ISSN 0022-4480
JOURNAL OF
Semitic Studies VOLUME LIV. NO. 2 AUTUMN 2009
SUBSCRIPTIONS A subscription to Journal of Semitic Studies comprises 2 issues. Prices include postage by surface mail, or for subscribers in the USA and Canada by airfreight, or in India, Japan, Australia and New Zealand, by Air Speeded Post. Airmail rates are available on request. Annual Subscription Rate (Volume LIII, 2 issues, 2008) Institutional Print edition and site-wide online access: £156//234/US$296 Print edition only: £185//222/US$281 Site-wide online access only: £148//222/US$281 Personal Print edition and individual online access: £51//77/US$102 Please note: £ Sterling rates apply in UK, / rates apply in Europe, US$ elsewhere There may be other subscription rates available, for a complete listing please visit www.jss.oupjournals.org/ subinfo. Full prepayment, in the correct currency, is required for all orders. Orders are regarded as firm and payments are not refundable. Subscriptions are accepted and entered on a complete volume basis. Claims cannot be considered more than FOUR months after publication or date of order, whichever is later. All subscriptions in Canada are subject to GST. Subscriptions in the EU may be subject to European VAT. If registered, please supply details to avoid unnecessary charges. For subscriptions that include online versions, a proportion of the subscription price may be subject to UK VAT. Personal rate subscriptions are only available if payment is made by personal cheque or credit card and delivery is to a private address. The current year and two previous years’ issues are available from Oxford University Press. Previous volumes can be obtained from the Periodicals Service Company, 11 Main Street, Germantown, NY 12526, USA. Email:
[email protected]. Tel: +1 (518) 537 4700. Fax: +1 (518) 537 5899. For further information, please contact: Journals Customer Service Department, Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, UK. Email:
[email protected]. Tel (and answerphone outside normal working hours): +44 (0)1865 353907. Fax: + 44 (0)1865 353485. In the US, please contact: Journals Customer Service Department, Oxford University Press, 2001 Evans Road, Cary, NC 27513, USA. Email:
[email protected]. Tel (and answerphone outside normal working hours): 800 852 7323 (toll-free in USA/Canada). Fax: 919 677 1714. In Japan, please contact: Journals Customer Services, Oxford University Press, 1-1-17-5F, Mukogaoka, Bunkyo-ku, Tokyo, 113-0023, Japan. Email:
[email protected]. Tel: (03) 3813 1461. Fax: (03) 3818 1522. Methods of payment. (i) Cheque (payable to Oxford University Press, to Oxford University Press, Cashiers Office, Great Clarendon Street, Oxford OX2 6DP, UK) in GB£ Sterling (drawn on a UK bank), US$ Dollars (drawn on a US bank), or EU/ Euros. (ii) Bank transfer to Barclays Bank Plc, Oxford Group Office, Oxford (bank sort code 20-65-18) (UK), overseas only Swift code BARC GB 22 (GB£ Sterling to account no. 70299332, IBAN GB89BARC20651870299332; US$ Dollars to account no. 66014600, IBAN GB27BARC 20651866014600; EU/ Euros to account no. 78923655, IBAN GB16BARC20651878923655). (iii) Credit card (Mastercard, Visa, Switch or American Express). Journal of Semitic Studies is published two times a year in April and October by Oxford University Press, Oxford, UK. Annual subscription price is £156//234/US$296. Journal of Semitic Studies is distributed by Mercury International, 365 Blair Road, Avenel, NJ 07001, USA. Periodicals postage paid at Rahway, NJ and at additional entry points. US Postmaster: send address changes to Journal of Semitic Studies (ISSN 0022-4480), c/o Mercury International, 365 Blair Road, Avenel, NJ 07001, USA. PERMISSIONS For information on how to request permissions to reproduce articles/information from this journal, please visit www.oxfordjournals.org/permissions. ADVERTISING Inquiries about advertising should be sent to Helen Pearson, Oxford Journals Advertising, PO Box 347, Abingdon OX14 1GJ, UK. Email:
[email protected]. Tel: +44 (0)1235 201904; Fax: +44 (0)8704 296864. DISCLAIMER Statements of fact and opinion in the articles in Journal of Semitic Studies are those of the respective authors and contributors and not of Journal of Semitic Studies or Oxford University Press. Neither Oxford University Press nor Journal of Semitic Studies make any representation, express or implied, in respect of the accuracy of the material in this journal and cannot accept any legal responsibility or liability for any errors or omissions that may be made. The reader should make his/her own evaluation as to the appropriateness or otherwise of any experimental technique described. © The University of Manchester 2008 All rights reserved; no part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise without prior written permission of the Publishers, or a licence permitting restricted copying issued in the UK by the Copyright Licensing Agency Ltd, 90 Tottenham Court Road, London W1P 9HE, or in the USA by the Copyright Clearance Center, 222 Rosewood Drive, Danvers, MA 01923. Typeset by Drukkerij Peeters, Herent, Belgium Printed by Bell & Bain Ltd, Glasgow, UK
JOURNAL OF
Semitic Studies V OLUME LIV. NO. 2 AUTUMN 2009
Contents articles CHRISTO VAN DER MERWE, Another Look at the Biblical Hebrew Focus Particle jGÌ JOSÉ MARTÍNEZ DELGADO, Lexicographical Arrangement of Masoretic Material MOSHE FLORENTIN, The Alternation u / o, Diphthongs, PataÌ furtivum and the 3m.s. Pronominal suffix in Samaritan Hebrew and Aramaic versus Tiberian Hebrew JASON BEMBRY, Plural Based Duals GEOFFREY HERMAN, Note on the Recently Published Discourse on Priesthood (BL Add. 18295, ff. 137b 140b) MARIA BULAKH, Nota Genitivi za- in Epigraphic Geez URI RUBIN, On the Arabian Origins of the Qur’an: The Case of al-furqan AHARON GEVA-KLEINBERGER and EFRAIM LEV, Language Passivity in the Medical Arabic and Judaeo-Arabic Prescriptions of the Cairo Genizah ALMOG KASHER, The Term ism in Medieval Arabic Grammatical Tradition: A Hyponym of Itself ABDULRAHMAN S. AL SALIMI, Themes of the Iba∂i/Omani Siyar AHMAD AL-JALLAD, The Polygenesis of the Neo-Arabic Dialects JAMES DICKINS, Relative Clauses in Sudanese Arabic
313 333
365 383 389 393 421 435 459 475 515 537
reviews Pier Giorgio BORBONE, Alessandro MENGOZZI and Mauro TOSCO (eds), Loquentes linguis. Studi linguistici e orientali in onore di Fabrizio A. Pennacchietti. Linguistic and Oriental Studies in Honour of Fabrizio A. Pennacchietti. Lingvistikaj kaj orientaj studoj honore al Fabrizio A. Pennacchietti (Eleanora CUSSINI)
575
Kevin J. CATHCART (ed.), The Correspondence of Edward Hincks. Volume 1 (1818 1849) (Stephanie DALLEY) Martha ROTH (ed.), The Assyrian Dictionary of the Oriental Institute of the University of Chicago Volume 18 Letter T; The Assyrian Dictionary of the Oriental Institute of the University of Chicago Volume 19 Letter ™ [Tet] (Jonathan TAYLOR) Lisa M. WRAY BEAL, The Deuteronomist’s Prophet: Narrative Control of Approval and Disapproval in the Story of Jehu (2 Kings 9 and 10) (Roger TOMES) W.J. HOUSTON, Contending for Justice: Ideologies and Theologies of Social Justice in the Old Testament (Eryl W. DAVIES) Robert P. GORDON (ed.), The God of Israel (Johannes DE MOOR) Jonathan D. LAWRENCE, Washing in Water: Trajectories of Ritual Bathing in the Hebrew Bible and Second Temple Literature (Hannah HARRINGTON) Lester L. GRABBE, A History of the Jews and Judaism in the Second Temple Period: Volume 1, Yehud: A History of the Persian Province of Judah; Volume 2: The Coming of the Greeks: The Early Hellenistic Period (335-175 BCE) (Philip R. DAVIES) Matt JACKSON-MCCABE (ed.), Jewish Christianity Reconsidered (John S. KLOPPENBORG) Michael RAND, Introduction to the Grammar of Hebrew Poetry in Byzantine Palestine (Ophir MÜNZ-MANOR) Mauro ZONTA, Hebrew Scholasticism in the Fifteenth Century: A History and Source Book (Resiannne FONTAINE) Martin ZAMMIT, ¨Enbe men Karmo Suryoyo: A Syriac Chrestomathy (Gillian GREENBERG) Sheila S. BLAIR, Islamic Calligraphy (Colin BAKER) Martin WITTINGHAM, Al-Ghazali and the Qur’an: One Book, Many Meanings (Shuruq NAGUIB) Thomas HERZOG, Geschichte und Imaginaire: Entstehung, Überlieferung und Bedeutung der Sirat Baibars in ihrem sozio-politischen Kontext (Albrecht FUESS) Armin SCHOPEN, Tinten und Tuschen des arabisch-islamischen Mittelalters: Dokumentation Analyse Rekonstruktion (Oliver KAHL) Nawal NASRALLAH (ed. and trans.), Annals of the Caliph’s Kitchens: Ibn Sayyar al-Warraq’s Tenth Century Baghdadi Cookbook. English Translation with Introduction and Glossary (David WAINES) Îusayn IBN ¨ALI DAÎIL ALLAH ABU AL-ÎASAN, Nuqus liÌyaniya min min†aqat al-¨ula: dirasa taÌliliya muqarana (María del Carmen HIDALGO-CHACÓN DÍEZ)
577
580 583 584 586 588
590 592 595 597 599 601 602 606 608
610
611
Heléne KAMMENSJÖ, Discourse Connectives in Arabic Lecturing Monologue (Gunvor MEJDELL)
614
short notes F. Rachel MAGDALENE, On the Scales of Righteousness: Neo-Babylonian Trial Law and the Book of Job (Anthony PHILLIPS) Tryggve N.D. METTINGER, The Eden Narrative: A Literary and Religio-Historical Study of Genesis 2 3 (Gary ANDERSON) Paul M. JOYCE, Ezekiel: A Commentary (Leslie C. ALLEN) Mark J. BODA, Daniel K. FALK, and Rodney A. WERLINE (eds), Seeking the Favor of God. Volume 2: The Development of Penitential Prayer in Second Temple Judaism (Gerbern S. OEGEMA) Christiana DE GROOT and Marion Ann TAYLOR (eds), Recovering Nineteenth-Century Women Interpreters of the Bible (Heather MCKAY) Roland BOER (ed.) Bakhtin and Genre Theory in Biblical Studies (Elzbieta LAZAREWICZ-WYRZYKOWSKA) Mark RONCACE and Patrick GRAY (eds), Teaching the Bible Through Popular Culture and the Arts (Martin O’KANE) Thomas AUDO, Treasure of the Syriac Language: A Dictionary of Classical Syriac (J.F. HEALEY)
619 620 621
622 623 624 624 625
ANOTHER LOOK AT THE BIBLICAL HEBREW FOCUS PARTICLE גּם1 C.H.J. VAN DER MERWE UNIVERSITY OF STELLENBOSCH
Abstract Despite the attention devoted to גּ ַםfrom various angles in recent years, there has been no exhaustive treatment in terms of all its occur rences in the Hebrew Bible. In this revision of the author’s own de tailed description of the focus particle in Genesis to 2 Kings, many of his earlier views are confirmed. However, it was also found that his earlier model needs some refinement. For example, although the most typical semantic function of גּ ַםis to signal that an entity x must be added to an entity y as far as a predication or argument z is con cerned, in a few instances (in mainly poetic texts) גּ ַםdoes have an af firmative connotation. Although even more rare, in a few instances גּ ַם also appears to be a near synonym of ְ ו. As far as a profile of its uses in terms of frequency is concerned, the most prototypical use (about 50%) is גּ ַם+ constituent.
1. Introduction The syntactic and semantic features of גּ ַםin Genesis to 2 Kings are described in Van der Merwe 1990. This study was followed by a description of its pragmatics (Van der Merwe 1993a: 181–99) as well as the description of ַרקand ( ָאְךthe ‘antonyms’ of )גּ ַםin Van der Merwe (1991: 297–311). The outcome of the above-mentioned research on גּ ַםwas summarized in Van der Merwe, Naudé and Kroeze (1999). Lyavdansky (2004: 231–50) focused on the use of גּ ַםas ‘a negative reciprocal in judgement speech’ in prophetic speech. Apart from these studies, which were limited to specific corpora within the Hebrew Bible, the only attempt towards providing an exhaustive description of ( גּ ַםin the sense of treating the entire corpus of Classical Hebrew) is the one that appeared in DCH (Clines 1995: 357–61). 1
The financial assistance of the National Research Foundation (NRF) towards this research is hereby acknowledged. Opinions expressed in this publication and conclusions arrived at are those of the author and are not necessarily to be attrib uted to the NRF. 313
ANOTHER LOOK AT THE BIBLICAL HEBREW FOCUS PARTICLE גּ ַם
The problem with DCH is that it first lists the translation values of גּ ַםas follows: ‘also, too, as well as, likewise, often repeated, both … and, (n)either … (n)or, indeed, again, (not) even, therefore and only’ and then proceeds to provide a taxonomy of all the lexemes (classified in terms of their word classes) that may follow גּ ַם. This taxonomy comprises the main part of the lexical entry, but does not provide either much insight into the syntax of the particle, or into the semantics of the different syntactic configurations it may occur in. In other words, although many of the translation values suggested for גּ ַםare correct, it is not clear at all when which type of translation value is involved (e.g. when which syntactic configuration occurs). Despite the attention גּ ַםhas received over the past two decades, there is still no exhaustive description of the particle. Making a contribution in this regards is the purpose of this study. Given recent developments in linguistics, an exhaustive description of גּ ַםwould need not only to look at all instances of גּ ַםin (at least) the Hebrew Bible, but also recent developments in linguistics (e.g. relevant insights from the field of cognitive semantics) (Van der Merwe 2006a: 87– 112; 2006b: 85–95; 2007: 238–77) need be taken into consideration. In this regard, the question concerning the more and the less prototypical uses of גּ ַםcall for special attention. Since none of the above-mentioned studies dealt systematically with the position in a clause where גּ ַםoccurs, this issue also needs to be investigated. It is therefore hypothesized that the responses to the following questions are crucial to providing a more complete description of גּ ַם. Firstly, in the light of all occurrences of גּ ַם, as well as recent developments in linguistics, must the categories used up until now to describe the word class and semantic-pragmatics dimensions of גּ ַםbe refined? Secondly, which categories of גּ ַםcan be regarded as the most typical. In other words, can the categories be better profiled in terms of their frequency of use in the Hebrew Bible? Thirdly, how must the position where גּ ַםoccurs in a clause be interpreted? This paper will not attempt to provide an exhaustive description of גּ ַם.2 Its focus is rather to answer the above-mentioned crucial questions. For these purposes, the paper will be structured as follows: I will commence with a brief summary of Van der Merwe, Naudé and Kroeze (1999: 314–17) and some general observations about the distribution of גּ ַםand its relationship with ַאף. Next, I will reconsider the status of the word class ‘focus particle’. Thirdly, I will provide a statistically-based profile of גּ ַםin the Hebrew Bible. Fourthly, the 2
For this purpose, see Van der Merwe and Naudé (forthcoming). 314
ANOTHER LOOK AT THE BIBLICAL HEBREW FOCUS PARTICLE גּ ַם
syntax of גּ ַםwill be dealt with. I will focus in particular on the question of where in a phrase or clause גּ ַםstands, and why? Fifthly, I would like to consider instances of גּ ַםthat appear to be problematic, since they cannot be accounted for in terms of my earlier model. In the light of this investigation, I will distinguish then, on the one hand, a few new semantic categories, and on the other hand, identify those instances that I still cannot explain. In conclusion, the findings of this investigation will be summarized. 2. Van der Merwe, Naudé and Kroeze (1999) In this work a distinction is made between the syntax (§2.1) and semantics and pragmatics of §( גּ ַם2.2). In terms of these two distinctions, sub-classes are then identified. 2.1 The Syntax of גּ ַם A syntactic feature of גּ ַםis that it can modify a word, a constituent or a clause. ( גּ ַםas opposed to )ַאףalmost always directly precedes the constituent or clause to which it refers (its domain). The following syntactic configurations are found: 2.1.1 גּ ַם+ Entity a. גּ ַם+ word in a word chain And this was good in the sight of all the people and also in the sight of Saul’s servants (1 Sam. 18:5).
1
b. גּ ַם+ constituent 2
But Er, Judah’s first-born, was wicked in the sight of the Lord; and the Lord slew him. And what he (Onan) did was displeasing in the sight of the Lord, and he slew him also (Gen. 38:7, 38:10).
c. גּ ַם+ clause 3 ְ … הִוא
315
I thought, there is certainly [lit. I just/only thought there is] no fear of God at all in this place, and they will kill me because of my wife. Moreover, she is indeed my sister (Gen. 20:11 12).
ANOTHER LOOK AT THE BIBLICAL HEBREW FOCUS PARTICLE גּ ַם
2.1.2 גּ ַם+ Entity // גּ ַם+ Entity When גּ ַםhas this syntactic configuration, it is also regarded as a double conjunction. a. גּ ַם+ word // גּ ַם+ word 4
And there went up with him both chariots as well as horsemen (Gen. 50:9).
b. גּ ַם+ clause // גּ ַם+ clause 5
And Abimelech said: I do not know who has done this thing. Neither did you tell me… Nor have I heard of it until today (Gen. 21:26).
2.2 The Semantics and Pragmatics of גּ ַם 2.2.1 Speakers or writers give an explicit indication to their audience that a specific something or someone must be added to something or someone referred to in the preceding context: also, even, moreover, even more so. a. If the reference to this entity occurs in a previous expression, an audience would find it strange if the speaker or writer did not indicate that the entity after גּ ַםhas to be added to the entity in the preceding expression. In other words, the use of גּ ַםcontributes toward constituting a well-formed text or discourse. See no. 2. b. If the entity that has to be added is something that is possible, but which one would not expect, i.e. an extreme case, גּ ַםis translated even. 6
Even before the fat was burned, the priest’s servant would come and say to the man who was sacrificing:… (1 Sam. 2:15).
c. Speakers wish to indicate clearly that an expression has to be supplemented with the expression (sentence) after גּ ַםto support a specific argument. In this way, the audience is left in no doubt about the connection between the two expressions. The expression after גּ ַםusually refers to a more persuasive or compelling argument than the preceding one. See no. 3. d. The specific inclusion of the entity or entities after גּ ַםusually reflects some special role that the inclusion has played. See no. 1. 316
ANOTHER LOOK AT THE BIBLICAL HEBREW FOCUS PARTICLE גּ ַם
2.2.2 Speakers make it clear that the inclusion of both entities preceded by גּ ַםis of special importance in a particular context: both, as well as. See no. 4. Sometimes the entities are two arguments that both carry equal weight in support of a statement. See no. 5. 2.2.3 Speakers may ask their audience for a corresponding reaction to something someone has already done. 7
So therefore, swear to me by the Lord that as I have dealt kindly with you, you also will deal kindly with my father’s house (Josh. 2:12).
3 Revisions 3.1 Introduction As far as the distribution of גּ ַםis concerned, its 769 occurrences are well distributed throughout the entire Hebrew Bible.3 This is in contrast to its near-synonym ַאף, which has a much less even distribution, viz. ַאףoccurs mainly in poetic texts and is concentrated in a particular locus, e.g. 30 out of 133 instances occur in Isaiah 26–48. The claim in Van der Merwe, Naudé and Kroeze (1999: 314) that גּ ַםnearly always immediately precedes the reference to the entity it governs, whether it is a member of a coordinated phrase (i.e. member of a word chain), a constituent (syntagm), or a sentence, is borne out by the data. In only 1 out of 769 instances,4 i.e. Job 2:10, is גּ ַם unambiguously not immediately preceding the entity it governs.5 3 No prototypical use of גּ ַםcan be linked with a particular diachronic stage or genre of BH. This is only possible for some of the more non typical uses, e.g. cases where גּ ַםgoverns more than one sentence tend to occur predominantly in poetic material or later texts (see §3.3.1e). The same applies to where גּ ַםis used as a near synonym of ְ ו. See §3.3.2.1 and §3.4. 4 These, however, exclude a number of other cases where the intepretation of גּ ַם is problematic, e.g. Isa. 40:24; Ezek. 10:16; Mal. 1:10; Ps. 133:1; Eccl. 7:22; 12:5; 2 Chron. 21:11. Furthermore, in the case of Gen. 6:3, Ezek. 21:32 and Ps. 84:7 textual critical considerations may come into play and in Prov. 17:26; 18:9; 19:2 and 20:11 a specialized use in a proverb collection may be involved. In Gen. ַ ַ גּ. Joel 44:10; 1 Sam. 12:16; 1 Kgs 14:14 the same construction is used, viz. עָתּה-ם 4:4 makes sense only if the scope of ג ַםis a cluster of sentences. 5 In the case of 2 Chron. 21:11 and Eccl. 2:14 it is also possible to argue that גּ ַם does not immediately precede its syntactic domain.
317
ANOTHER LOOK AT THE BIBLICAL HEBREW FOCUS PARTICLE גּ ַם
Compare the difference between the typical syntax of nos 8–9 with that of Job 2:10 in no. 10. 8
And his servants and also the woman urged him (1 Sam. 28:23). You must say, ‘Also your servant Uriah, the Hittite died’ (2 Sam. 11:21). But he replied, ‘You’re talking like one of the godless women would do! Should we receive what is good from God, and not also receive what is evil?’ (Job 2:10).
9 10
3.2 Word Class The category ‘focus particle’ is not regarded in typical traditional grammars as one of the basic word classes. In Van der Merwe, Naudé and Kroeze (1999: 311) גּ ַםis classified as a focus particle.6 The word class ‘focus particle’ is then treated under the broad heading ‘adverb’. In generative circles lexical items such as ַאף, גּ ַם, ַרקand ָאְךwould be regarded as quantifiers when they govern a member of a coordinated phrase or a constituent, and sentence adverbials when they govern a clause.7 In many recent grammars and respected linguistic publications, however, their equivalents in English and German, are often classified as ‘focus particles’ (König 1991; Bayer 1996; Dimroth 2004). Despite its few members, justifiable arguments are provided to distinguish this ‘basic’ word class. I therefore regard it as reasonable to maintain the category ‘focus particle’. However, the focus particle גּ ַםcan, like the members of many other word classes, also be used as a different word class, e.g. when it governs a sentence, it functions as a conjunctive adverb, or when it governs more than one sentence, it functions as a macro-syntactic connective.8 Those cases where גּ ַםgoverns two or more members of a coordinated phrase were described in Van der Merwe, Naudé and Kroeze 6 For a substantiation of distinguishing this word class, cf. Van der Merwe 1993b: 27 44. 7 Naudé (personal communication). 8 The empirical status of this category needs some further investigation. At this stage it represents an attempt to account for the use of a connection between units bigger than one sentence. Traditionally a word class is defined in terms of the mor phology and distribution of a lexeme within clauses and sentences. For the notion ‘connective’, cf. Caron 1997: 53 73.
318
ANOTHER LOOK AT THE BIBLICAL HEBREW FOCUS PARTICLE גּ ַם
(1999: 314) as double conjunctions (see §2.1.2). A term more widely used in recent times, at least among major grammars of English, is ‘correlative conjunction’ (Huddleston and Pullum 2002). In short, then, in contrast to regarding all instances of גּ ַםeither as focus particles or double conjunctions, I now postulate a basic category ‘focus particle’ (see §3.3.1a), which can also be used as a conjunctive adverb9 (see §3.3.1c), a correlative conjunction (see §3.3.1b), a macro-syntactic connective (see §3.3.1e) or even in a few rare instances as a simple coordination conjunction (see §3.3.2.1). 3.3 Syntax 3.3.1 Frequency of the Syntactic Configurations of ( גּ ַםPrototypes) a. גּ ַם+ constituent, where גּ ַםis a focus particle, is by far the most frequently occurring syntactic pattern. Nearly 55% of the instances of גּ ַםin the Hebrew Bible belong to this syntactic category.10 Typical examples of this category are nos 2 and 9.11 b. גּ ַם+ entity / גּ ַם+ entity (/ גּ ַם+ entity). These cases, where גּ ַםis used syntactically as a correlative conjunction, account for 21% of the occurrences of the lexeme in the Hebrew Bible. A distinction is made whether the ‘entity’ גּ ַםgoverns is a word or phrase or a sentence. (i) גּ ַם+ word or phrase // גּ ַם+ word or phrase12 (// גּ ַם+ word or phrase)13 (where גּ ַםis used syntactically as a correlative conjunctive) represent 13% of instances. Typical examples of this category are nos 5 and 11. Rule over us, both you, your son and the son of your son > your grandson (Judg. 8:22).
11
9
For the word class conjunctive adverb, cf. Bussmann (1996: 95). For an exhaustive list of all the instances of each category, cf. Van der Merwe and Naudé (forthcoming). 11 More typical examples are Gen. 3:6,22; Exod. 1:10; Lev. 25:45; Deut. 1:28; 1 Sam. 8:8,20; Isa. 1:15; Jer. 2:33. In some instances, e.g. Gen. 7:3; Zech. 9:2; Prov. 16:4, the predicate is ellipsed. Also included in this category are cases where the constituent following גּ ַםis refering to an agent(s) A of an action X, that is a corresponding reaction to the action(s) Y of an agent(s) B. Typical instances of the type גּ ַם+ NP + VP (The NP is typically an independent personal pronoun) are Gen. 44:9; Lam. 1:8; Ezek. 8:18; 9:10. Sometimes the גּ ַם+ NP follows the verb, e.g. Gen. 50:18; Hos. 4:6. 12 More typical examples are Exod. 5:14; Num. 18:3; Judg. 5:4; 19:19; 1 Sam. ֶ ֵ ׁנ-ם ְ ַ גּshould most probably be 2:26; 1 Kgs 3:13, 26; Jer. 14:18. The pattern יכם regarded on semantic grounds as belonging to this category. Cf. also Deut. 22:22; 23:19; 1 Sam. 25:43; Ruth 1:5. 13 Cf. also Judg. 8:22; Eccl. 9:6; 1 Chron. 11:2. 10
319
ANOTHER LOOK AT THE BIBLICAL HEBREW FOCUS PARTICLE גּ ַם
(ii) גּ ַם+ sentence // גּ ַם+ sentence14 (// גּ ַם+ sentence)15 (just more than 8% of instances). Typical examples of this category are nos 5 and 12. 12
Neither did you hear, nor did you know; nor was your ear open beforehand, in fact, I know you deal very treacherously, and you were called a rebel from your birth (Isa. 48:8). c. גּ ַם+ sentence. This category, where גּ ַםfunctions syntactically as a
conjunctive adverb, accounts for 12% of the occurrences.16 Typical examples are nos 3 and 13. 13
He said: ‘I have not found her, what’s more, the men of the place said: “There has been no prostitute here.”’ (Gen. 38:22).
d. גּ ַם+ member of a coordinated phrase17 (about 3% of instances).18 Typical examples are no. 1 and no. 8. e. גּ ַם+ sentences19 where גּ ַםfunctions as a macro-syntactic connective (about 2% of instances).20 14
Furthermore, what are you to me, O Tyre and Sidon, and all the regions of Philistia? Are you 14 Cf. also Gen. 14:16; Isa. 66:3 4; Jer. 6:15; Ezek. 24:5; Mal. 3:15; Ruth 1:12. 15 Cf. also Josh. 7:11; 1 Kgs 1:46 8; Isa. 48:8; Eccl. 9:11. 16 Cf. also Gen. 20:12; 2 Sam. 12:27; Zech. 3:7; Ps. 37:25; Song 8:1; 1 Chron. 10:13. 17 In Van der Merwe, Naudé and Kroeze (1999: 314), this category was labelled as ‘word in a word chain’ (see §2.1.1.a). The term ‘word chain’ is here replaced by the linguistically more adequate term ‘coordinated phrase’. For example, it better reflects the fact that both words and phrase are coordinated. The term ‘word chain’ implies only the coordination of words. 18 Cf. also 2 Sam. 2:2; 2 Kgs 24:4; Ps. 8:8; Eccl. 11:2; 1 Chron. 29:24; 2 Chron. 21:17; 24:12. 19 Category 3.3.1e is not distinguished in Van der Merwe, Naudé and Kroezé (1999). For the necessity to distinguish these categories and other new categories, see §3.4 below. 20 Isa. 43:13; 44:12; Jer. 5:28; 6:11; Ezek. 18:11; Hos. 7:9; Joel 4:4; Pss 19:12; 25:3; 85:13; Job 18:5; Eccl. 4:11; 6:5; 9:3; 2 Chron. 29:7. Most of these uses of גּ ַםare in poetic texts. The use in a prose text, viz. 2 Chron. 29:7 might be regarded as Late Biblical Hebrew.
320
ANOTHER LOOK AT THE BIBLICAL HEBREW FOCUS PARTICLE גּ ַם
paying me back for something? If you are paying me back, I will turn your deeds back upon your own heads swiftly and speedily (Joel 4:4).
The above-mentioned syntactic uses of גּ ַםaccount only for 93% of its occurrences in the Hebrew Bible. About 3% of instances cannot be accounted for in terms of my model. The remaining 4% are new non-prototypical uses. The reasons for distinguishing these categories will be evident from the discussion in the subsequent sections. 3.3.2 Syntax of the Various Categories In this section I would like to concentrate on the following questions: firstly, regarding the behaviour of גּ ַםin coordinated phrases, do גּ ַםand ו ְג ַםe.g. display the same distributional patterns in multi-member coordinated phrases, and do the differences in distribution correlate with semantic differences? Secondly, I would like to consider the syntagmatic distribution of גּ ַם+ constituent in clauses, e.g. does the unit גּ ַם+ constituent tend to occupy specific positions in a clause, and, if so, why? Thirdly, I want to look at the syntax of sentences that are governed by גּ ַם. E.g. while the typical pattern is גּ ַם+ verb or predicate, the question arises: how must גּ ַם+ x + verb cases be interpreted? 3.3.2.1 גּ ַםin Coordinated Phrases גּ ַםtypically occurs before the second member of two coordinated phrases (nos 1, 8 and 15). גּ ַםmay also precede the last member of a coordinated phrase with more than two members (no. 16). 15
They did not know Yahweh and also the work he had done for Israel (Judg. 2:10). So Saul and his three sons and his armour-bearer and all his men died together on the same day (1 Sam. 31:6).21
16
In a few instances גּ ַםis followed by more than one member of a coordinated phrase (no. 17). 17
(NRSV) The vine withers, the fig tree droops. Pomegranate, palm, and apple all the trees 21
1 Chron. 29:24. 321
ANOTHER LOOK AT THE BIBLICAL HEBREW FOCUS PARTICLE גּ ַם
of the field are dried up (Joel 1:12).22 When one compares the use of גּ ַםand ו ְג ַם, it appears that the latter
use dominates those cases where it can be argued that the ellipsis of a predicate is involved. These include instances where ellipsis is possibly involved, e.g. nos 1, 8 and 15,23 as well as instances where ellipsis is most probably involved, e.g. no. 18.24 (NRSV) They came up against Judah, invaded it, and carried away all the possessions they found that belonged to the king’s house, along with his sons and his wives, so that no son was left to him except Jehoahaz, his youngest son (2 Chron. 21:17).
18
In cases like no. 17 and nos 19–21 where ellipsis may be ruled out, גּ ַםtends to be more frequent. (NRSV) Hungry and thirsty, their soul fainted within them (Ps. 107:5). (NRSV) Drought and heat snatch away the snow waters; so does Sheol those who have sinned. (Job 24:19). (NRSV) The mandrakes give forth fragrance, and over our doors are all choice fruits, new as well as old, which I have laid up for you, O my beloved (Song 7:14).
19 20
21
I use the phrase ‘tends to’ above since a comparison of no. 16 and no. 22 suggests that גּ ַםis also used where ellipsis most probably did take place. 22
(NRSV) Thus Saul died; he and his three sons and all his house died together (1 Chron. 10:6).
In no. 17 and nos 19–21 it appears as if גּ ַםhas no role of pointing out the inclusion of an entity. In the only other case of this type, 22
2 Sam. 2:2; 2 Chron. 21:17. Also Gen. 6:4; 14:7; Judg. 2:10; 2 Sam. 2:2; Isa. 21:12; Eccl. 11:2. 24 Also Ezek. 21:14; 2 Chron. 24:12. 23
322
ANOTHER LOOK AT THE BIBLICAL HEBREW FOCUS PARTICLE גּ ַם
no. 21, however, it might be argued that גּ ַםis doing a little bit more than ‘and’. To complicate matters even further, in a case like no. 23, where ו ְג ַם is used, it is not immediately evident whether ו ְג ַםis doing more than ‘and’. 23
(NRSV) Everyone comes to shame through a people that cannot profit them, that brings neither help nor profit, but shame and disgrace (Isa. 30:5).
Before drawing some conclusions, consider the instances (nos 24– 7) where verb forms are connected by גּ ַם. (NRSV) Your nakedness shall be uncovered, and/yes your shame shall be seen (Isa. 47:3). ‘Away! Unclean!’ people shouted at them; ‘Away! Away! Do not touch!’ The fact is, they fled and/what’s worse they wander about; it was said among the nations, ‘They shall stay here no longer.’ (Lam. 4:15). By the rivers of Babylon there we sat down and wept, when we remembered Zion (Ps. 137:1). (NRSV) Who has heard of such a thing? Who has seen such things? Shall a land be born in one day? Shall a nation be delivered in one moment? Yet as soon as Zion was in labour she delivered her children (lit. Zion was in labour [and] also delivered her children) (Isa. 66:8).
24 25
26 27
While it is possible to interpret גּ ַםin both nos 24–6 as in nos 17, 19–20 as a near-synonym of ְ ו, such an interpretation is less likely in no. 27. In this case the ‘addition’ has a temporal connotation: ‘the moment Zion was in labour, she “also” delivered’. It also possible to argue that גּ ַםis not merely a near-synonym of ְ וin no. 24 and no. 25, since in no. 24 it may be interpreted as denoting an affirmation (‘yes’) and in no. 25 as a noteworthy addition (‘what’s worse’). This, however, not to be the case in no. 26 – unless this mourning ritual entails actions we do not fully understand. 323
ANOTHER LOOK AT THE BIBLICAL HEBREW FOCUS PARTICLE גּ ַם
In contrast, cases where ו ְג ַםis used, ‘noteworthy addition’ is as a rule denoted. See no. 28. Then the LORD said, ‘You are to entice him, and what’s more, you shall succeed; go out and do it’ (1 Kgs 22:22).25
28
Based on this rather meagre evidence, the only deductions that can be made are the following. 1. ו ְג ַםhas a more restricted distribution than גּ ַםin coordinated phrases. It tends to precede the last member of the phrase. In many of these instances, ellipsis of a predicate probably took place. 2. גּ ַםis used more often than ו ְג ַםin instances where ellipsis can be ruled out. 3. When גּ ַםis used in a coordinated phrase it is sometimes a nearsynonyms of ְ ו. The use of גּ ַםon its own (instead of )ו ְג ַם, however, is no sign that a near-synonym of ְ וis involved. 3.3.2.2 גּ ַם+ constituent In more than two-thirds of the occurrences of this category, גּ ַם+ constituent is fronted. This happens in the majority of the cases, but not exclusively, in reported speech. A typical example is no. 9. You must say, ‘Also your servant Uriah, the Hittite died’ (2 Sam. 11:21).
9
The reason for the fronting is in a sense obvious, viz. גּ ַם+ constituent represents the focus of the utterance, or, to formulate the notion ‘focus’ more precisely, גּ ַם+ constituent represents that part of the utterance that turns a discourse active proposition into a piece of information. In the case of no. 9 it means, the proposition ‘x died’ and x=soldiers was discourse active. In the utterance ‘Also your servant, Uriah the Hittite died’, Also your servant, Uriah the Hittite turns ‘x died’ into a piece of information. At this stage I would then like to hypothesize that the sentence initial position is the preferred position of the category גּ ַם+ constituent, since that is the clause position which is as a rule associated with constituents that turn a discourse active proposition into a piece of information. Although a two-thirds majority of instances where גּ ַם+ constituent is fronted may be considered as significant, one still has to have 25
See also 1 Kgs 21:19; Ps. 84:3; 2 Chron. 18:21. 324
ANOTHER LOOK AT THE BIBLICAL HEBREW FOCUS PARTICLE גּ ַם
an explanation for those instances where גּ ַם+ constituent is not fronted, e.g. nos 29–32. (NRSV) and he put him to death also (Gen. 38:10). Until Yahweh gives rest to your brothers as for you and they too have taken possession of the land (Josh. 1:15). (NRSV) Let us make him drink wine tonight also (Gen. 19:34). (NRSV) He said to him, ‘Very well, I grant you this favour too, and will not overthrow the city of which you have spoken’ (Gen. 19:21).
29 30
31 32
Although this type of ‘naked’ statistic may be misleading, it is nevertheless noteworthy that in 50 instances, clauses like no. 29 (introduced by a wayyiqtol) 26 occur, in 15 instances clauses like no. 30 (introduced by weqatal), 27 in 14 instances clauses like no. 31 (introduced by a directive)28 are used. In contrast, only ten instances with qatal 29 clauses (predominantly in poetic texts), three with weyiqtol,30 four with yiqtol,31, and one each with an infinitive 32 and a participle33 respectively, have been identified. These statistics give me the impression that the sentence-initial position is the default position of גּ ַם+ constituent, unless that position is needed for another expression, e.g. a wayyiqtol, weqatal or directive. This point is well illustrated by no. 33. Next Abner not only spoke privately to the Benjaminites, but
33 26
Gen. 3:6; 19:35; 20:6; 22:24; 26:21; 27:31; 29:30,33; 33:7; 38:10; Exod. 7:11 (2x); 8:28; Num. 11:4; Deut. 3:3; 9:19,20; 10:10; 12:30; Josh. 10:30; Judg. 3:22,31; 6:35; 8:9; 9:49 (2x); 1 Sam. 14:22; 19:20,21(2x),22,23,24(2x),25:13; 31:5; 2 Sam. 3:19; 7:19; 11:17; 1 Kgs 3:18; 14:23; 2 Kgs 21:11; Jer. 3:8; Job 1:6; 2:1; Est. 7:2; 9:15; 1 Chron. 10:5; 19:15; 24:31. 27 Gen. 3:22; 30:15; 32:21; 44:29; Exod. 1:10; 12:32; Num. 16:10; 27:13; Deut. 3:20; Josh. 1:15; Judg. 7:18; 1 Sam. 8:20; Jer. 48:26; Hos. 4:5; Zech. 9:7. 28 Gen. 19:34; 27:34,38; Num. 4:22; 18:28; 22:19; 2 Sam. 11:12; 17:5; 18:2,22; Obad. 1:13; Hab. 2:16; Zech. 8:21; Est. 9:13. 29 Gen. 19:21; Exod. 7:23; 2 Sam. 15:13; Isa. 7:13; Jer. 25:14; Hos. 5:5; Job 13:2; 33:6; Prov. 18:3; Eccl. 2:8. 30 Gen. 29:27; 30:3; Deut. 12:30; 2 Sam. 14:7. 31 Gen. 38:11; 2 Sam. 15:19; 17:12,13. 32 Judg. 10:9. 33 1 Sam. 8:8. 325
ANOTHER LOOK AT THE BIBLICAL HEBREW FOCUS PARTICLE גּ ַם
he [lit. Abner] also went to speak privately to David.… (2 Sam. 3:19).
However, for the remaining 19 instances, of which one half occurs in poetry, the only common denominator I could identify so far is that in most cases the constituent which גּ ַםgoverns contains a deictic expression, cf. no. 32. 3.3.2.3 גּ ַם+ sentence In about one third of the cases belonging to this category, גּ ַםimmediately precedes the verb (no. 34). To identify these instances is thus not difficult. The same applies to cases where גּ ַםprecedes the predicate of a nominal clause (no. 35). 34
I will bless her, what’s more, I will give you a son by her (Gen. 17:16). … for you are a foreigner, what is worse, you are an exile from your place > home (2 Sam. 15:19).
35
When a finite verb is modified by an infinitive absolute with the same root, גּ ַםimmediately precedes the infinitive absolute. גּ ַם+ inf. abs. may occupy the sentence-initial position (no. 36), but this is not necessarily the case.34 36
She said: ‘An Egyptian man rescued us from the shepherds; what’s more, he diligently drew water for us and watered the sheep and goats!’ (Exod. 2:19).
However, in a number of the instances belonging to this category גּ ַם is not followed by the verb of a verbal clause, but a non-verbal constituent (cf. nos 13, 37 and 38). On account of surface-level criteria, they are therefore not to be distinguished from the most typical category: גּ ַם+ constituent, discussed above (§3.3.2.2). 37
(NRSV) And Samuel said to him, ‘The LORD has torn the kingdom of Israel from you this very day, and has given it to a neighbour of yours, who is better than you.29 Moreover the 34
Note the use of the wayyiqtol in Gen. 31:15. 326
ANOTHER LOOK AT THE BIBLICAL HEBREW FOCUS PARTICLE גּ ַם
Glory of Israel will not recant or change his mind; for he is not a mortal, that he should change his mind (1 Sam.15:29). Therefore, be courageous and be valiant (lit. let your hands be strong and be sons of valiance); for Saul your lord is dead, (and) furthermore, the house of Judah has anointed me to be king over them’ (2 Sam. 2:6 7).
38
The question is, then, why are the constituents in those clauses fronted? When one analyses the information structure of each, it becomes evident that in most cases a shift in topic (no. 13),35 a shift in portrayal of the discourse active topic (no. 37),36 or focus of a constituent (no. 38)37 is involved. In other words, when גּ ַםgoverns an entire clause, sentence or sentences, the information structure of these sentences is marked in the same way as sentences without גּ ַם.38 3.4. Problematic Cases In this section, a distinction will be made between instances that at first appeared to be problematic, but for which there are now solutions, and those where the semantic contribution of גּ ַםstill cannot be established. In no. 14 גּ ַםonly makes sense if it governs a subsequent paragraph. It turned out that there are a number of other instances, predominantly in poetic texts, where גּ ַםgoverns, as in no. 14 and no. 39, more than one sentence (often a poetic line).39 39
(NRSV) Faithfulness will spring up from the ground, and right35 Also in Gen: 40:15; 42:22; Exod. 5:2; 1 Sam. 28:20; 1 Kgs 16:7; Ps.71:24; Eccl. 6:7; 7:21; Jer. 2:34; Neh. 5:16; 6:19. In Neh. 4:16; 5:14, 6:17 and 13:23 a shift in temporal frame is involved (and the temporal reference fronted). Interpret ing Neh. 5:13 is problematic; neither a shift in topic, temporal frame or focus of a constituent can be established with certainty. 36 Also Judg. 20:48. 37 Also Deut. 7:20; 2 Kgs 21:16; Isa. 57:7; Jer. 2:16; Ezek. 20:23, 25; Eccl. 5:16; 2 Chron. 21:13; 22:5. 38 For the ways in which the information structure of a text is overtly marked by means of the order of constituents, cf. Van der Merwe 2002/3: 70 86, Floor 2004 and Moshavi (forthcoming). 39 See also Isa. 43:13; 44:12; Jer. 5:28; 6:11; Ezek. 18:11; Hos. 7:9; Pss. 19:12; 25:3; Job 18:5; Eccl. 4:11; 6:5; 9:3; 2 Chron. 29:7.
327
ANOTHER LOOK AT THE BIBLICAL HEBREW FOCUS PARTICLE גּ ַם
eousness will look down from the sky. (Indeed, CvdM) The LORD will give what is good, and our land will yield its increase (Ps. 85:12 13).
No. 39 is the only example where גּ ַםgoverns more than one sentence, and it does not have an additive function, but an affirmative one. A number of instances, such as nos 40 and 41, however, justified distinguishing in addition to those in §3.3.1, a semantic category, viz. ‘affirmation’.40 40
(NEB) They surrounded me, yes, they surrounded me (Ps. 118:11). Judah gathered to seek [help] from Yahweh, indeed, from all the towns of Judah they came to seek Yahweh (2 Chron. 20:4).
41
Gen. 16.13 (no. 42) is the only instance where גּ ַםis used in this way. In other words, גּ ַםis used in a yes/no question in which the epistemic modality of an event is asserted by means of a rhetorical question. If one compares no. 42 with no. 43, it is not unreasonable to argue that גּ ַםfunctions in Gen. 16:13 like its near-synonym ַאףin Gen. 18:23, and 6 other similar instances in the Hebrew Bible. In other words, Gen. 16:13 is another instance where גּ ַםhas an affirmative connotation. (NRSV) So she named the LORD who spoke to her, ‘You are Elroi’; for she said, ‘Have I really seen God and remained alive after seeing him?’ (CvdM: Have I really seen here the one looking after me?) (Gen. 16:13).41 (NRSV) Will you really sweep away the righteous with the wicked? (Gen. 18:23).42
42
43
40 Also Deut. 2:15; 1 Sam. 24:12; Isa. 13:3; 47:3; Hos. 9:12; Mal. 2:2; Pss. 78:21; 118:11; 119:24; Job 30:8. 41 See Booij (1980: 1 7) for the text critical problems of this verse. 42 Also Gen. 18:13; 18:24; Amos 2:11; Job 34:17; 40:8. Gen. 3:1 lacks the question word.
328
ANOTHER LOOK AT THE BIBLICAL HEBREW FOCUS PARTICLE גּ ַם
In no. 44 and three other instances43 גּ ַםgoverns a clause, but the semantic relationship between the clauses it connects is contrastive in the sense of ‘a denial of an expectation’. Significant is that in all four cases גּ ַםis used (and not )ו ְג ַם. A possible explanation for this use of גּ ַם may be found in its rare use as near-synonym of ְ ו. As semantically empty coordinating conjunction, it can also connect clauses of which the content is contrastive. 44
Since my youth they have often attacked me, but they have not defeated me (Ps. 129:2).
While no. 45 and perhaps no. 46 too can be explained by analogy with ו ְַעָתּה, this is not the case in no. 47. (NRSV) Now therefore take your stand and see this great thing that the LORD will do before your eyes (1 Sam. 12:16). (NRSV) He said, ‘Even so; in accordance with your words, let it be: he with whom it is found shall become my slave, but the rest of you shall go free’ (Gen. 44:10). (NRSV) The LORD will raise up for himself a king over Israel, who shall cut off the house of Jeroboam today, even right now! (1 Kgs 14:14).
45
46
47
More problematic are instances like no. 48.44 (NRSV) How very good and pleasant it is when kindred live together in unity! (Ps. 133:1).
48
3.5 Summary The findings of this investigation can be summarized as follows: 1. In all but one instance in the Hebrew Bible גּ ַםprecedes the entity it governs. The referring expression may be a word, a phrase, a sentence or sentences. 43
Eccl. 4:16; Neh. 6:1; 2 Chron. 30:12. Also impossible to explain in terms of our current model are the following instances: Isa 40:24; Ezek. 10:16; Mal. 1:10; Prov. 17.26, 18:9, 19:2, 20.11; Eccl. 7:22; 12:5; 2 Chron. 21:11. 44
329
ANOTHER LOOK AT THE BIBLICAL HEBREW FOCUS PARTICLE גּ ַם
2. גּ ַםis primarily a focus particle that can be used from a syntactic perspective as a correlative conjunction, conjunctive adverb, macro-syntactic connective or a near-synonym of ְ ו. 3. The most typical semantic function of גּ ַםis to signal that an entity x must be added to an entity y as far as a predication or argument z is concerned. This happens typically in three syntactic environments. By far the most frequent (i.e. about 50% of its occurrences in the Hebrew Bible) is where גּ ַםgoverns a constituent.45 Since גּ ַם+ constituent represents the focus of an utterance, it tends to be fronted. However, it may also be ‘pushed’ out of the sentence-initial position by other constructions (typically wayyiqtol or weqatal ). The English translation value of this prototypical use of גּ ַםis ‘also’ or ‘too’. See no. 9. If the referent to be added represents an extreme or unexpected case, a translation value of ‘even’ would be preferable. See no. 6. When גּ ַםeither (a) is added to only one member of a coordinated phrase or (b) governs a sentence, or sentences, the use of גּ ַםsignals that the addition of an entity has one or other type of relevance in that context. While instances of (a) occur relatively seldom (i.e. about 3% of the occurrences in the corpus), instances of (b) are more frequent (governing a sentence: 12%; and more than one sentence: 2%). In (a) the translation value of גּ ַםis ‘also’ or ‘as well as’. See nos 1 and 8. In the case of (b), גּ ַםsignals information to be considered in addition to that which was given before (to be translated as ‘furthermore, besides, also’. See no. 38.). Sometimes this information represents the most conclusive information as far as an argument is concerned (to be translated as ‘moreover, what’s more, what’s worse’. See nos 3, 13 and 34–7). A feature of (b), i.e. גּ ַם+ sentence(s), is that the constituent immediately following גּ ַםis often the verb of a verbal clause or the predicate of a nominal clause. However, when any other constituent immediately follows גּ ַם, a shift in topic or the focus of that constituent is involved. See nos 13, 37 and 38. 4. Where גּ ַםis repeated before all the members of a coordinated phrase or consecutive sentences, a speaker or author signals the relevance of the inclusion of multiple entities as far as a predication or argument z is concerned. This relatively frequent use of גּ ַם 45
This figure, in contrast to the one (55%) referred to in §3.3.1, excludes in stances where the role of a referent of the constituent governed by גּ ַםin a corre sponding action is pointed out. See point 5 below. 330
ANOTHER LOOK AT THE BIBLICAL HEBREW FOCUS PARTICLE גּ ַם
(about 21% of instances in the corpus) is to be translated in English as ‘both x and y’, in the case of the referents of multiple phrases (see nos 4 and 11), and ‘(n)either… (n)or’ in the case of the propositional content of sentences (see nos 5 and 12). 5. A semantic category that has been described in Van der Merwe, Naudé and Kroeze (1999: 317), but only mentioned in this study (§2.2.3), is one that can be described as one pointing out the role of people or God in a corresponding reaction. It cannot be said that ‘also x’ does y, but what x does is a corresponding reaction to what z did. This use of גּ ַםrepresents 6% of the occurrences of the lexeme in the corpus, and can often be translated as ‘x in turn’ or ‘x too’. See nos 49–50.46 49
(NRSV) Now then, since I have dealt kindly with you, swear to me by the LORD that you in turn will deal kindly with my family. (Josh. 2:12).47 (NJPS) If you do this and God so commands you you will be able to bear up; and all these people too will go home unwearied (Exod 18:23).48
50
6. In a few instances ( גּ ַם1.6% of the corpus) does not have an additive, but an affirmative connotation. See nos 40–2. 7. Although rare, a few instances of גּ ַםappear to be used (in mainly poetic texts) as a near-synonym of ְ ו. See nos 17, 19–20. REFERENCES Bayer, J. 1996. Directionality and Logical Form: On the Scope of Focusing Particles and Wh in situ. (Dordrecht) Booij, T. 1980. ‘Hagar's words in Genesis xvi.13b’, VT 30, 1 7 Bussmann, H. 1998. Routledge Dictionary of Language and Linguistics. (London)
46
A number of these instances Lyavdansky 2004: 231 50 calls ‘a negative re ciprocal in judgement speech’. The cases Lyavdansky discusses all fit well into this category, e.g. Num. 22:33; Judg. 2:21; Jer. 4:12; 13:25 6; 31:37; Ezek. 5:11(2x); Ezek. 8:18; 9:10; 16:43; 16:52(2x); 20:15,25; 23:35; 24:9; Amos 4:6,7; Mic. 6:13; Mal. 2:9; Ps. 52:7. 47 Gen. 50:18; Exod. 7:11; Lev. 26:24; Josh. 2:12; 9:4; Judg. 1:3; 1 Sam. 28:22; Prov. 23:15; 26:4. 48 Gen. 44:9; 1 Sam. 1:28; 2 Sam. 2:6; 2 Kgs 22:19; Lam. 1:8; Zech. 3:7; 11:8; Ps. 71:22; Job 7:11; 40:14; Prov. 1:26; 11:25; 2 Chron. 34:27. 331
ANOTHER LOOK AT THE BIBLICAL HEBREW FOCUS PARTICLE גּ ַם
Caron, J. 1997. ‘Towards a Procedural Approach of the Meaning of Connectives’, in J. Costerrmans and M. Fayol (eds), Processing Interclausal Relationships: Stud ies in the Production of comprehension of text (New Jersey). 53 73 Clines, D.J.A. (ed.) 1995. The Dictionary of Classical Hebrew (DCH), II: ו ב. (Sheffield) Dimroth, C. 2004. Fokuspartikeln und Informationsgliederung im Deutschen. (Tübingen) Floor, S. 2004. ‘From Topic, Focus and Information Structure to Theme in Biblical Hebrew Narrative and Poetry’. Unpublished D. Litt. Thesis, University of Stellenbosch Huddleston, R.D. and G.K. Pullum. 2002. The Cambridge Grammar of the English Language. (Cambridge) König, E. 1991. The Meaning of Focus Particles: A Comparative Perspective. (Lon don) Lyavdansky, A. 2004. ‘Gam in the Prophetic Discourse’, Babel und Bibel 1, 231 50 Moshavi, A. forthcoming. Word Order in the Biblical Hebrew Finite Clause. (Eisenbrauns) Van der Merwe, C.H.J. 1990. The Old Hebrew Particle Gam: A Syntactic Semantic Description of Gam in Gn 2Kg. (Arbeiten zu Text und Sprache im Alten Testa ment 34. St. Ottilien) 1991. ‘The Old Hebrew “Particles” ‘ak and raq (in Genesis to 2 Kings)’, in W. Gross, H. Irsigler and T. Seidl (eds), Text, Methode und Grammatik: Wolfgang Richter zum 65. Geburtstag (St. Ottilien). 297 311 1993a. ‘Pragmatics and the Translation Value of Gam’, Journal for Semitics 4, 181 99 1993b. ‘Old Hebrew Particles and the Interpretation of the Old Testament Texts’, JSOT 60, 27 44 2006a. ‘Biblical Hebrew Lexicology: A Cognitive Linguistic Perspective’, Kleine Untersuchungen zur Sprache des Alten Testaments und seiner Umwelt 6, 87 112 2006b.‘Lexical Meaning in Biblical Hebrew and Cognitive Semantics: A Case Study’, Biblica 87:1, 85 95 2007.‘A Cognitive Linguistic Perspective on ִהנּ ֵהin the Pentateuch, Joshua, Judges and Ruth’, Hebrew Studies 47, 238 77 Van der Merwe, C.H.J., J.A. Naudé and J.H. Kroeze. 1999. A Biblical Hebrew Ref erence Grammar. (Sheffield) Van der Merwe, C.H.J. and E. Talstra. 2002/3.‘Biblical Hebrew Word Order: The Interface of Information Structure and Formal Features’, ZAH 15/16, 68 107 Van der Merwe, C.H.J. and J.A. Naudé. forthcoming. A Biblical Hebrew Reference Grammar: A Revised and Expanded Version
332
LEXICOGRAPHICAL ARRANGEMENT OF MASORETIC MATERIAL1 JOSÉ MARTINEZ DELGADO UNIVERSIDAD DE GRANADA
Abstract This study includes an edition, translation and analysis of three lists transmitted by the manuscript Firkovich I 4557 (ff. 4 9). The scribes attached this treatise to Îayyuj’s Kitab al Tanqi†/Treatise on Vocaliza tion and they also attributed it to him; but its contents confirm that this material predates Îayyuj. The tripartite division of this treatise shows that in the mind of the compiler the lexicographic criterion predominates and the analysis shows that the arrangement of these lists has a strong lexicographic nuance.
The lists that follow below form a clear link between the Masora and grammar, since they reflect the change both in mentality and in methodology.2 These Masoretic lists were found attached to two manuscript copies of the Kitab al-Tanqi†/Treatise on Vocalization by Îayyuj.3 This brief treatise, written in Cordoba in the tenth century, was a pioneer in both its topic and its language: it was the first time that the Tiberian Hebrew vocalization had been examined grammatically and one of the reasons for its success and widespread dissemination was its language, Arabic. 1
This study has been developed within the frame the research project HUM2006 02495/FILO (Lengua y Literatura del Judaísmo Clásico: Rabínico y Medieval). A shorter version was submitted to the program of the 20th Congress of the International Organization for Masoretic Studies (IOMS) held in association with IOSOT XIXth Congress 2007 in Ljubljana, Slovenia, taking place from 15 to 20 July, 2007. 2 On this topic see Aron Dotan, ‘De la Massora à la grammaire: les débuts de la pensée grammaticale dans l’hébreu’, Journal Asiatique 278, 1 2 (1990), 13 30. 3 The only complete copy of these lists is preserved in manuscript Firk. I 4557. Furthermore, Firk. I 2408 only preserves one sheet (6r 6v). For the description of these copies, see my edition of Îayyuj’s opuscule in José Martínez Delgado, ‘El opúsculo sobre la normativa vocálica de Îayyuy (edición y traducción)’, Miscelánea de Estudios Árabes y Hebraicos, Sección de Hebreo 54 (2005), 185 230.
333
LEXICOGRAPHICAL ARRANGEMENT OF MASORETIC MATERIAL
It seems that when one acquired the Treatise on Vocalization by Îayyuj from a scribe, the volume would also contain two other works attributed to Îayyuj. Five of the twelve manuscript copies of Îayyuj’s Treatise on Vocalization in the Firkovich collection include a work on the Masoretic accents attached as a colophon.4 This opuscule on accents is attributed to Îayyuj in all the copies. Indeed, it was even published in the nineteenth century as part of the translation of Îayyuj by Abraham Ibn ¨Ezra, suggesting that he also understood it to be part of the Treatise on Vocalization.5 However, a study of the contents shows that this material predates Îayyuj. In addition to this, in two of the twelve copies, another work based on lists transmitted by al-Masora, according to the headings, was also preserved. 6 Only one of the copies preserves the complete text while in the other (Firk. I 2408) we only find fragments of the beginning of the work which, by analogy, the copy also attributes to Îayyuj. The only certainty seems to be that in the Middle Ages, the scribes presented Îayyuj’s treatise with these appendices, creating a unit dedicated to Masoretic punctuation. Description of the lists These lists are made up of three essays that have been put together as a single unit, both thematically and stylistically, resulting in three chapters that are differentiated by their headings. The first is dedicated to the exchange of consonants, the second to homographic pairs and the third to punctuation. The work, coming from the Masoretic tradition, is anonymous. The method followed does not know the root theory established by Îayyuj in the second half of the tenth century and imposed on all the Arabophone communities. The lexicographic criterion applied to the lists is graphic. This factor confirms its origin in the Masora and 4 On the nature of this appendix on accents, see Ilan Eldar (2001) ‘Ha maqor sel ha tosafot le Kitab al tanqi† (Sefer ha niqud) li Yehudah Îayyuj’, Language Stud ies VIII, 141 91 and José Martínez Delgado, ‘Un Capítulo sobre los acentos bíblicos atribuido a Îayyuy’, Miscelánea de Estudios Árabes y Hebraicos, Sección de Hebreo 55 (2006), 475 513. 5 John William Nutt, Two Treatises on Verbs containing Feeble and Double Letters by R. Yehuda Hayug of Fez, Translated into Hebrew from the original Arabic by R. Moseh Gikatilia of Cordova; to which is added the Treatise on Punctuation by the same Author translated by Aben Ezra (London Berlin 1870). 6 Aron Dotan showed in his study, NiÒanim Riso’nim be Ìokmat ha milim (Tel Aviv 2005), pp. 110 11, that when speaking of al Masora al Kabira, medieval Jews were referring to what today is known as the Sefer OÈla we OÈla, edited by Fernando Díaz Esteban (Madrid 1975).
334
LEXICOGRAPHICAL ARRANGEMENT OF MASORETIC MATERIAL
guarantees its age because of its similarity to authors like Judah Ibn Qurays,7 David ben Abraham al-Fasi8 and Menachem ben Saruq,9 all from the first half of the tenth century and earlier than Îayyuj. One must distinguish in the case of these lists between the primary linguistic sources on the one hand, viz., the Bible in both its Hebrew and Aramaic portions, which are treated as a single linguistic unit, and, on the other, the secondary or metalinguistic sources, viz., lists transmitted by the Masora, lexicographic notes of unknown origin, and notes by the author himself or derived from Rabbinic literature. 1. Lexicographic analysis of the first list: exchanges Heading: אלקול פי חרוף אלאבדאל נקל מן אלמאסרה/ Chapter on the exchangeable
letters transmitted by the Masora This list is the most complex. It seems to be subdivided into three sections: 1.1. List of forty-five exchanges of healthy letters: zayin-Òade, sinzayin, Òade-sin, yod-’alef, gimel-qof, Ìet-¨ayn, Òade-†et, res-dalet, reslamed, nun-res, pe’-bet, mem-pe’, ’alef-‘ayn, Ìet-¨ayn, sin-Òade, kaf-gimel, dalet-zayin, he’-’alef, yod-’alef, nun-lamed, ’alef-¨ayn, bet-pe’, gimel-kaf, kaf-gimel, dalet-Òade, res-dalet, he’-Ìet, waw-yod, zayin-Òade, †et-taw, Ìet-kaf, lamed-res, mem-nun, sin-gimel, sin-zayin, Òade-†et, sin-sin (i.e. samek), mem-pe’, qof-Òade, Òade-sin, sin-sin, taw-†et, Ìet-he’, ¨ayn-Ìet, taw-res. 1.2. List of nine exchanges of the following servile letters: yod, taw, nun (twice), he’ (twice), sin and kaf (twice). 1.3. Lists of interchangeable words that are only differentiated by one of the following exchanges: ’alef-he’, bet-waw, gimel-he’, ’alef-Ìet, †e†-Òade, lamed-nun, ‘ayn-Ìet, qof-‘ayn (Aramaic), sin-Òade, and nuntaw. A more detailed analysis would seem to confirm that, in fact, the first section, that of the healthy letters (1.1.), could be created from two different lists: 1.1.1. List without any established order containing twenty exchanges: zayin-Òade, sin-zayin, Òade-sin, yod-’alef, gimel-qof, Ìet-¨ayn, 7
Dan Becker, Risala Yehudah Ibn Qurays (Tel Aviv 1984). Solomon Leon Skoss, The Hebrew Arabic Dictionary of the Bible Known as Kitab Jami¨ al Alfa (Agron) (New Haven 1936 45). 9 Ángel Sáenz Badillos, MaÌberet MenaÌem ben Saruq (Granada 1986). 8
335
LEXICOGRAPHICAL ARRANGEMENT OF MASORETIC MATERIAL
Òade-†et, res-dalet, res-lamed, nun-res, pe’-bet, mem-pe’, ’alef-‘ayn, Ìet¨ayn, sin-Òade, kaf-gimel, dalet-zayin, he’-’alef, yod-’alef, nun-lamed. 1.1.2. List made up of twenty-five exchanges that tries to follow an alphabetical order: ’alef-‘ayn, bet-pe’, gimel-kaf, kaf-gimel, dalet-Òade, res-dalet, he’-Ìet, waw-yod, zayin-Òade, †et-taw, Ìet-kaf, lamed-res, mem-nun, sin-gimel, sin-zayin, Òade-†et, sin-sin (i.e. samek), mem-pe’, qof-Òade, Òade-sin, sin-sin, taw-†et, Ìet-he’, ¨ayn-Ìet, taw-res. 1.1. Exchanges of healthy letters. 1.1.1. First twenty exchanges. The structure of what is considered the first part of the first list can be said to be made up of twenty exchanges. It lacks any alphabetical criterion or order. The set sequence is [exchange: case1 = case2], for example: . ויצעק ויזעק.אלזין פי מוצע אלצדי זעיר שם זעיר שם תקול בדלה צעיר אני לימים
Zayin instead of Òade: ָם$ָם ז ְֵעיר ש$( ז ְֵעיר שIsa. 28:10) is said instead of ָצִעיר מים ִ ָ אנ ִי לְי ֲ (Job 32:6); עק ַ צ ְ ִ ַ( ויּGen. 27:34) and עק ַ ְ ִ( ויַּ ז1Sam. 7:9). [4r:18 19]10
In exchange no. fourteen, the sequence changes, resulting in a new formula: [it has been said that: case1 = X: exchange]: . אלחא באלעין.וקיל אן אבחת חרב יריד בה אבעת חרב
It has been said that: ( ִאְבַחת חֶָרבEzek. 21:20) in this case means ִאְבַעת חֶרב ָ . Îet with ¨ayn. [4v:10]
The alteration of the sequence seems to indicate that this case does not come from the Masora. In fact, it is treated as a hapax legomenon of disputed authorship in all the tenth-century medieval dictionaries. Menachem ben Saruq attributes it to Judah Ibn Qurays and criticizes it severely, since more exchanges than those of the letters ’alef, he’, waw and yod are not accepted in the MaÌberet.11 This exchange can also be found attributed to R. Hay in medieval lexicographic literature.12 Another factor that may confirm the unity of these first twenty exchanges, except the last-mentioned, is that all have a real parallel in the Scripture, meaning that variants exist. In this way, lexicographically, it is established that ז ְֵעירwith zayin is a synonym of 10
The following numerations in brackets allude to the edition of the manu script Firk. I 4557 at the end of this study. 11 See Ángel Sáenz Badillos, MaÌberet MenaÌem, 19* 21* headword ’bÌ. 12 Thus, for example, in the comparative documents by YiÒÌaq Ibn Barun from Zaragoza: For his part, Rabbi Hay with God has his in His glory, understood it (’ibÌat) to be identical to ’ib¨at/terror, substituting ¨ayn (for Ìet). Pável K. Kokóvtsov, Kitab al muwazana bayn al luga al ‘ibraniyya wa l ¨arabiyya (Saint Petersburg 1890), 27. 336
LEXICOGRAPHICAL ARRANGEMENT OF MASORETIC MATERIAL
עיר ִ צ ָ with Òade, as are they all, except חֶרב ָ חת ַ ב ְ א ִ (Ezek. 21:20), since עת חֶָרב ַ ב ְ א ִ does not exist in the Scripture. For this reason, these cases
without parallel in the Scripture are understood to be subsequent glosses. In fact, the exchange Ìet-¨ayn had been previously recorded in [4v:16–17]: Îet with ¨ayn: ַ( ָסרוּחExod. 26:13) and רוּע ַ ש ָ (Lev. 21:18). A similar case can be found at the end of the first part: וקיל אן כי היה דבר המלך נחוץ יריד בה. נתע ולתע. ואלנון באללאם נשכה ולשכה לחוץ
Nun with lamed: ָכּה$( נ ְִשNeh.13:7) and ָכּה$( ִלְשNeh.13:5); ( נ ִַתּעcf. Job 4:10) and ִלַתּע. It has been said that ( ִכּי הָי ָה ְדבַר הֶַמֶּלְך נ ָחוּץ1 Sam. 21:9) in this case means ָלחוּץ. [4v:16 17]
1.1.2. Alphabetical list of the exchange of healthy letters. Regarding the structure of the second part of the first list, it can be said that it is much more complex than the first on all levels. Two factors lead us to separate this list into two parts: firstly, the second part is clearly different in nature and secondly, the second part tries to maintain an alphabetical order and its sequences differ from those in the earlier part. Additionally, its lexicographic complexity when compared to the first twenty exchanges is supported by the use of secondary sources to illustrate the exchanges. Four times the text reproduces passages in Hebrew. These glosses in Hebrew seem to come from a supposed original that is much older. This type of ‘clarification’ in Hebrew may vary in nature: a) they may be instructions: כל דבר שכתוב באלף ואין פירושו באלף כתקנו מפרשין אותו בעין … וכן כל דבר .שהוא בעין ואין פירושו בעין כתקנו מפרשין אותו באלף
All words written with ’alef, but which cannot be interpreted with ’alef, were corrected by the interpreters with ¨ayn […] Likewise, all words with ¨ayn, but which cannot be interpreted with ¨ayn, were corrected by the interpreters with ’alef. [4v:20 5r:1]
b) they may be annotations that are Masoretic in nature: .וכל לשון ויסירו נכתב בסמך ב'מ'א' כי גם אוי להם בשורי מהם
All words ( ויַּ ִָסירוּJudg. 10:16) are written with samek in the Scripture except ( ִכּי ג ַם או ֹי לֶָהם בְּש ֹוִּרי מֵֶהםHos. 9:12). [5v:9 10]
c) they may be annotations referring to the meaning: . מהללים מחללים לשון שבח. ואלחא באלהא עיר ההרס האומר לחרס
Îet with he’: ( ִעיר הֶַהֶרסIsa. 19:18) and ( ָהאוֵֹמר לֶַחֶרסJob 9:7). ְמַהְלִּלים (1Chron. 23:5) and ( ְמַחְלִּלים1Kgs 1:40), it means praise. [5v:17 18] 337
LEXICOGRAPHICAL ARRANGEMENT OF MASORETIC MATERIAL
. ופי תורת כהנים יקול צהוב למה הוא דומה לתבנית זהב. שער צהוב זהוב ער צָהוֹב ָ ֹש ֵ (Lev. 13:30) (means) ז ָהוֹב. Sifra’ (Lev.) equates צהו ֹב ָ with that which appears to have the quality of gold (Sifra Lev. 5,5). [5r:20 1]
The most common sequence in this second part is still [exchange: case1 = case2]: .ואלצאד באלשין צוררי שוררי
∑ade with sin: ( צוְֹרָריPs. 6:8) and וְֹרָרי$( שPs. 5:9). [5v:13 14]
Along with this, or derived from it, we find more complex formulas created from the basic sequence. These complex sequences are by necessity lexicographic, meaning that it is necessary to apply the exchange to understand the word. However, there is no Biblical parallel to authorize its use and the result is thus, for example, [exchange: case1 = X], where X is not supported: . בהצותם בחצותם.ואלהא באלחית הצו אשר חצו
He’ with Ìet: ( ִהצּוּNum. 26:9) (means) ר חִצּוּ$( ְבַּהצּוָֹתם ;ֲאֶשNum. 26:9) (means) ְבַּחצּוָֹתם. [5r:15 16]
With these two formulas combined, linguistic syllogisms or syllogistic formulas can be applied, developing two complex variants, depending on the difficulty of the line of proof: a) [exchange: case1 = X + Biblical prooftext → case2 = X → case2ª = X → case3 = X] לפני אפרים קיל יריד. ישב רוחו ונשבה בו.ואלבית באלפא נשפת ברוחך יריד נשבת . פרזלא ברזלא. וכדלך קול רך ויחיד לפני אמי לבני אמי.בה לבני
Bet with pe’: ְפָתּ ְברוּחֲָך$( נ ַָשExod. 15:10) means ְבָתּ$נ ַָש, (from) ֹב רוּחו$י ֵַשּ (Ps. 147:18) and ָֹבה בּו$( נ ְָשIsa. 40:7). ( ִלְפנ ֵי אְֶפַרי ִםPs. 80:3), it has been said that in this case it means ִֵלְבני. Likewise, the verse ַרְך ו ְי ִָחיד לְִפנ ֵי אִִמּי (Prov. 4:3) (means) לְִבנ ֵי אִִמּי. ( ַפְּרז ְָלאDan. 2:35) (means) ַבְּרז ְָלא. [5r:7 10]
b) [exchange: case1 = case2 → case2 = case1 → case2 = X + non-Biblical prooftext] ופי. שער צהוב זהוב. בין שורותם יצהירו ויריד יזהירו.ואלזאי באלצדי זוהר צוהר .תורת כהנים יקול צהוב למה הוא דומה לתבנית זהב
Zayin with Òade: ( זוַֹהרEzek. 8:2) and ( צוַֹהרGen. 6:16); וּרוָֹתם$ֵבּין ש ִ ְ ַ( יזDan. 12:3). ער צָהוֹב ָ ֹש ֵ (Lev. הירוּ ִ צ ְ ַ ( יJob 24:11) which means הירוּ 13:30) (means) ז ָהוֹב. Sifra’ (Lev.) denominates ָצהוֹבwith that which appears to have the quality of gold (Sifra Lev. 5:5). [5r:19 21] 338
LEXICOGRAPHICAL ARRANGEMENT OF MASORETIC MATERIAL
As can be seen, the exchange authorized by the basic sequence guarantees, furthermore, other analogous cases by means of this complex formula. Therefore, the sequence is multiplied, authorizing other cases that are supported by the tradition. This is, in short, the creation of analogous syllogisms that help with the interpretation of the Scripture. Finally, the (exchange: Biblical Hebrew [BH] = Rabbinic Hebrew [RH]) sequences are notable; they seem to contain the germ of comparison between the two phases of the language: .ואלסין באלזין זכוכית סכוכית
Sin with zayin: ( ז ְכוִֹכיתJob 28:17) and ְסכוִֹכית. [5v:6 7]
In other places, the Biblical exchange is attested or is confirmed by the Rabbinic literature. In these cases, the extra-Biblical prooftext appears along with the formula, but it is interesting that it is an independent formula [RH = X + Biblical prooftext + non-Biblical prooftext]. ופי לגה אלמשנה גמיעה יריד גמיאה מן הגמיאיני כקול החושש בשניו אל יגמע בהן .חומץ
In the language of the Misna ( ג ְִמיעָהShab. 8:1) means יאה ָ מ ִ ְ ג, from ִיאיני ִ מ ִ ְ הג ַ (Gen. 24:17), as with the passage הן אֶת ֶ ָמּע בּ ַ ַ ְנּ ָיו אַל יג$ש ִ ְ בּ$ש$ש ֵ ֹהחו ַ מץ ֶ ֹ( הַחוcf. Shab. 14:4). [5r:4 5]
1.2. Exchanges of servile letters. Composed of nine cases, the servile letters are understood to be, in this order, yod, taw, nun (twice), he’ (twice), sin and kaf (twice). The name given to these consonants is Ìuruf al-tafkhim ‘emphatic letters’, a term employed by all the tenth century eastern Hebrew grammarians, Karaites13 and Rabanites.14 The compiler explains that this refers to consonants that ‘are added to the word,’ i.e., that do not form part of the ‘root’ but are the result of the inflection. Here also are found formulas that try to create semantic pairs, since the formula provides both vocables with a single meaning. The basic formula is [case1 = case2]. This can be lengthened four times if necessary, thus: . לנחותם להנחותם. בדרך בהדרך.יוצדק יהוצדק ָ ֹ( י ְהוHaggai 1:1). בֶּדֶּרְך ַ (Gen. 24:27) and צָדק ָ ֹ( יוEzra 3:2) and צָדק הֶדֶּרְך ַ בּ ְ (Neh. 9:19). תם ָ ֹלנ ְחו ַ (Exod. 13:21) and תם ָ ֹהנ ְחו ַ ל ְ (Neh. 9:19).
[6r:1] 13
See, for example, Solomon Leon Skoss, The Hebrew Arabic Dictionary. See, for example, Aron Dotan, ’Or rison be Ìokhmat ha lason: sefer ∑aÌut leson ha ¨Ivrim la Rav Se¨adyah Ga’on (2 vols.) (Jerusalem 1997). 14
339
LEXICOGRAPHICAL ARRANGEMENT OF MASORETIC MATERIAL
There is also the basic formula, [case1=X], although in this case it entails the acceptance of a de facto qeré: .כנלותך יריד בנלותך ְ ֹבּנּ ְלו ַ . [6r:2 3] תָך ְ ֹכּנּ ְלו ַ (Isa. 33:1) means תָך
The exchanges of the servile letters may help to explain the analogical sequence applied by Îayyuj to the weak root, and is the only model established to date. Îayyuj sought the original from the explicit form. In this case, the process is the inverse, moving from the original form recorded in the Scripture to the explicit form via the syllogism [case1=case2 → case3 = X–Y]: . תהנזיק תהזיק תזיק.וכעת וכענת ֶ ְ( וּכEzra 4:10). הנזְ ִיק ַ תּ ְ (Ezra 4:13), הזּ ִיק ַ תּ ְ and עת ֵ כ ָ ְ ( וJudg. 13:23) and ענ ֶת תּזּ ִיק ַ . [5v:21 6r:1]
1.3. Interchangeable words. This list, which tries to follow an alphabetical order, is created from ten exchanges: ’alef-he’, bet-waw, gimel-he’, ’alef-Ìet, †e†-Òade, lamednun, ¨ayn-Ìet, qof-¨ayn (Aramaic), sin-Òade, and nun-taw. The lexicographic formulas are the two basic ones [case1 = case2] and [case1 = X] depending on the form analysed. Of all the sections, the one dedicated to words that can be interchanged by the exchange of ’alef-he’ is most noteworthy. This list is complex. First it presents the cases that follow the [case1=case2] formula and that appear to have the greatest prestige because they are recorded in the Bible. Next, it follows the cases expressed by means of the [case1=X] formula and returns again to the first [case1=case2] formula. The lexicographic importance of this list lies in the fact that of the sixteen cases, seven of them are collected in the introduction to the Book of Îayyuj.15 In fact, the similarity between the examples used by Îayyuj to establish his theory about the graphic and phonetic alternation of weak letters and the list itself is so close that it does not seem unreasonable to assert that Îayyuj knew this list or, less likely, that the list was formed from the Book of Îayyuj.16 15 These cases are: ( ֶהֶדרDan. 11:20) and ( ֶאֶדרMic. 2:8), י אֶג ְָאְלִתּי$( וכל מלבושIsa. 63:3) and ֶהג ְָאְלִתּי, ( ֶאְתַחַבּר2 Chron. 20:35) and ִהְתַחַבּר, ( ֵהיְךDan. 10:17) and ֵאיְך, תיוּ ָ ה ֵ (Jer. 12:9) and תיוּ ָ א ֵ (Isa. 56:12), אמוֹן ָ ה ָ (Jer. 52:15) and המוֹן ָ ה ֶ , כּים ֵ $ש ְ א ַ (Jer. 25:3) and ֵכּם$ַהְש. See Morris Jastrow, The weak and geminative verbs in Hebrew by Abû Zakariyyâ YaÌyâ ibn Dâwud of Fez, known as Îayyûg (Leiden 1897), 13 14; John W. Nutt, Two treatises, 9; and José Martínez Delgado, YaÌyá ibn Dawud, El libro de Îayyuy: versión original árabe del siglo X (Granada 2004), 45. 16 It is easy to see the relation between the two texts in the following example:
340
LEXICOGRAPHICAL ARRANGEMENT OF MASORETIC MATERIAL
2. Lexicographic analysis of the list of homographic words Heading: והדה אלאלפאט מזדוגה מכתלפה אלעבארה אלדי יגב עלי אלמתעלם אלוקוף עליהא וחפטהא פאנהא קריבה אלמעני וקד דונוהא אלאואיל פי אלמסורה . אלכבירה ורתבוהא עלי אלף בית ליסהל עלי אלטאלב חפטהא/ These words
are pairs with different interpretations, which require that the scholar stop and memorize them, since their meanings are very close. The ancients collected these in the Great Masora and arranged them alphabetically to facilitate their memorization by scholars. This is a glossary made up of ninety-three headwords ordered in a strict alphabetical order. It seems to be a faithful reproduction of the primitive Masoretic composition that, already in its origin, had taken on more the form of a glossary than of a simple list. This list is contained in the current edition of the Sefer OÈla we-OÈla (list 60 of the volume by Fernando Díaz Esteban)17 and Aron Dotan (2005) recently dedicated a complete monograph to it.18 The headwords follow a strict alphabetical order, except where scribal error can be identified. The headword criterion is graphic, i.e., the word is presented exactly as in the Bible, in such a way that its prefixes, when they occur, influence the order. These are, in short, ninety-three words that appear twice in the Bible, are graphically identical, but have a different meaning. Someone altered the original list for reasons of economy. With the first four words, the headword is repeated before each of the cases. However, beginning with the fifth, the headword comes only before the first case. After the headword, the letter bet appears, marked as an abbreviation. There are two possible ways to understand this abbreviation; on the one hand, it may have a numeric value and indicate the frequency of the form in the Bible; on the other, it may be the abbreviation of the Arabic expression, bi-qawl (‘in the verse’). This second possibility means that someone must have altered the original arrangement of the list. Îayyuy claims that ‘It was said ( ֵהָתיוּ לאכלהJeremiah 12,9), being its original form תיוּ ָ א ֵ , like תיוּ אקחה יין ָ א ֵ (Isaiah 56,12), because it is from תה בקר ָ א ָ (Is 21,12) and from מנים$תיוּ הש ָ א ֱ ֶ ( יSal 68,32)’, see Morris Jastrow, The weak and geminative, 13 or the Spanish version at José Martínez Delgado, YaÌyá ibn Dawud, 45; while the list recordered התיו אתיו. 17 Fernando Díaz Esteban, Sefer OÈla we OÈla (Madrid 1975), 103 6. 18 Aron Dotan, NiÒanim Riso’nim. 341
LEXICOGRAPHICAL ARRANGEMENT OF MASORETIC MATERIAL
Another characteristic of this glossary is its primitive way of citing the cases. The headword indicates the word that is going to be discussed. It is followed by one of the verses that contains the form. If the form is at the beginning of the verse, the headword forms part of the citation. However, when the headword is in the middle or at the end of the verse, the first two or three words from the beginning of the verse come after the headword although they do not contain the form, i.e., the allusion is to the passage, not to the form. This method guarantees that the reader understands the form in its authentic context. The established lexicographic formula that is applied to the ninety-three cases is [headword: case1 ≠ case2]. The relationship or conflict created between these ninety-three cases can have different values: a) Semantic: a.1. Phonetic relationship: this is the case with ַבַּמָּסּה, written once with samek in Deut. 6:16 (toponym) and another time without sin in 1 Chron. 15:22 (‘in transportation’) [6v:6–7]. a.2. Semantic proximity: certain pairs express similar or derivative concepts, for example, ז ְַקןin one case refers to an old man and in another to a beard [7r:17–18]. Something similar occurs with זונ ֶה, which in one case means to prostitute oneself and in another to desert [7r:17–18]. It seems, then, that this is the prototype of what later would be known as metaphoric sense, which is always derived from the basic sense.19 Another two types of semantic relationship with a strong presence in this list are: a.2.1. Onomastic: ָרכה ָ ב ְ appears as a proper noun in 1 Chron. 12:3, while Prov. 11:26 means ‘blessing’ [6v:18-19]. A semantic relationship can, therefore, be established between the noun and the proper noun [7r:13–14]. Similarly, ׁוֹהם ַ , at one point is a precious stone (onyx?) and at another a proper noun. Something similar occurs with ִַהְצּפוני, which at one point is a proper noun and at another is demonym [6v:16]. a.2.2. Toponym: the relationship is very similar to that of the onomastic: ֲחִליis at one point a toponym and at another a jewel [7r:2–3]. The toponym may come into semantic conflict with the onomastic and thus, ֶטֶלםis at one point a toponym and at another a proper noun [7r:3–4]. One of the most interesting cases is the com19
On this topic in Andalusian Hebrew lexicography see José Martínez Delgado, ‘Maimonides in the Context of Andalusian Hebrew Lexicography’, Aleph: Histori cal Studies in Science and Judaism 8 (2008), 15 40. 342
LEXICOGRAPHICAL ARRANGEMENT OF MASORETIC MATERIAL
parison of ָדּן, which in one case is a toponym in Israel and in another a foreign city [7r:6]. a.3. Antonyms: some pairs seem to come into authentic semantic conflict and become natural antonyms. For example, זורוּin one case expresses the action of compressing and in the other of twisting [7r:18–19]. The latter definition can be understood as an antonym of the former, creating a natural pair: compress-expand. ֲחלוּץin one case means ‘equipped’ while in another ‘barefoot’ [7r:21]; both concepts are also natural antonyms. b) Morphological: at times the pairs presented are homographic words that are different in nature. Thus ָאִמירis one case is the first person common singular of the imperfect hif ¨il of a verb with of a weak second radical, while in another it is a noun [6r:21–6v:1]. In other cases, the paradigms are totally different, such as ְלַתנּות, which in one case is a construct infinitive and in the other is an indirect object [7v:12–13], or מול, which in one case is a preposition and in another an absolute infinitive [7v:15–16]. One final example of this type involves the case where the form can represent an isolated word or a different word with a prefix, such as ְבָּשִרים, which means both ‘bodies’ and ‘above princes,’ i.e., besarim as a unit versus be-sarim as a compound [6v:8–9]. c) Theological: at times the pair ִבְּראותis in conflict [6v:7–8]. In the first case it is interpreted as ‘in the sight of ’ and in the second as ‘in the fear of.’ Morphologically, there is no differentiation; however, if this view is not accepted, the passage would have to be interpreted literally as saying that Uzziah ‘set himself to seek God in the days of Zechariah, who instructed him in the sight of God.’ Therefore, the anthropomorphism is discarded. This category may include the relationship established between the pair [ נ ֵַצח8r:6–7], which in one case is a name of God and in another the adverb ‘always,’ which can be understood as the ‘Eternal.’ d) Miscellaneous: at times the pairs have different reference points, somehow related. Thus, ָהָאְבניָ ִם, is at times a delivery room and at others a potter’s wheel [6v:13–14] and נ ְָבֶליָך, which on each of the two occasions in which it appears refers to a different musical instrument (lyres and bagpipes) [8r:5–6]. e) Casual: this term concerns those pairs which refer to totally different meanings and do not belong to different morphological paradigms. For example, ַמֲעַדנּותin one case is ‘delights’ and in another is ‘chains’ [8r:4]; יגָ ֶלin one case is ‘to roll’ and in another is ‘to exult’ [7r:4–5]. 343
LEXICOGRAPHICAL ARRANGEMENT OF MASORETIC MATERIAL
f ) Bilingual: on one occasion a pair made up of one Hebrew word and one Aramaic word, ָמִרים, is opposed [7v:21]. Since the Masora understands the Bible to be an inseparable unit, it would be more appropriate to include them in the group of morphological relationships (b). g) Lost: these are pairs that contemporary lexicography has rejected or omitted. However, the conflict inherent in their interpretation can be found in Ferrara’s Romance translation of the Bible. This is the case with [ ַהַתּנּ ִין6v:17], which in Isa. 27:1 is translated as the ‘culebro’ that accompanies Leviathan and means dragon, while in Neh. 2:13 it is translated ‘culebra’, which means serpent. Of the material that I have checked, this translation is the only case that presents this difference. The same occurs with [ ו ְַתַּכל7r:14–15] in Gen. 24:19 translated ‘y atemó’ (= he ended), while 2 Sam. 13:39 is the only case that interprets it as ‘y desseó’ (= ‘she’ wished), understanding the subject of the oration to be implicit, i.e., the soul (ו ְַתַּכל פש ָדּו ִיד ֶ ֶ )נ. h) Ambiguous: at times it is very difficult to determine the nuance intended by the pairing. [ ַכּחול7r:15–16] in Gen. 22:17 seems to have the clear meaning of ‘like sand,’ while in Job 29:18 it is more difficult to determine what is the appropriate interpretation. Current lexicography understands that this is identical to the earlier case, which is impossible here. In the Middle Ages, a clear allusion to the phoenix was seen in this passage and this is reflected in Ferrara’s translation. Another possibility is that it could be understood to mean ‘like a palm tree.’ Clearly, it is not easy to decide in favour of one or the other, for which reason this group is called ambiguous. 3. Analysis of the third list: Hebrew punctuation Heading: אלכלאם פי אלנקט/ Chapter on punctuation
Having completed a section dedicated to the graphic word in the material on exchanges, and having produced a glossary that reflects the nature of the phonological word, the compiler’s only remaining task is to offer a series of notations on the nature of the lexicogrammatical word. This section is by far the most important of the treatise, since it serves as a link in the process of the evolution of the Masora towards grammar. The chapter is dedicated to vowels and their alterations. However, the author only considers those that he believes may have semantic 344
LEXICOGRAPHICAL ARRANGEMENT OF MASORETIC MATERIAL
or morphological repercussions, that is, in this order, the segol-Òere and pataÌ-qameÒ binomials. The author studies the relationship that these pairs have with each other and with their environment, such as, for example, accents and dages. It may have been the nature of this last chapter that led scribes to include it alongside Îayyuj’s Treatise on Vocalisation and the chapter devoted to Masoretic accents, thus forming a very interesting and complete triptych devoted exclusively to the Hebrew cantillation system. The author’s conception is pre-grammatical. One example of this can be seen in his formulation of how a word is determined by its article, where the same definition used today is found, but the inverse: וכל הא גאת ללמערפה מקמוצה מתל העיר האיש אלא אן כאן בעדהא חית או דגש .פאנהא תפתח מתל החי החדש החרב המדבר הממלכה הגוי המלך המים
Every he’ that determines (a noun) has qameÒ, for example, ָהִעירand $איש ִ ה ָ . However, if it is followed by Ìet or a dages, it has pataÌ, for example, ַהחַי, $ַהֹחֶדש, ַהֶחֶרב, ַהִמְּדבָּר, ַהַמְּמָלָכה,ַהגּוֹי, ַהֶמֶּלְךand ַהָמּי ִם. [9r:7 9]
3.1. Vowel with /e/ pitch Beginning with the vowel with the /e/ pitch, he asserts that all words whose last syllable is accented and vocalized with /e/ followed by he’ have three points, that is, segol. However, if the same word is attached to another, it will have two points, that is, Òere: מתל.מא כאן עלי וזן משנה או מראה או מעשה או שה וכאן מקטועא פנקטה ג' נקט ואן כאן מוצולא אלי גירה כאן בנקטתין מתל למראֵה. שה כשבים.נחמד למראה שֵה עולותיך.עיניו.
Something with the form of נ ֶה$ִמְש, ַמְראֶה, ַמֲעֶש ֹהor ֶשֹה, being alone, has three points, for example, ( נ ְֶחָמד לְַמְראֶהGen. 2:9) and ( ֶשֹה כְָשִֹביםDeut. 14:3 sic!). But, if it is in contact with another (element) has two points, for example, ( ְלַמְראֵה עֵינ ָיוIsa. 11:3) and ( ֵשֹה עוֹלוֶֹתיָךDeut. 14:4). [8v:6 9]
As can be seen, on the one hand, the author does not allude at any time to the quantity of the vowel, but to its quality, and, furthermore, he is unaware of the notion of gadol and qa†on. On the other hand, he does not use, and perhaps is not aware of, the names of the vowels with /e/ pitch, but reduces everything to the number of points that reflects this sound. This vowel nomenclature is also used among others by al-Fasi in his dictionary and it is accepted even by very late grammarians such as David QimÌi. Following this same line of argument the conclusion is reached that the po¨el form of a verb with a weak third radical will have segol 345
LEXICOGRAPHICAL ARRANGEMENT OF MASORETIC MATERIAL
when it refers to a single individual, while if its meaning is generic, it will have Òere: ואן כאן כתירא.ומא כאן עלי וזן עושה וכאן ואחדא כאן בג' אבדא מתל היה רועֶה . כל עשֵֹה אלה.כאן בנקטתין מתל כל רועֵה צאן
Something with the form of עוֹשֶֹה, being singular, has always three points, like ( ָהי ָה רוֶֹעהGen. 37:2). But, if it refers to many (people) it has two points like ( ָכּל רוֵֹעה צֹאןGen. 46:34). [8v:9 11]
The pause is responsible for vocalic changes. Drawing inspiration from a common Masoretic formula, he heads his note with the particle kol: . לאנהא בלשון תחנה. וכל תראה בג' אלא ואחדה וכאשר תראֵה עשה
Every ִתְּראֶהhas three, except one, ְראה עֲֵשֹה ֵ ִר תּ$ש ֶ א ֲ כ ַ ְ ( וDan. 1:13), since it is a pause form. [8v:13 14]
This last clarification is in Hebrew, i.e., be-lason taÌana. The fact that he introduced the case with kol and ended it in Hebrew leads us to think that perhaps in this case also he is compiling Masoretic material as in the first list. A similar situation can be also found in the following passage: וקול תמר. אם תעשֵֹה עמנו רעה. ומה תעשֵה לשמך הגדול.וכל תעשה בג' גיר ארבעה ופי ישמעאל בן נתניה כי שקר אתה דובר אל.אלאמנון אל תעשה את הנבלה הזאת . ואלעלה פי דלך אנהא בלשון תחנה.תעשה
Every ַתֲּעֶשֹהwith three (points), except four, ְמָך הַגּ ָדוֹל$וּמה תֲַּעֵשֹה לְִש ַ (Josh. 7:9), ( ִאם תֲַּעֵשֹה עִָמּנוּ ָרעָהGen. 26:29), the sentence of Tamar to Amnon, ( ַתֲּעֵשֹה אֶת הַנּ ְָבָלה הַזֹּאת2 Sam. 13:12) and the sentence of Ismael b. Netania ֶֶקר אַָתּה דוֵֹבר אַל תֲַּעֵשֹה$( ִכּי שcf. Jer.40:16). The reason for this is because these are pause forms. [8v:17 21]
The author or compiler ends this section by explaining that every word that ends with yod and is a plural construct will inevitably be vocalized with Òere. The formula applied in all these cases seems to be [segol ≠ Òere]. 3.2. Vowel with /a/ pitch Here he does know the names of the vowels qameÒ and pataÌ. The first, when it appears in an absolute form, is transformed into the second whenever it is a construct form. As a gloss, he explains that, in the case of the root nun-he’-res, the initial sewa’ assimilates the /a/ sound from the second syllable and, much more importantly, he indicates that there is no phonetic difference between the absolute 346
LEXICOGRAPHICAL ARRANGEMENT OF MASORETIC MATERIAL
form and the construct form, providing confirmation that the author does not recognize any difference in quantity between this pair of vowels: נהרות אלנון במכרג אלהא תכרג נהרות ֲ ַ [ ;נ9r:4 5] הרוֹת ָ ְ נnun sounds with the sound of he’ i.e. הרו ֹת
The paseq accent has a very important semantic function and is used as an extragrammatical element equivalent to the construct and pausal forms [9r:13–17]. In this case as well, the formula is [qameÒ ≠ pataÌ] After introducing an interesting list of simanin [9r:17–18] on words that always have pataÌ and in which the Aramaic formula le† drawn from the Masoretic literature is used, he presents some examples of how the things he has discussed apply to the text. He concludes with some general rules in which he shows the relationship between qameÒ and a letter rafe and between pataÌ and dages. Conclusion The tripartite division of this treatise shows that in the mind of the compiler the lexicographic criterion predominates. In fact, it coincides with the division that is applied today to the lexicographic analysis of words. Contemporary lexicography understands that a word by its very nature has three aspects, viz., the ‘graphic word,’ ‘the phonological word or word-in-discourse’ and finally, the ‘lexicogrammatical word’. The first aspect is represented by the first list on exchanges, i.e., the graphic variants of the word. The second, phonological aspect, is represented by the second list on homographic words, i.e., those which sound the same, but have a different interpretation. The last aspect, the lexicogrammatical, is represented by the grammatical annotations in the third section. Therefore, the arrangement of these lists has a strong lexicographic nuance. English translation: Chapter on the exchangeable letters transmitted by the Masora The letters that can be exchangeable with others are: [First list] 1.
Zayin instead of Òade: ָם$ָם ז ְֵעיר ש$( ז ְֵעיר שIsa. 28:10) it is said instead of מים ִ ָ עיר אֲנ ִי לְי ִ צ ָ (Job 32:6); עק ַ ְ( ויַּ ִצGen. 27:34) and עק ַ ְ ִ( ויַּ ז1Sam. 7:9). 347
LEXICOGRAPHICAL ARRANGEMENT OF MASORETIC MATERIAL
Sin with zayin: ( נ ֱֶעָלָסהJob 39:13) and ָעַלס ;נ ֱֶעָלז ָהand ָעַלז.20 [4b] ∑ade with sin: ( נ ְָתצוּ2Kgs 25:10) and ( נ ְָתסוּJob 30:13). Yod and ’alef: ( ְבּלו ֹי ֵי הסְָחבוֹתJer. 38:11) and ( ְבּלוֵֹאי הְַסָּחבוֹתJer. 38:12). Gimel and qof: שוֵֹרגis ( שֹוֵֹרקJudg. 16:4); יה ָ רוּקּ ֶ ֹש ְ (Isa. 16:8) and שִֹּריג ִים ָ ה ַ (Gen. 40:12); ( נ ַָמקּוּPs. 38:6) and ( נ ָֹמגוּExod. 15:15). 6. Îet with ¨ayn: ַ( ָסרוּחExod. 26:13) and ַ( ָשֹרוּעLev. 21:18). 7. ∑ade with †et: י$( נ ְָקטָה נ ְַפִשJob 10:1) is י$ְקצה נ ְַפִש ָ ָ ;נ21 אקוּט ָ (Ps. 95:10) and אקוּץ ָ (cf. Lev. 20:23). 8. Res with dalet: ְלִהְתגֵֹּדּדand ְלִהְתגֵֹּרר.22 9. Res with lamed: ( ַאְלְמנָֹתיוIsa. 9:15) and ְרמנָֹתיו ְ א ַ (Amos 6:8); היר ִ צ ְ ה ַ ל ְ and היל ִ צ ְ ה ַ ל ְ (Ps. 104:5). 10. Nun with res: ( נוֵֹכלMal. 1:14) means רוֵֹכל: derived from ָרכיל ִ (Lev. 19:16); ( ִכּידו ֹןJob 41:21) and ( ַלִכּידוֹרJob 15:24); ( נ ְבוּכְַדֶראצַּרJer. 21:2) and ( נ ְבוּכְַדנ ֶַצּרDan. 1:18); ( ויַּ ְִתנ ְַכּלוּGen. 37:18) means ַרכּלוּ ְ ת ְ ִ ַ;ויּ מה בְִּרי ַ (Prov. 31:2) means ִבּני. ְ 11. Pe’ with bet: ( גּ ֵַפּיProv. 9:3) and ( גּ ֵַבּיJob 15:26); ( ִפּזּ ַרPs. 53:6) and ִבּזּ ַר (Ps. 68:31). 12. Mem with pe’: ( ז ַָעם י ְיNum. 23:8) and ְ( ז ַַעף ייMic. 7:9). 13. ’Alef with ¨ayn: ( ְמָתֵאבAmos 6:8) and ( ְמָתֵעבIsa. 49:7). 14. Îet with ¨ayn. It has been said that ( ִאְבַחת חֶָרבEzek.l 21:20) in this case means ִאְבַעת חֶָרב. 15. Sin with Òade: $( ִלנ ְֹתשJer. 1:10) and ( ִלנ ְתוֹץJer. 1:10); וְֹרָרי$( שPs. 5:9) and ( צוְֹרָריPs. 6:8). 16. Kaf with gimel: ( כּו ַֹבע1 Sam. 17:5) is גּוַֹבע. 17. Dalet with zayin: ַדּמּוִֹתי23 and ( ז ַמּוִֹתיJer. 4:28); ְדעף ַ ִ י24 and עף ַ ְ ִ( יזProv. 19:3). 18. He’ with ’alef: ( ָהָאמו ֹןJer. 52:15) and ( ֶהָהמוֹן1 Sam. 4:14); ַדבּר ֵ ְ כּם ו ֵ $ש ְ ה ַ (Jer. 7:13) and ֵכּם ו ְַדבֵּר$( ַאְשJer. 25:3). 19. Yod with ’alef: ( ַהַהג ְִריאִים1 Chron. 5:10) and יאים ;ַהַהג ְִרי ִים ִ ְרב ִ ע ַ ה ָ (2 Chron. 17:11) and ( ָהַעְרבִיים2 Chron. 26:7).25 2. 3. 4. 5.
20 Although the form נ ֱֶעָלז ָהis not recorded in the Bible, Ibn Janah claims in his Kitab al usul (root ¨ayin lam sad ) that ‘the meaning of these words with za’, sin and sad is the same, i.e. happiness, delights’ (Adolf Neubauer, The Book of Hebrew Roots [Oxford 1875], 530:29 30). 21 Maybe cf. ָקצה ָ נוּ$ש ֵ פ ְ ַ ( ו ְנNum. 21:5). 22 Both words are recorded in the Bible, but not with these forms. 23 The root is recorded in the Bible, but not in this form. 24 I find no way to explain this form; maybe it refers to יזִ ְַדֵּעף. In this case, the translation should be ‘ י ִַדֵּעףis ( יזִ ְַעףProv. 19:3)’. 25 Both cases are recorded by Hayyuj in his introduction to the weak Hebrew letters (Morris Jastrow, The weak and geminative verbs 14: 17 19).
348
LEXICOGRAPHICAL ARRANGEMENT OF MASORETIC MATERIAL
20. Nun with lamed: ָכּה$( נ ְִשNeh.13:7) and ָכּה$( ִלְשNeh.13:5); ( נ ִַתּעcf. Job 4:10) is ִלַתּע.26 It has been said that ( ִכּי הָי ָה ְדבַר הֶַמֶּלְך נ ָחוּץ1 Sam. 21:9) in this case means ָלחוּץ.27
[Alphabetical list] 21. ’Alef with ¨ayn: ( ֶלֶחם מְגאָלMal. 1:7) means in this case ר ל ֹא י ְִתגּ ַָאל$ֲאֶש ַ ָ תגּ ְ ִ לנוָּך ;ל ֹא י ְ א ַ ֵ מּה ג ֶ ְַרתּם בּ ֶ מ ַ א ֲ ַ ( וMal. 1:7) in this case גוֹעל ָ מ ְ . (Dan. 1:8) is על means ( ויַ ְגוֲֹאלוּ מִן הְַכֻּהנּ ָה ;ג ֵַעְלנוָּךEzra 2:62), in this case means ויַ ְגוֲֹעלוּ. All words written with ’alef, but which cannot be interpreted with ’alef, were corrected by the interpreters with ¨ayn, for example, ( ְמָתֵאבAmos 6:8) and עב ֵ ת ָ מ ְ (Isa. 49:7). Likewise, all words with ¨ayn, but which cannot be interpreted with ‘ayn, were corrected by the interpreters with ’alef, for example, ְ( ו ִַתְּתַפֵּלּל עַל יי1 Sam. 1:10) means ְֶאל הַנּ ַַער הַזּ ֶה הְִתַפָּלְּלִתּי ;ֶאל יי (1 Sam. 1:27); ( ֶאל אֵם הֶַדֶּרְךEzek. 21:26) means עַל אֵם. ֲהַשְֹמָתּ ִלְבָּך אֶל איּוֹב ִ בִדּי ְ ע ַ (Job 2:3) means על ַ , likewise the second one. In the language of the Mishnah ( ג ְִמיעָהShab. 8:1) means יאה ָ מ ִ ְ ג, from ִיאיני ִ מ ִ ְ הג ַ (Gen. 24:17), as with the passage נּ ָיו אַל יגְ ַַמּע בֶָּהן ֶאת הַחוֶֹמץ$ בְִּש$ש$( ַהחוֵֹשcf. Shab. 14:4). ( ֵאֶליהָ י ָבוֹאוּ רוִֹעים ו ְֶעְדֵריהֶםJer. 6:3) means יה ָ ל ֶ ע ָ . It had been said that ָוֹק ו ְֶהָעֶליה$( ֶאת הַשּ1 Sam. 9:24) in this case means יה ָ ל ֶ א ֵ ה ָ ְ ו. 22. Bet with pe’: ְפָתּ ְברוּחֲָך$( נ ַָשExod. 15:10) means ְבָתּ$נ ַָש, (from) ֹב רוּחו$( י ֵַשּPs. 147:18) and ָֹבה בּו$( נ ְָשIsa. 40:7); ( ִלְפנ ֵי אְֶפַרי ִםPs. 70:3), it had been said that in this case means ִֵלְבני. Likewise, the phrase ַרְך ו ְי ִָחיד לְִפנ ֵי אִִמּי (Prov. 4:3) (means) לְִבנ ֵי אִִמּי. ( ַפְּרז ְָלאDan. 2:35) (means) בְּרז ְָלא. ַ 23. Gimel with kaf: ( וּמִג ְָבּעוֹתExod. 28:40) from ֶת$( כוַֹבע נ ְחש1 Sam. 17:5). 24. Kaf with gimel: ( ויַּ ִָסְּכרוּ מְַעי ְנוֹת תְּהוֹםGen. 8:2) means ( ויַּ ִָסּג ְרוּNeh.13:19); ְרתּי ִ כּ ַ ס ִ ְ ( וIsa. 19:4) (means) סגּ ְַרתִּי ִ ְ ו. תם ָ ֹאֶרץ מְכוּרו ֶ ל ְ (cf. Ezek. 29:14) means תם ָ ֹמגוּרו ְ (Isa. 66:4). 25. Dalet with Òade: ( ַהפּוֶֹדהcf. Ps. 34:23) and ( ַהפּוֶֹצהPs. 144:10); ִ( ְפֵּדניPs. 26:11) and ִ( ְפֵּצניPs. 144:7). 26. Res with dalet: as in the phrase ( ֲהַדרGen. 36:39) and ( ֲהַדד1 Chron. 1:50); ( דוָֹדנ ִיםGen. 10:4) and ( רוָֹדנ ִים1 Chron. 1:7); ָבּן$( ֶחְמָדּן ו ְֶאְשGen. 36:26) and ָבּן$( ַחְמָרן ו ְֶאְש1 Chron. 1:41); ( ו ְִריפַתGen. 10:3) and ( ו ְִדיפַת1 Chron. 1:6).28 27. He’ with Ìet: ( ִהצּוּNum. 26:9) (means) ר חִצּוּ$( ְבַּהצּוָֹתם ;ֲאֶשNum. 26:9) (means) בַּחצּוָֹתם. ְ 26
Ibn Janah in his Kitab al usul (root lam ta’ ¨ayin) claims that ‘It had been said
תּעוּ ָ ִ פיִרים נ ִ ְנּ ֵי כ$ש ִ ְ ( וJob 4:10); its original form is תּעוּ ָ ל ְ ִ נbut they exchanged lam with
nun and assimilated it; but it is quite possible also that lam itself is being assimi lated without exchange with nun’ (Neubauer, The book of Hebrew Roots, 360: 14 16). 27 For this conclusion see Ibn Janah’s Kitab al usul (Neubauer, The book of He brew Roots, 426: 6 15). 28 The same cases keeping this order but following by another example can be found in Michael Wilensky, Sefer ha Rikma, vol. I, (Jerusalem 1964), 107:17 18. 349
LEXICOGRAPHICAL ARRANGEMENT OF MASORETIC MATERIAL
28. Waw with yod: ( ַעְלו ָן וּמָנ ַַחתGen. 36:23) and ( ַעְלי ָן וּמָנ ַַחת1 Chron. 1:40). עָקן ֲ ַ ( וGen. 36:27) and עָקן ֲ ַ ( יNum. 33:31). פוֹ ו ְאוֹנ ָם$ש ְ (Gen. 36:23) and פי ו ְאו ֹנ ָם ִ $ש ְ (1 Chron. 1:40). פּעוּ ָ ֹם עִירו$ש ֵ ְ ( וGen. 36:39) and עי ִ פּ ָ ֹ ם עִירו$ש ֵ ְ ( ו1 Chron. 1:50). וּעו ֹת$( ַמג ְִדּיל י ְש2 Sam. 22:51) and וּעוֹת$( ִמג ְדּוֹל י ְשPs. 18:51). 29. Zayin with Òade: ( זוַֹהרEzek. 8:2) and ( צוַֹהרGen. 6:16); וּרוָֹתם$ֵבּין ש ִ ְ ַ( יזDan. 12:3). ער צָהוֹב ָ ֹש ֵ (Lev. 13:30) הירוּ ִ צ ְ ַ ( יJob 24:11) means הירוּ (means) ז ָהוֹב. Sifra’ (Lev.) equates ָצהוֹבwith that which appears to have the quality of gold (Sifra Lev. 5:5). 30. ™et with taw: ( ְפַּסנ ְֵתִּריןDan. 3:5) and ( ְפַּסנ ְֵטִריןDan. 3:7). ַדבּר אְֶפַרי ִם ֵ כּ ְ ( ְרתֵתHos. 13:1) and חז ִיָקה ֱ ֶֶרטט ה ֶ ְ ( וJer. 49:24). In the language of the Mishnah ( ֲאֻלנ ְִתיתGen. Raba 80:9) and ( ֲאֻלנ ְִטיתShab. 22:5). 31. Îet with kaf: ִלים$( נ ְִכָש1 Sam. 2:4) and ִלים$( נ ֱֶחָשDeut. 25:18). [( נ ְִכָתּםJer. 2:22) and] ( נ ְִכָתּב ו ְנ ְֶחָתּםEsth. 3:12). 32. Lamed with res: ( ו ֲַאלוּDan. 2:31) and ( ו ֲַארוּDan. 7:2). מוֹרוֹתand מוֹלוֹת. 33. Mem with nun: ( נוָֹראExodus 15:11) and ( מוָֹראMal. 2:5). ( גּ ַם בְּנ ֵי נוֹףJer. 2:16) and ( מוֹף תְַּקבְֵּרםHos. 9:6). ֵָכר$( ִלְמסוְֹך שIsa. 5:22) (means) ִלנ ְסוְֹך כר ָ ש ֵ$ .
34. 35. 36. 37.
38.
39. 40. 41. 42.
43. 44. 45.
Sin with gimel: ( ְמסוּכָהcf. Mic. 7:4?) and ( ְמזוּג ָהsic). Sin with zayin: ( ז ְכוִֹכיתJob 28:17) and ( ְסכוִֹכיתsic). ∑ade with †et: עצהand עטה. Sin with sin (i.e. samek): ם עִשּ ֹוּ$( ָשEzek. 23:3) and ִָעים$( ו ְַעסּו ֶֹתם ְרשMal. 3:21). ְסְתּא$( ַאְרתְַּחַשEzra 7:1) and ְשְֹתּא$ְרתְּחַש ַ א ַ (Ezra 4:7). All the words סירוּ ִ ָ ַ( ויּJudg. 10:16) are written with samek in the Scripture except כּי ג ַם ִ ְ ִפן בּ ֶ ֶ חמוֹס כַּגּ ְ ַ ( יJob 15:33) and אבוֹת ָ הם ֶ ֵהם בְּשֹוִּרי מ ֶ ָ( אוֹי לHos. 9:12). ֹסרו ָ ָבּ. ( נ ָסוֹגוּIsa. 42:17) (means) נ ָשו ֹגוּ. סר ֶ ֹ( י ֹאכְלוּ בEzek. 18:2) (mean) שֹר Mem with pe’, when it said ( ִתְּרמְסוּEzek. 34:18) (means) ְרפשוּ ְ תּ ִ (cf. Ezek. 34:18). ( ַמְלָּטהPs. 116:4) and ( ַפְּלָּטהPs. 17:13); ( ַמְלּטוּJer. 48:6) and ( ַפְּלּטוּPs. 82:4). Qof with Òade: ( ָמֲחָקהJudg. 5:26) and ( ָמֲחָצהJudg. 5:26). ( ַבַּצָּלַּחתProv. 19:24) and ( ַבַקַּלַּחת1Samuel 2:14). ∑ade with sin: ( צוְֹרָריPs. 6:8) and וְֹרָרי$( שPs. 5:9). Sin with sin: ַקשּ ֹים ִ ַֹקש ְ ְריוֹן$ש ִ ְ ( ו1 Sam. 17:5) and ְָרין$שּ ִ ַ צְָדָקה כּ$בּש ַ ל ְ ִ ַ( ויּIsa. 59:17). Taw with †et: ( י ְַחטוֹףPs. 10:9) and ( י ְַחתּוֹףJob 9:12); ( ְכֶּחֶתףProv. 23:28) (means) כְֶּחֶטף. ( ִהְתעוּGen. 20:13) and ( ִהְטעוּEzek. 13:10). It said תו ֶֹעה (Gen. 37:15) and ( טוֶֹעהcf. Targum?) with the same meaning. Îet with he’: ( ִעיר הֶַהֶרסIsa. 19:18) and ( ָהאוֵֹמר לֶַחֶרסJob 9:7). ְמַהְלִּלים (1Chron. 23:5) and ( ְמַחְלִּלים1 Kgs 1:40), it means praise. ¨Ayn with Ìet: ( ֲעִריִריםLev. 20:20) and ( ֲחֵריִריםJer. 17:6). Taw with res: ֶגן$( ַפְּתֶשEsther 3:14) and ֶֶגן$( ַפְּרשEzra 4:11).
350
LEXICOGRAPHICAL ARRANGEMENT OF MASORETIC MATERIAL
As for the ‘emphatic letters’, I mean the attached letters, for example, when it said: 1.
$לְדרוֹש ִ (1 Sam. 9:9) and בר ָ הַָדּ$לַדְריוֹש ְ (Ezra 10:16) and their meaning is
2. 3.
כי ֵא ָ ת ְ ַפְרס ַ א ֲ (Ezra 4:9) and כי ֵא ָ ְפְרס ַ א ֲ (Ezra 5:6). עת ֵ כ ָ ְ ( וJudg. 13:23) and ענ ֶת ֶ ְ( וּכEzra 4:10). הנזְ ִיק ַ תּ ְ (Ezra 4:13) [61], הזּ ִיק ַ תּ ְ and תַּזּ ִיק. צָדק ָ ֹ( יוEzra 3:2) and צָדק ָ ֹ( י ְהוHaggai 1:1). בֶּדֶּרְך ַ (Gen. 24:27) and הֶדֶּרְך ַ בּ ְ (Neh.9:19). תם ָ ֹלנ ְחו ַ (Exodus 13:21) and תם ָ ֹהנ ְחו ַ ל ְ (Neh.9:19). מוּאל ֵ ְ$לש ִ (1 Sam. 3:9) and למוּאֵל ְ (Prov. 31:1). כם ֶ ת ְ ֹלְראו ַ (Deuteronomy 1:33) and תם ָ ֹלְראו ִ (Exodus 14:13). יע ַ $ש ִ ֹ ( יוIsa. 45:20) and ַיע$ש ִ ֹ( י ְהו1 Sam. 17:47). תָך ְ ֹ( כַּנּ ְלוIsa. 33:1) means תָך ְ ֹבּנּ ְלו. ַ לים ִ ְ מנ ִ (cf. Job 15:29) and לּים ִ מ ִ (Job 6: 26).
the same.
4. 5. 6. 7. 8. 9.
These words can be exchangeable (between them). You say: הֶדר ֶ (Dan. 11:20) and אֶדר ֶ (Mic. 2:8), תְראוּ ִ (Gen. 43:3) and תְּרהוּ ִ (Isa. 44:8), ( ינִ ְָאקוּJob 24:12) and ( ינִ ְָהקוּJob 30:7), ִָעים י ְֶהדּוֹף$( ו ְַהוּ ַת ְרשProv. 10:4) and ָך$( ְבָּכל אַוּ ַת נ ְַפְשDeut. 12:15), and their meaning is the same. תּי ִ ל ְ א ָ ְ ַי אֶג$לבּוּש ְ ַכ ל מ ָ ְ ( וIsa. 63:3) means תּי ִ ל ְ א ָ ְ הג. ִ יע ַ $ש ִ ֹ בִדּי או ְ ַבּית ָדּו ִד ע. ֵ (cf. 2 Sam. 3:18) for יע$הוִֹש. ( ֶאְתַחַבַּר2 Chron. 20:35) is הְִתַחַבּר. ( ַהנּזָ ֵרZech. 7:3) is אִנּזָ ֵר. ( ֶאְתֶכם אוִֹציאcf. Ezek. 11:7) is הוִֹציא. אתי הוֹן ִ צ ָ מ ָ (cf. Hos. 12:9) is אוֹן. ( הוֹי1 Kgs 13:30) is ( או ֹיNum. 21:29). ( ֵהיְך1 Chron. 13:12) and ( ֵאיְךGen. 44:34). ( ֵהָתיוּIsa. 21:14) and ( ֵאָתיוּIsa. 21:12). מון ֹ א ָ ה ָ (Jer. 52:15) and המוֹן ָ ה ֶ (1 Sam. 4:14). כּם ֵ $ש ְ ה ַ (Exod. 8:16) and כּם ֵ $ש ְ א ַ (Jer. 25:3). 2. ( ָתַּאְבִתּיPs. 119:40) and ( ַתֲּאו ִָתיcf. Ps. 38:10?). ִקצֵבי ְ ל ְ (Jon. 2:7) and ִֵקצוי ְ ל. ְ 3. ( ֵכָּההLev. 13:6) and ( גּ ֵָההProv. 17:22). 4. ( ו ַאֲרוֹז ִיםEzek. 27:24) and ( ֲחרוּז ִיםSong 1:10). ( ֲאמוּצִּיםZech. 6:3) is מוּצּים ִ ֲח. 5. ( נוֵֹטָרהSong 1:6) is נוֵֹצָרה. 6. ְטֶתּם$( ַאל פְַּש1 Sam. 27:10) is ְטֶתּם$אן פְַּש. ָ 7. ְִרבני ֵ ָָקה לִּי ע$ש ְ ע ָ (Isa. 38:14) is ָקה$ש ְ ָח. וּ$ חוּשis וּ$עוּש. 8. ( ַאְרָקאJer. 10:11) and ( ַאְרעָאDan. 2:35). 9. ְחו ֹרוֹת$( שSong 5:11) and ( ְצחוֹרוֹתJudg. 5:10). 10. יכם ֶ ת ִ ֹ( וּנ ְפוֹצוcf. Ezek. 11:17?) and תיכֶם ִ ֹוּתפו ֹצו ְ (Jer. 25:34).
1.
These words are pairs with different interpretations, which require that the scholar stop and memorize them, since their meanings are very close. The ancients collected these in the Great Masora and arranged them alphabetically to facilitate their memorization by scholars. 351
LEXICOGRAPHICAL ARRANGEMENT OF MASORETIC MATERIAL
אוֶֹרהin 1 Sam. 20:20 and Job 27:11. אוֹרוֹתin 2 Kgs 4:39 and Isa. 26:19. אגוּר ָ in Prov. 30:1 and Deut. 32:27. מיר ִ א ָ in Hos. 4:7 and Isa. 17:6. פָרת ְ א ֶ in Gen. 48:7 and 1 Chron. 2:19. בּאוִּרים ָ in 1 Sam. 28:6 and Isa. 24:15. לים ִ א ֵ בּ ָ in Exod. 15:11 and Isa. 57:5. ֹ בג ְדו ִ בּ ְ in Gen. 39:12 and Exod. 21:8. בוּלin 1 Kgs 6:38 and Job 40:20. ְקהלוֹת ֵ מ ַ בּ ְ in Num. 33:26 and Ps. 68:27. סּה ָ מּ ַ בּ ַ , one with samek in Deut. 6:16 and another with sin in 1 Chron.
15:22. בְּראוֹת ִ in 1 Chron. 21:28 and 2 Chron. 26:5. שִֹרים ָ ב ְ in Prov. 14:30 and Prov. 19:10. ִריאה ָ בּ ְ in Num. 16:30 and Hab. 1:16. גּ ָמוּלin Isa. 11:8 and Jer. 48:23. ברוֹת ְ דּ ֹ in Num. 27:7 and 1 Kgs 5:23. דּו ַֹדיin 1 Chron. 27:4 and cf. Song 2:16. בניָ ִם ְ א ָ ה ָ in Exod. 1:16 and Jer. 18:3. ָ $ש תּ ְ ב ַ ֹ הוin 2 Sam. 19:6 and Ps. 74:15. חול ֹ ה ַ in Lev. 10:10 and Prov. 27:3. ִצּפוֹני ְ ה ַ in Num. 26:15 and Joel 2:20. תּנּ ִין ַ ה ַ in Isa. 27:1 and Neh. 2:13. אד ֵ ְ וin Gen. 2:6 and Job 18:12. ואזב. ָרכה ָ וּב ְ in 1 Chron. 12:3 and Prov. 11:26. היא ִ ָ וin cf. Ezek. 2:10 and Ezek. 23:43. בל ֶ ח ֶ ְ וin Joshua 17:14 and Micah 2:10. ו ְחוּצוֹתin 1 Kgs 20:34 and Prov. 8:26. לי ִ ח ֲ ַ וin Josh. 19:25 and Prov. 25:12. לם ֶ ט ֶ ָ וin Josh. 15:24 and Ezra 10:24. ויַּ גָ ֶלin Gen. 29:10 and Ps. 16:9. ָ ויַּ גָ זin Num. 11:31 and Job 1:20. ו ְָדןin Judg. 5:17 and Ezek. 27:19. ויַּ ַָרעin Josh. 6:20 and 1 Kgs 16:25. תּח ַ פ ַ ְ ַ ויin Gen. 24:32 and 1 Kgs 7:36. קּב ֹ ִ ַ ויּin Lev. 24:11 and 2 Kgs 12:10.29 ו ְי ֵָרםin Num. 17:2 and 1 Sam. 2:10. ויַּ ֶֶרבin 2 Sam. 18:8 and Lam. 2:5. 29 Ferrara Romance translation of the Bible renders ‘y horadó’, i.e. ‘to bore through’.
352
LEXICOGRAPHICAL ARRANGEMENT OF MASORETIC MATERIAL
תּח ָ פ ְ ִ ו ְיin Judg. 11:1 and Josh. 15:43. וְּרדוּin Gen. 1:28 and Amos 6:2. $ ו ָֹראשin Gen. 46:21 and Lam. 3:19. והם ַ $ש ֹ ְ וin Exod. 28:20 and 1 Chron. 24:27. כל ַ תּ ְ ַ וin Gen. 24:19 and 2 Sam. 13:39.30 כחוֹל ַ ְ וin Gen. 22:17 and Job 29:18.31 ְרתּי ִ צ ַ ְ וin Exod. 23:22 and Isa. 29:3. ז ְַקןin Gen. 24:2 and Ps. 133:2. זוֹרוּin Isa. 1:6 and Ps. 58:4. זוֹנ ֶהin Hos. 4:15 and Ps. 73:27. $חֶרש ֶ in Josh. 2:1 and 1 Chron. 9:15. חלוּץ ֲ in Numb. 32:27 and Deut. 25:10. לּה ָ ח ִ in Deut. 29:21 and 2 Chron. 33:12. תּם ֶ ְ$חַרש ֲ in Judg. 14:18 and Hos. 10:13. טָרף ָ in Gen. 8:11 and Hos. 6:1. טִריּ ָה ְ in Judg. 15:15 and Isa. 1:6. ק$שּ ַ ִ יin Gen. 41:40 and cf. Prov. 24:26. ְך$שּ ָ ִ יin Deut. 23:20 and Prov. 23:32. נּ ָה$ש ֶ יו ִֹריin Numbers. 14:24 and Zech. 9:4. תם ָ כּ ְ in Isa. 18:5 and Ps. 78:72. אִרי ֲ כּ ָ in Isa. 38:13 and Ps. 22:17. אט ַ ל ְ in 2 Sam. 18:5 and cf. 2 Sam. 19:5. מד ֹ ל ָ in Jer. 12:16 and Zech. 2:6. ענֹּות ַ ל ְ in Num. 30:14 and Ps. 88:1. תנּוֹת ַ ל ְ in Judg. 11:40 and Mal. 1:3.32 ְדביר ִ ל ִ in Josh. 13:26 and 1 Kgs 6:16. מוֹלin Deut. 1:1 and Josh. 5:2. מנ ָה ָ in Num. 23:10 and 1 Sam. 1:5. מנּ ִית ִ in Judg. 11:33 and Ezek. 27:17. לים ִ לּ ְ ח ַ מ ְ in cf. 1 Kgs 1:40 and Mal. 1:12. ָרתה ָ מ ָ in Exod. 15:23 and Jer. 4:17. מִרים ָ in Exod. 15:23 and Dan. 5:19. מגוֵּרי ְ in Gen. 37:1 and Ezek. 21:17. תם ָ ח ָ $ש ְ מ ָ in Exod. 40:15 and Lev. 22:25. מעוֹג ָ in Ps. 35:1633 and 1 Kgs 17:12. 30 Ferrara Romance translation of the Bible renders ‘y desseó (alma)’, i.e. ‘and (the soul) wished’. 31 Ferrara Romance translation of the Bible renders ‘y como Fénix’, i.e. ‘and like Phoenix’. 32 Ferrara Romance translation of the Bible renders ‘culebros’, maybe meaning ‘dragons’ or ‘beast’. 33 Ferrara Romance translation of the Bible renders ‘escarnio’, i.e. ‘ridicule’, ‘de rision’.
353
LEXICOGRAPHICAL ARRANGEMENT OF MASORETIC MATERIAL
עַדנּוֹת ֲ מ ַ in 1Samuel 15:3234 and Job 38:31. ליָך ֶ ב ָ ְ נin Amos 5:23 and Isa. 14:11. סּי ִ ִ נin Exod. 17:15 and Isa. 49:22. צח ַ ֵ נin 1 Sam. 15:29 and Ps. 49:20. ֹ נגִּ ַשin Isa. 53:7 and cf. 1 Sam. 13:6+1 Sam. 14:24. יתי ִ ֵ ענ ֲ ַ נin Ezek. 14:4 and Ps. 119:107. מר ַ ס ָ in Ps. 119:120 and cf. Jer. 51:27. לּה ָ ס ִ in cf. 2 Kgs 12:21 and Lam. 1:15. טנ ָה ְ ֹש ִ in Gen. 26:21 and Ezra 4:6. על ָ in 2 Sam. 23:1 and Hos. 7:16. ענּוֹת ַ in Exod. 32:18 and Isa. 58:5. ֹ פּוֵֹרשin Prov. 29:5 and Lam. 4:4. צרוּרוֹת ְ in Exod. 12:34 and 2 Sam. 20:3. ֵקְרבוּin Hos. 7:6 and Ezek. 36:8. רוֶֹדהin Isa. 14:6 and 1 Kgs 5:4. עה ָ ֹ רוin Gen. 29:9 and Prov. 25:19. ֹ ברו ְ $ש ִ in Gen. 44:2 and Judg. 7:15. וֹנ ִים$ שin Esther 1:7 and Prov. 24:21. פָרה ְ $ש ִ in Exodus 1:15 and Job 26:13. תּכּוּ ֻ in Deut. 33:3 and Isa. 1:5. $חרוֹש ֲ ת ַ in Deut. 22:10 and Prov. 3:29.
Chapter on punctuation [Òere ≠ segol] Something with the form of נ ֶה$ִמְש, ְראה ֶ מ ַ , ש ֹה ֶ ע ֲ מ ַ or שֹה ֶ , being alone, has three points, for example, ( נ ְֶחָמד לְַמְראֶהGen. 2:9) and ( ֶש ֹה כְָשִֹביםDeut. 14:3 sic!). But, if it is in contact with another (element) it has two points, for example, ( ְלַמְראֵה עֵינ ָיוIsa. 11:3)35 and ( ֵשֹה עו ֹלוֶֹתיָךDeut. 14:4). Something with the form of עוֹשֶֹה, being singular, always has three points, like ( ָהי ָה רוֶֹעהGen. 37:2). But, if it refers to many (people) has two points like ( ָכּל רוֵֹעה צֹאןGen. 46:34). (As for) ( ָכּל עֵֹשֹה אֵֶלּהDeut. 18:12), it has two points because it is in contact with ;ֵאֶלּהif there is not ֵאֶלּה, it will have three. Every ִתְּראֶהhas three, except one, ְראה עֲֵשֹה ֵ ִר תּ$ש ֶ א ֲ כ ַ ְ ( וDan. 1:13), since it is a pause form.
1.
2.
3.
34
Ferrara Romance translation of the Bible renders ‘deleites’, i.e. ‘delights’. Both cases, ְראה ֶ מ ַ ְמד ל ָ ח ְ ֶ ( נGen. 2:9) and ְראה עֵינ ָיו ֵ מ ַ ל ְ (Isa. 11,3) are recordered in the list number 13th (Two words that appear twice with the same meaning) of Sefer OÈla we OÈla, (Fernando Díaz Esteban [ed.]), 33 5. 35
354
LEXICOGRAPHICAL ARRANGEMENT OF MASORETIC MATERIAL
4.
5.
6.
Every ִתְּהי ֶהhas one (sic!), except ( ַאל תְִּהי ֵה לִי לְִמִחָתּהJer. 17:17), since it is a pause form. For this reason I have showed the two (points) of תְּהי ֵה. ִ If this will not be the case, it will be like ( ִתְּהי ֶה בָּנוּ לְֵעָדהJosh. 24:27). בּנוּ ָ with dages. Every ַתֲּעֶשֹהis with three (points): except four, ְמָך הַגּ ָדוֹל$וּמה תֲַּעֵשֹה לְִש ַ (Josh. 7:9), ( ִאם תֲַּעֵשֹה עִָמּנוּ ָרעָהGen. 26:29), the sentence of Tamar to Amnon, ( ַאל תֲַּעֵשֹה אֶת הַנּ ְָבָלה הַזֹּאת2 Sam. 13:12) and the sentence of Ismael b. Netania ֶֶקר אַָתּה דו ֵֹבר אַל תֲַּעֵשֹה$( ִכּי שcf. Jer. 40:16). The reason for this is because these are pause forms. Every ending with yod has two points when it is a plural in contact with another (element) like ְ( ַמֲעֵשֹי ייPs. 107:24), ( עוֹשֵֹי2 Chron. 26:13), ( רוֵֹעיEzek. 34:2) and everything related to it.
[qameÒ ≠ pataÌ] Every qameÒ being by itself, when it is fixed, changes to pataÌ as shown by ָדּםand וָֹפר ;ַדּם$ שand וַֹפר$ן ;ש$ ָעָשand ן$ ָחָמס ;ֲעַשand ָחָתן ;ֲחַמסand ;ֲחַתן הרוֹת ָ ְ נnun sounds with the sound of he’ i.e. הרוֹת ֲ ַ מרוֹת ;נ ָ א ֲ (Ps. 12:7) alone and ִאֲמרוֹתin contact; נ ָָהרand נ ְַהר, same sound; ַמֲאָכלand ַמֲאַכל, בּת ָ $ש ַ and בּת ַ ַ$ש. 8. Every he’ that determines (a noun) has qameÒ, for example, ָהִעירand $איש ִ ה ָ . However, if it is followed by Ìet or a dages, it has pataÌ, for example, ַהחַי, $ַהֹחֶדש, ַהֶחֶרב, ַהִמְּדבָּר, ַהַמְּמָלָכה, ַהגּוֹי, ַהֶמֶּלְךand ַהָמּי ִם. When the sentence is an exclamatory, the point under he’ is pataÌ, ָראל ֵ ֹש ְ ִ בד י ֶ ע ֶ ה ַ (Jer. 2:14) and ִ( ַהֱאל ִֹהים ָאני2 Kgs 5:7). Its meaning is ‘indeed’ and similar. 9. ( ֲחָמת צוָֹבה2 Chron. 8:3), absolute like צוָֹבה, and ֲחַמתin construct. כּוָֹכב and כּוַֹכב, י ָםand י ַם. 10. Paseq has a relevant meaning, because it breaks the speech like ם$ֵש ד$ש ַ כ ְ ְרפּ ַ א ַ (1 Chron. 1:24). If there is not a paseq between ם$ש ֵ and ד$ש ַ כ ְ ְרפּ ַ א ַ , we could say that ם$ש ֵ is ד$ש ַ כ ְ ְרפּ ַ א ַ . Something similar happens with qameÒ and pataÌ, ( ָדַּרְך כּוָֹכבNum. 24:17) and יכם ֶ ה ֵ ֹ כב אֱל ַ ֹ( כּוAmos 5:26) and their relatives. 11. Siman: ַאכין, ַבּתין, גּ ַמין, ַרקין, ַתַהרייםand ְקנ ַזיים. All with pataÌ, except ֹבהוּ with qameÒ. 7.
[Miscellaneous] 12. ( ַמזּ ֶהExod. 4:2) is ז ֶה, with three (points). יהם ֶ ל ֵ ֲהזּ ֵה ע ַ (Num. 8:7) has two. Every [ ֶעדwith three], ו ְֵעדfrom ֵעדwith two. Every ְמַצוּ ֶהhas three, except ( נגָ ִיד וּמְַצוּ ֵהIsa. 55:4) with two. ֵבּןabsolute, ֶבּןin construct except one, i.e. ( ִבּן נוּןExod. 33:11), with one point being nun brigade. Likewise ֵלב, if the accent is on the first letter of the following word (it will have three points) like ֶלב-( ָרעProv. 25:20), but if the accent is far from it (i.e. from the first letter of the following word), or even on another letter, it will have two like ויַּ ְַך לֵב-( ָדּו ִד1 Sam. 24:5 +). 355
LEXICOGRAPHICAL ARRANGEMENT OF MASORETIC MATERIAL
13. Every qameÒ followed by lene and without a lene before it is qameÒ. 14. Every dages with a pataÌ before it and followed by no pataÌ is dages. This is so because of the reduplication of the following (element), except the four guttural letters after the accent, like תֲַּעשֹוּ-( ַמהJudg. 18:14 +).
It concludes what we wanted to clarify with the aid of God, of Him and of the kindness of His aid. He is the Highest one, the Praised one worthy of recognition always. Edition of the lists transmitted by the manuscript Firk. I 4557 (folios 4–9) Abbreviations ב / + ; · […] []ס"ב
= Firk. I 2408 (ff. 6r 6v) instead of addition omission the variant ends lines in the original manuscript gap Firk. I 2408 ends
[20–16:א4] אלקול פי חרוף אלאבדאל נקל מן אלמאסרה אלחרף אלדי יתבדל באלאכר פהו אלזין פי מוצע אלצדי זעיר ויצעק. שם זעיר שם תקול בדלה צעיר אני לימים . עלס עלז. ואלסין באלזין נעלסה נעלזה. ויזעק
20
אלו האותיות דאית.ב' ולה איצא אלקול פי חרוף אלאבדאל נקל מן · אלמאסרה .16 · אלקול פי חרוף אלאבדאל/ מפק · אות ומעיי לאות דאצון דמאן להדרי · ואילין הינון ב' ותקול פי.19 ; כקולך+ ב' באלחרף \ באלאכר ;ב' אלצדי.18 ;נקל מן אלמאסרה ; ]כנף[ רננים+ ב' באלזין.20 ; תקול בדלה/ בדלה 356
LEXICOGRAPHICAL ARRANGEMENT OF MASORETIC MATERIAL
]4ב[ אלצאד באלסין נתצו נתסו .אליא באלאלף בלויי הסחבות בלואי הסחבות .אלגימל באלקאף שורג שורק .שרוקיה השריגים. נמקו נמגו .ואלחית באלעין סרוח ושרוע .ואלצדי באלטית נקטה נפשי נקצה נפשי .אקוט אקוץ .ואלרא באלדאל להתגדד להתגדר .ואלרא באללאם אלמנותיו ארמנותיו .להצהיר 5 להצהיל .ואלנון באלרא נוכל יריד רוכל משתק מן רכיל .כידון לכידור .נבוכדראצר נבוכדנצר .ויתנכלו יריד ויתרכלו. מה ברי יריד בני .ואלפא באלבא גפי גבי .פזר בזר .ואל מים באלפא זעם יי זעף יי .ואלאלף באלעין מתאב מתעב. 10וקיל אן אבחת חרב יריד בה אבעת חרב .אלחא באלעין. ואלשין באלצאד לנתש ולנתוץ .שוררי צוררי .ואלכאף באלגימל כובע גובע .ואלדאל באלזין דמוני זמוני .ידעף ויזעף .ואלהא באלאלף האמון ההמון .השכם ודבר אשכם ודבר .ואליא באלאלף ההגריאים ההגריים .הערביאים הערביים. 15ואלנון באללאם נשכה ולשכה .נתע ולתע .וקיל אן כי היה דבר המלך נחוץ יריד בה לחוץ .ואלאלף באלעין לחם מגואל יריד בה מגועל .אשר לא יתגאל לא יתגעל. ואמרתם במה גאלנוך יריד בה געלנוך .ויגואלו מן הכהונה יריד בה ויגועלו .כל דבר שכתוב באלף 20ואין פירושו באלף כתקנו מפרשין אותו בעין כגון מתאב מתעב וכן כל דבר שהוא בעין ואין פירושו בעין
.1ב' ואלצאד /אלצאד; ב' ואליא /אליא; .2ב' ואל]גימל[ /אלגימל; .6ב' כידור לכידון ;כידון לכידור; .7ב' יריד; .9ב' באלפא +כק'; ב' מתאב +אנכי; .10ב' בהא /בה; .12ב' דמותי זמותי /דמוני זמוני; ] .16ס"ב[ וקיל אן כי היה; 357
LEXICOGRAPHICAL ARRANGEMENT OF MASORETIC MATERIAL
]5א[ כתקנו מפרשין אותו באלף כגון ותתפלל על יי יריד אל יי .אל הנער הזה התפללתי .אל אם הדרך יריד על אם. השמת לבך אל עבדי איוב יריד על מתל אלתאני ופי לגה אלמשנה גמיעה יריד גמיאה מן הגמיאיני כקול החושש בשניו אל יגמע בהן חומץ .אליה 5 יבואו רועים ועדריהם יריד עליה .וקיל את השוק והעליה יריד בה והאליה .ואלבית באלפא נשפת ברוחך יריד נשבת .ישב רוחו ונשבה בו .לפני אפרים קיל יריד בה לבני .וכדלך קול רך ויחיד לפני אמי לבני 10אמי .פרזלא ברזלא .ואלגימל באלכאף ומגבעות מן כובע נחשת .ואלכאף באלגימל ויסכרו מעינות תהום יריד ויסגרו .וסכרתי וסגרתי .לארץ מכורתם יריד מגורותם .ואלדאל באלצאד הפודה הפוצה .פדני פצני. ואלרא באלדאל כקול הדר הרר .דודנים רודנים .חמדן 15ואשבן חמרן ואשבן .וריפת ודיפת .ואלהא באלחית הצו אשר חצו .בהצותם בחצותם .ואלואו באליא עלון ומנחת עלין ומנחת .ועקן יעקן .שפו ואונם שפי ואונם .שם עירו פעו שם עירו פעי .מגדיל ישועות מגדול ישועות .ואלזאי באלצדי זוהר צוהר .בין שורותם 20יצהירו ויריד יזהירו .שער צהוב זהוב .ופי תורת כהנים יקול צהוב למה הוא דומה לתבנית זהב .ואלטית באלתו ]5ב[ פסנתרין פסנטרין .כדבר אפרים רתת ורטט החזיקה. ופי לגה אלמשנה אלנתית ואלנטית .ואלחית באלכאף נכשלים ונחשלים .נכתב ונחתם .ואללאם באלרא ואלו וארו .מורות מולות .ואלמים באלנון המוט תנוט נורא מורא .גם בני נוף מוף תקברם .למסוך שכר 5 לנסוך שכר .ואלסין באלגימל מסכה מזוגה .ואלסין באלזין זכוכית סכוכית .ואלצאד באלטית עצה עטה. ואלסין באלשין שם עשו ועסותם רשעים .ארתחשסתא ארתחששתא .וכל לשון ויסירו נכתב בסמך ב'מ'א' 10כי גם אוי להם בשורי מהם .יחמוס כגפן בסרו .אבות יאכלו בסר בשר .נסוגו נשוגו .ואלמים באלפא כקו' תרמסו תרפשו .מלטה פלטה .מלטו פלטו .ואלקוף באלצדי מחקה מחצה .בצלחת בקלחת .ואלצאד באל שין צוררי שוררי .ואלשין באלשין ושריון קשקשים 15וילבש צדקה כשרין .ואלתא באלטית יחטוף יחתוף. 358
LEXICOGRAPHICAL ARRANGEMENT OF MASORETIC MATERIAL
כחתף כחטף .התעו הטעו .ויקאל תועה וטועה ואלמעני ואחד .ואלחא באלהא עיר ההרס האומר לחרס .מהללים מחללים לשון שבח .ואלעין באלחת ערירים חרירים. ואלתא באלרא פתשגן פרשגן .ומן חרוף אלתפכים אעני 20אלחרוף אלזאידה כקו' לדרוש לְַדְריו ֹש $הדבר ואלמעני ואחד .אפרסתכיא אפרסכיא .וכעת וכענת .תהנזיק ]6א[ תהזיק תזיק .יוצדק יהוצדק .בדרך בהדרך .לנחותם להנחותם. לשמואל למואל .לראותכם לראותם .יושיע יהושיע .כנלותך יריד בנלותך .מנלים מלים .והדה אלאלפאת אלמבדלה ת]קו[ל אדר הדר .תראו תרהו .ינאקו ינהקו .והות רשעים יהדוף רק בכל אות נפשך ואלמעני ואחד .וכל מלבושי 5 אגאלתי יריד הגאלתי .ובית דוד עבדי אושיע הושיע. אתחבר התחבר .הנזר אנזר .ואתכם אוציא הוציא. מצאתי הון און .הוי אוי .היך איך .התיו אתיו .האמון ההמון .השכם אשכם .תאבתי תאותי .לקצבי לקצוי .כהה גהה .וארוזים 10וחרוזים .אמוצים חמוצים .נוטרה נוצרה .אל פשטתם אן פשטתם .עשקה לי ערבני חשקה .חושו עושו .ארקא ארעא .שחורות צחורות .ונפוצותיכם ותפוצותיכם. והדה אלאלפאט מזדוגה 15מכתלפה אלעבארה אלדי יגב עלי אלמתעלם אלוקוף עליהא וחפטהא פאנהא קריבה אלמעני וקד דונוהא אלאואיל פי אלמסורה אלכבירה ורתבוהא עלי אלף בית ליסהל עלי אלטאלב חפטהא .אורה .ואני שלשת החצים צדה .אורה אתכם ביד אל .אורות .ויצא אחד 20אל השדה .אורות יחיו מתיך נבלתי .אגור .דברי אגור בן יקה .אגור .לולי כעס אויב אגור .אמיר ]6ב[ כרבם כן חטאו לי .אמיר .ונשאר כן עוללות .אפרת ב' ואני בבואי מפדן .ותמת עזובה .באורים ב' וישאל שאול באלהים .על כן באורים .באלים ב' מי כמוכה. הנחמים באלים .בבגדו ב' ותתפשהו בבגדו .אם[ רעה[ בעיני .בול ב' ובשנה האחת .כי בול הרים .במקהלות 5 ב' ויחנו במקהלות .במקהלות ברכו אלהים .במסה ב' חד ס' לא תנסו .וחד ש' וכנניהו שר במשה .בראות ב' בעת ההיא בראות דויד .ויהי לדרוש .בשרים ב' חיי בשרים לב .לא נאוה לכסיל תענוג .בריאה ב' 359
LEXICOGRAPHICAL ARRANGEMENT OF MASORETIC MATERIAL
10ואם בריאה יברא .ומאכלו בריאה .גמול ב' ושעשע יונק על חור פתן .ועל קריתים גמול .דברות ב' כן בנות צלפחד .עבדי ירדו מן .דודי ב' ועל מחלוקת החדש השני .דודי לי ואני לו .האבנים ב' וראיתן .וארד אל בית היוצר .הובשת ב' ויבא יואב אל .אתה בקעת 15מעין ונחל .החול ב' ולהבדיל בין הקדש .כובד אבן. הצפוני ב' בני גד למשפחותם .ואת הצפוני ארחיק מע'. התנין ב' ביום ההוא יפקד יי .ואצאה בשער הגיא. ואד ב' יעלה .יהי רעב אונו .ואזב ב' .וברכה ב' הראש אחיעזר .וברכה לראש משביר .והיא ב' ויפרש בל ב' וח ֶ 20אותה לפני והיא .ואומר לבלה ניאופים ֶ . וידברו בני יוסף אל יהושע .כי לא זאת המנוחה. ]7א[ וחוצות ב' ויאמרו אליו הערים אשר .עד לא עשה ארץ וחוצות .וחלי ב' ויהי גבולם חלקת .נזם זהב וחלי כתם .וטלם ב' זיף וטלם ובעלות .ומן המשוררים אלישיב .ויגל ב' את האבן מעל פי .לכן שמח לבי .ויגז ב' שלוים .ויקם איוב ויקרע את מעילו. 5 ודן ב' גלעד בעבר הירדן .ודן ויון מאוזל בעזבוניך. וירע ב' וירע העם ויתקעו .וירע ויעשה עמרי הרע. ויפתח ב' את הגמלים .ויפתח על הלוחות ידותיה .ויקב ב' בן האשה .חור בדלתו .ו ְי ֵָרם ב' את המחתות .קרן משיחו. 10וירב ב' ותהי שם המלחמה .וירב בבת יהודה תאניה. ויפתח ב' הגלעדי .ויפתח ואשנה ונציב .ורדו ב' בדגת הים .עברו כלנה וראו .וראש ב' ובני בנימן בלע .זכר עניי ומרודי .ושוהם ב' והטור הרביעי תרשיש .בני מררי ליעזיהו בנו ושוהם .ותכל ב' 15להשקותו .ותכל דויד המלך .וכחול ב' כי ברך אברכך .ואומר עם קני אגוע וכחול .וצרתי ב' את צרריך .וחניתי כדור עליך .זקן ב' עבדו זקן ביתו .זקן אהרן שיורד .זורו ב' לא זורו ולא חבשו. זורו רשעים מרחם .זונה אם זונה אתה ישראל .הצמתה 20כל זונה ממך .חרש $ב' וישלח יהושע בן נון .ובקבקר חרש .חלוץ ועבדיך יעברו .ונקרא שמו בישראל. לה ב' ואמר הדור האחרון .וכהצר לו חלה את פני יי. ח ָ ]7ב[ ִ חרשתם ב' ויאמרו לו אנשי העיר .חרשתם רשע עולתה קצרתם .טרף ב' והנה עלה זית .כי הוא טרף וירפאנו. 360
LEXICOGRAPHICAL ARRANGEMENT OF MASORETIC MATERIAL
טריה ב' וימצא לחי חמור טריה .פצע וחבורה ומכה. ישק ב' כל עמי .שפתים ישק .ישך ב' נשך כל 5 דבר .אחריתו כנחש ישך .יורישנה ב' ועבדי כלב תם ב' כי לפני קציר כ ָ עקב היתה .הנה יי יורישנה ְ . כתם פרח .וירעם כתם לבבו .כארי ב' שויתי עד לאט ב' ויצו המלך את יואב. בקר .כי סבבוני כלבים ָ . 10והמלך לאט את פניו .למד ב' והיה אם למד ילמדו .ואמר אנה אתה הולך .לענות ב' כל נדר וכל שבועת אסר .לענות .על מחלת לענות .לתנות ב' מימים ימימה תלכנה בנות ישראל .ואת עשו שנאתי. לדביר ב' ומחשבון עד רמת המד' .ויבן עשרים 15אמה .מול ב' אלה הדברים אשר דבר משה .בעת ההיא אמר יי אל יהושע .מנה ב' מי מנה עפר יעקב .ולחנה יתן מנה אחת .מנית ב' ויכם מערוער ועד בואך מנית .יהודה וישראל המה. מחללים ב' ויעלו כל הארץ חדיו .ואתם מחללים 20אותו .מרתה ב' ויבאו מרתה .כשומרי שדיה. מרים ב' כי מרים הם .ודי הוה צבא הוה מרים. ]8א[ מגורי ב' בארץ מגורי אביו .זעק והילל בן אדם .משחתם ב' ומשחת אותם כאשר משחת .וזאת עשו להם וחיו ולא ימותו. מעוג ב' בחנפי לעגי מעוג .ותאמר חי יי אלהיך אם יש לי .מעדנות ב' אגג מעדנות .התקשר מעדנות כימה. נבליך ב' הסר מעלי המון שריך .הורד שאול גאונך 5 המית נבליך .נסי ב' יי נסי .ארים נסי .נצח ב' וגם נצח ישראל .עד נצח לא יראו אור .נגש ֹ ב' והוא נענה. כי איש איש מבית ישראל .נעניתי ב' לכן דבר אותם ואמרת אליהם .נעניתי עד מאד .סמר ב' מפחדך לה ב' ויקומו עבדיו ויקשרו ס ָ 10בשרי .שאו נס בארץ ִ . קשר .סלה כל אבירי .שטנה ב' ויקרא שמה שטנה. על ב' נאם הגבר הוקם על. ובמלכות אחשורושָ . ישובו לא על .ענות ב' קול ענות אנכי שומע .יום ענות אדם נפשו .פוֵרש ֹ גבר מחליק על רעהו .דבק 15לשון יונק אל חכו .צרורות ב' וישא העם את בצקו. ותהיין צרורות עד יום מותן .קרבו ב' כי קרבו כתנור לבם .ואתם הרי בית ישראל .רודה ב' רודה באף גוים. רועה ב' עודנו מדבר עמם ָ רודה בכל עבר הנהר . 361
LEXICOGRAPHICAL ARRANGEMENT OF MASORETIC MATERIAL
ברוֹ ואת גביעי שְ $ ורחל באה .שן רועה ורגל מועדת ִ . 20גביע הכסף[ .ויהי[ כשמע גדעון את מספר החלום. שונים ב' וכלים מכלים שונים .ירא את יי בני ומלך. ]8ב[ שפרה ב' ויאמר מלך מצרים למילדת .ברוחו שמים תּכוּ ב' והם תכו לרגלך .על מה תכו .תחרוש שפרה ֻ . ב' בשור ובחמור .אל תחרוש על רעך וגו'. אלכלאם פי אלנקט 5 מא כאן עלי וזן משנה או מראה או מעשה או שה וכאן מקטועא פנקטה ג' נקט .מתל נחמד למראה. שה כשבים .ואן כאן מוצולא אלי גירה כאן בנקטתין מתל למראה עיניו .שֵה עולותיך .ומא כאן עלי וזן עושה ֵ 10וכאן ואחדא כאן בג' אבדא מתל היה רועֶה .ואן כאן כתירא כאן בנקטתין מתל כל רועֵה צאן .כל עשֵ ֹה אלה. בתנתין אד כאנת מוצולה באלה .ולולא אלה כאנת בג'. וכל תראה בג' אלא ואחדה וכאשר תראֵה עשה .לאנהא בלשון תחנה .וכל תהיה בואחדה אלא אל תהי ֵה לי למחתה. פ Óית לי לתקדם אלתנתין 15לאנהא איצא בלשון תחנה .ולדלך ר Ô פי תהיה .ולולא דלך כאן מתל תהיה בנו לעדה .בנו מדגושה .וכל תעשה בג' גיר ארבעה .ומה תעשֵה לשמך הגדול .אם תעשֵֹה עמנו רעה .וקול תמר אלאמנון אל תעשה את הנבלה הזאת .ופי ישמעאל 20בן נתניה כי שקר אתה דובר אל תעשה .ואלעלה פי דלך אנהא בלשון תחנה .וכל מא כאן אכרה יוד בתנתין אד ]9א[ כאן גמיעה מנסובא אלי גירה מתל מעשי יי .עושי רועי .ונטיר דלך .וכל מקמוץ קאים בנפסה אדא שן אציף צאר בפתח גדל נטיר ָדם ַדם שופָר שופר עָ ָ מס חֲמס חתן חתן .נהרות אלנון במכרג אלהא עשן חָ ָ ֲ הר מרות מוצול .נ ָ ָ תכרג נהרות .אמָרות מקטוע .אִ ְ 5 מאכל .שבת שבת. ָ נ ְהר .אלמכרג ואחד .מאכָל וכל הא גאת ללמערפה מקמוצה מתל העיר האיש אלא אן כאן בעדהא חית או דגש פאנהא תפתח מתל החי החדש החרב המדבר הממלכה הגוי המלך המים. 10ואן כאן אלקול תמוה כאן אלנקט תחת אלהא פתח מתל העבד ישראל .האלהים אני .מענאה הל .ונטיר דלך. מה חמת צובָה .מוכרת .כמו וצובה חמת סמוך .חֵ ָ ֲ 362
LEXICOGRAPHICAL ARRANGEMENT OF MASORETIC MATERIAL
חמת .כוכָב כוכב י ָם ים .וללפסק מעני גליל והו ֲ אנה יפרק בין אלכלאם מתל ושם ארפכשד .לולא אל 15פסק בין שם ובין ארפכשד קלנא אן שם הו ארפכשד וכדלך אלקול פי אלקמץ ואלפתח דרך כוכָב. כוכב אלהיכם .ומא שאכלה .סימן אכין בתין גמין רקין תהריים קנזיים כלהם פתוחין ולית בהו קמץ. עד ]עד בג'[ .ו ָ ֵ מז ֶה הו ז ֶה בג' .הז ֵה עליהם בב' .וכל ֶ בן 20מן עֵד בב' .וכל מצו ֶה בג' .אלא ]נג[יד ומצו ֵה בב' ֵ . 36 נכרת בֶן נסמך .ב'מ'א' בן נון בנקטה לאנצגאט ]9ב[ אלנון .וכדלך ל]ב אן ג[נס אלטעם אליה פי אלחרף אל]אול[ 37 מן אלכלמה אלתי תליה מתל ולֶב ָרע .ואן הו בעד ]מנה[ ולו בחרף כאן בתנתין מתל ויך לֵב דוד .כל קמץ בעדה רפי וליס כל רפי קבלה קמץ .כל דגש קבלה פתח וליס כל פתח בעדה דגש .ואלעלה פי אן מא תדגש 5 מא בעד]ה[ אלא א'ה'ח'ע' בעד אלעטם מתל מה תעשו. כמל מא ארדנא אתבארה בעון אללה ומנה וחסן עונה והו תע' אלמחמוד אלמשכור דאימא. 10
(being assimilated).לאנצגאם Also and on theמנה There are two possibilities to fill this lacuna. On the one hand , depending on if the sentence is understood as verbal or nominal, i.e. ifדלך other is taken as a past verb or as a particule. In this case I prefer to take itבעד the word as a verb, but both readings are correct. I am deeply indebted to Professor Aharon Maman from the Hebrew University of Jerusalem who helped me to fill these lacunae in the manuscript. 36 37
363
THE ALTERNATION U / O, DIPHTHONGS, PATAÎ FURTIVUM AND THE 3M. S. PRONOMINAL SUFFIX IN SAMARITAN HEBREW AND ARAMAIC VERSUS TIBERIAN HEBREW MOSHE FLORENTIN TEL AVIV UNIVERSITY
Abstract Samaritan Hebrew and Aramaic, though sharing the same phonetic structure, differ mainly in the choice of options available in the gen eral Samaritan pronunciation. In the present study we have argued that the use of the auxiliary vowel in general, and specifically the use of the pataÌ furtivum was relatively widespread in Samaritan Hebrew and relatively rare in Samaritan Aramaic. Only this fact caused the large distribution of the vowel u in an open post tonic syllable in SH versus the clear preference of the vowel o in this position in Samaritan Aramaic. Morpho phonemic alongside historical analysis was used thor oughly in this short study. It helped us to prove that: (1) the vowel o ָ suso is of the 3m. s. attached to a plural noun in SH e.g. סוּסיו genuine in this language, being underlyingly a closed syllable; (2) the reason for the retention of the diphthong åw in the same pronominal is purely phonetic, suffix in Tiberian Hebrew e.g. סוּסָיוsusåw i.e. the long vowel before the semi vowel (*susaw).
* My colleague and friend Dr H. Mutzafi read the paper and supplied me with many beneficial remarks and data. Dr M. Morgenstern and Ms Keren Ben Shushan read the first draft and significantly contributed to its language style. In this paper I use the following abbreviations: DSA = A. Tal, A Dictionary of Samaritan Aramaic, Leiden 2000; GSH = Z. Ben Îayyim, A Grammar of Samari tan Hebrew (Jerusalem Winona Lake 2000); LSH = Late Samaritan Hebrew A Linguistic Analysis of its Different Types (Leiden Boston 2005); RPH = Z. Ben Îayyim, The Literary and Oral Tradition of Hebrew and Aramaic Amongst the Samaritans (in Hebrew), Vol. III, Part 2. (Jerusalem 1967); SA = Samaritan Aramaic; SE = Samaritan Elegies; SH = Samaritan Hebrew; SP = Samaritan Penta teuch; ST = A. Tal, The Samaritan Targum to the Pentateuch A Critical Edition, 3 vols. (Tel Aviv 1980 3); TH = Tiberian Hebrew; TM = Z. Ben Îayyim, Tibåt Mårqe A Collection of Samaritan Midrashim (in Hebrew) (Jerusalem 1988). 365
THE ALTERNATION U / O, DIPHTHONGS
1. The trigger for my study of the issues mentioned in the title of this paper was the discovery of a different spelling – — סרעיוof the Samaritan Hebrew (henceforth: SH) and Aramaic (henceforth: SA) word ‘ סרוחevil-doer’, for which the pronunciation saru is attested both in the recitation of the Samaritan Pentateuch (henceforth: SP) and in the Aramaic poems (see below). The corpus in which סרעיוwas found is also newly discovered: a Samaritan manuscript containing about sixty Hebrew and Aramaic elegies and poems which I found in the National Library of Russia1 in the summer of 2000. The first stanza of the first poem2 in SE is: אה געוז עלמה במה את מתרברב ואתה סרעיו חטאי וכל עובדיך בישים
Oh, you, who pass from the world about what do you boast, when you are an evil-doer, a sinner and all your deeds are bad?
The word סרעיוitself is already attested in a fourth century CE poem composed by the Samaritan ¨Åmråm DÇre: ברקה הוא חרבך דעתיד ‘ מוקד כל סרוחYour sword is the lightning which will burn every sinner’ (RPH, 47). Thus, סרוחand סרעיוare seemingly two orthographical variants of the same word: a participle Qal of the pattern qatol derived from the root סר"חwhose basic meaning is ‘stink’, thus ‘sin’, ‘guilt’ (see DSA, s.v. )סר"ח. While עinstead of an etymological חis well attested in Samaritan writings,3 one should pay attention to the orthographic representation of the last vowel of the word in question. סרוחin the ancient Samaritan poem quoted above is pronounced saru. However, the spelling סרעיו, ending with the letters –יו, cannot suit the vowel u, since the combination –יוin Samaritan writings reflects only the vowel o. Its roots as the marker of o are in the 3m. s. pronominal suffix attached to the plural forms of nouns in 1 The whole text will be published in Hebrew in the near future under the name מספד שומרוןSamaritan Elegies (henceforth: SE). 2 The whole poem of the genre אוכחוadmonition was published by me: ‘An Unknown Samaritan Elegy’, Leshonenu 67 (2005), 167 78. 3 Many occurrences of the verbs derived from the root ‘ חכ"םknow’ are written in ( עsee DSA, s.v. )חכם. This is, as is well known, due to the almost complete loss of the original gutturals in SH and SA. For the rules of the gutturals in the Samari tan pronunciation see GSH, §1.1.8. For examples of the different ways of repre senting gutturals in Samaritan writings, see, e.g. ST, Vol. III, 133. For the same phenomenon in the Isaiah Scroll, see, E.Y. Kutscher, The Language and Linguistic Background of the Isaiah Scroll (1Q1saa) (Leiden 1974, 507).
366
THE ALTERNATION U / O, DIPHTHONGS
SH,4 e.g. סוּסָיוsuso ‘his horses’ and in SA as well, e.g. פעליוfa}elo ‘his deeds’ (RPH, 74).5 Thus, the letters –יוbegan to mark the vowel o even in cases in which they had no etymological background, e.g. ¨ עמיוamo ‘one who sees’: ‘ כסי ועמי כל רז לא כות כל עמיוHe is hidden, but sees every secret, not as any seer’ (RPH, 101), פתחיוfato ‘one who opens’: (‘ פתחיו תרחי טובךYou) who open the gates of Your goodness’ (RPH, 93). Like סרעיו, עמיוand פתחיוare also Qal participles of the pattern qatol, which have no –יוin their etymological roots. Now, one should ask what is the genuine form: the existing recited saru spelled סרוח, or the one found in our manuscript, spelled סרעיוand doubtlessly reflecting the pronunciation *saro? Or, alternatively, are we dealing here with two different patterns of the same word? The dilemma is even more acute when one takes into consideration the fact that the vowel u in the Samaritan pronunciation stands normally in open syllables while o stands in closed syllables, i.e. these two vowels occur in complementary distribution being two allophones of a single phoneme, e.g. yeqom ‘ י ָקוּםhe stands’ / yequmu ‘ י ָקוּמוּthey stand’.6 As a result of this phonological rule, quite a few homophones appeared in SH and SA, e.g. sor = ‘ שורox’ = ‘ = שוּרwall’. This rule also affected the grammatical structure, neutralizing the difference between patterns such as ָקטוּל, participle passive of Qal, and ָקטולnominal and active participle of Qal, e.g. kÇtob = ‘it is written’ (‘ = )ָכּתוּבhe writes’ ()ָכּתוב.7 2. Thus, against this clear rule, we see the above mentioned pieces of evidence drawn from both the oral and written traditions: סוּסיו ָ suso, and { עמיוamo. To these two categories — the pronominal suffix and the pattern qatol in verbs III-Yod ()ל"י8 — we may add additional categories in which one finds o in an open syllable: (1) Participle passive Qal of the pattern ָקטוּלof verbs III-Yod in SH, e.g. tÇlo ָתּלוּי This orthographic rule was formulated by Ben Îayyim (see, e.g. TM, 17). Almost all diphthongs in the Samaritan pronunciation were contracted (GSH, §1.4.4.1). 6 See GSH, §1.2.0 and 1.2.4. 7 For these two patterns in SH, see GSH, §2.13.3; for the use and distribution of the pattern qatol in ancient and late Samaritan texts, see LSH, 221; for the merging of the two patterns ָקטוּלand ָקטוֹל, see A. Tal, ‘Observations on Word For mation in Samaritan Aramaic II: the ָקטוֹלpattern’, in M. Bar Asher (ed.) Studies in Mishnaic Hebrew, Scripta Hierosylimitana 37 (Jerusalem 1998), 349 64. 8 One may add to this category the following instances: bÇro ‘ ברויcreator’ (RPH, 101) nÇÒo ‘ נצוחvictorious’ (RPH, 183). 4 5
367
THE ALTERNATION U / O, DIPHTHONGS
‘hanged’ (Deut. 21:23), wgÇlo ‘ ו ְג ָלוּיopened’ (Num. 24:4, 16); (2) The pattern qatol in verbs III-Guttural in SA, e.g. fato (‘ פתחיוyou) open’ (RPH, 93), samo (‘ שמעיוyou) hear’ (RPH, 104); (3) The pattern qittul of nouns III-Yod, e.g.: sirro ‘ שריוbeginning’ (RPH, 101); zizzo ‘ זעזועtremor’ (RPH, 195). To the category of qittul > qitto one may add the written evidence of the ST, such as * נסיוnisso ‘test’, תניו *tinno ‘telling’, etc. In all other cases in SH and SA, u is found in an open post-tonic syllable, in accordance with the phonological rule: (1) In SH: susu ‘ סוּסוhis horse’ (representing the 3m. s. suffix attached to a s. noun); u ‘ אוor’; Ç’u ‘ ָאחוּreed-grass’; Çfu ‘ ֵאפואthen’ (Gen. 27:33); Çqu ַאקּו ‘wild goat’ (Deut. 14:5); (2) In SA: milgu ‘ מלגוout of ’ (RPH, 64); subÇru ‘ סוברוforgiveness’ (RPH, 47; representing all abstract nouns ending with u); wsabba’u ‘ ושבחוand praise!’ (RPH, 57). What causes then the difference between the ‘normal’ group represented by סוּסוsusu and having u in the final open post-tonic syllable, and the ‘aberrant’ group represented by סוּסיו ָ suso and tÇlo ָתּלוּיand having o in the same position? The answer is that actually the ‘aberrant’ group does not contradict the rule of distribution of u and o if we look at syllables morpho-phonemically.9 The final syllable of the Hebrew and Aramaic words suso (< *susaw), tÇlo (< *taluy), ¨amo (< *¨amoy), fÇto (< *fatoy) is phonetically open, but it is both historically and synchronically closed. In other words, forms like tÇlo and fÇto are underlyingly taloC and fatoC (C stands for consonant), and are automatically and subconsciously compared by the speaker to other forms in which the syllable is phonetically closed such as kÇtob, the plural tÇluww¢m (compare: nequww¢m ‘ נקיים = נקויםclean’, Gen. 44:10) and the variant optional form tÇluwwi (compare: rÇÒuwwi ‘ רצויfavourite’, Deut. 33: 24).10 One may consider all these syllables virtually closed. 9 See my article: ‘Distinction between Various Meanings and their Phonological Notation in Samaritan Hebrew’, in A. Dotan and A. Tal (eds), Te¨uda 11, Studies in Hebrew Language in Memory of Eliezer Rubinstein (Tel Aviv 1995), 107 16, and the reference in p. 114, n. 29. This important principal, first formulated by Ben Îayyim, was omitted in the Hebrew edition of LOT V, but re appeared in GSH, 44. In my article ‘The Object Suffixes in Samaritan Aramaic and the Ways they are attached to the Verb’, Abr Nahrain 29 (1991), 67 82, I have shown that III Guttural verbs though phonetically ending with an open syllable, e.g. sabba ‘he praised’, still don’t present a buffer consonant which is regularly inserted when the stem ends with a vowel. Thus, ‘ שבחךhe praised you’ sab´bak (RPH, 185) versus ‘ מלתהhe filled it’ mÇlÇte (RPH, 265). 10 For the behaviour of the descending diphthongs and the splitting into two vowels see GSH, §1.4.4.
368
THE ALTERNATION U / O, DIPHTHONGS
The newly discovered spelling סרעיוthus genuinely reflects the Aramaic pronunciation *saro parallel to the attested forms such as fato mentioned above. What then is the reason for the u in saru ?סרוח This word, with the same pronunciation, is attested already in the SP (Exod. 26:13). In SH it is the regular form of the passive participle ָקטוּלof Qal in III-Guttural verbs. Thus, one may find also faru רוּע ַ פּ ָ ‘disordered’ (Lev. 13:45) and fatu ‘open’ (Num. 19:15). If this saru in the Aramaic poem may be regarded as Hebrew influence,11 one may suggest that this kind of influence did not take place in the verbal participles samo and fato, probably because they are different from their Hebrew parallels attested in the Pentateuch such as sama.12 3. If the possibility of a Hebrew grammatical influence in a fourth century CE Aramaic poem is acceptable,13 one still has to explain the retention of u in the above mentioned SH passive forms since according to our rule they all stem from forms in which the last syllable was closed, thus o rather than u was expected: faru < *paru¨; fatu < *patuÌ ; saru < *saruÌ. The same applies to the vowel u in SH forms such as mad´du < *mad´du¨ ַ‘ ַמדּוּעwhy’ and mÇ´nu < *manoÌ ח ַ מנו ָ ‘rest’. In other words, what could cause the difference between mÇ´nu < *manoÌ ָמנוַחand fato < *patoÌ ?*ָפּתוַח The answer to this seemingly inconsistent behaviour of the same phonetic sequences is connected to the behaviour of the auxiliary vowel pataÌ furtivum in the Samaritan tradition. In this respect, two principles must be stressed: (1) The phonetic conditions in which this auxiliary vowel is inserted and the quality of this vowel in Tiberian Hebrew (henceforth: TH) are totally different from those existing in the Samaritan tradition; (2) In Samaritan tradition — much like Tiberian — the scope of the use of the pataÌ furtivum in Hebrew is significantly greater than in Aramaic. As is known, the pataÌ furtivum in TH is an auxiliary vowel a inserted between the (former) long vowels e, o, u and one of the gut11
However, this is not entirely certain (see, below, p. 371). See the patterns of the passive participle of verbs III Gutturals in GSH, p. 410. 13 For further discussion on the Hebrew influence on SA see A. Tal, ‘Languages in Contact; Hebrew and Aramaic in the Ancient Samaritan Community’, in Essays in Honour of G. D. Sixdenier New Samaritan Studies of the Société d’Études Samaritaines III and IV, A.D. Crown and L. Davey (eds), (Sydney 1995), 577 86; ‘Hebrew and Aramaic in Samaritan Literature’, Proceedings of the Israel Academy for Sciences and Humanities 7 (1988), 239 55 (in Hebrew). 12
369
THE ALTERNATION U / O, DIPHTHONGS
turals ע, ח, הat the end of a word. And as already mentioned by BenÎayyim,14 this auxiliary vowel was not of the same nature in all traditions of Hebrew. SH is different from TH in two aspects: (1) The auxiliary vowel is not just a but rather any vowel assimilated to the vowel standing before the guttural, e.g. ‘ ָמנוַחrest’ ma´nu’uÌ > *ma´nu’uÌ > *ma´nuÌ > mÇ´nu; יח ַ פ ִ ס ְ ‘aftergrowth’, *s¢´fi’iÌ > *s¢´fi’iÌ > * se´fiÌ > se´fi; ַ‘ ַתְּצִליחyou succeed’, *taÒli’iÌ > *taÒli’iÌ > *taÒ´liÌ > * taÒ´li; ְרדַּע ֵ פ ַ צ ְ *Òefar´da’a¨ > *Òefar´da’a¨ > *Òefar´da¨ > *Òefar´da; (2) It is inserted even before א, e.g. * ָמבואma´bu’u’ > *ma´bu’u’ > *ma´bu’ > mÇ´bu. In other words, the scope of this phenomenon in SH is much greater than its parallel in TH. This is not surprising if we take into consideration that the use of auxiliary vowels in SH is generally larger than in TH.15 The consequence of this almost entirely obligatory phenomenon in SH was the creation of ultima stressed and extra-long vowels as shown above: mÇ´nu, se´fi, Òefar´da, mÇ´bu. Note that extra-long vowels are preserved both in quality and length even in closed syllables, e.g. ‘ תדעוןyou (plural masc.)’*tidd¢¨un > *tiddu¨un > *tid´du¨un > tid´dun. We can now understand all the SH forms such as mÇ´nu ending in u rather than o though the vowel stood in syllables closed by gutturals. Indeed, the last syllable was formerly closed, i.e. *manoÌ; however in the pre-Samaritan stage, this last vowel became extra-long, and was therefore not subject to change. Thus, in SH we have many such words as ז ְרוַעzÇ´ru ‘arm’, נוּח ַ ָ יyÇ´nu ‘will rest’, ַ† ָטבוּחÇ´bu ‘slaughtered’. However, many SH words ending with a former guttural have an unstressed u in their ultima, e.g. תוּח ַ פּ ָ ´fatu (< *fa´tu) ‘open’, בוּע ַ ש ָ ´sibbu (< *sib´bu) ‘week’. It is quite plausible that all these words — alongside other formerly-ultima-stressed forms such as ִריח ַ בּ ְ ´bÇri (< *bÇ´ri ) ‘bar’, ´ גּ ִָביעgebi (< *ge´bi) ‘cup’ — followed the Samaritan tendency towards pen-ultima stressed automatically losing the extra long length of their final syllable.16 4. Let us now consider the use of this auxiliary vowel in SA. Note that in Biblical Aramaic the scope of the pataÌ furtivum is smaller than in TH: it is not inserted after the vowel e, e.g. ֵלהּleh ‘to him’. SA con14
GSH, §1.3.6. GSH, §1.3. 16 An assumption that these pen ultima stressed forms did not have at any stage the auxiliary vowel (i.e. *fatuÌ > fatu), will prevent a reasonable explanation for the retention of the final u. 15
370
THE ALTERNATION U / O, DIPHTHONGS
tinued to develop in the direction of Biblical Aramaic, going the opposite way from SH. While in SH the distribution of the auxiliary vowel was widespread, in SA it was barely used! Thus, *fatoÌ shifted directly to *fato (> fato) without the lengthening of the final vowel. The infrequency of the pataÌ furtivum in SA is clearly reflected in the non-existence of ultima-stressed forms with an extra-long vowel such as ** שקיעas´qi ‘found’ (only ´asqi שקיחis attested: RPH, 48, 80, 89), ** נביאne´bi ‘prophet’ (only ´nebi, 50), ** קביעqe´bi ‘fixed’ (only ´qebi, 73), ** מגיחmÇ´gi ‘fight’ (only ´mÇgi, 77; cp. SH מזריע maz´ri ‘yield seed’), ** תודיעtu´di ‘inform’ (only ´tudi, 100; cp. SH תצמיחtaÒ´mi ‘produce’), ** שליחsÇ´li ‘messenger’ (only ´sÇli, 170), ** שמיעsa´mi ‘submissive’ (only ´sami, 232), ** מלוme´lu ‘fullness’ (only ´melu, 205, 240; cp. SH mÇ´lu). 5. Now we can explain the form * סרעיוsaro versus the form * סרוחsaru. The former is a genuine Aramaic form of the pattern qatol stemming directly from *saroÌ, while the latter can be regarded as either: (1) a Hebrew form of the pattern qatul stemming from *saru’uÌ to *sa´ru and following the tendency of SH towards a penultima stress; or (2) an Aramaic form of the pattern qatol which went phonetically the same way, which is rare in Aramaic. One is tempted to adopt the former explanation, though the latter is not impossible. If indeed Hebrew influence occurred, one cannot determine whether this Hebrew influence was the result of a late penetration of a word recited in the Samaritan Hebrew Pentateuch or became part of SA while it was still a living language. The same explanations apply to other (rare) SA words, such as ´ מצבועmaÒbu ‘immersion’ (RPH, 260), and ´ מלוmelu ‘fullness’ (RPH, 152). 6. A completely different explanation for the distribution of u and o in this final position was given recently by E. Qimron.17 His approach is worth mentioning since it disregards several basic principles which are imperative for understanding the structure of SH and SA. 17
‘Substitution for Weak Consonants in the Samaritan Tradition’, in M. Bar Asher and M. Florentin (eds), Samaritan, Hebrew and Aramaic Studies Presented To Professor Abraham Tal (Jerusalem 2005), 73 87. 371
THE ALTERNATION U / O, DIPHTHONGS
Qimron states that SH and SA behave completely differently in the same phonetic environment and he criticizes other scholars who are insufficiently rigorous and assume that SA and SH are identical in their phonetic structure (p. 87). Thus, he argues that SH should have only u in an open post-tonic syllable while in SA — according to its different rules — it should have only o. This assumption forces him to regard the well attested Hebrew forms tÇlo and wgÇlo as merely abnormal. Naturally, he also has to explain the final o of the pronominal suffix. In this case he assumes that this o is not genuine Hebrew, since according to his rule סוּסיו ָ ‘his horses’ should have been developed in SH into **susu (rather than the existing suso). His solution is very simple and at the same time revolutionary: this vowel — so he believes — is actually the Aramaic morpheme borrowed by Hebrew in order to differentiate the plural form from the singular! 7. My arguments against this approach are as follows: I. SH and SA are indeed not entirely identical in their phonetic structure. However, one should note that these two languages are totally identical in their basic rules which determine the structure of their syllables, e.g.: (1) they share the same inventory of sounds and phonemes and neither has a schwa; (2) in both languages the same sounds do occur in the same environments, e.g. the guttural ¨ occurs only at the beginning of the word before the vowels a or å, and the vowels u and o are in most cases two allophones of a single phoneme; (3) they share the same syllable structure characterized by vowel-length totally dependent on the syllable being opened or closed, and they both have no light syllables; (4) they are both characterized by a general penultimate stress, the exceptions to which are principally the same (but see below!); (5) More particularly: in both languages the diphthongs were almost totally contracted, and the cases in which they are retained are the same. Accordingly, the mere assumption that in the same phonetic environment one language, SA, could have a certain phoneme (o), while the other, SH, could not, is in principal very difficult. II. At the same time SH and SA are phonetically distinct in two aspects: a. Obviously, they stemmed from different languages, and consequently — though sharing the same inventory of phonemes — may show different consonants and vowels in the same word. 372
THE ALTERNATION U / O, DIPHTHONGS
For example, ‘peace’ in SH is pronounced ´sÇlom * <( ָׁלוםsalom) while in SA it is pronounced ´selåm * <( ְָׁלםs¢lam). b. In several cases the two languages were different in preferring one of the possible options of Samaritan pronunciation. Thus, in our case the tendency to insert an auxiliary vowel was in SH much more common than it was in SA. The result of this tendency in SH was the relative multitude of forms such as fatu ַפּתוּח ָ ‘open’, stemming from forms with an auxiliary pataÌ furtivum, and nÇtuwwi ‘ נ ָטוּיinclined’, stemming from *natuy > *natuyi > *na†uyi> *natuwi > ndžuwwi. However, SH also went, though rarely, in another direction, i.e. the contraction of the diphthong without inserting any auxiliary vowel. This is the case of gÇlo. SA took generally the opposite way, i.e. not using an auxiliary vowel. Therefore it is relatively abundant with forms such as fato ‘ ָפּתוַחopen’ and presents only rarely forms ַ ‘spirit’ and (may stemming from auxiliary vowels, such as ru רוּח be) saru (see above) though presenting it regularly in the gentilic suffix, e.g. ‘ י ְִחיָדאיalone’ yi´dÇy < *yi´dÇyi. Thus, final o cannot be automatically considered ‘Aramaic’ even though relatively common in SA, and u cannot be considered ‘Hebrew’ even though relatively common in SH.18 III. Though bilingual speakers may borrow certain morphemes from one language to the other,19 the assumption that SH borrowed the morpheme o from SA is quite difficult if we take into consideration that the whole morpheme inventory of SH seems to be ‘clean’ of Aramaic elements.20 The assumption of borrow One may compare this principal mistake to the analysis of III Y forms like גּ ְִדי g¢∂i ‘kid’ in TH. Some scholars believe that these forms followed the Aramaic way in which originally monosyllabic nouns are stressed on the ultima, while in Hebrew they were developed to the segolates: ְכַּתלk¢†al ‘wall’ < *kutl; Hebrew: ֹכֶּתל. How ever, one may explain these nouns as genuine Hebrew: *´gidyu > *´gidy > *´gidiy > *´gidi > *g¢di (the stressed moving to a heavy ultima from a light penultima as is regular in Hebrew: נוּ$‘ י ְָשthey slept’ < *yas¢´nu < *ya´sinu. 19 One may consult, for instance, the scientific literature describing Arabic dia lects. Not a few Arabic dialects in the Arabic peninsula borrowed the 2nd f.s. suffixed pronoun is from Neo South Arabian (see W. Fischer and O. Jastrow, Handbuch der Arabischen Dialekte [trans. to Hebrew by R. Talmon], [Jerusalem 1991], 72; H.A. Qufishe, Yemeni Arabic Reference Grammar [Kensington 1992], 187). In several Ara bic dialects located in west and north west Syria an Aramaic substrat is reflected in the use of some Aramaic suffixed pronouns such as hun and hin (see P. Behnstedt, Sprachatlas von Syrien [Wiesbaden 1997], map 272, p. 545). I thank Dr Mutzafi for this reference and for the data regarding Eastern Neo Aramaic adduced below. 20 I agree with Ben Îayyim (GSH, § 2.013) that the SH 2nd fem. sing. affirmative ti is genuine Hebrew. 18
373
THE ALTERNATION U / O, DIPHTHONGS
ing seems to be more improbable when accompanied by the belief that the motivation for this borrowing was the differentiation between the singular and the plural. The assumption is that ‘ סוּסָיוhis horses’ suso did not shift to susu since susu סוּסוdenotes the singular, ‘his horse’. As will be shown below, I also do not accept this simplistic explanation for TH. In SH it can be even more strongly refuted, since in this language the difference between the plural and the singular was neutralized in the 1st person, i.e. ‘ סוּסִיmy horse’ is pronounced susi, and at the same time ‘ סוּסַיmy horses’ is pronounced also susi (due to the contraction of the diphthong ay). Is it reasonable to assume that whilst satisfied with this neutralization, the Samaritans still borrowed the Aramaic morpheme to preserve difference in other part of the paradigm?! Note that the Semitic languages generally go the opposite way from secondary differentiation between singular and plural pronominal suffixes, i.e. due to several reasons — phonetic or non-phonetic — they lose the original difference and do nothing to change it. This is the case, for instance, of the 2m. s. suffix attached to a plural noun in Biblical Aramaic: while shifted to –åÈ it was merged with the morpheme of the s., e.g. susåÈ stands both for ‘ סוסךyour horse’ and ‘your horses’ (in which case the Ketib is )סוסיך. Qimron himself adduces the case of the Hebrew of the Dead Sea Scrolls in which the differentiation between singular and plural pronominal suffix was neutralized, e.g. ‘ ידיוhis hand’ versus ‘ מצוותוhis commandments’ (p. 84, n. 40). The same tendency is well attested in Eastern NeoAramaic. Moreover, in this dialect the neutralization of the difference between singular and plural is not due to a phonetic accident but rather reflects a general tendency towards simplifying the system.21 IV. Qimron’s claim that every diphthong aw shifted to o and finally to u is also based on several basic misunderstandings: a. The conclusion regarding the existence of a shift in a certain language is not based on a simple counting of forms. Thus, the fact that aw shifted to o (and did not further shift to u) only in the pronominal suffix — though seemingly not in (three) other
21
Dr Mutzafi orally informed me about the ‘Jewish Neo Aramaic forms tori “my ox/oxen”, torox “your (m. s.) ox/oxen” and torax “your (fem. sing.) ox/oxen”, and likewise throughout the paradigm’. 374
THE ALTERNATION U / O, DIPHTHONGS
b.
a.
b.
c.
words — does not mean that the shift in question did not occur.22 Determining the exact phonetic conditions in which a certain shift took place is dependent on the source forms that one reconstructs. In our particular case we are dealing with the protoSamaritan Hebrew and Aramaic layers. Indeed, SH reflects a later layer than TH. However this does not mean — as it seems that Qimron believes — that the existing Tiberian forms are the exact representatives of proto-SH. We may adduce dozens of SH forms which could not have stemmed from attested TH forms. A few examples will suffice: ֵצאת, infinitive Qal of יצ"א, i.e. ‘to go’ is pronounced siyyåt in SH stemming from a nonTiberian form *ֵצַאת. 23 Now let us analyse the three Tiberian nouns ending in diphthongs which Qimron, incorrectly, believed to be the basic forms of the Samaritan forms. He regarded them as forms which prove the shift aw > o > u: חָדּיו ְ ַ ‘ יtogether’ yaddu. The Samaritan pronunciation of the Pentateuch does not merely reflect the language of the Samaritans. It also reflects their grammatical perception and the way in which they understood the text. It is quite clear that in this case the Samaritans saw in יחדוa form to which the singular morpheme is attached, as is proved by the only spelling of this adverb in the SP: יחדו, without yod of the plural. Thus, one cannot reconstruct *yaddaw > *yaddo > yaddu, but rather start directly from the last stage, without any diphthong contracted24. ָענו ָ ‘humble’ ¨anu may possibly have stemmed — as already noted by Ben-Îayyim (and rejected by Qimron, p. 84, n. 41) — from a passive participle of Qal, i.e. *¨anuw. Again, the proto-Samaritan form is different from the Tiberian one. ׁו ְא ָ ‘false’ su may be the Samaritan result of the following provenance: *sawˆ > *saww > *suww25 > *suw > su.
22 Note, for example, that the Hebrew nun was assimilated, according to the rule, only in one III Nun verb, nå†atti ִ ‘ נ ַָתִתּיI gave’, while in all other verbs of this category the phonetic rule was contradicted, thus: ַכנ ְָתּ$‘ ָשyou dwelt’, ‘ ָטַמנ ְִתּיI hide’, etc. 23 The whole issue was thoroughly discussed and exemplified by Ben Îayyim in the Introduction to GSH; see especially § 0.13 16. 24 The same applies to the nouns ‘ דמיוhis blood’ dammu and ‘ בעמיוhis people’ ¨ammu which were regarded by the Samaritans as singular nouns (GSH, §1.2.4). 25 For the assimilation of a to u before a double ww, compare in SH: ַֹצוּ ָארו *Òawwåro > ÒuwwÇru (see GSH, § 1.5.3.1 e).
375
THE ALTERNATION U / O, DIPHTHONGS
The pronominal suffix of the s., i.e. ‘ סוּסוhis horse’ does not stem in SH directly from the old reconstructed form *susaw. This is the basic form of the Tiberian word suso, which is the right representative of the proto-SH layer. The conclusion is thus inevitable: the only undoubted protoSamaritan form containing a final diphthong aw is the 3m. s. suffix attached to a plural noun. This diphthong has been contracted to o but did not shift to u, being regarded as occurring in a morpho-phonemically closed syllable.26 V. It is not difficult to assume that a syllable ending with a contracted diphthong is still regarded as ‘virtually closed’, or, in other words, underlyingly closed, as we argue concerning the SH o (< aw) of the pronominal suffix. A similar case is the ultima tonic I-Yod verbs ( )פ"יwith waw consecutive in TH, e.g. יׁן ַ ִ ַ ויּwayyi´sån (< *wayyiy´san) versus לְך ֶ ֵ ַ ויּway´yelék. Since only open syllables retained the penultima stress, we must assume that the syllable yi in the verb יׁן ַ ִ ַ ויּwas regarded as closed, even though it ended in a (proto-TH) long vowel. 8. Bearing in mind the idea that the morpheme of the Samaritan pronominal suffix resulted from firm phonological rules, we may now look at its Tiberian counterpart. A common explanation for the retention of the diphthong in סוּסָיוsusåw (versus סוּסוsuso) is again the motivation of differentiation between the two forms. However, even if one accepts this approach it is still necessary to find a reasonable solution for the retention of the diphthong in the forms discussed above ָׁו ְאsåw ‘false’ (and גּיַ ְאgåy ‘valley’). ְסָתוs¢†åw ‘winter’ and ָָענו ¨ånåw ‘humble’, are also relevant, though different in their basic forms. Therefore one has to prefer the phonological approach which supplies one explanation for all the non-contracted diphthongs. According to the historical version of this approach,27 in all these forms the semi-vowel closing the diphthong was geminated in consequence of assimilation of ˆ or y: סוּסיו ָ ׁו ְא ָ גּיַ ְא
susåw < *susaww < *susayw < *susayhu såw < *saww < *sawˆ gåy < *gayy < *gayˆ
26
See n. 9 above. The phonetic approach is found already in classical grammars of Hebrew (see e.g. Blau, A Grammar of Biblical Hebrew [Wiesbaden 1993], §7.3.2.2.). 27
376
THE ALTERNATION U / O, DIPHTHONGS
The only difficulty of this approach is the assumption that phonetically the semi-vowel y assimilated to the proceeding semi-vowel w, and that the glottal stop assimilated regressively to the semi vowels w and y. Such assimilations do not occur elsewhere in Hebrew. Therefore I would prefer to adopt a morpho-phonemic approach: in all these cases the semi-vowel closing the vowel — though not geminated phonetically — is part of a category which is phonetically geminated. Thus, the monosyllabic nouns såw and gåy belong — both historically and synchronically — to the pattern qatl. The pairs ׁו ְא ָ / אה ָ ׁוand גּיַ ְא/ גּ ֵיא- are considered by the speakers identical to and therefore behave exactly like the pairs ָמו ֶת/ מוִתיand ַבּי ִת/ יתי ִ בּ ֵ in which the stressed diphthong in each pair is retained. In other words, [såw] and [gåy] are phonemically /såww/ and /gåyy/. The retention of the diphthong in the nouns ְסָתוs¢†åw ‘winter’ and ָ¨ ָענוånåw has different roots. Contrary to the historically (and morpho-phonemically!) short QåmeÒ in ָשו ְאwhich is the result of assimilation of the short a to the proceeding w, the QåmeÒ in ְסָתו (< *sitawu) and in ָ¨* <( ָענוanaw < *¨anawu) was long in proto-TH — either originally long or lengthened in an open tonic syllable. Thus, before formulating a general rule determining the behaviour of all the diphthongs in Hebrew, we may state that a diphthong was not contracted if its vowel or semi-vowel were long. 9. Turning back to סוּסָיוand to its accepted provenance – susåw < *susaww < *susayw < *susayhu – we may suggest another process which phonetically seems to be less difficult: *susayhu > *susayw > *susaw (assuming ay > a, as in * < ְלָאןla´ˆayn) > susåw. According to this reconstruction, the diphthong was retained because its QåmeÒ was long. According to the above mentioned assumption, the diphthong was retained due to the gemination of the w, the QåmeÒ being a short vowel resulting from assimilation of a to w, like the QåmeÒ of ָׁו ְא. While we may not be certain about the exact historical provenance of the pronominal suffix in סוּסיו ָ , it seems that its morphophonemic status is quite clear: the suffix is part of a category — the inflected noun with pronominal suffixes — every member of which is characterized by a suffix preceded by a long element. Accordingly, we either find a diphthong, as in סוּסי ַ ‘my horses’ and in סוּסי ְִך ַ ‘your (fem.) horses’, or an historical (and synchronic) long vowel — resulting from a former diphthong ay — before the suffix, e.g. סוּסיָך ֶ suséÈå 377
THE ALTERNATION U / O, DIPHTHONGS
‘your (masc.) horses’, ָ סוּסֶיהsuséhå ‘her horses’,28 etc. Moreover, the so-called poetic 3rd suffix takes the form סוּסיהוּ ֵ , as in י ֵָדיהוּyå∂ehu ‘its hands’ (Hab. 3:10). This both historical and morpho-phonemic view of the pronominal suffixes reinforces my suggestion that the diphthong of סוּסיו ָ was retained because its vowel was long, resulting from the attraction of the diphthong ay. 10. Taking advantage of this study I would like to classify the diphthongs in proto-TH according to their phonetic value, which clearly affects their behaviour:29 A. ‘Super Heavy diphthongs’ are those consisting of three sounds, i.e. either a long vowel or a (morpho-phonemic) geminated semi-vowel. These diphthongs — standing in what may be called ‘super-closed’ syllables — were retained regardless of the phonetic environment. Examples: ְסָתיו, ַחוּ ָה, ַחי, ָשו ְא, סוּסיו ָ . B. ‘Heavy diphthongs’ are those containing a consonant after the semi vowel. Examples: *bayt ‘ ַבּי ִתhouse’, *mawt ‘ ָמו ֶתdeath’. The behaviour of these diphthongs — standing in what may be called ‘simple closed’ syllables — depends entirely on the stress: they were retained only when stressed (though broken later by an auxiliary vowel), and when unstressed they were contracted: ´bayit בּי ִת ַ / be´ti בּיתִי ֵ ‘my house’, ´mawvt מו ֶת ָ / mo´ti מותי ִ ‘my death’. C. ‘Light diphthongs’ are those consisting of only two sounds, i.e. a short vowel and a semi-vowel. These diphthongs — standing in 28
By stating that the vowel is synchronic long I do not mean that it was pho netically long. I regard it as long since it holds the penultima stress. A penultima stress was retained when the stressed syllable was closed, i.e. heavy, as in ָכַּתְבָתּ kå ´†aß tå ‘you wrote’ < *ka ´tab ta. Thus, the penultima stressed syllable of forms ֶ su sé Èå has to be regarded morpho phonemically heavy as well, i.e. con like סוּסיָך ָ תּ ִ כ ְ ה ִ : taining a long vowel: / su sé¯ Èå /. The same applies to pairs such as ָכַּתב/ יבה [hiÈ ´ti ßå] is morpho phonemically / hik´ ti bå / while [kå†aß ] is / kåtab / since its vowel is not retained: ָכְּתָבהk冢ßå. 29 I am based on the brilliant primordial description of Ben Îayyim in GSH, § 1.4.2. Yet, my view regarding the effect of the tone on vowel length is somewhat different, e.g. I do not believe that the second QåmeÒ in ָבּנ ָיוwas lengthened due to the stress. As mentioned above it either results from ay (> a) or from assimilation to the preceding ww. The (first) QåmeÒ of ַהָמּו ְָתהham´måw†å ‘the death’, also regarded long by Ben Îayyim, is discussed in the next note. The terminology offered is mine. 378
THE ALTERNATION U / O, DIPHTHONGS
open syllables — were contracted regardless of the stress: סוּסיָך ֶ su´séÈå ‘your (m. s.) horses’, סוּסיכֶם ֵ suseÈém ‘your (plural masc.) horses’.30 11. Classifying the diphthongs this way enables us to compare their behaviour to that of the short vowels a, i, and u:31 A. In syllables closed by geminated consonants, the three vowels completely retain both their original articulation and quantity: סּע ָ מ ַ ‘travel’ masså¨ < *massa¨u, פּיל ִ ה ִ ‘through’ hippil < *hippila, בּר ַ ׁ ֻ ‘was broken’ subbar. This group is parallel to the ‘Super Heavy diphthong’ group. B. In syllables closed by a simple consonant, the retention of the three vowels may be affected by the stress: a. the vowel a, while stressed was always retained, e.g. ‘ ָכַּתְבִתּיI wrote’ kå´†aßti, but while unstressed could be changed, e.g. ְדבּר ָ מ ִ ‘desert’ mi∂´bår ֶ ֶ ‘ יyour (plural masc.) hands’ yé∂´Èém < *ya∂Èém. < *madbar, ְדכם b. the vowel i, while stressed, was shifted either to e or to a, e.g. עץ ֵ ‘tree’ ¨eÒ < *´¨iÒu, בּר ֵ ‘ י ְַדhe will tell’ y¢∂ab´ber < *y¢∂ab´bir; תּי ִ ְ ׁנ ַ ָ ‘ יI slept’ yå´santi < *ya´sinti. c. the vowel u, while stressed, was always changed, e.g. ‘ דּבbear’ doß < *dub, ‘ י ְִכתּבhe will write’ yiÈ´tob < *yiÈ´tuß, and while unstressed was either retained, e.g. A ‘light’ diphthong is retained in ַהָמּו ְָתהbecause all forms containing direc tional ָהfollow exactly their basic forms (see E. Kautzsch [ed.], Gesenius’ Hebrew Grammar [sec. ed.; trans. A.E. Cowley], [Oxford 1910], §90i). Thus, ַהָמּו ְָתהfollows מו ֶת ָ as שֹק ֶ מּ ֶ ְדבָּרה ַדּ ַ מ ִ mi∂barå ‘to the wilderness of Damascus’ (1 Kgs 19:15) follows ַ מ ִ avoiding the QåmeÒ **midbårå expected in the open sylla the construct form ְדבּר ble. Thus, no doubt that the QåmeÒ of ַהָמּו ְָתהis short (cf. the end of the previous note). For the cause of this phenomenon, see J. Blau, ‘The Morphology of Nouns Terminating in Paragogic (Directional a)’, in M. Bar Asher (ed.), Language Studies V VI, Israel Yeivin Festschrift (Jerusalem 1992), 7 11. Blau convincingly showed that the roots of this interesting phenomenon are to be found in the old differences between pausal and contextual forms which resembled the difference between the emphatic and construct state in nouns. Yet, I do not believe that the diphthongs in תה ָ ְ בּי ַ , לה ָ ְ לי ַ , and תה ָ ְ מו ָ were retained because they were stressed (Blau, p. 93). לה ָ ְ עו ַ ¨aw´lå ‘injustice’ is not to be regarded as just an analogous form to the masculine עו ֶל ָ ¨åwél, but rather stemming from לה ָ ְ עו ָ * *´¨åwlå, a masculine noun derived from עו ֶל ָ with directional ָהas in תה ָ ְ מּו ָ ה ַ . One may assume that it was later understood as a feminine noun to which again! the directional ָהwas added, thus ( ַעו ְָלָתהEz. 28:15) and ( ֹעָלָתהJob 5:16). 31 I abstain from dealing with special and exceptions of the main rules stated below. 30
379
THE ALTERNATION U / O, DIPHTHONGS
לְך ָ ש ְ מ ֻ ‘thrown’ mus´låÈ or changed, e.g. תּי ִ כ ְ ל ַ ש ְ ה ָ ‘I was thrown’
hås´laÈti < *hus´laÈti. This group is parallel to the ‘Heavy diphthong’ group. C. In open syllables the three vowels lost completely both their original articulation and quantity: ‘ ָדָּברthing’ dåßår < *dabaru, בִרי ָ ‘ ְדּmy thing’ d¢ßåri < *dabari; ׁנ ָה ֵ ‘sleep’ senå < *sina, תי ִ ָ ׁנ ְ ‘my sleep’ s¢nå†i < * sinati; ‘ ָקְדָקדhead’ qå∂qå∂ < *qu∂qu∂u, ִתְּשְמרוּ ‘you (plural) will keep’ tism¢ru < *tismuru. This group is parallel to the ‘Light diphthong’ group. 12. Postscript A. A newly discovered ancient spelling — — סרעיוattests to a genuine pronunciation different from that known to us from the current recitation of the Samaritans. Thus, again, we realize that even the most reliable oral tradition has to be checked against other existing pieces of evidence. B. Samaritan Hebrew and Aramaic, though sharing the same phonetic structure, differ mainly in the choice of options available in the general Samaritan pronunciation. In the present study we have argued that the use of the auxiliary vowel in general, and specifically the use of the pataÌ furtivum was relatively widespread in Samaritan Hebrew and relatively rare in Samaritan Aramaic. Only this fact — and not alleged differences in the main phonetic structure of the two languages — caused the large distribution of the vowel u in an open post-tonic syllable in SH versus the clear preference of the vowel o in this position in Samaritan Aramaic. C. Morpho-phonemic — alongside historical — analysis was used thoroughly in this short study. It helped us to prove that: (1) the vowel o of the 3m. s. attached to a plural noun in SH — e.g. סוּסָיוsuso — is genuine in this language, being underlyingly a closed syllable; (2) the reason for the retention of the diphthong åw in the same pronominal suffix in Tiberian Hebrew — e.g. סוּסיו ָ susåw — is purely phonetic, i.e. the long vowel before the semi-vowel. In other words, it helped us to propose the following process: *susayhu > *susayw > *susaw > susåw = סוּסָיו. D. Tracing the exact shape of ‘proto-Samaritan Hebrew’ is of course impossible. What is certain, however, is that one cannot automatically assume that a Tiberian Hebrew form is the source form of a Samaritan word. Thus, the Samaritan adverb יחדוyaddu did not originate from the Tiberian י ְַחָדּיוyaÌdåw but rather was derived from יחדand the singular suffix u (< hu). 380
PLURAL BASED DUALS JASON BEMBRY EMMANUEL SCHOOL OF RELIGION
Abstract While most dual forms in the Hebrew Bible are built upon the singu lar base, there are several examples where the plural form is the base of the dual noun. The inserted a vowel, a characteristic of plurals in North West Semitic, appears in these words, indicating that it is the plural base upon which the dual ending is added. These forms are likely caused by an erroneous assumption about the word delet ‘door’, which constitutes the most examples of this phenomenon. The word delet was assumed to be a III he noun, not a qvtl noun. On analogy with other III he nouns that present a medial a vowel in their dual forms, delet assimilated to this same pattern. Select qvtl nouns with a liquid consonant in the middle position followed the pattern of delet and thus created these inconsistencies. The pattern was then applied to two other non qvtl nouns resulting in further inconsistencies.
That duals are built on the singular base is a common assumption among Hebrew grammarians.1 Such a consensus is well founded since the vast majority of the forms in this language are comprised of dual endings added to the singular base. Yet there are a number of forms in the Hebrew Bible that do not conform to this rule and they appear in sufficient number to merit further analysis. This paper addresses these exceptions in Hebrew and provides a likely explanation for them. It has been convincingly demonstrated that an -a- vowel has been inserted into the plural base of qvtl nouns in Hebrew.2 So Hebrew plurals on qvtl nouns are actually doubly marked, a feature that is one of the distinguishing marks of North-West Semitic.3 So in the 1 As reflected in Sabatino Moscati, An Introduction to the Comparative Grammar of the Semitic Languages; Phonology and Morphology (Wiesbaden 1980), 86. 2 H.L. Ginsberg, ‘The Northwest Semitic Languages’, in Benjamin Mazar, The World History of the Jewish People, Vol II (Piscataway, NJ 1970), 102. Thomas O. Lambdin and John Huehnergard, Historical Grammar of Classical Hebrew (Unpub lished Manuscript, Cambridge, MA, 1998), 27. See also Joshua Fox, Semitic Noun Patterns (HSS 52, Winona Lake 2003), 38, footnote 3. 3 John Huehnergard, ‘Remarks on the Classification of the Northwest Semitic
383
PLURAL BASED DUALS
form ְמָלִכיםthe plural of the segholate ‘ ֶמֶלְךking’, the qamets under the lamed is the -a- vowel that is inserted into the original singular form *malk which marks the word as plural in addition to the external ending ים.4 Such pluralization is standard among masculine and feminine qvtl nouns. The dual in Hebrew is normally built upon the singular base. Indeed, this is the general pattern seen most clearly in qvtl nouns where the third radical is a bgdkpt letter. So, for example, ‘ ֶעֶרבevening’, is ְרבּי ִם ַ ע ַ in all of the eleven occurrences of this word in the dual (Exod. 12:6; 16:12; 29:39, 41; 30:8; Lev. 23:5; Num. 9:3, 5, 11; 28:4, 8). Similarly, the word ‘ ֶבֶּרְךknee’ is normally ְרכּי ִם ַ בּ ִ in the dual.5 Such is the case seventeen times (ְרכּי ִם ַ בּ ִ in Deut. 28:35; Isa. 35:3; 66:12; ָ בּ ִ in Judg. 7:5; 1 Kgs Ezek. 7:17; 21:12; 47:4; Job 3:12; 4:4; ְרכּיו 8:54; 18:42; 2 Kgs 1:13; 2 Chron. 6:13; יה ָ ְרכּ ֶ בּ ִ in Judg. 16:19; 2 Kgs 4:20; ְרכּי ַ בּ ִ in Ps. 109:24; Dan. 10:10). Yet the word לת ֶ ֶדּ when used in the dual has the unexpected form ְדָּלַתי ִםin the following eleven examples: Deut. 3:5, 1 Sam. 23:7; Isa. 45:1; Jer. 49:31; Ezek. 38:11; Mal. 1:10; Job 38:8, 10; Eccl. 12:4; 2 Chron. 8:5; 14:6. The same -a- vowel appears in forms with pronominal suffixes in the following five examples: Josh. 26:6 (יה ָ ת ֶ ל ָ ;)ְדּ1 Kgs 16:34 (ָ ;)ְדָּלֶתיהIsa. 26:20 ( ְדָּלְתָךis the Qere while the Kethib is ;)דלתיך Zech. 11:17 ()ְדָּלֶתיָך: Job 31:32 ()ְדָּלַתי. Even in the forms where the word is in construct, the reminiscence of the historical -a- vowel is encoded in the noticeably absent daghesh in the taw in the following six forms: Josh. 2:19 (;)ַדְּלֵתי ְ ְ ;)וJob 3:10 Judg. 11:31 ( ;)ִמַדְּלֵתי1 Kgs 7:50 ( ;)ַדְּלֵתיPs. 78:23 (ַדלתי ( ;)ַדְּלֵתי2 Chron. 4:22 (ַדלֵתי ְ ְ )ו. One interesting example of a construct form is Job 41:6 where the form is ְדָּלֵתיin which the -a- vowel is present. One could reanalyse the six construct forms listed above as masculine plural constructs. Yet the normal plural for ֶדֶּלתis the feminine form ( ְדָּלתותconstruct plural being )ַדְּלתות. It seems, furthermore, that the biblical writers knew this distinction between dual constructs and plural constructs for ֶדֶּלתgiven that two of the six verses cited above have the feminine plural form ְדָּלתותin the same sentence (1 Kgs 7:50 and 2 Chron. 4:22). These dual forms of Languages’, in The Balaam Text from Deir ¨Allah Re evaluated. Proceedings of the In ternational Symposium held at Leiden 21 24 August 1989 (J. Hoftijzer and G. van der Kooij [eds], Leiden 1991), 284. 4 In this paper forms marked with an asterisk denote unattested probable his torical forms of the language. 5 There is one occurrence of ֶבֶּרְךin a dual construct without the daghesh in the ֵ בּ ִ in Judg. 7:6. kaph in the form ְרכי 384
PLURAL BASED DUALS
לת ֶ ֶדּdeviate from the expected pattern of *תּי ִם ַ ל ְ ַדּ, a form that never occurs. In fact, of all occurrences of ֶדֶּלתin the Hebrew Bible only one form, ( ְבַּדְּלתּו2 Kgs 12:10) has a daghesh in the taw indicating
the singular base. It is not only the Masoretic tradition where we see these exceptions with ֶדֶּלת. In the Babylonian tradition the same type of inserted -a- vowel is seen in the absolute form of the dual for ֶדֶּלת. Among the Cairo Genizah fragments TS A39.20 preserves the word in Job 38:8 with a Babylonian qamets with the lamed. Interestingly, the characteristic -a- vowel appears above the lamed and slightly to the left so that the vowel appears just above the right hand vertical of the following letter taw (see fig. 1). This is a consistent scribal practice in this text as demonstrated by other forms on the same page that may serve as controls on vowel placement. Note for example in the intervening verse the form ֲחֻתָלּתוattests the identical placement of the qamets again slightly to the left of the lamed’s vertical. Another fragment in this collection, TS A39.3, preserves the word ְדָּלַתי ִםin Deut. 3:5. Again, the Babylonian qamets is associated with the lamed.6 Unlike the example from Job 38:8 the qamets appears above the lamed slightly to the right of the upper vertical (see fig. 2). When compared with duals that conform to the general rule of expected paradigmatic vocalization seen in Masoretic Hebrew, one sees the same distinction between the expected morphology and that of the anomalous examples with an inserted a-vowel. TS AS 54.12 preserves the dual form אזנְ יַ ִם ָ . Likewise TS NS 251.41 preserves the dual form לי ִם ַ ְ ַרג. Unfortunately, the Hexaplaric evidence provides no relevant exemplars to aid the discussion.7
Fig. 1
Fig. 2
The dual form of ֶדֶּלתdoes not constitute all the exceptions to the general rule of duals being built upon the singular form. Another form with a bgdkpt letter in the third position that demonstrates the 6 I would like to thank Ben Outhwaite, head of the Taylor Schechter Genizah Research Unit, Cambridge University Library for his invaluable assistance in access ing and reading the Geniza fragments. 7 See the lists of words in Einar Brønno, Studien Über Hebräische Morphologie und Vokalismus (Leipzig 1943), 142 3; 145.
385
PLURAL BASED DUALS
same pattern is seen in two examples of ‘ ֶדֶּרְךway’ a qatl noun. Prov. 28:6 and 28:18 have ָרכי ִם ַ ְדּ, again evincing the characteristic -avowel associated with the plural base in qvtl nouns and the concomitant lack of daghesh in the kaph. The same pattern emerges for four occurrences of the dual ‘ ֶקֶרןhorn’. In Dan. 8:3 [2x], 6, 20 the form is ְקָרניַ ִםagain with the internal plural -a- vowel. The expected form ַקְרניַ ִםappears on two occasions as well, however (Amos 6:13 and Hab. 3:4).8 An additional attestation is seen in the noun ‘ ֶכַּרעleg’. On four occasions the dual form ָרעי ִם ַ כּ ְ is used (Lev. 1:13; 8:21; 9:14; 11:21 and Amos 3:12). The preceding examples all come from qvtl nouns in Hebrew. Yet they are not the only nouns that are pointed as duals formed on the plural base. To be added to this list are the four dual forms of חוָמה presented as חומַתי ִםin 2 Kgs 25:4; Isa. 22:11; Jer. 39:4; and 52:7 as well as the dual form of ַ לוּחpresented as ֻלחַתי ִםin Ezek. 27:5. Certainly the taw in the forms of חוָמהpreserves the older ending of feminine nouns but were considered as plural endings by the Masoretes. That is, the consonantal structure of the singular can be traced back to the form *חומת. To the Masoretes this final taw could be reanalysed as a marker of the feminine plural ôt. So while the forms of חומַתי ִםare clearly plural based duals, they may owe their look to improper Masoretic analysis of the taw. Even so, the taw in תי ִם ַ לח ֻ cannot be easily dismissed as an old marker of the feminine singular. ַ לוּחis listed in BDB as a masculine noun9 as it is in Ugaritic according to del Olmo Lete and Sanmartín.10 Indeed, its plural form in Hebrew is לוּחתbut it is clearly attested in several languages in the singular without the taw. So the dual form is added to a true plural ending. With all the exceptions to the general rule of duals laid out above, it remains to explain them. In his lengthy monograph on the subject of duals, Fontinoy only cursorily cites ְדָּלַתי ִםand ֻלחַתי ִםand places them in a category of art and architecture, yet says nothing about their irregular phonology.11 Bauer and Leander mention the forms תי ִם ַ לח ֻ and תי ִם ַ חומbut do not address the other examples and give no 8 Additional forms of the expected dual of )ַקְרניַ ִם( ֶקֶרןcould be cited here since the plural form is ַקְרנוֹת. That is, many construct forms like the form ֵ ַקְרניlikely at test a singular based dual in construct. 9 BDB, 431b. 10 Gregorio Del Olmo Lete and Joaquín Sanmartín, A Dictionary of the Ugaritic Language in the Alphabetic Tradition, translated by W.G.E. Watson (Leiden 2003), 494. 11 Charles Fontinoy, Le duel dans les langues sémitiques (Paris 1969), 51.
386
PLURAL BASED DUALS
explanation for these exceptions.12 Of the twenty-one examples which are qvtl nouns that I have cited above, all are a-class segholates with the exception of ֶכַּרע, the vowel of which is not indicated in the Hebrew Bible because of a lack of data. Comparative evidence is not decisive. Although the Akkadian cognate kuritum, ‘shin (of animals)’, might suggest it was u-class,13 the Aramaic ֶכַּרעand ְרעא ָ כּ ַ suggests an underlying qatal form, thus complicating the issue. Since the singular is not attested in Biblical Hebrew, it is possible that this form was an original qatal form. The other common feature in this class of nouns is the liquid consonant in the second position, in these cases לand ר. This is not to say that such criteria explain the phenomenon. Clearly the same criteria fit our first example of regular duals, ְרבּי ִם ַ ע ַ , which is based on an a-class segholate with the liquid consonant רin the second position. If there were only a handful of forms like this in the Hebrew Bible, one could plausibly suggest that they are mere pointing mistakes perpetuated by the ancient handlers of the texts through the centuries. Yet this pattern appears in nearly all of the dual forms of לת ֶ ֶדּand it is seen in a few other forms. Furthermore, these anomalous forms are scattered through the entire corpus, so attributing these forms to accidental pointing seems most unlikely. Although the lack of sufficient evidence precludes absolute explanations for these exceptions, I would suggest that the pattern of plural-based duals in Hebrew began with a word like ֶדֶּלת, which is actually a biradical root with a -t suffix and thus not a true qvtl.14 There is a single occurrence in the Hebrew Bible of ָדּלmeaning ‘door’ in Ps. 141:3 and Phoenician evidence also supports the biradical base. Even so, the word in Hebrew was vocalized in the singular as a qatl from an assumed root of דלת. So in the singular form the word could appear as a normal qvtl. Yet the dual form preserves the distinction ֶדֶּלתhas visà-vis other qvtls. Over time, this form would have blended with nouns that are now classified as III- הand so its dual form looks rather similar to that of words such as ָשָפה. Just as the dual form of פה ָ ש ָ is תי ִם ַ פ ָ ש ְ , so too לת ֶ ֶדּdisplayed the long a vowel in its dual form. Unfortunately, since there are no other dual forms of qvtl nouns that have taw as their third letter, it is impossible to demonstrate this phenomenon in another analogous noun. This appears to me to be the 12
Hans Bauer and Pontus Leander, Historische Grammatik der Hebräischen Sprache des Alten Testamentes (Wiesbaden 1962), 516. 13 CAD, vol 8, 560. 14 Fox 2003: 42. 387
PLURAL BASED DUALS
first step in determining why some Hebrew nouns in the dual are based on the plural form. I suggest that the initial situation was one in which the feminine endings created a set of biforms — 1a and 1b: 1a) *dalt (f.s.) ~ *daltaym (expected but unattested dual form) 1b) *dalat (f.s.) ~*dalataym (feminine dual)
While 1a provided the commonly attested singular form which was subsequently mis-analysed as masculine, its corresponding dual form fell out of use. In its place 1b supplied the only attested dual forms. 2a) *dalt> delet 2b) *dalataym > d¢latayim
The next step involved another analogy whereby some a-class segholates (qatl nouns) with a לor a רas a medial consonant followed the example of the noun ֶדֶּלת. 3) delet : d¢latayim :: derek : d¢rakayim
So the dual form of ֶדֶּרְך, ֶקֶרן, and perhaps ֶכֵּרעfollowed the pattern of ֶדֶּלת, at least in some examples, inserting the a-vowel of the plural and adding the dual ending resulting in several exceptions to the general rule. The explanation for ֻלחַתי ִםand חומַתי ִםseems to be a development independent from the qvtl exceptions. Even so, like the qvtl forms above, the process by which חומַתי ִםis derived employs analogy. As noted above חומַתי ִםis derived from an assumption by the Masoretes that the taw was a marker of the plural rather than an archaic feminine marker. One may surmise that ֻלחַתי ִםwas created on the basis of analogy with חומַתי ִם. 4) Ìômâh : Ìômôt : Ìômôtayim :: luaÌ : luÌôt : luÌôtayim
If the plural of חוָמהis חומתand the dual form is חומַתי ִם, and the plural of ֻלַחis ֻלחת, then the dual form could presumably be ֻלחַתי ִם.
388
NOTE ON THE RECENTLY PUBLISHED DISCOURSE ON PRIESTHOOD (BL ADD. 18295, FF. 137B–140B) GEOFFREY HERMAN HEBREW UNIVERSITY OF JERUSALEM
In a recent issue of the Journal of Semitic Studies a previously unpublished Syriac text on the priesthood kept at the British Library was discussed, published and translated by Adam H. Becker.1 It consists of two chapters: the first is written in the format of a dialogue between a Jew and a Christian on the subject of the abrogation of the Israelite priesthood by that of Jesus; the second is a short list of reasons why the Israelite priesthood has been superseded by the Christian priesthood. The text is found in a manuscript of West Syrian origin, dated A. Gr. 1914 (=1602/3 CE).2 Becker, had already discussed some of this text in an earlier publication where he sought to explain the concerns of this composition within the context of the early Islamic period.3 He was nevertheless unable to identify the composition, its author, provenance,4 nor to date it in any definitive way.5 He does observe, on the basis of internal allusions, that it formed a part of a larger work ‘which has not been identified or is no longer extant’.6 1 A.H. Becker, ‘The Discourse on Priesthood (BL Add 18295, ff. 137b 140b): an anti Jewish Text on the abrogation of the Israelite Priesthood’, JSS 51/1 (2006), 85 115. 2 W. Wright, A Catalogue of the Syriac Manuscripts in the British Museum (London 1872), part 3, 1184. 3 A.H. Becker, ‘Beyond the Spatial and Temporal Limes, Questioning the “Part ing of the Ways” Outside the Roman Empire’, in A. Becker, A.Y. Reed (eds), The Ways that Never Parted,, (Texte und Studien zum antiken Judentum 95, Tübingen 2003), 385 9; idem, ‘The Discourse on Priesthood’, 94, 96, n. 62. 4 Becker, ‘The Discourse on Priesthood’, 98 9. 5 Becker, ‘The Discourse on Priesthood’, 86 7: ‘For what it is worth, we can be certain then that P [= The Discourse on the Priesthood] was composed sometime between the seventh and the sixteenth centuries’. He favours a date in the early Islamic period. See, ibid, 85 8, 98; ‘Beyond the Spatial and Temporal Limes’, 385, no. 50. 6 Becker, ‘The Discourse on the Priesthood’, 88.
389
DISCOURSE ON PRIESTHOOD (BL ADD. 18295, FF. 137B 140B)
This discourse is, indeed, an excerpt of a larger work or rather, compilation, that, although hitherto unpublished, is extant in a number of manuscripts. These manuscripts, and their contents have been the subject of earlier scholarly examination. Of particular importance are the studies by Michel Breydy in the 1960–70s7 who discussed the relationship between Vat. Syr. 100;8 Vat. Syr. 581;9 Mingana 56;10 and Istanbul Maryemana 7;11 and others,12 all of 7
See, especially, M. Breydy, ‘Les compilations syriaques sur le sacerdoce au LX siècle’, Symposium Syriacum, 205 (1978), 267 93 where earlier literature is cited; idem, ‘Historisch literarische Daten zu den Hauptmanuskripten der Lehre vom Priestertum bei Mose bar Kepha’, Ostkirchliche Studien 25 (1976), 67 71; idem, Le concept du sacerdoce à la lumière des textes patristiques des Syro Maronites (Beirut 1964), 48 56 (non vidi). 8 See S.E. Assemani and J.S. Assemani, Biblothecae Apostolicae Vaticanae Codicum Manuscriptorum Catalogus (Rome 1758 [rep. Paris 1926]), 2, 543: ‘Responsio ad objectiones Judaeorum’. It is the third chapter of the first book. See, too, J.S. Assemani, Bibliothecae Orientalis, II (1721), 121 2; A Baumstark, Geschichte der syrischen Literatur (Bonn 1922), 277; W. Wright, A Short History of Syriac Literature (Amsterdam 1966 [first published, 1894]), 205. The discourse published by Becker appears in Vat. Syr. 100 on fol. 126 9. For a detailed analysis of the contents of this manuscript see P. Zingerle, ‘Aus dem handschriftlichen syrischen Werke des Johannes von Dara über das Priestertum’, Theologische Quartalschrift (Tübingen 1867), 183 205, ibid, 1868, 267 85. 9 A. van Lantschoot, Inventaire des Mss. Syriaques des fonds Vatican (490 631), Barberini orientali et Neofiti, Studi e Testi 243 (1965), 108 10. 10 A. Mingana, Catalogue of the Mingana Collection of Manuscripts, now in the pos session of the Trustees of the Woodbrooke Settlement, Selly Oak, Birmingham, 1, Syriac and Garshuni Manuscripts (Cambridge 1933), 153 5. According to the catalogue, this manuscript was written at Mosul in 1902 in West Syriac script. It has 140 leaves of double columns with thirty two lines to the column. It is ‘the book of Iwannis (John) of Dara’. It is divided into eleven memrei, subdivided into kephalia. The sec tion published by Becker corresponds with ff. 40b 43a, the fifth memra, described by Mingana as ‘divided into two kephalia, on priesthood. It contains a controversy between a Jew and a Christian’. According to Mingana (155) the original MS. from which the copy was made is ‘some seven hundred years old’ and a note in Garshuni describes the state of the original manuscript. Its colophon has disappeared and it was somewhat renovated in 1744. See Breydy, ‘Les compilations’, 277. 11 A. Vööbus, ‘Die Entdeckung einer neuen Schrift des Mose bar Kepha über das Priestertum’, Ostkirchliche Studien 23 (1974), 324 7; idem, ‘New Manuscript Discoveries for the Literary Legacy of Mose bar Kepha: the genre of Theological Writings’, Harvard Theological Review 68 (1975), 377 84, but cf. Breydy, ‘Historisch literarische Daten’. 12 For example, Harvard Semitic Museum Nr. 4054 (olim Syr. 119) which is now Syr 122. See M.H. Goshen Gottstein, Syriac Manuscripts in the Harvard College Library (Harvard Semitic Studies 23, Missoula 1979), 87. Goshen Gottstein does not refer to Breydy’s articles, but only to one by Vööbus, but states that ‘a prelimi nary comparison suggests that this MS. is particularly similar to Vat 100 and Mingana 56’. 390
DISCOURSE ON PRIESTHOOD (BL ADD. 18295, FF. 137B 140B)
which include the text that Becker has published, and discusses other similar Syriac and Arabic manuscripts (including many modern copies of older manuscripts) that deal with the priesthood, a subject of some interest in medieval Syriac literature. He also critically evaluates the attributions of these works to the mid-ninth century bishop, Iwannis of Dara and to Mose bar Kepha (d. 903 CE). Since, to my knowledge, no part of this larger composition had hitherto been published, Becker has done a great service in making a small part of it more widely available. Adam Becker writes: ‘I thank Geoffrey Herman for his discovery. The identification he has made confirms some of my speculations about this text, while obviating others. The literature from the ninth and tenth centuries has received little attention compared to that of the fourth through sixth and much work remains to be done. I look forward to examining the larger text of which the Discourse on Priesthood is only a part.’
391
NOTA GENITIVI ZA- IN EPIGRAPHIC GEEZ MARIA BULAKH RUSSIAN STATE UNIVERSITY FOR THE HUMANITIES, MOSCOW
Abstract The paper discusses the use of the analytic genitive construction with nota genitivi za in Epigraphic Geez. This construction is attested in Geez inscriptions, though considerably less frequently than the syn thetic genitive construction. In most cases za is used when the syn thetic construction is impossible or at least undesirable due to mor phological or syntactic reasons. Nota genitivi za is sometimes attested in a genitive construction without overt head. There are also a few examples of nota genitivi za used apparently as a free variant of the synthetic construction. These data are mostly in accordance with the facts of Classical Geez. However, they differ in one point: whereas for Classical Geez a separate function of the genitive construction with za has been established, namely, to express the so called genitive of material, there is no reason to assume that za constructions had such a function in Epigraphic Geez.
In the present contribution, the function of the nota genitivi za- in Epigraphic Geez (EG) shall be discussed.1 Epigraphic Geez is understood here as a group of varieties of Geez that are present in a number of inscriptions found in the territories of modern Eritrea, northern Ethiopia, Sudan and Yemen and dating from the first millennium CE. The language of these inscriptions is very similar to the form of Geez taught in traditional Ethiopian schools and described 1
A short version of this paper was read at the 16th Conference of Ethiopian Studies in Trondheim, Norway, July 2007. I am grateful to Hamburg University for granting me a scholarship for the project ‘Lexicon and Grammar of Epigraphic Geez’ and to the Deutsche Forschungsgemeinschaft (German Research Founda tion) for sponsoring my participation in the 16th Conference of Ethiopian Studies at Trondheim. Sincere thanks go to my mentor L. Kogan and to S. Loesov for read ing and improving the draft version of this paper. I would like to express my deep gratitude to A. Bausi who kindly agreed to read the preliminary version and whose insightful comments prompted me to reassess certain points and to correct certain discrepancies. Last but not least, the generous help of Prof. G. Fiaccadori, who read and amended the paper, is highly appreciated. Needless to say, I am responsible for all errors. 393
NOTA GENITIVI ZA- IN EPIGRAPHIC GEEZ
in several grammars by both Ethiopian and European scholars: here this form will be called Classical Geez (CG). CG was apparently created on the base of one or several varieties employed in EG or closely related to it. Moreover, EG is the only direct evidence of original, non-translated texts produced in Geez by native speakers. In fact, the Geez literature of the Aksumite period consists of translations;2 moreover, these literary texts are preserved in manuscripts the earliest of which date back to the twelfth/thirteenth century. Therefore, a thorough investigation of the differences between EG and CG is highly important both for the history of Geez and for comparative Semitic studies in general. The present contribution focuses on the use of the analytic genitive construction with the marker za- in an effort to demonstrate that EG provides new evidence on the development of this construction in Geez. In Geez, there are three competing morphological devices for expressing the possessive relation between two nouns: one synthetic, with the marker -a attached to the head noun; and two analytic: 1) by means of the preposition la- attached to the dependent noun and a possessive suffix (co-referent with the dependent noun) attached to the head noun; 2) by means of the marker za- attached to the governed noun (cf. Dillmann 1907: 324–8, 426–8, 459–70, Lambdin 1978: 23–4, 44–5, 107; Weninger 1993: 33; Tropper 2002: 168–73; and especially the detailed analysis by Schneider 1959: 43–64; cf. also Diem 1986: 241–2). Of these constructions, the synthetic is the most widely used and has the most general semantics: apart from possessive relations proper, it marks many other relations between two nouns. Thus I prefer to speak of a genitive, rather than of a possessive construction.3 The analytic constructions are less frequently employed. The first of them, using the preposition la- and a pronominal suffix, seems only to be employed for possessive relations proper. For the construction with the element za-, there seem to be no semantic restrictions, although the tendency to mark the relationship of the type ‘object’ – ‘material’ has been observed (cf. Schneider 1959: 49, 56–9, Kapeliuk 1994: 68).4 2
For the recent progress in the retrieval and investigation of the Aksumite texts see Bausi 2006. 3 Still, the semantic diversity of the genitive lies within the scope of semantic developments of the possessive relation, as shown by N. Nikiforidou (1991). 4 Two other, more peripheral, means of marking possessivity, namely, 3) prepo sition la with no possessive suffix attached to the head, and 4) preposition ? ¢m¢nna, are not considered here since their semantics deviate from the genitive relation. The former expresses a relation close, if not identical, to the dative; the latter has a partitive meaning (see Dillmann 1907: 471). The description found in 394
NOTA GENITIVI ZA- IN EPIGRAPHIC GEEZ
All three types exist in EG. In this contribution, I shall focus on the analytic construction with the marker za-. At first, I will give a brief summary of its functions and its relations with the synthetic construction in the corpus of CG. Then I shall discuss the data provided by EG. Finally, I shall compare the data of these two sources and draw conclusions on the evolution of the analytic construction with za- in Geez. Some terminological clarifications are in order: I shall refer to the whole morpho-syntactic device as the ‘za-construction’ whereas the genitive marker za- shall be called ‘nota genitivi’; the term ‘genitive construction’ shall be applied to both the synthetic and analytical constructions; the dependent noun (whether marked with a nota genitivi or not) shall be referred to as a ‘genitive modifier’. Za-construction in CG Morphology In CG, the nota genitivi za- can be used as an invariable marker of possessivity, or alternatively can agree with the head noun: za- for masc. sg., ? ¢nta for fem. sg., and ? ¢lla for pl. (Dillmann 1907: 328; within the present approach, za-, ? ¢nta and ? ¢lla are treated as members of a single paradigm). No morphological, syntactic or semantic conditions for the presence/absence of agreement have been established so far. The nota genitivi is morphologically fully identical with the relative pronoun za-. Indeed, it is often said that the nota genitivi is a specific function of the relative pronoun, but such terminology is misleading, for it implies that the primary function is that of the relative pronoun.5 In order to avoid this, within the present paper a distinction shall be consistently drawn between relative and genitive za-. Weninger 1993: 33 is somewhat simplified: both la and za are treated as simple notae genitivi that may be accompanied by a pronominal suffix. This is illustrated with an example of the rare construction 3) instead of the widely used 1). 5 In fact, there are no obvious arguments in favour of such reconstruction: the cognates of the Geez za are prevalent in West Semitic languages both as relative and genitive markers. Nor do we find clear evidence for the reverse development, from genitive to relative constructions (as proposed, e.g, in Goldenberg 1995: 12). Typologically, relative pronouns and markers of the genitive are known to be able to derive almost simultaneously from the same (deictic) elements (see, e.g., Aristar 1991: 14 15). 395
NOTA GENITIVI ZA- IN EPIGRAPHIC GEEZ
Syntax For CG, Dillmann (1907: 468–70, with the examples quoted below) describes four cases in which the za-construction is employed: a) with a proper name as the head noun: beta l¢Ìem za-y¢hud ‘Bethlehem of Judah’ (Matt. 2:5); ? ¢m-nazret za-galila ‘from Nazareth of Galilee’ (Mark 1:9; see also Schneider 1959: 59–60);6 b) with a third element intervening between the head and the genitive modifier: a possessive suffix (Matt. 26:28: dam¢ya za-Ìaddis s¢r! at ‘my blood of the New Testament’), a second genitive modifier (non-coordinate with the former, as in Gen. 35:8: xaba ! ¢∂a balan ? ¢nta laÌ ‘by the oak tree of lamentation’; see also Schneider 1959: 60), or any other lexeme governed by the head noun;7 g) with a chain of coordinate genitive modifiers (Enoch 41:5: mazag¢bta ∂aÌay wa-za-warx ‘the treasuries of the sun and of the moon’); d) with no overtly expressed head (Matt. 9:32: za-ganen ‘possessed’, lit. ‘[that] of a demon’). 6 According to Dillmann, constructions with heads unable to distinguish be tween status constructus and status absolutus (namely, nouns ending in e, o, a) also belong to this group, as well as constructions with heads in the accusative. How ever, as demonstrated by Schneider (1959: 55 6), most za-constructions with such heads should in fact be classified under b or g. Likewise, no convincing arguments have been made to include za constructions with proper names as genitive modifi ers in this group (the example from Matt. 2:16 quoted by Dillmann [1907: 468] in fact belongs to group b). 7 Dillmann’s group b subsumes chains of genitive constructions, i.e. sequences of nominal lexemes in which each element is connected with the following one by the genitive relation. Dillmann provides no illustration for this phenomenon. The material collected by Schneider (1959: 62 4) shows, firstly, that the za construc tion is by no means obligatory in such chains (cf. such examples as wangela mang¢sta ? ¢gzi? abÌer ‘Gospel of the kingdom of God’, mazgaba xiruta l¢bb¢ka ‘treasure of goodness of your heart’). Secondly, Schneider demonstrates that, in the majority of such examples, the use of the za construction can be explained by other morphological, syntactic or semantic reasons. The only two valid examples quoted in Schneider 1959 are Ò¢nfa saÒun za ? ¢bn ‘the extremity of the box of stone’ (Grébaut 1945: 17) and ? a! mada ? ¢bn za par¢nso ‘pillars of the stone par¢nso’ (Esth. 1:6). These two examples can hardly be considered sufficient evidence to claim that za constructions were used to break long chains of possessive construc tions. Either one can classify them as examples of a za construction in the ‘neutral’ context (see below), or seek the explanation for the employment of two different morphological devices in the semantic differences between the chained genitive constructions ? a! mada ? ¢bn (relation of the type ‘object’ ‘material’) and ? ¢bn za par¢nso (relation of classification, in which the head noun acts as the classifier of the dependent noun) (the latter explanation suggested by A. Bausi, personal communi cation).
396
NOTA GENITIVI ZA- IN EPIGRAPHIC GEEZ
Thus according to Dillmann’s description, the za-construction is employed only as a substitute for the synthetic construction in positions where the use of the latter is impossible or at least undesirable8 on morphological (proper names or nouns with possessive suffix as head nouns) or syntactic grounds. Schneider’s extensive study shows that in some cases the zaconstructions can also be employed in such contexts where no syntactic or morphological complexity prevents the use of the synthetic construction, that is, where the head is not a proper name and governs only one genitive modifier which immediately follows it (for the sake of brevity, I shall refer to such contexts as ‘neutral’). In all such examples, Schneider believes, the genitive modifier designates material. (Indeed, the same function has been detected by Dillmann himself in his Lexicon Linguae Aethiopicae [LLA 1032], cf. Polotsky 1961: 451). Schneider’s calculations of the genitive constructions of the type ‘object’ — ‘material’ for three texts gave the following results: MaÒÌafa m¢s†ira samay wa-m¢dr — the za-construction is employed seventeen times, the synthetic construction four times; Synaxarium (Taxsas ) — the synthetic construction is employed twice (for one of these examples the za-construction is attested as a variant); Revelation of John — the za-construction is employed eleven times, the synthetic construction ten times. As a cursory perusal of Schneider’s examples shows, at least some of his examples may in fact be explained as syntactically conditioned according to Dillmann’s criteria listed above as a, b, g (e.g., Mark 1:6: wa-È¢natu za-? adim ‘and his belt of leather’, cf. group b), or represent other deviations from ‘neutral’ genitive constructions (e.g., Perruchon 1903: 6: ÌaÒÒ wawarawat za-? ¢sat ‘arrows and javelins of fire’, with two coordinate head nouns). However, there is no doubt that he was indeed able to detect a number of passages where the za-construction is employed in a ‘neutral’ context: maswa! t za-? ¢sat ‘sacrifice of fire’ (ibid. 5); tabot za-? ¢sat ‘tabot of fire’ (ibid. 5); ? afras za-? ¢sat ‘horses of fire’ (ibid. 11); ba-È¢nat za-warÈ ‘with a belt of gold’ (Rev. 1:13 — var. È¢nata warÈ); ? aklil za-warÈ ‘crown of gold’ (Rev. 14:14, var. ? aklila warÈ); ? aklilat za-warÈ ‘crowns of gold’ (Rev. 4:4, var. ? aklilata 8 Indeed, for group g, one cannot claim that the synthetic construction is im possible in the same environment: use of the syntactic construction for two co ordinate genitive modifiers governed by the same head is not prohibited (and, it seems, not even rare): takla Ò¢dÈ wa r¢t! ‘the plant of righteousness and truth’ (Enoch 10:16). Still, this usage of the synthetic construction can be regarded as a ‘deviation’ from the norm, and this obviously accounts for the frequent employ ment of the za construction in this environment.
397
NOTA GENITIVI ZA- IN EPIGRAPHIC GEEZ
warÈ); wa-gera za-xaÒin ‘and helmet of iron’ (Perruchon 1903: 16); dabtara za-? ¢sat ‘tabernacle of fire’ (ibid. 5); gw¢dba (acc.) za-? ¢sat ‘axe of fire’ (ibid. 11) (see Schneider 1959: 55–9).9 It seems that the number of ‘neutral’ examples can be expanded whereas Schneider’s thesis of their being semantically restricted (expressing the ‘object’ – ‘material’ relation only) may be questioned. The systematic investigation of CG material lies beyond the scope of the present study. However, some evidence can be gleaned even from Dillmann’s grammar (1907). Among Dillmann’s examples of za-constructions without a head noun, several cases (e.g., Judg. 16:29) are quoted where the head is in fact present in the text. Such passages are obviously regarded by Dillmann as a secondary ‘reintroduction’ of a governing noun, where the complement za + noun is attached in apposition, not as a member of the genitive construction. This interpretation may be true for a few lexicalized headless constructions, such as za-lamÒ ‘leper’ and za-ganen ‘possessed by a demon’, which are frequently used with no head noun, but can also occur with such a noun: b¢? ¢si za-ganen ‘a man possessed by a demon' (Mark 1:23) where the relation between b¢? ¢si and za-ganen can indeed be appositional rather than complemental. However, at least one of these examples (wa-! azaÈta [acc.] za-w¢Èro ‘and pits of digging’ in Deut. 6:11) is likely to be interpreted as a za-construction in a ‘neutral’ context.10 Thus, the employment of the za-construction in ‘neutral’ contexts is likely not to be restricted to the cases where the genitive modifier designates material (although one cannot deny, in agreement with Schneider [1959: 56, 59], that the latter type of example does constitute the majority of the za-constructions in ‘neutral’ contexts in CG). 9 Another example quoted by Schneider, Mark 9:47 (gahannam za ? ¢sat ‘Gehenna of fire’) is probably to be discarded since the word gahannam was likely regarded as a proper name. Admittedly, the use of the synthetic construction in Matt. 5:22 and 18:9 (gahannama ? ¢sat, quoted by Schneider 1959: 58) speaks against such a possibility. However, according to Zuurmond's edition (2001: 62 3, 362 3, 439), the majority of manuscripts have either ? ¢sata gahannam or gahannam ? ¢sat in this passage (the syntax of the latter expression is difficult to in terpret: an apposition? an Amharic genitive construction?). 10 Goldenberg (1995: 4 5), in agreement with Dillmann, sees in all such collo cations appositional relationship rather than genitive. Moreover, he analyses nota genitivi za (and its analogues in other Semitic languages) as pronominal head of a synthetic genitive construction. Such an analysis, though very profitable for our un derstanding of the origin of the analytic construction in question, seems hardly jus tified from the synchronic point of view. At least within CG, the paradigm of the nota genitivi differs from that of the corresponding demonstrative pronoun, z¢, and cannot be explained as merely the status constructus form of the latter.
398
NOTA GENITIVI ZA- IN EPIGRAPHIC GEEZ
Caquot (1952: 493–4), in his study on the syntax of the Book of Enoch, explains the use of the registered za-constructions mostly with the factors of Dillmann’s b and g groups. However, he also proposes a further factor, namely, the attraction of relative and genitive za-elements: ! ¢∂ za-†¢bab za-? ¢m¢nnehu bal! u ? abuka wa-? ¢mm¢ka ‘the tree of wisdom from which your father and your mother ate’ (Enoch 32: 6). Unfortunately, it remains unclear to what extent the latter phenomenon was spread. Later grammatical descriptions of CG contain no new information on the topic. Weninger’s (1993: 33) and Tropper’s (2002: 168– 73) presentations of za-constructions obviously rely on Schneider 1959 and stress the importance of the za-construction, firstly, as a substitute for the synthetic construction when the latter is impossible on syntactic grounds, and, secondly, as a means of expressing the ‘object’ – ‘material’ relationship. Lambdin (1978: 107, following Dillmann 1907) mentions these two aspects as well, but holds the headless construction with za- to be primary, supposing that over the course of time its usage was expanded to non-substantivized attributes as well. To sum up, the majority of the za-constructions in CG seem to perform one of the following three functions. The first, most widely attested function is to substitute for the synthetic construction where its use is impossible or at least undesirable, for morphological or syntactic reasons. The second function is to introduce the genitive construction with no overt head.11 The third is to express the ‘object’ – ‘material’ relation, although such a relationship can also be expressed by the synthetic construction. Examples of the za-construction to which none of these functions can be attributed are rare in CG. Za-construction in EG In EG, the za-construction is also used less frequently than the synthetic construction (although noticeably more frequently than the periphrastic construction with la- and pronominal suffix).12 Through a systematic perusal of the EG corpus, I have collected thirty-seven 11
In fact, the headless constructions can be regarded as once again a case of a morphosyntactically conditioned substitute for the synthetic construction: the lat ter is obviously impossible in the absence of an overtly expressed head noun. 12 For CG, Schneider (1959: 43 4) reports a somewhat different picture: while the synthetic construction is still the most frequent, it is followed by the periphras tic construction; the za construction is thus the least widely employed. 399
NOTA GENITIVI ZA- IN EPIGRAPHIC GEEZ
examples of za-constructions of various degrees of certainty. Among these, twenty-two cases are more or less secure whereas fifteen examples are rather unclear but may still be interpreted as za-constructions.13 This number is definitely lower than that of the more common synthetic constructions: thus, no less than ninety-three secure examples of the synthetic construction are manifest in vocalized inscriptions alone. Morphology The paradigm za- (sg. masc.) — ? ¢nta (sg. fem.) — ? ¢lla (pl.) known from CG is hardly employed in EG. While the number of possessive constructions with za- is not insignificant (36), we find no evidence of ? ¢nta as nota genitivi,14 and only one case of ? ¢(l )la in this function, in a late inscription from Ham, RIE 232: 5 (? ¢la / sah¢l, most probably, a headless genitive construction in the plural, ‘those of mercy’).15 This evidence is perhaps not sufficient for us to conclude that the nota genitivi had an invariable form za- in EG of the earlier period. As far as the feminine form ? ¢nta is concerned, there is no context in the available EG corpus in which its use (either as nota genitivi or as a relative pronoun) would be obligatory.16 The situation 13 Genitive constructions as a whole (rather than single attestations of the nota genitivi za ) were counted. Thus, ? aflag / za sida / wa za takazi ‘the rivers of Sida and Takazi’ in RIE 189: 39 40 was counted as one za construction. 14 The element ? ¢nta is attested in EG as a preposition (RIE 187: 26; RIE 187: 28; RIE 195: 12). Pennacchietti (1974: 189) and Kapeliuk (2003: 227 8) consider this preposition to be a development of the relative/genitive marker, assuming that its origin lies in the collocation ‘preposition + relative marker’, where in certain contexts the ‘zero’ preposition was used. However, I shall refrain from claiming the presence of the feminine relative pronoun and/or nota genitivi ? ¢nta in EG based on this hypothesis alone. 15 Probably with an animate referent, cf. Fiaccadori 1990: 327. Alternatively, Kropp 1999: 168 70, 173, translates it as ‘[the years] of mercy’. His interpretation, while offering a plausible deciphering of symbols read by previous scholars as ? ¢nen, involves a number of morphological and syntactic problems in what concerns the rest of the phrase (Òarana instead of expected Òar¢na, absence of syntactic relation between ! amat and Òarana, plural form of nota genitivi before sah¢l ). In fact, Fiaccadori’s and Kropp’s interpretations can possibly be combined, ? ¢nen being un derstood as ciphers referring to the year (with Kropp), but the subsequent words translated as suggested by Fiaccadori (thus, ! amat / 590 / Òaraxu / ? ¢la / sah¢l / wa baÌtu / ba kama / Ò¢Ìuf ‘year 590; those of mercy cried; however, as it is writ ten…’). 16 No relative clauses or genitive constructions with animate feminine heads are observable in EG. As for inanimate nouns, the epigraphic data supplies practically no evidence of their agreement, whereas in CG inanimate nouns have fluctuating gender agreement (see Dillmann 1907: 285 6).
400
NOTA GENITIVI ZA- IN EPIGRAPHIC GEEZ
with the plural form ? ¢(l )la is somewhat more complex. There are established examples of za- as nota genitivi with the head in the plural, namely RIE 189: 31–2 (? asmata / ? ¢hgurihu / za-n¢dÈ ‘the names of his towns of stone’), RIE 189: 35–6 (? ¢hgura / noba / z[a-]Ìasar ‘Noba’s towns of straw’), RIE 189: 36 (? ¢hgura / nadaÈ / za-kasu ‘towns of stone of Kasu’). In all these examples the head is inanimate. One can suppose that absence of concord for number in inanimate nouns, allowed in CG (see Dillmann 1907: 501), was also possible (or even obligatory?) in EG.17 Unfortunately, the absence of reliable examples for concord of inanimate nouns (except for the above-mentioned) prevents me from making any definite statements. There are two examples in which the absence of ? ¢(l )la seems to speak against this hypothesis, i.e., examples where the nota genitivi apparently depends on an animate head in the plural: ∂ewa / ! ¢d / wa-? an¢st / za-Ì¢zba / ga! aza ‘the captives of men and women [i.e. the male and female captives] of the people of Ga!aza’ (RIE 188: 19) and È¢tla / ! ¢d / za-? a[fa]n / 503 / wa-? an¢st / 202 ‘the killed of men of [the kingdom of ] ?Afan — 503 and of women — 202’ (RIE 188: 17–18).18 However, the probability that the notae genitivi are governed in both cases not by the preceding plural nouns, but by the nouns ∂ewa ‘the captives’ and È¢tl ‘the killed, the killed people’, respectively, is high. Further, one can suppose that the latter two nouns, being morphologically singular, required singular agreement. At the same time, one has to keep in mind that the relative pronoun za- is in fact seen (only once — in RIE 189: 11) as referring to the animate head in the plural. Assuming that the nota genitivi za- behaves in a very similar fashion to the relative pronoun, and taking into account the evidence of CG, it seems likely that the nota genitivi za-, as well as the relative pronoun za-, could be used in EG as an invariable marker, but this usage was — certainly for the relative pronoun and hypothetically for the nota genitivi — very peripheral. Taking into consideration that relative clauses with animate plural head nouns are not rare in EG, whereas the za-constructions with such heads are hardly present at all, one can suppose that the absence of 17 The relative pronoun ? ¢l(l)a seems to fit into this hypothesis since it appears in EG with animate heads only (cf. RIE 187: 9; RIE 187: 23; RIE 189: 23; RIE 189: 25; RIE 189: 27; RIE 192: 13; RIE 192: 14; RIE 192: 29; RIE 192: 30; RIE 192: 33; RIE 192: 37; RIE 192: 40; RIE 232: 14). No reliable examples of relative clause with an inanimate head in the plural are attested in EG. 18 The latter example is not so obvious since one can suspect that ! ¢d ‘men’ could have a singular agreement (for a similar concord of ? ¢Ìzab cf. RIE 189: 18: ? ¢Ìzab¢ya / za wafara ‘my people who went out [sg. masc.]’).
401
NOTA GENITIVI ZA- IN EPIGRAPHIC GEEZ
the plural nota genitivi ? ¢(l )la in the EG texts of the early Aksumite period can be explained in terms of the peculiarity of texts and therefore cannot serve as a strong argument for the actual absence of this element in the language of that period.19 Syntax The contexts in which the za-construction is attested in EG can be classified as follows. a. The za-construction frequently appears in a sequence of coordinate genitive modifiers (altogether twelve secure examples from long royal inscriptions plus one less clear from a short inscription). In seven of these examples the synthetic constructions and the za-constructions are combined: the first genitive modifier is introduced as a 19
The diachronic evidence (presence of cognate plural notae genitivi ? l in Epi graphic South Arabian and lV in Modern South Arabian, see Pennacchietti 1968: 35) suggests that ? ¢(l )la as the plural form of nota genitivi in Geez is archaic rather than innovative. Further evidence may be in the element ? ¢l(l )e found in the proper names ? ¢l(l)e ! amida (RIE 188: 1; RIE 189: 2; RIE 189: 4; RIE 186: 1) and ? l ? ÒbÌ (presumably, to be vocalized as ? ¢lle ? aÒb¢Ìa or ? ¢lla ? aÒb¢Ìa; RIE 191: 7 8; RIE 192: 7; for the Greek transcriptions of this onomastic element, showing the gemination of l, see Fiaccadori 2004: 109 10). These proper names can be un derstood as fossilized headless genitive constructions, where the plural form of geni tive marker may be interpreted as the polite form, and the final e as a result of al ternation a/ e appearing in some prepositions and conjunctions in EG (m¢sla / m¢sle ‘with’; ? ama / ? ame ‘when’; cf. Bausi 2005a; 2005b: 160 1). This hypothesis, however, encounters several difficulties, most of them acknowledged by Bausi. The most important one is the distribution between the forms ? ¢l(l)e (manifest as the onomastic element only) and ? ¢l(l )a (present as the relative pronoun only) in EG. A further argument against the identification of the onomastic ? ¢l(l )e and relative/ genitive ? ¢l(l )a is that the final e of ? ¢l(l)e is unlikely to be of the same origin as e of the prepositions/conjunctions. The final e of the prepositions/conjunctions is most probably an archaism: cf. adverbial and prepositional element i, e in Akkadian (von Soden 1952: 163 [§113 k]); prepositions in e in Hebrew (Gesenius 1962: 314 [§ 103 n o]), pre suffixal forms of prepositions in ay in Syriac (Nöldeke 1966: 98 100), etc. In view of the Akkadian evidence, the Hebrew, Aramaic and Geez prepositional endings cannot be explained as forms built under the influence of the preposition *! alay ‘on, upon’ (a view expressed in Brockelmann 1908: 497 [§ 252 e] and popular to date). Rather, all these forms can be traced back to a proto-Semitic adverbial case *-ay, reconstructed by von Soden (1952: 163). As for the final e of the Geez relative/genitive marker ? ¢lle (found in this form in ancient manuscripts, cf. Bausi 2005a: 8), it can hardly go back to the adverbial ending; possibly, it is a result of a later process of analogy, i.e., of the expanding of the alter nation a/ e. For alternative interpretations of the onomastic element ? ¢l(l )e cf. Lusini 2004: 70 1 (criticized in Fiaccadori 2004: 109 10, Bausi 2005a), where its formal similarity to the relative pronoun is considered to be purely coincidental. 402
NOTA GENITIVI ZA- IN EPIGRAPHIC GEEZ
second member of the synthetic construction (with the head in status constructus) whereas the subsequent modifiers are marked with za-: n¢gusa / ? aksum / wa-za / Ì¢mer / [wa-]za / raydan / wa-za / saba? … ‘king of ?Aksum and of ΢mer and of Raydan and of Saba? …’ (RIE 188: 2–3). Among these, in five cases all modifiers except for the first are marked with za- and in two cases za- is attached to only some of them (RIE 185 I: 1–2 and RIE 191: 8–10). However, the nota genitivi can be attached to the first genitive modifier as well, as in RIE 189: 29 (? aflag / za-sida / wa-takazi ‘the rivers of Sida and Takazi’). Five examples of this type are known and it is worth noting that in four of them the head governs only two genitive modifiers.20 b. The za-construction is also found in cases where the same head noun governs two (or more) genitive modifiers that are not coordinate (and thus the corresponding genitive relationships differ in their semantics): ? ¢hgura / noba / z[a-]Ìasar ‘Noba’s towns of straw’ (RIE 189: 35–6). Two clear examples of this type are known. This group can be expanded with two examples where a possessive pronoun is attached to the head: ? ¢hgurihu / za-n¢dÈ / wa-za-Ìasar ‘his towns of stone and of straw’ (RIE 189: 18–19). Two further cases which I include in this group require a more detailed discussion. The first passage is RIE 188: 19: ∂ewa / ! ¢d / wa-? an¢st / za-Ì¢zba / ga! aza ‘the captives of men and women of the people of Ga!aza’. At first glance, the nota genitivi za- marks a genitive modifier governed by two coordinate head nouns: men and women of the people of Ga!aza.21 However, an alternative analysis can be made, according to which the phrase Ì¢zba / ga! aza is in fact dependent on the head ∂ewa (the captives of the people of Ga!aza). Such an analysis seems preferable since whereas the syntactic genitive relation between ∂ewa and ! ¢d wa-? an¢st is more or less evident (being paralleled by the obvious genitive construction È¢tla ! ¢d), the semantics of this construction is not possessive but rather appositional or adjectival (‘the captives — men and women’ or ‘the male and female captives’). The 20 Synthetic constructions with several coordinate genitive modifiers are also used in EG: wa mazag¢bta / ? ¢kl / wa †u† ‘and storehouses of grain and cotton’ (RIE 189: 20); mlk / ? kswm / w Ìmr m / w rydn m … ‘king of ?Aksum and ΢mer and Raydan …’ (RIE 185bis I: 1, with a duplicate in RIE 185bis II: 1, cf. also RIE 185 II: 1). 21 In CG, according to Dillmann 1907: 465, the synthetic construction can be used in such cases, either with the modifier repeated after each of the head nouns (la nagada ? ¢sra? el wa la Ì¢zba ? ¢sra? el ‘to a tribe and a family of Israel’) or with the first head noun in status absolutus (la nagad wa la Ì¢zba ? ¢sra? el). In the former case, the modifier can be replaced with a possessive suffix after the second noun (la nagada ? ¢sra? el wa la Ì¢zbu).
403
NOTA GENITIVI ZA- IN EPIGRAPHIC GEEZ
nouns ! ¢d and ? an¢st function here in all probability as markers similar to, e.g., the Amharic gender markers wänd and set (in such expressions as wänd l¢®ˇ®ˇ ‘son’, set l¢®ˇ®ˇ ‘daughter’). A noun phrase in such a function can hardly act simultaneously as a head of another genitive construction. Similar difficulty is encountered in the course of the analysis of the passage RIE 188: 17–18: È¢tla / ! ¢d / za-? a[fa]n / 503 / wa? an¢st / 202 ‘the killed of men of [the kingdom of ] ?Afan — 503 and of women — 202’. One could suspect (in agreement with Dillmann 1907: 468) that za- in this instance is used to break the sequence of two genitive constructions in the hypothetical *È¢tla ! ¢da ? afan.22 However, as in the previous case, an analysis seems to be preferable where the genitive modifier za-? a[fa]n is directly dependent on the head È¢tl (the killed of [the kingdom of ] ?Afan). c. There is one — not very certain — example of a za-construction with a proper name as its head: mamas : walda : ? aba dani? el : za-Òoma! ta : mara? a ‘Mamas, son of ?aba Dani?el of the cave ? Mara?a’ (RIE 252: 1–2). The phrase za-Òoma! ta is probably governed by the proper name ? aba dani? el, but one cannot exclude that its head is in fact mamas (in which case this example should be classified under group b). Given the fact that in CG only a za-construction is possible when the head is a proper name, it may be that this was also the case in EG. d. Headless genitive constructions, not uncommon in CG, are rare in EG. Of six possible examples, the only certain case is found in the late inscription from Ham (RIE 232: 5: ? ¢la / sah¢l ‘those of mercy’). Another possible example of a headless za-construction can be found in the inscription from Safra (RIE 183 II: 5–6): z-mÈbl ‘of the offering one’ (so Drewes 1962: 50–1; alternatively: ‘of the gain’). However, the usage of the za-constructions in this particular inscription seems to differ from the majority of the EG corpus. The re22
Let us observe that the use of the synthetic genitive construction in genitive chains is common in EG: [ba ]xayla / ? ¢gzi? a / samay ‘by the might of the Lord of Heaven’ (RIE 189: 1), sm > wld / ? gz? / bÌr ‘name of Son of God’ (RIE 191: 36) etc. One would have to seek the explanation for the presence of the nota genitivi za in the semantic difference between the two genitive constructions: È¢tla / ! ¢d and ! ¢d / za ? a[fa]n. Alternatively, one could suppose that the choice of za construction instead of the synthetic construction is conditioned by the complex, non linear, syntactic structure of the phrase: ? an¢st is obviously coordinate with ! ¢d and de pendent on È¢tla. The modifier za ? a[fa]n is inserted between two coordinate modifiers. The function of za construction in this case is possibly to distinguish one level of dependent nouns from another (and, perhaps, to stress that za ? a[fa]n, syntactically dependent on ! ¢d, logically modifies both ! ¢d and ? an¢st). 404
NOTA GENITIVI ZA- IN EPIGRAPHIC GEEZ
maining four examples are rather less certain, found in fragmentary or short inscriptions of obscure meaning where they appear as holophrastic expressions out of context (e.g., RIE 377: z-ÒrÌ, possibly ‘[vase] of the temple/palace’). Two of these examples are not vocalized which makes the identification of z- with the nota genitivi za- all the more uncertain. Moreover, three of them can be interpreted as proper names. In this connection, mention must be made of the proper name *za-y¢Ìann¢s (RIE 256: ? a[na] za-y-Ìan¢s-; RIE 260: 1–2: z-yÌns), where the interpretation of za- as an element identical to the nota genitivi za- appears likely a priori (for the interpretation of the onomastic element ? ¢(l)le as having a similar function see note 19).23 e. A ‘neutral’ context can be supposed for eleven examples but only one of them can be considered more or less reliable. It comes from the inscription RIE 200: z¢-? ¢bn : gabgab : za-lazen ‘this stone is the hall (alternatively, corvée) of [the clan of?] Lazen’ (for the variants of reading and interpretation see D.H. Müller 1894: 56; DAE p. 48; W.W. Müller 1972; Fiaccadori 2004: 122). Although various interpretations were suggested for the lexeme gabgab, the syntactic analysis of the text seems to be more or less clear: the phrase z¢-? ¢bn (read by W.W. Müller 1972 as za-? ¢bn) is the subject, the phrase gabgab za-lazen is the predicate, where the proper name lazen indicates the possessor of the gabgab. Note that if this analysis is correct, this is indicative of a true possessive construction, unlike the usage ‘object’ + ‘material’ in CG, where the genitive modifier is adjectivelike rather than possessor-like. A less reliable example is RIE 193 II: 6: b-Ì-r- : z-Ì-s-l- ‘land of Îsl’, where the text can be discerned on Littmann’s copy (DAE 13) but could not be read by the editors of RIE. Other cases of the za-construction in a ‘neutral’ context come from unvocalized early inscriptions, for which only tentative interpretations have been suggested so far, and can hardly be used as solid evidence. The obscure nature of the pertinent lexical items often prevents one from establishing whether the element z- is a nota genitivi or a relative pronoun introducing the verb of a subordinate clause. In any case, since seven examples come from the Safra inscription, the frequent use of z- as nota genitivi in a ‘neutral’ context (more or less certain at least in some of the cases, see Drewes 1962: 45) may be 23
A number of other short or broken inscriptions with initial z were not in cluded since it is impossible to tell whether the element z belongs to the root, or is a nota genitivi, or a relative pronoun. 405
NOTA GENITIVI ZA- IN EPIGRAPHIC GEEZ
considered as one of the grammatical peculiarities of this inscription. Interestingly, there are at least two possible employments of z- as nota genitivi in a ‘neutral’ context in two inscriptions published by L. Ricci, the palaeography and vocabulary of which suggest a similarity to the Safra inscription. An interesting question is, of course, whether synthetic constructions were also used in these texts. This question is not easy to answer, since the inscriptions are unvocalized, so that synthetic constructions are extremely difficult to discern. Note, however, that Drewes (1962: 35–6, 48) analyses several combinations of (presumably nominal) lexemes in the Safra inscription as being synthetic genitive constructions: RIE 183 I: 21–2 (w-gbt / ms ‘a gabata [as a measure unit] of hydromel’), RIE 183 IV: 6–7 (whb sl† ‘the giver of oil’), RIE 183 IV: 7–8 (w-whbt / ! m ‘the giver (fem.) of the wood’). The formation of the independent possessive pronouns in CG (zi? aya ‘mine’, zi? aka ‘thine’ etc.) is undoubtedly related to the zaconstruction: the elements to which the possessive suffixes are attached are very similar to those forming the paradigm of the nota genitivi (masc. sg. zi? a-, fem. sg. ? ¢nti? a-, pl. ? ¢lli? a-). However, combinations of these elements with possessive suffixes do not exhibit the same syntactic characteristics as combinations of the nota genitivi with nominal genitive modifiers. Indeed, independent possessive pronouns can be used in contexts where morphosyntactically conditioned za-modifiers appear: ba? ata ka! bat ? ¢nti? ahu ‘his double cave’ (Gen. 23:9; lit. ‘his cave of double amount’), and are frequently used in headless za-constructions: zi? aka y¢? ¢ti mang¢st ‘thine is the kingdom’ (Matt. 6:13), tas¢? ¢l¢wo ? ¢lli? ahu ‘his [followers] asked him’ (Mark 4:10). However, when used as a genitive modifier in a ‘neutral’ context, they either follow the head noun in status constructus or are marked with an additional nota genitivi: b¢? ¢sita ? ¢nti? aka or b¢? ¢sit za-? ¢nti? aka ‘the wife of thine’ (Dillmann 1907: 342). The only example of the corresponding construction in EG comes from an early non-vocalized inscription, RIE 218. One can tentatively interpret the passage in question as a nominal sentence, where the syntactic relation between the noun and the pronoun is predicative (RIE 218: 2–3: z-Ìwlt / z? h — to be vocalized as za/z¢-Ìaw¢lt zi? ahu ‘this stela is his’, cf. Kropp 2006). In this case, motivation for the employment of the analytic attachment of the possessive pronoun would be self-evident: neither the dependent noun of a synthetic construction, nor a possessive suffix can be used to denote a predicative possessive. An alternative interpretation is possible, treat406
NOTA GENITIVI ZA- IN EPIGRAPHIC GEEZ
ing this phrase as a part of a larger sentence: z-Ìwlt / z? h / sÌbw / ? gb / Ìzb — to be tentatively vocalized as za/z¢/za-Ìaw¢lta zi? ahu saÌab¢wa/wo ? gb Ì¢zb(u) ‘(his) people drew this stele of his to ? gb [i.e. of the king]’ (while still another interpretation regards the phrase z-Ìwlt / z? h as a direct object of the verb ÒÌf, cf. Conti Rossini 1942: 28, Littmann 1952: 6, Drewes 1962: 67). In any case, there is no reason to suppose that the use of the construction in EG was different from CG. Therefore, I prefer to treat independent possessive pronouns as constructions different from the za-constructions (although obviously related to them). Summing up, one can separate the use of the za-construction in EG into three groups. The first group, which includes the majority of clear examples of the za-constructions (groups a-c), consists of context-dependent variants of the synthetic construction — although one has to admit that the distribution of the synthetic and za-constructions is not strict. The second group is represented by the za-constructions without an overt head (group d). Headless za-constructions are few in EG: in fact, only one of them is more or less certain (though it, admittedly, comes from a late inscription). It is the CG evidence that makes it possible to interpret at least some earlier examples as headless za-constructions. The third group includes attestations of the za-constructions in ‘neutral’ contexts (group e). Only one reliable example from this group is known (RIE 200). Seven potential examples come from the Safra inscription and probably are to be regarded as a grammatical peculiarity; the possible occurrence of the element z- in the same function in two inscriptions published by L. Ricci in 1991 apparently points to their linguistic proximity to the Safra inscription. The picture attested in EG does not differ acutely from what is recorded for CG. Indeed, in some cases the data from EG and CG corroborate each other. Thus, the fact that no obvious examples of the za-construction substituting for the synthetic construction in linear chains of genitive constructions are found in either EG or CG leads me to seek an alternative explanation of the only unconvincing case (cf. group b). Similarly, we find very few examples in CG (see Schneider 1959: 56) and no examples in EG of a dependent noun preceding the head noun24 (which, inter alia, leads me to reject the 24
Littmann (DAE, p. 81) registers za constructions with modifiers preceding their heads in two inscriptions: DAE 40 (= RIE 388) and DAE 47 (= RIE 394). Both inscriptions are in fact monograms, which were not taken into consideration in the present study. It seems, in any case, that these two pieces of evidence are also 407
NOTA GENITIVI ZA- IN EPIGRAPHIC GEEZ
theoretically possible, but in fact very problematic interpretation of za-? ¢bn as a genitive attribute of gabgab in RIE 200). From the obvious presence of headless za-constructions and of za-constructions in ‘neutral’ context in CG one can postulate the existence of such constructions in EG also, despite the scarcity of convincing examples. The most important outcome of the present investigation is that the frequent use of the za-construction to express the ‘object’ – ‘material’ relation, established by Schneider for CG, is not supported by the evidence of EG. We have only one example of a za-construction with this function, namely RIE 185 II: 19–20 (msl / z-wrÈ / ? Ìd / 1 / w-z-brr / ? [Ì]d / 1 / w-z-ÒrÈ / slst / 3 ‘statue(s) of gold — one 1, and of silver — one 1, and of copper — three 3’, with duplicates in RIE 185 I: 18–19; RIE 185bis I: 20–1; RIE 185bis IIC: 27–31). However, the presence of the nota genitivi za- is more likely to be explained here on syntactic grounds — presence of three genitive modifiers depending on the same head noun (see group a); the syntactic structure is rendered somewhat unclear by the intervening numerals (also depending on the head noun msl) after each modifier. There are no ‘object’ – ‘material’ collocations among the za-constructions in ‘neutral’ contexts attested in EG. On the contrary, there are at least five secure examples of the synthetic construction employed with this function: RIE 189: 27 (È¢dada / b¢rur ‘an ornament of silver’); RIE 189: 27 (wa-Ì¢Èata / warÈ ‘and an ornament of gold’); RIE 189: 31 (? ¢hgura / n¢dÈ ‘towns of stone’); RIE 189: 36 (? ¢hgura / nadaÈ ‘towns of stone’); RIE 189: 40 (hagara / n¢dÈ ‘town of stone’). One can thereby claim that the frequent use of the zarather problematic: in DAE 40 (= RIE 388) (za y Ìa n¢s ga br, according to Littmann, ‘slave of Y¢Ìann¢s’) one can alternatively analyse za y Ìa n¢s as a proper name (cf. RIE 256: ? a[na] za y Ìan¢s , most probably ‘I am Za Y¢Ìann¢s’; cf. also RIE 260: 1 2: z yÌns). Littmann’s reading and interpretation of DAE 47 (= RIE 394) is far from certain (the editors of RIE, p. 485 read it in a different way but, unfortunately, provide no translation). Littmann’s statement that relative clauses could precede their heads is likewise not very convincing. The passage quoted by Littmann (DAE 11 = RIE 189: 40 1: za wahabani / ? ¢gzi? a / samay / ∂ewa …) can, in fact, be interpreted as a predicative construction: ‘[that] which the Lord of Heaven gave me [is:] the captives…’). The same analysis is possible in another ex ample from the same inscription, RIE 189: 25 6: wa ? ¢la / moto / magabt danokwe / 1 / dagale / 1 / ? anakwe / 1 / Ìaware / 1 / karkara / 1 / marihomu / 1 / … ‘those who died [are:] the chiefs Danokwe (1), Dagale (1), ?Anakwe (1), Îaware (1), Karkara (1), their priest (1) …’ (The alternative interpretation, ‘the chiefs that died [are] Danokwe (1), Dagale (1), ?Anakwe (1), Îaware (1), Karkara (1), their priest (1) …’ is less likely since it implies that the priest, mari, was included among the persons called magabt contrary to RIE 189: 27 8.) Adjectival modifiers follow their heads in the majority of contexts in EG. 408
NOTA GENITIVI ZA- IN EPIGRAPHIC GEEZ
construction to express the ‘object’ – ‘material’ relation is a late development. Further research on the za-constructions in texts from different stages of the Aksumite period will probably shed light on its chronology. At present, one can only remark that examples quoted by Schneider (1959: 59) from the Revelation of John allow us to date the emergence of this construction to the late Aksumite period at least — but only if we assume that it is the za-constructions and not the synthetic constructions (which are present as manuscript variants for the passages in question, cf. Hofmann 1967) that date back to the original Ethiopian text. A further question is the diachronic perspective. Which of the three above-mentioned functions are more archaic and which are innovative? Lambdin (1978: 107) reconstructs the development of the za-constructions as follows: from the most archaic headless constructions, the constructions of the type ‘object’ – ‘material’ developed, where the original appositional use of the dependent element is reinterpreted as a genitive modifier, with its subsequent spread to other cases, including the morphosyntactically conditioned usage. This picture can hardly be true since, as we have seen, no ‘object’ – ‘material’ usages are attested in EG, whereas both the first and the third element of Lambdin’s chain of development can be found in EG. At the same time, the hypothesis that the headless genitive constructions are the origin for other types of za-construction (or its analogues in other Semitic languages) seems highly promising. Such an approach can be found in Goldenberg (1995: 4–5), where all ‘neutral’ contexts are regarded as secondary, being analysed as headless constructions annexed through appositional relationship to a co-referent substantive. Moreover, Goldenberg analyses the headless constructions as synthetic genitive constructions with pronominalized heads. It is questionable whether such an analysis can be applied to the corresponding constructions in Semitic on a synchronic level (especially in Modern South Arabian languages, where the analytic construction corresponding to the Geez za-construction has become the main device to express the genitival relation). As for the diachronic aspect, one can claim with certainty that the source of nota genitivi za- in Geez and its cognates elsewhere in Semitic lies in the demonstrative pronoun *∂V- (see, e.g., Pennacchietti 1968: 147–51). The path of development, as reconstructed by Goldenberg, is highly plausible, but still not proved. Cognates of the Geez nota genitivi in other Semitic languages are attested both as genitive markers in ‘neutral’ contexts and as markers of headless genitive constructions. These two functions seem to be equally well represented throughout Semitic. The material 409
NOTA GENITIVI ZA- IN EPIGRAPHIC GEEZ
provided by Pennacchietti’s survey (1968) of the Semitic genitive constructions employing elements *∂V- or *sV does not show preference for this or that function in Semitic. Pennacchietti strives to prove that the headless genitive constructions with *∂V- or *sV are present throughout Semitic, whereas the respective genitive constructions with heads are missing in several languages. He still considers the unrestricted usage of *∂V-/*sV as nota genitivi to be original: in his view, the restricted use of the nota genitivi in Amorite, Dedanitic, Tigre, Tigrinya, Phoenician, Classical Arabic, and most modern Arabic dialects (where he registers only headless genitive constructions) is secondary and due to the emergence of a prepositional definite article (Pennacchietti 1968: 71–93). This hypothesis can hardly be upheld since definite articles preceding the respective nouns have been found in Modern South Arabian languages, alongside the unrestricted *∂Vconstructions (Simeone-Senelle 1997: 412). Nevertheless, one should refrain from automatically reconstructing the headless *∂V-constructions as though they were the original type of *∂V-construction in Semitic. The reason is that while claiming their universal presence in Semitic, Pennacchietti does not properly distinguish between the ‘classical’ headless genitive constructions with *∂V- (or *sV) functioning as nota genitivi, and the constructions where the element *∂V- (or *sV) is followed by a preposition (as a rule, *lV-). In the latter constructions, however, the original function of *∂V- (or *sV) seems to have been that of a relative pronoun (whereas the preposition functioned as nota genitivi): Judaic Aramaic naÌtoma ∂i-l-malka, *‘the baker who is of the king’ (N.baker REL-GEN-N.king) > ‘the baker of the king’ (cf. Pennacchietti 1968: 20, 56). Thus, if one looks closely at the material collected by Pennacchietti, one has to state that in Biblical Aramaic and Samaritan, the elements *∂V- in headless genitive constructions are followed by prepositions and can thus in fact be analysed as original relative pronouns. (In these languages the headless genitive constructions are represented only by cases where the genitive modifier is a possessive pronominal suffix, the whole construction functioning as an independent possessive pronoun: Samaritan dylk [dilak] ‘yours’, see Pennacchietti 1968: 21). Therefore, these constructions can hardly be counted among examples of *∂V- as nota genitivi without an overt head. It should not be excluded that both the headless genitive constructions and the ‘neutral’ genitive constructions with nota genitivi *∂V- were used as early as in Proto-Semitic (or at least in Proto-West-Semitic).25 25
If the nota genitivi was originally used in ‘neutral’ contexts, one should prob 410
NOTA GENITIVI ZA- IN EPIGRAPHIC GEEZ
As for the morphosyntactically conditioned analytic genitive constructions, I am only aware of its being postulated for Ugaritic (alongside the other two types, see Tropper 2000: 847–9) and in late Old Aramaic (Hug 1993: 94–5). However, in both cases the picture is not exactly identical to that of EG.26 Of course, in other Semitic languages the nota genitivi *∂V- can likewise appear in syntactically complex surroundings (e.g. Biblical Aramaic ma? nayya di-bet ? elaha di-dahaba w¢-kaspa, ‘the vessels of the house of God of gold and silver’, Ezra 5:14). However, to the best of my knowledge, no specific research has been done to answer the question whether such use in Biblical Aramaic or elsewhere in Semitic prevails over the usage of *∂V- in ‘neutral’ contexts and whether the synthetic constructions are forbidden/rare in the same context. The fact that the independent possessive pronouns in Geez can be employed as headless constructions and as syntactically conditioned constructions, but are never used in ‘neutral’ contexts, seems to indicate that the first two functions are interrelated in Geez. Comprehensive research on the relation between the za-constructions and the synthetic constructions in Semitic would shed more light on this matter.
ably seek for semantic or pragmatic factors motivating the distribution between synthetic and *∂V constructions. The choice between two types of genitive con struction based on a semantic or pragmatic factor is typologically plausible (cf. e.g. Schuh 1982 for Chadic languages) and has been proposed within Semitic for Bibli cal Aramaic by Garr (1990). As we have seen, the semantic (‘object’ ‘material’) motivation for the choice between the synthetic construction and za construction in Geez is a late innovation. Interestingly, the same motivation is suggested by Tropper (2000: 847) for Ugaritic: ‘der Gen. des Materials wird häufig mittels d(t) eingeführt’, but he quotes only two passages where such a relationship is present in a ‘neutral’ context (gpnm dt ksp, ‘harness of silver’ and rqd d sn, ‘rqd instruments of ivory’ [according to DUL 832, mrqd d sn, ‘?castanets of ivory’], alongside the syn thetic construction in bht ksp w xrÒ, ‘a building of silver and gold’), which makes it desirable to introduce further arguments supporting this statement. 26 Thus, in Ugaritic, unlike in EG, there are passages where the genitive modi fier precedes its head (Tropper 2000: 847f., 91.321a) or where the analytic con struction is used to break the chain of genitive constructions (91.321c). Conversely, the examples with several modifiers (xpn d iqni w smt , ‘xpn garment of violet and red purple’, w slm d Ìwtk [w d] b[t ml]k[k], ‘the well being of your land and of ? your royal palace’) are relatively rare and not acknowledged by Tropper as a sepa rate group (neither are such cases registered by Hug for late Old Aramaic). Of spe cial interest are examples collected by Tropper under 91.322 and analysed as head less genitive d constructions in apposition to proper names, which can be compared to the usage of the corresponding constructions in Geez and in late Old Aramaic with proper names as heads. 411
NOTA GENITIVI ZA- IN EPIGRAPHIC GEEZ
Appendix The analytic genitive construction with za-, present in the corpus of EG27 a. The za-constructions with several coordinate genitive modifiers NPHead NP NPGenitive modifier
NHead noun
(za +) NGenitive modifier1 + (za +) NGenitive modifier2 + … (za +) NGenitive modifierN
RIE 186: 1 4: n[g]s-m / ? ksm-m / w-z-m / Ìmr-m / w-z-m / rydn[-m] / wz-m / s[b]? -m / w-z-m / slÌn-m / w-z-m / Òym-m / [w-]z-m / bg-m / w-zm / ks-m ‘king of ?Aksum and of ΢mer and of Raydan and of Saba? and of SalÌen and of ∑¢yamo and of B¢ga and of Kasu’ RIE 188: 2 4: n¢gusa / ? aksum / wa-za / Ì¢mer / [wa-]za / raydan / wa-za / saba? / wa-za / salÌen / wa-za / Ò¢yamo / wa-za / b¢ga / wa-za / kasu ‘king of ?Aksum and of ΢mer and of Raydan and of Saba? and of SalÌen and of ∑¢yamo and of B¢ga and of Kasu’ RIE 189: 2 4: n¢gusa / ? aksum / wa-za / Ì¢me[r] / wa-za / raydan / wa-za / saba? / wa-za / salÌen / wa-za / Ò¢yamo / wa-za / b¢ga / wa[-za /] kasu ‘king of ?Aksum and of ΢mer and of Raydan and of Saba? and of SalÌen and of ∑¢yamo and of B¢ga and of Kasu’ RIE 187: 3 (with broken beginning): …] wa-za / b¢ga / wa-za / kasu ‘… and of B¢ga and of Kasu’ RIE 192: 5 7: ngs / [? ]ksm / w-∂-Ìmyr-m / w-∂-rdn / w-∂-sb? / w-∂-slf / w∂-bg / w-∂-ks / w-∂-Òym / w-∂-wytl ‘king of ?Aksum and of ΢mer and 27 The following potential examples were not included in the statistics: RIE 190: 6: w km z m / bÒÌkw m / ∂ m / dm obscure passage. RIE 192: B2: w mrt / ∂ †yb / ∂ ? mh / w[…] obscure passage. RIE 193: 29 30: dwdm : z bk[?] obscure passage. Ricci 1988: A a) 1: z wf… obscure passage. Ricci 1988: A a) 2: z wÈf obscure passage. Ricci 1988: F:16: w dll / (?) z ! lr/s obscure passage. RIE 185bis II: 10 11: xlÈ sb? m z-sdst ngst 4400 […] ‘the number of the people of the six kings’ in three other versions of the text (RIE 185 I, II, RIE 185bis I), distributive zz (to be vocalized as zaz(z)a ) is employed in the corresponding pas sages. Thus, this example was discarded as a misspelling instead of zz. If it were to be analysed as a genuine case of za construction, it would have to be vocalized as sab? omu za s¢d(d)¢stu nagast and interpreted as a construction with a pronominal suffix and genitive marker za (instead of usual la ). RIE 218: 2 3: z Ìwlt / z? h ‘this stela is his’ the usage of independent posses sive pronoun zi? ahu is considered here as a construction which is not identical to za construction, although related.
412
NOTA GENITIVI ZA- IN EPIGRAPHIC GEEZ
of Rdn [Raydan] and of Saba? and of Slf and of B¢ga and of Kasu and of ∑¢yamo and of Wytl’ RIE 185 I: 1 2: mlk / ? ksm-m / w-∂-Ìmr-m / w-rydn-m / w-Ìbst-m / w-sb? m / w-slÌ-m / w-Òym-m / w-ks-m / w-bg-m ‘king of ?Aksum and of ΢mer and Raydan and Îbst and Saba? and S¢lÌen and ∑¢yamo and Kasu and B¢ga’ RIE 191: 8 10: ngs / ? ksm / w-Ì[mr /] w-z-rydn / w-sb? / w-slÌn / w-†dm / w-z-ymnt / w-thmt / w-Ì∂rmwt / w-kl / ! rb-m / w-z-bg / w-nb / w-z-ks / w-Òym / w-z-drbt ‘king of ?Aksum and ΢mer and of Raydan and Saba? and SalÌen and ™dm28 and of Ymnt29 and Thmt30 and Îa∂ramawt and of all Arabs and of B¢ga and Noba and of Kasu and ∑¢yamo and of Drbt’ RIE 189: 31: ? ¢hgura / n¢dÈ / wa-za / Ìasar ‘towns of stone and of straw’ RIE 189: 28 9: ba-maxbarta / ? aflag / za-sida / wa-takazi ‘at the confluence of the rivers of Sida and Takazi’ RIE 189: 39 40: w¢steta / maxbarta / ? aflag / za-sida / wa-za-takazi ‘in the confluence of the rivers of Sida and of Takazi’ RIE 185 II: 19 20: msl / z-wrÈ / ? Ìd / 1 / w-z-brr / ? [Ì]d / 1 / w-z-ÒrÈ / slst / 3 ‘statue(s) of gold one 1, and of silver one 1, and of copper three 3’ (also in duplicate inscriptions: RIE 185 I: 18 19: msl / ∂-wrÈ-m / w-z-brwr / [? ]Ìd[-m…] ‘statue(s) of gold and of silver one’; RIE 185bis I: 20 1: m†l / ∂-wrÈ / 1 / w-∂-br[w]r / 1 / w-∂-ÒryÈ / slst ‘statue(s) of gold 1, and of silver 1, and of copper three’; RIE 185bis IIC: 2731: msl / z-wrÈ / ? Ìd / 1 / w-z-brr / ? Ìd / 1 / w-z-ÒrÈ slst / 3 ‘statue(s) of gold one 1, and of silver one 1, and of copper three 3’) RIE 191: 30: Ètl / z-? gzt / w-Ìst ‘the killed of ?Agwezat and Îst’ ? RIE 197: 1 3: dwl z-smsmy w-z-sb† ‘region of Smsmy and ∑b†?’ (cf. Drewes Schneider 1967: 96)
b. The za-constructions with several non-coordinate genitive modifiers NPHead NP
NPGenitive modifier
NHead noun NGenitive modifier1
za + NGenitive modifier2
RIE 189: 35 6: ? ¢hgura / noba / z[a-]Ìasar ‘Noba’s towns of straw’ RIE 189: 36: ? ¢hgura / nadaÈ / za-kasu ‘towns of stone of Kasu’ ? RIE 188: 19: ∂ewa / ! ¢d / wa-? an¢st / za-Ì¢zba / ga! aza ‘the captives of men and women of the people of Ga!aza’ ? RIE 188: 17 18: È¢tla / ! ¢d / za-? a[fa]n / 503 / wa-? an¢st / 202 ‘the killed of men of [the kingdom of ] ?Afan 503 and of women 202’ 28
< ESA *™awd M ‘highland’. < ESA *Yamnat/Yamanat. 30 < ESA *Tihamat ‘coastal areas’. 29
413
NOTA GENITIVI ZA- IN EPIGRAPHIC GEEZ
Two cases in which the first genitive modifier is expressed through a possessive pronoun can be included here as well: RIE 189: 18 19: ? ¢hgurihu / za-n¢dÈ / wa-za-Ìasar ‘his towns of stone and of straw’ RIE 189: 31 2: ? asmata / ? ¢hgurihu / za-n¢dÈ ‘the names of his towns of stone’
c. The za-constructions with proper names as head nouns NPHead NP
NPGenitive modifier
NProper name
za + NGenitive modifier
? RIE 252: 1 2: mamas : walda : ? aba dani? el : za-Òoma! ta : mara? a ‘Mamas, son of ?Aba Dani?el of the cave ?Mara?a’
d. The headless za-constructions NPHead NP
NPGenitive modifier
Ø
za + NGenitive modifier
RIE 232: 5: ? ¢la / sah¢l ‘those of mercy’31 ? RIE 183 II: 5 6: z-mÈbl ‘of the ?offering one’32 ? RIE 257: za-? angab33 ? RIE 337: za-? ¢taya34 31 Translation according to Fiaccadori 1990: 327; cf. also Kropp 1999: 168 70, 173. 32 Translation according to Drewes 1962: 50 1 (alternatively: ‘of the ?gain’). 33 Fragmentary inscription on a rock. Since it is found among monograms and other short inscriptions involving proper names, one might also suspect a proper name. The source of this name, however, would undoubtedly be a headless genitive construction (for genitive constructions as source for proper names cf. such Ethio pian names as za y¢Ìann¢s, za mika? el etc.): cf. Littmann (DAE 85, p. 72), who connects the element ? angab with the place name Aggabé attested in Adulitana (cf. RIE 277: 5; cf. also place name ? ¢ngabo in RIE 187: 5) and with the Geez word ? angabenawi ‘noble man’ (LGz. 29; LLA 1405), attested in RIE 263: 4 (also RIE 195 II: 15 as ? angabenaya and in RIE 189: 25 as ? ¢ngabenawe). Alternatively, the whole expression can be interpreted not as a proper name, but rather as a reference to a person's origin: ‘[he] of ?Angab’ or ‘[he] of the noble’. 34 Fragmentary inscription on the bottom of a potsherd; no translation is of fered in Schneider 1966. One can surmise that either ? ¢taya or za ? ¢taya is a proper name. In the first case, the whole phrase could be analysed as a headless genitive construction (‘[pot] of ? ¢taya’). In the second case, the proper name can be ana lysed as originating from a headless genitive construction (cp. note 33). Admittedly, the final vowel a is strange in a noun, even in a proper name. To explain this, an alternative hypothesis can be suggested, namely, interpretation of ya as a possessive
414
NOTA GENITIVI ZA- IN EPIGRAPHIC GEEZ
? RIE 377: z-ÒrÌ35 ? RIE 379: z-brh36
g. The za-constructions that are not syntactically or morphologically conditioned (‘neutral’ context) NPHead NP
NPGenitive modifier
NHead noun
za + NGenitive modifier
RIE 200: z¢-? ¢bn : gabgab : za-lazen ‘this stone is hall / corvée of [the clan of?] Lazen’37 ? RIE 193 II: 6: b-Ì-r- : z-Ì-s-l- ‘land of ?Îsl’38 ? RIE 183 I: 1 4: tzkr z-swt ‘memorandum of [= concerning] ?food’39 ? RIE 183 I: 3 4: Ìg z-dwly ‘law of my region?’40 suffix 1 sg., attached to an unknown element (za )? ¢ta . The analysis of the whole phrase as a headless genitive construction would also be possible in this case. 35 Inscription on a vase, in palaeo Ethiopian letters. Drewes (2000: 202) regards it as a proper name with unknown etymology. However, it is not impossible to ana lyse it as the nota genitivi and the noun ÒrÌ. The latter is also attested among the short inscriptions on the walls of the so called ‘Gabra Masqal’s tomb’, see RIE 212: 49: ÒrÌ; cf. CG Ò¢rÌ ‘chamber, fortress, palace, temple (LGz. 563; LLA 1273). It is thus quite possible that the inscription on the vase indicated its owner ‘[vase] of the temple/palace’. Note, however, that the meanings ‘temple’ or ‘palace’ are not attested in Epigraphic South Arabian cognates of this term (all of them demon strate the meaning ‘part of a building; upper store; building’). Sabaic ÒrÌ ‘?cult building’, quoted in SD 144 5, is too unreliable, based on emendation of ÒÌn to ÒrÌn in a broken context in RES 4767: 3. Perhaps one should mention Minaic ÒrÌ in RES 2771: 4: wywm / sl? / wdm / bÒrÌm / msnd / †yb (translation according to RES: ‘et au temps où il dédia à Waddum à ∑aÌrum une offrande de †ayb’), where †yb is interpreted by Sima (2000: 337) as ‘gold’ and ÒÌrm as either a proper name or as ‘vestibule of a temple’. 36 An inscription on a sherd of pottery (note that it remains unclear whether h is the end of the inscription). The natural suggestion is that the element brh is con nected with the well known Geez root brh: barha ‘shine, be bright’, b¢ruh ‘bright, shining’, barh ‘light’(LGz. 103; LLA 500). However, the translation of this frag ment as ‘[he] who shone’ or ‘[he] of light’ does not seem to make any sense (unless this is a beginning of a longer sentence). Perhaps the suggestion of Drewes (2000: 203), who considers brh to be a proper name derived from the same root, is more plausible in this case; then, the syntactic analysis of the phrase would be the same as that of RIE 337 (see note 34). 37 Cf. D.H. Müller 1894: 56; DAE 15, p. 48 9; W.W. Müller 1972; for the interpretation of lazen as a clan name see Fiaccadori 2004: 122. 38 Reading and translation according to DAE 13, p. 45. 39 Translation according to Drewes 1962: 45 6, 48. 40 This translation is, however, rejected by Drewes (1962: 38 9), who regards dwly as a place name: ‘law of Dwly’, ibid. 39, 48. 415
NOTA GENITIVI ZA- IN EPIGRAPHIC GEEZ
? RIE 183 I: 7: bg! / z-gb ‘sheep of [the age/state?] gb’41 ? RIE 183 I: 11 12: bg! / z-? †lt ‘sheep of [the age/state?] ? †lt’42 ? RIE 183 I: 19 20: w-lhm z-? ld ‘cow of [the age/state?] ? ld’43 ? RIE 183 I: 24 5: Èn / z-msg ‘a measure of butter?’44 ?? RIE 183 IV: 1 2: tzkr / z-whb ‘memorandum of gift?’45 ?? Ricci 1991 A: 2: t(?)]zkr / z-ngd / z-d… ‘memorandum of a ?merchant/ tribe of…’46 ?? Ricci 1991 B: 1: tzkr / z-dbr / … (?) ‘memorandum of ?Dabr … (?)’47 REFERENCES Aristar, A.R. 1991. ‘On dyachronic sources and synchronic pattern: An investiga tion into the origin of linguistics universals’, Language 67, 1 33 Bausi, A. 2005a. ‘Etiopico ? elle: a proposito di un’ ipotesi recente’, Scrinium 1 (= D. Nosnitsin [ed.],Varia Aethiopica. In Memory of Sevir B. Chernetsov (1943 2005)), 3 11 2005b. ‘Ancient features of Ancient Ethiopic’, Aethiopica 8, 149 69 2006. ‘The Aksumite Background of the Ethiopic “Corpus Canonum”’, in S. Uhlig (ed.), Proceedings of the XVth International Conference of Ethiopian Stud ies (Wiesbaden), 532 41 Brockelmann, C. 1908. Grundriss der vergleichenden Grammatik der semitischen Sprachen, 1: Laut und Formenlehre. (Berlin) Caquot, A. 1952. ‘Recherches de syntaxe sur le texte éthiopien d’Énoch’, JA 240, 487 96 Conti Rossini, C. 1942. ‘Un’iscrizione su obelisco di Anza’, Rassegna di Studi Etiopici 2:1, 21 8 Diem, W. 1986. ‘Alienable und inalienable Possession im Semitischen’, ZDMG 136, 227 91 41
Cf. Drewes 1962: 44 5. Cf. Drewes 1962: 44 5. 43 Cf. Drewes 1962: 44 5. 44 Translation according to Drewes 1962: 36 7, 49. Note that if the subsequent element, w †Ìn, is also the part of the genitive construction, this example should be classified under the group a although, in view of the frequency of za construc tions in ‘neutral’ contexts in Safra inscription, one can suggest that morpho syntactic conditions do not play any significant role in this language variety. 45 Translation according to Drewes 1962: 53 4 (cf. also Ryckmans 1962: 465). Note, however, that whb can be interpreted alternatively as a verb (cf. Drewes 1962: 54). Furthermore, it is not clear whether the phrase w lbs, which immedi ately follows this phrase, is also dependent on tzkr. In the latter case, the whole con struction could be classified under the group a (cp., however, note 44). 46 See Ricci 1991: 1298. 47 Translation according to Ricci 1991: 1303. Alternatively, one can venture an interpretation of dbr as a common noun, perhaps ‘work, corvée duty’, cf. Minaic dbr ‘travail; prestation de travail, redevances’ (LM 26), Sabaic ? dbr ‘?corvée duty’(SD 35). Note that the phrase z dbr occurs elsewhere within the same two in scriptions: A: 12 and B: 3. These lines are, however, damaged and can hardly clarify the meaning of dbr. 42
416
NOTA GENITIVI ZA- IN EPIGRAPHIC GEEZ
Dillmann, A. 1907 (repr. 1974). Ethiopic Grammar, ed. by C. Bezold, tr. with addi tions by J.A. Crichton. (London) Drewes, A.J. 1962. Inscriptions de l’Éthiopie antique. (Leiden) 2000. ‘Noms propres dans les documents épigraphiques de l’Éthiopie, II’, Semitica 50, 199 210 Drewes, A.J. and R. Schneider. 1967. ‘Documents épigraphiques de Éthiopie I’, Annales d’Éthiopie 7, 89 106 Fiaccadori, G. 1990 [1996]. ‘Epigraphica Aethiopica’, Quaderni utinensi 8 [15/16], 325 35 2004 [2005]. ‘Sembrouthes ‘gran re’ (DAE IV 3 = RIÉth 275). Per la storia del primo ellenismo Aksumita’, La Parola del Passato 59, 103 57 Garr, W.R. 1990. ‘On the Alternation between Construct and d Phrases in Bibli cal Aramaic’, JSS 35:2, 213 31 Gesenius, W. 1962. Hebräische Grammatik (revised by E. Kautzsch). (Hildesheim) Goldenberg, G. 1995. ‘Attribution in Semitic Languages’, Langues Orientales Anciennes: Philologie et Linguistique 5 6, 1 20 (repr. in G. Goldenberg. 1998. Studies in Semitic Linguistics. Selected Writings. [Jerusalem] 46 65) Hug, V. 1993. Altaramäische Grammatik der Texte des 7. und 6. Jh.s v. Chr. (Heidelberg) Kapeliuk, O. 1994. ‘Possessive and Determining Nominal Complexes in Semitic’, in G. Goldenberg and S. Raz (eds), Semitic and Cushitic Studies (Wiesbaden). 65 9 2003. ‘Some Remarks on the Etymology and Function of the Relative Mark ers in Ancient Ethiopic’, in M.L. Bender, G. Takács and D.L. Appleyard (eds), Selected Comparative Historical Afrasian Linguistic Studies in Memory of Igor M. Diakonoff (LINCOM Studies in Afroasiatic Linguistics 14, Munich). 219 32 Kropp, M. 1994. ‘Ein Gegenstand und seine Aufschrift. RIE 180 = JE 5’, in Yaqob Beyene et al. (eds), Etiopia e oltre. Studi in onore di Lanfranco Ricci (Napoli). 129 44 1999. ‘“Glücklich, wer vom Weib geboren, dessen Tage doch kurzbemessen, …!” Die altäthiopische Grabinschrift von Îam, datiert auf den 23. Dezember 873 n. Chr.’, Oriens Christianus 83, 162 76 2006. ‘Monumentalised accountancy from Ancient Ethiopia: The stele of Maryam Anza’, paper presented at the 2nd International Littmann Confer ence at Aksum 100 years German Aksum Expedition (DAE), 6 10 January 2006 Lambdin, T.O. 1978. Introduction to Classical Ethiopic (Ge! ez) (Ann Arbor, Mich.) Littmann, E. 1952. ‘L’iscrizione di Anza’, Rassegna di Studi Etiopici 11, 5 8 Lusini, G. 2004. ‘Note linguistiche per la storia dell’Etiopia antica’, in V. Böll et al. (eds), Studia Aethiopica: In Honour of Siegbert Uhlig on the Occasion of his 65th Birthday (Wiesbaden). 67 77 Müller, D.H. 1894. Epigraphische Denkmäler aus Abessinien, nach Abklatschen von J. Theodore Bent Esq. (Wien) Müller, W.W. 1972. ‘Zur aethiopischen Inschrift vor der alten Kathedralkirche in Axum’, Neue Ephemeris für Semitische Epigraphik 1, 129 33 Nikiforidou, N. 1991. ‘The meanings of the genitive: A case study in semantic structure and semantic change’, Cognitive Linguistics 2:2, 149 205 Nöldeke, Th. 1966. Kurzgefasste syrische Grammatik, revised by A. Schall. (Darmstadt) 417
NOTA GENITIVI ZA- IN EPIGRAPHIC GEEZ
Pennacchietti, F.A. 1968. Studi sui pronomi determinativi semitici. (Napoli) 1974. ‘Appunti per una storia comparata dei sistemi preposizionali semitici’, Annali dell’Istituto Universitatio Orientale di Napoli 24, 161 208 Polotsky, H.J. 1961. [Review of Schneider 1959], JSS 6, 251 6 Ricci, L. 1988. ‘Appunti archeologici’, Rassegna di Studi Etiopici 32, 129 65 1991. ‘Iscrizioni Paleoetiopiche’, in A.S. Kaye (ed.), Semitic Studies. In honor of Wolf Leslau. On the occasion of his eighty fifth birthday. November 14th, 1991, vol. 2 (Wiesbaden). 382 91 Ryckmans G. 1962. ‘Notes épigraphiques, sixième série. IX. Une contribution à l’épigraphie et à l’histoire de l’Éthiopie antique’, Le Muséon 75, 459 68 Schneider, R. 1959. L’expression des compléments de verbe et de nom et la place de l’adjectif épithète en Guèze. (Paris) 1966. ‘Quelques inscriptions céramiques de Ma†ara’, Rassegna di Studi Etiopici 22, 20 1 Schuh, R.G. 1982. ‘Types of genitive constructions in Chadic’, in: W.R. Leben (ed.), Précis from the 12th Conference on African Linguistics, Stanford, April 10 12, 1981 (= Studies in African linguistics, Supplement 8), 117 21 Sima, A. 2000. Tiere, Pflanzen, Steine und Metalle in den altsüdarabischen Inschriften. (Wiesbaden) 2003/2004. ‘Die “sabäische” Version von König !Ézanas Trilingue RIE 185 und RIE 185bis’, Archiv für Orientforschung 50, 269 84 Simeone Senelle, M. C. 1997. ‘The Modern South Arabian languages’, in R. Hetzron (ed.), The Semitic Languages (London). 378 423 von Soden, W. 1952. Grundriss der akkadischen Grammatik. (Roma) Tropper, J. 2000. Ugaritische Grammatik. (Münster) 2002. Altäthiopisch: Grammatik des Ge! ez mit Übungstexten und Glossar. (Münster) Weninger, S. 1993. G¢! ¢z (Classical Ethiopic). (Languages of the World / Materials 01, Munich Newcastle) SOURCES [DAE] Littmann, E. (ed.). 1913. Deutsche Aksum Expedition. IV: Sabäische, griechische und altabessinische Inschriften. (Berlin) Dillmann, A. (ed.). 1853. Biblia veteris Testamenti Aethiopica, in quinque tomos distributa. 1: Veteris Testamenti Aethiopici tomus primus, sive Octateuchus Aethio picus. (Leipzig) Esteves Pereira, F.M. (ed., tr.). 1911. Le Livre d’Esther. Version éthiopienne. (Paris) Grébaut, S. 1945. Le synaxaire éthiopien: Le mois de taÌsas (fin). (Patrologia orientalis 26, fasc. 1 = no. 125, Paris) Hofmann, J. (ed.). 1967. Die äthiopische Übersetzung der Johannes Apokalypse. (Cor pus Scriptorum Christianorum Orientalium 281, Scriptores Aethiopici 55, Louvain) Knibb, M.A. 1978. The Ethiopic book of Enoch. 2 vols. (Oxford) Perruchon, J. (ed.). 1903. Le Livre des Mystères du ciel et de la terre. (Paris) [RES] Répertoire d’epigraphie sémitique. (Paris), 1900 [RIE] Bernand, E. and A.J. Drewes and R. Schneider. 1991. Recueil des Inscrip tions de l’Éthiopie des Périodes pré Axoumite et Axoumite. Vol. 1. Les Documents. (Paris)
418
NOTA GENITIVI ZA- IN EPIGRAPHIC GEEZ
Zuurmond, R. (ed.). 1989. Novum Testamentum Aethiopice: The Synoptic Gospels. [I]: General Introduction; [II]: Edition of the Gospel of Mark. (Aethiopistische Forschungen 27, Stuttgart) 2001. The Gospel of Matthew. (Aethiopistische Forschungen 55, Wiesbaden) DICTIONARIES [DUL] Del Olmo Lete, G. and J. Sanmartín. 2003. A dictionary of the Ugaritic language in the alphabetic tradition. (Leiden) [LGz.] Leslau, W. 1987. Comparative Dictionary of Ge! ez (Classical Ethiopic). (Wiesbaden) [LLA] Dillmann, A. 1955. Lexicon Linguae Aethiopicae cum indice latino. (New York) [LM] Arbach, M. 1993. Le ma∂abien: Lexique, Onomastique et Grammaire d’une langue de l’Arabie méridionale pre´islamique. 1: Lexique ma∂abien. Thèse à doctorat Nouveau régime. Université de Provence Aix Marseille I. Centre d’Aix. (Aix en Provence) [SD] Beeston, A.F.L. et al. 1982. Sabaic Dictionary. (Louvain la Neuve)
419
ON THE ARABIAN ORIGINS OF THE QUR’AN: THE CASE OF AL-FURQAN URI RUBIN TEL AVIV UNIVERSITY
Abstract This article draws attention to hitherto unnoticed Arabic texts in which the word furqan features as a genuine Arabic term denoting ‘dawn’ or ‘morning’. It suggests that this genuine Arabic word is re flected in the Qur’an, especially in those passages in which the term furqan stands for revealed scriptures. The Qur’an calls these scriptures furqan in the sense that they are a guiding light that leads one out of darkness. The Syriac/Aramaic connotations of the term as encoun tered in other Quranic passages mainly those referring to Moses are also discussed, and their origin in the Syriac and Aramaic transla tions of the Book of Exodus is further clarified. The article shows that in these translations the Syriac / Aramaic purqan stands for redemp tion through separation, and this seems also to be the meaning of the Quranic furqan of Moses. The conflation of the Arabic and the Ara maic furqans in the Qur’an is also demonstrated, and finally the evi dence of the tafsir is analysed. It is shown that the exegetes are aware of the targumic sense of furqan (redemption through separation), but they tend to prefer the sense of the local Arabian furqan (light of dawn), so much so that they have derived from the sense of light as opposed to darkness a series of secondary meanings revolving around the idea of separation of truth from falsehood.
Introduction In a recent study of the Quranic term furqan, F.M. Donner proposes a new etymology for this term. In some cases he accepts its derivation from the Aramaic / Syriac word purqana (‘salvation’), as long assumed by many modern scholars, but argues that in other cases this derivation does not fit the Quranic context and suggests the Syriac puqdana as a possible origin for those cases.1 This means that according to Donner, the origin of the Quranic furqan is exclusively Aramaic/Syriac (either purqana or puqdana), 1 Fred M. Donner, ‘Quranic Furqan’, Journal of Semitic Studies 52 (2007), 279 300.
421
ON THE ARABIAN ORIGINS OF THE QUR'AN
and indeed from the very outset he gives up any attempt at finding Arabian origins for the term, asserting that, ‘furqan is not, to my knowledge, attested in early Arabic texts in contexts clearly independent of the Qur’an’.2 A similar argument against the Arabian origin of the word was already presented by A. Jeffery who said that ‘…there seems to be no evidence of the use of the word in Arabic earlier than the Qur’an’.3 But some scholars have not been so sure of the exclusively nonArabian origin of the Quranic furqan. For example, A.J. Wensinck spoke of two distinct words in the Qur’an. One is the Syriac purqana, meaning ‘salvation’ or ‘deliverance’, and the other is a genuine Arabic word meaning ‘distinction’, which the Qur’an uses for ‘revelation’.4 But Wensinck’s view was altogether rejected by Jeffery who said that ‘Wensinck seems to have been unduly influenced by the theories of the native commentators’.5 Wensinck, however, seems to have been correct in his basic assumption, although he was unable to provide sufficient evidence for the Arabian origins of the Quranic furqan. In what follows, such evidence is provided, which enables us to better understand the etymological history of the Quranic furqan. The Genuine Arabian furqan To begin with, the word furqan occurs in the available Arabic sources in contexts that are clearly independent of the Qur’an. The lexicographer al-Azhari (d. 370/980) adduces an utterance attributed to an unidentified ‘Bedouin’ (a¨rabi), who reportedly said: ‘I went on having sex with her (a∂fizu biha) until the furqan was shining (ila an sa†a¨a ’l-furqan)'. The word furqan is glossed by the words: ay al-saÌar, ‘that is to say, dawn’.6 The same utterance is recorded also in the Lisan al-¨arab by Ibn ManÂur (d. 711/1311),7 and in the Taj al-¨arus by al-Zabidi (d. 1205/1790).8 2
Donner ‘Quranic Furqan’, 280. Arthur Jeffery, The Foreign Vocabulary of the Qur’an (Baroda 1938), 228. 4 A.J. Wensinck, s.v. ‘FurÈan’, EI (old edition), II, 120. 5 Jeffery, Foreign Vocabulary, 228, Note 4. 6 MuÌammad b. AÌmad al Azhari, Tahdhib al lugha (ed. ¨Abd al Salam MuÌammad Harun, MuÌammad Abu ’l Fa∂l Ibrahim, ¨Ali MuÌammad al Bijawi, 16 vols. Cairo 1964 76), s.v. ‘∂.f.z.’ [XI, 488]. 7 Ibn ManÂur, Lisan al ¨arab (6 vols. Cairo n.d.), s.v. ‘∂.f.z.’ [IV, 2595 col. 1 2]. 8 MuÌammad Murta∂a al Zabidi, Taj al ¨arus (10 vols. Benghazi 1966), s.v. ‘∂.f.z.’ [IV, 46]. 3
422
ON THE ARABIAN ORIGINS OF THE QUR'AN
Needless to say, nothing is Quranic here. The recourse to the ‘Bedouin’ is a typical method used by the lexicographers in order to illustrate the significance and context of a given indigenous Arabic word or term. In this particular case, al-Azhari adduces the quotation from the Bedouin in order to illustrate the meaning of the verb ∂afaza (‘to have sexual intercourse with’), and furqan occurs here as a normal, though not necessarily widely-current, word for the light of dawn. A modified variation of the same utterance appears in al-Marzuqi’s (d. 421/1030) al-Azmina wa ’l-amkina, in a chapter dealing with Arabic terms describing the blackness of night. Since the sexual context is now irrelevant, the utterance of the Bedouin can be rephrased in a less offending manner. This time the Bedouin says: ‘I went on acting wildly (ata¨ssafu) in the darkness of night (al-hawalwal), till the furqan shone (sa†a¨a)'. When asked what al-hawalwal was, the Bedouin said: ‘The darkness of night’. When asked what the furqan was, he said: al-ÒubÌ (‘morning’).9 Furqan in the sense of ‘dawn’ recurs in a poetic verse recorded by al-Azhari as part of a report of Abu ¨Ubayda (d. 210/826). Its first hemistich runs as follows: Fa-ÒabbaÌat qabla adhani ’l-furqan: ‘They [i.e. the camels?] rose up early in the morning, before the call (adhan) to the furqan prayer’. Abu ¨Ubayda explains that furqan is ÒubÌ, ‘morning’.10 The passage from Abu ¨Ubayda recurs in Lisan al¨arab,11 and in Taj al-¨Arus.12 Here too the context — although clearly Islamic — is entirely non-Quranic. Furqan appears again in the sense of ‘dawn’, this time signalling the time of a well-known morning prayer, its legal time lasting from the crack of dawn till just before sunrise. More insight into the significance of the word furqan in the sense of ‘dawn’ is provided by al-Zabidi. Among the various meanings of the word, he records the report of the renowned BaÒran philologist Abu ¨Amr b. al-¨Ala’ (d. 154/771), who says that furqan means ‘morning’ (ÒubÌ) or ‘dawn’ (saÌar), and notes that from this stems the expression: ‘The furqan is shining (sa†a¨a)’, as well as the saying: ‘This is brighter (abya∂ ) than the furqan’.13 9 Abu ¨Ali AÌmad b. MuÌammad al Marzuqi, al Azmina wa ’l amkina (ed. Khalil al ManÒur, Beirut 1996), 433. 10 Al Azhari, Tahdhib al lugha, s.v. ‘b.d.’.’[XIV, 206]. 11 Lisan, s.v. ‘b.d.ˆ.’ [I, 224, col. 3]. 12 Taj al ¨arus, s.v. ‘b.d.’.’ [I, 43]. 13 Ibid., s.v. ‘f.r.q.’ [VII, 45 6]. See also E.W. Lane, An Arabic English Lexicon (2 vols., repr. Cambridge 1984), s.v. ‘f.r.q.’ [2385 col. 3]; Suliman Bashear, ‘The Title Faruq and its Association with ¨Umar I’, Studia Islamica 72 (1990), 50.
423
ON THE ARABIAN ORIGINS OF THE QUR'AN
Elsewhere al-Zabidi says that when signifying ‘dawn’, furqan is a metaphor (wa-huwa majaz),14 probably meaning that the word does not stand for dawn itself, only describes its function. Since the root f.r.q. signifies ‘division’, it seems that furqan is taken here as a verbal noun denoting splitting of darkness as light breaks through it. Another form derived from the root f.r.q., and which is likewise associated with dawn, is faraq. The expression faraqu ’l-ÒubÌ is explained by the lexicographers as signifying ‘the crack of dawn’. They say that it was called faraq because it has become separated (faraqa) from the darkness of night.15 In some Arab dialects, the form faraq was interchangeable with falaq. Qu†rub (MuÌammad b. al-Mustanir [d. 206/821]) says that the Arabs of the Tamim used to say faraqu ’l-ÒubÌ, while all others said falaqu ’l-ÒubÌ.16 The affinity between faraq and falaq is also noted by al-Zamakhshari (d. 538/1143) in his comments on Q 113:1. In this verse the ‘Lord of the falaq’ is invoked for protection. Al-Zamakhshari explains that falaq and faraq denote ‘morning’ (ÒubÌ), because night is split and separated from it (yuflaqu ¨anhu wa-yufraqu). He goes on to note that from here stems the saying: ‘The furqan is shining (sa†a¨a l-furqan)’, i.e. ‘the dawn has risen’ (†ala¨a ’l-fajru).17 In sum, the above texts assert the existence of the word furqan in the local, or indigenous, Arabic of early, and possibly, pre-Islamic times. It functioned as a metaphor of ‘dawn’, in the sense of splitting of, or separation from, darkness, much like the forms faraq and falaq. The Quranic furqan At this point the question arises whether the Quranic furqan might reflect this local Arabian furqan, rather than the Syriac/Aramaic one. It seems that the sense of ‘dawn’ or ‘separation from darkness’ fits particularly well into those passages in which the term furqan stands for, or describes, a revealed scripture. That these passages may reflect a local Arabian furqan was already suggested by Wensinck (see above), and in view of the above data we are able to see more clearly the relationship between this genuine Arabic word and the Quranic notion of revelation. The idea behind the term furqan as a revealed scripture seems to be that the scripture is a source of guiding light 14
Taj al ¨arus, s.v. ‘∂.f.z.’ [IV, 46]. Lisan, s.v. ‘f.r.q.’ [V, 3400 col. 1]. See also Lane, Lexicon, 2385 col. 1. 16 Al Marzuqi, Azmina, 243. 17 Al Zamakhshari, al Kashshaf ¨an Ìaqa’iq al tanzil (4 vols. Cairo 1966), IV, 300. 15
424
ON THE ARABIAN ORIGINS OF THE QUR'AN
that leads one out of the darkness of ignorance. This seems to be the case in Q 25:1 in which furqan stands for the Qur’an that has been revealed to MuÌammad: ‘Blessed is he who sent down the furqan upon his servant that he may be a warner to all beings’. This verse seems to say that God sent down to the Prophet an illuminating message which enables him to warn people and show them the light of guidance. This verse is contained in a sura which, according to the traditional Islamic chronology of the Qur’an, was revealed in Mecca. The same recurs in suras defined as Medinan. Thus in Q 2:185 the month of Rama∂an is mentioned as the time when the Qur’an was sent down, in order to bring ‘guidance to men and clear signs of the guidance and of the furqan’. The explicit combination in this verse between furqan and guidance as two complementary aspects of the ‘clear signs’ is best understood when the furqan is perceived as light. The word furqan appears side by side with guidance also in Q 3:3–4 [Medinan]: ‘He has revealed to you the book with truth, verifying that which is before it, and he revealed the Torah and the Gospels (Injil ) aforetime, a guidance for the people, and he sent down the furqan …’ Guidance goes hand in hand not merely with the term furqan but also with a more explicit word for light, i.e. nur. This is the case in Q 42:52 [Meccan] in which the book revealed to MuÌammad is said to have come as ‘light, guiding thereby whom we please of our servants’. The word ‘light’ (nur) as standing for the guiding message sent down to MuÌammad appears in further passages, Meccan (Q 7:157) as well as Medinan (Q 4:174; 5:15–16; 64:8). The guiding book itself is described as kitab munir, ‘luminous book’ (Q 22:8), which is also how books revealed to previous prophets are described (Q 35:25 [Meccan]; 3:184 [Medinan]). The book, or the Torah, that was revealed to Moses is also described as ‘light and guidance’ (Q 6:91 [Meccan]; 5:44 [Medinan]), and so is the Injil revealed to Jesus (Q 5:46 [Medinan]). The term furqan, then, as a name for, or description of, revealed scriptures, fits perfectly well into the Quranic symbolism of light. In fact, the association between furqan, as standing for the Qur’an, and light continued into post Quranic times. This is indicated by the phrase munir al-furqan (‘the luminous furqan’) by which the Qur’an was sometimes referred to by Muslims in the first Islamic era.18 But in the Qur’an the word furqan is not just a guiding light. When made part of God’s address to MuÌammad this local Arabian 18 Al Baladhuri, Jumal min ansab al ashraf (ed. Suhayl Zakkar and Riya∂ Zirikli, 13 vols. Beirut 1996), XIII, 87.
425
ON THE ARABIAN ORIGINS OF THE QUR'AN
word seems to have acquired further meanings which had come into the Qur’an from the Syriac/Aramaic. Western scholars already noted the Syriac/Aramaic parallel of the Quranic furqan,19 but have not yet provided the exact relationship between the two. The Syriac/Aramaic connotations of the word furqan reached the Qur’an together with the image of Moses. The latter brought in a furqan of his own, which consisted of a combination of two ideas: redemption and division. The history of this combination begins in biblical Hebrew, with the word pdut ()פדות.This word means ‘redemption’, as for example in Ps. 111:9: ‘He [i.e. God] sent redemption (pdut) unto his people’.20 However, the same word occurs also in Exod. 8:19 [23] where it is something which God sets between his own people and the people of Pharaoh. God says to Pharaoh: ‘I will set pdut between my people and your people'. The context is the redemption of the Israelites from the swarms of flies which will not reach the Israelites in the Land of Goshen, because, as stated in the previous verse, God has set apart the Land of Goshen from the rest of Egypt. This context has lent the word pdut the sense of ‘division', as if implying that the Israelites are redeemed when set apart from their foes. This is indeed how the Hebrew pdut of Exod. 8/18 [23] was perceived in the Syriac/Aramaic translations. In the Peshi†ta (Syriac Bible) pdut is rendered purshana (‘distinction'), which God will make between the Israelites and the Egyptians.21 In the Aramaic translation known as Targum Neophyti, pdut is translated purqan (‘redemption'). God says to Pharaoh: ‘And I will set purqan between my people and your people’.22 Since purqan is set here between the two parties, it signifies redemption that comes from division or separation. In the Targum Onkelos, pdut is again purqan, but here an extended paraphrase is provided. God says to Pharaoh: ‘I will make purqan for my people, and on your people I shall bring a plague [of flies]’.23 The same paraphrase appears in the Pseudo-Jonathan.24 19 E.g. Jeffery, Foreign Vocabulary, 225 9; EI 2, s.v. ‘FurÈan’ (R. Paret); Arne A. Ambros, A Concise Dictionary of Koranic Arabic (Wiesbaden 2004), 212; EQ s.v. ‘Names of the Qur’an’ (Mustansir Mir); Donner, 286 8. 20 See also Ps. 130:7. 21 Ch. Heller, Peshitta in Hebrew Characters with Elucidatory Notes, Part II: Exo dus (Berlin 1929), 78. 22 Alejandro Díez Macho, Neophyti 1. Tomo II. Éxodo (Madrid 1970), 47. 23 Israel Drazin, Targum Onkelos to Exodus (New York 1990), 97. 24 E.G. Clarke, Targum Pseudo Jonathan of the Pentateuch: Text and Concordance (Hoboken 1984), 74.
426
ON THE ARABIAN ORIGINS OF THE QUR'AN
This targumic purqan by which God discriminates between the Israelites and the Egyptians, seems to be the origin of the Quranic furqan of Moses. It occurs in Q 21:48 [Meccan], which reads: ‘We gave to Moses and Aaron the furqan and a light and a reminder for the God-fearing’. The close association between furqan and light (∂iya’) in this verse retains the basic sense of the local Arabian furqan (‘dawn’), but the fact that here it is given to Moses and Aaron in particular seems to allude to the targumic significance of the term as used for Exod. 8/19 [23]. The Qur}an, however, does not use it in any specific context of the life of Moses, so that it retains a general meaning of deliverance through separation. Hence the Quranic verse demonstrates a conflation between the local Arabian significance of furqan and the targumic one. The same conflation is demonstrated in Q 2:53 [Medinan] where the furqan is only given to Moses: ‘We gave Moses the book and the furqan that you might be guided'. The furqan of Moses has been associated here with the guiding book that was revealed to him, a link which again combines the targumic furqan of Moses with the local Arabian one. In Q 8:41 (Medinan) the word furqan seems to retain its pure targumic sense (‘redemption by separation') although now it is something that is given to the Muslims instead of to Moses. This verse alludes to ‘that which we sent down to our servant, on the day of the furqan, the day on which the two parties met…' The day of the furqan seems to stand for an Islamic military victory over the unbelievers [at Badr], and the furqan is evidently the ‘distinction’ which God made between the ‘two parties’, i.e. believers and unbelievers, which resulted in the redemption of the party of the latter. This is exactly the sense of the targumic furqan of Moses. The Qur’an itself seems to be aware of the relationship between the Islamic victory and the Israelites’ exodus from Egypt, because the locution ‘the two parties’ (al-jam¨an) recurs in Q 26:61 where it stands for the Israelites and the Egyptians who chase them.25 It is noteworthy that the following passage (Q 26:62–3) describes the parting of the sea, for which the roots f.l.q. and f.r.q. are being used: The sea is ‘split asunder’ (infalaqa), and each ‘part’ (firq) of it stands like a huge mountain. This implies a new connotation of the idea behind the Quranic furqan of Moses: it is not merely division between Israelites and Egyptians as in the targumic furqan, but also the parting of the sea which assured the deliverance of the Israelites. In the Aramaic Bible 25
This has already been noted by Donner, p. 289. 427
ON THE ARABIAN ORIGINS OF THE QUR'AN
purqan already features in connection with the redemption of the Israelites from Pharaoh’s army,26 and this may well be the origin of the implied relationship between the Quranic furqan of Moses and the parting of the sea. Another Medinan verse referring to furqan is Q 8:29 which defines the reward for piety: ‘O believers, if you fear God, he will give you furqan and do away with your evil deeds and forgive you…’. A solid basis for assessing the furqan here is provided in a parallel Medinan verse, Q 65:2, in which God says: ‘…and whoever fears God, he will give him a way out (makhraj )’. The reward for piety is here makhraj which replaces furqan. The idea is repeated in Q 65:4, and this time the reward for piety is ‘alleviation’ (yusr): ‘…and whoever fears God he will give him alleviation in his affair’. If makhraj and yusr be taken as synonymous with furqan, then the latter seems to denote a way out from hardships, much like the targumic furqan. In sum, the Quranic furqan reflects two basic etymological clusters, a local Arabian one which revolves around morning light as signifying splitting of, or separation from, darkness, and a targumic one which revolves around redemption achieved through division or separation. The local Arabian furqan can be discerned in passages in which this term conveys the sense of guidance, and is equated with a revealed book. The targumic furqan is discerned where this term is related to Moses, as well as to a victory of the Muslims over their foes. The Tafsir In Islamic tafsir the two etymological channels are reflected very clearly, but as is already the case in the Qur’an, the two are not always kept apart. As for the targumic furqan (redemption through separation), it is preserved in some interpretations of Q 8:29 in which, as seen above, the term furqan is a reward for piety, and is probably synonymous with makhraj and yusr, and denotes alleviation of worldly hardships. The exegetes are aware of the relationship between furqan and makhraj,27 as they explain it as ‘redemption’ (al-najat) and ‘victory’ (al-naÒr).28 Or as ¨Ikrima (d. 105/723) and al-Suddi (d. 128/746) reportedly said: ‘He gave you furqan, i.e. redemption, which means 26
Ibid., 289. See the traditions in al ™abari, Jami¨ al bayan fi tafsir al Qur’an (30 vols. Bulaq 1323/1905, repr. Beirut 1972) on Q 8:29 [IX, 148]. 28 E.g. Ibn al Jawzi, Zad al masir fi ¨ilm al tafsir (9 vols. Beirut 1984), III, 346. 27
428
ON THE ARABIAN ORIGINS OF THE QUR'AN
that God separates you from whom you fear, and so you are redeemed’.29 The targumic furqan is especially conspicuous in the exegesis of passages alluding to events in the life of Moses. Al-Mawardi (d. 450/ 1058) provides in his comments on Q 2:53 several interpretations for the furqan given to Moses, one of which is ‘victory (naÒr), by which God has distinguished between Moses and Pharaoh, till he saved (anja) Moses and his people and drowned Pharaoh and his people’.30 Another interpretation on al-Mawardi’s list takes furqan to stand for the splitting of the sea before the Israelites.31 This is inspired by the Quranic usage of the root f.r.q. for the parting of the sea (see above). The targumic furqan is reflected also in the interpretations of Q 21:48 in which the furqan is given to Moses and Aaron. Al-Mawardi has recorded the interpretation of al-Kalbi (d. 146/763), to the effect that this furqan stands for victory and redemption (al-naÒr wa ’l-najat); God has given victory to Moses and his supporters and destroyed Pharaoh and his followers.32 But in spite of the survival of the targumic furqan in the tafsir, the exegetes tend quite naturally to prefer the local Arabian sense of furqan (‘light of dawn’), and they endeavour to apply it to all the Quranic occurrences of the term, including those in which the term retains the targumic sense of redemption through division or separation. This is most conspicuous in some further interpretations of Q 8:29, in which furqan is promised as a reward for piety. The exegetes reinterpret the term as ‘light’, and take its synonymous makhraj to denote separation from some kind of mental darkness. For example, al-Ghazali (d. 504/1111) writes in his IÌya’ ¨ulum al-din, that the furqan promised to the believers in Q 8:29 was explained as denoting ‘light (nur) by which truth (al-Ìaqq) is distinguished from falsehood (al-ba†il ), and by which one gets out from doubts (wa-yukhraju bihi min al-shubuhat)’.33 29
Al WaÌidi, ¨Ali b. AÌmad, al Wasi† fi tafsir al Qur’an al majid (ed. ¨Adil AÌmad ¨Abd al Mawjud et al., 4 vols. Beirut 1994), II, 454. 30 Al Mawardi, al Nukat wa ’l ¨uyun fi tafsir al Qur’an (ed. ¨Abd al MaqÒud b. ¨Abd al RaÌim, 6 vols. Beirut 1992), I, 121. See also al Bay∂awi, Anwar al tanzil wa’l asrar al ta’wil (2 vols. Beirut 1988), I, 62. 31 Al Mawardi, Nukat, I, 122. See also Ibn al Jawzi, Zad al masir, on Q 2:53 [I, 81]; al Bay∂awi, on Q 21:48 [II, 72]. 32 Al Mawardi, Nukat, III, 450. See also Ibn al Jawzi, Zad al masir, V, 355. 33 Al Ghazali, IÌya’ ¨ulum al din (5 vols. Cairo 1967), III, 31. 429
ON THE ARABIAN ORIGINS OF THE QUR'AN
Ghazali’s words evince a reformulation of the idea of furqan and makhraj, which now are no longer redemption from worldly hardships, but rather ‘a way out from doubts’. This shift from the worldly to the spiritual meaning is evidently inspired by the local Arabian sense of furqan as light of dawn; it is encountered already in the comments of Muqatil b. Sulayman (d. 150/767) on Q 8:29. Muqatil defines the term furqan as a makhraj (‘way out’) which leads one out of shubuhat (‘doubts’).34 Muqatil has applied the same meaning to other Quranic occurrences of furqan in which it stands for a revealed scripture. This is seen in his comments on Q 2:185,35 3:4,36 and 25:1.37 Ghazali’s words also show how the exegetes have elaborated on furqan as ‘light’, adding to it a further secondary meaning derived from the root f.r.q., i.e. ‘distinction between truth and falsehood’. This derivation, too, is encountered already in earlier sources. In his comments on Q 8:29, al-Alusi (d. 1270/1853) says that by giving furqan to the believers God puts into their hearts ‘guidance and light’ (hidayatan wa-nuran) by which ‘you shall distinguish between truth and falsehood'. He traces this interpretation, which obviously draws on the local Arabian sense of furqan, back to Ibn Jurayj (d. 150/767) and Ibn Zayd (d. 182/798).38 A similar interpretation appears in some earlier tafsir compilations. Al-Tha¨labi (d. 427/1035) quotes Ibn Zayd who reportedly said that furqan in Q 8:29 is the distinction within the hearts of the believers between truth and falsehood.39 The earliest appearance of this interpretation can be traced back to ¨Urwa b. al-Zubayr (d. 94/713). This can be done thanks to Ibn Abi Îatim (d. 327/938) who has recorded in his Tafsir a tradition of Ibn IsÌaq (d. 150/768) which the latter quotes from ¨Urwa, through the latter’s nephew, MuÌammad b. Ja¨far b. al-Zubayr. ¨Urwa is said to have explained the term furqan of Q 8:29 as denoting ‘distinction (faÒl) between truth and falsehood, by which God brings into light (yuÂhiru) your truth (Ìaqqakum) and extinguishes (yu†fi’u) the falsehood (ba†il) of those who disagree with you (man khalafakum)'.40 This interpretation appears already in the Sira of Ibn 34
Muqatil b. Sulayman, Tafsir al Qur’an (ed. ¨Abdallah MaÌmud ShiÌata, 5 vols. Cairo 1979), II, 110. 35 Muqatil, I, 161. 36 Ibid., I, 262. 37 Ibid., III, 225. 38 Al Alusi, RuÌ al ma¨ani (30 vols. Beirut n.d.), IX, 196. 39 Al Tha¨labi, al Kashf wa’l bayan ¨an tafsir al Qur’an (ed. Abu MuÌammad b. ¨Ashur and NaÂir al Sa¨idi, 10 vols. Beirut, 2002), IV, 348. 40 Ibn Abi Îatim, Tafsir al Qur’an al ¨aÂim (ed. As¨ad MuÌammad al ™ayyib, 430
ON THE ARABIAN ORIGINS OF THE QUR'AN
IsÌaq, but without an isnad.41 Its terminology of light shows how the basic meaning of the local Arabian furqan (‘light of dawn’) was extended to denote the light of truth as distinct from the darkness of falsehood. The sense of distinction between truth and falsehood was the one which the exegetes liked most, and they applied it even to cases in the Qur’an in which furqan is explicitly related to Moses, i.e. reflecting the targumic purqan. They could do it quite easily because the Quranic furqan of Moses already combined the ideas of redemption and separation. For example, al-Mawardi records in his comments on Q 21:48 an interpretation of Ibn Zayd to the effect that the furqan of Moses and Aaron stands for the ‘proof ’ (burhan) which distinguished between the truth (Ìaqq) of Moses and the falsehood (ba†il) of Pharaoh.42 This is also how Muqatil already interprets the furqan given to Moses in Q 2:53. He says that it stands for ‘victory (naÒr), when God separated between truth and falsehood. He gave victory to Moses and destroyed Pharaoh’.43 Furthermore, in some interpretations the furqan of Moses and Aaron (Q 21:48) was entirely detached from the conflict with Pharaoh, and was equated with the Torah; as such it was explained as marking separation of the lawful from the unlawful. This interpretation was reported on the authority of Mujahid (d. 104/722) and Qatada (d. 117/735).44 It changes the meaning of the furqan of Moses from redemption to guidance, thus transforming it from targumic into local Arabian. Similarly, the furqan given to Moses in Q 2:53 was interpreted as the Torah which distinguishes between truth and falsehood. This view was reported on the authority of Ibn ¨Abbas.45 The sense of distinction between truth and falsehood also permeated the expression yawm al-furqan of Q 8:41. Muqatil explains it as 10 vols. Riyad 1997), V, 1686 (no. 8990). See also Tha¨labi, al Kashf wa ’l bayan, IV, 348. Cf. al ™abari, Tafsir, on Q 8:29 [IX, 148]: Printed yukhfi instead of the correct yu†fi’u. 41 Ibn Hisham, al Sira al nabawiyya (ed. MuÒ†afa al Saqqa, Ibrahim al Abyari, and ¨Abd al Îafi Shalabi, 4 vols. Beirut 1971), II, 325. The interpretation is in cluded in a chapter assembling the Quranic verses revealed in connection with the battle of Badr. 42 Al Mawardi, Nukat, III, 450. See also Ibn al Jawzi, Zad al masir, V, 355. 43 Muqatil, I, 107 8. 44 Al Mawardi, Nukat, III, 450. See also Ibn al Jawzi, Zad al masir, V, 355. And see the traditions in al ™abari, Tafsir, on Q 21:48 [XVII, 26]. 45 Al Mawardi, Nukat, I, 121. See also Ibn al Jawzi, Zad al masir, I, 81. And see the traditions in al ™abari, Tafsir, on Q 2:53 [I, 225 6]. 431
ON THE ARABIAN ORIGINS OF THE QUR'AN
standing for ‘the day of victory, on which God has separated truth from falsehood, and gave victory to the Prophet and defeated the polytheists at Badr’.46 Likewise, the furqan promised to the believers in Q 8:29 was interpreted as victory which separates the follower of truth (al-muÌiqq) from the follower of falsehood (al-mub†il).47 The sense of ‘division between truth and falsehood’ was especially appropriate for all those Quranic passages in which the term furqan stands for, or describes, a revealed scripture, and such an interpretation appears indeed in the various commentaries on these verses.48 Finally, another kind of elaboration on the basic meaning of light as inherent in the local Arabian furqan is seen in the interpretation of al-Zamakhshari of Q 8:29. Among the various optional explanations, al-Zamakhshari adduces the interpretation of furqan in the sense that ‘God will give you fame and will spread your reputation and achievements all over the earth’ (… yushhiru amrakum wayabuththu Òitakum wa-atharakum fi aq†ar al-ar∂ ). Al-Zamakhshari adds that this is derived from the saying: ‘I did so and so till the furqan shone (sa†a¨a)’, i.e. the dawn rose (ay †ala¨a ’l-fajru).49 Apart from a censured allusion to the above saying of the Bedouin, this interpretation shows another direction in which the sense of light was elaborated on. Here it symbolizes fame and reputation. Such derivation seems to be inspired by the saying (mathal ): ashhar min faraq al-ÒubÌ, ‘more famous than the crack of dawn’.50 The same interpretation has been repeated by al-Bay∂awi (d. c. 685/1286) on Q 8:29,51 as well as by al-Alusi.52 Summary and Conclusion The above study has demonstrated that the word furqan is not an exclusively non-Arabic loanword, as held by many, but has also pure 46 Muqatil, II, 116. See also al Mawardi, Nukat, II, 321; Ibn al Jawzi, Zad al masir, III, 361. And see the traditions in al ™abari, Tafsir, on Q 8:41 [X, 7 8]. 47 Al Bay∂awi, I, 381. 48 See the traditions to that effect in al ™abari, Tafsir, on Q 2:185 [II, 85]; 3:4 [III, 111]; 25:1 [XVIII, 136]. 49 Al Zamakhshari, Kashshaf, II, 154. 50 Abu Hilal al ¨Askari, Jamharat al amthal (ed. AÌmad ¨Abd al Salam and MuÌammad b. Basyuni Zaghlul, 2 vols. Beirut 1988), I, 459 (no. 1047). Var.: ashhar min falaq al ÒubÌ. See ibid., no. 1046. 51 Al Bay∂awi, I, 381. 52 Al Alusi, IX, 196.
432
ON THE ARABIAN ORIGINS OF THE QUR'AN
Arabian roots, discerned in non-Quranic texts in which it stands for the light of dawn. The Qur’an employs it mainly to describe revealed scriptures in their capacity as origin of guiding light. But the targumic (Syriac/Aramaic) connotations of the Quranic furqan are also present in passages in which furqan is related to Moses, and denotes ‘redemption by division or separation'. The exegetes of the Qur’an are aware of the significance of furqan as ‘redemption’, but they prefer to interpret the Quranic furqan in the sense of ‘distinction between truth and falsehood’. This is not merely a sterile manipulation designed to provide false Arabian origins to a non-Arabic loanword, as held by many,53 but is rather an honest exegetical effort revolving around the genuine sense of light as inherent in the purely local Arabian furqan. This means that even when faced with cases of loanwords, one must never give up the quest for the Arabian origins of the Qur’an. They are certainly out there, waiting to be unveiled.
53
See e.g. Donner, 299 n. 45, who suggests that ‘furqan as a verbal noun of faraqa “to separate, discriminate”, which we find in dictionaries, is itself a product of the exegetical work of early Qur’an scholars grappling with these Qurˆanic verses’. 433
LANGUAGE PASSIVITY IN THE MEDICAL ARABIC AND JUDAEO-ARABIC PRESCRIPTIONS OF THE CAIRO GENIZAH AHARON GEVA-KLEINBERGER and EFRAIM LEV UNIVERSITY OF HAIFA
Abstract A new study of the medical fragments at the T S Genizah collection has revealed 141 practical prescriptions. These are written in most cases in Arabic script (92) and the minority of them in Hebrew let ters, i.e. Judaeo Arabic (47), one text in Hebrew and one in Judaeo Persian. Most of the prescriptions were written on one page, usually on one side of a sheet of paper (very rarely vellum). The texts in Ara bic have a special technical language which uses verbs intensively in the passive voice (hence LP ‘Language Passivity’) instead of the nor mal ‘pure imperative’. This method of expression raises many ques tions about the reason for this formulation and about the alternative possibilities which exist in these prescriptions alongside LP. The fol lowing article surveys the LP types, makes some orthographical re marks that shed light on LP, singles out other substitutes for LP in the medical prescriptions in comparison with other medical formu laries in the Arabic of the Middle Ages. Finally, it tries to show, that from a merely linguistic phenomenon one can draw conclusions re garding other historical aspects, such as hints about the connection ‘physician patient pharmacist’ and the way to differentiate between normal theoretical medical texts and the pure practical medical pre scriptions.
Historical Background Over a thousand years ago, the Jewish community of Cairo (Fus†at) was one of the most important centres of Jewry in the world, particularly in the East. The community was a centre of considerable social, economic, and religious activity. The Palestinian Jews of Fus†at worshipped in the Ben Ezra synagogue and one of the rooms in their synagogue was utilized as a ‘Genizah’, or depository, from about the tenth to the nineteenth century. Almost every piece of writing that passed through its members’ hands, on vellum or paper, printed or manuscript, early or late, scholarly research or children's 435
LANGUAGE PASSIVITY IN THE MEDICAL ARABIC AND JUDAEO-ARABIC
writing exercises was consigned to the Genizah. Its extraordinary preservation for this long period against the ravages of time and decay was due to the exceptionally dry climate of Egypt. By the nineteenth century the fragments slowly began to reach the hands of private collectors and soon after that different academic institutes and libraries assembled their own collections. The main collection is the Taylor-Schechter (T-S) collection at Cambridge University library (150,000 items).1 The Genizah’s many collections have been studied since 1898.2 Much research focusing on a wide variety of matters has yielded a wealth of articles and books. Among the main fields that have been studied are various religious and biblical subjects, Jewish law, education, poetry, social life, trade and communal organization. Medical texts in the Genizah have been studied only as part of other subjects such as the different professional classes of the Jewish community in Old Cairo. A number of scholars have dealt with this topic, such as Goitein,3 Dietrich,4 Fenton,5 Cohen,6 Baker7 and Isaacs.8 In addition, the medical profession has been studied in works on the Genizah in general, and on the life of the Jewish communities and societies in the Mediterranean in particular.9 Medicine as a subject in the Genizah has been given due attention only in the last few years, with the publication of a catalogue of medical and para-medical manuscripts in the Cambridge Genizah collection by Isaacs and Baker.10 Short descriptions are given of 1,616 fragments; however, 1
B. Richler, Guide to Hebrew Manuscript Collections (Jerusalem 1994), 60 4. S.C. Reif, A Jewish Archive from Old Cairo (Surrey 2000), 1 22. 3 S.D. Goitein, ‘The Medical Profession in the Light of the Cairo Genizah Documents’, Hebrew Union College Annual 34 (1963), 177 94; S.D. Goitein, A Mediterranean Society: the Jewish Communities of the Arab world as Portrayed in the Documents of the Cairo Genizah (Berkeley, Los Angeles, London, 1967 88). 4 A. Dietrich, Zum Drogenhandle im islamischen Agypten (Heidelberg 1954). 5 P. Fenton, ‘The Importance of the Cairo Genizah for the History of Medi cine’, Medical History 24 (1980), 347 8. 6 M.R. Cohen, ‘The Burdensome Life of a Jewish Physician and Communal Leader: A Geniza Fragment from the Alliance Israelite Universelle Collection’, Jeru salem Studies in Arabic and Islam 16 (1993), 125 36. 7 C.F. Baker, ‘Islamic and Jewish Medicine in the Medieval Mediterranean World: the Genizah Evidence’, Journal of the Royal Society of Medicine 89 (1996), 577 80. 8 H.D. Isaacs, ‘The Impact of Western Medicine on Muslim Physicians and their Writing in the 17th Century’, Bulletin of the British Association of Orientalists 11 (1979 80), 52 7; H.D. Isaacs, ‘A Medieval Arab Medical Certificate’, Medical History 35 (1991), 250 7. 9 Goitein, A Mediterranean II, 240 72. 10 H.D. Isaacs (with the assistance of Colin F. Baker). Medical and Para medical Manuscripts in the Cambridge Genizah Collection (Cambridge 1994). 2
436
LANGUAGE PASSIVITY IN THE MEDICAL ARABIC AND JUDAEO-ARABIC
the book hardly discusses or analyses the evidence. Recent catalogues and further research on the T-S collection have provided information on more than 180 fragments referring to medicine in general and materia medica in particular.11 Our research project is based mainly on the T-S Genizah collection at Cambridge, since all other collections are much smaller and some have not yet been catalogued. To date, no specialist catalogue of medical materials in the other Genizah collections has been published.12 The literature on medicine in medieval Muslim countries in general13 and in Egypt14 in particular is vast but it mainly discusses theory, not practice. So far, sifting through Genizah fragments has yielded the names of more than fifty physicians.15 That a large number of Jews engaged in the medical profession in Egypt and other Muslim territories emerges from other historical sources as well, mainly the books by medieval biographers and historians of medicine such as Ibn Abi }UÒaybi¨a and Ibn al-Qif†i.16 The former mentions more than fifteen Jewish practitioners he met or knew of in Cairo in his time and be11 E. Lev and L. Chipman, Isaacs’ catalogue ‘Medical and Para medical Manu script in the Cambridge Genizah Collection’ New edition (Oxford forthcoming); F. Niessen and E. Lev, ‘Addenda to Isaacs Catalogue “Medical and Para medical Manuscript in the Cambridge Genizah Collection Together with the Edition of Two Medical Documents T S 12.33 and T S NS 297.56"', Hebrew Union College Annual 77 (2008), 131 65. 12 Catalogues of medical fragments in the Mosseri (Cambridge) and John Rylands (Manchester) collections will soon be published. 13 T. Arnold and A. Guillaume, Legacy of Islam (Oxford 1965); D. Campbell, Arabian Medicine and its Influence on the Middle Ages (Amsterdam 1926); L.I. Conrad, ‘Arab Islamic Medicine’, in W.F. Bynum and R. Porter (eds), Companion Encyclopedia of the History of Medicine. Vol. I, (London 1993); E.T. Hermann, ‘Early Arabian Medicine’, Bulletin of the Medical Library Association 25 (1936 7), 113 17; M. Levey, Early Arabic Pharmacology (Leiden 1973); E. Savage Smith, ‘Medicine’, in R. Rashed (ed.), Encyclopedia of the History of Arabic Sciences (London and New York 1996). Vol III, 903 62; M. Ullmann, Islamic Medicine (Edinburgh 1978); E. Savage Smith, ‘Tibb', EI, X, 452 60; P.E. Pormann and E. Savage Smith, Medieval Islamic Medicine (Washington 2007). 14 M.W. Dols (trans.), and A.S. Gamal, (ed.) Medieval Islamic Medicine, Ibn Ridwan’s Treatise on the Prevention of Bodily Ills in Egypt (Berkeley, Los Angeles, London 1984). 15 E. Lev, Work in Progress ‘The Research of Medical Knowledge in the Cairo Genizah Past, Present and Future’, in S. Reif (ed.), The Written Word Remains The Archive and the Achievement, Taylor Schechter Genizah Research Unit at Cam bridge University Library (Cambridge 2004), 37 51. 16 Ibn Abi }UÒaybi¨a, 1965. ¨Uyun al Anba}, fi ™abaqat al A†ibba}, Dar Maktab Îayat. Beirut. (Arabic); Ibn al Qif†i, 1903, Tarikh al Ìukama’. J. Lippert, (ed.). (Leipzig).
437
LANGUAGE PASSIVITY IN THE MEDICAL ARABIC AND JUDAEO-ARABIC
fore. The gap between the numbers of Jewish physicians mentioned in the historiographic literature and in the Genizah documents emphasises the importance and advantages of the Genizah as a primary source. Most of the fifty Jewish physicians found to date in the fragments lived and practised medicine in Cairo, with a few in Alexandria and other smaller cities in Egypt, between the eleventh and the thirteenth century. Their titles (all signifying ‘doctor’), according to the Genizah fragments, were al-muta†abbib, al-†abib, Ìakim and ha-rof1. For some of them we even have information on their specialization: eye doctors, a wound specialist and a physician who worked in a hospital. Anyone wishing to study and assess realistically the medical aspects of Mediterranean society in the Middle Ages has no choice but to check authentic, practical knowledge, mainly that which can be extracted from the prescriptions found in the Cairo Genizah. Such information uniquely and exclusively enables us to understand properly the practical medicine in that period.17 Medical treatment usually began with a patient visiting a physician in his clinic, continued with the latter writing a prescription, which was subsequently made up by a pharmacist in his pharmacy. In other cases, the physician saw patients in a rented room at the back of the pharmacy.18 The prescription stage is usually missing from historical records for various reasons: in some cases, the physician made up the formula himself so that no prescription existed, but in most cases, there was presumably no reason to keep the prescriptions, and they were torn up or thrown away. However, we were fortunately able to trace 141 prescriptions among other documents. The most important and interesting information regarding practical medicine is undoubtedly to be found in the prescriptions, as they reflect the medical reality that actually existed. A unique aspect of the information that emerges from the prescriptions is their originality.19 In most cases, they are written in Arabic script (92) and Arabic written in Hebrew script (Judaeo-Arabic) (47)20; very rarely Hebrew 17
E. Lev and Z. Amar, ‘Practice Versus Theory: Medieval Materia Medica Ac cording to the Cairo Genizah', Medical History 51 (2007), 507 26; E. Lev and Z. Amar, Practical Materia Medica of the Medieval Eastern Mediterranean According to the Cairo Genizah (Leiden 2008). 18 Isaacs, Medical, xii xvi. 19 E. Lev and Z. Amar, ‘“Fossils" of Practical Medical Knowledge from Medi eval Cairo', Journal of Ethnopharmacology 119 (2008), 24 40. 20 E. Lev, ‘Medieval Egyptian Judaeo Arabic Prescriptions (and Edition of Three Medical Prescriptions)', Journal of Royal Asiatic Society 18 (2008), 449 64. 438
LANGUAGE PASSIVITY IN THE MEDICAL ARABIC AND JUDAEO-ARABIC
(1) or Judaeo-Persian (1) is found. In a few cases, the prescription is written in Judaeo-Arabic but the benedictions that open and close it are written in Arabic script. A few prescriptions were copied identically, others with slight changes, from famous books such as Minhaj al-dukkan21 or Dustur al-Bimaristani.22 Most of the prescriptions were written on one page, usually on one side of a sheet of paper (very rarely vellum). Prescriptions were often written on reused paper (at times in the margin or in between the lines of other documents or even books). Linguistic Introduction Linguistically, one of the most prominent features which at first glance appears in the medical prescriptions written in Arabic and Judaeo-Arabic of the Genizah is the very frequent usage of passive verbs.23 Those verbs have actually the indirect semantic role of the imperative, yet they are not morphologically imperative but in most of the cases in the passive voice of the imperfect (mu∂ari¨ majhul ), normally in the third person, e.g. [ יסחק אלגמיע ויצא'ף אליהEverything should be ground and added] or in Arabic letters vHBË lOL'« vKG
5MJ »«d 5OË« vK ”d1Ë UOU vHBË ÍbM d9 r «—œ W&L' vK qLF&Ë [everything should be boiled andstrained over five dirhams
of tamarind and strained again and macerated with two uqiyyas of fumitory syrup and oxymel syrup and should be used].24 Therefore those prescriptions have in fact a sense of command, a sort of imperative mood, and in most cases are not in the pure imperative form that is used either in literary Arabic or in the Arabic dialects. This essay tries to determine what we can define as ‘Language Passivity’ (hence LP) in those prescriptions by reviewing the types of passive usage in them. It is worth mentioning that the Language of Passivity has been superseded or, at least, restricted in Early Middle Arabic.25 21 Abu ’l Muna Dawud b. Abi NaÒr Kohen al A††ar al Isra’ili, Minhaj al Dukkan. (Cairo, Bulaq 1940). (Arabic). 22 Daud Ibn Abi al Bayan, al Dustur al Bimaristani (Arabic), in Paul Sbath (ed.), ‘Le Formulaire des hôpitaux d’Ibn Abil Bayan’, Médicin du Bimaristan Annacery au Caire au XIIIe siècle, Bulletin de l’Institut d’Egypte, 1932 3, 15: 9 78. 23 Passive usage is a known phenomenon in Biblical Hebrew as well as in Clas sical Arabic; however the use is not intensive and prominent. 24 T S AS 155.277 verso (A) line 4. 25 See Joshua Blau, A Handbook of Early Middle Arabic. (Jerusalem 2002), 37, §36: ‘In living speech the internal passive has been superseded or, at least, restricted
439
LANGUAGE PASSIVITY IN THE MEDICAL ARABIC AND JUDAEO-ARABIC
Arguments In this article it is intended to define the exact and cardinal meaning of LP in the Arabic and Judaeo-Arabic medical prescriptions of the Cairo Genizah and their range, and by doing so to try and provide some possible answers to the following questions: • Is LP the sole method for expressing the imperative mood in Judaeo-Arabic medical prescriptions in the Cairo Genizah? • Is LP found in other Arabic medical texts in the parallel medieval period or is it merely a Judaeo-Arabic mode of expression? • Are there any other genres in Middle Arabic which resemble the medical prescriptions of the time or can we talk about an exclusive ‘doctors’ language’ or perhaps a unique pharmacological language in the triangular relationship of doctor-patient-chemist? • What are the reasons for using LP, and can one actually talk about motives for its use? • Can the LP way of writing medical prescriptions cast light indirectly on the medical life of the Jewish or perhaps Arab doctors of the early Middle Ages? The LP types of the prescriptions The medical prescriptions in Arabic from the Genizah collection contain a high rate of passive verbs in the first stem fa¨al, yet in the formulation of -u-a- in the prefix tenses of the imperfect.26 Thus we find verbs such as ישרב, יסחק,[ יוכדshould be taken, ground, and drunk]27 and sFË ¨jK0Ô ¨fLKË ¨lIMË [should be soaked, placed, combined and kneaded].28 Infrequently verbs in other conjugation (ASP §47). Therefore, on the one hand, reflexive verbal items may be used instead of the internal passive and, on the other hand, pseudo correct passive forms oc cur, as bÔO1Ó instead of classical bÔ4u Ó 1Ô “it will be found”’. See also Richard Gottheil and William H. Worrell, Fragments from the Cairo Genizah in the Freer Collection. (New York 1927), 44 57, letter 9 lacks passive verbs. 26 See Jan Retsö, The Finite Passive Voice in Modern Arabic Dialects. (Göteborg 1983), 21 2: ‘According to the common view, the “passive” in HA [=High Arabic] is formed by an interior vowel sequence opposed to another sequence in the “ac tive” verb. The same type of morphological opposition is found in many Indo Eu ropean languages and is there called apophony, “Ablaut” (AP). The term apophony has indeed been suggested for Semitic languages, and will be used here as a designa tion for this type of morpheme. This apophonic forms vary in different forms of Arabic. In HA, it is characterized by the vowel sequence u i in the suffix tense (the perfect) and u a in the prefix tenses (the imperfect)’. 27 See for example Or. 1081.J39 Prescription (recto) JA. 28 T S Ar.41.81 [on cloth]. 440
LANGUAGE PASSIVITY IN THE MEDICAL ARABIC AND JUDAEO-ARABIC
types are found, e.g. ¨‰ËUMÔ ¨·UC ¨V9: ¨qLF& ¨vIK ¨nu nH [should be…] and rarely lK9 [to be swallowed].29 As mentioned above, most of the cases include verbs in the third person, a linguistic device which creates vagueness. In most cases, even if the noun is feminine, this type of obscurity prevails as in WËœ_« o:& sFË [the medicines should be ground].30 Unusual is the usage of a substitute formula instead of the LP formulation. Thus we find r before the infinitive, which also tends to be used in the first stem, e.g. 5? r [earth should be put].31 Very rarely is there a correlation between the verb and the gender of the subject, e.g. wMOA —«œ …—UB sF1Ë o:&1 [cinnamon juice should be ground and kneaded].32 This tendency is also found after what one can call ‘semantically generalizing particles’ as [ אלגמיעall of ], e.g. [ יסחק ויצ'אף אלגמיע אליהall should be ground and added].33 Hardly ever does one ever find sentences which start in the singular and continue in the plural such as ותגמע [ אלגמיע ויסחקוall should be ground].34 This can be regarded as a rare phenomenon in the prescriptions, which directly represents mediaeval Judaeo-Arabic grammar and is influenced by vernacular speech.35 The prescriptions comprise a list of less than one hundred verbs, the majority of them, as mentioned, in the LP formulation. From all the entries, the verb which appears the most and decisively is the LP form c'R36 which semantically denotes an imperative mode, ‘take’, whereas literally it could be translated as ‘should be taken’. Ten verbs appear extensively in the prescriptions:
lIM ¨‰ËUMÔ ¨sF ¨a9D ¨V9: ¨qF ¨·UC ¨vHB ¨‚b qLF& o:& [ground, used, crushed, strained, added, put, made into pills, cooked, kneaded, taken and soaked],
all of them37 in LP formulation, i.e. in passive forms in the third person, yet there is no obvious order of verb appearance. Anyhow, the LP formulation c'R may appear several times in one prescription.38 29
See for example T S Ar.42.67 recto. T S NS 297.17. 31 T S Ar.40.141r:[blank verso] line 6. 32 T S Ar.41.125 line 4. 33 Or. 1081.J39 Prescription (recto) line 4. 34 TS 8J 14.3, lines 8 9. 35 See Blau, A Grammar of Mediaeval Arabic, p. 134, §191. 36 See Blau, A Grammar of Mediaeval Arabic, p. 83, §95, p. 85, §99. 37 [To be ground, used, pounded, strained, added, put, soaked, taken, kneaded, cooked, and formed into pills ]. 38 See for example NS 164.88 and T S Ar. 39.451 verso. 30
441
LANGUAGE PASSIVITY IN THE MEDICAL ARABIC AND JUDAEO-ARABIC
Seldom is the normal formulation of LP infringed. The infraction tends to be in the conversion of the imperfect to the perfect, yet the verb remains in the passive and in the third person, e.g. נקעÔ [macerated].39 Sometimes there is disorder in the appearance of the verbs, such as a passive verb in the perfect followed by a passive verb in the imperfect or vice versa, as in vHBË sË [kneaded and stained].40 In this specific sequence one might assume that the first action, that is to say the kneading, was done before, but one is not sure whether it was really like that, since in one prescription the particle b is found before a formula of a passive perfect verb such as wKH« b U0 dC'√ !U“«— tON [in water in which fresh fennel had already been boiled]41. The instance of lK9Ë V9:Ë [formed into pills and swallowed]42 might be explained in a different way, as the verb ‘to swallow’ has to appear semantically in a vague manner in the passive form, since the doctor cannot order the chemist to swallow the remedy. Another infraction of the LP is the infrequent appearance of active verbs in the medical prescriptions both in intransitive and transitive verbs, when a noun appears after the verb. Thus one finds vI9 vO lPdQ« [until a quarter remains]43 or YKSQ« vI9 Ê« vQ« [until a third remains]44 and …bF*« ÍuI U2 [which strengthens the stomach].45 Intransitive verbs are never used in LP, such as —UA or œU, e.g. חתי יציר עלי [[ אל]קואםuntil it reaches (the desired) consistency]46. Some verbs in the fifth and seventh stem have a semantically reflexive meaning, but they appear in the prescriptions in a morphologically active form, e.g. bIF vO [until it thickens].47 It is not clear in the case of some verbs whether they are active or have the LP formula, as for example [ ינצבעto be dyed].48 39
T S 12.33 recto only line 11. T S Ar. 41.71. verso only line 7. 41 T S NS J38 line 6. 42 T S Ar. 41.72 lines 3 4. 43 T S Ar. 41.71. verso only line 6. 44 TS Ar. 39.274 recto only. 45 T S Ar. 39.451 verso line 12. 46 T S 12.33 recto only line 9. 47 T S Ar. 30.99 line 6, verso only. See also Jan Retsö, The Finite Passive Voice in Modern Arabic Dialects. (Göteborg 1983), pp. 28 30: ‘The maõhul forms are not the only means possessed by HA for use in the passive construction. Although de nied by many great authorities on Arabic grammar, it is quite clear that the verbal forms with a t n prefix are used in passive constructions from the earliest periods of HA until the present day.’ 48 T S Ar. 42.15. 40
442
LANGUAGE PASSIVITY IN THE MEDICAL ARABIC AND JUDAEO-ARABIC
Some medical prescriptions of the Cairo Genizah use manÒub in an LP formulation, as for example יגמע בעד אלדקinstead of יגמע בעד ידקÔ *אן. [Should be mixed after being ground].49 Some interesting features of structures are the use of maf¨ul mu†laq after the passive verb in LP, e.g. or [ ויצרב צרבto be hit strongly].50 These patterns appear only in prescriptions written in Arabic with Hebrew letters. Only in texts written in Arabic letters do we find LP structures which are followed by qUN, e.g. ÃU: UX ‚b tœ [should be pounded, what needs to be pounded].51 Worth mentioning is the phenomenon that LP verbs successively tend to appear with changeable prepositions, especially in the case of the LP verb Ãd0 [literally: should be taken out] which can appear with various prepositions, as the following examples show: tOK Ãd0Ë [to be dissolved into it],52 tON Ãd0Ë [literally: to be taken in it],53 tM Ãd0Ë54 [and pour it out]. However, the translation of those variants does not differ significantly. The passivity appears not only in the above morphological form, but also in the passive participle in almost all of the medical prescriptions, as follows: מנזוע אלאעגם מנזוע אלאקמאע
…uHdQ« ŸËeMX ‚u:&X ÷UOP dJ
[cleaned of stones]55 [cleaned of thorns]56 [skinned of froth]57 [powdered white sugar]58
In the examples above, as in most of the collected data, the passive participle is also in the first stem. The entries serve as a substitute for attributes, a form which avoids from the usage of w99 XF.59 The
49 T S 43.338 line 10. Compare Joshua Blau, A Grammar of Mediaeval Judaeo Arabic, p. 180, §270ב: "’דק שי מן מדק" ‘כתש דבר במדוכה. 50 T S Ar.30.16. 51 See T S Ar.42.152 Prescriptions II line 5 and T S Ar.39.458 recto only lines 7 8. 52 T S Ar. 41.72 line 4. 53 T S Ar. 41.71. verso only line 9. 54 T S Ar.39.458 recto only lines 10 11. 55 Or. 1081.J39 Prescription (recto) line 6. 56 Or. 1081.J39 Prescription (recto). 57 T S Ar.42.67 recto line 4. 58 T S Ar.40.87, line 11. 59 See Joshua Blau. A Grammar of Mediaeval Judeo Arabic (Jerusalem 1961, rep. 1995), p. 137, footnote 40.
443
LANGUAGE PASSIVITY IN THE MEDICAL ARABIC AND JUDAEO-ARABIC
prescriptions include expressions like [ עסל מנזוע אלרגוהhoney skinned of froth]60 instead of writing *עסל רגותה מנזועה. Yet, the passive participle appears sometimes in the plural and even in the dual form, a device which is never used in the normal LP type of verb in the prescriptions. Thus we find forms such as a long list of materials, which follows after the passive participle in the plural form, as in the list of materials in T-S NS 327.97-line 6 which is prefaced by 5Qu&GX [washed], or in the long list of T-S Ar. 30.99 verso only 1-line 4, which is followed by the passive participle 5uK&X [boiled (pl.)], or: ÊUËeM*« dHAô«Ë wKPUJQ« ZKOKNQ« sX c'R [to be taken: cleaned chebulic and yellow myrobalam].61 Note that dual forms are usually not normative in practical medical Genizah fragments. While some prescriptions lack even one verb, a large number of prescriptions exhibit the phenomenon of double verbs relating to the same materials which appear before, for example [ ידק ויסחקto be pounded and ground]62 or a9DË vHBË [to be strained and cooked].63 It sometimes happens that the LP relates to a continuing description, e.g. U« wN lNdË vKG [should be boiled and put in a vessel].64 Rarely one finds even a trio of verbs one after the other, as for example:vKG ÊU:dP ”d1Ë vHBË [should be boiled and strained and macerated in basil]65 or even a four verbs, such as: lK9Ë V9:Ë sFË o:& [to be ground, kneaded, formed into pills and swallowed].66 Orthographical Remarks Various remarks can be made regarding the orthography and spelling of LP. In one Judaeo-Arabic prescription an inverted comma is found in the LP verb, thus 'ויוכד' = ויוכ.67 The usage of this orthographical sign can be regarded here as unnecessary since the comma, which denotes an abbreviated word, replaces here only one letter. Another 60
T S AS 148.22 Prescription (recto) line 4. T S Ar.42.67 recto line 2. 62 T S AS 148.22 Prescription (recto) line 4. 63 T S NS J38 line 4. 64 T S NS J34? line 4. 65 T S NS 306.76 recto line 4. 66 T S Ar.42.189 Prescription (recto) line 3. In many prescriptions one can find only one verb relating to a long list of materials, e.g. Or. 1081.J39 Prescription (recto) line 2, whereas various lists do not have any verb at all, though mentioning a vast number of materials without relating or describing the specific action. Such prescriptions do not include verbs at the beginning, for example, T S Ar.43.338 II Two Pres lines 1 11 or T S Ar. 34.239 recto only lines 1 6. 67 TS Ar. 30.65 line 16. 61
444
LANGUAGE PASSIVITY IN THE MEDICAL ARABIC AND JUDAEO-ARABIC
orthographical sign which is used in order to emphasize LP is the literary Arabic vowel pointing of ∂amma68 –u-a- in the prefix tenses of the imperfect, hence ויקטעÔ ,ויתנאולÔ ,נקעÔ ,יטבחÔ [cooked; that have already been macerated; used]69 or WËœô« jK0Ô [the medicines should be mixed].70 In a very rare usage of the pure imperative instead of LP one finds inverted commas which actually serve as a partition before the intervening object or after it, thus , וציר'ה,ציר'הם '[ ואכרגהmake them; make it; put it out!].71 This fragment was lately identified and described as a quasimedical-magical prescription for the establishment of the celestial sphere (pKHQ« WXU≈). Hebrew vowel letters are seldom found in the prescriptions to mark the LP, e.g. [ ויועמל; יועמלshould be done].72 Several spelling errors are found in Judaeo-Arabic medical prescriptions that are written in Hebrew letters, such as in the verb יתחן [to grind], and not as in the prescriptions with Arabic letters s:D.73 Another frequent spelling error in the Judeao-Arabic prescriptions with Hebrew letters is [ יצחקto pound] mirroring o:&.74 Metathesis exists in the Judaeo-Arabic medical prescriptions rarely as in אבטכהinstead of [ אטבכהcook it!].75 Variations exist in the hamza orthography of LP in Arabic letters. Thus we find UHB alongside the correct orthography of vHB [strained] and UI& versus vI& e.g. oO »«dAP tMX UI& [should be drunk with an old fragrant wine].76 With regard to the Arabic texts in Hebrew letters, there are LP forms of [ יגליto be boiled] and יצפי [to be strained] besides the the passives יגלאand יצפא.77 In a depilatory prescription for hairy women, the Cairo Genizah contains LP forms in negation, such as UHD rQ [was not extinguished].78 In general, negation forms are rare in the medical prescriptions under discussion. 68 Compare Moshe Gil, Documents of the Jewish Pious Foundations from the Cairo Geniza (Leiden 1976), p. 442f. This document 132 contains sometimes the pointing of ‹amma and some other times not, e.g. נקלÔ versus קבץ. 69 T S 12.33 recto only, lines 8+11+14. 70 T S Ar.41.81 [on cloth] line 6. 71 TS 8J? 14.3 lines 10!11!12!. 72 T S AS 169.297, lines 3 and 8. See Joshua Blau, A Grammar of Mediaeval Judaeo Arabic, pp. 21 2, §7. 73 See for example T S K25. 116 Prescriptions, recto only line 7. 74 See T S Ar.30.16 and NS 151.52. 75 TS 8J? 14.3 9 line 15. 76 T S Ar. 39.451 verso line 15. See Joshua Blau, A Grammar of Mediaeval Judaeo Arabic, pp. 24 5, §10. 77 T S Ar. 30.305. 78 T S Ar.11.22. The correct High Arabic form should be lam yu†fa and not lam yu†fa.
445
LANGUAGE PASSIVITY IN THE MEDICAL ARABIC AND JUDAEO-ARABIC
General Remarks Numerous prescriptions in Arabic from the Genizah begin typographically so that the verb is the first word in a new line in the written text itself.79 Thus we find for example in T-S Ar.42.189 Prescription (recto), that line 8 opens with the verb lLË [collected], and line 10 with the verb qLF&Ë [used].80 One can explain this phenomenon by supposing that the physician was attempting to explain to the chemist as clearly as possible the sequence of actions needed to implement the prescription. Several prescriptions open with a verb in a formula of greetings which appear above typographically, e.g. c'R tuË t1UcdP vK [to be taken with His blessings and help].81 Substitutes of LP Most of the texts of medical prescriptions from the Cairo Genizah that supply substitutes for the LP structure are written in Arabic with Hebrew letters. The first type of substitute is the use of pure imperative forms instead of using the normal LP formulation.82 Worth mentioning is the remarkable phenomenon that those prescriptions which use an alternative to the LP system once, do it several times in the same text. Therefore, we find in T-S NS 265.62 the following imperative forms: , דוק ואנכול, ושרב, וחבבה, ואעגינוה, דוק ואנכל, כוד, אלתי תכוד, אגעלהם,וצוב וכול, ואעגנוה, ואסקיה, ואגעל,ואעגן
[pour; put them; that take; take; pound and sieve; and knead it; make it into pills; drink; pound and sieve; knead; put into; make him drink; and knead it; and eat!].
The second substitute in the medical prescriptions is unusual and reminds us of Modern Hebrew usage. Instead of using the LP model or its alternative, the imperative, we sometimes find normal imperfect forms, such as ותטינה,'[ ותגפפהםand dry them; and put sand on it!].83 79
Even though this phenomenon is known in Classical Arabic, it is rarely found in texts that have a dialectal character. 80 See also T S Ar. 39.451 verso, lines 7 10 and TS Ar. 30.65, lines 12 13. 81 T S Ar.42.189 Prescription (recto) line 1. 82 The medieval doctor, Assaf ha Rofé, in his book Séfer Assaf ha Rofé, which was written in Hebrew, uses in most of his medical prescriptions pure imperative forms, i.e. קחand not יוקחor תקח. See S. Muntner, ‘Séfer Assaf ha Rofé’, Korot (1965, III, VII, VIII), pp. 706 16 (in Hebrew). 83 See TS 8J? 14.3 9 lines 9+11. 446
LANGUAGE PASSIVITY IN THE MEDICAL ARABIC AND JUDAEO-ARABIC
The third kind of substitution may be ‘a polite form of address',84 which involves the use of the jussive in the apodosis of a conditional sentence. This method of expression is in the third person but in the active:
“u4 c'UOKN œu√ tG9B Ê√ œ«—√ ÊUN [ and if he wants to dye it black, let him c'ROKN Ô .
take a nut]85 instead of *“u4
The Question of LP in Some Other Medieval Arabic Medical Prescriptions Al-AÒma¨i (eighth century CE), in his book describing various flowers, plants and trees, makes extensive use of the LP formula ‰ Ô UIÓÔ [it is said].86 In the Medical Formulary or Aqrabadhin of al-Kindi87 from the ninth century CE, we see that most of the prescriptions tend to open with the verb c'u [to be taken] as in the Genizah medical prescriptions in Arabic, e.g. c'u q0MQ«Ë ‚bQ« bFP c'u1 [should be takenshould be taken after pounding and sieving].88 Sometimes a list of several medical materials appears after this verb, without any other verbs.89 The LP of al-Kindi differs from that of the Genizah prescriptions in following the grammar of literary Arabic with regard to the dual. Where the Genizah’s LP ignores the dual, al-Kindi preserves it as follows:
U0 90UMFË dd(UP Êö:Ë ÊUI:& e4 bO«Ë qc sX su'ô« ÂœË w: `KX [dough salt one part, dragon’s blood one part. Both are pulverized and sieved with a silk cloth. Both are kneaded with water]91
As in the Genizah’s medical formulary, al-Kindi’s prescriptions use the alternative to LP, the pure imperative, e.g.: dFP nc c' wQUP 84 Compare Geoffrey Khan, Arabic and Administrative Documents in the Cam bridge Genizah Collections (Cambridge 1993), p. 441 on T S Ar. 40.41, line 1: oKD pQ– ‘Let that be granted’. 85 Ar. 40. 66. 86 al AÒma¨i, Abu ∑a¨id. Kitab al Nabat. (Cairo 1972). (Arabic). The aim of this book is not to survey prescriptions and medicaments. 87 See Martin Levey, The Medical Formulary or Aqrabadhin of Al Kindi (Madi son, Milwaukee 1966). 88 Ibid, pp. 35, 45, 67. 89 See for example pp. 41 3. 90 In this case we can see an example of the dual form of the verb which is not normative as in Classical Arabic. 91 Ibid., pp. 101, 109. al Kindi preserves the dual also regarding the na’ib fa{il, e.g. Ê«dD rHQ« wN l{u [two drops should be put in the mouth ].
447
LANGUAGE PASSIVITY IN THE MEDICAL ARABIC AND JUDAEO-ARABIC
[take a handful of decayed dung].92 He uses also the substitute form of addressing the reader of the prescription in the active in the second person, e.g.: UL dBFN WCXUO WUX—Ë …uKO WUX— c'U1 [You should take a sweet and sour pomegranate and extract both of them].93 Ibn Sina (980–1037 CE) in his book al-Qanun94 makes extensive use of the LP structure. In the prescriptions (WHA) there, we find a usage of the LP verb c'u, e.g.:
‰u9P »dAON ¨r —œ nB Íd9Q« ÂUL(« ¡d' c'R ≠ qI*«Ë ¨ZMO9J&Q« c'R 95 Æ—QHQ« ¡d' c'R Ë√ÆÆÆ—bON ¨‰UH?_« [(Those) should be taken: sagapenum, bdellium… should be taken also a part of a half a dirham of a wild pigeon and it should be drunk with childrens’ urine… or a part of the faeces of a mouse should be taken].
Ibn Sina varies his style sometimes from the normal LP formula by giving other medical prescriptions (WHA) which begin with Ê√, e.g.: ‰UISX vK WKUIQ« —eP sX »dA Ê√ [and drink one mithqal of lesser cardamon].96 It is noteworthy that when the author turns to ‘treatment’ (Ãö), he tends to use language that lacks the LP structure and rarely varies from the structure that uses Ê√.97 Ibn Ri∂wan of the eleventh century CE, in his translation of Galen’s Qa†i†riun, regarding medical transcriptions, makes abundant use of the style of sentence that begins in Arabic with Ê√ wG9M b or wN wG9M [you should], e.g.: 98
tQUO vK wU9Q« rOK&Q« nJQ« ”«— wN ôË√ X9S1 Ê« wG9M b‚
92
Ibid., p. 45. Ibid., p. 183. 94 See Ibn Sina, Abu ¨Ali al Îusayn. Al Qanun fi al ™ibb. V.2 (Beirut 1993). (Arabic). 95 Ibid., pp. 1572 3. 96 Ibid., pp. 1577 8. 97 See for example p. 849. 98 Malcolm Lyons, Corpus Medicorum Supplementum Orientale: Galeni. (Berlin 1963), p. 102. See also ¨Obadyah b. Abraham b. Moses Maimonides, The Treatise of the Pool (Al Maqala al Îaw‹iyya). Paul Fenton (ed.). (London 1981): the grandson of Maimonides uses rK≈ several times [you should know, literally: know!] in the imperative. As a substitute, yet rarely, he uses the structure pQ wG9M UX ‰Ë√ [First thing that you should do]. An extensive usage of wG9M is found also in M.C. Lyons, Kitab Tadbir al Amra∂ al Îadda li Buqra† (Hippocrates: Regimen in Acute Diseases). (Cambridge 1966), pp. IX, 8 9, 29 30. Lyons believes, that the original author of the book was ¨Isa bin YaÌya. The text does not show LP structures. Lyons, as an editor of another book of Hippocrates, Kitab Buqra† fi Îabl ¨ala Îabl (Hippocrates: 93
448
LANGUAGE PASSIVITY IN THE MEDICAL ARABIC AND JUDAEO-ARABIC
This sentence is translated into English as follows:
‘You must99 first take careful note of the elevation of the sound acromion process’. Ibn Ri∂wan uses the LP formulation infrequently and it seems that he imitates the passive voice that appears in the Greek source.100 This type of style expresses politeness, and it substitutes for the semantic imperative, whether the LP model or pure imperative. Maimonides (1138–1204 CE), in his book which includes two articles about coitus,101 uses the LP structure widely when referring to medical prescriptions. When the text is descriptive, e.g. when he describes problems of sexual intercourse, it lacks the LP structure. In his medical prescriptions, Maimonides also tends to begin with the LP verb c'R, e.g.: On Superfoetation). (Cambridge 1968), Page II, says: ‘This Arabic version is of in terest as exemplifying the work of a translator probably earlier, and certainly of less ability, than those, such as Îunain b. IsÌaq, IsÌaq b. Îunain and Îubaish, who made translation a science’. This source uses wG9M extensively for describing diseases. While referring to medicaments, e.g. p. 23, he uses the pure imperative forms, in a medical prescription given to strengthen women after giving birth:
»«dAP UNM«Ë «bO4 ULNI:«Ë ¡e4 bO«Ë qc sX ÊuLcË ÂuË ÊËdDË œu√ 51 sX «—UO9 UNQ lMA«Ë pQ– sX lMA«Ë »«dAP pQ– s«Ë ULF tI:«Ë °qEM(« —uA sX c' Ë√ «—UO9 pQ– sX lMA«Ë iOP√ ÆtKL% Ê√ U dXË V—√ dFË ·uBP «—UO9 [‘Make her a suppository of black figs, nitre, garlic and cumin, one part of each: grind them thoroughly, knead them with white wine and make a suppository of them; alternatively, take colocynth skins, grind them well, knead them with wine, make a suppository of them with wool and hare’s fur and order the woman to apply it’]. Yet, in p. 24, he varies his style, when he prescribes a treatment for women who have womb pains, by addressing the doctor in the imperative second person:
UNMF1Ë ULF Uœ UNb1Ë ¡e4 bO«Ë qc sX œu« 51Ë ÊuLcË ÊËdD c'Q1 Ê√ pQ–Ë «—UO9 UNQ lMB1Ë Æ»«dAP [‘You should also make her a suppository: take nitre, cumin and black figs, one part of each, bruise them very fine and knead them with wine’]. See also Lyons, the editor of Kitab Buqra† al Maruf bi Qa†i†riun ay Îanut al ™abib (Hippocrates: In the Surgery). (Cambridge 1968), p. 18: œœ—UN WcuNM*« ¡UC_« UX√ UNDP—« r dOSc ¡w `O:BQ« l{u*« sX UNQ [‘As for emaciated limbs, you should bring back to them a considerable amount of the sound parts and then bandage them’]. Here we can see that the translation from Greek into Arabic uses, after UX√ in the predi cate, the pure imperative form. This is not a mere prescription book, but a source that describes diseases. In another book which Lyons edited, Kitab Buqra† fi al Amra∂ al Biladiyya (Hippocrates: On Endemic Diseases [Airs, Waters and Places]) (Cambridge 1969), we witness a language pure of LP formulations. 99 A better translation would be: ‘You should…’ or ‘It is appropriate’. 100 Ibid, p. 171. 101 See Kroner H. Ein Beitrag zur Geschichte der Medizin des XII. Jahrhunderts an der Hand zweier medizinischer Abhandlungen des Maimonides. (Berlin 1906). 449
LANGUAGE PASSIVITY IN THE MEDICAL ARABIC AND JUDAEO-ARABIC
יוכ'ד' מן בזר אלראזיאנג' ובזר אלבאדרנג'בויה ויחבב כאלבנדק וימסך פי אלפם 102 לילא כאן או נהארא
[man nimmt103 von Fenchelsamen und Melissasamen und macht eine Pille gleich einer Haselnuss und haelt (sie) im Munde, mag es Nacht oder Tag sein.]; or: לטוך' אכ'ר יוכ'ד' אדמגה אלעצאפיר אלד'כראן ואנפחת אלארנב ואנפחת אלעזאל ואנפחת אלכ'יל ואנפכחת אלג'מאל אלערביה אג'זא סוא ידק אלג'מיע נאעמא ויעג'ן 104נאעמא ויעג'ן בדהן אלזנבק וילאך' אלקאן ומוצ'ע אלכליה
[Eine andere Salbe: man nimmt Hirn der maennlichen sperlinge, Kaeselab der Hasen und der Gazellen und der Pferde und der arabischen Kameele!, zu gleichen Teilen; es wird das Ganze sehr fein zerreiben und geknetet mit Jasminoel und es wird gesalbt der Teil unter den Hueften und die Stelle der Niere].
The Jewish chemist Cohen al-¨A††ar (twelfth century CE), uses all the styles to express the semantic imperative in his book Minhaj adDukkan. He can express it by the LP, by addressing the reader in the active in the second person and by the imperative. Some of these have complicated verb forms that require the use of pointing. The author uses also the manÒub in the passive, especially with the formula c'R Ê√ at the beginning of some prescriptions. This usage is unique to al-¨A††ar. A good example of a text which contains all those varieties would be: • passive manÒub formula
‰U?—« …dA wIMQ« dJ&Q« sX c'R Ê√ [ten ra†l of pure sugar should be taken] • a mainly pure imperative sequence
·uA Wd' sX tHÏAË Ó tÔÚËÎÏ dNÓ\UOU XbQ« vQ≈ Áb√Ë [make it gentle and strain it in a woollen cloth and put it back into the pot] • a mainly LP sequence
WMOQ —UMP ”d1Ë a9DË tKOQË Âu »c UX tQUSX« WFP—« wN lIMË [soak in four times its [quantity] of sweet water for a day and a night; then cook and macerate over a gentle fire]105 102 103
Ibid, p. 5, prescription N. Note that the German text does not translate the verbs as imperative or pas
sive. 104 105
Ibid, p. 6, prescription T. See Leigh N. Chipman and Efraim Lev, ‘Syrups from the Apothecary’s 450
LANGUAGE PASSIVITY IN THE MEDICAL ARABIC AND JUDAEO-ARABIC
Unlike other physicians, we find in the formulary of Ibn Abi 'l-Bayan from the thirteenth century, in his book al-Dustur alBimaristani,106 that the preface of most of the prescriptions is in the active voice, while the prescription itself is written in LP. The preface itself is a description of the diseases for which the described medicine is made:
r «—œ …dA bO«Ë qc sX qBPË Âu c'R \VKJQ« VKJQ« WC vK qF ¡«Ëœ … r d*Uc «œuF Ê« vQ« ÊUb [a medication administered against the bite of a dog, rabies: Garlic and onion should be taken each of them ten dirhams; they should be pounded until they become like an ointment …]
Is There Any Connection Between Medical Prescriptions and Cooking Recipes? While reading Maimonides in Arabic in his book on coitus, one notices that the recipes for increasing sexual desire, which are in Hebrew, rather resemble his medical prescriptions in Arabic. Yet, those ‘semi-culinary’ recipes present more flexible patterns, as Maimonides can use יוקחin Hebrew [literally: to be taken] instead of תקח. This form resembles the LP Arabic form of c'R e.g.: 'יוקח ביצי התרנגולת ג' אונק
[Man nehme107 Henneneier 3 Unzen]
Versus: תקח ד' בצלים ותצלם בתנור עד שיתבשלו
108
[Man nehme vier Zwiebeln und brate sie im Ofen bis sie durchgebraten sind]
Maimonides uses various styles in those recipes and can sometimes use the address form in the second person, instead of using active verbs, and also the LP structure, in one recipe, e.g.: 109…חלק
הזרעונים וישרם במי הארוקא עד שיהיו כדי שיסור קליפתם ויוקח ממנו
[Man nehme Erbsen und weich sie in Erucewasser solange, bis ihre Schale sich loest, und nehme davon ein Teil…] Shop: a Genizah Fragment Containing one of the Earliest Manuscripts of minhaj al Dukkan, Journal of Semitic Studies, 51:1(2006), pp. 144, 151. 106 Paul Sbath, Le Formulaire des Hôpitaux d’Ibn Abil Bayan (Cairo 1933). 107 Ibid., p. 55,c: Here we see the subjunctive in German. Compare:[man nimmt], ibid., p. 6. 108 Ibid., p. 22, VI. c4. 109 Ibid., p. 23, VII, b. 451
LANGUAGE PASSIVITY IN THE MEDICAL ARABIC AND JUDAEO-ARABIC
Having discovered that Maimonides used in his ‘semi-culinary’ recipes a variety of styles to express the semantic imperative, we investigated whether other cookbooks from the Middle Ages display the same patterns. An anonymous writer from the thirteenth century CE, who gives us numerous recipes from the Maghreb and Andalusia, was one of the sources investigated.110 In this source also a mixture of various styles is found to express the semantic imperative. Most of the book uses LP formulations, e.g.:
qFË ¨UIOœ lDIË
…oMFQ«
5L&Q« r:KQ« sX c'R¢ ∫WOAOA(U ·dF WFM 111 ÆWKBP ¡UX lX —b wN
[a dish called Îashishiyya: the [following components] should be taken: a fatty flesh out of the [animal’s] neck…it should be cut into small pieces and put in a pot of onion water]
Also when the anonymous author indicates explicitly that the socalled ‘recipes’ can be used as a ‘medical prescription’ (WHA), LP language is found, e.g.
a9DË ¨¡U*« sX ‰U?—√ W&L' vQ≈ ·UCË ¨q&FQ« sX q?— c'R ∫q&FQ« ¡UX WHA q&FQ« vI9Ë ¨¡U*« V c vO [a recipe for honey water: a ra†l of honey should be taken and added to five ra†ls of water; it should be cooked until the water evaporates and the honey remains]
Yet, we occasionally find that the author uses intact recipes, with the address form in the active imperfect, as for example: 110 Anonymous, Kitab al ™abikh fi al Maghrib wa l Andalus fi ¨AÒr al MuwaÌÌidin (Madrid 1961 2). See also Elizabeth R. MacGil, An Example of a Pri mary Source: This Booke! Of Sovereigne! Medicines1, in Bart K Holland, Prospecting for Drugs in Ancient and Medieval European Texts: A Scientific Approach. (Amster dam 1996), pp. 39 40: ‘This Booke of Sovereigne Medicines (ca. 1570) is a collection of medical recipes which survives in five manuscripts copies. …The remedies are presented in an or derly way with marginal titles and heading above the text… The style of treatments reflects the Benedictine Custom.’ In page 45, we find this text, which merely uses morphological imperatives in English: ‘Dropsy. Take half a handfull! [Old English] of sentuery, a handfull of wormwood, a handfull of Annyseedes!, and seeth them altogether in a pottell of running water vntil the half be consumed! streyne it! Drink it euery euening! morning bloud warme’. 111 Ibid., p. 104.
452
LANGUAGE PASSIVITY IN THE MEDICAL ARABIC AND JUDAEO-ARABIC
—UMQ« vK UNKF&Ë …UHOE ΠΫ—b c'Q1 [take a clean pot … and put it on the fire ]
Another source from Andalusia that provides us with culinary recipes from the mid thirteenth century is al-Tajibi’s cookbook.112 This book, like the former one, shows the same patterns of expressing the semantic imperative, i.e. LP formulations, besides pure imperatives, e.g.: 113
(LP)
VODQ« bOL&Q« s:D
[a good semolina should be ground]
Versus: 114
ÎU:KX tOK qF&Ë Î«—ËbX tFDI1Ë ÁdAI1Ë bd1 UX —b ÊU$–U9Q« sX c'Q1 (Addressing in the active imperative in the second person) [you should take a desired portion of eggplants and peel them and cut them into rounded pieces and put salt on it]
When al-Tajibi prescribes ‘semi-culinary’ recipes, i.e. recipes which serve also for medical purposes, he turns to the LP structure: 115
qHd sX qOUSX WFP—√ c'R ∫WLb« `z«Ëd« VcË Íb_« wIM ‰uU%
[a soap which cleans the hands and removes fatty smells: four mithqals of cloves should be taken].
Conclusions As was seen, the prevailing language of the medical prescriptions in Arabic from the Cairo Genizah is a ‘technical language’116 which uses verbs which have semantically the meaning of the imperative, but morphologically have the structure of the passive form of the imperfect tense in the third person. In the minority of the texts, the existence of a variety of other modes of expressing the semantic imperative was also attested. The language of the medical texts which 112 al Tajibi, Ibn Razin. Fa∂alat al Khuwan fi ™ayyibat al ™a¨am wa’ l Alwan2. (Beirut 1984). 113 Ibid., p. 79, recipe 28. 114 Ibid., p. 234, recipe 18. 115 Ibid., p. 279, recipe 17. 116 Compare D.R. Langslow, Medical Latin in the Roman Empire (Oxford 2002), p. 6, 1.2.1: ‘Technical Language and Technical Vocabulary: I have spoken thus far of technical language, and deliberately so. Some linguists have emphasized that, if we are to use the label “technical language” sensibly, we should characterize a technical variety at all levels of the grammar, and not just as a special lexicon’.
453
LANGUAGE PASSIVITY IN THE MEDICAL ARABIC AND JUDAEO-ARABIC
were translated by Arabs from Greek sources and do not serve as pure medical prescriptions, do not manifest the LP formula which so much characterizes our medical formulary, since they use other modes of address, especially the active voice. Yet, texts of Arab physicians, whether Muslims, Christians or Jews, do not differ significantly117 from the Genizah medical prescriptions in regard to LP when they prescribe, though when they describe diseases, they turn to a more active language which also obliges the use of active verbs and accusatives. Although the frequency of using the LP formulation in Arabic in the Genizah prescriptions is higher, texts written in Judaeo-Arabic with Hebrew letters tend more towards the pure imperative comparison with other medical prescriptions written in Arabic letters.118 In this article, we have found a linguistic similarity between the medical usage of LP and the culinary recipes, especially those that are termed as WHA, i.e. recipes that serve simultaneously not only as tasty dishes, but also offer medical benefits, such as dishes that strengthen sexual desire. What are the reasons for using the LP formulation so much in both the Genizah formulary and the other medical prescriptions written in Arabic in the early Middle Ages? There are several possible answers to this question. Firstly, the use of the passive in the medical formularies probably is a polite form of address; ‘you should do it’ and not ‘do it’. Secondly, the variant ways of expressing the imperative are introduced to avoid monotony of style. 117 See I. Friedlander, Der Sprachgebrauch des Maimonides: ein lexlikalischer und grammatischer Beitrag (Frankfurt 1902), p. IX: ‘Wir wissen, dass schon in sehr frueher Zeit das Arabische den Juden zur Umgangs und Muttersprache geworden war, dass aber das Hebraeische selbst in den guenstigsten Zeiten nicht die Grenzen einer wenn auch ziemlich weit verbreiteten Gelehrtensprache ueberschritt’. Com pare also Kees Versteegh, The Arabic Language (Edinburgh 1988), p. 121: ‘We do not know when the colloquial language of the Jews became Arabic, but it must have been rather early. The first literary works written in Arabic by Jews date from the period after the year 1000; the majority of them have been found in the Cairo Geniza’. 118 S.D. Goitein, A Mediterranean Society. (Berkeley 1999), p. 18: ‘Still, the Arabic of the Geniza records may be considered a special language. It was written in a script other than that of classical Arabic by persons who had not absorbed the Qur’an, the holy book of Islam, from childhood. This made for a language less gov erned by traditional Arabic grammar and vocabulary. The Genizah papers reflect the living language, and they constitute, in their great variety of styles and local idi oms, a first rate source for the history of the Arabic language’.
454
LANGUAGE PASSIVITY IN THE MEDICAL ARABIC AND JUDAEO-ARABIC
A third reason for using more LP than imperatives might be the complexity of the conjunction of the Arabic verb in the imperative, which sometimes obliges pointing in ∂amma or shadda, as seen in some medical prescriptions of the Genizah. A fourth reason may be as follows: the doctor writes the prescription to the chemist, yet some of the actions mentioned in the formulary address the patient and not the pharmacist. The physician guides the chemist how to prepare the formulary. It is the physician who instructs the patient to swallow the medicine, so the LP plays here a role in preventing misunderstanding. This is why the medical prescription is written in the third person on the one hand and in the passive on the other. The fourth reason mentioned above reinforces the view that in most cases the medical prescriptions of the Genizah written in Arabic involve a triangular relationship viz., the physician-the patient-the chemist: thus, we need a direct imperative in the physician-pharmacist relationship and an indirect imperative in the physician-chemist-patient relationship. If the medical texts had only the double form of address, i.e. physician-pharmacist or physician-physician (for example: information about diseases or instructions in a book or a notebook), then the direct (or ‘pure’) imperative could be used. Therefore, it has to be asked whether those prescriptions written in Arabic with Hebrew letters, which use the proper imperatives, are in a fact notebooks and not medical prescriptions. Overall, it can be assumed that the texts of medical prescriptions in Arabic from the Genizah do not differ much from the existing Arab style of the era under review. Hence remarks, such as Goitein’s, that there is no evidence of Jewish hospitals in Egypt at the time can be understood and it may therefore be assumed that the Jewish doctors tried to use the same technical language and style as their Arab counterparts.119 Thus, it can be seen how a purely linguistic phenomenon, such as the use of LP formulations, serves indirectly to help us understand better the professional aspect of the practice of medicine in eastern Mediterranean society in the Middle Ages. 119
Ibid, pp. 141 2: ‘No reference to a Jewish hospital has been found thus far in the Genizah material. This is somewhat surprising, for both the Genizah records and Arabic sources mention many Jewish doctors working in what could be called government hospitals, namely, those established by Muslim rulers; there are no Genizah references, however, to Jewish patients making use of the Muslim hospi tals. … Physicians treated indigent patients at no charge provided the sick could make their way to the physicians. Both Maimonides and his great great grandson David b. Joshua gave medical advice and help to poor persons with whom they had dealings in their capacity as heads of the Jews.’ 455
LANGUAGE PASSIVITY IN THE MEDICAL ARABIC AND JUDAEO-ARABIC
A Sample of a Prescription Two recipes and advice on the special diet to be taken by an invalid; the first recipe begins with a basmalah. The text is written in naskhi script on one stained leaf of paper (27.1 x 19.1 cm); 15 lines; verso is blank. There is a frequent use of LP in the text. (Selective examples of LP type appears in bold). T-S Ar. 42.189 Prescription (recto)
tuË tcdP vK c&R rOOdQ« sLOdQ« tKQ« r&P r —œ bO«Ë qc sX wMOA bË«—Ë w«dFQ« sX ÊuI—UH ‰UISX «dHON ×U V()Ë s+FË o)- o.«œ bO«Ë qc sX …œuL:XË ‚—“« qIXË UCOP «dOScË Êu&O« bU `9BQ« bM V-M WXbIX WOLO vK d:&Q« wN ‚u1 nB wN lK(Ë ÊU&Q r «—œ WFP—« bO«Ë qc sX ŸËeMX dHA« ZKOK Ë ŸËeMX wKPUc ZKOK … vK r —œ !U“«— —eP r «—œ WSK bO«Ë qc sX UPbM dAË ”u ‚dË wXU —u … ÷UOP dJ WOË√ vK wHBX ÊU:— —ePË œ—Ë »«d vK tMX lL+Ë W4U4œ Ë√ ·Ëd' r( e4 ëbOH« «bGKQ tMOFP WPdAQ« tUN nB wN qLF3-Ë ”œu'uD« r «—œ WSK ×U c&R WOHOA dOË UAQ« ÕU— lOL4 wN ‰bFË «dOScË »dAË r —œ bO«Ë qc sX 5ËeMX dHA«Ë wKPUc ZKOK Ë ÊuI—UHË wQUOQ WSK wN lK(Ë V()Ë s+FË o)-Ë r —œ nB bO«Ë qc sX UCOP dN qc wN WOQ«uX tKQ« sX ÊuJ dO)«
.1 .2 .3 .4 .5 .6 .7 .8 .9 .10 .11 .12 .13 .14
T-S Ar. 42.189 Prescription (recto) Translation: 1. 2. 3. 4. 5.
In the name of God, the Merciful, the Compassionate. It should be taken with His blessing and help. Hiera picra, one mithqal; Iraqi agaric and Chinese rhubarb, of each one dirham; Aniseed and white tragacanth and blue bdellium and scammony, of each one daniq; grind, knead and form into pills; Swallow in half a tawq at daybreak with advanced diet correlated in the morning… … cleaned chebulic myrobalan and cleaned yellow myrobalan, of each three dirhams; 456
LANGUAGE PASSIVITY IN THE MEDICAL ARABIC AND JUDAEO-ARABIC
6. 7. 8. 9. 10. 11. 12. 13.
14.
borage and liquorice and endive peel, of each three dirhams; fennel seed, one dirham; strained over one uqiyya of white sugar … and mix some of it with rose syrup and basil seed. For lunch: ceruse: one part, lamb or chicken and use in half, and that is the soup itself and balance in all the winter winds and in the summer heat; take hiera, three dirhams, lavender and agaric and chebulic and yellow myrobalan, both cleaned, of each one dirham and it should be drunk and white tragacanth, of each half a dirham. It should be ground and kneaded and formed into pills and swallowed on three consecutive nights in each month. Good is from God.120
120 Special thanks to our colleagues at the Taylor Schechter Genizah Research Unit at Cambridge University Library, Dr B. Outhwaite (director), and Prof. S. Reif (former director) who supported us. We thank Dr A. Shivtiel for sharing with us his enormous knowledge and experience. This research would not have taken place without the generous grant of St. John’s College, Cambridge that hosted Dr Efraim Lev as an Overseas Visiting Scholar (2003 4). The authors would like to thank the Syndics of Cambridge University Library for permission to publish the Cairo Genizah fragment presented in this article.
457
LANGUAGE PASSIVITY IN THE MEDICAL ARABIC AND JUDAEO-ARABIC
458
THE TERM ISM IN MEDIEVAL ARABIC GRAMMATICAL TRADITION: A HYPONYM OF ITSELF1 ALMOG KASHER BAR-ILAN UNIVERSITY
Abstract The term ism in the Arabic grammatical tradition is a hyponym of itself. In other words, it can refer to the entire class of words belong ing to the part of speech ism, or to just a subclass of this part of speech. In the latter sense it always stands in opposition to one of three other subclasses of this part of speech, viz. Òifa, maÒdar and Âarf; the subclass defined by the former is more focal than the one referred to by each of the latter. This interpretation resolves a difficulty stem ming from cases where ism and one of the three other terms refer to two mutually exclusive classes, although the latter always refers, in the contexts in question, to a subclass of the part of speech ism.
1. Introduction In an article on aspects of terminology in Arabic grammatical tradition, Peled (1999: 51) points out that ‘[modern scholars] lament the fact that the Arab grammarians’ phraseology fails to comply with modern scientific standards of technical vocabulary’, such as ‘[c]ases … where a number of terms are used to designate a single grammatical category, and others where one term designates more than one category, sometimes at different linguistic levels’. The phenomenon which will be discussed here, namely terms which can be their own hyponyms, is akin to the second type of difficulty noted by Peled, although it also possesses features which render it worthy of a study of its own. I will restrict myself here to the term ism,2 mainly in view This article is based mainly on Ch. 2.1 of my Ph.D. dissertation The Âarf in Medieval Arabic Grammatical Theory (Hebrew), which was submitted to Bar Ilan University in November 2006. The dissertation was carried out under the supervi sion of Dr Shlomit Shraybom Shivtiel (Bar Ilan University) and Dr Yishai Peled (Tel Aviv University). 2 The term ism is normally rendered by modern scholars as ‘noun’. As will be 1
459
THE TERM ISM IN MEDIEVAL ARABIC GRAMMATICAL TRADITION
of its relationship with the term Òifa.3 Due to Sibawayhi’s (d. c. 793) unquestionable prestige within the Arabic grammatical tradition, as well as the paucity of significant differences between his al-Kitab and later writings concerning the issues at stake here, I will first, in Section 2, examine the relationship between ism and Òifa in that work. Then, in Section 3, I will note several innovations introduced by later grammarians regarding this subject. In Section 4, I will discuss very briefly the relationship obtaining between ism and maÒdar and between ism and Âarf, in order to corroborate the interpretation presented in the two preceding sections. 2. Ism vs. Òifa in Sibawayhi’s al-Kitab The first topic Sibawayhi treats in his al-Kitab is the division of Arabic words (or morphemes) into three parts of speech (Sibawayhi, al-Kitab I: 1). His classification, albeit (as far as we know) unanimously endorsed by his followers, raises many theoretical problems,4 not least of which is the abundance of cases in al-Kitab, as well as in later grammatical treatises, where the phraseology implies that the classes referred to by the terms Òifa and ism form a mutually exclusive pair, despite the fact that Òifa refers (in the contexts in question) to a subclass of the part of speech ism.5 The following case will suffice as an illustration (further cases will be mentioned in the course of the discussion below): when Sibawayhi (ibid, II: 352) deals with the word pattern fay¨ulan, he states that it is utilized both for isms and Òifas. In other words, the terms ism and Òifa are regarded here as referring to two distinct word classes, although all the words which possess the pattern in question, including those classified as Òifas, are undoubtedly subsumed under the part of speech ism. I believe that the key to the resolution of this apparent contradiction is found in the following remark by Lyons (1977: 308): shown below, this translation is somewhat misleading. Moreover, since it is the quest for the distinct senses conveyed by this term which is the very goal of the present article, no single Western equivalent will be offered here. For a discussion of the various viewpoints of modern scholars towards the use of Western equivalents for indigenous terms, see Suleiman 1999: 9 11, n. 2. 3 This term also conveys, as a technical term, several distinct senses, two of which will be dealt with here. Its polysemy makes it impossible to translate it with a single Western equivalent. 4 See, for example, al Zajjaji’s (d. c. 951) treatment of the parts of speech, dis cussed by Suleiman (1990) and Versteegh (1995: passim). 5 Such a case was noted already by Diem (1974: 320), although the way he re solves it differs from the one which will be proposed below. 460
THE TERM ISM IN MEDIEVAL ARABIC GRAMMATICAL TRADITION
So far we have discussed hyponymy (and quasi-hyponymy) under the assumption that it is necessarily an irreflexive relation … But this assumption is questionable in the light of what has now been said about semantic marking: that the unmarked member of the opposition has both a more general and a more specific sense according to context. Since ‘dog’ is sometimes in contrast with ‘bitch’ and sometimes superordinate to it, it follows that in certain circumstances ‘dog’ can be a hyponym of itself.
In what follows I will try to show that the term ism is in fact a hyponym of itself; that is, alongside its denoting a whole part of speech, the term rather frequently refers to a subclass of this part of speech. In the latter sense, however, it always stands in contrast to another subclass of the part of speech ism, to Òifa in our case (in one of its two senses, as will be shown below), but also to maÒdar or Âarf in others.6 Put differently, the exact extension of the term ism in its more specific sense is not fixed but dictated by the other subclass to which it is contrasted, as well as by the context. Before we begin our main discussion of the relationship between the terms ism and Òifa in al-Kitab, two remarks concerning the latter are in order: First, two other terms are used by Sibawayhi as synonymous with Òifa, viz. waÒf and na¨t. I found no distinction whatsoever among these three terms in al-Kitab.7 Secondly, as has been mentioned above, the term Òifa is polysemous:8 it can denote either a word class or a syntactic function, roughly corresponding in the former sense to ‘adjective’, in the latter to ‘attribute’.9 As ‘adjective’ it applies not only to adjectival attributes but also to adjectives implementing other functions; similarly, in the latter sense, it applies to words (not necessarily adjectives), phrases and clauses functioning syntactically as attributes, e.g. verbs functioning as asyndetic relative clauses. The non-attributive use of adjectives is of relevance here:10 it follows that words classified as Òifas (at the word class level) may assume syntactic functions other than the one labelled Òifa, e.g. the predicate. 6
These two cases will be dealt with briefly in Section 4. For a different view, cf. Diem 1974: 315. 8 The word Òifa recurs as a metagrammatical intuitive term, whose semantic scope covers both its meaning as a technical term and the everyday concept under lying it, namely ‘description’ (see Peled 1999); this usage is beyond the scope of this article. 9 Cf. Mosel 1975: 287; Owens 1990: 66 9; Talmon 1997b. 10 For attributes not in the form of adjectives, see Mosel 1975: 295ff.; Owens 1990: 66 9. 7
461
THE TERM ISM IN MEDIEVAL ARABIC GRAMMATICAL TRADITION
For example, Sibawayhi (al-Kitab II: 344) illustrates the word pattern if ¨awl with the sentence innaha la-isÌawfu ’l-aÌalili (‘[the shecamel] is wide in the orifices of the teats’),11 in which a word of this pattern functions as predicate (more accurately, as its head).12 Moreover, a Òifa may also function as a circumstantial qualifier, i.e. Ìal (ibid, I: 65, 96, 162–3, 168, 169) and as a Âarf, a term which roughly corresponds here to ‘adverbial’ (ibid, I: 96). In view of these cases, as well as the numerous occasions on which Òifa in al-Kitab is treated as a word class (as we shall see below), the latter usage should not be regarded as secondary to its syntactic sense.13 Interestingly enough, it can happen that two tokens of this term occur within the very same discussion, one denoting ‘adjective’, the other ‘attribute’: right after he points out (ibid, I: 167) that it is ‘ugly’ to modify an indefinite noun with an attribute which is not a Òifa, here in the sense of ‘adjective’, Sibawayhi states that it is not ‘good’ to use nouns as Òifas, this time in the sense of ‘attribute’. Another such example appears in Ch. 415 (ibid, II: 181): Sibawayhi states that numerals between three and ten should not be annexed to waÒfs in their plural form, e.g. qurashiyyuna (‘Qurayshites’), waÒf here clearly denoting ‘adjective’. He then goes on to point out that this proves that in the phrase thalathatu nassabatin (‘three great genealogists’), with which he dealt earlier (ibid, II: 179), nassabat (classified as a Òifa) functions as if it were waÒfu ’l-mudhakkari (‘waÒf of a masculine noun’). That is, he posits the underlying structure (conveying the same meaning as the one above) thalathatu rijalin nassabatin (according to the previous discussion), in which the word in question functions as an attribute. Here the term waÒf pertains to the syntactic analysis of the phrase in question.14 That Sibawayhi was aware of the polysemy of Òifa can be inferred from the following case: he notes (ibid, I: 344–5) that independent personal pronouns may function as waÒfs, i.e. attributes, but feels the need to draw a distinction between this type of waÒf and waÒfs such as al-†awil (‘tall’), namely, those denoting a description, al-†awil thus representing here the adjectives. Here Sibawayhi clearly seems to distinguish between the two distinct senses of the term waÒf.15 11 See Lane 1863 93 IV: 1318, where an alternative translation of this phrase is provided. 12 See also, for example, Sibawayhi, al Kitab I: 201; II: 352. 13 For such an interpretation, cf. Diem 1974; Owens 1990: 68 89; Talmon 1997b. 14 Cf. Diem 1974: 314. 15 Talmon (1981: 283) regards this as a case of ¨a†f, although Sibawayhi never classifies it as such, as Talmon himself acknowledges.
462
THE TERM ISM IN MEDIEVAL ARABIC GRAMMATICAL TRADITION
But this ambiguity is by no means dichotomous: although there are occurrences of the term Òifa where one can easily see whether it denotes ‘adjective’ or ‘attribute’, al-Kitab abounds with instances where it is ambiguous. Moreover, often no distinction between the two senses seems to be drawn by Sibawayhi at all. For instance, right after he mentions the word turtab (‘a constant thing’) as an example of an ism of the pattern tuf¨al, he notes (ibid, II: 356) that speakers who utter the phrase amrun turtabun (with the same meaning) … ja¨alahu waÒfan. Does he mean here that turtab may function as an attribute? Or perhaps, in light of the fact that the chapter in which this statement appears deals with word patterns and is replete with illustrations of isms alongside Òifas, Òifa thus used in the sense of ‘adjective’, are we to render waÒf, in the above case, also as ‘adjective’ and totally disregard the attributive function of this word in the illustrative phrase? Such examples, I believe, have implications for the cases where the denotation of this term seems to be clear-cut: when Òifa seems, at first sight, to denote ‘adjective’, it nevertheless connotes a syntactic sense of ‘attribute’, and vice versa.16 The term’s ambiguity, in other words, is inherent, which explains Sibawayhi’s recurring practice of illustrating, in the course of purely morphological discussions, cases of Òifa (at the word class level) by using constructions in which the words in question function as adjectival attributes.17 That is, the classification of a word as a Òifa dictates, in general, its syntactic behaviour, so that its possible usage as an attribute suggests, though not necessarily, that it is a Òifa, in the sense of ‘adjective’.18 So far we have seen that Òifa, in one of its senses, denotes ‘adjective’. This does not, however, constitute proof that it is subsumed under the part of speech ism (although in view of the two other parts of speech it is certainly the most plausible candidate). Yet explicit indications that the term Òifa (in the sense of ‘adjective’) does indeed denote a subclass of the part of speech ism are innumerable.19 One such case is the formulation used as the heading of Ch. 92: hadha babu ma yantaÒibu mina ’l-asma’i ’llati laysat bi-Òifatin wa-la maÒadira li-annahu Ìalun … (‘this is the chapter about isms that are neither Òifas nor maÒdars, which are in the accusative since they are circumstantial qualifiers …’ – ibid, I: 165). Interpreting the term Òifa (and inevitably also the term maÒdar!) here as pertaining to the syntactic 16
Cf. Mosel 1975: 144. See, for example, Sibawayhi, al Kitab II: Ch. 512. 18 Cf. Diem 1974: 313 15; Mosel 1975: 144 5. 19 Cf. Mosel 1975: 148. 17
463
THE TERM ISM IN MEDIEVAL ARABIC GRAMMATICAL TRADITION
level would render its very appearance in this heading superfluous, since the isms dealt with in this chapter do not function as attributes but as circumstantial qualifiers. Nonetheless, as mentioned above, al-Kitab abounds with instances where these two terms, ism and Òifa, refer to two mutually exclusive classes.20 Both classes are subsumed under the part of speech ism, the former being more focal, or central, than the latter. That is, the subclass defined by ism can be said to refer to core cases of the part of speech in question, in contradistinction to the one referred to by Òifa. The distinction between the two subclasses, which finds its expression at the morphological and syntactic levels, originates, so it seems, at the semantic level, where words classified as isms are regarded as referring to objects, and those classified as Òifa as referring to qualities.21 In other words, this pair of linguistic terms evokes extra-linguistic situations in the grammarians’ minds, in which objects possess certain traits. This distinction fits the philosophical distinction between substance and attribute. As Rundgren (1976: 143) puts it: ‘Das Adjektivum heisst also bei Sibawayhi Òifatun … ein etwas philosophisch anmutender Terminus.’22 However, no further indication of any relationship between grammar and philosophy concerning this term is found in al-Kitab. A host of distinctions at the morphological level (morphophonemics and morphosyntax included) separate these two classes, of which the following are just a sample: the qalb (mutation)23 of /y/ into /w/ in words of the pattern fa¨la applies only to isms, not to Òifas (Sibawayhi, al-Kitab II: 426);24 indefinite isms of the pattern af ¨al are triptote, whereas Òifas of this pattern are diptote (ibid, I: 5; II: 1); there are isms and Òifas which have identical patterns in the singular but distinct plural forms (e.g. ibid, II: 211). Moreover, some chapters of al-Kitab deal with groups of word patterns sharing common traits; about many such patterns Sibawayhi states whether they are utilized for isms, for Òifas or for both.25 This practice can be illus20 Talmon (1997a: 280) considers the ‘semi independent status’ of the na¨t in Kitab al ¨ayn, inter alia ‘in its frequent contrast to ism’, as an indication that early grammarians may have been inspired by an older division into eight parts of speech. Talmon notes that this ‘is also observable in the Kitab, but in a smaller degree’. 21 Cf. Carter 2004: 106. 22 But cf. Carter (ed.) 1981: 317. For the various viewpoints on the origin of this term, see Versteegh 1977: 48 50; idem 1993: passim; Talmon 1997b; idem 2005: 515. 23 On this process see Bohas, Guillaume and Kouloughli 1990: 83 7. 24 See also Sibawayhi, al Kitab II: 411. 25 See, for example, ibid, II: Chs 512 14, 518 20, 522 3.
464
THE TERM ISM IN MEDIEVAL ARABIC GRAMMATICAL TRADITION
trated with the following excerpt: … fa-yakunu ’l-Ìarfu ¨ala af¨ala wa-yakunu li’l-ismi wa’l-Òifati fa’l-ismu naÌwu afkalin wa-ayda¨in waajdalin wa’l-Òifatu naÌwu abya∂a wa-aswada wa-aÌmara (‘… so that words assume [the pattern] af¨al, in which case they might be either ism or Òifa. Examples for ism are afkal [= a tremor], ayda¨ [= saffron] and ajdal [= a hawk]; examples for Òifa are abya∂, aswad and aÌmar [= colour names]’ – ibid, II: 344). That ism in all such cases denotes a subclass of the part of speech ism and not the entire part of speech denoted by the term ism, is inferred inter alia from formulations such as the following, also in a chapter dealing with word patterns: … fa-yakunu ’l-ismu ¨ala mithali fa¨awlalin fi ’l-ismi wa’l-Òifati (‘… so that isms assume the pattern fa¨awlal, as either ism or Òifa’ – ibid, II: 367). The only possible interpretation here is that a class referred to by the term ism (first token) subsumes two subclasses, referred to by the terms ism (second token) and Òifa. This interpretation is in line with Lyons’s remark quoted above: at the word class level the term ism can convey either a general sense, referring to the whole part of speech ism, or a more specific sense, corresponding to the term ‘noun’; in the latter sense it thus refers to a subclass of the very part of speech this term otherwise denotes. Thus, broadly speaking, it is only when ism stands in opposition to Òifa that it can be said to (approximately) correspond to the Western term ‘noun’. Syntactic consequences of this distinction between ism and Òifa as two subclasses of the part of speech ism are mentioned in al-Kitab, according to which certain functional positions are deemed to belong exclusively to one or the other. For example, Sibawayhi (ibid, I: 184) judges the sentence marartu bi-rajulin asadin shiddatan wajur’atan (‘I passed by a man who was a lion in strength and boldness’) to be ∂a¨if qabiÌ (‘weak and ugly’), since asad is ism lam yuj¨al Òifatan.26 Yet Sibawayhi (ibid, I: 265) clearly states that demonstrative pronouns are followed by isms functioning as attributes, or in his own words: tuÒafu bi’l-asma’i, asma’ here denoting a subclass of the part of speech ism which does not include adjectives; interpreting this term as referring to the entire part of speech ism would render such a description of the demonstrative pronouns superfluous.27 In other words, these definite isms, i.e. nouns, are classified by Sibawayhi as Òifas. But the picture turns out to be rather more complicated, since not only nouns (isms) may function as Òifa of these pro26 27
See also ibid, I: 167, 195, 235, 324. See also ibid, I: 59, 221. 465
THE TERM ISM IN MEDIEVAL ARABIC GRAMMATICAL TRADITION
nouns, but also adjectives (Òifas). Such Òifas acquire the manzila, i.e. status, of isms, so that they are not in the manzila of Òifas such as the one appearing in the sentence marartu bi-zaydini ’l-†awili (‘I passed by Zayd the tall’), viz. they are not adjectival attributes (ibid, I: 189). To sum up, these are Òifas (adjectives) with the status of an ism (noun) functioning as a Òifa (attribute)! Similarly, Sibawayhi condemns constructions in which Òifas assume functions which properly belong only to isms.28 On the other hand, he points out (ibid, II: 4) that some Òifas may function syntactically as isms, and therefore would not be in need of heads; he explains this as due to frequency of use.29 The following discussion (ibid, I: 96) embraces both cases: Sibawayhi, on the one hand, rejects constructions in which adjectives function as subjects of verbs in the passive voice — a function regarded as belonging exclusively to ism — e.g. in the sentence sira ¨alayhi shadidun (‘it was travelled on it [= the way] vehemently’); but he regards such usage as ‘better’ when the adjectives in question are followed by prepositional phrases functioning as their attributes, e.g. in the sentence sira ¨alayhi shadidun mina ’l-sayri (same meaning). Had the term ism in such cases referred to the whole part of speech ism such statements would become self-contradictory, since Òifa is a subclass of this part of speech. As mentioned above, alongside its denotation as ‘adjective’, Òifa also conveys the sense of ‘attribute’ at the syntactic level. It is worthy of note that in order to refer to the heads of such Òifas, Sibawayhi chooses not only the term mawÒuf (or man¨ut, when the attribute is dubbed na¨t) but also the term ism.30 3. Ism vs. Òifa in Later Treatises The picture portrayed in Section 2 applies, by and large, to later works as well. I will try to establish here that the innovations introduced by later grammarians boil down to rendering explicit what is implicit in al-Kitab. 28
See, for example, ibid, I: 96, 114, 122, 158; II: 179, 181. Such cases raise the question of which of the two levels, morphology or syn tax, prevails when they happen to clash, that is, when a certain word is allowed to assume functions peculiar to an ism despite possessing morphological characteristics of a Òifa. This interesting question will not be elaborated on here, since it has no bearing on the issues under consideration. 30 See, for example, ibid, I: 34, 153, 182, 183, 206, 271, 307, 314. 29
466
THE TERM ISM IN MEDIEVAL ARABIC GRAMMATICAL TRADITION
As we saw above, the term Òifa is used by Sibawayhi in two distinct senses, viz. ‘adjective’ and ‘attribute’, albeit with no terminological implications, although we did see a case which seemed to show that Sibawayhi was aware of this polysemy. Later grammarians occasionally feel the need to distinguish explicitly between Òifa in the sense of ‘attribute’ and Òifa in the sense of ‘adjectival attribute’; this they do through the use of special appellations for Òifa in the latter sense. Ibn Jinni (d. 1002) (Sirr: 357), for instance, refers to adjectival attributes as al-Òifa al-ÒariÌa, i.e. ‘pure’ Òifa, in contrast with definite nouns following demonstrative pronouns. He criticizes the construal of the latter as Òifas (which, as we have seen, is Sibawayhi’s position, adopted also by other grammarians, although Ibn Jinni gives no names),31 and maintains that one may apply this term to such constituents only majazan, but not Ìaqiqatan, viz. only figuratively.32 Al-Ba†alyusi (al-Îulal: 76) distinguishes between Òifa ma¨nawiyya and Òifa lafÂiyya, subsuming asyndetic relative clauses under the former, but not under the latter. The latter thus applies only to adjectival attributes, lafÂi apparently corresponding here to ‘morphological’, ma¨nawi to ‘(syntactico-)semantic’. Similarly, albeit in a more elaborate manner, al-Astarabadhi (d. 1287) (SharÌ II: 283–7), following Ibn al-Îajib (d. 1248), distinguishes between waÒf ¨amm (i.e. a ‘general’ waÒf ) and waÒf khaÒÒ (i.e. a ‘specific’ waÒf). The former pertains to the semantic level alone, thus roughly corresponding to ‘adjective’ (as a subclass of the part of speech ism), to the exclusion of, inter alia, asyndetic relative clauses and nouns functioning as Òifa. Such waÒfs, he asserts, can function not only as attributes, but also as circumstantial qualifiers (Ìal) and predicates. The latter, on the other hand, function only as attributes, though as a class they do not correspond to ‘attributes’: only isms (ism here in the sense of the whole part of speech) are included in this class, to the exclusion of, inter alia, clauses. In passing we may point out as well that the term waÒf ¨amm does not encompass what is included in the term waÒf khaÒÒ, since several classes are subsumed under the latter but not under the former, e.g. infinitives and nouns functioning as attributes.33 31 Ibn Jinni is not the only grammarian to oppose this classification. See also al Ba†alyusi (d. 1127), al Îulal: 106; Ibn Ya¨ish (d. 1245), SharÌ I: 130; Ibn Hisham (d. 1360), Mughni: 630 1. Cf. Owens 1990: 93 4. 32 On the Ìaqiqa majaz contrast, see Heinrichs 1984; Versteegh 1990: 282, 285 6. 33 Al Astarabadhi’s definitions of these two terms, which differ from those of Ibn al Îajib, render these appellations, i.e. ‘general’ and ‘specific’, inappropriate for
467
THE TERM ISM IN MEDIEVAL ARABIC GRAMMATICAL TRADITION
Such cases imply that the grammarians were fully aware of the problematic nature of polysemy in their nomenculture.34 It was noted above that the distinction between Òifa and ism parallels a philosophical distinction between substance and attribute; yet the only terminological indication for this in al-Kitab is the mere fact that Sibawayhi uses the term Òifa. That is, al-Kitab contains no philosophical terms which correspond to the other part of this opposition. His followers, on the other hand, do use such philosophical terms, namely, jawhar35 and dhat.36 Al-Mubarrad (d. 898) (al-Muqta∂ab IV: 220), for instance, states that the normal usage of words such as al†awil, in constructions of demonstrative pronouns, is as ‘concordants’ (tabi¨ [sg.]),37 i.e. attributes, of the jawhars following these pronouns, that is, they do not immediately follow the pronouns. It might be the case that al-Mubarrad here deliberately abstains from using the term ism, due to the problematic construction with which he deals (see above).38 From this treatise one can also adduce an illustration for the Òifa-dhat contrast: al-Mubarrad (ibid, I: 179) says, regarding the interrogative ma (‘what’): wa-hiya su’alun ¨an dhati ghayri ’l-adamiyyina wa¨an Òifati ’l-adamiyyina (‘it is designed to convey a question regarding either the substance of non-humans or the attributes of humans’).39 The only grammarian found to demonstrate some awareness of the problematic nature of formulations which imply that ism and Òifa refer to two mutually exclusive classes (a formulation which frequently occur in grammatical treatises) is, unsurprisingly, al-Astarabadhi. He interprets (SharÌ III: 372) such a usage of the term ism, on the part of Ibn al-Îajib, by stating that it means ghayr Òifa. these two classes, since the latter is not subsumed under the former but rather over laps it. Ibn al Îajib, on the other hand, defines the ‘general’ type only semantically, and the ‘specific’ type both semantically and syntactically. For him, therefore, the latter is included in the former. 34 See also, for example, Ibn al Sarraj (d. 928), al UÒul II: 27ff.; Ibn Jinni, al KhaÒa’iÒ II: 207; III: 193, 262 3; idem, Sirr: 361; Ibn Hisham, SharÌ: 141. 35 Cf. Van den Bergh 1965. 36 Cf. Rahman 1965. 37 According to Owens (1990: 97), al Mubarrad uses the verb tabi¨a in this sense ¨fairly consistently’. 38 In fact, Sibawayhi (al Kitab I: 235) uses the word jawhar in order to indicate that the word it refers to should not function as an attribute. Yet jawhar here might be interpreted, based on the context, as possessing its everyday meaning, i.e. ‘mate rial’. See also ibid, II: 236; al Mubarrad, al Muqta∂ab III: 258 60; Ibn Ya¨ish, SharÌ II: 75. But cf. Talmon 1997b; Carter 2004: 142. 39 See also al Mubarrad, al Muqta∂ab II: 51, 290; III: 339; idem, al Kamil I: 51; al Jurjani (d. 1078), al MuqtaÒid: 152 3; Ibn Ya¨ish, SharÌ III: 145; IV: 23; al Astarabadhi, SharÌ I: 179, 376 7; IV: 337; Ibn Hisham, Mughni: 196. 468
THE TERM ISM IN MEDIEVAL ARABIC GRAMMATICAL TRADITION
The most significant innovation made by Arabic grammarians with respect to the issues at stake here was their introduction of special appellations and formulations referring to the focal subclass of the part of speech ism, otherwise denoted by the term ism on its own (as we have seen above). Two examples suffice to illustrate this: Ibn Jinni’s (Sirr: 367) explanation of the double usage of the word sha¨ub (‘death’), i.e. with and without the definite article, is based on its aÒl (here in the sense of an original and basic trait of the entity in question)40 as a Òifa, in view of its pattern and etymology. Thus when it has a definite article it resembles proper nouns derived from adjectives, such as al-¨Abbas. On the other hand, those who use this word without the article regard it as an ism ÒariÌ (i.e. a ‘pure’ ism), not a Òifa, even though it retains, semantically, some of the latter’s colouring. Those who use the word without the article thus resemble those who use names such as ¨Abbas without the article. Our second example is taken from al-Jurjani. He states (al-MuqtaÒid: 258) that one of the types of khabar al-mubtada’ consists of ism maÌ∂ ghayr Òifa or just ism maÌ∂ (also indicating ‘pure’ ism). Judging from his illustrations of this type of predicate, e.g. rajul (‘man’) and akhuka (‘your brother’), these appellations correspond to ‘noun’.41 4. Ism vs. maÒdar and Âarf I have confined myself hitherto to the relationship between ism and Òifa. A similar relationship exists also between the members of the pairs ism-maÒdar and (though to a somewhat lesser extant) ism-Âarf, which I discussed more fully elsewhere.42 My findings there are summarized in outline in the following: As in the case of Òifa, the term maÒdar also possesses two meanings, at two distinct linguistic levels: on the one hand it corresponds to ‘infinitive’, at the word class level, as a subclass of the part of speech ism; on the other hand, it corresponds to ‘inner object’, i.e. al-maf¨ul al-mu†laq, at the syntactic level.43 As with Òifa, no clear-cut 40
On this meaning of aÒl see Baalbaki 1995: 9ff. On the relationship between aÒl and linguistic change in al Kitab, see idem 1988: 169 71. On this term as a component of ta¨lil, see Suleiman 1999. 41 See also Ibn Jinni, Sirr: 285; Ibn al Anbari (d. c. 1182), al InÒaf: 55 6; idem, Asrar: 72; Ibn Ya¨ish, SharÌ I: 88; III: 71; al Astarabadhi, SharÌ III: 461. 42 Kasher 2006: 92ff. 43 On the term al maf¨ul al mu†laq see Levin 1991. For a criticism on the part of a relatively late grammarian of the usage of this term at the syntactic level, see Carter (ed.) 1981: 344. 469
THE TERM ISM IN MEDIEVAL ARABIC GRAMMATICAL TRADITION
dichotomy exists between the two senses. Grammarians, albeit very rarely, use special appellations, such as maÒdar ÒaÌiÌ (i.e. ‘sound’ maÒdar), in order to distinguish between the use of the term maÒdar to denote an infinitive functioning as an inner object and its use to denote just an inner object, not necessarily implemented by an infinitive. As with Òifa, there are innumerable cases where the term maÒdar stands opposed to the term ism, or sometimes to special appellations such as al-ism al-ÒaÌiÌ, the latter referring to a subclass of the part of speech ism, regarded as more focal than the maÒdar’s subclass. From grammarians’ writings we can infer that at the semantic level words classified as isms, in the specific sense of this term, refer to concrete objects, whereas maÒdars refer to actions. As with Òifa, the grammarians mention morphological differences between these two subclasses; thus, for example, the word pattern if¨al is considered to be almost unique to maÒdars. At the syntactic level, the contrast maÒdar-ism often corresponds to the contrast between the inner object and other functions isms may assume. This relationship between ism and maÒdar is thus isomorphic to the relationship between ism and Òifa.44 The relationship between ism and Âarf is rather more complex, as the following (very concise) discussion shows. A fundamental trait of the term Âarf sets it apart from the terms Òifa and maÒdar as regards the issues at stake here: as we have seen above, Òifa and maÒdar can each refer either to a word class, irrespective of the function the word in question assumes, or to a syntactic function, which may be assumed by various types of words, phrases and clauses. This is not the case with Âarf: a word classified as a Âarf of necessity implements the syntactic function labelled Âarf. In other words, if a word assumes any function other than the one termed Âarf, such as a subject or a complement of a preposition, it is no longer classified as a Âarf.45 44 In Section 1 I pointed out that the exact extension of the term ism in its spe cific sense depends on the other subclass of the part of speech ism in opposition to which it stands. This can be illustrated with the triad ism Òifa maÒdar: in addition to numerous cases where Sibawayhi distinguishes between maÒdar and ism (e.g. the heading of Ch. 68 al Kitab I: 138), he nevertheless regards maÒdar as a subclass of ism, not only when the latter refers to the whole part of speech, but also when it is contrasted with Òifa (e.g. ibid, II: 353). That is, in both cases the term ism refers to a subclass of the part of speech ism, but each time to a rather different subclass: in the former this subclass excludes maÒdars, in the later Òifas. 45 Strictly speaking, the last two sentences cannot be applied to Sibawayhi, un less we replace the phrase ‘syntactic function’ with ‘syntactic state’: according to Sibawayhi’s theory of the Âarf, classifying a constituent as a Âarf does not preclude it from assuming syntactic functions other than the adverbial, mainly that of predi
470
THE TERM ISM IN MEDIEVAL ARABIC GRAMMATICAL TRADITION
Nevertheless, the term Âarf should not be taken to designate exclusively a syntactic function assumed by a semantically defined subclass of ism, since the grammarians often treat the term Âarf as designating a subclass of the part of speech ism.46 As with Òifa and maÒdar, the grammarians differentiate between the subclass Âarf and a more focal subclass of the part of speech ism, here also dubbed ism, or, alternatively, al-ism al-ÒaÌiÌ etc. This distinction originates from an important semantic trait of the Âarf: although words classified as Âarfs convey a spatial or temporal meaning,47 that is, they refer to objects as other isms do; the place or time is referred to only with respect to a given object or action. For instance, al-Mubarrad (al-Muqta∂ab III: 102) distinguishes between the sentences zaydun khalfaka (‘Zayd is behind you’) and alladhi zaydun fihi khalfuka (‘what Zayd is in is the area behind you’, literally: ‘your rear’). In both sentences the word khalf possesses a spatial meaning, yet it is only in the former that the place in question is viewed in terms of an object located in it, here a person named Zayd (referred to by the constituent zayd ). Al-Mubarrad states explicitly: wa-innama yakunu Âarfan idha ta∂ammana shay’an (‘it is a Âarf only when it embraces something’). In contrast, in the latter sentence the word khalf(uka) is classified as an ism, not as a Âarf. It should be noted that this sentence is the outcome of a syntactic transformation labelled ikhbar, elucidated by Goldenberg (1988: 67) as ‘the turning of an extracted part of a sentence into a predicate’. This transformation thus produces a construction, the refcate. Therefore, the second constituent in sentences such as zaydun khalfaka (‘Zayd is behind you’) is, for Sibawayhi, both Âarf and predicate. However, all Âarfs, irre spective of their syntactic function, are subsumed under what might be regarded as a syntactic ‘state’ of the Âarf (Cf. Carter 1972; Owens 1990: 111 15). For instance, one of the components of this ‘state’ is being in the accusative, as in the sentence above. However, if such a word is assigned the nominative, still functioning as a predicate, it is no longer to be classified as a Âarf (for a sentence, albeit seemingly artificial, with the word khalfuka used as predicate, see in what follows). On the other hand, later grammarians posit, for Âarfs which assume functions other than the adverbial, underlying verbal constituents, for which the Âarfs function as adverbials (Cf. Peled 2007: 162ff.), thereby initiating a reduction of these (surface) constructions to those in which the Âarfs function as adverbials. Here we can prop erly regard the term Âarf as denoting a syntactic function. 46 Many scholars have conceived the term Âarf as referring merely to a subclass of the part of speech ism (see, for example, Diem 1974: 325; Bohas, Guillaume and Kouloughli 1990: 52; Owens 1990: 141 51). Although this might be the case in some grammatical writings, most prepositional phrases are also classified as Âarfs. This point, which I discussed elsewhere (Kasher 2006: Part 1) will not be dealt with here, since it is not relevant to the main thrust of this paper. 47 In passing, not every such word is entitled to be a Âarf, e.g. bayt (‘house’). 471
THE TERM ISM IN MEDIEVAL ARABIC GRAMMATICAL TRADITION
erent of whose predicate, which is also the referent of the new subject, is conceived as an entity in itself;48 hence the inability of such predicates to be assumed by Âarfs. 5. Conclusions Our findings are in line with the model formulated by Bohas, Guillaume and Kouloughli (1990: 51–3), according to which the part of speech ism comprises a ‘hard core’ and two ‘margins’; one of these ‘margins’ contains the active and passive participles and some classes of adjectives, which can govern the nominative and the accusative and whose syntactic use is more restricted than the hard core’s. That is, the part of speech ism includes a focal subclass (the ‘hard core’, i.e. the noun) and several more ‘marginal’ subclasses, wherein one can find the Òifa, in the sense of adjective, as well as the maÒdar, in the sense of infinitive, and the Âarf. This focal subclass is referred to by the term ism or, alternatively, by special appellations such as ism maÌ∂, though these should by no means be regarded as fixed terms. This interpretation resolves, I believe, the apparent inconsistencies in grammatical writings, which have been presented above. It should come as no surprise if it turns out that the phenomenon portrayed here, of a term which is a hyponym of itself, is much more widespread than described in this paper,49 in view of its being a natural semantic characteristic of language, as was established by Lyons. REFERENCES Primary sources Al Astarabadhi, SharÌ = [Ra∂i al Din MuÌammad ibn al Îasan al Astarabadhi], SharÌ al Ra∂i ¨ala al Kafiya, ed. Yusuf Îasan ¨Umar. (Benghazi 1978) Al Ba†alyusi, al Îulal = Abu MuÌammad ¨Abd Allah ibn MuÌammad ibn al Sid al Ba†alyusi, Kitab al Îulal fi iÒlaÌ al khalal min Kitab al Jumal, ed. Sa¨id ¨Abd al Karim Sa¨¨udi. (Baghdad 1980) Ibn al Anbari, Asrar = Abu al Barakat ¨Abd al RaÌman ibn MuÌammad ibn Abi Sa¨id al Anbari, Kitab Asrar al ¨arabiyya, ed. MuÌammad Bahjat al Bay†ar. (Damascus 1957) Ibn al Anbari, al InÒaf = Kamal al Din Abu al Barakat ¨Abd al RaÌman ibn MuÌammad ibn Abi Sa¨id al Anbari, al InÒaf fi masa’il al khilaf bayna al naÌwiyyina al baÒriyyina wa’l kufiyyina, ed. MuÌammad MuÌyi al Din ¨Abd al Îamid, 4th edn. (Egypt 1961) 48 49
Cf. Lyons 1977: 693. Cf. Owens 1990: 62 83. 472
THE TERM ISM IN MEDIEVAL ARABIC GRAMMATICAL TRADITION
Ibn Hisham, Mughni = Jamal al Din ibn Hisham al AnÒari, Mughni al labib ¨an kutub al a¨arib, eds Mazin al Mubarak and MuÌammad ¨Ali Îamd Allah, 2nd edn. (n.p. n.d.) Ibn Hisham, SharÌ = Abu MuÌammad ¨Abd Allah Jamal al Din ibn Yusuf ibn AÌmad ibn ¨Abd Allah ibn Hisham al AnÒari, SharÌ Shudhur al dhahab fi ma¨rifat kalam al ¨arab, ed. MuÌammad MuÌyi al Din ¨Abd al Îamid, 10th edn. (Cairo 1965) Ibn Jinni, al KhaÒa’iÒ = Abu al FatÌ ¨Uthman ibn Jinni, al KhaÒa’iÒ, ed. MuÌammad ¨Ali al Najjar, 3rd edn. (Cairo 1986 88) Ibn Jinni, Sirr = Abu al FatÌ ¨Uthman ibn Jinni, Sirr Òina¨at al i¨rab, ed. Îasan Hindawi. (Damascus 1985) Ibn al Sarraj, al UÒul = Abu Bakr MuÌammad ibn Sahl ibn al Sarraj, al UÒul fi al naÌw, ed. ¨Abd al Îusayn al Fatli, 3rd edn. (Beirut 1988) Ibn Ya¨ish, SharÌ = Muwaffaq al Din Ya¨ish ibn ¨Ali ibn Ya¨ish, SharÌ al MufaÒÒal. (Egypt n.d.) Al Jurjani, al MuqtaÒid = ¨Abd al Qahir al Jurjani, Kitab al MuqtaÒid fi sharÌ al I∂aÌ, ed. KaÂim BaÌr al Marjan. (Baghdad 1982) Al Mubarrad, al Kamil = Abu al ¨Abbas MuÌammad ibn Yazid al ma¨ruf bi’l Mubarrad, al Kamil fi al lugha wa’l adab. (Beirut 1987) Al Mubarrad, al Muqta∂ab = Abu al ¨Abbas MuÌammad ibn Yazid al Mubarrad, al Muqta∂ab, ed. MuÌammad ¨Abd al Khaliq ¨U∂ayma, 2nd edn. (Cairo AH 1386 99) Sibawayhi, al Kitab = Sibawayhi, al Kitab, ed. Hartwig Derenbourg. (Paris 1881 9, Repr. Hildesheim 1970) Secondary sources Baalbaki, R. 1988. ‘A Contribution to the Study of Technical Terms in Early Arabic Grammar: The Term aÒl in Sibawaihi’s Kitab’, in A.K. Irvine, R.B. Serjeant and G. Rex Smith (eds), A Miscellany of Middle Eastern Articles in Memoriam Thomas Muir Johnstone, 1924 83. (Harlow). 163 77 1995. ‘Reclassification in Arab Grammatical Theory’, Journal of Near Eastern Studies 54, 1 13 Bohas, G., J. P. Guillaume and D.E. Kouloughli. 1990. The Arabic Linguistic Tradi tion. (London) Carter, M.G. 1972. ‘“Twenty Dirhams” in the Kitab of Sibawaihi’, Bulletin of the School of Oriental and African Studies 25, 485 96 (ed.). 1981. Arab Linguistics: An Introductory Classical Text with Translation and Notes. (Amsterdam) 2004. Sibawayhi. (London) Diem, W. 1974. ‘Nomen, Substantiv und Adjektiv bei den arabischen Gramma tikern’, Oriens 23 4, 312 32 Goldenberg, G. 1988. ‘Subject and Predicate in Arab Grammatical Tradition’, Zeitschrift der Deutschen Morgenländischen Gesellschaft 138, 39 73 Heinrichs, W. 1984. ‘On the Genesis of the Ìaqîqa majâz Dichotomy’, Studia Islamica 59, 111 40 Kasher, A. 2006. ‘The Âarf in Medieval Arabic Grammatical Theory’ (Hebrew). Ph.D. dissertation, Bar Ilan University Lane, E.W. 1863 93. Arabic English Lexicon. (London) Levin, A. 1991. ‘What is Meant by al maf¨ul al mu†laq?’, in A.S. Kaye (ed.), Semitic Studies: In Honor of Wolf Leslau on the Occasion of his Eighty fifth Birthday, November 14th, 1991. (Wiesbaden), II, 917 26 473
THE TERM ISM IN MEDIEVAL ARABIC GRAMMATICAL TRADITION
Lyons, J. 1977. Semantics. (Cambridge) Mosel, U. 1975. ‘Die Syntaktische Terminologie bei Sibawaih’. Ph.D. dissertation, University of Munich Owens, J. 1990. Early Arabic Grammatical Theory: Heterogeneity and standardiza tion. (Amsterdam) Peled, Y. 1999. ‘Aspects of the Use of Grammatical Terminology in Medieval Arabic Grammatical Tradition’, in Y. Suleiman (ed.), Arabic Grammar and Linguistics. (Richmond). 50 85 2007. ‘Problems in the Medieval Arabic Theory of Sentence Types’, in E. Ditters and H. Motzki (eds), Approaches to Arabic Linguistics: Presented to Kees Versteegh on the Occasion of his Sixtieth Birthday. (Leiden). 149 88 Rahman, F. 1965. ‘!at’, EI2, II, 220 Rundgren, F. 1976. ‘Über den griechischen Einfluss auf die arabische National grammatik’, Acta Societatis Linguisticae Upsaliensis 2, 119 44 Suleiman, M.Y. [=Y.] 1990. ‘Sibawayhi’s “Parts of Speech” According to Zajjaji: A New Interpretation’, Journal of Semitic Studies 35, 245 63 1999. The Arabic Grammatical Tradition: A Study in ta¨lil. (Edinburgh) Talmon, R. 1981. ‘Appositival ¨a†f: An Inquiry into the History of a Syntactic Category’, Arabica 28, 278 92 1997a. Arabic Grammar in its Formative Age: Kitab al ¨Ayn and its attribution to Îalil b. AÌmad. (Leiden) 1997b. ‘∑ifa: 1. In grammar’, EI2, IX, 551 2005. ‘Gaya, Òifa and al kalam al wasif in Ibn al Muqaffa¨’s Manual of Logic: New considerations about the beginning of Arabic grammar’, Jerusalem Studies in Arabic and Islam 30, 506 20 Van den Bergh, S. 1965. ‘$awhar’, EI2, II, 493 4 Versteegh, C.H.M. [=K.] 1977. Greek Elements in Arabic Linguistic Thinking. (Leiden) 1990. ‘Freedom of the Speaker? The Term ittisa¨ and related Notions in Arabic Grammar’, in K. Versteegh and M.G. Carter (eds), Studies in the History of Arabic Grammar II: Proceedings of the 2nd Symposium on the History of Arabic grammar, Nijmegen, 27 April 1 May 1987. (Amsterdam). 281 93 1993. Arabic Grammar and Qur’anic Exegesis in Early Islam. (Leiden) 1995. The Explanation of Linguistic Causes: Az Zaggagi’s Theory of Grammar. (Amsterdam)
474
THEMES OF THE IBA∆I/OMANI SIYAR ABDULRAHMAN S. AL SALIMI CHIEF EDITOR OF AL-TASAMOH JOURNAL
Abstract Most of the previous investigations into Omani history or Iba∂i literature have included siyar among the early sources without provid ing a complete and independent picture of them. Little or no atten tion has been paid to what these siyar represent. This study is there fore an attempt to fill some of these gaps in the scholarship on Omani siyar. The aims of this study of the siyar are: 1) to determine their basis and themes; 2) to discover which Iba∂i areas are more re ceptive to the introduction of the siyar; and 3) to determine how geo graphical distribution may have affected the structure and content of the siyar. Likewise, it is the aim of this article to show what the essen tial strategies of producing the siyar were, and to find out how these changed over time. In other words, this study explores the stages of the writing of the Omani siyar and seeks to explain how and why these stages were developed.
This paper aims at providing a broad overview of the themes of the Iba∂i/Omani siyar in political- religious correspondence. This should provide a foundation for further research on the topic, given that it is impossible to do justice to the huge corpus of documentary texts in a limited work such as this. Therefore, it explores the significance of producing the siyar, making an attempt to deal with each separate objective — e.g. theology, Omani socio-politics — as an independent theme by cross-analyzing the texts examined. In this context, it briefly considers the siyar as a whole or supplemental with other works i.e. Omani, Iba∂i and contemporary Eastern literature as a whole, using the latter mainly as sources of materials to complete the themes of the former. So far, Wilkinson has been the only scholar to produce a theory on this subject, although he includes the siyar among the Iba∂i fiqh works.1 Apart from that, Wilkinson has provided the background to the themes of siyar and made useful insights 1 J.C. Wilkinson, ‘The Omani Manuscript Collection at Muscat: Part II’, Ara bian Studies 6 (1978), 199 207.
475
THEMES OF THE IBA∆I/OMANI SIYAR
into the constitution of the Iba∂i community, its political and socioeconomic history, and the disputes within it over dogma.2 In what follows, a detailed study of theological aspects of the siyar is presented while allowing discussion of the dogmatic dispute between the Nizwa and Rustaq schools to shed light on the socio-political history of Oman. 1. Theological Aspects of the Siyar The siyar reveal the interesting conventions of early theological discussions among the Iba∂i imams in Basra. In order to understand the theological position of the Iba∂is in the siyar, it is necessary to analyse the compilations that have produced their doctrinal views from two angles: a comparative one in which Iba∂i theological views on Islamic doctrine are given, and the epistles substantiating arguments of the Iba∂i scholars. The feature that stands out of earlier theological compilations appears to be the growth of divergent separate dogmatic views within Islamic society with their accompanying literature. The arbitration settlement following the battle of ∑iffin between ¨Ali and Mu¨awiya had displeased all of the groups involved, especially Mu¨awiya, and remained a focal point of socio-political and theological attention for decades to come. The factors that provoked the battle of ∑iffin had existed before the murder of the caliph ¨Uthman b. ¨Affan (VII).3 According to Abu al-Mu’thir’s sira (LXI), people were divided into three parties: the ¨Uthmanis (the supporters of ¨Uthman), the anti¨Uthmanis, and those who were called the Shukkak, who neither supported nor opposed ¨Uthman.4 The development of separate doctrines in Islam happened after the ∑iffin arbitration, when ¨Ali’s supporters split into two groups: the Shi¨a and the Khawarij. Abu alMu’thir argues that, although the various Islamic sects seem different, they still had many beliefs in common.5 The siyar have preserved the Iba∂i arguments on the ∑iffin affair, but have also presented the op2
Ibid., 199 207. A.J. Wensinck, The Muslim Creed (Cambridge 1936), 36. On the ∑iffin arbi tration, see Martin Hinds, ‘The ∑iffin Arbitration Agreement’, JSS 17 (1972), 93 129. 4 For this development, a similar traditional view can be found with ¨Abd Allah b. MuÌammad al Nashi’ al Akbar; see J. van Ess, Frühe Mu¨tazilitische Häresio graphie (Beirut 1971), 16, 17, 19. 5 Al Siyar wa al jawabat, edited by Sayyida Kashif Isma¨il. (Muscat 1984), 2:309. 3
476
THEMES OF THE IBA∆I/OMANI SIYAR
position views. The siyar records of ∑iffin have been reported in alQalhati’s (seventh/thirteenth century) book, al-Kashf wa-al-bayan.6 These arguments are also given in the sira of Salim b. Dhakwan (XXV), and in ¨Abd Allah b. Iba∂i’s letter to ¨Abd al-Malik b. Marwan (XIV). Both the disputes concerning ¨Uthman’s actions (aÌdath) and the arbitration between ¨Ali and Mu¨awiya have been utilized as examples in the discussion of the concept of sin.7 In the epistles of Abu ¨Ubayda and Abu Mawdud, interesting perspectives are found on the connection between sin and actions. Abu ¨Ubayda and Abu Mawdud (XVIII) both hold that self-will is a strong motivation that pushes man to sin; this is the doctrine of deliberate sin.8 Watt, while tracing the beginning of the doctrine of predestination in Islam, found that the Khariji outlook was seen to be fatalistic and predestinarian. As Watt asserts, the development of Khariji doctrine on predestination is shown through the conception of the ‘God who demands righteousness from His creatures’. This conception led logically to the doctrine of human responsibility, with its corollary doctrine of predestination, wherein man has the power to perform the duties imposed on him.9 But the early Iba∂is themselves experienced a schism among themselves after Harith al-Iba∂i affirmed the doctrine of determination (al-qawl bi-al-qadar).10 Nevertheless, some Iba∂is agreed with Abu al-Hudhayl the Mu¨tazili, while the rest of the Iba∂is considered his view to be that ‘capacity precedes the act’.11 This subject was also considered by alÎajjaj b. Yusuf, the Umayyad governor of Iraq, who asked Jabir b. 6 See M. Kfafi, ‘The Rise of Kharijism according to Abu Sa¨id MuÌammad Sa¨id al Azdi al Qalhati’, Bulletin of the Faculty of Arts (Cairo) 14 (1952), 29 48. 7 See the first chapter of Kitab fihi aÌdath ¨Uthman, sirah VII; Abu al Qasim al Barradi, Risala fi taqyid kutub aÒÌabina: Dirasa fi Tarikh al Iba∂iyya, edited by M. ¨Azab and A. ¨Uwa∂, (Cairo 1994. This work is based on MS, Dar al Kutub al MiÒriya number 21791 B. Cairo) Abu al Qasim al Barradi, Al Jawahir al mun taqat fi ma akhalla bihi kitab al ™abaqat (Cairo 1302/1885), 219; Martin Hinds, ‘The Murder of the Caliph ¨Uthman’, International Journal of Middle East Studies 3 (1972): 450 69; Marin Hinds, ‘Kûfan Political Alignments and their Back ground in the Mid Seventh Century A.D.’, International Journal of Middle East Studies 2 (1971), 346 67. 8 Wensinck, Muslim Creed, 42. 9 W. Montgomery Watt, Free Will and Predestination in Early Islam (London 1948), 46. 10 J. Van Ess, Theologie und Gesellschaft im 2. und 3. Jahrhundert Hidschra: eine Geschichte des religiösen Denkens im frühen Islam (Berlin and New York 1992 7), 2:204. 11 Abu al Îasan {Ali al Ash¨ari, Maqalat al islamiyin, edited by MuÌyi al Din ¨Abd al Îamid. (Dar al Îikma, Damascus, 1415/1994), 20.
477
THEMES OF THE IBA∆I/OMANI SIYAR
Zayd to clarify this topic. Jabir’s reply referred to the Quranic verse: ‘Whom God doth guide there can be none to lead astray, and for whom God rejects from His guidance, there can be no guide.’12 Al-Ash¨ari (d. 324/935–6) describes the early Iba∂i parties in terms of the doctrine of qadar,13 and shows that the Iba∂i materials indicate early contact between the Iba∂is and the Qadariya. Îamza al-Kufi, according to Abu ¨Ubayda, shows that there was an arrangement with Ghaylan al-Dimashqi.14 This arrangement between the Iba∂is and the Qadariya perhaps occurred at the end of the first/ beginning of the eighth century, when al-Îarith al-Iba∂i founded his doctrine outside the Iba∂i traditional school in an attempt to link his doctrine with ‘Abd Allah b. Iba∂.15 As van Ess suggests, the Iba∂is were not opposed to the Qadariya, but determinism won out under Abu ¨Ubayda during the reign of al-ManÒur (r. 136–58/754–75).16 Furthermore, Crone and Zimmermann report Ibn Îazm’s claim that Ibn Iba∂ converted to Mu¨tazilism, suggesting that he may have been remembered as an adherent of a Qadariya system.17 It is important to note that the origin of the split among Abu ¨Ubayda’s pupils was the issue of qadar, and those whose views resembled qadari views were Shu¨ayb b. al-Ma¨ruf, ¨Abd Allah b. Yazid, and Harun b. alYaman.18 At the beginning of the third/ninth century, there was an Iba∂i debate between Harun b. al-Yaman and MaÌbub b. al-RaÌil; according to al-Salimi, Harun was among the Shu¨aybiya19 with whom the early Khawarij discussed the issue of qadar.20 At this point, it is worth examining the definition of Shu¨aybiya, since it appears in several Iba∂i works. There is confusion in the heresiographies between the Shu¨aybiya and the Sha¨biya, and many scholars misunderstand the differences between the stances of these two sub-sects. It is sug12
Qur’an, al Kahf 18:17. Al Ash¨ari, Maqalat, 20. Useful views in this case are given by van Ess, TG, 2:202, 268. 14 Van Ess, Anfänge Muslimischer Theologie (Beirut 1977), 230; EI2 s.v. ‘Ghaylan b. Muslim al DimashÈi’. 15 Van Ess assumes that al Îarith was alive in ca. 800, and hence was a contem porary of Abu Sufyan MaÌbub b. al RaÌil. TG, 2:212; Crone, Patricia and Zimmermann, Fritz), The Epistle of Salim b. Dhakwan, 202. 16 Van Ess, TG, 2:189, 202. 17 Crone and Zimmermann, The Epistle of Salim ibn Dhakwan, (Oxford 2001), 202 3; Ibn Îazm, al FiÒal fi al milal wa al ahwa’ wa al niÌal (Cairo 1317 21), 4:191. 18 Al Siyar wa al jawabat, 1:204. 19 A. al Salimi, TuÌfat al A{yan bi sirat ahl ¨Uman (Cairo 1961), 1:158. 20 Watt, Free Will and Predestination, 32. 13
478
THEMES OF THE IBA∆I/OMANI SIYAR
gested that the Shu¨aybiya is a sub-sect of the ¨Ajarida, and that the discussion between Shu¨ayb and Maymun regarding the topic of predestination led to schism among the ¨Ajarida in the last quarter of the first/seventh century.21 Moreover, conflict over this issue has been recorded between the Iba∂is and the ¨Ajarida. As for the Sha¨biya, it was a sub-sect of the Iba∂iyya that was formed gradually at the end of the second/eighth century and the beginning of the third/ninth century by ¨Abd Allah b. Yazid alFazari, Shu¨ayb b. al-Ma¨ruf, Ayyub al-∑awwaf and Abu al-Mu’arrij, then later by Harun b. al-Yaman.22 Subsequently, this development of Iba∂ism expanded in two ways. Politically, development occurred during the election to the Imamate of ¨Abd al-Wahhab in Tahart, in North Africa; the leaders of the Sha¨biya in this regard supported the Nukkaris against ¨Abd al-Wahhab. Theologically, expansion occurred during the debate between MaÌbub and Harun. Thus there emerged at this point a new sub-sect of Iba∂ism, although it is not known what motivated the Sha¨biya’s split nor after whom it was named. As mentioned above, al-Bisyawi refuted the opinions of the Sha¨biya as anti-Iba∂i, and listed the Iba∂i scholars who were the proponents of the Sha¨biya.23 Furthermore, al-Bisyawi notes that there were two Iba∂i sub-sects at the end of the second/eighth century: the ™urayfiya or ™arfiya (the eponym of ¨Abd Allah b. ™urayf who was a military leader with ¨Abd Allah b. YaÌya in Yemen (131/748) and the Sha¨biya.24 Supporting this suggestion, there is a letter by al-Rabi¨ b. Îabib concerning the leaders of the Sha¨biya, who were ¨Abd Allah b. ¨Abd al-¨Aziz, Abu al-Mu’arrij, and Shu¨ayb b. al-Ma¨ruf.25 In contrast, Abu Ghanim al-Khurasani, despite being one of their pupils who recorded Iba∂i traditions from them in Basra, did not support the Sha¨biya.26 Imam al-Muhanna’s epistle to Mu¨adh b. Îarb (XLIV) indicates a change in the conception of qadar in Oman, and his discussion of qadar caused a division in Omani society. One side, the Qadariya, claims that God does not create or form the actions of men. The other side believes that God creates belief (i.e. good) and disbelief (i.e. bad), and that He is responsible for people’s actions. The con21
Al Ash¨ari, Maqalat, 14. Al Siyar wa al jawabat, 1:204. 23 Ibid., 2:139. 24 Ibid., 1:204, 2:139. For the opinions of both schools, see ibid., 2:73, 86, 94, 138, 139. 25 A. Ennami, Studies in Iba∂ism (Cambridge University Ph.D. 1971), 10. 26 Ibid., 155. 22
479
THEMES OF THE IBA∆I/OMANI SIYAR
flict between the Qadariya and their opponents surfaced in Oman during the imamate of ¨Abd al-Malik b. Îumayd (208–26/824–41), when some of the Qadariya began to settle and flourish in the cities of Sohar and Tuwam. These suggestions from the siyar material are confirmed when their contribution to the discussion of qadar is considered. There is one significant piece of evidence: Abu ¨Ubayda, an antiQadari, endeavoured to pursue the logic of his teacher ∑uÌar b. al¨Abbas and to introduce to the creed of the anti-Qadari the idea that knowledge equals fate.27 Abu ¨Ubayda (XVIII) informs his followers that the arguments of predestination ‘…reflect the Qur’an, as much of it is about the virtue of knowledge of qadar’. Abu ¨Ubayda argues that ‘If He [God] did not predestine [people], His knowledge is still evident, though if He did predestine [them], this knowledge cannot be said to have shifted into predestination.’ But Abu ¨Ubayda continues, ‘… He [God] began with creation. He started by creating knowledge, since He showed in the Psalms what he was revealed, and signed his creation and order according to what he was confirmed.’ This indicates that Abu ¨Ubayda attributes the term qadar to God’s knowledge by following his teacher ∑uÌar b. al-¨Abbas al-¨Abdi’s commentary. In this way, the development of the concept of qadar became an issue in Oman. Imam al-Muhanna’s (226–37/841–51) epistle asserts that the action and destiny of people emanate from God’s knowledge. According to the commentary on this concept, the Iba∂is define qadar as God’s knowledge: God’s knowledge and His determination are two inseparable things. God requires His servants to do what He orders them to do according to their capacity, not their wish. The essential function of the doctrine of qadar is the portrayal of the ideal relationship between man and Creator. The discussion began among the early Iba∂is with the question of whether the origins of actions are attributed to God or to humans. Îamza alQadari asked Îajib, ‘Do you accept the statement that good is from God and evil is from humans?’ Îajib replied, ‘We accept the statement when it comes from people, but not when it comes from you.’28 This position is that of al-Îasan al-BaÒri: he affirmed that God creates only good, and that men’s evil acts come from themselves or from Satan; however, al-Îasan allowed that God’s guidance of men contained an element of succour, or grace (tawfiq).29 It is in27
Al Shammakhi AÌmad bin Sa¨id, Kitab al Siyar, 2:233; van Ess, Anfänge, 38; M. Cook, Early Muslim Dogma (Cambridge 1981), 142. 28 Al Siyar wa al jawabat, 1:386. 29 W. Montgomery Watt, Islamic Philosophy and Theology (Edinburgh 1985), 27. 480
THEMES OF THE IBA∆I/OMANI SIYAR
teresting to note that the siyar show that ¨Imran b. ¨Abd al-¨Aziz al-Nidabi [d. c. late second/eighth century], who was the imam of Nidabi30 mosque in Basra, said to al-Rabi¨ b. Îabib, ‘I cannot claim that God in his wisdom and justice asked his servants to do things which they are not able to do’. Al-Rabi¨ asked him, ‘Was not the guidance of God to Abu Bakr and ¨Umar not the same as to Abu Jahl’ (i.e to those who follow the right path and those who went astray) ¨Imran said, ‘No.’ Although ¨Imran insisted on his opinion, this did him no harm within Iba∂i circles. MuÌammad b. MaÌbub (d. 260/873) also reports the discussion of qadar in Iba∂i Ìalqas, (circles of the leading ¨ulama’ including Qadari scholars in Basra and an Iba∂i called al-Ghazzal, who came from Sohar to debate with them. Al-Ghazzal said to a man from the predestinarian group, ‘Which is good, the action of God or the action of humans?’ The man replied, ‘It is God’s action that is good.’ Then al-Ghazzal asked, ‘Is prayer part of God’s action or part of the servant’s action?’ The man replied, ‘It is part of the servant’s action.’ Then al-Ghazzal asked, ‘Is sleep part of God’s action or part of the servant’s action?’ He replied, ‘It is part of God’s action.’ Al-Ghazzal then said, ‘So the servant’s action is better than God’s action.’31 The Iba∂i views generated yet more discussion. They claimed that God did not force any of His creatures to obey or disobey Him, and they also called the Sunni (i.e. the majority who favoured determinism) ‘Jabriya’, since the majority view holds that God forces His servants to commit sins. MuÌammad b. MaÌbub comments that whoever does good or bad is simply executing God’s knowledge. Abu ¨Ubayda says that God commands obedience and makes it obligatory, and whoever obeys or disobeys, this is God’s knowledge. From this view, the debate on qadar is concerned with two main questions: who creates actions, and does God force people to do good or bad? An Iba∂i called Abu ¨Abd al-RaÌman asked Abu ¨Ubayda, ‘Does God force anyone to obey?’ Abu ¨Ubayda said, ‘I did not learn that God forces anyone either to obey or disobey. But you say that God forces the people of piety to be pious when he shows them its reward.’ Abu ¨Abd al-RaÌman said, ‘People flock to the sins they commit with God’s knowledge.’ Abu ¨Ubayda said, ‘By God, it is not like that; rather, their souls tempt them and Satan lures them to do these sins.’ He provides as evidence a verse in the Qur’an, which says, ‘…Satan 30
Nidab is a sub tribe of Azd in Oman; S. al ¨Awtabi, Ansab (Muscat 1981 and 1984), 2:243. 31 Ibn Baraka Abi MuÌammad ¨Abdullah, Taqyid (Al Salimi Library, Oman) MS 55.MNHC 481
THEMES OF THE IBA∆I/OMANI SIYAR
it was that tempted them.’32 He argued that God made people able to obey for the sake of reward, and prohibited them from disobeying for fear of punishment; He thereby encourages people to do what He prefers and chooses. So the servant chooses without any coercion on the part of God. As a result, he is either rewarded or punished for his actions. In the following centuries, the development of this subject gave rise to discussions about God’s will. Abu al-Mu’thir al-∑alt b. Khamis (LXI) tries to give a different interpretation of the concept of qadar. He stated that creation is equivalent to qa∂a’ (see EI 2), or more specifically, the qadar is the creation itself. Thus Abu al-Mu’thir refutes the attack of both the Qadariya and the Jabriya. In the middle of the fourth/tenth century, Abu al-Îasan al-Bisyawi (LXXXIX) asserted that qadar is ability (isti†a¨a). His view reflects the repudiation of the Mu¨tazili view that the specific ability to perform an action comes before action or simultaneously with action. In contrast, al-Bisyawi’s teacher, Ibn Baraka, argued that ability and power have the same meaning and have no independent existence, and that they are created in the servant who is given the ability and power at the point of acting, and is thus free to choose. But some Iba∂i scholars suggest that ability is created simultaneously with action and not before or after it, and it is not one type of ability but more than one, for each action has its own corresponding ability. Thus al-Bisyawi explains: if the ability were before action, there would be no motivation either to sin or to obey; so if a person were able, he would act. With respect to the notion of acquisition (kasb), it consists of every act that takes place with the ability to do that act. Ability means that people are accountable to God’s will (mashi’a). Ibn Baraka states that ‘the servant is not accountable for God’s creation and ability; rather, he is accountable for his committing sins and disobeying the orders of God’.33 So the question of fate can be laid aside until someone commits an action. Accordingly, the early Iba∂is generally recognize that determination and intention carry the same meaning. One is the intention in what God revealed to His prophets and in His guidance; the other is the intention in God’s knowledge of His creation. From this brief overview of Iba∂i authorities, it is clear that qadar was the subject of heated debate until the fourth/tenth century, with the aim of defining the conception of fate. The siyar are thus informative guides to the Iba∂i contribution to dogmatic discourse in 32 33
Qur’an, MuÌammad 47:25; Ibn Baraka, Taqyid, MS 58. Ibn Baraka, Taqyid, MS 56. 482
THEMES OF THE IBA∆I/OMANI SIYAR
Islam. Similar discussions defined the concept of movement (Ìaraka), where it is considered to be God’s creation in the person at the time of action. By tracing this material from the formative period of Iba∂ism, Ibn Baraka views it as one of the essences of God since will (irada) is one of the attributes (Òifa) of God. The Iba∂is formed their own opinions, and took up a middle position between those of the Qadariya and the Jabriya. Of course, the influence of the Qadariya was stronger than that of the orthodox Sunnis, yet for the Iba∂is in the centre of Basra, the topic of fate was under consideration until the end of second/eighth century, and only a minority of Iba∂i scholars were influenced by the Mu¨tazila then. Another theme in the siyar is sin and its connection to belief. It is not yet known for certain whether or not this creed was originally influenced by Christianity.34 In order to understand the term sin, examples of its use in Iba∂i literature are given, starting with ¨Abd Allah b. Iba∂’s letter to ¨Abd al-Malik. In this epistle, Ibn Iba∂ clarifies the differences regarding the concept of sin between the Khawarij and the Iba∂is, explains his own view, and introduces a refinement on the concept by distinguishing between the terms mushrik (idolater) and kafir (disbeliever).35 Throughout his reply, however, he employs the phrase kuffar al-ni¨am (ingratitude), which means that he equates sin with kuffar, disbelievers, since sin is against belief.36 Wensinck has shown that the connection between sin and disbelief is found originally in Prophetic tradition.37 Abu ¨Ubayda and Abu Mawdud (XVIII) investigated the topic in order to find out how sin leads to disbelief. They distinguished two types of sin: 1) that which is committed intentionally and contrary to the sinner’s better judgement, and 2) that which is committed unintentionally as a result of the sinner’s ignorance. The matter of sin and disbelief also involves the issue of innovation (bid¨a), of which there are two types: 1) that which involves polytheism and the ascription of partners to God, and 2) that which involves innovation in religion by monotheists. The issue of sin was discussed among the Iba∂is in Basra at the beginning of the third/ninth century. Perhaps the most outstanding example of this is the debate in which both Iba∂i leaders, MaÌbub 34
See Wensinck, Muslim Creed, 68; Cook’s discussion of Jacob of Edessa in Early Muslim Dogma, 145. 35 W. Montgomery Watt, The Formative Period of Islamic Thought (Edinburgh 1969), 29. 36 FarÌat al Ja¨bayri, al Bu¨d al Ìa∂ari lil ¨aqida al iba∂iya (Muscat 1987), 508. 37 Wensinck, Muslim Creed, 39. 483
THEMES OF THE IBA∆I/OMANI SIYAR
and Harun, attempted to classify the concept of sin. According to Harun, sin is of three types: 1) sin that implies the disbelief of its perpetrator, 2) sin in which it is irrelevant whether or not the perpetrator is a disbeliever, and 3) sin that God forgives. In short, Harun identifies three classes of sin; they are, respectively, the major sin, the minor sin, and the sin of suspended judgment (mawquf ¨anha). It seems that Harun’s intention was to link the conception of sin to the attitude of the grave sinner, and he deduced this from the conception of the intermediate position (al-manzila bayn al-manzilatayn) of the Mu¨tazilis. Perhaps he attempted to connect this issue with the doctrines of association (walaya), dissociation (bara’a), and suspended judgment (wuquf ).38 It should be made clear that the essence of the doctrines of association and dissociation has no connection with eschatological judgments, and concerns only the present world. Imam al-Muhanna b. Jayfar (XLIV) also addressed the concept of sin. He equates sin with kufr (disbelief ). Thus he investigated the actions that lead to sin, and classified sin into two types: 1) disbelief that consists of denial of revelation, and 2) disbelief that consists of disobedience and error in interpretation. Alongside the matters of will and sin, the third subject of theological discussion is divine, Godly matters (ilahiyat), often referred to as ‘belief in God’s attributes’. This issue led to a profound investigation by Muslim scholars. While anthropomorphism (tajsim) proved to be a major motive for the interpretation of the Qur’an, the metaphorical expressions in the Qur’an encouraged diverse theological opinions on the question of divine attributes. With regard to this development, Watt assumes that it most likely began at the end of the first/seventh century, and he connects it with the concept of qadar, since there were vigorous arguments about the Qur’an that arose from questions about the meaning of God’s speech.39 The articulation of Iba∂i views began with Abu ¨Ubayda in the first half of the second/eighth century. MaÌbub recalls that it was said to Abu ¨Ubayda that Muqatil b. Sulayman used to say that God created Adam in His own form. Abu ¨Ubayda said that Muqatil was wrong.40 MaÌbub and Harun both emphasized the metaphorical interpretation of the Quranic verses, and they adhered to the principle 38
On the development of the conception of wuquf, see Ignaz Goldziher, Die Richtungen der islamischen Koransauslegung (Leiden 1920), 178; van Ess, ‘Skepti cism in Islamic Religious Thought’, al AbÌath 21 (1968), 1 18; Watt, Formative Period, 30; Cook, Early Muslim Dogma, 44. 39 Watt, Islamic Philosophy and Theology, 49. 40 Al Siyar wa al jawabat, 1:299. 484
THEMES OF THE IBA∆I/OMANI SIYAR
of the unity of God’s essence (tawÌid ). The controversial theological issues that were under discussion were: 1) God’s attributes, 2) the vision of God on the Day of Resurrection, 3) the revelation by God on the night of 15 Sha¨ban, 4) God’s sitting on the throne (istiwa’ ). There were disagreements between MaÌbub and Harun. The central question was: Are the people who err in their conception of God’s attributes disbelievers or polytheists? The Iba∂is view God’s attributes in a similar way to the Mu¨tazila. According to them attributes did not have a kind of independent existence, but rather emerged from the unity of God’s being. For example, what God knows, he knows by Himself or by His essence, and not by any knowledge distinct from Him.41 In addition, the Iba∂i view is that the speech of God does not change; what changes is only its reading and recitation. As is well known, God’s attributes became a significant political issue, causing a crisis during the miÌna42 ( the inquisition concerning the creation of the Qur’an initiated by the Abbasid caliph al-Ma’mun (r. 196–218/812–33) when a debate developed over the nature of the Qur’an and whether the speech of God was created.43 Historically, the discussion of the creation of the Qur’an arrived in Oman during the period of Imam al-Muhanna, after the death of MaÌbub [220s/830s], but the imam himself did not join in this debate.44 For the Iba∂is in Iraq, however, it is necessary to give Ibn Baraka’s outline of this debate. The Omanis at the end of the second/eighth century asked Abu ∑ufra ¨Abd al-Malik b. Abi ∑ufra about the creation of the Qur’an, and he replied, ‘I understood your letter, but I have not heard any of our scholars mention that the Qur’an is created. What they say is that it is the speech of God.’ Ibn Baraka reports that he met an Iba∂i scholar in Baghdad called Ibn ¨Amrush, who had contact with al-Rabi¨ b. Habib’s contemporaries. Ibn ¨Amrush stated that he had not heard anything on this topic, and that was among the Iba∂is, but he added that he preferred people not to dissociate from those who say that the Qur’an is created. In addition, Abu ¨Abd Allah MuÌammad b. ¨Abd al-Îamid al-BaÌrani reported that he had been informed that Hayf b. YaÌya and ¨Adl b. Yazid discussed this 41
Watt, Islamic Philosophy and Theology, 49. EI2, s.v. ‘MiÌna’. 43 W. Madelung, ‘The Controversy on the Creation of the Koran’, in J.M. Barral (ed.), Orientalia Hispanica (Leiden 1974), 1:504 25. 44 A. al Salimi, Raw∂ al bayan ¨ala fay∂ al mannan fi al radd ¨ala man idda¨a qidam al Qur’an, edited by ¨Abd al RaÌman al Salimi (Biddiya 1994); al Salimi, TuÌfah, 1:155. 42
485
THEMES OF THE IBA∆I/OMANI SIYAR
topic.45 These outlines suggest that the discussion of the creation of the Qur’an arrived in Oman before Muhanna’s time, and the crisis probably began during the time when there was a strong influence from Baghdad. The siyar collections preserve two epistles dealing with God’s attributes during the middle of the third/ninth century.46 The first is from ¨Azzan b. al-∑aqr (d.268/882) (LV), who adopted a similar view to that propounded later by the Ash¨ariya. The second is from the Rustamid Imam Abu al-YaqÂan AflaÌ b. ¨Abd al-Wahhab to the Omanis (LVI), in which the Iba∂i imam of North Africa adopts the view that the Qur’an is created. Yet perhaps this creed would never have come to the fore without influence from elsewhere. In order to understand the origins of the debate over the creation of the Qur’an in Oman, al-Salimi has dealt with the doctrines of the Jahmiya and the Qadariya where they existed in Sohar.47 The early debate on this theological problem took place between MuÌammad b. MaÌbub and Hashim b. Ghaylan, and the Omani scholars, who concluded that God created all things, and without God nothing is created. Perhaps this position was suspended by Imam Muhanna. However, MuÌammad b. MaÌbub might have brought his view from Basra, where he grew up. ¨Abd Allah b. ¨Abd al-¨Aziz followed by {Abd Allah b. Yazid al-Fazari (both in the second half of second/eighth century) were the first Iba∂is to discuss this subject and believed that the Qur’an was created.48 MuÌammad b. MaÌbub’s view of the Qur’an, as recorded by his pupils before he came to Sohar in the beginning of second/ eighth century, was: ‘Do not say that the Qur’an is created nor that it is uncreated. Do not say that the Qur’an is God or not God. Rather say that it is the speech of God.’ The view of the Mashariqa school that the Qur’an was created has not been given wide attention, since it was noted and recorded for almost a century. As is clear from the above, it can be understood from their own materials. 45
Ibn Baraka, Taqyid, MS 25. Ennami, Studies in Iba∂ism, 357, records that al Barradi shows that there were three different opinions in Oman about the creation of the Qur’an: 1) the Qur’an is the word of God and it is not a attribute of God, nor is it His essence (this view was held by Abu ¨Ali Musa b. ¨Ali [d. 230 AH]); 2) the Qur’an is the word of God and His revelation to MuÌammad, and this issue is the one that may remain unknown (mimma yasa¨u jahluhu) (this view was held by MuÌammad b. MaÌbub); 3) rejection of the view that the Qur’an is created, and assumption of the attitude of making no decision (wuquf ). 47 Al Salimi, TuÌfa, 1:140. 48 Al Ju¨bayri, al Bu¨d al Ìa∂ari, 350. 46
486
THEMES OF THE IBA∆I/OMANI SIYAR
2. The Dynamics of Socio-Political History in Oman The relationship between the ‘ulama’, or religious scholars, and governments is not confined to Iba∂ism; it is an aspect of other Islamic groups and doctrines as well. The issues of the role of the imamate and the relationship between government and shari¨a goes back to the earliest stages of Islam.49 The siyar must be examined carefully for textual evidence of these concerns.50 In order to grasp the function of Islamic doctrine, this question must be addressed: What was the contribution of each Islamic sect to Islamic society and state throughout history? In Oman, the socio-political structure that has supported the imamate until the twentieth century has attracted the attention of several scholars,51 leading them to study subjects such as Oman’s religious learning, social structure, political history, and the imamate itself, which became an integral part of Oman’s culture.52 In this section, I will attempt to bring into perspective the relationship between these factors as portrayed in the siyar materials. This study is based on the chronology of events examined in the previous part of the article. Here the relationship between religion and state is discussed from the socio-political perspective, which combines both historical and social aspects.53 49
Cf. Patricia Crone and Martin Hinds, God’s Caliph: Religious Authority in the First Century of Islam (Cambridge 1986); I. Lapidus, ‘The Separation of the State and Religion in the Development of Early Islamic Society’, International Journal of Middle East Studies 6 (1975), 363 85. 50 Hamilton A.R. Gibb’s view is that there is no Arabic word for ‘state’ as a gen eral concept. Even for Ibn Khaldun the word dawla often explicitly means and al ways implies membership in the ruling family. Similarly, the term mamlaka com bines the concept of kingship and kingdom; for example, the interest of the state means the interest of the Umayyad family. See H.A.R. Gibb, ‘The Evolution of Government in Early Islam’, Studia Islamica, 4 (1955), 1 17. 51 Laura Veccia Vaglieri, ‘L’Imamate Iba∂iteta dell’ Omân’, Annali Istituto Orientale di Napoli 3 (1949), 245 82; Wilkinson, Water and Tribal Settlement in South East Arabia: A Study of the Aflaj of Oman (Oxford 1977), 138. 52 Dale Eickelman, ‘Religious Knowledge in Inner Oman’, Journal of Oman Studies 6 (1983), 163 72; J.C. Wilkinson, The Origins of the Omani State, in D. Hopwood (ed.), The Arabian Peninsula Society and Politics. (London 1972), 67 88; J.E. Peterson, ‘Oman’s Odyssey: From Imamate to Sultanate’, in B.R. Pridham (ed.), Oman: Economic, Social and Strategic Development (London 1987), 1 16; Thomas Bierschenck, ‘Religion and Political Structure: Remarks on Iba∂ism in Oman and the Mazb (Algeria)’, Journal of Islamic Studies 68 (1988): 107 27; J.B. Kelly, Sultanate and Imamate in Oman (Oxford 1959). 53 John Wilkinson’s theorizes (1987) that the Imamate of Oman is a social and tribal order rather than a legislative pattern. He tries to narrow the relationship 487
THEMES OF THE IBA∆I/OMANI SIYAR
In the century that followed the end of Umayyad rule, the Iba∂is succeeded in establishing imamat al-Âuhur (openly declared rather than hidden [‘kitman’ ]) in Oman (132/749) and Yemen (129–31/ 746–8), and some ten years later in North Africa (145/763).54 The activity of the Iba∂is in Basra influenced these regions through the pupils (carriers of knowledge)55 of Abu ¨Ubayda.56 Although the siyar do not preserve any exchange of letters between the leaders of Iba∂ism and their students, al-Kudami did present the role of these students in Oman: ‘We do not doubt that the Omanis were not included in the true religion (din al-istiqama), otherwise they would have followed the ∑ufriya. However, in the course of time, the scholars and learned men did go to Basra to obtain knowledge from Iraq, and they came back to teach the people what they had not already known of religion.’57 The Omanis elected al-Julanda b. Mas¨ud as the first imam, who brought to an end the rule of the Julandani dynasty, despite being himself a member of the family.58 The institutionalization of the ‘ulama’ and the political organization of the imamate in Oman was imposed by force, because first, the ousted Julandani dynasty revolted against their relative, the new imam al-Julanda,59 and second, the Omanis wanted independence from the Abbasids. The Iba∂i masters of Basra sent to Oman a group under the leadership of Hilal b. ¨A†iya al-Khurasani (XXIX), in order to try to discourage tribalism, lest it cause revolt against the imam. When the imamate had been established, Imam al-Julanda sent Abu ¨Ubayda and Îajib so as to establish contact with Basra.60 When the between imams and ¨ulama’ and considers that it was formed primarly by tribalism and Omani geography; see The Imamate Tradition of Oman (Cambridge 1987), 205 12. 54 EI2, s.v. ‘Iba∂iyya’. 55 A surviving list of Ìamalat al ¨ilm ila ¨Uman (carriers of knowledge to Oman) includes Musa b. Abi Jabir (c. 87 181/706 97), Abu al Mundhir Bashir b. al Mundhir (d. 178/794), Munir b. al Nayyir al Ja¨alani (c. 170 280?/786 893?) and MuÌammad b. al Mu¨ala al Kindi; see al ¨Awtabi, Ansab, 1981), 2:229; Abu Ghanim al Khurasani, Mudawwana al kubra (Muscat 1984), 2:306. 56 Cf. Khalifat, MuÌammad, Nasha† al Ìaraka al Iba∂iyya, (Amman 1981), 103 15; Ennami, Studies in Iba∂ism, 122; EI 2, s.v. ‘Iba∂iyya’; Wilkinson, ‘The Early Development of the Iba∂ite Movement in Basra’, in Juynboll (ed.), Studies on the First Century of Islamic Society (Carbondale 1982), 125 48. 57 Abu Sa¨id MuÌammad bin Sa¨id al Kudami, Al Istiqama (Muscat 1984), 2:91. 58 J.C. Wilkinson, ‘The Julanda of Oman’, Journal of Oman Studies 1 (1976), 97 108. 59 Al Salimi, TuÌfa, 1:93. 60 MaÌbub reflects this development of the Iba∂i movement in the beginning of the second/eighth century in his sira; Al Siyar wa al jawabat, 1:281. 488
THEMES OF THE IBA∆I/OMANI SIYAR
imamate of Julanda was defeated by the Abbasids, Shabib b. ¨A†iya (XXXI) took on the political role of muÌtasib (acting head of the community) due to his willingness to fill the gap that followed Imam Julanda, and so as to mediate with the centre of Iba∂ism in Basra. On the other hand, Shabib asked for zakat on Omani villages, but eventually ceded his rule to the sultan’s deputies when they came.61 Shabib’s sira directs all people to revolt in order to gain an independent state that revives the political spirit of early Iba∂ism. Nonetheless, relations with Iba∂i regions continued in Basra, as can be seen in the case of Thabit b. Dirham in Oman (XXVII). More precisely, the authority of Basra was firm over all the Iba∂i parties. This can be seen in ¨Abd al-¨Aziz al-Julandani’s case: even though the shurat (militia) in Oman killed him after the imamate of al-Julanda was defeated, the Julandanis nevertheless submitted their case to the imams of Basra who refused to intervene.62 Having traced the growing role of the Iba∂i spirit and its existence as a doctrine in Oman, the position of the Omani scholars, the Ìamalat al-¨ilm (carriers of knowledge) can now be considered. When the latter returned from Basra to Oman, their awareness of the need for moral strength was behind the resistance to the Abbasids. At first, all missionaries trained in Basra, but as the imamate became established, the ¨ulama’ developed their own local centres, perhaps influenced by tribalism. Musa b. Abi Jabir, the head of the carriers of knowledge, appeared again during this interregnum, and was among the people who elected al-Julanda to the imamate, but afterwards he became one of the missionaries. The new pupils who arrived from Basra wanted to take authority away from the former missionaries. Al-Salimi notes that there was a kind of confusion about the relationship between Musa and Shabib.63 According to al-Barradi, during this interregnum and before the imamate of alWarith (179–92/796–808), the Omanis revolted against the Abbasids under al-Kulanda b. al-Julanda until he was killed. Al-Kulanda seems to have been from the Julandanis, although he was trying by his revolt to occupy Yemen as well as Oman.64 It is necessary to briefly mention the structure of the Omani state since the re-establishment of the imamate of MuÌammad b. Abi ¨Affan (177–9/793–5). Given that the connection between the Iba∂is 61
Al Salimi, TuÌfa, 1:106. Ibid., 1:102. 63 Al Salimi, TuÌfa, 1:105; al Barradi also notes that Shabib was from the ∑ufriya; Risala, 54. 64 Al Barradi, al Jawahir al muntaqat, 170. 62
489
THEMES OF THE IBA∆I/OMANI SIYAR
and the clan of al-Azd continued in Basra,65 this relationship appears to have occasionally shaped the local politics of early Islamic states between place, tribe, and doctrine. In other words, the Omani state consisted of three pillars: the tribe of Azd, Iba∂ism, and the internal politics of Oman. These three elements have formed the historical pyramid of Oman. Iba∂i ideology helped to bridge the gap that originally existed between the Arabs and the other inhabitants of the land, whose previous rulers fell with the coming of Islam. In bridging this gap, Iba∂i ideology effectively put a stop to the economic decline of the Julanda period at the end of second/eighth century.66 Wilkinson adds that the Iba∂i ideology, in part, determined the form of Omani political unity, but that it is the physical rather than the cultural factors that have allowed an independent Omani state to survive in some form for nearly 1,200 years.67 It should be noted that the structure of the traditional Imamate system in Oman was distributed among several authorities: the executive authority, which lies in the Imamate, the legislative authority, which is represented by the ¨ulama’, and the military authority, which belonged to the shurat. In the latter part of the second/eighth century, the scholars living under the Imamate’s authority became arbab ahl al-Ìall wa-al-¨aqd, and the scholars were able to unite Omani tribal leaders to elect MuÌammad b. Abi ¨Affan to the Imamate. Musa b. Abi Jabir was worried about having as imam one of the Omani tribal leaders, because he expected that the scholars would then not have any influence on the elections to the Imamate. Moreover, he felt this would raise tribal pride among the Omanis themselves.68 Approximately two years later, Musa b. Abi Jabir discharged MuÌammad b. Abi ¨Affan from his office, and his statement (XXXIV) explains that he was dismissed for his excessive severity. Abu al-Mu’thir records that the scholars who elected MuÌammad b. Abi ¨Affan were themselves the ones who dismissed him. Musa b. Abi Jabir dismissed MuÌammad b. Abi ¨Affan from the Imamate because he had given his allegiance to al-Warith b. Ka¨b (179–92/796–808).69 He may also have been dismissed because he was from the shurat of Basra, and not actually from Oman.70 Abu al-Îasan al-Bisyawi (LXXI) justifies his dismissal for yet another reason: that the ‘ulama’ used him either as 65
Wilkinson, ‘Early Development’, 140. Wilkinson, Water and Tribal Settlement, 142. 67 Wilkinson, Origins, 69. 68 Abu Qăan’s sira; al Siyar wa al jawabat, 1:115. 69 Al Siyar wa al jawabat, 1:65. 70 Al Salimi, TuÌfa, 1:111. 66
490
THEMES OF THE IBA∆I/OMANI SIYAR
an imam of defence until the war ended, or merely as the commander of an army. The biggest success of the ¨ulama’ was in transferring the capital from Sohar to Nizwa so they could preserve both the imamate and the connection between the coast and the interior of Oman. Ever since this period, Nizwa has been known as bay∂at al-islam (the seat of the imamate). Indeed, it was these situations that naturally allowed the ¨ulama’ to become an active political organization in Omani history, rather than remote religious scholars.71 The remarkable achievement of the ¨ulama’ administration was that, when al-Warith died, they selected his successor before the tribal leaders assembled, in order not to lose their role.72 The siyar that were written during the imamate of Ghassan b. ¨Abd Allah (192–207/808–23) seem to have been the most useful method for disseminating legislative rulings, when the ‘ulama’ existed as muftis, counsellors, and viziers for the imams. The executive authority may, however, have been understandably reluctant to accept Munir b. al-Nayyir’s sira to Imam Ghassan b. ¨Abd Allah, when it directed the imamate of Oman to look outside its territory in an endeavour to expand the imamate’s authority overseas into the Indian sub-continent (ar∂ al-Hind). When the imam tried to protect the merchant traders of the Gulf, he had to establish an Omani navy to form a force against Indian pirates (bawarij al-Hind ).73 Attempts to dominate Gulf trade were a major reason for the emergence of Oman as a state.74 The interesting thing is that after Imam Ghassan’s death in 207/823, his successor was not recognized until 208/824. This means that the state was in an interregnum, and rule of the state was effectively under the control of the ¨ulama’.75 However, Wilkinson suspects that Imam Ghassan may have died in 208/824, or that his successor was elected in 207/823.76 When ¨Abd al-Malik b. Îumayd (208–26/824–41) succeeded to the imamate and was formally appointed by the ¨ulama’, Hashim b. Ghaylan (XXXVIII) indicated to him that he should keep an eye on the Qadariya and the Murji’a, whose missionaries had established their ideologies in Sohar and Tuwam. The flourishing of Omani cities was perhaps the main factor that attracted the doctrinal refugees, 71
See Wilkinson, Water and Tribal Settlement, 143. Al Salimi, TuÌfa, 1:111. 73 Ibid., 1:123; al Mas¨udi, al Tanbih wa al ishraf, ed. M.J. de Goeje (Leiden 1967), 355. 74 Wilkinson, Origins, 77. 75 Al Salimi, TuÌfa, 1:122. 76 Wilkinson, Imamate Tradition, 10. 72
491
THEMES OF THE IBA∆I/OMANI SIYAR
whether Iba∂is or others. The most important account occurs, for instance, in the sira of Musa b. ¨Ali (XLII), in which it is recorded that the ¨ulama’ agreed to dismiss Imam ¨Abd al-Malik b. Îumayd from office, as his advanced age had made him unable to control the country’s affairs;77 Abu Qăan adds that his mind was malfunctioning. This affair shows the inability of the ¨ulama’ to deal effectively with claims against high and powerful state officials. The ¨ulama’ discussed the idea of dismissing him, but Musa b. ¨Ali suggested that the scholars should control the army and run the state (LXVI). Thereafter, Oman had semi-independence, with complete autonomy over the military and the navy. Since the arrival of MaÌbub b. alRaÌil in Oman during the imamate of Ghassan b. ¨Abd Allah, his son MuÌammad b. MaÌbub had become a judge in Sohar, and was then succeeded by another member of his family.78 In light of these considerations, it is clear that families that had acquired religious knowledge in Oman, such as MaÌbub’s (RaÌili) and Musa b. ¨Ali’s (Banu Sama) families, became powerful authorities and continued to influence Oman’s political climate.79 The election to the imamate of al-Muhanna b. Jayfar (226–37/ 841–51) brought prosperity to Oman, since he was both the imam and a scholar, and this distinguished him from the previous imams.80 This can be seen in his sira to Mu¨adh b. Îarb (XLIV), in which he discusses Iba∂i theological and legal issues. During his reign, he restructured the imamate’s administration and reformed the military. A standing army replaced the militia (shurat), thus strengthening the executive power. Al-Salimi gives a description of the imamate’s army: the navy had three hundred ships, and the army had between 1,300 and 9,000 horses and camels, and also 10,000 foot soldiers, as well as a standing force composed of several ethnic minorities, including Indians.81 These developments affected the constitutional structure of the state, since the intention of the imam was to weaken the legislative power in state politics. Some of the ¨ulama’ were dissatisfied with these measures, and they decided to dismiss him. Musa b. ¨Ali, who was ra’s ahl al-Ìall wa-al-¨aqd, and who had the most influence, set out to discharge him from the office of the imamate. Imam alMuhanna told him, ¨O Abu ¨Ali (Musa), you are coming to me. I swear by God, if you obey the Omanis in what they wish, no imam 77
Al Siyar wa al jawabat, 1:116. Crone and Zimmermann, The Epistle of Salim ibn Dhakwan, 313. 79 Wilkinson, Imamate Tradition, 125, 174. 80 Al Salimi, TuÌfa, 1:107; Wilkinson, Imamate Tradition, 153, 159. 81 Al Salimi, TuÌfa, 1:107. 78
492
THEMES OF THE IBA∆I/OMANI SIYAR
would stay with them even for a year. Go back! I did not allow you to come, and you have not received permission to come.’82 Abu Qăan, in his sira, recognizes that MuÌammad b. MaÌbub and Bashir b. al-Mundhir knew that the conduct of al-Muhanna was incompatible with the qualifications required for retention of the imamate, and so they secretly renounced him.83 The first imamate in Oman was the golden age that ended with the imamate of al-∑alt b. Malik (237–72/851–86).84 Yet the sociopolitical theory of the time had no solution for dealing with an aged and invalid imam. Was it permissible to depose him from his office or not? This issue is similar to the one faced by the caliphate of ¨Uthman, and now the Omanis faced the same situation. The illustration of these relationships shows that all Omani imams in the first imamate, from the first, who was Julanda, to the last, who was either Rashid b. al-NaÂar or ¨Azzan b. Tamim, were more or less targeted for dismissal from the imamate. The only exception is Ghassan b. ¨Abd Allah.85 The real importance of the first imamate is that it consolidated Oman as a state.86 Before summing up this section, the role of the shurat in the siyar will be discussed, as it was mentioned that an armed force of Iba∂is was sent from Basra to establish the imamate of Julanda b. Mas¨ud. The shurat reappeared strongly during the interregnum period after the imamate of Julanda. However, when MuÌammad b. Abi ¨Affan held office as the imam, the shurat were perhaps joined and backed by young Omani tribesmen. The shurat were becoming a voluntary army inspired by Iba∂i ideology. Abu al-Îawari’s sira (LXXXIV) describes the actions of Sa¨id b. Ziyad, the head of the shurat during the imamate of MuÌammad b. Abi ¨Affan, when he burnt the property of opposing tribes. The case of Sa¨id b. Ziyad’s behaviour was debated among the ¨ulama’ in Oman and Basra.87 According to Ibn Baraka, Ibn Ziyad was exiled to Bahrain and later returned to Oman with the permission of Ghassan b. ¨Abd Allah.88 Imam MuÌammad 82
Ibid., 1:151. Al Siyar wa al jawabat, 1:117; al Salimi, TuÌfa, 1:158 9. 84 J.C. Wilkinson, Sources for the early history of Oman, Studies in the History of Arabia (Riyadh 1979), 54. 85 See al Salimi, TuÌfa: Julanda b. Mas¨ud (1:93), MuÌammad b. Abi ¨Affan (1:111), ¨Abd al Malik b. Îumayd (1:134), al Muhanna b. Jayfar (1:151), and al ∑alt b. Malik (1:195); he also discusses the imamate of Rashid b. al NaÂar and ¨Azzan b. Tamim. 86 Wilkinson, Origins, 54. 87 Abu al Îawari’s sira; al Siyar wa al jawabat, 1:341. 88 Ibn Baraka, Taqyid, 67. 83
493
THEMES OF THE IBA∆I/OMANI SIYAR
b. Abi ¨Affan himself was among the shurat, and was dismissed from the office of the imamate as a result of his behaviour. The shurat system, however, went much further than the irregular and voluntary, non-professional army of the old system; it entirely revolutionized the basic concept of an army. The early imams were faced with a particular problem. After the shurat ignored Imam al-Warith’s orders and broke into the prison at Sohar to kill the Abbasid general ¨Isa b. Ja¨far, who was a close relative of the caliph, in Dhu al-Îijja 189/ November 805, it was essential to develop a new system of shurat organization;89 it was necessary to bring the shurat under state authority. By the latter part of the imamate of Ghassan b. ¨Abd Allah, shurat members attacked and killed al-∑aqr b. MuÌammad b. Za’ida al-Julandani, even though he was under the protection of the imam’s governor in Sama’il; Abu al-Wa∂∂aÌ and Musa b. ¨Ali were neither able to protect him nor stop the attackers, because they were both afraid of the shurat.90 The sira of Munir b. al-Nayyir to Imam Ghassan b. ¨Abd Allah (XXXVI) is considered one of the most important documents that describes the early shurat. The author says: Their opinions united on the strength of the truth and the rule of religion to recruit a force of three to four hundred shurat leaders. We have been informed that some of these shurat leaders were extremely pious people to the extent that they would return the surplus dirhams or the two dirhams of their expenditure to the treasury.
He comments on the changing of the shurat: After this first generation, others came later, claiming to follow the path of the shurat, but they left the call of religion and proclaimed tribalism and adopted fanaticism amongst themselves. At the same time, the scholars were reluctant to oppose this trend, perhaps out of fear of them. Moreover, the shurat led recruits who had not been approved by the Imam.
89 The History of al ™abari: Volume XXX; The ¨Abbasid Caliphate in Equilib rium, trans. C.E. Bosworth (Albany 1989), 256; Al Baladhuri, AÌmad bin YaÌya, FutuÌ al buldan, 356; al Salimi, TuÌfa, 1:119; SirÌan, ‘Annals of Oman’, 14; Wilkinson, Imamate Tradition, 185. 90 SirÌan, ‘Annals of Oman’ (translated by E.C. Ross in the Journal of the Asiatic Society of Bengal [1874]), 16; al Salimi, on the other hand, justifies why the imam did not condemn the killing: the imam might have either secretly instructed the shurat to kill the man in order to avoid publicizing the involvement of the imam in this incident, or the murderer might have committed this killing because he knew he had the right to do so; see TuÌfa, 1:124.
494
THEMES OF THE IBA∆I/OMANI SIYAR
These statements show the transformation of the shurat’s affairs, and how criticism and disapproval caused the change in the shurat. One practical effect seen in these accounts is the encroachment of the shurat on the imamate’s authority, and how in the course of time fanaticism and dogmatism had appeared, especially when Imam al-∑alt was dismissed from the imamate. Imam al-∑alt describes the events (LXVI) saying that he commanded the shurat and all those concerned, such as the soldiers, to fight, but they refused, and then he ordered them to proceed. They lagged behind and discipline became so weak that al-∑alt became afraid of the potential for violence between the soldiers and the people, and the threat of war and bloodshed. He adds that this was the reason why he had to move to his son’s house without leaving the imamate. Wilkinson’s hypothesis concerning the shurat and the military power of the imamate may be correct. Since, in theory, the imam and his community are indissolubly bonded by a divine contract to support one another, it is the duty of the Muslims to obey their leader in the call to arms. Because the military capacity of the community is always at the imam’s disposal, he has no need of a standing army; indeed, he is not permitted to have one.91 In 272/885, Oman went through a period of turmoil that led to the collapse of the imamate after the departure of Imam al-∑alt from the Imamate (LXXXI). Abu Qăan outlines the reasons that led to the collapse of the state of the Omani imamate (LXVI): ‘In this state, there were people and youths with no piety, outwardly showing faith but concealing the love of life, and they displayed their religion for the sake of their life’. Omani tribal solidarity (¨aÒabiya) became involved in the issue, which turned into tribal schism and dogmatism, and then into five civil wars, until they themselves caused the downfall of their state, since the northern tribes were backed by the Abbasid governor in Bahrain, MuÌammad b. Nur (Thur or Bur?). The war and the imamate ended when Imam ¨Azzan b. Tamim was killed in 280/885.92 The ¨ulama’, the (legislative authority), were losing control, without being aware of the manipulation undertaken by 91
Wilkinson, Water and Tribal Settlement, 140. The History of al ™abari: Volume XXXVIII; The Return of the Caliphate to Baghdad, trans. Franz Rosenthal, 10; Abu al Qasim ¨Ali al Nashibi, ∑urat al ar∂, edited by J.H. Kramers, (Leiden 1938 9), 38; al IÒ†akhri, Abu IsÌaq, al Masalik wa al mamalik, edited by M.J. de Goeje, (Leiden 1870), 26. Ibn al Mujawir, Tarikh al mustabÒir, edited by Löfgren (Leiden 1951 4), 2:281; al Salimi, TuÌfa, 1:257. For an analysis of tribal affairs in the physical geography of Omani commu nity structure, cf. Wilkinson, Water and Tribal Settlement, 137, 188. 92
495
THEMES OF THE IBA∆I/OMANI SIYAR
tribal fanatics. Thus the relationship between the ¨ulama’ and the tribal leaders was transformed. Al-Fa∂l b. al-Îawari was the most prominent scholar backed by al-Îuddan and the northern tribes. AlFa∂l (LXXXI) supported ¨Abd Allah b. al-Îawari against Musa b. Musa, who had deposed Imam Rashid b. al-NaÂar. Furthermore, Musa b. Musa’s policy failed, because it was accepted by neither Imam Rashid b. al-NaÂar nor Imam ¨Azzan b. Tamim. The civil war ended with the killing of al-Fa∂l b. al-Îawari and Musa. It was really a conflict between the imams and the ¨ulama’, in spite of the fact that all the imams of that period were controlled by the ¨ulama’ and the tribal leaders. As a consequence of this fitna, the ¨ulama’ became divided over the theory of the physical nature of the imamate. The disagreement about deposing Imam al-∑alt b. Malik emerged as an extreme division between the moderate Nizwa and extreme Rustaq schools. The Nizwa school sought reconciliation over issues that had led to the crisis in the imamate at the end of the third/ninth century. The Rustaq school sought the legitimating of the YaÌmad imamate by excommunicating the party that deposed Imam al-∑alt in 272/885. In the contemporary records of the Nizwa school, the outlines of the disputes appear in the sira of al-Azhar b. MuÌammad b. Ja¨far (LXXVI), and in the comments in the sira of Abu ¨Abd Allah MuÌammad b. RawÌ (LXXVII); the main neutralist position is given in the sira of Abu ¨Abd Allah Nabhan b. ¨Uthman, who tried to sponsor the imamate of Abu al-Qasim Sa¨id b. ¨Abd Allah. The views of both schools came from the second generation, who were Abu MuÌammad b. Baraka and his pupil Abu al-Îasan al-Bisyawi for the Rustaqis, and Abu Sa¨id al-Kudami for the Nizwanis. In the field of Iba∂i technical legislation, the masters of both schools (Ibn Baraka and al-Kudami) consolidated and improved the Iba∂i Mashariqa school by reforming its constitution.93 The geographers al-IÒ†akhri and Ibn Îawqal report that after the Abbasids destroyed the Omani state, the Omani Iba∂is moved to the interior of Oman, where they elected their own imam in Nizwa.94 The importance of the ¨ulama’ was to preserve Iba∂i ideology, and to become pragmatic rather than dogmatic, transforming the imamate from Âuhur to difa¨ (defence). However, the alliance between both schools prevented any possible foreign intervention and the re-estab93
See their book list; Al Salimi, Abdullah bin Humayd, Al Lum¨a al Mar∂iya min al ashi¨at al Iba∂iyya, MNHC, Muscat, 1983. 94 Ibn Îawqal, ∑urat al ar∂, 38; al IÒ†akhri, al Masalik wa al mamalik, 26. 496
THEMES OF THE IBA∆I/OMANI SIYAR
lishment of the imamate. Abu al-Îawari (LXXXIV) recounts how the Omanis revolted in the interior, where they were able to kill BaÌira, the deputy of MuÌammad b. Nur. Moreover, Abu Qăan (LXVI) and Abu ¨Abd Allah MuÌammad b. RawÌ (LXXVII) give a list of imams who were elected during this forty-year interregnum: [The ¨ulama’ ] elected to the imamate Rashid b. al-NaÂar twice, ¨Azzan b. Tamim, al-∑alt b. al-Qasim al-KharuÒi twice, and al-Îawari b. ¨Abd Allah, ¨Abd Allah b. MuÌammad al-Îuddani [known as Abu Sa¨id alQurma†i]. Therefore, MuÌammad b. al-Îasan al-KharuÒi was elected as shari imam in 280/893, but soon abdicated and resigned from the imamate, then al-Îawari b. Mu†raf al-Îuddani twice, ¨Umar b. MuÌammad b. Mu†arrif al-Îuddani, MuÌammad b. Yazid, al-Îakam b. al-Mila twice, and ¨Azzan b. al-Hizabr.
Unfortunately, they did not name them in chronological order. MuÌammad b. RawÌ states that ‘the Omanis elected eight imams after Imam MuÌammad b. al-Îasan, although they elected some of them as shira’ and some as difa¨’.95 Abu al-Mu’thir (LIX) does not know whether the imams mentioned in the above list were just or unjust imams; nor is there any evidence concerning those imams who had backing from either the Rustaq or Nizwa schools. They were more likely supported by the Nizwa school, since all the Rustaq siyar fail to list or praise them. Abu al-Mu’thir’s sira (LXII) to MuÌammad b. Ja¨far questions why the latter allowed Imam ¨Azzan b. al-Hizabr to choose whether or not to fight, since he was an imam of shira’ and therefore should have fought. In the course of this interregnum in the imamate, the consequent adoption by Imam ¨Abd Allah b. MuÌammad al-Îuddani (Abu Sa¨id al-Qarmati) of the Carmathian heresy was regarded by the ¨ulama’ as just cause for his removal.96 95 SirÌan reports: ‘My own opinion is that none of the imams mentioned after al ∑alt b. Malik were universally recognised by the people of Oman, and that their rule did not extend over the entire country. They were recognised in certain dis tricts and not in others, by some of the tribes only, and not by all of them. For after the dissension which had arisen amongst them, the people of Oman had ceased to act in unison, and could not agree in the choice of an Imam’; SirÌan, ‘Annals of Oman’, 25. 96 See his details in al Salimi, TuÌfa,:1:266; SirÌan, ‘Annals of Oman’, 23; Ibn Ruzayq, History of the Imams and Seyyids, translated by Rev. G.P. Badger (Hakluyt Society 1871), 26. This imam (Abu Sa¨id) followed the Carmathians, so al Salimi clarified the statute of this imam: ‘[The ¨ulama’] elected him to the imamate before they discovered that he had left the Muslim faith and converted to the Carmathian heresy. The ¨ulama’ then elected al ∑alt b. al Qasim to the imamate’; al Salimi, TuÌfa, 1:266. See also the sira of Abu al Îawwari: ‘We disassociate ourselves from
497
THEMES OF THE IBA∆I/OMANI SIYAR
Until 320/939, the Omanis were united under Imam Abu alQasim Sa¨id b. ¨Abd Allah. Both the Rustaq and Nizwa schools were satisfied with his reign, and they formed an alliance. The restoration of the imamate had put great pressure on the Rustaq school to change their conceptions of the imamate; hence, they allied themselves with the Nizwa school, and this resulted in their preference for the pattern of elective imamate; thus this came to be the position of both Iba∂i schools. With respect to the ¨ulama’ in Oman, the election to the imamate of Abu al-Qasim Sa‘id b. ‘Abd Allah brought back the Iba∂i ideology to Oman. The ¨ulama’ had developed the concept of jababira (tyrant rulers) to describe outside influences or opposing foreign powers, and this concept gave rise to feelings of Omani nationalism.97 The ¨ulama’ presented the imamate as a vehicle for the unity of the people throughout this period. Abu al-Mu’thir ordered that houses belonging to followers of the Carmathians should be burnt down so they could not return.98 The domination of the Nizwa school is attested in Abu al-Îasan al-Bisyawi’s account (LXXXVIII) of Imam ÎafÒ b. Rashid’s war against al-Mu†ahhar b. ¨Abd Allah after he was elected twice to the imamate. Al-Salimi concludes that it is possible that this imam had a negative attitude towards the Rustaq school, with its extreme dogmatism regarding Musa b. Musa and Rashid b. al-NaÂar.99 If this statement is true, then the imam’s authority would still need the approval of the ¨ulama’ and their legislation. Occasionally, they produced siyar which were like masa’il or fatawa (legal opinions); they contain their comments and refutations regarding the nature and obligations of the imamate, as in the sira by Abu Bakr AÌmad b. ¨Umar alManaÌi.100 In 407/1016, the Omanis were able to restore the imamate of Âuhur. However, there is a stark different between the first and second imamates of Oman, in the sense that the second imamate could not re-establish the power of the first, due to lack of substantial resources. The essential role of the ¨ulama’ had made them aware that the state required them to integrate with the tribal leaders, even if these leaders had committed sins. Abu ¨Isa al-Sirri justified this by Abu Sa¨id al Qarma†i, also from his associates, those who suspended judgment on him and those who doubt his allegiance’; ibid., 1:269. 97 Wilkinson, ‘Water and Tribal Settlement’, 139. 98 Al Siyar wa al jawabat, 1:360. 99 Al Salimi, TuÌfa, 1:317. 100 Cf. the previous paper in the third period of the siyar. 498
THEMES OF THE IBA∆I/OMANI SIYAR
arguing that the Imamate needs someone who has power and wealth, even if he is known to be a grave sinner. In cases such as this, scholars should bring the sinner forward and ask him to repent; if he does, then they may elect him as imam. Abu ¨Isa provides proof of this by showing what the ¨ulama’ did when they elected al-Khalil b. Shadhan. Regarding the necessity of integrating a new state, Imam Rashid b. Sa¨id produced his manifesto in order to end the conflict between the Rustaq and Nizwa schools by declaring Rustaq dogma official. It is clear from the siyar and the literature mentioned in the new stage of the imamate that the ¨ulama’ attempted to reconstruct the policy of the imamate and to legislate on the constitution of the imamate. This is evident in the following siyar: Fi al-tawÌid wa-alimama kayfa hiya. Fi al-farq bayna al-imam al-¨alim wa-ghayr al¨alim, and Kitab al-Imama. The finest example of this legislation can be seen in al-MuÒannaf by Abu Bakr ¨Abd Allah b. Musa al-Kindi.101 It is certain that all of the authors of these works were scholars of the Rustaq school. Some siyar show that the imam’s policy and legislation had to be recommended by the ¨ulama’ throughout the imamate of Rashid b. Sa¨id, as the ¨ulama’ had to appear to control the state.102 The relationship between imams and ¨ulama’ in the subsequent period of the second imamate seemed to suffer a breakdown. Three points serve as evidence of this. 1. In the sira by Imam Rashid b. ¨Ali (CII), the imam submits his repentance (which was more like a manifesto) for the manner in which he had excommunicated the judge Nijad b. Musa for his behaviour without the ¨ulama’s prior approval. This sira also includes the ¨ulama’s acceptance of the repentance but also their approval of the excommunication. The evidence in this sira suggests that discord recurred within the ¨ulama’ in controlling the policies of the imamate, as some of the ¨ulama’ opposed the imam. Al-Salimi remarks that ‘the Rustaq’s party revolted against Imam Rashid b. ¨Ali, and they also made an effort to depose him from the office of the Imamate. The heads of this party were Musa b. Nijad and Abu Bakr AÌmad b. Abi Jabir, who started the mutiny from Rustaq city.’103 Increasingly, the clash within the ¨ulama’ developed into two rival factions: the ¨ulama’ of Jawf (in the interior of Oman), who tried to select their own 101
Wilkinson presents the point of view of K. Kanz al adib wa sulafat al labib (a work probably of the second half of the eleventh/eighteenth century) by Salim b. Sa¨id al Sa’ighi (MS Cambridge Univ. Library [add 2896]); see J.C. Wilkinson, ‘Ibadite Imams’, BSOAS 39 (1976), 534 51. 102 Al Salimi, TuÌfa, 1:312. 103 Ibid., 1:321. 499
THEMES OF THE IBA∆I/OMANI SIYAR
imam, and the ¨ulama’ who allied themselves with Imam Rashid b. ¨Ali. Eventually, the imam fought Musa b. Nijad, and the latter was killed in 496/1119.104 This account of a scholar being murdered by an imam reminds us of the great fitna (civil war) in Oman in 275 80 AH. 2. Upon investigation of the time period of the siyar, it appears that there were several imams at the same time in the ¨Umani miÒr, and that they controlled different regions of the country. Abu al-Îasan al-Bisyawi discusses a case in which there were two imams in one miÒr, one in Tuwam and the other in Sohar.105 Thus the Iba∂i theory of state, namely that there should be one imam in each miÒr, began to change.106 3. The election of an imam depended on the candidate’s affiliation with a certain school. This is evident in the sira of al-Îasan al-Bisyawi about ÎafÒ b. Rashid, and in the sira of Abu Bakr AÌmad b. ¨Umar. A third example is the debate in the sira of AÌmad b. MuÌammad b. ∑aliÌ al-Nizwani and his pupil Abu Bakr al-Kindi (CXVII), when the people of Nizwa refused to acknowledge the advice of the Rustaqi imam MuÌammad b. Abi ¨Affan expressed in his sira (CXIV). Another example is the local support of the Gha†afan in Ba†ina: they submitted their pledge to defend the imam in office (CXVIII). Some of the ¨ulama’ during this period seem to have been tribal leaders, and therefore controlled strong forces in the overall structure of the Imamate. This can be seen clearly in Nijad b. Musa’s family. On examining the siyar closely, it can also be seen that there was interference between the tribal elements and the Rustaq school, perhaps as a result of the manner in which the imam was elected. In any case, during this period, Imam Rashid b. ¨Ali’s conduct led to a split in the Rustaq party, but it would seem that other imams might have been present at the same stage. It is possible that one of these imams was MuÌammad b. Abi Ghassan.
Such situations led to the final clash at the end of this stage in 579/ 1183 between Imam Musa b. Abi al-Ma¨ali and MuÌammad b. Malik b. Shadhan (the leader of YaÌmad). This caused both the imamate and the school to split and be divided between the tribal leaders (CXIV). Al-Ba†ashi claimed that MuÌammad b. Malik might have been an imam, and probably the last imam of this era.107 If this 104
Ibid., 1:322. Al Siyar wa al jawabat, 2:191. 106 Al Salimi, TuÌfa, 1:336; al Salimi lists three imams of the second imamate who were not recognized: ¨Amir b. Rashid b. al Walid al KharuÒi, who was elected to the imamate in 476/1083; Rashid b. ¨Ali b. MuÌammad b. Ghassan al KharuÒi who remained in the imamate for eight and a half years; and al Khalil b. ¨Abd Allah b. ¨Umar b. MuÌammad b. al Khalil b. Shadhan. 107 Wilkinson, ‘The Omani and Iba∂i Background to the Kilwah Sirah: The 105
500
THEMES OF THE IBA∆I/OMANI SIYAR
suggestion is correct, then it is difficult to accept that both imams fought when Oman had already developed into a feudal system controlled by imams and tribal leaders, and the second imamate came to an end. In this period, the struggle between the imams and the ¨ulama’ can be observed, first between Rashid b. ¨Ali and Nijad b. Musa and later between Abu al-Ma¨ali Musa and MuÌammad b. Malik, and it can also be seen that the YaÌmad tribes were able to unite under the Banu KharuÒ. However, at the same time, the Azd of ¨Atik became prominent. Thus it can be seen that there was a transfer of tribal leadership from the Azd of YaÌmad to the Azd of ¨Atik. This was the first major change since the previous era began with the challenge to the authority of Julanda, and this heralded the appearance of the Nabhani dynasty. A specific date for when this happened has not been identified. Perhaps this dynasty came into being with the feudal system in the latter half of the fifth/eleventh century and then gradually developed.108 In the following five centuries, Oman was controlled by the Nabhani dynasty.109 However, it should be realized that, as stated by Wilkinson, there was more than one Nabhani family involved in the early history of this period. There is a period of at least 250 years for which there is no evidence that they ruled in Oman. Also, there seems to be no continuity between early and later Nabhani history.110 Nevertheless, this stage is characterized by a certain pattern of conduct in the relationship between rulers and religious scholars, now that the ¨ulama’ faced a dynasty that was both Omani and Iba∂i. The ¨ulama’ had not faced such a situation since the end of the Julandanis. Tracing the historical development of Oman throughout this period is something of a puzzle. Miles gives an account of the end of the sixth/twelfth century and the beginning of the seventh/thirteenth. For forty years, there appears to be no rivalry between the Demise of Oman as a Political and Religious Force in the Indian Ocean in the 6th/ 12th Century’, in R.B. Serjeant and G.R. Smith (eds), A Miscellany of Middle East ern Articles in Memoriam Thomas Muir Johnstone 1924 1983 (Longman 1989), 272 305. 108 See al Ba†ashi, Sayf bin Humud, ItÌaf al A¨yan fi Tarikh ba’∂ ¨ulama’ ¨Uman, 1:390. 109 For further information about the Nabhani dynasty, see Wilkinson, Imamate Tradition, 212 18. 110 Wilkinson, Imamate Tradition, 213. Wilkinson characterizes this age as ‘the Ibadite dark age’; Wilkinson, ‘Bio Bibliographical Background to the Crisis Period’, Arabian Studies 3 (1976), 155. 501
THEMES OF THE IBA∆I/OMANI SIYAR
Nabhanis and the Iba∂is.111 For approximately two and a half centuries, the ¨ulama’ seem to have lacked the ability to elect an imam. Historical sources do not show any imams between 579–809/1183– 1406. It was only in 809–32/1406–29 that the ¨ulama’ appointed Imam Malik b. al-Îawari (or al-Îawari b. Malik?).112 Perhaps the Nabhanis became rulers of a decentralized state system based on tribes and regions. Occasionally, the ¨ulama’ succeeded in electing several imams and controlling the interior of Oman.113 From this historical background, the relationship between the ¨ulama’ and the imamate can be clarified. In this stage, the subject matter of the siyar is based on political crises. For example, Abu MuÌammad b. Sulayman b. Mufarraj produced a sira in 887/1482 confiscating the property of the Nabhani family, since they were supporting tyrants who forcibly occupied people’s properties. Following this situation in 917/1511, Imam MuÌammad b. Isma¨il gave his own approval for confiscating the properties of the Nabhanis. The crisis extended to confiscating the properties of the RawaÌi tribe, since they had backed the Nabhani leaders, Sulayman b. Sulayman and MuÂaffar b. Sulayman, in their war against the imam in 909/ 1503.114 The siyar preserve valuable information about the relationships not only between the ¨ulama’ and the rulers, but also between the ¨ulama’ and the tribes. This can be seen, for example, in sira (CXXIV), which supports Imam Abu al-Îasan b. Khamis b. ¨Amir (839–46/1436–43) against the tribe of YaÌmad. What is interesting is that there were also schisms within the ¨ulama’, especially when Ibn Maddad produced his sira excommunicating Imam MuÌammad b. Isma¨il (906–42/1500–35) and his son Imam Barakat (942/1535). Some of the ¨ulama’ at this time wrote against this sira in support of the imam and his son. The origin of the schism among the ¨ulama’ was the issue of selecting an imam for each school. This can be seen in the war of Bahla castle between Imam Barakat and Imam ¨Abd Allah b. MuÌammad b. al-Qarn in 967/1560. At the beginning of the eleventh/seventeenth century, a new era in Oman started, and a new union appeared as a national reaction to the political unrest, when the settled groups of central Oman united 111
S.B. Miles, The Countries and Tribes of the Persian Gulf (London 1966), 135. Al Salimi, TuÌfa, 1:369; SirÌan, ‘Annals of Oman’, 31. 113 See the list of Omani imams in al Salimi, TuÌfa, 1:274; SirÌan, ‘Annals of Oman’, 24; Ibn Ruzayq, History of the Imams and Seyyids, cxxv. 114 For more on these siyar, see the previous paper in the fourth period of siyar; see also al Salimi, TuÌfa, 1:373. 112
502
THEMES OF THE IBA∆I/OMANI SIYAR
under Imam NaÒir b. Murshid al-Ya¨rubi. It is apparent that the Nizwa school maintained its moderate character and was able to eliminate the restrictive nature of the Rustaq school, and even flourished in the city of Rustaq itself. What is really distinctive in this series of events is the strong desire of the ¨ulama’ to construct and formulate an Islamic constitution for the imamate and to form a society based on an Islamic ideology. 3. Collaboration among the Iba∂is in Different Areas This section offers an overview of the collaboration involved in the disputes among the Iba∂i-ruled areas that the siyar record. The early inter-Iba∂i relationships and their development will be discussed and this collaboration already covered in part through examining the siyar literature will be outlined. Wilkinson has extensively studied local siyar to highlight the early development of Omani Iba∂is in East Africa.115 A similar approach is adopted here to examine the Iba∂is in Asia. The siyar are considered from three aspects: 1) the areas in which the Iba∂is were present, such as Iraq, North Africa, Yemen and the Îa∂ramawt, 2) the nature and relevance of Iba∂ism to the issues discussed, and 3) the development of Iba∂i ideology in each area where the siyar documents and materials originate. Iraq The Iba∂is originated in Basra after they split from the extremist Khawarij. As a result of the emergence of Iba∂i doctrines, the political organization of this new group contributed to the fragmentation of these communities. According to Ennami’s list, it is clear that during this stage the Iba∂i scholar Jabir b. Zayd produced and exchanged several letters and epistles with other early Iba∂i scholars.116 It is significant that Ibn Ja¨far mentions in his Jami¨ two epistles by Jabir (XV, XVI); the first is attributed to Jabir himself, and the second records his legal views. Perhaps these were written by his pupils; both seem to have been used in Oman in the early third/ninth century by the Ìamalat al-¨ilm. 115
Wilkinson, ‘Oman and East Africa: New Light on Early Kilwan History from the Omani Sources’, International Journal of African Historical Studies 6 (1981), 272 305; idem, ‘Omani and Iba∂i Background’. 116 Ennami, ‘A Description of New Iba∂i Manuscripts from North Africa’, JSS 15 (1970), 63. There are also seventeen letters by Jabir b. Zayd to his pupils be longing to the Ministry of National Heritage and Culture, Muscat (9188 204). 503
THEMES OF THE IBA∆I/OMANI SIYAR
The Iba∂i strategy in Basra was to use the epistolary form to communicate with the members of the missionary network about their policies and doctrines. The epistles of Basra indicate that Abu ¨Ubayda advised his followers to continue their activities secretly (XIX); these activities seem to have been directed towards the local people of Basra. Some of the early Iba∂i leaders emigrated to Basra (XX). These siyar reflect the fact that a more flexible policy existed throughout the period of kitman (concealment). Nonetheless, during the revolts in Arabia towards the end of the 120s/740s, the demands of the Iba∂is outside Mesopotamia prompted Abu Mawdud Îajib b. Mawdud (XXIII) to order the Iba∂is to organize themselves. ¨Abd Allah b. YaÌya al-Kindi, known as ™alib al-Îaqq (129–31/746–8), produced his epistle during the revolt he led in Arabia. The exact location of the drafting of this epistle is unknown; perhaps it was written in the Yemeni capital of Sanaa, as he did not go to the Hijaz. The religious committee of Basra was still controlling the revolt when A. Sufyan MuÌammad b. MaÌbub in his Kitab A. Sufyan claims that the roots of ¨Abd Allah b. YaÌya in the Yemen, of Abu al-Kha††ab and ¨Abd al-RaÌman b. Rustam in Maghrib (North Africa) and alJulanda b. Mas¨ud in Oman were all organized by Abu ¨Ubayda.117 The epistles provide detailed documentation of what was arranged in Basra when the imamate was in a state of concealment (kitman). These arrangements are interesting, for it can be seen that later Iba∂i history transformed their ideology in the kitman period and used the ¨izaba in North Africa.118 Since the development of early Iba∂i doctrine has already been studied,119 the study now focuses on how the early epistles of Basra were tailored to the different Iba∂i areas. It can be seen that the proto-Iba∂i policymakers of Basra influenced the Iba∂i communities in two ways: 1. Responsiveness. The Basran policymakers highlighted the need for the Iba∂i communities to implement suitable responses to events that arise unexpectedly. For example, Abu ¨Ubayda and Abu Mawdud’s letter (XXVI) to the people of the Maghrib concerning the murder of al-Îarith b. Talid and ¨Abd al-Jabbar orders the enforcement of the concept of suspended judgment (wuquf ) regarding their status between association (wilaya) and dissociation (bara’a). 117
Al Siyar wa al jawabat, 2:265. EI 2, s.v. ‘ÎalÈa’. 119 Ennami, Studies in Iba∂ism, 126. 118
504
THEMES OF THE IBA∆I/OMANI SIYAR
Perhaps Omani scholars (XXVII) were involved in the tribal crisis after the imamate of al-Julanda b. Mas¨ud had ended, since their authority was destroyed. The masters of Basra became involved in the community when they issued an order for the excommunication of the two tribal leaders, Qutayba b. Dirham and Sadus b. Yusuf, who worked in slave trading. For the Iba∂is of North Africa, although their state was founded by ¨Abd al-RaÌman b. Rustam, the question arose as to who should succeed him. At this time, the master of Basra was Rabi¨ (XXXIII), who supported the succession of ¨Abd al-RaÌman’s son ¨Abd al-Wahhab against his rivals. The authenticity of this epistle is debated, and van Ess doubts that Rabi¨ intervened in this matter.120 The North African scholars indicate that the recommendation of ¨Abd al-Wahhab for the imamate came when Abu ¨Ubayda al-Jannawni sent his letter to the Imam (XXVIII). Regarding van Ess’s point, one could agree with him, since Rabi¨ himself was a scholar rather than a politician. On the other hand, the Omani scholars needed a fatwa from Basra about Sa¨id b. Ziyad, who was their leader, when he burnt the property of opposing tribes. Despite the fact that Musa b. Abi Jabir was alive and was the leader of the Omani scholars, the fatwa came from Abu Ayyub Wa’il b. Ayyub, who considered Sa¨id b. Ziyad to be a tyrant.121 It is obvious from these siyar that they were usually written to clarify Iba∂i opinions on issues that resurfaced. The Iba∂i parties were widely dispersed, but they were still controlled by the centre at Basra. Occasionally, the Basrans sent with each of their letters a delegation responsible for producing a solution to controversial issues, as can be seen in their letter to the Omanis (XXVII). 2. Obligation. Based on the survey of the early siyar, it can be seen that the concept of obligation was in use at the time of Abu ¨Ubayda and Abu Mawdud Îajib. The siyar are meant to illustrate the duty of each party. These epistles were usually written in an exhortatory style to instruct followers. They were directed at those who were the arbab ahl al-Ìall wa-al-¨aqd. Yet it is not known whether the masters of the Iba∂is in Basra dispatched their theological epistles as a matter of policy, or rather as counsel to their pupils. In any case, the theme of obligation is echoed in the shurat system of the early Iba∂is, as is seen in 120 121
Van Ess, TG, 2:199. Ibn Baraka, Taqyid, 99. 505
THEMES OF THE IBA∆I/OMANI SIYAR
the manifesto of Khalaf b. Ziyad al-BaÌrani (XXX). Among the early Iba∂is, policy did not develop solely by the carriers of knowledge, but also by the shurat, who were the soldiers of God. This can be seen in the early revolts in Yemen and Îa∂ramawt, as well as in Oman.122 The policy of these early Iba∂is was perhaps to develop their doctrine outside Basra via the shurat, who were responsible for political organization, and also via the carriers of knowledge, who were responsible for holding and propagating doctrine. Proto-Iba∂i leaders ordered shurat to be vehicles for the support of the earliest revolts.123 While the shurat tended to appear in Arabia and Eastern Mesopotamia, in North Africa, revolution appears to have been under the control of the carriers of knowledge. It is clear that a wide debate was held among the leading Iba∂is at the end of the second/eighth and the beginning of the third/ninth centuries. This debate concerned the relationship between the centre at Basra and the periphery, and the adjustments needed when the balance of power shifted in favour of the periphery. Both masters of Basra, MaÌbub b. al-RaÌil and Harun b. al-Yaman, submitted the judgments to the Omani imam al-Muhanna b. Jayfar. This debate divided the Iba∂i followers in Yemen, Oman and Îa∂ramawt.124 MaÌbub, on the other hand, left Basra for Oman, where the Mashariqa were formed.125 Thus they succeeded in making a base for the imamate in Oman and North Africa.
122
On the shurat’s role, see Wilkinson, Imamate Tradition, 184. For this case, see the accounts that the Basran masters sent back to Îa∂ramawt’s and Yemen’s rebels, Abu Îamza al Mukhtar b. ¨Awf and Balj b. ¨Uqba, which discuss whether it is true or not that Imam al Julanda joined the revolution of ™alib al Îaqq in the Yemen. The Basran masters also sent to Oman a group of shurat to establish the first imamate of al Julanda b. Mas¨ud. In the first imamate of Âuhur in North Africa, ¨Abu ¨Ubayda recommended to the imamate Abu al Kha††ab ¨Abd al ¨Ala’ b. al SamÌ who was from the Yemen. 124 Al Salimi, TuÌfa, 1:158. 125 There are two claims regarding MaÌbub’s immigration to Oman. First, it is said that he joined his father in law al Rabi¨ when he left Basra (al Kharasini, Sa¨id b. AÌmad, Kitab Fawakih al {Ulum fi ta¨at al Îayy al Qayyum, 1:242). Second, it is said that he immigrated later, after his debate with Harun. The second claim is more widely accepted, since the debate between the Iba∂i leaders was in Basra at the time of Imam al Muhanna (226 37/841 51). Most likely, he went to Oman with al Rabi¨, and then returned to Basra, where he lived until the debate with Harun. 123
506
THEMES OF THE IBA∆I/OMANI SIYAR
The Yemen and the Îa∂ramawt The Yemen and the Îa∂ramawt adopted Iba∂i doctrine, which perhaps was already flourishing as early as the first/seventh century. There is little or no evidence of this in the siyar documents, neither in early works such as the sira of ¨Abd Allah b. YaÌya ™alib al-Îaqq (XXI), nor in the records of early missionaries. The Iba∂is were apparently never able to restore the imamate of Âuhur after the great Arabian revolt at the end of the Umayyad period; the imamate must have been destroyed with the revolt. Later, the Iba∂is of the Îa∂ramawt probably founded the Imamate in the valley (wadi) of Da¨wan.126 After the imamate was transformed from Âuhur to difa¨, Iba∂i revolutions occasionally occurred under the first Abbasids, when the Iba∂is were attacked by Ma¨an b. Za’ida, the governor of Yemen.127 There was also an imam, Sa¨id b. ¨Abd Allah, who succeeded ™alib al-Îaqq in the Yemen and the Îa∂ramawt, but the Iba∂is dismissed him and chose instead Khanbash. Thus Sa¨id b. ¨Abd Allah’s case was under consideration in Basra, and likewise the affair of the imamate of MuÌammad b. Abi ¨Affan in Oman. It is probable that both of these men were unacceptable to the Iba∂is in both places, because they may have been replaced by the leaders of Basra, and they were not elected by the people of Oman, the Yemen, and the Îa∂ramawt.128 Wilkinson proposes that the Iba∂i community of the Yemen and the Îa∂ramawt initially split from the Iba∂i centre at Basra during the time of ™alib al-Îaqq.129 Wilkinson’s suggestion regarding the schism among the Iba∂is in Basra (i.e. the affair between Harun and MaÌbub) is acceptable. When Abu al-Mu’arrij retreated from his Sha¨biya stance, the proto Iba∂i leaders were not satisfied with him. He had to go to Yemen and ask the Iba∂i people to abandon Sha¨biya doctrines and readopt Iba∂ism. Unfortunately, Abu al-Mu’arrij died before arriving in Qidam.130 Perhaps the 126
Wilkinson, Imamate Tradition, 151. ¨Abd al Baqi b. ¨Abd al Majid states that Ma¨an arrived in Yemen in 140/757 (or 142/759?) and killed more than fifteen thousand from the Îa∂ramawt, since many Iba∂is had remained after the battle of Qudayd (132/749); see ¨Abd al Baqi b. ¨Abd al Majid, Bahjat al zaman fi tarikh al Yaman, ed. MuÒ†afa Hijazi (Beirut 1966), 20. 128 Al Shammakhi, al Siyar, 1:83; al Salimi, TuÌfa, 1:112. I assume that there were two imams named Sa¨id b. ¨Abd Allah, and that both of them were elected as imams of difa¨ in Îa∂ramawt. The first was killed after defeating ™alib al Îaqq’s revolt in Îa∂ramawt. The second was the one whose case was under discussion by the leaders of Basra. 129 Wilkinson, Imamate Tradition, 151. 130 Al Siyar wa al jawabat, 2:25; Qidam is a place in the Yemen; see ¨Abd Allah 127
507
THEMES OF THE IBA∆I/OMANI SIYAR
Îa∂ramis collaborated with the Omani imamate during the dispute, rather than the Yemenis. This hypothesis is supported by al-Kulanda b. al-Julanda’s revolution. It is unknown whether he tried to support the Iba∂is or just occupy Yemen.131 Later, the Omani scholar Munir b. al-Nayyir informed Imam Ghassan b. ¨Abd Allah (192–207/808–23) about the Iba∂is of Khurasan and the Yemen.132 This leads us to suppose that the Omani imamate had gradually encouraged the Eastern Iba∂is to submit to their authority. Furthermore, Wilkinson believes it is possible that the Îa∂ramis opted for difa¨ rather than shira’, and that this policy of the Îa∂ramis made them more compliant with the Omani imamate.133 It seems that this alliance appeared when the Iba∂is of Basra were engaged in theological polemics, ultimately splitting the Iba∂is of the Yemen and the Îa∂ramawt at the beginning of the third/ninth century. This was the beginning of a continuing trend of the Îa∂ramis to follow the Omanis.134 The imamate of Oman occasionally tried to expand its authority into Southern Arabia ever since the Omanis installed the imamate of al-Julanda b. Mas¨ud, who made a treaty with the inhabitants of the island of Socotra.135 Additionally, they brought the territory of Mahra136 under Omani control, and took possession of Socotra (LIII). When MaÌbub’s family moved from Basra to Oman, the Iba∂i groups who followed his theological doctrines were drawn in turn to Oman, and they abandoned Basra. Wilkinson’s hypothesis, therefore, may well be correct, since it links the concept of the state (miÒr) with the Iba∂i circumstances in the Îa∂ramawt, although in Islamic legislation the state should ideally be the dar al-islam itself. Since there was no geographical obstacle, such as a sea, between Oman and Îa∂ramawt, they were regarded as a single, distinct miÒr. This is the case also because, as al-Mas¨udi points out, the majority of people in the Îa∂ramawt in 332/943 were Iba∂is.137 Indeed, Iba∂is existed in the b. ¨Abd al ¨Aziz al Bakri, Mu¨jam ma ista¨jam, ed. MuÒ†afa al Saqqa (Cairo, 1368/ 1949), 3:1052. 131 Al Barradi, al Jawahir al muntaqat, 170. 132 Al Siyar wa al jawabat, 1:247. 133 Wilkinson, The Imamate Tradition of Oman , 157. 134 Al Salimi, TuÌfa, 1:158. 135 Al Kindi, Abu Bakr AÌmad b. ¨Abd Allah, al MuÒannaf (MNHC, Muscat, 1979 84), 11:145. 136 EI2, s.v. ‘Mahra’. 137 Al Mas¨udi, ¨Ali b. al Îusayn, Muruj al dhahab, edited and translated by C.B. de Meynard and P. de Courteille (Paris 1861 77). 6:67. 508
THEMES OF THE IBA∆I/OMANI SIYAR
Yemen and the Îa∂ramawt until the ninth/fifteenth century.138 The first shari imam in Îa∂ramawt, in the fifth/eleventh century, was Abu IsÌaq al-Îa∂rami. He tried to revolt against the ∑ulayhids. This was the last recorded revolt by the Iba∂is of the Îa∂ramawt.139 As alÎamid notes, the ∑ulayhids occupied the Yemen and the Îa∂ramawt; their occupation destroyed Iba∂ism, and consequently the Iba∂is in the Îa∂ramawt were dispersed.140 Some details are given in the interesting letter sent by the people of the Îa∂ramawt to Imam al-∑alt b. Malik (237–72/851–86) and MaÌbub b. al-RaÌil, in response to the emergence of a schism with their imam of difa¨, AÌmad b. Sulayman. They wanted to depose their imam (sira L), so they submitted their case to the Omani imam and the ¨ulama’. Abu ¨Abd Allah MuÌammad b. MaÌbub replied, blaming the Îa∂ramis themselves, and advising them to follow the Omanis’ policy regarding the imamate. According to Abu alMu’thir’s sira, there was an imam of the Îa∂ramawt named Sulayman b. ¨Abd al-¨Aziz, who was possibly the father of AÌmad b. Sulayman.141 In the following period, during the fourth/tenth century, when the Omanis were in an interregnum, the Iba∂i leaders of the Îa∂ramawt sent a message to Abu al-Îawari (LXXXIV) in order to persuade him to approve the killing of Julanda’s family during the first imamate.142 The Julandanis had brought about a debate not only among the Omanis but also among the Îa∂ramis. Tribal factionalism may have been a subject of discussion among both Omanis and Îa∂ramis, since the Julandanis were originally a Yemeni tribe;143 the Îa∂ramis were still concerned about their destiny. But from the middle of the fourth/tenth century onward, the balance of power changed in the Îa∂ramawt due to the situation in Oman. This led to the development of a number of issues, three of which were addressed by Abu al-Îasan al-Bisyawi, namely the acceptance of gifts from tyrannical rulers, the impression that most of the Omani leaders of the time were oppressors, and legislation concerning the Îa∂ramis’ inquiries about the issue of pawning property. The Al Îamid, Tarikh Îa∂ramawt (Jeddah, n.d.), 1:272. See Abu IsÌaq Ibrahim b. Qays al Îa∂rami’s works MukhtaÒar al khiÒal (Muscat 1984) and Diwan Sayf al Naqqad (Kuwait, n.d.). 140 Al Îamid, Tarikh Îa∂ramawt, 2:403. 141 Al Siyar wa al jawabat, 1:50. 142 Ibid., 1:343. 143 Al ¨Awtabi, Ansab, 1:246; Wilkinson, ‘The Julanda of Oman’, Journal of Oman Studies 1 (1975), 97 108. 138 139
509
THEMES OF THE IBA∆I/OMANI SIYAR
documents reveal interesting aspects of the Iba∂is’ position in the Îa∂ramawt at the end of the fifth/eleventh century; these were given by Abu Zakariya YaÌya b. Sa¨id (d. 472/1079) in response to the need for an up-to-date policy for the people. Abu Zakariya’s epistle (CVIII) reveals the key points as he saw them: Your message has mentioned: 1) your experience of oppression and tyranny, 2) the dominion of the people of doubt over the people of certainty, and 3) the rulers are tyrants and they demand fees on the property of the orphans.
Abu Zakariya’s response suggests that the Hadramis should practice taqiya (dissimilation), and he describes its laws and types: obligatory, permissible, and impermissible. He also recommends the use of metaphorical language (wa-‘alaykum bi-ma‘ari∂ al-kalam) for rhetorical purposes. It should be borne in mind that this document is a letter, written during the time of Abu IsÌaq al-Îa∂rami’s revolution in the Îa∂ramawt. Their revolt was suppressed by the ∑ulayhids. The recipients of this letter were al-Nu¨man’s brothers AÌmad and MuÌammad, who were probably Da¨ar princes in the city of Shibam.144 Linking AÌmad b. al-Nu¨man with the last Da¨ar amir Rashid b. al-Nu¨man, who was killed in 605/1231, it is possible that AÌmad was the grandfather of Rashid. Abu IsÌaq A†fiyyash proposes that the Iba∂is in Yemen, after the debate of MaÌbub and Harun, gradually began to follow the Zaydiya,145 but the end of the Iba∂is in the Îa∂ramawt is still obscure. The correlation with the Yemenis seems to be indistinct, since the Omanis had lost contact with the Yemenis. But the Omani imams thus far had used Yemen as a base throughout their conflict, starting from the first half of the second/ eighth century when al-Julanda b. al-Kulanda fought against the Abbasids.146 Ibn al-Athir further notes that during the second imamate, Hafs b. Rashid, in his war with the Buyids, fled to the Yemen.147 North Africa From this brief discussion, it appears that the important relationship between North Africa and Oman began after MaÌbub left Al Îamid, Tarikh Îa∂ramawt, 2:423. Al Salimi, TuÌfa, 1:157. 146 Al Barradi, Abu Qasim Ibrahim, Al Jawahir al muntaqa fi ma akhalla bihi kitab al ™abaqat, 170. 147 Ibn al Athir ¨Izz al Din, Al Kamil fi al Tarikh, edited by ¨Abd al Wahhab al Najjar (Cairo 1348/1930 1349/1931), 7:58. 144 145
510
THEMES OF THE IBA∆I/OMANI SIYAR
Basra and after the establishment of the imamate of Âuhur in Oman and North Africa. The essential aspect of this era of Iba∂ism is the rise of new schools, in both fiqh and theology, which have since been known as the Mashariqa (the Eastern school in Oman and the Îa∂ramawt) and the Maghariba (the Western school in North Africa). Gradually, an altercation between the two new Iba∂i factions was initiated by the Rustamid imam AflaÌ b. ¨Abd al-Wahhab (208–58/823–71) regarding the issue of the creation of the Qur’an. The Iba∂i schools probably addressed this matter after the miÌna. Imam AflaÌ (LVI) adopted the Mu¨tazili view, and sent his monograph on the subject to Oman in order to refute the other party’s view. At this time, the debate about the createdness of the Qur’an was in its early stages in Oman, and no specific doctrine had been agreed upon. The imamates of Oman and North Africa coexisted, but it is interesting to find in this letter that the Rustamid imam did not communicate directly with Imam Muhanna in Oman; rather, he intended to make his theological position available to all Omanis. This material has not preserved any exchange between those imams before the end of both imamates at the end of the third/ninth century. Abu ¨Abd Allah MuÌammad b. MaÌbub, on behalf of Imam al∑alt b. Malik, replied to the people of North Africa (LI) regarding their question about imamate policy. This important document reflects the real development of Iba∂i political views after the centre at Basra dispersed. Abu ¨Abd Allah MuÌammad b. MaÌbub influenced the legislative policy of the imamate since he had been involved in its foundation. The epistle replies to four issues concerning the imamate: 1) the imam’s authority, 2) the state’s boundaries, 3) the election to the imamate, and 4) the right of the people to elect an imam if they are a minority. Abu ¨Abd Allah uses the state boundaries to determine the payment of taxes and alms, whereas he had been advised that the zakat should not be collected until the authority holds sway over all the people and is able to protect them from oppression and aggression. If the state does not satisfy these conditions, the people should revolt, because the rulers would be considered tyrants. Let us now look at this text in the light of the issue of collaboration. At first sight, it may seem that the imamate of the Rustamids still existed, though the sender was perhaps from Jabal Naffusa or Tripoli rather than Tahart. It might be asked why those people did not try to contact the Rustamid imams or the scholars. But there were several divisions in Jabal Naffusa, namely the Nukkariya, the 511
THEMES OF THE IBA∆I/OMANI SIYAR
Naffathiya, and the Khalafiya,148 who were unsure as to whether or not they should follow the Rustamid imams. The other relevant text here is the sira to Imam al-∑alt and the people of Oman. Its purpose is mainly to call the people of Oman to assemble. Internal evidence shows that it was written approximately at the beginning of the dispute concerning the dismissal of Imam al∑alt from office. The sira reminds them of the deniers (nukkar) who rejected the imamate of ¨Abd al-Wahhab. It compares them with the Julandanis, their competitors over the state. Also, the letter preserves interesting records about the end of the Iba∂i schism in Basra that was led by ¨Abd Allah b. ¨Abd al-¨Aziz, Shu¨ayb b. al-Ma¨ruf, and ¨Abd Allah b. Yazid al-Fazari. Both Iba∂i imamates collapsed at the end of the third/ninth century, so this naturally led to the reform of the institution of the imamate. In this era, their policies became dissimilar. The Omani policy frequently attempted to restore the Imamate. In contrast, North Africa adopted the policy of concealment (kitman).149 The inference from this is that the birth of the Mashariqa and the Maghariba was influenced by MaÌbub b. al-RaÌil. During the formative period of these Iba∂i schools, the siyar do not provide the necessary details to elaborate on their respective doctrinal systems, neither in theology nor fiqh.150 The various Islamic schools of law, in which such diversity of doctrine was crystallized, are different yet inseparable aspects of the same unity. In this respect, the Iba∂i schools’ endeavour was not to confine themselves to their particular local areas, but rather to expand extensively in legal practice. We may assert that the emergence of different constitutional policies in local Iba∂i areas is significant enough to make a distinction between the two groups, since the Mashariqa gave fatwas based on ra’y151 (opinion), and their productions were with their own opinions (athar), while the Maghariba controlled ra’y in their legal formulations ( fiqh). North African lists name the compilations of the Mashariqa, which include Omani siyar.152 It can be seen in these that the influ148
T. Lewicki, ‘Les Subdivisions de l’Iba∂iyya’, Journal of Islamic Studies 9 (1958), 71; Ennami, Studies in Iba∂ism, 249. 149 EI2, s.v. ‘Îalka’. 150 It is possible to trace the relationship between their scholars in both schools; see Oman in History (London 1995), 235 48. 151 EI2, s.v. ‘Ra’y’. 152 Al Barradi, al Jawahir al muntaqat, 281; Ennami, ‘A Description of New Iba∂i Manuscripts from North Africa’, JSS 15 (1970): 63 87; van Ess, ‘Unter suchungen zu einiger Iba∂itischen Handschriften’, ZDMG 126 (1976), 25 63. 512
THEMES OF THE IBA∆I/OMANI SIYAR
ence of the Rustaq school was greater on the North African school than that of Nizwa. This state of affairs started from the time of Ibn Baraka, al-Bisyawi, al-¨Awtabi, Abu Bakr al-Kindi, and Ibn al-NaÂar. This listing does not reflect the chain of transmission (isnad ) of this tradition nor does it demonstrate the authenticity of their compilations in North Africa. It merely illustrates their doctrinal affiliation in theology and legal practice. From the seventh/thirteenth century onwards, for four centuries during the dark ages of the Nabhanis, the effect of the relationship on both sides is obscure, and is not reflected in the siyar literature. Much more significant in this doctrinal collaboration is the change of outlook of the siyar at the beginning of the eleventh/seventeenth century to reveal a direct relationship between both schools. So far, the conflict among the Iba∂is of North Africa, that is, between the people of Naffusa and the Banu MuÒ¨ab (the Mazabis), led them to refer in their problems to the Iba∂is in Oman.153 At the same time, the Omanis were able to restore the imamate of the Ya¨ariba dynasty, and they advised the North Africans that they ought to follow their example and re-establish their imamate. The Iba∂is of the North African community, being scattered widely in isolated pockets along the northern border of the Sahara desert, submitted to the Omanis, who developed the institution of the imamate and became the centre of the Iba∂is. The birth of the Ya¨ariba drove the foreign Arab dynasties out of Oman, and the Portuguese out of the Omani and East African coasts, and started the modern history of the imamate. Summary The history of the Iba∂i schools going back to the early second/ eighth century has been considered from the time when the movement started from Basra by Abu Sha¨tha¨ Jabir b. Zayd, through Abu ¨Ubayda, until MaÌbub. With respect to the transmission of knowledge from Basra to Oman, this developed gradually, starting with the Iba∂i leaders themselves, then the shurat, and finally the carriers of knowledge (Ìamalat al-¨ilm). Al-Kharasini’s accounts are correct in that the stages of Iba∂i doctrine developed according to the Rasibis (¨Abd Allah b. Wahb al-Rasibi), the Iba∂is (¨Abd Allah b. Iba∂) and the MaÌbubis (MaÌbub b. al-Rahil).154 After the decline in the cen153
See al Salimi, ‘Identifying the (Iba∂i/Omani) Siyar’, section 4B. Forthcom
ing. 154
Al Kharasini, Fawakih, 3:135. 513
THEMES OF THE IBA∆I/OMANI SIYAR
trality of Basra, the Omanis developed political relations with the Yemenis and the Îa∂ramis, and tried to expand the role of the imamate and its influence under the concept of miÒr, in order to form the ideal state. The doctrines of Iba∂i teachings developed in Oman and North Africa, and continued through collaboration and doctrinal works which were exchanged between the two sides, thus forming a unique Islamic school.
514
THE POLYGENESIS OF THE NEO-ARABIC DIALECTS AHMAD AL-JALLAD HARVARD UNIVERSITY
Abstract This paper presents a framework within which the study of the devel opment of neo Arabic may be carried out. The present analysis inves tigates relative and demonstrative pronouns in peripheral neo Arabic in light of NW Semitic isoglosses and suggests that the contemporary dialects originated in two Old Arabic dialect clusters. However, the distribution of the relative and demonstrative features in most neo Arabic dialects indicates that there has been significant historical con tact between the proposed original clusters. Thus, a conventional tree diagram is not capable of describing their development. This analysis suggests that the neo Arabic dialects should be regarded as sedimen tary structures, containing both genetic and koine features. In light of this, the author proposes that the diachronic study of neo Arabic should be divided into two modules: the first is the study of the de velopment of innovations found in neo Arabic, regardless of dialect, from proto Central Semitic, and the second is the study of the spread of these features throughout the neo Arabic dialects through the mechanisms of contact presented here.
1. Introduction1 The development of neo-Arabic (NA)2 remains a topic of much interest to Arabicists and Semiticists alike; however, despite over a century of scholarly attention, and the equivalent in theoretical output, 1 This article is a revision of a paper of the same title delivered at NACAL 34 in Seattle on March 18, 2006. I am grateful for the feedback from the participants, especially Professor Wolfhart Heinrichs. All mistakes are, of course, my own. 2 In this paper, the distinction between neo Arabic and Classical Arabic is typo logical rather than chronological. For now, neo Arabic includes the modern dialects while Classical Arabic is specifically what is described by the Arab grammarians; the language of the eighth century CE and prior. Of course this is a working definition and as such subject to refinement as the argument progresses. The use of neo vs Classical, rather than old vs new, removes the implication of a historical linear de velopment from one language to the others.
515
THE POLYGENESIS OF THE NEO-ARABIC DIALECTS
there is very little consensus on a starting point. Scholars have offered several theories to account for the points of difference between the contemporary Arabic dialects and the earliest attested varieties. These approaches have been traditionally divided along the lines of monogenesis from a language qualified more or less as Classical Arabic (CA) or polygenesis and convergence.3 This paper will construct a polygenetic model of development for the neo-Arabic dialects. The current view is qualitatively different from its predecessors. Rather than providing a general account for the development of NA, the argument proposed here seeks to establish a general framework within which investigation into the NA language type can be carried out. The conclusions of this paper are not meant to act as a final verdict on the matter, but are instead a starting point from which detailed inquiry into the development of NA can take place. This paper is organized into five sections, including the introduction. Section 2 will categorize the many viewpoints expressed in the scholarly literature on the subject and, in light of this discussion, define the goals of the current effort. Section 3 will formulate a new methodology with which to approach the study of the development of Arabic. The conclusion of this analysis will be the basis of a new approach, in section 4, to construct a model of the general evolution of neo-Arabic. Finally, section 5 will summarize the results, and suggest future avenues for research. 2. Theories on Development of Arabic Any question having to do with diachronic development essentially involves an understanding of how the more recent form acquired the features that distinguish it from its ancestral form. This reasoning faces several obstacles when applied to the development of NA, since we lack a full understanding of an ancestral form and have a myriad of descendent forms. In order to successfully pursue an answer, we need to establish a starting point (ancestral form[s]) and, to deal with the lack of attested transitional stages, to identify subsequently a mechanism of change. By mechanism of change, it is meant that we should identify the forces of language evolution which are to be held accountable for generating the dialectal variation in question. The following discussion will review the two traditional approaches to this question, focusing on how they account for the two aforementioned conditions. 3
Versteegh 1997: 104. 516
THE POLYGENESIS OF THE NEO-ARABIC DIALECTS
2.1 Monogenesis The monogenetic view anchors itself in an assumed starting point, which — without any evidence to the contrary — takes the form of CA. This premise, by its very nature, shapes the rest of the analysis. If one assumes that CA is the direct ancestor of the NA dialects, then any theory constructed thereafter would draw its explanatory power from its proposed mechanism of change; that is, how well it can account for the points of difference between CA and NA. This paradigm has produced a variety of explanations to reconcile CA and NA. Perhaps the most notable is Ferguson’s notion of the military koine.4 For Ferguson, CA was more or less a single entity. He hypothesized that there developed an Arabic koine in the military camps throughout the period of the Arab conquests, and from this koine, the modern dialects evolved.5 Features shared by the NA dialects against CA were explained through the process of koineization. Although Ferguson’s theory was riddled with explanatory problems, he did take into account an important issue, which is the theme of this paper. He noticed that the dialectal variation, which undoubtedly existed on the Arabian Peninsula prior to the Arab expansion, must have continued after the conquests. Although he does not investigate the nature of this variation, it is an important observation nonetheless as it recognizes a possible condition for polygenesis. Ferguson, however, does not follow up with this thinking. Instead, having acknowledged the existence of variation, he restricts that variation to the classical norm, and uses the notion of levelling by means of koineization to account for the differences between NA and the various pre-Islamic dialects. For Ferguson, it is the variation in CA which led to uniformity in certain features of NA. 2.2 An Unattested Ancestor A striking point of difference between CA and NA is the presence of morphological case in the former and its absence in the latter. This fact has led some scholars to bypass the default CA starting point and present CA as the result of the same evolutionary forces responsible for NA. Corriente, an example par-excellence of this thinking, suggested the existence of a case-less variety of Arabic spoken on the 4
Ferguson (1959) outlines fourteen features the urban NA dialects share against CA. These features suggest to Ferguson that the Arabic dialects evolved from a koine centered in urban areas. 5 Ibid, 1959: 4. 517
THE POLYGENESIS OF THE NEO-ARABIC DIALECTS
border of Arabia and the Levant.6 This variety would have gained prestige following the Arab conquests and perhaps could be the ancestor of the modern Arabic dialects, rather than CA.7 I will return to the geographic implications of Corriente’s hypothesis in the next section; it is Corriente’s linguistic observations which are germane now. Corriente’s most significant suggestion in terms of the goals of this paper is that a different ‘dialect’ of old Arabic, absent in the accounts of the Arab grammarians, could have contributed to the development of NA. Corriente reconstructs this hypothetical dialect without the points of disagreement between NA and CA, especially concerning case. By doing so, his theory draws its explanatory power from its starting point. There is no need to rely on a mechanism of change to explain the absence of CA features in NA — they never existed to begin with. Owens’s ideas, as expressed in his most recent work (2006), are related to this second category of thinking. He not only bypasses a CA starting point, but reconstructs proto-Arabic primarily on evidence drawn from NA. Proto-Arabic becomes an abstraction able to handle supposedly irreconcilable points between CA and NA. Owens’s thinking is treated here as a stronger version of the polygenetic view, which marginalizes CA. This view is of special conceptual importance and will be discussed in more detail in 2.4. There have been many more approaches to this question, but the two described above capture the general polarization of those attempts — monogenesis vs. polygenesis, and normal language change vs. significant language contact. Although differing in their explanatory tools, the two opposing perspectives are similar in that they describe a binary opposition, with CA on one side and NA on the other. 2.3 A Note on Starting Points 2.3.1 Pre-Islamic Arabic Before entering into the main discussion of this paper, a brief treatment of two central themes in Arabic historical linguistics, namely, the accounts of the Arab grammarians and the role of case, is in order. 6 Corriente argues that the case system of CA had a relatively low functional yield (1971). He also points out inconsistencies in the usage of case in middle Arabic texts (1975). Owens follows Corriente and reconstructs a case less ancestor for NA. 7 Corriente (1976:88).
518
THE POLYGENESIS OF THE NEO-ARABIC DIALECTS
Following Owens, it is impossible to make any definitive statements about the linguistic landscape of Arabia prior to, and shortly after, the Arab conquests. Although vernacular Arabic of those times is described by the early Arab grammarians, their accounts cannot be considered conclusive. The reasons for this are plenty and generally beyond the scope of this paper. One significant remark by Owens, however, merits repetition and that is the approach taken by the Arab grammarians, most notably Sibawayh, in their description of Arabic. Sibawayh dealt with variation in the Arabic of his day not through dispassionate description, but by fitting it into his linguistic thinking.8 The non-occurrence of particular forms is apparently confirmed not in terms of what Sibawaih observed or tested, but rather in terms of the violation of his own general rule. What should by now be clear is that there is no pure ‘data’ to be found in Sibawaih. Everything he observes and writes about is filtered through his own grammatical thinking (Owens 1998: 65).
Clearly, any statements about vernacular Arabic of the eighth century CE based on the accounts of the grammarians are speculative. It must be stressed that this view of the classical accounts does not imply one should work on the assumption that the nature of NA’s ancestor is completely unknown, and pursue its reconstruction without a general guide. The fact that NA belongs to the Semitic language family should not be ignored, and reconstructions of proto-NA should take place within what is known of proto-Semitic, and subsequent sub-groupings, rather than be restricted to CA norms. Owens has ignored this fact to some extent as evident in his reconstruction of the proto-NA prefix conjugation lacking modal endings.9 This reconstruction is to be rejected because it contradicts the well-known fact that Proto-Central Semitic, the sub-grouping to which NA be8
This type of understanding is not exclusively a component of the Arabic grammatical tradition. The Hindu grammarians, for instance, restricted the use of the perfect tense in Sanskrit to the narration of facts not witnessed first hand. This description finds no support in the attested data (Whitney 1896: §821a). 9 Owens reconstructs three verbal types for the imperfect in pre Islamic times based on rules of epenthesis. He asserts that mode endings could not have existed on his reconstructed Hijazi and Nigerian types because the addition of a final mode vowel would delete a stem high vowel, which is not mentioned in the old Arabic sources. Owens’s no deletion, no epenthesis type, however, is not subject to this. Concerning a reconstruction of modal endings on that type, he states that ‘they could once have existed on an ancestral variety, but there is no historical linguistic trace allowing such an inference to be made’ (2006: 195 6). 519
THE POLYGENESIS OF THE NEO-ARABIC DIALECTS
longs, inflected the imperfect verb (prefix conjugation) with mood endings, rather than because it contradicts the accounts of the Arab grammarians. 2.3.2 The Position of Case Case inflection potentially belongs to the category of features that must be considered part of proto-NA due to their presence in ProtoCentral Semitic. Nevertheless, this category has uniformly disappeared in the developmental course of most Central Semitic languages. Case is given a special place in most approaches to the NA question, regardless of their starting point. This is despite recognition of the fact that the removal of the case and mood endings cannot transform CA into NA.10 The lack of case-endings in NA does not imply that they collectively descended from a case-less ancestor. Most of the Romance languages are case-less despite descending from Latin. The loss of case was not a single event that marked the transition from Latin to Romance. The break-down of the case system is observed first in spoken Latin with the evolution of analytical devices overlapping the case-system. Then, certain cases began to take on expanded usages until the overlap forced the extinction of redundant categories. The case-system of Classical Latin broke down into a bipartite distinction of nominative and oblique in late spoken Latin, a process which was gradual and not uniform.11 This distinction survived in French and Provençal until the twelfth or thirteenth century CE. The point is that the break-down of the case-system in Latin was a slow process which continued well into the Romance stage. The final result, however, is the lack of case in the descendent languages. If it were not for the attestation of the different stages of developments in the Romance languages, it would be impossible to discern the stages of its corrosion. Given that case was undoubtedly a feature of CS, the same process must have occurred at some point in the history of the NA dialects. The question of whether or not case was present in the Arabic of the seventh and eighth centuries CE is entirely separate from the question as to whether or not the ancestor(s) of NA possessed case. In fact, there is no direct evidence to suspect that the 10
Versteegh 1984: 19. Case in late spoken Latin was not expressed in a class nouns: compare nom: rósa and obl: rósa. Case is retained in the o class declension, however. There the oblique form of late spoken Latin is inherited from the ablative of Classical Latin: compare late spoken Latin oblique ánno to Classical Latin dative/ablative anno, in contrast to the accusative form annum (Penny 2002: 114). 11
520
THE POLYGENESIS OF THE NEO-ARABIC DIALECTS
ancestor(s) of NA lacked case. If the case-system was on the way out, it is not surprising to find it absent in the Arabic dialects by the time they are first attested. Even if the Arabic dialects went through a period in which they showed limited inflection, this might have been seen as sub-standard by the Arab grammarians, and might have escaped their description. Therefore, even if one does not view CA as the direct ancestor of NA, there are still no grounds to dismiss the presence of case in the elusive ancestor(s) of NA. 2.4 Goals Owens recently stressed the importance of incorporating the NA dialects in understanding the linguistic history of Arabic.12 One of the central questions addressed in this paper is how this can happen. Owens gives CA and NA equal footing in reconstructing protoArabic, the ancestor of both varieties in his view (2006:11). This method is expressed clearly in his reconstruction of the first person singular ending of the perfect. The reflex of this morpheme in most NA dialects is t. On this basis, Owens reconstructs an original *t and *tu.13 It is important to stop and consider the implications of this claim. If reconstruction of proto-Arabic were to be done without taking into consideration what is already known about antecedent Semitic forms, are we not then erroneously treating Arabic as a language isolate? Rather than reconstructing a single proto-from, Owens traces both of the main modern variants back to the protostage, even though one form is clearly derivable from the other. The understanding of the development of NA will constantly be a compromise between the NA evidence and what is known so far about Central and Proto-Semitic.14 A fair approach should involve both sides, weighing features differently according to their diachronic depth. It is not useful to apply comparative method blindly, without taking into account the comparative Semitic evidence. It is obviously true that CA is older than NA in chronological terms, but based on comparative Semitic evidence, CA is clearly more archaic in several important linguistic respects. To give both varieties an equal role in the reconstruction of proto-Arabic is therefore a mistake. Owen’s methodology leads to the reconstruction of proto-forms which are strikingly similar to the modern dialects; consequently, 12
Owens 2006: 8 13 Ibid. 14 Owens admits to this fact in his epilogue (2006: 267). 13
521
THE POLYGENESIS OF THE NEO-ARABIC DIALECTS
all that can be shown, if one were to accept this approach, is that the variety called NA is far older, in chronological terms, than expected. If NA developed in the seventh century, paralleling the Arab conquests, then it makes little sense to consider the language of the seventh century the proto-form of the modern spoken dialects because it is essentially the same language. The real question is how the NA language type developed. Whether this occurred in the seventh century or the fifteenth is not important. The loss of case and modal inflection is observed in many Semitic languages, but the indicative augment bv, for instance, is unique to NA. Reconstructing bv, or even the augmentative category itself, to the protolanguage is not useful because it is still an anomaly at the broader Semitic level. Two central questions emerge from this discussion: (1) from which antecedent linguistic situation did the NA dialects develop, and (2) how can one account for features peculiar to NA specifically in their broader Semitic context? The second question is at least partly dependent upon answering the first. To answer the first question, two conditions, already described, must first be met: a starting point and a mechanism of change. Considering the substantive points of this discussion, the next section will formulate a new model with which to answer question 1. 3. A Reconstructive Approach This section will attempt to define the values for the two prerequisites established in the previous section, a starting point and the mechanism of change. Although little is currently known about the linguistic landscape of pre-Islamic Arabia, it is not productive to assume that, due to the lack of attested evidence, no other varieties of pre-Islamic Arabic have modern descendents. If the NA dialects descended from a cluster of pre-Islamic dialects, then the evidence should be available. What is needed is a methodology with which to sift through the modern dialects in search of useful evidence. Because the understanding of the mechanism of change is dependent upon the starting point, and because the starting point is hidden under a millennium of change the details of which are currently unknown, we are confronted with a catch 22. Without a methodology, it is impossible to determine which features in any dialect could point to an ancestral type and which features are innovative or the result of borrowing. 522
THE POLYGENESIS OF THE NEO-ARABIC DIALECTS
3.1 Dialect Divisions in Northwest Semitic Arabic is most closely related to the Northwest Semitic subgroup.15 Accordingly, features indicative of dialect boundaries there might be helpful in discovering dialect variation in the ancestor of NA. Garr’s work on dialect variation in Syria-Palestine in the first half of the first millennium BCE will be important for this. The dialects in his study16 are divided into two groups according to several features, among them, their derivation of the relative and demonstrative pronouns. The relative and demonstrative pronouns are particularly suitable for analysis for several reasons: (1) they are a fundamental part of a language; therefore, their derivation must have occurred early on in the development of a specific dialect, (2) they are fairly stable and therefore there is little motivation for cross-dialectal borrowing, (3) their value can be determined as +/- *∂V, according to their derivation from the proto-West-Semitic relative-determinative *∂V.17 3.1.1 What Can Be Deduced? The values appear as follows: Table 1: REL DEMmasc.sing18
Byblian z zn
Phoenician Aramaic Samalian Moabite Hebrew ’s z z ’s ’sr z znh znh *zh19 zh
The first group, consisting of Hebrew, Moabite, and Phoenician, derives its demonstrative pronoun from the unmarked proto-Semitic relative-determinative *∂V, while the other, consisting of Aramaic, 15
For more on the classification of Semitic languages see Huehnergard 2004. These dialects are Phoenician, both standard and Byblian, Aramaic, Samalian, Moabite, and Hebrew (north and south). Garr treats Samalian as a sepa rate dialect island of Aramaic and Byblian of Phoenician. Both of these dialects ex hibit irregularities which exclude them from the general category of Aramaic and Phoenician. Garr’s dialect continuum splits the dialects into two clusters with Phoenician, Moabite, and Hebrew on the one end and Aramaic and Samalian on the other. 17 See Huehnergard (2006) for a discussion of the proto West Semitic relative determinative pronoun. 18 The masculine singular tends to be the unmarked form and therefore the suitable candidate for analysis here. 19 The masculine singular form of the demonstrative is unattested for Moabite, but it can be reconstructed as *zh from the Hebrew and Phoenician evidence. 16
523
THE POLYGENESIS OF THE NEO-ARABIC DIALECTS
Byblian, and Samalian, derives its relative pronoun from the unmarked *∂V.20 If the scope of the analysis is expanded to include the structure of the demonstrative compared to the relative, an interesting pattern comes forth. The dialects which derive their relative from the unmarked *∂V form the demonstrative from the *∂V plus a nasal. On the other hand, dialects which derive their relative from some other source derive their demonstrative directly from the unmarked *∂V. The notion of the unmarked *∂V is significant because it indicates an early divide in the primary use of the element. Pronouns formed by modifications to the *∂V can be viewed as secondary to its original function. It is apparent that the dialects divided themselves early on according to the primary use of the *dV, whether for deictic or relativization purposes. This split parallels an evolutionary split in the development of these two dialect clusters, and qualifies the derivational pattern of the relative and demonstrative as a marker for early dialect boundaries.21 3.2 Evidence for Dialectal Split in Pre-Islamic Arabic Before applying this methodology, a sample must be selected from the many NA dialects. On a theoretical basis, the sample chosen for analysis should not affect the results since the relative and demonstratives can have only two variable forms. However, for practical use of the sample in incorporating new variables at a future point, dialects isolated from the Arabophone world are particularly suitable. The following dialects have been selected due to their geographic distance, isolation, and/or easy chronological establishment: pre-Hilalian Moroccan, Chadian Arabic, Maltese, Cypriot Maronite Arabic, Anatolian Arabic, and Southern Najdi Arabic.22 Due to the 20
Garr 1985: 82, 85. The position of Byblian is problematic. It appears to be an amalgamation of features from both ends of the continuum. Garr understands Byblian to be an ar chaic dialect which did not participate in all the innovations common in Phoenician (Garr 1985: 230). The general thesis formulated in this work may also help in understanding this dialect’s unique position. 22 The pre Hilalian Moroccan dialects were brought to the Maghreb with the initial Arab conquests (Hearth 2002: 2). Arabic was brought to the Sudan in the fourteenth century (Kaye 1976: 1). Chadian Arabic can be seen as a conservative off shoot of Sudanic Arabic, which is isolated from the arabophone world. Maltese is the only NA variety to be considered a language in the socio political sense today. Arabic was brought to Malta as early as 870 CE and isolated from the Arabophone world in 1054 CE (Versteegh 1997: 210). Cypriot Maronite Arabic was brought to Cyprus from the Syrian coast in the twelfth century CE; it preserves many archaic 21
524
THE POLYGENESIS OF THE NEO-ARABIC DIALECTS
aforementioned factors, statements made about these dialects can also be accompanied by a relative date of settlement and/or isolation. The results are as follows: Table 2:
REL DEM GEN
PHM d /lli *∂V+ha dyal/d
CHD ’al *∂V ha hana
MLT li *∂V ha ta
CMA si *∂V+ha tel23
ANA ’al/illi *∂V +ha ∂el
SNJ alli/illi *∂V+/ ha Ìagg
The values for the relative pronoun are two: one derived from a substantive or definite article and one derived from the relative-determinative *∂V. The demonstrative pronoun also comes in two forms: one derived directly from the *∂V, and one accompanied by a mandatory deictic particle: ha. A third element is drawn into this analysis: the genitive. If the genitive is analysed on the basis of its derivation, it can be understood in terms of particles derived from the *∂V or from a substantive.24 Since the genitive is sometimes a frozen compound that includes the relative pronoun, it can be useful in detecting the relative pronoun’s original form in the dialect under investigation. 3.2.1 Correlations The demonstrative, where derived directly from the relative-determinative *∂V, co-occurs only with genitive particles derived from substantives. However, where the demonstrative has merged with the haparticle, it can co-occur with a genitive particle that consists of the frozen relative + dative construction. That is, genitive particles based features which have disappeared in the contemporary Levantine dialects (ibid, 1997: 212). The Anatolian Arabic dialects are part of the qeltu Mesopotamian clus ter, but following the Seljuk conquest of Anatolia became isolated (ibid, 1997: 213). The Bedouin dialects of the southern Najd of course never left the Arabian Peninsula and therefore are indicative of some of the innovations which occurred within the Arabian Peninsula (Ingham 1994). Their datable isolation from the Arabophone world reduces the likelihood of cross dialectal borrowing, and allows us to date innovative constructions. Although this point may not seem critical for the current argument, as the pronouns under investigation vary along the two men tioned forms throughout the Arabic dialects, this method will prove valuable when, in the future, more variables are brought under investigation. 23 The Cypriot genitive particle is cognate to the Anatolian one; devoicing can account for the difference: tel <*del< *∂el. 24 See Rubin (2005) for a discussion on the grammaticalization of genitive par ticles in neo Arabic. 525
THE POLYGENESIS OF THE NEO-ARABIC DIALECTS
on the relative pronoun ∂V never occur with -ha demonstratives. This brings us finally to the situation of the relative pronoun. The relative pronoun among most of the surveyed dialects is derived from the definite article, which may indicate that leveling occurred quite early on. The only case where the relative is not derived from the article, it co-occurs with the *∂V genitive and the ha+*∂V demonstrative; many Yemenite dialects parallel this.25 These findings match the dialectal split in Northwest Semitic. Dialects that primarily derive their relative pronoun from the unmarked *∂V form the demonstrative by means of a modification to the *∂V, perhaps to distinguish it from the relative. This modification takes the form of a prefixed ha deictic particle, mirroring, in terms of function, the nasal found in the Northwest Semitic dialects. Thus, the presence of an inalienable deictic particle may indicate that the relative function of *∂V was original in those dialects. Also, the data show that relative-based genitive particles are found in dialects which derive their relative pronoun from the unmarked *∂V. This cluster of developments is represented in Figure 1 as TYPE I. The second type, which derives the demonstrative from the *∂V and forms its relative and genitive by means of a substantive or modification to the *∂V, is represented in Figure 1 below as TYPE II. The demonstrative of this type can appear without the deictic prefix. The aforementioned correlations lead to the following reconstruction of proto-Arabic relative, demonstrative, and genitive particles. Figure 126
alla∂i?
These are: dem ha∂a and rel ∂i, ∂a (Kaye 1997: 287, 289). REL: relative pronoun; DAT: dative l ; GEN: genitive particle; DEM: proxi mal masculine singular demonstrative pronoun. 25
26
526
THE POLYGENESIS OF THE NEO-ARABIC DIALECTS
According to this reconstruction, the NA dialects stem from two original types as evidenced by their derivation of relative/demonstrative/genitive features from the relative-determinative *∂V.27 It is important to note that the existence of an analytical genitive at this early stage does not imply that the case system was defunct. Latin had many analytical devices that co-existed with its case system for centuries.28 The use of the *∂V as a demonstrative is a secondary derivation; it is distinguished from the relative by a modification — the prefixed ha-particle. In TYPE II, the demonstrative is formed by means of the unmarked *∂V. The relative here most likely derived from the definite article.29 The addition of the relative + demonstrative to this core is a reflection of what is found in CA, and therefore indicated in grey.30 Lastly, TYPE II grammaticalizes a substantive as a genitive particle.31 3.3 The Geography It is appropriate now to revisit Corriente’s hypothesis. Corriente postulated that the pre-Islamic dialects spoken in central Arabia were different from those spoken nearer to the Fertile Crescent, specifically Nabataean.32 Corriente contrasts these dialects based on the as-
27
Language ‘type’ rather than ‘language’ is used here because there is not enough evidence to argue for the existence of two original, homogeneous entities. These two types might be qualified as dialect clusters. Until more research is con ducted, however, it can only be said that they differed in terms of the derivation of the discussed features. 28 Spoken Latin used several analytical devices which overlapped morphological case categories. The genitive could be represented synthetically via the genitive case or through the analytical construction de+ablative (Penny 2002: 115). 29 Hebrew grammaticalized a substantive, ’aser ‘place’, into a relativizer, restrict ing the *∂V to the demonstrative series. Although it has been suggested that the relative se is derived from the devoiced reflex of *∂V, Huehnergard argues convinc ingly that se is in fact a reduced form of ’aser (Huehnergard 2006). This is signifi cant as it demonstrates that the *∂V is primarily deictic, and the relative is derived from a substantive. This matches the case of Arabic. 30 The CA relative consists of three components: definite article (¨al) + asserv ative particle (la) + the determinative (∂i) (Huehnergard 1983: 579). 31 See Rubin for a thorough discussion on the grammaticalization of genitive particles (Rubin 2005). 32 From early times, the Arabic of Nabataea was recognized as distinct from the language of the Bedouin Arabs. Nabataean jargon was a term used for what was perceived as less than grammatical Arabic. The poet Ibn Burd (714 84 CE) was accused of using Nabataean jargon for his incorporation of the colloquial in his poetry (Cachia 35: 2002). 527
THE POLYGENESIS OF THE NEO-ARABIC DIALECTS
sumed presence or absence of a functional case system.33 Suppose that Corriente’s hypothesis is correct, and that there were two general types of Arabic separated by geographic barriers, but that their differences were not necessarily related to the presence of case inflection. Instead, they are distinguished by the split in the derivation of features from the *∂V. Based on the distribution of TYPE I and II features in the NA dialects, it is reasonable to reconstruct an original TYPE II dialect in the Arabian Peninsula and TYPE I in the area of Nabataea and the Syrian Desert. The fact that non-Muslim dialects of Arabic bear more similarities to TYPE I than TYPE II supports this geographic arrangement. The majority of Christian Arab tribes inhabited the northern regions of the Arabian Peninsula prior to Islam. It makes more sense that Christian populations learned Arabic from Christian Arabs who came to live in the same neighbourhoods and were a part of the same ethno-religious community.34 TYPE II varieties probably originate south of TYPE I in central Arabia. Of course, this reconstruction is based on the distribution of modern Najdi dialects which exhibit many characteristically TYPE II features and probably did not move into the region from elsewhere. This reconstruction is represented in Map 1. Map 1:
33
Corriente 1976: 88. Religious barriers have often prevented linguistic permeation. For example, Bedouinization of the Mesopotamian dialects did not affect non Muslim groups (Versteegh 1997: 157). See also Blau 1961. 34
528
THE POLYGENESIS OF THE NEO-ARABIC DIALECTS
4. Polygenesis 4.1 The Mechanism of Change The evidence suggests that the NA dialects descend from two original dialect types, distinguished by their derivation of relative, genitive, and demonstrative elements. These two dialects were geographically isolated prior to Islam, with TYPE I in northern Arabia on the fringes of the Fertile Crescent and TYPE II in the Arabian Peninsula proper. These basic observations can be considered a working starting point. With this condition fulfilled, attention can be given to the second prerequisite: the mechanism of change. After the Arab conquests, TYPE I and II dialects came into contact with each other. This socio-linguistic situation is ideal for koineization.35 There is a tendency to understand koineization as it had occurred for Greek. However, if we are to understand how past koineization took place in the case of Arabic, we should look to its behaviour among the Arabic dialects today. The Arabic dialects of the Sinai are currently undergoing koineization. It is a bottleneck where many Bedouin tribes must live in close contact with each other, in addition to sedentary populations. As a result of koineization, features considered characteristically sedentary have been absorbed by Bedouin populations there. Bedouin dialects are also responsible for spreading borrowed sedentary features among each other so that it is possible for a Bedouin dialect to acquire characteristically sedentary features from another Bedouin dialect rather than directly from a sedentary one.36 An example of this phenomenon is the distribution of bv, the augment used to mark the indicative mood on the imperfect verb. This characteristically ‘sedentary’ feature has been adopted by several Bedouin dialects. Not only has it penetrated Bedouin dialects in contact with sedentary areas, but as a result of subsequent koineization, it has also diffused throughout neighbouring Bedouin dialects as well.37 Variation in the derivation of the genitive particle in Najdi Arabic might also be analogous to this.38 The linguistic record indi35 Hock gives three conditions for koineization to occur: (1) contact between closely related languages or mutually intelligible dialects; (2) these varieties are con sidered equally prestigious; (3) the absence of a link language (1991: 485). 36 De Jong 2000: 593 4. 37 Ibid, 2000: 478. 38 The genitive particle, mal, perhaps of Iraqi origin, is employed in the north ern Najdi dialects (Ingham 1994: 58). Its presence in the northern Bedouin dialects
529
THE POLYGENESIS OF THE NEO-ARABIC DIALECTS
cates that Bedouin dialects have not only absorbed sedentary features but have likewise influenced the sedentary dialects with Bedouin features. The gilit and sedentary Libyan dialects are two such examples.39 Having established that the Bedouin dialects are capable of absorbing sedentary features and spreading them as they move, one must ask what is to prevent the Bedouin from picking up a sedentary feature from one region and, in turn, influencing a sedentary dialect of another region with the same feature. It is very possible to view the Bedouin as conductors through which linguistic features are passed from region to region. Let this phenomenon be called the Bedouin-Urban Continuum (BUC). If this view is adopted, then it is possible to consider many of the NA dialects in constant contact throughout their history, and consequently, perpetually subject to gradual koineization.40 Koineization is a dynamic process which is multi-directional and exists so long as contact exists. The effects of koineization are not uniform. Some dialects might adopt a certain set of features while others might not. This is illustrated by the uneven distribution of koine features in the Sinai dialects.41 Understanding the way koineization works is crucial to the effort of understanding the evolution of Arabic as a whole. Aside from a few isolated cases, the Arabic dialects represent a continuum of contact — all are some way or another influenced by another variety, regardless of time or circumstance. 4.2 A Note on Genetic Affiliation The ubiquity of the indicative augment bv suggests that it is an early development. The Bedouin are notorious for its absence in their is most likely the result of contact with sedentary populations. Southern Bedouin dialects employ Ìagg, which is an Arabian coinage. 39 Blanc 1964: 6; Versteegh 1997: 165. 40 The fact that modern Bedouin dialects exhibit sedentary features in their speech cannot be regarded as an idiosyncrasy; rather, it should be considered the general condition of the linguistic relationship between Bedouin and sedentary speakers of Arabic. Thus, we cannot look to points of disagreement between Bedouin and urban dialects to reconstruct a Bedouin type vs sedentary type (or in Ferguson’s case, a military koine). It is impossible to discern whether any features shared by Bedouin and urban dialects are indicative of their respective proto lan guages or if they result from past contact. This is especially true once we consider that a Bedouin dialect can pass borrowed urban features on to other Bedouin dia lects. 41 De Jong 2000: 478. 530
THE POLYGENESIS OF THE NEO-ARABIC DIALECTS
speech; however, having witnessed its penetration into Bedouin Arabic in the Sinai, it is clear that there are no limits to which features can move where.42 Although Bedouin dialects may contain a number of features collectively inherited from a common ancestor, the indicative augment is surely not one of them. It is a recent borrowing. We know this because it is readily observable. Considering this, we must ask how one can determine which NA features are inherited from an ancestral language and which are cross-dialectal borrowings. The NA dialects, having a poor history of attestation, do not offer any information regarding their developmental stages. Grouping the dialects into genetic clusters based on superficial similarities is therefore erroneous. 4.3 Putting It All Together So far it has been shown that the NA dialects do not descend linearly from a single origin. The evidence instead suggests that they have at least two sources, and developed through a complex multi-directional process of koineization, which has kept them in constant flux. Through the on-going process of koineization and the notion of the BUC, a NA dialect is best understood as a sedimentary structure of features. That is, any dialect may consist of both features inherited from its original type, and features which are the product of koineization. Once koineization occurs, and new features are acquired by a specific dialect, it may in turn interact with other dialects, thereafter triggering yet another event of koineization. This process could have been repeated numerous times throughout the centuries — slowly passing features around the Arabic speaking continuum, and thus layering in any given dialect features that developed centuries apart under unique contact conditions.
42 Thomason and Kaufman (1988) present a general outline for the study of mixed languages, that is, languages having more than one contributing source. In such contexts, borrowing is not restricted to the lexicon but can penetrate the grammar as well. This model is appropriate for the understanding of how the NA dialects absorb features at all levels of the grammar. However, since as the NA dia lects are ultimately genetically related, that is, they are all Central Semitic lan guages, it is perhaps erroneous to consider them mixed languages in the traditional sense.
531
THE POLYGENESIS OF THE NEO-ARABIC DIALECTS
Figure 243
Figure 2 is a graphic representation of the distribution of features in NA according to what has been discussed till this point. The lowest tier represents the reconstructed starting point argued for in this paper. Their interaction triggered a koineization event in which various new, koineized varieties were produced. That koineization should produce a uniform variety is not supported by the evidence. These hybrid varieties exhibit a mix of features from both TYPE I and TYPE II, along with some innovative ones to reconcile the major points of difference between the two dialects, a very common process during koineization.44 The hybrids continue to interact with each other, redistributing features and innovating new ones under certain conditions; this is represented in Koineization II. The dialects emerging from Koineization II have inherited features from their original type in Koineization I as well as borrowed or innovated ones as a result of contact with other dialects in Koineization II. In Koineization III, all of the above can be restated but with yet another event to redistribute features acquired from the previous two koineization events. By Koineization III, dialects which have been separated in terms of their acquisition of features since their genesis in Koineization I are brought into contact. This redistributes features that developed under completely different conditions to reconcile differences perhaps non-existent in the present contact situation. This can be likened to 43
Each bubble labelled Koineization # represents a point of contact between different varieties; arrows emerging from each bubble represent new koineized types. 44 Hock notes the occurrence of ‘selective simplification’ in various African Koines. This process involves the reduction or elimination of grammatical catego ries which exhibit the greatest variation cross dialectally (1991: 488). 532
THE POLYGENESIS OF THE NEO-ARABIC DIALECTS
the Bedouin adopting the indicative modal prefix of sedentary dialects. The model reveals that there is clear asymmetry between the innovation of features and their distribution among the NA dialects. Therefore, it is not possible to understand the development of NA by treating an individual dialect as a homogeneous entity. Instead, it may be more productive to speak of the evolutionary history of the individual linguistic features found in the dialects, and subsequently investigate their distribution through koineization from a reconstructed source outwards — similar to the diffusion of areal features. Figure 2 does not represent an end-point of any sort. The mechanism of change continues to recycle itself so long as contact between different varieties remains. It is therefore not productive to look for well-defined points of koineization. Those indicated as such in the diagram are there for explanatory clarity. What is clear is that the distribution of NA innovations does not necessarily represent a genetic link between those dialects in which they are found; indeed, the notion of a genetic model to capture the relationship between entire dialects is most likely untenable. In light of this, the investigation into NA features requires two separate modules of study. The first should be the formal understanding of how the various NA innovations, such as an indicative augment or periphrastic interrogatives, developed. In theory, the linguistic environment responsible for producing these innovations can be reconstructed and associated with either or both of the two original types of Arabic. The second module involves tracking the distribution of these innovations, by means of the model presented in figure 2, in an attempt to identify their geographic source and their linguistic antecedents. 4.4 The Case of Case The previous sections avoided the question of case and its relationship to the two reconstructed ancestral types, but here we should like to return, if only briefly, to its discussion. Whether case was a feature found in TYPE I or TYPE II cannot be determined at this point; however, that it does not survive in any modern dialect is not reason enough to reconstruct a case-less origin. By analogy with Latin, and certainly several other world languages, case is a category that frequently disappears during the course of language evolution. Its disappearance in Arabic is by no means shocking. However, a discussion on the shift from synthetic structures to analytical ones is beyond the scope of this paper. Here, we would like to examine the 533
THE POLYGENESIS OF THE NEO-ARABIC DIALECTS
likeliness of the survival of case given the evolutionary conditions of Arabic. Two important points must be emphasized before approaching this question: (1) NA is best understood in terms of individual features, rather than entire dialects (2) koineization is selective; features which are common to both sides of the interaction are more likely to survive the process over points of significant difference.45 If case existed in both ancestral types and, through the course of their evolution, became extinct in a few dialects, at points of koineization between those dialects and case-bearing ones, it is not likely that case would survive. Considering this, if case dropped out of only a few varieties at first, the very nature of the evolution of Arabic would have accelerated the general trend in the remainder. It should come as no surprise that case disappeared fully in all the dialects, for not only is there a general trend in the loss of case in Semitic, but also, due to the mechanism of change of Arabic, selective forces played against it. 5. Conclusions The purpose of this paper is to establish an approach by which a productive investigation into the evolutionary history of Arabic may be carried out. Section II outlined the goals of this investigation in the proposal of two central questions. Through internal linguistic evidence, coupled with historical parallels, a working answer has been supplied for the first: a defined starting point and a general mechanism of change. In doing so, the very nature of the NA dialects has been redefined. Through the modified notion of koineization as well as the BUC, the NA dialects have been shown to be a part of a larger web of contact. Instead of regarding the NA dialects as individual entities descending from an ancestral language, languages, or even linguistic material, it is more reasonable, and perhaps even necessary, to view them as sedimentary structures, including features both inherited and koineized. Moreover, to investigate the evolutionary history of Arabic requires two separate endeavours: the first is to understand the evolution of those features which are characteristic of the NA language type, such as the indicative augment (Al-Jallad, forthcoming), and second, to account for their distribution according to the model discussed in 4.3. Both of these questions are equally impor45
Hock 1991: 486, 87. 534
THE POLYGENESIS OF THE NEO-ARABIC DIALECTS
tant for a holistic understanding of the history of Arabic. Investigation carried out in these terms will lead to an account of both the development of features that define NA as a language type and their distribution in the NA dialects through contact and koineization. REFERENCES Al Jallad, Ahmad M. Forthcoming. ‘The Etymology of the Indicative Augment b in Some Neo Arabic Dialects.’ in Charles Häberl (ed.), Afroasiatic Studies in Memory of Robert Hetzron. (Cambridge) Blanc, Haim. 1964. Communal Dialects in Baghdad. (Cambridge) Blau, Joshua. 1961. The Emergence and Linguistic Background of Judeo Arabic: A Study of the Origins of Middle Arabic.2 (Jerusalem) Cachia, Pierre. 2002. Arabic Literature an Overview. (New York) Corriente, Federico. 1971. ‘On the Functional Yield of some Synthetic Devices in Arabic and Semitic Morphology’, Jewish Quarterly Review 62, 20 50 1973. ‘Again on the Functional Yield of some Synthetic Devices in Arabic and Semitic Morphology’, Jewish Quarterly Review 63, 151 63 1976. ‘From Old Arabic to Classical Arabic through the Pre Islamic Koine: Some Notes on the Native Grammarians’ Sources, Attitudes and Goals’, JSS 21, 62 98 De Jong, Rudolf. 2000. A Grammar of the Bedouin Dialects of the Northern Sinai Littoral: Bridging the Linguistic Gap between the Eastern and Western Arab World. (Leiden) Diem, Werner. 1973. ‘Die nabatäischen Inschriften und die Frage der Kasusflexion im Altarabischen’, ZDMG 123, 227 37 Ferguson, Charles. 1959. ‘The Arabic Koine’, Language 35, 4, 616 30 Garr, W.R. 1985. Dialect Geography of Syria Palestine, 1000 586 B.C.E. (Philadel phia) Haeri, Niloofar. 2003. Sacred Language, Ordinary People. (New York) Heath, Jeffery. 2002. Jewish and Muslim Dialects of Moroccan Arabic. (London) Hetzron, Robert. 1976. ‘Two Principles of Genetic Reconstruction’, Lingua 38, 89 104 Hock, Henrich H. 1991. Principles of Historical Linguistics.2 (Berlin) Huehnergard, John. 1983. ‘Assertive *la and hypothetical *lu/law in Semitic’, Jour nal of the American Oriental Society 103, 3, 569 93 2004. ‘Afro Asiatic’, in Roger Woodard (ed.), Cambridge Encyclopedia of the World’s Ancient Languages (Cambridge) 138 59 2006. ‘On the Etymology of the Hebrew Relative se ,’ in S.E. Fassberg and A. Hurvitz (eds), Biblical Hebrew in its Northwest Semitic Setting: Typological and Historical Perspectives. (Jerusalem/Winona Lake, IN) 103 25 Ingham, Bruce. 1994. Najdi Arabic. (Amsterdam) Kaye, Alan S. 1976. Chadian and Sudanese Arabic in the Light of Comparative Arabic Dialectology. (The Hague) Kaye, Alan S. and Judith Rosenhouse. 1997. ‘Arabic Dialects and Maltese’, in R. Hetzron (ed.), The Semitic Languages. (Routledge Language Families Series, New York), 263 311 535
THE POLYGENESIS OF THE NEO-ARABIC DIALECTS
Owens, Jonathan. 1998a. ‘Case and Proto Arabic, Part I’, Bulletin of the School of Oriental and African Studies 61, 1, 51 73 1998b. ‘Case and Proto Arabic, Part II’ Bulletin of the School of Oriental and African Studies 61, 2, 215 27 2006. A Linguistic History of Arabic. (Oxford) Penny, Ralph. 2002. A History of the Spanish Language. (Cambridge) Rubin, Aaron D. 2005. Studies in Semitic Grammaticalization. (Winona Lake) Thomason, Sarah G. and Terrence Kaufman. 1984. Language Contact, Creolization, and Genetic Linguistics. (Berkeley). Versteegh, Kees 1984. Pidginization and Creolization: The Case of Arabic. (Amster dam Studies in Theory and History of Linguistic Science, IV: Current Issues in Linguistic Theory, 33, Amsterdam) 1997. The Arabic Language. (New York) Whitney, William D. 2003 (repr. From 1896). Sanskrit Grammar. (Mineola, NY)
536
RELATIVE CLAUSES IN SUDANESE ARABIC1 JAMES DICKINS UNIVERSITY OF SALFORD
Abstract This paper begins with a consideration of the nature of referentiality. A distinction is drawn between objective reference, which exists out side language, and subjective reference, which I argue may be imposed on the world by language. I argue that while agreement is a grammati cal phenomenon, as traditionally believed, it has close links with the notion of subjective reference. On the basis of this, I argue that defi niteness is a properly referential, and denotative phenomenon. I go on to show that the relative ‘elements’ such as alla∂i and illi which occur in various varieties of Arabic are not pronouns, but rather markers of definiteness, and that al in Sudanese Arabic can be regarded as a ‘defi 1
I thank Janet Watson for reading a draft of this paper and making very useful suggestions. All shortcomings are my own responsibility. I also thank the Leverhulme Trust for granting me a Research Award giving me relief from teaching and administration during the academic years 2002 4 in order to pursue work on a dictionary and grammar of Sudanese Arabic. This award has contributed to the production of this paper. For present purposes, Sudanese Arabic can be taken to have the following conso nant phonemes (cf. Dickins 2007: 24): /b/ voiced, bilabial, stop; /m/ bilabial, nasal; /w/ bilabial, glide; /f/ voiceless, bilabial, fricative; /d/ voiced, apico dental, stop; /t/ voiceless, apico dental, stop; /Z/ voiced, apico dental, fricative; /s/ voiceless, apico dental, fricative; /∂/ voiced, emphatic, apico alveolar, stop; /†/ voiceless, emphatic, apico alveolar, stop; /Â/ voiced, emphatic, apico alveolar, fricative; /Ò/ voiceless, em phatic, apico alveolar, fricative; /r/ (plain), apico alveolar, trill; /®/ emphatic, apico alveolar, trill; /l/ (plain), apico alveolar, lateral; /¬/ emphatic, apico alveolar, lateral; /n/ apico alveolar, nasal; /j/ voiced, dorso prepalatal, stop; /c/ voiceless, dorso prepalatal stop (marginal phoneme); /s/ voiceless, dorso prepalatal, fricative; /n/ dorso prepalatal, nasal (marginal phoneme); /y/ dorso palatal, glide; /g/ voiced, post dorso velar, stop; /k/ voiceless, post dorso velar, stop; /g/ voiced, post dorso post velar, fricative; /x/ voiceless, post dorso post velar, fricative; /¨/ voiced, pha ryngeal, fricative; /Ì/ voiceless, pharyngeal, fricative; /’/ voiced, glottal, fricative (sometimes described as glottal stop); /h/ voiceless, glottal fricative. For present purposes, Sudanese Arabic can be taken to have the following vowel phonemes (cf. Dickins 2007:25): /a/ open, unrounded, short vowel; /i/ front, close, unrounded, short vowel; /u/ back, close, rounded, short vowel; /u/ front, mid, unrounded, long vowel; /e/ back, mid, rounded, long vowel (see, however, Dickins 2007 for a critique of this account and an alternative analysis). 537
RELATIVE CLAUSES IN SUDANESE ARABIC
nite particle’. I argue that in Sudanese al is the head of its phrase, and consider the referentiality of al in the dialect. I then consider demonstratives as secondary relative clause markers in Sudanese, and non restrictive relative clauses in the dialect. I finally consider cases in which the al head of a relative clause is not co referential with the ele ment (particularly an adjectival element) which follows it. The theoretical analysis provided in this paper is based some what informally on the linguistic theory of axiomatic functional ism initially developed by J.W.F. Mulder and S.G.J. Hervey (e.g. Mulder 1968, 1989; Mulder and Hervey 1972, 1980; Hervey 1979, 1982), and subsequently extended on the basis of proposals put for ward by Michael Lamb in Dickins (1998). In an attempt to keep the main body of the paper as generally comprehensible as possible, I have confined technical comments relating to axiomatic functional ism to footnotes.2
1. General Background In this article, I shall use the term ‘Sudanese Arabic’ as shorthand for Central Urban Sudanese (Dickins 2007; elsewhere termed Khartoum Arabic [Dickins 2006]), that is, the dialect standardly spoken by long-term native Arabic-speaking residents of Greater Khartoum (Khartoum, Khartoum North, and Omdurman), and in other urban areas of central Sudan, roughly to the towns of Atbara in the north, Sennar on the Blue Nile, and Kosti on the White Nile. For details of the transcription system, see Dickins (2007). In this paper, I use a subscript to indicate a vowel which is deleted in liaison ‘readings’ of the material presented. The Sudanese examples for this article have been taken from three types of sources: 1. Examples constructed by myself. These are marked with a (C) after them, or sometimes an entire group of examples is noted as having been constructed before they are given.3 2 In fact, the syntactic analysis in this paper is much more like that of Mulder and Hervey than that implied whether in lexotatics or delotactics in Dickins (1998). In Dickins (1998), syntax (roughly as normally understood) or delotactics is an analysis involving content (semantic elements) only, whereas in Mulder and Hervey, syntactic analysis involves elements which have both form and content. In Dickins (1998), I have argued that Mulder and Hervey’s standard version of axio matic functionalism is coherently subsumed under the extended version of the theory proposed there (e.g. Dickins 1998: 250 1). Accordingly, the theoretical models of standard axiomatic functionalism are fully interpretable in terms of mod els within extended axiomatic functionalism. In extended axiomatic functionalist terms, Mulder and Hervey’s syntax is an analysis of relations at the allosemic level (what might be termed allosemotactics). 3 Some of the constructed examples are either elicited from Elrayah Abdelgadir
538
RELATIVE CLAUSES IN SUDANESE ARABIC
2. Material which I tape-recorded in 1985 and 1986, either from Omdurman Radio, or spontaneous conversations between Sudanese friends; or material which I wrote down from spontaneous speech. All these examples are marked with an (R) after them. 3. Material taken from a number of radio series recorded for me by Omdurman Radio in 1986. These are marked with the name of the series after them.
The definite article in Sudanese Arabic has the following allomorphs: After a consonant alad-, at-, az-, etc. After a vowel ld-, t-, z-, etc.
– before all non-apical and non-dorsal consonants: /b/, /m/, /w/, /f/. /y/, /g/, /k/, /x/, /g/, /¨/, /Ì/, /’/, and /h/. – before apical and dorsal consonants: /d/, /t/, /z/, /s/, /∂/, /†/, /Â/, /Ò/, /r/, /®/, /l / /¬/, /n/, /j/, /s/, and the marginal /c/ and /n/. – before all non-apical and non-dorsal consonants – before apical and dorsal consonants
In this article, I have used subscript forms to indicate vowels which ‘disappear’ through so-called re-syllabification in liaison with preceding or following words: thus, fihim ‘he understood’ in which the syllables are /fi/ and /him/ (with stress on the first syllable /fi/), but fihim al-kalam ‘he understood the statement/discussion’ (in which the first two syllables are /fih/ and /mal/). I have not used subscript forms with following pronouns — thus fihm-u ‘he understood it (m.sg.)’ (not fihim-u); also fihmu ‘they understood’ (same pronunciation as fihm-u ‘he understood it (m.sg.)’. I have also not used subscripts with the article al-. Thus fihmu l-kalam ‘they understood the statement/discussion’ (not fihmu al-kalam). We can establish the following provisional word and phrase categories for Sudanese Arabic: Word category noun adjective verb adverb
Phrase category nominal phrase adjectival phrase verb phrase adverbial
(NomP) (AdjP) (VerbP) (AdvP)
or produced by him without prompting on my part. Elrayah Abdelgadir acted as consultant in 2005 and 2007 for an Arabic/English Dictionary of Sudanese Arabic which I am working on. 539
RELATIVE CLAUSES IN SUDANESE ARABIC
For current purposes, I shall treat a phrase as any element consisting of more than one word (not counting the clitic al-). Where the element consists of only one word, I will refer to it simply as a noun, adjective, verb, or adverb. Where a noun (etc.) is preceded by al-, I shall refer to this at this point in the paper as a definite noun (etc.). We can also establish the notion of a bipartite clause, consisting of a predicand (which, for the moment, we can characterize as consisting of a noun or nominal phrase) and a predicate (which may be a noun, nominal phrase, adjective, adjectival phrase, verb, verb phrase, adverb, or adverbial.4 Consider the following in Sudanese Arabic in which a noun is followed by a member of a word category or phrase category. The ‘a’-list (1a, 2a, 3a, etc.) consists of indefinites, while the ‘b’-list (1b, 2b, 3b, etc.) consists of corresponding definites (all examples constructed by myself ). Indefinite initial noun Following Element Noun NomP Adjective
1a. zol muÒiba 2a. zol muÒiba kabira 3a. bet kabir 4a rajil za¨lan AdjP 5a. rajil za¨lan minn ak Verb 6a. rajil zi¨il 7a. rajil bifham VerbP 8a. rajil zi¨il minn ak 9a. rajil bifham al muskila Adverb 10a. rajil hinak AdvP 11a. rajil fi l bet Bipartite clause 12a. rajil bet u kabir 13a rajil bet u itkassar 14a rajil itkassar bet u
‘a terrible man’ [lit: ‘a disaster man’] ‘a very terrible man’ [lit: ‘a big disaster man’] ‘a big house’ ‘an angry man’ ‘a man who’s angry with you (m.sg.)’ ‘a man who got angry’ ‘a man who understands’ ‘a man who got angry with you (m.sg.)’ ‘a man who understands the problem’ ‘a man [who is] there’ ‘a man in the house’ ‘a man whose house is big’ ‘a man whose house was demolished’ ‘a man whose house was demolished’
4 The terms ‘predicand’ and ‘predicate’ are used in Bohas, Guillaume and Kouloghli (1990: 43), where they are employed as the translations of the traditional Arabic grammatical terms mubtada’ and xabar. They are subsequently used by Watson (1993). In respect of classical (and Standard) Arabic, the term ‘predicand’ in particular allows the term ‘subject’ to be employed as a translation of the tradi tional Arabic grammatical term fa¨il (i.e. the subject of the so called verbal sen tence). Although I do not believe that Sudanese Arabic makes a properly syntactic distinction between nominal and verbal sentences (see Dickins forthcoming), I have used the term ‘predicand’ in this paper to avoid the bizarre sounding descrip tion of the ‘non predicate’ element of an Arabic bipartite clause as the ‘subject’ of this clause, even where this predicate is clearly non verbal:. Thus in ar rajil fi l bet ‘the man’s in the house’, for example, where the predicate fi l bet ‘in the house’ is clearly non verbal, I want to avoid describing the ar rajil element as the subject.
540
RELATIVE CLAUSES IN SUDANESE ARABIC
Definite initial noun Following element Noun NomP Adjective
1b. az zol al muÒiba 2b. az zol al muÒiba l kabira 3b. al bet al kabir 4b. ar rajil az za¨lan AdjP 5b. ar rajil az za¨lan minn ak Verb 6b. ar rajil az zi¨il 7b. ar rajil al bifham VerbP 8b. ar rajil az zi¨il minn ak 9b. ar rajil al bifham al muskila Adverb 10b. ar rajil al hinak AdvP 11b. ar rajil al fi l bet Bipartite clause 12b. ar rajil al bet u kabir 13b. ar rajil al bet u itkassar 14b. ar rajil al itkassar bet u
‘the terrible man’ [lit: ‘the disaster man’] ‘the very terrible man’ [lit: ‘the big disaster man’] ‘the big house’ ‘the angry man’ ‘the man who’s angry with you (m.sg.)’ ‘the man who got angry’ ‘the man who understands’ ‘the man who got angry with you (m.sg.)’ ‘the man who understands the problem’ ‘the man [who is] there’ ‘the man in the house’ ‘the man whose house is big’ ‘the man whose house was demolished’ ‘the man whose house was demolished’
In the above, 1a–14a involve attribution (of a noun, nominal phrase, adjective, adjectival phrase, etc.) to an indefinite noun, while 1b–14b involve corresponding attribution to a definite noun. Examples 1a–14b illustrate various kinds of attributive structures in Sudanese Arabic, covering what are, from an English point of view, both adjectival attribution and relative clauses. What examples 1a– 14b appear to show is that whenever an initial noun is indefinite, as in examples 1a–14a (where the noun rajil does not have a preceding al-), the attributive element is indefinite, i.e. it also does not have a preceding al-. By contrast, whenever the preceding noun is definite, as in 1b–14b (where the noun ar-rajil has a preceding al-), the attributive element is also definite, i.e. it also has a preceding al-. This state of affairs holds true, regardless of whether the attributive element is a noun, nominal phrase, adjective, adjectival phrase, verb, verb phrase, adverb, adverbial, or a bipartite (predicand-predicate) clause. Where the bipartite clause consists of a predicand and a verbphrase predicate (as in 13a, 13b, 14a, and 14b), it holds true regardless of whether the predicand precedes the predicate (as in 13a and 13b (al-)bet-u itkassar, or whether the predicate precedes the predicand (as in 14a and 14b) (al-)itkassar bet-u, this latter perhaps being a marked word order). There thus appears to be no real structural distinction in Sudanese Arabic between attributive adjectival structures and relative clauses. In this paper, I will argue that this initial analysis is true, and explore some of the ramifications of the analysis. 541
RELATIVE CLAUSES IN SUDANESE ARABIC
2. Objective Reference vs. Subjective Reference In Sudanese Arabic, the masculine word Òaj and the feminine word doka are regionally based synonyms meaning ‘cooking-plate’ for kisra’ (a type of flat bread). Consider a situation in which person A (a woman, since women normally cook kisra) says wen-u↑5 ‘Where is it?’ (i.e. wen ‘where’ + -u↑ m.sg. suffix). In this situation, person B can legitimately reply, if in doubt about what is being referred to, bi-tagÒudi Ò-Òaj? ‘Do you mean the cooking-plate?’, in which the masculine word Òaj ‘echoes’ the masculine suffix - u↑. Person B cannot, however, legitimately say bi-tagÒudi d-doka? ‘Do you mean the cooking-plate?’, using the feminine word doka.6 Superficially, there is no mystery about this. If an object has been referred to as masculine through the use of a pronoun, we cannot then use a feminine noun to identify the same object. Why, however, is it not possible to do this? Saussure fairly famously said ‘C’est le point de vue qui crée l’objet’ (Saussure 1975: 23), which has been translated as ‘It is the viewpoint which creates the object’ Saussure 1959: 8), and ‘it is the viewpoint adopted which creates the object’ (Saussure 1983: 8). Anybody who has been hit by a bus or other large, hard object is likely to disagree with this statement, particularly if it is interpreted naively. However, what Saussure meant is that the way in which we conceive something in some respects at least gives that object what is for us its identity in relation to other objects of the same or different types. In a similar vein, Hjelmslev said, ‘… by virtue of the contentform and the expression-form, and only by virtue of them, exist respectively the content-substance and the expression-substance, which appear by the form’s being projected onto the purport, just as an open net casts its shadow down on an undivided surface’ (Hjelmslev 1963: 57; italics mine). We can understand content-form to be the abstract semantic organisation of the language, and expression-form to be the abstract formal organisation of the language (roughly phonology, if conceived as an abstract system). We can understand purport as be The vertical arrow ↑ here indicates a high tone (cf. Dickins 2007: 74). Because they are, in origin at least, regionally restricted forms Òaj is North ern Sudanese and doka is Western Sudanese these two words are not ideal forms to illustrate synonymy between masculine and feminine words referring to inani mate objects in Arabic. Better forms are provided by Standard Arabic mi†raq vs. mi†raqa, both meaning ‘hammer’, without, apparently, any dialectal like complica tions. However, as the focus of this paper is on Sudanese Arabic, I have preferred to use Sudanese Òaj and doka here. 5 6
542
RELATIVE CLAUSES IN SUDANESE ARABIC
ing extra-linguistic reality. Expression-substance is phonetic reality as informed by abstract linguistic analysis; for example not just a sound which can be represented as [p], but a sound [p] conceived as a function (realisation) of a phonological entity, for example /p/ in English. Content-substance is extra-linguistic reality as informed by an abstract semantic analysis of a particular language (English, Arabic, or whatever). Fairly interestingly, Hjelmslev’s analysis of the relationship between abstract semantic structure and purport can be taken to provide an account of the view that language does not simply describe pre-existing boundaries between concepts, but rather defines those boundaries, and therefore in a very real sense defines the concepts themselves (‘creates the objects’): the rope-like structure of the ‘open net’ imposes conceptual boundaries in the real world. There is, however, a further interpretation — or at least an additional possibility — which can be given to Hjelmslev’s analogy. If we think of a semantic account as a net which casts a shadow (necessarily because there is light shining through it), the shadow which it casts not only divides up the entities (the purport) onto which it shines; it also produces an impression (a shadow) of itself on these entities: the semantic reality which language as a net defines is not simply a divided up (conceptualized) version of external reality: it is a version of external reality onto which the trace of language is itself superimposed. What is the relevance of this to the example of Òaj and doka ‘cooking-plate’ in Sudanese Arabic? We cannot say bi-tagÒudi d-doka? ‘Do you mean the cooking plate’ if someone else says wen-u↑ ‘Where is it (m.sg.)?’ because the masculine form -u↑ only ‘picks out’ in the world those entities which are conceived masculinely. Doka conceives a cooking plate femininely; it casts a ‘feminine’ element from Hjelmslev’s open net of content-form onto the extra-linguistic reality (purport) of cooking-plates. Stretching Hjelmslev’s analogy somewhat, we may say that Òaj, by contrast, conceives a cooking plate masculinely; it casts a ‘masculine’ element from Hjelmslev’s open net of content-form onto the extra-linguistic reality (purport) of cookingplates. What I am suggesting is that we should regard masculine and feminine, even where these are purely a matter of so-called grammatical gender (they do not indicate maleness or femaleness in the real world), as having a kind of reference. This kind of reference is not objective: it does not exist (in any sense) ‘out there’ in the extralinguistic world independently of language. However, it is established by the ‘shadow’ of the ‘net’ of language itself, and only exists 543
RELATIVE CLAUSES IN SUDANESE ARABIC
in the context of that shadow. I suggest that reference of this kind be called subjective reference: it requires something in language as its ‘subject’ to bring it into existence (or at least in a certain restricted sense to bring it into existence). 3. The Subjective-Objective Continuum It seems clear that the masculineness of Òaj and the feminineness of doka ‘cooking-plate’ are a matter of subjective reference — a fact underlined by the synonymity of the two words in a denotative (or by extension truth-oriented) approach to semantics. It also seems quite clear that when masculine and feminine refer to entities (such as human beings) whose sex can be clearly identified, gender in Arabic is objective. Thus: labban labbana
‘male milk seller’ ‘female milk seller’
However, in some cases the boundary between objective and subjective reference is not clear. In English, some words for objects which involve a pairing of similar-shaped (and particularly roughly symmetrical) components are grammatically plural, e.g.: ‘scissors’, ‘tweezers’, ‘trousers’, ‘shorts’. Although the objects in question are most reasonably defined as singular, they have an element of objective componential plurality. Similarly all sand in fact consists of one or more grains of sand. Whether we refer to a pile of sand as ‘sand’ or ‘grains of sand’ is in one sense a matter of our perspective on it (‘the viewpoint creates the object’), rather than simply a question of the objective reality of the referent (‘purport’) itself. ‘Sand’ and ‘grains of sand’ are, of course, different from Sudanese Òaj and doka ‘cooking plate’. Masculine Òaj and feminine doka do not in any normal sense offer different perspectives on the object in question, while ‘sand’ and ‘grains of sand’ clearly do. However, the general point remains that language offers a variety of types of referring (reference), ranging from those which are clearly objective, to those which involve genuinely different perspectives on the same phenomena, to those which are purely subjective. 4. Agreement Both in traditional grammar and in modern approaches agreement (or concord) has been regarded as a matter of grammar (e.g. Corbett 544
RELATIVE CLAUSES IN SUDANESE ARABIC
2006: 1–8, and elsewhere). I believe that this view is essentially correct: we are, after all, inclined to regard forms involving wrong agreement, such as Òaj kabira ‘big (f.sg.) cooking-plate (m.sg.)’ (literally ‘cooking-plate big’) and doka kabir ‘big (m.sg.) cooking-plate (f.sg.)’ (literally ‘cooking-plate big’) as grammatically incorrect, even without the aid of an explicit linguistic theory which tells us that this is the case. However, the example wen-u↑ ‘Where is it?’, bi-tagÒudi Ò-Òaj? ‘Do you mean the cooking-plate?’ (not *bi-tagÒudi d-doka?) makes plain that agreement — or at least agreement-like phenomena — are not simply part of the domain of grammar. I will assume alongside traditional grammar and the great majority of modern approaches to linguistics that the maximum domain of grammar is the sentence. I also assume that a sentence is an abstract entity which is instanced (actually and potentially) in utterances (actual and potential) (a view implicitly shared by most approaches to linguistics, albeit that its consequences are often not coherently explored in linguistic theorising). The agreement-like features of an example such as wen-u↑ ‘Where is it?’, bi-tagÒudi Ò-Òaj? ‘Do you mean the cooking-plate?’, are not a matter of grammar for two reasons. Firstly, these features extend beyond the sentence (the maximum domain of grammar). Secondly, and by extension, the restrictions on gender occurrences (the impossibility/unacceptability of the reply bi-tagÒudi d-doka? ‘Do you mean the cooking-plate?’ vs. the possibility/acceptability of the reply bitagÒudi Ò-Òaj?) must be a feature of the utterance-level, rather than the sentence-level. Given that the sentence is the maximum domain of grammar, it is only utterances (as instances of sentences) rather than sentences themselves which can occur in sequence. We may also compare with the grammatical phenomenon of agreement (proper) another agreement-like feature, that of semantic contradiction, as in ‘We saw them tomorrow’. This is arguably simply pragmatically aberrant; in the world as we know it, events cannot be simultaneously in the past and in the future (perhaps there is a world, e.g. one instancing the idea of cyclical time and the ‘eternal return’ (cf. Eliade 1959) in which this state of affairs does not obtain). I earlier suggested that incorrect sentence-internal agreement yields a non-grammatical sentence (strictly speaking a non-grammatical putative — or would-be — sentence, since non-sentences are not properly a part of grammar). This seems at first sight a curious state of affairs in light of the argument that agreement-like phenomena arise from incompatible/contradictory subjective referents (as instanced by utterance-sequences such as wen-u↑ ‘Where is it?’, bi545
RELATIVE CLAUSES IN SUDANESE ARABIC
tagÒudi Ò-Òaj ? ‘Do you mean the cooking-plate?’). If such subjective incompatibility/contradiction between utterances is not a grammatical matter, we would imagine that similar incompatibility/contradiction within utterances would similarly be a non-grammatical matter. A partial — though also apparently circular — answer to this puzzle seems to be that such intra-utterance subjective semantic incompatibility/contradiction would be a non-grammatical matter, if such utterances could legitimately occur in a language. Such utterances cannot, by definition, however, legitimately occur, if we accept that utterances instantiate sentences, and that the corresponding wouldbe sentences are non-grammatical (i.e. properly speaking non-sentences), and therefore not part of the grammar. Further insight is provided by a comparison with real (objective) semantic (referential) incompatibility/contradiction, as in ‘We saw them tomorrow’. Here we can locate the problem of interpretation in the real world as we know it. In the case of ‘failed agreement’ such as Òaj kabira ‘big (f.sg.) cooking-plate (m.sg.)’ (literally ‘cooking-plate big’), there is no objective real-world incompatibility. However — suspending for the moment our disbelief in the grammatical possibility of Òaj kabira — an utterance involving the phrase Òaj kabira would involve a subjective semantic (referential) incompatibility. Because this incompatibility has no objective status in the world beyond language, it is, however, perceived as a form of ‘disturbance’ more than anything else. This is apparent in an exchange such as wen-u↑ ‘Where is it?’ (wen ‘where’ + -u↑ m.sg. suffix) / bi-tagÒudi d-doka? ‘Do you mean the cooking-plate (‘doka’ — f.sg.)?’ The exchange is unacceptable but not for any reason which is definable outside language. In ‘psychological pre-history’ as it were, one can see how ‘disturbed’ utterances (or parts of utterances) such as Òaj kabira ‘big (f.sg.) cooking-plate (m.sg.)’ are ruled out, along with the sentences which these utterances would instantiate. If the source of the ‘disturbance’ is not to be located outside language in the world, it must be located in language, i.e. in grammar. By virtue of such ‘disturbance’, a ‘sentence’ of this kind is ‘wrong’, i.e. not a sentence at all, and as such does not constitute part of the grammar of the language. 5. Definiteness and Reference Denotative difference is sometimes defined as difference in overall range of meanings between two words or phrases (in particular senses). A more generally adequate view, I believe, is that denotative 546
RELATIVE CLAUSES IN SUDANESE ARABIC
difference is demonstrated by difference in meaning or range of meaning between two words or phrases (in particular senses) in at least one context. That is to say, if there is one context (at least) in which two words or phrases (in particular senses) have a different meaning or range of possible meanings, these two words or phrases are different in denotation. Taking ‘denotation’ in the meaning just given, it can be demonstrated that ‘definite’ has a different denotation to that of ‘indefinite’ by consideration of the following two possible conversations between characters A and B. 15A. 15B1. OR 15B2.
I won a cake in the garden-party raffle. That’s funny; I also won a cake. That’s funny; I also won the cake.
Here B’s first answer (15B1) means that B won one cake out of one or more cakes to be won. (It also involves the defensible expectations that i. there was more than one cake to be won; and ii. that the cake which B won was different from the cake which A won.) B’s second answer (15B2) means that B won the same cake as A. (15B2 would be possible where, for example, winning the cake involved having a specific numbered raffle ticket, and A and B by some mishap happened both to have this number raffle ticket.) The answers 15B1 and 15B2 differ referentially in that 15B1 includes the possibility that B won a different cake from A (in fact it would standardly mean that B won a different cake from A), while 15B2 necessarily means that B won the same cake as A. I take it as obvious that ‘cake’ in 15B1 and 15B2 involves (an instantiation of ) the same word ‘cake’ in the same sense, and that ‘cake’ 15B1 and 15B2 are therefore (an instantiation of ) the same denotation. The difference in referential possibilities between 15B1 and 15B2, however, demonstrates that 15B1 and 15B2 are (instantiations of ) different overall denotations. Since this overall denotative difference is not a matter of (instantiations of ) different denotations of ‘cake’ in 15B1 and 15B2, it must be a matter of (instantiations of) different denotations of the other element in ‘a cake’ and ‘the cake’, i.e. ‘a’ and ‘the’. That is to say, ‘a’ and ‘the’ have different denotations. 6. Relative ‘Elements’ in Arabic as Non-Pronouns Relative ‘elements’ in Arabic, such as Standard Arabic alla∂i and even colloquial forms such as illi are sometimes referred to as relative pronouns (e.g. for Standard Arabic, Haywood and Nahmad 1965: 284; 547
RELATIVE CLAUSES IN SUDANESE ARABIC
for colloquial Arabic Yoda 2005: 131). In fact, such elements do not function much like pronouns, and certainly not like the relative pronouns of English and some other European languages. This is demonstrated by the following considerations: 1. The determinant of whether alla∂i or illi (etc.) occurs or not is whether the head noun-phrase to which the relative clause is attributed is definite or not. If the head noun-phrase is definite alla∂i or illi is obligatory. If it is not, it cannot occur (but see also Section 13). 2. Normally, alla∂i or illi does not occur instead of a pronoun in the main clause. Thus, in Standard Arabic ar-rajul alla∂i a¨rifu-hu ‘The man whom I know’, and not normally at least ar-rajul alla∂i a¨rif. There are occasional exceptions to this in Standard Arabic (cf. Badawi, Carter and Gully 2004: 493) and more frequent ones in Classical Arabic (cf. Wright 1971, vol. 2: 320), but not in most dialects (however, for non-occurrence of the object resumptive pronoun in Moroccan Arabic, see Heath 2002: 495).7 Thus the main body of the relative clause (alla∂i, illi, etc.) in Arabic comprises a ‘full’ form which is exactly like that of a corresponding independent clause (or other clause-like structure). There is an increasing tendency amongst scholars to recognize that forms such as alla∂i or illi are unlike the relative pronouns of English and similar languages, and terms are used such as ‘relative noun’ (e.g. Badawi, Carter and Gully 2004: 497) for Standard Arabic, or ‘relative particle’ (Dickins 2006) or 7
A third consideration in Standard Arabic is number and gender agreement, and, for the dual, case agreement. Agreement is with the head noun of which the relative clause is an attribute, rather than being a function of the status of the rela tive element (relative pronoun) within the relative clause, as it is in many European languages. Thus in, huma ar rajulan alla∂an ra’aitu huma ‘they are the two men whom I saw’ ar rajulan ‘the two men’ is dual m. nom. (as the predicate of the clause huma ar rajulan…). The relative element alla∂an ‘whom’ (at least in translation terms) agrees with this head noun, and is thus also dual m. nom. In English, by contrast, ‘whom’ is the object of the relative clause, having the ‘object’/’non nomi native’ suffix m. There is at least one situation in which Sudanese Arabic does not require a co referential pronoun in the relative clause. This is where the co referential element in the relative clause would be a temporal adverbial adjunct. An example is ayyi yom taji fi Ò ÒabaÌ talga hu ‘[on] any day, [when] you come in the morning, you’ll find him’. Here the relative clause taji fi Ò ÒabaÌ lacks a pronoun which is co referential with yom ‘day’ in the preceding main clause. Note, however, that it is not only a co referential pronoun (h)u which is missing here. Also missing is a preposition in this case fi ‘in/on’, to which the co referential pronoun would be attached. The alternative to ayyi yom taji fi Ò ÒabaÌ talga hu ‘[on] any day, [when] you come in the morning, you’ll find him’ with these additional elements is ayyi yom taji fi (h)u fi Ò ÒabaÌ talga hu. This is not then a case in which the Arabic lack of co referential pronoun within the relative parallels the lack of a co referential pronoun within the relative clause in a language like English. 548
RELATIVE CLAUSES IN SUDANESE ARABIC
‘marker of definiteness’ (Wise 1975: 88), or even ‘clausal definite article’ (Watson 1993: 230) for colloquial dialects.
7. Relative ‘Elements’ as Definite ‘Markers’ Particularly in Arabic dialects, it seems clear that relative ‘elements’, such as illi, are in fact definite ‘markers’ along the same lines as the definite article il- (or similar). Thus, in Egyptian Arabic, there is virtual complementary distribution between il- and illi. Il- seems to occur in some contexts, while illi seems to occur in others. Thus: pre noun pre adjective pre prepositional phrase
il bet il kibir illi fi l bet
pre verb
illi raÌ il bet
pre bipartite clause
illi ÒaÌb u fi l bet
‘the house’ (C) ‘the big one’ (C) ‘the one who is/ the ones who are in the house’ (C) ‘the one who went to the house’ (C) ‘the one whose friend is in the house’(C)
In fact, there is not full complementary distribution. Consider the following, which occurs in a short story by the Egyptian writer Yusuf Idris called al-Garib: il Ìa’’ mis hu illi gal†an … ana l gal†an
In fact/The fact is, it wasn’t him who was [the] wrong [one]; it was me/I was [the] wrong [one].
Before an adjective, it is thus possible in Egyptian Arabic to have both il- and illi. The usage illi gal†an seems more clausal than the usage (i)l-gal†an. This may, however, be simply a function of likely translation equivalents: ‘the one who was wrong’ vs. ‘the wrong one’. I will not pursue this question here (but cf. Dickins forthcoming, Section 4, on monopartite clauses in Sudanese Arabic). 8. The Definite Particle Sudanese Arabic, as spoken in Khartoum and the other big cities of Central Sudan, standardly uses al- in all the contexts described in Section 8 (the form alli, paralleling Egyptian illi, exists as a dialect form in the far north of the country. It is otherwise only heard in formalizing variants of the dialect, e.g. in radio interviews, where it mimics the Egyptian illi ). Thus: 549
RELATIVE CLAUSES IN SUDANESE ARABIC
pre noun pre adjective pre prepositional phrase
al bet al kabir al fi l bet
pre verb pre bipartite clause
al masa l bet aÒ ÒaÌb u fi l bet
‘the house’ (C) ‘the big one’ (C) ‘the one who is/ the ones who are in the house’ (C) ‘the one who went to the house’ (C) ‘the one whose friend is in the house’8 (C)
It seems obvious that this is the same form al- in all cases, and that it has the same denotation in all cases. In fact, even if one wanted to make some kind of distinction between pre-noun or pre-adjective alon the one hand, and pre-clausal/pre-semiclausal al- (i.e. pre-prepositional phrase, pre-verb, pre-bipartite clause al-) on the other, this distinction seems unsustainable. This is already suggested by the fact that the apparently pre-adjective vs. pre-clausal distinction between il-gal†an and illi gal†an in Egyptian Arabic, does not exist for Sudanese. All Sudanese has is al-galtan.9 That the distinction is quite imposed by outside presumptions — and not a feature of the language itself — is underlined by the following: ar rajil az za¨lan ar rajil az za¨lan minn ak al mara l gablana al mara l gablana l fikra di
‘the angry man’ (C) ‘the man who’s angry with you (m.sg.)’ (C) ‘the accepting woman/the woman who accepts’ (C) ‘the woman who accepts that idea’ (C)
Here az-za¨lan minn-ak ‘the [one] who’s angry with you’, where we might (from a translation point of view at least) want to regard al- as functioning clausally, is simply an expansion of the adjectival azza¨lan. Similarly, al-gablana l-fikra di ‘the [one (f.sg.)] who accepts 8
The use of al in contexts where one would expect a member of the alla∂i set of relative ‘elements’ is very old in Arabic. Examples quoted by Wright from ancient Arabic include ma anta bi l Ìakami t tur∂a Ìukumatu hu ‘you are not the judge whose judgement is accepted’ (Wright 1971: vol. 1, 269), where one would expect ma anta bi l Ìakami alla∂i tur∂a Ìukumatu hu. In the case of certain other struc tures in both Classical and Standard Arabic, al is obligatory, although its scope is unambiguously the entire following clause, rather than the immediately following element (in the following case an adjective); thus: al muhajirun al mutawaqqa¨ wuÒulu hum ‘the immigrants whose arrival is expected’. Here, the clause mutawaqqa¨ wuÒulu hum (lit: ‘expected arrival their’) is analysed in terms of tradi tional Arabic grammar as a predicand predicate structure (see Section 13), in which the predicate mutawaqqa¨ ‘expected’ is fronted. The al before mutawaqqa¨ wuÒulu hum ‘expected [is] their arrival’ provides agreement in definiteness with the previ ous al muhajirun ‘the immigrants’, and relates to the entire following clause mutawaqqa¨ wuÒulu hum ‘expected [is] their arrival’ rather than simply to mutawaqqa¨ ‘expected’. 9 Sudanese Arabic does not standardly have emphatic consonants in the envi ronment of /g/and /x/; hence the ‘t’ in galtan (cf. Dickins 2007: 41 2). 550
RELATIVE CLAUSES IN SUDANESE ARABIC
that idea’ is also an expansion of al-gablana. (Gablan is one of very few fa¨lan forms in Sudanese Arabic which can take a direct object.) In all cases al- functions as a definite in the same way. Given its wide range of uses in Sudanese Arabic, I will henceforth use the term definite particle (rather than definite article) to describe it. 9. Definite Particle as Syntactic Head Axiomatic functionalism works with a fairly simple dependency-type grammar, involving a distinction between morphology and syntax, and within syntax only heads and modifiers.10 There are various ways in which morphology is distinguished from syntax by different approaches to linguistics. Here I will initially define morphology as involving elements which can only be directly ordered with respect to one another in one specific way.11 Syntax involves elements which can be directly ordered with respect to one another in more than one way. Thus, taking unstoppable to consist of the elements un-, -stopand -able, unstoppable is a morphological complex: we can say ‘unstoppable’, but not ‘unablestop’, ‘stoppunable’, ‘stopableun’, or ‘ableunstop’, or ‘ablestopun’. By contrast, in ‘Cats hate dogs’, the relationship between cats, hate and dogs is syntactic: we can say ‘dogs hate cats’ as well as ‘cats hate dogs’. By extension, in ‘able to stop’ the relationship between able and to is syntactic and that between to and rest is also syntactic. This is shown by the fact that we can say not only ‘able to stop’ but also ‘willing and able to stop’ and ‘able to stop and rest’.12 10
What I have here termed a ‘head’ is more technically known as a ‘nucleus’ (e.g. Dickins 1998: 384), and what I have here termed a ‘modifier’ is more techni cally known as a ‘peripheral entity’ (e.g. Dickins 1998: 385). 11 Strictly speaking, one can add that morphology may involve elements which can be ordered differently with respect to one another, but that where this occurs, the difference is not taken to be relevant to grammatical identity. If, for instance, it were possible to say in English not only ‘unstoppable’, but also ‘stoppunable’ but if this made no difference to the denotation of the word in question, and if both ‘unstoppable’ and ‘stoppunable’ were perceived as no more than variants of a single word, unstoppable/stoppunable would still be analysable as a morphological, rather than syntactic, complex, the phonological variance being treated as a matter of allomorphy. 12 One obvious practical aspect of this definition is that the fact of being a pho nological word has no bearing on whether an entity is morphological or syntactic. Thus, ‘herself’, for example, can be shown to be syntactic (a syntactic complex) rather than morphological (a morphological complex) by the fact that English has forms such as ‘him and herself’ (or ‘him and herself ’, or ‘him and her self ’). Simi larly, ‘periodontist’ is shown to be syntactic rather than morphological by the fact 551
RELATIVE CLAUSES IN SUDANESE ARABIC
There are also various ways in which heads have been defined. Zwicky (1985), for example, puts forward eight possible criteria, summarized in Corbett, Fraser and McGlashan (1993: 1). Of particular interest here is Corbett, Fraser, and McGlashan’s sixth criterion: namely, that the head is ‘the distributional equivalent, that is, the constituent whose distribution is identical to that of the whole construct?’ (ibid.). Consider the following (all examples constructed by myself, and checked with informants): 16. 17 18 19 20 21
*sufta za¨lan sufta z-za¨lan sufta rajil *ma¨a za¨lan ma¨a z-za¨lan ma¨a rajil
‘I saw an angry one/man’ ‘I saw the angry one/man’ ‘I saw a man’ ‘with an angry one/man’ ‘with the angry one/man’ ‘with a man’
As examples 16–21 suggest, the distribution of al-phrases — i.e. the syntactic ‘slots’ in which they can occur in the sentence — is determined by the presence of al-. In fact, this represents a somewhat more complex situation than that described by Corbett, Fraser, and McGlashan’s characterisation of the head as ‘the constituent whose distribution is identical to that of the whole construct’, since al- cannot occur without its accompanying phrase: we cannot find situations in which al- occurs as a phrase on its own, such that we can say that the distribution of al- is the same as that of the whole construct. We can, however, say that the occurrence of al- determines the distribution of the entire al- phrase: that is to say, al- is the crucial element in determining what syntactic positions al-phrases occur in. Thus, al-phrases can occur as objects of verbs and prepositions (17 and 20 above), whereas otherwise identical phrases lacking al- (16 and 19 above) cannot. As 18 and 21 suggest, the general distribution of al-phrases is like that of bare nouns. Al-phrases can thus be described as nominal — where ‘nominal’ is distinguished from adjectival — cf. 16 and 19 — in contrast to the categorisation of traditional Arabic grammar which includes both nouns and adjectives in a single ‘name’ (ism) word-class. Examples 17 and 20 are potentially problematic as syntactic alphrases. Under the definition which I gave earlier in this section, for az-za¨lan to be syntactic at all, we have to show that al- and za¨lan can relate directly to one another as grammatical elements in more that we can say ‘ortho and periodontists)’ (cf. Arnstein 2005: 359). (I have, of course, omitted quite a lot of the argumentation needed to fully justify these analy ses here, but I hope that the examples given illustrate the general principles.) 552
RELATIVE CLAUSES IN SUDANESE ARABIC
than one way. That the al-adjective relation is syntactic is suggested by forms such as az-za¨lan minn-ak ‘the one who’s angry with you’ (‘the-angry from-you’). That the relationship between (az-)za¨lan and minn-ak is syntactic is demonstrated by the possibility of (az-)za¨lan minn-ak u-minna-na ‘[the one who is] angry with you and with us’. The phrase minn-ak u-minna-na ‘with you and with us’ is clearly an expansion of minn-ak ‘with you’ (i.e. it involves the additional subsidiary elements minna-na ‘to us’ added to minn-ak ‘with you’), and the whole phrase is clearly syntactic — as is demonstrated by what we can take to be the grammatical difference between minn-ak uminna-na ‘with you and with us’ and minna-na u-minn-ak ‘with us and with you’. Semantically, in az-za¨lan minn-ak ‘the one who is angry with you’, the initial al-element has as its domain za¨lan minnak ‘angry with you’. Although this does not fall within the definition of syntax given earlier in this section, such semantic considerations may be thought of as providing another means of demonstrating syntactic relations between two elements. A further, and perhaps more robust, demonstration of the syntactic (rather than morphological) nature of the relationship between al- and za¨lan — or more accurately between al and za{lan minn-ak — is the fact that al- can have as its domain more than one adjective phrase. Thus, as well as it being possible to say (az-zol) az-za¨lan minn-ak ‘(the man) who’s angry with you’, it is also possible to say the following: 22.
(az-zol ) az-za¨lan minn-ak u-ga¨id bara-u (C) (the man) the-angry from-you and-sitting on his own (the man) who’s angry with you and is sitting on his own
Here the al- at the start of az-Âa¨lan has as its domain not only za¨lan (minn-ak) but the entire coordinated phrase za¨lan minn-ak u-ga¨id bara-u. The relationship between al- and this syntactic phrase is by definition itself syntactic. The following is an attested example involving active participles: ana sayif nas katirin aj-jay min al-man†iga u-labis hidum al-man†iga u-daxil be-ha aj-jami¨ ‘I see a lot of people the one coming from the [industrial/artisanal] area, and wearing the clothes of the area, and entering in them the mosque’ (Mafahim 1986)
It is also worth noting that the adverb siwayya ‘a little’ may be fronted before its adjective. Thus, it is possible to say not only za¨lan siwayya ‘a little angry’, but also siwayya za¨lan ‘a little angry’. The 553
RELATIVE CLAUSES IN SUDANESE ARABIC
definite correspondents of these are az-za¨lan siwayya ‘the one who is a little angry’ and as-siwayya za¨lan ‘the one who is a little angry’. This last form is at least suggestive that the relationship between aland za¨lan is syntactic rather than morphological (but cf. the discussion of ‘scrambling’ in Dickins forthcoming, Section 5). A number of other adverbs may also be similarly fronted. In the following example gibel ‘a while ago’ is fronted before the participle sayif ‘seeing/ saw’: axwan-na l-¨azfin del al-gibel sayfinna-hum ‘Those friends [lit: ‘brothers’] of ours who were playing, whom we saw a while back’ (Murkab ¨ala ¬¬a 1986)
If the relationship between adjective phrases of this kind and al- is syntactic, as might be expected the relationship between verb phrases and al- is also syntactic. Thus: 23.
(az-zol) az-zi¨il minn-ak u-ga¨ad bara-u (C) (the man) the-he- got-angry from-you and-sat on his own (the man) who got angry with you and sat on his own
The following is an attested example involving past tense verbs, in which the scope of the al- before laget-u ‘I found him’ extends to include also ¨irift-u ‘I recognized him’: †ab¨an ger aj-jana da ¨ind-ak al-kabir al-ana ba¨rif-u, al-laget-u ¨irift-u ‘Of course, apart from the child [whom we have spoken about], you’ve got the big one, whom I know, whom I have met [and] recognized’ (R)
Corresponding forms show that the relationship between al- and prepositional phrases is syntactic: 24.
(az-zol) al-ma¨a-k u-ma¨a-na (C) (the man) the-with-you and-with-us the man who’s with you and us
The following is an attested example with two adverbials (the second Ìaliyyan a borrowing from Standard Arabic): as-sugul al-hassi Ìaliyyan ‘the work which is [being done] now [and] at present’ (R)
The fact that predicand-predicate clauses are syntactic (I will take this to be self-evident for current purposes) demonstrates that such clauses preceded by al- involve a syntactic relationship between aland the entire following clause: 554
RELATIVE CLAUSES IN SUDANESE ARABIC
25.
(az-zol ) az-zi¨il minn-u babikir (C) (the man) the-got-angry from-him Babikir (the man) whom Babikir got angry with
Similarly, in cases where the predicate is something other than a verb: 26.
(az-zol ) az-za¨lan minn-u babikir (C) (the man) the-angry from-him Babikir (the man) whom Babikir is angry with
The fact that the relationship between al- and an adjective with prepositional-phrase expansion (object) is syntactic indicates — I believe — that the relationship between al- and any adjective is syntactic. (I will not, however, try to demonstrate this proposal in greater detail here.) I have considered the relationship between (i) al- and adjectives/ adjective phrases, (ii) al-and verbs/verb-phrases, (iii) al- and prepositions/prepositional-phrases, and (iv) al- and predicand-predicate structures. The major word-class relationship which I have not looked at is that between al- and nouns/noun phrases. I shall turn to this now. There is no independent reason which I have come across to define the relationship between al- and a following noun as syntactic. That is to say, there do not seem to be any cases in which the domain of al- in relation to a following noun extends beyond the noun itself. Thus, for example, while (az-zol) az-za¨lan minn-ak u-ga¨id bara-u ‘(the man) who is angry with you and sitting by himself ’ is grammatically correct, a putative form such as *al-muhandis u-tarzi ‘the engineer and tailor’ is not grammatically correct. There is, of course, a direct parallelism in the realisational (‘surface’) ordering of the elements in al-muhandis ‘the engineer’, az-za¨lan ‘the angry one’, az-zi¨il ‘the one who got angry’, al-ma¨a-k ‘the one who’s with you’, az-za¨lan/zi¨il minn-u babikir ‘the one whom Babikir got/is angry with’, and in the surface ordering of any other form involving al-: in all cases al- comes first. There are two obvious ways of interpreting the nature of al-+ noun. Either we could say that the direct evidence (as noted) indicates that this is a morphological complex. Or, we can use analogy, to argue that if al-phrases in all other contexts are syntactic, the simplest analysis of al-+ noun in terms of the overall description of the language is to regard this too as syntactic. In fact, a similar procedure has already been adopted to the analysis of al-+ adjective. I have argued above that expanded al-+ adjective-phrases — e.g. az-za¨lan minn-ak ‘the one who’s angry with you’ — involve a syntactic rela555
RELATIVE CLAUSES IN SUDANESE ARABIC
tionship between al- and the expanded adjective phrase. By analogy, I have suggested that al-+ adjective-phrases should be regarded as syntactic, even where the adjective cannot be expanded. I believe that the use of analogy makes better sense in the case of al-+ noun phrases than would the view that such phrases are, uniquely out of all al-phrases, morphological complexes. I will not attempt to argue this further here (though I believe it would be well worth pursuing in more detail), but will simply adopt the analysis of al-+ noun as a syntactic (rather than morphological) complex.13 14 13 There is in fact a third possibility which is worth considering, although it has relatively marginal status in axiomatic functionalism as applied up till now. Michael Lamb has pointed out that on logical grounds, just as one can have unordered sets whose members are simple entities (as is the case in morphology and phonematics; e.g. Dickins 2007: 7 10), so one can have unordered sets whose members are themselves unordered sets. Such unordered sets of unordered sets have been termed ‘ themes’ (see Dickins 1998: 142 3, 379, 380, 381 for further discussion and for mal definitions). Heselwood, in particular, has shown how the notion of phonotheme can be used to provide a minimally complex analysis of initial conso nant clusters in English, as unordered sets of phonemes, on the basis that the order ing of the realisations of these phonemes is entirely predictable (cf. Heselwood 1987, 2007). There seems to be no problem in set theoretical terms in extending the notion of a theme to include not only unordered sets whose members (sub sets) are unordered sets, but also sets whose members, or some of whose members, are themselves ordered sets. Just as a theme in phonology, is a phonotheme, so we may refer to a theme in grammar (syntax) as a ‘syntheme’. One striking feature of al is that it always occurs phrase initially. Its clearest fea ture in relation to other elements in the sentence is that of difference of scope (whether it is one word, two words, a multi word clause, etc.). In this respect, al looks rather different from a noun like ‘cats’ or ‘dogs’ in English: the syntactic dif ference between ‘dogs hate cats’ and ‘cats hate dogs’, for example, is not primarily evidenced in a difference of scope for ‘cats’ or ‘dogs’ in the two sentences. It may be that it is possible to analyse the relationship between al and the elements which follow it in most cases as synthemic. Thus, it may be possible to say that al bet u kabir ‘the one whose house is big’, for example is a syntheme consisting of two unordered members: 1. al and 2. bet u kabir (bet u kabir itself being, of course, an internally ordered set). If this analysis is acceptable, it is preferable to one in which al is analysed as the head of the phrase al bet u kabir, because unordered sets are simpler than ordered ones. One elegant corollary of the syntheme analysis of al in most contexts would be that if there is no independent reason for regarding nominal forms, such as ar rajil ‘the man’, as being syntactic, we could analyse ar rajil as morphological, i.e. as in volving an unordered set one of whose members is al . The parallelism in analysis of al and what follows it would thus remain despite the fact that al + NOUN would be a morphological complex and al + ANYTHING ELSE a syntactic complex. In all cases, the relationship between al and the following element would be one of non ordering. The notion of syntheme, if appropriate here, would need to be properly integrated with the notion of coordination (see Dickins forthcoming; footnote 8). 14 The closeness of the association between al and its following noun is under
556
RELATIVE CLAUSES IN SUDANESE ARABIC
The analysis which I have put forward here that the definite article al- is the head of its phrase obviously bears similarities with the view taken in Chomskyan minimalism that the Determiner is the head of a Determiner-Phrase (e.g. Radford 1997: 187–91), and has been proposed in a generative context by a number of writers (e.g. Ouhalla 2004: 289). Two things should be noted, however. Firstly, the analysis of al- as the head of an al-phrase which I have presented here should not in itself be taken to imply that an indefinite noun (e.g. rajil ‘man’), adjective (e.g. za¨lan ‘angry’), etc. has a zero indefinite head. Secondly, the approach which I have adopted here is language specific, rather than universalist. That is to say, I regard it as perfectly possible that in some languages what might be called ‘determiner phrases’ are in fact noun-phrases, i.e. phrases in which the noun is the head and the determiner is a non-head dependent element. I believe these issues need to be considered on a language-bylanguage, case-by-case basis. 10. Referentiality in Attributive Adjectives and Relative Clauses In Section 2, I suggested that agreement has its ‘origin’ in subjective referential identity. In reply to the statement, wen-u↑ ‘Where is it?’ (i.e. wen ‘where’ + -u↑ m.sg. suffix), we can ‘legitimately’ reply bitagÒudi Ò-Òaj? ‘Do you mean the cooking-plate?’ using the masculine word Òaj ‘cooking-plate’, but we cannot ‘legitimately’ reply bi-tagÒudi d-doka? ‘Do you mean the cooking-plate?’ using the feminine word doka ‘cooking-plate’ because of identity of ‘subjective reference’ in the case of the masculine singular pronoun suffix -u↑ and the masculine word Òaj ‘cooking-plate’ and non-identity of ‘subjective reference’ in the case of the masculine singular pronoun suffix -u↑ and the feminine word doka ‘cooking-plate’. I suggested that in the case of agreement proper, subjective reference becomes grammaticalized. Thus, *doka kabir ‘a big cookingplate’ (in which the masculine adjective kabir fails to agree with the feminine noun doka) is not just ‘disturbed’ but grammatically wrong. As is well known, agreement in Arabic dialects typically involves not only gender and number (in complicated permutations which I shall lined by the fact that in principle it is possible to have relative clauses in which an al whose scope is the entire clause is followed by al marking a definite noun, for example al filim al al walad saf u ‘the film which the boy saw’ (C). In practice such double al s are avoided by postposing of the definite subject in the relative clause: al filim as saf u l walad (C). 557
RELATIVE CLAUSES IN SUDANESE ARABIC
not explore here), but also definiteness. Thus focusing on definiteness, compare the following lists of indefinite and definite (all examples constructed by myself ): Indefinite agreement Noun+Adj.
27a. bet kabir 28a. rajil za¨lan Noun+Verb 29a. rajil zi¨il 30a. rajil bifham (al muskila) Noun+Prep. 31a. rajil fi l bet Noun+Predicand Predicate clause 32a. rajil bet u kabir
‘a big house’ ‘an angry man’ ‘a man who got angry’ ‘a man who understands (the problem)’ ‘a man [who is] in the house’ ‘a man whose house is big’
Definite agreement Noun+Adj.
27b. al bet al kabir 28b. ar rajil az za¨lan Noun+Verb 29b. ar rajil az zi¨il 30b. ar rajil al bifham (al muskila) Noun+Prep. 31b. ar rajil al fi l bet Noun+Predicand Predicate clause 32b. ar rajil al bet u kabir
‘the big house’ ‘the angry man’ ‘the man who got angry’ ‘the man who understands (the problem)’ ‘the man [who is] in the house’ ‘the man whose house is big’
In all the above cases, indefinite agreement (27a–32b) is exactly paralleled by definite agreement: an indefinite noun is followed by an indefinite anything else, and a definite noun is followed by a definite anything else. Indefiniteness and definiteness are features not just of entities in context but of aspects of entities in context. Thus in a context in which I utter ‘The waiter is a bachelor’ the person whom I am talking about should already identifiably be a waiter, but he should not identifiably be a bachelor. The ‘waiter’ aspect of this person is known, but the ‘bachelor’ aspect of him is not. I have argued elsewhere (Dickins 1998: 109–10) that linguistic referentiality involves two things: a referent (or referential entity) and an ascription. If, I say of someone, ‘Genius’, I have referred to this person: he or she, however conceived, is the referent (or referential entity). However, more than that I have ascribed him or her to a class of people (the class of geniuses). The same is true, mutatis mutandis, however I refer to this person, whether ‘Idiot’, ‘Fatty’, ‘Lonely man’, or whatever. The situation with linguistic reference contrasts with that of pointing or even at looking someone in such a way that it is clear who the object of my gaze is (e.g. by staring). Here, there is a referent — the person ‘picked out’ by pointing or looking. However, 558
RELATIVE CLAUSES IN SUDANESE ARABIC
there is no ascription: we are not assigning this referent to a particular class. Consider the case of an individual — we can call him ‘John’ — who I want to talk about. Depending on the context, I may, for instance, correctly refer to John as ‘a man’ or ‘the man’ (assuming John is in fact a man). Whether I refer to John as ‘a man’ or ‘the man’, the referent is the same: this particular individual. However, there is a difference in meaning. This must be a difference in ascription since it is not a difference in referent. The ascription of John in ‘a man’ and ‘the man’ is the same: he is ascribed to the class of men. The difference in meaning between ‘a man’ and ‘the man’ when referring to John must therefore be a difference in the ascription of ‘a’ and ‘the’. Various dichotomies have been proposed to characterize meaning differences between ‘definiteness’ and ‘indefiniteness’. These include uniqueness vs. non-uniqueness, familiarity vs. novelty, strength vs. weakness, specificity vs. non-specificity (cf. Abbott 2005). I will not attempt to assess which, if any, of these distinctions most coherently identifies the distinction between definite and indefinite, although three things seem worth noting here: 1. It may not be fruitful to search for a universal meaning distinction between ‘definite’ and ‘indefinite’. Some languages seem either only to make marginal use of this kind of distinction (e.g. Lyons 1999: 48). Other languages make distinctions which seem (intuitively at least) to involve definiteness vs. indefiniteness, but which focus on other aspects of meaning (cf. Lyons 1999: 53). 2. Even in the case of languages like English and Arabic which seem to make a similar ‘definite’ vs. ‘indefinite’ distinction, there are differences. English superlatives, for example, are definite, e.g. ‘the most beautiful city’, while Arabic ones are indefinite, e.g. Sudanese Arabic ajmal madina (sufta-ha) ‘the most beautiful city (which I have seen)’ (cf. also Abbott 2005: 134). Similarly, English abstract generics are indefinite, e.g. ‘beauty’, while Arabic ones are definite, e.g. Sudanese Arabic aj-jamal. One possibility, in this respect, is that definiteness, like gender, has a subjective referential element (sections 2, 3). 3. It is possible that ‘definite’ (also ‘indefinite’) in a given language may be polysemous. That is to say, two or more uses of the ‘definite’ (or ‘indefinite’) in a given language may be genuinely contrastive, and thus constitute different senses of ‘definite’.
Having noted these issues, I will, for the sake of practicality, assume that definiteness and indefiniteness in English and Arabic are roughly the same as one another. I will also in practice assume that in both languages definiteness is a monosemous phenomenon — i.e. 559
RELATIVE CLAUSES IN SUDANESE ARABIC
that in some sense we can talk about the (one) meaning of definiteness in English and the (one) meaning of definitness in Arabic. I will refer to this meaning itself as ‘definite’. Assuming also that indefiniteness has only one meaning (sense) in the two languages, I will refer to this meaning as ‘indefinite’. Where an adjective is associated with a noun, the adjective is dependent in various ways on the noun. This is most obvious in the fact that whereas nouns have fixed gender, adjectives take their gender from the noun to which they relate (whether the adjective is used attributively or predicatively, or bears some other relevant relationship to the noun). Another sense in which adjectives are dependent on nouns, however, is that where an adjective has a gradable (Cruse 1986: 204) sense (e.g. ‘small’ vs. ‘big’) the interpretation of what this means is dependent on the noun to which the adjective relates. Thus a small elephant is still a big mammal, while a big shrew is still a small mammal. Just as adjectives are ‘domain-specific’ in this way, so is the definite article, as discussed earlier in this section with respect to ‘The waiter is a bachelor’. This phenomenon has been referred to by Kamp and Partee as the head primary principle, which they formulate as follows: ‘In a modifier-head structure, the head is interpreted relative to the whole constituent, and the modifier is interpreted relative to the local context created from the former context by the interpretation of the head’ (Kamp and Partee 1995: 161). Returning now to the distinction between referent and ascription, whether I refer to a particular individual as ‘the waiter’ or ‘a waiter’, the referent is still the same. What is different referentially between ‘the waiter’ and ‘a waiter’ is the ascription — not of ‘waiter’, of course, which is the same in both cases, but of ‘the’ vs. ‘a’. In the case of ‘the waiter’, I am ascribing to the individual concerned the status of being ‘definite’ with respect to being a waiter (though not in the case of an utterance ‘The waiter is a bachelor’ with respect to being a bachelor). In the case of ‘a waiter’, on the other hand, I am ascribing to the individual concerned the status of being ‘indefinite’ with respect to being a waiter. This analysis works quite simply for Arabic nouns whether definite or indefinite. When I refer to an individual as aj-jirison ‘the waiter’, I am ascribing to the individual concerned the status of being ‘definite’ with respect to being a waiter; while when I refer to the same individual as jirison ‘a waiter’, I am ascribing to them the status of being ‘indefinite’ with respect to being a waiter. Where a noun is defined attributively (by an adjective or other element), the situation appears rather more complex. Consider the following: 560
RELATIVE CLAUSES IN SUDANESE ARABIC
33. 34.
jirison mumtaz aj-jirison al-mumtaz
‘an excellent waiter’ (C) ‘the excellent waiter’ (C)
I will assume here that in both of these examples, the adjective is being used restricively (I will deal with non-restrictive relative clauses in Sudanese Arabic in Section 12). In 33, the individual referred to is both a jirison ‘waiter’ and mumtaz ‘excellent’ as a waiter (i.e. in respect of being a waiter; by contrast he may, for instance, be terrible as — in respect of being — a father). In 34, the individual referred to is similarly both a jirison ‘waiter’ and mumtaz ‘excellent’ as a waiter. However, he is also definite, not only in respect of being a waiter (there may be many waiters in this particular restaurant, for example), nor only in respect of being excellent (there may be many people in this context who are excellent in one way or another). Rather he is definite in respect of being an excellent waiter (he is the only excellent waiter in this context). What this implies is that the definite particle al- applies simultaneously to both jirison ‘waiter’ and mumtaz ‘excellent’ in combination. This situation seems to be transparently reflected in a language like English, where the phrase ‘the excellent waiter’ involves only one occurrence of the word ‘the’. However in Sudanese Arabic aj-jirison al-mumtaz ‘the excellent waiter’ involves two occurrences of al-. This seems to suggest, incorrectly, that al- refers twice independently — once to the person in question being a waiter, and the second time to the person being the excellent one (in some general sense, not merely excellent in respect of being a waiter). This would, in fact, be the case if the adjective here were used non-restrictively; see Section 12 below. However, it is not the case in respect of restrictive adjectives. The solution to the problem of the reference — more precisely the ascription — of al- in aj-jirison al-mumtaz ‘the excellent waiter’ seems to be to recognize that al- in aj-jirison here applies to jirison ‘waiter’, as one might expect it would, but only in respect of this waiter also being excellent, and that al- in al-mumtaz similarly applies to mumtaz ‘excellent’ but only in respect of this excellent [one] being also a waiter. 11. The Demonstrative da/di/del as Secondary Relative Clause Marker So-called circumfixing (cf. Bauer 2003: 325–6) structures in which a single element, or complementary elements, occur at both the beginning and the end of a construction are quite common in languages. 561
RELATIVE CLAUSES IN SUDANESE ARABIC
Examples are ne … pas in French (je ne vois pas ‘I don’t see’), and ma … -s in many Arabic dialects (e.g. ma katabs ‘he didn’t write’ in Egyptian Arabic). One of the striking features of al- in Sudanese Arabic is its phonological non-prominence. Not only is al- not stressable, it also has numerous allomorphs, in addition to the citation form allomorph al- (Section 1). These either assimilate the ‘l’-element to the following consonant, as in ar-rajil ‘the man’, or they involve the loss of the initial ‘a’-vowel, as in safu l-walad ‘they saw the boy’, or they do both, as in safu r-rajil ‘they saw the man’. In each case, and most strikingly in the last, the phonological prominence of the definite particle al- is reduced. Demonstratives in Sudanese Arabic fall into two groups: neutral and distal, as follows: neutral da di del
‘this, that, he, it’ (m.sg.) ‘this, that, she, it, these, they (inanimate)’ (f.sg.) ‘those, they’ (pl.)
distal dak dik delak also delak
‘that (over there/a long time ago)’ (m.sg.) ‘that, those (inanimate) (over there/a long time ago)’ (f.sg.) ‘those (over there/a long time ago)’ (pl.) ‘those (far over there/a very long time ago)’ (pl.)
Demonstratives may occur non-attributively: Ì ag¨ud ma¨a hum sahren talata ma¨a ‘I’ll stay with them two or three months with this one, two or three da, talata arba¨a sihur ma¨a da … months with that one’ (Basariyya bint al Ìajj 1986) ‘what’s that thing called, what do they di isím a sinu↑, bisammu ha sinu↑? call it?’ (Îiraf sa¨biyya 1986) dak ta¨ban u da bi∂Ìak ¨ale hu! ‘that person [over there] is exhausted, and this one’s laughing at him!’ (R) gabul dak ‘before that’ (R) nas zayy del ‘people like them’ (R)
As this last example in particular shows, da, di and del fairly frequently correspond to English pronouns. I will not pursue this in detail in this paper. However, it is indicative of the greater degree of neutrality of da, di and del as compared to English demonstratives — a factor which turns out to be important in assessing ‘book-ending’ use of demonstratives (i.e. their use as the second element in a circumfix-type structure) in Sudanese Arabic, as we shall see. 562
RELATIVE CLAUSES IN SUDANESE ARABIC
As well as occurring non-attributively, demonstratives in Sudanese Arabic can occur attributively: an nas al kiwaysin del bar∂ u az zol dak, al gadim da … al wakit da inti Òagayra ya¨ni? funjan al gahwa da bagaddid ¨iyun ak del taji u tarja¨ fi yom ak da
¨irifta muÌammad da? nas aÌmad del al muluxiyya di bigulu le ha sin u↑ bi l injilizi
‘these/those nice people’ (C) ‘also that man, that long serving one … ’ (R) ‘at that time were you young, in fact?’ (R) ‘that coffee cup’ (Basariyya bint al Ìajj 1986) ‘I'll poke your eyes out’ (more lit: ‘I’ll poke out those eyes of yours’) (R) ‘you come [here] and go back [again] on the same day’ (more lit: ‘… on this/that day of yours’) (R) ‘Do you recognize Mohammed?’ (i.e. the Mo hammed who is here at the moment) (R) ‘the family of Ahmad [that we have been talk ing about]’ (R) ‘What do they call muluxiyya [ the subject of our discussion ] in English?’ (more lit: ‘what do they call this/that muluxiyya in English?’) (R)
These examples illustrate the general principle that Sudanese Arabic demonstratives are attributive to definite elements, whether these elements be made definite by the article as in an-nas al-kiwaysin ‘the nice people’, or by a construct genitive in which the final element is definite, as in funjan al-gahwa da ‘that cup of coffee’, or nas aÌmad del ‘the/that family of Ahmad’. Demonstratives are thus grammatically definite, and where they function attributively, they agree in definiteness with the preceding elements of the attributive structure. As these examples also suggest, the demonstrative almost always comes last in an attributive structure in which it occurs in Sudanese Arabic. A number of the above examples, particularly towards the end of the list, illustrate that the use of demonstratives in Sudanese Arabic may be rather different from ‘this’, ‘that’, ‘he’ (etc.) in English. The use of demonstratives to convey the notion of current or just previous mention or relevance, as in nas aÌmad del ‘the family of Ahmad [that we have been talking about]’ is particularly noteworthy. Very commonly in Sudanese Arabic, long attributive phrases or definite phrases whose first element is not a definite noun are concluded by a demonstrative. The following are examples: following an adjective/adjectival da kull u fi t talata yom aj jayat del
‘that’s all going to happen [lit: ‘be’] in the coming three days’ (R) 563
RELATIVE CLAUSES IN SUDANESE ARABIC
al manajil al mayta bita¨t al gass di al itnen al axirat del kiwaysat
‘the blunt scythes for [lit: ‘of ’] grass’ (R) ‘the last two [bottles] are okay’ (R)
following an adverbial ad dukkan al be jay da al awwal yom da
‘that/the shop which is up ahead’ (R) ‘the one who [came] on the first day’ [lit: ‘the first day [one] that’] (R)
following a combination of adjectival and adverbial aÒ Òuda¨ al amis as sadid da al labis abya∂ fi n nuÒÒ da
‘yesterday’s bad headache’ (R) ‘the one who's in the middle and wearing white’ (R)
following a verb/verb phrase ana mag†u¨ al lisan? samiÌ. awwal gal al bahayim al bita∂báÌ a di kulla ha mar∂ana u fa†sana.
‘Am I tongue tied? Okay. First of all, he said that the animals which you slaughter are all ill and on the verge of death’. (Xalt i↑ ∑afiyya 1986)
following a combination of phrases including a verb phrase hidum ak al galyat al bitalbas in le sugul your expensive clothes which you wear to ad dukkan del ‘work in the shop’ [lit: ‘for the work of the shop’] (Dukkan wadd al baÒir 1986) following a bipartite clause al ism u maryud da Ìatta law fi nas yajibu axbar bi Ò Òura al humma15 galo ha di aftakir ma kutar, jayiz takun waÌda wila itnen
‘the/that one who’s name is Maryud’ ‘even if there are people who bring news in the manner which they claimed [lit: ‘said it’], I think they are not many; maybe they are one or two’ (Basariyya bint al Ìajj 1986)
Two general tendencies seem to be apparent. The first is that where there is no definite noun head the use of a final demonstrative is common. Examples are: al-ism-u maryud da ‘the/that one who’s name is Maryud’, and al-awwal yom da ‘the one who [came] on the first day’. The second is that the longer and more complex the elements are following the definite noun head, if there is one, the more likelihood there is for the whole phrase to be ‘bookended’ through inclusion of a final demonstrative. In apparent contrast to the use of demonstratives with the definite particle al- in Sudanese to make up for the lightness of this particle, the brevity of the definite particle seems sometimes to motivate the 15 The more standard form for ‘they’ in Sudanese Arabic is hum (m. and com mon plural).
564
RELATIVE CLAUSES IN SUDANESE ARABIC
use of definites, particularly in questions, where other languages which have less concise means of marking definiteness make do with standard (indefinite) usages. Examples of definite-phrases in questions in Sudanese Arabic are the following: al gal kida min u↑? ‘who said that?’/’who was it who said that?’ (R) ‘what happened?’ (R) al ÌaÒal/al ÌaÒil sin u↑? lakin ba¨ad kida l biÌÒal sin u↑? ‘but what’s going to happen after that?’ (R)
12. The Use of al+Pronoun in Non-Restrictive Relative Clauses English makes a grammatical distinction between restrictive and nonrestrictive relative clauses. Thus, in the sentence ‘My brother who lives in Birmingham is a fire-eater [but my brother who lives in Wolverhampton prefers ice-cream]’, the element ‘who lives in Birmingham’ is a restrictive relative clause, and defines which brother is meant. By contrast, in ‘My brother, who lives in Birmingham, is a fire-eater’, the element ‘who lives in Birmingham’ is a non-restrictive relative clause, and is typically analysed as being parenthetical (or appositional), and as such not in a true syntactic relation to the preceding nominal ‘my brother’ (e.g. Burton-Roberts 1998). Parenthetical (or appositional) relations of this kind contrast with the true syntactic attributive relation which occurs in restrictive relative clauses. Intonationally, a nonrestrictive relative clause in English involves a separate intonation unit from the main clause (Cruttenden 1986: 78), may be in a lower key than the main clause (Cruttenden 1986: 129), and may be less loud than the main clause (Cruttenden 1986: 179). A restrictive relative clause, by contrast, is intonationally integrated as a single tone group with the major sentence element of which it is a part. While the distinction between restrictive and non-restrictive is clear in respect of relative clauses in English, it is much less clear with respect to adjectives appearing before a noun. There seems to be no clear intonational difference between ‘My intrepid brother is a fire-eater’ in the restrictive sense ‘My brother who is intrepid [as opposed to my brother who is a total coward] is a fire-eater’, and in the non-restrictive sense ‘My brother, who is intrepid, is a fire-eater’. This probably mirrors the lack of distinction in pre-nominal adjectives in English between properly syntactic attribution and non-syntactic apposition. Regardless of whether the intention is restrictive or non-restrictive, such pre-nominal adjectives are traditionally — and I believe correctly — regarded as attributive. 565
RELATIVE CLAUSES IN SUDANESE ARABIC
In Sudanese Arabic the lack of distinction between restrictive and non-restrictive elements seems to include not only simple adjectives but any other elements which can appear attributively, such as adjectival phrases, adverbs, adverbials, verbs, verb phrases, bipartite clauses, and perhaps even nouns. Thus the elements occurring after an-najjar ‘the carpenter’ in the following examples seem to all be interpretable either restrictively or non-restrictively: an najjar as sagga¬ (fi l Ìilla) (C) an najjar al bistagil (fi l Ìilla) (C)
‘the carpenter, who works (in the village)’ or ‘the carpenter who works (in the village)’ ‘the carpenter, who works (in the village)’ or ‘the carpenter who works (in the village)’
In such cases, the obvious analysis of the element following the noun is that it is attributive (i.e. that this is a proper syntactic relation of attribution), regardless of whether the sense is restrictive or non-restrictive. This leaves some interesting problems with respect to the semantic analysis of al-. I have argued above (Section 10) that in a standard restrictive usage, e.g. an-najjar as-sagga¬ ‘the carpenter who works’ (as opposed to the one who doesn’t), the two al-s — before najjar and before saggal — refer to al- and relate simultaneously to both najjar ‘carpenter’ and saggal ‘working/who works’ in combination. If an-najjar as-sagga¬ is non-restrictive, however, al- on annajjar should relate only to najjar, while al- on as-saggal should relate only to saggal (since the person in question is definite by virtue of being the carpenter on the one hand, and the one who is working on the other). This analysis could, however, prove to be problematic should there turn out to be cases where the relationship seems to be neither clearly restrictive nor non-restrictive. While the attribution structure in Sudanese Arabic can apparently have either a restrictive or a non-restrictive sense, there is one form which is unambiguously non-restrictive. This is illustrated by the following examples: 35.
36.
jib †unn u-rubu¨ asmanti al-hu xamsa u-¨isrin kis ‘fetch a ton and a half of cement, which is/that is twenty bags’ (R) ana Ìa-atkallam ¨an ¨ada sudaniyya al-hi aktar ¨ada bi-tit¨allag bi-s-sudaniyin ‘I am going to talk about a Sudanese custom, which is the custom which has most to do with the Sudanese’ (R)
In 35 the indefinite noun asmanti is followed by the definite particle al- plus the pronoun -hu, while in 36 the indefinite ¨ada sudaniyya is similarly followed by al-hi. It seems that in most non-restrictive cases 566
RELATIVE CLAUSES IN SUDANESE ARABIC
of this kind, a pronoun which is co-referential with the preceding noun occurs after the definite particle al-. In some cases, however, the same basic form occurs without a co-referential pronoun: 37.
u-ana wiÒilta le mara mashura hinak al-isma-ha ‘at-taska’ ‘I came to a famous women there, whose name is At-Tashka’ (R)
The following example, in which the non-restrictive-type element occurs in a separate sentence, is particularly interesting: 38.a. fi Ìaja muhimma jiddan ‘There is something very very impor jiddan. tant’ b. al-hi? ‘Which is?/Namely?/What?’ (R)
Examples 35–8 raise the question of how the al- is to be treated semantically in such non-restrictive environments. One interpretation starts from the obvious fact that when an element treated as unknown, etc. in the context is first introduced, it is indefinite. Thus in 37, for example, mara mashura ‘famous woman’ is indefinite. Once introduced, however, an indefinite element becomes known. Were this same mara mashura to be further discussed in the text, she would subsequently be referred to by a definite form such as al-mara ‘the woman’, or hi ‘she’. The use of al- to introduce the parenthetical/appositional phrase in examples 35–7 can be taken as reflecting the fact that once introduced an indefinite element is indefinite no more, but becomes definite, since it is now known about. Accordingly, when referred to by a subsequent parenthetical/appositional phrase, the form al- is used co-referentially with the preceding element (mara mashura ‘a famous woman’ in the case of 37). This situation contrasts with that of restrictive attributes, as described in Section 10 above. In the case of restrictive attributes, such as the the definite ajjirison al-mumtaz ‘the excellent waiter’ or the indefinite jirison mumtaz ‘an excellent waiter’, it is the overall meaning of the noun+attribute which is definite or indefinite: the person in question is definite or indefinite in respect of being an excellent waiter, not just in respect of being a waiter, or of being excellent. Fairly frequently non-restrictive relative clauses in Sudanese have definite antecedents. Examples are: 39.
bita¨rif an-nas kulla-hum min at-tob al-qawmi al-hu mitmassikin be-hu ‘One knows [lit: ‘you know’] everybody from the national tob [sari-like garment worn by women], which they are attached/ committed to’ (Îiraf sa¨biyya 1986) 567
RELATIVE CLAUSES IN SUDANESE ARABIC
40.
ya¨ni Ìikayat at-tasriÌ al-hi sayfin-a ¨inda axwat-na del … ‘that is, the issue/question of [Western-style] hair-dressing, which our sisters [i.e. fellow-women] see’ (Îiraf sa¨biyya 1986)
Relative bipartite clauses in which the predicand is a pronoun are not confined to non-restrictive usages. The following are examples of standard restrictive relative clauses with a predicand pronoun: 41. 42.
43.
axtar al-lon al-hi ¨awza-hu ‘I choose the colour which she wants’ (Îiraf sa¨biyya 1986) katab le-na l-ugniya be-nafs a†-†ariga al-hu biganni-ha ‘He wrote the song down for us in the same way as he sings it’ (Murkab ¨ala ¬¬a 1986) Ìatta law fi nas yajibu axbar bi-Ò-Òura al-humma galo-ha di aftakir ma kutar, jayiz takun waÌda wila itnen ‘even if there are people who bring news in the manner which they claimed [lit: ‘said it’], I think they are not many; maybe they are one or two’ (Îiraf sa¨biyya 1986)
In all of these cases, the predicand pronoun (hi ‘she’ and hu ‘he’, and humma16 ‘they’) seems to be somewhat emphatic (although this is not particularly apparent in the English translations). All examples are contrastive. In 41 hi ‘she’ points up a contrast between what ‘I’ do and what ‘she’ does. In 42 there is a contrast between ‘us’ and ‘he’, and in 43 the contrast is between ‘people (in general)’ and another group ‘they’ which is making certain claims about what people (in general) do. 13. Non-co-referentiality between al- and the Following Element in Relative Clauses Where the attribute following a definite head involves al- + noun, both this al- and the attribute noun are co-referential with the head noun phrase. An example is: az zol al muÒiba
‘the terrible man’ [lit: ‘the disaster man’] (R)
Here all of al- (before zol), zol, al- (before muÒiba), and muÒiba have the same referent. Where the attribute following a definite head involves al- + adverbial, both this al- and the attribute adverb are similarly co-referential with the head noun phrase: ar rajil al fi l bet 16
‘the man in the house’ (C)
See footnote 15. 568
RELATIVE CLAUSES IN SUDANESE ARABIC
Here all of al- (before rajil), rajil, al- (before fi-l-bet), and fi-l-bet have the same referent. That fi-l-bet is to be regarded as having the same referent (at least viewed globally) as rajil is made plain by consideration of the following: ar rajil fi l bet
‘the man is in the house’ (C)
Where the attribute following a definite head involves al- + verb, this al- and the attribute verb must be co-referential with the head noun phrase, as in the following example: al walad az zi¨il
‘the boy who got angry’ (C)
Here all of al- (before walad), walad, al- (before zi¨il), and zi¨il have the same referent: the person described as a boy is the same person who got angry. Where the attribute following a definite head involves al- + verb phrase, this al- and the attribute verb phrase may be co-referential with the head noun phrase, as in the following example: al walad az zi¨il minn u
‘the boy who got angry with him’ (C)
Here all of al- (before walad ), walad, al- (before zi¨il minn-u), and zi¨il minn-u have the same u referent. What I mean by saying that zi¨il minn-u here has the same referent as walad (and also as the two al-s) is that the subject of zi¨il ‘he’ has the same referent as walad: the person who is referred to as a boy is the same person who did the getting angry. Another somewhat more precise way of describing the nature of the co-referentially here is to say that zi¨il minn-u is globally co-referential with walad. Where the attribute following a definite head involves al- + verb phrase, there is, however, also another possibility, as follows: al walad az zi¨il minn u
‘the boy whom he got angry with’ (C)
Here al- (before walad), walad, and al- (before zi¨il minn-u) all have the same referent. They refer to the ‘boy’ in question. Zi¨il minn-u, however, does not have the same referent — or at least not the same global referent. The global referent of zi¨il minn-u is another person — the ‘he’ of the English translation. It is tempting to regard alwalad az-zi¨il minn-u in the sense ‘the boy who got angry with him’, and al-walad az-zi¨il minn-u in the sense ‘the boy whom he got angry with’ as two different syntactic structures. Although such an analysis might be right, my suspicion is that it is not: what we have in the case al-walad az-zi¨il minn-u in the sense ‘the boy who got angry with him’, and al-walad az-zi¨il minn-u in the sense ‘the boy whom 569
RELATIVE CLAUSES IN SUDANESE ARABIC
he got angry with’ is a single syntactic structure with two different referential interpretations. The situation with adjectives and adjectival phrases interestingly parallels that with verbs and verb phrases. Where the attribute following a definite head involves al- + adjective, this al- and the attributive adjective must be co-referential with the head noun phrase, as in the following example: al walad az za¨lan
‘the angry boy’ (C)
Here all of al- (before walad), walad, al- (before za¨lan), and za¨lan have the same referent: the person described as a boy is the same person who is angry. Where the attribute following a definite head involves al- + adjectival phrase, however, this al- and the attributive adjectival phrase may or may not be co-referential with the head noun phrase, as in the following example: al walad az za¨lan minn u OR al walad az za¨lan minn u
‘the boy who is angry with him’ (C) ‘the boy whom he is angry with’ (C)
Another example of this second type in which the al- at the start of the attributive clause is not co-referential with the adjective which follows it comes from a radio interview with a Sudanese traveller and explorer: fi ba¨a∂ ar rif as saggal fi hu ‘in a part of the countryside in which he is/was work ing’ (Balilat mibasir 1986)
Here the al- before saggal is co-referential with rif, while saggal refers to the person who is working in this rif. Where the attribute following a definite head involves al- + attributive bipartite clause, this al- and the attribute bipartite clause are almost never co-referential with the following noun phrase. Thus: ar rajil al bet u kabir
‘the man whose house is big’ (C) (lit: ‘the man the house his big’)
Here ar- (before rajil ), rajil, and al- (before bet) all refer to the same person. However, bet, obviously enough, does not refer to a person at all. Note, however, that bet-u kabir ‘his house is big’ is globally co-referential with rajil (and also with the two al-s) in this phrase. This is obvious if we think of bet-u kabir as meaning not just ‘his house is big’, but, meaning more precisely ‘one whose house is big’ (cf. Dickins forthcoming, Section 3.6) — i.e. ‘a [indefinite] his house is a 570
RELATIVE CLAUSES IN SUDANESE ARABIC
big one’, or ‘a/one whose house is big’ (and thus constituting the indefinite form whose definite correspondent is al-bet-u kabir ‘the one whose house is big’). The referent of ‘one whose house is big’ is clearly the same as the referent of rajil: they both refer to the same person. 14. Conclusion and Prospects This article has proposed a novel distinction between objective and subjective reference as a grounds for analysing the Sudanese definite particle al- in semantic terms. On this basis, it has gone beyond traditional, exclusively grammatical understandings of definiteness agreement in Arabic to consider referential correlates of such agreement. I have shown that Sudanese al- is a definite particle rather than simple a definite article, and have argued, principally on distributional grounds, that al- functions as the head of its phrase. I have also looked at the use of demonstratives in Sudanese Arabic as secondary relative clause markers, and considered the use of al- in nonrestrictive relative clauses. I believe — as I have at times made plain in this article — that many of the analyses presented here apply not only to Sudanese Arabic, but also to other dialects and to Standard Arabic. In a dialect such as Egyptian, il- and illi are both markers of definiteness, just as is al- in Sudanese Arabic. Sudanese is, however, beautifully transparent, in having the same definite particle al-, regardless of the element which follows it. Similarly, I believe that in other Arabic dialects, iland illi (etc.) are to be analysed as the heads of their phrases as are aland alla∂i- in Standard Arabic (assuming the articles il- and al- are syntactic in these varieties). Non-restrictive relative clauses in a number of dialects (though not in Standard Arabic) also seem similar to those in Sudanese. REFERENCES Abbott, B. 2005. ‘Definiteness and indefiniteness’. In Horn, L.R. and Ward, G. (eds), Handbook of Pragmatics. (Oxford), 122 50 Arnstein, R. 2005. ‘Coordination of parts of words’. Lingua, Vol. 14, Issue 4: 35 93 Badawi, E., M.G. Carter, and A. Gully, 2004. Modern Written Arabic: a Compre hensive Grammar. (London and New York) Balilat mibasir. 1986. Omdurman radio series presenting dialect poetry and prose. Basariyya bint al Ìajj. 1986. Omdurman Radio series (drama). Bauer, L. 2003. Introducing Linguistic Morphology. (Edinburgh) 571
RELATIVE CLAUSES IN SUDANESE ARABIC
Bohas, J., J. P Guillaume and D.E. Kouloughli, 1990. The Arabic Linguistic Tradi tion. (London) Burton Roberts, N. 1998. ‘Language, linear precedence and parentheticals’, in The Clause in English: in Honour of Rodney Huddleston (Studies in Language Com panion Series). P. Collins and D. Lee (eds), (Amsterdam), 33 52 Collins, P. and Lee, D. 1998. The Clause in English: in Honour of Rodney Huddleston (Studies in Language Companion Series). (Amsterdam and Phila delphia) Corbett. C.G. 2006. Agreement. (Cambridge) Corbett, C.G., N.M. Fraser, and S. McGlashan. 1993. Heads in Grammatical Theory. (Cambridge) Cruse, D.A. 1986. Lexical Semantics. (Cambridge) Cruttenden, A. 1986. Intonation. (Cambridge) Dickins, J. 1998. Extended Axiomatic Linguistics. (Berlin and New York) 2006, ‘Khartoum Arabic’, in Encyclopaedia of Arabic Language and Linguistics (Leiden) (vol. 2), 559 71 2007. Sudanese Arabic: Phonematics and Syllable Structure. (Wiesbaden) forthcoming. ‘Basic sentence structure in Sudanese Arabic’. Journal of Semitic Studies Dukkan wadd al baÒir.1986. Omdurman Radio series (drama) Eliade, M. 1959. The Myth of the Eternal Return. (New York) Haywood, J.A. and H.M. Nahmad. 1965. A New Arabic Grammar of the Written Language. (London) Heath, J. 2002. Jewish and Muslim Dialects of Moroccan Arabic. (London and New York) Hervey, S.G.J. 1979. Axiomatic Semantics. (Edinburgh) 1982. Semiotic Perspectives. (London) Heselwood, B. 1987. ‘Functional phonology: a functional approach to the descrip tion of initial clusters in English’, in Belfast Working Papers in Linguistics 9: 77 95 2007. ‘A positional analysis of schwa in English’, in Transactions of the Philo logical Society, 15: 2 Îiraf sa¨biyya 1986. Omdurman Radio series on traditional crafts Hjelmslev, L. 1963. Prologomena to a Theory of Language. (Maddison) Horn, L.R. and G.G. Ward (eds). 2005. Handbook of Pragmatics. (Oxford) Kamp, H. and Partee, B. 1995. ‘Prototype theory and compositionality’. Cognition 57: 129 91 Lyons, C. 1999. Definiteness. (Cambridge) Mafahim.1986. Omdurman Radio series (drama). Mulder, Jan W.F. 1968. Sets and Relations in Phonology. (Oxford) 1989. Foundations of Axiomatic Linguistics. (Berlin) Mulder, Jan W.F. and Hervey, S.G.J. 1972. Theory of the Linguistic Sign. (The Hague) 1980. The Strategy of Linguistics. (Edinburgh) Murkab ¨ala ¬¬a 1986. Omdurman Radio series on Sudanese folk music Ouhalla, J. 2004. ‘Semitic relatives’. Linguistic Inquiry 35: 288 300 Radford, A. 1997. Syntactic Theory and the Structure of English. (Cambridge) Saussure, F. de. 1959. Course in General Linguistics (translated by Wade Buskin). (New York) 1975. Cours de linguistique générale. (Paris) 572
RELATIVE CLAUSES IN SUDANESE ARABIC
1983. Course in General Linguistics (translated by Roy Harris). (London) Watson, J.C.E. 1993. A Syntax of ∑an¨ani Arabic. (Wiesbaden) Wise, H. 1975. A Transformational Grammar of Spoken Egyptian Arabic. (Oxford) Wright W. 1971. A Grammar of the Arabic Language (2 vols. re published as single vol.). (Cambridge) Yoda, S. 2005. The Arabic Dialect of the Jews of Tripoli (Libya). (Wiesbaden) Xalt i↑ ∑afiyya. 1986. Omdurman Radio series (drama) Zwicky, A. 1985. ‘Heads’. Journal of Linguistics 21: 1 29
573
REVIEWS PIER GIORGIO BORBONE, ALESSANDRO MENGOZZI and MAURO TOSCO (eds), Loquentes linguis. Studi linguistici e orientali in onore di Fabrizio A. Pennacchietti. Linguistic and Oriental Studies in Honour of Fabrizio A. Pennacchietti. Lingvistikaj kaj orientaj studoj honore al Fabrizio A. Pennacchietti. Harrassowitz Verlag, Wiesbaden 2006. Pp. xii + 751. Price: /148.00 hardback. ISBN: 978 3 447 05484 3. This massive, seven hundred pages plus volume celebrates the sixty eighth birthday of one of Italy’s foremost Semitists, Fabrizio A. Pennacchietti, whose intellectual vi vacity, multifaceted interests and outstanding philological expertise, are only matched by his unique niceness and great sense of humour. The three editors, former students of Professor Pennacchietti at the University of Turin are Pier Giorgio Borbone (University of Pisa), Alessandro Mengozzi (University of Bergamo) and Mauro Tosco (University of Naples). They all share the credit of hav ing gathered seventy three contributions written in six different languages, English, French, German, Italian, Spanish and Esperanto the last, one of the many areas of interest and study of Professor Pennacchietti, remains a bit beyond the reach of average loquentes linguis. The authors are seventy nine well known scholars and younger researchers in the field of Semitics and Afro Asiatic languages, specifically focused on linguistics, literature and comparative literature, and Esperanto. As the editors stress in their preface (p. xi), these three main categories mirror the wide ranging field of scholarly research conducted by Professor Pennacchietti, encom passing as it does the study of Semitic linguistics, both from a theoretical viewpoint as well as its applications; the study of religious and popular literature; philological analysis and contributions to the field of epigraphy including the publication of inscriptions and the study of Aramaic dialects, in particular, of Neo Aramaic, to cite only a few examples, as well as a personal engagement with the usage and diffu sion of Esperanto. The seventy three articles appear in alphabetical order. However, due to the wide spectrum and the heterogeneity of themes treated, readers of the volume would perhaps have profited from a structured subdivision of the papers. This subdivision could have been done, for example, according to the aforementioned areas of interest and research of Professor Pennacchietti, as outlined by the editors in their preface. Given the large number of contributors I will refer to the main areas covered and the authors, and cite only a handful of the articles. Thirteen papers focus on linguistic analysis of Semitic and Afroasiatic languages (Blazek; Goldenberg; Israel; Jungraithmayr; Lipinski; Mazzini; Mettouchi; Satzinger; Savà; Takács; Tosco; Voigt; Zaborski). Among these, Lipinski discusses dissimilation of geminated phonemes, citing many examples, including the well known case of Aramaic korse’ ‘throne’ from an older form kussi’u, corresponding to Sumerian gisgu zí ‘throne’. A case of dissimilation in Indo European according to Lipinski, namely the Italian bastone from batto ‘I beat’, which would have taken place in ‘in an ancient non Romanic language, spoken once in the Apennine Penin sula’ (p. 437), seems unconvincing. Italian bastone ‘stick’, from Latin bastone may be related to bastum ‘pack saddle’, a denominative from the verb bastare ‘to hold a weight’, in turn a loan from Greek bastázein ‘to hold, sustain’. It is unlikely that it is connected to the Italian battere ‘to beat’, which derives from Latin battuere, with 575
REVIEWS
the same meaning. Only battere and battuere may therefore fit in with the theoreti cal framework developed by Bohas and cited by Lipinski, in which ‘a labial and dental “matrix” like bt expresses the basic meaning “to strike, beat”’ (p. 438) in many Semitic and Indo European languages. Three articles discuss lexical and etymological data (Boucherit; Braida and Destefanis; Rossi); seven address Neo Aramaic issues, such as studies of Neo Ara maic dialects and discussion of Neo Aramaic liturgical documents (Borghero; Kapeliuk; Khan; Mengozzi; Murre van den Berg; Odisho; Poizat). Kapeliuk analy ses the contacts between Neo Aramaic and Russian speakers in Iranian Azerbaijan based on translations of Gorky’s, Turgenev’s and Neverov’s works done by Qonstantin Marogulov in the years 1934 and 1935. Khan offers a discussion of compound verbal forms in the Christian Barwar dialect of Northern Iraq. The same dialect, and specifically the study of Barwar oral poetry, was the object of a 1976 study by Pennacchietti (‘Zmiryata d rawe: “stornelli” degli aramei kurdistani’, Festschrift Giuliano Bonfante, pp. 639 63). Bridging between Neo Aramaic and Syriac, Mengozzi publishes a Neo Aramaic version of a Syriac Christmas hymn. Six papers deal with Syriac (Borbone; Briquel Chatonnet, Desreumaux and Binggeli; Ciancaglini; Kallas, Pazzini; Saccagno). Pazzini presents his ongoing project de voted to the marking of diacritics and vocalization of the Syraic text of the Twelve minor prophets, according to the Western system. Eight papers are devoted to Arabic (Arnold, Blau; Diem; Giolfo; Jastrow; Kaye; Noja; Vrajitoru). Four are on Hebrew language and Hebrew documents (Aspesi, Garbini, Martone, Tamani). Among these Garbini analyses the meaning and origin of the Hebrew noun sele†, connecting it to the Semitic root sl† ‘to have power, author ity’. Instead of the interpretations ‘quiver’, ‘shield’, ‘spear’ or the like, he suggests the more convincing ‘staff of command’, that is the ‘king’s spear’. As is well known, a form sly† ‘empowered’ is widely attested in the Completion of Transaction clause in Aramaic legal documents, sometimes found as an alternative to rsy, with the same meaning, in contracts from the Dead Sea (Yadin et al., 2002). Martone announces his project to prepare an Italian translation of the Bar Kokhba letters with commen tary, now available in their entirety thanks to the lastest addition to earlier text edi tions with French translation and commentary (Benoit, Milik, de Vaux), and with English translation (J.A. Fitzmyer, D. Harrington, A Manual of Palestinian Aramaic Texts, Rome 1978) the monumental 2002 volume by Y. Yadin, J.C. Greenfield, A. Yardeni and B. Levine, with hand copies by Ada Yardeni, extensive commentary and English translation, publication of which was delayed by the untimely death of the two original editors, Yadin in 1984 after a long illness, and Jonas C. Greenfield who passed away suddenly in 1995. Martone gives the texts (after the cited editions) and Italian translations of the Hebrew and Aramaic letters, as well as one in Greek featur ing Bar Kokhba as sender. Among these, in P.Yadin 54 is the mention of a Palmyrene ‘Yeshua¨, son of the Tadmorean’, perhaps belonging to an originally Palmyrene Jewish family. Likewise, the roughly contemporary funerary inscriptions of Beth She¨arim (Israel) attest to settlements of Palmyrene Jews outside Palmyra. Particularly enjoy able for Semitists with a knowledge of Italian is the article by Tamani on sixteenth seventeenth century poems whose text may be read, simultaneously, in Hebrew and Italian. This clever baroque device, created a genre first initiated by Leon Modena (1571 1648), and his successors, other notable Italian Jewish men of letters such as Mosheh Hayyim Catalan (d. 1642) and Ishaq Hayyim Cantarini (1644 1723). Two more articles address issues pertaining to other languages: Sumerian pho nology by Macelaru, Armenian loans in Romany dialects by Orengo. Six articles 576
REVIEWS
deal with Esperanto (Baldi; Barandovská Frank; Frank; Gobbo; Minnaja; Silfer); three discuss the work of Orientalists and the contributions of sixteenth and seven teenth century merchants and travellers to the study of Near Eastern languages and culture (Bellino; Galletti; Lusini). Bellino writes on the work and teaching of David Morkos and the rise of Arabic and Semitic studies at the University of Turin; Galletti gathered and discussed the notes on Kurdish language by Italian travellers such as Gerolamo and Gian Battista Vecchietti, Pietro della Valle, and later the works by historians and philologians such as Lanza, Garzoni and others, who are considered among the founders of Kurdish studies. Ten articles focus on comparative literature (Albrile; Brugnatelli; Canova; Catastini; Khoury; Larcher; Marrassini; Piemontese; Tottoli; Tresso and Rivalta); three address the study of ancient societies or rituals (Fronzaroli and Catagnoti; del Olmo Lete; Roccati). Four papers discuss iconographic issues (Compareti; Cristo foretti; Favaro; Scarcia, Pellò). Finally, four papers present editions or investigation of inscriptions and documents: Amadasi Guzzo publishes a three line partly broken Phoenician dedicatory inscription to the god Melqart discovered in Ibiza in 2003 (accompanied by a clearly readable photograph); Avanzini discusses a Sabaic dedi catory boustrophedon nine line inscription (the quality of the photograph is poor) already published by Arbach in 2005. With the help of ‘various optical instruments’ and photographs Moriggi analyses a fifth century CE trilingual silver phylactery, in Greek, Hebrew and Aramaic, discovered in the 1874 8 excavations in the Mount Esquilino area in Rome, and recently restored. However, only a hand copy of the inscription accompanies the article. Finally, Crevatin presents a palaeographic analysis of a second century CE Egyptian papyrus of unknown provenance, in the Museum of the Etruscan Academy in Cortona, Italy, depicting the deceased before Osiris, with a series of hieroglyphic captions to accompany the images and a twelve line hymn in hieratic script (with photographs reproducing details). The volume would have benefited from a complete bibliography of Pennac chietti’s works, although skimming carefully through the individual bibliographies accompanying the articles, one may compile about thirty one titles (some cited sev eral times): two monographs and twenty nine articles, which represent only a por tion of Pennacchietti’s scientific production. Moreover, an additional, useful tool for users of the volume would have been a general bibliography, rather than only individual bibliographies. As for other minor shortcomings, one would have greatly profited from the presence of indexes, a general index, and other thematic indexes. These are only minor faults in a pleasing contribution dedicated to a scholar who, with much generosity, created a school, a scholar who continues scientific discus sion and exchange with his former students, now in turn well established figures in Italian academia. This bond is witnessed by the fond, grateful words they offer to their teacher. doi:10.1093/jss/fgp014
ELEONORA CUSSINI UNIVERSITÀ CA’FOSCARI VENEZIA
KEVIN J. CATHCART (ed.), The Correspondence of Edward Hincks. Volume 1 (1818 1849) University College Dublin Press, Dublin 2007. Pp.xii + 352. Price: /60.00/£45.00 hardback. ISBN: 978 1 904558 70 5. That Edward Hincks was a man of true genius to whom the basic decipherment of Akkadian cuneiform script and language should now be credited, is no longer in doubt. From much of his correspondence, deposited in the Griffith Institute in Oxford by his grandson, together with other letters tracked down in various institu 577
REVIEWS
tions in London, Dublin, Paris and Berlin, a picture emerges of extraordinary men tal energy and dedication in this Irish clergyman, with no lessening of drive and application as he grew older, a man who not only engaged with Egyptian hiero glyphs, Old Persian, Urartian and Akkadian cuneiform scripts and the languages they conveyed, but also found time for his parish duties. Far from being a self absorbed recluse dedicated only to study, Hincks made great efforts to obtain relief for his people when the Irish famine of 1846 8 struck. As an Irishman of outstand ing and precocious intellect, perennially short of money and isolated in Killyleagh, he was not a man to charm his peers and was unable to disguise his sharp criticism of other scholars, leading to frustrations early in his career. London, with its essen tially English institutions in the British Museum and the Royal Asiatic Society, was the main centre for scholarly interaction, publication and lectures, much more than Dublin. But his correspondence with several of the most eminent scholars of his day, including some in France, Germany and Italy, shows that his calibre was appre ciated very widely during his lifetime. It still astonishes Assyriologists that he was able to work out both the polyphonic and the logographic nature of Akkadian cuneiform script as well as making great strides in understanding the language. Kevin Cathcart held a professorship of Near Eastern languages at University College Dublin, and has long taken an interest in reviving the reputation of Hincks, as he demonstrated in his 1994 editing of The Edward Hincks Bicentenary Lectures. His biographical introduction in that volume is largely reworked for this book, with additional material arising from these letters, on two fronts: giving more information that Hincks was active in organizing relief committees in his district; and pinpointing the problems with Sir Henry Creswicke Rawlinson. He notes that although Hincks’s reputation as decipherer was acknowledged around the time of his death, it was subsequently diminished in favour of Rawlinson, whose fame was extolled by his brother, the ancient historian George Rawlinson, and by his col league Sir Wallis Budge in the latter’s book of 1925, The Rise and Progress of Assyriology. Budge positively denigrated Hincks, although R.W. Rogers had given Hincks his due in the Encyclopaedia Britannica (11th ed. 1911, s.v. cuneiform), and in his 1915 History of Babylonia and Assyria (6th ed., 225 44). In contrast to the social isolation of Hincks, Rawlinson moved in London’s high society and was able therefore to excite public interest in Assyriology alongside his own contributions to it. Cathcart himself is not a cuneiformist, but has had the benefit of researches by two Assyriologists who have come to the same conclusions largely independent of each other: Peter Daniels in the Bicentenary Lectures and elsewhere; and Mogens Trolle Larsen in his book of 1996 The Conquest of Assyria, and elsewhere. At the time when various scholars in Germany, France, England and Ireland were all com peting to be the first decipherers, the stakes were high, and Hincks with his pastoral duties, Rawlinson with his military, diplomatic and social affairs, both had to make time in busy lives for their researches. Getting access to new discoveries as they were published was sometimes an uncertain business, allowing Rawlinson to claim that he had made a breakthrough independently, when records later available show that his claims were dubious or untrue. Hincks was justifiably upset, and must have spent a precious part of his income trying to obtain copies of what he needed as quickly as possible. There is no doubt that his year’s work at the British Museum in 1853 4 produced a manuscript which remained unpublished but was divulged to Rawlinson, who persuaded the Trustees to deny Hincks further access to the collec tion. This episode, described in the Introduction, has also been thoroughly uncov 578
REVIEWS
ered thanks to the letters which will appear in the forthcoming second volume, and was not described in the introduction to the 1994 Bicentenary Lectures. That this type of friction was not restricted to Hincks and Rawlinson is evident in this vol ume, for example, in a letter written by George Cecil Renouard (p. 231), in which Grotefend’s apparent attempt to crib from Lassen and Burnouf is claimed. Much of the correspondence from the earlier part of Hincks’s career, in this first of three projected volumes, concerns his engagement in the decipherment of the Egyptian script and language. His early efforts in Dublin were largely frustrated by lack of general interest there in Egyptology, but his early correspondence shows how energetically he pursued his researches, perhaps invigorated by being thwarted, and it is engaging to read how openly he co operated with other scholars, anxious to ‘enlist recruits for Egyptian pursuits’. His interest in establishing the chronology of Egyptian dynasties and of certain Mesopotamian kings, and his understanding that certain recorded eclipses must underpin any attempt to create a framework, are the subject of letters especially written to William Hamilton, Professor of As tronomy at Trinity College Dublin. Hincks’s work on Egyptian inscriptions, to which many of the early letters refer, is nowadays best acknowledged on the fore court of the Cairo Museum, where he takes his place alongside Champollion and others. Another interesting correspondent was William Henry Fox Talbot, whose ener gies were directed towards decipherment as well as the development of photogra phy. The letters of the excavator of Nineveh, Austen Henry Layard, show how he came to appreciate the efforts and prompt publications of Hincks, and made the effort to pay him a visit in Killyleagh. The index facilitates a search for particular correspondents. Some of the terms used by Hincks and his contemporaries have changed their meaning, or else different terms are now used, such as enchorial for demotic, Me dian for Elamite, which was also called Second Persepolitan; First Persepolitan for Old Persian; Vannic for Urartian. The notes which Cathcart provides for each letter are models of precision, giving details of person referred to, and of previous publi cations relevant to an episode, throwing light on background of the different corre spondents. This fine volume, soon to be followed by the second, should help to restore the reputation of Hincks by showing his relationship with Rawlinson in a clearer light, with the letters of Layard especially revealing in this respect. It was most unfortu nate that the details that emerged from the work of Daniels and Larsen, which are further clarified in these letters, were not used at all by Lesley Adkins in her recent biography of Rawlinson, Empire of the Plain (2004), nor was the revised 1999 edition of Maurice Pope’s The Story of Decipherment (1999) able to change its main text attributing the decipherment of cuneiform very largely to Rawlinson, although his bibliography incorporated both Cathcart’s bicentenary volume of 1994 and Larsen’s book of 1996. We must hope that Cathcart’s painstaking work, worthy of its admirable subject, will be used by future scholars who look into those exciting days of discovery, the struggles of decipherment, and the race for recognition. Postscript. Vol. 2 is now published, and vol. 3 is in press. doi:10.1093/jss/fgp015
STEPHANIE DALLEY UNIVERSITY OF OXFORD
579
REVIEWS
MARTHA ROTH (ed.), The Assyrian Dictionary of the Oriental Institute of the Univer sity of Chicago Volume 18 Letter T (Assyrian Dictionary 18). Oriental Institute, Chicago 2006. Pp. xxx + 500. Price: $95.00 hardback. ISBN: 1 885923 42 2. MARTHA ROTH (ed.), The Assyrian Dictionary of the Oriental Institute of the Uni versity of Chicago Volume 19 Letter ™ [Tet] (Assyrian Dictionary 19). Oriental In stitute, Chicago 2006. Pp. xxxii + 167. Price: $75.00 hardback. ISBN: 1 885923 43 0. As the CAD nears completion the concluding volumes are appearing at an impres sive rate, now with these the ante penultimate and penultimate volumes. Volume T is far lengthier than ™ but the latter contains much anticipated entries such as †idu ‘clay’, †uppu ‘tablet’ and †upsarru ‘scribe’. tabalu: now attested at Mari, ARM XXVIII 10: 14: Ìa la aÒ [n]a ri im ù ta ba li im. tadu: see further J. M. Durand, ‘La divination par les oiseaux’, in Mari: annales de recherches interdisciplinaires 8 (Editions recherche sur les civilisations, Paris 1997), pp. 281 2. ta’tamu: see now D. Fleming, Democracy’s Ancient Ancestors: Mari and Early Collec tive Governance (Cambridge 2004), esp. pp. 207 8. tabilam: see now J. Fincke, Augenleiden nach keilschriftlichen Quellen: Untersuch ungen zur altorientalischen Medizin (Wurzburger medizinhistorische Forschungen 70. Wurzburg 2000), pp. 210 11. tahapsu: for the meaning ‘felt’, see Postgate, ‘Assyrian felt’, in P. Negri Scafa and P. Gentili (eds), Donum Anatolicum: studi presentati a Claudio Saporetti in occasione del suo 60. compleanno (Rome 2000), pp. 213 17. taÌlaptu: see S. Cole, OIP 114, no. 86: 16: túgtaÌ lap ti, attributed there (p. 184) to taÌluptu. taÌsâtu: additional attestations can be found in SAA IV 32: 6’ ([e zib sá KÚR 15 sú u 150 sú i na DAG].MES táÌ sa a ti ás bu ma ‘[Disregard that enemies] lie in [am]bush [at his right and left]’) and parallel 115: 5’. tahhitu: for the meaning ‘additional contribution’ see ARM XXVI/1 p. 191 note b to no. 47: 8. takkassu: for the use as a weight see H. Freydank, Mittelassyrische Rechtsurkunden und Verwaltungstexte IV. Texte aus Kar Tukulti Ninurta (Wissenschaftliche Ver offentlichungen der Deutschen Orient Gesellschaft 99. Wiesbaden 2003), 114: 14’ 15’. taktimu: also attested at Ugarit: M. Yon and D. Arnaud (eds), Études ougaritiques (Ras Shamra Ougarit 14. Paris 2001), p. 245 ad 88.2158: 37’: tak tim mu). talalu: on the meaning ‘form an alliance’ see now D. Charpin, ‘Sapîratum, ville de Suhûm’, in Mari: annales de recherches interdisciplinaires 8 (Paris 1997), p. 365 with refs. *tamsihu: for a new lemma ‘measurement’, see A. George, The Babylonian Gilgamesh Epic: Introduction, Critical Edition and Cuneiform Texts (Oxford 2003), p. 782 (ad I 23). tappûtu: for the meaning ‘party wall’ in CCT 3 20: 30 see C. Michel, Correspondance des marchands de Kanish au debut du IIe millenaire avant J. C. (Litteratures anciennes du Proche Orient 19. Paris 2001), p. 435 (translation Veenhof ). *taqdisu: add this word meaning ‘purification’; see K. Hecker et al., Kappadokische Keilschrifttafeln aus den Sammlungen der Karlsuniversitat Prag (Prague 1998), p. 265 ad 681: 26. 580
REVIEWS
taraqu: for the meaning ‘to relent’ see George, loc. cit., p. 888 ad XI 131 tardennu, terdû: cf. J. M. Durand, Florilegium marianum VII: le culte d’Addu d’Alep et l’affaire d’Alahtum (Memoires de N.A.B.U. 8. Antony: Societe pour l’etude du Proche orient ancien, 2002), p. 163 note to 10’’: ti ir di im tarimu: L. Sassmannshausen, Beitrage zur Verwaltung und Gesellschaft Babyloniens in der Kassitenzeit (Baghdader Forschungen 21. Mainz am Rhein 2001) p. 411 provides another attestation of this rare word: 1 mina of paint/glue for a gis[IG ta] ri m[i]. tarmitu: this rare word is now attested also at Mari: A.4280: 7 (D. Charpin, ‘Nomades et sédentaires dans l’armée de Mari du temps de Yahdun Lîm’, in C. Nicolle (ed.), Nomades et sedentaires dans le Proche Orient ancien: compte rendu de la XLVIe Rencontre assyriologique internationale, Paris, 10 13 juillet 2000 (Amurru 3. Paris 2004), pp. 83 94): […] ta ar mi tam la i ra [as si] ‘so that it (a contingent of troops) may not have permission to return home’. *tarrapitu: add this hapax from Ekallte, W. Mayer, Tall Munbaqa Ekalte II: die Texte (Ausgrabungen in Tall Munbaqa Ekalte 2. Wissenschaftliche Veroffent lichung der Deutschen Orient Gesellschaft 102. Saarbrucken 2001), 16: 11 and p. 10, apparently meaning ‘dyeing factory’. tassaku: for a possible Old Assyrian attestation see Michel, op. cit. p. 430 301d. tassistu: a further attestation (written ta as sí is tum) is found in BM 29638: 18 (N. Wasserman, ‘BM 29638: A New Ritual to Marduk from the Old Babylonian Period’, in ZA 96 (2006), pp. 200 11). tasimtu A: a 1’ b’) for the interpretation as a weighable object see Michel, op. cit. p. 313 214a. *tazziqu: add this lemma, attested once in an OB incantation; see A. George, ‘tazziqum, <
>”, NABU 1994: 27. têbu: for the meaning ‘distrust’ see T. Richter, ‘Akkadisch têb/pum’, NABU 1991 no. 46. tenisu: cf. [te] e ne es15 na ak ri su ‘the population of his enemies’, A. Shaffer and N. Wasserman, ‘Iddi(n) Sîn, King of Simurrum: A New Rock Relief Inscription and a Reverential Seal’, in ZA 93 (2003), pp. 1 52 ii 2 3. *teptêtu: add this word for an object made of precious materials; see Sassmanns hausen, op. cit. p. 428. terinnu: ad mng. 2 cf. K. Hecker, ‘Schultexte vom Kültepe’, in M. Mellink, E. Porada and T. Ozguç (eds), Aspects of Art and Iconography: Anatolia and its Neighbors: Studies in Honor of Nimet Ozguç (Ankara 1993), p. 290. tilmuttu: in favour of an association with Dilmun see D. Charpin and J. M. Durand ‘Notes de lecture: Texte aus dem Sînkasid Palast’, Mari: annales de recherches interdisciplinaires 7 (Paris 1993), p. 371 ad no. 113: 4. timiras: the colour applies also to cows; see Sassmannshausen, op. cit. p. 378. * tim’u: Hecker et al., loc. cit. p. 309 ad 735:21’ suggest derivation of a word writ ten tí im a am from tm’ ‘to swear’. tiqnu: ad mng. 2 see F. Zeeb ‘Das teqnitu in den Immobilienkaufurkunden aud Alalakh VII’, in M. Dietrich and O. Loretz (eds), Vom Alten Orient zum Alten Testament: Festschrift für Wolfram Freiherrn von Soden zum 85. Geburtstag am 19. Juni 1993 (Alter Orient und Altes Testament 240. Kevelaer / Neukirchen Vluyn 1995), pp. 541 9. tiqqûtu: cf. M. Geller, Renal and Rectal Disease Texts (Die babylonisch assyrische Medizin in Texten und Untersuchungen VII. Berlin 2005), no. 51 obv. 11’: a bar ra = te qú ú tu4 581
REVIEWS
**tiridanum: the Sumerian is actually ti rí da nu déhi(UM); see M. Civil, ‘The Forerunners of Marû and Îam†u in Old Babylonian’, in Riches Hidden in Secret Places. Ancient Near Eastern Studies in Memory of Thorkild Jacobsen (ed.) T. Abusch, pp. 63 71 (Winona Lake, Indiana 2002), pp. 65 8. tubqu: for the meaning ‘harem’ see J. M. Durand, ‘Documents pour l’histoire du royaume de Haute Mesopotamie II’, Mari: annales de recherches interdiscipli naires 6 (Paris 1990), p. 293 note c. tubuqtu B: perhaps more general in meaning than tubqu (‘harem’), referring to the palace interior as a locked space; see B. Lafont ‘La correspondance de Mukannisum trouvée dans le palais de Mari nouvelles pièces et essai d’évalua tion’, in D. Charpin, J. M. Durand (eds), Florilegium marianum VI: recueil d’etudes a la memoire d’Andre Parrot (Memoires de N.A.B.U. 7. Antony: SEPOA, Societe pour l’etude du Proche orient ancien, 2002), p. 373. tugurru: for this as a rural building with probably a non residential function, see L. Mori, Reconstructing the Emar Landscape (Quaderni di geografia storica 6. Rome 2003), pp. 65 70. tultu: ad 1b the worm affects also the eyes; see Fincke op. cit., p. 211. tuppanuru: this official also appears in RS 92.2007 (tup pa at nu ri/ra); Yon and Arnaud, op. cit., pp. 260 1. turtu B: refers to the restoration of leftover offerings to the despatching institution; see D. Fleming, Time at Emar: the Cultic Calendar and the Rituals from the Diviner’s Archive (Mesopotamian civilizations 11. Winona Lake, IN 2000), pp. 71 2. turuÌtanu: see now also Hecker et al., loc. cit. p. 266 ad 682: 4’ tù ru ùÌ ta nam KAS ‘container of beer’. tuÒâtu: for the meaning ‘sortie’ (in the phrase tuÒâtam wussurum ‘to attempt a sor tie’ [or simply ‘let flee’?]) see J. M. Durand ‘Le défaite de Zûzû, roi d’Apum’ in Mari: annales de recherches interdisciplinaires 5 (Paris 1987), p. 622 note b. †apalu: also attested in OA; see E. Bilgic and C. Günbatti, Ankaraner Kültepe Texte 3 (FAOS Beiheft 3. Stuttgart 1995), p. 150 ad no. 88: 55: e ta †í pì il5. †uÌÌû A: P. Vargyas, A History of Babylonian Prices in the First Millennium BC (Heidelberger Studien zum Alten Orient 10. Heidelberg 2001), pp. 40 1 sug gests that this term refers to scrap silver, examined and tested by a smith to ascer tain that the quality was sufficient for use as money. †uppi: see now H. Freydank, ‘†uppu in anderer sicht’, in AoF 34 (2007), pp. 225 36. †upsarru b3’a’ †upsar ekalli: on this figure see now M. Jursa and E. Payne, ‘Cuneiform Tablets in the Bristol Public Library’, JCS 57 (2005), pp. 120 3. †upsarratu: on female scribes see now B. Lion and E. Robson ‘Quelques textes scolaires paléo babyloniens rédigés par des femmes’, in JCS 57 (2005), pp. 37 54. These two volumes contain 15% (T) and 5% (™) words recognized as of non Akkadian origin. Almost all of those in ™ are loans from Sumerian words built around dub ‘tablet’; note also †al†altu ‘fugitive woman’ and †ulluma’u ‘liar, cheater’ (< Aramaic). As for T, around 5% of words come each from Sumerian and Hurrian, with a further 3% of indeterminate ‘foreign’ origin. Of the Semitic words, the fol lowing are labelled as ‘Aramaic’: tamimu ‘perfect’, taraÌu ‘door’, tensû ‘(a metal orna ment)’. There is also West Semitic tariru ‘(a profession)’ and Egyptian tawi ‘the two lands’, tiban ‘(an Egyptian measure of weight)’ and tinida ‘(a bowl or a brazier)’. Another two essential volumes of lasting value. doi:10.1093/jss/fgp016
JONATHAN TAYLOR BRITISH MUSUEM
582
REVIEWS
LISA M. WRAY BEAL, The Deuteronomist’s Prophet: Narrative Control of Approval and Disapproval in the Story of Jehu (2 Kings 9 and 10) (Library of Hebrew Bible/Old Testament Studies 478). T & T Clark, New York and London 2007. Pp. viii + 216. Price: £65.00 hardback. ISBN: 978 0 5670 2657 6. The account of Jehu’s revolt in 2 Kings 9 10 is one of the most gripping stories in the Hebrew Bible. It is also, according to Lisa M. Wray Beal, an important contri bution to the theology of the Deuteronomistic History, composed to express ap proval of Jehu for fulfilling the prophetic judgment on the dynasty of Ahab and for eradicating Baal worship in Israel, but disapproval for perpetuating the cultic inno vations of Jeroboam. She sets out to demonstrate both the artistry and the inten tions of the author through a narrative analysis of the story. Her narrative analysis assumes that the story is a coherent whole, any putative sources having been woven into the text as we have it. It pays attention to both the form and the content of the various scenes that make up the story. Thus, in the analysis of the opening scene (2 Kgs 9:1 4), the designation of Elisha (unusually) as ‘the prophet’ signals that the fulfilment of prophecy will be a dominant theme: the command to the young prophet to anoint Jehu and the subsequent disappearance of Elisha from the story indicate that the prophecies to be fulfilled will be those given through Elijah, an impression confirmed later in the story at 9:25 6, 36 7; 10:10, 17. The succession of imperatives emphasizes the authority of the prophet and the urgency of the mission, and the injunction to secrecy builds up suspense. Doubts about the procedure begin to arise, however, when the judgment placed in the young prophet’s mouth (9:7 10) is treated as an integral part of the story rather than as a redactional element, inserted as a link to the wider story of the northern kingdom (1 Kings 11 2 Kings 17). The Jehu story from 9:11 onwards (apart from 9:36 7) is free from the formulas which mark the judgments on previ ous dynasties (1 Kgs 14:7 11; 16:1 4; 21:20 4), and their execution by earlier usurpers (1 Kgs 15:29 30; 16:11 13), but 2 Kgs 9:7 10 is couched in very similar terms. On the other hand, the Jehu story has its unique features: it contains a dif ferent version of the oracle against Ahab (9:25 6) and, as Wray Beal herself points out, the play on different meanings of the word שלוםis without parallel in the Deuteronomistic History. The attempts to explain the young prophet’s speech oth erwise than as redactional addition is reminiscent of the way difficulties in the bib lical text were explained in pre critical days. That the text is a coherent whole is not a conclusion reached as the result of the analysis, but an arbitrary decision, taken in advance of the analysis and dictating the way the story is read. Wray Beal’s contention that the aim of the author was to show that God ap proved of Jehu for fulfilling the prophetic judgment on the dynasty of Ahab and for eradicating Baal worship from Israel, but disapproved of him for not departing from the wrong course taken by Jeroboam, seems to be based more on the sum mary of Jehu’s reign in 10:28 31 than on the analysis of the preceding story. That Jehu is presented as consciously fulfilling prophecy may be granted, on the evidence of 9:25 6 and possibly 10:10; but the eradication of Baal worship, though a con cern of Elijah, is nowhere explicitly prophesied. And the story is completely silent on the cultus established by Jeroboam. On the question that exercises modern read ers did the author approve of the methods Jehu employed: conspiracy, decep tion, murder, degradation? Wray Beal is disappointing. Although she insists that she is focussing on the author rather than the reader, it is difficult to imagine that the author and his original audience had no misgivings about whether the end jus tified the means. That surely is what makes the story so compelling. 583
REVIEWS
A word must be said about the presentation of the work. It evidently originated as a Ph.D. thesis, and one that undoubtedly deserved the award. Wray Beal cer tainly accomplished what she set out to do; the analysis is thorough and has yielded some interesting insights, whatever may be thought about the overall argument; the background reading has been extensive and the findings and views of other scholars are presented accurately and discussed fairly. But too much of the scaffolding still remains: ‘This study will show …’ must be the most repeated phrase in the book. It is ironical that one of the most riveting stories in the Bible should give rise to rather unexciting prose. A final criticism applies to the whole series to which the book belongs. There is an index of references and an index of authors but no subject index. I know that a subject index is more difficult to prepare, that it cannot be done mechanically but requires judgment, but it would have been a great help in navigating one’s way back through the book. For example, I have noted references to ׁלוםon pages 78 9, 81 3, 88, 93, 99, 122 (where we are told that it occurs eleven times in the Jehu narra tive), 156 67, 189 and 194. But how can I be sure I have picked up all the scat tered references? doi:10.1093/jss/fgp017
ROGER TOMES UNIVERSITY OF MANCHESTER
W.J. HOUSTON, Contending for Justice: Ideologies and Theologies of Social Justice in the Old Testament, Library of Hebrew Bible/Old Testament Studies 428, T & T Clark, London 2006. Pp xix + 272. Price: £70.00 hardback. ISBN: 978056704733. This volume, parts of which have already appeared elsewhere in published form, attempts to analyse a selection of Old Testament texts which deal with the theme of social justice, and it is written in the conviction ‘that the Bible has something to say on the subject which is meaningful and worth hearing in the modern world’ (p. 1). Given that ‘social justice’ occupies such a major place in the Old Testament, it was inevitable that the author had to be selective in his approach, and Houston himself concedes that the study is intensive rather than comprehensive in its coverage. After a brief opening chapter (which includes the now almost obligatory auto biographical note) the author evaluates four sociological models which may ac count for the socio economic developments and growth of class formation in Israel and Judah from the eighth down to the fifth centuries BCE. These include the theory of ‘rent capitalism’ (with special attention to the arguments advocated by Bernhard Lang), the (rather inelegantly named) ‘“ancient” class society’ model asso ciated with Hans Kippenberg (which holds that similar social processes took place in Israel and Judah of the biblical period as more or less contemporaneously in the city states of Greece and Italy); the tributary state model (advanced by Gottwald and others), which places most of the responsibility for the exploitation of the Isra elite peasant and the consequent transformation of society, on the state; and the patronage system model, which, according to Houston, was the normal way in which the relation between the classes was conceived in the Old Testament. The following chapter examines the prophetic indictments of social justice and considers to what extent the judgment of the prophets was the result of their own social and ideological commitments. Although there are brief discussions of the rel evant material in Isaiah, Micah, Jeremiah and Ezekiel, the focus of the chapter is on Amos, since this book is the ‘only one of the prophetic collections in which the pri mary concern is to present social injustice as the ground for national punishment’ 584
REVIEWS
(p. 58). There follows an illuminating account of those whom the prophets saw as the victims and those whom they saw as the oppressors and there is an analysis of what precisely the acts of oppression entailed (bloodshed, extortion, loss of free dom, violence and coercion, perversion of justice, and indulgence at the expense of the poor). However, among the prophets, there was no suggestion as to how to rec tify such injustice, and there was certainly no call to revolution, merely the an nouncement of God’s judgment. In chapter 4 Houston discusses (on the basis of Ezekiel 18) what it meant to be a just and righteous person in Israel, and he then analyses the injunctions of the law (especially Exod. 22:21 23:12) concerning proper behaviour towards the poor and powerless in society. The chapter also contains an examination of the relevant mate rial in Proverbs and the Book of Job. The common theme in all the texts examined in this chapter is that justice is dependent on individual commitment. None of the proposals encountered in the Old Testament for establishing social justice by com munal measures is expected to work on its own; rather, justice in society must depend on the actions of individuals, especially those with power and influence. A further chapter is devoted to the role of the king in maintaining social justice and there is a detailed examination of Psalm 72 and the ideology of royal justice. The prophetic critique of injustice (examined in chapter 3) is balanced in chapter 6 by an analysis of the prophetic vision of society as it should and could be ordered, and here Houston examines the relevant texts in Isaiah, Ezekiel and Zephaniah before turning to investigate the utopian character of justice reflected in Deuteronomy (the seven year release of debts and the seven year remission of slaves etc.) and in the Holiness Code (the law of the Jubilee year). There follows a chapter on ‘the God of Justice’ and here Houston argues that the action of God for justice in the Old Testament is conceived on the analogy of two central institutions in Hebrew society, namely, ‘kingship, which includes war, administrative action and judicial action, all of which can be used to right wrongs and deliver people from oppression; and patronage, which involves the offer of aid and protection, possibly up to and including the use of violence against oppressors’ (p. 204). The volume offers some illuminating insights and contains a well informed dis cussion of the relevant issues. The study could, however, have benefited from more careful proof reading: p. 20 line19 should read ‘even if it is successful’; p. 37 line 8 should read ‘could have been used’; p. 39 line 5 should read ‘indictment of oppres sion’ (to note the errors in chapter 2 only). Moreover, there are some inconsisten cies in the volume. For example, the author insists on p. 27 that the correct expres sion is ‘widows and fatherless children’ (rather than orphans, i.e. children who had lost both parents); yet, Houston himself refers often in the course of his study to the ‘widow and orphan’ (pp. 35, 177, 179, 189). Further, on p. 58 the author states that he is ‘not concerned…with final redactional modifications, or glosses’, but on the very next page observes that in Amos 1:3 2:16 the oracles against Tyre, Edom and Judah are probably ‘sixth century supplements’! Nevertheless, the volume pro vides an important and wide ranging discussion of the ideologies and theologies of social justice in the Old Testament, and not the least of its contributions is its inci sive and provocative critique of the notions of ‘justice’ and ‘equality’ embraced by contemporary Western culture. doi:10.1093/jss/fgp018
ERYL W. DAVIES BANGOR UNIVERSITY
585
REVIEWS
ROBERT P. GORDON (ed.), The God of Israel (University of Cambridge Oriental Publications 64). Cambridge University Press 2007. Pp. xvi + 307. Price: £55.00 hardback. ISBN: 978 0 521 87365 9. After a thoughtful introduction by the editor of this volume, Robert Gordon, fol low two major divisions, headed ‘General perspectives’ (7 contributions) and ‘Texts and themes’ (9 articles). Nathan MacDonald, ‘Aniconism in the Old Testament’ (pp. 20 34), starts with Mettinger’s well known monograph on the subject, but simplistically bagatellizes the historically proven tendencies towards aniconism in the ancient Near East long before the Israelite version of ‘monotheism’ started to appear. ‘Monotheism’ and aniconism are by no means the same thing. Egyptian theologians of the New King dom understood earlier than Israel that the Almighty cannot be adequately repre sented by any image or metaphor, yet they often represented Amun Re in images and religious language. One may compare the Christian representations of God in art and literature even though Christians do accept the Second Commandment. Despite the rather pretentious title of his contribution MacDonald soon narrows down his approach to the justification of aniconism in Ezekiel, Isaiah 40 8, and his favourite subject, Deuteronomy 4. John Barton defends his idea of ‘The Imitation of God in the Old Testament’ against recent criticism, especially from the side of Cyril Rodd, but apparently he is rethinking his earlier position and ends with the admission that the imitatio dei ‘is by no means a universal key to unlock the doors in this puzzling document’ meaning the Old Testament (pp. 35 46). Ronald Clements, ‘Monotheism and the God of many names’ (pp. 47 59) deals more with the rise of Israelite monotheism than with the names of God. His point of departure is ‘The claim that only one God actually exists can be interpreted inclu sively as a way of bringing together many traditions of worship and allegiance which were, at one time, separate.’ (p. 47) It seems to have eluded him that Egyptologists like Hornung and Assmann have demonstrated that oneness and multiplicity can coexist in one religious system. The idea that in ‘proto monotheism’ the universal, supra territorial, factor was lacking (p. 52, n. 14) is simply mistaken as far as the Amun Re worship and the exaltation of Marduk in Babylonia are concerned. Even earlier universal claims were made for Sumerian gods like Ninurta and Anu. Judiciously reviewing the literature on the place of creation in Old Testament theology, Katherine Dell, ‘God, creation and the contribution of wisdom’ (pp. 60 72) argues convincingly that creation is not just a late development but is a forma tive influence on Old Testament religion and theology which has found expression in historical, prophetic and wisdom literature. Diana Lipton’s contribution ‘By royal appointment: God’s influence on influenc ing God’ (pp. 73 93) is ‘a reassessment of the significance of kingship language and imagery when applied to God’ (p. 73). She concentrates on the darker aspects of God’s kingship (apparently without knowing W. Dietrich and Chr. Link. Die dunklen Seiten Gottes, Bd. 1, Neukirchen 42002: Bd. 2, Neukirchen 22004), the possibility of human intercession and some relevant passages in rabbinic sources, but inexplicably neglects the Tosefta Targum to Ezek. 1:1 and the Hekhalot litera ture. Walter Moberly counters Regina Schwartz’s thesis that monotheism is to blame for most violence in the modern world: ‘Is monotheism bad for you?: some reflec tions on God, the Bible; and life in the light of Regina Schwartz’s The Curse of Cain’ (pp. 94 112). 586
REVIEWS
Simon Sherwin, ‘Old Testament monotheism and Zoroastrian influence’ (pp. 113 24) takes issue with the popular theory that Zoroastrianism played a sig nificant part in the emergence of biblical monotheism, Janet Tollington, ‘God, women and children’ (pp. 127 38), studies the ‘interplay between God and individual women … particularly in relation to women as childbearers, or not’ (p. 127) from a feminist and synchronic point of view. Graham Davies provides a fine though not exhaustive study on ‘The exegesis of the divine name in Exodus’ (pp, 139 56), paying special attention to Exod. 34:5 8, 6:2 8, 3:3 15. Judith Hadley, ‘The de deification of deities in Deuteronomy’ (pp. 157 74), as sembles evidence for the thesis that names of Canaanite deities like Dagon, Tirath, Sheger and Astarte were deliberately reduced to common nouns in Deuteronomy. She has to admit, however, that Baal occurs in Deut. 4:3 and, in the plural, in 32:21. It has eluded her that several other deities seem to occur in the latter chapter (cf. my The Rise of Yahwism2 [Leuven 1997], 249 53). In recent times the morality of the God of Israel has been questioned not only from outside, but also from inside the Bible. Peter Williams, ‘Is God moral?: on the Saul narratives as tragedy’ (pp. 175 89), opposes the idea that the narrators criti cized the justice of God (Gunn; Exum), but leaves aside the matter of the evil spirit sent by God. On the strength of a number of parallels from Mari, Robert Gordon, ‘Standing in the Council: when prophets encounter God’ (pp. 190 204), argues that in texts like 1 Kings 22, Isa. 6:40; Jeremiah 23 and Zechariah 3 the prophets seem to have access to the Divine Council. His attempt to depict Abraham too as a prophetic intercessor in Genesis 18 (pp. 196 200) seems a bit forced. Both inside and outside Israel other human beings than prophets had access to the Divine Council (cf. my Rise2, 156, 353 6; Fs Loretz, 199 203). Klaus Koch, ‘Ugaritic polytheism and Hebrew monotheism in Isaiah 40 55’ (pp. 205 28), compares the two religions in rather broad terms, without taking into account the evidence of theological development in both Ilimilku’s and Deutero Isaiah’s thought. Philip Jensen discusses a limited number of scholarly ideas about the role of God in the Book of Jonah in his ‘Interpreting Jonah’s God: canon and criti cism’ (pp. 229 45) and rejects them all to advocate a canonical theological ap proach. James Aitken has an excellent article on ‘The God of the pre Maccabees: desig nations of the Divine in the early Hellenistic period’ (pp. 246 66) but he seems unaware of recent developments: the cultural and religious contacts between the Aegean and the Levant started much earlier than the fourth century and the ‘God of heaven’ (pp. 259 64) was a name of God already in the Persian period (Trotter). Finally William Horbury discusses the ‘Deity in Ecclesiasticus’ (pp. 267 92). He rightly sees a conjunction between the transcendence and immanence of God in Ecclesiasticus. Since about 1980 the number of studies about the God of Israel has increased far beyond what is justified on the basis of the rather scant evidence available. The rel evance of biblical data is difficult to assess as long as there exist fundamental differ ences of opinion about the origin and dating of the sources. Archaeology has deliv ered an incalculable mass of new data but the complexity of the picture appears to become greater with every major new find. In a situation like this any number of 587
REVIEWS
hypotheses is admissible and the volume under review adds a number of interesting ones. At the same time one would wish that authors would frankly admit the lim ited scope and the hypothetical nature of their proposals. doi:10.1093/jss/fgp019
J.C. DE MOOR KAMPEN THEOLOGICAL UNIVERSITY
JONATHAN D. LAWRENCE, Washing in Water: Trajectories of Ritual Bathing in the Hebrew Bible and Second Temple Literature. Society of Biblical Literature, Atlanta 2006. Pp. xix + 294. Price: $47.95 ISBN: 1 58983 199 3. Jonathan D. Lawrence has written a systematic study of washing in ancient Hebrew literature. His motivating question was the relationship of Christian baptism to the practice of Jewish ritual immersion. In his study of the texts and archaeological record he finds roots of Christian baptism in the Hebrew Bible and among the ‘di versity of bathing practices’ during the Second Temple (p. 21). He examines (if not in his text, in the helpful appendices) virtually every text on washing in ancient He brew literature. After the introductory chapter, Lawrence proceeds by describing ritual washing in the various texts of ancient Hebrew literature: Hebrew Bible (chapter 2), Second Temple literature (chapter 3), the Dead Sea Scrolls (chapter 4). Chapter 5 is de voted to archaeological discoveries which relate to ancient ritual bathing practices in Israel and Jordan. Chapter 6 gives a conclusion and further ideas for future stud ies. Lawrence also includes excellent pictures of miqva’ot and helpful appendixes. Appendix A compares purity practices by topic in texts from the Bible, Second Temple Literature and the Dead Sea Scrolls. Appendix B is a list of all references in the literature related to washing. Appendix C is a list of all miqva’ot discovered in Israel and Jordan. Lawrence repeatedly claims that since purification in water is seldom mentioned in the biblical text, it must not be assumed. Rather, other modes of purification are possible. He argues that viewing the biblical material through the lens of a purifica tion system, whether priestly (Milgrom) or rabbinic (Neusner), narrows the spec trum of diverse kinds of washing which are present therein. For example, references to purification in the Deuteronomistic History and in Ezra Nehemiah Chronicles could simply be metaphorical, a purging of the heart, as found in the Psalms (cf. Pss 26:6; 51) and prophets (cf. Jer. 4:14; Ezek. 36:25) (pp. 35 7). This diversity, he continues, comes to fruition in Second Temple texts. However, the references to purification in the Deuteronomic History which do not explicitly mention washing do not indicate that washing for purification was unknown in the time of the first Temple (p. 41). Lawrence asks why the writers give details of the temple’s construction but not of purification practices. He sug gests that purity practices were inserted later into Tabernacle descriptions by Sec ond Temple authors seeking to authenticate their own observance. In my view, this argument from silence is dangerous, especially since washing is a routine, usually mundane, activity which would not necessarily be of interest to the writer. Also, he overlooks the genre of these books, none of which are legal compilations or priestly handbooks. Furthermore, the authors do indicate a familiarity with the practice of washing for ritual purification (cf. 2 Sam. 11:2; 2 Kgs 5:10), a practice common not only to Israel but attested throughout the ancient world. Lawrence claims that there are new uses of washing in Second Temple times: before prayer, before Sabbath, washing after defecation, before initiation, and hand washing rituals, none of which are known in the Hebrew Bible (pp. 76f ). This claim 588
REVIEWS
seems overstated. For example, ritual washing before prayer is not so innovative when one realizes that prayer can be defined as meeting with God and therefore biblical precedents apply (cf. Exod. 19:10). Hand washing is known to the biblical writers (Lev. 15:11). Finally, washing before initiations is not a new phenomenon either but was already part of both the priestly and levitical initiations in the Hebrew Bible (Lev. 8:6; Num. 8:7). On the other hand, Lawrence is correct that Second Temple literature contains allegorical or symbolic explanations of ritual actions lacking in Scripture (cf. Philo’s Special Laws 1.263 9; Letter of Aristeas 304 6). Furthermore, atoning sacrifices are not a new source of purification. The Ìa††a’t is a purificatory offering already in Scripture (cf. Lev. 12:6; 14:19). Lawrence ar gues that the impure person who delays purification does not wash later but rather is required to bring a sacrifice. However, the person who has neglected purification is in violation of God’s law, a dilemma which only an atoning sacrifice will remedy, not just washing (Num. 19:20). Lawrence helpfully combines archaeological discovery with literary references in his argument that in Second Temple times purity was a strong Jewish concern. For example, he supports the Mishnah’s claim of divided staircases to separate the pure and impure in Jerusalem (m. Sheq. 8:2) in light of divided miqva’ot dating to the same period. Lawrence claims that the primary indicator of miqva’ot is steps and 158 ritual baths have been dated to the Hasomonean period, most in Judaea. He connects the sectarian Scrolls, in which purity is a major topic, with the site at Qumran where ten stepped pools have been found (p. 173). Lawrence helpfully surveys in a text by text fashion the data on washing found in the Dead Sea Scrolls. He notes new usages of purity terms: qds rather than htqds, and the increased usage of nzh and †bl. Lawrence correctly identifies the tendencies to abbreviate or use shorthand terms for larger concepts and to harmonize and strengthen purity laws of the various impurity bearers involving quarantine of some and longer periods of impurity (pp. 82 3, 99). Lawrence notes the priestly character of the Dead Sea Scroll community(ies), which requires everyone, not just priests, to perform ritual bathing. However, he incorrectly claims that in the Hebrew Bible ‘while the laypeople may have eaten parts of their sacrifices, only the priests ate their portions “in a holy place” and had the expectation of purity placed upon them’ (p. 129 n. 20). In fact, Scripture is clear that anyone, not just priests, who eats of the Lord’s sacrifices must be pure (Lev. 7:19 20; cf. 11:1 where pure food laws are addressed to all Israel). Another ‘difference’ Lawrence perceives between the Scrolls and the Bible is that the former require the menstruant to bathe. While Lawrence is correct that the menstruant’s bathing is not explicitly stated in the Bible, it must be assumed. First, impure persons with lesser impurity than menstruants were required to wash (e.g. those with seminal impurity, a one day impurity, Lev. 15:15 16). Second, even the bed and chair of the menstruant must be washed as well as those who touch them. Certainly, the menstruant touches her own bed and chair. Curiously, Lawrence as sumes a system of ritual washing underlying the Scrolls but not the Hebrew Bible (p. 150). Lawrence acknowledges that some concepts known from rabbinic literature were probably known to the Qumran authors (e.g. †ebul yom, Sabbath prohibitions) and that the data of the later sources sometimes help to clarify matters in the Scrolls. Thus, it is curious that he is hesitant to assume washing was required by Qumran texts even though they explicitly require purification in ritual contexts, e.g. before participating in sacred festivals and before handling newly harvested produce, and 589
REVIEWS
even though comparative evidence exists from Josephus and the Mishnah (cf. p. 115). The question arises then, what other method of purification was intended? Lawrence’s suggestions of alternate forms of washing (e.g. purification by the spirit, purification by knowledge and purification by affliction) do not seem applicable to the above contexts. Lawrence sees a development in the Scrolls akin to that of John the Baptist, who refers to a baptism of water and a baptism of the spirit. To be sure, the Scrolls, fol lowing the prophets, demonstrate that purification by God’s spirit was earlier than the New Testament. However, this does not diminish the necessity of ritual purifi cation by water. Washing as a metaphor for inner purification gains strength be cause there really was a physical purification process. For John the Baptist, as well as some Qumran authors, water purification anticipated divine forgiveness (John 1; cf. 1QS 3:4 9). Lawrence concludes that by demonstrating diverse uses of washing in the Second Temple period he has challenged Christian polemic which says baptism is spiritual while Jewish bathing is merely ritual (p. 201). I agree that purification by ritual washing facilitated spiritual renewal for many Jews and Christians. However, this does not mean the practice was bereft of religious significance in more ancient times or that it was done without water. Rather, both Jewish and non Jewish texts from the ancient Near East confirm that washing was a cross cultural ritual re quired before access to the gods. doi:10.1093/jss/fgp020
HANNAH K. HARRINGTON PATTEN UNIVERSITY
LESTER L. GRABBE, A History of the Jews and Judaism in the Second Temple Period: Volume 1, Yehud: A History of the Persian Province of Judah (Library of Second Temple Studies 47). T & T Clark, London 2004. Pp. xx + 471. Price: £55.00 hardback. ISBN: 0 567 08998 3; Volume 2: The Coming of the Greeks: The Early Hellenistic Period (335 175 BCE) (Library of Second Temple Studies 48). T & T Clark, London 2008. Pp. xxi + 434. Price: £75.00 hardback. ISBN: 987 0 567 03396 3. Four years (and £20!) separate these two volumes, which bring the four volume project to its halfway point. It must be doubtful whether the remaining years 174 BCE 70 CE can be covered in anything like as little space as 450 175 BCE, given the wealth of primary and secondary materials (including much of the New Testa ment?). But Lester Grabbe is the most prolific historian of Second Temple Jewish history of our times, thorough and (so far) indefatigable, and no skimping can pos sibly be foreseen. Both these volumes are divided into five parts: Introduction (Principles and Methods); ‘Sources’, beginning with archaeological, then literary sources divided into various categories; ‘Society and Institutions’ (subdivided differently in the two volumes); Historical Synthesis; and finally Conclusions A Holistic Perspective. In the first volume this includes a section ‘What We Know, What We Can Guess and Where Our Ignorance Is’ and is followed by an appendix on ‘Persian Influence on Jewish Religion and Thought’ and a bibliography of 1400 books (about 2000 in Volume 2). Half of these, as both Prefaces tell us, date from 1990 or later. Despite the author's many years of chairing the European Seminar on Historical Methodology and editing its publications, the introductory essay in vol. 1 on ‘Princi ples and Methods’ is quite brief. In a swipe against dogmatic methodologies, he seeks to ‘evaluate each case on its merits and consider the arguments advanced be 590
REVIEWS
fore deciding’ (p. 13). ‘Merits’ may be a slippery term: the principle leads him to be dubious of the historicity of Ezra but to accept the ‘Nehemiah memoir’ and he finds Nehemiah rather odious and ultimately a failure. But he admits the reasons for the latter conclusion are ‘somewhat subjective’, showing that there are exceptions to his principle! His approach could be and has been characterized as ‘positivistic’, adopting a view that the historian seeks to conclude ‘what happened’, unlike many modern historians who accept the need for interpretation, even imagi nation, to take us beyond ‘facts’ to the ‘meaning’ that all histories proclaim or imply. Grabbe has less interest than many in philosophical debates on such matters and he would probably accept that he is a determined anti postmodernist and in some respects quite traditional. But he is not ignorant of any of these issues, and in fact as he states very clearly this History is as much about what we do not know as what we do know (vol 1: 13 16; 351 3). He is also wary of ‘grand narratives’ and is more concerned with the skeleton of data than the skin of rhetoric. But Part Four of each volume comprises a ‘Historical Synthesis’, embedded in the forest of data and description and not imposed on it. These syntheses are episodic, tentative and sometimes conjectural, as Grabbe himself underlines. Yet he does claim that the lit tle known Persian period was the most important in the history of Judaism’s evolu tion. Because of our knowledge of the periods before and after (with the very im portant exception of the Ptolemaic, at least in central Palestine), we can deduce that hugely important developments were taking place in a poor and obscure corner of the Persian empire (as admittedly they were elsewhere, too). But the sources left to us do not let us see these events happening, except now and then. Narrative histo ries tend to expand those eras when we have most data (Palestine itself being a su preme example, its ancient history being dominated by the biblical texts), but that is a mirage: the ‘black holes’ (or rather, ‘grey holes’) often hold the most momen tous episodes. Conversely, so many of the texts are not interested in accurate de scription of events but rather in representing quasi mythical tropes like ‘exile’ and ‘restoration’ and in competing claims upon that elusive identity ‘Israel’ (which until recently so many historians and archaeologists naively took for granted). The ar chaeology of the Persian period is less help than in the earlier or later periods, its findings still poor, and poorly interpreted after decades of relative neglect. For all these reasons, we need a book that deals thoroughly with data, and especially what the data can and cannot mean: we therefore need narrative, but not too little or too much. The importance of the Persian period coupled with the poverty of good evidence certainly has generated a lot of hypothesizing over relations between Yehud and Samaria, over the role of the Temple and priesthood, over the emergence and role of torah and the origins of Judaism itself. On all these topics the reader finds the data enumerated and evaluated, scholarly opinions accumulated, and some narra tive suggested all in a quite relentlessly systematic way. But how else? The volumes differ minimally in response to the different source materials. ‘In troduction, Aims and Methods’ of the second volume singles out the problems of precise chronology, the emergence of a large Jewish Diaspora, the new methodology of ‘Post Colonial Theory’ (where he makes the interesting and important point that while Jews may have been the victims of colonialism in one sense, they are cultur ally the victors, since it is their version of history that has generally prevailed), the existence of a proper historiographical tradition (and its problems). Especially com mendable here is the chapter on ‘Hellensim and Jewish Identity’, which shows Grabbe at his best, combining a strict adherence to data, an awareness of a range of 591
REVIEWS
possible interpretations and a keen eye for scholarly concepts that obscure rather than clarify (he has a self confessed dislike of ‘consensus’, though is seldom in any way radical himself ); and, in the end, one feels a self imposed obligation to make a call, to make some sense of it all. While these volumes can be compared in many ways with the revised Schürer (and in fewer ways with the very uneven Cambridge History of Judaism), they have the advantage of the better and more consistent or ganization and equality of treatment that a single author allows, as well as more awareness of what doing history entails, even if less theorizing and less radicality than some would like to see. It would be easy to enumerate arguments, conclusions and opinions with which the reader (including this one) will disagree, yet such a reaction is on the whole un necessary and unfair. The logical way in which the sources, their interpretation, their possible historical value, what scholars think, and the historical reconstruction that they permit or imply, are scrupulously set out not only invites but even facili tates informed disagreement and alternative narratives. Only those who either trust biblical sources on principle or regard all histories as fiction will, on the other hand, find these volumes unhelpful. They will certainly replace a whole shelf of books in my own study. doi:10.1093/jss/fgp021
PHILIP R. DAVIES UNIVERSITY OF SHEFFIELD
MATT JACKSON MCCABE (ed.), Jewish Christianity Reconsidered (Rethinking Ancient Groups and Texts). Fortress Press Minneapolis 2007. Pp. x + 389. Price: $35.00. ISBN: 978 0 8006 3865 8. This volume is one of the products of the Society of Biblical Literature’s Consulta tion on Jewish Christianity, which set for its task a clarification of the terms ‘Jewish Christian’, ‘Jewish Christianity’, and ‘Christian Judaism’. To adapt John Marshall’s observation, except for the theoretically precise but deeply problematic use of the term by F.C. Baur and the Tübingen school, Jewish Christianity remains ‘an incon sistent, undertheorized, unilluminating, and ultimately detrimental category’ for the literature and groups usually so designated. The essays range from examinations of particular documents which have tradi tionally been supposed to be ‘Jewish Christian’, to essays which ask ‘in what sense is x Jewish Christian?’, to three probing and sophisticated efforts to problematize the category itself, and to expose some of its deeper ideological and theoretical invest ments. These three are methodologically and theoretically the most significant, though the other essays all qualify as valuable and well argued examples of histori cal criticism. Petri Luomanen (‘Ebionites and Nazarenes’) offers an incisive analysis of the reports about the Ebionites and Nazarenes by Irenaeus, Epiphanius and Jerome. Accepting the utility of the term ‘Jewish Christian’, Luomanen is only concerned to map the differences between Irenaeus’s Ebionites and those of Epiphanius. The former, Luomanen argues, derive ultimately from the ‘Hebrews’ of Acts 6 8 while the latter come from the Hellenistic ‘poor’ of Acts. Jerome’s Nazarenes were ‘mainly Christian’, pro Pauline, and used gospels written in Hebrew or Aramaic. ‘It seems unlikely that their Jewishness would have gone much beyond that’ (p. 116). Jonathan Draper (‘The Holy Vine of David made known to the Gentiles through God’s Servant Jesus’) argues that the Didache comes from a ‘Christian Jewish community’ which admitted Gentiles without requiring of them either cir 592
REVIEWS
cumcision or adherence to kashrut. Nonetheless, they held out full observance of the Torah as the ideal. Draper offers compelling exegeses of several pericopae, argu ing that the Didache gives us no evidence of a ‘parting of the ways’. Instead it evi dences a group of Jewish Jesus believers who found themselves in competition with the ‘growing influence and power of the rabbis’ (p. 269). In ‘The Pseudo Clementines’ Stanley Jones profiles the now lost Jewish Chris tian Circuits of Peter, reconstructable through a synoptic comparison of portions of the Recognitions and the Homilies. He isolates a document and a group which occu pied a middle point between ‘Hebrews’ and ‘Gentiles’, insisting on baptism but also embracing the purity commandments of Second Temple Judaism. Five essays try to establish criteria by which a group or document might (or might not) be called ‘Jewish Christian’. Craig Hill (‘The Jerusalem Church,’ pp. 39 56), though he acknowledges the dangers of supersessionism, nevertheless tries to establish the character of the Jerusalem church as Jewish Christian. For the Jew ish part of the term, he insists on a polythetic approach: the church held a belief in the restoration of Israel, it embraced the Torah, and it was focused on the temple. But when he asks the question, ‘were they Christian?’, Hill adopts a monothetic approach cultic devotion to Jesus. This is highly problematic, for not only is the term ‘Christian’ anachronistic as a denominator of any group prior to the end of the first century, but by dismissing the ethnic aspects of Judaism as irrelevant, Hill makes both ‘Christianity’ and ‘Judaism’ into commensurable ‘religious’ systems characterized by beliefs. Moreover, he assumes rather than demonstrates that ‘devo tion to Christ’ was an aspect of the Jerusalem church. Happily, Jeffrey Sumney (‘Paul and Christ believing Jews whom he opposes’) prefers ‘Christ believing Jews’ to ‘Christians’ (though he never explains in what sense Paul’s opponents believed in Christ or whether belief in Jesus was a key de nominator). Interestingly, he suggests that the ‘Judaizers’ in Galatia may simply have wanted to allow Gentiles to associate freely with non Christ believing Jews in synagogues and thus to avoid any accusation that Gentiles were claiming Jewish privileges. The key difference between Paul and these opponents, according to Sumney, was not particular beliefs about Christ but whether ‘belonging to Christ’ defined group identity, as it did for Paul, or whether covenantal membership was the overarching identity category. Patrick Hartin’s essay, ‘The Religious Context of the Letter of James’ is a wide ranging summary of current research on James, but makes the key point that James evidences a Jewish restorationist movement which was interested in ‘Jesus’ message rather than his person’ (p. 228). The Jacobean groups were fully within the house of Israel, embracing the whole Torah while acknowledging Jesus as the Messiah (1:1; 2:1). Hartin’s observations serve to highlight the problem involved in using the term ‘Christian’ in the sense that Hill wants to use it, since it is not at all evi dent that James believes in Jesus Hartin in fact takes ‘the faith of our Lord Jesus Christ’ in 2:1 to be a subjective genitive or that there is any ‘Christ devotion’ here. Warren Carter (‘Matthew’s Gospel: Jewish Christianity, Christian Judaism, or Neither’) referees the debate between Anthony Saldarini (Matthew’s Christian Jewish Community, Chicago 1994) and Donald Hagner (Matthew, Dallas 1993 5) as to whether Matthew’s group should be classified as a reformist sect seeking to reform Jewish society (Saldarini) or a Jewish ‘Christianity’ in which the Torah had been displaced by christology (Hagnar). Raimo Hakola (‘The Johannine Community as Jewish Christian?’) concludes that John displays so ambivalent an attitude toward 593
REVIEWS
the standard markers of Jewishness that labelling it as Jewish Christian is problem atic. Three essays go to the heart of the issue addressed by the consultation and confer on the volume its lasting significance. The editor’s ‘What’s in a name?’ offers a suc cinct survey of the various attempts to define Jewish Christianity, from F.C. Baur to Daniel Boyarin, James Dunn and Gabrielle Boccaccini. He shows that the category is both unworkably elastic and allows for the irony that Paul, who is perhaps most insistent on his Jewish credentials, is routinely excluded from the category of Jewish Christianity. This ought perhaps to lead us to abandon the category entirely as ana lytically useless. But Jackson McCabe seems loath to do so, preferring instead to call for a polythetic approach to classification as potentially more productive. It is in William Arnal’s essay (‘The Q Document’) and John Marshall’s ‘John’s Jewish (Christian?) Apocalypse’ that we see careful theorizing at work. Arnal gives a succinct account of recent redactional approaches to Q, observing importantly that at no stage is Jesus worshipped. Nowhere is Jesus described as the ‘Christ’ and in contrast to Paul’s thinking, Q does not treat Jesus’ death as salvific. Rejecting the term ‘Christian’ as ‘actively misleading’ (p. 132), Arnal points out that Q lacks features usually thought to constitute other Jesus groups the kerygma of the dying and rising Jesus, a confession of Jesus as ‘Christ’, a rejection of kashrut, Sabbath and circumcision, and the recognition of common cause with other Jesus groups, ‘the litmus test for the claim to ideological affiliation with Jesus movements’ (p. 135). Yet Arnal observes an odd phenomenon: tracking conceptual features against redaction history, he notes that Q becomes more ideologically Jew ish in its later stages (making greater appeals to the Tanak, insisting on tithing and the perdurance of the Torah) precisely as it also becomes more ‘Christian’. Q’s earli est stages express an ‘interest’ in Jesus, a belief in the God of Israel, and employs fictive family language, and in later stages it treats the Hebrew Bible as predictive, adds appeals to John the Baptist, and multiplies the titles assigned to Jesus. Never theless, ‘interest in Jesus and his “kingdom of God” in no way attenuates or even calls into question the self evident ethnic identity of the group responsible for the document and in fact provides the channel through which it expresses its commit ment to at least Israelite religiosity’ (pp. 151 2). Arnal’s findings thus cast into grave doubt any model which sets ‘Christianity’ in tension with ‘Judaism’ or which imagines of a developmental schema that sets ‘Christianity’ as the telos of group development. ‘Q’s odd history suggests instead that the Jesus people became more discursively and ideologically “Jewish” over time; that “Christian” self definition was in fact an act of Jewish sectarian identity formation, not the birth pangs of a new religion’ (p. 153). Marshall’s essay on the Apocalypse reiterates and extends some of Arnal’s points. The application of ‘Jewish Christian’ to the Apocalypse has functioned in the serv ice of two kinds of subordinations: it either highlights the way in which Judaism was subordinated to Christianity (so R.H. Charles’ magnum opus on the Apoca lypse), or it designates the Apocalypse as inferior theologically in the context of the canon of the New Testament (so Bultmann). Both of these tendencies are simply extensions of the developmental schema advanced by F.C. Baur, according to which the theological history of Christianity was epitomized in the conflict between the parochialism of Jewish Christianity and Paul’s universalism, both later absorbed and sublated in the Catholic synthesis. According to Marshall, the category ‘Jewish Christianity performs the dual function of masking the untenability of “Christian” while simultaneously acknowledging the appropriateness of the category “Jewish”’ 594
REVIEWS
(p. 236). The category, insofar as it reflects the value laden developmental schema, is an ‘effectively racist scheme of Christian development’ (p. 237). Marshall richly illustrates how the adoption of a schema of theological develop ment has led to bizarre assessments of the Apocalypse, with commentators relegat ing portions of the text to ‘shallow brained [Jewish] fanatics’ or to ‘Jewish Christian heretics’, or positing an author who was unable to maintain his Jewish identity, or simply using the term ‘Jewish Christian’ in an utterly promiscuous and untheorized fashion. The remedy, for Marshall, is not to rehabilitate ‘Christian’ but to read the Apoca lypse as a Jewish text. Like Arnal, Marshall insists on a polythetic approach, and observes that the Apocalypse displays ethnic mappings, calendrical and ritual prac tices, mythological appeals and spatial imagination, all consistent with a Jewish provenance. The consequence of Marshall’s approach is that the Apocalypse’s Jewishness (punkt!) is acknowledged without invoking the category of ‘Jewish Christian’, either as a way to subordinate the Apocalypse theologically to other con ceptual and ritual constellations of the Jesus movement, or to impose an extrinsic interpretive template that allows the interpreter to winnow the ‘Christian’ elements from the ‘Jewish’ chaff. Arnal and Marshall have persuasively shown how the term ‘Jewish Christian’ sys tematically distorts the understanding of the documents and groups to which it is applied. Its use moves in the realm, not of historical exegesis but of theological apologetics. doi:10.1093/jss/fgp022
JOHN S. KLOPPENBORG UNIVERSITY OF TORONTO
MICHAEL RAND, Introduction to the Grammar of Hebrew Poetry in Byzantine Pales tine (Gorgias Dissertations 22, Language and Linguistics). Gorgias Press, Piscataway, New Jersey 2006. Pp. xvii + 537. Price: $95.00 hardback. ISBN: 1 59333 348 X. The book, a revised version of the author’s Ph.D. dissertation (New York Univer sity, 2003), offers a descriptive grammar of the works of Elazar be rabbi Qillir, a prolific Hebrew poet of the early seventh century CE and an iconic representative of the Palestinian school of poetry (also known as Piyyut). Rand sets out to explore this poetic oeuvre in order to ‘build up a picture of the Qillirian dialect’ (p. xii), and by so doing to provide the reader with an ‘accurate representation of the grammar of Qillirian Piyyut’(ibid.). Rand’s decision to focus on a single poet is laudable. Until recently, all other treatments of the language of Piyyut and they were not too many addressed general questions in regard to that literary phenomenon. The only exception to this rule is Avshalom Kor’s unpublished doctoral dissertation (Tel Aviv University, 1988) on the language of the sixth century poet Yannai. Another noteworthy contribution is Naoya Katsumata’s monograph Hebrew Style in the Liturgical Poetry of Shmuel Hashlishi (Leiden 2003). Although Shmuel Hashlishi (= the Third, i.e., of the third rank in the Palestinian academy) composed his poetry in the eleventh century, Katsumata’s study is highly relevant since he faces similar methodological challenges. Several shortcomings of Katsumata’s methods were singled out by Rand (in a review he published in Zeitschrift der Deutschen Morgenländischen Gesellschaft 156 [2006]: 194 7), a fact that undoubtedly contrib uted to the refinement of his own work. In the book’s introduction, Rand discusses the state of the research (pp. 1 8) and elaborates briefly on the theoretical back ground of his project (pp. 8 14). His main source of influence, in that regard, is 595
REVIEWS
Roman Jakobson’s groundbreaking works on the intersection of linguistics and poetics. Following Jakobson, Rand concludes that ‘poetic language constitutes a legitimate object of linguistic inquiry’ (p. 13). Finally, the introduction is con cluded with a short discussion of the corpus (pp. 17 23), that consists mostly of Qillirian poems that survived in the Franco German liturgical rites of the Middle Ages alongside some materials from the Cairo Genizah. The remainder of the book can be divided into two parts; the first, which covers most of the book, offers a systematic structural analysis of the corpus, while the sec ond is devoted to the examination of the interdependence of poetics and linguistics, or in the author’s phrasing Rhetorical Figures. The first part of the book (consist ing of chapters one to three) is organized according to the traditional tripartite divi sion of phonology (pp. 25 9), morphology (pp. 31 242) and syntax (pp. 243 438). As can be expected from a skilful linguistic project, these chapters are rich in examples from the Qillirian corpus and provide a fair amount of comparative mate rials drawn mainly from Biblical, Dead Sea Scrolls and Mishnaic Hebrew but also from various Aramaic dialects and to some extent Greek. The comprehensive analy sis of the material leads Rand to one of the most important conclusions of the book. While Qillir’s Hebrew, Rand contends, ‘is largely based on BH [=biblical Hebrew],’ it is ‘not simply a slavish imitation of BH.’ (p. 16.) Furthermore, by a process of elimination Rand was able to arrive at a ‘(small) number of features with out any analogues’ that ‘may be considered to be the Qillir’s contributions… to the development of the Hebrew language’ (pp. 16 17). The fourth chapter of the book (‘Rhetorical Figures’, pp. 441 527) examines vari ous poetic features (for example paronomasia, metonymy and catachresis,) but also considers such phenomena as the incorporation and modification of biblical phrases in the poems or the manipulation of Aramaic roots and words. This chapter, arguably the most innovative contribution of the book, fuses poetics and linguistics (à la Jakobson) in a brilliant fashion and proves that ultimately the two cannot be separated. It can be concluded, then, that the book’s primary goal to study the poetic dialect of Elazar be rabbi Qillir from that dual perspective is entirely achieved. That being said, it is worthwhile to address a problematic aspect of the book that relates to methodology. Research, to greater or lesser extent, relies on induction. At the same time, induction cannot be taken for granted, it ought to be transparent and must be discussed and justified. In two instances in Rand’s book, however, one can sense a weakness in that regard. Take for example the disparity between the book’s extensive title (Introduction to the Grammar of Hebrew Poetry in Byzantine Palestine) and its limited scope (a detailed analysis of the works of a single poet). Does Rand maintain that the ‘Qillirian dialect’ encapsulates the linguistic traits of the entire school of early Palestinian Piyyut or that the study of the Qillir can serve as an introduction to the entire poetic tradition? In one place he seems to make this inductive claim when he talks of ‘the Qillir’s contributions, or those of the Byzan tine period payyetanic dialect in general’ (pp. 16 17). Since the author does not address this question in the open, the book’s title is somewhat misleading and the reader is left puzzled. A similar state of affairs is shown in the demarcation of the analysed corpus. In the introduction the author proclaims he seeks to analyse six comprehensive compositions that ‘constitute ca. 5% of the Qillirian oeuvre as it is presently known’ (p. 17) and adds that ‘in addition to the compositions cited above, which are collectively referred to as “the Corpus” and are analyzed exhaus tively in this grammar, occasional references are made throughout… to other Qillirian compositions’ (p. 22). However, there is no discussion whatsoever how 596
REVIEWS
five per cent of the material can stand for the entire Qillirian corpus (let alone, the entire poetical school). Furthermore, Rand then says that ‘the only really meaning ful criterion employed in its selection is difficulty, viz., all of the compositions in cluded here may be classed as belonging to the “difficult” end of the spectrum of Qillir’s piyyutim’, (p. 17) but again fails to explain that decision. I suspect that these flaws relate to the (rather notorious) process of converting a doctoral dissertation into a book. In fact, the title of the dissertation (Poetry and Grammar in the Works of Eleazr be rabbi Qillir A Grammatical Analysis of Selected Liturgical Poems) reflects the scope of the present work more accurately. At any rate, these remarks cannot undermine the fact that Rand’s impressive book opens a new chapter in the study of the history of Hebrew (poetic) language. One can only hope that the author will continue to pursue this course of study and that this publication will encourage other scholars to study the language of Piyyut as linguistic and poetic phenomena alike. doi:10.1093/jss/fgp023
OPHIR MÜNZ-MANOR THE OPEN UNIVERSITY OF ISRAEL
MAURO ZONTA, Hebrew Scholasticism in the Fifteenth Century: A History and Source Book (Amsterdam Studies in Jewish Thought 9). Springer Verlag, Dordrecht 2006. Pp. vii +318 + 70 (Hebrew). Price: £83.00 hardback. ISBN: 978 1 4020 3715 3. Mauro Zonta’s book on Hebrew Scholasticism is an impressive contribution to our knowledge of the history of medieval Jewish philosophy. Its aim is ‘to show the knowledge of Scholastic philosophical literature or, better, the use of Latin sources by some fifteenth century Jewish philosophers’ (Foreword). These philosophers are Abraham Bibago, Baruch Ibn Ya’ish, Abraham Shalom and Eli Habillo, all of whom wrote in the second half of the fifteenth century in the Iberian Peninsula, and Judah Messer Leon, who was active in Italy during the same period. Until very recently the study of this period in Jewish philosophy was a rather ne glected field. The past decade, however, has witnessed an increased interest in the subject, which has led to a number of seminal studies by, for example, Charles Manekin, Havah Tirosh Rothschild, Tamar Rudavsky, and last but not least, Zonta himself. Nonetheless, given the large amount of texts that still require further re search one cannot but agree with Zonta’s conclusion that it is still premature to write a general historical overview of the period under discussion, and that what is most needed now is editions and translations of unpublished material (Intro., p. 31). Therefore, the bulk of the book (Chs I IV, pp. 33 304) consists of English translations or paraphrases of selected sections of the writings of the aforemen tioned philosophers, as well as editions of some Hebrew texts by Baruch Ibn Ya’ish, Eli Habillo, and Judah Messer Leon (Hebrew section, pp. *1 *70). As the title indicates, the volume is thus a veritable Source Book. As for History, in his instructive Introduction (pp. 1 31) Zonta first provides an overview of the influence of scholasticism on fourteenth century Jewish philosophy as background, and then discusses specific features of ‘Hebrew Scholasticism’ in fifteenth century Spain and Italy respectively. Surveying the results of recent research, Zonta treats important questions pertaining to the phenomenon under discussion, such as the Hebrew authors’ knowledge of Latin and their participation in institutions of learn ing. In fact, the introduction provides a ‘state of the art’ overview of the field. The volume also contains indices of pre modern authors (up to 1600), modern authors and of manuscripts (pp. 305 18). 597
REVIEWS
The first three chapters treat exponents of Hebrew Scholastic thought in the Iberian Peninsula, whereas the fourth focuses on Italy. All chapters reveal the same structure: each chapter starts with the relevant biographical information and a list of works (Latin to Hebrew translations and original compositions) of the author(s) under discussion, and then presents one or more texts in the form of an English translation or a detailed paraphrase. For each text the Latin sources are listed, often with quotes provided in the footnotes, or, as in the last chapter, in parallel columns. Moreover, the discussions about the text contain important observations on termi nological issues and Latin glosses (see, for example, p. 183 and pp. *25 *33 in the Hebrew section). Chapter 1 offers a brief general sketch of the main contents as well as a 62 page English paraphrase of the Treatise on the Plurality of Forms (Ma’amar be ribbuy ha tzurot) by Abraham Bibago, whom Zonta describes as ‘the best known and most studied philosopher among the fifteenth century Spanish Jewish thinkers’ (p. 33). The author, who was engaged in anti Christian polemics, reveals a rather eclectic approach to Scholasticism and seems to have been interested primarily in Thomism and Scotism. The treatise is extant in a unique copy, perhaps an autograph. Among the sources quoted are writings by John the Canon and William of Ockham. Zonta notes that although Bibago’s relationship to Scholasticism is more limited than that of the other authors treated in the volume, the treatise reveals enough Scholastic patterns to be considered as forming part of Hebrew Scholasticism. The main subject of Chapter 2 is a commentary on the Nicomachean Ethics by a student of Baruch Ibn Ya}ish, a commentary that is based upon a Hebrew transla tion by Baruch and reflects the latter’s teaching. The text reveals a consistent use of Thomas’s Sententia Libri Ethicorum, the commentaries of Thomas on Aristotle be ing among the most important channels for the transmission of Scholastic thought to the Jewish philosophers. Other authors quoted include Albertus Magnus and the Spanish Thomist Pedro Martinez De Osma. Zonta presents the following passages: the introduction, the Incipits of the books I X compared to corresponding passages of Thomas’s commentary, and Book I.7. The chapter concludes with an English translation of a few brief sections of Ibn Ya’ish’s commentary on De Anima, the source of which is likewise Thomas’s commentary on the Aristotelian text. Eli Habillo, or Habilio, another exponent of Hebrew Thomism and Scotism, is the main protagonist of Chapter 3. The end of this chapter discusses his philo sophical correspondence with Abraham Shalom, who was more interested in nomi nalism, the third intellectual current in fifteenth century Spain. In comparison to the philosophers discussed in chapters I II Habillo’s Scholasticism was wider in scope and more far reaching. Zonta notes that Habillo had direct or indirect con tacts with institutional Latin scholasticism and may even have attended university courses in Paris. Habillo’s Latin to Hebrew translations include Antonius Andreas’s Quaestiones Super XII Libros Metaphysicorum and Johannes Versor’s Questions on the Physics of which Zonta provides the introductions. All these writings testify to the appropriation of numerous doctrines from con temporary Latin Scholasticism. Why should Jewish authors of the Iberian Peninsula turn to Latin Scholastic sources in the second half of the fifteenth century? Zonta suggests that their motivation was both religious and philosophical: knowledge of these sources was useful in Christian Jewish disputes, and at the same time Jewish philosophers became increasingly dissatisfied with ‘traditional Jewish Aristotelia nism, mainly based upon medieval Arabic Islamic philosophy’ (Intro., p. 1), which necessitated an alternative for Averroism that was considered too radical (see also Intro. p. 22 4). 598
REVIEWS
In Italy, however, the situation was not entirely the same, since Italy already wit nessed a period of ‘Hebrew Thomism’ between 1250 and 1350 (p. 3 5), and hence had a longer tradition of knowledge and use of Latin sources. Moreover, the attitude of Italian Jewish thinkers vis à vis Averroes was different to that of their Spanish colleagues, for ‘Averroes was still regarded as a valid authority and not as an old fashioned interpreter of Aristotle’ (p. 214). Furthermore, these thinkers were also influenced by Humanistic thought. Chapter IV pictures the philosophy of Judah Messer Leon, physician in Padua and ‘the foremost Italian “Hebrew Schoolman”’ (p. 25), as a Hebrew extension of the Paduan school and Hebrew Latin Averroism. Zonta presents the introduction and chapter I.1 of Messer Leon’s very lengthy Commentary on the Physics, which incorporates explanations of Averroes’ various commentaries, compared with parallels in Latin sources. Next follows a list of the Latin sources that are explicitly quoted by Messer Leon. This chapter, the longest of the four, also includes texts on the Isagoge that are attributed to Messer Leon. In sum, the volume not only demonstrates the existence of Hebrew Scholasti cism, it also shows that this was an innovative and important branch in Jewish thought. The texts presented support Zonta’s claim that ‘a fully fledged Hebrew Scholasticism characterized by the production of Hebrew philosophical works that use Latin Scholastic texts, doctrines and techniques in a way that is direct and clearly identifiable, systematic and mostly explicit does not come into being be fore the late fifteenth century in Spain and Italy’ (p. 13). The techniques derived from Latin sources concern the use of, for example, arguments pro and con, the expositio of Aristotle’s text, quaestiones disputatae, conclusiones, and reportatio by the Hebrew Schoolmen. Zonta emphasizes, however, that they were not ‘simply influ enced’ by their Latin colleagues: the relation of Hebrew Scholasticism to its Latin counterpart is not one of mere dependence, but rather one of parallelism (p. 2). The Hebrew texts also reveal criticism towards their sources (p. 44), and independ ent elaboration. Clearly, only a full exploration of all the hitherto unexplored texts can throw more light on the question of the relation of ‘influence’ vs. independence evinced by Hebrew Scholasticism, since the extent and nature of ‘influence’ may vary from one text to another. Zonta notes at the beginning of his study that the history of Hebrew Scholasti cism in the fifteenth century ‘is yet to be written’ (p. 2, n. 1). It is no exaggeration to say that Zonta’s own book constitutes a major step towards such a history. Hope fully, it will encourage further research in this promising field. doi:10.1093/jss/fgp024
RESIANNE FONTAINE UNIVERSITEIT VAN AMSTERDAM
MARTIN ZAMMIT, ¨Enbe men Karmo Suryoyo: A Syriac Chrestomathy (Gorgias Hand books). Gorgias Press, Piscataway, NJ 2006. Pp. xii + 207. Price: $65.00 hard back. ISBN: 1 59333 346 3. This chrestomathy will make an instant appeal to students of Syriac language and literature. It is beautifully produced, on good quality paper with clear and attractive fonts for the three scripts, and is an aesthetic pleasure. The book is intended primarily for students who have covered the essential fea tures of Syriac morphology and syntax. Twenty six extracts have been selected for inclusion. These extracts include thirteen readings printed in Estrangelo, ten in Serto, and three in East Syriac. Each of these entries is accompanied by a short in troduction in English and by extensive cross referenced grammatical notes. Since the book is not primarily aimed at first year students, but at readers who are further 599
REVIEWS
advanced, these grammatical notes do not discuss all basic matters, and a degree of competence is required to arrive at adequate translations. Dr Zammit has aimed at and achieved variety in his selection of passages: as Dr Brock says in his Preface, Dr Zammit has included extracts from authors and writings that are not widely known. These less well known texts take up a substan tial proportion of the book, about one third of the total. The obvious benefit of this selection is that undergraduates are introduced to material that they would prob ably not otherwise meet, but the price to pay for these choices is that the better known authors and texts are less well represented than they would otherwise be. For instance, the Peshitta passages are restricted to two, both from Luke’s gospel, taking up a little less than two pages. Interesting as these are, they are an inadequate representation of the Peshitta: the Old Testament Peshitta is mentioned only in passing. Future editions would be greatly improved by the inclusion of some Old Testament text: this is a document which is not only valuable to students of the ancient Versions and of the development of Christianity in the area of Mesopota mia, but also of great intrinsic interest. Some useful references to printed editions are given; references to the most im portant commentaries and critical studies would also be helpful to students wishing to expand their knowledge of the authors to whom Dr Zammit introduces them: for instance M.P. Weitzman, The Syriac Version of the Old Testament: An Introduc tion (Cambridge 1999); M. J. Pierre, Aphraate le Sage Persan (Paris 1988); H.J.W. Drijvers, Bardaisan of Edessa (Essen 1966) are all readily available. Perhaps because Dr Zammit was concerned not to overload his readers with in formation, some of the introductory passages are too short to give a rounded view of the works of these authors to readers who have no background knowledge. For instance, the introductory passage to Text 3, from Aphrahat’s fifth Demonstration ‘On Wars’ refers briefly to Aphrahat’s mention of Jewish practices but does not in dicate the importance of Aphrahat’s concern with Judaism or his attitude to it, a theme which runs so strongly throughout a number of the Demonstrations. The introductory passage to Text 4, from the works of St Ephrem, mentions the ques tionable attribution of some works to which Ephrem’s name is traditionally at tached; students would be helped by further information on this matter, specifying some of the more widely read of these texts of doubtful authorship. The Syriac English glossary is useful, covering as it does the wide range of the texts. The English Syriac glossary seems less valuable. Undergraduates do wel come it with delight, to some extent because it, like all this book, looks so attrac tive, and there are few such glossaries available. However, undergraduate students usually have little need of this approach, and the scope of the glossary is too re stricted for it to be useful to more advanced students who would probably work from a wider base or from a concordance suited to their particular interest. For instance, while ptgm’ is fully explained in the Syriac English glossary, in the Eng lish Syriac glossary ‘word’ is given only as mlt’, so the student searching for the appropriate word to render ‘the word of the Lord’ in working on much of the Old Testament Peshitta text would not find ptgm’. This glossary takes up almost one fifth of the total number of pages, an allocation which might be reconsidered in the next edition. The proof reading was not always adequate: for instance, on p. 5, line 7 of the notes, ‘uncommon is’ should read ‘uncommon in’; on p. 7, line 2 of the notes ‘weakend’ should read ‘weakened’; on p. 8, Introductory passage, para. 2, line 3 ‘he’ is omitted. 600
REVIEWS
At this price, students will probably look to their libraries for the opportunity to study this chrestomathy, rather than buying individual copies. doi:10.1093/jss/fgp025
GILLIAN GREENBERG UNIVERSITY COLLGE LONDON
SHEILA S. BLAIR, Islamic Calligraphy. Edinburgh University Press, Edinburgh 2006. Pp. xxxvii + 681. Price: £125.00 hardback. ISBN: 0 7486 1212 2. Within the study of Western manuscripts palaeography and codicology have devel oped as two distinct and independent areas of investigation. The description and codification of various script styles had long been established before recent research into the processes of manuscript production was recognised. This is not the case, however, in the field of Islamic manuscripts; both studies have progressed more or less side by side. Research into the production of handwritten books in Islamic soci ety and Islamic calligraphy are both relatively new disciplines that have become in extricably related and cannot be separated from each other. Sheila Blair’s monumental reference work Islamic Calligraphy is no exception, going beyond what its title at first suggests. Blair does not consider the study of Islamic script styles in isolation but, quite rightly, within an examination of the codex and the history of the art of the book in the Islamic world. How the differ ent scripts are used by scribes as part of the mise en page for example, or as both a decorative and functional tool within the illumination on the page, is extremely well documented, making this work very much a study of Islamic codicology in its wider sense. Hence, we find detailed sections on papyrus, parchment and paper; the preparation of inks and writing instruments; how sheets of quires were put together into book form; and the relationship of the scroll format to papyrus. To present the written word more often than not, the sacred word within an intricately executed calligraphy is an intrinsic feature of the visual aesthetic of Islamic art. While numerous examples from codices, albums and calligraphic sheets illustrate this point, Blair widens the scope by including any material that is used to carry calligraphic inscriptions, such as stonework, metalwork, textiles, ceramics and coinage even vegetation, such as the example of the technique among Ottoman calligraphers of writing on leaves and petals (p. 505, figure 11.13). Islamic manuscripts are of course the predominant source in this book for the study of calligraphy; particularly Qur’ans, on which Blair draws to illustrate varia tion of writing styles. This is hardly surprising as the history of Qur’an manuscripts is, to a great extent, the history of Arabic script. In their copying of the Qur’an cal ligraphers have not only made extensive use of wide ranging styles, but reflect dis cernable stages in the development of Arabic writing both geographically and his torically. Hence, Blair’s examples of calligraphy are taken not only from manu scripts produced in the Near and Middle East, but from the whole of the Islamic world, including India and the Islamic sub Continent; and South East Asia and the Malay Peninsula. The later dynastic styles of the Safavids and Qajars in Iran and Central Asia, and those of the Ottomans in Anatolia, the Balkans and the Eastern Mediterranean are also featured. This important reference work and detailed study is enhanced by some 150 col our illustrations including enlarged details of calligraphic features and illumina tion. In the section on ‘The standardization of round scripts under Ibn al Bawwab’, the opening and final pages of the Diwan of Salama ibn Jandal in the authenticated handwriting of Ibn al Bawwab (Topkapi Saray Museum, Istanbul), 601
REVIEWS
caught this reviewer’s attention (p. 170, figure 5.10 and p.172, figure 5.11). This Diwan of Salama ibn Jandal shares many similarities both in its layout and the jux taposition of scripts to the Diwan of al Îa∂ira (British Library, Add.26, 126) which, although claiming from the colophon to be also in the handwriting of Ibn al Bawwab, is nevertheless unauthenticated and rightly dismissed as a fourteenth century forgery (p.187). From such forgeries or suspicious attributions it becomes clear to what extent great calligraphers from whatever period were hugely revered and emulated. Indeed, what comes through Blair’s Islamic Calligraphy is a testament of tradition and repetition in the transmission of script styles once they have been developed, established and ‘canonised’ a testament that underlines the fact that the precise dating of undated Islamic manuscripts can never be a straightforward and simple undertaking. doi:10.1093/jss/fgp026
COLIN F. BAKER THE BRITISH LIBRARY
MARTIN WITTINGHAM, Al Ghazali and the Qur’an: One Book, Many Meanings (Culture and Civilization in the Middle East Series). Routledge, Abingdon 2007. Pp. viii + 154. Price: £70.00 hardback. ISBN: 978 0 415 37543 6. Martin Whittingham’s book grew out of a doctoral dissertation on the outstanding classical Muslim scholar Abu Îamid MuÌammad al Ghazali (450 505/1058 1111) under the supervision of Professor Carole Hillenbrand at Edinburgh. This is undoubtedly a well researched book which marks a milestone in the thriving study of al Ghazali and, more importantly, in the study of classical Muslim hermeneutics. While recent scholarship on al Ghazali has only secondarily dealt with his Quranic hermeneutics, the book under review is primarily devoted to this aspect of his thought. Nevertheless, Wittingham does not fail to engage at the same time with prevalent debates on other aspects of al Ghazali’s work, most significantly ‘the na ture of the relationship between the thought of al Ghazali and Ibn Sina (d.428/ 1037)’ (p. 8). Excluding the Introduction and Conclusion, the book comprises six chapters, neatly arranged into two parts which address al Ghazali’s hermeneutic theory (Part I) and practice (Part II). The dissertation form and style are evident in indi vidual chapters which are marked by introduction and conclusion sections, as well as orderly expositions on al Ghazali’s texts, sometimes but not always under the heading ‘Summary’. These sections are generally beneficial to the unfamiliar reader as a preliminary source of information on the contents of the texts examined, and particularly valuable for understanding the immediate textual context of al Ghazali’s discussion of hermeneutic issues. In Part I, Wittingham examines al Ghazali’s discussion of ta’wil in six different texts. In organising this Part, he distinguishes between non Sufi and Sufi influ enced texts alongside al Ghazali’s own epistemological distinction between mean ings pertinent to the activities of this world (e.g. law) and those pertinent to the activities of the Hereafter (e.g. the conditions of the heart) (p. 2). With a focus on non Sufi ta’wil, Chapters One and Two offer expositions of FayÒal al tafriqa bayna al Islam wa’l zandaqa and al MustaÒfa min ¨ilm al uÒul, respectively. Wittingham concludes that where FayÒal advocates toleration of ta’wil, defending its objectivity when informed by syllogistic logic, al MustaÒfa, a Shafi¨i polemic against Îanafi hermeneutics, seeks to restrict its practice (p. 36). Chapter Three, then, embarks on an investigation of ta’wil in four Sufi influenced texts concerned with the activities 602
REVIEWS
of the Hereafter and, hence, with hidden meanings of the Qur’an: (1) Kitab Adab tilawat al Qur’an (from IÌya’ ¨ulum al din); (2) Jawahir al Qur’an; (3) Mishkat al anwar; and finally (4) Kitab Qawa¨id al ¨aqa’id (also from IÌya’). Wittingham re veals how, in Adab, Jawahir and Mishkat, al Ghazali legitimates a Sufi approach to the Qur’an by underlining the complementarity of exoteric and esoteric interpreta tions (p. 39). In Qawa¨id, al Ghazali is then shown to ‘move between two models of ta’wil’, the Sufi model of complementarity, and the uÒuli framework for determin ing when a non literal (and non mystical) interpretation can displace a literal one (p. 64). With a shift in attention to al Ghazali’s actual interpretations of the Qur’an in Part II, Wittingham extensively re examines the Sufi influenced text of Jawahir al Qur’an in Chapter Four and that of Mishkat al anwar in Chapter Six. The analysis of Jawahir, in particular, highlights how al Ghazali employs hermeneutics to de velop his theological agenda; for example, to defend a classification of Quranic verses into pearls and jewels and to argue for the presence of the principles of all intellectual disciplines in the Qur’an and, hence, for a corresponding classification of all sciences into those of the pith and those of the outer shell. In Chapter Six, Wittingham investigates al Ghazali’s interpretation of the Light Verse (Qur’an 24:35), undertaking the unavoidable task of comparing al Ghazali’s interpretations in Mishkat with those of Ibn Sina in Kitab al isharat wa’l tanbihat and Fi ithbat al nubuwwat in order to assess the case for Mishkat’s borrowedness. Chapter Five introduces, for the first time, al Qis†as al mustaqim, a text which Wittingham deems to be primarily concerned with hermeneutic practice. A close reading of al Qis†as, however, casts doubt on the usefulness of the theory practice dichotomy adopted as the organising principle of the book. This, of course, is a well entrenched positivist distinction in modern thought. When employed in the study of the classical traditions of Islam in an effort to break down, for analytical purposes, the complex convergence and even fusion between theory and practice in Islamic thought, the resulting categorisation is, more often than not, reductive and forced. The text of al Qis†as highlights the limitations of this approach for it exemplifies how the discussion of ‘theory’, in this case al Ghazali’s syllogistic ta’wil, is inextricably embedded in the actual interpretations of what he considers to be cases of Quranic syllogisms. To his credit, Wittingham is aware that such ‘tidy dis tinctions… prove vulnerable under close scrutiny’ (p. 3) and, in the end, he con cedes that there is no ‘comprehensive theory’ of ta’wil to be gleaned from the dif ferent texts of al Ghazali (p. 128). The dichotomised approach, however, remains a feature of the book which, rather than facilitating clarity, engenders disruptive gaps in his analysis of the same text (e.g. al Ghazali’s interpretation of the Light Verse in Mishkat, Ch. 6, pp. 109 23 is discussed much later than Mishkat’s ‘Hermeneutical Theory’, Ch. 3, pp. 50 5); or of the same question in two differ ent but intertextually linked works such as FayÒal al tafriqa (Ch. 1) and al Qis†as al mustaqim (Ch. 5) where al Ghazali deals with syllogistic logic and ta’wil (see ‘Ta’wil and syllogistic logic’ in FayÒal, p. 25 and ‘The Qur’an and the syllogism in al Qis†as’, p. 89). Reservations regarding the dichotomous structure of the book aside, Witting ham’s achievement comes from his efforts to textually and historically contextualise al Ghazali’s hermeneutic discussions of ta’wil. He does this by pursuing a number of focal questions such as the purpose of the text examined and its target audience, the historical and intellectual background of its composition, and the development of debates within the literary genre to which it contributes. An illuminating exam 603
REVIEWS
ple is Wittingham’s informed and nuanced discussion of the purpose of al Qis†as al mustaqim, where he engages with and responds to Montgomery Watt’s questioning of al Ghazali’s motive for writing a work concerned with ‘trivialities’ such as Quranic syllogisms (pp. 82 3). Wittingham demonstrates, with textual evidence, that contrary to Watt’s view that al Qis†as was written for an unsophisticated read ership, this text was intended for an elite group of scholars (p. 86) and that, despite its formulation as a polemic discourse against an Isma¨ili opponent, its benefits are ‘more sublime’ than doctrinal correction as al Ghazali himself asserts. In addition to textual evidence, Wittingham also advances a contextual argument which locates al Qis†as in a classical intellectual climate hostile to syllogistic logic because of its foreign origins (p. 88). In this light, al Qis†as appears as an effort to promote syllo gistic logic amongst an unsympathetic scholarly elite by arguing for its grounded ness in the Qur’an, an effort reflecting at a more personal level al Ghazali’s double intellectual commitment to the Qur’an and to syllologistic logic which prompted him to integrate the two in this work (p. 101). Wittingham’s attention to contextuality moves beyond the immediate historical and intellectual questions arising from each text in order to consider al Ghazali’s thought in the larger scheme of his work. His elaborate analysis of Mishkat’s episte mology in Chapter Six deserves special note as it challenges the current scholarly position on this text, largely advanced through the work of Abrahamov and Davidson. Both these scholars concur that the Sufi language of al Ghazali’s inter pretation of the Light Verse in Mishkat conceals a theory of rational cognition in fluenced by Ibn Sina’s interpretation of the same verse in Kitab al isharat (p. 102). Acknowledging that there are several instances in the earlier work IÌya’ (surveyed on pp. 112 15) which suggest that al Ghazali perceived of no knowledge beyond rea son including knowledge by divine light, Wittingham upholds the possibility of intellectual development given the passage of time between IÌya’ and Mishkat (p. 116). He demonstrates that, already, in Ihya’ there appears to be more than one message about the paths to knowledge (see the discussion on al Ghazali’s notion of dhawq or direct experience in Ihya’, pp. 119 20). Wittingham further illustrates that in the later work of al Munqidh min al ∂alal, the superiority of the Sufi path over that of reason becomes more explicitly and genuinely articulated. Since al Munqidh was written only a few years after Mishkat, the dismissal of the latter’s Sufi language as a mere disguise by some scholars is shown to be much less justified than previously thought. A final issue which requires comment here arises from Wittingham’s preoccu pation with assessing the extent of al Ghazali’s borrowings from other thinkers, specifically Ibn Sina and the Ikhwan al ∑afa (the Brethren of Purity, c. tenth cen tury CE). But whereas the case for al Ghazali’s debt to Ibn Sina may be deemed evident, it is puzzling why the other ‘debt’ Wittingham pursues in this book is one to the Ikhwan al ∑afa and why, in this pursuit, he abandons the caution and rigour we see in his discussion of Ibn Sina’s influence. The argument for ‘an inter action with the Rasa’il of the Ikhwan al ∑afa’ as he puts it (p. 128) hinges upon a noted similarity between al Ghazali and the Ikhwan in their interpretation of Qur’an 13:17: ‘He sends down water from the sky making the riverbeds flow, each according to its measure. Then the torrent carries along swelling foam’. They concur that ‘water’ refers to the Qur’an and the ‘riverbeds’ to the hearts of men, but diverge on the meaning of foam; for al Ghazali, it is hypocrisy and dis belief and for the Ikhwan al ∑afa, it is the apparent meaning of the Qur’an itself (pp. 59 60). 604
REVIEWS
Given his disavowal of the Ikhwan al ∑afa in al Munqidh, al Ghazali, Witting ham argues, must have been familiar with their work and, in this case, he directly borrowed their interpretation but could not divulge his sources even though he was defending the displacement of the apparent meaning on the basis of transmitted in terpretation (riwaya, literally translated by Wittingham as ‘narration’, p. 60, but which technically refers to transmitted knowledge). A more plausible explanation is exactly the one Wittingham eliminates: that al Ghazali and the Ikhwan al ∑afa are drawing upon a ‘common source’ (p. 47). It is proposed here that this ‘common source’ is the hermeneutic tradition to which al Ghazali belonged. Norman Calder, whose work on classical hermeneutics1 is surprisingly omitted from this book (see bibliography), has masterfully delineated how in the process of meaning produc tion, classical exegetes were partaking in a tradition and drawing upon meanings ‘perpetually present’ therein. A quick look into this tradition proves the point. The tenth century Ibn Jarir al ™abari (d. 310/923) cites the early authority Ibn ¨Abbas (d.68/687) who proffers a non literal reading of ‘riverbeds’ in Qur’an 13:17 as sig nifying men’s hearts, and of foam as disbelief or doubt a meaning we find in al Ghazali but not in the Ikhwan al ∑afa.2 Al Sulami (d. 412/1021), al Ghazali’s pred ecessor and the author of the Sufi Qur’an commentary Îaqa’iq al tafsir, includes a reading of ‘water’ as the Qur’an attributed to the Sufi exegete Abu Bakr al Wasi†i (d.320/932),3 and later adopted by the famous Abu al Qasim al Qushayri (d. 465/ 1072) in his Sufi work La†a’if al isharat.4 The constant, though variant, recurrence of this non literal reading in earlier and later Sufi and non Sufi exegesis, all of which cannot be listed here, confirms that al Ghazali’s description of this reading as transmitted without mentioning his authori ties, is more likely to be indicating its wide currency rather than its dubious origin. In fact, it is not too farfetched to suggest that the Ikhwan al ∑afa were themselves drawing upon the hermeneutic tradition in their interpretations of the Qur’an. Other interpretative correspondences (Wittingham finds two less significant exam ples, p. 71 and p. 117) should be assessed in the light of the exegetical genre before any general claims, such as we find in this book, can be made regarding the Ikhwan al ∑afa’s influence on al Ghazali. This book is, nevertheless, a very valuable contribution to the study of classical Islamic thought and will be beneficial and accessible to students and scholars with diverse interests, ranging from Islamic philosophy, theology and law to Quranic and tafsir studies. doi:10.1093/jss/fgp027
SHURUQ NAGUIB UNIVERSITY OF LANCASTER
1 Particularly, Norman Calder’s two seminal articles: ‘Tafsir from ™abari to Ibn Kathir: Problems in the Description of a Genre, illustrated with Reference to the Story of Abraham’, in G.R. Hawting and Abdul-Kader A. Shareef (eds), Approaches to the Qur’an (Routledge, London 1993), 101 40; and ‘History and Nostalgia: Reflections on John Wansbrough’s The Sectarian Milieu’, Method and Theory in the Study of Religion, 9:1 (1997), 47 73. 2 MuÌammad ibn Jarir al-™abari, Jami¨ al-bayan ¨an ta’wil ay al-Qur’an (Dar al-fikr, 30 volumes, Beirut 1984 5/1405), Vol. 13, p. 135. 3 Al-Sulami, MuÌammad b. al-Îusayn b. Musa al-Azdi, Sayyid ¨Umran (ed.), Îaqa’iq altafsir, (Dar al-kutub al-¨ilmiyya, 2 volumes, Beirut 2001/1421), Vol. 1, p. 331. 4 Al-Qushayri, ¨Abd al-Karim ibn Hawazin ibn ¨Abd al-Malik, ¨Abd al-La†if Îasan ¨Abd al-RaÌman (ed), La†a’if al-isharat, (Dar al-kutub al-¨ilmiyya, 3 volumes, Beirut 2000/1420), Vol. 2, p. 105.
605
REVIEWS
THOMAS HERZOG, Geschichte und Imaginaire: Entstehung, Überlieferung und Bedeut ung der Sirat Baibars in ihrem sozio politischen Kontext (Diskurse der Arabistik 8). Harrassowitz Verlag, Wiesbaden 2006. Pp. xviii + 972. Price: /138.00. ISBN: 978 447 05089 0. Fact and fiction are not always strictly distinguished in historiography. This is even more true for the Mamluk era, as contemporary Arabic chronicles use narrative and anecdote throughout to make points. The reliability of these stories is hard to tell. On the other hand one has to remark that Mamluk literature does contain or is inspired by real events. Geschichte und Imaginaire by Thomas Herzog is settled right in the heart of the fact fiction complex and represents in every sense of the word a very impressive work. 972 pages disperse the spirit of profound scholarship. The work is the revised version of a Ph.D. thesis which was submitted at the Martin Luther Universität Halle Wittenberg and the Université de Provence, Aix en Provence. This Franco German undertaking is clearly enriching as Herzog combines convincingly modern French approaches in literary studies and social sciences with solid German philo logical skills. The subject of Herzog’s book is the Sirat Baibars (novel of Baybars) a popular Mamluk folk tale, written by anonymous authors, staring as hero the first great Mamluk sultan Baibars (r. 1260 77), who fought successfully against Mongols and Crusaders. Besides the depiction of epic fights against non Muslim enemies, the Sirat Baibars provides a real kaleidoscope of the contemporary Mamluk society. It has remained highly popular throughout modern times and is recited by storytellers in coffeehouses until our present days. Herzog has examined the Sirat Baibars on a remarkable textual basis. He has in tensively examined fifty one different manuscripts and two print editions of the saga. The manuscripts were written down between the sixteenth and twentieth cen turies, most of them dating from the seventeenth and eighteenth centuries. They come mainly from Egypt and Syria and bear witness to the high popularity the epos enjoyed among coffeehouse guests of the Arab world. The first aim of Herzog’s study is to ‘open the window of the mental worlds of the authors, narrators and audience’ of the Sira in order to understand ‘the imagi nary’ in it (p. 1). Herzog wants to examine where the work is situated and how these places are rated by the audience. He intends therefore to draw the mental maps behind the Sirat Baibars. Secondly he would like to examine the question of the origins and developments of the work through an analysis of the different manuscripts. The third point which is tackled by Herzog is to place the work within its socio political context in order to find out why it has been so popular for so long. In order to tackle these questions Herzog has undertaken the following. He presents in part A (pp. 31 49) the textual evidence and explains how the different versions of the Sira are interrelated. Part B (pp. 53 166) deals with the structure and main protagonists of the work and provides a short résumé of the story for those who are not that familiar with the work. Part B also contains a translation of the very important ‘Genoa Episode’. This translation is vital for the further analysis by the author. Moreover, Herzog explains in this part (pp. 157 66) the genre of the popular novel of heroes, the Arabic siyar literature. In part C (pp. 167 392) Herzog deals with the imaginaire (imaginary) in the Sirat Baibars and compares the protagonists and the stories of the novel with the actual socio political context of the Mamluk era, before he summarizes his findings 606
REVIEWS
in part D (pp. 393 424). Finally, he provides a huge appendix (part E, pp. 425 905), where he adds valuable details on the manuscripts and tables which show their textual dependencies by comparing the various subtitles of the different manuscripts. Well, this part is really something for Sirat Baibars enthusiasts as the author acknowledges himself. Others can only admire the amount of hard work required and might turn directly to the abridged version of the Sirat Baibars (pp. 861 905) where Herzog recalls the main story line of the Sira following the chapters of the Arabic print version. Concerning the authors of the Sirat Baibars, Herzog argues that they were cer tainly persons near to the Mamluk educational establishment as they cite works of Mamluk historiography, but they had also listened carefully to the popular stories circulating in the streets. A first version emerged by the early sixteenth century which was to be modulated over the following centuries. Herzog’s main achievement is that he provides us with a very good insight into the mental maps created by or incorporated in the Sirat Baibars by combining the methods of literary studies and historical studies. Very convincingly he shows the different spaces which are described in the Sirat. There is the space outside the Is lamic world, which is the region of danger, inhabited by Franks or Mongols who continuously attack or trick the Muslims. Needless to say that they fail as Baibars and his friends fight them back. The sea is quite often a metaphor for danger as the Christian pirates come from there. To fight these enemies is the duty of all Muslims and one wonders about the positive role the mighty Isma¨ilite warriors (pp. 176 81) play fighting side by side with Baibars, as the Sunni legal scholars of Mamluk times clearly regarded them as heretics, but in the memory of the common people their religious oddness was ap parently of minor importance compared to their historical role in fighting the Cru saders. The Islamic world is generally depicted as a safe haven against this outer world, but one should be mindful of the ‘false bottom' and the enemies from within. The town of Alexandria for example has an underground level consisting of tunnels and caves where Baibars was once taken captive in an underground church by his main enemy the mighty Jawan, who is a magician and crypto Christian in the disguise of a Muslim qa∂i. In particular, the first part of the Sira until Baibars’ ascendancy to the sultanate is a tale of the lower stratums of the society, who also constitute the majority of the audiences in the coffeehouses. Their world is depicted with sympathy and is full of anecdotes taken from daily life. The main protagonist here is Baibars’ friend ¨Uthman, the robber with the golden heart. Moreover, throughout the novel Sultan Baibars protects common people against abuses of his officials. Besides drawing these convincing images of the different spaces within the Sirat Baibars, a main achievement of Herzog lies in the historical contextualisation of the stories of the saga. For Herzog, the stories with an historic background, like the fight scenes against the Franks, date from the late fourteenth and early fifteenth century and are not immediately connected with the life of the historic Baibars. The time singled out by Herzog represents a time of constant threat by Frankish naval forces culminating in the sacking of Alexandria in 1365 by the Cypriote King Peter I of Lusignan (r. 1359 69) and coming to a an end only when Sultan Barsbay (r. 1422 38) finally subdued the kingdom of Cyprus in 1426 and reduced it to a vassal which had to pay annual tribute to the Mamluk sultan. For Herzog, the Genoa episode of the Sira when Baibars is taken as a captive to Genoa, triggering 607
REVIEWS
thereby a successful attack by Mamluk forces on the town, reflects this warfare against Cyprus. As someone who has worked on these aspects from the side of Mamluk historiography I fully agree with Herzog’s findings concerning the classifi cation of the historicity of events narrated in the Sira.5 However, what struck me most is the great extent to which the Christian Franks are apparently described in the Sira as an eternal threat for the Muslim realm at a time, especially after the mid fifteenth century, when the actual military threat had diminished to finally almost disappear after the Ottoman conquest of Cyprus in the 1570s. However, the image of the hostile Crusader lived on in the memory of the local Egyptians and Syrians and it was certainly strengthened after Napoleon launched his expedition against Egypt in 1798 and during the following period of French and British colonialism. To sum up: Herzog’s work will certainly remain for a long time ‘the’ reference work for the Sirat Baibars. It is of interest for any scholar of Mamluk and Ottoman times and also for scholars of comparative literary studies as it provides an open window to numerous aspects of the mentality and folk literature of the people of the Near East in Mamluk and post Mamluk times, or to put it more briefly: ‘Wel come to the Mamluk mind!’ doi:10.1093/jss/fgp028
ALBRECHT FUESS UNIVERSITÄT ERFURT UNIVERSITÉ FRANÇOIS RABELAIS, TOURS
ARMIN SCHOPEN, Tinten und Tuschen des arabisch islamischen Mittelalters: Doku mentation Analyse Rekonstruktion (Abhandlungen der Akademie der Wissen schaften zu Göttingen: Philologisch Historische Klasse, Dritte Folge, Band 269). Vandenhoeck & Ruprecht, Göttingen 2006. Pp. 264 + 23 plates. Price: /76.90 hardback. ISBN: 978 3 525 82541 9. In a time when interdisciplinary research is considered a yardstick of academic ‘ex cellence’ by many universities but for bureaucratic and political reasons starved of adequate funding and denied the necessary organisational framework, Armin Schopen is one of those scholars who actually live up to the proclaimed standards. Ironically he succeeds not because he is embedded in but rather because he operates on the margins of prevailing institutional structures. Schopen is an ethnologist by trade, and he combines the virtues of this discipline with a thorough knowledge of classical Arabic and a keen interest in all matters relating to the material culture of the Middle East, past and present. His latest book, reviewed here, is another very fine contribution, testifying also to the importance of protected time, personal space, scholarly integrity, humility, and a clear sense of purpose in achieving those high academic standards. The book deals by way of documentation, analysis, and reconstruction with the development and manufacture of black and coloured inks in Islamic civilisa tion. It is based on two clusters of evidence: first, on a critical examination of notably alchemical and technological but also general Arabic literary sources, many of them only preserved in the form of manuscripts and hitherto largely neglected; second, on the results obtained from a replication of the attested ink recipes under laboratory conditions, using natural raw materials in temperatures similar to those of the south 5 See for example Albrecht Fuess, ‘Was Cyprus a Mamluk Protectorate? Mamluk Influence on Cyprus between 1426 and 1517’, in Journal of Cyprus Studies, 11 (2005), [28/29], 11 28, and ‘Rotting Ships and Razed Harbours: The Naval Policy of the Mamluks’, in Mamluk Studies Review, 5 (2001), 45 71.
608
REVIEWS
ern Mediterranean. If the theoretical part of Schopen’s book is already a masterclass in source criticism, the practical part is a tour de force in its own right, bridging the often lamented gap between the social and natural sciences and bringing a dead sub ject back to life. Thus, Schopen’s book is of the greatest significance not only for the study of Islamic book culture and the scientific elements in Arabic alchemy but also, and perhaps more importantly, it confirms the overall accuracy of the information provided to us by the sources; and yes, we can identify the various vegetable and mineral substances occurring in these old texts and express them in Linnean terms or periodic symbols. Schopen has covered all aspects of the subject so well and explored the potential of the material so comprehensively that his book, I am sure, will be come the standard work on ink making in medieval Islam for generations to come. Schopen sets out with a brief but highly instructive introduction (pp. 9 17), dealing with the different types of ink and basic Arabic terms (midad, Ìibr, liqa, niqs), the different kinds of writing material (papyrus, parchment, paper), how and why ink making in medieval Islam was gradually developed into an art form, and how black and coloured inks were actually made. Next follows a very impressive account of the Arabic sources Schopen has studied in order to collect his ink recipes (pp. 19 32), listing (and describing!) in alphabetical order some three dozen texts (plus excerpts and fragments), more than half of which are manuscripts scattered all over the planet and many not exploited before; space does not permit to go into greater detail, so suffice to note a privately owned autograph copy of an important work entitled al Azhar fi ¨amal al aÌbar, written by the philosopher alchemist MuÌammad ibn Maimun al Marrakushi al Îimyari (fl. 1251 CE). The main part of Schopen’s book (pp. 35 180) consists of a meticulously annotated German translation of 165 ink recipes (not counting variant transmissions), arranged ac cording to different ink types: black and ‘India’ ink made from soot, charcoal, iron gall, or tannin (with numerous subdivisions); ‘mixed’ ink; white ink; and coloured ink (pink, red, orange, gold, yellow, green, blue, violet). Various glossaries (pp. 183 236), which in their high degree of elaboration look more like a collection of small articles, are dedicated to the base materials, binding agents, and additives which were employed in the production of black and coloured inks, ending with a short table of converted weights and measures. An extensive bibliography (pp. 237 46) is followed by various indices (pp. 247 64), covering substances (vegetable, mineral, animal) and utensils in German Arabic lists, then personal names and Linnean terms, and finally a reverse Arabic German register. The book concludes with 67 colour photos on 23 plates, among them a ‘colour chart’ reconstructed ac cording to the famous Arab alchemist Jabir ibn Îaiyan (fl. 800 CE); the majority of these illustrations show Arabic script samples written by Schopen himself on papy rus, parchment, or paper using the very inks he resurrected from the sources. The book is written in clear unpretentious German, printed on good paper, bound in an attractive format, and reasonably priced. The only minor flaw is that the ink recipes, which Schopen translated so carefully, do not appear in the original Arabic as well, especially since so many of them are not available in printed form alas, this would probably have blown Schopen’s task (and the price of the book) out of all proportion, and in any case his translation is furnished throughout with rel evant Arabic terms in bracketed transliteration. All in all, this book is a wonderful achievement and deserves to be on the shelf of anyone interested in the material culture of the Islamic world. doi:10.1093/jss/fgp029
OLIVER KAHL UNIVERSITY OF MANCHESTER
609
REVIEWS
NAWAL NASRALLAH (ed. and trans.), Annals of the Caliph’s Kitchens: Ibn Sayyar al Warraq’s Tenth Century Baghdadi Cookbook. English Translation with Introduction and Glossary (Islamic History and Civilisation 70). Brill, Leiden 2007. Pp. xii + 867 + 35 plates. Price: /139.00/$195.00 hardback. ISBN: 978 90 04 15867 2. The sub discipline of culinary history, within the broader rubric of Middle Eastern cultural history, has developed in interesting ways recently. Modest but important beginnings were made in the 1930s with articles in Arabic published in Beirut by Îabib Zayyat. He may, indeed, have influenced the subsequent and significant contributions of the late Maxime Rodinson during his sojourn in Lebanon in the 1940s. Following Dawud al Chalabi’s discovery and published edition (1934) of the thirteenth century cookbook by al Kha†ib al Baghdadi, A.J. Arberry’s transla tion (1939) of it marked another step forward by drawing the attention of a wider audience to the pleasures of the medieval Arab table. In the same translation he also noted the existence of a manuscript in the Bodleian Library, also a Kitab al ™abikh, by one Ibn Sayyar al Warraq; he judged the author to have composed (or, perhaps better, compiled) the work during the tenth century in the Abbasid era. It was not until 1987, however, when two scholars, Kaj Ohrnberg and Sahban Mroueh, published in Helsinki an edited version based on the two then available manuscripts. It remains to date the earliest extant cookbook in Arabic of what is obviously a long culinary tradition pre dating both the rise of Islam and the Per sian dynasties. And this brings us to the present volume under review. The Iraqi scholar, Nawal Nasrallah has taken medieval culinary studies further still with this award winning volume. It is, of course, a translation of al Warraq’s Kitab al ™abikh, but then, it is a good deal more as well. First, she has utilized a third manuscript lodged in the Topkapi Sarayi, Istanbul, which had been unavailable to the Finnish editors. This copy, used to verify the Helsinki edition, is an adaptation of it, in the sense of its being both abridged and augmented. Its description is contained in the first of a three part, sixty page introduction, the second contribution Nasrallah has made to this edition in general. The first part covers the sources and legacy of al Warraq’s cookbook; part two deals with the essentially bourgeois food culture not exclu sively courtly as sometimes said of Baghdad and its origins; part three, impor tantly reflecting the cookbook’s contents, is a discussion of Galenic humoral theory, pointing to the important interface between the medieval culinary and medical tra ditions, especially in the latter’s dietetic and pharmacological divisions. The book’s first twenty odd chapters deal with the humoral properties of a wide range of ingre dients used in cooking. The translation naturally covers the bulk of the volume, just over four hundred and fifty pages. The English text is marked with running pagination of the Arabic text in bold characters so that locating Arabic passages is straightforward. The vari ety of dishes and the detailed instructions provided for their preparation never ceases to amaze, in addition to the fact that, with a combined affection for the past and for present adventure, one can adapt the recipes to the modern palate and kitchen. Moreover, to the best of this reviewer’s knowledge, the medieval Arabic culinary corpus as a whole (al Warraq is but one of several works) is a documentary source for culinary history unparalleled in any other contemporary culture, western or eastern, for the period roughly between the tenth and fifteenth centuries. Nasrallah’s translation reads easily, including the not infrequent samples of poetry, interspersed among the recipes; often, too, a poem gives a clue as to the nature and purpose of the preparation, as we will see below. The translation ends with an ap 610
REVIEWS
pendix identifying the names of people and places mentioned in the text. This is very useful for both the historian and general reader alike. The final portion of the volume consists of what is modestly described as a ‘glos sary’; it is three hundred pages in length and represents an immense act of labour (and, one imagines, of devotion) on Nasrallah’s part. The bulk of the glossary com prises brief notes on all the ingredients of dishes, kitchen utensils, cooking tech niques, medical terms and ailments, personal hygiene and weights and measures. This is followed by a shorter English Arabic glossary intended to assist the reader in locating items in the preceding detailed Arabic English section. The indices com prise four sections, one each on people and places; ingredients, preparations, and implements; prepared foods and drinks; and medical terms, preparations, hygiene and perfumes. There is also a section on ‘works cited’ in the volume. Between trans lation and glossaries are thirty five coloured plates, almost exclusively drawn from medieval illustrated manuscripts, depicting people engaged in food preparation, serving and enjoying the fare. It would be a publishing miracle if a volume of this size and complexity were totally free of error. To mention any might appear to be quibbling, for they are few. However, in the event of further printings (which one must hope will be the case) the following may be taken into account. The English term ‘decorums’ (pp. 502, 505 for adab) is not plural and should be ‘etiquette’ or ‘manners’ as on page 508. On page 126 in the recipe for sawiq with almond has ‘mastic’ when the Arabic term is ‘misk’ (musk). On page 138 the ingredient ‘nutmeg’ is rendered as jawz bawwa and on page 643 it is jawz al †ib, without explanation, although in the index of in gredients the two Arabic terms are given as synonymous. There is also another slightly different version of afwah al †ib (atraf al †ib), p. 643, found in the anony mous cookbook Kanz al fawai’d, recipe 434. Finally, the description of the chapter on recipes for fuqqa’ (number 112) as ‘alcohol free beer’ does not appear warranted from the context. Nasrallah points out in a footnote (p. 453) that the beverage could become alcoholic if left longer than a few hours to ferment. However, in two poems on the beverage (pp. 458, 459), Ibn al Mu¨tazz describes the drink as ‘a cure for hangovers’ and Abu al Îasan al Kha†ib advises ‘drink it and the hangover fiz zles’. Anglo Saxon revellers will easily recognise this potion as the ‘hair of the dog’, a dose of alcohol to cure a hangover. The term ‘bubbly drink’ as Nasrallah renders fuqqa’ (p. 453) or alternatively, ‘frothy’, itself suggests the process of fermentation. That said, this is an impressive piece of scholarship and a joy to read both for pleasure and for more sober historical reflection on the ways medieval bourgeois Arab Muslims and also Christian and Jewish urbanites, enjoyed themselves at table. doi:10.1093/jss/fgp030
DAVID WAINES LANCASTER UNIVERSITY
ÎUSAYN IBN ¨ALI DAÎIL ALLAH ABU AL ÎASAN, Nuqus liÌyaniya min min†aqat al ¨Ula: dirasa taÌliliya muqarana. Wizarat al ma¨arif wakalat al wizara li al a†ar wa al mataÌif, ar Riya∂ 1423 H./2002 M. Pp. 395. ISBN: 9960 19 545 7. The settlement of al ¨Ula (a desert oasis in the northern part of Saudi Arabia, for merly known as ¨Dadanˆ, the Arabic name of the oasis which is better known in the West in the Biblical Hebrew form ¨Dedanˆ, (see A. Sima, ¨Zum antiken Namen der Stadt Dedanˆ, BN 104 [2000], 42 7) was a stopping place for caravans crossing the desert of the Arabian peninsula on the inter regional trade routes. Its rocks and stones are covered with inscriptions and graffiti the language of which (an an cient North Arabian dialect) came to be written down in the first millennium BCE. 611
REVIEWS
This is the author’s doctoral dissertation submitted to the University of Riyadh (Saudi Arabia). It consists of a study (decipherment, translation and commentary) of 151 new Dadanitic inscriptions (formerly Lihyanite, see Michael C.A. Macdonald, ¨Reflections on the Linguistic Map of pre Islamic Arabiaˆ, AAE 11 [2000], 33). These inscriptions were selected from a group collected by the author himself in the oasis of al ¨Ula and its environs, including the inscriptions found in the city of al Îigr, currently Madaˆin ∑aliÌ. The book is the second part of a new corpus of 347 Dadanitic inscriptions marked with the symbol Abu al Îasan and it contains the inscriptions from number 197 to number 347. The first 196 inscriptions are com piled in the master dissertation by the same author, Qiraˆa li kitabat liÌyaniya min gabal ¨Ikma bi min†aqat al ¨Ula (ar Riya∂ 1997). The book under review consists of a presentation, acknowledgments, prologue, abbreviations of foreign works and four main parts. The first part deals with inscriptions found in Gabal ˆUmm Darag (number 197 to 287). There appeared a letter form in the inscriptions of this area which was interpreted by some authors as the phoneme /Â/ and by others as the phoneme /†/. The result is that Abu al Îasan reads the verb ˆ†ll with /†/ and trans lates it as ¨to offer, to submit (a gift to the Gods)ˆ. By contrast, A. Sima, Die lihyanischen Inschriften von al ¨U∂ayb (Saudi Arabien) (Epigraphische Forschungen auf der Arabischen Halbinsel 1, Rahden/Westf. 1999) prefers to read the verb as ˆÂll and to translate it as ¨decken (den unterirdischen Wasserleitungskanal)ˆ. Fortu nately, in inscription number 197 (pp. 29 39), one of the largest and clearest of all the Dadanitic inscriptions, there appeared two different letter forms for these two phonemes. This supports the proposition of reading the above mentioned verb as w ˆÂlw and Âll in line 6, since the well know letter form for the phoneme /†/ is, without a doubt, that from the verb †¨n in line 8. However, Abu al Îasan reads the two letter forms as /†/, and therefore presents a translation very different from other authors. The second part is devoted to the inscriptions collected in Danan (number 288 to 299) and Abu ¨Awd (number 300 to 311). For the first time in the Dadanitic corpus appears the goddesses name ¨trgth in inscription number 288/2 (pp. 218 22). It is the goddess ¨A†tar, the Greek Atargatis. The third section places the inscriptions of Gabal ˆI†lib, in al Îigr (number 312 to 347). It is mostly graffiti of no more than three lines, many of them framed by a circle. Often mentioned in them and only in the graffiti area, is the expression n†r ddn, which is translated as ¨controller, observer for Ddn (Dadan)’. The name of the settlement Dadan also ap pears without being related, as is common, to a governor or monarch. The fourth part contains an analysis of the Dadanitic script and religion and politics in the oa sis of Dadan, an alphabetical index, the copies of the inscriptions, photos of some places where they appear, maps and tables. Finally, the author presents the bibliog raphy in two lists, one of works written in Arabic and the other of non Arabic works. It should be noted that in order to carry out a good epigraphical study, several points must be taken into consideration: Firstly, one must have the direct source in other words, the inscriptions, or good photos of them. In this regard, the book is useful because each inscription is presented with a colour photograph. Secondly, one must offer a suitable transliteration through the use of editorial conventions. As to the latter, it is remarkable that the author hardly draws attention, for instance, to the use of the conventions for the transliteration whose meaning must be deduced by the reader because they are not clarified in the book. Inscription number 221 (pp. 99 103) is shattered on the right side, so the reading of the beginning of the lines is illegible. The author uses the signs xx to express an undetermined epigraphi 612
REVIEWS
cal space. In inscription number 224 (pp. 110 13) two small fragments containing the beginning of the first and last lines are missing on the photograph. In this case, the author does not use the sign xx as one would expect, but just three dots. In in scription number 225 (pp. 113 15), which is broken on the upper left and the lower right side, the number of dots varies from two to three from one line to other, even those used in the same instance. As far as the reconstruction of words is concerned, in inscription number 212/1 (pp. 74 6), the author suggests that ¨bd should be the first element of a theophoric name in combination with the name of the deity ∂gbt. In the Dadanitic corpus there are several compound names with the god’s name ∂gbt, e.g. ˆs∂gbt, zd∂gbt, mrˆ∂gbt. Therefore, there is no reason to prefer one of them since the inscription is seriously damaged on the right side and there are no traces of any letter form. On the other hand, inscription number 242 (pp. 154 7), which was restored from two fragments, is probably two different inscrip tions. In the right part, the letters are incised, but in the left part the letters are carved in relief. It is unreasonable to think that this is due to a mason’s random decision or carelessness. A Dadanitic inscription for comparison is JSLih 52 in Jaussen, A. and Savignac, R., Mission archéologique en Arabie, II. El ¨Ela, d’Hégra à Teima, Harrah de Tebouk, plate LXXXI (Paris 1914). In this inscription the first lines are carved; however, only the last surviving line is incised. Michael C.A. Macdonald, ¨Ancient North Arabianˆ, in Roger D. Woodard (ed.), The Cambridge Encyclopedia of the World’s Ancient Languages (Cambridge 2004) pointed out the misconception of considering the Ancient North Arabian lan guages: ¨as a collection of early dialects of Arabic’ (p. 488). In the desire to see ‘the Arabicˆ in the Dadanitic inscriptions, many authors have forced the etymology of the word and their imagination to attach the text, which they had in hand, to the Arabic outline without bearing in mind other possible Semitic contexts. The situa tion becomes worse if the author hides from the reader the sources from which he has drawn. For instance, in inscription 202/2 (pp. 50 2), where the sequence hm∂ trk appears, the author has compared this with the Arabic mun∂u taraka but does not present any source on which he has based this nor an explanation of why he chose this translation. The inscription also presents problems in the sec ond line, since so many signs are unclear. Similarly, in inscription number 207/1 (pp. 62 4) the author has considered the verb st†r as an Arabic past tense and translated it as ¨to write, to compose, to registerˆ without citing any source to sup port his claim. Putting these details aside, the book deserves praise for introducing new sources for the study of the Dadanitic language. Thanks to these new 151 inscriptions, one is able to expand on one’s knowledge of this language. In the future one can hope that more inscriptions will be available for scholars to extend the information about this language and the culture and history of the people who used it. Finally, due to the fact that the book deals with a specified subject, and it is an Arabic edition, it was difficult to find it in libraries and universities in Europe. I am very grateful to the author who allowed me to have a copy of his work and Dr Salem Ahmad Tairan who handed it over to me. doi:10.1093/jss/fgp031
MARÍA DEL CARMEN HIDALGO-CHACÓN DÍEZ UNIVERSITY OF MARBURG
613
REVIEWS
HELÉNE KAMMENSJÖ, Discourse Connectives in Arabic Lecturing Monologue (Orien talia Africana Gothoburgensia 19). Acta Universitatis Gothoburgensis, Göteborg 2005. Pp: 219. ISBN: 91 7346 541 0. This ‘slightly revised version of a doctoral thesis’ has as its aim ‘to develop an alter native approach to the study of conjunctions, conjunctive adverbs, phrases and clauses in modern Arabic and to test it on a corpus of Formal Spoken Arabic’ (p. 17). The term ‘(discourse) connectives’ signals to the reader ‘that we are operating within the field of Discourse Analysis (or text linguistics) and not in traditional “sen tence grammar”’ (ibid.), and connectives are defined by their function of linking utterances in discourse, in combination with their utterance initial position, rather than by form. The other major analytic concept is discourse organization: ‘the basic segments of discourse and the functional relationships that hold between them’, and thus the work will examine ‘in what ways and to what degree the connectives contribute to signal these relationships’ (ibid.). The volume is organized in four sections and twenty two chapters: Background (chaps 1 5), Data (6 7); Methods (8 14); and Results (15 22); + Appendices (in fact only one, Appendix C, but there are references here and there in the text to fur ther appendices and charts, which I guess have been edited out); Abbreviations; Bib liography and Index. The background chapters discuss the treatment of Arabic connectives as found in traditional grammars Arabic and Western (chap. 2), didactic problems related to the neglect of the pragmatic use of connective elements, and introduce the terms ‘discourse connectives’, or ‘markers’, as developed in text linguistics and pragmatics (chap. 3). A theoretical framework is drawn up (chap. 4), rooted in the notions of Cohesion and Coherence, Relevance theory, and Grammaticalization. The theoretical underpinnings of Halliday’s distinction between ‘information structure’ and ’the matic structure’ is not introduced here, but rather in the section on Methods (chap. 11, pp. 99ff.). The two recent publications on Arabic connectives and discourse markers according to the ‘modern’ framework, al Batal (1994) and Sarig (1995), are discussed in chapter 5. Before the Arabic Lecturing Monologue [ALM] corpus is presented in chapter 7, Kammensjö summarily runs through the ‘classics’ on Arabic diglossia, language levels and style variation (chap. 6). In the Methods section, we are taken through short chapters on Data Collection and Transcription (chaps 8 9), while chapter 10 presents various approaches to and criteria for defining segments of discourse: sentence boundaries, punctuation, con text dependent text sentences (Lyons), syntactic, prosodic and semantic criteria; and an applied section/sample study on prosodic units in the corpus with prosodic boundaries in terms of relative pause length combined with falling and rising tone. The discussion concludes that ‘only a combination of different criteria can help us divide our data into meaningful units, which can be taken as basic in discourse analysis’ (p. 97). The author further opts for the ‘utterance’ as the preferred term for these units, which in the final analysis appears to be syntactically grounded. Chapter 11 introduces Halliday’s Systemic Grammar, with the three level struc tural systems and the thematic structure of the textual system, and briefly argues that the model may be applied to Arabic in spite of its alternating V(S) and SV word order.6 ‘Underlying assumptions about Discourse organization’ are repeated in 6 Mona Baker, In Other Words. A Coursebook on Translation (London 1992), 127f. mentions problems related to different thematic organization of Arabic and English, in the context of translating between the languages.
614
REVIEWS
the short chapter 12. In chapter 13 we are introduced to alternative models for dis playing discourse organization: the ‘hierarchy’ or ‘constituency’ model, and the ‘flow’ model, with utterance boundaries marked for either ‘flow sequence’ or ‘flow shift/break’. Then, in chapter 14, theoretical and methodological tools for analysing semantic relationships between text segments are shortly presented, from Halliday and Hasan’s classic 1976 work on cohesion (M.A.K. Halliday and R. Hasan, Cohesion in English [London 1976]) to ‘Rhetorical Structure Theory’ and ‘Semantic Structure Analysis’. In the Results section, the Arabic connectives in the ALM corpus are identified and counted (chap.15); chapter 16 shows their distribution over utterance bounda ries. The feature of topicalization is found to be a dominant feature of the discourse of the lectures. The author explores this at length in chapter 17, making the con nection to thematic structure through the concept of framing, which is also essen tial to her treatment of multiple connectives and their internal order in chapter 18. The claim that ‘topicalized themes, through their fronted exposition marked by framing, indeed belong in the “connective slot”’ (p. 123) is convincingly argued. Throughout, the results from the ALM corpus are compared with two sets of written ‘reference material’: ‘Modern Arabic Didactic Discourse’ (MAD) and ‘Old Arabic Didactic Discourse’ (OAD) a modern textbook on human geography and a ninth century medical treatise. In chapter 19, individual differences among lecturers are discussed while the charts and tables for each speaker announced as Appendix B are missing. In chapter 20, the role of the connectives are treated in relation to the ‘flow’ model, and in 21, their function as markers of semantic prag matic relationships. The major achievement of Kammensjö’s book is how she systematically and con sistently applies more universal sets of discourse analysis to an important field of Arabic grammar, and thereby opens up higher level perspectives on how a range of grammatical words, collocations and clauses, across grammatical categories, share certain functions and are distributed in texts for certain functions. Her contribu tion also demonstrates ways in which the Arabic language works, or rather is put to work, to achieve certain functions, in this case cohesion and coherence. The present reviewer found most rewarding the discussion and results on topicalization and the matic structure, while sometimes counts and classifications appear to be set up for the sake of comprehensiveness, and do not always yield significant information. The many charts comparing the ALM data with the modern and old written texts also do not, in my view, yield much new information, but do give evidential sup port for the historical development of Arabic written style towards a greater variety of connectives and more complex sentence structures. An explicit secondary aim has been to ‘throw more light on a modern variety of spoken Arabic, here called FSA (Formal Spoken Arabic)’ (p. 199). In chapter 7 the ALM corpus is defined as representing Formal Spoken Arabic, a sub category of Educated Spoken Arabic,7 and (with reference to Ryding’s textbook) supposed to represent a ‘standardized variant of spoken Arabic’. ALM also is called a ‘mixed va riety’. Discussing summarily the literature on ‘diglossia’, ‘language levels’, and the 7 HK refers to Mitchell’s divisio of ESA into formal vs. informal. I understand Mitchell to operate with formal vs. informal variants of features found in ESA. One should note that Mitchell excludes standard language negative markers as ‘too high flown’ for belonging to ESA. (T.F. Mitchell, ‘What is Educated Spoken Arabic?’, International Journal of the Sociology of Language 61 (1986), 7 32.)
615
REVIEWS
‘continuum’, Kammensjö concludes that ‘the only feasible classification seems to be to define varieties by the extra linguistic factors pertaining to the speech situation itself. Educated Spoken Arabic is a suitable term to cover any variety used by edu cated speakers in relatively formal situations, whether based in the colloquial or not’ (p. 65).8 The present reviewer, however, thinks that the author passes somewhat too lightly over the issue of the nature of her data. One of the defining features of vari ous kinds of ’mixed variety’ is the degree to which speech is pre planned or sponta neous, e.g. whether it is based on a written underlying text or it is produced extem poraneously. This is acknowledged on p. 69: ‘The collection of data for the present investigation confirms [that] whenever the lecturing style became more casual and spontaneous, the level would be close to the colloquial. Well planned and fully scripted lectures, on the other hand, use a higher level, closer to the MWA [Modern Written Arabic]. […] The ALM data has thus been drawn from planned and scripted lectures in History and Geography’. There is one mention of such shift (p. 88), where Arne9 is said to slip ‘into the colloquial variety of Arabic, and that happens whenever he leaves his manuscript to pick an example from real life (my ital ics)’. It is, however, rather evident from the excerpts taken from the speakers in the Appendix that some of these passages are extemporaneous, while others seem very close to written text style. In other words, features of orality may play very differ ently into the texts of various lecturers. For variation among speakers with regard to the use of the different classes of connectives, the charts are missing, as mentioned above, and the author’s comments are otherwise sparse and impressionistic. While the data investigated belongs to a didactic genre (university lectures), the presentation itself cannot be said to be a didactic masterpiece. As a reader, my ini tial enthusiasm was somehow gradually exhausted by the amount of theoretical and methodological approaches steadily added to the exposition, and in a very con densed way ending in an overload of notions and categories which I admit I found it difficult to keep track of. I also missed empirical exemplification along the way of abstract theory. Clarifying samples from the Arabic data would greatly en hance the readability. The various labels applied in the literature to Arabic varieties, be they perceived as stylistic levels or functional lects/genres, are not sorted out. To the well known MSA, ESA, and FSA the author adds her own: ALM, MWA and OWA, MAD and OAD [Modern/Old Arabic Didactic Discourse], SRD [Supraregional Dialect]. The list of abbreviations (p. 209) does not even include LA and OLA, SIL, IP, and RST and the reader sometimes has a hard time searching for their first mention and definition. With Discourse Connectives in Arabic Lecturing Monologue, Heléne Kammensjö has provided the field of Arabic linguistics with an important and very rich contri 8 The author summons the ‘diasystem’ approach launched by Weinreich (1954), in terms of ‘interpreting variation within a common integrated system’ in place of ‘diglossia’ (p. 65.). In order to improve our understanding of the mechanisms of Arabic mixed varieties, however, the ‘diasystem’ should incorporate the essential assumption of the ‘narrow diglossia’ approach namely that the two systems involved have unequal psycholinguistic status (in terms of being ‘primary’ and more entrenched). 9 That an informant sometimes is referred to by his pseudonym, sometimes with his nationality and teaching discipline, makes it unnecessary cumbersome to track down whether e.g. the speaker who uses dominant coll. variants of feature x is the same as the one with coll. variants of y. On the other hand, the informants’ typically Swedish pseudonyms render a comic effect: the Somali Sune or Jordanian Folke…
616
REVIEWS
bution, which challenges the sentence grammar and formal classes approach which still dominates much of the field. She encourages other researchers to apply the same tools from discourse analysis and functional grammar on a variety of texts and genres as well as from different periods, as a move towards constructing ‘a discourse history of Arabic’. If this important investigation is to be republished, it will benefit from consider able re editing, as well as from rewriting some of the very dense theoretical pas sages, with more explicitation perhaps also with more focus. Some lesser points: p. 79, l.17 19 under Transcription, the case of word initial hamza is discussed, and its marking rejected, not only for reasons of economy, or problems of deter mining whether it is pronounced or not, but ‘because its presence as a phoneme word initially is doubtful in FSA […] In ALM, word forms from two linguistic va rieties often occur in the same portion of discourse. Thus, to write ¨idha (LA) with a hamza, would require a hamza also in front of iza, a practice which would appear somewhat articificial’. Why? p. 92, l.7 below the table: ‘Only in one case was the relative pronoun repeated after wa’ but the sample does not show this. p. 92 3: Tables 10.1 and 10.2 are referred to as 11.1 and 11.2. p. 93, l. 4 has two commas which distort the meaning. p. 125, l. 4 refers to 12.1, which does not exist (Halliday’s categories). doi:10.1093/jss/fgp032
GUNVOR MEJDELL UNIVERSITY OF OSLO
617
SHORT NOTES F. RACHEL MAGDALENE, On the Scales of Righteousness: Neo Babylonian Trial Law and the Book of Job (Brown Judaic Studies 348). Brown University, Providence 2007. Pp. xiv + 365. Price: $59.95 hardback. ISBN: 978 1 930675 445. It has long been recognized that the book of Job has a legal background. Magdalene believes this to lie in the Neo Babylonian litigation records of the late seventh fifth centuries BCE in which period the book of Job was composed, and whose legal tra ditions Israel shared. In successive chapters Magdalene examines the relationship between suffering and law in the ancient world, sets out her methodology and her understanding of Neo Babylonian courts and their legal procedures, and in the bulk of the work examines the litigation metaphors in Job. Essential to her thesis is the unity of the book which is unquestioned. Magdalene argues that Satan brings a suit in the Divine Court against Job accus ing him of blasphemous intent. Only the blessings God has unjustifiably given him have prevented him from committing this crime. Satan’s intent is not only to mount a ‘sting operation’ against Job, but to dethrone God for disrupting the moral economy by wrongly protecting Job. God must then agree to Satan’s requests not only to vindicate Job but himself too. Satan hopes to hear Job blaspheme God to his face. Ignorant of what has gone on in the Divine Council, Job counter sues God charging him with initiating a false suit against him, violating the basic principles of justice and failing to create an orderly universe. Unaware of Satan’s charge against him, Job has to defend himself on multiple grounds but never mentions blasphe mous intent. For Magdalene the trial is still in the future. Job is seeking a pre trial settlement. The friends believe Job deserves his suffering. He should determine the nature of his offence and throw himself on God’s mercy. But they fail to offer definitive proof of Job’s guilt, never accusing him of blasphemy. It is Elihu’s intervention which is crucial for Magdalene. She sees him as the second accuser alongside Satan, neces sary for the success of the prosecution. Elihu believes that in answering the friends, Job has blasphemed as Satan said he would. God has to defend himself against both Satan and Job. In his defence, God seeks to undermine Job’s credibility. He asserts he does maintain social justice and con trols all facets of existence but does not counter Job’s charge of a false suit believing Job to be innocent. Eventually Job concedes that he cannot win against God. But instead of punishing Job for blasphemy, God condemns the three friends: their actions have bordered on the blasphemous. Job and God are reconciled. Un der the restoration agreement Job receives double damages. God always believed Job innocent. He had to prove Satan’s allegation wrong. He had never accused Job of blasphemy. Magdalene concludes that the book of Job ‘mimics to some degree’ the Neo Babylonian trial transcripts. Many scholars would not agree with Magdalene’s unargued acceptance of the cohesive unity of the book essential to her thesis. Indeed the Prologue and the Dialogue seem to have different concerns, the possibility of disinterested religion and the nature of God man relationship. While the Prologue shows that the former 619
SHORT NOTES
is possible, the Dialogue confirms that while man may know God, he cannot know everything about him. He can only be an agnostic believer. It is this that the friends fail to recognize. Indeed in the Prologue itself it seems that the issue of blasphemy is decided in Job’s refusal to carry out his wife’s suggestion to ‘curse God, and die’, that is to commit suicide. This explains why it is not an issue in the Dialogue and God can assert that in contrast to the friends, Job has spoken ‘what is right’. Further Elihu’s intervention, which adds little to the narrative, is more likely to have been prompted by a later writer to soften the apparent impropriety of Job forcing God’s hand through his self curse. While the book of Job clearly reflects the milieu of the author’s contemporary legal world, to tie it so closely to a Neo Babylonian model seems speculative. doi:10.1093/jss/fgp034
ANTHONY PHILLIPS FALMOUTH
TRYGGVE N.D. METTINGER, The Eden Narrative: A Literary and Religio Historical Study of Genesis 2 3. Eisenbrauns, Winona Lake 2007. Pp. xvii + 165. Price: $29.50. ISBN: 978 1 57506 141 2. The problem of how to read Genesis 2 3 has bedevilled scholars for generations. Current issues under debate include: Is the story a literary unity? If so, is it to be read in a diachronic or synchronic fashion? To what degree are ancient Near East ern sources necessary for fleshing out the narrative? This book is a very measured and balanced approach to these and similar ques tions having to do with methodology. Mettinger places a great deal of importance on diachronic questions. In his view the Eden Narrative is late: it post dates both Ezekiel and P. But strikingly he believes that the story itself is a literary unity; where one scholar will see an aporia, Mettinger sees a deeper level of meaning. His discus sion does not simply concern issues that have been the stock and trade of the bibli cal guild (e.g. the documentary hypothesis) but goes into great depth in contempo rary literary criticism as practiced by literary theorists. The evidence of the ancient Near East is also important in his analysis only a familiarity with the Adapa myth and the Gilgamesh epic will provide one with a window into the significance of the wisdom and eternal life motifs that are so basic to the Eden narrative. Mettinger views the story as a combination of the two pieces of tradition: the Adam myth (derived from Ezekiel 28 and Job 15:7 8) and the Deuteronomistic theology of law. From the former he gets the motif of a figure who is expelled from an idyllic setting by virtue of an act of hubris, from the latter the motif that obedi ence to the law will lead to a life of unlimited blessing. The key theme to the story is the test that God puts Adam and Eve through (though the word ‘test' is, of course, absent from the text). Should Adam and Eve prove obedient to the com mand then God will give them access to the tree of life. The end of the story, as intended by God, is a life defined by both wisdom and eternal life. Unlike some recent interpretations of the tale, God is a morally upstanding character in the end. One of the more interesting readings in the story is the treatment of the two trees. As is well known, the tree of life is only mentioned twice, once at the very beginning (2:9) and again at the end (3:22, 24). Only in these texts does one learn about two trees. In the intervening material we read solely of the tree of the knowl edge of good and evil. For many scholars this is clear evidence of two different sources. As early as 1883 with the work of Karl Budde we find the proposal that the original story had only the tree of knowledge; the tree of life was added later and 620
SHORT NOTES
constitutes a secondary addition. Mettinger finds this implausible for a couple of reasons. First of all the pairing of wisdom and immortality in ancient Near Eastern myth is well documented. There is every reason to expect that a tree of life and knowledge would be found in a single narrative. Secondly, and more importantly, the presumption of a unitary literary work gives us a deeper window into the dynamics of the narrative itself. Mettinger believes that we can understand the peculiar way these two trees function in the story if we take care to distinguish what is known to the author/reader (and ultimately, God) and what is known to the characters of the story. Adam and Eve only know about the tree of knowledge. Hence in the central part of the narrative (from 2:16 to 3:21) when we see matters through their eyes, only one tree is evident. The outer frame of the tale represents the world of the narrator. To him, and to him alone, is there an awareness of the tree of life in the very middle of the Garden. This, Mettinger argues, provides the reader with a delicious irony. Only he is aware of the stupendous prize that would have awaited Adam and Eve had they persevered in obedience. This interpretation, curiously enough, is nearly exactly the same as that of Ephrem the Syrian (see his Hymns on Paradise), though not noticed by Mettinger. Overall this is an excellent work. My only caveat is the book does not take time to compare the plot of the Eden Narrative to the rest of the primeval history. In my view, the stories that comprise Genesis 2 11 have very similar themes and purposes that need their own hearing. doi:10.1093/jss/fgp035
GARY ANDERSON UNIVERSITY OF NOTRE DAME
PAUL M. JOYCE, Ezekiel: A Commentary (Library of Hebrew Bible//Old Testament Studies 482). Continuum, London 2007. Pp. xi + 307. Price: £70.00 hardback. ISBN: 978 0 567 02685 9. Joyce laid the foundation for this commentary in his doctoral dissertation, which was published in a revised form as Divine Initiative and Human Response in Ezekiel (JSOT Supplements 51, 1989). He disclosed there that he was preparing a com mentary for the New Century Bible, a series evidently no longer in active produc tion. It is good that this stimulating work has seen the light of day in the present edition. The former series has dictated the approach: not intended to be exhaustive and based on the NRSV, which is corrected as considered necessary. The author has stayed current in his field, keeping up with recent studies and contributing his own articles, as both the bibliography and the scholarly references in the course of the commentary attest. Indeed, the comprehensive bibliography will be a gift to gradu ate students entering the field. While the commentary is alert to historical and liter ary matters, its focus is the theology of the book of Ezekiel in its own terms. In the Introduction, which takes up a quarter of the volume, by far the longest section is on theological themes. To a large extent the section recapitulates the author’s earlier work on the radical theocentricity of Ezekiel and the collective, rather than indi vidualistic, thrust of the human response. These themes are developed in their par ticular contexts in the subsequent chapter by chapter commentary. Repetition of the sectioning and descriptive headings in the outline at the close of the Introduc tion would have provided clearer guidance for readers. While individualistic aspects are not ruled out, they are seen to play a minor role. A collective interpretation is urged even in 18:21 32. Exceptionally, 20:38 is judged individualistic and called ‘an anomalous case’ (p. 154). A wise insight is that repentance, part of the human 621
SHORT NOTES
response, should be interpreted differently according to whether it occurs in a pre judgement or post judgement setting. Thus it should be given its full value in Ezekiel’s post 587 BCE messages and a lesser meaning in his pre 587 messages of doom. In fact, a similar pre and post judgement differentiation applies to the book of Jeremiah. This criterion does lay great weight on making historical decisions about particular passages. For Joyce a basic principle is that the book as far as 33:20 follows a consistent trajectory of judgement, apart from redactional additions, which are judged to be comparatively few. Even the oracles against foreign nations negatively continue a judgement theme, rather than implying a positive future for Israel. In line with this thematic structuring, 3:16 21; 11:14 21 (despite the posi tive tone); 14:1 11; 18; and 33:1 20 are assigned settings before the fall of Jerusa lem, rather than during Ezekiel’s post 587 ministry. This places the commentary firmly on one side in an academic debate. Here are some other examples of exegetical decisions. The theophany in chapter 1 is regarded positively as indicating the divine presence among the exiles; evidently the judgement theme has not yet begun. In 3:26 R.R. Wilson is followed in seeing a reference to an arbitrator and so intercessor. In 18:31 the (pre fall) challenge to the exiles to shoulder their own responsibility by getting a new heart and a new spirit is contrasted with the (post fall) free gift of 36:26 7. In chapters 40 2 S.S. Tuell is followed in finding their origin in a vision of a heavenly temple seen by Ezekiel. If so, one wonders whether the missing link with the present visionary nar rative is a statement like ‘I saw the new Jerusalem coming down out of heaven from God’ (Rev. 21:2). Two especially interesting sections in the Introduction are entitled ‘Place in the Canon’ and ‘After Ezekiel: Ezekiel in Tradition.’ The book concludes with two use ful indexes, an index of biblical and other literary references and an index of au thors. doi:10.1093/jss/fgp036
LESLIE C. ALLEN FULLER THEOLOGICAL SEMINARY
MARK J. BODA, DANIEL K. FALK and RODNEY A. WERLINE (eds), Seeking the Favor of God. Volume 2: The Development of Penitential Prayer in Second Temple Judaism (Early Judaism and Its Literature. Volume 22). Society of Biblical Literature, Atlanta 2007. Pp. xv + 281 hardback. Price: $39.95. ISBN: 978 1 58983 278 7. This is an important collection of essays on a much neglected topic. After a Preface by Mark J. Boda, Daniel K. Falk, and Rodney A. Werline (pp. xi xv) and an im pressive introduction by Eileen Schuller (‘Penitential Prayer in Second Temple Judaism: A Research Survey’, pp. 1 15), we find contributions by Rodney A. Werline (‘Prayer, Politics, and Social Vision in Daniel 9’, pp. 17 32); Pieter M. Venter (‘Daniel 9: A Penitential Prayer in Apocalyptic Garb’, pp. 33 49); Michael H. Floyd (‘Penitential Prayer in the Second Period from the Perspective of Baruch’, pp. 51 81); LeAnn Snow Flesher (‘The Use of Female Imagery and Lamentation in the Book of Judith: Penitential Prayer or Petition for Obligatory Action?’, pp. 83 104); Judith Newman (‘The Form and Settings of the Prayer of Manasseh’. pp. 105 25); Daniel K. Falk (‘Scriptural Inspiration for Penitential Prayer in the Dead Sea Scrolls’, pp. 127 57); Russell C.D. Arnold (‘Repentance and the Qumran Covenant Ceremony’, pp. 159 75); Esther G. Ghazon (‘The Words of the Luminaries and Penitential Prayer in Second Temple Times’, pp. 177 86); Bilhah Nitzan (‘Traditional and Atypical Motifs in Penitential Prayers from Qumran’, pp. 187 208), and Rodney A. Werline (‘Reflections on Penitential 622
SHORT NOTES
Prayer: Definition and Form’, pp. 209 25). The book ends with an Afterword by Eileen Schuller (pp. 227 37), a Bibliography (pp. 239 54), a List of Contributors and Indices on Ancient Sources, Modern Authors and Subjects (pp. 255 81). These are the papers that came out of a special consultation during the Society of Biblical Literature Annual Meetings of 2003 to 2005 (Consultation on Penitential Prayers). They are introduced by Eileen Schuller, who offers a history of research, from which survey it becomes clear, why penitential prayers are such an important subject. After initial efforts to define and analyse the genre ‘penitential prayer’, which started only some 20 years ago (to be mentioned here are James H. Charlesworth, David Flusser, as well as the much earlier Joseph Heinemann), from 1989 to 2003 there was first a SBL Consultation and Working Group: ‘Prayer in the Greco Roman World’. In 2003 the newly formed International Society for Deuterocanonical and Cognate Literature dedicated its first conference to prayer. Furthermore, at the same time several surveys and studies on prayer in the Dead Sea Scrolls were published. However, with all the prayer material between the canonization of the Hebrew Bible and the Greco Roman period found in the Apocrypha and Pseudepigrapha, the Dead Sea Scrolls, Philo and Josephus a systematic analysis of the material has yet to be offered. In order to be able do so, scholarship needs first to agree on a categorization of penitential prayer. Here the present volume takes as starting point the definitions offered by Claus Westermann’s Praise and Lament in the Psalms (At lanta 1981) and Rodney A. Werline (here, pp. 209 25) and then tackles some of the other problems identified by Schuller, such as the classic features of the genre penitential prayer. A large portion of the book is dedicated to Qumran and espe cially The Words of Luminaries. In conclusion, this is an important and well bal anced contribution to a much neglected subject. doi:10.1093/jss/fgp037
GERBERN S. OEGEMA MCGILL UNIVERSITY
CHRISTIANA DE GROOT and MARION ANN TAYLOR (eds), Recovering Nineteenth Century Women Interpreters of the Bible (Society of Biblical Literature Symposium Series). Society of Biblical Literature Symposium Series 38. Pp: ix + 244. Price: $34.95 paperback. ISBN: 978 1 58983 220 6. Following an excellent Introduction by the editors that succinctly explains the de velopment of the Ideology of Separate Spheres following the Industrial Revolution, the book presents a collection of essays evaluating the scriptural interpretations pro vided by thirteen nineteenth century British and North American women. These are: Sarah Trimmer (Heather. E. Weir), Mary Cornwallis (Marion Ann Taylor), Sarah [Ewing] Hall (Bernon P. Lee), Catherine McAuley (Elizabeth M. Davis), Mary Anne Schimmel Penninck (Lissa M. Wray Beal), Harriet Beecher Stowe (Marion Ann Taylor), Florence Nightingale (Christiana de Groot), Josephine Butler (Amanda W. Benckhuysen), Elizabeth Rundle Charles (Marion Ann Taylor), Christina Rossetti (Amanda W. Benckhuysen), Elizabeth Wordsworth (Rebecca G.S. Idestrom), Annie Besant (Christiana de Groot) and Etty Woosnam (Donna Kerfoot). Each essay outlines the social location and education of the women, often daughters or wives of clergy or academics, and painstakingly identifies their contri bution to women’s understanding of the role and meaning of scripture in their daily lives. The women were not particularly ‘feminist’ in their writings, only Annie Besant was unequivocal in her rejection of scripture as having anything other than negativity towards women, but adopted the current Christian view of women as 623
SHORT NOTES
helpmeets to men and obedient to them, yet sovereign in their separate sphere, not claiming ‘equality’ but, rather, independent and comparable value. The authors make clear that these women exegetes offered insights independent to those of (male) academics and that while these were intended originally to be of value to women who reared and taught their children and gave moral guidance to family, friends and servants, later, these womanly styles of reading scripture, especially those of the women who applied recent biblical scholarship and accounts of travel in the Middle East to develop their understanding of biblical stories and characters, percolated through the wider world of biblical studies. It is for this widening of the nature of the field that their work is particularly honoured here. doi:10.1093/jss/fgp038
HEATHER A. McKAY EDGE HILL UNIVERSITY
ROLAND BOER (ed.), Bakhtin and Genre Theory in Biblical Studies (Society of Bibli cal Literature Semeia Studies 63). Society of Biblical Literature, Atlanta 2007. Pp. viii + 238. Price: $25.95 paperback. ISBN: 978 1 58983 276 3. This collection grew out of a session entitled ‘Bakhtin and Genre’, which took place at the 2004 SBL meeting in San Diego. Out of eleven essays, two set the theoretical scene for further discussions (M.J. Buss and C.A. Newsom), five are dedicated to the Hebrew Bible (C. Mitchell on Chronicler, B. Green on Jonathan, J. Fentress Williams on Tamar and Judah, C. Mandolfo on Lamentations and Psalms of Lament, D.M. Valeta on Daniel), one revisits the definition of apocalyp tic (M.E. Vines), two address the New Testament texts (C.C. Fuller on the geneal ogy in Matthew's Gospel, P.N. Anderson on Johannine misunderstanding dia logues), and one deals with biblical allusions in a contemporary literary text (B. Maddison on Toni Morrison’s Beloved). The book opens with the editor’s Intro duction and closes with two responses, K. Bodner’s to the HB essays and V.K. Robbins’s to the remaining ones. In this volume Bakhtin’s ideas and categories are put in dialogue with the achievements of form criticism. The contributors take Bakhtin’s key categories, such as dialogic and monologic narrative, chronotope and heteroglossia, and explore their intersection with the problem of genres in biblical texts. Readings thus pro duced are refreshing and in most cases convincing, even though one could perhaps wish for a little more of critical engagement with Bakhtin’s categories. All in all, this book can be recommended to any reader interested in applying critical theories in the study of the Bible. doi:10.1093/jss/fgp039
ELZBIETA LAZAREWICZ-WYRZYKOWSKA UNIVERSITY OF MANCHESTER
MARK RONCACE and PATRICK GRAY (eds), Teaching the Bible Through Popular Cul ture and the Arts (Society of Biblical Literature Resources for Biblical Study 53). Society of Biblical Literature, Atlanta 2007. Pp. vi + 393. Price: $37.95 paper back. ISBN: 978 1 58983 244 2. This volume, ‘devoted to the various media that serve as intertexts for study along side the biblical writings’, consists of five sections (representing the medium of music, film, art, literature and ‘other media’, i.e. comic strips, television pro grammes etc). Each chapter contains a very substantial catalogue of works and this is followed by suggested strategies as to how students might engage with the differ ent media as interpretations of the biblical narrative. The reader may be familiar 624
SHORT NOTES
with many of the works cited but is also introduced to a range of lesser known pieces and while some of the strategies suggested for learning and teaching might prove useful, others are quite superficial and uncritical. The works discussed in each chapter are organized generally according to the order in which their subject matter appears in the canon and in some cases, such as the section on biblical subjects in art, links to websites are provided. There is no doubt that this volume provides a compendium of very useful information and will prove useful as reference material for beginner courses on the Bible and the Arts. However, it probably attempts too much in addressing five such very diverse media categories within one volume and the result is that many of the introductions to significant works of art (in each me dium) are simply too short and condensed to impart any meaningful information. From a presentational point of view, considering that its aim is to draw attention to the value of the arts, the book itself suffers from a rather dense style and contains no images that might usefully have served to break the monotony of the small, compressed font used throughout the book. doi:10.1093/jss/fgp040
MARTIN O’KANE UNIVERSITY OF WALES, LAMPETER
THOMAS AUDO, Treasure of the Syriac Language: A Dictionary of Classical Syriac (Gorgias Historical Dictionaries 9). Gorgias Press, Piscataway, NJ 2008. Pp. Vol. 1: 15* + 492; Vol. 2: 648. Price: $293.00 hardback. ISBN: 978 1 59333 572 4. This is a reprint of a Syriac Syriac dictionary published in the Eastern Syriac script (often designated ‘Nestorian’) in Mosul by the Dominican Fathers’ Press in 1897 (second volume 1901, though bearing the imprint 1897). Its author, Thomas (Toma) Audo (c. 1853 1918: dates vary in different sources), was Chaldaean Bishop of Urmia, ultimately losing his life among tens of thousands of other Assyrians during the First World War, and his contribution to Syriac lexicography was enormous. The dictionary, with its Syriac title Simta dlesana suryaya, has the special feature of being written by someone familiar with the forms of Syriac used later than the great classical authors down to Barhebraeus and contains many lexical forms which are otherwise not found in the western dictionaries. This has been re marked on recently by S. Brock (Aram 1 [2003], 167, 176 7). George Kiraz, in the introduction to this reprint, notes that in the case of the root ÎW™ Audo lists about twenty derivatives which do not appear in Payne Smith’s Thesaurus (which Audo greatly admired) or Brockelmann’s Lexicon. To cite a specific example which was important recently in my own research, the noun raÒopa, ‘mosaicist, pavement layer’, which appears in one of the earliest Syriac inscriptions (dated 194 CE), is not found in the western dictionaries (except in a special medical sense), but it is found in Audo. This dictionary will continue to be an important research tool. There have been other reprints of Audo in a single volume (Chicago 1978; Stockholm 1979; Glane/Losser [Netherlands] 1985). This reprint in two volumes contains a valuable introduction by Kiraz and an extremely helpful list of Audo’s abbreviations there is none in the original by Rabban Yuyoqim Unval. The reprint belongs to a series of ‘Gorgias Historical Dictionaries’ (GHD) being pub lished by Gorgias Press of Piscataway, New Jersey. This is GHD 9. Others in the series include reprints of Payne Smith (GHD 7) and of E. Manna’s Chaldean Arabic Dictionary (GHD 6). There are also reprints of important Arabic dictionaries: Freytag (GHD 2), Dozy (GHD 3) and Kazimirski (GHD 4). Gorgias Press is a remarkable phenomenon, the brainchild of George Kiraz, a 625
SHORT NOTES
Syriac scholar in his own right and an expert in Semitic script computing. Since it published its first book in November 2001, it has made an enormous contribution to Syriac Studies and in more recent years has also expanded into related fields of Ancient Near Eastern, Biblical, Classical and Arabic and Islamic Studies, and even occasionally further afield. Over 1100 volumes (some of which constitute multi volume works) had appeared by September 2008. It has specialized primarily in reprints, making available anew classic editions such as E.A.W. Budge’s text and translation of Barhebraeus’ Chronography. It is cur rently preparing a new edition and translation of Michael the Syrian’s Chronicle based on the sole manuscript in Aleppo (otherwise available only in J. B. Chabot’s 1899 1924 edition, Le Chronique de Michel le Syrien). The full range of publica tions can be seen at http://www.gorgiaspress.com/bookshop/. Beyond the sphere of reprints and reissues is a wide ranging Gorgias Dissertation Series and a Gorgias Handbooks Series, which includes several introductory works on the Syriac language as well as invaluable introductions to Syriac Studies in gen eral and the Syriac Bible in particular by Sebastian Brock. But it is to be empha sized that Gorgias has now extended itself well beyond Syriac Studies: there is a Ugaritic series and there are over twenty items, for example, in its Women’s Studies list. Gorgias Press currently publishes 100 books per year and is still expanding. As will be seen from the price of Audo above, the products are not cheap, but reprints of dictionaries and other classic works are always going to find a market. Through Gorgias (and also through the older enterprise of Georg Olms of Hildesheim) it is now very easy for a library or an individual scholar to create a good collection of reference tools. Except for the nostalgic, a well bound reprint of a major dictionary will be more attractive than a dog eared second hand copy which is in danger of falling apart. A good original of Payne Smith (almost too big to use except on a library table) is at the time of writing on sale for £700 in a U.K. bookshop; the Gorgias reprint is $553, a little over £300. I understand that the principle on which Gorgias manages to maintain this level of scholarly output is that of short print runs. Overheads in production are also minimized through extensive use of computer processing and, importantly, profit is not the main motive. Gorgias is in fact performing a great service to the scholarly community with no compromise on quality. And its service to the community ex tends to a series of annual book prizes for students. All this said, one of my favourites among the Gorgias books I have acquired over the last few years is George Kiraz’s own The Syriac Alphabet for Children (ISBN 1 59333 112 6). This publication reflects another of Gorgias’s concerns, service to the Syrian Orthodox and other Syriac using communities in the Middle East and in the diaspora. doi:10.1093/jss/fgp041
JOHN HEALEY UNIVERSITY OF MANCHESTER
626
BOOKS RECEIVED Juan Antonio Álvarez Pedrosa Núnez, Judeo árabe (’Ilu. Revista de Ciencias de las Religiones. Anejos: Serie de monografías). Universidad Complutense de Madrid, Madrid 2004. Pp. 106. ISBN: 84 95215 71 3. Pinchas Giller, Shalom Shar’abi and the Kabbalists of Beit El. Oxford University Press, Oxford 2008. Pp. viii + 199. Price: £38.99 hardback. ISBN: 978 0 19 532880 6. Andreas Görke, Das Kitab al Amwal des Abu ¨Ubaid al Qasim b. Sallam (Studies in Late Antiquity and Early Islam). The Darwin Press, Princeton 2004. Pp. ix + 204. Price: $39.94 hardback. ISBN: 0 87850 146 0. Robert G. Hoyland and Philip F. Kennedy, Islamic Reflections Arabic Musings: Stud ies in Honour of Alan Jones. The E.J.W. Gibb Memorial Trust, 2004. Pp. 300. Price: £45.00 hardback. ISBN: 0 906094 50 X Peter Y. Medding, Sephardic Jewry and Mizrahi Jews (Studies in Contemporary Jewry XXII). The Avraham Harman Institute of Contemporary Jewry, The Hebrew University of Jerusalem and Oxford University Press, Oxford 2008. Pp. xvi + 343. Price: £29.99 hardback. ISBN: 978 0 19 534097 6. Mohammed Nekroumi, Interrogation, polarité et argumentation: Vers une théorie structurale et énonciative de la modalité en arabe classique (Bonner Islam 1). EB Verlag, Schenefeld 2003. Pp. x + 231. ISBN: 3 936912 5. Elia di Nisibi, Arabic text edited by Samir Khalil Samir, Translation and notes by Anna Pagnini, Il Libro per Scacciare la Preoccupazione (Kitab daf¨ al hamm) tomo 1 (Patrimonio Culturale Arabo Cristiano 9). Silvio Zamorani Editore, Turin 2007. Pp. 473 (Italian and Arabic). Price: /48.00 paperback. ISBN: 978 88 7158 143 9. Il Libro per Scacciare la Preoccupazione (Kitab daf ¨ al hamm) tomo 1 (Patrimonio Culturale Arabo Cristiano 9). Silvio Zamorani Editore, Turin 2007. Pp. 473 (Italian and Arabic). Price: /48.00 paperback. ISBN: 978 88 7158 143 9. Michael Streck, Die akkadischen Verbalstämme mit ta Infix (Veröffentlichungen zur Kultur und Geschichte des Alten Orients und des Alten Testaments Band 303). Ugarit Verlag, Münster 2003. Pp. xii + 163. ISBN: 3 934628 35 4. Reinhard Weipert, ‘Ein Unglück kommt selten allein’: Vier arabische Synonymen sammlungen zum Worfeld dahiya (Beiträge zur Lexikographie des Klassischen Arabisch Nr. 16). Bayern Akademie der Wissenschaften, München 2004. Pp. 240. ISBN: 3 7696 1630 8. Peter Willey, Eagle’s Nest: Ismaili Castles in Iran and Syria (The Institute of Ismaili Studies Ismaili Heritage Series 10). I.B. Tauris, London 2005. Pp. xxvii + 321. Price: £25.00 hardback. ISBN: 1 85043 464 6. Barbara Michalak Pilulska, Modern Poetry and Prose of Oman 1970 2000. The Enigma Press, Cracow 2002. Pp. 440. ISBN: 83 86110 44 9. Dionisio Candido, I Testi del Libro di Ester: Il Caso dell’Introitus TM 1, 1 22 LXX A1 17; 1 22 Ta A1 18; 1, 1 21 (Alaecta Biblica 160). Editrice Pontificio Istituto Biblico, Rome 2005. Pp. iii + 483. Price: /30.00 paperback. ISBN: 88 7653 160 2. International Review of Biblical Studies Volume 53 (2006 2007). Brill, Leiden 2008. ISBN: 978 90 04 16551 9. 627
BOOKS RECEIVED
Anna Porter, Kasztner’s Train. Constable, London 2008. Pp. ix + 548. Price: £9.99 paperback. ISBN: 978 1 84529 717 6. Old Testament Abstracts. Vol. 30 no. 3 October 2007. Vol. 31 no 1 February 2008; Vol. 31 no. 2 June 2008. ISSN: 0364 8591. Vol. 31 David Flusser (Translated by Azzan Yadin, Foreword by David Bivin), Judaism of the Second Temple Period. Volume 1: Qumran and Apocalypticism. Wm. B. Eerdmans, Grand Rapids, The Hebrew University Magnes Press and Jerusalem Perspective, Jerusalem 2007. Pp. xiii + 356. Price: £19.99/$36.00 hardback. ISBN: 978 0 8028 2469 1. Adele Berlin, The Dynamics of Biblical Parallelism: Revised and expanded edition (The Biblical Resource Series). Wm B. Eermanns/Dove Booksellers, Grand Rapids/Dearborn 2008. Pp. xxiv + 200. Price: $21.00/£11.99 paperback. ISBN: 978 0 8028 0397 9. Bruce Chilton, The Roots of Violence in Judaism, Christianity, and Islam. Doubleday, New York 2008. Pp. 259. Price: $24.95 hardback. ISBN: 978 0 385 52027 0. S.J. Badakhchani, Paradise of Submission: A Medieval Treatise on Ismaili Thought. A New Persian Edition and English Translation of Nasir al Din Tusi’s Rawda yi talslim (Ismaili Texts and Translations Series 5). The Institute of Ismaili Studies / I.B. Taurus, London 2005. Pp. xix + 287 + 220. Price: £35.00 hardback. ISBN: 1 86064 436 8. Teodoro Abu Qurra. Introduzione, traduzione, note ed indici a cura di Paolo Pizza, testo arabo a cura di Samie Khalil Samir, La Libertà (Patrimonio Culturale Arabo Cristiano 6). Silvio Zamorani Editore, Turin 2001. Pp. 260. Price: /30.00 paperback. ISBN: 88 7158 102 4. Th. Emil Homerin, The Wine of Love and Life: Ibn al Farid’s al Khamriyah and al Qaysari’s Quest for Meaning (Chicago Studies on the Middle East). The Middle East Documentation Center, Chicago 2005. Pp. xxvi + 126. Price: $59.95 hardback. ISBN: 0 9708199 2 7. Hamid Haji, Founding the Fatimid State: The Rise of an Early Islamic Empire. An Annotated English translation of al Qa∂i al Nu¨mans’s IftitaÌ al Da¨wa (The Institute of Ismaili Studies. Ismaili Texts and Translations Series 6). I.B. Tauris, London 2006. Pp. xvii + 256. Price: £29.95 hardback. ISBN: 978 1 85043 885 4. Arieh Bruce Saposnik, Becoming Hebrew: The Creation of a Jewish National Culture in Ottoman Palestine. Oxford University Press, Oxford 2008. Pp. xii + 314. Price £34.99 hardback. ISBN: 978 0 19 533121 9. Kathy Ehrensperger and R. Ward Holder (eds), Reformation Readings of Romans (Romans Through History and Cultures Series). Continuum, London 2008. Pp. 222. Price: £25.00 paperback. ISBN: 978 0 567 02714 6. Daniel F. Moor, Jesus, An Emerging Jewish Mosaic: Jewish Perspectives, Post Holocaust (T. & T. Clark Jewish and Christian Texts Series 2). T. & T. Clark, London 2008. Pp. x + 326. Price: £75.00 hardback. ISBN: 978 0 567 02738 2. Jonathan Karp, The Politics of Jewish Commerce: Economic Thought and Emancipa tion in Europe, 1638 1848. Cambridge University Press, Cambridge 2008. Pp. x + 379. Price: £50.00 hardback. ISBN: 978 0 521 87393 2. Naftali Brawer, A Brief Guide to Judaism: Theology, History and Practice. Constable and Robinson, London 2008. Pp. xv + 270. Price: £8.99 paperback. ISBN: 978 1 84529 601 8. Roland Boer, Last Stop before Antartica: The Bible and Postcolonialism in Australia (Society of Biblical Literature Semeia Series 24). Society of Biblical Literature, Atlanta 2008. Pp. xiv + 201. Price: $24.95 paperback. ISBN: 978 15893348 7. 628
BOOKS RECEIVED
Farhad Daftary, Ismaili Literature: A Bibliography of Sources and Studies. I.B. Tauris in association with The Institute of Ismaili Studies, London 2005. Pp. xviii + 469. Price £35.00 hardback. ISBN: 978 185043 4399. Farhad Daftary, Ismailis in Medieval Muslim Societies (The Institute of Ismaili Studies Ismaili Heritage Series 12). I.B. Tauris in association with The Institute of Ismaili Studies, London 2005. Pp. xii + 259. Price: £29.50 hardback. ISBN: 1 84511 091 9. Massimo Campanini, An Introduction to Islamic Philosophy. Edinburgh University Press, Edinburgh 2008. Pp. xiii + 186. Price: £18.99 paperback. ISBN: 978 0 7486 2608 3. Leo Perdue, The Sword and the Stylus: An Introduction to Wisdom in the Age of Empires. Wm. B. Eerdmans, Grand Rapids 2008. Pp. x + 502. Price: $38.00 paperback. ISBN: 978 0 8028 6245 7. Patrice Cressier, Maribel Fierro and Luis Molina (eds), Los Almohades: Problemas y Perspectivas (2 vols) (Estudios Árabes e Islámicos Mongrafias). Consejo Superior de Investigaciones Científicas, Madrid 2006. Pp. li + 1230 (Vols 1 and 2). ISBN: 84 00 08395 4 (vol. 1); 84 00 08395 6 (vol. 2). Jan G. Van der Watt (ed.), Review of Biblical Literature 2008. Society of Biblical Literature, Atlanta 2008. Pp. vi + 596. ISSN: 1099 0046. L’opera di Giovanni Garbini: Bibliografia degli scritti 1956 2006. Paideia Editrice, Brescia 2007. Pp. 124. David Shepherd (ed.), Images of the Word: Hollywood’s Bible and Beyond (Society of Biblical Studies Semeia Series 54). Society of Biblical Literature, Atlanta 2008. Pp. vi + 227. Price: $25.95 paperback. ISBN: 978 1 58983 275 6.
629
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.
CHRISTO VAN DER MERWE
Another Look at the Biblical Hebrew Focus Particle jGÌ
313
JOSÉ MARTÍNEZ DELGADO
Lexicographical Arrangement of Masoretic Material
333
MOSHE FLORENTIN
The Alternation u / o, Diphthongs, PataÌ furtivum and the 3m.s. Pronominal suffix in Samaritan Hebrew and Aramaic versus Tiberian Hebrew
365
JASON BEMBRY
Plural Based Duals
383
GEOFFREY HERMAN
Note on the Recently Published Discourse on Priesthood (BL Add. 18295, ff. 137b–140b)
389
MARIA BULAKH
Nota Genitivi za- in Epigraphic Geez
393
URI RUBIN
On the Arabian Origins of the Qur’an: The Case of al-furqan
421
AHARON GEVA KLEINBERGER and EFRAIM LEV
Language Passivity in the Medical Arabic and Judaeo-Arabic Prescriptions of the Cairo Genizah
435
ALMOG KASHER
The Term ism in Medieval Arabic Grammatical Tradition: A Hyponym of Itself
459
ABDULRAHMAN S. AL SALIMI
Themes of the Iba∂i/Omani Siyar
475
AHMAD AL JALLAD
The Polygenesis of the Neo-Arabic Dialects
515
JAMES DICKINS
Relative Clauses in Sudanese Arabic
537
REVIEWS
575
SHORT NOTICES
619