Arabic DialectologyStudies inSemitic Languages and LinguisticsEdited byT. Muraoka and C.H.M. VersteeghVOLUME 53Arabic DialectologyIn honour ofClive Holes on the Occasion ofhis Sixtieth BirthdayEdited byEnam Al-Wer and Rudolfde JongLEIDEN BOSTON2009This book is printed on acid-free paper.Library ofCongress Cataloging-in-Publication DataArabic dialectology : in honour ofClive Holes on the occasion ofhis sixtieth birthday / edited by Enam Al-Wer and Rudolfde Jong.p. cm. (Studies in Semitic languages and linguistics ; v. 53)Includes a bibliography ofClive Holes published works.Includes bibliographical references and index.ISBN 978-90-04-17212-8 (hardback : alk. paper)1.Arabic languageDialects. 2.SociolinguisticsArab countries. I. Al-Wer, Enam. II. Jong, Rudolfde. III. Holes, Clive, 1948- IV. Title. V. Series.PJ6709.A76 2009492.77--dc22 2009014371ISSN 0081-8461ISBN978 90 04 172128Copyright 2009 by Koninklijke Brill NV, Leiden, The Netherlands.Koninklijke Brill NV incorporates the imprints Brill, Hotei Publishing,IDC Publishers, Martinus NijhoffPublishers and VSP.All rights reserved. No part ofthis publication may be reproduced, translated, stored ina retrieval system, or transmitted in any form or by any means, electronic, mechanical,photocopying, recording or otherwise, without prior written permission from the publisher.Authorization to photocopy items for internal or personal use is granted by Koninklijke Brill NVprovided that the appropriate fees are paid directly to The Copyright Clearance Center,222 Rosewood Drive, Suite 910, Danvers, MA 01923, USA.Fees are subject to change.printed in the netherlandsco1i1s vCONTENTSAcknowledgements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viiIntroduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .ixBibliographv of Clive Holes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiiiPoem: On Your Sixtieth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .xixby Said Abu AtheraTOPICS IN HISTORICAL LINGUISTICSIndeterminacv and the Comparative Method: Arabic as a Model for Understanding the Historv of Aramaic . . . . . . . 3Ionathan OwensFrom qltu to glt: Diachronic Notes on Linguistic Adaptationin Muslim Baghdad Arabic. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17Heikki PalvaTe g/-question in Egvptian Arabic Revisited . . . . . . . . . . . . . . 41Manfred Voidich and Liesbeth ZackDESCRIPTIVE DIALECTOLOGYWords and Tings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63Peter BehnstedtTe Arabic Dialect of a aww Communitv of Northern Oman . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77Domenyk EadesTe Dialect of the Euphrates Bedouin, a Fringe Mesopotamian Dialect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99Bruce InghamOuelques Donnes Sociolinguistiques sur lArabe Parl a Damas a la Fin des Annes Mille Neuf Cent Soixante-dix109Ierome LentinCONTACT PHENOMENAContact, Isolation, and Complexitv in Arabic . . . . . . . . . . . . . . . 173Peter Trudgillco1i1s viLoan Verbs in Arabic and the DO-construction . . . . . . . . . . . . . 187Kees VersteeghSOCIAL DIALECTOLOGYWhen Najd Meets Hijaz: Dialect Contact in Ieddah . . . . . . . . . 203Aziza Al-EssaBig Bright Lights Versus Green and Pleasant Land:: Te Unhelpful Dichotomv of Urban Versus Rural in Dialectologv. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223David BritainTe Variable (h) in Damascus: Analvsis of a Stable Variable249Hanadi IsmailCODE MIXINGTe Varietv of Housewives and Cockroaches: Examining Code-choice in Advertisements in Egvpt . . . . . . . . . . . . . . . . 273Reem BassiouneyIndex. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285co1i1s viiACKNOWLEDGEMENTSTe contributors to this volume have supported this project at short noticefromitsinceptionandthroughout.Wethankthemwhole-heartedlv for their academic spirit and generositv.Notmanvpeoplegetapoemfortheirbirthdav.HatsoftoSaid for composing a splendid poem and for making this tribute to Clive trulv special.From the Universitv of Essex, we thank Mike Iones for his helpful suggestions and support.We are also verv grateful to the series editor, Kees Versteegh, who hasbeen,asalwavs,generouswithhistimeandinsightfulinhis comments.FromBrillwethankLiesbethKanisforheremciencvindealing with manv complications, and for her help and patience.Itiscustomarvtoleaveoneextraspecialthankvoutotheend line. In this case it goes to Deidre. According to Saids poem Deidre hasknockedthirtvvearsofClivesage.Weagreewithhim,of course,andwouldliketoaddourheartfeltgratitudeforherhelp witheditingandsupplvingmaterial.Wethankheralsoforrespon-dingpromptlvanddiscretelvtoouremailsandphonecalls,while Clive remained blissfullv ignorant!co1i1s viiii1vouUc1io ixINTRODUCTIONArabic Dialectology is a collection of articles written bv leading schol-ars and distinguished voung researchers. In print, thev come together to pav homage to Clive Holes on the occasion of his sixtieth birthdav. Asbeftsthehonoureescontributionsandinterests,thecollection covers a wide spectrum of the feld of Arabic linguistics and linguis-tics in general.Clives earliest research on Arabic, in the late 1970s, was in socio-linguistics, specifcallv in Labovian variationist sociolinguistics, as it is nowadavs known. At the time, sociolinguistics was certainlv thri-ving, but sociolinguists were still thin on the ground. In Europe, the Labovianapproachtookaleapforwardatthehandsofoneofthe contributors in this volume, Peter Trudgill, whose studv of his belo-ved hometown has made the English citv of Norwich itself almost a part of sociolinguistics. Not manv places earn this privilege, but just over a decade afer the Norwich Studv, and thousands of miles awav fromEngland,atinvislandintheArabianGulfwasgivenequal prominenceinsociolinguisticresearch,thistimeatthehandsof CliveHoles,thenanalienresearcherintheState.Clivesresearch onlanguagevariationandchangeinBahrainisthefrstworkon Arabicdialectswhichisvariationisttothecore,executedwiththe passion of an ethnographer and the knowledge of an insider. His take on variation in Arabic has lent a fresh and realistic perspective to the subject. As such, his work in the feld has not onlv led but also shaped subsequent studies on Arabic dialects.In his research, Clive is a true nomad, but unlike the nomad ances-tors of the people in whom he is interested, he has ventured into new landwhiletheoldgroundwasstillabundant.Tisjournevhasled him to manv other areas which he has made fertile: language manuals, descriptive grammars, historical linguistics, culture and societv and, mostrecentlv,popularliteratureandpoetrv.Hisworkthroughout hasremainedgroundedinfeldlinguistics,exploratorv,locallv focused and sociallv sensitive.His research on popular literature and Bedouin poetrv gives centre stagetodeservingpoetsworkingundercover.Inadditiontodocu-menting a wealth of vocabularv and structures, folklore, local values i1vouUc1io xand local traditions, it ingeniouslv uncovers a 4,000-vear old literarv linkgoingbacktoSumerian.VanquishedbvthemodernArab societv,muchofthisculturalheritagehasbeenunfairlvconsigned to oblivion, while the rest is being continuallv eroded.Outside academia, Clives work on contemporarv Bedouin poetrv, satirisingcurrentinternationalpoliticshasbeencited inthemedia. An Oxford dons research into traditional poetrv seems an unlikelv placeforGeorgeGallowav1tocropup.Butheis,professorClive Holeshasdiscovered,thesubjectofastirringversetributefroma Bedouintribesman,writesMaevKennedvinTeGuardian (12Februarv2008),withapictureofajubilantGallowavandthe captionInscribedintheannalsofhonourhisname!takenfrom ClivestranslationofalHajavaspoemAnOdetoGeorgeGalloway. His analvses of Bedouin poetrv uncover a wealth of adversarial sati-ristswhosepoemsareabitingsatireonsocietvandpoliticsinthe ArabMiddleEast;suchpoemscouldnothavebeenpublishedin Arabcountries.TotheArabreader,thepoemsareenjovableand funnv, ofen painfully so in their original text, but in manv parts thev are incomprehensible without Clives decoding, as one of the editors, a native speaker of Arabic with Bedouin roots, can testifv. In one of his articles, Clive astutelv describes popular poetrv as an exercise of free speech in the modern Arab World.Keepingtothetraditionalmetreandrhvmes,Clivereproduces Bedouinpoetrvwithitsdrvwittransposingitsimagervtomakeit accessibletotheEnglish-speakingreader.Asanexample,wecite some verses from Y Kundalzza Rys! (Hev Condoleezza Rice)2. In thispoem,GeorgeW.Bushisthespeaker;Cliverendersthepoem in the Texan dialect:Mah moods good, vall, on mah lips a smile is crackinAhm reallv in high spirits cos the bad guvs we sent packin!Sav, Powell, come here, n fetch some liquor in a fagon!And Rumsfeld get a fautist (Ah aint stavin on the wagon!),An hire a rebec-plaverfrom the Gulf where thev like singin,Where the desert A-rabs nod an to mah evrv word thevre clingin1George Gallowav is the British MP for Bethnal Green and Bow, known particu-larlv for his opposition of the Iraq war. Te poem bv al-Hajava was written following Gallowavs win in the 2005 elections. 2Published (with S. Abu Athera) George Bush, Bedouin Poet, 2007. Te poem was composed in late 2003 bv the poet M. F. al-Hajava when it seemed the Ameri-cans had won the Iraq war.i1vouUc1io xiBring some gin with vall, and some whiskv and some beer,With good ole Condoleezza:bring em all over here!As epitomised in Clives publications, true scientifc research refects a mixture of fascination for the subject, an eve for detail and a thirst for knowledge. If Arabic culture is a well in a drv land, then the studv oftheArabiclanguageistheropeandpailtoquenchonesthirst. Not onlv the thirst to research the Arabic language, but also to dis-cover the mechanics of language in general and, perhaps even more so,tounderstandthecultureofapeoplebvwhichonewasdrawn to that well in the frst place.Cliveknowsthatanvonestrivingtounderstandwhatmakesthe Arabmindtickshouldfrstofallbeabletocommunicatewith Arabs.Toachievethis,readingbooksaboutIslamandnewspaper articlesoncurrenteventsintheArabworldissimplvnotenough. Oneshouldfrstandforemostbeabletotrulvcommunicatewith people: listen, and then answer in Arabic, of course! Onlv this can beasoundbasisforunderstandingaculture,whichisstillviewed bvmanvashighlvexotic.Indeed,Clivesinterestsgodeeperthan the studv of the Arabic language alone. Tis language is the kev to a world of culture, of which its popular manifestations have caught his heart as well.Discussing linguistics with Clive is alwavs much more than a drv exchange of ideas on language. Te same livelv interaction that cha-racterisesthelanguagethatisdiscussed,alsotvpifesthenatureof theexchangeofideasitself.Andweremembermanvatimewhen these exchanges were far from drv!TecontributionsinthisbookfromoutsidethefeldofArabic linguisticsrefectthegrowingrealisationoftheimportancetolin-guistics of engaging with the insights from Arabic data in linguistics. Tere can be no doubt that this promising endeavour is in large part a result of Clives publications over the past three decades.Apart from his own academic achievements, Clive has been uns-tintinginhissupportforothers,throughhisencouragementof voung researchers to his outspoken advocacv of fairness in access to education.Aseditors,weareindeedpleasedtohavegatheredtogetherthe mostdistinguishedofscholarsinthiscollectiontobringaftting tribute to Clive Holes as a highlv acclaimed linguist, a distinguished i1vouUc1io xiiprofessorandaFellowoftheBritishAcademv.But,mostofallwe were driven bv the wish to honour a precious friend.Tebookbeginswithapoem,composedespeciallvfortheocca-sionbvClivesfriendandassociate,thepoetSaidAbuAthera.We apologisefornotprovidingatranslationofthispoem,butwefelt that nobodv but Clive could do it justice!Enam and RudolfColchesterAmsterdam niniiocv.vuvoiciiviuoiis xiiiBIBLIOGRAPHY OF CLIVE HOLES1980PhonologicalvariationinBahrainiArabic:the[j]and[v]allophonesof /j/, Zeitschri fur Arabische Linguistik 4. 72-89.1983aBahrainidialects:sectariandialectsandthesedentarv/nomadicsplit, Zeitschri fur Arabische Linguistik 10. 7-37.1983bPatterns of communal language variation in Bahrain, Language in Society 12/4. 433-457. 1983cReviewofAl-Tajir,M.A.LanguageandLinguisticOriginsinBahrain, Bulletin of the School of Oriental and African Studies 46/3. 552-553.1984aColloquial Arabic of the Gulf and Saudi Arabia. London: RKP (reprinted 1986, 1992, 1994, 2000, 2004. 319 pp + cassette).1984bBahrainidialects:sectariandiferencesexemplifedthroughtexts, Zeitschri fur Arabische Linguistik 13. 27-67.1984c Textual approximation in the teaching of academic writing to Arab stu-dents: a contrastive approach.Swales, I. and Mustapha, H. (eds) English for Special Purposes in the Arab Vorld, Aston Universitv. 228-242.1985a Review of Talmoudi, F. Diglossia in ^orth Africa, Bulletin of the School of Oriental and African Studies 48/3. 549-550.1985bReview of Nakano, A. Folktales of Lower Egypt, Iournal of Semitic Studies 30/2. 332-334.1986aTe social motivation for phonological convergence in three Arabic dia-lects, International Iournal of the Sociology of Language, 61. 33-51.1986bVariationinthemorphophonologvofArabicdialects,Transactionsof the Philological Society 84. 167-190.1986c Communicative function and pronominal variation in Bahraini Arabic, Anthropological Linguistics 28/1. 10-30. 1986d Principles of Arabic language course design. Proceedings of theBRISMES/MESA International Conference on Middle Eastern Studies. 9-18.1986eReviewofVersteegh,C.H.PidginizationandCreolization.TeCaseof Arabic. Bibliotheca Orientalis 43. 218-222.1987aLanguageVariationandChangeinaModernisingArabState.Librarvof Arabic Linguistics Series, Monograph No 7. London & New York: Kegan Paul International. 214 pp.1987bReview of Sowavan, S.A. ^abati Poetry, BRISMES Bulletin 13/2. 219-220.1987c Review of Brockett, A.A. Te Spoken Arabic of Khabura on the Batina of Oman, Bulletin of the School of Oriental and African Studies 50/3 (1987). 558-560.1988a(withBaker,M.)TeuseofcomputerisedtextconcordancinginArabic language teaching and translation.Proceedings of the 2nd Symposium on the Teaching of Arabic with the Computer. Universitv of Leeds. 5-29.1988b TetvpologvofOmaniArabicdialects.ProceedingsoftheBRISMES International Conference on Middle Eastern Studies. 12-21.1988cReviewofKav,E.AnArabicDictionaryofCivilEngineering.BRISMES Bulletin 14/2. 264-265.1988dReview of Kave, A. ^igerian Arabic-English Dictionary. Iournal of Semitic Studies 33/2. 337-338. niniiocv.vuvoiciiviuoiis xiv1988eReviewofKarmi,H.Al-MughniAl-Akbar(English-ArabicDictionarv), BRISMES Bulletin 15. 175-177.1989aTowards a dialect geographv of Oman. Bulletin of the School of Oriental and African Studies 52/3. 446-462. 1989bReviewofHaddad,E.W.ADictionaryofDataProcessingandComputer Terms (English-French-Arabic). BRISMES Bulletin 16/1. 103-104.1990aGulfArabic.CroomHelmDescriptiveGrammarsSeries.London: RoutledgeCurzon. 302 pp.1990bA multi-media, topic-based approach to universitv-level Arabic language teaching.DiglossicTension.TeachingArabicforCommunication.Folia Scholastica (Universitv of Leeds). 36-41.1990cReviewofVoigt,R.M.DieinfrmenVerbaltypendesArabischenunddas Biradikalismus-Problem. Iournal of Royal Asiatic Society. 141-144.1990dReview of Kave, A. English-^igerian Arabic Dictionary. Iournal of Semitic Studies 35/2 (1990). 374-375.1990eReviewofBavvumi,H.,Kalfat,Kh.,al-Shafi,A.MujamTasrifal-Afal al-Arabiyya, BRISMES Bulletin 17/1. 113-114.1991aKashkasha and the fronting and africation of the velar stops revisited: a contribution to the historical phonologv of the peninsular Arabic dialects. Kave, A. (ed.) Semitic Studies in Honor of Volf Leslau, Vol. 1. Wiesbaden: Harrassowitz. 652-678.1991bReview of Feghali, H.I. Moroccan Arabic Reader. Bulletin of the School of Oriental and African Studies 54/2. 370.1991cReviewofFeghali,H.I.ArabicAdeniReader.BulletinoftheSchoolof Oriental and African Studies 54/3. 634.1991dReview of Oafsheh, H.A. Yemeni Arabic I, Al-Arabiyya 24. 148-149.1992a(withHarris,R.&Autv,N.)BreakthroughArabic.London:Macmillan. 243 pp + 3 C90 cassettes; CD version, 2005, McGraw-Hill.1992bReviewofIrvine,A.K.,Sergeant,R.B.,andSmith,G.R.AMiscellanyof Middle Eastern Articles, In Memoriam Tomas Muir Iohnstone 1924-1983. Bulletin of the School of Oriental and African Studies 55/1. 132-134.1992cReviewofIohnstone,B.RepetitioninArabicDiscourse.Bulletinofthe School of Oriental and African Studies 55/3. 556-557.1992dReviewofSchall,A.ElementaArabica.IournalofSemiticStudies37/2. 348-349.1992eReviewofSowavan,S.TeArabianOralHistorical^arrative.anEthno-graphic and Linguistic Analysis. British Iournal of Middle Eastern Studies 19/2. 221-222.1992fReview of Feghali, H.I. Arabic Adeni Textbook and Arabic Adeni Reader. Al-Arabiyya 25. 139-140.1993a(withEid,M.,eds.)PerspectivesonArabicLinguisticsVol.5=Current issues in Linguistic Teorv No 101. Amsterdam: Benjamins. 347 pp.1993bTeusesofvariation:astudvofthepoliticalspeechesofGamalAbd al-Nasir. Eid, M. and Holes, C.D. (eds) Perspectives on Arabic Linguistics Vol. 5. Amsterdam:Benjamins. 13-45. 1993cReview of Eid, M. and McCarthv, I. Perspectives on Arabic Linguistics, Vols I o II. Bulletin of the School of Oriental and African Studies 56/1. 132-134.1993dReview of Rouchdv, A. Te Arabic Language in America. Al-Arabiyya 26 (1993). 150-154.1994aMainarticleArabic,andarticlesonthelanguagesituationinBahrain, Lebanon, Libva, Oman, Saudi Arabia, Svria, United Arab Emirates. Asher, R.E. et al (eds) Te Encyclopaedia of Language and Linguistics (10 niniiocv.vuvoiciiviuoiis xvVols.),Pergamon/AberdeenUniversitvPress:Vol.1.191-194,300;Vol. 4. 2068-2069, 2207; Vol. 5. 2872; Vol. 7. 3662; Vol. 8. 4500; Vol. 9. 4836. 1994bDesigningEnglish-Arabicdictionaries.DeBeaugrandeR.,ShunnaqA., Heliel,M.(eds)Language,DiscourseandTranslationintheVestand Middle East. Amsterdam:Benjamins. 161-180. 1994cReviewofDevenvi,K.&Ivanvi,T.(eds)ProceedingsoftheColloquium on Arabic Grammar, Budapest 1-7 September 1991. Iournal of Royal Asiatic Society. 268-9; Iournal of Semitic Society 39/1. 129-130.1995aModern Arabic. Structures, Functions and Varieties. Longman Linguistics Librarv. London & New York: Longman. xiii + 343pp.1995b Te structure and function of parallelism and repetition in spoken Arabic: a sociolinguistic studv, Iournal of Semitic Studies, 40/1. 57-81.1995cCommunitv,dialectandurbanisationintheArabic-speakingMiddle East. Bulletin of the School of Oriental and African Studies 58/2. 270-287.1995dTe Rat and the Ships Captain: a dialogue poem (muHawara) from the Gulf,withsomecommentsonthesocialandliterarv-historicalback-ground of the genre, Studia Orientalia 75. 101-120.1995eTepassiveinOmaniArabic.Proceedingsofthe2ndConferenceofthe InternationalAssociationofArabicDialectologyAIDA).Cam bridge. 69-74.1995fReviewofAbu-Haidar,F.ChristianArabicofBaghdad.Bulletinofthe School of Oriental and African Studies 58/1. 140-141.1995gReview of Cadora, F.I. Bedouin, Village and Urban Arabic. an Ethnolinguistic Study. Bulletin of the School of Oriental and African Studies 58/1. 142-144.1995hReview of Watson, I. A Reference Grammar of Sanani Arabic. Bulletin of the School of Oriental and African Studies 58/3. 555-556.1995iReview of Muller, K. Und der Kalif lachte, bis er auf den Rucken fiel. Ein BeitragzurPhraseologieundStilkundedesklassischenArabisch2Vols). Iournal of Royal Asiatic Studies. 407-408.1995jReviewofSelim,G.D.Arabic-EnglishandEnglish-ArabicDictionariesin theLibraryofCongress.BritishIournalofMiddleEasternStudies22. 224-225.1996aTeDisputeofCofeeandTea:adebate-poemfromtheGulf.Smart,I. (ed.) Tradition and Modernity in Arabic Language and Literature. London: Curzon Press. 302-315.1996bTe Arabic dialects of south-eastern Arabia in a socio-historical perspec-tive, Zeitschri fur Arabische Linguistik 31. 34-56. 1996cReview of Kurpershoek, M. Oral Poetry and narratives from Central Arabia Vol. I. the Poetry of al-Dindan, a Bedouin Bard in southern ^ajd. Iournal of American Oriental Society 116.1. 155-156.1996dReview of Ingham, B. ^ajdi Arabic. Bulletin of the School of Oriental and African Studies 59/3. 561-563.1996eReviewofSuleiman,Y.(ed.)ArabicSociolinguistics.BritishIournalof Middle Eastern Studies 23/1. 102-104.1997aReviewofDevenvi,K.,Ivanvi,T.,andShivtiel,A.Proceedingsofthe ColloquiumonArabicLexicographyandLexicology,PartOne.Iournalof Semitic Studies 42/1. 196-198.1997bReviewofMuller,K.DerBeduineunddieRegenwolke.EinBeitragzur ErforschungderaltarabischenAnekdote.IournaloftheRoyalAsiatic Society, 3rd series 7/1. 118-119.1997cReview of Eid, M. (ed.) Perspectives on Arabic Linguistics Vol VII. Bulletin of the School of Oriental and African Studies 60/3. 560-561.niniiocv.vuvoiciiviuoiis xvi1997dReview of Bergtsson, P. Two Arabic Versions of the Book of Ruth. Iournal of Royal Asiatic Society 3rd series 7/3. 434-436.1998aTe Debate of Pearl-Diving and Oil-Wells: a poetic commentarv on socio-economicchangeintheGulfofthe1930s,ArabicandMiddleEastern Literatures Vol 1 No 1. 87-112. 1998b Retention and loss of the passive verb in the Arabic dialects of northern Oman and eastern Arabia. Iournal of Semitic Studies 43/2. 347-362. 1998cReviewofWatson,I.Sbahtu!ACourseinSananiArabic.Bulletinofthe School of Oriental and African Studies 61/1. 208-209.1998dReviewofKurpershoek,P.M.OralPoetryand^arrativesfromCentral Arabia Vol II. the Story of a Desert Knight. Te Legend of Shlewih al-Atawi andotherUtaybahHeroes,IournalofAmericanOrientalSociety118.1. 106-108.1998eReview of Ingham, B. Arabian Diversions, Iournal of Royal Asiatic Society 3rd series 8/2. 261.1999aSocio-economic change and language change in the eastern Arab World. Etudes Asiatiques 53/1. 45-74. 1999bMain article on Arabic Literature in English Translation and articles on theKoran,theMuallaqt,andtheMuqaddima.France,P.(ed.) Te OxfordGuidetoLiteratureinEnglishTranslation.Oxford:OxfordUni-versitv Press. 139-149. 1999cReviewofVersteegh,K.TeArabicLanguage,IournalofIslamicStudies 10/1. 101-104.1999dReview of Shrvock, A. ^ationalism and the Genealogical Imagination. Oral History and Textual Authority in Tribal Iordan, Edebiyat 10. 167-172.2000aUman: modern Arabic dialects, Encyclopaedia of Islam, Vol X. 817-818. 2000bAl-Wahiba, Encyclopaedia of Islam, Vol XI. 47. 2000cRefexes of CLA qad and qat in the Arabic dialects of eastern and central Arabia. In Aspects of the Dialects of Arabic Today. Proceedings of the 4th Conference of the International Association of Arabic Dialectology AIDA), Marrakesh. 88-97.2000dReview of Edzard, L. Language as a Medium of Legal ^orms. Implications of the Use of Arabic as a Language in the United ^ations System, Die Velt des Islams 40. 104-106.2000eReviewofKurpershoek,P.M.OralPoetryand^arrativesfromCentral Arabia3.BedouinPoetsoftheDawasirTribe,IournalofRoyalAsiatic Society, 3rd Series, 10/2. 222-224.2001aDialect,CultureandSocietyinEasternArabia,VolumeI.Glossary. Handbuch der Orientalistik Series. Leiden, Boston: Brill. lxiii + 573pp.2001bDialogueoftheDeaf .InstituteofTranslationandInterpretingBulletin, December 2001. 28-30.2002aNon-Arabic Semitic elements in the Arabic dialects of eastern Arabia, in Arnold,W.andBobzin,H.(eds)SprichdochmitdeinenKnechtenara-misch,wirverstehenes!0BeitrgezurSemitistik.FestschriftfurOtto Iastrow zum 0. Geburtstag. Wiesbaden: Harrassowitz. 269-280.2002bReview of Suleiman, Y. (ed.) Language and Society in the Middle East and ^orth Africa, Iournal of Islamic Studies 13/1. 100-106.2003aReview of Suleiman, Y. Te Arabic language and ^ational Identity, Times Literary Supplement, 2 Mav. 29.2003bReviewofdeIong,R.AGrammaroftheBedouinDialectsof^orthern Sinai Littoral, Iournal of Semitic Studies48. 413-416.niniiocv.vuvoiciiviuoiis xvii2003cReviewofBorg,A.(ed.)TeLanguageofColorintheMediterranean, Ierusalem Studies in Arabic and Islam 28. 425-428.2004aModernArabic.Structures,FunctionsandVarieties.RevisedEdition.George townClassicsinArabicLanguageandLinguisticsSeries. Washington D.C.: Georgetown Universitv Press. xix + 420pp. (revised and updated edition of the 1995 book of the same title).2004bZawdj, Encyclopaedia of Islam, Supplement. 842-843.2004cOuadriliteralverbsintheArabicDialectsofEasternArabia.Haak,M., DeIong,R.andVersteegh,K.(eds)ApproachestoArabicDialects.A Collection of Articles presented to Manfred Voidich on the Occasion of his Sixtieth Birthday. Leiden, Boston: Brill. 97-116.2004dArabianGulfhiyyabiyya,IewishBabvlonianfarfsa,ChristianSicilian sepolcri:popularcustomswithacommonorigin:IournalofSemitic Studies 49/2. 275-287.2004eTanwnintheArabicDialectsofEasternArabia.Woidich,M.and Aguad,I.(eds)FestschriinHonourofthe0thBirthdayofPeter BehnstedtEstudios de Dialectologia ^orteafricana y Andalusi, 8. Zaragoza [2006]. 89-97.2004fReviewofKurpershoek,P.M.OralPoetryand^arrativesfromCentral Arabia 4. A Saudi Tribal History, Iournal of the Royal Asiatic Society, 3rd Series, 14/3. 273-276.2005aDialect,CultureandSocietyinEasternArabia,VolumeII.Ethnographic Texts.HandbuchderOrientalistikSeries.Leiden,Boston:Brill.xli+ 347pp.2005bTe Birth of Orientalism: Sir William Iones. France, P. and Gillespie, S. (eds)TeOxfordHistoryofLiteraryTranslationinEnglish,VolIII 190-170. Oxford: Oxford Universitv Press 1. 443-455. 2005cTemeXoftheverbintheArabicdialectsofeasternArabia.Khan,G. (ed.) Semitic Studies in Honour of Edward Ullendorf. Leiden, Boston: Brill. 115-125.2005dDialectandnationalidentitv:theculturalpoliticsofself-representation in Bahraini musalsalt. Paul Dresch and Iames Piscatori (eds) Monarchies and^ations.GlobalisationandIdentityintheArabStatesoftheGulf. Reading:I.B. Tauris. 52-72.2006aBahrainiArabic.Versteegh,K.,Eid,M.,Elgibali,A.,Woidich,M.and Zaborski, A. (eds) Encyclopedia of Arabic Language and Linguistics Vol 1: A-Ed. Leiden, Boston: Brill. 241-255.2006bGulf States. Versteegh, K., Eid, M., Elgibali, A., Woidich, M. and Zaborski, A.(eds)EncyclopediaofArabicLanguageandLinguisticsVol1:A-Ed. Leiden, Boston: Brill. 210-216.2006cTeArabicdialectsofArabia.ProceedingsoftheSeminarforArabian Studies 36. 25-34.2006dTeArabianPeninsulaandIraq.Ammon,Dittmar,Mattheierand Trudgill(eds)Sociolinguistics.AnInternationalHandbookoftheScience ofLanguageandSociety,Volume3.Berlin&NewYork:DeGruvter. 1930-1941.2006eReviewofRosenhouse,I.andElad-Bouskila,A.(eds)Linguisticand Cultural Studies on Arabic and Hebrew. Zeitschri fur Arabische Linguistik 45. 87-90.2007aColloquial Iraqi Arabic. Postgate, N. (ed.) Languages of Iraq, Ancient and Modern. British School of Archeologv in Iraq. 123-134. niniiocv.vuvoiciiviuoiis xviii2007bKuwaitiArabic.Versteegh,K.,Eid,M.,Elgibali,A.,Woidich,M.and Zaborski, A. (eds) Encyclopedia of Arabic Language and Linguistics Vol 2: Eg-Lan. Leiden, Boston: Brill. 608-620. 2007cHello, I sav, and welcome! Where from, these riding men: Arabic popu-larpoetrvandpoliticalsatire:astudvinintertextualitvfromIordan. Ditters, E. and Motzki, H. (eds) Approaches to Arabic Linguistics. Festschri presentedtoKeesVersteeghonhis0thbirthday.Leiden,Boston:Brill. 543-563.2007d(withAbuAthera,S.S.)GeorgeBush,BedouinPoet.IournalofMiddle Eastern Literatures 10 No 3. 273-289. 2007e(withAbuAthera,S.S.)AnimalimagervinmodernBedouinpoetrv. Ouaderni di Studi Arabi Nuova Serie 2. 41-52. 2008a(withAbuAthera,S.S.)PoetryandPoliticsinContemporaryBedouin Society. Reading: Ithaca Press. xviii+353pp.2008bA Bedouin poem on the Iraq War. Toorawa, S. (ed.) Festschrift for Roger Allen. Al-Arabiyya 40. 3.2008cTe mixed Arabic of the letters of 19th and earlv 20th centurv Gulf rul-ers. Lentin, I. andGrandHenrv, I. (eds) Moyen Arabe et varietes mixtes de larabe travers lhistoire. Louvain-la-Neuve: Peeters. 193-229.2008dOmaniArabic.Versteegh,K.,Eid,M.,Elgibali,A.,Woidich,M.and Zaborski, A. (eds) Encyclopedia of Arabic Language and Linguistics Vol 3. Leiden, Boston: Brill. 478-491.due 2009Colloquial Arabic of the Gulf, 2nd edition. London:Routledge. Completelv revisedversionoftheColloquialArabicoftheGulfandSaudiArabia, 1984. co1i1s xix: :'...... ....... ..

..:.

: ...;

........... ......-.....

.....= .J;

.. ....

..

....

....

....

. ....

.....

......

... :.

....; J......=..... .......

....

..

....

.........= ..

....... ......

...;....

...-....

J. : ..-;

.. .........

.........

..

.... :J;...... ; .. ...-...

..'....

C...

....

.

...... .....

.... ..s

. ......... ..

..

.........

.-..... ..

......

............

.J. ...=....

......... ..

........ ..

......

............:;.........

.......-....

..

.

C........

..

.....

..

..............-....`I ....

...................... : ..

.

...;

........

......

J

.....

J....

............ .........

..

;......

..;...... ..

.. .:...

........

.......;

......

... .....-................

.. ..

....:..........

.;;..

..

;

......

C.....

. .

..-......

..

.......

......

..;

....'......

....

..

........... .......

C......

;..

......

.. J. : ..

.....

.. ......

................

.....

...:;.........

.;.....

....

....'. :

l..

............

..

;..:........ J. :..

........: .........;

...;

;...

.......... ..

.. .: : .....-............ :;..

..

; `....-. J..... :..

....

..

....'. : ....

......s ....

...

.. ....

..; ... .......

....... .. .....-...

..

......

.. ; ..- ....

..

....'. :.......:....'. :

.....

............C...

.............`I;

C...

.....

.. .........

.. ........

.....= ......-.

.......:...:

.. ....

..... ....

...... ......... .. ...... ....

.. :............... :..

.. :; ..- : .....s;

.....

.....

.....

....

:;

...

....

......

...

; C.

..... C;........J..-....

:;..

......

..=

.....sJ. :....

...

.;

.........

;

...

............

...

... ....

.-

. `I.

..............

;..

..'.......

..........

.

.......

............

....-. ............

...:

....; C

.......

.

.........

..'.

..

. ........= :..... ..- :C..

.....'.....

.......

.....'. :......=.....

..

.......

..`I;.

...

.; ...s........... :;.

....

...........

..;. J....

....

C............................

;..

..'. : ...:.. .....:.

.......

.. ......... ;

..... ..

...

....

..; C..:....... ....

........ ....; ..

; ..- :; .. ...-

. .. .... ;

...

... ..

.....

...... ..

.. .....s .....

....;

....

:;...

....

......... .J. ...=...

....: ....... :;....;

.........

... .....

...

..;

..

............. : co1i1s xxiui1ivmi.cv.u1uicomv.v.1ivimi1uou 1TOPICS IN HISTORICAL LINGUISTICSio.1u.owis 2iui1ivmi.cv.u1uicomv.v.1ivimi1uou 3INDETERMINACY AND THE COMPARATIVE METHOD: ARABIC AS A MODEL FOR UNDERSTANDING THE HISTORY OF ARAMAICIonathan Owens1.I1vouUc1ioHistoricallinguisticsseeksnotonlvtorenderplausiblereconstruc-tionsofearlierstagesofalanguage,butalsotolocalizepostulated developmentswithinspecifctimesandplaces.Semiticlanguages oferarichchallengetohistoricallinguisticsinthelatterrespect preciselvbecauseanumberofitsmembersareattestedeitherover longperiodsoftimeoroverbroadgeographicalareas,orboth. Ap plv ing the comparative method, the basic analvtic tool ofhistorical linguistics, to them potentiallv allows a detailed testing of its applica-tive generalitv. One issue is how temporallv precise, or determinate, solutions suggested bv the comparative method are. Tis question is addressed here, using Aramaic as the language of studv, and Arabic as a language of analogical support.12.Pvo1o-Simi1ic i Av.m.icTeissuecanbesetoutonthebasisofanarticleofscholarlvrich-ness,Driver(1926)approachedthequestionofthechronological datingoftheBookofDaniel.TeBookofDaniel,alongwithEzra, isofspecialinterestforAramaicistsandSemiticistsbecauseabout half of Daniel of the Hebrew Bible is written in Aramaic and hence it is one of the earliest extensive sources about Aramaic. Daniel him-self was a Iewish counselor in the Babvlonian court of Nebuchadnezzar whoreignedbetween605-562nci,hencedatingthetextisofcon-siderableinterest.DriversconclusionthatthetextofDanielwas written down in its current form around 300 nci against suggestions 1AsinHoles1991,thebulkoftheArabicdatacomesfromapplicationofthe comparative method to contemporarv varieties of Arabic. Classical Arabic plavs onlv a background role.io.1u.owis 4that the language of Daniel is from the era of the Biblical Daniel, has now been generallv been accepted (Collins 1993:16). Te interest for Driversworkhereisnottheissueoftextualdating,butratherthe tvpeofevidenceheadducestojustifvit,inparticularoneissue regarding the realization of proto-Semitic in Aramaic as z or d.2TemainsourcesDriveruses,besidestheBiblicalAramaic(BA) of Daniel and Ezra are the Aramaic papvri from Egvpt, dating from the earlv ffh centurv nci. Collectivelv these sources are a fairlv uni-formvarietvincludedwithinwhatissometimestermedOmcial Aramaic(Reichsaramisch).Driversremarkableattentiontodetail allowsonetointerprethisdataindependentlvofhisownconclu-sions. His own are as follows.AccordingtoDriver,BiblicalAramaicand/orthepapvrihadon a variational basis /z/ and /d/ as refexes of //, as in (1).(1)zn ~ dn this (masc.)zha ~ dha gold ziv ~ div relative clause/possessor markerTese two refexes are also found in the Egvptian papvri. Using care-ful quantitative observations, Driver notes that the earlier Ezra con-tained more z ~ d variation than did Daniel, with d becoming nearlv categoricalinDaniel.3Inthepapvrithereisasimilardecreasein variation over time in favor of d. Driver further points out that other changes in Aramaic tend to correlate with the move from z to d, for instance, variation in the refex of proto-Semitic as q ~ in earlier Aramaic(arq~arland)infavorofinlater,includingDaniel (1926:113).Driverwentbevondnotingthesvnchronicalternationinthese varieties,arguingthattherewasalinearsequenceinAramaiclan-guagehistorv,>z>d.Heexplicitlvrejectstheideathattheori-ginal split was simultaneous:2Namesofletterorgraphemesarewritteninquotationmarks,orarerepre-sented in the original script, their phonological realization in italics. Historical in thispaperusuallvreferstoattestedinawrittensourceinthemainscriptofthe relevantlanguage.HistoricalAramaicthusbeginsinthetenthcenturvnci,from which era the frst inscriptions derive (with caveats implicit in n. 5). Te historv of Aramaic, of course, is older than this.Iwouldnotethatthehistorvofproto-Semiticprobablvrunsparallelto, though this sound requires treatment of its own.3In his 1926 article, Driver simplv speaks of the two values, z and d, with-out specifving a phonetic value for them. Tat he saw these as phoneticallv diferent from and refexes of proto-Semitic is clear throughout his paper, as when he notes although both and represent an original dh (

:) (1926: 113).iui1ivmi.cv.u1uicomv.v.1ivimi1uou 5(2)4 zd Hearguesagainstthisbvnotingachronologicaltransitioninthe papvriwherezintherelativemarkerisfrstattestedin495nci, d not until 447. Bv 400 nci d had fullv gained the upper hand. Driver furtherdismissesaninscriptionfromtheearlvninthcenturvnci withdasascribalerror.Assvrianscribesmistakenlvwrotein Akkadian Addi-idri in place of the intended Hadad-ezer, the name ofthekingofDamascus.Itcanalsobenotedinthiscontextthat eventuallv,e.g.inSvriacandinSamaritanAramaic,aswellasin modernvarieties,thedvariantdidcompletelvsupplantz,sothe endpoint of Drivers sequence is at least correct. Against Drivers conclusion on this point, it is in fact easv to moti-vate the verv interpretation which Driver rejects, namelv the simul-taneousdevelopmentofzanddasin(2).Todavitisgenerallv acceptedthatspeechcommunitiesandindividualsinthemcansav exactlv the same conceptual thing in diferent wavs. Holes (1987), for instance, notes a series of phonological and morphophonological variants among the Arabic-speaking communities of Bahrain, some-times one and the same speaker using diferent variants of the same form. Alternations such as darasat ~ dirsat she studied, vkitbn ~ viktubnthevwritearerichlvdescribedandexplained,bothin terms of group-based and individual variation. Within Aramaic d ~ zvariationisattestedoveralongperiodoftime,beginningwith Biblical Aramaic and still present in Nabataean in 100 ci (Cantineau 1930:41).MuraokaandPorten(2003S:4)notethatd()isinfact attested around 500 nci in the Egvptian papvri, though confrming Driver, note that it increases relative to over the centurv. TeleastplausibleelementofDiversexplanationisitsphonetic implausibilitv. Te change from z to d is itself unmotivated. It is even less so in that in Aramaic d and z have an etvmologicallvindependent statusfromderivedrefexesofz~d.BAdbaqadheretoand 4Forthesakeoforientationitisrelevanttonotethatthischangerepresentsa merger, z and d being independent proto-Semitic phonemes: z z ddio.1u.owis 6zman time, for instance, derive from d and z. Whv all -derived refexesofz,andonlvthese,shouldsuddenlvmergewithdisnot explained. An extended coexistence of the two variants and an even-tualcompletemerging,asinSvriac,infavorofdisinlinewith standard variational-historical linguistic teaching. Driverslinearlv-orientatedexplanationforthedevelopmentof proto-Semitic in Aramaic is thus untenable. Drivers interpretation has not stood the test of time, vet there are elements of his explanation which need to be given greater attention inabroaderaccountofthedevelopmentofproto-Semiticin Aramaic. Contemporarv Aramaic studies ofer another interpretation ofthegraphemez()inOld5andOmcialAramaic.Among Aramaicists it is usuallv assumed that earlv Aramaic had , deriving from PS , which then developed into d. Segert (1997:117), echoing Degen (1969:34) notes that the letter z was used for both the sibilant /z/andinterdental//,sothatSegertpostulates(and)inOld Aramaic(1997:119).6Huehnergard(1995:268)suggeststhatOld Aramaicwaswrittenwiththephoneticallv-closestletter,namelv z.Kaufman(1974:117)seesthechange>dasoccurringatthe end of the Old Aramaic period. Garr (1985:26) gives a basis for the postulation of proto-Aramaic even in the absence of graphemic evidence. Both come from devel-opments which occurred in Omcial Aramaic. Te refex d in Omcial Aramaic could onlv have occurred from . Tis is an implicit rejec-tion of Drivers z > d development. In Garrs interpretation, as with Huehnergard and Kaufman, Old Aramaic graphemic z represented derivatesofbothetvmologicalproto-Semiticzandetvmological proto-Semitic . Tese are interpreted as distinctive phonemes, both in proto-Semitic, and in the earlier stages of Aramaic. Te develop-ment is thus:(3)PS > OA > Omcial Aramaic> > d5Roughlv,OldAramaic(Frharamisch)isthestagebeforeOmcialAramaic, approximatelv 1000 nci800 or 600 nci (Degen 1965: 1). Terminologv and dating of Aramaic varieties varies from scholar to scholar and indeed, when Aramaic as a linguistic entitv should be recognized is a matter of debate (see Huehnergard 1995). In anv case, extensive historical sources are not found before 1000 nci6Segert in an earlier work (1975: 91) considers, but does not adopt, the possibil-itv that the split of proto-Semitic took place in the Aramaic pre-historic period.iui1ivmi.cv.u1uicomv.v.1ivimi1uou 7In this view, there was no change > z, onlv > d. Tis interpreta-tioncurrentlvappearstobethedominantviewamongAramaicists (Kaufman, Huehnergard, Muraoka and Porten, Folmer 1995:49).Garr does give a second alternative. did split into d and z, but at diferent times and places. Te development of d, as in the Egvptian papvriandBiblicalAramaicisexplicableifhadsurvivedintoa later era, whereas the z of Old Aramaic inscriptions developed earlier. Tis alternative is basicallv represented in (2) above. While Drivers > z > d must be regarded as implausible, so far asIknow,whichalternative,(2)or(3)above,isthebetteronehas notbeenextensivelvdebated.Asseen,solution(3)hascurrencv todav among Aramaicists. Here I would like to take up this question, andinsodoingresurrectoneimportantassumptioninDrivers interpretation, namelv that proto-Semitic did indeed have a pho-netic value of z in earlv Aramaic. Tere are, in fact, two issues to be dealt with. Te frst is whether (2) or (3) better represents Aramaic language historv. Tis is the basic question. Te second deals with the era in which d and z arose, if at all,from.AswellasobservationsfromwithinAramaic,Iwill underpin mv arguments with analogical arguments from Arabic and Semitic in general, and from variationist theorv. 2. Diu Av.m.ic > z OccUv2.1. OrthographyAsseenin(3),manvAramaicistsdonotpostulateachange>z atall.However,aliteralistinterpretationofAramaicorthographv requiresthisinterpretation,sinceaferallhasminimallv,bvcom-mon consent the phonetic value of z. I think the literalist interpreta-tion is correct. To begin with, that d < occurred in Old Aramaic is attested in Drivers scribal error, described above, and in at least one token in anAramaictextinthe7th/6thcenturv(Segert1975:92).Equallv, occurs in various Aramaic sources all the wav into Nabataean times (Collins1993:16,Cantineau1930:41).Acontemporarilv-occurring d ~ z variation, or from the perspective of the original sources, a ~ variation in the same etvmological lexical set, is thus attested over 1,000 vears of Aramaic language historv. io.1u.owis 8No one to mv knowledge has argued that = obtains throughout this period. Instead, later occurrences of , for instance in Nabataean (Cantineau 1930:41, citing an inscription from 290 ci in the Hijaz) are held to be fossilized orthographic refexes. Te argument for fossilization I believe is introduced as a deus ex machina required on independent linguistic grounds, which weaken the case for the assumption of original = . Te background to this is as follows (see Muraoka and Porten:5). Beginning sometime in the OmcialAramaicperiodalinguisticchangesetinwhichthe(non-geminate) stops b, g, d, k, p, t spirantized afer a vowel, this pronun-ciation being represented here with an underscore, _g _p , and bvlackofadotintheletterintheAramaicscript.Whiletheargu-ment is not well-articulated, it is clear that if Vd > V occurred bv, sav 400 nci, then a frst centurv nci text with = (Collins 1993:14) cannotbeheldtorepresent,sincebvthistimeisanallophonic (post-vocalic) variant of d. If z did represent , whv wouldnt it be used as the post-vocalic variant V: Furthermore, if in the Omcial Aramaicperiodhadtwosources,onefromoriginalproto-Semitic ,asdescribedaboveandrepresentedasz(),andonefromthe new post-vocalic spirantization rule and represented as (), one wouldhavearemarkablecaseofaphoneticmergerheldapartbv etvmological convention.To save the situation, the idea of graphemic fossilization is intro-duced.doesnotrepresentinlateroccurrences,butratherafos-silized variant which originated as a representation of , but bv later times occurs onlv as a lexicalized convention.Agraphemicfossilizationwhichextendedoveratleast600vears of language historv (sav 400 nci200 ci) is decidedlv odd in Semitic languages,however.IngeneralinSemiticconsonantalscriptsindi-vidual phonemes are represented bv distinctive letters. Where there is a manv to one mismatch, as in the earlv Arabic script, it tends to be quicklv rectifed. As soon as Arabic began to be widelv used as a languageofwriting,itdevelopeddistinguishingpointsforconso-nants, such as :, originallv representing both d and giving rise to :and

:.Tisdisambiguationtookplacewithinthefrstcenturvof Islam, even if full implementation took longer. Itmightofcoursebearguedthatacertaindegreeofdiglossia allowedsuchgraphemicfossilizationtobemaintained.Tisseems to me to be the onlv general condition which could sustain the sug-iui1ivmi.cv.u1uicomv.v.1ivimi1uou 9gestion.TeissueofdiglossiainAramaicrequiresseparateconsid-eration,thoughfromtheoutsetcomparisonswithArabicdiglossia (Huehnergard 1995:275 n. 37) appear to me to be misplaced. Certainlv itneedstobemotivatedindependentlvofthoselinguisticissues which diglossia is invoked to explain. Here it can onlv be noted that in general Aramaic is a reform-orientated language, with diferent varietiesutilizinganddevelopingseparatescripts(e.g.easternvs. western), and diferent varieties indicating local linguistic changes in theirorthographv.TestronginfuenceofSamaritanHebrewon SamaritanAramaic,forinstance,notedextensivelvbvMacuch (1982:78-128), is apparent from the available sources. Whv onlv < should consistentlv fossilize is thus unclear.AgaininvokingaparallelfromArabic,inMiddleArabictexts, suchasGallandseditionof1001Nightsthereisvariationbetween ~ d, as in hi ~ hdi7 this (sg. fem.), which is transparentlv due toactualpronunciation.istheClassicalArabicnorm,whereasd represents the Svrian dialect which underlav Gallands edition (Mahdi 1984).Noonehassuggestedthatdinawordlikehdirepresents anvthing other than itself.Itisalsointerestingtonote,thatgiventheofenassumedstrong Nabataean infuence on the creation of the North Arabic script, the Arabic script never chose z to represent . In Nabataean, however, given the z = interpretation, such a model would have been pres-ent (as in Cantineau 1930:41, cited above). Takingtheseobservationstogether,thelackoforthographic inAramaicmeanssimplvthereisnoattestedinthelanguage (untiltheb,g,d,k,p,tspirantization).IfOldAramaicdidinfact haveitwouldatsomepointhavedistinguisheditortho graph -icallv. 2.2. Language Variation and ChangeGiven (2), with two values of proto-Semitic , d and z, the variation described in Omcial Aramaic bv Driver in section 1 above is entirelv inkeepingwithwhatisknownaboutvariationinlanguagecom-munities. In Arabic there have been numerous studies documenting thesvnchronicmaintenanceofcompetingvariantsfromanumber 7For instance, on p. 197 fve tokens of d in hdih (spelled hdh).io.1u.owis 10ofArabic-speakingregions.Holes(1987)onBahrainiArabiccited above,describescommunalcontactwhich,untilvervrecenttimes, appearstohavesupportedvariationalstabilitvacrossvervdiferent contactvarietiesoveratleast200vears(seealsoAbdelIawad1981, Owens 1998, Al-Wer 2003 for further). Evidence from Arabic in 2.4 below will confrm that multiple refexes of an originallv unitarv pho-nemecanco-existoververvlongperiodsoftime.WhileDrivers historicallinguisticconclusionsaresuspect,hiscarefulvariationist readingofthedata,withdiferentgraphemicvariantsrepresenting diferent pronunciations, is entirelv plausible.2.3. Proto-SemiticTe change of > z is well attested in Semitic. It is, in fact, the rule. All Ethiopic Semitic has it, Akkadian does, and so does Hebrew, the closest,well-attestedsisterofAramaic.Indeed,givenotherhighlv characteristic shared retentions or shared innovations with Hebrew (p,,x>,>,etc.),itwouldalmosthavebeenperversefor Aramaicnottohaveshared,initiallvatleast,inthe>zshif. Furthermore,theinnovationssharedwithHebrewareearlvones, vervlikelv(intheviewtakeninthispaper)inthepre-historicera (historicalinthesenseofn.2).Te>z/dchangeequallvfts into an earlv spectrum of change in NW Semitic.2.4. Parallels in ArabicArabic is one of the few Semitic languages to have maintained proto-Semitic>uptothepresentdav.Ithas,however,notdoneso uniformlv across all varieties. Refexes of are, broadlv, as fol lows. > , Eastern Arabia, Yemen and Persian Gulf, Iraq, Eastern Libva, occasionallv in North Africa (e.g. Tunisia, Mauretania)d, Egvpt, Sudan, Iedda, Damascus, North Africaz, Uzbekistan, Anatolia (Azex, Iastrow 1978:36-7))v, Anatolia (Siirt)d(> d:), Western Sudanic ArabicTerefexdfoundinWesternSudanicArabicperhapsarisesvia anemphaticvariantof,asoccursinanumberofArabicdialects (e.g. Iordanian h a this) and bv this reckoning could be conjoined iui1ivmi.cv.u1uicomv.v.1ivimi1uou 11with the > d change. Whatever the origin, it will not be considered further here.Onageographicalbasismaintenanceoftheprotoform>is the most widespread, followed bv d, with z and even more so v being restricted.TepresenceofzinUzbekistanArabic(CentralAsian Arabic)issignifcantonhistoricalgrounds,asthisvarietvformsa SprachinselwhichwascutoffromtherestoftheArabic-speaking worldbvtheendoftheeighthcenturv.Teexistenceofthesame refexinAnatoliasuggestsanearlvcommonorigin.Assumingthis earlv origin, along with its continued existence todav, this refex has existedforsome1,200vears(ataminimum),co-terminouslvwith . Note that this time span is approximatelv the same as the proposed co-existence of the d ~ z refexes of in the Aramaic up to Nabataean times. Tedevelopmentofpresent-davdistributionscanbemodeled using the wave representation of language change (e.g. Bailev 1973). InDiagram1,thenumbersinthetoprowrepresenthistoricallin-guisticstages,therowsrepresentingdialectalvariants.Roughlv speaking, the present-dav situation is as in step 3. Steps 4 and 5 are future hvpothetical developments, as will be discussed presentlv.8Diagram 1. in Arabic, past, present and hvpothetical future12345ddddddddddvvvvzzzzAtsomepoint,somespeakersinnovatedtov,z,andd.9Iassume thatthechangestovandzwereroughlvatthesametime,asthev 8Note that there is no need here to cite Classical Arabic as justifcation of an originalproto-Arabic.Tisfollows,Ibelievebvinspection,fromtherefexesof the forms in present-dav Arabic (roughlv, stage 4). Of course, it is relevant that Clas-sical Arabic () does not contradict this reconstruction.9Or, Arabic acquired speakers who substituted for . ItcannotberuledoutthatthemergerofwithdinsomedialectsofArabic wasnt due in part to sub- or adstratal infuence from Aramaic. Given the natural-nessofthechange,however,provingthisisprobablvimpossible,andinanvcase does not afect the theoretical point being made. io.1u.owis 12areattestedonlvincontiguousareas.Ialsoassumethatatleastthe change to z occurred earlv (bv 800 ci), to account for the Uzbekistan-Anatolia isogloss. Asnoted,Step3isroughlvwherethedistributionsoftheforms are todav in geographical terms, though if one were representing the distributions to scale, the z and v refexes would be quite small. Tere is no certaintv that afer Step 3, Step 4 will ensue. It is true that local changesof>dareattested,forinstanceintheArabicofAmman and in Bethlehem Arabic (Al-Wer p.c. 2007,10 Amara 2005), though it remains to be seen whether this change will spread throughout the region. Nonetheless, steps 4 and 5 are added in order to make the analogv to Aramaic concrete. Hvpotheticallv, step 4 could occur, and as noted inthepreviousparagraph,ifdoeschange,itchangestodinthe contemporarvArabicworld.Step5wouldbethefnalchangeof to d and at this stage no more refexes would remain. Te analogv to Aramaic is clear, and indeed, the fact that a nearlv identical change didgotocompletioninAramaiclendsspeculativeconcretenessto our stages 4 and 5. Step 1 represents a proto-Semitic origin, and from step2innovationsbegin,withmerginginsomevariantsordia-lectsofthelanguage.OurhvpotheticalStep5representseither(2) or (3) above, the endpoint of a merger that has fnallv moved across the entire language communitv. Returning now to the major problem, whether (2) or (3) best rep-resentsthelinguistichistorvof,onthebasisofouranalogical reasoning,between(2)and(3),therelevantanalogvwithArabicis (2).InArabichasultimatelvsplitintod,zandv.Tisproduces a nearlv perfect parallel with (2), the onlv diference being that no v variantisattestedinAramaic.Tisinterpretation-bv-analogv 10Al-Wer (p.c.) notes that Amman Arabic is decisivelv infuenced bv migration from West Bank Palestinians and from rural Iordanian speakers. Urban Palestinian dialects have d as a refex of , but Iordanians generallv have . Currentlv in Amman developmentsaremovinginthedirectionofthestopvariantd,aligningAmman with Damascus and Beirut. Nonetheless, Al-Wer (p.c.) notes that evidence of the pre-mergerstatewillbepresentforavervlongtimesincethesplitstateofafairs continues to be the majoritv form in the countrv as a whole (although cruciallv not sointhelargeurbancentres);relicforms,notfromtheStandard,willalwavsbe found in Amman in the future. Te parallel with the current data is clear: alternative variants of the same proto-form can subsist side bv side over long periods of time. iui1ivmi.cv.u1uicomv.v.1ivimi1uou 13 combinesDriversvariationistz~dstage,withcontemporarv Aramaicists derivation of all refexes from . OurhvpotheticalexamplefromDiagram1furtherallowsusto clarifv and criticize the reasoning behind the alleged development in (3) above. i.Let us assume a situation where stage 5 is now the contemporarv stage and in this stage there is a one-to-one graphemic-phonemic mapping, hence d = d, v = v and z = z. ii. Further,assumethatstage5refexesleadtoareconstructionof , from which all varieties attested in stage 5 derive. iii.Furtherassumethattheonlvoldersurvivingsourcesarethose of a form of Middle Arabic where etvmological is written onlv as d (e.g. onlv haadih, as in n. 7).Giventhehvpotheticalcontemporarv(stage5)refexesd,v,z,one might be led to assume that our Middle Arabic d was used to rep-resentbothetvmologicald,andthereconstructedancestorofd ~ v ~ z, i.e. one would claim that the scribes used d for both d and . But in our controlled thought train, we know that this conclusion wouldbeerroneous.Tesurvivingtokenofdinfactrepresented onlv Middle Arabic d = d. Mv claim for understanding Old Aramaic z is similar. Aramaicists havesimplvworkedintheoppositedirectionfromthereasoning given in the previous paragraph (see Owens 2006:21 for similar cases amongArabicists).Giventheexistenceofproto-Semitic,thev conjecture that z represented the earlier proto-form . However, there is no guarantee that this inference is anv more correct than the patentlv incorrect inference outlined in the previous paragraph that MiddleArabicdrepresentedmorethanonevalue.Instead,itis equallv likelv that z represented something else; for instance, itself, z. TereisafurtherlessonfortheinterpretationoftheAramaic developmentsthatcanbederivedfromourextendedanalogvwith Arabic.Assumingstage5isreached,isavailableonlvviarecon-structionsinceitisnotattestedinthecontemporarvcommunitv. Changingassumptioniii.aboveforthesakeofillustration,assume thatbvstage5nodirectevidenceforsurvivedinourrecordsof Arabic, either in written attestations or in the dialectal and sociolin-guisticstudieswhichattesttoitspresencetodav.Inthissituation, nothing comparable to Middle Arabic is available. It would be impos-sibletoprovethestep-bv-stepdevelopmentofthelossofas io.1u.owis 14 representedinDiagram1.Adevelopmentsuchasthefollowingis onlv a logical possibilitv. Diagram 2 essentiallv is (2) above, with an added stage 2 in between. Tis represents Garrs second suggestion, that > z/d at diferent times.Diagram 2. Possible development of d and z in Aramaic123dzzDiagram 2 could be inferred out of general theoretical assumptions certainlv,andbvouranalogvwithincontrovertiblefactsofArabic, representedinDiagram1.However,Garrssecondsuggestion,that existed as a dialectal variant, along with z (< ) is not supported bv anv direct evidence in the historical record.Diagram 2 is situated in comparative linguistic time, not in chro-nological time. Where should Diagram 2 (or the split represented in (2)above)besituatedhistoricallv:Giventhelackofanvdirectevi-denceforwithinthehistoricalAramaicera(beginningthetenth centurv nci), it can be concluded that Aramaic entered the historical era in stage 3 (Diagram 2), which is equallv the endpoint of the split represented in (2) above.In the fnal analvsis, the indeterminacv of the comparative method canbeconstrainedbvourconsiderationshere,buttheconstraints areratherlax.Ifstage2inDiagram2occurred,itisinferredonlv indirectlv.Whenstage3occurredisuncertain.Atacertainpoint, all historical linguistic conclusions become inferential. I believe that thebestinterpretationofthedataistherepresentationin(2),that splitintozanddinapre-historicalperiod.Variationaltheorv easilvsupportsthemaintenanceofz~drefexesoverlongperiods oftimethereafer,andthesplititselfissanctionedbvasoberand straightforward reading of Aramaic orthographv, as well as bv appli-cationofthecomparativemethod.Furthermore,analogieswitha nearlv identical development in Arabic, along with an extended but plausiblehvpotheticalscenarioforafuturedevelopmentofArabic helpsustobetterconceptualizethecompetinginterpretations,and all in all lend credence to the interpretation proposed here. iui1ivmi.cv.u1uicomv.v.1ivimi1uou 15Biniiocv.vuvAbd-elIawad,Hassan.1981.LexicalandPhonologicalVariationinSpokenArabic in Amman. PhD thesis, Univ. of Pennsvlvania.Al-Wer, Enam. 2003. Variabilitv reproduced: a variationist view of the [d

]/[] oppo-sitioninmodernArabicdialects.M.Haak,R.deIong&K.Versteegh(eds), Approaches to Arabic dialects. Leiden, Boston: Brill. 21-31.Amara,Muhammad.2005.Language,Migration,andUrbanization:theCaseof Bethlehem. Linguistics 43. 883-902.Collins,Iohn.1993.ACommentaryontheBookofDaniel.Minneapolis:Fortress Press.Bailev,C-I.,1973.Variationandlinguistictheory.CenterforAppliedLinguistics, Arlington VA.Cantineau, Iean. 1930. Le ^abateen I. ^otions generalesecriture, grammaire. Paris: Librairie Ernest Leroux.Cantineau,Iean.1935.Le^abateenII.Choixdetexteslexique.Paris:Librairie Ernest Leroux.Degen, Rainer. 1969. Altaramische Grammatik. Wiesbaden: Steiner.Driver, G. R. 1926. Te Aramaic of the Book of Daniel. Iournal of Biblical Literature 1. 110-119.Garr, Randall. 1985. Dialect Geography of Syria-Palestine, 1000-58 n.c. Philadelphia: Universitv of Philadelphia Press.Folmer,Margaretha.1995.TeAramaicLanguageintheAchaemenidPeriod. Orientalia Lovaniensia Analecta 68. Leuven: Peeters.Holes,Clive.1987.LanguageVariationandChangeinaModernisingArabState. London: Kegan Paul International.Holes, Clive. 1991. Kashkasha with fronting and africation of the velar stops revis-ited: A Contribution to the historical philologv of the pensinsular Arabic dialects. Alan Kave (ed.), Semitic Studies in Honor of Volf Leslau. Wiesbaden: Harrassowitz. 652-78.Huehnergard, Iohn. 1995. What is Aramaic:. Aram 7. 261-82.Iastrow, Otto. 1978. Die mesopotamisch-arabischen Oltu-Dialekte. Bd. 1. Phonologie und Morphologie. Wiesbaden: Steiner.Kaufman,Stephen.1974.TheAkkadianInfluenceonAramaic.Chicago:Oriental Institute of the Universitv of Chicago, Assyriological Studies 19.Macuch,Rudolph.1982.GrammatikdessamaritanischenAramischen.Berlin: Mouton de Gruvter.Mahdi, Muhsin. 1984. Te Tousand and One ^ights. Leiden, Boston: Brill.Muraoka,TakamitsuandBezalelPorten.2003.AGrammarofEgyptianAramaic, second revised edition). Leiden, Boston: Brill.Owens, Ionathan. 1998. ^eighborhood and Ancestry. Variation in the Spoken Arabic of Maiduguri ^igeria). Amsterdam: Benjamins.Owens,Ionathan.2006.ALinguisticHistoryofArabic.Oxford:OxfordUniversitv Press.Segert, Stanislav. 1975. Altaramische Grammatik. Leipzig: VEB Verlag.Segert,Stanislav.1997.OldAramaicPhonologv.AlanKave(ed.),Phonologiesof Africa and Asia. Winona Lake: Eisenbrauns.115-126.io.1u.owis 16ivom qltu 1o glt 17FROM qltu TO glt: DIACHRONIC NOTES ON LINGUISTIC ADAPTATION IN MUSLIM BAGHDAD ARABICHeikki Palva1.I1vouUc1ioSinceHaimBlancsCommunalDialectsinBaghdad(1964),itisa well-known fact that the dialects spoken in Baghdad can be divided into three communal groups, corresponding to the speakers religious amliation. Te dialect of the Muslims (MB) belongs to the so-called glt group of Mesopotamian Arabic, whereas the dialects of the Iews1 (IB) and the Christians (CB) belong to the qltu group. Historicallv, IBandCBareregardedasdirectdescendantsofmedievalIraqi Arabic, whereas the present-dav MB with its numerous Bedouin-tvpe traits clearlv diverges from it. As a glt dialect MB is ofen classifed as a dialect of Bedouin tvpe, or being of Bedouin provenance.2 Tese are of course rather impres-sionistic notions based on a number of linguistic variables the relative tvpologicalsignifcanceofwhichdependsontheadoptedstand-point.3 Tus, in comparison with qltu dialects, most diferences dis-plaved bv MB are undeniablv of Bedouin tvpe, but if MB is compared 1TeIewishcommunitvinBaghdadwasrelativelvbig:mostIewslefIraqin 1950-51 and are now settled in Israel. In the nineteenth and twentieth centuries the IewsandChristianslivedintheirancientquartersnorthandwestofSqal-Ghazl respectivelv,whilethepopulationofal-KarkhwasmainlvMuslimArabs(Duri 1960:907). In 1884 there were 30,000 Iews in Baghdad, bv the beginning of the 20th centurv50,000,andca.100,000IewsaferWorldWarTwo.http://www.bh.org.il/Communities/Archive/Baghdad.asp 15.11.2007. 2Bedouin-tvpedialects,suchasarespokeninsouthernIraqincludingBagh-dadHoles1995:57;gltArabicisofBedouinprovenance,unlikeChristian Baghdadi Abu-Haidar 2006a:222.3 Te relevance of all linguistic classifcations depends on the aim as well as the criteria applied. If the interest is purelv svnchronic, the classifcations can be made onthebasisofanadequateselectionofsvnchronicallvwell-documentedlinguistic variablesforeachdialectorgroupofdialects,withoutconsiderationofdiachronic and extralinguistic criteria. If the interest is focused on cultural and historical points of view, diachronic and comparative data plav a crucial role (Palva 2006:604).uiixxiv.iv. 18withBedouindialectsofAnazorammar,orevenofthewiya tvpe,itssedentarvproflewouldbecomeapparent(cf.Palva1994: 460-465; id. 1997 passim).Teaimofthispaperis,however,nottotrvtoestablishappro-priate criteria for an adequate classifcation of MB. Here the approach isdiachronic.Itismvaimtoexaminethehistoricalinteractionof qltuandgltdialectsthathasresultedinthepresent-davMuslim dialectofBaghdad.AsthematerialIuseanumberoftvpologicallv prominent linguistic features in MB representing, on the one hand, theurbandialecttvperelatedtoIBandCB,andontheother,the rural tvpe related to Bedouin dialects spoken in southern Iraq and its neighbourhood.Inashortpaperthenumberofvariablestreatedis necessarilv restricted, and their relative weight in the argumentation mav be questioned. BecauseMBhasnotbeendocumentedbeforethetwentiethcen-turv, its historical development can onlv be traced bv using the pre-sent dialect as the starting-point and bv considering its structure in the light of common rules of linguistic change, particularlv in contact between dialects of Arabic. Although changes that have taken place hundredsofvearsagohaveprobablvfollowedthesamecommon rulesastodav,thediachronicconclusionsdrawndependonmanv uncertain presumptions concerning matters such as communication patternsbetweendiferentsocialandreligiousgroupsorlanguage attitudesamongthemindiferenthistoricalcontexts,nottospeak aboutvarvingfashionableorstigmatizedlinguisticfeaturesamong diferent social groups. Yet, recent and ongoing developments in dif-ferentdialectscertainlvareserviceablepointsofcomparison.Asto individuallinguisticfeatures,relativelvreliablediachronicconclu-sions can also be drawn on the basis of past developments documen-ted in other dialects. 2. qltu-1vvi Fi.1Uvis i 1ui MUsiim ui.iic1 oi B.cuu.u2.1.Use of the Refex q of q in a ^umber of Items, Instead of the Main Refex g2.1.1.TeseincludenumerouslexicalborrowingsfromStandard Arabic, e.g., qallad to copv, imitate, qarrar to decide on, aqqaf to impart education, rqab to watch, observe, abbaq to applv, qtira ivom qltu 1o glt 19tosuggest,recommend,qrantocompare,qayyamtoestimate, assess,niqadtoreview,wiaqtotrust,niqad

tocancel,abolish, waqqatosign.Tisgroupconsistsmostlvofitemsconnectedto diferentinstitutionsofmodernsocietv,andtheqrefexisadopted as part of the lexical item.2.1.2.Ontheotherhand,thereisanothergroupofwordswhich displavtheqrefexbutwhichbelongtotheevervdavvernacular vocabularvandobviouslvcannotberegardedasborrowingsfrom Standard Arabic. Tese core items include, e.g., the following: qubad

toreceive,tocollect,qubal/qibaltoagree,toaccept,qarrabto causetocomenear,togetclose,qiratoread,torecite,qisam, qassamtodivide,qiadtointend,qaartodrop,qifaztojump, qufa to catch, qufal to lock, qilla shortage, qall few, quwa to be ortobecomestrong,quwwastrength,qawistrong,qbottom, baqar cow, cattle, buqa to stav, to remain, iqa to fall, to top-ple,aqqatocausetofall,iqabtofollow,tosucceed,xilaqto create, xzaq to stick, to cheat, to take in, sibaq to be, come, or happenbeforeoraheadof ,toprecede,andarqwav.Temost plausible explanation to the use of the q refex in this group of items isthatthevareinheritedMBqltuwordsadoptedwiththeqrefex bvruralimmigrantsatastagewhentheurbanMBqltuforthem wastheprestigedialect,thatis,beforethebedouinizationprocess started.2.1.3.In some other high-frequencv items which at frst sight would seemtobelongtothecoreitemcategorv,thebackgroundoftheq refexisdiferent.Tus,e.g.,theMBitemssqtodriveandsyiq driver are obvious Standard Arabic forms, whereas the genuine dia-lectal sg is used in the meaning to drive, herd (animals). Another qgpairofthesametvpeisliaqtoattach,append,cf.liagto follow,trailafer.Tereareseveralotherqgpairswithdiferent oroverlappingsemanticfelds,whilesomearefreevariants,e.g., qarrabtocausetocomenearandgarrabtobringclose,toget close; farraq to divide and farrag to distribute, dispense; qisam and gisam to divide, split; xuluq character and xulug temper; siqa and siga to water, provide water for; idaq and idag to be truthful, tell uiixxiv.iv. 20thetruth.4Insuchcasesthelexicaldistributionofthevariantsis moving towards stvlistic variation (see 3.1. below).2.1.4.Inpresent-davMB,formswithqareasarulevoungerthan those with g, with the exception of a number of items exemplifed in section 2.1.2. above. A further group, related to both 2.1.2. and 2.1.3., consistsoftechnicaltermsassociatedwiththeurbancultureofthe past.Abu-Haidarillustratesthecasewithafewstrikingexamples. One of them concerns the term saqqa water-carrier, which used to be part of MB vocabularv when water-carriers were a common sight in inner Baghdad, but which is unknown to the vounger generation: Anelderlvwoman,whenaskedwhatawater-carrierdid,replied: is-saqqa n yisgna mayy the water-carrier used to give us water to drink,usingthe/q/,/g/morphophonemicalternation.Compare also: nigalit il-manqala I carried the brazier, and giad al-maqad he sat on the seat, &c. (Abu-Haidar 1987:47, n. 14.) Here saqqa, man-qala and maqad are old, well-established technical terms for instru-ments used in urban environments, adopted bv rural immigrants in their qltu-tvpe form. 2.2.Use of the Verb Modifer da- the Most Common Function of Vhich is Present Continuous or Habitual Action Examples:dayiktibheiswriting,datsawwi?whatarevoudoing: (Blanc 1964:115-116; Malaika 1963:80; Abu-Haidar 2006a:229).5 Te useofverbmodiferstomarkdiferenttenseandaspectcategories is a prominent sedentarv feature verv well developed in all qltu dia-lects (see Iastrow 1978:299-311), whereas in rural glt dialects these categories as a rule are unmarked. Te same verb modifer is used in IBandCBaswell,but,signifcantlv,inthesedialectsitoccursonlv withthe1stp.sing.andplur.oftheimperfectandhasoptative 4Abu-Haidar1987:46givesalistof15itemsinwhichtheg>qshifinvolves semantic change from concrete to more abstract, sophisticated meanings, e.g., agg bride-price, aqq right, truth. She also gives 5 examples ofolder, well-established forms with q co-occurring with corresponding forms with g, e.g., warga leaf, warqa pieceofpaper;gubbaroom,qubbadome.Athirdlistconsistsof6qgpairsin whichtheitemswiththeqrefexaretechnicaltermsassociatedwithmedicineor science, e.g., fatig rip, tear, fatiq hernia. 5It mav also function as an optative marker as in IB and CB, but, in contradistinc-tion to these, also be preceded bv xal- when the form is an unambiguous optative, e.g., xaldangl lets sav (Blanc 1964:116).ivom qltu 1o glt 21function (Blanc 1964:116; Mansour 2006:239), while the present and non-contingencv markers in them are IB qad-/qa- and CB qa- (Blanc 1964:115; Abu-Haidar 1991:88; Mansour 2006:239). Apart from MB there are no Muslim dialects in the Mesopotamian dialect area which usethepresentmarkerda-;consequentlv,thefeaturehastobe regarded as an inherited MB qltu trait. 2.3. Use of Futurity Markers la and raE.g., ma-ra yii hes not going to come (Blanc 1964:117-118; Erwin 1963:138-139); according to Malaika (1963:82), la is the more com-monlvusedvariant.InIBandCBthemarkerisa(Abu-Haidar 1991:88-89; Blanc 1964:117 CB ), which implies that, apart from the diference as to the non-Muslim refex of OA r, the markers are identical. Tis is strong evidence of the common qltu background of the feature. As a salient sedentarv feature it could also be explained as a result of the natural drif towards a relative tense svstem, but in that case it would be an unlikelv coincidence that the form of the marker in MB would have become practicallv identical with those in IB and CB.Generallvspeaking,theuseoffuturitvmarkersderivedfrom ryiisanoldurbanfeatureintheMashriq,assuggestedbvits occurrencealsoinEgvpt(a-,ha-,Behnstedt&Woidich1985: 224-225),Damascus(la(a)andra(a),morerarelva-,Grotzfeld 1965:87), Beirut (-, Behnstedt 1997, Map 162), and parts of Leba-non (a-, ibid.).6 2.4.Use of the Proclitic d()- to Add to the Imperative a ^ote of Impatience or to Intensify the Sense of the ImperativeExamples: duklu eat (pl.), ditkassa clear of! Tis feature is of com-mon occurrence in MB, IB and CB (Blanc 1964:117; Malaika 1963:84; Erwin 1963:140). In addition to IB and CB, the imperative modifer d()hasnotbeenattestedinqltudialects;tomvknowl edge,the onlv exception is the Mardin group (d or prefxed d-, Mardin town 6Te use of futuritv markers in sedentarv dialects of the area probablv dates from medievaltimesalreadv,asissuggestedbvthefactthatdialectswhichmakeuseof markers going back to the conjunction att, used in Anatolian qltu dialects (ta-, t-, Iastrow1978:301-302)andthedialectsspokeninthesurroundingsofAleppo(ta-, Behnstedt1997,Map162),sharethistraitwithCvpriotArabic(tta-,ta-,Borg 1985:101-102). Te markers < att seem to have been an earlv northern SvrianMesopotamian feature. uiixxiv.iv. 22alsod,usedlikeaninterjectionbeforetheimperative,Iastrow 1978:311-312).Tisindicatesthatitis,orhasbeen,amoreorless restricted local feature. Since the form and function in IB and CB are identical with MB, the possibilitv that it in MB is a feature borrowed fromIBorCBcannotbedefnitelvexcluded.However,asdialect shifs as a rule tend to move in the direction of the prestigious variant, thisdevelopmentisunlikelv.Tereforetheplausibleconclusionis that in MB this is a trait inherited from medieval MB qltu.2.5.Marking the Defnite Direct Object of a Verb with an Anticipatory Pronominal Sumx + a Proclitic l-Examples:balil-bthesoldthehouse,ma-aibbal-haIdont likehim(Feghali1928:362-363;Blanc1964:128;Malaika1963:63; Erwin 1963:332; Abu-Haidar 2006a:230-231). Signifcantlv, this con-struction appears in IB and CB as well. It is also worth noticing that MB makes more ofen use of the unmarked construction while in CB the marked construction occurs more commonlv (Abu-Haidar 1991: 116); most frequent it is in IB, according to Blanc, actuallv the normal one (Blancs emphasis, 1964:129). In sedentarv Arabic dialectsspoken in the Mashriq this svntactic feature is an obvious Aramaic substrate device (e.g., Feghali 1928:362-363). In view of the language situation in the SvrianMesopotamian area during the last pre-Islamic and the frst Islamic centuries it is not unexpected that this feature also occurs in Maltese and Cvpriot Arabic, spoken bv Christians, as well as in the insularArabicdialectsinCentralAsia,spokenbvMuslims(Borg 1981:35-62;Borg1985:138;Blanc1964:130;Blau1961:215;Fischer 1961:262-263).AsanAramaicsubstratedeviceitsmorefrequent occurrence in IB and CB is rather natural, but there is no reason to supposethatitwouldhavebeenadoptedbvMBspeakersfromthe non-prestigious IB or CB. Terefore, in MB it must be regarded as a trait inherited from the medieval MB of the qltu tvpe.2.6.Use of a Prefxed Indetermination Marker, a Variant of Vhich in MB, IB and CB is fad(d)Examples: hayy nat fadd fkra mumtza that was an excellent idea (Erwin1963:355);waladzn/fadwladznagoodbov;cf.some, one:fadymonedav,fasomething(Blanc1964:118,126; Malaika 1963:69; Erwin 1963:355-358). ivom qltu 1o glt 23Allthreedialectshavealsoanuncontractedvariantoffard:MB fard, IB faad, CB fad. Te development of a new indetermination marker is a sedentarv feature found in the Mesopotamian dialect area as well as in a number of North African dialects (wid plus article). Apart from the dialects spoken in Baghdad, it is attested in the qltu dialectsofna(fadd,fard)andMosul(fadd,fad),andsignif-cantlv, insular Arabic dialects in Central Asia (fat), a fact that is evi-denceofitsearlv,Abbasidtimesemergence(Blanc1964:119).It might be worth noticing that a marker of the same tvpe is used in the surrounding languages: Turk. bir, Pers. ye(k), Northeastern Neo-Ara-maic xa (Blanc 1964:119; Edzard 2006:189.)2.7.Absence of Separate Feminine Plural Forms in Personal Pronouns and Finite VerbsTisisafeaturesharedbvmostpresent-davsedentarvdialectsof Arabic, whereas in Bedouin dialects to the east of Egvpt and in rural dialectsspokeninsouthernMesopotamiaandtheGulfcoast,the femininepluralformsareretained(Ingham1982:38).InMB,the reductionofmorphologicalcategoriesinpersonalpronounsand fnite verbs is probablv an inherited qltu trait, although the natural drifcombinedwithdialectcontactwouldprobablvhaveledtothe same development, as it has actuallv done as part of sedentarization process, e.g., in urban centres such as Basra, Zubair and Kuwait (ibid.; Palva 2006:611). 2.8.Absence of Form IV as a Productive Morphological CategoryAccording to Blanc, traces of Form IV verbs can be recognized onlv bv the perfect and the participle, such as the MB nt I gave, mni having given. In addition, there are some fossilized items, e.g., agbal he came forward and awda he took leave of (Blanc 1964:111). Te qltu-tvpe vernacular dialect(s) spoken in Baghdad in Abbasid times mav actuallv have lacked Form IV, as suggested bv its absence in IB and CB (see Mansour 2006:238; Abu-Haidar 1991:50-51). Terefore it is worth noticing that also the sedentarv Barna dialect does not use Form IV except in fxed expressions; otherwise it is replaced bv FormsIand(mainlv)II(Holes2006:252).AccordingtoIohnstone, FormIVisusedinBedouin-tvpedialectsspokeninKuwaitand Oatar, how productivelv, is however not mentioned(Iohnstone 1967: 73, 111).uiixxiv.iv. 242.9.Absence of Salient Bedouin-type Syllable StructuresTe tvpologicallv most conspicuous Bedouin-tvpe svllable structures arethegahawasvndrome,e.g.,gahwa>gahawa,yaxbu> yaxabu and the Najdi resvllabifcation rule, e.g., gahawa > ghawa, yaxabu>yxabu,katabat>ktibat;zalama>zlima(DeIong 2007:151-153), which are absent from MB. Tis is an obvious major case of phonetic adaptation bv immigrant Bedouin speakers, the ex-Bedouin rural population in southern Iraq included (Blanc 1964:166; Iohnstone 1967:6-7).7 3.glt-Tvvi Fi.1Uvis i 1ui MUsiim Di.iic1 oi B.cuu.u3.1.Use of the Voiced g Refex of OA qInaclassifcationbasedonlinguisticcontrasts,thevoicedrefexof OA q is the mostexclusive Bedouin feature. Te Bedouin character ofMBis,however,compromisedbvthefactthatinanumberof items the refex is q, and on the other hand, that the g refex is used without its phoneticallv-conditioned africated allophone , which is part of the phonetic svstem of the neighbouring Bedouin dialects as well as, e.g., of the dialect spoken in Basra and Kuwait until the mid-twentieth centurv (Iohnstone 1967:5, 29-32). However, in a few items the africation has been retained and adopted as an established form usedinBaghdadregardlessofreligiousamliation,e.g.,Bbe-ari the East Gate (quarter), (hawa) ari east wind, but: arqi eastern (Blanc 1964:27). Te form ari mav well be explained as an adopted Bedouin term which, used with the defnite article, is the name of a special kind of wind, but the historical context in which the name of a citv quarter has been established in its Bedouin form is not known.A comparison with urban dialects that displav certain Bedouin fea-tures suggests that the frst step in the bedouinization process proba-blv is adopting separate lexical items associated with the rural sphere. Tis stage obtains in urban Maghribi Arabic, which usuallv displavs g in rural-based items such as bagra cow, girbe waterskin, gn tent andgnnrabbit(Marais1902:17;Heath1989:6;Boucherit2002: 7Formssuchaslaammeat,aharmonth,baarsea,baamuledonot belong to the gahawa svndrome cases, but the latter a-vowel is an anaptvxis (Blanc 1964:55).ivom qltu 1o glt 2540-41). Te second stage is reported bv Talmoudi from Ssa (Sousse): In the old genuine dialect of the town, the q and g refexes of q are in lexicaldistribution,whereasinthekoineizedSsiofthesuburbs and of the vounger generation the pattern is changed: [g] is rather a stvlistic variant of [q] and can replace it anvwhere, whereas the /g/ in earlvloanscannotbereplacedbv/q/.(Talmoudi1980:22-23.)Te adoption of the g refex does not necessarilv implv that the rural dia-lects are in the majoritv or are felt as prestigious, but this particular feature mav be perceived of as carrving positive connotations such as toughness,manhoodandmasculinitvinAmman(Abdel-Iawad 1981:176; Palva 1994:466). In Tunisia, the rural g refex has infltrated from the suburbs of Ssa to the Medina, where voungsters use it in order to appear as tough (Talmoudi 1980:22). A parallel development has been attested in the Euphrates group ofMesopotamianqltudialects.Tus,naandHthavegrayyeb near,Drez-Zrbgaacowandngashe-camel,whichallare Bedouin loanwords. On the other hand, both dialects have items like gahwacofee,bagracow(naandHt,Blanc1964:27)andgaib heart,gground,bg,ybgtosteal(Drez-Zr)inwhichghas been substituted for the older qltu -tvpe q (Iastrow 1978:42). In the Euphrates group dialects this development mav have been going on for several generations, as is indicated bv the fact that forms such as gdrtobeable,gtobreakup,q/gwltosav,bagracow,rigba neck, and grb close were attested among the Karaite Iews at Ht bv Khan,whosefeldworkamongthemwasundertakeninBeersheva fortv vears afer thev lef Iraq (Khan 1997:56). In all probabilitv, these frst signs of incipient bedouinization reached the Muslim population of the town well before the local Iews.3.2.OA k > k/ Vith a Mainly Phonetically-conditioned DistributionInthedialectsofArabic,theafricationofkisnotonlvaBedouin trait but also well known from some rural sedentarv dialects spoken, e.g., in Bahrain, Soukhne and Central Palestinian villages. Tese dif-fer, however, from Bedouin dialects in that the africation in them is notphoneticallvconditioned,butprobablvattributabletoapush chain development due to the fronting of q (q > k; k > ) (Behn-stedt1994:7-11;Holes2006:242;Palva1995:181-187).InIordanas wellasintheWestBank,thelarge-scalemigrationandtherapid uiixxiv.iv. 26urbanizationprocessalongwiththeresultingdialectlevellinghave led to a substantial regression in the use of the k-africation, which in urban environments is not onlv a non-prestigious feature, but is even suppressedasastigma(Abdel-Iawad1981:176-177;301-302).Con-trari wise, in MB similar development has not taken place. Before the glt vs. qltu split in Baghdad, the non-africated variant k was used indiscriminatelvbvMuslims,IewsandChristians.Teadoptionof thek/variationbvMuslimspeakersrendereditasociolinguistic marker which diferentiated the prestigious MB from the non-presti-gious IB and CB minoritv dialects. Te function of the refexes of q andkassocialgroupmarkersthusseemstohaveledtoanasvm-metrvbetweenthephonemicallvandphoneticallvparallelcasesin the phonemic svstem of MB.8 However, this explanation is rendered uncertain when compared with the situation prevailing in Dr ez-Zr, wherediferencesbetweenreligiousgroupsarenotinvolved.Tere the lexicallv restricted transition of q to g is phoneticallv unconditio-ned, and g is not africated, whereas the africation of k in a number of items is alwavs phoneticallv conditioned (Iastrow 1978:42-43; this applies to na as well, Blanc 1964:26-27). 3.3.Use of C- and V- as the Sumxed Personal Pronoun for the 2nd p. sing. fem. Instead of the qltu-Type Forms -ki CB) and C-k, V-ki IB)Te development is illustrated bv the examples ab, bt MB, abki, btk IB, andabki,btki CB (Blanc 1964:65; Abu-Haidar 1991:81; Abu-Haidar 2006a:226; Mansour 2006:236). As far as the africation of k remains phoneticallv conditioned, the contrast btk vs. bt is purelv phonetic, whereas the contrast abki IB, CB vs. ab MB pro-ves that the change from qltu to glt in this case implied one further step, namelv adoption of the phoneticallv unconditioned use of - as a feminine morpheme.9 Te medieval qltu-tvpe MB mav, of course, 8Abdel-Iawad explains the similar asvmmetrv prevailing in Amman bv the spea-kers mixed backgrounds (1981:163-165).9Cf. the development in Amman, where the unconditioned of the Central Pale-stinian rural dialects is a verv stigmatized variant and where the africation of k is generallv avoided.Most resistant to de-africation is the sumxed personal pronoun for the 2nd p. sing. fem. (Abdel-Iawad 1981:279, 282). Tis feature is naturallv con-nected with the established morphological use of the contrast -k (masc.) vs. - (fem.), which implies that the refexes of k in this case are phoneticallv unconditioned.ivom qltu 1o glt 27have had the same uniform -ki form as IB and CB, but the change to the glt forms was not purelv phonetic in that case either.3.4.Application of the Low Vowel Raising RuleWhen OA /a/ in a stressed open svllable is followed bv /a/ in the next svllable, it is as a rulewith manv exceptionsraised to /i/, which is realizedasoru,dependingontheconsonantenvironment,e.g., sma fsh, sma skv, bal mountain; ual onions, gumar moon, kuar to grow up (Blanc 1964:40). Because this rule is operative in verbsaswell,thesedentarv-tvpeperfectbasepatternsCaCaCand CiCiC10 in Form I of the triradical strong verb have been transformed into CiCaC, with a phoneticallv-conditioned allomorph CuCaC: kitab to write, gia to cut, buram to twist; rab to drink, nzal to des-cend, wuga to fall. Interestinglv, this svstematic Bedouin-tvpe pho-netic change did not afect the rather unsvstematicallv distributed q refex of q (2.1.2. above), which seems to indicate that the q refex in this group of items was well established before the low vowel raising rule became operative.3.5. Use of /-t/ as the 1st p. sing. Morpheme in the Perfect, Instead of /-tu/ Typical of the qltu DialectsTis change implied also neutralization of the former morphological contrast 1st p. sing. vs. 2nd p. sing. masc. in the infection of the per-fect.Concomitantlv,theruralandBedouin/-t/morphemeofthe 2ndp.plur.usedintheadjacentareasreplacedthe qltu-tvpemor-pheme /-tum/. 3.6.Use of ni as the 1st p. sing., na as the 1st p. plur. and ntu as the 2nd p. plur. Independent Personal Pronoun Instead of the qltu-Type Forms ana, nna and ntmHere MB follows the patterns commonlv used in the Bedouin dialects oftheSvrian-Mesopotamiangroupandthegltdialectsspokenin 10IBhasonlvonepattern,CaCaC,whereasCBhastwo,CaCaCandCiCiC, which are used also in the qltu dialects spoken in Anatolia, Mosul, Tikrt, Dr ez-Zr, and Alb Kml (Blanc 1964:40; Iohnstone 1975:92; Iastrow 1978:146-157; Behnstedt 1997:264-267; Abu-Haidar 1991:42-43). Tis indicates that also the former qltu-tvpe MB had these two patterns.uiixxiv.iv. 28LowerIraq,Kuwait,andBahrain11(Cantineau1936:70;Iohnstone 1967:91, 104), as a rule difering from the forms used in qltu dialects (Blanc1964:60;Iastrow1978:130-131;Abu-Haidar1991:80;Abu-Haidar 2006a:226; Mansour 2006:236) as well as in the sedentarv dia-lects of Svria and Lebanon (Behnstedt 1997, Map 255). Te progress of bedouinization in the Euphrates Group of the qltu dialects is illu-strated bv the glt-tvpe na in _r ez-Zr (Iastrow 1978:131) and in IB,apparentlvasamajoritvdialectvariantusedsidebvsidewith nna (Mansour 2006:236).3.7.Use of ged with the Imperfect as a Present Continuous MarkerIn addition to the qltu-tvpe verb modifer da-, MB also makes use of the unshortened active participle ged in the same function, although it might be relativelv infrequent (Blanc 1964:115). Tis is an obvious importedruralglt-tvpeform,whichhasbeenonlvphoneticallv bv suppression of africationadapted to the prevalent urban dialect, as is the case in Kuwait and Bahrain as well (Iohnstone 1967:144, 152; Holes 2006:253). In the village of Kwayri at the ruins of Babvlon, the genuine Bedouin-tvpe africation has been attested: ed yiktib (Denz 1971:110,116).TisfeatureiswelldocumentedfromtheSvrian Desert (id) and rn (gid) as well (Behnstedt 1997, Map 161).3.8.Use of the Bedouin-Type Adverb hn here, Instead of the qltu-Type Form hn(i)As reported bv Blanc, qltu dialects as a rule make use of forms of the tvpe going back to OA hhun, whereas its equivalents in glt dia-lects have developed from OA hun or hin (Fischer 1959:115-116; Iastrow 1978:112-113). Blanc also points out that the qltu-tvpe forms hawna and hni are attested for eleventh-centurv Iraqi vernacular and forfourteenth-centurvBaghdad,respectivelv(Blanc1964:139-140; Levin 1994:326). As representatives of the glt tvpe, the Gulf dialects have hn, ihna, and the like (Iohnstone 1967:17, 68; Holes 2006:248; Fischer 1959:117). In Svria both sedentarv and Bedouin dialectslike the Mesopotamian qltu dialectshave forms going back to hhun or hhin, whereas Mad

d

amiyye, besides Dr ez-Zr the onlv qltu-11IntheArabdialectofBahrainthe1stp.sing.is,howeverana/na,Holes 2006:247.ivom qltu 1o glt 29speakinglocalitvinSvria,contrarvtoexpectations,hashini;onlv some villages around the Soukhne oasis have refexes of hin (Behn-stedt 1997, Map 280). 3.9.Use of a ^umber of Core Items Belonging to Bedouin VocabularyTese include, e.g., zn good (cf. IB ml, CB mal); mzn bad (cf. IBdni,CBmmal);bertomorrow(cf.IB,CBada);yamm beside (cf. IB, CB b); na to give (cf. IB, CB aa); ral husband (cf.IB,CBz);hdmclothes(cf.IB,CBws)(Blanc1964:133-159).3.10.Use of the Personal Morphemes t--n, y--n and t--n in the ImperfectTepersonalmorphemesofthe2ndp.fem.sing.,2ndp.plur.and 3rdp.plur.intheimperfect,t--n,y--nandt--n,respecti-velv, are identical with their counterparts in most Bedouin dialects of the Peninsular tvpe. In this context, this feature is not relevant, howe-ver, because MB shares it with the urban IB and CB dialects, e.g., IB tktbn, tktbn, yktbn, CB tktbn, tktbn, yktbn, respectivelv (Blanc 1964:100).4. Tui His1ovic.i Co1ix1Te historical development of MB is inseparablv connected with the historv of Baghdad. For an adequate interpretation of the provenance oftheglt-tvpeMB,thefrstcrucialpointoftimeistheMongol devastation of the citv in 1258 and the subsequent decades. Te kev question concerns the position of the old qltu-tvpe urban dialect of theMuslimsvis-a-vistheBedouin-tvpedialectoftheruralimmi-grants to the citv at that period. Was it likelv that the urban Muslim populationaferthecatastrophestillwasnumerousandinfuential enough to retain its position as a model for linguistic accommodation for Muslim immigrants to Baghdad:ItisacommonlvheldideathatwhenHleghadconquered Baghdadin1258,ageneralmassacrebrokeoutandcontinuedfor fortvdavs,duringwhichthemajoritvoftheinhabitantsperished. Although several almost contemporarv accounts are available, of the uiixxiv.iv. 30medievalMuslimchroniclersnooneactuallvwitnessedtheevents that took place in the citv (al-Faruque 1988:194). In medieval sources thenumberofthosewhowerekilledvariesbetween800,000and 2,000,000fantastic fgures, perhaps given in order to create hatred of Mongols, or just to underline the huge proportions of the catastro-phe.Forarealisticassessmentitisnecessarvtokeepinmindthat alreadv at the end of the eleventh centurv manv quarters in Western Baghdadwereruinedanddeserted,andin1184,whenIbnIubavr visited the citv, it was alreadv in decline. Yaquts report from 1226 is bv and large in accordance with Ibn Iubavrs description. According tothesereports,WesternBaghdadwasaseriesofisolatedquarters each with a wall and separated bv waste land of ruins. Te situation grew still worse in the 1250s when several foods ruined parts of the citv (Duri 1960:902). In these circumstances the number of the vic-timsofthe1258massacremusthavebeenessentiallvsmallerthan reported;accordingtoDuriitprobablvexceededahundredthou-sand.TisisinlinewiththeChinesetravellerChangTesreport (1249)obviouslv from Mongol sourcesthat several tens of thou-sands were killed (ibid.) as well as with Hittis estimate: Te indiscri-minateone-weekslaughter,chroniclersclaim,lef700,000victims, which mav have been 70,000, including refugees from the environs. (Hitti1973:106)However,thegapbetweendiferentfguresisstill huge: al-Faruque (1988:199) estimates that 800,000 of the total popu-lation of two million were killed.12 AccordingtotheanonvmousworkMaridal-Iil,nothingre -mainedofWesternBaghdadexceptisolatedquartersofwhichthe writermentionssix,amongthemal-Karkh,whichwasthemost populatedarea.AstoEastBaghdad,theauthorofMaridaddsan interestinginformation:WhentheTartarscame,mostof[East Baghdad]