Journal of Semitic Studies, Volume 54, Number 1, Spring 2009

ISSN 0022-4480 JOURNAL OF Semitic Studies VOLUME LIV. NO. 1 SPRING 2009 INFORMATION FOR AUTHORS Contributions, books...

Author: Editorial Committee

107 downloads 1913 Views 7MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

ISSN 0022-4480

JOURNAL OF

Semitic Studies VOLUME LIV. NO. 1 SPRING 2009

INFORMATION FOR AUTHORS Contributions, books for review, and other editorial communications should be addressed to: The Editors Journal of Semitic Studies Middle Eastern Studies School of Languages, Linguistics and Cultures University of Manchester Manchester M13 9PL UK Article Length Editorial policy is normally to favour shorter articles, i.e. no longer than 10,000 words. Typescript layout Articles and reviews should be typed on one side of the paper (A4 or 8.5 x 11 inches) with double spacing and generous margins on one side of the paper only. Authors of articles should submit TWO COPIES, and disk, if available. Non Roman script is to be written as clearly as possible and used as sparingly as possible. Notes should be numbered consecutively, double spaced and placed at the end. The Editors will not accept for publication any typescripts which do not conform to these standards. For detailed instructions please see the style sheet. Abstracts Authors of articles are required to provide an abstract, which should be a single para graph, not exceeding 150 words. Submission We will also normally require the final revised version to be supplied in hard copy and electronically. Whilst we can accept most wordprocessor formats the preferred option is Word. The files should be submitted in the wordprocessor's native format and as an RTF (Rich Text Format). Illustrations Electronic submission of illustrations. We can accept 3.5 inch floppy disks, CD ROMs and email attachments. High quality print outs of all figures are still essential, as they will be scanned for reproduction in the event of problems with the electronic files. Figures must be saved as high resolution TIF files (at least 300 pixels per inch for photographs, and 1200 pixels per inch for black and white line drawings). Offprints and Proofs Authors of articles will receive one copy of the issue containing their article and twenty five offprints of their article free of charge. Alteration on the proofs, other than the correction of printer's errors, may be disallowed. Copyright and Permissions Manuscripts submitted will be expected to contain original work and should not have been published in abridged or other form elsewhere. It is a condition of publication in the Journal that authors grant an exclusive licence to The University of Manches ter. Any requests from third parties to reproduce articles are handled on behalf of the journal by Oxford University Press. This will also allow the article to be as widely dis seminated as possible and will protect the rights of the author and OUP. In granting an exclusive licence, authors may use their own material in other publications pro vided that the journal is acknowledged as the original authority for publication, and Oxford University Press is notified in writing and in advance.

JOURNAL OF

Semitic Studies V OLUME LIV. NO. 1 SPRING 2009

Contents articles JAN KEETMAN, Wechselwirkung von Vokalen und Gutturalen im Semitischen unter dem Einfluss anderer Sprachen: die Beispiele des Akkadischen und Hebräischen NA’AMA PAT-EL, The Development of the Semitic Definite Article: A Syntactic Approach GORDON J. HAMILTON, A Proposal to Read the Legend of a Seal-Amulet from Deir Rifa, Egypt as an Early West Semitic Alphabetic Inscription VINCENT DECAEN, Theme and Variation in Psalm 111: Metrical Phrase and Foot in Generative Perspective ROBERT HOLMSTEDT, Word Order and Information Structure in Ruth and Jonah: A Generative-Typological Analysis FRANÇOIS BRON, Notes sur les inscriptions néo-puniques de Henchir Medeina (Althiburos) TZVI NOVICK, The Modality of Òarik in Tannaitic Hebrew SHLOMY RAISKIN, Talmudic Aramaic Fauna Names: Murzema and Shaqi†na NADEZHDA VIDRO, A Newly Reconstructed Karaite Work on Hebrew Grammar GERRIT BOS and Y. TZVI LANGERMANN, The Introduction of Sergius of Reshäina to Galen’s Commentary on Hippocrates’ On nutriment ALESSANDRA AVANZINI, Origin and Classification of the Ancient South Arabian Languages AARON D. RUBIN, The Functions of the Preposition k- in Mehri HILLA PELED-SHAPIRA, From Conventional to Personal, or: What Happened to Metaphor under the Influence of Ideology The case of Gha’ib Tu¨ma Farman

1 19 51 81 111 141 149 161 169 179 205 221

227

reviews Georges BOHAS and Mihai DAT, Une théorie de l’organisation du lexique des langues sémitiques: matrices et étymons, Philippe CASSUTO and Pierre LARCHER (eds), La formation des mots dans les langues sémitiques (Edward LIPINSKI) Aaron D. RUBIN, Studies in Semitic Grammaticalization (David STEC) Jean-Jacques GLASSNER, Translated and Edited by Zainab BAHRANI and Marc VAN DE MIEROOP, The Invention of Cuneiform: Writing in Sumer (Jon TAYLOR) J.A. HALLORAN, Sumerian Lexicon: A Dictionary Guide to the Ancient Sumerian Language (J.N. POSTGATE) Stefanie U. GULDE, Der Tod als Herrscher in Ugarit und Israel (W.G.E. WATSON) Hallvard HAGELIA, The Tel Dan Inscription: A Critical Investigation of Recent Research on Its Palaeography and Philology (Bob BECKING) Ernst JENNI, Studien zur Sprachwelt des Alten Testaments II (Martin F.J. BAASTEN) Robert D. MILLER II S.F.O., Chieftains of the Highland Clans: a History of Israel in the 12th and 11th Centuries B.C., Ann E. KILLEBREW, Biblical Peoples and Ethnicity: an Archaeological Study of Egyptians, Canaanites, Philistines, and Early Israel 1300-1000 B.C.E. (Eveline J. VAN DER STEEN) Bernard S. JACKSON, Wisdom Laws: A Study of the Mishpatim of Exodus 21:1 22:16 (Jeffrey STACKERT) Hilary LIPKA, Sexual Transgression in the Hebrew Bible (Jenni WILLIAMS) Kevin A. WILSON, The Campaign of Pharaoh Shoshenq I into Palestine (K.A. KITCHEN) Thomas B. DOZEMAN and Konrad SCHMID (eds), A Farewell to the Yahwist? The Composition of the Pentateuch in Recent European Interpretation (William JOHNSTONE) Alice HUNT, Missing Priests: The Zadokites in Tradition and History (Deborah ROOKE) Lena-Sofia TIEMEYER, Priestly Rites and Prophetic Rage: Post-Exilic Prophetic Critique of the Priesthood (Deborah ROOKE) Nicholas P. LUNN, Word-Order Variation in Biblical Hebrew Poetry: Differentiating Pragmatics and Poetics (Robert D. HOLMSTEDT) D. GOODBLATT, Elements of Ancient Jewish Nationalism (Doron MENDELS) Beverly P. MORTENSEN, The Priesthood in Targum Pseudo-Jonathan: Renewing the Profession (Simon ADNAMS LASAIR) Alexander SAMELY, Forms of Rabbinic Literature and Thought: An Introduction (Günter STEMBERGER)

251 253 254 255 258 259 261

265 270 272 274

276 278 281 283 285 287 289

Emanuela TREVISAN SEMI, Jacques Faitlovitch and the Jews of Ethiopia (Steven KAPLAN) Alessandro BAUSI and Alessandro GORI, Tradizioni orientali del «Martirio di Areta». La prima recensione Araba e la versione Etiopica. Edizione critica e traduzione (Michael A. KNIBB) John C. LAMOREAUX (translator), Theodore Abu Qurrah (Hugh GODDARD) Garth FOWDEN, QuÒayr Ämra: Art and the Umayyad Elite in Late Antique Syria (Andrew MARSHAM) Paul STARKEY, Modern Arabic Literature (Ami ELAD-BOUSKILA)

292

294 296 298 302

short notes Steven E. FASSBERG and A. HURVITZ (eds), Biblical Hebrew in its Northwest Semitic Setting: Typological and Historical Perspectives (J.F. HEALEY) Hélène LOZACHMEUR, La Collection Clermont-Ganneau: ostraca, épigraphes sur jarre, étiquettes de bois (J.F. HEALEY) Karel JONGELING, and ROBERT M. KERR, Late Punic Epigraphy: an Introduction to the Study of Neo-Punic and Latino-Punic Inscriptions (J.F. HEALEY) Jean-Claude HAELEWYCK, Grammaire comparée des langues sémitiques: élements de phonétique, de morphologies et de syntaxe (J.F. HEALEY) Suzanne SCHOLZ, Introducing the Women’s Hebrew Bible (Deborah ROOKE)

307 308 309 310 311

WECHSELWIRKUNG VON VOKALEN UND GUTTURALEN IM SEMITISCHEN UNTER DEM EINFLUSS ANDERER SPRACHEN: DIE BEISPIELE DES AKKADISCHEN UND HEBRÄISCHEN JAN KEETMAN ISTANBUL

Abstract In this article it is argued that the sound shifts (*ai >) e > i and pharyngeal + a > e > e (both long and short) from Imperial Akkadian to Old Babylonian prove the existence of two vowel qualities of e in Imperial Akkadian. This may well be explained if we regard the sec ond sound shift as the shift of the pressing of the pharynx from a consonant to a vowel, probably under Sumerian influence, resulting in a vowel near a, like German ä, which may have existed in Sumerian. While in Akkadian some consonants where lost but left an imprint on the vowels, in Hebrew auxiliary vowels evolved which sus tained the pronunciation of nearly the same consonants. While Akkadian made a compromise, Hebrew resisted when its daily use faded or was even reconstructed in some communities in antiquity.

a) Akkadisch Zu den Verdiensten des Buches von Rebecca Hasselbach über die syllabischen Texte des sargonischen Akkadischen1 gehört die nahezu lückenlose Beweisführung dafür, dass das Sargonic Akkadian (hier hinfort „Reichsakkadisch“) ein Vorläufer des Altbabylonischen (und mithin auch des Ur III-Akkadischen) ist.2 Übereinstimmungen mit 1

Hasselbach 2005. Neben Stärken hat diese Arbeit leider auch zwei grundsätzliche Schwächen. Die erste ist die Vernachlässigung der Personennamen. Zwar hat Hasselbach Recht damit, dass das Material der Personennamen vom übrigen sprachlichen Material zu trennen ist, doch sollte auch es möglichst vollständig gesammelt und für sich er schlossen werden. Manche Probleme und auch Missverständnisse kommen daher, dass diese Arbeit nicht unternommen wurde. Die andere Schwäche ist die zum Teil ungenügende Rezeption ihrer Vorgänger. Man lese sich dazu z. B. die zum Teil 2

1

DIE BEISPIELE DES AKKADISCHEN UND HEBRÄISCHEN

dem Assyrischen erklärt Hasselbach als Archaismen, die im späteren Babylonischen ausgefallen sind.3 L. Kogan und K. Markina haben eine Reihe berechtigter Einwände gegen diesen Standpunkt formuliert. Unserer Meinung nach wird Hasselbachs These dadurch aber insgesamt nicht widerlegt. Da dieser Artikel eigentlich schon abgeschlossen war, kann ich auf Kogan und Markina nur knapp eingehen und empfehle ihre Rezension zu Hasselbach 2005 auf jeden Fall zur Lektüre.4 Mit Hasselbach und Sommerfeld ist der Autor der Meinung, dass bei den Verba II inf. reichsakkadisch juken nur mit altbabylonisch (aB) ukin zu verbinden ist. Eine Angleichung von (j)uken an das starke Verbum als uka’’ in wie altassyrisch (aA) wäre zwar denkbar, ist aber wegen des Schwundes des Stimmabsatzes im Akkadischen unwahrscheinlich. Also ist die Nähe zur babylonischen Form, anders als Kogan und Markina meinen, für die Dialektgeschichte relevant. Der pleonastische Gebrauch der „Subjunktivendungen„ u + ni > -uni wie im Assyrischen ist zwar innovativ, doch das Argument wird dadurch geschwächt, dass –uni auch aA selten ist, reichsakkadisch auch nur –u belegt ist und –ni aB auch nach Vokal fast vollständig ausgefallen ist. Die „assyrische“ Form des Verbaladjektivs Gt pitarsum statt pitrusum stammt aus der Liebesbeschwörung MAD 5, 8, die obwohl in Kis gefunden, von den anderen reichsakkadischen Texten ebenso wie von der babylonischen und assyrischen Tradition des Akkadischen stark abweicht.5 Daher ist der Text aus der Betrachtung auszuscheiden. massive Kritik an Sommerfeld auf den Seiten 59 60 (betrifft die Zeichen DU und GU) durch und das was Sommerfeld 1999, 21 zu dem Thema geschrieben hat (zu den Belegen vgl. auch noch aus Hasselbachs Index pataqum und sadadum). Nicht erkennbar ist auch, warum Hasselbach auf S. 87 betont, dass in ihrer Arbeit (im Ge gensatz zu Gelb, 1961) I = /yi/ gebraucht wird. Vgl. dazu GAG3 §22c*, §75e* u. a. Wie es bei diesem Thema kaum anders sein kann, vertritt der gegenwärtige Autor an vielen Stellen von Hasselbach abweichende Auffassungen. Doch solche Differen zen schmälern nicht den Wert der von Hasselbach vorgelegten Arbeit. 3 Vgl. die Zusammenstellung bei Sommerfeld 2003: 572 3. Der dort genannte Plural m. cas. obl. Reichsakkadisch und Assyrisch e, babyl. i geht wohl ebenfalls auf ai zurück. Vgl. Hasselbach 2005: 179 80 Anm.100; Cross 2003: 355 6. Noch offensichtlicher ist dies beim Stativ der Verbaladjektive der Verben II inf. Vgl. paris zu ken (assyr. Reichsakkad.), kin babyl. 4 Siehe Kogan und Markina 2006. 5 Eben deshalb ist die Interpretation schwierig. Als Abweichungen kann man sehen: 1) Verbalpräfix ti statt ta (sonst nur Ebla und Mâri). Einzige Darstellung von etymologischem *g durch GA: ru GA tim „des Geifers“ (rgw) MAD 5, 8, 12, cf. Kogan und Markina 2006, 567. 3) Akkadisch sonst nicht belegte Wurzeln a) Òawarum „Hals“ MAD 5, 35; 36, cf. AHw 1087; b) duarum „um jemanden herum laufen“ (arab. dur) MAD 5, 8, 21; 22. 2


Mit der Einführung von Dativsuffixen sind beide Dialekte innovativ. Die assyrischen Formen in einigen Königsinschriften und Eigennamen sind also echte Assyriasmen und keine Archaismen. Es sind aber auch Suffixe nach babylonischem Muster belegt.6 Gegenüber diesen spärlichen und nie ohne babylonische Variante auftretenden Assyriasmen verbinden die Formen der Imperative, sowie Infinitive und Verbaladjektive von D- und S-Stamm, purris, supris, purrus und suprus das Reichsakkadische mit dem babylonischen Dialekt.7 Ebenso der beginnende Schwund von –ni und die Morphologie der Verba II inf. Nähe zum Altbabylonischen ist auch aus anderen Gründen zu erwarten, schließlich ist das durch Textfunde zugängliche Verbreitungsgebiet des Reichsakkadischen weitgehend identisch mit der späteren Verbreitung des Altbabylonischen. Bezüglich des bereits im Reichsakkadischen eingetretenen Lautwandels ai > e führt Hasselbach aus, dass er nicht notwendig eine Isoglosse zum Assyrischen darstelle. Der Lautwandel könnte im Babylonischen in zwei Schritten vollzogen worden sein, nämlich ai > e > i.8 Der zweite Schritt setzt aber im Babylonischen offenbar den ersten voraus, denn sonst bleibt e erhalten. Ein Lautwandel kann aber kaum auf eine Form Rücksicht nehmen, die längst nicht mehr gesprochen wurde. Hier liegt das größte Problem, das Hasselbachs Ansatz aufwirft. Nun gibt es für e im Babylonischen außer ai im wesentlichen nur zwei Quellen, die in das Reichsakkadische zurückreichen: 1) a > e in Nachbarschaft zu den alten Pharyngalen Ì, äin und vielleicht von g9 und Längung. 2) Fernwirkung eines Pharyngals (oder von g) mittels der babylonischen Vokalharmonie.10 6

Cf. Kogan und Markina 2006, 578 80. Siehe Hasselbach 2005, Abschnitt 4. 5. 8 auf S. 210 und 4. 5. 10 mit Anm. 178 auf S. 211. Beim Imperativ des S Stammes der Verben I w ist im Reichsakkadischen nur su belegt, nicht se wie im Assyrischen normal. Bei den übrigen Formen ergibt sich je doch ein anderes Bild. Im Süden wird mit su gebildet, in einem Text aus Kis wie im Altbabylonischen überwiegend mit sa , ansonsten von Kis an nach Norden, wie im Assyrischen mit se . Doch ist die Bildung des Vokals nach dem s des S Stammes bei den Verben I w im Akkadischen durchweg durch konkurrierende Formen und Unstetigkeiten geprägt. Z. B. gehört wabalum altbabyl. zur a Klasse, mittelbabyl. zur e Klasse. Vgl. Belege und Diskussion bei Hasselbach 2005: 224 6. 8 S. 91 Anm. 186. 9 Vgl. Kogan 2001. Trotz der Einwände bei Keetman 2004, 7 Anm. 7 ist Kogan sicher darin zuzustimmen, dass die auslösende Wirkung von g auf a > e ungenü gend belegt ist. 10 Wohl noch nicht reichsakkadisch (Hasselbach 2005: 121), aber auf dem un ter 1) genannten Prozess fußend. Bei der historischen Betrachtung solcher Prozesse 7

3


Die mögliche Entwicklung a > e in Gegenwart von r und l ist altbabylonisch, aber nicht altassyrisch und reichsakkadisch belegt und kann hier übergangen werden. Dass die unterschiedlichen Prozesse nicht in einem Schritt zur gleichen Vokalqualität führen, ist durchaus Einsichtig. Die Verschmelzung von ai liefert am ehesten einen langen Vokal in der Mitte zwischen a und i, den wir e schreiben wollen und der vermutlich auch späterem e im Akkadischen entsprach. Pharyngale bedeuten per definitionem eine Engebildung im Rachen, die auch mit der hinteren Zunge möglich ist. Diese Position der Zunge kommt einem der beiden Merkmale des Kardinalvokals a nahe, nämlich der zurückgezogenen Zunge, die damit hinten eine Enge bildet. Es fehlt das zweite Merkmal, die maximale Mundöffnung. Dies führt zu einer Annäherung an die Stellung von ä. Man kann dies auch von der akustischen Wirkung her betrachten. Eine Engebildung im hinteren Teil des Ansatzrohres führt zu einer Anhebung der ersten beiden Formanten. Die für a charakteristische Öffnung vorne hebt ebenfalls den ersten Formanten, senkt aber den zweiten. D. h. der erste Formant erreicht nicht ganz sein Maximum, wie bei a, der zweite ist etwas höher, wie bei den vor a artikulierten Vokalen. Denkt man sich also einen Übergang zwischen einem konsonantischen Pharyngal und dem Vokal a, so mag ein etwas nach vorne verlagerter Vokal ä herauskommen. 11 Weiter vorne ist der Grund für die hintere Enge, die zurückgezogene Zunge nicht mehr gegeben. Daher dürfte ä der Startpunkt des Lautwandels im Akkadischen gewesen sein. Der Lautwandel wurde wahrscheinlich durch einen entsprechenden Vokal im Sumerischen begünstigt.12

muss auch berucksichtigt werden, dass zwischen Sarkalisarri von Akkad und der zweiten Halfte der Regierungszeit Sulgis von Ur eine Spanne von 100 oder 150 Jah ren liegt, fur die das Akkadische nur außerst sparlich dokumentiert ist. 11 Dazu ausführlich und mit Erörterung der akustischen Wirkung, Keetman 2004. 12 Vgl. den Exkurs am Ende dieses Artikels. 4


Zur leichteren Orientierung ist das Verhalten der beiden unteren Formanten F 1 und F 2 in der obigen Graphik angegeben, die nur die Tendenz aufzeigen soll, die Formanten, ebenso wie die Vokale sind ja ohnehin als Mittelwerte zu verstehen. Für das Hebräische wäre Segol zwischen a und e anzunehmen und PataÌ entweder bei a oder zwischen a und Segol. Bei einer Verschmelzung eines Pharyngales mit a könnte also ä herauskommen, während ein Vokal mit mittlerer Mundöffnung eine hintere Enge erschwert13 und außerdem weiter von der Mundhaltung bei a entfernt ist.14 Nehmen wir als Ergebnis dieser Prozesse zunächst ä bzw. langes ä (hier behelfsmäßig ä+ geschrieben) an und für ai > e ein mehr geschlossenes e, so können wir die weitere Entwicklung zum Altbabylonischen als parallele Erhöhung und Vorverlagerung der beiden langen Vokale beschreiben: ä+ > e, e > i. Da ein kurzes e nicht auf ai zurückgehen kann, gab es nur ein kurzes ä und folglich nur ä > e, aber keinen e > i parallelen Lautwandel e > i. Als Motivation für die letzten Schritte kann man den Versuch sehen, bei den kurzen Vokalen gleiche Abstände zwischen a, e, i herzustellen. Die Bereinigung bei den langen Vokalen wäre dann dieser Entwicklung gefolgt. Man kann sich dies noch in zwei Schritte aufgeteilt denken. Zunächst e in der Mitte zwischen a und i, sowie ä+ nahe an a. Dann ein Ausgleich, der gleiche Abstände zwischen den vier Vokalen herstellt. Schließlich in Analogie zu den kurzen Vokalen die Bereinigung des Systems zu 3 Positionen. Dadurch wird das Ausweichen von älterem e vor e aus ä+ etwas plausibler. Die babylonische Vokalharmonie wäre dann zunächst eine Angleichung von a an ä gewesen, vermutlich unter dem Einfluss der Wurzelharmonie des Sumerischen.15 Im Assyrischen, wo es diese Angleichung nicht gab und ä+ folglich schwächer vertreten war, wäre dann ä+ einfach an e angeglichen worden und die Anhebung und Vorverlagerung von e aus ai zu i ist unterblieben. Wenn wir annehmen, dass semitische Pharyngale unter sumerischem Einfluss durch einen pharyngalisierten Vokal kurz vor a er13 Diese könnte, statt mit dem hinteren Teil der Zunge auch mit den Pharynx muskeln erzeugt werden. Doch wir gehen von nicht absichtlich herbeigeführten Effekten bei der Bildung der Vokale aus und betrachten deshalb nur die Bewegung der Zunge. 14 Die zentralisierende Wirkung der Pharyngale auf Vokale belegt auch das Patach Furtivum im Hebräischen (s. u.). 15 Zur Vokalharmonie innerhalb sumerischer Wortwurzeln vgl. Keetman 2005: 11 12.

5


setzt wurden, so lässt sich auch die Fernwirkung von Pharyngalen im Babylonischen erklären. Wir nehmen an, in dem Wort *qamÌum „Mehl“, sei das nicht in Nachbarschaft eines Vokales stehende Ì durch ein kurzes ä ersetzt worden. Durch die babylonische Vokalharmonie habe dieses auf das a der ersten Silbe gewirkt. Später fiel es wie der Stimmabsatz aus,16 wodurch die erste Silbe zur Erhaltung ihrer Quantität gelängt wurde. Also: *qamÌum > *qamäum > *qämäum > qemum. Wie zu erwarten tritt dieser Effekt im Assyrischen nicht ein, weil es keine Angleichung von a an e gibt (bzw. ursprünglich a an ä).17 Obwohl babylonisch auch r und l den Lautwandel a > e verursachen können, haben sie nicht die gleiche silbenübergreifende Wirkung wie die Pharyngale. Z. B. qeberum „beerdigen“ aber qabrum „Grab“.18 Auch dies ist nach unserem Modell zu erwarten, weil an ihrer Stelle kein Vokal gebildet wird. Unsere Vermutung, der Lautwechsel a > e in Gegenwart eines etymologisch anzunehmenden Pharyngals sei durch sumerischen Einfluss zu erklären, muss aber erst noch in Einklang mit der geschichtlichen Entwicklung des Lautwandels gebracht werden. Hasselbach schreibt: „(…) the change of /a/ and /i/ to /e/ in any environment was more advanced in the northern periphery, that is in the Diyala region, than in northern and southern Babylonia. The sound change probably originated in the North and then gradually spread further South.“19 Das steht im offenen Widerspruch zu der Untersuchung von Piotr Steinkeller zu der Wurzel b¨l „Herr“ in Personennamen der Fara-Zeit und prasargonischen Zeit, mit Belegen insbesondere aus Fara und Isin, also Sudbabylonien.20 Die Träger dieser Namen scheinen zum guten Teil längere Zeit in dem Gebiet ansässig gewesen zu sein. Bei der Wurzel sm¨ „hören“ ergibt sich sogar, dass die älteren Namen aus Babylonien bereits e haben, während es in der sargonischen Zeit einen Rückgriff auf a in den nun verfügbaren Textbelegen gibt.

16

Vor diesem kurzen ä ist natürlich ein Stimmabsatz zu denken, der hier der Übersichtlichkeit wegen ausgelassen wurde. 17 Tropper 1999 schlägt eine andere Erklärung, nämlich Metathese des Pharyngals vor. Eine Metathese gerade dieser Konsonanten kurz vor ihrem Ausfall, die im wesentlichen auf das Babylonische beschränkt bliebe, ist jedoch nicht sehr wahrscheinlich. Vgl. auch Keetman 2004: 11 mit Anm. 14. 18 GAG3 §9b. 19 Hasselbach 2005: 120. 20 Steinkeller 2004: 12 14. 6


In Lagas arbeitete unter Lugalanda und Irikagina ein in den Urkunden häufig genannter Brauer mit Namen ì-lí-be6(PI)-lí.21 Vgl. auch das Lehnwort be6-lu5-da < belutum „Herrschaft“.22 In einer Fara-Zeitlichen oder wenig jungeren Kaufurkunde wird mehrfach ein „Bauer“ (engar) is-me-ì-lum erwähnt,23 der beim Verkauf dabeisitzt. Vermutlich ein Nachbar,24 der mit seiner Anwesenheit bekräftigt, dass er keine Ansprüche erhebt. Jedenfalls jemand, der nicht auf der Durchreise ist. Im gleichen Text kommt der Name bé-li-li vor.25 Der Träger ist Zeuge des Feldverkaufs und mithin mit Sicherheit ortsansässig. Dieser meist BE-lí-lí geschriebene Name steht fast sicher für bel-ili oder beli-ili „Mein Gott ist der/mein Herr“26 Das Zeichen BI = bé ist sowohl im Sumerischen als auch im Reichsakkadischen das normale Zeichen für be.27 Der Name begegnet auch in Fara selbst als Bé-li-li28 in einer Rationenliste(?), die Leute mit verschiedenen Berufsbezeichnungen wie „Schreiner“, „Baumeister“, „Schreiber“ auffuhrt. Soweit erkennbar sind die ubrigen Namen fast alle sumerisch und mehrere haben als theophores Element die Stadtgottheit von Fara/Suruppak dsùd. Die Alternative, in Bé-li-li einen „banana name“ zu sehen, ist un21

VS 25, 41 v 10; AWL 43 iii 8; 64 ii 4; Nik I 22 vii 2 passim. Ukg. 4 vii 26 = 5 vii 9. Vgl. Selz 1998: 324 mit Anm. 195 und Literatur dort. 23 ELTS 15 i 24; ii 23; iii 21 passim. 24 Gelb, Steinkeller und Whiting 1991: 237 8 stellen fest, dass ENGAR hier für ENGAR.US steht und dass ihnen und dem „(field) scribe“ bei Hausverkäufen der „master house surveyor“ und der „street herold“ gegenüberstehen. Daraus schließen sie, der ENGAR(.US) müsse „a high administrative official in charge with agricultural activities“ sein. Andererseits ist die Bedeutung US = ús „angrenzen, be nachbart sein“ altsumerisch gut bezeugt (Behrens und Steible 1983: 361 2). Neh men wir nun eine Parallele zwischen dem „master house surveyor“ (um mi a lú é és ∞ar) und dem „(field)scribe“ (dub sar( gána)) an, so steht der „street herald“ (ni∞ir sila) dem engar ( ús) gegenüber. Während die ersten beiden Berufe mit Vermessung, bzw. Niederschrift zu tun haben, dient der ni∞ir sila wohl der Bekanntmachung der neuen Verhältnisse im Viertel. Auf dem Land gibt es weniger Nachbarn und es reicht vielleicht, wenn man sie einfach bittet, doch beim Kauf anwesend zu sein. 25 ELTS 15 xii 20; xiii 4. 26 Vgl. Di Vito 1993: 94, mit Hinweis u. a. auf sargonisch EN ì lí, be lí DINGIR und Ur III EN.DINGIR.MU. Als ältere Form zum präsargonisch in Lagas belegten ì lí be6 lí ist beli ili (bzw. ba‘li ili) wegen der Veränderung der Wort stellung zu erwarten. Siehe Di Vito 1993: 289 90. Die einzige reale Möglichkeit für einen Irrtum wäre, dass uns ein „banana name“ einen Streich spielt. Zwar gibt es in ELTS 15 einen ähnlichen „banana name“ Ki lí lí, doch findet sich bé/be6 li li auch in Fara, wo diese Namensbildung ziemlich unublich ist. 27 Cf. Sommerfeld 2003: 572. 28 NTSS 569 Rs.(?) iii 6’. 22

7


wahrscheinlich, weil „banana names“ in Fara sehr selten sind.29 Aus Fara gibt es dann noch die Schreibung des Namens als be6-li-li.30 In den syllabischen Texten der Akkad-Zeit gibt es hingegen in Südbabylonien nur einen Beleg für a > e in Gegenwart von Ÿ. Der Verfasser des betreffenden Briefes ist ein Mann namens Me-zi.31 Zugleich ist Me-zi auch der einzige Absender, der Briefe in Akkadisch und Sumerisch schreibt. Einen Brief fängt er mit akkadischer Einleitungsformel an, schreibt den eigentlichen Brief aber auf Sumerisch.32 Es ist daher möglich, dass Mezi aus einer an Zweisprachigkeit gewöhnten Gegend stammte und folglich auch sein Akkadisch eher dem Akkadischen des Südens entsprach. Zugegeben, die Belege für den Vollzug des Lautwechsels a > e in Gegenwart von Ÿ in Namen von sicher ortsansässigen Personen aus Süd- und Mittelbabylonien vor dem Reich von Akkad sind spärlich. Die Träger von sargonischen Namen, die a > e dokumentieren, etwa aus Nippur und Umma, könnten mit den Sargoniden gekommen sein, auch wenn ihre Namen nicht den Dialekt der Königsinschriften dokumentieren. Doch zwei Faktoren werten unsere Beispiele auf: Sie stammen aus einem eher konservativen sprachlichen Material und so weit der Autor sieht stehen ihnen keine Belege für den Erhalt von a in Gegenwart von Ÿ in präsargonischen Personennamen aus dem Süden gegenüber. Die Beobachtungen passen ferner zu Walter Sommerfelds „vorläufigem Ergebnis“ , dass das geschriebene Akkadisch der Akkad-Zeit, „konsequent als offizielle Sprache eingeführt wurde“.33 Das Akkadisch, welches wir in Babylonien, vor allem im Süden finden, scheint das Akkadisch einer mit den Sargoniden gekommenen, eventuell nur dünnen Schicht zu sein, das dem ebenfalls mehr archaischen Akkadisch der Königsinschriften nahe stand.34 29 Der im gleichen Text Rs.(?) i 3’ genannten Namen ha li li ist wegen ha lí lum, WF 22 ix, vermutlich als hal ili „Mein Gott ist ein Onkel (mütterlicherseits)“ zu deuten. Für ähnliche Namen vgl. AHw 314b oben. 30 RTC 12 iii 2. 31 Ad 3, 17 19: BE lí É wa a ti [l]i /is me\ „Mein Herr möge mein Wort hö ren!“ Das letzte Wort ist stark beschädigt, aber nach der Kopie hinreichend sicher. 32 Kienast undVolk 1995: 43; cf. Michalowski 1993: 28 30. 33 Sommerfeld 2003: 585; ähnlich Westenholz 1999: 33. Allerdings spielt bei beiden Autoren auch noch der Ubergang aj > e statt i wie im Altbabylonischen eine Rolle bei der Trennung des Reichsakkadischen vom lokalen Akkadisch Babyloniens. 34 Eine ähnliche Analyse erwägt auch Hasselbach 2005: 232, zieht aber nicht den Schluss, dass mithin die angenommene Vorreiterrolle der Diyala Region, nicht zu beweisen ist.

8


b) Eine alternative Erklärung für a > e im Akkadischen? In einem kürzlich35 erschienen Artikel behandelt N.J.C. Kouwenberg die Veränderung protosemitischer Gutturale im Altassyrischen.36 Da dieser Artikel, wenn auch eher beiläufig, eine alternative Erklärung für den Lautwandel a > e (bzw. ä) nennt, müssen wir auf ihn kurz eingehen. Kouwenberg kommt, insbesondere durch Beobachtungen an schwachen Verben, zu dem Ergebnis, dass alter Stimmabsatz und äin einen Stimmabsatz im Altassyrischen ergeben. Zu den gutturalen Frikativen stellt Kouwenberg fest: „h, Ì have been dropped and replaced by a long vowel or a glide, depending on their position in the word“.37 Diese Resultate werden überzeugend begründet.38 Im Zusammenhang mit den hier vertretenen Thesen muss aber ein Zwischenschritt in Kouwenbergs Argumentation erörtert werden. Unter Hinweis auf Blake39 führt Kouwenberg zu Ì aus: „the fact that it causes Ecolouring suggests that before being lost it underwent palatalization, at least after a vowel“.40 Wenn E-Färbung von a nur durch Palatalisierung des Konsonanten zu erklären wäre, dann müsste dies auch für äin gelten. Dies ist mit der von Kouwenberg festgestellten Entwicklung dieses Konsonanten zu einem laryngalen Verschlusslaut nicht vereinbar und wird von ihm auch nicht behauptet. Eine Palatalisierung vor und hinter a ist außerdem nicht gerade eine wahrscheinliche Entwicklung. Schließlich sollte *aj im Altbabylonischen am ehesten wie *ai zu i und nicht wie *aÌ zu e führen.41 35 Das Manuskript zu diesem Artikel wurde bis auf kleine Änderungen im Januar 2007 abgeschlossen. 36 Kouwenberg 2006. 37 Kouwenberg 2006: 175. 38 Man könnte nur einwenden, dass sich von Beobachtungen an schwachen Verben nicht in jedem Fall allgemeine Lautgesetze ableiten lassen. Doch ein Gegen satz zwischen *’, *¨ und h, Ì im Altassyrischen ist hinreichend klar und kann kaum als Sonderentwicklung bei schwachen Verben erklärt werden. 39 Blake 1945. 40 Kouwenberg 2006: 151. Kouwenberg scheint auf Blakes Argument selbst nicht allzu sehr zu vertrauen, jedenfalls diskutiert und gebraucht er es kaum. Kurz nach der zitierten Stelle schreibt Kouwenberg überdies: „In intervocalic position E colouring doubtless created a palatal glide“ (Kouwenberg 2006: 151). Gemeint ist E Färbung vor Ì, obwohl Blake Ursache und Wirkung gerade umgekehrt sieht. 41 Vgl. *baitum > bitum „Haus“, *äinum > inum „Auge“ etc. Hingegen *raÌmum > remum „Erbarmen“. Ferner *jadum > jidum (reichsakkadisch) > idum „Seite“ und *Ìamum > emum „Schwiegervater“ etc.

9


Die E-Färbung ist daher besser von dem Ort her zu erklären, an dem die beiden Pharyngale ursprünglich artikuliert wurden, als durch eine hypothetische und zugleich weite Verschiebung eines dieser Laute. Ein Gleitlaut mag dabei als Folge der Vokalfärbung an manchen Stellen hinterher entstanden sein. Außerdem kann man, ausgehend von Kouwenbergs Beobachtung, dass äin mit dem Stimmabsatz zusammenfiel, argumentieren dass deshalb Verben, die ursprünglich h oder Ì enthielten, einem anderen Paradigma angeglichen wurden. Das wäre eine Verteilung auf ein bereits vorgegebenes System entsprechend dem Kriterium: Unterbrechung des Luftstromes oder keine Unterbrechung. Die Palatalisierung von Ì steht auch im Widerspruch zum orthographischen Befund der reichsakkadischen Texte. Wegen des bereits beginnenden Übergangs a > E (ä) wäre hier j für Ì insbesondere in Gegenwart von a zu erwarten. Die anschließende Tabelle gibt einen Überblick über den Zeichengebrauch:42 ’V

A

E

Ì

Ù, Ú

¨V

A, Á

E

Ì

Ú

hV

A, Á, É

-

-

-

ÌV

É

E

E

-

jV

Ì+A

E

I, E?

U (Ú)

Das Fragezeichen nach E für [ji] stammt von Hasselbach. Ú für [ju] wurde vom Autor eingeklammert, da er es für wahrscheinlich hält, dass die wenigen Texte, die das Präfix ju- als ú- schreiben, den beginnenden Lautwandel [ju-] > [u-] dokumentieren, der schon in der Ur III-Zeit ganz vollzogen ist.43 Da dieser Wandel offenbar auch die Lesung des Zeichens I von ji zu i verändert, ist es möglich, dass der beginnende Abfall von j zuerst beim Präfix [ju-] zu beobachten ist.44 42 Cf. Hasselbach 2005: 34. E für *hi wurde als unsicher ausgelassen. Vgl. ibid. 79. Nach ibid. 81 wurde E für *Ìe ergänzt (bei Hasselbach wohl ausgelassen, weil Ì hier reichsakkadisch wahrscheinlich bereits ausgefallen ist). 43 In Ur III Texten gibt es keine für [ji], [ju] reservierte Zeichen mehr. Dazu Hilgert 2002: 120 1, cf. GAG3 §22c*. 44 Die spätere Festlegung von I auf [i] beruht natürlich auch auf der Beobach tung, dass I nicht nur beim Präfix, sondern auch an anderen Stellen für [i] gebraucht wird und älteres ì hier weitgehend verdrängt. Doch das mag mit Verzögerung gesche hen sein oder es ist bei den wenigen Texten, bei denen der Abfall von anlautendem j eintritt, nur zufällig nicht bezeugt, was gut möglich ist, da i am Silbenanfang weit

10


Unsicher ist U für *Ìu weil der einzige Beleg li-ip-te-u-ma (ptÌ) sehr wahrscheinlich als liptajuma „sie sollen öffnen“ mit einem nach dem Abfall von Ì und Färbung des Vokals entstandenen Gleitlaut zu interpretieren ist.45 Die Tabelle zeigt, dass das Reichsakkadische j relativ klar von allen übrigen Zeilen trennt. Die einzige Ausnahme ist das zugleich am wenigsten spezifische Zeichen E. Man könnte einwenden, dass Kouwenberg die Palatalisierung nur nach Vokal zwingend verlangt hatte. Doch dann brauchen wir einen anderen sprachlichen Mechanismus, um Ìa > E (ä) zu erklären und wenn ein solcher existiert, gibt es keinen Grund mehr, die Palatalisierung nach einem Vokal anzunehmen. Das etwas ältere Syllabar der Ebla-Texte ist weniger systematisch aufgebaut als das reichsakkadische Syllabar. Davon heben sich die lexikalischen Texte mit ihrer mehr einheitlichen Schreibweise ab. Außerdem ist die Interpretation in den lexikalischen Listen zuverlässiger als bei einsprachigen Texten oder gar in den Eigennamen. Daher scheint es geraten das Syllabar der lexikalischen Texte zum Vergleich heranzuziehen.46 ’V

A, Ì

-

Ì

Ù

‘V

A

-

Ì

Ù

hV

É, A

-

I

U9

ÌV

É, Ì

-

I, Ì

U9

jV

A

-

I

U9, U4

Insgesamt zeigt die Tabelle die von Kouwenberg genannte Übereinstimmung in den letzten beiden Spalten. Doch die erste Spalte seltener ist als das Verbalpräfix ji bzw. i . Dazu kommen orthographische Besonder heiten, wie die noch lange übliche Schreibung von ili „mein Gott“ als NI.NI = ì lí. 45 Cf. Hasselbach 2005: 81 und Kouwenberg 2006: 151 mit der gleichen Inter pretation. 46 Die Tabelle folgt der Zusammenstellung bei Rubio 2006: 115 16. Auf einige unsichere Lesungen wurde verzichtet. Der Lautwert ji für I ist auch lexikalisch belegt und zwar in dem Wort jidum, das in Ebla offenbar nicht wie sonst im Akkadischen für „Arm“, „Seite“ steht, sondern (noch) wie westsemitisch jad, „Hand“ bedeutet. Vgl. die Belege bei Pettinato 1982 unter VE 515, VE 531, VE 557, VE 012 und kontrastierend ì sa tù (isatum) „Feuer“, VE 783. Für die interessante Gleichung VE 802: den líl = i li lu (besser: ji li lu) vgl. den möglichen Gebrauch von EN zur Schreibung von [je/in], cf. Sommerfeld 1999: 20 Anm. 29. Auf die strittige Etymo logie dieses Namens kann hier nicht eingegangen werden. Siehe zuletzt Sommerfeld 2006: 74. Cf. auch VE 799a/b dEN.ZI/ZU sú i nu (= sú ji nu) für den semitischen Namen des Mondgottes (ganz genau: [tsujin] oder [tsujain] > [tsujin]?). 11


ergibt das umgekehrte Ergebnis. Insbesondere das Zeichen, welches h und Ì mit der größten Sicherheit anzeigt, É wird weder in Ebla noch reichsakkadisch auch für jV gebraucht. Wenn die Schreibung von *Ìi und *Ìu eine Palatalisierung anzeigen sollte, so würde die Orthographie dies für h sogar noch besser belegen. Dies steht im Widerspruch dazu, dass h den Lautwandel a > E normalerweise nicht auslöst. Kouwenbergs Untersuchung zeigt, dass wir insbesondere den Gegensatz zwischen hartem und weichem Ansatz auch bei der Wahl der Zeichen im Eblaitischen und älteren Akkadischen als einen Aspekt unter mehreren berücksichtigen müssen. Tendenzen bei der Zeichenwahl können anzeigen, dass für semitische Muttersprachler eine Ähnlichkeit in der Aussprache bestand. Da aber kein System mit einer eindeutigen Zuordnung von Phonem und Graphem geschaffen wurde, sind Schlüsse nur in einer Richtung einigermaßen zuverlässig: Wenn die Grapheme zwei aus der Etymologie erschlossene Phoneme eindeutig trennen, dann ist auch eine phonetische Unterscheidung sehr wahrscheinlich. Unsere Kritik betrifft Kouwenbergs Feststellungen nur in einem für seine eigene Analyse unwichtigen Nebenpunkt, den er selbst nicht weiter ausdiskutiert. Dass den Assyriologen bisher nichts besseres als die leicht angreifbare Idee von Blake eingefallen ist, zeigt nebenbei auch, wie schwierig es ist, den plötzlichen Verlust zahlreiches semitischer Phoneme als innersprachliche Entwicklung zu deuten. Das Akkadische erfährt einen Kahlschlag seiner Phoneme außerdem nicht nur bei den Gutturalen. Reichsakkadisch beginnt auch der Verlust des im Arabischen als À erscheinenden Phonems47 und des Gleitlautes j außer zwischen Vokalen.48 Dass so viele, in unterschiedlicher Weise und an verschiedenen Stellen artikulierte Phoneme fast gleichzeitig verschwinden, ist am besten durch äußeren Einfluss zu erklären. c) Hebräisch Wie im Akkadischen geraten die hinteren Konsonanten ebenfalls unter Druck. Dies lässt sich u. a. an der Regel ablesen, dass sie nicht verdoppelt werden dürfen und an zahlreichen Verwechslungen, insbesondere in den Qumran-Texten. Doch hat das Hebräische diesem Druck bis zu einem gewissen Grad widerstanden. Dabei tritt eine eigenartige Erscheinung auf, das PataÌ furtivum und die Îatef-Vokale. 47 48

Hasselbach 2005: 143. Siehe oben. 12


Beschränken wir uns der Einfachheit halber zunächst auf das PataÌ furtivum. Schließen die Pharyngale ‫ע‬, ‫ ח‬oder der Laryngal ‫ה‬ eine Silbe, die einen Vokal anders als a, a enthält, so wird ein kurzes a eingeschoben. Also *ruÌ > ruaÌ „Geist“ etc. Erklärt wird diese Erscheinung als phonetischer Zwang oder als Versuch einer deutlicheren Aussprache.49 Aber warum kommen diese Vokale dann im Arabischen nicht vor und warum scheinen die Pharyngale im Hebräischen a zu stärken, während sie im Akkadischen meist a > e herbeiführen? Letzteres wird besonders offensichtlich, wenn man bedenkt, dass bei der Aufspaltung von Doppelkonsonanz am Wortende (Segolierung) vor ‫ ע‬und ‫ ח‬und auch häufig nach diesen Pharyngalen im Hebräischen statt üblichem Segol (æ) PataÌ (a) eintritt, während im Akkadischen die gleichen Pharyngale in einer Silbe normalerweise a > e bewirken. Rekapitulieren wir zunächst die Bildung des Vokals a. Dieser Vokal ist einerseits durch eine maximale Öffnung des Mundes gekennzeichnet, andererseits durch einen Rückzug der Zunge, die damit hinten eine Enge bildet. Diese Enge hat der Vokal a mit den pharyngalen Konsonanten gemeinsam. Bei ‫ ע‬ist es ein stimmhafter Verschluss, bei ‫ ח‬eine stimmlose Friktion. Stellen wir uns nun einen Sprecher vor, der an die Aussprache von Pharyngalen nicht gewöhnt ist und folglich auch die Pharynxmuskeln bis auf die Hilfestellung des am Zungenbein ansetzenden Constrictor pharyngis medius beim Zurückziehen der Zunge zum Sprechen nicht einsetzt. Will dieser einen Pharyngal bilden, so wird er dies vor allem oder ausschließlich mit der Rückverlagerung der Zunge tun. Dabei ist anzunehmen, dass er dies im Zusammenspiel der Muskeln tut, wie er es für die Rückverlagerung der Zunge gewohnt ist, d. h. bei der Bildung des Vokals a. Bei der Bildung des Pharyngals durchläuft er also die Mundstellung für a. Vergleichen lässt sich die Übernahme von ungewohnten Lauten aus einer fremden Sprache. Sehr häufig wird versucht, den Laut mit Hilfe des gewohnten Inventars irgendwie nachzustellen. Z. B. wird langes französisches ü im Englischen häufig zu [iu] oder [yu]. Viele Sprecher des Deutschen geben französisches balcon [balkõ] durch [balko∞] wieder. D. h. obwohl das Schriftbild im Deutschen eine Aussprache [balkon] nahe legt und ∞ im Deutschen nur als sekundärer Laut aus ng existiert, wird versucht, den hinteren Nasalvokal õ durch den entsprechenden nichtnasalen Vokal und einen hinteren Nasal nachzubilden. 49 Z. B.: „ein willkürlicher Gleitlaut (…) vgl. in deutschen Mundarten iach für ich, Buech für Buch.“ Jenni 1977: 36.

13


Schließt der Pharyngal direkt an einen Vokal an, so sind ein Luftstrom von der Lunge und schwingende Stimmbänder vorhanden, weswegen tatsächlich kurz a hörbar wird. Steht der Pharyngal hingegen am Silbenanfang, so ist die Bildungsphase nicht hörbar. Das Lösen der hinteren Enge mag leichter fallen als die Bildung der Enge, weswegen vor Vokal kein PataÌ furtivum zu hören ist. Sieht man in PataÌ und dann auch in PataÌ furtivum kein a, sondern eher ein ä, so lässt sich dies durch Abschwächung der für die Bildung einer hinteren Enge nicht gebrauchten, aber für a wichtigen vorderen Öffnung erklären.50Wir wären dann in einer Situation ähnlich dem Akkadischen, nur dass die im Akkadischen anzunehmende weitere Vorverlagerung zu einem in der Mitte zwischen a und i stehenden e durch Segol blockiert war. Segol selbst wird von ∑ere am weiteren Ausweichen nach vorne gehindert. Bei der Segolierung, bei der es keinen Grund gibt, den Vokal anders als passend zu wählen, kann man im Prinzip den gleichen Effekt wie beim PataÌ furtivum beobachten, wobei auch hier die Wirkung vor dem Pharyngal größer ist. Einwenden könnte man, dass ‫ ה‬eigentlich kein Pharyngal ist. Doch es ist möglich, dass ‫ ה‬leicht pharyngalisiert gesprochen wurde, insbesondere in einem sprachlichen Umfeld, in dem dieser Konsonant nicht üblich war, was die Tendenz zu einer Art Betonung gefördert haben mag. Vielleicht erklären sich so auch einige der gelegentlichen Unregelmäßigkeiten beim Landwandel a > e im Akkadischen.51 Weil bei den Îatef-Vokalen nicht sogleich ein Vokal folgt, der praktisch ein Loslassen der Zungenstellung bewirkt, tritt hier die Begünstigung eines Vokals nahe a bei den Pharyngalen und h wohl ebenfalls ein. Wohl weil es mit Schwa praktisch kaum noch wahrnehmbar wäre, wurde auch ‫ א‬einbezogen. Um das Auftreten von QameÒ Ìa†uf unter den Îa†ef-Vokalen in unserem Modell zu erklären, muss man annehmen, dass es aus Reduktion von u (oder o) entstanden ist. Von u nach a steigen beide Formanten kontinuierlich an. Eine hintere Enge hebt beide Formanten und verschiebt einen hinteren Vokal also in Richtung a. Kommt hinzu, dass durch die Flüchtigkeit des Lautes und den angenommenen Gebrauch der Zunge zur Bildung der hinteren Enge, die Mundstellung von u wohl auch insgesamt nicht mehr ganz eingenommen werden kann. Ein kurzes å könnte sehr wohl das Ergebnis dieser Veränderung sein. 50

Ich lasse die Frage offen, denn die Nichtverwendung von QameÒ lässt sich auch damit erklären, dass QameÒ in unbetonten, geschlossenen Silben å und nicht a vertrat. 51 Für solche Unregelmäßigkeiten vgl. Kogan 2001. 14


Fassen wir unsere Betrachtung noch einmal etwas anders zusammen. Wir haben die Effekte im Akkadischen grob gesprochen vom Konsonanten her zum Vokal betrachtet und im Hebräischen vom Vokal her zum Konsonanten. Beide Betrachtungsweisen passen zu den jeweiligen Ergebnissen. Im Akkadischen fallen die Pharyngale aus, hinterlassen aber einen Abdruck im Vokalismus, nämlich den Lautwandel a > e. Im Hebräischen dient eine Veränderung im Vokalismus offenbar dazu, die betreffenden Konsonanten zu erhalten. d) Exkurs zu den Silbenzeichen Das System der akkadischen Silbenzeichen beruhte fast ausschließlich auf aus dem Sumerischen geborgten Lesungen. Diese mögen beim Gebrauch in akkadischen Worten gegenüber der rein sumerischen Aussprache modifiziert worden sein. Doch auf diesem Wege wurde keine phonologisch oder gar phonetisch eindeutige Darstellung des Akkadischen erreicht. Daher können wir nicht sicher erwarten, dass die oben angenommenen Verschiebungen der Vokale auch durch eine Veränderung in der Schrift eindeutig angezeigt wurden. Z.B. mag das Zeichen E altbabylonisch ein zwischen a und i liegendes [e] bezeichnen. Es könnte aber reichsakkadisch noch durchaus ein nahe an a liegendes [ä] bezeichnet haben. Aufgrund der Präfixharmonie des Sumerischen hat Arno Poebel für das Altsumerische 6 Vokale postuliert, darunter ein geschlossenes e „similar to the (first) e in German sehen, reden, mehr“52, sowie ein offenes e entsprechend deutschem ä. Poebels Ansatz ist nicht ohne Alternative aber auch nicht unwahrscheinlich.53 Wegen der Statistik der Vokale würde der Autor einige von Poebels Vokalen eher etwas verschieben, was uns hier aber nicht zu interessieren braucht.54 Nicht betroffen hiervon wäre Poebels offenes e (ä). Eric Smith55 hat mittlerweile vorgeschlagen, die Präfixharmonie als [-ATR] mit 7 Vokalen zu beschreiben, d. h. auf Pharyngalisierung zurückzuführen, was zu unserem Ansatz, den Ersatz von Pharyngalen durch eine Veränderung der Vokale im Akkadischen auf sumerische Sprechgewohnheiten zurückzuführen, ebenfalls passen würde.

52

Poebel 1931: 3. Dazu ausführlich Keetman 2005. 54 Vorläufig Keetman 2005. Zwei weitere Artikel des Autors hierzu sind in Vor bereitung, einer in WdO. 55 Smith 2007. 53

15


Was die Silbenzeichen betrifft, so lässt sich vermuten, dass zunächst das Zeichen É (sum. „Haus“), weil es als [ä] gesprochen wurde,56 aber im Sumerischen nicht als Silbenzeichen benutzt wurde, ausgesondert wurde, um Pharyngale im Semitischen zu schreiben. Zu vergleichen wäre PI = be6 zur Schreibung von wa, wi, wu. Wie Walter Sommerfeld herausgefunden hat, unterscheidet die reichsakkadische Orthographie die Vokale e und i deutlicher als die Keilschrift späterer Epochen.57 Von den entsprechenden Zeichenpaaren finden einige, nämlich e/ì, bé/bí, me/mi, sè/si auch in der sumerischen Präfixharmonie Verwendung und bezeichnen dort den Unterschied zwischen offenem e ( = [ä]) und sumerischem i.58 Mangels anderer Alternativen müssen diese Zeichen im Akkadischen auch e < ai übernehmen.59 Später werden die Kä-Zeichen entweder ganz aufgegeben, zu Ke-Zeichen verschoben oder genutzt, um eine genauere Abgrenzung der Konsonanten zu erreichen. Vermutlich wegen der Stellung von sumerischem i hinter akkadischem i aber über späterem akkadischem e werden nun sumerische i-Zeichen verstärkt auch für akkadisches e verwandt, wodurch die bekannte Ungenauigkeit bei der Wiedergabe von e und i in der Keilschrift entstanden ist. BIBLIOGRAPHIE Behrens, H. und H. Steible. 1983. Glossar zu den altsumerischen Bau und Weihinschriften. (FAOS 6. Wiesbaden) Blake, F.R. 1945, ‘Studies in Semitic Grammar III’, Journal of the American Oriental Society 65, 111 16 Cross, F.M. 2003. ‘Some Problems in Old Hebrew Orthography with special Attention to the Third Person Masculine Singular Suffix on Plural Nouns [ âw]’in F.M. Cross (ed.), Leaves from an Epigrapher’s Notebook (Winona Lake). 351 6 Gelb, Ignace, J. 1961. Old Akkadian Writing and Grammar. (2. Aufl. Chicago) Gelb, I.J., Steinkeller, P. und R. Whiting. 1989. Earliest Land Tenure Systems in the Near East: Ancient Kudurrus. (OIP 104. Chicago) Hasselbach, R. 2005. Sargonic Akkadian: A Historical and Comparative Study of the Syllabic Texts. (Wiesbaden) Hilgert, M. 2002. Akkadisch in der Ur III Zeit. (IMGULA 5. Münster) Jenni, E. 1977. Lehrbuch der hebräischen Sprache des Alten Testaments. (Basel und Stuttgart) 56

Trotz der von sumerisch é gal „Palast“ abgeleiteten westsemitischen Wörter mit h kannte das Sumerische sehr wahrscheinlich kein h oder einen ähnlichen Laut. Cf. Keetman 2006. 57 Sommerfeld 1999: 18 19; 2003: 572. 58 So Poebel 1931, Keetman 2005 und etwas anders formuliert auch Smith 2007. 59 Jedenfalls beim Plural m. im casus obliquus. Beispiele bei Sommerfeld 1999: 19. 16


Keetman, J. 2004. ‚Der Verlust der „Kehllaute“ im Akkadischen und der Lautwan del a > e’, Altorientalische Forschungen 31, 5 14 2005. ‚Die altsumerische Vokalharmonie und die Vokale des Sumerischen’. Journal of Cuneiform Studies 57, 1 16 2006. ‚Gab es ein h im Sumerischen?’ Babel und Bibel 3, 9 30 Kienast, B und K. Volk. 1995. Die sumerischen und akkadischen Briefe. (FAOS 19. Stuttgart) Kogan, L. 2001. ‘*g in Akkadian'. Ugarit Forschungen 33, 263 7. (Ergänzungen in UF 34) Kogan, L. und K. Markina 2006, Rezension zu Hasselbach 2005. Babel und Bibel 3, 555 88. Kouwenberg, N.J.C. 2006. ‘The Proto Semitic Gutturals in Old Assyrian’, in G. Deutscher und N.J.C. Kouwenberg (Hrsg.) The Akkadian Language in its Semitic Context, Studies in the Akkadian of the Third and Second Millennium BC (Leiden). 150 76 Michalowski, P. 1993. Letters from Early Mesopotamia, translated by Piotr Michalowski, edited by Erica Reiner. (Atlanta, Georgia) Pettinato, G. 1982. Teste lessicali bilingui della bibliotheca L. 2769. (MEE 4, Napoli) Poebel, A. 1931. The Sumerian Prefix Forms e and i in the Time of the Earlier Princes of Lagas. (AS 2, Chicago) Rubio, G. 2006. ‘Eblaite, Akkadian and Eastsemitic’, in G. Deutscher und N.J.C. Kouwenberg (Hrsg.) The Akkadian Language in its Semitic Context, Studies in the Akkadian of the Third and Second Millennium BC (Leiden), 110 39 Smith, E. 2007. ‘[ ATR] Harmony and the Vowel Inventory of Sumerian’. Journal of Cuneiform Studies 59, 19 39 Soden, W. von 1995. Grundriss der Akkadischen Grammatik. (3. Auflage. AnOr 33. Rom.) (= GAG3) Sommerfeld, W. 1999. Die Texte der Akkade Zeit. I. Das Diyala Gebiet: Tutub. (Münster) 2003. ‘Bemerkungen zur Dialektgliedrung, Altakkadisch, Assyrisch und Babylonisch’, in G. Selz (Hrsg.) Festschrift für B. Kienast. (AOAT 274, Mün ster). 569 86 2006. ‘Die ältesten semitischen Sprachzeugnisse eine kritische Bestandsauf nahme’, in G. Deutscher und N.J.C. Kouwenberg (Hrsg.) The Akkadian Language in its Semitic Context, Studies in the Akkadian of the Third and Second Millennium BC (Leiden). 30 75 Steinkeller, P. 2004. ‘On the Writing of belum in Sargonic and Earlier Sources’. Nouvelles Assyriologiques Brèves et Utilitaires 2004/1, 12 14 Tropper, J. 1999. ‘Zur Etymologie von Akkadisch sukênu/suhehhunu’. Welt des Orients 30, 91 4 Vito, R. di. 1993. Studies in Third Millennium Sumerian and Akkadian Personal Names. The Designation and Conception of the Personal God. (Studia Pohl SM 16. Rom)

17

THE DEVELOPMENT OF THE SEMITIC DEFINITE ARTICLE: A SYNTACTIC APPROACH* NA}AMA PAT-EL HARVARD UNIVERSITY

Abstract This paper aims at explaining the development of the Semitic definite article through an examination of its attested syntactical features. The paper will try to show that the original function of the definite article was not to mark definiteness, that it was first attached to the at tribute, not the noun, and that only later was it transferred to the noun and interpreted as an article.

1. Introduction Several important studies have been published lately concerning the historical development and the function of the West-Semitic definite article. Naturally, this topic has attracted much scholarly attention in the past, and some of its aspects are still a matter of great contention. There is almost no dispute in the field that the origin of the definite article is an attributive demonstrative.1 The exact identification of this * Several people have commented on earlier versions of this paper. I would like especially to thank John Huehnergard for his continued support and interest and for his many thought provoking comments. I would also wish to thank Wolfhart Heinrichs, Geoffrey Khan and Aaron Rubin for their helpful comments. Any errors are mine alone. Abbreviations used in this paper are: Adj adjective; AdN adnominal; Akk. Akkadian; Amh. Amharic; Arb. Arabic; Arm. Aramaic; BH Biblical Hebrew; C Christian; CA Classical Arabic; Can. Canaanite; CS Central Semitic; Def. definite; Dem demonstrative; ES East Semitic; G G¢¨¢z; IE Indo European; Indef. indefinite; J Jewish; LateA Late Aramaic; M Muslim; MidA Middle Aramaic; MH Mishnaic Hebrew; N noun; Nor. Norwegian; OA Old Aramaic; OfA Official Aramaic; OSA Old South Arabian; Ph. Phoenician; Poss. = possessive suffix; PS = proto Semitic; P WS proto West Semitic; TO Targum Onqelos; Ug. Ugaritic. 1 In this paper, the term ‘definite’ refers to the grammatical marker, and not to its pragmatic or semantic features, unless specifically noted, as the emphasis of the cur rent paper is syntax. The term ‘demonstrative’ is a deictic expression; ‘attributive/ 19

THE DEVELOPMENT OF THE SEMITIC DEFINITE ARTICLE

demonstrative is what essentially divides scholars (Rubin 2005: 72). In their most recent contributions to the discussion, Voigt (1998) and Tropper (2001) reconstructed *han for both Hebrew and Arabic but Rubin (2005: 72–6) argued for two separate demonstratives as the origin for the Arabic and Hebrew articles. According to Rubin, the CS article developed from two Semitic demonstratives *han- (Can., Arm., OSA) and *’ul (Arb.). The doubling caused by attaching the article is a result of an assimilation of the final n/l to a following consonant. Typologically, most languages develop articles from attributive demonstratives (Heine and Kuteva 2002: 109–11) and this typology is the basis for most of the work done on the Semitic article.2 However, there are several substantial discrepancies with this general approach and its application to the Semitic languages.3 The first problem is the position of attributive demonstratives relative to their head noun. Most Semitic languages show a N-Dem word order except G., OSA and occasionally Arab. This position is in agreement with the regular position of attributes in Semitic: adjectives, relative clauses and other adnominal modifiers normally follow their head noun. Note the following examples which portray the typical word order with attributive demonstratives in the Semitic languages:4 Akk.: G.: OSA: BH: Ph.:

kaspam anniam ‘this silver (acc.)’ ba-za hagar ‘in this city’ 5 dn ’l-n-hn ‘these two gods’ (R 2923/7) hay-yom haz-ze (Gen. 17:26) ‘this day’ ml’kt z’ ‘this work’ (KAI 10:14)

pronominal demonstrative’ is a pronominal paradigm which is used to mark deixis (Diessel 1999: 2). Attributive demonstratives occur with a coreferential noun, and pronominal demonstratives may syntactically substitute a noun. This paradigm dis tinguishes (at least) distal and proximal and has the same morphological distinction as any other adjective in Semitic. 2 This is especially evident in Rubin (2005). Zaborski also maintains that the demonstrative origin of the article ‘is the only reasonable solution since usually defi nite articles go back to demonstratives’ (2000: 25). 3 Lambdin is the only scholar I am aware of who objected to the assumption that a demonstrative is the origin of the Semitic article. He nicknamed it ‘the particle theory’ (1971: 315, n. 1). Lambdin recognized and pointed to many of the problems I will discuss below, but his worries were largely ignored by later scholars. While I do not agree with parts of the solution Lambdin offers, his observations are insightful and astute. It is regrettable that this important article, despite its flaws, was shunned. 4 The underlined examples mark languages where the demonstrative precedes the head noun. 5 Note also that the relative pronoun, which is identical to the near demonstra tive, stands after the noun (Beeston 1984:41): brktn ∂t ¨rn ‘the cistern of the castle’ (Höfner 1943: 43). 20


spr hnd ‘this book (m.acc.)’ (2.19, 9); ’alpm sswm hnd ‘these 2000 horses (m.pl)' (2.33, 32); ml’akty hnd ‘this mission (f.) of mine’ (2.33, 35) Arm.: dmwt’ z’t ‘this statue’ (Fekh. 15)6 Arb.: hada l-malik / l-malik ha∂a ‘this king’ zayd ha∂a ‘this Zeid’ ïbadi ha’ula’i ‘these servants of mine’ Ug.:

In Arab. Dem-N is permissible only when the noun is marked with the definite article, which is a late feature, at least later than proto-CS. In any case, the Dem-N word order is conditioned by the syntax of the noun.7 G. is under heavy influence of non-Semitic languages, about which we know very little.8 Thus, since the preferred word order in both ES and WS is N-Dem, it should be considered both a PS and a PWS feature, while the article is only common CS (Rubin 2005: 65). This is a major weakness in the assumption that the article developed from an attributive demonstrative. Note also, that according to Rubin’s reconstruction, the Canaanite and Aramaic articles both go back to the same demonstrative (*han), yet the article is post-positive in Aramaic and pre-positive in Canaanite.9 While the syntax of the article is consistent in all the CS languages, which no doubt points to a similarity in the development of their article, the differences in its position cannot be ignored.10 6

This word order changed by LateA and later to become Dem. N. See Pat El (forthcoming). 7 The situation in OSA is hard to evaluate. There are not many examples of the attributive demonstrative and the information regarding their behaviour in different syntactic environments is therefore partial. 8 The order of N modifier in G. is variable, that is adjectives, genitive and relative clauses may be positioned before, or after their head noun; however, in modern Ethio Semitic, most languages exhibit modifier N word order, although some variation is possible (Little 1974: 79, fig. 7). This is a strong indication that there has been sub stantial shift of nominal modifiers from post position to pre position. Some elements are more likely to gain a fixed position than others. Already in G. quantifying adjec tives, such as b¢zuÌ ‘much’, and ordinal numbers are regularly pre posed: sab¢äwit hagar ‘the seventh city’. Thus, the pre position of the demonstrative in G. is a result of a process of change, rather than an indication of the original situation. 9 A similar point was made by Lambdin (1971: 316). Firmage’s rejection (2002: 37, n. 14) of Lambdin in this regard should not be accepted. Lambdin is correct in noting that demonstratives appear after the noun in Semitic. Firmage uses Arabic as evidence that Lambdin is wrong, but Arabic is a clear innovation, and a late one at that, and even there we find the demonstrative regularly after the noun. 10 It is well known by now that the order of morphemes does not necessarily reflect their word order prior to grammaticalization (Harris and Campbell 1995: 199 200); however, in this case, it seems that the development of the article is much later than the stabilization of the demonstrative word order. 21


The second problem is the exact function of the Semitic demonstrative pronoun *ha(-n). First, we should note that *ha(-n) does not occur as an independent demonstrative in Semitic,11 but rather is always prefixed to a demonstrative as in the following: Akk. Ug. BH MH Arb. Arm. Amh.

annûm < *hanni12 hnd haz-ze, hallaze halla ha∂a hdyn, hlyn13 ¢nnäzzih (cf. bä-zzih ‘in this’)

Hasselbach (2007: 21)14 argues that the particle *ha and its extension *ha-n are originally adnominal, basing her conclusions on their tendency to be attached to attributive demonstratives when two separate paradigms (adnominal and pronominal) exist. Some dialects of Aramaic show this clearly. In other languages, forms with the prefix *ha(-n) are primarily or exclusively attributive: TO. Ug. BH

ha-∂a15 (attributive): ’arä ha∂a (Gen. 12:7) ‘this land’. da (pronominal): da ’arä (Num. 34:2) ‘this is the land’. hn-d and hndn are always used attributively: Ìp† hndn ‘his Ìup†usoldier’ (KTU 2.72).16 ha-’is hallaze (Gen. 24:65) ‘this man’; ha-’is haz-ze (Gen. 24:58) ‘this man’.17

11

Once more, Lambdin is right to note that ha is not an independent demonstra tive in Hebrew and Phoenician (1971: 315). Here too, we must reject Firmage’s dis missal of Lambdin who asserts that we simply do not have enough early material in these two languages to make such a claim (2002: 37, n. 14). We do indeed have comparative evidence which clearly corroborates Lambdin’s claim. 12 Lieberman (1986: 591) argues that this form is constructed out of several ele ments: ’an , the first element in independent nominative pronouns (*’an a, ’an taˆ, ’an ti etc.) and ni, an Afroasiatic pronoun. For Hasselbach (2007: 25) *han is a prefix, while Rubin (2005: 76) claims that the origin of this form is a nisbe *hanni, which was reduced when cliticized. All these scholars agree that the Akkadian form is not an inflected *han, but either a derivation of it or a complex form with a prefix *han. 13 Starting from MidA (Cook 1992: 10). See more below. 14 I would like to thank Rebecca Hasselbach for making her paper available to me pre publication. 15 Even if we assume that the Targumic Aramaic ’arä ha ∂a is an imitation of the Hebrew ha ’areÒ haz zo(’)t, it still behoves us to note that while the Hebrew definite article haC is represented by the Aramaic definite article a, the Aramaic translator clearly identified haC before a demonstrative in Hebrew with ha in Aramaic. 16 Other forms (hndt, hnk, hnkt) appear in unclear contexts. 17 In MH the head noun is not necessarily marked with an article: s¢losa d¢ßarim hallalu ‘these three things’ (Av. 5:19). See discussion in section 3.4. 22


ze seper tol¢∂o† ’a∂ am (Gen. 5:1) ‘this is the book of the history of man’ Akk. (*han >) annûm is primarily used attributively, though not exclusively.

Given that *ha(-n) is prefixed to attributive elements, it is strange that in Aram. and OSA, *ha(-n) is a suffix.18 Hasselbach also notes that these forms’ inflection, if it exists at all, is secondary (Hasselbach 2007: 21). If indeed these forms were not inflected, it is hard to regard them as pronominal. While the phonological arguments for reconstructing a form *ha(-n) as the underlying form of the article are solid, there are some indications that *ha(-n) could not have been a demonstrative. Demonstratives by definition mark degrees of proximity (Diessel 2005: 170–1). Yet the element *ha(-n) not only lacks such a distinction19, it may also be cliticized to various elements indiscriminately:20 the proximal demonstrative in CA (ha∂a) and Akk. (anni-); the distal demonstrative in BH (hallaze); to both in TO (ha-hu, ha-∂a), Syriac (hanna, haw < ha-w), and Mand. (hahu, hax).21 A slightly less perplexing problem is the lack of inflection. The definite article in CS does not exhibit gender-number distinctions.22 This is quite puzzling since even languages which have lost gender distinctions in the nominal paradigm (e.g. Dutch23) still maintain gender dis18

Even Hasselbach who supports Rubin’s analysis notes that ‘already in early Se mitic there existed the possibility to mark the demonstrative pronouns as adnominal by affixing the element *ha(n)’ (2007: 21). With no evidence to the contrary, ‘affix ing’ should be understood as prefixing. 19 Hasselbach (2007: 21) argues that n is probably a near deixis element in Se mitic, but ha itself does not seem to have any such attribute. 20 This lack of proximity distinction is found with the relative determinative pro noun as well; the determinative is, of course, not a demonstrative. 21 In addition, from the very beginning, the article and the demonstrative co occur: Arb. ha∂a al malik ‘this king’, BH ha ’ama haz zo(’)t ‘this maidservant’ (Gen. 21:10). In a process of grammaticalization of Dem>article, there is usually a period where the presence of the article excludes the demonstrative; this did not happen in CS. In most Indo European languages it is still the case, and the demonstrative and definite article are in complimentary distribution: *this the king. 22 Zaborski (2000) argues that the definite article was originally inflected for gen der and number (m. ha n, f. ha t, pl. ha l), but subsequent assimilation levelled the differences. However, there is no direct evidence that this is so and the entire premise lies heavily on the assertion that the article must have its origin in a demonstrative. 23 The adnominal categories in the Germanic languages were more gender dis tinctive than the nominals. This is especially evident in the article and demonstratives (Dekeyser 1980: 98). While in English the noun, the definite article and even the indefinite article (slightly later) lost all gender number case distinctions in Late Old English, the demonstrative maintained a basic number marked paradigm. In Dutch 23


tinctions on the article; in fact in many languages the rise of the article is assumed to be related to the collapse of the nominal system which leads to the article carrying the relevant nominal inflection.24 In Semitic, with its rich nominal morphology, there is no reason that the attributive demonstrative should lose its inflection as it becomes a definite article.25 Thus, it would be highly problematic to reconstruct the CS definite article from a demonstrative. The only way that a non-inflecting adnominal prefix *ha(-n) could have been the origin of the CS article is if it was attached to the attribute first, and only later to the noun by analogy, as it is quite unlikely that an ad-nominal prefix will be attached to a substantive. This point is crucial to the understanding of the syntax of the article and its later developments in Semitic. While many of the syntactical features of the article are shared, there are some native, and apparently old, patterns in each language which do not quite fit what we think we know about the article’s syntax. The importance of relics to historical reconstruction has been highlighted repeatedly by historical linguists dealing with morphology,26 but is where the distinctions were lost much later than English, the demonstrative and defi nite article are the only forms which maintain a distinction between genders (com mon de and neuter het). 24 Lyons (1999: 67) mentions only one inflected language (Hungarian) where the article is frozen, while the demonstrative and noun show inflection. However, Hun garian has fewer nominal features than Semitic (e.g. it has no gender). Note also that the innovated article in Amharic distinguishes gender: m.s.: l¢g u ‘the boy’ / bära w ‘the ox’; f.s: l¢g itwa ‘the girl’ / lam wa or lam itu or lam itwa ‘cow’; c.pl.: n¢gus occ u ‘kings’ (Leslau 1995: 155 6). The tendency of articles in world languages to carry more grammatical load than the noun is so strong that it was the basis of a theory that the innovation of articles is a way to ensure that the nominal phrase will carry the grammatical markers, even when the morphological expression is lost (Lyons 1999: 324). There are of course known exceptions, as English, where the noun distinguishes number but the article does not. 25 Compare, for example, the definite article in Central Neo Aramaic, which origi nated from pronouns, where distinctions between the genders were not levelled: ™ur. b u gab ano (Jastrow 1992, 6.10:8) On this side (m.); b i xaÒra†e (Jastrow 1992, 12.13:11) On this peg (f.). Typological studies have shown that attributive demonstratives are the last grammatical entities in the nominal system to lose inflec tion. Moravcsik’s proposed universal is: if the adnominal adjective agrees with the noun, so does the adnominal demonstrative (1997: 317). Nevertheless, it is possible that in the process of grammaticalization, a demonstrative may lose its inflection. The difficulty in Semitic is that even outside the ‘article’ position hn does not show inflec tion. 26 Hetzron (1976: 92ff.) labelled it ‘the principle of archaic heterogeneity’, ac cording to which the relatively most heterogeneous system must be the most archaic. Hetzron used this principle mostly in comparative historical linguistics to determine the relative archaism of one system over the other. This, however, may very well apply 24


equally valid in syntax.27 We should mention in this context Meillet’s note that we reconstruct on exceptions, not rules (Meillet 1954: 27). Therefore, these relics will be an important part of the following discussion, along with a reassessment of the syntactic rules and the explanations previously given to them. In the following discussion I will use the term ‘article’ to refer to the form, not the function, of the so-called definite article in CS. 2. The Syntax of the Definite Article The syntax of the definite article in CS is quite peculiar yet consistent across the CS languages. This common syntax was recently discussed by Huehnergard (2005: 184–6) as a part of his discussion of common CS features. Although Huehnergard does not believe the article to be a result of shared innovation,28 he does mention a list of common syntactical features that are found in Canaanite, Aramaic, Arabic and possibly OSA and thus may point to a regional phenomenon: the article appears on the last member of a construct; it does not occur with proper nouns or nouns with pronominal suffix; predicative adjectives are not marked with the article and attributive adjectives agree in definiteness with the noun (2005: 185). To Huehnergard’s list one should add that the definite article allows nominalization of adjectives: BH haÒ-Ò¢ïra (Gen. 19:38) ‘the young (sister)’.29 The fact that all languages exhibit exactly the same peculiar syntax should at least point to a common process. The similarities in syntax may help us figure out the origin of the pattern, by tracing back what triggered this behaviour. In order to do that we need to assume, as we do in morphological reconstruction, that the attested situation is a result of a process, not the original state. The common explanations suggested hitherto to account for the syntactic similarities are mostly unsatisfactory and do not account for to internal reconstruction as well. See for example Greenberg: ‘a highly irregular for mation which has withstood analogy must be very old’ (1957: 51). 27 Watkins noted that ‘[t]here are frequently more highly bound or restricted con structions in the syntax of a language, just as in its morphology, and it is the duty of the historian to look precisely here, even though such constructions may be of little interest to the descriptivist’ (1976: 312). 28 The lack of an article in Deir {Alla, Samalian and Ug. makes the assumption of it being a common CS feature improbable. 29 Many Semitic languages do not require any special apparatus to nominalize adjectives. However, all CS languages which use the article use it to nominalize much more than ø marking, if they use the latter at all. Note that with the exception of some Neo Aramaic dialects, no Semitic language uses the demonstrative, or any other independent pronoun, in order to nominalize attributes. 25


all of the data as a system. The redundancy of marking the article on both noun and adjective has been explained as a feature of nominal agreement (Rubin 2005: 83). The question of agreement is a complex issue which I shall not delve into here; suffice it to mention Arabic examples such as manzilatun qafrun ‘desolate abode’, nisa’un ÒaliÌatun ‘pious women’ (Reckendorf 1921: 71; Fischer 2001: 72) to note that agreement does not necessarily imply similarity of form. Moreover, other nominal clitics, such as the possessive suffixes, do not ‘move’ from the substantive to the adjective. Thus, the redundancy is still unresolved. The exclusion of nouns with possessive suffixes is usually compared to IE languages, where it is not normally common for the article to be attached to nouns with possessive pronouns (Rubin 2005: 82). While it is true that in many of the world’s language possessivity excludes the definite article, in many cases it is so because possessivity is marked through a possessive determiner. In French, mon/ma/mes are syntactically equivalent to le/la/les and carry similar morphology. This type of possessivity occupies the slot of the determiner, like quantifiers and attributive demonstratives, and therefore the article, being a determiner itself, is excluded.30 Since in Semitic the possessive suffixes are not a determiner (i.e. an article or a demonstrative pronoun), but rather suffixes, there is no apparent reason the article should be excluded. In addition, in all the Semitic languages the attributive demonstrative is permissible with nouns with suffixes, thus the article, if indeed from a demonstrative origin, should not be exluded either.31 For comparison, IE languages which exclude the article from nouns with possessive pronouns, exclude the attributive demonstrative as well. The genitive construction has been described and explained many a time, but for our purposes, it is important to note that it expresses the attributive relation syntactically, where the regens is the head and the rectum is the attribute (Goldenberg 1995: 3). The argument that the construct equals nouns with possessive suffixes (Brockelmann 1910: 30 Note that in Greek, where possession is expressed through pronouns with ad jectival declension (emos, eme, emon), the definite article, which originated from a demonstrative, is not excluded from nouns with possessive pronouns: ta hemetera biblia ‘our books’ and similarly, in Italian la mia casa ‘my house’. In this sense, French is a determiner genitive language, while Italian is an adjectival genitive language. See Haspelmath (1999: 233 4) where he attempts to explain the exclusion of articles with performance economy. 31 In Mehri, where the definite article is from a different source than the CS lan guage, the article does not exclude possessive suffixes. Note also that independent possessive articles in Semitic (Heb. sel , Aram. dil , Iraqi Arab. mal etc.) require the article: it talafon mal ak ‘your telephone’ (Rubin 2004: 333).

26


§178; Rubin 2005: 82), while true, does not by itself account for the fact that only the second element is marked with the definite article; pronouns and nouns do not have the same syntax nor the same level of definiteness (Khan 1984: 470). 3. Relics The following discussion focuses on native patterns in specific languages which do not exhibit the expected syntax but which seem to be very established. The last section of the discussion deals with nounadjective combinations where only the adjective is marked with an article. This pattern occurs in all languages which developed an article. 3.1 Arabic 3.1.1 Improper Annexation (al-’i∂afa gayr al-Ìaqiqiyya) Improper annexation is a special construct which contains an adjective as the regens and a noun as the rectum: Ìasanu l-waghi ‘beautiful (lit. pretty of appearance)’ (cf. BH y¢pa† to’ar). In this pattern, the rectum is normally marked with the definite article, yet the construct is not syntactically definite, i.e. an adjective modifying it would not be marked with an article and attributive demonstratives are excluded. In a regular type of annexation, whenever the rectum is marked with an article, the annexation is syntactically definite. However, in the improper annexation, if the construct is to be definite, the article must precede the regens, i.e. the adjective, as well as the rectum:32 ga¨du s-saäri ‘curly of hair, a person with curly hair’ al-ga¨du s-saäri ‘the curly of hair, the person with curly hair’

It is hard to understand why a definite rectum does not mark the entire construct as definite as it does with any normal construct pattern. 3.1.2 The Relative Pronoun alla∂i The Arabic relative pronoun (alla∂i, allati) is a combination of the definite article al-, a particle la and the PS relative-determinative pronoun *∂u (Akk. sa, BH zu, Aramaic zi). There is no doubt that the 32 This is not the case in Hebrew, where the improper annexation imitates the syntax of the regular annexation: s¢ßa¨ hap paro† y¢po† ham mar’e ‘the seven healthy cows’ (Gen. 41:4). Note, however, that most examples in the Bible are predicative and there are in effect very few examples of the improper annexation with a definite article.

27


relative-determinative pronoun ∂u was used in other dialects in the same way that alla∂i is used in CA.33 In Quranic and Classical Arabic, only definite antecedents may take alla∂i to introduce their relative clause. However, other types of Arabic use alla∂i with indefinite as well as definite nouns.34 This is an indication that the use of alla∂i was originally independent of nominal definiteness. This is further substantiated by comparative evidence, since the relative construction developed much earlier than the definite article and it exists in all the Semitic languages, whether they use an article or not. For example, the BH defunct relative pronoun zV occurs with both definite and indefinite heads: Indef.: hodiëni derek zu ’elek (Ps. 143:8) ‘let me know which road I should walk’. Def.: hinne ’elohenu ze qiwwinu lo (Isa. 25:9) ‘here is our lord for whom we have waited’.

Akk. and Ug., which did not develop an article, also use the relative pronoun indeterminately. It is therefore very likely that the preference of some dialects of Arabic to position alla∂i only before definite heads is secondary and was a result of differentiating two patterns which originally served the same function. 3.2 Hebrew 3.2.1 The Definite Article as a Subordinating Particle The definite article may appear on participles whose head noun is not definite: s¢pipon äle ’oraÌ han-noseÈ ïqq¢ße sus (Gen. 49:17) ‘a snake on a road biting a horse’s heels’. b¢-ya∂ mal’aÈim hab-ba’im Y¢rusalayim (Jer. 27:3) ‘by the hand of messengers coming to Jerusalem’.

The lack of article on the head noun causes great difficulty to Hebraists. Davidson states that in this pattern the head noun is definite, and even when it is not formally marked, ‘the preceding word is really definite’ (1902: 133, §99). 33 This basic pronoun is still in use in some Modern dialects, like the Zbala dialect in Morocco. See Brustad (2000: 109 10). 34 Hopkins (1984: 240, §288 n. 1) notes that the cases where an asyndetic relative clause follows an intrisincally or formally definite antecedent are much rarer than alla∂i clauses following an indefinite noun.

28


This pattern is very often termed ‘a relative pattern’ (Gesenius 1910: 440, §126a; Waltke and O’Connor 1990: 246, 621). This is of course not impossible, since participles are a nominal form with verbal functions. Nevertheless, the ‘relative participle’ gave rise to the assumption that the definite article is a relative pronoun. This idea was suggested previously by a number of scholars (Goshen-Gottstein 1946: 29; Wright 1967: II 232, §95f; Siloni 1995; Shlonsky 1997: 36; Firmage 2002: 35). Lately, it was used again by Gzella (2006), in his discussion of the Phoenician definite article, who claimed that the definite article was not used to mark definiteness, but was added to substantives that were already definite. According to him, the definite article’s original function must have been subordination: Allein die Annahme, dass sich Syntagmen wie das in KAI 14:9 [h’lnm hqdsm] aus einer alten relativischen Apposition wie *‘die Götter, die heilig [sind]’ durch Breviloquenz über ‘die Götter, die heiligen’ entwickelt haben, vereinigt die einzelnen linguistischen Fakten in einer plausiblen Zusammenschau. (Gzella 2006:11)

In BH, the position of the definite article before participles with an indefinite head noun points, according to Gzella, to a subordinating function: ‘Ganz deutlich verweist jedoch mitunter die Setzung des Artikels beim hebräischen Partizip auf eine relativische oder unterordnende Funktion…’ (Gzella 2006: 17). There is a lot of evidence against such an analysis: finite verbs were originally excluded from this pattern,35 it cannot be negated, the participle can predicate only the head noun,36 and lastly the article can be 35

See Gesenius (1910: 447i); Goshen Gottstein (1946: 29); Peretz (1967: 75, 108 9); Waltke and O’Connor (1990: 339d). There are some examples of finite verbs; however, all the examples with finite verbs occur in late books, except Josh. 10:24: q¢Òine ’an¢se ham milÌama he hal¢Èu(’) ’itto (Josh. 10: 24) ‘officers, men of war, who travelled with him’. This pattern with finite verbs disappeared all together in later stages of Hebrew, while the participle with the definite article is used from the earliest sources until Israeli Hebrew. In addition, note that almost all the finite verbs found preceded by the definite article begin with h, and thus may be explained as cases of dittography. Therefore, it is possible that the pattern with finite verbs is a late analogy to the pattern with participles and developed after the participle became a part of the verbal paradigm, thus ’aser+predicative participle: ’aser+ finite verb:: haC+participle:? > haC+finite verb. 36 That is, there can never be an overt subject in this ‘relative clause’ and there can never be constructions of the type **mal’aÈimi hab ba’imj ’elehemi ‘messengers to whom people come’. The Arabic na¨t sababi pattern allows the attribute to refer to a different referent than the head noun. However, the na¨t is clearly a nominal pattern, as its negation is the nominal negator gayr rather than the verbal negators ma/la (Polotsky 1978: 169, 171 2). Note also that the predicate of the alleged relative clause is as 29


attached only to the participle and never to any other element. All these are substantial differences between the relative clause and the participle clause, the latter being highly restricted compared to ’aser relatives.37 Besides the lack of agreement in definition, the N-participle combination is syntactically identical to the N-Adj. combination (for which see section 3.4). In addition, there is no reason to treat an article on participles as having a function distinct from the article on substantives. Therefore hqdsm in h’lnm hqdsm cannot be a sentence, if h’lnm is not a sentence. Moreover, if Gzella’s analysis is correct then the PS relative-determinative *∂V could have similarly transferred to the substantive from its attributive position, but this did not happen.38 Relative clauses, like adjectives, are exponents of the attributive relation, but this does not imply that their internal syntax is identical to that of the adjective, nor that the article is identical to the relative pronoun, as I hope the discussion above illustrates. signed case by the head noun, not by its function within the clause, that is, it bears regular nominal agreement: …l adwiyati l gayri muÒarraÌn bi bayï ha qanunan… ‘drugs, the sale of which is prohibited by law’ (Qanun as sur†a 36, apud Polotsky 1978: 171, e.g. 40). 37 Siloni (1995: 448, n. 3) acknowledges the possibility that participles can func tion as nouns or adjectives, but maintains that they are verbal in this pattern, because they take direct objects, while nouns cannot take direct objects. This approach is a bit circular, as Siloni takes this to be a given, while this is what needs to be proven and in any case, examples where participles take direct objects when they are non verbal exist: ki m¢ragg¢lim ’e† ha ’areÒ (Gen. 42:30) ‘as spying over the land’; l¢ no†¢rim ’e† piryo (Song 8:12) ‘people guarding its fruit’. Siloni’s sole reason to analyse the article as a relativizer is her inability to explain the lack of article on the head noun, despite the fact, that even according to her parameters, there is only one minor similarity between relative clauses and participle clauses (see her table in p. 452). It is also re grettable that Siloni’s data and analysis are so misleading. Many of the examples she provides as ungrammatical may be grammatical with other lexemes; she uses sen tences with the presentative hinne which show a different syntax than regular nomi nal sentences in the language. She ignores some important facts concerning partici ples in Hebrew, e.g. participles can occur without definite article in Modern Hebrew, only when they have been fully substantivized, thus qonim ‘shoppers’ (her example) is acceptable, but not ’oÈlim ‘diners’, mitnaÌalim ‘settlers’ but not yosßim ‘residers’. Finally, she refers to Arabic na¨t sababi construction to prove that in some languages, the participle may predicate a non subject element. Here too the participle is no different than the attributive adjective and merits no special consideration. Moreover, Siloni ignores the clear indication of case in Arabic, which undeniably points to a nominal agreement with the head noun. 38 In some Neo Arabic dialects, the definite article is used as a relativizer, probably due to contamination with the relative pronoun illi; this is, however, a secondary and late innovation and should not be taken to represent the original function of the Arabic article. 30


Note also that like the Arabic improper annexation, here too, the pattern is not definite unless the head noun is also definite: äl ham-mizbeaÌ hab-banuy (Judg. 6:28) ‘on the already-built altar’. ha-älma hay-yoÒe† lis’oß (Gen. 24:43) ‘the maiden going out to draw (water)’.

This pattern occurs in Mishnaic Hebrew with the same syntax: indefinite noun is followed by a definite participle (e.g. Er. 9:4). If indeed the function of the particle *han was to mark definiteness, a definite article on the attributive participle should have been enough to render the entire construction definite. 3.3 Distribution: Phoenician and Old Aramaic In both Phoenician and OA the definite article appears mostly on modified nouns, i.e. nouns followed by attributive demonstratives or relative clauses. In OA, out of 54 nouns in Lambdin’s sample over half were followed by an attributive demonstrative and only 10 were unmodified (Lambdin 1971: 318). This is very similar to the situation in Phoenician. Firmage went as far as to suggest that before relative clauses the article is a discontinuous morpheme: h… ’s (2002: 33–4). 3.4 Common deviation: Noun-[Article-Adjective] In all languages, there are cases where only the adjective is marked with the definite article, in contradiction to the rule of noun-adjective agreement. It should be noted that the meaning of these clauses remains identical to their meaning were the noun to be marked with the article as well. 3.4.1 Arabic In Arabic, this pattern was largely reanalysed as a construct chain and the adjective was marked with a genitive case (Ewald 1833: II 29; Brockelmann: II 209, §132b–c; Wright 1967: II 232, §95f ).39 Most Arabists agree that the reinterpretation of this pattern as construct, which is more common in the modern dialects than in Middle Arabic, is later and arose due to the external similarity of the two (Hopkins 1984:§186, n. 1). This pattern is quite common with numerals: 39 Note that this syntactic re interpretation goes against the meaning of the clause: yawmu s sabiï is ‘the seventh day’, not ‘the day of the seventh N’, daru l aÌirati is ‘the other world’, not ‘the world of the other N’ and so on. In addition, note that in the noun and adjective in this pattern are always in agreement, which is not the case for the construct: (Lebanese Arb) sint il ma∂ye ‘last year’.

31


baytu l-muqaddasi ‘the holy temple’ Òalatu l-’ula ‘the first prayer’ yawmu s-sabiï ‘the seventh day’.

Grotzfeld (2000) suggested that the pattern was used to denote binary alternatives: ‘the big N’ as opposed to ‘the small N’. This opposition is marked both morphologically, with the elative ’akbar-kubra, and syntactically, with a pattern where only the adjective is marked with a definite article while the noun is not. Grotzfeld, like some other Arabists, is of the opinion that this pattern is an example of secondary article deletion. He estimates that the rise of the elative caused the merger of the binary pattern and its subsequent disappearance. Thus, the pattern must have been common before the rise of the elative.40 This pattern is found in all attestations of Middle Arabic (Christian, Jewish and Muslim). The pattern appears even when quoting Quranic expressions which show regular determination originally, like ◊«d rOI*« Òira† l-mustaqim ‘the straight path’ (Hopkins 1984: 182 §186).41

dOJ« fK ‰UIA ‘the weight of the large fils’ (Hopkins 1984: 182 §186) WO «d)« w{«—« ‘the lands where kharaj is paid’ (Hopkins 1984: 182 §186)42 wU « Âu Ë ‘the second day’ (Hopkins 1984: 203 §212) In Judaeo-Arabic, the appearance of the article only on the adjective is a general phenomenon. Blau (1952: 33; 1995: 168, n. 20) suggests that it originated from a pattern with ordinal numbers, which is common in Judaeo-Arabic: ‫‘ יום אלחמסין‬the fiftieth day’. ‫‘ פי טריקת אלמסתקימת‬on the straight path’ (Blau, 1995: 161 §229) ‫‘ מן סכה אלגדידה‬from the new coin (f.)’ (Blau, 1995: 161 §229)

Blau proposes two possibilities: either the N-Adj. was understood as one word, a kind of compositum, or the combination was understood as a construct. The former suggestion does not account for the appearance of the article on the second element, instead of on the entire phrase: **al[N-Adj]; the second suggestion seems circular: the noun is understood as the regens because only the adjective is marked with an article. The head noun of a relative clause is occasionally not marked with an article, even when the clause opens with alla∂i (Blau, 1995: 233 §355, §357). 40 Another possibility which Grotzfeld puts forward is that both patterns split from a single pattern. 41 Cf. aÒ Òira† l mustaqim (Qur’an 6:71). 42 The final ya’ is either a genitive or perhaps a construct (Hopkins 1984: §186, n. 4).

32


Similarly in Christian Arabic, the adjective is commonly marked only on the adjective (Blau 1966: 356, §239), with the same distribution as other Middle Arabic dialects:43

WbI*« ÷—« ‘the holy land’ WIOF« dH« ‘the old books’ WOô« WFO ‘human nature’ The pattern is very common in Neo-Arabic, particularly in both the Gulf and the Levant, but is found in other dialects as well.44 As many Arabists dealing with modern dialects have noticed, speakers tend to correct themselves and avoid using the pattern when asked to produce a definite N-Adj. combination.45 Thus, the exact spread of the pattern is not quite clear from the literature. • Levantine Arabic: bayt l-maÌru¨ ‘the burnt house’ ba††iÌ ¢l-’aÌ∂ar ‘the green watermelon’ su’ ¢l-äti’ ‘the old marketplace’ sint il-ma∂ye ‘last year’ (cf. Modern Hebrew [rel.+verb] has-sana se-äßra) sint ig-gaye ‘next year’ (cf. Modern Hebrew [def.+participle] has-sana habba’a)

Feghali (1928: 135) suggests that the pattern originated from an older, classical, pattern which the Arab grammarians attempted to explain as either an annexation where the adjective is understood as a ‘real’ substantive,46 as a kind of compositum (ämu l-’awwali > ämulawwali ‘last month’) or as an annexation with a missing noun (Òalatu l-’ula < Òalati l-saäti l-’ula).47 43 Blau (ibid.) notes that since the article on the adjective is the one marking the opposition between predicative and attributive, the article on the noun was felt to be redundant. Note also the positioning of the article on the noun alone is much rarer (Blau, ibid §240). 44 See also Rhodokanakis (1911: II §89a) for Omani Arabic. In Maltese, the pat tern is common mainly in toponyms and adverbials: wied il kbir ‘the great valley’; blata l bajda ‘the white rock’. See Borg (1989) for details and discussion. 45 Grotzfeld (2000: 10) notes that speakers of Lebanese Arabic tend to ‘correct’ this pattern in writing according to what they think is normative, thus adding l before the noun, while in speech the pattern with no l before the noun is productive. Similarly, my colleague SN explained ktab ¢z zgir ‘small book’ in his Southern Leba nese dialect as ‘incorrect fast speech’. See also Borg (1989: 78, n. 19) who quotes a local scholar explaining a similar pattern in Maltese as ‘orthographical abbreviation’. 46 See also Grotzfeld: ‘adjektivische Attribute, die unterscheidenden Wert haben, sind gewöhnlich wie untergeordnete Substantive konstruiert: selber mit Artikel bei einem Beziehungswort ohne Artikel’ (1965: 93). 47 Note also Borg (2000: 30) who suggests that BH patterns such as Òom ha äsiri

33


• Iraqi Arabic The pattern is quite common and productive in the Christian and Jewish dialects of Baghdad, but can be found also in the Muslim dialect, though there it occurs mostly in fixed combinations (Blanc 1964: 126; Mansour 1991: 44).48 The tendency to reanalyse it as a construct is even more pronounced in this dialect, as can be seen in examples with feminine nouns, where the construct differs morphologically from the non-construct. All Iraqi dialects use the pattern for place names, like bab es-sarji ‘the Eastern Gate (a name of a city quarter)’, and some recurrent expressions, such as id el-yemna ‘the right hand’49 as well as for temporal expressions, such as sahr ej-jay ‘next month’. C: J: M:

ïd le-kbig ‘the great feast (Easter)’ qondart el-lexxi ‘the other shoe’ Òlat l¢-jdidi ‘the new synagogue’ dfater el-ëttaq ‘the old note-books’ ïd ec-cebir ‘the great feast (Feast of the Sacrifice)’ ënab l-aswad ‘the black grapes’

The regular relative construction is le-blad s-sefnaha ‘the country we saw’. However, in all dialects, though particularly in the Jewish dialect, there are examples where the head noun is not marked (Blanc 1964: 127). Here too, the pattern is occasionally reinterpreted as a construct. In the Christian and Muslim dialects the pattern is again common with temporal expressions, while in the Jewish dialect it is very productive: C/M: sant ej-jina ‘the year we came’ waqt el-kent bej-jays ‘the time you were in the army’ J: sen es-suwwet-u ‘the thing you did’ ßaÌur ella∂i qa-yezzawwaj ‘the young man who is getting married’ ma-ynam be-mkan el-yebgadlu ‘he doesn’t sleep where [lit. in a place that] he’ll be cold’

Since this pattern occurs in CA and other modern dialects of Arabic, Blanc (1964: 128) cautiously suggests that the Jewish and Chris(Zech. 8:19) may be normal constructs with two nouns and an attributive adjective, i.e. not the tenth fast, but rather the fast of the tenth (month). 48 Geoffrey Khan noted (in a personal communication) that in the Eastern Arabic dialects one cannot exclude a possible influence of a Persian substratum, where the izafe pattern is similar to the pattern discussed here. 49 Note Syriac ’i∂eh da ymina ‘the right hand’ where the N Adj combination is reinterpreted as a genitive construction. 34


tian dialect preserve traces of an older usage, while the Muslim dialect shows widespread levelling. 3.4.2 Hebrew The BH article may precede the attribute of a formally unmarked noun (Lambdin 1971: 321; Waltke and O’Connor 1990: 260). Note that the noun is not in the construct state, thus, unlike Arabic, this pattern in Hebrew is synchronically understood as a noun-adjective construction:50 seßa¨ paro† ha†-†oßo† (Gen. 41:26) ‘the seven healthy cows’ qane ha†-†oß (Jer. 6:20) ‘the good cane’ ÌaÒer hag-g¢dola (1 Kg. 7:12) ‘the big courtyard’ ruaÌ ha-raä (1 Sam. 16:23) ‘the evil spirit’

As in Arabic, this pattern is quite common with ordinals (Brockelmann 1961: II 209, §132c–d):51 miy-yom ha-rison äd yom has-s¢ßiï (Exod. 12:15) ‘from the first to the seventh day’. ’el mabo has-slisi (Jer. 38:14) ‘the third entrance’. ä∂ m¢qom saär ha-rison (Zech. 14:10) ‘as far as the place of the first gate’.

Lambert (1895) notes that the word yom is always unmarked with an article when followed by an ordinal number, while the number is always marked. There are only two exceptions, both in late books (Neh. 8:18 and Dan. 10:12). In the vast majority of the cases the word yom is preceded by a preposition, where the definiteness is only revealed in the vocalization. Lambert (1895: 280) further claims that the masoretes added the article wherever they could, without changing the consonantal text, although originally no article was there. The original pattern is, according to Lambert, a construct. This situation is even more widespread in MH, where it seems that when the article is used, the combination N-definite Adj. is more common than definite N-definite Adj (Sarfatti 1989:156).52 The consonantal text makes some examples ambiguous, e.g. b¢ dereÈ ha† †oßa w¢ ha ysara (1 Sam. 12:23) ‘the good and upright way’. 51 Note that examples, where the noun is marked with article or possessive suffix and the ordinal is not, are not definite: ha ros ’eÌa∂ (1Sam. 13:18). Similarly: ’aÌiÈem ’eÌa∂ (Gen. 42:19) ‘one of your brothers’, despite the suffix on the noun. 52 There is also a pattern where only the noun is marked with the article, however in Sarfatti’s corpus there were only 7 such occurrences, compared to 63 of the pattern under discussion. Sarfatti (1989: 157) suggests that the pattern where only the noun is marked was a substandard form. Ben Hayyim (1992: 434) analyses the adjective in 50

35


naära ham-m¢’orasa (San. 7:4) ‘a betrothed maiden’ li-mqom mayim ha-raïm (Avot 1:11) ‘to a place of evil water’

While Kutscher (1982: 217) claims that this pattern is the general rule in MH, Sarfatti (1989: 161) prefers to call it ‘a tendency’ as there are no rules that explain the preference of one pattern over the other. Sarfatti (1989: 164–5) concludes that the use of this pattern probably expanded since it was secondarily understood as a construct. Ben Hayyim (1992: 438) suggests that this pattern is a subject-less relative clause: naära ham-m¢’orasa > ‘a girl who is engaged to be married’.53 This of course applies, then, to phrases like han-naära hamm¢’orasa. Like Gzella (2006), Ben Hayyim bases his suggestion diachronically on the occurrences of haC before a finite verb in the Bible.54 For him, haC and ’aser have the same syntactical function. 3.3.3 Phoenician There are relatively few examples of noun-adjective combinations in Phoenician and only one example of an adjective marked with an article while the noun is not: ’lnm hqdsm ’l ‘these holy gods’ (KAI 14:22).55 Firmage (2002: 41) suggests that since a definite noun typically came before the demonstrative, in this example, the adjective, which is positioned immediately before the demonstrative, took the article by analogy. 3.3.4 Syriac In Syriac, the noun is in the absolute state, instead of the emphatic one, when it is preceded by cardinal numbers.56 However, the adjective is in the emphatic state: hannon tla†a gaßrin zaddike ‘these three righteous men’ (Aph. Fide 29: 19 20)

Thus, it seems that the adjective is the one carrying the state distinction (a remnant of the definite-indefinite opposition), rather than the noun (Brockelmann 1961: II 210, §132e) the latter pattern as an adverbial clause: ba äbur hay yele∂ Ìay (2 Sam. 12:21) ‘the boy, (when he was) alive’. 53 Ben Hayyim does not discuss the semantics of such a pattern, so it is hard to tell whether he thinks it carries a definite meaning or not. 54 See my discussion above in section 3.2.1. Note also Pat El and Treiger (2008) for the rejection of subject less relative clauses in Semitic. 55 Another possible example is ’rÒt dgn h’drt ‘rich lands of grain’ (KAI 14:19), which depends on the reading of dgn; if it is the name of the god, the construct is definite, if not the article on the adjective does not agree with the noun. 56 The absolute state is the normal state of the noun after ordinals in Syriac. 36


3.3.5 Summary All the languages surveyed above, except perhaps Phoenician, exhibit this pattern as part of their system, not as an accidental form, despite the fact that it is in contradiction with the regular agreement rules in these languages. Several explanations have been given to account for its existence: compositum (Feghali 1928; Borg 2000), construct (Blau), relative (Ben Hayyim 1992; Gzella 2006) and syntactic elative (Grotzfeld 2000).57 The assumption of compositum is not in accordance with Semitic syntax: if the entire combination is understood as one word, why is the article medial and not prefixed? The construct explanation is not in accordance with the meaning of the pattern, and the regular gendernumber agreement between the noun and the adjective indicates that the adjective modifies the noun and is not dependent on it. The claim for a relative origin I have dealt with above. Grotzfeld’s suggestion is interesting, but it does not resolve the historical origin of the pattern. 4. The Function of the Article: Re-evaluation Analysis of the data presented above indicates that the article’s original function was not to mark definiteness and that it belongs with the attribute. The following is a summary of the arguments, based on the examples and discussion above. 4.1 The origin of the CS article is not a demonstrative Word Order While the CS article may be pre-positive or post-positive, the demonstrative in the Semitic languages consistently occupies an attributive position after the noun and the adjective. It is unreasonable to assume an unmotivated demonstrative movement from a nounphrase-final position to a noun-phrase-initial position, when we have no proof for such a development. Inflection The article has no inflection in any CS language. This is a unique situation for a Semitic pronominal form. If indeed *ha(n) was a demonstrative, the loss of its inflection is unaccounted for and even unlikely; the CS languages maintained at least gender/number distinc57 Borg’s explanation may also be closer to the one made by Grotzfeld. Borg (1989, 2000) claims that this pattern did not originate from the nominal construct and is independent of it. He suggests that the pattern with what he calls ‘initial article’, [N article Adj.], antedates the regular pattern [article N article Adj.], and was used for specific combinations in order to mark the entire phrase as one lexeme.

37


tion in the nominal paradigm. Even the article in Amharic, a language devoid of most nominal inflectional distinctions, exhibits gender distinctions. If PS *han did not have an inflection, as is very likely the case, it cannot be regarded as a demonstrative pronoun. Definiteness The existence of patterns like the improper annexation in Arabic or the ‘relative participle’ in Hebrew, where the definite article does not render the noun definite, casts doubt on the assumption that the article marks definiteness. The fact that CA ga¨du s-saäri and BH mal’aÈim hab-ba’im are indefinite, despite having an article, can be explained if the article was not originally a definite marker. This point also substantiates our rebuttal of the demonstrative origin: the demonstrative is the epitome of definite markers. Languages which do not have an article, like Russian or Ug., use the demonstrative to render a noun specifically definite. However, if *han is inconsistent in rendering elements definite, it seems unlikely that it arose from a demonstrative. 4.2 The article was first positioned on the attribute PS *han The existing data show that *ha(n) is not an independent demonstrative, but some kind of adnominal element. When it appears in a language with a morphological distinction between attributive and independent demonstrative, it is attached to the attributive demonstratives. We never find ha-Dem. as a pronominal demonstrative where a bare demonstrative functions as an attributive. This is not a coincidence. Predicative Adjective The article distinguishes the attributive adjective from the predicative adjective. If we assume that the article marked attributes, it will explain its exclusion from predicative position. Possessive Pronoun The Semitic possessive pronouns are not determiners, as in IE languages, and yet nouns with these suffixes exclude the article. This phenomenon should be taken with the exclusion of the article from construct noun. Both N+N and N+Poss. are basically the same pattern, where the possessive pronoun is a pronominalized N (Goldenberg 1995: 3–4). In both patterns, the second element is the attributive of the head N. While in N+N, the article appears on the attributive (the rectum), in N+Poss., this is not possible, as the possessive suffix is a clitic, not an independent word.58 58

Note that in some languages, a noun marked with an article is semantically equivalent to a noun with possessive suffixes: Aram. ’aba, ’imma = ‘the father/mother, my father/mother’ (similarly with brother). 38


Double-marking The ‘double’ marking on the noun and adjective is commonly explained as a case of analogy; all scholars assume that the adjective takes the article due to its agreement with the noun. However, whenever an influence in the nominal system can be traced, it is normally the adjective that influences the noun (N < Adj).59 Typically in language change, the regular influences the irregular. Moreover, if movement of features from the noun to its modifier was indeed possible or common, the possessive suffixes would have been moved as well; if *han-malku kabiru yielded *han-malku hankabiru, then *malku-na kabiru ‘our great king’ might be expected to have yielded **malku-na kabiru-na. Yet no Semitic language, ancient or modern, has ever taken that path.60 Rules of noun-adjective agreement do not imply identical morphology in Semitic.61 Nominalization One of the important features of the article is its capacity to nominalize adjectives:62 al-kabiratu ‘the big woman’, qaddisa ‘the holy man’ etc.63 Excluding some Neo-Aramaic dialects, the demonstrative never nominalizes in Semitic. If this function of the article is original, which it seems to be, there is no reason it would be attached to nouns. The Semitic relative pronoun is also an adnominal marker and a nominalizer. The difference between the article and the relative pronoun is the grammatical elements they nominalize: *ha(n)- nominalizes all elements with nominal inflection: adjectives, participles and attribu59

The plural endings are a case in point: some form of broken plural is known to exist in proto WS, but in NWS, excluding a group of relics, all nouns take regular adjectival endings. Furthermore, the anding * at for f.pl. and * un for m.pl. are pro ductive and predictable on nouns in NWS, despite the fact that nouns normally do not have predictable plural forms. The Semitic feminine ending t/ at also likely origi nated from the adjective. It is most likely that in PS substantives were not morpho logically marked for gender, e.g. Arb. ’umm ‘mother’, BH ’eßen ‘stone’ and pairs such as Arb. gamal ‘camel’ / naqa ‘she camel’, where natural gender is marked lexically. 60 Even Amharic, which uses the suffix pronoun to mark definiteness, does not mark both noun and adjective with the article. 61 Arabic has fsg agreement with inanimate plural nouns of both genders, while collective nouns with singular morphology usually take plural adjectives, though they may occasionally be found with singular adjectives (Reckendorf 1921: 59). Some sporadic examples of lack of agreement in case are also found. See Reckendorf (1921: 60 §41.8). 62 The article is not strictly speaking needed to nominalize adjectives. Yet, beyond a small set of adjectives, it is hard to find an adjective without an article outside the attributive position. Some adjectives may be fully substantivized, ex. BH Òaddiq ‘right eous man’ and then their syntax is like that of substantives, i.e. no article is needed. 63 Note that even in Amharic the definite article nominalizes adjectives: t¢ll¢q u ‘the big one’ (Leslau 2000: 49 §44.5). 39


tive demonstratives, while *∂u nominalizes elements without nominal inflection: prepositional phrases, adverbs and sentences. These patterns appear in all CS languages which developed an article.64 In short, the two forms are in complementary distribution. When adjectives are nominalized, it may be argued that the article is their head: al-kabir, hag-g¢dola ‘the older one’. In fact, the article does not determine the syntax of the nominalized element; case, number and gender are already marked in the latter’s morphology.65 Conversely the relative pronoun carries all these traits when nominalizing prepositional phrases, adverbs and sentences (Pat-El and Treiger, 2008). The pronoun *∂V is the head of what follows it (note the genitive on the noun in CA du l-qarnayni), but *ha(n) never shows this behaviour: the case of al-kabir- is determined by the function of the noun to which al- is affixed. 5. Origin The most likely origin of the article is a deictic particle *ha and its derivatives, *han and *hal, as is suggested by Hasselbach (2007).66 One or both of these particles exist in all the Semitic languages and mostly function as presentatives:67 Akk. allû, annû (anna?) Amarna. allû, annû ‘presentative’ (Rainey 1988) BH. hinne, hen (he’?)68 64

The pattern with PS *∂u should be reconstructed to Proto Semitic (Pat El and Treiger 2008). 65 In Arabic, the article excludes nunation, but it does not determine the case of the form; the case is determined by the syntax of the noun, regardless of definiteness or the presence of the article. 66 Rubin (2005:75) emphasized that *han and *’ul are not etymologically re lated. He also did not connect any of them to *ha. Other scholars, like Tropper (2001), assumed a phonological process *han > *hal. 67 The development presentative > article is found elsewhere as well: In Creole French, the French presentative la is used as a definite article, while the old article le/ la/les is not used as a definite article (Grant 1995). Sissala, a Niger Congo language, spoken in Burkina Faso and Ghana, uses a derivation of the verb ‘to see’ as a definite article (Blass 1990: 197). This form also functions as a presentative and is attached to demonstratives to indicate spatial distance (Blass 1990: 198 9). Finally, in French the demonstrative ce/cette/ces are compound forms which originate from an attach ment of an uninflected Latin presentative ecce to demonstratives: ce < ecce hoc, cet/ cette < ecce iste, celui/celle < ecce ille (Harris 1980: 147). 68 Brown (1987) suggests that halo’ represents two separate particles: ha+lo’ (inter rogative with the negative particle) and *halu’ etymologically related to Aram. hlw/’lw and Akk. allû. He notes that translating halo’ as ‘is it not? nonne?’ cannot be applied 40

THE DEVELOPMENT OF THE SEMITIC DEFINITE ARTICLE 69 Arm. h’/ha/, ’alu (BA), hlw (EgA) (Muraoka and Porten 1998: 329) Ug. hl /halli-/,70 hn Arb. hal, ’inna (G. h¢yya ‘there’ perhaps connected to BH he, na-yya ‘here (I am)’, nahu ‘presentative’)71

This function of the presentative may be attested in Ug. as well, though Ug. did not develop an article. There are examples of a relative pronoun preceded by the prefiix ha: ’anykn dt l’ikt mÒrm hndt b Òr mtt (KTU 1.19.10 13) ‘The group of ships that you sent to Egypt, which was wrecked off Tyre’.

Pardee analyses hndt as an independent demonstrative referring to the f. noun ’any (2003/2004: 136) against Tropper’s reading it as a locative adverb (230). Pardee takes hnd to be the basic demonstrative in Ug. However, it is quite possible that hnd and its derivatives are themselves complex entities: hn-d, hn-dt, as in other Semitic languages, and as can be seen in other Ug. deictic particles, like hnk etc.72 This is also evident in the fact that hn is a presentative, and only when it takes a demonstrative extension, like d, k, m, it is a demonstrative pronoun. This pattern occurs with substantives as well:73 hn alpm sswm hnd (RS 16.402 31 2 and 37 8) ‘these 2000 horses’.

Both Tropper (2000: 233) and Pardee (2003/2004: 125, n. 455) suggest that hn here functions as a locative adverb. This is possible, but consistently to all the biblical examples (e.g. 1 Sam. 20:37; Ps. 54:2; Prov. 26:19). Brown (1987: 219, n. 103) further suggests that the ’ was added to *halu’ because it was confused with the homophonous halo’. 69 To this Brown (1987: 211 12) adds ’aru. The semantic similarities between hlw/’lw and ’rw should not be confused with etymological connection; therefore, I see no reason to include this form in the list. Another possible form is Syr. harka ‘here’, which Rubin (2007: 123) suggests is derived from hn ka, which underwent *#n > #r (Testen 1985). 70 See also hlk, hlny (in some contexts interchangeable with hnny) in the opening formula in epistolary usage (Brown 1987: 203 4; Tropper 2000: 738; Pardee 2003/ 2004:365), hlm (Tropper 2000: 797, Huehnergard and Pat El 2007: n. 55). For evidence of doubling see syllabic spelling in Huehnergard (1987: 33): al li ni ya. 71 Na/na is a very common presentative in G. and Ethio Semitic, and later became the suppletive imperative ‘come’ in Amh. For more forms see Leslau (1987: 380 1). 72 Pardee himself suggests that the forms hnhmt (RS 15.128:8) and hnmt (94.2965:20) are a combination of presentative hn and independent pronoun. 73 Note the similarity to the syntax of OA and Phoen., described in Lambdin, where the definite article mostly appears with attributive demonstrative and relative clauses. 41


may still show us an early stage of the development which led in other languages to the grammaticalization of the article.74 The presentative hn may also precede a noun followed by an attributive demonstrative: hn bns hw ‘that servant’ (RS 96.2039:8, 10, 15). In one example, the spelling hbt ‘(the) house’ (RS 29.093:16) either points to an assimilation, or to the use of ha, instead of han.75 The connection between the prefix ha and the presentative is apparent in other patterns as well. In Huehnergard and Pat-El (2007), we have shown that Arab. hada and BH hinne ze have very similar syntax, which we argued is cleft, where the presentative is marked as the predicate: ha-da bnuka wajadnahu bi-’a¨la makkata (Ibn Hisam Sirah 113) ‘Here is you son whom we have found in the mountain of Mecca’.76 w-hinne-ze mal’aÈ nogea¨ bo (1Kg 19:5) ‘Then an angel touched him’.

This pattern is used to predicate the excitement of the speaker, similar to Fr. voici que, voila que, which also shows the pattern [Presentative]-[Subordinate clause]. In both languages, the presentative may be syntactically separated from the relative pronoun, though in Arabic this separation is less common than in Hebrew.77 6. A Proposed Scenario The article was first attached to non-predicative adjectival forms, i.e. to adjectives/participles/demonstratives, either nominalized or attributive.78 These forms derive their attributes externally, i.e. from an external referent. It is possible that the incentive for marking non-predica74 In this I take Pardee’s analysis against Tropper’s (2000: 233, §42.73 4). While Tropper argues that hn (which he vocalizes as hannV ) in the examples quoted above functions as a demonstrative, Pardee (2003/2004: 135, 365 and elsewhere) insists that there is no substantiation for a demonstrative function as ‘hn functions uniquely as a presentative particle in Ugaritic, and the same is true of the expanded forms hnn and hnny’ (p. 365). Pardee (ibid, n. 494) further suggests that French voici works best for hn. 75 Pardee is certainly right that there are too few examples in Ug. to substantiate Tropper’s claim for an article. 76 In Christian Arabic, ha∂a is regularly used as a presentative (Blau 1966: 463, §363). 77 Note that in the article cited above, which dealt with the Semitic determinative *∂V, we have claimed that the patterns are similar syntactically, though using etymologically different elements. My claim here is that these elements are not etymologically different. 78 This was already alluded to in Lambdin (1971: 324).

42


tive adjectival forms is that nominal phrases and nominal sentences have the exact same syntax: Akk. sarrum dannum could be interpreted as ‘the king is strong’ or ‘the strong king’.79 Thus, the basic opposition is predicative (unmarked) vs. attributive (marked). The process nonpredicative > definite should be reconstructed as follows:80 Stage 1: *ha-∂u, *ha†-†ab- ‘this one (m.s)’, ‘a/the good one’ >

This was expanded to positions where a head noun precedes: Stage 2: *kalb- ha-∂u, *kalb- han-†ab- ‘a/the dog, this one’, ‘a/the good dog’ >

Due to its frequent appearance on attributives, at a later stage, it was understood as an adnominal marker. As a result there is movement of the article to the noun due to appositional relation, though stage-2 patterns were never lost: Stage 3: *han-kalb- ha-∂u, *han-kalb- han-†ab- ‘a/the dog, this one’, ‘a/ the good dog’ >

Note that the transference of article did not occur in construct, because the nouns there are not in apposition. An opposition between marked (+feature) and unmarked (ø-feature) is formed. Although the opposition was originally not differentiated semantically, it was reanalysed as distinguishing definiteness, due to the nature of the prefix (< presentative) and the elements to which it was attached: Stage 4: *han-kalb- han-†ab- ‘the good dog’ as opposed to kalb- †ab- ‘a good dog’.

Presentatives are typically excluded from predicative position, so their use as a marker of non-predicative function is logical. The pattern discussed in Huehnergard and Pat-El (2007) exemplifies that the assignment of a predicate function to a presentative can be done only through cleaving. The process adnominal > article may be compared to the development of the relative pronoun in Semitic. Subordination was originally done by dependency: construct head+sentence, similarly to nominal construct: construct head+noun. The determinative pronoun expresses the appositional relation: non-construct head=∂u+sentence, where the 79 A similar proposal was made in Pat El and Treiger (2008), where the authors argued that the incentive to mark prepositional phrases with the relative determina tive pronoun was to mark these phrases as specifically adnominal, as opposed to predi cative, where they are un marked. 80 Note that all the stages are attested in most languages.

43


head noun and the relative pronoun are in apposition (Goldenberg 1995; Pat-El and Treiger 2008). Later in Arabic, the determinative ∂u was understood to be a marker of definiteness: N-Sentence > (adnominal marker) N=∂u-Sentence > (marker of definite head) N ∂u-Sentence as opposed to N Sentence (non definite head). Thus Proto-CS may have had the following pairs, where the ø option was inherited from PS: Non-nominal attributes: ø: *ga’a farisu yarkabu äla farasin ‘a horseman who rides a horse came’. adN: *ga’a farisun ∂u yarkabu äla farasin ‘a horseman who rides a horse came’ Def. > ‘the horseman who rides a horse came’.

Nominal attributes ø: *farisun kabirun ‘a/the strong horseman’ adN: *farisun ha-kabirun ‘a/the strong horseman’ Def.: > ‘the strong horseman’

Note that in both constructions the attributive complex is marked for agreement: in the determinative pattern the head pronoun carries agreement, while in the ‘article’-pattern the nominal attribute does. Thus, both have the same phrasal syntax, regardless of their internal syntax: Ni-[Att.]i (Goldenberg 1995). The complementary distribution with the relative-determinative pronoun is easily explained: both forms mark adnominals, but the relative-determinative pronoun marks forms without nominal inflection (sentences, adverbs and prepositional phrases), while the ‘article’ marks adnominals with nominal inflection (adjectives and demonstratives). The relative-determinative pronoun is inflected in complete agreement (number-gender-case) with its head noun and with no relation to its syntactical function in the relative clause; the pronoun carries nominal inflection for the entire phrase: faris-un ∂-u rakaba ‘horseman who rode’ or faris-un ∂-u bi-bayt-in ‘horseman at home’.81 However, there is no need for inflection on the ‘article’, because the pattern shows complete nominal inflection on the attribute: faris-un al-kabir-un ‘strong horseman’. Nominal agreement in gender-number-case is a feature of adnominal relations.

81 This feature of the determinative pronoun may be the reason why it was never transfered to the noun like the article.

44


6.1. Aramaic and OSA In Aramaic and OSA the process through which the prefix *han- became the article must have been somewhat different, because the end result is a suffixed article, despite the fact that most likely the origin of the article of these two languages is the same as that of the BH article, i.e. the prefix *ha(-n).82 We may assume a similar process to what was described for BH and Arabic, thus the entire pattern was originally N ha-Dem, as is the case in other Semitic languages. The prefix then was attached to the preceding noun through a process of reanalysis which resulted in N-han Dem (for the phonological process see Rubin 2005: 79–80):83 *malkV han-∂V > *malk-han ∂V > *malka’ ∂V > OA-OfA malka z¢na

This is partially substantiated by evidence from OA where definite nouns are commonly found with a following attributive demonstrative lacking ha- (Lambdin 1971: 318): b-spr’ znh ‘in this inscription’ (Sef. I B:28); spry’ ’ln ‘these inscriptions’ (Sef. II C:9); b-nÒb’ znh ‘in this monument’ (Sef. I C:17). This is the case for OSA as well, where the demonstrative pronouns lack a ha- prefix: ∂n, ∂t, ’ln (Beeston 1984: 41). Except G¢¨¢z, all other Semitic languages use ha- prefix on their demonstrative pronouns.84 Aramaic and OSA are therefore a notable exception. 82

Here, too, the comparison to IE languages, used especially in Rubin (2005: 69 71) and Zaborski (2000: 25), does not stand. All Romance languages use a reflex of a Latin demonstrative, ille or ipse, as their article. Of all the Romance languages, only Rumanian has a post positive article. However, the reason for this difference does not lie in internal historical developments in Romance, but is rather due to contact. A post positive article is a common feature of the Balkan Sprachbund, of which Rumanian is a part: Rumanian castel ul ‘the castle’, Bulgarian ruklja ta ‘the dress’, Albanian vazja t ‘the girl’, Macedonian nedela ta ‘the week'; (Tomic 2004: 15). Another important fea ture which Rumanian shares with members of the group, but not with other Romance languages, is the loss of the infinitive. In short, a post positive article in Romance is not a regular option that may happen at random. Similarly, the post positive article in Old Norse and later in the Scandinavian languages arose from a post positive demonstra tive: *ulfr hinn ‘this wolf ’ > ulfrinn ‘the wolf ’ (> Nor. ulv en ‘the wolf ’), while the article in German arose from a pre positive demonstrative. Thus, none of these examples can be easily compared with the Semitic situation, where N Dem is a shared feature and no contact can explain the difference between Hebrew and Aramaic. 83 Zaborski (2000: 32) has a different interpretation: he suggests that the final a on demonstrative pronouns like ha zena was reinterpreted as belonging to the defi nite article and was transferred to the noun. This explanation does not account for the absence of ha in all early Aramaic dialects. Zaborski also ignores the fact that znh appears only after ms. nouns, but the suffixed definite article a is common to all nouns, regardless of gender or number. 84 Phoenician does so sporadically (Hackett 2004: 376). 45


The prefix ha- was reattached to the attributive demonstrative again in MidA and spread thenceforth (Cook 1992: 10).85 Note that the prefix ha- is attached to different bases in different dialects and therefore should be considered a secondary development in Aramaic: Hatran hdyn, OSyr. hn’ ‘this (ms)’. Some MidA dialects, like Nabatean and Palmyrene, do not use the prefix ha at all, which may indicate that it is still an innovation.86 However, all LateA dialects use it, though not always with all demonstrative forms.87 Thus, since most Semitic languages show the prefix *ha- on their attributive demonstratives, Aramaic and OSA ha-less demonstratives are unusual. The transference of the ha from the demonstrative to the noun may explain its absence in earlier dialects. 8. Summary and Conclusions The common explanation for the CS article has been shown to be insufficient and to rely heavily on typological evidence instead of Semitic evidence. Comparative data and the attested syntax indicate that its origin is not an attributive demonstrative. It was further shown that many of the relics found in both classical and Neo-Semitic languages are very consistent and should be assumed to go back to a common ancestor. Based on the syntax of these relics and the common 85 We may perhaps speculate that the return of the attributive prefix ha is related to the process of article weakening. This process is evident in Eastern Aramaic, but may have had its origins already in MidA. Thus, the prefix ha may have been added to demonstratives to further strengthen their deictic features. This process may be compared to a certain innovation in Creole French, where in a large part of the nominal lexicon, the original French article was agglutinated to the following noun and understood as a part of it (la vie > lavi ‘life’, les yeux > lizye ‘eyes’). These languages then innovated an article based on the deictic particle la, which is suffixed in all forms of French: Fr. le chien > Mauritian Creole lisye ‘a dog’, lisye la ‘the dog’ (Grant 1995). Posner (1997: 384) suggests that this process in Creole French can be seen as a culmination of the process that led to the creation of the French definite article. Possibly as a result of the collapse of the French case system, ille, a semantically bleached demonstrative, was increasingly being used as a noun marker. Ille became the default marker in the absence of other determiners. Posner posits that this is merely a redistri bution of elements, rather than a structural change. 86 Note for example that more conservative dialects, like Q, do not use this prefix (Cook 1992: 16). There is one occurrence of hdn in 4QAmramb, but since the text after it is fragmentary it is hard to evaluate its exact meaning and analysis. 87 Partial list: JPA hdyn, hhyn, hhw’, hhy’; SamA hdh, hdyn, hhw’, hhy’; CPA hlyk, hlyn, hdn, hd’, hlyn; Syr. hana, ha∂e, halen; Mand. hazin, haze, halin; JBA hada, haden, ha’ilen.

46


syntax of the article in CS, I have argued that the article was first attached to a non-predicative form, and was then reanalysed as an adnominal marker. The existence of two patterns, one with a marker and one lacking it, led to the understanding of the prefix as marking a definite nominal as opposed to non-definite noun with ø-prefix. This process happened with the Semitic relative-determinative, whose original function had nothing to do with definiteness, but was reanalysed as marking a relative clause with a definite head. These arguments are substantiated by empirical data and explain the existing syntax and the relic syntax. I have further suggested that the article was derived from the presentative, which is a non-predicative particle. The presentative, or its reflexes han/hal, exists in all the languages. In Ug., which did not develop an article, the presentative hn may be attached to relative pronouns and demonstratives, a stage that I have argued preceded the full development of the article. While Huehnergard’s claim that the article is not a common CS feature stands, we can see the beginning of the process described above in Ug. although it did not lead to the creation of an article there. However, since the preliminary stages exist already in Ug., the use of ha or its reflexes to mark attributes should be considered a common CS feature. REFERENCES Beeston, A.F.L. 1984. Sabaic Grammar. (Manchester) Ben Hayyim, Z. 1992. ‘The Definite Article in the Combination Noun Adjective’ [Hebrew] in idem, The Struggle for a Language. (Jerusalem). 434 9 Blanc, H. 1964. Communal Dialects in Baghdad. (Cambridge) Blass, R. 1990. Relative Relations in Discourse. (Cambridge) Blau, J. 1952. ‘Numerals in Judeo Arabic [Hebrew]’. Tarbiz 23:4, 27 35 1966. A Grammar of Christian Arabic, based mainly on South Palestinian texts from the first millennium. (Louvain) 1974. ‘Studies in Semitic Pronouns (Including the Definite Article)’ [Hebrew]. in E.Y. Kutscher et al. Henoch Yalon Memorial Volume (Ramat Gan). 29 34 1978. ‘On Some Arabic Dialectal Features Paralleled by Hebrew and Aramaic’. JQR 76:1, 5 12 1995. A Grammar of Mediaeval Judeo Arabic [Hebrew]. (Jerusalem) Borg, A. 1989. ‘Some Maltese Toponyms in Historical and Comparative Perspective’, in P. Wexler, A. Borg and S. Somekh (eds) Studia Linguistica et Orientalia Memoriae Haim Blanc Dedicata (Wiesbaden). 62 85 2000. ‘Some Observations on the ‫ יום הששי‬Syndrome on the Hebrew of the Dead Sea Scrolls’, in T. Muraoka and J.F. Elwolde (eds) Diggers at the Well (Leiden). 26 39 Brockelmann, C. 1961. Grundriss der vergleichenden Grammatik der semitischen Sprachen. (Hildesheim) 47


Brown, M.L. 1987. ‘“Is it not?” or “Indeed!”: HL in Northwest Semitic’, Maarav 4:2, 201 19 Brustad, K.E. 2000. The Syntax of Spoken Arabic: A Comprehensive Studies of Moroc can, Egyptian, Syrian, and Kuwaiti Dialects. (Washington) Cook, E.M. 1992. ‘Qumran Aramaic and Aramaic dialectology’, in T. Muraoka (ed.) Studies in Qumran Aramaic (Louvain). 1 21 Davidson, A.B. 1902. Introductory Hebrew Grammar; Hebrew Syntax. (Edinburgh) Dekeyser, X. 1980. ‘The Diachrony of the Gender Systems in English and Dutch’, in J. Fisiak (ed.) Historical Morphology (The Hague). 97 111 Diessel, H. 1999. Demonstratives: Form, Function and Grammaticalization. (Amster dam) Ewald, H. 1833. Grammatica Critica Linguae Arabicae. (Leipzig) Feghali, M.T. 1928. Syntaxe des parlers arabes actuels du Liban. (Paris) Firmage, E. 2002. ‘The Definite Article in Poenician’, Maarav 9, 33 53 Fischer, W. 2001. A Grammar of Classical Arabic. trans. Jonathan Rodgers. (New Haven) Garr, W.R. 2004 [reprint of 1985]. Dialect Geography of Syria Palestine, 1000 586 B.C.E. (Winona Lake, In.) Goldenberg, G. 1995. ‘Attribution in the Semitic Languages’, Langue Orientales Anciennes: Philologie et Linguistique 5/6, 1 20 Goshen Gottstein, M.H. 1946. ‘Definiteness in the Phoenician and Punic Inscrip tions’. Leshonenu 14, 19 38 Grant, A.P. 1995. ‘Article Agglutination in Creole French: a wider perspective’, in P. Baker (ed.) From Contact to Creole (Westminster). 149 76 Greenberg, J.H. 1957. Essays in Linguistics. (New York) 1978. ‘How Does a Language Acquire Gender Markers?’ in J.H. Greenberg (ed.) Universals of Human Language (Stanford). 3: 48 82 Grotzfeld, H. 1965. Syrisch arabische Grammatik (Dialekt von Damaskus). (Wiesbaden) 2000. ‘Rabi¨ al äwwal and Nahr el kibir. The Notion of Dichotomy and its Expression in Arabic’, ZAL 38, 7 14 Gzella, H. 2006. ‘Die Entstehung des Artikels im Semitischen: eine “phönizische” Perspektive’, JSS 51:1, 3 18 Hackett, J. A. 2004. ‘Phoenician and Punic’, in R.D. Woodard (ed.) The Cambridge Encyclopedia of the World’s Ancient Languages (Cambridge). 365 85 Harris, M.B. 1980. ‘The Marking of Definiteness in Romance’, in J. Fisiak (ed.) Historical Morphology (The Hague). 141 56 Haspelmath, M. 1999. ‘Explaining Article Possessor Complementarity: Economic Motivation in Noun Phrase Syntax’, Language 75, 227 43 Hasselbach, R. 2007. ‘Demonstratives in Semitic’, JAOS 127:1, 1 27 Heine, B. and T. Kuteva 2002. World Lexicon of Grammaticalization. (Cambridge) Hetzron, R. 1976. ‘Two principles of Genetic Reconstruction’, Lingua 38, 89 198 Höfner, M. 1943. Altsüdarabische Grammatik. (Leipzig) Hopkins, S. 1984. Studies in the Grammar of Early Arabic based on papyri datable to before 300 A.H./912 A.D. (Oxford) Huehnergard, J. 1987. Ugaritic Vocabulary in Syllabic Transcription. (Atlanta) 2005. ‘Features of Central Semitic’, in A. Gianto (ed.) Biblical and Oriental Essays in Memory of William L. Moran (Rome). 155 203 and N. Pat El. 2007. ‘Some Aspects of the Cleft in Semitic Languages’, in E. Cohen and T. Bar (eds) Studies in Semitic and General Linguistics in Honor of Gideon Goldenberg (Münster). 325 42 Jastrow, O. 1992. Lehrbuch der ™uroyo Sprache. (Wiesbaden) 48


Khan, G. 1984. ‘Object Markers and Agreement Pronouns in Semitic Languages’, BSOAS 47, 468 500 Kutscher, E. Y. 1982. A History of the Hebrew Language. (Jerusalem) Lambdin, T. 1971. ‘The Junctural Origin of the West Semitic Definite Article’, in H. Goedicke (ed.) Near Eastern Studies in Honor of William Foxwell Albrigh (Balti more). 315 33 Lambert, M. 1895. ‘Le mot ‫ יום‬suivi des nombres ordinaux’, Revue des Etudes Juives 31, 279 81 Leslau, W. 1987. Comparative Dictionary of Geëz (classical Ethiopic): Geëz English/ English Geëz with an Index of the Semitic Roots. (Wiesbaden) 1995. Reference Grammar of Amharic. (Wiesbaden) Lieberman, S. 1986. ‘The Afro asiatic Background of the Semitic N Stem: Towards the Origins of the Stem Afformatives of the Semitic and Afro Asiatic Verb’, BiOr 43, 577 628 Lyons, C. 1998. ‘The Origins of Definiteness Marking’, in J.C. Smith and D. Bent ley (eds) Historical Linguistics 1995 (Amsterdam/Philadelphia). 1: 223 41 1999. Definiteness. (Cambridge) Mansour, J. 1991. The Jewish Baghdadi Dialect: Studies and Texts in the Judaeo Arabic Dialect of Baghdad. (Or Yehuda) Meillet, A. (1954). La méthode comparative en linguistique historique. (Paris) Moravcsik, E.A. 1997. ‘Parts and Wholes in the Hungarian Noun Phrase A Typo logical Study’, in B. Palek (ed.) Typology: Prototypes, Item Orderings and Universals (Prague). 307 24 Pardee, D. 2003 4. ‘Review of Tropper Ugaritische Grammatik’, AfO 50. Online Version Pat El, N. and A. Treiger. 2008. ‘On Adnominalization of Prepositional Phrases and Adverbs in Semitic’, ZDMG 151: 265 352 Peretz, Y. 1967. The Relative Clause [Hebrew]. (Tel Aviv) Posner, R. 1997. Linguistic Change in French. (Oxford) Reckendorf, H. 1921. Arabische Syntax. (Heidelberg) Rainey, A.F. 1988. ‘Some Presentation Particles in the Amarna Letters from Canaan’. Ugaritische Forschung 20, 209 20 Rhodokanakis, N. 1911. Der vulgärarabische Dialekt im Δofâr (Åfâr). (Vienna) Rubin, A. 2004. ‘Notes on the Genitive Exponents of Some Modern Arabic Dialects’, Folia Orientalia 40, 327 36 2005. Studies in Semitic Grammaticalization. (Winona Lake, Indiana) 2007. ‘On Syriac harka and Aramaic r < *n’, JNES 66:2, 123 4 Sarfatti, G. B. 1989. ‘Definiteness in Noun Adjective Phrases in Rabbinic Hebrew’ [Hebrew], in M.Z. Kaddari and S. Sharvit (eds) Studies in the Hebrew Language and the Talmudic Literature (Ramat Gan). 153 67 Shlonsky, U. 1997. Clause Structure and Word Order in Hebrew and Arabic: an essay in comparative Semitic syntax. (New York) Siloni, T. 1995. ‘On Participial Relatives and Complementizer D0: a case study in Hebrew and French’, Natural Language and Linguistic Theory 13, 445 87 Testen, D. 1985. ‘The Significance of Aramaic r > *n’, JNES 44:2, 143 6 Tomic, O.M. 2004. ‘The Balkan Sprachbund Properties: An Introduction’, in O.M. Tomic (ed.) Balkan Syntax and Semantics (Amsterdam/Philadelphia). 1 55 Tropper, J. 2000. Ugaritische Grammatik. (Münster) 2001. ‘Die Herausbildung des bestimmten Artikels im Semitischen’, JSS 46, 1 31. 49


Voigt, R. 1998. ‘Der Artikel im Semitischen’, JSS 43, 221 58 Waltke, B.K. and M.P. O’Connor. 1990. An Introduction to Biblical Hebrew Syntax. (Winona Lake, Indiana) Watkins, C. 1976. ‘Towards Proto Indo European Syntax: Problems and Pseudo Problems’, in S.B. Steever, C.A. Walker and S.S. Mufwene (eds) Papers from the Parasession on Diachronic Syntax (Chicago). 302 26 Wright. 1967. A Grammar of the Arabic Language3. (London) Yahuda, A.S. 1906. ‘Bagdadische Sprichwörter’, in C. Bezold (ed.) Orientalische Studien Theodor Nöldeke (Gieszen). 1: 399 416 Zaborski, A. 2000. ‘Inflected Article in Proto Arabic and Some Other West Semitic Languages’, Asian and African Studies 9, 24 35

50

A PROPOSAL TO READ THE LEGEND OF A SEAL-AMULET FROM DEIR RIFA, EGYPT AS AN EARLY WEST SEMITIC ALPHABETIC INSCRIPTION GORDON J. HAMILTON HURON UNIVERSITY COLLEGE

Abstract The legend of a seal amulet, UC 51354, from Deir Rifa, Egypt re mains unclassified. Here the case is made that it is written in Proto Canaanite alphabetic script and reads as a short West Semitic text in dicating the name and title of its owner.

Dr S. Quirke, Curator of the Petrie Museum of Egyptian Archaeology, University College London, brought to my attention a fine domed seal-amulet, UC 51354 in that museum’s collection, that is incised with Egyptian-looking signs but does not render any sense in Egyptian.1 A century ago Petrie found this seal-amulet, together with a set of perforated beads, in his excavations of the several cemeteries at Rifeh/ Deir Rifa, Egypt.2 By the style of some of the other contents of those cemeteries, Petrie thought that they contained burials from the First Intermediate Period to the New Kingdom, mostly from the Twelfth and Eighteenth Dynasties.3 Although Petrie transmitted no record of 1

I would like to express my gratitude to Dr Quirke for drawing this inscription to my attention, allowing me to examine it at the Petrie Museum in August 2005, ar ranging photographs of it to be taken, and granting permission to publish two of them (pls. 1 and 2 below). I would also thank Profs W.E. Aufrecht and F.M. Cross and Mr M.T. Bolmer for their valuable comments on a draft of this paper. 2 W.M.F. Petrie, Gizeh and Rifeh vol. 1 (London 1907), 13, pl. 23.1 (hereafter cited as Petrie, Rifeh). For maps locating Rifeh, which is known as Deir Rifa in more recent publications, approximately six miles/ten kilometres south of Asyut in the ex treme south of Middle Egypt, see J. Baines and J. Malek, Cultural Atlas of Ancient Egypt2 (Abingdon 2000), 121 and J. Bourriau, ‘The Second Intermediate Period’, in I. Shaw (ed.), The Oxford History of Ancient Egypt (Oxford 2000), 200. 3 Petrie, Rifeh, 1, 11, 13. More recently, Bourriau (‘Second Intermediate Period’, 203 5) has examined the contents of Cemetery S from that site, the burials of Nubians known as the ‘pan grave’ people, dating their presence there from the beginning of the Thirteenth Dynasty, c. 1750 BCE to the mid Fifteenth Dynasty, c. 1600 BCE. I 51

A PROPOSAL TO READ THE LEGEND OF A SEAL-AMULET

the context in which this button-shaped seal was discovered, he did publish a reasonably accurate drawing of it, with its five signs arranged largely as a horizontal line (fig. 1 below). I propose to show that the inscription incised on this seal-amulet was arranged mostly as a short vertical column, that its handwriting belongs to the Proto-Canaanite script tradition, and that it can be read as a West Semitic personal name followed by its owner’s title: l qn Ìz, ‘(Belonging) to Qn, (the) Seer’. That name has good parallels documented from both West and South Semitic sources, as does that office especially from later biblical and West Semitic epigraphic texts. If this proposal is accepted by others, the legend on UC 51354 would constitute the fourth ProtoCanaanite inscription found in Egypt, joining two recently published early alphabetic rock inscriptions from Wadi el-Îol4 and a recently relocated heddle jack incised with a West Semitic personal name that was discovered long ago by Petrie at Lahun.5 would like to thank Ms Bourriau for checking her records that this seal amulet is not recorded as having been found in one of those graves. See Section 7 below regarding possible connections between the owner of this seal and the Nubians buried at Deir Rifa. 4 See especially J.C. Darnell, ‘Die frühalphabetischen Inschriften im Wadi el Îôl’, in W. Seipel (ed.), Der Turmbau zu Babel, Ursprung und Vielfalt von Sprache und Schrift vol. 3A (Vienna 2003), 165 71; J.C. Darnell et al., Two Early Alphabetic In scriptions from the Wadi el Îôl: New Evidence for the Origin of the Alphabet from the Western Desert of Egypt (AASOR 59, Boston 2005); F.W. Dobbs Allsopp, ‘Asia, An cient Southwest: Scripts, Earliest’, in K. Brown (ed.), Encyclopedia of Languages and Linguistics2 (Oxford 2006), 495 7; and the palaeographic analysis contained in G.J. Hamilton, The Origins of the West Semitic Alphabet in Egyptian Scripts (CBQMS 40, Washington 2006), with a fuller bibliography on pp. 324, 327 8 (hereafter cited as Origins, plus page or figure number). Based on the documented presence of Semites at Wadi el Îol late in the Middle Kingdom, the choice of locations of these alpha betic inscriptions on a rock wall, most of whose Egyptian texts are assigned to the late Middle Kingdom and Second Intermediate Period, and the minimal development of their handwriting from Egyptian scripts, I would concur with the fairly wide estimate of the time they were written proposed by Darnell et al., c. 1850 1700 BCE (ibid. 87 90, 102 6; Origins, 295 6), against the much later assignment, c. 1300 BCE, advo cated by B. Sass, ‘The Genesis of the Alphabet and Its Development in the Second Millennium B.C. Twenty Years Later’, de Kêmi à Birit Nari 2 (2004/2005), 150 2 (also see nn. 5, 54, 57 below). 5 First published by W.M.F. Petrie (Kahun, Gurob, and Hawara [London 1890], pl. 27.85), the wood of this weaving instrument has been examined by C. Cartwright, H. Granger Taylor, and S. Quirke, ‘Lahun Textile Evidence in London’, in S. Quirke (ed.), Lahun Studies (Reigate 1998), 92 111. M. Dijkstra (‘The So Called ’AÌî†ub Inscription from Kahun (Egypt)’, ZDPV 106 [1990], 55 6) suggested a new deci pherment for the last part of the name of its owner and the present writer has recently proposed a new reading of its initial element (Origins, 62, n. 49). See p. 298 of the latter for interpretations of the recent radiocarbon testing of the wood of this heddle jack, the results of which were early, c. 2140 1940 BCE (95 per cent probability), and 52


Plate 1: UC 51354 with beads (copyright, The Petrie Museum of Egyptian Archaeology, UCL)

Plate 2: Legend of UC 51354 (copyright, The Petrie Museum of Egyptian Archaeology, UCL) 53


1. The Artefacts This seal-amulet is made of steatite, glazed after its inscription was incised. It measures approximately 18 mm. in diameter, almost threequarters of an inch (drawn at a ratio of 1:1 in Petrie, Rifeh, pl. 23.1, reproduced in fig. 1 below). It has the form of a small button, flat on the incised side and domed on the other (see pls 1 and 2). After the cylinder-shaped form of seals known from the Old Kingdom was largely discontinued in Egypt, this button-form was one of several shapes of seal-amulets employed in the First Intermediate Period and Middle Kingdom.6 A spiral decoration occurs around the edge of UC 51354, augmented by two opposing ankh-signs.7 These motifs are also well known from seal-amulets, particularly those in the shape of scarabs, found in Egypt and Palestine that are assigned to the First Intermediate Period, Middle Kingdom, and Second Intermediate Period.8 This seal is perforated throughout, with the ends of the perforation located roughly near the ankh-signs (pl. 2 and fig. 2). pp. 330 1 for a select bibliography on this relocated artifact (EA 70881 in the British Museum). Based on the floruit of Lahun late in the Middle Kingdom, the use of this type of weaving instrument in loom types known from that era (and not those docu mented from the New Kingdom), and the minimal state of the development of its incised Proto Canaanite alphabetic forms from Egyptian scripts, I would agree with the most probable dating for this object proposed by Quirke (ibid. 92), c. 1850 1700 BCE (Origins, 296 9). The combination of those elements, including the more objec tive radiocarbon dating results, precludes the possibility that consonantal alphabetic writing began some four to five hundred years later, in the late fourteenth century BCE, as recently argued by Sass, ‘Genesis of the Alphabet’. 6 See especially T.G.H. James, ‘Ancient Egyptian Seals’, in D. Collon (ed.), 7000 Years of Seals (London 1997), 33 4, 37. For multiple examples of these button shaped seal amulets, see G. Brunton, Qau and Badari vol. 1 (London 1927), pls 32.37 96; 33.97 108. For a vestigial continuation of the older type of cylinder seal, see the fine faience seal of Queen Sobekneferu of the late Twelfth Dynasty (photo: Bourriau, ‘Second Intermediate Period’, opposite p. 192). 7 See Darnell et al., Two Early Alphabetic Inscriptions, 83, and Origins, 328, on the crude ankh sign inscribed beside Wadi el Îol Text 2. 8 Concerning the longevity of scroll and spiral motifs on scarabs, W.A. Ward (Studies on Scarab Seals vol. 1 [Warminster 1978], 55) wrote: ‘All the basic patterns, ulti mately used in varying degrees of complexity, which underlie the unlimited variations of the Twelfth Dynasty and later, existed during the First Intermediate Period’. More recently, see D. Ben Tor, ‘The Relations between Egypt and Palestine in the Middle Kingdom as Reflected by Contemporary Canaanite Scarabs’, IEJ 47 (1997), 165, n. 19. For examples of spiral motifs on scarabs from Egypt assigned to the Middle Kingdom, see ibid. figs 1.4, 1.9, 1.16; of the double ankh motif, figs 1.5, 1.13 14. Both motifs continue in the Palestinian series (ibid. figs. 3 6), the emergence of which Ben Tor (ibid. 164) dates to the Thirteenth Dynasty. 54


With this seal Petrie (Rifeh, pl. 23.1) found a set of seventy-nine small beads, probably of glazed ‘faience’.9 Since each of these ball beads is also perforated, in all likelihood they formed a necklace surrounding the perforated seal-amulet when worn by its owner (pl. 1). The entire set of beads and button-seal was restrung with modern string at the Petrie Museum of Egyptian Archaeology (pls 1, 2). The reconstruction of this arrangement of beads and seal has an impact on how one understands the direction of writing originally intended for its legend.

Figure 1 Drawing of UC 51354 From Petrie, Rifeh, pl. 23.1 Scale: Actual Size

Figure 2 Drawing of UC 51354 Traced from a transparency of a digitized photograph (see pl. 2) and collated with the original 9

Dr Quirke noted that while these beads have not been tested and are badly soiled, glazed ‘faience’ would be their more probable material (email correspondence, April 23, 2007). Concerning all of the beads found in these cemeteries, Petrie (Rifeh, 13) made only a summary statement: ‘But the greater part were purely of the XIIth dynasty style of ball beads of blue or green glaze, carnelian, or amethyst. A few strings of small garnet beads were found…’. 55


2. The Direction of Writing of the Legend Petrie published the legend on this seal-amulet arranged basically as a horizontal line of five signs, with the ankh-signs at the top and bottom (Rifeh, pl. 23.1; fig. 1 above). It seems much more likely, however, that its incised text should be read largely as a vertical column of letters, with the ankh-signs at the sides,10 as that legend would have appeared when the seal and beads were originally strung and presumably worn around the owner’s neck (pl. 2; fig. 2).11 Reconstructing the largely vertical arrangement of this short text has an impact on how one understands the stances of the letters as they have been received or developed from Egyptian scripts.12 3. Palaeography This seal contains five clear letters in Proto-Canaanite script,13 only one of which is damaged. The consonantal values of four of those let10 The positions of the ankh signs provide no clue concerning the direction in which to read this text, since one of them would be upside down in the horizontal arrangement supposed by Petrie (fig. 1 above) and both would be situated sideways in the vertical arrangement that I am proposing (fig. 2). See R.H. Wilkinson (Reading Egyptian Art: A Hieroglyphic Guide to Ancient Egyptian Painting and Sculpture [Lon don 1992], 177) regarding the ankh sign as a symbol for ‘life’ and ‘a potent amulet throughout all periods of Egyptian history’. 11 I am indebted for this interpretation to the arrangement of these objects in the photograph taken by Ana Dori of the Petrie Museum (pl. 2). Following Egyptian practice, early alphabetic letters could be arranged either as vertical columns or hori zontal lines, although the former is much more commonly attested (see the figures in B. Sass, The Genesis of the Alphabet and Its Development in the Second Millennium B.C. [ÄAT 13, Wiesbaden 1988] and the drawings in Origins, 324 99). 12 The long standing issue of whether West Semites borrowed hieroglyphic and hieratic forms of a limited number of Egyptian signs to use as forms of the letters in their consonantal alphabet may now be answered affirmatively. Concerning the two new inscriptions from Wadi el Îol, see: G.J. Hamilton, ‘W.F. Albright and Early Alphabetic Epigraphy’, NEA 65 (2002), 39, n. 14; Darnell, ‘Die frühalphabetischen Inschriften’, 165 71; Sass, ‘Genesis of the Alphabet’, 150; Darnell et al., Two Early Alphabetic Inscriptions, 75 87; and Dobbs Allsopp, ‘Ancient Southwest Scripts’, 496 7. See Origins, 29 268 for a detailed reconstruction the nearly complete repertoire of specific Egyptian forms that were adopted, most of which can be ascertained with certitude in part on the basis of evidence from those recent discoveries or other relo cated epigraphs, the Lahun Heddle Jack (n. 5 above), Sinai 375a and 375c (ibid. 1, 374 8 [new photographs: G.J. Hamilton et al., ‘Three Recently Relocated Early West Semitic Alphabetic Texts: A Photographic Essay’, Maarav 14 (2007), pls 1 8]). 13 Various terms have been used to name these scripts and inscriptions, e.g. ‘Proto Canaanite’; ‘Old Canaanite’; ‘Proto Sinaitic’; ‘proto alphabetic’; ‘early alphabetic’ (see

56


Figure 3: Script Chart 57


ters can be established with confidence. The identification of the consonantal value of the remaining one, the highest letter on this seal, can be narrowed to one of two possibilities. The script chart in fig. 3 has the following format: the name of each letter on this seal;14 the sign numbers for, and drawings of exemplars of actual hieroglyphic and hieratic forms of Egyptian signs15 that served as the certain or probable prototypes for these early West Semitic letter forms; a drawing of each letter on UC 51354 (the letter whose consonantal value is debatable is given twice for ease of comparisons); and drawings of the closest formal parallels to each letter’s form in known Proto- and Old Canaanite scripts.16 The letters will be analysed below in reverse, as they were cut on this seal-amulet (simply to facilitate accurate drawings of them); of the brief overview in Origins, 4). I shall employ the term ‘Proto Canaanite’ in refer ence to early alphabetic scripts and epigraphs written before c. 1400 BCE, no matter whether they were found in Egypt, the western Sinai, or the southern Levant (with no connotations intended regarding the ethnicity of those who wrote them) and ‘Old Canaanite’ for those dated to between c. 1400 to 1050 BCE. 14 I shall employ the Hebrew acrophones, simply because they are the best known to those interested in West Semitic epigraphy. See Origins (29 253) for a reconstruc tion of the original letter names that integrates several epigraphic and traditional sources: an incomplete list of clipped names from Ugarit; mostly clipped forms twice tran scribed into syllabic cuneiform on a recently published tablet in The British Museum (F.M. Cross and J. Huehnergard, ‘The Alphabet on a Late Babylonian Cuneiform School Tablet’, Or 72 [2003], 223 8); recently published transcriptions of several of the full Aramaic acrophones into cuneiform; and the letter names transmitted in Greek, Ethiopic, several levels of Hebrew and Aramaic, and Arabic traditions. With few ex ceptions, these names can now be shown to correlate closely with the earliest graphic forms of the letters and their antecedents in the Egyptian sign list (summarized in Origins, 283 9; contrast the minimal reconstruction recently proposed by W.W. Hallo, ‘Again the Abecedaries’, in C. Cohen, A. Hurvitz, and S.M. Paul [eds], Sefer Moshe: The Moshe Weinfeld Jubilee Volume: Studies in the Bible and the Ancient Near East, Qumran and Post Biblical Judaism [Winona Lake 2004], 285 302). 15 The sources of the drawings of these Egyptian attestations, the majority of which were done by Egyptologists (whose accuracy was checked whenever possible against the originals), will be cited on a letter by letter basis below. The asterisks preceding two Egyptian forms indicate tracings of modern fonts by myself. For a fuller listing of Egyptian hieroglyphic and hieratic forms as they pertain to these and the other early alphabetic letters, see Origins, 29 268, especially fig. 2.77, ‘Summary Script Charts’ (254 68). 16 Apart from two lameds excerpted from Sinai 357, the drawings in the two col umns on the right were done by the present writer, in most instances based on an exami nation of the original inscriptions (the sources for these tracings, mostly from new pho tographs of the glass negatives of the originals in the collection of the Harvard Semitic Museum, are listed in the figures of Appendix 1, Origins, 324 98; any others will be referenced below). The numbers in the second column from the right refer to Proto Canaanite inscriptions from the western Sinai, often termed the ‘Proto Sinaitic’ texts. 58


course, they would have faced the opposite direction on an impression.17 The first letter on UC 51354 could be either a Proto-Canaanite lamed or a waw. When this text is viewed as a vertical column (see above), the form of this letter shows a closed, triangular loop that is attached to an almost horizontal line that juts upwards slightly. The loop on this letter is fairly close to that seen on some hieratic forms of the only certain source of lamed in the Egyptian sign list, V1, ‘coil of rope’.18 One hieratic form of V1 has a slightly open, almost triangular head and another shows a completely closed, circular one.19 The straighter shaft on this letter also has analogues in both hieratic and in semi-cursive hieroglyphic forms of V1.20 Those Egyptian writings show a more oblique angle to their downstrokes, whereas this letter is situated in a more horizontal manner. That basically horizontal stance is also found on one of the Proto-Canaanite lameds of Sinai 357, a mine inscription from Serabi† el-Khadim,21 and twice on Sinai 17 For the first impressions of seal amulets on Egyptian documents dating to the early Middle Kingdom, see James, ‘Ancient Egyptian Seals’, 38. 18 K. Sethe first derived lamed from V1 (see his collected essays: Der Ursprung des Alphabets: Die neuentdeckte Sinaischrift; Zwei Abhandlungen Entstehungs geschichte unserer Schrift [Berlin 1926], 442). See Origins (126 44, figs. 2.35 8) for the exten sive overlap now attested between forms of that Egyptian sign and this early alpha betic letter, which leave the source of only two occurrences of early lamed uncertain (ibid. 134 5, fig. 2.39). 19 The hieratic form set highest in fig. 3 dates to the Eleventh/Twelfth Dynasties and the middle one to the latter period (both reproduced from G. Möller, Hieratische Paläographie: Die äegyptische Buchschrift in ihrer Entwicklung von der fünften Dynastie bis zur römischen Kaiserkeit vol. I [Leipzig 1927; reprinted Osnabrück/Bissendorf 1965], sign number 632. My sincere thanks to Biblio Verlag for permission to repro duce figures from Hieratische Paläographie, which is hereafter cited as Möller, volume number, and sign number. The lowest hieratic form in that figure was traced by the present writer from T.G.H. James, The ÎeÈanakhte Papers and Other Early Middle Kingdom Documents (Publications of the Metropolitan Museum of Art Egyptian Ex pedition 19, New York 1962), Pal. 13. These papyri are now dated to the early Twelfth Dynasty (recently see G. Callender, ‘The Middle Kingdom Renaissance’, in I. Shaw (ed.), The Oxford History of Ancient Egypt [Oxford 2000], 162). 20 This differs from standard hieroglyphic forms of V1 that have reverse curls to their downstrokes (examples of which are given in Origins, fig. 2.35). Both semi cursive forms in fig. 3 were traced from P. Lacau, Sarcophages antérieurs au Nouvel Empire (Catalogue général des antiquités égyptiennes du Musée du Caire, Cairo 1904), pl. 17, a sarcophagus from Meir assigned to the Twelfth Dynasty by H.G. Fischer (‘Archaeological Aspects of Epigraphy and Palaeography’, in R. Caminos and H.G. Fischer (eds), Ancient Egyptian Epigraphy and Palaeography [New York 1976], 42). 21 The drawings of the lameds from Sinai 357 are from I. Beit Arieh, ‘Investigations in Mine L’, Tel Aviv 5 (1978), fig. 6, who has examined this inscription in situ. I would like to thank Prof. Beit Arieh and J. Dekel for permission to reproduce these drawings.

59


352, a steliform panel from the same site.22 One of the Proto-Canaanite lameds on Sinai 352 also shares with the proposed l from Deir Rifa an upward thrust to its largely horizontal stem. That stance would represent a relatively small development away from the positioning of hieratic writings of this letter’s source in Egyptian scripts, V1.23 The only difficulty in identifying this letter as a lamed is the exact point at which the right end of the loop joins the upwardly thrusting line. One would expect it to connect at more of a right angle to judge from those other Proto- and Old Canaanite writings of l and not to continue at a slant; but compare the angle on the slightly ‘open-headed’ hieratic form of V1 excerpted as the highest writing in fig. 3 above. Waw is an alternate reading for this letter. The letter waw descends from T3, a mace usually with a pear-shaped head, exemplified in fig. 3 by a standard hieroglyphic form on a Twelfth Dynasty stele from Abydos.24 Less often T3 has an elliptical head, as illustrated by a slightly damaged semi-cursive hieroglyph on a papyrus dating to the TwelfthThirteenth Dynasties from the Ramesseum.25 Although differing in their stances, both of these hieroglyphic exemplars of mace-signs show ‘V’-shaped connections between their heads and stems that are similar to that found on the first letter of the seal from Deir Rifa. In terms of Proto-Canaanite inscriptions, Sinai 351 witnesses a horizontally situated waw with a pear-shaped head that is formed in a similar, but not identical manner to this letter. The waw on Wadi el-Îol Text 2,26 although manifesting a different shape for its head, has an upswing to its 22

Note the more developed stance of l found on a bowl with Old Canaanite script found at Lachish (drawn from G.R. Driver, Semitic Writing3 [London 1976], pl. 43), which is now assigned to the twelfth century BCE based on a lowered chronology for Ramesses II and III (F.M. Cross, Leaves from an Epigrapher’s Notebook: Collected Pa pers in Hebrew and West Semitic Palaeography and Epigraphy [= LEN; HSS 51, Winona Lake 2003], 209, n. 11). 23 Changes in the stances of early alphabetic letters from their Egyptian graphic prototypes are often greater than developments to their forms (see the summary in Origins, 276 83). 24 See ibid. 86, n. 83 for a brief overview of the history of scholarship about this letter and pp. 86 92 for the graphic overlap between forms of the Egyptian sign T3, ‘mace’ and early West Semitic waw. The first hieroglyphic form of T3 is reproduced from E.A.T.W. Budge et al. (eds), Hieroglyphic Texts from Egyptian Stelae, etc., in the British Museum (London, 1911 22), part 2, pl. 12 hereafter cited as HT, part number, and plate number. I am indebted to the British Museum Company for per mission to reproduce this and other signs from HT, 1 6, the accuracy of each of which has been checked either on the original inscription or on a photograph of it in the archives of the Department of Ancient Egypt and the Sudan, The British Museum. 25 This semi cursive hieroglyph of T3 was traced from A.H. Gardiner, Ramesseum Papyri: Plates (Oxford 1955), pl. 22. 26 See Darnell et al., Two Early Alphabetic Inscriptions, pl. 7. 60


shaft that is much like that which is seen on the first letter of UC 51354. While a horizontal stance for the Egyptian sign T3, ‘mace’ is attested,27 the upward thrust seen on the stem of this letter would represent a minor development away from the positioning of such an Egyptian hieroglyph. Waw is thus a possible reading for this letter. The difficulty in discerning whether the writer of this seal-amulet intended a waw or lamed here should not obscure the fact that either identification of this letter has good graphic parallels in known ProtoCanaanite inscriptions. The formal change from either potential Egyptian prototype, T3, ‘mace’ for waw, or, V1, ‘coil of rope’ for lamed, would represent a relatively minor development in stance, the upswing of its horizontal part. In fact, that small development helps to identify this as a West Semitic letter (and not one of either of those two closelyrelated forms of Egyptian signs).28 One can identify the second letter on this seal-amulet as an early alphabetic qôp with certainty given the similarity of its form with other Proto-Canaanite writings of that letter.29 It has a three-part form: an 27

The horizontal positioning of T3 is much more rarely attested than its usual upright posture; see Origins (87, fig. 2.23) for a horizontal T3 from Wadi Îammamat. 28 One is probably dealing with ‘look alikes’ here (rare for these two letters in early alphabetic scripts [ibid. 134 n. 156]). In many systems of writing some forms come to appear so similar that they can only be differentiated by their linguistic contexts. See Z. Zába (Rock Inscriptions of Lower Nubia [Czechoslovak Institute of Egyptology in Prague and Cairo Publications 1, Prague 1974], sign list: D4 and D21) for exam ples of four such ‘look alikes’ in Egyptian scripts. See Cross, LEN, 11, 29, for exam ples of well known ‘look alikes’ in much later Jewish Aramaic scripts (i.e. d/r, w/y). 29 I am following an insight of Driver (Semitic Writing3, 167 9) regarding the meaning of the acrophonic name *qop as ‘monkey’ back into Egyptian scripts to isolate its probable, but not certain graphic prototype as E32, ‘sacred baboon’ (see Origins, 209 21). E32 has both an upright posture (as exemplified in fig. 3 by an idealized font traced from Wilkinson, Reading Egyptian Art, 73) and a horizontal one where a monkey is portrayed on all four limbs (illustrated by a hieratic form dating to the Twelfth Dynasty [Möller, I, Nachträge]). The form of qôp from Deir Rifa presum ably descends from a horizontal form of E32 in which the monkey’s limbs have been omitted; see Origins (217 8) for discussion of the omission of feet and legs on this letter and on Egyptian signs during the late Middle Kingdom and Second Intermedi ate Period. One awaits the discovery of actual hieroglyphs of E32, particularly semi cursive ones, from Middle Kingdom inscriptions to test out this probable derivation. The identification of these letter forms as qôps has recently been questioned, unneces sarily in my view, by S. Wimmer and S. Wimmer Dweikat, ‘The Alphabetic Texts from Wadi el Hôl: A First Try’, GM (2001), 108, 111, fig. 3 and in a more nuanced fashion by F. Briquel Chatonnet, ‘Les inscriptions protosinaïtiques’, in D. Valbelle and C. Bonnet (eds), Le Sinaï durant l’antiquité et le Moyen Âge: 4000 ans d’histoire pour un désert: actes du colloque “Sinaï” qui s’est tenu l’UNESCO du 19 au 21 Septembre 1997 (Paris 1998), 57. On its occurrence in nqb in the complete Proto Canaanite inscription of Sinai 346a, see especially: W.F. Albright, The Proto Sinaitic Texts and 61


angular head on the left; a slightly larger body on the right; and an oblique line situated above its right end. The latter likely represents a short tail, whose top happens to touch the preceding letter.30 The total shape of this qôp is most comparable to a two-part qôp with a closed neck and ovular body on Sinai 351. Insignificant contrasts can be seen in the formations of the heads of those two letters, the Deir Rifa form showing a blunter profile, that Sinaitic writing being shaped more like a diamond. The short tail on the q of this seal may be compared with less well-preserved occurrences of that feature on Proto-Canaanite qôps from Sinai 346b and 349. A curled tail is well preserved on the reduced form of the certain q on the Grossman Seal, less well so on a new reading of q on the abecedary line of the Ïzbet ∑ar†ah Ostracon.31 Although the third letter on the seal-amulet from Deir Rifa is damaged on its left, it may still be identified with confidence as a ‘viper’ form of nûn with one ‘horn’.32 This nûn shows four parts: a short head set high on the right side; an oblique ‘horn’ of medium length on top of that head; a longer slanting line for its neck; and part of the body that is situated lower, almost on a horizontal plane. Each of those features on this letter matches forms of ‘horned vipers’ found in Egyptian Their Decipherment (HTS 22, Cambridge, MA 1966), 17; D. Pardee, ‘Proto Sinaitic’, OEANE vol. 4 (1997), 354; and, without positing an emendation to that text, Ori gins, 401. On its occurrence in the complete text of the Grossman/Goetze/St Louis Seal, see: F.M. Cross, ‘The Evolution of the Proto Canaanite Alphabet’, BASOR 134 (1954), 21, n. 24 and ‘The Origin and Early Evolution of the Alphabet’, EI 7 (1967), 10* (= LEN, 312, n. 24; 319, n. 16); Albright, The Proto Sinaitic Texts, 11; and Origins, 309, 398. For new photographs of this text and an impression of it, see Hamilton, ‘W.F. Albright’, 38. 30 See fig. 4 below for other letters on this seal (Ì, z), Wadi el Îol Text 2 (t, y), Sinai 352 (t, l), and the Lachish Dagger (∂, l) whose lines occasionally touch or over lap. This phenomenon is relatively rare and should not be taken as evidence of a script system that employs ligatured forms. 31 See the discussion of this new reading in Origins (218, n. 282), where it is pos ited that q was written out of order after r on this beginner’s tablet (compare the certain mistake in the ordering of z and Ì, and perhaps also of p and ¨, in this abecedary; contrast A. Demsky, ‘Abecedaries’, in W.W. Hallo and K.L. Younger [eds], The Context of Scripture: Canonical Compositions from the Biblical World vol. 1 [Leiden 1997], 363). 32 Both the Egyptian sign I9 and its Proto Canaanite descendants show variants with two, one, and no ‘horns’ (Origins, 154 61). Nûn can now be shown to have two sources in the Egyptian sign list, both I9, ‘horned viper’ and I10, ‘cobra in repose’ (with E. Puech, ‘Quelques remarques sur l’alphabet au deuxième millénaire’, Atti del I congresso internazionale di studi fenici e punici vol. 2 [Rome 1983], 579, fig. 8; Darnell et al., Two Early Alphabetic Inscriptions, 94, n. 44; see: Origins, 154 68; against, among others, Sass, Genesis, 125 [contrast 106, n. 76] and W.V. Davies, ‘Egyptian Hieroglyphs’, Reading the Past: Ancient Writing from Cuneiform to the Alphabet [London 1990], 131 2, Tables 2, 3). 62


inscriptions. Except for a slighter longer neck, its basic body shape corresponds closely to a two-‘horned’ writing of hieroglyphic I9 incised on an inscription dating to the First Intermediate Period from Koptos (highest example in fig. 3).33 Its depiction with just one ‘horn’, executed by an oblique line, is analogous to two occurrences of that sign on an Egyptian stele assigned to the Middle Kingdom, whose provenance is unknown.34 The most comparable, well-preserved ProtoCanaanite nûn with one ‘horn’ occurs in Sinai 363, although the neck and body parts have merged on that particular writing. The distinct neck and body on the form from Deir Rifa more closely match those features on a ‘hornless’ nûn from Sinai 347. One can thus find very similar models of nûn in previously recognized Proto-Canaanite epigraphs and trace the clear parts of the nûn on UC 51354 to an Egyptian hieroglyphic prototype, a variant of I9, with no need to postulate any significant changes. This snake-pictograph was incised to face in the opposite direction of the preceding letter, qôp (no matter how one conceives of the stance of the n as having come about — see the next paragraph). The use of opposing orientations for these two letters with fronts and backs is another small indication that one is seeing here an early West Semitic alphabetic, and not an Egyptian inscription.35 One then needs to account for the clearly incised, slightly curving oblique line that is found high on the left side of this letter and the outlined section extending beyond its body located lower on that side. I would speculate that the seal-cutter initially incised the higher line as the beginning of a nûn, its head and neck, intending to cut it to face towards the left (as in the preceding qôp). A small chip of the steatite then appears to have flaked off below that incision (see pl. 2, figs. 2, 3). After that damage apparently occurred, the cutter then incised the complete form of nûn, now oriented towards the right (presumably selecting that option rather than throwing away the already fashioned, decorated, and partially-incised seal). The outlined figure on the lower left could then be accounted for as the preliminary etching for the rest of the tail of the complete nûn, which the seal-cutter did not finish incising (perhaps not wishing to risk extending the damaged section of the steatite).36 While the preceding is a speculative reconstruction of what 33 Traced from H.M. Stewart, Reliefs and Paintings from the Petrie Collection, Part Two: Archaic Period to Second Intermediate Period (Warminster 1979), pl. 11.1. 34 Reproduced from HT, 3, pl. 16. 35 See Origins (279 82) regarding letters with fronts and backs facing opposite directions, on both vertically and horizontally arranged Proto Canaanite texts as developments away from the usage of orientations in Egyptian scripts. 36 Compare the small fragment numbered Sinai 372a incised with a door sign

63


might have happened in the formation of this letter, it would at least provide a tentative explanation for the incised and outlined marks as they may relate to the damage on the left side of this n. No matter how one accounts for that small area, there is no doubt that the larger extant incisions to their right read as a Proto-Canaanite ‘viper’ type of nûn with one ‘horn’. The horizontal placement of the lowest two letters, Ì and z, side by side on a text whose three other letters are arranged as a vertical, singlefile column has parallels in three other Proto-Canaanite columnar inscriptions: Wadi el-Îol Text 2 [bis]; Sinai 352; and the dagger from Lachish (‘double letters’ on each of these vertical texts are highlighted by arrows in fig. 4). The arrangement of two letters side by side on UC 51354 is almost certainly intentional. This arrangement of an alphabetic text likely continues one of the varieties of placing signs on narrow columns of Egyptian inscriptions.37 It represents an archaic arrangement of letters, one that was abandoned fairly early in the transmission of the West Semitic alphabet. Directly below the head of the nûn is situated what will be counted, somewhat arbitrarily, as the fourth letter of this seal. This letter may be assigned the consonantal value of *Ì on the basis of its clear resemblance to linear forms of the Egyptian prototype of that letter, O42, ‘fence outside of primitive shrine O19’, and its similarity to several occurrences of Ìêt in Proto- and Old Canaanite scripts.38 It is situated to the right of and touches the zayin (pl. 2; contrast their separation in the rendering by Petrie, fig. 1 above).39 It has roughly the shape of a modern number sign (#), although there is no extension of the two bars on its right side, which runs into part of the spiral decoration. This Ìêt has rotated a quarter turn from the position of its prototype in Egyptian scripts, semi-cursive hieroglyphic or hieratic forms of O42. whose bottom lines were initially etched but apparently never completed (photo graph: R.F. Butin, ‘The Proto Sinaitic Inscriptions’, HTR 25 [1932], pl. 26; Sass, Genesis, fig. 120). 37 See Origins (Appendix 2, 401 6) concerning the various ways of arranging let ters on early alphabetic texts, from group writings to placements as single files and the likely origins of those arrangements in Egyptian practices. 38 Ibid. 97 102. With this derivation I have extended back into Egyptian writing the understanding of the name(s) of Ì as ‘fence’ by F.M. Cross and T.O. Lambdin, ‘A Ugaritic Abecedary and the Origins of the Proto Canaanite Alphabet’, BASOR 160 (1960) 26 (= Cross, LEN, 316). This required a separation of early writings of Ìêt and dalet, which have often been confused but can now be differentiated on the basis of their Egyptian graphic prototypes (Origins, 61 3). 39 The first and second letters on this seal also touch (fig. 2), as do at least two letters on Wadi el Îol Text 2, Sinai 352, and the Lachish Dagger (fig. 4). 64


Figure 4: Vertical Columns with Two Letters Arranged Horizontally

Linear forms of that sign typically show four verticals with long horizontal lines cutting through their middles and running just below their bases, as illustrated by two semi-cursive hieroglyphic exemplars and a hieratic writing excerpted from texts dating to the Twelfth Dynasty.40 A conservatively derivative ‘four-bar’ form of Ìêt has recently been discovered on Sinai 375a, a plaque from Serabi†.41 A similar form of that letter with only three bars occurs on the dipinto with Old Canaanite 40 The hieroglyphic exemplars are reproduced from HT, 4, pl. 50, a statuette from Deir el BaÌri. The hieratic form is excerpted from Möller, II, 368. 41 See Origins (98 100, figs. 2.26 7) for a discussion of this new reading, based on a fresh collation of the original plaque that was recently relocated in the collection of the Harvard Semitic Museum. This letter in fig. 3 was traced from a transparency of an unpublished colour photograph of Sinai 375a made by Drs M. Lundberg and B. Zuckerman, West Semitic Research: black and white photographs of that inscrip tion are now published in Hamilton et al., ‘Three Recently Relocated Texts’, pls 6, 7.

65


writing from Zarephath.42 Both of those Ìêts have rotated a quarter turn from the position of their Egyptian prototype, O42, as has the Ìêt on the seal-amulet from Deir Rifa.43 After being artificially rotated for formal comparison (marked by an asterisk in fig. 3), the latter shows only two vertical bars. Although atypical, ‘three-bar’ forms of the antecedent hieroglyph O42 do occur, as illustrated by a hieroglyphic writing with a small circle at the top of each ‘thread’ of the fence,44 the further reduction in the Ìêt on UC 51354 to a form with only two bars needs to be acknowledged as a significant West Semitic innovation. This reduction would appear to have been an intentional change (and not just the accidental omission of a stroke), since there was room at either the bottom or top of that figure to have incised another stroke (pl. 2; fig. 2). Although this ‘two-bar’ form of Ìêt is unique in Protoand Old Canaanite scripts, it is valuable in terms of palaeographic typology in that it supplies an intermediate stage between previously attested ‘three-bar’ forms (e.g. the Ìêt from Zarephath) and the heretofore unexplained ‘one-bar’ form of Ìêt, which has almost the shape of an ‘I’, attested on two arrowheads from el-Kha∂r.45 While the unique form and already attested secondary stance of this letter on UC 51354 do not interfere with identifying it as a Ìêt, they represent the largest developments away from Egyptian precursor signs that can be traced for any of the letters in the script of this short text. 42

Traced from F.M. Cross, ‘Early Alphabetic Scripts’, in idem (ed.), Symposia Celebrating the Seventy fifth Anniversary of the Founding of the American Schools of Oriental Research (Cambridge, MA 1979), 113, fig. 2. 43 Quarter turn rotations for at least thirteen letters of the Proto Canaanite alpha bet from stances received from their Egyptian antecedents can now be charted. Those turnings probably represent an extension of a pattern of seven letters that inherited both upright and horizontal postures from their Egyptian prototypes (Origins, 278 9). 44 From a late Middle Kingdom stele of unknown provenance (HT, 3, pl. 28). Also see the variants with three bars numbered O42E, O42F, and O42G in J. Hallof, H. van den Berg and G. Hallof (eds), Hieroglyphica Sign List Liste des Signes Zeichenliste 2 (Publications Interuniversitaires de Recherches Egyptologiques Informatisees 1; Utrecht/Paris 2000), http://www.ccer.theo.uu.nl/ccer/apps/hiero/ hiero.html. An earlier edition of that work is excerpted as an ‘extended library’ in R. Hannig, Großes Handwörterbuch Ägyptisch Deutsch (Mainz 1995), 1117 68. 45 The latter are reproduced, with kind permission, from Cross, LEN, fig. 32.5 (with a small correction made to the Ìêt on the left to indicate a tiny extension of the vertical above the top horizontal, based on a personal examination of the original in the Harvard Semitic Museum). Note the continuation of two bar forms of that letter in chronologically distant inscriptions from Palestine: an incised bowl rim from tenth century BCE Tell Batash/Timnah (G.L. Kelm and A. Mazar, Timnah: A Biblical City in the Sorek Valley [Winona Lake 1995], 111, fig. 6.4); and twice on jars from es Samoa assigned to the same century (Z. Yeivin, ‘The Mysterious Silver Hoard from Eshtemoa’, BAR 13/6 [1987], 43). 66


The fifth letter on this seal-amulet can be identified as a zayin descending conservatively from a hieroglyphic form of T7A, ‘axe of more recent type’.46 T7A could be written with a handle, as in the idealized form traced from Hieroglyphica Sign List, or without,47 as in two exemplars reproduced from a sarcophagus from the Middle Kingdom or Second Intermediate Period.48 The form of zayin on the Deir Rifa seal continues the variety of this axe-sign without a handle. It has the shape of an hourglass set on its side. This is the first well-preserved descendant of a hieroglyphic form T7A to appear in a Proto-Canaanite inscription; another descendant (possibly with a handle) was recently found on Sinai 375a from Serabi† el-Khadim, but that letter is very damaged.49 The abecedary line of the Ïzbet ∑ar†ah Ostracon contains a lightly incised form of zayin of the same type, but with an upright stance.50 The value *z assigned to that letter, written slightly out of order in that abecedary,51 assures its consonantal value on this seal. Line 4 of that Old Canaanite ostracon also contains a z from the same 46

I am following the identification of this variant form of T7, ‘axe’, as T7A in Hieroglyphica Sign List; it is listed as T7* by Gardiner (Egyptian Grammar3 [London 1957], 511). The isolation of both hieratic forms of T7 and hieroglyphic forms of T7A as sources for early alphabetic forms of zayin was achieved through the conver gence of lines of research by several scholars (Origins, 92, n. 95). A clear derivative of a hieratic form of that sign has also been identified on Sinai 345 (ibid. 93 4, fig. 2.24). Only the hieroglyphic alphabetic correspondences are charted in fig. 3. 47 Gardiner, Egyptian Grammar3, 511, T7*, n. 2. 48 Reproduced from Lacau, Sarcophages antérieurs au Nouvel Empire, pl. 39. 49 See Origins (94 5) for a discussion of this new reading, and Hamilton et al., ‘Three Recently Relocated Texts’, pls 6, 7 for photographs of it on p. 19, section 5, end of first full paragraph, after ‘horizontal line’, add: (read either from right to left on the seal itself [pl. 2, fig. 2] or from left to right on an impression of it). 50 This follows the graphic interpretation of that letter, which was lightly incised, by F.M. Cross, ‘Newly Found Inscriptions in Old Canaanite and Early Phoenician Scripts’, BASOR 238 (1980), figs 10, 11 (= idem, LEN, 221, fig. 32.6; 223, fig. 32.7), confirmed on the original by both E. Puech, ‘Origine de l’alphabet’, RB 93 (1986), fig. 3 and the present writer, Origins, fig. 2.25; cf. the more limited perceptions of it by M. Kochavi, ‘An Ostracon of the Period of the Judges from Ïzbet ∑ar†ah’, Tel Aviv 4 (1977), fig. 3, and Sass, Genesis, fig. 175. 51 After the merger of *z and *∂ in transmitting tradents of the alphabet, this ostracon contains only the zayin grapheme (Origins, 95 6). I would doubt that its Ìêt zayin sequence should be taken as evidence of ‘pluriform primers’ (so B. Byrne, ‘The Refuge of Scribalism in Iron I Palestine’, BASOR 345 [2007], 18) even given the same ordering on the recently discovered abecedary from Tel Zayit (R. Tappy et al., ‘An Abecedary of the Mid Tenth Century B.C.E. from the Judean Shephelah’, BASOR 344 [2006], 5 45). Both of these beginners made multiple mistakes in ordering the letters (cf. Byrne, ‘Scribalism’, 21, n. 77). 67


model that is only moderately more developed.52 The presence on UC 51354 of an alphabetic derivative of the Egyptian sign T7A, ‘axe of more recent type’, is chronologically significant. Since that updated form of axe-sign was only introduced into Egyptian scripts in the Twelfth Dynasty,53 West Semites could not have borrowed it to use as the letter zayin before that time. That adoption constitutes one of several chronologically diagnostic palaeographic indicators of when West Semitic alphabetic writing began.54 The borrowing of this letter form also provides a terminus ante quem for when this seal-amulet was incised, no earlier than the beginning of the Twelfth Dynasty. Summary of Palaeographic Findings The script of this seal is Proto-Canaanite. All five letters—l/w, q, n, Ì, z—have close graphic parallels in Proto- and Old Canaanite alphabetic inscriptions, although the consonantal value of one may be debated. The forms of the ‘horned viper’ type of nûn and ‘hourglass’ shape of zayin are primitive, graphically very close to their Egyptian hieroglyphic forebears, I9, ‘horned viper’ and T7A, ‘axe of more recent type’ respectively. One would need to posit only a small change, primarily in stance, to derive the highest letter on this seal from either a hieratic form V1, ‘coil of rope’, thus achieving the reading of lamed, or, from a hieroglyphic form T3, ‘mace’, yielding an identification as waw. While the graphic development of its qôp, ‘monkey’, cannot be charted because its Egyptian predecessor, E32, ‘sacred baboon’, is only probable, one can say that its form with a head, outlined body, and short tail is similar to other writings of that letter in Proto-Canaanite texts from the western Sinai and does not show a reduction to a single-line rendition of the body as seen on the qôps of the Grossman Seal or Ïzbet ∑ar†ah Abecedary. The writing of Ìêt on UC 51354 exhibits the most graphic development, a major change in the loss of at least one bar from its form (or perhaps even two) and a minor change in its quarter-turn rotation from the stance received from its certain Egyptian antecedent, O42, ‘fence’. 52 See the tracings by Puech, ‘Origine’, fig. 3 and the present writer, Origins, fig. 2.25. 53 Gardiner, Egyptian Grammar3, 439, 511, T7*, n. 1; Fischer, ‘Archaeological Aspects of Epigraphy and Palaeography’, 34, n. 31. 54 See Origins (289 90) for a discussion of this and two other palaeographic indi cators of when Proto Canaanite writing started, c. 1900 BCE (similarly and independ ently, Darnell et al., Two Early Alphabetic Inscriptions, 90; contrast Sass, ‘Genesis of the Alphabet’, 157).

68


This short text also manifests both archaic and innovative arrangements of its letters. It continues an archaic arrangement of the letters in which Ìêt and zayin are written side by side in an otherwise singlefile column of characters. In contrast, the two letters with fronts and backs, qôp and nûn, are arranged to face in opposite directions, an arrangement of characters that is against Egyptian norms, but well attested in other Proto-Canaanite inscriptions. 4. Dating Bereft of a report by Petrie of the other contents of the tomb in which this seal-amulet and its accompanying beads were found at Deir Rifa, one has mostly palaeographic means by which to date it. Its terminus ante quem can be established with certainty. This button-shaped seal could not have been cut before the Twelfth Dynasty since one of its letters, zayin, unambiguously descends from an updated hieroglyphic form, T7A, ‘axe of more recent type’ that was only introduced into the Egyptian sign list at that time.55 Assigning this seal to a specific time thereafter depends largely on two palaeographic elements: (a) to what degree its handwriting has developed from Egyptian scripts; and (b) which other early alphabet text(s) manifest(s) a similar rate of change. The short text on UC 51354 shows one major and three minor developments from Egyptian scripts: one letter, Ì, exhibits a major change, the loss of at least one bar from its Egyptian prototype; Ì has also developed in a minor way through rotation from such an antecedent; the stance of another letter, l/w, is also slightly different from either of its most likely Egyptian precursors; and the two letters with fronts and backs, q and n, were written to face in opposite directions (the latter perhaps made necessary by chipping in part of the steatite but nonetheless tolerated by the seal-cutter on the final product which was finished by glazing). The relatively small number of developments in the script of UC 51354 is most comparable with the Proto-Canaanite handwriting on the dagger from Lachish, which also has five letters: One major development (a shortened form of l); and three minor ones (a slightly developed stance to l; mirroring of a hieratic form of n; a t

55 See the references in n. 53. The beginning of the Twelfth Dynasty has been recently estimated at c. 1938 BCE (Baines and Malek, Cultural Atlas2, 9, 40), or, c. 1985 BCE (I. Shaw [ed.], The Oxford History of Ancient Egypt [Oxford 2000], 480, 459 [a select bibliography on the difficulties of Middle Kingdom chronology]).

69


that has rotated from the position of its hieratic forebear).56 Found in Tomb 1502 at Lachish, that dagger can now be dated by co-ordinating typological analyses of the ceramics found with it and its early alphabetic script to c. 1725 BCE (± 25 years).57 Given the large formal change seen in the Ì of UC 51354, I would estimate that the legend on this seal was inscribed at least somewhat later than the inscription on that dagger, c. 1700 BCE. Yet one also needs to acknowledge that unlike that dagger, this sealamulet’s discovery was so poorly recorded and published that one is dealing with an artefact that is almost without an archaeological provenance (so Quirke, personal communication). I would indicate that unfortunate reality by leaving a very wide margin for error in estimating its time of writing, c. 1700 BCE (± 150 years). It was most likely cut in the middle of Egypt’s Thirteenth Dynasty, a time when we know that other West Semites began having seals engraved in Egypt and scarab-shaped amulets start to appear in Palestine.58 Yet this seal could have been inscribed as early as the two early alphabetic graffiti from Wadi el-Îol or the West Semitic name on the heddle jack from Lahun, c. 1850–1700 BCE.59 It seems very unlikely, however, that such a highquality seal of glazed steatite and probably ‘faience’ beads would have been acquired by a West Semite and incised with Proto-Canaanite letters after the expulsion of the Semitic Hyksos from Egypt, c. 1550 BCE (or shortly thereafter).60

56

Origins, 303 4. See especially Tubb’s analysis of the contents of that tomb (ibid. n. 37), includ ing the elimination of a scarab thought to be of Neferhotep II, whose presence had previously required a later dating for the whole assemblage. The continuing scepti cism about the Proto Canaanite classification of the script on this dagger by Sass (Genesis, 54; ‘Genesis of the alphabet’, 150) is no longer warranted; for good parallels in Proto Canaanite and Egyptian scripts for each letter in the order in which it oc curs, see Origins, 147 9; 132, n. 153; 223 5; 157 60; 248 51. For a new proposal as to the meaning of this text in West Semitic, see ibid. 149, n. 181. The secure assignment of this artefact to the Middle Bronze Age, which Sass acknowledges (ibid.), makes his lowering of the origin of alphabetic writing to c. 1300 BCE not just unlikely but impossible (see also nn. 4, 5 above). 58 For the seal of an ‘Asiatic’ in Egyptian from the Thirteenth Dynasty, see M. Bietak, Avaris, The Capital of the Hyksos, Recent Finds at Tell el Dabä (London 1996), 41, fig. 35.12 and Section 6 below. Regarding the emergence of scarabs in Palestine at that time, see Ben Tor, ‘Relations between Egypt and Palestine’, 164. 59 Also see the references in nn. 4 and 5. 60 The end of the Second Intermediate Period has recently been estimated at c. 1550 BCE, specifically 1532 28 BCE for the conquest of Avaris (Bourriau, ‘Second Intermediate Period’, 185), or, c. 1520 BCE (Baines and Malek, Cultural Atlas2, 9, 42). 57

70


5. A Proposed Linguistic Decipherment I would propose reading the text of this button-seal as three parts: probably a prepositional lamed indicating possession, ‘(Belonging) to’ (much less likely an initial copulative or emphatic waw, ‘and, even, also’); a m. proper name ‘Qn’; and a m. sg. G participle, Ìz used as a substantive to identify the owner’s title or occupation, ‘(the) Seer’. Under this proposed reading, the letters indicating the lamed of possession and the name, Qn, were arranged as a vertical column, whereas the characters indicating his title, Ìz, ‘(the) Seer’, were arranged as a horizontal line. Given two possible readings for the first letter in terms of paleography, l or w, the more likely reading on a seal-amulet is l, understood as a prepositional lamed specifically indicating possession.61 An exact parallel for such a usage can be found on the complete Proto-Canaanite text of the Grossman Seal from c. 1400 BCE (± 100 years): l-b†/s ¨rqy, ‘(Belonging) to B†/s, (the) Arkite’.62 Reading the first letter of UC 51354 as a w, while possible in terms of palaeography, would yield a very improbable use of common Semitic waw, either as a copulative ‘and’, or, as an emphatic-declarative ‘also, even, namely’ at the beginning of an inscribed seal.63 But to communicate that small uncertainty I will render the translation of the much more linguistically likely possessive lamed in italics: ‘(Belonging) to’. I would take the second and third letters as constituting a m. sg. proper name, qn, related to personal, clan, and place names attested in Biblical Hebrew and Epigraphic South Arabian.64 The Biblical He61

On this common Semitic preposition, see especially the compilations in HALOT (2), 507 and DUL (= G. del Olmo Lete and J. Sanmartín, A Dictionary of the Ugaritic Language in the Alphabetic Tradition [2 vols, Leiden 2003]), 481. See James (‘Ancient Egyptian Seals’, 38) regarding impressions of some inscribed amulets used as seals early in the Middle Kingdom. 62 I am following the boustrophedon reading of this seal proposed conjointly by Cross, ‘Evolution’, n. 24 (= idem, LEN, 312, n. 24) and Albright, The Proto Sinaitic Texts, 11, against the several directions for reading it envisioned by Puech, ‘Origine’, 182 (most of which fail because they require turning its legend at right angles to the iconographic scene). On the dating of this unprovenanced seal, see Origins, 309. I would follow a similar possessive use of lamed that has been proposed by A.G. Lundin for the earlier Lachish Dagger (as cited by B.E. Colless, ‘The Proto Alphabetic In scriptions of Canaan’, AbrN 29 [1991], 35), although we differ in our understandings of the personal name that follows it (Origins, 149, n. 181). 63 For those translations in longer West Semitic narrative texts, see HALOT (1), 258 and DUL, 940 1. 64 Taking the second, third and fifth letters together to form an alternate reading, the name qnz, ‘Kenaz’, is blocked because of the spelling of the cognate name qn∂ at 71


brew cognates are: qayin, ‘Cain’, the name of a character in biblical mythological and genealogical texts (e.g. Gen. 4:1, 17); the place name haqqayin (Josh. 15:57); and the clan or tribal name that, although twice spelled qayin/qayin (Num. 24:22; Judg. 4:11), is usually transmitted with a nisbeh ending, haqqênî, ‘the Kenites’ (e.g. Num. 24:21).65 As is also seen in the vocalization of haqqênî, this proposed name from Deir Rifa likely transmits a form in which an original *ay diphthong was contracted: *qayn- > *qên-.66 Halpern noted the closest cognates to the biblical evidence: ‘The name Cain is attested both as a Nabatean personal name (J.T. Milik and J. Starcky apud ARNA 146:23; 157:107:3; 108:2) and, with a preformative ’alep, as a Sabean clan name (DOSA, 454)’.67 These names are usually related to later Aramaic and Arabic nouns meaning ‘smith’ or ‘metal-worker’,68 although appeals are sometimes made either to words written with the same consonants that refer to aspects of singing, particularly laments,69 or to another meaning of qaynu/qayna in Arabic, ‘slave’.70 While I would favour a meaning of ‘metal-worker’ or ‘smith’ for qn in a West Semitic text from the first half of the second millennium BCE,71 the minimal Ugarit (DUL, 705 6), although such a personal and clan name is attested in biblical literature (BDB, 889; HALOT [3], 1114) and Punic (F.L. Benz, Personal Names in Phoenician and Punic Inscriptions [Rome 1972], 178, 405). 65 For the full listings, see BDB, 883 4, HALOT (3), 1097 8. See also the exten sive discussions by R.S. Hess, ‘Cain’, in ABD vol. 1 [1992], 806 7, and B. Halpern, ‘Kenites’, in ibid. vol. 4, 17 22 (and the literature therein). 66 Note that the *ay diphthong is consistently contracted in tradents of three acrophonic letter names, *bêt , *Ìê† , and *mêm (Origins, 285), which may well have originated among West Semites at a time and place close to the writing of this seer’s name. More directly, perhaps compare the vocalization of the personal name qen in the Samaritan Pentateuch and the spellings of some of the various South Semitic names cited in HALOT (3), 1097 (qn, qnt, qn’l). The usually meticulous Hess (‘Cain’, 806) erred in interpreting Biblical Hebrew qáyin as a *qatil form; the accent on the first syllable assures that it is a standard Masoretic triphthongal reflex of an original *qayl formation. 67 B. Halpern (‘Kenites’, 18), citing F.V. Winnett and W.L. Reed (eds), Ancient Records from North Arabia (Toronto 1970) and J. Biella, Dictionary of Old South Arabic (HSS 25, Chico 1982). 68 See especially: BDB, 883; Hess, ‘Cain’, 806; Halpern, ‘Kenites’, 17 8; and L.E. Stager, ‘Forging an Identity: The Emergence of Ancient Israel’, in M.D. Coogan (ed.), The Oxford History of the Biblical World (Oxford 2000), 147 (the latter compar ing both Aramaic qayyan and the biblical description of a mythological figure Tubal cain, ‘who made all kinds of bronze and iron tools’ [Gen. 4:22]). 69 In particular, see Hess, ‘Cain’, 806 and Halpern, ‘Kenites’, 17 18. 70 See especially the references for this option compiled in HALOT (3), 1097 and BDB, 883. 71 I would base that preference on a tentative understanding of qnm at the begin ning of a complete text from Serabi†, Sinai 380 (Origins, 385 6), as ‘metal workers’. 72


context of this noun used as a proper name on a seal does not allow one to sort out those semantic options with any certainty. Thus I will transliterate the name as Qn. F.M. Cross has recommended reading the lowest two letters, Ìz, as a m. sg. G active participle traced to the root *Ìzy, ‘to see’, used as a substantive to indicate the occupation of this seal’s owner, ‘(the) Seer’.72 Although one would expect such a third-weak form to be spelled with a final y if the nominative case ending was /u/, i.e. *Ìaziyu, Cross would explain its absence here as being due to a genitive /i/ following the prepositional lamed, contracting /*iyi/ to /*î/, and left unmarked in the orthography of this period, i.e *la/i-Ìaziyi > *la/i-Ìazî.73 Since words traced to the root *Ìzy are well documented in later West Semitic languages and occupations following personal names are well attested on seals belonging to West Semites, I shall follow Cross’s proposal for deciphering this word as ‘(the) Seer’ on UC 51354.74 Participial forms traced to the root *Ìzy are well known from both Biblical Hebrew and epigraphic sources in Aramaic and a Transjordanian dialect.75 Of some 129 occurrences of words assigned to this root in For evidence of metal working in the mines at that site, above the entrance to one of which Sinai 380 was written (Sass, Genesis, figs. 100, 103, 104), see I. Beit Arieh, ‘Serâbî† el Khâdim: New Metallurgical and Chronological Aspects’, Levant 17 (1985), 89 116. Halpern (‘Kenites’, 18 21) treated some of the biblical evidence as possibly re flecting persons, places, and groups of the late second millennium BCE, but empha sized that most biblical traditions about them stem from the Iron Age, c. 700 BCE or later. Hess (‘Cain’, 806) noted that I. Ephäl (The Ancient Arabs [Jerusalem 1982], 194, 211 12, 226 7) would date some of the South Semitic epigraphic attestations to as early as the fifth century BCE. 72 Email correspondence, September 27, 2006. Prof. Cross corrected my attempt to read the last two letters in the opposite direction, zÌ, as part of a short sentence name, ‘(The clan) Cain moves (away)’ (‘to move’ or ‘to move away’ are meanings assigned to *zwÌ/zÌÌ in later Aramaic, especially Syriac, and Arabic). That proposal fails given the lack of attestations of that verb in the West Semitic onomasticon. Cross’s superior interpretation is communicated with his kind permission. 73 Email correspondence, September 29, 2006. 74 The Arabic cognate Ìazay, ‘perceive with an inner vision’ (BDB, 302) assures that one is dealing in this root with an etymological *Ì (and not *Ì) and *z (not *∂). To our knowledge, *Ì is differentiated from *Ì and *z from *∂ in Proto Canaanite inscriptions (for a discussion of their different graphemes and originally distinct letter names, see Origins, 97 102, 57 60; 92 7, 145 54). 75 Although dated, see the still valuable review by A. Jepsen, ‘chazah’, TDOT (IV), 280 90. For parallels of titles on seals and sealings of the much later Iron Age, see B. Sass in N. Avigad, Corpus of West Semitic Stamp Seals (Jerusalem 1997), 466 8, espe cially four or five titles composed of m. sg. G active participles, hspr, ‘the Scribe’, hÒrp, ‘the Goldsmith’, (h)khn, ‘(the) Priest’, hrp’, ‘the Healer’, and probably lÌs ‘Charmer/ Whisperer of incantations’ (467 8, 509). 73


Biblical Hebrew, seventeen are G participial forms, Ìozeh, Ìozîm, ‘seer(s)’.76 Although words containing that root occur at all levels of the biblical texts, from archaic poems usually assigned to the premonarchic period (e.g. Num. 24:4, 16), to prophetic texts traced to the monarchic era (e.g. Hos. 12:11; Mic. 3:7), to figures of the Exile, even if their writings were edited later (e.g. Ezek. 12:22, 27), to works that were clearly written in the post-exilic period (e.g. 1 Chron. 21:9), they are almost always seen as loanwords from Aramaic into Hebrew.77 The earliest instance of a G participial form traced to *Ìzy in an Aramaic inscription occurs on the Zakkur stele, dated to just after 800 BCE.78 Millard has recently translated the relevant portion of that ‘memorial’ text: ‘Ba¨lshamayn [spoke] to me through seers [Ìzyn] and diviners’.79 In the Deir Älla Plaster Text dating to the same century,80 Balaam son of Beor is emphatically termed a ‘seer’: Ìzh ’lhn h’, ‘a seer of the gods was he’.81 These selections from epigraphic Transjordanian, Aramaic and Biblical Hebrew texts may suffice to illustrate the presence of G participial forms of the verb *Ìzy/Ìzh in West Semitic languages of the first millennium BCE and thus the likely presence of such a form in the West Semitic lexicon of the second millennium BCE as proposed for the legend of this seal-amulet from Deir Rifa. 76 Jepsen, ‘chazah’, 282. For more recent literature particularly concerning attesta tions in various levels of Aramaic, see HALOT (5), 1872. 77 Jepsen, ‘chazah’, 285 9. For a probable occurrence of the cognate noun Ìzn, ‘vision’ (less likely ‘inspector’ or ‘supervisor’) in a broken context in a Hebrew inscrip tion from Lachish, see F.W. Dobbs Allsopp et al., Hebrew Inscriptions: Texts from the Biblical Period of the Monarchy with Concordance (New Haven 2005), 337 8 and J.P. Kang, ‘A Dictionary of Epigraphic Hebrew’, unpublished Ph.D. dissertation (Princeton Theological Seminary 2006), 136. I would like to thank Dr Kang for both of these references. 78 Recently see A. Millard, ‘The Inscription of Zakkur, King of Hamath’, in W.W. Hallo and K.L. Younger (eds), The Context of Scripture: Monumental Inscrip tions from the Biblical World vol. 2 (Leiden 2000), 155 and C. L. Seow, ‘West Semitic Sources’, in M. Nissinen, Prophets and Prophecy in the Ancient near East (SBLWAW 12, Atlanta 2003), 203 7, n. a. 79 Millard, ‘Zakkur’, 155. 80 See the various dates reviewed by M. Dijkstra, ‘Is Balaam also among the Prophets?’ JBL 114 (1995), 45 6, B. Levine, ‘The Deir Älla Plaster Inscriptions’, in W.W. Hallo and K.L. Younger (eds), The Context of Scripture: Monumental Inscrip tions from the Biblical World vol. 2 (Leiden 2000), 141, and Seow, ‘West Semitic Sources’, 207. 81 Similarly, Levine, ‘Deir Älla’, 142; Dijkstra, ‘Balaam’, 47. Both a cognate verb, wyÌz, ‘he saw’ and noun, mÌzh, ‘a vision’ can be reconstructed securely at the end of line 1 (with J. Hackett, The Balaam Text from Deir Älla [HSM 31, Chico 1984], 33, against, among others, Dijkstra ‘Balaam’, 52, n. 20 and Seow, ‘West Semitic Sources’, 209 10).

74


6. Functions Classifying UC 51354 as a seal-amulet owned by a West Semite which reads, l qn Ìz, ‘(Belonging) to Qn, (the) Seer’, fits both the sigillary and magical functions that are known for this type of object during the wide period in which it is estimated to have been inscribed, c. 1700 BCE (± 150 years). Its sigillary function finds its best parallels with Egyptian name seals. Its magical function has its closest correspondences in funerary amulets found in the burials of other West Semites in the eastern Delta and Palestine. Egyptian parallels to its legend suggest that this seal was very likely owned by a person of high social status. The closest parallel naming specifically a West Semite and giving a title on a seal-amulet found in Egypt occurs on a scarab written in Egyptian found in an extremely wealthy burial at Tell el-Dabä that is assigned to the Thirteenth Dynasty: ‘The Deputy Treasurer, Äamu (= “the Asiatic”)’.82 James provided an overview of the use of Egyptian seals with names and titles: ‘During the Twelfth and Thirteenth Dynasties personal signet scarabs were made in large numbers. They are usually of glazed steatite and bear official titles and the names of the title holders. Their use is confirmed by the discovery of many sealings with comparable impressions in secular sites in Egypt and the Sudan’.83 James differentiated such name seals from decorative scarabs or related objects that probably belonged to ‘lower-grade officials who were not qualified to possess seals with their personal names and titles included in the legends’.84 These Egyptian parallels would suggest that, in all likelihood, the seer Qn was considered among the social élite of his time.85 In a more speculative vein, one then needs to ask what this name seal was meant to seal. During the Middle Kingdom, Egyptian seals were used to seal wine jars, stoppers, boxes and other containers, as 82 Bourriau, ‘Second Intermediate Period’, 191; Bietak, Avaris, 41, fig. 35.1. From that site see as well a damaged amethyst scarab mounted on a gold ring and inscribed with both a name and title, apparently in corrupt hieroglyphs (ibid. 26, fig. 22.1). Cylinder seals from the Levant and Mesopotamia inscribed with names and titles are also common; for titles indicating ‘Musician’, ‘Scribe’, ‘Servant’, ‘Royal constable’, and ‘Sergeant’, see D. Collon, ‘Ancient Near Eastern Seals’, in idem (ed.), 7000 Years of Seals (London 1997), 17, 24 9, figs. 1/6, 1/17 19. 83 James, ‘Ancient Egyptian Seals’, 35. He chose to illustrate that pattern by a steatite scarab from the cemetery at el Lisht (late Middle Kingdom, c. 1800 BCE) whose legend reads: ‘Overseer of dispute, Iy’ (ibid. 42, fig. 2/10a). 84 Ibid. 40. 85 See Bourriau (‘Second Intermediate Period’, 190) for indications from archi tectural and burial patterns of increasing social stratification among West Semites at Tell el Dabä.

75


well as documents of rolled or folded papyrus (e.g. the ÎeÈanakhte papers and related documents).86 But since the legend of UC 51354 is written in Proto-Canaanite script and its language is West Semitic, it could not readily have fulfilled those functions in an Egyptian speaking-context such as Deir Rifa. One has to suppose that either it functioned within such a linguistic context in a way comparable to the many Egyptian seals without legends, simply as a recognizable device,87 and/or that it served another purpose. This seal’s West Semitic legend strongly suggests that it may well have been employed to seal documents that were also written in Proto-Canaanite alphabetic writing intended to be read by other members of the West Semitic-speaking communities in Egypt.88 If that line of speculation is cogent, then this name seal may bear indirect witness to longer early West Semitic alphabetic documents that were written on more perishable materials, such as papyrus, that have not survived to modern times.89 When Qn was alive, he likely wore this alphabetic seal-amulet and beads as a necklace and used it as his personal seal. After Qn’s death the second function of this artefact, as a funerary amulet, then appears to have come into effect. This amulet-seal necklace was found in one of the cemeteries at Deir Rifa, presumably with the remains of Qn. That would represent the same funerary context in which amulet-seals dating from the beginning of the Thirteenth Dynasty, habitually in the shape of scarabs, are found in the tombs of other West Semites at Tell el-Dabä in the eastern Delta and only slightly later from many sites in Canaan.90 86

James, ‘Ancient Egyptian Seals’, 38. On the use of seal amulets without legends, the majority of which are shaped as scarab beetles, James commented (ibid. 34): ‘Engraved with a distinctive device it could be used as an individual badge of possession, a very suitable instrument for marking and sealing purposes…It did not have to be “read” in its strictest sense, but simply recognised’. 88 For a map showing the dozen sites in Egypt with Palestinian Middle Bronze Age cultural finds, see Baines and Malek, Cultural Atlas2, 41. Also see n. 91 below. 89 The longest early alphabetic rock inscription to surface to date is Sinai 349, originally with seven lines (Origins, 339 40). For an overview of semi cursive hiero glyphic and hieratic documents on papyrus, in particular see R. Parkinson and S. Quirke, Papyrus (London 1995). 90 Bietak, Avaris, 41. Ben Tor (‘Relations between Egypt and Palestine’, 187 8) concluded that: ‘[T]he custom of using scarabs primarily as funerary amulets was transmitted by the Asiatics who settled in the eastern Delta during the late Middle Kingdom. The Tell el Dabä material culture clearly demonstrates that these settlers adopted Egyptian customs, including the use of scarabs as funerary amulets, and most probably initiated the import of the latter into Palestine’. One may now con sider looking further south in Egypt for evidence of the adoption of that custom. 87

76


7. Historical Significance and Possible Context This seal-amulet is historically significant in three different ways. First, this artefact provides documentary evidence for the presence of a West Semite named Qn in Egypt during the late Middle Kingdom or Second Intermediate Periods.91 This will likely be of interest to biblical scholars seeking philological parallels to the names qyn, ‘Cain’ and hqyny, ‘the Kenites’. Secondly, the pattern of its legend would indicate that its owner, termed a seer, was very likely considered among the social élite of his time, thus providing further evidence of social stratification among West Semites in Egypt either near the end of the Middle Kingdom or during the Second Intermediate Period.92 This may also be of interest to those charting the social status of prophetic figures among West Semites during the Middle Bronze Age and later. Thirdly, this incised seal from Deir Rifa constitutes the fourth witness to early alphabetic writing from Egypt proper, joining the wooden weaving instrument inscribed with its owner’s name found at Lahun (a site that is also technically in Middle Egypt but much further north) and the two graffiti in Proto-Canaanite script recently discovered at Wadi el-Îol (on a track through the Qena Bend of the Nile in Southern Upper Egypt).93 Surprisingly, each of these alphabetic inscriptions comes from a location outside of the Nile Delta, the area with the densest concentration of West Semites in the late Middle Kingdom and Second Intermediate Period. Given the surprising place of discovery of this seal-amulet, one might also wish to speculate on what a West Semitic visionary was doing at Deir Rifa during one of those periods. There is a reasonable chance that Qn may have been associated with the Nubians known from their distinctive ‘pan-graves’, the ceramic contents of which indicate their presence at this site from early in the Thirteenth to the mid-Fifteenth Dynasties.94 Bourriau cautiously speculated that when Egypt was 91

For a recent overview of Egypt’s relations with the southern Levant from pre dynastic times to the First Intermediate Period, in particular see G. Mumford, ‘Tell Ras Budran (Site 345): Defining Egypt’s Eastern Frontier and Mining Operations in South Sinai during the Late Old Kingdom (Early EBIV/MBI)’, BASOR 342 (2006), 52 8. On the documented presence of West Semites in Egypt from late in the First Intermediate Period and Middle Kingdom, most recently see Darnell et al., Two Early Alphabetic Inscriptions, 87 90, 97 9, nn. 106 17. Concerning West Semites in Egypt from the Thirteenth to early Eighteenth Dynasties, see especially Bourriau, ‘Second Intermediate Period’. 92 Ibid. 190. 93 See nn. 4, 5. 94 Ibid. 190. See also J.D. Bourriau, ‘Some Archaeological Notes on the Kamose 77


divided during the Second Intermediate Period, the Nubians at Deir Rifa guarded the southern extremity of the area controlled by West Semites for the Hyksos, whereas another group of Nubians located at Mostagedda, a site almost directly across from Deir Rifa on the opposite bank of the Nile, guarded the northern limits of the territory controlled by Egyptians for Thebes.95 If the chronological estimates are accurate for the presence of those Nubian mercenaries particularly at Deir Rifa, c. 1750–1600 BCE, arrived at mostly through ceramic typological means, and for the legend of this seal, c. 1700 BCE (± 150 years), arrived at largely by means of palaeographic typological analysis, it is possible to speculate that the seer Qn may have been part of that sustained military presence guarding the southern border of the Egyptian territory ruled by other West Semites. While admittedly this is a conjectural reconstruction of Qn’s reason for being at Deir Rifa, consulting seers, diviners, and oracles during crises, especially times of military conflict, was common throughout the ancient Near East.96 On a more certain level, this button-seal provides evidence for an attachment by its owner to two cultures. The attachment of Qn to his West Semitic culture can be seen in his choice of having his personal name seal written in Proto-Canaanite alphabetic script with a legend that includes his title, Ìz, ‘(the) Seer’, in West Semitic. Stated negatively, he did not choose to have that title translated and written in Egyptian, or, his Semitic name rendered in Egyptian syllabic orthography. But Qn also had an attachment to aspects of Egyptian culture as can be seen in his selection of a distinctively Egyptian button-form for his seal-amulet, one that was decorated with the Egyptian motifs of spirals and ankh-signs. In very small ways those choices witness both the deeply ‘Egyptianizing’ tendencies among West Semites, particularly those living in Egypt during the late Middle Kingdom and Second Intermediate Period, and a prominent West Semite’s assertion of his own cultural and linguistic heritage.97 If others accept my classification of the handwriting of UC 51354 as Proto-Canaanite script and consider the linguistic decipherment of its legend proposed above as viable, the seal-amulet belonging to Qn, ‘the Seer’ would represent the earliest seal incised with alphabetic Texts’, in A. Leahy and J. Tait (eds), Studies on Ancient Egypt in Honour of H.S. Smith (Egypt Exploration Society Occasional Publications 13, London 1999), 43 8. I would like to thank Ms Bourriau for the latter reference. 95 Bourriau, ‘Second Intermediate Period’, 190. 96 Millard, ‘Zakkur’, 155, n. 10 (with literature). 97 For other examples of those tendencies, see Origins, 320. 78


letters that is presently attested, assigned to c. 1700 BCE (± 150 years), a date lamentably arrived at almost exclusively on palaeographic grounds.98

98

Previously, the cylindrical Grossman Seal has been the earliest seal certainly inscribed with alphabetic letters (see n. 62 above). I remain uncertain about the clas sification of the writing on a seal amulet from Palestine proposed by K.A. Kitchen, ‘An Early West Semitic Epigraph on a Scarab from Tell Abu Zureiq?’, IEJ 39 (1989), 278 80. The problem in that proposal is palaeographic his proposed lamed does not correspond to any early forms attested for that well known letter (cf. Origins, 128 9, figs. 2.35 8). 79

THEME AND VARIATION IN PSALM 111: PHRASE AND FOOT IN GENERATIVE-METRICAL PERSPECTIVE1 VINCENT DECAEN UNIVERSITY OF TORONTO

Abstract This programmatic paper approaches the metre of biblical poetry as a problem in generative metrics. Recalling the earlier proposal of Kury¥owicz (1972, 1975), it is argued that the organizational princi ples of the Tiberian liturgical chant (syllable, foot, and, crucially, pho nological phrase) are also the metrical principles of biblical poetry or at least that of Job, Proverbs and a fair portion of the Psalms. When the musical transformations of the poetic accent system are taken into account, Psalm 111 conforms to Kury¥owicz’s 2+2 phrases per verse. However, the 2+2 analysis only scratches the surface: the distribution of foot , word and line types in Psalm 111 is also regu lated by prosodic principles. 1 A special word of thanks to Elan Dresher, to whom this paper is dedicated, and whose continued support and encouragement through the years have kept me working away at generative grammar and Tiberian Hebrew phonology (or as I pre fer, ‘generative Masoretics’). It should be obvious that Dresher’s seminal 1994 study is my constant inspiration throughout. I must also recognize John Hobbins and his invaluable online resources, espe cially his detailed bibliographies and history of research (see http://ancienthebrew poetry.typepad.com). He was quick with voluminous and insightful feedback on earlier drafts, and the study is many times better as a result. I am especially grateful for his pointing out those crucial footnotes in Dhorme’s commentary on Job, and also for patiently explaining the details of Fokkelman’s syllable counting. Thanks also to Jim Price who, in addition to offering feedback on an earlier draft, generously supplied me with data for nesiga before silluq. Specific insights are noted in the relevant footnotes. Of course, these gentlemen can in no way be held responsible for the views let alone the errors contained herein. I should also thank an anonymous reviewer for helpful suggestions that have been incorporated in this final version, especially for alerting me to Loretz’s count for Psalm 111. Finally, my work is made possible in part through a generous grant from the nonprofit GRAMCORD Institute, dedicated to computer assisted analysis of bibli cal languages ( http://www.gramcord.org).

81

THEME AND VARIATION IN PSALM 111

It seems that in the present stage of research the accentuation system may not be excluded from the competition of prosodical systems of Old Hebrew poetry, but its application should be substantially limited, perhaps to the poetical works from the first half of the first Millenium B.C. (Segert 1960: 286).

‘The psalms are for singing’, intones Mowinckel, ‘and singing implies a constriction of the rhythm called metre’ (1962, II: 159). This declaration echoes that of Gustav Hölscher, from whom Mowinckel borrowed his metrical theory wholesale: ‘Dichtung ist metrisch geformte Rede’ (1920: 93). Mowinckel’s bold statement should be amended, however, to add a subtle but crucial distinction: the psalms are for chanting. Jacobson distinguishes ‘logogenic’ liturgical chanting, with a rhythm ‘determined by the natural cadences of speech’, from ‘melogenic’ song in which ‘the words are fitted to the music, rather than vice versa’ (2002: 14, citing C. Sachs). Similarly, Dresher describes liturgical chant, both in its Tiberian and Gregorian forms, as ‘a stylized form of intonation’ projected ‘from the tendencies inherent in ordinary speech’ (forthcoming: 1). The psalms are for chanting, then, and Tiberian chanting implies the delimitation and hierarchical organization of phonological phrases to which one of four ‘tropes’ (fixed musical phrases) is assigned (Jacobson 2002, Portnoy and Wolff 2000, 2001; see further Weil 1995). The phonological phrase, the basic building block of the liturgical chant, is at the very same time the primary domain of Tiberian Hebrew (TH) postlexical phonology: the so-called ‘sandhi rules’ of spirantization; external gemination; and, crucially, the rhythm rule or nesiga (Dresher 1994: esp. §3.2, 10–11; see further Revell 1987) that reorganizes syllables into (typically) binary feet,2 reinforcing the characteristic TH iambic rhythm.3 It is worth underscoring this point, since the phonological phrase rarely gets the attention deserved by its fundamental role in TH phonology. 2 ´ Khan furnishes a prototypical example, tô´äßo†êhém (Ez. 6:9); see further Khan (1987: 54ff ). 3 Specifically, a ‘loose iambic’ rhythm. The loose iambic foot is analogous to deuterium (one proton, one neutron) with its two isotopes, tritium (one proton, two neutrons) and standard hydrogen (one proton, zero neutrons). For a general review of the theory of iambic meter and its many variations, see, e.g., Steele (1999). For comparable loose iambic verse in German and Russian, see further Tarlinskaja (1993); see also Segert (1960: 289).

82


Moreover, ‘poets “measure” those elements of their language that are most essential to its nature and structure’ (Steele 1999: 13, emphasis added). It is a commonplace that the verse of a tone language like Chinese organizes tones; and that the verse of quantity-sensitive languages like ancient Sanskrit, Greek and Latin organizes by long versus short syllables, whereas quantity-insensitive languages like French, Italian and Japanese simply count syllables. It only stands to reason in this light that the phonological phrase, both the primary domain for TH phonological rules and the basic building block of the liturgical chant, should play the key organizing role in ancient Hebrew poetry. In this perspective, it is curious that Mowinckel apparently never refers to the special mode of chanting devised for the Pss, together with the poetry of Proverbs and Job (Price 1990, II: Chs. 10ff; 1996, vol. 5; see also Flender 1992). This second, less well-known system of cantillation has ‘its own rules and associated grammar — similar in structure [to the so-called prose system] but different in content’ (Price 1990: 18). Indeed, on the one hand it has an impoverished phrase structure — less articulated, seemingly more primitive — yet, on the other hand, is exquisitely sensitive to the number and nature of syllables.4 The linguist Jerzy Kury¥owicz is apparently alone in assuming that the poetic accent system is the foundation of biblical metre (1972: Ch. 10, §§14–34, 166–78; 1975: Ch. 12, §§6–17, 215–25; see further Cooper 1976),5 tentatively proposing a metrical analysis of Biblical Hebrew (BH) poetry based on the phonological phrase (his ‘accented word-complex’) on the analogy of the Old Germanic ‘bar’.6 4 Representative here is the choice among three conjunctives to serve the dis junctives great revia and azla legarmeh; see further Price (1990: 250 1, 260 2). In terestingly, it is not simply a matter of full versus schwa syllables, but, crucially, sur face long versus surface short vowels: thus, the nature of the pretonic vowel deter mines the choice between the conjunctive sinnorit mahpak (pretonic open syllable with surface long vowel) versus illuy (pretonic closed syllable with surface short vowel). In Psalm 111, compare v. 9aa (with sinnorit mahpak serving great revia) and v. 10aa (with illuy serving great revia). 5 Closely allied to Kury¥owicz’s line of investigation is the work on colometry in light of TH accentuation, e.g., LaSor (1979). On the so called Kampen School of colometry, see the history of research provided by de Hoop (2000: 48 50). See fur ther the recent study by Thomas Renz (2003) that pays special attention to the syn tax of the accent system. 6 Interestingly, the family resemblance between Biblical Hebrew and ancient Germanic metres crops up in the recently published notes to Lambert’s grammar (Lambert 2005): ‘La prosodie hébraïque (comme la prosodie germanique) est fondée sur le nombre des syllables accentués …’ (appendix II: §1, p. 9).

83


Vance’s superficial and unsympathetic treatment of Kury¥owicz’s proposal (Vance 2001: Ch. 2, 166–73) fails to do justice to his genuine insights.7 As is so often the case, this seems to be a reinvention of the wheel. Dhorme’s footnotes to his discussion of metre in Job (1926: cxlviii, notes 2-3) point to an obscure paper by Paul Vetter (1897).8 Vetter’s ‘elementarste und grundlegende Gesetz’ summarizes his approach to the metre of Job guided by the poetic accents: Jeder Vers des Buches Job enthält, wenn er ein Distichon ist, drei Cäsuren: eine Haupt- [major disjunctive] und zwei Nebencäsuren [minor disjunctives]. Ist es aber ein Tristichon, dann zählt er fünf Cäsuren, nämlich zwei Haupt- und drei Nebencäsuren. Diese Scheidung des Verses in vier bezw. sechs Cäsurgruppen beruht auf logischer Grundlage (Vetter 1897: 399f, emphasis original).

This paper adopts a generative-grammatical framework to reformulate and extend this phonological-phrase approach to BH poetry, in the first instance to those portions of the biblical corpus marked out by the poetic cantillation: the three so-called ‘Books of Truth’ (from the Hebrew acronym for Job, Proverbs and Psalms; or alternatively and appropriately ‘Twin’). This is in keeping with the recommendation of Hölscher, who for a number of methodological reasons confines his study to ‘nur Texte der jüngsten Zeit’: ‘Als Gegenstand metrischer Untersuchung empfehlen sich darum in erster Linie Psalmen, Sprüche, Hiob, Hoheslied, Klagelieder und Sirach’ (1920: §17, p. 99, emphasis his; cf. Segert 1960: 286). 7 To be fair, Kury¥owicz evinces a lack of sophistication when it comes to the accent system. A glaring case is found in his parade example of Ps. 1:1, comparing his parse to that of Sievers (Kury¥owicz 1972: §29, 175f): here he counts mahpak legarmeh, a disjunctive (D3f ) clearly marked by the paseq (vertical bar), as simply mahpak, a conjunctive, giving a count of one phrase instead of the correct two. Another case, noted by Vance (2001: 169), is Kury¥owicz’s reading of the tetragrammaton as vowel final y¢howâ instead consonant final ’adonay. This reading produces an exceptions list for the Psalms (1972: §19, 170) where all but one ex ception have the tetragrammaton on a conjunctive (of note among his ‘exceptions’ is Ps. 21:14a, where BHS is incorrectly missing the dagesh); as a result, his analysis is systematically thrown off by one in §20 (p. 171) in six of his 26 examples. Furthermore, that lone exception not involving the tetragrammaton, 26:9b, is the output of a transformation that converts an underlying disjunctive to a conjunctive (Price 1990: 170f, 209ff, et passim) Kury¥owicz’s ‘rule’ that he borrows here unattributed from Gesenius Kautzsch §15 o (though see 1975: 223, n. 13). He turns around and invokes this rule of a virtual disjunctive, extrapolating from 29:6b incorrectly in the event to cover the putative exceptions with the tetragrammaton as well. 8 Thanks again to John Hobbins for pointing out this obscure reference.

84


This programme in generative phonology is explicitly situated within the interdisciplinary study of generative metrics, where formal music theory and literary linguistics engage theoretical phonology and generative grammar generally (Fabb 1997, 2002; see further Dresher and Friedberg 2006). Within this formal framework the linguistically significant generalizations of Hölscher, Horst, Mowinckel and Segert9 on the metrical ‘foot’ on the one hand, and Vetter, Kury¥owicz and Cooper on the metrical ‘phrase’ on the other, can be captured, integrated and thereby more adequately reformulated. 1. Psalm 111 and Counting Regularities Pursuing the strategy of divide and conquer, the scope of this programmatic paper is drastically limited to just the twenty-two lines of Psalm 111 and those lines only (a traditional romanization with phrase divisions added is provided in appendix 1). The paper explains how the prosodic representation of the liturgical chant regulates the poetic metre of this psalm. The analysis is proposed as the basis of an extended research programme in generative metrics and BH poetry. Psalm 111 has been selected, first and obviously, because it is a poem that has been annotated with the poetic accents, and so presumably is representative of the metrical tradition. Second, it is an acrostic, which removes the fiddle-factors in lineation as well as possible intercalation of extraneous lines (of potentially different metrical structure). Third, there are no emendations worth troubling about. Fourth, it is short for a biblical acrostic, presenting a more tractable database. Finally, and most importantly, Psalm 111 emerges as the most regular poem in the seminal syllable-counting study of Culley (1970). Furthermore, because it is an acrostic, it also receives a detailed statistical analysis in Vance’s massive doctoral study (2001) as an additional check. Culley (1970) proposes a ‘strictly descriptive approach’ to BH poetry. He adopts a syllable count as the better measure of the length and the contour of the colon. He provides a detailed analysis of several key poems, the product of the application of his relatively crude statistical approach. Vance, however, is dismissive of the value of Culley’s paper (Vance 2001: Ch. 2, 182–4) — despite the broad overlap of the two approaches. 9 Segert (1960) provides an overview of the ‘alternating stress approach’, with references for the history of this line of research (see esp. his note 1, p. 284).

85


Culley is careful not to declare that this demonstrates meter, since it surely does not, but that these data might prove useful for the question of meter. One is hard pressed to see how. All that seems to be demonstrated is that the lines (cola, verses, or whatever) are approximately the same length. But this has never been in disputed and hardly points conclusively to some underlying meter. Culley admits this latter point but fails to show what value syllable counts do have if one is not going to propose a syllabic meter (p. 184).

On the contrary, the ‘value’ of Culley’s syllable counts is at least threefold. First, the resulting taxonomy (his ‘groupings and distinctions’) ‘permits some distinctions to be made within the corpus of Classical Hebrew’ (p. 28): a non-trivial diagnostic tool with implications for source criticism (see further Fokkelman 2000: 9–11 on ‘normative numbers’). Second, the several ‘significant patterns’ strongly suggest ‘that restrictions have been imposed upon the poetic structure’—or perhaps better, superimposed—implying a ‘metrical structure of some sort’ (p. 27), pace Vance. Third and more pertinent in the present context is Culley’s implied identification of Psalm 111 as the most regular among his most regular group, the group characterized by a statistical mode of eight syllables per line. This eight-syllable-modal group includes, significantly in the context of poetic cantillation here, representative chapters from Job (chs. 6 and 9) and, not surprisingly, the twin acrostic of Psalm 112 (as well as three other psalms).10 Specifically, Culley provides a statistical ‘summary’ of Psalm 111: a range of 7–10 syllables per line, a ‘significant range’ of 7-9 syllables, with a ‘most frequent length’ (or mode) of 8 syllables (p. 18). 10

Interestingly, select lines from the so called P Source (Numbers 23 4) are grouped here too. In a corpus linguistic study of biblical dialectology (DeCaen 2000 1), I suggest that the Pss congregate at the extreme end of the dialect spec trum together with the P Source. À propos the significance of the eight syllable count, Fokkelman (2000) writes, ‘And what if we should find more than 40 poems in the Psalter that score this per fect 8.00 as the average number of syllables per colon? In that case, the proposition that the occurrence of such a round number is a coincidence has become totally untenable… I am toying with the idea that the 148 poems of the Psalter contain 1193 strophes, 2695 verses…, 5714 cola, 18,944 words and 45,733 syllables. If we divide this last number by the number of cola we get 8.0036, a highly remarkable number as it only deviates from the perfect number 8 (8.00) in its third decimal. This is such a minimal deviation that we should consider whether the number 8 might not have been a guideline or normative figure for the poets in constructing cola’ (Fokkelman 2000: 9 10). 86


Vance provides a check on Culley’s rough analysis: Vance’s range is 69 syllables, with a mode and median of 8 syllables and an arithmetic mean, rounding to one decimal place, of 7.7 (Vance 2001: Ch. 3, 421-427). Appendix 2 is provided as a comparative guide to counting. The Masoretic Text (MT) counts for word11 and syllable are provided first (notice the interesting average word count of 3.0). Culley’s count (1970: 18) reduces MT by four syllables by reading the divine name as bisyllabic yahwe instead of the trisyllabic ’a∂onay (but then he adds a syllable in v. 2b, apparently a typographical error). Vance (2001: 425) and Fokkelman (2003: 369) are more or less following Friedman’s system of syllable counting, discounting various shwa vowels. Vance obtains a slightly higher figure by counting the segholate öme∂e† as trisyllabic in verses 3b and 10c. The lettercounting system developed for ready comparison with Ugaritic verse is provided as an additional check (Loretz 1979: 168; see further Loretz and Kottsieper 1987: esp. 23, 25–6, 39–40); an average of 13.5 letters is roughly what might be expected for an average of 8 syllables. (In appendix 3 and elsewhere, the anaptyctic vowels12 are isolated by the use of <pointed brackets>; similarly, the post-tonic syllables are so marked. There does not appear to be any reason here prima facie to discount any such syllable in the count, as will become clearer below, pace Vance and Fokkelman. Indeed, the post-tonic syllable of the segholates substantially improves the rhythm and strengthens the generalization regarding anapests, as explained below; and the post-tonic syllable seems to be absolutely required metri causa in v. 10b: s1Èel †ôß´ ). This exceedingly narrow range around the average of eight syllables (mean, median and mode) surely demands an explanation in terms of a syllable-organizing unit or ‘foot’ that permits minor variations. The remainder of this paper is devoted to such an explanation in terms of higher-level organizing units: the phonological phrase and metrical foot.

11 ‘Word count’ here discounts white space counter indicated by the TH hyphen or maqqef. The word count for the purely consonantal text increases to 72 by eliminating the five maqqefs, giving a mean of 3.3 words instead. 12 For a theoretical explanation of the nature and distribution of TH anaptyxis in terms of syllable contact and sonority constraints, see DeCaen (2003).

87


2. Theme and Variation: Continuous Dichotomy and the Prosodic Hierarchy ‘Meter is a contract between the poet and the reader. The poet declares what he or she is going to do in the opening lines of the poem, and this in turn, sets up the reader’s expectation’ (Vance 2001: 491). Following Vance’s suggestion, the heuristic adopted for this research programme is that the first line and first stanza must establish the underlying metrical theme of the Psalm. Further, the ‘fulfilling of the contract may involve permissible variations to which the reader is sensitive and that give heightened pleasure for the reader’ (491). Accordingly, the ‘permissible’ variation will be sought in subsequent lines. At the highest level, we observe two ‘bars’ or phonological phrases in 111:1a (see the accentual parse provided in appendix 3). Indeed, the poem generally approaches, in Kury¥owicz’s terms, a 2 + 2 verse.13 However, making due allowance for the lawful transformations of disjunctives into virtual disjunctives, made explicit in Price’s phrase-structure analysis (1990, 1996),14 the acrostic becomes consistently 2 + 2. (The ‘virtual’ disjunctives are marked below by an exclamation mark (!), and all such cases are so indicated in appendix 3, section (B).) This observation of 2 + 2 metrical regularity hardly scratches the surface, however, and still might be compatible with the syntax-only approach of conventional wisdom (see references in Vance 2001). The linguistically significant generalization is the double dichotomy of the accentual parse of 111:1a given in (1).15 13 Cf. Ps. 44:5 27 (Kury¥owicz 1975: §15, pp. 222f ), Prov. 8:22 31 (Cooper 1976: 199ff ); and Job generally (Vetter 1897). The apparent cases of 4 + 2 in verses 9 and 10 are reanalysed here as tripartite 2 + 2 + 2. 14 The Law of Transformation is similar to rhythm rule or nesiga, just one pro sodic level higher: the clash of back to back disjunctives is resolved by demoting a disjunctive to its ‘virtual’ disjunctive counterpart (an appropriate conjunctive). Schematically, we could characterize all instances in Psalm 111 as C D D → C C D. (Indeed, musically, the reality might even be more nesiga like: X Y Y → Y X Y. This might explain why putative conjunctives, reserved exclusively for the output of such a transformation, graphically resemble disjunctive counterparts in the prose system: conspicuously, the conjunctive tarcha, which is identical to the prose D1f tiphcha.) 15 The standard notation employed here is borrowed from Dresher (1994), ulti mately from Cohen (1969). Conjunctives (C) are distinguished from disjunctives (D). The relative rank of the disjunctive is given {0 3}, where 0 marks the highest/ strongest grade among the disjunctives and 3 the lowest/weakest; D0 is verse final silluq, for example, and Dn will be the general variable, where n ranges over {0 3}.

88


(1)

1 2f

1

C

D2f

’ô∂1h

yahwé17

M16

D1

beÈol

leß0ß

Significantly, the double dichotomy in (1) captures the very essence of the TH trope (fixed musical phrase): C Dn+1f C Dn. Each such phrase is characterized by its own basic conjunctive. For the athnach trope (D1) in (1), for example, the phrase is dominated by the conjunctive munach: munach dechi munach athnach. The silluqtrope (D0) in 111:6b (given by way of contrast in (2)) is dominated by the conjunctive mereka: mereka revia-mugrash mereka silluq. (Similarly, D2 is dominated by sinnorit-mahpak, and so on.) (2)

0 1f

0

C

D1f

C

D0

la-†1†

la-hém

na-≤Ìa-≥lá†

go-yî¬m

There is yet another significant generalization one level further down. Psalm 111 is certainly not a series of random collections of more or less eight syllables distributed over two ‘bars’ or phrases. In terms of theme, it is not a coincidence that the very first line 111:1a instantiates exactly the average eight syllables and the average three words; nor is it a coincidence that the syllables alternate weak-strong The additional f for ‘final’ separates the minor disjunctives (Dnf ) from the major disjunctives (simply Dn); in Price’s terminology, Dnf is the ‘subordinate’ disjunc tive, while Dn is the ‘remote’ disjunctive. Finally, as a conjunctive place filler in the representations, I am using M to mark maqqef. 16 The maqqef (M) is assumed throughout to do duty for the trope’s conjunc tive: a ‘virtual munach’, as it were, in this case. 17 The rendering of the divine name is conventional, and the doubt regarding the actual pronunciation of this name does not raise substantive questions in this study, as will become clear in the sequel. 89


in the iambic rhythm, suggestive of a syllable-organizing metrical ‘foot’. The prosodic representation in (3) is given as the explanation (for details, see further Dresher 1994). (3) INTONATIONAL PHRASE (I)

I f

PHONOLOGICAL PHRASE (f)

f

PHONOLOGICAL WORD (w)

w

w

w18

w

FOOT (F)

F

F

F

F

SYLLABLE (s)

s

s

s

s

s

s

s

s

’ô

∂1h

yah

wé

be Èol

le

ß0ß

In (3) we observe the general isomorphism between the TH accentual parse and the underlying prosodic hierarchy of generative phonology (Dresher 1994: esp. (4), p. 8; see further Nespor & Vogel 1986). Notice that the intonational phrase (I), the domain governing pausal phonology, corresponds on this view to the poetic line governed by the major disjunctive. Notice further how the continuous dichotomy characteristic of TH prosody extends in this thematic line all the way down to the level of the syllable. Finally, it is worth noting the correspondence of prosodic ‘foot’ (F) in (1)–(3) to the TH ‘accent’ (virtual or otherwise). The mere observation that this line instantiates a regular iambic rhythm (Hölscher’s Silbenalternation) misses this significant linguistic generalization regarding the global nonlinear, dichotomous organization. The diagram in (3) is the Platonic form of Hölscher’s regular eight-syllable, four-foot, two-phrase Doppeldipodie or Vierheber (1920: §C, 99–101, esp. 100). Let us consider this first line and its double dipod as, following Vance, the ‘contract between the poet and the reader’, necessarily declared in the opening line, which ‘sets up 18 The prosodic level of phonological word (w) is ignored in the sequel where redundant. The questions of ‘prosodic word’ versus ‘clitic group’ and the proper analysis of TH cliticization generally are tangential to the goals of this paper, and are consequently left to the side. For a theoretical overview, see Dresher (2000, in press).

90


the reader’s expectation’ (Vance 2001: 491). Once this theme is firmly established, the way is cleared for ‘permissible’ minor variation, to which we now turn. 3. Variation: Foot-Syncopation and Foot-Substitution The superficial observation that Ps. 111:1b is also composed of exactly eight syllables misses the prosodic family resemblance with 111:1a. The analysis of v. 1b in terms of the prosodic hierarchy is given in (4). In order to transform the theme (v. 1a) into this variation (v. 1b), we need only posit one rule of foot-syncopation (missing foot marked Ø). (4)

I f

f

F s

F s

s

sô´∂ ye

be

s

Ø

F

s

s

´ sa rîm

s

s

we ë

∂â¬

The mirror-image foot-syncopation is found in 111:2a, which is provided in (5) by way of contrast (extrametrical anaptyxis is marked with <pointed brackets>). (5)

I f

f

Ø

F s

s

F s

ge ∂o lî¬m 91

F

s

s

s

s

ma¨

sê´

yah

wé


We can accordingly characterize these rule-governed metrical variants in the first stanza in terms of constrained (a) foot-syncopation and (b) foot-substitution. 3.1. Foot-Syncopation The upper end of the range is necessarily four feet per line, given the dipodic nature of the phonological phrase. The variants mark the lower end of the range as three feet per line. Two questions arise once we admit the tripodic line under syncopation. First, how do we keep the lower range from dropping to two feet per line, since permitting syncopation generally would obviously overgenerate in Psalm 111? In other words, how do we prevent syncopation from applying to both phrases in the line? Second, more technically, what is the prosodic nature of the tripodic input and output of the ‘Law of Transformation’ (note 14) governing the TH virtual disjunctives? First, we obviously need a basic phrase-structure rule that introduces asymmetry and syncopation. Accordingly we introduce optionality (indicated by the round brackets) into a generalized phrase-structure rule, obtaining thereby X → (W) S (read ‘some prosodic unit X maps to an optional weak unit W and an obligatory strong unit S on the lower prosodic level’). This rule permits only the two mappings between levels shown in (6): the iambic structure in (6a) and the so-called ‘degenerate’ structure in (6b). (6)

(a)

X W

(b) X S

S

To avoid overgenerating (i.e., obtaining the reduction to just two feet per line through syncopation in both phrases), we require a general constraint against ‘stress clash’. The constraint is formulated in (7), abbreviated *CLASH (read ‘star clash’), to ensure that two feet are insufficient to project the two phonological phrases demanded by the poetic line. (Let us assume further that the repair strategy would result in the structure in (8).) (7)

*CLASH X

X

*S

S

92


(8)

X

→ W

S

The clash constraint in (7)–(8) suggests in turn the proper treatment of the TH Law of Transformation, whereby a disjunctive is demoted to its ‘virtual’ counterpart (note 13). The input in (9), with back-to-back strong nodes, would obviously violate *CLASH; the output must be similar to (8). I propose that the asymmetrical left-recursive structure in (10a) captures the Janus-faced behaviour of the virtual disjunctives, contrasting with the flat structure in (10b). (9)

I fW

fS *FS

FW (10a)

FS I

(10b)

f fW FW

I f

FS

F W FW F S

FS

In favour of the proposal in (10a) is the asymmetric ternary structure observed at other levels of the hierarchy: in the F/s mapping (the anapest, see below) and the U/I (verse/line) mapping. The entire TH verse or prosodic utterance (U) (consistent with Dresher’s isomorphism between TH accent system and prosodic hierarchy) permits a 4 + 2 in verses 111:9 and 10, which according to the major TH disjunctives should be parsed as the asymmetric 2 + 2 + 2 in (11).19 The diagram in (11) captures Dhorme’s intuition on 19

This suggestion opens up an interesting way to explore the asymmetric nesting of TH phonological phrases with consequences for Dresher’s Discrete Level Hypothesis versus the Strict Layer Hypothesis (Dresher 1994, esp. p. 39). Specifi cally, the following diagram shows the maximum left recursion possible on the view proposed here, treating the Strict Layer Hypothesis more as the licensing condition for such recursion. This diagram as conjecture makes a testable prediction regarding 93


the three-line construction: ‘Selon nous, le vers tristique n’existe point. Il y a vers accompagné d’un hémistiche ou demi-vers’ (1926: cxlviii, n. 2). (11)

U IS

UW IW

IS

Against this proposal of left-recursion in (10a) is the apparent lack of phonological effect of the asymmetry and virtual disjunctive. Indeed, the data from the sandhi rule of spirantization in the Pss — otherwise diagnostic of phonological phrasing — are consistent rather with the flat structure in (10b). On the other hand, the data from nesiga (stress retraction) cut the other way in favour of (10a). There is an apparent failure of nesiga in 111:9c, where the word-pair wenôr0’ semô´ might be expected to cre´ This failure ate a clash (repaired by stress retraction as wenôra’ semô). can be readily explained by the asymmetric structure (10a), but not by the flat structure (10b). While phonologically-analogous l¢È0 does in fact trigger nesiga (Revell 1987), this isolated, problematic example in 111:9c is not much to hang one’s hat on. Unfortunately, data on nesiga in the Pss20 are hard to come by, and minimal pairs are virtually nonexistent. There are, nevertheless, at least two apparent minimal pairs that capture the conjecture’s prethe distribution and depth of TH phonological phrase nesting, derived from first principles. U U I I

I

I f f f f F 20 Many thanks again to James Price, who kindly provided me with nesiga data from his marked up database. His data relating to the environment before athnach and silluq in the three poetic books are consistent with the complementary distribu tion predicted by the conjecture here. 94


dicted asymmetry, contrasting retraction in (12) under normal phrasing with the failure to retract stress on the virtual disjunctive in (13) despite the analogous stress environment. The geometry and the weak (W) versus strong (S) contrast conjectured here considerably simplify the characterization of the necessary environment for nesiga,21 eliminating the need to invoke syntactic categories and/or structures with Revell (1987). (12) (a)

f

FW

FS

s

s

s <s>

’1

lî

’0t

(b)

tâ

(Ps. 22:11)

f FS

FW s

s

s

h0

yû

lî´

(Ps. 139:22)

f

(13) (a) fW

FS FS s

kî-yaÌî´∂

s

s

s <s>

w¢ ä

nî´

’0

nî

(Ps. 25:16)

21 A quick summary of Praetorius’s description of the environment is supplied by Khan (1987: 62); see Revell (1987) for an exhaustive treatment.

95


f

(b) fW

FS FS

yaÌs¢ßû¬

s

s

s

ra

¨0

lî´

(Ps. 41:8)

There are more subtle effects that could likewise be attributed to the asymmetrical geometry. For instance, the regular conjunctive mereka before silluq predictably alternates with the marked munach according to syllable-count. That alternation disappears — indeed, is neutralized in favour of the marked alternative munach — where that conjunctive represents the virtual disjunctive revia-mugrash (Price p.c.). In other words, the necessary (but not sufficient) condition for mereka is the structure in (12), not (13). 3.2. Nature and Distribution of Feet There can be no doubt that TH prosody admits the ternary anapest in some form as a foot. The alternative would predict the projection of degenerate feet from shewa-syllables willy-nilly. The asymmetric analysis in (14c), however, has considerable explanatory power, able to constrain and explain the distribution of shewa-syllables across TH prosodic structure: the shewa is restricted to weak nodes in (14). (14) (a) one syllable

(b) two syllables

w

w

F

F

sS

sW

sS

hô¬∂

’o

∂1h

96


three syllables22

(c)

(d) more than three syllables

w

w FW

FS sS

FW sW

sS

ye

sa rî¬m

FS

sW

sS sW

sS

te

hil la

†ô´

Of considerable interest is the distribution of the anapest above in (4)–(5), indeed throughout the psalm (see especially appendix 3). We observe that an anapest in Psalm 111 appears on the disjunctives only, not the conjunctives. Indeed, the anapest is much preferred on the major disjunctives (the generalization is strengthened by reading ’a∂onay for the divine name), and is virtually de rigueur on the minor disjunctives in line types B–C. Consider further that this sort of distribution and prosodic head-dependent asymmetry is consistent with the general prosodic approach here, but remains unexplained under the conventional, crucially non-prosodic, syntax-only approach.23 Simple foot-substitution in the double dipod, constrained by the proposed asymmetry, would therefore generate a principled range of 8-10 syllables per line as result; syncopation would drop the lower limit, generating 6–10 syllables with an average of 8 syllables per line. We therefore explain Culley’s findings from first principles. 22 It should be pointed out that the Tiberian cantillation distinguishes ‘short’ versus ‘long’ trisyllabic words (Dresher in press; 1994: §6.2, 34 6), whereas there is no distinction here in Psalm 111 (they are all short anapests). Specifically, lîre’aw (5a) and ne’emanîm (7b) are both ‘long words’ by rule, but this does not register in the metre or rather, the Tiberian prosody results in supernumerary feet. This very subtle mismatch is worth exploring as a diagnostic and perhaps key in future study. 23 While this ‘head dependent asymmetry’ in Psalm 111 is quite striking, in cross linguistic perspective it is rather commonplace. A theoretical treatment within current generative phonology can be found in Dresher and van der Hulst (1998); they specifically treat TH under ‘visibility’ asymmetries (§2.1, 328ff ). (Notice fur ther that this head dependent ‘complexity’ asymmetry here in Psalm 111 may in part shed some light on the original differences in stress and trope between con junctives and disjunctives.)

97


However, there is much more going on terms of distribution of foot- and word-types. Consider, for example, the distribution of the monosyllabic or degenerate word-feet (14a). In five lines (111:3a, 4a, 5a, 6a and 10b, indicated by Ø in appendix 3) the weak syllable of the foot is itself syncopated, as it were, at the beginning of the line; notice further that in all cases we are dealing with a D2 dechi line. In itself, such an iambic line, known as a ‘clipped’ (or ‘truncated’) line with initial syncopation (and ‘broken-backed’ with medial syncopation), is rather commonplace in iambic metre. For instance, the clipped line can be found throughout the English tradition, beginning with Chaucer among other worthies (e.g., Steele 1999: Ch. 2, §6, pp. 84ff ).24 (Notice again the explanatory superiority of the prosodic versus syntactic theories with respect to the distribution of prosodic phenomena.) Consequently, the range in Psalm 111 ex hypothesi is now 5–10 syllables per line when all permissible variation is accounted for. It is worth noting, however, that the actual range obtained in Psalm 111 in fact avoids these extremes: 6–9 syllables per line (consistent with Vance 2001: 426). It is possible to distinguish other patterns in the distribution of words and feet generally, though this takes us beyond the current remit. Suffice it to suggest that the distribution of ‘long’ (14d) versus ‘short’ words (Dresher in press; 1994: §6.2, 34–6) may prove to be rule-governed. Notice, for example, that the two-foot ‘long word’ (weÒi∂qa†ô (3b), leniple’o†aw (4a), tehilla†ô (10c)) appears exclusively on the D1 accent in Psalm 111. We know independently that the lesser accents D2 and D3 are somehow more complex — or at least more discriminating — perhaps reflecting a difference in free ‘initial’ versus fixed ‘cadence’ (see further Fabb 1997: §4.1.2.). 4. Distribution of Line-types over Stanzas The three varieties of poetic-line are assigned, in the appendix 3 below, the sigla A (the majority four-foot double dipod), B (medial syncopation with resulting virtual disjunctive) and C (initial syncopation). The distribution across Psalm 111 is marked out in appendix 1, with three movements indicated (I-III).25 24

Indeed, probably Chaucer’s most famous line, the first line from the general prologue to the Canterbury Tales, begins: ø / x / x / x / x / <x> Whan that Aprill with his shoures soote. 25 The simple arrangement of lines into quatrains and so on is the result of applying the principle of continuous dichotomy bottom up, and is consistent with 98


The first thing to notice is the restricted distribution of the problematic B-line. Generally it is restricted to the second line of a verse: AB (vv. 1 and 4) and CB (v. 5; cf. v. 9); the exception is the BA in v. 3. Notice, further, that such second-position B-lines are all dominated by the D0 silluq-trope. Staying with the distribution of B, the second thing to notice is the movement-initial quatrain pattern XBYY, where X is either A or C, and Y is the remaining choice: hence the quatrain patterns ABCC (vv. 1–2) and CBAA (vv. 5–6). Pursuing this line of analysis in the spirit of Dhorme (1926), we can interpret verses 9-10 as the rulegoverned amplification of the CBAA pattern: CB+AA+ = CBBAAA (in this light, see again (11) above). Further, this last observation could be understood as consistent with the general idea that the overall structure of the first verse should be replicated in some fashion elsewhere, especially in the poem’s conclusion. As a final note on global structure, notice that the first verse of the quatrains in the opening movements (vv. 1, 3, 5, 7) are all characterized by the count 6/7: six phonological words and seven metrical feet. 5. Conclusion: Theme and Variation in Psalm 111 The metre of Psalm 111 can be characterized, therefore, in terms of the double-dipod ‘contract’ in the opening line of v. 1a (and again in the concluding line 10c), with the ‘permissible’, rule-governed variation of syncopation licensed in and by the following three lines that make up the opening quatrain (all four lines with 8 syllables, according to Culley)26. We can capture this metrical generalization with the diagram in (15): a prosodic template with asymmetry (the F/s choice) and optionality (indicated again by parentheses), augmented by a metrical-grid interpretation beneath (the height of the columns representing relative stress)27; the tree in (15) does not encode the additional suggestions in the secondary literature (see further Scoralick 1997: 191 193). Fokkelman finds Scoralick’s recent study ‘unsatisfactory’ (2003: 213, n. 1). His analysis is admittedly inconsistent with that of the twin Psalm 112, as well as the MT and ‘all subsequent translations and commentaries’, yet Scoralick is chided for not spotting it in her ‘unsound’ analysis (p. 213). Indeed, his version is so obvious that he can ‘leave it to the reader to find good poetical reasons’ for altering MT (2002: 170)! 26 Assuming that the typographical error in his count is corrected. 27 This metrical grid can naturally and simply explain the relative length of vowels in TH orthoepy. Khan gathers the details from the masoretico grammatical literature with special reference to the accentuation (1987: esp. 59 66). 99


possibility of ‘clipping’. Once the ‘reader’s expectation’ is fixed, further ‘permissible variations to which the reader is sensitive and that give heightened pleasure for the reader’ (Vance 2001: 491), leading sonnet-like up to the spectacular double three-liner as envoi in vv. 9–10. (15) Template for Psalm 111 I fW

fS

(FW) sW

*

FS sS

* *

F/sW

*

(FW) sS * * *

sW

*

FS sS

* *

F/sW

sS

*

* * * *

On the basis of this study we can hazard of number of claims. Crucially, we can make the strong claim that the definition of the BH poetic line (half-verse or colon28) is simply a pair of phonological phrases. Similarly, the poetic verse or prosodic utterance (U) is a pair of intonational phrases (or a triple by left-recursion), the right edge of which is marked by pausal phonology. The yoking of such prosodic phrases generates top-down, as a first approximation, a range of 3–4 feet per line, hence 6–8 feet per verse. Consequently, the combinatorics can in principle explain the narrow range of syllables per line in Psalm 111 that was observed by Culley (1970), generating the sort of statistical profile provided by Vance (2001). On this view, the prosodic-word count (Dresher 1994: esp. §3.1, p. 9) and syllable count are both epiphenomenal, though still diagnostic29 of the underlying organizing units of phrase and foot. 28 In his study of the prose accents and colometry, de Hoop similarly concludes, ‘it was demonstrated that … colons generally end with a so called “major” accent … provided that they are preceded by the own fixed minor disjunctive accent [near accent]. This pattern of a major accent being preceded by its fixed minor disjunc tive accent generally leads to the so called “dichotomic structure”: a colon subdi vided into two parts by the fixed minor distinctive accent’ (2003: 33). 29 The prosodic-word (w) count is derived as follows, assuming the phrase-

100


There are a variety of important implications flowing from this programmatic study. First, because prosody is the syntax-phonology interface, it might still be possible to explain the structure of Psalm 111 in terms of the no-phonology, syntax-only consensus position (again, see Vance 2001 for summaries with references). However, the prosodic approach, inspired by the TH chanting, uncovers and explains linguistically significant distributional facts that can only be stated in terms of syllables and relatively weak/strong positions. Such distributions would be the merest coincidences in the non-phonological approach. The non-trivial distribution of monosyllabic words, for example, and the phenomenon of clipping cannot be captured by a syntaxonly theory. Similarly, syntax has nothing to say about syllables, whereas the metrical grid developed here can generate both the count’s range and average. Or again, the distribution of word types is apparently governed by the rank of the disjunctive accent, not by syntactic structure per se. On the view proposed here, then, it is the metrical phonology that explains the syntactic regularities in BH poetry (to the extent that there are any genuine regularities), not vice versa. (It is more likely, then, that we are dealing with a ‘grab bag’ of syntactic fragments that happen to match a given metrical grid.) Second, there are also interesting implications of a more technical nature: tentative answers, if we extrapolate the findings here, to the questions posed in passing by Dresher (1994). For example, why is the domain of pausal phonology (the intonational phrase (I)), not explicitly marked in the Tiberian system(s)? Answer: if liturgical chanting originated with the structures in Psalm 111, then the intonational phrase was already encoded in the distinction between major/non-final/remote D2–D0 (crucially, not D3, or rather D3f, significantly encoded as quasi-conjunctives graphically with paseq) versus minor/final/near disjunctives (cf. de Hoop 2000: 61f, 67f ), coinciding predictably and redundantly with major pause on the half-verse (DeCaen 2005; see also Sanders 2003). Or, again, why four degrees of disjunction? Why does the system run out at D3? Why not D4 or D5? There may be a variety of linguistic explanations: for instance, it is well known that stress systems structure rule above, restated here: X → (W) S. The minimum of the range would equal the number of phonological phrases in the layer above, while the maximum of the range would be capped by the number of feet permitted in the layer below. In Psalm 111 the range is predictably 2 4 prosodic words per line. 101


typically distinguish at most four degrees of stress. There are also just four basic elements, for example, in the Latin chant (initium, tenor, mediant, finalis). There are in fact many sources of four without resorting to numerology. Nevertheless, in Psalm 111 the maximal expansion of the system to handle the final tricola only requires branching to the depth of a ‘minor’ D3 (or D3f ): surely not a coincidence, on the view presented here. Finally, why ‘twin’ Tiberian cantillation systems? Perhaps because the simpler poetic system was designed to handle the metrically simple and regular 2 + 2 verse that characterizes ‘truth’ (Job, Proverbs and much of the Psalms), presumably with deep roots in the Second Temple liturgy. Subsequently, it might be conjectured, the primitive system30 had to be greatly expanded and articulated (including crucially the promotion of athnach to D0 status) to handle the challenge of reading extended prose passages in a similar fashion. The next steps in this programme are the analysis of the remaining acrostic psalms and the parallel track of investigating the socalled 2 + 2 mashal metre (4 + 4 in terms of ‘beats’ or feet) of the presumably homogeneous corpus in Job (Vetter 1897). Also on the agenda is the relation of the two Tiberian systems to each other31 and to the non-Tiberian accent systems. A particular question arising in the present discussion is why the Law of Transformation (note 14) should stand as an iron law for all degrees of disjunctive in the TH poetic system, whereas the analogous transformation is confined to geresh (D3f ) in the prose system.32

30 Revell points out (p.c.), in this regard, that the Palestinian and Babylonian systems bear a strong family resemblance to the poetic, not the prose, Tiberian sys tem. 31 Price points out that the prose in Job 32:1 6a is accentuated according to the poetic system, and that ‘these verses exhibit minor deviations from the rules of ac centuation due to their non poetic structure’ (1990: note 1, p. 161). These ‘minor deviations’ might be a good place to begin. 32 A startling implication follows from this difference: two different liturgical readings are possible, given identical Hebrew inputs. The difference would be con spicuous in terms of spirantization (or lack thereof ).

102


Appendix 1: Transcription of Psalm 111 [Line-internal phrase divisions are marked by the vertical bar |. The annotation in the right-hand column refers to the number of words, the number of feet, and finally the three-way contrast in line type (A–C) in appendix 3. The sigla I–III demarcate the three movements.] I 1a 1b 2a 2b

’ô-∂1h yah-wé33 | be-Èol le-ß0ß be-sô¬∂ ye-sa-rî´m | we-ë-∂â¬ ge-∂o-lî´m | ma¨-a-sê¬ y yah-wé de-rû-sî´m | le-Èol Ìep-Òê-hém

3/4/A 3/3/B 3/3/C 2/3/C

3a 3b 4a 4b

hô∂ we-ha-∂0r | pó-ö-lô¬ we-Òi∂-qa-†ô¬ | ö-mé-∂e† la-¨á∂ zé-Èer äsâ¬ | le-nip-le-’o-†0w Ìan-nû¬n we-ra-Ìû¬m| yah-wé

3/3/B 3/4/A 3/4/A 3/3/B

II 5a 5b 6a 6b

†é-rep | na-†0n lî´-re’-0w yiz-k3r le-¨ô-l0m | be-rî-†ô¬ k3aÌ ma-ä-s0w | hig-gî´∂ le-am-m¨ô¬ la-†1† la-hém | na-Ìa-lá† gô-yî´m

3/4/C 3/3/B 4/4/A 4/4/A

7a 7b 8a 8b

ma-ä-sê¬ ya-∂0w | ’e-mé† û-mis-p0† ne-’e-ma-nî´m | kol piq-qû-∂0w se-mû-Èî´m la-¨á∂ | le-¨ô-l0m ä-sû-yî´m | be-’e-mé† we-ya-s0r

4/4/A 2/3/C 3/3/C 3/3/C

III 9a 9b 9c

pe-∂û¬† | sa-láÌ le-äm-mô¬ Òiw-wâ¬ le-¨ô-l0m | be-rî-†ô¬ qa-∂ô¬s we-nô-r0’| se-mô¬

3/3/C 3/3/B 3/3/B

6/6 9/9

4/4/A 3/4/A 3/4/A

7/8 10 / 12

10a re’-sî† ÌoÈ-mâ¬| yir-’á† yah-wé 10b s1-Èel †ô¬ß |le-Èol ö-se-hém 10c te-hil-la-†ô¬ | ö-mé-∂e† la-¨á∂

33

Here and throughout, one can read ’ãdonay instead. 103

6/7 5/6

6/7 6/7

6/7 8/8

6/7 6/6


Appendix 2: Syllable Counts for Psalm 111 Verse

MT Word

MT Syllables

Culley (1970) p. 18*

Vance (2001) p. 425**

Fokkelman (2003) p. 369

Loretz (1979) p. 168***

1a 1b 2a 2b 3a 3b 4a 4b 5a 5b 6a 6b 7a 7b 8a 8b 9a 9b 9c 10a 10b 10c total average

3 3 3 2 334 3 3 3 3 3 4 4 4 2 3 3 3 3 3 4 3 3 68 3.0

9 8 9 8 7 9 9 8 7 8 9 9 10 8 8 9 7 8 7 9 8 9 183 8.3

8 8 8 +9 7 9 9 7 7 8 9 9 10 8 8 9 7 8 7 8 8 9 180 8.2

8 8 7 8 6 +9 8 7 6 8 8 8 9 7 8 9 7 8 7 8 7 +9 170 7.7

8 8 7 8 6 8 8 7 6 8 8 8 9 7 8 9 7 8 7 8 7 8 168 7.6

14 13 13 15 11 13 14 13 12 14 15 14 16 14 14 13 11 13 12 17 14 12 297 13.5

*Culley is counting MT syllables, but correcting for the divine tetragrammaton: counting the bisyllabic yahwe instead of the trisyllabic ’adonay. He appears to have made a typographical error in the count for verse 2b. **Both Vance and Fokkelman are more or less following Friedman’s syllable counting system. Vance, however, obtains a slightly higher figure by counting the segholate öme∂e† as trisyllabic (an inconsistent treatment of the segholate class) in verses 3b and 10c. ***Loretz is counting letters in the consonantal text as an alternative to counting syllables; the regularity is equally obvious. 34

The maqqef here with hô∂ is a variant output of the virtual transformation, creating an apparent mismatch between prosodic word and phonological word. This is the only case in Ps 111, and so the count is quietly corrected from 2 to 3. 104


Appendix 3: Accentual Parse of Psalm 111 (A) Full Double Dipod

n

n+1 C D1f 3b 6b 10c D2f 1a 4a 6a 7a 10b D3f 10a

D0

n Dn+1

C

revia mugrash we-Òi∂qa-†ô la-†e† la-hem te-hilla-†ô dechi ’ô-∂eh yah-wé Ø ze ä-sâ¬ Ø k3aÌ ma¨-< a-> saw ma¨-< a-> sê ya-∂aw Ø se †ôß legarmeh re’-sî† ÌoÈ-mâ

D1

D2

(B) Medial Syncopation

D1f! D0 1b 4b 5b 9c D2f! D1 3a 9b

silluq ö-me la-ä∂ na-la† gô-yîm ö-me la-ä∂ athnach be-Èol le-ßaß le-niple’-o-†aw hig-gî∂ le-am-m¨ô ’e-me† û-mis pa† le-Èol ö-sé-hém revia gadol yir-’a† yah-we

n

n+1 C

Dn

n Dn+1 !

Dn

virtual revia mugrash! be-sô∂ ye-sa-rîm Ìan-nûn we-ra-Ìûm yiz-kor le-¨ô-lam qa-∂ôs we-nô-ra’ virtual dechi! Ø hô∂ we-ha-∂ar Òiw-wâ le-¨ô-lam

105

silluq we-ë-∂â yah-we be-rî-†ô se-mô athnach pó¨-lô be-rî-†ô


(C) Initial Syncopation

n

n+1

n Dn+1

D1f 2b 7b 8b D2f 2a 5a 8a D3f 9a

D0

D1

D2

C

revia mugrash

Dn silluq

de-rû-sîm ne-ma-nîm ä-sû-yîm dechi ge-∂o-lîm Ø †e se-mû-Èîm35 legarmeh pe-∂û†

le-Èol Ìep-Òê-hém kol piq-qû-∂aw be-me† we-ya-sar athnach ma¨-sê yah-wé na-†an lî-re’aw la-ä∂ le-¨ô-lam revia gadol sa-laÌ le-äm-mô

BIBLIOGRAPHY Anstey, Matthew. 2006. ‘The Grammatical lexical Cline in Tiberian Hebrew’. Jour nal of Semitic Studies 51:1, 59 84 Collins, Terry. 2005. Review of Colometry and Accentuation in Hebrew Prophetic Poetry by Thomas Renz. Journal of Semitic Studies 50:2, 384 5 Cohen, Miles B. 1969. The System of Accentuation in the Hebrew Bible. (Minneapolis) Cooper, Alan Mitchell. 1976. ‘Biblical Poetics: A Linguistic Approach’. Unpub lished Ph.D. thesis, Yale University Culley, R.C. 1970. ‘Metrical Analysis of Classical Hebrew Poetry’, in J.W. Wevers and D.B. Redford (eds), Essays on the Ancient Semitic World (Toronto Semitic Texts and Studies, Toronto). 12 28 35

Technically, this line, as parsed by the TH accents, is the output of the vir tual disjunctive transformation (that is, technically line type B). However, I am reading here this munach as a virtual disjunctive dechi (hence, line type C) metri causa. This reading removes the glaring irregularity of the trisyllabic semûÈîm on a conjunctive, and removes an awkward B line from the strophe pattern. Regardless, nothing substantive is at stake as far as the essential claims put forward in this pa per; it does raise, however, the vexing question of just what sort of mismatches might be permitted between the TH accents and a regular metre. (This question of mismatches is not unrelated to the problem of finding an algorithm that can gener ate the TH accentuation from the raw morphosyntactic input [and what sort of mismatches result], and how this algorithm might differ between the two accent systems[prose versus poetic].) 106


DeCaen, Vincent. 2000 1. ‘Hebrew Linguistics and Biblical Criticism: A Minimalist Programme’. Journal of Hebrew Scriptures 3 [Republished in Perspec tives on Hebrew Scriptures (2006), edited by E. Ben Zvi.] 2003. ‘Hebrew Sonority and Tiberian Contact Anaptyxis: The Case of Verbs Primae Gutturalis”. Journal of Semitic Studies 48:1, 35 46 2004. ‘The Pausal Phrase in Tiberian Aramaic and the Reflexes of *i’. Journal of Semitic Studies 49:2, 215 24 2005. ‘On the Distribution of Major and Minor Pause in Tiberian Hebrew in the Light of the Variants of the Second Person Independent Pronouns’. Journal of Semitic Studies 50:2, 321 7 de Hoop, Raymond. 2000. ‘The Colometry of Hebrew Verse and the Masoretic Accents: Evaluation of a Recent Approach (Part 1)’. Journal of Northwest Semitic Languages 26:1, 47 73 2003. ‘“Trichotomy” in Masoretic Accentuation in Comparison with the De limitation of Units in the Versions: With Special Attention to the Introduction to Direct Speech’, in Korpel and Oesch (eds). 33 60 Dhorme, Paul. 1926. Le livre de Job. (Études Bibliques, Paris). [English translation 1967.] Dresher, B. Elan. 1994. ‘The Prosodic Basis of the Tiberian Hebrew System of Accents’. Language 70:1, 1 52 2000. ‘Cliticization and Phrasing in Tiberian Hebrew’. Paper presented at LSA 2000 Forthcoming. ‘Between Music and Speech: The Relationship Between Gregorian and Hebrew Chant’. [Chambers Festschrift] In press. ‘The Word In Tiberian Hebrew’. To appear in K. Hanson and S. Inkelas (eds) The Nature of the Word: Essays in Honor of Paul Kiparsky, edited by Kristin Hanson and Sharon Inkelas. (Cambridge MA) Dresher, B. Elan, and Nila Friedberg (eds). 2006. Formal Approaches to Poetry: Recent Developments in Metrics. (Phonology and Phonetics 11. Berlin) Dresher, B. Elan and Harry van der Hulst. 1998. ‘Head Dependent Asymmetries in Phonology: Complexity and Visibility’. Phonology 15, 317 52 Fabb, Nigel. 1997. Linguistics and Literature: Language in the Verbal Arts of the World. (Blackwell Textbooks in Linguistics 12. Oxford) 2002. Language and Literary Structure: The Linguistic Analysis of Form in Verse and Narrative. (Cambridge) Flender, Reinhard. 1992. Hebrew Psalmody: A Structural Investigation. (Yuval Monograph Series 9. Jerusalem) Fokkelman, J.P. 1998. Major Poems of the Hebrew Bible: At the Interface of Hermeneutics and Structural Analysis. Vol. 1, Ex. 15, Deut. 32, and Job 3. (Studia Semitica Neerlandica 37. Assen) 2000. Major Poems of the Hebrew Bible: At the Interface of Hermeneutics and Structural Analysis. Vol. 2, Ex. 15, 85 Palms and Job 4 14. (Studia Semitica Neerlandica 41. Assen) J. P. 2002. The Psalms in Form: The Hebrew Psalter in its Poetic Shape. (Tools for Biblical Study 4. Leiden) 2003. Major Poems of the Hebrew Bible: At the Interface of Hermeneutics and Structural Analysis. Vol. 3, The Remaining 65 Psalms. (Studia Semitica Neerlandica 43. Assen) 2004. Major Poems of the Hebrew Bible: At the Interface of Hermeneutics and Structural Analysis. Vol. 4, Job 15 42. (Studia Semitica Neerlandica 47. Assen) 107


Hölscher, Gustav. 1920. ‘Elemente arabischer, syrischer und hebräischer Metrik’, in K. Marti (ed.), Beiträge zur alttestamentlichen Wissenschaft: Karl Budde zum siebzigsten Geburtstag am 13. April 1920 [Buddefestschrift] (Beihefte zur Zeit schrift für die alttestamentliche Wissenschaft 34. Giessen). 93 101 Jacobson, Joshua R. 2002. Chanting the Hebrew Bible: The Art of Cantillation. (Philadelphia) Khan, Geoffrey. 1987. ‘Vowel Length and Syllable Structure in the Tiberian Tradi tion of Biblical Hebrew’. Journal of Semitic Studies 32:1, 23 82 1996. ‘The Tiberian Pronunciation Tradition of Biblical Hebrew’. Zeitschrift für Althebraistik 9:1, 1 23 Korpel, Marjo, and Josef Oesch (eds). 2003. Unit Delimitation in Biblical Hebrew and Northwest Semitic Literature. (Pericope 4. Assen) Kury¥owicz, Jerzy. 1972. Studies in Semitic Grammar and Metrics. (Prace jπzykoz nawcze [Polska Akademia Nauk, Komitet Jπzykoznawstwa] 67. Wroc¥aw) 1975. Metrik und Sprachgeschichte. (Prace jπzykoznawcze [Polska Akademia Nauk, Komitet Jπzykoznawstwa] no 83. Wroc¥aw) Lambert, Mayer. 2005. Termes massorétiques, prosodie hébraïque et autres études: Appendices à la Grammaire hébraïque. (Hautes Études Orientales 39, Moyen et Proche Orient 2, Geneva) LaSor, William Sanford. 1979. ‘An Approach to Hebrew Poetry through the Masoretic Accents’, in A.I. Katsh and L. Nemoy (eds), Essays on the Occasion of the Seventieth Anniversary of the Dropsie University (1909 1979), (Philadelphia). 327 53 Loretz, Oswald. 1979. Die Psalmen: Beitrag der Ugarit Texte zum Verständnis von Kolometrie und Textologie der Psalmen. Vol. 2, Psalm 90 150. (Alter Orient und Altes Testament 207/2. Neukirchen Vluyn) Loretz, Oswald, and Ingo Kottsieper. 1987. Colometry in Ugaritic and Biblical Poetry: Introduction, Illustrations and Topical Bibliography. (Ugaritisch Biblische Literatur 5. Altenberge) Margalit, B. 1975. ‘Studia Ugaritica I: Introduction to Ugaritic Prosody’. Ugarit Forschungen 7, 289 313. Mowinckel, Sigmund. 1962. The Psalms in Israel’s Worship. 2 vols. (Translated from the Norwegian by D.R. Ap Thomas. New York). [Revision of Offersang og Sangoffer (1951)] Nespor, Marina, and Irene Vogel. 1986. Prosodic Phonology. (Studies in Generative Grammar 28. Dordrecht) Portnoy, Marshall and Josée Wolff. 2000. The Art of Torah Cantillation: Ta amei Hamikra [Ta’ame ha Mikra]: A Step by Step Guide to Chanting Torah. (New York) 2001. The Art of Torah Cantillation Volume 2: Ta amei Hamikra [Ta’ame ha Mikra]: A Step by Step Guide to Chanting Haftarah and M’gillot. (New York) Price, James D. 1990. The Syntax of Masoretic Accents in the Hebrew Bible. (Studies in the Bible and Early Christianity 27. Lewiston NY) 1996. Concordance of the Hebrew Accents in the Hebrew Bible. 5 vols. (Studies in the Bible and Early Christianity 34A E. Lewiston NY) Renz, Thomas. 2003. Colometry and Accentuation in Hebrew Prophetic Poetry. (Kleine Untersuchungen zur Sprache des Alten Testaments und seiner Umwelt [KUSATU] 4. Waltrop) Revell, E. J. 1987. Nesiga (Retraction of Word Stress) in Tiberian Hebrew. (Textos y estudios. Madrid) 108


Sanders, Paul. 2003. ‘Pausal Forms and the Delimitation of Cola in Biblical Hebrew Poetry’, in Korpel and Oesch 264 78 Scoralick, Ruth. 1997. ‘Psalm 111 Bauplan und Gedankengang’. Biblica 78:2, 190 205 Segert, Stanislav. 1960. ‘Problems of Hebrew Prosody’, in Congress Volume, Oxford 1959 (Supplements to Vetus Testamentum 7. Leiden). 283 91 Steele, Timothy. 1999. All the Fun’s in How You Say a Thing: An Explanation of Meter and Versification. (Athens, Ohio) Tarlinskaja, Marina. 1993. Strict Stress Meter in English Poetry Compared with German and Russian. (Calgary, Alberta) Vance, Donald R. 2001. The Question of Meter in Biblical Hebrew Poetry. (Studies in Bible and Early Christianity 46. Lewiston NY) Vetter, Paul. 1897. ‘Die Metrik des Buches Job’. Biblische Studien 2:4, 377 464 Weil, Daniel Meir. 1995. The Masoretic Chant of the Bible. (Jerusalem)

109

WORD ORDER AND INFORMATION STRUCTURE IN RUTH AND JONAH: A GENERATIVE-TYPOLOGICAL ANALYSIS* ROBERT D. HOLMSTEDT UNIVERSITY OF TORONTO

Abstract The relationship between syntax and information structure is an in creasingly popular subject of research within Biblical Hebrew studies. However, there exist two asymmetries within current approaches taken as a whole: first, the only theoretical linguistic frameworks em ployed are situated somewhere within the functional approach to lin guistics (in contrast with formal, and specifically, generative ap proaches); second, a Verb Subject typological classification for Bibli cal Hebrew is assumed without empirical justification. Yet, the rela tionship between syntax and semantics, on the one hand, and prag matics, on the other, is primarily unidirectional; in other words, prag matics necessarily accesses the syntactic and semantic features of a text, but not vice versa. It stands to reason, then, that any model of information structure can only be as accurate as the syntactic and se mantic model upon which it builds. This study presents a typological and generative linguistic analysis of the data in Ruth and Jonah, an Subject Verb classificiation for Biblical Hebrew and an Subject Verb based model of information structure.

Introduction How an author communicates the message of a text can only partially be accounted for by analysing the formal syntactic and semantic features used within. A great deal of communicative information is conveyed by the manipulation of linguistic features beyond lexical meanings and constituent relationships. The study of pragmatics * An earlier, shorter version of this article (‘Topic and Focus in BH if BH is SV’) was presented at the annual meeting of the Society of Biblical Literature, Philadelphia, November 20, 2005. I thank Cynthia L. Miller and John A. Cook for reading multiple drafts and providing insightful feedback. Finally, I thank Randall Buth, whose opposing paper (‘Topic and Focus in BH if BH is VS’) provoked fur ther refinement in my analysis. All opinions expressed are solely mine, as of course are any errors. 111

WORD ORDER AND INFORMATION STRUCTURE IN RUTH AND JONAH

takes into account this information beyond syntax and semantics and the analysis of the pragmatic structure of stretches of discourse or texts is referred to as information structure. The relationship between syntax and semantics, on the one hand, and pragmatics, on the other, is primarily unidirectional; in other words, pragmatics necessarily accesses the syntactic and semantic features of a text, but not vice versa. It stands to reason, then, that any model of information structure can only be as accurate as the syntactic and semantic model upon which it builds. Herein lies a current problem in Biblical Hebrew (BH) studies. While the growth of interest in information structure of Hebrew texts continues apace,1 two asymmetries mar this endeavour: first, the only theoretical linguistic frameworks employed are situated somewhere within the functional approach to linguistics (in contrast with formal, and specifically, generative approaches); second, a verb-subject (VS) typological classification for BH is assumed without empirical justification. Using a typological and generative linguistic approach to the data from Ruth and Jonah, I assert in this article a different understanding of BH word order, that it is a subject-verb (SV) language, and sketch a SV-based model of information structure. This essay has four parts. In the first section I describe the typological study of word order and consider the relevant features from Ruth and Jonah. In the second part I describe a specifically genera1 See, among others, R. Buth, ‘Word Order Differences between Narrative and Non Narrative Material in Biblical Hebrew’, in D. Assaf (ed.), Proceedings of the 10th World Congress of Jewish Studies, Division D, Vol. 1 (Jerusalem 1990), pp. 9 16; C.H.J. van der Merwe, ‘The Function of Word Order in Old Hebrew with Special Reference to Cases where a Syntagmeme Precedes a Verb in Joshua’, JNSL 17 (1991), 129 44; idem, ‘Towards a Better Understanding of Biblical Hebrew Word Order (re view of Walter Gross’s Die Satzteilfolge im Verbalsatz alttestamentlicher Prosa)’, JNSL 25:1 (1991), 277 300; B.L. Bandstra, ‘Word Order and Emphasis in Biblical He brew Narrative: Syntactic Observations on Genesis 22 from a Discourse Perspective’, in W.R. Bodine (ed), Linguistics and Biblical Hebrew (Winona Lake, IN 1992), pp. 109 24; N.A. Bailey, ‘What’s Wrong with my Word Order? Topic, Focus, Informa tion Flow, and Other Pragmatic Aspects of Some Biblical Genealogies’, Journal for Translation and Textlinguistics 10 (1998), 1 29; J M. Heimerdinger, Topic, Focus and Foreground in Ancient Hebrew Narratives (Sheffield 1999); A. Moshavi, ‘The Prag matics of Word Order in Biblical Hebrew: A Statistical Analysis’, unpublished Ph.D. thesis (Yeshiva University 2000); K. Shimasaki, Focus Structure in Biblical Hebrew: A Study of Word Order and Information Structure (Bethesda, MD 2002); C.H.J. van der Merwe and E. Taalstra, ‘Biblical Hebrew Word Order: The Interface of Information structure and Formal Features’, ZAH 15/16 (2002/2003), 68 107; S.J. Floor, ‘From Word Order to Theme in Biblical Hebrew Narrative: Some Perspectives from Infor mation Structure’, JS 12:2 (2003), 197 236; idem, ‘Poetic Fronting in a Wisdom Poetry Text: The Information Structure of Proverbs 7’, JNSL 31:1 (2005), 23 58.

112


tive approach to the syntactic features of both books. In the third part I introduce the information structure model that I then apply to the data from Ruth and Jonah in the fourth and final section. The Typological Study of Word Order The typological study of word order is most often traced back to Joseph Greenberg’s 1963 article, ‘Some Universals of Grammar with Particular Reference to the Order of Meaningful Elements’.2 This essay set in motion a rich comparative linguistic method with the goal of discerning morphological and syntactic ‘universals’3 among human languages. In the first section in Greenberg’s essay he focused on ‘certain basic factors of word order’ and proposed using three criteria to identify the basic word order of any given language:4 (1) the use of prepositions versus postpositions; (2) the relative order of subject, verb and object in declarative sentences with nominal subject and object; (3) the position of qualifying adjectives, either preceding or following the modified noun.

Although these three criteria have been modified as the typological program has matured, they still reflect the fundamental questions involved in determining how a language patterns: does a head (i.e., the constituent being modified) precede or follow its modifier? For each basic type of head (noun, verb and adposition), there are three basic types of modifiers (complement, adjunct and specifier). Complements are constituents that complete the head, and are thus obligatory for forming a larger grammatical item. For instance, transi2

In J.H. Greenberg (ed.), Universals of Language (Cambridge, MA 1963), 73 113. The quotations marks around the word ‘universals’ simply serve to distinguish the typological notion of language universals (which are rarely without exception) and the Chomskyan generative concept of universals (which are, as principles of Universal Grammar, taken to be without exception and part of the human language faculty). 4 Ibid., 76. It should be noted, with regard to the nature of Greenberg’s universals, that Greenberg himself lists exceptions in his footnotes. In defence of this loose approach to language universals (which some now call ‘tendencies’ rather than ‘universals’), Thomas Payne suggests that ‘[l]anguages which deviate from Greenberg’s ideal types do not ‘violate’ Greenberg’s universals. They are simply in consistent with the ideal type. Since the majority of languages of the world are in consistent, it may be more appropriate to dub a perfectly consistent language as a violation of expectations!’ (T.E. Payne, Describing Morphosyntax: A Guide for Field Linguists [Cambridge 1997], 90 1). 3

113


tive verbs require complements, often in the form of direct objects, but sometimes also in the form of prepositional phrases, and so forth. Adjuncts, in contrast, are non-obligatory modifiers, such as adjectives, adverbs, and non-complement prepositional phrases. The specifier category includes constituents that ‘specify’ salient features of the head; so, for example, subjects specify the agent/theme/patient of verb phrases and articles specify the definiteness of noun phrases. Using these three syntactic categories, the basic oppositions for a typological study are ‘head-complement versus complement-head’, ‘head-adjunct versus adjunct-head’ and ‘head-specifier versus specifier-head’.5 The goal is to determine if and how a language exhibits strong tendencies for each grammatical category and each syntactic relationship. The table in (4) illustrates a simple typological analysis for English, asking the ‘head-initial’ or ‘head-final’ question for each syntactic category. (4) Heads Nouns Verbs

Complements Adjuncts (≈ obligatory modifiers) (≈ optional modifiers) destruction of the city big cities, cities in Africa destroy the city run quickly, quickly run

Adpositions in the city

runs very quickly

Specifiers the/that/our city They destroyed cities Straight down the street

The examples in (4) illustrate that English is strictly head-initial for the order of head and complement, strictly head-final for the order of head and specifier, and both head-initial and head-final for the order of head and adjunct, with greater weight given to the headfinal examples because they occur in less-restricted environments (see below, examples [14] and [15]). English exhibits no one order for all grammatical categories, but is fairly consistent within each syntactic category; in this way, English is a typical SVO language by typological standards. Compare the BH data in (5)–(9). (5) Preposition + Nominal Complement (= Preposition + Object): head-initial ba’areÒ ’el hayyam ‘in the land’ (Ruth 1:1) ‘to the sea’ (Jon. 1:4) 5

See M. Dryer, ‘On the Six Way Word Order Typology’, Studies in Language 21:1 (1997), 69 103; A. Alexaidou, ‘Introduction’, in A. Alexiadou (ed.), Theoreti cal Approaches to Universals (Linguistik Aktuell/Linguistics Today, Amsterdam 2002), 1 13. 114


(6) Noun + Nominal Complement (= Construct Phrase): head-initial s¢dê mo’ab yark¢tê hass¢pinâ ‘territory of Moab’ (Ruth 1:1) ‘rear parts of the ship’ (Jon. 1:5) (7) Verb + Complement (= Verb + Object): head-initial wattisse’nâ qolan wayhwh he†il rûaÌ g¢dolâ ‘and they lifted their voice’ ‘and Yhwh cast a great wind’ (Ruth 1:9) (Jon. 1:4) (8) Noun + Adjunct (= Noun + Adjective): head-initial naärâ mo’abiyâ haïr hagg¢dolâ ‘a Moabite girl’ (Ruth 2:6) ‘the great city’ (Jon. 1:2) (9) Verb + Adjunct (= Verb + Adverb/Adverbial constituent): a. head-initial ki al¢tâ raätam l¢panay wattibkênâ öd ‘and they wept again’ (Ruth 1:14) ‘because their evil has come up before me’ (Jon. 1:2) b. head-final hê†eb Ìarâ li w¢ko tidbaqin ïm naärotay ‘and thus you shall stick close ‘rightly it has angered me’ (Jon. 4:9) with my girls’ (Ruth 2:8)

As the examples demonstrate, the head-complement relationship in BH (5)–(7) is firmly head-initial while the head-adjunct relationship (8)–(9) shows some inconsistency. Nominal heads precede their adjuncts,6 illustrated in (8), but verbal heads, illustrated in (9), exhibit both head-initial and head-final tendencies. Overall, the examples in (5)–(9) demonstrate that BH is a strongly head-initial language, though the verb allows some variation with its non-complement arguments. The final category, specifiers, also manifests both head-initial and head-final order, as the examples in (10) illustrate. (10) Noun + Specifier: a. head-final (determiner-noun) ha’is haïr ‘the man’ (Ruth 1:2) ‘the city’ (Jon. 1:2) b. head-initial (noun-determiner/possessive) raätam hannaärâ hazzot ‘this girl’ (Ruth 2:5) ‘their evil’ (Jon. 1:2)

While articles precede nouns (10a), demonstrative and possessive pronouns follow nouns (10b). The pattern of verbal heads and their 6

Numerals can both precede and follow their nominal heads in BH and are thus an exception to the strong head adjunct order otherwise exhibited by nominal heads. The divergence of numerals from other types of nominal modifiers, though, is quite common in languages. 115


specifiers (i.e. subjects) is similarly inconsistent. The salient issue is whether the SV/head-final example in (11a) or the VS/head-initial example in (11b) represents the basic order.7 (11) Verb + Specifier a. head-final (subject-verb) w¢sadday hera¨ li ‘and Shaddai has done ill to me’ (Ruth 1:21) b. head-initial (verb-subject) ki hemar sadday li m¢’od ‘because Shaddai has embittered me much’ (Ruth 1:20)

wayhwh he†il rûaÌ g¢dolâ ’el hayyam ‘and Yhwh cast a great wind to the sea’ (Jon. 1:4) ki äl¢tâ raätam l¢panay ‘because their evil has come up before me’ (Jon. 1:2)

When a basic order for any of the categories is not immediately apparent within a linguistic corpus, which (11a) and (11b) demonstrate for Ruth and Jonah, linguists commonly use four criteria to sort the matter out: clause type, frequency, distribution and pragmatics.8 Clause Type The criterion of clause type builds upon Greenberg’s initial approach; however, typologists have since considerably refined the definition of the appropriate basic clause. One example is in Anna Siewierska’s study Word Order Rules, in which she defines the basic clause as ‘stylistically neutral, independent, indicative clauses with full noun phrase (NP) participants, where the subject is definite, agentive and human, the object is a definite semantic patient, and the verb represents an action, not a state or an event.’9 Admittedly, clauses of the sort described by Siewierska may not occur in abundance in a typical 7 The overwhelming majority opinion is that VS is basic in BH. Advocates of an SV approach are rare; see P. Joüon, Grammaire de l’Hebreu biblique (Rome 1923); V. DeCaen, ‘On the Placement and Interpretation of the Verb in Standard Biblical Hebrew Prose’, unpublished Ph.D. thesis (University of Toronto 1995); R.D. Holmstedt, ‘The Relative Clause in Biblical Hebrew: A Linguistic Analysis’, unpublished Ph.D. thesis (University of Wisconsin 2002); idem, ‘Word Order in the Book of Proverbs’, in R.L. Troxel, K.G. Friebel, and D.R. Magary (eds), Seeking Out the Wisdom of the Ancients: Essays Offered to Honor Michael V. Fox on the Occa sion of His Sixty Fifth Birthday (Winona Lake, IN 2005), 135 54; idem, ‘Issues in the Linguistic Analysis of a Dead Language, with Particular Reference to Ancient Hebrew’, The Journal of Hebrew Scriptures 6:11 (2006), 1 21. 8 M. Dryer, ‘Word Order’, in T. Shopen (ed.), Language Typology and Syntactic Description, Vol. I: Clause Structure, 2nd edn (Cambridge 2007), 61 131. 9 A. Siewierska, Word Order Rules (London 1988), 8.

116


text or discourse due to the nature of human communication.10 For example, in languages that allow subject pronouns to be omitted (that is, ‘pro-drop’ languages, such as Spanish, Italian, and Hebrew), illustrated in (12a), clauses with full noun phrase subjects, as in (12b), will be difficult to isolate. (12a) Ruth 1:2 wayyabo’û s¢dê mo’ab ‘and (they) entered the territory of Moab’ (12b) Ruth 1:6 ki paqad yhwh ’et ämmô latet lahem laÌem ‘that Yhwh had considered his people by giving them bread’

That clauses with full noun phrase subjects are statistically less common in languages like Hebrew in no way invalidates the identification of basic clauses in a text. Rather, it simply indicates that this criterion cannot be used alone but only in careful coordination with the other three criteria.11 Frequency Those who assign primacy, or at least significance, to statistics use the frequency criterion. Whereas with the first criterion a basic clause type may be in the statistical minority, the frequency approach demands that the basic order designation be assigned to a statistically dominant pattern.12 This criterion is one of the most common in BH studies and Takamitsu Muraoka summarized it succinctly in his study of emphatic structures in BH: ‘[W]e are not interested in discussing the theory that [VS] order is normal because action is the most important piece of information to be conveyed by this sentence type called verbal clause. In other words, by saying that V-S is the normal word-order we do not mean that it is logically or intrinsically so, but simply statistically.’13 The problems in using simple statistical dominance to determine basic word order are weighty, however. For example, given a 2:1 ratio of VS to SV order in a given corpus, are we justified in classifying that language as VS? The problem is exacerbated when the statistics 10

Ibid., 8 14. For a concise summary of the basic issues involved in the typological quest for determining ‘basic word order’ in any given language, see F.J. Newmeyer, Lan guage Form and Language Function (Cambridge, MA 1998), 330 7. 12 See, for example, J.A. Hawkins, Word Order Universals (New York 1983). 13 T. Muraoka, Emphatic Words and Structures in Biblical Hebrew (Jerusalem 1985), 30. 11

117


are even closer, as Matthew Dryer notes: ‘In the Auk dialect of Tlingit, for example, a text count … for the order of subject and verb revealed VS outnumbering SV by 177 to 156. In a case like this, the difference in frequency is sufficiently small for it not to seem reasonable to say that VS is more frequent than SV or that VS is basic.’14 Additionally, since a given text type, or genre, may be associated with specific clause types, such as the strong association between BH narrative and the wayyiqtol form, clause type frequency must be qualified appropriately and may not necessarily represent the basic word order in the grammar of that language (as opposed to the narrower ‘grammar’ of a specific text or genre). For instance, consider the raw numbers from Ruth and Jonah, provided in (13). (13) Simple Count of Verbs and Clauses in Ruth and Jonah Verb Count from the Book of Ruth

Verb Count from the Book of Jonah

420 Verbs 313 Finite Verbs 138 wayyiqtol 97 Perfective 78 Imperfective

201 Verbs 149 finite verbs 84 wayyiqtol 38 Perfective 27 Imperfective

Clauses with Explicit Subjects 47 VS clauses 19 VS wayyiqtol 26 XVS (10 modal, 3 X, 13 C) 2 VS 22 SV 20 SV 2 XSV

Clauses with Explicit Subjects 44 VS clauses 35 VS wayyiqtol 8 XVS (5 modal, 3 C) 1 VS 17 SV 13 SV 1 XSV

The relative numbers from both books are quite similar. Of the total verbs, seventy-five percent are finite verbs, the type which are typically identified as necessary for the basic clause type. Of the finite verbs,15 forty-four percent in Ruth and fifty-six percent in Jonah are the past narrative wayyiqtol. It is thus immediately clear that one particular form is strongly associated with narrative. More important are the clauses with overt subject constituents. At first 14

Dryer, ‘Word Order’, 74. Since finite verbs can be associated with syntactic subjects, they meet the minimal qualification for a word order study that is interested in the relative order of the subject and verb. Non finite verbs (i.e. infinitives, imperatives, and partici ples) exhibit additional syntactic complications and should be initially excluded in basic word order analyses. 15

118


blush it appears that both books favour VS order by a 2:1 ratio over SV. However, this ratio covers a great number of complicating factors involving both the basic clause type and pragmatics criteria. At this point, it will suffice to point out that the chart in (13) makes it clear that when we consider only clauses without subordinating constituents (‘C’), fronted phrases (‘X’), or modal verbs, SV is favoured by 10:1, if not more, in both books. Distribution The third approach is based on the test of distribution. Given two or more alternatives for a syntactic construction, the one that occurs in the greater number of environments is the basic order. Note that this is not the same as statistical dominance, because the issue at hand is not ‘occurrence’ but ‘environment’. For instance, in English, manner adverbs may both precede and follow the verb, as in (14)–(15): (14) a. Ethan slowly walked into the room. ADV V b. Ethan walked into the room slowly. V ADV (15) a. ?*Ethan is slowly walking. ADV V b. Ethan is walking slowly. V ADV

Although both options exist in English, based on the distributional patterns, it can be argued that V-ADV order is basic because there are environments in which the order ADV-V is not used or is less felicitous, e.g. (15a).16 Pragmatics Finally, we come to the criterion of pragmatics. Attention to the pragmatic features of clauses is particularly significant for so-called ‘free-order’ languages like Hebrew, that is, languages exhibiting a great deal of word order variation and no immediately apparent basic order. At the core of this approach is the recognition that the majority of language data contains pragmatically ‘marked’, or ‘non-neutral’, clauses. Even for languages that have a more rigid word order, such as English, pragmatics can produce extreme but grammatically

16

Ibid., 69. 119


acceptable examples, as in (16a) and (17a), in contrast with the basic constructions in (16b) and (17b):17 (16) a. Mary, I saw. b. I saw Mary. (17) a. Into the room came the Prime Minister. b. The Prime Minister came into the room.

The analysis of pragmatic features of clauses and texts has added a necessary layer to the investigation of basic word order.18 For the analysis of BH word order, frequency is the most problematic of the four criteria. The word order profile of much of the Hebrew Bible is distorted by the dominance of the past narrative verb form, the wayyiqtol, which is necessarily VS and which is clearly associated with a particular discourse type. Thus, the other three criteria, particularly clause type and pragmatics, will be given methodological priority in this study. A Generative Orientation to Biblical Hebrew Word Order When the issue of basic word order in BH, specifically the order of the verb and its specifier, the subject (i.e. VS or SV), is approached from a generative perspective, constituent movement becomes a critical feature in the analysis. By way of a brief orientation, generative analysis has determined that initial derivations (we could call these ‘clauses-in-the-making’) start with the subject preceding the verb. Since within the generative approach many constituents in the clause ‘move’ from this starting position to higher positions in the clause (that is, towards the front of the clause), it is possible for this deriva17

Ibid., 76. Marianne Mithun (‘Is Basic Word Order Universal?’, in D.L. Payne (ed.), Pragmatics of Word Order Flexibility [Amsterdam 1992], 15 61) questions whether some languages can be assigned to a typologically word order category. In particu lar, for languages with an apparently ‘free word order’, Mithun argues that we should not be looking for a basic word order in terms of the position of subject, verb and modifiers. Rather, she suggests that in these languages it might be the case that the syntactic role of an item (subject, object, etc.) is less important than its dis course role (e.g. topic hood, identifiability, ‘newsworthiness’). Thus, the order of the constituents, subject noun phrase, verb, complements, etc., will change in a ‘ba sic clause’, depending on the information status of the constituents. I do not think that Mithun’s observations obviate a basic word order discussion for such lan guages. I suggest that in clauses in which the constituents all share the same prag matic marking, e.g. all the constituents are ‘new’, such as in presentative clauses, the observable order could be identified as basic. 18

120


tion to result in a clause with VS or SV order (hence, the theoretically passé but still pedagogically useful distinction between ‘deep’ structure and ‘surface’ structure).19 The typological study of basic word order, when performed within the paradigm of generative linguistics, is able to identify the salient features of constituent order on more than one linguistic level. A typologically-minded generativist recognizes the value of cross-linguistic analysis, the nuanced discussion of which clause type best approximates the basic clause type, the identification of a variety of discourse types, and the typological obsession with compiling vast sets of data. At the same time, the generative approach does not view the final, or ‘surface’, product as the sole object of syntactic study;20 in other words, the basic distinction between deep structure and surface structure allows the generativist to identify features potentially relevant to a discussion of word order variation in such a way that a non-generativist cannot.21 19 On the conceptual changes brought about the Minimalist Program, with par ticular reference to ‘deep structure’ and ‘surface structure’ as components of the model, see A. Marantz, ‘The Minimalist Program’, in G. Webelhuth (ed.), Govern ment and Binding Theory and the Minimalist Program: Principles and Parameters in Syntactic Theory (Oxford 1995), 349 82. 20 Although Chomsky’s earliest comments on the value of ‘statistical studies’ are somewhat dismissive in tone (he took a slightly more positive approach toward ty pology within the Principle and Parameters approach in the mid 1980s), the basic critique of Greenberg’s initial study has not changed: ‘Insofar as attention is re stricted to surface structures, the most that can be expected is the discovery of sta tistical tendencies, such as those presented in Greenberg 1963’ (Aspects of a Theory of Syntax [Cambridge 1965], 118). Additionally, while Frederick Newmeyer has re cently proposed a method by which generativists can make use of typology (see above, n. 11), he earlier made the following sceptical observation concerning the linguistic relevance of typology to determining universal grammar: ‘[T]here is no evidence that ‘the collection of valuable facts’ has ever led or could lead to the dis covery of any generalizations other than the most superficial sort. For example, the seven year long Stanford University Language Universals Project (whose results are now published as Greenberg, Ferguson and Moravcsik 1978) carried out Li’s pro gram to perfection yet has not led, as far as I know, to any substantial theoretical revisions. The problem is that the fairly shallow generalizations and statistical corre lations described in the project’s reports were far too sketchily presented to be of much use in ascertaining even the grammatical structure of the individual lan guages treated, much less shed any light on universal grammar’ (Grammatical Theory: Its Limits and Its Possibilities [Chicago 1983], 71). Newmeyer has recently again expressed this scepticism in ‘Typological Evidence and Universal Grammar’, Studies in Language 28:3 (2004), 527 48. 21 From a BH studies perspective, see J.A. Naudé for a similar critique of sur face level approaches (‘A Syntactic Analysis of Dislocations in Biblical Hebrew’, JNSL 16 [1990], 115 30).

121


It is perhaps easiest to illustrate the implications of constituent movement for our understanding of BH word order by first considering examples from languages in which the choice between SV and VS order is largely dictated by the presence of other constituents. Consider German main clauses, where we find the phenomenon known as ‘verb-second’;22 this is a syntactic constraint that requires the verb to be in second position in main clauses, as the three examples in (18) demonstrate. (18) German a) Hans kaufte den Ball Hans bought the ball b) Den Ball kaufte Hans gestern yesterday Hans bought the ball c) Gestern kaufte Hans den Ball yesterday Hans bought the ball

(S-V-O) (O-V-S-ADV) (ADV-V-S-O)

The SV example in (18a) is the typical order in simple main clauses, while the VS orders in (18b)–(18c) illustrate that a non-subject constituent preceding the verb affects the relative order of the subject and verb. Within a constituent movement framework, examples like the German clauses in (18b)–(18c) are taken to be derivations from a common source, presumably (18a).23 Within a movement account, the motivation for some types of movement remains within the domain of syntax. Thus, we find constituent movement in English interrogative clauses (19).

22

German, Dutch and standard Afrikaans are considered to be ‘well behaved V2 languages’ (i.e., they never allow the co occurrence of a complementizer and V2 verb in complement clauses of matrix verbs); Danish, Swedish, Norwegian and Fri sian are examples of languages categorized as ‘limited embedded V2 languages’ (i.e. only with bridge verbs such as ‘know that…’); Yiddish and Icelandic are examples of ‘general embedded V2 languages’ (i.e. they permit the co occurrence of complementizers and V2 in all the complements of all matrix verbs). English is an example of a ‘residual V2’ language, in which we find V2 phenomena in, e.g. ques tions, as in example (19) in the main text above. See S. Vikner, Verb Movement and Expletive Subjects in the Germanic Languages (New York 1995). 23 The derivational approach to clause construction has been shown over the last four decades to fulfill the requirements of descriptive and explanatory adequacy as well as theoretical economy. For instance, a grammar that includes a base word order with a few operations motivating constituent movement, which in turn may result in further movement, is inherently less burdensome than a taxonomic gram mar, that is, one that simply lists the numerous, almost infinite permutations. And if the derivation based grammar explains all and only the grammatical and felici tous examples, then it is to be highly preferred. 122


(19) English (a) She saw the new Hummer. (S-V-O) (b) [Which new car] did she see? (O-V-S, where the tensed auxiliary = V) (c) [What] did she see? (O-V-S, where the tensed auxiliary = V)

The indicative statement in (19a) represents the base word order in English: SVO. But in both interrogative statements in (19b)– (19c), we find not only the fronting of the WH-phrase (‘which new car’ and ‘what’), but also the inversion to VS order (where the salient verb is the tensed form of do). This variation is often labeled ‘residual verb-second’ and, given the history of English, it is not surprising that it has retained some Germanic syntactic features. Constituent movement may also be semantically driven, as in (20), where the examples from Kru, a Niger-Congo language, exhibit a switch from SVO order in typical declarative clauses (20a) to SOV in clauses with a negative function word (20b).24 (20) Kru (Niger-Congo) (a) Nyeyu-na bla nyino-na (S-V-O) man-the beat woman-the the man beat the woman (b) Nyeyu-na si nyino-na bla (S-O-V) man-the NEG woman-the beat the man did not beat the woman

Additionally, some African languages vary the word order depending on the tense and aspect of the verb used; the Sudanic languages Lendu, Moru, Mangbetu, and the Gur languages Natioro and Bagassi exhibit SVO order with the perfective verbs and SOV with imperfective verbs.25 Similarly, the Sudanic language Anyuak/Anywa appears to switch from SVO in the present tense to SOV in the past and future.26 A helpful comparison for BH is that of the formal registers of modern Israeli Hebrew: when a constituent precedes the subject and verb, the normal SV order is inverted to VS, as in the clauses in example (21).27 24

T. Givón, On Understanding Grammar (New York 1979), 124 5. A. Siewierska, Word Order Rules (London 1988), 95. 26 C. Perner, Anyuak: A Luo Language of the Southern Sudan (New Haven, CT 1990); M. Reh, Anywa Language: Description and Internal Reconstructions (Köln 1996). 27 L. Glinert, The Grammar of Modern Hebrew (Cambridge 1989), 417; idem, Modern Hebrew: An Essential Grammar, 3rd edn (New York 2005), 162 4. 25

123


(21) Modern Hebrew (a) hak¢lavim navÌu baqeÒev ‘the dogs barked in rhythm’ (b) ’aÌar kakh navÌu hak¢lavim baqeÒe ‘afterwards the dogs barked in rhythm’

(S-V) (ADV-V-S)

The general process of one constituent motivating the movement of another constituent is often referred to as ‘triggering’; in modern Israeli Hebrew a fronted constituent triggers the inversion of the basic SV order to VS order.28 It is possible, if not likely, that a similar process of triggered inversion operates in BH. A high percentage of VS clauses in the Hebrew Bible, like the example in (22), also contain an initial constituent (e.g. the relative function word ’aser); thus, these VS clauses may reflect triggered inversion. (22) Ruth 4:11(REL-V-S) k¢raÌel ûk¢le’â ’aser banû stêhem ’et bêt yisra’el ‘like Rachel and Leah, who the both of them built the house of Israel’

In contrast, the same is not true of many SV clauses: for instance, in (23) there is no initial constituent that could act as a trigger for the movement of the subject and verb. (23) Ruth 4:18b (S-V; = Basic; Non-Triggered)29 pereÒ holid ’et ÌeÒron ‘Perez begat Hezron’

When we examine the BH data and ask whether the majority of VS and SV clauses fit a triggered inversion account, the answer is yes. The set of potential triggers in BH includes syntactic members, such as relative words (22), interrogatives (24), causal words (25), as well as semantic members, such as modal operators (whether overt [26]30 or covert [27]) and negative operators (28). 28 U. Shlonsky and E. Doron, ‘Verb Second in Hebrew’, in D. Bates (ed.), The Proceedings of the Tenth West Coast Conference on Formal Linguistics (Stanford 1992), 431 45; E. Doron, ‘Word Order in Hebrew’, in J. Lecarme, J. Lowenstamm, and U. Shlonsky (eds), Research in Afroasiatic Grammar: Papers from the third conference on Afroasiatic Languages (Amsterdam 1996), 41 56; U. Shlonsky, Clause Structure and Word Order in Hebrew and Arabic (Oxford 1997). 29 Note that the use of SV order to mark narrative transitions, e.g. stages in the plot development or ‘scene’ changes, is a literary device not a formal pragmatic op eration. 30 Jon. 1:6 is the sole example in Ruth and Jonah of a clause with an explicit modal word with a VS clause. For examples of triggered VS order with explicit modal words outside of Ruth and Jonah (with the function word ’im ‘if ’), see Gen. 47:16, 18; Exod. 22:2; Lev. 13:56; Num. 14:8; 21:9; 30:6; 1 Sam. 21:5; 2 Sam.

124


(24) Gen 44:7 (WH-V-S) lammâ y¢dabber ’adoni kadd¢barim ha’ellê ‘why does my lord speak according to these words?’ (25) Jon. 1:10 (CAUS-V-S) ki yad¨û ha’anasim ki millipnê yhwh hû’ boreaÌ ‘because the men knew that he was fleeing from Yhwh’ (26) Jon. 1:6 (MODAL-V-S) ’ûlay yitässet ha’elohim lanû ‘perhaps God may bear us in mind’ (27) Ruth 1:831 (VMODAL-S; ‘modality’ = covert Trigger) yaäs yhwh ïmmakem Ìesed ‘may Yhwh do kindness with you’ (28) Ruth 4:10 (NEG-V-S) w¢lo’ yikaret sem hammet meïm ’eÌayw ûmissaär m¢qomo ‘and the name of the dead man will not be cut off from his kinsmen or the gate of his place’

Additionally, we can easily account for the fixed placement of the past narrative wayyiqtol, illustrated in (29), in a triggered-inversion analysis. (29) Ruth 1:3 (Ø-V-S) wayyamot ’elimelek ’is noömi ‘and Elimelek, the husband of Naomi, died’

The gemination in the wa-y-yiqtol has traditionally been understood as a fully assimilated function word (represented above by Ø)32 and, if this is so, then the assimilated function word within the wayyiqtol form would naturally trigger VS inversion. 5:6; Isa. 4:4 5; 6:11; 24:13; Ezek. 16:48; Job 31:9; 37:20; Prov. 23:15; Song 7:13; Eccl. 10:10. 31 See also 1:9; 2:4, 12 (2x), 19; 4:11, 12, 14; Jon. 1:11, 12; 3:8b, 9. 32 What the function word within the wayyiqtol was originally is unknown. That it is a function word best explains the phenomenon of this form: the raised verb (due to triggering) is then morphologically fused with the unknown function word that the gemination represents as well as the introductory simple conjunction ‘and’. This accounts for the fixed VS order we always see with the wayyiqtol. For surveys of both classical and modern proposals regarding the history and semantics of the underspecified function word present in the wayyiqtol form see L. McFall, The Enigma of the Hebrew Verbal System: Solutions from Ewald to the Present Day (Sheffield 1982), 217 18; B.K. Waltke and M. O’Connor, An Introduction to Bibli cal Hebrew Syntax (Winona Lake, IN 1990), 544 5; W.R. Garr, ‘Driver’s Treatise and the Study of Hebrew: Then and Now’ (Preface to reprint of S.R. Driver’s A Treatise on the Use of the Tenses in Hebrew and Some Other Syntactical Questions. Grand Rapids, MI 1998), xviii lxxxvi. 125


Returning to the typological criteria for determining basic word order, it becomes clear that no clause exhibiting triggered inversion to VS order qualifies as basic. We can use, therefore, only clauses that do not have any constituent other than the verb or subject at the front of the clause, and from these examples we must determine which word order, VS or SV, is the base from which all other orders derive. Here is where we must draw in the criterion of pragmatics: if, for instance, SV is the base in BH and all other orders are derivative, then a VS clause without a discernible syntactic or semantic trigger, as in (30), must reflect pragmatically motivated triggered inversion. (30) Ruth 4:17 (V-S) yullad ben l¢noömi ‘a son has been born for Naomi’

Information Structure In order to identify formal features of this pragmatic layer of ancient Hebrew grammar, I have developed a working model of information structure that includes four core concepts in two layers: Theme and Rheme constitute the first layer, and Topic and Focus the second layer. The Theme is the constituent in a sentence that adds the least information to the communicative setting.33 It is the existing information that provides an anchor for added information, and is often described as the information in a discourse that is ‘old’, ‘known’, or ‘given’. The Rheme is the information that is being ‘added’; this can be new information or information that has been put aside, so to speak, earlier in the discourse and is now being re-invoked.34 Consider the clause in (31) as the initial component in an anecdote. (31) Abigail and Benjamin were drinking juice

33 J. Firbas, ‘On Defining the Theme in Functional Sentence Analysis’, Travaux Linguistiques de Prague 1 (1966), 267 80, esp. p. 272; idem, ‘ On the Delimitation of the Theme in Functional Sentence Perspective’, in R. Dirven and V. Fried (eds), Functionalism in Linguistics (Amsterdam 1987), 137 56; idem, Functional Sentence Perspective in Written and Spoken Communication (Cambridge 1992), 72 3; see also M.A.K. Halliday, An Introduction to Functional Grammar (London 1985), 38. 34 See J. Lyons, Semantics (Cambridge 1977), 509; K. Lambrecht, Information Structure and Sentence Form: Topic, Focus, and the Mental Representations of Dis course Referents (Cambridge 1994), 206 18. For an explicitly generative formula tion, see M.S. Rochemont, Focus in Generative Grammar (Amsterdam 1986), 9 10.

126


As the initial statement in a new discourse, there can be nothing contrastive about any of the constituents. There is also no Theme, presuming that we have not already been discussing Abigail, Benjamin, or juice. Rather, everything is new, or rhematic. Now add, as the second statement, the clause in (32). (32) Abigail wanted to drink another cup

Since Abigail is now old information, this constituent is considered a Theme, while wanted to drink another cup is the rhematic information. But since the information state of the participants in a discourse is always ‘in-process’ until the discourse is complete, what we do not know is if there is another statement following that could affect our understanding of the full pragmatic function of Abigail. Consider the continuation in (33a): (33a) And Benjamin wanted another one as well

While the thematic status of Abigail in (32) and Benjamin in (33a) remains the same, the juxtaposition of the two statements provokes the addition of a second layer of information: the Topic. Both Abigail and Benjamin in (32) and (33a) ‘orient’ the listener to which of three thematic entities (Abigail, Benjamin, or cup of juice) information is being added. In contrast, consider an alternate continuation in (33b), with (32) repeated. (32) Abigail wanted to drink another cup (33b) But Benjamin wanted milk35

With (33b) the pragmatic context changes significantly, and, accordingly, so does the total pragmatic information conveyed by both Abigail and Benjamin. The situation is now a contrastive one, with the entities Abigail and Benjamin set over against each other. This is Focus. 35 Although I have only identified Focus on the constituent Benjamin in (33b) for the sake of the comparison and flow of the presentation, it is clear that milk would also carry Focus (i.e., it is contrasted to ‘another cup of juice’ that Abigail desires in [32]. For discussion of multiple, discontinuous Focus structures like that which (33) exhibits, as well as the similar constructions that correspond with multi ple WH questions (e.g. Who bought what?), see M.L. Zubizarreta, Prosody, Focus, and Word Order (Cambridge, MA 1998). Her ‘assertion structure’ is a novel pro posal by which we may account for those propositions which have not one, but two open variables (thus two separate constituents marked for focus): ‘The A[ssertion] S[tructure] contains two ordered assertions representing the focus presupposition of a statement; the first assertion is the existential presupposition provided by the con text question; the second assertion is the equative relation between a definite vari able and a value’ (4).

127


In summary, we have four pragmatic concepts to take into account when dealing with most clauses in any given discourse. They are summarized in (34).36 (34) Theme Rheme Focus Topic

Old/known (or presupposed) information New/added (or re-invoked) information Information contrasted with possible alternatives Thematic information used to 1) isolate one among multiple Themes, or 2) set the scene (e.g. time, place)

Note that Topic has two basic functions: in (32)–(33a) we already considered examples of a Topic orienting the reader to which Theme constituent is being modified. In (35) we see the other use: orienting the reader/listener to scene-setting information (time or place adverbials). (35) Yesterday Abigail and Benjamin were drinking juice37

The time adverbial yesterday establishes a temporal setting for the utterance and is considered thematic since it is assumed that the two parties in the communicative setting both share knowledge of the referent of yesterday. (This is an example of a non-contextually defined Theme.) Finally, it is important to recognize that Topic is restricted to thematic information, but Focus can affect both Themes and Rhemes. Consider the examples of this in (36)–(37). (36) Abigail and Benjamin were drinking juice (All rhematic, no Topic or Focus)

36 For discussions of Topic and Focus, particularly those set weakly or strongly within a generative framework, see Rochemont, Focus; E. Vallduví and E. Engdahl, ‘The Linguistic Realization of Information Packaging’, Linguistics 34:3 (1996), 459 519; E. Vallduví and M. Vilkuna, ‘On Rheme and Kontrast’, in P. Culicover and L. McNally (eds), The Limits of Syntax (San Diego 1998), 79 108; Zubizarreta, Prosody. See also G. Rebuschi and L. Tuller (eds), The Grammar of Fo cus (Amsterdam 1999); A. Meinunger, Syntactic Aspects of Topic and Comment (Am sterdam 2000). 37 Neither Ruth nor Jonah contains ‘Scene setting’ Topic VS (triggered) exam ples of the simple type like Gen. 8:14 (ûbaÌodes hassenî b¢sib¨â w¢ësrîm yôm laÌodes yabsâ ha’areÒ ‘And in the second month on the twenty seventh day of the month the earth dried up’). The only qualifying example is perhaps Ruth 1:17 with the adverb kô ‘thus, in this manner/way’, e.g. ko yaäsê yhwh lî ‘In this way Yhwh will act to wards me’.

128


(37a) As for the juice, Abigail loved it (but not Benjamin) (juice = Topic; Abigail = Focus) (37b) And they both smiled at their father (not their mother) (father = Rheme + Focus)

Given the information and entities introduced in (36), the clause in (37a) presents an initial Topic, the juice, which is obviously a Theme carried over from the previous context, as well as a Focus, Abigail, for which the contrast with Benjamin can either be inferred from intonational stress or the addition of the parenthetical phrase.38 In contrast, the clause in (37b) presents mostly rhematic information, the only Theme being the pronoun they, which refers back to the compound subject, Abigail and Benjamin, from (36). Significantly, either intonational stress on father or the parenthetical phrase not their mother makes it clear that part of the new information their father is also a Focus constituent. With this framework in hand, we are ready to consider the Hebrew data from the books of Ruth and Jonah. Word Order and Information Structure in Ruth and Jonah In order to illustrate the manifestation of Topic and Focus in BH, we will now turn to the books of Ruth and Jonah. Since pragmatic concepts are heavily context-dependent, it is advisable to begin the construction of an information structure model for BH by analysing discrete narrative units. Once we have developed a working model, it may be tested and refined against other, larger textual units. Of course, it is linguistically plausible that multiple models will be required to describe all of the texts within the Hebrew Bible, given the diversity of texts, time periods and discernible linguistic influences. Thus, for this study, the results are to be taken as descriptive of the information structure utilized by the authors of Ruth and Jonah only, with the intention that these results later be tested against other corpora. As a first step, the data we look for are basic SV clauses, namely those that lack the complicating influence of Topic or Focus 38

For discussion of multiple fronting structures like (37a), see L. Haegeman and J. Guéron, English Grammar: A Generative Perspective (Oxford 1999), 333 43, 520 4; N. Erteschik Shir, ‘Focus Structure and Scope’, in Rebuschi and Tuller, Grammar of Focus, 119 50; P. Beninca and C. Poletto, ‘Topic, Focus, and V2: De fining the CP Sublayers’, in L. Rizzi (ed.), The Structure of CP and IP (Oxford 2004), 52 75. 129


fronting. But we find very few, and the simple reason for this is that narratives are informationally complex. They always contain multiple themes as they develop, and so the only place that can contain a clause without at least a Topic constituent is at the beginning of the narrative or scenes with new characters.39 A parade example is the first verse of the book of Job, provided in (38). (38) Job 1:1 (S-V-PP; All Rhematic, no Topic or Focus) ’is hayâ b¢’ereÒ ¨ûÒ ‘a man was in the land of Uz’

We have only one example like this in the book of Ruth, at the outset of the genealogy in chapter four, given in (39), repeated from (23). (39) (=23) Ruth 4:18b (S-V-O; S=Theme, V-O=Rheme; No Topic or Focus) pereÒ holid ’et ÌeÒron ‘Perez begat Hezron’

The SV clause presents us with one old entity, pereÒ, and two new pieces of information, a new verb and object. Critically, we cannot analyse the subject pereÒ as a Topic, because there are no other thematic entities from which to choose. The NP pereÒ is the only agentive entity available from the preceding clause to serve as the subject of the verb holid. Topic does not function redundantly in this way. Aside from these examples, we, therefore, have clear cases of basic SV word order in Ruth and Jonah, but, as I have explained, this is expected. In accordance with the features of narrative, the remaining SV examples present us with either Topic or Focus information. Consider example (40): (40) Ruth 4:1 (S-V-ADV; S=Topic) ûboäz älâ hassaär ‘and Boaz went up to the gate’

This SV example orients the reader to which character is acting at a major transition in the book: Boaz. At the beginning of this new scene, the use of Topic-fronting indicates which of the known characters will carry the plot forward. Similarly, consider (41). (41) Ruth 3:4 (S-V-PP-O; S=Topic) w¢hû’ yaggid lak ’et ’aser taäsin ‘and he will tell you what you should do’ 39 Besides the initiation of narratives or new scenes with new characters, genealogies and proverbs are the only other consistent source of basic SV clauses.

130


This example makes it clear that personal pronouns used with finite verbs do not always present contrasts (or better, Focus). In fact, most of the occurrences of personal pronouns are used to present Topics (see also Ruth 1:22, 2:13). Example (42) adds an important piece of information: some lexical and grammatical items do not function as syntactic operators, that is, they do not trigger inversion; hinnê is one of these non-operators (see also Ruth 3:2, 4:1, and example [48], below). (42) Ruth 2:4 (S-V-PP; S=TOPIC) w¢hinnê boäz ba’ mibbêt leÌem ‘and surprise! Boaz came from Bethlehem’

As in Ruth, the first two SV clauses in the book of Jonah present us with Topic entities: in 1:4a it is Yhwh that is a fronted Topic, in 1:4c it is the ship upon which Jonah was sailing. Consider (43), which provides 1:4c: (43) Jon. 1:4c (S-V-O-PP; S=Topic) w¢ha’oniyyâ Ìiss¢bâ l¢hissaber ‘and the ship was about to break up’

Here we have a case of a thematic entity, the ship (introduced already in v. 3), fronted in order to orient the reader to a new Topic. The narrator had been talking about Yhwh, but has now shifted to a new Topic, the ship. The only other simple Topic-fronting clause in Jonah is in 3:3, given in (44). (44) Jon. 3:3 (S-V-PREDNOM; S= Topic) w¢ninwê haytâ ïr g¢dolâ le’lohim ‘and Nineveh was a great city to the gods’

Nineveh has been a thematic constituent since the second verse of the book, and it is again employed immediately preceding this clause, but as a oblique argument. In (44) it takes on a nominative subject role and is fronted as a Topic entity, marking Nineveh as the item out of all the possible thematic constituents to be modified by a predication. While (38)–(44) illustrate a subject-Topic structure, in (45) we see an example of an object undergoing Topic-fronting. (45) Ruth 4:3 (O-V-S; O=Topic) Ìelqat hassadê ’aser l¢’aÌinû le’elimelek makrâ noömi ‘Naomi is selling40 the portion of the field that belongs to our kinsman, Elimelek’ 40 I take makrâ in this clause as a perfect verb used performatively, i.e. ‘Naomi hereby puts up for sale’.

131


The object in (45) is fronted, in the mouth of Boaz, in order to orient the other redeemer and the elders, to the important Topic at hand: the fact that a plot of land belonging to their extended family is being sold. Certainly this is not a Focus — there is nothing to contrast the field with; no other tracts of land are mentioned or relevant in the context. Note how the pragmatic fronting of the object-Topic triggers VS inversion. Another possible example of object-fronting for Topic exists in Ruth 3:17: (46) Ruth 3:17 (O-V-PP; O=Topic) ses hass¢örim ha’ellê natan li ‘he gave me these six (measures of) barley’

In example (46), Ruth simply begins her description of what Boaz had done for her at the threshing floor by specifying the gift of barley. This is not a Theme shared by Ruth and Naomi within the world of the narrative, but it is a Theme known to the audience of the narrative. Interestingly, it is used by the narrator to convey the entirety of what transpired between Boaz and Ruth, since Ruth does not share any of the other details with Naomi in the narrative. Thus, not only is it thematic, for this particular exchange it is the Theme.41 Before we move on to Focus-fronting examples, we should consider a final, common type of Topic-fronting: that which involves a prepositional phrase. Note that while example (47) does not contain an explicit subject, it does illustrate the use of Topic-fronting nonetheless (‘C’ stands for ‘complementizer’, which is the syntactic category label for most subordinators). (47) Ruth 1:16 (C-PP-V; PP=Topic) ki ’el ’aser telki ’elek ‘because wherever you go, I will go’

The PP ’el ’aser telki after the initial function word presents a Topic isolating the two directions of motion within the narrative up to this point: going back to Judah with Naomi or going back to Moab to family. Orpah has made her choice, now Ruth is making hers clear. In this elegant statement of loyalty, the choice that Ruth makes is fronted to orient Naomi to which of the two directions 41

It is possible that the clause below, from Ruth 3:11, is another example of a fronted Topic object, but is not entirely clear whether the fronted object functions as the Topic or a Focus (see also Ruth 3:5). Ruth 3:11 (O V PP; O=Topic) kol ’aser to’mri ’eësê llak ‘all that you say I will do for you’ (Ruth 3:11) 132


Ruth will comment upon. This Topic-fronting allows the narrator to contribute to Ruth’s character development. Ruth could have made a negative statement (e.g. ‘I will not leave you, I will not follow Orpah’), but instead she responds to Naomi’s plea to leave her with positive assertions about precisely where she will go, stay, and die. The next three clauses in this extended statement of loyalty maintain the same pragmatic pattern. Significantly, it might be tempting to read the initial PP as a Focus, contrasting wherever you go with the understood opposite wherever I go (by myself/without you), but the absence of explicit personal pronouns in the clause prohibits such a reading. Note also that while there is no subject, example (47) provides valuable information about the complexity of the clause structure at the front of the ancient Hebrew clause. Inside the Complementizer Phrase (CP) — the part of the clause to the left of the subject and verb, where both subordinators and fronted phrases are located — the PP has been Focus-fronted.42 This suggests the structure of the CP is multi-layered, with a Topic Phrase and Focus Phrase residing within the CP domain. Let us now turn our attention to Focus-fronting, specifically clauses that exhibit single Focus-fronted items, as in (48). (48) Ruth 1:15 (V-S; V=Focus) hinnê sabâ y¢bimtek ’el ämmah w¢’el ’elohêha ‘look! your sister-in-law has returned to her people and gods’

The VS clause in (48) presents us with precisely the type of clause at the centre of the SV versus VS debate: in the VS analysis this clause could be basic. But even the context suggests otherwise (not just the demands of an SV framework). In this example, the verb within the quoted speech is focused and therefore moved to the front of the clause (note again that interjections like hinnê, which, along with items like vocatives, are not part of the syntax of the clause proper and do not trigger VS inversion). The Focus presents a contrast between the action of the sister-in-law, she returned, and its logical opposite, ‘staying’. One might be tempted to read this clause, at least in English, with contrastive stress, and hence the Focus, on the noun y¢bimtek, resulting in something like ‘Your sister-in-law has returned so you return as well’. The problem with this reading of the verse is that for the 42

On the general nature of complementizer and complementizer phrases, see A. Radford, Syntactic Theory and the Structure of English: A Minimalist Approach (Cambridge 1997), 54 8, 95 6. 133


subjects to be focused like this would require the pronoun ’att to exist in the second half. The pronoun does not exist in this clause, thus such a reading is not available. Instead, Naomi, the speaker, is contrasting the courses of action that the two daughters-in-law have taken: one returned, one stayed. As a last example of single-item Focus-fronting, consider (49), in which we see a Focus-fronted adverb. Notice, by the way, that the Focus-fronting of the adverb triggers VS inversion. (49) Ruth 1:21a (ADV-V-S; ADV=Focus) w¢rêqam hesibani yhwh ‘(and I full went away) but empty Yhwh returned me’

The adverb in (49) is Focus-fronted to highlight the contrast between the manner in which Naomi left Israel, ‘full’, and in her opinion the manner in which Yhwh has brought her back from Moab, ‘empty’. And, of course, this contrast establishes a dominant motif in the book as a whole. In example (50), the initial function word ki should trigger VS inversion, yet we have SV order. (50) Ruth 4:15 (C-S-V; S=Focus) ki kallatek ’aser ’ahebatek y¢ladattû ‘because your daughter-in-law who loves you bore him’

It is clear that Topic and Focus-fronting are movement operations that occur after the syntactic triggering process that produces VS inversion. So, in this case, the Focus-fronted subject phrase kallatek ’aser ’ahebatek is moved to its position after VS inversion, a move that results in a surface order of SV. For example (51), the subject hammawet is fronted to contrast it not with contextual alternatives, but with logical alternatives — those established solely from the shared knowledge of the speaker-listener outside of a particular discourse. (51) Ruth 1:17 (C-S-V-PP; S=Focus) ki hammawet yaprid bêni ûbênek ‘indeed (only) death will separate me and you’

So hammawet is contrasted with, basically, anything else that typically might be a reason for a widowed daughter-in-law to leave her mother-in-law, such as other family or new marriage. The addition of the English restrictive adverb only better captures this particular Focus structure for us than simply giving heavy stress to the word hammawet. Crucially, the ability of the ancient Hebrew CP to contain both a subordinating function word, such as ki in (51) and a 134


fronted Focus item reinforces the complexity of the ancient Hebrew CP and is further evidence for the existence of a Topic Phrase and Focus Phrase residing within the CP domain. As with Topic-fronting, prepositional phrases can be Focus-fronted, as we see in (52) and (53) (see also Ruth 2:15 and perhaps Ruth 3:11). (52) Ruth 1:10 (C-PP1-V-PP2; PP1=Focus) ki ’ittak nasûb l¢ämmek ‘Indeed! We shall return with you to your people’ (53) Ruth 2:21 (PP-V; PP=Focus) ïm hann¢äîm ’aser lî tidbaîn ‘with the lads that are mine you should stick close’

The book of Jonah contains three examples of single Focusfronting, in 1:5, 2:5, and 4:10. The short clause in 2:5 is a simple case of Jonah contrasting his actions (‘but I …’) with Yhwh’s. This type of Focus in chapter two is instrumental in establishing the accusatory and snivelling tone of the psalm in the Jonah’s mouth. At the end of chapter four Focus is used in Yhwh’s mouth to contrast his own concept of compassion with Jonah's. However, it is the example in 1:5, provided in (54), that is linguistically most interesting. (54) Jon. 1:5 (S-V-PP; ENTIRE CP = FOCUS) w¢yônâ yaad ’el yark¢tê hass¢pînâ ‘and Jonah went down into the rear of the ship’

In this verse we are presented with an SV clause that, in juxtaposition to the preceding statement that the ship’s sailors feared for their lives and were frantically trying to keep the ship afloat, asks the reader to contrast Jonah’s actions. It is not just the subject that is contrasted with the possible alternatives (e.g. the sailors) but also the predicate (e.g. with ‘lightening the ship’). Therefore, we should view this as a case of an entire clause (CP) being moved to the Focus domain. While the majority of clauses in narrative fit into one of the two categories we have covered, viz. cases of a single Topic or single Focus-fronted constituent, there are a few examples (and numerous examples in poetic texts) of multiple fronting. We have a single occurrence in the book of Ruth, given in (55): (55) Ruth 1:21a (SPRO-ADV-V; SPRO=Topic; ADV=Focus)43 ’anî m¢leâ halaktî ‘I went away full (but Yhwh returned me empty)’ 43

The possibility and existence of double fronting in ancient Hebrew dictates that we cannot classify this language as a V2 language, strictly speaking (as DeCaen, ‘Placement and Interpretation’, does). The presence of multiple items before the 135


The clause in (55) presents us with the first part of Naomi’s complaint, the second part of which we examined in (49). Not only is the adjective m¢le’â (used adverbially here) placed before the verb, so too is the personal pronoun ’ani. The pronoun orients the reader to the desired Theme — in the preceding verse, Naomi had just stated ‘Shaddai has made me very bitter’, which means that there were at least two possible thematic items from which to select. The fact that Naomi changes from Shaddai to herself as the subject of the next clause is the motivation for the Topic pronoun referring to herself. The adverbial m¢le’â is then Focus-fronted and is used to create the contrast between m¢le’â and rêqam that the next clause (presented in [49]) completes. While the author of Ruth did not frequently employ multiplefronting, the author of Jonah clearly found it to be a useful strategy. There are three examples of double fronting in the shorter book of Jonah (1:14; 2:9; 4:11), all of which are similar to what we saw in Ruth 1:21. More complex is the example of what is likely a triplefronting in Jon. 2:10, provided in (56). (56) Jon. 2:10a (S-PP-V-PP; S=Topic; PP=Focus; V=Focus) wa’anî b¢qôl tôdã ’ezb¢Ìã la ‘but I, with a voice of thanks, shall sacrifice to you’

In (56) the unambiguously modal verb ’ezb¢Ìâ is preceded by both the subject pronoun ’ani and prepositional phrase b¢qol todâ. Why? First, the Topic-fronted subject pronoun orients the reader to the fact that the next predication will concern the referent of ’ani (which is, of course, Jonah), not the previously modified ‘adherents of worthless idols’. Then the Focus-fronted PP b¢qol todâ presents a contrast between what Jonah’s manner of action and the implied manner of thanklessness or silence of those in the previous clause, Jon. 2:9. Finally, the verb ’ezb¢Ìâ also carries Focus in order to contrast how those from v. 9 ‘abandon’ their faithfulness while Jonah not only remains faithful but intends to ‘offer a sacrifice’ to God.44 verb obviously means that the verb cannot be in the second syntactic position in the clause. Hence, the motivation for using the more general reference ‘triggered inversion’ (after Shlonsky, Clause Structure). 44 It is possible that Ruth 1:14b is also a case of double Focus, with both Ruth and the verb ‘clung’ marked for a contrast with Orpah and ‘kissed’. However, in order for the verbal contrast to make sense logically, we must assume semantic gapping and reconstruct it: Orpah kissed her mother in law and left, but Ruth did not kiss (?) her mother in law and clung to her. Perhaps, but I am doubtful. Rather, this clause is best taken as a case of subject Focus fronting, or perhaps Focus on both the subject and verb. 136


From the data in Ruth and Jonah, it appears that for these authors the domain of fronting, what is also called the ‘left periphery’, has the structure given in (57), which nicely accords with a great deal of current research on the architecture of the left periphery and fronting phenomena. 3

(57)

0 0 0

CP

0 0

0

TopP

2

C

Domain of Triggers for raising V over canonical S position 0 0 0

Focus-fronted phrase(s)

1

FocP

0

Topic-fronted phrase(s)

TP core SV clause

Since at least the grammars of Ruth and Jonah exhibit multiple Focus-fronting, it must be that the Focus Phrase can project at least two distinct levels. What requires further study is whether ancient Hebrew exhibits multiple Topic-fronting, and whether we can discern an order within the Topic and Focus fields when multiple Topic and/or Focus phrases exist.45 Conclusion The majority of the clause types in Ruth and Jonah (or in any biblical book) do not fulfil the typological criteria used to determine basic word order, especially the basic clause type criterion. This means that at the centre of the VS versus SV basic word order argument stands a small set of clauses, but this is an important set if we want to classify ancient Hebrew typologically and account for information structure accurately. If we start with the VS position, a necessary position is that no SV clause lacks a Topic or Focus operator, but VS 45

For example, Beninca’ and Poletto, ‘Topic, Focus, and V2’, argue for Italian that ‘the encoding of informational relations in the syntax of the left periphery fol lows a very precise semantic path’, which they identify as: [Topic [Hanging Topic [Scene Setting Constituents [Left Dislocation [List Interpretation]]]]][Focus [Contr. Adverbs/Objects [Contr. Circum./Quant. Adverbs [Informational Fo cus]]]]. 137


clauses may be pragmatically neutral. In contrast, within the SV framework developed in this study, a few SV clauses may actually be basic and thus pragmatically neutral, but any VS clause without a syntactic or semantic trigger must contain a Topic or Focus operator. A case in point is the VS example from (30), repeated in (58). (58) Ruth 4:17 (V-S; V=FOCUS) yullad be l¢noömî ‘a son has been born for Naomi’

The conventional VS position would analyse this as a clause with no pragmatic marking. But from the context we already know that the boy has been born, and that Naomi has taken for herself some sort of caretaker role. Moreover, the specific context — an exclamation by the women of Bethlehem — suggests that this is no simple clause; rather, it is a statement of surprise, and what could be more surprising than old Naomi having a ‘son’.46 Thus, the reason for the Focus-fronting of the verb in (58) is to present a counter-expectation statement: Naomi, a widow who is presumably beyond the age of child-bearing (at least according to her impassioned assertion in 1:12), has, contrary to all expectations (including her own), ‘given birth’, and is thus, in the larger narrative, finally redeemed. Both this explanation of the VS example in (58), along with the fact that such clauses are very rare in Ruth and Jonah47 (as well as within all narratives in the Hebrew Bible), suggests that an SV model such as I have proposed here has greater descriptive and explanatory adequacy than VS models. One who adheres to the SV position has a rational explanation for the rarity of simple VS clauses: verbs are rarely focused in discourse; rather, nominal participants (whether agent or patient) or verbal modifiers (e.g. manner, location) are overwhelmingly the focused items. Thus, verbs are rarely raised to the 46 Note that if there were any Focus marking for Naomi, which is tempting to read, then we should have the active verb with Naomi as an explicit subject. In stead, the use of the passive clause in Ruth 4:17 makes it clear that the Focus is on the event, not the participants. 47 The sole VS example in Jonah is in the poem, in 2:6a: Jon. 2:6a (V S; chiasm) apapûnî mayim äd nepes t¢hôm y¢sob¢benî ‘(the) waters have encompassed me, up to (my) throat/life, the deep has surrounded me’ In this example, the syntactically and semantically non triggered verb is not Focus or Topic fronted either. Rather, this is a perfect example of chiasm at its most el egant. The first colon presents V S order and the second presents S V order, with the PP ‘up to throat’ acting as a Janus member, facing both cola.

138


left-periphery unless there exists a focus item that has triggered such movement; in other words, VS order is rare unless something else precedes the verb. In this article I have sketched a working model for investigating issues of word order variation in BH, a model that is built upon both typological and generative linguistics. Significantly, I have concluded that the data suggest a SV analysis for BH rather than the conventional VS analysis. Taking this empirically-driven SV conclusion as a starting point for an analysis of information structure, I described a framework for understanding the interaction of four core pragmatic concepts: Theme, Rheme, Topic, and Focus. The resulting model of core BH syntax, and the left periphery in particular, allows for the type of flexibility, including multiple-fronting structures, that BH exhibits.

139

NOTES SUR LES INSCRIPTIONS NEO-PUNIQUES DE HENCHIR MEDEINA (ALTHIBUROS) FRANÇOIS BRON EPHE, PARIS

Abstract Nouvelle étude de l’inscription néo punique de Henchir Medeina = Althiburos (KAI 159), suivie de notes sur la bilingue latine et néo punique provenant du même site (KAI 160).

Le site de Henchir Medeina, l’antique Althiburos, à trente-cinq kilomètres à l’ouest de Maktar, a livré jusqu’à présent quatre inscriptions puniques ou néo-puniques. Deux furent trouvées dès 1873 et expédiées au Louvre: l’une est une inscription néo-punique (AO 5106 = NP 124 = KAI 159). L'autre (AO 5184) est punique et a été réétudiée récemment par M. Sznycer.1 En 1908, lors de fouilles régulières, fut découverte une inscription bilingue, latine et néo-punique, malheureusement très fragmentaire (CIL VIII, 27774 = KAI 160). Enfin, en 1968, des fouilles tunisiennes ont mis au jour une inscription néo-punique, publiée de manière préliminaire.2 L’inscription AO 5106 est gravée sur un bloc mesurant 54cm sur 32cm; elle comporte un texte de sept lignes en écriture néo-punique, auquel ont été ajoutées deux lignes d’une autre main, d’une graphie très fruste. Mentionnée tout d’abord par J. Derenbourg dans une séance de l’Académie,3 elle a été publiée, de manière très succincte, par Halévy,4 puis par Euting5 et reprise par Sainte Marie.6 Mais les seules études approfondies sont celles de Ph. Berger.7 Par la suite, elle 1 M. Sznycer, ‘Une inscription punique d'Althiburos (Henshir Médéina)’, Semitica 32 (1982), 57 66 et pl. 8. 2 M. Ennaïfer, La cité d’Althiburos et l’édifice des Asclepieia (Tunis 1976), 17 18 et pl. IV. 3 J. Derenbourg, CRAI (1874), 306. 4 J. Halévy, ‘Appendice aux inscriptions libyques Inscription d’Altiburos’, JA (1874), 592 5. 5 J. Euting, ‘Inschriftliche Mittheilungen’, ZDMG 29 (1876), 235 40. 6 E. de Sainte Marie, Mission à Carthage (Paris 1884), 108 11. 7 Ph. Berger, ‘Note sur la grande inscription néo punique et sur une autre ins

141

NOTES SUR LES INSCRIPTIONS NEO-PUNIQUES

a été mentionnée par Clermont-Ganneau dans une étude plus générale:8 il a, en particulier, dans une note, indiqué plusieurs corrections de lecture et proposé de nouvelles interprétations, qu’il n’a plus eu l’occasion de développer par la suite. C’est lui également qui a proposé d’interpréter comme un samech le signe qui avait été lu jusqu’alors comme un tsadé.9 Cette étude est parue trop tardivement pour que Lidzbarski puisse en tenir compte dans son Handbuch et n’a été prise en considération ni par Cooke dans son Textbook,10 ni par W. Röllig dans KAI11 ni par K. Jongeling et R.M. Kerr dans leur récent opuscule.12 Pourtant, la photographie publiée par M. Ennaïfer vient confirmer l’exactitude de la plupart des lectures de ClermontGanneau.13 1) l-’dn B¨l Îmn b-’ltbrs ndr ’s ndr’ ¨bdmlqrt kns bn Kns’¨n w2) M¨rys bn Tbrsn w-S†mn bn Yksltn w-Mshb’ bn Lylÿ w-Ggm bn Ssy¨t w3) Msgm¨ bn Tbrsn w-Y¨smzgr bn Sbg w-’dnb¨l bn Yll w-Gzr bn Knzrmn w-M¨rys 4) bn Lbw’ w-Z¨lgm bn S†w¨n w-Y¨stbgw bn Mshb’ w-Ìbrnm hmzrÌ w5) n/tsmrn/t bn’t w-’yspn ¨lt mqdsm b-yrÌ Krr st Yll hzbÌ bn .g†¨n b 6) sp†m Mshb’ bn Yzrm w-¨zrb¨l bn Brk w-S..sln bn Z¨zbl w-Mbyw hspr ’s 7) ¨lnm bn Y¨†mn w-khn l-B¨l Îmn Wrwsn bn ’rs k’ sm¨ qlm brkm 8) ’s h¨l’ k’ ¨lt’ w-mnÌt b-mqds 9) ’s ¨d mlk sm ndb’ 1) ‘Au seigneur Baäl Îammon, à Althiburos, vœu qu’ont voué Äbdmilqart Kns (?), fils de Kns’¨n, 2) M¨rys, fils de Tbrsn, S†mn, fils de Yksltn, Mshb’, fils de Lylÿ, Ggm, fils de Ssy¨t, 3) Msgm¨, fils de Tbrsn, Y¨smzgr, fils de Sbg, Adonibaäl, fils de Yll, Gzr, fils de Knzrmn, M¨rys, cription d’Altiburos’, JA (1887), 457 71; idem, Inscription néo punique d'Altiburos (Ligne 8 et 9) (Paris 1891). 8 Ch. Clermont Ganneau, ‘Le “mazrah” et les curiae, collegia ou ordines cartha ginois dans le Tarif des sacrifices de Marseille et dans les inscriptions néo puniques de Maktar et d’Altiburos’, RAO, t. III (Paris 1900), 22 40. 9 Ibid, 332 3 et n. 1. 10 G.A. Cooke, A Textbook of North West Semitic Inscriptions (Oxford 1903), no 55. 11 H. Donner et W. Röllig, Kanaanäische und aramäische Inschriften (Wiesbaden 1962 4), no 159. 12 K. Jongeling et R.M. Kerr, Late Punic Epigraphy (Tübingen 2005), 39 40. See also now K. Jongeling, Handbook of Neo Punic Inscriptions (Tübingen 2008), 155 7. 13 M. Ennaïfer, op. cit. à la n. 2, pl. V. Nous remercions très vivement madame P. Kalensky, du Musée du Louvre, qui nous a procuré une nouvelle photographie de cette inscription (Fig. 1). 142


4) fils de Lbw’, Z¨lgm, fils de S†w¨n, Y¨stbgw, fils de Mshb’ et leurs collègues, (qui composent) le mzrÌ. 5) … sur les sanctuaires, au mois de Krr de l’année de Yll, le sacrificateur, fils de .g†¨n, avec 6) les suffètes Mshb’, fils de Yzrm, Äzarbaäl, fils de Barik et S.sln, fils de Z¨zbl; et Mbyw, le scribe, qui 7) est sur eux (?), fils de Y¨†mn; étant prêtre de Baäl Îammon Wrwsn, fils de Aris; parce qu’il a entendu leur voix, il les a bénis. 8) (Eux) qui ont offert ici son holocauste et la minÌa dans le sanctuaire 9) et qui, par surcroît, ont offert spontanément là un sacrifice molek.’

L. 1–4: après la dédicace à Baäl Îammon, le texte commence par l’énumération des douze dédicants, membres du mzrÌ. A quelques exceptions près, leurs noms et patronymes sont libyques et il est le plus souvent difficile de leur trouver des parallèles.14 Le premier porte l’épithète kns, dont l’interprétation reste controversée: s’agit-il d’un nom propre épithète — mais cette pratique est inhabituelle dans ce genre d’onomastique — ou bien d’un titre, formé sur la racine KNS, ‘rassembler’, qui désignerait le chef du mzrÌ — mais on attendrait l’article? On peut proposer quelques corrections de lecture: – à la l. 3, le premier nom est à lire Msgm¨ plutôt que M’gm¨, ce qui nous donne un nom de la série bien connue commençant par Ms-;15 – à la l. 4, les estampages aussi bien que la photographie invitent à lire le quatrième nom Y¨stbgw plutôt que Y¨st’n. (Euting lisait déjà Y¨Òtbgw). L. 5: après la liste des membres du mzrÌ, l’inscription se poursuit, dans l’interprétation de Berger, adoptée par Röllig, par une phrase nominale: ‘Et NSMRN, fils de ’T (??) et ’YSPN (étaient préposés) aux sanctuaires’. Outre que les noms propres semblent très étranges, une telle construction serait tout à fait inhabituelle. Cela avait conduit Clermont-Ganneau à y voir plutôt ‘une nouvelle petite phrase, où se cachent peut-être deux verbes, reliés entre eux par la conjonction waw’.16 Cependant, il ne s’est pas risqué à en donner une traduction, ‘(se proposant) d’y revenir à une autre occasion’, ce qu’il n’a jamais fait. L’intuition de Clermont-Ganneau était sans doute correcte, mais l’interprétation de ce passage reste très difficile: la lecture 14

On consultera K. Jongeling, Names in Neo Punic Inscriptions (Groningue 1984). 15 Cf. K. Jongeling, op. cit., 68 71. 16 Ch. Clermont Ganneau, op. cit. à la n. 8, p. 32 et n. 1. 143


du premier mot est loin d’être assurée, vu l’ambiguïté de plusieurs des caractères qui le composent. Est-ce une forme d’un verbe smr, ‘clouer’, attesté dans une inscription phénicienne de Chypre (KAI 43/13)? Faut-il lire ensuite un nom bn’t, formé sur la racine BNY, ‘construire’? Dans le mot qui suit, Clermont-Ganneau voyait une forme de yiphil, avec la graphie ’y-, bien connue en néo-punique,17 d’un verbe SPN, ‘couvrir’, qui a donné en phénicien les noms mspn (IPT 22/2 = KAI 122) et mspnt (KAI 10/6), ‘toit’. A la même ligne, lire, avec Clermont-Ganneau, pour le nom du sacrificateur, Yll au lieu de Bll. La fin de la l. 6 et le début de la l. 7 constituent une nouvelle crux. Voici la lecture et l’interprétation de Röllig: w-Mbyw hÒp’ s 7)¨l kmr Ny¨†mn, ‘et (c’était) MBYW, le voyant, qui (était préposé) aux prêtres de NY¨™MN’. La mention d’un Òp’ serait un hapax en phénico-punique. Quant à kmr, mot typiquement araméen, dont la lecture est due à Euting, sa présence sur une amphore phénicienne trouvée à Chypre et aujourd’hui perdue (RES 1519 B) semble très douteuse.18 Enfin la mention de la déesse égyptienne Neith, idée due, semble-t-il, à Cooke, paraît tout à fait incongrue. L’identification avec un dieu libyco-berbère Motmanius, connu par une inscription latine de Lambèse (CIL VIII, 2650), suggérée par E. Lipinski, n’est pas plus séduisante.19 Là aussi, Clermont-Ganneau avait proposé une interprétation plus sobre: Mbyw porte le titre bien connu hspr, ‘le scribe’;20 le resh, quoique ténu, se distingue sur les estampages. La ligne se termine par le pronom relatif ’s. Au début de la ligne 7, ClermontGanneau hésitait entre les lectures ¨lkm, ¨ltm ou ¨lnm. Cette dernière lecture, composée de la préposition ¨l, suivie du suffixe de troisième personne du pluriel, déjà attestée sur une inscription de Carthage (CIS 3920/4), semble la plus vraisemblable. On ne sait à qui se réfère le pronom suffixe: aux trois suffètes? à l’ensemble des membres du mzrÌ? Vient ensuite ‘l’indispensable patronymique’ (les italiques sont de Clermont-Ganneau) bn Y¨†mn, qui nous procure un nom libyque tout à fait acceptable. L. 7, lire Wrwsn, avec les Hollandais, et non pas WrwÒn, vieille lecture conservée par Röllig. 17 J. Friedrich, W. Röllig, M.G. Amadasi Guzzo et W.R. Mayer, Phönizisch Punische Grammatik (Rome 1999), par. 147a. 18 Cf. M.G. Amadasi Guzzo et V. Karageorghis, Fouilles de Kition III. Inscrip tions phéniciennes (Nicosie 1977), 184 5 (F 2). 19 E. Lipinski, Dieux et déesses de l'univers phénicien et punique (Louvain 1995), 373. 20 Clermont Ganneau, op. cit. à la n. 8, p. 332 3, n. 1.

144


Les l. 8–9 ont été l’objet d’une étude particulière de Berger,21 puis elles ont été réexaminées à plusieurs reprises par J. Février.22 Pour Berger, le rapport avec le texte précédent est clair: ‘c’est la constatation de l’accomplissement du vœu’. Cela est possible; cependant, à une époque où il est à la mode de dénoncer des faux, ces deux lignes ont de quoi susciter les soupçons: tout d’abord, leur graphie très maladroite surprend, à la suite d’un texte officiel. Ensuite, la présence d’une forme de hiphil est un hapax en phénico-punique. Enfin, le mot ¨lt, ‘holocauste’, n’est pas connu par ailleurs en phénico-punique, pourtant riche en termes relatifs aux sacrifices; en revanche, la paronomase heëlah ¨ôlah, ‘offrir un holocauste’, est bien attestée en hébreu biblique. L’inscription bilingue (KAI 160), découverte en 1908, a connu un sort funeste: exposée à l’air libre sur le site, elle a été effacée par les intempéries.23 Par la suite, la pierre a été transportée au Musée du Bardo et l’inscription republiée, à partir des photographies anciennes, par Z. Ben Abdallah.24 L’inscription latine avait été signalée tout d’abord par A. Merlin,25 puis le texte punique a été étudié par E. Vassel.26 Chabot y est revenu brièvement dans le Bulletin archéologique de 1944.27 Mais la notoriété de cette inscription est due à un article de Février, au titre frappant.28 Par la suite, malgré une note de Levi Della Vida,29 son interprétation a été reprise dans la première édition de KAI et par diverses publications archéologiques.30 Seules les deux premières lignes du texte latin se prêtent à une restitution: QVOD BONVM FAV[stum, fel]IXQVE SIT SOD[alibus

Le terme crucial est celui de sodalibus qui désigne les membres d’une corporation, d’un sodalicium. 21

Op. cit. à la n. 7. J. G. Février, ‘Molchomor’, RHR 143 (1953), 8 18; idem, ‘Essai de reconsti tution du sacrifice molek’, JA 248 (1960), 167 87. 23 Cf. M. Ennaïfer, op. cit. à la n. 2, p. 20 et n. 51. 24 Z. Ben Abdallah, Catalogue des inscriptions latines païennes du Musée du Bardo (Rome 1986), 148, no 380. 25 A. Merlin, BAC (1908), 234 et pl. 45. 26 E. Vassel, ‘Etudes puniques V. Sur la bilingue d’Althiburos’, Revue tunisienne (1916), 278 85. 27 J. B. Chabot, BAC (1944), 289 90. 28 J. G. Février, ‘Une corporation de l’encens à Althiburos’, Semitica 4 (1951 2), 19 24. 29 G. Levi Della Vida, ‘Iscrizione punica da Lepcis’, ANLR, ser. VIII, 10 (1955), 550 61: cf. p. 556, n. 2. 22

145


Quant au texte punique, voici la lecture qu’en propose Février: 1) 2) 3) 4) 5) 6)

’]s ytn’ lyt ’ktrt bÿm sk¨rnm ’]s ’ys lnqsmy kytb wbyt s¨t ….Ì]Ò’ nk’t swmyws kns’ s¨t ……]†’ S¨†rnyn’ bn S¨†r …….Ì]sb s¨t Brkb¨l bn Mtn ……..]s¨t

Voici maintenant sa traduction des deux premières lignes: 1) ‘(Ce) qu’a offert l’association de l’encens, avec le surplus (?) de leurs gains; 2) ce que chacun selon sa part a inscrit et à la Maison des Aromates (a apporté?)’. Février a voulu retrouver l’équivalent de sodalibus dans lyt, qui serait un nom formé sur la racine LWY, ‘accompagner’, et comparable à l’hébreu tardif lewiyyâ, ‘suite (d’un personnage), caravane’. Ensuite, ’ktrt serait pour ’q†rt, ‘l’encens’, avec confusion des emphatiques avec les sourdes. Enfin, il explique ÿm par un hapax hébreu, dans Is 11/ 15, qui aurait le sens de ‘violence’ ou ‘excès’. A la l. 2, il rapproche s¨t du minéen m†ÿ et de l’ougaritique †ÿ, ‘offrande d’aromates’. Tout cela paraît bien invraisemblable, en particulier l’interprétation de la première ligne. Levi Della Vida l’avait bien vu, qui, acceptant avec hésitation la lecture ’ktrt de Février, proposait de comprendre: ‘…à qui ils donnèrent le chapiteau durant leur vie, en mémoire d’eux’. La traduction ‘chapiteau’ (h. koteret) lui était inspirée par la présence du mot kt¨rt dans une inscription de Lepcis, IPT 31/2 = KAI 119. La graphie ÿm pour Ìym se retrouve dans KAI 134/3. L’orthographe sk¨r pour sk¨r n’est pas impossible (PPG3, par. 48b). La fin de la ligne évoque une stèle funéraire d’Athènes (KAI 53/1), mÒbt skr bÌym, ‘stèle du souvenir parmi les vivants’. Cependant, d’après notre connaissance actuelle du punique, ly ne peut signifier que ‘à moi’ et non pas ‘à lui’, qui s’écrirait l’ (PPG3, par. 254a); c’est ce qu’a bien vu Krahmalkov, dont voici la traduction: ‘tout homme qui m’a accordé le turban (de chef de la corporation), récompense-les durant leur vie’.31 Ktrt correspondrait à h. keter. Enfin, ne faut-il pas lire plutôt ’ktbt, ‘le document écrit’, terme déjà attesté dans le tarif de Marseille (KAI 69/17, 18) et dans une inscription de Lepcis (IPT 26/3 = KAI 124)? On obtiendrait ainsi la traduction: ‘ceux qui m’ont accordé le document écrit, durant leur vie, à titre de mémorial’. 30

Cf. M. Ennaïfer, op. cit. à la n. 2, p. 20 1; A. Krandel Ben Younès, La pré sence punique en pays numide (Tunis 2002), 213. KAI mentionne l’interprétation de Levi Della Vida en appendice, à partir de la seconde édition, t. II, 340 1. 31 Ch. Krahmalkov, A Phoenician Punic Dictionary (Louvain 2000), 247. 146


Les lignes suivantes ne sont pas plus faciles: le mot clé en est probablement s¨t. L’étymologie proposée par Février n’est guère convaincante, d’autant plus que le sens exact de la racine ™Ÿ est obscur, aussi bien en ougaritique qu’en minéen. Chabot avait suggéré que ce pourrait être, ‘dans un dialecte indigène, l’équivalent de sodalis’.32 C’est aussi l’idée de Krahmalkov, qui compare un akkadien se’u, ‘ami’.33 Ne serait-ce pas plutôt le correspondant d’ar. sîät-, ‘parti, réunion d’hommes’, sabéen s¨t, ‘communauté, partisans’? A la l. 2, byt s¨t désignerait donc ‘la maison de la corporation’. A la même ligne, précédé de l-, nqsmy ne peut être qu’un infinitif construit niphal, suivi d’un pronom suffixe. A la fin de la l. 3, il faut peut-être lire k-ns’ s¨t, ‘car la corporation a offert’, ce qui correspondrait au in] templo po[nere du texte latin. A la l. 4, Clermont-Ganneau lisait S¨†rnyn’ bn S¨†r, avec un samech, ce qui paraît justifié, vu la longueur de la hampe.34 S¨†rnyn’ est déjà attesté dans une inscription de Sardaigne (KAI 173/7), alors que S¨†rnyn’ n’est pas connu par ailleurs; les deux orthographes se rencontrent pour S/S¨†r. Enfin, à la l. 5, on aurait mention d’un ‘comptable de la corporation’, Ìsb ou m]Ìsb s¨t.

Fig. 1. 32

J. B. Chabot, op.cit. à la n. 27. Krahmalkov, op. cit., 476. 34 Lettre du 28 juillet 1913, citée par E. Vassel, op. cit. à la n. 26, p. 279. 33

147

THE MODALITY OF ∑ARÎK IN TANNAITIC HEBREW TZVI NOVICK NOTRE DAME

Abstract The Paper argues that the word Òarîk in tannaitic Hebrew does not, despite claims to the contrary, indicate permissibility. Analysis of apparent counter examples illustrates the relationship between tele ological necessity and weak deontic obligation.

In tannaitic Hebrew, modal Òarîk ordinarily indicates ‘necessary’. However, commentators have identified instances in which Òarîk seems to mean something like ‘permissible’. In this essay I review these and other relevant instances from the perspective of modern linguistics. My analysis establishes that the notion of permissibility does not belong to the semantics of Òarîk; its apparently anomalous usages yield to pragmatic explanations. The pragmatic range of Òarîk in fact bears witness to a modal phenomenon that has received insufficient attention, namely, the close relationship between the categories of necessity in teleological modality and weak obligation (as expressed, e.g., by English ought) in deontic modality. Usages of Òarîk Typical and Atypical The etymological meaning of the root Ò-r-k lies in the notion of lack or need.1 From here it is a short step to the core modal usage of Òarîk * I thank Elitzur Bar Asher, Dr Eliezer Diamond and Professor Christine Hayes for their comments on earlier drafts of this article. This essay was made possible in part by a grant from the Memorial Foundation for Jewish Culture. 1 See Ben Yehuda’s dictionary (A Complete Dictionary of Ancient and Modern Hebrew, [New York 1959]), 5631 6). For discussion of Ugaritic and Aramaic cog nates see E.Y. Kutscher, A History of the Hebrew Language (Jerusalem 1982), §123; Isaac Gluska, Hebrew and Aramaic in Contact During the Tannaitic Period: A Socio linguistic Approach (Tel Aviv 1999), 57 n. 338. For a full account of the verbal forms of the root in tannaitic Hebrew see Menahem Moreshet, A Lexicon of the New Verbs in Tannaitic Hebrew (Ramat Gan 1980), 312 13. 149

THE MODALITY OF ∑ARÎK IN TANNAITIC HEBREW

in rabbinic literature, teleological necessity: to accomplish the goal at hand, one thing ‘needs’ (Òarîk) another.2 Thus, for example, m. Îag. 2:3: ‘Vessels produced in purity need (Òerîkîn) immersion before being used for holy things but not before being used for the heave offering’.3 In most cases, the telos that defines the necessity is left unstated, to be inferred from context. Òarîk also extends to express deontic obligation, as in m. Maksh. 2:8: ‘One who finds an object in a place with a majority of gentiles need not (’ênû Òarîk) announce it; if in a place with a majority of Jews, he needs (Òarîk) to announce’.4 In some cases in tannaitic literature, deontic Òarîk appears to indicate not obligation but permissibility. I know of seven such cases identified by other commentators: t. Maäs. Sh. 5:16; t. Ëruv. 5:10; t. Ëruv. 5:18 (bis); t. Shab. 13:7; Sifre Num. 68; Sifre Deut. 16. Of the seven, three (t. Maäs. Sh. 5:16; t. Ëruv. 5:10; t. Ëruv. 5:18) are positive, and four (t. Ëruv. 5:18; t. Shab. 13:7; Sifre Num. 68; Sifre Deut. 16) involve negation. Leaving aside the positive instances for later, and the one negative instance (in t. Ëruv. 5:18) that occurs together with a positive Òarîk, we may take up first the three independent negative instances. T. Shab. 13:7 addresses what food may be rescued from a house in case of fire on the Sabbath. The principle underlying the pericope is that one may only rescue such food as one intends to use on that day. According to ms. Vienna, the pericope reads: ‘if he saved fine bread, he is not permitted (’ên rasa’y) to save coarse bread’ (on the assumption that there is enough fine bread to suffice for the Sabbath). But ms. Erfurt reads ’ên Òarîk instead of ’ên rasa’y. If ’ên Òarîk is not a mistake — the fact that it is the lectio difficilior argues in its favour — the phrase indicates, like ’ên rasa’y, impermissibility. The more com2 Such conditional sentences have been called ‘anankastic’, from Greek ânágkj. See George Henrik von Wright, Norm and Action (London 1963), 10; Kai von Fintel and Sabine Iatridou, ‘What to Do If You Want to Go to Harlem: Notes on Anankastic Conditionals and Related Matters’ (Cambridge, Mass. 2004), 1 15 (ms. available at http://web.mit.edu/fintel/www/harlem.pdf ). 3 Unless otherwise qualified, citations to the Mishnah use ms. Kaufman; to the Tosefta, ms. Vienna; to Mek. R. Ish., ms. Oxford 151; to the Sifra, ms. Assemani; to Sifre Num. and Sifre Deut., ms. Vatican 32. For another case of teleological Òarîk see m. Kel. 5:7, where Òarîk occurs in response to the question: ‘how does one pu rify an impure oven?’ See also MenaÌem Zevi Kaddari, ‘On Deontic Modality in Mishnaic Hebrew’, in Moshe Bar Asher (ed.), Studies in Mishnaic Hebrew (Scripta Hierosolymitana 37, Jerusalem 1998), 207 8, 213. 4 Outside of this case, lehakrîz ‘to announce’ is always preceded in tannaitic lit erature by Ìayyab.

150


mon meaning of ’ên Òarîk, ‘he need not’, would be nonsensical here, as it is in no case obligatory to save one’s food from a fire.5 The next instance, Sifre Num. 68 (Horovitz ed., p. 63), involves a debate about the identity of the characters in Num. 9:6 who had incurred corpse impurity and therefore did not participate in the Passover sacrifice. R. Ishmael opines that these were the bearers of Joseph’s coffin, while, in R. Akiva’s view, they were Mishael and ElÒafan, who had removed the bodies of Nadav and Avihu from the Temple (cf. Lev. 10:4). R. Isaac counters: ’ên Òarîk ‘that is not permissible (to say)’, because any of these figures would have contracted corpse impurity long enough before the Passover that ‘they could have purified themselves’. Rather, these were individuals who, immediately before Passover, had encountered and buried an abandoned corpse. As R. Isaac attempts to refute the views of R. Ishmael and R. Akiva, and not simply to offer his own alternative, it is clear that, with ’ên Òarîk, he means to claim that their views are not simply not necessary, but in fact not supportable.6 In Sifre Deut. 16 (Finkelstein ed., pp. 26–7), the last pericope alleged to attest negative Òarîk ‘not permissible’, it is reported that R. Ishmael, when acting as judge in a case between an Israelite and a non-Israelite, would always decide in favour of the Israelite, whether by applying Torah law or the law of the nations. R. Shimon b. Gamaliel comments: ’ênû Òarîk ‘it should not be done so’, but rather the judgment should be fair, under whatever law the litigants agree upon. Given that R. Shimon champions a different practice, we may rightly conclude that he means to characterize R. Ishmael’s approach not simply as unnecessary, but as incorrect.7 I have found many other clear instances of negative Òarîk with the approximate sense ‘not permissible’ in tannaitic literature. Two examples will suffice. In t. Hor. 1:7, an interpretation of a biblical verse is 5 On this passage see Saul Lieberman, Tosefta Ki Fshu†ah (3rd ed.; New York 2001), 2.211. 6 On this passage see Menahem I. Kahana, ‘The Halakhic Midrashim,’ in S. Safrai (ed.), The Literature of the Sages, Second Part (Philadelphia 2006), 37 8 and n. 157. In the Bavli parallel (b. Suk. 25b), the printed editions and ms. Munich 95 omit ’ên Òarîk, but mss Harley 5508 and Ebr 134 include it. 7 On this passage see Saul Lieberman, Tosefet Rishonim (Jerusalem 1937 9), 1.155; Kahana, ‘Halakhic Midrashim’, 37 n. 157. The medieval commentary of Pseudo Rabad (Herbert W. Basser [ed.], Pseudo Rabad: Commentary to Sifre Deuter onomy [Atlanta 1994], 19) glosses ’ênû Òarîk with ’ênû yakôl in his explication of the passage: ‘R. Shimon disagrees and says that if they came before an Israelite and said, judge us according to Israelite law, he is not permitted (’ênû yakôl) to judge according to gentile law to make the Israelite victorious’.

151


offered, then challenged, then defended by citation of a second verse. The conclusion follows: ‘thus one should not (’ên Òarîk) say according to the latter view (the challenge to the interpretation) but according to the former view (the original interpretation)’.8 The point is surely not that the challenge need not be accepted — it has, after all, been refuted — but that it should not be accepted. The other instance occurs in Midr. Tannaim Deut. 6:6 (Hoffmann ed., p. 26). According to R. Eliezer, the word ’eleh in Deut. 6:6 indicates that, in the liturgical recitation of the surrounding pericope, one need have conscious intent only until one reaches Deut. 6:6. R. Akiva responds: ’ênû Òarîk, for the continuation of the verse indicates that the section as a whole, even beyond Deut. 6:6, requires conscious intent. The phrase ’ênû Òarîk evidently indicates not that R. Eliezer’s view is unnecessary, but that it is incorrect.9 Thus ms. Erfurt; ms. Vienna has ’ên Òorek. See infra note 13. Other clear examples include Sifre Num. 78 s.v. nôs¨îm ’anaÌnû (Horovitz ed., p. 75); Midr. Tannaim Deut. 6:4; 33:7; 33:26; 34:5 (Hoffmann ed., pp. 24, 214, 221 2, 224). I also think it likely (pace the translation in Jacob Neusner, The Tosefta: Translated from the Hebrew with a New Introduction [2 vols; Peabody 2002 (1977 86)], 537) that t. Sheq. 3:22 provides another instance. R. Shimon is assert ing not that the proselyte need not set aside a bird offering in the absence of the Temple, but that he should not. Evidence for this interpretation comes not only from the fact that the reason supplied by R. Shimon argues positively against set ting aside the offering (and not merely against a requirement to set aside the offer ing), but also from the Bavli and Yerushalmi parallels (collected in Lieberman, Tosefta Ki Fshu†ah, 4.712), which have R. Shimon ‘nullify’ the practice of having proselytes set aside bird offerings. Another probable example of negative Òarîk in the sense of ‘not permissible’ occurs in Mek. R. Ish. Ba Ìôdes 2 (Horovitz ed., p. 210): ‘Rabbi says: we ought not (’ên … Òerîkîn) to make Moses great if it requires that the Holy One, Blessed be He, retracted his word’. See the identical statement in Mek. R. Ish. Ba Ìôdes 4 (Horovitz ed., pp. 217 18). For another possible example, see m. Yoma 4:1, which concerns the Temple ritual on Yom Kippur. According to the first, anonymous view, the high priest is supposed to declare, over the he goat se lected by lot as a sin offering, ‫‘ לי"י חטאת‬for God as a sin offering’. R. Ishmael re sponds: ‘he should not have (lô’ hayâ Òarîk) said “as a sin offering” but only “for God”’. From the Mishnah text alone, one could probably, with equal plausibility, render lô’ hayâ Òarîk in R. Ishmael’s response as: ‘he need not have said’. But from the Sifra parallel (’AÌarê Môt 2:2) (Weiss ed., 81a) it is clear that R. Ishmael not only licenses the omission of ‘a sin offering’ but precludes it outright: ‘He should not have said “as a sin offering”; he says only “for God”’. (The Mishnah pericope is also paralleled in t. Yoma 2:10.) It is therefore more plausible to interpret lô’ hayâ Òarîk as signalling preclusion, at least in the Sifra version. The same construction occurs in very similar contexts in two other pericopes in the Mishnah, at m. Taän. 2:3 and m. So†. 9:6. Finally, a discussion in Midr. Tannaim Deut. 24:16 (Hoffmann ed., p. 160) reaches the conclusion that children under the age of thir teen are penalized neither by the earthly court nor by the heavenly. The question then arises: ‘If this is the case, then infants should not (lô’ Òerîkîn) die, and why 8 9

152


Accounting for the Atypical Negative Usages Is there a philological basis for Òarîk with the meaning ‘permissible’? A semantic shift from the notion of need or necessity to permissibility is, to my knowledge, cross-linguistically unattested. The closest approach to the latter notion in the context of tannaitic Hebrew comes with the word Òôrek. The Aramaic cognate, Òerôk, is amply attested with the meaning ‘power, capability’ in Targum Jonathan (TJ). Thus, for example, Jer. 14:22 ‘are there, among the vanities of the peoples, rain-bringers’, is rendered in TJ as: ‘behold there is no power (Òerôk) in the errors of the peoples to rain down rain’. Likewise, the king of Israel’s lament in 2 Kgs 5:7, ‘am I God, that I may kill and enliven?’, occurs in TJ as: ‘is there in me the power (Òerôk) from God to kill and enliven?’.10 In tannaitic Hebrew, the clearest instance of Òôrek ‘power, capability’ occurs in Sifre Zuta Num. 5:31 (Horovitz ed., p. 239). In this pericope it is supposed that an adulterous woman who, because of her other virtues, merited to pass the test of the So†ah waters, will tell her friends: ‘Do not hesitate to sin. I drank and the waters did not harm me. It seems that they have no power (Òôrek)’. The Hebrew construction here, ’ên bahen Òôrek ‘they have no power’, precisely matches the Aramaic lêt/’ît behôn Òerôk ‘they have (no) power’ in the above instances from TJ. I have also found three other pericopes in which, as in TJ, foreign gods are denied Òôrek .11 then do they die?’ This case is quite certain, but its status as tannaitic is question able. I have collected five instances in post tannaitic literature where Lieberman makes a claim for Òarîk ‘possible, permissible’. Three (y. Ëruv. 2:3 [20a]; b. Shab. 121a; b. Yoma 88a) involve negation and two (y. Pes. 1:1 [27a]; y. Gi†. 6:2 [48a]) are positive. See Saul Lieberman, Yerushalmi Ki Fshuto (Jerusalem 1934), 252, 369 71; idem, Tosefet Rishonim, 1.155; idem, Tosefta Ki Fshutah, 2.822; 3.855). The three negative cases are plausible. On the two positive instances see n. 22 below. 10 The use of Òerôk in connection with divine power, as in the above examples, is very prevalent in TJ: when a prophet rejects the divine status of foreign gods or kings (e.g., Isa. 37:19; Ezek. 28:2; Hos. 8:6), TJ denies of them Òerôk. See also TJ to Judg. 6:31; 1 Kgs 18:21; 18:24; 2 Kgs 19:18; Isa. 41:23; Jer. 2:11; 10:11; 16:20; Jon. 1:5; Hab. 2:18. In all of these cases, where Òerôk is not followed by an infinitival complement, it is possible to interpret the word as ‘benefit’ rather than ‘power’, but in light of the examples cited in the body of the text, where an infinitival complement follows, and the meaning ‘power’ seems clearly called for, it seems reasonable to assign the meaning ‘power’ in these cases also. 11 Mek. R. Ish. Ba Ìôdes 6 s.v. lô’ yihyeh (ed.Horovitz, p. 223) = Midr. Tannaim Deut. 5:7 (Hoffmann ed., p. 20); Mek. R. Ish. Ba Ìôdes 6 s.v. kî ’anôkî (Horovitz ed., p. 226); Sifre Deut. 318 (Finkelstein ed., p. 364) = Midr. Tannaim Deut. 153


Semantic development from capability to permissibility is attested in such roots as y-k-l. Can the semantics of Òôrek / Òerôk, then, support the assignment of the meaning ‘permissible’ to Òarîk? There are reasons to doubt. First, Òôrek is never attested with the meaning ‘permissibility’ in tannaitic Hebrew (or, to my knowledge, in any other stratum of Hebrew). Second, even Òôrek ‘power, capability’ is, to my knowledge, attested only in the above instances, so that it may be bracketed as an Aramaic calque that achieved limited circulation, at least in the register of literary tannaitic Hebrew.12 Finally, what is true of Òôrek is not necessarily true of Òarîk.13 32:17 (Hoffmann ed., p. 195). In the second of the three cases, Òorek alternates with koaÌ ‘strength, power’. Another possible example occurs in Sifre Deut. 43 (Finkelstein ed., p. 98), where a king instructs his son, who is about to visit a ban quet: ‘do not eat more than your Òorek, do not drink more than your Òorek’. The son ignores his father, and eats and drinks ‘more than his Òorek’, and ends up vomiting. It seems more reasonable to assume that Òorek here indicates one’s capability rather than one’s need, as one can ordinarily eat and drink far in excess of one’s need with out vomiting. Presumably, the semantic development of Òôrek and its Aramaic cog nate proceeded in two stages: ‘a want, a need’ > ‘that which supplies a need, some thing that confers benefit’ > ‘power, capability’. The first stage is widely attested in Hebrew and Aramaic, for example, in Sifre Deut. 324 (Finkelstein ed., p. 375) ‘fat things in which there is Òôrek’, which Reuven Hammer (Sifre: A Tannaitic Commen tary on the Book of Deuteronomy [New Haven 1986], 336) renders as ‘fat things which are nourishing’; in TJ, e.g., at Jer. 13:7; and in Samaritan Aramaic, as per Z. Ben Hayyim, The Literary and Oral Traditions of Hebrew and Aramaic Amongst the Samaritans (Jerusalem 1957 79), 3.2.179, 197. I have not found evidence of the second stage outside TJ and the noted rabbinic sources, and it appears to have gone unnoticed by lexicographers. The entries under Òôrek in Ben Yehuda’s dictionary (Dictionary, 5640 6) and under Òôrek and Òerôk in that of Marcus Jastrow (A Dic tionary of the Targumim, the Talmud Babli and Yerushalmi, and the Midrashic Litera ture [New York 1992], 1271, 1300) do not include the meaning ‘power, capability’. Nor is this meaning recognized in translations of Targum Jonathan. Thus Daniel J. Harrington and Anthony J. Saldarini (Targum Jonathan of the Former Prophets: In troduction, Translation and Notes [Aramaic Bible 10; Wilmington 1987], 273), for example, render the above phrase from 2 Kgs 5:7 thus: ‘Is there need from before the Lord for me to kill and to let live?’ 12 Note that in the clearest case, Sifre Zuta Num. 5:31, it occurs in the mouth of a woman conversing with her friends, i.e., in a colloquial context. 13 There is some confusion, in the transmission of Ben Sira and tannaitic litera ture, between Òôrek and Òarîk. See Z. Ben Hayyim, ‫מסורת השומרונים וזיקתה למסורת‬ ‫הלשון של מגילות ים המלח וללשון חז"ל‬, Lesonenu 22 (1958), 229 n. 15‫ ;א‬Hanoch Yalon, Introduction to the Vocalization of the Mishna (Jerusalem 1964), 102 4. It is note worthy, however, that, while mss Kaufman and Parma of the Mishnah usually fa vour ’ên Òôrek lômar over ’ên Òarîk lômar (Yalon, Introduction, 103), in the three Mishnah pericopes cited above, supra at note 9, as attesting negative Òarîk with the meaning ‘not permissible’, namely, m. Taän. 2:3; m. Yoma 4:1; and m. So†. 9:6, mss Kaufman and Parma consistently record Òarîk. 154


Given the absence of a solid philological basis for Òarîk ‘permissible’, we should explain negative Òarîk ‘not permissible’ on pragmatic grounds. What is occurring in the above cases is that negative Òarîk ‘not necessary’ is pragmatically strengthened, by an accepted convention of usage, so that hearers understand it to convey not only that the predicate is not required (which is what it literally says), but that the predicate is, moreover, impermissible, excluded.14 The strengthening convention is presumably rooted in considerations of politeness: it is more civil to make the weaker assertion that one need not rather than the stronger assertion that one should not. Further support for this conclusion comes from a closer look at one of the examples examined above, from t. Hor. 1:7: ‘thus one should not (’ên Òarîk) say (judge) according to the latter view but according to former view’. This statement, in a slightly different version, constitutes a standard terminus in the exegesis of the school of R. Ishmael: ‘thus you should not (’ên ‘aleyka) say (or judge) according to the latter view but according to former view’.15 The Tosefta’s ’ên Òarîk corresponds to ’ên ‘aleyka in the latter terminus. If we explain ’ên Òarîk ‘not permissible’ by supposing that Òarîk is semantically ambiguous between ‘necessary, required’ and ‘permissible’, then we must likewise suppose that the construction äl x is semantically ambiguous between ‘it is incumbent upon so-and-so’ and ‘it is permissible for so-and-so’, but the latter meaning is never attested. Rather, in both cases, the ambiguity is a product of the pragmatic strengthening convention that operates upon the negative collocation. The resulting pragmatic ambiguity is attested in parallel constructions in other languages. Thus, for example, the Italian sentence, Gianni non deve andare a Roma, can mean either that Gianni need not go to Rome (with no pragmatic strengthening) or that Gianni should not go to Rome (with pragmatic strengthening).16 In modern Hebrew, too, hû’ lô’ Òarîk la-leket can indicate that the subject need 14 For extensive discussion of the pragmatics at work see Laurence R. Horn, A Natural History of Negation (Chicago 1989), 337 61. Syntactic explanations of the phenomenon (which in syntactic terms is called ‘NEG raising’) are also available for discussion see Ferdinand de Haan, The Interaction of Modality and Negation: A Typological Study (New York 1997), 13 14 but the pragmatic explanation is to be preferred because negative Òarîk can and often does indicate the absence of a re quirement. The availability of the strengthened interpretation (‘impermissible’) de pends on the discourse context. 15 See, e.g., Mek. R. Ish. Nezîqîn 2 (Horovitz ed., p. 252); Sifre Num. 24 (Horovitz ed., p. 30). 16 See De Haan, Interaction, 93, and see generally 86 109.

155


not go or that he ought not go.17 We may likewise think, in English, of not necessary in such contexts as the irked parent who responds to his toddler’s overturning a dinner plate with: ‘That really wasn’t necessary’, i.e., the toddler should not have done that. The ease with which negative Òarîk allows for pragmatic strengthening suggests that Òarîk, in its deontic usage, often conveys weak obligation, like English ought, rather than strong obligation, like English must, for this sort of pragmatic strengthening is cross-linguistically much more common with weak universal modals rather than with strong ones.18 Many of the examples of ’ên Òarîk ‘not permissible’ collected above are more naturally rendered ‘ought not’ than ‘may not’. Likewise, the exegetical formula ‘it ought not say’ (i.e., the verse in question should be written differently) can be expressed either with Òarîk (as in Sifra BeÌûqôtay 1:1 [Weiss ed., p. 111a]) or with ra’ûy ‘should’ (as in t. So†. 6:11; Sifra Mîlu’îm 1:2 [Weiss ed., p. 45b]; Sifre Num. 86 [Horovitz ed., p. 85]). The usage of a term with roots in teleological necessity to express weak obligation is readily comprehensible. Teleological necessity appears ‘weaker’ than deontic obligation in that it is less peremptory: it does not issue directives, but instead purports to make objective claims about relationships between means and ends. Positive Òarîk ‘Ought’ If the above analysis is correct, then we should expect to find cases of positive Òarîk with the meaning ‘ought’. Let us turn first to the three 17

Laurence R. Horn, ‘Remarks on Neg Raising’, in Peter Cole (ed.), Syntax and Semantics, Vol. 9: Pragmatics (New York 1978), 199. The example of Greek xrß ‘need, necessity’ is also instructive. (For the translational equivalence of xrß and Ò r k, compare Tob. 3:1 in Aramaic [4Q197 3, 1 Ò]ryk lk] and Greek xreían ∂xeiv].) In Od. 3.14, Athena informs Telemachus that he need feel no shame (‘there is no need of shame for you’) on account of his father’s absence, because, having travelled to Nestor’s home, Telemachos is now in a position to find out about his father. Nega tive xrß here clearly indicates the absence of a need; Athena’s point is that, until now, Telemachus had reason to feel shame, but this reason no longer obtains. Thus the Murray Dimock translation [The Odyssey: Books 13 24 (LCL 105; Cambridge, MA 1995), 81] clarifies by adding a temporal marker: ‘Telemachus, no longer need you feel shame’. (Cf. ∫piqen in 2.270, 278.) But compare Il. 9.496, where Phoenix here tells Achilles: ‘So, Achilles, master your great spirit, nor is there need for you to have a pitiless heart’. The first clause indicates that negated xrß in the second clause has the force of ‘should not': Achilles is being asked to control his passions and show pity. 18 See Horn, Natural History, 324 8. As Horn notes (328), this generalization is weaker in the case of deontological as opposed to other kinds of modality. 156


cases of alleged Òarîk ‘permissible’ listed above. T. Ëruv. 5:10 begins by indicating that if a father and his dependents live in different homes in a single courtyard, ‘they do not need (’ên Òerîkîn) to make an eruv’ in order to carry. The reason is that, given the others’ dependence on the father, their homes do not constitute legally distinct domains, so that the entire courtyard counts as a single, private domain owned by the father. However, ‘their alleyway [into which their courtyard leads] Òarîk a post and crossbar’. The parallel in the Bavli (b. Ëruv. 73a) reads: ‘their alleyway is rendered permissible (nîtar) [for carrying] by means of a post and crossbar’. Thus Saul Lieberman concludes that the force of Òarîk in t. Ëruv. 5:10 is ‘becomes permissible’.19 However, Òarîk here can more easily be assigned its basic sense of teleological necessity: the Tosefta is indicating that the alleyway requires a post and crossbar if one wishes to carry. The second alleged instance of positive Òarîk ‘permissible’ occurs in t. Ëruv. 5:18. This pericope treats of a situation in which one or more households in the courtyard has not participated in the eruv. For carrying to be permitted in this circumstance, such households must let their homes to one of the households that has participated, either through a legitimate rental arrangement or by a nominal declaration ceding control over the home (a legal fiction). The Tosefta states a basic distinction between Jews and gentiles in this regard: ‘A Jew cedes control, but in the case of a gentile, [the eruv will not work] until he rents’. Only Jews, not gentiles, can take advantage of the legal fiction. Immediately prior to this line, the Tosefta declares, according Lieberman’s rendering: ‘a Jew who publicly desecrates the Sabbath may not (’ên Òarîk) cede control, and a Jew who does not publicly desecrate the Sabbath may (Òarîk) cede control’.20 But it is equally or more plausible to render these modals instead as, respectively, ‘should not’ and ‘should’. The Jew who publicly desecrates the Sabbath may, technically, make use of the legal fiction, but the observant Jews who use the eruv should not rely on this technicality, but should instead treat him like a gentile, and enter into a rental arrangement with him. Likewise, the observant Jew ‘should’ cede control, that is, not only can he do so, but he should do so, and not either refuse to cede control or insist on rent. The Tosefta conveys the same quasi-obligation in two earlier pericopes: ‘it is incumbent (miÒwâ) upon one to cede control’ (t. Ëruv. 5:11); ‘it is incumbent (älayw) on a member of the courtyard who forgot to participate in 19 20

Lieberman, Tosefta Ki Fshu†ah, 2.396. Lieberman, Tosefta Ki Fshu†ah, 2.402. 157


the eruv, to cede control’ (t. Ëruv. 5:12).21 Thus, while the distinction between Jew and Gentile occurs in teleological universals — the Jew can make the eruv work by ceding control, the Gentile cannot — the distinction between observant Jew and apostate occurs in weak deontics: the former should cede control, and the latter should not. The last pericope, t. Maäs. Sh. 5:16, tells a story involving R. Eliezer, who owned a vineyard near Lod. Produce of a vineyard’s fourth year must be consumed in Jerusalem. Grapes grown at some distance from Jerusalem may be redeemed on money, which must then be taken to Jerusalem and spent there on food. Rabbinic decree at first permitted redemption only for vineyards beyond a day-trip’s distance from Jerusalem, but later extended this license to all vineyards beyond the city’s walls (t. Maäs. Sh. 5:14–15). The Tosefta tells that R. Eliezer ‘did not want’ to redeem his vineyard. His students told him: ‘now that, by decree, all vineyards up to the wall of Jerusalem are subject to redemption, you are Òarîk to redeem it’. R. Eliezer then arose, and harvested and redeemed his crop. On Lieberman’s account, R. Eliezer, a traditionalist, was reluctant to rely on the license to redeem. But his students informed him that the license to redeem entailed a stricture — it prohibited the sale of one’s fruit within Jerusalem — and therefore R. Eliezer could (Òarîk) redeem, and need have no qualms about relying on a lenient ruling.22 Lieberman’s reading requires extensive inference beyond the plain text. It seems far easier to suppose that Òarîk here indicates propriety. R. Eliezer’s students tell him that because the rabbis have now decreed redemption permissible even to the wall of Jerusalem, it is meet that he do so.23 21

On t. Ëruv. 5:11 see Lieberman, Tosefta Ki fshuta, 2.397 8. Lieberman, Tosefta Ki Fshu†ah, 1.785. For R. Eliezer’s tendency toward strin gency in legal practice, see Yitzhak D. Gilat, R. Eliezer ben Hyrcanus: A Scholar Outcast (Ramat Gan 1984), 60 7. Lieberman’s exegesis seems in part motivated by the aim of aligning the Tosefta story with its parallel in the Babylonian Talmud (R. ha Sh. 31b; BeiÒ. 5a). According to the Bavli, R. Eliezer wished to convey his vine yard to the poor. His students informed him that his colleagues had already ‘per mitted’ (hatîrûhû) his vineyard. For further explanation see Rashi, ad loc., adopted by Lieberman. Lieberman sees an equivalence between ‘permitted’ in the Bavli ver sion and the Tosefta’s ‘you are Òarîk to redeem it’. 23 The two cases of positive Òarîk ‘permissible’ in the Yerushalmi alleged by Lieberman (see above n. 8) are likewise dubious. On y. Pes. 1:1 (27a), two alterna tive interpretations of the question in this passage, ‘as for dark alleyways, need/may (Òarîk) one inspect them ab initio by candlelight?’ are described ad loc. in the com mentary Qorban Ha ëdâ, and there seems no reason to favour one over the other. The solution to the other positive Òarîk, y. Gi†. 6:2 (48a) (‘a minor girl who is Òerîkâ 22

158


In this last case, Òarîk shades almost into the notion of permissibility. David Daube’s analysis of Latin oportere offers an enlightening parallel: Oportere signifies ‘it is correct, proper’. Hence it may refer not only to what is necessary, to duties, but also to what is permitted, to rights. I may say ‘The correct thing is to wear evening dress’, i.e. it is necessary; but I may also say, ‘It is perfectly correct to wear a lounge suit’, i.e. it is permitted.24

Thus, for example, the jurist Ulpian determines that if a party wall is defective, ‘it was assuredly proper (oportuit) to pull it down’. But Daube acknowledges that ‘even when oportet refers to rights, it does not express complete freedom of choice. “The proper” always retains a trace of “the requisite”’.25 In t. Maäs. Sh. 5:16 too, R. Eliezer’s students do not simply state that R. Eliezer may redeem his fruits, but that he may properly do so. They are not offering legal advice, but attempting to persuade him by characterizing redemption of the fruit as the ‘correct thing’ to do. Positive Òarîk ‘ought’ occurs in a number of other cases. For example, R. Judah b. Baba is quoted (t. Ter. 5:10) as insisting that ‘courts should (Òerîkîn) conclude that all become nullified in one hundred and one’, i.e., that heave-offerings of all species lose their status when they fall into a mixture one hundred times larger than it.26 Likewise, t. B.Q. 9:29 dictates that one who is injured by another should (Òarîk) seek God’s mercy on behalf of the injurer even if the injurer has not sought it. to be divorced’), is probably to be found in y. Yev. 13:2 (13c), where the same pericope is introduced in connection with a structurally parallel line from the Mishnah (‘a minor girl who is Òerîkâ to refuse’) that uses Òarîk in the ordinary sense of ‘necessary’. 24 David Daube, Forms of Roman Legislation (Oxford 1956), 12. 25 Daube, Forms, 14. 26 Roger Wertheimer (The Significance of Sense: Meaning, Modality, and Moral ity [Ithaca 1972], 120 note *) observes that ought is rare in contemporary legislative English. The one exceptional case that he cites seems quite analogous to t. Ter. 5:10: ‘The grand jury ought to find an indictment when all the evidence before it, taken together, is such as in its judgment would, if unexplained or uncontradicted, warrant a conviction by trial jury’. Wertheimer explains the significance of ought here: ‘Given the special role of the grand jury … the intent and rationale of this section seem clear enough. While the grand jury is not bound (‘must’ or ‘shall’) to find an indictment in such cases, it is not simply empowered or permitted to do so. Nor should one say that it is left to the grand jury’s discretion. Rather the grand jury is provisionally or presumptively bound: i.e., they will find an indictment in such cases, but if, in their judgment, there is some good reason for not doing so, they are permitted to refrain’. 159


Conclusion I have argued that Òarîk, as a modal word, stretches from teleological necessity (‘necessary’) to strong obligation (‘must’) to weak obligation (‘ought’). This range of usages, all readily explainable on pragmatic grounds, makes Òarîk ambiguous. Does negative Òarîk, in any given case, indicate the absence of teleological necessity (or of a strong obligation) with respect to the act in question, or a weak obligation not to perform that act? Does positive Òarîk indicate teleological necessity, strong obligation, or weak obligation? I have not attempted to identify how, if at all, tannaitic Hebrew distinguishes contextually among these usages.27 In most cases the possibility of confusion is minimal. But it is worth noting that almost all the cases of Òarîk ‘ought’ collected here, positive and negative, occur in the Tosefta and the tannaitic midrashim, and very often in conversation. This distribution suggests that Òarîk ‘ought’ has a colloquial flavour, as already implied by its origin in the pragmatics of politeness.28

See Kaddari, ‘Deontic Modality’, 211, on Òarîk: ‘Sometimes society is the source of the obligation and its essence is good advice, as in … “Three things a man must [Òarîk] say in his house on Sabbath eve, towards dusk” [m. Shab. 2:7]; … “for a man must [Òarîk] satisfy people just as he has to satisfy the Almighty’ [m. Sheq. 3:2]”.’ While many examples of Òarîk ‘ought’ do have an interpersonal ele ment, it is questionable whether society should be described as ‘the source of obli gation’; we may note that m. Sheq. 3:2 cites two scriptural prooftexts. In any case, both of the pericopes cited by Kaddari begin with Òarîk ’adam ‘one should’. It is probable that the occurrence of ’adam as subject plays at least some role in signal ling the weak deontic force of these declarations. Statements beginning Ìayyab ’adam ‘one must’ likewise, though to a lesser extent, tend to convey obligations that are in one or another sense weak; see, e.g., m. Ber. 9:5: ‘One must bless over the bad as one blesses over the good’. 28 It seems that Òarîk expressing strong obligation also tends to occur with slightly greater frequency in the Tosefta than in the Mishnah. Thus, for example, in the Mishnah, leässer is preceded in three pericopes by Ìayyab (m. Dem. 6:11 12; m. Gi†. 5:4) and never by Òarîk. In the Tosefta, Òarîk leässer occurs in nine pericopes (t. Dem. 8:1 [paralleling m. Dem. 6:12]; t. Maäs. 2:5; 2:10 12; t. Maäs. Sh. 3:7 8; t. Shab. 15:14) and Ìayyab leässer in only two pericopes, one paralleling Ìayyab leässer in the Mishnah (t. Dem. 1:4 [paralleling m. Dem. 6:11]; t. Maäs. 2:1). Likewise, t. Pes. 5:7 corresponds to m. Pes. 8:2, but the former twice has ‘one need (Òarîk) not perform the Second Passover’ while the latter twice has ‘one is exempt (pa†ûr) from performing the Second Passover’. More generally, in the parallel dis cussion of the Second Passover in t. Pes. 7 8 and m. Pes. 8 9, the Tosefta without exception uses ’ên Òarîk while the Mishnah equally consistently uses pa†ûr or an imperfect verb. This distribution may suggest that Òarîk ‘must’ was also perceived as colloquial, though not to the same degree as Òarîk ‘ought’. 27

160

TALMUDIC ARAMAIC FAUNA NAMES: MURZEMA AND SHAQI™NA SHLOMY RAISKIN BAR-ILAN UNIVERSITY

Abstract This article investigates the origin and meaning of the Aramaic fauna names murzema and shaqi†na, mentioned in the Babylonian Talmud (Tractate Îullin, p. 63a). Beginning with Akkadian and Sumerian, the article surveys various roots in Semitic languages, as well as Persian. Various commentaries on the Talmud as well as seveal more contemporary sources indicate that the two birds can both be identified as the Greater Flamingo. The final part of the article concludes that the murzema is indeed the Greater Flamingo, which was called by hunters in Arabic al mirzam.

Introduction Tractate Îullin of the Babylonian Talmud, pages 59a–65a, presents a discussion regarding the permissibility of various fauna for consumption and at times provides lists of specific types which are or are not permissible. Îullin 63a quotes the great Amoraic Sage, Rav Yehuda, regarding a bird named shaqi†na: ‫ שקיטנא אריכי שקי‬:‫אמר רב יהודה‬ ...‫ מורזמא‬:‫ וסימניך‬,‫‘ וסומקי שריא‬Rav Yehuda said: Shaqi†na which is long-legged and red – is permitted, and the [mnemonic] sign [for this] is: murzema’. R. Shlomo Yitzhaki (Troyes [France], 1040–1105, commonly known as Rashi), in his Talmudic commentary explains that it is permitted to eat a bird named shaqi†na whose legs are long and red (since there are two types of this species, each having distinct external features), and this halachic Law was given a mnemonic Sign: murzema, since murzema is also the name of a permitted bird with long legs and a red body. This article will attempt to uncover the meaning of the names shaqi†na and murzema and the identity of these birds, which may or may not be synonymous.1 1

Various manuscripts exist regarding this excerpt. According to mss Hamburg 161

TALMUDIC ARAMAIC FAUNA NAMES: MURZEMA AND SHAQI™NA

The origin of the word ‘Shaqi†na’ A bird named saqatu appears in Akkadian sources, according to CAD (the Chicago Assyrian Dictionary) ‘a wading bird’.2 The word’s Sumerogram is gìr.gíd.damusen, which can be read sà-qa-tum, but also se-ep-sú a-rik or se-ep a-rik (i.e. sep arik). Reference to this bird is made in Niek Veldhuis’s book on Sumerian bird names (under the name gir3-gid2-damusen): This bird name literally means ‘long foot,’ and may refer to a wader. Although the item appears only once in the old Babylonian period, it is well known from first millenium lexical lists where it is rendered saqatu or sep(su) arik in Akkadian.3

The bird’s name, saqatu (similar to saqi†na), and its meaning ‘long foot’, both indicate a high likelihood this is indeed the Talmud’s saqi†na. Moreover, the second name, sep(su) arik, is similar to the Talmud’s description of this bird as being arikhei saki — meaning: long legged, and perhaps this is what the Talmud was aiming at — describing the bird saqatu and perhaps its sub-type sep(su) arik which is red, and is known to be Kosher. The name shaqi†na may also be derived from the Hebrew and Aramaic root s-q-†, which means ‘peaceful’ or ‘quiet’.4 The origin of the word murzema Persian: murg-zimc (red and long bird) The tenth-century lexicon of R. Nathan ben Jehiel of Rome (c. 1035–1106) — the Arukh — stated that murzema is a Persian 169, Munich 95 and Vatican 123 it would seem that murzema is not designated to be a mnemonic sign at all, but rather a separate issue altogether from shaqi†na. This is deduced from the fact that the Hamburg and Munich mss have a mnemonic sign appearing before the word murzema which is ‫‘ אדומין כשרין‬red are kosher’, and in Vatican 123 there is no mention of a mnemonic sign (‫ )וסימנך‬at all. Regarding the word murzema itself in the Vilna edition it appears as ‫מורזמא‬, in the Vatican 123 ms. it appears as ‫( מזרזמא‬mezarzema, however, there are additional changes in this ms.), in Vatican 122 it appears as ‫( מורמזא‬murmeza), and in the Munich ms. it is '‫( מורזימ‬murzim[a]). 2 CAD = The Assyrian Dictionary of the Oriental Institute of the University of Chicago (John A. Brinkman et al. eds), Chicago Il., 1984, Vol. 15 (S), 168. 3 Niek Veldhuis, Religion, Literature, and Scholarship: The Sumerian composition Nanse and the birds, with a catalogue of Sumerian bird names (Leiden 2004), 250. 4 Michael Sokoloff, A Dictionary of Jewish Babylonian Aramaic of the Talmudic and Geonic Periods (Ramat Gan 2002), 1173, s.v. ‘‫’שקט‬. 162

TALMUDIC ARAMAIC FAUNA NAMES: MURZEMA AND SHAQI™NA

word, since ‘anything long and red is called murzema’. Alexander Kohut (1842–94), in his Arukh Completum, said that murzema or murzima, with the above meaning, is not attested in Persian, and suggests that it is composed of two Persian words: murg (⁄d) — a bird, and zimc (ê“) — red and large, and ‘that is what a long red bird is called’. 5 Dov Geiger (1856–1943) rejected Kohut’s conjecture, stating that there is no basis for this etymology, and deemed it far-fetched. 6 Akkadian: amursanu (pigeon) Several sources mention a Sumerian bird named amar-sagmusen, in Akkadian amursanu, which according to CAD is possibly ‘a type of pigeon’.7 The word amar-sag, according to Veldhuis, is used for ducks and once for a raven, and ‘by the old Babylonian period, the word has come to refer to a kind of dove (Akkadian amursanu), although in the raven incantations… amar-sag is still used in the sense of ‘young'. In Old Babylonian Sumerian, the word for chick or fledgling is amar’.8 This bird-name, amursanu, is similar to the Talmudic bird-name murzema. Amharic derivative: r-z-m (long) murzema could possibly mean ‘long’, due to the bird’s long legs, as is the meaning of the Amharic word räzzim ARTICLE took place in each of the CS languages, and regards this as one of the simplest examples of grammaticalization. An appendix to the chapter briefly discusses definite articles that have developed in some modern Se mitic languages. The second development to be explored is that of direct object markers, or notae accusativi. These can be divided into two groups, one of which uses a particle related to Hebrew ’et / ’et and the other an object marker identical to the dative/allative preposition ‘to, for’, such as the Aramaic l (the ‘prepositional accusative’). The situ ation in Aramaic is discussed in particular detail, since both object markers are found across the various dialects. Rubin traces out a three stage development: Old Aramaic inherited a non obligatory object marker ’yt from Proto Northwest Semitic. Then when an eastern variety of Aramaic began to be spread across the Near East, especially under the Persian empire, ’yt was lost and l began to appear as the direct object marker. Finally, when there was a noticeable split between Western and East ern dialects, the former retained a reduced version of the older form, i.e., yat, while in the latter the construction with l became widespread. Rubin observes that in the earliest Aramaic dialects to use the prepositional accusative its use was restricted to animate (specifically human) objects, and only later extended to definite inanimate objects. He offers three possible explanations for why the preposition used to indi cate the indirect object came to be used to mark the direct object. The first has to do with respect. The direct object was replaced by an indirect one when the object was 253

REVIEWS

something to be respected, i.e. animate beings. The second explanation has to do with the desire for explicitness. With free word order and the lack of case marking in Aramaic there was the need to avoid ambiguity about subject and object, which could easily arise, particularly when both were animate beings. The third explana tion is that indirect objects are more often humans. An indirect object is often con nected with verbs of telling, giving, asking, showing or bringing, and therefore likely to be human. Thus whether for reasons of respect, disambiguity or due to a general linguistic tendency, the indirect object marker came to mark direct objects, origi nally only animate objects, then later all definite nouns as direct objects. Various explanations for the origin of the Hebrew object marker ’et and Aramaic ’yt / yat are discussed but found unconvincing, and any possible connection with the preposi tion 'et is ruled out. While Rubin hypothesizes that Hebrew ’et and Aramaic yat arose via grammaticalization, he admits that their origins are shrouded in obscurity. He points out, however, that these forms are further grammaticalized in some unex pected ways. Thus Syriac has essentially lost yat in its function as a direct object marker, but it does have a form yat , used with pronominal suffixes, as a reflexive pronoun. In Mishnaic Hebrew and also in Christian Palestinian, Samaritan and (very rarely) Jewish Palestinian Aramaic, the accusative particle ’et / yat has acquired a new demonstrative function. The new demonstrative can have a general meaning of ‘that’, or a more emphatic meaning of ‘that very, this very, the same’. In Modern Hebrew ’et with suffixes is still found with the meaning of ‘that very’ or ‘the same’, though it is also the normal way of expressing a pronominal object. Finally, the book looks at present tense markers in modern dialects of Aramaic and Arabic. These are classified into two main types. In the first type, all but one of the present tense markers derives from an auxiliary verb originally meaning ‘be’, ‘sit’ or ‘stand’, which can be used to describe a locative position, translatable as ‘to be present’. The one exception, Aramaic b , derives from a locative preposition, and thus fits well within this category. In the second type, the present tense marker (Arabic b ) derives from a subordinating particle marking a temporal clause, the particle itself ultimately deriving from a preposition. The finite verbal form follow ing the Arabic b suggests that we are dealing with an original subordinate clause which has taken on the function of a main verb. This book is clearly intended more for Semitists than linguists, but should also be valued by linguists for drawing their attention to work done in the comparative Semitic field. Quotations from all the languages are presented in transliteration. Parallel developments in modern (non Semitic) languages are sometimes cited for the purpose of comparison. doi:10.1093/jss/fgn053

DAVID M. STEC UNIVERSITY OF SHEFFIELD

JEAN JACQUES GLASSNER. Translated and edited by ZAINAB BAHRANI and MARC VAN DE MIEROOP, The Invention of Cuneiform: Writing in Sumer. The Johns Hopkins University Press, Baltimore 2003. Pp. xvii + 266. Price: £31.00 hard back. ISBN: 0 8018 7389 4. Glassner’s book is dedicated to a fascinating topic of interest far beyond the com munity of Near Eastern specialists. The origins and impact of writing have attracted the attention of scholars in a great variety of fields in recent years. Writing in Sumer has aroused particular interest, since this is traditionally held to be the earliest in vention of writing. Here Glassner reassesses the Near Eastern material in the light of research conducted in other fields. 254

REVIEWS

Before launching into the main thrust of the book, Glassner rightly uses the first chapter to detail native traditions about the invention of their writing system. And here he proposes new interpretations of some of those traditions. Glassner brings new ideas elsewhere too, some of them provocative (such as the hypothesis on p. 199 that writing in Mesopotamia might have been invented for the purposes of divination). Occasionally, greater support would have been desirable for certain claims, as for example when the author states: (p. 224) ‘Especially with the invention of the Greek alphabet, writing lost its autonomy and placed itself purely at the service of the spoken language’; (p. 225) ‘Next to prayers, offerings, and sacrifices, men address letters or written requests to the gods, which the gods are more likely to respond to than to oral requests’; or (p. 204) ‘to them [scribes and scholars], all phenomena became first of all a graphic sign’, which signs were ‘perceived as the source from which power emanated' and to which ‘the entirety of reality was ascribed’. Chapter 10 sees the author expound on the power of writing during the third millennium BCE. In a lengthy section dedicated to the significance of writing the divine determinative in front of certain royal names, the author hypothesises that (p. 200) ‘The use of the determinative tends to bring the human ruler near to the divine sphere, while the title digir/ilum [“god”] draws him away from it, as no divinity ever bore it.’ Thus while rulers are described as ‘god of his city’ (pp. 200 1), ‘A god is always the “king”, lugal/sarrum, of his city or state’. Furthermore (p. 201): ‘As that term [“god”] was never used for gods themselves, it removed Naram Sin from the divine sphere and inserted him into the vocabulary of human institutions’. The reviewer would see the titulature more as a unifying than as a dis tinguishing feature. Rulers are, of course, most commonly referred to as kings rather than gods of their city; meanwhile gods may be particular to a ruler (‘his/my god’), bear an attribute (‘angry goddess’, ‘goddess with patient mercy’, ‘goddess whose station shines from pure heaven’) or be associated with a place. Note ensí ke4 digir uru na ke4 rá zu im ma bé ‘the ruler prayed to the gods of his city’ (or ‘the ruler [Gudea, whose name is written without a divine determinative], the god of his city, prayed’; Gudea cylinder B i 15); ki tus maÌ zu sè digir ki lagaski a im si gam e dè es ‘the gods of the land of Lagash bow down before your august residence’ (Ishme Dagan B 26); or á ba uruki ba digir bé ne ki bi sè ba an gam e es ‘On ac count of its might, the gods of those cities bow down towards it’ (Lugal e 40). More generally note that the epithet digir kalam ma ‘god(s) of the Land’ is not un common and may be applied either to a deity (such as Haia or Ninshubur) or to a king (such as Shu Sin or Ibbi Sin). The reviewer agrees with the translators’ concluding sentiments (p. xvii). For while this book will draw criticism (see most recently and most stridently Englund’s review in JAOS 125 [2005] pp. 113 16), it challenges readers to con template their views on writing, and the basis on which those views are held. doi:10.1093/jss/fgn054

JON TAYLOR THE BRITISH MUSEUM

J.A. HALLORAN, Sumerian Lexicon: A Dictionary Guide to the Ancient Sumerian Language. Logogram Publishing, Los Angeles 2006. Pp. xiv + 318. Price: $70.00 hardback. ISBN: 0 978642 91 0. It is a familiar fact to teachers and learners of Sumerian that one of its charms is the absence of a usable dictionary. The only modern dictionary, The Sumerian Diction 255

REVIEWS

ary of the University Museum of the University of Pennsylvania (PSD for short) achieved B, then A in three volumes, and was then overtaken by the electronic age and transmogrified itself into a website. There are of course good reasons for this unsatisfactory state of affairs. Principal among these are the difficulty of eliciting from a given cuneiform sign an idea of its pronunciation which can be rendered in a romanized transcription. Such transcrip tions will not of course be accurate phonetic realizations, but should have a demon strable validity in terms of the system of transcription agreed by Assyriologists, which has been constructed through the mediation of the Akkadian syllabary and uses the Latin alphabet to represent the Sumerian phonemes, which our transcrip tions attempt to identify and approximately render. Sometimes, especially in earlier texts of the 3rd millennium, we simply have no idea of the ‘pronunciation’ of a given sign in a given context, or even in any context, and then we are reduced to referring the sign by its sequential number in a modern list (e.g. LAK 321, meaning A. Deimel, Liste der archaischen Keilschriftzeichen, No. 321). When we add to this the discrepancies between different modern transcriptions of the same word, either as a result of changing opinions through time, or simply disagreement among scholars, either about proper transcription procedures or about the substantive ‘pro nunciation’, it is clear that the identity of a Sumerian ‘word’ can be rather elusive, and that it is difficult to construct a dictionary along the lines of the romanized alphabet. This was pointed out to me many years ago by Edmond Sollberger, who himself produced a very innovative glossary in his TCS 1, when he said that the great advantage of Deimel’s Sumerisches Lexikon was its arrangement by the cuneiform signs, since they do not change with the whims of modern Sumerolo gists, unlike romanized versions which are subject to frequent change in response to genuine advances in our knowledge of the language, and fashions in the principles of transcription. Reasons there may therefore be for the lack of a dictionary, but it remains unsat isfactory. Until the PSD went on line, apart from glossaries to groups of texts, such as Römer's Sumerische Königshymnen der Isin Zeit, or Sollberger’s TCS 1, the stand ard short cut for finding the meaning of a transcribed Sumerian word was to look in one of the Akkadian sign lists (primarily R. Labat [and F. Malbran Labat], Manuel d’Épigraphie akkadienne, or R. Borger, Assyrisch Babylonische Zeichenliste). These show Sumerian words, but only when the cuneiform signs were used as logo grams ‘Sumerograms’ to write Akkadian words. Sumerian words were also fundamental to the rich lexical tradition of indigenous Mesopotamian scholarship, and much of our knowledge of the meaning and usage of the Sumerian vocabulary comes from lexical texts and bilinguals. These equivalences do not however feature in the Akkadian sign lists, unless they were also used as logograms. In recent years to track them down the publication of B. Hübner and A. Reizammer, Inim Kiengi (1985) has been a useful aid, since it brought together readings and meanings of Sumerian words from the recent scholarly literature, including the lexical and bilin gual sections of the two major Akkadian dictionaries, and other glossaries. It did not have pretensions to be a dictionary, but was a very handy index which enabled the Sumerologist to track down the hard facts behind the current accepted realiza tion of a word. There was therefore no such thing as a modern Sumerian dictionary in book form, and clearly the volume under review seeks to fill the gap. It derives from 20 years of work by its author, John Alan Halloran, during which he put together a web site offering a ‘Sumerian lexicon’. Like Inim Kiengi, it started not so much 256

REVIEWS

from the Sumerian texts themselves, but as an index to modern Sumerological scholarship; but it uses a great many more sources, and each entry now has a nor malized form of the lemma (with appropriate cross references) and meanings, sometimes accompanied by explanatory comments. In some respects this dictionary is the equivalent of the Concise Dictionary of Akkadian, of which the reviewer is one of the editors, and it has some of the same positive and negative qualities. It is handy to use when one is not on line, and at least it is complete from A to Z. On the other hand, like the CDA, it does not cite textual sources for meanings and usages. This is not serious in the case of common words, but the necessity to save space does not allow either dictionary to indicate the level of uncertainty which might surround a rarely attested word, or give the reader any way to check the validity of a meaning. In the case of the Akkadian dic tionary, the users are advised that for such questions it is necessary to refer to one of the two major Akkadian dictionaries (the Chicago Assyrian Dictionary and von Soden’s Akkadisches Handwörterbuch, of which the CDA is a translated abbrevia tion). With Sumerian the problem is that there is no such definitive thesaurus avail able. It does of course exist for words beginning with A and B, in the printed volumes of the PSD. On line now at http://psd.museum.upenn.edu one can also look up the rest of the alphabet, and although this does not give you completed dictionary entries, it does give a lot of helpful detail, including information on the graphics (which signs are used), a version of the cuneiform sign, some citations from the Sumerian texts, and references to mentions of the word in the secondary Sumerological literature. Provided one has online access, both for the Sumerologist and for those less expert in the language, the PSD website is therefore the most convenient and most informative place to consult in the first instance. This is un questionably the best way to track down the currently accepted readings, meanings, and textual attestations of a Sumerian word. It is difficult to see why one would prefer to use the ‘Sumerian Lexicon’ under review, because even if it were more cor rect in any given instance, there is no way to check this out. Some colleagues have suggested that the Concise Dictionary of Akkadian is ‘dan gerous’ for students of the language in their early stages, because it cannot convey the level of certainty modern scholarship has achieved as to the meaning and some times even the existence of any given word. The same is true of this Sumerian Lexi con, only more so because of the nature of our knowledge of the Sumerian language and its graphic practices. In our current state of knowledge about Sumerian one needs to check back with the primary sources much more often than with Akkadian, and currently accepted opinion is liable to change much more fre quently. Our inability to do this is therefore likely to be more damaging for Sumerian than it is for Akkadian, and unfortunately reduces the value of the book for the unwary or inexperienced user. It will not mislead those who are already us ers of the existing aids to Sumerian lexicography, principally the PSD or ePSD, but there is a risk that those not familiar with all the intricacies of Sumerian and its writing system may treat its contents as more definitive than they are. Experience indicates that if there is a simple and handy dictionary to a language, students will use it and often look no further; this one is admirable in its ambition to make Sumerian accessible to a wider audience, and it will certainly be used, but for seri ous students this must always be accompanied by the injunction to check any un certainties with the (e)PSD. doi:10.1093/jss/fgn055

J.N. POSTGATE UNIVERSITY OF CAMBRIDGE

257

REVIEWS

STEFANIE U. GULDE, Der Tod als Herrscher in Ugarit und Israel (Forschungen zum Alten Testament 2 Reihe 22). Mohr Siebeck, Tübingen 2007. Pp. xiv + 283. Price: /54.00. ISBN: 978 3 16 149214 3. As indicated by the title, the subject of this dissertation is the portrayal of death as a ruler in Israel, set against an ancient Near Eastern background, chiefly the Ugaritic texts. There is also reference to the evidence, including glyptic material, from Anatolia, Egypt, Mesopotamia and Phoenicia. The book has three principal parts: a lengthy introduction, which runs to 62 pages, on the main topic, i.e. death as ruler, an analysis of all the relevant texts and lastly, a short summary. The first part includes an account of how the death of man is portrayed in the OT: death is negative, makes men equal and belongs in the context of worshipping a single god. Next comes a section on figurative (bildsprachlich) elements in general: metaphor, personification and so on, and then there is an excursus on the concept of myth in OT studies. The conclusion is that myth and metaphor are similar in function, differing only in context. Myth may be defined as a narrative metaphor. This is followed by a brief survey of studies on Death as a figure or character (Figur) in the ancient Near East, including the OT. The main part of the book is an account of how Death is represented, first in ancient Near Eastern texts, with par ticular emphasis on the Ugaritic texts, and then in OT texts. There, Death is a glut ton (Hab. 2:5), a robber (Jer. 9:20), a shepherd (Ps. 49:15) and a partner in a cov enant (Isa. 28:15, 18). A second excursus is on texts depicting the underworld as greedy, without reference to Death, showing that this concept is not alien to an cient Hebrew thought. The texts discussed are Deut. 11:6, Isa. 5:14; 9:19; Num. 16; 26:10 11; Prov. 1:12; 30:15 16; Pss 73:9, 106:17 and 124:3. The short con cluding section is followed by a bibliography and indexes. The ancient Near Eastern texts represent Death either as a creeping and furtive being, especially in Mesopotamia and Egypt, or as a powerful monster, as in Ugaritic mythology, a depiction that is the result of a long process of development. In Northwest Semitic texts, Death belongs to the present world, not to the afterlife. An unexpected conclusion is that in respect of Death as a ruler, the direction of in fluence was from Syria Palestine to Egypt and Mesopotamia and not the other way round (p. 240). In the course of the book quite a few passages are considered, chiefly texts in Ugaritic and Hebrew. In addition, where relevant, parallels are drawn between the two sets of texts. Surprisingly, there is little reference to KTU 1.161, even though it deals specifically with the death of kings (but see p. 187, n. 298). Curiously, there is no discussion of the sea monster Tnn’s epithet sly†, commonly translated ‘tyrant’ (p. 89), which would seem to belong to the main topic of the book. Two items not mentioned are P. Xella, Archeologia dell’inferno (Verona 1987) and my ‘Love and Death Once More (Song of Songs VIII 6)’, VT 47, 1997, 385 7. On the Akkadian text from Ras Shamra (discussed on pp. 105 7) see now A. Chalmers, ‘RS 25.460 and Early Hebrew Poetry’, UF 36 (2004), 1 9. The Hebrew expression äsa Ìozê in Isa. 28:15b (discussed pp. 215 16) remains difficult, but my proposed solution (Bib 59 [1978], 133) is not considered. There are a few mistakes: p. 77: ‘Cunéi forme’ for ‘Cunéiformes’; p. 90, n. 77: the references to Del Olmo Lete, Diccio nario, 106 and to a Hebrew cognate are incorrect; p. 99: ‘sdatrt’ for ‘sda†rt’; p. 157, n. 207 ‘canaanite’ for ‘Canaanite’; p. 274: ‘Urprünge’ for ‘Ursprünge’; p. 276: ‘La Terra die Baal’ for ‘La Terra di Baal’. Some new suggestions are buried in the footnotes, many of which are quite long (e.g. p. 200, n. 328). For example, it is proposed that Death’s title mdd il, ‘beloved 258

REVIEWS

of Ilu’, which he shares with Yam, the sea god, is not to be considered as euphemis tic or ironic but as conferring legitimacy of status in the cosmic order (p. 110, n. 135). Also interesting are the Hittite parallel to Ps. 49:1 13 and 17 21 (cited p. 210, n. 354) and the explanation of the many votive sea anchors found in Ugarit (p. 81, n. 39). There is much material in this survey that will provoke readers to re examine the evidence on Death as a ruler in the ancient Near East and perhaps to resolve the many unanswered questions that remain. doi:10.1093/jss/fgn056

WILFRED G.E. WATSON NORTHUMBERLAND

HALLVARD HAGELIA, The Tel Dan Inscription: A Critical Investigation of Recent Research on Its Palaeography and Philology (Studia Semitica Upsalensia 22). Uppsala Universitet, Uppsala 2006. Pp. 250. Price: $59.50 paperback. ISBN: 91 554 6613 3. By the time the three inscribed fragments were excavated from Tell Dan/Tell el Qadi and published admirably quickly by Biran and Naveh in the early nine ties of the twentieth century CE, Old Testament scholarship was in historiographic despair. Minimalists were standing opposed to maximalists, the tendency to date greater parts of the Hebrew Bible in the Persian era or even in Hellenistic times, was more and more favoured. Scholars were uttering serious doubts as to whether we could have guaranteed knowledge about the pre exilic phase of ‘Ancient Israel’. To the minimalists, the ‘house of David’ inscription seemed an unpleasant disso nant voice. To others the inscription was a welcome piece of argument for the his toricity of King David sometimes extending to the narratives on him in Samuel and 1 Kings. This momentum provoked a heated discussion on every thinkable de tail or feature of the inscription. Was the text a forgery? Do the three fragments fit together in the way the editors suggested? In what language was the text written? Does the morpheme ‫ ביתדוד‬refer to the territory of the Davidic dynasty or to the temple of a deity Dod? Much scholarly and emotional energy was invested in the debate that was not in all cases played solely by the rules of a scientific debate. A few years ago, George Athas offered a valuable contribution to the discussion with a thorough analysis of the available data.1 After a very thorough investigation of the archaeological contexts of the finds and a painstaking epigraphical and palaeographical analysis of the inscription, Athas arrived at the conclusion that the three fragments have been part of one large monumental inscription. Fragment A contains the remnants of the upper part of the inscription while fragments B1+2 should be placed below fragment A at about 20 25% from the bottom.2 This well argued position is so convincing that I had to withdraw my earlier opinion that the fragments had been part of two separate inscriptions.3 Nevertheless, Athas’s analysis undermines definitively the view that the fragments can be joined as has been pro posed by Biran and Naveh. Next to that, Athas arrives at the view that the greater inscription should be seen as a royal inscription comparable with the Mesha in scription. In this second full monograph on the Tell Dan inscription, Hallvard Hagelia re considers Athas’s views arriving at a different position. Hagelia starts his investiga tion with a display of some 20 different (re)constructions of the inscription(s) in 1 G. Athas, The Tel Dan Inscription: A Reappraisal and a New Interpretation (JSOT Sup 360), Sheffield 2003. 2 Athas, Tel Dan Inscription, 189 91; the idea had previously been suggested. 3 B. Becking, ‘The second Danite Inscription: Some Remarks’, BN 81 (1996), 21 30.

259

REVIEWS

cluding their translations. He then offers a painstaking paleographic analysis of the inscription ending in his presentation of the text. Basic to his understanding is that he accepts the join of the fragments A and B1+2 as proposed by Biran and Naveh. In a way he seems to ignore the arguments of Athas by simple repeating the phrase that ‘a physical join can be seen’. This implies that Hagelia accepts the text as re constructed by Biran and Naveh including the restoration of the names of the Isra elite kings Jehoram and Ahaziah. Hagelia further offers a helpful palaeographic comparison leading to the assumption that the script of the Tell Dan inscription can easily be compared with the script of Aramaic, Moabite and Hebrew inscrip tions from the ninth and early eighth centuries. This leads him to a date in the late ninth century for the inscription, it being a royal commemorative inscription writ ten not long after the events narrated. Hagelia’s grammatical and linguistic analysis of the inscription is very interesting. He displays in good detail the discussion about the presence or absence of narrative wyq†l forms in the inscription. He convincingly argues that the core of the text is to be seen as a narrative chain indicating that this literary procedure not only occurs in Hebrew and Moabite, but also in Old Aramaic. He then correctly refers to the absence of a waw before a preformative verb especially in the form ‫יהן‬, ‘he went’, in line 3. The presence of this form should not be seen as an argument against the presence of a narrative chain in the inscription. Hagelia opens the possibility of construing the form as a qualifying yq†l, informing the reader in an explanatory way about the way in which the narrator’s father went to his rest. His further analysis of forms and vocabulary underline the growing consensus that the language of the in scription is basically Old Aramaic with a few local dialectical features, such as the /t/ in the verb qtl, ‘to kill’, instead of the more common /†/ [q†l]. Hagelia’s remarks on the historical implications of the inscription and its connec tions with biblical evidence are characterized by shortness. He connects the con tents of the inscription with events in the interlude between the battle at Qarqar and the rebellion of Jehu. In his opening section, Hagelia more or less promised to present the debate on the Tell Dan inscription. This promise is only partly delivered. Of course, Hagelia presents various, sometimes opposed, positions about the various aspects of the dis cussion. This display would, however, have gained from a more thorough approach. In my opinion the debate about Tell Dan is not solely about how to read and inter pret an inscription. There is more at stake. At the background of the various posi tions are different historiographical views. The whole panorama from positivistic to post modern can be detected. Some contributions to the debate seem to be based on a Collingwoodian understanding of the past, others tend in the direction of Hayden White and a few are merely naïve on the question of how to (re)construct the past. Besides, the fierceness of the debate visible in the ad hominem argu ments might have been fuelled by the stand toward the epidemic of biblio phobia.4 Hagelia has not convinced me. In my view the proposal by Athas better explains the difficulties between the fragments A and B1+2. Even on the facsimile that Hagelia publishes on page 53 one can see immediately that there is no clear con tinuation of the readable lines in the join as proposed by Biran and Naveh. Next to that, I have a few minor points of criticism. 4 On this diagnostic term see H.M. Barstad, ‘The Strange Fear of the Bible: Some Reflections on the “Bibliophobia” in Recent Ancient Israelite Historiography’, in L.L. Grabbe (ed.), Leading Captivity Captive (ESHM 2; JSOT Sup., 278; Sheffield 1998), 120 7.

260

REVIEWS

• Even if one accepts the classical join, Hagelia’s reconstruction of line 1 is more than problematical: [ ] ‫ מר ע]די[ וגזר‬.[‫[ …‘ ]א‬sa]id …? tr[eaty …] and he cut [ ]'. In my view, there is too much room between [‫( ע]די‬Frag. A) and ‫( וגזר‬Frag. B1+2) to see them as parts of one sentence. Besides, the waw before ‫ גזר‬makes it impossible to read here the syntagma ‫גזר עד‬, ‘to make an agreement’, as parallel to Hebr. ‫כרת ברית‬. • Hagelia (pp. 60 2) follows the generally accepted construction of Tel Dan in scription lines 4 5: ‫[אנה‬. ‫[א]יתי‬.] ‫הדד‬. ‫… ו[יהמלך‬, ‘Hadad made [me] king, yes me!’, interpreting the clause as a legitimation formula for a king who did not re ceive the throne by inheritance. There are, however, a few problems with this view. The addition ‫א]יתי‬, supplying the object ‘me’ is far from certain. Within the parameters of the joined text the addition ‫ א]בי‬resulting in the translation ‘Hadad made [my] fa[ther] king’, would yield an intelligent reading. In my view hmlk should be construed as a Haph. ‘perfect’ 3.m.s. of the verb mlk: ‘he made king’. For syntactical reasons it is plausible to suggest that the subject of this clause preceded the verb and that hdd is the object: ‘X made Hadad … king’.5 Since the word divider after hdd is not certain, Hadad most probably should be construed as the theophoric element of a personal name belonging to the person who was made king. Suggesting that I read ‘Hadad’ as the divine name would indeed turn the idea upside down, but Hagelia is misrepresenting my view (p. 62). • In his discussion on phonemes, Hagelia (p. 125) states that in the Tell Dan in scription the phoneme /ê/ is found. This is correct for the final open syllable in ‫בה[תלחמה‬, and ‫אלפי‬, but incorrect in the form ‫ אמרי‬and ‫אלפי‬. This noun should be construed as masc. plur. det. and read as, e.g., ’asirayya(’ ), ‘the prisoners’. • Page 134: construing ‫הם‬.‫ ארצ‬as a construct chain, ‘the land of them’, might ac count for the presence of a word divider. • I would prefer to label ‫ אית‬as the Aramaic nota objectiva, and not as a nota accusativum (p. 159), since the Aramaic language does not have cases. • The book is disfigured by some typographical errors of which I only mention a few: ‘Mobite’ (p.3); ‘Ahutiv’ (p.3); ‘stratiography’ (p. 122); ‘Joün’ (p. 136). At page 68 the word dividing dots have disappeared. Despite my critical remarks, I would like to thank Hallvard Hagelia for sharing with the scholarly world his insights on the various aspects of the Tell Dan inscrip tion. Nobody interested in West Semitic epigraphy or ancient Near Eastern history can leave this book unread. doi:10.1093/jss/fgn057

BOB BECKING UNIVERSITY OF UTRECHT

ERNST JENNI, Studien zur Sprachwelt des Alten Testaments II. Kohlhammer, Stuttgart 2005. Pp. 351. Price: /45.00 hardback. ISBN: 3 17 018746 5. This is the second volume of collected essays by the well known Swiss Hebraist Ernst Jenni edited by Jürg Luchsinger, Hans Peter Mathys and Markus Saur. Eight years after the first one came out, the present collection of solid studies testifies to J.’s unabated scholarly activity and, according to the editors’ preface, should again be considered as a present from the author ‘to the friends of the Hebrew Bible and its language’. The gift is highly appreciated.

5 See

also Becking, ‘Second Danite Inscription’, 26. 261

REVIEWS

As explained in the author’s preface, the overall purpose of the studies was not to corroborate or illustrate a specific theory, but rather inductively to take the Biblical Hebrew text and its grammatical peculiarities as point of departure and from there to arrive at more general conclusions on the basis of a methodical approach. True to his usual style, J. refrains from making his point by means of some eclectic exam ples, but endeavours to be comprehensive in his taxonomy and exhaustive in his listing of the features in question, a fact that adds considerably to the value of his writings. In all, the collection comprises eleven essays from the period 1998 2004, one of which is published here for the first time. Several articles are tightly or loosely con nected with J.’s earlier work on the preposition lamed; they refine at some points the conclusions arrived at in the author’s famous monograph. Others are concerned with the use of particles, whereas quite a few delve into the realm of modality. The first study, entitled ‘Vollverb und Hilfsverb mit Infinitiv Ergänzungen im Hebräischen’ (pp. 11 35; originally published in ZAH 11 [1998], 50 67), dis is cusses cases in which a modal verb such as ‫י ָֹכל‬, ‫נ ַָתן‬, ‫ָאָבה‬, ‫ֵמֵאן‬, ‫ הוִֹסיף‬or ‫ֵהֵחל‬ used together with an infinitive either with or without the preposition l , as in Deut. 10:10 ‫אבה יהוה השחיתך‬-‫ לא‬or Deut. 23:6 ‫בלעם‬-‫אבה יהוה לשמע אל‬-‫ולא‬, a fea ture that standard grammars indiscriminately juxtapose as alternatives. J. shows that infinitives without l should be considered the direct object of the modal verb, which itself functions as a full fledged main verb, whereas constructions containing infinitives with l are to be regarded as a single complex predicate, in which the fi nite verb serves as an auxiliary. Accordingly, the construction without l is used for general utterances, while, that with the preposition l refers to a specific, concrete action or state. The modal verbs used with l are divided into verbs of partial reali zation (Teilverwirklichung), possibility and volitivity. A critical reader may be left with the question of whether indeed diachronic considerations play no role here, but that is not something that will easily come up within J.’s strictly synchronic ap proach. The second essay, ‘Epistemische Modalitäten im Proverbienbuch’ (pp. 36 47; previously published in the Fs. Hans Peter Müller, Mythos im Alten Testament und seiner Umwelt [1999], 107 11), is concerned with about 80, predominantly late, cases of infinitive construct with l that should be considered to be the predicate of a non verbal clause. The function of such infinitives is usually adding a deontic modal nuance to the predication (i.e. necessity or certain expectation), as in Eccl. 3:15 ‫‘ ואשר להיות כבר היה‬what must happen, has already happened’. In a small number of cases in Proverbs, however, we should assume epistemic modality. For instance, Prov. 16:30 ‫ עצה עיניו לחשב תהפכות‬could be translated as ‘whoever narrows the eyes is likely to contrive froward things’. To the literature quoted in this article should now be added A. Gianto, ‘Mood and modality in classical Hebrew’, Israel Oriental Studies 18 (1998), 183 98 and id., ‘Some notes on evidentiality in Biblical Hebrew’ in Biblical and Oriental Essays in Memory of William L. Moran (2005), 133 53. The third article, entitled ‘Einleitung formeller und familiärer Rede im Alten Testament durch ’mr ’l und ’mr l ’ (pp. 48 64; published in the Fs. Georg Sauer, Vielseitigkeit des Alten Testaments [1999], 17 33) suggests that, in the main, the dis tinction between ‫ אמר אל‬and ‫ אמר ל‬has to do with different levels of politeness; even though some of the nuances seem to escape the modern reader, a tendency to prefer ‫ אל‬over ‫ ל‬when ‘talking up’ is apparent. For a quite different approach, see M. Malessa, ‘Biblisch Hebräisch ‫ְל‬/‫ ִדּבֶּר אֶל‬und ‫ֵאת‬/‫ ִדֶּבּר עִם‬im Vergleich’ in the Fs. 262

REVIEWS

Muraoka, Hamlet on a Hill (2003), 333 40 and id., Untersuchungen zur verbalen Valenz im biblischen Hebräisch (Assen 2006; revision of diss. Leiden, 2003), esp. 167 205; there is doubtless room for further discussion. The fourth essay, ‘Subjektive und objektive Klassifikation im althebräischen Nominalsatz’ (pp. 65 76; previously in Theologische Zeitschrift 55 [1999], 103 11), aims to determine the functional difference between classifying non verbal clauses, such as Gen. 44:16 ‫‘ הננו עבדים לאדני‬behold, we are my lord’s servants’, and what J. calls ‘re classifying’ non verbal clauses such as Gen. 50:18 ‫‘ הננו לך לעבדים‬behold, we are as your servants’, in which the predicate is preceded by the preposition l . J. concludes that in the latter case there is no complete identity between the subject (‘we’) and the predicate class to which the subject is re classified (‘servants’). In other words, the identity is not yet real, but present in the mind of the speaker; the latter clause could therefore be paraphrased as ‘as of now, we may be considered to be or become your servants’; the same interpretation holds true for many other such cases. The fifth article, ‘Aktionsarten und Stammformen im Althebräischen: das Piël in verbesserter Sicht’ (pp. 77 106; originally ZAH 13 [2000], 67 90), in fact con stitutes an appendix to J.’s well known monograph Das hebräische Piël (1968), in which the author revises earlier conclusions and takes issue with other publications on the subject, such as Kouwenberg’s Gemination in the Akkadian verb (1997) and Joosten’s ‘The function of the Semitic D stem’, Orientalia 67 (1998), 202 30. J. abandons the division into transitive and intransitive verbs as a pivotal feature and reformulates the function of the Piël on the basis of Vendler’s (‘Verbs and times’ [1957]) four lexical or situation aspects (Aktionsarten): state, activity, accom plishment and achievement. With verbs expressing complex actions, the Piël leaves the process (Hergang) of the action implicit (unausgedrückt): with stative verbs it serves as a factitive, with accomplishment verbs a resultative, with verbs expressing complex movements or sound utterances it denotes the successive. Achievement verbs and transitive activity verbs have no Piël. With verbs of complex activity, a repetitive action in the Piël is rendered as a punctual achievement. The sixth contribution, ‘Textinterne Epexegese im Alten Testament’ (pp. 107 17; previously in Fs. Odil Hannes Steck, Schriftauslegung in der Schrift [2000], 23 32) discusses the use of lamed with infinitive construct to modify a preceding utter ance, such as ‫לאמר‬: ‘by saying, …’ or Gen. 2:3 ‫ברא אלהים לעשות‬-‫מלאכתו אשר‬-‫מכל‬ ‘from all the work which He had created by making (it)’ (i.e. the Latin gerund faciendo), with special attention to the logical semantic relation between the main verb and the infinitive. In roughly one fifth out of 300 cases of such infinitives with lamed we have to do with ‘epexegesis’; the infinitive explains the main verb by specifying it or adding precision. In the seventh essay, ‘Eine hebräische Abtönungspartikel: äl ken’ (pp. 118 33; originally in Fs. Klaus Seybold, Prophetie und Psalmen [2001], 201 15), the author scrutinizes the semantic range of ‫כן‬-‫על‬. Whereas the expression usually (144 out of 155 times) introduces a logical consequence and hence may be rendered as ‘there fore’, the constructions ‫כן‬-‫( כי על‬10×) and ‫כן‬-‫( אשר על‬1×) rather convey an under lying reason (German: da nun einmal; Latin: quia enim). J. deftly places the two functions within a single theoretical framework (with an interesting detour to Swiss German dialectology!). Essay number eight, ‘Semantische Gesichtspunkte des Hebräischen und deutscher Übersetzungen am Beispiel von Num 10, 29 31’ (pp. 134 50; previously pub lished in W. Gross [ed.], Bibelübersetzung heute. Geschichtliche Entwicklungen und 263

REVIEWS

aktuelle Herausforderungen. Stuttgarter Symposion 2000 [2001], 209 33) reiterates and expands on some points also discussed in other studies included in the volume under review (viz. ##3, 1, 4 and 7 respectively). As the title clearly reveals, the ninth article, ‘Presidential Address: Höfliche Bitte im Alten Testament’ (pp. 151 65; A. Lemaire [ed.], Congress Volume Basel 2001 [VTSup 92; Leiden 2002], 1 16), is the written version of the presidential address at the 2001 IOSOT conference in Basel and discusses the ways in which a polite re quest is made in Biblical Hebrew, within the linguistic framework of expressing deference. Languages generally have ways of mitigating the potentially harsh tone of orders or requests: intonation, modal particles, tags (‘…, will you?’). In Biblical Hebrew this may be achieved by a) the normal (‘short’) imperative; b) the long im perative ending in â (the ‘adhortative’); c) the addition of the particle nâ’; d) a combination of b) and c). Through an examination of all imperative forms in Gen esis 2 Kings, as well as a discussion with recent studies mainly by Kaufman (‘An emphatic plea for please’, Maarav 7/1 [1991], 195 8) and Fassberg (Sugiyot be taÌbir ha Miqra [1994], 13 35), J. arrives at the following hypothesis: the short imperative (e.g. ‫ )ַהגּ ֵד‬is generally used when the fulfilment of the order is obvious; the long imperative ([‫ )ַהגּ ִיָדה ]ִלּי‬expresses a polite request; the particle nâ’ (‫נ ָא‬-‫)ַהגּ ֶד‬ marks a vivid, insistent request, possibly against the will or the expectation of the hearer; the long imperative combined with nâ’ (‫נּ ָא‬-‫ )ַהגּ ִיָדה‬is accordingly used for polite requests that are likely to meet resistance from the side of the hearer. Essay number ten, ‘Untersuchungen zum hebräischen Kohortativ’ (pp. 166 226; originally ZAH 15/16 [2002/3], 19 67), examines the use of the so called ְ ‫א‬ ֶ cohortative, i.e. first person ‘imperfect’ verbal forms ending in â, such as ‫ְקטָלה‬ and ‫ְקטָלה‬ ְ ִ ‫נ‬. J. concludes that these do not merely express volitivity (‘I/we want …’); such nuances, after all, are already denoted by normal imperfect forms. The cohortative is used in order to express a need for (or expectation of ) assent or ap proval from the hearer (or, in the case of self adhortation, from the speaker him self ). Accordingly, in cases where such approval is not necessary or already secured, the cohortative is not used. For plural forms a distinction is made between an ‘in clusive plural’ use (i.e. ‘I/we and you’), an ‘exclusive plural’ (‘we, but not you’) and an ‘identical plural’ or ‘reflexive’ one (‘all of us, we ourselves’). The study contains a useful exhaustive list of occurrences. The eleventh, final and by far the longest contribution to the volume under re view is a hitherto unpublished study entitled ‘Verwendungen des Imperativs im Biblisch Hebräischen’ (pp. 227 315). It resumes the subject of essay #9 (partly also that of #10) and purports to be an exhaustive study of the use of imperative forms in Biblical Hebrew. The result is a penetrating piece of research, a reference work of which the conclusions are obviously difficult to summarize. The article comes with a comprehensive and highly detailed database (23 pp.) containing each and every imperative form, classified according to all distinctions made by the author: gram matical form (short or long imperative, nâ’ etc.), the role of the speaker and hearer; register and style (rhetorical order, exhortation, …), syntactic context (e.g. ‫)קוּם לְֵך‬, performative verbs and the nature of the speech act (decree, proposal, adhortation, counsel, …). At the end of the volume under review, there is a list of references to the original publications, a joint bibliography for all the articles in the volume and an index of biblical passages. There can be little doubt that Jenni’s meticulous and penetrating studies on Hebrew syntax and semantics are important contributions to the field. If the collec 264

REVIEWS

tion of essays by this learned octogenarian constitutes a gift to the academic com munity, it is a gift that should be gratefully acknowledged. doi:10.1093/jss/fgn058

MARTIN F.J. BAASTEN UNIVERSITEIT LEIDEN

ROBERT D. MILLER II S.F.O., Chieftains of the Highland Clans: a History of Israel in the 12th and 11th Centuries B.C. (The Bible in its World). Wm. B. Eerdmans, Grand Rapids 2005. Pp. xix + 186. Price: $28.00 paperback. ISBN: 0 8028 0988 X. ANN E. KILLEBREW, Biblical Peoples and Ethnicity: an Archaeological Study of Egyp tians, Canaanites, Philistines, and Early Israel 1300 1000 B.C.E. (Society of Bibli cal Literature Archaeology and Biblical Studies 9). Society of Biblical Literature, Atlanta 2005. Pp. xx + 362. Price: $39.95 paperback. ISBN: 1 58983 097 0. Robert Miller’s book, Chieftains of the Highland Clans: a History of Israel in the 12th and 11th Centuries BC, proposes a model for the social and territorial organization of the Highlands of Israel and the West Bank in the Early Iron Age. It starts with a demarcation of the period, area and population that is the subject of research. Israel is defined as an ethnic group by the homogeneity of its material culture, which continues into the IA II period. The second chapter is a short discussion of the function and limitations of model building in the writing of history, and a defence of the use of anthropologi cal models from unrelated cultures. Chapters 3 and 4 introduce the ‘Complex Chiefdom Model’ adopted by Miller, and the methodology to test it. The definition is largely taken from two studies: T. Earle (ed.), Chiefdoms (Cambridge 1991), and C. Peebles and S. Kus, ‘Some ar chaeological correlates of Ranked societies’, American Antiquity (1977), 47:421 48, adaptations of Elman Service’s classic model. Characteristics of the complex chiefdom, according to these studies, are a multilevel hierarchy and ‘mobilization’ of subsistence goods: the various production goods are not redistributed by the chief but used to control his people by means of trading or converting them into services to supporters. ‘Cycling’ is another characteristic of the complex chiefdom: the periodic breakdown of the system, which occurs when the limits of productivity have been reached and the chief fails to satisfy the demands of his supporters. A rebellion follows and the system breaks down into several smaller subsystems, each headed by one of the sub chiefs. Eventually one of these will rise to the position of chief, incorporating the second level subchiefs, and a new cycle starts. The methodology used to test the model includes Gravity Modelling, Thiessen polygons, and size statistics, to detect and classify clusters. These clusters are then compared to the archaeological record to define the nature of the interaction. The next chapter, by far the longest in the book, is the actual ‘testing’ of the model, applying the methodology to the primary data, and arriving at a hypothesis for the social organization of early Israel. This consists of clusters or systems, some of which are defined as complex chiefdoms, others as ‘simple’ chiefdoms. The clus ters are defined by applying the Gravity Model compared to Thiessen polygons to define the borders and to Central Place Models to define the number of hierarchi cal levels. A large part of the chapter is devoted to the description of import goods in the various sites in order to define trade and the flow of sumptuary goods; to a description of the ecology of the region; of the temporal changes in the settlement pattern; mortuary evidence, and ‘other archaeological correlates’ such as architec ture. 265

REVIEWS

Chapter 6 investigates written sources, mainly the Merneptah stela; and chapter 7 defines the hypothesis, in the form of a ‘narrative’. The last chapter (apart from the very short conclusions at the end) looks at the biblical narrative and compares it with the hypothesis. This book is a laudable effort to use archaeological and anthropological models in the study of a highly charged and dark period, applying a strict methodology. Unfortunately, the result is the opposite of what the author intended: the method ology is too fraught with difficulties and uncertainty in the basic data to lead to a convincing hypothesis. Apart from that, the method as applied by the author, is also flawed. My main problem with Miller’s application of the complex chiefdom model is that instead of using it as a hypothesis, which is subsequently tested for its validity, it is used as the basis for a hypothetical ‘narrative’ that defines the boundaries of the various sys tems. In other words, the question is not whether Early Israel consisted of complex chiefdoms but rather where the boundaries were. The Gravity Model, used by Miller for his primary analysis, is a way of quantify ing the idea that interaction between sites (villages, towns, camps) is negatively cor related to the distance (measured as walking time) between them, and positively to the product of their size. This model has been tested by archaeologists and anthro pologists, mainly in the New World, and a summary of results can be found in Plog (S. Plog, ‘Measurement of prehistoric interaction between communities’, in K. Flannery [ed.], The Early Mesoamerican Village [New York 1976], 255 71) or Hodder and Orton (I. Hodder and C. Orton, Spatial Analysis in Archaeology [Cam bridge 1976]). These summaries show that the model can quantify interaction, and predict artefact distributions. There are difficulties however. Miller sums them up for the Highlands: the size of multiperiod sites is far from certain, surveys have of ten missed smaller sites and there are serious dating issues. Even so, whilst the Gravity Model is useful to explain distribution patterns of artefacts, Hodder and Orton (Spatial analysis, 195 6) state explicitly that it cannot be used to analyse so cial (or ethnic) patterning or territoriality, as artefact distribution patterns do not necessarily follow ethnic boundaries. Miller uses the Gravity Model to create maps of interaction clusters, often com plex systems with three or more tiers. These systems, however, are purely theoretical constructs, created by applying a mathematical formula to the assumed size and relative location of every site and then drawing lines between those with the highest interaction factor. This ‘interaction factor’ varies and can be high, or relatively low (as long as it is the highest outcome for that specific site), something that is not expressed in Miller’s maps, where every line drawn between sites has the same weight. His application of the model also implies that every site necessarily interacts with at least one other, introducing a factor of artificiality. Miller then skips a few steps, and defines the centres and subcentres of these systems as centres of administrative control (p. 30). So he jumps from a geographical model to a social one, without any explanation. Thiessen polygons are then drawn, as corroborating evidence for the distribution of the clusters. As Thiessen polygons are based on the physical distance between sites, and therefore a simplified version of the Gravity Model, they necessarily con firm the boundaries of the various systems, and their corroborative value is zero. A supportive method quantifying bimodal distribution as an indication of a complex chiefdom, namely the plotting of the site size against the number of sites, is ren 266

REVIEWS

dered unreliable by the small number of sites involved in each system (56 for the largest system of Balatah, if I counted right, and much smaller for the others) mak ing the use of statistics meaningless. In other words, the application of the Gravity Model, its interpretation, and its ‘support’ from other models, are all highly flawed. There is no evidence here for complex chiefdoms. Miller’s extensive (and very useful) enumeration of import artefacts shows evi dence for outside contacts, possibly for markets and various forms of distribution, but there are too many anomalies in this record to use as an argument for complex chiefdoms. His decision to include the archaeological data from Benjamin, as they are better documented than those from the area of study, is not a happy one, as it creates an extra assumption: that the population in Benjamin came from the same back ground as that in the study area and was organized in the same way. An extensive description of the ecology of the region shows that every ‘bailiwick’ defined by Miller extended over a variety of ecological environments, suggesting a mixed economy. It outlines the interaction between the various ecological regions for the exchange of subsistence goods and is useful regardless of whether the ‘baili wicks’ actually existed or not. The other areas of evidence, mortuary data and architectural data, are too scanty to be of use. Miller’s postulate that monumental architecture necessitates conscriptive labour, and therefore the existence of chiefdoms, seems outdated, as it becomes clearer that public monumental projects (such as town walls, or temples) could well be accomplished by corporate groups, such as tribes (see G. Philip, ‘The Early Bronze I III Ages’, in B. MacDonald, R. Adams and P. Bienkowski [eds], The Archaeology of Jordan [Sheffield 2001], 177 8; or E.J. van der Steen, Tribes and Territories in Transition [Leuven 2004], 111 12 for recent examples), whereas conscriptive labour can well indicate a feudal system. The remainder of the book explores the written sources, particularly the mention of ‘Israel’ in the Merneptah Stela and the biblical narrative. I agree with Miller that the main conclusion from the Stela is the fact that Israel must have been an impor tant factor, to be mentioned at all. However, the ‘people’ determinative seems to me to have more significance than Miller grants it, and unfortunately, one of the things the Stela does not provide is the geographic location of ‘Israel’. The biblical narrative is laid beside Miller's own ‘narrative’ (chapters 7 and 8). Miller is extremely careful to make clear that he does not take the biblical narrative at face value, but that there may still be truth in it. Nevertheless Miller has missed a chance here: the classic studies of Lord and Parry (A. Lord and A. Parry, The singer of tales [Cambridge, Mass 1960]), or, closer to home, the study of Bedouin oral tra ditions (P. Heath, The Thirsty Sword. Sirat Antar and the Arabic popular epic [Salt Lake City 1996], B. Connelly, Arab folk epic and identity [Berkeley 1986]), or A. Shryock’s, Nationalism and the genealogical imagination. Oral history and textual au thority in Tribal Jordan [Berkeley 1997], a study of the oral and written traditions of the Adwan and the Abbadi in central Transjordan), would have been enlightening. The complex chiefdom ‘narrative’ is then used to explain some of the biblical narra tive. It makes for an interesting story but is hardly proof of the complex chiefdom society that Miller suggests here. One of the main problems with most of these studies, that use anthropological models in the explanation of Early Israel or the archaeology and history of the re gion in general, is that they use models that apply to a number of ‘primitive’ cul tures, often from the New World, Africa and Oceania. This is where these models 267

REVIEWS

have been developed and tested, and there is nothing comparable for the southern Levant. Nevertheless, this approach is fraught with problems: not only are the cultures themselves unrelated to those under study but so are the climates, ecosystems and geology. This renders applications of the models, such as Miller does, theoretical and tenuous. One area of study that may remedy that shortcoming is that of Levantine eighteenth early twentieth century societies described in travel accounts and by Arab historians. They do not provide ready made models, but they describe a society that has more in common with that of the Iron Age, or any archaeological period, than those of, for example, Tahiti. There are some minor irritations in the book. The bibliography is very incom plete: many publications mentioned in the text are missing or misquoted in the bibliography; there is the consistent misspelling of names, such as Deir Allah (Deir Älla); or Theissen polygons (Thiessen polygons). It does contain useful informa tion, especially in the overviews of material culture and ecology. However, as a model, it remains unconvincing. Ann Killebrew’s book is a study of the various peoples that populate the Hebrew Bible, their history, culture and social organization, as can be construed from the archaeology and the written sources. The book deals with Egyptians, Canaanites, Philistines and Israelites. It starts with an introduction into the thirteenth century breakdown of the Egyptian empire and the consequent collapse of the Late Bronze Age society. The frame of reference used is World Systems Theory, and the various cores and periph eries (and semiperipheries) are outlined. Canaan is defined as a periphery, largely of Egypt. The ‘crisis’ is viewed in terms of systems collapse, although Killebrew prefers to see the end of the Late Bronze Age as a ‘transformation’ rather than a ‘crisis’. I tend to agree with her. The ‘crisis’ was rather a restructuring of elements of the sub systems into a new, less integrated formation which was the Early Iron age. The next chapter looks at Egypt in Canaan in the thirteenth century and asks if this was a case of imperialism or colonialism. An extensive survey of the evidence includes written Egyptian sources both in Egypt and in Canaan, architecture, pot tery and other artefacts and burial customs. Egyptian presence in Canaan fluctu ated and was strongest during the reign of Ramesses III. The purpose of Egypt’s presence in Canaan was to collect a maximum tribute and the military presence at sites such as Beth Shean, Deir el Balah and others, was to maintain order and se cure the loyalty and cooperation of the local rulers. Chapter 3 looks at the Canaanites, a culturally not very well defined group. Killebrew calls them a geopolitical entity, a ‘multiethnic’ group that shared a lan guage, a common culture, and formed a ‘social boundary’. They are the ancestors of various Iron Age sociopolitical groups, such as the Phoenicians, Ammonites, Moabites and Israelites. Consequently, Killebrew focuses on the history and archae ology of the Canaanites during the Middle and Late Bronze Ages, as she sees here the roots of the later Iron Age polities. An extensive survey of the architecture, pal aces, temples and domestic structures, as well as burial customs, pottery and other artefacts underline the wide cultural variety. One of the weaknesses of this chapter is that it focuses entirely on the region west of the Jordan, excluding a large part of the Canaanite territories, namely Transjordan. Consequently, some conclusions about Canaanite culture are incorrect, such as the remark that there were no Canaanite walled towns. There were several towns in Transjordan with Late Bronze town walls, such as Sahab and Abu Kharaz, and there was a Late Bronze market 268

REVIEWS

centre on the Amman plateau. The region east of the Jordan formed a political and economic entity with the western region in the Late Bronze Age and should not be excluded. Chapter 4 deals with Israel, the ‘Mixed Multitude’. It has a large section on re gional surveys, on the basis of which Killebrew concludes that the collapse of Egyp tian control must have created a power vacuum resulting in the settling of nomadic populations in the region. The Transjordanian Plateau is dealt with in this section but, unfortunately, the Jordan Valley is not. This is an important omission, as the Jordan Valley was not only the link between the two highlands but also a focus of political and ethnic activity both in the Late Bronze and Early Iron Ages (van der Steen, Tribes and Territories, 301 8). One of the problems with interpreting the material culture of the Early Iron Age in the Highlands, as pointed out by Killebrew, is that most interpretations and analyses are governed by a need to ‘discover’ Early Israel in it. This chapter is there fore as much a survey of interpretations as of material remains. It looks extensively at village planning, the pillared building, silos and terraces; traditional ‘markers’ of Early Israel. After a brief survey of existing hypotheses Killebrew adds her own, the ‘Mixed Multitude’ theory. This defines Israel as a heterogeneous collection of largely Canaanite tribal and kin based groups, comprising nomads, farmers, Apiru and Shasu, and gradually merged into an ethnic entity embracing an epic narrative, recounted in the Bible, and a shared religion, that of Yahweh. Chapter 5 deals with the Philistines. Killebrew looks at the phenomena of cul tural diffusion versus colonization, and concludes that the Philistines were colo nists, originating probably in Cyprus but with roots in the eastern Aegean and per haps Anatolia. The material culture is analysed using a ‘pilot site’, Tel Miqne / Ekron, which has been excavated and published extensively, and which has the longest uninterrupted occupation history of any Philistine site. The synthesis looks at the origins of the Philistines, and at the chronology here the ‘low chronology debate’ is touched upon briefly but no conclusions are drawn although Killebrew seems to favour the ‘middle chronology’. The book concludes with a summary of the reviewed evidence and an overall reiteration of the conclusions of the various chapters. The ‘Mixed Multitude’ model has much to recommend it. It incorporates the different strands of material culture of Early Israel as well as the widely varying tra ditions that lay at the root of the biblical narrative and it has the flexibility to adapt to new evidence, or new interpretations of evidence, without becoming meaning less. Whether the Yahweh cult was the core ideology of Early Israel seems to me doubtful, as the Yahweh cult may have been the religion of one specific group from Edom (Shasu?), and there is evidence for other cults as well. This in fact illustrates my point that the model is versatile. Interestingly enough, it is Miller’s section on imported goods in the highland sites (pp. 45 51) rather than Killebrew’s own section on the archaeology of the highlands that sums up material evidence for her theory: localized Transjordanian influences in Cisjordanian sites, the presence of Egyptian objects, the bull on the Bull site, copper objects from the Wadi Arabah, and even the locally produced Phil istine pottery from Tell en Nasbeh. These may all be evidence for trade, as Miller assumes, or they may have been brought and produced by small groups from a heterogeneous background, Killebrew’s ‘mixed multitude’. The question that is not answered satisfactorily is ‘why?’. Why did this ‘mixed multitude’ of social and ethnic groups, migrants, runaway slaves, peasants and 269

REVIEWS

herders, merge into new ethnic or socio ethnic entity? Was it, as suggested in the Bible, the presence of a common enemy such as the Philistines that made them stick together? Was it, indeed, a sweeping religious mass conversion comparable to that of early Islam? Or was it simply a question of power and conquest, one group subjecting all the others? And what about the other polities in the region, such as the Ammonites and the Moabites? Did their development into unified polities fol low the same paths? These questions still need answers and, despite the vast amounts of literature on the subject (as provided in Killebrew’s bibliography), we are far from having those answers. The question of ethnicity in this context is an important one. Accepting Killebrew’s definition of ethnicity as group identification, the question remains: why do people identify with a certain group and how and why does that self iden tification change over time? This has been an important area of study in recent an thropology. The work of Anthony Smith (A.D. Smith, National Identity [London 1991], The Nation in History: historiographical debates about ethnicity and national ism [Cambridge 2000]) has been particularly useful in this context and may pro vide some of the answers. Both books, Miller’s study of the complex chiefdoms and Killebrew’s mixed mul titude, present a model for the Early Iron Age in the highlands. Both use the same sources but their conclusions are radically different. In my opinion, Killebrew’s model is the more convincing, largely because its point of departure is the archaeological record (including the written sources), whereas Miller’s point of departure is the model itself to which the data are sub jected. Because Miller does not look at the history or the origins of his ‘complex chiefdoms’, they remain an isolated phenomenon with no roots in reality. Killebrew’s mixed multitudes have names and histories in the Late Bronze Age as well as a history of their own; therefore, they are more real, and as such, more con vincing. doi:10.1093/jss/fgn059

EVELINE J. VAN DER STEEN UNIVERSITY OF LIVERPOOL

BERNARD S. JACKSON, Wisdom Laws: A Study of the Mishpatim of Exodus 21:1 22:16. Oxford University Press, Oxford/New York 2006. Pp. xv + 552. Price: £98.00 hardback. ISBN: 978 0 19 826931 1. Bernard S. Jackson’s Wisdom Laws: A Study of the Mishpatim of Exodus 21:1 22:16 is the author’s fullest explication of his view, already described elsewhere, that the so called Mishpatim of the Covenant Code represent the collection and expansion of originally oral, customary rules that are expressive of wisdom values. As such, the thesis of Wisdom Laws is primarily concerned with the earliest Sitz im Leben of the Mishpatim; however, Jackson also reconstructs a complex redactional history for the final form of the Book of the Covenant and describes the differing interests exhib ited in these hypothesized intermediate stages. All of these points are argued through a detailed exegesis of the Mishpatim that is aided by comparison with judi cial situations in biblical narrative, comparable laws in the other biblical legal cor pora, the ancient Near Eastern legal traditions and biblical wisdom literature. Part I contains a helpful Forschungsbericht for the Covenant Code that not only reviews but also categorizes the different scholarly approaches to biblical law and especially the Book of the Covenant (e.g., the legal/scientific model of Raymond Westbrook; the literary models of John Van Seters and David P. Wright). Jackson then introduces his view of the legislation of the Covenant Code as ‘wisdom laws’ 270

REVIEWS

and contrasts this conceptualization with other scholarly models. Of primary im portance for Jackson’s wisdom model is a ‘narrative’ vs. ‘semantic’ approach to the words of the laws themselves. By ‘narrative’ approach, Jackson means that the sense of the laws is not bound in the literal sense of their words; rather, the laws ‘evoke images of situations from known social contexts’ (25) and should be understood in light of typical social scenarios. Such a ‘narrative’ approach allows for common sense elaboration of ‘restricted code’: the sparseness of the Covenant Code’s lan guage must be informed and elaborated by common ‘values and understandings’ from these laws’ original social location. The Mishpatim can thus also function as paradigmatic cases for comparison with slightly differing disputes. This narrative approach accords well with Jackson’s understanding of the earliest Mishpatim as oral sayings meant for private dispute resolution, not public adjudica tion. As evidence for this claim, Jackson highlights the self executing character of the Mishpatim: rather than requiring third party adjudication, these originally oral sayings allowed the private parties of a dispute to resolve their differences through negotiation. To facilitate such negotiation, the ‘wisdom laws’ often advance arbi trary (i.e., objective and publicly observable) tests to determine responsibility. For example, Exod. 21:21 absolves from homicide charges a slave owner who beats his slave if the beaten slave is able to stand for a day or two after the attack. Yet, should the slave die a week after his beating, why should this elapsed time eliminate the the owner’s attack as cause for his death? The duration of a day or two is an arbi trary standard that allows for private dispute resolution between the parties. Similar is the case of killing an intruder: it is permitted if it occurs at night but not during the day (Exod. 22:1 2a). Part II comprises detailed exegetical treatments of the laws in Exod. 21:1 22:16, which Jackson divides into ten sections according to legal topics. In these chapters, the author attempts especially to highlight aspects of the legal paragraphs that ac cord with his ‘wisdom laws’ thesis, while assigning contrary details (such as third party adjudication) to later compositional layers. In this reviewer’s opinion, the real strength of this volume is Part II’s detailed consideration of the Mishpatim them selves (even if some of the material will be familiar to readers who have read Jackson’s previous works). Regardless of whether one accepts Jackson’s larger thesis concerning the wisdom origins of the Mishpatim, the author’s keen insights and sensitivity to multiple options for interpretation make this exegetical treatment of the text very valuable for serious study of the Covenant Code. Particularly benefi cial is Jackson’s extensive interaction with the scholarly literature, often contained in considerable footnotes (which are printed in a very small font size). The volume is also thoroughly cross referenced, which is especially useful in a study of this size and complexity. On the basis of the exegetical insights in Part II, Part III presents Jackson’s views of the development of legal institutions in ancient Israel and their intersection with the textual legislation of the Book of the Covenant. Jackson’s view of the literary development of the Mishpatim tracks closely that of Eckart Otto: originally oral sayings were committed to writing in legal paragraphs, which were subsequently joined into intermediate collections of Mishpatim. These intermediate collections of Mishpatim are the ‘urban’ collection, Exod. 21:2 27 (although the slave laws in 21:2 11 are later than the rest of the Mishpatim [originating from ‘a royal libera tion edict’], 21:15 17 may be secondary, and 21:22 5 are clearly an interpolation), and the ‘rural collection’, Exod. 21:28 22:14. At the final stage of development, these intermediate collections were combined with other intermediate collections of 271

REVIEWS

laws to form the Covenant Code. Jackson also suggests that the two sections of the Mishpatim likely existed independently before being combined with the rest of the Covenant Code. In his view, each of these stages of compilation like all biblical law is influenced by wisdom circles. It is ultimately difficult to evaluate the main thesis of Wisdom Laws, for the re construction of earlier redactional strata through indirect evidence and especially the recovery of originally oral sayings is a provisional exercise. To his credit, Jackson is careful to note that the institutional and literary histories that he pro poses are speculative at points and open to charges of circular reasoning (e.g., 387; 433). Nonetheless, the complexity of the compositional history that he proposes for the Covenant Code and his precise identification of Deuteronomic and Priestly re visions seem overreaching. Also problematic is the lack of a strong thesis regarding the relationship between the ancient Near Eastern legal collections and the Mishpatim. Jackson employs the matically related ancient Near Eastern laws in his exegesis of the Covenant Code; however, it is not clear how he understands the nature of their similarities. In his case for emending Exod. 21:29’s ‫‘ ישמרנו‬he has not guarded it’ according to the LXX’s reading ‘he has not destroyed it’ (implied ‫)ישמדנו‬, Jackson argues against fol lowing the ancient Near Eastern parallels (e.g., CH §251 reads alapÒu la usanniq ma ‘but he does not control his ox’), stating that such ancient Near Eastern parallels cannot ‘control one’s judgment of the semantics and institutional context of the biblical text’ (271 n. 78). He also ‘doubts’ that the original wisdom sayings of the Mishpatim were formulated casuistically (31). This means that the current form of the Mishpatim that corresponds with other ancient Near Eastern legal paragraphs is independent of the content that is likewise parallel among these legal corpora. Such a view seems unlikely. Nevertheless, these objections hardly undermine this work’s significant contribution to the study of the Covenant Code, and Jackson is to be commended for his sophisticated and nuanced analysis. doi:10.1093/jss/fgn060

JEFFREY STACKERT UNIVERSITY OF CHICAGO

HILARY LIPKA, Sexual Transgression in the Hebrew Bible. Sheffield Phoenix Press, Sheffield 2006. Pp. xii + 285. Price: £60.00 hardback. ISBN: 978 1 905048 34 2. Lipka’s excellent monograph examines the nature and consequence of sexual trans gression in three boundary areas: religious, communal and personal. Lipka argues that, contrary to common assumption, personal boundary violation was recognized and acknowledged in the Hebrew Bible. In the two transgressions she examines, rape and adultery, she argues that different texts portray differently the nature of the offence, the impact on the victim, whose rights are violated, who should punish the transgressor and what the punishment should be. Although there are places in the discussion where questions of redaction are raised and she also uses other ANE sources to illuminate where they can, the main focus of her approach is essentially text based, with a series of close readings and comparisons of words via intertexts. Although her approach intends to infer some thing about Israelite sexual attitudes ‘in certain times and in certain places’, she fo cuses on the community as portrayed within the Hebrew Bible, and its Sitz im Leben. This approach ensures the text does not float free of its moorings. It looks for the implied author rather than the empirical, although Lipka herself does not use this terminology. She underscores the importance of using all genres of writings 272

REVIEWS

to command as representative a view as can be possible of sexuality in ancient Israel, and recognizing dissonance where it exists. After an introductory reflection and caveat on the nature of sexuality as cultural construct, she begins with transgressions against religious boundaries, illustrated by adultery, variously argued by scholars to be either a tort violation, or an offence against God. Lipka argues that the view of adultery varies in different texts, and can be both these things. She illustrates first with pertinent laws, and shows the differ entiation of understanding between the Decalogue and H (adultery as solely a reli gious transgression) and D (adultery as both a religious transgression and a crime against the husband). She then moves into narrative, poetic and wisdom views of adultery. The section on transgressions of communal boundaries concerns itself with the aftermath of sexual transgression, and includes some debate into the ongoing ques tion of the semantic range of ‫ בתולה‬and also the importance of virginity to a girl’s marriage prospects. Transgressions such as sexual intercourse between two unbetrothed people are framed in terms of a violation against the community. The chapter begins the discussion on rape by examining its legal ramifications. The final section on personal boundaries also concerns itself with rape, and con trasts the narrative depiction of the victim’s devastation with the law’s focus on compensation. The importance of correct resolution for the well being of both vic tim and community is underlined, and consequences of a lack of such resolution underlined in the Tamar Amnon narrative. There is a useful appendix summarizing the definition used of the lexical items under discussion Overall the book is well argued and clearly written. Although it might have been good to explore the apparent dichotomy between the laws a group in society for mulates and the stories the society tells, it is nevertheless very useful to have the two in proximity in discussion, allowing perspective on both that either could not have alone. In particular it allows for the proximity of the informal and formal sanctions (laws) she identifies by which a society regulates its behaviours, and works out in practice her caveat against simply equating the legal corpora with sexual norms in ancient Israel. She rightly suggests that this material was not a legal code for the courtroom. Nevertheless, in relegating the legal material to a theological construct on how society functions ideally, she overlooks the possibility of some of the legal material as perspective on how matters were to be settled in the community with out recourse to the courts (as discussed by B.S. Jackson with reference to the Cov enant Code). The lexical approach works well generally, but most of all when Lipka combines its usage with other literary strategies which underscore the valuable points she makes. More discussion of why virginity is as important as she believes would also have been useful. Extensive use of footnotes means that a considerable contribution is made to de bates in the scholarly community about such areas as the nature of adultery as char acterized in the Hebrew Bible and the vexed question of which verbal markers de note rape and which do not (in particular in respect of Genesis 34). But the footnoting of the more detailed argument and the clear style of the main text also means that this book will be of interest to experienced students in biblical studies interested in exploring the portrayal of sexual transgression in more detail, and the proximity of discussion on law and narrative will be particularly valuable to them. doi:10.1093/jss/fgn061

JENNI WILLIAMS UNIVERSITY OF MANCHESTER

273

REVIEWS

KEVIN A. WILSON, The Campaign of Pharaoh Shoshenq I into Palestine. (Forsch ungen zum Alten Testament, 2. Reihe, 9). Mohr Siebeck, Tübingen 2005. Pp. viii+ 180. Price: £39.00 paperback. ISBN: 3 16 1482270 0. Students both in Egyptology and of the Hebrew Bible have made frequent use of the triumphal scene and Palestinian geographical list left at Karnak by the Libyan founder of Egypt’s 22nd Dynasty, Shoshenq I (c. 945 24 BCE), both for Egypt’s for eign relations then and as background to the attack by ‘Shishak’ (better, Shushaq) upon Rehoboam’s Judah reported in Kings and Chronicles. A compact and clearly articulated monograph by Kevin Wilson has now attempted a reassessment of that monument (and others), of the Hebrew allusions and of how all these should be understood. The undercurrent running through his work is mildly minimalist to wards both Egyptian and Hebrew data. But more serious is its author’s too limited study and use of merely four cases of lists with scenes, and not of the full corpus. This leads to somewhat flawed results. In ‘Asiatic’ triumph scene lists, there is a discernible progression from the first hand lists of Thutmose III into compound lists (Sethos I; Ramesses II in part) con taining both derived older material and entries that do reflect actual campaigning by 19th Dynasty kings, ending with the long, purely ‘literary’, lists of Ramesses III that derive almost wholly from their predecessors. In this context, the later list by Shoshenq I stands out a mile; its list is not derived from any prior list, as 90% of preserved names are entirely new; the other 10% of names are largely the strategi cally ‘unavoidable’ places that nobody overrunning Palestine could ignore. That fact requires a proper explanation; in Wilson’s book, none is forthcoming. In Chapter 1, Wilson essays a critique of previous studies, with much that is per tinent. However, not every objection is valid. He insists that the Shoshenq list indi cates a campaign against Judah, but not against Israel. But from 1 Kgs 11:40, 12:2, it is clear that the fugitive Jeroboam was a client of Shoshenq and (politically) was his protégé and thus his puppet. The stela of Shoshenq at Megiddo marks pre cisely his impact there, and the sequences in the list make military sense and are not arbitrary. See the base map given in Kitchen, Third Intermediate Period in Egypt (all editions), p. 434, fig. 9, using universally accepted identifications almost through out. In Chapter 2, Wilson shows only limited understanding of the textual elements in triumph scenes, with their own history, and sometimes historical; cf. K.A. Kitchen, G.A. Gaballa, ZAS 96(1969), 23 8 (§4), and on Shoshenq’s text, Kitchen, Poetry of Ancient Egypt (Documenta Mundi: Aegyptiaca I), Jonsered 1999, 433 40, §64. In his Chapter 3, Wilson reviews the elements in triumph scenes, and rightly stresses their symbolic nature, as regards the visual ideogram of ‘victorious king smites foes before deity’, the often generalized nature of the triumphal addresses and of superscriptions to the lists. He then jumps to the conclusion (pp. 64 5) that, like all the other elements here, the actual topographical list is entirely sym bolic of victory, and of no direct historical value or reference hence (although of interest for geography) it cannot be used to reconstruct the course of Shoshenq I’s campaign. But this represents a massive non sequitur, and is readily falsified by the clearly historical element (the building text) in the triumphal text in this scene, and his failure to note the unique nature of the topographic list in this case. In his Chapter 4, Wilson reviews four other Egyptian inscriptions (three under Shoshenq I) which have been associated with that king’s Palestinian campaign. First, the stela Gebel Silisila 100, of Year 21 of Shoshenq I, explicitly commission 274

REVIEWS

ing his known major works at Karnak including, thus, the great triumphal relief, as Wilson admits (p. 66). Second, the Karnak stela of Shoshenq I is definitely a mili tary record, gives a reason for armed action (‘vile things’ somebody [else] had done), followed by [the king going out], ‘his chariots following him’ (it would not be anybody else!) And so ‘His Majesty made great slaughter’ of the foe, and men tion is made of the Bitter Lakes. This is the E. border with a foreign foe, not some minor local squabble (W.’s Dakhla ‘parallel’ is not apposite). Third, the biographical allusion of Hori (rather than Heru) to following the king on campaign in Retenu may reflect his presence in the Levant with Shoshenq I but (with Wilson, p. 70) not necessarily; we can agree that it is indecisive. Fourth, the Megiddo fragment of a former royal stela of Shoshenq I. Here, again, Wilson wriggles away from the obvious significance of its presence in Megiddo. Such stelae, whether erected in towns or on rock faces always indicate Egyptian overlordship, and often direct control, including the direct result of campaigning by pharaohs at the time. So, the Shoshenq I stela can be no different: it represents, however briefly, that king’s dominance over Megiddo and over its local ruler that is, Jeroboam, king of Israel, Shoshenq’s protégé and intended vassal. In Chapter 5, Wilson deals with the two biblical references. He accepts the essen tial features in these, but fails to understand that mention of a Shoshenq reduction of Israel to vassaldom is not necessary to the purposes of the Kings writer(s), still less in Chronicles. That the campaign was against Jerusalem alone is also contra dicted by the presence of the Negeb and south Judean sites included in the List. Contrary to Wilson (very careless here), the Chronicler does not claim either that Shoshenq’s conquest ‘devastated… the entire [Judean] countryside’ or that it ‘en compassed all of Judah’ (p. 85). Chapter 6 brings us to the book’s final conclusions. These begin with the errone ous statement that the List can offer nothing on the campaign, merely because the Relief is merely a formal work. He is right to say that the Megiddo stela does not imply a Shoshenq destruction of Megiddo; quite the contrary, it was there to re mind a live population where their ultimate loyalty should go. It does indicate Egyptian overlordship. Finally, as an Appendix, Wilson usefully provides a full conspectus of the surviv ing place names in the List, and of possible readings and interpretations. P. 103 n. 8: contra Wilson, the El Hiba relief was definitely a triumph scene, but is 98% lost. For List No. 12, in agreement with W, M[aq(q)eda?] is possible epigraphically because the first bird is not 3 but m. Gezer is ruled out, and Rainey (as cited by W.) cannot overrule the Chicago edition on this. However, M[aqqeda] is rather off track; M[aresha] might be geographically superior. With Nos. 15 18, 22 7, 32 5, 38 9, we would be largely in happy agreement. Nos. 53 6 form a clear group (p3 ndjskt, error in hieratic for p3 n( )skt, ‘one of Succoth’?). For Nos. 57 9, Zemaraim is beyond reproach; restoring [Ti]rzah seems inevitable, hence ‘the Tower’ should be Shechem, known from Judg. 9:46; No. 65 ‘the Vale’, that of Esdraelon. Again, most of this is agreed widely, leaving lost entries aside. Thereafter, in rows VI X, it is broadly agreed that we have names in the Negeb and around, but not all solvable in detail. Not surprisingly, W. objects to my reading ‘height of David’ in Nos. 105 6, where he rides rough shod over both the reading and the comparative data. The reading of hadibiyat, ‘heights’ in No. 105 is perfectly fairly restorable on the total known traces. In No. 106, Dwt is also certain. I had never thought of reading it as ‘David’ (Dwd) until noticing ‘the psalms of Dawit’ in a context of Psalm 65 in an Ethiopic victory inscription of Abreha in Old South Ara 275

REVIEWS

bia. Here, Dawit IS ‘David’, and cannot be any other. If one Afro Asiatic language can do this, so can others. And they do. In Egyptian, voiced terminal consonants in Semitic loanwords can become voiceless e.g., Hebrew hereb, ‘sword’, appears as H r p in Egyptian. See J.E. Hoch, Semitic Words in Egyptian Texts of the New King dom and Third Intermediate Period, (Princeton 1997), 233 5, No. 324. So, there is no practical problem whatsoever in Dwd becoming Dwt here. On Nos. 110 12, all three should be taken as one (against Na’aman), as ‘Arad of Beth Yeroham’; there is no ancient rule against triplets! The interpretations of Nos. 121, 124 7, 130, 133, 139, 150 find wide acceptance as cited by Wilson (omitting, again, lost and dubi ous entries). Finally, in row X, 2a 3a have wide acceptance; then, Äin goren (Wilson) for 4a is certainly feasible, if still unattested. Space limits preclude further comments on this topic. Wilson’s work is a useful discussion, but is not sufficiently au fait with the full available knowledge about topographical lists to provide a fully balanced view of that of Shoshenq I. doi:10.1093/jss/fgn062

K.A. KITCHEN UNIVERSITY OF LIVERPOOL

THOMAS B. DOZEMAN and KONRAD SCHMID (eds), A Farewell to the Yahwist? The Composition of the Pentateuch in Recent European Interpretation (Society of Bibli cal Literature Symposium Series 34). Society of Biblical Literature, Atlanta 2006. Pp. viii + 197. Price: $24.95 paperback. ISBN: 1 58983 163 2. This collection is the fruit of a special session of the Pentateuch Seminar of the So ciety of Biblical Literature held in November 2004. Six of the papers included were read at that meeting (those by Thomas Römer, Konrad Schmid, Jan Christian Gertz, Thomas B. Dozeman, Christoph Levin and David M. Carr); three (by Albert de Pury, Erhard Blum and John Van Seters) have been added. The session was organized by Dozeman, Römer and Schmid in the wake of the publication of the similarly entitled controversial series of essays edited by Gertz, Schmid and Markus Witte, Abschied vom Jahwisten (BZAW 315; Berlin 2002), itself prompted by monographs by its three editors. The present work provides not a mere transla tion of some of the German contributions to that earlier collection but new essays (in English) by several of its contributors and new responses to them that are in tended to carry the discussion forward. Compact but widely ramifying in content, this collection thus provides in relatively brief compass valuable surveys of ‘the state of the question’ in key areas of current Pentateuchal criticism. The hope of the editors is that this collection will ‘facilitate communication be tween European and North American scholars’ (perhaps one should say, ‘some mainland European scholars’; North America includes both USA and Canada). The marked difference between these Continental and North American approaches is already signalled in the title of the book: the question mark in the title of the American published volume is missing in the 2002 German published collection (though one was present in its originally projected form; its eventual omission pre sumably marked increasing conviction about the thesis). The view promoted by Römer, Schmid, de Pury, Gertz and Blum, in a variety of ways and on the basis of a variety of biblical sources, is that it was the Priestly au thor, or at least the author of PG, the ‘Grundschrift’ that probably ends at Exodus 40, not ‘J’ in any of its guises, who was the first to combine in literary form, in the exilic period or shortly after, the up till then independent primeval history and tra ditions of the ancestors contained in Genesis with the alternative and competing story of Israel’s origins in the exodus under Moses in Exodus. The creation of ‘the 276

REVIEWS

master narrative of salvation history’ is the work of P; the non P material that pre supposes this master narrative is thus necessarily post P redactional. Some examples of the Continentals’ argumentation may be given. There are sur prisingly few overt literary links between Genesis and Exodus (and beyond) and these turn out to be late (e.g., Genesis 15) (Schmid). Hosea 12 demonstrates pre exilic rivalry between the Jacob and Moses traditions (de Pury). The respect enjoyed by Joseph in Egypt and the enslavement of the Israelites there are mutually exclu sive; the attempt to link Gen. 50:24 with Exod. 1:6 produces tragicomic results (Gertz). Undoubtedly, the motif of the burial of Joseph’s bones interlinks Gen. 50:25, Exod. 13:19, Josh. 24:32, with its further relation to the purchase of ground in Gen. 33:19, but that does not necessarily imply early literary connection (Blum, in a change of mind from his earlier published position). The continuation of Exod. 2:23aa in Exod. 4:19 shows that the burning bush episode, Exod. 3:1 4:18, that in any case presupposes Israel’s cry in Exod. 2:23 5, is a post-P insertion (Schmid); the first and third signs in Exod. 4:1 9 are modelled on P’s plagues (Blum). Such views, and many besides, are taken up point by point and met with vigor ous counterarguments by, especially, the Americans. Levin too diverges from his fel low Continentals by defending the existence of a coherently edited collection of non P material to which he attaches the label ‘J’. Essentially, for the Americans at least, non P is pre P. Dozeman presents a strong argument that Moses’ ‘wilderness’ vocation in Exod. 3:1 4:18 is prior to his ‘Egypt’ vocation in P in Exod. 6:2 7:13 and provides thematic links for a pre P narrative running forward into Numbers. Defending his view of J as author and historian, Van Seters launches a telling as sault on the notion of ‘redactor’. In its current form, he traces that notion to Rendtorff ’s application in redactio historical terms of Noth’s traditio historical ‘block’ model of originally independent Pentateuchal traditions. This reviewer, an off shore European, will not seek to redress the balance; indeed would suggest that the points made by the Americans should be taken considerably further. A striking feature is the generally eirenic tone in which the discussion is conducted and much credit is due to the organizers of the Seminar for the patient maintenance of debate where views are so divergent and the stakes are so high. It may be that there is a fundamental difference in mindset between the two groups. The commitment of the Continentals is to analytical study: the ambition is to give as complete an analysis of the literature as possible as a necessary preliminary to his torical reconstruction. Thus even slight divergence and subtle distinction in the text provide valuable evidence for the complexity of a process of literary growth that must not be overlooked but be given due weight. No premature synthesis can do this complexity justice. As Carr remarks with regard to Schmid and Gertz: ‘the starting point is that the texts are separate until “proven combined”’ (pp. 162f.). These Continental contributions are to be respected as challenges to easy synthesis. But, it seems to this reviewer, the results of analysis often go against the grain of the biblical narrative. The separation of Exodus 1 and 2, for instance, by K. Schmid (and others; cited p. 83) leaves the motivation for Moses’ exposure on the Nile un clear. He surmises shame on his mother’s part at an illegitimate birth (see his earlier work Erzväter und Exodus, 1999, p. 154). Thus a slur on the Aaronic matriarch is proposed to promote a theory which is highly likely never to have corresponded to the way in which she was regarded at any point in the evolution of the text. An unsatisfactory feature of the Continental consensus evident in this collection is the term ‘non P’. The material in the Pentateuch that originates in independence 277

REVIEWS

of P deserves a positive appellation, particularly if, as in the view of this reviewer, ‘non P’ is pre P. It is here that Deuteronomy, conspicuous by its marginality in this collection, is of relevance. Where parallels between Deuteronomy and the preced ing Tetrateuch are occasionally brought in, as, e.g., by Carr in his tabulation on p. 173, they are instantly helpful in detecting pre P coherence, a narrative that is presupposed by, even affiliated to, D. I am amused by de Pury’s defence of the ex istence of a Jacob tradition presupposed by Hosea 12, ‘The Jacob story does not have to be retold by the Hosean poet…: obviously it suffices for his audience to hear one word or allusion to make the connexion, immediately, with the well known story’ (p. 62), for this is precisely the argument I have used for the existence of a Patriarchal narrative in Genesis presupposed by D: ‘The story is so well known that it does not need to be retold; so vast indeed, that it cannot be retold’ (FS E.W. Nicholson, 2003, p. 227). But in contrast to ‘one word or allusion’ in Hosea, Deu teronomy, at a conservative estimate, alludes to the oath to the Patriarchs in some twenty nine widely scattered verses. The identification of affiliation is not simply through the use of similar terms, syntax, motifs, forms, or literary structure (so Dozeman echoing Michael Fishbane) but also through contrasting views of Realien, e.g., diverging concepts of itinerary, chronology and institution, especially Festivals. In this case, very far reaching de ductions can be made. I have argued that the version of the Tetrateuch affiliated to Deuteronomy is not only echoed in the reminiscences in Deuteronomy 1 11 but is the narrative counterpart of the legislation in Deut. 12:1 16:17 (it is striking that Deut. 16:18 is headed by the rubric ‘Judges’, just as Deut. 17:14 turns to the ques tion of a ‘King’; cf. 1 Samuel). Thus the legislation on treatment of the Hebrew slave in Deut. 15:12 18, for instance, is key to understanding why the Book of the Covenant, as part of a composition affiliated to D, begins with the release of He brew slaves. The Joseph story in Genesis 37; 39 50* presents precisely the treat ment of the indigent Hebrew sojourner who ought not to be treated like a slave that points up the outrages of Exodus 1 15* and thus is necessarily part of the same narrative. Repeatedly in this collection Wellhausen’s proposal is cited that Exod. 3:1 4:18 is an insertion because it breaks the connection between 2:23 and 4:19. But it is surely evident that 4:18 26 is the insertion: 4:27 resumes precisely from 4:17. Furthermore, 4:22 3 contains reuse of material transposed from the D version of Plague X from Exod. 11:4. This phenomenon of the transposition of D material by P, which this reviewer has argued for in many other contexts, e.g., Exod. 15:22* 19:2* and Numbers 13* 14*; 16*, suggests that P is the later writer who, in the process of adding substantial new materials, is also editing, sometimes drastically, an already existing version. Levin has suggested that my use, in my con tribution to the original Abschied volume, of the legislation in Deut. 12:1 16:17 to help to recreate a D affiliated version in Exodus ‘schießt über das Ziel’ (TR 69, 2004, p. 341, a copy of which he kindly sent me). But not, I should submit, if one is into the study of synthesis. doi:10.1093/jss/fgn063

WILLIAM JOHNSTONE UNIVERSITY OF ABERDEEN

ALICE HUNT, Missing Priests: The Zadokites in Tradition and History (Library of Hebrew Bible/Old Testament Studies 452). T. & T. Clark, New York/London 2006. Pp. xiv + 218. Price: £65.00 hardback. ISBN: 978 0 567 02852 5. This book’s central proposition is that the supposed leading role of the Zadokite priesthood throughout the first millennium BCE in Israel is a scholarly fiction based 278

REVIEWS

on unsubstantiated assumptions that date from pre critical scholarship and were re inforced in the late nineteenth century by Wellhausen’s synthesis of the history of the priesthood which adopted the same assumptions. Hunt argues that even though the ‘sons of Zadok’ were, and still are, assumed to have dominated the Jerusalem Temple from the time of David down to the Hasmonaean period, there is simply no evidence to this effect, either in the Old Testament/Apocrypha or in other re lated literature of the period such as the Letter of Aristeas, Jubilees and Josephus. The term ‘sons of Zadok’ appears four times in Ezekiel 40 8, and apart from that is found only in the Dead Sea Scrolls, where it appears in the Community Rule (CD/ 1QS), the Rule of the Congregation (1QSa) and the Blessings (1QSb) and in 4QFlorilegium in a midrash on Ezekiel 44. Hence, it cannot be assumed that a priestly group known as ‘sons of Zadok’ even existed at a period significantly prior to the date of the Dead Sea Scrolls, let alone that such a group dominated the cleri cal hierarchy of the First and Second Temples. The strategy adopted in order to demonstrate this proposition is as follows. The first chapter is a brief review of modern scholarship together with a statement of the issues, and declares, ‘The present study makes an assessment of the Zadokites by locating them historiographically in biblical scholarship, by analysing historiographic methods used in this scholarship, and then by offering a portrait of the Zadokites through historiographic, literary, and social scientific analysis’ (p. 12). Chapter 2 then gives a much broader review of scholarship on the Zadokites and chapter 3 presents a discussion of historiography, ending with a cou ple of pages on social scientific methodologies, before the textual discussion begins. Chapters 4 and 5 cover the biblical evidence for Zadokites, chapter 6 covers the extra biblical evidence (1 and 2 Esdras, Josephus and the Dead Sea Scrolls), chapter 7 offers ten pages on the Zadokites at Qumran, and then finally chapter 8 discusses Gerhard Lenski’s macro sociological theory and how it might be fitted to the Hasmonaean kingdom, before concluding on the final page that the ‘sons of Zadok’ were a sectarian group that developed in the Hasmonaean period. As a piece of scholarship, this is less than impressive. As already stated, the main thrust of the work is that the supposed centuries long dominance of the Zadokites as an identifiable group is purely an assumption with no supporting textual evi dence; and it is true that there is very little direct textual support for such a con struction. Inasmuch as it challenges scholars to think again about received para digms instead of simply accepting them unquestioningly, the book serves a useful function. However, there is little else to commend it. In fact, the overwhelming impression is that this one basic insight has been padded out to book length pro portions by the addition of large amounts of extraneous material, which in the final reckoning add little or nothing to the main case being argued. Chapter 3 on historiography, for example, reviews the work of Lemche, Thompson, Dever and others, and attempts to show how their approaches and viewpoints are relevant to a study of the Zadokites; but the supposed links between these scholars’ ideas and the Zadokite question are truisms. Consider the following insights gleaned from Lemche’s work: ‘[B]iblical references to the Zadokites… reflect the socio political milieu of the time in which the references were written. … The evidence of the presence and role of the Zadokites depends heavily upon the dating of the pericopes’ (p. 59). Perhaps I am missing something here, but these seem to me to be basic tenets of sound biblical scholarship, rather than startling new methodo logical insights that need pages of introduction and explanation. (But then again, perhaps that is an unwarranted assumption on my part.) The textual treatments in 279

REVIEWS

the following four chapters are equally disappointing. Calling them ‘analyses’ would be a misnomer, as more often than not they consist of a quotation of the relevant passage in Hunt’s own translation a self indulgence that is not justified by any apparent scholarly or exegetical rationale in the subsequent discussions plus a re statement of the content of the passage with no additional critique or comment. Sometimes there is a review of scholarly opinions about a given quotation, but the amount of actual analysis and original thought about the texts is disappointingly small. The most disappointing chapter from this point of view is chapter 5, on the ref erences in Ezekiel to the sons of Zadok. Given that this is a crucial chapter in the book, in that it deals with the only occurrences of the phrase ‘sons of Zadok’ that appear in the Old Testament, I would have expected a rigorous analysis and lively interaction with existing scholarly work on the topic. Instead, in a twenty page chapter, what we get is a review of scholarship on Ezekiel 40 8 (6 pages), emphasis on the crucial nature of the Ezekiel material plus author’s translation of Ezek. 40:45 6, 43:18 19, 44:6 16 and 48:9 11 (2 pages), further reviews of the schol arship on Ezekiel produced by Blenkinsopp and Cook (four and a half pages), a suggestion to read Ezekiel 44 in the light of Ezekiel 23 plus another author’s trans lation of Ezek. 44:6 16 (one and a half pages), author’s translation of Ezek. 23:1 49 plus extensive bibliographical footnote on the whore metaphor in Ezekiel 16 and 23 (3 pages), followed by paraphrasing of the content of the translated chapter 23 (half a page), and finally the suggestion (accompanied by more author’s transla tions of Ezek. 23:38 9 and 44:6 9) that Ezekiel 44 is a ‘midrash’ on Ezekiel 23 and that it should be linked with concerns about foreigners serving in the Temple (1 page). The remaining page or so of the chapter consists of summaries of the case so far, in which Hunt comments of the Ezekiel material, ‘Whether we place it in the late sixth century BCE or the Hasmonean period, we must conclude that at the time of writing there was some discussion of the duties and rights of priests and that the author of the material considered priests called ‘sons of Zadok’ as central to service of YHWH’ (p. 142) hardly an earth shattering conclusion. Where are the historiographical and social scientific methodologies trumpeted so fervently earlier on in the book? How does all that discussion in chapter 3 feed into the tex tual treatments in chapters 4 to 7? As far as I can see, not at all; Hunt is doing nothing more than many historiographically unregenerate but intellectually aware biblical critics have been doing for decades. True, she has altered her starting (or concluding?) assumption on the basis of her observations about the textual evi dence, but the way in which she treats the texts is certainly not revolutionary, not least because in places she hardly treats them at all. Lenski’s social scientific model does eventually make an appearance in chapter 8, which is headed ‘Conclusions’ does she not know that conclusions are not the place to introduce new mate rial? but even here the model is applied to the Hasmonaeans without any men tion of the Zadokites, and its purported relevance to the Zadokites is only finally made semi explicit on the final page of the book in a summary of the whole work. Hunt is also unable to contemplate any other explanation but her own thesis for the absence of the designation ‘sons of Zadok’ from the textual record that has sur vived, and strenuously denies the possibility of any group within the priesthood being recognized even informally as ‘sons of Zadok’, to the point that she is forced to misrepresent my own published comments on the subject by taking some of them out of context and giving a false description of the context of others of them. While this misrepresentation is disappointing, a more serious flaw in her scholar 280

REVIEWS

ship is that she makes no attempt to address the issue of the ideology of Samuel Kings; such an examination might suggest an alternative reason for the dearth of references to priestly houses in general and sons of Zadok in particular in the pre exilic record, which is the most significant absence of evidence for the so called ‘tra ditional view’. But Hunt is content simply to treat this Deuteronomistic silence as evidence of absence, and move on to other texts without a backward glance hardly the act of a responsible historiographer, one would have thought. My initial assumption about this book was that it originated as a doctoral thesis. However, this was not stated explicitly in the front matter, and upon encountering the elementary level flaws discussed above, together with the book’s excessive reli ance on often quite lengthy quotations from secondary works and a raft of typo graphical and grammatical errors and obscurities, I found myself questioning this assumption. But further investigation proved my initial assumption correct (vae tibi, Vanderbilt!), thereby confirming the principle that absence of evidence should not necessarily be taken as evidence of absence. If only the same principle could have been applied to Hunt’s investigation of the sons of Zadok. doi:10.1093/jss/fgn064

DEBORAH ROOKE KING’S COLLEGE LONDON

LENA SOFIA TIEMEYER, Priestly Rites and Prophetic Rage: Post Exilic Prophetic Cri tique of the Priesthood (Forschungen zum Alten Testament 2. Reihe 19). Mohr Siebeck, Tübingen 2006. Pp. xvii + 318. Price: /59.00. ISBN: 3 16 149059 2. The relationship between priests and prophets is a perennial source of speculation for scholars, and although thankfully we have now moved away from the denigra tion of ritual that once pervaded Old Testament studies, so that prophetic criticism of cultic activities is no longer automatically assumed to be opposition to the cult per se, there is still the issue of how the two areas of religious specialism related to each other, and precisely how any prophetic critique of cultic activities should be interpreted. These are the questions addressed in the present work. A revision of Tiemeyer’s 2002 Oxford doctoral thesis, the book offers a close textual reading, in the historical critical tradition, of four post exilic prophetic writings which Tiemeyer argues contain criticism of the priesthood: Haggai, Zechariah 1 8, Malachi and Trito Isaiah. The reading offered of these four central texts is sup ported by comparison with additional biblical material, both pre and post exilic, in which the same criticisms of the priesthood appear as in the central four. Follow ing the mandatory review of scholarship, Tiemeyer begins by defining and defend ing her view of precisely which parts of the four central texts should be understood as a critique of the priesthood, before moving on to examine the nature of this cri tique. First, she interprets Trito Isaiah and Malachi to show that the priests them selves are conscious of no wrong doing, and then she sets out what she sees as the content of the critique against the priests: they lack knowledge and have failed to fulfil their duty to teach the people (Isa. 56:9 12, Mal. 2:1 9); they are encourag ing social injustice (Isa. 58:3 5; Zech. 5:1 4; Mal. 3:5; cf. Nehemiah 5); they are carrying out unorthodox sacrificial rites (Isa. 57:6 8; Isa.65:3 4; Isa. 66:3); they have intermarried with foreign women (Ezra Nehemiah; Mal. 2:10 16); they are supporting foreign political alliances (Isa. 57:9 10); they are neglecting the Yahwistic cultic regulations (Mal. 1:6 14; Isa. 61:8); and they are themselves ritu ally impure and so unable to function to purify the people (Hag. 2:10 14; Zechariah; Malachi). However, despite this raft of criticisms, argues Tiemeyer, most of the prophets examined do not urge the destruction of the priesthood, but have a 281

REVIEWS

range of views on what should happen to it. Zechariah and Malachi envisage a purified and cleansed (Levitical) priesthood that will function as God intended it to; and although Isa. 66:1 6 predicts the destruction of the priesthood, other voices within Trito Isaiah advocate changes in its composition: a generalized priest hood of all Judaeans (Isa. 61:6), a priesthood that can include proselytes as well as native Jews (Isa. 56:1 8), and a priesthood that includes Gentiles (Isa. 66:18 24). As a revision of a doctoral thesis, the book bears the hallmarks of the thesis genre in the rather self conscious reviewing of scholarly opinions at every step and the minute setting out of each stage in the argument. There is a good deal of close tex tual discussion, which does not always make for easy reading, and an almost text book like breaking down of the subject matter into bite sized paragraphs demar cated by headings and sub (sub sub )headings, with surveys of arguments before they commence and summaries of them once they have been presented. While the evident desire to present the material in a clearly organized fashion is commend able, and the format does allow the reader to home in on precisely the paragraph that may be of interest for whatever reason, it does not really encourage one into the sweep of the argument, and has a rather laboured feel to it. In particular, the introductory review of scholarship on Isaiah 56 66 I found somewhat abstruse, owing in part to the fact that the way in which the headings are used gives an erro neous impression of the direction of the argument; and the highly detailed discus sion in chapter 2 about which parts of Trito Isaiah address the priests is broken down using as many as five levels of headings in places, an unnecessary extra fussi ness which serves to atomize, rather than to clarify, the flow of thought. Easier reading are the parts of the book where Tiemeyer is drawing conclusions and dis cussing the implications of the detailed argumentation here there is a good sense of what is driving the thesis and how the exegetical work feeds into the overall pic ture. Indeed, where the work is able to keep sight of the provisional nature of the conclusions that can be drawn as conclusions about the text rather than necessarily about the reality behind the text, it is careful and solidly based, and the result is generally convincing. My main reservation about this project is, however, its use of prophetic critique to establish ‘facts’ about the priesthood. It is not that Tiemeyer is unaware of the difficulties involved in her chosen approach; indeed, she addresses them overtly in the Introduction, where she says, ‘[W]hile the result of our enquiry will tell us a lot about how the priests were perceived by the prophetic writers, we shall learn signifi cantly less about the actual behaviour of the priests in this era’ (p. 2). She even re peats this sentiment at the end of chapter 4, where, having argued that despite the prophetic critique the priests believed themselves to be in the right, she moves on to explore precisely what the prophetic accusations consisted of, with the proviso that we must keep in mind that accusations are not the same as unbiased accounts of historical events (p. 112). Yet in the conclusions to the following chapters that deal with particular aspects of the prophetic critique, there is often a tacit assump tion that the priests actually are in the wrong and the prophets are right, an as sumption that comes through most clearly in the concluding remarks for the vol ume as a whole. In the course of the book, Tiemeyer has compared the post exilic critique of the priests with similar critiques found in the pre exilic prophets, for example, in Amos and Ezekiel, and in the concluding remarks she comments, ‘Rather than attributing the post exilic priests with orthodoxy, reformed through their suffering following the destruction of Jerusalem and/or through their exilic experience, I have endeavoured to show that the emerging new clergy were not so 282

REVIEWS

very different from their pre exilic predecessors’ (p. 287). There seems to be here the implied conception of a kind of pristine agreed orthodoxy from which pre exilic clergy deviated, to which the exile was intended to recall them, and according to which they are now legitimately being judged and condemned by the prophets. However, maybe what is evidenced by the continuation of the prophetic critique from pre exilic to post exilic times is that orthodox Yahwism/Judaism as tradition ally understood took much longer to prevail than is generally assumed, and that during the sixth and fifth centuries BCE it was still being negotiated between oppos ing groups with different conceptions of what was legitimate. Certainly, to that ex tent Tiemeyer is correct to say that the post exilic priests seem not to be so very different from their pre exilic predecessors, but in assessing the implications of that position righteousness should perhaps not quite so readily be assumed to lie with the prophets. On the other hand, the ease with which such an assumption is so of ten made in all sorts of scholarly contexts, not just here certainly demon strates the potency of the prophetic rhetoric, which was arguably one of the factors in creating the concept of ‘orthodoxy’ as we know it (or assume it). A final niggle about the production of the volume: as well as a bibliography, it includes three indexes, Source Index, Subject Index and Authors Index. However, none of the three indexes covers material in the footnotes, which given that foot notes are a major location of author and text citations seems odd to say the least, and limits the usefulness of the indexes. In this computerized age of book produc tion it should surely be possible to create indexes that cover the whole text of the book including the footnotes. doi:10.1093/jss/fgn065


NICHOLAS P. LUNN, Word Order Variation in Biblical Hebrew Poetry: Differentiating Pragmatics and Poetics (Paternoster Biblical Monographs). Paternoster, Bletchley 2006. Pp. xxii + 373. Price: £24.99 paperback. ISBN: 978 1 84227 423 1. The raison d’être of Nicholas Lunn’s monograph (a revision of his 2004 doctoral thesis) is to demonstrate that Biblical Hebrew (BH) prose and poetry share the same basic syntax. This would seem to be an obvious and therefore banal thesis, yet the supposed irregularity of poetic word order and the all too common practice of disregarding syntactic variation in, e.g., the Psalms, as a reflex of ‘poetic style’ high lights the importance of this research question. Lunn’s work fills a niche, if for no other reason than to put an end to the avoidance of serious engagement with the syntax of BH poetry. The book has 11 chapters, which fall into two primary sections. In the first sec tion Lunn defines the problem (chapters 1 2) and establishes his linguistic model for analysing information structure (chapter 3). In the second section he uses this model to explain the various syntactic and pragmatic issues in BH poetic lines (chapters 4 8) and even whole texts (chapter 9). The last two chapters discuss the merits and faults of alternative approaches (chapter 10) and provide a summary (chapter 11). At the heart of Lunn’s study is the thesis that for all of the ‘defamiliarization’ techniques used in poetry, any word order variation ‘has to remain within the limi tations imposed by the syntactic constraints of the language’ (p. 5). Thus, the ma jority of this work represents the working out of the grammar of BH word order variation and its manifestation in biblical poetry; Lunn has set for himself the task of distinguishing between syntactic or pragmatic influences on the word order in a 283

REVIEWS

given poetic line and other poetic influences, such as chiasm. To do so, he builds especially on an information structure model proposed by Knud Lambrecht (Infor mation Structure and Sentence Form: Topic, Focus, and the Mental Representation of Discourse Referents, Cambridge 1994), which he summarizes and applies to BH prose in chapter 3 before applying it to biblical poetry in the following chapters. I find most of Lunn’s explanations in chapters 4 9 to be plausible and many to be likely. For example, on p. 123 he explains the structure of the three lines in Ps. 138:6 ki ram yhwh // wesapal yir’eh // wegaboah mimmerÌaq yeyeda¨ ‘Because Yhwh is exalted // yet he sees the lowly // but the haughty he knows at a distance’ (my translation). The contrast between the first two lines is apparent (though God is exalted he still notices the humble), but it is the third line that illustrates the essen tial correctness of Lunn’s distinction between syntactic constraints and techniques of poetic defamiliarization. He convincingly argues that the O PP V order of the third line is not due to defamiliarization but rather has a syntactic pragmatic expla nation; specifically, the initial object NP gaboah is a contrastive topic, the fronted PP mimmerÌaq carries focus, and the final V yeyeda¨ is presuppositional, since it is broadly related to the preceding verb yir’eh. The result of such pragmatically moti vated reordering of the syntax is that, ironically, while God from his high position takes notice of the ‘low’, he maintains his distance from the ‘high’, or as Lunn ex plains, ‘God does know these people, just as he knows all human beings, but these haughty ones he knows from afar’ (p. 123; italics are Lunn’s). That said, I question Lunn’s wholesale acceptance of Lambrecht’s model of infor mation structure; Lambrecht’s theory, while admirable in its scope and refinement, has serious problems, especially relating to the definition of focus. For instance, Lunn (following the work of his thesis advisor, Heimerdinger) categorizes the S V clause hannaÌas hissî’anî ‘the serpent deceived me’ (Gen. 3:13, my translation) as an ‘event reporting’ clause with ‘sentence focus’ and asserts that this ‘is not a comment about [Eve’s] own action (What did you do?), but a clause in which both subject and predicate share equal focus (What happened?)’ (p. 80). I have two issues with this analysis. First, if BH is a V S language (as Lunn asserts on p. 4), why are event reporting clauses not V S? Lunn never addresses this oddity that basic word or der does not fit into the very category of ‘focus structure’ (sentence focus) that lacks any ‘pragmatic presuppositions’. Into which of Lambrecht’s three focus structures, then, does a basic V S clause fit? Second, it is highly counterintuitive that all mean ingful propositions fit into one of Lambrecht’s three focus structures, or to put it another way, that there are no ‘sans focus’ meaningful propositions (note that Lambrecht explicitly distinguishes focus from ‘new’ information). It is a question able model that does not allow for the existence of a basic word order, pragmatically neutral clause (which are admittedly rare but nonetheless possible). Lunn notes that parallel S V clauses, such as pî Òaddîq yehgeh Ìokma // ûlesônô tedabber mispa† ‘the mouth of the righteous utters wisdom // and his tongue speaks justice’ in Ps. 37:30 (my translation), are common in Proverbs, likely due to the lack of significant prag matic context (p. 132; also p. 78). Why these clauses should be marked for focus is unclear (contrast Lunn’s analysis of such clauses with the analysis I provide in my ‘Word Order in the Book of Proverbs’ [Pp. 135 54 in L. Troxel, D.R. Magary and K.G. Friebel (eds), Seeking out the Wisdom of the Ancients: Essays Offered to Honor Michael V. Fox on the Occasion of His Sixty Fifth Birthday. [Winona Lake, Ind. 2005]). Lunn’s work is indeed valuable, if for no other reason than its defence of the es sential syntactic sameness of BH prose and poetry, which I take to be its primary 284

REVIEWS

contribution. The book should be read and recognized for taking the critical next step in our study of ancient Hebrew grammar applying advances in BH prose syntax to BH poetry and in the process to distinguish pragmatic influences on word order from poetic influences. Unfortunately, I cannot agree with Heimerdinger’s ef fusive recommendation of the book on the front cover, back cover, and in the fore word: the devotion to Lambrecht’s information structure model, which Lunn care fully and skillfully applies to the BH poetic data, is also the primary weakness of his analysis. I therefore suggest that it is time for those interested in pragmatics, infor mation structure, and BH word order to set Lambrecht’s book aside and look for other, more insightful models within general linguistics. It may very well be the other lasting contribution of Lunn’s study that we in BH linguistics relegate Lambrecht’s book to the status of supporting material rather than the central theo retical role it plays in the BH word order and information structure monographs by Heimerdinger, Shimasaki and Lunn. doi:10.1093/jss/fgn066

ROBERT D. HOLMSTEDT UNIVERSITY OF TORONTO

D. GOODBLATT, Elements of Ancient Jewish Nationalism. Cambridge University Press, Cambridge 2006. Pp. 276. Price £45.00. ISBN: 9780521862028. As I already stated in the blurb of the book, Goodblatt’s study is a fine piece of work. The book has many insights, new ideas, and contributes a great deal to re search on the issue of nationalism. My positive attitude towards Goodblatt’s book does not contradict the fact that I have some minor reservations concerning it that I will briefly relate here: 1. Nationalism as a phenomenon. It is not necessary in this short review to enter again ad nauseam the arguments whether we can speak at all of nationalism in An tiquity. G., sensibly, agrees with me on that point (D. Mendels, The Rise and Fall of Jewish Nationalism. Jewish and Christian Ethnicity in Ancient Palestine2. Grand Rap ids Michigan 1997; the latter view accepted by Anthony D. Smith, Nationalism and Modernism. London and New York 1998). What both of us actually say about Jewish nationalism in the Second Temple period is that it is a somewhat different form of the one that emerged in the nineteenth century. One should emphasize that nationalisms (in plural!) differ throughout the course of history in their con tent and the way they are embedded in societies. This is also the case with, say, an cient imperialism that is so different from the modern phenomenon in various as pects; hence one who uses the last mentioned term in his own book quite freely (S. Schwartz, Imperialism and Jewish Society: 200 BCE to 640 CE, Princeton, NJ 2001), should take back his criticism concerning the use of the term nationalism in Antiquity. It should also be noted that nationalism as a phenomenon should be set by its researchers against the background of other nations. Although G. mentions in passing the Nabataeans and other peoples, he should have dealt in depth with other nations, either of the same period or those which emerged later in history. One cannot deal with nationalism in a vacuum, as G. in fact does. 2. The main argument of G.’s book is the use of language, Hebrew, as an indis pensable aspect of nationalism. Basically I think that he is right. However, in line with Samuel P. Huntington’s observations concerning the use of a lingua franca (The Clash of Civilizations, chapter 3), I would say that the use of Aramaic, espe cially by bilingual people (see G. p. 57), did not necessarily suppress Jewish nation alistic ideas, being merely a medium for communication of Jews with others in the region. The use of a lingua franca even sharpened the separate identity of the Jews 285

REVIEWS

using Hebrew (comparable to a present bilingual Indian whose use of English for communication does not make him English in terms of nationalistic feelings). Interestingly, G. pays no attention at all to the Greek and later Latin speaking Jewish Diaspora (which was not bilingual). Does he assume that all of these Greek speaking Jews were not ‘nationalistic’? If he does, he is right; this latter view would support the thesis of A. Edrei and myself (published recently in the Journal for the Study of Pseudepigrapha 16.2 [2007], 91 137 and 17.3 [2008], 163 87) claiming that the Greek speaking Jewish Diaspora lost touch with the Eastern Hebrew speaking Judaism and its rabbis as a result of the difference in language (e.g. that the Western Jews did not know Hebrew or Aramaic, and the latter did not care to translate their corpus of Halakhic literature into Greek or Latin). Thus, the disconnectedness with the eastern (nationalistic) Jews was even stronger due to the ‘lack’ of nationalistic ideas in the Western Jewish Diaspora. Later in the history of Europe the use of vernacular turned out to be a crucial element in the construction of nationalistic ideas. This process can be observed in the actions of Jan Hus (and later Martin Luther). These and other reformers translated Holy Scripture as well as significant theological books into the vernacular. Such translations were an impor tant factor in arousing national feelings. Bohemian Czechs during the fourteenth and fifteenth centuries crystallized their national feelings against the background of the German segment there. Much of this creation of Czech nationalism was based on texts in the vernacular circulating already in the fourteenth century, a process culminating in the foundation of the Bethlehem chapel in Prague in 1394 and the vernacular preaching for which it was founded. Jan Hus became rector of the Beth lehem chapel in Prague in addition to being a professor at the university. He was conscious of the national and linguistic differences between the Czech and German populations. Thus all his life he associated this nationalistic aspect with his new re ligious reform (M. Lambert, Medieval Heresy3. Oxford 2002, chapters 15 16). It would have been useful for G. to present some analogies to this use of language, taking us beyond the Jews in the Second Temple period. 3. G.’s enormous efforts to distinguish between the three different ‘titles’ of na tionalism, Israel, Judah and Zion are very interesting indeed (chapters 5 7). How ever, if his description of the three types (and therefore, a distinction between dif ferent stages of nationalism in the history of the Second Temple period) is true, then we hardly hear what the differences between these ‘types’ were in the context of the nationalistic idea, and what expression they had on the ground (namely among the Jewish people). If they are not after all different in their content (ideo logically and practically), and are just synonyms of the same thing, then the histori cal evaluation of nationalism in ancient Israel has gained nothing at all, and the chapters remain but a discussion, however important, of the nomenclature. 4. The discussion concerning the antiquity of much of the biblical corpus (in chapter 5) is in itself important and should be addressed even more frequently by Hellenistic Judaists. In developing this argument it would have been helpful to in clude an in depth discussion of the Pseudepigrapha and the remnants of Jewish Hellenistic literature. In this literature, which quite often rewrites the Bible, we can actually see very clearly how the Jews lived with their Bible in daily life. An elabora tion on some of this literature would have supported G.’s thesis about the Bible as a canonic text, which helped build the Jewish nation (Eupolemus uses only ‘Judea’ for describing the Jews). For G.’s case it would not matter whether Jubilees was written in the early Maccabean period (VanderKam) or later in the century (Mendels), but it shows how ‘Israel’ used its heritage to endow the new created na 286

REVIEWS

tion in the Hasmonean period with legitimacy. G. would have profited from a dis cussion of this particular work (By the way, both ‘Zion’ and ‘Israel’ are used in the Book of Jubilees. For instance, 1:28; 2:31). In conclusion, G’s book has strengthened in many of its points what has been said about Jewish nationalism in recent years; the issue of its ‘fall’ (chapter 8) should not bother one too much since, as G. rightly argues, there was a continuity of the concept after the destruction of the Temple by what I have called ‘passive’ nationalism. Here one should add a general comment. It seems to me that at present scholars in the humanities should avail themselves to a much greater extent of the research carried out by social scientists. For instance, it would have been use ful in this particular book to deal extensively (and not just in passing, as on p. 115) with the problem of memory in its public and social context, in line with recent approaches of social scientists (see for a very recent overview: D. Mendels [ed.], On Memory. An Interdisciplinary Approach, Oxford and Vienna 2007), or to tackle in depth some of the (wrong) ideas of Benedict Anderson (in his Imagined Communi ties. London and New York 1991; see for example the sharp criticism of A. Hast ings, The Construction of Nationhood. Cambridge 1997, p. 194). doi:10.1093/jss/fgn067

DORON MENDELS THE HEBREW UNIVERSITY OF JERUSALEM

BEVERLY P. MORTENSEN, The Priesthood in Targum Pseudo Jonathan: Renewing the Profession (Studies in Aramaic Interpretation of Scripture). Brill, Leiden 2006. Pp. Vol. 1 xii + 452, Vol. 2 iv + 484. Price: /269.00 hardback. ISBN: 978 9004145 82 5. From the outset of this book Mortensen offers a bold interpretation of Targum Pseudo Jonathan. According to Mortensen Targum Pseudo Jonathan is a priestly document, written by priests, for priests. Dividing Pseudo Jonathan into three types of material, ‘translation (from the Hebrew Bible to Aramaic), shared expan sion (interpretive material that parallels added material in other targums), and unique expansion (interpretive material added to the Biblical translation that only one targum contains)’ (p. 1), Mortensen focuses almost exclusively on the last kind, calling it the Pseudo Jonathan Unique source, or PJu. It is this source that Mortensen argues contains the bulk of Pseudo Jonathan’s priest centered message. She thus spends most of her first volume examining almost all the PJu material, at tempting to demonstrate that it bears out her priestly origin and priestly audience hypothesis. She does this by organizing the PJu material into various categories that are also used as her main chapter headings: ‘The Priestly Heritage (Completely Priest Centered Expansions)’; ‘The Temple Job (Completely Priest Centered Ex pansions)’; ‘Expansions that Support the Priest Centered Message’; ‘Considerations of the Priest’s Function Where No Temple Exists’; and ‘Esoteric Knowledge What Makes the Priest Special’. In her second volume Mortensen catalogues the PJu ma terial as well as the other source material in Pseudo Jonathan, thus making available to the reader all the evidence that has apparently informed her hypothesis. Yet there is no commentary in the second volume, so it is left to the reader to deter mine how the statistical data and specific passages recorded in that volume fit into Mortensen’s overall argument. The second volume also contains a section entitled ‘Selected References’. But again it is left to the reader to determine how the various works recorded in the ‘Selected References’ section relate to specific parts of Mortensen’s theory, especially given the surprising scarcity of footnotes in the first volume. Since the bulk of Mortensen’s argument is articulated in the 287

REVIEWS

first volume, most of my comments will be directed toward that portion of this book. The book manifests a number of methodological problems. Many have to do with Mortensen’s concrete engagements with her sample PJu passages. Yet one of the more significant problems becomes evident on the first page of this book. As was implied by the introductory paragraph above, one of the main distinctions in this book is between Targum Pseudo Jonathan as a whole and the Pseudo Jonathan unique material or PJu. As Mortensen states in her first sentence, ‘Targum Pseudo Jonathan projects a single overriding purpose: it speaks to priests’ (p. 1). A few sen tences later she writes, ‘This [priestly] agenda only stands out when the unique Pseudo Jonathan material is separated from the rest’ (p. 1). By distinguishing and separating the material she will later call PJu from Pseudo Jonathan as a whole, Mortensen hopes to produce ‘a clear view of the obviously focused agenda of the targumic writer’ (p. 1). Although this move may initially seem straightforward, upon inspection it can be seen to complicate and confuse Mortensen’s analysis of both Pseudo Jonathan and PJu. On the one hand she attempts to argue that Pseudo Jonathan as a whole has a priest centered agenda. Yet on the other hand she states that this agenda only becomes evident when the PJu material is examined in isolation from the other Pseudo Jonathan material. In making this move, Mortensen undermines the force of her argument before she even begins making it. In effect she admits that when taken as an integral whole, Pseudo Jonathan oc cludes the priest centered agenda that is manifest in PJu. Mortensen does not address this particular methodological problem, and gener ally seems to give methodology very little discussion. She does explain the process by which she isolated the PJu material, and also gives full chapter treatment to ‘Pseudo Jonathan’s Unique Character’ (Chapter 1). However, she does not explain how the priest centred agenda of PJu expands to include the entirety of Pseudo Jonathan. In fact in her concrete analyses of specific passages she often conflates PJu with Pseudo Jonathan as a whole as if the two have a single identity. In the con cluding chapter of volume one, ‘Implications of Unique Pseudo Jonathan’s Priest Centered Agenda’, she partially explains this conflation when she discusses the composition of Pseudo Jonathan. She states that in Pseudo Jonathan ‘scholar Cohens undertake to resurrect the profession of the priesthood in a fresh way. They translate the Pentateuch. They underline all relevant passages, repeating them where possible, and, where they find fertile ground, they insert current ideas and their own insights. As with all targums, they make no indication of interpolations or sources, allowing the entire targum to serve as the “translated word of God”. They appoint a final editor who fashions the document into the smooth text we have to day hence the single ‘targumist’ referred to above’ (p. 448). Although this hypoth esis of a single targumist has the advantage of explaining how Pseudo Jonathan might have had a priestly origin, it by no means indicates how a priest centred agenda was made manifest throughout Pseudo Jonathan as a whole. This is espe cially the case when considering the fact that much of Pseudo Jonathan, including much PJu material, has parallels in rabbinic literature. Mortensen mentions some of these parallels in several footnotes in her chapter on Esoteric Knowledge (cf. e.g. p. 401, n. 8), but nowhere does she discuss these parallels at length. The most ex tended discussion of PJu’s parallels appears in a footnote on page 435 where she states in three sentences, ‘While many expansions unique to Pseudo Jonathan ap pear in other rabbinic texts as well, the genius of the compiler lies in the selection and arrangement of this material. Since the parallels occur in a wide variety of 288

REVIEWS

documents a simple transfer of ideas does not explain PJu. The independent agenda of the targumist makes PJu a text deserving of attention as a separate phenomenon’ (p. 435, n. 1). In other words Pseudo Jonathan’s targumist has redacted these ex pansions to fit with his priest centered agenda. Mortensen’s statements concerning the redaction of the PJu material provide some explanation for what could be perceived as a significant problem with her study. Yet they do not account for the most problematic aspect of her book. In short, much of Mortensen’s evidence does not seem to confirm the hypothesis of a priestly origin and priestly audience for Pseudo Jonathan, or even PJu. In fact many of the passages Mortensen examines only appear to be priestly because Mortensen says they are so. In many of Mortensen’s examples there are no textual markers to indicate that the passages in question are directed toward priests or someone of a priestly family. Rather a priestly origin for such passages seems to be assumed sim ply because they are found in PJu. Some passages studied by Mortensen do indeed appear to have a priest centred message (cf. e.g. many of the passages in her ‘Tem ple Job’ chapter). However these passages are the minority of cases examined by this book. Mortensen could have provided a partial corrective for this problem by con ducting a survey of priestly beliefs and practices in post temple Judaism, thus dem onstrating the breadth of priestly interest and engagement. Yet given that informa tion concerning such beliefs and practices is scarce, if not non existent, Mortensen is placed in the difficult position of arguing that PJu and Pseudo Jonathan are priestly documents with no evidence to corroborate her claims other than that which is found in these targums. As a result, it is difficult to determine how Mortensen can demonstrate her hypothesis with any great level of certainty. Despite the flaws with this work, Mortensen has provided scholars with an inter esting hypothesis that deserves further exploration in another context, using a differ ent methodology. Not only does it articulate a new theory concerning the origins of Targum Pseudo Jonathan, but it also raises the methodological question of how to examine source material in conjunction with a completed document. For these rea sons this volume will be of interest to targum scholars who desire to enter into the discussion concerning the problem of Targum Pseudo Jonathan’s origins, as well as the discussion of what methodology is the most appropriate to examine this problem. doi:10.1093/jss/fgn068

SIMON ADNAMS LASAIR EDMONTON

ALEXANDER SAMELY, Forms of Rabbinic Literature and Thought: An Introduction. Oxford University Press, Oxford 2007. Pp. 279. Price: £50.00 hardback. ISBN: 978 0 19 929673 6. Alexander Samely, Professor of Jewish Thought at Manchester University, is well known for his pioneering systematic description of early rabbinic hermeneutics, Rabbinic Interpretation of Scripture in the Mishnah (Oxford 2002). His new book is again a major contribution, and, although his approach is as usual strongly philo sophical, it will be much more accessible to students and even general readers. Compared to modern non fiction writing with its explicitly delimited discourse, rabbinic texts seem to be deficient. This is frequently explained by historical acci dent, not deliberate composition. Against such an approach, S. accepts the texts as they are, and sees much in them as deliberate and meaningful. He therefore asks what these rabbinic forms of discourse mean. He intends to offer ‘a companion to the close reading of rabbinic texts’ (p. 4), ‘a basic account of all the major rabbinic literary formats’ (p. 5). 289

REVIEWS

Samely first describes the smallest self contained literary forms and the diverse ways how statements are linked to each other by mere juxtaposition, series of say ings varying a common theme, or clusters with an unclear thematic continuity. Since criteria of coherence are infinitely flexible, the mutual relationship between statements mostly remains unclear; the statements are open to connections of all sorts. Syntheses of rabbinic thought on whatever topic thus have no direct basis in the primary evidence. Although there are features of unity across rabbinic literature (parallels, names of rabbis, biblical quotations), to claim its unity is a hermeneutic and not necessary assumption, especially in the extreme form of an encompassing Oral Torah. In order to understand a text and its single terms, we have to decide whether the text presents an open or a closed set of statements, if it is exhaustive or offers only examples. Can we take document boundaries as evidence for a unified rabbinic message, as Jacob Neusner proposes, or must we assume open clusters so that in the extreme everything in rabbinic literature is connected with everything else? This is a basic hermeneutic problem when reading rabbinic texts. Many rabbinic statements have multiple thematic links and may change their primary focus with the chang ing environment; it is up to the reader how far (s)he can go in making connections. As Samely rightly emphasizes, the ambiguity of rabbinic textuality is frequently un derestimated because the interpretation of the Bavli or Rashi etc. is taken for granted (p. 62). The following chapters on the use of the Bible in rabbinic literature mainly sum marize Samely’s Rabbinic Interpretation, expanded to cover all of rabbinic literature. In a very dense analysis of the surface of the Mishnaic text, Samely then discusses the constant literary device of quoting rabbis. These speech reports and small narra tives always refer to a reality outside the text, offer some kind of story including disagreements among the rabbis. The amoraim assume that the anonymous voice or stam represents the unanimous or majority opinion since the presenter sides with it and only rarely explicitly takes sides. The dispute format where a single rabbi counters the stam, signals that it ‘must have been public before the text was created’ (p. 106). The stam seems to have a constant identity throughout the text; thus ‘that voice claims implicitly an authority greater than that of any single rabbi. As a result, the stam suggests itself as the voice of rabbinic knowledge’ (p. 107). When the anonymous voice of later documents binds together parts of the same and of differ ent documents, their united stams give the impression of a coherent halakhah. Rab binic texts have been arranged and composed into documents, but their composi tion ‘is de emphasized on every page by the very literary structures which constitute it’; the reader thus ‘becomes seduced into thinking that the documents are really not constituted as fixed texts at all’ (p. 111). Quoting rabbis depicts them as speaking and turns them all into contemporaries of the reporting voice and of each other. This becomes more complex in later texts which require an oral transmission before a saying reaches the reporting voice. The literary form thus emphasizes the oral mediation of rabbinic knowledge. Oral reci tation provides its own hermeneutic context, a certain place and time. It also de mands dividing large texts for oral use on communal occasions, the size depending on the category of text, longer for narrative texts, smaller for halakhic discussions. The oral setting also influences the ordering of mishnaic sentences, complementing pre existing clusters or series and commenting on them. A cluster like m. Ber 2 3 looks like a first draft, fitting a conversation with several speakers: ‘No one’s person constant outlook, experiences, or preferences would act as a filter for what is re 290

REVIEWS

corded’ (p. 132). Why has this been maintained in the redaction of the tractates? One might point to the inertia of existing structures in absence of an urgent incen tive to change them. But this certainly is only a possible hypothesis, as Samely con cedes. Much of the Mishnah and other halakhic texts are characterized by a large scale use of the hypothetical legal case. The reader assumes the individual halakhic cases to be compatible with each other; it is up to him to decide how they are linked to each other. Verbal articulations limit the meaning of existing practices and replace the former openness inherent in the practice; every new case developed on the basis of an earlier one also limits the meaning of the former case. But the texts ‘do not give descriptions of Jewish life, but prescriptions’ (p. 152); Jewish practice at any given time or place cannot be deduced from them. A chapter is devoted to the Talmud as conversation and repository. Principles of text arrangements tend to become weaker towards the end of a chapter: ‘This sug gests that the Gemara to some extent merely stores, rather than organizes, the infor mation it contains’ (p. 158). It is problematical to assume the unity of the docu ment ‘Babylonian Talmud’; its unity is less in the text than in the rabbinic tradition. Of particular interest is the description how quotations other than bibli cal or Mishnaic passages are usually presented as if they came from a pool of tradi tion without textual frame, thus deliberately distinguishing between the Mishnah and other material used in the Gemara. Using baraitot as commentary on the Mishnah, the Talmud suggests a unity of tradition. ‘The creators of the Gemara may have (in my view, must have) obliterated whatever textual constitution their own sources had in order to reconstitute them as a Mishnah commentary made up from quotations… a literary conceit of the highest order’ (pp. 170f). In the Gemara the Mishnah assumes a much greater complexity than in the original. Its gaps are exploited and filled in to such an extent that the reader, once used to the Gemara, sees the Mishnah no longer on its own, but as a pointer to a much fuller text, as a summary of its commentary; we thus witness a ‘reversal of roles between text and commentary’ (p. 173). A last chapter deals with ‘Hermeneutic Models of Story and History’. Here Samely describes the relationship between Midrash and biblical history and the way later events are understood as repetitions of biblical paradigms. The present is seen through biblical eyes. Biblical events are constructed as if they were first hand per sonal history; coherent accounts of the contemporary world thus become unneces sary. Every chapter is followed by a list of select further reading. Briefly annotated sample texts of rabbinic literature illustrate what the main text describes and analyses. A glossary, bibliography and indices conclude this concise and very dense description of literary forms of rabbinic literature and their deeper meaning. The book is highly innovative in its description of how rabbinic texts function as texts and constitute meaningful wholes. It describes them in a way which for the most part is accessible without knowing the texts or their specific problems, thus transferable to other sets of texts as well. As such it works very well for the outsider and the beginner. But it is even more useful for somebody who already has experi enced the frustration when trying to find connections in rabbinic texts and to un derstand how a text is structured. To such a reader Samely’s description not only explains the causes of one’s frustration but also helps to find meaning in such texts with so many open ends. The relationship between text and tradition which alone is the real context of the single works, is very well described, the much discussed 291

REVIEWS

question of the redaction of rabbinic texts receives much illumination. This impor tant book will be of great help for beginners in rabbinic studies. But it is even more recommended to everybody who already has some experience with rabbinic texts or who has even studied them for a long time. People dealing with similar texts from other cultures will also profit from its insights. doi:10.1093/jss/fgn069

GÜNTER STEMBERGER UNIVERSITY OF VIENNA

EMANUELA TREVISAN SEMI, Jacques Faitlovitch and the Jews of Ethiopia. Vallentine Michael, Edgware 2007. Pp. xvii + 204. Price: £19.99. ISBN: 978 0 85303 655 5. Despite the existence of a vast and ever growing literature on the Beta Israel (Falasha) of Ethiopia, there are still significant gaps in the scholarly literature about this group. None of these has been as significant as the lack of a proper scholarly biography of Jacques (Ya’acov) Faitlovitch. Faitlovitch, more than any other figure in the twentieth century, studied their traditions, shaped their image and promoted their cause. In many ways, the aliyah of over 75,000 ‘Ethiopian Jews’ and their descendants and relatives are the fulfillment of Faitlovitch’s dreams. It is hard to think of any scholar better qualified than Emanuela Trevisan Semi of the University of Venice to write about Faitlovitch. She has already written on Ital ian fascist anthropology about the ‘Falasha’, edited a book of Taamrat Emmanuel’s letters, and co edited several volumes of conference papers from the Society for the Study of Ethiopian Jewry. She has a broad knowledge of the sources, as well as a command of numerous relevant languages including Italian, Hebrew, French and German. The volume she has prepared follows Faitlovitch from his early years as a student of the great Orientalist and fellow Falasha phile, Joseph Halévy, through his nu merous journeys to Ethiopia. It chronicles his many years of public activity on be half of ‘his people’. It brings together the relevant published sources as well as archi val sources and newspaper and journals. Moreover, it departs sharply from the romanticized vision of Faitlovitch as the ‘Father of the Falasha’ and offers a nuanced portrait of this scholar, activist, educator, reformer, missionary and fundraiser. To her credit, Trevisan Semi does not attempt to resolve the contradictions in Faitlovitch’s character, but allows him to remain as both a romantic and an oppor tunist, a champion of the ‘Falasha’ and one of their harshest critics. To be sure even a scholar as well prepared as Trevisan Semi was presented with a number of challenges in preparing this volume. Given Faitlovitch’s many roles and the breadth of activities, a simple chronological narrative would not do justice to the many larger issues which arose during half a century of work. Most of the four teen chapters in the book explore a particular period, but go further by examining a particular theme in Faitlovitch’s life. Thus, throughout the book, but particularly in Chapter 6, Trevisan Semi stresses the anti missionary focus of much of Faitlovitch’s activity. Indeed, the missionaries were probably the first to treat the ‘Falasha’ as Jews in the broader sense of the term. However, her reference to the ‘Church Mission among the Jews’ is confusing. The relevant mission society was historically known as the ‘London Society for the Pro motion of Christianity amongst the Jews’, it later became the Church Ministry among the Jewish People. It was not the Church Missionary Society (pp. 162, 181), which was founded a decade earlier, although representatives of the CMS were present in Ethiopia in the early nineteenth century. 292

REVIEWS

In Chapter 7, Trevisan Semi examines Faitlovitch’s fourth and fifth trips to Ethiopia and his fundraising activities. She devotes considerable attention to his role as a schnorrer. (Defined in the glossary as a ‘beggar’, but clearly carrying more negative connotations in this context.) As Trevisan Semi points out, in an era of lit tle accountability it is difficult to determine how much of the money Faitlovitch raised went to support his travel and expenses and how much was directed to the welfare of the ‘Falasha’. His shabby treatment of many of the students he took out of Ethiopia (pp. 97 108) is not to his credit. Indeed, his own brother accused him of exploiting the ‘Falasha’ for his own purposes (p. 75). Chapter 10 looks at Faitlovitch’s difficult relationship with his best known stu dent, Taamrat Emmanuel, and the larger question of his lack of sensitivity to opin ions other than his own. Here, she characterizes Faitlovitch as a ‘Western Colonialist’. The term is itself rather clumsy. ‘Western’, in the period with which Trevisan Semi is concerned, would seem to be redundant when paired with ‘colonialist’. Elsewhere (pp. 59, 68, 117) her use of ‘Western’ in phrases such as ‘western Jews’ or ‘Western Judaism’ is not always clear. On p. 61, for example, ‘western jewry’ is used with reference to Jews from Turkey, Yemen and Aden! ‘Colonialist’, is also a curious term. Does she wish to say that Faitlovitch was a supporter of colonialism? His attitude to the Italian presence in Eritrea was cer tainly ambivalent, and it is hard to see him as a supporter of the Fascist conquest of Ethiopia. The attitudes he exhibited were certainly paternalistic and ethnocentric. Trevisan Semi is correct in noting that racism and colour prejudice (p. 28) underlay the attitudes of many of Failovitch’s opponents. However, Faitlovitch’s own support for the Beta Israel was rooted in his conviction that they were racially Jews and in herently different from and superior to other Ethiopians. While the translation of the volume from the original Italian is serviceable, it is far from elegant. There are a few clear misreadings in which the translation from a source language to Italian and from Italian to English have led to peculiar readings. Thus the Ge’ez text, Gadla Abraham, becomes the Death and Will of Abraham (p. 52), rather than the usual designation as the Testament of Abraham. Similarly, the statement that converts are ‘ke shapahat’, p. 155 becomes ‘like impetigo’. It must also be remembered that the readership for this book includes both schol ars of Ethiopia, many of whom have only a limited knowledge of Judaism or Jewish history; and scholars of Judaism, who while more numerous, generally have even less knowledge of Ethiopian history and culture. There are, accordingly, details in the book which will be unclear to some readers. While attempts are made to explain these background issues, they are not always successful. The four maps provided to chart Faitlovitch’s journeys are too small to be of any real help to anyone not already familiar with Ethiopia’s geography and topography. Moreover, several of the defini tions which appear in the glossary at the end of the book are problematic including: buda defined as a ‘sorcerer’ (but p. 54 as a ‘witch doctor’ and p. 55 as ‘sorcery’), frendji as a ‘foreigners’ (but cf. p. 49 where it is interpreted as meaning ‘non jews’), and the Talmud as a ‘written interpretation of the Bible’. Moreover, temqat is cer tainly a ‘ritual immersion as performed in Ethiopia’. However, the reference in the text (p. 85) to ‘a Temqat ceremony’ in January almost certainly describes the holiday of Epiphany, rather than a baptism or a ritual immersion per se. It should also be noted that Trevisan Semi repeatedly (pp. 17, 39, 49, 65) uses Amharic terms (kahen, debtera, frendji, tomare) in the singular, when the plural is intended. Trevisan Semi would have also been well served had she had a more active editor. Page 3 contains a sentence with runs on for more than 50 words. There are, moreo 293

REVIEWS

ver, numerous inconsistencies in the volume, which an editor should have caught. Mesguid (p. 17), but mesgid (p. 60), Aescoly, (p. 52), but Aeskoly, (pp. 3, 4) Halévy (passim) but on the back cover of the book Halévi. In the glossary tevet is not capitalized as are other Hebrew months and Ger/gerim are ‘convert/s’ not ‘con vert’s’. H.J. Polozky appears, (p. 3) but as readers of this journal know, the name almost always appears as Polotsky. Martin Wurmbrand is mentioned (pp. 3, 6, 49), but he is usually referred to as either Max or Mordechai in his publications. There is, moreover, no consistency in the presentation of personal names. In some cases full names are provided, in others family names with initials; in still others just last names. Thus in the first paragraph of the first page we have Nahum Slouschz and Joseph Halévy, but M. Nurock and Rabbi (no first name) Berman. If the overall tone of this review appears somewhat critical, it is not because I doubt the tremendous value of Trevisan Semi’s work; quite the opposite. Tudor Parfitt is certainly correct when he writes in the forward (xiv) that this work ‘is likely to remain the definitive work [on Faitlovitch] for some time to come.’ Every serious scholar of the Beta Israel will read and cite this work. Accordingly, it must be held to a very high standard. doi:10.1093/jss/fgn070

STEVEN KAPLAN HEBREW UNIVERSITY OF JERUSALEM

ALESSANDRO BAUSI e ALESSANDRO GORI, Tradizioni orientali del «Martirio di Areta». La prima recensione Araba e la versione Etiopica. Edizione critica e tradu zione. Presentazione di PAOLO MARRASSINI (Quaderni di Semitistica 27). Dipartimento di Linguistica, Università di Firenze, Florence 2006. Pp. xxviii + 306. Price: /50.00. ISBN: 88 901340 8 9. In 522 or 517, depending on the chronology adopted the Jewish Himyarite king Yusuf As’ar Ya†’ar, known in Arabic sources as Δu Nuwas, adopted Judaism as the state religion and began to persecute the Christian population. In the fol lowing year he besieged and captured Nagran, where Christianity was well estab lished, and killed large numbers of Christians, including, amongst a group of nota bles, the saintly Areta (Arabic Îari†, Ethiopic Îirut, the alleged head of Nagran). In response to these events, and at the request of the emperor Justin, the Ethiopian ruler Kaleb Ella AÒbeÌa, who claimed suzerainty over Saba and Himyar, led a mili tary expedition to the Yemen, defeated and killed Yusuf, and installed a Christian viceroy before returning to Ethiopia. This cluster of events represents one of the best known phases in the history of pre Islamic South Arabia and of the Ethiopian Aksumite kingdom, and the impact of them, both at the time and subsequently, is reflected in the number and variety of the sources in which they are mentioned: epigraphic material from South Arabia and Ethiopia; Byzantine authors such as Procopius; and a substantial collection of hagiographic writings in Syriac, Greek, Arabic, Ethiopic, Armenian and Georgian. Included amongst this third group is the Martyrdom of Areta, which, apart from versions in Armenian and Georgian, is known from a Greek version (Bibliotheca Hagiographica Graeca 166 (and metaphrastic revisions of this)), two Arabic versions probably made from the Greek, an Ethiopic version and a Karshuni version. The present work provides edi tions of the first Arabic recension and the Ethiopic version of the Martyrdom, while editions of the second Arabic recension and the Karshuni version are to be published in a subsequent volume. The work forms part of a larger research project ‘Le due sponde del Mar Rosso: archeologia e linguistica per la riconstruzione delle fasi culturali più antiche 294

REVIEWS

dell’area sud semitca’, which is led by Paolo Marrassini. In a short preface Professor Marrassini discusses the significance of the volume, highlighting the value of pub lishing the Ethiopic version with the Arabic and the importance of the fact that the editions of the Arabic and Ethiopic very properly, as he argues follow stand ard procedures for the production of critical editions of texts and do not, as is often the case in the editing of oriental texts, rely on the concept of a ‘base manuscript’. The introduction is the work of Alessandro Bausi and provides, after a short ac count of the events, a masterly survey of all the sources relating to the events at Nagran and of the research that has been undertaken on them. There then follow editions and translations of the two texts. The first Arabic recension of the Martyrdom, although known and discussed, had not previously been edited. The edition and translation by Alessandro Gori (pp. 19 89) is based on four manuscripts: BL Or. 5019 (eleventh cent.); Sin. Ar. 469 (thir teenth cent.), which is damaged and contains only a fifth of the text; Birmingham, Selly Oak Colleges, Mingana Chr. Ar. Add. 236 (thirteenth cent.), a fragment of two leaves that forms the immediate continuation of the text in Sin. Ar. 469; and Sin. Ar. 535 (thirteenth cent.), a copy of BL Or. 5019. The apparatus is divided into two registers (in the upper register a list of the witnesses, in the lower the vari ant readings), but there are only a few brief notes attached to the translation. An edition and Portuguese translation of the Ethiopic version of the Martydom of Areta was produced by Pereira in 1899, but the edition and translation by Alessandro Bausi published here (pp. 91 306) places study of this text on a totally different level. The ‘History of the men of Nagran and martyrdom of Saint Îirut and his companions’ forms one of the writings included in the traditional Ethiopic collection Gadla sama¨tat (‘Contendings of the Martyrs’). The Ethiopic text of the Martyrdom is clearly based on an Arabic Vorlage, more precisely on an Arabic ver sion that was related to, but certainly not identical with, the first Arabic recension. Bausi’s edition is based on eleven manuscripts, which range in date from the four teenth to the eighteenth century, and Bausi argues that the text that we have of the Martyrdom is not too far removed from the original Ethiopic composition. The apparatus to the text is divided into four registers: first, a list of the witnesses; second, extracts from other Ethiopic writings that refer to Areta and the events at Nagran, particularly extracts from the reading in the Synaxarium for 26 Îedar, the day on which the martyrdom of Areta and the men of Nagran was com memorated; third, the variant readings; and finally, an apparatus of the punctua tion in the MSS. This last item is not normally included in editions of Ethiopic texts, but can be important for the assessment of the MSS and the interpretations they offer. Bausi has attached to the translation some fairly extensive notes on the philologi cal, historical and literary problems raised by the text. He makes full reference to the other versions of the Martyrdom and to the extensive secondary literature, and for the Greek version of the Martyrdom (BHG 166) he was able to make use of the valuable new edition of the text prepared by Marina Detoraki in her Sorbonne doc toral dissertation of 2000. The number and variety of the sources that bear on the cluster of events con nected with the martyrdoms at Nagran raise a whole host of problems of interpreta tion that have attracted the attention of scholars for the last century and a half. The present work quite deliberately does not attempt to offer new solutions to all these problems, but, much more modestly, only to make available critical editions of two witnesses of the oriental tradition of the Martyrdom. In fulfilling this aim the two 295

REVIEWS

authors have set high standards and provided a very valuable basis for further study of these texts and of the issues raised by the events at Nagran. doi:10.1093/jss/fgn071

MICHAEL A. KNIBB KING’S COLLEGE LONDON

JOHN C. LAMOREAUX (translator), Theodore Abu Qurrah (Library of the Christian East 1). Brigham Young University Press, Provo, Utah 2005. Pp. xxxvii + 278. Price: $29.95 hardback. ISBN: 0 934893 00 4. Over the last few years Brigham Young University has begun an ambitious pro gramme of publishing critical editions and English translations of a number of Middle Eastern texts, Jewish, Christian and Islamic. The Islamic series currently includes seven titles, one by Ibn Sina (Avicenna), two by al Ghazali, two by Ibn Rushd (Averroes), one by Suhrawardi and one by Mulla Sadra, and the Jewish series, which so far concentrates exclusively on the medical works of Moses Maimonides, has two volumes, one consisting of some of his medical aphorisms and the other outlining his views on asthma (though it actually addresses a far wider set of medical issues than simply that disease). The Christian series also currently includes two volumes, one being a critical edi tion and translation by S.H. Griffith of the tenth century Syrian Orthodox scholar Yahya b. Ädi’s ethical treatise The Reformation of Morals and the other being this translation of almost the complete works of the eighth ninth century Chalce donian (or Greek Orthodox) bishop of Haran (a town traditionally thought of as being in northern Syria), but currently, thanks to a 1939 border adjustment, in Turkey, Theodore Abu Qurrah, by John Lamoreaux of Southern Methodist Univer sity. The first volume is part of the ‘Eastern Christian Texts’ series, which includes the text in the original language as well as a translation, and the second is the first volume of the ‘Library of the Christian East’, which provides translations for a wider readership. Given the widespread ignorance in the West today, and perhaps particularly in the English speaking world, of the fact that there is such a thing as living Middle Eastern Christianity, any attempt to make the works of significant Middle Eastern Christian thinkers more widely known is very welcome. The Chair of the Advisory Board for the whole project is Sidney Griffith of the Catholic University of America, and the editor of the Library of the Christian East series is David Taylor of the University of Oxford. The latter’s editorial introduction makes it clear that the intention of the series is quite explicitly to complement or supplement the other existing series of early and medieval Christian texts which usually concentrate exclusively on authors who wrote in either Latin or Greek, to indicate that in a Christian context ‘oriental’ does not necessarily mean ‘marginal’, ‘peripheral’ or ‘unsophisticated’ (p. viii), and thus to help ‘all those who are curious to comprehend the true diversity of the early church’ (p. vii), and to highlight the fact that since the publication was funded through the support of the U.S. Con gress and the Library of Congress this makes very clear the pertinence of the project to policy making as well as the academic community. Theodore Abu Qurrah (d.c. 829 CE) wrote in both Greek and Arabic (and in deed Syriac, though none of his works in this language are extant), and it is his role in this pivotal stage of the development of Middle Eastern Christianity, shifting from the language of the New Testament to the language of the Qur’an, which makes this collection of his works in both languages so valuable. Almost all of his works in both languages are included in this volume, with the translator indicating clearly (pp. xxvi xxvii) his rationale for the exclusion of some works, which is nor 296

REVIEWS

mally on the basis either that the work is not actually by Theodore (e.g. the Summa Theologiae Arabica), or that a good English translation has recently appeared else where (e.g. Sidney Griffith’s translation of his work defending icons), or that the work is so technical in its language that it is likely to be of interest only to real spe cialists (e.g. the Letter to David the Monophysite [in Arabic]) and his discussion (in Greek) of philosophical names. What remains, therefore, is nineteen works, five (in Part I) originally in Arabic, grouped under the title ‘Discerning the True Religion’, directed particularly against Judaism but also including reference to Magians (Zoroastrians), Samaritans, Manicheans, Marcionites, Muslims and others; six (in Part II), some originally in Arabic and some in Greek, on ‘Discerning the True Church’, mainly directed against the Armenians but also including discussion of the views of the Nestorians and the Jacobites, the main alternative Christian groupings in Haran; five (in Part III), all originally in Arabic and intended mainly for an internal (i.e. Christian) readership while also beginning to reflect the beginnings of inter religious discus sion of some of the themes, on ‘Topics in Controversial Theology’, that is, episte mology, natural theology, the Trinity, and two texts on free will; and finally two, both in Greek, linked under the title of ‘The Byzantine Legacy’, which include guidance for Christians about how to relate to other religious communities, the ‘Refutation of the Saracens’ (as reported by John the Deacon) and the Greek Frag ments (which, despite its title, is, the editor suggests, an integral composition prob ably compiled within a century of Theodore’s death). The translator’s introduction consists of a helpful six page outline of what is known (which is not a lot) of Theodore’s biography, an eight page summary of the main arguments found in the works, and ten pages giving fuller details of the manuscripts, editions, and other translations of the works included in the volume. There is a full bibliography, index of scriptural (including twenty Quranic) citations and index. The cover is beauti fully illustrated with a picture of the four evangelists from a Syriac lectionary from Mosul produced in 1499 CE, a timely reminder of the contemporary relevance of the book in the light of the death, following kidnap, of Paulos Faraj Rahbo, the Chaldean (Catholic) Archbishop of that city, in March 2008. Specialists in Christian Muslim relations will look particularly to Chapter 17, the short text where Theodore responds to questions from a Muslim enquirer about the Christian view of free will (pp. 207 8), a theme which is also discussed on pp. 196 8, Chapter 18, the Refutation of the Saracens, in which Theodore dis cusses both Christian beliefs (about God, Christ and Muhammad) and practices (such as the Eucharist and monogamy), and also pp. 238 41, where Theodore an swers the questions in quick succession of the Arabs, of an unbeliever, and of an Agarene (i.e. Muslim). There is much else of interest, however, not least as evidence of the thoroughly religiously plural context of Haran around the end of the eighth century, and it is also very interesting to see which of Theodore’s works were writ ten in each language, and therefore primarily for whom, Greek for a primarily Christian audience both inside and beyond the world of Islam, and then Arabic for a more mixed, even a multi religious, audience within the world of Islam. I have one technical quibble, namely that the arrangement of the footnotes is not quite as helpful as it might be, since although they are easily accessible at the foot of each page, they are run together, rather than each note starting on a new line of its own, and this means that an individual note is not quite as easy to locate as it might be. But this is a small quibble when put alongside the translator’s achieve ment in making available for the first time in book form in English the collected 297

REVIEWS

works of an author who, in the translator’s words, was ‘arguably one of the most creative and imaginative Christian theologians of the early Middle Ages’ (p. xii). Put another way, John of Damascus, who is commonly described in a Western con text as ‘the last of the Church Fathers’, did not represent the end of Chalcedonian Christianity, or indeed Christianity as a whole, in the Middle East. Rather it lived on, even if under radically different circumstances, especially after the transition from Umayyad rule (with Damascus as capital) to Äbbasid rule (with Baghdad as capital) in 750 CE, and the works of Theodore Abu Qurrah are powerful evidence of this, both as one of the first Christians to write in Arabic and as one of the first to undertake a sustained theological defence of Christianity against the rival claims of Islam. It is therefore excellent to have an English language collection of most of his works published in order to help make his name and works much more widely known. doi:10.1093/jss/fgn072

HUGH GODDARD UNIVERSITY OF NOTTINGHAM

GARTH FOWDEN, QuÒayr Ämra: Art and the Umayyad Elite in Late Antique Syria (The transformation of the classical heritage 36). University of California Press, Berkeley and Los Angeles, 2004. Pp. xxix + 390. Price: $55.00 hardback. ISBN: 0 520 23665 3. QuÒayr Ämra (‘the little castle of Ämra’) is located in the basalt desert known as the Balqa’, in modern Jordan, about 60 kilometres southeast of Ämman. It is a small, well preserved example of the many Romano Syrian and early Islamic quÒur (conventionally, ‘desert castles’), most of which are found on the arid inner edge of the Syrian half of the Fertile Crescent, on the border with the Syrian and Arabian desert. QuÒayr Ämra’s fame stems from the murals that adorn the interior of the bathhouse and audience hall (ruins of other structures, including a mosque and caravanserai stand nearby). The figural representations of profane subjects associ ated with Near Eastern kingship are startling exceptions to an ‘Islamic’ artistic tra dition that is usually taken to be predominantly abstract and aniconic. Scholarly discussions of the building have tended either to be art historical, and concerned with the murals and their patron, or archaeological, treating the building’s function and its wider geographical and architectural context. Fowden’s analysis is art historical. He seeks to situate this single, particularly full, example of Umayyad visual culture in its late antique, Syrian context, and to use the frescoes to understand more about late Umayyad court culture and ideology. Questions about the quÒur as a category of buildings and of their function and con text are by no means ignored, but they are subordinated to questions of patronage and artistic interpretation. The result is a timely and thought provoking contribu tion to late antique studies and Islamic art history. Fowden assembles and synthe sizes a very large amount of evidence pertinent to interpretation of QuÒayr Ämra’s frescoes. In so doing, he brings out the historical distinctiveness of the decades be fore 750, when a newly victorious Îijazi Arabian nobility ruled a world empire from the former Roman diocese of Oriens. Fowden’s primary concern is with six prominent groups of paintings within the audience hall: nudes and entertainers; the hunt; the enthroned prince; the reclining woman with children; the six kings; the bathing woman. Studies of each of these in turn form the core of the book: chapters two, three, four, six, seven and eight, respectively, are devoted to each image or set of images. The remainder of the book’s ten chapters set the publication in the context of previous investigations 298

REVIEWS

(chapter one), examine other contexts for interpreting the building and its murals (Greek culture and the archaeology of other quÒur, in chapter nine) and present some conclusions about ‘Umayyad self representation’ (chapter ten). The book piv ots around the attribution of the building to its patron, and the implications of this proposed provenance (chapter five). Much of the interpretation stands or falls on this question of patronage. Fowden is not only sure that the building is indeed late Umayyad, he is near certain that it was constructed in c. 735 by the crown prince al Walid b. Yazid, later caliph al Walid II (r. 743 4). As he acknowledges, in the absence of a full foundation inscription such an attribution will never be absolutely certain, but he builds a fairly strong case. The (by now widely recognised) termini post and ante quem are 711 12 and 750 (the fall of Roderic, king of Spain, who appears in one of the frescoes paying homage to the hall’s prince, and the Abbasid revolution, respectively). It is argued that biographical evidence makes al Walid b. Yazid the strongest candidate for patron by virtue of his association with the Balqa’, including his exile there during the caliphate of his uncle Hisham, and his reputation for cultural sophistication and hedonism (the other two main candidates are al Walid’s father, Yazid b. Äbd al Malik, and his uncle and namesake al Walid b. Äbd al Malik). Fowden proposes the beginning of al Walid b. Yazid’s exile from Hisham’s court in c. 735 8, when he is known to have resided in the Balqa’, as the most likely moment for QuÒayr Ämra’s construction. Indeed, he argues, after Imbert, that it is possible (though very far from certain) that a very damaged inscription within QuÒayr Ämra refers to its patron as wali ähd al muslimin wa’l muslima, ‘heir apparent of the Muslim men and women’. In other places, he also notes al Walid’s reputation as a provider of services to the pil grims, and that he is said to have led the Ìajj in 735. It may be worth adding that leadership of the campaigns against Rome and of the Ìajj were two very prestigious public appointments, frequently given to the crown princes or to candidates for the succession, and that the only qaÒr that can definitely be linked to an heir apparent (by a clear inscription) is QaÒr Burqu’, also on a route between Syria and Arabia. This was developed by al Walid b. Äbd al Malik in 700 1, the first year in five that he had not campaigned against the Romans and the year in which his junior brother (and thus rival), Sulayman, led the Ìajj; QuÒayr Ämra may fit into a pat tern of Marwanid succession politics, where the outcomes of heirs’ (and potential heirs’) contests depended in part upon a physical presence among the Syrian tribes and on the Ìajj route. Although Fowden claims that the written evidence is brought to test the inde pendently formulated interpretation of the visual material, he occasionally appears in fact to have been led by the later Arabic literary sources. He acknowledges that his interpretation of the image of the reclining woman with two children as Umm al Îakam, the concubine mother of al Walid II’s nominated heir (chapter six), and the large bathing nude woman as Shah i Afrid, the mother of his rival, Yazid b. al Walid (chapter eight), are ‘speculative compared [to the interpretation of other im ages in the hall]’ (p. 250). In each case, anecdotes about Hisham’s court (r. 724 43) from al Mada’ini (d. c. 830 50) and others of his generation, preserved only in late ninth and tenth century compilations, are the basis of the reading. Furthermore, as Fowden notes in his discussions of these particular images, for his interpretations to stand, these murals would have to have been painted some years after the pro posed foundation date of c. 735. In chapter four, interpretation genuinely seems to be led by the art history and then corroborated by the texts and as a result it is much more persuasive. Through 299

REVIEWS

a wide ranging analysis of the painting of the seated monarch, which would have appeared above the prince’s actual throne, Fowden demonstrates the importance of the Roman and Sasanian visual heritage to Umayyad expressions of their authority. As he reminds us (p. 123), ‘the prince in his alcove is, first and foremost, a reflec tion of the late Roman iconography of imperium’ (or, in the East, perhaps basileia, or mulk). In this connection, he notes parallels with Roman consular diptychs, but also those between QuÒayr Ämra’s prince and the ruler on the post Sasanian Qazwin plate, who is also depicted in a fusion of Roman and Iranian styles, which owes more to Sasanian Iran than the image at QuÒayr Ämra. He also notes evidence from pre Islamic poetry. (Here one might add al Nabigha’s pre Islamic panegyric for the Lakhmid king, al Nu¨man.) However, for Fowden, the most important precedents for such images are East Roman images of Adam and Christ as world rulers. Adam was much less problem atic than Christ in Quranic terms, and he also neatly solved the impossibility of depicting God in Islam: ‘an Umayyad prince was neatly circumventing the obstacle Islam had placed in the way of depictions of divine sanction or investiture, by al luding to an icon that was widely known in late antique Syria and depicted the common father of all mankind, whom God had created in his own image’ (p. 140). The argument works very well, free of any textual support, but Fowden adduces quite a compelling piece of textual evidence in the letter about the caliphal succes sion that al Walid II is said to have had publicly proclaimed in Iraq and Khurasan in 743, in which the Quranic verse that portrays Adam as the first khalifa fi’l ar∂, ‘the caliph on the earth’ has a central place as the model of the caliph (i.e. ‘succes sor’, or ‘deputy’). Circumstantial evidence from the extant Syrian and Arabian literary tradition suggests that Fowden may well be right to emphasize the importance of Adam (alongside David, Solomon and other biblical and Quranic patriarchs) to Umayyad self representation as ‘the ultimate symbol of legitimacy’ (p. 138). In much of the early Arabic Islamic historical tradition (eighth tenth century), as in late antique Syriac texts, such as the Book of the Cave of Treasures (probably sixth century or later), Adam’s role as the holder of the first covenant with God, and hence as the first ruler of the world is emphasised, as it is in East Roman visual evidence ad duced by Fowden. For the Arabians, as for many other late antique peoples (nota bly, the Armenians, at the other end of the Roman Iranian frontier), the idea of God’s covenant as uniting all true monotheists was central to political theory, as was the notion that divinely ordained rulership could be traced from the Creation, via great historical figures, down to their sovereignty in the present. (The parallel importance of Gayomard/Jamshid in the Iranian world, and then in Islam, should also be explored further.) It is to exactly this effect that the letter of al Walid II in vokes the Quranic Adam: the reference is to the ‘primordial covenant’ (ähd or mithaq Allah) between God and Humanity which underpins (and is in some sense coterminous with) the covenant between the Muslims and the Umayyad caliph, who is now khalifat Allah fi’l ar∂ ‘the representative of God on earth’. This ‘cov enant’ (ähd, mithaq, Ìabl, et al.) was absolutely central to all Umayyad self repre sentation in poetry and prose. For Fowden, al Walid is very much the driving personality behind the imagery at the palace complex: the composition of the images ‘suggests an original, imagina tive mind at work … likely … the patron’s rather than the artists’ (p. 254); ‘al Walid … chose Adam, the lord of creation, as the model for his portrait’ (p. 323). QuÒayr Ämra is evidence of an ‘explicit and self confident vision’ of Arabian 300

REVIEWS

monotheist monarchy (p. 324), which uses Adam (and, as he notes, also David) as visual symbols of Umayyad authority. However, elsewhere he concedes that it is also a rare chance survival, and most likely not that unusual (pp. xxv, 275f.). The prob lem here is that once again the later, highly literary portrait of the eloquent and im aginative figure of al Walid that we find in the ninth , and especially the tenth cen tury tradition appears to determine the interpretation of the images. One wonders what precedents for Umayyad use of the prophets and patriarchs we would find if more of the Umayyad Dar al Imara south of the Temple Mount in Jerusalem still stood, or if anything of Muäwiya’s al Kha∂ra’ remained in Damascus. Certainly, the use of the patriarchs as models was not new at all. Just as Adam was God’s first earthly delegate and covenantee, David’s particular value, for the Umayyads as for the Roman emperors, was in his status as God’s first dynast; Solo mon’s was as David’s successor (also, like Adam, a lord over the elements). The Umayyad poetry makes David and Solomon’s (and others’, including Jacob’s) sig nificance to the political theory of succession plain, and both the evidence of the pre Islamic poetry and now Robin’s work (2005) on the South Arabian inscriptions prove that this understanding of Hebrew/Old Testament kingship was not new to the peoples of the Arabian peninsula in the seventh century. How this Umayyad self representation was received is a second problem that ab sence of evidence makes almost unanswerable. For Fowden, al Walid II’s innovative use of Adam in image and text is ironic in that it was formulated on the eve of the fall of the Umayyads (p. 324). He argues that because ‘Quranic exegetes active un der the Umayyads, when called upon to explain the … word khalifa as applied to Adam, never with one possible exception suggested any connection between this term and the head of the Islamic state [khalifa, ‘caliph’]’, the ‘Muslim Estab lishment’ (his capital ‘e’) never accepted Umayyad legitimation of their rule (p. 140). However, in the absence of sources, the existence in the Umayyad period of a ‘Muslim Establishment’ must remain an assumption. It is an argument from silence that al Qadi (1988), who compiled the Islamic exegetical evidence on which Fowden depends, was loathe to make: the Umayyad interpretation lost out, and with it many of the arguments and interpretations of any ‘establishment’ that they had patronised; indeed, the Qur’an itself suggests that the Umayyads’ reading of the tradition in this instance was not particularly weak or wilful. Although he is not very explicit about it, it would appear that Fowden’s ‘Muslim Establishment’ existed in the Iraqi garrisons of Kufa and Basra and in the old capi tal of Medina (we have to wait until page 315 for this to become apparent). This may be after all, Medina and Kufa were centres of early traditionist religious law but there is perhaps a hint of teleology here. ‘Muslim Establishment’ implies a rather homogenous and static community with shared, normative values and ideals perhaps the ninth and tenth century ülama’, who developed and preserved the common tradition and values of what would become Sunni Islam. It might be ar gued that this scholarly ‘Establishment’ was only embryonic in this period com peting with other epistemic communities and not yet ‘established’. The Umayyad caliphate fell not so much as a result of an ideological failure to persuade this con stituency (‘the Muslim Establishment’s refusal to legitimize the excessively “kingly” Umayyads’ in the face of which al Walid was ‘defiant’ [p. 138]), but because of structural weaknesses in the organisation of its tribal armies, material grievances in the populations of its frontier provinces and a fatal fragility in its very patrimonial (and Arabian) institutions of succession. The representatives of the victorious strains of Islamic tradition then wrote the only extant histories, in which the 301

REVIEWS

Umayyads were represented as despotic and impious tropes of late antique historiography throughout the monotheist world and in which they oppressed the righteous representatives of pious Islam in the province of Iraq (now the me tropolis). However, because safely overthrown, the Umayyads could also, somewhat paradoxically, become entertaining, and even didactic, literary types: hero (Masla ma); pious imam (Ümar II); strict administrator (Hisham); tragi comic poet and hedonist (al Walid II). This later image of al Walid II (with its echoes of the elitist, even somewhat antinomian, court culture of the Abbasids, where its literary form may have been shaped) has arguably sometimes over determined Fowden’s reading of the frescoes. In his use of the Arabic literary sources, Fowden builds on Oleg Grabar’s syn thetic interpretations of Umayyad material culture and later Arabic texts. Like Grabar, Fowden sees Islamic history in the context of ‘late antiquity’, notably as the continuation of monotheist empire. This perspective has proved very productive (not least in Fowden’s own Empire to Commonwealth [1993]). However, in QuÒayr Ämra, Fowden’s emphasis on the patronage and intentions of one particular prince is sometimes at the expense of bringing out the long continuities in the culture and symbolism of Near Eastern kingship that the frescoes reflect. He also perhaps ac cepts too much at face value the ninth and tenth century traditionists’ accounts of Islam’s origins and history, not only their vivid accounts of al Walid himself, but also their claims to represent a dominant ‘established’ Islam. Nonetheless, the great value of QuÒayr Ämra lies in the author’s erudition: he has extensive knowledge of the Roman and Iranian cultural traditions that contributed to the decoration of the audience hall, and draws on a very wide range of primary and secondary sources in many languages. Through these sources, he vividly brings to life the late antique, Syro Mesopotamian visual expression of what may be seen as a distinctive moment in Near Eastern court culture. Finally, it is worth noting here that Fowden publishes QuÒayr Ämra almost si multaneously with a journal article, which seeks to set the portrait of the prince at QuÒayr Ämra in the context of aniconism in the seventh and eighth century Near East, and a book, which repeats some of the text of QuÒayr Ämra alongside a dis cussion of Christian sites in the Umayyad period (G. Fowden, ‘Late antique Art in Syria and its Umayyad Evolutions’, Journal of Roman Archaeology, 17 [2004], pp. 282 304; G. Fowden and E.K. Fowden, Studies on Hellenism, Christianity and the Umayyads, Athens, 2004). The IFPO publication of the restored images, some of which Fowden used when in preparation, is also now published (Cl. Vibert Guigue, Gh. Bisheh and Fr. Imbert, Les peintures de Qusayr Ämra. Un bain omeyyade dans la bâdiya jordanienne [BAH 179], Beirut 2007). doi:10.1093/jss/fgn073

ANDREW MARSHAM UNIVERSITY OF MANCHESTER

PAUL STARKEY, Modern Arabic Literature. Edinburgh University Press, Edinburgh 2006. Pp. 222. Price: £50.00 hardback. ISBN: 978 0 7486 1291 8. This book examines modern Arabic literatures, or more precisely the development of modern Arabic literatures over the last two hundred years or so, and is both im portant and timely: important because it is a comprehensive description and is well organized; timely since recent books dealing with this issue were mainly published in the early 1990s. Paul Starkey chooses to describe the development of modern Arabic literatures from two points of view: first he examines the background of the issue, and second 302

REVIEWS

he divides the development of modern Arabic literatures from the chronological point of view. Therefore he opens with the ‘Revival’ or ‘Renewal’ (al nah∂a) of modern Arabic literatures, which as he quite rightly points out, runs from the mid dle of the nineteenth century to the end of the First World War. After discussing the background and revival of modern Arabic literatures, Starkey dedicates the three following chapters to poetry: ‘Neo classical Poetry’, ‘Romanti cism in Arabic poetry’ and ‘Poetry: the Modernist’. The next three chapters are dedicated to prose literature: ‘Prose Literature: Early Developments’, ‘Prose Litera ture: the Period of Maturity’ and ‘The Sixties Generation and Beyond’. In chapters nine and ten he deals with Arabic drama: ‘Drama: Early Experiments’ and ‘Drama: the Period of Maturity’. The last chapter is dedicated to the conclusion. It is correct to divide modern Arabic literatures (or any literature for that matter) into their various literary genres, and when it is made very clear by the author that ‘[…] The particular readership that I have had in mind has been that of the univer sity undergraduate coming to the subject for the first time…’ (p. vii), it means that he is describing the development of modern Arabic literatures according to both the chronological aspects and those of the various literary genres. The question raised because of this division is first and foremost why the author prefers not to include more literary genres such as Arabic cinema. Furthermore, dealing with the role of other media genres such as radio, television and more re cently the internet, can elaborate upon a broad spectrum of the development of modern Arabic literatures. Nonetheless, the author succeeds not only in covering most of the prominent genres but also most of the central issues appearing in the modern Arabic literatures, such as censorship and freedom of speech, the clash be tween ‘East’ and ‘West’, writing in exile, women’s and/or feminist literatures, the use of language in modern Arabic literatures, and so on. Let us examine some of these issues and begin with the question of censorship and several aspects of freedom of speech. Starkey rightly deals with this sensitive is sue as it is reflected not only in literary genres such as drama (pp. 189, 192), but indicates that the definition of freedom of speech as it is taken for granted in West ern societies, simply does not exist in the Middle East (p. 155). In addition, the author quite rightly mentions formal and informal censorship as well as govern mental censorship (pp. 190, 193, 194, 200). It is very true that the question of cen sorship and freedom of speech is connected with three taboos: religion, politics and sex, so therefore, several literary works can be noted which were and still are banned by some Arab regimes. The most prominent examples of this are Îikayat Zahra (The Story of Zahra, 1980), by the Lebanese writer Îanan al Shaykh (b. 1945), and the Sudanese writer al ™ayyib ∑aliÌ (b. 1929), in his famous novel Mawsim al Hijra ila al Shamal (Season of Migration to the North, 1966), in both cases for moral and sexual reasons. In the case of the novel Awlad Îaratina (Children of Gebelawi, 1959), by the Egyptian Nobel Prize laureate Najib MaÌfuÂ (1911 2006), it is for religious reasons. One of the most central issues in modern Arabic literatures is the ‘clash’ between ‘East’ and ‘West’, which is discussed randomly in this book, and not necessarily in a comprehensive way. Some of the novels dealing with this issue are mentioned, such as in the case of some Egyptian writers such as YaÌya Îaqqi (1905 92), in his short novel Qindil Umm Hashim (The Lamp of Umm Hashim, 1944) (p. 112) and Äbd al Îakim Qasim (1935 90), in his novel MuÌawala li'l Khuruj (An Attempt to Get Out, 1980). More unique is Mawsim al Hijra ila al Shamal by al ™ayyib ∑aliÌ or the Tunisian author al Îabib al Salimi (b. 1951), in his novel Matahat al Raml 303

REVIEWS

(The Sand’s Labyrinth, 1994), and this is the case with the Moroccan writer MuÌammad Zifzaf (1945 2001), in his novel al Mar’a wa’l Warda (The Woman and the Rose, 1972), among other Arabic literary works. Therefore, this issue is the encounter between the culture of the ‘East’ with that of the ‘West’, or in other words between the colonizer and the colonized or the oppressor and the oppressed. The corpus of Arabic literary works and particularly novels, which may be called ‘between East and West’, can be divided into two main categories: the first includes those novels that meet the definition of the Eastern hero leaving for the ‘West’ who in most cases returns to his own country. The second category includes novels in which the subject of ‘East’ and ‘West’ is employed only as a partial axis of the plot. Another important issue of modern Arabic literatures is writing outside the Arab world, or in other words writing in exile. Starkey describes the question of exile, or more precisely the phenomenon of political, cultural, economic and problems of freedom of speech, as leading some prominent poets and writers from the begin ning of the writing of modern Arabic literatures to leave their territory in the Arab world for other territories, and particularly to North and South America. These writers created a very unique literature, either in Arabic or in the local language, i.e. English, Spanish and Portuguese, that is known as Adab al Mahjar, and Jubran Khalil Jubran (1883 1931) is the most famous figure among other prominent writ ers such as Mikha’il Nuäyma (1889 1988) and Ilya Abu Ma∂i (1889 1957) who were its exponents. While most of the Mahjar writers were Lebanese and Syrians, the Arab writers who were forced to emigrate outside the Arab world during the nineteenth and the beginning of the twentieth centuries were Egyptians and Iraqis. It is worthy of note that since the second half of the twentieth century a new wave of Arab intellectuals has emigrated outside the Arab world and may be termed post Mahjar writers. They came both from the Arab Mashriq and Maghrib and mostly preferred to live in Europe. Some of them continue to write in Arabic (al ™ayyib ∑aliÌ, Îanan al Shaykh, Zakariyya Tamir, Salim Barakat, Adunis), while others preferred to write in the language of their new territory, such as French (al ™ahir b. Jallun, Asia Jabbar) or German (Rafiq al Shami). Paul Starkey uses the term ‘Western Reader’ (pp. 12, 14, 17 and so on), and the question that arises is what about other readers, such as those from Asia and Africa? It is quite natural that each of ‘us’ writes for ‘his/her’ reader, but we should bear in mind that readers from other territories are part and parcel of ‘our’ addressees, par ticularly if like the author we state that ‘[…] I hope that it may also prove of inter est to readers interested in comparative literature, as well as to those whose prime interests may not be literary, but who are in one way or another concerned with different aspects of the Middle East a region of obvious and growing importance in the world today’ (p.vii). In addition, one should be aware that most of the bibliography of this book is in English with only a few sources in Arabic, and it must be said that this is somewhat strange since the research under discussion deals with modern Arabic literatures, and especially because this book will also be very useful in the Arab world and out side it. I should mention some mistakes and errors that have occurred in this book, first and foremost repetition of the same information. This is the case on the same page (p. 110), where it is mentioned that the Egyptian journal al Fajr was founded in 1925. Other cases include the repetition of mentioning the lifetime of writers such as MaÌmud Sami al Barudi (1839 1904), on page 43 and again on page 44, and so on. One more remark regards the need to update information connected with some 304

REVIEWS

writers, such as in the case of the Syrian writer Ümar Abu Risha who was born in 1910 (p. 72), but it should also be noted that he died in 1990. This is also the case with the Egyptian writer Äbd al Îamid Juda al SaÌÌar who was born in 1913 (p. 100), but it is not mentioned that he died in 1974. The same applies to the Tunisian writer MaÌmud al Masädi who was born in 1911 (pp. 121, 194), but it is not mentioned that he died in 1981. Other errors occur in the publication dates of several literary works. This is the case with the Egyptian writer MaÌmud ™ahir Lashin’s short story ‘Îadith al Qarya’, that was published on 19 November, 1928, not in 1929 (p. 111). This is also the case in the short novel by YaÌya Îaqqi Qindil Umm Hashim that was pub lished in 1944 and not in 1946 (p. 111). Other errors occur such as place, in the case of the Palestinian poet MaÌmud Darwish who was born in al Birweh (or al Berweh) but not Barwa (p. 92), and that of the Palestinian writer Emil Îabibi who was awarded the Israel Prize for literature in 1992, not in 1972 (p. 151). Paul Starkey is very right when he points out regarding Maghribian literature ‘[…] that of nah∂a in North Africa has been seriously under studied, at least by English speaking researchers and deserves more serious consideration’ (p. 37). I have three comments to add on this issue: first, that it is characteristic not only of the nah∂a but also after it, and as a matter of fact until the present day. Second, it is not sufficiently studied not only by English speaking scholars but also by research ers in Arabic from the Mashriq, and this is the case of composing anthologies of modern Arabic literatures that are edited either in Arabic by Mashriqian authors or in English and other languages. And third, although Starkey devotes a relatively large part of his book to the Maghribian literatures (pp. 36 8, 57, 121 2, 133 5, 159 9, 193 7), his study of the Mashriq literatures still proves his own statement. Nonetheless, it should be said that the tremendous contribution of modern Moroc can literature to the North African literatures is also noteworthy since it does not occupy a marginal place compared with the Tunisian and Algerian literatures, as is incorrectly indicated in this study. In conclusion, Paul Starkey's book is very useful and well organized. Although, as he states, ‘the particular readership that I have had in mind has been that of the university undergraduate coming to the subject for the first time, I hope that it may also prove of interest to readers interested in comparative literature […]’ (p. vii). He truly achieves this goal. doi:10.1093/jss/fgn074

AMI ELAD-BOUSKILA BEIT BERL COLLEGE

305

SHORT NOTES STEVEN E. FASSBERG and A. HURVITZ, (eds), Biblical Hebrew in its Northwest Semitic Setting: Typological and Historical Perspectives (Publication of the Institute for Advanced Studies 1, The Hebrew University Jerusalem). Hebrew University Magnes Press/Eisenbrauns, Jerusalem/Winona Lake, IN. 2006. Pp. 324. Price: $45.00 hardback. ISBN: 1 57506 116 3. The papers contained in this volume arose from a meeting of an international re search group on Biblical Hebrew in its Northwest Semitic Setting convened at the Institute for Advanced Studies at the Hebrew University of Jerusalem during the 2001 2 academic year. The Preface provides the setting: a new look at the research which led to William L. Moran’s seminal paper on ‘The Hebrew Language in Its Northwest Semitic Background’, published in the Albright Festschrift edited by G. Ernest Wright (in 1961, though the article was finalized in 1958). So much new evidence has been discovered and published since then that a new evaluation was deemed necessary. The editors mention the Deir Älla text, the Tell Fekheriye bilin gual, the Yavneh Yam, Arad and Kuntillet Äjrud inscriptions and (of later date) a vast quantity of ‘Judaean Desert’ material (Dead Sea Scrolls, Bar Kosiba letters, etc.). The essays included cover a wide variety of topics. They tend to be specific, per haps preparing some of the ground for a new summary to rival Moran’s, but readers will not really find a new synthesis here. There are twenty papers. Though they are printed in alphabetical order of the name of the author, they may be grouped as follows: A. Lemaire’s contribution on ‘Hebrew and Aramaic in the First Milennium B.C.E. in the Light of Epigraphic Evidence’ corresponds directly to the agenda of updating Moran. Lemaire is a master of the epigraphic environment and in this sur vey shows how epigraphy can help in the understanding of the development of He brew. Another contribution by him, produced jointly with A. Yardeni, is focussed on the publication for the first time of new Hebrew ostraca from c. 600 BCE. Most of the other papers are directly related to the agenda of reviewing the lin guistic setting of Biblical Hebrew, but deal with very specific aspects: M. Bar Asher on the qal passive of geminate verbs; S.E. Fassberg on command sequences; W.R. Garr on paragogic nun; J. Huehnergard on the etymology of se (connection with âser); A. Hurvitz on prostration formulae; J. Joosten on the disappearance of the iterative weqatal (as a diachronic marker in Classical and Late Biblical Hebrew); M.Z. Kaddari on homonymy and polysemy in the new Modern Hebrew Lexicon of Biblical Hebrew; G. Khan on aspects of the copula; M. Mishor on the language and text of Exodus 18; A. Moshavi on object/adverbial fronting (Genesis as cor pus); A. Niccacci on the verbal system evidenced in poetry; F.H. Polak on epic for mulae; E. Qimron on pausal pataÌ; G.A. Rendsburg on Israelian Hebrew in the Song of Songs. Some of the contributions are of wider Semitic interest but have little to do with the theme. That of J. Blau is concerned with nominal inflection in Arabic. Two con cern inner Ugaritic issues, the forms and functions of the finite verb in Ugaritic nar rative verse (E.L. Greenstein) and onomastics in Ugaritic poetry (M. O’Connor). J.A. Emerton’s contribution is concerned with tracing the origins of Israelite histori cal writing. 307

SHORT NOTES

The volume reflects the vitality of research in this area and the editors are to be congratulated on a fine production which will ultimately point the way for a new Moran. doi:10.1093/jss/fgn075

JOHN F. HEALEY UNIVERSITY OF MANCHESTER

HÉLENE LOZACHMEUR, La Collection Clermont Ganneau: ostraca, épigraphes sur jarre, étiquettes de bois (Mémoires de l’Académie des Inscriptions et Belles lettres 35). Vol. 1 and vol. 2. De Boccard, Paris 2006. Pp. Vol. 1 556, Vol. 2 vii + 347. ISBN: 2 87754 164 9. There is no doubt that this work deserves a longer review. It is, however, of such a substantial size and complexity that it seems better to bring it to public notice quickly than to add further years to a delay which has already celebrated its cente nary! It would also in an academic review be extremely difficult to know where credit should be allocated for the thousands of details, since so many great scholars have had a part in this story. What is undeniable is that Hélène Lozachmeur has done scholarship an immense service for which Elephantine studies and Aramaic epigraphists will be permanently in her debt. At the end of the nineteenth century, the island of Elephantine became the object of intense archaeological research. The German concession at the site was to produce the vast treasure of the Elephantine papyri which threw unexpected light on the Jewish military colony in the service of the Achaemenid Persians. It was a typically nineteenth century perspective to give principal attention to the papyri with reli gious content and the evidence of a Jewish temple at Elephantine was sensational. Charles Clermont Ganneau (1846 1923) had been in 1878 the first scholar to rec ognize that the few papyri known by that date related to the Persian period. Already active in archaeological fieldwork (it was he who had brought to scholarly attention the equally important Moabite Inscription), Clermont Ganneau pressed for and ob tained a French concession at the site of Elephantine from 1906. In 1907 there were published statements of what Clermont Ganneau had in mind. The title ‘Jehovah à Éléphantine’ gives the clue: his hope was to find material of direct Old Testament relevance, perhaps even a copy of the Hebrew Bible dated to the fifth century BCE. There were four French campaigns between 1906 and 1911. The results were somewhat disappointing, given the high hopes: the Germans had the concession to the best bit of the site. But the excavations produced over 300 ostraca and other minor epigraphical finds, mostly bearing brief texts in Aramaic. These texts are ex tremely difficult, but the length of delay before they were all finally published (in the present volume) requires explanation. Apart from the intrinsic difficulties, the material passed through the hands of a number of great scholars. Clermont Ganneau died in 1923, not long after World War I. Sir Arthur Cowley (1861 1931) had a role in the study of the texts and eventually the material fell to be dealt with by André Dupont Sommer (1900 83). Dupont Sommer (again with an inter ruption caused by World War II) did most of the systematic work on the material and published a few individual items from the collection (which is housed in the Cabinet of the Corpus Inscriptionum Semiticarum). The items published earlier re appear in the fourth volume of the Textbook of Aramaic Documents from Ancient Egypt by B. Porten and A. Yardeni (1999), alongside over a hundred similar items in other collections. Dupont Sommer left behind extensive notes and these are the foundations of the present publication, brought to fruition by his collaborator, Hélène Lozachmeur. 308

SHORT NOTES

She has done a remarkable job, displaying her archaeological and epigraphic ex pertise in full measure. Inevitably the editorial task involved much archival research and this research is well represented in the volumes, with detailed accounts of the history of the study of the ostraca, the archives of the excavations and other files. Indeed the book contains an important collection of evidence on the history of French scholarship during the twentieth century. All of the above is covered in the Introduction, along with summary information on the genres of the ostraca, a bibliography of items already published and a reca pitulation of the work done by the different contributors named above. There is also a chapter devoted to the study of the material (ceramic) aspects of the ostraca, a modern feature which would certainly not have appeared in the 1920s! We then move to what must here be a very brief description of the edition of the texts. Chapter II is devoted to palaeography. There follows in Chapter III the full publication of the 280 items (mostly ostraca, some jar inscriptions) of the main collection. Then there are 36 belonging to subsidiary collections of unknown origin (items X1 X33 and Y1 Y3) and 10 joined fragmentary texts are re listed as a sup plement. There are also five ‘étiquettes’ written on wood. For each item, there is a full physical description and transliterations showing the readings of the different scholars who have worked on the material. Better readings have been achieved by Lozachmeur through the use of digital technology (Adobe Photoshop). There are letters, lists of names, delivery notes, writing exercises, etc. Translations are provided, though these are often insufficiently complete to estab lish the precise nature of the message or transaction. A large proportion concern foodstuffs, textiles, utensils, tools, building materials, etc. A few allude to the Sab bath and ritual impurities. Chapters V and VI contain an ‘approche prosopographique’ and brief grammati cal comments. After a short conclusion, indexes, including an index of Aramaic words, complete the first volume. The second volume is mainly devoted to palaeographic tables and photographic plates and drawings (and some digital enhancements) of the texts. These are, how ever, preceded by some coloured plates illustrating ceramic groups, drawings of ce ramics and a collection of plates of archival documents related to the earlier work which had been done by Clermont Ganneau, Cowley and Dupont Sommer. This is a magnificent publication. Both the Académie des Inscriptions et Belles lettres and the wider scholarly community of Aramaicists will be ever in debt to Hélène Lozachmeur, who has brought to splendid completion the work of many great scholars, while making her own significant (and modern) contribution to their labours. The publication is welcomed in a preface by M. Jean Leclant as a model of publication: that is what it is. doi:10.1093/jss/fgn076


KAREL JONGELING, and ROBERT M. KERR, Late Punic Epigraphy: an Introduction to the Study of Neo Punic and Latino Punic Inscriptions. Mohr Siebeck, Tübingen 2005. Pp. x + 115. Price: /19.00 paperback. ISBN: 3 16 148728 1. This small book is an impressive (and unique) introduction to Neo Punic epigra phy, designed for classroom use, and has appeared at a time when there appears to be a revival of interest in Phoenician and Punic. In recent years a Phoenician Punic grammar and dictionary have appeared from the pen of C.R. Krahmalkov (2001 and 2000), reviewed rather negatively by Jongeling and Kerr in Folia Orientalia 39 309

SHORT NOTES

(2003), 193 201 and Orientalia 71 (2002), 173 81. The best grammar remains that by J. Friedrich and W. Röllig (revised by M.G. Amadasi Guzzo) (Phönizisch Punische Grammatik3, 1999). Until now the revival was, however, largely limited to interest in Phoenician rather than the later Punic: John Gibson’s well known Text book of Syrian Semitic Inscriptions series included a Phoenician volume (1982), but Punic (which had formed part of Cooke’s Text book of North Semitic Inscriptions, 1903) was not covered. This relative neglect is now rectified: the introduction here under review was followed in 2008 by the publication of Jongeling’s comprehensive Handbook of Neo Punic Inscriptions (Mohr Siebeck). The latter gives the impression of being the foundation on which the introductory volume of 2005 was based and Jongeling’s involvement in Neo Punic research goes back at least as far as his 1984 Groningen thesis on Names in Neo Punic Inscriptions. (In addition, of course, to his part in the J. Hoftijzer and Jongeling Dictionary of the North West Semitic Inscrip tions of 1995.) It is to be noted that Late Punic Epigraphy (like the Handbook of Neo Punic In scriptions) is concerned with the subset of Punic epigraphy which is called ‘Late’ and written in a distinctive script of cursive origin (called ‘Neo Punic’). There are also many Late Punic inscriptions in Latin and Greek script and a selection is in cluded. The time limitation is that the inscriptions all postdate 146 BCE, the date of the destruction of Carthage at the end of the Third Punic War. The inscriptions come mostly from North Africa, though it was no doubt a special pleasure for Jongeling, who lectures on Welsh as well as Semitics, to include the sole Neo Punic inscription from north of the Alps, one from Holt in north east Wales. Both the introduction under review here and the bigger Handbook are important contributions to the study of this neglected branch of the Semitic languages. doi:10.1093/jss/fgn077


JEAN CLAUDE HAELEWYCK, Grammaire comparée des langues sémitiques: élements de phonétique, de morphologies et de syntaxe (Langues et cultures anciennes 7). Éditions Safran, Brussels 2006. Pp. 191. Price: /45.00 paperback. ISBN: 2 87457 003 6. This volume is the product of twenty years of experience in teaching Aramaic, Hebrew and Comparative Semitics. It is marked by an attractive style and clear presentation. After outlining the historical and current evidence of the Semitic lan guages (following a more or less conventional geographical framework for classifica tion, while drawing attention in footnotes to some of the controversies), the author devotes major sections to phonology, morphology and syntax. Each section is punc tuated by tables of forms which present clearly the basic evidence on which the synthesis is based. Almost inevitably, the syntax section is short, though a substan tial amount of space is devoted to the Hebrew waw consecutive. This is under standable if the starting point is Hebraic, though experience suggests that it is rather difficult to get non Hebraists animated about this problem, limited as it is to one Semitic language in its early phase! There is a selective, but satisfactory, bibli ography. Comparison with the Introduction to the Comparative Grammar of the Semitic Languages by S. Moscati and others (1964) is inevitable. Haelewyck covers syntax, which is not even attempted in Moscati and his treatment is much more up to date in relation to newer discoveries (e.g. Ebla) and changed paradigms (e.g. Central Semitic). The discussions are generally much simpler and more decisive than those 310

SHORT NOTES

in Moscati (which sometimes gives the impression of having been written by a committee). The presentation of the data is certainly excellent. Though Moscati has been translated into Arabic, it has not, so far as I can see, been translated into French. Especially because of its tables, Haelewyck’s book will be useful even for those (students) who will struggle with the French, but, sadly, an English transla tion will be desirable if an American or British student readership is to be reached. doi:10.1093/jss/fgn078


SUSANNE SCHOLZ, Introducing the Women’s Hebrew Bible (Introductions in Feminist Theory 13). Continuum, London 2007. Pp. 142. Price: £18.99 paperback. ISBN: 978 0 567 08257 2. Although there is currently an abundance of general introductions to the Hebrew Bible, there are far fewer introductions relating to specialist areas within Hebrew Bible study, including feminist criticism; and so this slim volume is a welcome ad dition to the available resources. In it, Scholz provides not only an introduction to feminist biblical criticism but a contextualization of it. First she reviews the history of feminist approaches to the Hebrew Bible, and then presents brief biographies of four contemporary feminist biblical critics who are from very different back grounds, in order to demonstrate the diversity of approach by which the field is characterized. Next, Scholz discusses the methodologies of feminist criticism historical, literary and cultural and finally she moves on to demonstrate femi nist criticism at work by discussing from a feminist perspective two issues in the Hebrew Bible that continue to be of major concern to feminist biblical scholars: sexual violence, and the ideologically driven portrayal of ‘foreign’ (i.e. non Israelite) women. There is thus an amazing range of material in the book, but what is per haps the book’s most striking aspect is its personal character. By including biogra phies of feminist scholars as well as discussing their ideas, Scholz demonstrates how feminist criticism of the Hebrew Bible is no (mere) intellectual pursuit, but is by its very nature a deeply personal enterprise which unlike other forms of criticism makes no attempt to hide its situatedness. Altogether, this is a distinctive and stimulating introduction that will be very useful for students and scholars alike. doi:10.1093/jss/fgn079


311

SUBSCRIPTIONS A subscription to Journal of Semitic Studies comprises 2 issues. Prices include postage by surface mail, or for subscribers in the USA and Canada by airfreight, or in India, Japan, Australia and New Zealand, by Air Speeded Post. Airmail rates are available on request. Annual Subscription Rate (Volume LIII, 2 issues, 2008) Institutional Print edition and site-wide online access: £156//234/US$296 Print edition only: £185//222/US$281 Site-wide online access only: £148//222/US$281 Personal Print edition and individual online access: £51//77/US$102 Please note: £ Sterling rates apply in UK, / rates apply in Europe, US$ elsewhere There may be other subscription rates available, for a complete listing please visit www.jss.oupjournals.org/ subinfo. Full prepayment, in the correct currency, is required for all orders. Orders are regarded as firm and payments are not refundable. Subscriptions are accepted and entered on a complete volume basis. Claims cannot be considered more than FOUR months after publication or date of order, whichever is later. All subscriptions in Canada are subject to GST. Subscriptions in the EU may be subject to European VAT. If registered, please supply details to avoid unnecessary charges. For subscriptions that include online versions, a proportion of the subscription price may be subject to UK VAT. Personal rate subscriptions are only available if payment is made by personal cheque or credit card and delivery is to a private address. The current year and two previous years’ issues are available from Oxford University Press. Previous volumes can be obtained from the Periodicals Service Company, 11 Main Street, Germantown, NY 12526, USA. Email: [email protected]. Tel: +1 (518) 537 4700. Fax: +1 (518) 537 5899. For further information, please contact: Journals Customer Service Department, Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, UK. Email: [email protected]. Tel (and answerphone outside normal working hours): +44 (0)1865 353907. Fax: + 44 (0)1865 353485. In the US, please contact: Journals Customer Service Department, Oxford University Press, 2001 Evans Road, Cary, NC 27513, USA. Email: [email protected]. Tel (and answerphone outside normal working hours): 800 852 7323 (toll-free in USA/Canada). Fax: 919 677 1714. In Japan, please contact: Journals Customer Services, Oxford University Press, 1-1-17-5F, Mukogaoka, Bunkyo-ku, Tokyo, 113-0023, Japan. Email: [email protected]. Tel: (03) 3813 1461. Fax: (03) 3818 1522. Methods of payment. (i) Cheque (payable to Oxford University Press, to Oxford University Press, Cashiers Office, Great Clarendon Street, Oxford OX2 6DP, UK) in GB£ Sterling (drawn on a UK bank), US$ Dollars (drawn on a US bank), or EU/ Euros. (ii) Bank transfer to Barclays Bank Plc, Oxford Group Office, Oxford (bank sort code 20-65-18) (UK), overseas only Swift code BARC GB 22 (GB£ Sterling to account no. 70299332, IBAN GB89BARC20651870299332; US$ Dollars to account no. 66014600, IBAN GB27BARC 20651866014600; EU/ Euros to account no. 78923655, IBAN GB16BARC20651878923655). (iii) Credit card (Mastercard, Visa, Switch or American Express). Journal of Semitic Studies is published two times a year in April and October by Oxford University Press, Oxford, UK. Annual subscription price is £156//234/US$296. Journal of Semitic Studies is distributed by Mercury International, 365 Blair Road, Avenel, NJ 07001, USA. Periodicals postage paid at Rahway, NJ and at additional entry points. US Postmaster: send address changes to Journal of Semitic Studies (ISSN 0022-4480), c/o Mercury International, 365 Blair Road, Avenel, NJ 07001, USA. PERMISSIONS For information on how to request permissions to reproduce articles/information from this journal, please visit www.oxfordjournals.org/permissions. ADVERTISING Inquiries about advertising should be sent to Helen Pearson, Oxford Journals Advertising, PO Box 347, Abingdon OX14 1GJ, UK. Email: [email protected]. Tel: +44 (0)1235 201904; Fax: +44 (0)8704 296864. DISCLAIMER Statements of fact and opinion in the articles in Journal of Semitic Studies are those of the respective authors and contributors and not of Journal of Semitic Studies or Oxford University Press. Neither Oxford University Press nor Journal of Semitic Studies make any representation, express or implied, in respect of the accuracy of the material in this journal and cannot accept any legal responsibility or liability for any errors or omissions that may be made. The reader should make his/her own evaluation as to the appropriateness or otherwise of any experimental technique described. © The University of Manchester 2008 All rights reserved; no part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise without prior written permission of the Publishers, or a licence permitting restricted copying issued in the UK by the Copyright Licensing Agency Ltd, 90 Tottenham Court Road, London W1P 9HE, or in the USA by the Copyright Clearance Center, 222 Rosewood Drive, Danvers, MA 01923. Typeset by Drukkerij Peeters, Herent, Belgium Printed by Bell & Bain Ltd, Glasgow, UK

JAN KEETMAN

Wechselwirkung von Vokalen und Gutturalen im Semitischen unter dem Einfluss anderer Sprachen: die Beispiele des Akkadischen und Hebräischen

1

NA’AMA PAT EL

The Development of the Semitic Definite Article: A Syntactic Approach

19

GORDON J. HAMILTON

A Proposal to Read the Legend of a Seal-Amulet from Deir Rifa, Egypt as an Early West Semitic Alphabetic Inscription

51

VINCENT DECAEN

Theme and Variation in Psalm 111: Metrical Phrase and Foot in Generative Perspective

81

ROBERT D. HOLMSTEDT

Word Order and Information Structure in Ruth and Jonah: A Generative-Typological Analysis

111

FRANÇOIS BRON

Notes sur les inscriptions néo-puniques de Henchir Medeina (Althiburos)

141

TZVI NOVICK

The Modality of Òarik in Tannaitic Hebrew

149

SHLOMY RAISKIN

Talmudic Aramaic Fauna Names: Murzema and Shaqi†na

161

NADEZHDA VIDRO

A Newly Reconstructed Karaite Work on Hebrew Grammar

169

GERRIT BOS and Y. TZVI LANGERMANN

The Introduction of Sergius of Reshäina to Galen’s Commentary on Hippocrates’ On nutriment

179

ALESSANDRA AVANZINI

Origin and Classification of the Ancient South Arabian Languages

205

AARON D. RUBIN

The Functions of the Preposition k- in Mehri

221

HILLA PELED SHAPIRA

From Conventional to Personal, or: What Happened to Metaphor under the Influence of Ideology – The case of Gha’ib Tu¨ma Farman

227

REVIEWS

251

SHORT NOTICES

307

Journal of Semitic Studies, Volume 54, Number 1, Spring 2009

Journal of Semitic Studies, Volume 54, Number 2, Autumn 2009

Journal of Semitic Studies, Volume 55, Number 1, Spring 2010

Journal of Semitic Studies, Volume 55, Number 2, Autumn 2010

Asian Music - Journal Volume 42, Number 1, Winter - Spring 2011

Journal of Biblical Literature, Vol. 128, No. 1 (Spring 2009)

JWSR Volume XV, Number 1, 2009

New Testament Studies, Volume 55, Number 1 (January 2009)

East-West Journal of Mathematics (Volume 1, Number 1)

Journal of Workplace Learning, Volume 17 Number 1 2 2005

Journal of Chiropractic Humanities: Volume 16, Issue 1, (December 2009)

Journal of Cuneiform Studies (JCS 61) 2009

Journal of Ancient Near Eastern Religions (Volume 9, Number 1, 2009)

American Theological Inquiry, Volume 2, Number 1 2009: Biannual Journal of Theology, Culture & History

Jewish Social Studies - Volume 16, Number 3, Spring Summer 2010

October, journal No.1 Spring (1976)

Textual Practice, Volume 6, Number 1 (Textual Practice Journal)

Textual Practice, Volume 6, Number 1 (Textual Practice Journal)

Cell Journal, Volume 144, Number 1, January 7, 2011

New Testament Studies, Volume 55, Number 4 (October 2009)

Contemporary Pragmatism. Volume 6, Number 1. June 2009

New Testament Studies, Volume 55, Number 2 (April 2009)

New Testament Studies, Volume 55, Number 3 (July 2009)

Cultural Studies: Volume 3, Issue 1 (Cultural Studies Journal)

New Testament Studies, Volume 54, Numbers 1-4 (2008)

American Mathematical Monthly, volume 116, number 1, january 2009

Textual Practice: Journal, Voume 8, Number 1 (Textual Practice Journal)

Газета Завтра 790 (54 2009)

Journal of Biblical Literature, Vol. 126, No. 1 (Spring 2007)

Journal of Biblical Literature, Vol. 127, No. 1 (Spring 2008)

Journal of Semitic Studies, Volume 54, Number 1, Spring 2009

Journal of Semitic Studies, Volume 54, Number 2, Autumn 2009

Journal of Semitic Studies, Volume 55, Number 1, Spring 2010

Journal of Semitic Studies, Volume 55, Number 2, Autumn 2010

Asian Music - Journal Volume 42, Number 1, Winter - Spring 2011

Journal of Biblical Literature, Vol. 128, No. 1 (Spring 2009)

JWSR Volume XV, Number 1, 2009

New Testament Studies, Volume 55, Number 1 (January 2009)

East-West Journal of Mathematics (Volume 1, Number 1)

Journal of Workplace Learning, Volume 17 Number 1 2 2005

Journal of Chiropractic Humanities: Volume 16, Issue 1, (December 2009)

Journal of Cuneiform Studies (JCS 61) 2009

Journal of Ancient Near Eastern Religions (Volume 9, Number 1, 2009)

American Theological Inquiry, Volume 2, Number 1 2009: Biannual Journal of Theology, Culture & History

Jewish Social Studies - Volume 16, Number 3, Spring Summer 2010

October, journal No.1 Spring (1976)

Textual Practice, Volume 6, Number 1 (Textual Practice Journal)

Textual Practice, Volume 6, Number 1 (Textual Practice Journal)

Cell Journal, Volume 144, Number 1, January 7, 2011

New Testament Studies, Volume 55, Number 4 (October 2009)

Contemporary Pragmatism. Volume 6, Number 1. June 2009

New Testament Studies, Volume 55, Number 2 (April 2009)

New Testament Studies, Volume 55, Number 3 (July 2009)

Cultural Studies: Volume 3, Issue 1 (Cultural Studies Journal)

New Testament Studies, Volume 54, Numbers 1-4 (2008)

American Mathematical Monthly, volume 116, number 1, january 2009

Textual Practice: Journal, Voume 8, Number 1 (Textual Practice Journal)

Газета Завтра 790 (54 2009)

Journal of Biblical Literature, Vol. 126, No. 1 (Spring 2007)

Journal of Biblical Literature, Vol. 127, No. 1 (Spring 2008)

Recommend Documents