Journal of Semitic Studies-Vol. 1-2009 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 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 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
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) 251
Aaron D. RUBIN, Studies in Semitic Grammaticalization (David STEC) 253
Jean-Jacques GLASSNER, Translated and Edited by Zainab BAHRANI and Marc
VAN DE MIEROOP, The Invention of Cuneiform: Writing in Sumer (Jon TAYLOR) 254
J.A. HALLORAN, Sumerian Lexicon: A Dictionary Guide to the Ancient Sumerian
Language (J.N. POSTGATE) 255
Stefanie U. GULDE, Der Tod als Herrscher in Ugarit und Israel (W.G.E. WATSON) 258
Hallvard HAGELIA, The Tel Dan Inscription: A Critical Investigation of Recent
Research on Its Palaeography and Philology (Bob BECKING) 259
Ernst JENNI, Studien zur Sprachwelt des Alten Testaments II (Martin F.J. BAASTEN) 261
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) 265
Bernard S. JACKSON, Wisdom Laws: A Study of the Mishpatim of Exodus
21:1 22:16 (Jeffrey STACKERT) 270
Hilary LIPKA, Sexual Transgression in the Hebrew Bible (Jenni WILLIAMS) 272
Kevin A. WILSON, The Campaign of Pharaoh Shoshenq I into Palestine
(K.A. KITCHEN) 274
Thomas B. DOZEMAN and Konrad SCHMID (eds), A Farewell to the Yahwist?
The Composition of the Pentateuch in Recent European Interpretation
(William JOHNSTONE) 276
Alice HUNT, Missing Priests: The Zadokites in Tradition and History (Deborah
ROOKE) 278
Lena-Sofia TIEMEYER, Priestly Rites and Prophetic Rage: Post-Exilic Prophetic
Critique of the Priesthood (Deborah ROOKE) 281
Nicholas P. LUNN, Word-Order Variation in Biblical Hebrew Poetry:
Differentiating Pragmatics and Poetics (Robert D. HOLMSTEDT) 283
D. GOODBLATT, Elements of Ancient Jewish Nationalism (Doron MENDELS) 285
Beverly P. MORTENSEN, The Priesthood in Targum Pseudo-Jonathan: Renewing
the Profession (Simon ADNAMS LASAIR) 287
Alexander SAMELY, Forms of Rabbinic Literature and Thought: An Introduction
(Günter STEMBERGER) 289
Emanuela TREVISAN SEMI, Jacques Faitlovitch and the Jews of Ethiopia (Steven
KAPLAN) 292
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) 294
John C. LAMOREAUX (translator), Theodore Abu Qurrah (Hugh GODDARD) 296
Garth FOWDEN, QuÒayr Ämra: Art and the Umayyad Elite in Late Antique Syria
(Andrew MARSHAM) 298
Paul STARKEY, Modern Arabic Literature (Ami ELAD-BOUSKILA) 302
short notes
Steven E. FASSBERG and A. HURVITZ (eds), Biblical Hebrew in its Northwest
Semitic Setting: Typological and Historical Perspectives (J.F. HEALEY) 307
Hélène LOZACHMEUR, La Collection Clermont-Ganneau: ostraca, épigraphes sur
jarre, étiquettes de bois (J.F. HEALEY) 308
Karel JONGELING, and ROBERT M. KERR, Late Punic Epigraphy: an Introduction
to the Study of Neo-Punic and Latino-Punic Inscriptions (J.F. HEALEY) 309
Jean-Claude HAELEWYCK, Grammaire comparée des langues sémitiques: élements
de phonétique, de morphologies et de syntaxe (J.F. HEALEY) 310
Suzanne SCHOLZ, Introducing the Women’s Hebrew Bible (Deborah ROOKE) 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.
2
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
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 formu-
liert. Unserer Meinung nach wird Hasselbachs These dadurch aber
insgesamt nicht widerlegt. Da dieser Artikel eigentlich schon abge-
schlossen 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 star-
ke Verbum als uka’’ in wie altassyrisch (aA) wäre zwar denkbar, ist
aber wegen des Schwundes des Stimmabsatzes im Akkadischen un-
wahrscheinlich. 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 pitrus-
um 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 inno-

vativ. Die assyrischen Formen in einigen Königsinschriften und Ei-
gennamen 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 babyloni-
schen Dialekt.7 Ebenso der beginnende Schwund von –ni und die
Morphologie der Verba II inf.
Nähe zum Altbabylonischen ist auch aus anderen Gründen zu er-
warten, schließlich ist das durch Textfunde zugängliche Verbrei-
tungsgebiet des Reichsakkadischen weitgehend identisch mit der spä-
teren Verbreitung des Altbabylonischen.
Bezüglich des bereits im Reichsakkadischen eingetretenen Lautwan-
dels 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.
7
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
3
Die mögliche Entwicklung a > e in Gegenwart von r und l ist alt-

babylonisch, 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 Ver-
schmelzung 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 Ra-
chen, 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 Mund-
öffnung. 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 kon-
sonantischen 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 zurückgezo-
gene 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 Mund-
haltung 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 Alt-
babylonischen 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 her-
zustellen. Die Bereinigung bei den langen Vokalen wäre dann dieser
Entwicklung gefolgt. Man kann sich dies noch in zwei Schritte auf-
geteilt 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 Aus-
weichen von älterem e vor e aus ä+ etwas plausibler.
Die babylonische Vokalharmonie wäre dann zunächst eine Anglei-
chung von a an ä gewesen, vermutlich unter dem Einfluss der
Wurzelharmonie des Sumerischen.15 Im Assyrischen, wo es diese An-
gleichung 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 sumeri-
schem Einfluss durch einen pharyngalisierten Vokal kurz vor a er-
13
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 Vokal-
harmonie 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 ety-
mologisch anzunehmenden Pharyngals sei durch sumerischen
Einfluss zu erklären, muss aber erst noch in Einklang mit der ge-
schichtlichen 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 Ur-

kunden 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 Ver-
kauf dabeisitzt. Vermutlich ein Nachbar,24 der mit seiner Anwesen-
heit 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. Die-
ser 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 so-
wohl 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 erkenn-
bar 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 un-
21
VS 25, 41 v 10; AWL 43 iii 8; 64 ii 4; Nik I 22 vii 2 passim.
22
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’.
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 Zu-
gleich ist Me-zi auch der einzige Absender, der Briefe in Akkadisch
und Sumerisch schreibt. Einen Brief fängt er mit akkadischer Ein-
leitungsformel an, schreibt den eigentlichen Brief aber auf Sume-
risch.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 Sü-
den gegenüber.
Die Beobachtungen passen ferner zu Walter Sommerfelds „vorläu-
figem 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, even-
tuell nur dünnen Schicht zu sein, das dem ebenfalls mehr archai-
schen 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 Zusammen-
hang 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 E-
colouring suggests that before being lost it underwent palatalization,
at least after a vowel“.40
Wenn E-Färbung von a nur durch Palatalisierung des Konsonan-
ten zu erklären wäre, dann müsste dies auch für äin gelten. Dies ist
mit der von Kouwenberg festgestellten Entwicklung dieses Konso-
nanten 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 die-
ser 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 be-
reits vorgegebenes System entsprechend dem Kriterium: Unterbre-
chung des Luftstromes oder keine Unterbrechung.
Die Palatalisierung von Ì steht auch im Widerspruch zum ortho-
graphischen 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 Zeichen-
gebrauch: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 begin-
nenden Lautwandel [ju-] > [u-] dokumentieren, der schon in der Ur
III-Zeit ganz vollzogen ist.43 Da dieser Wandel offenbar auch die Le-
sung des Zeichens I von ji zu i verändert, ist es möglich, dass der be-
ginnende 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 in-
terpretieren 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 hat-
te. 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 le-
xikalischen Texte mit ihrer mehr einheitlichen Schreibweise ab. Au-
ßerdem 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 Über-

einstimmung 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 Ge-
gensatz 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 Zeichen-
wahl können anzeigen, dass für semitische Muttersprachler eine
Ähnlichkeit in der Aussprache bestand. Da aber kein System mit ei-
ner eindeutigen Zuordnung von Phonem und Graphem geschaffen
wurde, sind Schlüsse nur in einer Richtung einigermaßen zuverläs-
sig: Wenn die Grapheme zwei aus der Etymologie erschlossene Pho-
neme eindeutig trennen, dann ist auch eine phonetische Unterschei-
dung 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 semiti-
scher 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 un-

ter Druck. Dies lässt sich u. a. an der Regel ablesen, dass sie nicht
verdoppelt werden dürfen und an zahlreichen Verwechslungen, ins-
besondere in den Qumran-Texten. Doch hat das Hebräische diesem
Druck bis zu einem gewissen Grad widerstanden. Dabei tritt eine ei-
genartige Erscheinung auf, das PataÌ furtivum und die Îatef-Vokale.
47
Hasselbach 2005: 143.
48
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 of-
fensichtlich, 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 Vo-
kal ist einerseits durch eine maximale Öffnung des Mundes gekenn-
zeichnet, 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 Pharynx-
muskeln 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 ge-
wohnt 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 sekundä-
rer 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 Luft-
strom von der Lunge und schwingende Stimmbänder vorhanden,
weswegen tatsächlich kurz a hörbar wird. Steht der Pharyngal hinge-
gen am Silbenanfang, so ist die Bildungsphase nicht hörbar. Das Lö-
sen 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, son-
dern 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 ähn-
lich dem Akkadischen, nur dass die im Akkadischen anzunehmende
weitere Vorverlagerung zu einem in der Mitte zwischen a und i ste-
henden 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 an-
ders 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, insbe-
sondere 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 Be-
günstigung eines Vokals nahe a bei den Pharyngalen und h wohl
ebenfalls ein. Wohl weil es mit Schwa praktisch kaum noch wahr-
nehmbar wäre, wurde auch ‫ א‬einbezogen.
Um das Auftreten von QameÒ Ìa†uf unter den Îa†ef-Vokalen in un-
serem Modell zu erklären, muss man annehmen, dass es aus Redukti-
on von u (oder o) entstanden ist. Von u nach a steigen beide Formanten
kontinuierlich an. Eine hintere Enge hebt beide Formanten und ver-
schiebt 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 zusam-

men. 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 Voka-
lismus offenbar dazu, die betreffenden Konsonanten zu erhalten.
d) Exkurs zu den Silbenzeichen
Das System der akkadischen Silbenzeichen beruhte fast ausschließ-

lich auf aus dem Sumerischen geborgten Lesungen. Diese mögen
beim Gebrauch in akkadischen Worten gegenüber der rein sumeri-
schen Aussprache modifiziert worden sein. Doch auf diesem Wege
wurde keine phonologisch oder gar phonetisch eindeutige Darstel-
lung des Akkadischen erreicht. Daher können wir nicht sicher erwar-
ten, 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 liegen-
des [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 mitt-
lerweile 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 Verände-
rung der Vokale im Akkadischen auf sumerische Sprechgewohn-
heiten zurückzuführen, ebenfalls passen würde.
52
Poebel 1931: 3.
53
Dazu ausführlich Keetman 2005.
54
Vorläufig Keetman 2005. Zwei weitere Artikel des Autors hierzu sind in Vor
bereitung, einer in WdO.
55
Smith 2007.
15
Was die Silbenzeichen betrifft, so lässt sich vermuten, dass zu-

nächst das Zeichen É (sum. „Haus“), weil es als [ä] gesprochen wur-
de,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 Zeichen-
paaren finden einige, nämlich e/ì, bé/bí, me/mi, sè/si auch in der su-
merischen Präfixharmonie Verwendung und bezeichnen dort den
Unterschied zwischen offenem e ( = [ä]) und sumerischem i.58 Man-
gels 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 rela-
tive to their head noun. Most Semitic languages show a N-Dem word
order except G., OSA and occasionally Arab. This position is in agree-
ment with the regular position of attributes in Semitic: adjectives, rela-
tive 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.: kaspam anniam ‘this silver (acc.)’
G.: ba-za hagar ‘in this city’
5
OSA: dn ’l-n-hn ‘these two gods’ (R 2923/7)
BH: hay-yom haz-ze (Gen. 17:26) ‘this day’
Ph.: 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
Ug.: 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’
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 P-
WS 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 con-
sistent 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 demonstra-

tive pronoun *ha(-n). First, we should note that *ha(-n) does not occur
as an independent demonstrative in Semitic,11 but rather is always pre-
fixed to a demonstrative as in the following:
Akk. annûm < *hanni12
Ug. hnd
BH haz-ze, hallaze
MH halla
Arb. ha∂a
Arm. hdyn, hlyn13
Amh. ¢nnäzzih (cf. bä-zzih ‘in this’)
Hasselbach (2007: 21)14 argues that the particle *ha and its exten-
sion *ha-n are originally adnominal, basing her conclusions on their
tendency to be attached to attributive demonstratives when two sepa-
rate 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. ha-∂a15 (attributive): ’arä ha∂a (Gen. 12:7) ‘this land’.
da (pronominal): da ’arä (Num. 34:2) ‘this is the land’.
Ug. hn-d and hndn are always used attributively: Ìp† hndn ‘his Ìup†u-
soldier’ (KTU 2.72).16
BH 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 exclu-
sively.
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 indica-
tions 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 demon-
strative 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 defi-
nite article in CS does not exhibit gender-number distinctions.22 This
is quite puzzling since even languages which have lost gender distinc-
tions in the nominal paradigm (e.g. Dutch23) still maintain gender dis-
18
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 defi-
nite 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 at-
tached 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 dis-
cussion, along with a reassessment of the syntactic rules and the expla-
nations 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 syn-
tactical features that are found in Canaanite, Aramaic, Arabic and pos-
sibly 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 syntacti-
cally 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 suf-
fixes, 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 pro-
nouns, 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 lan-

guages which do not exhibit the expected syntax but which seem to be
very established. The last section of the discussion deals with noun-
adjective 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 syn-
tactically definite, i.e. an adjective modifying it would not be marked
with an article and attributive demonstratives are excluded. In a regu-
lar type of annexation, whenever the rectum is marked with an article,
the annexation is syntactically definite. However, in the improper an-
nexation, 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 pat-
tern.
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 pro-
noun *∂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 indica-
tion that the use of alla∂i was originally independent of nominal defi-
niteness. This is further substantiated by comparative evidence, since
the relative construction developed much earlier than the definite arti-
cle 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 origi-
nally 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 mes-
sengers 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 defi-
nite, 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 func-
tions. 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 func-
tion 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 parti-
ciple 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-partici-
ple 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-determina-
tive *∂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 rela-
tion, 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 un-
modified (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 agree-
ment. 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 oppo-
sition 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 com-
mon 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 appear-
ance 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 distribu-
tion 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 hab-
ba’a)
Feghali (1928: 135) suggests that the pattern originated from an
older, classical, pattern which the Arab grammarians attempted to ex-
plain 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 dia-
lect, 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: ïd le-kbig ‘the great feast (Easter)’
qondart el-lexxi ‘the other shoe’
J: Òlat l¢-jdidi ‘the new synagogue’
dfater el-ëttaq ‘the old note-books’
M: 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 pro-
ductive:
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 mar-
ried’
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 Ara-
bic, 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 construc-
tion: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 (Brockel-
mann 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 conso-
nantal text, although originally no article was there. The original pat-
tern 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 com-
mon than definite N-definite Adj (Sarfatti 1989:156).52
50
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
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 mar-
ried’.53 This of course applies, then, to phrases like han-naära ham-
m¢’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 arti-
cle 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 posi-
tioned immediately before the demonstrative, took the article by anal-
ogy.
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 distinc-
tion (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 syn-
tax: 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 gender-
number 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-posi-
tive, the demonstrative in the Semitic languages consistently occupies
an attributive position after the noun and the adjective. It is unreason-
able to assume an unmotivated demonstrative movement from a noun-
phrase-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 distinc-
57
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 dis-
tinctions. 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 arti-
cle 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 ori-
gin: the demonstrative is the epitome of definite markers. Languages
which do not have an article, like Russian or Ug., use the demonstra-
tive to render a noun specifically definite. However, if *han is incon-
sistent 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 demonstra-
tives. We never find ha-Dem. as a pronominal demonstrative where a
bare demonstrative functions as an attributive. This is not a coinci-
dence.
Predicative Adjective The article distinguishes the attributive adjec-
tive 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 deter-
miners, 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 posses-
sive 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. How-
ever, 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 han-
kabiru, 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 agree-
ment 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’, qaddis-
a ‘the holy man’ etc.63 Excluding some Neo-Aramaic dialects, the de-
monstrative 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 pro-
noun is the grammatical elements they nominalize: *ha(n)- nominalizes
all elements with nominal inflection: adjectives, participles and attribu-
59
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 preposi-
tional 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
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 at-
tributive 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 appar-
ent 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 predi-
cate:
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, simi-
lar to Fr. voici que, voila que, which also shows the pattern [Presenta-
tive]-[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 attribu-
tive.78 These forms derive their attributes externally, i.e. from an exter-
nal referent. It is possible that the incentive for marking non-predica-
74
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 non-
predicative > 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, be-
cause 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 pre-
fix (< 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 dis-
cussed in Huehnergard and Pat-El (2007) exemplifies that the assign-
ment of a predicate function to a presentative can be done only through
cleaving.
The process adnominal > article may be compared to the develop-
ment 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 op-
tion 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 rela-
tive-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 com-
plete 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- be-
came 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 demonstra-
tive 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 there-
fore 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 lan-
guages show the prefix *ha- on their attributive demonstratives, Ara-
maic 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 lan-
guages are very consistent and should be assumed to go back to a com-
mon 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 at-
tached 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 origi-
nal 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 pre-
sentative, which is a non-predicative particle. The presentative, or its
reflexes han/hal, exists in all the languages. In Ug., which did not de-
velop an article, the presentative hn may be attached to relative pro-
nouns 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 fea-
ture 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. How-
ever, 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 Egyp-
tian.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 oth-
ers, the legend on UC 51354 would constitute the fourth Proto-
Canaanite 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 three-
quarters 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 Intermedi-
ate 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 un-
derstands 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 bot-
tom (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 let-
ters, with the ankh-signs at the sides,10 as that legend would have ap-
peared 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 un-
derstands the stances of the letters as they have been received or devel-
oped 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 let-
10
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 con-

sonantal 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 for-
mal 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 impres-

sion.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 writ-
ings show a more oblique angle to their downstrokes, whereas this let-
ter 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 repre-
sent a relatively small development away from the positioning of hier-
atic 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 Twelfth-
Thirteenth 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 situ-
ated 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 al-
though 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 rep-
resent a minor development away from the positioning of such an Egyp-
tian 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 Proto-
Canaanite inscriptions. The formal change from either potential Egyp-
tian 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 closely-
related 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 re-
duced 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 dam-
aged 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 fea-
tures 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 in-
cised 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 Proto-
Canaanite 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 epi-
graphs and trace the clear parts of the nûn on UC 51354 to an Egyp-
tian 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 com-
plete form of nûn, now oriented towards the right (presumably select-
ing that option rather than throwing away the already fashioned, deco-
rated, 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 incis-
ing (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 ex-
tant 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, single-
file column has parallels in three other Proto-Canaanite columnar in-
scriptions: 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 alpha-
betic text likely continues one of the varieties of placing signs on nar-
row columns of Egyptian inscriptions.37 It represents an archaic ar-
rangement of letters, one that was abandoned fairly early in the trans-
mission 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 resem-
blance 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 hori-
zontal 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 dis-
covered 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 ante-
cedent hieroglyph O42 do occur, as illustrated by a hieroglyphic writ-
ing 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 Proto-
and 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 hereto-
fore 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 de-

scending 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 exem-
plars 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 descend-
ant of a hieroglyphic form T7A to appear in a Proto-Canaanite in-
scription; 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 chrono-
logically diagnostic palaeographic indicators of when West Semitic
alphabetic writing began.54 The borrowing of this letter form also pro-
vides 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 alpha-
betic inscriptions, although the consonantal value of one may be de-
bated. 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, pri-
marily 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 prob-
able, 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 Egyp-
tian 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 arrange-
ments of its letters. It continues an archaic arrangement of the letters
in which Ìêt and zayin are written side by side in an otherwise single-
file 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 at-
tested 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 anteced-
ent; the stance of another letter, l/w, is also slightly different from ei-
ther 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 alpha-
betic 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 seal-
amulet’s discovery was so poorly recorded and published that one is
dealing with an artefact that is almost without an archaeological prov-
enance (so Quirke, personal communication). I would indicate that
unfortunate reality by leaving a very wide margin for error in estimat-
ing 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 high-
quality seal of glazed steatite and probably ‘faience’ beads would have
been acquired by a West Semite and incised with Proto-Canaanite let-
ters after the expulsion of the Semitic Hyksos from Egypt, c. 1550 BCE
(or shortly thereafter).60
56
Origins, 303 4.
57
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).
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 hori-
zontal 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 par-
allel 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 He-
61
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 trans-
mitted 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 Ara-
maic 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 lan-
guages and occupations following personal names are well attested on
seals belonging to West Semites, I shall follow Cross’s proposal for de-
ciphering 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 pre-
monarchic 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 Ara-
maic inscription occurs on the Zakkur stele, dated to just after 800
BCE.78 Millard has recently translated the relevant portion of that ‘me-
morial’ text: ‘Ba¨lshamayn [spoke] to me through seers [Ìzyn] and di-
viners’.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 pres-
ence of G participial forms of the verb *Ìzy/Ìzh in West Semitic lan-
guages 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 Egyp-
tian 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 Dy-
nasty: ‘The Deputy Treasurer, Äamu (= “the Asiatic”)’.82 James pro-
vided 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 con-
firmed 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 speak-
ing-context such as Deir Rifa. One has to suppose that either it func-
tioned 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 docu-
ments 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.
87
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.
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 Sec-
ond 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 Sur-
prisingly, 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 oppo-
site 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 nega-
tively, he did not choose to have that title translated and written in
Egyptian, or, his Semitic name rendered in Egyptian syllabic orthogra-
phy. 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, particu-
larly those living in Egypt during the late Middle Kingdom and Sec-
ond 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 im-
plies 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 ‘deter-
mined 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 He-
brew (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 de-
served 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, empha-
sis added). It is a commonplace that the verse of a tone language like
Chinese organizes tones; and that the verse of quantity-sensitive lan-
guages like ancient Sanskrit, Greek and Latin organizes by long ver-
sus 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 fur-
ther Cooper 1976),5 tentatively proposing a metrical analysis of Bib-
lical Hebrew (BH) poetry based on the phonological phrase (his ‘ac-
cented 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 pro-

posal (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 disjunc-
tives]. 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 refor-
mulate 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 recommenda-
tion of Hölscher, who for a number of methodological reasons con-
fines 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 lin-
guistically 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 pro-
grammatic 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 ex-
tended 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 presum-
ably is representative of the metrical tradition. Second, it is an acrostic,
which removes the fiddle-factors in lineation as well as possible inter-
calation of extraneous lines (of potentially different metrical struc-
ture). 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 regu-
lar poem in the seminal syllable-counting study of Culley (1970).
Furthermore, because it is an acrostic, it also receives a detailed statis-
tical analysis in Vance’s massive doctoral study (2001) as an addi-
tional check.
Culley (1970) proposes a ‘strictly descriptive approach’ to BH po-
etry. He adopts a syllable count as the better measure of the length
and the contour of the colon. He provides a detailed analysis of sev-
eral 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 ap-
proaches.
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 demon-
strated 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 distinc-
tions’) ‘permits some distinctions to be made within the corpus of
Classical Hebrew’ (p. 28): a non-trivial diagnostic tool with implica-
tions for source criticism (see further Fokkelman 2000: 9–11 on
‘normative numbers’).
Second, the several ‘significant patterns’ strongly suggest ‘that re-
strictions 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 im-
plied 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, signifi-
cantly in the context of poetic cantillation here, representative chap-
ters 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 sylla-
bles, 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 6-

9 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 letter-
counting 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 syl-
lable of the segholates substantially improves the rhythm and
strengthens the generalization regarding anapests, as explained be-
low; 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 sylla-
bles (mean, median and mode) surely demands an explanation in
terms of a syllable-organizing unit or ‘foot’ that permits minor varia-
tions. 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 de-
clares 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 un-
derlying 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 transforma-
tions of disjunctives into virtual disjunctives, made explicit in Price’s
phrase-structure analysis (1990, 1996),14 the acrostic becomes con-
sistently 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 M16 D1
’ô∂1h yahwé17 beÈol leß0ß
Significantly, the double dichotomy in (1) captures the very es-

sence 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 silluq-
trope (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
PHONOLOGICAL PHRASE (f) 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 accen-

tual parse and the underlying prosodic hierarchy of generative phonol-
ogy (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 corre-
spondence of prosodic ‘foot’ (F) in (1)–(3) to the TH ‘accent’ (vir-
tual or otherwise). The mere observation that this line instantiates a
regular iambic rhythm (Hölscher’s Silbenalternation) misses this sig-
nificant linguistic generalization regarding the global nonlinear, di-
chotomous 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 varia-
tion, 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 varia-
tion (v. 1b), we need only posit one rule of foot-syncopation (missing
foot marked Ø).
(4) I
f f
F F Ø F
s s s s s s s s
be sô´∂ ye sa rîm
´ 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 F F
s s s s s s s
ge ∂o lî¬m ma¨<a> sê´ yah wé
91
We can accordingly characterize these rule-governed metrical vari-

ants 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 syncopa-
tion. 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 syn-
copation 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 intro-
duces 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 pro-
sodic 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 (b) X
W S S
To avoid overgenerating (i.e., obtaining the reduction to just two
feet per line through syncopation in both phrases), we require a gen-
eral 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 treat-
ment of the TH Law of Transformation, whereby a disjunctive is de-
moted to its ‘virtual’ counterpart (note 13). The input in (9), with
back-to-back strong nodes, would obviously violate *CLASH; the out-
put must be similar to (8). I propose that the asymmetrical left-recur-
sive structure in (10a) captures the Janus-faced behaviour of the vir-
tual disjunctives, contrasting with the flat structure in (10b).
(9) I
fW fS
FW *FS FS
(10a) I (10b) I
f f
fW FS F W FW F S
FW FS
In favour of the proposal in (10a) is the asymmetric ternary struc-

ture 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
UW IS
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 consist-
ent 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’ semố might be expected to cre-
ate a clash (repaired by stress retraction as wenôra’ semố). This failure
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 exam-
ple 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 pre-
the 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 phras-

ing 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 tâ (Ps. 22:11)
(b) f
FW FS
s s s
h0 yû lî´ (Ps. 139:22)
(13) (a) f
fW FS
FS
s s s s <s>
kî-yaÌî´∂ 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
(b) f
fW FS
FS
s s s
yaÌs¢ßû¬ 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 projec-
tion of degenerate feet from shewa-syllables willy-nilly. The asym-
metric analysis in (14c), however, has considerable explanatory
power, able to constrain and explain the distribution of shewa-sylla-
bles 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
(c) three syllables22 (d) more than three syllables

w w
FS FW FS
FW sS sW sS sW sS
sW sS te hil la †ô´
ye sa rî¬m
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 ap-
proach.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 syncopa-
tion), is rather commonplace in iambic metre. For instance, the
clipped line can be found throughout the English tradition, begin-
ning with Chaucer among other worthies (e.g., Steele 1999: Ch. 2,
§6, pp. 84ff ).24 (Notice again the explanatory superiority of the pro-
sodic versus syntactic theories with respect to the distribution of pro-
sodic 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 re-
mit. 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 be-
low, the sigla A (the majority four-foot double dipod), B (medial
syncopation with resulting virtual disjunctive) and C (initial synco-
pation). The distribution across Psalm 111 is marked out in appen-
dix 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 prob-

lematic 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 domi-
nated 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 rule-
governed 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 fash-
ion 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 character-
ized 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 varia-
tion of syncopation licensed in and by the following three lines that
make up the opening quatrain (all four lines with 8 syllables, accord-
ing 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 metri-
cal-grid interpretation beneath (the height of the columns represent-
ing 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) FS (FW) FS
sW sS F/sW sS sW 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 ex-
plains 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-phono-
logical approach.
The non-trivial distribution of monosyllabic words, for example,
and the phenomenon of clipping cannot be captured by a syntax-
only 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 lin-
guistic 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 re-
sorting 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 coin-
cidence, on the view presented here.
Finally, why ‘twin’ Tiberian cantillation systems? Perhaps because
the simpler poetic system was designed to handle the metrically sim-
ple 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 cru-
cially 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 remain-
ing acrostic psalms and the parallel track of investigating the so-
called 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 move-
ments.]
I
1a ’ô-∂1h yah-wé33 | be-Èol le-ß0ß 3/4/A
1b be-sô¬∂ ye-sa-rî´m | we-ë-∂â¬ 3/3/B 6/7
2a ge-∂o-lî´m | ma¨-a-sê¬ y yah-wé 3/3/C
2b de-rû-sî´m | le-Èol Ìep-Òê-hém 2/3/C 5/6
3a hô∂ we-ha-∂0r | pó-ö-lô¬ 3/3/B

3b we-Òi∂-qa-†ô¬ | ö-mé-∂e† la-¨á∂ 3/4/A 6/7
4a zé-Èer äsâ¬ | le-nip-le-’o-†0w 3/4/A
4b Ìan-nû¬n we-ra-Ìû¬m| yah-wé 3/3/B 6/7
II
5a †é-rep | na-†0n lî´-re’-0w 3/4/C
5b yiz-k3r le-¨ô-l0m | be-rî-†ô¬ 3/3/B 6/7
6a k3aÌ ma-ä-s0w | hig-gî´∂ le-am-m¨ô¬ 4/4/A
6b la-†1† la-hém | na-Ìa-lá† gô-yî´m 4/4/A 8/8
7a ma-ä-sê¬ ya-∂0w | ’e-mé† û-mis-p0† 4/4/A

7b ne-’e-ma-nî´m | kol piq-qû-∂0w 2/3/C 6/7
8a se-mû-Èî´m la-¨á∂ | le-¨ô-l0m 3/3/C
8b ä-sû-yî´m | be-’e-mé† we-ya-s0r 3/3/C 6/6
III
9a pe-∂û¬† | sa-láÌ le-äm-mô¬ 3/3/C
9b Òiw-wâ¬ le-¨ô-l0m | be-rî-†ô¬ 3/3/B 6/6
9c qa-∂ô¬s we-nô-r0’| se-mô¬ 3/3/B 9/9
10a re’-sî† ÌoÈ-mâ¬| yir-’á† yah-wé 4/4/A

10b s1-Èel †ô¬ß |le-Èol ö-se-hém 3/4/A 7/8
10c te-hil-la-†ô¬ | ö-mé-∂e† la-¨á∂ 3/4/A 10 / 12
33
Here and throughout, one can read ’ãdonay instead.
103
Appendix 2: Syllable Counts for Psalm 111
Verse MT MT Culley Vance Fokkelman Loretz

Word Syllables (1970) (2001) (2003) (1979)
p. 18* p. 425** p. 369 p. 168***
1a 3 9 8 8 8 14
1b 3 8 8 8 8 13
2a 3 9 8 7 7 13
2b 2 8 +9 8 8 15
3a 334 7 7 6 6 11
3b 3 9 9 +9 8 13
4a 3 9 9 8 8 14
4b 3 8 7 7 7 13
5a 3 7 7 6 6 12
5b 3 8 8 8 8 14
6a 4 9 9 8 8 15
6b 4 9 9 8 8 14
7a 4 10 10 9 9 16
7b 2 8 8 7 7 14
8a 3 8 8 8 8 14
8b 3 9 9 9 9 13
9a 3 7 7 7 7 11
9b 3 8 8 8 8 13
9c 3 7 7 7 7 12
10a 4 9 8 8 8 17
10b 3 8 8 7 7 14
10c 3 9 9 +9 8 12
total 68 183 180 170 168 297
average 3.0 8.3 8.2 7.7 7.6 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 n
C Dn+1 C Dn
D1f D0 revia mugrash silluq

3b we-Òi∂- qa-†ô ö-me- <∂e†> la-ä∂
6b la-†e† la-hem na-<Ìa->la† gô-yîm
10c te-hilla-†ô ö-me- <∂e†> la-ä∂
D2f D1 dechi athnach
1a ’ô-∂eh yah-wé be-Èol le-ßaß
4a Ø ze- <Èer> ä-sâ¬ le-nip- le’-o-†aw
6a Ø k3aÌ ma¨-< a-> saw hig-gî∂ le-am-m¨ô
7a ma¨-< a-> sê ya-∂aw ’e-me† û-mis pa†
10b Ø se- <Èel> †ôß le-Èol ö-sé-hém
D3f D2 legarmeh revia gadol
10a re’-sî† ÌoÈ-mâ yir-’a† yah-we
(B) Medial Syncopation n
n+1 n
C Dn+1 ! Dn
D1f! D0 virtual revia mugrash! silluq

1b be-sô∂ ye-sa-rîm we-ë-∂â
4b Ìan-nûn we-ra-Ìûm yah-we
5b yiz-kor le-¨ô-lam be-rî-†ô
9c qa-∂ôs we-nô-ra’ se-mô
D2f! D1 virtual dechi! athnach
3a Ø hô∂ we-ha-∂ar pó¨-<o->lô
9b Òiw-wâ le-¨ô-lam be-rî-†ô
105
(C) Initial Syncopation n
n+1 n
Dn+1 C Dn
D1f D0 revia mugrash silluq

2b de-rû-sîm le-Èol Ìep-Òê-hém
7b ne-<’e>ma-nîm kol piq-qû-∂aw
8b ä-sû-yîm be-<’e>me† we-ya-sar
D2f D1 dechi athnach
2a ge-∂o-lîm ma¨-<a->sê yah-wé
5a Ø †e <-rep> na-†an lî-re’aw
8a se-mû-Èîm35 la-ä∂ le-¨ô-lam
D3f D2 legarmeh revia gadol
9a pe-∂û† 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)
Structural Analysis. Vol. 3, The Remaining 65 Psalms. (Studia Semitica
Neerlandica 43. Assen)
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
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 par-

tially be accounted for by analysing the formal syntactic and seman-
tic 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 be-
tween 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 informa-
tion structure can only be as accurate as the syntactic and semantic
model upon which it builds. Herein lies a current problem in Bibli-
cal Hebrew (BH) studies. While the growth of interest in informa-
tion 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 ap-
proaches); 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 typo-
logical study of word order and consider the relevant features from
Ruth and Jonah. In the second part I describe a specifically genera-
1
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 es-
say set in motion a rich comparative linguistic method with the goal
of discerning morphological and syntactic ‘universals’3 among hu-
man languages. In the first section in Greenberg’s essay he focused
on ‘certain basic factors of word order’ and proposed using three cri-
teria 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 typologi-
cal 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). Com-
plements are constituents that complete the head, and are thus ob-
ligatory for forming a larger grammatical item. For instance, transi-
2
In J.H. Greenberg (ed.), Universals of Language (Cambridge, MA 1963), 73 113.
3
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).
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 ad-
jectives, 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 ex-
hibits strong tendencies for each grammatical category and each syn-
tactic 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 Complements Adjuncts Specifiers
(≈ obligatory modifiers) (≈ optional modifiers)
Nouns destruction of the city big cities, cities in the/that/our
Africa city
Verbs destroy the city run quickly, quickly They destroyed
run cities
Adpositions in the city runs very quickly 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 or-
der of head and specifier, and both head-initial and head-final for the
order of head and adjunct, with greater weight given to the head-
final 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 typologi-
cal 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
wattibkênâ öd ki al¢tâ raätam l¢panay
‘and they wept again’ (Ruth 1:14) ‘because their evil has come up before
me’ (Jon. 1:2)
b. head-final
w¢ko tidbaqin ïm naärotay hê†eb Ìarâ li
‘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 relation-
ship (8)–(9) shows some inconsistency. Nominal heads precede their
adjuncts,6 illustrated in (8), but verbal heads, illustrated in (9), ex-
hibit both head-initial and head-final tendencies. Overall, the exam-
ples in (5)–(9) demonstrate that BH is a strongly head-initial lan-
guage, though the verb allows some variation with its non-comple-
ment 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)
hannaärâ hazzot raätam
‘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 wayhwh he†il rûaÌ g¢dolâ ’el hayyam
‘and Shaddai has done ill to me’ ‘and Yhwh cast a great wind to the sea’
(Ruth 1:21) (Jon. 1:4)
b. head-initial (verb-subject)
ki hemar sadday li m¢’od ki äl¢tâ raätam l¢panay
‘because Shaddai has embittered ‘because their evil has come up
me much’ (Ruth 1:20) 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) demon-
strate for Ruth and Jonah, linguists commonly use four criteria to
sort the matter out: clause type, frequency, distribution and prag-
matics.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 ‘sty-
listically 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 repre-
sents 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 identifi-
cation 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 de-
mands 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 dis-
cussing 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.
11
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.
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 asso-
ciated with specific clause types, such as the strong association be-
tween 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 201 Verbs
313 Finite Verbs 149 finite verbs
138 wayyiqtol 84 wayyiqtol
97 Perfective 38 Perfective
78 Imperfective 27 Imperfective
Clauses with Explicit Subjects Clauses with Explicit Subjects
47 VS clauses 44 VS clauses
19 VS wayyiqtol 35 VS wayyiqtol
26 XVS (10 modal, 3 X, 13 C) 8 XVS (5 modal, 3 C)
2 VS 1 VS
22 SV 17 SV
20 SV 13 SV
2 XSV 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.
15
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.
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 constitu-
ents (‘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 distribu-
tional 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 major-
ity of language data contains pragmatically ‘marked’, or ‘non-neu-
tral’, 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 problem-
atic 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 cri-
teria, particularly clause type and pragmatics, will be given methodo-
logical 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 criti-
cal 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 deriva-
17
Ibid., 76.
18
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.
120
tion to result in a clause with VS or SV order (hence, the theoreti-

cally 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 sur-
face 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 consid-
ering 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 exam-
ples in (18) demonstrate.
(18) German
a) Hans kaufte den Ball (S-V-O)
Hans bought the ball
b) Den Ball kaufte Hans gestern (O-V-S-ADV)
yesterday Hans bought the ball
c) Gestern kaufte Hans den Ball (ADV-V-S-O)
yesterday Hans bought the ball
The SV example in (18a) is the typical order in simple main
clauses, while the VS orders in (18b)–(18c) illustrate that a non-sub-
ject constituent preceding the verb affects the relative order of the
subject and verb. Within a constituent movement framework, exam-
ples like the German clauses in (18b)–(18c) are taken to be deriva-
tions 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 con-
stituent 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 depend-
ing 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 ex-
ample (21).27
24
T. Givón, On Understanding Grammar (New York 1979), 124 5.
25
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.
123
(21) Modern Hebrew

(a) hak¢lavim navÌu baqeÒev (S-V)
‘the dogs barked in rhythm’
(b) ’aÌar kakh navÌu hak¢lavim baqeÒe (ADV-V-S)
‘afterwards the dogs barked in rhythm’
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 con-
tain 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 under-
stood 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 in-
formation to the communicative setting.33 It is the existing informa-
tion 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 consid-
ered a Theme, while wanted to drink another cup is the rhematic in-
formation. 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, ac-
cordingly, 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 ac-

count when dealing with most clauses in any given discourse. They
are summarized in (34).36
(34)
Theme Old/known (or presupposed) information
Rheme New/added (or re-invoked) information
Focus Information contrasted with possible alternatives
Topic 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 de-
fined Theme.)
Finally, it is important to recognize that Topic is restricted to the-
matic 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 informa-
tion, the only Theme being the pronoun they, which refers back to
the compound subject, Abigail and Benjamin, from (36). Signifi-
cantly, 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 con-
cepts are heavily context-dependent, it is advisable to begin the con-
struction of an information structure model for BH by analysing dis-
crete 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 re-
quired 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 multi-
ple 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 the-
matic 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 ex-
amples 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 char-
acters 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 fi-
nite 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 lexi-
cal and grammatical items do not function as syntactic operators,
that is, they do not trigger inversion; hinnê is one of these non-op-
erators (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 al-
ready 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 con-
trast 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 bar-
ley. 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 nar-
rative. 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 con-
sider 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 none-
theless (‘C’ stands for ‘complementizer’, which is the syntactic cat-
egory 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 fol-
low 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 main-
tain 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 ab-
sence of explicit personal pronouns in the clause prohibits such a
reading. Note also that while there is no subject, example (47) pro-
vides valuable information about the complexity of the clause struc-
ture 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 con-
trast 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, result-
ing 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 ex-
ist in the second half. The pronoun does not exist in this clause, thus
such a reading is not available. Instead, Naomi, the speaker, is con-
trasting 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 be-
tween the manner in which Naomi left Israel, ‘full’, and in her opin-
ion 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 in-
version, 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 in-
version. 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-lis-
tener 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 typi-
cally 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 con-
tain 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 Focus-
fronting, 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 accu-
satory 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 juxtaposi-
tion 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 Fo-
cus-fronted constituent, there are a few examples (and numerous ex-
amples in poetic texts) of multiple fronting. We have a single occur-
rence 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 com-
plaint, 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 multiple-
fronting, 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 triple-
fronting 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 con-
trast 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.
(57)
3
0
CP Domain of Triggers
0
0
for raising V over
0
canonical S position
0
2
C TopP
0
0
0
0
0
Topic-fronted FocP
1
phrase(s)
Focus-fronted TP
phrase(s)
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 dis-
cern 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 bibli-
cal book) do not fulfil the typological criteria used to determine ba-
sic 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 po-
sition 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 exclama-
tion 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 over-
whelmingly 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 conven-
tional 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 inscrip-
tions 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 régu-
lières, 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 inscrip-
tion 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
a été mentionnée par Clermont-Ganneau dans une étude plus géné-

rale: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 pro-
posé d’interpréter comme un samech le signe qui avait été lu jus-
qu’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 Clermont-
Ganneau.13
1) l-’dn B¨l Îmn b-’ltbrs ndr ’s ndr’ ¨bdmlqrt kns bn Kns’¨n w-
2) M¨rys bn Tbrsn w-S†mn bn Yksltn w-Mshb’ bn Lylÿ w-Ggm bn Ssy¨t w-
3) 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 w-
5) 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 collè-
gues, (qui composent) le mzrÌ.
5) … sur les sanctuaires, au mois de Krr de l’année de Yll, le sacrifica-
teur, 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 ra-
cine KNS, ‘rassembler’, qui désignerait le chef du mzrÌ — mais on
attendrait l’article? On peut proposer quelques corrections de lec-
ture:
– 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 con-
duit Clermont-Ganneau à y voir plutôt ‘une nouvelle petite phrase,
où se cachent peut-être deux verbes, reliés entre eux par la conjonc-
tion waw’.16 Cependant, il ne s’est pas risqué à en donner une tra-
duction, ‘(se proposant) d’y revenir à une autre occasion’, ce qu’il n’a
jamais fait. L’intuition de Clermont-Ganneau était sans doute cor-
recte, 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-puni-
que. 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 Lam-
bèse (CIL VIII, 2650), suggérée par E. Lipinski, n’est pas plus sédui-
sante.19 Là aussi, Clermont-Ganneau avait proposé une interpréta-
tion plus sobre: Mbyw porte le titre bien connu hspr, ‘le scribe’;20 le
resh, quoique ténu, se distingue sur les estampages. La ligne se ter-
mine par le pronom relatif ’s. Au début de la ligne 7, Clermont-
Ganneau 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 lec-
ture 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 constata-
tion 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 ma-
ladroite 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-puni-
que, 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 ancien-
nes, 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 archéologi-
que de 1944.27 Mais la notoriété de cette inscription est due à un ar-
ticle 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 res-
titution:
QVOD BONVM FAV[stum, fe-
l]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.
22
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.
145
Quant au texte punique, voici la lecture qu’en propose Février:

1) ’]s ytn’ lyt ’ktrt bÿm sk¨rnm
2) ’]s ’ys lnqsmy kytb wbyt s¨t
3) ….Ì]Ò’ nk’t swmyws kns’ s¨t
4) ……]†’ S¨†rnyn’ bn S¨†r
5) …….Ì]sb s¨t Brkb¨l bn Mtn
6) ……..]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 se-
rait 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’interpréta-
tion de la première ligne. Levi Della Vida l’avait bien vu, qui, accep-
tant avec hésitation la lecture ’ktrt de Février, proposait de compren-
dre: ‘…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’or-
thographe 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 pro-
bablement s¨t. L’étymologie proposée par Février n’est guère convain-
cante, 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 rencon-
trent pour S/S¨†r.
Enfin, à la l. 5, on aurait mention d’un ‘comptable de la corpora-
tion’, Ìsb ou m]Ìsb s¨t.
Fig. 1.
32
J. B. Chabot, op.cit. à la n. 27.
33
Krahmalkov, op. cit., 476.
34
Lettre du 28 juillet 1913, citée par E. Vassel, op. cit. à la n. 26, p. 279.
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 insuf-
ficient attention, namely, the close relationship between the catego-
ries of necessity in teleological modality and weak obligation (as ex-
pressed, 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 ex-
press 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) an-
nounce 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 indi-
cate 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 assump-
tion 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 mis-
take — the fact that it is the lectio difficilior argues in its favour —
the phrase indicates, like ’ên rasa’y, impermissibility. The more com-
2
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 in-
curred 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 per-
missible (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, imme-
diately 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 exam-
ples 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 accord-
ing to the latter view (the challenge to the interpretation) but accord-
ing 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 in-
stance 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 re-
sponds: ’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
8
Thus ms. Erfurt; ms. Vienna has ’ên Òorek. See infra note 13.
9
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
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 permissibil-
ity 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 at-
tested 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’. Like-
wise, 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 in-
stance of Òôrek ‘power, capability’ occurs in Sifre Zuta Num. 5:31
(Horovitz ed., p. 239). In this pericope it is supposed that an adul-
terous 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, sup-
port the assignment of the meaning ‘permissible’ to Òarîk? There are
reasons to doubt. First, Òôrek is never attested with the meaning ‘per-
missibility’ 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 ‘permissi-

ble’, 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 conven-
tion 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 strength-
ening convention is presumably rooted in considerations of polite-
ness: 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 ver-
sion, constitutes a standard terminus in the exegesis of the school of
R. Ishmael: ‘thus you should not (’ên ‘aleyka) say (or judge) accord-
ing 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 ex-
plain ’ên Òarîk ‘not permissible’ by supposing that Òarîk is semanti-
cally 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 per-
missible 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 colloca-
tion.
The resulting pragmatic ambiguity is attested in parallel construc-
tions 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 nec-
essary’, i.e., the toddler should not have done that.
The ease with which negative Òarîk allows for pragmatic strength-
ening 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-lin-
guistically much more common with weak universal modals rather
than with strong ones.18 Many of the examples of ’ên Òarîk ‘not per-
missible’ 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 ex-
pressed 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 obliga-
tion 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 rela-
tionships 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’ de-
pendence on the father, their homes do not constitute legally distinct
domains, so that the entire courtyard counts as a single, private do-
main 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 permis-
sible’.19 However, Òarîk here can more easily be assigned its basic
sense of teleological necessity: the Tosefta is indicating that the alley-
way 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 dec-
laration 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, respec-
tively, ‘should not’ and ‘should’. The Jew who publicly desecrates the
Sabbath may, technically, make use of the legal fiction, but the ob-
servant Jews who use the eruv should not rely on this technicality,
but should instead treat him like a gentile, and enter into a rental ar-
rangement with him. Likewise, the observant Jew ‘should’ cede con-
trol, that is, not only can he do so, but he should do so, and not ei-
ther 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
Lieberman, Tosefta Ki Fshu†ah, 2.396.
20
Lieberman, Tosefta Ki Fshu†ah, 2.402.
157
the eruv, to cede control’ (t. Ëruv. 5:12).21 Thus, while the distinc-
tion 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 vine-
yards 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 stu-
dents 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 li-
cense to redeem entailed a stricture — it prohibited the sale of one’s
fruit within Jerusalem — and therefore R. Eliezer could (Òarîk) re-
deem, 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 de-
creed 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.
22
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
158
In this last case, Òarîk shades almost into the notion of permissibil-
ity. 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 stu-
dents 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 at-
tempting 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 exam-
ple, 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 distri-
bution suggests that Òarîk ‘ought’ has a colloquial flavour, as already
implied by its origin in the pragmatics of politeness.28
27
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’.
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 consump-
tion 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
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 conjec-
ture, 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 Tal-
mudic 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<rzm or räzzämä which
means ‘be tall, long’.9
Andalusi Arabic: r-z-n (sedate)

The Andalusi Arabic root r-z-n10 means ‘sedate’ (e.g.: razána is trans-
lated: ‘sedately’), and can certainly be connected with the Aramaic
bird-name shaqi†na which could indicate quiet and calm (from the
Hebrew root s-q-†).
5
A. Kohut, Arukh Completum, Vol. V (Vienna 1926), 247.
6
B. Geiger, Arukh Completum, IX, 269.
7
CAD, Vol. 1 (A), 95.
8
Niek Veldhuis, op. cit., p. 218.
9
Alexander Militarev, ‘Once more about glottochronology and comparative
method: the Omotic Afrasian case’. Aspektx komparativistiki 1 (Aspects
of comparative linguistics 1), Festschrift S. Starostin: Orientalia et Classica II (Rus
sian State University for the Humanities, Moscow 2005), 367. See also Getahun
Amare, ‘Issues of Time and Place Adverbs in Amharic’, African Languages and Cul
tures, Vol. 8, No. 2 (1995), 123 36; Edward Ullendorff, The Semitic Languages of
Ethiopia: A Comparative Phonology (London 1955), 141.
10
See Federico Corriente, A Dictionary of Andalusi Arabic (Leiden 1997), 207.
163
Hebrew: r-m-z or r-z-m (to wink, to hint)

The origin of murzema may lie in the Hebrew root r-m-z (and
through metathesis — r-z-m),11signifying hinting or winking, as in
the book of Job 15:12: ‫ַמה יּ ִָקֲּחָך לִֶבָּך וּמַה יּ ְִרז ְמוּן עֵינ ֶיָך‬, and as Rashi’s
commentary explains: (‫ירזמון – כמו ירמזון כמו שמלה שלמה )שמות כב‬,
namely — yirzemun is a metathesis of yirmezun. The Hebrew Bibli-
cal Encyclopedia, s.v. ‘‫’רזם‬, quoted this verse, noting the possibility of
metathesis, with the root r-m-z as an Aramaic loanword, while r-z-m
appears in the Lakhish letters.12
Identifying the birds murzema and shaqi†na
Ludwig Lewysohn (1819–1901), in ‘Zoologie des Talmuds’,13 con-

cludes that both murzema and shaqi†na are synonymous with the
[Greater] Flamingo, Pheonicopterus ruber. Jacob Levy’s (1819–92)
lexicon does not suggest any identification for murzema,14 but makes
reference to Lewysohn’s conclusion under the word shaqi†na.15
11
The root r m z is Arabic as well, with a similar meaning. Regarding the various
meanings of r m z see at length in C.E. Bosworth et al. (eds), The Encyclopedia of
Islam: New Edition Vol. VIII (Leiden 1995), 426 30.
12
See Menahem Zvi Kaddari, Milon Ha’ivrit Hamikra’it (Ramat Gan 2006), 998;
Wilhelm Gesenius, A Hebrew and English Lexicon of the Old Testament: with an ap
pendix containing the biblical Aramaic (Francis Brown, S.R. Driver and Charles
A. Briggs, [eds], Oxford 1906), 931, s.v. ‘‫’רזם‬. See also: W.F. Albright, ‘Notes on
Egypto Semitic Etymology’, American Journal of Semitic Languages and Literatures,
34, 2 (1918), 84. An opposite meaning exists in R. Chaim Dov Shevel’s article ‘‫פירוש‬
(‫’רבינו מיוחס בן אליהו על ספר איוב)ב‬, Hadarom, 23 (Nisan 1966), 95, item 12 and foot
note 14: ‫ ואין לו עד במקרא‬,‫"ומה ירזמון מפני מה עיניך פקוחות ומביטות בעזות‬, namely the
meaning is not ‘winking’ or ‘hinting’ eyes, but rather ‘open eyes gazing fiercely’. Eliezer
Ben Yehuda (1858 1922), in his dictionary Milon Ha’ivirt Hayeshana VeHahadasha
(N.H. Tur Sinai [ed.]), Vol. 13 (14) (Jerusalem 1952), 6517, footnote 7, refers to the
Arabic, with the meaning ‘weak' or ‘weary', and the verse’s meaning is accordingly:
‘why have your eyes grown weary of sight’. However, see tenth century commentator
Jehuda ben Koreisch, Epistola de Studii Targum Utilitate (J.J.L. Barges and D.B.
Goldberg [eds], Paris 1857), 57: ‫ומה ירזמון עיניך… יריד בה ירמזון והו רמז בלשון משנה ורמיזה‬
‫ורמיזות וכדאך הו באלערביה רמז וארמאז ותפסירה ואשארה וקד דכר מתל ומה ירמזון ומה ירזמון פי‬
…‫חרוף אלתקדים ואלתאכיר מתל כבשה וכשבה ושמלה ושלמה‬.
13
Ludwig Lewysohn, Die Zoologie des Talmuds: eine umfassende Darstellung der
rabbinischen Zoologie, unter steter Vergleichung der Forschungen alterer und neuerer
Schriftsteller (Frankfurt a. M. 1858), 170 1.
14
Jacob Levy, Wörterbuch uber die Talmudim und Midraschim, Vol. III ‫ע מ‬
(Berlin Wien 1924), 247 8.
15
Jacob Levy, ibid., Vol. IV ‫ת פ‬, p. 602, s.v. ‫ְקטנ ָא‬
ְ ‫ש‬ִ or ‫ִקיטנ ָא‬
ְ ‫ש‬ ְ . Incidentally, David
Golomb suggested that the origin of the word shaqi†na is from the Hebrew maskit
164
Marcus Jastrow (1829–1903) also translated shaqi†na16 as flamingo,

but did not do so under murzema,17 where he just stated that it is a
name of a permissible bird, long-legged and red, and finally con-
cluded that the meaning of the Talmudic statement ‫וסימנך מורזמא‬
(‘and the [mnem.] sign is: murzema’) is unclear, and thus evidently
Jastrow did not view murzama as being synonymous with shaqi†na.
No identification or etymology was proposed in Gustaf
Dalman’s18 (1855–1941) or Krupnik’s19 (1889–1972) lexicons.
Michael Sokoloff wrote that shaqi†na is ‘a name of a bird’,20 and that
murzema is ‘a permitted bird’.21 In both instances Sokoloff stated
that the etymology is uncertain.
The zoologist Israel Aharoni (1882–1946), wrote in his book,
Memoirs of a Hebrew Zoologist, the following description:
sheqi†an22 (‫שִקיטָן‬
ְ ) (Phoenicopterus ruber antiquorum). An aquatic
bird, red and with the longest legs among our land’s birds. In tractate
Îullin 63a it was called also murzema (‫מא‬ ָ ְ ‫)מוְּרז‬.23
R. Moshe-Henech Bernstein, a Shochet from London, identified
the Talmudic shaqi†na with the flamingo,24 and this opinion was
later upheld in Israel-Meir Levinger’s article in the journal Sinai re-
garding traditionally permissible birds.25 Nevertheless, R. Bernstein
did not mention murzema, and therefore it is uncertain whether or
not he thought this was synonymous with shaqi†na.
(‫)משכית‬, which seems highly unlikely (see D. Golomb, Targumna, part III: Leviticus,
Vol. II (Warsaw Jerusalem 1938), 433).
16
Marcus Jastrow, Dictionary of the Targumim, the Talmud Babli and Yerushalmi,
and the Midrashic Literature (New York Berlin 1926), 1622, s.v. ‘‫ִקיטנ ָא‬ ְ ‫ש‬ ְ ’.
17
Marcus Jastrow, ibid., p. 749, s.v. ‘‫’מוְּרז ְָמה‬, referred from ‘‫( ’ַמְרז ְָמא‬ibid., p. 840).
18
Lexikon Der Abbreviaturen (Gesamellt von G.H. Händler, ‘Anhang zum
Aramäisch neuhebräisches Wörterbuch' von Gustaf H. Dalman, Frankfurt a. M.
1897), 241, s.v. ‘‫מרזמא‬, ‫’מורזמא‬, simply left a question mark next to the word murzema,
and with regard to shaqi†na (ibid., p. 413, s.v. ‘‫ִקיטנ ָא‬
ָ ‫ש‬ ְ ’) simply notes ‘eine Vogelart’
‘a type of bird’, with no specific identification.
19
Krupnik’s lexicon just states that murzema is a type of bird, see: Baruch Krupnik,
Milon Shimushi LaTamud LaMidrash VeLatargum (A Practical Dictionary for the Tal
mud, Midrash and Targum), Vol. II (London 1927), 19, s.v. ‘‫’מוְּרז ְָמא‬.
20
Michael Sokoloff., op. cit. (footnote No. 4), p. 1174, s.v. ‘‫שקטנאה‬, ‫’שקיטנאה‬.
21
Michael Sokoloff, op cit., p. 649, s.v. ‘‫’מורזמא‬.
22
Sheqitan is the Modern Hebrew term for flamingo.
23
Israel Aharoni, Memoirs of a Hebrew Zoologist (Jerusalem 2001), 254 (Hebrew).
24
Moshe Henech Bernstein, Olam Hatushiya (World of Wisdom, London 1937),
66.
25
Yisrael Meir Levinger, ‘‫’עוף טהור נאכל במסורת‬, Sinai, Vol. 64, No. 1 6/386 91
(Tishre Adar 1969), 269.
165
murzema – al-Marzam in Arabic sources (the flamingo)

A bird named al-marzam is mentioned in the Arabic version of the
ancient book of Indian fables ‘Kalilah and Dimnah’.26 In the Hebrew
text the name is given as ‫( מרזם‬mrzm).27 François de Blois made ref-
erence to this word (i.e. mrzm) in his book Burzoy’s Voyage to India
and the Origin of the Book of Kalilah wa Dimnah, stating that ‘This
bird does not seem to have found its way into the lexica, though
these do list mirzam as the name of two stars’. According to Ibn
ManÂur’s (1232–1311) dictionary,28 Mirzam or al-Mirzam (Â“d or
Â“dL«) is the Arabic name of Bellatrix, one of the major stars in
Orion. One of the meanings of this Arabic word is ‘the roarer’ or ‘the
announcer’. Damiri, in his zoological dictionary, defines this as ‘a
kind of aquatic bird with long legs, a long neck, a crooked beak and
a black colouring on the ends of its wings. It mainly eats fish. It is
permissible to eat it’.29
An actual identification of this bird can be found in the Encyclo-
paedia of Islam, under ‘Nuham’, a word which in ancient Arabic texts
denotes the Greater Flamingo, ‘Phoenicopterus ruber roseus or anti-
quorum’:30
The term nuham, drawn from the root n-h-m, which evokes the no-
tion of growling, was given to this large graceful bird on account of its
discordant cries composed of howls and bellows. The same applies to
mirzam, another mediaeval name for the flamingo, as the root r-z-m
also contains the notion of growling.
The name mirzam was used by hunters,31 but apparently not ex-
clusively, since this name also appears in ancient Arabic poetry,
where ‘the only mention of the flamingo is found in the work of the
poet ∑afi al-Din Al-Hilli [1278–1349 – S.R.]… who calls it mirzam,32
in a list of the fourteen ‘obligatory birds’ contained in a long urdjuza
26
François de Blois, Burzoy’s Voyage to India and the Origin of the Book of Kalilah
wa Dimnah (London 1990), 12, n. 3.
27
See Ion Grant Neville Keith Falconer, Kalilah and Dimnah: or The fables of
Bidpai, being an account of their literary history (Cambridge 1885), xxxvii.
28
Muhammad ibn Mukarram Ibn ManÂur, Lisan l ärab, Vol. XII (Beirut 1955
6), 240. De Blois, ibid., makes reference to this source.
29
Kamal ad Din Muhammad ibn Musa ad Damiri, Hayatu l hayawani l kubra,
Vol. II (Bulaq 1867 8), 380; De Blois, ibid., 12, footnote 3.
30
EI, op. cit, 110.
31
Ibid., 111.
32
∑afiaddin Al Hilli, Diwan (Beirut 1962), 231. For further information regard
ing Al Hilli, see C.E. Bosworth, The Mediaeval Islamic Underworld, Part I: The Banu
Sasan in Arabic Life and Lore (Leiden 1976), 132 49.
166
of twenty-nine stanzas each… dedicated to the chaliph al-NaÒir li-

Din Allah (…1180–1225)….' 33
In conclusion, the flamingo was called al-mirzam in Arabic, from
the root r-z-m, which ‘contains the notion of growling’, which there-
fore brings us to the conclusion that the murzema (i.e. al-marzam or
al-mirzam in Arabic) is the Greater Flamingo.
33
EI, op. cit.
167

A NEWLY RECONSTRUCTED KARAITE WORK

ON HEBREW GRAMMAR
NADEZHDA VIDRO
UNIVERSITY OF CAMBRIDGE
Abstract
Karaite grammatical thought originated in the ninth or even the

eighth century CE. It had its roots in Masoretic literature and in the
Arabic tradition of grammar. Scholarly study of Karaite grammatical
tradition was given new impetus when the second Firkovitch Collec
tion held in the National Library of Russia in St Petersburg and con
taining the majority of extant Karaite linguistic manuscripts became
open to international scholars. Hitherto a number of Karaite gram
matical treatises have been edited and the origins, periodization and
characteristic features of this tradition studied. Yet further surviving
treatises must be reconstructed and grammatical concepts investi
gated in more detail before a complete account of the Karaite linguis
tic teachings can be given.
This article reconstructs from unpublished manuscripts the Karaite
grammatical treatise Kitab al üqud fi taÒarif al lugha al ïbraniyya tra
ditionally attributed to Abu al Faraj Harun. The newly reconstructed
text shows that Kitab al üqud was not composed by Abu al Faraj
Harun and must presently be considered anonymous. Additionally, it
provides evidence that the word üqud here is a previously unknown
technical term meaning ‘derivational relations, rules’ rather than
‘pearl strings’ as was suggested before.
In his article ‘An Unknown Grammatical Work by Abul-Faraj Harun’

H. Hirschfeld1 published the only hitherto identified fragment of
Kitab al-üqud fi taÒarif al-lugha al-ïbraniyya, a Karaite treatise on the
grammar of Biblical Hebrew written in Judaeo-Arabic described by
its author as a second and a particularly concise abridgment of
al-Kitab al-kafi fi al-lugha al-ïbraniyya2 by Abu al-Faraj Harun ibn
al-Faraj:
1
JQR n.s. 13 (1922 3), 1 7.
2
Edited in G. Khan, M. Á. Gallego and J. Olszowy Schlanger (eds), The Karaite
Tradition of Hebrew Grammatical Thought in its Classical Form: A Critical Edition
and English Translation of al Kitab al kafi fi al lugha al ïbraniyya by Abu al Faraj
Harun ibn al Faraj (2 vols, Leiden 2003).
169
A NEWLY RECONSTRUCTED KARAITE WORK ON HEBREW GRAMMAR
‫ מנה‬$‫תצארא ישמל אבואבה אלא אלשאד‬$‫תצרת אלכאפי פי אללגה אכ‬$‫קד כנת אכ‬
‫תצר‬$‫מנה אלכאפי וסאל סאיל אן אכ‬$‫סוא מא חצל פיה מן אלזיאדאת מא לם יתצ‬
‫מל פיה‬$ ‫כור יעקד פיה עקוד פי אלתצאריף ויג‬$‫תצר אלמד‬$‫ מן אלמכ‬$‫ר אוגז‬$‫תצר אכ‬$‫מכ‬
‫ע אליהא תסתועב פי אליסיר מן אלזמאן‬$ ‫מלא מנהא תכון אצולא ירג‬$ ‫ג‬
I had prepared an abridgment of the Sufficient Book on the Hebrew
Language which comprised all its chapters, with a few exceptions, as
well as some additions not included in the Sufficient Book. Then some-
one asked me to write another abridgment even more concise than the
one mentioned above in which the rules pertaining to conjugations
would be established and all conjugations would be brought together
as a basis to which one could refer and which could be studied in a
short time (Or. 5565 E, fol. 13r.).
The fragment discovered by Hirschfeld in the British Library
Genizah Collection consists of two leaves (folios 13 and 14 of Or.
5565 E, henceforth MS O) and contains the following parts of the
treatise: (1) the title (fol. 13r.); (2) the beginning of the introduction
(fol. 13v.); (3) the end of the chapter on words belonging to one or
more parts of speech (fol. 14r.); (4) the beginning of the chapter on
the conditions for forming morphological patterns (fol. 14r.–14v.)
Until now it was believed that the above-mentioned fragment was
all that survived of Kitab al-üqud. However, when editing an anony-
mous Karaite grammatical treatise named in the colophon al-
MukhtaÒar,3 this work being carried out as part of my Ph.D. research
at the University of Cambridge under the supervision of Prof. G.
Khan, I was able to prove its identity with Kitab al-üqud on the
grounds of the following evidence.
The most important MSS containing the text of al-MukhtaÒar are
II Firk. Evr.-Arab. I 2581 (henceforth MS A, 7 folios), and II Firk.
Evr.-Arab. I 2591 (henceforth MS B, 110 folios) preserved in the
second Firkovitch collection in the National Library of Russia in St.
Petersburg.4 These MSS clearly belong together, MS B being the
continuation of MS A. Together they contain the text of al-
MukhtaÒar from some place in the middle of the treatise to its end
(with a few lacunae). A comparison of these MSS with MS O reveals
3
This treatise was first discovered by M.N. Zislin. Cf. M.N. Zislin, ‘Eastern
School of Jewish Grammarians in X XII cent. (Résumé)’ (in Russian), in Semitskie
yazyki 2:2 (1965), 246 9, esp. 246 7. See also Geoffrey Khan, ‘The Medieval
Karaite Tradition of Hebrew Grammar’, Boletín de la Asociacion Española de
Orientalistas 38 (2002), 51 76, esp. 65; Khan et al. al Kitab al kafi, vol. 1, xxx.
4
Two other MSS I am currently aware of are MSS II Firk. Evr. Arab. I 2724
(23 folios) and II Firk. Evr. Arab. I 2594 (2 folios). They contain text copied from
several different places in the treatise.
170
that the last six lines of MS O, fol. 13v. are nearly identical with the
first seven lines of MS A, fol. 1v. Moreover, the text of MS O, fol.
14r.–14v. is contained in MS B, fol. 6r.–7r. with minor variations in
wording and orthography. These two overlaps suggest that MS O,
fol. 13–14 and MSS A and B are fragments of one and the same
grammatical work. The only obstacle to this identification is the fact
that MS O, fol. 13–14 is entitled Kitab al-üqud fi taÒarif al-lugha al-
ïbraniyya whereas the text in MSS A and B is referred to as al-
MukhtaÒar in the body of the treatise as well as in the colophon.
However, as was already mentioned, Kitab al-üqud was intended as
an abridgment of al-Kitab al-kafi so that the term al-MukhtaÒar
should probably be regarded as a description of the nature of the
treatise rather than its title.
All the above leads to the conclusion that Kitab al-üqud, which
was previously considered lost except for the fragment published by
Hirschfeld, has in fact survived almost entirely and can be recon-
structed as follows:
Title and beginning of the Introduction MS O, fol. 13r. 13v.

End of the Introduction and Chapters MS A
One to the middle of Chapter Three
Middle of Chapter Three to the end of MS B (parallel texts in II Firk. Evr. Arab.
the treatise I 2724, 2595 and MS O, fol. 14r. 14v.)
Preliminary investigation of the text of the newly reconstructed

Kitab al-üqud shows that important information pertaining to the
author, the title and the purpose of the work can be found in the
body of the treatise. Hirschfeld believed that Kitab al-üqud was
composed by Abu al-Faraj Harun,5 a prominent Karaite grammar-
ian, who was active in Jerusalem in the first half of the eleventh cen-
tury.6 Following Hirschfeld’s opinion, Kitab al-üqud is generally
included in the list of works written by Abu al-Faraj Harun.7 More-
over, the first abridgment of al-Kitab al-kafi mentioned in the intro-
5
Hirschfeld, ‘An Unknown Grammatical Work’, 2.
6
For recent accounts on Abu al Faraj Harun’s contribution to the development
of Karaite grammatical tradition, cf. for example, G. Khan, ‘Abu al Faraj Harun
and the Early Karaite Grammatical Tradition’, JJS 48 (1997), 314 34; idem. ‘The
Medieval Karaite Tradition’, 62 9; Khan et al., al Kitab al kafi, vol. 1, xi xlvi.
7
Cf. for example, Khan et al., al Kitab al kafi, vol. 1, xii; N. Basal, ‘Excerpts
from the Abridgment (al MuÌtaÒar) of al Kitab al kafi by Abu al Faraj Harun in
Arabic Script’, Israel Oriental Studies 17 (1997), 197 225, esp. 201.
171
duction to Kitab al-üqud 8 also appears on this list under the title al-
MukhtaÒar.9 However, Abu al-Faraj Harun’s authorship of these
books must now be questioned. Indeed, it was already shown10 that
the treatise contained in MS B cannot have been written by Abu al-
Faraj Harun, for this grammarian is mentioned in its text as a con-
temporary of the author:
‫ הרון בן אלפרג‬$ ‫ אבי אלפרג‬$‫א והו אלשיך‬$‫ודה פי זמאננא הד‬$ ‫כר מן כתבה מוג‬$‫וקד ד‬
…‫חרסה אללה‬
Someone, whose books are present in our time, namely the master
Abu al-Faraj Harun ibn al-Faraj, may God protect him, mentioned…
(MS B, fol. 9r.)
In as much as it has now been established that this treatise is a
part of Kitab al-üqud both al-MukhtaÒar and Kitab al-üqud must be
considered anonymous until further evidence is discovered.
Along the same lines, the generally accepted opinion on the mean-
ing of the book’s title as well as its purpose should now be reconsid-
ered. Hirschfeld translated Kitab al-üqud fi taÒarif al-lugha al-
ïbraniyya as Pearl-Strings on the Grammatical Inflections of the Hebrew
Language and thought that the intention of the author in writing this
book was to compose ‘a vade mecum in which the bare grammatical
facts were strung together’.11 However, the term üqud and its singular
form äqd are frequently used in the newly identified sections of the
book12 to describe rules concerning derivational relations between
different forms of a verb, referred to as üqud la tanÌallu, literally
‘knots that cannot be untied’ (sg. äqd la yanÌallu), e.g.:
‫לך אלאמר אבדא נקטתין‬$‫ר מלוך ד‬$‫הא גיר אצלי יכון אכ‬ֵ ‫רה‬$‫ואעלם אן כל אמר אכ‬
‫א עקד לא‬$‫ פהד‬.‫ ַרבֶה צבאך ו]צ[אה‬.‫ְדעה חכמה לנפשך‬ֶ ‫עין והמא כן‬$‫מא סוא מוצ‬
‫י פי‬$‫לך אלאמר אלד‬$‫ ועקד תֿאן הו אן אלנקטתין אלתי פי ד‬.‫ע‬$‫א אלמוצ‬$‫ינחל פי הד‬
‫ ועקד תֿאלתֿ הו‬:‫בר אליחיד אבדא אלי קאמצה‬
ַ ‫ע‬
ָ ‫הא גיר אצלי תנקלב פי‬
ֵ ‫רה חרף‬$‫אכ‬
‫ה עקוד לא‬$‫ר עבר אלמונתֿ אבדא בתו והד‬$‫רה הא גיר אצלי יכון אכ‬$‫אן כל אמר אכ‬
‫תנחל‬
8
See quotation above.
9
This al MukhtaÒar is to be distinguished from the text of the same title
discussed above.
10
Khan, ‘The Medieval Karaite Tradition’, 65; Khan et al., al Kitab al kafi, vol.
1, xxx; Zislin, ‘Eastern school’, 24; M.N. Zislin (ed.), Me’or Äyin (Moscow 1990),
esp. 17.
11
Hirschfeld, ‘An Unknown Grammatical Work’, 2. Hirschfeld’s opinion was
based on the author’s remark in the introduction ‫( יעקד פיה עקוד פי אלתצאריך‬MS O,
fol. 13v.).
12
This fact provides additional evidence in support of our identification.
172
Take note that the last vowel of each imperative ending in a non-root
heh is always Òere except in two passages, i.e. ‫( כן ְדעֶה חכמה לנפשך‬Prov.
24:14) and ‫( ַרבֶה צבאך וצאה‬Judg. 9:29). This is a rule which cannot be
dissolved regarding this matter. The second rule is that the Òere in the
end of such imperatives ending in a non-root heh always turns into a
qamaÒ in the masculine singular past verb form. The third rule is that
every feminine singular past verb form derived from an imperative
ending in a non-root heh always ends in a taw. These are rules which
can never be dissolved (MS B, fol. 18r.).
This passage shows that the Judaeo-Arabic ‫ עקד‬must indeed be
read äqd (lit. ‘knot’) and not ïqd (‘pearl-string’) and translated in
the technical sense of ‘link, connection, relation, rule’. To the best of
my knowledge, no other work on grammar uses the word äqd -üqud
with the same meaning. A parallel Hebrew term qeser (pl. qisronim)
occurs in Me’or Äyin, an anonymous Karaite grammatical treatise
composed in Hebrew in the second half of the eleventh century,
where it is used to describe dependences between grammatical facts
similar to those found in Kitab al-üqud.13 However, considering that
Kitab al-üqud is one of the main sources of Me’or Äyin,14 qeser here
is in all probability a translation of äqd.
If this meaning of the term üqud is established, the title of our
treatise must be translated Book of Rules Regarding the Grammatical
Inflections of the Hebrew Language rather than Pearl-Strings on the
Grammatical Inflections of the Hebrew Language, even though it was
probably intentionally ambiguous for rhetorical effect. In this case it
seems likely that it was the purpose of the author not simply to col-
lect grammatical facts pertaining to the conjugational patterns of
Hebrew verbs but to link them to one another and in doing so to
create a set of practical rules to facilitate the acquisition of the system
of Hebrew verbal derivation.
An additional purpose of the author, announced in the introduc-
tion in the words ‫‘( ויגמל פיה גמלא מנהא‬all conjugations would be
brought together’ [MS O, fol. 13v.]), was to discuss all conjugational
patterns of Biblical Hebrew. He did this following the method of
symbols common in Karaite linguistics before the twelfth century.15
13
Zislin, Me’or Äyin, 139 44: for Zislin's interpretation of qeser as ‘relationship
between grammatical forms’, cf. ibid., 104.
14
Ibid., 17.
15
On this method, cf. ibid., 17 19, 20; D. Becker, ‘The “Ways” of the Hebrew
Verb according to the Karaite Grammarians Abu al Faraj Harun and the Author of
Me’or Äyin’ (in Hebrew), in M.A. Friedman (ed.), Teüdah 7 (Studies in Judaica,
Tel Aviv 1991), 249 75; A. Maman, ‘Medieval Grammatical Thought: Karaites
versus Rabbanites’ (in Hebrew), Language Studies 7 (1996), 79 96, esp. 87 96.
173
The idea of this method is to divide Hebrew verbs into groups such
that all verbs in a group have one particular vowel in the first (or the
last) syllable of their imperatives and one particular vowel in the first
(or the last) syllable of their past forms, which is different from that
of the imperative. These classes of verbs are then assigned mnemonic
symbols (Ar. älama), i.e. disyllabic Hebrew words such that the first
vowel of the symbol word corresponds to the vowel common to all
imperatives and the last vowel of the symbol word corresponds to the
vowel common to all past forms in the group. For instance, the sym-
bol ‫ ִשיָרה‬represents verbs in which the first vowel of the imperative is
Ìireq and the first vowel of the past form is qamaÒ, e.g. ‫ָשם‬, ‫נ ָכון‬, ‫נ ַָמק‬,
etc. By their definition, the symbols can comprise verbs of different
types so further division into conjugational patterns (Ar. taÒrif ) is
performed in order to group together verbs with a similar structure.
Then each conjugational pattern is described by citing a number of
forms of a representative verb with this pattern. Apart from Kitab al-
üqud, this method is currently known from Abu al-Faraj Harun’s
books al-Kitab al-mushtamil and al-Kitab al-kafi as well as from
Me’or Äyin.
A comparative study of the accounts found in these four works is
beyond the scope of this article. However, it is worth noting that
Kitab al-üqud contains a much more elaborate description of the
verbal system of Biblical Hebrew than that found in al-Kitab al-kafi,
of which it was intended to be an abridgment. Indeed, Kitab al-
üqud presents twelve symbols comprising fifty-eight conjugational
patterns built on the basis of the first vowels of the imperative and
the past, as opposed to seven symbols and twenty-one conjugations
found in al-Kitab al-kafi. It also mentions five symbols with thirty-
five conjugational patterns built on the basis of the last vowels of the
imperative and the past, as opposed to three symbols with fifteen
conjugations found in al-Kitab al-kafi. Moreover, Abu al-Faraj
Harun limits his description of a conjugational pattern to four forms
of a representative verb, namely masculine singular imperative, third
person masculine singular past, active participle masculine singular
and passive participle masculine singular. This number was signifi-
cantly augmented by the author of Kitab al-üqud, who lists eighteen
forms, i.e. masculine singular imperative, third person masculine sin-
gular past, active participle in both genders and numbers, passive
participle in both genders and numbers, feminine singular impera-
tive, third person feminine singular past, masculine plural impera-
tive, masculine plural past and four forms of the future in the order
of the prefix letters ‫אינת‬. This set of uninflected forms is often sup-
174
plemented by numerous series of forms with pronominal suffixes,

which considerably increases the total number of forms in each con-
jugational pattern. It can thus be concluded that Kitab al-üqud can-
not be regarded as a mere abridgment of al-Kitab al-kafi. Rather, it
represents a substantial expansion and revision of the material on
verbal inflections of Hebrew contained in Abu al-Faraj’s work.
One can gain an even better idea of the nature of the relation be-
tween Kitab al-üqud and al-Kitab al-kafi by comparing the contents
of these books. Indeed, of the thirty-nine chapters and sections16 of
Kitab al-üqud only twenty-one closely follow and summarize the
text in the respective chapters of al-Kitab al-kafi. These are the fol-
lowing:
1. Introduction;
2. chapter ‘On the categories of words used in natural speech’;
3. section ‘On words belonging to one or more parts of speech’;17
4. chapter ‘On the infinitive’;
5. section ‘On the differences between verbal nouns and nouns called
infinitives’;
6. section ‘On true and pseudo-infinitives’;
7. chapter ‘On transitive and intransitive verbs’;
8. chapter ‘On the first and second imperative’;
9. chapter ‘On types of nouns’ apart from its sections ‘On pronouns’
and ‘On suffixed object pronouns’;
10. chapter ‘On conjoining’;
11. chapter ‘On the connective’;
12. chapter ‘On the attribute’;
13. chapter ‘On the emphatic and the permutative elements’;
14. chapter ‘On the initial item and the predicate’;
15. chapter ‘On true and pseudo-verbs’;
16. section ‘On the attribute resembling the active participle’;
17. section ‘On the order of the verb, the agent and the patient’;
18. chapter ‘On the meaningful arrangement of parts of speech’;
19. chapter ‘On the functional categories of passive participles’;
20. chapter ‘On some of the functions of masculine letters’;
21. section ‘On ambiguities in language’.
It can easily be seen that most if not all chapters in this group are
dedicated to syntactical issues.
Another eight chapters and sections address topics discussed in al-
Kitab al-kafi but deal with them in a different way.
16
The hierarchy of chapters and sections is not very strict in Kitab al üqud so
that chapters sometimes include sections on deviant topics.
17
Titles of sections given in italic are mine, not the author’s.
175
1. The chapter ‘On masculine and feminine letters’ divides the

letters of the alphabet into root and functional letters and calls
them ‘feminine’ and ‘masculine’ respectively,18 a terminology
not found in al-Kitab al-kafi.
2. The section ‘On four types of letters, namely root letters, auxiliary
letters, built-in letters and affixed letters’ introduces a new gram-
matical concept of an ‘auxiliary letter’19 which in modern terms
corresponds to the first root letter of ‫פ"ל‬, ‫ פ"נ‬and ‫ פ"י‬verbs.
3. The chapter ‘On conditions for forming morphological pat-
terns’ shows a significantly different arrangement of the mate-
rial.
4. The chapters (1) ‘On imperatives which differ in their first
vowel from the past verb forms derived from them and other
related matters’; (2) ‘On imperatives which coincide in their
first vowel but differ in their last vowel [from the past verb
forms derived from them] and on other related matters’; and
(3) ‘On imperatives which have no past form’ not only deci-
sively expand the contents of a single chapter on verbal conju-
gations in al-Kitab al-kafi but also introduce the notion of
üqud, i.e. regular derivational relations between different forms
of verbs.
5. The chapter ‘On types of active participles’ discusses the matter
from a different, i.e. morphological rather than comparative
and semantical point of view.
6. The section ‘On pronouns’ suggests a different definition of
the term ‘pronoun’ denoting only pronominal suffixes as pro-
nouns (Ar. ∂ama’ir) and calling independent pronouns nouns
(Ar. asma’ Âahira).
Contrary to the previous group, these chapters and sections are

mainly morphological.
18
For other works on grammar using this terminology, cf. N. Allony. ‘Qarqaöt
ha Diqduq by Joshua ben Abraham: Three Judaeo Persian Fragments from the
Cairo Genizah’ (in Hebrew), in Y.D. Gilat, Ch. Levine and Z.M. Rabinowitz (eds),
Studies in Rabbinic Literature, Bible and Jewish History (Jerusalem 1982), 291 311,
esp. 297 8.
19
The introduction of this concept was previously attributed to Me’or Äyin;
Zislin, Me’or Äyin, 14; Maman, ‘Medieval Grammatical Thought’, 81 4; idem,
‘The Hebrew Alphabet as a Grammatical Mnemotechnic Framework: Introduction
to al Kitab al mushtamil, Part III’ (in Hebrew), Language Studies 8 (2001), 95 139,
esp. 103.
176
Still another ten chapters and sections of Kitab al-üqud have no

parallels in al-Kitab al-kafi.20 These are:
1. Chapter ‘On establishing root letters and other related mat-
ters’;
2. section ‘On the origins of language’;
3. section ‘On types of imperatives’;
4. chapter ‘On imperatives which differ neither in their first
vowel nor in their last vowel from the past verb forms derived
from them’;
5. chapter ‘On conjugational patterns belonging to mnemonics
‫הם‬ֵ ‫ה‬
ָ , ‫הנן‬
ְ , or ‫מת‬
ֵ ‫ה‬
ָ ‫ה‬
ֶ ’;
6. chapter ‘On establishing the form of an imperative in diffi-
cult cases’;
7. chapter ‘On passive participles which belong to a conjuga-
tional pattern’;
8. chapter ‘On stripping words of added letters in order for a
word to return to its basic form without additions’;
9. chapter ‘On infiäl and iftiäl’;
10. chapter ‘On forms of active and passive verbs’.
Again, the majority of chapters and sections in this group deal
with morphological matters and often aim to teach methods of mor-
phological analysis to beginners.
A pattern emerges from this breakdown of the contents of Kitab
al-üqud reflecting the nature of revisions undertaken by its author in
his attempt to make the material contained in al-Kitab al-kafi more
easily accessible to learners of Biblical Hebrew. He summarized the
chapters on syntax, considerably expanded or modified chapters on
verbal morphology and added a number of new pedagogically orien-
tated morphological chapters. By doing this he produced a treatise
heavily relying upon al-Kitab al-kafi in its syntactical part but largely
independent from it in its account of the verbal morphology of He-
brew. Considering that the morphological chapters constitute the
central part of the book (c. 80 folios out of the total of 118) it can be
stated that even though Kitab al-üqud is described by its author as
an especially concise abridgment of al-Kitab al-kafi it must in fact be
regarded as a treatise in its own right closely related to rather than
dependent upon the book of Abu al-Faraj Harun.
20
This of course does not mean that all of them are completely novel and do
not rely on other books by Abu al Faraj Harun or other authors.
177
Conclusions
In this paper I have shown that the fragment published by H.

Hirschfeld in his article ‘An Unknown Grammatical Work by Abul-
Faraj Harun’ and the Karaite grammatical treatise al-MukhtaÒar con-
tained in the MSS II Firk. Evr.-Arab. I 2581, 2591, 2724, 2594 are
parts of one and the same work, namely Kitab al-üqud fi taÒarif al-
lugha al-ïbraniyya, which has survived in the above-mentioned MSS
nearly in its entirety. It can now be proved that Kitab al-üqud as well
as the other epitome of al-Kitab al-kafi called al-MukhtaÒar were not
composed by Abu al-Faraj Harun ibn Faraj. The author of these
works must thus be considered anonymous. The title Kitab al-üqud
fi taÒarif al-lugha al-ïbraniyya should probably be translated Book of
rules regarding the grammatical inflections of the Hebrew language. It
seems that the purpose of the author was to compose a pedagogically
orientated work presenting the student with an expanded system of
symbols, an enlarged number of detailed representative conjugations
as well as a set of principles linking together different facts of He-
brew verbal morphology.
Further investigation of the text of Kitab al-üqud and comparing
it with earlier and later Karaite works on grammar is expected to
deepen our knowledge of Karaite theory of Hebrew verbal morphol-
ogy and to contribute to the elucidation of the course of develop-
ment of Karaite grammatical thought.
178

THE INTRODUCTION OF SERGIUS OF

RESH{AINA TO GALEN'S COMMENTARY
ON HIPPOCRATES' ON NUTRIMENT
GERRIT BOS and Y. TZVI LANGERMANN

BAR-ILAN UNIVERSITY
Abstract
MS New York, Jewish Theological Seminary of America 2761, pre

serves an Arabic version in Hebrew characters of Galen’s commentary
to Hippocrates On foods, including an introduction by Sergius of
Reshäina, who translated the Greek into Syriac. Neither this transla
tion nor the introduction had been known to exist. The introduction,
which follows the canonical eight fold formula of accessores ad
auctores, is an extremely rich document for the history of commentar
ies to scientific texts. Sergius’ observations on the similarities in elusive
writing between Hippocrates and Plato are particularly interesting.
Introduction
MS New York, JTSA 2761

MS New York, Jewish Theological Seminary of America 2761 is an
important codex containing parts of several medical or medical-
philosophical treatises, some of them unique, all of them transcrip-
tions of Arabic texts into Hebrew letters.
The texts are written in two columns per page, each of about 30
lines. Unfortunately, the upper left hand corner has been torn off, so
that the exterior column has only about 27 legible lines; for some
lines, only one or two words can be deciphered. A full and definitive
description of the contents of the manuscript must be deferred to
some other occasion. For now the following brief account must suf-
fice. The first ten or so folios contain remnants of a didactic poem
on medicine and philosophy, as well as the opening pages of what
appears to be a tract dealing chiefly with medicine. The introduction
to this tract, however, which one of us (Langermann) plans to study
in a separate article, recounts a very interesting and unusual cos-
mogony. There follow some portions of the Galenic corpus, includ-
179
THE INTRODUCTION OF SERGIUS OF RESH{AINA
ing the end of On Medical Experience (al-Tajriba ’l-†ibbiyya), and the

beginning of Galen’s commentary to Hippocrates’ On Nutriment
(Kitab ’l-Ghidha’). The great bulk of the codex is taken up by por-
tions of al-Zahrawi’s Kitab ’l-TaÒrif.
The text which we present here, and which is found (in the cur-
rent pagination) on ff. 8b–10a, is the rendering into Arabic of the
introduction of Sergius of Reshäina to his Syriac version of Galen’s
commentary on the Hippocratic treatise On Nutriment. As far as we
have been able to determine, Sergius’ translation is not recorded in
the literature, and there are no other copies known to us, in either
Syriac or Arabic, of this text.1 The translation into Arabic is attrib-
uted to one Yusuf ibn Khaj (licet al-Khuri?), also otherwise un-
known. The manuscript was inspected by a scholar whose identity is
unknown to us (perhaps Alexander Marx?) in the 1920s. His atten-
tion was attracted by the very odd name of the translator into Ara-
bic, and he addressed a query concerning this matter to one of the
leading scholars of the epoch, Gotthelf Bergsträsser. Bergsträsser’s re-
ply was copied onto one of the first leaves. It reads:
‘Prof Bergstrasser writes (17.X.26): Was den Uebersetzungnamen
ihrer Handschrift anlangt, so kommt wohl nur Jusuf an-Naqil im
Betracht. Ibn UÒaibiä I 201f.: qUM« verstümmelt zu ‰Uá« und dies
dann falschlich als Vatername angelasst.’
Abu Yaäqub Yusuf bin Ïsa, known as an-Naqil (‘the translator’)
was from Khuzistan and a student of bin ∑ahrbakht.2 Ibn Abi
UÒaibiä had a low opinion of his work. In our view, the translator
whose name is garbled in the opening sentence is likely to have been
Yusuf al-Khuri, who is known to have translated Galen’s book on
simples into Arabic.3
1
Îunayn’s version of Galen’s commentary, not extant, is noted in Fuat Sezgin,
Geschichte des arabischen Schrifttums, vol. 3 (Frankfurt 1970), 137. The latest inven
tory of Syriac versions of Galenic writings is Rainer Degen, ‘Galen im Syrischen:
Eine Ubersicht uber syrische Überlieferung der Werke Galens’, in Vivian Nutton
(ed.), Galen: Problems and Prospects (London 1981), 131 66.
2
Ibn Abi UÒaibiä, {Uyun al anba} fi †abaqat al a†ibba (Beirut: Dar al thaqafa,
third printing, 1981), part 2, 173.
3
Concerning Yusuf al Khuri, cf. Îunayn ibn IsÌaq, Risala, no. 53 (Ed. G.
Bergsträsser, Îunain ibn IsÌaq. Über die syrischen und arabischen Galen Übersetz
ungen. (Abhandlungen für die Kunde des Morgenlandes hrsg. von der Deutschen
Morgenländischen Gesellschaft. XVII. Band. No. 2). Leipzig 1925). According to
Îunayn his translation of On simples is of a poor quality (khabitha radi’a). See as
well G. Strohmaier, ‘Der syrische und der arabische Galen’, Aufstieg und Niedergang
der Römischen Welt, hrsg. v. W. Haase u. H. Temporini, Teil II, Bd. 37, 2. Berlin,
New York 1994, 1987 2017, at p. 2001.
180
Sergius asserts that he has translated this book for a ‘brother’ who,
out of ‘love for all knowledge that the soul requires’, requested
Hippocrates’ book On Nutriment in particular. The ‘order’ to write
— given here, though, not by a superior but by an equal — is a very
common theme in introductions. Sergius goes on to say that he did
not want to undertake the task out of fear of incurring ‘the envy of
those who are not satisfied by anything other than amassing money.’
Apologies of this sort are common in Syriac literature.4 However, we
have not found any other case where the writer expresses his fear of
avaricious envy; does this mean that Sergius was well-paid for this
translation, and feared the envy of his rivals?
The ‘Alexandrian’ Introduction

The bulk of the introduction concerns the merits and utility of On
Nutriment, the Hippocratic corpus and the place of On Nutriment
therein, and Hippocrates’ mode of exposition in this and other
works. These and other related topics are developed within the struc-
ture of the seven- or eight- part introduction that is traditional
among ‘the Greeks’, as Sergius explains: ‘It is the custom of the
Greeks, when beginning commentaries to the books of the ancients,
to lay down at the beginning of the books, eight headings, by means
of which will become clear their aim, the reasons [for writing them],
their strength, their limit, their author, their number (?), how they
were executed, and where they are to be read [i.e. place in the cur-
riculum?] by means of which [information] they encourage people to
read it, once they have been given to hear what wonderful things,
hidden secrets, and enormous benefits are in it.’
The introduction divided into seven or eight parts is indeed a tra-
dition that developed in late antiquity among commentators to
books of science and philosophy.5 The tradition was continued by
Syriac, Arabic, Hebrew and Latin authors.6 Jaap Mansfeld in particu-
lar has made some very close studies of this literary phenomenon;
however, he has not paid much attention to the medical literature.7
4
Eva Riad, Studies in the Syriac Preface (Uppsala 1988), 190 208.
5
L.G. Westerink, ‘The Alexandrian commentators and the introductions to
their commentaries’, in Richard Sorabji (ed.), Aristotle Transformed: The Ancient
Commentators and their Influence (London 1990), 325 48. See also the following
notes.
6
See now Sara Klein Braslavy, ‘The Alexandrian Prologue Paradigm in
Gersonides’ Writings’, Jewish Quarterly Review 95 (2006), 257 89, and the litera
ture cited there.
181
Galen lived several centuries before Philoponus and other commen-

tators of late antiquity. However, Mansfield has included Galen’s two
essays, De Ordine librorum and De Libris propriis, in his study. He
observes that Galen does not apply a schema isagogicum and only in-
cidentally avails himself of a technical vocabulary. Nonetheless, one
may conclude that the following introductory issues are relevant in
relation to his own works: (a) general matters, such as the distinction
between the sects, the systematic classification of the treatises, the
work that has to be studied first (which is not the same for all read-
ers); (b) particulars, for example, the purpose, or theme, of each in-
dividual work, or the author’s motive for writing it (which some-
times explains the book’s utility, as in the On Proof), the question of
authenticity, the position of the treatise or monograph in the order
of study and the reason for or explanation of its title.8
A little further on Mansfeld summarizes: ‘Summarizing this evi-
dence concerned with both general and particular issues to be found
in Galen, I find that several general and the great majority of the par-
ticular issues of the later schemata isagogica are anticipated in his
autobibliographies and commentaries, and sometimes in his other
works as well. The clarification of what is unclear, the various modes
of expression or elucidation, the aims and qualifications of the exegete,
the proper qualifications of the students, the theme of the work, the
intention of the author (of the treatise or of the commentary), the
utility of the work, its authenticity (or that of its parts), its title and its
division into books or sections, the systematic ordering of the indi-
vidual treatises as well as the order of the study of the works of
Hippocrates (and Galen himself ) are discussed whenever necessary.’9
In the passage that we have just cited, Sergius refers explicitly
twice to the eight-part introduction, but, as we shall see, he goes
through the gamut of topics to be covered two or three times, and
there is some inconsistency in the number of topics to be covered
7
Jaap Mansfeld, Prolegomena: Questions to be Settled before the Study of an Au
thor or a Text (Leiden 1994), and idem, Prologemena mathematica from Apollonius of
Perga to Late Neoplatonism (Leiden 1998). The second book, as the title indicates, is
devoted to mathematical writings; there are a few sparse references to Galen, listed
in the index s.v. There is an extensive discussion of the ‘Alexandrian Introduction’
in Arabic literature in Gregor Schoeler, ‘Der Verfasser der Augenheilkunde K. Nur
al üyun und das Schema der 8 Präliminarien im 1. Kapitel des Werkes’, Der Islam
64 (1987), 89 97. See now also Hinrich Biesterfeldt, ‘Palladius on the Hippocratic
Aphorisms’, in Cristina D’Ancona (ed.), The Libraries of the Neoplatonists (Leiden
and Boston 2007), 385 97.
8
Mansfeld, Prologomena (1994), 130 1.
9
Ibid., 173.
182
and their rubrics. As we shall see, Palladius, in his unpublished com-

mentary to Hippocrates’ Aphorisms, also presents multiple and in-
consistent lists, and so also does Abu ’l-Faraj Äbdallah Ibn ’l-™ayyib
in the introduction to his commentary on Aristotle’s Categories.
Sergius first exhibits a list of eight themes or ‘headings’ (ru’us, a
literal translation of the Syriac kefalaia, itself a Greek loan-word) that
‘the Greeks’ traditionally cover in their introductions to commentar-
ies. The themes have names but no explanation, and their meaning is
not always clear. They are the following:
1) their purpose (ghara∂)
2) their ‘causes’ (asbab)
3) their strength (quwwa) {utility??}
4) their ‘limit’ (muntaha) {order??}
5) their author (mu’allif)
6) their number (ïdda)
7) how they were executed [method of exposition?] (kayfiyyat ämaliha)
8) where they are to be read (maw∂iü qira’atiha)
All prologues begin with the purpose. ‘Causes’ may refer to the
reason or cause (aîtía) for the title, also a regular feature.10 Rubric
five, ‘author’, most likely refers to the authenticity of the work. If the
sixth heading is to be read ïdda (so we prefer), then it refers to the
number of chapters into which the book is divided. The eighth item
most likely refers to the place of the work in the order of study of the
writings of a particular author.11
A little further on Sergius presents the list again, this time with a
short explanation of each rubric. Unfortunately, only the last three
are legible in our fragment, but their intentions are clear enough,
and they conform to the full discussion that follows in the introduc-
tion. These are: (6) the parts into which the book is divided; (7) to
which branch of knowledge the science (taught in the book to be
commented upon) belongs; and (8) which type of instruction the
author will employ. After twice displaying lists of topics that he plans
to cover, Sergius begins the introduction proper; he works his way
through the list, going into as much or as little detail as he sees fit.
Published studies on this topos examine a wide variety of literary
creations but, as we have already said, medical works have for the
most part been neglected. Mansfeld briefly discusses Stephanus of
10
See, e.g., Philoponus’ prologue to his commentary to Aristotle’s Categories,
cited by Riad, Syriac Preface, 45; Syriac ëllata, as in the book of Nestorius, cited by
Riad, 48.
11
Riad, Syriac Preface, 62, referring to prologues to commentaries on the Aris
totelian corpus.
183
Athens’s commentary on Hippocrates’ Prognostics (written roughly be-

tween 550–650), whose introduction reflects the schema found in
commentaries to Aristotle. It opens with this sentence: ‘Let us also
this time begin with the eight headings which are the customary pro-
legomena for each work: the intention, the usefulness and so on’. Prac-
tically the same sentence stands at the beginning of the Palladius/
Stephanus commentary on On Fractures, where each of the points is
then discussed at appropriate length.12 It is thus particularly appropri-
ate to bring into the discussion here, and to discuss at some length,
an as yet unpublished introduction to a Hippocratic text, namely
Palladius’ commentary on the Aphorisms. This text is not extant in the
original Greek, but an Arabic translation, as well as two different
Hebrew versions prepared on the basis of the Arabic, are available.13
Palladius goes through no less than three lists of topics to be cov-
ered. He says that they number eight, and, in fact, we find eight
items in the first two lists; however, it seems that only seven are cov-
ered in the third list. In fact, all three lists differ from each other in
some way. A similar inconsistency is found in the introduction of
Abu ’l-Faraj Äbdallah Ibn ’l-™ayyib to his commentary on Aristotle’s
Categories. Abu ’l-Faraj lists ten items that are traditionally brought
up for discussion in the introductions to Aristotle. However, in his
rather lengthy exposition, items four and five switch places.14
In both Palladius and Sergius, the first list displays simply eight
topics — single words whose meaning in this context is not always
clear. Palladius then presents a second list, in which the eight topics
are clarified by means of short phrases (ten words or less); this fea-
ture too has a parallel in Sergius. These are the same eight topics that
Palladius names in the first list, but their order is different. Finally,
the third time around, Palladius — again, just like Sergius — gives
not just a list, but a full discussion of the topic.
For purposes of comparison with Sergius’ incomplete text, as well
as the intrinsic importance of this hitherto unpublished material, we
shall survey here the eight topics that appear in Palladius’ introduc-
tion; some of this will prove to be useful when we return to Sergius
and sketch out the full treatment he gives in his third listing. Our
translations in Palladius’ first list are influenced by his later elabora-
12
Mansfeld, Prologemena (1994), 53.
13
On this important text, whose publication, we hope, will not be long de
layed, see most recently Biesterfeldt, ‘Palladius’.
14
Cleophea Ferrari, Der Kategorienkommentar von Abu l Farag Äbdallah Ibn
a† ™ayyib: Text und Untersuchungen (Leiden and Boston 2006), 95 101 (German),
2 11 (Arabic).
184
tion upon these rubrics. In list II our translations are somewhat free,
since a literal English translation would be dense and in some cases
simply impossible.
list I
1 importance (himma)
2 benefit (manfaä)
3 relationship (nisba)
4 genre (Òanf )
5 composition (ta’lif )
6 soundness (ÒiÌÌa)
7 pedigree (isnad)
8 method (tadbir)
list II
1 it should have some importance
2 it should also be of benefit
3 it should be related to something
4 (6) it should be sound
5 (4) it should belong to a known literary genre
6 (5) it should be composed of different parts
7 it should trace back to one of the aspects of wisdom
8 it should have a method that is manifest (literally ‘described’)
list III
(1) The importance of the Aphorisms lies in its comprehensiveness.
(2) Its utility comes to the fore in its suitability for both skilled and
unskilled readers. (3) Palladius states at the outset that it (the Apho-
risms) is related to Hippocrates’ ‘method’ (tadbir). He explains that
the Hippocratic corpus divides up into four categories: (a) books con-
cerned with the action of people, e.g., On Prognosis, which explains
how we may foretell medical developments on the basis of symptoms
now present; (b) books concerned with the action of substances, e.g.
On Barley Water; (c) books that concern that which affects us from
without, e.g. Airs, Waters, and Places; (d) a method for writing a
book, such as the Aphorisms, which is divided into short statements.
Clearly the criterion for establishing the fourth category differs from
that employed for the first three. (4) ‘Soundness’ may mean here au-
thenticity, one of the standard topoi, since Palladius remarks here that
the soundness is connected with Hippocrates’ reputation. He men-
tions four books — the same four noted under the previous rubric as
exemplars par excellence of the Hippocratic corpus — that are utterly
indispensable and have no substitute.
185
(5) The Aphorisms is the key to medicine, and it must be read first.
Hence this rubric addresses the common theme of the place of the
book in the order of study. (6) Composition: Hippocrates wrote the
book in the form of concise aphoristic statements, and Galen later
organized it into seven sections. (7) The Aphorisms trace back to the
branches of medicine, theory and practice. (8) The method is subtle
(la†if ). Hippocrates employs three methods: (a) prolix narrative
(Ìadith); (b) concise, aphoristic remarks; and (c) a combination of
the first two.15
Let us return to Sergius. After displaying the second list of eight
topics, Sergius describes briefly the range of material covered in On
Nutriment — not just nutrition, but also bone-setting and embryol-
ogy. Perhaps this is related to the first topic, the book’s utility; but it
is only after these remarks that Sergius informs us that he will now
list the book’s uses, and they are many:
i. Some patients avoid drugs completely.
ii. Foods penetrate and work from within in a fundamental way that
drugs do not.
iii. Foods are appropriate for sound bodies as well as for the sick.
iv. The physician who knows how to make proper use of foods needs
nothing else.
v. The indigent have no choice but to rely upon foods.
vi. Mistakes in nutrition are not as damaging as those involving drugs.
vii. Travellers, or someone who finds himself in a place where drugs are
unavailable, must rely upon foodstuffs.
Sergius concludes that On Nutriment is indispensable, much as
Palladius had said about the Aphorisms in (4) above, but for different
reasons.
The second item is the reason for the book’s title. This is a very
common theme in the prologues, but Sergius’ explanation here is not
clear to us. The third topic is the division of the Hippocratic corpus.
According to Sergius, some Hippocratic writings discuss natural
things (humours, chyme, foetus, etc.), while others treat of unnatural
things (fevers, fractures, etc.). On Nutriment deals with both. This
rubric is thus close in structure to Palladius’ ‘relation’ (3), but the
contents are entirely different. The fourth rubric is preserved very
15
Mutandi mutatis this remains the scholarly evaluation of the Hippocratic cor
pus. Jacques Jouanna, Hippocrates (Baltimore and London 1999), writes (p. 57):
‘the diversity of the corpus derives from the various audiences aimed at by the
treatises…Certain treatises are addressed to a broad readership composed of both
specialists and lay people; others are aimed at a specialist readership.’
186
imperfectly, but it addresses the place of the book within the

Hippocratic corpus.
The fifth item is the book’s authenticity, again a common theme,
and one that Palladius seems to have addressed in his (4). Sergius
cites the book’s conciseness and depth, the strength of the arguments
as well as the testimony of Galen — all testify to the book’s authen-
ticity. The sixth concerns the book’s division, or composition
(Palladius sixth topic too); but rather than discussing the chapters,
Sergius talks about the range of subject materials covered: nutrition
in general and the nutrition of individual organs; classifications of
foodstuffs; embryology; and more on food. The seventh is similar to
Palladius’ ‘lineage’ (seventh on his list as well); Sergius speaks about
theory and practice. The eighth item, again just like the eighth item
in Palladius’ list, is the mode of composition. Exactly like Palladius
Sergius says that some Hippocratic writings are concise, others pro-
lix, yet others, a mixture of the two. On Nutriment is concise, and
deliberately difficult of comprehension. It is addressed only to a seri-
ous audience, to those willing to invest the time and effort to fully
fathom the subject. Hippocrates employs here the same style as
Plato, and for the same reasons; we shall devote the next section to a
closer examination of this comparison between the two great au-
thorities. This judgment is confirmed by Jacques Jouanna, who de-
scribes On Nutriment as a collection of ‘intentionally enigmatic
aphorisms.’16
One key topos of the ‘Alexandrian introduction’ receives very brief
treatment on the part of Palladius, and none at all from Sergius (in
the part of the introduction that we have recovered), and that is the
order of study, in particular the order of study of the texts that make
up the Hippocratic corpus. Interestingly enough, we do have a de-
tailed account of this from an early Arabic source, the Kitab al-Nafi¨
of Ali ibn Ri∂wan (d. c. 1061). The one manuscript accessible to
us, Dublin, Chester Beatty Ar. 4026, is missing the relevant section,
but we do have a reliable report based upon another manuscript,
Cairo, Egyptian National Library ™ibb 483, in the monograph of
Joseph Schacht and Max Meyerhof.17 According to Ibn Ri∂wan,
16
Jouanna, Hippocrates, 401.
17
J. Schacht and M. Meyerhof, The Medico Philosophical Controversy Between
Ibn Butlan of Baghdad and Ibn Ridwan of Cairo: A Contribution to the History of
Greek Learning Among the Arabs (Cairo 1937), 23 4. A small section of Kitab al
Nafi¨ was translated into English by M.C. Lyons, ‘The Kitab al Nafi¨ of Ali ibn
Ri∂wan’, Islamic Quarterly 6 (1961), 65 71, who made use of a collation of the two
manuscripts prepared by the late A.J. Arberry.
187
Hippocrates left us no instructions concerning the proper order of

study of his writings, since he (Hippocrates) preferred to rely upon
oral instruction. Ibn Ri∂wan then proposes what seems to him the
proper course of study; it comprises the following texts (the titles are
given exactly as they appear in the monograph), in this order: On the
Foetus; Nature of Man; Airs, Waters, and Places; Aphorisms; Prognostic;
Gynaecology; Epidemics; On Humours; On Nutriment; The Physician’s
Office; On Joints; On Wounds of the Head.
Plato and Hippocrates

In the section of the introduction that treats of the mode of exposi-
tion, Sergius draws a parallel between the very concise, hence at
times enigmatic or even esoteric style, that Hippocrates chose to em-
ploy in On Nutriment, and Plato’s style:
‘Hippocrates [composed] his books according to this method only
because that was what he had to do. Hippocrates said, “Let books be
written for people according to their capacities". Plato said, “The
two opposites should not be joined together. Let not the pure be-
come contaminated by the impure. For when you mix sweet water
with water that contains mud, you putrefy the water, and it is not
good anymore. If you teach the wicked (†alan?! licet zallam), he will
mock and ridicule you. In general, only those possessing sound eye-
sight benefit from light”.’18
The connection between Hippocrates and Plato and, in particular,
the dependence in some measure of the latter upon the former, has
been much discussed. The link between the two has a solid basis in a
passage from the Phaedrus 270 C-D, where Plato cites Hippocrates
— more precisely, ‘Hippocrates and the truth’ — for the principle
that the physician must know the nature of the human body, and
especially the precise nature of each of its parts, in order to treat the
patient successfully. This idea has important implications for Plato’s
philosophy.19 It was Galen who pressed the connection between the
two hoary authorities most vigorously. Overall Galen felt that Plato
took some of the most important points of his doctrine from
Hippocrates.20 He wrote an extensive work on the agreement be-
18
It is not certain where the quotation from Plato ends; it is quite possible that
the last sentence in our citation is due to Sergius.
19
See Fritz Steckerl, ‘Plato, Hippocrates, and the Menon Papyrus’, Classical
Philology 40 (1945), 166 80.
20
Phillip De Lacy, ‘Galen’s Platonism’, American Journal of Philology 93 (1972),
27 39, esp. pp. 36 8.
188
tween the doctrines of the two thinkers, which was translated into
Arabic, though this translation is no longer extant.21
Nonetheless, we have not been able to find the apothegm attrib-
uted to Plato in any genuine Platonic writing, or in any of the many
doxographies and other repositories of Platonic wisdom. This is not
very surprising; there are many examples of sayings attributed to
Plato that display what seems to be authentic Platonic teachings, but
whose precise formulation is not found in any of Plato’s writings.22
One can find remarks that are suggestive of the saying found in our
text. In Phaedo 67 B, for example, Socrates argues that if we main-
tain our own purity, ‘we shall probably reach the company of others
like ourselves and gain direct knowledge of all that is pure and un-
contaminated — that is, presumably, of truth.’23 The gist of this
statement is found in Abu ’l-Faraj’s introduction to his own com-
mentary to the Categories: ‘As for Plato, it is found in statement, “Let
only the pure approach the pure”.’24 Al-Biruni cites from the Phaedo
passages that have no exact equivalent in the Greek text; his editor
and translator, Edward Sachau, speculates that he may have seen
some as yet unidentified recension.25 We may if we like indulge in a
similar speculation concerning Abu ’l-Faraj’s source.
Abu ’l-Faraj’s introduction is very relevant for another reason. He
too cites Hippocrates and Plato back to back. However, there is an
important difference in the point which each author ascribes to the
two sages. The passage from the Phaedo quoted by Abu ’l-Faraj con-
cerns the individual’s preparation for his own study, and the impuri-
ties that are warned against are those of the body. The individual
is warned to restrain his impulses and desires. The sayings of
21
See Sezgin, GAS, VI,105 6.
22
On this see the classic study of Franz Rosenthal, ‘On the Knowledge of Pla
to’s Philosophy in the Islamic World’, Islamic Culture 14 (1940), 387 422.
23
Edith Hamilton and Huntington Cairns, The Collected Dialogues of Plato,
Bollingen Series LXXI (Princeton 1961), 49; the translation is by Hugh
Tredennick.
24
Ferrari, Kategorienkommentar, 10 (Arabic).
25
See E. Sachau, Alberuni’s India, vol. 2 (London 1910), 277 8. Some of the
citations from Phaedo agree well with the Greek, but others do not; Sachau suggests
that they derive ‘from a work in which text and commentary were mixed together,
and the original form of the dialogue changed into a simple relation.’ This is of
course pure speculation; no such recension of the Phaedo is known to exist. In an
earlier passage from the same book, Sachau encountered three citations from Plato,
only one of which matched the Greek dialogues. He queried the great historian of
Greek philosophy E. Zeller, who suggested in response that the problematic pas
sages are likely to derive from a commentary to the Timaeus ‘by some Christian au
thor, as e.g. John Philoponus’.
189
Hippocrates and Plato in Sergius’ introduction, by contrast, address

the issues of writing for and teaching others. Immediately after citing
from Plato, Sergius adds, ‘But as for Hippocrates, it is found in his
statement that unclean bodies only add evil, the more they are nour-
ished'.26 In his usage, nourishing unclean bodies is a metaphor for
‘feeding’ the ignorant ideas that they are not able to cope with.
The Hippocratic Corpus

There are several short lists of Hippocrates’ books in Sergius’ intro-
duction, as well as occasional references to single works. Some of the
titles pose problems. Titles generally begin with fi, corresponding to
the Greek peri; if this preposition is followed by a string of nouns,
we cannot always be sure if one or more different books are in-
tended. Thus, for example, [On] the Nature of Man is surely an inde-
pendent treatise, even though the fi is missing. However, in the case
of On Blood, Semen, The Composition of Organs, and The Formation
of the Child (Fi ’l-Dam wa-l-nu†fa wa-tarkib al-a¨∂a’ wa-kawn ’l-†ifl ),
we are not sure how many different books are mentioned. Here fol-
low the names of the books and our identifications or conjectures.
They appear in our list in the order in which they appear in the text.
We refer whenever possible to Jacques Jouanna, Hippocrates (n. 15
above) and to the third volume of Fuat Sezgin, Geschichte des
arabischen Schrifttums (n. 1 above) for further bibliography.
On Seasons. Probably identical with Fi Ma¨rifat fuÒul as-sana; Sezgin,
p. 43.
On Temperaments. (Fi ’l-Amzija). This is most likely to be the known
work On Temperaments, called in Arabic Kitab ’l-AÌla† (Sezgin,
35).
On Blood, Semen, The Composition of Organs, and The Formation of
the Child (Fi ’l-Dam wa-l-nu†fa wa-tarkib al-a¨∂a’ wa-kawn
’l-†ifl ). The first of these (‘On Blood’) is likely to be Sayalan
’l-dam, not known in the Greek, but known (by title alone) in
Arabic (Sezgin, 46). On Semen is probably identical with the
Greek Peri Gones (On Offspring, or On Seed) and the Arabic:
(Fi ’l-Ajinna, or, al-Janin; Sezgin, p. 38, no. 10).
On Water, Air, Earth and Countries (Fi ’l-Ma’ wa-l-hawa’ wa-l-ar∂ wa-
l-buldan). This is the very famous Airs, Waters, and Places, which
26
Ibid. The citation, which is from Aphorisms II.10, appears in a variety of late
antique and Islamic sources, which are listed by Ferrari in her note 3. To these may
be added Galen, On simples (Kühn XI, 457), who adds that the same applies to
souls: ‘nourishing words’ are very harmful to impure souls.
190
went under different names in antiquity. Galen preferred to call

it Dwellings, Waters, Seasons and Countries.27 The title as it ap-
pears in Sezgin, 36, is Kitab ’l-Ahwiya wa-l-azmina wa-l-miyah
wa-l-buldan. See also Jouanna, 374–5.
On Bones and Bandages (Fi ’l-IÂam wa-l-a∂mida). This work also ap-
pears in the list of Ibn Ri∂wan (above, p. 188). We tentatively
suggest that it is probably identical with the work called in
Greek Kat’ Ietreion (In the Surgery) which, according to Galen,
might be more appropriately called Notes on Bandaging.28 The
traditional Arabic titles are Kitab Qa†ya†riyun or Îanu†at-†abib
(Sezgin, 36).
On Ulcers (Fi ’l-QuruÌ). Jouanna 415–6; this text is possibly identi-
cal with the one cited by al-Razi, K. al-Îawi XIII, 21 as: K.
Abuqra† fi l-quruÌ wa-kitabuhu fi l-jiraÌat; cf. Sezgin, p. 45.
On All Types of Pains (Fi Jami¨ anwa¨ ’l-awja¨ ).
Ameliorating Uteri ([Fi?] IÒlaÌ ’l-arÌam). This would seem to be one
of Hippocrates’ books (or perhaps a section thereof ) on sterile
women (volume 8 in the edition of Littré).
On the Nature of Man ([Fi] ™abiät ’l-insan). Sezgin, 37–8.
Aphorisms (Kitab al-FuÒul) (Sezgin, 28–32).
Prognosis (Kitab Taqdim al-ïlm). The usual title for this book in
Arabic is Taqdimat al-ma¨rifa.
Bonesetting (Kitab al-Jabr). This work went under several titles in
Arabic — al-Jabr, al-Kasr, and al-Khal¨ (Sezgin, 44). See next
item.
On Joints (Kitab al-MafaÒil). See Jouanna’s list, no. 43, pp. 402–3,
where On Fractures/On Joints is listed as a single entry. Kitab al-
MafaÒil is known in the Arabic literature as a separate treatise
(Sezgin, 45).
Epidemics (Kitab Abidhimiya). Sezgin, 34.
Acute Diseases (kitab al-‘ilal al-Ìadda). See Sezgin, 33, whose sources
all call this treatise al-amra∂ al-Ìadda. See next item.
On Barley Water (Fi ’l-Kashk; note that the usual title for this work in
Arabic is Kitab Ma’ al-shaïr, registered by Sezgin, 33). This is the
title sometimes given to On Acute Diseases, in Greek as well as in
Arabic (See Sezgin, 33); however, from the context it seems that
Sergius has in mind a separate treatise, perhaps a section of On
Acute Diseases that circulated independently.
On Nature. Unknown.
27
Jouanna, Hippocrates, 356.
28
See the translator’s introduction by E.T. Withington in Hippocrates, vol. III
(Loeb Classical Library, Cambridge and London 1928), p. 54.
191
‫‪THE INTRODUCTION OF SERGIUS OF RESH{AINA‬‬
‫‪Acknowledgments‬‬
‫‪Langermann’s research was supported by the Israel Science Founda-‬‬
‫‪tion (Grant no. 192/05). Publication of the text from the unique‬‬
‫‪manuscript is Courtesy of the Library of the Jewish Theological‬‬
‫‪Seminary.‬‬
‫‪The Text‬‬
‫‪The text has been copied out, line by line and column by column.‬‬
‫‪Lacunae are indicated by three dots (…). On a few occasions we sug-‬‬
‫‪gest a correction to the text within square brackets [ ] immediately‬‬
‫‪after the word or words that appear to be corrupt.‬‬
‫]‪ [1a‬הדא כתאב אלגדא‬

‫לאבקראט שרח גאלינוס נקל יוסף אבן כאג)?(‬
‫אלמתרגם קאל סרכס נבדא בעון‬
‫אללה פנכתב מא פסרה גאלינוס‬
‫מן כתאב אבקראט פי אלגדא‪ .‬קאל איהא אלאך‬
‫אלחביב אנך למא סאלתני וטלבת אלי מרארא‬
‫כתירה אן אתרגם לך בעץ כתב אליונאניין‬
‫אלי אלסריאניה וכאצה אלמקאלה אלדי וצעה‬
‫אבקראט פי אלגדא ודלך לחבך לגמיע מא‬
‫יחתאג אליה אלנפס מן אלעלם אעתנית‬
‫]ו[לא אחרמך מא סאלתני וטלבת אלי פיה‬
‫לאנה ממא יחתאג אליה ולם אכן אחב אן‬
‫אפעל שיא מן דלך למערפתי במא ילזמני‬
‫מן אלחסדה אלדין לא יגניהם אלא גמע אל‬
‫מאל מן אלעתב ואלמלאמה ודלך אנהם‬
‫לא יטלבון מנפעה טאלבי אלעלם אלדין ליסת‬
‫ענדהם אמואל תגניהם עלי]ען[ דלך ואנה איצי‬
‫אטלב אליך אלא תגהל ען טלב נפעה עלם‬
‫‪[1b]….‬‬
‫‪…..‬‬
‫‪…..‬‬
‫‪…..‬נט’ר פי‬
‫כתבה אלתי וצ’עהא פי גמיע אלאנואע מנהא‬
‫פי אלפצול ומנהא פי אלאמזגה ומנהא פי אל‬
‫דם ואלנטפה )?(ותרכיב אלאעצ’א וכון אלטפל‬
‫ופי אלמא ואלהוא ואלארץ ואלבלדאן ופי גבר‬
‫אלעט’אם ואלאצ’מדה‪ .‬ואלקרוח ופי גמיע‬
‫‪192‬‬
‫אנואע אלאוגאע ואצלאח אלארחאם וטביעה‬

‫אלאנסאן ולמא אלף הדא ראי אנה קד בקא מנהא‬
‫ארפעהא ואפצ'להא והו מא לא בד מנה ולא‬
‫גני ענה והו אלגדא‪ .‬אלדי בה קאום ]קואם[ כל חי לאן‬
‫בה יקבל אלגנין אלכון ואלנמו‪ .‬ובה תט’הר‬
‫אלאעצ'א אפעאלהא ובה תתבת אלצחה‬
‫וילבת גוהר אלאנסאן באקיא אלי אכ'ר עמרה‬
‫אלדי אליה ינתהי ולא ימכן אן ילבת אלחיואן‬
‫בליא מן אלגדא פלדלך מא ]אנמא?[ אלף הדא אלכתאב‬
‫וצירה כאלדכירה או כאלעלק אלנפיס ולאן‬
‫מן עאדה אליונאניין אדא אבתדוא פי תפאסיר‬
‫כתב אלקדמא אן יצעוא בדיא ללכתב תמאניה‬
‫רוס תטהר פיהא גרצהא ‪.‬ואסבאבהא‪ .‬וקותהא‪.‬‬
‫ומנתהאהא‪ .‬ומולפהא‪ .‬ועדתהא‪ .‬וכיפיה‬
‫עמלהא‪ .‬ומוצ'ע קראתהא‪ .‬ותחץ’ אלנאס‬
‫עלי קראתהא אדא אסמעוא מא פיהא מן אל‬
‫אמור אלמעגבה ואלאסראר אלכפיה‪ .‬ואל‬
‫פואיד אלעט'אם פלנדכר אלאן פי הדא אלכתאב‬
‫בתפסיר הדה אלרווס ונעלם מא אלדי יאמרנא‬
‫בה כל ואחד מנהא‪ .‬ואלרוס הי הדה‬
‫אלאול יכברנא בראי אבקראט אלתאני‬
‫…]‪[2a‬‬
‫…‬
‫‪…..‬‬
‫‪…..‬‬
‫אלו’ עלי כל גז יתפצל הדא אלכתאב‪ .‬אלז’‬
‫יבין פיה מן אי אנואע אלעלם הו‬
‫הדא אלכתאב אלח’ יכבר באי אנוא]ע[ אלתעלים‬
‫יסתעמל אברקאט הדא אלכתאב אי אם אבוקראט‬
‫יריד אן יבין פי הדא אלכתאב עלי טלב עלם‬
‫אלגדאה וגרץ’ אנואעה ותגיירהא וקואהא ופעלהא‬
‫ומנאפעהא ומצרהא ולא יגד בדא מן אלכרוג‬
‫ען חדה‪ .‬ואלקול פי אנואע אלאמראץ’ אלתי‬
‫תך מן כתרה אלגדא או מן רדאתה או מן‬
‫כיפיתה או מן כמיתה באבינה ‪.‬וקול איצ'י פי‬
‫גבר אלעטאם אלמכסורה ועדה אלאיאם אלתי‬
‫תברא פיהא ופי כון אלגנין ואנחבאלה‪ .‬ופי‬
‫תחרכה פי אלרחם ותמאמה עלי אלכמאל‬
‫ואבאן אלכרוגה מע עדה איאם כל ואחד מן‬
‫הדה תם ירגע פיכמל עלם הדא אלמקאל והדה‬
‫הי אלמנאפע אלתי ינאלהא מן קרא הדא אלכתאב‬
‫‪193‬‬
‫לאנא נרי נאסא כתירא אדא מרצ'וא ליס‬

‫אנמא לא יסתעמלון אלאדויה פקט ולכן לא‬
‫יקדרון אן יסמעוא בדכרהא ואלמתטבב‬
‫למערפתה באנואע אלגדא קד יואפקהם עלי‬
‫שהותהם ויגדוהם במא ינבגי וינתפע בה‬
‫ולו סקוא פי אמראץ’ אלחמי ווגע אלכבד ואל‬
‫טחאל ואלראס ואלמעדה וגמיע אלאעצ'א אלדאכלה‬
‫גמיע אלאדויה אלמואפקה ולם יך להם גדא‬
‫מואפק לדלך לבקית אלאוגאע עלי חאלהא לאן‬
‫אלדוא אדא עולג’ בה מן כארג’ לא יצל אלי דאכל‬
‫]‪ [2b‬יחל קותה פאמא אלגדא אלמואפק תחול‬

‫כימוסא מחמודא ואד'א אנצרף אלי אל‬
‫‪..‬אלוגע נפעה … אלמס במזאגה‪ .‬והדא‬
‫יצלח מע … אלשביב ותקדירה…‬
‫עלי מא יגב פלדלך תחפט’ אלצחה באנואע‬
‫אלאטעמה ואיצ'י פאן אלאדויה לא תסתקים‬
‫ללאגסאד אלצחיחה כמא קאל אבוקראט פי‬
‫כתאב אלפצול פינבגי אדא" אן תעד הדא אלכתאב‬
‫אפצ'ל כתב אלטב לאנה אדא ערפה אלטביב‬
‫וקדר אן יסתכרג’ מנה אלמנאפע לם יחתג’‬
‫אלי סאיר אלאשיא ואיצ'י פפי אלנאס אלפקרא‬
‫ומן לא יקדר עלי אתכאד אלאדויה למוונהא‬
‫וגלאיהא ולעדמהא פיסתגני באלגדא ען אלאדויה‬
‫ואיצ'י פאן אלכטא אדא וקע קבל אלאדויה לם‬
‫יכד אלעליל יפלת‪ .‬ואן וקע מן אלאגדיה פפי‬
‫אלנאט'ר ]אלנאדר[ מא יעטב אלעליל‪ .‬וקד ינתפע בהדא‬
‫אלכתאב איצ'י אלמסאפרון פי אלבר ואלבחר‬
‫אלדין לא יקדרון עלי סאיר אלאנואע‪ .‬ואן אתיה‬
‫איצ'י ללמר מוצ'ע לא יך פיה מא יחתאג אליה‬
‫מן אלאדויה פיחתאג’ אלי אלאגדיה צ'רורא"‬
‫פואגב אדא" ללטביב אן יעלם הדא אלכתאב‬
‫ואנמא יקע אלמדח ללטביב אדא אסתגני ען‬
‫אלאדויה ידל)?( לאגדיה ]יבדלהא באלאגדיה[ ודלך הו אלטביב אלמאהר‬
‫ולאן יעאני אלטביב באלאגדיה אפצ'ל ואכמל‬
‫פי צנאעה אלטב מן אן יעאני במא פיה אלגדאיה‬
‫ואלדואייה לו אמכן דלך פאן אלעלאג’ אנמא‬
‫נפע במא פיה האתאן אלקותאן‪ .‬ולאן יעאני במא‬
‫פיה קוה גדאה ודואיה כיר מן יעאני במא פיה‬
‫קוה דואייא" צרפה ולדלך ימדח אטבא אלהאכל)?(‬
‫לאנהם מתי וגדוא אלגדא יבלג להם מא אראדוא‬
‫‪194‬‬
‫]‪ [3a‬ואן קצר ען אלפעל לם יעאנוא באלאדויה‬

‫בתא ‪.‬ולדלך לא יגב ללטביב אן יעוד לאגס]אם[‬
‫אדא חלת בהא אלאמראץ’ אלאדויה‪ .‬פאנמ]א[‬
‫מדחוא אלטביב אדא צנע באל]אגדיה[ מא לא‬
‫יקדר גירה עליה באלאדויה מן שפא אלאמראץ’‬
‫ותדביר אלמרצ’י‪ .‬פקד ערף אלאן אן הדא אל‬
‫כתאב ליס ממא ינתפע בה כסאיר אלכתב פקט‬
‫בל הו ממא לא בד מן אקתנאיה‪ .‬וראס הדא אלכתא]ב[‬
‫מכתוב פי אלגדא לאבקראט ‪.‬כאן יבגץ’ כתרה‬
‫אלכלאם ולו אראד אן יקץ אנואע אלאגדיה‬
‫ואשכאצהא כתירה גדא לא יכאד יחאט בהא‬
‫פמן דא אלדי יקדר אן יצף אלאנואע ואלאשכאץ‬
‫פלדלך מא אסתעמל אבקראט אלדי קד עלם‬
‫אלטב ומע אלטב אלעלם באלכלאם הדא אלחד‬
‫ווסם הדא אלכתאב בהדא אלסמה‪ .‬ואנמא‬
‫סמי בהדא אלאסם לאן נאסא" כתירה עדדהם‬
‫וצ’עוא כתאבא כתירה פלולם יסמו אסמאה‬
‫‪...‬אואיל כתבהם ואסמא אלכתב‪ .‬ואלמראד בהא‬
‫מא כאן גל אלנאס יערפוהא ולא יעלמון מא‬
‫יחתאגון אליה מנהא‪ .‬פקד ינבגי אן תעלם‬
‫]א[ן גמיע כתב אבקראט אנמא וצ’עת עלי נחוין‬
‫אחדהמא ינבי ען אלאמור אלטביעיי]ה[ אי‬
‫אלאכלאט ואלקוי ואלגנין וטביעה אלאנסאן‪.‬‬
‫ואלכימוסאת ואנואע אלגדא ממא ישבה הדא‪.‬‬
‫ואלנחו אלאכר פי אלאשיא אלכארג’ה ען אלטביעה‬
‫באנואע אלחמי וכסר אלע’אם ותקדים אל‬
‫עלם פי אלעטב ואלסלאמה ותקלב אלאמזגה‬
‫ובאלגמלה סאיר אלאמור אלתי תשבה הדא‬
‫וכדלך וצ’ע הדא אלכתאב על עלי הדין הדין אלנחוין‬
‫פי אמור אלטביעה ואלכארג’ה ענהא‪ .‬ואלאצחא‬
‫…]‪[3b‬‬
‫…‬
‫…‬
‫…‬
‫פלדלך הדא אלכתאב הו אלראבע מן כתב‬
‫אבקראט ובעדה יקרא‪ .‬ואעלם אנה יסמי‬
‫באסם אבקראט סלתה נפר ]תלתה עשר[ וקד וצ’ע‬
‫כתבא בעצ’הא יכאד ישבה בעצ’א" פלדלך מא‬
‫שך מן פסר כתבהם וכאדוא לא ימיזון בעצ’הא‬
‫מן בעץ’‪ .‬פאמא נחן פנקול והו אלחק אן הדא‬
‫אלכתאב הו לאבקראט אלדי וצ’ע כתאב תקדים‬
‫‪195‬‬
‫אלעלם וכתאב אלפצול וכתאב אלגבר וכתאב‬

‫אלמפאצל וכתאב אפידימיא וכתאב אלעלל‬
‫אלחאדה ואנמא עלמנא דלך מן איגאזה וחרפה‬
‫אלתטויל מע גמוץ’ תפהמה וקלה פצ'ולה ופי הדא‬
‫מא הו אצדק ואכד פי אלחג'ה‪ .‬והו שהאדה‬
‫גאלינוס ותפסירה איאה מן בינהא כלהא‪ .‬והדא‬
‫אלכתאב ינקסם אלי ג’ מעאני‪ .‬אלאול פי גדא‬
‫אלעאמה ופי קותה ופעלה‪ .‬ופי אלאעצ’א אלתי‬
‫תגתדי בה ופי מנפעתה ומצ’רתה‪ .‬ואלתאני הו‬
‫פי תפציל אנואע אלגדא‪ .‬ופיה מע הדא ביאן מא‬
‫כאן מנה רטבא" ומא כאן מנה יאבסא" ואלמערפה‬
‫מן אלרואיח ואלטעום ופי תפציל אלאעצ’א‬
‫אלתי תצ’רהא או תנפעהא‪ .‬ופי אלאמראץ’ אלתי‬
‫תתולד ענהא‪ .‬ואלתאלת פי כון אלגנין ופי‬
‫עדד אלאיאם אלתי יכמל פיהא‪ .‬ועדד אלאיאם‬
‫אלתי יתחרך פיהא פי אלרחם ועדד אלאיאם‬
‫אלתי יתולד פיהא ותביין מא בין אלדכר ואל‬
‫אנתי תם רגע מן דלך אלי גמלה אמרה ויצף‬
‫אנואע אלגדא ויביין מנאפעה ומצ’רה ויקדר‬
‫…]‪[4a‬‬
‫…‬
‫…‬
‫ממן וצ’ע אלכתב … יבין…‬
‫ג’ אגזא מן אלעלם לאן מן אלעלם מא ידל עלי‬
‫אלמנטק ופהם אלעקל‪ .‬ומנה מא ידל עלי עמל‬
‫אלאיד … אלג’ עלי אסתעמאל אלנועין ועלי‬
‫הדא אלחאל תגד אכתר כתב אבוקראט פאן מנהא‬
‫מא הו פי טביעה אלאנסאן ומנהא מא הו עלי‬
‫גיר הדא אלכתאב אלכסר ואלגבר ואמא אלדי פי‬
‫אלכשך פידל עלי נועין לאן הדא אלכתאב‬
‫אנמא הו מערפה אלחמי ואלקיאס עליהא מן‬
‫אפעאלהא תם אלחכם עליהא במא יוגבה אלקיאס‬
‫תם בעד דלך דכר אלכשך ועמלה וסאיר אלאשיא‬
‫אלתי תעמל באלאידי פפי הדא אלכתאב נועאן‬
‫מן אלעלם ואלעמל לאנה אדא קאל פי אלגדא‬
‫ואנצראפה ליגדוא אלאעצ’א וינמיהא פי וקת אל‬
‫נמו ויכלף עליהא מא תחלל מנהא פי וקת דלך‬
‫אנמא יסתעמל פיה אלנט’ר ואלקיאס ואדא קיל‬
‫חלו ליס בחלו חלו באלקוה חלו באלפעל ומא‬
‫ישבה הדא אנמא יסתעמל איצ’י טריק אלקיאס‬
‫ואמא אלצ’מאדאת ואלאדהאן וגמיע מא יעצב‬
‫‪196‬‬
‫וירבט ויעאלג במתל הדה אלצ’רוב פאנמא הו‬

‫מן עמל אלאידי ואקעא" תחת אלקיאס לאנה‬
‫תאן לה וקד ערף אן הדא אלכתאב קד אחתוי‬
‫עלי הדה אלאגזא ואמא אלאטבא אלדין מן זמאן‬
‫גאלינוס פאנהם אסתעמלוא אלעלם אכתר מן‬
‫הדה אלקסמה פאמא אלדין כאנוא פי זמאן אבקראט‬
‫פאנמא אסתעמלוא אלעלם עלי אנחא תלאתה‬
‫אמא אן יקלוא אלאלפט’ ויגמצ’וא אלמעאני‬
‫]‪ [4b‬ליסתעאן עליהא באלפהם ואמא אן יכברוא‬

‫אללפט’ ויטילוהא וירידוא בה כלאף מא יט’ן פי‬
‫אלאול ואמא אן יכלטוא הדין וקד וצ’ע‬
‫אבקראט כתבא עלי הדא אלנחו ככתאב‬
‫אל… ואלמפאצל וכתאב אפידימיא‬
‫וקד וצ’ע כתבא פי אמראץ’ מרת’ ]מרת[ בה ועאינהא‬
‫וכתב כתבא אכר תשבה הדא מן כתרה‬
‫אטנאב אלקול ותטוילה ואדא כתר אלקול‬
‫לם יכד יחתאג אלי תפהם כתר ]כתיר[ פאמא‬
‫כתאבה פי אלטביעה וכתאב אלאכלאט‬
‫וכתאב אלפצול פאנה אעמק הדא ואגמצ’הא‬
‫גדא לאנה וצ’עהא באיגאז ואכתצאר פאמא‬
‫כתאב אלכשך ואלדי קיל פי אלקרוח אלראס ואלדי‬
‫פי אלקרוח אלתי פי גמיע אלגסד ואלמקאלה‬
‫אלב’ ואלו’ מן כתאב אפידימיא פאנהא‬
‫מכתלטא ]מכתלטה[ מן אלנחוין כלהמא אלא אן פיהא‬
‫כלאמא" לא יחתאג אלי פהם כתיר וקד יגב‬
‫אן תעלם אן הדא אלכתאב הו מן אלכתב‬
‫אלתי וצ’עת באיגאז פאנמא וצ’ע אלחכמא‬
‫כתבהם עלי הדא אלנחו ליס לשרארתהם‬
‫וצ’נהם באלעלם לכן יסתעמל כל אמר…‬
‫ולא יתכל עלי אלכלאם אלבין ולאן אלכסל‬
‫אדא ורד עליה אלוגיז מן אלקול מל ושגלה‬
‫אלשב ]אלשכ[ ענה והדא חכים לא יעד פי צנף‬
‫אלחכמא פאנמא וצ’ע אבקראט כתבה‬
‫עלי הדה אלנחו לאנה ממא ינבגי אן יפעלה‬
‫וקד קאל אבקראט ליוצ’ע ללנאס עלי מקאדירה‬
‫וקאל אפלאטון לא ינבגי אן יגתמע אלצ’דאן‬
‫פלא ידנא אלטאהר מן אלנגס לאנך מתי‬
‫כלטת מאא" עדבא" במא פיה חמאה אנתנת‬
‫]‪ [5a‬אלמא ולם תצלח כדלך אן אנת עלמת‬

‫אלזלאן ]כתוב מעל ‪:‬אלטלאן[ אסתהזא‪ ..‬בך וסכר מנך ובאלגמלה‬
‫‪197‬‬
‫לא ילתד באלנור אלא מן כאן צחיח אלבצר‬

‫ולדלך אהל אלחרץ עלי אלטלב)?( ואלתפהם‬
‫אדא יט’פרוא במתל הדא אלכתאב גדלוא וסרוא‬
‫בהא למא פיהא מן אלאסראר אלכפיה ואל‬
‫ ואנא אביין‬.‫אמור אלגאמצ’ה אלמסתורה‬
‫לך האהנא טרפא" תערף בה אלגנס ואלצור‬
‫והו אנה אנמא יסמי אלגנס ללשי אלדי תכץ‬
‫פיה אשיא כתירה מתבאינה אלטבאיע בעצ’ הא‬
‫מן בעץ’ לאן אלתור ואלפרס ואלחמאר‬
‫תשמלהא ותעמהא אלחיאה ותתפצל ותתמיז‬
‫בפצולהא וכדלך אלתין ואלבלוט ואלכבז‬
‫תדעא הדה כלהא גדא" וכל ואחד מנהא‬
‫ינפצל ען צאחבה פאפהם הדא אלא תגלטן‬
‫ והדא‬.‫ כמל צדר סרכס‬.‫פיה‬
.‫אבתדא כתא’ אבוק’ שרח גאלינוס‬
Translation
We present a fairly literal translation of the text. In those places

where the text is abstruse, but we feel that we can suggest its inten-
tion, we offer a paraphrase within square brackets. Lacunae in the
manuscript, and obvious corruptions or passages that could not be
translated have been indicated by an ellipsis (…). Other than those
paraphrases, and the analysis offered in the introductory essay, we
add no annotations.
[1a] This is Hippocrates’ book On Nutriment with the commen-
tary of Galen, translated by the translator Yusuf ibn Khaj [licet al-
Khuri?]. Sergius said: Let us begin, with the help of God, to write
what Galen explained concerning Hippocrates’ book On Nutriment.
He [Sergius] said: Dear brother, since you asked me and requested of
me many times that I translate for you some of the books of the
Greeks into Syriac, but especially the treatise that Hippocrates com-
posed about foods, and this was on account of your love for all
knowledge that the soul requires, I assented and will not deny you
that which you asked for and requested; for it is one of the things
that are required.
I did not want to do any of this, because I know the censure and
rebuke that will come upon me due to the envy of those who are not
satisfied by anything other than amassing money. This is so because
they do not seek the same gain as those who are in search of knowl-
edge [and] who do not have material possessions that might be a
198
substitute for it [i.e. seeking knowledge]. I also beseech you not to

neglect seeking beneficial knowledge…
[1b] …inquired into the books that he wrote in all their varieties.
Among them, On Seasons, and among them, On Temperaments, and
among them, On Blood, Semen, The Composition of Organs, and The
Formation of the Child; On Water, Air, Earth and Countries; On Set-
ting Bones and Bandages; On Ulcers; and On All Types of Pains; Amel-
iorating Uteri; On the Nature of Man. After he had written this, he
realized that there still remained the loftiest and most excellent of
them, namely, that which cannot be done without, and no one is
free of need of it: that is to say, nutriment. Maintenance (reading
qawam) of every living thing depends upon it. From it, the embryo
receives its coming-to-be and growth; by its means, the organs mani-
fest their actions; through it, health is sustained, and the substance
of man endures until the end of the life-span to which it can reach.
An animal cannot endure if it is worn out due to a lack of food.
Therefore, (?) he composed this book, making it a treasure or a pre-
cious object of value. Since it is the custom of the Greeks, when be-
ginning commentaries to the books of the ancients, to lay down at
the beginning of the books, eight headings, by means of which will
become clear their aim, their causes, their strength, their limit, their
author, their number (?), how they were executed, where they are to
be read [i.e. place in the curriculum?], by means of which [informa-
tion] they encourage people to read it, once they have been given to
hear what wonderful things, hidden secrets, and enormous benefits
are in it.
Let us therefore talk about the explanation of these headings in

connection with this book, so that we may know what each one of
them makes incumbent upon us. These are the headings: The first, it
informs us of Hippocrates’ opinion. The second
[2a]…
…
The sixth, regarding every part into which this book is divided.
The seventh will make clear to which of the branches of knowledge
this book belongs. The eighth will declare which type of instruction
Hippocrates employed in this book, that is, whether he wished to ex-
plain in this book the cause [?, i.e. final cause] of food, the purpose
of its kinds, their transformations, powers, action, benefits and
harm, but he had no choice but to go beyond its boundary [strict
boundary of this field, as explained in the next sentences, e.g., deal-
ing with bone-setting and embryology]. And the doctrine concern-
199
ing the types of maladies that come about on account of an excess of

food, or on account of its poorness, its quality, its quantity, in its
clearest [formulation?]. And the doctrine concerning the setting of
broken bones, the number of days [required] for them to heal, and
on the coming to be of the fetus and its conception. And on its
movement within the womb, the completion of its perfection; and
he explained its coming out [of the womb], along with the number
of days, [he explained] each of these. He then went back and com-
pleted this topic (?, i.e. alimentation). These are the benefits pro-
cured by the person who reads this book. For we see many people
who, when they take sick, not only do they not take drugs — they
cannot bear to hear them mentioned! Due to his knowledge of the
types of food, the doctor is able to accommodate their desire, to feed
them what is needed, so that they are helped by it. Indeed, were they
given to drink all of the appropriate drugs for feverish maladies, pain
of the liver, spleen, head, stomach, and all of the internal organs, but
they had not the food appropriate for these, their pains would stay as
they are. Even though drugs cure from the outside, they do not pen-
etrate [9ab] [so that?] their power [may] set in. However, food that is
fitting transforms into laudable [i.e. salubrious] chyme. If it is ap-
plied to … pain, it is beneficial … as necessary. Therefore, health is
preserved by means of the types of foods. In addition, drugs are not
proper for sound bodies, as Hippocrates said in the Book of Apho-
risms. Therefore, one ought to count this book as the most excellent
of the books on medicine, because, if the physician knows it and is
able to extract from it useful things, he does not need anything else.
Moreover, there are some indigent among the populace, and those
who cannot take drugs because of the trouble involved (li-ma’uniha,
lic. ma’unatiha?), their high cost, or their unavailability, and they can
make do with food instead of drugs. Also in case of a mistake: if it
happens in connection with drugs, the patient can hardly escape
[death or serious injury]. However, should it happen in connection
with foods, then, the patient rarely will perish. Voyagers on land and
at sea, who cannot [obtain] any of the other types [of remedy], may
profit from this book. Moreover, if the person comes to a place
where the drugs that he requires are not to be found, then he will
necessarily have need of foodstuffs. Therefore, it is requisite upon the
physician to have knowledge of this book.
Indeed, praise is due to the physician who does without drugs, but
replaces them with foodstuffs: that is the skilled physician! Perhaps it
would even be better and more excellent for the physician to spend
his effort upon foodstuffs than to spend his effort upon that which
200
has [both] alimentary and pharmaceutical [properties], were that

possible; for [otherwise] the cure will be effective only insofar as it
possess both of these powers. And perhaps it would be better to
spend his efforts on things possessing both alimentary and pharma-
ceutical [properties] than it would be to spend his effort upon what-
ever has a purely pharmaceutical power. For this reason the physi-
cians of [al-hakil] were praised, since, whenever they found the food-
stuff [that had the sought-after therapeutic properties?], they had
what they wanted, [3a] and even if it fell short of the [desired] ac-
tion, they exerted no effort at all on drugs. [This last sentence is
clearly mangled, and we have no idea what Sergius means to say
here.] Therefore, the physician should not habituate bodies to drugs
when illnesses set in them. They have praised the physician when he
does with foodstuffs what another cannot do with drugs with regard
to curing illnesses and taking care of the sick.
Now, then, it is known that this book is not just beneficial like
the other books, but rather it is something that one simply must
possess. The heading of this book is written, ‘On Nutriment, by
Hippocrates’. He [Hippocrates] disliked prolixity. Even had he
wanted to relate the types of food, and their individuals are very
many [i.e. the very many classes of individual foods], he would
hardly have been able to encompass them. Who indeed is capable of
describing the classes and individuals? Therefore, Hippocrates, who
knew medicine, and, along with medicine, the science of how to
express oneself [literally: the science of speech], did not apply this
definition, but rather stamped this book with this name. He called
it by this name only because people, many in number, composed
many books, even if (?) they had not just given names to the begin-
nings of their books and names to their books. [There are many
books on foods, and it is important to know that this one is by
Hippocrates?] He meant by it, that which most people recognize,
but they do not know what they require of it. [People will immedi-
ately know that this is a book on an important subject that probably
conveys information that is much needed.]
You ought to know that all of Hippocrates’ books were composed
in two different ways. One of these is to inform about natural things,
that is, humours, powers, the fetus, the nature of man, chymes, the
types of food, and the like. The other way concerns things that are
unnatural [i.e. beyond the bounds of healthy nature], such as the
types of fever, broken bones, prognosis for perdition or sound health,
the fluctuation [i.e. inversion] of temperaments, and, in general, all
things similar to these. Accordingly, he composed this book accord-
201
ing to these two ways, concerning natural things, and those that are
unnatural [literally: ‘lie outside of nature’]. Healthy people…
[3b]…
… Therefore, this book is the fourth of Hippocrates’ books, after
which he [the medical student?] reads [advanced texts?]. Know that
he [it?] was called by the name Hippocrates…He composed books,
some of which are almost like others. This is the reason why the
commentators had some doubts and could hardly distinguish the
one from the other.
As for us, we say — and it is the truth — that this book is by the
Hippocrates who wrote On Prognosis, the Aphorisms, On Bone-Set-
ting, On Joints, Epidemics, and Acute Diseases. We know this, because
of its conciseness and wariness of prolixity, along with the depth of
understanding and the rarity of anything superfluous, yet concerning
this [issue], there is argument that is more correct and more firm.
This is the testimony of Galen and his commentary to it, among all
of them.
This book is divided among three issues. The first is the nutrition
of the general populace, their food and its action, the organs that are
nourished by it, its benefit and harm. The second is the classification
of the types of food. Included in this is the clarification, which of
them are moist and which dry, knowledge of smells and tastes; the
classification of the organs that are harmed by them or benefit from
them; and the diseases that are produced by them [foodstuffs]. The
third is the formation of the embryo, how many days it takes for it
to be completed, how many days it moves within the womb, and in
how many days it will be birthed, as well as the explication of the
difference between male and female. He then went back to the gen-
eral topics, describing the types of foods, and explaining their ben-
efits and harms. He could…
[4a]
…
…from among those who wrote books …he explained
…three parts of science. For part of science is that which points to
logic and intellectual understanding. Part of it points to work with
the hands [surgery]. The third concerns making use of both kinds.
You will find most of Hippocrates’ books to be according to this de-
scription. One (?) of them is On the Nature of Man, and some are
different from that book, [such as] [On] Fractures and Bone-setting.
However, that which is comprised by On barley water indicates two
kinds. For that book comprises knowledge of fever and reasoning
202
about it on the basis of its actions, then the judgement about them
as to what reasoning mandates, and then, after that, mention of the
barley water, its preparation, and the other things that are do be
done with the hands.
Now this book has two kinds of science and practice. When he
talks about food and its utilization for nourishing the organs, causing
them to grow during the period of growth, and replenishing that
which wastes away of them, during the period that that [is applica-
ble, i.e. maturity], [all of ] this is to be done on the basis of investiga-
tion and reasoning. When ‘sweet’, ‘not sweet’, ‘sweet in potentia’,
‘sweet in actu’, and the like are mentioned, this too is to be done by
way of reasoning. However, the pastes, oils, and all the like that [are
used to] bandage, bind and cure, they are the work of the hands,
[though] falling under [the rule?] of reasoning, for it [reasoning] is
secondary to it [practice in these matters].
It is known that this book encompasses these parts. The physi-
cians at the time of Galen made a greater division of science [or:
knowledge]. However, those who were in the time of Hippocrates
made use of knowledge only according to three divisions — either
because they were sparing with words but deep in their concepts
[4b], so that one would rely upon one’s understanding [as in the
English expression, ‘a word to the wise…’], or else they were prolix
and lengthy in their words, intending the opposite of what was
thought in the first [case]; or they combined both of these.
Hippocrates […] composed books according to this method, such
as On [Fractures] and Joints and Epidemics; and he composed a book
on diseases that passed by him and he inspected. He composed an-
other book [perhaps: other books], similar to this with regard to
their great prolixity and length. Since their discourse is lengthy, not
much understanding is needed [since everything is set forth at
length]. However, his On Nature, On Temperaments, and Aphorisms
are much deeper and much more abstruse, because he made them
concise and succinct. But as for On Barley Water, and what was said
about ulcerations in the head, and ulcerations in the entire body, as
books two and six of Epidemics, they comprise a combination of
both methods; however, they contain a discourse that does not re-
quire great understanding.
You ought to know that this book [i.e. On Nutriment] is one of
the books that he wrote with conciseness. Scholars did not compose
their writings according to this tendency out of their stinginess and
meanness with regard to knowledge. Instead, they did all of this [de-
liberately], not relying upon a lucid discourse. For when a lazy per-
203
son is given a succinct discourse, he becomes impatient, and his mis-

givings divert him from it. Such a person is not to be counted in the
file of scholars. Hippocrates [composed] his books according to this
method only because that was what he had to do. Hippocrates said,
‘Let books be written for people according to their capacities’. Plato
said, ‘The two opposites should not be joined together. Let not the
pure become contaminated by the impure. For when you mix sweet
water with water that contains mud, you putrefy [5a] the water, and
it is not good anymore.’ If you teach the wicked (†alan!? licet
zallam??), he will mock and ridicule you. In general, only those pos-
sessing sound eye-sight benefit from light. For this reason, when peo-
ple who are eager to seek out and to understand get hold of this
book, they gladden and rejoice in it, because of the hidden secrets
and concealed, profound things that it contains.
I will clarify a little here, and you will thus know the genus and
forms [perhaps one ought to read Òanf, ‘species’]. That is, he calls that
thing ‘genus’ which characterizes many things whose natures are dis-
tinct, one from the other. For the ox, the horse, and the ass all have
in common and share in vitality [being animals is more precise here],
but are different in species. So also the fig, acorn, and bread are all
called (?) food, but each one is different from the other. Know this,
so that you do not become mistaken in it.
Sergius’ introduction is complete. This is the beginning of the
book of Hippocrates with the commentary of Galen.
204

ORIGIN AND CLASSIFICATION OF

THE ANCIENT SOUTH ARABIAN LANGUAGES
ALESSANDRA AVANZINI
UNIVERSITY OF PISA
Abstract
ASA (= Ancient South Arabian) documentation is testimony to a

lengthy linguistic history in southern Arabia which predates the earli
est written attestations. The hypothesis attributing the origin of ASA
culture to immigration from the north is hard to endorse. The QAT
(= Qatabanic) verb system and ASA more in general have strong par
allels with the verb system of the north west of the second millen
nium. Just as the hypothesis of a recent wave of immigration to south
Arabia is open to debate, so must the general idea of an ASA belong
ing to central Semitic as opposed to archaic southern Semitic be re
examined.
The states of south Arabia originated in one of the crucial transition

periods in the history of the Near East between the late Bronze Age
and the Iron Age. Recent historic debate on the various models ex-
plaining the formation of the early first millennium state elsewhere
in the ancient Near East has produced significant results but few
studies have faced the issue of the formation of the state in southern
Arabia.1
Notwithstanding slightly different standpoints, two scholars
Nebes (Nebes 2001) and Garbini (Garbini 2004, Garbini 2006:
235–44)2 have recently suggested a large wave of migration of ethnic
1
Despite south Arabian culture being rightfully within the history of the an
cient near east, the history of south Arabia at the beginning of the first millennium
has often been neglected by the many general treatises on this period. This lacuna is
why the need voiced by Liverani (Liverani 2005 b: 171) for some mention of the
history of ancient south Arabian kingdoms in the history of the near east is so im
portant.
2
The standpoints of the two scholars are assuredly not identical. That of N.
Nebes is much more subtle and problematic, rich in interesting linguistic observa
205
CLASSIFICATION OF THE ANCIENT SOUTH ARABIAN LANGUAGES
groups at the end of the second millennium from Palestine towards

Yemen as the main factor in the birth of ASA states.
Taking a migratory model as the main or even sole cause of the
birth of a new culture in Southern Arabia is debatable from a general
historical standpoint. Much of the Near East historiography which
has emerged in recent years has deemed the exogenous model for the
formation of a culture methodologically obsolete. The establishment
of the land of Israel is certainly no longer linked to the arrival of the
twelve tribes in the promised land (Master 2001; Joffe 2002).
It may not be possible to consider the origin of ASA culture
wholly from within a process of internal development with no exter-
nal influence at all. Clearly, pastoral groups already present in the
area, turning into settlers, and some infiltration from the fringe of
the desert were factors not only in the formation stage of south Ara-
bian states but also during the whole of ASA history.
I assuredly do not reject the importance of contact with the states
of the north — suffice it to recall the ninth century ASA ibex from
Hama (Scigliuzzo 2003; Avanzini 2005a). I also agree with what de
Maigret said regarding the significance of contact with the Assyrian
empire (de Maigret 1999), just as I fully agree with some of
Liverani’s remarks (Liverani 2005: 236) on the positive effect that
the Assyrian empire had on the peripheral areas (while it was assur-
edly much more destructive on nearer areas). All these, however, do
not offer a scrap of proof of a recent arrival of populations.
The crucial aspect which seems to be particularly doubtful is to
think that south Arabian culture came exclusively from the outside
with a group of shepherds and nomads from the north. Take, for ex-
ample, real immigrants like the Philistines in Palestine whose cultural
traits were perfectly coherent with their foreign origin. There was
nothing cultural that was ‘wholly’ external in early first millennium
southern Arabia but what there was had already been wholly ab-
sorbed into a new, typically south Arabian, cultural context.
The hypothesis of ASA culture stemming from an exogenous
source was also favoured by archaeologists involved in south Arabia,
at least until a few years ago. Yemen lacks the great culture of the
Bronze Age — a palatine society similar to that of Palestine and
Syria. The scanty Bronze Age settlements investigated by de Maigret
(de Maigret 1981; 1990) did not seem highly credible as the fore-
runners of ASA culture either for the cultural level of technology or
tions while Giovanni Garbini’s is more provocatively extremist as is his argumenta

tive vigour.
206
for the number of inhabitants. The results of the mission to Sabir

pointed in the same direction — they seemed to indicate relations
with Africa rather than with the future Sabaean culture (Vogt and
Sedov 1998).
Only in recent years has the study of the Bronze Age culture in
south Arabia seemed to open a new horizon. The Chicago mission to
the high plateau has highlighted a different cultural state of affairs
from that which had been hypothesized earlier by de Maigret
(Wilkinson, Edens and Gibson 1997; Wilkinson 2005); ASA letters
carved in the pottery of Sabir indicate a possible link with the later
period. The architectural decorations in the Tihama megalithic site
of al-Midamman are absolutely identical with the decorations in a
number of temples of Jawf (Keall 2004: 47–9, figs 15–18).
Archaeologically, the issue remains unsolved,3 but in the Bronze
period there was assuredly an autochthonous culture in the whole of
Yemen.
I feel, however, that the linguistic data are totally irreconcilable
with the notion of an exogenous formation of ASA culture.
We all know that we cannot reconstruct peoples’ histories just by
studying the languages they spoke. So although the historical signifi-
cance of the linguistic datum is not immediate, it is important for
reconstructing the process that produced it. A general historic model
must take account of linguistic data; and these, in turn, must obvi-
ously fit the hypothetical model. Otherwise the model must be con-
sidered invalid.
However, as already pointed out, the model of an exogenous de-
velopment of ASA culture came from two linguists, G. Garbini and
N. Nebes.
Garbini’s theory of a recent arrival of south Arabian languages in
Yemen (Garbini 2004) is closely linked to his firm belief in the short
chronology for the beginning of ASA culture. Despite recent signs of
accepting the long chronology (Garbini 2006: 240), he has not alto-
gether abandoned his idea of several groups of nomads from Meso-
potamia stopping first in Palestine and then happily arriving in that
desolate area bordering the desert at the foot of the Yemeni high pla-
teau to find their Arabia Felix.
Garbini’s hypothesis is largely based on the history of the script, as
was that of J. Pirenne earlier. Personally, I am not convinced that the
3
‘The evidence is not yet sufficiently developed to identify the causes of these
late second millennium developments, even on the question of immigration from
northern Arabia (an interpretation favoured by many scholars of South Arabian ori
gins) versus emergence from indigenous antecedents’ (Edens 2002: 52).
207
alphabetical order of southern Semitic writing was created in the

north-west Semitic area,4 but whether this is true or not, it is a sec-
ondary issue. Even if we accept the alphabetical order as a north-
western creation (Garbini 2006: 235) it does not mean that south
Arabian culture was brought by groups who came from Palestine.
Nobody has ever attributed the origin of the Greeks to a mass move-
ment of peoples from the East despite the alphabetical order of
Greek clearly deriving from Phoenician.
There is also, in Garbini’s reconstruction and sometimes in re-
marks made by archaeologists, a broad and (in my view) sometimes
insufficiently clear use of the term ‘north Arabian’. 5
What I also find completely mistaken in Garbini’s latest recon-
struction is that writing has been taken as a criterion for defining an
ethnic group. This is completely wrong for a language, and all the
more so for writing. In Garbini’s view ‘north Arabs’ and ‘south Arabs’
were two ethnic groups who had wandered through the desert, be-
fore both arriving in South Arabia.
My opinion of Nebes’s standpoint is rather that he gave too much
priority to Sabaic documentation over that of other south Arabian
languages,6 and that his idea was essentially based on a general prob-
lem of the classification of Semitic languages.7
From studies by V. Christian (Christian 1919–20: 729–39) and

J. Cantineau onwards, and up to the recent definition of Central
Semitic, one morphological isogloss has been stressed almost to the
point of becoming the only one supporting this classification of
Semitic languages: Hetzron considers the imperfect yaqtulu as an in-
4
The history of south Semitic writing is still to be written. Some interesting
parallels between the writing on the oldest ASA sticks and Dedanite writing were
suggested by M.C.A. Macdonald during the 10th Rencontre sabéenne in Saint
Petersburg. Assuredly, ASA writing has no clear genealogical relationship with
north Semitic alphabetical writing but it does have a ‘link by inspiration or anal
ogy’, cf. Healey 1990: 61.
5
M.C.A. Macdonald (Macdonald 2000) has provided an accurate definition of
this and what the exact subdivisions of the term are from the epigraphic stand
point.
6
One of the main steps forward in south Arabian studies in recent years has
been to consider ASA as a number of separate languages forming a linguistic group
instead of a single language broken into a series of dialects.
7
For an in depth discussion on Nebes’s hypothesis see Mazzini 2005: 223 5.
208
novation of Central Semitic, as compared with the older yVqattVl,

preserved in Akkadian, Ethiopian Semitic, MSA (= Modern South
Arabian).
Hetzron’s contribution of removing Arabic from its role of an ar-
chaic language has been crucial in defining Central Semitic and
therefore for classifying Semitic languages. A geographically marginal
archaic Semitic made up of Akkadian, MSA and Ethiopic was a
counterpoint to Central Semitic.8
This whole reconstruction is based on one isogloss alone9 and,
what is more, one that is charged with uncertainty. The isogloss has
been variously considered by scholars. Suffice it here to consider the
contrary stance of Voigt (Voigt 1990), for whom yVqattVl is a pri-
mary and hence constitutes a potential isogloss, and Marrassini
(Marrassini 2003: 373–5);10 the latter refers to the studies by Cohen
(Cohen 1984: 62–78).11
Nebes reached the correct conclusion that a yaqattal form was
missing in ASA (Nebes 1994, Nebes 1997),12 and he then accepted
the idea that Sabaic was distinct from southern Semitic, which re-
mained archaic and conservative. The tardy arrival of ASA languages
in the archaic enclave of southern Semitic languages seemed the
most convincing explanation for the absence of yaqattal. Hence the
model of a migration of peoples from the north conjectured by
Nebes.
8
ASA was also included in Central Semitic (see Voigt 1987).
9
Appleyard’s obvious embarrassment (Appleyard 1996, Appleyard 2002), in ac
cepting the incongruence between this isogloss that distinguishes ASA from south
ern Semitic and the other many isoglosses that they share, can also be seen in his
interesting articles on this issue.
10
For Marrassini, although the verb form was primary, it would be a ‘conserva
tion’ and not an ‘innovation’ and thus useless for classifying south Semitic lan
guages.
11
Personally, I find D. Cohen’s theory perfectly plausible according to which
yVqattVl is an independent development in MSA and Ethiopic. Cohen reached this
conclusion after studying the dynamics by which recent languages (for example
neo Aramaic) developed the verb system. See Knudsen 1998: 5 6: ‘The crucial
problem in the discussion has been the comparative assessment of the relationship
between the Akkadian present and the Ethiopic and Modern South Arabian
imperfects. A majority of scholars maintain that these forms are genetically related
and that they reflect a common Proto Semitic form, whereas a relatively large
minority of scholars dispute that view. […] In recent years the majority view has
prevailed in the literature and maybe Marrassini is right that ‘one has got rid too
hastly of the Rundgren hypothesis on the réemploi de l’intensif (Rundgren 1959)’
(Marrassini 1991: 1020)’.
209
But the study of ASA languages overall in my view makes Nebes’s

hypothesis historically unlikely. The issue has to be framed in a new
perspective.
The answer to the question whether or not ASA forms a linguistic
family must obviously be affirmative. Precisely because languages are
part of the history of a culture and since it is archaeologically easy to
acknowledge the existence of an ASA culture in order to prove that
there is not an ASA linguistic family, we are obliged to seek marked
elements showing a typological difference similar for proving that
Etruscan was within the family of languages of ancient Italy.
Not only does ASA have the same characteristic script13 but there
is also a series of phonetic,14 morphological,15 syntactical16 and lexi-
cal17 isoglosses which mark ASA as being a family within Semitic.
12
Some of my earlier remarks (Avanzini 1991) are certainly to be corrected on
the basis of Nebes’s studies.
13
Using the same monumental script not only implies sharing the same writing
technique but also the same textual models, text syntax throughout the whole of
ASA culture.
14
The best know isogloss common to every ASA language is the three unvoiced
sibilants.
15
The pronoun system with the same ∂ base for the relative and demonstrative
pronouns.
The determination of the noun made by the suffix n is the isogloss that charac
terizes the ASA family within the languages of Arabia as was pointed out by
Beeston (Beeston 1982: 178 86) and repeated by M.C.A. Macdonald (Macdonald
2000) in his classification of the languages of Arabia. In my opinion, the laryngal h
in the HAD suffix hn is a mater lectionis despite what Tropper believes (Tropper
2001: 11 13). The m suffix in the substantive which is also defined in Stein’s
grammar as a non determination suffix, has a complex history. It was originally
used for adverb forms and for indicating subordination as well as for the syntactical
dependence of elements already marked as definite. See the interesting remarks in
Gzella 2006 on determination in Phoenician.
16
Syntax: the order of the words (verb subject object) is the same in all ASA
languages; the syntax of the prepositions is the same, but different compared to
Arabic, Ethiopic, MSA while, certainly by no coincidence, similar to Ugaritic. Cf.
Pennacchietti 1974 based on the V. Brøndal’s theory of prepositions.
17
The study of the ASA lexicon is also extremely important for the hypothesis
suggested here of a lengthy history of ancient south Arabian languages prior to its
earliest written attestations. For example, the juridical lexicon in ASA languages has
a great ‘technical’ wealth and specificity which only evolved within ASA and is not
found in other Semitic languages. In historical terms, too, a developed detailed ju
ridical lexicon has interesting applications and in my view presumes a centralised
bureaucratised state power in at least some periods of the long history of South
Arabia.
210
As well as isoglosses common to all ASA languages there are some

which are exclusive to one or another of the ASA languages. The fol-
lowing are some of these but the list could certainly be longer.18
Morphology:
the suffix pronouns, in h in SAB (= Sabaic), in s1 in non-SAB lan-
guages;19
the construct dual of the substantive: SAB y; QAT w;
the plural of the relative pronoun formed in QAT and in HAD
(= Hadramitic) on the same basis as the singular instead of the ’l
basis (Mazzini 2006);
the suffixes of the causative in h- in SAB, in s1 in non-SAB languages;
l-yf ¨lwn in QAT vs. l-yf ¨lw in SAB;
the infinitive suffix: SAB n; QAT m (for the rules of morphological
and syntactic use of the two suffixes, see Nebes 1988; Stein 2002);
the feminine plural of suffixed conjugation: SAB: f ¨ly QAT: f ¨ln
(Robin 1983: 176 82); perhaps also the feminine plural of the pre-
fixed conjugation: SAB: yf ¨ln, MIN (= Minaic): tf ¨ln. Beeston in
his grammar (Beeston 1984: 15) reconstructs a yf ¨ln form and Stein
(Stein 2003: 178 9) tf ¨ln; some examples seem to me to lend
greater credibility to Beeston’s hypothesis of reconstruction (Avan-
zini 2006 b).
Syntax:
the hypothetical phrase (Sima 2001);
the relative phrase: the relative clause in Qatabanic can be expressed
through an asyndetic phrase whose antecedent is a mimated sub-
18
Even the matres lectionis were used differently in the various ASA epigraphic
corpora. SAB, especially in its most archaic phase, used a very limited number of
matres lectionis (Stein 2003: 41 7), while there were many more in non Sabaic ASA
languages and in late SAB (Avanzini 2006c: 257 8). This is very interesting for the
cultural history of ancient south Arabia. On the one hand, epigraphic texts are very
homogeneous both in their graphic layout and in their palaeographic development,
creating an epigraphic ‘style’ common to texts in every ASA language and on the
other there are graphic phenomena such as the different way of writing vowels
which points to different writing schools in the single regions.
One interesting aspect is the habit of writing vowels in non SAB languages
using two graphemes. For example, some plural suffixes in the substantives: QAT:
hw; HAD, MIN: hy.
19
In QAT and HAD two attached pronouns s1ww s1yw attached to a dual or
to an external plural are attested. It is not easy to explain the long script, but it as
suredly depends on the accent a probably pre suffix stress and consequent pre
suffix vowel length (Gragg 1997) as the following HAD example: Shabwa
chantier V, 1975 6 7 (Pirenne 1990: 76): s1qnyt ’Ì∂ ¨lh s1ww h wfy grb s1 ‘the dedi
cation that He took over him for the safety of his person’ shows with the long suffix
after the preposition ¨lh (the laryngal h surely indicates a vowel).
211
stantive without the use of the ∂ base pronoun (CSAI I, 197=R

3688, 2: kbrm b-ykbr ‘the official who acts as the official’) (see:
Nebes 1997; Mazzini 2006), a syntactic construct unknown in the
other ASA languages.
So we have a group of common isoglosses and others differenti-
ated among the ASA languages.
There can only be two explanations for this. Either different lan-
guages were made highly homogeneous by dissemination of com-
mon traits from a common centre of great cultural prestige which
was probably Saba. The possible ‘sabaeization’ of south Arabian cul-
ture is an important historical phenomenon, but in my view the
linguistic data proves that it was secondary. I have sought to prove
how the onomastics, which at first sight could have seemed Sabaic in
all pre-Islamic Yemen, was in point of fact archaic and not Sabaic
(Avanzini 2006a). Or there was a long common period upstream of
the written attestation of ASA languages, a proto-ASA phase within
Yemen, from which the historically attested languages emerged sepa-
rately — we should certainly not imagine ASA languages to be de-
scended from a single, limited common proto-language, but rather
languages and dialects in contact within an area. A proto-ASA phase
summarizes our assumptions, views, and starting points if nothing
else. It is a purely heuristic model.
One could think of a linguistic stratum common to all ASA lan-
guages from which particular features have been selected to become
the ‘distinctive’ part of the system of ASA languages. On a general
level, this hypothesis overturns the distinction between the ‘most re-
cent’, ‘latest arrived’ Sabaic and the other more ‘conservative’ ASA
languages. In many ways Sabaic is probably the closest language to
second millennium Semitic and the least subjected to neo-
morphologization phenomena (Avanzini 2005 c).
One interpretative key for reconstructing the classification of ASA is

its verb system.
I found it useful to begin the study of the ASA verb system start-
ing from the non-Sabaic languages such as, for example QAT
(Avanzini 2004), in which the neo-morphologization of a form b-yf ¨l
makes the system clearer.
In his grammar book of 1962, in which he goes into the matter at
greater length, Beeston states with regard to the difference in the
212
QAT verb system between the use of yf ¨l and b-yf ¨l: ‘b- is character-
istic of the indicative in QAT, while a jussive (and perhaps also a sub-
junctive) use is characterized by the absence of b-. Such a formula-
tion would bring the QAT imperfect usage very close to that of the
Syro-Palestinian dialect of Arabic’ (Beeston 1962: 24–5).20
A study of the QAT corpus has shown that no modal opposition21
can be surmised for the yf ¨l vs. b-yf ¨l opposition, but the b-yf ¨l
b-yf ¨lwn (pl.) forms are used to specify the present-future (Avanzini
2005b), in relative clauses in particular.
The interesting point in general comparison is that QAT has a
morphological opposition within the verbal prefix forms between
two masculine plural terminations –w / -wn, the first used for the
preterite: CSAI 1, 203 = RES 3858 4–5: w-ygb’ w-h[†]b Y∂mrmlk ’byt
w-’r∂ty w-’qny Qtbn ‘Y∂mrmlk gathered and returned the houses, the
lands, and the possessions of Qataban’. CSAI 1, 115 = Ärbach-
Sayun 1 l. 3: w-yhrgw w-s1lqÌ Î∂rmwt ‘and then they scattered death
and destruction on the Îa∂ramawt’, the second for the imperfective:
CSAI I 205 = R 4337: w-l-yÌr† Ìms1y wrqm l-mlk Qtbn w-¨hr S2mr
ys1mÂ’wn ‘must pay 50 pieces of gold to the king of Qataban and the
magistrate of S2mr whose task this is (whose task it was and will be
to receive payment of the fine)’.
QAT later morphologized a b- prefix for the form with the plural
in -wn to indicate the present-future.
In QAT we would have proof of two prefixed forms for the past: a
short form for the preterite: ‘he did’, and the other, long form for a
narrative imperfective action ‘he was doing’.
The QAT verb can therefore be reconstructed in accordance with
the following pattern:
f ¨l w-yf ¨l yf ¨l b-yf ¨l
w-yf ¨lw (pl.) yf ¨lwn (pl.) b-yf ¨lwn (pl.)
(suffix)- preterite (prefix)-preterite narrative-imperfective present-future
Not only QAT but all ASA languages have a prefix form for the
past. N. Nebes and P. Stein are assuredly correct in finding some of
Tropper’s conclusions hurried (Tropper 1994): there is no proof of a
20
Description was summarised in Beeston 1984: 64: ‘The indicative imperfect
normally shows a prefix b , e.g. R 3688 2 kbrm b ykbr. Without this prefix the im
perfect is usually a jussive, or conditional, or in some other way non indicative’.
21
This same modal opposition is probably recent and is not found in archaic
languages; cf. Petracek 1981.
213
yf ¨l (preferite), yf ¨ln (present/energic) opposition in Sabaic, but there

is an obvious past significance for the prefix form in both old and
middle SAB.
I mention two examples from RES 3945, the inscription of
Karibil Watar, from the beginning of the seventh century: RES 3945
14 w-ygn’ gn’m ‘and he built a wall’; RES 3945 16: w-hgrn Ns2n
yhÌrm bn mwf†m ‘and the city of Nashshan was annihilated with
fire’.22
There are many attestations of the forms w-yf ¨ln/w-yf ¨lnn to indi-
cate the past in the dedicatory inscriptions from MaÌram Bilqis.23
In the many MIN legal texts and in HAD there are many attesta-
tions of yf ¨l forms for the past without the form necessarily having
the w- prefix. For example: M 356 = RES 3695: kl ’klh ys1¨rbn byt
Wd k-s1m ‘all comestibles marked in the bayt Wadd, whether belong-
ing to them’ (Beeston 1978: 142).
In HAD for example: Rb I/89 n. 279, 12–14: ∂-’l ys3b h-∂t Îmym
∂t ynÒf ‘he who did not offer to ∂t Îmym that which he had to pro-
vide according to the rite’ (Frantsouzoff 2001: 128–32). There is also
a possible example of prefix-preterite in Musée d’al-Mukalla n. 161
12: ys1m¨-s1m Yhw∂¨. I find Frantzousoff ’s translation (Frantsouzoff
2005: 248–9) ‘que Yhw∂¨ les entende’ hard to accept both in cul-
tural24 and in grammatical terms; the optative form in ASA is always
l-yf ¨l. We could suggest: ‘Yhw∂¨ was witness of their transaction
(mentioned in previous lines)’.
There is also a –m suffix attested in HAD for the prefixed form. The
examples of the –m suffix forms are too few to establish a grammatical
rule, but in any case I find the following examples interesting: Rb I/84
n. 196a-d+201 a-b: b¨br ∂t ynÒfm ’b-s1 Qs3mm ∂-ynwÌm b-rys3n w-ybÒÌ
s1wd’-s3 ‘since his father Qs3mm was performing the rite, he who was
mourning, according to the rule, and then he achieved Her issue’
(Frantsouzoff 1995: 19). Rb I/89 n. 279 10: ywm ygs1mwnm RÌbn
‘when they would build RÌbn’ (Frantsouzoff 2001: 128–32).
We could surmise two forms of past: one for the preterite and the
other for the imperfective in –m.
What is interesting from an historical standpoint is that all the
ASA languages maintained broad usage of the prefix form to indicate
the past; this past-tense significance for the prefix form was reduced
22
Differently from my earlier stance (Avanzini 1995: 36) I have a marked pref
erence for this translation. Interesting comparisons for the ‘technical’ meaning of
the verb Ìrm can be drawn with biblical Hebrew and Moabite (London 2004:
149 50).
23
See some interesting stylistic syntactic remarks in Gruntfest 1999.
214
to little more than a linguistic relic in Biblical Hebrew (the ‘con-

verted’ waw); all that remains in Arabic is in lam yaf äl, and it is
completely absent in modern languages with no trace of the past pre-
fixed form in MSA or in Ethiopic. Not only did ASA maintain its
prefix form for the past, but it has two distinct prefix forms with
aspectual and temporal differences.
This reconstruction of the ASA verb suggests an obvious compari-
son and is proof of Tropper’s reconstruction (Tropper 1994, 1995,
2000) of the western Semitic verb system of the second millennium.
suffix conjugation prefix conjugation

qatala yaqtul, yaqtulu,
qatalu (plural) yaqtulu (plural) yaqtuluna (plural)
preterite preterite present
The verb has three forms one with suffixes and two with prefixes.
The verb system which initially had a marked aspectual value soon
ran into difficulties. The short prefix form (preterite-aorist) became
integrated with the suffix form; the long form (present-narrative)
went through many phases of morphological specialization in which
a stronger temporal component than that of the early aspectual con-
notation was inserted thus becoming more specific.
This verb scheme is found upstream of all attested ASA languages;
indeed we could say that it is the proto-ASA verb system, a proto-
western not only a proto-north-western verb system25.
To drastically curtail the importance of the presumed loss of the
imperfect yVqattVl in ASA modifies the general frame of the history
of southern Semitic.
Having proved that south Arabian languages are testimony to a
history in the longue durée within south Arabia and do not come
from a recent colonization (late second-early first millennium) on
the part of populations from the north opens a broad historic and
linguistic debate. For instance, a repeated datum, which in my view
has been merely reiterated without being subjected to further discus-
24
The expression is frequent in late texts (see the examples mentioned by
Fransouzoff 2005: 251 2), but always has the divinity as the subject ‘that the god
listens to him, does his wish’. Frantsouzoff ’s explanation: ‘Il paraît que les auteurs
de notre inscription aient employé cette formule avec ys1m¨ sans en comprendre
profondement le sens réel’ does not seem very probable.
25
‘Typologically the reconstruction of the prefixed verb system of Proto ancient
South Arabian is very close to that of Northwest Semitic of the II Millennium’
(Mazzini 2007).
215
sion or critique, is that MSA does not derive from ASA.26 Separating
ancient south Arabian from modern south Arabian, again on the ba-
sis of the iprus-iparras/yenger-yenager isogloss and thus on the recent
arrival of populations who spoke ASA risks leading us astray histori-
cally and linguistically.
After lending support to the idea of non-affiliation between MSA
and ASA Simeone-Senelle had a meaningful lapsus and repeated that
the MSA languages were spoken in south Arabia ‘bien avant la
pénétration de l’Arabe, venu du nord’ (Simeone-Senelle 2002: 393).
Nobody had ever thought that MSA derived from classical Arabic
and it is therefore likely that Simeone-Senelle was getting ASA and
Arabic mixed up.
As we well know from studying the Romance languages which ob-
viously come from Latin, the former were born from a colloquial
language and not from Cicero’s Latin. If we compared a Latin in-
scription from the Republican period with a recording of spoken
Rumanian the difference would be no less great than if we were to
compare Sabaic and Mehri.
Epigraphic South Arabic is the most antique documentation of
southern Semitic. Why, then should we do without it, in principle at
least?
If my theory of an endogenous origin of ASA languages is correct,
it would have far-reaching consequences for the hitherto accepted
classification of Semitic languages: Hetzron’s map of the distribution
of Semitic languages would have to be redrawn. The oldest phase of
southern Semitic would be less southern because of the many
isoglosses between ASA and second millennium north-western
Semitic, and central Semitic would be less central and ‘innovative’,
comprising practically the whole western part of the Semitic world
from Syria to Yemen.
26
This hypothesis is asserted by some French scholars and I find the position
held by V. Porkomovkij (Porkomovkij 1997) more convincing. See also Huehner
gard 2005: 160 1: ‘Nebes demonstrated that none of the languages for which there
is sufficient evidence Sabaean, Minaean, Qatabanian exhibits the form
yVqattVl; the imperfective form of the verb is, instead, yaqtulu. […] It also means
that none of these languages can be the ancestor of either the Modern South
Arabian languages or the Ethiopian Semitic languages’.
Huehnergard’s article is very interesting for other reasons. He hypothesizes a
Central Semitic divided into two main groups, North Central Semitic and South
Central Semitic.
216
REFERENCES
Appleyard D.L. 1996. ‘Ethiopian Semitic and South Arabian towards a Re exami
nation of a Relationship’, in Sh. Izra’el and Sh. Raz (eds), Israel Oriental Studies
16. Studies in Modern Semitic Languages (Leiden). 203 28
2001. ‘New Finds in the 20th century: the South Semitic Languages’, in Sh.
Izra’el (ed.), Israel Oriental Studies 20. Semitic Linguistics: the State of the Art at
the Turn of the 21st Century (Tel Aviv). 401 30
Avanzini A. 1991. ‘Linguistic Data and Historical Reconstruction: between Semitic
and Epigraphic South Arabian’, in A.S. Kaye (ed.), Semitic Studies in Honor of
Wolf Leslau (Wiesbaden). 107 18
1995. Inventario delle iscrizioni sudarabiche 4. As Sawda’. (Paris Rome)
2004. Corpus of South Arabian Inscriptions I III. Qatabanic, Marginal
Qatabanic, Awsanite inscriptions. (Pisa)
2005a. ‘Some Thoughts on Ibex on Plinths in Early South Arabian Art’, Ara
bian Archaeology and Epigraphy 16, 144 53
2005b. ‘Some Brief Observations on the Qatabanic Verb System and on the
Relationship between the Ancient South Arabian and the Modern South Ara
bian’, in A.V. Sedov and I.M. Smiljanskaja (eds), Arabia Vitalis. Studies in Hon
our of V.V. Naumkin (Moscow). 318 23
2005c. ‘Some Remarks on the Classification of Ancient South Arabian Lan
guages’, in P. Fronzaroli (ed.), 10th Meeting of Hamito Semitic Linguistics,
Firenze 18 21 Aprile 2001 (Florence). 117 25
2006a. ‘Ancient South Arabian Anthroponomastics: Historical Remarks’. Pro
ceedings of the Seminar for Arabian Studies 36, 79 85
2006b. ‘To accompany a Recently Published Sabaic Text: Historical and
Grammatical Remarks’, in P.G. Borbone, A. Mengozzi and M. Tosco (eds),
Loquentes Linguis. Festschrift F. Pennacchietti (Wiesbaden). 35 44
2006c. ‘A Fresh Look at Sabaic'. Review of P. Stein. Untersuchungen zur
Phonologie und Morphologie des Sabäischen. (Rahden, Westf ) 2003. Journal of
American Oriental Society. 126: 2, 253 60
Beeston A.F.L. 1962. A Descriptive Grammar of Epigraphic South Arabian. (London)
1978. ‘A Minean Market Code’, Bulletin of the School of Oriental and African
Studies 41, 142 5
1982. ‘Languages in Pre Islamic Arabia’, Arabica 28, 178 86
1984. Sabaic Grammar. Journal of Semitic Studies Monograph 6. (Manches
ter)
Christian V. 1919 20. ‘Akkader und Südaraber als ältere Semitenschichte’,
Anthropos 15 16, 729 39
Cohen D. 1984. La phrase nominal et l’évolution du système verbal sémitique. (Paris)
Edens C. 2002. ‘Looking for Connections: Southwest Arabia in Late Prehistory’,
Man and Environment 27, 45 55
Frantsouzoff S. 1995. ‘The Inscriptions of the Temples of Δhat Îimyam at
Raybun’, Proceedings of the Seminar for Arabian Studies 25, 15 28
2001. ‘Le “tailleur de pierre” (grby n/hn) dans les inscriptions sudarabiques’,
Raydan 7, 125 43
2005. ‘Une nouvelle inscription Îa∂ramoutique datée par éponyme’, in
A.M. Sholan, S. Antonini and M. Arbach (eds), Sabaean Studies: Archaeological,
Epigraphical and Historical Studies in Honour of Y.M. Äbdallah, A. De Maigret
and Ch.J. Robin (Naples, ∑anä’). 247 58
217
Garbini G. 2004. ‘The Origins of South Arabians’, in A.V. Sedov (ed.), Scripta
Yemenica. Studies in Honor of M.B. Piotrovkij (Moscow). 203 9
2006. Introduzione all’epigrafia semitica. (Torino)
Gragg G. 1997. ‘Old South Arabian Phonology’, in A.S. Kaye (ed.), Phonology of
Asia and Africa 1 (Winona Lake, IN). 163 5
Gruntfest Y. 1999. ‘The Consecutive Imperfect in Semitic Epigraphy’, in Y.
Avishur and R. Deutsch (eds), Historical, Epigraphical and Biblical Studies in
Honor of Prof Michael Heltzer (Tel Aviv). 171 80
Gzella H. 2006. ‘Die Entstehung des Artikels im Semitischen: eine “Phönizische”
Perspektive’, Journal of Semitic Studies 51, 1 18
Healey J.F. 1990. The Early Alphabet. (London)
Huehnergard J. 2005. ‘Features of Central Semitic’, in A. Gianto (ed.), Biblical and
Oriental Essays in Memory of William L. Moran (Rome). 155 203
Joffe A. H. 2002. ‘The Rise of Secondary States in the Iron Age Levant’, Journal of
the Economic and Social History of the Orient 45, 425 67.
Keall E.J. 2004. ‘Possible Connections in Antiquity between the Red Sea Coast of
Yemen and the Horn of Africa’, in P. Lunde and A. Porter (eds), Trade and
Travel in the Red Sea Region. (Oxford). 43 55
Knudsen E.E. 1998. ‘Central Semitic *yaqtulum Reconsidered. A Rejoinder to J.
Tropper’, Journal of Semitic Studies 48, 1 9
Liverani M. 2005a. ‘Imperialism’, in S. Pollock and R. Bernbeck (eds), Archaeologies
of the Middle East. Critical Perspectives (Oxford). 223 43
2005b. Review of M. Van de Mieroop, A History of Ancient Near East (Ox
ford 2003), Orientalia 74, 171 3
Macdonald M.C.A. 2000. ‘Reflections on the Linguistic Map of Pre Islamic Ara
bia’, Arabian Archaeology and Epigraphy 11, 28 79
de Maigret A. 1981. ‘Two Prehistoric Cultures and a New Sabaean Site in the East
ern Highlands of North Yemen’, Raydan 4, 194 204
1990. The Bronze Age Culture of Hawlan al Tiyal and al Hada. (Rome)
1999. ‘The Arab Nomadic People and the Cultural Interface between the
“Fertile Crescent” and “Arabia Felix”’, Arabian Archaeology and Epigraphy 10,
220 4
Marrassini P. 1991. ‘Some Observations on South Semitic’, in A.S. Kaye (ed.),
Semitic Studies in Honor of Wolf Leslau (Wiesbaden). 1016 23
2003. ‘“Genetic Subgrouping” in Semitic’, in P. Marrassini (ed.) Semitic and
Assyriological Studies Presented to Pelio Fronzaroli (Wiesbaden). 369 80
Master D.M. 2001. ‘State Formation Theory and the Kingdom of Israel’, Journal of
Near Eastern Studies 60, 117 31
Mazzini G., 2005. ‘Ancient South Arabian Documentation and the Reconstruction
of Semitic’, in P. Fronzaroli (ed.), 10th Meeting of Hamito Semitic Linguistics,
Firenze 18 21 Aprile 2001 (Florence). 215 38
2006. ‘∂ base Pronominal System in Qatabanic’, in P.G. Borbone, A.
Mengozzi and M. Tosco (eds), Loquentes Linguis. Festschrift F. Pennacchietti
(Wiesbaden). 475 88
2007. ‘The System of Prefixed Verb in Proto ancient South Arabian’, in M.
Moriggi and M. Cassarino (eds), Proceedings of the 12th Meeting of Hamito Se
mitic Linguistic (Rubettino). 255 61
Nebes N. 1988. ‘The Infinitive in Sabaean and Qatabanian Inscriptions’, Proceed
ings of the Seminar for Arabian Studies 18, 63 78
218
1994. ‘Verwendung und Funktion der Präfixkonjugation im Sabäischen’, in

N. Nebes (ed.), Arabia Felix, Beiträge zur Sprache und Kultur des vorislamischen
Arabien. Festschrift Walter W. Müller zum 60. Geburtstag (Wiesbaden). 191 211
1997. ‘Stand und Aufgaben einer Grammatik des Altsüdarabischen’, in R.G.
Stiegner (ed.), Aktualisierte Beiträge zum 1. Internationalen Symposion Südarabien
interdisziplinär an der Universität Graz mit kurzer Einführung zur Sprach und
Kulturgeschichte (Graz). 111 31
2001. ‘Zur Genese der altsüdarabischen Kultur. Eine Arbeitshypothese’, in R.
Eichmann and H. Parzinger (eds), Migration und Kulturaltransfer. Der Wandel
vorder und zentralasiatischer Kulturen im Umbruch vom 2. zum 1. vorchristlichen
Jahrtausend. Akten des Internationalen Kolloquiums. Berlin 23. bis 26. November
1999 (Bonn). 427 35
Pennacchietti F. 1974. ‘Appunti per una storia comparata dei sistemi preposizionali
semitici’, Annali dell’Istituto Orientale di Napoli 32, 161 208
Petracek K. 1981. ‘Le système de l’arabe dans une perspective diachronique’,
Arabica 28, 162 77
Pirenne J. 1990. Fouille de Shabwa I. (Paris)
Porkomowskij V. 1997. ‘Modern South Arabian Languages from a Semitic and a
Hamito Semitic Perspective’, Proceedings of the Seminar for Arabian Studies 27,
219 23
Robin J.Ch. 1983. ‘Complements à la morphologie du verbe en sudarabique
épigraphique’, Matériaux arabes et sudarabiques, Recherches en cours 1, 163 85
Routledge B. 2004. Moab in the Iron Age. Hegemony, Polity, Archaeology. (Philadel
phia)
Rundgren F. 1959. Intensiv und Aspektkorrelation. (Uppsala and Wiesbaden)
Scigliuzzo E. 2003. ‘A South Arabian Ivory Vessel from Hama Reconsidered’,
Ugarit Forschungen 35, 629 47
Sima A. 2001. ‘Altsüdarabische Konditionalsätze’, Orientalia 70, 283 312
Simeone Sennelle, M.C. 2001. ‘Les langues sudarabiques modernes à l’aube de l’an
2000’, in Sh. Izra’el (ed.), Israel Oriental Studies 20. Semitic Linguistics: the State
of the Art at the Turn of the 21st Century (Tel Aviv). 379 400
Stein P. 2002. ‘Zur Morphologie des sabäischen Infinitivs’, Orientalia 71, 393 414
2003. Untersuchungen zur Phonologie und Morphologie des Sabäischen.
(Rahden/Westf)
Tropper J. 1994. ‘Present *yaqtulum in Central Semitic’, Journal of Semitic Studies
39, 1 6
1995. ‘Das altkananäische und ugaritische Verbalsystem’, in M. Dietrich and
O. Loretz (eds), Ugarit. Ein ostmediterranes Kulturzentrum im Alten Orient
(Münster). 159 70
2000. Ugaritische Grammatik. (Münster)
2001. ‘Die Herausbildung des bestimmten Artikels im Semitischen’, Journal
of Semitic Studies 46, 1 31
Vogt B. and A.V. Sedov. 1998. ‘The Sabir Culture and Coastal Yemen during the
Second Millennium BC The Present State of Discussion’, Proceedings of the
Seminar for Arabian Studies 28, 261 70
Voigt R.M. 1987. ‘The Classification of Central Semitic’, Journal of Semitic Studies
32, 1 21
1990. ‘The Germination of the Present imperfect Forms in Old Ethiopic’,
Journal of Semitic Studies 35, 1 18
219
Wilkinson T.J. 2005. ‘The Other Side of Sheba: Early Towns in the Highlands of
Yemen’ Bibliotheca Orientalis 62, 5 14
Wilkinson T.J., C. Edens and M. Gibson. 1997. ‘The Archaeology of the Yemen
High Plains: a Preliminary Chronology’, Arabian Archaeology and Epigraphy 8,
99 142
220

THE FUNCTIONS OF THE PREPOSITION

K- IN MEHRI
AARON D. RUBIN
PENNSYLVANIA STATE UNIVERSITY
Abstract
The Mehri preposition k ‘with’ is used in a variety of idiomatic func

tions, some of which are not fully treated in, or are absent from, the
exisiting grammatical sketches of Mehri. These include expressions of
possession, of physical or environmental conditions, and of time. Not
only is a description of these important for our understanding of
Mehri, but some of these idioms, due to their peculiar nature, are also
valuable for cross linguistic study.
In addition to its basic meaning ‘with’, the clitic preposition k- in

Mehri has a number of idiomatic functions, some of which have
been hitherto undescribed.1 Description of these functions is of value
not only to those interested in Mehri and comparative Semitics, but
also to those who work on general linguistic typology.
Before addressing the various functions of k-, we will look briefly
at the form. The preposition appears as k- before a nominal object,
but when pronominal suffixes are to be attached, the preposition has
the allomorph s-. The following table gives the complete set of suf-
fixed forms:2
sing. dual plural

1c say s¢´ki sin
2m suk s¢´ki sik¢m
2f says sik¢n
3m s¢h s¢´hi sih¢m
3f sis sis¢n
1
I am very grateful to Fabrizio Pennacchietti and Gary Rendsburg, who both
kindly provided me with very valuable comments to an earlier draft of this article.
2
All forms and examples come from the texts of T.M. Johnstone, as published
in Stroomer (1999), unless otherwise noted.
221
THE FUNCTIONS OF THE PREPOSITION K- IN MEHRI
Note that the 2ms form has the rare variant sik, formed presum-
ably on analogy with tik, the 2ms form of the pronominal direct ob-
ject marker t-.
As already noted, the basic meaning of this preposition, with or
without suffixes, is ‘with’, as illustrated in the following sample sen-
tences:
ho m¢g¢tayr k¢-Ìam¢ý ‘I will speak with my mother.’ (9:2)3
w¢kub k-'agawz ‘He went in with the old woman.’ (75:7)
t¢syer say ‘You will go with me.’ (55:2)
agayg s¢h sx¢w¢lul ‘The man (who was) with him stayed.’ (68:5)
s¢biw agayg sih¢m ‘They took the man prisoner with them.’ (69:2)
Etymologically, Mehri k- is connected with the widely attested
Semitic preposition k- ‘like, as’. On the correspondence of the Mehri
meaning to that found in the rest of Semitic, already Bittner (1914:
8) saw the parallel of Ethiopic m¢sla ‘with’ and Arabic mi†la ‘like’.4
A similar (but by no means identical) development of the preposi-
tion k- is also found in South Ethiopic. In Amharic, for example, the
preposition kä- has a wide variety of meanings, including ‘with’.5 At
this point, the idiomatic uses of this preposition shall be examined in
turn.
1. The ‘Have’ Construction
The most important idiomatic use of k- is in the construction which

corresponds to the English verb ‘to have’.6 Like most other Semitic
languages, Mehri has no verb ‘to have’ and must resort to a peri-
phrastic construction. An explicit past tense is completely absent
from this construction, and must be gleaned from context. Examples
are:
say n¢xlit ‘I have a palm tree.’ (77:5)
s¢h gigen u g¢g¢not ‘He had a boy and a girl.’ (22:1)
3
Numbers in parentheses following all examples refer to text and line numbers
in Stroomer (1999).
4
On Ethiopic m¢sla, see also Pennacchietti (1974: 189 n. 69). Pennacchietti
also briefly notes the importance of this parallel Ethiopic and South Arabian devel
opment (p. 198).
5
See further in Pennacchietti (1974: 195). On the uses of kä in Amharic, see
Leslau (1995: 605 7). Note here that in the meaning of ‘with’, the Amharic
kä usually combines with another pre or postposition.
6
This use was discussed very briefly by both Bittner (1913: 54 Anm. 1) and
Thomas (1937: 251).
222
sis woz ‘She had a goat’. (49:1)

gayg s¢h rikeb ‘A man had a riding-camel.’ (lit. ‘A man, he had a riding-
camel’) (12:1)
A future tense can be made explicit by an imperfect form of the
verb wiÈa ‘be, become’:
taÈa suk maws ‘You will have a razor.’ (75:11)7
The ‘have’ construction is negated as any other nominal sentence,
that is, with the particles ’¢l … la:
’¢l sih¢m Èawt la ‘They had no food.’ (11:2)
’¢l sin È¢raws la ‘We had no money.’ (91:2)
The grammaticalization of a preposition ‘with’ into a possessive is
found in other Semitic languages, including in neighbouring Arabic
dialects, as I have discussed in detail elsewhere.8 It is also well at-
tested cross-linguistically.9 Therefore, the Mehri construction is, gen-
erally, unremarkable. However, there is an additional feature of this
Mehri construction which merits closer attention. Within the ‘have’
construction, a noun indicating a close family member (father,
mother, son, daughter, brother, sister, wife, brother-in-law) will ap-
pear with a redundant possessive suffix:
say Ì¢bráyti ‘I have a daughter.’ (48:7) (lit. ‘I have my daughter.’)
s¢h agah ‘He had a brother.’ (85:1) (lit. ‘He had his brother.’)
say Ìam¢ý 'agawz ‘I have an old mother.’ (65:7)
agah s¢h Ì¢br¢´t¢h ‘His brother had a daughter.’ (85:1)
s¢h t¢´†¢h r¢Ìáym¢t ‘He has a (beautiful) wife.’ (19:1)
s¢h Ì¢bánhe y¢bayt ‘He had 7 sons.’ (50:1)
s¢h Ìaym¢h ‘He had a brother-in-law.’ (64:10)
sik¢m †ad gak¢m ‘Do you have a brother?’ (Müller 1902: 140, l. 38)
The special treatment of familial terms in connection with posses-
sion is well known from elsewhere, for example, from Italian, where
one typically says la mia faccia ‘my face’, la mia vicina ‘my neigh-
bour’, but mia mamma ‘my mother’.10 The difference in the case of
Mehri is that we are dealing not with a possessive phrase of the type
7
The noun maws ‘razor’ is feminine, and so the verb taÈa is a 3fs imperfect, in
agreement with its logical subject maws. Johnstone (1987) does not indicate that
maws is feminine, but the Arabic word from which it is borrowed (Classical musa,
dialectal mus) is indeed feminine. On the history of this word, with discussion of its
gender in Arabic, see Pennacchietti (2005).
8
Rubin (2005: 58 9).
9
Heine and Kuteva (2002: 88).
10
Maiden and Robustelli (2000: 162 3).
223
‘my mother’, ‘your brother’, but rather with a pseudo-verbal ‘have’

construction.
2. Expressions of Physical or Environmental Conditions
The preposition k-, in suffixal forms, is also used in a variety of ex-

pressions relating to physical or environmental conditions. From the
former category, we find:
say (sis, sin) Ì¢bur ‘I’m (she’s, we’re) cold.’ (lit. ‘I have coldness.’)
say (sis, sin) ÌarÈ ‘I’m (she’s, we’re) hot.’
This type of idiom is well known from the Romance family (e.g.,
Spanish tengo frío ‘I’m cold’) and from many other languages, though
it is not typical of Semitic.
Many more examples fall into the category of environmental con-
ditions. This type of expression is rather more unusual. For example,
the expression ‘it is raining’ will vary depending on context; specifi-
cally, its expression depends on who is experiencing the rain. If relat-
ing one’s own experience, one would say say m¢wse ‘It’s raining (for
me)’ (53:2; 62:1). But if relating a story in the third person plural,
one would say sih¢m m¢wse ‘It’s raining (for them)’. Following are
several other examples to illustrate this use of k-:
say ag¢llet ‘It was misty.’ (47:4) (lit. ‘Mist was with me.’ or ‘I had
mist.’)
sin Ìarit ‘There was a moon.’ (81:1)
say ¢ktiw ‘It was dark.’ (91:16)
s¢h awákt ‘He was hot (at night).’ (91:19)
sis¢n ariÌ ‘There is wind.’ (27:22)
say aÈaÒ¢m ‘It’s cool, cold.’ (94:35)
s¢h ag¢llet w-am¢wse ‘It was misty and rainy.’ (98:1)
sin am¢wse ‘It’s been rainy.’ (98:3)
It should be mentioned that there are sometimes other ways to
express environmental conditions. For example, there is a G-stem
verb from the root lsw ‘to rain’.11 This verb is used impersonally in
the feminine singular, as in ¢wsut ‘it rained’ (10:16). However, in
Johnstone’s texts, the construction with k- is quite common, while
the verb is rare.
These expressions of condition, like the ‘have’ construction dis-
cussed above, lack any explicit reference to tense, which must be
11
This is the same root found in the word m¢wse ‘rain’.
224
gathered from the context So, just as s¢h Èawt can mean ‘he has food’
or ‘he had food’, likewise say ÌarÈ can mean ‘I am hot’ or ‘I was
hot’.
3. Expressions of Time
Finally, the preposition k- is also used idiomatically in a few com-

mon expressions relating to time: k¢-sob¢Ì ‘in the morning’, k-a∂¯ahr
‘at noon; in the afternoon’, k-amg¢rab ‘in the evening' (rare), and,
optionally (and rarely) in k-'aÒ¢r ‘in the evening’. This same k- seems
also to be incorporated into the form kalayn(i) ‘in the evening’.12
With k¢-sob¢Ì, in fact, the k- is really likewise inseparable, and often
k¢-sob¢Ì is best translated simply as ‘morning’ (as in the second ex-
ample below). Examples are:13
téˆ k¢-sob¢Ì agayg s¢wkuf ‘Then, in the morning, the man fell asleep.’
(22:65)
hatom téˆ k¢-sob¢Ì ‘He stayed until morning.’ (9:11)14
n¢hor †ayt kalayni È¢†ot abok¢r ‘One day, in the evening, the camel got
tired.’ (37:17)
kalayni wátx¢f¢k ‘I arrived in the evening.’ (38:8)
téˆ k-'aÒ¢r †a† b¢-Ì¢lláy’ ¢nkot †¢b¢ráyn ‘Then one evening, at night, a hy-
ena came.’ (81:1)
ta [= téˆ] k-a∂¯ahr 'aym¢l ¢fse ‘Then at noon, he made lunch.’ (12:4)
Bittner (1914: 8) refers to the similarity of this temporal use of k-
with the Hebrew preposition k-, but gives no Hebrew examples. Pre-
sumably he was thinking of passages such as k¢-Ìom hay-yôm ‘at the
heat of the day’ (Gen. 18:1), ka-harimi qoli ‘when I raised my voice’
(Gen. 39:18), and the like.15
4. Conclusions
This short article has outlined the various functions of the common
Mehri preposition k-. The particular possessive construction used for
12
This suggestion is found in Bittner (1914: 9) and in Johnstone (1987: 208),
under the heading kl¨n(y).
13
For examples of k occurring with several other temporal nouns, see Bittner
(1914: 8 9).
14
This sentence, and the one preceding, illustrate two different meanings of téˆ,
‘then’ and ‘until’. In fact, téˆ is also used as a temporal relative ‘when’. The phrase k¢
sob¢Ì is nearly always preceded by téˆ, in one of these three meanings.
15
Joüon (1996: §133g, §166m).
225
family members, and the idiomatic expressions dealing with physical

or environmental conditions, have on the whole gone undescribed
until now. They are of value for the scholar who seeks to better un-
derstand the Mehri tongue, and, in addition, they display some in-
teresting typological phenomena which are of equal value to the gen-
eral linguist. Other Mehri phenomena, of equal or greater value to
the Semitist or general linguist, surely await description.
REFERENCES
Bittner, Maximilian. 1913. ‘Studien zur Laut und Formenlehre der Mehri Sprache
in Südarabien. III. Zum Pronomen und zum Numerale’, Sitzungsberichte der
Kaiserlichen Akademie der Wissenschaften in Wien, Philosophisch Historische Klasse
172/5
1914. ‘Studien zur Laut und Formenlehre der Mehri Sprache in Südarabien.
IV. Zu den Partikeln’, Sitzungsberichte der Kaiserlichen Akademie der Wissen
schaften in Wien, Philosophisch Historische Klasse 174/4
Heine, Bernd, and Tanya Kuteva. 2002. World Lexicon of Grammaticalization.
(Cambridge)
Johnstone, T.M. 1987. Mehri Lexicon and English Mehri Word List. (London)
Joüon, Paul. 1996. A Grammar of Biblical Hebrew. 2 vols. Revised and translated by
T. Muraoka. (Rome)
Leslau, Wolf. 1995. Reference Grammar of Amharic. (Wiesbaden)
Maiden, Michael, and Cecilia Robustelli. 2000. A Reference Grammar of Modern
Italian. (Chicago)
Müller, David Heinrich. 1902. Die Mehri und Soqo†ri Sprache. I. Texte. Süd
arabische Expedition IV. (Vienna)
Rubin, Aaron D. 2005. Studies in Semitic Grammaticalization. (Winona Lake, IN)
Pennacchietti, Fabrizio. 1974. ‘Appunti per una storia comparata dei sistemi
preposizionali semitici’, Annali dell’Istituto Orientale di Napoli 34, 161 208
2005. ‘Sull’etimologia di arabo musa “rasoio”’, in A. Mengozzi (ed.), Studi
Afroasiatici: XI Incontro Italiano di Linguistica Camitosemitica (Milan), 231 37
Stroomer, Harry (ed.) 1999. Mehri Texts from Oman. Based on the Field Materials of
T.M. Johnstone. (Wiesbaden)
Thomas, Bertram. 1937. ‘Four Strange Tongues from Central South Arabia’, Pro
ceedings of the British Academy 23, 231 331
226

FROM CONVENTIONAL TO PERSONAL, OR:

WHAT HAPPENED TO METAPHOR UNDER
THE INFLUENCE OF IDEOLOGY –
THE CASE OF GHA}IB TU{MA FARMAN*
HILLA PELED-SHAPIRA
BAR-ILAN UNIVERSITY
Abstract
This article discusses metaphors derived from the animal world, and
used to describe relations between intellectuals and the authorities in
mid twentieth century Iraq, against the backdrop of the conservative
and conformist use of metaphor in Classical Arabic literature. The
topic in question is examined through the narrative works of the
twentieth century exiled Iraqi Communist writer Gha’ib Tu¨ma
Farman (1927 90), who used animal metaphors as an artistic device
through which he depicted himself as a Leftist intellectual persecuted
by the government. The examples from Farman’s own works are con
sidered in light of the use which other twentieth century Arab writers
make of animal metaphors, and the artistic needs which the latter
serve.
Metaphors from various domains are introduced into works of litera-

ture for artistic and rhetorical enrichment and variety, especially for
the purpose of enabling a writer to express his or her message in a
personalized and unique manner. Metaphors in a work of literature
may be uniform or heterogeneous, deriving from various domains
such as religion, colours, water, and so on. The present article will
discuss metaphors derived from nature or, more precisely, metaphors
from the animal world, used to describe relations between intellectu-
als and the authorities in mid-twentieth-century Iraq, against the
backdrop of the conservative and conformist use of metaphor in
Classical Arabic literature. The topic in question will be examined
through the works of a twentieth-century émigré Iraqi Communist
writer, Gha’ib Tu¨ma Farman (1927–90), who used animal meta-
phors as an artistic device through which he depicted himself as a
* I wish to thank Prof. Shmuel Moreh for having read this paper and made
some illuminating comments.
227
THE CASE OF GHA}IB TU{MA FARMAN
Leftist intellectual persecuted by the government. The examples from

Farman’s own works will be examined in the light of the use which
other twentieth-century Arab writers make of animal metaphors, and
the artistic needs which the latter serve.
Animal imagery can be found, for example, in the Boston-born
lyric writer and poet Edgar Allen Poe’s (1809–49) well-known poem
‘The Raven’. In this poem, written in 1845, the bird’s black colour
represents the hopelessness of waiting and the poet’s bitter realization
that his beloved Lenore will never return.1
Poe, of course, is only one of many writers who have borrowed
images from the animal kingdom in order to create an atmosphere,
characterize a hero, or build a world of unique meanings, whether
fictional or documentary. Some of Poe’s successors have done so
as well: in England Katherine Mansfield (1888–1923) managed
to create such an atmosphere in her short story ‘The Fly’
(1922), where this insect, drowning in a pool of ink, is used with
much symbolism as an allegory for a grieving father’s mind;2 later
in Italy Elsa Morante (1918–85) in her History: A Novel used a de-
scription of a laboratory animal, whose ‘handsome little olive-
colored eyes gleamed, immobile, like the eyes of the mad’,3 as a way
to characterize the novel’s hero Useppe, a sick child not expected to
live long.
Regarding intellectuals in the Arab world, we must point out that
as a group they developed on the back of the class of ülama’ or cler-
ics, in the wake of that world’s encounter with the West towards the
end of the nineteenth century and following the winds of secularism
it brought with it. The new Arab intellectuals, among them atheists,
were exposed to Western ideas and tried to instill them into their
own society, but often suffered from feelings of alienation and an in-
ability to communicate even with those closest to them, because of
the intellectual and cultural gap between the two sides. The collision
between the two worlds, occasioned by this gap, was not long in
1
Edgar Allan Poe, ‘The Raven’, The Complete Tales and Poems of Edgar Allan
Poe (New York 1938), 945:
Fancy unto fancy, thinking what this ominous bird of yore
What this grim, ungainly, ghastly, gaunt, and ominous bird of yore
Ment in croaking ‘Nevermore.’
[…] Quaff, oh quaff this kind nepenthe and forget this lost Lenore!
Quoth the Raven, ‘Nevermore.’
2
Katherine Mansfield, ‘The Fly’ in Bowen, Elizabeth (selected), 34 Short Stories
Katherine Mansfield (London 1957 [reprinted 1970]), 350 5.
3
Elsa Morante, History: A Novel (translated from the Italian by William
Weaver), (New York 1984), 425 6.
228
coming.4 On the one hand intellectuals were pushed to the sidelines

by their conservative society, which did not understand the novel
Western terminology they used, and on the other hand the authori-
ties were less than pleased to hear about Western political ideas, espe-
cially when these challenged the legitimacy of their own despotic
rule, and reacted with repression and persecution. Secular Arab intel-
lectuals were thus left out in the cold. Paradoxically, when in a
number of Arab countries self-declared Socialists came to power, as
happened for example in Egypt, Iraq and Syria, intellectuals found it
even more difficult to criticize the authorities, since they now found
themselves attacking those supposedly in favour of an ideology
which they supported,5 since the rulers’ actions were at odds with
their own ideology’s declared principles. Arab intellectuals were thus
driven to the margins, both socially, despite their desire to introduce
social change, and politically, despite their ambition to inculcate a
different political culture, in which the welfare of the individual was
paramount. Secular Arab intellectuals thus live in a conservative
Muslim society, whose rulers prefer to keep it as it is in order to pre-
vent revolts; in such a society there is no room for the Western vision
of welfare and individual freedoms, which Arab intellectuals, in par-
ticular those of a Socialist bent, espouse. The alienation which secular
Arab intellectuals feel as a result of this inter-cultural encounter is
expressed in their writings. It is to be expected, naturally, that the
sharper the collision between the intellectual’s ideology and that of
the rulers or of society, the harsher will be its literary expressions and
the more pointed and profound the metaphors.
Before we examine how metaphors help modern Arab intellectuals
describe their lives in their countries or the ideal state to which they
aspire, the differences between Classical Arabic metaphors and those
used in the modern West must be considered, in order to understand
how and why as time went on some Arab authors gradually gravi-
tated from the former to the latter. That will provide the background
against which we shall show how Gha’ib Tu¨ma Farman used meta-
phors in their modern form.
The imagery used in Arabic literature has changed considerably
from classical to modern times. An important role in this was played
by the transition from a belief in Allah to an adherence to such mod-
ern Western ideologies as Communism. For example, classical au-
4
Menahem Milson, Medieval and Modern Intellectual Traditions in the Arab
World (Tel Aviv 1973), 36ff. (Hebrew).
5
Milson, Medieval and Modern Intellectual Traditions, 41.
229
thors debated the question of ‘what is majaz (“U)?’ In early Qur’an

exegesis majaz provided an explanation for various Quranic expres-
sions. The need for majaz was explained as follows: the Qur’an was
revealed in ‘clear Arabic language’ (5 Ì Ô wÌÒdÓÓ ÌÊUÓK, Q 26:195) and,
since the Arabs’ language is full of unique expressions, it is only to be
expected that such expressions should appear also in the Qur’an.
Over the years the term majaz has come to possess additional mean-
ings. Among the Mu¨tazilites, for example, majaz came to mean a
figurative as against a true and literal expression (in contrast to
Ìaqiqa – WIOI ), used in order to avoid a personification of God.6
A similar case is presented by the semantic transformation of the
term badi¨ (l
b), whose literal meaning is ‘something new, original’
but which over time has come to connote rhetorical forms and meta-
phors. The rhetorical images which flourished in particular during
and after the Abbasid period,7 ran the gamut from a simple transfer
from one semantic field to another, to an opposite interpretation
(taˆwil – q
ËQ), which the Mu¨tazilites used for the purpose of ex-
plaining away personifying expressions referring to God as no more
than figures of speech. The first to have dedicated a complete work
to the subject of badi¨ was Äbd Allah Ibn al-Mu¨tazz (d. 206/908);
in it he described the various types of badi¨ and provided examples
from the Qur’an, the Ìadith and both pre-Islamic and Islamic
poetry.8 After Ibn al-Mu¨tazz the use of badi¨ declined, as part of the
relative stagnation into which Arab culture and literature fell
between the thirteenth and the nineteenth century.
6
W.P. Heinrichs, ‘majaz’, in Julie Scott Meisami and Paul Starkey (eds), Ency
clopedia of Arabic Literature, London, 1998, 495 6; B. Reinert, ‘Madjaz’, E.I2
(Leiden 1986), vol. V, pp. 1025 6. For more on the history of the term majaz see
the chapter ‘“U:« WHK' in MuÌammad Äbd al Mu††alib, al Balagha wa’l uslubiyya
(Cairo 1984), 57 70, esp. pp. 69 70, on the necessity of using majaz in order to
express spiritual needs.
7
On the widespread use of badi¨ in the Abbasid period and on the prominent
writers of that period who made frequent use of metaphor (…—UF «), simile (tOA),
juxtaposition (‚U), paronomasia (”UM) and a variety of other rhetorical devices,
see: AÌmad Ibrahim Musa, al ∑abgh al badiï fi al lugha al ärabiyya (Cairo 1969),
50 116. On the use of these rhetorical devices by Òufi poets after the Abbasid pe
riod see MuÌammad Äbd al Munïm Khafaji, al Adab fi al turath al Òufi (Cairo
1980). ∑ufi poets would occasionally express themselves in indirect terms (‘ÊËdR
`
dB « ÊËœ ÒLK « v« ÊËbLF
Ë ¨…—UF« vK …—Uô«'), in order to keep their thoughts and
philosophy secret (ibid., p. 250). This gave their poetry a symbolic dimension lack
ing in other poets, who preferred more concrete and direct language.
8
W.P. Heinrichs, ‘badi¨’, in Meisami and Starkey (eds), Encyclopedia of Arabic
Literature (London 1998), 122 3. See also Aghna†ius Kratshqufski, Kitab al badi¨
li Äbd Allah bin al Mu¨tazz (Damascus, no date).
230
As a matter of fact in medieval times majaz was used in a concrete

sense, without the symbolism commonly favoured by today’s writers.
Furthermore, in those days a correspondence between the described
object and its metaphor or majaz was required, with respect to form,
colour and size (teeth, for example, were compared to pearls), but
similarity by association or through the object’s abstract symbolic as-
pects was ignored.
Since the beginning of the twentieth century Arab poets, in par-
ticular those who emigrated to America, began to incline towards
romanticism; Arabic poetry thus became more personal, expressing
poets’ inner feelings. No longer did poems consist of superficial
descriptions detached from their symbolic meaning, but came to in-
clude both symbols and allusions.9 No longer did poets choose
silent, direct and rational ways of expression, but preferred more per-
sonal and subjective modes. Greater stress was now laid on meaning
rather than on florid language, partly due to the Arabic language’s
contacts with languages which do not make great use of the rhetori-
cal devices included in the concept of badi¨.10
In the following examples we shall see how animal imagery bears
different meanings in the works of different poets, each in accord-
ance to his own ideology, the message he wishes to convey to the
reader in the course of literary communication, and his style. The
range of uses to which modern Arab writers put these images is very
variegated, extending from correspondences to features of the literary
figures to covert and explicit ideological messages. To take one exam-
ple, the Sudanese author al-™ayyib ∑aliÌ (b. 1929) makes frequent
use of animal imagery in order to stress some external feature of a
character, often its laughter. The feature is more often than not
anomalous. Thus the laughter of the main character in Ürs al-Zayn
(Zayn’s Wedding, Beirut, 1967) is compared to the bray of a donkey
(—UL(« oON);11 the laughter of the deaf Äshamneh, another one of the
novel’s eccentric characters, is described as a rooster’s cry (ÕUO
ÃUb«);12 the lame Musa is described as living on the margins, as do
old stray dogs (WUC« …“uF« »öJ« iF gOF UL).13 ∑aliÌ’s images
9
Shmuel Moreh, ‘An Outline of the Development of Modern Arabic Litera
ture’, in Shmuel Moreh, Studies in Modern Arabic Prose and Poetry (Leiden 1988),
67, 71 3; Moreh, ‘The Influence of Western Poetry, Particularly that of T.S. Eliot
on Modern Arabic Poetry (1947 70)’, in Moreh, Modern Arabic Poetry 1800
1970 (Leiden 1976), 217, 245.
10
Musa, al ∑abgh al badiï fi al lugha al ärabiyya, 465.
11
al ™ayyib ∑aliÌ,Ürs al Zayn (Beirut 1967), 13, 23.
12
∑aliÌ, Ürs al Zayn, 26.
13
∑aliÌ, Ürs al Zayn, 27.
231
are usually derived from farm animals, since most of his stories take
place in a rural setting.
The Lebanese woman writer Layla Ba¨labakki (b. 1934) describes
the heroine’s mental world and the status of women in the Arab
world in general using animal metaphors, as in the stories ‘Kuntu
muhra wa-Òirtu fa’ra’ (‘I was a Filly and became a Mouse’) and
‘al-Qi††a’ (‘The Cat’) (both in Safinat Ìanan ila al-qamar, Beirut
1964).14
When the Syrian woman writer Ghada al-Samman (b. 1942)
wants to describe the anxiety felt by the main character in the short
story ‘al-Fajr ïnda al-nafidha’ (‘Dawn in the Window’), she makes
use of images borrowed from the avian world:15 the heroine fears
that her sensuous neighbour’s visits to her home every evening may
wreck her family. The neighbouring woman appears to the heroine
as a predator intent on preying on fearful forest birds flying around
hither and thither in frightened schools, which of course symbolize
the heroine herself.
The Egyptian woman writer and physician Nawal al-Sa¨dawi
(b. 1931) also uses many metaphors taken from the animal king-
dom. In her short story ‘Îubbi al-waÌid’ (‘My Only Love’)16 a black
dog shows more empathy towards the main character than does her
husband. In this way al-Sa¨dawi communicates to the reader her
criticism of Middle Eastern husbands’ indifference to and contempt
for their spouses, in contrast to the loyalty shown her by the dog. In
her Mudhakkirat †abiba (Memoirs of a Woman Doctor) al-Sa¨dawi
again uses metaphors taken from the animal world to describe the
heroine’s feelings in the alienated and unfriendly place where she
finds herself and her fears for her fate: the hospital at night seems to
her imagination to be like a wild hyena17 and her domineering hus-
band’s ears are described as large and flat as a rabbit’s, perhaps sym-
bolizing his cowardice and his concern for his ‘honour’ and status,
which hide under a cover of confident manhood.18 When a young
girl comes to her clinic in fear of her relatives, who threaten to mur-
der her for having brought ‘dishonour’ on her family, the doctor de-
14
On metaphorical images from the animal world in Ba¨labakki’s writings see
Olga Bramson, Revolutionary Attitudes of the Arab Women According to Layla
Baalbakki’s Narrative Works, M.A. Thesis (Bar Ilan University 1990), 94 9.
15
Ghada al Samman, ‘al Fajr ïnda al nafidha’, in Äynaka qadri (Beirut, 1985),
155.
16
Nawal al Sa¨dawi, ‘Îubbi al waÌid’, in LaÌÂat Òidq (Cairo 1962), 57 8, 65.
17
Al Sa¨dawi, Mudhakkirat †abiba (Beirut, 1988 [1st published in 1965]), 35.
18
Al Sa¨dawi, Mudhakkirat †abiba, 70.
232
cides to save her from the clutches of her relatives, whom she per-
ceives as ‘wild animals, snakes, rats and cockroaches’.19
Among the writings of Iraqi author Ädnan Ra’uf (1926–98)
metaphors taken from the animal kingdom are particularly in evi-
dence in his short story ‘Duda’ (‘Worm’, 1946),20 which is reminis-
cent of Franz Kafka’s The Metamorphosis. The story, a parable on
loneliness and the search for the meaning of life, revolves around the
worm, which symbolizes the pettiness and triviality of human exist-
ence and man’s feelings of insignificance in the face of the grand
forces which surround and move him and which he cannot affect in
any way. Ra’uf thus uses metaphors borrowed from the animal world
in order to raise and discuss existential philosophical questions.
Similar to the previously mentioned writers, the Iraqi Gha’ib
Tu¨ma Farman, who was born in Baghdad in 1927 and died in Mos-
cow in 1990, puts images from the animal world to a variety of uses
in unique ways. Whereas the authors mentioned above exploited ani-
mal metaphors mainly for characterizing their figures’ inner world
and outer looks, Farman expands their use to cover a great number
of topics, feelings and figures, thereby achieving four distinct aims.
First of all, the images he uses enable him to build up a complex,
detailed picture of how Leftist intellectuals in mid-twentieth-century
Iraq were persecuted and depicts their struggles against the authori-
ties. Secondly, animal metaphors make it possible for Farman to give
very special artistic expression to the educated individual’s feeling of
alienation from a society which refuses to accept him although he
grew up in it. The third function of images from the animal king-
dom in Farman’s works is to draw a picture of Baghdad as he remem-
bered it, a stifling city with narrow, winding alleys. Lastly, Farman
uses animal metaphors in order to describe unattainable peace and
happiness, as an allegory for his own life as a ceaseless wanderer who
never found a true place of rest and eventually died on foreign soil.
In the present article we shall examine these functions in Farman’s
works, in the context of the lives of Leftist Iraqi intellectuals in the
1950s, whom this writer represents so well.
Gha’ib Tu¨ma Farman began his writing career as a poet, then went
on to write short stories, which were published in two collections,
ÎaÒid al-raÌa (Grindstone Harvest, Baghdad 1954) and Mawlud akhar
(Another Newborn, Baghdad 1959), as well as eight novels, some of
which describe the Iraqi poor and their tribulations in different peri-
19
Al Sa¨dawi, Mudhakkirat †abiba, 86.
20
First published in the Iraqi journal al Waqt al ∂a’i¨ in 1946.
233
ods, while others deal with young Iraqi intellectuals, most of whom
had Socialist tendencies and were therefore persecuted, as was Farman
himself. He also wrote a historical essay on Iraq entitled al-Îukm al-
aswad fi al-Ïraq (The Rule of Darkness in Iraq, Cairo 1957) and trans-
lated over thirty novels from Russian into Arabic. He wandered for
decades throughout the Arab world, in Europe and in Asia, partly
from choice, in order to study and work, and partly from necessity,
since his own government mistrusted him for his Socialist beliefs and
the Communist slogans he wove into his works, and did not let him
make a decent living in his own homeland.
While Iraq was still a monarchy Communists were already treated
badly. However, the situation worsened after it became a republic in
1958, and especially after the execution of Äbd al-Karim Qasim in
1963. The latter had been sympathetic towards the Communists,
who were subsequently treated harshly by the authorities, out of fear
that they would attempt to overthrow the regime and set up a Com-
munist-style ‘dictatorship of the proletariat’. The steps which the
authorities took against the Communists, and especially the intellec-
tuals, covered the gamut from economic sanctions and revoking their
citizenship in milder cases, to expulsion, imprisonment, torture and
even execution in the case of more intransigent opponents, some-
times involving a public trial ending in a hanging. Countless inno-
cent people fell victim to such actions. It even happened that the gov-
ernment sentenced a person to death, then reduced the sentence to
life imprisonment under public pressure, and later again sentenced
him to death on trumped-up charges, as in the case of the secretary of
the Communist Party in 1947, known as ‘fahd’ (‘panther’), according
to the memoirs of the Iraqi poet MuÌammad Mahdi al-Jawahiri.21
Gha’ib Tu¨ma Farman in his works describes these dark episodes in
the history of Iraq not only through the plot’s events, but also by
means of metaphors taken from the animal world, which the reader
can compare with the situation in Iraq during those years.
Leftist intellectuals’ sense of being persecuted – the individual versus the
authorities
As previously mentioned, the first aim of the imagery in question is
to make the reader aware of Leftist intellectuals’ sense of being perse-
cuted by the authorities to whose rule they were opposed. Their ob-
jections to the regime were expressed in ideologically driven writing
which the authorities perceived as seditious, and occasionally also in
positive acts of resistance, such as violent demonstrations, open criti-
21
MuÌammad Mahdi al Jawahiri, Dhikrayati vol. II (Damascus 1991), 53.
234
cism of the government in literary works and in the press, strikes,

and calls for social and class solidarity.
Farman’s allegorical short novel Alam al-sayyid Ma¨ruf (Mr.
Ma¨ruf ’s Pains, Beirut 1982) is a biting depiction of the problematic
relationship between Leftist intellectuals and the Iraqi authorities.
It tells of an educated man named Ma¨ruf who works in a govern-
ment office and finds it difficult to fit into his backward and illiter-
ate surroundings. The target of the allegory is the arrogant and
paralyzing attitude of those in power in Iraq towards the country’s
intellectuals: Ma¨ruf ’s superiors try their best to repress him, just as
the Iraqi authorities try to repress their Leftist intellectuals. Here we
find many examples of Farman’s use of metaphors taken from the
animal world in order to convey to the reader an intellectual’s sense
of oppression:
øa
—U K TU)« —uB « «c „¡U s
« s ∫d
b*«
øÍbO U
¨dA« s X« ÆwUO s ¨w
—U s UO tO b« ô wMô ∫·ËdF
Æ…dA ¨Ê«uO ¨XJ« […] ÆdA« s p« w p« √b ∫d
b*«
ÆÍbO U
¨rJ
—U s 5 FM« s
cN wM OH «cJË ∫·ËdF
22ÆÃd« ¨Ãd« ÆŸud'« UN
« ÜNK UOb« s pOHU ∫d
b*«
The manager: Where did you get this mistaken conception of history?
Ma¨ruf: Because I do not find in it anything of my own history, of my
life. Am I not human, sir?
The manager: I did begin to doubt whether you are human. […]
Silence animal, insect!
Ma¨ruf: And with these two epithets you threw me out of your history,
sir.
The manager: I’ll throw you out of this whole world, you rat. Get out
of here.
In this dialogue Ma¨ruf is equated with various animals, and de-
prived of his association with humankind and its history. When he
says to the manager that ‘you threw me out of your history’ he has
removed himself from the rest of mankind, to whom the manager
belongs. The manager treats the story’s hero with disdain, and refers
to him insultingly as an ‘animal’, ‘insect’ and ‘rat’. The use of the
names of animals which are considered pests is intended to mark in-
tellectuals as socially and politically harmful beings, who pose a dan-
ger to the very foundations of the regime.
The passage quoted above tells the reader that those in power os-
tracize and criminalize intellectuals; in other words, they refuse to
22
Gha’ib Tu¨ma Farman, Alam al sayyid Ma‘ruf (Beirut 1982), 113 14.
235
recognize their legitimacy and their relevance. If intellectuals are ani-

mals or insects, they clearly cannot be considered as thinking people,
capable of making decisions. In this way the regime in Iraq signals to
its Leftist intellectuals that they have no share in their country, that
they are outsiders with no right to exist, and therefore have no share
in human history, either. Leftist intellectuals in the government’s
view are nothing more than a kind of temporary outgrowth of which
no account should be taken. The short novel’s constant preoccupa-
tion with human history derives from the fact that after all Commu-
nists strive to change local history by means of a world-wide prole-
tarian revolution which shall replace the existing regime with a dicta-
torship of the proletariat. Since this is the case the regime, in order to
prevent the intellectual from changing history through revolution,
drives him out of history.
It must be remembered that intellectuals in general, in any society,
usually perceive themselves as the promulgators of social change.
They do not necessarily express this in ways which have an effect on
everyday life; they do not always come out with declarations or lead
social or other rebellions. Sometimes it is only their writings which
sow the seeds of novel ideas for the evolution of society. Once an in-
tellectual has been robbed of his legitimacy, he can no longer fulfil
his role of social catalyst. His motivation to improve his society and
make it progress is dampened when he is told he is no longer a part
of that society. In the passage above Farman gives an artistic descrip-
tion of the Iraqi regime’s attempt to push intellectuals away from the
centre of social action.
Other passages in this work describe relations between the edu-
cated classes and the Iraqi authorities in terms of predator and prey
in the wild. In the following excerpt the main character, Ma¨ruf,
speaks about the history of jerboa hunting among the Arabs:
Ê« vA« ô ¨Ÿud'« Ü«Ë Æ¡UA ULM
«Ë ¨¡UA U v Ë ¨¡UA U —b ¨Áb ¨qCH
vA«Ë Æ…—UF« `
dB ¨p– vK n¬ XM uË ¨wH 5Ë wMO tu« U «c Ë« ¨qƒ«
wMF
ÆÆÆbOI dO u«Ë ¨bOL dO UO « Ê« ¨d{U(« Xu« w ULO ôË ¨ÁUA« U
23
ÆŸud'U
Please, hunt it as much as you want, whenever you want and wherever
you want. I, the jerboa, am not afraid of being eaten, or that’s what
I say to myself, although I regret it, in all honesty. What I am most
afraid of, especially at present, is that I shall be unwanted, and die
unmissed, just like a jerboa.
23
Farman, Alam al sayyid Ma¨ruf, 123.
236
Ma¨ruf ’s declaration that he is not afraid of being eaten symbolizes

the sense of persecution and victimization felt by Iraqi intellectuals,
whom their rulers perceive as dangerous. He is prepared to be ‘eaten’,
that is to be downtrodden by the authorities, but not to give up his
principles. In the foregoing passage the animal knows that it will be
caught and eaten, just as the intellectual knows that the government
will repress and perhaps even liquidate him, because of the danger he
represents. This is a story of foreordained death. The irony embodied
in the story is that whereas the bosses liken the main character to an
animal, in fact the opposite is the case: he turns out to be more hu-
man than the rest, and it is they who let their darkest animal in-
stincts come to the fore.
The Syrian writer Zakariya Tamir (b. 1931) has said that the vio-
lence in his stories is no more than a reflection of an Arab intellectu-
al’s everyday life, which is filled with violence and bloodshed, impris-
onment and disappointment.24 Farman, just like Zakariya Tamir, ar-
gues through his books’ characters that intellectuals, and anyone else
who dares express opposition to the regime, are still being hunted
down, like animals that are hunted and then killed. This provides yet
another example of Iraqi intellectuals’ sense of being persecuted by
their government. Farman believes that Arab leaders systematically
liquidate the intellectuals produced among their own people. The
animal’s death is known beforehand, as is the intellectual’s persecu-
tion. Farman is thus saying that the lives of intellectuals in the Arab
world are as precarious as those of prey, as long as they do not toe the
line dictated by the regime.
Imagery borrowed from the animal kingdom appears also in de-
scriptions of characters’ looks and behaviour, as in the following two
examples:
wMO w È«— ¨WuD*« V J« s WOU WFœ t ÂbIO rU bO« V J s Uœ 5 Ë
¨…b «Ë WE!K ¨tUO s X! Wd H WU « ¨«b W
d ¨d« Ÿu s WU « —d;«
25
ÆW —U« WU «
As he approached Mr. Hashim’s desk in order to give him a second
bundle of typed letters he saw a different kind of smile in the editor’s
eyes, a very strange, predatory smile which for a moment erased yester-
day’s smile from his memory.
24
Robert B. Campbell (ed.), Contemporary Arab Writers: Biographies and
Autobiographiess (Beirut 1996), 394 5 (Arabic).
25
237
ÆÆÆWOM ôË ÜNO …Ëö ô ÆW"O)« WU ô« fHM tO« Ê«dEM rU bO« UMO XU
26
Æ”«d «
Mr. Hashim’s eyes looked at him with that same evil smile, completely
lacking in sweetness or pity … rapaciousness.
Farman’s choice of animal life, especially of predators in the forest,
for his imagery is not random. A predatory smile represents a preda-
tory person, in this case the boss lying in wait for the story’s hero.
This metaphor is not meant by the author to describe that person’s
actual smile, but rather to show the real intentions hiding behind the
smile; it is not a description of a specific kind of smile, but an at-
tempt to reveal the evil intentions it masks.
Farman’s novel al-Markab (The Ship, Beirut 1989) deals, among
other things, with dilemmas concerning the old versus the new, Is-
lamic tradition versus secularization, relations among young people,
and the status of women in the Arab world today. One of the novel’s
central characters is Suham, an educated young woman who must
cope with attempts by the management at her workplace to prevent
her promotion, using various means including the spread of false ru-
mours and slanders about her. Just as in Alam al-sayyid Ma¨ruf, here,
too, the directors in Suham’s workplace can be understood as alle-
gorically representing Iraq’s political leadership, which uses informers
to keep an eye on every stratum of society. Jabir is the ‘spy’ sent to
‘sniff ’ around Suham and to follow her around. When she sees him
she describes him as a predator lying in wait for its prey:
27
ÆÂœ s ÊU FI ÁUMOË ¨W!OI« t U « r
Ád¬ w UH«Ë t
«d ¨Ã—b« bF« XM
I went up the staircase and then saw him standing at the top, smiling
his ugly smile, his eyes two spots of blood.
The two bloodspots (literally: ‘pools of blood’) in the eyes evoke
the image of a predator lying in wait for its prey, of a beast whose
eyes are bloodshot with full veins as it thirsts for blood. Iraq’s blood-
thirsty intelligence services have been widely described: the Iraqi
Communist poet Badr Shakir al-Sayyab (1926–64), for example,
wrote a famous poem entitled ‘al-Mukhbir’ (‘The Informer’), about
the Iraqi regime’s spying activities directed against its own people.
Some have gone so far as to call Iraq a ‘police and spy state’.28 Spies
26
27
Farman, al Markab (Beirut 1989), 113.
28
Salim Fattal, In the Alleys of Bagdad (Jerusalem 2003), 310 11 (Hebrew).
238
reached into people’s very homes, so that family members were afraid
to express their thoughts even within their own four walls lest a fam-
ily member inform on his brother or father.
In addition to metaphors derived from the world of predators
Farman also mentions some wily creatures who entrap their prey,
such as the spider. When the hero of Alam al-sayyid Ma¨ruf spoke
with his superior,
29
Æt«— ‚u oKF uJM ZO ¡«—Ë s tOK qD
œUMF« ÊUDO È«—
He saw the devil of stubbornness observing him from behind a spider-
web hanging above his head.
In the Qur’an the spider symbolizes an unreliable crutch, on
which those who followed other gods instead of Allah leaned,30 a
creature which wages war against its potential prey by weaving its
web to entrap it. In the same way Farman’s hero feels as if he were
prey about to fall into a waiting spider-web. There is an obvious as-
sociation between the manager, symbolizing the Iraqi regime looking
for a chance to pounce on any man of conscience standing in its way,
and the spider, a crawling predator which entraps insects and de-
vours them after they are caught in its web.
The phrase ‘hanging above his head’ creates an association with
the hangman’s rope, with which not a few Leftist intellectuals made
their acquaintance during the second half of the twentieth century.
Furthermore, the image of the spider-web adds to the sense of hope-
less entanglement, a feeling which accompanied Farman throughout
his years of wandering, for he had the unpleasant experience of los-
ing his citizenship, and of course his passport, too, because of his
political opinions. To a large measure his Leftist stand drove him
into homelessness.
By his skillful manipulation of similes and metaphors borrowed
from the domain of predators and relations in the wild, Farman
manages to convey the sense of persecution felt by Leftist intellectu-
als in Iraq, who live under a regime which constantly threatens their
freedom of expression and of action, and often their very lives. In
fact, it can in fairness be said that the use of metaphors borrowed
from the animal world enables the writer to express his displeasure,
to say the least, at the regime’s despotic attitude towards Iraqi intel-
lectuals, in a way which circumvents the censorship which is often
imposed on written works; this is so because as long as the criticism
29
30
Qur’an 29:41.
239
is in the form of metaphor and not explicit, the authorities have little
excuse to attack the author, although writers have been subject to
limitations even with no apparent justifiable cause in Iraq and in
other Arab countries. Metaphors thus serve in this case both as a
medium of expression and as a kind of armour, protecting the author
from being stifled by the authorities.
A sense of alienation and of not belonging — the experience of exile
Farman’s second aim in his use of imagery borrowed from the animal
world is to express the kind of feelings of alienation and of being a
stranger which can be experienced by a private individual even within
his own family and people. Exile, physical or metaphorical, has been
the fate of many Leftist intellectuals in Iraq throughout the second
half of the twentieth century. Many went into exile; some were forced
to leave, others went of their own volition, but often it was a combi-
nation of the two. The process of looking for a way out of the intellec-
tual siege under which the Iraqi regime placed its intellectuals in-
volved much pain, which began in the homeland, where intellectuals
felt they were in internal exile, since they were unable to find work
because of orders from above, they were persecuted for their political
opinions, and their freedom of expression was curtailed. The next step
was exile in another Arab country; there the intellectual would try to
earn a living, but usually found that the restrictions on his freedom of
expression were no less stifling than before. Some intellectuals there-
fore left the Arab world and went into exile in Western Europe. There,
despite being relatively free to express themselves, they pined for their
homeland and their native tongue and did not therefore become fully
integrated into the local society and its culture. Farman’s friend, the
Jordanian-born Saudi author Äbd al-RaÌman Munif (1933–2004),
added that as if these difficulties were not enough the host country, if
it did not extradite the exiles to their country of origin, placed restric-
tions on their activity that turned them into virtual prisoners or hos-
tages.31 At any rate, the usual path which the life of an intellectual
took was: internal exile → exile in another Arab country → exile in
the West.
The experience of exile, with its attendant feelings of alienation
and of being an outsider, is often expressed in Farman’s writings by
means of metaphors borrowed from the animal kingdom. In the
novel al-Makha∂ (Labour Pains, Baghdad 1974) Farman tells the
story of Da’ud Karim, an Iraqi student abroad. The reader is exposed
31
Äbd al RaÌman Munif, al Katib wa’l manfa humum wa afaq al riwaya al
ärabiyya (Beirut 1992), 89.
240
to the persecution of Iraqi Communist students through Karim’s

reminiscences, upon returning to his homeland, of his own student
days. The novel tells of students arrested, and sometimes also ex-
pelled from the university, by the authorities for their political and
ideological views. The feelings which accompany the novel’s main
character throughout, and which are passed on to the reader, are of
alienation and exile (ghurba) within one’s own homeland. Thus to
the external, physical exile of a student in a far land, as described in
other works by Farman, an internal, mental exile is added. When the
hero of the novel returned from a journey to southern Iraq he had
the feeling that:
32ÆÊ«uFôU w Ë— ‰u n «Ë se« ‰UD «Ë ¨W
œUF W
d Êb*« bË
The cities seemed strange and hostile, and time extended and wrapped
itself around my soul like a viper.
In this case the image of the viper clearly reflects the student’s feel-
ings of being completely alone, unable to find his childhood home
which had been destroyed in his absence. Loneliness, alienation, the
sense of being a stranger, of not belonging, deprivation and distress,
are all felicitously expressed by metaphors taken from the domain of
predatory crawlers; these images thus serve as a tool for reflecting the
character’s inner emotions.
The confusion and lack of confidence which Farman’s heroes ex-
hibit are also expressed through the metaphor of the snake in his
novel Åilal äla al-nafidha (Shadows on the Window, Beirut 1979),
which tells the story of a family which moves out of one of Bagh-
dad’s poorer neighborhoods to a better part of the city. One of the
children, Majid, had studied engineering abroad. His stay abroad he
perceived as banishment (‘wHMU Ã—U)« v« dH« ÊU’), but upon return-
ing to Iraq he felt like an exile inside his own home. His feeling of
strangeness and discomfiture are expressed through images borrowed
from the animal world:
33
ÆwUL« w WLzU XU WU ŸU« XKLKL
… poisonous snakes asleep in my depths moved restlessly.
Majid can find no peace, either at home or outside. In an internal
monologue he says that his years in exile make him feel as if he were
walking on broken ground (‘…dË ÷—« w dO« wMU wdFA WdG« «uM’),
32
Farman, al Makha∂ (Baghdad 1974), 37.
33
Farman, Åilal äla al nafidha (Beirut 1979), 183.
241
whereupon he tries to renew the lost sense of intimacy (‘WH√’) be-

tween him and the places he left.34 He does not find his place within
Iraqi society, nor even within his own family. When he does make
new friends he finds it difficult to understand their language, which
is derived from their life experiences, and these differ from his own,
because of his extended stay abroad. He compares their strange lan-
guage to an explosive (‘…dH …œU’); in other words, they can hardly
communicate with him. He feels that his own self is not natural, and
grows like an abnormal thyroid gland or gets crushed like a miserable
bug (‘WKO– …dA ”d WËdN Ë« ¨WO—b« …bG« rC WLC «c«’).35
Farman evokes metaphors based on the animal world twice on the
same page for the purpose of describing a feeling of estrangement, a
lack of self-esteem, and the haunting memories of the past.
The image of the snake is repeated elsewhere in the novel, as part
of the hero’s soul-searching. There the snake symbolizes the hero’s
helplessness, which he feels burning inside him:
36 ÆUbK«Ë ÂuL« »Ëd{ wMI
c wUôU eF« s Ê«uô« Ác qË
And all these shades of impotence, like snakes, give me a taste of
venom and bites.
Farman quite purposefully chooses a crawling, poisonous creature
to reflect the mental distress of an intellectual who returns to his
homeland and cannot find his place in it.
Yet another image from the animal kingdom which Farman uses
to convey the individual’s feeling of impotence is that of a turtle ly-
ing on its back:
¨¡«uN« ⁄«d w wK—U f—« ¨…UO(« TU »d UdN# vK WuKI …UH!K Ëb« U«Ë
wF{Ë v« œu« Ê« v ¨TUA« q— vK ¨”QO Õe e«Ë ¨bU'« ¡UL« tË wMKUI
37ÆÆÆwFOD«
I seem like a turtle on its back near the shores of life, kicking my feet
up in the empty air, greeted by the silent face of the sky. I move in
desperation on the sand of the shore, wishing I would return to my
natural position …
Such are the feelings of a person who returns to his homeland and
does not find his place: he is detached from his surroundings and
34
Farman, Åilal äla al nafidha, 182 3.
35
Ibid.
36
Farman, Åilal äla al nafidha, 285.
37
242
subsists helplessly on life’s margins, at their side and not within. The
animal-world imagery also reflects the author’s personal experiences
in a natural way: like the hero of Åilal äla al-nafidha, Farman was
not welcomed by the authorities when he visited Iraq after having
been exiled from it, and was therefore unable to renew his contacts
there and become reintegrated into its society. Farman was always
considered an undesirable and, as he often admitted to his friends,
he was constantly afraid that his passport would not be renewed.38
The travails, the alienation and the loneliness of the intellectual in
Farman’s world can be seen in most of his works. His characters of-
ten feel trapped and in a mental dead-end. Despite the optimism
which he was so careful to preserve in his conversations with his
friends even after decades of exile, the depressive mood and eco-
nomic and physical difficulties inherent in a life of wandering like
his become evident through his imagery. His works are full of auto-
biographical allusions, which can be deciphered by means of the
metaphors he uses:
¨¡uC« XU ¨qOK« w ¨W d{ô« ÊuJ sU XO« Æl Ub t√ WU t KI «
v« VI s W
d! »—UIF« qI MË ¨WLKE« U
«Ë“ w dO«dB« wMG ¨¡UA ô« sH
39ÆVI
His mother’s clock greeted him with nine strokes. The house is as quiet
as a grave at night, with weak light and rotten innards. Grasshoppers
sing in the dark corners and scorpions run freely from one hole to an-
other.
Here Farman creates a pun, since the word äqarib means both
‘hands of a clock’ and ‘scorpions’. Throughout the story the hero is
overwhelmed by his feeling of being persecuted by superiors and
of living in conditions which make it impossible for him to progress,
just as the scurrying scorpions symbolize an immediate and grave
danger.
A person’s inner feelings are often projected unto his surroundings,
and here the hero’s distress is reflected also in what he sees:
40 ÆWd H U«uO ”Ëƒ— q" uO« WO« ‚u s qD —Uô«Ë
The trees tower above the fences of the houses like the heads of beasts
of prey.
38
Mentioned by the writer Äbd al RaÌman Munif at a gathering held after
Farman’s death: Munif, al Katib wa’l manfa, 118.
39
40
243
The hero feels out of place in his family and within Iraqi society.
When he walks along the street he feels that even the Baghdadi land-
scape, the trees lining its streets, are threatening and fraught with
danger. The sense of pressure and hopelessness is expressed also
through pseudo-legendary images, such as that of the monster
tinnin, a metaphor for the fictional narrator’s claustrophobia, which
takes away the air from the hero’s lungs:
s t U oM
œUJ
Ë ¨5 zd« s ¡«uN« h 1Ë ¨fHM« q«œ gOF
5M XLB«Ë
41
Æq«b«
Silence is like a tinnin living inside the soul, drawing the air out of the
lungs and almost choking their owner from the inside.
This metaphor induces a sense of suffocation and of being
hemmed in. The tinnin, a snake-like or fish-like sea monster, which
is here represented as a creature whose thirst knows no bounds, is
considered a harbinger of misfortune. According to one of the beliefs
associated with it, if a sick person dreams about a tinnin it means
that his end is drawing near.42 The novel’s hero is in internal exile,
surrounded by a sense of being a stranger in his own country; his
silence feels to him like a tinnin within the soul, sucking away all his
strength.
In this way Farman uses the tinnin and the other creatures in or-
der to describe the feelings of estrangement and alienation which
surround the characters in his writings even inside their homes and,
through them, he characterizes his own feelings as well.
The image of the city in Farman’s works
Samia Mehrez has described Nagib MaÌfuz (1911–2006) and Gamal
al-Ghi†ani (b. 1945) as historians of Cairo, since both in their writ-
ings have described that city over an extended period, in a sense
documenting its history.43 Like these two, Farman can be considered
a documenter of Baghdad. One of the ways in which he conveys the
way he remembers the city is by the use of imagery borrowed from
the animal world, through which he depicts Baghdad’s winding al-
leys as a place which strangles and imprisons, and in which the indi-
41
42
Kamal al Din MuÌammad bin Musa bin Ïsa al Damiri, (notes and intro
duction by AÌmad Îasan Basj), Îayat al Ìayawan al kubra (Beirut, 1994), vol. I,
238 9.
43
Samia Mehrez, ‘Re Writing the City: The Case of Khitat al Ghitani’, in
Egyptian Writers between History and Fiction (Cairo, 1994), 58 77.
244
vidual feels lost. By doing so Farman gives artistic expression to the

tendency among Socialist writers to describe the city as an oppressive
place, where authority strips the individual of his rights.44
The novel Khamsat aÒwat (Five Voices, Beirut 1967) depicts the
lives of five intellectuals living in Baghdad, each with his own unique
personality and his own life. Each chapter in the book relates the life
of one, and when the five meet, the reader is given a picture of their
common experiences in café society in Baghdad of the nineteen-
fifties. As Saïd, one of the five, walks the streets of Baghdad looking
for the address of a woman who asked him for help, the winding
alleys give him the feeling of being trapped within a maze:
45
ÆhK K UNu q ed üJM WJ w …Q XDI WK/ q" …
Like an ant which suddenly fell into a spider-web and then mustered
all its strength to escape.
Saïd feels ‘tossed’ from one alley to another, in an expression of
protest against the lack of planning in the life and government of his
country, which results in complications and helplessness. He feels
driven through alleys which come together, split apart, become ever
narrower and turn around each other, like a giant octopus from
whom escape is impossible:
46
ÆqzU ◊uD« q—« q" W“ô« t –UI
The alleys tossed him like the arms of a tremendous octopus.
In these passages the old, unplanned city of Baghdad, in which
many originally temporary structures have become permanent fea-
tures, is described as a maze which mocks and confuses those who
walk its streets. As in the case of the spider, the octopus here repre-
sents the many imaginary arms which surround those who walk in
the city; this city, where the hero was born and in which he lives,
occasionally closes in on him and makes him feel claustrophobic.
The sense of stifling is even stronger in Alam al-sayyid Ma¨ruf, where
the hero feels lost in a maze and the city is compared to a cocoon
which completely shuts in the larva:
WIdAU ÜNH vK WH K*« t M
b w ¨—«b n« ¡«—Ë s lM fLA« XU bI
47ÆW“ô« WIOC« ¨Ê«—b'« WIö *«
44
Shmuel Moreh, ‘Town and Country in Modern Arabic Poetry from Shawqi
to Al Sayyab’ in Studies in Modern Arabic Prose and Poetry, (Leiden 1988), 136 60.
45
Farman, Khamsat aÒwat (Beirut 1967), 8.
46
Farman, Khamsat aÒwat, 7.
47
245
The sun would rise from behind a thousand walls in his city which
wound around itself like a cocoon, with its closefitting walls and nar-
row alleys.
The metaphor of the cocoon is somewhat misleading, since a co-
coon protects the pupa as it is transformed into a butterfly, whereas
the phrase ‘behind a thousand walls’ indicates fear, reservation and
confinement, in a way which is reminiscent of a number of lines by
the Iraqi poet Buland al-Îaydari (1926–96), also an exile:
œ«bG wœ—UD
wdU%
48Æ…¬d*« U
«Ë“ q w
Baghdad is pursuing me
It is besieging me
In every angle of the mirror.
In his very first novel, al-Nakhla wa’l-jiran (The Palm Tree and the
Neighbours, Beirut 1965), Farman tells the story of Salima the baker,
her stepson Îusayn and other wretched characters living on the mar-
gins of Baghdadi society during World War II. This work is consid-
ered the first mature novel ever written in Iraq, addressing the com-
plex inner world of its characters. As the main character, Îusayn,
walks through the streets of Baghdad at night the narrator provides
the reader with a depiction of poverty, filth, disorder and negligence:
q bMË ÆWULI« Â«uU …¡uK2Ë ¨Wu« …UO*U WKKË …—cË ¨WA u qOK« w œ«bG XU
49ÆWU{ WDË ¨VzU VK Wu
Baghdad at night was desolate, filthy, moist with dirty water and filled
with heaps of garbage. At every heap there was a stray dog and a lost
cat.
In this passage Farman makes use of animals to stress the neglect
found in Baghdad, the piles of garbage which attract hungry stray
beasts. Farman emphasizes the city’s unsightliness, in the tradition of
Socialist writers who view the city as a place where individuals are re-
pressed by the authorities, where the neglected masses are incapable
of taking their fate into their own hands and to make for themselves
48
Buland al Îaydari, ‘Fi †ariq al hijra min Baghdad’, al A¨mal al kamila (Ku
wait, 1992), 671.
49
Farman, al Nakhla wal jiran (Baghdad [?] 1988), p. 134 (first published in
Beirut, 1965).
246
a life in which man, not fate, decides.

To conclude, Farman inserts metaphors and similes borrowed
from the animal world into his descriptions of Baghdad. Through his
choice of beasts the reader is introduced to the author’s negative view
of the city, a view which is in keeping with the tradition of romantic
Socialists.
Unattainable peace and happiness
The previous examples represent dark periods in the lives of the
characters and of Farman himself. In contrast, very occasionally one
also finds in his works imagery borrowed from the animal world and
used to represent unrealized happiness. Such happiness is almost
within the hero’s reach, but in the end the attempt to attain it fails
and he is left mute and alone. The animals in this case are usually
with wings — birds or butterflies, which symbolize a freedom which
Farman unfortunately did not possess.
Such metaphors are encountered in Alam al-sayyid Ma¨ruf, where
the hero ‘meets’ the sunset every evening, treating it as a lover:
s tKI « «–« ÊU ÆUNK UOb« w Èd« WM UNœUF ô ·ËdF bO« Èb WM »ËdGK
—uOD« ô ¨tu ULO d« ¡w Í« t
uN
bF
r dNM« vK qD*« vNI*« w ÍËeM*« tM—
s VOG v ¨WUHA« ¡UL« w …bFB ¨dNM« —b vK WCIM tu ·d XU w «
50[ÆÆÆ]UNUMË v« ¨dô« d« w ¨œuF —uOD« Ê« ·dF
tô ÜOb«
The sunset possesses a charm for Mr. Ma¨ruf which no other charm in
the world can equal. Whenever he encountered it from his isolated
corner in the coffee-shop overlooking the river nothing else around
him could lure him: not the birds which hovered above him and dived
into the river, rising high into the clear skies to disappear from the
world, for he knew that the birds would eventually return to their
nests […]
Birds have nests to which they return and from which they fly
away whenever they want. But Farman is denied this right; every
time he comes back to Iraq for a visit he is unsure whether the au-
thorities will let him leave again. Iraqi intellectuals were usually not
permitted to make private journeys abroad, and were never certain
whether they could enter or leave again.51
Another metaphor for unrealized freedom and happiness taken
from the animal world is the butterfly, symbolizing beauty, lightness
50
51
Sulayman al Äskari, ‘al Muthaqqaf al ïraqi min al ÌiÒar al siyasi ila al manfa
wa’l mawt’, al Ärabi, 2001, issue 511, 8 13.
247
and mobility. Ironically, Farman, who throughout his adult life wan-
dered from one country to another until he settled in Moscow, did
so without the freedom enjoyed by the unfettered butterfly:
b
d
ôË ¨ÂUM*«Ë WEIO« 5 Á¬— rK w tQË ¨Ác WLU« …uAM« WE( qOD
Ê« b
d
ÊU
nOD« tM VOG
ô v ¨dU rK ô« u U p– Ê« dJH
Ê« b
d
ôË ¨tOMO ` H
Ê«
Ë« Æb
dA« ‰UO)« «eHË ¨”UM« ÂU “ 5 ¨se« s WKH w ÜHD t d Íc« VO(«
Ê« ·U
uË ¨wU“ Õb w UNF{u ¨WuK W«d t
b
w XFË qH q" ÊU
UNd Æu9Ë W«dH« oM Ê« WU ÕbI« oM vK tH lC
Ê« UC
« vA
Ë ¨XKH
52Æ»dNË ¨W! H« v« Íb N Ê« v« ÕbI« w »dDC
He wanted to extend this moment of pale intoxication, in which he

was as if in a dream he saw between waking and sleep. He did not
want to open his eyes, nor did he want to think it was nothing but a
passing dream, lest he lose hold of the beloved vision which passed by
him quickly, out of the forgetfulness of time, among the crowds of
people and the leaps of displaced imagination. Or as if he were like a
child into whose hands dropped a colourful butterfly, and who then
put it into a glass jar in fear that it would slip away, but also feared to
place the palm of his hand on the neck of the jar, lest the butterfly suf-
focate and die. He let it move about inside the jar until it found the
opening, and escaped.
In addition to freedom, the butterfly also symbolizes something
slippery which cannot be easily obtained, a wish which escapes be-
tween one’s fingers in seconds and does not come true. Within the
present narrative context it serves as an anticipatory clue which tells
us the readers that the Leftist intellectual’s mission to base the state
in which he lived on care for the weak is bound to fail. A butterfly is
very short-lived and therefore, although it symbolizes freedom, its
impending death provides yet another hint at failure — the failure of
the main character’s and Farman’s social mission.
To conclude, medieval writers used metaphors which fit the repre-
sented object only with respect to its external aspects; object and
metaphor were fully conformable to each other in size, form and col-
our. With time metaphors changed their character, especially after
European romanticism entered Arabic literature, and writers moved
from concrete imagery to metaphors intended to express their inner
world.
To take metaphors borrowed from the animal world as an exam-
ple, in addition to the use which modern Arab authors make of
them, for example for describing their character’s external features,
52
248
Gha’ib Tu¨ma Farman also exploits them for conveying to the reader
as authentic a sense as possible of his own inner world and his frus-
trations as an Arab intellectual, as well as a window into some topics
which were of paramount importance to him as a Communist Iraqi
writer. Unlike medieval Arab authors, Gha’ib Tu¨ma Farman’s im-
agery is an instrument in the service of his ideology. In other words,
Farman provides a case-study of how Arab metaphor has changed
under the influence of Western culture, especially Communism.
First of all, Farman’s animal-world imagery provides a concise,
precise and fascinating description of the relationship between intel-
lectuals and the regime in mid-twentieth-century Iraq. It emphasizes
the sense of persecution which Leftist Iraqi intellectuals suffered dur-
ing the period of intense anti-Communism, and also the regime’s
arrogant behaviour towards Leftist intellectuals, whom it demonized
and de-legitimized. Farman with his metaphors conveys to the reader
the emotional world of an intellectual who wishes to deliver his
country from its backwardness and his people from its degenerate
religious beliefs, by means of an ideology which would free it from
the shackles of the past. This man was persecuted inside his own
homeland and as a result was forced into exile, just as Farman him-
self was.
It is not by chance, therefore, that the political leadership is com-
pared to a beast of prey, since the Iraqi authorities treated Farman
and his colleagues in a ‘predatory’ way, just as they treated Iraq’s reli-
gious minorities. Nor is it by chance that Farman’s heroes are com-
pared to hunted prey. The hunt is a metaphor for the regime’s ‘cap-
ture’ of intellectuals, whom it does not permit to disseminate their
ideas, to make a living in Iraq, and to freely publish their works.
The species of animals chosen by Farman are also intentional. It is
not by chance that the animals he uses bite, lure their prey cleverly
into a web, or possess many long arms. All these are intended to
serve his purpose, namely to provide a concrete description of the
lives of Leftist Iraqi intellectuals from the nineteen-fifties on, who
were trapped in the talons of the regime. In addition, the animal
metaphors can be said to have provided the author with a modicum
of protection against accusations by the authorities. After all, if noth-
ing explicit is said, and all is expressed in the form of metaphor and
simile, the regime supposedly cannot charge the author with any-
thing, although in fact Arab regimes have never been in need of con-
crete proof in order to arrest or exile a writer.
Secondly, animal metaphors help Farman describe the alienation
and solitude felt by his characters, in Iraq and abroad. An intellectu-
249

REVIEWS
GEORGES BOHAS and MIHAI DAT, Une théorie de l’organisation du lexique des langues
sémitiques: matrices et étymons (Collection Langages). ENS Éditions, Lyon 2007.
Pp. 235. Price: /24.00 paperback. ISBN: 978 2 84788 076 2.
PHILIPPE CASSUTO and PIERRE LARCHER (eds), La formation des mots dans les langues
sémitiques (Langues et langage 15). Publications de l’Université de Provence,
Aix en Provence 2007. Pp. 202. Price: /25.00 paperback. ISBN: 978 2 85399
660 0.
Two French publications dealing with word formation in Semitic languages ap
peared in 2007, one by G. Bohas and M. Dat, the other one by a team of ten
authors, who have participated in a syposium held in 2003 at Aix en Provence. As
expected, the first work provides further developments and illustrations of Bohas’s
theory on the ‘matrix’ and the ‘etymon’ (cf. JSS 50, 2005, pp. 205 7), especially in
Arabic and Hebrew. The starting point is the existence of primitive biconsonantal
bases C1C2, called ‘etymons’. The third radical is an ‘increment’ serving to express
a particular connotation of the basic meaning and leading to the formation of
three consonantal radicals. This is realized by various means, in particular by gemi
nation, prefixion, suffixion, metathesis and reduplication. Chapter I presents ex
amples in Arabic (pp. 41 68), while Chapter II examines the question in Hebrew
(pp. 69 97). An important change in Bohas’s theory is the welcome recognition
that also w and y may be constituent phonemes of the ‘etymon’ (p. 49, n. 13). The
‘matrix’ consisting of reversible phonological elements is then analysed in Chapters
III (pp. 99 132), IV (pp. 133 46), and V (pp. 147 54) on the basis of several ex
amples in Arabic. Chapter VI deals thereafter with similar cases in Hebrew (pp.
155 96). The consequences of the theory are developed in Chapter VII (pp. 197
222). They lead to the conclusion that a ‘matrix’ obtaining a similar meaning in
various language families does not prove their common origin. In fact, the ‘matrix’
as linguistic sign preserves some perceptible features of the object or action signi
fied, unless the ‘mimophonic’ weight of the sign is lost in the course of linguistic
evolution, distorting the phonological ‘mimicry’ of the ‘matrix’. In other words,
the original linguistic sign is not arbitrary, contrary to Saussure’s opinion, but has
onomatopoeic origins. The theory is based on correct observations, unnoticed by
R. Weipert (ZDMG 150, 2000, pp. 656 60). However, it operates with abstrac
tions, since exclusively consonantal ‘matrices’, ‘etymons’, ‘increments’, and ‘radi
cals’ cannot shape a living language. The question is how vocalic phonemes at
least a and i/u can be integrated in this system. The use of epenthetic vowels,
common in Maghrebine Arabic and in Gurage dialects, does not answer the ques
tion, for vowels can be structural elements of the language, just like the semi
vowels w and y.
Related questions are discussed by some authors in the second work, divided in
four sections. In the first one, Ph. Cassuto deals with the classification of words ac
cording to mediaeval grammarians of Hebrew (pp. 15 43). V. Porkhomovsky then
presents the conception of root and word formation according to earlier Russian
Semitists (pp. 45 52), in particular A.M. Gazov Ginzberg, who in 1965 antici
pated Bohas’s theory on the onomatopoeic origins of the Proto Semitic roots, N.V.
Yushmanov (1896 1946), who assumed that Proto Semitic had archiphonemes or
251
REVIEWS
‘diffused’ phonemes, S. Mayzel (1900 52), who accepted the triconsonantal origin
of the Semitic root, and I.M. Diakonoff, whose Afrasian Languages (Moscow 1988)
reflects a mainly diachronic perspective. Asking in his study Whether Language was
Imitative in Its Origins? (Moscow 1965 [in Russian]), Gazov Ginzberg distinguishes
four types of word formations by imitation of natural sounds: the ‘internal imita
tion’, based on human sounds, like snoring, laughing, crying, etc., the ‘external imi
tation’ of animal voices and all kind of noises, the gestures of oral organs, like open
ing or shutting of the mouth, and the babbling or nursery words. The question is,
of course, whether onomatopoeia was the principal source in the origin of language
or has played only a limited part. Attested phonetic differences provide the starting
point of Yushmanov’s research. In his posthumous Studies in General Phonetics,
Semitics, and the Morphology of Classical Arabic (Moscow 1998, pp. 126 90 [in
Russian]) he assumes that five ‘archiphonemes’ existed originally in Proto Semitic: a
dental (*t), a velar (*k), a liquid (*r/l), a labial (*p), and a nasal (*n). This means, for
instance, that the distinction of voiced, unvoiced, and emphatic consonants repre
sents a differentiation of the basic sounds in the course of linguistic evolution. The
same ought to be said about the spirantized and palatalized occlusives, changes that
can happen merely in order to ease the articulation. Yushmanov’s hypothesis is thus
basically diachronic and it could be reformulated nowadays in the broader frame of
the Afro Asiatic language family. Therefore no fundamental difference seems to ex
ist between Yushmanov’s theory and Diakonoff ’s approach, contrary to what
Porkhomovsky would seem to suggest (p. 52).
The next contribution, by J. F. Prunet, provides a critical overview of linguistic
approaches to the problem of the Semitic root, adding a large bibliography
(pp. 53 80). Prunet considers the ‘consonantal root’ and the ‘vocalized root’ to be
a single, indivisible morpheme, but he seems to regard the choice of the vowel in
particular roots as an empirical factor, which is at least questionable. In the second
section, the accurate analysis of roots and morphemes by Chr. Touratier (pp. 83
95) shows that the grammatical system of Semitic consonantal roots results from
an improper projection of the consonantal script into the realm of spoken lan
guage. P. Larcher’s study of the interrelation between root and pattern leads to the
conclusion that these grammatical categories do not suffice to explain the forma
tion and meaning of particular words (pp. 97 112), while L. Edzard examines
blends and compounds in the word formation of modern Semitic languages,
i.e. Arabic, Hebrew and Amharic (pp. 113 47). In the third section, M. Tosco
discusses the crisis of Semitic morphology in Maltese (pp. 151 63), while A.
Zaborski formulates some risky hypotheses concerning verbal formations in
Semitic and Cushitic (pp. 165 72). The last section of the work contains a discus
sion of problems related to the gender and number of nouns by H. Gabrion
(pp. 175 84) and an examination of the construct state in Babylonian by R.
Mugnaioni (pp. 185 200).
The study by G. Bohas and M. Dat contains a comprehensive bibliography
(pp. 223 31) and an index (pp. 233 5), while each contribution to the proceedings
of the Aix en Provence symposium is followed by a bibliography. Both works have
a highly theoretical profile, but provide some useful material and new insights for
the understanding of spoken and written Semitic languages.
doi:10.1093/jss/fgn052 EDWARD LIPINSKI
UNIVERSITY OF LEUVEN
252
REVIEWS
AARON D. RUBIN, Studies in Semitic Grammaticalization (Harvard Semitic Studies

57). Eisenbrauns, Winona Lake 2005. Pp. xv +177. Price: $32.95 hardback.
ISBN: 1 57506 923 7.
The aims of the book are stated as being to help bridge the gap between linguistics
and Comparative Semitics by looking at a range of historical developments within
Semitic from the perspective of a particular linguistic process, and to help shed
light on some well known developments that have hitherto not been satisfactorily
explained. The author gives his own definition of grammaticalization as ‘the change
whereby lexical items and constructions come in certain linguistic contexts to lose
their lexical meaning and serve grammatical functions, or, the change whereby a
grammatical item develops a new grammatical function’.
In a chapter on Grammaticalization in Semitic a wide variety of examples of
grammaticalization is presented from both ancient and modern Semitic languages
to illustrate this phenomenon, and these clearly demonstrate how pervasive
grammaticalization is in the development of language. The rest of the book is taken
up with an examination of three kinds of grammatical development in much
greater detail. The first of these is the origin of definite articles in the Central Se
mitic (CS) languages. This group of languages has both prefixed articles (Hebrew
and Arabic) and suffixed ones (Aramaic and Old South Arabian). But as there is no
definite article in Akkadian or Ethiopic, there was almost certainly not one in
Proto Semitic or Proto West Semitic. Thus the definite article must be a develop
ment or a series of independent developments within CS. Rubin accepts the widely
held view that the CS articles can be traced back to a demonstrative particle. The
Akkadian near demonstrative meaning ‘this, those’ has the base anni , which Rubin
derives from an original base of *hanni , and the far demonstrative ‘that, those’ has
the base ulli , from an original *ulli . Rubin argues that the articles of Canaanite,
Aramaic and Old South Arabian derive from the same Proto Semitic source as the
Akkadian *hanni series of demonstratives, while the Arabic article is from the same
source as the Akkadian *ulli series. He looks briefly at how the change of DE
MONSTRATIVE > 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 Reflec-
tions 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
cusses cases in which a modal verb such as ‫י ָֹכל‬, ‫נ ַָתן‬, ‫ָאָבה‬, ‫ֵמֵאן‬, ‫ הוִֹסיף‬or ‫ֵהֵחל‬ is
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
cohortative, i.e. first person ‘imperfect’ verbal forms ending in â, such as ‫ְקטָלה‬ ְ ‫א‬ ֶ
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.
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
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
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
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
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.
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.
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.
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.
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: psc@periodicals.com. 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: jnls.cust.serv@oxfordjournals.org. Tel (and answer-
phone 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: jnlorders@oxfordjournals.org. 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: oku-
daoup@po.iijnet.or.jp. 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 Interna-
tional, 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: helen@oxfordads.com. 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 Akkadi-
schen 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-Vol. 1-2009 PDF

Încărcat de

Informații document

Titlu original

Drepturi de autor

Formate disponibile

Partajați acest document

Partajați sau inserați document

Opțiuni de partajare

Vi se pare util acest document?

Este necorespunzător acest conținut?

Drepturi de autor:

Formate disponibile

Journal of Semitic Studies-Vol. 1-2009 PDF

Încărcat de

Drepturi de autor:

Formate disponibile

ISSN 0022-4480

Contributions, books for review, and other editorial communications should be

WECHSELWIRKUNG VON VOKALEN UND

Zu den Verdiensten des Buches von Rebecca Hasselbach über die

dem Assyrischen erklärt Hasselbach als Archaismen, die im späteren

Mit der Einführung von Dativsuffixen sind beide Dialekte inno-

Die mögliche Entwicklung a > e in Gegenwart von r und l ist alt-

Zur leichteren Orientierung ist das Verhalten der beiden unteren

setzt wurden, so lässt sich auch die Fernwirkung von Pharyngalen im

In Lagas arbeitete unter Lugalanda und Irikagina ein in den Ur-

wahrscheinlich, weil „banana names“ in Fara sehr selten sind.29 Aus

b) Eine alternative Erklärung für a > e im Akkadischen?

In einem kürzlich35 erschienen Artikel behandelt N.J.C. Kouwenberg

Insgesamt zeigt die Tabelle die von Kouwenberg genannte Über-

ergibt das umgekehrte Ergebnis. Insbesondere das Zeichen, welches

Wie im Akkadischen geraten die hinteren Konsonanten ebenfalls un-

Beschränken wir uns der Einfachheit halber zunächst auf das

Fassen wir unsere Betrachtung noch einmal etwas anders zusam-

d) Exkurs zu den Silbenzeichen

Das System der akkadischen Silbenzeichen beruhte fast ausschließ-

Was die Silbenzeichen betrifft, so lässt sich vermuten, dass zu-

Behrens, H. und H. Steible. 1983. Glossar zu den altsumerischen Bau und

THE DEVELOPMENT OF THE SEMITIC

This paper aims at explaining the development of the Semitic definite

Several important studies have been published lately concerning the

demonstrative is what essentially divides scholars (Rubin 2005: 72). In

pronominal demonstrative’ is a pronominal paradigm which is used to mark deixis

The second problem is the exact function of the Semitic demonstra-

equally valid in syntax.27 We should mention in this context Meillet’s

2. The Syntax of the Definite Article

The syntax of the definite article in CS is quite peculiar yet consistent

all of the data as a system. The redundancy of marking the article on

The following discussion focuses on native patterns in specific lan-

relative-determinative pronoun ∂u was used in other dialects in the

This pattern is very often termed ‘a relative pattern’ (Gesenius 1910:

3.3 Distribution: Phoenician and Old Aramaic

3.4 Common deviation: Noun-[Article-Adjective]

baytu l-muqaddasi ‘the holy temple’

Similarly in Christian Arabic, the adjective is commonly marked

na¨ara ham-m¢’orasa (San. 7:4) ‘a betrothed maiden’

4. The Function of the Article: Re-evaluation

4.1 The origin of the CS article is not a demonstrative

tion in the nominal paradigm. Even the article in Amharic, a language

4.2 The article was first positioned on the attribute

Double-marking The ‘double’ marking on the noun and adjective is

tive demonstratives, while *∂u nominalizes elements without nominal

The article was first attached to non-predicative adjectival forms, i.e.

tive adjectival forms is that nominal phrases and nominal sentences

head noun and the relative pronoun are in apposition (Goldenberg

6.1. Aramaic and OSA

The prefix ha- was reattached to the attributive demonstrative again

8. Summary and Conclusions

The common explanation for the CS article has been shown to be

Beeston, A.F.L. 1984. Sabaic Grammar. (Manchester)

Brown, M.L. 1987. ‘“Is it not?” or “Indeed!”: HL in Northwest Semitic’, Maarav

Khan, G. 1984. ‘Object Markers and Agreement Pronouns in Semitic Languages’,

Voigt, R. 1998. ‘Der Artikel im Semitischen’, JSS 43, 221 58

A PROPOSAL TO READ THE LEGEND OF

The legend of a seal amulet, UC 51354, from Deir Rifa, Egypt re

Dr S. Quirke, Curator of the Petrie Museum of Egyptian Archaeology,

the context in which this button-shaped seal was discovered, he did

Plate 1: UC 51354 with beads

Plate 2: Legend of UC 51354