, although the Hebrew ‫ַּׁשף‬ ָ ‫‘‏א‬conjurer’ entered either via Aramaic ‫ אָשַׁ ף‬or Akkadian (w)āšipu, the Hebrew word is the only version of the word that reflects gemination of the root’s middle consonant, which is wine, and to be a derivative of a vb.-root qahiya ‘to have no appetite.’ Some have conjectured that it is a foreign, perh. African, word disguised, and have thought it connected with the name of Kaffa in the south Abyssinian highlands, where the plant appears to be native. But of this there is no evidence, and the name qahwah is not given to the berry or plant, which is called bunn, the native name in Shoa being būn. The European langs. generally appear to have got the name from Turkish kahveh, about 1600, perh. through It. caffè; cf. F., Sp., Pg. café, Ger. ka­fee, Da., Sw. kaffe. The Eng. coffee, Du. koffie, earlier Ger. coffee, koffee, Russ. kophe, kopheì, have o, app. representing earlier au from ahw or ahv. 8.  Note that personal names are a special case since they do not necessarily have a continued presence in the language. In fact, while they are often adapted to the phonology and morphology of the recipient language, personal names used of specific individuals within a ְ ‫ א ְַר‬in Gen 14:1, 9; contra Young and Rezetko 2008: 1.307–8) are argutext (for example, ‫יֹוך‬ ably not cases of borrowing because they may simply match the individual’s foreign-origin name. This is in contrast to, say, a personal name that becomes a common name within the speech community, such as ‫מ ְריָם‬, ִ the name of Moses’ sister.

106

Robert D. Holmstedt

likely because the word was imported as a qattāl-pattern noun, the nominal morphological category used for ‘nouns of profession’ (Joüon and Muraoka 2006: §88Ha and n. 49; also §87d). The problem that adaptation and accommodation raise is the problem of identification: most cases of discernible borrowings in Hebrew come from other Semitic languages, which share a similar phoneme inventory and the triradical root morphology. For all practical purposes, then, we should begin by limiting ourselves to words the shape of which falls outside the paradigmatic margins. As Campbell suggests, “Words which violate the typical phonological patterns (canonical forms, morpheme structure, syllable structure, phonotactics) of a language are likely to be loans” (Campbell 2004: 70). An example of a borrowed word that has been phonologically adapted but must have been borrowed is ‫ּכ ָתב‬. ְ The vocalization tradition for the word indicates that it was not affected by the Canaanite shift (ca. 1400 b.c.e.) in which ā > ō. If it had been affected, it would be vocalized like ‫* ksy/ksh, Piel) is due to the shortening of a geminated consonant before a šēwa (see Joüon 1993: §18m). 9.  It is possible to see the final -ω in σεννημω in the Secunda at Ps 35:16 as an abnormal representation of -ū, similar to what we have in ιαμουθω (Ps 49:11, MT ‫‘ יָמּותּו‬they will die’), θαμωγ (Ps 46:7, MT ‫‘ ָּתמּוג‬she wavers’), and βωσα (Ps 89:46, MT ‫ּבּוׁשה‬ ָ ‘shame’)—compare Brønno 1943: 363 and Janssens 1982: 126–27. Then, if we interpret the final -ū in MT ‫ְיכ ְַסיֻמּו‬ (Exod 15:5) as resulting from assimilation of the final vowel to the 3mp verbal suffix -ū, we may suppose that the Secunda reflects a Hebrew dialect in which the 3mp suffixed pronoun -mw was normally vocalized -mū, whereas the MT reflects a dialect in which this form was normally vocalized -mō (the assumption of a dialectal difference in this regard between the Secunda and the MT was voiced earlier on by Brønno [1943: 202] and mentioned as a possibility by Joüon [1993: §61i]). In contradistinction, it would be unjustified to assume that the final -ου in λαμου in the Secunda at Ps 28:8; 49:14 and in βεταμου in the Secunda at Ps 49:12 is an abnormal representation of -ō, because the only clear instance of -ου in the Secunda reflecting historical -ō is ζερουωθαϊ in Ps 18:35 (MT ‫‘ זְרֹוע ָֹתי‬my arms’), where the form *zĕrūʿōtāy could have developed by vowel dissimilation (Brønno 1943: 170–71), and it is impossible to posit a development of the same kind for the forms λαμου and βεταμου.

152

Yigal Bloch

would have been part of the regular phenomenon of the elision of final short vowels in Hebrew (Robertson 1972: 66). However, it is difficult to derive the form -mū reflected in the Secunda from the postulated original *-humā. Moreover, it is not at all clear whether the form -humā of the 3du suffixed pronoun, known from Classical Arabic, ever existed in the linguistic stratum out of which Hebrew developed as a distinct language. A recent study of dual pronouns in the Semitic languages by Elitzur A. Bar-Asher (2009: 32–37) suggests that the Proto-Semitic form of the 3du suffixed pronoun was *-šumī or *-šumay, whereas the form of the 3du independent pronoun in Proto-Semitic was *šumā. According to this reconstruction, during the development of the attested Semitic languages, the vocalic endings of these forms were leveled, so that for all the languages for which the vocalization of 3du suffixes is known, the independent and the suffixed 3du pronouns end in identical vowels—as in Classical Arabic, with independent 3du humā vs. suffixed 3du -humā, and in the earliest stages of Akkadian (prior to ca. 1900 b.c.e.), with independent 3du šunī(ti) vs. suffixed 3du -šunī(ti) in the genitive and the accusative and -šunīšim in the dative. 10 Now, it is reasonable to suppose that the leveling of the final vowel of the 3du pronouns took place also in the Northwest Semitic linguistic milieu of the late second millennium b.c.e. out of which the Hebrew language had developed. But in the absence of direct indications of the vocalization of the 3du pronouns in Northwest Semitic dialects of the relevant period, it is impossible to tell whether the final vowel of those pronouns was -ā or -ī (see Tropper 2000: 228). If anything, the fact that the El-Amarna letters from Canaan employ the form šunī for the 3du independent pronoun and the form -šunī for the 3du suffixed pronoun in the genitive and the accusative, 11 centuries after these forms went out of use in Akkadian that was written in its Mesopotamian homeland (Moran 1973; Rainey 1996a: 66, 83–85), suggests that the 3du pronouns in Northwest Semitic dialects of the late second millennium b.c.e. ended in -ī rather than in -ā. However, inference of this sort is admittedly conjectural. Returning to the 3mp suffixed pronoun -mw in Biblical Hebrew: once Robertson’s proposal to derive it from the 3du suffixed pronoun *-humā is recognized as problematic, it appears most reasonable to accept the view of Bauer and Leander (1922: §§21j, 29pʹ), according to which all forms of the 3mp suffixed pronoun in Biblical Hebrew developed from the original 3mp suffixed 10. For the studies that identified the forms of 3du suffixed pronouns in the earliest stages of Akkadian, see Whiting 1972 and Hasselbach 2005: 149–50, 157–58. The forms humā and -humā of 3du pronouns in Arabic result from the general shift *š > h, which took place outside the verbal and nominal roots and outside the prefix of the t-infixed causative stem in most attested West Semitic languages, or in the linguistic strata out of which those languages developed (see Voigt 1988). 11.  Only the genitive/accusative form of the 3du suffixed pronoun (as opposed to the dative form) is attested in the El-Amarna letters from Canaan.

The Third-Person Masculine Plural Suffixed Pronoun -mw

153

pronoun *-humū̆, where the final vowel was an anceps—that is, a vowel articulated sometimes as short and sometimes as long. According to Bauer and Leander, the common forms of 3mp suffixed pronouns in Biblical Hebrew: ‫ם‬-, ‫הֶם‬-,and -‫הֵם‬, developed from the form *-humū̆, in which the final anceps vowel was perceived as short and elided, whereas the form -‫ מּו‬developed from the original *-humū̆, with the final anceps vowel perceived as long (*-humū > -mū, elision of h). The form -‫ מֹו‬developed under similar conditions but with dissimilation of the final vowel (*-humū > *-humō > -mō). 12 The above-mentioned Canaanite forms ta-aḫ-ta-mu and ma-aḫ-sí-ra-mu in El-Amarna letters from Shechem and Jerusalem testify that a form of the 3mp suffixed pronoun with the elision of h and of the vowel following it had already developed by the fourteenth century b.c.e., but the cuneiform spelling does not allow one to tell whether the final vowel of this form was u, ū, or ō. The proposal of Bauer and Leander assumes that the forms ‫הֵם‬-/‫הֶם‬- of the 3mp suffixed pronoun, which did not undergo elision of the consonant h, experienced the phonological influence of the 3fp suffixed pronoun in the vowel following h: 3mp *-humū̆ > ‫הֵם‬-/‫הֶם‬- (and not *‫הֻם‬-), similar to 3fp *-hinnā̆ > ‫הֵן‬-/‫הֶן‬- (Bauer and Leander 1922: §29pʹ). 13 The influence of the 3fp suffixed pronoun can also explain why, in some forms of the 3mp suffixed pronoun that have preserved the final vowel, this vowel appears as ā—for example, ‫‘ ֵאלֵי ֵהמָה‬their (m.) pillars’ (Ezek 40:16); ‫‘ ָּב ֵהּמָה‬in/by them (m.)’ (Exod 30:4; 36:1; Hab 1:16); ‫‘ ָּכ ֵהּמָה‬like them (m.)’ (Jer 36:32); ‫‘ ָל ֵהּמָה‬to / for them (m.)’ (Jer 14:16); ‫‘ ֵמ ֵהּמָה‬from them (m.)’ (Jer 10:2; Qoh 12:12). In the forms ‫ ֵהמָה‬‎-/ ‫ ֵהּמָה‬‎-, not only the vowel following h but also the final vowel appear to have been influenced by the final anceps vowel of the 3fp suffixed pronoun *-hinnā̆ (perceived in this case as long). 14 The same influence of the final vowel of the 3fp suffixed pronoun appears to underlie the forms ‫מה‬- and ‫המה‬- of the 3mp suffixed pronoun, which are attested frequently in Qumran Hebrew (see Qimron 1986: 62). 15 12.  It should be noted that, although Bauer and Leander postulated *-humū̆ as the ProtoSemitic form of the 3mp suffixed pronoun, the Proto-Semitic form should probably be reconstructed as *-šumū̆, assuming that the shift *š > h in the initial consonant of this form took place after the Proto-Semitic stage (see above, n. 10). 13.  Like *-humū̆, the form *-hinnā̆ was probably a development out of Proto-Semitic *-šinnā̆ (see n. 12). 14.  The final vowel of the original 3fp suffixed pronoun *-hinnā̆ is preserved in some forms of the 3fp suffixed pronoun in the Hebrew Bible: ‫ק ְר ֶּבנָה‬-‫ֶל‬ ִ ‫‘ א‬into them (f.)’ (Gen 41:21 [2×]); ‫‘ ְלי ְַח ֵמּנָה‬to arouse them (f.) sexually’ (Gen 30:41), ‫‘ ּגְִוּיֹתֵ י ֶהנָה‬their (f.) bodies’ (Ezek 1:11); ‫‘ ָב ֵהּנָה‬in/by them (f.)’ (Lev 5:22; Num 13:19; Jer 5:17); ‫‘ ָ ּכ ֵהּנָה‬like them (f.)’ (Gen 41:19; 2 Sam 12:8 [2×]; Job 23:14); ‫‘ ָל ֵהּנָה‬to / for them (f.)’ (Ezek 1:5, 23 [2×]; 42:9; Zech 5:9); ‫‘ ֵמ ֵהּנָה‬from/out of them (f.)’ (Lev 4:2; Isa 34:16; Jer 5:6; Ezek 16:51, 42:5; Ps 34:21; 1 Chr 21:10). 15.  The forms of the 3mp suffixed pronoun ‫ ֵהּמָה‬-/‫ ֵהמָה‬- and ‫מה‬- make it difficult to accept the proposal of Fassberg (2009) that the vowel in the 3mp suffixed pronoun ‫הֵם‬-/‫הֶם‬-

154

Yigal Bloch

Based on the above, it is possible to classify the 3mp suffixed pronoun -mw (‫מּו‬-/‫מֹו‬-) as typologically archaic, given that this form reflects a stage of development of the Hebrew language when the 3mp suffixed pronoun was still unaffected by the phonological influence of the 3fp suffixed pronoun. 3.  Archaism or Archaization? The Distribution of the 3mp Suffixed Pronoun -mw in the Hebrew Bible Establishing the typologically archaic character of a given linguistic feature in Biblical Hebrew is not sufficient to conclude that the literary units exhibiting this feature were composed significantly earlier than the bulk of the material included in the Hebrew Bible. It is possible that a typologically archaic feature continued to exist in the Hebrew language down to the end of the biblical period alongside features reflecting a more advanced stage of linguistic development. In order to determine that a typologically archaic linguistic feature is significant for dating some literary units in the Hebrew Bible to the thirteenth–tenth centuries b.c.e., one must demonstrate that the feature in question does not appear in literary works that can be dated on nonlinguistic grounds to a later period, or that the use of this feature in late literary units is only an artificial attempt at archaization that fails to comply with its original use, grammatically or semantically. Robertson was aware of this requirement, and set out to delineate criteria that would distinguish between the truly archaic uses of the 3mp suffixed pronoun -mw and the uses that continued well into the first millennium b.c.e. To this end, Robertson distinguished between the uses of the form -mw with nouns and verbal forms, on the one hand, and with prepositions, on the other hand. The use of the 3mp suffixed pronoun -mw with nouns and verbal forms was not attested, according to Robertson, in what he termed “standard poetic Hebrew”—the language of the poetic works preserved in the Hebrew Bible whose dates of composition can be established to fall in the eighth century b.c.e. or later (Robertson 1972: 65–66). developed without the influence of the 3fp suffixed pronoun, by an internal vowel dissimilation: *-humū̆ > *-himū̆ > *-him (final anceps vowel perceived as short and elided) > ‫הֵם‬-/‫הֶם‬-. The development proposed by Fassberg does not account, in itself, for the final vowel ā in the forms ‫ ֵהּמָה‬-/‫ ֵהמָה‬-, and ‫מה‬-. The independent 3mp personal pronoun ‫ ֵהּמָה‬could derive from the oblique 3mp personal pronoun hmt, attested in Ugaritic and Phoenician (Fassberg 2009: 328 n. 8), but no suffixed pronoun *-hmt is attested in Northwest Semitic languages. Perhaps the 3mp suffixed pronouns ‫ ֵהּמָה‬-/‫ ֵהמָה‬- may be explained as having developed by analogy with the 3mp independent pronoun ‫ ֵהּמָה‬. But first, this would yield an overly complicated explanation for those suffixed forms (involving first the dissimilation of the internal vowel, then the change of the final vowel by analogy with the 3mp independent pronoun, rather than a single change of vowels in the suffixed forms by analogy with the 3fp suffixed pronoun *-hinnā̆). And second, for the Qumran Hebrew form of the 3mp suffixed pronoun ‫מה‬-, analogy with an independent 3mp personal pronoun ‫ ֵהּמָה‬would be less than straightforward.

The Third-Person Masculine Plural Suffixed Pronoun -mw

155

With regard to prepositions, the use of the 3mp suffixed pronoun -mw with the preposition -‫‘ ל‬to / for’ in the form ‫ לָמֹו‬is attested in works dated to the sixth century b.c.e.: Lamentations (Lam 1:19, 22; 4:10, 15) and Deutero-Isaiah (Isa 43:8; 44:7, 15; 48:21; 53:8). Obviously, the use of the form ‫ לָמֹו‬cannot be considered evidence of an early dating of the poetic units in which it occurs. However, the use of the 3mp suffixed pronoun -mw with other prepositions is not attested, according to Robertson, in poetic works dating to the eighth century b.c.e. and later. For this reason, these forms would constitute valid evidence for dating the poems in which they occur to the thirteenth–tenth centuries b.c.e. (Robertson 1972: 68–69). In fact, a significant concentration of examples of the suffixed pronoun -mw with a preposition other than -‫ ל‬occurs only in the poetic part of the book of Job, where the preposition ‫‘ עַל‬upon’ appears with this suffixed pronoun, in the form ‫ ָעלֵימֹו‬, eight times (Job 6:16; 20:23; 21:17; 22:2; 27:23; 29:22; 30:2, 5). This concentration of the suffixed pronoun -mw with the preposition ‫ עַל‬in Job, together with the single attestation of the same suffixed pronoun with a noun in this book—namely, ‫( ַכּפֵימֹו‬Job 27:23)—moved Robertson to argue that the poetic part of Job is to be dated to the eleventh–tenth centuries b.c.e. along with other poems considered to belong to the earliest stratum of Biblical Hebrew poetry (Robertson 1972: 68–69, 146, 153–55). 16 However, since the book of Job is usually dated by biblical scholars to the Second Temple period (late sixth century b.c.e. or later), Young and Rezetko argued that Robertson’s conclusion concerning the very early date of the poetic section of Job exemplifies the problematic nature of an attempt to date Biblical Hebrew poetry by linguistic criteria (Young and Rezetko 2008: 333–34). 17 What is missing from Robertson’s discussion as well as Young and Rezetko’s is close attention to the specific uses of the suffixed pronoun -mw in the poetic section of Job. In several instances in which the suffixed pronoun -mw is used, the context of the relevant verses indicates clearly that the referent of this suffixed pronoun is in fact singular, not plural. This is true of the form ‫ַכּפֵימֹו‬ 16.  It should be noted that, in order to arrive at such an early date for the poetic part of Job, Robertson also used other typologically archaic linguistic features appearing in this literary work, including the use of (short) prefixed verbal forms without the conjunction w- to signify complete situations in the past. However, with regard to most prefixed verbal forms used without w- in the poetic section of Job, Robertson admitted that these forms can in fact be analyzed as imperfective (“frequentative,” in his terminology—see Robertson 1972: 40), and hence, we must add, belonging to the long prefix-conjugation *yaqṭulu. Moreover, as will be argued immediately, the use of the 3mp suffixed pronoun -mw in the poetic section of Job suggests that this literary work is to be dated late, most likely to the Second Temple period, and not early. 17.  A number of recent studies unmentioned by Young and Rezetko have made a convincing case for a late dating of Job. One should note especially Knauf 1988; Halpern 2002; and Greenstein 2003: 652 n. 6.

156

Yigal Bloch

‘his (sic) hands’ in Job 27:23 and of the form ‫‘ ָעלֵימֹו‬upon him (sic)’ in Job 20:23; 22:2; 27:23 (GKC §103g, n. 3). These instances suggest that the use of the suffixed pronoun -mw by the author of the poetic section of Job was nothing but an artificial, and sometimes misguided, attempt at archaization made when the original semantic content of this form (the 3mp suffixed pronoun) was already partly forgotten by Hebrew speakers and literati. In any event, even the use of the form ‫ ָעלֵימֹו‬with a 3mp referent of the pronominal suffix cannot be interpreted as evidence for an extremely early date of the literary units in which it occurs. A Qumran pesher to the book of Isaiah, 4Q163, quoting Isa 14:8, spells ‫‘ עלימו‬upon them (m.)’ instead of the MT ‫ָעלֵינּו‬ ‘upon us’ (Allegro 1968: 19). If the text quoted by 4Q163 is more original than the MT, 18 it indicates that the form ‫ ָעלֵימֹו‬could be used at the time when the prophetic speeches of Isaiah were composed—that is, in the late eighth century b.c.e.; and if the text quoted by 4Q163 represents a scribal innovation, it indicates that the form ‫ ָעלֵימֹו‬could still be used in the Hellenistic–Roman period. Thus, the only usages of the 3mp suffixed pronoun -mw that may indicate that the poetic units in which they occur were composed in a very early period are the usages in which the pronoun is suffixed to nouns or verbs. 19 In all the poetic works attributed by Albright and his followers to the thirteenth–tenth centuries b.c.e., these forms are attested only in Exodus 15 (9×, with verbal forms only), Deuteronomy 32 (4× with nouns) and Deuteronomy 33 (1× with a noun). 20 None of these three poems uses the form -mw to the exclusion of other forms of the 3mp suffixed pronoun. Exod 15 and Deut 33 each use the form 18.  This possibility is supported, to some extent, by the biblical scroll 4QIsae, which reads ‫‘ עליהם‬upon them (m.)’ instead of the MT ‫ ָעלֵינּו‬in Isa 14:8 (Skehan and Ulrich 1997: 96). It is plausible that the scribe of 4QIsae or his predecessor, having found the typologically archaic form ‫ עלימו‬in his source text, replaced it with the more standard ‫עליהם‬. However, all the other textual witnesses to Isa 14:8, including the Great Isaiah Scroll (1QIsaa), support the MT version, ‫ ָעלֵינּו‬. 19.  The only instance of the 3mp suffixed pronoun -mw used with a preposition other than -‫ ל‬or ‫ עַל‬is the preposition ‫‘ ֵאלֵימֹו‬to them (m.)’ in Ps 2:5. The significance of Psalm 2 for clarifying the implications of the use of the 3mp suffixed pronoun -mw for dating the works of biblical poetry is explored in §4 below. 20.  Exodus 15: ‫‘ ְיכ ְַסיֻמּו‬they covered them (m.)’ (Exod 15:5); ‫אכלֵמֹו‬ ְ ֹ ‫‘ י‬he ate them (m.) up’ (v. 7); ‫‘ ִּת ְמ ָלאֵמֹו‬let her be sated with them (m.)’; ‫ּתֹוריׁשֵמֹו‬ ִ ‘let her dispossess them (m.)’ (v. 9); ‫‘ ִּכּסָמֹו‬he covered them (m.)’ (v. 10); ‫‘ ִּת ְב ָלעֵמֹו‬she swallowed them (m.) up’ (v. 12); ‫חז ֵמֹו‬ ֲ ‫‘ יֹא‬he seized them (m.)’ (v. 15); ‫‘ ְּת ִבאֵמֹו‬you (2ms) brought them (m.)’; ‫‘ ְו ִת ָּטעֵמֹו‬and you (2ms) planted them (m.)’ (v. 17). The last verbal form appears originally to have been part of a sequential construction expressing a complete situation in the past; and. following the rules of Tiberian Masoretic vocalization, this construction should be vocalized *‫( ו ִַת ָּּטעֵמֹו‬see Bloch 2009: 54–56, esp. n. 83). Deuteronomy 32: ‫‘ ָצרֵימֹו‬their (m.) enemies’ (Deut 32:27); ‫‘ עֲנָבֵמֹו‬their (m.) grapes’ (v. 32); ‫‘ אֱלֹהֵימֹו‬their (m.) gods’ (v. 37); ‫‘ ְז ָבחֵימֹו‬their (m.) sacrifices’ (v. 38). In addition, Deuteronomy 32 uses the 3mp suffixed pronoun -mw 3× with prepositions: ‫‘ ָעלֵימֹו‬upon them’ (v. 23); ‫‘ לָמֹו‬to / for them (m.)’ (vv. 32, 35).

The Third-Person Masculine Plural Suffixed Pronoun -mw

157

‫הֶם‬- with a preposition: ‫‘ עֲלֵיהֶם‬upon them (m.)’ (Exod 15:16), ‫‘ ָּבהֶם‬with them (m.)’ (Deut 33:17). Deut 32 copiously uses both the form ‫ם‬- and the form ‫הֶם‬-, with verbal forms, nouns, and prepositions. 21 It has been demonstrated above (section 2) that the forms ‫הֶם‬-/‫הֵם‬- of the 3mp suffixed pronoun are typologically later than the form -mw, having undergone a vowel change under the influence of the 3fp suffixed pronoun (‫הֶן‬-/‫הֵן‬-). 22 Hence, the fact that the form -mw always co-occurs in Biblical Hebrew poetry with the form ‫הֶם‬-/‫הֵם‬- means that during the historical period when the biblical poems were composed, phonological influence of the 3fp suffixed pronoun on the 3mp suffixed pronoun had already taken place, even though the form -mw, which is free of such influence and hence had presumably originated at an earlier stage in the history of the Hebrew language, was retained to a larger or a smaller degree. 4.  Why the 3mp Suffixed Pronoun -mw Does Not Indicate Composition in the Second Millennium b.c.e.: The Evidence of Psalm 2 The crucial question concerning the chronological significance of the use of the 3mp suffixed pronoun -mw with nouns and verbal forms is whether this use is attested in at least one poetic unit preserved in the Hebrew Bible whose date of composition can be established as no earlier than the eighth century b.c.e. If the answer to this question is affirmative, it will demonstrate that, contrary to Robertson’s conclusions, the use of the 3mp suffixed pronoun -mw with nouns and verbal forms continued at least through the eighth century b.c.e., Deuteronomy 33: ‫‘ ּבָמֹותֵ ימֹו‬their (m.) backs’ (Deut 33:29). In addition, Deuteronomy 33 uses the 3mp suffixed pronoun -mw twice with a preposition: ‫( לָמֹו‬Deut 33:2 [2×]). However, both occurrences of this form in Deut 33:2 are textually dubious (see BHS ad loc.). 21.  The form ‫ם‬-: ‫‘ מּומָם‬their (m.) blemish’ (Deut 32:5); ‫יתם‬ ָ ‫ֲר‬ ִ‫‘ ַאח‬their (m.) future’ (vv. 20, 29); ‫‘ ּבָם‬in/by them (m.)’ (vv. 20, 23, 24); ‫‘ א ְַקנִיאֵם‬I shall provoke them (m.) to jealousy’; ‫‘ א ְַכ ִעיסֵם‬I shall arouse their (m.) anger’ (v. 21); ‫ִכרָם‬ ְ ‫‘ ז‬the memory of them (m.)’ (v. 26); ‫צּורם‬ ָ ‘their (m.) Rock’ (vv. 30, 31); ‫‘ ְמ ָכרָם‬he has sold them (m.) out’; ‫‘ ִה ְס ִּגירָם‬he has handed them (m.) over’ (v. 30); ‫‘ ּג ְַפנָם‬their (m.) vine’ (v. 32); ‫‘ י ֵינָם‬their (m.) wine’ (v. 33); ‫‘ רַ ְגלָם‬their (m.) foot’ (v. 35); ‫ְסיכָם‬ ִ ‫‘ נ‬their (m.) libation’ (v. 38). The form ‫הֶם‬-: ‫‘ ֵמהֶם‬from them (m.)’ (v. 20); ‫‘ ְּבה ְַבלֵיהֶם‬with their (m.) vanities’ (v. 21); ‫‘ א ְַפאֵיהֶם‬I shall bring an end to them (m.)’ (v. 26); ‫‘ ָּבהֶם‬in them (m.)’ (v. 28). 22.  This runs contrary to Robertson’s assessment that the 3mp suffixed pronoun ‫הֶם‬- was inherited by Hebrew from the Canaanite linguistic milieu of the second millennium b.c.e. Robertson’s view on this matter was based on the appearance of the 3mp suffixed pronoun -hm in Ugaritic (Robertson 1972: 66). However, the crucial question of the vocalization of the suffixed pronoun -hm in Ugaritic was not considered by Robertson. It should be noted that, while the use of the 3mp suffixed pronoun -mw with prepositions cannot be diagnostic of an early dating of the literary units in which it occurs, the use of the 3mp suffixed pronoun ‫הֵם‬-/‫הֶם‬- with any part of speech does demonstrate that the phonological influence of the 3fp suffixed pronoun on the 3mp suffixed pronoun had already taken place by the time of the composition of the literary units in which the use is attested. In this regard, there is no reason to distinguish between the use of the 3mp suffixed pronoun ‫הֵם‬-/‫הֶם‬- with nouns, verbal forms, or prepositions.

158

Yigal Bloch

and hence their use in Exodus 15 and Deuteronomy 32–33 is not sufficient reason to date these poems to the thirteenth–tenth centuries b.c.e. Psalm 2 contains the 3mp suffixed pronoun -mw on two nouns: ‫מֹוסרֹותֵ ימֹו‬ ְ ‘their (m.) bonds’, and ‫‘ עֲבֹתֵ ימֹו‬their (m.) twisted ropes’ (v. 3); and one verbal form: ‫הלֵמֹו‬ ֲ ‫‘ ְי ַב‬he frightens them (m.)’ (v. 5). In addition, the suffixed pronoun -mw is used in Psalm 2 on two prepositions: ‫‘ לָמֹו‬at them (m.)’ (v. 4), and ‫ֵאלֵימֹו‬ ‘to them (m.)’ (v. 5). Thus, with regard to the parts of speech to which the 3mp suffixed pronoun -mw is appended, the usage in Psalm 2 is more comprehensive than in Exodus 15 (where the form -mw is not used on prepositions) and in Deuteronomy 32 (where the form -‫ מֹו‬is not used on verbal forms). 23 In contrast, Psalm 2 uses the 3mp suffixed pronoun ‫ם‬- on two verbal forms: ‫ְּתרֹעֵם‬ ‘you (ms) shall smash them (m.)’, and ‫‘ ְּתנ ְַּפצֵם‬you (ms) shall break them (m.) in pieces’ (v. 9). But, since the form ‫ם‬- of the 3mp suffixed pronoun developed out of the original form *-humū̆ with the elision of h, the vowel following h, and the final anceps vowel (perceived in this case as short), we cannot determine whether the development of this form had undergone the influence of the 3fp suffixed pronoun, such as is evident in the forms ‫הֵם‬-/‫הֶם‬-. Hence, although the 3mp suffixed pronoun ‫ם‬- is much more common in Biblical Hebrew than -mw, it cannot be considered typologically less archaic than the latter. Thus, in qualitative terms, the use of the 3mp suffixed pronoun -mw in Psalm 2 is no less archaic than its use in Exodus 15 or Deuteronomy 32–33. The numerical prevalence of the forms -mw in Exodus 15 and Deuteronomy 32 over those in Psalm 2 does not outweigh this observation. Now, when was Psalm 2 composed? Given the explicit mention of Zion in Ps 2:6 and the absence of any linguistic feature or content that could be reasonably construed as linking Psalm 2 with the Northern Kingdom of Israel, the Judahite origin of this psalm seems assured. Then, since the content of Psalm 2 marks it as a work of royal propaganda, glorifying the king installed in Zion as the son of God, it is reasonable to infer that this psalm was composed before the destruction of the Kingdom of Judah in 586 b.c.e. Concerning the upper limit of the period of time during which Psalm 2 was likely composed, several details suggest that this psalm was composed later than ca. 750 b.c.e. First, in Ps 2:9, ‫ׁשבֶט ַּב ְרזֶל ִּכ ְכ ִלי יֹוצֵר ְּתנ ְַּפצֵם‬ ֵ ‫‘ ְּתרֹעֵם ְּב‬You shall smash them with an iron scepter, break them in pieces like a potter’s vessel’, the form ‫ְּתרֹעֵם‬ is an Aramaism, because ‫ רעע‬is the regular Aramaic reflex of Proto–West Semitic *ŕṣ́ṣ ‘to smash, shatter’, of which the Hebrew reflex is ‫( רצצ‬HALOT 1270b–71a, 1285b–86a, 1983b). The most plausible historical background for Aramaic linguistic influence on Judahite Hebrew is the use of the Aramaic language as a vehicle of diplomatic and administrative communication under 23.  No nouns with a 3mp suffixed pronoun appear in the Song of the Sea (Exod 15:1– 18), so one cannot tell anything about the use of the 3mp suffixed pronoun with nouns in the language of the author(s) of this poetic unit.

The Third-Person Masculine Plural Suffixed Pronoun -mw

159

the auspices of the empires that controlled the ancient Near East from the second half of the eighth century b.c.e. onwards (see Bloch 2008: 34 n. 45, and the earlier studies cited there). To be sure, the LXX of Ps 2:9 reads ποιμανεῖς αὐτοὺς ‘you will shepherd them’, which implies Hebrew *‫ּת ְרעֵם‬, ִ and some scholars consider the reading of the LXX to be more original (thus, for example, Wilhelmi 1977). However, in none of the occurrences of the verb ‫רעע‬ ‘to smash, shatter’ in the Hebrew parts of the Bible 24 does the LXX translate it with a Greek verb of comparable meaning; hence, it appears that the translators of the LXX were simply unaware of the existence of this verb in Hebrew and tried to make the best sense they could of the consonantal spelling ‫ תרעם‬in Ps 2:9. Moreover, the pseudepigraphic Psalms of Solomon 17:23b–24a provide a paraphrase of Ps 2:9 that employs the infinitive συντρῖψαι ‘to smash’, reflecting a good understanding of the Hebrew ‫ ְּתרֹעֵם‬by a different Greek translator (see Pietersma 2004: 56). Thus, the MT ‫ ְּתרֹעֵם‬should be considered original. Second, the mention of an ‘iron scepter’ (‫ׁשבֶט ַּב ְרזֶל‬ ֵ ) with which the king of Judah would smash the nations of the world appears to refer to a scepter in the form of a mace, since a mace was a standard attribute with which kings were portrayed in the ancient Near East (see ANEP, nos. 296, 298, 414, 439, 442, 447, 461, 529, 537). A considerable number of mace-heads that had probably originally belonged to scepters of this kind and that were cast partly in iron are known from the eighth century b.c.e., but not earlier (Lemaire 1986). Third, an analysis of the literary representation of the king in Psalm 2 presented recently by Eckart Otto shows that the psalm draws partly on literary motifs whose earliest datable appearance in ancient Near Eastern literary works is in Neo-Assyrian royal inscriptions spanning the period from the mid-eighth to the mid-seventh century b.c.e. The motifs in question are the conspiracy of the foreign nations to throw off the yoke of the king and the king smashing his enemies like pieces of pottery (Otto 2002: 34–51; 2004). 25 In fact, even the 24. The following appearances of ‫‘ רעע‬to smash, shatter’ are specified in HALOT (1270b–71a) as certain: qal—Isa 24:19, Jer 15:12, Ps 2:9, Job 34:24; Hithpael (Hithpolel)— Isa 24:19, Prov 18:24. BDB 949b–50a add to these the following occurrences in Qal: Jer 11:16, Prov 25:19 (and prefer the reading of the LXX in Ps 2:9). 25.  The literary dependence of Psalm 2 on Neo-Assyrian royal propaganda motifs has already been demonstrated by Ringgren (1983) and Becking (1990). Becking’s study shows that the literary motifs employed in Psalm 2 are first attested in the inscriptions of the Assyrian kings of the second half of the eighth century b.c.e., Tiglath-pileser III (745–727 b.c.e.) and Sargon II (722–705 b.c.e.). The attempt by Koch (1999: 255–59) to connect the conspiracy of foreign nations in Ps 2:2–3 with Syro-Canaanite mythical tradition is not convincing, because the notion of such a conspiracy (which would imply that the king was justly granted power over the foreign nations, and their efforts to free themselves from his power were criminal) is not attested in Syrian or Levantine epigraphic sources from the pre-Hellenistic period. In any event, Koch (1999: 267–68) acknowledges at least a partial dependence of Psalm 2 on the motifs of Assyrian royal propaganda from the period spanning ca. 750–640 b.c.e.

160

Yigal Bloch

representation of the king as a son of God, linked by Otto to the Egyptian royal protocol of the fifteenth–thirteenth centuries b.c.e., may well be drawn from Neo-Assyrian sources (see Tigay 2003: 247*–48*). The evidence presented above suggests that Psalm 2 was composed in the period from ca. 750 to 586 b.c.e. Since the use of the 3mp suffixed pronoun -mw in Psalm 2 is no less archaic, in qualitative terms, than the use of the same suffixed pronoun in Exodus 15, Deuteronomy 32, and Deuteronomy 33, it appears that the use of the 3mp suffixed pronoun -mw in the latter three poems does not constitute compelling evidence for dating these poems prior to the eighth century b.c.e. 5.  From the Specific to the General: The Weight of Cumulative Evidence in Dating Biblical Hebrew Poetry In the preceding discussion, I established that the use of the 3mp suffixed pronoun -mw cannot be used to establish that poems dated by Robertson to the thirteenth–tenth centuries b.c.e. were indeed composed in that period. In an earlier study, I demonstrated that the only other linguistic feature that Robertson considered diagnostic for dating some biblical poems to the thirteenth– tenth centuries b.c.e.—the use of (short) prefixed verbal forms without the conjunction w- to signify complete situations in the past—cannot serve this purpose either (Bloch 2009). Nevertheless, one should consider the possibility that, while no linguistic feature is sufficient on its own to date any poetic text in the Hebrew Bible to the late second millennium b.c.e., a significant concentration of several typologically archaic features in a given poetic text may justify this sort of early dating. This possibility was considered by Robertson himself. He singled out several linguistic features that, although not diagnostic for dating any given biblical poem to the thirteenth–tenth centuries b.c.e. when considered separately, could nevertheless support to a significant degree an early date for those poems in which they are clustered: But, though these forms considered one at a time are rarely valuable for dating, an accumulation of several of them within one poem may be evidence of an early date. We are postulating that all were common features of early poetry. Therefore, one would expect that in genuinely early poems several often occurred together in one poem, in other words, that a clustering effect would result. But, on the other hand, since all are rare in standard poetic Hebrew, no clustering, rather an isolated form here and there, would be anticipated. (Robertson 1972: 135)

The relevant features, the origins and distribution of which in Biblical Hebrew poetry were discussed extensively by Robertson (1972: 57–65, 69–110), can be summarized as follows: 26 26.  The following summary presents only features attested frequently in the Hebrew Bible. The features that are extremely rare in Biblical Hebrew, even in the poems considered

The Third-Person Masculine Plural Suffixed Pronoun -mw

161

1.  Preservation of w/y in a syllable-opening position in nouns and verbal forms derived from roots III-w/y—for example, ‫‘ ְיכ ְַסיֻמּו‬they covered them (m.)’ (Exod 15:5). 2.  The use of the relative pronoun ‫ זּו‬or ‫ זֶה‬instead of the standard ‫ֲׁשר‬ ֶ ‫—א‬for example, ‫ָל ָּת‬ ְ ‫זּו ָּגא‬-‫‘ עַם‬the people whom you (ms) have redeemed’ (Exod 15:13); ‫ִית‬ ָ ‫זּו ָקנ‬-‫‘ עַם‬the people whom you (2ms) have created’ (Exod 15:16). 3.  The nonassimilation of the 3ms object suffix -hû to the preceding verbal suffix -(ū)n—for example, ‫‘ ַואֲר ְֹמ ֶמנְהּו‬and I shall glorify him’ (Exod 15:2). 4.  The so-called compaginis vowels ‫ֹו‬- and ‫ִי‬- appended to nominal forms, for example, in the participle ‫ֶאָּדִרי‬ ְ ‫‘ נ‬glorious’ (Exod 15:6), which may be interpreted as remnants of Proto–Northwest Semitic case endings. 5.  The so-called enclitic mem, which appears to be related to the particle -m in Ugaritic, probably expressing focus (Cohen 2004: 232 n. 6). The unvocalized final -m was not recognized as a distinct morpheme by the Masoretes; hence locating this morpheme in any given instance in the Hebrew Bible requires text-critical analysis, which suggests that the extant morphological structure of a specific word as indicated in the Masoretic Text should be emended (see Cohen 2004, and the earlier studies cited there). However, for the prepositions -‫ב‬, -‫כ‬, -‫ ל‬with the ending ‫מֹו‬- (that is, in the forms ‫ּבמֹו‬, ְ ‫ּכמֹו‬, ְ ‫)למֹו‬, ְ the ending ‫מֹו‬was not assigned any specific morphological function by the Masoretes (this ending is obviously different from the 3mp suffixed pronoun -mw). Hence, the explanation of the ending ‫מֹו‬- in these forms as an enclitic focus marker is most reasonable and does not require any corrections to the Masoretic Text (Robertson 1972: 109; Cohen 2004: 240). An example of the enclitic mem used with one of the prepositions -‫ב‬, -‫כ‬, -‫ ל‬is provided by the form ‫( ְּכמֹו‬Exod 15:5, 8).

In light of the arguments offered in §§2–4 above and in Bloch 2009, two more linguistic features can be added to this list: 1.  The use of short prefixed verbal forms expressing a complete situation in the past without the conjunction w-, for example, ‫‘ ִּתּפֹל‬she fell’, and ‫ִּדמּו‬ ְ ‫‘ י‬they (m.) became silent’ (Exod 15:16). 2.  The use of the 3mp suffixed pronoun -mw with nouns or verbal forms—for example, ‫‘ ְיכ ְַסיֻמּו‬they covered them (m.)’ (Exod 15:5).

While these features on their own are not diagnostic for dating a given biblical poem to the thirteenth–tenth centuries b.c.e. (contra Robertson 1972: 148), they are typologically archaic and must be taken into account when one considers the clustering of typologically archaic features in some units of Biblical Hebrew poetry. early by Robertson (1972: 111–33), are not included in the following discussion, because, given the rarity of their use, they may be mere manifestations of idiosyncratic tendencies of individual authors rather than linguistic features common during a particular historical stage of the Hebrew language. In any event, Robertson found that adding information about extremely rare features to his counts of more common features (listed in the text after this note) did not alter his conclusions significantly (Robertson 1972: 148–49).

162

Yigal Bloch

The list of typologically archaic linguistic features presented above has included examples from the Song of the Sea (Exod 15:1–18), and this is not by chance. It is the single poetic unit in the whole Hebrew Bible within which all the relevant features are concentrated (Robertson 1972: 138). This situation dovetails with the fact that, insofar as the use of the 3mp suffixed pronoun -mw with nouns and verbal forms is concerned, the Song of the Sea evidences more instances than any other literary unit in the Hebrew Bible (see above, §3). But are these instances sufficient reason to conclude that the Song of the Sea was composed several centuries earlier than the main bulk of poetic material in the Hebrew Bible? Or, to phrase the question differently: What is the significance of the clustering, and should it be explained by hypothesizing diachronic processes? The answer is that the clustering of typologically archaic linguistic features in a given literary unit is a quantitative phenomenon. As such, it cannot be considered a decisive criterion for the dating of the relevant literary unit if this literature exhibits no features that are qualitatively distinct for a specific stage in the development of the Hebrew language. For example, in Deutero-Isaiah (Isaiah 40–66), composed around the time that the Neo-Babylonian empire conquered by Cyrus the Great (539 b.c.e.), the following typologically archaic linguistic features are attested: 27 1.  Preservation of w/y in what was originally a syllable-initial position in nouns and verbal forms derived from roots III-w/y: ‫‘ ְּתדַ ְּמיּון‬you (pl.) will liken’ (Isa 40:18); ‫‘ ְתדַ ְּמיּונִי‬you (pl.) will liken me’ (Isa 40:25, 46:5); 28 ‫ֱתיּון‬ ָ ‫‘ ַוּיֶא‬and they (m.) came’ (Isa 41:5).  29 2.  Two occurrences of the relative pronoun ‫זּו‬: ‫‘ יהוה זּו ָחטָאנּו לֹו‬YHwh, against whom we have sinned’ (Isa 42:24); ‫‘ עַם זּו יָצ ְַר ִּתי ִלי‬the people whom I have created for Myself’ (Isa 43:21). 27. The question whether one should consider only Isaiah 40–55 to be the work of Deutero-Isaiah and postulate a separate origin for Isaiah 56–66 (Trito-Isaiah) is immaterial to the present discussion, because all the typologically archaic linguistic features detailed here appear in Isaiah 40–55. 28.  According to the rules of syllabification in the Masoretic Text, a consonant followed by šĕwa mobile is joined to the following syllable containing a full vowel, if there is one; in such a case, a syllable beginning with two consonants is formed (Joüon 1993: §27da). However, the situation discussed here is one that must have obtained before the elision of vowels in some originally open syllables that resulted in the Masoretic šĕwa mobile (see Joüon 1993: §30dg). At that stage, the forms appearing in Isa 40:18, 25; and 46:5 would supposedly have been pronounced *tudammiyūn and *tudammiyūnî. 29.  The construction ‫ֱתיּון‬ ָ ‫ ַוּיֶא‬, employing a prefixed verbal form ending -ûn, which had originally belonged to the 2 and 3mp forms of the long prefixed conjugation, nevertheless expresses a complete situation in the past and thus must be considered, from the syntactic viewpoint, an inept attempt at archaizing (Bloch 2009: 62 n. 112). However, morphologically, the retention of the radical y in a syllable-initial position in this form is no less archaic than the retention of this radical elsewhere.

The Third-Person Masculine Plural Suffixed Pronoun -mw

163

3.  Eight instances of the original enclitic mem with prepositions: ‫‘ ְּבמֹו‬in, by’ (Isa 43:2; 44:16, 19), ‫‘ ְּכמֹו‬like’ (Isa 41:25 [2×], 44:7, 46:9, 51:6). 4.  Six short prefixed verbal forms signifying complete situations in the past without the conjunction w- in a single judgment speech, Isa 41:1–5 (see Bloch 2009: 61–64).

Of these features, only the form ‫ ְּכמֹו‬appears in the prophetic books of Ezekiel, Jeremiah, Haggai, and First Zechariah (chaps. 1–8), which were likewise composed in the sixth century b.c.e. (Jer 13:21, 15:18, 30:7, 49:19, 50:26, 44; Ezek 5:9, 16:57; Hag 2:3; Zech 5:3 [2×]). In addition, in Jeremiah and Ezekiel there are a few instances of the vowel endings ‫ָה‬- (unstressed) and ‫ִי‬(stressed) that can be understood as remnants of original case endings: ‫ישבתי‬ ‘she who dwells’ (Jer 10:17, Kethiv); ‫ מקננתי‬,‫‘ ישבתי‬she who nests’ (Jer 22:23, Kethiv); ‫‘ ׁש ְֹכנִי‬he who resides’, ‫‘ ּת ְֹפ ִׂשי‬he who holds’ (Jer 49:16); ‫‘ שכנתי‬she who resides’ (Jer 51:13, Kethiv); ‫‘ הישבתי‬she who dwells’ (Ezek 27:3, Kethiv); ‫ַׁש ַמלָה‬ ְ ‫‘ ַהח‬electrum (or, the color of amber)’ (Ezek 8:2). However, the overall concentration of typologically archaic linguistic features in these books is much less pronounced than in Deutero-Isaiah. Nevertheless, it would be wrong to claim, on these grounds, that DeuteroIsaiah is older than Jeremiah and Ezekiel, or significantly older than Haggai and First Zechariah. In part, this is due to historical considerations, which clearly relate the main core of the books of Jeremiah and Ezekiel to the period shortly before and shortly after the destruction of the Kingdom of Judah in 586 b.c.e.; and the literary units of Deutero-Isaiah, Haggai, and First Zechariah to the end of the Babylonian Exile and the beginning of the Restoration under the auspices of the Persian Empire in the 530s–510s b.c.e.  30 But also from the linguistic viewpoint, all the typologically archaic traits appearing in DeuteroIsaiah are attested in other works of biblical poetry originating no earlier than the Babylonian Exile: 1.  Radical y is preserved in a syllable-initial position in the verbal form ‫ִׁשלָיּו‬ ְ‫י‬ ‘may they prosper’ (Ps 122:6). 31 Psalm 122 is one of the Psalms of Ascent (Psalms 120–34), which appear to have been composed as a single literary work during the Second Temple period (Hunter 1999). 30.  The main core of the books of Jeremiah and Ezekiel does, in all likelihood, date to the period of the activity of the respective prophets, although the two books appear to have undergone considerable later redaction (see Schmid 2006: 335–48, 351–59; Rom-Shiloni 2008: 110–13, and the earlier studies cited there). For the dating of Deutero-Isaiah, see Eng 2004; Schmid 2006: 328–33. For the dating of Haggai and Zechariah 1–8, see Schmid 2006: 393–97. 31.  Of course, the verb was originally šlw; compare ‫ׁשלְַו ִּתי‬ ָ ‘I was (not) at ease’ (Job 3:26). However, III-w verbs were normally perceived as III-y in Biblical Hebrew, including the verbal forms ‫( ְיכ ְַסיֻמּו‬Exod 15:5) and ‫ֱתיּון‬ ָ ‫( ַוּיֶא‬Isa 41:5), derived from what were originally the verbs *ksw and *ʾtw (see HALOT, 102a, 487b–88a).

164

Yigal Bloch

2.  The relative pronouns ‫ זֶה‬and ‫זֹו‬, derived from the same base as the relative pronoun ‫זּו‬, namely, the Proto–West Semitic *ḏ (Lipiński 1997: 324–26), appear in the constructions ‫ְּת ּבֹו‬ ָ ‫ׁש ַכנ‬ ָ ‫צּיֹון זֶה‬-‫ַר‬ ִ ‫‘ ה‬Mount Zion, where You have dwelt’ (Ps 74:2); and ‫אל ְַּמדֵ ם‬ ֲ ‫‘ ְועֵד ִֹתי זֹו‬and My decrees that I teach them’ (Ps 132:12). Psalm 74 is commonly understood as a psalm of lament in the wake of the destruction of the Jerusalem Temple in 586 b.c.e. (see, for example, Hossfeld and Zenger 2005: 243–44, but contrast the earlier dating and the Northern Israelite origin of the main body of the psalm proposed by Weber 2000). Psalm 132 is a Psalm of Ascent and thus probably dates to the Second Temple period. 3.  The use of the form ‫ ְּכמֹו‬in Ezekiel, Jeremiah, Haggai, and First Zechariah has already been noted above, in the present section. 4.  The use of short prefixed verbal forms to signify complete situations in the past without the conjunction w- is attested in the exilic Psalm 44 (Bloch 2009: 64–66).

On the other hand, the noun ‫ֲרב‬ ָ ‫‘ ַמע‬west’ (Isa 43:5, 45:6, 59:19) appears, apart from Deutero-Isaiah, only in Daniel (8:5), Chronicles (1 Chr 7:28; 12:16; 26:16, 18, 30; 2 Chr 32:30; 33:14), and a few psalms that provide no content clues to use for dating (Ps 75:3, 103:12, 107:3). Furthermore, the prepositional phrase ‫ ְכ ֶאחָד‬is used with the adverbial meaning ‘together’ in Isa 65:25, and elsewhere only in Ezra–Nehemiah (Ezra 2:64, 3:9, 6:20; Neh 7:66), Chronicles (2 Chr 5:13), and Qoheleth (11:6), which is commonly considered a postexilic book by biblical scholars. Thus, ‫ֲרב‬ ָ ‫‘ ַמע‬west’ and ‫‘ ְכ ֶאחָד‬together’ should be understood as features limited qualitatively to postexilic literary works in the Hebrew Bible, and they provide the decisive criterion for postexilic dating of Deutero-Isaiah in linguistic terms. 32 The example of Deutero-Isaiah demonstrates that the quantitative prevalence of typologically archaic linguistic features in a given literary unit in the Hebrew Bible does not suffice to date its composition earlier than the composition of other literary units in which the concentration of the same linguistic features is less pronounced, if none of the features involved is limited qualitatively to a specific period in the history of the Hebrew language. Consequently, the clustering of typologically archaic linguistic features in the Song of the Sea (Exod 15:1–18) cannot be considered decisive evidence for dating its composition to the thirteenth–tenth centuries b.c.e. Perhaps the Song of the Sea does indeed date somewhat earlier than the mid-eighth century b.c.e. But in view of the fact that the typologically archaic linguistic features attested in the Song of the Sea appear also in poetic and prophetic texts composed in the eighth– 32.  The implications of ‫“ ַמעֲָרב‬west” and ‫“ ְכ ֶאחָד‬together” for dating the language of Deutero-Isaiah have been noted by Ehrensvärd (2003: 181), with reference to earlier studies by Avi Hurvitz. Ehrensvärd considered Psalms 75 and 107, which use the noun ‫( ַמעֲָרב‬Ps 75:7, 107:3), to be pre-exilic, and hence expressed a reservation concerning the classification of ‫ ַמעֲרָב‬as a qualitatively distinct characteristic of exilic or post-exilic Hebrew. However, he did not provide any specific arguments or references for a pre-exilic dating of those psalms.

The Third-Person Masculine Plural Suffixed Pronoun -mw

165

sixth centuries b.c.e. (for example, Psalm 2 and Deutero-Isaiah), dating the Song of the Sea to the late second millennium b.c.e. on linguistic grounds is unwarranted. The same is also true, by logical inference, of the poetic units in Genesis–Samuel where the quantitative concentration of typologically archaic linguistic features is less salient than in Exodus 15. 6. Conclusions The conclusions of this essay are twofold. First, I have demonstrated that, while the 3mp suffixed pronoun -mw (vocalized in the Masoretic Text as either ‫מֹו‬- or ‫מּו‬-) is indeed archaic from the viewpoint of the development of the Hebrew language, it remained in use well into the first millennium b.c.e. With regard to prepositions, the use of the 3mp suffixed pronoun -mw is attested as late as the Second Temple period, in Deutero-Isaiah, Lamentations, Job (where this pronoun is sometimes used in an artificial, archaistic manner by an author who no longer understood the rules of its proper use—see Job 20:23, 22:2, 27:23), and the Qumran pesher to Isaiah 4Q163 (if the latter does not reflect the original text of Isa 14:8). Furthermore, the use of the 3mp suffixed pronoun -mw with nouns and verbal forms is attested in Psalm 2, and I have provided arguments for dating the composition of Psalm 2 between the mid-eighth and the early sixth centuries b.c.e. Thus, the use of the 3mp suffixed pronoun -mw cannot be considered diagnostic for dating any Biblical Hebrew poem to the thirteenth–tenth centuries b.c.e. Second, I have considered the significance of the quantitative concentration of typologically archaic linguistic features in poetic texts allegedly belonging to the corpus of early Hebrew poetry (specifically, in the Song of the Sea, Exod 15:1–18, where the clustering of these features is most salient). I have shown that some of the typologically archaic features in question also appear in Deutero-Isaiah (Isaiah 40–66), a prophetic work composed in the late sixth century b.c.e., after the end of the Babylonian Exile. The concentration of typologically archaic linguistic features in Deutero-Isaiah is greater than their concentration in such prophetic books as Jeremiah, Ezekiel, Haggai, and First Zechariah (Zechariah 1–8), which also should be dated, on historical grounds, to the sixth century b.c.e. Thus, the concentration of typologically archaic linguistic features in a certain literary unit within the Hebrew Bible is not a sufficient reason to date the unit significantly earlier than other literary works in which the concentration of the same features is less prevalent. In dating Biblical Hebrew texts on linguistic terms, decisive weight should be attributed to qualitatively distinctive linguistic features—that is, to the features that are fully limited to a specific period in the history of the Hebrew language. When focusing on qualitatively distinct linguistic features, one should be aware of the limitations of the available evidence. Because there is no direct indication of what Hebrew looked like in the closing centuries of the second

166

Yigal Bloch

millennium b.c.e., comparison with Ugaritic or with Canaanite forms in the El-Amarna letters can only provide a set of features that may have existed in Hebrew at the stage of its inception as a distinct language. At the same time, one must reckon with the possibility that, in some aspects, Hebrew was from its very beginning innovative compared with the linguistic milieu out of which it emerged. For this reason, the argument presented by Young and Rezetko (2008: 336–38)—who listed several features in Exod 15:1–18 that are typologically late compared with Ugaritic and with El-Amarna Canaanite but are standard for Biblical Hebrew sources dating to the eighth century b.c.e. and later—is not sufficient to demonstrate that Exod 15:1–18 should be dated to the same general period as the rest of the Hebrew Bible. 33 However, Young and Rezetko are right to stress that, in order to establish a possible range of dates for Exod 15:1–18 (or any other biblical literary unit), one must consider not only the linguistic features connecting that unit with the linguistic milieu of Canaan in the fourteenth–thirteenth centuries b.c.e. but also the features that connect it with other parts of the Hebrew Bible. And one should add that establishing the distribution of a given linguistic feature in biblical literary units that can be dated on nonlinguistic grounds is prerequisite to considering the possibility of the use of that feature to date the literary units whose period of composition is not clear. With this notion in mind, it is possible to turn now to a final evaluation of Robertson book (1972) in light of my conclusions in this essay and in Bloch 2009. Robertson’s analysis established that only two linguistic features attested in some of the poetic works preserved in the books of Genesis–Samuel—(1) the use of (short) prefixed verbal forms to express complete situations in the past without the conjunction w-, and (2) the use of the 3mp suffixed pronoun -mw with nouns and verbal forms—could be considered diagnostic for dating these poetic works to the late second millennium b.c.e. However, I conclude that the use of the 3mp suffixed pronoun -mw with nouns and verbal forms continued at least until the mid-eighth century b.c.e., and I have also demonstrated (Bloch 2009) that the use of short prefixed verbal forms to express complete situations in the past without the conjunction w- continued at least until the sixth 33.  Actually, with regard to one of the features listed by Young and Rezetko, Biblical Hebrew is conservative rather than innovative compared with Ugaritic and the regular usage of El-Amarna Canaanite. The prefix y- in 3mp verbal forms belonging to prefix-conjugations can be traced back to Proto-Semitic (Voigt 1987: 12–13), even though the 3mp prefixed verbal forms in Ugaritic and El-Amarna Canaanite employ the prefix t-, which had probably developed by analogy with the 3fp prefix t- (Tropper 2000: 432–38; Rainey 1996b: 43–45). It appears that the prefix y- in 3mp prefixed verbal forms, inherited from Proto-Semitic, survived all along in Canaanite dialects of the fourteenth–thirteenth centuries b.c.e., even though during this period its distribution was severely limited by the use of the prefix t- with 3mp prefixed verbal forms (for two attestations of the prefix y- in 3mp prefixed verbal forms in the El-Amarna letters from Byblos, see Rainey 1996b: 44).

The Third-Person Masculine Plural Suffixed Pronoun -mw

167

century b.c.e. Consequently, I conclude that Robertson’s proposal for dating the composition of the poetic units in Genesis–Samuel to the thirteenth–tenth centuries b.c.e. cannot be supported by linguistic evidence. References

Albright, W. F. 1944 The Oracles of Balaam. JBL 63: 207–33. 1945 The Old Testament and Canaanite Language and Literature. CBQ 7: 5–31. 1968 Yahweh and the Gods of Canaan: A Historical Analysis of Two Contrasting Faiths. Garden City, NY: Doubleday. Allegro, J. M., with the collaboration of A. A. Anderson 1968 Qumrân Cave 4, I: 4Q158–4Q186. DJD 5. Oxford: Clarendon. Bar-Asher, E. A. 2009 Dual Pronouns in Semitics and an Evaluation of the Evidence for Their Existence in Biblical Hebrew. Ancient Near Eastern Studies 46: 32–49. Bauer, H., and Leander, P. 1922 Historische Grammatik der hebräischen Sprache des Alten Testasmentes, vol. 1: Einleitung, Schriftlehre, Laut- und Formenlehre. Halle: Max Niemeyer. Becking, B. 1990 “Wie Töpfe Sollst Du Sie Zerschmeißen”: Mesopotamische Parallelen zu Psalm 2,9b. ZAW 102: 59–79. Bloch, Y. 2008 Should Parallelistic Structure Be Used as Evidence for an Early Dating of Biblical Hebrew Poetry? JANES 31: 23–45. 2009 The Prefixed Perfective and the Dating of Early Hebrew Poetry: A Reevaluation. VT 59: 34–70. Brønno, E. 1943 Studien über hebräische Morphologie und Vokalismus auf Grundlage der Merkatischen Fragmente der zweiten kolumne der Hexapla des Origenes. Leipzig: Brockhaus. Cohen, C. 2004 The Enclitic-mem in Biblical Hebrew: Its Existence and Initial Discovery. Pp.  231–60 in Sefer Moshe: The Moshe Weinfeld Jubilee Volume. Studies in the Bible and the Ancient Near East, Qumran, and Post-Biblical Judaism, ed. C. Cohen, A. Hurvitz, and S. M. Paul. Winona Lake, IN: Eisenbrauns. Cross, F. M. 1973 Canaanite Myth and Hebrew Epic: Essays in the History of the Religion of Israel. Cambridge: Harvard University Press. Cross, F. M., and Freedman, D. N. 1975 Studies in Ancient Yahwistic Poetry. SBLDS 21. Missoula, MT: Scholars Press. Ehrensvärd, M. 2003 Linguistic Dating of Biblical Texts. Pp. 164–88 in Biblical Hebrew: Studies in Chronology and Typology, ed. I. M. Young. JSOTSup 369. London: T. & T. Clark.

168

Yigal Bloch

Eng, M. 2004 What’s in a Name? Cyrus and the Dating of Deutero-Isaiah. Pp. 216–24 in Inspired Speech: Prophecy in the Ancient Near East. Essays in Honor of Herbert B. Huffmon, ed. J. Kaltner and L. Stulman. JSOTSup 378. London: T. & T. Clark. Fassberg, S. E. 2009 Vowel Dissimilation in Plural Pronouns in Biblical Hebrew. Or 78: 326–35. Freedman, D. N. 1997 Archaic Forms in Early Hebrew Poetry. Pp. 5–12 in Divine Commitment and Human Obligation: Selected Writings of David Noel Freedman, vol. 2: Poetry and Orthography, ed. J. R. Huddlestun. Grand Rapids, MI: Eerdmans. Greenstein, E. L. 2003 The Language of Job and Its Poetic Function. JBL 122: 651–66. Halpern, B. 2002 Assyrian and Pre-Socratic Astronomies and the Location of the Book of Job. Pp. 255–64 in Kein Land für sich allein: Studien zum Kulturkontakt in Kanaan, Israel/Palästina und Ebirnâri für Manfred Weippert zum 65. Geburts­tag, ed. U. Hübner and E. A. Knauf. OBO 186. Freiburg: Universitätsverlag / Göttingen: Vandenhoeck & Ruprecht. Hasselbach, R. 2005 Sargonic Akkadian: A Historical and Comparative Study of the Syllabic Texts. Wiesbaden: Harrassowitz. Hossfeld, F.-L., and Zenger, E. 2005 Psalms 2: A Commentary on Psalms 51–100, ed. K. Baltzer, trans. L. M. Maloney. Hermeneia. Minneapolis: Fortress. Hunter, A. G. 1999 The Psalms of Ascents: A Late Festival Recovered? Pp. 173*–87* in Proceedings of the Twelfth World Congress of Jewish Studies: Jerusalem, July 29–August 5, 1997, Division A: The Bible and Its World, ed. S. Aḥituv et al. Jerusalem: World Union of Jewish Studies. Janssens, G. 1982 Studies in Hebrew Historical Linguistics Based on Origen’s Secunda. Leuven: Peeters. Joüon, P. 1993 A Grammar of Biblical Hebrew, trans. and rev. T. Muraoka. Subsidia Biblica 14/1–2. Rome: Pontifical Biblical Institute. Knauf, E. A. 1988 Hiobs Heimat. WO 19: 65–83. Knudtzon, J. A. 1907–15  Die El-Amarna-Tafeln. 2 vols. Leipzig: Hinrichs. Koch, K. 1999 Israel im Orient. Pp. 242–71 in Religionsgeschichte Israels: Formale und materiale Aspekte, ed. B. Janowski and M. Köckert. Gütersloh: Chr. Kaiser. Lemaire, A. 1986 “Avec un sceptre de fer”: Ps. II,9 et l’archéologie. BN 32: 25–30.

The Third-Person Masculine Plural Suffixed Pronoun -mw

169

Lipiński, E. 1997 Semitic Languages: Outline of a Comparative Grammar. OLA 80. Leuven: Peeters. Moran, W. L. 1973 The Dual Personal Pronouns in Western Peripheral Akkadian. BASOR 211: 50–53. Otto, E. 2002 Politische Theologie in den Königspsalmen zwischen Ägypten und Assyrien: Die Herrscherlegitimation in den Psalmen 2 und 18 in ihren altorientalischen Kontexten. Pp.  33–65 in “Mein Sohn bist du” (Ps 2,7): Studien zu den Königs­psalmen, ed. E. Otto and E. Zenger. SBS 192. Stuttgart: Katholisches Bibelwerk. 2004 The Judean Legitimation of Royal Rulers in Its Ancient Near Eastern Contexts. Pp. 131–39 in Psalms and Liturgy, ed. D. J. Human and C. J. A. Vos. JSOTSup 410. London: T. & T. Clark. Pardee, D. 1997 Ugaritic Myths: The Baʿlu Myth: Dawn and Dusk. Pp. 241–83 in The Context of Scripture, vol. 1: Canonical Composition from the Biblical World, ed. W. W. Hallo and K. L. Younger. Leiden: Brill. Pietersma, A. 2004 Empire Re-affirmed: A Commentary on Greek Psalm 2. Pp. 46–62 in God’s Word for Our World, vol. 2: Theological and Cultural Studies in Honor of Simon John De Vries, ed. J. H. Ellens et al. JSOTSup 389. London: T. & T. Clark . Qimron, E. 1986 The Hebrew of the Dead Sea Scrolls. HSS 29. Atlanta: Scholars Press. Rainey, A. F. 1978 El Amarna Tablets 359–379: Supplement to J. A. Knudtzon, Die El-AmarnaTafeln. AOAT 8. 2nd rev. ed. Kevelaer: Butzon & Bercker / NeukirchenVluyn: Neukirchener Verlag. 1996a Canaanite in the Amarna Tablets: A Linguistic Analysis of the Mixed Dialect Used by the Scribes from Canaan, vol. 1: Orthography, Phonology, Morphosyntactic Analysis of the Pronouns, Nouns, Numerals. HO 25/1. Leiden: Brill. 1996b Canaanite in the Amarna Tablets: A Linguistic Analysis of the Mixed Dialect Used by the Scribes from Canaan, vol. 2: Morphosyntactic Analysis of the Verbal System. HO 25/2. Leiden: Brill. Ratner, R. 1988 Does a t-Preformative Third Person Masculine Plural Verbal Form Exist in Biblical Hebrew? VT 38: 80–88. Ringgren, H. 1983 Psalm 2 and Bēlit’s Oracle for Ashurbanipal. Pp. 90–95 in The Word of the Lord Shall Go Forth: Essays in Honor of David Noel Freedman in Celebration of His Sixtieth Birthday, ed. C.  L. Meyers and M. O’Connor. Winona Lake, IN: Eisenbrauns.

170

Yigal Bloch

Robertson, D. A. 1972 Linguistic Evidence in Dating Early Hebrew Poetry. SBLDS 3. Missoula, MT: Society of Biblical Literature. Rom-Shiloni, D. 2008 Deuteronomic Concepts of Exile Interpreted in Jeremiah and Ezekiel. Pp. 101–23 in Birkat Shalom: Studies in the Bible, Ancient Near Eastern Literature, and Postbiblical Judaism Presented to Shalom M. Paul on the Occasion of His Seventieth Birthday, ed. C. Cohen et al. Winona Lake, IN: Eisenbrauns. Schmid, K. 2006 Hintere Propheten (Nebiim). Pp. 303–401 in Grundinformation Altes Testament: Eine Einführung in Literatur, Religion und Geschichte des Altes Testaments, ed. J. C. Gertz. Göttingen: Vanderhoeck & Ruprecht. Skehan, P. W., and Ulrich, E. 1997 4QIsae. Pp. 91–97 in Qumran Cave 4, X: The Prophets, ed. E. Ulrich et al. DJD 15. Oxford: Clarendon. Tigay, J. H. 2003 Divine Creation of the King in Psalm 2:6. ErIsr 27 (Miriam and Hayim Tadmor volume): 246*–51*. Tropper, J. 2000 Ugaritische Grammatik. AOAT 273. Münster: Ugarit-Verlag. Voigt, R. M. 1987 The Classification of Central Semitic. JSS 32: 1–21. 1988 Die Personalpronomina der 3. Personen im Semitischen. WO 18: 49–63. Weber, B. 2000 Zur Datierung der Asaph-Psalmen 74 und 79. Bib 81: 521–32. Whiting, R. M. 1972 The Dual Personal Pronouns in Akkadian. JNES 31: 331–37. Wilhelmi, G. 1977 Der Hirt mit dem eisernen Szepter: Überlegungen zu Psalm ii 9. VT 27: 196–204. Young, I. M.; and Rezetko, R.; and Ehrensvärd, M. 2008 Linguistic Dating of Biblical Texts, vol. 1: An Introduction to Approaches and Problems. London: Equinox.

The Kethiv/Qere ‫הוא‬, ִ Diachrony, and Dialectology Steven E. Fassberg 1. Introduction The Kethiv/Qere perpetuum ‫ ִהוא‬of the 3fs independent pronoun is attested 120 times in the Pentateuch, as well as 3 times in the Prophets and Writings (Rendsburg 1982: 351; Tropper 2001: 159). The phenomenon has long perplexed scholars and raised questions that have not yet been satisfactorily answered: Does the Kethiv reflect a pronunciation in which the 3ms and 3fs independent pronouns were realized identically as [hū]? Why are there at least 11 exceptions, in which the 3fs pronoun is spelled as expected—‫ ?היא‬1 And why is the Kethiv/Qere ‫ ִהוא‬limited essentially to the Pentateuch? I propose to reinvestigate ‫ ִהוא‬in light of the phenomenon of Kethiv/Qere in general, the relative date of the crystallization of the Pentateuch, and comparative data that, to date, have not been brought to bear on the discussion. I believe the Kethiv/Qere ‫ִהוא‬ has implications for the diachrony and dialectology of Biblical Hebrew. 2.  History of Research Most nineteenth- and twentieth-century grammarians attributed the phenomenon of Kethiv/Qere ‫ ִהוא‬to the scribal process by which medial matres lectionis were introduced into the consonantal biblical text. It has been widely accepted that scribes inserted a waw systematically whenever they encountered ‫ הא‬in the consonantal text until they reached the book of Joshua, at which point they realized that not all cases of ‫ הא‬reflected the m.s. [hū]. According to this scenario, at a later period the Tiberian Masoretes eliminated the grammatical solecism by vocalizing it with a ḥireq. 2 There is no explanation, however, 1.  The examples usually cited include Gen 14:2; 20:5; 38:25; Lev 11:39; 13:10, 21; 16:31; 20:17; 21:9; Num 5:13, 14. Tropper (2001: 159) collected another 7 examples in BHS where one finds ‫היא‬: Gen 19:20; 26:7; 40:10; Exod 1:16; Lev 5:11; 13:6; 20:18. He pointed out that in 6 of the verses where ‫ היא‬occurs as expected, ‫ הוא‬is also attested (Gen 19:20; 20:5; 26:7; 38:25; Num 5:13, 14). 2. See Rendsburg 1982: 351–52 for a brief survey of opinions. Yeivin (1985: 1103) cites one example (Deut 11:10) of Kethiv/Qere in a Babylonian vocalized manuscript.

171

172

Steven E. Fassberg

for the fact that the scribes got it right in the 11 cases in the Pentateuch where they added a yod (‫)היא‬, but in 3 other cases in the Prophets and Writings they incorrectly added a waw, when the 3fs pronoun was intended. 3 Some scholars have offered a different orthographic explanation, one which blames the 3fs Kethiv ‫ הוא‬on manuscripts that did not distinguish between waw and yod (see Emerton 2000: 267 for bibliography). For example, Cross (1998: 222–23) writes: “There are also bits of evidence not hitherto used which tend to support an early-first-century c.e. date for the Recension. There is the bizarre phenomenon of the Qere perpetuum in the Pentateuch where the feminine pronoun hîʾ is spelled hwʾ in the Kethiv. The most plausible explanation for this is that the manuscript or manuscripts copied for the pentateuchal recension was a manuscript in which waw and yod were not distinguished in the Jewish script. This occurs at only one time in the development of the Jewish scripts: in the Early Herodian period (30–1 b.c.e.).” Few have considered the possibility that the orthography ‫ הוא‬actually reflected a common pronunciation [hū] of the 3rd-person independent pronoun for the reason that Semitic languages maintain a distinction between 3ms and 3fs forms (GKC §32l). A notable exception is Lambert (1946: 34 n. 3): “L’ancienne explication d’après laquelle ‫ הּוא‬et ‫ נַעַר‬ont été employés pendant un certain temps pour les deux genres nous paraît toujours la meilleure.” He was preceded in this by Green (1872: 96) and more recently has been followed by Rendsburg (1982), Tropper (2001), and Morgenstern (2007: 49–50). After a long period in which the enigmatic ‫ ִהוא‬was largely ignored, Rendsburg took up the subject briefly in 1980 and then returned to it in more detail in 1982. He suggested that the pentateuchal orthography ‫ הוא‬reflected a 3rd-person common pronoun [hū] that was the result of a Hittite and Hurrian substratum in the Israelite hill country; both languages had a 3cs pronoun. Moreover, he pointed out the existence in Old Babylonian of a 3rd-person common oblique form [šuʾāti] (accusative/genitive) / [šuʾāšim] (dative), which, he hinted, was also the result of Hurrian influence. 4 Rendsburg’s bold proposal reawakened interest in the subject, though somewhat belatedly: two decades 3.  In two of the non-pentateuchal occurrences, one wonders if the existence of consecutive 3rd-person pronouns is responsible somehow for the Kethiv/Qere: ‫ותלך ותעשה כדבר‬ ‫‘ אליהו ותאכל ִהוא־והּיא וביתה ימים‬She went and did as Elijah had said, and she and he and her household had food for a long time’ (1 Kgs 17:15); ‫־הוא זמה והּיא עון פלילים‬‎ ִ ‫‘ כי‬For that would have been debauchery, a criminal offense’ (Job 31:11). The text and language of the third example is difficult: ‫גם־הוא למלך הוכן העמיק הרחב מדורתה‬ ִ ‫‘ כי־ערוך מאתמול תפתה‬The Topheth has long been ready for him; he too is destined for Melech, his firepit has been made both wide and deep’ (Isa 30:33 njps). 1QIsaa reads ‫ היה‬in this passage: ‫כי ערוך מאתמול תפתח‬ ‫גמ היה למלך יוכן הכיני והעמקי הרחיבי מדורתה‬. 4. Compare the separate Old Babylonian masculine and feminine nominative forms: ms [šū] and fs [šī].

The Kethiv/Qere ‫הוא‬, ִ Diachrony, and Dialectology

173

passed before Emerton (2000) responded, rejecting Rendsburg’s hypothesis because he considered the older orthographic explanation more likely. A year after Emerton wrote, Tropper (2001) proposed a new solution no less challenging than the suggestion advanced by Rendsburg: the pentateuchal spelling ‫ הוא‬for the 3fs possibly indicated that one early Hebrew dialect possessed two distinct nominative forms of the 3s independent pronoun, 3ms ‫הוא‬ and 3fs ‫היא‬, but only one oblique form ‫[ הוא‬huʾā] (< *huʾā̌ t). Tropper surmised that it was the more frequent common oblique form ‫[ הוא‬huʾā] that underlay the Kethiv/Qere ‫הוא‬. ִ The impetus for this suggestion came from the Old Babylonian 3cs oblique form [šuʾāti] (accusative/genitive) / [šuʾāsim] (dative) that Rendsburg had previously noted. 5 Morgenstern (2007: 49–50) reacted to Tropper’s paper in a discussion of Qumran 3ms ‫ הואה‬and 3fs ‫ היאה‬independent pronouns; he expressed skepticism over Tropper’s solution, preferring to interpret the Kethiv of the Masoretic Text at face value as a common pronoun. In the same year that Morgenstern wrote, Cohen (2007: 113–15, 311–13) also touched on the subject in a general work on Kethiv/Qere. He reconstructed the following stages in the development of the 3fs pronoun on the basis of orthography and vocalization: *hiʾa-tu > *hiʾat > *hiʾa > *hiwa > *hiya > *hiy > hî. According to Cohen, the Kethiv ‫ הוא‬reflected the fourth stage, in which the 3ms form was realized as [hiwa]. He did not elaborate on what might have motivated the shift of ʾ > w. 3.  The Phenomenon of Kethiv/Qere and the Crystallization of the Pentateuch The most recent treatment of the Kethiv/Qere phenomenon by Ofer (2008– 9) reaches the conclusion that the Qere is not the result of late editorial activity on the part of the Tiberian Masoretes but instead, an older tradition that was passed on orally and only written down at the time when the vowel and accent signs were also given graphic representation. Ofer points out the existence of parallel traditions of biblical words—one written and the other read aloud— that are mentioned in the Talmud (see also Tov 2001: 59; Yeivin 2003: 56–57). Furthermore, he stresses that the antiquity of both the Kethiv and the Qere traditions is demonstrated in the Septuagint and the biblical Dead Sea Scrolls, where sometimes one finds a reading similar to the Kethiv of the Masoretic Text, and, at other times, a reading that is similar to the Qere. 6 The distribution of Kethiv/Qere ‫ ִהוא‬appears to be a diachronic feature of Biblical Hebrew, since it is limited, with three exceptions (n. 3 above) to the 5.  Tropper (2001: 162) also observed that there are additional common forms: the 3cs [šâti]/[šâši(m)] as well as the 2cs [kâti]/[kâšim]. 6.  See before him Kutscher 1974: 519–22 and Yeivin 2003: 57. Tov (2004: 217–18) noted that there are no Kethiv/Qere notations in any of the Dead Sea Scrolls.

174

Steven E. Fassberg

Pentateuch. According to most biblical scholars, the Pentateuch was the first section of the Hebrew Bible to have crystallized since reference to it (‫תורת‬ ‫ )משה‬is thought to be found in both the Prophets and the Writings. 7 Several scholars date the crystallization of the Pentateuch to the period of Ezra in the mid-fifth century b.c.e., basing their dating on, among other things, internal biblical evidence such as Neh 8:1 8 and the fact that, by the time of the Samaritan schism, the Samaritans had already accepted pentateuchal authority.  9 If the crystallization of the Pentateuch does indeed predate that of the Prophets and the Writings, then a linguistic feature attested only in the consonantal text of the Pentateuch logically reflects an older stage of Hebrew than the stage found in the consonantal text of later books. In addition to the pentateuchal Kethiv/Qere ‫הוא‬, ִ one also finds another uniquely pentateuchal Kethiv/Qere, ‫ֲר‬ ָ ‫‘ נַע‬youth, servant’, which is attested 22 times as against 1 occurrence of the orthography ‫ֲרה‬ ָ ‫( נַע‬Deut 22:19); in the later books of the Hebrew Bible, the consonantal and vocalized texts agree: ‫ֲרה‬ ָ ‫נַע‬. 10 As noted above (§2), Lambert believed that the pentateuchal orthography reflected an epicene form [naˤar] as did ‫[ הוא‬hū]. 4.  Possible Phonetic Realizations of the 3fs Kethiv ‫אוה‬ In light of the 3fs pronoun [hī] in other Hebrew corpora, the majority of scholars believe that the consonantal orthography of the 3fs pronoun ‫ הוא‬in the Masoretic Text does not reflect the phonetic realization of the form.  11 Only a small minority (§2 above) have interpreted the Kethiv of ‫ הוא‬as indicating that, during the period before the crystallization of the Pentateuch, there was in 7.  The work of Ryle (1892) has been influential in establishing this idea. Barton (1996: 68) summarizes the current consensus: On the face of it there is agreement among scholars on only one matter concerning the canonization of the Hebrew Scriptures: that the present threefold division into Law (tōrâh), Prophets (nbîʾîm) and Writings (ktûbîm) provides a rough guide to the relative date at which these collections were regarded as “canonical scripture.” For recent thoughts on the biblical canonization process, see Chapman 2003. 8.  ‫ויאפסו כל־העם כאיש אחד אל־הרחוב אשר לפני שער־ המים ויאמרו לעזרא הספר להביא את־‬ ‫‘ ספר תורת משה אשר־צוה ה’ את־ישראל‬the entire people assembled as one man in the square before the Water Gate, and they asked Ezra the scribe to bring the scroll of the Teaching of Moses with which the Lord had charged Israel’. 9.  Nonetheless, see Barton 1969: 8: “Now it is initially puzzling that such a simple consensus turns out to conceal very wide divergences of opinion, once we move from relative to absolute dating.” 10.  The word ‫ נַעֲרָה‬is attested 25 times in the Prophets and Writings as a common noun in addition to 3 occurrences as a proper noun (1 Chr 4:5, 6 [2×]). Cohen (2007: 27–29) suggested that ‫ נער‬reflects [naˤara] whereas ‫ נערה‬represents a later pronunciation [naʿarā] with lengthening of the final vowel. He adduces no evidence to substantiate his claim. 11. Scholars are divided over whether or not the orthography ‫ היאה‬in the Dead Sea Scrolls was realized [hīʾa] or was merely a lengthened, festive written form. For bibliography and a recent discussion, see Morgenstern 2007: 49–52.

The Kethiv/Qere ‫הוא‬, ִ Diachrony, and Dialectology

175

at least one dialect of Biblical Hebrew a 3c pronoun [hū]. As previously mentioned, objections to a common pronoun [hū] have been based on the belief that the neutralization of the gender in the 3s pronoun is unattested in Semitic languages. Rendsburg and Tropper, however, both pointed out that neutralization is attested in the Old Babylonian 3cs oblique [šuʾāti]/[šuʾāšim], and Tropper further noted the examples of the Old Babylonian 3cs [šâti]/[šâši(m)] and 2cs [kâti]/[kâšim]. Prima facie, two additional interpretations of the data come to mind. The first is that the pentateuchal 3fs Kethiv ‫ הוא‬reflects a form [*hīw 12] (< *hīwa [hūw]), which was then extended from the 3ms pronoun to the 3fs (*hī + w), creating an opposition of 3ms [*hūw] and 3fs [*hīw]. Although such a distinction is unattested elsewhere in Hebrew, it is not impossible: innovation and analogy have frequently reshaped Semitic independent pronouns. For example, in the Neo-Aramaic village dialects of the Ṭuroyo cluster (province of Mardin, southeastern Turkey) and in the dialect of nearby Mlaḥsô, the independent pronouns 3ms [hīye] and 3fs [hīya] are found. 14 Both pronouns are neologisms based on the fs *hī(y)-, to which were added the corresponding pronominal suffixes -e (3ms) and -a (3fs). 15 5.  A Noteworthy Kethiv/Qere Parallel in Biblical Aramaic There is another set of relevant pronominal Kethiv/Qere readings in the Old Testament that have not been brought into previous discussions: the 3fp pronouns in Biblical Aramaic. As demonstrated by Ben-Ḥayyim (1951), the consonantal text of Biblical Aramaic on the whole lacks 3fp pronominal and verbal forms. Of the 15 vocalized forms of the 3fp in the book of Daniel, the consonantal text in 11 cases has a corresponding 3mp form: 12.  Cohen’s reconstruction (2007: 113–15) of a final short vowel in *hiwa is problematic since all final vowels should have fallen away already at an early stage of Proto-Hebrew. Morag (1954; 1988: 156–57) argued that the final vowel in the Qumran forms ‫ הואה‬and ‫היאה‬ was a reflex of the oblique forms *hūʾat, *hīʾat. Qimron (1986: 57 n. 56), on the other hand, thinks the final vowel may be a relic of the Proto-Semitic short vowel, as does Morgenstern (2007: 51). 13.  One expects a yod glide instead of waw when contiguous to the vowel ī, as in Classical Arabic ‫ ِه َي‬. For a discussion of waw glides in Hebrew, see Qimron 2001. 14.  Jastrow 1985: 33; 1994: 28. For a general inventory of forms in Semitic, see Barth 1913: 1–22. 15.  This was suggested by Nöldeke (1881: 225) for the Midyat forms of the Ṭuroyo cluster: [hūwe] < hū + -e and [hīya] < hī + -a. Jastrow (1990: 94) believes that the existence of [hūwe] and [hīye] in Mardin Arabic contributed to the preservation of the Midyat forms. Note that the 3 cp independent pronoun in Mlaḥsô was also created on the basis of the fs: hīyĕn.

176

Steven E. Fassberg

Pronominal suffixes: ‫‘ ִמ ְּנהֵון‬of them’, Dan 2:33 (2×); ‫ ִמ ְּנהֵון‬2:41 (2×); ‫ ִמ ְּנהֵון‬2:42; ‫‘ ּבֵיּנֵיהֵון‬between them’, 7:8; ‫ָּלהֵון‬ ְ ‫‘ ּכ‬all of them’ 7:19 Verbs: ‫ֻרון‬ ָ ‫ ְיד‬Dan 4:9; ‫ ְנפ ַָקו‬5:5; ‫ֲק ָרו‬ ָ ‫ֶתע‬ ְ ‫ א‬7:8; ‫ ְנ ַפלָו‬7:20.

In two of the remaining examples, although the vocalization reflects the 3fp, Ben-Ḥayyim thought that the consonantal text allowed for masculine forms, similar to the examples cited above: ‫ִׁש ְּכנָן‬ ְ ‫‘ י‬they will dwell’, Dan 4:18 (defective orthography for ‫ִׁש ְּכנֻן‬ ְ ‫ )י‬and ‫ֶויָן‬ ְ ‫‘ ֶלה‬they will be’, Dan 5:17 (waw/yod confusion for underlying ‫) ֶלהֱֹון‬. In the editions to which he had access, BenḤayyim also found a Kethiv/Qere with the 3fp independent pronoun ‫ ִאּנִון‬in Dan 7:17. 16 Moreover, he pointed out that there were additional examples of the 3mp pronoun for an expected 3fp, for which, surprisingly, there was no 3fp Qere. 17 Ben-Ḥayyim believed that Biblical Aramaic was not the only corpus that lacked 3fp forms. A glance at the Official Aramaic material from Elephantine as well as the later Middle Aramaic inscriptional material from Palmyrene and Nabatean known at the time convinced him that the same situation obtained in these corpora. He considered this to be the case also for Old Aramaic as reflected in the inscription from Sujin (Sefire). 18 New inscriptions in Official Aramaic, Nabatean, and Palmyrene have not changed the validity of Ben-Ḥayyim’s description. 19 The same is not true for Old Aramaic, however, 16.  This is also the reading in BDB, 1081b. Codex Leningrad B19a reads the fp ‫אּנִין‬. ִ 17.  ‫לֹוהי ִדּי פ ְַר ְזלָא ְוח ְַס ָפּא ְוהַדֵ ֶּקת‬ ִ ‫ּומחָת ְלצ ְַלמָא עַל־רַ ְג‬ ְ ‫ְת עַד ִדּי ִה ְת ְּגזֶרֶת ֶאבֶן ִדּי־לָא ִבידַ יִן‬ ָ ‫ה ַוי‬ ֲ ‫ָחז ֵה‬ ‫‘ ִה ְּמֹון‬As you looked on, a stone was hewn out, not by hands, and struck the statue on its feet of iron and clay and crushed them’ (Dan 2:34; ‫ ִהּמֹון‬resumes the f.du. ‫לֹוהי‬ ִ ‫הוֵית עַד ;)רַ ְג‬ ֲ ‫ָחז ֵה‬ ‫ּי־ּמִריטּו ַג ַפּיּה‬ ְ ‫‘ ִד‬As I looked on, its wings were plucked off’ (Dan 7:4; ‫ ְּמִריטּו‬should agree with ‫ ַג ַפּיּה‬, which is fp, as can be seen in ‫‘ ְולַּה ַג ִּפּין א ְַר ַבּע‬and she had four wings’ [Dan 7:6]); ‫ּושאָר‬ ְׁ ‫ְהיבַת ְלהֹון עַד־ ְזמַן ְו ִעָּדן‬ ִ ‫ש ְל ָטנְהֹון ְוא ְַרכָה ְב ַחּיִין י‬ ָׁ ‫ֶע ִדּיו‬ ְ ‫ָתא ה‬ ָ ‫‘ חֵיו‬The dominion of the other beasts was taken away, but an extension of life was given to them for a time and season’ (Dan 7:12; ‫ׁש ְל ָטנְהֹון‬ ָ and ‫ ְלהֹון‬resume the fp ‫ָתא‬ ָ ‫)חֵיו‬. 18.  Because of the generally defective orthography, Ben-Ḥayyim admitted that he could not prove that the spelling ‎‫ן‬‎‎- in ‎th‎e Sefire inscriptions reflects the 3mp plural suffix [ūn], as he wished to argue (e.g., ‫‘ יגזרן רבוה‬may his nobles be cut in two!’ [Donner and Röllig 2002: 222, line 40]) and not a 3fp suffix [ān] (‫‘ יהינקן‬if they will give suck’ [2002: 222, lines 22–23]). Since Ben-Ḥayyim wrote, Huehnergard (1987) has collected the examples of 3mp and 3fp imperfect forms and believes that ‫ יהינקן‬reflects a jussive suffix [-nā] as opposed to a 3fp indicative [-ān]. According to Huehnergard, the latter morpheme is attested at Sefire in ‫‘ יעררן‬they will be stripped’ (Donner and Röllig 2002: 222 A, line 41) and ‫‘ יקחן‬they will be taken’ (2002: 222 A, line 42). 19.  Ben-Ḥayyim overlooked the 3fp independent pronoun in the Ashur ostracon from the mid-7th century b.c.e.: ‫‘ הצדא הני מליא אלה‬Are these words true?’ (Donner and Röllig 2002: 233, line 12; see Hug 1993: 56, 144). No 3fp pronominal forms have shown up in other Official Aramaic sources, in Nabatean, or in Palmyrene. For Egyptian Aramaic, see Muraoka and Porten 2003: §§11–12, 24. There is, however, one example of a 3fp pronominal suffix [-hen] in a Hatran inscription that differs from the 3mp suffix [-hon]: ‫‘ כולהין‬all of them’ (Beyer 1998: 103, 130).

The Kethiv/Qere ‫הוא‬, ִ Diachrony, and Dialectology

177

since there is now incontrovertible evidence for 3fp imperfect verbs as well as fp suffixed pronouns. 20 Ben-Ḥayyim was reluctant to attribute the neutralization of the 3mp and 3fp in Official Aramaic sources to any one phenomenon, though he wondered whether foreign influence was partially responsible. 21 He felt that, even though the literary language of Official Aramaic did not preserve a difference between masculine and feminine forms, the spoken vernaculars at the time must have done so since a masculine/feminine distinction is present in later Aramaic sources. 22 I believe that the Biblical Aramaic 3fp Kethiv/Qere forms offer an important and instructive parallel to the Hebrew 3fs independent pronoun Kethiv/ Qere ‫ ִהוא‬in the Pentateuch. The Biblical Aramaic phenomenon is corroborated by extrabiblical Aramaic corpora and therefore cannot be attributed to scribal error. Since the Aramaic 3fp Kethiv has been shown to represent a genuine linguistic phenomenon, there is every reason also to take the Hebrew 3fs Kethiv ‫ הוא‬as an authentic form (either [*hū] or [*hīw]) and not a form that was erroneously created by scribes who willy-nilly added a waw to the defective orthography ‫הא‬. Kethiv and Qere forms in Hebrew and Aramaic reflect old traditions. 23 6. Conclusion In light of the fetching parallel found in the 3fp pronouns and verbs of Biblical Aramaic, where both the Kethiv and Qere forms are supported by comparative evidence, and because scholarship has shown in general that Kethiv and Qere readings are antique and authentic, there is no reason to doubt the validity of the 3fs Kethiv ‫הוא‬. This orthography may reflect either [hū], a neutralization of the 3ms and 3fs pronouns, or possibly a 3fs [*hīw]. Because the 3fs Kethiv 20.  For a discussion of 3mp and 3fp forms prior to the publication of the Tell Fekherye inscription in 1981, see Degen 1969: §§35–36, 49. There are no examples in the Old Aramaic corpus of 3fp perfect forms. Attested 3fp pronominal forms include Sefire ‫ויזרע בהן‬ ‘and he sewed in them’ (Donner and Röllig 2002: 222, line 36) and Tell Fekherye ‫‘ מת כלן‬the lands, all of them’ (2002: 303, lines 3–4); the 3mp pronominal suffix occurs with final -m. 21.  He must have had Persian in mind, and in this respect anticipated the approach of Rendsburg, who also looked to a non-Semitic substratum for the cause of neutralization. 22.  Ben-Ḥayyim went on to comment that the Mishnaic Hebrew reflected in manuscripts, as opposed to that of printed editions, evidences a similar neutralization of plural masculine and feminine forms (in both perfect and imperfect verbs, and pronouns with final -n), and he hesitatingly suggested that it should be attributed to the influence of Official Aramaic. He acknowledged that, if true, this meant pushing back the roots of Mishnaic Hebrew to an early period. 23.  As noted above (§3), the Kethiv and Qere traditions are paralleled already in the Dead Sea Scrolls and are mentioned in the Talmud. For a discussion of the antiquity of two Kethiv/Qere traditions (pronominal suffixes and II-w/y participles) in Biblical Aramaic, see Fassberg 1989.

178

Steven E. Fassberg

‫ הוא‬is for all intents and purposes limited to the Pentateuch, and because the Pentateuch crystallized earlier than the Prophets and the Writings, one must deduce that the Kethiv ‫ הוא‬is evidence for an early dialectal form that later disappeared in Biblical Hebrew. References

Barth, J. 1913 Die Pronominalbildung in den semitischen Sprachen. Leipzig: Hinrichs. Barton, J. 1996 The Significance of a Fixed Canon. Pp. 67–82 in vol. 1/1 of Hebrew Bible, Old Testament: The History of Its Interpretation, ed. M. Saebø. 2 vols. Göttingen: Vandenhoeck & Ruprecht. Ben-Ḥayyim, Z. 1951 The Third Person Feminine Plural in Old Aramaic. ErIsr 1 (Schwabe volume): 135–39. [Hebrew] Beyer, K. 1998 Die aramäischen Inschriften aus Assur, Hatra und dem übrigen Ostmesopotamien. Göttingen: Vandenhoeck & Ruprecht. Chapman, S. B. 2003 How the Biblical Canon Began: Working Models and Open Questions. Pp. 29–51 in Homer, the Bible, and Beyond: Literary and Religious Canons in the Ancient World, ed. M. Finkelberg and G. G. Stroumsa. Leiden: Brill. Cohen, M. 2007 The Kethib and Qeri [sic] System in the Biblical Text: A Linguistic Analysis of the Various Traditions Based on the Manuscript ‘Keter Aram Tsova’. Jerusalem: Magnes. [Hebrew] Cross, F. M. 1998. The Stabilization of the Canon of the Hebrew Bible. Pp. 219–29 in From Epic to Canon: History and Literature in Ancient Israel. Baltimore: Johns Hopkins University Press. Degen, R. 1969 Altaramäische Grammatik der Inschriften des 10.–8. Jh. v. Chr. Abhandlungen für die Kunde des Morgenlandes 38/3. Mainz: Deutsche Morgenländische Gesellschaft. Donner, H., and Röllig, W. 2002 Kanaanäische und aramäische Inschriften, vol. 1/5: Erweiterte und überarbeitete Auflage. Wiesbaden: Harrassowitz. Emerton, J. A. 2000 Was There an Epicene Pronoun hūʾ in Early Hebrew? JSS 45: 267–76. Fassberg, S. E. 1989 The Origin of the Ketib/Qere in the Aramaic Portions of Ezra and Daniel. VT 39: 1–12. Green, W. H. 1865 A Grammar of the Hebrew Language. New York: Wiley.

The Kethiv/Qere ‫הוא‬, ִ Diachrony, and Dialectology

179

Huehnergard, J. 1987 The Feminine Plural Jussive in Old Aramaic. ZDMG 137: 266–77. Hug, V. 1993 Altaramäische Grammatik der Texte des 7. und 6. Jh.s v.Chr. Heidelberger Studien zum Alten Orient 4. Heidelberg: Heidelberger Orientverlag. Jastrow, O. 1985 Laut- und Formenlehre des neuaramäischen Dialekts von Mīdin im Ṭūr ʿAbdīn. 3rd ed. Wiesbaden: Harrassowitz. 1990 Personal and Demonstrative Pronouns in Central Neo-Aramaic: A Comparative and Diachronic Discussion Based on Turoyo and the Eastern Neo- Aramaic Dialect of Hertevin. Pp. 89–103 in Studies in Neo-Aramaic, ed. W. Heinrichs. HSS 36. Atlanta: Scholars Press. 1994 Der neuaramäische Dialekt von Mlaḥsô. Semitica Viva 14. Wiesbaden: Harrassowitz. Kutscher, E. Y. 1974 The Language and Linguistic Background of the Isaiah Scroll (1Q Isaa). STDJ 6. Leiden: Brill. Lambert, M. 1946 Traité de grammaire hébraïque. Paris: Presses universitaires de France. Morag, S. 1954 The Independent Pronouns of the Third Person Masculine and Feminine in the Dead Sea Scrolls. ErIsr 3 (Cassuto volume): 166–69. [Hebrew] 1988 Qumran Hebrew: Some Typological Observations. VT 38: 148–64. Morgenstern, M. 2007 The System of Independent Pronouns at Qumran and the History of Hebrew in the Second Temple Period. Pp. 44–63 in vol. 1 of Shaʿarei Lashon: Studies in Hebrew, Aramaic, and Jewish Languages Presented to Moshe Bar-Asher, ed. A. Maman, S. E. Fassberg, and Y. Breuer. 3 vols. Jerusalem: Bialik Institute. [Hebrew] Muraoka, T., and Porten, B. 2003 A Grammar of Egyptian Aramaic. 2nd ed. HO 32. Leiden: Brill. Nöldeke, T. 1881 Review of Der neu-aramäische Dialekt des Ṭûr ˤAbdîn von Eugen Prym und Albert Socin. ZDMG 35: 218–35. Ofer, Y. 2008–9  Ketiv and Qere: The Phenomenon, Its Notation, and Its Reflection in Early Rabbinic Literature. Leš 70: 55–73; 71: 255–79. [Hebrew] Qimron, E. 1986 The Hebrew of the Dead Sea Scrolls. HSS 29. Atlanta: Scholars Press. 2001 ‫ו לסימון הגה מעבר‬′′‫וי‬. Pp. 362–75 in Homage to Shmuel: Studies in the World of the Bible, ed. Z. Talshir, S. Yona, and D. Sivan. Jerusalem: Bialik Institute/ Beer-sheva: Ben-Gurion University of the Negev Press. [Hebrew] Rendsburg, G. 1980 Late Biblical Hebrew and the Date of “P.” JANES 12: 65–80. 1982 A New Look at Pentateuchal HWʾ. Bib 63: 351–69.

180

Steven E. Fassberg

Ryle, H. E. 1892 The Canon of the Old Testament: An Essay on the Gradual Growth and Formation of the Hebrew Canon of Scripture. London: Macmillan. Tov, E. 2001 Textual Criticism of the Hebrew Bible. 2nd ed. Minneapolis: Fortress / Assen: Van Gorcum. 2004 Scribal Practices and Approaches Reflected in the Texts Found in the Judean Desert. STDJ 54. Leiden: Brill. Tropper, J. 2001 Das genusindifferente hebräische Pronomen HWʾ im Pentateuch aus sprachvergleichender Sicht. ZAH 14: 159–72. Yeivin, I. 1985 The Hebrew Language Tradition as Reflected in the Babylonian Vocalization. Jerusalem: Academy of the Hebrew Language. [Hebrew] 2003 The Biblical Masorah. Studies in Language 3. Jerusalem: Academy of the Hebrew Language. [Hebrew]

Discerning Diachronic Change in the Biblical Hebrew Verbal System Martin Ehrensvärd 1.  Introduction and Objectives Diachronic change took place in ancient Hebrew, just as it probably occurred in all other languages used for hundreds of years. This change is evident in the biblical texts, some of which are written in Early Biblical Hebrew (EBH), and some of which are written in Late Biblical Hebrew (LBH). However, it seems that at least some later authors and/or scribes were able to write EBH, a fact that makes linguistic dating on the basis of the evidence we currently have extremely difficult or impossible (Young et al. 2008). Moreover, the extant biblical texts are not original manuscripts. In the course of transmission, substitutions of vocabulary and syntax were made in the manuscripts and many of these substitutions affected the features that we use to decide whether a text is LBH or EBH, making the linguistic dating of Hebrew biblical texts more problematic. In this essay, I shall look at several verbal features used to date texts linguistically, showing that, when analyzed closely, the specific feature is not an unambiguous chronological indicator, as has been assumed. Consequently, use of the feature for diachronic analysis merits more caution than might otherwise seem necessary. 1 2.  Prefix Conjugation Forms of the Second-Person Masculine Singular with Simple Waw (Wetiqtol) In a 2005 article, Jan Joosten provides an analysis of the differences in frequency between the EBH and LBH use of second-person masculine singular Author’s note:  Some of the data and discussion in this paper were prepared in conversation with Robert Rezetko and therefore express our joint research and conclusions. Thanks are also due to Ian Young and Søren Holst for reading and commenting on the article. Any mistakes are mine. 1. For discussions of the diachronic aspects of many verbal features, see especially Eskhult 1990, 2000, 2003, 2005; and Joosten 1989, 1992, 1999, 2002, 2005, 2006, 2007. For an alternative look at many of these features and many more, see Ehrensvärd 1999, 2003, 2006; Rezetko 2003, 2007; Young 2003; and Young et al. 2008.

181

182

Martin Ehrensvärd

prefix-conjugation forms with simple waw. These forms are rare since usually either weqataltá or a waw plus imperative is used instead. In other words, the same information is conveyed by all 3 constructions, and the latter 2 were preferred. 2 Joosten presents 2 examples in EBH prose (Exod 19:3, ‫כה תאמר לבית‬ ‫‘ יעקב ותגיד לבני ישראל‬Thus shall you say to the house of Jacob and declare to the children of Israel’; and Num 17:25) and 9 examples in LBH (for instance, Dan 9:25, ‫‘ ותדע ותשכל‬You must know and understand’; the remaining examples are listed in the following analysis). 3 This difference in frequency is striking, especially since the core EBH corpus of Genesis–Kings is nearly 3 1⁄2 times larger than the core LBH corpus of Esther–Chronicles. Further, Joosten (2005: 331–34) argues that this disparity is understandable in terms of the LBH use of verbs where, he argues, the jussive can be used in nonvolitive statements just like weqatal and hence can take its place, as in wetiqtol. These facts seem to provide evidence in favor of what many scholars regard as the change or collapse/deterioration of the EBH tense system in LBH (Ehrensvärd 2003: 172–74, with additional sources). Joosten’s research here, as always, is inspiring. However, it is notoriously difficult to make sense of the Biblical Hebrew verbal system, and any general theory must usually account for a large number of exceptions to the rule. Therefore, caution is necessary, especially when dealing with fairly uncommon features. In fact, the feature to which Joosten points is less common in LBH than his figures suggest. He writes that there are 9 instances in LBH, but he only mentions 8 of them, and 8 is in fact all there are: Daniel has 4 instances in 2 verses (Dan 9:25; 12:13), and Chronicles has 3 instances in 2 verses (2 Chr 20:9, 20). To these, Joosten adds Neh 9:28, but this verse resembles poetry and is laid out as poetry in BHS. If this verse is to be included, there are at least 36 poetic instances that might be included, most of which are found in EBH texts. 4 The LBH attestations in 4 verses within 2 LBH books, to which an additional attestation in Neh 9:28 might be added, do not amount to a great deal of material, nor do they provide a strong distribution, especially if the construction is also found in 2 verses in 2 different EBH texts. The total number of data are 8 instances in LBH and 2 in EBH. The total is small, and the distribution hardly definitive, especially if poetic texts are included. Nevertheless, the feature of 2ms

2. Although Joosten discovered this, it was noted independently by Blum 2008. 3. However, Blum (2008: 110–13) argues that ‫ ותגיד‬is to be understood as x-tiqtol—in other words, understanding a ‫ כה‬between ‫ ו‬and ‫)וכה תגיד =( תגיד‬. This possibility cannot be excluded. 4.  2 Kgs 19:25; Ps 2:12; 5:12; 7:10; 31:4; 50:15; 65:5; 71:2, 21; 104:30; 138:7; 144:5; Prov 5:20; Isa 1:29; 12:1; 37:26; 38:16; 40:27; 41:15, 23; 43:10; 46:5; 49:18; 58:10; 64:11; Jer 2:22; 6:27; 50:11; 51:46; Lam 3:66; 4:21; Ezek 24:27; 26:21; 32:28; Mic 6:14; 7:19.

Diachronic Change in the Biblical Hebrew Verbal System

183

prefix-conjugation verbs with simple waw may be considered an indication that the verbs were used differently in LBH, but should be treated with caution. I would like to stress that it would not be surprising to find differences between EBH and LBH in the use of verbs. But this does not alter the fact that we find a number of late texts that do not exhibit LBH traits with any greater frequency than the core EBH texts do (Young et al. 2008: passim). And the syntactic features that are characteristic of LBH are almost always a proliferation of syntactic features or usages already attested in EBH. In other words, LBH in certain respects seems to differ in its verbal usage compared with EBH, but in almost all instances the difference is one of degree. This makes it more difficult to judge whether or not a text from outside the core EBH or LBH corpus is written in LBH, since the texts with which we deal are rarely long and hence do not offer enough material for a proper analysis. In his contribution to this volume, E. Dresher makes an important point: language change happens when a new linguistic feature is introduced and used sparsely at first. It then proliferates and, at some point, the original feature barely remains in use. But this apparently did not prevent Israelite scribes from keeping the EBH tradition alive and from producing texts in this form of Hebrew in periods when texts were also produced in EBH. In other words, some scribes in the centuries following the exile seemed to know what EBH was, and they used it for genres that required this kind of language. Examples of these genres are prophetic literature (Haggai, Zechariah 1–8, Malachi, Joel 2–3, Isaiah 40–66, if these are all postexilic); any expansions of the Pentateuch; and perhaps in the writings of P (see Ben Zvi 2009; 2010). It is possible that the story of Ruth, if it is postexilic, was considered appropriate for a LBH treatment because of its chronological setting. These scribes allowed themselves to experiment in other texts; hence the uneven linguistic character of the core LBH books. They did not invent new syntax but used what they knew about EBH from the EBH texts, playing with it more, throwing in vocabulary that was sometimes innovative and was felt to be more in fashion due to its use in cognate languages and, very likely, in the vernacular; or maybe it was not a matter of scribal experimentation but simply a matter of using a new and less-restrained style when writing texts that fell outside the above-mentioned genres. 3.  Lamed + Infinitive Construct Mats Eskhult (2005: 359–60) argues, as do some other scholars, that ‫ל‬-qetol increases in LBH when compared with EBH. The main arguments for this view are the increased frequency of ‫ל‬-qetol in Chronicles when compared with Samuel–Kings, including substitutions of ‫ל‬-qetol in synoptic Chronicles for other verb forms in synoptic Samuel–Kings, and the trajectory of increased frequency of ‫ל‬-qetol in postbiblical Hebrew, especially in Mishnaic

184

Martin Ehrensvärd Table 1.  Frequency of ‫ל‬-qetol in Biblical Books (courtesy of Robert Rezetko) Book

Number ‫ל‬-Qetol Minus ‫לאמר‬

Percentage ‫ל‬-Qetol to All Words, Minus ‫לאמר‬

Genesis

205

0.71

Exodus

191

0.80

Leviticus

90

0.52

Numbers

143

0.61

Deuteronomy

296

1.46

Joshua

100

0.68

Judges

140

0.99

Samuel

334

0.95

Kings

295

0.81

Isaiah 1–39

91

0.67

Isaiah 40–55

35

0.61

Isaiah 56–66

37

1.00

Jeremiah

244

0.81

Ezekiel

164

0.62

Hosea

16

0.50

Joel Amos Obadiah

5

0.38

19

0.67

2

0.51

Jonah

13

1.32

Micah

7

0.36

Nahum

1

0.13

Habakkuk

12

1.32

Zephaniah

7

0.67

Haggai

5

0.56

Zechariah 1–8

22

0.89

Zechariah 9–14

8

0.40

Malachi

5

0.42

Psalms

161

0.63

Job

46

0.42

Proverbs

62

0.69

Ruth

23

1.26

6

0.35

93

2.19

Song of Songs Qoheleth

Diachronic Change in the Biblical Hebrew Verbal System

Lamentations

12

0.58

Esther

93

1.99

Daniel

43

1.22

Ezra

55

1.47

Nehemiah

77

0.96

Chronicles, including 1 Chronicles 1–9

406

1.13

Chronicles, excluding 1 Chronicles 1–9

398

1.31

Chronicles synoptic, including 1 Chronicles 1–9

171

1.18

Chronicles synoptic, excluding 1 Chronicles 1–9

171

1.19

Chronicles nonsynoptic, including 1 Chronicles 1–9

235

1.10

Chronicles nonsynoptic, excluding 1 Chronicles 1–9

227

1.41

Total (including only total number 406 in Chronicles)

3,564

0.84

185

Hebrew where “the infinitive construct is found only when it is preceded by a ‫( ”ל‬Kutscher 1974: 41). 5 Nonetheless, insofar as Biblical Hebrew is concerned, several scholars have cast doubt on this view. 6 The following facts demonstrate that the frequency of ‫ל‬-qetol fluctuates inconsistently throughout Biblical Hebrew and that the construction cannot be considered a characteristic of LBH without a good deal of caution.  7 The ‫ל‬-qetol is more frequent in Chronicles than in Samuel and Kings (see table 1). This is true for Chronicles as a whole, particularly in the nonsynoptic parts, and it is true for ‫ל‬-qetol in proportion to all qetol forms, all verbs, 5.  For other statements of this view, see Corwin 1909: 27 (§35); compare 27–32 (§§35– 41); Driver 1913: 538–39; Joüon and Muraoka 2006: 405 n. 1 (§124l); Kieviet 1997: 45–73; Kropat 1909: 72; van Peursen 2004: 257–58, 296–97; Polzin 1976: 60–61; compare 71–72, 92, 94–95, 100–101, 111, 113–14; Sáenz-Badillos 1993: 145; Verheij 1990: 60–62, 67–71, 120; Waltke and O’Connor 1990: 603 n. 22 (§36.2.2). 6.  Rendsburg 1980: 71–72; Ehrensvärd 2003: 179; Young 2003: 297–98; compare Young et al. 2008: 1.163; 2.79. Young (2003) argues that the ratio of ‫ב‬-, ‫כ‬-, and ‫ל‬-qetol forms in Hebrew inscriptions matches LBH rather than EBH. 7.  The figures in the following include all infinitive constructs with ‫ ל‬except for ‫לאמר‬ which is excluded on the grounds that it is a fixed expression. The argument holds, however, even if one does not make this functional distinction. Other forms could be excluded too, but weighing all ‫ל‬-qetol forms in the Hebrew Bible is beyond the scope of this investigation.

186

Martin Ehrensvärd

all words (= graphic units), and all ‫ב‬- and ‫כ‬-qetol forms. So, for example, 1 Kings has an average 0.81% ‫ל‬-qetol forms per 100 words (= graphic units), Samuel has 0.95, and nonsynoptic Chronicles (minus 1 Chronicles 1–9 8) has 1.41. Other scholars have rightly observed the ratios, but their conclusions about Biblical Hebrew in general, which are based mainly or only on this observation, are off the mark. Other measures must be considered as well. The frequency of ‫ל‬-qetol in nonsynoptic Chronicles (minus 1 Chronicles 1–9) as a percentage of all words (= graphic units) is very similar to the Hebrew sections of the book of Ezra (1.47 per 100), but Deuteronomy has 1.46 per 100 and is situated between these two LBH books in terms of percentages. From another angle, the difference between Samuel–Kings and nonsynoptic Chronicles (minus 1 Chronicles 1–9) is similar to the change from Leviticus to Judges, where no diachronic difference is expected. Nehemiah sits between Genesis, Exodus, and Kings, on the one hand, and Deuteronomy and Judges, on the other. Finally, it is interesting to observe the use of ‫ל‬-qetol in synoptic Samuel– Kings and Chronicles. On the one hand, on 25 occasions Chronicles has a ‫ל‬-qetol form that is parallel to another verb form in Samuel–Kings. 9 On the other hand, on 11 occasions Samuel–Kings has a ‫ל‬-qetol form that is parallel to a different verb form in Chronicles. 10 On the face of it, this seems to support the view that the increased frequency of ‫ל‬-qetol in Chronicles represents a linguistic development. Rezetko (2007) has studied four ‫ל‬-qetol appearances closely, all related to 2 Samuel 6 and synoptic Chronicles (see chart, top of p. 187). On the basis of literary and text-critical evidence, Rezetko (2007: 124–26, 149, 151–53, 176–77) argues that the ‫ל‬-qetol forms in 1 Chr 13:9, 12 are primary, whereas the ‫ל‬-qetol forms in 2 Sam 6:10 and 1 Chr 15:25 are secondary. Thus, for example, in 2 Sam 6:9, 4QSama and the LXX of both Samuel and Chronicles reflect the infinitive, all opposed to MT Samuel’s wayyiqtol form. A similar argument can be made for the wayyiqtol form in 2 Sam 6:6. In other 8.  In contrast to Verheij 1990, I do not include 1 Chronicles 1–9 here (despite the fact that the instances in these chapters constitute 15% of 1 and 2 Chronicles), since there are only 25 qetol forms there, and only 10 of those are ‫ל‬-qetol forms. 9.  2 Sam 5:11 // 1 Chr 14:1; 2 Sam 6:6 // 1 Chr 13:9; 2 Sam 6:9 // 1 Chr 13:12; 2 Sam 6:12 // 1 Chr 15:25; 2 Sam 7:12 // 1 Chr 17:11; 2 Sam 7:27 // 1 Chr 17:25; 2 Sam 7:29 // 1 Chr 17:27; 2 Sam 10:3 (qetol) // 1 Chr 19:3 (‫ל‬-qetol); 2 Sam 10:6 // 1 Chr 19:6; 2 Sam 24:1 // 1 Chr 21:1; 2 Sam 24:18 // 1 Chr 21:18; 2 Sam 24:18 // 1 Chr 21:18; 1 Kgs 5:30 // 2 Chr 2:17; 1 Kgs 6:1 // 2 Chr 3:1; 1 Kgs 8:29 // 2 Chr 6:20; 1 Kgs 8:64 (qetol)// 2 Chr 7:7 (‫ל‬-qetol); 1 Kgs 18:33 // 2 Chr 32:13; 1 Kgs 22:49 // 2 Chr 20:36; 2 Kgs 12:6 // 2 Chr 24:5; 2 Kgs 12:12 // 2 Chr 24:12; 2 Kgs 14:10 // 2 Chr 25:19; 2 Kgs 18:19 // 2 Chr 32:9; 2 Kgs 18:35 // 2 Chr 32:14; 2 Kgs 18:35 // 2 Chr 32:14; 2 Kgs 22:9 // 2 Chr 34:16. 10.  2 Sam 6:10 // 1 Chr 13:13; 2 Sam 10:11 // 1 Chr 19:12; 2 Sam 24:4 // 1 Chr 21:4; 2 Sam 24:21 // 1 Chr 21:22; 2 Sam 24:21 // 1 Chr 21:22; 1 Kgs 3:9 // 2 Chr 1:10; 1 Kgs 3:11 // 2 Chr 1:11; 1 Kgs 5:19 // 2 Chr 2:3; 1 Kgs 7:16 // 2 Chr 3:16; 1 Kgs 8:52 // 2 Chr 6:40; 1 Kgs 12:24 (‫ל‬-qetol) // 2 Chr 11:4 (‫מן‬-qetol).

Diachronic Change in the Biblical Hebrew Verbal System

187

2 Sam 6:6

‫ויעחז‬

‫לעחז‬

1 Chr 13:9

2 Sam 6:9

‫ויאמר‬

‫לאמר‬

1 Chr 13:12

2 Sam 6:10

‫ולא־אבה דוד להסיר‬

‫ולא־הסיר דויד‬

1 Chr 13:13

2 Sam 6:12

‫וילך דוד ויעל‬

‫ויהי דויד וזקני ישראל ושרי‬ ‫האלפים ההלכים להעלות‬

1 Chr 15:25

words, the differences between these texts in terms of the presence or absence of ‫ל‬-qetol do not seem to be linguistic. Consequently, it is a mistake to argue that the specific ‫ל‬-qetol forms in 1 Chr 13:9, 12; 15:25 are an indication of a more general diachronic linguistic development in Biblical Hebrew. Many core EBH and core LBH books fit the standard chronological model of Biblical Hebrew in that the core LBH books use ‫ל‬-qetol more often than the core EBH books. Nevertheless, as observed above (compare Deuteronomy and Nehemiah), it is a mistake to generalize from this observation to the rest of the Hebrew Bible. This is well illustrated by the books of Haggai, Zechariah, and Malachi. These three books, dated to the early postexilic period on the basis of internal references, exhibit very different trends in the rates of occurrence of ‫ל‬-qetol. If we look at ‫ל‬-qetol forms as a percentage of all words in all biblical books, Haggai, Zechariah, and Malachi individually are more conservative in this regard than all the books of the Pentateuch and Former Prophets with the exceptions of Leviticus and Numbers. Moreover, postexilic Joel stands nearly at the bottom of the scale, whereas preexilic Habakkuk, which is important because it has much more evidence, stands near the top of the scale (with, for example, Jonah and Chronicles), unlike most of the other books of the Twelve. In conclusion, increased use of ‫ל‬-qetol is not a linguistic characteristic of all the core LBH books, and the usage rates of this feature do not support the presupposition of the traditional diachronic model of Biblical Hebrew that has interpreted the distribution pattern to contrast between preexilic EBH and postexilic LBH books. 4. Passive Infinitive Construct with Prepositive Subject Frank Polak (2006: 127) mentions 4 examples of a passive infinitive construct with a prepositive subject 11 in postexilic books (Hag 1:2; Zech 8:9; Esth 8:13; 9:1) and says: Although I have not found any parallels in the Aramaic documents of this period, in Akkadian, constructions of this type are not uncommon when the infinitive is 11.  Note that this is a different issue from an infinitive with prepositive object (see Young et al. 2008: 2.127–28).

188

Martin Ehrensvärd

an intransitive verb, for example, libbī ana marāsim bēlī lā inaddin (‘My Lord should not let my heart ache,’ ARM X 141, 30). Thus the inference is justified that in Hebrew this construction reflects the administrative register of the Achaemenid period.

To Polak’s list, I add Esth 3:14 and Josh 5:8: Hag 1:2

‘the house of Yahweh to be built’

Zech 8:9

‘the temple to be built’

Esth 3:14

‘the copy of the writing to be given’

Esth 8:13

‘the copy of the writing to be given’

Esth 9:1

‘the time drew near when the king’s command and decree were to be executed’

Josh 5:8

‘all the nation was finished being circumcised’a

‫בית יהוה להבנות‬ ‫ההיכל להבנות‬ ‫פתשגן הכתב להנתן‬ ‫פתשגן הכתב להנתן‬ ‫הגיע דבר־המלך ודתו‬ ‫להעשות‬ ‫תמו כל־הגוי להמול‬

a.  The Niphal of ‫ מול‬here can be read as passive or reflexive.

Next, I present two similar constructions, in which the verbs are unmarked morphologically because the Niphal is not used: Josh 2:5

‘and it was (time for) the gate to be shut (Qal)’

Esth 7:4

‘for we have been sold, I and my people, to be exterminated (Hiphil), to be killed (Qal), and to be destroyed (Piel)’

‫ויהי השער לסגור‬ ‫נמכרנו אני ועמי‬ ‫להשמיד להרוג ולאבד‬

Although they are unmarked morphologically, since the Niphal is not used, these are examples in which the subject is prepositive, and the infinitive construct has a passive sense. 12 Given this distribution, the passive infinitive construct with prepositive subject probably cannot be considered an LBH linguistic feature, because it is not typical of the core LBH books of Esther-Chronicles, being attested in only 2 different phrases (‫ פתשגן הכתב להנתן‬in Esth 3:14/8:13 and ‫הגיע דבר־המלך‬ ‫ ודתו להעשות‬in Esth 9:1), and since it is also attested in core EBH in Josh 5:8. 5. Predicative Infinitive Absolute The predicative use of the infinitive absolute means the use of the infinitive absolute as a finite verb other than an imperative. This usage is well attested in early Northwest Semitic texts. According to the standard chronological paradigm, this usage increases in LBH. However, it is rare in Qumran Hebrew and 12.  These examples are cited in Joüon and Muraoka 2006: 409 (§124t); Waltke and O’Connor 1990: 603 (§36.2.1f).

Diachronic Change in the Biblical Hebrew Verbal System

189

does not occur in Mishnaic Hebrew (see, for instance, Eskhult 2000; Joüon and Muraoka 2006: 400; Fassberg 2008). 13 In the following presentation of the relevant data, I distinguish between what I consider unambiguous examples of the phenomenon and ambiguous examples. Here is a suggestion as to which infinitives absolute to include in this category: Gen 41:4; Exod 8:11; 36:7; Lev 6:7; 25:14; Num 6:5; 15:35; 30:3; Deut 3:6; 14:21; 15:2; Josh 9:20; Judg 7:19; 1 Sam 1:9; 2:28; 22:13; 25:26, 33; 1 Kgs 9:25; 22:30 (2×); Isa 5:5 (2×); 22:13 (3×); 31:5; 37:19; 42:20 (Qere), 20, 22, 24; 57:17; 58:6 (4×), 7; 59:4 (4×); Jer 3:1; 4:18; 7:9 (6×), 18; 8:15; 13:16 (Qere); 14:5, 19; 19:13; 22:14, 19 (2×); 23:14 (2×); 31:2; 32:33, 44 (3×); 36:23; 37:21; 44:17, 18, 19; Ezek 1:14 (2×); 7:14; 23:30, 47; 36:3; Hos 4:2 (5×); 10:4 (2×); Mic 6:13; Hab 2:15; 3:13; Hag 1:6 (4×), 9; Zech 3:4; 7:5; 12:10; 14:12; Ps 17:5; 35:16; 65:11 (2×); Job 15:3; 15:35 (2×); 26:9; 40:2; Prov 12:7; 13:20; 15:22; 17:12; Qoh 4:2; 7:25; 8:9; 9:11; Dan 9:5, 11; Esth 2:3; 3:13; 6:9; 8:8; 9:1, 6, 12, 16 (3×); 9:17 (2×), 18 (2×); Neh 8:8; 9:8, 13; 1 Chr 5:20; 15:22; 16:36; 2 Chr 7:3; 18:29 (2×); 28:19; 31:10 (3×)—altogether 142 instances. Uncertain examples are, for example, 1  Kgs 22:30 (2×), which are here counted as predicative infinitives absolute and included in the list; and 2 Kgs 4:43 (2×), which are counted as substituting for imperatives and excluded from the list. Note the following synoptic material: Kings plus/nonsynoptic: 1 Kgs 9:25; Chronicles nonsynoptic: 1  Chr 15:22; 2  Chr 7:3; 28:19; 31:10 (3×); Kings/Chronicles synoptic: 1  Kgs 22:30 (2×)  // 2  Chr 18:29 (2×); Psalms/ Chronicles synoptic: Ps 106:48 (imperative) // 1 Chr 16:36. What stands out in this survey of the data is the following: the predicative infinitive absolute is used most often in Jeremiah (29×, not a core LBH text), and then in postexilic Esther (14×), Second Isaiah (13×; 8× in First Isaiah) and Chronicles (10×). However, similar to some EBH prophetic books (Amos, Obadiah, Nahum, Zephaniah), there are also no predicative infinitives absolute in the books of Joel, Jonah, Malachi, Ruth, Song of Songs, and Ezra, all of which are probably or definitely postexilic. This distribution is not what we would expect from a clear-cut LBH feature. Apparently, it was also a feature that some writers preferred and others did not, 14 although the books that use it the most are primarily texts that in the traditional view stem from the exile and after. It is hard to explain, however, 13.  Rendsburg 2002: 76–79 is another useful treatment and includes a fairly comprehensive bibliography, but his distinction between northern, southern, non-Judahite, and other uses seems ad hoc. A detailed functional analysis of the infinitive absolute as finite verb is a desideratum (see Smith 2000), but preliminary indications are that such an analysis would not affect the present argument. 14.  This is what Fassberg (2008 argues), but he sees this use as an archaizing feature when found in LBH texts.

190

Martin Ehrensvärd

why there would be so few features in Ezra, Nehemiah, and Daniel if this really were a common LBH feature. 6. Conclusion The only basis for reliable scholarship is research that takes into account all of the relevant Hebrew data—linguistic, textual, and literary. There are clear differences between EBH and LBH, but when it comes to features that are sparsely attested (wetiqtol, passive infinitive with prepositive subject), caution is necessary; and when it comes to features that are well attested (‫ל‬-qetol, predicative infinitive absolute), it is important to examine all biblical texts in order to determine the full picture before drawing conclusions about how these features should be classified and where they fit into the history of the language. References

Ben Zvi, E. 2009 The Communicative Message of Some Linguistic Choices. Pp. 269–90 in A Palimpsest: Rhetoric, Ideology, Stylistics, and Language Relating to Persian Israel, ed. E.  Ben Zvi, D.  Edelman, and F.  Polak. Perspectives on Hebrew Scriptures and Its Contexts 5. Piscataway, NJ: Gorgias. 2010 Reconstructing the Intellectual Discourse of Ancient Yehud. SR 39: 7–23. Blum, E. 2008 Das althebräische Verbalsystem: Eine synchrone Analyse. Pp.  91–142 in Sprachliche Tiefe – Theologische Weite, ed. O. Dyma and A. Michel. Biblischtheologische Studien 91. Neukirchen-Vluyn: Neukirchener Verlag. Corwin, R. 1909 The Verb and the Sentence in Chronicles, Ezra and Nehemiah: Part of a Dissertation. Borna near Leipzig: Noske. Driver, S. R. 1913 An Introduction to the Literature of the Old Testament. 9th ed. Oxford: Clarendon. Ehrensvärd, M. 1999 Negating the Infinitive in Biblical Hebrew. ZAH 12: 146–64. 2003 Linguistic Dating of Biblical Texts. Pp. 164–88 in Biblical Hebrew: Studies in Chronology and Typology, ed. I. Young. JSOTSup 369. London: T. & T. Clark. 2006 Why Biblical Texts Cannot be Dated Linguistically. HS 47: 177–89. Eskhult, M. 1990 Studies in Verbal Aspect and Narrative Technique in Biblical Hebrew Prose. Acta Universitatis Upsaliensis, Studia Semitica Upsaliensia 12. Uppsala: Upp­sala University. 2000 Verbal Syntax in Late Biblical Hebrew. Pp.  84–93 in Diggers at the Well: Proceedings of a Third International Symposium on the Hebrew of the Dead Sea Scrolls and Ben Sira, ed. T. Muraoka and J. F. Elwolde. STDJ 36. Leiden: Brill.

Diachronic Change in the Biblical Hebrew Verbal System

191

2003 Markers of Text Type in Biblical Hebrew from a Diachronic Perspective. Pp.  153–64 in Hamlet on a Hill: Semitic and Greek Studies Presented to Professor T. Muraoka on the Occasion of His Sixty-Fifth Birthday, ed. M. F. J. Baasten and W. T. van Peursen. OLA 118. Leuven: Peeters. 2005 Traces of Linguistic Development in Biblical Hebrew. HS 46: 353–70. Fassberg, S. 2008 The Infinitive Absolute as Finite Verb and Standard Literary Hebrew of the Second Temple Period. Pp.  47–60 in Conservatism and Innovation in the Hebrew Language of the Hellenistic Period: Proceedings of a Fourth International Symposium on the Hebrew of the Dead Sea Scrolls and Ben Sira, ed. J. Joosten and J.-S. Rey. STDJ 73. Leiden: Brill. Joosten, J. 1989 The Predicative Participle in Biblical Prose. ZAH 2: 128–59. 1992 Biblical Hebrew Weqatal and Syriac Hwa Qatel Expressing Repetition in the Past. ZAH 5: 1–14. 1999 Pseudo-Classicisms in Late Biblical Hebrew in Ben Sira, and in Qumran Hebrew. Pp. 146–59 in Sirach, Scrolls and Sages: Proceedings of a Second International Symposium on the Hebrew of the Dead Sea Scrolls, Ben Sira, and the Mishnah, Held at Leiden University—15–17 December 1997, ed. T. Muraoka and J. F. Elwolde. STDJ 33. Leiden: Brill. 2002 Biblical Hebrew as Mirrored in the Septuagint: The Question of Influence from Spoken Hebrew. Text 21: 1–19. 2005 The Distinction between Classical and Late Biblical Hebrew as Reflected in Syntax. HS 46: 327–39. 2006 The Disappearance of Iterative WEQATAL in the Biblical Hebrew Verbal System. Pp. 135–47 in Biblical Hebrew in Its Northwest Semitic Setting: Typological and Historical Perspectives, ed. S. E. Fassberg and A. Hurvitz. Publication of the Institute for Advanced Studies 1. Jerusalem: Magnes / Winona Lake, IN: Eisenbrauns. 2007 The Syntax of Volitive Verbal Forms in Qoheleth in Historical Perspective. Pp. 47–62 in The Language of Qohelet in Its Context: Essays in Honour of Prof. A. Schoors on the Occasion of His Seventieth Birthday, ed. A. Berlejung and P. van Hecke. OLA 164. Leuven: Peeters. Joüon, P., and T. Muraoka 2006 A Grammar of Biblical Hebrew. Translated and revised by T. Muraoka. SubBi 27. 2nd ed. Rome: Pontifical Biblical Institute. Kieviet, P. J. A. 1997 The Infinitive Construct in Late Biblical Hebrew: An Investigation in the Synoptic Parts of Chronicles. Dutch Studies on Near Eastern Languages and Cultures 3: 45–73. Kropat, A. 1909 Die Syntax des Autors der Chronik verglichen mit der seiner Quellen. BZAW 16. Giessen: Alfred Töpelmann. Kutscher, E. Y. 1974 The Language and Linguistic Background of the Isaiah Scroll (1QIsa a). STDJ 6. Leiden: Brill.

192

Martin Ehrensvärd

Peursen, W. T. van 2004 The Verbal System in the Hebrew Text of Ben Sira. Studies in Semitic Languages and Linguistics 41. Leiden: Brill. Polak, F. 2006 Sociolinguistics: A Key to the Typology and the Social Background of Biblical Hebrew. HS 47: 115–62. Polzin, R. 1976 Late Biblical Hebrew: Toward an Historical Typology of Biblical Hebrew Prose. HSM 12. Missoula, MT: Scholars Press. Rendsburg, G. A. 1980 Late Biblical Hebrew and the Date of “P.” JANES 12: 65–80. 2002 Israelian Hebrew in the Book of Kings. Occasional Publications of the Department of Near Eastern Studies and the Program of Jewish Studies, Cornell University 5. Bethesda, MD: CDL.  Rezetko, R. 2003 Dating Biblical Hebrew: Evidence from Samuel–Kings and Chronicles. Pp.  215–50 in Biblical Hebrew: Studies in Chronology and Typology, ed. I. Young. JSOTSup 369. London: T. & T. Clark. 2007 Source and Revision in the Narratives of David’s Transfer of the Ark: Text, Language and Story in 2 Samuel 6 and 1 Chronicles 13, 15–16. Library of Hebrew Bible/Old Testament Studies 470. London: T. & T. Clark. Sáenz-Badillos, A. 1993 A History of the Hebrew Language, trans. J. Elwolde. Cambridge: Cambridge University Press. Smith, M. S. 2000 The Infinitive Absolute as Predicative Verb in Ben Sira and the Dead Sea Scrolls: A Preliminary Survey. Pp. 256–67 in Diggers at the Well: Proceedings of a Third International Symposium on the Hebrew of the Dead Sea Scrolls and Ben Sira, ed. T. Muraoka and J. F. Elwolde. STDJ 36. Leiden: Brill. Verheij, A. J. C. 1990 Verbs and Numbers: A Study of the Frequencies of the Hebrew Verbal Tense Forms in the Books of Samuel, Kings, and Chronicles. SSN 28. Assen: Van Gorcum. Waltke, B. K., and O’Connor, M. 1990 An Introduction to Biblical Hebrew Syntax. Winona Lake, IN: Eisenbrauns. Young, I. 2003 Late Biblical Hebrew and Hebrew Inscriptions. Pp.  276–311 in Biblical Hebrew: Studies in Chronology and Typology, ed. I. Young. JSOTSup 369. London: T. & T. Clark. Young, I.; Rezetko, R.; and Ehrensvärd, M. 2008 Linguistic Dating of Biblical Texts. 2 vols. London: Equinox.

The Archaic System of Verbal Tenses in “Archaic” Biblical Poetry Tania Notarius 1. Introduction In recent decades, diachronic analyses of Biblical Hebrew (BH) have concentrated mostly on comparisons between Classical Biblical Hebrew (CBH) and Late Biblical Hebrew (LBH). 1 As part of this trend, studies of verbal morphosyntax have also focused on differences between these two periods (see, for example, Eskhult 2000 or Cohen 2008; important progress was made in Joosten 2005, 2006, 2007, 2009). In this essay, however, focusing on the problem of tenses, I draw scholarly attention once more to Archaic Biblical Hebrew (ABH), which is presumably attested in what is conventionally labeled “archaic” poetry (see also the recent publications of Bloch 2009 and 2010). 2 Postulating an additional stage in the development of the verbal system that is typologically distinct from both CBH and LBH contributes to the general discussion about the history of the Hebrew verb. 3 A semantic analysis of verbal tenses in the corpus of “archaic” poetry reveals quite a diverse picture. Some uses are conditioned by specific discourse modes. Other uses are consistent with CBH and LBH. One example of the 1.  See the publications of the symposium in Hebrew Studies 2005 (“Can Biblical Texts Be Dated Linguistically?” pp. 321–76) and 2006 (pp. 83–2010); and see Young 2003; see also Young, Rezetko, and Ehrensvärd 2010; and the review in Naudé 2011. As Naudé 2011 notes, the complex linguistic situation is not limited to the CBH and LBH dichotomy. 2.  See the essays by Bloch and Cohen in this volume. 3.  The term archaic poetry is used here conventionally to denote poetic passages incorporated into the prosaic books of the Torah and Former Prophets (the most significant among them are the Song of Moses, Song of the Sea, Song of Deborah, Song of David, Blessing of Jacob, Oracles of Balaam, Blessing of Moses, and Song of Hannah); see the bibliographic review in Bloch 2009: 34–37. Not all parts of the “archaic” corpus equally demonstrate the archaic type of verb system; for further discussion, see Notarius 2010a: 266–68. The archaic verb system was typologically more conservative than CBH and LBH and therefore preceded them in terms of relative chronology. In this paper, I describe several salient features of the archaic verb system. A full reconstruction of the archaic language type as well as discussion about the date of composition of the poems that consistently represent the archaic language type are outside of the scope of this paper.

193

194

Tania Notarius

latter is the use of the predicative participle in Hannah’s Song (see Notarius 2010a: 260): ‫ׁשאֹול ַוּיָעַל‬ ְ ‫מֹוריד‬ ִ ‫ּומ ַחּיֶה‬ ְ ‫יְהוָה מ ִֵמית‬ (1)  ​ The Lord kills and brings to life; he brings down to Sheol and raises up. (1 Sam 2:6) 4 Unlike this example, several texts unmistakably reveal an archaic type of verbal system: these are the Song of Deborah, the Song of the Sea, the Song of Moses (albeit there are some textual and linguistic fluctuations in the closing part of the Song), both narrative songs in 2 Samuel 22 (vv. 5–20 and 33–46), and probably select portions of the Blessings of Moses and the Blessings of Jacob (see Notarius 2007; 2010a; 2011b). Since I explored the participle and volitive modal forms elsewhere (Notarius 2010a; 2010b), in this essay I focus on the indicative forms in two discourse modes: conversation and poetic narrative. 5 In particular, I explore two phenomena characteristic of the archaic language type: (1) the use of the imperfective ‫ יקטל‬as the present progressive tense in conversation; (2) the extremely narrow semantic scope of the perfect tense ‫ קטל‬in poetic narrative. 2.  Imperfective ‫ יקטל‬as the Present Progressive Tense in Conversation Biblical poetry rarely represents events as present progressive—that is, as taking place at the very moment of speech and generally associated with conversation. As J. Joosten (1989: 128–59) has shown convincingly, in CBH the semantic category of present progressive is expressed by the predicative participle (for example, ‫ ְמ ַבּקֵׁש‬in Gen 37:16, example [3] below). I previously demonstrated that the predicative participle is not encountered in the corpus that represents the archaic language type (Notarius 2010a: 245–48). In this essay, I claim that the prefix form ‫( יקטל‬a reflex of the Old Canaanite imper-

4.  The translation of the examples follows the nrsv, unless an alternative translation is suggested. 5.  I presented the principles of the discursive and semantic analysis of verbal tenses in biblical poetry in Notarius 2008; 2010b; and 2011a; and there is no need to repeat them here. The method presupposes the discourse modes typology (C. Smith 2003), a two-component aspect theory (C. Smith 1997) that distinguishes among (1) the situation type, expressed by a verb lexeme; (2) the grammatical viewpoint aspect; and (3) the (neo)-Reichenbachian approach to the problems of time in language and text that operates with three temporal locations: Speech Time (in which the speech production is located), Reference Time (which is defined in relation to Speech Time), and Event Time (which is defined in relation to Reference Time). See Reichenbach 1947; Comrie 1985; Hornstein 1990; Giorgi and Pianesi 1997; C. Smith 2004; 2005; 2007; 2008; C. Smith, Perkins, and Fernald 2007; Demirdache and Uribe-Etxebarria 2004; and 2005.

Verbal Tenses in “Archaic” Biblical Poetry

195

fective aspect yaqtulu) is used as the present progressive tense in the archaic language type (see Rainey 1986). 2.1.  Imperfective ‫ יקטל‬as the Present Progressive Tense in ‫ הֲלֹא‬Sentences To illustrate the imperfective ‫ יקטל‬as the present progressive, I will discuss an example from the Song of Deborah: ַ ‫ֶׁשנָב מ‬ ְ ‫יסרָא ְּבעַד ָהא‬ ְ ‫ִׁש ְקפָה ו ְַּתיַּבֵב אֵם ִס‬ ְ ‫ ְּבעַד ַהחַּלֹון נ‬28 (2)  ​ ‫ַּדּוע ּבֹׁשֵׁש ִר ְכּבֹו‬ 29 ‫־היא‬‎ ִ ‫רֹותי ָה ּתַ עֲנֶיּנָה ַאף‬ ֶ ‫ׂש‬ ָ ‫ ח ְַכמֹות‬:‫בֹותיו‬ ָ ‫לָבֹוא מַּדּו ַע ֶאחֱרּו ַּפעֲמֵי מ ְַר ְּכ‬ ‫ׁשלַל‬ ְ ‫חמָתַ יִם ְלרֹאׁש ֶּגבֶר‬ ֲ ַ‫ׁשלָל רַ חַם ר‬ ָ ‫ִמ ְצאּו ְיח ְַּלקּו‬ ְ ‫הֲלֹא י‬30 :‫א ָמרֶי ָה לָּה‬ ֲ ‫ׁשיב‬ ִ ‫ָּת‬ ‫ׁשלָל‬ ָ ‫ָעים ִר ְקמָה ֶצבַע ִר ְקמָתַ יִם ְלצ ְַּוארֵי‬ ִ ‫ׁשלַל ְצב‬ ְ ‫יסרָא‬ ְ ‫ָעים ְל ִס‬ ִ ‫ְצב‬ Out of the window she peered, the mother of Sisera gazed through the lattice: “Why is his chariot so long in coming? Why tarry the hoofbeats of his chariots?” Her wise lady makes an answer, indeed, she answers the question to her: “They are surely finding and dividing the spoil: a girl or two for every man; spoil of dyed stuffs for Sisera, spoil of dyed stuffs embroidered, two pieces of dyed work embroidered for the neck as spoil.” (Judg 5:28–30) Both Fokkelman and Knauf noticed that the ‫ יקטל‬forms in v. 30 (‫ִמ ְצאּו ְיח ְַּלקּו‬ ְ ‫)י‬ express the present progressive: in the words of Knauf, “right now, they are gaining and distributing booty.” 6 These scholars, however, did not consider the syntactic and semantic status of the ‫ הֲלֹא‬sentence and did not entertain the possibility that the verbal usage here is conditioned by its seeming character as an interrogative. 7 As Joosten (2002: 54) pointed out, in interrogative sentences, ‫ יקטל‬may still preserve its present progressive meaning even in CBH, as the follow example illustrates: ִ ‫ִׁש ָאלֵהּו ה‬ ְ ‫ִמ ָצאֵהּו ִאיׁש ְו ִהּנֵה תֹעֶה ַּבּשָׂדֶה ַוּי‬ ְ ‫ַוּי‬ (3)  ​ ‫ וַּיֹאמֶר‬:‫ָאיׁש לֵאמֹר מַה ְּת ַבּקֵׁש‬ ֹ ֹ :‫ידה ּנָא ִלי אֵיפה הֵם ר ִֹעים‬ ָ ‫אֶת ַאחַי אָנ ִכי ְמ ַבּקֵׁש ה ִַּג‬ A man found him wandering in the fields; the man asked him, “What are you seeking?” He replied, “I’m looking for my brothers. Can you tell me where they are grazing their flocks?” (Gen 37:15–16) One can maintain that ‫ הֲלֹא‬in Judg 5:30 introduces a genuine or rhetorical question and is, therefore, not a legitimate example of ‫ יקטל‬for the present progressive. 6.  Knauf 2005:174: “yiqtol-L (yaqtulu) is used for the true present.” See Fokkelman 1995: 587 n. 7: “In 30a [an embedded text that quotes Canaanite ladies] yqtl is the simple present [of the speakers].” 7.  This verse is translated as a question in the nrsv, kjv, jps, and many other versions.

196

Tania Notarius

As has been argued by many scholars, ‫ הֲלֹא‬is not necessarily an interrogative but is commonly an asseverative or representative particle. 8 The arguments were summarized in two articles by A. Moshavi. 9 To begin with, Judg 5:30, being an answer to a genuine question in v. 28, is not a genuine question itself. Theoretically, it can be a negative rhetorical question, as in Ps 94:9–10: (4)  ​

:ַ‫יֹוכיח‬ ִ ‫ הֲלֹא‬. . . ‫ הֲיֹסֵר ּגֹויִם‬:‫ִׁשמָע ִאם יֹצֵר ַעיִן הֲלֹא י ִַּביט‬ ְ ‫הֲנֹטַע אֹזֶן הֲלֹא י‬ He who planted the ear, does he not hear? He who formed the eye, does he not see? He who disciplines the nations, . . . does he not chastise? (Ps 94:9–10)

For other possible examples of an interrogative use of ‫הֲלֹא‬, see Jer 3:1, Ezra 8:14; and perhaps also Ps 85:7, Mic 2:7, Prov 24:12, and Ruth 3:1. Moshavi (2007: 54) suggests syntactic criteria for disambiguating the negative-interrogative and noninterrogative ‫הֲלֹא‬, but unfortunately, no such criteria are available for the ‫הֲלֹא‬‎-verb sentences, as in Judg 5:30. She also elaborates pragmatic considerations for distinguishing between negative rhetorical and noninterrogative ‫( הֲלֹא‬see Moshavi 2011). Applying the pragmatic criteria suggested by Moshavi to Judg 5:30 reveals that in this case ‫ הֲלֹא‬is not a negative rhetorical. First, the real negative rhetorical questions in BH either imply an assertion obvious to the speaker and already known to the listener or are used in a persuasive context, marking a truth that the listener does not want to accept. However, noninterrogative ‫ הֲלֹא‬is used to introduce new information, as in announcements and predictions (see Moshavi 2011: 95–97). In my view, although Judg 5:30 contains information about the conduct of war that implies worldly knowledge shared by both speaker and audience, it is neither previously known nor persuasive. The attendant of Sisera’s mother is providing a plausible explanation for Sisera’s delay that is intended to calm the mother’s concern. 10 Second, this ‫ הֲלֹא‬clause is actually an answer to the genuine question of Sisera’s mother in v. 28 and, therefore, according to Moshavi, is not an interrogative itself (Moshavi 2011: 97–99; compare Sivan and Schniedewind 1993: 218–19). Hypothetically, rhetorical questions can be answers to genuine questions but usually only with a nuance of irony and annoyance, which is not the case in Judg 5:30, where an inferior is speaking to a superior. Both these 8.  BH noninterrogative ‫הֲלֹא‬/‫ הֲלֹוא‬is cognate with Ugaritic hl, El-Amarna allū, Old Aramaic ‫הלו‬, Biblical Aramaic ‫אֲלּו‬, and Samaritan Hebrew ălū (see Brown 1987: 218–19; Sivan and Schniedewind 1993: 210–11; and Moshavi 2011: 91 n. 2 for a fuller list of references). 9. See Moshavi 2007; 2011; I thank Adina Moshavi for making available to me her previously unpublished article (now Moshavi 2011) and for discussing this case with me. 10.  Compare with the interpretation in Brongers 1981: 183; here, Brongers renders ‫הֲלֹא‬ ‘surely’.

Verbal Tenses in “Archaic” Biblical Poetry

197

pragmatic factors suggest individually and in combination that the ‫ הֲלֹא‬sentence in Judg 5:30 is not a negative rhetorical but a noninterrogative sentence; therefore, one should not expect the ‫ יקטל‬forms to have a present progressive meaning in this case. 2.2. The Participle and ‫ יקטל‬Verb in Noninterrogative ‫הֲלֹא‬ Sentences in CBH and LBH Prose and Poetry Further examination of verbal tenses in ‫ הֲלֹא‬sentences reinforces the conclusion of the preceding section. In ‫ הֲלֹא‬sentences in both CBH and LBH prose, the predicative participle is consistently used as present progressive, sometimes even as present habitual, while ‫ יקטל‬is consistently used to indicate future tense, volitive modality, iterative aspect, and occasionally some epistemic modal uses. 11 The same picture is generally maintained in most parts of the poetic corpus, except the book of Job (for example, Job 8:10), which is therefore not included in the examples below. 12 I now illustrate each of these usages. 2.2.1 The Participle in ‫ הֲלֹא‬Sentences The predicative participle is widely attested as a present progressive in ‫הֲלֹא‬ sentences: ָ‫ֶׁש ָלחֲך‬ ְ ‫ִׂש ָראֵל אֶל יֹוסֵף הֲלֹוא ַאחֶיךָ ר ִֹעים ִּב‬ ְ ‫וַּיֹאמֶר י‬ ְ ‫ׁשכֶם ְלכָה ְוא‬ (5)  ​ ‫אלֵיהֶם וַּיֹאמֶר לֹו ִהּנֵנִי‬ ֲ And Israel said to Joseph, “Look, your brothers are pasturing the flock at Shechem. Come, I will send you to them.” He answered, “Here I am.” (Gen 37:13) 13

It is also used as present habitual:

11.  Not all the ‫ הֲלֹא‬sentences collected below are necessarily noninterrogatives. The data reveal no difference in ‫ יקטל‬distribution in different ‫ הֲלֹא‬sentences, whether they are potentially negative rhetorical or noninterrogative. 12.  As demonstrated by Mats Eskhult in a lecture to the Department of Hebrew Language at the Hebrew University in 2009, the predicative participle is not part of the verbal system in the book of Job. Except for Job, ‫ יקטל‬as present progressive in ‫ הֲלֹא‬sentences is sporadically attested in other poetic compositions, which might be influenced by archaic features, for example: ‫ּותבּונָה ִּתּתֵ ן קֹולָה‬ ְ ‫ָכמָה ִת ְקרָא‬ ְ ‫( הֲלֹא ח‬Prov 8:1); perhaps also Ps 139:21. Ruth 3:1 provides an interesting case, since the form of ‫א ַבּקֵׁש( יקטל‬ ֲ ) may be interpreted as present progressive; however, the meaning of the form ‫א ַבּקֵש‬ ֲ is different from the parallel ‫ ְּת ַבּקֵׁש‬in Gen 37:15 (see example [3] above). Both lexically and grammatically, the verb in Ruth 3:1 expresses an emotional state of desire, rather than the physical effort of seeking, as in Gen 37:15. This fact suggests a habitual interpretation. Perhaps, both the negative rhetorical and noninterrogative uses merged to a certain extent (Sivan and Schniedewind 1993: 226). 13.  See also 1 Sam 23:19; 26:1; 2 Sam 11:10; 2 Kgs 11:41 (and dozens of other ‫הלא הם‬ ‫ כתובים‬cases); Jer 23:24 (a verbal adjective of the static verb used as a predicate); Ezek 21:5; Ps 54:2; Prov. 26:19; 2 Chr 32:11.

198

Tania Notarius

‫ַּטאת רֹבֵץ‬ ָ ‫יטיב ַלּפֶתַ ח ח‬ ִ ֵ‫ ִאם לֹא ת‬. . . ‫הֲלֹוא‬ (6)  ​ Behold . . . if you do not do well, sin is lurking at the door. (Gen 4:7) 14 2.2.2.  A Range of Meanings for ‫ יקטל‬in ‫ הֲלֹא‬Sentences in CBH and LBH The most commonly attested meaning for ‫ יקטל‬in ‫ הֲלֹא‬sentences is future, as in Num 23:12: ‫ֶׁשמֹר ְלדַ ּבֵר‬ ְ ‫ ְּב ִפי אֹתֹו א‬′‫ָׂשים ה‬ ִ ‫ֲׁשר י‬ ֶ ‫ַוּיַעַן וַּיֹאמַר הֲלֹא אֵת א‬ (7)  ​ He answered, “Indeed, I will take care to say what the Lord puts into my mouth.” (Num 23:12) 15 There are also several examples in which ‫ יקטל‬in ‫ הֲלֹא‬sentences is used as a volitive modal, that is, it refers to a wish or desire to do something:  16 ‫ִקרָא ָדִוד אֶל ָהעָם ְואֶל א ְַבנֵר ּבֶן נֵר לֵאמֹר הֲלֹוא תַ עֲנֶה א ְַבנֵר‬ ְ ‫ַוּי‬ (8)  ​ David called to the army and to Abner son of Ner, saying, “Abner! Would you answer!?” (1 Sam 26:14a) 17 The word ‫ יקטל‬is also attested in iterative and present habitual sayings, partly overlapping with participle in this sense: ‫א ָלפָיו ְו ָדִוד‬ ֲ ‫ ַּב‬ ⟩‫ׁשאּול ⟨באלפו‬ ָ ‫הֲלֹוא ָלזֶה יַעֲנּו ב ְַּמחֹלֹות לֵאמֹר ִהּכָה‬ ‫⟨ברבבתו⟩ ְּבִר ְבב ָֹתיו‬ It is of him they sing to one another in dances, “Saul has killed his thousands, and David his ten thousands.” (1 Sam 21:12) 18

(9)  ​

This semantic distribution between the predicative participle and ‫יקטל‬ forms in the ‫ הֲלֹא‬sentences indicates unambiguously that the present progressive use of imperfective ‫ יקטל‬in Judg 5:30 is a rare and typologically archaic phenomenon, reflecting the use of ‫ hybrid construction > one-unit pronoun

Noted that, in principle, it is also possible to have a direct shift from two-unit pronouns to a one-unit pronoun without the intermediate stage. 3.3.  The Diachronic Development of the Pronominal Constructions As argued above, from a typological point of view, there is good reason to take the two-unit construction, similar to the example found in Akkadian, as the original stage. Thus, I propose the following development: 22.  In some languages only the second element declines for case, as, for example, in Russian. 23.  In the context of Biblical Hebrew, Jay (2009) proposes that ‫ איש‬functions as a quantifier. This analysis relies on the hypothesis regarding English that a sentence such as “The cats tickle each other” derives from “The cats each tickle the other(s)” (this observation has a long history in the literature; see Dougherty 1970; Belletti 1982; Heim et al. 1991). This theory encounters various problems, however, even in English. For example, the truth conditions are not the same in the two sentences: if there are many “cats,” the first sentence is true even if not all cats in fact were tickled, but the second sentence is not. More importantly, in many languages, including Biblical Hebrew, the pronominal elements, unlike in English, are not quantifiers. Although Jay (2009; based on Stein 2008) demonstrates that ‫ איש‬most often does not appear in its “lexical” meaning, he does not demonstrate that it is a quantifier anywhere in the biblical corpus. In fact, Stein has failed to notice the significant function of ‫ איש‬when it appears with its correlative ‫ ;רעהו‬see below, §4.4, on this issue.

220

Elitzur A. Bar-Asher Siegal

Stage I: {N(oun)P(hrase)1, NP2 ...NPn}.NOM

verb.sg reciprocal-pronoun1 NOM.SG

reciprocal-pronoun2 ACC.SG

Topic

Subject

Object

(6) Old Akkadian aḫ-um ana aḫ-im lā inappuš brother-NOM to brother-GEN NEG 3.SG.DUR.make.a.claim One will not make a claim against the other. (TCL 19 63:45) Stage II: {NP1, NP2 . . . NPn3}.NOM

Topic

VERB.PL

reciprocal-pronoun1 NOM.SG

Subject

reciprocal-pronoun2 ACC.SG

Object

The main change is that the verb is plural in form. While one might initially assume that the plural agreement reflects a reanalysis of the nominative topic as the subject, it seems that the plural appeared for another reason. Plural agreement can also occur when there is no explicit topic, and the reciprocal pronouns are anaphoric expressions referring to an earlier sentence in the context. Thus, this seems to be a good example of what is known in the literature (Corbett 2006: 155–60) as semantic agreement instead of syntactic agreement (“the committee have met” versus “the committee has met”) due to the fact that part of the meaning of a reciprocal relation is that more than one member occupies the subject position. Occasionally, this development appeared already in the Old Babylonian dialect of Akkadian: (11)  aḫ-um aḫ-am lā ibaqqarū brother-NOM brother-ACC NEG 3MPL.DUR.raise.a.claim None should raise claims against the other. (YOS 8 99:19f, OB) This is also what we encounter in Standard Arabic in sentences with twounit pronouns, as the following example demonstrates: baʿḍ-u-na (12)  ​a.  naḥnu nuwaddiʿu we 1.C.PL.IMP.say.farewell some-NOM.1.C.PL   baʿḍ-an  some-ACC.INDF We bade farewell to each other. (Cantarino 1975: 137) b.  linusāʿida baʿḍ-u-na baʿḍ-an 1.C.PL.SBJV.assist some-NOM-1.C.PL.POSS some-ACC.INDF Let us assist each other. (ar-ar.facebook.com/tohelpeachother)

Hebrew Pronominal Reciprocal Constructions

221

Compare this with examples from the Qurʾān during the classical period, where the agreement is with the pronoun: stamtaʿa baʿḍ-u-na (13)  a.  ​ rabb-ana lord-POSS.1.PL 3.M.SG.PST.make.profit some-NOM-1.PL.POSS   bi-baʿḍ-in   in-some-GEN.INDF Our Lord! We made profit from each other. (6:128)

b.  yakfuru baʿḍ-u-kum bi-baʿḍ-in 3.M.SG.IMP.deny some-NOM-2.M.PL.POSS in-some-GEN.INDF You will disown each other. (29:25) Similarly, in Biblical Hebrew, another strategy for expressing reciprocity appears: reciprocity is expressed, not pronominally (see above [2c]), but by repeating the same noun phrase in two positions in the sentence. For the purpose of this discussion, it is important to note that, even in sort of case, where there is clearly no other antecedent, the verb is plural: ‫ָׁשלּו‬ ָ ‫ִּכי ִגּבֹור ְּב ִגּבֹור ּכ‬ (14)  ​ One warrior will stumble over another. (Jer 46:12)  24 Stage III: {NP1, NP2 . . . NPn}.NOM

Subject

VERB[PL]

indefinite-pronoun1-indefinite-pronoun2 ACC.SG 24

Object

The last stage is a typical example of the topic reanalyzed as the subject and, consequently, the two separate pronouns are conceived as one unit—namely, the one-unit reciprocal pronoun. Before we turn to the constructions found in the history of Hebrew, it is important to note the following: 1.  This process may explain phonological fusion, in which two-unit pronouns become a one-unit pronoun, such as, for example, in German einander or in the Syriac merger of ḥad ḥad into ḥdādē. Similarly, it may account for the development in Standard Arabic (example [7b]) in which only one element of the pronoun remains (both cases can be considered tokens of haplology, with differing degrees of phonological deletion: ḥadḥad > ḥdādē; baʿḍuna baʿḍan > baʿḍuna). 2.  These stages reflect different syntactic relations. Thus, one language may have both types of construction simultaneously. 24.  In languages in which the verb is always in final position (as is the case in Akkadian), there is no relocation of the verb.

222

Elitzur A. Bar-Asher Siegal

Returning to the biblical formulas—as I have noted (see [2e] above and [15] below), there are a few examples of stage I, such as the following: ‫ְו ִדּבֶר חַד אֶת ַאחַד‬

(15)  ​ One said to the other. (Ezek 33:30)

But in almost all other instances with pronominal constructions, the verb is plural. Thus, returning to the question at the beginning of this essay: the difference between the constructions in terms of verbal agreement reflects (1) two different grammatical relations in reciprocal constructions, and (2) two historical stages. This suggests that the sentences in Biblical Hebrew with a singular form of the verb are relics of an older formula. After following languages where it is possible to track such a development, I propose a similar development in the history of Hebrew, of which only a few vestiges remain. It is now possible to turn to the question of whether the standard biblical construction, which consists of ‫ איש‬and ‫רעהו‬/‫אחיו‬, represents stage II or stage III. Clearly, from the formal perspective, this is not completely stage III, because it is possible to insert elements between the components of the formula, as in (16): (16)  ​

‫ִׁש ְמעּו ִאיׁש ְׂשפַת רֵ עֵהּו‬ ְ ‫לֹא י‬ They will not understand one another’s speech. (Gen 11:7)

However, it is possible that, semantically, Biblical Hebrew was already at stage III because there is a known phenomenon in the process of grammaticalization in which the functional process precedes its formal representations. 25 The support for this sort of analysis would be the few examples in which an explicit subject appears in a verb-initial sentence, and the NP cannot be analyzed as a topic, such as, for example, in (17): (17)  ​

‫ׂשרֵי ִג ְלעָד ִאיׁש אֶל רֵ עֵהּו‬ ָ ‫אמרּו ָהעָם‬ ְ ֹ ‫וַּי‬ The people, the leaders of Gilead, said to one another. . . . (Judg 10:18)

However, there are very few examples of this sort and they are all, to some extent, indecisive. 26 25.  This is the assumption of most studies on grammaticalization: functional processes motivate formal developments. In other words, once the functional process occurs the formal development can follow; hence, the assumption that there is also an order in time. See Bybee et al. 1994: 9–12, 19–21, among others. See also Hopper and Traugott (2003: 53–55) regarding the development of the French future. 26.  In example (17), the verb is in initial position because it is a waw-consecutive construction and, according to various analyses, the verb moves to this position only late in the derivation. Another example is the following: ‫ָחיו‬ ִ ‫חרֵי א‬ ֲ ‫( ִּכי אָז ֵמהַּב ֶֹקר נַעֲלָה ָהעָם ִאיׁש ֵמ ַא‬2 Sam 2:27). However, it is very possible from the context (see v. 26) that the reciprocal expression

Hebrew Pronominal Reciprocal Constructions

223

Returning to the question of the variety of reciprocal constructions in Biblical Hebrew, introduced at the beginning of this paper—so far, I have briefly discussed the typological distinction between verbal and pronominal constructions (2a–b vs. 2d–g) and provided a diachronic explanation for the difference between (2d) and (2f–g) in terms of verbal agreement. Now I turn to the various options for the components of pronominal constructions in Biblical Hebrew. 3.4.  Variations of Pronominal Reciprocal Constructions in Biblical Hebrew 3.4.1.  The Standard Constructions The sentences in (2f–g) and (4a–b) above demonstrate the two regular formulas of two-unit constructions to express reciprocity pronominally in Biblical Hebrew. In both, the first element is ‫‘ איש‬man’, and the second is either ‫אחיו‬ ‘his brother’ or ‫‘ רעהו‬his friend’. This presents us with an interesting situation in that, in one language, there are simultaneously two alternative constructions. Other languages also have more than one set, but often they are either typologically different (as in English “one . . . another” versus “each other”) or belong to different registers (as is the case in Modern Hebrew, ‫אחד את השני‬ [standard], ‫[ זה את זה‬higher level]). 27 Surveying the biblical data, we have a belongs to another sentence, in which case there is an ellipsis and one should add ‫ שבו‬to translate: ‘Surely then the people would have turned away from following each other in the morning’. Another sentence to consider is: ‫ִד ְּברּו י ְִראֵי יְהוָה ִאיׁש אֶל ֵרעֵהּו‬ ְ ‫‘ אָז נ‬Then those who feared God spoke to one another’ (Mal 3:16). The complement of this verb with the preposition ‫ אֶל‬is unexpected. In fact, it seems to be an example of a more general phenomenon in which the expression ‫אחיו‬/‫ איש אל רעהו‬appears with verbs that do not usually take the preposition ‫אֶל‬, such as in Gen 43:33; Jer 36:16; Ezek 24:23; Isa 13:8. One may consider that in these contexts this expression has an associative meaning. This function is expressed in various way cross-linguistically, most commonly with adverbs such as “together.” In other languages, it is expressed with a prefix, such as the Latin prefix com-. Interestingly, very often this is the same expression used in these languages for a reciprocal action (see Zaliznjak and Shmelev 2007; see also Evans et al. [2007], who argue that ‘act jointly’ is part of the meaning of the “prototypical reciprocal clauses”). This conceptual connection is also realized in the discontinuous construction (discussed above in n. 13), where the relationships between the participants of the reciprocal relations are expressed by the associative meaning. In light of this, the following line in Jeremiah should be reconsidered: ‫ָחיו‬ ִ ‫ְו ִנּפ ְַצ ִּתים ִאיׁש אֶל א‬ ‫ו ָהאָבֹות ְו ַה ָּבנִים י ְַחָּדו‬.ְ Instead of translating the verse ‘I will smash them one against the other, parents and children together’ (Jer 13:14), one could tie together the final two expressions with the meaning ‘together.’ In this context, note that the adverb ‫‘ יחדו‬together’ also appears several times in reciprocal contexts: Deut 25:11; 2 Sam 2:16; Isa 45:21; Amos 3:3; Ps 71:10, and others. In the future, I will consider these various phenomena further. See also Halevy 2011b: 406. 27.  Nedjalkov (2007b: 157) mentions a few other languages with more than one set of pronominal expressions. He also comments that this phenomenon is often the result of borrowing from other languages. See also Haas (2010: especially chap. 5) regarding competing strategies in English. In our context, it is interesting to note that, on the one hand, ‫איש־רעהו‬

224

Elitzur A. Bar-Asher Siegal

hard time finding a clear distribution for the two constructions either textually or historically. Books such as Judges, which uses ‫רעהו‬/‫ איש‬exclusively, or such as Ezekiel, which uses ‫אישה אחותה‬/‫ איש אחיו‬exclusively, are rare. 28 Isaiah and Jeremiah seem to be aware of this alternation and combine both expressions in their poetry to create parallel expressions: (18)  ​

‫חזָק‬ ֲ ‫ָחיו יֹאמַר‬ ִ ‫ּולא‬ ְ ‫  ִאיׁש אֶת רֵ עֵהּו י ְַעזֹרּו‬29 They help each other and say one to another, “Be strong!” (Isa 41:6)

(19)  ​

‫ָחיו‬ ִ ‫אמרּו ִאיׁש עַל רֵ עֵהּו ְו ִאיׁש אֶל א‬ ְ ֹ ‫ּכֹה ת‬ Thus will each of you say to his neighbor and to his brother. . . . (Jer 23:35)

Before surveying the less common constructions in Biblical Hebrew mentioned in the introduction (2c–e), I would like to add another comment regarding possible signs of internal grammaticalization within Hebrew of the components of these constructions. As noted, sentences (2c) and (14) demonstrate that in Biblical Hebrew the repetition of any noun could express reciprocity.  30 Thus, one could speculate the following stages in the grammaticalization of the pronouns: 1.  Originally, there was no single word dedicated to expressing reciprocity. At this stage, it is reasonable to assume that the words ‫ איש‬and ‫רעהו‬/‫ אחיו‬were used only in contexts where the meaning of ‘brother/friend’ was relevant. 2.  Later, through a process of grammaticalization, the semantics of the words ‫איש‬ and ‫רעהו‬/‫ אחיו‬went through a process described in the literature of grammaticalization as semantic bleaching—that is, the loss of lexical content and the retention (or acquisition) of grammatical content. Thus in our contexts, these lexemes became pronouns in the context of reciprocal constructions.

is a calque of an Aramaic formula and, on the other hand, the ‫ אח‬component in the alternative formula is a cognate of the same lexeme used for similar functions in Akkadian. 28.  There are some short books, such as Joel, with only one occurrence of this sort of construction; they are irrelevant to our discussion. 29.  Note the difference in agreement. While the first is plural, as is standard in Biblical Hebrew, the second verb is singular. 30.  Regarding the cross-linguistic phenomenon of repeating a noun phrase to express reciprocity, see Nedjalkov 2007: 154; Bar-Asher Siegal 2011; for evidence of this phenomenon in Biblical Hebrew, see Jay 2009: 7 n.14. It is impossible to characterize the text in which such a formula occurs in diachronic or stylistic terms. It is also hard to determine whether the occurrences of this sort of formula are merely relics of an older stage or are evidence that at some level the repetition was still a productive means of expressing reciprocity. See Bar-Asher Siegal 2011 for the Akkadian evidence with regard to this question.

Hebrew Pronominal Reciprocal Constructions

225

In fact, however, it seems to be more accurate to say that at first only the first component, ‫איש‬, was the regular indefinite pronoun. 31 Thus, the extension was the addition of a second correlative component: either ‫ אחיו‬or ‫רעהו‬. Both are nouns with a genitive pronominal suffix that refers to the first component, ‫איש‬. One can even imagine that this process of semantic bleaching began when ‫רעהו‬/‫ אחיו‬were being used in contexts similar to the following: (20)  ​

‫ָחיו ְו ִאיׁש אֶת רֵ עֵהּו‬ ִ ‫חנֶה ְו ִה ְרגּו ִאיׁש אֶת א‬ ֲ ‫ּשׁעַר לָׁשַ עַר ַּב ַּמ‬ ַ ‫ִע ְברּו וָׁשּובּו ִמ‬ ‫ְו ִאיׁש אֶת ְקרֹבֹו‬ Go back and forth through the camp from one end to the other, each killing his brother and friend and neighbor. (Exod 32:27)

This is a context in which the original meaning of ‘brother’ is probably still relevant in the listing, even though the list as a whole (‘brother’, ‘friend’, and ‘neighbor’) reflects ‘everyone’. In other contexts, one can imagine that all three of these words were perceived as a general expression for such a relation. 32 Thus, in the early stage, only ‫ איש‬was used as an indefinite pronoun with various other participants (‘brother’, ‘friend’, etc.), and then ‫ אחיו‬or ‫ רעהו‬became a pronoun as well. In an early stage, ‫ איש‬and ‫רעהו‬/‫ אחיו‬were probably used only for people, but once these words were grammaticalized, we find the expected development of their being used in the context of animals as well: 33 ֻ ‫ֶׁשת ְו ַאיִל ְמ‬ ֶ ‫ׁשּל‬ ֻ ‫ֶׁשת ְועֵז ְמ‬ ֶ ‫ׁשּל‬ ֻ ‫ֶגלָה ְמ‬ ְ ‫וַּיֹאמֶר ֵאלָיו ְקחָה ִלי ע‬ (21)  ​ .‫ׁשּלָׁש ְותֹר ְוגֹוזָל‬ ְ ‫ַּתו‬ ‫ֶך ַוּיִּתֵ ן ִאיׁש ִּב ְתרֹו ִל ְקרַ את רֵ עֵהּו‬ ָ ‫ַוּיִּקַ ח לֹו אֶת ּכָל ֵאּלֶה ַו ְיבַּתֵ ר א ָֹתם ּב‬ ‫ָתר‬ ָ ‫ְואֶת ה ִַּצּפֹר לֹא ב‬ So he said to him: “Bring me a heifer, a goat and a ram, each three years old, along with a dove and a young pigeon.” And he brought all these, cut them in the middle and arranged the halves opposite each other; the birds, however, he did not cut in half. (Gen 15:9–10) Later, the same terms were even used with inanimate objects: (22)  ​

‫חמֵׁש ַהיְִריעֹת ִּת ְהיֶיןָח ְֹברֹת ִאּשָׁה אֶל אֲח ָֹתּה‬ ֲ The five curtains should join each other (Exod 26:3)

31. See Stein 2008, among others. 32.  As we shall see below (§4.4), the use of these pronouns in Biblical Hebrew is more general, and thus we may assume that this development occurred independently of the reciprocal use of these pronouns. 33.  For a similar description, see Jay 2009: 7.

226

Elitzur A. Bar-Asher Siegal

Another development that might be expected would be lack of gender agreement. Biblical Hebrew does not exhibit lack of gender agreement in these constructions, but signs of this development are found in other languages. 3.4.2.  Other Constructions As noted in the introduction, there are also some other less-common pronominal constructions in Biblical Hebrew. The first is a repetition of the words ‫אחת‬/‫‘ אחד‬one’ (see above, [2e]): (23)  ​

‫ָסים‬ ִ ‫ַו ְי ַחּבֵר אֶת ַהיְִריעֹת ַאחַת אֶל ַאחַת ּב ְַּקר‬ He joined the curtains to one another with the clasps. (Exod 36:13)

It is interesting to note that the repetition of the lemma ‘one’ in the two-unit pronouns is common in all dialects of Late Aramaic. In fact one may consider that the following example in Ezekiel is the Aramaic expression (note the Aramaic form ‫ )חַד‬with its Hebrew gloss (‫ָחיו‬ ִ ‫)איׁש אֶת א‬: ִ  34 (24)  ​

‫ָחיו לֵאמֹר‬ ִ ‫ְו ִדּבֶר חַד אֶת ַאחַד ִאיׁש אֶת א‬ One said to the other, saying to each other. . . . (Ezek 33:30)

Note that the use of ‘one’ in this sort of construction is common crosslinguistically, and, while it is common in Late Aramaic, it is occasionally found in other Semitic languages as well. 35 However, since it is only occasionally found in Hebrew, it is reasonable to infer that this was a linguistic variation that was probably introduced due to influence from other languages. Even less common is the repetition of the demonstrative (2d), as examples (25)–(26) demonstrate, though this construction is very important to this historical discussion: (25)  ​

‫־זֶה ְו ָאמַר‬‎‫ו ָקרָא זֶה ֶאל‬ . . . ְ ‫ָפים ע ֹ ְמ ִדים ִמ ַּמעַל לֹו‬ ִ ‫ְׂשר‬ Seraphim stood above Him . . . and they called one to the other and said. . . . (Isa 6:2–3)

(26)  ​

‫ָמים‬ ִ ‫ׁש ְבעַת י‬ ִ ‫ַוּיַחֲנּו ֵאּלֶה נֹכַח ֵאּלֶה‬ For seven days they camped opposite each other. (1 Kgs 20:29)

Notice that, from a typological point of view, both (24) and (25) illustrate the first stage in the development of the pronominal constructions, because the 34.  Moshkovitz 1985: 264. Interestingly, the Septuagint translates only the Hebrew sentence. 35.  Nöldeke (2001: 354) also provides references for a similarly rare formula in Arabic. In Bar-Asher Siegal 2011, I mention one Akkadian example from the Neo-Assyrian period in which the two-unit pronouns consist of the repetition of ištēn ‘one’. I speculate that this occurred due to the influence of Aramaic.

Hebrew Pronominal Reciprocal Constructions

227

verb is still singular. 36 Interestingly, it appears with the nonstandard formulas (that is, constructions that do not consist of ‫ איש‬and ‫רעהו‬/‫)אחיו‬. Furthermore, unlike the other constructions, as the contrast between (25) and (26) illustrates, the pronouns agree in number with their antecedents (‫‘ זה‬this’ versus ‫אלה‬ ‘these’). I continue the discussion about number agreement in §3.5. From a typological perspective, (25) illustrates the older type, which is rare in the biblical corpus. However, from the perspective of the history of Hebrew, examples (25)–(26) provide the earliest attestations of the elements that became the regular construction in later Mishnaic Hebrew, when the most common construction was the repetition of demonstratives (Segal 1927: 208, §433). Thus, it is not unlikely that the few examples of this construction in Biblical Hebrew reflect early signs of this mishnaic construction back in the biblical period. Because the relationship between the biblical and mishnaic constructions is the topic of the next section of this essay, I turn now to introducing reciprocal pronouns in the rabbinic corpus. 3.5  Mishnaic Hebrew Pronominal Reciprocal Construction(s) The following examples demonstrate the above-mentioned standard construction of Mishnaic Hebrew: (27)  a.  ​

‫השוכר את האומנים והיטעו זה את זה ואין לו זה על זה אלא‬ ‫תרעומת‬ If one engages craftsmen and they deceive one another, they have only resentment against each other. (m. B. Meṣiʿa 6:1)

b. 

‫אין דנין לא זה את זה ולא זה עם זה ולא זה על זה ולא זה בפני‬ ‫ ואין מעידין לא זה את זה ולא זה עם זה ולא זה על זה ולא זה בפני‬.‫זה‬ ‫זה‬ They do not judge each other, with each other, concerning each other, or in the presence of each other; and they do not testify against each other, with each other, or in the presence of each other. (t. Sanh. 5.4)

c. 

‫ בזמן שמקצתן רואים‬.‫שתי חבורות שהיו אוכלות בבית אחד‬ ‫אלו את אלו הרי אלו מצטרפין לזימון‬ If two separate parties have dined in the same house, if some of each party are able to see some of the other company, they may join to say the grace of the meal together‎. (m. Ber. 7:5)

36.  See also Greenberg 1997: 686. It is interesting to compare the ancient translations of this sentence. While the Septuagint relies strictly on the Hebrew text and is singular, the Aramaic targum uses the plural for both verbs, which is also the standard in the rest of the Hebrew Bible.

228

Elitzur A. Bar-Asher Siegal

In (27c), there is agreement in number between the sets participating in the reciprocal relation. It should be noted that in both (27a) and (27c) the antecedent of the pronouns is plural (‘craftsmen’ and ‘two parties’); however, only sentence (27c) has the plural form of the demonstrative. The reason for this is that in (27a) the reciprocal relationship is between individuals (one craftsman deceives the other), and in (27c) the reciprocal relationship is between sets (one party sees the other). Thus, it is clear that this agreement is not syntactical but semantic. 37 However, in a few other places, we encounter the singular demonstrative ‫ זה‬where the plural form ‫ אלו‬is expected, indicating that the grammaticalization of the demonstratives as the components of the pronominal reciprocal construction affected the grammar: (28)  ​

‫הקמחין והסלתות מעלין זה את זה‬ The flours and fine meals may neutralize each other. (t. Ter. 6.6)

The shift from the biblical formula to the mishnaic is nicely illustrated in the following example, taken from a legal midrash that cites a reciprocal expression in the Bible but then uses the rabbinic formula in its own phrasing of the law: (29)  ​

‫” אלא איש בעונות‬.‫ “איש באחיו‬,]‫” אינו או[מר‬.‫”וכשלו איש באחיו‬ ‫ מלמ[ד[ שכל יש[ראל] ערבין זה לזה‬.‫אחיו‬ “They shall stumble over one another”—the verse does not say “over one another” [lit., a man over his brother] but “one over the sins of the other.” This teaches us that all the Israelites are a pledge for each other [lit., this to this]. (Sipra Beḥuqotay 7.5)

In a few instances, we do find the biblical formula in Mishnaic Hebrew. The following example appears in tractate ʾAbot, which is known for its archaic style and, thus, shows affinity with the language of the Bible:  38 (30)  ​

.‫ הווי מתפלל (ל)[ב]שלמה שלמלכות‬.]‫ חנניה סגן הכהנים או[מר‬′‫ר‬ .‫שאילולי מוראה איש את רעהו חיים בלענו‬ R. Ḥanina, the chief priest, said: “Pray for the welfare of the ruling power, since, but for the fear thereof, men would engulf each other alive.” (m. ʾAbot 3:2)

37.  Glinert (1989: 69) already noted this distribution in the context of Modern Hebrew, where this construction appears in situations with a higher register of language and preserves this syntactic-semantic distinction. Regarding the uniqueness of these agreement rules crosslinguistically, see Heine and Miyashita 2008: 169–70. 38.  For the relationship between the language of tractate ʾAbot and Biblical Hebrew, see Sharvit 2006: 32–59. For a discussion of the issue of reciprocal expressions, see specifically p. 48 there.

Hebrew Pronominal Reciprocal Constructions

229

In this same context, it is interesting to note that at Qumran one regularly finds the biblical components ‫רעהו‬/‫איש–אחיו‬, while the components ‫זה–זה‬, which are common in Mishnaic Hebrew, are used only occasionally (compare, for example, 1QS 2:20–21 with 1QS 5:23). 39 This is the expected distribution, when there is such a difference between Biblical and Mishnaic Hebrew. So far, the differences between the biblical and the mishnaic formulas are mostly formal. While in the former the construction consists of the pronominal expressions ‫רעהו‬/‫איש–אחיו‬, the latter consists of a repetition of the demonstrative ‫זה‬. Syntactically, both constructions seem to be similar with to the fact that, despite the single number of the components of the construction, the verb is plural. 4.  The Biblical and Mishnaic Pronominal Constructions and Their Diachronic Relation 4.1.  Biblical and Mishnaic Hebrew Before proceeding to the specific issue of the relationship between the reciprocal constructions in each period, we should note one fact with regard to the relationship between the languages of the two corpora. While it is clear that the texts of the biblical era and the mishnaic era are from different periods, it is not always accurate to speak of them as successive linguistic periods. In other words, encountering a certain grammatical function expressed by X in Biblical Hebrew and by Y in Mishnaic Hebrew, we are not always correct methodologically in inferring that X changed/developed into Y and then positing an explanation for the change. There are strong arguments in support of the notion that Mishnaic Hebrew evolved from an early dialect that probably coexisted with the biblical dialect. It is reasonable to suppose that the appearances of the pronominal reciprocal construction that consisted of the repetition of the demonstratives reflected this situation. 40 For the purposes of the functional comparison to be discussed below (§§4.3–4), it is sufficient to note that, for some speakers, this sort of development occurred. Thus, the goal of this part of the paper is to describe the nature of this shift. 4.2.  The Origin of the Mishnaic Construction From a typological perspective, the mishnaic and the biblical formulas are the same, because both have two-unit pronouns, and the verbs are plural. However, it is not a simple matter to determine how to characterize the change between the periods. Presumably, since pronominal reciprocity is a syntactical phenomenon, this is a syntactical change. But it is difficult to claim this definitively because, from a typological perspective, nothing changed on the 39.  See also Mor 2009: 238–39. 40.  See, inter alia, Bar-Asher 2006: 573; and Rendsburg’s article in this volume.

230

Elitzur A. Bar-Asher Siegal

syntactic level. It was not a morphological shift either, because no morphological changes occurred. Thus, one could potentially argue that, although this is a grammatical issue, it involved nothing more than a lexical change from ‫רעהו‬/‫ איש–אחיו‬to ‫זה‬-‫זה‬. As we saw in §3.4.2, the mishnaic formula had already appeared in a few examples in the biblical corpus; thus it is possible that this is another example of a dialectal variation that became the standard in the later period (§4.1). One could also propose that Aramaic contributed to this shift and consider it to be some sort of a calque of an Aramaic formula, because similar vocabulary is found in the equivalent Biblical Aramaic construction, where pronominal reciprocity is also expressed by repeating the demonstrative pronoun: ‫ׁשן‬ ָ ‫ָק‬ ְ ‫ְוא ְַרכֻּבָתֵ ּה ָּדא ְל ָדא נ‬ (31) a.  ​ and his knees were striking one another. (Dan 5:6) b.  ‫ְולָא ֶלהֱו‍ֹן ָּד ְב ִקין ְּדנָה ִעם ְּדנָה‬ and they will not adhere to one another. (Dan 2:43) If there was Aramaic influence on Mishnaic Hebrew in this matter, we should ask: where and when did it happen? The examples in (31) are from Biblical Aramaic, specifically from the book of Daniel, the provenance of which is debatable. Interestingly, repeating the demonstratives is not the standard way to express reciprocity pronominally in Galilean Aramaic, where it is customarily expressed (as in the late eastern dialects) by repeating the word ‫‘ חד‬one’:  41 (32)  ​

‫ואינון פליגין חדא על חדא‬ They are at variance with each other. (y. Ḥal. 3.2)

The construction with a repetition of the demonstratives also appears at Qumran (Muraoka 2011: 51) and in the Samaritan dialect (Stadel 2011: 39–40). However, it is still unclear whether these are only attestations of a higher level of language, imitating the biblical style, or whether the formulation was a genuine phenomenon in these dialects. Similarly, in the late Palestinian translation, Tg. Yerushalmi (Pseudo-Jonathan), occasional sentences that are not a translation of a biblical verse also use the following formula: 41.  There are rare examples of demonstratives used for the reciprocal function: ‫גבר אית‬ ‫‘ ליה בר וחורגא ואינתו אית לה בר וחורגה ואסבון דין לדין‬a husband has a son and a stepdaughter and the wife has a son and a stepdaughter and they married each other’ (y. Yebam. 10.6; and in a similar context in y. B. Meṣiʿa 2.5); ‫ והוו מגפפין דין לדין ומנשקין‬.‫גלי גרמיה וחכים דין לדין‬ ‫‘ דין לדין‬he uncovered himself, and they recognized each other, and they embraced each other and kissed each other’ (Lam. Rab. 1:46). Since such sentences are rare, they may be examples of the use of an archaic formula, or they may reflect a variation that was retained in certain dialects.

Hebrew Pronominal Reciprocal Constructions (33)  ​

231

‫וקריבו דין לדין‬ And they approached each other. (Tg. Ps.-J. on Num 21:14)

Although it is possible that this is a reflection of another variation in Palestinian Aramaic, in this case it might instead be an archaism. As long as we do not have a full survey of the pronominal reciprocal constructions in the history of Aramaic, it is difficult to determine where and whether the mishnaic form is exhibiting a borrowing from Aramaic. It could have occurred in the older period of Official or Middle Aramaic, or it could be reflecting a local Palestinian dialect that appears only in the late translations. With regard to the question introduced at the beginning of this essay, the various components of the pronominal constructions may be explained as the result of a calque that literally translated the components of the pronominal construction of another language. Although structures similar to (2d) and (2e) appear in Aramaic, it is significant that they are not found in the same Aramaic dialects, suggesting that there may have been various times or locations that exercised influence. 4.3.  The Pronouns in Pronominal Reciprocal Constructions The beginning point for the current discussion was a common assumption in the typological literature that a reciprocal construction is a defined phenomenon that can and should be treated independently—in other words, that there are forms, the function of which is to represent reciprocal situations. As such, these forms can be compared cross-linguistically, and they can also be subjected to historical analysis. I begin with a fact observed in the literature on reciprocal constructions, which is that the so-called reciprocal pronouns are also used in asymmetric relations in which reciprocity is impossible: (34)  ​

‫ָחיו‬ ִ ‫ִּכי אָז ֵמהַּבֹקֶר נַעֲלָה ָהעָם ִאיׁש ֵמ ַאחֲרֵי א‬ then the people would have gone away in the morning, each from following the other. (2 Sam 2:27) 42

If X follows Y, Y cannot follow X. The pronouns in these asymmetrical contexts convey the fact that X follows Y, Y follows Z, and so on. This is definitely not a reciprocal relation; and clearly, while the objects in the middle fulfill the role of both leaders and followers, this is not true of the first and the last. In light of this, it is better to describe the function of these pronouns in the following way: These are pronominal expressions used in relationships between two sets or more without specifying which of the sets occupies which position. 42.  Concerning the interpretation of this verse, see above n. 28.

232

Elitzur A. Bar-Asher Siegal

Thus, these pronouns may serve to mark both reciprocal relationships and asymmetrical relationships. 43 Furthermore, in Biblical Hebrew, the use of the pairs ‫ איש‬and ‫רעהו‬/‫אחיו‬ should be considered more broadly. I submit that this usage is actually what stands behind the situations in the legal corpora that list casuistic laws:  44 ‫ָרמָה‬ ְ ‫ָרגֹו ְבע‬ ְ ‫ְו ִכי יָזִד ִאיׁש עַל רֵ עֵהּו ְלה‬ (35) a.  ​ If a man acts presumptuously toward his neighbor to kill him treacherously. . . . (Exod 21:14) b.  ‫ׁשמֹר‬ ְ ‫ׂשה ְוכָל ְּב ֵהמָה ִל‬ ֶ ‫ִּכי יִּתֵ ן ִאיׁש אֶל רֵ עֵהּו חֲמֹור אֹו ׁשֹור אֹו‬ If a man gives his fellow a donkey, an ox, a sheep, or any other animal to keep for him. . . . (Exod 22:9) c.  ‫ּורצָחֹו נֶפֶׁש ּכֵן הַָּדבָר ַהּזֶה‬ ְ ‫ֲׁשר יָקּום ִאיׁש עַל רֵ עֵהּו‬ ֶ ‫ִּכי ַּכא‬ This case is like that of someone who rises against his fellow and murders him. (Deut 22:26) These examples are not reciprocal; even if the verb allows for a reciprocal reading elsewhere, this is not the case in these legal contexts (35a–c). These examples fit the proposed definition above, though the “sets” usually contain only one member each. In fact, the sentences in (35) are regular examples of lexicalized nonspecific indefinite pronouns. Thus, although above, the semantic bleaching of relevant nouns was thought to occur only in the context of reciprocal constructions, it now becomes clear that this “bleaching” is relevant to more situations because these nouns appear as “pronouns” in more contexts. There are, however, some differences between these examples (35) and the examples above (16–22) with the reciprocal use of these expressions (and to some extent this is also true in the case of the asymmetric examples). First, there is no specification in any of examples (35a–c) about which of the sets occupies which position; in reciprocity, however, the lack of determination with regard to the referent of each component of the pronouns (‫ איש‬and ‫רעהו‬/‫)אחיו‬ is only within a defined domain (the extension of the sets in the topic/subject 43.  Dalrymple et al. 1998 survey the various logical relations that can be expressed by the so-called reciprocal pronouns in English. They examine the various semantic components that determine the specific logical relations between the various sets that hold the relation expressed by the predicate. 44.  The following sentence from outside of the legal corpus perhaps is also relevant for our analysis: ‫ֲׁשר יְדַ ּבֵר ִאיׁש אֶל ֵרעֵהּו‬ ֶ ‫ׁשה ָּפנִים אֶל ָּפנִים ַּכא‬ ֶ ֹ ‫‘ ְו ִדּבֶר יְהוָה אֶל מ‬The LORD used to speak to Moses face to face, just as a man speaks to his friend’ (Exod 33:11). The fact that the verb is in singular suggest that only God is speaking, and what the second clause describes is only the fact that God and Moses are facing each other, as is the case in a regular conversation.

Hebrew Pronominal Reciprocal Constructions

233

of the sentence). Thus, in (35) the pronouns are not limited to a specific group, because they do not refer to any constituent in the specific context. Second, in reciprocity each set that is a member of the topic/subject set is supposed to participate more than once in the relation/action of the verb, each time as a different argument, whereas in nonreciprocal contexts each participant participates only once. As discussed above (§3.2), the reason for the plural form of the verb is either that the subject is always plural or that the plurality of the events results in semantic agreement; consequently, it is not surprising that in (35) the verb is singular. Let us compare this distribution with similar contexts in Mishnaic Hebrew and examine how all the functions expressed by the biblical pronouns ‫ איש‬and ‫רעהו‬/‫ אחיו‬are expressed in Mishnaic Hebrew. The repetition of demonstratives is the regular expression for relations between sets of two or more without specifying which of the sets occupies which position. Thus, this formula appears also in nonreciprocal/asymmetrical contexts, as example (36) illustrates: (36)  ​

‫ואפה בשלשה תנורין זה אחר זה‬. He baked in three ovens, one after the other. (t. Pesaḥ. 2.1)

However, these are not the pronouns used in casuistic laws. In Mishnaic Hebrew, instead of the pair ‫רעהו‬-‫איש‬, the regular expression in such contexts consists of the following pair of indefinite pronouns: ‫חבירו‬-‫אדם‬: ‫התקינו שיה(ו)א אדם שואל את שלום חבירו בשם‬ (37) a.  ​ It was directed that every man should greet his friend in the name of the Lord. (m. Ber. 9:5) b. 

‫ ס[י]מא‬.‫ בין ער בין ישן‬.‫ בין שוגג בין מזיד‬.‫אדם מועד לעולם‬ ‫ משלם נזק שלם‬.‫את עין חבירו ושיבר את הכלים‬ A man is always accounted as noxious, regardless of whether he causes damage intentionally or unintentionally, when awake or asleep. If one blinded the eyes of his fellow, or broke his vessels, he must pay full damage. (m. B. Qam. 2:6)

In fact, it is even more accurate to say that in (37b) the first element of the blinding relationship (‫ )סימא את עין חבירו‬is implied by the agent of the verb. Indeed, in Mishnaic Hebrew it is very common for the first element of these sentences to be implied by a participial form: (38) a.  ​

‫המסכך את גפנו על גבי תבואתו של חבירו הרי זה קידש וחייב‬ ‫באחריותו‬ If one allows his vine to overtop the grain crop of his fellow. . . . (m. Kil. 7:4)

234

Elitzur A. Bar-Asher Siegal b.  ‫ או שלווה המינו‬.‫הגוזל את חבירו‬ If one has robbed aught from his fellow or borrowed from him. . . . (m. B. Qam. 10:6)

Interestingly, unlike Biblical Hebrew, where the process of grammaticalization also allows for the application of these indefinite pronouns to inanimate objects (as demonstrated above in [22]), Mishnaic Hebrew usually does not, as the following example illustrates: (39)  ​

.‫מטבילים מגב לגב ומחבורה לחבורה ביום טוב‬ One may immerse from one purpose to another, and from one company to another. (m. Beṣah 2:3)

For animate expressions, these indefinite pronouns are used, but for inanimate expressions, a full repetition of the relevant noun phrase is required. 45 This contrast is evident in the following example, where the word ‫‘ חטאת‬a sinoffering’ is repeated, while in the case of ‫‘ תינוק‬a child’, the second element is the pronoun ‫חבירו‬: (40)  ​ .‫לא היו עושין (א)לא חטאת על גב חטאת [ו]לא תינוק על גב[י] חבירו‬ It was prohibited to prepare a sin-offering by virtue of [the purification made for] another sin-offering, or [to make use of] a child by virtue of [the purification made for] his fellow. (m. Parah 3:4) There seem to be occasional exceptions, however, in which such pronouns appear for inanimate objects as well, such as the following: (41)  ​

‫אין משחיזין את הסכין אבל משיאה על גבי חברתה‬ One may not whet a knife, but one may sharpen one against the other. (m. Beṣah 3:7)

Thus, if Mishnaic Hebrew had employed indefinite pronouns for the reciprocal constructions, the two-unit pronouns would have included the nouns ‫אדם‬ and ‫( חבירו‬Segal 1927: 208–9, §434). Instead, we have an example of a hybrid construction in the Mishnah. In example (42), the first element is the biblical ‫איש‬, and the second is the mishnaic ‫חבירו‬: (42)  ​

‫והחזנים מזרקים לפניהם והן מחטפים ומכים איש את חבירו‬ The superintendents threw them before them, and they snatched at them, and struck one another. (m. Sukkah 4:4)

45.  See n. 30 for a discussion on the use of repetition in similar contexts in Biblical Hebrew and cross-linguistically.

Hebrew Pronominal Reciprocal Constructions

235

With regard to the origin of this construction, note that, while ‫ אדם‬as a lexicalized indefinite pronoun was found already in Biblical Hebrew (see, for example, Lev 13:3 [Segal 1936: 64–65, §101]) and was very common in Second Temple literature such as Ben Sira, the use of ‫ חבר‬instead of the biblical lexeme ‫ רע‬should be regarded in light of what is found in Aramaic. There were signs of this construction already in Biblical Aramaic, and this is the lexeme that was used for this function in all Late Aramaic dialects.  46 4.4.  What Accounts for the Different Distributions of Pronouns in Biblical Hebrew and Mishnaic Hebrew? Surveying the data in Biblical and Mishnaic Hebrew introduces the paramount question at hand: What is the reason for this difference between Biblical Hebrew and Mishnaic Hebrew? This question can be asked from two directions: Why in the former are the same expressions used in both reciprocal contexts and nonreciprocal casuistic laws? Conversely, why in Mishnaic Hebrew is reciprocity expressed with demonstratives, while indefinite pronouns are used in nonreciprocal legal contexts? I propose that the answer lies in the fact that, while the biblical construction used nouns that bleached in certain contexts and were used as indefinite pronouns in certain contexts, the mishnaic construction contains demonstratives that are part of the grammar and are anaphoric expressions—that is, expressions that by their nature refer to other nominal expressions. 47 As anaphoric expressions, they are linked to an antecedent. Therefore, the context clarifies their reference. This requirement is met only when there are antecedents, which is the case in most reciprocal contexts. This requirement is not met in generic contexts such as casuistic laws. This is not the case with indefinite pronouns, which are not necessarily anaphoric, because they stand by themselves and do not refer to other nominal expressions in the context. Therefore, these indefinite pronouns can be used in contexts such as casuistic laws where such referents are not specified. Continuing in this line of thought, one may assume that, on the one hand, the fronting of the nominal expression in a reciprocal relationship (as, for example, in sentences like “to be patient with each other is a good thing”) could have been expressed easily only in Biblical Hebrew. Although I could not locate such an example in the biblical corpus, ironically, I found an example that uses the biblical construction in the mishnaic corpus: 46.  In Dan 7:20, it is used in the meaning of ‘other’. 47.  To avoid confusion, I do not use the term anaphora in the sense in which it was used in government and binding contexts. That framework distinguishes between anaphora and pronominals—the former must have strictly local antecedents (i.e., the anaphor and its antecedent are in the same sentence), while the antecedents of the latter need not be local. In this essay, as is common in typological literature, the term pronoun is used more broadly in that it does not need an antecedent, and the term anaphora is used in its more classic sense as a pronoun that has an antecedent, but its antecedent is not necessarily local.

236 (43)  ​

Elitzur A. Bar-Asher Siegal ‫קשה שנאת איש את רעהו לפני המקום‬ Mutual hatred is unfavorable in the eyes of God. (t. Menaḥ. 13.22).

On the other hand, it was probably impossible to express this sentence using the mishnaic formulation without an explicit topic, such as “for people to hate each other is unfavorable in the eyes of God.” Admittedly, it is hard to prove this, but the requirement of binding explains the absence of these pronouns in the mishnaic casuistic laws, making this conclusion reasonable. In sum, having given consideration to the pronominal reciprocal constructions from a typological point of view and in isolation from the other uses of these pronouns, we see that the relationship between the standard constructions in Biblical Hebrew and the constructions in Mishnaic Hebrew is identical, even though they have different lexical elements. Once these constructions are examined from a structural perspective, exploring the function of each element as it also appears in various other syntactical constructions, we can see that they are clearly different. The Biblical Hebrew construction employs indefinite pronouns, while the Mishnaic Hebrew construction is built by repeating the demonstrative pronouns. Returning to the methodological concerns considered at the beginning of the discussion regarding a functional comparison of two linguistic periods (§2), we see clearly that describing the phenomena of reciprocal constructions is not trivial. Assuming a diachronic relationship between Biblical Hebrew and Mishnaic Hebrew, one could posit that the Aramaic construction found in Daniel was reflected in the Hebrew construction. But, instead of describing the differences between Biblical Hebrew and Mishnaic Hebrew by focusing on the reciprocal construction, one will find it more productive to observe the pronouns used in each period. While Biblical Hebrew uses indefinite pronouns with and without explicit antecedents, Mishnaic Hebrew uses indefinite pronouns only nonreferentially—that is, without explicit antecedents. Thus, indefinite pronouns can appear in casuistic laws but not in reciprocal constructions. When the antecedent is a set and the verb describes a relation that connects them, then demonstratives appear in Mishnaic Hebrew, because they are anaphoric. As a result, pronominal reciprocal constructions have a different shape in each period. The older construction, which appears in Biblical Hebrew, has indefinite pronouns, and the later construction, which appears in Mishnaic Hebrew, has demonstrative pronouns. Accordingly, the similarity between the constructions of both periods (such as the plural form of the verbs and the way that the constructions provide reciprocal readings) is part of other, more-general principles that are used in order to produce the reciprocal meaning cross-linguistically. This observation might be related to another phenomenon. In Biblical Hebrew, ‫ איש‬can be specific and have a referent, as is in the following verse:

Hebrew Pronominal Reciprocal Constructions

237

ְ ‫ַוּיֵל‬ ‫ֶך ִאיׁש ִמּבֵית לִֵוי ַוּיִּקַ ח אֶת ּבַת לִֵוי‬

(44)  ​

A man from the house of Levi went and married a daughter of Levi. (Exod 2:1) In contrast, ‫ אדם‬in Mishnaic Hebrew is used only nonspecifically. When ‫איש‬ has a specific referent, it may also be used in Mishnaic Hebrew, and in the Mishnah it is almost always followed by ‫‘ פלוני‬so-and-so’: (45)  ​

‫ איש פלוני מגרש את‬.‫היה עובר בשוק ושמע קול הסופרים מקרים‬ ‫פלונית ממקום פלוני‬ If a person passing through a street hears the voice of public notaries [dictating to their clerks or pupils], saying, “So-and-so divorces so-and-so living at a certain place. . . .” (m. Giṭ. 3:1)  48 Table 1.  Comparison of the Distribution of Indefinite Pronouns in Biblical and Mishnaic Hebrew

Biblical Hebrew

‫רעהו‬/‫איש־אחיו‬    specific/non-specific (for inanimate as well) Mishnaic Hebrew ‫איׁש‬ specific (occasionally also nonspecific)

‫חבירו‬-‫אדם‬ ‫זה‬-‫זה‬ nonspecific (mostly for animate) anaphoric

Thus, borrowing from Aramaic changed not only the forms of the pronouns but also the function of each of the pronouns and the differences between them. Regarding reciprocal constructions: formally, the elements are different but, from a typological point of view, there is a great deal of similarity. 5. Conclusion In this essay, I deal with various aspects of pronominal reciprocal constructions in Biblical and Mishnaic Hebrew. I focus on the phenomena related to various aspects of historical syntax, and in §2 I briefly mention various methodological considerations. Throughout this paper, I illustrate different types of syntactic developments that occurred in the history of the pronominal reciprocal constructions in Hebrew, among them the following: 48.  The fact that indefinite pronouns are used referentially in some languages whereas in other languages they are not is well known in the typological literature (see Haspelmath 1997; among others).

238

Elitzur A. Bar-Asher Siegal

1.  Changes from one type of construction to another.  These changes are the result of reanalyses of the grammatical relations between the components of the constructions. Thus, for example, the topic of the sentence is reanalyzed as its subject, and this development reveals itself through verbal agreement, whether the verb agrees with the nouns participating in the reciprocal relationship or with their pronouns (§3.3). 2.  Changes in the range of application of a certain construction.  Thus, many of the pronominal constructions consist of nominal expressions that became grammaticalized pronouns used in various constructions. This process involved a semantic bleaching in which nouns lost their lexical meaning and only retained their grammatical content. As expected, this was a gradual process; thus, for example, there are sentences from Mishnaic Hebrew in which certain nouns are used only for animate expressions (§4.3), while similar expressions in Biblical Hebrew were (already) being used for a wider range, and were used for inanimate expressions as well (§3.4.1). 3.  Changes in the components of the various constructions.  In addition, it is possible that some changes in the components of the various constructions reflect the influence of Aramaic and may be considered examples of calques with regard to the function of the demonstratives (§§3.4.2, 4.2).

Returning to the methodological discussion about historical syntax in §2— the first two types of development are easier to follow, because there is similarity in form in the various periods. In contrast to this, in positing a third type of development, we find that the influence of one language on another relies on a functional similarity, which is more delicate. Section 4.4 was devoted to examining one case of such a development: the shift from the biblical components of the pronominal reciprocal construction to the mishnaic construction. Dealing with the question of how to characterize this shift reveals, I believe, the risk of considering syntactic issues in isolation. As became clear, what at first seemed to be a lexical shift from ‫רעהו‬/‫ איש–אחיו‬to ‫זה‬-‫זה‬, was in fact a rather significant change from a structural perspective. The terms /‫איש–אחיו‬ ‫ רעהו‬are indefinite pronouns and therefore can be used in legal texts without references. The component ‫ זה‬is a demonstrative and therefore needs to have a referent in its context. In light of these types of changes, it is now possible to conclude with the question that opened this paper about the coexistence of the various reciprocal constructions in Biblical Hebrew, and to summarize the explanations that were given throughout this essay regarding their origin and their relationship with similar constructions in Mishnaic Hebrew (see table 2).

Hebrew Pronominal Reciprocal Constructions

239

Table 2.  Summary of the Origin and Nature of Reciprocal Constructions in Biblical and Mishnaic Hebrew Example

Type

Comments

I.

‫אמ ְַציָהּו‬ ֲ ‫ִת ָראּו ָפנִים הּוא ַו‬ ְ ‫ ַוּי‬Verbal construction (2 Kgs 14:11; §3.1)

T-stem, common among the Semitic languages

II.

‫ וַּיֹאמֶר ִנָּועֵד אֶל ּבֵית ָהאֱל ִֹהים‬Verbal construction (Neh 6:10; §3.1) ‫ׁש ִּתים‬ ְ ‫ּופ ִל‬ ְ ‫ִׂש ָראֵל‬ ְ ‫ וַּתַ עֲר ְֹך י‬Nominal construc‫ֲרכָה ִל ְקרַ את ַמעֲָרכָה‬ ָ ‫ ַמע‬tion—repetition of the (1 Sam 17:11; §3.3, §4.3) nouns

N-stem found in other Semitic languages (such as Akkadian)

III.

IV.  a. ‫ְולֹא ָקרַ ב זֶה אֶל זֶה‬   (Exod 14:20)  b. ‫השוכר את האומנים והיטעו‬ ‫זה את זה‬ (m. B. Meṣiʿa 6:1; §3.3, §3.4.2–3.5, §4.2)

Pronominal construction Two-unit pronouns— singular agreement (in Biblical Hebrew [a]) reflects an older stage in the grammaticalization process of pronominal reciprocal constructions. Similar to the standard biblical construction, mishnaic construction (b) has plural verbal agreement

Assumed to be relics of a typologically older encoding of reciprocity The construction with demonstratives is standard in Mishnaic Hebrew. A few examples appear in Biblical Hebrew. Found also in Biblical Aramaic and in Tg. Pseudo-Jonathan Function: Pronominal expression used in relating two sets or more without specifying which set occupies which position. This construction contains demonstratives, which, as anaphoric expressions, require explicit antecedents (therefore not used in formulation of casuistic law, for example).

V.

‫ ְו ָקרַ ב א ָֹתם ֶאחָד אֶל ֶאחָד‬Pronominal (Ezek 37:17, §3.4.2) construction

The use of ‘one’ in a pronominal construction is found in the standard Late Aramaic formula and occasionally in other Semitic languages.

VI.

‫ִדחָקּו‬ ְ ‫ָחיו לֹא י‬ ִ ‫ ְו ִאיׁש א‬Pronominal (Joel 2:8, §3.3–3.4.1) Construction Two-Unit pronouns – the plural agreement reflects at least the second stage in the grammaticalization process of pronominal reciprocal constructions

One of the two standard constructions in Biblical Hebrew. Function: Pronominal expression used in relations between two sets or more without specifying which set occupies which position. Consisting of “bleached nouns”, these are indefinite pronouns and do not require explicit antecedents. Therefore, they are used in casuistic laws as well.

VII.

‫חזִקּו ִאיׁש ְּברֹאׁש ֵרעֵהּו‬ ֲ ַ‫ ַוּי‬Similar to VI (2 Sam 2:16, §3.3–3.4.1)

Similar to VI In Mishnaic Hebrew it appears as a marked construction, usually with some connection to the Biblical corpus.

240

Elitzur A. Bar-Asher Siegal

References Amberber, M. 2002 Verb Classes and Transitivity in Amharic. LINCOM Studies in Afroasiatic Linguistics. Munich: Lincom Europa. Arnold, M. A. 2005 Categorization of the hitpaʿel of Classical Hebrew. Ph.D. dissertation, Harvard University. Bar-Asher, E. A. 2009 A Theory of Argument Realization and Its Applications to Features of the Semitic Languages. Ph.D. dissertation, Harvard University. Bar-Asher, M. 2006 Mishnaic Hebrew: An Introductory Survey. Pp.  567–95 in The Literature of the Sages, Second Part, ed. M. E. Stone and S. Safrai. Amsterdam: Van Gorcum / Philadelphia: Fortress. Bar-Asher Siegal, E. A. 2011 Notes on Reciprocal Constructions in Akkadian in Light of Typological and Historical Considerations. Semitica et Classica 4: 23–42. Behrens, L. 2007 Backgrounding and Suppression of Reciprocal Participants: A CrossLinguistic Study. Studies in Language 31: 327–408. Belletti, F. A. 1982 On the Anaphoric Status of the Reciprocal Constructions in Italian. The Linguistic Review 2: 101–38. Bybee, J. L.; Perkins, R. D.; and Pagliuca, W. 1994 The Evolution of Grammar: Tense, Aspect and Modality in the Languages of the World. Chicago: University of Chicago Press. Cantarino, V. 1975 Syntax of Modern Arabic Prose, vol. 3. Bloomington: Indiana University Press. Comrie, B. 1988 Topics, Grammaticalized Topics, and Subjects. Berkeley Linguistics Society 14: 265–77. Corbett, G. G. 2006 Agreement. Cambridge: Cambridge University Press. Dalrymple, M.; Kanazawa, M.; Kim, Y.; Mchombo, S.; and Peter, S. 1998 Reciprocal Expressions and the Concept of Reciprocity. Linguistics and Philosophy 21: 159–220. Dougherty, R. C. 1970 A Grammar of Coordinate Conjoined Structures, Part I. Language 46: 850–98. Evans, N.; Gaby, A.; and Nordlinger, R. 2007 Valency Mismatches and the Coding of Reciprocity in Australian Languages. Linguistic Typology 11: 541–97.

Hebrew Pronominal Reciprocal Constructions

241

Everaert, M. 1999 Types of Anaphoric Expressions: Reflexives and Reciprocal. Pp.  63–83 in Reciprocals: Forms and Functions, ed. Z. Frajzyngier and T. S. Curl. Amsterdam: Benjamins. Faarlund, J. T. 1990 Syntactic Change: Toward a Theory of Historical Syntax. Berlin: de Gruyter. Frajzyngier, Z., and Curl, T. S., eds. 1999 Reciprocals: Forms and Functions. Amsterdam: Benjamins. Gelb, I. J. 1957 Notes on von Soden’s Grammar of Akkadian. BO 12: 93–111. Givón, T. 1984 Syntax: A Functional-Typological Introduction, vol. 1. Amsterdam: Ben­jamins. Glinert, L. 1989 The Grammar of Modern Hebrew. Cambridge: Cambridge University Press. Goldenberg, G. 1991 ‘Onself’, ‘one’s own’ and ‘one another’ in Amharic. Pp. 531–49 in Semitic Studies in Honor of Wolf Leslau, ed. A. S. Kaye. Wiesbaden: Harrassowitz. Greenberg, M. 1997 Ezekiel 21–37: A New Translation with Introduction and Commentary. AB 22a. Garden City, NY: Doubleday. Haas, F. 2010 Reciprocity in English: Historical Development and Synchronic Structure. Routledge Studies in Germanic Linguistics 15. New York: Routledge  / London: Taylor & Francis. Halevy, R. 2010 Reciprocal Constructions in Hebrew. Pp. 265–86 in Judaic Studies in Memory of Moriah Liebson, ed. A. Maman and R. Bliboim. Jerusalem: Hebrew University. [Hebrew] 2011a The Grammaticalization of Bipartite Reciprocal Markers in Hebrew. HS 52: 7–18. 2011b Reciprocal Constructions between Syntax and Lexicon. Leš 73: 401–22. Harris, A. C., and Campbell, L. 1995 Historical Syntax in Cross-Linguistic Perspective. Cambridge Studies in Linguistics 74. Cambridge: Cambridge University Press. Harris, M. 1978 The Evolution of French Syntax: A Comparative Approach. London: Longman. 1997 Indefinite Pronouns. Oxford: Clarendon. 2007 Further Remarks on Reciprocal Constructions. Pp. 2087–2115 in Reciprocal Constructions, ed. V. P. Nedjalkov. 5 vols. Amsterdam: Benjamins. Heim, I.; Lasnik, H.; and May, R. 1988 Reciprocity and Plurality. Linguistic Inquiry 22: 63–102. Heine, B. 1999 Polysemy Involving Reflexive and Reciprocal Markers in African Languages. Pp. 1–29 in Reciprocals: Forms and Functions, ed. Z. Frajzyngier and T. S. Curl. Amsterdam: Benjamins.

242

Elitzur A. Bar-Asher Siegal

Heine, B.; and Miyashita, H. 2008 The Intersection between Reflexives and Reciprocals: A Grammaticalization Perspective. Pp. 169–223 in Reciprocals and Reflexives: Theoretical and Typological Explorations, ed. E. König and V. Gast. Berlin: de Gruyter. Hopper, P. J., and Traugott, E. C. 2003 Grammaticalization. 2nd ed. Cambridge: Cambridge University Press. Jay, J. 2009 Reciprocal Constructions in Biblical Hebrew. GIALens: Electronic Notes Series 3/1. http://www.gial.edu/. Kemmer, S. 1993 The Middle Voice: A Typological and Diachronic Study. Amsterdam: Benjamins. Khan, G. 2008 The Neo-Aramaic Dialect of Barwar. 3 vols. Leiden: Brill. Kikusawa, R. 2003 The Development of Some Indonesian Pronominal Systems. Pp. 237–68 in Historical Linguistics 2001: Selected Papers from the 15th International Conference on Historical Linguistics, Melbourne, 13–17 August 2001, ed. B. J. Blake, K. Burridge, and J. Taylor. Amsterdam: Benjamins. König, E., and Gast, V., eds. 2008 Reciprocals and Reflexives: Theoretical and Typological Explorations. Berlin: de Gruyter. König, E., and Kokutani, S. 2006 Towards a Typology of Reciprocal Constructions: Focus on German and Japanese. Linguistics 44: 271–302. Kremers, J. 1997 How Arabs Speak to Each Other about Themselves: A Study of nafs and baʿḍ in Modern Standard Arabic. M.A. thesis, University of Nijmegen. Lipiński, E. 1997 Semitic Languages: Outline of a Comparative Grammar. OLA 80. Leuven: Peeters. Macuch, R. 1965 Handbook of Classical and Modern Mandaic. Berlin: de Gruyter. Maslova, E. 2007 Reciprocal and Polyadic: Remarkable Reciprocals in Bantu. Pp. 335- 52 in Reciprocal Constructions, ed. V. P. Nedjalkov. 5 vols. Amsterdam: Benjamins. 2008 Reflexive Encoding of Reciprocity: Cross-Linguistic and Language-Internal Variation. Pp. 225–57 in Reciprocals and Reflexives: Theoretical and Typological Explorations, ed. E. König and V. Gast. Berlin: de Gruyter. Mathesius, V. 1930 On Linguistic Characterology with Illustrations from Modern English. Pp.  56–63 in Actes du Premier Congrès International de Linguistes à La Haye, du 10-15 avril 1928. Leiden: Sijthoff.

Hebrew Pronominal Reciprocal Constructions

243

Mor, U. 2009 The Grammar of the Epigraphic Hebrew Documents from Judaea between the First and the Second Revolts. Ph.D. dissertation, Ben-Gurion University of the Negev. Moshkovitz, Y. T. 1985 Sēfer Yĕḥezkēl: Daʿat Mikrâ. Jerusalem: Rabbi Kook Institution. Muraoka, T. 2011. A Grammar of Qumran Aramaic. Ancient Near Eastern Studies Supplement 38. Leuven: Peeters. Nedjalkov, V. P. 2007a Overview of the Research: Definitions of Terms, Framework, and Related Issues. Pp. 3–114 in Reciprocal Constructions, ed. V. P. Nedjalkov. 5 vols. Amsterdam: Benjamins. 2007b Encoding of the Reciprocal Meaning. Pp. 147–208 in Reciprocal Constructions, ed. V. P. Nedjalkov. 5 vols. Amsterdam: Benjamins. Nedjalkov, V. P., ed. 2007 Reciprocal Constructions. 5 vols. Amsterdam: Benjamins. Nöldeke, T. 1875 Mandäische Grammatik. Halle: Waisenhaus. 2001 Compendious Syriac Grammar. Appendix trans. P. T. Daniels. Reprinted Winona Lake, IN: Eisenbrauns. Plank, F. 2008 Thoughts on the Origin, Progress, and Pronominal Status of Reciprocal Forms in Germanic, Occasioned by Those of Bavarian. Pp. 347–73 in Reciprocals and Reflexives: Theoretical and Typological Explorations, ed. E. König and V. Gast. Berlin: de Gruyter. Rendsburg, G. A. 1990 Diglossia in Ancient Hebrew. New Haven, CT: American Oriental Society. Saussure, F. de 2006 Writings in General Linguistics, trans. C. Sanders and M. Pires. London: Oxford University Press. Segal, M. H. 1927 Grammar of Mishnaic Hebrew. Oxford: Clarendon. 1936 Diqduq lešon ha-mišna. Tel-Aviv: Devir. Sharvit, S. 2006 Language and Style of Tractate Avoth through the Ages. Beer-Sheva: Ben Gurion University of the Negev Press. Siloni, T. 2001 Reciprocal Verbs. Proceedings of the Israel Association for Theoretical Linguistics 17. http://linguistics.huji.ac.il/IATL/17/TOC.html. 2008 The Syntax of Reciprocal Verbs: An Overview. Pp. 451–99 in Reciprocals and Reflexives: Theoretical and Typological Explorations, ed. E. König and V. Gast. Berlin: de Gruyter. Soden, W. von 1952 Grundriss der akkadischen Grammatik. Rome: Pontifical Biblical Institute.

244

Elitzur A. Bar-Asher Siegal

Sokoloff, M. 2002 A Dictionary of Jewish Babylonian Aramaic of the Talmudic and Geonic Periods. Ramat-Gan: Bar Ilan University Press / Baltimore: Johns Hopkins University Press. Stadel, C. 2011 The Morphosyntax of Samaritan Aramaic. Ph.D. dissertation, Hebrew University. [Hebrew] Stein, D. E. S. 2008 The Noun ‫( ִאיׁש‬ʾîš) in Biblical Hebrew: A Term of Affiliation. JHS 8/1. http:// www.jhsonline.org/. Traugott, E. C., and Heine, B. 1986 Introduction. Pp. 1–14 in Approaches to Grammaticalization, ed. E. C. Traugott and B. Heine. 2 vols. Amsterdam: Benjamins. Zaliznjak, A. A., and Shmelev, A. D. 2007 Sociativity, Conjoining, Reciprocity and the Latin Prefix com-. Pp. 209–29 in Reciprocal Constructions, ed. V. P. Nedjalkov. 5 vols. Amsterdam: Benjamins.

Syntactic Aramaisms as a Tool for the Internal Chronology of Biblical Hebrew Naʿama Pat-El 1. Introduction The question of dating the language of the books of the Hebrew Bible has been debated for many years. While the majority of scholars agree that biblical texts can and should be dated, a small group has consistently rejected this idea on the basis of various arguments, linguistic and otherwise. One of the major points of contention is whether the effects of language contact, primarily contact with Aramaic, have any merit when one is determining the date of a text. The crux of the matter thus far in the debate has been the lexicon: scholars who support the possibility of dating interpret the existence of Aramaic lexemes, subject to certain conditions, as an indication of lateness, while scholars who argue dating is not feasible claim that there could be other explanations for the occurrence of these lexemes in the Hebrew text. In the last decade, the debate has come to an unfortunate standstill: little new material is introduced into the debate, and no new angles are advocated. While the influence of Aramaic on the lexicon of Biblical Hebrew has been scrutinized repeatedly, the influence of Aramaic on the syntax of Biblical Hebrew has not been investigated extensively. Two reasons for this may be: (1) the syntax of Aramaic itself is not completely understood, and so it has not been fully described; and (2) many researchers may be unaware of significant advances made in language contact studies during the last 20 years. 1 Therefore, before I deal with the syntactic data of chronogical significance, a few preliminary methodological remarks are in order. In §2 below, I outline Author’s note:  I wish to thank Jo Ann Hackett, James Jumper, Noam Mizrahi, and Gary Rendsburg for their constructive comments on earlier drafts of this essay. I am also grateful to the editors for many useful comments that have improved the text and argumentation. All remaining errors are, of course, my own. Abbreviations used in this paper include: OSA = Old South Arabian; REL = relative pronoun. 1.  It should be noted that relevant studies in Semitic linguistics are also frequently overlooked to the detriment of the debate. I will occasionally remark on some of the more glaring omissions.

245

246

Naʿama Pat-El

the principles of both historical linguistics as it bears on syntactic change, and language contact theory as it bears on syntactic transference. In §3, I provide two case studies illustrating the application of these principles and their contribution to Hebrew chronological studies. In this essay, I propose that we move away from concentrating on Aramaic lexemes in Hebrew. Although this approach has contributed much to the discussion of dating, its benefits have been exhausted by now—not so much because all Aramaic lexemes have been exposed, but because the value of lexical interference in general is not strong enough to substantiate conclusively the relative dating of a text. Interference in other parts of the grammar constitutes much more solid evidence of interference. Indeed, syntactic change requires more extensive contact and is normally found only after some borrowed lexemes have already been naturalized. Therefore, syntactic interference is a better indicator of long and extensive contact (Moravcsik 1978). 2. Methodology 2.1.  Aramaic Influence on Hebrew: The State of the Debate The influence of Aramaic on Biblical Hebrew has been the subject of scholarly investigation for over a century. The focus so far has been mostly lexical, with scholars concentrating on lexical replacement, semantic calques, and loan words, many of which have already been identified (Eskhult 2003; Joosten 2005: 329; Wagner 1966; Hurvitz 2003). Texts showing a significant number of Aramaisms have been claimed to exemplify Late Biblical Hebrew. Simply identifying Aramaic lexemes in Hebrew texts, however, is not sufficient for dating. As one main advocate for linguistic dating of texts has noted, the identification of relevant Aramaisms for dating should be restricted to cases where lexical replacement can be textually verified and must be subject to a number of conditions (Hurvitz 1968; 2003). 2 The relevance of Aramaic to the chronology of Biblical Hebrew rests on three factors: (1) There are not enough extrabiblical Hebrew sources from the Iron Age through the Persian period (or from the tenth through the fourth centuries b.c.e.) to compare with the Biblical Hebrew text; hence the data that they provide are insufficient for dating the biblical material confidently; (2)  Aramaic as a dialect cluster is better attested than Epigraphic Hebrew. 2.  Thus, scholars convinced that a history of BH can be fleshed out from the text itself do not claim that every Aramaic lexeme is significant but, rather, that they carry some weight in dating texts, subject to a number of principles and in addition to other criteria. Unfortunately, Young, Rezetko, and Ehrensvärd (2008: vol. 1) are quick to dismiss the use of Aramaisms for dating. Some of their claims have merit and deserve serious intellectual engagement, but some are extraneous and misleading. In the following, I try to address some of the problematic aspects of their work. While criticizing specific studies is legitimate and potentially useful, dismissing the entire endeavor of linguistic dating out of hand is both simplistic and goes against supported linguistic practices (see more below).

Syntactic Aramaisms as a Tool

247

(3) The dating of Aramaic texts is, by and large, not controversial and thus serves as an extrabiblical textual source for periodization (see Sokoloff’s essay in this volume). Consequently, in the quest to date biblical texts, Aramaic can sometimes provide the requisite “linguistic peg.” Several scholars have expressed doubts about the validity of Aramaisms, especially lexical Aramaisms, as indicators of textual date. Some have pointed out that there may be other explanations for the existence of Aramaic lexemes in the biblical text (Rendsburg 2002; 2003; 2006) or have rejected loans as an indication of language change (Naudé 2003: 204). Young (1998) suggested, on the basis of 2 Kgs 18:26, that the intellectual literate elite were conversant in Aramaic more than a century prior to the fall of Jerusalem, and hence some Aramaic influence is evident even before the Babylonian Exile. Young, Rezetko, and Ehrensvärd (2008: vol. 1) go even further and argue that using Aramaic to determine internal biblical chronology is futile. They suggest, following Young (1993; 2002), that neither Canaanite nor Aramaic is a distinct linguistic entity, and thus each could theoretically use elements from the other. 3 As additional support for this proposal, they use Deir ʿAlla, which is neither Canaanite nor Aramaic yet contains a number of features found in both Canaanite and Aramaic. For Young, Rezetko, and Ehrensvärd (2008: 1.207–8), Deir ʿAlla is an example of a distinct linguistic entity that nevertheless shares linguistic features from other distinct linguistic entities. 4 The relationship of Deir ʿAlla to other Northwest Semitic languages has already been discussed at length by Huehnergard (1991: 291), who observed that the assumption that there was a mid-second-millennium dichotomy only between Canaanite and Aramaic is “unnecessary and unsupportable.” 5 Huehnergard further notes that Aramaic, Canaanite, and Ugaritic share a set of innovations that define them as distinct linguistic entities, and so each one differs from the language from which it split. 6 What are left after the split of these dialect clusters are dialects that have not been affected by the innovations found 3.  See, for example, their claim that “beneath the surface of Standard Aramaic a neat distinction between Aramaic and Canaanite becomes more problematic” (Young, Rezetko, and Ehrensvärd 2008: 1.203). 4.  Some of the arguments presented in Young, Rezetko, and Ehrensvärd (2008: 1.203–7) regarding the mixed nature of Deir ʿAlla were also used by Rainey (2007) to argue for the genetic identity between Aramaic and Hebrew (on the basis of the distribution of pʿl / ʿbd and other mostly lexical features). In Hackett and Pat-El forthcoming, we offer a lengthy explanation why these arguments are invalid in historical linguistics. 5.  Huehnergard (1991) is not included in the extensive bibliography in Young, Rezetko, and Ehrensvärd (2008: vol. 2). 6. For a discussion of the principles that underlie the linguistic division between branches, see Hetzron 1976; this is another important work that is absent from the bibliography in Young, Rezetko, and Ehrensvärd 2008. An instructive application of Hetzron’s prin­ ciples is Huehnergard 1995, which defines Aramaic according to its innovative features— that is, what makes Aramaic unique relative to Northwest Semitic.

248

Naʿama Pat-El

in Aramaic, Canaanite, and Ugaritic but still share some inherited features with them. Such a language is Deir ʿAlla. The features shared by Deir ʿAlla, Canaanite, and Aramaic are all inherited features, not innovations. 7 Deir ʿAlla is, therefore, not a language with a mixed system and thus does not support the hypothesis outlined in Young, Rezetko, and Ehrensvärd (2008: vol. 1). Young, Rezetko, and Ehrensvärd (2008: 1.208–12) list a number of supposed Aramaisms (all lexical and involving irregular sound changes) to prove that they exist in all strata of Biblical Hebrew and are not confined solely to late books. They further suggest that only words attested after the Babylonian Exile can be used as indicators of Aramaic influence on Hebrew, while words appearing in Old Aramaic texts, such as ʾiggeret, are “irrelevant to the question of the word’s chronology within Hebrew” (2008: 1.220). 8 They conclude that, because so many linguistic and cultural factors are at play, “the value of Aramaisms as a chronological marker is extremely dubious” (2008: 1.221). It is essential to note that the authors’ comments illustrate their unfamiliarity with the basic tenets of historical linguistics and the field of language contact, as well as with relevant studies of other, similar contact situations. In what follows, I illustrate how current linguistic theory in these two areas can help us evaluate the data useful for working out the relative and absolute chronology of Biblical Hebrew. 2.2.  Historical Linguistics The subject of historical linguistics is language change over time within populations. 9 As with evolutionary biology and the study of fossils, it is very 7.  The most important of these features are lack of a definite article, preterite prefix conjugation, suffix conjugation 3fs -t, and the N-stem. 8.  This argument makes little sense when viewed from the perspective of linguistic contact and change. For example, the third-person plural pronoun in English, they, is a borrowing from Old Norse. It is the common Scandinavian pronoun, so it obviously existed far before any contact with English. The fact that Old English had a different pronoun, hi, cements the chronology of the contact. In order to prove interference, we need to show that the feature was not present in the receiving language before contact with the source language but was present in the source (Thomason 2001: 94). If we cannot prove that a feature was present in Aramaic but not in Hebrew prior to contact, how can we prove which is the source language and which is the recipient? Hurvitz, in his many papers reiterating this point, did not devise a new linguistic criterion; most of his criteria are what all linguists dealing with contact-induced change use. Were scholars to adopt the criteria used by Young, Rezetko, and Ehrensvärd to evaluate loanwords, they would also need to exclude Hebrew dāt < Persian dāt, because it is attested in Pahlavi before the sixth century b.c.e., as well as all Aramaic words in Arabic. 9.  I am using terms regularly employed in Historical Linguistics literature, particularly in Historical Syntax. This is different from the terminology used by the generative school of Historical Linguistics, as used for example, in Naudé 2003. For Naudé (2003: 197–98), change relates to an individual speaker, whereas the spread of change to the entire population is called diffusion (2003: 199). However, it is widely accepted that languages cannot change

Syntactic Aramaisms as a Tool

249

rare to have at our disposal every language stage between two attested stages of a certain language. Even a dialect cluster as well attested as the Romance languages is fragmentary: we have records of several Latin dialects from earlier periods, but they are not the direct ancestors of French, Italian, Spanish, and 44 other Romance languages and dialects. 10 This is where historical linguistic methodology is useful, providing tools that allow us to explain linguistic phenomena and to suggest mechanisms of change by which languages such as French evolved from a hypothesized Latin-like language. Identification of linguistic variation is not proof of change but, rather, a descriptive assessment. Syntactic change is a relatively new subfield within historical linguistics. The approach used here posits that only three mechanisms are responsible for syntactic change: reanalysis, extension, and borrowing (Harris and Campbell 1995). Reanalysis is a mechanism that alters the underlying structure of a syntactic pattern while the surface manifestation remains unchanged. For example, in BH bôʾăkā gayʾ (1 Sam 17:52) ‘toward the valley’, literally ‘your [2ms] coming [to the] valley’, the originally infinitival form still carries a 2ms pronominal suffix (-kā) and has not lost its morphological features as an infinitive construct. The new structure has a different function (a preposition governing a noun) and constitutes a synchronically unanalyzable morphological form. Thus, the surface manifestation, its morphology, is that of an infinitive with a suffixed pronoun, while its function is that of a preposition. Extension involves a change of surface manifestation that does not involve immediate modification of underlying structure. For example, BH yiqtol of the sound verb reflects two historical morphological forms that are synchronically only distinguished syntactically (yiqtol vs. way-yiqtol). In the perfect, a similar distinction exists, although not on the basis of historic reality but, rather, as an extension of the duality in the prefix conjugation (qatal vs. wĕ-qatal). Borrowing is a mechanism in which a replication of a syntactic pattern is incorporated into the borrowing language (also called the “receiving language”) through the influence of a native pattern found in a contact language (also called the “source language”). These three mechanisms are not mutually exclusive, and in many cases more than one is responsible for a single case of syntactic change (Thomason 2001: 91). Borrowing, however, differs from reanalysis and extension, because it is an external mechanism, influenced (or driven) more by outside pressure than by internal structural factors. It can also affect change in both underlying and so fast that speakers of succeeding generations lose communicability; thus, a generational change cannot be a structural change and hence of no importance to historical linguists (Harris and Campbell 1995: 53). Generally, a localized change exhibited by an individual “represents what may but need not become a change” (Fischer 2007: 32–33, italics mine). 10. See Posner (1996: chap. 3) for a lengthy discussion of the relationship between Latin (and its variants) and the medieval and modern Romance languages.

250

Naʿama Pat-El

surface structures simultaneously because it relies on an outside system. Borrowing and the way it operates in Biblical Hebrew are the focus of this study. 2.3.  Languages in Contact Linguistic contact is a situation that results in linguistic borrowing. In the past 20 years, the effects of contact have been researched extensively, and a number of important works have been published. While some researchers focus on the speaker and the mental processes at play, others are more concerned with identifying contact outcome and what enables it to take place (MyersScotton 2002; Thomason 2001). According to recent studies, contact-induced change is not an isolated phenomenon and should be discernible in a language as a whole—in its phonology, morphology, and even syntax. Thomason (2001: 91–95) stresses a number of necessary conditions for identifying contactchange: (1) a source language must be identified; (2) both source and receiving languages should share phonological, morphological, or syntactic features; (3) a shared feature that was not attested in the receiving language prior to contact must be identified; and (4) this same shared feature must be attested in the source language prior to contact. 11 A special case of language contact involves a Sprachbund, a geographical area where three or more languages share a number of structural features as a result of contact (Thomason 2001: 99). 12 This term does not apply to a situation of contact between only two languages; otherwise, most languages in the world would be part of a Sprachbund. In order properly to identify such a linguistic area, there need to be some diagnostic features common in every language in the group within the area, regardless of the group’s genetic affiliation (Aikhenvald and Dixon 2006: 12). 13 Therefore, Hebrew and Aramaic on their own cannot be considered a Sprachbund; rather, they are languages in contact. The issue at hand is whether the structural similarities between them are due to shared inherited features or to contact. Borrowing is a useful concept in historical linguistics. Borrowed lexemes, or loanwords, are frequently used to help document older stages of a language. A well-known example is Germanic loanwords in Finnish that reflect an older stage of Germanic than the stage preserved in attested Germanic dialects. For example, Finnish kuningas ‘king’ provides evidence for the reconstructed proto-Germanic nominative singular ending *-az, which is not directly doc11. Contra Young, Rezetko, and Ehrensvärd (2008: 1.220), who argue that such lexemes cannot be used to indicate Aramaic influence because “the evidence from Aramaic does not indicate this to be a word exclusively of the postexilic period.” See also n. 9 above. 12.  This term was not applied as such to Hebrew and Aramaic, but the description of the relationship between these languages in Young, Rezetko, and Ehrensvärd (2008: 1.221) accords with this phenomenon. 13. See Tomić (2004) for a list of diagnostic traits of the Balkan Sprachbund, possibly the most famous example of a linguistic area.

Syntactic Aramaisms as a Tool

251

umented in any known dialect of Germanic (see Campbell 1999: 69–70 for details). This datum gives more credence to what otherwise would be only linguistic conjecture. Furthermore, borrowing is very often a window into the social dealings of speakers at the time when the text was composed. In other words, the quantity of borrowed elements is an indication of both the length of contact and the level of bilingualism, of the quantity and quality of speakers’ exposure to the contact language (Thomason and Kaufman 1988: 46ff.). 14 So, on what basis can we argue that a change is the result of contact rather than internal pressure? The answer involves evaluating attested data and the probability of all possible explanations. When causes other than contact are less likely and when evidence for contact exists, it may be the best explanation (Thomason 2008: 47). 15 This cannot always be fully ascertained; however, if a certain change, particularly a structural change, appears in a language lacking any evidence of development leading to it—“out of the blue” as it were—a contact-induced change is a reasonable interpretation. Syntactic borrowing is fundamentally different from lexical and phonological borrowing. Thomason (2001: 70–71) constructs a borrowing scale, suggesting four grades of contact, and in each she specifies which part of the grammar is likely to be affected and to what extent. According to the scale, the lexicon is affected with even casual contact (grade 1). Phonology is affected with more-intense contact (grade 2), a situation described as having a minority of speakers who are fluent bilinguals. Syntax is affected only in grade 3, which is defined as a situation of increased bilingualism in which both speaker attitude and social factors favor borrowing. In other words, syntactic changes are possible in situations of continuous and prolonged close contact between speakers of two languages, unlike lexical changes, which may occur even with passing contact. Moreover, syntactic interference happens in a situation when the source language is considered dominant over the receiving language for any number of reasons. Thus, syntactic change is an indicator of extensive contact as well as a socially favorable environment for language change through contact. Examining syntax rather than the lexicon allows researchers to avoid the type of suspicion raised in Young (1998)—namely, that a small group of bilinguals was responsible for Aramaic lexical items. While this may explain the existence of sporadic Aramaic lexemes in preexilic texts, it cannot account for 14.  These two indicators are essentially social rather than linguistic. Thomason has argued in various publications that the primary predictor of contact is social. See especially Thomason 2008 and the works cited therein. 15.  As was noted above, this does not mean that contact is the sole explanation. Some linguists argue that frequently more than one mechanism is involved in any particular change (Harris and Campbell 1995: 52). Many changes are a result of a chain-reaction process triggered by borrowing but spread through other mechanisms (Thomason 2008: 47–48).

252

Naʿama Pat-El

the appropriation of Aramaic syntactical patterns. These structural changes were the result of a social and linguistic situation known to have existed only after the Babylonian Exile. My approach in this essay therefore goes beyond the usual suspects: lexical items and irregular sound changes. I examine the results of extensive and protracted contact in a late historical period in contrast to an earlier time (CBH) when this sort of contact, if indeed it occurred, was minor and restricted. 3.  Syntax as a Tool in Linguistic Dating Syntax has been the focus of several studies of linguistic changes attested in the Bible (Kropat 1909; Polzin 1976). In a number of recent papers, Joosten (2005; 2006; 2008) has pointed to various syntactical changes, primarily in the verbal system, that distinguish CBH from LBH. All these changes are attributed to processes of decay and renewal—processes well attested in other languages—that, according to Joosten, happened in Biblical Hebrew independently due to internal pressure. Little attention has been given to syntactical change in Biblical Hebrew due to the external pressure of Aramaic. 16 In the following subsections, I present two case studies illustrating this. The first involves a known change in syntax, the underlying cause of which has not yet been explained. The second case involves a syntactical feature in Hebrew that has not been recognized as “borrowed” heretofore. In light of the discussion in §2.3 above, in order to prove contact-induced change, I need to show that (1) the source of the feature under discussion is Aramaic, (2) the feature is attested in both Aramaic and Hebrew, (3) the feature is not attested in CBH, and (4) the feature is attested in Aramaic texts earlier than LBH. Since some scholars have raised doubts about whether LBH is an actual linguistic entity, I will rephrase (4) as follows: the feature is attested at least in Official Aramaic or earlier. This formulation does not assume the existence of LBH. 3.1.  Object Markers and Object Suffixes Both Kropat (1909: 36) and, later, Polzin (1976: 28–31) note that the use of the direct object marker ʾet with pronominal suffixes has been radically reduced in the language of the Chronicler. 17 Polzin (1976: 29) notes that ʾet with suffixes appears 150 times in Chronicles with no parallel in Kings; however, in the vast majority of these cases, there is no other available option to mark a 16.  In a recent study (Pat-El 2008a), I illustrate the importance of Aramaic dialectal variation to the diachrony of the biblical text. The methodology used there is applied here as well. 17.  Polzin (1976: 30) evaluates the ratio of verbal suffixes to ʾet in the Chronicler at 8 : 1 in favor of verbal suffixes (not counting repeated forms). Note, however, that at no point was ʾet + suffixes more common than verbal suffixes; the claim made by Polzin is not about dominance but about frequency.

Syntactic Aramaisms as a Tool

253

pronominal object, 18 so the use of ʾet is obligatory rather than optional. Only in three cases does the Chronicler freely use ʾet with a pronoun when another option is possible. 19 Thus, the use of pronominal object suffixes has increased significantly at the expense of the object marker ʾet. 20 Polzin offers no explanation for the decrease of ʾet with suffixes in LBH but specifically lists this as one of the features of LBH not attributable to Aramaic influence (Polzin 1976: 28). 21 While the facts are indisputable, I disagree with Polzin about the source of the change and argue that this feature should be attributed to contact with Aramaic. Old Aramaic has a similar direct object marker, ʾyt (Sefire) or wt- (Hadad), 22 which was undoubtedly inherited from NWS (Garr 2004: 115–16). In Old Aramaic, this particle is attested with both nouns (Sefire II C 5f) and pronouns (Hadad 28); however, in Official Aramaic and in later dialects, this particle has only one attestation in an unclear context in the Assur Ostracon, a very early Official Aramaic text. In Official Aramaic, finite verbs regularly take suffixed pronominal direct objects. An innovation of Official Aramaic is the use of the preposition l- as a direct object marker, primarily when the object is animate and definite (Folmer 1995: 340). 23 In the book of Daniel, whose Aramaic is later than Official Aramaic, yāt occurs with a pronominal suffix only once. 24 In Late Aramaic, only l- is used as an object marker; hence, it is clear that the use of yt- was abandoned prior to Late Aramaic but began fading out as an object marker much earlier, probably in Official Aramaic. The Peshiṭta has very few instances of yt (yāt). This particle was so foreign to Syriac readers that it was interpreted as the noun ‘being’ (Payne-Smith 2007 [1879]: 1639–41; Sokoloff 2009: 586–87). This interpretation binds this word in the mind of the reader to the existential ʾīt, which has a different etymology. 18.  For example, when another pronoun is already attached to the verbal form (e.g., 1 Chr 8:8, ‫ן־ׁש ְלחֹו א ָֹתם‬ ִ ‫מ‬, ִ or when the verbal form is a 2mp perfect form, as in 2 Chr 12:5, ‫ַּתם‬ ֶ‫א‬ ‫עֲז ְַב ֶּתם א ִֹתי‬, which only rarely takes an object pronominal suffix throughout the Bible. For a full list of biblical references, see Polzin 1976: 77 n. 3. 19.  2 Chr 8:2, ‫ׁשלֹמֹה א ָֹתם‬ ְ ‫ ; ָּבנָה‬2 Chr 24:25, ‫ֲלּיִים ִּכי־ ָעזְבּו אֹתֹו‬ ִ ‫;ּב ַמח‬ ְ 2 Chr 28:23, ‫הֵם מ ְַעזְִרים‬ ‫אֹותם‬ ָ . 20. In Mishnaic Hebrew, finite verbs normally take pronominal suffixes. The particle ʾet with pronominal suffixes occurs in the following cases: with participles; when the object precedes the verb; with III-weak verbs and in some inconsistent cases (Azar 1995: 65–66). 21.  Others also considered the preference of object pronouns suffixed to ʾot- as a sign of a late stage of Hebrew. See, for example, Dan (1996: 360–61) on a similar feature in Jonah, which is taken as an indication of its late date. 22.  This particle is vocalized in Middle and Late Aramaic as yāt. 23.  There are two examples with a definite inanimate object preceded by l- (Cowley 1923: C 12:2, 5). 24.  In the Hebrew portion of the book of Daniel, ʾet does not occur with pronouns at all, while there are 22 occurrences of object pronominal suffixes (Polzin 1976: 31).

254

Naʿama Pat-El

This particle did not completely disappear in Aramaic: it was used in western Middle and Late Aramaic as an object marker and as a demonstrative. 25 Similarly, it did not completely disappear from Hebrew either: ʾet was used increasingly before subjects (Polzin 1976: 32). 26 Thus, while ʾet is still used in Chronicles its syntax is different than that of earlier stages of the language in that it is not normally used with object pronouns. 27 This seems to be in accordance with the loss of this same function of yt in Official Aramaic. While ʾet is regularly used in CBH as an accusative marker, this function is possibly no longer productive in the language of Chronicles, where the suffixed object marker is the common strategy used to mark pronominal objects. Furthermore, Aramaic shows a changing distribution of the accusative marker yt, culminating in its replacement and removal in Official Aramaic. The change in Official Aramaic is a result of the introduction of a new object marker; however, this process did not take place in Hebrew. In other words, there is no obvious motivation for the change in Hebrew. It seems that all four conditions outlined in §3 above apply here. Therefore, the reduced distribution of the objective function of ʾet in Chronicles can and should be attributed to Aramaic influence rather than to internal processes in Biblical Hebrew.  28 3.2. Causal Subordination Particle on the Basis of šelIn both Biblical Hebrew and Old Aramaic, the regular causal conjunction is kī. While Hebrew maintained kī as a causal conjunction throughout its history, in Aramaic a new causal conjunction, b-zy, replaced kī already in Old Aramaic (Cowley 1923: C 30:23; C 37:4). Outside Aramaic, the conjunction b-zy is 25. See Rubin 2005 for a suggested course of development. 26.  Note, however, that Rendsburg (1980: 66) has demonstrated that the position of ʾet before nominatives is sufficiently attested in preexilic texts to suggest that the distribution of this usage has not changed. 27.  Polzin (1976: 37) further argues that ʾet was pushed away by the increased use of l- to introduce direct objects, the former having developed under Aramaic influence. A demonstrative function is also regularly attested in Mishnaic Hebrew (Rubin 2005: 122–25). 28.  Morag (1974: 313–15) notes another change in the syntax of ʾet + suffixes, especially in Ezekiel and Jeremiah. The inflected object marker ʾotî is used where the inflected preposition ʾittî is expected. Morag (1974: 314) attributes the change to language contact (Neo-Babylonian in the case of Ezekiel, and Aramaic in the case of Jeremiah). Since the uninflected form ʾet stands for both the accusative marker and the preposition and since neither Aramaic nor Babylonian uses an independent accusative marker, bilingual speakers conflated the two paradigms and used them indiscriminately. This type of change, sometimes called “grammatical accommodation,” involves reinterpretation of a morpheme on the model of a phonetically similar but syntactically different morpheme in another language (Aikhenvald 2003: 2). Morag further suggests that “hypercorrection” (without using the phrase itself) came into play: speakers perceived ʾotî to be a more correct form, because they were aware of its existence but not of the rules governing its use and thus preferred it over ʾittî. See also Blau (1970: 15) for the phenomenon.

Syntactic Aramaisms as a Tool

255

also used in some other Semitic languages with a causal meaning, although it is not equally common in all of them; 29 see, for example, Old South Arabian: 30 (1)  OSA: ʿlhn nhpn wbnyhw . . . w-[yr]m ʾymn mlk Sbʾ hqnyw šymhmw tʾlb rymm . . . tltn-hn ʾṣlmn ḏḏhbn ḥmdm bḏt hwšʿhmw bstkmln kl ṭyb w-ṣrf tnṭʿw ʾAlhan Nahfan and his sons . . . and [Yari]m Aiman, king of Seba, dedicated to their patron deity, Taʾlab Riyam . . . these three golden statues as a praise, because Taʾlab helped them in completing (the planting of) all the Ṭyb-incense and Ṣrf-incense they had planted. (CIS 308:1–4) Unsurprisingly, in Biblical Hebrew, there are several examples of a causal ba-ʾăšer, though this conjunction is usually used with a spatial meaning, a regular function of the preposition bĕ-. 31 (2)  Spatial

ְ ‫ַוּיֵל‬ ‫ִמצָא‬ ְ ‫ֲׁשר י‬ ֶ ‫ְהּודה לָגּור ַּבא‬ ָ ‫ָעיר ִמּבֵית ֶלחֶם י‬ ִ ‫ָאיׁש ֵמה‬ ִ ‫ֶך ה‬ way-yēlek hā-ʾîš mē-hā-ʿîr mib-bêt leḥem Yĕhûdā lāgûr ba-ʾăšer yimṣāʾ The man went from Bethlehem in Judah to live wherever he could find a place. (Judg 17:8)

(3)  Causal 32

‫ׁשּתֹו‬ ְ ‫־א‬ ִ ‫ֲׁשר א ְַּת‬ ֶ ‫ם־אֹות ְך ַּבא‬ ָ ‫ְולֹא־חָׂשַ ְך ִמ ֶּמּנִי ְמאּומָה ִּכי ִא‬ wĕ-lōʾ ḥāśak mimmennî mĕʾûmā kî ʾim ʾôtāk bā-ʾăšer ʾatt-ʾištô He has not kept back anything from me, except you, because you are his wife. (Gen 39:9)

This pattern is also found as a causal conjunction in Late Biblical Hebrew, with the relative particle šeC: (4)  ‫ִׁשּכָח‬ ְ ‫ָאים הַּכֹל נ‬ ִ ‫ָמים ַהּב‬ ִ ‫ׁש ְּכבָר ַהּי‬ ֶ ‫ִכרֹון ֶל ָחכָם ִעם־ה ְַּכ ִסיל ְלעֹולָם ְּב‬ ְ ‫ִּכי אֵין ז‬ kî ʾên zikrôn le-ḥākām ḥim hak-kĕsîl lĕ-ʿôlām bĕ-šek-kĕbār hayyāmîm hab-bāʾîm hak-kol niškāḥ The wise, like the fool, will leave no enduring remembrance, because everything is long forgotten in the days to come. (Qoh 2:16)

29.  It is not found in Arabic; in Akkadian and Ethiopic, it has a similar yet not identical pattern. 30.  All translations in this section are my own. 31.  Note also Deir ʿAlla, where the conjunction carries spatial meaning: b-ʾšr rḥln yybl ḥṭrʾ ʾrnbn (I 11), ‘where (there are) ewes, a staff will lead hares’. 32.  See also Gen 39:9 with šaC-.

256

Naʿama Pat-El

As is obvious from all these examples, a causal conjunction *bi-ḏV- is attested in various Semitic languages and therefore is not indicative of contact between Aramaic and Hebrew. In Middle Aramaic, however, we find an innovative conjunction, b-dyl d- ‘because’, primarily in western Aramaic dialects: (5)  Palmyrene ṣlmʾ dnh . . . dy ʾqym lh bny šyrtʾ dy slq ʿmh mn prt w-mn ʾlgšyʾ b-dyl-dy špr lhwn This monument . . . which the members of the caravan who went with him from Forat and Vologesia erected for him, because he was good to them. (Cowley 1923: C 3916:1–3) (6)  Qumran Aramaic w-ydʿ ʾnh dy lʾ ykwl Rʿwʾl lmklyh mnk b-dyl dy huʾ ydʿ And I know that Reuel cannot hold it back from you, because he knows. (4Q197 frg. 4 ii 4) (7)  Samaritan Aramaic b-dyl d-ḥlp mly mʾbd ʾnʾ ḥyyw Because he changed my words, I will kill him. (Ben Hayyim 1988: TM 152b) In some Aramaic dialects, a preposition b-dyl ‘because of’ is also attested, primarily with suffixed pronouns. 33 This is significant, because prepositions are by far the most productive source of conjunctions in Aramaic. In all attested cases of homomorphic prepositions and conjunctions, the development of a preposition preceded the conjunction—never the other way around (Pat-El forthcoming). Thus, in dialects such as Qumran Aramaic, the conjunction bdyl d- ‘because’ should be considered a secondary development that probably began before the attested stage of this dialect. Furthermore, since this conjunction is already a regular part of the grammar of Middle Aramaic, it is safe to assume that it originated in some pre-Middle Aramaic dialect, possibly at the time of Official Aramaic. The specifics of the development of b-dyl d- in Aramaic do not concern us here, but we should note that dyl-, which is attested in the earliest sources of Aramaic (for example, Sef. III 20 from the mid-eighth century b.c.e.), has no causal meaning in Aramaic (Degen 1969: 59). Furthermore, the combined form b-dy-l is not attested in any other Semitic language. This must be, then, an Aramaic innovation that we should date to pre-Middle Aramaic, very likely to the time of Official Aramaic and geographically to western Aramaic dialects of that time. 33. In Tg. Onqelos: liṭā ʾarʿā bĕdilāk ‘The land is cursed because of you’ (Gen 3:17). In Qumran Aramaic: w-šbq lhwn ḥṭʾyhwn bdylh ‘He forgave their sins because of him’ (11QTgJob 38:3).

Syntactic Aramaisms as a Tool

257

Two Biblical Hebrew texts attest a similar combination, bĕ-šel-, with a causal meaning both as a preposition (bĕ-šel- [Jonah 1:7, 8, 12]) and as a subordinating particle (bĕ-šel ʾăšer [Qoh 8:17]). 34 In this context, it is important to note that the combination ʾăšer with the preposition l- is always possessive in Classical Biblical Hebrew. 35 There is no example in CBH of the pattern bĕ-šel with a causal meaning. 36 (8)  . . . . . . . . . ‫ָרעָה הַּזֹאת לָנּו‬ ָ ‫ֲׁשר ְל ִמי־ה‬ ֶ ‫ּבא‬ ba-ʾăšer lĕ-mî hā-rāʿā haz-zōʾt lānû On account of whom did this evil come upon us? (Jonah 1:8) ‫ׁשל‬ ֶ ‫ֲׂשה תַ חַת־ ַהּשֶׁמֶׁש ְּב‬ ָ ‫ֲׁשר נַע‬ ֶ ‫ֲׂשה א‬ ֶ ‫ָדם ִל ְמצֹוא אֶת־ ַה ַּמע‬ ָ ‫ִּכי לֹא יּוכַל ָהא‬ ‫ִמצָא‬ ְ ‫ָדם ְל ַבּקֵׁש ְולֹא י‬ ָ ‫ֲׁשר יַעֲמֹל ָהא‬ ֶ‫א‬ kî lōʾ yûkal hā-ʾādām limṣôʾ ʾet ham-maʿăśe ʾăšer naʿăśā taḥat haššemeš bĕ-šel ʾăšer yaʿămol hā-ʾādām lĕbaqqēš wĕ-lōʾ yimṣāʾ Humans cannot understand what is done under the sun; no matter how much they may labor to seek (an answer), they will not find (it). (Qoh 8:17)

(9) 

While Qoheleth is widely agreed to be a late book, the status of Jonah is still debated (Dan 1996: 344–45). I am not the first to suggest that bĕ-šel ʾăšer is part of the linguistic evidence supporting a late date (Dan 1996: 361). The question, however, is whether this particular feature is characteristic of a northern preexilic dialect of Hebrew rather than of a postexilic dialect.  37 Considering the evidence in Aramaic, we see that b-dyl d is indeed a western dialectal 34. This interpretation is hardly innovative. It was mentioned already in Ginsburg’s commentary on Qoheleth (Ginsburg 1861: 408), where the verses from Jonah are also mentioned. I thank Dr. Noam Mizrahi for the reference. More-recent notices of this feature in Jonah appear in Brenner 1979: 404; Landes 1982: 153*; and Dan 1996: 361. Note also Muraoka (2012), who suggests that the alternation between -‎ ‫ׁש ְּל‬ ֶ ‫( ְּב‬Jonah 1:7) and -‫ֲׁשר ל‬ ֶ ‫ַּבא‬ (Jonah 1:8) is an example of diglossia. 35.  For example, kōl ʾăšer lô ‘everything he had, all of his property’ (Gen 12:20). This pattern is common Semitic; see Pat-El and Treiger 2008. 36.  Note also that in the books of Qoheleth and Jonah, where the causal preposition / conjunction is used, the combination ʾăšer l- is not used with a possessive function (pace Waltke and O’Connor 1990: 335 n. 15). Landes (1982: 153*) claims that this pattern in Jonah is “modeled on Classical Hebrew b-ʾšr l,” though such a pattern is not attested in Classical Hebrew. For more information about Aramaic features in Jonah, see Dan 1996 and the bibliography therein. 37.  Northern, or Israelian Hebrew, refers to the relics of what is believed to be the dialect spoken in the biblical Kingdom of Israel. Because the biblical text was redacted in Jerusalem, many of the linguistic features of the northern dialect were leveled. Rendsburg (2002: 35–46; 2006), following Hurvitz (1968: 234–40), argues that certain Aramaic-like features are in fact charectaristic of the northern Hebrew dialect, rather than of Aramaic, and are therefore not indicative of a late text.

258

Naʿama Pat-El

feature, which normally would have been in contact with northern Hebrew; however, it is attested in Aramaic much later than the fall of Israel, and therefore must be late. 38 There is one notable difference between Hebrew and Aramaic in this respect: in Aramaic, normally when the prefix conjugation follows b-dyl d-, the conjunction carries a result meaning, 39 while a causal b-dyl d- is followed by a suffix conjugation or a nominal sentence. Two issues need to be borne in mind: first, the result meaning is common in eastern dialects (and Onqelos), not in western dialects. Second, shared features need not have the same distribution in the receptor language that they have in the source language (Thomason 2001: 93). All the elements of the preposition and conjunction shown above in Jonah and Qoheleth are original Hebrew features, but their new structural configuration as well as the causal function are the result of language contact. This is the most reasonable conclusion, because they are not attested anywhere else with this combination and function except in Aramaic, where the innovation can be accounted for historically (Pat-El forthcoming). Furthermore, had this particle been an original Hebrew innovation, it would probably not require an additional relative marker (bĕ-šel, ʾăšer), since in Hebrew, as in most Semitic languages, the relative particle is not required when a subordinating particle is used. This is indeed the case in Classical Hebrew. The requirement that a relative particle should introduce any type of subordinated clause, including adverbial clauses already marked as subordinated is an innovation of Aramaic (Pat-El 2008b). This feature is also found in Mishnaic Hebrew but was not fully integrated into the grammar of Hebrew beforehand. 40 Thus, the syntax of ba-ʾăšer lĕ- is typical of Aramaic and post-Biblical Hebrew but not of Classical Hebrew. Ba-REL lĕ- is only found in Jonah and Qoheleth and never appears in earlier books. Here too, it seems that the conditions specified in §2.3 apply: the source of the pattern is Aramaic, the pattern is attested at a certain point in both languages but not earlier than the assumed time of contact, and the feature probably developed during Official Aramaic. This distribution makes it a perfect candidate for an Aramaism: it replaces an earlier particle and has an

38.  A similar position is promoted in Dan (1996: 365–66), who argues that several Aramaic features in Jonah are not attested before Official Aramaic and hence are unlikely to be indicative of Israelian Hebrew. He further concludes that Jonah could not have been composed in an early period, though he is reluctant to attribute to the book a more specific date. 39.  See, for example, in Samaritan Aramaic: yhbt lh šmy bdyl d-lʾ ydḥl ‘I gave him my name so that he would not be afraid’ (Ben Hayyim 1988: TM 16b). 40. There are some instances of both this feature in earlier biblical books, but it is very rare.

Syntactic Aramaisms as a Tool

259

identifiable source. 41 The common causal conjunction kî is still used in both books. The use of an innovative and an older form synchronically is not uncommon in contact situations. The slow disappearance of kî is evident in texts from Qumran (Brin 1978: 20; Mor 2009: 196 n. 1097), the Judean Desert (Mor 2009: 195, 276), and Mishnaic Hebrew (Segal 1927: 146). Beyond replacing earlier forms, which could be a case of lexical replacement, the syntactic arrangement of the elements ba-REL lĕ- is not attested in CBH. 4.  Summary and Conclusion In this essay, I aim to illustrate the value of an overlooked aspect of language contact as a tool for internal dating. I argue that evidence of syntactic borrowing from Aramaic provides a far stronger case for a text’s postexilic date than lexical borrowing. As I argued above, syntactic borrowing is only possible when there is extensive contact between languages, while lexical borrowing may occur even when contact is casual, without widespread bilingualism. Thus, when an instance of syntactic borrowing is identified, it should be assumed to reflect a contact situation involving prolonged and intense linguistic contact as well as favorable social factors. Although contact between speakers of Aramaic and Hebrew existed throughout the history of Hebrew, the sort of contact that allows syntactic borrowing did not take place prior to the Babylonian Exile. Following Thomason, I list (above in §3) four conditions required to establish that a syntactic feature is indeed borrowed and not the result of internal change. Two cases are used to demonstrate these points. In the first case, the preference for using suffixed object pronouns over using ʾot- was a significant grammatical change that is observable in Chronicles and a number of other books. The disappearance of the object marker was not gradual; it happened abruptly and with no apparent internal motivation. I suggest that contact with Aramaic, which lost the marker much earlier, was the cause. Although Hebrew lacks evidence of the syntactic process, it is attested in Aramaic. The second case analyzed above, b-dīl d-, was an Aramaic innovation, the development of which can be traced within the language. The syntactical pattern, the process that created it, and its function are unique to Aramaic and are not attested in other Semitic languages. Hebrew borrowed the end result wholesale, without going through the same long process that took place in Aramaic and without showing any obvious motivation for the change.

41.  The common causal preposition in CBH is ʿal (for example, Gen 20:3; Num 22:32), which can also appear with ʾăšer as a causal conjunction (for example, 2 Sam 3:30). A more frequent causal conjunction in CBH is yaʿan (for example, 1 Sam 15:23). Yaʿan is not used in either Jonah or Qoheleth, and the causal meaning of ʿal may be attested once in Jonah 4:2 in the adverb ʿal-kēn ‘therefore’.

260

Naʿama Pat-El

So far, the debate has concentrated solely on Hebrew and its own changes. What I show is that the more we know about the syntax of Aramaic the better we can date its influence on Hebrew. This, of course, was already implied by Hurvitz (1968). If we suspect contact, we should carefully study the entire grammar and diachrony of both languages; otherwise, the set of features that we are using (lexicon, phonology) ends up being extremely small, and the features we identify will only be the most obvious (lexemes, phonemes). In fact, syntactical features, which are harder to detect, complete the picture of a vast and comprehensive contact between these two languages. Finally, my essential argument here is that much of the discussion on the history of Hebrew is detached from the general field of historical linguistics and languages in contact, to its great detriment. The phenomenon of postexilic Hebrew provides a fascinating case of linguistic contact and standardization. Analysis of this variety of Hebrew can rely on similar diachronic phenomena in other languages, and it can offer useful insights for the analysis of similar linguistic situations within them. References Aikhenvald, A. Y. 2003 Mechanisms of Change in Areal Diffusion: New Morphology and Language Contact. Journal of Linguistics 39: 1–29. Aikhenvald, A. Y., and Dixon, R. 2006 Areal Diffusion and Genetic Inheritance: Problems in Comparative Linguistics. Oxford: Oxford University Press. Azar, M. 1995 The Syntax of Mishnaic Hebrew. Jerusalem: The Academy of the Hebrew Language and University of Haifa. [Hebrew] Ben-Ḥayyim, Z. 1988 Tibat Marqa. Jerusalem: Academy of the Hebrew Language. Blau, J. 1970 On Pseudo-Corrections in Some Semitic Languages. Jerusalem: Israel Academy of Sciences and Humanities. Brenner, A. 1979 The Language of the Book of Jonah as a Criterion to Determine Its Date of Composition. Beth Miqra 24: 396–405. [Hebrew] Brin, G. 1978 Notes about the Temple Scroll. Leš 43: 20–28. [Hebrew] Campbell, L. 1998 Historical Linguistics: An Introduction. Cambridge, MA: MIT Press. Coleman, J. 1995 The Chronology of French and Latin Loan Words in English. Transactions of the Philological Society 93: 95–124.

Syntactic Aramaisms as a Tool

261

Cowley, A. E. 1923 Aramaic Papyri of the Fifth Century b.c. Oxford: Clarendon. Dan, B. 1996 The Language of Jonah in the Literature: A Reevaluation and Assessment. Beth Miqra 41: 344–68. [Hebrew] Degen, R. 1969 Altaramäische Grammatik: Der Inschriften des 10.–8. Jh. v. Chr. Wiesbaden: Deutsche Morgenländische Gesellschaft. Eskhult, M. 2003 The Importance of Loanwords for Dating Biblical Hebrew Texts. Pp. 8–23 in Biblical Hebrew: Studies in Chronology and Typology, ed. Ian Young. London: T. & T. Clark. Fischer, O. 2007 Morphosyntactic Change: Functional and Formal Perspectives. Oxford: Oxford University Press. Folmer, M. L. 1995 The Aramaic Language in the Achaemenid Period: A Study in Linguistic Variation. OLA 68. Leuven: Peeters/Departement Oosterse Studies. 2008 The Use and Form of the nota objectivi in Jewish Palestinian Aramaic Inscriptions. Pp.  131–58 in Aramaic in Its Historical and Linguistic Setting, ed. H. Gzella and M. L. Folmer. Wiesbaden: Harrassowitz. Garr, W. R. 2004 Dialect Geography of Syria–Palestine, 1000–586 b.c.e. Winona Lake, IN: Eisenbrauns. [Original: Philadelphia: University of Pennsylvania Press, 1985.] Ginsburg, C. D. 1861 Coheleth. London: Longman. Hackett, J. A., and Pat-El, N. forthcoming  On Canaanite and Historical Linguistics: A Rejoinder to Anson Rainey. In The North American Conference on Afroasiatic Linguistics, 39 February 2011, Austin, TX. Harris, A. C., and Campbell, L. 1995 Historical Syntax in Cross-Linguistic Perspective. Cambridge: Cambridge University Press. Hetzron, R. 1976 Two Principles of Genetic Reconstruction. Lingua 38: 89–198. Huehnergard, J. 1991 Remarks on the Classificatioon of the Northwest Semitic Languages. Pp. 282–93 in The Balaam Text from Deir ʿAlla Re-evaluated, ed. J. Hoftijzer and G. van der Kooij. Leiden: Brill. 1995 What Is Aramaic? Aram 7: 261–82. Hurvitz, A. 1968 The Chronological Significance of “Aramaisms” in Biblical Hebrew. IEJ 18: 234–40. 2003 Hebrew and Aramaic in the Biblical Period: The Problem of ‘Aramaisms’ in Linguistic Research on the Hebrew Bible. Pp. 24–37 in Biblical Hebrew: Studies in Chronology and Typology, ed. I. Young. London: T. & T. Clark.

262

Naʿama Pat-El

Jeffrey, A. 2007 The Foreign Vocabulary of the Qurʾān. Leiden: Brill. Joosten, J. 2005 The Distinction between Classical and Late Biblical Hebrew as Reflected in Syntax. HS 46: 327–39. 2006 The Disappearance of Iterative WEQATAL in the Biblical Hebrew Verbal System. Pp.  135–48 in Biblical Hebrew in Its Northwest Semitic Setting, ed. S. E. Fassberg and A. Hurvitz. Winona Lake, IN: Eisenbrauns. 2008 Diachronic Aspects of the use of wayĕhî in Biblical Narrative. Pp. 165–72 in Festschrift Avi Hurvitz, ed. by S. E. Fassberg and A. Maman. Jerusalem: Hebrew University. [Hebrew] Kropat, A. 1909 Die Syntax des Autors der Chronik verglichen mit der seiner Quellen. Giessen. Landes, G. M. 1982 Linguistic Criteria and the Date of the Book of Jonah. ErIsr 16 (Orlinsky volume): 147*–70*. Mor, U. 2009 The Grammar of the Epigraphic Hebrew Documents from Judea between the First and the Second Revolts. Ph.D. dissertation. Ben Gurion University of the Negev. [Hebrew] Morag, S. 1974 On the Historical Validity of the Vocalizaion of the Hebrew Bible. JAOS 94: 307–15. Moravcsik, E. A. 1978 Language Contact. Pp. 93–122 in Universals of Language Contact, ed. J. H. Greenberg, C. A. Ferguson, and E. Moravcsik. Stanford: Stanford University Press. Muraoka, T. 2012 A Case of Diglossia in the Book of Jonah? VT 62: 129–31. Myers-Scotton, C. 2002 Contact Linguistics: Bilingual Encounters and Grammatical Outcomes. Oxford: Oxford University Press. Naudé, J. A. 2003 The Transitions of Biblical Hebrew in the Perspective of Language Change and Diffusion. Pp. 189–214 in Biblical Hebrew: Studies in Chronology and Typology, ed. I. Young. London: T. & T. Clark. Nöldeke, T. 1903 A Review of Die Aramäische im Alten Testament Untersucht. ZDMG 57: 412–20. Pat-El, N. 2008a Traces of Aramaic Dialectal Variation in Late Biblical Hebrew. VT 58: 650–55. 2008b Historical Syntax of Aramaic: A Note on Subordination. Pp. 55–76 in Aramaic in Its Historical and Linguistic Setting, ed. H. Gzella and M. L. Folmer. Wiesbaden: Harrassowitz. forthcoming  Studies in the Historical Syntax of Aramaic. Piscataway, NJ: Gorgias.

Syntactic Aramaisms as a Tool

263

Pat-El, N., and Treiger, A. 2008 On Adnominalization of Prepositional Phrases and Adverbs in Semitic. ZDMG 158: 265–83. Payne-Smith, R. 2007 [1879]  Thesaurus Syriacus. Piscataway, NJ: Gorgias. Polzin, R. 1976 Late Biblical Hebrew: Toward an Historical Typology of Biblical Hebrew Prose. Missoula, MT: Scholars Press. Posner, R. 1996 The Romance Languages. Cambridge: Cambridge University Press. Rainey, A. F. 2007 Redefining Hebrew: A Transjordanian Language. Maarav 14/2: 1–15. Rendsburg, G. A. 1980 Late Biblical Hebrew and the Date of “P.” JANES 12: 65–80. 2002 Some False Leads in the Identification of Late Biblical Hebrew Texts: The Cases of Genesis 24 and 1 Samuel 2:27–36. JBL 121: 23–46. 2003 Hurvitz Redux: On the Continued Scholarly Inattention to a Simple Principle of Hebrew Philology. Pp. 104–28 in Biblical Hebrew: Studies in Chronology and Typology, ed. I. Young. London: T. & T. Clark. 2006 Aramaic-Like Features in the Pentateuch. HS 47: 163–76. Rubin, A. D. 2005 Studies in Semitic Grammaticalization. HSS 57. Winona Lake, IN: Eisenbrauns. Segal, M. H. 1927 A Grammar of Mishnaic Hebrew. Oxford: Clarendon. Sokoloff, M. 2009 A Syriac Lexicon. Winona Lake, IN: Eisenbrauns / Piscataway, NJ: Gorgias. Thomason, S. G. 2001 Language Contact: An Introduction. Edinburgh: Edinburgh University Press. 2008 Social and Linguistic Factors as Predictors of Contact-Induced Change. Journal of Language Contact: THEMA 2: 42–56. http://www. jlc-journal.org. Thomason, S. G., and Kaufman, T. 1988 Language Contact, Creolization, and Genetic Linguistics. Berkeley: University of California Press. Tomić, O. M. 2004 The Balkan Sprachbund Properties: An Introduction. Pp.  1–55 in Balkan Syntax and Semantics, ed. O. M. Tomić. Amsterdam: Benjamins. Wagner, M. 1966 Die lexikalischen und grammatikalischen Aramaismen im alttestamentlichen Hebräisch. BZAW 96. Berlin: Alfred Töpelmann. Young, I. 1993 Diversity in Pre-Exilic Hebrew. Tübingen: Mohr. 1998a Israelite Literacy: Interpreting the Evidence—Part I. VT 48: 239–53. 1998b Israelite Literacy: Interpreting the Evidence—Part II. VT 48: 408–22. 2002 The Languages of Ancient Samʾal. Maarav 9: 93–105. Young, I.; Rezetko, R.; and Ehrensvärd, M. 2008. Linguistic Dating of Biblical Texts. 2 vols. London: Equinox.

The “Linguistic Dating of Biblical Texts” Comments on Methodological Guidelines and Philological Procedures Avi Hurvitz 1.  Introduction and Objectives In recent years, a new thesis has been introduced into the academic discourse concerning the diachronic study of Biblical Hebrew (= BH). The thesis requires that scholars engaged in studying the linguistic history of BH reexamine and reassess certain time-honored methodologies long considered reliable for dating purposes. Most of these methodologies focus on clarifying major or minor semantic shifts of a given word, different words with the same meaning used in identical contexts, verbs and nouns from the same consonantal root that function similarly in different conjugations and nominal patterns, different syntactic structures that fulfill the same purpose, and more. Linguists and philologists attempt to explain these sorts of phenomena in various ways. Sometimes the best explanation, based on patterns of distribution within biblical compositions from different periods, is the historical explanation. For more than a century of modern research, beginning in the first decades of the twentieth century, there has been a broad consensus that it is possible to distinguish between Early and Late Biblical Hebrew and to explain examples of the above-mentioned phenomena as the result of diachronic changes. The new thesis has been formulated as follows: “[W]e suggest a new synthesis where ‘EBH’ [Early BH] and ‘LBH’ [Late BH] are understood as coexisting styles of literary Hebrew throughout the biblical period” (Young, Rezetko, and Ehrensvärd 2008: 2.72). The challenge of this nondiachronic approach lies in its repeated objection to the current scholarly use of the hypothesis of diachrony to explain the differences between EBH and LBH. It maintains Author’s note:  I would like to extend a special word of thanks to my research assistant, Aaron Hornkohl, who has made invaluable suggestions in the course of my preparing this essay for publication. My sincere thanks are also due to Hani Davis, whose expertise and constructive comments helped to improve the manuscript. Finally, I am grateful to Ziony Zevit and Cynthia Miller-Naudé for their meticulous care and academic acumen in bringing this article to its final publication.

265

266

Avi Hurvitz

that faulty assumptions underlying the prevailing methodologies and a general misunderstanding of how historical research in language history ought to proceed undermine the validity of linguistic conclusions and historical claims about Hebrew. The thesis of “coexisting” styles, a nondiachronic alternative explanation for all the relevant data, rests on these claims. With this in mind, I would like to set forth the insights I have gained from reviewing the methodological and philological guidelines required for a meaningful diachronic investigation of BH texts. In order to assess the viability of the new thesis quoted above, I will reexamine the BH nouns ‫ ִא ֶּגרֶת‬and ‫מ ְד ָרׁש‬, ִ and consider the data from the perspective of both the diachronic and the new, non-diachronic, approaches. 1 In what follows, these lexemes serve as concrete test-cases for probing the validity of the new thesis. 2 I will also briefly underscore the importance of the Imperial Aramaic scribal formulary from the Persian period for understanding the development of LBH. 2.  The Case of ‫   ִא ֶּגרֶת‬3 Table 1.  ‫ִא ֶּגרֶת‬

I.  The Biblical Evidence 10 occurrences in LBH: Esther (2×); Nehemiah (6×); Chronicles (2×) 3 occurrences in BA: Ezra; for instance: Esth 9:29 [ versus [1 Kgs 21:11

Queen Esther . . . His townsmen . . . Jezebel had in  did as structed them . . .

wrote a . . . letter

(‫)אגרת‬

about Purim

in the

letters

(‫)ספרים‬

she had sent

] ]

Neh 6:19:

Tobiah

sent

letters

(‫)אגרות‬

to intimidate me

2 Chr 30:1 [ versus [2 Sam 11:14 In the morning, [1 Kgs 21:7–8 His wife [2 Kgs 10:1 [2 Kgs 20:12

Hezekiah . . .

wrote

letters

(‫)אגרות‬

to Ephraim

David Jezebel . . . Jehu King Berodach . . .

wrote a wrote wrote sent . . .

letter letters letters letters

(‫)ספר‬ (‫)ספרים‬ (‫)ספרים‬ (‫)ספרים‬

] to Joab ] in Ahab’s name ] . . . to Samaria ] . . . to Hezekiah ]

1. “The most impressive illustrations of diachronic development in Biblical Hebrew come from the lexicon [emphasis mine­—A.H.]” (Joosten 2005: 328). 2.  For earlier discussions on the nature of LBH and its historical position within the Hebrew language, as well as on the specific methodological strictures to be adopted in diachronic studies, see Hurvitz 1997; 2000; 2006. 3. See Hurvitz 1997: 311–13; 2000: 150–51; 2006: 200. The arguments of those who challenge the traditional diachronitc research (to be discussed below) do not always do justice to the views and opinions expressed in the three aforementioned articles and require a brief overview. For further discussions and bibliographical references, see the articles listed in this note.

The “Linguistic Dating of Biblical Texts”

267

II.  The Extrabiblical Sources Elephantine Aramaic (5th cent. b.c.e.): ‘If then it be . . . good to my lord, let a letter (‫ )אגרת‬be sent . . . to . . . the officer’ (Driver 1957: 33, Letter X, line 2) Bar-Kokhba (134–135 c.e.): ‘Letter (‫ )אגרת‬of Shimʿon Bar Kosiba, peace to Yehonatan son of Baʿaya’ (Yadin 1961: 43, no. 4) Palmyrene Inscription (137 c.e.): ‘Germanicus Caesar, in the letter (‫)אגרתא‬ which he wrote to Statilius, explained that . . .’ (Cooke 1903: 329, Tariff IIC, line 5) Mishnaic Hebrew: ‘R. Judah said: Letters (‫ )איגרות‬once came from beyond the sea to the sons of the High Priests’ (m. ʾOhal. 17:5) Targumic Aramaic and Syriac (Peshiṭta): 2 Sam 11:14 (MT): ‘In the morning, David wrote a letter (‫ )ספר‬to Joab’ 2 Sam 11:14 (Tg. J.): ‘In the morning, David wrote a letter (‫ )אגרתא‬to Joab’ 2 Sam 11:14 (Syriac): ‘In the morning, David wrote a letter (‫ )אגרתא‬to Joab’

2.1.  Distribution Pattern within BH The noun ‫—א ֶּגרֶת‬meaning ִ ‘written communication, letter, epistle, edict’— is attested ten times in BH, all in the distinctive LBH corpus and all dated to the Second Temple period: Esther (2×), Nehemiah (6×), and Chronicles (2×). In addition, the word is recorded three times in the Aramaic chapters of the book of Ezra, which also reflect the late stratum of biblical literature.  4 2.2.  Linguistic Contrast In Classical Biblical Hebrew (CBH), also known as Standard Biblical Hebrew (SBH), the function of ‫ ִא ֶּגרֶת‬is consistently fulfilled by its ancient equivalent ‫ ֵספֶר‬, which refers not only to letters but also to documents of various lengths, from letters and missives to extensive narratives, chronicles, and collections of laws (Exod 24:7, Deut 24:1, 2 Sam 1:18, 2 Kgs 22:11). Similarly, in preexilic Epigraphic Hebrew (the Lachish Letters; see Donner and Röllig 2002: no. 193, lines 5, 9, 10, 11, 19; no. 195, lines 6–7; no. 196, lines 4, 14), as well as in Late Bronze Age Ugaritic (del Olmo Lete and Sanmartín 2004: 767–69, s.v. spr [II]), the word used for ‘letter’ is ‫ ֵספֶר‬, not ‫א ֶּגרֶת‬. ִ The extensive and varied range of linguistic data adduced above—biblical and extrabiblical, literary as well as epigraphic, Hebrew and Aramaic alike—provides striking evidence that we are dealing here with a case of chronological development, in which the (more recent) LBH term for ‘letter’, ‫א ֶּגרֶת‬, ִ gradually pushed aside and eventually replaced its (more ancient) CBH counterpart, ‫ ֵספֶר‬, in its specific meaning ‘letter’ (see, for instance, BDB 8b; Bendavid 1967: 64; Polzin 1976: 126; Sáenz-Badillos 1993: 117; Eskhult 2003: 22; Polak 2006:120). 5 4.  Rezetko (2007: 399) presents this linguistic material with certain inaccuracies in that he mixes Hebrew and Aramaic references in a list of “terminology for official ‘letter’ in BH.” 5.  Needless to say, ‫ ֵספֶר‬, in its more common usage as ‘book, literary composition’, continued to be part and parcel of the standard Hebrew vocabulary in subsequent years as well.

268

Avi Hurvitz

2.3.  Extrabiblical Sources Besides appearing in the OT, ‫ ִא ֶּגרֶת‬is frequently used in the Aramaic papyri of the Persian period, in texts contemporary with the LBH corpus. The term ‫ ִא ֶּגרֶת‬is also well documented in postbiblical Aramaic (the Bar-Kokhba Letters, “Jewish Aramaic,” Syriac, and more) and Rabbinic Hebrew (RH—in both tannaitic and amoraic writings; see above). Although attested in an Old Aramaic text dated to the seventh century b.c.e. (Folmer 1995: 712; Mankowski 2000: 24; Eskhult 2003: 13 n. 7; Rezetko 2007: 399), ‫ ִא ֶּגרֶת‬is widely considered an unmistakable hallmark of the Northwest Semitic (NWS) linguistic milieu prevailing during and after the Second Temple period—that is, beginning in the last quarter of the sixth century b.c.e. 2.4.  Objections and Response The adherents of the nondiachronic approach attempt to demonstrate the methodological inadequacy of the guidelines underlying the diachronic investigation of ‫א ֶּגרֶת‬: ִ the word is attested already in Aramaic texts from the Neo-Assyrian period . . . i.e., contemporary with EBH. Therefore, the evidence from Aramaic does not indicate this to be a word exclusively of the postexilic period. The Aramaic evidence is thus strictly irrelevant to the question of the word’s chronology within Hebrew [emphasis mine—A.H.]. (Young, Rezetko, and Ehrensvärd 2008: 1.220)

The fact that ‫ ִא ֶּגרֶת‬is attested in Old Aramaic is very interesting but irrelevant to the argument. The crucial question here is not “When was ‫ ִא ֶּגרֶת‬first used in Aramaic?” but, rather, “When is it first attested in BH?” The most that we can deduce from the presence of ‫ ִא ֶּגרֶת‬in Old Aramaic is that it was theoretically possible for Hebrew to have borrowed it from Aramaic prior to the exile. When that possibility was actually realized is another matter altogether. An examination of the term’s distributional pattern within BH—namely, the actual documented instances of ‫ ִא ֶּגרֶת‬in the individual books of the Hebrew Bible (see table 1)—indicates that there are ten occurrences, limited exclusively to Esther, Nehemiah, and Chronicles, all of which were composed during the Second Temple period. This is the sort of concrete, textual evidence that is indispensable to any philological endeavor, because it is grounded in fact rather than conjecture. A striking parallel, in which a group of ancient words native to Akkadian was borrowed by BH late in its history, is provided by the names of the months in the Babylonian calendar. The introduction into BH of the loanwords ‫ּכ ְסלֵו‬, ִ ‫טבֵת‬ ֵ , ‫ׁשבָט‬, ְ ‫ֲדר‬ ָ ‫א‬, ‫נִיסָן‬, ‫סיוָן‬, ִ and ‫ אֱלּול‬is assigned in scholarly literature to the postexilic period: “The Imperial Aramaic names are clearly derived from the NB/LB calendar” (Kaufman 1974: 114). However, within Akkadian itself, these names are attested in pre-Israelite times! The ancient attestation of the month names in Akkadian does not, then, exclude, a priori, the possibility

The “Linguistic Dating of Biblical Texts”

269

that their penetration as loanwords into Hebrew took place hundreds of years later under changed historical and cultural circumstances. Scholarly consensus maintains that 70 years of exile in Babylonia and the return to Zion after the Persian conquest are the circumstances that brought these names into regular use in LBH. 6 Interestingly, even the sages of talmudic literature, who ordinarily did not approve of foreign influence on Jewish life and customs, recognized that the names of the months came up with them [the exiles returning to Palestine] from Babylonia: at first [they wrote], “in the month of Ethanim” [1 Kgs 8:2] . . . “in the month of Bul” [1 Kgs 6:38] . . . “in the month of Ziv” [1 Kgs 6:1, 37], but from the time that they came up [to Jerusalem from Babylonia] they wrote, “and it came to pass in the month of Nisan” [Neh 2:1] . . . ; “in the month of Kislev” [Neh 1:1] . . . “in the tenth month, which is the month of Tevet” [Esth 2:16]. 7

Needless to say, these sages were specialists in neither historical linguistics nor comparative Semitics, yet their philological perception was sharp enough to detect the non-Hebraic origin of the month names as well as the historical period during which they were borrowed. The role of Aramaic in discussions of EBH, LBH, and postbiblical Hebrew is of cardinal importance. It was a major Semitic language in early antiquity and the foreign language with which Hebrew speakers in the land of Israel had the most contact from the tenth century b.c.e. until the Hellenistic period. By the eighth century, it was serving as the international diplomatic language throughout the Levant, and because many of the exiles of 597 and 586 b.c.e. were settled in Aramaic-speaking areas in Babylonia, bilingualism was common among the exiles who returned to Zion. However, indiscriminate references to “Aramaic” in general, without distinguishing among dialects, geography, and periodization cannot yield meaningful results for historical linguistics. Any study that draws on comparative evidence between Hebrew and Aramaic must factor in these sorts of data (see Hurvitz 2003). This also applies to the Old Aramaic inscription in which ‫ ִא ֶּגרֶת‬is first documented (see above, §2.4). The contention of those who challenge the diachronic analysis, that the “Aramaic evidence is . . . strictly irrelevant to the question of the word’s chronology within Hebrew” (Young, Rezetko, and Ehrensvärd 2008: 1.220), requires much revision.

6.  See the standard biblical lexicons (under the different names of the months) and encyclopedias (in entries dealing with the calendars). 7.  Y. Roš Haš. 1.2, 56d (Venice ed.): ’‫ בראשונה ‘בירח האיתנים‬.‫ עלו בידם מבבל‬′‫שמות חדשי‬ ‫ מיכן והילך ‘ויהי בחדש ניסן שנת עשרים’; ‘ויהי‬. . .’‫ בראשונה ‘בירח זיו‬. . .’‫ בראשונה ‘בירח בול‬. . . ‫בחדש כסליו שנת עשרים’; ‘בחדש העשירי הוא חדש טבת‬. Note that the line of argumentation underlying the talmudic sages’ statement fulfills the above-mentioned third criterion, “Linguistic Contrast” (§2.2).

270

Avi Hurvitz

They also raise a second objection: “[T]he distribution and usage of these words [‫ ִא ֶּגרֶת‬and ‫— ֵספֶר‬A.H.] in Esther and Chronicles are not haphazard. The noun [‫—א ֶּגרֶת‬A.H.] ִ is found in contexts that deal with letter-writing in a socioreligious context. This type of situation is not attested in EBH literature [emphasis mine—A.H.]” (Young, Rezetko, and Ehrensvärd 2008: 1.72). Following R. Bergey (1983: 148–49), who distinguished between two variations in the meaning of BH ‫( ִא ֶּגרֶת‬one a “political” document and the other a “socioreligious”), Rezetko (2007: 399–400) has argued that the difference between these two usages cannot be matched up with an earlier synonym. Linguistic contrast, our third criterion for dating purposes, requires that we demonstrate a “contrast” between two seemingly identical alternatives (see above, §2.2); this does not imply, however, that every semantic nuance or shade of meaning is necessarily indicative of a datable linguistic shift. Each case must be evaluated on the basis of concrete evidence. It is therefore regrettable that those who challenge the diachronic approach fail to mention Bergey’s conclusion, which is that ‫ ִא ֶּגרֶת‬in Hebrew is a late lexeme regardless of its exact semantic nuances. Bergey (1983: 149) explicitly states: “‫ אגרת‬penetrated the literary Hebrew lexical stock at some point in the post-exilic period where it shared, together with the already commonly used ‫ספר‬, the semantic sphere ‘letter.’” 8 In other words, regardless of the precise semantic sense suggested for the lexical item ‫ ִא ֶּגרֶת‬in different BH texts, the word’s very emergence on the biblical scene as a viable alternative to ‫ ֵספֶר‬serves as a late chronological marker indicative of postexilic times. Therefore, Bergey’s conclusion corresponds with our diachronic analysis (§§2.1–2.3). Demonstrating the status of the two lexemes along the axis of time in three consecutive stages of the Hebrew language, table 2 graphically illustrates this line of linguistic development. 3.  The Case of ‫ ִמ ְדרָׁש‬and ‫  ָּדרַ ׁש‬9 The noun ‫ ִמ ְד ָרׁש‬is attested twice in the Hebrew Bible, both times in Chronicles (2 Chr 13:22, 24:27). Moreover, both verses appear in “parallel passages” 8.  There are at least two additional problems with Rezetko’s line of argumentation. First, while the usage of ‫ ִא ֶּגרֶת‬in Esther and Chronicles may be restricted to “socioreligious” contexts, as noted by Rezetko (2007: 400 n. 111), its use in Nehemiah—namely, for “official correspondence, to / from King Artaxerxes and to / from ‘governor’ Nehemiah or Judean ‘nobility’ and their ‘opponents’ Sanballat and Tobiah . . .”—is strikingly “political” and therefore provides direct “linguistic contrast” for the “political” use of ‫ ֵספֶר‬in early contexts. Second, contrary to the contention that ‫ ֵספֶר‬was not used for “sociopolitical” correspondence in CBH—in Rezetko’s words (2007: 400 n. 111): “As far as I can tell, Samuel, Kings, Isaiah and Jeremiah do not attest letter-writing in a socio-religious context”—one can cite Jeremiah’s letter (‫ ; ֵספֶר‬Jer 29:1) to the exiles in Babylon. 9.  Hurvitz 1972: 131–34; 1995: 1–10; 2006: 201.

The “Linguistic Dating of Biblical Texts”

271

Table 2.  Use of ‫ ֵספֶר‬and ‫ ִא ֶּגרֶת‬in Different Historical Phases of Hebrew ‫ִא ֶּגרֶת‬

‫‘( ֵספֶר‬letter’)

CBH



+

LBH

+

+

RH

+



1.  In earlier biblical writings, the attested term for ‘letter’ is ‫ ֵספֶר‬exclusively. 2.  In late biblical writings, both classical ‫ ֵספֶר‬and the more recent ‫ ִא ֶּגרֶת‬may be found side by side. 3.  In postbiblical literature (RH), ‫ ֵספֶר‬meaning ‘letter’ is no longer current and is consistently replaced with its late counterpart ‫א ֶּגרֶת‬. ִ

in Kings (1 Kgs 15:7, 2 Kgs 12:20[19]). However, the parallel passages lack the word ‫( ִמ ְד ָרׁש‬see table 3, pp. 272–73). 10 These data alone do not necessarily imply that the word is a postexilic innovation, because many more than two occurrences of the word in a single book are required to make a compelling argument for linguistic lateness. Two further pieces of linguistic evidence substantiate the lateness of the term ‫מ ְד ָרׁש‬: ִ the semantic range of ‫ דרׁש‬in Hebrew generally and the root’s distribution pattern in BH in particular. 3.1.  Semantic Range As far as the semantic range of ‫ ָּדרַ ׁש‬is concerned, a clear distinction is discernible between its broad connotations in BH, ‘look for, seek, search for (objects, persons, and particularly God [or his oracular word])’, and its narrow, specialized connotation in MH, where ‫ ָּדרַ ׁש‬is associated specifically with God’s law (‫)ּתֹורה‬ ָ and commandments (‫)מ ְצֹות‬, ִ conveying the potential semantic nuance of ‘study, investigate, interpret (a written [scriptural] text)’. This late semantic development is also well attested in the Dead Sea Scrolls, the vocabulary of which shares a considerable number of postclassical isoglosses with LBH. As noted, for instance, by Qimron and Strugnell (1994: 107): “Broadly speaking, the language of the Qumran sectarian literature is similar (especially in phraseology and syntax) to the language of those biblical books that were written in the Post-exilic period.” 3.2.  Distribution Pattern As for ‫’דרׁש‬s distribution pattern within BH—in both its nominal (‫)מ ְד ָרׁש‬ ִ and verbal (‫)ּדרַ ׁש‬ ָ manifestations—the basic meaning ‘look for, seek’ is found consistently throughout biblical literature, but the potential instances of its 10.  In one case, CBH ‫ ֵספֶר‬in Kings is replaced by ‫ ִמ ְדרָׁש‬in Chronicles; in the other, ‫ ִמ ְדרָׁש‬has no counterpart in Kings and is simply an addition. The Chronicler, influenced by the literary standards of his time, was obviously responsible for recasting the early phrase.

272

Avi Hurvitz Table 3.  Noun ‫ ִמ ְדרָׁש‬/ Verb ‫ ָּדרַ ׁש‬+ Commandments/Torah

I.  The Biblical Evidence

LBH: ‫ מדרש‬appears twice: Chronicles (2×) 2 Chr 13:22

the . . . acts of Abijah . . . are written in the Midrasha

2 Chr 24:27

. . .

[

of the prophet Iddo

are written in the Midrash on the book of the Kings

versus

[1 Kgs 15:7

] the . . . acts of Abijam . . . are . . . written in the ________

[2 Kgs 12:20(19) the    acts of Joash . . .

book of the Chronicles ]

are . . . written in the ________

book of the Chronicles ]

LBH: ‫ דרש‬occurs twice: Ezra (1×); Chronicles (1×) Ezra 7:10

For Ezra had set his heart

to studyb

1 Chr 28:8

observe (‫)ׁש ְמרּו‬ ִ and

keep (‫ׁש ְמרּון‬ ְ ‫)ׁשמֹור ִּת‬ ָ

the law

of the Lord

(

′‫אֶת ּתֹורַ ת ה‬

‫(ל ְדרוׁש‬ ִ

  seek outb all the commandments

of the Lord

(

′‫  ִמ ְצֹות ה‬. . .

‫(ו ִד ְרׁשּו‬ ְ

____

the commandments

of the Lord

( . . .′‫אֶת ִמ ְצֹות ה‬

_____)]

____ . . .

his testimonies

[ versus ] [Deut 6:17 [

(

‫ְועֵד ָֹתיו‬

_____)]

(

____ . . .

his statutes

‫ֻּקיו‬ ָ ‫ְוח‬

_____)]

[1 Sam 9:9

when a man went

to inquire

________

of God . . .

( ‫____ אלהים‬

‫](ל ְדרוׁש‬ ִ

[Gen 25:22

so she went

to inquire

________

of the Lord

(

′‫ אֶת ____ ה‬. . .‫]) ִ ל ְדר ֹׁש‬

[Hos 10:12

it is the time

to seek

________

the Lord

(

′‫אֶת ____ ה‬

[

[1 Kgs 22:5 Jehoshaphat said . . .

inquire . . . for the word c of the Lord

(

‫])ל ְדרוׁש‬ ִ

′‫ אֶת ְּדבַר ה‬. . .‫])ד ָרׁש‬ ְּ

a.  In Chronicles, the noun ‫ ִמ ְד ָרׁש‬is rendered ‘story’ (2 Chr 13:22 rsv/njps; 24:27 njps) and ‘commentary’ (24:27 rsv). b.  In Ezra, the verb ‫ ָּדרַ ׁש‬is rendered ‘to study’ (Ezra 7:10 [rsv/Njps]); in Chronicles, it is translated ‘seek out’ (1 Chr 28:8 [rsv]) and ‘apply yourselves to’ (1 Chr 28:8 [njps]). c.  ′‫ דבר ־ה‬here “implies a one-time ‘saying’ of God, whereas his ‘sayings’ as written in the Law represent, of course, instructions and commandments of a permanent nature” (Hoftijzer, in Hurvitz 1995: 7 n. 19).

specialized sense ‘study, interpret’ are confined exclusively to the distinctive LBH corpus (3× in Chronicles, 1× in Ezra) and other works in postbiblical Hebrew (see table 3 above). It is widely recognized that this semantic shift in the use of ‫ ָּדרַ ׁש‬reflects a late development in the history of biblical literature. M. Gertner (1962: 3) formulates this notion very well as follows: “Instead of priest and prophet came rabbi and scholar, and instead of prophecy and the ‘word of God’ [′‫]ּדבַר־ה‬ ְ came interpretation and the ‘word of Torah’ [‫־ּתֹורה‬ ָ ‎‫]ּד ַבר‬.” ְ Therefore, whatever the exact English translation of the texts cited above may be (see notes a, b,

273

The “Linguistic Dating of Biblical Texts” Table 3.  Noun ‫ ִמ ְדרָׁש‬/ Verb ‫ ָּדרַ ׁש‬+ Commandments/Torah (cont.)

II.  The Extrabiblical Sources 1QS 6:6 [

someone must always of the be engaged in study

Law, day and night

)‫יומם ולילה‬

law

(‫] )____ובתורתו יהגה יומם ולילה‬

d

‫(דורש בתורה‬

versus

] his

[Ps 1:2

and on

he meditates day and night

1QS 8:15

This means the expounding of the Law (‫)מדרש התורה‬, decreed by God through Moses.

M. Šeqal. 1:4 The priests used to expound this scripture (‫ )דורשים מקרא זה‬to their advantage. Mek. R. Ish., This is one of the three expressions in the Torah (‫ )בתורה‬which R. Ishmael used to interpret Exod 21:19 (‫ דורש‬. . . ‫ )היה‬as being figurative. (Lauterbach 1935: 53)

d.  The wording of 1QS here clearly echoes the formulation used in Ps 1:2, as well as in Josh 1:8 (see Hurvitz 1972: 132), all of which employ the three phraseological components ‫ּתֹורָה‬, ‫יֹומָם ָו ַל ְילָה‬, and a verb conveying the notion of ‘being engaged in, busy with, paying special attention to’ the ‫ ּתֹורָה‬and ‫מ ְצֹות‬. ִ However, it is only in the postbiblical text from Qumran that we find the root ‫ דרש‬in this connection; here, it replaced its predecessor ‫הגה‬ ‘reflect, meditate, utter, recite (?)’ in the biblical Vorlage.

and c in table 3 above), it is clear that the new collocation ‘to seek God’s law / commandments’ ‫מ ְצוֹות‬/‫ה‬ ִ ‫ּתֹור‬ ָ + ‫ ִל ְדרֹׁש‬at first competed with but eventually replaced the earlier idiom, ‘seek God/God’s (oracular) word’ ′‫ּב( ה‬/‫ֶת‬ ְ ‫ל ְדרֹׁש (א‬  ִ / ֹ ′‫)דבַר־ה‬ ְ ‫)ּב‬ ִ ‫( ִל ְדרׁש‬Gertner 1962: 3–6). We are unable to determine whether the ‫ ָּדרַ ׁש‬expression in Chronicles and Ezra reflects a concept of Scripture that was identical with what was current among the rabbinic sages or, for that matter, among the Qumran sectarians (Schiffman 1975: 54–60). However, the aim of the present discussion is not to ascertain the exact semantic nuance of LBH ‫ ָּדרַ ׁש‬but, rather, to draw attention to the remarkable fact that the very appearance on the biblical scene of ‫ ָּדרַ ׁש‬together with (or, syntactically, governing) terms related to the written pentateuchal law (‫)ּתֹורה‬ ָ and commandments (‫)מ ְצֹות‬ ִ is absent from CBH. This indicates that the LBH application of ‫ ָּדרַ ׁש‬is a linguistic innovation of later generations. In other words, regardless of what the precise semantic nuance and suggested interpretation of the noun ‫ ִמ ְד ָרׁש‬and the verb ‫ ִל ְדרֹׁש‬were in LBH (and QH), the debut of ‫ ִמ ְד ָרׁש‬and the use of ‫ ִל ְדרֹׁש‬in a characteristically late idiom— with a distinctively late meaning—heralded the profound transition in meaning and function that ‫ ָּדרַ ׁש‬underwent in postexilic times. These words may well be regarded as linguistic forerunners of the postbiblical halakhic terminology that was later employed in Tannaitic and Amoraic texts composed in both Hebrew and Aramaic.

274

Avi Hurvitz

3.3.  Objections and Response What are the specific considerations and doubts raised in objection to the diachronic investigation of ‫מ ְד ָרׁש‬, ִ as presented above? Rezetko (2007: 405) writes: The verb ‫ דרש‬occurs 41 times in Chronicles (1⁄4th of all BH occurrences). Consequently it is unsurprising that ‫מדרש‬, a distinctive hallmark of Rabbinic literature, should appear in this book only.

The statistical fact that ‫ ָּדרַ ׁש‬appears 41 times in Chronicles is not much of a help for dating purposes, because it fails to account for the semantic range of the root in the various biblical passages under examination. As noted above, ‫ ָּדרַ ׁש‬had undergone an unmistakable semantic shift in its linguistic history within BH; the crucial question regarding this development is revealed in the use of ‫ ָּדרַ ׁש‬and ‫ ִמ ְד ָרׁש‬in conjunction with the written text of God’s law (‫)ּתֹורה‬ ָ and commandments (‫)מ ְצֹות‬. ִ It has also been emphasized that this innovative aspect of ‫ ָּדרַ ׁש‬found expression in its later manifestations, where the root syntactically governed not God himself or his one-time oracular saying but, rather, his written law and commandments (see above, table 3). It is precisely here, therefore, that the use of ‫ ָּדרַ ׁש‬in Ezra and Chronicles crosses the linguistic borderline between CBH and LBH. Unfortunately, however, these semantic and syntactic aspects of ‫ ָּדרַ ׁש‬find no place in Rezetko’s discussion. Consequently, the seemingly impressive figure of 41 occurrences with which he opens his discussion on the potential lateness of ‫ ִמ ְד ָרׁש‬is irrelevant for purposes of dating. Statistical tallies are significant only when coupled with a detailed, perceptive philological-linguistic analysis, which involves more than counting. This notion was expressed long ago by S. R. Driver (1882: 203), one of the greatest Hebraists in the history of biblical scholarship who, in this connection, criticized scholars who “number words instead of weighing them.” 11 4.  Excursus: The Imperial Aramaic Scribal Formulary of the Persian Period The term ‫ ִא ֶּגרֶת‬is not an isolated example of its kind in LBH. Rather, it should be recognized as “the tip of the iceberg,” symbolizing a whole set of technical terms and expressions that were used extensively in Imperial Aramaic administrative-governmental correspondence of the Achaemenid regime (see Polak 2006). This Aramaic formulary exerted a profound linguistic influence throughout a vast territory of the ancient Near East, extending—in the words of the book of Esther—“from India to Ethiopia over one hundred and twenty-seven provinces” (Esth 1:1). This branch of Aramaic, which served as the standard vehicle for official communication at that time, played a decisive 11. The continuation of Driver’s quotation is also instructive: “and when individual cases are examined, some cause which cannot be tabulated may appear for the presence or absence of a given word in a particular writing” (italics original).

The “Linguistic Dating of Biblical Texts”

275

role in shaping the linguistic profile of contemporary LBH (Kutscher 1982: 72–73, §100; 81, §117). It should be remembered, however, that although a lexical term that belonged to the distinctive Imperial Aramaic formulary may at times have appeared already in Old Aramaic (see the discussion of ‫ ִא ֶּגרֶת‬in §2.4 above), the antiquity of any linguistic element within Aramaic itself (or, for that matter, in any other foreign language) does not reveal when exactly it was imported into or became characteristic of (certain segments of) BH. This consideration likewise applies to loanwords introduced into Imperial Aramaic from Persian. 4.1.  Persian Loanwords In an instructive discussion of “Persianisms,” C. L. Seow (1996: 646) records 22 items that are “all commonly recognized Persian loanwords in the Hebrew Bible.” M. Eskhult’s (2003: 12–14) study of loanwords in BH lists only some 15 lexemes, and recently R. M. Wright (2005: 113–20), in a concise but well-balanced survey on this topic, presents the same 22 loanwords as Seow. The vast majority of scholars in our field note that these loans are a hallmark of the Writings (‫ )כתּובים‬section of the Bible, which continues the latest compositions of biblical literature. Not a single lexeme of undisputable Persian origin is recorded in either the Pentateuch or the Prophets. This striking correlation between the majority of biblical books whose historical setting unequivocally reflects the Persian period (Esther, Daniel, Ezra, Nehemiah, and Chronicles) and their unmistakable linguistic affinities with the Persian lexicon cannot be ignored or simply dismissed as coincidental. This correlation strongly suggests that the borrowing of Persian loanwords into BH precisely in this time period was the result of “languages in contact” brought about by the Persian conquest. In other words, Persian loanwords appear to serve well as chronological markers of lateness in the recorded history of BH. 5.  Concluding Remarks My objective in this study has been to reassess select concrete examples the linguistic lateness of which in BH is widely acknowledged in traditional diachronic research but questioned or denied by some scholars who argue that such research is not based on valid data or viable assumptions. Using these examples, I have revisited the methodological guidelines and philological procedures underlying the historical-comparative analysis employed in this essay in order to determine whether or not the combined evidence of our four suggested criteria suffice for generating valid diachronic statements:  12 12.  Once it has been established that the collocation of ‫ ָּדרַ ׁש‬with ‫ּתֹורה‬ ָ or ‫ ִמ ְצֹות‬is a characteristically late idiom, it may be argued, in turn, that, as a chronological marker, its presence in biblical texts of debatable diachronic provenance may indicate the late composition of those texts (if it fulfills our fourth criterion of ‘Accumulation’ [see below, §5 (4)]. Consider the case of Psalm 119:

276

Avi Hurvitz

1.  Linguistic Distribution (above, §2.1) guarantees that the appearance of the element suspected of being a late feature is indeed confined—exclusively or predominantly—to late biblical writings. Obviously, late occurrences of any given feature do not, in and of themselves, prove that we are dealing with a post-Classical linguistic innovation, but they do serve as a philological prerequisite that justifies an attempt to question the antiquity of the feature under examination. 2.  Linguistic Contrast (above, §2.2) establishes the existence of CBH alternatives available to early writers prior to the penetration of a suspected late feature. Thus, the argumentum ex absentia (from CBH) cannot be faulted for being ex silentio. 3.  Extrabiblical Sources (above, §2.3) refers to evaluating relevant biblical data in their broader synchronic and diachronic linguistic contexts. These nonbiblical sources serve as “control data” for the internal BH diachronic analysis so as to ensure that a suspected late feature is indeed broadly characteristic of the late linguistic milieu in general. 4.  Finally, accumulation is applied whenever an attempt is made to classify chronologically biblical texts of unknown or disputed date. This criterion demonstrates that we are not dealing with a text that includes one or two isolated cases of nonstandard CBH only by chance but, rather, with a text, the language of which as a whole is marked unambiguously as late owing to an assortment of features characteristic of the postexilic period. In other words, the non-Classical elements in the text can indeed be classified as post-Classical.

It goes without saying that these four criteria do not constitute a magic formula by means of which every single chronological difficulty in BH can be resolved. The scope of both the biblical corpus and the comparative material from extrabiblical sources is limited and hence limiting. Nevertheless, a rigorous application of these four criteria in tandem provide researchers with a dependable tool with which it is possible to resolve many chronological issues and thereby to generate reliable data useful for writing a history of Biblical Hebrew. Ps 119:155 [

the wicked. . .

they do not

seek thy

statutes

(‫דרשו‬

‫)חקיך  לא‬

____ Lord

(‫דרשו‬

‫ לא‬′‫____ ה‬

versus

‫]ײ‬

[Isa 9:12(13) The people

did not . . . seek the

Ps 119:45

. . .

I have

sought thy precepts

(‫דרשתי‬

‫פקדיך‬

)

Ps 119:94

save me, for

I have

sought thy precepts

(‫דרשתי‬

‫פקודיך‬

 )

___ the Lord (‫דרשתי‬

‫____אדני‬

____ )]

‫])ואת‬

[ versus ] [Ps 77:3(2)

In . . . my trouble I

seek

Here, it should be noted that ‫ּקּודים‬ ִ ‫ ָּדרַ ׁש ִּפ‬is equivalent to ‫ּדרַ ׁש ִמ ְצֹות‬, ָ because the root ‫ פקד‬is the Aramaic equivalent of the root ‫( צוה‬see, for instance, the renderings of Hebrew ‫ ִצּוָה‬in Tg. Onqelos). On the linguistic lateness of Psalm 119 in general, see Hurvitz 1972: 131–52.

The “Linguistic Dating of Biblical Texts” References

277

Bendavid, A. 1967 Biblical Hebrew and Mishnaic Hebrew, vol. 1. Tel-Aviv: Dvir. [Hebrew] Bergey, R. L. 1983 The Book of Esther—Its Place in the Linguistic Milieu of Post-exilic Biblical Hebrew Prose: A Study in Late Biblical Hebrew. Ph.D. dissertation. Dropsie College. Cooke, G. A. 1903 A Text-Book of North-Semitic Inscriptions. Oxford: Clarendon. Donner, H., and Röllig, W. 2002 Kanaanäische und aramäische Inschriften. 5th ed. Wiesbaden: Harrassowitz. Driver, G. R. 1957 Aramaic Documents of the Fifth Century b.c. Abridged and rev. ed. Oxford: Clarendon. Driver, S. R. 1882 On Some Alleged Linguistic Affinities of the Elohist. Journal of Philology 11: 201–36. Eskhult, Mats 2003 The Importance of Loanwords for Dating Biblical Hebrew Texts. Pp. 8–23 in Biblical Hebrew: Studies in Chronology and Typology, ed. I. Young. London: T. & T. Clark. Folmer, M. L. 1995 The Aramaic Language in the Achaemenid Period: A Study in Linguistic Variation. Leuven: Peeters. Gertner, M. 1962 Terms of Scriptural Interpretation: A Study in Hebrew Semantics. BSOAS 25: 1–27. Hurvitz, A. 1972 The Transition Period in Biblical Hebrew: A Study in Post-Exilic Hebrew and Its Implications for the Dating of Psalms. Jerusalem: Bialik Institute. [Hebrew] 1995 Continuity and Innovation in Biblical Hebrew: The Case of ‘Semantic Change’ in Post-Exilic Writings. Pp. 1–10 in Studies in Ancient Hebrew Semantics, ed. T. Muraoka. Abr-Nahrain Supplement 4. Louvain: Peeters. 1997 The Historical Quest for ‘Ancient Israel’ and the Linguistic Evidence of the Hebrew Bible: Some Methodological Observations VT 47: 301–15. 1999 The Relevance of Biblical Hebrew Linguistics for the Historical Study of Ancient Israel. Pp. 21*–33* in Proceedings of the Twelfth World Congress of Jewish Studies, Division A: The Bible and Its World. Jerusalem: World Union of Jewish Studies. 2000 Can Biblical Texts Be Dated Linguistically? Chronological Perspectives in the Historical Study of Biblical Hebrew. Pp. 143–60 in Congress Volume: Oslo 1998, ed. A. Lemaire and M. Sæbø. VTSup 80. Leiden: Brill. 2003 Hebrew and Aramaic in the Biblical Period: The Problem of ‘Aramaisms’ in Linguistic Research on the Hebrew Bible. Pp. 24–37 in Biblical Hebrew: Studies in Chronology and Typology, ed. I. Young. London: T. & T. Clark.

278

Avi Hurvitz

2006 The Recent Debate on Late Biblical Hebrew: Solid Data, Experts’ Opinions, and Inconclusive Arguments. HS 47: 191–210. Joosten, J. 2005 The Distinction between Classical and Late Biblical Hebrew as Reflected in Syntax. HS 46: 327–39. Kaufman, S. A. 1974 The Akkadian Influences on Aramaic. Chicago: University of Chicago Press. Kutscher, E. Y. 1954 New Aramaic Texts. JAOS 74: 233–48. 1982 A History of the Hebrew Language, ed. R.  Kutscher. Jerusalem: Magnes  / Leiden: Brill. Lauterbach, J. Z. 1935 Mekhilta de-Rabbi Ishmael: A Critical Edition, vol. 3. Philadelphia: Jewish Publication Society. Mankowski, P. V. 2000 Akkadian Loanwords in Biblical Hebrew. HSS 47. Winona Lake, IN: Eisenbrauns. Olmo Lete, G. del; and Sanmartín, J. 2004 A Dictionary of the Ugaritic Language in the Alphabetic Tradition, trans. W. G. E. Watson. 2 vols. 2nd rev. ed. Leiden: Brill. Polak, F. 2006 Sociolinguistics: A Key to the Typology and the Social Background of Biblical Hebrew. HS 47: 115–62. Polzin, R. 1976 Late Biblical Hebrew: Toward an Historical Typology of Biblical Hebrew Prose. HSM 12. Missoula, MT: Scholars Press. Qimron, E. 1986 The Hebrew of the Dead Sea Scrolls. HSS 29. Atlanta: Scholars Press. Qimron, E., and Strugnell, J. 1994 Qumran Cave 4: V : Miqṣat Maʿaśe ha-Torah. DJD 10. Oxford: Clarendon. Rabin, C. 1962 Foreign Words. Pp. 1070–80 in vol. 4 of Encyclopaedia Biblica. Jerusalem: Bialik Institute. [Hebrew] Rezetko, R. 2007 ‘Late’ Common Nouns in the Book of Chronicles. Pp. 379–417 in Reflection and Refraction: Studies in Biblical Historiography in Honour of A. Graeme Auld, ed. R. Rezetko, T. H. Lim, and W. B. Aucker. VTSup 113. Leiden: Brill. Sáenz-Badillos, A.l 1993 A History of the Hebrew Language, trans. J. Elwolde. Cambridge: Cambridge University Press. Schiffman, L. H. 1975 The Halakhah at Qumran. Leiden: Brill. Seow, C. L. 1996 Linguistic Evidence and the Dating of Qohelet. JBL 115: 643–66.

The “Linguistic Dating of Biblical Texts”

279

Wright, R. M. 2005 Linguistic Evidence for the Pre-Exilic Date of the Yahwistic Source. London: T. & T. Clark. Yadin, Yigael 1961 The Expedition to the Judean Desert, 1960: Expedition D. IEJ 11: 36–52. Young, I.; Rezetko, R.; and Ehrensvärd, M. 2008 Linguistic Dating of Biblical Texts. 2 vols. London: Equinox.

The Evolution of Literary Hebrew in Biblical Times The Evidence of Pseudoclassicisms Jan Joosten 1. Introduction Historical research on Biblical Hebrew operates, as does most research in the humanities (and perhaps in the natural sciences as well), in the framework of a “grand narrative.” One might call it a paradigm or, more simply, a working hypothesis: a set of suppositions permitting scholars to make sense of a great variety of disparate data. For almost 200 years, reckoning from the publication of Gesenius’ Geschichte der hebräischen Sprache und Schrift in 1815, the grand narrative of knowledgeable Hebraists has been as follows: •  Most of the Hebrew Bible—including the bulk of the Pentateuch and Former Prophets—was written in the monarchic era (between the tenth and the sixth centuries b.c.e.) in the official literary language practiced at the court and in the temple of the state of Judah. •  A sizable corpus, however, containing notably the books of Ezra–Nehemiah, Chronicles, Esther, Daniel, and Qoheleth came into being only after the Babylonian Exile. •  The Classical Hebrew (CBH) of the older, preexilic books differs from the Late Biblical Hebrew (LBH) of the postexilic books, with the differences being due mostly to diachronic development. •  A corollary of this theory is that biblical texts may in principle be dated, approximately, on the basis of their language. For instance, although tradition attributes Qoheleth and Daniel to the tenth and the sixth centuries, respectively, the language of these books shows that they were written after the exile.

Of course, many important questions are left open in this scheme. Books that are basically preexilic may have been edited, supplemented, or partially rewritten in later times. There may be more varieties of Hebrew than just classical and late: dialectal subtypes, professional jargon, higher and lower stylistic registers. A transitional period may have to be postulated between pre- and 281

282

Jan Joosten

postexilic Hebrew. Poetic texts will be hard to classify, because poets exploit a much wider range of expressions than is usual in prose. Some old poems may reflect the Hebrew language of the premonarchic period. Grand narratives, however, are made to incorporate these sorts of open questions. They provide an initial framework within which scientific observations can meaningfully be debated. Grand narratives do not fall from the clear sky. They are not arbitrary constructs established by mere convention. In the present case, the adoption of the basic scheme of preexilic versus postexilic Hebrew owed much to the authority and obvious competence of Wilhelm Gesenius (Joosten forthcoming). When he published his Geschichte, he was only 29 years old, but by this tender age he had already authored both a dictionary and a grammar of Biblical Hebrew—he could know. His dictionary and grammar contained many observations confirming the approach. The theory also accorded with the general state of biblical studies at the beginning of the nineteenth century. De Wette had recently laid the foundation for a critical approach to the dating of biblical texts: Deuteronomy, introducing the idea of cult centralization, belonged to the time of Josiah, roughly the seventh century b.c.e.; writings familiar with the idea of cult centralization were later than Deuteronomy; while writings unfamiliar with this idea were earlier (De Wette 1805). Gesenius’s scheme received a major boost when Hebrew inscriptions from the monarchic period came to light. The language of these inscriptions turned out to conform closely to the Hebrew of the biblical texts dated to the earlier period. 1 Later generations of Hebraists adopted the approach because in their opinion it continued to function as a valuable reference point for their own research. Dominant narratives are not proven facts, however. They are not observable. They are therefore open to challenge. A recent challenge to the Hebraists’ scheme was developed first by Ernst-Axel Knauf and Philip Davies and later by Ian Young, Robert Rezetko, and Martin Ehrensvärd and some others. Roughly, what these scholars argue is that the language considered to be preexilic in the earlier approach is in reality a dialect, or style-form, that could freely be selected by authors writing in different periods, including the postexilic period. 2 Claims of this sort do more than tweak the related elements of a working hypothesis. They are meant to shatter the grand narrative to pieces. According to a recent study by Young illustrating the new approach, Classical (or Standard, or Early) Biblical Hebrew is found not only in the Pentateuch and Former Prophets but also in Haggai and Zechariah, which belong to the 1.  When the Siloam inscription—the first Judean inscription from biblical times to be discovered—came to light in 1880, scholars were happy to observe that its language was close to preexilic Hebrew. Note the opinion of Driver (1912: x): “The Hebrew is as idiomatic and flowing as a passage from the Old Testament.” Similarly Cooke (1903: 16): “The style is pure and idiomatic, and reads like a good prose passage out of the O.T.” 2.  For the recent debate, see the articles in Young 2003.

The Evolution of Literary Hebrew in Biblical Times

283

early Persian period, and in the Qumran Pesher Habakkuk from the Hellenistic period (Young 2008). The upshot is that linguistic analysis can contribute nothing to the dating of biblical texts. Several biblical scholars have been happy to embrace this new approach, which is congenial to their own impression of the date of biblical literature (Ben Zvi 2010). Hebraists, however, have not found the theory apposite. This could be due to the natural and well-documented tendency of scholars to cling to old ideas, particularly ideas to which they committed themselves in writing. 3 But it is equally possible that the challenge is ill founded. The issue can be resolved only by a renewed analysis of the facts. In this essay, I contribute to this renewed analysis by drawing attention to one particular aspect of Hebrew texts dated by common consent to the Persian period or later. 2.  Matching up Arguments and Evidence Among Hebrew writings for which a postexilic date is not controversial, some texts conform to what one expects in the LBH corpus: they use Persian loanwords, lexical and grammatical Aramaisms, and other elements unattested in CBH; others, however, remain closer to CBH usage. The variation between texts diverging from CBH and texts conforming to it arises at least in part because of stylistic choices made by the authors. This is not a new insight, but it is certainly true. 4 Young bases his demonstration on observations about the language of Pesher Habakkuk, mostly in the area of vocabulary, but also in the area of grammar (Young 2008). Although the method is perhaps skewed somewhat toward numbering—linguistic features ought to be weighed as much as counted—Young develops a fruitful avenue of research. The same method could be applied to other Qumran texts, to Ben Sira, and to the late books of the Hebrew Bible. It would be interesting to know which writings are relatively more similar to CBH and which works use other modes of expression more freely, and to ask why the difference arises. The successful demonstration of this point does very little, however, to counter the idea that linguistic differences can be used to date an ancient Hebrew text. Pesher Habakkuk may be written in a language that strongly re­sembles Classical Biblical Hebrew, but this quality does not impede the diachronic approach. If Pesher Habakkuk were an undated text, impossible to date on the strength of its contents, a Hebraist trained in historical linguistics would have little trouble establishing its relatively late date. 5 The central word of 3.  This insight is often attributed to Thomas Kuhn, but it was already well known to Max Planck: “Eine neue wissenschaftliche Wahrheit pflegt sich nicht in der Weise durchzusetzen, daß ihre Gegner überzeugt werden und sich als belehrt erklären, sondern vielmehr dadurch, daß ihre Gegner allmählich aussterben und daß die heranwachsende Generation von vornherein mit der Wahrheit vertraut geworden ist” (Plank 1948: 2). 4.  See, for example, Joosten 1999 and the literature gathered there. 5.  When medieval manuscripts of the Damascus Document were discovered in the Cairo Geniza, at the end of the nineteenth century, some scholars proposed dating the writing to

284

Jan Joosten

Pesher Habakkuk, the Hebrew word ‫ פשר‬itself, is a late loanword that reached Hebrew through Aramaic. Avi Hurvitz has argued that four criteria need to be observed in order to argue for the lateness of a given text (Hurvitz 1972): 1.  Distribution: elements of the text to be dated should align with late books exclusively, being unattested in the CBH corpus. 2.  Extrabiblical confirmation: extrabiblical parallels in Hebrew and other languages, especially Aramaic, can strengthen the diachronic validity of certain features. 3.  Opposition: the lateness of a given element can be confirmed if an exact equivalent can be identified in the CBH corpus. 4.  Accumulation: a late text should manifest an accumulation of late linguistic features.

The word ‫ פשר‬answers to all these criteria: 6 1.  In Biblical Hebrew, it is attested only in Qoh 8:1. 2.  It is also found 34 times in the Aramaic of Daniel. 7 3.  In the CBH corpus, other words are used to express the meaning ‘interpretation’. 8 4.  The word ‫ פשר‬occurs dozens of times throughout 1QpHab.

This feature alone (use of the word ‫ )פשר‬would suffice to indicate that the text is postexilic. In addition, the pesher contains the late Persian loanword ‫‘ רז‬mystery’, and the classical word ‫ קץ‬is used in the late sense of ‘period’. 9 Case closed. It all comes down to what one wants to demonstrate. Due to an interest in language typology, one may undertake to show that “Late Biblical Hebrew” is not a homogenous state of language. There are important linguistic differences among the various books that compose the LBH corpus. 10 To describe the the Middle Ages. Solomon Schechter, the editor of the text, dated it to the Second Temple period, invoking among other things the evidence of its language: The language of the ms is for the most part pure Biblical Hebrew. The first three pages rise even to the dignity of Scriptural poetry, though a good deal of it is obscured by the unfortunate condition in which the text is at present. But there are in it terms and expressions which occur only in the Mishna or even only in the Rabbinic literature dating from the first centuries of the Middle Ages. (Schechter 1910: xi)  6.  The fourth criterion is normally taken to refer to the accumulation of different features, not the repeated attestation of one word. The word ‫ פשר‬occurs so often in 1QpHab, however, and is so central to the argument that it can hardly be explained as a gloss or a word that entered the writing in the course of later revision. 7.  The word is of Akkadian origin but almost certainly reached Hebrew through the medium of Aramaic. 8.  For the CBH equivalents of ‫פשר‬, see Gen 40:12, 18; and Judg 7:15. 9. Note 1QpHab 7:2: ‫‘ גמר הקץ‬the end of the period’. 10.  This is not something that has escaped scholarly notice: in a plenary debate on the diachronic approach of Biblical Hebrew at the 14th World Congress of Jewish Studies in

The Evolution of Literary Hebrew in Biblical Times

285

diversity among the late biblical books more precisely is worthwhile. It may lead to interesting new results. If, however, the aim is to date undated texts, other considerations come into play. Mainstream Hebraists are not in the least perturbed by the use of CBH features in Pesher Habakkuk or other late writings. Authors of the Persian and Hellenistic periods knew much of the classical literature intimately and wished to situate their own writing in continuity with it. They made a conscious effort to model their literary production on the earlier texts (Greenfield and Naveh 1984: 120–22). Therefore, their use of what appears to be classical Hebrew does not affect the question of date. For dating purposes, only their use of post­ classical elements, such as ‫ פשר‬or ‫רז‬, is significant. The viability of the approach hinges in this case on the notion of imitation, or artificial continuation. This is indeed an important postulate of Hebrew studies: after the exile, when some of the writings that would later become part of the Bible had acquired a measure of authority, authors wrote religious literature in a style that intended to remain close to the earlier texts. 11 In the remainder of this essay, I point to evidence showing that this postulate is justified. Later authors did indeed imitate Classical Hebrew, which they knew from studying old texts. The result, however, was not always felicitous. The later authors did not fully master the classical language but “tripped up” in small details. Even the late books with the most instances of classicizing language are not classical but pseudoclassical. 3.  Pseudoclassicism in Pesher Habakkuk: Confusion of Literary Registers The Pesher Habakkuk selected by Young is as good a place to begin as any. Many words and expressions that appear in the commentary part of this writing are biblical but unattested in the core corpus of Late Biblical Hebrew. But the use of classical elements does not mean the Hebrew itself is classical. Even where the vocabulary is the same, the usage is subtly different. One of the words classified as “Early Biblical Hebrew” by Young is the noun ‫‘ עריץ‬violent (?)’ in 1QpHab 2:6. The word appears 20 times in the biblical corpus but never in LBH. The mere use of the word, however, does not imply that the pesher is written in “Early” BH. In the Bible, ‫ עריץ‬is a poetic word that is found in Psalms, Job, Proverbs, Isaiah, Jeremiah, and Ezekiel, exclusively in poetic passages: 12 2005, David Talshir challenged the scholars in the room to name one feature that is found in all the core LBH books but in no classical text; no one took up the challenge. 11.  “Post-exilic texts seem to reflect the more or less artificial employment of the preexilic ‘classical’ literary language, which no longer existed as a spoken medium” (Elwolde 2006: 136). 12.  The existence of poetic words in Biblical Hebrew was already well known to Gesenius (1815: 22). Living languages also tend to use certain words in poetry only: in French,

286

Jan Joosten

​ (1) 

‫ׁש ִעים עֲֹונָם‬ ָ ‫ַל־ר‬ ְ ‫ּופָקַ ְד ִּתי עַל־ּתֵ בֵל ָרעָה ְוע‬ ‫ַׁש ִפיל‬ ְ ‫יצים א‬ ִ ‫ָר‬ ִ‫אוַת ע‬ ֲ ‫ׁשּב ִַּתי ּגְאֹון ז ִֵדים ְו ַג‬ ְ ‫ְו ִה‬ I will punish the world for its evil, / and the wicked for their iniquity; I will put an end to the pride of the arrogant, / and lay low the insolence of tyrants. (Isa 13:11)

In Pesher Habakkuk, however, the usage is not poetic: ‫על הבו]גדים לאחרית א‬ [ ‫וכן פשר הדב‬ ‫ המה עריצ˙[י הבר]ית אשר לוא יאמינוא‬.‫הימים‬ ˙ ‫בשומעם את כול הב ˙א[ות‬ ‫ע]ל[ ]הדור האחרון מפי‬ The interpretation of the word [concerns the trai]tors in the last days. They are the violator[s of the coven]ant who will not believe when they hear all that is going [to happen t]o the last generation. (1QpHab 2:5–7)

(2)  ​

The use of a poetic word in a prosaic context seems to reflect artificial reuse. The Qumran writers found ‫ עריץ‬in Scripture and decided to use it in their own writings but failed to pay attention to the word’s literary register. If the use of ‫ עריץ‬were an isolated case, it could be brushed off as an accident. But there are several other examples of words and forms that appear only in poetry in the biblical texts but turn up in prose in the pesher. The noun ‫הון‬ ‘abundance, wealth’ is found 26 times in the Bible, always in poetical contexts, with a strong concentration in wisdom literature. In the pesher, as in Qumran Hebrew generally, ‫ הון‬is used in prosaic discourse and means simply ‘money’. 13 Other poetic features of Biblical Hebrew that are recycled in the pesher’s prose are the verb ‫‘ פעל‬to do’ (1QpHab 8:13; 12:8) and the “archaic” form ‫‘ למו‬for them’ (1QpHab 5:6). The confusion of levels of style typifies language use that is non-native. While it is relatively easy to study the vocabulary and grammar of a foreign language, it is much harder to develop a feel for register. A foreigner may write nearly perfect English, yet the stilted or pedantic quality of his or her style may give him or her away. Similarly, the use of poetic words in nonpoetic contexts shows that Biblical Hebrew was not a native idiom of the Qumran authors but an acquired language. The slight misuse of CBH elements in Pesher Habakkuk indicates, as surely as do the postclassical elements, the relatively late date of the composition.

words such as aquilon for the north wind, couche in the meaning ‘bed’, and épancher ‘to pour’ are used only in poetry or in high-flown prose. 13.  The word alternates with nonbiblical ‫ ממון‬in a way that suggests that the latter was the normal word for ‘money’ while ‫ הון‬was a more prestigious synonym (see, for example, 1QS 6:2 and 19).

The Evolution of Literary Hebrew in Biblical Times

287

4.  Semantic Pseudoclassicisms in Qumran Hebrew The pseudoclassical nature of Qumran Hebrew can also be demonstrated from semantics. Some CBH words and expressions are used in later writings with a meaning that reflects the interpretation of old texts, not the natural continuity of the language. An interesting example, not attested in Pesher Habakkuk but in the Damascus Document, is the verb ‫( אות‬Joosten 2003). Close study of the contexts where it occurs, in Gen 34:15, 22, 23; and 2 Kgs 12:9, indicates that the verb means ‘to consent’ in the Bible. 14 In the Damascus Document, however, the verb should be translated ‘to associate with, to pool resources’: ‫אל {{ית}} יאות איש עמו בהון ובעבודה‬ (3)  ​ No one should associate with (the one who is slack in the fulfillment of the instructions of the upright) in wealth or work. (CD 20:7) 15 The divergence between biblical and Qumranic usage most likely indicates that the original meaning of the verb, ‘to consent’, was forgotten in later times. The new meaning, ‘to associate with’, arose from the interpretation of Genesis 34, a chapter where the notion of sharing economic resources plays an important part. The interpretation underlying the Qumranic use of the verb ‫אות‬ also surfaces in the targums on the Pentateuch. In Tg. Neofiti and in a Geniza fragment, the verb is translated ‫אתערב‬, lit., ‘to intermingle’ and thence ‘to share, to associate with’. This same verb is used in the Rule of the Community (1QS) in a close parallel to the passage from the Damascus Document (CD) quoted above: ‫ולוא יתערב איש מאנשי הקודש בהונו ועם עצתו לכול דבר‬ (4)  ​ None of the men of holiness should associate with his goods (that is, the goods of the one who breaks a word of the law of Moses) or his advice on any matter. (1QS 8:23) 16 The Qumran authors borrowed the verb ‫ אות‬from Scripture, but they used it in a sense that was arrived at secondarily on the basis of interpretation. At first sight, the use of the verb in CD 20:7 may create the impression that the text is written in pure Classical Hebrew. But closer analysis indicates that this is not so. The language is not classical but pseudoclassical. Several other pseudoclassicisms have been uncovered in the course of the linguistic analysis of the scrolls (Joosten 1999; Kister 2009):

14.  A different reading of the evidence is proposed by Kottsieper 2010; see, however, Kister 2009: 568. 15.  The verb also appears in a fragment of the Damascus Document discovered at Qumran: 4Q266 frg. 1 1:15. 16.  See also 1QS 9:8.

288

Jan Joosten •  ‫‘ שחת‬pit’ is used with the meaning ‘perdition’ based on the root ‫‘ שחת‬to destroy’ (Wernberg-Møller 1957: 81); •  ‫‘ מעוז‬refuge’ (from ‫ ) עוז‬is used with the meaning ‘strength’ on the root ‫עזז‬ ‘to be strong’ (for example, 1QHa 16:23–24); •  ‫‘ אפעה‬serpent’ is used with the meaning ‘wickedness’, from contextual exegesis (Joosten 1999: 150–51).

All these “biblical” words are used in the scrolls with new meanings based on etymological or contextual exegesis. The phenomenon shows that the Qumran writings were composed in a period far removed from the original setting of the biblical books, the language of which they borrowed. 5.  Pseudoclassicisms in Late Biblical Hebrew The reuse of archaic expressions of the type illustrated is characteristic of Qumran Hebrew, but it is not exclusive to this literature. Instances of the phenomenon are found sporadically in the late biblical books as well. An interesting example is the use of the idiomatic expression ‘to fill one’s hand’ (‫ מלא‬Piel + ‫ את יד‬+ X). In the Pentateuch and Former Prophets, this expression means ‘to ordain to a sacred office’ or something very similar: ‫ָפים‬ ִ ‫ּותר‬ ְ ‫ָאיׁש ִמיכָה לֹו ּבֵית אֱל ִֹהים ַוּיַעַׂש אֵפֹוד‬ ִ ‫ְוה‬ ‫ְהי־לֹו ְלכֹהֵן‬ ִ ‫ַו ְי ַמּלֵא אֶת־י ַד ַאחַד ִמ ָּבנָיו ַוי‬ This man Micah had a shrine, and he made an ephod and teraphim, and installed one of his sons, who became his priest. (Judg 17:5) 17

(5)  ​

In Ezek 43:26, the expression refers to the consecration of an altar. The etymology of the meaning ‘to ordain’ is obscure. In addition, the expression is attested three times in Chronicles. Modern translations and dictionaries tend to treat the LBH passages the same way as the others. Attention to the context shows, however, that this is problematic (Paran 1989: 283–84; Joosten 1999:152–53). Two of the Chronicles passages have nothing whatsoever to do with consecration or induction to a priestly office. In 1 Chr 29:5, David appeals to the people to give gold, silver, and precious stones for the building of the temple. After having enumerated all he is willing to donate himself, he goes on to say: ‫ּומי ִמ ְתנַּדֵ ב ְלמַּלֹאות יָדֹו הַּיֹום לַיהוָה‬ ִ (6)  ​ Who, then, will offer willingly, filling their hands today to the Lord? (1 Chr 29:5b) The nrsv has rendered this ‘Who then will offer willingly, consecrating themselves . . .’, but there is no contextual warrant for this translation. The Israelites who contribute to the building of the temple are not priests, nor are they trying 17.  See also Exod 28:41; 29:9, 29, 33, 35; Lev 8:33; 16:32; 21:10; Num 3:3; Judg 17:12; 1 Kgs 13:33.

The Evolution of Literary Hebrew in Biblical Times

289

to become priests. The passage is concerned with donations. The expression ‘to fill their hand’ is a figure of speech implying generosity. One might render the half verse: ‘Who then will offer willingly, filling his hands today with gifts for the Lord?’ Similarly, the subject matter in 2 Chr 29:31 is giving, not ordination. In the third passage, the notion of ordination does come up, but it is doubtful whether the idiom ‘to fill the hand’ expresses it: 18 ‫ֶת־ּבנֵי ַאהֲרֹן ְוה ְַלִוּיִם וַּתַ עֲׂשּו ָלכֶם‬ ְ ‫הנֵי יְהוָה א‬ ֲֹ ‫הֲלֹא ִהּדַ ְח ֶּתם אֶת־ּכ‬ ‫ֵילם‬ ִ ‫ָקר ְוא‬ ָ ‫הנִים ְּכ ַעּמֵי ָהאֲרָצֹות ּכָל־ ַהּבָא ְל ַמּלֵא יָדֹו ְּבפַר ּבֶן־ּב‬ ֲֹ ‫ּכ‬ ‫ׁש ְבעָה ְו ָהיָה כֹהֵן ְללֹא אֱל ִֹהים‬ ִ Have you not driven out the priests of the Lord, the descendants of Aaron, and the Levites, and made priests for yourselves like the peoples of other lands? Whoever comes, filling his hand with a young bull or seven rams, will be a priest of what are no gods. (2 Chr 13:9)

(7)  ​

As the reference to the ‘young bull or seven rams’ shows, the ‘filling of the hand’ concerns the gift that an apostate Israelite brings in order to be inducted into the priesthood. Those who ‘fill their hand’ with a big enough gift are ordained as priests. The gift is a prerequisite of the ordination, not the ordination itself. None of the passages in the Pentateuch or Former Prophets mentions this sort of gift, and none seems to be implied there. We may conclude that the meaning of the expression ‘to fill the hand’ is different in Classical BH and Late BH. 19 While in CBH it means ‘to ordain to a priestly office’, in LBH it means ‘to bring an offering’. The latter meaning probably reflects interpretation on the basis of the component parts. By the time of the Chronicler, the old idiomatic expression had fallen into disuse and its global meaning had been forgotten, at least by some readers of the Hebrew Scriptures. 20 18.  2 Chr 13:9 is loosely parallel to 1 Kgs 13:33, where the idiomatic expression is used in the classical way. It appears, however, that the Chronicler has here freely rewritten his source text, thus enabling him to use the expression with the pseudoclassical meaning. 19.  Young and Rezetko (2008: 106–9) argue, on the basis of secondary literature, that the use in Chronicles is the same as in the CBH corpus. Attentive study of the Hebrew passages shows that this is not so. 20. Similar reanalysis of old idiomatic expressions occurs frequently in our own time. In French, the expression faire long feu (literally ‘to make long fire’) derives from the practice of lighting a canon: if the fire was “long,” the shot didn’t go off. Originally the expression meant ‘to fail’. In modern times, however, the practice of lighting canons has fallen out of use, and the idea underlying the expression is no longer understood. Faire long feu is often used in the meaning ‘to succeed over a long period of time’, usually with a negation: ce projet ne fera pas long feu (‘that project will fail quickly’). In English, expressions such as “to wet (instead of ‘to whet’) one’s appetite” and “the point is mute (instead of ‘moot’)” also show that old idiomatic expressions are poorly understood (although I cannot testify to their being used with a different meaning).

290

Jan Joosten

The claim that the Chronicler’s use of the expression rests on exegesis can be strengthened by a comparison with the targums. In the targums, beginning with Tg. Onqelos to the Pentateuch, the expression ‘to fill the hand’ is systematically rendered ‘to bring an offering’ (‫)קרב קורבנא‬. The end of the verse from Judg 17:5 quoted above is translated in Tg. Jonathan: ‫וקריב ית קרבן חד מבנוהי והוה ליה לכהין‬ (8)  ​ He brought the offering of one of his sons, who became his priest. While the Hebrew text makes no reference to an offering, the targum introduces this notion on the basis of the idiomatic expression. The targum indicates that the interpretation known to the Chronicler enjoyed wider currency among learned Jews during the Second Temple period. The Chronicler not only knew this interpretation but went so far as to revive the ancient expression in his own writing—with the new meaning. Pseudoclassicisms are less frequent in the LBH corpus than in the Qumran scrolls, but they are not rare. Several other examples have been identified in Chronicles, Nehemiah, and Daniel (Joosten 1999). They show that later authors studied earlier texts diligently and tried to link up with the earlier language and style in their own writing. They also show that, by the time the late biblical books were composed, CBH had already become a “dead language” to a certain extent. 6. Conclusion In a general essay on the dating of biblical literature by linguistic means, Ron Hendel envisions the possibility of a work written in flawless CBH during the Persian period (Hendel 2005: 110). Through perfect imitation, a late author would be able to create a work, the postexilic date of which would be impossible to detect. Although this eventuality cannot be excluded, the study of pseudoclassicisms shows how improbable it is. Even authors who strongly manifest their intention to stick to classical models occasionally go astray. This is not just a matter of including one or two modern expressions, such as the word ‫פשר‬. Even when words and forms were taken straight from earlier texts, the way they were used was not always in accord with classical usage. The ostensible linguistic similarity between CBH and texts of the Second Temple period such as Chronicles or Pesher Habakkuk is deceptive. Although the words are the same, the subtle interplay between them is not. In some instances, the differences may be due to natural evolution of the language. In other instances, some of which have been discussed here, the changes do not reflect organic development but artificial recycling. The presence of pseudoclassicisms in the Hebrew literature of the Second Temple period strongly favors the chronological approach developed by Wilhelm Gesenius and continued in our times by Avi Hurvitz and others.

The Evolution of Literary Hebrew in Biblical Times

291

* * * Recent revisionist proposals would take Hebrew studies back to premodern times, when Biblical Hebrew was regarded as a synchronic—or pan-chronic— whole, unyielding to historical analysis. This is not a paradigm shift most Hebraists would care to embrace. A more serious problem is the growing distance between Hebrew linguistics and biblical studies. The diachronic markers established with much acumen and sophistication by Hebraists play little or no part in the equally subtle game of dating biblical texts as practiced by exegetes. Attacks on the documentary hypothesis in the late 1970s have inaugurated a period in which it has become commonplace to date substantial parts of the Pentateuch to the Persian period or later. To the mainstream Hebrew linguist, these proposals are difficult, or even absurd. The language of, say, Genesis 15 is different from that of Ezra–Nehemiah in a way that makes it almost impossible to imagine that the two texts come from the same general period. On this point, a great deal of work remains to be done. Linguists should sharpen their criteria, and exegetes should reflect more on their methodology. But mostly what is called for is more dialogue between the two groups. References

Ben-Zvi, E. 2010 Reconstructing the Intellectual Discourse of Ancient Yehud. Studies in Religion / Sciences Religieuses 39: 7–23. Cooke, G. A. 1903 A Text-Book of North-Semitic Inscriptions: Moabite, Hebrew, Phoenician, Aramaic, Nabataean, Palmyrene, Jewish. Oxford: Clarendon. Driver, S. R. 1912 Notes on the Hebrew Text and the Topography of the Books of Samuel. Oxford: Oxford University Press. De Wette, W. L. M. 1805 Dissertatio critico-exegetica, qua Deuteronomium a prioribus Pentateuchi libris diversum, alius cuiusdam recentiori auctoris opus esse monstratur. Jena. Elwolde, J. 2006 Language and Translation of the Old Testament. Pp. 135–58 in The Oxford Handbook of Biblical Studies, ed. J.  W. Rogerson and J.  M. Lieu. Oxford: Oxford University Press. Gesenius, W. 1815 Geschichte der hebräischen Sprache und Schrift. Leipzig: Vogel. Greenfield, J. C., and Naveh, J. 1984 Hebrew and Aramaic in the Persian Period. Pp. 115–29 in The Cambridge History of Judaism, ed. W. D. Davies and L. Finkelstein. vol. 1. Cambridge: Cambridge University Press.

292

Jan Joosten

Hendel, R. 2005 Remembering Abraham: Culture, Memory, and History in the Hebrew Bible. Oxford: Oxford University Press. Hurvitz, A. 1972 Ben lashon le-lashon: Le-toledot leshon ha-miqra bimey bayit sheni (The Transition Period in Biblical Hebrew: A Study in Postexilic Hebrew and Its Implications for the Dating of the Psalms). Jerusalem: Bialik Institute. Joosten, J. 1999 Pseudo-classicisms in Late Biblical Hebrew, in Ben Sira, and in Qumran Hebrew. Pp. 146–59 in Sirach, Scrolls and Sages: Proceedings of a Second International Symposium on the Hebrew of the Dead Sea Scrolls, Ben Sira, and the Mishnah, Held at Leiden University, 15–17 December 1997, ed. T. Muraoka and J. F. Elwolde. Leiden: Brill. 2003 Minuah kitati u-pharshanut ha-miqra: Horaʾat ha-poʿal ‘ʾwt’ bekitve Qumran. Meghillot 1: 219–26. forthcoming  Wilhelm Gesenius and the History of Hebrew in the Biblical Period. In the Proceedings of the Colloquium on Wilhelm Gesenius Held in Halle in 2010, ed. S. Schorch. Kister, M. 2009 Some Lexical Features of the Writings from Qumran. Pp.  561–69 in The Qumran Scrolls and Their World, ed. M. Kister. Vol. 2. Jerusalem: Yad BenZvi. [Hebrew] Kottsieper, I. 2010 ‫ אות‬und ‫ערב‬. Ein Beitrag zum hebräischen Lexikon und zum Verhältnis von CD zu 1QS. RevQ 95: 405–19. Paran, M. 1989 Forms of the Priestly Style in the Pentateuch. Jerusalem: Magnes. Planck, M. 1948 Wissenschaftliche Selbstbiographie: Mit einem Bildnis und der von Max von Laue gehaltenen Traueransprache. Leipzig: Barth. http://en.wikiquote.org/ wiki/Max_Planck. Schechter, S. 1910 Fragments of a Zadokite Work. Documents of Jewish Sectaries 2. Cambridge: Cambridge University Press. Wernberg-Møller, P. 1957 The Manual of Discipline. Leiden: Brill. Young, I. 2008 Late Biblical Hebrew and the Qumran Pesher Habakkuk. The Journal of Hebrew Scriptures 8, Article 25 (http://ejournals.library.ualberta.ca/index.php/ jhs/article/view/6223). Young, I., ed. 2003 Biblical Hebrew: Studies in Chronology and Typology. JSOTSup 369. London: T. & T. Clark. Young, I., and Rezetko, R., with the assistance of Martin Ehrensvärd 2008 Linguistic Dating of Biblical Texts: An Introduction to Approaches and Problems. 2 vols. London: Equinox.

Signs of Late Biblical Hebrew in Isaiah 40–66 Shalom M. Paul 1. Introduction 1.1.  It is generally recognized that the late sixth century b.c.e. marks a watershed in the history of Biblical Hebrew. This diachronic development, the principal telltale markers of which are reflected in the lexicon, syntax, and grammar, is anchored primarily in the biblical books composed in the Persian period. However, in the course of writing my commentary on Isaiah 40–66 (Paul 2012), prophecies that are clearly dated to the transitional period between the late exilic and early postexilic eras, it became clear to me that the seeds of linguistic change gradually and sporadically had already begun to make their debut. 1 These innovative elements may now be viewed as representative of an intermediate link between what has been termed Standard or Classical Biblical Hebrew, or Early Biblical Hebrew (EBH)—that is, preexilic Hebrew—and Late Biblical Hebrew (LBH)—that is, exilic and postexilic Hebrew—when a merger of old and new components occurred, forming what has been called a “mixed language.” 1.2.  I am very much aware of the vigorous debate between scholars who speak of parallel registers, synchronic variety, different scribal schools, and oral versus written language. I am also cognizant of discussions that consider the implications introduced by stylistic, social, geographical (western and eastern compositions), as well as by diglossia (North Israelian) and dialectal factors, and the criteria set by Avi Hurvitz, who bases his numerous studies 1.  This present essay, which includes only a partial list of the prophet’s neologisms, does not refer to the epigraphical documentation from the First and Second Temple periods or to the comparative Semitic analogues, both of which are of minimal significance for a study of the distinctive characteristics of LBH as compared with EBH. The first examples of each of the linguistic elements drawn from Isaiah 40–66 are followed by additional witnesses, when applicable, drawn from the later books of the Bible, Ben Sira, the Aramaic targums, the Dead Sea Scrolls, and rabbinic literature. The fact that several of these attestations also appear in EBH, though infrequently, does not invalidate the primacy of their overall later documentation (see the contributions of B. E. Dresher, J. Cook, and R. D. Holmstedt in this volume).

293

294

Shalom M. Paul

on four considerations for determining LBH: linguistic distribution pattern, contrastive classical equivalents, extrabiblical attestation (as in the writings from Elephantine, Ben Sira, the Dead Sea Scrolls, and rabbinic literature), and the accumulation of examples. All this is well known to scholars in our field and need not be clarified further here. 2 1.3.  My essay concentrates primarily on the lexicon of Isaiah 40–66 (a complex of prophecies that has hardly been subject to such analysis as of yet), with a few examples drawn from syntax and grammar in order to exemplify what Jan Joosten has aptly called “linguistic stratigraphy” (Joosten 2005: 328). Though these prophecies are still well anchored in Classical Hebrew, the beginnings of what would become common in Late Biblical Hebrew and even later forms of Hebrew have left their imprint in the text. The relevant data are presented below without any attempt to explain how or why these innovations occurred. They are important because their attestation in the prophecies of Second Isaiah provides scholars of the history of Hebrew with a date.  3 2.  Syntactic Features 2.1. The infinitive absolute in lieu of a finite verb appears frequently in LBH. See, for example, Ezek 23:30 (‫(עָׂשה‬, 36 )‫(ו ַהּגֵד‬, ְ 47 )‫ ;)ּובָרֵ א‬Hag 1:6 (‫)ּפָנֹה‬, 9 (‫ ;) ָהבֵא‬compare also Zech 3:4, 7:5, 12:10; Job 15:35; Qoh 4:2 (  ‫)וׁשַ ּב ֵַח‬, ְ 8:9 (‫)ונָתֹון‬, ְ 9:11 (‫;)ו ָראֹה‬ ְ Esth (14×), for example, 9:16–18 (7×); Dan 9:5, 11 (‫;)וסֹור‬ ְ Neh 7:3; 8:8 (‫;)וׂשֹום‬ ְ 9:8, 13; 1 Chr 5:20; 2 Chr 28:19 (‫ ;)ּומָעֹול‬31:10. It also appears often in Deutero-Isaiah’s prophecies, for example: 42:20: . . .  ַ‫ּפָקֹוח‬ (Qere) ‫;ראֹות‬ ָ 59:4: ‫ והולֵיד‬. . . ‫ הָרֹו‬. . . ‫ וּדַ ּבֶר‬. . . ‫טֹוח‬ ַ ‫ ָּב‬. 2.2.  A finite verb with a pronominal suffix instead of a verb followed by a direct object marker and an object. Compare particularly 2 Chr 22:9 ‫ויקברֻהו‬ ׂ ְ ‫ ַוּי‬. with 2 Kgs 9:28 ‫ ;ויקברו אֹתו‬2 Chr 25:28 ‫ִׂשאֻהּו‬ ּ ָ ‫ ַוּי‬with 2 Kgs 14:20 ‫ִשאּו אֹתו‬ The lateness of this feature is very well exemplified in Isaiah 40–66, where it is ְ ‫יֹוע‬, ְ ‫ ;י ִַּציל‬53:3, 4 ‫;חֲׁשַ ְבנֻהּו‬ attested 300 times—for instance, in Isa 57:12 ‫ילּוך‬ ִ 13 ‫ֻך‬ ָ 55:5: ‫ְדעּוך‬ ָ ‫—י‬as opposed to a mere 5 times with the direct object: 41:16 ‫ָּת ִפיץ‬ ‫;אֹותם‬ ָ 42:9 ‫ ;אשמיע אתכם‬50:1 ‫ ;מכרתי אתכם‬65:3 ‫יסים אֹותי‬ ִ ‫ ַה ַמ ְּכ ִע‬, 12 ‫ּומָניתי אתכם‬. 2.3.  Syntactic construction of ‫ היה‬plus a participle. See, for example, Isa 59:2 ‫היו מ ְַב ִּד ִלים‬. Compare in Aramaic, for example, Dan 4:26 ‫הוָה‬ ֲ ‫;מ ַהּל ְֵך‬ ְ 5:19 ‫ֲלין‬ ִ ‫ְעין ְו ָדח‬ ִ ‫ ;הֲוֹו זָי‬6:4 ‫הוָה ִמ ְתנַּצַח‬ ֲ , 15 ‫ׁשּתַ ּדַ ר‬ ְ ‫הוָה ִמ‬ ֲ ; Ezra 6:6 ‫יקין הֲוֹו‬ ִ ‫ ;רַ ִח‬etc. It is also very frequent in Rabbinic Hebrew, as in Mek. Beshallaḥ 4 ‫ ;היה מהלך‬m. Ber. 2.  See the contributions of M. Eskhult, A. Hurvitz, F. Polak, G. A. Rendsburg, and R. M. Wright in part 1 of Young 2003 (“Studies within the Chronological Framework of Pre-Exilic Standard Biblical Hebrew and Post-Exilic Late Biblical Hebrew”), and P. R. Davies, M. Ehrensvärd, J. A. Naudé, R. Rezetko, D. Talshir, and I. Young in part 2 of Young 2003 (“Challenges to the Chronological Model”), as well as Young, Rezetko, and Ehrensvärd 2008. 3.  Whether or not one should distinguish between a Second and Third Isaiah is irrelevant to this discussion because there is a consensus that the periods concerned are within the time frame mentioned above.

Signs of Late Biblical Hebrew in Isaiah 40–66

295

2:1 ‫ ;היה קורא‬3:5 ‫ ;היה עומד‬4:5, ‫היה רוכב‬, 6 ‫ ;היה יושב‬m. Kil. 5:7 ‫ ;היה עובר‬and in Qumran, for example, CD 6:12 ‫ ;ויהיו מסגירי הדלת‬1QS 1:18 ‫ מברכים‬. . . ‫;יהיו‬ 1QM 8:11–12, 17:15 ‫ מריעים‬. . . ‫יהיו‬. 3. Grammar 3.1.  Increased use of the Piel stem of the verb ‫ הלך‬in LBH, appearing some 25 times, almost always in late books, as opposed to the Qal stem that occurs over 1,000 times throughout the Bible. It appears in Isa 59:9 ‫ ִלנְגֹהֹות‬. . . ‫נְקַ ּוֶה‬ ֲ ‫ ַּב‬but is totally absent from First Temple prophetic books. See also, ‫אפֵלֹות ְנ ַהּל ְֵך‬ for example, Ezek 18:9; Hab 3:11; Prov 6:11, 28; 8:20; Job 30:28; Lam 5:18; Qoh 4:15, 8:10, 11:9; and several late psalms, such as Ps 104:10, 26; 142:4 (but only once in classical prose, 1 Kgs 21:27). Thereafter, it is frequent in Mishnaic Hebrew, for example: ‫ מהלך‬m. ʿErub. 4:1, 7; Yoma 5:1; Ned. 2:1; ‫מהלכין‬ m. ʿErub. 5:8; Naz. 5:5; B. Qam. 3:4; ‫ ְל ַה ֵל ְּך‬m. Roš Haš. 1:9; and in Aramaic, for example, Tg. Onq. Gen 7:18 ‫ַלכָא‬ ּ ְ ‫ּומה‬ ְ (Heb. ‫ ;)ותלך‬and once at Qumran, 1QM 7:12 ‫יהיה מהלך‬. 3.2.  Compare likewise the Piel of ‫פחד‬, which appears only twice in the Bible (Isa 51:13 ‫ ;ו ְַּת ַפחֵד‬Prov 28:14 ‫)מ ַפחֵד‬ ְ as opposed to 22 times in the Qal. 4.  Lexical Features 4.1. Aramaic 4.1.1.  ‫‘ ְּכ ֶאחָד‬together’ (Isa 65:25) is a calque of ‫ ַּכח ַָדא‬, the Aramaic translation of Heb. ‫יחדו‬. Compare especially Isa 11:6 ‫יחּדו‬ ָ ‫ ופרה ודֹב ִת ְרעֶינָה‬with 65:25 ‫זאב וטלה י ְִרעּו כאחד‬. See also Qoh 11:6; Ezra 3:9, 6:20; Neh 7:66; 2 Chr 5:13. It is standard in the Aramaic targum to Isaiah 40–66, appearing 16 times, in Tg. Ps.-J. 40:5; 41:19, 20, 21; 43:9, 17; 45:16, 20, 21; 46:2; 48:13; 52:8, 9; 60:13; 65:7; 66:17 (‫ֲדא‬ ָ ‫) ַּכח‬. Compare also its frequent occurrence in rabbinic literature, as in t. Neg. 1.11; y. Sanh. 6.6.23c, ʿAbod. Zar. 1.3.39c; b. Sanh. 18b, ʿAbod. Zar. 76a, Qidd. 6a (‫ֲדא‬ ָ ‫ ;) ַ ּכח‬and in Qumran: 4Q223–24 2 ii 48 (‫;)ּכ ֶאחָד‬ ְ 4Q254 5–6:3, 4Q271 3:4 (‫)ּכ ַאחַת‬. ְ 4.1.2.  ‫‘ ִמָּדה‬tax, tribute’ (Isa 45:14) ‫( אנשי ִמָּדה‬read ‫‘ נ ֹ ְׂשאֵי ִמָּדה‬bearing tribute’); compare Ezra 4:20, 6:8; Neh 5:4; and ‫ְדּה‬ ָ ‫ ִמנ‬Ezra 4:13, 7:24; and in rabbinic literature (‫ְּדה‬ ָ‫)מנ‬, ִ compare, for instance, b. Ned. 62b; Gen. Rab. 64:29. It almost completely replaces earlier ‫מַס‬, which appears 23 times in the Bible and only 3 times in LBH—Esth 10:1; 2 Chr 8:8, 10:18. 4.1.3.  ‫‘ גׁשׁש‬to grope’ (Isa 59:10 [2×]) serves as the Aramaic equivalent of Hebrew ‫ ;מׁשׁש‬for example, Tg. Ps.-J. Gen 27:12 ‫ׁשינַנִי‬ ִ ‫ַׁש‬ ְ ‫( ְיג‬Heb. ‫ּׁשנִי‬ ֵ ֻ‫ ;) ְימ‬so, too, in vv. 21, 22; Tg. Neof. Deut 28:29 ‫( מגׁששין‬Heb. ‫ ;)ממשש‬in m. Ḥal. 2:2 ‫הספינה‬ ‫‘ גוששת‬The ship gropes the ground (in the harbor)’; in b. Šabb. 152b ‘He felt him (‫( )גְׁשַ ׁשֵיּה‬and) saw that there was substance in him’; and at Qumran: CD 1:9; 4Q268 1:16 ‫כעְורים וכמגששים‬ ִ ‫ ;ויהיו‬4Q306 2:4 ‫מגששים‬ ‫יהיו‬ ‫נפשם‬ ‫ובכל‬ [‫ד]רך‬. (Interestingly, the verb ‫ משש‬does not appear at all in the Qumran scrolls.)

296

Shalom M. Paul

4.1.4.  ‫‘ אׁשׁש‬be strong/firm’ (Isa 46:8) is a hapax legomenon verb (‫ ׁששּו‬ ָ ֹ ‫)התא‬ from the Aramaic noun ‫‘ אֻּׁשַ ּיָא‬foundation’ (Ezra 4:12, 6:3; and maybe also Jer 50:15). Compare also rabbinic literature, for example, Gen. Rab. 67 ‫והיכן‬ ‫‘ נתאוששו בידו‬When were (the blessings) confirmed?’; Cant. Rab. 4:2 ‫הלכות‬ ‫‘ המאוששות‬well-founded/firmly established decisions’. 4.1.5.  Aramaic verbs also serve as the b-word in parallel poetic stichs with their Hebrew counterparts in Isaiah 44–66: 4.1.5.1.  ‫ נטל‬// ‫( נׂשא‬Isa 63:9) appears only here in the Piel (‫) ַו ְינ ְַּטלֵם ַו ְינַּשְׂאֵם‬. For ‫‘ נטל‬to lift, raise’, see Dan 4:31 ‫ִטלֵת‬ ְ ‫ׁש ַמּיָא נ‬ ְ ‫ ַע ְינַי ִל‬, 7:4 ‫ְטילַת מן ארעא‬ ִ ‫ּונ‬. It serves as the Aramaic translation of Heb. ‫ נׂשא‬in many verses, such as Tg. Onq. Exod 10:13, Deut 32:11; Tg. Ps.-J. Ps 134:2. It occurs extensively in Rabbinic Hebrew as both a noun and a verb, as well as twice at Qumran: CD 11:10 = 4Q271 5 i 6 ‫אל יטול בבית מושבת סלע ועפר‬. 4.1.5.2.  ‫ סבל‬// ‫‘ נׂשא‬to carry a load/burden’ (Isa 46:4, 7; 53:4, 11). For the verb ‫ סבל‬in LBH, see Lam 5:7; Ps 144:14; Qoh 12:5. It serves as the Aramaic translation of ‫נׂשא‬, for example, in Deut 24:15, 32:11; Job 21:3; and appears in Rabbinic Hebrew and Aramaic as both a noun and a verb; as well as at Qumran, twice as a verb 4Q200 2:2 (‫ ;)סבול‬4Q525 5:12 (‫ ;)יסבילו‬and once as a noun 4Q161 2–4:10 (‫)סֹבֶל‬. 4.1.5.3.  ‫ סגד‬// ‫‘ השתחוה‬to bow down’ (Isa 44:15, 17, 19; 46:6). For the verb alone in Aramaic, see Dan 2:46; 3:5, 6, 7, 10, 11, 12, 14, 15 (2×), 18, 28. See, too, the multiple examples of the targumic translation of the Hebrew verb—for example, in Tg. Onq. Gen 24:26, Exod 34:8; Tg. Neof. Num 25:2, Deut 19:3. It occurs extensively in rabbinic literature in both Aramaic and Hebrew, as in b. Sanh. 61b, Šabb. 72b, Ḥul. 62b. 4.1.5.4.  ‫ בחר‬// ‫‘ צרף‬to test’ (Isa 48:10) is a hapax legomenon homonym of ‫‘ בחר‬to choose’. See also, for example, Tg. Ps.-J. Judg 7:4 ‫ָרנּון‬ ִ‫ֶבח‬ ְ ‫( ְוא‬Heb. ‫ֶצ ְרפֶּנּו‬ ְ ‫)וא‬. ְ It is also the equivalent of Heb. ‫בחן‬, for example, in Tg. Ps.-J. Jer 17:10 ‫ָתא‬ ָ ‫ּכּולי‬ ְ ‫בחַר‬ ּ ָ (Heb. ‫ ;)ּבֹחֵן ְּכלָיֹות‬Zech 13:9 ‫ֲרינּון‬ ִ‫ֶבח‬ ְ ‫( ְוא‬Heb. ‫ְּתים‬ ִ ‫)ּוב ַחנ‬. ְ (It is interesting to note that the Hebrew expression ‫[ כסף נבחר‬Prov 10:20] is translated in the LXX as πεπυρωμένος ‘refined in fire’.) 4.1.5.5.  ‫ מדד‬// ‫ שקל‬// ‫‘ כול‬to measure’ (Isa 40:12). ‫ כול‬is a hapax legomenon. See, for instance, Tg. Onq. Exod 16:18, Deut 25:14, Lev 19:35; Tg. Ps.-J. Ruth 3:15. It is also used extensively in rabbinic literature, in Aramaic, as in b. Menaḥ. 53b, ‘In the final analysis he is only measuring (‫ )כייל‬a tenth’; Ḥul. 105b ‘Anything that is tied, sealed, measured (‫)וכייל‬, and counted’. 4.2.  Hebrew Verbs 4.2.1.  ‫עור‬, in the Hiphil stem, ‘to raise up a conquering nation, individual, or instrument of destruction’: Isa 41:2 (‫ֶדק‬ ֶ ‫ צ‬or possibly read ‫ָּדיק‬ ִ ‫)צ‬, 25 (referring to Cyrus); 42:13 (‫;)ק ְנאָה‬ ִ 45:13 (referring to Cyrus). See, for example, Isa 13:17 (a late prophecy): ‫ָדי‬ ָ ‫ ;מ‬Jer 50:9 ‫;קהַל גוים‬ ְ 51:1 ‫רוח משחית‬, 11 ‫רוח מלכי‬

Signs of Late Biblical Hebrew in Isaiah 40–66

297

ְ ‫ה ַבי‬ ‫ ;מָדַ י‬Ezek 23:22 ‫ִך‬ ֲ ‫;מ ַא‬ ְ Hag 1:14 ‫ ;רוח זרֻבבל‬Dan 11:2 ‫מלכות יָוָן‬, 25; Ezra 1:1 ‫ ;רוח ּכֹרֶׁש‬1 Chr 5:26 ‫ ;רוח ּפּול‬2 Chr 21:16 ‫רוח הפלשתים‬. So, too, in Qumran, 4Q161 2–4:8 (with ‫ ;)ׁשֹוט‬4Q386 1 iii 8 (with ‫) ֵחמָה‬. Compare also at Qumran, in the Polel stem: CD 19:7 ‫( עורר‬with ‫ ;)חרב‬1QHa 14:29 ‫( יעורר‬with ‫)אף‬. It also appears similarly in Rabbinic Hebrew and Aramaic. This verb generally replaces the earlier ‫ קום‬in the Hiphil stem in similar contexts. See, for example, Judg 2:16 ‫ׁשֹפטים‬, 18 ‫ ;הַּׁשֹפֵט‬3:15 ‫ ;ומושיע‬2 Sam 12:11 ‫ ;רעה‬1 Kgs 11:14, 23 ‫;ׂשטָן‬ ָ Amos 6:14 ‫גוי‬. Also note that the Hithpael stem appears in Biblical Hebrew only in Isa 51:17, 64:6; Job 17:8, 31:29. 4.2.2.  ‫בין‬, in the Hiphil stem, ‘to teach’, Isa 40:14. So also in Ps 119:27, 34, 73, 125, 130, 144, 169 (a late psalm); Job 6:24, 32:8; Dan 1:17; 8:16, 27; 9:22; 10:14; 11:33; Ezra 8:16; Neh 8:8; 2 Chr 35:3; and compare the meaning ‫מבין‬ ‘teacher’ in 1 Chr 15:22; 25:7, 8; 27:32. 4.2.3.  ‫‘ גאל‬to (be) defile(d)’ (Isa 59:3, 63:3), a by-form of ‫געל‬, appears elsewhere only in late texts and never in EBH. See Zeph 3:1; Mal 1:7 (2×); Lam 4:14; Ezra 2:62 = Neh 7:64; Dan 1:8 (2×). It also occurs with this meaning at Qumran, CD 12:16 ‫ ;והאבנים והעפר אשר יגואלו בטמאת אדם‬1QM 9:8: ‫להתגאל‬ ‫ ;בדם טמאתם‬see also 4Q274 1:6; 4Q379 3:5; 4Q513 13:3, 4; 11Q19 47:13. 4.2.4.  ‫‘ עמד‬to abide, endure’: Isa 66:22 ‫כי כאשר השמים החדשים והארץ‬ ‫ כן יעמד זרעכם ושמכם‬. . . ‫ עֹמדים ְל ָפנַי‬. . . ‫החדשה‬. For additional examples, see Jer 32:14 ‫בכלי ֶחרֶׂש למען יעמדו ימים רבים‬ ִ ‫;ונתּתם‬ ָ Hag 2:5 ‫;ורוחי עֹמדת בתוככם‬ Ps 102:27 ‫ ;המה יֹאבֵדו ואתה תעמֹד‬Esth 3:4 ‫ ;היעמדו דברֵ י מרדכי‬Dan 10:17 ‫ואני‬ ‫מעתה לא יַעֲמָד בי כח‬, 11:17 ‫לא תעמד ולא לו תהיה‬. Contrast this last example with Isa 8:10 ‫ּדַ ברו דבר ולא יקום‬, 14:24 ‫וכאשר יעצתי היא תקום‬. Moreover, as is well known in the latest books of the Bible, the root ‫עמד‬ often takes on the basic meaning previously held by the root ‫קום‬. Compare, for example, Exod 1:8 ‫ ויקם מלך חדש‬with Dan 8:23 ‫ ;יעמֹד מלך עַז ּפָנים‬Num 25:7 ‫ ויקם‬. . . ‫ וירא פינחס‬with Ps 106:30 ‫ ;ויעמֹד פינחס‬Deut 19:15 ‫ דבר‬. . . ‫יקום‬ with Esth 3:4 ‫ דברי מרדכי‬. . . ‫הֲיעמדו‬. See also, for example, Dan 8:22; 12:1, 13; Ezra 2:63 = Neh 7:65; and compare, for instance, m. Sanh. 10:3 ‫אינן עמדים‬ ‫בדין‬, paraphrasing Ps 1:5 ‫ במשפט‬. . . ‫ָקמּו‬ ֻ‫ ;לא י‬Mek. Yithro 9 ‫נביא עתיד אני‬ ‫להעמיד מהם‬, paraphrasing Deut 18:18 ‫נביא אקים להם‬. See also the interchange of the two verbs in CD-A 7:19–20 (citing Num 24:17) . . . ‫וקם שבט מישראל‬ . . . ‫ובעמדֹו‬. 4.2.5.  ‫‘ מצץ‬to suck’ (Isa 66:11) is a hapax legomenon by-form of ‫מצי‬. Compare Tg. Ps.-J. Ps 75:9 ‫( ְימ ְַּצצּון‬for Heb. ‫ִמצּו‬ ְ ‫)י‬. It appears also in rabbinic literature, as in m. Parah 9:3 (‫)מוצצת‬, Šabb. 19:2 (‫)ומוצצין‬. 4.3.  Late Hebrew Nouns 4.3.1.  ‫‘ שרב‬parched ground’ (Isa 49:10) appears elsewhere only in 35:7 (a chapter usually considered to be related to Deutero-Isaiah) with the meaning ‘burning heat’. For the former, see, for example, Sir 43:22; and for the latter,

298

Shalom M. Paul

see Tg. Onq. Gen 31:40 ‫ ;ׁשַ ְר ָּבא‬and Tg. Ps.-J. Gen 31:40 ‫ׁשריבא‬ ְ (for Heb. ‫;)חֹרֶב‬ see also y. Sanh. 10.5.29c; Gen. Rab. 17:4. 4.3.2.  ‫‘ צאצאים‬offspring’ (Isa 44:3, 48:19, 61:9, and 65:23) appears elsewhere only in Job 5:25, 21:8, 27:14, and 31:8. Compare also at Qumran 17 times, as in 1QM 10:13; 1QHa 10:9; 4Q185 1–2 ii 15; and in Rabbinic Hebrew, for instance, in b. B. Meṣiʿa 107a, Taʿan. 6a. The term replaces EBH ‫זֶרַ ע‬, ‫בנים‬, and ‫בני בנים‬. 4.3.3.  ‫‘ מסמרים‬nails’ (Isa 41:7) also appears in 1 Chr 22:3; and ‫( מסמרות‬Jer 10:4; Qoh 12:11 [‫ ;]מׂשמרות‬2 Chr 3:9); and frequently in rabbinic literature in Hebrew (for example, b. B. Bat. 7b; Gen. Rab. 68:13) and Aramaic (for example, y. Ḥag. 3.1.78d, Pesaḥ. 5.3.32b (top); and once at Qumran (CD 12:17 ‫)וכל כלי מסמר או יתד‬. 4.3.4.  ‫‘ זיקות‬firebrands, sparks’ (Isa 50:11) is a biblical hapax legomenon (similar to ‫ִּקים‬ ִ ‫ז‬, Prov 26:18) occurring in Sir 43:13, parallel to ‫‘ ברק‬lightning’; at Qumran: 1QM 6:3; 1QHa 9:12; 4Q286:3, 4 (‫ ;)זקים‬and in rabbinic literature, for example, y. Ber. 5.1.9a; b. Ḥul. 137b (‫)זיקוקין דנור‬. 4.3.5.  ‫‘ לאחור‬future’ (Isa 41:23, 42:23) also appears in Jer 7:24 and Ps 114:3, 5 (similar in meaning to ‫)אחרית‬. Compare Sir 6:28 ‫כי לאחור תמצא‬ ‫מנוחתּה‬, 12:12 ‫ָרי‬ ָ ‫אמ‬ ֲ ‫ ;ולאחור תׂשיג‬and m. Ḥag. 2:1 ‫מה לפנים ומה לאחור‬. (Though the words ‫ אחור‬and ‫ לאחור‬appear more than 30 other times in the Bible, their meaning is ‘backward, back part, behind’.) 4.3.6.  ‫‘ עולמים‬eternity’ (Isa 45:17 [2×], 51:9), prevalent in LBH, replaces ‫עולם‬, ‫נצח‬, and ‫עד‬. See Ps 145:13 and Qoh 1:10; compare Dan 2:4 ‫ָל ִמין‬ ְ ‫מ ְַלּכָא ְלע‬ ְ ‫ְחי אֲדֹנִי ַה ֶּמל‬ ‫ חֱיִי‬with 1 Kgs 1:31 ‫ ְלעֹלָם‬. . . ‫ֶך‬ ִ ‫י‬, 2:44, 6:27, 9:24; 2 Chr 6:2 (= 1 Kgs 8:13). Compare the targums, such as Tg. Onq. Exod 15:18 ‫לעלמא ולעלמיא‬ (Heb. ‫)לעולם ועד‬, Deut 32:40 ‫( לעלמין‬Heb. ‫ ;)לעולם‬Tg. Ps.-J. Isa 25:8 ‫לעלמין‬ (Heb. ‫)לנצח‬, Mic 7:18 ‫( לעלמין‬Heb. ‫ ;)לעד‬and is used extensively at Qumran, appearing dozens of times, for instance, in 1QM 13:7 ‫;שמכה נברכה לעולמים‬ 1QS 4:22 ‫ברית עולמים‬, which partially replaces ‫ברית עולם‬, as in Gen 9:16, 17:7; Exod 31:16; Lev 24:8. 4.3.7.  ‫ֲרב‬ ָ ‫‘ ַמע‬west’ (Isa 43:5, 45:6, 59:19) also appears in Ps 78:7, 103:12, 107:3 (all of which have late features); Dan 8:5; 1 Chr 7:28; 12:16; 26:16, 18, 30; 2 Chr 32:30, 33:14, a late term replacing classical ‫ים‬. Compare, for example, Josh 12:7 ‫ בעבר הירדן ימה‬with 1 Chr 26:30 ‫ ; ֵמעֵבר לירדן מערבּה‬and is prevalent in the targums, such as Tg. Onq. Gen 28:14 ‫ ;ומערבה‬as well as Sam. Tg. Gen 28:14 ‫( למערב‬for Heb. ‫ ;)ימה‬occurring extensively in rabbinic literature, as in m. Maʿaś. Š. 5:2 ‫לוד מן המערב‬, ʿErub. 3:5 ‫ מערב‬. . . ‫ ;מזרח‬Sipre Numbers 73 ‫ ;תוקע לצפון ולמערב‬and at Qumran (12×) as a noun, as in 3Q15 8:11; 4Q274 1:2; 11Q19 30:7; and (5×) as an adjective in 3Q15 3:10; 6:12; 10:8, 13; 11:16.

Signs of Late Biblical Hebrew in Isaiah 40–66

299

5.  Phrases and Expressions 5.1.  ‫‘ בן בטן‬child of the womb’ (Isa 49:15) also appears in Job 19:17. Its Aramaic equivalent is ‫( בר בטני‬Prov 31:2). The Hebrew expression replaces EBH ‫פרי בטן‬, for example, in Gen 30:2, Deut 28:4, Isa 13:18, and Mic 6:7. 5.2.  ‫‘ עם הק ֹדש‬the holy people’ (Isa 62:12, 63:18 [‫)]עם קדשך‬, later occurring in Dan 12:7 (‫ ;)עם ק ֹדש‬compare Dan 8:24 (‫ ;)ועם קדושים‬Qumran 1QM 12:1 (‫ ;)עם קודש‬4Q171 3–4 iii 8 (‫ ;)עם קדשו‬4Q248:9 ([‫ ;)עם הק[דש‬1QM 14:12 (‫)עם קדשכך‬. In EBH, the parallel expressions are ‫( עם קדוש‬Deut 7:6; 14:2, 21; 26:19; 28:9) and ‫( גוי קדוש‬Exod 19:6). 5.3.  ‫ צבאות‬′‫‘ אמר ה‬says Yhwh of Hosts’ (Isa 45:13). Compare 1 Sam 15:2, 2 Sam 7:8 as against Jeremiah (51×), Haggai (7×), Zechariah (21×), Malachi (21×), and 1 Chr 17:7. 5.4.  ‫‘ רוח הק ֹדש‬the holy spirit’ (Isa 63:10, 11 [‫ )]רוח ָקדשו‬appears once in Ps 51:13 (‫ ;)רוח קדשך‬and extensively in rabbinic literature—for example, in m. Soṭah 9:6, 15; t. Soṭah 12:5, 13:3; and at Qumran, for example, in 1QHa 4:26, 6:13, 8:11, 16, 21; 1QSb 2:24; CD 2:12, 5:11, 7:4; 1QS 4:21, 8:16, 9:3. 5.5.  ‫‘ זבֻל קדשך‬Your holy heights’ (Isa 63:15) is a hapax legomenon; compare Qumran,for example, 1QHa 11:34 (‫ ;)זבול קדשו‬4Q256 19:1 (‫;)מזבול קדשו‬ 1QS 10:3 (‫ ;)מזבול קודש‬1QM 12:1 (‫ ;)מזבול קודשכה‬4Q491 5–6:1 ([‫בזבול‬ ‫)קוד[שכה‬. 5.6.  ‫‘ עיר הקדש‬the holy city, city of the temple’ (Isa 48:2, 52:1, 64:9 [‫ערי‬ ‫ ;]קדשך‬appearing later in Dan 9:24 [‫]עיר קדשך‬, 26 [‫ ;]והעיר והק ֹדש‬Neh 11:1, 18); compare Sir 36:18 (‫ ;)קרית קדשך‬and at Qumran: CD 20:22; 4Q176 8–11:2; and twice in the Talmud: b. Qam. 97b, Soṭah 47a. 5.7.  ‫‘ בית קדשנו‬our holy temple’ (Isa 64:10); appearing later in 1 Chr 29:3 (‫ ;)בית הק ֹדש‬2 Chr 3:8, 10 (‫ ;)בית ק ֹדש הקדשים‬and at Qumran: 1QS 8:5, 9:6; 4Q176 16:3. The expression ‫ בית קדש הקדשים‬also appears in rabbinic literature, for example, in m. Mid. 4:7; y. Ber. 4.5.8c, Kil. 8.4.31c. 5.8.  ‫‘ בית תפלה‬house of prayer’ (Isa 56:7 [2×]) is a hapax legomenon. See also, for instance, y. Ber. 4.5.8c; Midr. Cant. Zuṭa 1.1. References

Joosten, J. 2005 The Distinction between Classical and Late Biblical Hebrew as Reflected in Syntax. HS 46: 327–39. Paul, S. M. 2012 Isaiah 40–66: Translation and Commentary. Eerdmans Critical Commentary. Grand Rapids, MI: Eerdmans. Young, I., ed. 2003 Biblical Hebrew: Studies in Chronology and Typology. JSOTSup 369. London: T. & T. Clark. Young, I.; Retzetko, R.; and Ehrensvärd, M. 2008 Linguistic Dating of Biblical Texts. 2 vols. London: Equinox.

Language Variation, Discourse Typology, and the Sociocultural Background of Biblical Narrative Frank H. Polak 1.  Introduction and Objectives This essay concerns basic distinctions between different styles in biblical narrative prose that can be described linguistically and that suggest a complex interaction between oral and written discourse. But no less important is the issue of the sociocultural contexts within which these narratives were written and the prevalent modes of communication described in them. In order to discuss these matters systematically and to demonstrate their importance, I propose a four-step analysis of biblical narrative prose that will lead (a) from syntactic-stylistic analysis to discourse typology, (b) from typology to communicative situation, (c) from communicative situation to sociocultural background, and (d) from sociocultural to sociohistorical background and, thereby, to periodization. 1 At the outset, I want to clarify some of my presuppositions. The language of biblical narrative knows much variation, with manifold roots in geography and historical circumstances (Young 1993). 2 What is of most interest for the 1.  I speak of periodization because the periods involved do not define an unequivocal division on the time line. Overlapping is always present since what a scribal apprentice learned in his youth may accompany him long after decisive sociopolitical change and may also be transmitted to his apprentices (see n. 4 below). Hence, it would be methodologically unsound to invoke an absolute chronology. 2.  In the present essay, I do not wish to touch upon geographical variation, which is only barely visible in epigraphic remains. Some phonological and morphological variation (contraction of the /ay/ diphthong; assimilation of /n/ before the female ending /t/) is reflected in the Samaria ostraca (Young 1993: 29, 114–15). The Deir ʿAlla text combines features that are uniquely Aramaic with phenomena that can only be described as Canaanite/Hebrew (Hoftijzer 1976: 287–93; Young 1993: 50–54); they give this text a unique place within the Syro-Palestinian dialect continuum (Garr 1985: 229). The suggestion that this mixture may represent Aramaic archaizing poetic language (Hoftijzer 1976: 301–2) has lost none of its attractiveness.

301

302

Frank H. Polak

periodization of Biblical Hebrew is the connection between language variation and the particular uses of language in terms of social circumstances and social forces. This is the subject matter of the various branches of sociolinguistics, the aims of which are to analyze “the characteristics of language varieties, the characteristics of their functions, and the characteristics of their speakers as these three constantly interact, change, and change one another within a speech community” (Fishman 1970: 4). When viewed in this perspective, language usage is connected to the sociocultural milieu and thus, indirectly, to the sociohistorical and sociopolitical background. Because the history of Israel was subject to a number of acute constituent crises, including the Babylonian conquest, exile, and Babylonian-Persian domination, these sociohistorical and sociopolitical conditions could not fail to affect language in a most significant way. 3 Thus, the sociolinguistic analysis of stylistic variation can reveal important features of the social, historical, and political milieus in which biblical texts came into being. The position of literary language in such an analysis needs clarification. An important distinction sets the local/regional vernacular apart from the supraregional administrative language, which must be distinguished from literary language because of poetic elements (such as the traces of parallelism in Genesis 1–3), which are expected in literary but not administrative language. One of the sources of change in the supraregional administrative and literary languages is the adoption of vernacular features, but this type of change is countered by scribal education which promotes conservation of the basic features of the supraregional language, a phenomenon described by sociolinguists as language maintenance (Fishman 2006; Saville-Troike 2003: 224–26; Fasold 1987: 213–42). 4 The language of oral literature shares the oral medium with colloquial language, but it is a supraregional literary creation, as indicated by 3.  Kofoed (2006: 106–12) discusses the dating of texts within the sociocultural continuum of preconquest Anglo-Saxon English, but he disregards the fact that the Norman invasion of 1066 formed a major break in the sociolinguistic and sociocultural structure of the British island (see Baugh and Cable 2002: 98–112), which led to the ongoing “erosion” of the Old English poetic tradition (Amodio 2004: 79–98). In spite of the continuity of literary copying (Strang 1970: 240–41), Irvine (2004: cxxxix–clxvi) shows that, 50 years after the invasion, the language of the Anglo-Saxon chronicles (the Peterborough chronicle) lost important morphological and syntactic distinctions (see also Horobin and Smith 2002: 126–29, 142–45; Baugh and Cable 2002: 146–55; Strang 1970: 314–16). 4.  In the ancient Near Eastern context, the conservative influence of scribal schooling stands out in the Akkadian literary texts in the Standard Babylonian dialect, as described by Kouwenberg in the present volume. Kofoed (2006: 103 n. 21) adduces the “Great Prayer to Ishtar,” which von Soden assigned to the Neo-Babylonian period but which also appeared in Boghazköy (Reiner and Güterbock 1967), as evidence for the weakness of linguistic evidence in dating literary sources. Actually in this case, scholars erred by assigning a late date to an early text and not by the attribution of an early date to a late text. However, Reiner and Güterbock (1967: 261–63) also show that the Neo-Babylonian text contains expansions

Sociocultural Background of Biblical Narrative

303

the use of epic formulas that preserve parallelism and are shared by biblical Hebrew, Ugaritic, Aramaic, and Old Babylonian. 5 Actually, these distinctions pertain less to language itself than to language usage, and thus to the situation in which language is used for communication. I distinguish three main corpora: 1.  The Achaemenid corpus, representing the (late) Neo-Babylonian and Persian period (Ezra–Nehemiah, Esther, 1–2 Chronicles, and, in the Seleucid period, Daniel 1). 2.  The Judean corpus, consisting mainly of Deuteronomy and Deuteronomistic texts (Deuteronomy 1–3, 9–10, 34; Joshua; 1–2 Kings; the Jeremiah Vita; large sections of the so-called Priestly corpus). This corpus reflects the Judean monarchy of the seventh and early sixth century. 3.  The Medial corpus, consisting of texts that cannot be related to the first two corpora. This corpus includes narratives such as the tales of the patriarchs, Samuel, Saul and David, Elijah and Elisha, and parts of the tales of Moses and the savior-judges.

What sets the Medial corpus apart from the Judean and Achaemenid texts is its preservation of many features that are characteristic of oral language and its multifaceted interaction with texts in writing. In the study of early mediaeval European literature, characteristics of this sort are subsumed under the headings Vokalität and Medialität (Schaefer 1992), because the oral and written media were employed similarly for both literary and nonliterary types of communication. Bäuml (1993) shows how in the medieval period oral texts were written down, whereas poetic texts in writing were delivered orally. In my view, a complex symbiosis of this kind is embodied by the Bible’s Medial corpus. The criteria for this breakdown into three corpora are partly linguistic. Verheij (1990: 56–83) indicates a characteristic differentiation between the frequency of word classes (verbs, nouns and participles/infinitives) in biblical historiography. Thus he sets the language usage of Chronicles apart from that of Kings, which in its turn is distinct from that of Samuel. Other criteria are factual. Texts are considered to belong to the Achaemenid corpus because of vis-à-vis the older text (for instance, lines 42–50) and thus forms a rewritten version of the prayer rather than a late copy. 5.  I refer to such recurring phrases as: ‘he lifted his eyes and saw’ (‫;)וישׂא עיניו וירא‬ ‘he lifted his voice and called out’ (‫‘ ;)וישׂא קולו ויקרא‬he fell . . . and did homage’ (. . . ‫ויפל‬ ‫ ;)וישתחו‬see Cassuto 1973–75: 2.69–109; del Olmo Lete 1981: 36–37, 54–58; Polak 1989; 2006. The phrase ‘he lifted his eyes and saw’ is attested in Old Babylonian Gilgamesh and the Hittite Ullikummi myth (Polak 1996: 95–96; 2006: 287; George 2003:1.177 [Pennsylvania Tablet iii, line 137]; Güterbock 1951: 157). The phrases involving ‘he did homage’ include ‫השתחוה‬, a verb in the Št-stem (Hishtaphel) that is not productive in Hebrew (Cohen 2004; Smith 1994: 168 n. 95). The formulaic phrase ‘he spoke up and said’ (‫ )ויען ויאמר‬is matched by Aramaic (Ahiqar 110, 118–19; see Polak 2006: 294–95).

304

Frank H. Polak

explicit references to Persian domination or references to realia such as Persian (or Greek) daric coins of gold, ‫אֲדַ ְר ּכ ֹנִים‬‎‫( ‏‬1 Chr 29:7; Zadok 1982: 296). Exegetical considerations bearing on the theology of a text are sometimes used to favor the postexilic dating of a text. But one set of these ideational considerations, more often than not, can be counterbalanced by another set that argues for different or even the opposite conclusions. Consequently, such considerations cannot provide a primary framework for literary periodization. The steps for determining the sociocultural background of a text from its language are more complicated than the often relatively straightforward determination of grammatical borrowing or isoglosses. Language usage includes stylistic features that are to a large extent a matter of free choice, rhetoric, and literary design. In a literary context, style is conceived of as “the overall organization of meaning through form . . . by the working together of sentence structure, vocabulary, figures of speech, rhythm and many other elements” (Brooks and Warren 1958: 375–78). However, Josephine Miles (1964: 213– 22) has shown that the language usage of English poetry (and prose), though basically free, reveals certain noun/verb/adjective proportions that are more or less fixed for different periods. Features of this sort, then, form a syntacticstylistic pattern, which can be summarized in a discourse profile. In the present study, I sketch a discourse profile of the three corpora in biblical narrative, always taking into account the need to test whether and to what extent this profile is sensitive to rhetoric variation. I first offer a short review of the methods used and the results attained by them. Then I survey additional linguistic markers and sociocultural factors that set the Medial corpus apart from the Judean and Achaemenid corpus. Finally, in the last part of this study, I consider the question of sociohistorical periodization and the status and characteristics of narratives that seem exceptional. 2.  From Syntactic-Stylistic Classification to Discourse Typology 2.1.  Syntactic-Stylistic Features Language usage has always been an important aspect of the study of the Hebrew Bible. The distinction between Priestly, Deuteronomic, and other literary complexes is basically dependent on lexical analysis and in particular on phraseology. However, syntactic criteria can be at least as important. Lohfink (1999: 41–42) and Römer (2000: 179–80) recognized that in the Deuteronomic style infinitive clauses are far more frequent than in the Holiness Code. For the book of Judges, Richter (1964: 141) demonstrated the “dissolution of the ancient forms and their pregnant syntax” as against “the advance of parenesis . . . with its greater verbosity.” However, what was sorely lacking in these statements was an objective overall analysis of syntactic features that would be able to go beyond impressionistic notions.

Sociocultural Background of Biblical Narrative

305

In order to overcome this limitation, I have devised a discourse profile based on a clause-by-clause analysis, according to three standard features and two additional features that are often diagnostic (see §§2.1.4–5 below). The features are all connected to explicit, lexicalized sentence constitutents (ELC’s) and a scale of syntactic complexity. 6 2.1.1.  Sentence Constituents of Independent Clauses A central feature of the analysis I propose relates to the number of explicit, lexicalized sentence constituents (ELC) apart from the predicate, such as the subject (noun or independent pronoun), object, or indication of place, time, manner, reason, or goal (arguments and adjuncts). When the subject is only indicated by the obligatory prefix or affix of the finite verb, it is considered part of the verb, and is not counted as an ELC. 7 By the same token, the object suffix is not considered an ELC. It is true that, from a purely syntactic point of view, the object and subject are sufficiently indicated by a suffix or prefix/affix. But anaphoric reference and, thereby, the discourse structure is affected when only the suffix or prefix/affix is used to indicate the participants (Longacre 1989: 141–42). After all, noun constructions are far more explicit than are references by suffixes or prefixes/affixes (Fox 1987: 111–18, 139–44). Even indications by independent pronouns (or particles with suffixed pronouns) carry more emphasis than the enclitic constructions. Notably, unlike prefix/affix or object suffix, an independent pronoun can be fronted and carries an accent in its own right. From a discourse point of view, it is highly significant that many narrative tales prefer short clauses, consisting of the predicate only (with subject prefix/affix and object suffix) or of the predicate with one ELC (a slash indicates a clause boundary): 8 6.  For more details about the categories and methods of analysis, see Polak 1998: 71–87; 2010: 38–41. Syntactic analysis provides the full picture behind the statistics of the Noun/ Verb–Finite/Nominal Verb distribution (Polak 1998: 71), in which many data are lost. Heuristically, these statistics formed only the point of departure for the syntactic analysis and were useful for establishing a broad picture as backdrop. However, contrary to the way in which Young, Rezetko, and Ehrensvärd (2008: 1.33–35) describe my method, such statistics have no place in the discourse typology that I propose. 7.  Hebrew has two distinctions that do not exist in modern languages such as English, French, or German: (1) In these languages, the subject pronoun is always an independent pronoun, whereas Hebrew can use the verbal forms with prefixed/affixed pronominal subjects only or can expand the verbal construction by means of an independent pronoun. In modern languages, the translation of these sentences is not able to distinguish between a construction with separate, independent pronouns and a construction with prefixed/affixed pronominal subject only. (2) In Biblical Hebrew, object ellipsis permits the deletion of the object, which must be supplied in modern languages (in translation, these will be indicated by parentheses). 8. The English translations below usually follow the njpsv, sometimes with slight modifications. The parentheses indicate additions for the sake of English syntax and idiom.

306

Frank H. Polak

Neh 2:1:

ְ ‫ֶּתנָה ַל ֶּמל‬   ‫ִיתי רַ ע ְל ָפנָיו‬ ִ ‫ ְולֹא־ ָהי‬/‫ֶך‬ ְ ‫ ָוא‬/‫ֶּׂשא אֶת־ ַהּיַיִן‬ ָ ‫ָוא‬ I took the wine / and gave (it) to the king. / I had never been out of sorts in his presence.

I assign clauses of this type to a class called “short clauses.” A second class comprises all clauses with at least two ELC’s: ‫ָוא ְֵצאָה ְבׁשַ עַר־ ַהּגַיא ַל ְילָה‬ (1)  ​ I went out (1) by the Valley Gate (2) at night. (Neh 2:13) 2.1.2. Clause Subordination The clause subordination category, which is indicated by a hyphen, includes all clauses in hypotaxis in general, such as relative clauses, infinitive clauses, participial clauses; and clauses introduced by a conjunction to indicate, for example, perception, indirect speech, time, place, manner, reason, or goal—such as: ‫למען‬, ‫יען‬, ‫פן‬, ‫כאשר‬, ‫באשר‬, ‫כי‬: 9 ‫ָבים‬ ִ ‫ָחים— ֵמעֲלֹות הַשַ ּׁחַר—עַד צֵאת הַּכֹוכ‬ ִ ‫ִיקים ָב ְּרמ‬ ִ ‫חז‬ ֲ ‫ֶציָם ַמ‬ ְ ‫‏ ְ‏וח‬ (2)  ​ and half held lances—from the break of day—until the stars appeared. (Neh 4:15[21]) Subordinate clauses are considered syntactic constituents of the main clause. 2.1.3.  Groups of at Least Two Nouns within a Syntactic Slot The category of more than one noun in a syntactic slot includes constructions such as a construct phrase, a kernel with an attribute or apposition, and a syndetic/asyndetic noun junction (single nouns are not counted): ‫ָעיר‬ ִ ‫ּולחֹומַת ה‬ ְ ‫ֲׁשר־ ַל ַּביִת‬ ֶ ‫ן־לי ע ִֵצים ְל ָקרֹות אֶת־ׁשַ עֲרֵי ה ִַּבירָה א‬ ִ ‫ִּת‬ ֶ ‫ֲׁשר י‬ ֶ‫א‬ ְ ֹ ‫ן־לי ַה ֶּמלֶך ְּכי ַד־אֱלהַי הַּטֹובָה ָעלָי‬ ִ ‫ִּת‬ ֶ ‫ ַוּי‬/‫ֲׁשר־אָבֹוא ֵאלָיו‬ ֶ ‫ְו ַל ַּביִת א‬ “to give me timber for roofing the gates of the acropolis and for the city walls and for the house I shall occupy.” The king gave me these, thanks to my God’s benevolent care for me. (Neh 2:8)

(3)  ​

2.1.4.  Complex Hypotaxis Apart from the category “clause subordination,” which includes all subordinate clauses (simple or complex), special attention is due the subcategory of clauses that are dependent on subordinate clauses or that include two lexi-

9. However, motive clauses that begin with ‫ כי‬are regarded as main clauses, since clauses of this sort can be interrogative (Gen 44:34; 50:19). In the Mesha Stele (line 17), a clause of this type is preceded by a sentence divider. By the same token, adversative clauses are considered main clauses.

Sociocultural Background of Biblical Narrative

307

calized constituents or a noun group (the equal sign indicates the embedded clause): 10 ְ ‫ֶך ה ִַחיצֹונָה—לֵאמֹר ַל ֶּמל‬ ְ ‫חצַר ּבֵית־ ַה ֶּמל‬ ‫ֶך‬ ֲ ‫ְ‏‏ו ָהמָן ּבָא ַל‬ (4)  ​ ‫ֲׁשר־ה ִֵכין לֹו‬ ֶ ‫ָרּדֳכַי עַל־ ָהעֵץ = א‬ ְ ‫= ִל ְתלֹות אֶת־מ‬ Haman had just entered the outer court of the royal palace to speak to the king about having Mordecai impaled on the stake he had prepared for him. (Esth 6:4)

The entire stretch ‫ לאמר–לו‬is dependent on the main clause (level 1); it includes the embedded clause ‫( לתלות את־מרדכי על־העץ‬level 2), which in its turn includes a relative clause that is dependent on ‫העץ‬: ‫( אׁשר־הכין לו‬level 3). 2.1.5. Long clauses, containing at least three ELC’s. These merit special attention, since they indicate particular complexity: ‫יׁשית אֶת־נַעֲרֹו‬ ִ ‫ֲמ‬ ִ ‫ִׁשלַח ֵאלַי ַס ְנ ַבּלַט ּכַָּדבָר ַהּזֶה ַּפעַם ח‬ ְ ‫‏‏ ַוּי‬ (5)  ​ (1) Sanballat sent (2) me (3) the same message (4) a fifth time by (5) his servant. (Neh 6:5) 2.2.  Discourse Profile The syntactic-stylistic classification of the various clauses according to the above categories makes it possible to construct a discourse profile of a given text and thus to appraise the degree of linguistic complexity of various text groups. I calculate the percentage of cases in the following categories, relative to the number of clauses in the entire text: 1.  Short clauses (0–1 ELC) 2.  Long clauses (2+ ELC) 3.  All hypotaxis (simple and complex)

These three classes must add up to 100.00% (or 99.99/100.01%, due to rounding off procedures). 4.  The percentage for noun groups is calculated in the form of mean noun pairs (the number of all grouped nouns divided by 2) relative to the number of clauses 5.  The percentages of clauses in complex hypotaxis 6.  The percentage of clauses with 3 or more ELC’s

2.3.  How It Works In the following tables, I present the analysis of all clauses in a given text, and indicate the category and the number of grouped nouns. The acronym ELC followed by a digit indicates the number of explicit lexicalized constituents 10.  See also the example of Neh 2:8 above, which is in itself an instance of indirect speech.

308

Frank H. Polak Table 1a.  1 Samuel 4:1b–18 (Sample) v. 10

‫ׁש ִּתים‬ ְ ‫ ַו ִּי ָּלחֲמּו ְפ ִל‬1 ELC ‫ִׂש ָראֵל‬ ְ ‫ ַו ִּיּנָגֶף י‬1 ELC ‫ ַוּיָנֻסּו ִאיׁש ְלא ָֹהלָיו‬2 ELC ‫ ו ְַּת ִהי ַה ַּמּכָה ְּגדֹולָה ְמאֹד‬1 ELC ‫ׁשים ֶאלֶף רַ ְג ִלי‬ ִ ֹ ‫ׁשל‬ ְ ‫ִׂש ָראֵל‬ ְ ‫ ַוּיִּפֹל ִמּי‬2 ELC ‫ִל ָקח‬ ְ ‫ ַואֲרֹון אֱל ִֹהים נ‬1 ELC

v. 11 v. 12

NG2 NG3 NG2

‫ּופי ְנחָס‬ ִ ‫ָפנִי‬ ְ ‫ּוׁשנֵי ְבנֵי־ע ִֵלי מֵתּו ח‬ ְ 1 ELC

NG5

‫ֲרכָה‬ ָ ‫ָמן ֵמ ַה ַּמע‬ ִ ‫יׁש־ּב ְני‬ ִ ‫ָרץ ִא‬ ָ ‫ ַוּי‬2 ELC ‫ׁשלֹה ַּבּיֹום הַהּוא‬ ִ ‫ ַוּיָבֹא‬2 ELC

NG2 NG2

‫ֻעים‬ ִ ‫ ּומַָּדיו ְקר‬1 ELC ‫ֲדמָה עַל־רֹאׁשֹו‬ ָ ‫ ַוא‬1 ELC v. 13

‫ ַוּיָבֹוא‬0 ELC ְ ‫ ְו ִהּנֵה ע ִֵלי יֹׁשֵב עַל־ה ִַּכּסֵא יד ֶּדר‬3 ELC ‫ֶך ְמ ַצּפֶה‬ ‫ ִּכי־ ָהיָה ִלּבֹו חָרֵ ד עַל אֲרֹון ָהאֱל ִֹהים‬2 ELC

NG3 NG2

‫ָאיׁש ָּבא‬ ִ ‫(לה ִַּגיד) ְוה‬ ְ 2 ELC ‫ ְלה ִַּגיד ָּב ִעיר‬SUB ‫ָעיר‬ ִ ‫ ו ִַּת ְזעַק ּכָל־ה‬1 ELC

NG2

Table 1b.  Discourse Profile of 1 Samuel 4:1b–18 (All 66 Clauses) Phenomenon

Number Percentage

Phenomenon

0–1 ELC

35

53.03

mean noun pairs

2+ ELC

27

40.91

4

6.06

subordination

Number Percentage 34.5

52.27

3+ ELC

8.0

12.12

complex hypotaxis





apart from the predicate, NG indicates a noun group consisting of the number of nouns indicated by the ensuing digit (must have at least one group of 2, or NG2). SUB refers to subordination and CMPL to complex hypotaxis. 11 Table 1a presents a syntactic-stylistic analysis of one part (vv. 10–13) of the tale of the battle near Ebenezer (1 Sam 4:1b–18, which contains 66 clauses total) in order to provide a sample analysis. 12 Table 1b provides the discourse profile of the entire text, which is a summary of the percentages in the different categories. 11.  The verbal group consisting of ‫ היה‬with a noun predicate is counted as a single predicate. 12.  My analysis of v. 13 is based on the conjectural reading ‫ יד דרך ִמ ְצּפה‬but is not dependent on this reading.

309

Sociocultural Background of Biblical Narrative Table 2a.  2 Kings 22:1–14 (Sample)

ְ ‫ׁשנָה ַל ֶּמל‬ ‫אׁשּיָהּו‬ ִ ֹ ‫ֶך י‬ ָ ‫ֶׂשרֵ ה‬ ְ ‫ׁשמֹנֶה ע‬ ְ ‫ְהי ִּב‬ ִ ‫ ַוי‬1 ELC ְ ‫ׁשלַח ַה ֶּמל‬ ‫ׁשּלָם הַּסֹפֵר ּבֵית יְהוָה‬ ֻ ‫ֶן־מ‬ ְ ‫אצ ְַליָהּו ב‬ ֲ ‫ֶת־ׁשפָן ֶּבן־‬ ָ ‫ֶך א‬ ָ 3 ELC (‫)לֵאמֹר‬

v. 3

NG4 NG 2 NG 6

‫ לֵאמֹר‬SUB ‫ֶל־ח ְל ִקּיָהּו הַּכֹהֵן ַהּגָדֹול‬ ִ ‫ עֲלֵה א‬1 ELC

v. 4

NG 3

(‫ ְוי ַּתֵ ם אֶת־ ַה ֶּכסֶף (הַּמּובָא‬1 ELC ‫ הַּמּובָא ּבֵית יְהוָה‬CMPL NG 2 ‫ָספּו ׁש ֹ ְמרֵ י ַהּסַף ֵמאֵת ָהעָם‬ ְ ‫ֲׁשר א‬ ֶ ‫ א‬CMPL NG 2 ‫ִּתנֻהּו עַל־י ַד עֹׂשֵ י ה ְַּמלָאכָה‬ ְ ‫ ְוי‬1 ELC

v. 5–6a

NG 3

‫ֻפ ָק ִדים בֵית יְהוָה‬ ְ ‫ ַהּמ‬CMPL NG 2 ‫ׁשים‬ ִ ‫ָר‬ ָ ‫(ל ַחּזֵק) ֶלח‬ ְ ‫ֲׁשר ְּבבֵית יְהוָה‬ ֶ ‫ִּתנּו אֹתֹו ְלעֹׂשֵ י ה ְַּמלָאכָה א‬ ְ ‫ ‏ ְוי‬4 ELC ‫ְולַּבֹנִים ְולַּג ְֹדִרים‬

NG 4

‫ֶדק ַה ָּביִת‬ ֶ ‫ ְל ַחּזֵק ּב‬CMPL NG 2 v. 6b

(‫)ל ַחּזֵק‬ ְ ‫ ְ‏ו ִל ְקנֹות ע ִֵצים ְוא ְַבנֵי מ ְַחצֵב‬CMPL NG 3

v. 7

‫ ְל ַחּזֵק אֶת־ ַה ָּביִת‬CMPL ‫ׁשב ִא ָּתם ַה ֶּכסֶף‬ ֵ ‫ א ְַך לֹא־י ֵ ָח‬2 ELC ‫ָדם‬ ָ ‫ִּתן עַל־י‬ ָ ‫ ַהּנ‬SUB ‫ ִּכי ֶבאֱמּונָה הֵם ע ֹ ִׂשים‬1 ELC

v. 8

‫ַל־ׁשפָן הַּסֹפֵר‬ ָ ‫ וַּיֹאמֶר ִח ְל ִקּיָהּו הַּכֹהֵן ַהּגָדֹול ע‬2 ELC

NG 2 NG 2

‫ָאתי ְּבבֵית יְהוָה‬ ִ ‫ַּתֹורה ָמצ‬ ָ ‫ ֵספֶר ה‬2 ELC

NG 2 NG 2

‫ֶל־ׁשפָן‬ ָ ‫ ַוּיִּתֵ ן ִח ְל ִקּיָה אֶת־ ַה ֵּספֶר א‬3 ELC ‫ִק ָראֵהּו‬ ְ ‫ ַוּי‬0 ELC

Table 2b.  Discourse Profile of 2 Kings 22:1–14 (All 45 Clauses) Phenomenon

Number Percentage

Phenomenon

0–1 ELC

12

26.67

mean noun pairs

2+ ELC

17

37.78

3+ ELC

subordination

16

35.55

complex hypotaxis

Number Percentage 51.5

114.44

8.0

17.78

10.0

22.22

The discourse profile for this tale indicates a striking preference for short clauses in parataxis (predicate only or 1 ELC), as against a clear disinclination for long clauses, hypotactic constructions, and long noun groups. More than

310

Frank H. Polak Table 3.  The Profiles Compared

Text

Class Number Clauses 0–1 ELC 2+ ELC

All Hypotaxis

Noun Pairs

Complex Hypotaxis 3+ ELC

Ebenezer

66

53.03

40.91

6.06

52.27



12.12

Josiah

45

26.67

37.78

35.55

114.44

22.22

17.78

50% of the text consists of paractactic short clauses. This is the lean, brisk style that, according to my analysis, is characteristic of the Medial corpus. On the other hand, the tale of Josiah’s cleansing of the temple (2  Kgs 22:1–14; tables 2a–b) stands out because of the frequency of its long clauses, subordinate clauses, and long noun groups. The discourse profile indicates that this narrative is characterized by a preference for long, elaborate clauses, hypotaxis (often in complex constructions), and long noun groups. Indeed, the number of mean noun pairs (half the number of grouped nouns) exceeds the number of clauses. This fact is expressed by the high percentage of pairs of this sort (114.44 %), indicating that, in the mean, every clause contains one noun pair and that more than 10% of all clauses contain more than one pair. This text exemplifies the intricate, elaborate style, which, as I will show later, is characteristic of the Judean corpus and the Achaemenid corpus. Before discussing the discourse structure involved in these styles, I first treat the extent of the differences. 2.4. What Discourse Typology Reveals The radical difference between the lean, brisk style of the Ebenezer tale and the intricate, elaborate style of the Josiah narrative is highlighted in table 3. The difference between the Ebenezer and Josiah text profiles reveals a broad stylistic differentiation that is evidenced by more than 200 analyzed narrative units, almost all of them containing more than 30 clauses. 13 Previous investigations have shown that the lean, brisk style predominates in the Medial corpus (120 samples), whereas the elaborate, intricate style is characteristic of the Judean and Achaemenid corpuses (50 and 30 samples, respectively; Polak 1998, 2003, 2006b, 2010). In order to learn whether or not the linguistic style is a function of a narrative’s theme, I investigated a sample of 87 narrative units with at least 30 clauses, in 4 thematic groups: Cultic/Festive Meal; Anointment/Public Honor; Battle Account; Prophetic Appearance/Religious Discourse. Within the Medial corpus, I recognize 2 subtypes, mainly because of small differences in 13.  In previous studies, I have published the full analysis or the summarized discourse profile of 132 samples: 75 samples in the Medial corpus, 38 in the Judean corpus, and 19 in the Achaemenid corpus. In addition, I use the analyses of 56 samples from the Medial corpus, 11 samples from the Judean corpus, and 10 from the Achaemenid corpus that have not yet been published.

311

Sociocultural Background of Biblical Narrative Table 4.  General Discourse Typology Corpus

Class Number Clauses

0–1 ELC

All Noun Complex 2+ ELC Hypotaxis Groups Hypotaxis 3.59

3+ ELC

Medial (1)

25

55.57

33.53

11.14

41.85

8.44

Medial (2)

21

44.17

41.73

14.93

42.31

6.44

7.75

Judean

21

31.72

41.88

28.91

89.72

14.23

14.78

Achaemenid

17

30.52

43.12

29.56

86.59

18.53

12.86

a.  This table, in which the figures should not add up to 100%, shows the median for the classes involved. In statistics, the median is defined as the middle value of the set; half of the data appear above the median and half below the median. Thus, the median value of 1, 17, 22, 36, and 45 is 22. In a set that consists of an even number of values, the median is the arithmetic mean of the two middle values.

the frequency of short clauses. These findings have already been published in detail elsewhere (Polak 2006b: 134–48; 2010: 54–59, 70–73). 14 Here I limit myself to a general overview of all corpuses (see table 4). By and large, the resulting picture is quite clear. In all 4 thematic groups within the Judean and the Achaemenid corpuses, less than 1⁄3 of the text typically consists of short clauses (0–1 ELC). Almost 1⁄3 of the clauses are subordinate to the main clause, and almost every clause includes a noun group. Complex hypotaxis and long clauses are found in ca. 15% of all clauses. The frequency of infinitive clauses in the Deuteronomistic style, recently noted by Lohfink and Römer (see above, §2.1), is but one of the characteristics of this style. In the Medial corpus, hypotaxis (ca. 13%) appears far less frequently than in the other 2 corpuses, and noun groups are found in less than 1/2 of all clauses. Less than 10% of the text consists of long, elaborate clauses, and complex hypotaxis is even less frequent. With regard to short clauses, I note a certain differentiation between 2 subtypes. Whereas in Medial 1 the percentage of short clauses hovers around 50%, in Medial 2 this percentage is slightly lower and ranges from ±40% to 45%. These figures show that the Medial corpus is characterized by a lean, brisk style that is not encumbered by the intricacies characteristic of the elaborate style of the Judean and Achaemenid corpuses and that style is not determined by the thematic context of the narrative in which it is employed. These findings raise the question how to account for the radical differentiation between these two styles. My proposals for answering this question take 14.  Here and there, individual samples contain an exceptional feature (Polak 2006: 136, 146; 2010: 56–58), but these cases are far and few and do not distract from the general picture. For example, in the book of Samuel the lean, brisk style appears in 82% of the sample, whereas these partial exceptions affect no more than 18% of the units.

312

Frank H. Polak

into account linguistic typology and social context. Although these categories seem worlds apart, both are associated with the modes of communication that are connected with the varieties of discourse structure. This issue will be broached in §2.5 of my study. 2.5.  From Discourse Profile to Typology 2.5.1  Aspects of Discourse Structure Generally speaking, the extensive, elaborate clauses, the long noun groups, and the high number of subordinate clauses of narratives in the intricate style, such as the Josiah narrative, pack a great deal of information into single, complex sentences. Hence this style is characterized by the density of its lexical information (Halliday 1989: 61–67; Polak 2006a: 299–301). By contrast, in the lean, brisk style of the Ebenezer tale the information is spread out over a relatively large number of consecutive clauses. As a result, this style stands out because of its lexical sparsity (Halliday 1989: 79–81; Polak 2006a: 301–3). A particularly salient characteristic of the lean, brisk style is the frequency of sequences of short clauses, both in quoted discourse and in the narrator’s domain: 15 ‫ּׁשעֲנּו ּתַ חַת ָהעֵץ‬ ָ ‫ ְו ִה‬/‫ ְורַ חֲצּו רַ ְגלֵיכֶם‬/‫יֻּקַ ח־נָא ְמעַט־ ַמיִם‬ ‫ ַאחַר ּתַ עֲבֹרּו‬/‫ ְו ַסעֲדּו ִל ְּבכֶם‬/‫ֶקחָה פַת־ ֶלחֶם‬ ְ ‫ְוא‬ Let a little water be brought;/ bathe your feet/ and recline under the tree./ And let me fetch a morsel of bread,/ and refresh yourselves;/ then go on. (Gen 18:4–5)

(6)  ​

‫ֲׂשי עֻגֹות‬ ִ ‫ ַוע‬/‫לּוׁשי‬ ִ /‫ׁשלֹׁש ְס ִאים ֶקמַח סֹלֶת‬ ְ /‫ֲרי‬ ִ‫‏ ַמה‬ (7)  ​ Quick,/ three seahs of choice flour!/ Knead /and make cakes! (Gen 18:6) In the Ebenezer tale, we note the sequence in v. 14: /‫ מֶה קֹול ֶההָמֹון ַהּזֶה‬/‫ וַּיֹאמֶר‬/‫ָקה‬ ָ ‫‏‏ ַוּיִׁשמַע ע ִֵלי אֶת־קֹול ה ְַּצע‬ ‫ ַוּיַּגֵד ְלעֵלי‬/‫ ַוּיָבֹא‬/‫ָאיׁש ִמהַר‬ ִ ‫ְוה‬ Eli heard the tumultuous outcry/ and asked/ “What is the meaning of this tumult?”/ The man hurried/ and came/ and told Eli. (1 Sam 4:14)

(8)  ​

2.5.2.  Discourse and Information Structure 16 The two styles use different syntactic means to build a text. In this section I undertake to respond to the following questions: what kind of discourse is 15. See Polak 2006b: 301–3; 2006b: 142–44. 150; 2010: 38, 51–52; So also, for example, in exposition: 1 Sam 1:1–2; 9:1–2a; 25:2–3; in discourse: Gen 32:27–30; 2 Sam 1:6b–10a; 9:2; 20:16; and in action narrative: 1 Sam 16:12; 17:4–8a; 17:51; 2 Sam 11:17–18. 16. By information structure I mean, following Erteschik-Shir (2007), the “Syntax– Discourse Interface,” but I take this term in a very broad sense, for we are not to limit

Sociocultural Background of Biblical Narrative

313

made possible by the lean brisk style as opposed to the intricate, elaborate style? How is the information structured when the narrator uses the intricate, elaborate style? How is the information structure affected by the use of the lean, brisk style? As we will see, the answers to these questions, presented below, affect our understanding of the communicative situation and, thereby, of the implied sociocultural context of the text. 1.  The inclination toward short clauses favors a particular status for indications of, for example, place and time and even an indirect object (Eskhult 1990: 116–18). In the elaborate style, the narrator may specify subject, object, indirect object, place, and time all in the same clause: ‫ַל־ּפי יְהוָה‬ ִ ‫ׁשה ֶעבֶד־יְהוָה ְּב ֶארֶץ מֹואָב ע‬ ֶ ֹ ‫ׁשם מ‬ ָ ‫ַוּיָמָת‬ (9)  ​ So (1) Moses the servant of Yhwh died (2) there, (3) in the land of Moab, (4) at the command of Yhwh. (Deut 34:5) By contrast, the Medial corpus is characterized by the frequency of short core clauses (with subject and/or object), whereas concomitant data (“satellites”) are placed in different clauses, such as the setting in the opening of a narrative, or as additional information in a circumstantial clause: ‫ְהּודה‬ ָ ‫ֵׁשת־י‬ ֶ ‫ַת־ׁשּוע א‬ ַ ‫ ו ַָּתמָת ּב‬/‫ָמים‬ ִ ‫‏ ַוּי ְִרּבּו ַהּי‬ (10)  ​ A long time afterward/ Shua’s daughter, the wife of Judah, died. (Gen 38:12) The indication of time “a long time afterward” is expressed in a separate clause, ‫ָמים‬ ִ ‫ ַוּי ְִרּבּו ַהּי‬, which precedes the information concerning Shua’s daughter’s death. The note on Naboth’s execution states the place and means in separate clauses: ‫ ַוּיָמֹת‬/‫א ָבנִים‬ ֲ ‫ִס ְקלֻהּו ָב‬ ְ ‫ ַוּי‬/‫ָעיר‬ ִ ‫‏וַּי ִֹצאֻהּו ִמחּוץ ל‬ (11)  ​ Then they took him outside the town/ and stoned him/ and he died. (1 Kgs 21:13) A circumstantial clause describes Esau’s fatigue: ‫ ְוהּוא ָעי ֵף‬/‫ַּׂשדֶה‬ ָ ‫ֵׂשו ִמן־ה‬ ָ ‫ ַוּיָבֹא ע‬/‫‏ ַוּיָזֶד יַעֲק ֹב נָזִיד‬ (12)  ​ Once when Jacob was cooking a stew,/ Esau came in from the open,/ famished. (Gen 25:29) While the elaborate style systematically conceptualizes and specifies the relationship between the various pieces of information, the lean, brisk style tends to do so sparingly and often prefers to leave the relationship implicit.  17 the interface of Biblical Hebrew syntax and discourse to the topic-focus construction as in Erteschik-Shir’s study. 17.  In order to comply with the grammatical norms and sensibilities of twentieth century American English, modern English translations render the separate Hebrew clauses by hypo-

314

Frank H. Polak

2.  A second aspect of the preference for short clauses is the inclination to spread information pertaining to one action over two clauses: a pre-clause and a core clause (GKC §120a, d, g; Longacre 1989: 71–72; Polak 2006a: 288–97; 2009: 166–67, 187–97): /‫ָקר רָץ א ְַב ָרהָם‬ ָ ‫ְ‏‏ואֶל־ ַהּב‬ ‫ ַו ְי ַמהֵר ַלעֲׂשֹות אֹתֹו‬/‫ ַוּיִּתֵ ן אֶל־ ַהּנַעַר‬/‫ָקר רַ ְך וָטֹוב‬ ָ ‫ַוּיִּקַ ח ּבֶן־ּב‬ Then Abraham ran to the herd/ took a calf, tender and choice,/ and gave it to a “boy,”/ who hastened to prepare it. (Gen 18:7)

(13)  ​

Following the indication of the agent, Abraham, and the place of action in the first clause, the second clause states the object involved in the main action: the handing over of the calf to the slave-boy. The second is the core clause for which the preceding clause, ‘took a calf’, prepares the background. One notes Abraham’s request, ‘Let a little water be brough; bathe your feet’ (v. 4): the water is the instrument that enables the guests to wash their feet. By the same token, bread is what will refresh them (v. 5: ‘let me fetch a morsel of bread,/ and refresh yourselves’). In all these cases, the information is spread over two clauses in a biclausal construction. A construction of this type is also used in the epic formula that describes Abraham’s perception of his guests (v. 2:‫ִּׂשא‬ ָ ‫‏ ַוּי‬ ‫ָבים ָעלָיו‬ ִ ‫ָׁשים ִנּצ‬ ִ ‫אנ‬ ֲ ‫ׁשה‬ ָ ֹ ‫ׁשל‬ ְ ‫עֵינָיו ַוּי ְַרא ְו ִהּנֵה‬, modernized as ‘looking up, he saw three men standing near him’, njpsv). Motion verbs can be used to introduce the main action: ְ ‫ׁשכָב ַוּיֵל‬ ‫ִׁשּכָב‬ ְ ‫ֶך ַוּי‬ ְ ‫ָאתי ׁשּוב‬ ִ ‫א־קר‬ ָ ֹ ‫וַּיֹאמֶר ל‬ (14)  ​ But he replied, “I didn’t call you; go back to sleep.” So he went and lay down. (1 Sam 3:5) ְ ‫ָקם ַוּיֵל‬  ַ‫חז ֵק ְּבקַ ְרנֹות ה ִַּמ ְזּבֵח‬ ֲ ַ‫ֶך ַוּי‬ ָ ‫ׁשלֹמֹה ַוּי‬ ְ ‫ַואֲד ִֹנּיָהּו יָרֵא ִמ ְּפנֵי‬ (15)  ​ Adonijah, in fear of Solomon, went at once and grasped the horns of the altar. (1 Kgs 1:50) In these cases, the verb ‫ הלך‬serves to indicate the inception of the specific activity indicated by the following verbs or clause (Polak 2009: 189–91). Constructions of this type are largely characteristic of the Medial corpus (Polak 2009: 168–73). 3.  In the lean, brisk style, the exact identification of the participants is mainly limited to the setting, after which their identity is often assumed and does not require renewed specification, even when the subject changes during the interaction. The subjects are underdetermined. 18 The account of the tactic clauses (“Once when Jacob”) or circumstantial modifiers (“famished”); other translations into modern languages use similar techniques. 18.  Underdetermination is studied by Longacre 1989: 162–74; and de Regt 1999: 28–32, 43–48. I have discussed pragmatic aspects (Polak 2002: 1–12, 97–102).

Sociocultural Background of Biblical Narrative

315

negotiations between Abraham and Abimelech exemplifies such underdetermination. The sequence opens with the king’s question: “And Abimelech said to Abraham, ‘What mean these seven ewes that you have set apart?’ He said (‫)ויאמר‬, ‘You are to accept these seven ewes from me’” (Gen 21:29–30). The shift of speakers is not marked by specification of the subject, and thus the identification of the speaker is left to the individual reader (or the audience), who must understand what is happening in light of the interaction between the participants. 19 In the elaborate style, by contrast, the basic preference is for a clear presentation and explication of the individuals involved in the action, event, or situation. In the Esther tale, almost all actions are marked by indication of the acting subject. 20 4.  Specification and clear identification are also involved in the contrast between parataxis and hypotaxis. Complex sentences specify the logical, temporal, and spatial relationship between the clauses. By contrast, paratactic clause chaining is dominated by the free interplay between the clause complex of the setting, clauses indicating the action/event sequence, and circumstantial clauses. The individual reader (or the audience) must figure out the logical connection in accordance with the context. In short, the elaborate, intricate style makes explicit the relationships between the participants, specifies the referents, and makes logical connections explicit. To a large extent, the lean, brisk style leaves these matters to the reader (or the audience) and his/her understanding of the situation. In other words, the difference between the lean, brisk and the intricate, elaborate style pertains to the communicative situation. This aspect of the issue at hand is brought out by Koch and Oesterreicher (1994: 587–89), who describe the lean, brisk style as “proximity language,” whereas elaborate diction is connected with distance. 2.6. From Discourse Typology to Communicative Situation and Sociocultural Background 2.6.1.  Speaking, Writing, and the Communicative Situation In this section, I examine the communicative situations implied by the lean, brisk and the intricate, elaborate style. Typologically, the differences between these two styles match the cross-linguistic differentiation between oral and written discourse, analyzed in particular by Wallace Chafe, Michael Halliday, 19.  Some additional examples: Gen 18:1–3, 7–10, 27–32; 33:4–8, 11–13; 46:29; Judg 15:19; 1 Sam 15:27; 16:11–12, 21; 2 Sam 20:10; 2 Kgs 5:14–16, 19, 21–22, 23–26; 6:6–7; 8:14–15; 9:4–6, 9:19, 24–25. In some cases the change is marked by the shift from singular to plural or from masculine to feminine. 20.  For instance: Esth 4:1a, 5–10, 12–17; 6:3–7, 10–13; 7:1–9 (including the unique mention of Harbonah). In the terms of Miller (1996: 250–52) one notes the mention of the ְ ‫מל‬ subject in the second turn, or second and third round in 4:7, 10, 13, 15, 17; 6:4–5 (‫ֶך‬ ּ ֶ ‫נַעֲרֵ י ַה‬.‫‏‬ repeating 6:3b); 6:7 (Haman, repeating v. 6a,b); 6:10–11a (as against vv. 11bc); 7:5–6, 8.

316

Frank H. Polak

and their respective schools. 21 The communicative modes of writing and speaking involve different situations and different modes of operation. The technique of writing allows for planning, rereading, and correction. Speaking is always instantaneous; and even when planned, it does not allow for correction (“editing”). 22 According to this analysis, which is based on wide crosscultural and cross-linguistic research, the discourse profile of the elaborate style fits the characteristics of formal, written language. 23 It thus represents the work of the trained scribe, the product of the scribal chancery and its education. Its complexity and precision are more than just a stylistic phenomenon related to individual style. These features reflect the exactitude of a scribe who has learned to draft detailed official records and legal contracts. Such documents must be able to bridge distances in space and time with a minimum of misunderstandings or ambiguities, and therefore everything relevant is explicitly and exactly indicated. Thus, Biblical Hebrew texts couched in the intricate, elaborate style presuppose an advanced scribal education and administrative culture, as evidenced by the Siloam inscription and the ostraca from Jabneh/ Jamnia, Lachish, and Arad (see §2.6.6 below). In contrast, the characteristics of the lean, brisk style are close to the crosslinguistic characterization of unplanned, spontaneous, spoken language. 24 And where written texts can bridge the distance in space and time, spoken discourse presupposes the co-presence, here and now, of speaker and addressee, who hear one another, share a common background for the interaction, can ask for clarification when necessary, and are able to clear up misunderstandings. This is exactly the sort of situation that is presupposed by the discourse structure of texts in the lean, brisk style, as analyzed above (§2.5.2). On the basis of this typology, I conclude that the lean, brisk style points to an oral substratum or to 21.  See in particular Behaghel 1900; Chafe 1982, 1985, 1994; Halliday 1989; Miller and Weinert 1998. 22.  The speaker can correct himself (“self-repair”), but his correction forms an additional utterance in the spoken interaction, whereas in writing a correction of the erroneous expression replaces the latter. What in speaking is a succession of two utterances—error and correction—is in writing matched by a single, corrected expression. 23.  See in particular Katzenberger 2004; Ochs 1979; Pawley and Syder 1983; Chafe and Danielewicz 1987; Biber 1995; Miller and Weinert 1998; Biber and Conrad 2001; 2009: 226–34, 253–69; and see the surveys of Tannen 1982; Chafe and Tannen 1987. Tannen (1982; 1985; Chafe and Tannen 1987) shows that the relationship between oral and written language is not only connected to syntactic structure but also to personal involvement (see also Besnier 1988). However, these considerations do not invalidate the structural component. 24.  The precise term “spontaneous spoken language” (Miller and Weinert 1998: 14–15, 23, 38) is crucial, because spoken language can be intricate when used by persons with experience of formal language, such as clergymen, jurists, and other academics, including students. Hence, the objections of Biber (1988) to the differentiation, on the basis of a corpus of discussions between students, is rejected by Miller and Weinert (1998: 19–21) and relativized by Biber and Conrad (2009: 261–64).

Sociocultural Background of Biblical Narrative

317

close contact with oral literature, and also to strong and many-faceted interactions with performances by oral poets and narrators (Polak 2006a: 290–304). 2.6.2.  The Written, the Oral, and Language Usage in the Medial Corpus My inference of an oral background behind texts in the lean, brisk style has been countered by the argument that style is a matter of free, literary choice and that gifted narrators can always choose the style that fits their narrative (Young, Rezetko, and Ehrensvärd 2008: 2.83; Niditch 2010: 91). In this view, all syntactic-stylistic variation is to be subsumed under the heading of literary, rhetorical design. This hypothesis, however, as illustrated above in §2.4, is contradicted by my systematic analysis of 87 narratives in 4 theme groups (Polak 2006b: 135– 36, 143–48; 2010: 55–62, 72–74). The conclusion of that analysis is twofold: 1.  Narratives from different corpuses preserve the discourse profile of the corpus to which each tale belongs, even when they do not share the same thematic content. 2.  Conversely, narratives of different thematic content represent similar typology when they belong to the same corpus.

It is therefore impossible to explain these differences in discourse profile as a matter of free literary design and rhetorical shaping controlled or influenced by the topic or theme of the narrative. Additional factors from outside the literary framework other than stylistic choice must be considered in order to explain this phenomenon. Sociolinguistic typology provides a link between language usage and extra-literary, sociocultural conditions (Fennell and Bennett 1991). The correspondence between the linguistic characteristics of spoken and written discourse and the discourse typologies of biblical narrative points to the distinction between the oral and the scribal media as a main factor. This conclusion is supported by various sociocultural aspects of biblical narrative. 2.6.3.  Linguistic Typology and Sociocultural Conditions The importance of the sociocultural sphere is indicated by the analysis of Josephine Miles concerning the distribution of nouns, verbs, and adjectives in English poetry over the centuries (for prose, she adds connectives, including conjunctions and prepositions). Miles recognizes a “clausal” or “predicative” style, characterized by a relative preponderance of verbs; a “classical” style, which is dominated by nouns; and a “sublime” style, characterized by the prominence of adjectives (Miles 1964: 2–18, 216–22). This distinction enables her to discern a “predicative” style in most medieval literature, and, in particular, in authentic, orally based ballad poetry (Miles 1964: 100–107, 517–20). Significantly, the ballad style was considered too “low” for literary purposes. When writing poets composed literary ballads by imitation, they did not preserve the stylistic characteristics of the oral ballad (Miles 1964: 18–19,

318

Frank H. Polak

103–8). The predicative predominance decreased with time and the advancement of learning between the sixteenth and the nineteenth century (Miles 1964: 520–22; 1967: 22–61). This analysis provides an important example for the development from an orally based style to a writerly style along the centuries, hand in hand with the advancement of education and sociocultural knowledge. A similar picture is manifested by the development of the style of Old English law. Schwyter (1998) describes a gradual growth of legal formulation from the plain style of the first legal corpora, which were entirely dependent on unwritten law, to the development of an elaborate legal style, which manifests some of the intricacies known from Latin legal language. By and large, then, discourse typology goes hand in hand with sociocultural conditions and development. 25 2.6.4.  From Sociocultural Conditions to Communicative Situation: The Scribal Roots of the Achaemenid Corpus An important aspect of the sociocultural context of biblical narrative is the theme of “writing.” In the Achaemenid corpus, references to writing and written documents are widespread. 26 A particularly notable point is the reference to the message that Hiram sent to Solomon in writing as described by the Achemenid-era Chronicler: ‘And Chiram, king of Tyre, said in writing ְ ‫ֶל־שלֹמֹה ֶמל‬ and dispatched to Salomon’ 27 (2 Chr 2:10, ‫ֶך־צֹר ִּב ְכ ָתב‬ ְׁ ‫חּורם א‬ ָ ‫וַּיֹאמֶר‬ ‫ּשלַח‬ ְׁ ‫) ַוִי‬. On many occasions, the narrative highlights the scribe’s function in the formulation and execution of royal orders. 28 Notably, when the narrator wishes to indicate Ezra’s role in the construction of Judean religious practice, 25.  Biber and Finegan (1989) show that, in the course of the history of English literature, the development was from prose that tended toward an increasingly complex, writerly style (in the seventeenth and eighteenth centuries) to prose with a style closer to the immediacy of oral language (in the nineteenth century). They acknowledge, however, that this development was related to esthetic preferences and the rise of realism rather than to language usage as such (Biber and Finegan 1989: 516), but it would be extremely difficult to ascribe ideologically inspired realism to a piece of biblical narrative. Moreover, their criteria relate to abstract qualities rather than to concrete syntactic data. 26.  Forms of the root ‫כתב‬: Esth 2:23, and passim; Ezra 3:2; 8:34; Neh 6:6; 7:5; 8:14–15; 13:1; 1 Chr 16:40; 24:6; 29:29; 2 Chr 23:18, and passim; the noun ‫ּכ ָתב‬: ְ Esth 1:22, and passim; 1 Chr 28:19; 2 Chr 2:10; 35:4; the noun ‫א ֶּגרֶת‬: ִ Esth 9:26, 29; Neh 2:7–9; 6:5, 17, 19; 2 Chr 30:1, 6; the noun ‫ ֵספֶר‬: Esth 1:22, and passim; Dan 1:4, 7; Neh 7:5, and passim. 27.  So also the LXX. The quotation is from the nets. The njps packs all the information into one sentence and thus succeeds in getting rid of ‘said’: ‘Huram, king of Tyre, sent Solomon this written message in reply’. 28. Esth 3:12; 8:9; Neh 13:13; 1 Chr 24:6; 27:32; 2 Chr 24:11; 26:11; 34:13. In order to avoid misunderstandings, let me add that in Samuel–Kings the reference to the sōpēr is to one of the highest royal ministers (2 Sam 8:17// 20:25; 1 Kgs 4:3; 2 Kgs 12:11; 18:18, 37; 19:2; 22:3, 8–10, 12; 25:19; Isa 36:3, 22; 37:2). The mention of ‫שבֶט סֹפֵר‬ ֵ ׁ ‫‏מ ֹ ְׁש ִכים ְּב‬in the Song of Deborah (‘such as hold the marshal’s staff’, Judg 5:14) is to be explained as an honorific title for tribal leaders, because the ‘staff’ doesn’t fit the scribe. One is reminded of the role

Sociocultural Background of Biblical Narrative

319

he uses the epithet “the scribe” (Ezra 7:11–12; 7:6; Neh 8:1, 4, 9, 13; 12:36) in addition to his priestly title (Ezra 10:10, 16; Neh 8:2). This usage testifies to the high status of the scribes in the sociocultural milieu that gave rise to the writings in the Achaemenid corpus. Accordingly, the scribal roots of this corpus are unmistakable. The scribal background of this corpus is also evidenced by the Persian and Aramaic loanwords in the administrative field (Rosenthal 1963: 58–59; Polak 2005: 596–98), such as, for example, ‫‘ גזבר‬treasurer’ (Ezra 1:8); ‫‘ גנזים‬treasuries’ (Esth 3:4; 7:9); 29 ‫‘ גנזך‬treasury’ (1 Chr 28:11); ‫‘ דת‬instruction’ 30 (Esth 1:8 and passim; Ezra 8:36); ‫‘ ִפ ְתגָם‬decision, announcement’ (e.g., Qoh 8:11; Esth 1:20; Ezra 4:17; 6:11). 31 The term ‫‘ פרדס‬park’ (Neh 2:8) is also related to the administration, because it is used to indicate the royal forests, which served as a source for wood. 32 Among the borrowings from Aramaic, one notes, for instance, the verbs ‫‘ פקד‬to order’ (as against ‫;)צּוָה‬ ִ ‫‘ ִקּבֵל‬to receive’ (as against ‫‘ אסף ;)לקח‬to gather’ (as against ‫ ;כנס‬Polak 2006b: 122). The use of ‫לשון‬‎ (//‫גוי‬, Isa 66:18; Zech 8:23; Aramaic ‫ּׁשן‬ ָ ‫)ל‬ ִ  33 to indicate an ethnic group by its language likewise belongs to the administrative level. Aramaic influence on the Hebrew of the Achaemenid corpus is completely explainable by the position of Official Aramaic as the language of the imperial administration in Western Asia (and in general), as indicated, for example, by of the spr in the rule of the city of Ugarit and the Akkadian title šāpiru ‘overseer/governor’ (CAD S 453–58). 29.  Different from ‫גנזי‬, Ezek 27:24, which seems related to Ethiopian (HALOT, s.v.). 30.  Young, Rezetko, and Ehrensvärd (2008: 1.303–4) point to the Qere, ‫( אש דת‬Deut 33:2). However, the translation ‘law’ does not really fit the context, which evokes a theophany. Hence the suggestion to interpret ‫( דת‬Steiner and Leiman 2009) as a verbal form from the root ‫ דאי‬remains quite attractive, all the more so, because it is reflected in the Palestinian Targum. 31.  Young, Rezetko, and Ehrensvärd (2008: 1.302) doubt the derivation of ‫ ִפ ְתגָם‬from Old Persian *pati-gāma because this lexeme has been reconstructed from Middle Persian; Young (1993: 71) considers it to be derived from an ancient Indo-European lexeme such as Greek φθέγμα ‘sound, voice, utterance’. However, the reconstructed form of the Old Persian (Seow 1996: 650) is supported by a loanword in Elamite (Tavernier 2007 :410, noting Middle Persian patigam). 32.  Notably, the term pardēsu was already in use in Neo-Babylonian (CAD P 182). This borrowing is probably related to the Babylonian import of wood from the Iranian mountains after the fall of Elam. Like Seow (1996: 649–50), Young, Rezetko, and Ehrensvärd (2008:1.301–2) think that the /s/ in ‫ פרדס‬is irregular in comparison with the supposed Old Persian lexeme *paridaiḏa (par-te-taš in the Persian Fortification Tablets). However, the /s/ is reflected by the Neo-Babylonian form and the Greek word παράδεισος, used by Xenophon, who would have heard the authentic Persian pronunciation around 404–401 (Chantraine 1968–77: 857). 33.  ‫‏כל־עממיא אמיא ולׁשני‬: Dan 3:4 and passim (Polak 2005: 589–90; 2006b: 117; and see Stolper 1984: 310). Akkadian lišānu with the meaning ‘speech community, ethnolinguistic group’, already attested in the Assyrian period, is used in many inscriptions from the Persian era to indicate the extent of the Empire.

320

Frank H. Polak

the contracts and bullas found in Wādi Dāliyeh, at Tell el-Qôm (Makkedah), and at Jericho, and by the documents from Elephantine (Polak 2005: 591–92; 2006b: 122–23). 34 In this context, I am assuming that the contemporary language of the provincial chancery of Jerusalem was also Aramaic, and thus all official documents, such as contracts, were written in Aramaic, as they were in Samaria (witness the findings at Wādi Dāliyeh). In consequence, Aramaic rather than Hebrew was the preferred language for scribal education: the children and pupils of the scribes had to learn the official language of the administration in order to have a career. 35 Thus the scribes and other literati imbibed Aramaic lexical items and language patterns from their very youth. Moreover, because of its widespread use in the administration, Aramaic became a prestige language that everyone who had business with the authorities needed to be able to understand and use. In this way, Aramaic entered the spoken language of the property-owning classes. I infer, then, that Hebrew-Aramaic bilingualism was a common societal phenomenon in the Neo-Babylonian/Persian era rather than a limited characteristic of some officials, as was true around 700 b.c.e. (2 Kgs 18:26; Schniedewind 2004: 165–82; 2006; Polak 2005: 591–606). 2.6.5.  The Judean Corpus and the Royal Chancery Numerous scholars have already observed the many references to the technique of writing in the books of Joshua and Kings and in the Jeremiah Vita, where the role of Baruch “the scribe” is mentioned frequently. The connection between this corpus and the expansion of scribal knowledge in the seventh century has also been widely documented (Niditch 1996: 95–105; Schniedewind 2004: 110–14, 134–36; Schaper 2004: 103–5; 2005: 329–35; Leuchter 2010: 124–25). 36 The epigraphic finds from the late seventh and early sixth century b.c.e., the ostraca from Lachish, Arad, and Jabneh/Jamnia, attest the intricate, elaborate style of the Judean corpus in all its aspects (Polak 1998: 103–4; 2006b: 137–38). 37 An additional aspect of scribal expertise that is not taken 34.  See the articles by Sokoloff and Pat-El in this volume. According to Young, Rezetko, and Ehrensvärd (2008: 1.215–17, 293–303), it is impossible to exclude the presence of Iranian and Aramaic loanwords or other language influences in the preexilic period. However, this view disregards the high number of administrative terms among the Aramaisms and Parsisms. Does anyone really believe that Iranian (or Aegean or Sanskrit) administrative terminology influenced Hebrew before the Persian era? 35.  Note the Aramaic proverbs in the collection of Ahiqar found at Elephantine; and the scribal excercise from Hellenistic Maresha (Eshel, Puech, and Kloner 2007). 36.  I concur with Millard (2010: 156–58) that the number of sherds found with writing on them is a matter of random accident; even before the seventh century, writing must have been a normal activity in non-royal, non-temple circles. But the seventh-century use of seals with personal name and patronymic only (seals in earlier periods always included an image) shows that at this juncture literacy had spread beyond the limits of small “elites” (Millard 2001). 37.  Young, Rezetko, and Ehrensvärd (2008: 1.143–71) maintain (following Young 2003: 292) that Epigraphic Hebrew contains many LBH features. However, the most convincing

Sociocultural Background of Biblical Narrative

321

into account by Young, Rezetko, and Ehrensvärd (2008: 2.22) is Deuteronomy’s systematic use of rhetorical patterns that are known from Neo-Assyrian texts (Weinfeld 1972: 14–37; Steymans 1995; Otto 1999:1–90). These patterns show that the authors behind Deuteronomy were acquainted, either directly or through Aramaic mediation, with some of the content of Assyrian treaties and political propaganda (Römer 2005: 83–90) and used the language for their own purposes (Morrow 2010). Even though this acquaintance may have been indirect, it shows that Judean literati had more than casual contact with some highly placed royal court scribes and the royal chancery. The intricate style of Deuteronomy, in particular, could not have arisen in the context of the provincial, Aramaic-oriented chancery postulated for the Neo-Babylonian/Persian era. The Judean corpus, then, must be positioned in the context of the royal chancery of the Judean Kingdom. 38 Scribal education in general, even in a private context, was ultimately connected with the standards of the scribal work place, either at the royal or the provincial/local chancery. 39 Scribal activity and education were dependent on the royal chancery, even if one particular scribe was not directly associated with the court. Association with the temple may be presumed for much of what has been assigned to the “Priestly” writings (including Leviticus 18; 23–24). Some strata of P texts, however, seem closer to the Medial corpus (Genesis 1; 17:1– 8; Leviticus 19–22; 25–27), whereas the typology of other texts, notably Numbers 35–36, aligns with the Achaemenid corpus. These data indicate that the Priestly corpus originated in oral tradition, was written down and elaborated on by the temple scribes under the late Judean monarchy, and was expanded in the Neo-Babylonian/Persian period. The present article, however, is not the appropriate place to develop this subject further.

piece of evidence, the alleged use of the verb ‫‘ רצה‬to be pleased by’ in the Mishnaic/Modern Hebrew meaning ‘to want’ (Young 2003: 292; Young, Rezetko, and Ehrensvärd 2008: 1.167) is dependent on reconstruction. What is presented is the graph ‫[צה‬. . .], preceded by an extensive vacat (approximately 11 spaces; Arad 40:6–7; Aharoni 1986: 72). Other features include the phrase ‫( במאתי[ם ו]אלף אמה‬Siloam, 5) in which the smaller number precedes the larger one, as in “Priestly” and LBH texts (Young 2003: 296; Young, Rezetko, and Ehrensvärd 2008: 1.155). However, Polzin (1976: 95–96, 100–101, 112–13) has been unable to tie “P” language unequivocally to LBH, and Hurvitz (1982; 2000) has virtually disproven the late dating of “P.” On the contrary, the Siloam inscription shows that the order of the numerals is no argument for the late date of “P.” 38.  If the possible northern roots of Deuteronomy are taken into account, one must also think of scribal chanceries in the Kingdom of Israel. 39.  Young, Rezetko, and Ehrensvärd (2008: 2.97) believe that “scribes were drawn (as far as we know) from local people using their own local literary language.” Do the authors then deny the presence of scribal education or a supraregional administrative language? Do they recognize no differentation between literary and administrative language?

322

Frank H. Polak

2.6.6.  Writing, Oral Performance, and the Medial Corpus By contrast, in the Medial corpus, references or allusions to writing are not prominent. Writing is not mentioned in connection with the marriage agreement of Jacob and Laban (Gen 29:18–19, 25–27). Their border agreement (Gen 31:44–54) was based on an elaborate ceremony that included the setting up of a ‘pillar’ (a ‫)מצבה‬, solemnia verba, and a festive meal on the stone heap (‫)על הגל‬, but no writing whatsoever (Gandz 1935). Writing is represented as a divine activity or the work of the mediator (Exod 24:4; 34:27–28; 1 Sam 10:25). Human writing is represented, not as a normal, administrative activity, but as the work of hostile, cruel authorities, and was fraught with danger (“the oral mentality,” Judg 8:14; 2 Sam 11:14–15; Schaack 1998: 32–43, 55–64). With respect to the position of oral and written communication, then, the text world of the Medial corpus fits the stylistic profile of the narratives contained in it. Accordingly, I conclude that the narratives in the Medial corpus were set in writing in a sociocultural situation with closer connections to oral narrative than to the developed scribal chancery reflected by the Judean corpus. Niditch (1996: 78–88, 94–98) describes these connections in terms of the oral-written continuum. And indeed, when we limit our thinking to the frequency of language features only, or to narrative characteristics, we can place the oral and the written on one and the same scale. In this sense, Lord (1995: 222–37) speaks of “transitional texts,” written down but dependent on oral literature. Such texts, when written by poets who were “inside” the oral culture, can preserve many of the oral features. However, the terms literacy and orality represent human activity, rather than abstract qualities. Orality as such always implies the appearance of a performing poet or narrator before an audience (Zumthor 1984a: 38–40; Bäuml 1998: 248–49; Niles 1999: 79–119). As medievalists have often pointed out, even written texts such as the Chanson de Roland and the Nibelungenlied are intended for oral delivery by the performing artist (Zumthor 1984a: 30–36, 48–54, 69–106; 1984b: 70–77, 83–90; Bäuml 1984; 1986; 1993; 1998). These texts, then, represent a tertium quid and should be conceived of as voiced, even though this voice is not audible anymore. Oral features in medieval written texts must be viewed in this context (Niles 1999: 102–7, 127–29, 196–200). Foley (2010: 21) speaks of “voices from the past” that are now hidden from us. Hence, instead of the “rugged landscape” of the “oral-written continuum” and its many dimensions (Ranković 2010: 39), we should, with Amodio (2004: 128–32), speak of the “oral-literate nexus.” 3.  Stylistic Profile, Social Context, and the Periodization of Biblical Narrative 3.1.  The Medial Corpus and Its Sociocultural Milieu Thus, we see that the literary milieu that gave rise to the Medial corpus is distinct from that of the Judean/Achaemenid Corpus. The Medial corpus

Sociocultural Background of Biblical Narrative

323

represents a “transitional text” by insiders who were part and parcel of the oral culture (Lord 1995: 223–35), who witnessed the performances of the oral singers and narrators, had intimate knowledge of the style of performance, accepted this style as normative, and preferred the norms of the oral performance to those of the scribal chancery. They were more indebted to the oral arena than to scribal education. Thus the syntactic-stylistic difference between the Medial corpus and the Judean/Achaemenid corpus implies a difference in social context. If the latter is at home in the scribal chancery, the former is rooted in the oral world—a world that honored oral performance as a prestigious literary discourse and a matter of the “higher classes” no less than written literature in the scribal milieu. The attribution of narratives such as the tales of the patriarchs and the Samuel-Saul-David cycle to “high” social strata fits the mentality exhibited by these narratives. The patriarchs and matriarchs are depicted as the honored ancestors of the Israelite people—noble, wise men and women who knew how to persevere in adverse circumstances. Royalty are represented in court, in military action, and in cultic activity. The language of these tales, in all its sparsity, preserves reminiscences of epic poetry (for instance in the epic formulas, for which, see §2.5.2 above). Hence, I conclude that the Medial corpus originated in a milieu in which orality was considered prestigious and was closely allied to the “higher classes.” This sociocultural context was the period before the development of the royal bureaucracy as a center of power—that is to say, before the seventh century b.c.e. 3.2.  The Text World of the Medial Corpus and Its Social Context The world depicted in many of the narratives in the Medial corpus is quite different from the organized state with a strong administration that is found in Judah in the late eighth–early seventh century (Lipschits, Sergi, and Koch 2010; Schniedewind 2004: 68–75, 93–106). The Samuel-Saul-David narratives present a royal authority that is weak and open to challenge (2 Samuel 19–20) and thus an inchoate kingdom rather than the full-blown monarchy portrayed in 2 Kings (Lichtenberger-Schäfer 1996: 95–105; Niemann 1993: 51–53, 61–71). By the same token, the patriarchs, even when formally presented as owners of herds and flock (Gen 12:5, 16; 13:2; 24:35), are subject to ruin by drought and poverty (Gen 12:10; 26:1; 41:56; 43:1) and may even be fugitives, with physical strength as their only capital (29:10). Nevertheless, their behavior is both noble and wily (Gen 14:16–24; 32:4, 14–21; 33:8–15). This picture hardly fits the ancestors of a mighty kingdom. Accordingly, the text world evoked by the narratives in the Medial corpus fits the sociocultural milieu implied by the lean, brisk style that dominates this corpus. The main texts of the Medial corpus came into being before the development of a broad royal apparatus and thus preceded the Judean corpus, albeit with a certain amount of overlap.

324

Frank H. Polak

Consideration of the religious world leads to similar conclusions. The Medial corpus is rich with the residues of animistic, dynamistic, and magical beliefs that do not square with the religious notions of Deuteronomic monotheism (Polak 2003: 67–79). Compare, for example, the tale of the outpouring of the spirit on the elders (Numbers 11) with the rationalistic versions in the Deuteronomic recast (Deut 1:9–17) and the Jethro narrative (Exod 18:13–27). It suffices to recall the tales of Jacob’s struggle with the angel (Gen 32:25– 32) and the divine attack on Moses that was remedied by Zipporah’s magical knowledge (Exod 4:24–26). The mythical connotations of the tale of the revelation on Mount Sinai (Loewenstamm 1981: 173–89; Polak 1996: 129–38) contrast sharply with the rationalistic version in Deuteronomy 4–6, which imposes severe limitations on the concrete descriptions in the Exodus tale (Sommer 1999: 426–36). Reference to cultural clues, then, warrants the scholarly conclusion that the evidence points to the precedence of the Medial corpus. A third consideration relates to literary development. Many tales in the Medial corpus are matched or are continued by versions that have been recognized as later recasts or as reader reaction (Fortschreibung), formulated in the intricate, elaborate style. This is the relationship between the tale of Moses’ call (“Priestly” Exodus 6), in contrast to the burning bush in Exodus 3–4; of Isaac’s “Priestly” blessing of Jacob (Gen 28:1–10), as against the ensuing narrative; and of the tales of Gideon’s heroic exploits (Judges 7–8), as opposed to the Deuteronomistic summary (Judg 8:27–35). In the tale of the revelation at Mount Sinai, one notes the difference between the general tale and the divergent note of Exod 19:9. In table 5, I note the discourse profile of the narratives in the Medial corpus (the source tale) in comparison with the later recast (Judean or Achaemenid corpus). The source tale is noted at the left, whereas the corresponding recast is indented. Comparison of the sets of figures shows that the relationship between the lean, brisk diction and the elaborate, intricate style is not simply a relationship of continuity, gradual transition, or coexistence (see also Gumbrecht 1983; Assmann 1983). While the secondary recast (Judean or Achaemenid corpus) is couched in the elaborate style, the lean, brisk style is characteristic of the source text (Medial corpus). Thus, the exegetical differentiation between these corpora underlines the stylistic distinction between them. 4.  The Chancery Scribe and the Lean, Brisk Style 4.1.  The Social Position of the Scribe and His Style The preceding discussion shows that the lean, brisk style of the Medial corpus represents a period in which the oral literature was associated even with the highest classes and in which the royal bureaucracy was not yet widely developed. Therefore, I argue, the tales in this corpus should be set apart from

325

Sociocultural Background of Biblical Narrative Table 5.  Source and Later Recast (Fortschreibung) Text

Class

Jacob: Gen 28:10–22   Gen 28:1–9, “P” Moses: Exod 3:1–15   Exod 6:2–13;   6:26–7:6, “P”

0–1 ELC 45.61 22.58 43.66 25.00

All 2+ ELC Hypotaxis 28.07

26.32

41.94 38.02

54.39

35.48 19.72

47.37

Noun Pairs

83.88 55.63

27.63

70.39

Complex Hypotaxis 3+ ELC 1.75 35.48 8.45

52.59

32.76

14.65

73.26

6.03

Judg 7:9–22

45.58

20.59

5.88

73.53

1.47

Exod 19:3–8,10– 19; 20:18–21   Exod 19:9

37.50

45.23 —

45.83

33.33 —

16.67

13.10

104.17

37.50 —

2/ 50.00

14.47 9.48 13.23

4.17

6.82

9.68 5.63

19.74

Judg 7:23–8:25   Judg 8:27–35,  Dtr

12.28

20.83

10.71

3/ 75.00 1/ 25.00

the narratives in the Judean or Achaemenid corpus. In these corpuses, the intricate, elaborate style indicates a thorough scribal education, either at the royal chancery (the Judean corpus) or at the Aramaic-centered provincial chancery of Achaemenid Yehud. Is it possible, however, that scribes with a chancery education and expertise composed narratives that consistently maintained the lean, brisk style? I have to admit that some narratives from the seventh century or the Neo-Babylonian/ Achaemenid period indeed contain clear traces of the lean, brisk style or are even dominated by this style. This possibility is evidenced by some of the tales in the opening chapters of Deuteronomy and by the tale of Jonah. In the present section, I discuss some features of the discourse profile of these narratives. However, the a priori assumption of a general and systematic use of the lean, brisk style by scribes from the seventh century or the Neo-Babylonian/ Achaemenid era raises severe difficulties, in particular in the societal field.  40 Ultimately, this is a matter of prestige and power relations. An important aspect of style is the social position that is revealed by its usage (Eckert 2002: 123–26; Coupland 2007: 121–32, 146–63). Scribes who used the intricate, 40.  On bilingual authors in the modern period, such as Joseph Conrad, see provisionally my remarks in Polak 2010: 35.

326

Frank H. Polak

elaborate style positioned themselves, not only as experts, but also as persons with authority. In this connection, it is important to recall that the scribal chancery, the locus of the bureaucratic apparatus, was an important center of power, as indicated by the role of Shaphan the scribe in the “discovery” of the lawbook. The scribes who compiled and shaped the book of Kings had detailed knowledge of the data concerning the royal dynasties of Israel and Judah and contrived to synchronize them in an integrated dynastic chronicle (Halpern and Lemaire 2010). This example shows how in ancient Judea scribal expertise and power went hand in hand, much as in Assyria, Egypt, and Persia. Accordingly, scribes who used the elaborate style presented themselves as experts associated with the milieu of the scribal chancery and thus with the center of bureaucratic power. Indirectly, then, use of the scribal style implied a claim to the prestige and authority emanating from the power center. By contrast, in a world dominated by the royal chancery, a narrator using the lean, brisk style foregoes the claim to any connection with the bureaucratic power center. He presents himself as someone associated with social strata that are remote from the world of the chancery and the official administration—or, in other words, with the “lower” strata of society. It is difficult to imagine that a chancery scribe would adopt this approach. On the contrary, the style of Deuteronomy and of many sections associated with temple ritual show that the scribes of these periods only rarely deviated from the intricate, elaborate style of the chancery, as was only to be expected. A second aspect of the issue pertains to the psychology of language use. Systematic education in the scribal style inculcated scribes with the syntactic and lexical patterns that dominated stylistic formulations. The effect of underlying patterns is well known in connection with second-language acquisition. The morphosyntactic patterns of the first language continue to interfere with the second language (Gabryś-Barker 2008; Bolonyai 2009). Bilingual speakers suffer from similar problems (Miccio, Scheffner Hammer and Rodríguez 2009; Wi 2009); neurolinguistic influences have likewise been discussed (Kutas, Moreno, and Wicha 2009). This issue pertains to scribal education as well. Carr (2005: 159–61) highlights the memorization of texts that then turned the texts into literary templates. Of course, the teaching of intricate, elaborate stylistic patterns in scribal education must have had a similar effect. Hence, the assumption that the absorption of these patterns did not affect the literary style of the chancery scribe or the literati educated by the chancery and its scribes goes against the grain. 4.2.  Elements of the Lean, Brisk Style in the Judean/Achaemenid Corpus As I already noted, the Judean/Achaemenid corpus includes tales with features of the lean, brisk style. These passages merit detailed analysis. The open-

327

Sociocultural Background of Biblical Narrative Table 6.  The Jonah Tale Text

Class Number Clauses

0–1 ELC

2+ ELC

All Hypotaxis

Noun Complex Pairs Hypotaxis

3+ ELC 3.57

Jonah 1:1–16

84

52.38

32.14

15.48

24.40

7.14

Jonah 3

43

46.51

37.21

16.28

44.19

9.30

6.98

Jonah 4

51

27.45

41.18

31.37

30.39

15.69

11.76

ing chapters of Deuteronomy are set off by a distinctly oral style. The divine instructions to Moses represent many of the features of the lean, brisk style, including the use of motion verbs to introduce the main action (see §2.5.2): ‫ּוסעּו ָלכֶם ּובֹאּו הַר ָהאֱמ ִֹרי‬ ְ ‫( ְּפנּו‬Deut 1:7, ‘Start out and make your way to the hill country of the Amorites’); ‫ּורׁשּו אֶת־ ָה ָארֶץ‬ ְ ‫ ּבֹאּו‬,‫( ְראֵה נָתַ ִּתי ִל ְפנֵיכֶם אֶת־ ָה ָארֶץ‬v. 8, ‘See, I place the land at your disposal. Go, take possession of the land’). However, these passages are followed by typical long noun groups (v. 7b) and intricate subordinate clauses (v. 8b). Thus, despite a degree of preference for short clauses (39.54%), the opening narrative of Deut 1:6–17 is characterized by a large number of noun groups (81.40%), subordinate clauses (25.58%), clauses in complex subordination (11.63%), and long clauses (16.27%). A similar phenomenon is attested in the Deuteronomic version of the tale of the spies (1:19–46), in which the frequency of short clauses is even higher (42.74%, like type 2 of the lean, brisk style). Subordination, however, is even more conspicuous than in the introductory narrative (32.26%), as is complex subordination (20.97%). Similarly, chaps. 2–3 and 9–10 manifest the full-blown, elaborate style. These data support the inference that, although expert scribes may have intended to compose texts in the lean, brisk style, they were unable to carry out their plan and reverted back to their normal, elaborate style. More consistency is revealed by the Jonah tale. Linguistic analysis indicates its clear LBH features (Young, Rezetko, and Ehrensvärd 2008: 2.44–45), which place it in the Achaemenid corpus. 41 Nevertheless, large parts of its narrative are couched in the lean, brisk style that, in my view, would be characteristic of much earlier periods (see table 6). However, despite these data, the Jonah tale cannot serve as evidence for a scribe in the Neo-Babylonian/ Achaemenid period who was writing in the style of the Medial corpus. I note a few problematic issues. First, if the narrator uses the lean, brisk style, he is, by implication, attentive to spoken language and open to influences from the vernacular, as indeed revealed by the LBH features of this tale that are there for 41.  Both Athalya Brenner (1979) and George Landes (1982) conclude that this tale reflects the transitional, exilic period, a conclusion that fits with Dobbs-Allsopp’s (1998) dating of Lamentations and the simple fact that ‘the great town’ (Babylon, of course, in a kind of anagram: ‫נינוה‬/‫ )בבל‬is still in full power.

328

Frank H. Polak

all to see. Second, the statistics in table 6 indicate a gradual development from a clearly lean, brisk diction in the opening tale to a strictly elaborate style in the closing chapter. A third point is that the themes figuring in the present tale are truly exceptional. In particular, the narrative is set in a foreign context, including a Mesopotamian town, and features not only a ship manned, most likely, by Phoenician sailors but also a tremendous storm at sea that threatens to sink the ship. The swift development of the events and the outburst of panic on the ship supply sufficient explanation for the use of the swift, brisk style. Thus, the Jonah tale demonstrates that exceptional themes may affect the stylistic design of a narrative (Polak 2006: 134 n. 77; 2010: 57). Hence, both in Deuteronomy and in the Jonah tale the use of a style that is foreign to the scribe from a cultural point of view, as he crosses from the scribal register into the register of orality, remains imperfect. The chancery scribe may abandon his customary style under the influence of the vernacular of his own time or of literary examples and templates but soon falls back into his scribal style. Aside from these few exceptions, it is possible to state that the narratives gathered under the rubrics Medial corpus, Judean corpus, and Achaemenid corpus belong to their own period and are distinguished by their own characteristic use of language. 5. Conclusion The point of departure for my study is the finding, based on the syntacticstylistic analysis of more than 200 examples, that different corpora in biblical narrative are dominated by different language forms. The dominant language form of texts in the Medial corpus is quite different from the language form of the Judean corpus, which originated in the seventh or late sixth century b.c.e., and the Achaemenid corpus, which I attribute to the Persian and Neo-Babylonian era. The distinction between these styles indicates different sociocultural backgrounds and, hence, different sociohistorical contexts. The tales in the intricate, elaborate style of the Judean corpus reveal all the characteristics of written language (long clauses, long noun groups, and a relatively high frequency of clauses in hypotaxis) and thus were composed in the scribal style of the chancery of the Judean monarchy. The characteristics of the elaborate style are also found in the Judean epigraphic texts. The language of the Achaemenid corpus is even more complex than that of the Judean corpus and manifests clear signs of Aramaic and Persian influence, characteristically in the administrative register. The scribal character of the latter two corpora indicates their sociohistorical context: these corpora were rooted in the scribal chancery and thus were composed by people close to the royal or governmental bureaucracy and the centers of power. In this sense, use of the scribal style implies not only scribal expertise but also prestige and authority.

Sociocultural Background of Biblical Narrative

329

A milieu of a very different character is implied by the lean, brisk style of the Medial corpus. The characteristics of this style with its short, mostly paratactic clauses are close to the cross-linguistic characteristics of spontaneous spoken language. Hence, this style suggests roots in oral narrative. In other words, the tales in the Medial corpus were committed to writing by people who had witnessed the oral performance of this tale by the narrator/singer or had intimate knowledge of the oral style and considered this style to be normative for literary discourse.  42 This means that, in the writers’ milieu, the language of oral performance was considered prestigious, and the power center was not that of the royal chancery but the higher strata in society before the development of the advanced bureaucracy. While the seventh–sixth century was home to the Judean corpus, and the fifth (or fourth) century to the Achaemenid corpus, the tenth through the eighth centuries b.c.e. were home to the texts of the Medial corpus. 42.  The notion of coexistence and symbiosis fits the use of the lean, brisk style by scribes in the period preceding the full-blown, widespread royal bureaucracy (Polak 2003: 82–84). A model for such symbiosis in this period is supplied, it seems, by the Mesha Stele, which embodies many scribal features but nevertheless maintains a strongly paratactic style (Polak 2006: 159). Here, we witness the work of a scribal framework, probably representing a kingdom in its inception, which was close to the oral performance.

References

Aharoni, Yohanan 1986 Arad Inscriptions. 2nd ed. Jerusalem: Bialik Institute and Israel Exploration Society. [Hebrew] Amodio, Marc C. 2004 Writing the Oral Tradition: Oral Poetics and Literate Culture in Medieval England. Notre Dame, IN: Notre Dame University Press. Assmann, Aleida 1983 Schriftliche Folklore. Pp. 175–93 in Schrift und Gedächtnis: Beiträge zur Archaeologie der literarischen Kommunikation, ed. A. Assmann, J. Assmann, and C. Hardmeider. Munich: Fink. Assmann, Aleida; Assmann, Jan; and Hardmeier, Christof, eds. 1983 Schrift und Gedächtnis. Beiträge zur Archaeologie der literarischen Kommunikation. Munich: Fink. Bäuml, Franz 1984 Medieval Texts and the Two Theories of Oral-Formulaic Composition: A Proposal for a Third Theory. New Literary History 16: 31–49. 1986 The Oral Tradition and Middle High German Literature. Oral Tradition 1: 398–445. 1993 Verschriftlichte Mündlichkeit und vermündlichte Schriftlichkeit. Pp. 254–66 in Schriftlichkeit im frühen Mittelalter, ed. U. Schaefer. Scriptoralia 53. Tübingen: Narr.

330

Frank H. Polak

1998 Autorität und Performanz: Gesehene Leser, gehörte Bilder, geschriebener Text. Pp.  248–73 in Verschriftung und Verschriftlichung. Aspekte des Medienwechsels in verschiedenen Kulturen und Epochen, ed. C. Ehler and U. Schaefer. Scriptoralia 94. Tübingen: Narr. Baugh, Albert C., and Cable, Thomas 2002 A History of the English Language. 5th ed. London: Methuen. Behaghel, Otto 1900 Geschriebenes Deutsch und Gesprochenes Deutsch. Wissenschaftliche Beihefte zur Zeitschrift des Allgemeinen Deutschen Sprachvereins 17–18: 213–32. Besnier, Niko 1988 The Linguistic Relationships of Spoken and Written Nukulaelae Registers. Language 64: 707–36. Biber, Douglas 1988 Variation across Speech and Writing. Cambridge: Cambridge University Press. 1995 Dimensions of Register Variation: A Cross-Linguistic Comparison. Cambridge: Cambridge University Press. Biber, Douglas, and Conrad, Susan 2001 Register Variation: A Corpus Approach. Pp. 175–96 in The Handbook of Discourse Analysis, ed. Deborah Schiffrin, Deborah Tannen, and Heidi E. Hamilton. Oxford: Blackwell. 2009 Register, Genre and Style. Cambridge: Cambridge University Press. Biber, Douglas, and Finegan, Edward 1989 Drift and the Evolution of English Style: A History of Three Genres. Language 65: 487–517. Brenner, Athalya 1979 The Language of the Book of Jonah as a Means of Establishing the Date of Its Composition. Beth Miqra 79: 396–405. [Hebrew, with English summary] Bolonyai, Agnes 2009 Code-Switching, Imperfect Acquisition, and Attrition. Pp.  253–69 in The Cambridge Handbook of Linguistic Code-Switching, ed. B. E. Bullock and A. J. Toribio. Cambridge: Cambridge University Press. Brooks, Cleanth, and Warren, Robert P. 1958 Modern Rhetoric. 2nd ed. New York: Harcourt, Brace and Company. Bullock, Barbara E., and Toribio, Almeida Jacqueline, eds., 2009 The Cambridge Handbook of Linguistic Code-Switching. Cambridge: Cambridge University Press. Carr, David M. 2005 Writing on the Tablet of the Heart: Origins of Scripture and Literature. Oxford: Oxford University Press. Cassuto, Umberto M. 1973–75  Biblical and Oriental Studies, trans. I.  Abrahams. 2 vols. Jerusalem: Magnes.

Sociocultural Background of Biblical Narrative

331

Chafe, Wallace L. 1982 Integration and Involvement in Speaking, Writing and Oral Literature. Pp. 35–53 in Spoken and Written Language: Exploring Orality and Literacy, ed. D. Tannen. Norwood, NJ: Ablex. 1985 Linguistic Differences Produced by Differences between Speaking and Writing. Pp.  105–23 in Literacy, Language and Learning: The Nature and Consequences of Writing and Reading, ed. D.  R. Olson, N.  Torrance, and A. Hildyard. Cambridge: Cambridge University Press. 1994 Discourse, Consciousness and Time: The Flow and Displacement of Conscious Experience in Speaking and Writing. Chicago: University of Chicago Press. Chafe, Wallace L., and Danielewicz, Jane 1987 Properties of Spoken and Written Language. Pp. 83–113 in Comprehending Oral and Written Language, ed. Rosa Horowitz and S.  Jay Samuels. San Diego: Academic Press. Chafe, Wallace L., and Tannen, Deborah 1987 The Relation between Written and Spoken Language. Annual Review of Anthropology 16: 383–407. Chantraine, Pierre 1968–77 Dictionnaire étymologique de la langue greque. Histoire des mots. Paris: Klincksieck. Cohen, Chaim 2004 The Saga of a Unique Verb in Biblical Hebrew and Ugaritic: ‫‘ השתחוה‬to Bow Down’–Usage and Etymology. Pp. 321–44 in Textures and Meaning: Thirty Years of Judaic Studies at the University of Massachusetts, Amherst, ed. L. Ehr­lich et al. Amherst: Department of Judaic and Near Eastern Studies, University of Massachusetts. http://www.umass.edu/judaic/anniversaryvolume/. Coupland, Nikolas 2007 Style: Language Variation and Identity. Cambridge: Cambridge University Press. Dobbs-Allsopp, F. W. 1998 Linguistic Evidence for the Date of Lamentations. JANES 26: 1–36. Eckert, Penelope 2002 Style and Social Meaning. Pp. 119–26 in Style and Sociolinguistic Variation, ed. P. Eckert and J. R. Rickford. Cambridge: Cambridge University Press. Eckert, Penelope, and Rickford, John R.  2002 Style and Sociolinguistic Variation. Cambridge: Cambridge University Press. Ehler, Christine, and Schaefer, Ursula, eds. 1998 Verschriftung und Verschriftlichung: Aspekte des Medienwechsels in verschiedenen Kulturen und Epochen. Scriptoralia 94. Tübingen: Narr. Erteschik-Shir, Nomi 2007 Information Structure: The Syntax-Discourse Interface. Oxford: Oxford University Press. Eshel, Esther; Puech, Émil; and Kloner, Amos 2007 Aramaic Scribal Exercises of the Hellenistic Period from Maresha: Bowls A and B. BASOR 345: 39–62.

332

Frank H. Polak

Eskhult, Mats 1990 Studies in Verbal Aspect and Narrative Technique in Biblical Hebrew Prose. Stockholm: Almqvist & Wiksell. Fasold, Ralph 1987 The Sociolinguistics of Society, Oxford: Blackwell. Fennell, Barbara A., and Bennett, John 1991 Sociolinguistic Concepts and Literary Analysis. American Speech 66: 371–79. Fishman, Joshua A. 1970 Sociolinguistics: A Brief Introduction. Rowley, MA: Newbury. 1989 Language and Ethnicity in Minority Sociolinguistic Perspective. Clevedon, U.K.: Multilingual Matters 2006 Language Maintenance, Language Shift, and Reversing Language Shift. Pp. 406–36 in The Handbook of Bilingualism, ed. Tej K. Bhatia and William C. Ritchie. Oxford: Blackwell. Foley, John Miles 2010 Verbal Marketplaces and the Oral-Literate Continuum. Pp. 17–37 in Along the Oral-Literate Continuum: Types of Texts, Relations and Their Implications, ed. S. Ranković. Utrecht Studies in Medieval Literacy 20. Turnhout: Brepols. Fox, Barbara A. 1987 Discourse Structure and Anaphora: Written and Conversational English. Cambridge: Cambridge University Press. Gandz, Solomon 1935 Oral Tradition in the Bible. Pp.  249–69 in Jewish Studies in Memory of George A. Kohut, ed. S. W. Baron and A. Marx. New York: Alexander Kohut Memorial Foundation. Garr, W. Randall 1985 Dialect Geography of Syria–Palestine, 1000–586 b.c.e. Philadelphia: University of Pennsylvania Press. [Repr. Winona Lake, IN: Eisebrauns, 2004] George, Andrew R. 2003 The Babylonian Gilgamesh Epic: Introduction, Critical Edition, and Cuneiform Texts. 2 vols. Oxford: Oxford University Press, 2003. Gabryś-Barker, Danuta, ed. 2008 Morphosyntax in Second Language Acquisition. Clevedon, U.K.: Multilingual Matters. Gumbrich, Hans Ulrich 1983 Schriftlichkeit in mündlicher Kultur. Pp. 158–174 in Schrift und Gedächtnis: Beiträge zur Archaeologie der literarischen Kommunikation, ed. A. Assmann, J. Assmann, and C. Hardmeider. Munich: Fink. Güterbock, Hans G. 1951 The Song of Ullikummi: Revised Text of the Hittite Version of a Hurrian Myth. JCS 5: 135–61. Halliday, Michael A. K. 1989 Spoken and Written Language. 2nd ed. Oxford: Oxford University Press.

Sociocultural Background of Biblical Narrative

333

Halpern, Baruch, and Lemaire, André 2010 The Composition of Kings. Pp. 123–54 in The Books of Kings: Sources, Composition and Reception, ed. A. Lemaire, B. Halpern and M. J. Adams. VTSup 129. Leiden: Brill. Hoftijzer, J. 1976 Grammatical Survey of the Language the Texts Are Written In. Pp. 283–302 in Aramaic Texts from Deir ʿAlla, ed. J. Hoftijzer and G. van der Kooij. Leiden: Brill. Horobin, Simon, and Smith, Jeremy 2002 An Introduction to Middle English. Edinburgh: Edinburgh University Press. Hurvitz, Avi 1982 A Linguistic Study of the Relationship between the Priestly Source and the Book of Ezekiel. CahRB 20. Paris: Gabalda. 2000 Once Again the Linguistic Profile of the Priestly Material in the Pentateuch and Its Historical Age: A Response to J. Blenkinsopp. ZAW 112: 180–91. Irvine, Susan, ed. 2004 The Anglo-Saxon Chronicle: A Collaborative Edition, vol. 7: ms  E. Cambridge: Brewer. Katzenberger, Irit 2004 The Development of Clause Packaging in Spoken and Written Texts. Journal of Pragmatics 36: 1921–48. Koch, Peter, and Oesterreicher, Wulf 1994 Funktionale Aspekte der Schriftkultur / Functional Aspects of Literacy. Pp.  587–604 in Schrift und Schriftlichkeit / Writing and Its Use, vol.  1, ed. Hartmuth Günther and Otto Ludwig. Berlin: de Gruyter. Kofoed, Jens Brun 2006 Using Linguistic Difference in Relative Text Dating: Insights from Other Historical Linguistic Case Studies. HS 47: 93–114. Kutas, Marta; Moreno, Eva; and Wicha, Nichole 2009 Code-Switching and the Brain. Pp. 289–305 in The Cambridge Handbook of Linguistic Code-Switching, ed. B. E. Bullock and A. J. Toribio. Cambridge: Cambridge University Press. Landes, George M. 1982 Linguistic Criteria and the Date of the Book of Jonah. ErIsr 16 (Orlinsky volume): 147*–70*. Lemaire, André; Halpern, Baruch; and Adams, Matthew J., eds. 2010 The Books of Kings: Sources, Composition and Reception. VTSup 129. Leiden: Brill. Leuchter, Mark 2010 The Sociolinguistic and Rhetorical Implications of the Source Citations in Kings. Pp. 119–34 in Soundings in Kings: Perspectives and Methods in Contemporary Scholarship, ed. Mark Leuchter and Klaus-Peter Adam. Minneapolis: Fortress. Levin, Harry 1972 Observations on the Style of Ernest Hemingway. Pp. 322–37 in Essays in Stylistic Analysis, ed. Howard S. Babb. New York: Harcourt Brace Jovanovich.

334

Frank H. Polak

Lichtenberger-Schäfer, Christa 1996 Sociological and Biblical Views of the Early State. Pp. 78–105 in The Origins of the Ancient Israelite States, ed. Volkmar Fritz and Philip R. Davies. JSOTSup 228. Sheffield: Sheffield Academic Press. Lipschits, Oded; Sergi, Omer; and Koch, Ido 2010 Royal Judahite Jar Handles: Reconsidering the Chronology of the lmlk Stamp Impressions. Tel Aviv 37: 3–32. 2011 Judahite Stamped and Incised Jar Handles: A Tool for Studying the History of Late Monarchic Judah. Tel Aviv 38:5–41 Loewenstamm, Samuel E. 1981 Comparative Studies in Biblical and Ancient Oriental Literatures. AOAT 204. Kevelaer: Butzon & Bercker / Neukirchen-Vluyn: Neukirchener Verlag. Lohfink, Norbert 1999 Was There a Deuteronomistic Movement? Pp. 36–66 in Those Elusive Deuteronomists: The Phenomenon of Pan-Deuteronomism, ed. Linda S. Schearing and Steven L. McKenzie. JSOTSup 268. Sheffield: Sheffield Academic Press. Longacre, Robert E. 1989 Joseph: A Story of Divine Providence. A Text Theoretical and Textlinguistic Analysis of Genesis 37 and 39–48. Winona Lake, IN: Eisenbrauns. Lord, Albert Bates 1995 The Singer Resumes the Tale, ed. Mary Louise Lord. Ithaca, NY: Cornell University Press McKitterick, Rosamond 1990 Conclusion. Pp. 319–33 in The Uses of Literacy in Early Mediaeval Europe, ed. Rosamond McKitterick. Cambridge: Cambridge University Press. Miccio, Adele W.; Scheffner Hammer, Carol; and Rodríguez, Bárbara 2009 Code Switching and Language Disorders in Bilingual Children. Pp. 241–52 in The Cambridge Handbook of Linguistic Code-Switching, ed. B. E. Bullock and A. J. Toribio. Cambridge: Cambridge University Press. Miles, Josephine 1964 Eras and Modes in English Poetry. 2nd ed. Berkeley: University of California Press. 1965 The Continuity of Poetic Language: The Primary Language of Poetry, 1540′s—1940′s. New York: Octagon. 1967 Style and Proportion: The Language of Prose and Poetry. Boston: Little, Brown. Millard, Alan R. 2001 The Corpus of West-Semitic Stamp Seals: Review Article. IEJ 51: 76–87. 2010 Books and Writing in Kings. Pp. 155–62 in The Books of Kings: Sources, Composition and Reception, ed. A. Lemaire, B. Halpern, and M. J. Adams. VTSup 129. Leiden: Brill. Miller, Cynthia L. 1996 The Representation of Speech in Biblical Hebrew Narrative. HSM 55. Atlanta: Scholars Press. Miller, Jim, and Weinert, Regina 1998 Spontaneous Spoken Language: Syntax and Discourse. Oxford: Clarendon.

Sociocultural Background of Biblical Narrative

335

Morrow, William 2010 ‘To Set the Name’ in the Deuteronomic Centralization Formula: A Case of Cultural Hybridity. JSS 55: 365–83. Niditch, Susan 1996 Oral World and Written Word: Ancient Israelite Literature. Louisville: Westminster John Knox. 2010 Epic and History in the Hebrew Bible: Definitions, “Ethnic Genres,” and the Challenges of Cultural Identity in the Biblical Book of Judges. Pp. 86– 102 in Epic and History, ed. David Konstan and Kurt A. Raaflaub. Oxford: Blackwell. Niemann, H. M. 1993 Herrschaft, Königtum und Staat: Skizzen zur soziokulturellen Entwicklung im monarchischen Israel. FAT 6. Tübingen: Mohr Siebeck. Niles, John D. 1999 Homo Narrans: The Poetics and Anthropology of Oral Literature. Philadelphia: University of Pennsylvania Press. Ochs, Elinor 1979 Planned and Unplanned Discourse. Pp.  51–80 in Discourse and Syntax, ed. T. Givón. Syntax and Semantics 12. San Diego: Academic Press. Olmo Lete, Gregorio del 1981 Mitos e Leyendas de Canaan segun la Tradicion de Ugarit. Madrid: Cristiandad. Olson, David R.; Torrance, N.; and Hildyard, A., eds. 1985 Literacy, Language and Learning: The Nature and Consequences of Writing and Reading. Cambridge: Cambridge University Press. Otto, Eckhart 1999 Das Deuteronomium: Politische Theologie und Rechtsreform in Juda und Assyrien. BZAW 284. Berlin: de Gruyter. Pawley, Andrew, and Syder, Frances H. 1983 Natural Selection in Syntax: Notes on Adaptive Variation and Change in Vernacular and Literary Grammar. Journal of Pragmatics 7: 551–79. Polak, Frank H. 1996 Theophany and Mediator: The Unfolding of a Theme in the Book of Exodus. Pp. 117–47 in Studies in the Book of Exodus: Redaction-Reception-Interpretation, ed. M. Vervenne; BETL 126. Leuven: Leuven University Press. 1998 The Oral and the Written: Syntax, Stylistics and the Development of Biblical Prose. JANES 26: 59–105. 2002 On Dialogue and Speaker Status in Biblical Narrative. Beth Miqra 48: 1–18, 97–119. [Hebrew with English summary] 2003 Style Is More Than the Person: Sociolinguistics, Literary Culture and the Distinction between Written and Oral Narrative. Pp. 38–103 in Biblical Hebrew: Studies in Chronology and Typology, ed. Ian M. Young. JSOTSup 369. London: T. & T. Clark. 2005 Sociolinguistics and the Judean Speech Community in the Achaemenid Empire. Pp.  589–628 in Judah and Judaeans in the Achaemenid Period, ed. O. Lipschits and M. Oeming. Winona Lake, IN: Eisenbrauns.

336

Frank H. Polak

2006a Linguistic and Stylistic Aspects of Epic Formulae in Ancient Semitic Poetry and Biblical Narrative. Pp. 285–304 in Biblical Hebrew in Its Northwest Semitic Setting, ed. S. Fassberg and A. Hurvitz. Jerusalem: Magnes / Winona Lake, IN: Eisenbrauns. 2006b Sociolinguistics: A Key to the Typology and the Social Background of Biblical Hebrew. HS 47: 115–62. 2009 Verbs of Motion in Biblical Hebrew: Lexical Shifts and Syntactic Structure. Pp. 161–97 in A Palimpsest: Rhetoric, Ideology, Stylistics, and Language Relating to Persian Israel, ed. E. Ben-Zvi, D. Edelman, and F. H. Polak. Piscataway, NJ: Gorgias. 2010 The Book of Samuel and the Deuteronomist: A Syntactic-Stylistic Analysis. Pp. 34–73 in The Books of Samuel and the Deuteronomists, ed. C. SchäferLichtenberger. BWANT 188. Stuttgart: Kohlhammer. Ranković, Slavica 2010 The Oral-Literate Continuum as Space. Pp. 39–71 in Along the Oral-Literate Continuum: Types of Texts, Relations and Their Implications, ed. S. Ranković. Utrecht Studies in Medieval Literacy 20. Turnhout: Brepols. Ranković, Slavica, ed. 2010 Along the Oral-Literate Continuum. Types of Texts, Relations and Their Implications. Utrecht Studies in Medieval Literacy20. Turnhout: Brepols. Regt, Lenart J. de 1999 Participants in Old Testament Texts and the Translator: Reference Devices and Their Structural Impact. SSN 39. Assen: Van Gorcum. Reiner, Erica, and Güterbock, Hans G. 1967 The Great Prayer to Ishtar and Its Two Versions from Boǧazköy. JCS 21: 255–66. Richter, Wolfgang 1964 Die Bearbeitungen des “Retterbuches” in der deuteronomistischen Epoche. BBB 21. Bonn: Hanstein. Römer, Thomas C. 2000 l’École deutéronomiste et la formation de la bible hébraïque. Pp. 179–91 in The Future of the Deuteronomistic History, ed. Thomas C. Römer. BETL 147. Leuven: Leuven University Press 2005 The So-Called Deuteronomistic History: A Sociological, Historical and Literary Introduction. London: T. & T. Clark. Rosenthal, Franz 1963 A Grammar of Biblical Aramaic. Wiesbaden: Harassowitz. Saville-Troike, Muriel 2003 The Ethnography of Communication: An Introduction. Oxford: Blackwell. Schaack, Thomas 1998 Die Ungeduld des Papiers: Studien zum alttestamentlichen Verständnis des Schreibens anhand des Verbums katab im Kontext administrativer Vorgänge. BZAW 262. Berlin: de Gruyter. Schaefer, Ursula 1992 Vokalität: Altenglische Dichting zwischen Mündlichkeit und Schriftlichkeit. Scriptoralia 39. Tübingen: Narr.

Sociocultural Background of Biblical Narrative

337

Schaefer, Ursula, ed. 1993 Schriftlichkeit im frühen Mittelalter. Scriptoralia 53. Tübingen: Narr. Schaper, Joachim 2004 A Theology of Writing: The Oral and the Written, God as Scribe and the Book of Deuteronomy. Pp. 97–119 in Anthropology and Biblical Studies: Avenues of Approach, ed. L. J. Lawrence and M. I. Aguilar. Leiden: Deo. 2005 Exilic and Post-exilic Prophecy and the Orality/Literacy Problem. VT 55: 324–41. Schniedewind, William M. 2004 How the Bible Became a Book: The Textualization of Ancient Israel. 2006 Aramaic, the Death of Written Hebrew, and Language Shift in the Persian Period. Pp.  137–47 in Margins of Writing, Origins of Cultures, ed. Seth L. Sanders. Chicago: The Oriental Institute of the University of Chicago. Schwyter, Jürg R. 1998 Syntax and Style in the Anglo-Saxon Law-Codes. Pp.  189–231 in Verschriftung und Verschriftlichung: Aspekte des Medienwechsels in verschiedenen Kulturen und Epochen, ed. C. Ehler and U. Schaefer. Scriptoralia 94. Tübingen: Narr. Seow, Benjamin D. 1996 Linguistic Evidence and the Dating of Qoheleth. JBL 115: 643–66. Smith, Mark S. 1994 The Ugaritic Baal Cycle: Volume 1: Introduction with Text, Translation and Commentary of KTU I.1–I.2. VTSup 55. Leiden: Brill. Sommer, C. L. 1999 Revelation at Sinai in the Hebrew Bible and in Jewish Theology. JR 79: 422–51. Steiner, Richard C., and Leiman, Sid Z. 2009 The Lost Meaning of Deuteronomy 33:2 as Preserved in the Palestinian Targum to the Decalogue. Pp. 157–66 in Mishneh Todah: Studies in Deuteronomy and Its Cultural Environment in Honor of Jeffrey H. Tigay, ed. Nili Sacher Fox, David A. Glatt-Gilad, and Michael J. Williams. Winona Lake, IN: Eisenbrauns. Steymans, Hans Ulrich 1995 Deuteronomium 28 und die âdê zur Thronfolgeregelung Asarhaddons: Segen und Fluch im Alten Orient und in Israel. OBO 145. Göttingen: Vandenhoeck & Ruprecht / Freiburg: Universitätsverlag. Stolper, Matthew W. 1984 The Neo-Babylonian Text from the Persepolis Fortification. JNES 43: 299–310. Strang, Barbara M. H. 1970 The History of English. London: Methuen. Tannen, Deborah 1982 Oral and Literate Strategies in Spoken and Written Narratives. Language 58: 1–21. 1984 Conversational Style: Analyzing Talk among Friends. Norwood, NJ: Ablex.

338

Frank H. Polak

1985 Relative Focus on Involvement in Oral and Written Discourse. Pp. 124–47 in Literacy, Language and Learning: The Nature and Consequences of Writing and Reading, ed. D. R. Olson, N. Torrance, and A. Hildyard. Cambridge: Cambridge University Press. Tavernier, Jan 2007 Iranica in the Achaemenid Period (ca. 550–330 b.c.): Lexicon of Old Iranian Proper Names and Loanwords, Attested in Non-Iranian Texts. OLA 158. Leuven: Peeters. Verheij, Arian J. C. 1990 Verbs and Numbers: A Study of the Frequencies of the Hebrew Verbal Tense Forms in the Books of Samuel, Kings, and Chronicles. SSN 28. Assen: Van Gorcum. Wi, Longxing 2009 Code-Switching and the Bilingual Mental Lexicon. Pp. 271–88 in The Cambridge Handbook of Linguistic Code-Switching, ed. B. E. Bullock and A. J. Toribio. Cambridge: Cambridge University Press. Young, Ian M. 1993 Diversity in Pre-exilic Hebrew. FAT 5. Tübingen: Mohr Siebeck. 2003 Late Biblical Hebrew and Hebrew Inscriptions. Pp. 276–311 in Biblical Hebrew: Studies in Chronology and Typology, ed. I. M. Young. JSOTSup 369. London: T. & T. Clark. Young, Ian M.; Rezetko, Robert; and Ehrensvärd, Martin 2008 Linguistic Dating of Biblical Texts: An Introduction to Approaches and Problems. 2 vols. London: Equinox. Zadok, Ran 1982 Remarks on Ezra and Nehemiah. ZAW 94: 296–98. Zumthor, Paul 1984a La poésie et la voix danse la civilisation médiévale. Essais et Conférences Collège de France. Paris: Presses Universitaires de France. 1984b The Text and the Voice. New Literary History 16: 67–92.

Northern Hebrew through Time: From the Song of Deborah to the Mishnah Gary A. Rendsburg 1. Introduction Over the course of the past 25 years, I have devoted a series of books and articles to the subject of the northern dialect of ancient Hebrew. The majority of these studies have been concerned with the portions of the Bible written in Israelian Hebrew (IH), while several studies have treated Mishnaic Hebrew (MH). The latter represents not only a colloquial variety of ancient Hebrew (Rendsburg 1990) but also a northern dialect, in keeping with the location of Sepphoris and Tiberias, the major centers of Tannaitic activity (Rendsburg 1992; 2003b). 1 This essay addresses the question whether any diachronic change is discernible within the approximately 1300-year span of IH–MH. This period stretches from ca. 1100 b.c.e., an oft-suggested date for the Song of Deborah in Judges 5, to ca. 200 c.e., the approximate date of the redaction of the Mishnah and related texts. 2 Given that a great amount of diachronic change is discernible for texts emanating from Judah, reflected in the major features that distinguish Standard Biblical Hebrew (SBH) from Late Biblical Hebrew (LBH), one would assume prima facie that discernible diachronic change is reflected in the IH–MH continuum as well. Author’s note:  It is my pleasant duty to thank the Oxford Centre for Hebrew and Jewish Studies at Yarnton Manor for bestowing upon me Visiting Scholar status for the period 4 October 2010 through 15 February 2011, during which time the present article was written. In addition, I thank Ziony Zevit for his excellent comments on an earlier version of this paper. 1.  For a general sense of the role played by these two cities, one may consult a standard reference work such as Stemberger 1996: 76–81, with sketches of individual rabbis such as R. Meir at Tiberias, R. Yose ben Halafta at Sepphoris, R. Nehorai at Sephoris, and of course R. Judah ha-Nasi at Bet Sheʿarim and Sepphoris. Note as well already R. Yose ha-Gelili (a second-generation Tanna), plus the location of the Sanhedrin in Usha and Shefarʿam, before it relocated to Bet Sheʿarim, then Sepphoris, and finally Tiberias. 2.  I do not engage here the question of priority between the Mishnah and Tosefta (or of individual passages or tractates in the Tosefta and Mishnah), nor are we able to pinpoint the date of the equally important Midreshe Halaka texts (Mekhilta, Sipra, Sipre BeMidbar, Sipre Devarim, etc.). For the purposes of the present study, one must accede to the view that the corpus of Tannaitic texts arises ca. 200 c.e., with some material presumably composed before this date, even if all or most of the redactional activity takes place after this date.

339

340

Gary A. Rendsburg

Two factors arise, however, that make our task more complicated. First, the amount of data at our disposal for IH is limited. The corpus of Israelian texts in the Bible is smaller than the Judahite corpus, with the former comprising 149 chapters out of a total of 920 (Hebrew) chapters in the Bible, or approximately 16% of the canon (Rendsburg 2003a: 8). Perhaps more importantly, the vast majority of northern texts date to a more limited timeframe, ca. 1000–720 b.c.e. (see §2 below), thus making diachronic analysis more challenging. True, we have a very large corpus of Tannaitic literature from almost a millennium later, but the nature of MH, notwithstanding its northern provenance, is quite different from IH (see further below). Coincidentally (or perhaps not so), the quantitative difference between IH and Judahite Hebrew (JH) is also true of the epigraphic remains from ancient Israel. As witness thereto, note that Shmuel Aḥituv’s standard textbook on ancient Hebrew inscriptions (Aḥituv 2008) devotes 232 pages to Judah (with more than 20 sites represented: Siloam, Arad, Lachish, Meṣad Ḥashavyahu, Ḥurvat ʿUzza, etc.), with, by contrast, only 84 pages devoted to Israel (with only 7 sites represented: ʿIzbet Sartah, Gezer, Kuntillet ʿAjrud, Samaria, Kalah, Hazor, Kinneret). Of these, note that Gezer is on the border of northern Israel and southern Judah, Kuntillet ʿAjrud is far into the Sinai Desert, Kalah is in Mesopotamia—and the ʿIzbet Sartah epigraph is an abecedary. This leaves only Samaria, Hazor, and Kinneret in true northern Israel—though for Hazor we have extremely little, while for Kinneret we have only a single two-word inscription (fortunately, however, we are able to avail ourselves of this very limited evidence; see below, §5). This leaves only the Samaria ostraca as a sizable corpus of texts from a northern site, indeed, from the capital of the Northern Kingdom—and yet even these texts reveal so little, given their very formulaic nature. Apart from numerals and personal names, we have perhaps only 6 lexemes (‫שת‬, ‫נבל‬, ‫ין‬, ‫ישן‬, ‫שמן‬, ‫רחץ‬, ‫ )כרם‬in the main corpus. Only in Samaria ostracon 111 do we gain more than the usual formulas, with 5 other lexemes present (the verbs ‫שלם‬, ‫קשב‬, ‫ ;מנה‬the nouns ‫‘ רעם‬shepherds’ and ‫‘ שערם‬barley’). From all of this, we learn very little about northern Hebrew, notwithstanding the use of the noun ‫‘ שת‬year’ and the monophthongization represented by ‫‘ ין‬wine’, both well discussed in the literature. The second complicating issue is the Aramaic factor. While it is true that many LBH developments are internal to Hebrew, with no connection to Aramaic influence, the fact remains that the shadow of the latter looms everpresent, with many LBH features resulting from the status of Aramaic as the lingua franca of the Near East, especially during the Persian period. Thus, to cite just 2 examples (1 lexical, 1 grammatical) from among dozens: (a) the verb ‫‘ קבל‬take, receive’ appears 10 times in Job 1–2, Esther, Ezra, and Chron-

Northern Hebrew through Time

341

icles (and only 1 time elsewhere, namely, Prov 19:20, though note that Proverbs constitutes an IH composition, as per §2 below); and (b) the abstract suffix ‫ות‬- becomes more common in postexilic texts—for example, ‫‘ ע ְַבדּות‬slavery’ 3 times in Ezra–Nehemiah (replacing SBH ‫ָדים‬ ִ ‫‘ ּבֵית עֲב‬house of slavery’ most of the time). The problem for our treatment of IH arises from recognition of the fact that Aramaic influence is felt on northern Hebrew throughout its history. To be more specific, the imprint of the former on the latter is not quite “Aramaic influence” per se, in the sense of the effect that Aramaic would have on all Hebrew beginning in the 6th century b.c.e. but, rather, influence that arose due to consistent language contact across the Aram-Israel border (an amorphous border, to be sure) for centuries. This contact was similar to the contact that existed between IH and Phoenician in the northern reaches of Israel, and between IH and Ammonite and Moabite in the Transjordanian portions of Israelite settlement. From the data available to us—and here I anticipate one of my conclusions—generally speaking, at least for the biblical period, there is no perceptible diachronic development within IH qua IH. 3 Which is to say, IH features that occur in the earliest texts (including Judges 5) are still present in the latest texts (namely, Nehemiah 9 and Qoheleth). Change does occur, though as we shall see, the modifications across time are the same modifications that occurred throughout the Hebrew language, so that IH simply followed JH in this regard, as the latter morphed from SBH into LBH. 2.  The Sources of Israelian Hebrew Before proceeding to specific documentation that will serve to substantiate the general deductions presented in the previous paragraph, let us first review the sources of IH, with an attempt at a general dating schema. (a) The sole text from the premonarchic period (ca. 1100–1000 b.c.e.) is Judges 5. (b) Texts from the early monarchic period (ca. 1000–860 b.c.e.) include the blessings to the northern tribes in Genesis 49 and Deuteronomy 33; Deuteronomy 32; Lev 25:13–24; select stories in Judges (especially Deborah, Gideon, Jephthah); and 2 Sam 23:1–7. (c) Works from the middle monarchic period (ca. 860–720 b.c.e.) include the Elijah and Elisha cycles; material concerning the kings of Israel; Amos; Hosea; Micah 6–7; and Proverbs. (d) Next come the northern psalms, dated to the post-kingdom period of Israel

3.  By “perceptible,” I mean issues from the realms of morphology, syntax, and lexicon, which may be discerned in the written form of our texts. Slight changes in pronunciation, for example, are less likely to be detected within the literary shape of our compositions.

342

Gary A. Rendsburg

(ca. 720–550 b.c.e.). 4 (e) Finally, Nehemiah 9 and Qoheleth represent the late period (ca. 550–400 b.c.e.). 5 Implied in the above schema is the assumption that, even after the destruction of the Kingdom of Israel in 721 b.c.e., the northern dialect of Hebrew persisted. Notwithstanding the archaeological evidence, which reveals a greatly reduced population in the areas of Samaria and the Galilee—a reduction that arose from (a) deportations by the Assyrians, and (b) migrations southward to Judah—clearly some speakers of Hebrew remained in the region (see, for example, Japhet 1983: 104–5). With no evidence to the contrary, the patois of these Israelites must have remained the northern dialect. As is well known, major social and political upheavals (such as the events of 745–721 b.c.e.) typically cause major changes in language, 6 so it is only natural to expect IH to have undergone certain transformations during this period. Changes of this sort may have included the end of the literary standard and the adaptation of a formerly colloquial register for literary purposes. This would explain, for example, the language of Qoheleth, a book that is both late and northern (for the former, see Seow 1996; for the latter, see Davila 1990), and the eventual emergence of MH, even if our evidence for MH derives from centuries later. At the same time, however, one must note that Nehemiah 9, also dated to the Persian period, is more literary in style (that is, when compared with Qoheleth), though, to be sure, it reflects LBH developments (see further below, §§9.1.2, 9.1.3, 9.2). In addition to the works listed above, the Song of Songs also needs to be incorporated in some fashion. This work is replete with IH features (see Noegel and Rendsburg 2010: 3–62), though the date of this composition remains elusive. Most scholars would assign its authorship to the later end of the spectrum, though in our recent coauthored monograph, Scott Noegel and I proposed an early-monarchic setting for the poem (Noegel and Rendsburg 2010), based chiefly on the anti-Solomonic tenor that we find therein. For our present purposes, I remain agnostic on this matter; in any case, we have sufficient material in the IH corpus aside from the Song of Songs to proceed with our analysis. 4.  Any number of the northern psalms may belong to the previous period as well, though for our present purposes I am content to classify these poems as post–Kingdom of Israel. 5.  Within the limited space of this article, I cannot substantiate the dates of each composition, though I trust that the reader will accede to these broad categorizations (except, perhaps, for those who subscribe to the approach of Young, Rezetko, and Ehrensvärd 2008). 6.  The best example in the history of the English language is the Norman Conquest of 1066, which marks the transition from Old English to Middle English. As for the next major transition, note that the beginning of Early Modern English corresponds more or less to the end of the Plantagenets and the rise of the House of Tudor in 1485, aided by the introduction of the printing press into England by William Caxton in 1476, which served to spread the London standard into other parts of England.

Northern Hebrew through Time

343

3.  Grammatical Features Appearing in Judges 5 and Beyond The main point to be made here is that the same IH features tend to appear throughout the IH corpus, from the earliest attestation of this dialect (even if Judges 5 is its sole representation) through the larger middle three groups delineated above, and into the two Persian-period works. 3.1.  Reduplicated Plural of Geminate Nouns A parade example is the reduplicated plural form of geminate nouns, known from Aramaic and attested in IH from Judges 5 through Nehemiah 9, including an array of texts between them. Examples in relative chronological order, based on the above dating schema, include (Rendsburg 2003a: 14–15): 7 Judg 5:14 ָ‫‘ עֲ ָממֶיך‬your people’; Judg 5:15 ‫‘ ִח ְקקֵי‬decisions of’; Deut 33:15 ‫ה ְַררֵ י‬ ‘mountains of’; Prov 29:13 ‫ָכים‬ ִ ‫‘ ְּתכ‬oppressions’; Ps 36:7, 50:10, 76:5, 87:1, 133:3 ‫‘ ה ְַררֵ י‬mountains of’; Ps 77:18 ָ‫ח ָצצֶיך‬ ִ ‫עֲמ‬ ֲ ‘your arrows’; Neh 9:22 ‫ָמים‬ ‘peoples’; Neh 9:24 ‫‘ ע ְַממֵי‬peoples of’ (see also Song 2:17, 4:6 ‫ָלים‬ ִ ‫‘ ְצל‬shadows’; Song 4:8 ‫‘ ה ְַררֵ י‬mountains of’). In SBH, by contrast, the forms are ‫ע ִַּמים‬ ‘peoples’, ‫‘ הָרֵ י‬mountains of’, etc. 3.2.  Feminine-Singular Nominal Ending ‫ֹות‬A second example of an IH feature attested in Judges 5 and then for centuries afterward is the fs nominal ending ‫ֹות‬-, witnessed in the following forms (Rendsburg 2003a: 14): Judg 5:29 ‫‘ ח ְַכמֹות‬wise lady’; 2 Kgs 6:8 ‫‘ ּתַ חֲנ ִֹתי‬my camp’; Ps 45:1 ‫‘ י ְִדידֹת‬love’; Ps 45:16 ‫‘ ְׂשמָחֹת‬joy’; Ps 53:7 ‫‘ ְיׁשֻעֹות‬salvation’; Ps 73:22 ‫‘ ְּבהֵמֹות‬beast’; Ps 132:12 ‫‘ עֵד ִֹתי‬my testimony’; Prov 1:20, 9:1, 24:7 ‫ָכמֹות‬ ְ ‫‘ ח‬wisdom’; Prov 14:1 ‫‘ ח ְַכמֹות‬wise lady’; Prov 24:7 ‫‘ ָראמֹות‬high’ (adj.), Prov 28:20 ‫‘ אֱמּונֹות‬faith’; Qoh 1:17, 2:12, 7:25, 9:3 ‫‘ הֹולֵלֹות‬madness’. In these forms, not only has /-at/ been retained in the absolute state (see, for example, 2 Kgs 9:17 ‫ׁש ְפעַת‬ ִ ‘multitude’; Hos 7:5 ‫חמַת‬ ֲ , Ps 16:6 ‫חלָת‬ ֲ ַ‫‘ נ‬inheritance’) without shifting to /-ā/, as occurs in SBH, but the short /a/ vowel has shifted to /o:/, exactly as obtains in Phoenician. 3.3.  Relative Pronoun -‫שׁ‬ ֶ A third grammatical element attested in northern compositions stretching from the Song of Deborah to Qoheleth is the use of the relative pronoun -‫שׁ‬ ֶ (with alternative pronunciation - ַ‫שׁ‬/-‫)שׁ‬, ָ attested as follows (Rendsburg 2003a: 12–13): Judg 5:7 (2×) ‫ ;ׁשַ ּקַ ְמ ִּתי‬Judg 6:17 ‫ַּתה‬ ָ ‫;ׁשא‬ ָ Judg 7:12, 8:26 ‫;ׁשעַל‬ ֶ 2 Kgs 6:11 ‫ּׁשּלָנּו‬ ֶ ‫;מ‬ ִ Ps 133:2, 133:3 ‫ׁשּיֹרֵ ד‬. ֶ And then, of course, the morpheme -‫שׁ‬ ֶ appears 67 times in Qoheleth (2× with the vocalization -‫)שׁ‬ ְ and every time (except for the superscription) a relative pronoun of this sort is needed in the 7.  Here and below, I cite only Rendsburg 2003a, which (a) presents the clearest summary of my research into IH, and (b) directs the reader to additional works on the subject. For more recent publications, see Rendsburg 2008; 2009; and Noegel and Rendsburg 2010: 3–62.

344

Gary A. Rendsburg

Song of Songs. 8 This contrasts with the standard relative pronoun in the Bible, namely, ‫ֲשׁר‬ ֶ ‫א‬, which appears not only in JH but in IH as well (suggesting that -‫שׁ‬ ֶ is limited to certain subdialects within IH). 3.4.  ‫‘ ִמן‬from’ before an Anarthrous Noun One final illustration of a feature attested in Judges 5 and then long afterward is the use of ‫‘ ִמן‬from’ before an anarthrous noun, which also appears in Aramaic and in Deir ʿAlla (I 3). 9 Examples of this feature in IH sources are the following (Rendsburg 2003a: 23): Judg 5:20 ‫ן־ׁש ַמיִם‬ ָ ‫‘ ִמ‬from heaven’; Judg 7:23 ‫‘ ִמן־ ָאׁשֵר‬from Asher’; Judg 7:23 ‫ַּׁשה‬ ֶ ‫ָל־מנ‬ ְ ‫‘ ִמן־ּכ‬from all of Manasseh’; Judg 10:11 ‫ן־ּבנֵי עַּמֹון‬ ְ ‫‘ ִמ‬from the Ammonites’; Judg 10:11 ‫ׁש ִּתים‬ ְ ‫ן־ּפ ִל‬ ְ ‫‘ ִמ‬from the Philistines’; 2 Kgs 15:28 ‫‘ ִמן־חַּטֹאות‬from the sins of’; Prov 27:8 ‫ן־קּנָּה‬ ִ ‫‘ ִמ‬from its nest’; Ps 45:9 ‫ֵיכלֵי‬ ְ ‫‘ ִמן־ה‬from the palaces of’; Ps 73:19 ‫‘ ִמן־ ַּבּלָהֹות‬from terrors’; Ps 116:8 ‫ן־ּד ְמעָה‬ ִ ‫‘ ִמ‬from tears’ (see also Song 4:15 ‫ן־לבָנֹון‬ ְ ‫‘ ִמ‬from Lebanon’). This feature eventually penetrated Judahite Hebrew in a major way, so that it emerges as an LBH trait as well, as demonstrated primarily by the 51 instances of ‫ ִמן‬before an anarthrous noun in Chronicles (21 cases of ‫ן־ּבנֵי‬ ְ ‫‘ ִמ‬of/ from the sons of’ [1 Chr 4:42, 5:18, 9:3, etc.]; a disproportionate number of examples with proper names, such as 1 Chr 11:22 ‫‘ ִמן־קַ ְב ְצאֵל‬from Kabzeel’; and select other instances, for example, 2 Chr 31:3 ‫ן־רכּוׁשֹו‬ ְ ‫‘ ִמ‬from his property’). For an example derived from parallel passages, compare 2 Sam 6:12 ‫ִמּבֵית עֹבֵד‬ ‫‘ אֱדֹם‬from the house of Obed-edom’ with 1 Chr 15:25 ‫‘ ִמן־ּבֵית עֹבֵד־אֱדֹם‬from the house of Obed-edom’. The SBH form is exemplified by ‫ּׁש ַמיִם‬ ָ ‫‘ ִמ‬from heaven’ (Isa 14:12, Ps 12:2, etc.), with the nun of ‫ ִמן‬assimilated to the following consonant. 3.5.  An Interim Summary The first three features (§§3.1–3) span about 700 years of literary composition, from ca. 1100 b.c.e. to ca. 400 b.c.e., that is, from the approximate date of the Song of Deborah to the time of either Nehemiah 9 or Qoheleth. For the fourth feature, we cannot reach into the Persian period in Israelian texts (since this trait is not attested in either Nehemiah 9 or Qoheleth), though clearly the use of ‫‘ ִמן‬from’ before an anarthrous noun continued throughout this epoch, as witnessed by the 51 occurrences in Chronicles. More importantly for the present enterprise, these four elements may serve as testimony to the general state of affairs. IH features that appear in our earliest composition, the Song of Deborah, continue to appear for centuries later, at about the same pace, scattered here and there in narratives about the northern judges, annalistic material concerning the kings of Israel, northern psalms, 8.  On the history of -‫שׁ‬, ֶ see Holmstedt in the present volume (see pp. 113–119). 9.  On the identification of Deir ʿAlla as a Canaanite dialect with specific links to IH (or to be more specific, with Gileadite, an IH subdialect used in Transjordan), see Rendsburg 1993.

Northern Hebrew through Time

345

Proverbs, Qoheleth, the Song of Songs, etc. This pattern suggests that IH qua IH reflects little in the way of perceptible diachronic change with regard to the features that I have investigated. 10 4.  Additional Grammatical Features, Attested from the Early Monarchic Period Onward The general standpoint presented in §3.5 is confirmed by a host of other illustrations, though for what follows, the chronological span begins not with the Song of Deborah (which is, after all, a relatively short text with only limited data) but with compositions emanating from the early monarchic period. Serviceable features for analysis include the following. 4.1. Second-Feminine-Singular Independent Pronoun ‫אתי‬ The 2fs independent pronoun ‫ אתי‬appears as the Kethiv in the following IH passages (Rendsburg 2003a: 11–12), with the speaker indicated in parentheses: Judg 17:2K (Micah of Ephraim), 1 Kgs 14:2K (Jeroboam I), 2 Kgs 4:16K, 8:1K (Elisha), and 2 Kgs 4:23K (husband of the Shunammite woman). The presumed pronunciation of this form is ‫א ִַּתי‬, which corresponds well with the Samaritan pronunciation of the 2fs independent pronoun ‫( אתי‬written thus, with yod) åtti (Ben-Ḥayyim 2000: 226). So, while we lack an explicit attestation of this form in an IH composition from the postmonarchic or Persian period, the tradition maintained by the Samaritans confirms the continuation of this feature into the fifth century b.c.e. (and beyond) in the territory that was once the heart of the Northern Kingdom of Israel. 4.2. Feminine-Singular Demonstrative Pronoun ‫זֹו‬/ ‫זֺה‬ The next relevant item is the fs demonstrative pronoun ‫זֹו‬/ ‫ זֺה‬, which appears in the following northern texts (Rendsburg 2003a:13): 2 Kgs 6:19 ‫;זֹה‬ Hos 7:16 ‫ ;זוֹ‬Ps 132:12 ‫ ;זוֹ‬Qoheleth (6×) ‫זֹה‬. Once more, the attestations span centuries, in this case, from the early-monarchic-period Elisha narrative until the Persian-period book of Qoheleth, with two instances in the interval. 4.3.  Infinitive Absolute in Place of the Finite Verb The use of infinitive absolute in place of the finite verb is attested as early as Byblos Amarna and Ugaritic (both fourteenth century b.c.e.), as well as in Phoenician (especially the Karatepe inscription, eighth century). One may, accordingly, assume an early appearance in northern Hebrew as well, even if this usage is not attested until IH texts dating to the early monarchic period (and not in Judges 5, for example). The relevant cases are the following (Rendsburg 2003a: 22): Lev 25:14 ָ‫יתך‬ ֶ ‫ֲמ‬ ִ ‫‘ אוֹ ָקנֹה ִמּיַד ע‬or buy from the hand of your friend’; Judg 7:19 ‫‘ ְונָפֹוץ ַהּכ ִַּדים‬and they shattered the jugs’; 1 Kgs 22:30 10.  See above, n. 3.

346

Gary A. Rendsburg

‫‘ ִה ְת ַחּפֵׂש וָבֹא ב ִַּמ ְל ָחמָה‬I will disguise myself and go into the battle’; 2 Kgs 3:16 ‫‘ עָׂשֹה ַהּנַחַל ַהּזֶה ּג ִֵבים ּג ִֵבים‬I will make this wadi full of pools’; 2 Kgs 4:43 ‫אָכֹל‬ ‫‘ ְוהֹותֵ ר‬they shall eat and have some left over’; Amos 4:5 ‫ּתֹודה‬ ָ ‫מ ָחמֵץ‬ ֽ ֵ ‫‘ ְוקַ ּטֵ ר‬and ְ ‫‘ ה‬the wicked are burn a todah-offering from leaven’; Prov 12:7 ‫ׁש ִעים‬ ָ ‫ָפֹוך ְר‬ overthrown’; Prov 15:22 ‫ֲׁשבֹות‬ ָ ‫‘ ָהפֵר ַמח‬plans are undone’; Prov 17:12 ‫ּפָגֹוׁש ּדֹב‬ ‫(‘ ׁשַ ּכּול ְּב ִאיׁש‬better) that a man meet a bereaved bear . . .’; Neh 9:8 ‫ְוכָרֹות ִעּמֹו‬ ‫‘ ה ְַּב ִרית‬and you made a covenant with him’; Neh 9:13 ‫ּׁש ָמיִם‬ ָ ‫‘ ְודַ ּבֵר ִע ָּמהֶם ִמ‬and you spoke with them from heaven’. Again, observe how a particular IH feature is attested during a time span that bridges early texts, such as Leviticus 25 and Judges 7, and late texts, such as Nehemiah 9,with ample attestations from Kings, Amos, and Proverbs in the interval. 4.4.  Interrogative Series ‫ אוֹ‬. . . -ֲ‫ה‬ The syntagma ‫ אוֹ‬. . . -ֲ‫ ה‬used to mark two successive questions occurs in the following passages (Rendsburg 2003a: 24): Judg 18:19 ‫ֱיֹותךָ כֹהֵן ְלבֵית‬ ְ ‫הֲטֹוב ה‬ ‫ִׂש ָראֵל‬ ְ ‫ׁש ָּפחָה ְּבי‬ ְ ‫ּול ִמ‬ ְ ‫ׁשבֶט‬ ֵ ‫ֱיֹותךָ כֹהֵן ְל‬ ְ ‫‘ ִאיׁש ֶאחָד אֹו ה‬Is it better for you to be a priest to the house of one person, or to be priest to a tribe and clan in Israel?’ (Danites ְ ‫ָך אֶל־ ַה ֶּמל‬ ְ ‫הי ֵׁש ְלדַ ּבֶר־ל‬ to Micah of Ephraim); 2 Kgs 4:13 ‫ֶך אֹו אֶל־ׂשַ ר ַה ָּצבָא‬ ֲ ‘Can someone speak on your behalf to the king or to the commander of the army?’ (Elisha to Shunammite woman); 2 Kgs 6:27 ‫ָקב‬ ֶ ‫ֲמן־הַּגֹרֶן אֹו ִמן־ ַהּי‬ ִ ‫‘ ה‬From the threshing floor or from the wine press?’ (Jehoram, king of Israel, to woman of Samaria); Qoh 2:19 ‫ִהיֶה אֹו ָסכָל‬ ְ ‫‘ ֶה ָֽחכָם י‬. . . whether he will be a wise man or a fool?’; Qoh 11:6 ‫הזֶה אֹו־זֶה‬ ֲ ‘. . . whether this-one or that-one?’. In this instance, the same usage is attested from the period of the early monarchy (Judges, Elisha) through the late period (Qoheleth)—as well as in the Deir ʿAlla inscription (II 9), which is situated chronologically more or less in the middle of these two extremes. 4.5.  T-Stem to Express Passive Voice The use of the T-stem verb to express the passive voice occurs widely in Aramaic (especially since the N-stem is lacking in all dialects and varieties of the language). The parallel use of the Hithpael (and related verbal patterns) in Hebrew is a feature of IH; hence, this grammatical trait constitutes an excellent example of an isogloss shared by IH and Aramaic. Note the following examples (Rendsburg 2003a: 18–19): 11 Mic 6:16 ‫ָמִרי‬ ְ ‫ִׁשּתַ ּמֵר חֻּקֹות ע‬ ְ ‫‘ ְוי‬and the laws of Omri are observed’; Prov 31:30 ‫‘ ִהיא ִת ְת ַהּלָל‬she is to be praised’; Qoh 8:10 ‫ָעיר‬ ִ ‫ִׁשּתַ ְּכחּו ב‬ ְ ‫‘ ְוי‬and they are forgotten in the city’. 11.  For some additional examples, including examples that are not necessarily in IH texts, see Talshir 2003: 275. However, see the response and further explication by Moyer 2009: 103–5.

Northern Hebrew through Time

347

An additional relevant example appears in the Balaam oracles, with their heavy Aramaic tinge: Num 23:9 ‫ַּׁשב‬ ָ ‫ִתח‬ ְ ‫‘ ּובַּגֹויִם לֹא י‬and among the nations is not to be reckoned’. 12 If one includes the Numbers passage, then the feature treated here spans the early monarchic period, the middle monarchic period, and the Persian era, demonstrating once more how an IH trait persists over the course of centuries. In JH, by contrast, the T-stem serves only for the reflexive and the reciprocal. 5.  Lexical Items There are also numerous lexical items that bridge the centuries of IH compositions. Three examples, from among many, 13 will suffice: (a) ‫‘ יֶרַ ח‬month’ (Deut 33:14, 2 Kgs 15:13; contrast JH ‫( ;)ח ֶֹדשׁ‬b) ‫‘ ּכַד‬jug’ (Judges 7 [4×]; 1 Kgs 17:12, 17:14, 17:16; 1 Kgs 18:34; Qoh 12:6; perhaps contrast other JH terms for vessels of wine, water, etc., 14 such as ‫ ָג ִּבי ַע‬and ‫( ַּב ְקּבֻק‬see also Genesis 24 [9×], where ‫ ּכַד‬serves to enhance the style-switching technique in a story set in Aram)—in addition to the Tel Kinneret jar fragment inscription: ‫ ;כד השער‬and (c) ‫‘ מַּתַ ת‬gift’ (1 Kgs 13:7; Prov 25:14; Qoh 3:13, 5:18; contrast JH ‫ַּתנָה‬ ָ ‫)מ‬. 6.  Mishnaic Hebrew Grammatical Features Until this point, I have concentrated solely on the biblical evidence. As noted at the outset, however, the northern dialect of ancient Hebrew endured into Roman times, in the form known as Mishnaic Hebrew (after the foundational text) or Tannaitic Hebrew (after the individuals who created the corpus; for the major texts, see above, n. 2). Accordingly, I now turn our attention to this variety of ancient Hebrew, with a special focus on the features that span both the earlier IH and the later MH. Of the IH features canvassed thus far, four are characteristic of MH as well: (a) the relative pronoun -‫שׁ‬ ֶ (§3.3); (b) the fs demonstrative pronoun ‫זֹו‬/‫( זֺה‬with the spelling ‫ זֹו‬used in MH) (§4.2); (c) the Tstem to express the passive (§4.5); and (d) the lexeme ‫‘ ּכַד‬jug’ (one of dozens that could be cited; see §5). Many more traits that link IH and MH augment the picture. In the remainder of this section (§6) I present four additional grammatical features, with each subsection presenting first the IH data and then the MH examples; in the following section (§7), I offer a sampling of relevant lexical items. 12.  The many Aramaic-like features in Numbers 22–24 serve collectively as part of the overall style-switching technique. See briefly Rendsburg 2006, and in much greater detail Moyer 2009: 77–192. 13.  See the long list collected in Rendsburg 2003a: 25–31. 14.  Note the comment by Davila (1990: 86): “It [that is, ‫‘ ּכַד‬jug’] may be a word peculiar to the northern dialect, although I cannot suggest a southern synonym for it.”

348

Gary A. Rendsburg

6.1. The ‫‘ זֶה הַּיֹום‬This Day’ Construction In this syntagma (attested also in Phoenician and Aramaic), the demonstrative pronoun appears before the noun, though the sense is not ‘this is the day’ but simply ‘this day’. Examples in IH texts are as follows (Rendsburg 2003a: 21): 1 Kgs 14:14 ‫‘ זֶה הַּיֹום‬this day’; 2 Kgs 6:33 ‫ָרעָה‬ ָ ‫‘ זֹאת ה‬this evil’; Song 7:8 ‫‘ זֹאת קֹֽומָתֵ ְך‬this your stature’. The construction appears in MH as well (Segal 1936: 51; Pérez Fernández 1999: 23), though at times the definite article is omitted. 15 Thus, for example: 16 m. Ketub. 4:8 ‫‘ זֶה ִמ ְד ָרשׁ‬this midrash’; m. Naz. 3:7 (2×) ‫ידים‬ ִ ‫‘ אֵּלוּ ְמ ִע‬these testifiers’; 17 m. Naz. 7:2 ‫‘ אֵיּלוּ ה ְַּטמָאוֹת‬these impurities’. 6.2.  Third-Person Feminine-Singular Suffix-Conjugation III-y Verbs Ending in -at In contrast to 3fs forms of III-y verbs in SBH, such as ‫ְתה‬ ָ ‫( ָּבנ‬to use a paradigm verb), IH attests forms that end in -at, as follows (Rendsburg 2003a: 16): Lev 25:21 ‫ָׂשת‬ ָ ‫‘ ְוע‬and it [i.e., the land] shall do [i.e., produce]’; 2 Kgs 9:37 Kethiv ‫‘ והית‬and it [i.e., the carcass] shall be’. An additional example is to found in the Siloam Tunnel inscription (line 3), written by a recent émigré from the region of southern Samaria to Jerusalem (Rendsburg and Schniedewind 2010): ‫‘ הית‬it [i.e., the fissure] was’. This trait is one of the defining characteristics of MH (Kutscher 1982:128), with the following passages serving as illustrative examples only (see further Segal 1936: 152; and Haneman 1980: 342–43): m. Ketub. 5:4 ‫שׁים‬ ִ ‫ֳד‬ ָ ‫ּשׁה ח‬ ָ‫שׁ‬ ִ ‫עָשַׂ ת‬ ‫‘ ִל ְפנֵי ַה ַּבעַל‬she did [i.e., lived] six months with the husband’; m. Qidd. 2:7 (2×) ‫יעית ָהיָית‬ ִ ‫ּושׁ ִב‬ ְ ‫ּשׁ ָּלהֶן ָהיָית‬ ֶ ‫‘[ ִמ‬it’, i.e., the basket of figs] was [i.e., belonged] to them, and it was seventh-year produce’; m. Nid. 1:4 ‫שּׁלֹא ָראַת ָּדם ִמּיָמֶי ָה‬ ֶ ‫‘ ּכָל‬she who has not seen blood in her days’. 6.3.  The Double Plural Construction Whereas in SBH, a plural construct chain is comprised of a plural nomen regens and a singular nomen rectum (such as Exod 19:8–9 ‫‘ ִּד ְברֵ י ָהעָם‬the words of the people’), both IH and MH also employ the double plural construction, in which both nouns appear in the plural. IH examples include (Rendsburg 15.  The lack of the definite article in two of these examples is part of a general tendency in MH for less frequent use of the definite article, in comparison with BH (see, for example, the constructions cited in Segal 1927: 201, e.g., m. Šebu. 3:7 ‫‘ ִּכּכַר זוֹ‬this loaf’; m. Menaḥ. 13:9 ‫‘ שׁוֹר זֶה‬this bull’). I attribute this phenomenon to influence from the spoken dialect, though an isogloss with Phoenician may also be at play. 16.  Here and below, readings and numerations follow ms Kaufmann A50 (Budapest), except that for the sake of simplicity, I have elected to omit the rafe mark even where the manuscript includes it. 17.  To be completely accurate, on the first occasion, the demonstrative pronoun is written plene as ‫אֵיּלוּ‬, as in m. Naz. 7:2.

Northern Hebrew through Time

349

2003a: 21): 2 Kgs 15:25 ‫ָדים‬ ִ ‫‘ ְּבנֵי ִג ְלע‬children of the Gileadites’ > ‘the people of Gilead’; Ps 47:10 ‫‘ נ ְִדיבֵי ע ִַּמים‬nobles of the peoples’ > ‘nobles of the people’; Ps 74:13 ‫‘ ָראׁשֵי תַ ּנִינִים‬heads of the Tanninim’ > ‘heads of the Tannin’. MH examples include (see Kutscher 1982: 129): m. Roš Haš. 1:1 ‫שׁנִים‬ ָ ‫ָראשֵׁי‬ ‘heads of the years’ > ‘heads of the year’ (i.e., New Year festivals); m. Ter. 11:10 ‫‘ ְּבבָּתֵ י ְכנֵסָיוֹת‬in the houses of assemblies’ > ‘in the houses of assembly’ (i.e., synagogues); m. Ter. 11:10 ‫וּבבָּתֵ י ִמ ְד ָרשׁוֹת‬ ְ ‘and in the houses of studies’ > ‘and in the houses of study’ (i.e., academies). 6.4.  Passive Participle with Active Voice On two occasions in the Bible, the passive participle is used with active voice; both of these passages occur in IH texts (Rendsburg 2003a: 22): 2 Kgs 6:9 ‫ְח ִּתים‬ ִ ‫‘ נ‬descending’; Song 3:8 ‫א ֻחז ֵי ֶחרֶב‬ ֲ ‘grasping the sword’. This usage is more widespread in MH (Segal 1927: 160–61; 1936: 133–34; Pérez Fernández 1999: 139–40), as indicated by the following illustrative passages (for the most detailed treatment, see Blau 1953): m. Peʾah 2:6 ‫אנִי‬ ֲ ‫ְמקוּ ַּבל‬ ‫ָישׁא‬ ָ ‫ֶּבי ְמי‬ ִ ‫‘ ֵמר‬I received (it) from Rabbi Meyasha’; m. Ketub. 2:10 ‫ֵשׁת‬ ֶ ‫בא‬ ּ ְ ‫אנִי‬ ֲ ‫זָכוּר‬ ‫‘ ְּפלוֹנִי‬I recall (a woman), wife of so-and-so’ 18; m. B. Meṣiʿa 1:3 ‫ָהיָה ָרוכֻב עַל‬ ‫‘ ַ ּגּבֵי ְב ֵהמָה‬he was riding on the back of the beast’; 19 m. Kelim 1:14 ‫ִיכנַס‬ ְ ‫שׁאֵין נ‬ ֶ ‫ּשׁם ֶאּלָא ְרחוּץ יָדַ יִים ְורַ ְג ַליִם‬ ָ ‫‘ ַל‬that one does not enter there unless he has washed (his) hands and feet’. 7.  Mishnaic Hebrew Lexical Traits There are also many lexical links shared by IH and MH, with the following list (with first nouns and then verbs) merely illustrative (Rendsburg 2003a: 25–31; 2003b). The mishnaic references represent the first attestation in the canonical order of the tractates. In some cases, the lexical items listed here are quite common; ‫לוֹחית‬ ִ ‫‘ ְצ‬dish’, for example, occurs 43 times in the Tannaitic corpus. (a) ‫‘ ֵספֶל‬bowl’ (Judg 5:25, 6:38; m. Sukkah 4:9); (b) ‫‘ ַעיִר‬village’ (Judg 10:4; m. Demai 5:7 ‫ ;)עֲיָירוֹת‬20 (c) ‫‘ ּפַּקוּעוֹת‬wild gourds’ (2 Kgs 4:39; m. Šabb. 2:2); (d) ‫ּבּורים‬ ִ ‫‘ ִצ‬piles, heaps’ (2 Kgs 10:8; m. Ber. 5:5); 21 (e) ‫לוֹחית‬ ִ ‫‘ ְצ‬dish’ (2 Kgs 2:20; m. Šabb. 8:2); (f) ‫‘ ִצּנָה‬cold’ (Prov 25:13; m. Yoma 3:5); (g) ‫קַ ב‬ ‘qab (unit of measurement)’ (2 Kgs 6:25; m. Peʾah 6:1); (h) ‫ה‬-‫ר‬-‫‘ א‬pluck’ 18.  In this case, I cite the reading of ms Parma A (de Rossi 138), which is clearer and well executed. In ms Kaufmann, this passage is written in the margin by the vocalizer—and thus the reading is imperfect: ‫ישׁ פלוני‬ ְ ‫ֵשׁת ִא‬ ֶ ‫ִיתי ְבא‬ ִ ‫זָכוּר {הי} ַהי‬. 19.  The vocalization as passive participle, notwithstanding the Kethiv in ms Kaufmann, is confirmed by the (unvocalized) reading ‫ היה רכוב על גבי בהמה‬in ms Parma A (de Rossi 138). 20.  The MH form is always feminine and plural, though clearly we are dealing with the same word; see Rendsburg 2000: 38–41. 21.  In this case, the connotation of the word has shifted, because in MH the basic meaning is ‘public, gathering’.

350

Gary A. Rendsburg

(Ps 80:13; Song 5:1; m. Šeb. 1:2); (i) ‫שׁ‬-‫ל‬-‫‘ ג‬flow’ (Song 4:1, 6:5; b. Pesaḥ. 37b ‘boil’); 22 ( j) ‫ב‬-‫ה‬-‫ב‬-‫‘ ה‬singe, roast’ (Hos 8:13; m. Šabb. 2:3); (k) ‫ך‬-‫ר‬-‫ח‬ ‘roast, singe’ (Prov 12:27; m. Šabb. 16:5); (l) ‫ף‬-‫נ‬-‫‘ ט‬soil, make dirty’ (Song 5:3; m. Makš. 4:5); (m) ‫ב‬-‫ב‬-‫‘ י‬whine, shrill’ (Judg 5:28; m. Roš Haš. 4:9 ‫י ַ ָּבבוֹת‬ ‘blasts, shrills’); 23 (n) ‫ל‬-‫ל‬-‫‘ מ‬crush, squeeze, rub’ (Prov 6:13; m. Maʿaś. 4:5); (o) ‫ב‬-‫צ‬-‫‘ ק‬cut’ (2 Kgs 6:6, Song 4:2; m. ʿArak. 1:3); (p) ‫ס‬-‫נ‬-‫‘ שׁ‬gird’ (1 Kgs 18:46; m. Kelim 26:1 ‫שׁנֶץ‬ ֶ ‘strap, lace, thong’). 24 8.  Constancy within the IH–MH Continuum The larger picture that emerges from all these interlocking data is the following. Throughout the history of northern Hebrew, even if our sources are at times limited and with centuries intervening, the same lexical and grammatical features are employed consistently and repeatedly. These traits must have served to distinguish the northern and southern dialects of Hebrew throughout antiquity, during both the biblical and the postbiblical periods.  25 A stellar example is afforded by the noun ‫‘ ֵספֶל‬bowl’, attested in Judg 5:25 (our oldest IH text) and Judg 6:38 (another early text), and then not again until it appears 18 times in the Tannaitic corpus. 26 Clearly, this lexeme must have been in use in northern Israel during the millennium or more that separates the 2 biblical occurrences and the 18 Tannaitic appearances, even if the sources at our disposal do not happen to use the word. The same is true for a much rarer word, the verbal root ‫‘ יבב‬whine, shrill’, which appears in Judg 5:28 as a hapax legomenon and its nominal derivative ‫‘ יַּבָבוֹת‬blasts, shrills’, and is limited to a single attestation in m. Roš Haš. 4:9. 27 The aggregate data presented in the preceding sections reveal a rather consistent northern dialect of Hebrew (the IH–MH continuum) spanning about 1,300 years. To repeat the statement adumbrated above, there appears to be little or no diachronic change reflected in the numerous lexical and grammati22.  Normally, I cite only Tannaitic material (MH1) in my research into MH, though on this occasion, quite strikingly, the key source for the continuation of ‫שׁ‬-‫ל‬-‫‘ ג‬flow’ into MH is an Amoraic (MH2) compilation. Note further that the root g-l-ṯ ‘swell, flow’ appears in Ugaritic. For full documentation, see Tuell 1993. 23.  While the root occurs as a verb in Judg 5:28, in the cited Mishnah passage, the related form is a plural noun. 24.  As per the information conveyed in n. 23, while the root occurs as a verb in 1 Kgs 18:46, in the Mishnah the lexeme occurs as a noun, with interchange between /s/ and /ṣ/ as well. Note that the verb occurs already in Ugaritic. 25.  This may be the appropriate time to recall the evidence of Matt 26:73 // Mark 14:70, which indicates that Peter’s λαλιά ‘speech’ revealed him to be a Galilean. 26.  Note also the presence of spl in Ugaritic: KTU 1.104:8, 4.123:17, 4.385:3 (del Olmo Lete and Sanmartín 2003: 766). Indeed, some Ugaritic lexemes “skip” the Bible altogether, only to emerge about a millennium and a half later in MH (see Levine 1962), representing yet another sign of the northern dialect cluster. 27.  The verb appears again in an Amoraic source, y. Yebam. 16.5 (15d).

Northern Hebrew through Time

351

cal traits specific to this regional variety of ancient Hebrew. The traits that are attested early on, even as early as Judges 5, are still present in the northern dialect of ancient Hebrew more than a millennium later. This in itself is a remarkable conclusion of our investigation, even if this volume is about diachronic change and not about diachronic stability. 9.  Diachronic Change We are led to ask, then: Can it be that the northern Hebrew dialect remained constant for such a long period of time? Especially given the upheavals of 745–721 b.c.e., followed by the presence of the Assyrians, Babylonians, Persians, Greeks, Ptolemies, Seleucids, and Romans in the land? 28 The answer is clearly no. For as anyone who knows the primary sources will readily realize, notwithstanding all that I have presented, MH is not the same as IH—and even a chapter such as Nehemiah 9 (not to mention Qoheleth) is not the same as the earlier prose material in Kings or Judges. It is to these matters that I now turn. 9.1.  Methods Developed by Frank Polak Applied to IH Texts The most important diagnostic tool (or actually set of tools) for the dating of biblical texts is the tool(s) developed by Frank Polak. Beginning with several groundbreaking studies during the 1990s and continuing to the present day (including in the present volume), Polak has created a typology of biblical prose texts that permits one to observe the development from the more “oral, verbal, simple, rhythmic” style of early narratives (that is, dated to the early monarchy) to the more “written, nominal, complex, annalistic” style of later narratives (that is, dated to the Persian period)—with several transition points between them. 29 Polak has not provided data specific to IH texts, but from within his extended analyses one may nevertheless extract, or at least extrapolate, the relevant IH data. 9.1.1.  Noun-Verb Ratios and Nominal-Finite Ratios While in his more recent studies, Polak has developed more sophisticated methods with an eye to more-detailed analysis, for our present purposes, it will be convenient merely to present the summary information conveyed in Polak 1998: 70 regarding Noun-Verb and Nominal-Finite ratios. 30 28.  This second question is crucial, for in truth a language or dialect can remain essentially the same over the course of one millennium. The most well-known example is Icelandic, which has evolved very little since its earliest written attestation in ca. 1100 c.e. However, in this case the speakers of the language were largely isolated on an island in the North Atlantic Ocean, albeit with periodic contact with their close cousins on the Scandinavian mainland. 29.  For the simplest statement, see Polak 1998: 69. For his latest contribution to the subject, see pp. 307–315 in the present volume. 30.  As the term implies, Noun-Verb ratio expresses the proportion of nouns and verbs in a given text. Nominal-Finite ratio takes the number of verbs and then subdivides them

352

Gary A. Rendsburg

The Elijah and Elisha cycles (all IH) are among the most “oral, verbal” of all biblical narratives, with both having exceedingly low Noun-Verb ratios of 0.600 and 0.589, respectively, and very low Nominal-Finite ratios of 0.133 and 0.148, respectively. These figures indicate that the chapters about Elijah/Elisha are among the earliest prose texts in the Bible. Selected chapters from the book of Judges (almost all of them with northern settings) fall into the transitional subclass; that is, the data reveal slightly higher Noun-Verb ratios, though in both cases, quite remarkably, extraordinarily low Nominal-Finite ratios. For reasons that I have not been able to ascertain, Polak lumped the Deborah-Barak and Abimelekh accounts (Judges 4, 9) together as one group, and the Ehud, Gideon, and Jephthah material (Judges 3, 6–8, 11–12) as another group. Of these, note that only Ehud is not in a purely northern (including Transjordanian) setting, so that by and large the figures for these eight chapters from Judges provide information relevant to the investigation of IH. The Noun-Verb ratios for the two groups are 0.662 and 0.671, respectively, while the Nominal-Finite ratios are 0.119 and 0.128, respectively. Representative of the next phase, pertaining to the late pre-exilic and exilic periods, is 2 Kings 11–16, about 40% of which concerns the Northern Kingdom of Israel (13:1–25, 14:11–16, 14:23–29, 15:8–31). For this material, the Noun-Verb ratio is 0.736, while the Nominal-Finite ratio is 0.190. Observe that both ratios are higher than the previously cited figures for earlier compositions. Finally, we may consider the data presented by Polak for Nehemiah 8–10, with the IH prose of Neh 9:6–37 representing about 45% of these chapters. 31 Again we observe a high Noun-Verb ratio of 0.731, this time coupled with an extremely high Nominal-Finite ratio of 0.322. Table 1 summarizes all these figures conveniently (though again, one needs to bear in mind that some of these corpora include some Judahite material). 9.1.2.  Hypotaxis in Nehemiah 9 We remain with Nehemiah 9 here to present several passages that reflect the more complex prose style, which includes the greater use of hypotaxis (that is, subordination, as indicated by particles such as ‫ ִּכי‬and relative markers such as ‫ֲׁשר‬ ֶ ‫א‬, or their English equivalents ‘for’, ‘because’, ‘that’, ‘which’, etc.). Consider these passages, for example, each of which includes double subordination: into nominal verbs (participles and infinitives) and finite verbs (suffix-conjugation, prefixconjugation, and imperatives). 31.  Polak (1998: 69) listed the following verses for his Nehemiah 8–10 pericope: Neh 7:72–10:1, 10:29–11:3 (he obviously excluded Neh 10:2–28 because these verses are all very short and are composed mainly of personal names). The IH material is Neh 9:6–37 (even though I have used the term “Nehamiah 9” above for simplicity’s sake)—that is, 32 out of the total 71 verses, or 45%.

353

Northern Hebrew through Time Table 1 Noun-Verb Ratio

NominalFinite Ratio

Elijah

.600

.133

Elisha

.589

.148

Judges 4, 9

.662

.119

Judges 3, 6–8, 11–12

.671

.128

2 Kings 11–16

.736

.190

Nehemiah 8–10

.731

.322

ְ ‫ָדיו‬ ָ ‫ּובכָל־עֲב‬ ְ ‫ו ִַּתּתֵ ן אֹתֹת ּומ ְֹפ ִתים ְבּפ ְַרעֹה‬ (1)  ​‫ּובכָל־עַם א ְַרצֹו ִכּי יָדַ ְע ָּת ִכּי ֵהזִידּו‬ ‫עֲלֵיהֶם‬ and you gave signs and wonders against Pharaoh and against all his servants and against all the people of his land, for you knew that they were-insolent against you. (Neh 9:10) ָ‫ֲשיבָם ֵאלֶיך‬ ִׁ ‫ֲשר־ה ִֵעידּו בָם ַלה‬ ֶׁ ‫ְביאֶיךָ הָרָגּו א‬ ִ ‫ְואֶת־נ‬ (2)  ​ and they killed your prophets, who admonished them in order to return them to you. (Neh 9:26) 9.1.3.  Extended Noun Groups in Nehemiah 9 Polak (2006: 128–29; 2009) has also observed the manner in which extended noun groups characterize LBH texts. In contrast to SBH, where typically even just one noun may stand alone, in LBH nouns are strung together, in imitation of the chancellery style characteristic of Persian-era administrative documents. Examples from our chapter include: ֶׁ ‫ָל־צ ָבאָם ָה ָארֶץ ְוכָל־א‬ ְ ‫ַש ַמיִם ְוכ‬ ּׁ ָ ‫ַש ַמיִם ְׁשמֵי ה‬ ּׁ ָ ‫ׂית אֶת־ה‬ ָ ‫ָש‬ ִ‫ע‬ (3)  ​‫ֲשר ָעלֶי ָה ַהי ַ ִּּמים‬ ‫ֲשר ָּבהֶם‬ ֶׁ ‫ְוכָל־א‬ And you made the heavens, and the heavens of the heavens, and all their host; the land and all that is upon it; the seas and all that is in them. (Neh 9:6) ‫ּובכָל־עַם א ְַרצֹו‬ ְ ‫ָדיו‬ ָ ‫ּובכָל־עֲב‬ ְ ‫ְּבפ ְַרעֹה‬ (4)  ​ against Pharaoh and against all his servants and against all the people of his land. (Neh 9:10)

354

Gary A. Rendsburg

‫ובים‬ ִ ֹ ‫ּומ ְצֹות ט‬ ִ ‫ֻּקים‬ ֱ ‫ְשִרים ְותֹורֹות‬ ָׁ ‫ו ִַּתּתֵ ן ָלהֶם ִמ ְׁש ָפ ִּטים י‬ ִ ‫אמֶת ח‬ (5)  ​ And you gave them just laws and true instructions, rules and good commandments. (Neh 9:13) ‫ית ָלהֶם‬ ָ ‫ֻּקים ְותֹורָה ִצִּו‬ ִ ִ ‫ּומ ְצֹוות ְוח‬ (6)  ​ And commandments and laws and instruction you commanded them. (Neh 9:14) ִ ‫ּירׁשּו ּב‬ ְ ‫ֲדמָה ְׁש ֵמנָה ַוִי‬ ָ ‫ָרים ְּבצֻרֹות ַוא‬ ִ‫ַוִי ְּל ְּכדּו ע‬ (7)  ​‫ָּתים ְמל ִֵאים־ּכָל־טּוב ּבֹרֹות‬ ‫אכָל לָרֹב‬ ֲ ‫ֵיתים ְועֵץ ַמ‬ ִ ‫ָמים ְוז‬ ִ ‫ֲצּובים ְּכר‬ ִ ‫ח‬ And they captured fortified cities and rich land; and they inherited houses full of all goods, quarried cisterns, vineyards and olive groves, and trees for eating in abundance. (Neh 9:25) ִ ‫הנֵינּו ְו ִלנ‬ ֲֹ ‫ּולכ‬ ְ ‫שׂרֵינּו‬ ָ ‫ר־מ ָצא ְַתנּו ִל ְמ ָלכֵינּו ְל‬ ְ ‫ֲש‬ ֶׁ ‫אֵת ָכּל־ה ְַּת ָלאָה א‬ (8)  ​‫ְביאֵנּו ְו ַלאֲבֹתֵ ינּו‬ ָ‫ּולכָל־ ַעּמֶך‬ ְ all the suffering that has come upon us—upon our kings, our officers, and our priests and our prophets and our fathers, and all of your people. (Neh 9:32) ָ‫ָתך‬ ֶ ‫הנֵינּו ַואֲבֹתֵ ינּו לֹא עָׂשּו ּתֹור‬ ֲֹ ּ‫שׂרֵינּו כ‬ ָ ‫ֶת־מ ָלכֵינּו‬ ְ ‫ְוא‬ (9)  ​ and our kings, our officers, our priests, and our fathers have not followed your Torah. (Neh 9:34) 9.2.  Additional LBH Features in Nehemiah 9 In addition, there are other LBH features present in Nehemiah 9 (Rendsburg 1991a: 363), such as the following: 1.  The form ‫‘ ָחיָה‬live’ (v. 29) in contrast to SBH ‫( חַי‬Hurvitz 1982: 47). 2.  ‫‘ חַּנּון ְורַ חּום‬gracious and compassionate’ (vv. 17, 31) in contrast to the reverse order of this word pair in SBH (Hurvitz 1972: 104–6). 3.  ‫‘ עַד־הָעֹולָם‬until eternity’ (v. 5) as opposed to the SBH form without the definite article: ‫( עַד־עֹולָם‬Hurvitz 1972: 158–59). 4.  Radically reduced use of nota accusativi -‫ את‬+ pronominal suffix, with an attendant increase in the pronominal suffix attached directly to the verb—the data for Nehemiah 9 are 0 of the former versus 23 of the latter (Polzin 1976: 30). 5.  Use of various plural nouns in place of their SBH singular equivalents, for example, ‫‘ ִע ִּתים‬times’ (v. 28) instead of ‫‘ עֵת‬time’ (Polzin 1976: 42).

9.3.  Toward a Conclusion The upshot of the two preceding sections (§§9.1–9.2)—one section drawing on Polak’s research, the other using earlier investigations into the nature of

Northern Hebrew through Time

355

LBH—is the following. IH does indeed change over time, but it does so in the same manner as the changes that affect the Judahite variety of LBH. In other words, whatever forces generated the shift from SBH to LBH in the larger Judahite corpus at our disposal were also at work in the northern dialect of ancient Hebrew. On the whole, however, these changes had little or no effect on the identifiable IH traits in Persian-period northern texts such as Nehemiah 9 and Qoheleth. For as we have seen above, features such as the reduplicated plural of geminate nouns (§3.1), the fs demonstrative pronoun ‫ זֺה‬/‫§( זֹו‬4.2), the use of the infinitive absolute in place of the finite verb (§4.3), the interrogative series ‫ אוֹ‬. . . -ֲ‫§( ה‬4.4), the lexeme ‫‘ ּכַד‬jug’ (§5), and many more are still present in these two representatives of late IH. One of the main forces alluded to in the previous paragraph was the major impact of Aramaic during the period of Achaemenid rule, the results of which are seen at every turn in LBH. But since, as we noted at the outset of this essay, connections with Aramaic are seen throughout the history of IH, it is more difficult to judge explicit Aramaic influence on Persian-period IH. The sources, as noted above, are limited, with Nehemiah 9 and Qoheleth being the only two texts. 32 There appears to be little or no Aramaic influence over Nehemiah 9 (note that we have referred to none in our discussion above). The question of Qoheleth, of course, is much thornier. 9.4. Qoheleth On the one hand, any single Aramaism in Qoheleth could be attributed to the provenance of the book in northern Israel. This is undoubtedly the case with examples such as ‫חֹורים‬ ִ ‘freemen’ (10:17); ‫‘ ּכַד‬jug’ (12:6); ‫‘ ְמ ִדינָה‬province’ (2:8, 5:7); and ‫‘ שׁוּק‬street’ (12:4–5), all of which are attested in earlier IH sources (1 Kings 17–18, 20–21; Judges 7; Proverbs; Song of Songs). On the other hand, the frequency of Aramaisms in Qoheleth far outstrips the incidence of IH-Aramaic isoglosses in other Israelian sources—even if this statement is not validated by statistical analysis here (for the same judgment, see Seow 1996: 650–54). In light of the increased frequency of Aramaisms in Qoheleth, especially given the overlap between many of these usages and the Aramaisms found in Imperial Aramaic sources (again, see Seow 1996: 650–54), one is fully justified in dating the book to the Persian period. Naturally, not every linguistic peculiarity in the book of Qoheleth is to be ascribed to late usage. Any number of grammatical traits may be due to the idiolect of the writer, who may have chosen to write in a (chatty?) personal style to convey to his readership his personal musings. Nonetheless, LBH influence plainly is present, and in this way Qoheleth reflects diachronic change, especially in contrast to earlier northern compositions.

32.  Though some, to be sure, would add the Song of Songs here; see above, §2.

356

Gary A. Rendsburg

9.5.  Mishnaic Hebrew Finally, we return to the issue of MH. As we have seen above, the language of the Tannaitic sources reflects a continuation of the IH dialect used during the Iron Age and the Persian period. There are, as I have demonstrated, sufficient lexical and grammatical continuities between the two corpora (IH texts and MH sources) to justify this claim. At the same time, however, MH represents a sharp break with the earlier literature. For, while the biblical IH texts are written in the literary standard (with the exception of Qoheleth), the rabbis elected to fashion their compilations in a more colloquial Hebrew. Features of MH—such as (a) gender neutralization, via the use of epicene forms for 2p and 3p pronouns and verbs; (b) the expression of the adjectival clause without the definite article on the noun, for example, ‫‘ נֶפֶשׁ ַהּיָפָה‬the good appetite’ (m. Ḥul. 4:7); (c) the use of the independent possessive pronoun ‫שׁל‬ ֶ ‘of’ (even if it is prefixed frequently, especially in the manuscripts—see the next example); and (d) the anticipatory pronominal suffix, as in ‫שׁ ֶּל ָענִי‬ ֶ ‫‘ יָדוֹ‬the hand of (him) the poor man’ (m. Šabb. 1:1)—are attested in other spoken varieties of Semitic. These features thus point the way toward our considering MH to be a colloquial dialect (Rendsburg 1991b). Presumably, this register was used out of convenience and expediency, since the contents of the Mishnah, Tosefta, and related texts emerged from oral discussions held by the rabbis in the yeshivot, discussions that no doubt took place in a spoken variety of Hebrew.  33 10. Conclusion In this article, I have dealt with a series of texts bridging approximately 1,300 years. We have seen a remarkable amount of constancy in the IH–MH continuum, with certain lexical and grammatical features attested at both extremes of the chronological range. This finding may at first seem surprising, though parallels are known from other languages. Of the hundreds of regional words in English that have persisted for centuries, I cite here but one. The word mistal ‘cow shed’ is first attested in the Depositions of the Castle of York (1673) in the following context: “Henry Cordingley, of Tonge, saith, that .  .  . he sawe the said Mary Sikes riding upon the backe of one of his cowes. And he endeavouring to strike att her stumbled and soe the said Mary flewe out of his mistall window.” The word continues to be used 300 years later in Yorkshire, as exemplified by its occurrence in the crime novel Night Is a Time to Die (1972), written by John Wainwright, a native of Leeds who served 20 years as a police officer in the West Riding Constabulary, Yorkshire: “The cows were waiting in the mistal” (p. 7). 34 But notwithstanding the use of 33.  One should note that the variety of Hebrew present in other texts, such as 3Q15, 4QMMT, and the Bar Kokhba letters, is quite close to MH. 34. Citations from the Oxford English Dictionary, s.v. mistal. For the complete text of the former citation, see also http://www.archive.org/stream/depositionsfromc00grea/ depositionsfromc00grea_djvu.txt.

Northern Hebrew through Time

357

mistal in both texts, no one would argue that late seventeenth-century English is the same as contemporary English. Indeed, the same changes that affected English throughout Great Britain and beyond (for example, the shift from saith to says) occured in Yorkshire English, notwithstanding its very identifiable dialect, even as mistal has endured and still may be heard in the county. And so it is with IH: certain features, both lexical and grammatical, persisted for centuries, even for more than a millennium when one introduces MH into the picture; nonetheless, profound changes occurred. These linguistic changes, to summarize the above exposition, result from two main thrusts: (a) the influence of LBH (that is, more or less the same LBH witnessed in Judean sources) evident in Nehemiah 9 and Qoheleth, both dated to the Persian period; and (b) the major shift in mode of expression among the rabbis, with the result that MH reflects a colloquial variety that was used even in writing. References

Aḥituv, S. 2008 Echoes from the Past: Hebrew and Cognate Inscriptions from the Biblical Period. Jerusalem: Carta. Ben-Ḥayyim, Z. 2000 A Grammar of Samaritan Hebrew. Jerusalem: Magnes / Winona Lake, IN: Eisenbrauns. Blau, J. 1953 Benoni Paʿul be-Horaʾa ʾAqtivit. Leš 18 (5713): 67–81. Davila, J. R. 1990 Qoheleth and Northern Hebrew. Pp. 69–87 in Sopher Mahir: Northwest Semitic Studies Presented to Stanislav Segert (= Maarav 5–6), ed. E. Cook. Winona Lake, IN: Eisenbrauns. Haneman, G. 1980 Torat ha-Ṣurot šel ha-Mišna: ʿAl Pi Masoret Ketav-Yad Parma (De Rossi 138). Tel-Aviv: Tel-Aviv University Press. Hurvitz, A. 1972 Ben Lashon le-Lashon. Jerusalem: Bialik. 1982 A Linguistic Study of the Relationship between the Priestly Source and the Book of Ezekiel: A New Approach to an Old Problem. CahRB 20. Paris: Gabalda. Japhet, S. 1983 People and Land in the Restoration Period. Pp. 103–25 in Das Land Israel in biblischer Zeit, ed. G. Strecker. Göttingen: Vandenhoeck & Ruprecht. Kutscher, E. Y. 1982 A History of the Hebrew Language. Jerusalem: Magnes / Leiden: Brill. Levine, B. A. 1962 Survivals of Ancient Canaanite in the Mishnah. Ph.D. dissertation, Brandeis University.

358

Gary A. Rendsburg

Moyer, C. J. 2009 Literary and Linguistic Studies in Sefer Bilʿam (Numbers 22–24). Ph.D. dissertation, Cornell University. Noegel, S. B., and Rendsburg, G. A. 2010 Solomon’s Vineyard: Literary and Linguistic Studies in the Song of Songs. SBL Ancient Israel and Its Literature. Atlanta: Society of Biblical Literature / Leiden: Brill. Olmo Lete, G. del, and Sanmartín, J. 2003 A Dictionary of the Ugaritic Language, ed. and trans. W. G. E. Watson. 2 vols. Handbook of Oriental Studies 1/67. Leiden: Brill. Pérez Fernández, M. 1999 An Introductory Grammar of Rabbinic Hebrew. Leiden: Brill. Polak, F. H. 1998 The Oral and the Written: Syntax, Stylistics and the Development of Biblical Prose Narrative. JANES 26: 59–105. 2006 Sociolinguistics: A Key to the Typology and the Social Background of Biblical Hebrew. HS 47: 115–62. 2009 Parallelism and Noun Groups in Prophetic Poetry from the Persian Era. Pp. 199–235 in A Palimpsest: Rhetoric, Ideology, Stylistics, and Language Relating to Persian Israel, ed. E. Ben Zvi, D. Edelman, and F. Polak. Perspectives on Hebrew Scriptures and Contexts 5. Piscataway, NJ: Gorgias. Polzin, R. 1976 Late Biblical Hebrew: Toward an Historical Typology of Biblical Hebrew Prose. HSM 12. Missoula, MT: Scholars Press. Rendsburg, G. A. 1990 Diglossia in Ancient Hebrew. AOS 72. New Haven: American Oriental Society. 1991a The Northern Origin of Nehemiah 9. Bib 72: 348–66. 1991b Parallel Developments in Mishnaic Hebrew, Colloquial Arabic, and Other Varieties of Spoken Semitic. Pp. 1265–77 in Semitic Studies in Honor of Wolf Leslau on the Occasion of His Eighty-Fifth Birthday, November 14th, 1991, ed. Alan S. Kaye. 2 vols. Wiesbaden: Harrassowitz. 1992 The Galilean Background of Mishnaic Hebrew. Pp. 225–40 in The Galilee in Late Antiquity, ed. Lee I. Levine. New York: Jewish Theological Seminary. 1993 The Dialect of the Deir ʿAlla Inscription. BO 50: 309–29. 1996 Linguistic Variation and the “Foreign” Factor in the Hebrew Bible. IOS 15: 177–90. 2000 Notes on Israelian Hebrew (II). JNSL 26: 33–45. 2003a A Comprehensive Guide to Israelian Hebrew: Grammar and Lexicon. Orient 38: 5–35. 2003b The Geographical and Historical Background of the Mishnaic Hebrew Lexicon. Orient 38: 105–15. 2006 Aramaic-Like Features in the Pentateuch. HS 47: 163–76. 2008 Qetaʿ Shenat ha-Yovel (Vayyiqraʾ 25:8–24) ke-Ḥibbur Ṣefoni. Meḥqarim beLashon 11–12 (= Sefer ha-Yovel le-Avi Hurvitz): 297–308.

Northern Hebrew through Time

359

2009 Israelian Hebrew Features in Deuteronomy 33. Pp. 167–83 in Mishneh Todah: Studies in Deuteronomy and Its Cultural Environment in Honor of Jeffrey H. Tigay, ed. N. S. Fox, D. A. Glatt-Gilad, and M. J. Williams. Winona Lake, IN: Eisenbrauns. Rendsburg, G. A., and Schniedewind, W. M. 2010 The Siloam Tunnel Inscription: Historical and Linguistic Perspectives. IEJ 60: 188–203. Rofé, A. 1988 The Vineyard of Naboth: The Origin and Message of the Story. VT 38: 89–104. Segal, M. H. (M. Z.) 1927 A Grammar of Mishnaic Hebrew. Oxford: Clarendon. 1936 Diqduq Leshon ha-Mishna. Tel-Aviv: Devir. Seow, C. L. 1996 Linguistic Evidence and the Dating of Qohelet. JBL 115: 643–66. Stemberger, Günter 1996 Introduction to the Talmud and Midrash. Edinburgh: T. & T. Clark. Talshir, D. 2003 The Habitat and History of Hebrew during the Second Temple Period. Pp.  251–75 in Biblical Hebrew: Studies in Chronology and Typology, ed. I. Young. London: T. & T. Clark. Tuell, S. S. 1993 A Riddle Resolved by an Enigma: Hebrew ‫ גלׁש‬and Ugaritic glṯ. JBL 112: 99–104. Young, I.; Rezetko, R.; and Ehrensvärd, M. 2008 Linguistic Dating of Biblical Texts. 2 vols. London: Equinox.

Diachrony in Biblical Hebrew Lexicography and Its Ramifications for Textual Analysis Chaim Cohen 1. Introduction The purpose of this essay is to demonstrate how the diachronic method goes hand-in-hand with proper textual analysis including occasionally, when absolutely justified, minor textual emendations of the MT (even when no alternative Hebrew version is extant). Two cases, each illustrating different aspects of proper diachronic analysis, are presented in this article. The first is a longer textual example based on this sort of minor textual emendation; the second is much shorter and provides an obvious demonstration of how and why even the best scholars may sometimes err. Both will be discussed in conjunction with the following basic diachronic principles for dating texts in Biblical Hebrew. (a) Scholars conventionally divide Hebrew in the Bible into three periods: Ancient/Archaic Biblical Hebrew (ABH), Standard Biblical Hebrew (SBH), and Late Biblical Hebrew (LBH; see, for example, Kutscher 1982: 12; Hurvitz 2002: 35; Zevit 2005: 323; and most recently Blau 2010: 7–9 [§1.3]). 1 While Hurvitz has emphasized the validity of this chronological division especially for the purpose of pinpointing the elements that may be considered identifying features of LBH (as opposed to SBH), 2 in the present study, I deal with some specific identifying features of all three periods. 3 1.  Here it should be noted that the competing, essentially bipartite nomenclature that emphasizes mainly EBH and LBH (see Young, Rezetko, and Ehrensvärd 2008: passim [despite p. 8 and chap. 12]) is less appropriate because it assumes a priori that ABH (usually considered part of EBH there) is not worthy of a separate, equally exclusive designation. Note also the indecision expressed by Hurvitz (2000: 145 n. 9) on this issue, which was later replaced with his renewed acceptance of the aforementioned tripartite division (Hurvitz 2002: 35). Finally, see also the discussions of this issue in Rendsburg 1991: 81–83; Zevit 2005: 323. And, for abbreviations unique to this essay, see pp. 373–375 below. 2.  See the recent excellent summary of A. Hurvitz’s major contributions in this area in Young, Rezetko, and Ehrensvärd 2008: 1.12–23 (§2.2) and a comprehensive listing of his relevant publications in Young, Rezetko, and Ehrensvärd 2008: 2.242–45 (66 items from 1961 to 2007). 3.  Isolating identifying features from each period is the best way to demonstrate the separate existence of the three periods from a linguistic point of view. Such features may

361

362

Chaim Cohen

(b) Archaic features characteristic of ABH may surely be used later on as well but generally much less frequently or to a much lesser degree, 4 unless there is reason to assume intentional archaizing (as in the book of Job)  5 or in the few cases (called “mirage forms” in Kutscher 1982: 38–39 [§54]) where it can be demonstrated that the archaic feature was retained in Aramaic and reintroduced into BH under Aramaic influence as a feature of a later dialect (usually LBH). 6 No less significant in characterizing ABH is the absence of also occur later on but much less frequently or under special circumstances. With regard to the specific case of ABH, see paragraph (b) above. 4.  See, for example, the discussion below (§2.1) with regard to the almost total absence of the nota accusativi in ABH, for which additional corroborating evidence is provided by the existence of this same phenomenon in the Old Byblian dialect of Phoenician (KAI 1–8), which dates to the eleventh–tenth centuries b.c.e. 5.  See, for example, the detailed discussion in Young, Rezetko, and Ehrensvärd 2008: 1.333–34, 2.53–56, and the bibliography cited there. One fascinating case of intentional archaizing that does not seem to have been fully understood previously is the usage of the archaic terminology ‫ האתנים‬/ ‫ בול‬/ ‫ בירח זו‬in 1 Kgs 6:37, 38; 8:2, referring in all three verses to the date of the completion and dedication of the First Temple by King Solomon in the tenth century b.c.e. Thus the author (writing after the destruction of the First Temple in 587/6 b.c.e., which itself is referred to in 2 Kings 25) intentionally used the three early Hebrew month names together with the archaic Poetically Semantic Equivalent B-word ‫ירח‬ ‘month’; the common, everyday term is the synonymous A-word ‫חודש‬. (For Poetic Semantically Equivalent B-words and their corresponding A-words, see C. Cohen 2008: 458–64.) It may be assumed that this archaizing is the result of the author’s quoting from historical documentary sources (contemporaneous with the events described therein) or from oral tradition, where of course the designation of the month names as well as the archaic term for ‘month’, ‫ירח‬, would have been in accordance with the early First Temple BH/Phoenician calendar. As a result, the later author of the book of Kings felt a need to provide a double explanatory gloss in two of the three aforementioned verses as follows: ‫ הוא החודש‬. . . ‫בול‬/‫בירח האתנים‬ ‫השמיני‬/‫‘ השביעי‬in the month of ‫בול‬/‫ האתנים‬. . . that is, the seventh/eighth month’ (1 Kgs 6:38; 8:2). On these explanatory glosses, see, for example, Avishur 1979: 142; Hurowitz 1992: 228–29; 1994: 69–70; Mulder 1998: 230–32, 283, 379–80; Cogan 2001: 236–37, 247, 278. Compare with Rendsburg 2002: 127–28. Only in the case of the early Hebrew month of ‫ זו‬did the author deviate somewhat from this formula, using the phrase ‫ בירח זו‬in 1 Kgs 6:37 without any explanatory gloss, but in 1 Kgs 6:1 he provided the explanation: ‫חודש זו הוא‬ ‫‘ החודש השני‬the month of ‫—זו‬that is, the second month’. In all of the other four BH verses in which we find an early Hebrew month name, all referring to the month of ‫( אביב‬Exod 13:4; 23:15; 34:18; Deut 16:1 [2×]), the phrase used is ‫‘ חודש האביב‬the month of ‫( ’אביב‬without any explanatory gloss). The only plausible explanation for this distinction is that these four verses are all from the Torah literature and were therefore naturally written (without archaizing) in the First Temple SBH dialect, when the early Hebrew month names were still in use (the later Babylonian month names were first used in LBH), but when the archaic term for ‘month’, ‫ירח‬, had already been replaced by the everyday term ‫חודש‬. Contrast Cassuto 1975: 224–25 and nn. 44 and 45. 6.  A good example of these “mirage forms” is the twice attested 2fs form ‫‘ ׁשַ ּקַ ְמ ִּתי‬that you arose’ in the ABH Song of Deborah (Judg 5:7). The ‫ּתי‬ִ suffix for 2fs forms (instead of the regular ‫ּת‬ְ suffix) is elsewhere attested (most often as the Kethiv form) especially in the

Diachrony in Biblical Hebrew Lexicography

363

regular SBH and LBH features (such as the nota accusativi ‫ את‬discussed below in §2.1) in the contexts where they would be expected had the text been written in one of the later periods. Compare, for example, Exod 15:4 (ABH without nota accusativi), with reference to the Egyptian cavalry and chariot corps as definite direct objects of the verb, with Exod 14:6, 25, 28 in SBH, with ‫ ;את‬see also Judg 4:7, 13, 15). (c) Identifying features of LBH are the features that are more frequent in LBH than in the previous periods. While extrabiblical data from Aramaic may often provide key evidence in addition to what may be learned from internal BH evidence, Akkadian evidence (which may sometimes contradict the Aramaic evidence) should never be overlooked or underestimated.  7 (d) Although internal Biblical Hebrew evidence must always be considered paramount, additional corroborating evidence from extrabiblical sources is often of crucial importance for making the case. Such additional evidence may come from extrabiblical Hebrew inscriptions dating to the First Temple period, from Rabbinic Hebrew, from the book of Ben Sira and the Hebrew Dead Sea Scrolls dating to the Second Temple period, or from the various dialects of the other ancient Semitic languages dating from the biblical period. One important source that is especially significant for certain identifying features of ABH but that has so far been neglected in this regard is the Old Byblian dialect of Phoenician (KAI 1–8), which dates to the eleventh–tenth centuries b.c.e. 8 (e) Finally, the above principles must be considered in the framework of Hurvitz’s four tests for determining the historical period of a lexical or grammatical feature: biblical distribution, extrabiblical sources, linguistic opposition (or contrast), and accumulation. 9

much later books of Jeremiah and Ezekiel (for example, Jer 2:20, 33; 3:4, 5; 31:21; 46:11; Ezek 16:13, 18, 22, 31 (2×), 43 (2×); 47, 51; see also Ruth 3:3, 4). This suffix is the regular form in Akkadian (and is so vocalized in Ugaritic and Phoenician as well, which corroborates its aforementioned occurrence in the ABH Song of Deborah); but it is also the regular form in Aramaic, which explains its later attestations in Jeremiah and Ezekiel as an Aramaism. See earlier, GKC 121/§44h and n. 1; Kutscher 1982: 38–39 (§54); and most recently, Blau 2010: 55, 161. Contrast the much less satisfactory explanation accepted by I. Young, which at best could only justify the usage of this form in Ruth 3:3, 4. See Young, Rezetko, and Ehrensvärd 2008: 1.191. 7.  Example (2) in the present study is a reminder of the importance of using Akkadian evidence as a check against automatically assuming a late Aramaic influence on BH. 8. This important source is not even mentioned in the chapter on ABH in Young, Rezetko, and Ehrensvärd 2008: 1.312–40 (chap. 12). The Old Byblian dialect of Phoenician (KAI 1–8) provides important corroborating evidence for example (1) in the present study. 9.  See most conveniently Hurvitz 2000: 148–53. See also the detailed discussion of these four classical criteria in Young, Rezetko, and Ehrensvärd 2008: 1.20–23.

364

Chaim Cohen

2.  Example (1)  ‫ זֹאת! ָה ָארֶץ‬for MT ‫־ ָה ָארֶץ‬‎‫( ְו ֶאת‬Gen 49:15a) MT:

‫־ ָה ָארֶץ ִּכי נָ ֵעמָה‬‎‫ַוּי ְַרא ְמנֻחָה ִּכי טֹוב ְו ֶאת‬

njps: When he (Issachar) saw how good was security, and how pleasant was the country . . . My Suggested Emendation: ‫ַוּי ְַרא ְמנֻחָה ִּכי טֹובָה! זֹאת! ָה ָארֶץ ִּכי נָ ֵעמָה‬ My Translation: When he saw how good was the (allotted) haven, and how pleasant was this, the land . . . At first glance, the njps translation seems perfectly reasonable, based as it is on synonymous parallelism of the abc//b′c′ type (the verb ‫ ַוּי ְַרא‬that appears in the first clause stands for both clauses). The only apparent problem is the feminine noun ‫ ְמנֻחָה‬together with the masculine adjective or stative verb ‫טֹוב‬ ‘to be good’ (the expected form is ‫טֹובָה‬, just as in the second clause the parallel feminine form ‫ נָ ֵעמָה‬appears together with the feminine noun ‫) ָה ָארֶץ‬. This problem can easily be solved by simply reading !‫ ַוּי ְַרא ְמנֻחָה ִּכי טֹובָה‬together with the Samaritan Torah and (as translated by) the LXX and the Vulgate.  10 2.1.  Biblical Distribution of the Nota Accusativi ‫אֵת‬/-‫ אֶת‬versus the Homonymic Preposition ‫אֵת‬/-‫‘ אֶת‬with’ An additional problem that is relevant to the issue of diachrony in BH is the use here of the nota accusativi form -‫ ְואֶת‬in ABH. In all of the 10 ancient poems of ABH in the Torah and the Former Prophets (a total of approximately 250 verses), the term ‫ אֶת‬occurs only 5 times in 4 Torah verses, as clearly indicated in table 1. In the first additional occurrence (besides the presently discussed verse), Gen 49:25, the term ‫ ְואֵת‬is definitely not the nota accusativi. God is the one giving the blessing, and he is not its recipient! There is no reason to adopt the emendation accepted by BHS from ‫ ְואֵת‬to !‫‘ ְואֵל‬and God’ (the two components of the compound divine name ‫ אֵל ׁשַ ּדַ י‬have already been divided among the two stichs in this verse!). 11 As long ago demonstrated by Ibn Ezra, the prefixed preposition ‫‘ ֵמ‬from’ at the beginning of the first word of this verse, ‫ ֵמאֵל‬, stands also for the first word of the second clause, ‫( ְואֵת‬as if the text read ‫‘ *ּו ֵמאֵת ׁשַ ּדַ י‬and from Shaddai’). 12 For the parallelism of ‫ ֵמ‬//‫ ֵמאֵת‬, see 10.  See BHS; Cross and Freedman 1975: 86 n. 54; and most recently Zipor 2005: 607. As noted by Zipor, the MT masculine reading ‫ ִּכי טֹוב‬may have been influenced by the stock phrase ‫ ַוּי ְַרא אֱל ִֹהים ִּכי טֹוב‬, which is repeated 6× in Gen 1:4, 10, 12, 18, 21, 25 (but note also, in contrast, Gen 6:2). 11.  This emendation was earlier proposed by Cross and Freedman 1975: 91 n. 79. 12.  For the latest edition of Ibn Ezra’s commentary on Genesis, see M. Cohen 1999: 185. See also Weinfeld 1975: 312. Note that there is no valid internal textual evidence that the BH preposition ‫אֵת‬/-‫‘ אֶת‬with’ ever meant ‘from’ (as opposed to Akkadian itti ‘with’, which together with certain verbs does occasionally mean ‘from’; see, for example, CDA, 136). Contrast, for example, Hamilton 1995: 682 n. 18; de Hoop 1999: 206–7 and the bibliography

Diachrony in Biblical Hebrew Lexicography

365

Table 1 Ten ABH Poems

‫אֵת‬/-‫ אֵת‬Attested

No. of Verses

Gen 49:2–27

26.0

49:15 (‫־ ָה ָארֶץ‬‎‫)ו ֶאת‬ ְ 49:25 (‫)ואֵת ׁשַ ּדַ י‬ ְ

Exod 15:1b–18

17.5

0

Num 23:7–10, 18–24; Num 24:3–9, 15–19, 20b, 21b–22, 23–24

27.0

23:10 (‫ִׂש ָראֵל‬ ְ ‫־רֹבַע י‬‎‫) ֶאת‬

Deut 32:1–43

43.0

0

Deut 33:2–29

28.0

33:9 (‫־ ֶאחָיו‬‎‫)ו ֶאת‬ ְ 33:9 ([Qere] ‫־ ָּבנָיו‬‎‫)ו ֶאת‬ ְ

Judg 5:1–31

31.0

0

1 Sam 2:1–10

10.0

0

2 Sam 1:18–27

10.0

0

2 Sam 22(= Ps 18):2–51

50.0

0 a

7.0

0

2 Sam 23:1–7

a.  In two other verses in which ‫ א ִֹתי‬and -‫ ְואֶת‬occur in 2 Sam 22:20, 28, the alternative readings of these two verses in Ps 18:20, 28, respectively, without the nota accusativi should be preferred.

Ps 24:5. This word ‫אֶת‬, however, is not the nota accusativi but the preposition ‫ אֶת‬meaning ‘with’. The use of ‫ ֵמאֵת‬is identical to using ‫( מ ִֵעם‬as in Job 1:12, in comparison with Job 2:7). That the two terms ‫ אֶת‬are two completely different words—in fact, homophones—is proven by their two different declensions: ָ‫אֹותך‬, ‫אֹותי‬, ִ ‫ אֹותֹו‬for the nota accusativi; but ‫א ִּתי‬, ִ ָ‫א ְּתך‬, ְ ִ ‫ ִאּתֹו‬for ‫ אֶת‬meaning ‘with’. The latter preposition ‫ אֶת‬occurs in Akkadian as itti beginning with Old Akkadian in the second half of the third millennium b.c.e. (see CAD I/J 302–3; CDA, 136) and is written in Phoenician as ‫( את‬see PPD, 48), usually written differently from the Phoenician nota accusativi ‫( אית‬see PPD, 88–89); it appears later as ‫ י ַת‬in the Aramaic of Tg. Onqelos (see DJPA, 246–47). Given this distribution, the BH preposition ‫ אֶת‬meaning ‘with’ is surely diachronically appropriate for the ABH lexicon, but as we shall soon see, based especially on Phoenician comparative diachronic evidence, the BH nota accusativi ‫ אֶת‬most definitely is not! In a study published more than 30 years ago, I labeled the second additional occurrence of ‫ אֶת‬in ABH, the phrase ‫ִׂש ָראֵל‬ ְ ‫ אֶת־רֹבע י‬in Num 23:10 and listed in these two studies. The classic article demonstrating that such double-duty elements in BH and Ugaritic poetic parallelism must appear in the first (not the second) clause is Greenstein 1974: 89–96.

366

Chaim Cohen

its justified emendation to !‫ּב ַע‬ ֻ ‫ ּתַ ְר‬or !‫ ּתַ ְרּבֻעֹת‬meaning ‘dustcloud’ “perhaps the most celebrated case of an emended biblical hapax legomenon.” 13 Regarding the third and fourth additional occurrences in Deut 33:9ab, I have little positive to suggest other than the fact that, if the emendation of Gen 49:15a proposed here is accepted, the former verse would then be the only verse of approximately 250 verses in ABH in which the nota accusativi is attested. 14 2.2.  The Evidence from Extrabiblical Sources Besides this internal BH evidence for the near absence of the nota accusativi in the ABH dialect, there is also important corroborating comparative diachronic evidence from Phoenician—or more specifically, the Old Byblian dialect of Phoenician (= KAI, inscriptions 1–8), dating to ca. eleventh–tenth centuries b.c.e., precisely the period of the ABH dialect. According to M. Giu­ lia and A. Guzzo (1999: 196–97 [§275]), the Phoenician nota accusativi (appearing in the three forms ‫ת‬/‫את‬/‫ )אית‬never occurs in the Old Byblian dialect of Phoenician. Its earliest attested occurrence is in the Standard Phoenician dialect at the end of the ninth century b.c.e. (KAI 30:3) and then often in the Azitawada inscription in the middle of the eighth century b.c.e. (for example, KAI 26A:I:3; III:3, 14/15, 19 [3×]). In contrast, the Phoenician preposition ‫את‬ ‘with’ (only in this form), which is cognate with the Akk. preposition itti appearing first in the Old Akkadian period (ca. 2500–2000 b.c.e.; see also §2.1), appears in KAI 1:1 (usually considered the earliest Old Byblian inscription) as an element in the accepted reconstructed PN of the son of ‫אחרם‬: ‫]א]תבעל‬, which is usually analyzed as ʾItto-Baʿal, meaning ‘Baʿal is with him’.  15 It is thus clear both that in the MT of the poems written in ABH there are very few cases of the nota accusativi and that this paucity of attestations paral13.  See C. Cohen 1978: 38. For complete details regarding the emendation of ‫רֹבע‬-‫אֶת‬ to !ַ‫ּבע‬ ֻ ‫ ּתַ ְר‬or !‫ ּתַ ְרּבֻעֹת‬meaning ‘dustcloud’ based on the similar usage of Akk. tarbŭ῀/tarbūtu with that meaning, see C. Cohen 1978: 37–39, 60–63, and most recently Ges18, 1216. For the different forms of the Akkadian term and some additional occurrences, see now CDA, 400 and especially CAD T, 485. The main justification for this emendation is not that this term must have one of the same noun patterns as its Akkadian semantic equivalent, but rather (as long ago suggested by W. F. Albright) the possibility of thus also solving the problem of this single attestation of the nota accusativi in this ABH poem. See C. Cohen 1978: 62, n. 76; contra Morag 1981: 9, n. 27. 14.  I cannot accept the nihilistic proposal of Cross and Freedman (1975: 112, n. 28) to simply ignore all of verses 8–10 based on the following sweeping conclusion: “How much, if any, of vs. 8–10 belongs to the original blessing must remain a question. The passage is rejected in toto as a late addition by some scholars.” Similarly, I cannot accept the brilliant but completely unjustifiable emendations to this verse suggested by N. H. Tur-Sinai (1962: 237). Note that even after all these emendations, the two attestations of the nota accusativi still remain intact. 15.  See, for example, PPD, 89 which also compares the PN ‫ֶת ַּבעַל‬ ְ ‫( א‬1 Kgs 16:31), albeit in its MT misvocalized form.

Diachrony in Biblical Hebrew Lexicography

367

lels the situation in Phoenician. What remains to be discussed is the positive evidence for my new, suggested textual emendation in Gen 49:15a (see the beginning of §2 above). The historical validity for this textual emendation cannot be proven absolutely. A strong circumstantial case in its favor can be made, however. 16 It is not necessary to show that there was not a single case of the nota accusativi in ABH. It is sufficient to demonstrate that there were almost no cases, and this has already been shown. Moreover, I am not the first scholar to analyze Gen 49:15a textually in this way, trying to arrive at a reasonable reconstruction of the text that would eliminate the nota accusativi. 17 What I attempt to demonstrate is merely that my emendation is somewhat more reasonable than others that have been proposed. 2.3.  Three Previous Reconstructions of Gen 49:15a In the following paragraphs, I discuss the three main reconstructions of Gen 49:15a (MT = ‫ ָה ָארֶץ ִּכי נָ ֵעמָה‬-‫ ) ַוּי ְַרא ְמנֻחָה ִּכי טֹוב ְואֶת‬that have been suggested before now. (a)  Cross and Freedman (1950) ‫י ְִר ֶא ְמנֻח ֹ ִּכ טֹב י ִַּבט א ְַרצ ֹ ִּכ נָ ֵע ָמ‬ Translation: He sees for himself a resting-place which is good, He beholds for himself a land which is pleasant. Besides removing the matres lectiones, Cross and Freedman (1975: 74 and 86 nn. 53–56) assume that “the nota accusativi and article reflect a later revision of the text; these may have been introduced to balance the meter in a defective line, since something has apparently dropped out.” They further assume that what dropped out was the verbal form ‫י ִַּבט‬, based on partial haplography of the last two words of the first stich (‫—ּכי טֹב‬without ִ the ‫ו‬′′‫ ו‬but with the ‫ד‬′′‫)יו‬. There is evidence from Old Byblian, however, for the occasional use of the definite article in the ABH period. 18 Furthermore, the reconstruction of ‫יביט‬ ‫‘ ארצו‬he beholds for himself a land’ is without precedent in BH. All such constructions include the preposition ‫ל‬/‫ֶל‬ ְ ‫‘ א‬to’ (cf. Isa 5:30; 8:22; 51:6; Ps 102:20; 16.  I hereby thank my good friend and former teacher Shalom Paul for discussing this issue with me and helping me to arrive at the above balanced conclusion and formulation regarding the inability of textual scholars to “prove” these sorts of emendation absolutely. 17.  Contrast the following forced argumentation by de Hoop against this sort of emendation: “the use of a nota accusativi in an ancient poem [that is, ABH—C. C.], is no reason for a different translation [in other words, a textual emendation—C. C.]. First of all, it is possible that this nota accusativi was inserted later on. Or the text may not have been as old as . . . supposed; or the whole presupposition that the presence of ‫ אֶת‬is impossible in older poetry was wrong” (see de Hoop 1999: 157). 18. See KAI 4:2–3: ‫‘ הבתם אל‬these temples’. Compare ‫‘ הארץ זאת = זאת הארץ‬this land = this, the land’, which is part of my own aforementioned reconstruction of Gen 49:15a (see above).

368

Chaim Cohen

104:32; Job 28:24). The emendation of one letter from ‫ ו‬to ‫ ז‬seems a much simpler and more reasonable reconstruction. (b) S. Gevirtz (1975: 109–11)

‫ ַוּי ֵֶתא רָץ ִּכי נָ ֵעמָה‬/ ‫ֱתה‬ ֶ ‫ַוּי ְַרא ִמ ְנחָה ִּכי טֹוב ַוּיֶא‬ Translation: And he saw how advantageous was offering, And he found how expedient was tribute. Gevirtz’s reconstruction (based primarily on his very dubious rejection of the parallelism ‫ ָה ָארֶץ‬//‫‘ ְמנֻחָה‬the (allotted) haven//the land’, for which, see §2.4 below) is so extreme that it hardly requires refutation. I simply quote de Hoop’s fully justified criticism: “Nothing in the textual tradition suggests a major corruption and the text is understandable as it stands. Finally, the saying becomes fully incomprehensible in Gevirtz’s translation: Issachar saw that offering was advantageous and for that reason became a slave?” (de Hoop 1999: 57). (c) S. Geller (1979: 66–67)

‫(ו)א ְַרצ ֹ ִּכ נָ ֵע ָמ‬ ְ ‫י ְִר ֶא ְמנֻח ֹ ִּכ טֹב‬ Translation: When he saw that his resting-place was (so) good, His land was (so) pleasant, . . . Geller’s reconstruction is an adaptation of Cross and Freedman’s suggestion, without adding the verb ‫יביט‬. Thus, Geller’s solution to the problem that the nota accusativi appears in such an early poem is simply to delete it, without feeling any need to explain why just in this case (and in almost no other in the MT among 250 ABH verses) the nota accusativi was later added to the original text. Once again, the emendation of just one letter from ‫ ו‬to ‫ ז‬seems clearly to be preferable. 2.4.  The Special Usage of BH ‫ ְמנֻחָה‬in Gen 49:15a The positive evidence for the emendation ‫ זֹאת! ָה ָארֶץ‬in place of MT ‫־‬‎‫ְו ֶאת‬ ‫ ָה ָארֶץ‬in Gen 49:15a depends first of all on a proper understanding of the special use of the term ‫ ְמנֻחָה‬meaning ‘(allotted) haven’ in this context referring to the Issachar tribe’s satisfaction with the apportioned land it received. My translation of the first clause of Gen 49:15a,  19 !‫‘ ַוּי ְַרא ְמנֻחָה ִּכי טֹובָה‬When he saw how good was the (allotted) haven’, in accordance with this special usage, is based on the following philological evidence: (a) Deut 12:9–10 ְ ‫ אֱלֹהֶיךָ נֹתֵ ן ל‬′‫ֲׁשר ה‬ ‫ָך׃‬ ֶ ‫חלָה א‬ ֲ ַ‫ָּתה אֶל ה ְַּמנּוחָה ְואֶל ַהּנ‬ ָ ‫ָאתם עַד ע‬ ֶ ‫ִּכי לֹא ב‬ ֹ ‫ִיח ָלכֶם‬ ַ ‫ֶתכֶם ְו ֵהנ‬ ְ ‫ְחיל א‬ ִ ‫ אֱלהֵיכֶם ַמנ‬′‫ֲׁשר ה‬ ֶ ‫ַועֲב ְַר ֶּתם אֶת ַהּי ְַרּדֵ ן ִויׁשַ ְב ֶּתם ָּב ָארֶץ א‬ ‫ָביב ִויׁשַ ְב ֶּתם ֶּבטַח׃‬ ִ ‫ִמּכָל א ְֹיבֵיכֶם ִמּס‬ 19.  For the emendation of MT ‫ טֹוב‬to ‫טֹובָה‬, see the beginning of §2 above and n. 10.

Diachrony in Biblical Hebrew Lexicography

369

njps: because you have not yet come to the allotted haven that the Lord your God is giving you. When you cross the Jordan and settle in the land that the Lord your God is allotting to you, and He grants you safety from all your enemies around you and you live in security. . . . (b) Other Related BH Verses Compare also the related usage of ‫ ְמנּוחָה‬referring to Zion as the land chosen by God as His “resting-place”—that is, his place of residence (Ps 95:11; 132:8 [cf. 2 Chr 6:41]; contrast Isa 66:1) and even once ‫זֹאת‬ ‫ָתי‬ ִ ‫‘ ְמנּוח‬this is my resting-place’ (Ps 132:13–14; see. the discussion of ‫‘ זֹאת ָה ָארֶץ‬this [is] the land’ immediately below). (c) Tg. Onqelos on Gen 49:15a (‫) ַוּי ְַרא ְמנֻחָה ִּכי טֹוב‬:

‫ָקא אֲרֵי טָב‬ ָ ‫חזָא חּול‬ ֲ ‫‘ ַו‬On perceiving that the allotted portion was good’ (and compare the commentaries of Rashi, Ibn Ezra, Rashbam, and Ḥizquni). 20 Thus, this positive evidence indicates that, in parallelism with the term ‫‘ ָה ָארֶץ‬the land’ in the second clause of the present context, the njps translation of ‫ ְמנֻחָה‬as ‘security’ in Gen 49:15a is surely inadequate; and even the rendering ‘resting place’ does not quite fit. The special usage suggested here ‘(allotted) haven’ in Gen 49:15a on the basis of all the above evidence may also mark the ABH origin of this usage throughout BH. 2.5.  The Special Use of BH ‫ זֹאת! ָה ָארֶץ‬in Gen 49:15a The most dramatic positive evidence, however, for the emendation !‫זֹאת‬ ‫ ָה ָארֶץ‬in place of MT ‫־ ָה ָארֶץ‬‎‫ ְו ֶאת‬in Gen 49:15a is the key BH usage of the technical phrase ‫‘ זֹאת ָה ָארֶץ‬this (is) the land’. 21 In all 5 of its occurrences in the 20.  For all these medieval commentaries and Tg. Onqelos on Gen 49:15a, see Katznelenbogen 1987: 268–69. See also de Hoop 1999: 156–57 n. 471. Rashbam was apparently the first to compare the verse with Deut 12:9 specifically, but this comparison was already implied by the targum’s translation ‫ָקא‬ ָ ‫‘ חּול‬the allotted portion’. It consistently rendered ‫ חּולָק ְוא ְַח ָסנָה‬for the BH phrase ‫‘ חלק ונחלה‬a share in the inheritance’ (Gen 31:14; Deut 10:9; 12:12; 14:27, 29; 18:1; see Aberbach and Grossfeld 1976: 35 n. 37; and 48, last paragraph [= the inadvertently unnumbered n. 54]). 21.  In the discussion following my NAPH lecture, on which this essay is based, T. Muraoka rightly claimed that the difference in usage between the 5 attested MT occurrences of ‫ זֹאת ָה ָארֶץ‬meaning ‘this is the land’ (see above) and my proposed emendation ‫ זֹאת! ָה ָארֶץ‬in Gen 49:15a with the meaning ‘this the land’ (= ‘this land’ with ‫ זֹאת‬preceding the noun being used adjectivally = ‫ ) ָה ָארֶץ הַּזֹאת‬must be specifically addressed. In response to Muraoka, I note especially the following select cases of this admittedly rare adjectival usage cited from among the list of examples in Joüon and Muraoka 2006: 500 (§143i): Gen 2:23 (‫ָדם‬ ָ ‫וַּיֹאמֶר ָהא‬ ‫‘ זֹאת ַה ַּפעַם‬Then the man said, “This time” [as opposed to the previous attempts]’; compare Exod 8:28; 9:14; 2 Sam 17:7; Jer 10:18; 16:21); Exod 32:1, 23 (‫ָאיׁש‬ ִ ‫ׁשה ה‬ ֶ ֹ ‫־זֶה מ‬‎‫‘ ִּכי‬for that man Moses’; compare Exod 11:3); 1 Kgs 14:14 (‫ָּתה‬ ָ ‫‘ זֶה הַּיֹום ּומֶה גַם ע‬this very day and even now’; see Exod 12:14, 17); 2 Kgs 6:33 (′‫־זֹאת ָה ָרעָה ֵמאֵת ה‬‎‫‘ ִהּנֵה‬This calamity is from God’;

370

Chaim Cohen

MT as follows, this phrase refers only to the land apportioned to the tribes and to the promised land for all of Israel (in Josh 13:23, the technical phrase ‫זֹאת‬ ‫חלַת‬ ֲ ַ‫‘ נ‬this (is) the allotted portion of the land’ replaces ‫ זֹאת ָה ָארֶץ‬in 13:2 [see below] and appears 11 more times in Joshua 13–19):23 Num 34:2: ‫ָאים אֶל ָה ָארֶץ ְּכנָעַן זֹאת‬ ִ ‫ַּתם ּב‬ ֶ ‫א ֵלהֶם ִּכי א‬ ֲ ‫ִׂש ָראֵל ְו ָאמ ְַר ָּת‬ ְ ‫־ּבנֵי י‬‎ ְ ‫צַו ֶאת‬ ‫חלָה ֶארֶץ ְּכנַעַן ִל ְגבֻל ֶֹתי ָה‬ ֲ ַ‫ֲׁשר ִּתּפֹל ָלכֶם ְּבנ‬ ֶ ‫ָה ָארֶץ א‬ Njps: Instruct the Israelite people and say to them: When you enter the land of Canaan, this is the land that shall fall to you as your portion, the land of Canaan with its various boundaries. Num 34:13: ‫ֲׁשר ִּת ְתנַחֲלּו א ָֹתּה‬ ֶ ‫ִׂש ָראֵל לֵאמֹר זֹאת ָה ָארֶץ א‬ ְ ‫־ּבנֵי י‬‎ ְ ‫ׁשה ֶאת‬ ֶ ֹ ‫ַו ְיצַו מ‬ ‫ַּטה‬ ֶ ‫ֲצי ַהּמ‬ ִ ‫ׁשעַת ַהּמַּטֹות ַוח‬ ְ ‫ לָתֵ ת ְל ִת‬′‫ֲׁשר ִצּוָה ה‬ ֶ ‫ְּבגֹורָל א‬ njps: Moses instructed the Israelites, saying: This is the land you are to receive by lot as your hereditary portion, which the Lord has commanded to be given to the nine and a half tribes. Deut 34:4: ‫ּוליַעֲק ֹב‬ ְ ‫ִצחָק‬ ְ ‫ַע ִּתי ְלא ְַב ָרהָם ְלי‬ ְ ‫ִׁשּב‬ ְ ‫ֲׁשר נ‬ ֶ ‫ ֵאלָיו זֹאת ָה ָארֶץ א‬′‫וַּיֹאמֶר ה‬ ‫ׁשּמָה לֹא תַ עֲבֹר‬ ָ ‫יתיךָ ְבעֵינֶיךָ ְו‬ ְ ‫לֵאמֹר ְלז ְַרעֲךָ א‬ ִ ‫ֶר ִא‬ ְ ‫ֶּתנֶּנָה ה‬ Njps: And the Lord said to him, “This is the land of which I swore to Abraham, Isaac, and Jacob: I will assign it to your offspring. I have let you see it with your own eyes, but you shall not cross there” (compare with the boundaries of the promised land listed in vv. 1–3). Josh 13:2: ‫ְׁשּורי‬ ִ ‫ׁש ִּתים ְוכָל ַהּג‬ ְ ‫ְלילֹות ה ְַּפ ִל‬ ִ ‫ִׁש ָארֶת ּכָל ּג‬ ְ ‫זֹאת ָה ָארֶץ ַהּנ‬ Njps: This is the territory that remains: all the districts of the Philistines and all [those of] the Geshurites. (What follows is a geographical description of the parts of the land that still await distribution to the nine and a half tribes, ending with v. 7: ‫חלָה‬ ֲ ַ‫־ ָה ָארֶץ הַּזֹאת ְּבנ‬‎‫ְוע ַָּתה ַחּלֵק ֶאת‬ ‫ּשׁבֶט ה ְַמנַּשֶׁה‬ ֵ ‫ֲצי ַה‬ ִ ‫ָטים ַוח‬ ִ ‫ּשׁב‬ ְ ‫ׁשעַת ַה‬ ְ ‫‘ ְל ִת‬Therefore, divide this territory into hereditary portions for the nine tribes and the half tribe of Menasseh’.) 22

see Judg 20:3, 12; 1 Kgs 9:9; Ezek 6:10). Note, however, that the phrase ‫( זֶה הַּיֹום‬which in 1 Kgs 14:14 above means ‘this very day’) also appears with the meaning ‘this is the day’ in Ps 118:24; Lam 2:16. This is also the case with the phrase ‫זֹאת ָה ָארֶץ‬. In the 5 verses discussed below in this same subsection, it does indeed mean ‘this is the land’; but in the emended Gen 49:15a, its original meaning in the very same context in ABH was ‘this the land’ (= ‘this land’). 22.  Note the subsequent substitution of the technical phrase ‫ זֹאת ָה ָארֶץ‬by the no-lesstechnical phrase ‫חלַת‬ ֲ ַ‫‘ זֹאת נ‬this (is) the allotted portion of the land’ (Josh 13:23, 28; 15:20; 16:8; 18:20, 28; 19:8, 16, 23, 31, 39, 48).

Diachrony in Biblical Hebrew Lexicography

371

Ezek 48:29: ‫קֹותם ְנאֻם‬ ָ ‫ִׂש ָראֵל ְו ֵאּלֶה מ ְַח ְל‬ ְ ‫ׁש ְבטֵי י‬ ִ ‫חלָה ְל‬ ֲ ַ‫ֲׁשר ּתַ ִּפילּו ִמּנ‬ ֶ ‫זֹאת ָה ָארֶץ א‬ ‫ אלהים‬′‫ה‬ Njps: That is the land which you shall allot as a heritage to the tribes of Israel, and those are their portions—declares the Lord God. In conclusion, the positive results of this example clearly demonstrate the need for a proper diachronic perspective in BH textual analysis, especially as an additional justification for textual emendations of the MT (even for minor textual emendations). 3. Example (2) The ‫ּות‬- Suffix for Abstract Nouns 3.1.  Biblical Distribution There are approximately 65 distinct BH abstract nouns in the MT with the ‫ּות‬- suffix. 23 Of these, about 15 occur in the Torah (the words marked with an asterisk below appear only in the Torah): ,‫ ְּפדּות‬,‫ עֵדּות‬,‫*מ ְסּכֵנּות‬ ִ ,‫ מ ְַלכּות‬,‫ ְּכִריתּות‬,‫*ּכבֵדּות‬ ְ ,‫ זְנּות‬,‫ ְּדמּות‬,‫ *ּג ְַבלֻת‬,‫א ְַלמָנּות‬ .‫ *ּתַ ְרּבּות‬,‫ׁשִרירּות‬ ְ ,‫ׁשבּות‬ ְ ,‫*קֹומ ִמּיּות‬ ְ ,‫*צ ִמיתֻת‬ ְ Eleven of these 65 nouns occur in the LBH books—Qoheleth, Esther, Daniel, Ezra, Nehemiah, 1 Chronicles, and 2 Chronicles (the nouns that are marked with an asterisk appear only in these books): ,‫ עֵדּות‬,‫ *ע ְַבדּות‬,‫ׂש ְכלּות‬/ ִ ‫*ס‬ ִ ,‫ מ ְַלכּות‬,‫ י ְַלדּות‬,‫*ה ְתח ְַּברּות‬ ִ ,‫ *הֹולֵלּות‬,‫ְּדמּות‬ .‫*ׁש ְפלּות‬ ִ ,‫ *ׁשַ חֲרּות‬,‫*רעּות‬ ְ The remaining 42 BH abstract nouns occur neither in the Torah nor in the aforementioned LBH books. This distribution is clear evidence that, diachronically, the ‫ּות‬- suffix for abstract nouns cannot be considered an identifying feature of LBH. 23.  No attempt will be made here to provide full documentation for the 65 BH abstract nouns with the –‫ ּות‬suffix referred to in the present study. All such documentation and full philological discussion of all contexts will be provided by Hilla Mikhaʾel in her Ben-Gurion University Ph.D. thesis, The Phenomenon of Biblical Hebrew Abstract Nouns with the -‫ּות‬ Suffix in Light of Akkadian Abstract Nouns with the -ūtu Suffix (in Hebrew), which is currently in preparation under my supervision. I have already confirmed all 65 forms and can vouch for their existence in BH. The only other modern study of the biblical distribution of all abstract nouns with the -‫ ּות‬suffix known to me is that of Rezetko 2003: 224. His figure of “71 unique common nouns of this type in the Bible” (no list is provided) is in my opinion a bit inflated, perhaps as a result of including such cases as ‫‘ ּבָכּות‬weeping’, which occurs only as part of the proper name of a tree (Gen 35:8), and some other cases, such as ‫‘ ה ְַלמּות‬hammer’ (Judg 5:26), which despite having the -‫ ּות‬suffix is surely not an abstract noun. The only full-length study known to me is Gulkowitsch 1931, which deals (often indiscriminately) with all abstract nouns (with or without the -‫ ּות‬suffix) in all Hebrew dialects (biblical and postbiblical) and therefore makes no real contribution to the subject at hand.

372

Chaim Cohen

3.2.  Extrabiblical Sources While it is certainly true that the ‫ּות‬- suffix for abstract nouns is very common in Aramaic dialects of the Second Temple period, the Akkadian -ūtu suffix for abstract nouns is no less common; it occurs in all Akkadian dialects, beginning with Old Akkadian in the second half of the third millennium b.c.e. Compare, for example, šarrūtu ‘kingship’ (from Old Akkadian on [see CDA, 361]; semantically equivalent to BH ‫( ;)מ ְַלכּות‬w)ardūtu ‘slavery’ (from OA/OB on [CDA, 434]; semantically equivalent to BH ‫ ;)ע ְַבדּות‬šībūtu ‘old age; testimony; witness’ (from OA/OB on [CDA, 371]; partially semantically equivalent to BH ‫ ;)עֵדּות‬almānūtu ‘widowhood’ (SB and Nuzi [CDA, 13]; cognate with BH ‫ ;)א ְַלמָנּות‬kabtūtu ‘majesty; heaviness’ (SB [CDA, 140]; cognate with BH ‫)ּכבֵדּות‬. ְ  24 Most striking perhaps is the West Semitic hapax legomenon loanword ripu῀tu ‘health’ (cognate with BH ‫[ ִר ְפאּות‬Prov 3:8]), which is attested only in EA 269:17 from the fourteenth century b.c.e. (see Nili 2004: 211–12 and my note quoted there). This important Akkadian comparative evidence is a further indication that, diachronically, the ‫ּות‬- suffix for abstract nouns cannot be considered an identifying feature of LBH. 3.3.  Linguistic Opposition (or, Contrast) Use of the criterion of opposition (or, contrast) in fact leads to the same conclusion as in the two previous sections. Of course, there are some cases in which a particular abstract noun with ‫ּות‬- suffix attested only in LBH replaces a semantic equivalent with a different morphological structure in SBH (for example, LBH ‫ׂש ְכלּות‬/ ִ ‫ ִס‬and ‫הֹולֵלּות‬, both meaning ‘folly’, replacing their SBH semantic equivalent ‫)אֶּולֶת‬. ִ But these same LBH terms also replace the 2 SBH semantically equivalent parallel terms ‫ ְּפתַ ּיּות‬and ‫( ְּכ ִסילּות‬each appears only in Prov 9:13), both of which are abstract nouns with an ‫ּות‬- suffix. It is thus clear that LBH and SBH both represent the concept of ‘folly’ by using two different abstract nouns with an ‫ּות‬- suffix. Similarly, the most widely attested and most often cited LBH abstract noun with an ‫ּות‬- suffix is the term ‫‘ מ ְַלכּות‬kingship, reign’ (occurring 91 times in BH, the vast majority in LBH). In the very helpful table in Young, Rezetko, and Ehrensvärd (2008: 2.197 [item #191]), it is noted that 1 of the 4 terms that LBH ‫ מ ְַלכּות‬replaces is SBH ‫( מ ְַמלָכּות‬occurring 9 times in BH, e.g., 1 Sam 15:28; 2 Sam 16:3). Thus, once again it is clear that LBH and SBH each represent the concept of ‘kingship’ by using different abstract nouns with an ‫ּות‬- suffix. This evidence does not warrant the conclusion that the ‫ּות‬- suffix for abstract nouns is an identifying feature of LBH. Thus, all three criteria (§§3.1, 3.2, and 3.3 above) lead to the inescapable conclusion that, as opposed to almost all mod24. For Akkadian abstract nouns with -ūtu suffix, see in general Huehnergard 1997: 124–25 (§14.4); for the Akkadian-BH parallels, see C. Cohen 1978: 79–80 n. 170.

Diachrony in Biblical Hebrew Lexicography

373

ern scholarship on this issue, 25 the BH abstract nouns with ‫ּות‬- suffix ( just as in Akkadian) cannot be used to identify any individual period of BH. 4. Conclusion In conclusion, the two examples analyzed in this article have demonstrated the importance of proper diachronic analysis, not only for BH lexicography in general (example [2]), but also, more specifically, for correct textual analyses of the MT (example [1]). In both cases, the key was the correct (example [1]) or incorrect (example [2]) usage of assumed identifying features of the different periods of BH (ABH in example [1]; LBH in example [2]). Perhaps the most significant methodological lesson to be learned here is the importance, not only of determining and then considering the original period of a particular BH text before attempting its textual analysis (example [1]), but also the need to reinvestigate critically the assumed identifying features of the three periods (example [2]). 25.  See, for example, Hurvitz 1972: 79–82; Kutscher 1982: 43, 81, 84; Qimron 1986: 66 (note however the end of n. 89 there). Only Joüon and Muraoka (2006: 242–43) and Rezetko (2003: 224; see n. 23 above) were on the right track.

References

Aberbach, M., and Grossfeld, B. 1976 Targum Onkelos on Genesis 49. SBL Aramaic Studies 1. Missoula, MT: Scholars Press. Avishur, Y. 1979 Phoenician Inscriptions and the Bible, part 1. Jerusalem: Rubenstein. [Hebrew] Blau, J. 2010 Phonology and Morphology of Biblical Hebrew: An Introduction. LSAWS 2. Winona Lake, IN: Eisenbrauns. Cassuto, U. 1975 Biblical and Oriental Studies, trans. I Abraham. Vol. 2. Jerusalem: Magnes. CDA 2000 Black, J.; George, A.; and Postgate, N., eds. A Concise Dictionary of Akkadian. 2nd ed. Wiesbaden: Harrassowitz. Cogan, M. 2001 1 Kings. AB 10. New York: Doubleday. Cohen, C. 1978 Biblical Hapax Legomena in the Light of Akkadian and Ugaritic. SBLDS 37. Missoula, MT: Scholars Press. 2008 New Directions in Modern Biblical Hebrew Lexicography. Pp.  441–73 in Birkat Shalom: Studies in the Bible, Ancient Near Eastern Literature, and Postbiblical Judaism Presented to Shalom M. Paul on the Occasion of His Seventieth Birthday, ed. C. Cohen et al. Winona Lake, IN: Eisenbrauns.

374

Chaim Cohen

Cohen, M. 1999 Mikraʾot Gedolot ‘Ha-Keter’: Genesis, part 2. Ramat-Gan: Bar-Ilan University Press. [Hebrew] Cross, F. M., and Freedman, D. N. 1975 Studies in Ancient Yahwistic Poetry. SBLDS 21. Missoula, MT: Scholars Press. DJPA 2002 Sokoloff, M. A Dictionary of Jewish Palestinian Aramaic of the Byzantine Period. 2nd ed. Ramat-Gan: Bar-Ilan University Press / Baltimore: Johns Hopkins University Press. Geller, S. A. 1979 Parallelism in Early Biblical Poetry. HSM 20. Missoula, MT: Scholars Press. Ges18 1987–2009  Gesenius, W. Hebräisches und aramäisches Handwörterbuch über das Alte Testament, ed. R. Meyer and H. Donner. 18th ed. Berlin: Springer. Gevirtz, S. 1975 The Issachar Oracle in the Testament of Jacob. ErIsr 12 (Nelson Glueck Memorial Volume): 104–12. Giulia, M., and Guzzo, A. 1999 Phönizisch – Punische Grammatik. 3rd ed. AnOr 55. Rome: Pontifical Biblical Institute. Greenstein, E. L. 1974 Two Variations of Grammatical Parallelism in Canaanite Poetry and Their Psycholinguistic Background. JANES 6: 87–105. Gulkowitsch, L. 1931 Die Bildung von Abstraktbegriffen in der hebräischen Sprachgeschichte. Leipzig: Pfeiffer. Hamilton, V. P. 1995 The Book of Genesis: Chapters 18–50. NICOT. Grand Rapids, MI: Eerdmans. Hoop, R. de 1999 Genesis 49 in Its Literary and Historical Context. OtSt 29. Leiden: Brill. Huehnergard, J. 1997 A Grammar of Akkadian. HSS 45. Atlanta: Scholars Press. [2nd ed., Winona Lake, IN: Eisenbrauns, 2005] Hurowitz, V. (A.) 1992 I Have Built You an Exalted House. JSOTSup 115. Sheffield: Sheffield Academic Press. 1994 In the Month of ‫זו‬, That Is, the Second Month. Pp. 69–70 in ʿOlam Ha-Tanakh: I Kings, ed. B. Oded. Tel-Aviv: Davidson-Attai. [Hebrew] Hurvitz, A. 1972 The Transition Period in Biblical Hebrew. Jerusalem: Bialik Institute. [Hebrew] 2000 Can Biblical Texts Be Dated Linguistically? Chronological Perspectives in the Historical Study of Biblical Hebrew. Pp. 143–60 in Congress Volume: Oslo 1998, ed. A. Lemaire and M. Saebo. VTSup 80. Leiden: Brill.

Diachrony in Biblical Hebrew Lexicography

375

2002 The Archaeological Debate on the Antiquity of the Hebrew Bible in Light of Linguistic Research of the Hebrew Language. Pp. 34–46 in The Controversy over the Historicity of the Bible, ed. L. I. Levine and A. Mazar. Jerusalem: Yad Ben-Zvi. [Hebrew] Joüon, P., and Muraoka, T. 2006 A Grammar of Biblical Hebrew. Revised English Edition. Subsidia Biblica 27. Rome: Pontifical Biblical Institute. Katznelenbogen, M. L. 1987 Torat Chayyim: Genesis, part 2. Jerusalem: Mossad Harav Kook. [Hebrew] Kutscher, E. Y. 1982 A History of the Hebrew Language. Jerusalem: Magnes / Leiden: Brill. Morag, S. 1981 ‘Layers of Antiquity’: Some Linguistic Observations on the Oracles of Ba­ laam. Tarbiz 50: 1–24. [Hebrew] Mulder, M. J. 1998 I Kings. Historical Commentary on the Old Testament. Leuven: Peeters. Nili, Y. 2004 Hapax Legomena in the Akkadian Language. Ph.D. Thesis. Ramat-Gan: BarIlan University. [Hebrew] PPD 2000 Krahmalkov, C. R. Phoenician–Punic Dictionary. Leuven: Peeters. Qimron, E. 1986 The Hebrew of the Dead Sea Scrolls. HSS 29. Atlanta: Scholars Press. Rendsburg, G. A. 1991 The Strata of Biblical Hebrew. JNSL 17: 81–99. 2002 Israelian Hebrew in the Book of Kings. Bethesda, MD: CDL. Rezetko, R. 2003 Dating Biblical Hebrew: Evidence from Samuel–Kings and Chronicles. Pp.  215–50 in Biblical Hebrew: Studies in Chronology and Typology, ed. I. Young. JSOTSup 369. London: T. & T. Clark. Tur-Sinai, N. H. 1962 Peshuto Shel Miqraʾ, vol. 1. Jerusalem: Kiryat Sepher. [Hebrew] Weinfeld, M. 1975 The Book of Genesis with a New Commentary. Tel-Aviv: Gordon. [Hebrew] Young, I.; Rezetko, R.; and Ehrensvärd, M. 2008 Linguistic Dating of Biblical Texts. 2 vols. London: Equinox. Zevit, Z. 2005 Introductory Remarks: Historical Linguistics and the Dating of Hebrew Texts. HS 46: 321–26. Zipor, M. 2005 The Septuagint Version of the Book of Genesis. Ramat-Gan: Bar-Ilan University Press. [Hebrew]

Outline of Aramaic Diachrony Michael Sokoloff 1. Introduction In the present essay, I survey select features found in Aramaic from the beginning of its historical appearance (ca. tenth century b.c.e.) 1 through their acceptance and integration into the great literary dialects of the first millennium c.e. (ca. 200–1000 c.e.). 2 Apart from literary texts such as the Proverbs of Aḥiqar, preserved in a fifth-century b.c.e. copy of an earlier text, and the Aramaic material in the biblical books of Ezra and Daniel, all of the firstmillennium b.c.e. texts are original and can be dated definitely either on internal or on external grounds. Hence, from the point of view of a diachronic study, questions of textual transmission are irrelevant. 3 2.  Dialects Examined 2.1.  Old Aramaic (= OA; ca. 950–600 b.c.e.) Inscriptions inscribed on stone and other hard materials from the land of Israel in the west to Iran in the east have been found from this period. The most important (from west to east) are: 1.  The Tel Dan Inscription—A text set up by Hazael of Aram 2.  Samalian—A western dialect from the Kingdom of Samʾal used in a small number of royal inscriptions Author’s note:  Since the submission of this essay in June 2011, two important volumes on Aramaic have appeared: Katell Berthelot and Daniel Stökl Ben Ezra, Aramaica Qumranica: Proceedings of the Conference on the Aramaic Texts from Qumran in Aix-en-Provence 30 June–2 July 2008 (STDJ 94; Leiden: Brill, 2010); and T. Muraoka, A Grammar of Qumran Aramaic (ANES Supplement 38; Leuven: Peeters, 2011). 1. Since the Arameans are definitely attested historically from the thirteenth century b.c.e. (see Lipiński 1997: 45–50), it can be assumed that Aramaic was already by then an independent language, although there is no direct evidence for it in the second millennium b.c.e. 2.  For a general survey of the Aramaic dialects, see Beyer 1986; Huehnergard 1995; Lipiński 1997: 61–70; Creason 2008. For a detailed diachronic analysis of Aramaic that relies too heavily on reconstruction, see Beyer 1984: 77–153; 1994: 37–56; 2004: 45–66. 3.  In order to follow the discussion presented here, the reader should consult the list of abbreviations. See also the Aramaic texts provided on p. 401 below.

379

380

Michael Sokoloff

3.  Deir ʿAlla—A text mentioning the biblical figure of Balaam  4 4.  Texts from the core Aramean area—Most of the texts (for example, the Sefire inscriptions) come from present-day Syria in the area around Damascus and Aleppo 5.  Tell Fakhariya (Fakh)—A bilingual Akkadian-Aramaic inscription on a statue that was originally set up in a temple in Gozan 6.  Mesopotamia—Texts on clay tablets 7.  Bukān—A dedicatory inscription found in Iranian Azerbaijan

2.2.  Official Aramaic (= OfA; ca. 600 b.c.e.–Third Century b.c.e.) The largest number of Aramaic texts from antiquity come from the Old Persian period, during which Aramaic was the lingua franca of the Achaemenid Empire. The material may be divided into several parts (see Greenfield 2001: 111–20): (1) Official Aramaic, the dialect of the administrative documents, which are mostly known from texts found in Egypt; and (2) Standard Literary Aramaic, 5 the dialect of the literary texts (such as Aḥiqar). 2.3.  Early Western Aramaic (Fifth Century b.c.e.) The Hermopolis papyri are a group of letters written in northern Egypt but found in Hermopolis that display some isoglosses found later in firstmillennium c.e. western Aramaic texts. 2.4.  Late Standard Literary Aramaic (= LSLA; ca. 200 b.c.e.–70 c.e.) 1.  Biblical Aramaic (= BA) 6—The language of the Aramaic portions of Ezra and Daniel transmitted in the Bible 2.  Qumran Aramaic (= QA)—A large number of literary texts written by Jews in the region of Israel during the Second Temple period were found in caves along the Dead Sea near Qumran.

2.5.  Judean Aramaic (ca. 150 b.c.e.–150 c.e.) Judean Aramaic was a language used mainly for administrative purposes by Jews in Judea from the Hasmonean to the Bar Kochba periods. 2.6.  Literary Dialects of the First Millennium c.e. 7 1.  Western (= WA) a.  Jewish Palestinian Aramaic (= JPA) b.  Samaritan Aramaic (= SA) c.  Christian Palestinian Aramaic (= CPA) 4. See Hackett 1984; and Hoftijzer and van der Kooij 1991. Because of the problematic nature of this text, it will not be cited in this survey. 5.  For this term, see Greenfield 2001: 111–20. 6.  On the dating of the Aramaic of Daniel, see Kitchen 1965. 7.  For a discussion of the formation of these dialects and their geographic relationships, see Boyarin 1981: 613–49.

Outline of Aramaic Diachrony

381

2.  Eastern (= EA) a.  Syriac (= Syr) b.  Jewish Babylonian Aramaic (= JBA) c.  Mandaic (= Ma)

3.  Writing Systems 3.1.  Adaptation of the Canaanite Alphabet The earliest Aramaic inscriptions are written in the Phoenician script, which was probably already being borrowed toward the end of the second millennium b.c.e. By the fifth century b.c.e., a typical Aramaic script had emerged, and from this point on its historical development can be followed from a large number of original documents that have been preserved (see Naveh 1970; 1982: 78–89). The Jews who were exiled to Mesopotamia after the destruction of the First Temple in 586 b.c.e. exchanged their national script for the Aramaic script, which eventually developed into the Jewish square script that is still being used to this day. 3.2.  Nearly all of the Aramaic material from the first millennium b.c.e. is written with this alphabetic script, with two major exceptions: 1.  The Uruk Incantation (see Müller-Kessler 2002)—This text, which dates from the second century b.c.e. is a forerunner of the Aramaic incantation texts known later at Babylonia from the first millennium c.e. 2.  Papyrus Amherst 63 (see Nims and Steiner 1983; 1985; Steiner 1985; Vlee­ ming and Wesselius 1985; 1990)—A group of Aramaic texts written in the Demotic script and dating to the fourth century b.c.e.

3.3.  In the first millennium c.e., each literary dialect employed its own particular script. Some of the dialects eventually developed vocalization and diacritics to aid in the proper reading of the texts (Morag 1961). 8 4.  The Use of Matres Lectionis 4.1.  The earliest West Semitic alphabetic texts, beginning with Ugaritic dating to the last third of the second millennium b.c.e. and continuing with the earliest Canaanite inscriptions, were written in a purely alphabetic script, in which neither final nor medial vowels were indicated in the writing system. Contrary to this usage, already in the earliest Aramaic inscriptions dating to the ninth century b.c.e., the graphemes and were used to indicate final long vowels (see Degen 1969: 25ff.), and this system was expanded in the later periods. The major exception to this rule in OA is the ending of mp nouns with a 3ms suffix, which is consistently spelled (see Degen 1969: 56) instead of the expected . Since this morpheme is always written (that is, -awhi) in later texts (see, for example, Leander 1928: 29ff.), this is most 8.  Since many of these texts were composed centuries before they were vocalized, the assumption that they actually reflect the phonology and morphology of the various dialects at the time of the composition of the texts is open to question.

382

Michael Sokoloff

likely a remnant of an unattested archaic Aramaic orthography from a period before OA when no final vowels were as yet indicated. Since, in OA, ʾalep was still pronounced in syllable-final position, for example, in the prefix conjugation of I-ʾ verbs and in the morpheme /-aʾ/ of determination, this grapheme could not be employed as a vowel letter. However, with the loss of syllablefinal consonantal ʾalep, from the OfA Period on, it too began to be used as a vowel letter (see Muraoka and Porten 1998: 30). 4.2.  The obligatory use of matres lectionis to indicate final long vowels was expanded throughout the history of the Aramaic language, though the vowels indicated by specific graphemes changed from one period to another. Thus, in OfA, indicated [ē], indicated [ū] or [ō], indicated [ā], [ē], and [ō], and indicated [ā] or [ē]. 9 In later periods, modifications occurred as the result of phonetic changes. Thus, with the contraction in various dialects of the diphthong [ay] > [ē], final was used to indicate [ē], first in mp construct forms and then later in other usages (for example, the ms participle of III-y verbs). When word-final ʾalep was lost in pronunciation, , which had already indicated the ubiquitous post-positional definite article [-aʾ], was now used alongside to indicate final [-ā] elsewhere (for example, in 3ms perfect forms of III-y verbs). Its use alongside to indicate word-final [ē] derives from its loss as a consonant in word-final position in participles of III-ʾ verbs. 4.2.1.  The manner in which final matres lectionis were employed differed diachronically. Thus, on the one hand, in early western Aramaic, was consistently employed for [-ā]. This usage is found sporadically in LSLA and consistently in the first millennium c.e. in both JPA and SA. On the other hand, in Syr and JBA as well as other dialects, is consistently used for this phone. 10 4.3.  According to the earliest Aramaic epigraphic evidence, was often used as a mater lectionis to indicate internal [ū] (see Kutscher 1983: 9–11). 11 By the OfA Period this was nearly obligatory, and was also used to indi9.  Occasionally, archaic orthographies are retained in spite of the general rules. A prime example is the use of word-final in OfA to indicate [-nā] in both the independent 1p pronoun spelled nearly always ‫( אנחן‬instead of the expected ‫ )אנחנא‬and in the 1p perfect verbal morpheme (see Kutscher 1972: 11; Muraoka and Porten 1998: 45). The same is probably true regarding the 2ms pronoun ‫( אנת‬see Muraoka and Porten 1998: 43, and compare the occasional spellings ‫ אנתה‬in the MT of BA). 10.  Note also that, in Syr, final and were consistently written where they originally indicated long (unaccented) unpronounced vowels, for example, in the morphemes of the 1s possessive pronoun and in the 3mp perfect verbal ending. In QA, a special system was used to distinguish between word-final [-ū] and [-ō] and word-final [-ī] and [-ē] that was apparently borrowed from the same usage in Qumran Hebrew texts; see the addition of . Thus, = [ē] and = [ī] (see Sokoloff 1974: 14). This usage had neither precedent nor following. 11. Contra Degen (1969: 27–28), who denied the use of internal matres lectionis for long vowels in OA. In addition to the evidence from Sefire, note additionally now, for example, [ū]: ‫ ישמוה‬Bukān 3; ‫ שורה‬Bukān 5; ‫ אדקור‬Fakh 3; ‫ ססנורי‬Fakh 7; ‫ דמותא‬Fakh 15; ‫ נשון‬Fakh 22.

Outline of Aramaic Diachrony

383

cate [ō] < [aw]. Additionally, by this period, was used to indicate original [ī] (see Muraoka and Porten 1998: 32–33), 12 which was obligatory from this time onward. 13 The use of to indicate original internal [ā] first occurs in the QA texts (see examples in Fitzmyer 2004: 264). This usage, however, was not employed in the first-millennium c.e. western Aramaic dialects, but it was very common in the eastern dialects of both JBA and Mandaic (see Kara 1983: 22–23; Nöldeke 1875: 3). 14 4.4.  The indication of internal short vowels by matres lectionis in the later dialects took place over a long period of time. As in other Semitic languages, Aramaic originally had a tripartite vocalic system in which the three basic vowels were phonemically distinguished both qualitatively and quantitatively, that is [a i u ≠ ā ī ū]. However, by sometime in the second half of the first millennium b.c.e., this system had evolved from a phonemic point of view into a system in which the original quantitative distinctions had now been replaced by qualitative distinctions, and this was maintained afterward in all of the firstmillennium c.e. dialects. 15 4.5.  Two periods should be distinguished: Internal original short vowels were first indicated in Aramaic in the QA texts where represents [u] and [o]. 16 This practice was continued in the first-millennium c.e. literary dialects, where additionally was extensively employed to indicate original short [i] in some dialects (see, for example, for CPA in Müller-Kessler 1991: 37). 4.6.  The diphthongs [ay] and [aw] are indicated from OA onward by and , respectively (see Degen 1969: 28; Muraoka-Porten 1998: 36–38). In various first-millennium c.e. dialects, [ay] is also indicated either by or by . 5. Phonology 5.1.  The Original Aramaic Phonemic Inventory and Its Mergers 17 5.1.1.  The phonemic inventory of OA is considered to be near what is conjectured to have existed in Proto-Semitic. However, since Aramaic was written 12.  Note that a clear exception to this rule in OfA is the common indication of the mp by the archaic OA (instead of the expected ). That this ending should be read [-īn] and not [*-ān] is proved by the facts that the orthography does occasionally occur and that the plural construct is written (see Kutscher 1969: 67–8). 13.  Thus in the LSLA texts (see Sokoloff 1974: 13) and in all of the Aramaic dialects of the first millennium c.e. 14.  The phenomenon is also practically nonexistent in Syr. 15. On this issue, see Kaufman 1983, who deals mainly with the reduction of short vowels to ultra-short vowels. For the situation as it later evolved in the first millennium c.e., see Khan 1997: 107–10; Birkeland 1947; Morag 1961: 45–59; Lipiński 1997: 161–62. 16.  Many examples can be found in Fitzmyer 2004: 263–67. The orthographic practices in these texts parallel the practices found in the Qumran Hebrew texts, in spite of the fact that the Aramaic texts are not sectarian in nature. 17.  For more details and examples, see Degen 1969: 31–38; Muraoka and Porten 1998: 2–10.

384

Michael Sokoloff

with the borrowed Canaanite alphabet, which contained only 22 graphemes, and since the scribes did not invent additional graphemes or add diacritics, the phonemic inventory of OA can only be restored on the basis of comparative evidence. 5.1.2.  The OA scribes solved the problem of writing all of the existing phonemes by using individual graphemes to represent two or occasionally three different consonants. The basis for this assumption is the fact that, by the time of the OfA period, the orthography of certain consonants in many words had shifted but in a manner that was consistent with the comparative assumptions concerning the original consonantal inventory of the Semitic languages. Thus, some words written in OA with are written in OfA with while others are written with . Those that retain correspond to words having /q/ in the other Semitic languages, while those in which occurs correspond to cognates that have /ḍ/ in Arabic and /ṣ/ in Hebrew. Were we to assume that had only one phonetic value in OA, this partial shift could not be explained. The other OA graphemes that do compound duty of this sort are: , , , 18 and . 5.1.3.  The following examples illustrate these changes: 19 1.  = [ḏ] > [d], for example, OA ‫ > זהב‬OfA ‫‘ דהב‬gold’ 2.  = [ẓ] > [ṭ], for example, OA ‫ > צבי‬OfA ‫‘ טבי‬gazelle’ 3.  = [ḍ ] > [ʿ ], 20 for example, OA ‫ > ארק‬OfA ‫‘ ארע‬land’ 4.  , = [ ṯ] > [t], for example, OA ‫עשר‬, Fakh ‫ > עסר‬OfA ‫‘ עתר‬wealth’

5.1.4.  Two other graphemes, and , are retained without change between OA and OfA, but it may be assumed that they also did double duty in the earlier period. 21 5.1.5.  In OfA texts, the scribes occasionally still employed the historical OA spellings and but, from hypercorrect orthographies, we deduce that this did not reflect the actual phonetic state of affairs (see Kutscher 1954: 235). After the OfA period, this archaic orthography disappeared completely (Muraoka and Porten 1998: 3–9). 22 5.1.6.  It is uncertain, however, whether the shift [ś] > [s] was completed by the OfA period, since in this period still dominated for /ś/ in the texts. 18.  This grapheme does double duty only in Fakh for what is expressed in all the other texts by . 19. For details and additional literature, see Muraoka and Porten 1998: 3–9; Folmer 1995: 49–74. 20.  The original phonetic value of this consonant is not completely certain; see Steiner 1977: 57–120. 21.  For Biblical Hebrew, see Blau 1982. 22.  A rare exception is found in BA ‫א ְַר ָקא‬, Jer 10:11. The phonetic explanation for the preservation of and in various words in Mandaic is unclear (see Nöldeke 1875: 43, 72).

Outline of Aramaic Diachrony

385

However, on the basis of the evidence from QA, it is clear that by 200 b.c.e. the shift had been completed in LSLA, even though scribes in the first millennium c.e. occasionally continued to use . 23 5.2.  Phonematization of the [bgdkpt] Consonants As a result of the above phonetic shifts, the number of consonants in Aramaic was reduced to either 22 or 23 by the second half of the first millennium b.c.e. However, about this time, due to partial assimilation of the stop consonants to preceding vowels, a double pronunciation of the six stop consonants developed either independently in Aramaic and in Hebrew or in the latter under the influence of the former, so that two allophonic series developed: (1) stops [bgdkpt] before vowels or when these consonants were geminated; (2) fricatives [v ḡ ḏ x f ṯ] or aspirates [bhghdhkhphth] elsewhere (see Kutscher 1982: 21–22; 1987: 117; Muraoka and Porten 1998: 19–20; Blau 2010: 78–81). Since these two sets of consonants were in complementary distribution, they did not increase the Aramaic phonemic inventory. However, this eventually changed over time, and partial phonematization of these consonants eventually took place. Thus, based on the Masoretic vocalization of BA, [ṯ] and [ḏ] were occasionally independent phonemes, as shown by the minimal pairs [ʾištĕḵaḥaṯ] ‘it (f.) was found’ as opposed to [ʾištĕkaḥat] ‘you were found’; [ʿavḏāḵ] ‘your servants’ [Qere; spelled ‫ ]עבדיך‬as opposed to [ʿavdāḵ] ‘your servant’ (see Sokoloff 1970); and in Syr, all 6 fricatives had a phonemic load (see Muraoka 1997: 12). 5.3. Vowels 24 5.3.1.  The Problem 5.3.1.1.  Because of the nature of the Aramaic writing system, in which vowels were only partially represented in a nonexclusive manner by matres lectionis, the definitive establishment of the phonology of a particular morpheme or lexeme is often nearly impossible for the earlier periods. Moreover, the tendency of West Semitic scribes to employ archaic orthography (“historical spellings”) often obscured the actual pronunciation. 5.3.1.2.  Even for dialects such as BA and Syr that have a vocalization system, establishing the real phonetic values of the grammatical and lexical morphemes at the time of their composition is problematic, since in most cases the vocalization in the manuscripts was added by Masoretes or grammarians at a period much later than the establishment of the consonantal text, and, thus, it often reflects subsequent phonetic and morphemic realities. 23.  For the early Syr inscriptions, see Beyer 1966: 243; for JPA, see DJPA 570–73; for JBA, see DJBA 1188–90. 24.  See Creason 2008: 114–17.

386

Michael Sokoloff

5.3.1.3.  In addition to the Aramaic vowel system becoming completely qualitative from the LSLA period onward, there is also evidence for the reduction of some of the short vowels in open syllables to shewa in some of the dialects, a phenomenon which was completed in nearly all of the dialects by the first millennium c.e. (see Kaufman 1983). 5.4.  Assimilation and Dissimilation 5.4.1. [-nC-] > [-CC-] and [-CC-] > [-nC-] (see Folmer 1995: 74–94) 5.4.1.1.  In OA, as in other West Semitic languages (for example, Hebrew), the assimilation [-nC-] > [-CC-] is unconditional within a root morpheme (for example, ‫‘ ]*אנפין [k] before [ṣ] 28 5.4.2.1.  In first-millennium b.c.e. texts, there are only a few examples of this phenomenon: 1× in an OA text from Samʾal (‫‘ כיץ‬summer’; see Degen 1969: 42), 3× in the Aḥiqar proverbs (‫‘ כצף‬anger’; ‫‘ כצר‬harvest’; ‫‘ כשיט‬honest’), 1× in a text from Elephantine (‫‘ כצת‬part’), and 1× in a text from Babylonia (‫‘ כרץ‬calumny’; see Muraoka and Porten 1998: 18). 5.4.2.2.  Because of the alleged similarity between this dissimilation and a similar dissimilation found in Akkadian as described by Geers’s law, 29 influence on OA by Akkadian has been asserted here. However, this has been rejected, since the dissimilation in Aramaic does not follow the Akkadian rules (see Kaufman 1974: 121–22). Additionally, nearly all of the examples occur in Aramaic texts otherwise not influenced by Akkadian. 30 5.4.3.  Dissimilation in Roots with Two Laryngeal Consonants (see Sokoloff 2000) 5.4.3.1.  In OA, there do not seem to be any roots with two laryngeal consonants. 31 However, as a result of the phonetic shift [ḍ] > [ʿ], which took place sometime in the middle of the first millennium b.c.e. (see above, §5.1.2), a number of roots of this type came into existence. The fact that most roots of this type were still written in OfA with  32 may indicate that the above shift was delayed in these cases because it would have resulted in roots with two laryngeal consonants. 5.4.3.2.  Nevertheless, by the time of LSLA, the phonetic shift [ḍ] > [ʿ] had taken place universally, and a number of lexemes with two laryngeal consonants came into existence, such as: ‫‘ חעך‬to laugh’, ‫‘ חען‬to embrace’, ‫‘ עבע‬to hasten’, ‫‘ עלע‬rib’, ‫‘ עע‬tree, wood’, ‫‘ עעיתא‬rampart’, ‫‘ ערע‬to meet’, ‫‘ רחע‬to wash’, and ‫‘ רעע‬to crush’. 33 While there is possible evidence for this 28.  The nature of “emphasis” (see Creason 2008: 118, pharyngealized consonants) in the ancient Semitic languages is still a debated issue, but since the present discussion deals with the encounter of two emphatic consonants in the same morpheme leading to a change in one of them, its exact nature may be disregarded here. 29. See Geers 1945, where the rules for dissimilation of one of the emphatic consonants in a root morpheme are delineated. 30.  Thus, the Aḥiqar examples are all in the proverbs and not in the framework story (see Greenfield 2001: 164). The former goes back to a western provenance, while the latter shows strong Akkadian influence (see 2001: 115). The only first-millennium c.e. dialect to show dissimilation of this type is Mandaic. 31.  Some scholars have assumed that this dissimilation took place in OA also in the root *mḥḍ > *mḥʿ > mḥʾ (see, for example: Kutscher 1972: 18). However, since there is no unequivocal evidence for this dissimilation in OA, the occurrence of *mḥʾ in OA points to its being an independent root (see Degen 1969: 42). 32.  They are: ‫‘ עק‬tree, wood’, ‫‘ לעבק‬quickly’, ‫‘ לערק‬toward’. The only exceptions are ‫‘ עלע‬rib’ and ‫‘ רחע‬to wash’ (for documentation, see Muraoka and Porten 1998: 8–9). 33.  For details, see Sokoloff 2000: 748–50.

388

Michael Sokoloff

dissimilation in one lexeme of this group already in the fourth century b.c.e., 34 the evidence from QA texts seems to indicate that the two laryngeals in these words were retained until the first-millennium c.e. dialects, when the dissimilation of one of them took place universally in all of these lexemes in all of the Aramaic dialects. 35 6. Morphology 6.1.  Nominal Inflection 6.1.1.  Case Endings and Mimation 6.1.1.1.  Case endings are expressed in the Semitic languages in word-final position by means of the three basic short vowels. Since they are not represented in the nonvocalized writing systems, evidence regarding them must be obtained either indirectly (for example, transcriptions in other writing systems) or from particular nouns (see below). Thus, even though scholars agree that case endings once existed in Aramaic, it cannot be determined whether they still existed in OA, and if so, when they were eventually lost. 6.1.1.2.  In light of this situation, several proposals have been made. Due to the absolute form ‫‘ אנתת‬woman’ that occurs in OfA, Kutscher suggested that the accusative case ending still existed in a particular syntactic formation (see Kutscher 1972: 114; 1977: 40). Wesselius has argued that the ending in fs participles in the Hermopolis papyri, such as ‫‘ הוי יהבת‬do give’, as opposed to ‫יהבה‬, was retained to express the accusative function (see Wesselius 1980; and the discussion in Folmer 1995: 252–57). 6.1.1.3.  Samalian distinguishes between mp nouns, for example, [ū] (nominative) ≠ [ī] (accusative; see Dion 1974: 144–45, 227–28), and the apparent nominative plural form occurs in OA only in ‫‘ נשון‬women’ (Fakh 21, 22). 6.1.1.4.  Remnants of the old ms nominative case ending were retained, however, in all syntactic positions for the biradical nouns ‫ אב‬and ‫ אח‬before possessive pronouns (for example, ‫אבוך‬, ‫ )אחוך‬throughout the history of Aramaic. 36 6.1.1.5.  The noun pVm ‘mouth’ presents interesting evidence with regard to case endings. Though this word functions as a biradical noun in Aramaic (see Nöldeke 1910: 177), etymologically it is derived from the monoradical 34.  This occurs in the word spelled ‫‘ אק‬wood’ in an Aramaic text from Idumea (see Notarius 2006). 35.  Two more examples from the later period that are not attested as yet in the LSLA period may be added here: (1) Syr ‫ܐܥܦ‬, JBA ‫ אפף‬, ‫‘ עפף‬to fold, double’ < * ʿʿp < PS *ḍ ʿp (see DJBA 874; SyrLex 81); (2) Syr ‫ܐܦܥܐ‬, JBA 2# ‫‘ אפא‬hyena’ < *ʿpʿ < PS *ḍp/bʿ (see DJBA 154; SyrLex 88). 36.  Note in OA, however, the defective orthography ‫ אחך‬in the Sefire inscription (see Degen 1969: 36). For OfA, see, for example, Muraoka and Porten 1998: 48–49. For BA, see Rosenthal 1961: 30. For first-millennium c.e. Aramaic, see the examples in JPA (see DJPA 31, 45). The later dialects also have -‫‘ חמו‬father-in-law’ (see, for example, for Syr, Nöldeke 1904: 91).

Outline of Aramaic Diachrony

389

root p(V)- + mimation. 37 Since in OA and OfA, all attestations of this word are spelled defectively, the value of [V] cannot be ascertained directly. 38 However, from the form pu-um-mi-e in the Uruk Incantation (see Hoftijzer and Jongeling 1995: 917), it is clear that in eastern Aramaic the internal vowel was the original nominative [u], and it was retained in all of the later first-millennium c.e. eastern dialects: Syr ‫( ܦܘ�ܡܐ‬SyrLex 1165), JBA ‫( ּפּומָא‬DJBA 889). On the other hand, the form in all of the first-millennium c.e. western Aramaic dialects is pem/pim- (see Kutscher 1976: 20–22; DJPA 437), indicating that in the west the original genitive case ending [ī] dominated in this word throughout, as in Hebrew. 6.1.2.  Number and Gender 6.1.2.1.  As in all of the Semitic languages, Aramaic has three numbers (singular, plural, and dual) and two genders (masculine and feminine). Since internal [ī] was not indicated in OA, the nominal ending ‫ין‬- in this dialect definitely indicates the m.du. [-ayn], such as ‫‘ רחין‬mill’ (Bukān 9); ‫‘ שמין‬sky, heaven’ (Hoftijzer and Jongeling 1995: 1161). While the dual was retained throughout the first millennium b.c.e., in the first millennium c.e. it either merged with the older plural morpheme [-ayyā] (for example, Syr ‫ ;ܫܡܝܐ‬SyrLex 1572) or the later eastern morpheme [-ē] (for example, ‫ ;̈ܪܔ�ܠܐ‬SyrLex 1434). 6.1.2.2.  While the regular plural morpheme for singular nouns ending in a consonant or [-e] is , that is [-īn], there are many exceptions to this rule, especially in the first-century c.e. dialects, such as ‫א ָבהָן‬ ֲ (pl. of ‫‘ אַב‬father’). In Samalian and in the OA Sefire inscription, the old fp absolute ending — that is [-āt]—occurs (see Garr 1985: 94–95). However, elsewhere in OA and in all the later Aramaic dialects, the ending is —that is [-ān]—which developed in analogy to the masculine ending [-īn]. 6.1.2.3.  The mp emphatic ending was [-ayyā] during the entire first millennium b.c.e. 39 While this form was retained in the western dialects in the first millennium c.e., it was replaced by [-ē] in the eastern dialects in most syntactic situations. 6.2.  Pronouns 6.2.1.  Independent Personal Pronouns The 1sc form is ‫אנא‬/‫ אנה‬throughout the entire history of Aramaic. 40 In OA, the 2ms pronoun is ‫ את‬with the expected assimilation of the original [-n-]. In OfA, the dissimilated forms ‫( אנת‬m) and ‫( אנתי‬f) appear. Although we are not 37.  Compare Hebrew ‫ּפֶה‬, Ugaritic p. 38.  The value of the internal vowel is unknown. 39. For Aḥiqar and the Uruk Incantation, see Muraoka and Porton 1998: 39 n. 186; on the alleged influence here of Akkadian, see Kaufman 1974: 127. 40.  Samalian has the additional form ‫אנכי‬/‫( אנך‬see Dion 1974: 150 for a discussion of its relationship to Canaanite).

390

Michael Sokoloff

completely certain, it is likely that the masculine form was pronounced with final [-ā]. 41 In all of the first-millennium c.e. dialects, the final vowels of these two pronouns were reduced to -∅, and they merged to [ʾatt]. 42 The 3m and the 3fs pronouns were both written ‫ הא‬in OA but ‫ הו‬and ‫ הי‬in OfA and ‫ הוא‬and ‫היא‬, respectively, in LSLA. The actual pronunciation of these older forms is debatable, but by the first millennium c.e., they were definitely pronounced [hu] and [hi] in all of the dialects, regardless of orthography. The 1p form )‫ אנחנ(ה‬first appears in OfA, and it was reduced in the first millennium c.e. to either ‫( אנן‬for example, JPA) or ‫( ܚܢܢ‬for example, Syr). The 2mp pronoun occurs once in OfA as ‫אנתם‬, but in QA, it was already being written ‫( אנתון‬see Beyer 1984: 519), 43 the form that dominated in the first millennium c.e. The 3mp pronoun is written ‫ הם‬in OA but ‫המו‬, ‫המון‬, or ‫ אנון‬in OfA and LSLA. The base -‫ המ‬disappears in the first millennium c.e. dialects, and all of the forms then have the base -‫הנ‬/-‫אנ‬, such as Syr ‫( ܗܢܘܢ‬nom.), ‫( ܐܢܘܢ‬acc.). 6.2.2. Demonstrative Pronouns In OA and OfA, both the near and far singular demonstrative pronouns are built around the deictic element [ḏ]/[d]—that is, early , later . Gender is distinguished by a vowel—[e] for masculine and [ā] for feminine. In the first millennium c.e., the demonstrative particle hā can be prefixed to these pronouns, such as JPA ‫הדין‬, ‫הדה‬. Note, however, that in both the eastern and western first-millennium c.e. dialects, new demonstratives without the [d-] element also occur, for example, JPA ‫אהן‬, ‫ ההן‬and Syr ‫ܗܢܐ‬. The plural forms in first-millennium b.c.e. dialects are built around the deictic element -‫( אל‬for example, OA ‫אל‬, ‫)אלן‬, and in the first millennium c.e. this is preceded by the demonstrative particle ‫( הא‬for example, JPA ‫האליין‬, Syr ‫)ܗܠܝܢ‬. The far deictic pronouns can be constructed from the near pronouns either by adding the deictic element ‫ ך‬or by adding (or not adding) the demonstrative element ‫ הא‬to the independent third-person personal pronouns. 44 6.3. Verbs 6.3.1.  Verbal Roots 6.3.1.1.  The number of verbal roots in any particular Semitic language is determined by several factors: 1.  The number of consonants in the root—In historical times, the vast majority of Semitic roots had three consonants. 2.  The number of consonants in the language—All things being equal, a Semitic language that has more consonants in its inventory can form more roots by the permutation of any three of them. 41.  The orthography ‫ אנתה‬occurs in LSLA (see Beyer 1984: 518). 42.  Syr, however, retains the archaic OfA orthography ‫ ܐܢܬܝ‬for the feminine pronoun. 43.  Again in Syr, it is written with a historically unpronounced . 44.  For OfA, see Muraoka and Porten 1998: 58–60; for Syr, see ‫ < ܗܢܘܢ‬hā + ʾennon.

Outline of Aramaic Diachrony

391

3.  Incompatibility rules—Investigation of all of the Semitic languages has shown that there are incompatibility rules that forbid the formation of certain roots (see Greenberg 1990: 365–87). 45 4.  Nonexistence of roots—Many roots that are theoretically possible do not exist.

6.3.2.  Verbal Stems 6.3.2.1.  For any given number of roots in a Semitic language, compound verbal lexical morphemes can be created by interdigitalizing them with specific, discontinuous stem morphemes. The stem morphemes are composed of either vowels alone (G), vowels and length following the second root consonant (D), or vowels and a specific consonant before the first root consonant (for example, H, Š, Y, N). 46 6.3.2.2.  Additionally, any basic stem can be transformed from active to passive (p) meaning by changing the stem vowels, and an additional meaning can be associated with each stem by adding the consonant [t] either before or after the first root consonant. In this manner, each basic stem can give rise to two derived stems. Since Aramaic has three basic stems (see G, D, H/ʾ], the maximum number of stems that can occur in the language is nine, although, as will be seen, this number is generally never reached in any one particular dialect. 6.3.2.3.  The stems in OA are: 47 G D Gp – tG tD –

H – –

Notes

1.  The Gp forms are limited to the prefix conjugation (see Fitzmyer 1995: 194). 2.  The Gt stem is found at Fakh and Sefire. 48

6.3.2.4.  The stems in OfA are: G D Gp Dp [part.] tG tD

H/A – tA

45.  To take an extreme example, a root with three identical consonants (C-C-C) cannot exist in a Semitic language. In spite of this, we do occasionally find such roots which were artificially created, such as the Syr denominative root ‫‘ ܨܨܨ‬to fasten with a nail’ (SyrLex 1300). 46.  Most grammars attribute a semantic value to a particular stem, but though this may occasionally be true, more often it is not. Determining the meaning of a compound verbal morpheme is the task of the lexicographer, not the grammarian. 47.  They are: G = Peʿal; D = Paʿʿel; H/A = H/Afʿel. 48. See ‫ יגתזר‬Fakh 23; ‫ יתשמע‬Sefire I A 29.

392

Michael Sokoloff

Notes

1.  In this period, intervocalic [h] in the H-stem (-VhV-) begins to assimilate to the surrounding vowels, and often this is not expressed in writing. As a result, this stem became associated with the vowel [a] before C1, giving rise to the secondary Afʿel-stem (for details, see Muraoka and Porten 1998: 113–16). 2.  Gp occurs in the suffix conjugation and the participle but essentially never in the perfect conjugation. 49 3.  Dp is limited to participial forms. 4.  There are no examples of the tH stem. It is assumed that [h] regressively assimilated to the preceding [t], which was then lengthened.

6.3.2.5.  The stems in LSLA are: G D H/A Gp Dp [part.] H/Ap tG tD (tA)

Notes

1.  The Gp stem occurs in LSLA only in the suffix conjugation and the part. 2.  H/A stems are used indiscriminately in BA. In QA, the H > A shift can easily be followed: In 11QTgJob, the A-stem is consistently used in intervocalic [-VhV-] position, while the H-stem is used elsewhere. On the other hand, in later texts (for example, the Genesis Apocryphon), only the A-stem is used (for details, see Sokoloff 1974: 15–16). 3.  The Hp stem occurs in BA, 50 while the Ap stem occurs often in QA (see Beyer 1984: 467). 4.  The existence of the tA stem in BA is problematic and is only based on a textual emendation (see Koehler and Baumgartner 2000: 1929, s.v. ‫ נזק‬H).

6.3.2.6.  The stems in first-millennium c.e. Aramaic are: 51 G D A 52 Gp [part.] Dp [part.] Ap [part.] tG tD tA

Notes

1.  The H-stem is now completely replaced by the A-stem. 53 2.  Nonparticipial p-forms have now completely disappeared.

49.  The only two examples that appear during this period are in early western Aramaic: (1) ‫‘ יובל‬let it be delivered’ in the Hermopolis papyri occurring at the end of letters (for example: TAD A 2.7:5); (2) ‫‘ ימנע‬it will be withheld’ (TAD C 1.1:36 [Aḥiqar] 136). 50. See ‫ָת ְקנַת‬ ְ ‫ ה‬Dan 4:33. 51.  The table is valid for all of the known dialects. 52.  Note, however, the unique survival of an H-form in Judean Aramaic ‫‘ התית‬I brought’ (see DSA 33). 53.  Though the common verb ‫‘ הימן‬to believe’ is diachronically derived from ‫ אמן‬H, synchronically it is a quadriliteral root in all of these dialects.

Outline of Aramaic Diachrony

393

3.  The tA-stem is now more prominent, and the t-stems are often employed semantically in place of the p-stems.

6.3.3.  Less-Common Verbal Stems 6.3.3.1.  Alongside the common verbal stems, we find a number of stems that are much rarer, either because they never became productive in Aramaic or because they were limited to particular root patterns. The following examples have in common the fact that they have quadriliteral stems. 1.  Šafʿel (see Rabin 1969; Kaufman 1974: 123–24) 54

In first-millennium c.e. Aramaic, Š-stems are attested only in several verbs in BA. 55 The number of Š-stems greatly increased in first-millennium c.e. dialects, both in the west and in the east. 56 2.  Denominative quadriliterals

There seem to be no examples of denominative quadriliteral roots in firstmillennium b.c.e. Aramaic, but they are abundantly attested in first-millennium c.e. Aramaic dialects. 57 6.3.3.2.  Expanded Triliterals with a Vowel There are no certain examples of expanded triliterals with a vowel in firstmillennium b.c.e. Aramaic. 58 In first-millennium c.e. dialects, expanded triliterals with either [o]/[aw] or [ay] 59 are found in the various dialects. 6.3.4.  Base Expansion of Biliteral Verbal Roots 6.3.4.1.  Since in the Semitic languages, in general, and in Aramaic, in particular, the great majority of the verbal roots are triliteral, there is strong analogic pressure to adapt biliteral roots to the dominant pattern by expansion. Three first-millennium b.c.e. Aramaic roots belong to this category: 54. Both Rabin and Kaufman are of the opinion that the Š-stem was borrowed into Aramaic from a western “Amorite” dialect and not from Akkadian. 55. See ‫‘ שיזב‬to save’; ‫‘ שיציא‬to destroy’; ‫‘ שכלל‬to complete’. 56.  This is especially true in Syr. For examples, see Rabin 1969: passim. Additionally, there are also several Safʿel forms (for example, Syr ‫‘ ܣܪܗܒ‬to hasten’ SyrLex 1043) which may be related. 57.  For TA and JPA, see Dalman 1905: 251–52; for Syr, see Nöldeke 1904: 131. For the conjugation of these verbs, see Heidel 1936: 9–10. 58.  On the form ‫הימן‬, see above, n. 52. The only example of an expanded triliteral with a vowel is BA ‫‘ סובל‬to carry’, possibly derived from ‫( סבל‬see Koehler and Baumgartner 2000: 1936). However, others see in this root a Safʿel from the well-known Aramaic root ‫( יבל‬see Bauer and Leander 1927: 92). 59.  On expanded triliterals with [o]/[aw], see, for example, JPA and TA ‫‘ סובר‬to carry’, ‫‘ סופק‬to provide’, ‫‘ רוקן‬to empty’ (see Dalman 1904: 250–51); Syr ‫‘ ܫܘܬܦ‬to partake’. On expanded triliterals with [ay], see, for example: ‫‘ ܒܝܖܠ‬to be intoxicated’, ‫‘ ܩܝܢܢ‬to work as a smith’, ‫‘ ܫܝܚܠ‬to make intoxicated’ (see SyrLex, s.v.)

394

Michael Sokoloff

1.  ‫( *בוש‬Semitic *bwṯ 60) ‘to be ashamed’—By the OfA period it should have been ‫*בות‬, and in all of the first-millennium c.e. dialects it occurs as ‫( בהת‬for example, JPA ‫בהת‬, DJPA 86). 2.  ‫‘ עוד‬to remember’—This root is still unattested in OA, but it was retained in CPA. However, in Syr, it appears as ‫( ܥܗܖ‬see SyrLex 1074). 3.  ‫( רוץ‬Semitic *rwẓ) ‘to run’—This root is attested in OA as ‫( רוץ‬see Hoftijzer and Jongeling 1995: 1064), in QA as ‫( רוט‬see 11QTgJob 39:21; 1QApGen ar 2:19; and Sokoloff 1974: 155), and in all of the first-millennium c.e. dialects as ‫( רהט‬see DJPA 517).

6.3.4.2. II-w/y verbs are not employed in OA in either the D- or the tDstems. In the first millennium b.c.e., only a few examples are found from OfA onward. 61 However, in all of the first-millennium c.e. Aramaic dialects, these forms are very common. 62 6.3.4.3. II-w/y verbs are rarely adapted to the triliteral model by reduplication of their second radical in first-millennium b.c.e. Aramaic, 63 but this is very common in both TA and JPA in the first millennium c.e. (see Dalman 1904: 326). 6.3.4.4.  Geminate roots are adapted to the triliteral model in the D-stem in first-millennium b.c.e. Aramaic from OA on, 64 and many examples are found in the various dialects. 65 Additionally, these roots occur in some of the firstmillennium c.e. dialects with additional internal [aw] or [o]. 66 6.3.4.5. In first-millennium c.e. Aramaic dialects, II-w/y and geminate roots are often expanded to quadriliteral roots by reduplication of the biliteral base. 67 No examples of this phenomenon are found in the first-millennium b.c.e. dialects. 6.3.5.  Inflectional Categories The following discussion centers around several points of the morphology of the various conjugations but does not discuss the contentious question of their use as either tenses or moods. 60.  Compare H ‫בוׁש‬. 61.  For OfA: ‫( חיב‬see Muraoka and Porten 1998: 132); for BA: ‫( קים‬see Bauer and Leander 1927: 149); for QA: ‫קים‬, ‫( אתקיאם‬see Beyer 1984: 681). The forms from 11QTgJob that Beyer analyzed as ‫ דון‬D (Beyer 1984: 552, s.v. ‫)דין‬, should be interpreted differently (see Sokoloff 1974: ad loc.). 62.  See, for example, for JPA: Dalman 1905: 322–26; for Syr: Nöldeke 1904: 126. 63.  For OA: ‫‘ כנן‬to set up’ (Fakh 10); for BA: ‫רומם‬, ‫( התרומם‬see Bauer and Leander 1927: 146–47); for QA: ‫( התבונן‬see Beyer 1984: 532). 64.  For OA, see Degen 1969: 73; for OfA, see Muraoka and Porten 1998: 134; for BA, see Bauer and Leander 1927: 167; for QA, see Schattner-Rieser 2004: 82. 65.  For example, for JPA, see Dalman 1904: 335; for Syr, see Nöldeke 1904: 128. 66.  For example, JPA ‫קורר‬, ‫אתקורר‬, Syr ‫ܩܘܪܪ‬, ‫‘ ܐܬܩܘܪܪ‬to cool, be cooled’ (see DJPA 507; SyrLex 1346). 67.  For TA and JPA, see Dalman 1904: 326–27, 336–37; for Syr, see Nöldeke 1904: 131.

Outline of Aramaic Diachrony

395

6.3.5.1.  Suffix Conjugation 6.3.5.1.1.  In the first millennium b.c.e., the morphemes for 2ms/fs forms are written ‫ת‬-/‫תי‬- (see Muraoka and Porten 1998: 97–99). On the basis of comparative Semitics and the plene spelling ‫תה‬- in LSLA, the original contrast is interpreted as [-tā]/[-tī] (see Bauer and Leander 1927: 101). However, in LSLA, the final vowel of the 2ms suffix was disappearing, and it became [-t] in all of the first-millennium c.e. dialects. The 2fs suffix in LSLA most likely remained [-tī], since this form was kept in all of the first-millennium c.e. western dialects, 68 though it merged with the 2ms to [-t] in the eastern dialects. 69 6.3.5.1.2.  From a comparative Semitic point of view, the Aramaic 3fp morpheme of the suffix conjugation was originally [-ā], although it is not yet attested in OA (see Degen 1969: 64; Muraoka and Porten 1998: 101–2). In both OfA 70 and in BA (Kethiv; see Bauer and Leander 1927: 101), this morpheme is spelled [= ū]—that is, just as had happened in standard Biblical Hebrew, it had merged with the 3mp morpheme. 71 However, in BA (Qere) and in QA, this morpheme was retained as [-ā], indicating that it had not yet completely disappeared in the first millennium b.c.e. 72 In the first-millennium c.e. dialects, there is considerable phonological variation in this morpheme, but the 3p merged form [-ū] for masculine and feminine is not employed in any of the dialects. The original [-ā] morpheme is retained in both TA (see Dalman 1904: 255) and JBA (see Kara 1983: 146), 73 while the unstressed 3mp/fp morpheme [-V] is reduced to -∅ in Syr (see Nöldeke 1904: 104], JBA (see Kara 1983: 146), and Mandaic (see Nöldeke 1875: 223). In the western dialects, a new 3fp morpheme [-ī/ē(n)] appears in all of the dialects. 74 6.3.5.2.  Long-Prefix Indicative Conjugation 6.3.5.2.1.  It has been argued that, similar to the suffix conjugation, the 3fp form in OA merged also with the 3mp in the long prefix conjugation (see Ben-Ḥayyim 1951). Although the OA orthography /y-n/ is defective for both 68.  For example, in CPA (see Müller-Kessler 1991: 152–53). 69.  For example, Syr, where the historical orthography is kept (see Nöldeke 1904: 104). 70. See ‫‘ מטו אגרתא‬the letters arrived’ TAD A 4.7:12. The merger is also attested for Nabatean (see ‫ וכליבת ברתה לנפשהם‬. . . ‫[‘ עבדו כמכם ברת ואלת‬the sepulcher] which PN1 and PN2 made for themselves’ [CIS 2/1 198:1; and Cantineau 1930: 76]) and Palmyrene (see ‫ לא אסקו‬. . . ‫‘ עבידן שגין‬many things were not entered’ [Trf (I):5; and Rosenthal 1936: 58]) 71.  For a discussion of the issue, see Ben-Ḥayyim 1951; Kutscher 1977: 118–19; Sokoloff 1974: 19–20. 72.  Two examples occur in the Genesis Apocryphon: ‫(‘ חבלא בזיתא‬the winds) damaged the olive tree’, 13:16; ‫‘ עשר שנין שלמא‬ten years have passed’, 22:28; 73.  Occasionally, the morpheme [-ān] also occurs, for example, ‫‘ נתרן עיניה‬his eyes fell out’ b. Soṭah 13a (30) (see Epstein 1960: 34–35). 74.  For CPA, see Müller-Kessler 1991: 152; for SA, see Macuch 1982: 143; for JPA, see Fassberg 1990: 165. For a summary table of the morpheme in the various Aramaic dialects, see Fassberg 1990: 233 n. 10.

396

Michael Sokoloff

the 3mp and the 3fp morpheme in the long prefix conjugation (see Degen 1969: 65) and therefore inconclusive on this point, Kutscher pointed out that the OA jussive form of the 3mp is yqṭlw (for example, ‫ יסרו‬Sefire I C 15; ‫ ישמו‬Sefire I C 23), while that of the 3fp is yqṭln (for example, ‫ אל יהרגן‬Sefire I A 24). 75 This clearly indicates that the OA 3fp form ended in [-ān] (see Kutscher 1972: 29; 1977: 146; but see Huehnergard 1987). Similarly, nearly all of the evidence from OfA and LSLA shows that the 3fp form was yiqtĕlān (see Muraoka and Porten 1998: 102–3; Sokoloff 1974: 19–20). 76 All of the first-millennium c.e. dialects clearly distinguish between these two forms (such as Syr ‫)ܢܩܜܠܘܢ– ܢܩܜܠܢ‬. 6.3.5.3.  Prefix Modal Conjugation 6.3.5.3.1.  In the first-millennium b.c.e. texts, modal forms are characterized by several points: 1.  Some dialects 77 employ the prefix l- in third-person forms, possibly under the influence of the Akkadian precative liprus-form (see Kaufman 1974: 124–26). Additionally, the stem vowel between C2 and C3 often changes. 78 2.  Other dialects, such as OA, use the regular y- prefix. 3.  Final nun is apocopated in the 2fs, 2mp, and 3mp forms where it occurs (see Degen 1969: 65; Muraoka and Porten 1998: 104–5). 4.  For forms ending in a root consonant, [-inn-] is not added before an accusative pronominal suffix. 79 5.  For verbal forms ending in [-ūn], [n] is deleted before an accusative pronominal suffix.

As a result of the restructuring of the tense system in the first millennium c.e., the above structure was completely remodeled. 80 6.3.5.4.  Internal Passives 6.3.5.4.1.  As was pointed out above, OA employs the conjugated Gp-stem in the prefix conjugation, while in OfA, SLA, and LSLA, essentially only forms of the suffix conjugation are found. 81 6.3.5.4.2.  There is no certain evidence for the existence of conjugated Dp forms in Aramaic, 82 but beginning with the OfA period, Dp participial forms 75.  Note, however, the 3fp jussive forms that have merged with 3mp jussive forms in the Bukān inscription: ‫‘ יאפו‬they will bake’, line 7; ‫‘ ימלאוהי‬they will fill it’, line 8. 76.  The only first-millennium b.c.e. dialect in which there was a merger with the 3mp form is that of the Aḥiqar proverbs (see Kottsieper 1990: 138). 77.  See Samalian, the Assur letter, Fakh, SLA (only the root ‫)הוי‬. 78.  This can be clearly seen only in II-y and III-weak roots. 79.  For example, ‫‘ ימחאהי‬may it smite him’ (Bukān line 13). 80.  For details, see §7.2. 81.  For the two exceptions, see above, §6.3.2.4. 82.  All of the examples referred to in Beyer 2004: 331 should be rejected: (1) BA ‫ְיבַּקַ ר‬ in Ezek 4:15 is a D form (see HALOT 1837); (2) BA ‫ ְּכ ִפתּו‬in Dan 3:21, emended to ‫ֻּפתּו‬ ִ ‫*ּכ‬ on the basis of ‫ָתא‬ ָ ‫ ַּכּפ‬v. 20 (see Bauer and Leander 289 n. 1). This is unnecessary since the

Outline of Aramaic Diachrony

397

occur (see Muraoka and Porten 1998: 120; Bauer and Leander 1927: 112), 83 and they are frequent in all of the first-millennium c.e. dialects. 84 Hp suffixed forms are not attested in OA, are rare in OfA, 85 but are very common in LSLA. 86 Hp participial forms are common in all of the first millennium c.e. dialects. 6.3.5.5.  Infinitives 87 6.3.5.6.1.  Like Biblical Hebrew, OA had two morphologically distinct infinitival forms for the absolute and construct infinitives. The former occurs only in the Sefire inscriptions and serves to strengthen a following conjugated verbal form. 88 While this syntagm is known later from first-millennium c.e. Aramaic dialects, only one morpheme was employed in these dialects for both infinitival constructions. 89 6.3.5.6.2.  In all of the dialects, the infinitives construct may be divided into two groups: the G-stem and the remaining basic stems. Since an m-prefix form of the G-stem is found in all of the dialects from OfA on, it was once thought that the non-m-prefix forms in this dialect were influenced by Hebrew. 90 However, with the non-m-prefix forms’ occurrence in the OA inscriptions and in Samalian, it became clear that it was a genuine Aramaic form. On the other hand, the occurrence now of the m-prefix infinitive of the G-stem in Fakh (see Kaufman 1982: 151) shows that this form already existed in OA alongside the shorter form. root ‫ כפת‬G is widely attested in the meaning ‘to bind’ in many first-millennium c.e. dialects (see, for example, DJPA 268; SyrLex 645); (3) QA ‫‘ כסית‬it was covered’ may reasonably be interpreted as Gp, since ‫ כסי‬G is attested in first-millennium c.e. Aramaic (see, for example, DJPA 265 [pass.]; SyrLex 639). 83.  A possible Dp participle may occur in Bukān: [‫‘ ארקה תהוי ממלח[ה‬may his soil be salted’, line 9 (see Sokoloff 1999: 112–13). 84.  Besides the many examples in the literary dialects, note the following epigraphic examples from earlier Judean Aramaic: ‫‘ משלם‬paid’ (TDT1 131:10); ‫‘ מקבל‬received’ (DSA 76); ‫‘ מעניה‬afflicted one’ (DSA 72); ‫‘ מכסיא‬clothed’ (DSA 57). 85. See ‫ הפקדו‬TAD B 2.9:7. 86.  For BA, see ‫ הּובַד‬Dan 7:11; ‫ֵיתיּו‬ ָ ‫ ה‬Dan 3:13; ‫ הּוסֲפַת‬Dan 4:33; ‫ָרבַת‬ ְ ‫ ָהח‬Ezek 4:15; ‫ָה ְנחַת‬ Dan 5:20; ‫ ֻהּסַק‬Dan 6:24; ‫ ֻהעַל‬Dan 5:13; ‫ֳקימַת‬ ִ ‫ ה‬Dan 7:4; ‫ָת ְקנַת‬ ְ ‫ ה‬Dan 4:33; for QA, see ‫אדבקת‬ Beyer 1984: 546; ‫ אחזית‬Beyer 1984: 575; ‫ אחלמת‬Beyer 1984: 580; ‫ אחלפת‬Beyer 1984: 580; ‫ החשיו‬Beyer 1984: 586; ‫ הובל‬Beyer 1984: 592; ‫ אעברת‬Beyer 1984: 651; ‫ ארחקת‬Beyer 1984: 694; ‫ אשלטת‬Beyer 1984: 710; ⸢‫ אחוא⸣ת‬Beyer 1994: 393; ‫ אודעת‬Beyer 1994: 407. 87.  For a survey with previous literature, see Folmer 1995: 189–98. 88.  For example, ‫‘ הסכר תהסכרהם‬you must surely hand them over’, Sefire III 2 (see Fitzmyer 1995: 144; Degen 1969: 116–17; Kutscher 1972: 32). 89.  For JBA, see Margolis 1910: 86; for Syr, see Nöldeke 1904: 235–38. 90.  This was especially the case regarding the form ‫לאמר‬, for introducing direct speech, which is often used in the Elephantine papyri written by Jews (see Cowley 1923: 13: ‫לאמר‬: a Hebraism, commonly used to introduce the business). Leander (1928) does not even cite the form.

398

Michael Sokoloff

6.3.5.6.3. The infinitives of the derived conjugations (D, H, Š, Gt, Dt, Ht) may be divided into three groups: 1.  Without m-prefix and without a vocalic ending (see Kaufman 1982: 150–51; Muraoka 1983–84: 99–100)—These forms are found only in OA (for example, ‫‘ לחיי‬to give life’ [Fakh 7]). 2.  Without m-prefix and with the [-āCā] ending—This is the dominant form in OfA and LSLA (that is, ‫קטלה‬, etc.), and it was retained in the first millennium c.e. in TA (‫קטלא‬, etc.). Before pronominal suffixes, these forms have a construct form ending in ‫ת‬- or ‫ות‬-. 3.  With the m-prefix and the [-āCā/-āCū] ending—These forms occur in early western Aramaic in the Aḥiqar proverbs 91 and in the Hermopolis papyri (for example, ‫[ למחתה‬TAD A 2.5:6]; ‫[ למושרתהם‬TAD A 2.2:13]). Similar forms appear in first-millennium c.e. Aramaic in JPA (‫ )מקטלה‬and in Syr (‫)ܡܩܜܠܘ‬. 92

7. Syntax 7.1.  Expression of the Construct State and Its Replacement with the Genitive Particle zy/dy (see Kaufman 1974: 130–32; Kutscher 1977: 95–96) 7.1.1.  Like other Semitic languages, Aramaic can express a relationship between two adjacent nouns by placing the first in the genitive state, such as ‫ שר חזרך‬wall of GN (Zakkur A 10). This syntagm is used exclusively in all of the OA inscriptions, 93 except for the bilingual Fakh inscription, where the genitive zy-construction is employed on the Akkadian model known already from the Old Babylonian period. 94 By the OfA period, this usage had become ubiquitous in Aramaic and had also begun to spread to SLA and early western Aramaic (see Folmer 1995: 259–325; Muraoka and Porten 1998: 228–32). 7.1.2.  Closely related to this syntagm is the periphrastic genitive construction with a proleptic pronoun appended to the first member of the phrase, agreeing in gender and number with the second member. Attempts to find semantic distinctions between the construct state syntagm and its replacement with the the genitive particle zy/dy have not been entirely successful (see Muraoka 1966; Garr 1990).

91.  Only with pronominal suffixes, for example, ‫ למנחתותה‬TAD C 1.1:171. 92.  For the significance of these forms in Syr, see Boyarin 1981: 622–32. The forms with m-prefix in CPA are corruptions from Syr (see Müller-Kessler 1991: 170, 177). Infinitives have essentially disappeared in this dialect. Nouns formed from infinitives of derived stems (for example, ʾkn ʿwtkwn; Müller-Kessler 1991: 177) show that they derive from forms without m-prefix. 93.  The phrase ‫( מלכיא זי סחרתי‬Sefire III 7/8) is not an exception (see Kaufman 1974: 130 n. 74). Translate: ‘all the kings who are of my surrounding area’. 94.  For example, ‫( מאניא זי בת הדד‬line 16/17) = unūtē ša bīt Addu (line 27) ‘vessels of the Temple of Hadad’.

Outline of Aramaic Diachrony

399

7.1.3.  In all of the first-millennium c.e. Aramaic dialects, the breakup syntagms dominate and the construct state recedes (see, for example, for Syr: Nöldeke 1904: 161–63). While the construct state did not entirely disappear, it became restricted mainly to phrases beginning with bet ‘house’ and bar ‘son, member of’ (see, for example, Sokoloff 2009: 145–50, 178–85). 7.2.  Incorporation of the Participle into the Verbal System 7.2.1.  OA had two main non-modal verbal conjugations (see Li 2010): 1.  Suffixed—This expressed various past actions (for example, simple past, past perfect) and, in conditional clauses, actions not localized in the past. 2.  Prefixed—This expressed a variety of actions, for example, future, present, generic.

7.2.2.  In OA, active participles formed from verbal roots expressed substantives. 95 However, by the OfA period, they began to take on the additional temporal function of the present tense (see Muraoka and Porten 1998: 203). In LSLA, these usages were expanded to express the present tense and acquired various progressive and imperfective functions (see Li 2010: 76–85). In the first-millennium c.e. Aramaic dialects, the participle is an integral part of the verbal conjugation system. This can be clearly seen, for example, from the synthetic forms of the participle with nominative pronominal suffixes in JBA: Singular 1.  ‫ קטילנא‬ 2.  ‫ קטלת‬ 3.  ‫קטיל‬, f. ‫ קטלא‬

Plural ‫קטלינן‬ ‫קטליתו‬ ‫קטלי‬, f. ‫ן‬/‫קטלא‬

Semantically, these forms can express a variety of tenses (for example, preterite, present, future mainly in independent clauses), and they can even be used modally. 8. Vocabulary 8.1.  Original Lexical Items: Loss and Gain Like all other aspects of language, vocabulary is subject to change over time. Older words become obsolete and disappear from a language while new words appear from both indigenous and foreign sources. 96 Until the nineteenth century, with the exception of BA, Aramaic was known exclusively from first millennium c.e. literary sources. Since then, archaeological discoveries have greatly broadened our perspective of this language, going back to the ninth century b.c.e. Many previously unattested Aramaic vocabulary items 95.  For example, ‫‘ כל עלל בית מלך‬all who enter the palace’ (Sefire I A 6). 96.  For the appearance of many new verbal roots in Mishnaic Hebrew that were unknown in Biblical Hebrew, see Moreshet 1980.

400

Michael Sokoloff

have appeared in these new texts, and it is now possible to follow their existence diachronically. The following examples from OA exemplify this issue without attempting to be exhaustive. 8.1.1.  Several OA vocabulary items show that words that had been known from other West Semitic languages were employed in Aramaic in this period but disappeared later on. The following OA examples are culled from DNSWI: ‫‘ א(י)ש‬man’ (DNSWI 115), later ‫אנש‬, ‫‘ הרג ;גבר‬to kill’ (DNSWI 293), later ‫‘ חפץ ;קטל‬affair’ (DNSWI 396), later ‫‘ חץ ;מלה‬arrow’ (DNSWI 397), later ‫גר‬, det. ‫‘ לקח ;גירא‬to take’ (DNSWI 581), later ‫נסי‬, ‫נסב‬, ‫‘ מחנה ;נטל‬camp’ (DNSWI 614), later ‫משרי‬, det. ‫משריתא‬. 8.1.2.  Other words that were employed until OfA also disappeared in the later dialects, for example, ‫‘ כהל‬to be able’ (DNSWI 489), later ‫השכח‬, ‫יכל‬ (see Folmer 1995: 634–40); ‫‘ לוד‬to efface’ (DNSWI 568), later ‫‘ לחי ;מחק‬evil’ (DNSWI 571), later ‫באיש‬. 8.2.  Suppletive Verbal Roots 8.2.1.  OA and OfA originally had two roots in the G-stem for ‘to walk’: ‫ אזל‬and ‫*הוך‬. The former was apparently used in all the verbal conjugations, 97 while the latter only occurred in the prefix conjugation and in the infinitive (see Porten and Lund 2002: 84). By the second half of the first millennium b.c.e., ‫ אזל‬was no longer employed in the imperfect and the infinitive of the G-stem where it was suppleted by ‫*הוך‬. This situation obtained throughout the first millennium b.c.e. However, in the first millennium c.e., in most of the Aramaic dialects, ‫ אזל‬was again used, and ‫ *הוך‬disappeared. 98 8.2.2.  Similarly, OA and OfA had two roots in the G-stem for ‘to give’ which were originally used in all verbal conjugations: ‫ נתן‬and ‫( יהב‬Folmer 1995: 641–48). 99 Eventually, in the second half of the first millennium b.c.e., ‫ יהב‬was no longer employed in the imperfect and infinitive where it was suppleted by ‫נתן‬. This situation remained the same in all of the Aramaic dialects in the first millennium c.e. 100 Additionally, in OfA, both roots were used in the tG conjugation (see Porten and Lund 2002: 146, 237). 101 However, in the first-millennium c.e. Aramaic dialects, only ‫ יהב‬tG was used (see, for example, DJPA 236; SyrLex 566).

97.  Thus, we find it in the prefix conjugation form ‫( תאזל‬Sefire I B 39) and in the infinitive ‫( למאזל‬TAD D 7.1:9). 98.  In the first millennium c.e., it was still used in TA and SA. 99. On ‫ יהב‬in the imperative, see Degen 1969: 74 n. 73; on ‫ נתן‬in the perfect, see Porten and Lund 2002: 236. 100.  The original root ‫ נתן‬became ‫ ܢܬܠ‬in Syr and ‫ נתב‬in JBA. 101.  BA employs only ‫ יהב‬Gt (see HALOT 1889).

Outline of Aramaic Diachrony

401

9. Conclusion The Aramaic language is attested continuously for almost three millennia from a variety of sources: epigraphic texts, literary compositions, and spoken contemporary dialects. The present survey deals only with evidence from first-millennium b.c.e. and first-millennium c.e. texts and has shown the development of select phonological, morphological, syntactic, and lexical items. Because the provenience and date of nearly all of the first-millennium b.c.e. sources are known, the diachrony of these items can be followed. Where apparent discrepancies appear, they can often be accounted for by our understanding of the sources of these texts occasionally reflecting different contemporary dialects in the past. Additionally, it is not difficult to follow the later progression of the various diachronic aspects of the Aramaic language in the first millennium c.e., when Aramaic split into two major dialect areas, both subdivided along confessional lines. Indeed, on the basis of the diachronic information at our disposal, it appears that there will be no problem in classifying any new unprovenienced Aramaic text. Aramaic Texts CAL Comprehensive Aramaic Lexicon. http://cal1.cn.huc.edu. Bukān CAL 14200 Fakh CAL 11200 Trf CAL 41301 References

Bauer, H; and Leander, P. 1927 Grammatik des biblisch-aramäischen. Halle: Max Niemeyer. Ben-Ḥayyim, Z. 1951 The Feminine Plural in Ancient Aramaic. ErIsr 1 (Schwabe volume): 135–39. Beyer, K. 1966 Der reichsaramäische Einschlag in der ältesten syrischen Literatur. ZDMG 116: 242–54. 1984 Die aramäischen Texte vom Toten Meer. Göttingen: Vanderhoeck & Ruprecht. 1986 The Aramaic Language. Göttingen: Vanderhoeck & Ruprecht. 1994 Die aramäischen Texte vom Toten Meer: Ergänzungsband. Göttingen: Vanderhoeck & Ruprecht. 2004 Die aramäischen Texte vom Toten Meer. Vol. 2. Göttingen: Vanderhoeck & Ruprecht. Birkeland, H. 1947 The Syriac Phonematic Vowel System. Pp. 13–39 in Festskrift til Professor Olaf Broch. Oslo: Norwegian Academy of Sciences. Blau, J. 1982 On Polyphony in Biblical Hebrew. Jerusalem: Israel Academy of Sciences and Humanities.

402

Michael Sokoloff

2010 Phonology and Morphology of Biblical Hebrew: An introduction. LSAWS 2. Winona Lake, IN: Eisenbrauns. Boyarin, D. 1981 An Inquiry into the Formation of the Middle Aramaic Dialects. Pp. 615–49 in Bono Homini Donum: Essays in Historical Linguistics in Memory of Alexander Kerns, ed. Y. L. Arbeitman and A. R. Bomhard. Amsterdam Studies in the Theory and History of Linguistic Science 4; Current Issues in Linguistic Theory 16. Amsterdam: Benjamin. Cantineau, J. 1930 Le Nabatéen. 2 vols. Paris: Leroux. Cowley, A. E. 1923 Aramaic Papyri of the Fifth Century b.c. Oxford: Clarendon. Creason, S. 2008 Aramaic. Pp.  391–426 in The Ancient Languages of Syria–Palestine and Arabia, ed. R. D. Woodward. Cambridge: Cambridge University Press. Dalman, G. 1905 Grammatik des jüdisch-palästinischen Aramäisch. Leipzig: Hinrichs. Degen, R. 1969 Altaramäische Grammatik. Wiesbaden: Deutsche Morgenländische Gesellschaft. Dion, P. E. 1974 La langue de Ya’udi. Waterloo: Academic Studies in Religion in Canada. Epstein, J. N. 1960 A Grammar of Babylonian Aramaic. Jerusalem: Magnes. [Hebrew] Fassberg, S. E. 1990 A Grammar of the Palestinian Targum Fragments from the Cairo Geniza. Atlanta: Scholars Press. Fitzmyer, J. A. 1995 The Aramaic Inscriptions of Sefire. 2nd rev.  ed. Rome: Pontifical Biblical Institute. 2004 The Genesis Apocryphon of Qumran Cave I: A Commentary. 3rd ed. Rome: Pontifical Biblical Institute. Folmer, M. L. 1995 The Aramaic Language in the Achaemenid Period. Louvain: Peeters. Garr, W. R. 1990 On the Alternation between Construct and dī Phrases in Biblical Aramaic. JSS 35: 213–31. 2004 Dialect Geography of Syria–Palestine, 1000–586 b.c.e. Repr. Winona Lake, IN: Eisenbrauns. Geers, F. W. 1945 The Treatment of Emphatics in Akkadian. JNES 4: 65–67. Greenberg, J. H.; Denning, K. M.; and Kemmer, S. 1990 On Language: Selected Writings of Joseph H. Greenberg. Stanford, CA: Stanford University Press. Hackett, J. A. 1984 The Balaam Text from Deir ʿAlla. HSM 31. Chico, CA: Scholars Press.

Outline of Aramaic Diachrony

403

Heidel, A. 1940 The System of the Quadriliteral Verb in Akkadian. Chicago: The University of Chicago Press. Hetzron, R., ed. 1997 The Semitic Languages. London: Routledge. Hoftijzer, J., and Jongeling, K. 1995 Dictionary of the North-West Semitic Inscriptions. 2 vols. Leiden: Brill. Hoftijzer, J., and van der Kooij, G. 1991 The Balaam Text from Deir ʿAlla Re-evaluated: Proceedings of the International Symposium Held at Leiden, 21–24 August 1989. Leiden: Brill. Huehnergard, J. 1987 The Feminine Plural Jussive in Aramaic. ZDMG 137: 266–76. 1995 What Is Aramaic? Aram 7: 262–82. Kara, Y. 1983 Babylonian Aramaic in the Yemenite Manuscripts of the Talmud. Jerusalem: Hebrew University. [Hebrew] Kaufman, S. A. 1974 The Akkadian Influences on Aramaic. Chicago: University of Chicago Press. 1982 Reflections on the Assyrian-Aramaic Bilingual from Tell Fakhariyeh. Maarav 3: 137–76. 1983 Aramaic Vowel Reduction. Pp. 47–55 in Arameans, Aramaic, and the Aramaic Literary Tradition, ed. M.  Sokoloff. Ramat-Gan: Bar-Ilan University Press. Khan, G. 1997 Jewish Palestinian Aramaic Phonology. Pp. 103–13 in Phonologies of Asia and Africa, ed. A. S. Kaye and P. T. Daniels. Winona Lake, IN: Eisenbrauns. Kitchen, K. A. 1965 The Aramaic of Daniel. Pp. 31–79 in Notes on Some Problems in the Book of Daniel, ed. D. J. Wiseman. London: Tyndale. Kottsieper, I. 1990 Die Sprache der Ahiqarsprüche. Berlin: de Gruyter. Kutscher, E. Y. 1954 New Aramaic Texts. JAOS 74: 233–48. 1972 A History of Aramaic, Part 1. Jerusalem: Academon. [Hebrew] 1976 Studies in Galilean Aramaic. Ramat-Gan: Bar-Ilan University Press. 1977 Hebrew and Aramaic Studies. Jerusalem: Magnes. Leander, P. 1928 Laut- und Formlehre des Ägyptisch-Aramäischen. Gothenburg: University of Gothenburg. Li, T. 2010 The Function of the Active Participle in the Aramaic of Daniel. Pp. 69–104 in Aramaic in Postbiblical Judaism and Early Christianity: Papers from the 2004 National Endowment for the Humanities Summer Seminar at Duke University, ed. E. M. Meyers and P. V. M Flesher. Winona Lake, IN: Eisenbrauns. Lipiński, E. 1997 Semitic Languages: Outline of a Comparative Grammar. Louvain: Peeters.

404

Michael Sokoloff

Margolis, M. 1910 A Manual of the Aramaic Language of the Babylonian Talmud. Munich: Beck. Morag, S. 1961 The Vocalization Systems of Arabic, Hebrew, and Aramaic. The Hague: Mouton. Moreshet, M. 1980 ‫( לקסיקון הפועל שנתחדש בלשחון התנאים‬A Lexicon of New Verbs in Tannaitic Hebrew). Ramat-Gan: Bar-Ilan University Press. Müller-Kessler, C. 1991 Grammatik des Christlich-Palästinisch Aramäischen. Hildesheim: Olms. 2002 Die aramäische Beschworung und ihre Rezeption in den Mandäischmagischen Texten. Res Orientalis 14: 193–208. Muraoka, T. 1966 Notes on the Syntax of Biblical Aramaic. JSS 11: 151–67. 1983–84  The Tell-Fekherye Bilingual Inscription and Early Aramaic. Abr-Nahrain 22: 79–117. 1997 Classical Syriac: A Basic Grammar with a Chrestomathy. Wiesbaden: Harrassowitz. Muraoka, T., and Porten, B. 1998 A Grammar of Egyptian Aramaic. Leiden: Brill. Naveh, J. 1970 The Development of the Aramaic Script. Jerusalem: Israel Academy of Sciences and Humanities. 1982 Early History of the Alphabet. Jerusalem: Magnes. Nims, C., and Steiner, R. 1983 A Paganized Version of Psalm 20:2–76. JAOS 91: 261–74. Nöldeke, T. 1875 Mandäische Grammatik. Halle: Waisenhauses. Nöldeke, T., and Crichton, J. A. 1904 Compendious Syriac Grammar. London: Williams & Norgate. [Repr. Winona Lake, IN: Eisenbrauns, 2001] Notarius, T. 2006 ?q(n) ‘Wood’ in the Aramaic Ostraca from Idumea: A Note on the Reflex of Proto-Semitic /*ṣ/ in Imperial Aramaic. Aramaic Studies 4/1: 101–9. Porten, B., and Lund, J. 2002 Aramaic Documents from Egypt: A Key-Word-in-Context Concordance. Comprehensive Aramaic Lexicon Project 1. Winona Lake, IN: Eisenbrauns. Porten, B., and Yardeni, A. 1986–99  Textbook of Aramaic Documents from Ancient Egypt. 4 (A–D) vols. Jerusalem: Academon. Rabin, C. 1969 The Nature and Origin of the Šaf   ʿel in Hebrew and Aramaic. ErIsr 9 (Albright volume): 148–58. Rosenthal, F. 1936 Die Sprache der palmyrenischen Inschriften und ihre Stellung innerhalb des aramäischen. Leipzig: Hinrichs.

Outline of Aramaic Diachrony

405

Schattner-Rieser, U. 2004 L’araméen des manuscrits de la Mer Morte. Lausanne: Zèbre. Sokoloff, M. 1970 /t-ṯ/–Two Phonemes in Biblical Aramaic. Leš 34: 230 [Hebrew]. 1974 The Targum to Job from Qumran Cave XI. Ramat-Gan: Bar-Ilan University Press. 1999 The Old Aramaic Inscription from Bukan: A Revised Interpretation. IEJ 49 (1/2): 105–15. 2000 Qumran Aramaic in Relation to the Aramaic Dialects. Pp.  746–54 in The Dead Sea Scrolls: Fifty Years after their Discovery, ed. L. Schiffman et al. Jerusalem: Israel Exploration Society and The Shrine of the Book. 2002a A Dictionary of Jewish Babylonian Aramaic of the Talmudic and Geonic Periods. Ramat-Gan: Ba- Ilan University Press / Baltimore: Johns Hopkins University Press. 2002b A Dictionary of Jewish Palestinian Aramaic of the Byzantine Period. 2nd ed. Ramat-Gan: Bar-Ilan University Press / Baltimore: Johns Hopkins University Press. 2003 A Dictionary of Judean Aramaic. Ramat-Gan: Bar-Ilan University Press. 2009 A Syriac Lexicon. Winona Lake, IN: Eisenbrauns / Piscataway, NJ: Gorgias. Steiner, R. 1977 The Case for Fricative-Laterals in Proto-Semitic. New Haven, CT: American Oriental Society. 1985 Ashurbanipal and Shamash-shum-ukin: A Tale of Two Brothers from the Aramaic Text in Demotic Script. RB 92: 60–79. Vleeming, S. P., and Wesselius, J. W. 1985–90  Studies in Papyrus Amherst 63, I–II. Amsterdam: Juda Palache Instituut. Yardeni, A. 2000 Textbook of Aramaic, Hebrew and Nabataean Documentary Texts from the Judaean Desert, A: The Documents. Jerusalem: Dinur Center for Research in Jewish History. [Hebrew]

Diachrony in Ugaritic Joseph Lam and Dennis Pardee 1. Introduction To what extent is diachronic change discernible in Ugaritic? The significance of this question is self-evident for Ugaritologists, given that the linguistic dating of texts has ramifications for both historical reconstruction and textual interpretation. However, it is also of potential interest to scholars of Biblical Hebrew, for from the earliest days of the study of the Ugaritic texts, there has been a tendency to view them in comparison with the biblical texts. Despite the excesses wrought by the application of the comparative approach in the past, it remains true that biblical scholars have much to gain from models derived from other languages and text-corpora from the ancient Near East, including Ugaritic. Given the importance that diachronic arguments have held in the history of biblical scholarship, in this essay we take up the topic of diachrony in Ugaritic, focusing primarily on the Ugaritic data on their own terms but with an eye toward their potential significance for understanding diachronic change in Biblical Hebrew. At the outset, a clarification needs to be made on the scope of this paper with respect to the term diachrony. Ugaritic is a Northwest Semitic language and shares important features with both the “Amorite” language continuum of the third to second millennia 1 and the later first-millennium dialects, including Hebrew, Phoenician, and Aramaic. 2 The precise classification of Ugaritic 1.  This category would include the texts from Mari, Emar, and various other sites from northern Syro-Mesopotamia that were written in Akkadian but in certain features (primarily proper names) show hints of the underlying local West Semitic stratum. 2.  One must also not neglect the Canaanite glosses in the Amarna letters, which predate the Ugaritic texts by about a century (see the discussion of chronology in §2.4). Arabic is more distant chronologically but not insignificant for comparative purposes; on the continuities between Ugaritic and Arabic, see Kaye (1991), who proposes that Ugaritic not be classified as Northwest Semitic but as a separate branch of Central Semitic. His reliance on phonological criteria, however, is problematic, because the phonological similarity of Ugaritic to Arabic might be due to the relatively archaic nature of Ugaritic as compared with the other Northwest Semitic languages (that is, the phonological similarities between Ugaritic and Arabic would be shared retentions and not innovations).

407

408

Joseph Lam and Dennis Pardee

within Northwest Semitic—especially vis-à-vis Canaanite—has been a topic of considerable debate (Voigt 1987; Huehnergard 1991; Tropper 1994; 2000; Pardee 1997). For our present purposes, it is enough to note that, while Ugaritic provides us with an unparalleled glimpse into the state of Northwest Semitic in the Late Bronze Age, there are difficulties with viewing it as a direct descendant of or antecedent to these other languages. On the one hand, despite the clear continuities between Amorite and Ugaritic culture (Goetze 1941; Villard 1986; Durand 1993; Bordreuil and Pardee 1993), the fact that virtually all of our knowledge of Amorite is derived from proper names culled from syllabic cuneiform texts precludes the possibility of any detailed diachronic comparison. On the other hand, even if one were to classify Ugaritic as an archaic member of the Canaanite subgroup (Tropper 1994; 2000), it remains too distinct to be counted merely as an earlier stage of any of these languages.  3 (Incidentally, perhaps many of the excesses of earlier generations of Ugaritic scholarship can be seen as the result of making too direct and unqualified a linguistic correlation between Ugaritic and Hebrew. 4) The precise relationship between Ugaritic and the other Northwest Semitic languages certainly merits careful consideration as part of the task of comparative Semitic reconstruction, but strictly speaking, it cannot be said to belong to the domain of diachronic analysis. For these reasons, the focus of this essay will be on inner-Ugaritic diachrony—that is, the extent to which diachronic developments can be discerned within the attested span of the Ugaritic language itself. In light of this sort of task, the question that immediately arises is the suitability of the corpus to diachronic analysis. We examine this question in the following section (§2), offering comparisons with the Biblical Hebrew corpus as a way of highlighting the distinctive features of the Ugaritic data. Then, after some brief reflections on the methodology behind the isolation of diachronic stages (§3), we present two case studies (§§4–5)—examples that represent potential avenues for diachronic analysis in Ugaritic. 2.  The Nature of the Ugaritic Corpus A number of inherent characteristics of the Ugaritic corpus distinguish it from Biblical Hebrew in terms of amenability to diachronic analysis. The first 3. J. Tropper, one of the most articulate proponents of the classification of Ugaritic as Canaanite, stresses the same point in unambiguous terms: “Sollte das Ug[aritische] mit den kanaanäische Sprachen im engeren Sinn zu einer gemeinsamen nordwestsemitischen Sprachgruppe zusammenzufassen sein, die man als ‘ugarito-kanaanäisch’, als ‘levantinisch’ oder als ‘kanaanäisch’ (im weiteren Sinn) bezeichnen könnte, wäre es gewiß als eigenständiger Zweig innerhalb des Kanaanäischen zu betrachten” (Tropper 2000: 5; italics added). 4.  Perhaps one of the most egregious examples is that of the prepositions, particularly the wholesale transference of translational values in Ugaritic to Hebrew contexts (see the summary of Pardee 1975: 330–37).

Diachrony in Ugaritic

409

two to be discussed can be viewed as beneficial to diachronic analysis, while the other two represent potential barriers to it. 2.1. Discreteness It may appear obvious, but the discrete nature of the Ugaritic corpus is an important factor in relation to diachronic analysis. 5 The Hebrew Bible comes to us as a single composite text, shaped through layers upon layers of redaction, and further mediated through centuries of scribal transmission. Thus, the identification and isolation of distinct text-stages (or sources) for diachronic analysis is itself hypothetical. By contrast, the Ugaritic data consist of distinct tablets (approx. 2,000 texts at last count), each of which (more or less) represents a single event of inscription in antiquity. In theory, then, it ought to be possible to arrange them in a chronological sequence. To put it another way, the “sources” have already been delineated. The question is whether the attested time span of the Ugaritic texts is broad enough to allow the discernment of diachronic changes within them (see §2.4). 6 2.2.  Geographical Provenance The Ugaritic texts hail almost entirely from two closely related sites, Ras Shamra and nearby Ras Ibn Hani, which would seem to preclude the presence of a great deal of dialectical variation. 7 Though dialectical difference has been invoked on occasion as an explanation of peculiarities in Ugaritic (for example, R. E. Friedman [1980: 194] on RS [Varia 14]), it has not been viewed as a major factor by most scholars. By contrast, since it is widely recognized that the sources and traditions behind the biblical text are varied with regard to geographical provenance (within ancient Israel), the dialectical component for Biblical Hebrew is not at all negligible. 8 For the Hebrew Bible, then, this adds complexity to the task of diachronic analysis, since it is not always clear 5.  By this general statement, we are not asserting that the mythological texts, for example, lack a textual prehistory but only that the actual exemplars we possess come to us from the hand of the ancient scribe, unmediated. 6.  Assuming that the diachronic span is sufficiently broad, a related question would be whether one can define bodies of “early” and “late” texts great enough in number to provide adequate data for comparison. Such is the theoretical shortcoming of the diachronic study of Tropper (1993), which was based on a single “late” text (RS 34.126 [KTU 1.161]); even if one were to accept his interpretation of the various features described, one could not be certain whether they represented truly “late” features or simply linguistic or scribal idiosyncrasies. 7.  For an assessment of Ras Ibn Hani and its function, see Lagarce and Lagarce 1995. 8.  The Shibboleth incident in Judg 12:5–6 is a mere anecdotal reflection of what must have been a much wider phenomenon in antiquity. Even in the meager corpus of Epigraphic Hebrew, one finds dialectical differences such as the reflex of the Proto-Semitic /*ay/ diphthong in northern (Samarian) Hebrew (yn for /yên/ [‘wine’]) versus southern (Judean) Hebrew (yyn for /yayn/).

410

Joseph Lam and Dennis Pardee

whether a given feature of linguistic variation is due to diachronic or to dialectical factors. But dialectical variation can be assumed to be less of a factor for Ugaritic than for the biblical texts. 9 2.3.  Genres Attested Another factor in the diachronic analysis of the Ugaritic texts is their distribution in terms of genre. In particular, the laconic, formulaic, and often repetitive nature of many texts makes them less amenable to detailed linguistic analysis. For instance, by far the most frequently attested genre (whether by number of texts or amount of preserved text) is administrative texts, which consist mostly of various types of lists and thus provide relatively little that is of use for detailed grammatical comparison (such as verbs). Similarly, the ritual texts, another significant prose genre, consist largely of lists of sacrifices in which deity names and animal types are often repeated, interspersed with laconic descriptions of ritual acts to be performed. This is not to say that these genres are devoid of value for linguistic analysis nor to deny the usefulness of the other genres, such as the prose of the letters and the legal texts (which are relatively sparse) or the narrative poetic texts. However, the limited data furnished by these genres as well as the marked stylistic distinctions between them represent a hindrance to diachronic analysis, certainly when compared with the more uniform religious content of the Biblical Hebrew corpus. 2.4.  Span of Attestation The three characteristics that we have noted so far relate only to the general amenability of the Ugaritic data set to diachronic analysis. However, what can be said more specifically concerning the chronological span of these texts? Traditionally, it has been asserted that the Ugaritic texts cover a chronological range of roughly two centuries, from about 1400 b.c.e. to the final destruction of Ugarit ca. 1185 b.c.e. 10 In recent decades, however, it has become increas9.  The evidence for dialectical variation in Ugaritic is sporadic and generally limited to small groups of definable texts, thus making the interpretation of the data difficult. Take the example of the texts inscribed in the “short” cuneiform alphabet, discovered at Ugarit and at other sites throughout Syria–Palestine (Dietrich and Loretz 1988: 145–275; Tropper 2000: 73–80). This alphabet shows clear similarities to the standard Ugaritic system but seems to have encompassed a smaller inventory of 22 signs. (No abecedary in the “short” alphabet has yet been discovered.) The nature of the texts and their geographical distribution suggest that the “short” alphabet may have represented the application of the cuneiform-alphabetic principle to the phonemic inventory of a Canaanite dialect (or various such dialects) contemporary to Ugaritic, but the data are too few to make a more certain identification (Tropper 2000: 78–79). Neither is there any clear evidence of a diachronic development between the “long” and “short” alphabets. 10.  A fourteenth-century date for the beginning of the Ugaritic texts is assumed in many of the standard grammars (Gordon 1965: 1; Segert 1984: 15; Sivan 1997: 1). The date of 1185 b.c.e. for the city’s destruction is also approximate, but no scholar to our knowledge

Diachrony in Ugaritic

411

ingly clear that most of the extant textual material was inscribed in the final decades of occupation of the Late Bronze Age city. This situation can be seen as a function of the nature of ancient archives more generally. Tablets were kept only for as long as they were useful. For administrative and economic texts, this might have been just a short time; on the other hand, legal texts in certain instances might have been kept for generations. But on the whole, only isolated texts from earlier periods were deemed worthy of preservation; tablets of a more ephemeral nature would have been discarded or recycled.  11 Not only are most of the Ugaritic texts to be attributed to the last stage of occupation at Ugarit, but the few texts that have traditionally been associated with an earlier period have had their dating challenged more recently on independent grounds. First, the extant tablets of the major mythological cycles (Baal, Aqhat, and Kirta) have conventionally been dated to the mid-fourteenth century b.c.e. based on colophons that mention a scribe named Ilimilku; 12 in the most fully preserved of these colophons, he is reckoned to the service of a King Niqmaddu, 13 long associated with Niqmaddu II, 14 who reigned in the presumes a date outside of the range 1200–1175 b.c.e. For more on the arguments related to this dating, see Singer 1999: 729–30; and Freu 1998. 11.  See van Soldt 1991: 47–48 and the references cited therein for further discussion of these points. 12.  The colophons in question are three: RS 2.[008]+ (KTU 1.4) VIII left edge (where Ilimilku’s name is restored) and RS 2.[009]+ (KTU 1.6) VI 54–58, both from the Baal cycle; and RS 3.325+ (KTU 1.16) VI: left edge, from the Kirta epic. However, based on close similarities in script, all of the extant tablets of the Baal, Aqhat, and Kirta cycles are conventionally attributed to Ilimilku (see the epigraphic observations in Herdner 1963: 1, 58, 78, et passim). 13.  However, others have suggested that the title ṯʿy nqmd mlk u͗grt ‘the ṯāʿiyu-priest of Niqmaddu, king of Ugarit’ (RS 2.{009}+ [KTU 1.6] VI 57) might have applied not to Ilimilku himself but to his teacher, Attenu (van Soldt 1988; Smith and Pitard 2009: 728), a fact that would have implications for dating. 14.  In this essay, references to the Ugaritic kings follow the traditional system, in which numbers are assigned after each royal name in the chronological order that they appear in the extant historical sources: thus Niqmaddu “I” corresponds to the king in the patronym of Yaqaru (once thought to be the founder of the dynasty), as attested on his dynastic cylinder seal (Nougayrol 1955: xli–xliii), while Ammistamru “I” is the fourteenth-century king who is known as the sender of Amarna Letter EA 45. Evidence from the Ugaritic “king list” RS 24.257 (KTU 1.113; Pardee 1988: 165–78) has generally been excluded, despite the many homonymous royal names found there, due to the fragmentary nature of the text and the uncertainties involved in the chronological placement of those kings (Singer 1999: 611 n. 22). More recently, on the basis of additional (duplicate) Akkadian king lists discovered in 1988 and 1994, D. Arnaud has proposed a wholesale renumbering of the kings, taking into account the newly attested royal names (see Arnaud 1999: esp. p. 163). However, because these newer king lists also do not appear to cover the entire dynasty (Pardee 2002b: 195–204, esp. pp. 196–99), we retain the older system, though readers should be aware that Arnaud’s

412

Joseph Lam and Dennis Pardee

fourteenth century (Albright 1951: 31). 15 However, in 1992, a fragment of a text with mythological themes (RS 92.2016) was discovered at Ras Shamra, the content of which is obscure, but which also contains a colophon that must have belonged to the same Ilimilku (Caquot and Dalix 2001). 16 The archive in which the tablet was discovered is associated with a high official named Urtenu and contains texts that mention the last kings of Ugarit from the late thirteenth and early twelfth century. This raises the possibility that the king “Niqmaddu” whom the famous Ilimilku served was not Niqmaddu II from the fourteenth century but, rather, Niqmaddu III from the very end of the thirteenth century, the second-to-last king of Ugarit. 17 If so, all of the tablets that Ilimilku inscribed would have to be dated to that time as well. To maintain the view that Ilimilku operated in the fourteenth century, one would need to assert that the text found in Urtenu’s archive was indeed a copy preserved from a much earlier time—not an impossible scenario but not the most likely situation either.  18 Another text that has traditionally been an anchor for a fourteenth-century date for the Ugaritic cuneiform alphabet is RS 11.772+ (KTU 3.1), a treaty between Niqmaddu and a Hittite king named Šuppiluliuma (  ṯpllm). This text, discovered in 1938, is almost universally viewed by scholars in connection with Šuppiluliuma I of the second half of the fourteenth century, in light of similar Akkadian treaties from Ras Shamra. 19 However, as before, the repetition of royal names in antiquity has created chronological ambiguity: recently, A.-S. Dalix has argued that the royal figures in question were not Niqmaddu II and Šuppiluliuma I but Niqmaddu III and Šuppiluliuma II (Dalix 1997; 1998). 20 numbering has been adopted in some other recent studies (thus our “Niqmaddu II” = “Niqmaddu III” in Arnaud’s system, etc.). 15.  This date was viewed to be consistent with the earliest Akkadian texts from Ras Shamra, some of which also make reference to this fourteenth-century Niqmaddu. There is an Akkadian scribe from Ras Shamra whose name is written dingir-lugal, who is sometimes associated with the Ilimilku of the Ugaritic mythological texts (van Soldt 1991: 27– 29), though this is not at all certain. 16.  The name Ilimilku is reconstructed, but the preservation of the titles {[. . . š]˹b˺ny . lmd . a͗ tn . prln} ‘the Šubbanite, disciple of Attenu the divining-priest’ (RS 92.2016: 40′), which are the same as the titles found in the colophon of the sixth tablet of the Baal cycle, makes the identification all but certain. 17.  For lists of the kings of Ugarit with potential synchronisms with other kings in the ancient Near East, see the foldout in Singer 1999, which is part of a detailed essay on the political history of Ugarit, and the chart in Freu 1998: 37, which covers only the thirteenth century onward (both of these works predate the publication of Arnaud 1999 and thus adhere to the “older” system of numbering the Ugaritic kings; see n. 14 above). 18.  Note, however, that the discovery of a legal text in the Urtenu archive dating explicitly to the reign of Ammistamru II (RS 94.2168) shows that there was at least some diachronic variation in the texts found there, spanning perhaps the latter half of the thirteenth century. 19.  For a recent study of this text with bibliography, see Pardee 2001a. 20.  In fact, Dalix (1997) goes so far as to question the fourteenth-century provenance of the corresponding Akkadian material, though this part of her thesis has been convincingly

Diachrony in Ugaritic

413

Perhaps the strongest argument in favor of this view is archaeological, for the “Western Archive” of the palace in which this text was found contains material almost exclusively from the late thirteenth and early twelfth century (van Soldt 1991: 57–58; Dalix 1998). A third Ugaritic text for which the fourteenth-century provenance has been challenged is RS 18.113 A (KTU 2.42). This is a letter that was thought to refer to Amenophis III of Egypt based on the word nmry that appears in line 9, which is interpreted as his cognomen Nimmuria. 21 In this case, however, A. F. Rainey (1974: 188) has suggested an alternate interpretation for the word nmry as a common noun meaning ‘splendor’, eliminating the fourteenth-century historical reference altogether. Further support for the later dating lies in the fact that all of the other texts coming from the same royal palace archive date to the later period (Singer 1999: 677–78). If one accepts the validity of these three instances of redating of texts, the earliest (externally) datable Ugaritic texts would then hail from the early- to mid-thirteenth century (see §4 below), making the span of attestation of the texts only about one century at the most. Moreover, as we have noted, most of the textual evidence is chronologically clustered around the final decades of the Late Bronze Age. The upshot, then, is that Ugaritic would justifiably be called a one-period language. 22 While a century is not necessarily too short for observable diachronic changes to occur, neither would one expect a priori any drastic diachronic developments over such a relatively short period. 3.  Synchronic versus Diachronic Variation: Methodological Considerations It is a truism that all languages are constantly in flux. Thus, the existence of variant features is characteristic of any sociolinguistic system—a phenomenon that has been called “orderly heterogeneity” (Weinreich, Labov, and Herzog 1968). Whether it is due to different (contemporary) language users’ producing refuted by Freu (2004). One of the principal results of the most recent study of RS 11.772+ (Pardee 2001a) is to demonstrate that the differences between this text and the Akkadian versions are significant enough that one cannot view the Ugaritic text as a direct translation of any Akkadian exemplar that we currently possess. 21. The cognomen Nimmuria appears frequently throughout the Amarna correspondence from the fourteenth century; the identification was already proposed in the editio princeps of RS 18.113 A (Virolleaud 1965: 15). 22.  It is worth noting that earlier scholars have spoken of the “late Ugaritic” of the prose documents (for example, Albright 1958) as opposed to the language of the narrative poetic texts. In light of the archaeological and prosopographical indications just discussed, however, it would be more methodologically sound to define the mythological texts as “archaic” (or “archaizing”) in contrast to the remaining texts. To put it another way, most of the Ugaritic for which we have evidence is “late” Ugaritic, of which two principal registers are preserved (see §5). Compare Tropper (1993), who supposes a more gradual development of the language over the period of the attested texts.

414

Joseph Lam and Dennis Pardee

different diachronic forms or due to a single language user’s producing inconsistent forms, evidence of diachronic change “in process” can be discerned even in a single stage of a language. In Ugaritic, then, one finds features of language variation that can be interpreted as evidence of diachronic change in process. For example, at the level of orthography, one can point to the phenomenon of “historical writing”—that is, cases where the spelling of a word varies between phonemic and phonetic representations. W. R. Garr (1986) has invoked this sort of explanation for the sporadic alternation of voiced and unvoiced pairs of consonants, which in his view was based on regular phonological processes but was not represented consistently in the orthography. In such cases, where a phonemic representation exists alongside forms that belie further phonological changes, it is reasonable to assume that the latter represents the actual pronunciation of the form more closely, while the former gives evidence of the inherent conservatism of writing systems. 23 These are clearly instances of historical (phonological) change in process. Given that the cuneiform alphabetic script almost never indicates vowels, however, not all such orthographic variations permit unambiguous interpretations. Take the relative-determinative element 24 d/dt, of which the distribution is somewhat puzzling, since both spellings are attested in the fs as well as in the mp and fp. This was given a two-tiered explanation by Tropper (2000: 234–38) in which declinable (d for ms; dt for the fs and pl.) 25 and indeclinable (d for all numbers and cases) variants coexisted. Alternatively, however, it is also possible to reconstruct a simpler situation, in which the -t component was interpreted (via reanalysis) as an optional enclitic /-ti/ in the fs and the pl. forms, giving rise to free variants spelled d and dt in those cases (Pardee 2003–4: 137–39). 26 In either explanation, the distribution of forms would represent an intermediate stage in the transition from a fully inflected relative-determinative marker (reconstructable along the lines of Arabic and Old Akkadian) to an uninflected relative complementizer (Holmstedt 2008: 26). Another example can be cited, this time relating to paleography, in which the distribution of the variant is confined to a specific group of texts but the typological relationship between the two linguistic forms is uncertain. The fea23.  It is also possible, of course, that the alternate spellings represent actual phonetic differences that were recorded by the scribes; they would then signal individual instances in which a given phonetic change had or had not occurred. 24.  Here we avoid the terminology “relative pronoun” in light of the criticisms of Holmstedt (2008). 25. In light of comparative Semitic considerations, Tropper (2000: 235) reconstructs /dū/ (nom.), /dī/ (gen.), /dā/ (acc.) for the ms; /dātu/ (nom.), /dāti/ (gen.), /dāta/ (acc.) for the fs; and /dūtu/ (nom.), /dūti/ (obl.) for the pl. One should note the syllabic spelling {du-ú} in RS 20.123+ II 29′, which represents the ms nominative form. 26.  One possible set of vocalizations would be /dā(ti)/ for the fs and /dū(ti)/ for the pl. (Pardee 2003–4: 138).

Diachrony in Ugaritic

415

ture in question is an alternate form of the sign {ẓ} discussed by Freilich and Pardee (1984). In examining a number of texts in which others have posited the replacement of {ẓ} by {ṭ}, it was ascertained that most of these actually represented an alternate sign-form for {ẓ} earlier identified by Dietrich, Loretz, and Sanmartín (1975), one that shows some resemblance to the standard {ṭ} but is clearly distinguishable from it. 27 In this case, all of the texts in which this variant {ẓ} occur come from the “House of the Priest with Lung and Liver Models” (or “House of the Hurrian Priest”) located in the “Sud-Acropole” and excavated in the 24th campaign of 1961. 28 Though the dating of the archive is inconclusive, all indications are consistent with the final period of Late Bronze occupation (van Soldt 1991: 201). It appears, then, that what we have is synchronic paleographic variation in the sign {ẓ} in the late period, perhaps idiosyncratic to a small group of individuals who inscribed these particular texts.  29 Furthermore, at present there is no firm basis for determining which form of {ẓ} was typologically earlier, nor even whether one developed directly from the other. From this brief survey of examples, it seems that linguistic variation in itself is not sufficient to identify a diachronic development. For this purpose, at least two additional criteria are necessary. First, for any given feature that shows variation, one must have some basis for asserting that one form is earlier than the other (rather than just a stylistic difference), whether it is because of typological or comparative considerations. Second, for the texts that contain the variant feature, one must ideally have external reasons (whether archaeological, prosopographical, or the like) for separating them chronologically from texts that lack the feature. If there is convergence in the delineation of texts by diachronic and external criteria, then it appears justified to identify a development as diachronic, and possibly to use the feature for the dating of other texts. 30 Methodologically speaking, it is difficult to avoid circularity in such 27.  For a clear comparison of the standard {ṭ} and the alternate form of {ẓ}, see Freilich and Pardee 1984: 28. The writing of {ẓ} for etymological {ṭ} has only been identified in three texts: RS 5.194 (KTU 1.24), which exhibits another “archaizing” feature in which the relative-determinative element is spelled {ḏ} (see Blau 1968); RS 5.259 (KTU 1.25); and RIH 78/14 (CAT 1.163), a badly damaged text containing lunar omens (see Pardee 2000: 859, 866, 870–71). 28.  The texts containing this variant form of {ẓ} are RS 24.244 (KTU 1.100), RS 24.245 (KTU 1.101), RS 24.251+ (KTU 1.107), RS 24.258 (KTU 1.114), RS 24.260 (KTU 1.115), RS 24.263 (KTU 1.117), RS 24.271 (KTU 1.123), RS 24.293 (KTU 1.133), RS 24.649 A (KTU 1.152), RS 24.655 A (KTU 7.184), plus the two abecedaries RS 24.281 (KTU 5.20) and RS 24.288 (KTU 5.21). 29.  Since the abecedary RS 24.281 (KTU 5.20) consists of two inscriptions of the alphabetic inventory (by teacher and pupil, respectively), one is obliged to assert that this special form of {ẓ} was taught in some scribal circles. 30.  We speak of “convergence” here with awareness of R. E. Friedman’s (2003: 27–31) use of the same term in his articulation of the Documentary Hypothesis; he points out that the appeal of the theory is not only that it resolves perceived tensions or contradictions (such

416

Joseph Lam and Dennis Pardee

a procedure, for the identification of distinguishing linguistic features always happens parallel with considering external dating criteria. But the best explanation is accounting for all the data in the simplest and most economical way. With these ideas in mind, we now present two potential examples of these converging lines of data in Ugaritic. The first involves the combination of paleographical criteria with archaeological and historical criteria; the second invokes a wide variety of grammatical and stylistic features in conjunction with separating the texts according to genre. Together these data illustrate both the potential benefits of diachronic analysis in refining our understanding of the Ugaritic texts and the challenges involved in this approach. 4.  The Texts of Ammistamru II We have already discussed the evidence against a fourteenth-century date for the invention of the Ugaritic cuneiform alphabet. Though this picture is subject to revision in response to new evidence, at present it appears to be the most sensible hypothesis. However, if one rules out the fourteenth century for the invention of the cuneiform alphabet, which texts are the earliest texts in Ugaritic? Given the lack of firm chronological data from Ugaritic prosopography in general, one is reliant on the preservation of the names of the kings of Ugarit for possible historical synchronisms. It turns out that the earliest datable Ugaritic texts all have a plausible connection to Ammistamru II, who reigned in the mid-thirteenth century (ca. 1260–1235/30). 31 A number of texts make direct historical mention of his name: these include three sources containing impressions from his royal seal (the Akkadian legal text RS 16.270 [KTU 6.23; see fig. 1], the bulla RIH 83/21 [CAT 6.75], and the letter RIH 78/12 [CAT 2.82]) 32 as well as three legal texts in which he was the granter of property or associated rights (RS 15.111 [KTU 3.2], RS 16.382 [KTU 3.5; see fig. 2] and RS 94.2168). In addition, two other texts can be dated to his reign based on plausible restorations of his name: RIH 78/03+ (KTU 2.81) is a partially preserved draft of a letter to the king of Egypt in which Ammistamru’s name is almost certainly to be restored, 33 while RS 15.117 (KTU 7.63) could contain a list of titles held by Ammistamru as crown prince during the latter part of his father as different divine names) but that the delineation of sources by different “lines of evidence” (doublets/triplets, divine names, varying terminology) leads essentially to the same results. Whether one agrees with Friedman’s source identifications or not, one must grant the validity of his methodological point—that it is the convergence of multiple, independent criteria that justifies defining two distinct sets of linguistic data. 31.  Two alternative dates are given by Singer (1999: foldout) and Freu (1998: 37), respectively. 32.  For the impressions on RS 16.270 and RIH 83/21, see Pardee 2007; for RIH 78/12, see Bounni, Lagarce, and Lagarce 2008. 33.  The restoration would be {[ʿmṯtm]r . ʿbdk} (RIH 78/03+: 5); text published in Bordreuil and Caquot 1980: 356–58.

Diachrony in Ugaritic

417

Niqmepa’s reign (Pardee 2010). Together these texts would be datable to the early to mid-thirteenth century b.c.e.—a period that in turn would represent the time of the invention of the Ugaritic cuneiform alphabet to the best of our current knowledge (Pardee 2007). 34 4.1.  Distinguishing Features of the Seal Impressions Although the seal impressions are the shortest texts in this group, within the four-word inscription mἰšmn ʿmyḏtmr mlk u͗grt ‘Seal of ʿAmmîyiḏtamru, king of Ugarit’, one finds a cluster of three archaic-looking features. Most obvious is the name of the king himself, which is normally spelled ʿmṯtmr (as in RS 15.111 and RS 16.382); the form ʿmyḏtmr (perhaps to be vocalized /ʿammîyiḏtamru/ [case form] or /ʿammîyiḏtamir/ [absolute]) would represent the original shape of the name before the contraction of the triphthong /îyi/ and the devoicing of /ḏ/ to /ṯ/ via contact with /t/. 35 A second potentially archaic feature, though less obvious than the first, concerns the word for ‘seal’, mἰšmn. In Ugaritic, one finds two additional spellings of this word: ma͗šmn and mšmn. The latter is clearly the most typologically developed of the three, because it shows quiescence (and lack of orthographic representation) of ʾalep (Pardee 2003–4: 26–27). 36 A fairly straightforward interpretation of the other two spellings would consider mἰšmn the earliest of the three forms (/maʾšamānu/ < *maqtal + -ān), while ma͗šmn would reflect the insertion of an epenthetic vowel after the ʾalep (as an intermediate step toward ʾalep-quiescence), that is, /maʾašamānu/ or /maʾašmānu/ (with subsequent elision of the following short vowel). 37 Unfortunately, the textual distribution of the three forms cannot be arranged in an unambiguous chronological sequence; in particular, m͗ıšmn is attested also in RS 15.125 (KTU 2.19), which explicitly mentions Niqmaddu III from the late thirteenth century. 38 4.2.  The Two-Wedged {g} While the two features just discussed are limited to the seal impressions, a third distinguishing feature is found in most of the Ammistamru II texts 34.  In this list, we exclude RS 24.257 (KTU 1.113), RS 34.126 (KTU 1.161), and the Hurrian text RS 24.274 (KTU 1.125), all of which refer to the name “Ammistamru” as a deified king. 35.  The meaning of the name would be ‘may my divine paternal grandfather (ʿammî) protect (yiḏtamir)’, with the verb in the name coming from the root ḎMR ‘to guard, protect’ (Bordreuil and Pardee 1984: 13). 36.  This is not to assert that quiescence of ʾalep was a regular feature of Ugaritic. 37.  In light of the existence of three spellings, it is likely that ma͗šmn and mšmn reflect actual phonetic realizations of the noun, while m͗ıšmn could either be a phonetic spelling or a historical spelling. 38. The typologically “late” form mšmn is attested in RS 17.025 (KTU 6.66): 1 and RS 7.088 (KTU 6.69): 1–2, while the “intermediate” form ma͗šmn is found in RS 14.023 (KTU 6.17): 1, and only possibly in RS 17.442 (KTU 4.318): 1 (partially restored).

418

Joseph Lam and Dennis Pardee

listed and thus holds greater potential for identification as a potential archaism, though the usual complications and ambiguities are present here as well. The feature in question is a special form of the sign {g} that is inscribed with two wedges (one vertical and one horizontal) instead of the usual lone vertical wedge (see figs. 1–2). The interpretation of this paleographic feature as an archaism is bound up with the invention of the Ugaritic cuneiform alphabet; the fact that most of the abecedaries from Ugarit retain an order of signs essentially identical to the later Phoenician/Hebrew order (apart from later phonetic mergers) suggests that the cuneiform alphabet was inspired by a contemporary proto-Canaanite (linear) prototype (Pardee 2007: 189). 39 And since the shape of the gimel in the first-millennium Phoenician script consists of two strokes connecting at right angles, it is tempting to view the two-wedged cuneiform alphabetic {g} as an imitation of what would have been the proto-Canaanite gimel. Of the Ammistamru II texts listed above, the seal impressions (all made by a single seal) and the legal texts RS 15.111 and RS 16.382 attest clear tokens of two-wedged {g}; RS 94.2168 contains what looks to be an imitation of the two-wedged form (but inscribed with a single motion); as for RIH 78/03+ and RS 15.117 (possibly from the reign of Ammistamru’s father, Niqmepa), these contain {g}’s with broad top edges and somewhat rounded right edges but are certainly not formed with two wedges. 40 RS 15.117 may represent the oldest text of the lot, but it lacks the feature attested in the time of Ammistamru II; this is not in itself a problem, because the {g} in other texts from the time appears in the single-wedged form. Apart from the texts that mention Ammistamru explicitly, the two-wedged {g} is attested in a number of other texts from Ras Ibn Hani—a fact that strengthens the potential usefulness of this feature as a criterion for dating because all of these texts can be linked archaeologically to the time of Ammistamru II.  41 Some of the texts containing two-wedged forms of {g} (at varying levels of preservation) are the administrative texts RIH 77/18 (CAT 1.175) and RIH 78/19 (CAT 4.775), the incantation RIH 78/20 (CAT 1.169), the mythological fragment RIH 78/26 (CAT 1.176), and the smaller fragments RIH 78/22 (CAT 7.218) and RIH 87/02 (CAT 9.529). 42 Apart from this tenuous paleographical 39.  Moreover, additional letters ({͗ı}, {u͗ }, {s̀ }) which did not exist in the linear alphabet were appended to the end of the alphabet. 40.  For details, see Pardee 2012. 41. The latest treatments attribute the establishment of Ras Ibn Hani, especially the “North Palace” (where the texts were found), to Ammistamru’s reign (Lagarce and Lagarce 1995; 2008). 42. All of these are mentioned in Pardee 2012. One additional text contains a twowedged {g}, the halaḥam abecedary RS 88.2215 (Bordreuil and Pardee 2001a), which was discovered in a clearly “late” archaeological context. However, given the uniqueness of this

Diachrony in Ugaritic

419

Figure 1.  Photograph and copy of the seal impressions of Ammistamru on Akkadian legal text RS 16.270. The seal reads {m͗ıšmn . ʿmyḏtmr . mlk u͗ grt}. The {g} is visible in the lower left corner of the impression as it reads on the tablet, in the negative (above the scale); in the lower right corner in the flipped version, as it would have read on the seal itself (below the scale). Photograph is reproduced courtesy of the Projet PhoTEO, Mission de Ras Shamra. Drawing by Dennis Pardee.

420

Joseph Lam and Dennis Pardee

Figure 2.  Copy of the Ugaritic legal text RS 16.382 obverse. A {g} is visible in the word {u͗ grt}, line 4; another in the word {gth}, line 7. Drawing by Dennis Pardee.

Diachrony in Ugaritic

421

indicator, one is hard-pressed to find any other linguistic features that unify these texts in distinction from “later” texts. 43 4.3. Assessment What, then, are we to make of the three typologically archaic features mentioned? In light of inconsistencies in the distribution of these features, one might wish to classify them as attempts at archaizing rather than true archaisms. This sort of description appears particularly applicable in the case of the seal impressions, for one could imagine the ideological function of placing archaic linguistic features on a king’s personal seal. As for the two-wedged {g}, the “archaizing” explanation is one way of accounting for the apparent lack of this feature in RS 15.117, despite its potential earlier dating. Perhaps more detailed studies in the future will yield new results, but at present, it is difficult to identify any marked diachronic shift in the Ugaritic language between the time of Ammistamru II and the final destruction of the city. 4.4.  The Texts of the Scribe Thabilu A final point is worth making regarding the texts from the period of Ammistamru II, one that does not involve the separation of diachronic stages by means of internal criteria alone but that nonetheless is germane to the relative dating of texts. This is a comparison of the hand of the well-known scribe Ilimilku and that of another scribe, named Thabilu, at least part of whose career may be dated to the time of Ammistamru II. Tablets inscribed by the latter have been discovered at Ras Ibn Hani and are thus dated by the archaeologists to the time of this king (see n. 42); but two tablets signed by him and several others showing the same hand have been found at Ras Shamra (references in text both conceptually and paleographically, it is difficult to date with certainty and thus would not necessarily overturn the picture described thus far. 43.  To this point, we have not mentioned the paleographic feature of extra wedges. This is the phenomenon in which certain signs in the cuneiform alphabet could have been inscribed with a greater number of wedges than the standard form but only when the addition of wedges did not create ambiguity (that is, the addition was made to the sign already having the most wedges in a given series, such as {n} in the series {t}, {a͗ }, {n}). Tropper (2000: 26) suggested that the phenomenon of extra wedges might have been a relatively archaic feature, since it occurs in RS 16.382 (as well as other texts from Ras Ibn Hani, including RIH 78/26). However, the lack of this feature in the seal impressions of Ammistamru II (which contain the “archaic” {g}) seems to militate against the use of extra wedges as a firm criterion for dating (Pardee 2002a: 60–61). The presence of forms with extra wedges in the texts inscribed by Ilimilku is, to the extent that the later dating of this scribe is accepted, an indicator that the phenomenon was optional at all periods. The relative frequency of signs with extra wedges in the texts from Ras Ibn Hani may indicate, nevertheless, that the phenomenon reflects a period of experimentation at a time when the writing system was relatively new—as opposed to the hypothesis that these signs were characteristic of the system as invented. Of interest then is the observation that Thabilu (see §4.4), whose works have been discovered at both Ras Shamra and Ras Ibn Hani, does not use such forms.

422

Joseph Lam and Dennis Pardee

Pardee 2012). Thabilu’s graphic inventory did not, however, include the twowedged {g} and this criterion is thus not applicable in dating his works. It is impossible to say what length of time may have separated the two scribes. It may have been as much as half a century, if each one’s work was narrowly circumscribed in time, or their careers may even have overlapped, if each worked for a quarter century or so and Thabilu began his career late rather than early in the reign of Ammistamru. The two hands are very different, but nothing narrowly palaeographical contributes to the earlier dating of Thabilu, which must be established archaeologically. The same is true of the content of their works: both were interested in mythology, and they say different things about the gods, but it would be difficult to assign any chronological importance to the differences. The one feature that does stand out is the fact that Thabilu set down texts not only in Ugaritic but probably also in Hurrian and in Akkadian, both of the latter groups also in the alphabetic script. In particular, the attempt to write Akkadian in alphabetic script, virtually without parallel at Ugarit, may indicate a time of experimentation, perhaps shortly after the invention of the alphabetic writing system (an idea of Robert Hawley’s that he will develop in future publications). 5.  The Distinction between Poetry and Prose No statement on diachrony in Ugaritic would be complete without a discussion of the differences between poetry and prose. In a way reminiscent of the arguments for the antiquity of certain poetic texts in the Hebrew Bible, scholars of Ugaritic have long assumed a chronological distinction between the poetic dialect (particularly the language of the mythological texts) and the prose texts, viewing the poetry as relatively archaic (Albright 1958; Liverani 1964; Parker 1970; Mallon 1982; Segert 1984: 15). Furthermore, it was often hypothesized that, among the major mythological cycles, the Baal story was the oldest, followed by Aqhat and Kirta (Albright 1958: 36). As mentioned above, one of the reasons for asserting the relative antiquity of the narrative poetry had to do with the dating of the scribe Ilimilku to the fourteenth century (Liverani 1964: 173; Gibson 1977: 1). This reason is now called into question by the discovery of the Ilimilku colophon on RS 92.2016 in the “House of Urtenu” (§2.4). However, it must be emphasized that the presumed chronological priority of the poetic texts was never predicated entirely on the dates of the tablets. Rather, the more fundamental consideration involved the theory behind the composition of these texts: namely, that they represented the committing to writing of much older oral traditions. 44 Thus, whatever the 44.  This general point is not really disputable; for example, the theme of the battle between the Storm-god and the Sea, one of the most conspicuous episodes in the Ugaritic Baal Cycle, is attested at Mari in the eighteenth century b.c.e. (Durand 1993; Bordreuil and Pardee 1993), which furnishes tangible proof of the preexistence of this tradition.

Diachrony in Ugaritic

423

date of the tablets themselves, their content could be viewed as more archaic. When one adds to the picture some of the obvious affinities between Ugaritic and Biblical Hebrew poetry (especially the modes of parallelism), as well as the apparent cultural and ideological significance of the mythological texts (thus justifying their preservation), the result is a compelling argument for a “Canaanite” (broadly defined) 45 or “Levantine” poetic tradition that predated the attested forms of both Ugaritic and Hebrew as we have them. In the context of diachrony, however, the issue is not merely whether there exist archaic elements or themes in the poetic texts. Instead, the question is more focused: do the poetic texts preserve an older form of the language than the prose texts? This question is more complicated, because it speaks to the distinction between language (form) and content. Assuming that the narrative poetic texts do represent the recording of older traditions, it is nonetheless possible that they underwent linguistic updating. After all, if they were inscribed (by Ilimilku) and read during the “late” period, the language could not have been so archaic as to have eluded comprehension. Keep in mind that linguistic updating in this case would not necessarily be inconsistent with a distinct poetic dialect, for the distinction may have been entirely stylistic and generic as opposed to diachronic. In fact, one might wish to quarrel with the very framing of the distinction as a distinction between “poetry” and “prose.” Because the mythological texts were among the first texts discovered at Ras Shamra, while texts from later campaigns consisted predominantly of prose texts of various types, the opposition between these two categories came to be definitive in discussions of diachrony. However, still-more-recent excavations have unearthed examples of what can be deemed “late poetry.” Two examples that come immediately to mind are the royal funerary ritual RS 34.126 (KTU 1.161), datable to around the time of the ascension of Ammurapi (the last king of Ugarit) to the throne, and RS 92.2014 (Bordreuil and Pardee 2001b), an incantatory text that was not only discovered in the “House of Urtenu” but mentions Urtenu’s name (and thus was probably written for him). While the content of these “late” poetic texts is clearly different from that of the mythological texts, the fact that the former are also structured (roughly) by poetic parallelism makes them potentially conducive to diachronic comparison with the latter, because they allow one to avoid the “interference” of stylistic factors. Unfortunately, the category of “late poetic texts” is probably not yet sufficiently large to permit meaningful distinctions based on this category alone. In light of these potential challenges, the only way forward is nonetheless to examine the potentially archaic features that have been proposed and evaluate them for their plausibility. As an illustration of the principles involved, we will 45.  On the use of the term Canaanite in this broader sense, see Day (1994) and Pardee (2001b).

424

Joseph Lam and Dennis Pardee

now undertake a brief assessment of some of the most salient distinguishing features of narrative poetry vs. prose. The following discussion draws from the detailed summary of Smith (1994: 29–58), which was written in relation to the Baal cycle specifically but represents a thorough overview of proposed archaic features in Ugaritic poetry (relating to style, content, and language). Readers interested in more detail are advised to consult Smith’s volume for further discussion and bibliography. 5.1.  The Verbal System in the Narrative Poetic Texts The most oft-cited distinction between the poetic and prose dialects of Ugaritic relates to the verbal system (Mallon 1982). The relevant features are both semantic and morphological. First of all, in poetry, the prefix-conjugation (YQTL) form of the verb seems to be used for actions occurring in past, present, and future time, in marked contrast to prose, where not a single YQTL form with past-time reference can be certainly identified. Thus, it has been thought that the verbal system in narrative poetry represents an older situation in which multiple prefix-conjugation forms with different “mood” vowels (*yaqtul, *yaqtulu, *yaqtula, plus the energics) had different semantic functions—a situation resembling Akkadian and what can be reconstructed of Amarna Canaanite. In particular, *yaqtul (with ∅-ending) in poetry would represent a perfective verbal form with both preterite (past-time) and jussive (volitive) meanings, in contrast to prose usage, where the preterite function was taken over by the suffix-conjugation *qatala. 46 Recently, however, Greenstein (2006) conducted a study of all YQTL forms in the Baal, Aqhat, and Kirta cycles, challenging the assumption that the function of *yaqtul as preterite was part of narrative poetic usage. Two of his arguments are particularly compelling. First, in III-y roots, Greenstein notes the lack of any discernible difference in meaning between YQTL forms with and without the third root -y. 47 Second, with III-ʾ roots, he finds that every single attested nonjussive singular YQTL form is spelled with final {u͗ }, indicating that it was *yaqtulu. (If *yaqtul-preterite were truly a productive form in Ugaritic narrative verse, one would have expected at least a few vestiges of it in III-ʾ roots.) For these and other reasons, Greenstein argues that the *yaqtul-preterite was not part of regular usage in narrative poetry. The distinctions between apocopated and nonapocopated III-y forms 48 and between forms with and without 46.  Elsewhere in the central Semitic languages, the preterite function of *yaqtul can be seen only in vestiges such as the Arabic negation structure lam + yaqtul and the Hebrew narrative form wayyiqṭōl. 47.  In theory, the absence or presence of the third root -y could be seen as an orthographic indicator of *yaqtul vs. *yaqtulu, as with ybk for /yabki/ ‘he wept’ (perfectivepreterite) versus ybky for /yabkiyu/ ‘he weeps’ (imperfective). 48.  In keeping with his thesis, Greenstein suggests that the apocopated III-y forms may in fact be original *yaqtulu forms with a final contracted triphthong, as with ybk for /yabkû/ < /*yabkiyu/ (2006: 84–85).

Diachrony in Ugaritic

425

-n are purely stylistic in his view. On balance, one must note that, since Greenstein restricted his analysis to the Baal, Aqhat, and Kirta texts, there may exist genuine retentions of *yaqtul-preterite in other poetic texts. 49 Moreover, this is not to deny any hint of diachronic development in the YQTL forms; Greenstein himself grants the possibility that the concentration of forms with -n may have had a diachronic significance (confirming the traditional chronological ordering of Baal, then Aqhat, then Kirta [Greenstein 2006: 89]). However, his overall argument is compelling and would point to the notion of stylistic distinction as the determinative distinction, at least in relation to the function of YQTL verbal forms. Despite the doubts surrounding the existence of a *yaqtul-preterite, however, one does find variants of certain morphological features of the verb for which the distribution falls out along the lines of poetry versus prose. One is the 3ms suffix-conjugation (*qatala) form of III-w/y roots, which sometimes shows retention of the third root consonant (ʿly /ʿalaya/ ‘he went up’) and sometimes not (ʿl /ʿalâ/ or /ʿala/). While the contracted forms occur in both poetry and prose, the uncontracted (and typologically earlier) forms are so far found only in poetry (Pardee 2003–4: 323; cf. Tropper 2000: 655, 664–65). 50 Also relating to the prefix-conjugation verb is another apparent archaism in the poetic texts: in the Baal cycle in particular, the prefix letter of the 3mp YQTL forms is attested as both y- and t-; the former is clearly the original form, based on comparative considerations. Elsewhere, however, only the tprefix is certainly attested (Tropper 2000: 432–38; Pardee 2003–4: 2). 5.2.  Other Distinguishing Features Besides the verbal system, there are other features that differ in poetry and prose. While they are somewhat isolated and sporadic due to the paucity of the evidence, nonetheless they can be seen as contributing cumulatively to a view of the relative antiquity of the poetic texts. The Ugaritic noun for ‘brother’ varies between spellings that reflect the historical form (namely a͗ḫy [‘my brother’] for all three nominal cases) and spellings that show vowel harmony between the initial /a/ and the long contracted case vowel (for example, u͗ḫy / ʾuḫûya/ [‘my brother’] in the nominative; ͗ıḫy / ʾiḫîya/ in the genitive; Tropper 2000: 67, 176; Pardee 2003–4: 33–35). What is remarkable is that the forms that explicitly indicate vowel harmony (that is,

49.  The form tsp͗ı ‘it devoured’ in RS 22.225 (KTU 1.96):3 (an incantation against the “evil eye”) could represent one instance. 50.  Tropper (2000: 655) lists two contracted forms from narrative poetry: dw (KTU 1.16 II 20, 23) and šrnn (KTU 1.14 IV 50), both from Kirta. However, his analyses are questionable, for the two occurrences of dw are probably infinitives (since in both instances the form follows the preposition k), while the interpretation of šrnn (an emendation of šrna͗ in the text) is uncertain and could represent a class other than III-w/y, such as a hollow root.

426

Joseph Lam and Dennis Pardee

which have initial {u͗ } or {͗ı}) occur only in prose. 51 It is uncertain whether this situation reflects one in which the phonological change had occurred everywhere but was only indicated orthographically in the prose text or whether the orthography reflects a genuine difference in pronunciation between a more conservative (“correct”) older form and a newer form. 52 One also finds some lexicographical features that distinguish poetry and prose. For instance, the noun a͗dn / ʾadānu/ 53 is attested in poetry with the meaning ‘lord’, as in later Phoenician and Hebrew. In prose, however, it appears to have the meaning ‘father’. In light of the comparative evidence, the semantic development to ‘father’ is probably best viewed as an innovation within Ugaritic. A number of other common lexical items show distinct patterns of distribution (summarized by Smith 1994: 57–58). M.  Held (1959: 174–75) notes the distinction between nouns in poetic usage such as ͗ımr (‘sheep’) and ġlm (‘youth’), as compared with š (‘sheep’) and nʿr (‘youth’) in the prose texts. These and other examples like them (including nouns for sacrificial animals) could represent more stylistic differences rather than strict diachronic ones, especially given the lack of any clear typological relationships between the poetic and prose forms. 5.3. Assessment The points examined here in brief represent somewhat of a mixed bag. On the one hand, certain distinguishing features are determined to be mere stylistic differences between poetry and prose. On the other hand, other features appear plausibly to be the result of diachronic change. Even for the potential diachronic features, however, the coexistence of certain variant features in one or the other language stratum (such as y- and t- prefixes for the 3mp YQTL in poetry or the contracted and uncontracted 3ms *qatala forms from III-w/y roots) may open up the additional possibility (not surprisingly) that what we have is in fact archaization rather than a true archaism. 54 In fact, given the 51.  For the occurrences, see Tropper 2000: 176; all of the occurrences listed are from letters and administrative texts. 52.  This kind of vowel harmony also appears to be the explanation for the two attested examples of the 1cs prefix-conjugation forms of the verb BWʾ ‘to enter’; both occurrences are found in poetic texts: u͗ba͗ / ʾubūʾa/ in RS 24.244 (KTU 1.100): 72, and u͗bu͗ / ʾubūʾu/ in RIH 78/20 (CAT 1.169): 18. At the same time, there are analogous forms in other II-weak roots that give no explicit indication of vowel harmony, such as a͗mt / ʾamūtu/ (‘I will die’) in RS 2.[004] (KTU 1.17) VI 38bis (Aqhat). Data such as these provide barriers to the formulation of a more general phonological (diachronic) rule governing such forms in prose-versuspoetic contexts. 53.  Cognate with Biblical Hebrew ʾādôn. 54.  The phenomenon of archaization is not always applied with absolute consistency everywhere; this would explain the simultaneous occurrence of y- and t- prefix forms in 3mp YQTL forms. Regarding the coexistence of contracted and uncontracted III-w/y *qatala forms in prose, one could say that the prose situation was one of historical writing (with

Diachrony in Ugaritic

427

probability that the poetic texts were inscribed contemporary with the prose texts, it is unlikely that archaization would have been a non-factor, though the distinction between the two phenomena remains difficult to make from our modern vantage point. 6. Conclusion By no means do we presume to have proclaimed the last word on diachrony in Ugaritic. But two concluding comments are in order. First, in approaching the Ugaritic corpus, all current indications point to a relatively short period (around a century) for the inscribing of the texts. While this does not mean that we should halt all attempts at diachronic analysis in Ugaritic, it does imply that we should temper our expectations appropriately when approaching the data. Second, from the two extended examples cited (the texts of Ammistamru II and the narrative poetic texts), it should be evident that considerations of stylistic difference and archaization represent significant factors in the process of identifying true archaisms—a situation that has close parallels in contemporary discussions regarding diachrony in Biblical Hebrew. In light of this, scholars of either field would do well to familiarize themselves with the arguments commonly adduced in the other field, for it is in the making of these distinctions (between archaism, archaization, and stylistic difference) that questions of methodology become all the more important. many “phonetic” forms written) whereas, in poetry, scribes tended toward the more “correct” historical form.

References Albright, W. F. 1951 The Old Testament and the Archaeology of the Ancient East. Pp. 27–47 in The Old Testament and Modern Study: A Generation of Discovery and Research, ed. H. H. Rowley. Oxford: Clarendon. 1958 Specimens of Late Ugaritic Prose. BASOR 150: 36–38. Arnaud, D. 1999 Prolégomènes à la redaction d’une histoire d’Ougarit II: Les bordereaux de rois divinisés. SMEA 41/2: 153–73. Blau, J. 1968 On Problems of Polyphony and Archaism in Ugaritic Spelling. JAOS 88: 523–26. Bordreuil, P., and Pardee, D. 1984 Le sceau nominal de ʿAmmīyiḏtamrou, roi d’Ougarit. Syria 61: 11–14. 1993 Le combat de Baʿlu avec Yammu d’après les textes ougaritiques. MARI 7: 63–70. 2001a Abécédaire (no 32). Pp.  341–48 in Études ougaritiques: I.  Travaux 1985– 1995, ed. M. Yon and D. Arnaud. RSO 14. Paris: Éditions Recherche sur les Civilisations.

428

Joseph Lam and Dennis Pardee

2001b Une incantation (no 52). Pp. 387–92 in Études ougaritiques: I. Travaux 1985– 1995, ed. M. Yon and D. Arnaud. RSO 14. Paris: Éditions Recherche sur les Civilisations. Bounni, A.; Lagarce, É.; and Lagarce, J. 2008 La tablette RIH 78/12 et le sceau nominal de ʿAmmishtamru. Pp. 153–58 in D’Ougarit à Jérusalem: Recueil d’études épigraphiques et archéologiques offert à Pierre Bordreuil, ed. C. Roche. Orient and Méditerranée 2. Paris: de Boccard. Caquot, A., and Dalix, A.-S. 2001 Un texte mythico-magique (no 53). Pp.  393–405 in Études ougaritiques: I. Travaux 1985–1995, ed. M. Yon and D. Arnaud. RSO 14. Paris: Éditions Recherche sur les Civilisations. Dalix, A.-S. 1997 Ougarit au XIIIe siècle av. J.-C.: Nouvelles perspectives historiques. CRAIBL, 819–24. 1998 Šuppiluliuma (II?) dans un texte alphabétique d’Ugarit et la date d’apparition de l’alphabet cunéiforme: Nouvelle proposition de datation des “Archives Ouest.” Semitica 48: 5–15. Day, J. 1994 Ugarit and the Bible: Do They Presuppose the Same Canaanite Mythology and Religion? Pp. 35–52 in Ugarit and the Bible: Proceedings of the International Symposium on Ugarit and the Bible, Manchester, September 1992, ed. G. J. Brooke, A. H. W. Curtis and J. F. Healey. Ugaritisch-Biblische Literatur 11. Münster: Ugarit-Verlag. Dietrich, M., and Loretz, O. 1988 Die Keilalphabete: Die phönizisch-kanaanäischen und altarabischen Alphabete in Ugarit. ALASP 1. Münster: Ugarit-Verlag. Dietrich, M.; Loretz, O.; and Sanmartín, J. 1975 Untersuchungen zur Schrift- und Lautlehre des Ugaritischen (III). UF 7: 103–8. 1976 Die keilalphabetischen Texte aus Ugarit. AOAT 24/1. Kevelaer: Butzon & Bercker / Neukirchen-Vluyn: Neukirchener Verlag. 1995 The Cuneiform Alphabetic Texts from Ugarit, Ras Ibn Hani and Other Places (KTU: 2nd, enlarged ed.). ALASP 8. Münster: Ugarit-Verlag. Durand, J.-M. 1993 Le mythologème du combat entre le dieu de l’orage et la mer en Mésopotamie. MARI 7: 41–61. Freilich, D., and Pardee, D. 1984 {ẓ} and {ṭ} in Ugaritic: A Re-examination of the Sign-Forms. Syria 61: 25–36. Freu, J. 1998 La fin d’Ugarit et de l’empire hittite: Données nouvelles et chronologie. Semitica 48: 17–39. 2004 Šuppiluliuma I ou Šuppiluliyama (II)? Res Antiquae 1: 111–24. Friedman, R. E. 1980 The MRZḤ Tablet from Ugarit. Maarav 2/2: 187–206.

Diachrony in Ugaritic

429

2003 The Bible with Sources Revealed: A New View into the Five Books of Moses. New York: HarperCollins. Garr, W. R. 1986 On Voicing and Devoicing in Ugaritic. JNES 45: 45–52. Gibson, J. C. L. 1977 Canaanite Myths and Legends. 2nd ed. Edinburgh: T. & T. Clark. Goetze, A. 1941 Is Ugaritic a Canaanite Dialect? Language 17: 127–38. Gordon, C. H. 1965 Ugaritic Textbook. AnOr 38. Rome: Pontifical Biblical Institute. Greenstein, E. L. 2006 Forms and Functions of the Finite Verb in Ugaritic Narrative Verse. Pp. 75– 102 in Biblical Hebrew in Its Northwest Semitic Setting, ed. S. E. Fassberg and A. Hurvitz. Jerusalem: Magnes Press / Winona Lake, IN: Eisenbrauns. Held, M. 1959 mḫṣ/*mḫš in Ugaritic and Other Semitic Languages: A Study in Comparative Lexicography. JAOS 79: 169–76. Herdner, A. 1963 Corpus des tablettes en cunéiformes alphabétiques découvertes à Ras Shamra– Ugarit de 1929 à 1939. MRS 10. Paris: Imprimerie Nationale and Geuthner. Holmstedt, R. 2008 The Relative Clause in Canaanite Epigraphic Texts. JNSL 34: 1–34. Huehnergard, J. 1991 Remarks on the Classification of the Northwest Semitic Languages. Pp. 282–93 in The Balaam Text from Deir ʿAlla Re-evaluated: Proceedings of the International Symposium Held at Leiden (21–24 August 1989), ed. J. Hoftijzer and G. van der Kooij. Leiden: Brill. Kaye, A. 1991 Does Ugaritic Go with Arabic in Semitic Genealogical Sub-classification? Folia Orientalia 28: 115–28. Lagarce, J., and (du Puytison-)Lagarce, É. 1995 Ras Ibn Hani au Bronze Récent: Recherches et réflxions en cours. Pp. 141–54 in Le pays d’Ougarit autour de 1200 av. J.-C.: Actes du Colloque International, Paris, 28 juin–1er juillet 1993, ed. M. Yon, M. Sznycer, and P. Bordreuil. RSO 11. Paris: Éditions Recherche sur les Civilisations. 2008 Remarques sur le matériel égyptien et égyptisant de Ras Shamra (“Maison aux Albâtres”) et de Ras Ibn Hani à la lumière de données récentes sur la chronologie de la fin d’Ugarit. Pp. 153–64 in The Bronze Age in the Lebanon: Studies on the Archaeology and Chronology of Lebanon, Syria and Egypt, ed. M. Bietak and E. Czerny. Österreichische Akademie der Wissenschaften, Denkschriften der Gesamtakademie 50. Vienna: Österreichischen Akademie der Wissenschaften. Liverani, M. 1964 Elementi innovativi nell’ugaritico non letterario. Atti della Accademia Nazionale dei Lincei, Rendiconti della Classe di Scienze morali, storische e filologiche 8/19: 173–91.

430

Joseph Lam and Dennis Pardee

Mallon, E. D. 1982 The Ugaritic Verb in the Letters and Administrative Documents. Ph.D. dissertation, Catholic University of America. Nougayrol, J. 1955 Textes accadiens et hourrites des archives est, ouest et centrals. MRS 6. Paris: Imprimerie Nationale and Klincksieck. Pardee, D. 1975 The Preposition in Ugaritic. UF 7: 329–78. 1988 Les textes para-mythologiques de la 24e campagne (1961). RSO 4. Paris: Éditions Recherche sur les Civilisations. 1997 Review of Ugarit and the Bible: Proceedings of the International Symposium on Ugarit and the Bible, Manchester, September 1992, ed. G. J. Brooke, A. H. W. Curtis, and J. F. Healey. JAOS 117: 375–78. 2000 Les textes rituels. RSO 12. Paris: Éditions Recherche sur les Civilisations. 2001a Le traité d’alliance RS 11.772+. Semitica 51: 5–31. 2001b Canaan. Pp.  151–68 in The Blackwell Companion to the Hebrew Bible, ed. L. G. Perdue. Oxford: Blackwell. 2002a RIH 77/27, RIH 77/12, RIH 78/26 et le principe de l’écriture cunéiforme alphabétique. Syria 79: 51–63. 2002b Ritual and Cult at Ugarit. SBLWAW 10. Atlanta: Society of Biblical Literature. 2003–4  Review of Ugaritische Grammatik, by J. Tropper. AfO 50: 1–404. http:// orientalistik.univie.ac.at/publikationen/archiv-fuer-orientforschung/. 2007 The Ugaritic Alphabetic Cuneiform Writing System in the Context of Other Alphabetic Systems. Pp. 181–200 in Studies in Semitic and Afroasiatic Linguistics Presented to Gene B.  Gragg, ed. C.  L. Miller. Studies in Ancient Oriental Civilization 60. Chicago: Oriental Institute. 2010 RS 15.117 et l’origine de l’alphabet cunéiforme d’Ougarit: Rapport de collation. Or 79: 55–73. 2012 {g} as a Paleographic Indicator in Ugaritic Texts. Pp. 111–26 in Palaeography and Scribal Practices in Syro-Palestine and Anatolia in the Late Bronze Age, ed. E. Devecchi. Leiden: Nederlands Instituut voor het Nabije Oosten. Parker, S. B. 1970 Studies in the Grammar of Ugaritic Prose Texts. Ph.D. dissertation, The Johns Hopkins University. Rainey, A. F. 1974 The Ugaritic Texts in Ugaritica 5. JAOS 94: 184–94. Segert, S. 1984 A Basic Grammar of the Ugaritic Language with Selected Texts and Glossary. Berkeley: University of California Press. Singer, I. 1999 A Political History of Ugarit. Pp. 603–733 in Handbook of Ugaritic Studies, ed. W. G. Watson and N. Wyatt. Leiden: Brill. Sivan, D. 1997 A Grammar of the Ugaritic Language. HO 1/28. Leiden: Brill.

Diachrony in Ugaritic

431

Smith, M. S. 1994 The Ugaritic Baal Cycle, vol. 1: Introduction with Text, Translation and Commentary of KTU 1.1–1.2. Leiden: Brill. Smith, M. S., and Pitard, W. T. 2009 The Ugaritic Baal Cycle, vol. 2: Introduction with Text, Translation and Commentary of KTU/CAT 1.3–1.4. Leiden: Brill. Soldt, W. H. van 1991 Studies in the Akkadian of Ugarit: Dating and Grammar. AOAT 40. Kevelaer: Butzon & Bercker; Neukirchen-Vluyn: Neukirchener Verlag. Tropper, J. 1993 Morphologische Besonderheiten des Spätugaritischen. UF 25: 389–94. 1994 Is Ugaritic a Canaanite Language? Pp.  343–53 in Ugarit and the Bible: Proceedings of the International Symposium on Ugarit and the Bible, Manchester, September 1992, ed. G. J. Brooke, A. H. W. Curtis, and J. F. Healey. Ugaritisch-Biblische Literatur 11. Münster: Ugarit-Verlag. 2000 Ugaritische Grammatik. AOAT 273. Münster: Ugarit-Verlag. Villard, P. 1986 Un roi de Mari à Ugarit. UF 18: 387–412. Virolleaud, C. 1965 Le palais royal d’Ugarit V: Textes en cunéiformes alphabétiques des archives