Jonathan Owens - Handbook of Arabic Linguistics.pdf

The Oxford Handbooks in Linguistics The Oxford Handbook of Arabic Linguistics Edited by Jonathan Owens Print Publication Date: Sep 2013 Subject: Linguistics Online Publication Date: Dec 2013

The Oxford Handbooks in Linguistics The Oxford Handbook of Applied Linguistics Second edition Edited by Robert B. Kaplan The Oxford Handbook of Case Edited by Andrej Malchukov and Andrew Spencer The Oxford Handbook of Cognitive Linguistics Edited by Dirk Geeraerts and Hubert Cuyckens The Oxford Handbook of Comparative Syntax Edited by Gugliemo Cinque and Richard S. Kayne The Oxford Handbook of Compounding Edited by Rochelle Lieber and Pavol Štekauer The Oxford Handbook of Computational Linguistics Edited by Ruslan Mitkov The Oxford Handbook of Compositionality Edited by Markus Werning, Edouard Machery, and Wolfram Hinzen The Oxford Handbook of Field Linguistics Edited by Nicholas Thieberger

The Oxford Handbooks in Linguistics The Oxford Handbook of Grammaticalization Edited by Heiko Narrog and Bernd Heine The Oxford Handbook of the History of Linguistics Edited by Keith Allan The Oxford Handbook of Japanese Linguistics Edited by Shigeru Miyagawa and Mamoru Saito The Oxford Handbook of Laboratory Phonology Edited by Abigail C. Cohn, Cécile Fougeron, and Marie Hoffman The Oxford Handbook of Language Evolution Edited by Maggie Tallerman and Kathleen Gibson The Oxford Handbook of Language and Law Edited by Peter Tiersma and Lawrence M. Solan The Oxford Handbook of Linguistic Analysis Edited by Bernd Heine and Heiko Narrog The Oxford Handbook of Linguistic Interfaces Edited by Gillian Ramchand and Charles Reiss The Oxford Handbook of Linguistic Minimalism Edited by Cedric Boeckx The Oxford Handbook of Linguistic Typology Edited by Jae Jung Song The Oxford Handbook of Translation Studies Edited by Kirsten Malmkjaer and Kevin Windle The Oxford Handbook of Sociolinguistics Edited by Robert Bayley, Richard Cameron, and Ceil Lucas The Oxford Handbook of Arabic Linguistics Edited by Jonathan Owens The Oxford Handbook of the History Of English

The Oxford Handbooks in Linguistics Edited by Terttu Nevalainen and Elizabeth Closs Traugott

[UNTITLED] The Oxford Handbook of Arabic Linguistics Edited by Jonathan Owens Print Publication Date: Sep 2013 Subject: Linguistics Online Publication Date: Dec 2013

Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide. Oxford New York Auckland Cape Town Dar es Salaam Hong Kong Karachi Kuala Lumpur Madrid Melbourne Mexico City Nairobi New Delhi Shanghai Taipei Toronto With offices in Argentina Austria Brazil Chile Czech Republic France Greece Guatemala Hungary Italy Japan Poland Portugal Singapore South Korea Switzerland Thailand Turkey Ukraine Vietnam Oxford is a registered trademark of Oxford University Press in the UK and certain other countries. Published in the United States of America by Oxford University Press 198 Madison Avenue, New York, NY 10016 © Oxford University Press 2013 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, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, by license, or under terms agreed with the appropriate reproduction rights organization. Inquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above. You must not circulate this work in any other form and you must impose this same condition on any acquirer. Library of Congress Cataloging-in-Publication Data

[UNTITLED] Owens, Jonathan. The Oxford handbook of Arabic linguistics / Jonathan Owens. p. cm.—(Oxford Handbooks) Includes bibliographical references and index. ISBN 978–0–19–976413–6 (alk. paper) 1. Arabic language—Grammar. I. Title. II. Title: Handbook of Arabic linguistics. PJ6106.O96 2013 492.7’5—dc23 2012027454 1 3 5 7 9 8 6 4 2 Printed in the United States of America on acid-free paper

Dedication The Oxford Handbook of Arabic Linguistics Edited by Jonathan Owens Print Publication Date: Sep 2013 Subject: Linguistics Online Publication Date: Dec 2013

Dedication To the once and future study of the Arabic language

Contributors The Oxford Handbook of Arabic Linguistics Edited by Jonathan Owens Print Publication Date: Sep 2013 Subject: Linguistics Online Publication Date: Dec 2013

Contributors Enam Al-Wer Essex University

Elabbas Benmamoun University of Illinois

Ramzi Baalbaki American University of Beirut

Peter Behnstedt PD Hamburg University

Abdelali Bentahila Abdelmalek Essaadi University, Tangier

Sami Boudelaa United Arab Emirates University

Tim Buckwalter University of Maryland

Lina Choueiri American University of Beirut

Peter T. Daniels Independent Scholar

Contributors

Eirlys Davies Abdelmalek Essaadi University, Tangier

Everhard Ditters Nijmegen (Radboud) University

Lutz Edzard Oslo University

Mohamed Embarki University of Besançon

Sam Hellmuth University of York

Clive Holes Oxford University

Maarten Kossmann Leiden University

Pierre Larcher Université de Provence

Stefano Manfredi University of Naples “L’Orientale” and University of Turin

Daniel L. Newman University of Durham

Jonathan Owens Bayreuth University

Dilworth B. Parkinson Brigham Young University

Robert Ratcliffe Tokyo University of Foreign Studies

Jan Retsö University of Göteburg

Contributors

Karin Christina Ryding Georgetown University

Solomon I. Sara Georgetown University

Yasir Suleiman Cambridge University

Mauro Tosco University of Turin

Manfred Woidich University of Amsterdam

Abbreviations1 The Oxford Handbook of Arabic Linguistics Edited by Jonathan Owens Print Publication Date: Sep 2013 Subject: Linguistics Online Publication Date: Dec 2013

Abbreviations1 AH/CE, e.g., Ibn Sῑda (458/1066) = Islamic calendar = AH 458/Gregorian calendar = CE 1066, (*X) token not correct with material in brackets *(X) token not correct without material in brackets 〈〉 encloses root consonants or specified phonological element 〈〉 orthographic unit e.g. [Larcher, “ALT II”] Refer to the article “Arabic Linguistic Tradition II” written by Pierre Larcher in this volume. 1

first person

second person

third person

4M model of codeswitching

ACC

accusative

ADV

adverb

AGT

Arabic grammatical tradition

Abbreviations1

ALT

Arabic linguistic tradition

ANT

active participle

adjective phrase

ATR

advanced tongue root

AUX

auxiliary

consonant

Classical Arabic

computer linguistics

CLLD

clitic-left dislocation

CMC

computer-mediated communication

CMPL

completive

CollA

colloquial Arabic

COMP

complementizer

CONJ

conjunction

COP

copula

complementizer phrase

codeswitching

construct state

consonant vowel

determiner

Abbreviations1

DEF

definite

determiner phrase

dependency grammar

in syntax, indicates structurally relevant gap

EEG

electroencephalography

EgA

Egyptian Arabic

elative

embedded language

EMA

electromagnetic midsagittal articulometry

ENS

educated native speaker

EXS

existential

F1 (etc.)

first formant

feminine

functional projection

future

fMRI

functional magnetic resonance imaging

Fr.

French

GEN

genitive

GPA

Gulf Pidgin Arabic

HAB

habitual

Abbreviations1

Hz.

Herz

Iraqi Arabic

immediate constituent

IDF

indefinite

IMPF

imperfect

INDC

indicative

inflectional phrase

IPA

International Phonetic Alphabet

IRR

irrealis

Jordanian Arabic

JSV

jussive

native language

second language

Lebanese Arabic

LAD

language acquisition device

masculine

Moroccan Arabic

MEG

magnetoencephalography

matrix language

MLD

Moroccan-flavored Dutch

MLF

Matrix Language Frame model

Abbreviations1

ms.

milleseconds

noun

Nigerian Arabic

NEG

negative

NegP

negation phrase

NLP

natural language processing

NOM

nominative

noun phrase

Num

numeral

NumP

number phrase

Old Arabic

OBJ

object

optimality theory

phrase

Palestinian Arabic

P/C

Pidgin/Creole

PET

positron emission tomography

plural

preposition

prepositional phrase

PRE

preformative vowel of imperfect

Abbreviations1

PRED

predicate

PRES

present

PROG

progressive

Pron

pronoun

PRT

participle

PSSD

possessed

PSSR

possessor

PST

past

PSV

passive

processability theory

quantifier

quantifier phrase

R&P

root and pattern (theory)

REL

relative

resumptive pronoun

Standard Arabic (also known as Modern Standard Arabic)

SBJ

subjunctive

singular

SILL

strategy inventory for language learning

SLA

second-language acquisition

Spec

specifier

Abbreviations1

SQUID

superconducting quantum inference device

SUBJ

subject

SVO

subject–verb–object

tense

TAD

traditional Arabic dialectology

TAFL

teaching of Arabic as a foreign language

TJA

Tripolitanian Jewish Arabic (Libya)

target language

TMA

tense, mode aspect marker

TMS

transcranial magnetic stimulation

tense phrase

universal grammar

UPSID

University of California–Los Angeles Phonological Segment Inventory Database

verb

vowel

VOS

verb–object–subject

VOT

voice onset time

verb–subject

WAD

Wortatlas der arabischen Dialekte

WFR

word formation rules

WSA

Western Sudanic Arabic

Abbreviations1

Journals, Book Series, and Organization Abbreviations ACTFL

American Council for the Teaching of Foreign Languages

AIDA

Association Internationale de la dialectologie Arabe

Anthropological Linguistics

BSOAS

Bulletin of the School of Oriental and African Studies

CERES

Centre d’Études et de Recherches Économiques et Sociales

EALL

Encyclopedia of Arabic Language and Linguistics

EI2

Encyclopedia of Islam/Encyclopédie de l’Islam, nouvelle édition

GLECS

Groupe Linguistique d’ Études Chamito-Sémitique

GURT

Georgetown University Round Table on Languages and Linguistics

HSK

Handbücher zur Sprach- und Kommunikationswissenschaft/Handbooks of Linguistics and Communication Science

IFAO

Institut Français d’Archéologie orientale.

ILR

Interagency Language Roundtable

INALCO

Institut National des Langues et Civilisations Orientales

IULC

Indiana University Linguistics Club

JAOS

Journal of the American Oriental Society

MAS Gellas

Matériaux Arabes et Sudarabiques

PAL

Perspectives on Arabic Linguistics (book series, Benjamins)

ZAL

Zeitschrift für arabische Linguistik

ZDMG

Zeitschrift der deutschen morgenländischen Gesellschaft

Abbreviations1

Notes: (1) Because Chapter 9 on Arabic computational linguistics introduces a plethora of abbreviations, which are listed in a separate appendix in that chapter, only key abbreviations from that chapter are included in the current list. In a few cases abbreviations are ambiguous; for example, JA = both Jordanian Arabic and Juba Arabic. The context of the article always provides clear disambiguation in these cases.

Transcription and Transliteration Equivalences The Oxford Handbook of Arabic Linguistics Edited by Jonathan Owens Print Publication Date: Sep 2013 Subject: Linguistics Online Publication Date: Dec 2013

Transcription and Transliteration Equivalences aa = ā ii = ῑ = iy uu = ū = uw x = ẖ = kh ħ = ḥ ʕ = ‘ ʔ = ˒ θ = ṯ = th đ = ḏ = dh ʃ = š = sh Ʒ = ž dƷ = dž = ğ = dj γ = ġ = gh= gh

“Emphatics”

(depending on reflex value) Other emphatic consonants indicated by dot under consonant, for example, ṃ, ṛ, ḷ

Maps The Oxford Handbook of Arabic Linguistics Edited by Jonathan Owens Print Publication Date: Sep 2013 Subject: Linguistics Online Publication Date: Dec 2013

Maps Map 1.1 Countries with Arabic as a majority language. 17 Map 1.2 The Arab league. 18 Map 1.3 Arabic as official language. 19 Map 1.4 Arabic as minority language. 20 Map 22.1 Geographical distribution of Arabic-based Pidgins and Creoles (major centers and areas of distribution). 496

A House of Sound Structure, of Marvelous form and Proportion: An Introduction Jonathan Owens The Oxford Handbook of Arabic Linguistics Edited by Jonathan Owens Print Publication Date: Sep 2013 Online Publication Date: Dec 2013

Subject: Linguistics, Languages by Region DOI: 10.1093/oxfordhb/9780199764136.013.0001

Abstract and Keywords This article begins by defending the so-called Proposition 1, which claims that for the linguist Arabic is the most interesting language in the world. It then describes the handbook’s scope and choice of topics, followed by a discussion of the real world of research on Arabic. It suggests that the study of a language must be more than the sum of its parts. However, as far as Arabic goes, a holistic linguistic tradition remains an unrealized desideratum. A number of factors continue to militate against this development, including the fact that Arabic is a very large language, the stovepiping characteristic of contemporary academia and the clash of academic and cultural traditions. Finally, the article discusses how the only approach that does justice to Proposition 1 is one grounded in radically open-minded empiricism. Keywords: Arabic, language, linguistics, Proposition 1, empiricism

1.1 The Interest of Arabic: Proposition 1 ARGUABLY, for the linguist, Arabic is the most interesting language in the world. I will term this “Proposition 1.” This

claim will certainly strike most as either arrogant or woefully wrong-headed, otiose, and lacking any measurable basis of substantiation. It furthermore runs afoul of deeply embedded beliefs in linguistics itself. In particular is the assumption that all languages are, for purposes of linguistic analysis and insight they give into the universal properties of language, equal. Indeed, on this basis one can agree only that there is no a priori reason to think that the structure of Arabic will tell us more about language than will, say, the structure of Dγweďe, a Central Chadic language spoken by perhaps 40,000 speakers. In terms of its grammatical properties alone, Arabic has no more claim to the attention of linguists than does any other language. To hypothetically formulate a second objection, it might be argued in some circles that Arabic should have a special linguistic place due to being the language of Quranic revelation. While this position may have its partisans among some, it in fact has no inherent connection to its status within linguistics, as indeed was recognized by many of the Classical Arabic grammarians themselves (e.g., Ibn al-Nadim, cited in [Owens, “History”]). A third objection is simply that there is no basis for defining what “interesting” means. This brings us to the defense of the proposition.

1.1.1 The Geographical, Demographic, Chronological, Cultural Gestalt First, and most basically, once one factors away the grammatical, semantic, pragmatic, and formal aspects of languages, it is clear that not all languages are equal.

A House of Sound Structure, of Marvelous form and Proportion This can be measured first of all with simple quantitative standards. There are large languages and small languages. Arabic is one of the world’s largest, spoken natively by about 300 million speakers, and as a second language (L2) by perhaps another 60 million. It is by a large margin the largest language in Africa (nearly 200 million speakers) and one of the biggest in Asia (120 million). It has been estimated to be the fifth largest language in the world in terms of native speakers. Strength of numbers alone guarantees it communicative centrality in the world language system (de Swaan 1998, 2001). Arabic is equally spoken over one of the largest land areas of any native language. It is spoken continuously in the east from Iraq and Khuzistan in southwest Iran, all the way to Morocco and to northeastern Nigeria in the west, an area covering nearly a seventh of the latitudinal distance of the globe. In addition, a number of Arabic-speaking Sprachinseln can be found outside of this area (see Map 1.4 in the Appendix at the end of this chapter). Arabic is furthermore the language of the Quran, Islam being the only one of the large religions whose holy book is revealed in a specific language. Hence, it is learned to one degree or another for religious, ritual, cultural, and legal purposes by nearly all Muslims,1 and equally important, is therefore revered as the purveyor of God’s word. It is the language of the great texts of Arabic–Islamic culture. “Arabic” thus binds the communicative, intellectual, and emotional in one linguistic gestalt, in a way perhaps no other language in the world today does. The history, both written and orally reconstructible, of the Arabic-speaking peoples is, compared with most languages, well documented, even if from the specialist’s perspective gaps in the history are perhaps more prominent than what is available. The first reference to Arabs, which may be inferred to be a reference to Arabic speakers, dates from 853 BCE, North-Arabic clan names are mentioned even earlier (Lipiński 2000: 101, 457), and Arabic begins spreading with great rapidity out of its core Middle East location in the Arabian peninsula, Iraq, and Syrian and Jordanian desert at the beginning of the Islamic era (nominally, 622 CE). By 92/711, relatively large and self-contained groups of Arabic speakers stretch from Uzbekistan in the east to Spain (Andalusia) in the west. A further significant expansion out of Upper Egypt into the Lake Chad area around 800/1400 extends this region. With the exception of Spain, and allowing for modern, “global” diasporas, this essentially defines the limits of the Arabic-speaking world until today (see Owens 2009, chapter 1, for broader summary). The linguistic consequences and challenges of this geo-history are self-evident. While Arabic has even in preIslamic times always been dialectally diverse (Rabin 1951), this diversity has probably increased in the wake of the great Arab–Islamic expansion. If till today simple models for classifying Arabic dialects elude us [Behnstedt and Woidich, “Dialectology”], it is no doubt in large part because an originally diverse proto-situation has continued to diversify across the vast geographical region where Arabic is spoken. Hand in hand with cataloguing the dialectal diversity goes the challenge of developing an historical linguistic model that accounts for the present-day situation. If, as argued in this volume [Owens, “History”], traditional accounts of Arabic language history have generally failed to provide linguistically adequate models of historical development, work on a comprehensive account is largely in its incipient stages. Not surprisingly, in its expansion across a seventh of the earth’s latitudinal distance, speakers of Arabic have come into contact with a large number of languages. The degree to which spoken Arabic itself has been globally affected by this contact is a matter of ongoing debate, with some scholars, such as Versteegh (particularly 1984), arguing that the effects have been profound, whereas others, including Kossmann [“Borrowing”], see Arabic often as the dominant, hence imposing, language in contact situations. Certainly the latter perspective receives support from those well- or fairly well-documented extreme situations where unquestionably, or arguably, new varieties arise from the contact. One of these concerns the emergence of Pidgin and Creole varieties in the Sudanic region and East Africa, varieties that emerged from a common ancestor in the 19th century, today variously known as Turku, Juba Arabic, and Nubi or Kinubi. Since Versteegh’s (1984) argument that the structure of Arabic dialects is to be accounted for by having passed through a stage of Pidginization, a counterconsensus ([Tosco and Manfredi, “Creoles”]) has developed that these Pidgin/Creole varieties are indeed entirely new languages, following the classical model of creolization, with little implication for understanding mainstream historical developments of contemporary Arabic dialects. Relatively underdebated are Uzbekistan and Afghanistan Arabic, spoken by very small populations. Whereas these varieties have classic features of Arabic verbal morphological structure, in other areas of grammar they display marked deviations from any other variety of Arabic, for instance, in having a fixed subject–object–verb (SOV) word order. All deviations are readily explicable as influence from the Dari, Tajik, and Uzbek adstrates, and therefore the question can be raised as to whether these varieties are typologically mixed

A House of Sound Structure, of Marvelous form and Proportion languages ([Tosco and Manfredi, “Creoles”]). Before adducing more evidence in favor of Proposition 1, it is relevant here to take stock of the argument to this point. Beginning from older, classical perspectives on language, issues in Arabic dialectology and language history are multifarious, the challenge of building a comprehensive descriptive database remains high, and questions of language contact all along the vast geographical expanse of Arabic are open. Each of these domains represents a significant linguistic challenge, certainly descriptively but also methodologically and theoretically: what is the role of contemporary dialects in reconstructing language history; what determines direction of influence (van Coetsam 2000); what domains of language are more liable to contact influence; why do ostensibly similar global social conditions among communities of Arabic speakers lead to radically different linguistic outcomes (Owens 2000: 23); indeed, does a definable construct “Arabic” exist [Retsö, “Arabic?”]. But matters become even more interesting linguistically when the two peripheral varieties, Juba Arabic/Nubi and Uzbekistan Arabic, are added. Arabic is the only language in the world from which have emerged both Creole varieties and, arguably, mixed-language varieties. Arabic thus provides a living model for linguistics as a whole to address classic questions of historical and contact linguistics: what happens structurally to a language in the case of normal transmission (in general, the end product of the contemporary dialects) versus, by way of comparison, extreme situations of sociopolitical upheaval or cases of intense contact in a minority situation (Thomason and Kaufman 1988). Interim positions along the continuum formed by these poles can be integrated into linguistic typologies (e.g., Maltese, Kormakiti Arabic in Cyprus, Anatolian Arabic). Certainly, in the domains of phonology and morphology and also to some degree syntax, rigorous measures of core (necessary, not sufficient) Arabic could be constructed. Lurking in the background is the question of how inferences can be drawn from today’s situations to interpret issues of Arabic historical linguistics and how, proceeding from contemporary sociolinguistics methodologies, determining factors in such developments can be extrapolated.

1.1.2 The Classical Language, the Linguistic Tradition The factors summarized in the previous section alone are of enticing interest to linguistics, without mention even having been made of what is unquestionably the most central icon of Arabic: the classical language. It is remarkable that what today is for some the form of Arabic—the ʕ Arabiyya, or the Fuṣħaa, popularly known as Standard or Modern Standard Arabic—is by and large identical to the form of Arabic broadly described by the late 2nd-/8th-century grammarian Sibawaih. The functions of the ʕArabiyya are legion. Most centrally, it is, roughly, the variety of Quranic revelation. It is the variety that came to symbolize the remarkable intellectual and cultural flowering in the Islamic era and the variety around which the Arabic script developed [Daniels, “Writing”]. It is the variety that became a central cultural and political pillar of the Arabic nahḍa “renaissance” movement of the 19th century ([Newman, “Nahḍa”]) and enjoys the status of official language in 23 nation states today (see Map 1.3 in the chapter Appendix) with its concomitant importance in modern educational systems, it is the variety typically taught in non-Arab universities [Ryding, “Acquisition”], and it continues to be an essential element in any debate on Arab identity [Suleiman, “Folk Linguistics”]. Each and every one of these associations implies linguistic issues of different types: descriptive, historical, political, second-language acquisition. What is most remarkable, however, is the Arabic linguistic tradition itself, which was built on the basis of one of the true classics of linguistics, the Kitaab of Sibawaih (Baalbaki 2008; [“ALT I”]). The very first book on Arabic grammar (so far as our documented record of transmission goes) is a comprehensive (nearly 1,000 pages) descriptive work built on a highly elaborated grammatical theory. While opinions differ as to the origin of the post-Sibawaih Arabic linguistic tradition, it is clear that a highly sophisticated and differentiated theoretical grammatical and pragmatic discourse continued to develop for at least the next 500 years [Larcher “ALT II”]. No less interesting and significant is the voluminous lexicographical tradition that developed in tandem with the grammatical [Sara, “Classical Lexicography”]. Students of Arabic therefore deal not only with the varieties of Arabic themselves but also with a metadiscourse, as it were, which was established within Arabic–Islamic culture. Arabic texts were passed down to us, along with a theoretical framework for analyzing them, constitutive of the Arabic–Islamic tradition, which continues to be of central importance in the contemporary teaching of Arabic and which challenges the interpretive acumen of linguists studying this tradition.

A House of Sound Structure, of Marvelous form and Proportion Thus, with respect to Proposition 1, it is not only that Arabic is one of the few languages of the world within which developed a linguistic tradition; also, it is a tradition that continues to exercise its influence on today’s Arabs and Arabic society and beyond to Islamic society.

1.1.3 Arabic and Arab Identities The two previous points set the stage for the inherent language tension that exists in contemporary Arabic societies. Arabic, the mother tongue of its approximately 300 million speakers, is not the same Arabic as the Arabic that is codified and has official political status and cultural centrality through its association with the Quran and with pan-Arab identity. On one hand, these two broadly defined varieties can be represented as mutually opposed: official versus unofficial, written versus spoken, formal versus informal, pan- versus local, learned formally versus acquired as a first language (L1). The functional contrasts were made famous by Ferguson (1959). Equally, one can emphasize the complementarity of the codes. The native colloquial is the language not only of home and friends but also of all that is informal, unofficial, spontaneous, and intimate. The growing entertainment industry in its diverse media manifestations is thus wholly dominated by the colloquials, as is the informal world of texting and twittering [Holes, “Orality”]. Blogging, a domain awaiting comprehensive linguistic research, appears to cover a spectrum of styles. The difference between the two is also one of ideology versus practice, of ideal versus real. The fuṣħaa, even if in its perceptions and usage it is a variety of fuzzy contours (Kaye 1972; Parkinson 1991) and is rarely2 used in the real world in its prescribed form, is the variety of preeminent cultural importance [Suleiman, “Folk Linguistics”]. Sociolinguistics, a subdiscipline of linguistics of relatively recent provenance closely related to the older dialectology, shows the degree to which ideal and real can differ in the realm of spoken Arabic. The careful microdocumentation of speech communities consistently has shown (studies from the Arabian–Persian Gulf, Saudi Arabia, Jordan, Damascus, Bethlehem, Cairo, Casablanca, and northeast Nigeria) that features of spoken colloquial varieties are what drive language change [Al-Wer, “Sociolinguistics”]. Moreover, when Arabic meets other languages bilingually, it is again the colloquial that always forms the basic matrix of contact [Davies et al., “Codeswitching”]. Even in mixed colloquial–fuṣħaa exchanges such as on media talk shows, the colloquial can have a dominant role. The vibrant co-existence of quite differentiated varieties, a situation hardly unique to Arabic, nonetheless takes on a special, perhaps unique status in the world’s languages, precisely because each variety, beyond its linguistic profile, embodies a different history, a different symbolism, a different legitimization. While these differences are of central interest to students of linguistics, they extend beyond the academic lecture hall to the real world of language teaching and language policy. To which variety, for instance, should a program of second-language teaching be tailored, or, if the varieties have different cognitive profiles, what are the implications for L1 teaching? These are questions best not answered by policy fiat. Indeed, the experience of Arabic in post-9/11 America represents probably the sorriest example ever of huge resources expended for developing language teaching programs, largely divorced from the fundamental research on the language being taught that would make for a more rational and efficient teaching program [Ryding, “Acquisition”]. Research from across the spectrum of linguistics is implicated in any academicization of Arabic teaching, whether as an L1 or L2.

1.1.4 Grammar Arabic is thus a language of rare breadth and extension in the world, a language like perhaps no other in the degree to which it embodies the culture and politics of its speakers. It is, however, a language, and it has been studied from a number of classical grammatical perspectives. Even here Arabic has structural features that set it apart from many, sometimes most, of the world’s languages. The phenomenon of emphasis (pharyngealization) of consonants is a hallmark of the language and has engendered numerous studies both in phonetics [Embarki, “Phonetics”] and in phonology [Hellmuth, “Phonology”]. What is emblematic of Arabic, however, hardly exhausts the interest of Arabic for linguistics. As Hellmuth points out, for instance, stress in Arabic has been of central interest in phonological theory.

A House of Sound Structure, of Marvelous form and Proportion In morphology, an ongoing debate surrounding Arabic and many other Semitic languages is the status of the consonantal root as a morphemic element. As Ratcliffe [“Morphology”] points out, the Arabic grammatical tradition itself viewed the stem, not the root, as the basis of morphology, and arguments from within contemporary morphological theory have been developed for this as well. But equally, psycholinguistic studies on the basis of carefully constructed experiments have interpreted the consonantal root as having a crucial role in morphological processing [Boudelaa, “Psycholinguistics”]. Besides the Arabic grammatical tradition itself (1.1.2), there are two further prominent approaches to Arabic grammar. The older one is the philological tradition [Edzard, “Philology”], with which the study of Arabic grammar in the West began. Besides its general interest in Arabic grammar, this tradition incorporates cultural issues and has been present at the interface of Arabic texts of all genres and language. The other is more recent and is based on the precepts of theoretical grammar, particularly syntactic theory in the generative tradition, which endeavors to locate what is specifically Arabic within a broader program of universal grammar [Benmamoun and Choueri, “Syntax”].3 All of the formal grammatical domains feed into the growing domain of computational linguistics and into the broader field of natural language processing [Ditters, “Computational”]. Finally, the classical lexicographical tradition has its counterpart in contemporary lexicography, a field increasingly drawing the vast online publishing industry in Arabic for its sources [Buckwalter and Parkinson, “Modern Lexicography”]. Here again one experiences the special challenges confronting the Arabic lexicographer, for instance, whether to lemmatize according to root or stem, how to sublemmatize parts of speech, and whether to lump polysemously or to differentiate identical forms. The articles in this handbook describe a language that, when looked at in its totality, is of rare thematic linguistic differentiation.

1.2 Scope and Choice of Chapter Topics Proposition 1 encapsulates an ideal. The handbook is intended to reflect the full breadth of research on Arabic linguistics in the West. Realistically, this implies that it includes only chapters on topics judged to have a critical mass of background research. The reader will therefore miss domains that might be expected in a linguistics handbook. Asymmetries will be noticeable. There is a chapter on L2 acquisition but none on L1 acquisition, a chapter on sociolinguistics, but none on oral discourse, a number of chapters on grammar but none on semantics. The gaps are regrettable but unavoidable so long as the focus of the chapters is on the domains of Arabic linguistics that do indeed enjoy a fairly broad and deep coverage rather than on Arabic-flavored general linguistics, as it were. 4 The chapters themselves reflect domains of research with great disparities of detail. In some cases the chapter is able to cover nearly all of the published research on a given domain, for instance, the chapter on Pidgins and Creoles and even, surprisingly (see remark at end of 1.1.3), work on L2 Arabic language acquisition. In others the breadth of available material has meant that authors could summarize only broad lines of research, illustrating the topic in greater detail with selected examples. Arabic language contact, particularly as reflected in loanwords, for instance, has a very large literature; the research on Arabic dialects is immense, and the research on the Arabic grammatical tradition is large. As far as Western research goes, these disparities to some degree reflect the relative age of the subdomain. In general, codeswitching, psycholinguistics, sociolinguistics, and pidgin and creole linguistics, for instance, are barely 30 or 40 years old as independent specializations of linguistics. Research on Arabic dialects, on the other hand, was already well established in the 19th century. This does not, however, imply that any domain of Arabic linguistics has been exhaustively treated. As Behnstedt and Woidich point out [“Dialectology”], many dialects, for instance, are poorly described, and the integration of dialectology and sociolinguistics, an essential element of sociolinguistics in the West, has seen only modest progress in the case of Arabic, while historical dialectology, a part of the general field of Arabic historical linguistics, is meager at best. Gaps should certainly be seen as a challenge to open up wider avenues of research.

1.3 The Real World of Research on Arabic: a Critical Look

A House of Sound Structure, of Marvelous form and Proportion Given the current state of research on Arabicist it may be asked: if Proposition 1 is correct, does the linguistic research match the inherent interest of the language? Here I would answer with only a very conditional “yes.” On one hand, as noted in the previous section, there are areas of research with a large literature and well-established research tradition. On the other hand, there are topics central to the study of any language with only modest research traditions in Arabic. Studies on spoken Arabic discourse are rare (see note 4), while more recent domains of linguistics such as psycho-linguistics, sociolinguistics, or the study of spoken Arabic pragmatics, though growing, are still in their incipient stages. Ultimately, however, the study of a language must be more than the sum of its parts. It will be suggested here that, as far as Arabic goes, a holistic linguistic tradition remains an as yet unrealized desideratum. In the past and currently, a number of factors militate against this development. Four factors can be identified.

1.3.1 Arabic Is Large The first is simply the immensity of the field itself. Arabic presents prima facie anything but a unified domain of inquiry. Consider, for instance, the two basic media that Arabic linguistics works with: the written and the spoken word, the former of which is associated with the Classical and Standard language and the latter with the dialects. These two media are in important respects of a different nature. The written domain is a learned domain, one that itself continues a heritage dating back to the 2nd/8th century, whose standard and norms have been long established. While one might be able to change certain aspects of the Standard language, such as the idiomatic domain ([Newman, Kossmann]), one cannot change its morphology or syntax. The spoken domain, on the other hand, is beholden to contemporary methods of descriptive and field linguistics, associated with, inter alia, corpus collection and language documentation, work with expert consultants, and instrumental phonetics of the spoken language. Norms, such as there are in this domain, emerge from the individual research studies undertaken in it. Experience, moreover, has shown that in the Western tradition these two domains exist largely in parallel universes, with scholars linked to one or the other but not both. Those concerned with the written language, for instance, to the extent that they move outside the field of the linguistics of the written varieties, gravitate toward the other literary domains of Arabic such as Arabic literature, law, and medical texts. Many such individual cases could be cited, but quite typical in this respect is Carl Brockelmann, whose Grundriss der semitischen Sprachen (1908, 1913) remains a standard reference work. After publishing this work, he went on to write another well-regarded book, Geschichte der islamischen Völker (1943) (History of Islamic Peoples). Brockelmann never studied a spoken variety of Arabic, and his Grundriss, while a work of compendious scholarship, is marked by a decided antipathy toward theoretical issues in historical and contact linguistics (Owens 2009: 43), precisely two areas where Arabic is particularly implicated, as discussed already.5 Those working in the realm of the spoken language, on the other hand, are faced initially with a plethora of challenges, for instance, which aspect of language to concentrate on or which varieties of Arabic to try to delineate. Finding a format to integrate these in turn with the Classical or Standard varieties may imply defining variables that are central to neither tradition. Edzard [“Philology”] correctly notes that there is in principle no contradiction between a philological (written) orientation and a “theoretical” linguistic one; experience has nonetheless shown that relatively few scholars not only work in both domains but also, more importantly, attempt a synthesis of the two.

1.3.2 Stovepiping The problem is at once abetted and exacerbated by the stovepiping characteristic of contemporary academia. Whereas 30 years ago one could claim to be a linguist, today it is more likely that one will be a sociolinguist, psycholinguist, or general or specialized syntactician. Certainly these developments follow their own internal logic, as methods and theoretical perspectives have become more specialized during this period. At the same time, in this there is the danger that the academic apparatus defines the language rather than the language being served by the apparatus. To take an example from sociolinguistics, one can ask how many studies are needed to define the social status of

A House of Sound Structure, of Marvelous form and Proportion the “qaf” variable. On one hand, the fact that there have been fruitful studies on this variable means that it provides a necessary and interesting comparative breadth; on the other hand, certainly many other variables, some of broad comparative potential and others of particular local interest, await treatment. Added to this, embedding the findings on a comparative basis in the vast Arabic world is a challenge that has received relatively little attention from Arabic sociolinguists.6 Beyond this is the ever-present danger of calling the game over as soon as a sociolinguistic phenomenon has been studied from within a particular theoretical perspective, as often as not one initially defined from outside of the Arabic-speaking world. Al-Wer’s perspective in [“Sociolinguistics”] is better; she shows that ultimately constructs need to be interpreted within a context that does justice to the particularities of a given part of the Arabic world, illustrating her point with the interpretation of the ostensibly universal or at least very general “education” variable as a proxy for other, community-immanent variables.

1.3.3 Clash of Traditions Complementing the two previously defined issues is that academic and cultural traditions provide ready barriers for synthetic perspectives. Within the West, for instance, Carter (1988: 207) attempts to dissociate Arabic linguistics from Arab linguistics. “… ‘Arabic linguistics’.… detaches the language entirely from its environment so that it becomes a pure abstraction.” On the other hand, Arab linguistics, the legitimate study of the Arabic language, is “… the vast and continuing output of traditional works, both editions of texts and secondary sources, which remain wholly within the historical norms of Islamic scholarship” (ibid.). In Carter’s terms, a handbook of Arabic linguistics that has at its core questions about the Arabic language, however defined, is suspect from the start. To be fair, one of Carter’s objections to an Arabic linguistics deserves attention. “Solving” a problem in Arabic within a general linguistic theory runs the danger of importing an issue, a technique of inquiry, a focus on a grammatical construction whose ultimate interest is dictated from outside of Arabic and whose “solution” offers little to those interested in the complex structure of Arabic. At the same time, however, as noted already, trivial an observation though it is, Arabic is simply a language, so linguistic approaches will want to understand it within general theories of language. Moreover, as argued in Sections 1.1.1 and 1.1.2, Arabic itself has unique geographical, social, historical, and cultural properties that have, as it were, pushed the language in directions hardly encountered elsewhere. Linguistic theory can hardly avoid it, even if, in practice, non-Arabicist linguists often do so (see, e.g., criticisms in [Tosco and Manfredi, “Creoles”] or Ryding [“Acquisition”] on the barriers confronting researchers of second-language acquisition due simply to lack of language knowledge). It is easy to formulate a solution to this problem: practitioners need to be as well versed in Arabic in all its linguistic ramifications as they are in the methodologies and theories of linguistics. Nonetheless, its implementation implies a commitment of both individual and institutional time and intellectual resources, which are not necessarily easy to come by. Perhaps more pernicious than the delegitimization of a linguistic approach to Arabic is Mahdi’s (1984: 37) admonition to study dialects to be rid of its debilitating influence on the Standard (fuṣħaa).7 This perhaps wellintentioned perspective derives most directly from a normative 19th-century tradition (see [Newman, “Nahḍa”], which attempts to lay the blame for the ill learning of the Standard language on the use of dialects and can justify the study of dialects only against a possible benefit for the Standard. Such a perspective is not uncommon in the Arabic world.8 Leaving aside the cultural and political issues inherent in this position [Suleiman, “Folk Linguistics”], adopting this perspective would necessarily mean excluding Chapters 10, 12, 13, 14, 15, and 22 from this volume while requiring severe reductions in most others, since the dialect is nothing less than the mother tongue. It is not so much an approach foreign to general linguistic inquiry as it is a rejection of the scientific and empirical study of the world, defining in narrow political-cultural terms the goals of research on one of the most ineffable and undefined domains of human experience: language.

1.4 Attitudes The reader may be confused at this point. On one hand, Proposition 1 claims that Arabic is, for the linguist, an intellectual challenge like no other. On the other, this challenge is often met by traditions, theories, academic structures, and attitudes that at best ensure a fragmented understanding of the language and at worst succeed in a holistic characterization of “Arabic” only at the expense of defining whole domains of language experience into nonexistence.

A House of Sound Structure, of Marvelous form and Proportion It can be suggested, without exaggerating the professional and even ideological differences that accrue in the study of Arabic, that the only approach that does justice to Proposition 1 is one grounded on radically open-minded empiricism. For precedence, one need go no further than the medieval Arabic grammatical tradition itself. The following quote from the mid-4th-/10th-century Zajjaji in one of the earliest works of metareflection in the Arabic tradition. In Chapter 5 (al-baab al-xaamis), he reflects on the nature of linguistic causes. After identifying three types of linguistic causes (pedagogical, analogical, theoretical-speculative; see Versteegh 1995: 89), he approvingly summarizes the approach to language study attributed to al-Xalil ibn Aħmad, the polymath contemporary of and teacher of Sibawaih (see [Sara, “Classical Lexicography”]). In the passage, Xalil is said to have likened the scholar trying to ascertain the nature of (Arabic/language) to one trying to understand a house construction:

If a wise person were to enter a house of sound structure, of marvelous form and proportion, whose builder’s wisdom appeared correct to him according to reliable information, the unmistakable lines of proof and clear arguments. And each time the man stopped and pondered a part of the house, he said, “[the builder] did it this way for such and such a reason and such and such a cause.” That is what occurred to him and appeared reasonable to him. Now it might be that the builder did build it for the reason the man inspecting the house thought, but it is equally possible he did it for another reason, even if the inspector’s reason might be correct. If a different grammatical reason should occur to another person than myself, which is more appropriate than my explanation, so let him suggest it. (p. 66) With this passage, there are obviously interpretative issues that go beyond an introductory chapter. In particular, the passage is enticingly ambiguous as to what a “more appropriate” explanation might be. The history of the Arabic tradition itself shows that an explanation in the 5th/11th century might be more nuanced than one in the 3rd/9th and that one in the 6th/12th century might add further elements [Larcher, “ALT II”], not to mention the classic grammar-internal differences of the Basrans and Kufans (Sibawaih vs. Farra’; Owens 1990). It would be a grave mistake, however, to stop with the classical tradition. The recent history of linguistics is marked not only by the continual reappraisal of classic linguistic ideas and traditional issues but also by new theoretical, methodological, and, increasingly, technical advances, many described in this volume, that promise to transform, expand, and enrich the very idea of grammatical explanation to such an extent that a genius such as Xalil, if he were alive today, would be envious. Xalil’s metaphor unmistakably sets a basic ground rule for linguistic research, namely, that no possible explanatory aspect be excluded on a priori grounds. Since explanations are, ultimately, explanations of linguistic substance, facts, observations, summaries of data, measurements, and reinterpretations of previous explanations, Xalil’s approach implies setting no preconditions as to what comes under the purview of Arabic linguistics. 9 It is in this spirit that the current handbook should be read; it is a reference work that brings together different approaches and scholarly traditions, an invitation to the reader to explore the multifaceted world of Arabic linguistics. The articles in this volume expertly explore the nature of the house of Arabic from many angles. Many argue for specific points of view, others give descriptions of synoptic breadth, while others provide exhaustive overviews of the state of the art. The parts may or may not come together to describe a common structure; they do provide blueprints for a better understanding of it.

Note to References Chapters 9 “Issues in Arabic Computational Linguistics” and 13 “Dialects and Dialectology” have very comprehensive bibliographies. They are, however, too large to be included in their entirety in the print version of

A House of Sound Structure, of Marvelous form and Proportion the handbook. Rather than edit away this very valuable resource, it was decided to include the complete bibliographies to these two chapters in the online version of the handbook while including a selected bibliography in the print version. This Appendix gives basic background information about Arabic as well as a brief discussion of the transcription and transliteration conventions used in this book.

A.1 Maps The bulk of the native Arabic-speaking population lives within countries with majority Arabic-speaking populations. Sizable non-Arabic minorities include Berbers (Amazigh), with large minorities in Algeria and Libya and up to half the population of Morocco, where in fact Arabic shares its status as official language with Berber. Other minorities are speakers of the various South Arabian languages in Yemen (and a small population in Oman) and Kurds and Aramaic speakers in Iraq and to a lesser degree Syria. Even after the South Sudan, which has few native Arabic speakers, recently split off from the North, the Sudan has a large and diverse linguistic minority population. Finally, Mauretania has a not insignificant non–Arabic-speaking population (Wolof, Fulfulde) in the south of the country. Map 1.1 shows countries with majority Arabic-speaking populations. It can be noted that although the main lingua franca of South Sudan, Juba Arabic, historically derives from Arabic, by linguistic measures it is a different language [Tosco and Manfredi, “Creoles”] and therefore is not included on Map 1.1. Maps 1.2 and 1.3 illustrate the lack of complete isomorphy between political status of a language and the native language of its inhabitants. The Arab League ( , Map 1.2) comes close to being composed entirely of countries with Arabic as a majority language. There are only two exceptions: Somalia, where the native language of the vast majority of the population is Somali, a Lowland East Cushitic language genetically very distantly related to Arabic; and the Comoro Islands, whose native Bantu language is closely related to Swahili. Besides being the official language of all countries in the Arab League, Arabic is also the official language of Eritrea (majority native language Tigrinya; Hailemariam 2002: 75), a country with a tiny population of Arabic native speakers. In addition it is, along with French, an official language in Chad, which does have a sizable native Arabicspeaking minority. In these two countries, Arabic attained official status under quite different circumstances and at different times. In Eritrea, for instance, it was during the brief British rule from 1941 to 1952 that Arabic was introduced as the official language, a status it has maintained until today, whereas in Chad Arabic was adopted as an official language well after independence (1960) in the 1990s, and only after considerable debate (de Pommerol 1997). Finally, Map 1.4 shows that for the most part Arabic-speaking minorities live on the political borders of majority Arabic-speaking countries. Even the exceptions in this regard, the tiny Arabic-speaking populations of Uzbekistan, Afghanistan, and Khorasan in eastern Iran were, at the time of their settlement in the 2nd/8th century, a part of a continuous migration of Arabs into Central Asia. It can be noted that, while from the perspective of genetic linguistics Maltese can be considered a variety of Arabic (Owens 2010), on a sociopolitical basis and as an official language of the European Union it is an independent language.

A.2 Genetic Affiliation of Arabic While a definitive classification of Arabic within a Stammbaum representation may be impossible [Retsö, “Arabic”], within traditional genetic models the following two models are the most widely discussed (based on Faber 1997: 5, 6): (Afro-Asiatic) Cushitic Omotic Chadic Ancient Egyptian

A House of Sound Structure, of Marvelous form and Proportion Berber Semitic: Variant 1 East Semitic: Akkadian, Eblaitic West Semitic Northwest Semitic Canaanite: Hebrew, Phoenician, Moabite Aramaic South Semitic Arabic Southeast Semitic Modern South Arabian: Jibbali, Mehri, Harsŭsi, Soqotri Ethio-Saebean OSA: Sabean, Qatabanian, Hadramauti, Minean Ethiopian Semitic Variant 2 East Semitic Akkadian Eblaite West Semitic Central Semitic Arabic Northwest Semitic Ugaritic Canaanite: Hebrew, Phoenician, Moabite, Ammonite, El-Amarna Aramaic DeirAlla South Semitic Eastern Soqotri Mehri, Harsŭsi, Jibbāli Western

A House of Sound Structure, of Marvelous form and Proportion Old South Arabian Ethiopian Semitic North Ethiopic: Ge’ez, Tigre, Tigrinya South Ethiopia Transverse SE Amharic, Argobba Harari, East Gurage Outer SE n group: Gafat, Soddo, Goggot tt group Muher West Gurage

A.3 Transcription and Transliteration Conventions The representation of Arabic in Latin script is beholden to different conventions. Rather than try to force standardization in this volume, the various systems used are taken over intact in different chapters. Having said this, the editor is strongly biased toward the use of the International Phonetic Alphabet (IPA), or modified IPA symbols, for representing any spoken text. Nothing, moreover, speaks against using it for transliterated written texts, though here other traditions have developed different conventions.

Click to view larger Map 1.1 Countries with Arabic as a majority language.

Click to view larger Map 1.2 The Arab league.

A House of Sound Structure, of Marvelous form and Proportion

Click to view larger Map 1.3 Arabic as official language.

Click to view larger Map 1.4 Arabic as minority language.

Ultimately, moreover, justification can be asked of each set of conventions. For instance, representing a long “i” as [ī, i:, ii, or iy] implies different phonological interpretations of the nature of vowel length. It can be noted that IPA conventions themselves should hardly be regarded as sacrosanct. The multiexponential phenomenon of “emphasis,” for instance, is now represented by C + ʕ, such as t ʕ, that is, C + pharyngealization. As the two relevant articles in this volume make clear, however ([Embarki, “Phonetics”; Hellmuth, “Phonology”]), pharyngealization (tongue retraction toward pharynx, pharyngeal constriction) is but one gesture defining the phenomenon and is not necessarily the most prominent one.10 Equally relevant would be, for instance, a symbol based on the articulatory metaphor developed in the Arabic tradition of likening the flattened tongue body to a plate or pot cover (iṭbaaq, muṭbaq). In any case, the multiplicity of transcription/transliteration conventions means that the reader’s indulgence is needed for the treatment of proper Arabic names, where the same person will appear in difference orthographic guises, according to the conventions of the chapter, Ibn Jinni, Ibn Jinnī, Ibn Ğinnī, Ibn Ğinni. Would that he could comment on the matter.

References Amara, Mohammad. 2005. Language, migration and urbanization: The case of Bethlehem. Linguistics 43: 883–901. Baalbaki, Ramzi. 2008. The legacy of the Kitaab. Leiden: Brill. Brockelmann, Carl. 1908–1913. Grundriss der vergleichenden Grammatik der semitischen Sprachen. 2 vols. Berlin: Reuther and Reichard. ——. 19772 (1943). Geschichte der islamischen Völker. Hildesheim: Olms. Carter, Michael. 1988. Arab linguistics and Arabic linguistics. Zeitschrift für Geschichte der arabisch-islamischen Wissenschaften 4: 205–218. de Pommerol, Patrice. 1997. L’arabe tchadien: émergence d’une langue véhiculaire. Paris: Karthala. de Swaan, Abram. 1998. A political sociology of the world language system (1): The dynamics of language spread.

A House of Sound Structure, of Marvelous form and Proportion Language Problems and Language Planning 22: 63–75. ——. 2001. Words of the world. Cambridge, UK: Polity Press. Faber, Alice. 1997. Genetic subgrouping of the Semitic languages. In The Semitic languages, ed. Robert Hetzron, 3–15. London: Routledge. Faiq, Said. 2006. Coherence. In Encyclopedia of Arabic Language and Linguistics, vol. I, ed. Kees Versteegh, Associate Editors: Mushira Eid, Alaa Elgibali, Manfred Woidich, Andrzej Zaborski, 427–430. Leiden, Netherlands: Brill. Ferguson, Charles. 1959. Diglossia. Word 15: 325–340. Hachimi, Atiqa. 2007. Becoming Casablancan: Fessis in Casablanca as a case study. In Arabic in the city: Issues in language variation and change, ed. Catherine Miller, Enam Al-Wer, Dominique Caubet, and Janet Watson, 97– 122. London: Routledge. Haeri, Niloofar. 1996. The sociolinguistic market in Cairo: Gender, class, and education. London: Kegan Paul International. Hailemariam, Chefna. 2002. Language and education in Eritrea. Amsterdam: Aksant. Holes, Clive. 1987. Language in a modernising Arab state: The case of Bahrain. London: Kegan Paul International. Kaye, Alan. 1972. Remarks on diglossia in Arabic: Well-defined vs. ill-defined. Linguistics 81: 32–48. Khalil, Esam. 2006. Cohesion. In Encyclopedia of Arabic Language and Linguistics, vol. I, ed. Kees Versteegh, Associate Editors: Mushira Eid, Alaa Elgibali, Manfred Woidich, Andrzej Zaborski, 430–433. Leiden, Netherlands: Brill. Lipiński, Edward. 2000. The Aramaeans, their ancient history, culture, religion. Leuven: Peeters. Mahdi, Muhsin (ed.). 1984. Alf Layla wa Layla. Leiden: Brill. Al-Marzuqi, Munsṣif. 2011. ʒ ayy luƔa sayatakallam al- ʕArab al-qarn al-muqbil? Jazira Net, November 6, 2011. Mejdell, Gunvor. 2006. Mixed styles in spoken Arabic in Egypt. Leiden: Brill. Owens, Jonathan. 1990. Early Arabic grammatical theory: Heterogeneity and consolidation. Amsterdam: Benjamins. ——. 1995. Language in the graphics mode: Arabic among the Kanuri of Nigeria. Language Sciences 17: 181–199. ——. 1998. Neighborhood and ancestry: Variation in the spoken Arabic of Maiduguri (Nigeria). Amsterdam: John Benjamins. ——. 2000. Introduction. In Arabic as a minority language, ed. J. Owens, 1–43. Berlin: Mouton de Gruyter. ——. 2009. A linguistic history of Arabic. Oxford: Oxford University Press. ——. 2010. Review article: “What is a language?: Review of Bernard Comrie, Ray Fabri, Elizabeth Hume, Manwel Mifsud, Thomas Stolz & Martine Vanhove (eds.), Introducing Maltese Linguistics. Selected papers from the 1st International Conference on Maltese Linguistics, Bremen, 18–20 October 2007.” Journal of Language Contact (Varia): 103–118. Owens J., and Alaa Elgibali (eds.). 2010. Information structure in spoken Arabic. London: Routledge. Parkinson, Dilworth. 1991. Searching for modern Fusṣḥā: Real life formal Arabic. Al-’Arabiyya 24: 31–64. Procházka, Stephan. 2006. Arabic. In Encyclopedia of language and linguistics, vol. I, ed. Keith Brown, 423–431. Oxford: Elsevier.

A House of Sound Structure, of Marvelous form and Proportion Rabin, Chaim. 1951. Ancient West Arabian. London: Taylor’s University Press. Sallam, A. 1980. Phonological variation in educated spoken Arabic. BSOAS 43: 77–100. Thomason, Sarah, and Terrence Kaufman. 1988. Language contact, Creolization, and genetic linguistics. Berkeley: University of California Press. van Coetsam, Frans. 2000. A general and unified theory of the transmission process in language contact. Heidelberg: Winter. Versteegh, Kees. 1984. Pidginization and Creolization: The case of Arabic. Amsterdam: Benjamins. ——. 1995. The explanation of linguistic causes. Amsterdam: Benjamins. http://arabicorpus.byu.edu/index.php http://en.wikipedia.org/wiki/List_of_languages_by_number_of_native_speakers#30_to_50_million_native_speakers “List of languages by number of native speakers.”

Notes: (1) For native speakers, Procházka’s (2006) estimate of 280 million strikes us as reasonable, if perhaps slightly low. In addition, Arabic is spoken fluently as a second-language lingua franca in particular in Algeria, Morocco, Mauretania, Libya, Yemen, Chad, Tunisia, and the Sudan. An estimate of 452 million “total” speakers, such as found at http://en.wikipedia.org/wiki/List_of_ languages_by_number_of_native_speakers#30_to_50_million_native_speakers, should be treated with great caution. Estimating total number of speakers in a language like Arabic begs the question of what a language speaker is. In a survey carried out among Kanuri, one individual reported to me that she uses Arabic “often” (Owens 1995). When I thereupon addressed her in Arabic, she could not understand a word. She explained that she began many acts with bi sm illaahi (“in the name of God”). Defining “total” (of what?) is no less a slippery task than defining “often.” (2 ) The crucial adverb rarely should be understood as follows. Arabic is spoken by, conservatively, 300 million individuals. Each individual, probably conservatively, speaks for two hours per day, at 10,000 words per hour (slightly low probably), giving 6 trillion words of Arabic per day. The only forums where a normative, spoken Standard Arabic is used are certain media broadcasts (e.g., the excellent news channels al-ʕ Arabiyya or alJaziyra, national and commercial channels mainly for information-orientated topics such as news and documentaries) and in various official meetings, including some but hardly all educational formats (see Mejdell 2006; also [Holes, “Orality”]). Of the 300 million speakers, only a tiny minority of them are engaged at any one time in a function prescribing the use of Standard Arabic. Otherwise, for most individuals nearly always, and for all at some time, the basis of everyday speech is a colloquial variant. (3 ) Chapter 6 is a double chapter; the original intent was to have two separate chapters, one on the standard language and the other on dialects. Individual circumstances required conflating the two into one. (4 ) For instance, the justifiably well-regarded Encyclopedia of Arabic Language and Linguistics has a chapter on “Cohesion” (Khalil 2006) with nine non-Arabic items in the bibliography and ten on Arabic. Unfortunately, this breakdown realistically reflects the dearth of material on spoken Arabic discourse, for instance, only one booklength work, an edited volume (Owens and Elgibali 2010), which is too little in the editor’s view to merit a separate chapter here. The article preceding Khalil’s on “Coherence,” a central topic equally in literary and spoken texts, treats the subject only as it is reflected in the Classical literary tradition (Faiq 2006). The limitation is regrettable but does reflect the unbalanced state of the art in this domain. (5) Indeed, it is striking that while comparative Semitic and comparative Indo-European literature both came of age in the same era, the 19th century, and to a large degree in the same region—Central Europe— the theoretical contribution of the former to the development of general principles of historical linguistics was negligible whereas that of the latter was essential. 6

A House of Sound Structure, of Marvelous form and Proportion (6 ) For instance, despite relatively well-documented accounts of “qaf” variation covering thirty years of research in the Arabic world from the Gulf to Morocco (e.g., Sallam 1980; Holes 1987; Haeri 1996; Amara 2005; Hachimi 2007), no studies have synthesized these accounts with a view toward defining the extent to which a common social dynamic lies behind “qaf” usage. It is, for instance, no sociolinguistic accident that the “qaf” variable is of such marginal interest in Nigerian Arabic, a distinctly minority language in northeast Nigeria, that it was not included as a variable in Owens (1998). (7 ) Mahdi speaks of the sicknesses of the dialects, which require treatment . The passage in fact comes in the Introduction to a well-edited edition of 1001 Nights, which left the original “Middle Arabic” style intact rather than classicizing out its authenticity, as is the current custom (e.g., the version on arabicorpus). Another popular approach is the regulation of language use by legal fiat. Munṣif al-Marzuqi, who writes an occasional column for Jezira Net, for instance, would (article of Nov. 6, 2011) criminalize the use of what he terms “Creole” Arabic, by which he intends, in the parlance of contemporary linguistics, a codeswitched variety of Arabic (tajriym istiʕmaal luƔat al-kriyuwl). (8 ) For instance, generally speaking, “Arabic” in Arabic departments in the Arabic world stop with the classical language. (9 ) An extreme though in today’s world by no means uncommon situation is when Arabic needs to be studied in tandem with other languages in the domain of codeswitching [Bentahila et al., “Codeswitching”; also Kossmann, “Borrowing”; Newman, “Nahḍa”]. (10 ) For instance, Embarki [“Phonetics”], summarizing Al-Ani (1970), identifies four traits of consonantal “emphasis,” only one of which involves pharyngeal space. Jonathan Owens Jonathan Owens, University of Bayreuth

Phonetics Mohamed Embarki The Oxford Handbook of Arabic Linguistics Edited by Jonathan Owens Print Publication Date: Sep 2013 Online Publication Date: Dec 2013

Subject: Linguistics, Phonetics and Phonology, Languages by Region DOI: 10.1093/oxfordhb/9780199764136.013.0002

Abstract and Keywords Phonetics is a linguistic field that studies speech in terms of production, transmission, and reception. The three domains of speech study the speaker (production), the hearer (reception), and what takes place between the two (transmission). To this purpose, phoneticians use methods derived from the science of physiology for production, from physics for transmission, and from psychology for reception. In this article, the first section deals with the principal phonetic descriptions of the Arab system produced by the early Arab grammarians of the classical period (2nd/8th––5th/11th). The second section presents the consonant and vowel systems of modern Arabic. The third section deals with the contribution of experimental phonetics to the specificities of the consonant and vowel Arabic systems focusing in particular on (1) pharyngeal consonants; (2) pharyngealized consonants; (3) temporal aspects (vocalic and gemination quantity); and (4) consonant and vowel variation. Keywords: production, transmission, reception, phoneticians, Arabic phonetic system, consonant, vowel, modern Arabic, experimental phonetics

2.1 Introduction PHONETICS is a linguistic field that studies speech in terms of production, transmission, and reception. To simplify, as Lodge (2009: 2) says, the

three domains of speech study the speaker (production), the hearer (reception), and what takes place between the two (transmission). To this purpose, phoneticians use methods derived from the science of physiology for production, from physics for transmission, and from psychology for reception. Thus, as Ladefoged and Johnson note (2010: 2), there are different types of phoneticians. However, all phoneticians are conscious that the sounds they describe from these three perspectives are utilized to encode linguistic information. The same object, the sounds of words, has a concrete continuous material face and an abstract, cognitive, and categorical face. The first is phonetic; the second is phonological. Hence, even without mentioning it explicitly, phonetic descriptions have one foot in the domain of phonetics and another in phonology. The border between the two disciplines is very narrow, and the interface is multifarious (see Scobbie 2007: 17–52 for a detailed description of interfaces and overlaps between phonetics and phonology). The phonetic–phonology interface is triadic according to Kingston (2007: 401). First, phonetics defines the distinctive features. Second, it explains the phonological patterns. Third, it implements phonological representations. At the incipience of the phonetic discipline in the arabophone area (as early as the 2nd/8th century), the boundary between phonetics and phonology for the early Arab grammarians was extremely difficult to delimit (see Chapter 3). Their descriptions simultaneously included articulatory, acoustic-auditory, and phonological criteria. The first section of this chapter will deal with the principal phonetic descriptions of the Arab system produced by the early Arab grammarians of the classical period (2nd/8th to 5th/11th century). The second part will present the consonant and vowel systems of Modern Arabic, while the third part will deal with the contribution of experimental phonetics to the specificities of the consonant and vowel Arabic systems focusing in particular on (1) pharyngeal consonants, (2) pharyngealized consonants, (3) temporal aspects (vocalic quantity and gemination), and (4) consonant and vowel variation. As for prosodic structure, which I will not deal with here, one may refer to the reference works published on Arabic (Al-Ani 1970: 89–95; Watson 2002: 79–121; Canepari 2005: 327–329; Ryding 2005: 35–39; see also [Hellmuth, “Phonology”]).

2.2 The Arabic Phonetic System As Ladefoged (2003, quoted in Chelliah and de Reuse 2011: 251) quite justly remarks, there is “nothing more ephemeral than the sounds of a language. The sounds will live only as long as the language is spoken. When the sounds are those of elderly speakers whose children belong to another world, then soon those sounds will be gone forever. All that can remain are whatever records we have been able to archive.” How to collect and interpret phonetic data is dealt with in detail in Bowern (2008: 63–72) and Chelliah and de Reuse (2011: 251–278), both of which recommend using all types of evidence; written data have a big importance here. Phoneticians and phonologists working with Arabic have at their disposal an abundance of resources bequeathed by the early Arab grammarians. Reference to these works will allow us to compare in this chapter the early phonetic descriptions with those of modern

Phonetics researchers and to verify if more recent data based on the use of sophisticated techniques and instrumentations validate or invalidate the early descriptions.

2.2.1 Classical Arabic The early Arab grammarians whose work has reached us are not numerous; there are no more than 350 names in the classical period listed by Al-Suyuṭi (963/1556– 1009/1601) in his work al-Muzhir. From the phonetic descriptions that we know from this period, the consonant system of Classical Arabic included either 28 or 29 consonants, the number varying according to whether a phonetic value is given to the first letter of the alphabet (’alif (|)) or not. The consonantal phonemes were described as ’uṣūl [usʕ u:l] (primary), among which 25 were nonvocalic [ṣiħaaħ], described according to an articulatory region [ħajjiz] and an aperture [madraža, “degree”] while 4 were ) [ʔajwaf] because they are characterized by a relatively unhindered exhalation of air [hawaaʔijja] [Sara, vocalic (lit. “hollow” “Classical Lexicography”]. Table 2.1 Classification of classical Arabic consonants according to Al-Khalīl (d. 786) in Kitāb al-’Ayn (The book of the ’Ayn) Phonemes

Plain

Hollow

Region

Description

guttural

uvular

arched

apical

alveolar

interdental

apical

labial

vocalic

IPA

ʕ ħ h

x Ɣ

q k

ž ʃ dʕ

sʕ s z

tʕ d t

đʕ θ đ

r l n

f b m

a: ʔ w j

Letter-sound

The first region, the most backed articulatory area,1 is composed of three consonants: ‘“ayn” /ʕ/, “ħā”’ /ħ/, and “hā”’ /h/. The /x/ and “ġayn” /Ɣ/, which are qualified as guttural [ħalqijja]. The third second region is composed of two consonants, “khā”’ [lahawijja]:“qāf” /q/ and “kāf” /k/. The fourth region consists of three arched consonants region consists of two uvulars [šažrijja]: “ğīm” /ž/; “šīn” /š/; and “ḍād” /dʕ/. The fifth region contains the apicals [ʔasalijja] “ṣād” /sʕ/; “sīn” /s/; /z/. The sixth region has three alveolars [nitʕʕijja]: “ṭā”’ /tʕ/; “dāl” /d/; and “tā”’ /t/. The seventh region is and “zāy” [liθawijja]: “ẓāʔ” /đʕ/; “thā”’ /θ/; and “dhāl” /đ/. The eighth region consists of three apical composed of interdentals [đawlaqijja]: “rā”’ /r/; “lām” /l/; and “nūn” /n/. The ninth region consists of te labials [šafawijja]: “fā”’ consonants /f/; “bā”’ /b/; and “mīm” /m/. The two consonants “‘alif” /a:/- “hamza” /ʔ/ (these two last consonants are considered to be independent phonemes by Al-Halīl, but not so by his successors), “wāw”

/w/, and “yā”’

/j/ are vocalic

[hawa:ʔijja], and

thus they don’t have a precise articulation point in the oral caving (Roman 1977). Table 2.1 synthetizes these different articulations (the International Phonetic Alphabet [IPA] symbol corresponds, however, to the articulation of the sound–letter in Modern Arabic). Besides the 28 or 29 normative phonemes of “the language of Arabs” (i.e., ‘Al-Arabiyya), the treatises of the early Arab grammarians gave considerable attention to variation. Owens (2001: 422) qualifies Sībawayhi’s al-Kitāb (177/793) as a grammatical work that has institutionalized the variation the most. For example, Sībawayhi describes 29 phonemes as “primary” (’uṣūl), to which are added six [xafi:fa] (/n/ of articulations “secondary” (furū‘), referred to as “good” (mustaḥsana)—among which can be cited the light “nūn” assimilation), the “šīn”

/ʃ/ pronounced as “ğīm”

/ž/, the “ṣād”

referred to as “bad” (ġayr mustaḥsana)—among which are the “ğīm” /tʕ/ and the “ṣād” /đʕ/, which merge in “thāʔ”

/sʕ/ pronounced, respectively, as “tā”’

/ṣ/ pronounced as “zāy” /ž/ pronounced either as “kāf” /t/ and “sīn”

/z/—and seven articulations /k/ or as “šīn”

/s/, as well as the ḍād

/š/, the “ṭā”

/dʕ/ and the “ḍā”

/θ/.

Ibn Jinnī’s work al-Khaṣa:ʔiṣ (1002) was the first to describe in detail the Classical Arabic vowel system, including the clear distinction among the three short vowels /i, u, a/ and the three long vowels /i: u: a:/. He was also the first grammarian to rid himself of the phonetic spelling trap to describe them. The Classical Arabic phonological system clearly reveals the membership of the language in a Semitic family whose principal characteristic is a reduced vowel system limited to three cardinal qualities (with length opposition) and a rich consonantal system often exceeding 29 consonants. A second characteristic is the organization of the consonants according to the morphophonological constraints of “root” and “pattern” [Ratcliffe, “Morphology”; Hellmuth, “Phonology”]. A third characteristic is the triadic organization of certain consonantal oppositions (Watson 2002: 2–3). The triad is composed of three consonants sharing several properties and the most frequent pattern of one voiceless consonant, one voiced consonant, and a corresponding pharyngealized one, whether voiced or not. For Classical Arabic, Watson (2002: 3) gives a stop (also termed “plosive”) triad composed of /t, d, tʕ/, a sibilant triad composed of /s, z, sʕ/, and a fricative triad composed of /θ, đ, đʕ/. It should be noted that the consonants that constitute a triad cannot figure in the same etymon of the word (Boudelaa and Marslen-Wilson 2001; Watson 2002), a point discussed in some detail in this volume [Hellmuth, “Phonology”]. This organization in triads allows for a better understanding of the complex relations between consonants as well as their evolution in Modern Arabic. The pharyngealized stop consonant of Modern Arabic /dʕ/ is inserted in a palatal triad. The discussion of the evolution of the pronunciation of the uvular stop /q/ allows for a better understanding of the nature of the relation that it maintained and that it continues to maintain with /k/ on one hand and /Ɣ/ on the other (Jakobson 1957; Bonnot 1976; Roman 1981).

Phonetics 2.2.2 Modern Standard Arabic 2.2.2.1 The Consonant System Modern Arabic has 28 consonants (summarized in Table 2.2). The consonants on the left of the column are voiceless; the ones on the right are voiced. The system is based on a set of basic contrasts. The differences with the consonant system of Classical Arabic described by the early Arab grammarians seem to be minimal. However, the articulation of some segments seems to have evolved from Classical Arabic to Modern Arabic. /ž/, the ḍād /dʕ/, the “šīn” /ʃ/, and “sīn” /s/ (Beeston 1962; Murtonen 1966; Al-Wer 2003; We can cite, inter alia, the “ğīm” Embarki forthcoming). Modern Arabic is characterized by a rich consonantal system that places it slightly over the average of 22.8 consonants derived from the University of California–Los Table 2.2 Consonantal system of Modern Standard Arabic Bilabial

Plosive

Nasal

Labiodental

Dental

Postalveolar

Palatal

t, d

Velar

Uvular

Pharyngeal

Glottal

Trill Fricative

Dentoalveolar

r f

θ đ

s z

Approximant

š ž

x Ɣ j

Lateral

ħ ʕ

approximant Pharyngealized

tʕ dʕ

plosive Pharyngealized

đʕ

sʕ

fricative Angeles (UCLA) phonological segment inventory database (UPSID 317 )2 (Maddieson 1984). As Newman (2005: 185) notes, Arabic is distinguished by the presence of certain consonants (up to ten) that are not common to other UPSID-based languages—such as the pharyngealized stops and fricatives. Arabic consonants are opposed to one another in terms of different articulatory manners and places of articulation. Arabic has a large number of places of articulation: (1) five different places of articulation for stops against an average of three in other UPSID-defined languages; and (2) seven different places of articulation for fricatives against an average of four in other languages. Eight pairs of obstruent consonants contrast in terms of voicing, and there is an overall of 15 voiced consonants and 13 unvoiced. The effects of voicing on consonants have been well described in the literature (Port et al. 1980; Mitleb 1984a). Among Arabic obstruents, the duration of the voiceless is longer than that of the voiced cognate. This temporal pattern is reversed for the contiguous vowel: the duration of the vowel before a voiceless segment is short, whereas that of the vowel accompanying a voiced consonant is long. Mitleb’s (1984a) study, carried out on eight speakers of Jordanian origin, confirms this temporal pattern. The frequency of phonemes reveals several functional aspects of language. For instance, it is linked to the linguistic change and to the representation of the mental lexicon (Bybee 2003: 11–12; [Hellmuth, “Phonology”]). It appears that the most frequent words are pronounced differently from the less frequent words and that they are more subject to reduction phenomena (Gordon 2007: 73). A brief exploration of the Arabic lexicon in speech situations shows that the frequency of anterior consonants

Phonetics

Table 2.3 Frequency of 26 consonants in modern Arabic Sound

11.77

6.18

5.14

4.66

4.49

3.35

3.34

3.11

2.56

2.53

2.50

2.06

2.13

Sound

sʕ

tʕ

dʕ

1.85

1.79

1.35

0.96

0.95

0.91

0.80

0.67

0.65

0.64

0.53

0.36

0.20

Source: Newman (2005: 191).

(labials, dentals, and

dento-alveolars) is important. Indeed, 9 of the most frequent 10 consonants are articulated at the front of the vocal tract (see Table 2.3). On one hand, this contrasts strongly with the common perception of Arabic as a “guttural language,” and, on the other hand, it explains the coarticulatory patterns of the language (a better motor control of the front of the vocal tract).

2.2.2.2 The Vowel System Modern Arabic has one of the most elementary types of vowel systems, composed of three cardinal vowels /i u a/, which are common to a very large majority of natural languages. A reduced vocalic system correlates with a richer consonantal system (Flemming 2001). The frequency in the lexicon of these three vowels is unequal: the vowel /a/ has a frequency that slightly exceeds 60%; /i/ just below 25%; and /u/ slightly below 15% (Newman 2005: 205). This order probably follows a general tendency in the languages of the world. The three cardinal vowel qualities are doubled in number in Modern Arabic by a contrastive lengthening opposing the short vowels to the long vowels (Cantineau 1960; Watson 2002: 22–23). All in all, Modern Arabic has six vowel /i i: u u: a a:/. These vowels contrast essentially via two parameters: (1) the height of the body of the tongue or high vs. low, and (2) the front-back position of the tongue or front vs. central vs. back. The vowels /i/ and /i:/ are high front vowels, /u/ and /u:/ are high back, while /a/ and /a:/ are low central vowels. Lip rounding is not contrastive in Arabic; only high back vowels are produced with slight lip rounding. As explained in Section 2.3.4, these vowels show strong variation, determined by the linguistic context, prosodic position, and geographical origin of the speakers.

2.3 The Contribution of Experimental Phonetics As Heselwood and Hassan (2011) indicate in the introduction of their collective work Instrumental Studies in Arabic Phonetics, the early Arab grammarians gave very detailed phonetic descriptions of Arabic sounds. From the end of the 12th to the beginning of the 13th century, they supposedly produced the first representation of the vocal tract with the main consonantal articulations (ibid.). This representation is distinguished by its modernity because it shows precisely both place of consonantal articulation and the articulators. Figure 2.1 provides a diagram of the vocal tract borrowed from Bakalla (1982) and quoted by Heselwood and Hassan. Although phonetics was not yet a structured discipline and tools for observation did not exist, the early Arab grammarians still gave us ample, precise indications about the articulatory characteristics of consonants as well as on their acoustic and perceptive properties (Bonnot 1976; Heselwood and Hassan 2011). Thanks to the most modern equipment, researchers today can compare their own observations with those of the early Arab grammarians. This, however, is not without disagreement: for example, Sībawayhi distinguished between the opposing mažhu:ra lit. “made loud” and mahmu:sa “whispered,” which many modern phoneticians and phonologists translate with voiced versus voiceless, with the exception of Jakobson (1957), who uses lenis versus fortis to describe this contrast. Instead of doing a chronological presentation of the main research in Arabic experimental phonetics, I have chosen to present the most important conclusions of the works, which pertain to (1) pharyngeal consonants, (2) pharyngealized consonants, (3) length, and (4) consonantal and vocalic variation.

Click to view larger Figure 2.1 Vocal tract diagram titled Ṣūrat makhārij al-ḥurūf (Picture of the points of articulation of the letters) from Miftāḥ al-‘Ulūm (The key to the sciences) by Al-Sakkāki. Dotted line indicates the nasal passage with a nostril above the lip (from Bakalla 1982: 87, quoted by Heselwood and Hassan 2011: 7).

2.3.1 Pharyngeal Consonants Modern phoneticians qualify the two guttural consonants

“‘ayn”

/ʕ/ and “hā”’

/ħ/ in Al-Khalīl as pharyngeal. However, not all

Phonetics researchers agree that they belong phonologically to a so-called natural class (see Zawaydeh 2003 for an extensive review of that point). Yet Sībawayhi does not use the same terms to describe these two consonants. The sound /ħ/ is described as fricative (raxw), while the sound /ʕ/, which is situated between the stop (ʃadi:d) and fricative, is produced with [tardi:d], according to Sībawayhi’s terminology. Ghali (1983) and, following him, Hassan (2011) use the quality (taraddudijja) “frequentative” to designate the consonant /ʕ/. Al-Ani (1970) is considered the first experimental work in Arabic phonetics. His research is based on cineradiographic data (x-ray), which give accurate images of the surface of the vocal tract, lips, tongue, uvula, and pharyngeal movements combined with acoustic data to describe the consonants and vowels of Modern Arabic. Relying on the productions of four Iraqi native speakers, Al-Ani (ibid., 59–60) confirms Sībawayhi’s description of /ħ/ as a fricative voiceless consonant. If he accepts the pharyngeal place of the consonant /ʕ/, he describes it, however, as a voiceless stop in all positions (initial, medial, final) whether singleton or geminated (ibid., 62–63). If we carefully examine the mid-sagittal sections from the cineradiographic films (Al-Ani 1970 72–74) and the spectrograms (ibid., 65–71), we can observe that /ʕ/ presents a constriction lower in the pharynx with the body of the tongue in more retracted position compared with /ħ/ (see Figure 2.2). On the acoustic level, /ʕ/ does not have the profile of a stop or even that of a fricative as is /ħ/; one can, however, see that it is clearly voiced. The stop articulatory manner of /ʕ/ described by Al-Ani was not often followed by other phoneticians. Using the same techniques (x-ray and acoustic measurements), Ghali (1983: 440) chooses the feature frequentative for /ʕ/, which Sībawayhi also uses, and assigns it the “trill” articulatory manner. Besides / ʕ /, Ghali (ibid., 441) classifies four further consonants in the trill category: the alveolar /r/; the two uvulars /x/; the /Ɣ/; and the glottal /ʔ/. In the last two decades, more sophisticated technologies such as ultra-fast imaging have been used. Some researchers successfully applied these techniques to perfect our knowledge of Arabic consonants. Zawaydeh (2003) uses the endoscopic technique to visualize articulatory adjustments during the production of these two consonants in Jordanian Arabic. The results indicate that, during the production of /ħ/ and /ʕ/ as well as during the production of pharyngealized consonants, the distance between the epiglottis and the pharyngeal wall is reduced (Zawaydeh 2003: 287). These results are similar to those obtained by Ghazeli (1977), who uses cineradiography to study pharyngealized consonants. In a recent study of Iraqi Arabic, Hassan et al. (2011) employs ultrafast laryngoscopy (an imaging technique using endoscopy), combined with electroglottography (EGG). This technique captures vocal fold vibrations by positioning two electrodes on the neck on both sides of the thyroid cartilage.

Figure 2.2 X-ray tracing of the articulation of /ħ/ (dotted line) and /ʕ/ (plain) in the context of /Ci/ (from Al-Ani 1970: 72).

These physiological data were combined with acoustic data to complete our knowledge of the features of /ħ/ and /ʕ/. While confirming their pharyngeal place of articulation, Hassan et al. (2011: 834) confirm that these two consonants are pronounced by Iraqi speakers as aryepiglotto-epiglottal fricatives, transcribed as voiceless /H/ and voiced /ʢ/, and are considered as variants of /ħ/ and /ʕ/, respectively. Heselwood (2007: 5), relying on Laufer (1996: 114), indicates that /ʕ/ in the production of 21 speakers from 11 different Arab countries is never pronounced as a fricative consonant; it does not inherit this characteristic except through the fact that it is phonologically paired with /ħ/, which is a real fricative. On the basis of articulatory and acoustic data, Heselwood (2007: 9–28) describes /ʕ/ as a “tight approximant,” which he proposes to represent using the following symbol . On the basis of acoustic and articulatory data from Moroccan speakers, Yeou and Maeda (2011: 155) conclude that /ħ/ and /ʕ/ are real approximants, since, unlike certain fricatives such as /s/, the two pharyngeals have a larger articulatory constriction and the turbulence is present only for the voiceless consonant.

2.3.2 The Pharyngealized Consonants Arabic has a specific phonological contrast that opposes plain dental or dentoalveolar consonants to their pharyngealized cognates. Modern Arabic has four pharyngealized consonants /tʕ dʕ đʕ sʕ/; some modern Arabic dialects have slightly more, while others have less. Ferguson (1956) shows that allophonic phraryngealized variations exist in Modern Arabic for the /l/. I exclude the consonant /q/ from this correlation and adopt Bonnot’s (1976) point of view, who dedicates a long chapter to the relation between the two stops /k/ and /q/ and concluded, based on articulatory and acoustic data, the absence of the pharyngealized feature during the production of the consonant /q/. The pharyngealized consonants /tʕ dʕ đʕ sʕ/ existed in Classical Arabic with presumably a slightly more backed place of articulation and a different articulation manner for some of them (see Al-Wer 2003: 28–29 for the evolution of /dʕ/; see Roman 1981 for the evolution of the emphatic among the guttural consonants). These consonants were often designated by a plurality of Arab terms such as “istiʔla:ʔ,” “tafxi:m,” “‘itʕba:q,” or “‘iħsʕa:r,” which modern linguists translated by “emphatic” (see Bonnot 1976, esp. chapter dedicated to emphasis, 84–118). The main dental or dento-alveolar articulation of these consonants is not a major point of disagreement among researchers, but the same

Phonetics cannot be said of their pharyngealized secondary articulation. According to Ladefoged and Maddieson (1996: 365–366), the place of constriction of the secondary pharyngeal articulation is formed midway between the uvula and the epiglottis. The sagittal sections presented by Al-Ani (1970: 57–58; see Figure 2.3) show that the back of the tongue has a rather flat position and that its root has a more backed position for the pharyngealized consonant compared with its non-pharyngealized counterpart. The narrowing at the origin of the constriction seems to be produced in the median region of the oropharynx. The acoustic data deal more with the effects of the adjacent vowel showing that the secondary pharyngealized articulation leads to a rise of the frequency of the first formant, F1, and a lowering of the frequency of the second formant, F2 (Al-Ani 1979: 44–56). Shahin (1997) interprets this acoustic pattern in phonological terms. The observed rise of F1 associated with pharyngealized consonants is shared by other guttural consonants (the glottals, pharyngeal, and uvulars); the author regroups them in a class called pharyngealization harmony. The lowering of F2, however, concerns only pharyngealized consonants, and Shahin proposes regrouping them in a different class called uvularization harmony. Ghazeli’s (1977) study using the same instrumentation as Al-Ani (1970) shows that the main characteristic of Arabic pharyngealized consonants on the articulatory level is a retraction of the root of the tongue and a flattening of its posterior part in the shape of a plateau, a tightening of the pharyngeal cavity above the epiglottis, and a slight labial protrusion. Based on the cineradiographic data of a Saudi speaker, Bonnot (1976: 369) determines that the constriction of a pharyngealized consonant goes from the uvula region up to the deepest level of the pharynx. Compared with its non-pharyngealized counterpart, a pharyngealized consonant is distinguished by a more backed place of articulation and a superior articulation strength as well as by a slight increase in its length and a shortening of the adjacent vowel (ibid., 472– 473). These data are confirmed in Elgendy’s (2001) study on pharyngealization.

Click to view larger Figure 2.3 X-ray tracing of the articulation of /t/ (dotted line) and /tʕ/ (plain) in the context of /Ci/ (from Al-Ani 1970: 57).

Several chapters of Hassan and Heselwood (2011) examine the articulatory and acoustic properties of pharyngealized consonants using modern techniques such as nasoendoscopy, videofluoroscopy, electromagnetic midsagittal articulometry (EMA), and ultrasound. In addition to the retraction of the tongue body and the flattening of its posterior part, these data show that pharyngealized consonants are different from their non-pharyngealized counterparts in the volume of buccal and pharyngeal resonance cavities; the adjustments of the body of the tongue and its root; the height of the lower jaw; and the different positions of the hyoid bone, the epiglottis, the aryepiglottic cords, and the larynx. According to Al-Ani (1970), acoustic data focus the effects of pharyngealized consonants on their phonetic environment. These effects manifest themselves through an important modification of the first two formants of the vowels, shown by a significant increase of F1 and a substantial decrease of F2. Jongman et al. (2011: 89) report significant effects on the third formant, F3, in the production of 12 Jordanian speakers. The frequency of F3 increases significantly when the vowel is placed next to a pharyngealized consonant. Bonnot (1976: 451) shows that F1 and F2 are very close in a pharyngealized consonantal environment compared with a non-pharyngealized environment. Embarki et al. (2011b) analyzes the influence of pharyngealization on Jordanian, Kuwaiti, Moroccan, and Yemeni speakers by comparing V1CʕV2 sequences, where C is /tʕ dʕ đʕ sʕ/, with similar V1CV2 sequences containing the non-pharyngealized cognates /t d đ s/. The frequency measures of the first two formants F1 and F2 as well as the distance F2 – F1 (Fv) taken at three different landmarks of the vowel (onset, midset, and offset) confirm the frequency differences indicated in the literature, that is, the increase of F1, lowering of F2, and closeness of the two formants (ibid., 146). The study also shows that the values of Fv in the environment of a nonpharyngealized consonant are on average greater by 348 Hz than the corresponding pharyngealized one. Also, the differences of Fv between the two contexts (non-pharyngealized– pharyngealized) are stronger at the onset of V2 than at the offset of V1 (Embarki et al. 2011b: 147). Embarki et al. conclude that the influence exercised by a pharyngealized consonant on the vocal environment is stronger than its non- pharyngealized cognate. Jongman et al. (2011: 88–89) show that the effects of the pharyngealized consonant in initial or final position of a word significantly impact the three first formants of the adjacent vowel, influencing the vowel in a constant way from the onset to the offset via the midset. While confirming the effects of pharyngealization on the two first formants of the adjacent vowels, Ghazeli (1981: 275) shows that the direction of the pharyngeal coarticulation has a more regressive or persistent (carryover) (left–right) nature than a progressive or anticipatory (right–left) one. Ali and Daniloff (1972) use cinefluographic data in Iraqi Arabic to highlight this characteristic: a left–right effect being more important than a right–left effect both in the magnitude of the retraction gesture of the tongue and in the number of segments affected by the spread of pharyngealization. However, in a study based on articulatory data (EMA) of a Tunisian speaker, Embarki et al. (2011a: 210) indicate that the effect of a pharyngealized consonant in C2 position (medial position) starts with the first vowel of the word (V1) and continues above the stationary part of the vowel of the second syllable (V2), indicating that the pharyngealization has carryover as well as anticipatory effects. The acoustic data of the study of Jongman et al. (2011: 91) show that the pharyngealized consonant placed in the final position of a word has significant effects on the non-pharyngealized consonant placed at initial position. On the other hand, the same pharyngealized consonant placed in initial position of a word has no significant effect on the final consonant. This shows that the anticipatory effects (right–left) of pharyngealization are more important than the carryover effects (left–right).

Phonetics Another interesting piece of data in Ghazeli’s (1981) work is the spread of pharyngealization in relation to morphemic boundaries. Ghazeli confirms that the spread of phrayngealization effects stops at the boundaries of the word (ibid., 275). These effects do not seem to cross from one word to the other. The coarticulatory effects of the pharyngealized consonant on adjacent vowels were measured, among other ways, by a linear regression (i.e., the locus equation; see Lindblom 1963) quantifying the coarticulation degree between the consonant and the vowel between extremes: 0 for a null coarticulation; 1 for a maximal coarticulation. Yeou (1997) shows that the value of the slope of a pharyngealized consonant in Modern Arabic produced by Moroccan speakers is weaker than the value of the slope of its nonpharyngealized counterpart. This same pattern is confirmed in Embarki’s et al. (2011a) study of the production of pharyngealized consonants in Modern Arabic and Arabic dialect of 16 native speakers from four different countries, Jordan, Kuwait, Morocco, and Yemen.

2.3.3 Duration Duration is a phonetic parameter that is specific to all linguistic units, consonants, vowels, syllables, words, and sentences. As Coates (1980: 4) says, “Time is vital in the understanding of phonological processes and processing. A fortiori, it is vital in phonological representation too.” Traditionally, the rubric duration deals with contrastive length that pertains not only to vowels in a large number of natural languages (Ladefoged and Maddieson 1996: 320) but also to consonants, although to a lesser degree. Arabic is among the languages that use quantity contrast both for consonants and vowels. In addition to these two categories, I will include voice onset time (VOT), which is also mainly a temporal phenomenon. However, I will not talk about the effects of consonantal voicing on the duration of both the consonant and the vowel. In this respect I refer the reader to the works of Port et al. (1980) and Mitleb (1984a).

2.3.3.1 Vowel Length As for vowels, most languages using vowel quantity have a duration opposition between two vowel categories, the short and the long ones. Ladefoged and Maddieson (1996: 320–321) explain that quantity can oppose three or even four vowel categories in some languages. The ratio of duration between vowels (the duration of the long vowel divided by the duration of the short vowel) varies enormously among languages. Some languages use a low ratio, for example 1.3, while others use a significantly longer ratio, such as 3.2 (Lehiste 1970). In Arabic, studies show that the quantity ratios between vowels vary a great deal. Al-Ani (1970: 75) shows that the relative duration of short isolated vowels is 100 to 150 ms, while with the long ones it is 225 to 350 ms, which makes the ratio long to short more than two to one. Port et al. (1980) present a ratio of 2.6 for Egyptian, Iraqi, and Kuwaiti speakers. Mitleb’s (1984b: 231) study on Jordanian Arabic showed that the Arabic long vowel is 65% longer compared with its short counterpart, a 1.5 ratio. Belkaid (1984) presents a ratio slightly greater than 2 for speakers of Tunisian origin. Studying three speakers of different dialectal origin, Abou Haidar (1991) presents varying ratios, but an average of around 2.6. Alioua (1992) finds a mean ratio of 2 for three Moroccan speakers. Jomaa (1994) proposes an intermediate ratio of 2.4 for several dialects. These ratios are, nonetheless, less than those of Modern Arabic and are between 1.3 and 2, with higher relationships in eastern dialects and lower ones in dialects from the Maghreb (Jomaa 1994). The contrastive vowel length is conveyed essentially through duration (Lehiste 1970). In Al-Ani’s (1970: 22–25) study, it seems that quantity contrast is accompanied, in an insignificant way, by vowel quality (see Section 2.2.2.2). The length ratio between short and long vowels is affected by other linguistic parameters. For example, Mitleb (1984b) shows differences that are inherent to the nature of the adjacent consonant, whether it is singleton or geminated (see the previous discussion). De Jong and Zawaydeh (2002: 319) show that long, stressed vowels in Arabic were lengthened 120% by native Jordanian speakers in contrast with their short counterparts. Canepari (2005: 319) indicates that, in a unstressed position, long vowels in Arabic are realized like semi-long vowels. However, the sensitivity is limited by the theory of acoustic invariance, which is based on the hypothesis that invariable acoustic properties correspond to a segment or to phonetic features, independently of context, speaker, and language (Lahiri et al. 1984; Pickett et al. 1999). Thus, according to Zawaydeh and de Jong (1999), contrastive vowel length is maintained in Arabic fast speech. Port et al. (1980) and Mitleb (1984b: 233) indicate that the length domain in Arabic is determined at thesegmental level, while in other languages it is determined at the syllable level such as in Swedish or at the word level such as in English.

2.3.3.2 Gemination Consonant length is treated in languages in terms of gemination. Duration and gemination refer to different aspects of articulation. Quantity is a matter of length, while gemination applies to the repetition of the same articulation. The question of whether Arabic consonants are really geminated or simply long has been discussed by researchers. In his study of pharyngealization, Bonnot (1976: 225) uses cineradiographic data to prove that the closure release of the geminated /tʕ/ occurs only at the final occlusion, which leads him to conclude that gemination in Arabic is not present with stops in a two-phase articulation but rather in one single phase (ibid., 450). According to Bonnot, the most important criterion is an increase in duration, and the so-called geminated consonants are in reality long consonants. Languages such as Arabic, which combine both vowel and consonant quantity, are less numerous. In Modern Arabic, the distribution of the 28 consonants is completely regular, with each consonant occupying three positions: initial, medial, and final. All consonants can be singleton or geminated (Kaye 2009: 563). Contrary to the majority of languages where stops are geminated preferably in the medial position of the word (Ladfoged and Maddieson 1996: 92–93), dentals and dento-alveolars in Arabic can be geminated in initial position as well, as with all so-called solar consonants. Gemination is thus phonological in Arabic, and it is highly contrastive in distinctions of morphological nature (Watson 2002; see, under the morphology subsection of the chapter of Arabic, Kaye 2009: 572–574; [Ratcliffe, “Morphology”]). Al-Ani (1970: 75–77) shows that the duration of geminated consonants increases until it reaches twice the duration of its singleton counterpart. This ratio between a consonant and its

Phonetics geminated counterpart varies slightly in the literature to the point that sometimes overlaps are noticed between the lowest average durations for a geminated consonant and the highest durations for its singleton counterpart. Bonnot (1976) indicates that the geminate pharyngealized stop /tʕ tʕ/ is distinguished from its singleton counterpart /tʕ/ essentially through the duration of the complete closure, which is longer for the geminated one than for its singleton counterpart, with overlapping zones. Al-Ani (1970: 33) indicates that the duration of the geminated consonant is twice that of its singleton counterpart. In Arabic, the geminated consonant can be preceded by a short vowel (V) or a long vowel (V:); the temporal pattern can also be globally affected without reducing a long vowel to the point of confusing it with a phonologically distinct short vowel (Hassan 2003: 46). Khattab (2007: 156) shows that the geminate– singleton ratio in Lebanese is higher when the preceding vowel is short (2.5) compared with a long vowel context (2.09). Hassan (2003) indicates that the temporal pattern is different when a long vowel is followed by a singleton consonant (V:C) compared with the short vowel context followed by a geminate (VC). Mitleb (1984b) finds the same distributional pattern. Bonnot (1976: 235) uncovers a difference in the closure duration of the consonant when it is preceded by a short vowel or a long vowel. In the first case, the closure can have a longer duration of up to 50 ms compared with the second case. Basing his findings on electromyography (EMG) data for Estonian and English, Lehiste et al. (1973: 146–147) indicate that singleton and geminated consonants are different through the duration of the closure and the amplitude of the peaks. Al-Tamimi and Khattab (2011: 214– 215) show that the differences between singleton and geminated consonants of Lebanese speakers included, in addition to the duration, other acoustic parameters such as F0, intensity, and the degree of voicing of the consonant. Bonnot (1976) notices articulatory differences illustrated by the lowering of the lower maxillary. Indeed, the lower maxillary is lowered less during the production of the singleton consonant /t: tʕ:/ (ibid., 255, 346). Bonnot also indicates that the position of the tongue varies for a singleton pharyngealized consonant and its geminated counterpart. Unlike the front part of the tongue, which does not show differences, the geminated consonant causes a tightening of the posterior part of the back of the tongue (ibid., 371). Other results show that the geminated stop is distinguished from its singleton counterpart through the VOT duration.

2.3.3.3 VOT VOT is defined as the temporal difference between the release of the complete closure and the onset of quasi-periodical vibrations of the vocal folds. This parameter applies only to stop consonants. It is described as positive when the first voiced periodical resonance starts immediately after the release of the consonant, as is the case of voiceless stops. It is described as negative when the vibrations of the vocal folds begin before the closure release, as is the case with voiced stops. Lisker and Abramson’s (1964) classic study, based on the examination of stop consonants in 11 languages, showed that this temporal interval, which is the VOT, allowed for the distinction among three categories of stops in those languages: (1) voiceless unaspirated stops, with a positive VOT between 0 and 25 ms, or short lag; (2) voiceless aspirated stops with a positive VOT of 60 to 100 ms, or long lag; and (3) voiced stops, with vibrations beginning before the closure release. Some languages use the three patterns to oppose stops, while other languages use a binary opposition only between the two patterns. Lisker and Abramson (1964) show that the VOT duration varied according to the place of articulation of the consonant: longer for the velars, shorter for the labials, and intermediate for dentals. Cho and Ladefoged (1999: 213) mention six criteria as the origin of the VOT variation, including the cavity volume in front of and behind the constriction, the movement of the articulators, and the contact zone between the articulators. Al-Ani (1970: 76) indicates that Arabic has a negative VOT for voiced consonants; the duration of this prevoicing varies between 50 and 300 ms according to the position of the consonant (initial, medial, or final) and its nature (singleton or geminated). On the other hand, Arabic has a positive VOT for voiceless consonants that varies between 20 and 40 ms for unaspirated and 35 and 60 for aspirated variants. In their study of Lebanese Arabic, Yeni Komshian et al. (1977: 38) indicate that stops are characterized by a binary VOT, a long prevoicing, or negative VOT for the voiced consonants /b d dʕ/ varying between 40 and 80 ms and a short interval or positive VOT for the voiceless consonants /t tʕ k q/ between 15 and 35 ms. This study did not examine the VOT of the glottal consonant /ʔ/. Al-Ani (1970: 60–62) describes the latter with a short VOT of 15 to 20 ms. Al-Ani’s (1970) study shows VOT differences according to pharyngealization contrast. The VOT of /t/ is longer than that of its pharyngealized /tʕ/ counterpart, by 40–45 ms for the first and only by 20–30 ms for the second (Al-Ani 1970: 44–45). Yeni-Komshian’s et al. (1977: 42) results show differences between pharyngealized consonants /tʕ dʕ/ and corresponding non-pharyngealized /t d/ presented in the form of overlapping zones of 0 to 30 ms. The VOT of voiceless consonants /t tʕ/ appears to be different; it is clearly shorter for the pharyngealized consonant (ibid., 40). Ghazeli (1977) confirms this distribution: the VOT of /tʕ/ is positive although very short (15 ms) compared with the double (30 ms) for the non-pharyngealized consonant /t/. Zeroual et al. (2007: 400) also show that the voiceless pharyngealized stop /tʕ/ has a positive VOT, 14 ms shorter than its non-pharyngealized counterpart /t/ (48 ms). In Arabic phonology, the question of whether there is a phraryngealized relation that links the consonants /k/ and /q/ is amply discussed (see Section 2.2.1). VOT seems to be one of the elements taken into consideration. Al-Ani (1970: 32) found the same pattern in Iraqi speakers: a longer VOT for /k/ between 35 and 44 ms; and a shorter VOT for /q/ varying between 20 and 26 ms. On the other hand, Yeni-Komshian et al. (1977: 42) presented averages of positive VOT that seem similar for the two consonants: 25 to 30 ms for /k/; and 25 to 35 ms for /q/. This is probably because Lebanese speakers utter a /k/ that is close to a uvular consonant. The dominant VOT pattern for /k/ and /q/ could be explained by the idea according to which the relation linking these two consonants is of the same nature as the one linking /t/ and /tʕ/—that is, a pharyngealized relation that materializes, inter alia, through a long VOT for /t k/ opposed to a short VOT for /tʕ q/. Lisker and Abramson (1964) indicated that the duration of the VOT varies according to the place of articulation of the consonant: longer for velar consonants; shorter for labial consonants; medial for dentals. However, this is not the case here, and, despite a more backed place of constriction than that of /k/, /q/ inherits a shorter VOT. The explanation is given in part in Cho and Ladefoged (1999: 213), which explains the VOT variation in terms of the volume of the cavity in front and behind the constriction, the movements of the articulators, and the contact zone between the articulators. Basing his data on cineradiographic data, Bonnot (1976: 440) gives details on the

Phonetics articulation of the two consonants and on the contact zones between articulators; these details explain the long release of /k/ and nearly simultaneous release of /q/. Besides the variation according to the place of articulation of the consonant, Yeni-Komshian et al. (1977: 43) show that the duration of the VOT with Lebanese speakers varies according to the adjacent vowel: the VOT is longer with front vowels.

2.3.4 Consonant and Vowel Variation Variation concerns all segmental units of Arabic—consonants as well as vowels. The best-known phenomenon for consonants is that of assimilation. Kaye (2009: 564) indicates several consonant assimilation cases. The assimilation in Arabic concerns all consonants and can be partial or total. The hamza, the glottal stop /ʔ/, is considered by some as a consonant that gets completely assimilated by the solar adjacent consonant when it is at a word initial position (Canepari 2005: 325). The Arabic linguistic tradition, on the other hand, considers this purely graphic hamza as a latent consonant. Assimilation as a phonetic phenomenon was well studied by early grammarians, who precisely described the assimilation of /n/ in [ŋ], [nʕ], or [m] before /q, k,ʃ, j, s, z, sʕ, d, t, tʕ, dʕ θ, đ, đʕ, f/. They underlined the dependency of the nasal expansion on the place of articulation, thus showing that the guttural (stops and pharyngeal fricatives /ħ/, /ʕ/, and glottals /ʔ/ /h/) blocked this assimilation (Bakalla 1983). Consonant assimilation was also explored in its phonological dimension (Abu Salim 1988). As presented already, early Arab grammarians emphasized the allophonic variants of consonants (see Sībawayhi’s description of secondary articulations, mustahḥsana and ġayr mustahḥsana). Embarki et al. (2011a) explain that the differences for locus equations of pharyngealized consonants between Modern Arabic and Dialectal Arabic and among the four countries used in the study (Jordan, Kuwait, Morocco, and Yemen) were due to a weakening of the pharyngealization gesture. Indeed, pharyngealized consonants tend to be articulated like their nonpharyngealized counterparts with very few retraction effects in the back of the tongue; this tendency is very clear in the realization of the consonant /sʕ/ (ibid., 204). As explained in Section 2.2.2.2, the Arabic vowel system consists of three cardinal qualities that contrast in terms of length: /i u a/ versus /i: u: a:/. This configuration is absolute and does not consider the allophonic realization of phonemes, which is slightly richer. Early Arab grammarians such as Sī bawayhi, described these variations, such as the precision of the imala phenomenon (cf. Sara 2007; [Sara, “Classical Lexicography”]). Kaye (2009: 565) explains that variation affects Arabic short vowels more than long ones. He lists a total of 16 different allophones for the six basic phonemes; Al-Ani (1970: 23–24) lists 17 allophonic realizations. Canepari (2005) illustrates on a diagram the principal allophonic realizations of six vowels in Modern Arabic (see Figure 2.4).

Figure 2.4 Articulatory space of vowels in modern Arabic and their allophones (from Canepari 2005: 317).

These allophonic realizations essentially depend on the phonetic context (the nature of the adjacent consonant) and the prosodic nature (stressed vs. unstressed syllable). The aforementioned studies, which are specific to pharyngealization, show without exception that in a pharyngealized context the frequency of F1 increases and the frequency of F2 decreases noticeably. The phonetic contiguity of certain consonants pushes the cardinal vowels toward less peripheral frequencies. Embarki et al. (2006) show that the formants of three short cardinal vowels of Modern Arabic presented acoustic signs of nasalization when these vowels were used in the morphologic process of tanwin (_VN#). In a literature review, Newman (2005) lists the different formant values (F1 and F2) in Modern Arabic in a non-pharyngealized context (see Table 2.4). These values show a rather important dispersion that indicates a possible free variation around the six Arabic vowels. Can we really talk in such a case of free variation? The very different values listed in the literature can be explained by the nature of the corpus, the phonetic context and the geographical and dialectal differences of the speakers in the studies in question. The speakers are either from a unique dialectal origin, Iraqi (Al-Ani), Tunisian (Belkaïd), or Egyptian (Newman), or are from different countries (Abou Haidar; Ghazeli). The question that poses itself is how the subjects perceive this variation. Jongman et al. (2011: 93) show that the vowels extracted from phonetic sequences that were affected by pharyngealization and used in perception tests obtain high identification rates. This demonstrates that the Arab subjects have internalized this vowel dispersion and that they use it optimally in lexical treatment tasks.

Phonetics

Table 2.4 Vowel dispersion according to the frequency of the first two formants F1/F2 i:

Al-Ani

285

2200

290

2200

285

775

290

800

675

1200

600

1500

Ghazeli

310

2225

455

1780

330

900

450

1125

Belkaïd

285

2195

355

1830

310

790

340

995

425

1720

400

1640

Haidar

315

2230

485

1750

335

835

500

1120

690

1500

675

1585

Newman

390

1870

435

1790

465

1075

480

1170

620

1455

615

1460

Abou

Source: Newman (2005).

2.4 Conclusion This chapter shows the range and uniqueness of experimental phonetic studies. The studies cited here indicate that the Arabic language is a distinct language with specific phonetic contrasts and that it shares universal phonetic features with other languages. If we can be reasonably satisfied with the degree of precision of the phonetic descriptions, exclusively sensorimotor here, much remains to be done in the field of speech perception in Arabic. Detailed studies on perception should be conducted to uncover not only robust contrasts but also subtle articulatory and acoustic features. An important part of this chapter was dedicated to the articulatory and acoustic features of pharyngeal and pharyngealized consonants, where the literature is extensive. Little data, however, have been published on the perception of these consonants or on their vowel environment. Our phonetic knowledge should be enhanced by techniques such as functional magnetic resonance imaging (fMRI), which have not been used here. Research still needs to be conducted to complete our understanding of listeners’ and speakers’ use of contrasts specific to Arabic to work out, convey, and interpret the most diverse meanings as well as the mobilization of these contrasts in morphology and psychology.

References Abou Haidar, Laura. 1991.Variabilité et invariance du système vocalique de l’arabe standard. PhD diss., Université de Franche-Comté, Besançon. Abu Salim, I. M. 1988. Consonant assimilation in Arabic: An autosegmental perspective. Lingua 74: 45–66. Al-Ani, Salman H. 1970.Arabic phonology: An acoustical and physiological investigation. The Hague: Mouton. Al-Tamimi, Jalal and Ghada Khattab. 2011. Multiple cues for the singleton-geminate contrast in Lebanese Arabic: Acoustic investigation of stops and fricatives.Proceedings of the 17th International Congress of Phonetic Sciences, Hong Kong, 212–215. http://www.icphs2011.hk. Al-Wer, Enam. 2003. Variability reproduced: A variationist view of the [dʕ]/[đʕ] opposition in modern Arabic dialects. In Approaches to Arabic dialectology, ed. Martine Haak, Rudolf de Jong, and Kees Versteegh, 21–31. Amsterdam: Brill. Ali, Latif and Raymond Daniloff. 1972. A contrastive cinefluorographic investigation of the articulation of emphatic–non emphatic cognate consonants. Studia Linguistica 26: 81–105. Alioua, Ahmed. 1992. De la corré lation entre la duré e et l’aperture des voyelles brèves en arabe littéral. Travaux de l’Institut de Phonétique de Strasbourg 22: 1–8. Bakalla, Muhammad H. 1982. Ibn Jinnī: An early Arab Muslim phonetician. Taipei: European Language Publications. —— 1983. The treatment of nasal elements by early Arab and Muslim phoneticians. In The history of linguistics in the Near East, ed. C. H. M. Versteegh, Koerner Konrad, and Hans-Josef Niederehe, 49–71. Amsterdam:John Benjamins. Beeston, Alfred. 1962. Arabian sibilants. Journal of Semitic Studies 7: 222–233. Belkaïd, Y. 1984. Les voyelles de l’arabe littéraire moderne: analyse spectrographique.Travaux de l’Institut de Phonétique de Strasbourg 16: 217–240. Bonnot, Jean-François. 1976. Contribution à l’Etude des consonnes emphatiques de l’Arabe à partir de méthodes expérimentales. PhD diss., Universitédes Sciences Humaines de Strasbourg.

Phonetics Boudelaa, Sami and William Marslen-Wilson. 2001. Morphological units in the Arabic mental lexicon. Cognition 81: 65–92. Bowern, Claire. 2008. Linguistic fieldwork: A practical guide. New York:Palgrave MacMillan. Bybee, Joan. 2003. Phonology and language use. Cambridge, UK: Cambridge University Press. Canepari, L. 2005. A handbook of pronunciation. München: Lincom Europa. Cantineau, Jean. 1960. Etudes de linguistique arabe. Paris: Klincksieck. Chelliah, Shobhana L. and Willem De Reuse. 2011. Handbook of descriptive linguistic fieldwork. New York: Springer. Cho, Taehong and Peter Ladefoged. 1999. Variation and universals in VOT: Evidence from 18 languages. Journal of Phonetics 27: 207–229. Coates, Richard. 1980. Time in phonological representations.Phonetics 8: 1–20. de Jong, Kenneth J. and Bushra Zawaydeh. 2002. Comparing stress, lexical focus, and segmental focus: patterns of variation in Arabic vowel duration. Journal of Phonetics 30: 53–75. Elgendy, Ahmad M. 2001. Aspects of pharyngeal coarticulation. In Proceedings of the Institute of Phonetic Sciences, University of Amsterdam, vol. 24,197–199. Embarki, Mohamed. 2014. Evolution et conservatisme phonétiques dans le domaine arabe. Diachronica 31(1). Embarki, Mohamed and Christelle Dodane (eds.). 2011. La Coarticulation des indicesàla representation. Paris: L’Harmattan. Embarki, Mohamed, Christian Guilleminot, and Melissa Barkat-Defradas. 2006. Expansion nasale en arabe standard: Indices acoustiques d’une coarticulation anticipatoire. Revue Parole 39–40: 209–234. Embarki, Mohamed, Christian Guilleminot, Mohamed Yeou, and Sallal Al Maqtari. 2011a. Agression coarticulatoire des consonnes pharyngalisées dans les séquences VCV en arabe moderne et dialectal. In La Coarticulation des indices à la representation. Ed. Mohamed Embarki, and Christelle Dodane, 143–54. Paris: L’Harmattan. Embarki, Mohamed, Slim Ouni, Mohamed Yeou, Christian Guilleminot, and Sallal Al Maqtari. 2011b. Acoustic and electromagnetic articulographic study of pharyngealisation: Coarticulatory effects as an index of stylistic and regional variation in Arabic. In Instrumental studies in Arabic phonetics, ed. Zeki M. Hassan and Barry Heselwood, 193–215. Amsterdam: Benjamins. Ferguson, Charles A. 1956. The emphatic l in Arabic. Language 32: 446–452. Flemming, Edward S. 2001. Scalar and categorical phenomena in a unified model of phonetics and phonology. Phonology 18:7–44. Ghazeli, Salem. 1977.Back consonants and backing coarticulation in Arabic. PhD diss., University of Texas, Austin. ——. 1981. La coarticulation de l’emphase en arabe.Arabica 28: 251–277. Ghali, M. M. 1983. Pharyngeal articulation. Bulletin of the School of Oriental and African Studies 46:432–444. Gordon, Matthew. 2007. Functionalism in phonology. In The Cambridge handbook of phonology, ed. Paul de Lacy, 61–77. Cambridge, UK: Cambridge University Press. Hassan, Zeki M. 2003. Temporal compensation between vowel and consonant in Swedish and Arabic in sequences of CV:C and CVC: and the word overall duration. PHONUM 9: 45–48. Hassan, Zeki M., John Esling, Scott Moisik, and Lise Crevier-Buchman. 2011. Aryepiglottic trilled variants of /ʕ, ħ/ in Iraqi Arabic. In Proceedings of the 17th International Congress of Phonetic Sciences, Hong Kong, 831–834. http://www.icphs2011.hk. Hassan, Zeki M. and Barry Heselwood. 2011. Instrumental studies in Arabic phonetics. Amsterdam: John Benjamins. Heselwood, Barry. 2007. The “tight approximant” variant of the Arabic “ayn.” Journal of the International Phonetic Association 37: 1–32. Heselwood, Barry and Zeki Hassan. 2011. Introduction. In Instrumental studies in Arabic phonetics., ed. Zeki Hassan and Barry Heselwood, 1–25. Amsterdam: Benjamins. Jakobson, Roman. 1957. Mufaxxama—the emphatic phonemes in Arabic. In Studies presented to Joshua Whatmough, ed. E. Pulgram, 105– 115. The Hague: Mouton. Jomaa, Mounir. 1994. L’opposition de duré e vocalique en arabe: essai de typologie. Actes des XXèm es Journées d’Etudes sur la Parole (JEP) 20: 395–400. Jongman, Allard, Wendy Herd, Mohammad Al-Masri, Joan Sereno, and Sonja Combest. 2011. Acoustics and perception of emphasis in Urban Jordanian Arabic. Journal of Phonetics 39: 85–95. Kaye, Alan J. 2009. Arabic. In The world’s major languages, ed. Bernard Comrie,560–577. London: Routledge. Khattab, Ghada. 2007. A phonetic study of gemination in Lebanese Arabic. In Proceedings of 16th International Congress of Phonetic Sciences, Saarbrücken, 153–158. http://www.icphs2007.deand

Phonetics Kingston, John. 2007. The phonetics–phonology interface. In The Cambridge handbook of phonology, ed. Paul de Lacy, 401–434. Cambridge, UK: Cambridge University Press. Ladefoged, Peter. 2003. Phonetic data analysis: An introduction to fieldwork and instrumental techniques. Malden, MA: Blackwell. Ladefoged, Peter and Keith Johnson. [2001] 20102 . A course in phonetics. Boston: Wadsworth. Lahiri, A., L. Gewirth, and S. Blumstein. 1984. A reconsideration of acoustic invariance for place of articulation in diffuse stop consonants: Evidence from a cross-language study. Journal of the Acoustical Society of America 76: 391–404. Laufer, Asher. 1996. The common [ʕ] is an approximant and not a fricative.Journal of the International Phonetic Association 26: 113–117. Lehiste, Ilse. 1970. Suprasegmentals. Cambridge, MA: MIT Press. Lehiste, Ilse, Katherine Morton, and Marcel Tatham. 1973. An instrumental study of consonant gemination. Journal of Phonetics 1: 131–148. Lindblom, B. 1963. On vowel reduction. Report 29, Royal Institute of Technology, Stockholm, Speech Transmission Laboratory. Lisker, Leigh and Arthur Abramson. 1964. Cross-language study of voicing in initial stops: Acoustical measurements. Word 20: 384–422. Lodge, Ken. 2009. A critical introduction to phonetics. London: Continuum. Maddieson, Ian. 1984. Patterns of sounds. Oxford: Oxford University Press. Mitleb, Fares. 1984a. Voicing effects on vowel duration is not an absolute universal. Journal of Phonetics 12: 23–28. ——. 1984b. Vowel length contrast in Arabic and English: A spectrographic test. Journal of Phonetics 12: 229–235. Murtonen, A. 1966. The Semitic sibilants. Journal of Semitic Studies 11: 135–150. Newman, Daniel L. 2005. Contrastive analysis of the segments of French and Arabic. In Investigating Arabic: Current parameters in analysis and learning, ed. Alaa Elgibali, 185–220. Leiden: Brill. Owens, Jonathan. 2001. Arabic sociolinguistics. Arabica 48: 419–469. ——. [2006] 20092 . A linguistic history of Arabic. Oxford: Oxford University Press. Pickett, E. R., S. Blumstein, and M. Burton. 1999. Effects of speaking rate on the singleton/ geminate consonant contrast in Italian.Phonetica 56: 135–157. Port, R. F., S. Al-Ani, S. and S. Maeda. 1980. Temporal compensation and universal phonetics. Phonetica 37: 235–252. Roman, André. 1977. Les zones d’articulation de la koinè arabe d’après l’enseignement d’Al-Halīl [Al-Khalil]. Arabica 24: 58–65. Roman, André. 1981. De la langue arabe comme un modèle général de la formation des langues sémitiques et de leur évolution. Arabica 28: 127–161. Ryding, Karin C. 2005. A reference grammar of modern standard Arabic. Cambridge, UK: Cambridge University Press. Sara, Solomon. 2007. Sībawayh on ʔimāla (Inclination). Text, translation, notes and analysis. Edinburgh: Edinburgh University Press. Scobbie, Jim. 2007. Interface and overlap in phonetics and phonology. In The Oxford handbook of linguistic interfaces, ed. Gillian Ramchand and C. Reiss, 17–52. Oxford: Oxford University Press. Shahin, Kimary N. 1997. Postvelar harmony: An examination of its bases and crosslinguistic Variation. PhD diss., University of British Columbia, Vancouver. Watson, Janet. 2002. The phonology and morphology of Arabic. Oxford: Oxford University Press. Yeni-Komshian, Grace H., Alfonso Caramazza, and Malcom Preston. 1977. A study of voicing in Lebanese Arabic. Journal of Phonetics 5: 35– 48. Yeou, Mohamed. 1997. Locus equations and the degree of coarticulation of Arabic consonants. Phonetica 54: 187–202. Yeou, Mohamed and Shinji Maeda. 2011. Airflow and acoustic modelling of pharyngeal and uvular consonants in Moroccan Arabic. In Instrumental studies in Arabic phonetics, ed. Zeki M. Hassan and Barry Heselwood, 141–162. Amsterdam: Benjamins. Zawaydeh, Bushra A. 2003. The interaction of the phonetics and phonology of gutturals. In Phonetic interpretation. In Papers in Laboratory Phonology, vol. 6, ed. John Local, Richard Ogden, and Rosalind Temple, 279–292. Cambridge, UK: Cambridge University Press. Zawaydeh, Bushra A. and Kenneth de Jong. 1999. Stress, phonological focus, quantity, and voicing effects on vowel duration in Ammani Arabic. In Proceedings of the 14th International Congress of Phonetic Sciences, Vol. 1,451–454. Berkeley: University of California Press. Zeroual, Chakir, Philip Hoole, Suzanne Fuchs, and John Esling. 2007. EMA study of the coronal emphatic and non-emphatic plosive consonants of Moroccan Arabic. In Proceedings of 16th International Congress of Phonetic Sciences, Saarbrücken, 397–400. Zeroual, Chakir, Philip Hoole, Suzanne Fuchs, and John Esling. 2011. Contraintes articulatoires et acoustico-perceptives liées à la production

Phonetics de /k/ emphatisé e en arabe marocain. In La Coarticulation des indicesàla representation, ed. Mohamed Embarki, and Christelle Dodane, 227– 40. Paris: L’Harmattan.

Notes: (1) The articulatory descriptions of the Arabic grammarians always conventionally started from the back of the articulatory tract and worked their way forward to the lips. (2 ) The UCLA Phonological Segment Inventory Database was developed by Maddieson (1984). In its initial version, the database contained phonological information on 317 languages, representing all of the world’s language families. An augmented version with 451 languages was published in 1991. Mohamed Embarki Mohamed Embarki, University of Besançon

Phonology Sam Hellmuth The Oxford Handbook of Arabic Linguistics Edited by Jonathan Owens Print Publication Date: Sep 2013 Online Publication Date: Dec 2013

Subject: Linguistics, Phonetics and Phonology, Languages by Region DOI: 10.1093/oxfordhb/9780199764136.013.0003

Abstract and Keywords Phonology is the study of systematic patterning in the distribution and realization of speech sounds within and across language varieties. Arabic phonology features heavily in the work of the Arab grammarians, most notably in the Kitaab of Sibawayh. Sibawayh provides phonetic descriptions of the articulation of individual speech sounds, which are accompanied by an analysis of the patterning of sounds in Arabic, which is indisputably phonological in nature. This article sets out five important strands of phonological research on Arabic, taking in work on the language-particular phonological properties of Arabic as well as research that exploits fine-grained variation among spoken varieties of Arabic for theoretical gain. The discussion is structured to move from segmental phonology (the properties of individual speech sounds) to suprasegmental phonology (the properties of larger domains such as the syllable, word, or phrase). Keywords: Arabic phonology, speech sounds, language, Arab grammarians, Kitab of Al-Sibawayh, articulation, phonological research, segmental phonology, suprasegmental phonology

3.1 Introduction PHONOLOGY is the study of systematic patterning in the distribution and realization of speech sounds within and across

language varieties. The phonology of Arabic features heavily in the work of the Arab grammarians, most notably in Al-Sībawayh’s Kitab (Harun 1983). Sībawayh provides phonetic descriptions of the articulation of individual speech sounds [Embarki, “Phonetics”; Sara, “Classical Lexicography”], which are accompanied by an analysis of the patterning of sounds in Arabic that is indisputably phonological in nature. Sībawayh was thus among the first in a long line of phonologists to work on the phonology of Arabic, and in this chapter we set out some of the key strands in that research, highlighting the contribution made by the study of Arabic to our understanding of phonology itself. It is relevant to ask in what sense Sībawayh is doing phonology, as it is understood today. For one thing, Sībawayh takes pains to carefully identify which consonants and vowels made up the phonological inventory of Arabic, classifying individual sounds ([ħuruuf] “letters”) as either basic or “derived” (Al-Nassir 1993: 17–20). Some recent research in this area is discussed in Chapter 1 (Section 3.2.1). Sībawayh quite clearly uses the concept of underlying forms, from which positional variants are derived in an orderly and predictable fashion, and carefully describes the operation of a number of named phonological processes. For example, in the case of imaalah “inclination” (raising/fronting of [a:]/[a]), he describes in detail the range of contexts that do or do not trigger the process and formulates it explicitly as derivational and iterative: a long vowel [a:] will raise/front (to [ie]; Owens 2006) when the preceding syllable contains a short front high vowel [i] kasrah, and the raised/fronted long vowel will itself then trigger imaalah in a following syllable (Al-Nassir 1993: 91–103).

Phonology The notion of markedness is applied indirectly in the work of the grammarians in all areas of the grammar (Owens 1988), and, in phonology, Sībawayh identifies ṣaħiiħ“strong” and muʕtall “weak” elements in many of the phonological oppositions that he proposes. For example, he distinguishes sounds as either mutaħarrik (CV, followed by a short vowel) or saakin (C, closing a syllable), and argues that the mutaħarrik is the strong member of the pair. Identification of syllable-initial (onset) position as strong and syllable-final (coda) position as weak is argued to account for the differing range of phenomena observed in each position (Al-Nassir 1993: 111), mirroring contemporary approaches to onset-coda asymmetries in general (e.g., Lombardi 1999) and directly matching claims made about the underlying syllabic structure of Arabic (Lowenstamm 1996; see Section 3.2.3). Other phonological phenomena discussed by Sībawayh range from the optimal size of the verbal root (3–5 ħuruuf) (Al-Nassir 1993: 26) to variation in the realization of particular sounds or lexical items across definable groups of speakers (Al-Nassir 1993: 116–117; cf. Owens 2006: ch.7). We even find discussion of the potential phonological effects of word frequency (“they dare change what occurs more frequently in their speech”; Al-Nassir 1993: 117) that prefigure exemplar-based approaches to phonology (see Section 3.2.3). A pattern we see in Sībawayh, repeated throughout phonological work on Arabic, is that Arabic proves to be an interesting object of study in two ways: (1) because the language has phonological features that are themselves typologically relatively unusual; and (2) because its phonological features vary, minimally but systematically, across different varieties of the language. The Arab grammarians would not have sought to compare the particular properties of Arabic with those of other languages; in the context of modern linguistics, however, the typologically unusual properties of Arabic present a genuine challenge to theories that have oft en been shaped for the most part by work on Indo-European languages. In turn, the fine-grained variation observed among spoken Arabic dialects has proved a rich seam of research, in particular in generative phonology, that seeks to model surface variation in terms of a limited set of underlying structural differences, whether parameters or ranked constraints. In Broselow’s (1992: 7) words, “The dialects of Arabic provide an ideal testing ground …, since most of the dialects are similar enough to provide a basis for meaningful comparison, but taken as a whole they exhibit a wide range of variation.” In the main body of this chapter, we set out five important strands of phonological research on Arabic, taking in work on the language-particular phonological properties of Arabic as well as research that exploits fine-grained variation among spoken varieties of Arabic for theoretical gain. The discussion is structured to move from segmental phonology (the properties of individual speech sounds) to suprasegmental phonology (the properties of larger domains such as the syllable, word, or phrase).

3.2 State of the Art

3.2.1 Phonotactic Restrictions on Consonant Co-occurrence in Arabic Verbal Roots The Arabic lexicon displays nonconcatenative (i.e., discontinuous) morphology. This has led some authors to analyze Arabic words in terms of a triliteral (three-term) consonantal root, with words generated from the lexical root by internal rearranging of the sequence of consonants and vowels (see [Ratcliffe, “Morphology”] for competing views of the structure of Arabic morphology): (1) katab

kutib

kitaːb

kutub

“he wrote”

“it was written”

“book”

“books”

In McCarthy’s (1979) root + template analysis of Arabic morphophonology, consonantal features are represented on a separate tier (or plane) from vocalic features, and different grammatical classes are generated using fixed templates that define the sequence of consonantal and vocalic positions:

Phonology

(2)

Since the majority of Arabic roots are composed of three consonants, one might expect the consonants in any particular lexical root to be selected at random from the consonantal inventory of Arabic and freely combined. This is not the case, however, and, apart from some mention by Arab grammarians of which consonants resist cooccurrence,1Cantineau (1946: 133–136) is first to note that not all possible combinations of consonants seem in fact to be observed. Building on this, Greenberg (1950) subjects the possible combinations of consonants appearing in initial (C1), medial (C2), and final (C3) position to a systematic quantitative survey, in a corpus of 3775 Arabic verbal roots. The data were taken from two 19th-century Arabic dictionaries (by Western authors), with qualitative comparison to the facts of other languages within the Semitic family. For Arabic (and for Semitic in general), Greenberg (1950) establishes two key patterns: (3) Complete absence of roots with adjacent identical consonants in C1–C2 position *〈mmd〉, contrasting with no restriction on adjacent identical consonants in C2–C3 position 〈mdd〉2 (4) Varying degrees of restriction on the occurrence of homorganic consonants, which share place of articulation, within a root (C1–C2, C2–C3, or C1–C3) Working from the generalizations observed in ( 4), Greenberg motivates a grouping of the Arabic consonantal inventory into four “sections,” or natural classes, shown in (5).3 Within a single verbal root, Arabic consonants are seen to freely occur with those in other sections, which have a different place of articulation, but are subject to restrictions on co-occurrence with members of the same section. Although examples of verbal roots containing two consonants from the same section can be found, the number of such roots is relatively small.4 (5) Back

Liquids

Front

Labial

x Ɣ ħ ʕ h ʔ k g q

l r n

θ ð t d ṭ ḍ s z ṣ ẓ∫

b f m

The generalizations expressed in (3) and (4) have each inspired a large body of research, which has proven significant for an accurate understanding of the phonology not only of Arabic but also of human language in general. McCarthy (1979, 1981) proposes an explanation of the asymmetry inherent in (3) that combines a root + template morphological analysis of Arabic with the insights of autosegmental phonology (Goldsmith 1976, 1981). In autosegmental phonology, individual speech sounds are represented as (bundles of) features associated with syllabic structure. The surface realization of an underlying feature is determined by how it is linked to the prosodic structure. For example, a set of vocalic features will be realized as a short versus long vowel, depending on whether they are linked to one syllabic position only or are allowed to “spread” to two syllabic positions. For McCarthy (1979, 1981), the ill-formedness of identical consonants in C1–C2 position versus the wellformedness of identical consonants in C2–C3 is due to two facts: (1) the strict observance in Arabic of the obligatory contour principle (OCP), which places a ban on adjacent identical segments; and (2) a requirement in Arabic that features spread autosegmentally only from left to right.5 Underlyingly a verbal root such as 〈mdd〉 thus comprises two feature bundles only, representing 〈m〉 and 〈d〉. The features of 〈d〉 spread left to right, filling the third syllabic consonantal position in the template and yielding surface forms analyzable as triliteral 〈mdd〉. Since spreading proceeds only left to right, an empty syllabic position cannot be filled by left ward spreading of 〈m〉, resulting in an absence of verbal roots such as *〈mmd〉. The analysis is illustrated in (6):

Phonology

(6)

An analysis in terms of underlying forms is warranted because there is no ban on adjacent identical consonants in derived, morphologically complex forms; compare the lack of a verbal root *〈ttk〉 with permitted forms such as [tatakallam] “you conversed” (McCarthy 1986: 209). Bohas (1997) proposes a more radical approach (cf. earlier work by Voigt 1988; Ehret 1989), suggesting that a biliteral root morpheme, called the “etymon,” underlies all triliteral roots, not only those with identical C2–C3; each proposed etymon captures a regular form-meaning correspondence between a consonant pair and a semantic field. Other proposals that extend from the templatic nature of Arabic morphology include the claim that Arabic is a language in which all syllables are comprised of sequences of CV pairs (see Section 3.2.3). The analysis in (6) has been challenged due to a nontrivial assumption it makes about the nature of phonology: namely, that some aspects of phonological knowledge—such as the ban in Arabic on adjacent, identical segments in C1–C2 position but not in C2–C3—are encoded as restrictions on possible underlying lexical representations (i.e., as morpheme structure constraints). McCarthy (1998, 2005) argues against his own earlier analysis for this reason. A core assumption of optimality theory (OT; Prince and Smolensky 2004) is that all grammar, including phonology, determines surface realizations (“outputs”) only. This claim is formulated in OT as the “richness of the base”: the phonological grammar must generate all and only those forms observed on the surface of the language without stipulating restrictions on possible inputs to the grammar (i.e., on the properties of the lexicon). Greenberg’s asymmetry in (3) represents a serious challenge to theories of this kind. Solutions to this problem, in the OT literature, mostly appeal to the notion of paradigm uniformity, whereby surface forms are preferred if they bear structural resemblance to other surface forms in the same morphological paradigm (McCarthy 1998, 2005; Gafos 1998, 2001, 2003; Rose 2000). In some of these analyses (Gafos 1998; Rose 2000), the doubled final consonant in the surface form of a root like [madda] “he stretched” results from reduplication and is permitted because, in the OT framework, the OCP constraint can be outranked by other competing constraints. In another approach (Gafos 2001, 2003), the idea of separation of consonants and vowels onto different tiers is rejected, and the underlying form of all Arabic verbs is proposed to be a CVCC “stem,” such as /madd/.6 What about Greenberg’s other generalization in (4) that although there is no complete ban on adjacent homorganic consonants in C2–C3 position their occurrence is highly restricted? In a series of papers, McCarthy (1986, 1988, 1994) extended an OCP analysis to these facts also. In contrast to the categorical nature of the asymmetry in (3), the restrictions described in (4) are gradient in nature: the key fact to explain is why there are some roots containing homorganic consonants rather than none. McCarthy adopts feature geometry (Clements 1985; McCarthy 1988), in which the bundle of features that map onto an individual speech sound is represented as a hierarchically grouped tree structure rather than as an unordered matrix. For example, laryngeal features (e.g., [voice] or [spread glottis]) appear under a different node from place features (e.g., [labial], [coronal], or [dorsal]). McCarthy (1988) extends this idea by representing individual place features on separate tiers, as in (7). If the OCP (which bans adjacent identical elements) operates on individual tiers, then the place restrictions are explained.

(7)

This analysis predicts a complete ban on homorganic consonants within a single verbal root but cannot account for the small but nontrivial number of verbal roots that do contain homorganic consonants and that are too

Phonology numerous to be treated as exceptions. An additional complication is that within the larger classes of consonants, such as the coronals (Greenberg’s “front” section), consonants are observed to co-occur somewhat more freely than those within the smaller natural classes (Padgett 1995). A competing approach (Pierrehumbert 1993; Frisch et al. 2004) addresses this problem by taking the gradient tendency in (4) as its starting point. The approach is based on a quantitative, probabilistic model of phonology, contrasting fundamentally with the categorical models of generative grammar embodied in the work of McCarthy and Gafos. The analysis in Frisch et al. (2004) is based on a corpus of 2674 verbal roots from a contemporary dictionary of Modern Standard Arabic (Cowan 1979). The distribution of roots is analyzed in terms of the ratio of observed consonant combinations to those that would be expected if consonants were allowed to occur freely in particular positions of the root (Pierrehumbert 1993). For example, given the frequency of occurrence of [d] in the C1 position and of [t] in the C2 position, Frisch et al. expect there to be 2.3 roots of the form 〈d t X〉 (where X is any consonant); in fact, there are no such roots, and the observed–expected (O–E) ratio is thus “0,” which the authors describe as “the strongest degree of under-representation” (Frisch et al. 2004: 185). In contrast, there are four roots of the form 〈d g X〉, where only 3.3 such roots are expected to occur, giving an O–E score of 1.21, a case of overrepresentation. Frisch et al. (2004) argue that an adequate analysis of the gradient generalization described in (4) must capture both under- and over-representation of particular consonant co-occurrences. They propose that a principle of “similarity avoidance” operates productively in the Arabic lexicon. To model this principle, they calculate the degree of similarity between two consonants in the Arabic inventory in terms of the number of natural classes in which the two consonants share and do not share membership.7 The metric thus encodes not only raw phonetic similarity (number of shared features) but also the relative size of the consonantal inventory at each place of articulation (cf. Padgett 1995). The analysis represents a hybrid approach in which the phonological knowledge of native speakers is claimed to combine traditional phonological factors (here, natural classes defined by place of articulation) with frequency effects: “the native speaker knows an abstract but gradient OCP-Place constraint.… based on generalization over the statistical patterns found in the lexicon” (Frisch et al. 2004: 216). Frisch et al. suggest that a general cross-linguistic tendency to avoid repetition of similar sounds, based on a preference for maximally salient adjacent sounds (Boersma 1998), is heightened in Arabic since consonants are adjacent in lexical representations if a root-based analysis of Arabic morphology is assumed. To prevent gradient similarity avoidance from applying in the case of roots with identical consonants in C2–C3—in the case of the generalization in (3)—Frisch et al. (2004) acknowledge that they must adopt some form of “categorical override.” They noted that prior approaches to (3), whether autosegmental (McCarthy 1979, 1981, 1986) or paradigm-based (Gafos 2003), both treat the double consonant in a root like 〈mdd〉 as licensed by the appearance of a singleton consonant in a related form: the related form in the autosegmental approach is the underlying form (/md/); in a paradigm-based approach the related form is another surface realization of the same root ([madda] “he stretched” ∼ [madadtu] “I stretched”). Frisch et al. thus proposed that some kind of “related form override” must account for C2–C3 identical roots. Although Frisch et al. (2004) are formally agnostic about how related forms license geminate roots in Arabic, they cite a range of behavioral data that they suggest supports an analysis in terms of roots rather than stems. Frisch and Zawaydeh (2001) asked native speakers of Arabic to judge the relative well-formedness of nonsense words; roots with identical C1– C2 such as [tatafa] were universally rejected, but roots containing homorganic consonants in other positions and combinations, such as in C1–C2 in the nonsense word [tasafa], display the same variant restrictions as observed in the general Arabic lexicon. For Frisch et al., this confirms that similarity avoidance is a productive part of the phonological grammar for Arabic. Similarly, Davis and Zawaydeh (2001) argue that truncation patterns in Arabic hypocoristics (nicknames) display effects that can be explained only by appealing to the consonantal root in lexical representation. Finally, evidence from the speech errors of an aphasia patient who is bilingual in French and Arabic (Prunet et al. 2000; Idrissi et al. 2008) shows different patterns of errors in the two languages that are consistent with vowels being present in lexical representations in French but absent from lexical representations in Arabic. In contrast to Frisch et al., Gafos (2001) argue that psycholinguistic data of this kind provide evidence of root-based language processing (only) but maintain the claim that the grammar of Arabic is stem based (see Ratcliffe [“Morphology”]). In conclusion, then, the patterning of consonants in Arabic verbal roots show both categorical and gradient effects

Phonology (in the generalizations set out in (3) and (4), respectively), and theories of phonological knowledge need to be able to account for both types of effects. Arguably, therefore, a hybrid approach of some kind (e.g., Pierrehumbert 2006) is to be favored. It is worth noting, however, that almost all of the previously discussed analyses treat the distribution of consonants in verbal roots only and in Modern Standard Arabic. Studies of patterning of consonants in the nominal system (Faust and Hever 2010) or in spoken dialects are even more rare. An exception is Herzallah’s (1990) study of phonetically velar sounds [k x Ɣ] in Palestinian Arabic, which behave phonologically as uvular in their co-occurrence restrictions (cited in Davis 1995). Further, and broader, empirical work in this area may yet reveal a more complete picture for which phonological theory must account.

3.2.2 Postvelar Consonants and Emphasis All varieties of Arabic share the property of having a small vowel inventory and a relatively large consonantal inventory (cf. Maddieson 2011). In particular, the consonant inventory has a large proportion of “guttural” consonants with postvelar place of articulation [q χ ʁ ħ ʕ h ʔ] and a set of “emphatic” coronal consonants that displays postvelar secondary articulation [ṭ ḍ ṣ ẓ] and contrasts with plain counterparts [t d s ð].8 The postvelar(ized) segments influence the phonetic realization of neighboring segments, both vowels and consonants. The most salient effect is backing (F2 lowering) of immediately adjacent vowels: compare [taːb] “he repented” with [ṭaːb] “he recovered.” In some dialects, other consonants, such as [r l m b], can also trigger the same effect in certain contexts. The domain of this “emphasis spread” is nonlocal in character, reaching the entire word in some spoken dialects and even beyond it into adjacent words in some cases. Emphasis spread is typologically unusual and is shared with only a few other language families.9 Work on emphasis represents a large proportion of both past and current research in Arabic phonology. Although the phonetics and phonology of emphatics are inextricably linked, the discussion here focuses on phonological issues.10 The key phonological issues that have exercised the research community with regard to emphasis include what the nature—and name—of the phonological feature used to represent emphasis should be, whether this feature is a property of individual segments or larger domains, and how to explain the differing domain and directionality of emphasis spread observed in different dialects of Arabic (see Bellem 2007: 26–33 for a summary). The Arab grammarians use a range of terms to describe the properties and effects of the emphatic and guttural consonants. Sībawayh describes the emphatic coronals [ṭ ḍ ṣ ẓ] as muṭbaq “covered, enclosed,” contrasting with plain coronals such as [t d s], which are munfatiħ “open” (Al-Nassir 1993). The muṭbaq consonants are characterized by raising of the tongue dorsum toward al-ħanak al-aʕlaa “the roof of the mouth,” with no mention of a role for the pharynx in the articulation (ibid.). Sībawayh also describes a class of seven consonants—the four muṭbaq emphatic coronals plus the three uvulars [q χ ʁ]—that share the feature of being mustaʕlin “elevated” and are identified as a natural class because they all block imaalah (see Section 3.1). Ibn Jinnī contrasts the seven mustaʕlin consonants with all other consonants, which are munxafiḍ “lowered” (ibid.). Sībawayh uses the term mufaxxam “made grand” to describe a raised and backed realization of /a:/ alif and /a/ fatħa, as [aː] and [a], respectively, in the context of a mustaʕlin consonant (ibid., 103).11 Sībawayh thus does not categorize the pharyngeals [ħ ʕ] as eliciting tafxiim. Watson (2002) analyzes “pharyngeal” gutturals and “pharyngealized” emphatics, in Cairene and San’aani, as a single group (cf. Broselow 1976), characterized phonologically by the feature [guttural]12 in a nonprimary position in the feature geometry. She sees the gutturals as pharyngealized counterparts of nonpharyngealized sounds, just as emphatic coronals are pharyngealized counterparts of plain coronals, as illustrated in (8) (ibid., 42–44). Watson describes emphasis spread as pharyngealization and notes that it is accompanied in some dialects by varying degrees of labialization:

Phonology

(8)

Watson (2002) analyzes pharyngeal and pharyngealized sounds with the same feature specification (nonprimary [guttural]) but argues that both the phonetic realization and the extent of pharyngealization will differ in each subgroup, due to the difference in the relationship between the primary and nonprimary features in each case. In pharyngealized coronals, the tongue dorsum realizes nonprimary [guttural] at the same time as the tongue tip/blade realizes primary [coronal]; the tongue is thus under tension, and it takes time for the articulators to move back to nonemphatic settings. Th is results in greater, more nonlocal spread of pharyngealization from the emphatic coronals than from the gutturals (Watson 2002: 273). Watson’s featural representation can also explain the realization of Classical Arabic /q/ in spoken dialects: in Cairene, */q/ lost primary [dorsal] and nonprimary [guttural] was promoted, yielding /ʔ/ (Watson 2002: 45 n. 18). In contrast, McCarthy (1994: 202–218) argues that the gutturals (Watson’s pharyngeal) must receive a different featural analysis from the emphatic coronals (Watson’s pharyngealized), because even though there is evidence for grouping gutturals and emphatic coronals together, in that they both block imaalah (ibid., 218–220), there is also ample evidence that the gutturals form a natural class to the exclusion of the emphatic coronals. Key evidence in Arabic for the natural class of gutturals comes from co-occurrence restrictions within lexical roots (see Section 3.2.1) and from vowel lowering and metathesis (gahawah syndrome) in the vicinity of gutturals (ibid., 202– 218). McCarthy (1994: 221) thus proposes a distinct feature representation for the various relevant classes of sound, as in (9). He uses the feature [pharyngeal], which is in most respects parallel to Watson’s (2002) [guttural].13

(9)

Other featural analyses of emphasis include Davis’s (1995) treatment of two varieties of Palestinian Arabic (PA), in which a pharyngeal node is argued to host two features: [RTR] “retracted tongue root”; and [CP] “constricted pharynx.” Davis argues on articulatory grounds that [RTR] is active in the uvular gutturals [χʁ], whereas [CP] is active in the pharyngeals [ħ ʕ]; the emphatic coronals have nonprimary [RTR]. In a study that compares data in PA and St’àt’imcets Salish, Shahin (2003) argues for a cross-linguistic distinction between uvularization (spreading of nonprimary RTR, in PA from emphatic coronals) and pharyngealization (spreading of primary RTR, in PA from gutturals). In contrast to the articulatorily defined features described thus far, Bellem (2007) analyzes emphasis using a system of psychoacoustically defined features (after Harris and Lindsey 1995) that includes just three resonance features: A, I, and U. For Bellem, A spreading (emphasis spread) competes in the phonology with I spreading ( imaalah) and U spreading (labialization). A key area of variation across different varieties of Arabic is in the domain of emphasis spread. All of the spoken

Phonology dialects studied thus far appear to share the asymmetry that leftward spreading is less restricted than rightward spreading and that spreading is greater from emphatic coronals than from gutturals. Spreading may be blocked by intervening palatal vowels or consonants in some dialects in one or both directions. Davis (1995) analyzes blocking of RTR spreading by high vowels and palatal consonants by means of a rule grounded in the natural antagonism between [RTR] (retracted) and a [+high] (advanced) tongue positions. Bellem (2007) suggests that variation in the domain of emphasis spread is better analyzed as blocking by intervening emphatics/gutturals of an active imaalah (palatalization) process than as blocking of tafxiim (emphasis) by intervening palatals. For most authors (e.g., McCarthy 1994; Watson 2002), emphasis spread is viewed as autosegmental spreading of a feature from a consonant to adjacent vowels and consonants. However, if separate V/C tiers are assumed, as in (2), the exact mechanism by which a spreading feature is able to spread not only to adjacent consonants but also to vowels is not fully spelled out. Working in OT, Shahin (2003) analyzes spreading using alignment constraints on surface (output) realizations. To a limited extent, this echoes Beeston’s (1970: 19) view that iṭbaaq is a prosody (in the Firthian sense14 )rather than a “component of the four velarized alveolar consonants” and, thus, a property of domains rather than of segments. This brief summary includes only the most influential or innovative work on emphasis in Arabic phonology. A large number of studies of emphasis in individual spoken dialects exist (Bellem 2007 provides a recent survey), and only a few studies have attempted to analyze the patterns of emphasis across dialects or across Semitic in general (Hayward and Hayward 1989; McCarthy 1994; Bellem 2007). In particular, Bellem offers an analysis linking the relative strength of features in different dialects, and thus their capacity to participate in active feature spreading processes such as emphasis, with more general phonological properties of each dialect such as the number of laryngeal contrasts in the phonological inventory. It is comparative work of this kind that is most likely to reveal the full range of phonological representations underlying the surface phenomena collectively known as emphasis.

3.2.3 Syllabification and Syllable Structure Cross-dialectal variation in syllabification across Arabic dialects is perhaps most clearly exemplified by the differing realization of sequences of three consonants (CCC). Such sequences commonly occur when a consonant-initial suffix such as [lu] “to him” is added to a CC final word, such as [qult] “I said.”15 Although a small number of dialects tolerate a surface CCC cluster (e.g., Moroccan [qultlu]), most dialects insert an epenthetic vowel to break up the CCC sequence, and dialects vary as to where the vowel is placed, CvCC versus CCvC: Iraqi [gəlitlu] but Cairene [ʔultilu]. The pattern is robust and is also reflected in the realization of CC-initial loanwords: “Fred” is [fi.rɛd] in Iraqi but [if.rɛd] in Cairene (Broselow 1983). The basic two-way distinction gives rise to an informal nomenclature of “gəlit” versus “qəltu” dialects (Blanc 1964), or, more recently, of VC versus CV dialects (Kiparsky 2003). Two competing explanations of the VC ∼ CV epenthesis facts emerged in the 1980s and 1990s. Ito (1989) proposed a directional difference between the dialects: Iraqi syllabifies from right to left, while Cairene syllabifies from left to right. Mester and Padgett (1994) offer an OT implementation of this directional approach but also notes its limitations, not least in explaining different syllabification patterns in word-initial CC clusters (as in the realizations of “Fred” already noted). In contrast, Broselow (1992) proposes that a structural parameter determines the syllabic affiliation of stray consonants in different dialects, building on an earlier suggestion by Selkirk (1981) that “stray” consonants can be syllabified as “degenerate” (vowelless) syllables. Broselow suggests that Arabic dialects vary in how they treat the “stray” third consonant in a CCC sequence: in Iraqi, a stray consonant is syllabified into a (temporarily) vowelless rhyme; in Cairene, a vowelless rhyme is not permitted so the stray consonant is syllabified as an onset. Epenthesis fills in the empty vowel positions, inserting a vowel before a rhymal consonant and after an onset consonant. Broselow further argued that parallel variation across CV ∼ VC dialects in the treatment of word-internal CVVC syllables is due to a similar structural parameter. In VC dialects, word-internal CVVC syllables are tolerated ([baabha] “her door”), despite breaching a more general preference in Arabic for bimoraic syllables; 16 in CV dialects, a word-internal CVVC sequence is not permitted at all and is instead repaired by closed syllable shortening (CSS; [bab-ha] “her door”). Broselow proposed that in VC dialects the final consonant in a CVVC word-internal syllable is incorporated into the preceding rhyme (satisfying the general preference for bimoraicity), by a process of Adjunction-to-Mora. Th is process applies exactly in dialects (the VC dialects) that permit syllabification of the stray

Phonology consonant in a CCC sequence into the rhyme. Kiparsky (2003) reframes Broselow’s structural analysis as mora licensing. Dialects vary in whether they permit semisyllables, which contain a mora that is unlicensed (i.e., unaffiliated to any syllabic position, neither onset nor rhyme), and, if so, at what level of representation the unlicensed material is permitted. Kiparsky maintains a basic two-way divide, grouping C (Moroccan) dialects with VC (Iraqi) and contrasting them to CV dialects (e.g., Cairene).17 The broad distinction is that VC dialects permit unlicensed moras, whereas CV dialects do not; the distinction between VC and C dialects is that VC dialects permit unlicensed moras at the lexical level but forbid them at postlexical level, whereas in C dialects unlicensed moras are permitted at all levels of representation, both lexical and postlexical. Kiparsky’s account is formulated within OT in terms of ranked constraints: in VC dialects the constraint LICENSEμ (which requires a mora to be affiliated to syllabic structure) is low ranked, whereas in CV dialects it is highly ranked. Kiparsky assumes a stratified model of OT, allowing the VC ∼ C dialects distinction to be modeled as promotion of the constraint LICENSEμ to a higher ranked position at the postlexical level in VC dialects (only). Kiparsky (2003) suggests that variation in the position of the epenthetic vowel in CCC sequences covaries not just with availability of CSS (found in CV but not in VC, as argued by Broselow) but also with variation in a range of other syllabification phenomena, including metathesis (found in VC but not CV dialects) and the distribution of CC clusters (which occur only word finally in CV dialects and only word initially in VC dialects). Working from a typologically enlarged data set, Watson (2007) argues that the covariance between epenthesis and other syllabification patterns is not so clear-cut. The match is quite good for VC dialects, but a significant subset of CV dialects (as classified by epenthesis in CCC sequences) turns out to behave rather more like VC dialects, for example, by tolerating word-internal CVVC syllables. Watson thus argues for a further dialect group, the “Cv” dialects,18 and incorporates Broselow’s adjunction-by-mora parameter into Kiparsky’s OT account as a NOSHAREDMORA constraint. The constraint is low ranked in VC and Cv dialects, explaining shared patterns of behavior across the two groups. A completely different approach to syllabification in Arabic, which happens to share the notion of vowelless syllables with Selkirk (1981), developed out of attempts during the 1990s to capture all and only the observed range of template shapes in Arabic verbal and nominal forms (Guerssel and Lowenstamm 1990; Idrissi 1997). In the resulting “CV-only” analysis (Lowenstamm 1996, 2003),19 no consonants are syllabified as codas in Arabic at all, and instead every syllable-final consonant is analyzed as the onset of a vowelless syllable (cf. Yoshida 1993; Bellem 2007). The somewhat atypical syllabification patterns observed in Moroccan Arabic (MA) have inspired their own strand of research. Along with other Maghreb dialects, MA permits a wider range of word-initial onset clusters than observed in other Arabic dialects, such as [kteb] “he wrote,” [glih] “he grilled,” and [qleb] “he knocked over” (Gafos et al. 2011: 30). These are analyzed as branching (complex) onsets by some authors (e.g., Benkirane 1998) but more commonly as sequences of simplex onsets separated by empty vocalic positions in both “standard” moraic syllable theory (Kiparsky 2003) and CV-only approaches (Boudlal 2001). Recent work by Shaw et al. (2009) and Gafos et al. (2010, 2011) within the broad model of articulatory phonology (Browman and Goldstein 1986) proposes a model of the mapping between syllabic structure and articulatory evidence (the fine-grained temporal alignment of the articulatory gestures of the two consonants) that favors analysis of MA word-initial clusters as sequences of simplex onsets. The authors remain agnostic as to which phonological representation of such sequences best matches the articulatory facts (though a formal analysis is proposed by Gafos 2002, 2006), but future work of this kind in MA and in other dialects may shed further light on the typology of syllabification in Arabic. Finally, recent work on variation in the rhythmic properties of different Arabic dialects suggests that the typology of variation in syllabification across all dialects may prove to be even more fine-grained. Work on cross-linguistic rhythmic typology has shown that a two-way divide between “stress-timed” and “syllable-timed” languages oversimplifies (Roach 1982; Nolan and Asu 2009). Instead, there appears to be a continuum of rhythmic variation across languages, arising from independent variables affecting syllabification, including incidence of vowel reduction and the syllabification phenomena discussed in the VC ∼ CV literature outlined already. This rhythmic continuum is observed cross-dialectally in Arabic; although all dialects are stress timed, a comparison across six dialects displays that they move from more to less stress timed as one travels west to east from Morocco to the Levant (Ghazali et al. 2002, 2007). This surface variation correlates with the permitted syllabification patterns and syllable types observed in different dialects (Hamdi et al. 2005). Further investigation of the rhythmic properties of a

Phonology wider range of dialects, employing the most stable rhythm metrics (Wiget et al. 2010) and perhaps also incorporating a survey of the incidence of vowel reduction (cf. Cantineau’s differential vs. nondifferential parameter20 ), may reveal new avenues of research on Arabic syllabic structure and syllabification.

3.2.4 Word Stress and Metrical Theory Arguably the most important contribution made by the study of Arabic dialectal variation to the advancement of phonological theory has been in the area of metrical phonology, which seeks to account for the position of word stress in words of different syllabic structures. The position of stress is predictable in all Arabic dialects and can usually be reduced to a simple stress assignment algorithm. Although word stress received no attention in the work of the Arabic grammarians, probably because stress assignment is largely predictable (Watson 2011), there is a very rich body of research on Arabic word stress in contemporary linguistics. This has been fueled by an ample supply of data, in the form of good descriptions of the word stress patterns of a wide range of dialects (dating from the early 20th century onward), and by theoretical advances, which have led to continual reanalysis in the field, with some of the advances prompted by the facts of particular Arabic dialects. Watson provides a thorough survey of this literature and points out that much of the interest in the study of Arabic word stress lies in the fact that the surface stress patterns of Arabic dialects share many key properties but vary nontrivially in others (see also Kager 2009). Stress assignment in all Arabic dialects is sensitive to syllable weight, that is, quantity sensitive. The basic stress algorithm for all Arabic dialects assigns stress to a final super-heavy syllable (CVVC or CVCC) or otherwise to a nonfinal (usually penultimate) Table 3.1 Stress assignment in words with varying syllabic structure, as realized in different dialects21 Standard Arabic22

Palestinian Arabic

Lebanese Arabic

Cairene Arabic

Negev Bedouin

Gloss

kaˈtabt

kiˈtabt

I wrote

kiˈta:b

book

ˈmaktab

office

ˈkaːtib

ˈkːtib

ˈkaˈtib

writer

ˈᴣamal

ˈgamal

ᴣiˈmal

camel

ˈmaktaba

ˈmaktabi

makˈtaba

ˈmaktabah

library

heavy syllable (usually CVC or CVV).23 A heavy syllable in word-final position does not attract stress, however. This pattern makes sense if we adopt the notion of consonant extrametricality, whereby a word-final consonant is excluded from calculations of syllable weight. The disjunction that a final super-heavy syllable attracts stress whereas a final heavy syllable does not has been observed in all Arabic dialects described to date, and most authors agree that some form of extrametricality holds in all dialects. A rare exception to the “stress a final superheavy syllable” generalization is found in Sanaani Arabic in which a nonfinal heavy syllable (CVV or CVG) may attract stress away from a word-final super-heavy syllable (Watson 2002). For the most part, then, cross-dialectal variation in stress assignment is seen in words without either a final superheavy or a penultimate heavy. Table 3.1 gives data from Standard Arabic and four spoken dialects illustrating in (a)–(b) the shared properties observed in words containing a final super-heavy or penult heavy and in (c)–(d) the variation observed in two examples of words of other prosodic shapes.

Phonology The examples in (a) in Table 3.1 contain a final super-heavy CVCC or CVVC, which attracts stress in all of the dialects shown. In (b), stress falls on a heavy syllable (CVC or CVV) in the penultimate position, even if it is followed by a heavy syllable, in all dialects. Dialectal differences are seen in (c)–(d). The examples in (c) show stress assignment in a disyllable with a light penultimate, with Negev Bedouin Arabic the odd one out (stress on the final syllable rather than the first). In (d) we see stress assignment in a word with a heavy antepenultimate syllable followed by two light syllables (“HLL”), with Cairene Arabic the odd one out (stress on the light penultimate rather than the heavy antepenultimate). Table 3.2 Summary of Hayes’s (1995) metrical stress theory analysis of Arabic dialects Dialect

Foot Type

Foot Construction

Extrametricality

Degenerate Feet

Classical

unbounded

left-headed

consonant

N/A

Bani Hassan24

moraic trochee

left-to-right

foot

permitted

Palestinian

moraic trochee

left-to-right

foot25

absolute ban

Cairene

moraic trochee

left-to-right

consonant

absolute ban

Lebanese

moraic trochee

right-to-left

syllable

absolute ban

Bedouin Hijazi

moraic trochee26

right-to-left27

syllable

absolute ban

Negev Bedouin

iamb

left-to-right

foot

permitted

Cyrenaican Bedouin

iamb

left-to-right

foot

absolute ban

The surface variation in cases like (c)–(d) in Table 3.1 can be ascribed to underlying structural variation, and different theoretical approaches propose different potential parameters of variation. A widely adopted approach is the metrical stress theory (Hayes 1995), which reduces surface variation in Arabic to underlying variation in foot type, in the direction of foot construction within the word, in the size of the prosodic constituent targeted by extrametricality (segment vs. syllable vs. foot), and in treatment of syllables that cannot be grouped into a foot of the preferred type (known as “degenerate feet” and linked to restrictions on the size of the minimal word; see Section 3.2.3). A summary of Hayes’ analysis of a range of Arabic dialects is provided in Table 3.2. A metrical foot is a grouping of one or more syllables in which one syllable is designated as the head. Classical Arabic has an unbounded foot (containing any number of unstressed syllables along with the head). The majority of dialects in Hayes’ (1995) survey show a trochaic (left-headed) foot,28 with just a few dialects of Bedouin origin displaying an iambic (right-headed) foot. The difference in foot type is seen clearly in disyllables (Table 3.1c): moraic dialects have initial stress [ˈᴣamal]; iambic dialects have final stress [ᴣiˈmal]. In fact, the iambic pattern is probably equally widely distributed, at least in geographic terms, as it is found in dialects in Chad, Cameroon, and also Nigeria. All dialects of Arabic show extrametricality, but some of the surface variation can be ascribed to differences in the size of prosodic constituent which the stress algorithm treats as extrametrical. Hayes (1995) analyzes both Cairene and Palestinian Arabic as building trochaic feet left to right through the word, but they differ in how stress is assigned in a word that contains a heavy syllable followed by two light syllables, as in (d) in Table 3.1. Hayes attributes this difference to the operation of foot extrametricality in Palestinian Arabic [ˈmaktaba] (the final foot comprising two light syllables is ignored for the purposes of stress) versus consonant extrametricality only in ˈ

Phonology Cairene Arabic [makˈtaba] (with no effect in this case). In a similar fashion, differences observed in words of different syllable structures are analyzed as evidence of variation in the direction of foot construction and in the treatment of degenerate feet (see ibid. for details). Van der Hulst and Hellmuth (2010) point out that the minimal stress pairs that crucially distinguish one dialect from another are relatively infrequent, occurring only in words of certain prosodic shapes. This in turn means that, for the bulk of words, more than one set of parameters could account for the data, allowing for variation in which parameter settings language learners infer from them. These apparently minor surface differences between dialects have on occasion been instrumental in the development of metrical theory. For example, Watson (2011) describes how the particular patterns observed in morphologically complex words in Bedouin Hijazi Arabic (AlMozainy et al. 1985) led to the proposal of the bracketed metrical grid (Halle and Vergnaud 1987; Hayes 1995), a representation that encodes prosodic constituency at different levels. In a similar way, Arabic dialects show variation in the sensitivity of stress assignment to morphological structure (Brame 1973, 1974), and these facts were instrumental in developing theories such as lexical phonology (Kiparsky 1982) and stratal OT (Kiparsky 2000). Equally, the interaction of stress assignment with segmental processes is well-known in Arabic for giving rise to cases of opacity, in which the triggering context for a phonological process is not apparent in the surface form of the word. Such cases present a particular challenge to nonderivational theories of phonology, such as classic OT, and a sizeable body of literature has sought ways to analyze such cases of opacity (McCarthy 2003; Elfner 2009). Finally, the literature includes one or two interesting cases of dialects in which citation form word stress assignment patterns are subject to variation in connected speech. In Sanaani Arabic, for example, stress may be attracted to the initial syllable of a word when it occurs in a postpausal (phrase-initial) position (Watson 2002). Similarly, in the Casablanca dialect of Moroccan Arabic, the word-stress algorithm observed in words in citation form appears to disappear in connected speech and is replaced by word-final stress on all words (Boudlal 2001). The interaction of word stress with phrasal stress and other suprasegmental phenomena is probably the least well-documented aspect of the metrical phonology of Arabic dialects and as such is likely to yield important results in future.

3.2.5 Intonation Work on intonation in spoken Arabic dialects is an emerging field of research, and the body of literature discussed here is much smaller than that discussed in earlier sections of this chapter. We define intonation here as comprising the following phenomena (Halliday 1967): the chunking of utterances into prosodic phrases (tonality); the distribution of prosodic prominences (tonicity); and the shape of the pitch contour observed on and around those prominences (tone). A range of competing theoretical positions exists to account for each of these (Gussenhoven 2004; Ladd 2008). For example, for some authors prosodic phrasing (tonality) is derived directly from syntactic structure, whereas for others it reflects an intervening level of representation, the prosodic hierarchy (Inkelas and Zec 1995). The autosegmental-metrical (AM) theory of intonation offers a formal phonological representation of intonation, in which the pitch contour is modeled as a series of high (H) or low (L) pitch targets, associated autosegmentally with either the heads or edges of prosodic (“metrical”) constituents (Gussenhoven 2007). As for all less widely researched languages, there is a descriptive gap in work on intonation in Arabic, since standard grammars generally lack detailed discussion of intonational properties. As a result, there are relatively few descriptions of the intonational phonology of individual Arabic dialects, and even fewer studies make comparisons across dialects. Chahal (2009) provides a secondary analysis of a number of descriptions of individual dialects. She concludes that all of the dialects studied to date display postlexical use of pitch only (no dialects of Arabic have lexical tone) and that in all cases the observed intonational patterns require analysis in terms of both prominence-lending and demarcative pitch events (pitch accents on stressed syllables and boundary tones at phrase edges, in AM terms). Chahal’s survey finds that dialects do vary in the inventory of possible nuclear tones observed (the nuclear tone being the last and most prominent pitch accent in an intonational phrase, together with any following tonal configuration, such as a final rise or final fall). As in other areas of phonology, however, we might expect to find greater variation across dialects once more finely grained parameters of variation are identified. Cross-linguistic prosodic typology is as yet in its infancy but already suggests that the scope of cross-linguistic intonational variation is not limited to variation in the inventory of

Phonology possible nuclear tone configurations (Jun 2005). Ghazali et al. (2007) offer a very preliminary overview of intonational variation in Arabic based on qualitative analysis of a small sample of parallel data in six dialects.29 More data and analysis informed by known crosslinguistic parameters of variation are likely to yield further insights. For example, intonational languages are known to vary in the distributional density of pitch accents (Vigario and Frota 2003; Jun 2005; Ladd 2008). This crosslinguistic variation is also observed cross-dialectally in Arabic: Lebanese Arabic shows at least one intonational pitch accent in every “intermediate” prosodic phrase, whereas Egyptian Arabic displays an intonational pitch accent on every prosodic word; these two dialects also vary in whether they permit deaccentuation (Hellmuth 2007; Chahal and Hellmuth forthcoming). This parameter would match with the generalization observed by Ghazali et al. that “flat hat” patterns, with relatively sparse modulations in pitch across the utterance, were observed only in Eastern (Levantine) dialects and not in Western dialects (which included Egyptian Arabic). In another potential parameter of prosodic variation, languages are known to vary in whether their intonational system uses both prominence-lending and demarcative pitch events (pitch accents and boundary tones) or demarcative (boundary tones) only (Jun 2005). Although Chahal’s (2009) survey suggests that all Arabic dialects studied up to that point displayed both pitch accents and boundary tones, further descriptive work may reveal that the generalization is not correct. For example, the stress migration facts of Moroccan Arabic (Mitchell 1993; Boudlal 2001) are open to reanalysis as an intonational system that uses boundary tones only. Again, this is consistent with Ghazali et al.’s (2007) observation that North African dialects displayed a single rise + fall across each utterance (assuming partition of their utterances into two prosodic phrases each). In intonational phonology, as in other areas of phonological investigation we have seen in this chapter, a range of competing theoretical frameworks is available so that similar surface facts are open to reanalysis. Rifaat (2005) argues, from analysis of Modern Standard Arabic used in Egyptian broadcast media, that the intonational phonology of Standard Arabic is typologically unusual in its simplicity compared with other languages. El Zarka (2011) develops Rifaat’s analysis for colloquial Egyptian Arabic and suggests that different intonational configurations map directly to pragmatic functions such as focus and topic. Other work on the prosodic realization of pragmatic functions in Arabic includes work in functional grammar (Brustad 2000) and laboratory phonology (Hellmuth 2009, 2011). There is little work on the syntax–phonology interface in Arabic, with work on Egyptian Arabic being an exception (Hellmuth 2004, 2010, 2012). As in all areas of Arabic phonology, then, good theoretical modeling depends on continuing availability of good descriptions of the empirical facts of a range of Arabic dialects. Unlike word stress, these suprasegmental issues did not escape the attention of the Arab grammarians, with discussion of both pre- and postpausal phenomena (Cantineau 1946) as well as the role of prosody in disambiguation (Al-Harbi 1991). These issues can be expected to yield further theoretical gains in the years to come, once descriptive data are available in a sufficiently wide range of varieties of Arabic.

3.3 Conclusion The main body of this chapter outlines some of the most influential, or in our view important, areas of Arabic phonological research, and in all of these areas work is ongoing. Other aspects of Arabic phonology have received only limited attention to date and promise to be equally fruitful. One clear example is work on vowels, perhaps due in part to the small size of the vowel inventory of Arabic. Th is contrasts to Sībawayh, who treated the “alifs of imaalah and tafkhiim” with primary focus on imaalah; further research on the phonology of vowel fronting/raising in Arabic, within and across dialects, might serve to contextualize the extensive body of work that exists on emphasis (cf. Bellem 2007). Similarly, work on establishing the true size of the phonological vowel inventory of spoken varieties of Arabic is needed (cf. Youssef 2010). The motivation for continued research in all areas of Arabic phonology is partly theoretical, with reanalysis of existing data triggered as new theories of phonology are proposed and developed. In other cases as we have seen, new data in Arabic have motivated theoretical innovation in the past and can be expected to do so again. The contribution made to Arabic phonological research by the availability of detailed descriptions of the phonetics and phonology of a range of Arabic varieties cannot be underestimated (Rosenhouse 2011), yet the facts of many aspects of the phonology of many varieties of Arabic are still unknown. Our understanding of Arabic phonology,

Phonology and of phonology itself, will continue to benefit from fieldwork that adds descriptions of further dialects and registers of Arabic to the data set for which phonological theory must account.

References Al-Harbi, L. 1991. Formal analysis of intonation: The case of the Kuwaiti dialect of Arabic. Edinburgh: Herriot-Watt University. Al-Mozainy, Hamza, Robert Bley-Vroman, and John McCarthy. 1985. Stress shift and metrical structure. Linguistic Inquiry 16: 135–144. Al-Nassir, A. A. 1993. Sibawayh the phonologist: A critical study of the phonetic and phonological theory of Sibawayh as presented in his treatise Al-Kitab. London: Kegan Paul International. Beeston, A. F. L. 1970. The Arabic language today. London: Hutchinson University Library. Bellem, Alex. 2007. Towards a comparative typology of emphatics: across Semitic and into Arabic dialect phonology. PhD diss. School of Oriental and African Studies, University of London. Benkirane, T., 1998. Intonation in Western Arabic (Morocco). In Intonation systems: a survey of twenty languages, ed. D. Hirst and A. Di Cristo, 345–359. Cambridge, UK: Cambridge University Press. Blanc, Haim. 1964. Communal dialects in Baghdad. Cambridge, MA: Harvard University Press. Boersma, Paul. 1998. Functional phonology. The Hague: Holland Academic Graphics. Bohas, Georges. 1997. Matrices, etymons, racines. Leuven: Peeters. Boudlal, A. 2001. Constraint interaction in the phonology and morphology of Casablanca Moroccan Arabic. Rabat, Morocco: Université Mohammed V. Brame, M., 1973. Stress assignment in two Arabic dialects. In A festschrift for Morris Halle, ed. S. Anderson and Paul Kiparsky, 14–25. New York: Holt, Rinehart & Winston. —— 1974. The cycle in phonology: stress in Palestinian, Maltese and Spanish. Linguistic Inquiry 5: 39–60. Broselow, Ellen. 1976. The phonology of Egyptian Arabic. Unpublished PhD diss., University of Massachusetts, Amherst. ——. 1983. Nonobvious transfer: On predicting epenthesis errors. In Language transfer in language learning, ed. S. Gass and L. Selinker, 269–280. Rowley, MA: Newbury House. ——. 1992. Parametric variation in Arabic dialect phonology. InPerspectives on Arabic Linguistics IV: Papers from the 4th annual symposium on Arabic Linguistics, ed. Ellen Broselow, Mushira Eid, and John McCarthy, 7–45. Amsterdam: John Benjamins. Browman, C. and L. M. Goldstein. 1986. Towards an articulatory phonology. Phonology Yearbook 3: 219–252. Brustad, Kristen. 2000. The syntax of spoken Arabic: A comparative study of Moroccan, Egyptian, Syrian and Kuwaiti dialects. Washington, DC: Georgetown University Press. Cantineau, J. 1939. Remarques sur les parlers de sédentaires syro-libano-palestiniens. Bulletin de la Sociétéde Linguistique de Paris 40: 80–88. ——. 1946. Esquisse d’une phonologie de l’Arabe Classique. Bulletin de la Société de Linguistique de Paris 126: 93–140. Chahal, Dana. 2009. Intonation. In EALL 2, ed. Kees Versteegh. Associate Editors: Mushira Eid, Alaa Elgibali, Manfred Woidich, Andrzej Zaborski, 395–400. Leiden: Brill.

Phonology Chahal, Dana and Sam Hellmuth. Forthcoming. Comparing the intonational phonology of Lebanese and Egyptian Arabic. In Prosodic typology, vol. 2. Oxford: Oxford University Press. Clements, G. N. 1985. The geometry of phonological features. Phonology Yearbook 2: 223–250. Cowan, J. (ed.). 1979. Hans Wehr: A dictionary of Modern Written Arabic. Wiesbaden: Otto Harrassowitz. Davis, Stuart. 1995. Emphasis spread in Arabic and grounded phonology. Linguistic Inquiry 26: 465–498. Davis, Stuart and Bushra Zawayadeh. 2001. Arabic hypocoristics and the status of the consonantal root. Linguistic Inquiry 32: 512–520. Ehret, Christopher. 1989. The origins of third consonants in Semitic roots: An internal reconstruction (applied to Arabic). Journal of Afroasiatic Languages 3: 109–202. El Zarka, Dina, 2011. Leading, linking and closing tones and tunes in Egyptian Arabic—What a simple intonation system tells us about the nature of intonation. In PAL: Papers from the annual symposia on Arabic Linguistics volume XXII-XXIII: College Park, Maryland, 2008 and Milwaukee, Wisconsin, 2009, ed. Ellen Broselow and Hamid Ouali, 57–74. Amsterdam: John Benjamins. Elfner, Emily. 2009. Syllabification and stress-epenthesis interactions in harmonic serialism. MS thesis, University of Massachusetts, Amherst. Faust, Noam and Ya’ar Hever. 2010. Empirical and theoretical arguments in favor of the discontinuous root in Semitic languages. Brill’s Annual of Afroasiatic Languages and Linguistics 2: 1–38. Firth, J. R. 1948. Sounds and prosodies. Transactions of the Philological Society 1948: 127–152. Frisch, Stefan and Bushra Zawayadeh. 2001. The psychological reality of OCP-Place in Arabic. Language 77: 91– 106. Frisch, Stefan, Janet Pierrehumbert, and Michael Broe. 2004. Similarity avoidance and the OCP. Natural Language & Linguistic Theory 22: 179–228. Gafos, Adamantios. 1998. Eliminating long-distance consonantal spreading. Natural Language & Linguistic Theory 16: 223–278. ——. 2001. The initial state and verbal stems in Arabic. MS thesis, Utrecht Institute of Linguistics, OTS (Onderzoeksinstituut voor Taal en Spraak). ——. 2002. A grammar of gestural coordination. Natural Language & Linguistic Theory 20 269–337. ——. 2003. Greenberg’s asymmetry in Arabic: a consequence of stems in paradigms. Language 79: 317–357. ——. 2006. Dynamics in grammar: A comment on Ladd and Ernestus and Baayen. In Papers in Laboratory Phonology 8: Variaties of phonological competence, ed. L. M. Goldstein, D. H. Whalen, and C. T. Best, 51–79. Berlin: Mouton de Gruyter. Gafos, Adamantios, Philip Hoole, and Chakir Zeroual. 2011. Preliminary study of Moroccan Arabic word-initial consonant clusters and syllabification using electromagnetic articulography. In Instrumental studies in Arabic phonetics. Amsterdam, ed. Zeki Hassan and Barry Heselwood (eds.). Instrumental studies in Arabic phonetics, 29–45. Amsterdam: John Benjamins. Gafos, Adamantios, Philip Hoole, Kevin Roon, and Chakir Zeroual, 2010. Variation in timing and phonological grammar in Moroccan Arabic clusters. In Papers in Laboratory Phonology 10: Variation, detail and representation, ed. Cécile Fougeron, Barbara Kühnert, Mariapaola D’Imperio, and Nathalie Vallée, 657–698. Berlin: Mouton de Gruyter. Ghazali, S., R. Hamdi, and M. Barkat. 2002. Speech rhythm variation in Arabic dialects. Aix-en-Provence: Laboratoire Parole et Langage. In Proceedings of the Speech Prosody 2002 Conference, 11–13 April 2002, ed. B.

Phonology Bel and I. Marlin, 127–132. Ghazali, S., R. Hamidi, and K. Knis. 2007. Intonation and rhythmic patterns across the Arabic dialects continuum. In PAL: Papers from the annual symposium on Arabic Linguistics volume XIX: Urbana, Illinois April 2005, ed. Elabbas Benmamoun, 97–122. Amsterdam: John Benjamins. Goldsmith, J. 1976. Autosegmental phonology. Bloomington: Indiana University Linguistics Club. Greenberg, J. 1950. The patterning of root morphemes in Semitic. Word 6: 162–181. Guerssel, Mohamed and Jean Lowenstamm. 1990. The derivational morphology of the Classical Arabic verbal system. Ms.: Université du Québec à Montréal/Universite de Paris 7. Gussenhoven, C. 2004. The phonology of tone and intonation. Cambridge, UK: Cambridge University Press. ——. 2007. The phonology of intonation. In The Cambridge handbook of phonology, ed. Paul de Lacy, 253–280. Cambridge, UK: Cambridge University Press. Halle, Morris and Jean-Roger Vergnaud. 1987. An essay on stress. Cambridge, MA: MIT Press. Halliday, M. A. K. 1967. Intonation and grammar in British English. The Hague: Mouton. Hamdi, Rym, Salem Ghazali, and Melissa Barkat-Defradas. 2005. Syllable structure in spoken Arabic: A comparative investigation. In Proceedings of Interspeech 2005, 2245–2248. Harris, John and Geoff Lindsey, 1995. The elements of phonological representation. In Frontiers of phonology: Atoms, structures, derivations, ed. Jacques Durand and Francis Katamba, 34–79. London: Longman. Harun, ‘Abd S. M. 1983. Kitab Sibawayh: Tahqiq wa-sharh. Beirut: ‘Aalam al-Kutub. Hassan, Zeki and Barry Heselwood (eds.). 2011. Instrumental studies in Arabic phonetics. Amsterdam: John Benjamins. Hayes, Bruce. 1995. Metrical stress theory: Principles and case studies. Chicago: University of Chicago Press. Hayward, Katrina and Richard Hayward. 1989. “Guttural”: Arguments for a new distinctive feature. Transactions of the Philological Society 87: 179–193. Hellmuth, Sam. 2004. Prosodic weight and phonological phrasing in Cairene Arabic. In Proceedings of the 40th meeting of the Chicago Linguistics Society: The main session. 97–111. Chicago: Chicago Linguistic Society. ——. 2007. The relationship between prosodic structure and pitch accent distribution: Evidence from Egyptian Arabic. Linguistic Review 24: 289–314. ——. 2009. The (absence of) prosodic reflexes of given/new information status in Egyptian Arabic. In Information structure in spoken Arabic, ed. Jonathan Owens and Alaa Elgibali, 165–188. London: Routledge. ——. 2010. Functional complementarity is only skin deep: Evidence from Egyptian Arabic for the autonomy of syntax and phonology in the expression of focus. In The sound patterns of syntax, ed. Nomi Erteschik-Shir and Lisa Rochman. Oxford: Oxford University Press. ——. 2011. Acoustic cues to focus and givenness in Egyptian Arabic. Instrumental studies in Arabic phonetics. Amsterdam, ed. Hassan, Zeki and Barry Heselwood (eds.). Instrumental studies in Arabic phonetics, 299–324. Amsterdam: John Benjamins. ——. 2012. Variable cues to phrasing: Finding edges in Egyptian Arabic. In Prosody matters: essays in honor of Lisa Selkirk, ed. Toni Borowsky, Shigeto Kawahara, and Mariko Sugahara, 237–279. London: Equinox. Herzallah, Rukayyah. 1990. Aspects of Palestination Arabic phonology: A non-linear approach. Ithaca, NY: Cornell University.

Phonology Idrissi, Ali, 1997. Plural formation in Arabic. In PAL, Papers from the annual symposium on Arabic linguistics, volume X, ed. Mushira Eid and Robert Ratcliffe, 123–146. Amsterdam: John Benjamins. Idrissi, Ali, Jean-François Prunet, and Renée Béland. 2008. On the mental representation of Arabic roots. Linguistic Inquiry 39: 221–259. Inkelas, S. and D. Zec, 1995. Syntax-phonology interface. In The handbook of phonological theory, ed. J. Goldsmith, 535–549. Oxford: Blackwell. Irshied, Omar and M. Kenstowicz. 1984. Some phonological rules of Bani-Hassan Arabic: A Bedouin dialect. Studies in the Linguistic Sciences 14: 109–148. Itô, Junko. 1989. A prosodic theory of epenthesis. Natural Language & Linguistic Theory 7: 217–260. Jacobs, Haike. 1990. On markedness and bounded stress systems. Linguistic Review 7: 81–119. Jun, Sun-Ah, 2005. Prosodic typology. In Prosodic typology: The phonology of intonation and phrasing, ed. Sun-Ah Jun, 430–458. Oxford: Oxford University Press. Kager, René. 2009. Stress. In EALL, vol. 4, ed. Kees Versteegh. Associate Editors: Mushira Eid, Alaa Elgibali, Manfred Woidich, Andrzej Zaborski, 344–353. Leiden: Brill. Kaye, Jonathan. 1990. “Coda” licensing. Phonology Yearbook 7: 301–330. Kiparsky, Paul, 1982. From cyclic phonology to lexical phonology. In The structure of phonological representations, ed. Harry van der Hulst and N. Smith, 131–175. Dordrecht: Foris. ——. 2000. Opacity and cyclicity. Linguistic Review 17: 351–367. ——. 2003. Syllables and mora in Arabic. In The syllable in optimality theory, ed. Caroline Féry and Reuben van de Vijver, 147–182. Cambridge, UK: Cambridge University Press. Ladd, D. R. 2008. Intonational phonology. Cambridge, UK: Cambridge University Press. Lass, R. 1984. Phonology. Cambridge, UK: Cambridge University Press. Lombardi, Linda. 1999. Positional faithfulness and voicing assimilation in optimality theory. Natural Language & Linguistic Theory 17: 267–302. Lowenstamm, Jean, 1996. CV as the only syllable type. In Current trends in phonology, models and methods, ed. Jacques Durand and Bernard Laks, 419–443. Salford: European Studies Research Institute. ——. 2003. À propos des gabarits. Recherches linguistiques de Vincennes 32: 7–30. Maddieson, Ian. 2011. Consonant inventories. In The world atlas of language structures online, ed. Matthew Dryer and Martin Haspelmath. Munich: Max Planck Digital Library. McCarthy, John. 1979. Formal problems in Semitic phonology and morphology. Cambridge, MA: MIT Press. ——. 1981. A prosodic theory of nonconcatenative morphology. Linguistic Inquiry 12: 373–418. ——. 1986. OCP effects: Gemination and antigemination. Linguistic Inquiry 17: 207–263. ——. 1988. Feature geometry and dependency: A review. Phonetica: International Journal of Speech Science 43: 84–108. ——. 1994. The phonetics and phonology of Semitic pharyngeals. In Phonological structure and phonetic form: Papers in laboratory phonology III, ed. P. Keating, 191–233. Cambridge, UK: Cambridge University Press. ——. 1998. Morpheme structure constraints and paradigm occultation. In Proceedings of the Annual Meeting of Chicago Linguistic Society 32, ed. Catherine Gruber, Derrick Higgins, Kenneth Olson, and Tamra Wysocki, 123– 150. Chicago: Chicago Linguistics Society.

Phonology ——. 2003. Sympathy, cumulativity, and the Duke-of-York gambit. In The Syllable in optimality theory, ed. Caroline Féry and Ruben van de Vijver, 23–76. Cambridge, UK: Cambridge University Press. ——. 2004. Headed spans and autosegmental spreading. Ms. University of Massachusetts, Amherst. Available at http and works.bepress.com/john_j_mccarthy/60. ——. 2005. Optimal paradigms. In Paradigms in phonological theory, ed. Laura Downing, Tracy A. Hall, and Renate Raffelsiefen, 170–210. Oxford: Oxford University Press. Mester, Armin and Jaye Padgett, 1994. Directional syllabification in generalized alignment. In Phonology at Santa Cruz 3, ed. Jason Merchant, Jaye Padgett, and Rachel Walker, 79–85. Santa Cruz, CA: Linguistics Research Center. Mitchell, T. F. 1993. Pronouncing Arabic 2. Oxford: Oxford University Press. Nolan, Francis and Eva-Liina Asu. 2009. The Pairwise variability index and coexisting rhythms in language. Phonetica: International Journal of Speech Science 66: 64–77. Owens, Jonathan. 1988. The foundations of grammar: an introduction to medieval Arabic grammatical theory. Amsterdam: John Benjamins. ——. 2006. A linguistic history of Arabic. Oxford: Oxford University Press. Padgett, Jaye. 1995. Stricture in feature geometry. Stanford, CA: CSLI Publications. Pierrehumbert, Janet. 1993. Dissimilarity in the Arabic verbal roots. Proceedings of the North East Linguistics Society 23: 367–381. ——. 2006. The next toolkit. Journal of Phonetics 34: 516–530. Prince, A. and P. Smolensky. 2004. Optimality theory: constraint interaction in generative grammar. Malden, MA: Blackwell Publishing. Prunet, Jean-François, Renée Béland, and Ali Idrissi. 2000. The mental representation of Semitic words. Linguistic Inquiry 31: 609–648. Rifaat, Khalid, 2005. The structure of Arabic intonation: A preliminary investigation. In PAL XVII-XVIII: Papers from the seventeenth and eighteenth annual symposia on Arabic linguistics, ed. Mohammad Alhawary and Elabbas Benmamoun, 49–67. Amsterdam: John Benjamins. Roach, P. 1982. On the distinction between “stress-timed” and “syllable-timed” languages. In Linguistic controversies, ed. D. Crystal, 73–79. London: Edward Arnold. Rose, Sharon. 2000. Rethinking geminates: Long-distance geminates and the OCP. Linguistic Inquiry 31: 85–122. Rosenhouse, Judith. 2011. Trends of development in Arabic dialectology in the 20th century: A survey. Zeitschrift für arabische Linguistik 54: 42–66. Scheer, Tobias. 2004. A lateral theory of phonology: What is CVCV, and why should it be?. Berlin: Mouton de Gruyter. Selkirk, Elisabeth O. 1981. Epenthesis and degenerate syllables in Cairene Arabic. In Theoretical issues in the grammar of Semitic language, ed. Hagit Borer and J. Aoun, 209–232. Cambridge, MA: Department of Linguistics, MIT. Shahin, Kimary. 2003. Postvelar harmony. Amsterdam: John Benjamins. Shaw, Jason, Adamantios Gafos, Philip Hoole, and Chakir Zeroual. 2009. Temporal evidence for syllabic structure in Moroccan Arabic: Data and model. Phonology 26: 187–215. van der Hulst, Harry and Sam Hellmuth, 2010. Word accent systems in the Middle East. In A survey of word accentual patterns in the languages of the world, ed. R. Goedmans and H. van der Hulst, 615–646. Berlin: Mouton

Phonology de Gruyter. Vigario, Marina and Sonia Frota. 2003. The intonation of Standard and Northern European Portuguese: A comparative intonational phonology approach. Journal of Portuguese Linguistics 2: 115–137. Voigt, Rainer M. 1988. Die infirmen Verbaltypen des Arabischen und das Biradikalismus-Problem. Stuttgart: Steiner. Watson, Janet C. E. 2002. The phonology and morphology of Arabic. Oxford: Oxford University Press. Watson, Janet C. E. 2007. Syllabification patterns in Arabic dialects: Long segments and mora sharing. Phonology 24: 335–356. Watson, Janet C. E. 2011. Word stress in Arabic. In The Blackwell companion to phonology, vol. 5, ed. Marc Oostendorp, Colin Ewen, Elizabeth Hume, and Keren Rice, 2990–3018. Oxford: Blackwell. Wiget, Lukas, Laurence White, Barbara Schuppler, Izabelle Grenon, Olesya Rauch, and Sven L. Mattys. 2010. How stable are acoustic metrics of contrastive speech rhythm? Journal of the Acoustical Society of America 127: 1559– 1569. Yoshida, Shohei. 1993. Licensing of empty nuclei: The case of Palestinian vowel harmony. Linguistic Review 10: 127–159. Youssef, Islam. 2010. Against underlying mid vowels in Cairene Arabic. Zeitschrift für arabische Linguistik 52: 5– 38.

Notes: (1) Greenberg (1950: n. 2) notes that lists of “incompatible consonants” are provided by Jalāl ad-Dīn Suyūṭī and Ibn Jinni. (2 ) Identical consonants in C1–C3 position are relatively rare but nonetheless observed, for example, 〈qlq〉. (3 ) The table shows consonants only in the inventory of Arabic; Greenberg also includes consonants found in other Semitic languages, such as [p]. (4 ) In (5), and throughout this paper, the emphatic coronals are represented using symbols not from the International Phonetic Alphabet (IPA): [ ] appear as [ṭ ḍ ṣ ḍ], respectively. Although current practice is to represent “emphasis” using the IPA uvularization diacritic [ ], the phonetic realization of emphasis varies more widely across dialects than this representation implies and is defined by a complex of articulatory gestures, only one of which is uvularization (see Section 3.2.2; also [Embarki, “Phonetics”]). (5) Directionality of autosegmental spreading is usually argued to vary by rule and by language (see discussion in McCarthy 2004). (6 ) See Ratcliffe [“Morphology”] for a discussion of root-versus stem-and word-based approaches to Arabic morphology. (7 ) Similarity is calculated as the number of shared natural classes divided by the sum of the number of shared and unshared natural classes (Frisch et al. 2004: 198). (8 ) Few varieties display the full set of emphatics; see Embarki [“Phonetics”] for discussion. (9 ) Also found in some other Semitic languages, Caucasian, and languages of the Pacific North West (McCarthy 1994; Shahin 2003); the phonetic realization of the effect varies greatly across languages and dialects. (10 ) See Embarki [“Phonetics”] for a detailed overview of research on the articulatory and acoustic properties of Arabic emphatics. (11) Sibawayh also notes that alif is realized as [a] in the Hijazi dialect (Al-Nassir 1993: 103).

Phonology

(12 ) This is a monovalent feature in Watson’s analysis; monovalent features are either present or absent from the phonological representation of speech sound, with no binary ± settings. (13 ) Both authors ascribe their use of [guttural]/[pharyngeal] to the proposal made in Hayward and Hayward (1989) for a feature [guttural], which is there argued to be more broadly defined in terms of the zone of constriction. Note that McCarthy treats /q/ separately from the uvular fricatives. (14 ) See Firth (1948) or Lass (1984: 163–166) for a brief overview of Firthian prosodic analysis. (15) Compare Classical Arabic [qult] “I said” with [qult] in Morocco, [gult] in Baghdad, and [ʔult] in Cairo. (16 ) Broselow (1992) adopts a fairly standard moraic view of syllable structure, in which a mora, a unit of syllable weight, is assigned to rhymal constituents (vowels and coda consonants) but not to onset consonants. Arabic dialects display a general preference for bimoraic syllables, that is, with a heavy rhyme, either VC or VV. (17 ) “C dialects” are those that, like Moroccan, allow surface CCC clusters. (18 ) The Cv dialects are mostly found in Yemen but also include Meccan. (19 ) Lowenstamm’s CV-only analysis for Arabic has its roots in government phonology, in which word-final codas are analyzed as the onset of a vowelless CV syllable (Kaye 1990), and has inspired strict-CV phonology, which extends the CV-only analysis to all languages (Scheer 2004). (20 ) In differential dialects, unstressed vowel deletion (syncope) targets only high vowels [i u], whereas in nondifferential dialects, syncope affects all short vowels, including [a] (Cantineau 1939). (21) The algorithms for each dialect, as described in Hayes (1995), are applied to derive a parallel set of examples. (22 ) These are “pausal” forms of the words, that is, as produced utterance-finally, without case-marking vowels. (23 ) In Sanaani Arabic, only CVG (closed by geminate) and CVV count as heavy (Watson 2002). (24 ) Bani Hassan Arabic is a Bedouin dialect spoken in northern Jordan (Irshied and Kenstowicz 1984). (25) Jacobs (1990) analyzed Palestinian Arabic with syllable extrametricality rather than foot extrametricality. (26 ) McCarthy (2003) reanalyzed Bedouin Hijazi as having iambic feet, in parallel with other Bedouin varieties. (27 ) Hayes (1995: 181) suggests that Bedouin Hijazi could also be analyzed with left -to-right foot construction. (28 ) Hayes distinguished two types of trochaic foot: the moraic trochee (comprising two mora); and the syllabic trochee (comprising two syllables). All the trochaic Arabic dialects use moraic trochees. (29 ) They also present the results of a larger, and more methodologically robust, quantitative study of rhythmic variation in Arabic dialects, which is discussed in Section 3.2.3. Sam Hellmuth Sam Hellmuth, University of York

Morphology Robert R. Ratcliffe The Oxford Handbook of Arabic Linguistics Edited by Jonathan Owens Print Publication Date: Sep 2013 Online Publication Date: Dec 2013

Subject: Linguistics, Morphology and Syntax, Languages by Region DOI: 10.1093/oxfordhb/9780199764136.013.0004

Abstract and Keywords This article discusses the study of Arabic morphology. It first considers the root-and-pattern theory, which has become the orthodox approach to Arabic synchronic morphology. It then details the paradigm shift in the mid1980s, when students of Arabic morphology reached the conclusion that a rigidly reductionist root-and-pattern analysis is fundamentally inadequate as a descriptive tool. This has led to a variety of alternative models, which can be loosely grouped under the rubric of word based or stem based. All such models have in common the idea that many or all morphological regularities in Arabic can be best described in terms of derivational processes operating on words or stems rather than in terms of combinations of roots and patterns. Keywords: Arabic, linguistics, language, morphology, root-and-pattern theory, word based, stem based

4.1 General Issues MORPHOLOGY in general refers to the study of form. Hence, linguistic morphology can be loosely defined as the study

of the form of words. But behind this deliberately vague formulation there is a world of controversy. Depending upon one’s focus and theoretical point of view, morphology can be defined as the study of the internal structure of words, or as the study of the processes for forming words, or as the study of the formal similarities and interconnections among words. The study of Arabic morphology cannot be separated from these larger theoretical and definitional issues. Broadly speaking, the history of morphology over the last century can be described as shift from one of these three definitions to another, briefly as a trajectory from structure to process to relationship. Early 20th-century structuralism focused on analyzing words into minimal units of form and sense (signs), for which the term morpheme was coined. The structuralist approach can be characterized as analytical and reductionist. The rise of generative grammar led to an emphasis on process. Unlike earlier structuralists, the generativists were explicitly interested in describing the mental behavior of speakers. It was thought that evidence for mental processes was most likely to be found in aspects of language that were productive (capable of producing novel forms) or regular (describable in terms of a rule). This led to a morphological theory centered on the notion of word formation rules (WFRs). McCarthy (2008: 297) says, “Morphology is the study of word formation.” In recent times, an alternative to the rule-based approach has emerged from neurocognitive linguistics, in the form of models based on a network of connections among words in a lexicon (Lamb 1998). This approach appears likely to converge with word-based and paradigmatic models that have developed organically from older generative and structuralist approaches. As Gafos (2009: 338) says, “… Linguistic morphology is primarily concerned with systems of relations between words.” If one were to characterize the earlier history of Arabic morphological theory in terms of these 20th- and 21st-

Morphology century currents, it might be said that the medieval Arabic grammatical tradition was focused on process. The central concept of Arabic “ṣarf” (lit. “change,” generally translated “morphology”) is taṣriyf (lit. “causing something to change,” generally translated “derivation” [Baalbaki, “ALT I”]). By contrast, the 19th-century comparative Semitic tradition, with antecedents in medieval biblical philology, focused on analysis of words into smaller units.

4.2 The Structuralist Approach and Its Antecedents One of the great idées reçues in linguistics, often repeated in general textbooks and introductory grammars, is that words in Semitic languages are uniquely formed by combining a consonantal root, which indicates core meaning, and a syllabic-vocalic pattern, which indicates grammatical function [Hellmuth, “Phonology”]. No doubt this does have a basis in the way that most Arabic dictionaries have been organized since medieval times [Buckwalter and Parkinson, “Modern Dictionaries”]. Students of Arabic soon learn how to identify a root in a word to look it up in the dictionary. However, as Larcher (1999, 2006) observes, the Arabic grammatical tradition, as opposed to the lexicographical tradition, does not make use of derivations of words from roots. The term ʔaṣl “root, source” as used in this tradition refers to a word form perceived to be the base for another, usually the maṣdar (verbal noun, lit. “source”). “Ġalāyīnī tells us that the imperative uktub ‘write!’ is derived from the imperfect yaktub, ‘he writes, he will write’ the imperfect yaktub from the perfect katab, and the perfect katab from the maṣdar kitaaba” (Larcher 2006: 575). Owens (1988: 89–124) makes it clear that the grammarians allowed analysis of words into smaller units only in the case where such units occurred in sequence. Thus, the 13th-century grammarian Astarabadhi analyzed a word like muslimuuna “Muslims” as consisting of two kalima (“words,” but in this context perhaps “morphemes”): muslim- (“muslim”) and -uuna (masc. pl. nom.). The same author, however, explicitly rejects the idea of analyzing a word like kulayb “small dog” (diminutive) into two kalima correlating with the root “dog” and the pattern “diminutive.” The tradition of rigorously analyzing all words into roots and patterns may well be rooted in medieval Hebrew comparative philology, which was in turn influenced by the Arabic lexicographical tradition (Maman 2004). For Biblical Hebrew and Aramaic, where conditioned sound changes have considerably complicated the surface synchronic phonology, the assumption that a three-consonant root necessarily underlies every surface word allowed for the discovery of regular phonological changes that might otherwise have remained unobserved. Because the history of comparative Semitics has been strongly shaped by researchers primarily interested in the Biblical languages, root-and-pattern (R&P) theory became the dominant model in comparative Semitics. From there it was adapted into reference and teaching grammars of the individual Semitic languages, including Classical Arabic (CA). Exactly when the R&P theory became the orthodox approach to Arabic synchronic morphology is not clear. Larcher (1999) and other Francophone linguists (Bohas 1993) attribute the origin of the idea to Cantineau (1950a, 1950b). But the theory is already incorporated in textbooks in other European languages from the same period (e.g., Cowan 1958). The prosodic analysis of the London School linguists (e.g., Firth 1948; Palmer 1970) and the long component analysis developed within American structuralism (e.g., Harris 1944) assume an R&P model as the departure point for theories designed to extend the notion of morpheme to discontinuous sequences within a word. However, these theories were principally developed on the basis of Semitic languages other than Arabic. As an explicit synchronic theory influenced by the structuralist ideal of maximal analytical reductionism, the attribution to Cantineau seems as good a starting point as any. Certainly Cantineau offers an explicit and maximally strong version of R&P theory: Les racines et les schèmes constituent deux grands systèmes croisés, enveloppant dans leurréseau toute la masse du vocabulaire sémitique. Cantineau (1956a, cited in Bohas 1993: 45) … Tout mot est entièrement défini sans ambiguité par sa racine et son schème …. Cantineau (1956b, cited in Bohas 1993: 45)

Morphology This theory works quite elegantly up to a point, as we can see by organizing the following basic verb forms into a croisement, or grid, as Cantineau suggests: (1)

pattern〉

Imperfect

Perfect

Active PRT

Passive PRT

-CCuC

CaCaC

CaaCiC

maCCuuC

root∨ ktb

“write”

ya-ktub

kataba

kaatib

maktuub

drs

“study”

ya-drus

darasa

daaris

madruus

Figure 4.1 Root, template, vowel melody.

From this point it is possible to further extend the reductionist analysis. One way is by looking for meaningful groups in consonantal strings shorter than the root. This is a venerable tradition in Arabic and comparative Semitic linguistics (see Zaborski 1991), whose most recent variation is the etymon theory of Bohas (1997). Another possibility is to further subdivide the pattern. This is the approach of the highly influential work of McCarthy (1979, 1981, 1983). McCarthy’s aim is to adapt the analytical apparatus of autosegmental phonology, a theory developed for tonal phenomena, to the morphological analysis of Arabic. Although he essentially takes for granted the traditional R&P analysis, his particular innovation is to separate the traditional pattern into two parts, a vowel melody and a syllabic template (or CV skeleton), as shown in Figure 4.1. Each of these three elements was said to be a morpheme on a separate tier. Words were formed through combining these morphemes through a process of tier conflation. And the “nonconcatenative” morphology of Arabic could be reanalyzed as morpheme based and combinatorial—although McCarthy (1981: 375) treats all of this as redundancy rules in a word-based lexicon. One happy result of this analysis was that it brought out commonalities among patterns. Certain patterns like diminutive kulayb “little dog” and plural kilaab “dogs” (sg. kalb) share a CV skeleton (CvCvvC) but differ in vowel melodies. Derived verb stems II (yuCa CCiC) and III (yuCaaCiC) have different CV skeleta but share a vowel melody (u-a-i). Indeed the perfective of all the derived stems can be said to share a vowel melody, simply -a- with automatic spreading.

4.3 State of the Art: The Word-Based Turn Since the mid-1980s the study of Arabic morphology has undergone something of a paradigm shift. Many students of Arabic morphology have reached the conclusion that a rigidly reductionist root-and-pattern analysis à la Cantineau is fundamentally inadequate as a descriptive tool. This has led to a variety of alternative models, which can be loosely grouped under the rubric of word-based or stem-based. All such models have in common the idea that many or all morphological regularities in Arabic can be best described in terms of derivational processes

Morphology operating on words or stems rather than in terms of combinations of roots and patterns. No one would deny that forms like yaktubu and kataba are related. But there are various ways this relationship can be described without assuming derivation from a lexically listed root k-t-b. There are several motivations for this shift. The most important has to do with technical problems of formal description. Many subsystems of Classical Arabic morphology that appear opaque or excessively complex within an R&P analysis (plurals, diminutives, tense/aspect and voice apophony, the derived verb system) yield to a much simpler and more rational analysis within a framework that incorporates the idea of word-to-word or stem-to-stem derivation (McCarthy and Prince 1990a, 1990b, 1995; McCarthy 1993; McOmber 1995; Ratcliffe 1997, 2003a, 2005; Benmamoun 1999, 2003; Ussishkin 2003). A second motivation, emphasized recently in work by Gafos (2003, 2009), Chekayri and Sheer (1996), and Chekayri (2007) but with antecedents reaching back to Schramm (1962, 1991), is that word-based derivation provides a better framework for exploring the relationship of phonology to morphology, especially the phonology of geminates and glides. A third motivation has to do with patterns of synchronic variation in dialectal spoken Arabic. Many descriptive linguists (see, e.g., Heath 1987; Holes 2004; Watson 2006) have concluded that the R&P framework is too restrictive to accommodate what speakers actually appear to be doing when they manipulate and innovate new word patterns. A fourth motivation, related to the previous, is diachronic change and analogy (Heath 1987; Carter 1996; Ratcliffe 1998, 2001a, 2003b, 2006). Many patterns in dialectal Arabic are not found in CA and therefore appear to be innovations. How do speakers create new patterns? A fifth motivation relates to meaning and semantic interpretation (Larcher 1995, 2006; Watson 2006). Simply put, if one expects that the meaning of a word can be determined by the meaning of its root combined with the meaning of its pattern, one is likely to be disappointed. Against all of this, the principal critique has come from the side of psycholinguists, who argue that speakers seem to be aware of the relationships between words sharing a root (Boudelaa and Marslen Wilson 2005; Boudelaa 2006) and that speakers seem to be able to manipulate root consonants independently of vowels (Prunet, Beland, and Idrissi 2000).

4.3.1 Problems of Formal Description McCarthy (1981: 375) observes that the traditional R&P analysis provides “no general treatment of relations between vowel patterns except as instantiated on a particular root.” Within a framework that aimed “to capture significant generalizations,” that is, to describe any sort of recurrent pattern in the data in terms of abstract rulelike statements, this situation was clearly unsatisfactory. As noted already, the division of patterns into CV skeletons and vowel melodies brought out commonalities among patterns that the traditional analysis ignores. Thus, paradoxically McCarthy’s reductionism (dividing words into ever smaller pieces) led to a more holistic approach (recognition of regularities linking words or patterns across the lexicon). But it quickly becomes apparent, or it became apparent to McCarthy and his collaborator Alan Prince, that other general patterns of cross-word regularity cannot be stated within a framework that assumes derivation from a root. This led them to propose a significant role for stem-based derivation, notably in the case of the so-called broken, or stem-internal, plurals, where they propose that “the stem rather than the root is the base of pluralization” (McCarthy and Prince 1990a: 251). As an example of the sort of problem that motivated this shift consider the singular–plural pairs in (2). (2)

Morphology

ʕaskar

ʕasaakir

“soldier”

maktab

makaatib

“office”

ḍamiir

ḍamaaʔir

“pronoun”

qaalab

qawaalib

“mold”

The plural forms in the second column all have the same syllable structure and vowel pattern. It is most economical to describe them as being formed on the same plural pattern, CaCaaCiC. But R&P theory does not allow this because the string of consonants that fills the pattern does not constitute a root in any but the first case. Under standard R&P analysis we are forced to recognize four different patterns: CaCaaCiC plus root ʕ-s-k-r; maCaaCiC plus root k-t-b; CaCaaʔiC plus root ḍ-m-r; and CawaaCiC plus root q-l-b. Furthermore, this analysis obscures the systematic relationship between singular and plural. The plural maCaaCiC always reflects singular with initial m-, CaawaCiC always reflects a singular with long vowel in the first syllable, and CaCaaʔiC always reflects a singular with a long vowel in the second syllable. Simply put, every onset and coda (every letter in the Arabic script, Ratcliffe 2001b) in the singular is mapped onto a C position in the plural. To put the problem another way, if we define a pattern as a vocalic-syllabic shape consistently associated with a meaning or grammatical function, then the plurals in (2) are clearly formed by combing a pattern with a consonantal string, but this consonantal string is not a root. If we insist that all words contain a root, then the residue does not fit the definition of the pattern. Actually, the McCarthy and Prince (1990a) analysis goes beyond the patterns in (2). They suggest that plurals like those in (3) can also be accounted for by the same process as those in (2): (3) SG

qidħ

qidaaħ

“spear”

sulṭaan

salaaṭiin

“sultan”

In all these cases, the plural has an initial CvCaa sequence, and what follows this—C, CvC or CvvC—is determined by the syllable structure of the singular. McCarthy and Prince (1990a) propose to capture this generalizaton through the theory of prosodic circumscription: bracketing the first two-mora CvX sequence and mapping this to an iambic CvCaa. template, with the syllabic residue carried over from singular to plural. As they acknowledge (1990a: 217–218), this analysis represents a radical break with R&P theory: “Although the defining iambic sequence has a clearly templatic character, the familiar resources of root-and-template morphology are quite inadequate to the task of representing it. The fault lies not in the notion of template but in its presumed dependence on the consonantal root; for the iambic plural systematically reflects aspects of the singular that the consonantal root does not determine.” McCarthy and Prince (1990b) and McCarthy (1993) propose a different, but also stem-based, mechanism for the verbal morphology. They derive stem II CaCCaC and stem III CaaCaC from the basic stem CaCaC through affixation of an empty mora (designated μ) after the intial Cv sequence 〈Ca〉μCaC, which is then filled by spreading of either the preceding vowel yielding CaaCaC or of the following consonant yielding CaCCaC. As Ratcliffe (1997, 2003a) and Benmamoun (2003) point out, these analyses are not contradictory. The decision to treat the verb one way and the noun another is quite arbitrary. If we take the imperfect form of the verb as basic and further assume word-based derivation, verbal and nominal morphology are surprisingly parallel. For the most frequent type of verbs and nouns, those with a three-consonant stem (CvCCun, yaCCvC-v), the most productive

Morphology “internal” derivations in both nominal and verbal morphology involve the same core operation, which can be described either as mapping of a bimoraic CvX sequence to an iambic CvCvX template or as affixing of a vμ sequence to this same CvX base. Since prosodic circumscription explicitly allows reference to a phonologically, rather than morphologically, defined part of a word, the fact that the initial heavy syllable of verbs is heteromorphemic (person/number/gender prefix ya-, ta-, etc. + first consonant of stem CCVC) is not a problem. (The minimal freely occurring forms of CA nouns, indefinite (nominative), and verbs, jussive, are taken as the starting point.) (4) CvCCvC

〉〉

CvCvxCvC

〈CvX〉

〉〉

〈CvCvX〉

ya k t u b

yu k a t t i b

(I 〉〉 II) “cause to write”

ya k t u b

yu k aa t i b

(I 〉〉 III) “write to (s.o.)”

ka l b u n

k i l aa bun

(plural) “dogs”

ka l b u n

ku l ay bun

(diminutive) “little dog”

In fact, once we admit the principles of prosodic circumscription and moraic affixation into the analysis, we virtually eliminate the need for the idea of the fixed pattern. Ratcliffe (1997, 2003a) argues that all productive morphology appearing to be purely pattern based (not containing an additional affix) can be accounted for by the same rule of moraic affixation to a bimoraic base, interacting with other affixation processes. These include the active participle (CaaCiC), deverbal adjectives and nouns (CvCiiC, CvCuuC, CvCaaC), and the stem IX color verbs (iCCaCC, yaCCaCC) with stem final gemination: (5)

Morphology

active participle source:

aμ affixation

prefix deletion

coda filling

〈yvkaμ〉tvb

〉〉 kaμtib 〉〉

kaatib “writer, having written”

imperfective 〈yak〉tub 〉〉

deverbal nouns and adjectives source:

aμ affixation

coda filling

perfective 〈kat〉ab 〉〉

〈katvμ〉ab

〉〉

katiib, etc.

aμ affixation

prefix addition

coda filling

〈zurvμ〉q

〉〉 ya〈zVrv μ〉q 〉〉

yazraqq “become blue”

color verbs source: color noun 〈zur〉q 〉〉

This analysis obviously depends upon what is taken as the source of the derivation, but the choice is not arbitrary. There is a logical connection between imperfective aspect and active (participles), as between perfective and passive (verbal adjectives). In the imperfective aspect since the action is ongoing, the agent must be present, although the patient—the effect or result of the action—may not yet exist. In active constructions the action is predicated of the agent (i.e., the agent is the subject). The focus in both categories is therefore on the agent or subject. In the perfective since the action is complete, the agent is no longer present, but the result or effect of the action remains. Likewise, in passive constructions the action of the verb is predicated of the patient (i.e., the patient is the subject). The focus here is on the patient or object. For the expression of color, nouns and adjectives are arguably more basic than verbs. The assumed direction of coda filling (rightward spread yielding a long vowel in nouns; leftward spreading yielding a geminate consonant in verbs II and IX) is consistent with other differences in the nominal and verbal morphology—nouns are predominately suffixing, verbs predominately prefixing. Ussishkin (2003) argues that all the derived verb stems can be analyzed as formed by affixation, with no need for fixed templates. Adopting the idea that stems II and III involve affixation of an empty mora, he then uses an optimality theory (OT) analysis to argue that prosodic features of the derived stems, such as foot structure and stress placement, follow from interaction of phonological constraints applicable to all words. This is in contrast to McCarthy and Prince’s (1990a, 1990b, 1995) hypothesis that morphological templates are defined in terms of units of prosody, which would imply that prosody is imposed (in some cases) by the morphology independent of the phonology. Another technical problem where a word-based approach has promise is in the stem vowel alternations in the basic or stem-I verb. In an R&P analysis, these vowels can be analyzed only as having a direct referential value (or as being part of a pattern that has direct referential value). But there appears to be an implicational relationship between the vowel of the perfect and imperfect: the stem vowel of the imperfect is either the same as that of the perfect (ya-ðhabu-ðahaba “go,” ya-kburu-kabura “be big”) or is different (ya-ḍribu-ḍaraba “beat,” yaktubu-kataba “write,” ya-lbasu-labisa “wear”). If it is different, then one vowel is high (u,i) and the other is low (a). Alternation occurs about 75% of the time (McOmber 1995). The nonalternating types can be explained by phonological factors in the a-a case (presence of a postvelar consonant) and by semantic factors in the u-u case

Morphology (stative meaning). Taking these factors into account, McOmber (1995) shows that the assumption of imperfect-toperfect directionality gives 90+% predictability for the verbs in Wehr (1979) versus, at best, only 72% for the alternative perfect-based derivation. Ratcliffe (1997, 2003a) adopts this directionality on the basis of the argument that the imperfect stem shows a variety of (phonologically nonpredictable) syllabic shapes (-CvC, -CvvC, and CCvC), consistent with the notion that it is a lexical entry, while the perfect shows a consistent (templatic) pattern (CvCvC-) (subject to regular phonological rules), which is consistent with the idea that this is a derived form whose shape can be predicted. An alternative analysis follows the Arab grammatical tradition in taking the perfect as the base of derivation. Guerssel and Lowenstamm (1996) propose that the alternation follows an apophonic path Ø 〉 i 〉 a 〉 u 〉 u. This analysis requires the assumption that there is no underlying a in forms like ḍar(a)ba, ya-ḍribu but that there is one in kataba, ya-ktubu and ignores the a-a apophony. However, this pathway is supposedly a typological universal, found in other Afro-Asiatic and also Germanic languages (see Bendjaballah 2006 and references cited therein).

4.3.2 Interaction of Phonology and Morphology R&P theory allows phonological explanations for morphological irregularity. For example, the observation that glide deletion in the environment a_a is a regular phonological process in Arabic [Baalbaki, “ALT I”] allows the inference that the 3rd sg. perfective qaama “he stood” reflects * qawama—same pattern, CaCaC, as kataba “he wrote” (Voigt 1988). However, this sort of phonological interference in morphology is the only type of phonology– morphology interaction that is acknowledged. A word-based approach permits a richer understanding of phonology–morphology interaction and offers a new solution to many old problems traditionally grouped under the label of “weak” or “biconsonantal roots.” A central idea is that there may be morphological explanations for some kinds of phonological irregularity. For example, there are a number of semantically basic nouns with only two consonants in the underived, singular form. These always acquire a third consonant (its quality determined by the surrounding vowels) in derived forms, such as broken plurals or denominal verbs: dam “blood,” pl. dimaaʔ (pattern CiCaaC, with the final C filled by glottal stop); ism “name,” pl. ʔasmaaʔ (pattern ʔaCCaaC); derived verb samma(y)a, yusammiy “to name” (pattern CaCCaC, yuCaCCiC, final C filled by /y/). By contrast, words with more than four consonants must lose a consonant to form broken plurals: zanbarak “(mechanical) spring,” pl. zanaabik; barnamij “program,” pl. baraamij. There is no phonological rationale for the absence of the glide or glottal stop in the underived forms in the first set of cases or for the absence of /m/ or /n/ in the derived forms in the second set of cases. Ratcliffe (1997, 2003a) explains these exceptions through the assumption that underived words are not formed on patterns while derived forms are. (Templatic constraints restricted to derived forms are indeed a phenomenon widely attested across world languages; Ratcliffe 2003a.) Thus, a word like dam, which has no further internal structure, must acquire a third consonant when mapped to a plural pattern. Since the pattern is by definition an invariant shape, composed of slots for a fixed number of consonants, it is the pattern that imposes or requires triliteralism (or quadriliteralism) and hence the “root.” In other words, if a language uses fixed patterns to express morphological categories, the phenomenon of the “root” will emerge naturally as a result of derivational processes and need not be specified as a separate, independent morphological category. Ratcliffe (1997, 2003a) extends this analysis to the verb and suggests that it explains the anomaly of the so-called 1-w verbs, like waṣala, ya-ṣilu “arrive.” R&P theory forces the supposition that the imperfect reflects underlying *ya-wṣilu, but as Voigt (1988) observed there is no phonological reason for deletion of –w- in these verbs. Not all verbs with –w- in the imperfect show absence of –w- in the perfect. Stem-initial y- in verbs never deletes. Under a word-based analysis, these verbs are the mirror image of biconsonantal nouns like dam: the imperfect twoconsonant form is the basic, underived form, and the –w- of the perfect is a default consonant required to fill out the perfect template (as well as the templates of other derived forms such as participles and derived verbs). The insertion of the default to the left of the stem in verbs and to the right in nouns is consistent with the general directional bias of the two categories, as observed already. Some of these verbs even appear to be very old derivations from primitive biconsonantal nouns, for example, wasama “to mark” (cf. ism “name”) (Ratcliffe 2001a). For other so-called weak verbs, those assumed to have an underlying glide in second or third position, Chekayri’s (2006: 168) conclusion is quite intriguing: “it has been shown that the distribution of [y] and [w] is predictable. That

Morphology is, the glide appearance in some forms of a given verb is the output of a derivation originating in a lexical vowel.” Simplifying somewhat for the purposes of exposition, the problem is that while R&P would predict two types each of 2nd and 3rd weak verbs (those with w and y, respectively) or perhaps six types (two glides times three possible stem vowels), what one finds is basically three types correlating with the lexical stem vowel of the imperfect (yaquulu “say,” yasiir “go,” yanaam “sleep”; yadʕuu “call,” yarmii “throw,” yalqaa “meet”). In particular, for verbs of the type ya-naamu, naama, it is difficult to make the case for an underlying “root glide.” One would expect imperfective * ya-nwamu or *ya-nyamu, as there is no phonological restriction against such sequences (cf. the normal color adjectives ʔaswad, “black,” ʔabyaḍ “white”). For verbs containing a geminate consonant, the R&P approach leads to an analysis of a verb like yamuddu “extend” as reflecting underlying *yamdudu, with metathesis. Gafos (2003) shows that the alternations in this category can be explained most naturally in an approach that takes the surface geminated stem as the basic form. The larger question in all of this is what speakers are actually doing when they identify a consonantal string for mapping to a template: identifying a root morpheme stored in the lexicon or extracting a phonologically defined string. A morpheme is standardly identified based on its recurrence in a set of words, like the -mit in English permit, submit, and remit. If the consonantal root is defined as a morpheme, then a root would be identifiable only where there was more than one word containing it. If the string is defined phonologically, however (such as “consonants in a word or stem”), a root is in principle extractable from any word. That speakers are identifying phonological strings rather than recurrent elements is supported by the fact that loanwords like bank “bank” (〉〉pl. bunuuk) or sijill “register” (〈Latin sigillum), (〉〉 sajjal “to register”) have been integrated into the system. Bat-El (1994) first makes this argument for Hebrew. If root extraction is approached as a problem of phonological parsing, we must ask what phonological features are relevant. Ratcliffe (2003a, 2004) proposes that the criterion for distinguishing a consonantal string for mapping purposes is relative sonority, as defined for Arabic by Angoujard (1990). Relative sonority, determines the potential occurrence of segments in different syllabic positions. (Nuclei must have a higher sonority, than onsets; codas must have a sonority, less than or equal to the nucleus.) And this in turn defines the distinction between consonants and vowels. Some relatively high sonority, segments (in Arabic the high vowels/glides u/w and i/y (IPA j)) can occur in both onset and nucleus position and hence are both consonants and vowels. Normally the stripping process will extract and carry over all segments with a sonority, lower than u/i. Thus, in ma∫ruuʕ 〉〉 pl. ma∫aariiʕ “plan” the coda /u/ will not map to an onset in the output because there are already four lower sonority, segments in the input and no output template has more than four C slots. If the input has fewer than three segments below sonority, u/i, speakers have the option of raising the sonority, bar to include vocalic codas. Thus, ya-muut “die” 〉〉 yu-mawwit “cause to die” the coda /u/ maps to an onset position (represented orthographically as w). If a word has more than four segments below sonority, of u/i, then the sonority, bar can be lowered to further exclude high sonority, consonants such as approximants and nasals (as in the cases like barnamij 〉〉 pl. baraamij). The theory predicts that problems will arise when the coda is maximum sonority /a/, which cannot be an onset. One solution is to map the coda /a/ to a default consonant, usually w ( qaalab 〉〉 pl. qawaalib “mold,” baab 〉〉 pl. ʔabwaab “door”). But there is abundant evidence that speakers have trouble processing CaaC type inputs, whether nominal or verbal, for purposes of derivation. About half of all noun stems with this pattern fail to undergo broken plural formation (Levy 1971). The total number of verbs with this stem shape in the imperfect is only around 30, versus over 200 with –CuuC and –CiiC (Chekayri 2007).

4.3.3 Variation in Dialectal Arabic Heath’s (1987: 12) treatment of Moroccan Arabic remains the most thorough attempt to describe a variety of Arabic without the assumption of derivation from a consonantal root: “underived noun, verb and other stems have a simple linear representation (or perhaps two or three slightly distinct ones, in the event of ambiguity). The linear representations include both Cs and Vs.” Heath proposes three specific mechanisms for derivation: fixed template; template plus projection; and local rules (ibid., 3). The first mechanism accounts for cases where traditional R&P would also work, while the latter two roughly parallel (and prefigure) McCarthy and Prince’s partial mapping and moraic affixation, respectively (1990a, 1990b, 1995). Heath (1987: 5) offers various reasons for taking this stance, but one of the most important in retrospect is that it

Morphology provides a framework for handling variation: “I often suggest that two or more distinct models… may have some degree of psychological validity for native speakers. Essentially I argue for the extension of variation models from the study of the social distribution of low-level surface forms … to the study of abstract representations and rules.” Heath is in many cases able to elicit a variety of alternative output patterns of a given category from a single input, especially in cases where less frequently encountered words or less frequently employed derivations were involved. Owens (1998: 199–204) reports a similar result in his work on Nigerian Arabic. As one example of the sort of problems Heath uncovered, consider the two alternate diminutives for kħəl “black”: kwħiħəl and kwħiyəl (153). In an R&P analysis, these forms would have to be described in terms of distinct patterns CCiyəC versus CCiCəC. However, Heath argues that they are better analyzed as different examples of the same pattern CCiCəC but with different strategies for mapping the input word to the pattern—spreading of a consonant versus insertion of a default consonant. Another kind of variation emerges as a result of diachronic changes such as regular loss of phonemic glottal stop. Verbs like ʔakala, yaʔkulu “eat” can be analyzed unproblematically as containing a three-consonant root ʔ-k-l in CA. The Moroccan imperfect yakel, with stable /a /after the person/number prefix reflects the regular phonological development ( yaʔkul 〉〉 yaakul 〉〉 yakəl). But elsewhere, the paradigm appears to incorporate two new roots: participle wakəl (like waṣəl “arriving,” apparent root w-k-l), and perfective kla (3rd sg.) klit (1st sg.) (like bda bdit “began,” with apparent root k-l-y). In other words, speakers seem to have resorted to various strategies for adding a third consonant to expand the inherited two-consonant sequence k-l to fill out the patterns of the derived forms. In so doing, they have created several new apparent “roots” from a single word (Heath 1987: 80, 168 n. 15). On the other side of the Arabic dialect continuum, Holes (2004: 99) wrestles with the distinct strategies that speakers of Eastern Arabian dialects have for deriving quadriliteral verbs. Quadriliteral gaṣgaṣ means to “chop something up … into small pieces” (= intensive action), whereas theme II gaṣṣaṣ is “to do a lot of cutting” (= extensive over time), both verbs being derived from gaṣṣ “cut.” Other examples such as hajjal and hajwal, both “to kick out,” seem to show that speakers of this dialect too have alternative strategies for mapping a threeconsonant input to a quadriliteral output. Watson (2006) describes a similar variety of strategies for forming quadriliteral verbs in San’ani. Another area where one sees multiple strategies for deriving output patterns is in the hypocoristic formations in Palestinian and Levantine dialects described by Davis and Zawaydeh (1999, 2001). (6) Name

Hypocoristic nawwuur

ʔanwar

nawwuur

diyma

damduum

dyaana

dayyuun

marwa

marmuur

raanya

rannuun

The hypocoristic is consistently CaCCuuC, but there appear to be a variety of ways for filling the C-slots. Recourse to an underlying root will work in the first case but not elsewhere. Ratcliffe (2004) shows that the variety of extraction strategies here can be rationalized in a sonority stripping analysis. As the CaCCuuC hypocoristic illustrates, the dialects not only show various strategies for mapping input word to output patterns but also a variety of apparently innovative patterns (not found in CA). These include new patterns for old categories, such as the quadriliteral diminutives in Moroccan (Heath 1987: 113–132; Ratcliffe 2001), exemplified previously, the Moroccan nouns of profession CCaCCi (Heath 1987: 139–152) and new broken plural patterns from Iraq to Morocco (Ratcliffe 2003b). They also include new patterns for categories not found in CA,

Morphology such as Moroccan diminutive verbs tCiCCeC (Heath 1987: 76), the presumably independent San’ani diminutive verbs (t)CayCaC (Watson 2006), new derived stems (e.g., CooCaC, tiCooCaC, CeeCaC, tiCeeCaC, yintiCaCCaaC) in Najdi (Ingham 1994), and the Levantine hypocoristics (Davis and Zewaydeh 1999, 2001), discussed above. The mechanism by which speakers generate these innovative patterns are of tremendous interest for what they reveal about the creative aspects of language and mind. Yet a theory that says that words are formed exclusively by combination of preexistent roots and patterns cannot explain them.

4.3.4 Historical Change When the dialects exhibit patterns, words, and roots that are not attested in CA, the most plausible assumption in most cases is that these forms were simply absent from the spoken forms of Arabic ancestral to the dialects during the period (8th–9th centuries) when the classical grammarians and lexicographers were active in codifying the language, that is, that such forms are innovations. Under this assumption two general trends can be said to characterize the historical morphology of Arabic (Carter 1996; Ratcliffe 2001a, 2003b, 2006). First, most inherited biconsonantals are reanalyzed to have three consonants in all parts of the paradigm. For example, nouns like CA ∫if-a “lip” (pl. ∫ifaah, ∫ifawaat) often become geminated like Iraqi ∫iffa (pl. ∫ifaaf, ∫ifaayif); CA verbs like ya-ṣil generally acquire a stable glide as in Egyptian yiwṣal. When new biconsonantals emerge as a result of sound change, these too are reshaped on the triconsonantal pattern by adding defaults, sometimes creating multiple new “roots” from a single word, as in the Moroccan “eat” case. Second, new patterns develop. One subtrend here is the generalization of quadriliteral patterns where CA has both triliteral and quadriliteral patterns, depending upon the structure of the input, as in the Moroccan diminutives (Ratcliffe 2001, 2006): (7) CA

Moroccan

kalb 〉〉 kulayb

kəlb

kliyəb

“dog”

kitaab 〉〉 kutayyib

ktab

ktiyəb

“book”

These developments presuppose the operation of analogy, which requires reasoning over sets of words. They show that morphological patterns, and therefore roots, function as abstract structures rather than as concrete items or objects. There is reason to think that similar processes of analogy and restructuring have given rise to new roots and patterns in the prehistory of Arabic. Larcher (2003: 576–577) presents a number of cases where words containing a long -aa have been reanalyzed to generate new apparent “roots.” For example, baal “mind,” where the tradition would expect a root b-glide-l, appears to be the source of baalaa “to mind” (stem III of root b-l-y). Also, ʔi∫aara “sign” (pattern ʔiCCaCa), verbal noun of ʔa∫aara (stem IV) “to indicate,” root ∫-w-r, seems to have been reinterpreted as pattern CiCaaCa from a root ʔ-∫-r, giving rise to a new form II verb ʔa∫∫ara “to indicate.” This type of reanalysis is a major source of apparent biconsonantal etyma (Zaborski 1991). Ratcliffe (1998) argues in detail that most of the patterns of the Classical Arabic broken plurals are the result of analogic processes. The greatest degree of allomorphy is found for plurals of singulars containing a long vowel: CvvCvC and CvCvvC. Ambiguity in parsing long vowels would explain, for example, why one gets both triliteral (ṭawiil 〉〉 ṭiwaal “tall”) and quadriliteral (ḍamiir 〉〉 ḍamaaʔir “pronoun”) plurals from inputs of the same stem shape.

4.3.5 Semantic Interrelationships Finally, another critique of R&P analysis emerges from considerations of meaning and interpretation (Larcher 1995, 2006; Watson 2006). While it is often noted that words kataba “he wrote,” kitaab “book,” maktab “office, desk,”

Morphology and maktaba “library, bookstore” all have something to do with writing and all contain the “root” k-t-b, Larcher (1995) points out that the semantic connections among these forms is subtler and more precise. Both maktab and maktaba are nouns of place, but the former is a place where writing is done while the latter is a place where books are kept. In other words, maktab derives its meaning from kataba and maktaba from kitaab. There is no noun of place for the root k-t-b. Moreover, knowing that kataba means “write” would not allow one to predict that kitaab means “book.” (Etymologically it should mean something like “letter.”) When a word acquires a specific, nonpredictable meaning like this, standard morphological theory assumes that it must be lexically listed. But R&P theory prohibits words from having a lexical listing. Watson (2006: 7) makes a similar point about derivation in San’ani. For example, taħaymar “to act like a donkey” is semantically derived from ħimaar “donkey,” not ʔaħmar “red.”

4.4 The Psycholinguistic Critique The principal critique of the word-based approach has come not in the form of offering alternative analyses of the technical problems that motivate this approach but rather in the form of arguments trying to prove the “legitimacy” or “psychological reality” of the consonantal root (see Prunet 2006 for a survey). The fact that the debate has taken this turn reveals a deep philosophical problem in linguistics regarding the ontological status of the abstract terms we use to describe linguistic phenomena. Personally, I am convinced by arguments of neurocognitive linguists like Lamb (1998) that such terms as root, morpheme, or even word have no neurobiological reality and therefore no objective existence. They are convenient fictions, roughly analogous to the equally fictional units of measurement that scientists use to describe the physical world. One can debate whether feet or meters are more appropriate for measuring the height of Mt. Everest. One can debate the technical definition of such units. And one can make falsifiable statements about reality using such fictions (“Mt. Everest is higher than 7000 meters”). But logically, one cannot claim that the height of Mt. Everest provides evidence for the reality of the meter as opposed to the foot (or vice versa). What many psycholinguists are actually doing is presenting data that they analyze in terms of a theory that rootmorphemes are stored in a mental lexicon, but without showing that alternative theories cannot explain the same data as well or better. This is something like saying that since we have chosen to measure Mt. Everest in feet, feet are real and no alternative measurement is possible. Nonetheless, this research raises a number of valid issues, notably: 1. Speakers seem to be aware of the relationship between words that share a root; see, for example, Boudelaa and Marslen-Wilson’s (2005) priming experiments. 2. Speakers manipulate (root?) consonants independently of vowels; see Prunet, Beland, and Idrissi’s (2000) discussion of aphasic errors. Both of these points seem reasonable on the basis of other evidence as well. That speakers are aware of relationships instantiated through the root is apparent from the phenomenon of root-echo responses (Stewart 1996): mabruuk “congratulations” 〉 allah yibaarik fiik “God bless you.” That speakers can manipulate consonants independently of vowels is clear from language games involving consonantal inversion (McCarthy 1981; Heath 1987: 184–197). (However, Wolfer’s [2011] survey of ludlings—secret languages made by changing words in systematic ways—shows that consonantal inversion is far from being the only way that speakers of Arabic dialects have to deconstruct and manipulate words.) Boudelaa and Marslen-Wilson (2005: 232) propose to account for the first type of phenomena through “a morphological parsing process in which morphemic units are extracted and located at sub-lexical level.” However, they argue that their experiments suggest a close link between words that share a purely phonological “root” (without semantic commonality) like kitaab “book,” katiiba “regiment.” This raises the question: Why not assume that parsing works on phonological principles (identifying consonants in a word) rather than by reference to a lexically listed “morphemic unit”? The standard reason for listing is that the relationship between form and meaning is arbitrary and hence must be learned. But if the root is simply a phonologically defined form without meaning, why does it need to be listed?

Morphology Prunet et al. (2000) argue that the domain of aphasic metathesis errors is the consonantal root. But as Ratcliffe (2004) points out, the domain of errors is not in fact the root as defined by lexicographical convention. Even the very small set of words in Prunet et al.’s sample exhibit the problems discussed in Sections 3.1 and 3.2. “Secondary” non-root consonants like the /w/ in plurals like qawaalib participate in metathesis errors. Glides participate only when there is a sonority contour present. Hypothetically underlying “root” glides do not participate. A sonority-based parsing analysis would seem to work as well or better here too. While these new sources of data are welcome, the basic problems of description and analysis remain.

4.5 Future Prospects There is probably much more to be said about the morphology of Arabic dialects—both in terms of basic documentation and diachronic explanation. Perhaps the most important impact of the word-based turn is that it has encouraged scholars to look for and try to analyze morphological phenomena in the dialects that cannot be easily explained in the R&P framework traditionally applied to CA. There is also probably more to be said about Arabic morphology in typological perspective—beyond the comparative work on apophony discussed by Bendjaballah (2006) and the work on fixed-output effects in various languages cited in Ratcliffe (2003a). The point here is not that all languages have to be the same but rather that as linguists we want a theoretical apparatus that allows us to describe similar phenomena in different languages in the same way. Language-particular traditions of analysis, like R&P, tend to obscure such similarities. No doubt there is quite a lot more to be said about Arabic morphology and mind. An important development in recent years has been the emergence of connectionism (Lamb 1998; Plaut and Gonnerman 2000). Connectionists argue that a grammar based on symbols and rules is not compatible with what is known about the brain and argue instead for a model of morphology as a network of interconnections analogous to a network of neurons. R&P, if taken literally as a theory of mind, represents a particularly extreme version of the symbols and rules approach: the lexicon contains only abstract (unpronounceable) symbols, and all surface words must be derived by rules combining these symbols. Word-based approaches are less abstract, and some theorists like Gafos (2003, 2009) are already describing Arabic morphology in terms of relationships rather than rules. But the main point of convergence is that both the word-based approach and connectionism eschew unique decompositionality and allow words to participate in multiple relationships: Traditional theories of lexical processing assume that … words are built out of discrete units called morphemes, and a given word either is or isn’t decomposed into constituent morphemes at each stage of processing …. Distributed connectionist modelling offers an alternative. Words are represented as alternative patterns of activity over multiple groups of units, and subpatterns of those patterns can exhibit varying degrees of stability and independence in a natural way. (Plaut and Gonnerman 2000: 478) Compare: “Morphological rules may be able to reference various phonologically-defined parts of the input word whether syllables or strings of syllabic constituents such as consonants and vowels” (Ratcliffe 1997: 151–152). Replace Ratcliffe’s “morphological rules” with “patterns of activity” and replace Plaut and Gonnerman’s “subpatterns” with Ratcliffe’s phonological specifics, and the road toward a conciliation of these approaches lies open. In the end, the success or failure of the word-based approach (as of any theoretical approach) must be evaluated not by such questions as “Is it true?” and “Should I believe it?” but by the questions “What (old problems) does it explain?” and “What (new problems) does it reveal?”

References Angoujard, Jean-Pierre. 1990. Metrical structure of Arabic. Dordrecht: Foris.

Morphology Bat-El, Outi. 1994. Stem modification and cluster transfer in Modern Hebrew. Natural Language and Linguistic Theory 12: 571–596. Bendjaballah, Sabrina. 2006. Apophony. In EALL I, ed. Kees Versteegh. Associate Editors: Mushira Eid, Alaa Elgibali, Manfred Woidich, Andrzej Zaborski, 119–123. Leiden: Brill. Benmamoun, Elabbas. 1999. Arabic morphology: The central role of the imperfective. Lingua 108: 175–201. ——. 2003. Reciprocals as plurals in Arabic. In Research in Arabic grammar II, ed. Jacqueline Lecarme, 53–62. Amsterdam: John Benjamins. Bohas, Georges. 1993. Diverses conceptions de la morphologie arabe. In Développements récents en linguistique arabe et semitique, ed. Georges Bohas, 45–59. Damascus: Institut français du Damas. ——. 1997. Matrices, étymons, racines: éléments d’une théorie lexicologique du vocabulaire arabes. Leuven: Peeters. Boudelaa, Sami. 2006. Cognitive linguistics. In EALL I, ed. Kees Versteegh. Associate Editors: Mushira Eid, Alaa Elgibali, Manfred Woidich, Andrzej Zaborski, 421–427. Leiden: Brill. Boudelaa, Sami and William Marslen-Wilson. 2005. Discontinuous morphology in time: Incremental masked priming in Arabic. Language and Cognitive Processes 20 (1–2): 207–260. Cantineau, Jean. 1950a. Racines et schèmes. In Melanges W. Marçais, 119–124. Paris: G.P. Maisonneuve et Cie. ——. 1950b. La notion de “schème” et son altération dans diverses langes sémitiques. Semitica 3: 73–83. Carter, Michael. 1996. Signs of change in Egyptian Arabic. In Understanding Arabic: Essays in honor of El-Said Badawi, ed. Alaa Elgibali, 137–144. Cairo: American University of Cairo Press. Chekayri, Abdellah. 2007. Glide. In EALL, II ed. Kees Versteegh. Associate Editors: Mushira Eid, Alaa Elgibali, Manfred Woidich, Andrzej Zaborski, 164–169. Leiden: Brill. Chekayri, Abdellah and Tobias Sheer. 1996. The apophonic origin of glides in the verbal system of Classical Arabic. In Studies in Afro-Asiatic grammar, ed. Jacqueline Lecarme, Jean Lowenstamm, and Ur Shlonsky, 62–76. The Hague: Holland Academic Graphics. Cowan, David. 1958. Modern literary Arabic. Cambridge, UK: Cambridge University Press. Davis, Stuart and Bushra Adnan Zawaydeh. 1999. A descriptive analysis of hypocoristics in colloquial Arabic. Langues et Linguistique 3: 83–98. ——. 2001. Arabic hypocoristics and the status of the consonantal root. Linguistic Inquiry 32: 512–520. Elgibali, Alaa (ed.). 1996. Understanding Arabic: Essays in honor of El-Said Badawi. Cairo: American University of CairoPress. Firth, John R. 1948. Sounds and prosodies. Transactions of the Philological Society 1948, 127–152. Reprinted in Frank Palmer 1970, Prosodic analysis, 1–26. London: Oxford University Press. Gafos, Adamantios. 2003. Greenberg’s asymmetry in Arabic: A consequence of stems in paradigms. Language 79: 317–355. ——. 2009. Stem. In EALL IV, ed. Kees Versteegh. Associate Editors: Mushira Eid, Alaa Elgibali, Manfred Woidich, Andrzej Zaborski, 338–339. Leiden: Brill. Guerssel, Mohand and Jean Lowenstamm. 1996. Ablaut in Classical Arabic measure I active verbal forms. In ed. Jacqueline Lecarme, Jean Lowenstamm, and Ur Shlonsky. Studies in Afro-Asiatic grammar, 123–34. The Hague: Holland Academic Graphics. Harris, Zellig. 1944. Simultaneous components in phonology. Language 20: 181–205.

Morphology Heath, Jeffrey. 1987. Ablaut and ambiguity: Phonology of a Moroccan Arabic dialect. Albany: State University of New York Press. Holes, Clive. 2004. Quadriliteral verbs in the Arabic dialects of Eastern Arabia. In Approaches to Arabic dialects, ed. Martine Haak, Rudolph de Jong, and Kees Versteegh, 117–134. Leiden: Brill. Ingham, Bruce. 1994. Najdi Arabic. Amsterdam: John Benjamins. Kaye, Alan, ed. 1991. Semitic studies in honor of Wolf Leslau. 2 vols. Wiesbaden: Harrasowitz. Lamb, Sydney M. 1998. Pathways in the brain. Amsterdam: John Benjamins. Larcher, Pierre. 1995. Où il est montré quʾen arabe classique la racine nʾa pas de sense et quʾil n y a pas de sens à dériver dʾelle. Arabica 41: 291–314. ——. 1999. Vues “nouvelles” sur la dérivation lexicale en arabe classique. In Tradition and innovation: Norm and deviation in Arabic and Semitic linguistics, ed. Lutz Edzard and Mohammad Nekroumi, 103–123. Wiesbaden: Harrasowitz. ——. 2006. Derivation. In EALL, I, ed. Kees Versteegh, associate ed. Mushira Eid, Alaa Elgibali, Manfred Woidich, and Andrzej Zaborski, 573–579. Leiden: Brill. Lecarme, Jacqueline, ed. 2003. Research in Arabic grammar II. Amsterdam: John Benjamins. Lecarme, Jacqueline, Jean Lowenstamm, and Ur Shlonsky (eds.). 1996. Studies in Afro-Asiatic grammar. The Hague: Holland Academic Graphics. Levy, Mary M. 1971. The plural of the noun in Modern Standard Arabic. PhD diss., University of Michigan. Maman, Aharon. 2004. Comparative Semitic philology in the Middle Ages. Trans. David Lyons. Leiden: Brill. McCarthy, John. J. 1979. Formal problems in Semitic phonology and morphology. PhD diss., Massachusetts Institute of Technology, Cambridge, MA. ——. 1981. A prosodic theory of non-concatenative morphology. Linguistic Inquiry 12: 373–418. ——. 1983. A prosodic account of Arabic broken plurals. In Current trends in African linguistics, ed. I. Dihoff, 289– 320. Dordrecht: Foris. ——. 1993. Template form in prosodic morphology. In Papers from the third annual formal linguistics society of Midamerica conference, ed. L. Smith Stvan, 127–218. Bloomington: Indiana University Linguistics Club. ——. 2008. Morphology. In EALL, III, ed. Kees Versteegh, associate ed. Mushira Eid, Alaa Elgibali, Manfred Woidich, and Andrzej Zaborski, 297–307. Leiden: Brill. McCarthy, John. J. and Alan Prince. 1990a. Foot and word in prosodic morphology: The Arabic broken plural. Natural Language and Linguistic Theory 8: 209–283. ——. 1990b. Prosodic morphology and templatic morphology. In PAL II, ed. Mushira Eid and John McCarthy, 1–54. Amsterdam: Benjamins. ——. 1995. Prosodic morphology. In The handbook of phonological theory, ed. John Goldsmith, 318–366. Malden, MA: Blackwell. McOmber, Michael L. 1995. Morpheme edges and Arabic infixation. In PAL VI, ed. Mushira Eid, 173–189. Amsterdam: Benjamins. Owens, Jonathan. 1988. The foundations of grammar: An introduction to medieval Arabic grammatical theory. Amsterdam: John Benjamins. ——. 1998. Neighborhood and ancestry: Variation in the spoken Arabic of Maiduguri, Nigeria. Amsterdam: John

Morphology Benjamins. Palmer, Frank R. (ed.). 1970. Prosodic analysis. London: Oxford University Press. Plaut, David and Laura Gonnerman. 2000. Are non-semantic morphological effects incompatible with a distributed connectionist approach to lexical processing? Language and Cognitive Processes 15: 445–485. Prunet, Jean François. 2006. External evidence and the Semitic root. Morphology 16: 41–67. Prunet, Jean François, Renee Beland, and Ali Idrissi. 2000. The mental representation of Semitic words. Linguistic Inquiry 31: 609–638. Ratcliffe, Robert. 1997. Prosodic templates in a word-based morphological analysis of Arabic. In PAL X, ed. Mushira Eid and Robert Ratcliffe, 147–171. Amsterdam: Benjamins. ——. 1998. The “broken” plural problem in Arabic and comparative Semitic: Allomorphy and analogy in nonconcatenative morphology. Amsterdam: John Benjamins. ——. 2001a. Analogy in Semitic morphology: Where do new roots and new patterns come from? In New data and new methods in Afroasiatic linguistics: Robert Hetzron in memoriam, ed. Andrzej Zaborski, 90–105. Paris: INALCO. ——. 2001b. What do “phonemic” writing systems represent: Arabic huruuf, Japanese kana and the Moraic principle. Written Language and Literacy 4: 1–14. ——. 2003a. Toward a universal theory of shape-invariant (templatic) morphology: Classical Arabic reconsidered. In Explorations in seamless morphology, ed. Rajendra Singh and Stanley Starosta, 212–269. London: SAGE. ——. 2003b. The historical dynamics of the Arabic plural system: Implications for a theory of morphology. In ed. Lecarme, 339–362. ——. 2004. Sonority-based parsing at the margins of Arabic morphology. Al-Arabiyya 37: 53–75. ——. 2005. Semi-productivity and valence marking in Arabic—the so-called “verbal themes.” In Corpus approaches to sentence structure, ed. Toshiro Takagaki, Susumu Zaima, Yoichiro Tsuruga, Francisco Moreno Fernandez, and Yuji Kawaguchi, 179–190. Amsterdam: John Benjamins. ——. 2006. Analogy. In EALL I, ed. Kees Versteegh, associate ed. Mushira Eid, Alaa Elgibali, Manfred Woidich, and Andrzej Zaborski, 74–82. Leiden: Brill. Schramm, Gene. 1962. An outline of Classical Arabic verb structure. Language 38: 360–375. ——. 1991. Semitic morpheme structure typology. In ed. Kaye, 1402–1408. Stewart, Devin J. 1996. Root-echo responses in Egyptian Arabic. In ed. Alaa Eligbali, 157–180. Ussishkin, Adam. 2003. Templatic effects as fixed prosody: The verbal system in Semitic. In ed. Lecarme, 511–530. Voigt, Rainer. 1988. Die infirmen Verbaltypen und das Biradikalismus-Problem. Stuttgart: F. Steiner. Watson, Janet C. E. 2006. Arabic morphology: Diminutive verbs and diminutive nouns in Sanʾani Arabic. Morphology 16: 189–204. Wehr, Hans. 1979. A dictionary of modern written Arabic. Ed. J. Milton Cowan. Wiesbaden: Harrasowitz. Wolfer, Claudia. 2011. Arabic secret languages. Folia Orientalia 47: 7–49. Zaborski, Andrzej. 1991. Biconsonantal roots and triconsonantal roots in Semitic: Solutions and prospects. In ed. Alan Kaye, 1675–1703. Robert R. Ratcliffe

Morphology Robert R. Ratcliffe, Tokyo University of Foreign Studies

Arabic Linguistic Tradition I: Naḥw and ṣarf Ramzi Baalbaki The Oxford Handbook of Arabic Linguistics Edited by Jonathan Owens Print Publication Date: Sep 2013 Online Publication Date: Dec 2013

Subject: Linguistics, Languages by Region DOI: 10.1093/oxfordhb/9780199764136.013.0005

Abstract and Keywords The Arabic linguistic tradition (also termed Arabic grammatical tradition) is the most extensive among the Arabic linguistic sciences. This article focuses on the two major branches of the grammatical tradition: nahw (which refers to grammar in general but more specifically to syntax); and sarf (morphology). Sections of the article cover early grammar and the origins of the grammatical theory, early works and Sibawayhi’s Kitaab, grammar from the 3rd/9th century onward, and the study of morphology in the Arabic linguistic tradition. Keywords: Arabic, inguistic, grammatical tradition, syntax, grammar, nahw, sarf, morphology

5.1 Introduction THE Arabic linguistic tradition (ALT, also termed Arabic grammatical tradition) is certainly the most extensive among

the Arabic linguistic sciences. Although the distinction between the two groups referred to as luġawiyyūn and naḥwiyyūn is often artificial and simplistic, as in Zubaydῑ’s (d. 379/989) Ṭabaqāt al-naḥwiyyῑn wa-l-luġawiyyῑn, the first group largely refers to philologists or lexicographers who, as of the 2nd/8th century, were concerned with collecting linguistic data and exploring the meaning of words in ġarῑb (strange usage) material and dialectal variants. This trend in linguistic study, which has surely contributed to the early appearance of lexicons as well as lexical collections based on subject rather than root (e.g., plants, animals, human body organs, weapons, natural phenomena; [Sara, “Classical Lexicography”]), survived in works such as al-Ṣāḥibῑ by Ibn Fāris (d. 395/1004), alMuẖaṣṣaṣ by Ibn Sῑda (d. 458/1066), and as late as the 10th century al-Muzhir by Suyūṭῑ (d. 911/1505). In contrast, the term naḥwiyyūn refers to the grammarians par excellence, that is, scholars whose interest mainly lies in the realm of syntax, morphology (including morphophonology), and, to a lesser extent, phonetics. A third group of scholars is known as the balāġiyyūn (rhetoricians, [Larcher, “ALT II”]). In spite of the great affinity between their discipline and grammar so much that one of their most prominent authors, Ğurğānῑ (d. 471/1078), bases his whole theory of word order and semantic and syntactic interrelationships (i.e., naẓm) on the meanings that naḥw provides (Dalāʾil 64, 176, 282, 310, 403–404), the approach of the balāġiyyūn to linguistic analysis and their subject matter differ significantly from those of the naḥwiyyūn (Baalbaki 1983). To the exclusion of the fields of philology, lexicography, and rhetoric, this chapter deals with the two major branches of the grammatical tradition, namely, naḥw (which refers to grammar in general but more specifically to syntax) and ṣarf (morphology).

5.2 Early Grammar and the Origins of the Grammatical Theory Most biographical sources attribute the founding of Arabic grammar to Abū l-Aswad al-Duʾalῑ (d. 69/688). Other early figures credited with laying the foundations of grammar include Nasṣr b. ʿĀṣim (d. 89/708) and ʿ Abdalraḥmān b. Hurmuz (d. 117/735) (Sῑrāfῑ, Aẖbār 13; Zubaydῑ, Ṭabaqāt 21–27). But although these reports and the anecdotal material related to them can in no way be substantiated, they serve to highlight the link between early grammatical

Arabic Linguistic Tradition I activity and the need to attend to the “corruption” of speech (laḥn) that was blamed on nonnative speakers of Arabic as a result of the futūḥāt (conquests). Furthermore, the subjects that were claimed to have captured the interest of these early scholars, namely, the three parts of speech and the particles that govern nouns and verbs, are strongly connected with laḥn and argue that early grammatical activity aimed primarily at adherence to “correct” usage and avoidance of error. An oft cited anecdote, for instance, has it that what prompted Abū lAswad to author a book (waḍaʿa kitāban) on grammar is that his daughter addressed him by saying: (1) mā

a-šadd-u

l-ḥarr-i

what

EL-severer-NOM

DEF-heat-GEN

“What type of hot weather is most severe”?

In fact, she wanted to exclaim over the hotness of the weather and should have used the exclamatory phrase: (2) mā

a-šadd-a

l-ḥarr-a

what

EL-be severe

DEF-heat-ACC

“How hot it is!” (Sῑrāfῑ, Aẖbār 19; cf. Zubaydῑ, Ṭabaqāt 21).

At this stage it is clear that there was no real distinction between the various linguistic disciplines and that early interest in syntax and morphology was strongly related to other areas of Islamic scholarship and, in particular, the qirāʾāt (Quranic readings), ḥadῑṯ (prophetic tradition), fiqh (jurisprudence), and tafsῑr (exegesis). It is noteworthy that most of the grammarians before Sῑbawayhi were readers (qurrāʾ) and that many of them narrated ḥadῑṯ or were formally trained in it (Baalbaki 2008: 5). Sῑbawayhi (d. 180/796) himself comments on several qirāʾāt in his Kitāb, and, according to the biographical sources, an error he made in reciting a ḥadῑṯ prompted him to seek a discipline that would guard against linguistic error, hence his pursuit of grammar and subsequent authorship of the Kitāb (Zubaydῑ, Ṭabaqāt 66; Ibn al-Anbārῑ, Nuzha 54–55). But it is the influence of fiqh and tafsῑr on naḥw that has largely occupied scholars in the last few decades. Carter (1972) advanced the theory that the origins of Arabic grammar can be traced in the Islamic science of law and that “grammar has no meaning if it cannot be related to the practicalities either of Islamic doctrine or the power and influence of the grammarians in Islamic society” (Carter 1991: 9). His most compelling argument is that Sῑbawayhi’s criteria of linguistic correctness are expressed by ethical terms such as ḥasan (good) and mustaqῑm (right) and that a host of grammatical terms such as badal (substitute), šarṭ (condition), ḥadd (limit), aṣl (origin), and niyya (intention) can best be understood in the light of their employment in legal contexts. Carter’s “legal thesis” contributes significantly to our understanding of the interrelatedness of grammar and law, but it should be pointed out that the grammatical terms he cites are predominantly methodological terms rather than categorical ones (cf. Versteegh 1993: 35). Further clarification of the origins of grammatical terminology may be sought in the earliest extant Quranic commentaries. In this respect, Versteegh (1993: 196–197) demonstrates how several terms used by early exegetes constitute the link between everyday vocabulary and the later technical terminology. Examples include ẖabar (predicate), naʿt (attribute), māḍῑ (past tense), ism (noun), istifhām (question), taʿağğub (admiration), and iḍmār (deletion). It should be borne in mind, however, that our knowledge of the terminological situation of early grammar is quite limited. Baalbaki (2006), for example, discusses the set of previously unknown outlandish morphological terms that Muʾaddib, a 4th-century grammarian, uses in his Daqāʾiq al-taṣrῑf and argues that they belong to an earlier period. Whether or not these terms are peculiar to a group of scholars whose main preoccupation was morphological ʿilal (causes) remains an open question. But it is certain that a better

Arabic Linguistic Tradition I understanding of early grammatical terminology—hopefully with the emergence of hitherto unknown texts such as Muʾaddib’s—would have huge implications on our reconstruction of the beginning of grammatical activity, as many of the assumptions related to various aspects of early grammar (e.g., the Basran–Kufan polarization) rely heavily on our present knowledge of early terminology. Theories that ascribe the genesis of Arabic grammar to foreign influence also rely in part on terminological evidence. Merx (1889) argues that Arabic grammar is based on the Greek model that was available to the Arabs in translation of Greek treatises. He thus traces the terms for the three parts of speech, ism (noun), fiʿl (verb), and ḥarf (particle) to ónoma, rhèma, and súndesmos, respectively, and he links iʿrāb (declension) and ẖabar (predicate) to hellènízein/hellènismós and katègoroúmenon, respectively. According to Rundgren (1976), several basic grammatical terms, including naḥw, ṣarf, and qiyās, are direct translations of Greek terms or are inspired by Greek notions. The “Greek thesis” was harshly attacked by a number of scholars, most notably Carter, who convincingly argues that the absence of any reference to foreign influences in the indigenous accounts of Arabic grammar is a major flaw in the theory (Carter 1972: 72). Versteegh (1993: 25) proposes the possibility that the Arabs only borrowed some of the elements of Greek grammatical teaching since they “became acquainted with Hellenistic culture and scholarship in a watered down version as it was being taught in schools all over the Byzantine empire.” As far as terminology is concerned, Troupeau (1981) demonstrates that, based on Ibn alMuqaffaʿ’s (d. 142/759) epitome of the Hermeneutics, there is no conformity between primitive Arabic grammatical terminology and terminology of Greek logic. In short, attempts to trace the emergence of Arabic grammar to foreign influence, be it Greek, Syriac, or Indian, have shown the existence of interesting parallels in terminology and grammatical categories between Arabic grammar and foreign traditions but have largely failed to demonstrate the existence of massive borrowing from foreign sources in the early period. That certain terms and notions might have been borrowed from other traditions is of relatively little importance given that the distinctive characteristics of the ALT—including its essential terminology of syntactic position and its fundamental notions of dependency, hierarchy, suppletive insertion (taqdῑr), and original form or pattern (aṣl)— argue for predominantly native origins. Most of our primary knowledge about the pre-Sῑbawayhi grammarians comes from the Kitāb itself. Sῑbawayhi makes no reference to the early figures who later biographical sources credit with the establishment of grammar (i.e., Abū l-Aswad and his contemporaries). He quotes, however, the views of several later authorities who were either his predecessors or his contemporaries. ʿAbdallāh b. Abῑ Ishāq (d. 117/735) is quoted 7 times, ʿĪsā b. ʿUmar (d. 149/766) 20 times, Abū ʿAmr b. al- ʿAlāʾ (d. 154/770) 57 times, and Abū l-Haṭṭāb al-Aẖfaš al-Kabῑr (d.177/793) 58 times. But the two scholars who had the greatest impact on Sῑbawayhi were Yūnus b. Ḥabῑb (d. 182/798) and alHalῑl b. Aḥmad (d. 175/791). In fact, a note found in a copy of the Kitāb derived from Hārūn b. Mūsā (d. 401/1010) suggests, contrary to the biographical sources, that Yūnus and al-Halῑl were the only two “real” teachers of Sῑbawayhi (wa-muʿallimā Sῑbawayhi l-Halῑl wa-Yūnus; cf. Humbert 1995: 9). Yūnus is quoted 217 times in the Kitāb. From these, it is clear that he had a sophisticated system of grammatical analysis whose main features include (1) extensive use of taqdῑr as an analytical tool, (2) formulation of grammatical “rules” of universal validity, (3) reliance on anomalous examples in drawing conclusions or formulating “rules,” and (4) description of usage by employing terms that are characteristic of Sῑbawayhi’s appraisal of his own data, such as qabῑḥ (ugly), qalῑl (infrequent), ẖabῑṯ (repugnant), kaṯῑr (frequent), and wa[001]h (correct or better usage) (Baalbaki 1995: 126–129, 2008: 14–16). For his part, al- Halῑl was Sῑbawayhi’s principal and most influential teacher. There are 608 references to al- Halῑl in the Kitāb, and the amount of data that Sῑbawayhi reports on his authority is overwhelming indeed. Accordingly, it is practically impossible to examine the terminology and analytical tools and methods of either Sῑbawayhi or al-Halῑl in isolation of one another. Al-Halῑl’s immense influence on Sῑbawayhi is highlighted in the tradition. Al-Faẖr al-Rāzῑ (d. 606/1210), for example, on the authority of Suyūṭῑ(Iqtirāḥ 205–206), is reported to have written that Sῑbawayhi put together (ğamaʿa) in his book the data (ʿulūm, lit. sciences) that he derived (istafādahā) from al- Halῑl. Sῑbawayhi also mentions 21 times an anonymous group to which he refers as naḥwiyyūn. Views differ widely as to the exact meaning of the term (cf. Carter 1972: 76; Talmon 1982), but it is obvious that Sῑbawayhi almost invariably criticizes their views and at times even the views of Yūnus when he sides with them. The reason for this is most probably that he disapproves of their speculative approach and their keen interest in creating hypothetical forms and constructions that do not occur in speech in spite of their resemblance to attested usage. This may be corroborated by the fact that in his Kitāb al-ʿAyn al-Halῑl mentions a similarly anonymous group of scholars whom he calls naḥārῑr (pl. of niḥrῑr, skillful or learned) and whom he explicitly accuses of creating words that do conform

Arabic Linguistic Tradition I to Arabic word composition and patterns but that are neologisms that are not permissible (lā tağūz) in actual usage (Baalbaki 2008: 20).

5.3 Early Works and Sῑbawayhi’s Kitāb It is evident that the Kitāb does not emerge from a vacuum but rather builds on previous grammatical activity that Sῑbawayhi himself, let alone the later sources, refer to. Sῑbawayhi’s contemporaries as of the second half of the 2nd/8th century until the early 3rd/9th century (such as Aṣma ʿῑ, who is quoted in the Kitāb and died in 216/831 at the ripe age of 91) are reported in the biographical sources to have tirelessly collected linguistic data from the Bedouin—a process known as ğamʿ al-luġa “collection of variants.” There are no reports that Sῑbawayhi himself made any journey to the desert for data collection from the Bedouin, but his frequent references to them in statements like samiʿnā l-ʿArab (we heard the Arabs) or min afwāh al-ʿArab (from the mouths of the Arabs) strongly indicate that he had direct contact with native speakers whom he regarded as the ultimate source of the data he seeks, that is, kalām al-ʿArab (speech of the Arabs). Unlike the Kitāb, the extant sources authored by Sῑbawayhi’s contemporaries mainly belong to the lexicographical tradition but incidentally touch upon grammatical issues. These sources include the two root-based lexica by al- Halῑl, Kitāb al-ʿAyn, and by Abū ʿAmr al-Šaybānῑ (d. 206/821), Kitāb al-Ğῑm, and the thematically arranged lexicon of Abū ʿUbayd al-Qāsim b. Sallām (d. 224/838), alĠarῑb al-Muṣannaf. In addition to several less extensive works on various aspects of lexicography—such as laḥn al-ʿāmma (common errors), aḍdād (words with two contrary meanings), ištiqāq (derivation particularly of proper nouns), nawādir (rare usage), and amṯāl (proverbs)—two groups of books deserve special attention (cf. Baalbaki 2008: 26). The first of these are three linguistically oriented exegetical works that include a sizable body of grammatical material. These are Farrāʾ’s (d. 207/822) Maʿānῑ l-Qurʾān, Abū ʿUbayda Maʿmar b. al-Muṯannā’s (d. 209/824) Mağāz al-Qurʾān, and al-Aẖfaš al-Awsaṭ’s (d. 215/830) Maʿānῑ l-Qurʾ ān. Being Quranic commentaries, these works are structurally different from the Kitāb since their grammatical content is determined by the text they interpret, none of them offering a comprehensive and systematic study of grammar. The second group comprises two grammatical works attributed to contemporaries of Sῑbawayhi’s. The first, titled al-Ǧumal fῑ l-Naḥw, is erroneously attributed to al- Halῑl, whereas its real author is most probably Ibn Šuqayr (d. 317/929), as Tanūẖῑ (d. 442/1050) asserts (Tārῑẖ 48). The other title is Muqaddima fῑ l-Naḥw, and its attribution to Halaf al-Aḥmar (d. 180/796) is extremely doubtful and is not supported by the later grammatical or biographical sources. Even more doubtful is the attribution of grammatical works to pre-Sῑbawayhi grammarians, such as Šarḥ al-ʿilal to ʿAbdallāh b. Abῑ Isḥāq and Ikmāl and Ǧāmiʿ to Īsā b. ʿUmar. One can safely conclude that the Kitāb is the first unquestionably authentic book on Arabic grammar. Whether its approach represents a departure from earlier grammar is open to question. Talmon (2003) compares the grammatical teaching of al-Halῑl and Sῑbawayhi with the extra-Kitābian linguistically oriented sources of the 2nd/8th and 3rd/9th centuries and concludes that the two grammarians considered their teaching distinct from the main grammatical theory up to their time. He calls this theory “The Old Iraqi School of Grammar,” whose main exponent is the Kufan Farrāʾ. The theory was not restricted to the Kufan milieu since the two Basrans, Abū ʿUbayda and alAẖfaš al-Awsaṭ, also represent a grammatical tradition that is not identical with Sῑbawayhi’s. Accordingly, Talmon speaks of the “innovations” and “reformation” or even “revolution,” which al- Halῑl and Sῑbawayhi introduced to grammatical study. Baalbaki (2005: 413–416) points out that, although Talmon admits that the corpus of material presented in the sources is of “fragmentary character” and often talks of the “absence of concrete textual evidence” and of the data being “too meager for definite conclusions,” his hypothesis claims a global interpretation of the grammatical activity and the relations among the grammarians in the early period of Arabic grammar. To be sure, the efforts of al- Halῑl and Sῑbawayhi represent a major development in the history of Arabic grammar, yet it is certainly an exaggeration to talk of a “revolution” that took place at their hands or to claim that the theoretical differences between them and other grammarians are due to the different grammatical “traditions” or “schools” at this stage. Whatever the case may be, the issues presented in the Kitāb dominated the grammatical tradition for centuries after Sῑbawayhi. Largely overshadowing other works, the Kitāb was viewed with a kind of sanctity and was often referred to as Qurʾān al-naḥw (the Quran of grammar; cf. Abū l-Ṭayyib, Marātib 106), in a rare instance of associating the word Quran with something other than the Revealed Book. To quote Versteegh (1997: 39), “Without exaggeration one could say that the entire linguistic tradition in Arabic is nothing but a huge commentary on the Kitāb Sῑbawayhi.” With the surge in interest in Arabic linguistics over the last few decades, there has been increasing awareness of the centrality of the Kitāb to the ALT as a whole and

Arabic Linguistic Tradition I consequently serious appreciation of the ALT itself in the history of linguistic ideas. The bulk of the Kitāb comprises a large body of transmitted data that Sῑbawayhi reports and analyzes. This body of attested material generally known as samāʿ(lit. hearing) represents for him kalām al-ʿArab (speech of the Arabs), which falls under four major categories: the Quran; the prophetic traditions (ḥadῑṯ); the speech of the Bedouin (including proverbs and speech patterns or idiomatic expressions); and poetry. There are, of course, methodological problems associated with each of these categories (cf. Baalbaki 2008: 35–47), but Sῑbawayhi is keen to accommodate the peculiarities of each type within his overall system of grammatical analysis. In the case of poetry, for example, he devotes an early chapter to poetic license ( Kitāb I, 26–32) to highlight the inherent differences between šiʿr (poetry) and kalām (in the narrow sense of prose). He also refers repeatedly throughout the Kitāb to the difference between the two genres and discusses in separate chapters phenomena that are unique to poetic usage (ibid., I, 269–274; IV, 204–216). But Sῑbawayhi is not interested merely in reporting and describing attested linguistic phenomena; rather, he tries to justify them, to examine the relationships that exist among the constituents of structure, and to propose theoretical origins (aṣl, pl. uṣūl) from which forms and patterns might have developed. In this respect, he resorts to the notion of ʿamal (government) in interpreting various syntactical relations among the various constituents of structure and to the notion of taqdῑr (suppletive insertion; also implied in contexts in which the terms iḍmār (suppression), ḥaḏf (elision), and niyya (intention) appear in ascribing various formal and semantic aspects of the utterance to elements that are not uttered but are essential in analyzing structure). A detailed examination of his analytical methods is beyond the present scope, but a glimpse at some of the more fundamental of these is essential for achieving an appreciation of the uniqueness of his approach in the tradition as a whole. Sῑbawayhi adopts a number of methods and strategies in the analysis of his linguistic data. The distinction between norm and anomaly is obviously a vital preliminary step in the process of organizing the vast and often conflicting material at his disposal. His priority is to defend the norm, but since he almost universally accepts attested usage he also has to interpret and justify (at times even criticize but not dismiss) anomalous material. Commenting on the Tamῑmῑ usage ḏahaba ams-u (“Yesterday-NOM has gone”) instead of ḏahaba amsi, which occurs in all other dialects, he says that the norm in ams is to have a final –i (kasra), hence amsi. However, the Tamῑmῑs have modified this norm to make the word a triptote, and hence it takes nominative case in the Tamῑmῑ usage when it is agent (Kitāb III, 283). In another instance, he describes the form minhim (“from them”) in the dialect of Rabῑʿa, instead of minhum, as bad usage (luġa radῑʾa) but justifies its occurrence on phonological grounds (ibid., IV, 196– 197). By adopting the concept of “basic rule” (cf. Baalbaki 2008: 134–152), which is implicit throughout the Kitāb, Sῑbawayhi ensures that the usage he considers to be most common or most representative of, for example, a form, pattern, or particle is recognized as the actual manifestation of accepted norm and is not undermined by attested material that does not conform to it. The recognition of “basic rules” allows Sῑbawayhi to make his data much more manageable than would have been the case had he adopted an indiscriminate approach that lends equal weight to the normal and anomalous. The practical side to the distinction between norm and anomaly is that Sῑbawayhi allows the generation of material through qiyās (analogy) on the basis of the former but not the latter. His approach to the norm–anomaly dichotomy may be viewed as part of his overall effort to demonstrate the coherency of his data and consequently the ability of his analytical system to highlight this coherency. In this system that not only describes usage but also analyzes and interprets it, a ʿilla (cause) has to be sought for the major phenomena as well as for the minutest details of attested speech. A relatively small number of ʿilal are used to unveil what the later linguists refer to as the ḥikma (wisdom) that underlies kalām al-ʿArab. Among these ʿilal are kaṯra (frequency), ẖiffa (lightness), ḥaml ʿalā l-akṯar (analogy based on the more frequent usage), saʿat al-kalām (latitude of speech), and ʾilm al-muẖāṭab bihi (the addressee’s knowledge of an implied element). For instance, taẖfῑf (lightening) is cited as the reason for tarẖῑm (euphonic elision), that is, the elision of the final parts of certain words, mainly in the vocative. An example of this is in the proper noun Ḥāriṯ, which may be changed to Ḥāri without the final consonant in yā Ḥāri “O Ḥāri”(Kitāb II, 239–241). Until a few decades ago, early Arabic grammar, including Sῑbawayhi’s Kitāb, was judged in the light of later sources and commentaries rather than the early sources themselves. An unfortunate consequence of this method is the assumption that Sῑbawayhi, like later authors, is largely preoccupied with lafẓ (form) at the expense of maʿnā (meaning). As scholars started to focus on examining the text of the Kitāb, there emerged an increasing awareness of the interrelatedness of form, lafẓ and meaning, maʾnā in Sῑbawayhi’s method of analysis.

Arabic Linguistic Tradition I Expressions of the type maʾnā l-naṣb/al-fiʾl/ al-fāʾ/al-tanwῑn/al-taʿağğub (the meaning of the accusative/the verb/the conjunctive fāʾ/the nunation/the exclamation; Kitāb I, 320, 310; III, 68; II, 229; I, 328, respectively) clearly demonstrate the inseparability of form and meaning. A most telling example is the chapter on the fāʾ, which is followed by the subjunctive Kitāb III, 28–41). Sῑbawayhi proposes the construction (3) lā

ta-ʾtῑ-nῑ

fa-tu-ḥaddiṯ-a-nῑ/

fa-tu- ḥaddiṯ-u-nῑ

not

you-come-me

and-you-converse-SBJ-me/

and-you-converse-INDC-me

and ascribes two possibilities of meaning to each of the subjunctive and indicative. In the subjunctive, the sentence means either “You do not visit me, so how can you converse with me”? or “You visit me often, but you do not converse with me,” whereas in the indicative it means either “You neither visit me nor converse with me” or “You do not visit me, and you are conversing with me now” (cf. Baalbaki 2001). In an essentially experimental fashion, he then introduces various changes to the model sentence and illustrates how the case of the verb after fāʾ is linked to a specific meaning in each construction. For example, he discusses the meaning of both the subjunctive and indicative in each of the following related constructions: (4a) mā

ta-’tῑ-nā

fa-takallam-a/ fa-takallam-u

illā

not

you-come-us

and-you-speak-SBJ/ and you-speak-INDC

except

bi-l-ğamῑl-i with-DEF-courtesy-GEN “You never visit us and speak but courteously” (4b) lā

ta-’tῑ-nā

fa-tu-ḥaddiṯ-a-nā/ fa-tu- ḥaddiṯ-u-nā

not

you-come-us

illā

izdad-nā

fῑ-ka

raġbat-an

except

increase-our

in-you.M

interest-ACC

and-you-converse-SBJ-us/ and-you- converse-INDC-us

“You never visit us and converse with us without us becoming more interested in you” and (4c)

Arabic Linguistic Tradition I

wadd-a

law ta-’tῑ -hi

fa-tu- ḥaddiṯ-a-hu

wished-he if

you-come-him

and-you-converse-SBJ-him/

fa-tu-ḥaddiṯ-u-hu and-you-converse-INDC-him “He wished that you would visit him so that you converse with him.”

The claim that Sῑbawayhi’s grammar lacks any systematic semantic component (Itkonen 1991: 148) is thus contradicted by the fact that maʿnā is inextricably linked to lafz in his analysis of structure. Furthermore, maʾnā in the Kitāb is directly related to the intention of the speaker (mutakallim) and the message he seeks to impart to his addressee or listener (muẖāṭab). The role of the linguist is thus to trace the mental operations that accompany the utterance and that dictate its formal and semantic characteristics. Sῑbawayhi’s analysis of language as interaction between the speaker and the listener, that is, as social behavior that takes place in a specific context, is unmatched in the ALT. Successful communication becomes a type of social obligation whose fulfillment depends on the speaker’s competence in deciding which utterance can best express his intentions as well as on the listener’s competence in the analysis of the utterances he is addressed with and, if necessary, in responding to them correctly. The issue of whether the theoretical frameworks proposed by Sῑbawayhi (and largely adopted by later grammarians; see Section 5.4) resemble certain modern linguistic theories has been hotly contested. From the typological point of view, for example, Bohas et al. (1990: 38) argue that Sῑbawayhi’s grammatical system belongs to a class which analyzes utterances “in terms of operations performed by the speaker in order to achieve a specific effect on the allocutee.” This is generally true of Sῑbawayhi’s analysis of speech and its context of situation. It should be remembered, however, that he at times examines constructions of considerable complexity that he formulates (cf. Kitāb II, 404–406) but that certainly are not used and hence have absolutely no communicative value. One such example is as follows: (5) ayy-u

man

ya-’ti-nā

nu-‘ṭi-hi

nu-krim-u-hu

which-NOM

who

he-come.JSV-us

we-give.JSV-him

we-honor-INDC-him

(roughly translated as “Whom—if he comes to us we give him—shall we honor”?)

Scarce as they are, such constructions—which seem to be precursors of grammatical drills extensively used by Mubarrad (d. 285/898) and other later grammarians—are probably meant to illustrate the correct syntactical positions that the various components of the structure ought to occupy (Baalbaki 2008: 217) and hence do not negate Sῑbawayhi’s general purpose of describing kalām al-ʿArab so that it can be replicated correctly. In the later tradition, one finds pedagogical examples in which extended sentence structures are constructed showing a large panoply of governed and governing elements in their respective functions. In the following example from the 4th-/10th-century grammarian Ṣaymarī (Tabṣira I, 123), the verb is shown governing nine different clausal dependents, eight accusative and the nominative governance of the agent. (6)

Arabic Linguistic Tradition I

a‘lam-tu zayd-an ‘amr-an munṭaliq-an i‘lām-an yawm-a l-ǧumu‘at-i ‘ind-a-ka ḍāḥik-an ḥidār-a šarr-i-hi informI

ZaydACC

Amr-ACC leave-ACC

knowing-ACC day-ACC

Friday-GEN atACC-your

laughing-ACC fearing-ACC evil-GEN-his

“Friday at your place laughing, I seriously informed Zayd that Amr was leaving out of fear of his evil.”

From a different perspective, Carter (1973) argues that Sῑbawayhi, in reducing the language to a set of functions, uses a method that has intrinsic similarities with immediate constituent analysis. Carter, however, insists that the parallels between Arabic grammar and Western linguistics—as in comparing reconstructed underlying forms and the ordered rules of deriving the passive verb in Arabic with notions of deep structure and transformation—are purely coincidental similarities between unrelated linguistic systems (Carter 1994: 386). For his part, Owens (1988) concludes that the Arabic grammatical theory of governance is based on a form of dependency grammar, although there are differences between the two. Whereas interesting, at times even remarkable parallels do exist between the Arabic grammatical model and modern linguistics, the question is how relevant the latter is for understanding the former. For the purpose of ordering the Arabic linguistic theory among the many approaches known to linguistic description (cf. Owens 2000a: 119), comparison with modern linguistics can yield results that should interest any student of the history of linguistic ideas. The danger remains, however, that in looking for parallels with modern linguistics, researchers may be tempted to fully identify the Arabic grammatical theory with a particular Western approach and become oblivious to the unique sociohistoric context of Arabic grammar and the issues that are specific to the data on which it is based and to the Islamic disciplines to which it is related.

5.4 Grammar from the 3rd/9th Century Onward Sῑbawayhi’s basic analytical tools and methods were never really challenged throughout the tradition, except for Ibn Maḍāʾ’s (d. 592/1196) unique attempt to refute some of the basic axioms of the grammatical theory, including ʿamal (government), qiyās (analogy), taqdῑr (suppletive insertion), and the speculative type of ῑilal. There were, of course, questions on which individual grammarians did not see eye to eye with Sῑbawayhi, perhaps the most famous of which are the 134 masʾalas (issues) that Ibn Wallād (d. 332/944) in his Kitāb al-Intiṣār (cf. Bernards 1997) interprets as disagreement between the two Basran grammarians Sῑbawayhi and Mubarrad (d. 285/898). The Basran–Kufan divide is interesting in this respect. Although biographical and grammatical sources of the 4th/10th century onward often generalize the differences between two established maḏhabs (“schools”; cf. Carter 2000 for characteristics of schools), about a third of Ibn al-Anbārῑ’s (d. 577/1181) masʾ alas in his Inṣāf are actual points of disagreement between the Kufan grammarian Farrāʾ (d. 207/822) on one hand and the two main Basran figures, Sῑbawayhi and Mubarrad, on the other (Baalbaki 1981). This notwithstanding, Farrāʾ and the other Kufans use the same analytical tools and methods as the ones used by Sῑbawayhi. Thus, while they might differ with him or other Basrans on how rigorously in a certain case qiyās should be applied or samāʿ should be admitted, they never challenge the very use of qiyās or samāʿ and indeed any of the other tools of analysis—such as taqdῑr, ʿilla, and aṣl —Sῑbawayhi so methodically applies in the Kitāb. Grammarians after Sῑbawayhi viewed his Kitāb as the ultimate, often irrefutable source of truth in grammatical matters, but this does not mean that the ALT was static after Sῑbawayhi [Larcher, “ALT II”]. We propose here to highlight the major developments that took place after Sῑbawayhi, particularly in the realm of naḥw. Ṣarf, which witnessed relatively little change throughout the tradition, will appropriately be discussed in a separate section. A marked development in approach, but always within the parameters of Sῑbawayhi’s grammatical theory, is already obvious in the first major work after Kitāb, namely, Mubarrad’s Muqtaḍab. Mubarrad generally relies more than Sῑbawayhi on qiyās and thus often dismisses as impermissible usage that is attested through samāʿ (e.g.,

Arabic Linguistic Tradition I Muqtaḍab I, 269–270; II, 146, 249, 336). He even dismisses qirāʾāt (Quranic readings) for being incompatible with qiyās and ridicules the qurrāʾ (readers) who introduced them (ibid., I, 123; II, 134, 316; IV, 105, 124–125, 195). Compared with Sῑbawayhi, furthermore, there is in Mubarrad’s approach to the notions of ʿamal (government) and taʿlῑl (causation) an increase in the level of “complexity” and “sophistication.” His terminology related to ʿamal reveals a tendency to classify and describe the ʿawāmil (operants) and to refer to them as a class in its own right. His terms include ʿawāmil al-afʾāl (operants that govern verbs), ʿawāmil al-asmāʾ (operants that govern nouns), taṣarruf al-ʿāmil (plasticity of the operant), al-ʿaṭf ʿalā ʿāmilayni (here, elision of two operants after a conjunction), and bāb al-ʿawāmil (class of operants; ibid., II, 6, 10; IV, 195, 300,(cf. Mubarrad, Kāmil I, 287; III, 99), 317, respectively). Similarly, the term ʿilla, in addition to its original sense of “cause,” acquires the sense of “quality” or “detail” (Baalbaki 2008: 246–247) and can thus be used to refer to linguistic phenomena rather than to their causes. Thus, an expression like wa-hāḏā yušraḥ fῑ bāb ʿalā ḥiyālihi bi-ğamῑ:ʿ ʿ:ilalihi (and this will be explained with all its details in a separate chapter; ibid., II, 275) shows how ʿilal become inseparable from the phenomena for which they are recognized as causes. Kitāb al-Uṣūl fῑ l-naḥw, authored by one of Mubarrad’s students, Ibn al-Sarrāğ (d. 316/929), represents a remarkable departure from the Kitāb, not in terminology and content but in arrangement and style of presentation and argumentation. In the book’s introduction, Ibn al-Sarrāğ distinguishes between two types of ʿilal, the first of which comprises grammatical facts through which the speech of the Arabs is learned (e.g., that every agent is nominative), whereas the second—called ʿillat al-ʿilla (meta- ʿilla)— unveils the ḥikma (wisdom) embedded in that speech (e.g., for what reason is the agent nominative; Uṣūl I, 35). In a conscious effort at classifying grammatical material, he systematically separates the main or fundamental issues of a certain bāb or topic (hence, uṣūl) from the subsidiary or less essential questions or problems (hence, furūʿ). The distinction between uṣūl and furūʿ goes back to Mubarrad’s teacher, Māzinῑ (d. 249/863), in his book on morphology called Taṣrῑf. Mubarrad himself generalizes this distinction to the realm of syntax in his Muqtaḍab, but it is Ibn al-Sarrāğ who opens the door wide for the systematic organization of the material according to the principle of uṣūl and furūʿ. In this sense one can appreciate the famous statement that, by applying his uṭūl, Ibn al-Sarrāğ rationalized a previously “insane” grammar (mā zāla l-naḥw mağnūnan ḥattā ʿaqqalahu Ibn al-Sarrāğ bi-uṣūlihi; Yāqūt, Muʿğam VI, 2535). The interest that both Mubarrad and Ibn al-Sarrāğ show in taʿlῑl and uṣūl has surely contributed to the emergence of the two closely related genres of grammatical writing that, unlike the more widespread surveys of syntax and morphology, deal exclusively either with ʿilal assigned by the grammarians practically for every linguistic phenomenon or with the uṣūl, that is, the theoretical and methodological issues on which the discipline of grammar rests. One of the first books on ʿilal is authored by a Muʿtazilite, Zağğāğῑ (d. 337/949) and is titled al-Īḏāḥ fῑ ʿilal alnaḥw (cf. Versteegh 1995 for translation and comments). Zağğāğῑ distinguishes between three types of ʿilal, namely, taʿlῑmiyya (pedagogical), qiyāsiyya (analogical), and ğadaliyya naẓariyya (argumentational-theoretical; Īḍāḥ 64–66). The first type is descriptive and comprises the grammatical rules of Arabic such as might be used in classroom instruction. For example, in the construction inna Zayd-an qāʿim-un (Indeed Zayd is standing), the accusative in the noun of inna and the nominative in its predicate are both attributed to inna itself. The second type provides explanations for the rules, based on the principle of analogy (qiyās). Hence, the fact that inna causes its noun to be in the accusative is said to be due to its resemblance to transitive verbs that likewise cause the direct object to be in the accusative. Theoretical arguments related to grammatical debates are represented by the third type. One can thus debate here the nature of inna’s resemblance to transitive verbs and the reason for the occurrence of the accusative noun (Zayd-an) linearly before the nominative one (qāʿim-un). Another Muʿtazilite, Ibn Ǧinnῑ (d. 392/1002), raises in his Haṣāʿ iṣ fundamental issues of methodology and epistemology and tirelessly tries to demonstrate the inherently organized and harmonious nature of Arabic and the ḥikma that underlies the speech of the Arabs in an almost unprecedented manner in the tradition (for the main features of Haṣāʾiṣ, see Guillaume 2000: 276–279). He also argues that ʿilal of the grammarians are closer to those used by the theologians (mutakallimūn) than to those used by jurists (mutafaqqihūn; Haṣāʾ iṣ I, 48). In spite of the adherence of Zağğāğῑ and Ibn Ǧinnῑ to Muʿtazilite theology and logic, they both fully subscribe to the principle of the autonomy of grammar. In fact, Zağğāğῑ clearly demarcates the objectives of the logicians from those of the grammarians (Īḍāḥ 48), and Ibn Ǧinnῑ explicitly distinguishes between ta ʿlῑl in grammar and jurisprudence but recognizes certain similarities between the two neighboring disciplines of grammar and logic with respect to the notion of ta ʿlῑl on which they converge (Suleiman 1999: 66). Also worthy of mention under the genre of ta ʿlῑl are Ibn al-Warrāq’s (d. 381/991) ?Ilal al-naḥw, Ibn al-Anbārῑ’s (d. 577/1181) Asrār al-ʿArabiyya, and ʿUkbarῑ’s (d.

Arabic Linguistic Tradition I 616/1219) al-Lubāb fῑ ʿilal al-iʿrāb wa-l-bināʿ, all of which exhaustively cite causes for the phenomena of naḥw and, to a lesser extent, ṣarf. Much of the material known as uṣūl al-naḥw appears in the previously cited works on ta ʿlῑl. Among the more famous works that specifically deal with uṣūl are Ibn al-Anbārῑ ’s Lumaʿ al-adilla fῑ uṣūl al-naḥw and al-Iġrāb fῑ ğadal al-iʿrāb and Suyūṭῑ’s (d. 911/1505) al-Iqtirāḥ fῑ ʿilm uṣūl al-naḥw. Both authors define uṣūl al-naḥw as a discipline that investigates the adilla (proofs) of grammar (Lumaʿ 80; Iqtirāḥ 27). According to Ibn al-Anbārῑ (Lumaʿ 81 ff.), the three adilla that make up uṣūl al-naḥw are naql (transmitted data), qiyās (analogy), and istiṣḥāb al-ḥāl (presumption of continuity; i.e., continuing to apply a certain ʿilla until it can be proven no longer applicable). To the other two components, naql (or samāʿ) and qiyās, Suyūṭῑ adds iğmāʿ (consensus) on the authority of Ibn Ǧinnῑ. Accordingly, methodological issues related to these notions prevail in the uṣūl tradition, such as the following questions: What dialects are admissible to the corpus? What are the time limits for accepting transmitted data? Is consensus among grammarians an irrefutable proof of sound judgment? Should analogical extension apply to anomalous material? Which has precedence in case of conflict: samāʿ or qiyās? In addition to such issues, Suyūṭố identifies 24 types of ʿilla (Iqtirāḥ 115–119) and devotes a special section of his book to what he calls masālik alʿilla, that is, methods for establishing a ʿilla, such as reliance on explicit or implicit mention of a certain ʿilla by a trustworthy native speaker (ibid., 137–149; see detailed study in Suleiman 1999: 182–194). To turn back to the 4th/10th century, it was clear by that time that grammar as a discipline was becoming increasingly speculative and that grammatical argumentation was impacted by the introduction of logic and philosophy to the linguistic sphere. A most striking example from the period is the Muʿtazilite grammarian Rummānῑ (d. 384/994), better known for his extensive commentary on the Kitāb in which he appears to be heavily influenced by logic. In his biography on Rummānῑ, Ibn al-Anbārῑ (Nuzha 233–235) notes that he, Rummānῑ, used to mix his writings with logic and cites Abū ʿAlῑ al-Fārisῑ (d. 377/987) as saying that his own grammatical teaching and Rummānῑ’ s cannot be the same thing. The period from the 5th/11th century onward abounds with famous grammarians, some of whose works are still used today in some traditional institutions as textbooks for teaching grammar. Among these are Ibn Bābashāḏ (d. 469/1077), Šantamarῑ Another Muʿtazilite, Ibn Ǧinnī (d. 392/1002), raises in his Haṣāʾiṣ fundamental issues of methodology and epistemology and tirelessly tries to demonstrate the inherently organized and harmonious nature of Arabic and the ḥikma that underlies the speech of the Arabs in an almost unprecedented manner in the tradition (for the main features of Haṣāʾiṣ, see Guillaume 2000: 276–279). He also argues that ʿilal of the grammarians are closer to those used by the theologians (mutakallimūn) than to those used by jurists (mutafaqqihūn; Haṣāʾiṣ I, 48). In spite of the adherence of Zağğāğī and Ibn Ǧinnī to Muʿtazilite theology and logic, they both fully subscribe to the principle of the autonomy of grammar. In fact, Zağğāğī clearly demarcates the objectives of the logicians from those of the grammarians (Īḍāḥ 48), and Ibn Ǧinnī explicitly distinguishes between taʿlīl in grammar and jurisprudence but recognizes certain similarities between the two neighboring disciplines of grammar and logic with respect to the notion of ta‘līl on which they converge (Suleiman 1999: 66). Also worthy of mention under the genre of taʿlīl are Ibn al-Warrāq’s (d. 381/991) ʿIlal al-naḥw, Ibn al-Anbārī’s (d. 577/1181) Asrār al-ʿArabiyya, and ʿUkbarī’s (d. 616/1219) al-Lubāb fīʿilal al-iʿrāb wa-l-bināʾ, all of which exhaustively cite causes for the phenomena of naḥw and, to a lesser extent, ṣarf. Much of the material known as uṣūl al-naḥw appears in the previously cited works on taʿlīl. Among the more famous works that specifically deal with uṣūl are Ibn al-Anbārī’s Lumaʿ al-adilla fī uṣūl al-naḥw and al-Iġrāb fī ğadal al-iʿrāb and Suyūṭ’s (d. 911/1505) al-Iqtirāḥ fố ʿilm uṣūl al-naḥw. Both authors define uṣūl al-naḥw as a discipline that investigates the adilla (proofs) of grammar (Lumaʿ 80; Iqtirāḥ 27). According to Ibn al-Anbārī (Lumaʿ 81 ff.), the three adilla that make up uṣūl al-naḥw are naql (transmitted data), qiyās (analogy), and istiṣḥāb al-ḥāl (presumption of continuity; i.e., continuing to apply a certain ʿilla until it can be proven no longer applicable). To the other two components, naql (or samāʿ) and qiyās, Suyūṭī adds iğmāʿ (consensus) on the authority of Ibn Ǧinnī. Accordingly, methodological issues related to these notions prevail in the uṣūl tradition, such as the following questions: What dialects are admissible to the corpus? What are the time limits for accepting transmitted data? Is consensus among grammarians an irrefutable proof of sound judgment? Should analogical extension apply to anomalous material? Which has precedence in case of conflict: samāʿ or qiyās? In addition to such issues, Suyūṭī identifies 24 types of ʿilla (Iqtirāḥ 115–119) and devotes a special section of his book to what he calls masālik al-

Arabic Linguistic Tradition I ʿilla, that is, methods for establishing a ʿilla, such as reliance on explicit or implicit mention of a certain ʿilla by a trustworthy native speaker (ibid., 137–149; see detailed study in Suleiman 1999: 182–194). To turn back to the 4th/10th century, it was clear by that time that grammar as a discipline was becoming increasingly speculative and that grammatical argumentation was impacted by the introduction of logic and philosophy to the linguistic sphere. A most striking example from the period is the Muʿtazilite grammarian Rummānī (d. 384/994), better known for his extensive commentary on the Kitāb in which he appears to be heavily influenced by logic. In his biography on Rummānī, Ibn al-Anbārī (Nuzha 233–235) notes that he, Rummānī, used to mix his writings with logic and cites Abū ʿAlī al-Fārisī (d. 377/987) as saying that his own grammatical teaching and Rummānī’s cannot be the same thing. The period from the 5th/11th century onward abounds with famous grammarians, some of whose works are still used today in some traditional institutions as textbooks for teaching grammar. Among these are Ibn Bābashāḏ (d. 469/1077), Šantamarī (d. 476/1084), Zamaẖšarī (d. 538/1144), Ibn al-Dahhān (d. 569/1174), Ibn al-Anbārī (d. 577/1181), Suhaylī (d. 581/1185), Ibn Yaʿīš (d. 643/1245), Ibn al-Ḥāğib (d. 646/1249), Ibn ʿUsfūr (d. 669/1271), Ibn Mālik (d. 672/1274), Astarābāḏī (d. after 686/1287), Ibn al-Nāẓim (686/1287), Abū Ḥayyān (d. 745/1344), Ibn Hišām (d. 761/1360), Ibn ʿAqīl (d. 769/1367), Ušmūnī (D.C. 900/1495), and Suyūṭī (d. 911/1505). Broadly speaking, works from this period are characterized by a growing interest in formal (lafẓī) considerations at the expense of meaning. There was thus a considerable shift in focus, and Sībawayhi’s originality and his approach to language as a social interaction between a speaker and a listener in a defined context gave way to a pedantic and largely uninspiring approach. Sībawayhi’s terminology, analytical tools, and even his šawāhid (attested and often ascribed material) were largely preserved, but his insight into the pragmatic role he assigns to the speaker and listener and the delicate balance he establishes between form and meaning were no longer on the minds of most grammarians. The stereotypic nature of the basic grammars notwithstanding, some authors such as Ibn al-Ḥāğib and Astarābāḏī did show significant interest in speech acts [Larcher “ALT II”], while others, notably Suhaylī, strove to demonstrate that meaning has a larger role in determining case endings than is normally admitted in the mainstream theory (Baalbaki 2008: 290–297).

5.5 Morphology The terms ṣarf or taṣrīf (also ʿilm al-ṣarf/al-taṣrīf) designate the study of morphology in the ALT. Sībawayhi’s Kitāb is in roughly two equal parts, with naḥw preceding ṣarf. Most grammar books follow the same arrangement (e.g. Zağğāğī’s Ǧumal and Zamaẖšar’s Mufaṣṣal), although there are exceptions, such as Mubarrad’s Muqtaḍab, which starts with ṣarf but does not fully separate chapters that deal with it from those that deal with naḥw. Early in the tradition, Māzinī (d. 249/863) devotes his book entitled Kitāb al-Taṣrīf to morphological topics in isolation of syntax. This practice survived along with the more prevalent practice of including both naḥw and ṣarf under the same title. Among the most well-known works that are exclusively devoted to ṣarf are as follows: Ibn Ǧinnī’s (d. 392/1002) Munṣif, Sirr ṣināʿat al-iʿrāb, and al-Taṣrīf al-Mulūkī; Muʾaddib’s (d. after 338/949) Daqāʾiq al-taṣrīf; Ibn al-Ḥāğib’s (d. 646/1249) Šāfiya; Ibn ʿUṣfūr’s (d. 669/1271) Mumtiʿ; and Astarābāḏī’s (d. c. 686/1287) Šarḥ Šāfiyat Ibn alḤāğib. Throughout the tradition, the study of ṣarf was much less susceptible than naḥw to controversy or to differences in approach or rule formulation among the grammarians. Indeed, the same methods of morphological analysis adopted by Sībawayhi and Māzinī are applied by the later grammarians to yield very similar results. Ibn Ǧinnī (Munṣif I, 4) typically defines the domain of ṣarf as the fixed form of words (anfus al-kalim al-ṯābita) in contrast to naḥw, which deals with the changes that words undergo (aḥwālihi l-mutanaqqila) due to the effect that the operants have on their case endings. He illustrates this by citing the following three constructions: (7a) qāma

Bakr-un

stood up Bakr-NOM “Bakr stood up.”

Arabic Linguistic Tradition I (7b) raʾay-tu Bakr-an saw-I

Bakr-ACC

“I saw Bakr.” (7c) marar-tu bi-Bakr-in passed-I with-Bakr-GEN “I passed by Bakr.”

As far as naḥw is concerned, each of the three different operants causes a change in the case ending of Bakr. Yet as a word or lexical unit (kalima), Bakr remains the same throughout since it is otherwise unaffected by the various operants. But there are instances where the boundaries between naḥw and ṣarf are difficult to delineate. One such case is that of the passive participle, which as a morphological pattern belongs to ṣarf but which also belongs to naḥw by virtue of its being an operant that governs the agent of a passive verb (nāʾib ʿan al-fāʿil). It is interesting, for example, to see how the chapter on the agent of a passive verb in commentaries on Alfiyya (cf. Ibn ʿAqīl, Šarḥ 219–225) straddles both perspectives from which passive participles are examined. At another level, no distinction is made between phonology and morphophonology. In fact, phonetics is introduced in conjunction with one aspect of morphophonology, namely, assimilation (idġām/iddiġām), as in the Kitāb (IV, 431; cf. Owens 2000b: 296; [Embarki, “Phonetics”]; [Hellmuth, “Phonology”]; [Edzard, “Philology”]). Hence, Sībawayhi dealt with the points and manner of articulation in order to explain how certain characteristics of articulation can cause gemination or prevent it (Kitāb IV, 436, ll. 15–17). The study of morphology is generally divided into two parts, explicitly expressed in later works (e.g., Ibn ʿUṣfūr, Mumtiʿ I, 31–33; cf. Bohas and Guillaume 1984: 17; Bohas et al. 1990: 73). The first of these, taṣrīf (derivation), includes the total range of morphological patterns of verbs and nouns [Ratcliffe, “Morphology”]. Words are presumed to be derived from consonantal roots and belong to fixed patterns that often have semantic values. The pattern mafʿūl, for instance, is typically a passive participle, as in maktūb “written.” Note that some monographs, such as works on faʿala and afʿala, deal specifically with the meanings of certain patterns: for example, Aṣmaʿī’s (d. 216/831) Faʿala wa-afʿala (form I and form IV verbs); Ṣaġānī’s (d. 650/1252) Kitāb al-infiʿāl (form VII verbs). To express the relationship between a root and the various patterns derived from it, grammarians represent the root by the triliteral template f- ʿ- l, where the f stands for the first radical (fāʾ al-fiʿl), the ʿ for the second (ʿayn al-fiʿl), and the l for the third (lām al-fiʿl). The root is regarded as the “basic” component of a word, to which are added vowels and, at times, prefixes, infixes, and suffixes ([Hellmuth, “Phonology”] vs. [Ratcliffe, “Morphology”]). The notion of mīzān ṣarfī (morphological measure) is introduced to identify the patterns to which words belong. The mīzān takes into account the root of the word, its vowels, and, where applicable, its affixes. These 10 affixes, usually carrying morphemic value, are known as hurūf al-ziyāda (augmented consonants) and are combined in mnemonic phrases such as al-yawma tansāhu (Today you shall forget it), saʾaltumūnīhā (You asked me it), and hawītu l-simān (I loved the plump women). Accordingly, verbal and nominal patterns are identified based on the mīzān. For example, beginning with the form I verb “he hit,” ḍaraba, iḍṭaraba, istaḍraba, ḍārib, ḍarrāb, and munḍarib belong to the patterns faʿala, iftaʿala, istafʿala, fāʿil, fāʿʿāl, and munfaʿil, respectively, the added consonants indicated here in boldface. Numerous conventions also apply to the use of the mīzān (cf. Astarābāḏī, Šarḥ I, 10). For example, the fourth and fifth radicals in quadriliterals and quinqueliterals are expressed by reduplicating lām; hence, jaʿfar “brook” and safarğal “quince” are of the patterns faʿlal and faʿallal, respectively. Another rule is that, if the augmented consonant is a reduplication, it is represented by the relevant consonant in the template f- ʿ- l and not by the actual consonant of the word itself. Hence, qaṭṭaʿa “to tear apart” from qaṭaʿa

Arabic Linguistic Tradition I “cut,” is of the pattern faʿʿala and not *faʿṭala. In contrast, augmented consonants that are not reduplications are retained in the pattern, as in mustaqbal (future) and ʿanbas (lion), whose patterns are mustafʿal and fanʿal, respectively. The second part of morphology comprises the morphophonological rules that account for the change words undergo from a supposed aṣl (origin) without an accompanying change in meaning. In many cases, the aṣl is posed to demonstrate the rules of vowel mutation (iʿlāl). For example, qāma and yaqūmu are said to have as their aṣl *qawama and *yaqwumu. The surface form yaqūmu is thus presumed to have been derived from *yaqwumu through an intermediate stage involving metathesis, as in (8): (8) *yaqwumu = underlying form (aṣl)〉 *yaquwmu via metathesis (naql) of u-w 〉 yaqūmu = surface form

Although the likes of *qawama and *yaqwumu are almost never used as surface forms, their assumption, according to the grammarians, is based on analogy to their “sound” counterparts, such as kataba and yaktubu, which display no discrepancy in this instance between the basic (aṣl) and surface forms. It is further supported by the fact that the wāw surfaces in other derivatives such as qawwama “erect,” qāwama “resist,” and taqāwama “mutually resist.” Similarly, the aṣl of maqūl and mabīʿ are *maqwūl and *mabyūʿ, both on the analogy of the “sound” counterpart (e.g., maktūb), their surface form explicated by various morphophonological changes analogous to those in (8). Other than iʿlāl, these morphophonological processes include ibdāl (substitution), ziyāda (augmentation), haḏf (omission), idġām (gemination), qalb (mutation of glides), naql (metathesis between glides and vowels; see (8)), and waqf (pause). As far as the historicity of the proposed uṣūl is concerned, Ibn Ǧinnī clearly stated (Munṣif I, 190–191) that when grammarians consider *qawama to be the aṣl of qāma and *istaqwama the aṣl of istaqāma they simply state that the proposed aṣl would have been expected had the norm applied and do not refer to any prior stage of the language. Grammarians apply to ṣarf several of the strategies they employ in their study of naḥw. Section 5.3 pointed out that the notion of “basic rules” helps make the data more manageable, while at the same time anomalous material is usually interpreted and admitted to the corpus of acceptable linguistic usage. This is closely paralleled in ṣarf by the concept of ilḥāq(appending), which embraces extremely complex rules through which triliterals are said to be appended to quadriliterals and quinqueliterals and quadriliterals are appended to quinqueliterals (Baalbaki 2001– 2002). For example, ğadwal “creek” of the triliteral root ǦDL is appended to ğaʿfar of the quadriliteral root ǦʿFR, and ğa‘anfal “thick-lipped” (ǦḤFl) is appended to safarğal (SFRǦL): (9) Operation of ilḥāq on two words

As a result, the number of the basic morphological patterns that are acknowledged within a closed system is reduced considerably as several patterns are condensed into one and rare or deviant examples are appended to the nearest frequent or “normal” pattern available. For example, hammariš “ill-tempered old woman,” the only quadriliteral example of the pattern faʿʿalil, is appended to quinqueliterals of the type qahbalis “huge woman,”

Arabic Linguistic Tradition I ğaḥmariš “old woman,” and Ṣahṣaliq “vehement voice” (cf. Sībawayhi, Kitāb IV, 302, 330). It should be noted that appended consonants, such as n in ğaḥanfal and m in hammariš, are not individually morphemic in contrast to consonantal affixes (e.g., m and t in mustaqbal). Beyond the spheres of ištiqāq and ilḥāq, Ibn Ǧinnī introduced a general principle for justifying morphologically deviant examples. He argued that such examples serve as an indication (manbaha) of the original from which preceded a change that took place in the pattern it represents (Haṣāʾiṣ I, 159–163, 256–264; Munṣif I, 190–191). For example, the attested form istarwaḥa “breathe” is anomalous because, contrary to qiyās, it was not changed to istarāḥa. Its occurrence, however, according to Ibn Ǧinnī serves a specific purpose, namely, to indicate that the origin of verbs like istaqāma “stand erect” and istaʿāna “seek help” are *istaqwama and *istaʿwana. Through this interpretation, Ibn Ǧinnī not only exploited the occurrence of an anomaly to reinforce, rather than to undermine, the norm for the pattern involved but also highlighted the wisdom (ḥikma; cf. Munṣif I, 277) that characterizes Arabic in retaining anomalous forms. The notion of ḥikma is also strongly implied in the newly discovered genre of writing that deals exclusively with morphological taʿlīl. The publication of Muʾaddib’s (d. after 338/949) Daqāʾiq altaṣrīf brought to light the existence in the ALT of attempts that exhaustively provide justification for morphological phenomena (Baalbaki 2006). Muʾaddib’s reference (Daqāʾiq 25, 343) to a group of scholars of morphology, whom he calls aṣḥāb al-taṣrīf (morphologists) or mutaʿāṭū hāḏihi l-ṣināʿa (practitioners of this craft), strongly suggests that by the early 4th/10th century ṣarf was increasingly gaining independence from naḥw, although it shared the key strategies of grammatical analysis. aḍdād words with two contrary meanings adilla (pl. of dalīl) proofs aḥwāl mutanaqqila changes that words undergo due to the effect the operants have on their case endings ʿamal government amṯāl (pl. of maṯal) proverbs aṣḥāb al-taṣrīf morphologists aṣl (sg. of uṣūl) origin, original form, pattern; the main or fundamental issue of a certain bāb or topic ʿaṭf ʿalā ʿāmilayni elision of two operants after a conjunction ʿawāmil operants, governors ʿawāmil al-afʿāl operants that govern verbs ʿawāmil al-asmāʾ operants that govern nouns badal substitute (appositive) balāġiyyūn rhetoricians fʿl three root consonants representing basic morphological template fiʿl verb fiqh jurisprudence furūʿ subsidiary or less essential questions or problems, marked structures futūḥāt

Arabic Linguistic Tradition I conquests ğamʿ al-luġa collection of linguistic data ġarīb strange usage ḥarf particle (part of speech), letter, sound ḥikmah wisdom ḥurūf al-ziyāda augmented consonants ẖabīṯ repugnant usage ibdāl substitution of one consonant for another idġām/iddiġām assimilation, gemination iḍmār deletion, suppression iğmāʿ consensus, esp. consensus of experts ilḥāq appending, analytic morphophonological process deriving expanded form on basis of another existing pattern (wazn) ‘illa (sg. of ʿilal) cause, quality, explanation, detail; in phonology, conditioned change of long vowels ʿillat al-ʿilla meta-ʿilla ʿilal ğadaliyya naẓariyya argumentational-theoretical explanation ʿilal qiyāsiyya explanation by analogy ʿilal taʿlīmiyya explanation for pedagogical purposes ʿiʿrāb declension, case, and mode endings/categories ism noun istifhām question istiṣḥāb al-ḥāl presumption of continuity ištiqāq derivation kalām prose; in later usage, sentence kalām al-ʿArab speech of the Arabs kalima word, lexical unit, morpheme kaṯīr frequent usage kaṯra frequency (of use)

Arabic Linguistic Tradition I lafẓ form lafẓī formal laḥn “corruption” of speech laḥn al-ʿāmma common errors, errors of common people lā yağūz not permissible luġa radīʿa bad usage luġawiyyūn philologists, lexicographers maḏhab (sg. of maḏāhib) school of thought/grammar mādī past tense maʿnā meaning manbaha indication masʾala (sg. of masā’il) issue mīzān ṣarfī morphological measure or form, including both vowels and consonants muẖāṭab addressee, listener mustaqīm right, correct mutafaqqihūn jurists mutakallim speaker mutakallimūn theologians naḥārīr (pl. of niḥrīr) skillful, learned naḥw grammar in general, more specifically syntax naḥwiyyūn grammarians nāʾib ʿan al-fāʿil agent of a passive verb naql transmitted data, irregular but sanctioned data; in phonology, metathesis between glides and vowels naṣb accusative (in nouns), subjunctive (in verbs) naʿt attribute, adjective nawādir rare usages naẓm theory of word order and semantic and syntactic interrelationships

Arabic Linguistic Tradition I niyya intention qabīḥ ugly usage qalb metathesis; in phonology qalīl infrequent usage qirāʾāt Quranic readings qiyās analogy qurrāʾ Quranic readers saʿat al-kalām latitude of speech samāʿ (lit.) hearing, body of attested material heard directly from native speakers of Arabic ṣarf morphology, in particular lexically fixed form of words šarṭ condition, conditional sentence šawāhid attested and oft en ascribed material, esp. from poetry and the Quran šiʿr poetry taʿağğub admiration, astonishment tafsīr exegesis taẖfīf lightening (usually by phonological reduction) taʿlīl causation tanwīn nunation taqdīr suppletive insertion tarẖīm euphonic elision taṣarruf al-ʿāmil plasticity of the operant ʿulūm data, (lit.) sciences wağh correct, better usage waqf pause, pausal position ziyāda augmentation to basic morphological form

References

Arabic Linguistic Tradition I Primary Sources Abū l-Ṭayyib, Marātib = Abū l-Ṭayyib ʿAbdalraḥmān b. ʿAlī al-Luġawī, Marātib al-naḥwiyyīn, ed. Muḥammad Abū lFaḍl IbrāhĪm. 2nd ed. Cairo: Dār Nahḍat Miṣr, 1974. Astarābāḏī, Šarḥ al-Šāfiya = Raḍī al-Dīn Muḥammad b. Ḥasan al-Astarābāḏī, Šarḥ Šāfiyat Ibn al-Ḥāğib, ed. Muḥammad Nūr al-Ḥasan, Muḥammad al-Zafzāf and Muḥammad Muḥyīl-Dīn ʿAbdalḥamīd. 4 vols. Reprint from the Cairo edition. Beirut: Dār al-Kutub al-ʿIlmiyya, 1982. Ğurğānī, Dalāʿ il = Abū Bakr ʿAbdalqāhir b. ʿAbdalraḥmān al- Ğurğānī, Dalāʾ il al-iʾ ğāz, ed. Muḥammad Rašīd Riḍā. Reprint from the Cairo edition, Beirut: Dār al-Maʿrifa, 1981. Ibn al-Anbārī, Iġrāb = Abū l-Barakāt ʿAbdalraḥmān b. Muḥammad al-Anbārī, al-Iġrāb fī ğadal al-i ʿrāb, ed. Sa ʿīd alAfġānī. 2nd ed. Beirut: Dār al-Fikr, 1971. Ibn al-Anbārī, Inṣāf = Abū l-Barakāt ʿAbdalraḥmān b. Muḥammad al-Anbārī, al-Inṣāf fī masāʾ il al-ḫilāf bayna lnaḥwiyyīn al-Baṣriyyīn wa-l-Kūfiyyīn, ed. Muḥammad Muḥyīl-Dīn ʿAbdalḥamīd. 2 vols. Cairo: al-Maktaba lTiğāriyya, 1955. Ibn al-Anbārī, Lumaʿ = Abū l-Barakāt ʿ Abdalraḥmān b. Muḥammad al-Anbārī, Lumaʿ al-adilla fī uṣūl al-naḥw, ed. Sa ʿīd al-Afġānī. 2nd ed. Beirut: Dār al-Fikr, 1971. Ibn al-Anbārī, Nuzha = Abūl-Barakāt ʿAbdalraḥmān b. Muḥammad al-Anbārī, Nuzhat al-alibbāʿ fī ṭabaqāt al-udabāʿ, ed. Ibrāhīm al-Sāmarrāʿī. Baghdad: Maktabat al-Andalus, 1970. Ibn ʿAqīl, Šarḥ = Bahāʾ al-Dīn ʿAbdallāh b. ʿAbdalraḥmān Ibn ʿAqīl, Šarḥ Ibn ʿAqīl ʿalā Alfiyyat Ibn Mālik, ed. Ramzī Munīr Baʿalbakī. Beirut: Dār al- ʿIlm li-l-Malāyīn, 1992. Ibn Ġinnī, Ḫaṣāʿ iṣ = Abū l-Fatḥ ʿUṯmān Ibn Ğinnī, al-Ḫaṣāʾ iṣ, ed. Muḥammad ʿAlīal-Nağğār. 3 vols. Cairo: Dār alKutub al-Miṣriyya, 1952–1956. Ibn al-Sarrāğ, Uṣūl = Abū Bakr Muḥammad b. Sahl Ibn al-Sarrāğ, al-Uṣūl fī l-naḥw, ed. ʿAbdalḥusayn al-Fatlī. 3 vols. Beirut: Muʾassasat al-Risāla, 1985. Ibn ʿUṣfūr, Mumtiʿ = Abūl-Ḥasan ʿAlī b. Muʾ min Ibn ʿUṣfūr, al-Mumtiʿ fī l-taṣrīf, ed. Faḫr al-Dīn Qabāwa. 4th ed. 2 vols. Beirut: Dār al-Āfāq al- Ğadīda, 1979. Mubarrad, Kāmil = Abūl-ʿAbbās Muḥammad b. Yazīd al-Mubarrad, al-Kāmil, ed. Muāammad Abūl-Faḍl Ibrāhām, and al-Sayyid Šiḥāta. 4 vols. Cairo: Maktabat Nahḍat Miṣr, 1956. Mubarrad, Muqtaḍab = Abū l-ʿAbbās Muḥammad b. Yazīd al-Mubarrad, al-Muqtaḍab, ed. Muḥammad ʿAbdalḫāliq ʿUḍayma. 4 vols. Cairo: Dār al-Taḥrīr, 1965–1968. Sībawayhi, Kitāb = Abū Bišr ʿAmr b. ʿUṯmān Sībawayhi, al-Kitāb, ed. ʿAbdalsalām Muḥammad Hārūn. 5 vols. Cairo: al-Hayʾ al-Miṣriyya l- ʿĀmma, 1977. Sīrāfī, Aḫbār = Abū Sa ʿīd al-Ḥasan b. ʿAbdallāh al-Sīrāfī, Aḫbār al-naḥwiyyīn al-Baṣriyyīn, ed. Fritz Krenkow. Beirut: Imprimerie Catholique, 1936. Suyūṭī, Iqtirāḥ = Ğalāl al-Dīn Abū l-Faḍl ʿAbdalraḥmān b. Abī Bakr al-Suyūṭī, al-Iqtirāḥ fī ʿilm uṣūl al-naḥw, ed. Aḥmad Muḥammad Qāasim. Cairo: Maṭba ʿat al-Sa ʿāda, 1976. Ṣaymarī, Tabṣira = Abū Muḥammad ʿAbdallāh b. ʿAlī al-Ṣaymarī, al-Tabṣira wa-l-taḏkira, ed. Fatḥī Aḥmad Muṣṭafā ʿAlī al-Dīn. 2 vols. Makka: Umm al-Qurā University, 1982. Tanūḫī, Tārīḫ = Abūl-Maḥāsin al-Mufaḍḍal b. Muḥammad b. Misʿar al-Tanūḫī, Tārīḫ al- ʿulamāʾ al-naḥwiyyīn, ed. ʿAbdalfattāḥ Muḥammad al-Ḥulw. Riyad: Maṭābi˓ Dār al-Hilāl, 1981. Yāqūt, Muʿğam = Šihāb al-Dīn Abū ˓Abdallāh Yāqūt b. ˓Abdallāh al-Rūmī al-Ḥamawī, Muʿğam al-udabā˒, ed. Iḥsān ʿAbbās. 7 vols. Beirut: Dār al-Ġarb al-Islāmī, 1993.

Arabic Linguistic Tradition I Zağğāğī, Īḍāḥ = Abū l-Qāsim ʿAbdalraḥmān b. Isḥāq al-Zağğāğī, al-Īḍāḥ fī ʿilal al-naḥw, ed. Māzin al-Mubārak. Cairo: Dār al- ʿUrūba, 1959. Zubaydī, Ṭabaqāt = Abū Bakr Muḥammad b. al-Ḥasan al-Zubaydī al-Andalusī, Ṭabaqāt al-naḥwiyyīn wa-lluġawiyyīn, ed. Muḥammad Abū l-Faḍl Ibrāhīm. 2nd ed. Cairo: Dār al-Ma ʿārif, 1973.

Secondary Sources Baalbaki, Ramzi. 1981. Arab grammatical controversies and the extant sources of the second and third centuries A.H. In Studia Arabica et Islamica: Festschrift for Iḥsān ʿAbbās on His Sixtieth Birthday, ed. Wadād al-Qāḍīi, 1–26. Beirut: American University of Beirut. ——. 1983. The relation between naḥw and balāġa: A comparative study of the methods of Sībawayhi and Ğurğānī. ZAL 11: 7–23. ——. 1995. The book in the grammatical tradition: Development in content and methods. In The book in the Islamic world, ed. George N. Atiyeh, 123–139. New York: State University of New York Press. ——. 2001. Bāb al-fāʾ [fāʾ + subjunctive] in Arabic grammatical sources. Arabica 48: 186–209. ——. 2001–2002. Ilḥāq as a morphological tool in Arabic grammar. Journal of Arabic and Islamic Studies 4: 1–25. ——. 2005. Review of R. Talmon’s Eighth-century Iraqi grammar: A critical exploration of pre-Ḫalīlian Arabic linguistics. Journal of Semitic Studies 50: 413–416. ——. 2006. Unfamiliar morphological terminology from the fourth century A.H.: Muʾaddib’s Daqā’iq al-taṣrīf. In Grammar as a window onto Arabic humanism: A collection of articles in honour of Michael G. Carter, ed. Lutz Edzard and Janet Watson, 21–50. Wiesbaden: Harrassowitz. ——. 2008. The legacy of the Kitāb: Sībawayhi’s analytical methods within the context of the Arabic grammatical theory. Leiden: Brill. Bernards, Monique. 1997. Changing traditions: al-Mubarrad’s refutation of Sībawayh and the subsequent reception of the Kitāb. Leiden: Brill. Bohas, Georges and Jean-Patrick Guillaume. 1984. Étude des théories des grammairiens arabes. I. Morphologie et phonologie. Damascus: Institut Français de Damas. Bohas, Georges, Jean-Patrick Guillaume, and Djamel Eddin Kouloughli. 1990. The Arabic linguistic tradition. London: Routledge. Carter, Michael G. 1972. Les origines de la grammaire arabe. Revue des études islamiques 40: 69–97. ——. 1973. An Arab grammarian of the eighth century A.D. Journal of the American Oriental Society 93: 146–157. ——. 1991. The ethical basis of Arabic grammar. al-Karmil 12: 9–23. ——. 1994. Writing the history of Arabic grammar. Historiographia Linguistica 21 (3): 385–414. ——. 2000. The development of Arabic linguistics after Sībawayhi: Baṣra, Kūfa and Baghdad. In History of the Language Sciences, vol. 1, ed. Sylvain Auroux et al., 263–272. Berlin: Walter de Gruyter. Guillaume, Jean-Patrick. 2000. La nouvelle approche de la grammaire au IVe -Xe siècle: Ibn Ǧinnī (320/932– 392/1002). History of the Language Sciences, ed. Sylvain Auroux et al., I, 273– 280. Berlin and New York: Walter de Gruyter. Humbert, Geneviève. 1995. Les voies de la transmission du Kitāb de Sībawayhi. Leiden: E.J. Brill. Itkonen, Esa. 1991. Universal history of linguistics: India, China, Arabia, Europe. Amsterdam: J. Benjamins. Merx, Adalbertus. [1889] 1966. Historia artis grammaticae apud Syros. Reprint, Nendeln, Liechtenstein: Kraus.

Arabic Linguistic Tradition I Owens, Jonathan. 1988. The foundations of grammar: An introduction to medieval Arabic grammatical theory. Amsterdam: J. Benjamins. ——. 2000a. On club membership: A reply to Kouloughli. Histoire Épistémologie Langage 22(2): 105–126. ——. 2000b. The structure of Arabic grammatical theory. In History of the language sciences, vol. 1, ed. Sylvain Auroux et al., 286–300. Berlin: Walter de Gruyter. Rundgren, Frithiof. 1976. Über den griechischen Einfluss auf die arabische Nationalgrammatik. Acta Universitatis Upsaliensis 2: 119–144. Suleiman, Yasir. 1999. The Arabic grammatical tradition: A study in ta ʿlīl. Edinburgh: Edinburgh University Press. Talmon, Rafael. 1982. Naḥwiyyūn in Sībawayhi’s Kitāb. ZAL 8: 12–38. ——. 2003. Eighth-century Iraqi grammar: A critical exploration of pre-Ḫalīlian Arabic linguistics. Winona Lake, IN: Eisenbrauns. Troupeau, Gérard. 1981. La logique dʾIbn al-Muqaffaʿ et les origines de la grammaire arabe. Arabica 28: 242–250. Versteegh, Kees. 1993. Arabic grammar and Qurʾ ānic exegesis in early Islam. Leiden: E.J. Brill. ——. 1995. The explanation of linguistic causes: az-Zağğāğī’s theory of grammar. Amsterdam: J. Benjamins. ——. 1997. Landmarks in linguistic thought III: The Arabic linguistic tradition. London: Routledge. Ramzi Baalbaki Ramzi Baalbaki, American University of Beirut

The Syntax of Arabic From A Generative Perspective

The Syntax of Arabic From A Generative Perspective * Elabbas Benmamoun and Lina Choueiri The Oxford Handbook of Arabic Linguistics Edited by Jonathan Owens

Abstract and Keywords Research on Arabic varieties within modern syntactic approaches has tracked the debates that have preoccupied the field of generative linguistics in its different incarnations throughout the last six decades. The debates centered on the nature of linguistic categories, syntactic configurations and their constituents, syntactic alternations and processes that alter the order of constituents, and dependencies between members of the syntactic representations. This article considers the main issues within Arabic syntax and the influential approaches that have been advanced. It focuses on debates surrounding phrase structure and word order, the syntax of the noun phrase, subjects and subject agreement, negation, long A’-dependencies, and wh-in situ constructions. Keywords: syntax, phrase structure, word order, noun phrase, subjects, negation, long A’-dependencies, wh-in situ constructions

6.1 Introduction RESEARCH on Arabic varieties within modern syntactic approaches has tracked the debates that have preoccupied

the field of generative linguistics in its different incarnations throughout the last six decades. The debates centered on the nature of linguistic categories, syntactic configurations and their constituents, syntactic alternations and processes that alter the order of constituents, and dependencies between members of the syntactic representations. The discussion focused on problems of word order and sentence structure, phrasal structure, agreement, tense, negation, questions, relative constructions, and clitic-left dislocation. Related to those specific issues are the questions about the nature of the linguistic system, its components and how they interface with each other, and its universal and language specific characteristics. Arabic has figured in the debates within generative linguistics in varying degrees, though the focus started with the syntax of Standard Arabic (SA) and has since been expanded to include the spoken varieties and, more recently, the processing and the acquisition of Arabic. While we cannot possibly do justice to all the research on Arabic syntax within the generative paradigm, we hope, in the pages to follow, to familiarize the reader with the main issues within Arabic syntax and the influential approaches that have been advanced.

6.2 Phrase Structure and Word Order Since the phrase marker plays a central role in deriving syntactic generalizations within generative syntactic paradigms, research on Arabic has focused extensively on this area. The differences that Arabic displays in this regard compared with, for example, Germanic and Romance languages and the different word orders available in the various Arabic varieties present challenges for theories of phrase structure and their quest for isolating universal aspects of language. In that regard, one important challenge that Arabic varieties present for syntactic approaches has dominated the

The Syntax of Arabic From A Generative Perspective debate. This revolves around so-called verbless sentences such as (1) from SA, where nouns and predicates are marked for case, and Moroccan Arabic (MA), where nouns and predicates are not.1 (1) a.

ʔal-bayt-u

kabiir-un

DEF-house-NOM

big-NOM

“The house is big.” b.

ʔ ar-ražul-u

kaatib-un

DEF-man-NOM

writer-NOM

“The man is a writer.” c.

ʔar-ražul-u

fii

l-bayt-i

DEF-man-NOM

DEF-house-GEN

“The man is in the house.” (2) a.

ḍ-ḍar

kbir-a

DEF-house

big-F

“The house is big.” b.

r-ražəl

nəžžar

DEF-man

carpenter

“The man is a carpenter.” c.

r-ražəl

f-ḍ-ḍar

DEF-man

in-DEF-house

“The man is in the house.”

In this respect, Arabic is quite different from English and French in having main sentences ostensibly without a verb. The equivalent constructions in English contain a verb that seems to form a constituent with the predicate. In other words, there is a verbal constituent in (3) that seems to be missing in (1) and (2). (3) a. The house is big. b. The man is a teacher. c. The main is in the house. This is a classic case of language variation related to clause structure. Two broad approaches were contemplated

The Syntax of Arabic From A Generative Perspective by students of Arabic within the generative paradigm (Bakir 1980; Ayoub 1981; Jelinek 1981; Eid 1983, 1991; Moutaouakil 1987; Heggie 1988; Fassi Fehri 1993; Bahloul 1994; Benmamoun 2000; Aoun et al. 2010). 2 The first approach claims that there is indeed a verb in the structure but that it is phonologically null, that is, an abstract verb that is syntactically present but does not have phonological content. This approach tapped into the idea that has been well accepted within the generative paradigm that some categories may be phonologically null, either because they lack phonological content or because they have undergone a process of deletion. According to this approach, the main feature that distinguishes Arabic from, for example, English and French is that the copula verb can be null in main finite clauses such as (1) and (2). One reason that this may be plausible is that, in past tense and future tense clauses, the copula is overt, that is, has phonological content as illustrated in (4) and (5) from SA and (6) and (7) from MA.3 (4) a.

kaana

ʔal-bayt-u

kabiir-an

be.PST

DEF-house-NOM

big-ACC

“The house was big.” b.

kaana

ʔ ar-ražul-u

kaatib-an

be.PST

DEF-man-NOM

writer-ACC

“The man was a writer.” c.

kaana

See Also

저가항공사의 함정?! - 영국 어학연수 후기

ħkii-na

maʕ

[wala mʕallme]i

ʔablma

student-her

DEF-bad NEG

talked-1.PL

with

no teacher

before

huwwe

yuusal

arrive.3

“Her bad student, we didn’t talk to any teacher before he arrived.”

To see how reconstruction for bound variable anaphora works, we provide schematic representations of the sentences in (82–83):

The Syntax of Arabic From A Generative Perspective (84) a. Reconstruction Available … [Antecedent …. proni. … … ]j. …. … ([CP). …. … … QPi. … ([CP) …. RPj … (]) (]) … b. Reconstruction unavailable * … [Antecedent …. proni. … … ]j … … ([Island ) … … … QPi… ([Island ) …. RPj … (])(]) … Resumption, it is concluded, is not a unitary phenomenon. It can be derived from two different sources: the first one, true resumption (see Aoun, Choueiri, and Hornstein 2001), has the antecedent base-generated in its surface position and related to a pronominal inside the sentence via some interpretive mechanism available in the grammar (85a). In the other derivation, termed apparent resumption (see ibid.), the antecedent is generated together with the resumptive element, and it undergoes movement to its surface position leaving a copy or trace in the base (85b): (85) a. True resumption Antecedenti … … … … RPi b. Apparent resumption Antecedenti … … … … Antecedenti-RP The bound variable reading obtains only when a sentence can be derived from the source in (85b). In that case, the lower copy of the antecedent contains a copy of the bound pronoun, which will be c-commanded by the quantifier phrase within the sentence—hence the bound variable reading. The representation in (85b) depends on the availability of movement: since movement from within islands is prohibited, the sentences in (82b)–(83b) cannot have such a representation. It is not surprising then, that the bound variable reading resulting from reconstruction is not available in those sentences. Recent work by Guillot and Malkawi (2006) and Malkawi and Guilliot (2007) challenges the tight connection that Aoun, Choueiri, and Hornstein (2001) assume to exist between reconstruction and movement. For the latter, reconstruction is the result of interpreting the lower copy of a moved element. Malkawi and Guilliot (2007) use data from JA to show that reconstruction needs to be dissociated from movement. (86a) shows that reconstruction can take place inside strong islands in JA, and (86b) illustrates the fact that, despite the absence of islands, reconstruction is still not available: (86) a. ṭaalib-[ha]i

l-kassul

l-mudiira

ziʕl-at

laʔannu [kul mʕalmih]i

student-her

DEF-bad

DEF-principal

upset-3.F

because every teacher

šaaf-at-uh

(hu)

Ɣašš

li- mtiħaan

saw-3.F-him

cheated.3

DEF-exam

“Her bad student, the principal got upset because every teacher saw him cheating in the exam.” b. ʕalamit

kariim

b-it-fakkir

ʔinnu

lazim

i-Ɣayyar-ha

grade

Karim

INDC-2 -think

that

must

3-change-it.F

“Karim’s grade, you think that he must change (it).”

Assuming that bound variable anaphora requires a c-command relation between the quantifier phrase and the

The Syntax of Arabic From A Generative Perspective relevant pronoun, cases like (86a) clearly indicate that, even in the absence of movement, reconstruction must be available. Coreference between kariim and the embedded subject is possible in (86b). If reconstruction were forced, we would expect coreference to fail as a result of a principle C violation.46 Since this is not the case, we can therefore conclude that reconstruction is not necessary even when movement is available. Malkawi and Guilliot (2007) further argue that the contrast between the presence of reconstruction and its absence cannot be explained away by stipulating that reconstruction is merely optional. Such an assumption would not account for the contrast they observe between weak pronouns (86b) and strong pronouns (87) with respect to principle C effects: (87)* ʔaχu

laylai

brother

Layla

proi

gaal-at

ʔinnu

saafar

said-3.F

that

left.3

“The brother of Layla, she said that he/the idiot left.”

In (87), coreference between Layla and the subject of the main clause is not possible, indicating that reconstruction of the clitic-left dislocated element has taken place and a principle C violation ensued. The pattern of reconstruction in resumptive contexts is even more complex in JA, but this suffices to show how it challenges a tight link between movement and reconstruction. Guilliot and Malkawi (2006) and Malkawi and Guilliot (2007) argue that reconstructed readings follow, not necessarily from movement, but from the operation COPY, which is independently available in the grammar of all languages. The interaction between COPY and the syntax of pronouns in JA accounts for the data in (86)–(87). The approach in Guilliot and Malkawi (2006) and Malkawi and Guilliot (2007) to the syntax of pronouns in Arabic differs from that set out in Ouhalla (2001). They assume that weak pronouns are definite determiners that can appear in two different structures: one where the definite determiner is followed by a NP that later undergoes deletion under identity with an antecedent (88a); and the other where the definite determiner is not branching (88b) (see Elbourne 2001): (87)* Weak pronouns ]

a. [the/it

b. [the/iti]

It is the availability of both structures that accounts for the optional availability of reconstructed readings with weak pronouns (88). In (86a), the weak resumptive pronoun has the structure in (88a), with the NP being a copy of the clitic-left dislocated noun phrase that gets deleted. Thus, a copy of the pronoun contained within the clitic-left dislocated noun phrase is present in the resumption site for interpretation. This accounts for the availability of the bound variable reading in (86a). In (86b), the availability of the structure in (88b) accounts for the absence of principle C effects. Malkawi and Guilliot (2007) follow Aoun, Choueiri, and Hornstein (2001) in assuming that strong pronouns are full DPs that can be adjoined to another DP resulting in constructions like (89): (89) saami

huwwe

nəse

l-mawʕad

Sami

forgot.3

DEF-appointment

“Sami, he, forgot the appointment.”

The Syntax of Arabic From A Generative Perspective In (87) then, the clitic-left dislocated phrase is initially generated in an adjunction structure with the strong pronoun and then moved to its sentence initial position. Therefore, a copy of the name layla is present in the resumption site, leading to the observed principle C effect. The discussion of resumption in long A’-dependencies highlights the complexity of the syntactic microvariation that we observe between the various spoken dialects of Arabic. Needless to say, further research is required to understand what underlies this variation. Nevertheless, the investigation of resumption and its properties within the generative paradigm has led to an interesting discussion of the diagnostics of A’-movement. It has also raised questions regarding the nature of resumptive elements.

6.7 Wh-In Situ Similar concerns have led students of Arabic syntax to examine wh-in situ constructions in the dialects where they are available. We have already observed that, unlike other Arabic varieties, EgA does not make use of the gap strategy in wh-questions (see Table 6.1). In fact, direct questions in EgA are generally formed by keeping the whword in situ. Soltan (2010) provides the sentences in (90) and (91) to illustrate the in situ strategy in EgA: (90) a. ʔinta

šuf-t

miin

ʔimbaarih

you.M.SG

saw-2 .M

who

yesterday

“Who did you see yesterday?” b. ʔaħmad

ha-yi-saafir

fein/ʔimtaa/ʔizzaay/leih

Ahmad

FT-3-travel

where/when/how/why

“Where/When/How/Why will Ahmad travel?”

(91) a. (huwwa)

ʔaħmad

ʔal-l-ak

ʔin

mona

ʔištar-it

ʔeih

(he)

Ahmad

said.3 -to-you.M

that

Mona

bought-3 .F

what

“What did Ahmad tell you that Mona bought?” b. (huwwa)

ʔaħmad

ʔal-l-ak

?in

mona

safir-it

fein

(he)

Ahmad

said.3 -to-you.M

that

Mona

traveled-3 .F

where

“Where did Ahmad tell you that Mona traveled?”

As seen in (90)–(91), argument as well as adjunct wh-pronouns can remain in situ in EgA. Interestingly, the wh-in situ can have matrix scope whether the wh-element occurs in the matrix clause (90) or the embedded clause (91).

The Syntax of Arabic From A Generative Perspective Following initial observations in Wahba (1984), Soltan (2010) shows that wh-in situ interrogatives are not sensitive to island constraints in EA. The examples in (92) illustrate this generalization: (92) a. ʔinta

ʔabil-t

ʔil-bint

illi

ʔitgawwiz-it

miin

you

met-2

DEF-girl

that

married-3.F

who

“You met the girl who married who?” (=“Who did you meet the girl that married?”) b. huda

miš y-it

ʔabl-ma

ʔaħmad

yi-ʔaabil

miin

Huda

left -3.F

before

Ahmad

3-meet

who

“Huda left before Ahmad met who” (=“Who did Huda leave before Ahmad met?”)

Wahba (1984) and Soltan (2010) propose different accounts of wh-in situ interrogatives in EgA. Focusing on the interpretation of sentences like (90)–(91) as matrix questions, Wahba (1984) argues that in situ elements move covertly to the matrix Comp to mark the scope of the question. The acceptability of the sentences in (92) is attributed to the assumption that islands do not constrain LF movement (see, e.g., Huang 1982). Soltan (2010) criticizes this stipulation and claims that wh-in situ questions in EgA are generated without any movement and that they are interpreted via the mechanism of unselective binding (see Pesetsky 1987). Thus, a question like (90a) would have the representation in (93), whereOp, a base-generated operator in the matrix clause, marks the scope of the wh-question: (93) [C P Opi [TP ʔinta šuft miini ʔimbaariħ]] Examining wh-in situ interrogatives across the various Arabic varieties complicates the picture substantially. SA and MA do not allow wh-in situ questions (Nouhi 1996; Al-Ghalayini 2003). LA and Iraqi Arabic (IA), on the other hand, are languages that allow optional wh-in situ, but unlike EgA they impose restrictions on those constructions. In fact, wh-in situ questions in those varieties are more constrained than questions involving wh-fronting, as illustrated in (94) from IA (see also Wahba 1991; Ouhalla 1996b): 47 (94) a.* ʕurf-ut

mona

il-bint

[illi

ištar-at

šeno]

knew-3.F

mona

DEF-girl

that

bought-3.F

what

“What(x) Mona knew the girl who bought (x)?” b.* tawwar-at

mona

[Ali

ištara

šeno]

thought-3.F

Mona

Ali

bought.3

what

“What did Mona think Ali bought?”

The Syntax of Arabic From A Generative Perspective c.* šeno

tǣawwar-at

?mona

[Ali

ištara e]

what

thought-3.F

Mona

Ali

bought.3

“What did Mona think Ali bought?”

In IA, unlike EgA, wh-phrases cannot appear within islands (94a). In fact, wh-in situ elements are also prohibited in an embedded tensed clause, when the question has matrix scope (94b). (94a) contrasts with (95), which involves wh-fronting from within a relative clause island but induces only a mild subjacency violation. This contrast in acceptability indicates that wh-fronting and wh-in situ in IA involve different derivations. (95) ??šeno

ʕurf-ut

mona

il-bint

[illi

ištar-at e]

what

knew-3.F

Mona

DEF-girl

that

bought-3.F

“What(x) Mona knew the girl who bought (x)?”

Ouhalla (1996b) proposes an analysis of IA wh-in situ questions that advances the two claims in (96): (96) a. There are two types of wh-elements: the bare type and the compound type. The compound type shows overt phi-features. b. Bare wh-elements are long distance A’-anaphors whereas compound wh-elements are local A’anaphors. (96a) underlies the contrast one observes between IA wh-words and EgA wh-words. Thus, in IA, the wh-words meno “who” and šeno “what” can be divided into two parts: one carries the wh-morpheme, i.e. men- and šen-, and the second carries phi-features (i.e.,– o). The pronominal element found in wh-words in IA, functions also as an accusative pronoun in the language (97): (97) šuf-t-o saw-1.SG-him “I saw him.” In EgA, wh-words are bare in the sense of Ouhalla (1996b) since they do not carry phi-features. Being of the compound type, IA wh-words are local A’-anaphors (96b). This means that they must have an antecedent in the minimal tensed clause in which they occur. This requirement readily accounts for the unacceptability of (94b). In (94b), the domain in which the wh-word needs to have an antecedent is the embedded clause. However, the embedded Comp is not marked [+wh] in that sentence. Hence, the wh-word fails to have an antecedent in the minimal tensed clause in which it occurs and the resulting sentence is unacceptable. (94a) has a somewhat different explanation: the wh-word in situ there is related to a local antecedent in the minimal tensed clause in which it occurs. This is the embedded Comp, which is marked [+wh], in the case of relative clauses. Thus, the whword in (94a) cannot be related to the matrix Comp, and the sentence cannot have a matrix scope reading. Ouhalla’s (1996b) analysis of IA wh-in situ can be extended to EgA, where wh-pronouns are of the bare type. They are thus expected to behave as long distance A’-anaphors. This means that they need not have a local antecedent in the minimal finite clause in which they occur, and they need not be bound to the nearest potential antecedent. This accounts for the fact that, in EgA, wh-in situ can occur in embedded tensed clauses and inside islands as well. Such an analysis has the advantage of reducing this case of syntactic (micro)variation to the lexical properties of the wh-words themselves.

The Syntax of Arabic From A Generative Perspective A systematic exploration of wh-in situ questions across Arabic dialects is still lacking. Some important theoretical issues remain open: what factors are responsible for the presence or absence of wh-in situ questions in some Arabic varieties? Ouhalla’s analysis addresses the properties of wh-elements in situ independently of movement. However, the question remains regarding the relation between wh-in situ and wh-fronting.

6.8 Other Related Approaches to the Syntax of Arabic Generative approaches to the syntax of Arabic have provided important generalizations to probe various aspects of the theoretical apparatus of the generative paradigm. The study of SA and various other Arabic dialects has made important contributions to the ongoing research on the computational component of the grammar, its properties, and how it interacts with other aspects of the grammar. The dialects that have particularly been the focus of attention represent four main linguistic groupings (Maghrebi, Levantine, Egyptian, and Gulf). However, many varieties within these regions remain underresearched, and other varieties of Arabic spoken by minorities in sub-Saharan Africa and Asia have yet to receive the same degree of attention within the generative paradigm. In addition to the focus on synchronic syntactic phenomena, the field has also engaged issues related to the acquisition of language, language deficits and disorder, language change, and language contact, among many others. Recently, there has been increasing interest in the psycholinguistic and neurolinguistic aspects of Arabic syntax and morphology. In addition, research in syntax proper now increasingly relies on the use of a diverse set of research tools and data, including production and comprehension experiments, searchable electronic corpora, and of course elicited data and narratives from native speakers. The goal is to compare the acquisition of some syntactic and morphosyntactic patterns (e.g., the acquisition of inflectional morphology, questions, negation) in Arabic with the acquisition of similar patterns in other languages. This kind of research is informed by recent developments in syntax and informs both syntactic research as well as the study of the cognitive aspects of language. For example, Aljenaie (2001, 2010) investigates the acquisition of the morphosyntax of Kuwaiti Arabic. She collected extensive data from Kuwaiti children over a long period of time. Her findings have so far confirmed that Kuwaiti children use the imperfective as a default form, which is consistent with recent proposals in Arabic syntax and morphology that have demonstrated that the imperfective does not carry any temporal or aspectual information and is rather a “nonfinite” form of the verb. Kuwaiti children improperly use the imperative form, an inflected form, in a way that is again consistent with the status of the imperative as a reduced form of the imperfective that is devoid of person agreement. Working with Saudi children with language impairment, Abdullah (2002) uncovered similar results. Similarly, Lorimor (2007) undertook a sentence completion experiment to test how Lebanese speakers perform on agreement in the context of coordination, which is still generating a lively debate in syntax. Her findings were consistent with what descriptive studies of close conjunct agreement have shown, namely, that there is a tendency to have close conjunct agreement when the conjuncts follow the verb but not when they precede the verb. However, she also uncovered an interesting result that previous theoretical studies did not discuss. When Lebanese speakers are confronted with a situation where the two conjuncts must be between an auxiliary verb and another predicate, such as an adjective, there is a tendency to have close conjunct agreement with the auxiliary that precedes the two conjuncts and full agreement with the predicate that follows the two conjuncts. As discussed in Benmamoun et al. (2010), this finding has important implications for theoretical analyses of close conjunct agreement. The biclausal/gapping account advanced in Aoun et al. (1994) cannot be easily extended to these facts. Rather, the facts seem to argue for the effect of adjacency on agreement. This in turn has led to a different conception of the relation between syntax and morphology. This suitably illustrates how experimental work can inform and be informed by theoretical syntactic studies. The same can be said about corpus studies, field workbased studies, and studies that use the old fashioned but still critically relevant introspective data. It is encouraging that this type of research is being carried out on Arabic.

References Abdullah, F. 2002.Specific language impairment in Arabic-speaking children: deficits in morpho-syntax. PhD diss., McGill University,Canada.

The Syntax of Arabic From A Generative Perspective Al-Ghalayini, Mustafa. 2003. Źaamiʕu d-duruusi l- ʕarabiyyati. Beirut: al-Maktaba al-ʕaṣriyya. Aljenaie, Khawla. 2001.The emergence of tense and agreement in Kuwaiti Arabic child language. PhD diss., University of Reading, UK. Aljenaie, Khawla. 2010. Verbal inflection in the acquisition of Kuwaiti Arabic. Journal of Child Language 37: 841– 863. Aoun, Joseph and Elabbas Benmamoun. 1998. Minimality, reconstructions, and PF-movement. Linguistic Inquiry 29 (4):569–597. Aoun, Joseph, Elabbas Benmamoun, and Dominique Sportiche. 1994. Agreement and conjunction in some varieties of Arabic. Linguistic Inquiry 25: 195–220. ——. 1999. Further remarks on first conjunct agreement. Linguistic Inquiry 30: 669–681. Aoun, Joseph, Elabbas Benmamoun and Lina Choueri. 2010. The syntax of Arabic. Cambridge, UK: Cambridge University Press. Aoun, Joseph and Lina Choueiri. 1996. Resumption and last resort. Unpublished manuscript, University of Southern California,Los Angeles. ——. 1999. Modes of interrogation. In PAL XII ed. Elabbas Benmamoun, 7–26. Amsterdam: John Benjamins. ——. 2000. Epithets. Natural Language and Linguistic Theory 18: 1–39. Aoun, Joseph, Lina Choueiri, and Norbert Hornstein. 2001. Resumption, movement, and derivational economy. Linguistic Inquiry 32: 371–403. Aoun, Joseph and Yen-Hui Audrey Li. 1990. Minimal disjointness. Linguistics 28: 189–203. ——. 1993a.Syntax of scope. Cambridge, MA: MIT Press. Ayoub, Georgine. 1981. Structure de la phrase verbale en Arabe Standard. PhD diss.,Universitéde Paris VII, France. Bahloul, Maher. 1994. The syntax and semantics of taxis, aspect, tense and modality in Standard Arabic. PhD diss., Cornell University,Ithaca. Bahloul, Maher and Wayne Harbert. 1993. Agreement asymmetries in Arabic. In Proceedings of WCCFL 11, ed. Jonathan Mead, 15–31. Stanford, CA: CSLI. Baker, Mark C. 1988. Incorporation: A theory of grammatical function changing. Chicago: University of Chicago Press. Bakir, Murtadha. 1980. Aspects of clause structure in Arabic. PhD diss.,Indiana University, Bloomington. Benmamoun, Elabbas. 1992. Inflectional and functional morphology: Problems of projection, representation and derivation. PhD diss., University of Southern California,Los Angeles. ——. 1997. Licensing of negative polarity in Moroccan Arabic. Natural Language and Linguistic Theory 15: 263– 287. ——. 1999. The syntax of quantifiers and quantifier float. Linguistic Inquiry 30: 621–642. ——. 2000. The feature structure of functional categories: A comparative study of Arabic dialects. Oxford: Oxford University Press. ——. 2003. Agreement parallelism between sentences and noun phrases: A historical sketch. Lingua 113: 747– 764. ——. 2006. Licensing Ccnfigurations: The puzzle of head negative polarity items. Linguistic Inquiry 37 (1): 141–

The Syntax of Arabic From A Generative Perspective 149. Benmamoun, E., Abunasser, M., Al-Sabbagh, R., Bidaoui, A. & Shalash, D. Forthcoming. The Location of Sentential Negation in Arabic Varieties. Brill’s Annual of Afroasiatic Languages and Linguistics. Benmamoun, Elabbas and Heidi Lorimor. 2006. Featureless expressions: When morphophonological markers are absent. Linguistic Inquiry 37 (1): 1–23. Borer, Hagit. 1984. Restrictive relatives in Modern Hebrew. Natural Language and Linguistic Theory 2: 219–260. ——. 1995 The ups and downs of Hebrew verb movement. Natural Language and Linguistic Theory 13: 527–606. ——. 1996. The construct in review. In Studies in Afroasiatic grammar, ed. J. Lecarme, J. Lowenstamm, and U. Shlonsky, 30–61. The Hague: Holland Academic Graphics. ——. 2008. Compounds: The view from Hebrew. Unpublished manuscript, University of Southern California,Los Angeles. Brustad, Kristen. 2000. The syntax of spoken Arabic. Washington, DC: Georgetown University Press. Chomsky, Noam. 1977. On Wh-movement. In Formal syntax, ed. P. Culicover, T. Wasow, and A. Akmajian, 71–132. New York: Academic Press. ——. 1981. Lectures in government and binding. Drodrecht: Foris. ——. 1995. The minimalist program. Cambridge, MA: MIT Press. Choueiri, Lina. 2002. Issues in the syntax of resumption: Restrictive relatives in Lebanese Arabic. PhD diss., University of Southern California,Los Angeles. Cinque, G. 1994. On the evidence for partial N-movement in the Romance DP. In Paths towards universal grammar: Essays in honor of Richard Kayne, ed. G. Cinque, J. Koster, J.-Y. Pollock, and R. Zanuttini, 85–110. Washington, DC: Georgetown University Press. Déchaine, Rose-Marie. 1993. Predicates across categories. PhD diss., University of Massachusetts, Amherst. Demirdache, Hamida. 1991. Resumptive chains in restrictive relatives, appositives, and dislocation structures. PhD diss., Massachusetts Institute of Technology, Cambridge. ——. 1997. Dislocation, resumption, and weakest crossover. In Materials on left dislocation, ed. Elena Anagnostopoulou, Henk van Riemsdijk, and Frans Zwarts, 193–231. Amsterdam: John Benjamins. Doron, Edith. 1996. The predicate in Arabic. In Studies in Afroasiatic grammar, ed. J. Lecarme, J. Lowenstamm, and U. Shlonksy,77–87. Leiden: Holland Academic Graphics. Doron, Edit and Caroline Heycock. 1999. Filling and licensing multiple specifiers. In Specifiers, ed. David Adger, Susan Pintzuk, Bernadette Plunkett, and George Tsoulas,69–89. Oxford: Oxford University Press. Eid, Mushira. 1983. The copula function of pronouns. Lingua 59: 197–207. ——. 1991. Verbless sentences in Arabic and Hebrew. In PAL III, ed. Bernard Comrie and Mushira Eid,31–61. Amsterdam: John Benjamins. ——. 1993. Negation and predicate heads. In Principles and prediction: The analysis of natural language, ed. Mushira Eid and Gregory K. Iverson,135–152. Amsterdam: John Benjamins. Eisele, John. 1988. The syntax and semantics of tense, aspect, and time reference in Cairene Arabic. PhD diss., University of Chicago. Elbourne, Paul. 2001. E-type anaphora as NP deletion. Natural Language Semantics 9: 241–288. Fassi Fehri, Abdelkader. 1988. Agreement in Arabic, binding and coherence. In Agreement in natural language:

The Syntax of Arabic From A Generative Perspective Approaches, theories, description, ed. M. Barlow and C. Ferguson, 107–158. Stanford, CA: CSLI. ——. 1993. Issues in the structure of Arabic clauses and words. Dordrecht: Kluwer. ——. 1999. Arabic modifying adjectives and DP structures. Studia Linguistica 53: 105–154. Gelderen, Elly van. 1996. Parametrising agreement features in Arabic, Bantu languages, and varieties of English. Linguistics 34: 753–767. ——. 2008. Negative cycles. Linguistic Typology 12: 195–243. Guilliot, Nicolas and Nouman Malkawi. 2006. When resumption determines reconstruction. In Proceedings of the 25th West Coast Conference on formal linguistics, ed. Donald Baumer, David Montero, and Michael Scanton, 168– 176. Somerville, MA: Cascadilla Proceedings Project. Halila, Hafedh. 1992. Subject specifity effects in Tunisian Arabic. PhD diss., University of Southern California,Los Angeles. Hallman, Peter. 2000. The structure of agreement failure in Lebanese Arabic. In Proceedings of the 19th West Coast Conference on Formal Linguistics, ed. Roger Billerey, 178–190. Somerville, MA: Cascadilla Press. Harbert, Wayne and Maher Bahloul. 2002. Postverbal subjects in Arabic and the theory of agreement. In Themes in Arabic and Hebrew syntax, ed. U. Shlonsky and J. Ouhalla, 45–70. Dordrecht: Kluwer. Heggie, Lorie. 1988. The syntax of copular structures. PhD diss., University of Southern California,Los Angeles. Hoyt, Frederick. 2002. Impersonal agreement as a specificity effect in rural Palestinian Arabic. In PAL XIII-XIV, ed. Dil Parkinson and Elabbas Benmamoun, 111–141. Amsterdam: John Benjamins. Huang, C. T. James. 1982. Move WH in a language without WH movement. Linguistic Review 1: 369–416. Jelinek, Eloise. 1981. On defining categories: Aux and predicate in Egyptian Colloquial Arabic. PhD diss., University of Arizona, Tuscon. Kenstowicz, Michael. 1989. The null subject parameter in Modern Arabic Dialects. In The null subject parameter, ed. Osvaldo Jaeggli and Ken Safir, 263–275. Dordrecht: Kluwer. Kihm, Alain. 1999. Towards a predication theory of Semitic construct state nominals. Unpublished manuscript, CNRS,Paris. Koopman, Hilda and Dominique Sportiche. 1991. The position of subjects. Lingua 85: 211–258. Lorimor, Heidi. 2007. Conjunctions and grammatical agreement. PhD diss., University of Illinois at UrbanaChampaign. Malkawi, Nouman and Nicolas Guilliot. 2007. Reconstruction and islandhood in Jordanian Arabic. In PAL XX, ed. Mustafa Mughazy, 87–104. Amsterdam: John Benjamins. McCarthy, John. 1979. Formal problems in Semitic phonology and morphology. PhD diss., Massachusetts Institute of Technology,Cambridge. McCloskey, James. 1990. Resumptive pronouns,Ā-binding and levels of representation in Irish. In Syntax and semantics 23: Syntax of the modern Celtic languages, ed. Randall Hendrick, 199–248. San Diego: Academic Press. Mohammad, Mohammad. 1988. On the parallelism between IP and DP. In Proceedings of WCCFL VII, ed. Hagit Borer, 241–254. Stanford, CA: CSLI. ——. 1989. The sentence structure of Arabic. PhD diss., University of Southern California,Los Angeles. ——. 1999a. Word order, agreement and pronominalization in Standard and Palestinian Arabic. Amsterdam: John

The Syntax of Arabic From A Generative Perspective Benjamins. ——. 1999b. Checking and licensing inside DP in Palestinian Arabic. In PAL XII, ed. Elabbas Benmamoun, 27–44. Amsterdam: John Benjamins. Mouchaweh, Lina. 1986. De la syntaxe des petites pré positions. PhD diss., Universitéde Paris VIII, Paris. Moutaouakil, Ahmed. 1987. min qaḍaayaa r-raabiṭ fii l-lugha l-’arabiyya. Casablanca: ‘Ocaadh. Mughazy Mustafa. 2004. Subatomic semantics and the active participle in Egyptian. PhD diss.,University of Illinois, Urbana-Champaign. Munn, Alan. 1999. First conjunct agreement: Against a clausal analysis. Linguistic Inquiry 30 (4): 643–668. Nouhi, Youssef. 1996. Wh-constructions in Moroccan Arabic. MA thesis,University of Ottawa, Ottawa. Ouali, H. and C. Fortin. 2007. The syntax of complex tense in Moroccan Arabic. In PAL XIV, ed. Elabbas Benmamoun,175–190. Amsterdam: John Benjamins. Ouhalla, Jamal. 1988. The syntax of head movement: A study of Berber. PhD diss., University College London. ——. 1993a. Negation, focus and tense: The Arabicmaa and laa. Rivista di Linguistica 5: 275–300. ——. 1993b. Subject-extraction, negation, and the anti-agreement effect. Natural Language and Linguistic Theory 11: 477–518. ——. 1994. Verb movement and word order in Arabic. In Verb movement, ed. D. Lightfoot and N. Hornstein, 41–72. Cambridge, UK: Cambridge University Press. ——. 1996a. The construct state in Berber. In Studies in Afroasiatic grammar, ed. Jacqueline Lecarme et al., 278– 301. The Hague: Holland Academic Graphics. ——. 1996b. Remarks on the binding properties of wh-pronouns. Linguistic Inquiry 27: 676–707. ——. 2001. Parasitic gaps and resumptive pronouns. In Parasitic gaps, ed. Peter Culicover and Paul Postal,147– 180. Cambridge, MA: MIT Press. ——. 2002. The structure and logical form of negative sentences in Arabic. In Themes in Arabic and Hebrew syntax, ed. J. Ouhalla and U. Shlonsky, 299–320. Dordrecht: Kluwer. Pesetsky, David. 1987. Wh-in-situ: Movement and unselective binding. In The representation of (in)definiteness, ed. Eric Reuland and Alice ter Meulen. Cambridge, MA: MIT Press. Ritter, Elizabeth. 1988. A head-movement approach to construct state noun phrases. Linguistics 26: 909–929. Roberts, Ian and Ur Shlonsky. 1996. Pronominal enclisis in VSO languages. In The syntax of the Celtic languages, ed. Robert D. Borsely and Ian Roberts, 171–199. Cambridge, UK: Cambridge University Press. Rouveret, Alain. 1994. Syntaxe du Gallois. Paris: CNRS Editions. Ryding, Karin. 2005. A reference grammar of Modern Standard Arabic. Cambridge, UK: Cambridge University Press. Sells, Peter. 1984. Syntax and semantics of resumptive pronouns. PhD diss., University of Massachussets, Amherst. Shlonsky, Ur. 1991. Quantifiers as functional heads: A study of quantifier float in Hebrew. Lingua 84: 159–180. ——. 1992. Resumptive pronouns as last resort. Linguistic Inquiry 23: 443–468. ——. 1997. Clause structure and word order in Hebrew and Arabic: An essay in comparative Semitic syntax. Oxford: Oxford University Press.

The Syntax of Arabic From A Generative Perspective ——. 2004. The form of Semitic noun phrases. Lingua 114: 1465–1526. Siloni, Tal. 2001. Construct States at the PF interface. Linguistic Variation Yearbook 1: 229–266. Soltan, Usama. 2007. On formal feature licensing in minimalism: Aspects of Standard Arabic morphosyntax. PhD diss., University of Maryland. ——. 2010. On licensing wh-scope: Wh-questions in Egyptian Arabic revisited. Unpublished manuscript,Middlebury College, Vermont. ——. 2011. On strategies of question-formation and the grammatical status of the Q-particle huwwa in Egyptian Arabic Wh-questions. University of Pennsylvania Working Papers in Linguistics 1: Article 24. Available at http://repository.upenn.edu/pwpl/vol17/iss1/24. Sproat, Richard and Chilin Shih. 1988. The cross-linguistic distribution of adjective ordering restrictions. In Interdisciplinary approaches to language, ed. C. Georgopoulos and R. Ishihara, 565–593. Dordrecht: Kluwer Wahba, Wafaa. 1984. Wh-constructions in Egyptian Arabic. PhD diss., University of Illinois, Urbana-Champaign. ——. 1991. LF-movement in Iraqi Arabic. In Logical structure and linguistic structure, ed. C.-T. J. Huang and R. May, 253–276. Dordrecht: Kluwer Academic Publishers.

Notes: (*) Elabbas Benmamoun’s research for this paper was supported in part by National Science Foundation (NSF) grant BCS 0826672. (1) In this respect, MA patterns with the modern spoken Arabic varieties, which also lack morphological case marking. (2 ) There is also a debate about the status of the so-called pronominal copula found in contexts such as (i) from MA: ((i)) ħna huma

l-xəddama

we they

DEF-workers

“We are the workers.”

Notice that the pronounhuma “they” agrees with the subject pronounħna “we” only in number and gender. They obviously do not agree in person. The critical problem about this type of construction, found across Arabic varieties including SA, is the syntactic status of the pronoun. See Eid (1991) for an overview of its distribution and analysis. (3 ) As is well-known, the nature of the temporal and aspectual systems of Arabic, and Semitic in general, is not settled. We will refer to the so-called perfective verb ( maaḍii in the Arabic tradition) as past. With regard to generative approaches, the critical issue is whether there is a grammatical category that encodes tense and aspect and that interacts with nominals, such as the subject. This issue will be discussed in more detail later on. (4 ) For example, the English present tense is null, but it is associated with nominative case. (5) Note that the modal in SA in (8) is a particle that does not inflect. In MA, for instance, the modal inflects for agreement with the subject (i). ((i))

The Syntax of Arabic From A Generative Perspective (a.) yəqdər

ykuun

f-ḍ-ḍar

may

in-DEF-house

“He may be in the house.”

(b.) *yəqdər

f-ḍ-ḍar

may

in-DEF-house

(6 ) The same analysis extends to the interaction between tense and the verb in English. The tense needs to be merged with the verb, but that merger may be blocked by an intervening element such as negation, resulting in a repair mechanism, such as Do-support. This has been a mainstay of generative syntax since its inception. Putting aside the details and the merits of the analysis, the main point, which is not controversial, is that members of the sentence interact, and this may get disrupted if other elements intervene. (7 ) However, the argument may lose its force if it turns out that the merger between negation and the predicate takes place not in the syntax but in some postsyntactic component. Such a component would have the vocabulary and primitives to capture generalizations that are sensitive to the phonological content of the members of the syntactic phrase marker. At this stage, the nature of the postsyntactic component is still vague and cannot be properly subjected to scrutiny. At any rate, the fact that the minimality effects from the negative constructions align neatly with the arguments from modals and case is significant. (8 ) See Eisele (1988) for a detailed study of tense in Egyptian Arabic (EgA). See also Mughazy (2004) for a study of participles in EgA and the temporal and aspectual properties of the clauses they head. (9 ) There is also disagreement about whether the vocalic melody on active verbs in SA carries tense or aspect. McCarthy (1979) assumes that it does, but Benmamoun (2000) and Ouali and Fortin (2007) adopt the opposite view. For Benmamoun (2000), the fact that in MA there is no discernable vocalic melody casts doubt on attributing any temporal or aspectual properties to vowels in Arabic in general. The other problem is that, even in SA, vocalic melodies differ for no temporal or aspectual reason. In addition, the vocalic melodies on singular nouns seem to have no grammatical content, which raises further suspicion about the vocalic melody in Arabic morphology in general. However, the vocalic melody in the passive verb and statives does seem to have semantic content but this is not the case in most of the modern dialects. Further research is obviously needed to understand the grammatical content of the vocalic patterns. (10 ) Ouali and Fortin (2007) argue that the prefixes/proclitics found on imperfective verbs in MA carry tense. They also argue that when the auxiliary verb is present the sentence has a biclausal structure that contains two tense projections that dominate aspectual projections, which in turn dominate the main verbal projection. (11) For a detailed description of the structure of the noun phrase in SA, see Ryding (2005) and references therein. (12 ) In the Q__NP order, the noun phrase (NP) in SA appears with genitive case. In the NP__Q order the quantifier carries a clitic agreeing with the NP in number and gender. In addition, the quantifier agrees in case with the preceding NP. (13 ) While in some dialects of Arabic the order definite article-cardinal number-noun is acceptable, this is not possible in SA: ((i))

The Syntax of Arabic From A Generative Perspective (a.) * ʔaθ-θalaaθ-at-u

ʔawlaad-in

DEF-three-F-NOM

boys-GEN

“the three boys”

(b.) l-tleet

wleed

DEF-three

children

“the three children”

(c.) * l-teelit

walad

DEF-third

child

“the third child”

(14 ) NumP conflates ordinal and cardinal numbers. Shlonsky (2004) makes a distinction between the two, and our NumP corresponds to his Card#P, the projection that hosts cardinal numbers. (15) The original argument for N-movement within the Semitic noun phrase can be found in Ritter (1988). There, Nmovement was also motivated on the basis of unifying the syntax of simplex and complex noun phrases. (16 ) The examples in (30) correspond to Fassi Fehri (1999: 107, (1)–(2)). (17 ) For arguments that adjectival modifiers are indeed generated prenominally in Arabic, see Fassi Fehri (1999) (18 ) Shlonsky (2004) extended the XP-movement analysis previously outlined to account for the alternation between the pronominal and postnominal order of numerals and quantifiers, illustrated in (21)–(27) (see also Shlonsky 1991). In a study of the syntax of quantifiers in Arabic, Benmamoun (1999) argued, however, that the Q___NP ordering (23) and the NP___Q ordering (26) are radically different and that they involve different derivations. In that case, the prenominal order and postnominal order of modifiers would have two different syntactic sources, and the perceived alternation does not constitute evidence for XP-movement within the Arabic noun phrase. Of course, the picture is more complicated than our discussion would lead one to believe: both Fassi Fehri (1999) and Shlonsky (2004) present arguments for the limitations of massive XP-movement within the Arabic NP. Our main point here is to show how the syntax of the noun phrase in Arabic weighs in on an important debate in the generative literature concerning movement within the noun phrase. (19 ) For a more exhaustive list of properties, see Borer (1996), Mohammad (1999b), and Fassi Fehri (1999). (20 ) The genitive marker varies from dialect to dialect. Brustad (2000: 72) provides the table in (i) showing the genitive exponents in four different dialects representing four major dialect groups (Maghreb, Egyptian, Levantine, and Gulf Arabic):

The Syntax of Arabic From A Generative Perspective ((i)) Genitive Exponents Masculine

Feminine

Plural

Moroccan

dyal/d

–

Egyptian

bitaaʕ

bitaʕit

bituuʕ

Syrian

tabaʕ

–

(tabaʕul)

Kuwaiti

maal

(maalat)

(maalot)

(21) The examples in (i) from LA are consistent with this generalization. ((i)) (a.) l-χams

kətub

DEF-five

books

“the five books”

(b.) *l-χams

kətub

l-walad

DEF-five

books

DEF-child

“the five books of the child”

(c.) kətub

l-walad

l-χamse

DEF-books

DEF-child

DEF-five

“The five books of the child”

While in some dialects of Arabic the order definite article-cardinal number-noun is acceptable (see also footnote 13), this order is not available with construct state nominals. The only order possible in that case is the one where the cardinal number word follows the construct state nominal. (22 ) Lexical merger, as Benmamoun (2000) discusses, can be quickly dismissed since members of the construct state nominal are syntactically and semantically compositional. For more on this issue, see Siloni (2001) and Borer (2008). Benmamoun (2000) provides, in addition, arguments against syntactic merger between the two members of a construct state nominal. Due to limited space, we will not discuss those arguments here. (23 ) Benmamoun (2003) further investigates the nominal origins of the VS sequence that has led to this parallelism. 24

The Syntax of Arabic From A Generative Perspective (24 ) See Fassi Fehri (1993: 91). (25) As pointed out by the traditional Arab grammarians, some speakers deployed full agreement in both orders, which is similar to the pattern we find in the modern spoken dialects. However, the asymmetry is a property of the language or at least the varieties studied by the grammarians and therefore is a legitimate topic of study. To claim otherwise would be to make Arabic appear to be different from any other language that displays variation among its speakers and dialects, which is an indication of change in progress that could be the result of a loss of property or the emergence of a new property (innovation). Incidentally, agreement asymmetries relative to word order are attested in a number of languages (e.g., participial agreement in French). Therefore, the situation in SA is not unique. (26 ) The picture is, of course, more complicated. So-called null pronominals can be licensed in rich discourse contexts where person agreement may not be present. However, absent such rich contexts, person agreement seems critical. It should be mentioned that the syntactic status of the category null pronominal remains undetermined within minimalist approaches. This in turn has something to do with the still unclear relation between thematic structure, the notion of subject, and the syntactic configuration. One reason for positing a null pronominal was the need to satisfy the thematic criterion or to the fulfill the subject requirement of clauses (so-called extended projection principle) in addition to providing binders for anaphors and place holders for circumventing island conditions in syntactic dependencies. (27 ) See also Doron and Heycock (1999), who argue for two types of subjects for Arabic, the so-called broad subject and the so-called narrow subject. (28 ) The fourth option, namely, neither movement of the subject nor movement of the verb, may also be possible (see Borer 1995). (29 ) If full agreement is the realization of a pronominal, the issue of the asymmetry dissolves, but then one must contend with the fact that full agreement is found in a variety of contexts where it would be hard to argue that it is a pronominal. For example, it can be found on both an auxiliary verb and a main verb when it seems that there is one source for the pronoun in the specifier of the lexical projection of the main verb. ((i)) kun-na

ya-ʔkul-na

be.PST-3.F.PL

3-eat-F.PL

“They were eating.”

Also, a lexical subject in the same context fully agrees with the lexical verb and partially agrees with the auxiliary: ((ii)) kaan-at

ṭ-ṭaalib-aat-u

ya-ʔkul-na

be.PST-3.F

DEF-student-F.PL-NOM

3-eat-F.PL

“The students were eating.”

It would be hard to maintain the claim that the full agreement in (ii) on the main verb is a pronoun given that there is already a subject between the verb and the auxiliary. (30 ) Some spoken dialects, such as PA and LA, do display an agreement asymmetry relative to word order in the context of indefinite subjects (see Hallman 2000; Hoyt 2002).

The Syntax of Arabic From A Generative Perspective

(31) The structure of coordination and the computation of the features of its members must also be part of any analysis of agreement in the context of coordination. See in this connection Benmamoun (1992), Bahloul and Harbert (1993), Munn (1993), Harbert and Bahloul (2002), and Soltan (2007). (32 ) Ouhalla (1993a) provides strong arguments that the negative maa in SA, which is associated with focus, is located in a higher projection specified for focus features. (33 ) We are putting aside important details for lack of space. Soltan (2007) relies crucially on the notion that noninterpretable features on negation drive the checking relation. He assumes that there is a tense feature on negation that compels it to enter into a checking relation with the tense feature on the verb. (34 ) See Brustad (2000). (35) The picture is actually more complex. We do find the circumfix in the context of nonverbal predicates and the independent negative in the context of verbs and pseudo-verbs (Eid 1993; Brustad 2000; Soltan 2007). (36 ) There are many constructions that exhibit long A’-dependencies, including wh-interrogatives, restrictive relatives, free relatives, cleft s, (clitic-)left dislocation, topicalization, tough-movement constructions, parasitic gaps, and comparative clauses, among others. In this chapter, we will focus on wh-interrogatives, illustrated in (58), restrictive relatives (ia), and clitic-left dislocation (CLLD) (ib). ((i)) (a.) haʔulaaʔ

šurakaaʔ-u-na

allaðiina

kun-na

na-dʕuu

min duuni-ka

those

partners-NOM-our

that.M.PL

were-1.PL

1.PL-invite

without-you

“Those are our partners that we used to invite without you.”

(b.) ʔinna

haaðaa

l-xiṭ

li-r-raʔiis

Bush

kaana

yumkinu

COMP

this

DEF-speech

to-DEF-president

Bush

was.3

be.possible.3

ʔan

yaktuba-hu

masʔuul-un

al-Likud

that

write-3.M.SG

official-NOM

DEF-Likud

“This speech by President Bush, an official from the Likud Party could have written it.”

(37 ) Adjunct wh-words in Egyptian Arabic can marginally appear in the left periphery related to a gap within the sentence (i): ((i))

The Syntax of Arabic From A Generative Perspective

??fein/ʔimtaa/ʔizzaay/leih

ʔaħmad

ha-yi-saafir

where/when/how/why

Ahmad

FT-3-travel

“Where/When/How/Why will Ahmad travel?”

(38 ) Soltan (2011) suggests that the constraint in (i) is at work in EgA: ((i)) A’-positions must be resumed. (39 ) Ouhalla (2001) uses the term pronoun, since he discusses only strong and weak pronouns as resumptive elements. Due to space limitations, we will not extend the discussion to epithets. For an analysis of epithets as resumptive pronominals in Arabic see Aoun and Choueiri (2000) and Aoun, Choueiri, and Hornstein (2001). (40 ) Weak pronouns, which occur in all nonsubject positions, are affixed to heads, for example, V, N, or P (as seen in (72)); strong or tonic pronouns, which usually occur in subject positions, are free morphemes (as seen in (73)). (41) Following Roberts and Shlonsky (1996), Ouhalla (2001) assumes that nonsubject clitics in Arabic are agreement affixes, having the representation in (ib). ((i)) (a.) šəf-t-/kteeb-/minn-o saw-1SG-/book-/from-him/his “I saw him/his book/from him.” (b.) X+cl [DP pro] (b.) X+cl [DP huwwa] The enclitic pronominal, which can be manifested on the categories V, N, and P, as seen in (ia), is in fact an agreement affix identifying a null pronoun argument. In certain contexts, Ouhalla (2001) further claims, the null pronoun can be replaced by an overt strong pronoun, as illustrated in (ic). This analysis brings together strong pronouns appearing in object position with those that appear in subject position, since the latter usually co-occur with agreement morphemes on the verb. (42 ) Here we are making the assumption that wh-words like miin “who” (in (73a) and (75a)) are operators, while wh-phrases like šmen ṭalib? “which student” (in (78)) are not (see Pesetsky 1987 for a defense of such an assumption). (43 ) Shlonsky (1992) observes that the only position where gaps are allowed and resumptives prohibited in PA is the highest subject position (HSR), as seen in (i): ((i)) l-bint

ʔilli(*hiy)

raayħ-a

ʕa-l-beet

DEF-girl

that (*she)

going-F.SG

to-DEF-house

“the girl that is going home”

Aoun and Choueiri (1996) present evidence that HSR doesn’t apply in LA restrictive relatives. (44 ) See Shlonsky (1992) for a possible analysis of the alternation between the resumptive and gap strategies in SA restrictive relatives that relies on the agreement properties of the relative marker.

The Syntax of Arabic From A Generative Perspective (45) To illustrate the generalization, we are using here examples of adjunct islands. It is important, however, to note that the generalization extends to other types of islands (e.g., relative clause islands and wh-islands) as well. (46 ) Principle C of the binding theory (Chomsky 1981) blocks identity between a referential expression and a ccommanding pronoun (i). ((i)) She said that Mary left. (She≠Mary) (47 ) For a discussion of wh-in-situ questions in LA, see Aoun and Choueiri (1999) and Aoun, Benmamoun, and Choueiri (2010). Elabbas Benmamoun Elabbas Benmamoun, University of Illinois

Lina Choueiri Lina Choueiri, American University of Beirut

The Philological Approach to Arabic Grammar Lutz Edzard The Oxford Handbook of Arabic Linguistics Edited by Jonathan Owens Print Publication Date: Sep 2013 Online Publication Date: Dec 2013

Subject: Linguistics, Language Contact DOI: 10.1093/oxfordhb/9780199764136.013.0007

Abstract and Keywords This article is devoted to the philological analysis of selected features of Classical Arabic and Modern Standard Arabic as well as selected features of Middle Arabic. The sections on the writing system and phonology as well as morphology are held rather brief, as the aspect of “cultural embeddedness” seems to be less relevant in these realms of grammar. First and foremost, it reflects a nonexhaustive overview of some cases or even causes célèbres in Arabic syntax and semantics. The article concludes with a short philological analysis of a Middle Arabic (here: Judeo-Arabic) text. Keywords: philological analysis, Classical Arabic, Modern Standard Arabic, Middle Arabic, syntax, semantics

7.1 Introduction The term philology can be narrowed down to the study of language in written historical texts, encompassing aspects of history, literary studies, and linguistics. The philological approach can be language specific or comparative and can also refer more specifically to the decipherment, edition, and preservation of texts. This chapter provides an overview over some important ideas in the Western philological tradition of the description of Arabic on its levels of phonology, morphology, syntax, and lexicon. Thereby, the term philological implies a databased approach, not to the exclusion of linguistic approaches, as long as these are oriented at word lists and texts as opposed to exclusively linguistic theory. The philological approach to Arabic also considers comparative Semitic evidence and other comparative data. Indeed, a typological perspective is an important ingredient of the philological approach to grammar. Keeping in mind that the traces of the Western philological tradition are deeply rooted in boThthe Arab and the European Middle Ages (cf., e.g., Fück 1955; Bobzin 1992, 1995; Hamilton 2006), this chapter takes at its point of departure and terminus post quem the Arabic grammar by Sylvestre de Sacy dating from 1831 rather than the work of Petrus Venerabilis or other important figures in the Middle Ages (for a motivation of this divide, cf., e.g., Bobzin 1992: 155). In a European context, philological has also been understood as referring specifically to the philology of Greek and Latin (“Altphilologie”), or philology par excellence. Of course, the philological description of Arabic cannot be treated in isolation from the Classical tradition. Regarding the Arabic grammatical tradition, Versteegh (1977) argues for strong influence of or dependency on the Greek tradition, pointing, for instance, to the same set of examples of prototypical nouns ( “man,” /faras “horse,” and “wall”) in both the writings of Dionysios Thrax and the Arabic grammarians (Edzard 2001a). The evidence of Greek influence, usually transmitted via an intermediate Syriac–Aramaic stage, is even stronger in the realms of the sciences like philosophy, medicine, astronomy, and music theory. The philological approach to Arabic grammar is, of course, only a small part of what can be considered Arabic philology in general. Noteworthy examples of anthologies covering aspects of Arabic philology include the Cambridge History of Arabic Literature (notably the 1990 volume Religion, Learning and Science in the ‘Abbasid Period) and the Grundriß der Arabischen Philologie (three volumes, 1982–1992 edited by Wolfdietrich Fischer and

The Philological Approach to Arabic Grammar Helmut G ätje). The first volume of the latter series treats the whole history of the Arabic language, from pre-Islamic times up to the modern dialects, and examines special features of the lexicon and onomastics. The same volume also includes in-depth information on script history, epigraphy, numismatics, papyri, and features of Arabic manuscripts in general, areas that constitute indispensable tools for the diachronic and synchronic analysis of Arabic texts. The second volume in the series is devoted to Arabic literature, in a meaningful sequence that reflects boThchronology and scholarly dependence on respective other genres of literature: poetry, the Quran, lexicography, grammatical theory, poetology, rhetorics and metrics, ʾadab literature, artificial prose, popular literature, and modern literature in general. Further chapters cover historiography and geography, religious literature, and scientific literature (philosophy, mathematical sciences, medicine, and other realms) and the development of subject-specific language and terminology (cf. notably Endress 1992). While the Quran is considered to be sui generis in the Muslim tradition [Larcher, “ALT II”] (even though it certainly has poetic features), the other genres bear witness to the mentioned scholarly dependency. Both lexicography and grammatical theory had to rely on poetry and the Quran as essential databases: poetology, rhetorics, and metrics obviously were dependent on all of the previous genres; all of the other genres depend on the former, reflecting the deeply entrenched intertextuality in Arabic culture. Finally, the genre of ʾinšā’ literature is an important research area for philologists, relating to boThform and style of Arabic administrative documents. The latter is closely linked to the field of Arabic papyrology, which has boomed in recent years and among whose major modern exponents are Werner Diem (e.g., Diem 2006), Simon Hopkins (e.g., Hopkins 1984), and Geoff rey Khan (e.g., Khan 2011). As the philological approach is primarily concerned with a diachronic analysis, the definition of a terminus post quem for what constitutes Arabs and Arabic is important. According to Retsö (2003), the term Arab can for the first time be identified as one of the participants in the battle of Qarqar in Syria (853 BC). As far as written sources are concerned, the inscription of an-Namāra dating from 328 AD is often cited as one of the first documents with clearly identifiable Arabic features, such as the definite article (’)l- in line two of the inscriptions: 〈ʾ 1 ʾ š d y n 〉 “the two ʾŠD” (cf. Retsö 2003: 467–470; [Retsö, “Arabic?”]). The present chapter is devoted to the philological analysis of selected features of Classical Arab and Modern Standard Arabic on one hand and selected features of Middle Arabic on the other; in principle, however, the philological approach could also be meaningful as applied to corpora of modern (or older) dialectal data in transcription. The following subchapters regarding the writing system and phonology as well as morphology are rather brief, as the aspect of cultural embeddedness seems to be less relevant in these realms of grammar. First and foremost, the following reflects a nonexhaustive overview of some cases or even causes célèbres in Arabic syntax and semantics, selected by the author. This chapter closes wiTha short philological analysis of a Middle Arabic (here Judeo-Arabic) text.

7.2 Writing System and Phonology As stated already, proper decipherment and edition of texts is a cornerstone of philology, and this is no trivial endeavor in the case of early Arabic inscriptions and Arabic manuscripts in general. The Nabatean script (attested from c. 100 BC until 350 AD) counts as the genetic ancestor of the Arabic alphabet (for an overview, cf. Gruendler 2006). As the phonemic inventory of Classical Arabic is slightly larger than that of Aramaic, and, since the systematic use of diacritics (naqṭ) emerged only in the second half of the first century after the Hiǧra, a lot of ambiguities exist when it comes to the reading of early documents. A building inscription on a dam of Muʿāwiya (from the year 677 AD) and Quran manuscripts in Ḥiǧāzi script are among the first documents with consistent use of diacritical marks (ibid., 151). Research on phonology, especially in the older stages of Arabic, obviously depends on a precise reading of the historical sources. For instance, the correct understanding of the opposition pair maǧhūr versus mahmūs or the historically voiced quality of both/ṭ/ and /q/ would not be known today. Neither would one be aware of lateral quality of /ḍ/ (e.g., compare Arabic al-qāḍī “the judge” as reflected in Spanish alcalde “mayor” wiThreanalysis of the Arabic definite article as part of the word) if not for the information contained in Chapters 565 through 571 of Sībawayhi’s Kitāb and other historical sources (cf., e.g., Fleisch 1958a, 1958b; Al-Nassir 1993; Edzard 2001b). Sībawayhi’s declared rationale in Chapter 565 for providing extremely precise descriptions of pronunciation is to

The Philological Approach to Arabic Grammar explain why certain consonants (can) undergo assimilation (ʾidġām or iddiġām; [Baalbaki, “ALT I”]) within and across words and why other consonants cannot. One of the points addressed by Sībawayhi are assimilation and resyllabification in forms V, VI, VII, and VIII. Sībawayhi quotes a number of alternative forms of the standard diatheses V, VI, VII, and VIII, all of which feature assimilation and resyllabification. In forms V and VI, haplological syllable ellipsis occurs in cases like fa-lā(t)tanāǧaw! “don’t whisper to each other.” Sībawayhi also quotes assimilated verbs of form VII (standard /inC1aC2 aC3 a/) that are not part of the standard language, such as immaḥā “he was effaced.” (1) gives derivations of the nonstandard examples of forms V, VI and VII: (1) Derivation of nonstandard forms (V, VI, and VII) tatamannawna “You (m. pl.) wish.”

→

tamannawna

fa-lātatanāǧaw! “Don’t whisper to each other!”

→

fa-lā(t)tanāǧaw!

inmahḥā “He was effaced.”

→

immahāā

The situation in form VIII is more complex, as Sībawayhi cites an array of forms that exceeds the well-known cases of partial and total assimilation that may occur in this form. Example (2) presents an overview of the nonstandard output forms (masculine plural of perfect, imperfect, and participle) of the verb {q-t-l} “to kill” in form VIII (“to kill each other”) that normally do not undergo any assimilatory change: (2) Nonstandard assimilation and metathesis (VIII) iqtatalū

→

qittalū

yaqtatilūna

→

{yaqattilūna, yaqittilūna}

muqtatilūna

→

{muqattilūna, muqittilūna}

The ordering in the set brackets indicates that people who say yaqattilūna will also say muqattilūna, and so on. Interestingly, these forms amount to an assimilation of the infixes (-t-) to the middle radical, as is also obvious in the participle form murtadifūna → muruddifūna “[they (m.) are] directly following,” which also features vowel harmony with respect to u. The driving force behind these forms appears to be the wish to avoid a sequence of equal or similar CV syllables. As a result, the underlying form of the diathesis (VIII) is quite opaque in these cases. Forms with total phonological merger (progressive, regressive, or “reciprocal” assimilation, that is, “compromise” on a phonetically intermediate consonant) arise in the case of verbs whose first radical is a voiced or velarized sibilant or a voiced or velarized alveolar stop: (3) Nonstandard progressive and regressive assimilation (VIII) isṣṭ “He was patient.”

→

iṣṣabara

iṣṭaǧara “He was angry.”

→

iḍḍaǧara

iẓṭalama “He suff ered injustice.

→

iṭṭalama

The last output form clearly demonstrates that /ṭ/ historically was voiced [Embarki, “Phonetics”]; otherwise, the

The Philological Approach to Arabic Grammar modern concept of partial regressive assimilation with respect to the feature [±voiced] would not make sense in this case. Most of the forms cited by Sībawayhi are also attested in the Quran (for a collection of forms, cf. Vollers 1906: 111– 122). A careful consideration of the textus receptus of the Cairene Quran edition also yields important information on other phonological features, both segmental and suprasegmental, of Classical Arabic, including assimilation and haplological syllable ellipsis. Special signs indicate partial or total assimilation to be observed in proper taǧwīd “Quranic recitation.” The following example illustrates both haplology and assimilation features with assimilated/elided positions in boldface: (4) Haplology and assimilation in the Quran (Q 12:2) ʾinnā ʾanzalnā-hu qurʾānan ʿarabīya lla ʿalla-kum ta ʿlamūna instead of ʾinna-nā ʿanzalnā-hu qurʾānan ʿarabīyan laʾalla-kum taʿlamūna “We [God] have indeed sent it [the Quran] down as an Arabic Quran; maybe you will know.” Also, special Quranic signs, that is, minuscule wāws and yāʾs above a word-final 〈h〉, indicate that the possessive and object suffix of the third-person masculine singular 〈-h〉 is long after a short syllable (CV), that is, -hū/-hī, and short of a long syllable (C[001] or CvC), that is, -hi/-hi, a circumstance also known from Classical Arabic metrics. Guttural phonology can also be mentioned as a realm, where a comprehensive evaluation of attested form is essential for coming to sound conclusions, for example, regarding the distribution of imperfect (“theme”) vowels. McCarthy (1991) provides ample documentation for Arabic in this context.

7.3 Morphology A philological approach, that is, a close evaluation of texts and forms in context, is indispensable in deciding on the status of a certain morphological category. As an example, the status of the so-called pluralis paucitatis (ǧamʾ alqilla), characterized by a prefixed ʾa- (e.g., Fischer 2006: 57 (= Nöldeke 1896)) has recently come under discussion. Relevant forms include ʾaʿyun “(few) eyes” instead of ʿuyūn “(more than a few) eyes.” A philological approach is also essential when it comes to detecting new morphological patterns and wordformation strategies (or the further development of preexisting ones) through careful examination of written sources. More recent developments in this context include compound formations such as the phenomenon of blends, where one or both of the constituting elements are shortened (e.g., Badawi et al. 2004: 58, 751). Relevant examples include substantives and adjectives with prefixes as well as structures with an internal appositional structure, such as lā-nihāʾīya “infinity,” šibh-rasmī “semiofficial,” or šarq-ʾawsaṭī “Middle Eastern.” Blending is also productive in modern times (e.g., Versteegh 2001: 181–183) and surfaces in such forms as kahraṭas “electromagnetism” (composed of kahrabāʾ “electricity” + maġnāṭṭῑs “magnet”). Loan words for scientific vocabulary may be fully integrated in the Arabic nominal system or retain (part) of their original structure. An example of a fully integrated technical term is the neologism raskala “recycling” (/C1aC2 C3 aC4 a/). An example of a partially integrated technical term is the neologism kibrītīd “sulfide,” where a European-style suffix -īd is attached to the Arabic equivalent of “sulfur,” kibrīt (cf. Badawi et al. 2004: 741). New patterns continue to emerge, such as the pattern /C1awC2 aC3 a/ for denominal verbs from nouns with a long first syllable: ʿawlama “globalization” from ʿālam “world” (cf. Badawi et al. 2004: 762). Some patterns have become prominent by way of qiyās [Baalbaki, “ALT I”], [Larcher, “ALT II”] through one word, such as /C1āC2 ūC3 / for technical instruments, a pattern “triggered” most likely by the noun ḥasūb “computer” (cf. also Edzard 2007; [Newman, “Nahda”]; [Kossmann, “Borrowing”]). Innovations can also be observed in the verbal system. A noteworthy example is the neologism ʾaslama, yuʾaslimu “to islamize” and its reflexive form taʾaslama, yataʾaslamu “to be(come) islamized,” as opposed to ʾaslama, yuslimu “to render oneself [to God], to become a Muslim,” resulting in the creation of a minimal pair ʾa-slama (/ʾaC1C2 aC3 a/) versus ʾaslama (/C1aC2 C3 aC4 a/). The former verb is modeled in its surface structure after verbs like ʾamraka “to Americanize” and its reflexive form taʾamraka “to be(come) Americanized.”

7.4 Syntax

The Philological Approach to Arabic Grammar 7.4.1 Overview Syntax is another realm of Arabic grammar that has seen considerable progress in the past decades. As far as the philological approach is concerned, Diem (1998, 2002, 2011) and Manfred Ullmann (1988, a detailed study on irreal conditional clauses) are two central authors in this respect. Peled (1992) attempts successfully to combine a philological with a (historical-)linguistic approach to Arabic syntax. As far as Diem’s studies are concerned, their particular strength lies in the detailed reference to the types of verbs involved and their specific context (for Diem 1998, see Section 7.3.3). Diem (2002) investigates the precise distribution of verbs with double accusative (“translocative” in more recent grammatical terminology) throughout the history of Arabic (for verbs wiThtwo pronominal suffixes, see also Gensler 2003). The basic examples are as follows: (5) Translocative constructions in Arabic ʾaʿṭay-tu

Zayd-ani

l-kitāb-a

gave-I

Zayd-ACC

DEF-book-ACC

“I gave Zayd the book.” ʾaʿṭay-ty

l-kitāb-a

li-Zayd-in

gave-I

DEF-book-ACC

to-Zayd-GEN

“I gave the book to Zayd.”

Diem (2002: 5) identifies the following semantic verb groups in this context (translocation of a thing from the translocator to the recipient or vice versa): (a) “to give” (e.g., ʾaʿṭā “to give”); (b) “to impose” (e.g., ʾalzama “to oblige s.o.”); (c) “to transmit a message” (e.g., ʿarrafa “to inform”); and (d) “withdraw” (e.g., saraqa “to steal from s.o.”). In his study on kayfa as a conjunction, for example, in sentences like ʾa-lam tara kayfa faʿala rabbu-ka biʾaṣḥābi l-fīli “have you not seen how God has dealt with the owners of the elephant,” Diem (2011) pursues a similar strategy. The relevant constructions are analyzed according to different verbal categories: (a) emotion; (b) perception; (c) communication; (d) intellect; and (e) affectual expressions. In this context it is noteworthy that neither Classical nor Modern Standard Arabic constitutes a monolithic block. Willmsen (2010), for instance, raises the issue that object pronouns in Egyptian newspapers syntactically are treated differently from those in Syrian and Jordanian newspapers. Syntax is a realm of grammar that especially warrants a typological perspective. Diem’s (1986) work on alienable and inalienable possession in Semitic is an example of research that clearly owes its depThof insight to such a typological perspective. From a broader perspective, it is important to note that Classical Arabic in terms of linguistic word order universals (cf. Greenberg 1966) is a model of a verb–subject–object (VSO) language, exhibiting the following features (“〉〉” signifies “implication”): 1) VSO 〉〉 prepositions 2) VSO 〉〉 prefixed (prespecifying) definite article 3) VSO 〉〉 adjectives, genitives, and relative clauses succeed head noun 4) VSO 〉〉 auxiliary before main verb 5) VSO 〉〉 standard of comparison after adjective Here are some illustrating examples (6): (6) Predicted morphosyntactic features of Classical Arabic, patterns (6a)–(6e): a)

The Philological Approach to Arabic Grammar

bi-l-bayt-i

ʿalā

l-ʾarḍ -i

qabla

l-ġadāʾ-i

in-DEF-house-GEN

DEF-earth-GEN

before

DEF-lunch-GEN

“in the house”

“on earth”

“before lunch”

b) al-bayt-u

l-kabīr-u

DEF-house-NOM

DEF-big-NOM

“the big house” (subject) c) bayt-u-n

kabīr-u-n

al-bayt-u

l-kabīr-u

house-NOM-IDF

big-NOM-IDF

DEF-house-NOM

DEF-big-NOM

“a big house” (subject)

“the big house” (subject)

bayt-u

walad-i-n

bayt-u

l-walad-i

house-NOM

child-GEN-IDF

house-NOM

DEF-child-GEN

“a child’s house”

“the house of the child”

walad-u-n

ǧāʾa

al-walad-u

llaḏī ǧāʾa

child-NOM-IDF

he.came

DEF-child-NOM

who.MSG he.came

“a child that came”

“the child that came”

d) ya-ksir-u

ǧaʿala

ya-ksir-u

he-breaks-INDC

he.made

he-breaks-INDC

“He breaks”

“He began to break”

The Philological Approach to Arabic Grammar

ʾaṭwal-u

min Muḥammad-i-n

taller-NOM

from Muḥammad-GEN

“taller than Muḥammad” ʾaḥsan-u

min-hā

more.beautiful-NOM

from-her

“more beautiful than she”

Case marking (ʾiʿrāb) allows for a certain degree of syntactic variation, not in the (morpho-) syntax as outlined already but in the position of the basic components verb, subject, and object. Indeed, there has been discussion of whether word order in the history of Arabic allows one to determine at which point in history ʾiʿrāb was lost. Fück (1950: 2) adduces, inter alia, the following examples from the Quran, where the object appears before the subject and where a reading without ʾiʿrāb might give way to misunderstanding (7): (7) Classical Arabic sentences with VOS word order: ka-ḏālika ʾinna-mā yaxša llāha min ʿibādi-hī l-ʿulamāʾu (Q 35:28) “Out of his worshipers, only the scholars love God.” wa-ʾiḏi btalā ʾibrāhima rabbu-hū bi-kalimatin (Q 2:124) “when his Lord put Abraham to test” On the other hand, Wehr (1952: 181) argues that the Egyptian colloquial also featured occasional VOS word order, without causing semantic problems. Thus, it appears that word order per se is no argument for deciding on a timeline for the loss of ʾiʿrāb. Philological research also has to take into consideration whether subject and object are animate or inanimate and other factors. Verbal syntax and semantics continue to be an intriguing field in Arabic and Semitic linguistics. The difference between mā faʿala and lam yafʿal, has been a notorious problem, already treated by Wehr (1953), Larcher (1994), and Dahlgren (2006) [Benmamoun and Choueri, “Syntax”]. Early analyses tried to prove the existence of an aspectual opposition between mā faʿala (resultative) and lam yafʿ al (past). Wehr (1953) makes a strong case that mā often expresses “affected” involvement of the speaker and is almost mandatory after oaths and the like. Birnstiel (forthcoming) argues that the primary function of the former is to negate a constituent, whereas the primary function of the latter is to negate the predicative relationship between subject and predicate (nexus negation). Examples (8a) versus (8b) illustrate the contrast: (8) The opposition mā faʿ ala vs. lam yafʿ al a. wa-mā kafara sulaymānu wa-lākinna š-šayāṭῑna kafarū (Q 2:102) “For it was not Solomon who denied the truth, but it was the devils who disbelieved.” (constituent negation) b. wa-sawāʾun ʿalay-him ʾa-ʾanḏarta-hum ʾam lam tunḏir-hum lā yuʾminūn (Q 36:10) “It is the same to them whether you warn them or do not warn them: they will not believe.” (nexus negation) The opposition hypothetical versus counterfactual in irreal conditional structures (cf. Peled 1992: 40) is another case in point in this context. This opposition is encoded not by means of different particles but by the use of tense. Hypothetical conditions (i.e., conditions that can be realized) can be reflected by a protasis wiTha verb in the prefix conjugation (imperfect). The protases of counterfactual conditional structures (i.e., structures whose conditions cannot be realized) regularly feature kāna + suffix conjugation. Examples of boThcases are found in (9): (9) The opposition factual vs. counterfactual in conditional clauses

The Philological Approach to Arabic Grammar a. law našāʾu ǧaʿalnā-hu ʾuǧāǧan (Q 56:70) “If we [really] wanted (factual-hypothetical), we would make it [the drinking water] bitter.” b. law kuntum daʿawtumū-nāʾaṭaʿnā-kum “If you had called us (counterfactual), we would have obeyed you.”

7.4.2 Case Study: Arabic Mā as an Interrogative Element and a Negator In Classical Arabic, the status (pars orationis) of a given word may not always be clear, and a meticulous philological approach is essential when it comes to analyzing passages that defy an unambiguous interpretation. A famous passage in the sīra nabawīya concerns the interpretation of the particle mā in mā ʾaqraʾu (e.g., Brünnow and Fischer 2008: 41, Arabic text). According to the Ḥadīṯ tradition, this passage is to be understood as “I cannot/will not read/recite.” The translations of both Weil (1864, vol. 1: 114) and Rotter (1976: 44) follow this interpretation; Guillaume (1955: 106), however, translates “What shall I read.” Rubin (2005: 50) suggests that historically the Arabic interrogative pronoun mā has been grammaticalized as a negative marker. Instead of “grammaticalization,” one could also speak of an “illocutionary transgression” in such cases. At any rate, one may well argue that the interpretation “I cannot/will not read/recite” does not necessarily differ logically from the interpretation “what shall I read.” Generally speaking, the rhetorical statement “what can I do [in this matter]?” appears to be coextensive with the statement “I cannot do anything [in this matter].” Regarding the transmission of the report on Muḥammad’s first revelation, Schoeler (2010: 70) carefully shows how the original ambiguous statement mā ʾaqraʾu, as transmitted by az-Zuhrī on the authority of ʿUrwa, was later in Muslim tradition turned into the comparatively unequivocal statement mā ʾana bi-qāriʾin “I am not one to recite.” Wehr (1953: 36) adduces more examples from the Quran that are a priori ambiguous wiThrespect to the understanding of mā (10): (10) Ambiguous examples (interpretation of mā) from the Quran: mā ʾaġnā ʿan-hu mālu-hū wa-mā kasab (Q 111:2) “What was the benefit of his wealth for him and what did he acquire?” or “His wealth did not benefit him and he did not acquire.” wa-mā yubdiʾu l-bāṭilu wa-mā yuʿīdu (Q 34:49) “And what does the wrong produce and what does it help?” or “And the wrong does not produce anything, nor does it help.” The basic question to ask is whether these commonalities are only of a superficial phonological nature or of a deeper semantic nature. Even some book titles seem to imply a logical connection between these two categories, such as Bergsträsser’s (1914) Verneinungs- und Fragepartikeln und Verwandtes im Ḳurʾān. Ein Beitrag zur historischen Grammatik des Arabischen. Nekroumi (2003: 72) also highlights the distributional, or logical parallels, between interrogation and negation (11): (11) Distributional parallels between interrogation and negation hal ǧāʾa

Zayd-un

rākib-an

ʾamsi

Q came

Zaud-NOM

riding-ACC

yesterday

“Did Zayd come yesterday by horse?” mā ǧāʾa

Zayd-un

rākib-an

ʾamsi

not came

Zayd-NOM

riding-ACC

yesterday

“Zayd did not come yesterday by horse.”

In logical grammar, boThinterrogations and negations are simply defined as “modified forms of the

The Philological Approach to Arabic Grammar affirmative/declarative sentences” and thus form a semantic group. The most intuitive construction in this context is probably found in Wehr (1953: 36), who argues convincingly that one can hardly, if at all, distinguish interrogation from negation in negated exceptive clauses (12): (12) Interrogation or negation in negated exceptive clauses mā waʿada-nā llāhu wa-rasūlu-hū ʾil-lā ġurūran (Q 33:12) “What have God and his apostle promised us except deception?” = “God and his apostle have promised us nothing but deception” Another example concerns the grammarian Muḥammad ibn al-Mustanīr (Qutrub), who used to visit Sībawayhi early in the morning and to whom Sībawayhi would say (13): (13) Interrogation or negation in negated exceptive clauses mā ʾanta ʾillā quṭrubu laylin “What are you except a night-quṭrub.” = “You are nothing but a night-quṭrub.”

7.4.3 Case Study: Subject and Predicate in Arabic Sentences It is elucidating, for instance, to compare the approach of modern grammar to the sentence types of Arabic with that of native Arab(ic) grammar. Peled (2009) analyzes, inter alia, the following two sentence types, which were identical for many of the Arab grammarians (14): (14) Subject and predicate in Arabic sentences a. ʿAbd-u-llāhi

ḍaraba

zayd-an

ʿAbd-NOM- llāh

hit

zayd-ACC

“ʿAbdullāh hit Zayd.” b. Zayd-un

ḍarab-tu-hū

Zayd-NOM

hit-I-him

“I hit Zayd.” (“Zayd I hit him.”)

Whereas a “Western” analysis would classify (14b) as a topicalized (“left-dislocated”) variant of ḍarabtu zaydan, the majority of Arab grammarians classified (14b) as structurally identical with zaydun munṭaliqun “Zayd departs [“is departing”]” (“Zayd is somebody, whom I hit”), that is, ḍarabtu-hū in (14b) was treated as predicate, and not zaydun as left-dislocated topic. Peled (2009: 105) argues that the modern description of (14b) can either follow the left-dislocation or topicalization model (e.g., Holes 1995: 205–209) or the native Arabic approach without necessarily abandoning the concept of topicalizion. Badawi, Carter, and Gully (2004: 346) take a comparable position in their grammar of modern written Arabic. The following quotation, which refers to ḍarabtu zaydan “I hit Zayd” vs. Zaydun ḍarabtu-hū “idem” (“Zayd, I hit him”), is instructive: What happens to be inversion of the agent and verb [cf. (14a)] is actually a variety of topic-comment sentence, in which the topic, the agent of the comment-verb and the binding pronoun all happen to be identical (coreferential). In other words, there is no true inversion of agent and verb on the western pattern. This is confirmed by the fact that the verb + agent sequence contains only two elements while its apparent conversion contains three, a noun (= topic), a verb, and a pronoun agent (acting as both logical agent and binding pronoun).

The Philological Approach to Arabic Grammar 7.4.4 Case Study: The ʾIḍāfa ġayr Ḥaqīqīya and the Naʾt Sababī Construction Another case in point for an approach that combines an investigation of native Arabic grammatical theory and a modern philological-linguistic perspective is the improper annexation (ʾiḍāfa ġayr ḥaqīqīya) and the semantically linked qualifier (naʾt sababī), both of which have no direct equivalent in modern European languages. Already the grammar by de Sacy (1831, vol. 2: 275–280 ) provided a useful analysis. Here are examples of the ʾiḍāfa ġayr ḥaqīqīya in Classical Arabic (15): (15) The ʾiḍāfa ġayr ḥaqῑqῑya in Classical Arabic a. ḍāribu ʾaxī-hi “one who beats (or: has beaten) his brother” aḍ-ḍāribu ʾaxī-hi the one who beatsaten) his brother” b. raǧulun karīmu n-nasabi “a man of noble descent” ar-raǧulu l-karīmu n-nasabi “the man of noble descent” Whereas type (15a) is statistically rarer, one does find examples such as [a]l-muqῑmῑ ṣ-ṣalāti “the ones who perform prayer” (Q 22:35) or [a]š-šātimay ʾirḍ-ī “the two who are smearing my reputation” (verse 74 of the Muʿallaqa of ʿAntara). As Diem (1986: 248f.) has shown, European Arabists have referred to type (15b) only as “improper annexion,” while native Arab grammatical theory (notably az-Zamaxšarī, Mufaṣṣal, § 111) classifies both type (15a) and type (15b) as ʾiḍāfa lafẓῑya “(purely) formal annexation” or ʾiḍāfa ġayr ḥaqīqīya—as opposed to the ʾiḍāfa maʿnawīya, the “proper annexation.” The second related type, the semantically linked qualifier (naʿt sababī), exhibits the same rules regarding determination as can be observed in the ʾiḍāfa ġayr ḥaqīqīya (e.g., Fischer 2006: 194, 210). In addition, the phenomenon of “case attraction” is remarkable in this context (16), where ḥasan-an is linked morphologically through the agreeing accusative –an to the preceding noun, though it is semantically a predicate to the following waǧh: (16) The naồt sababī construction in Arabic raʾay-tu

mraʾat-a-n

ḥasan-a-n

waǧh-u-hā

saw-I

woman-ACC-IDF

beautiful-ACC-IDF

face-NOM-her

“I saw a woman with a beautiful face.” raʾay-tu

l-mraʾat-a

l-ḥasan-a

waǧh-u-hāa

saw-I

DEF-woman-ACC

DEF-beautiful-ACC

face-NOM-her

“I saw the woman with the beautiful face.”

The diachronical derivation of this type of construction, according to Diem (1998: 10, following Reckendorf 1898), can be represented as follows (17): (17) Derivation of the surface of a naʾt sababī structure,

The Philological Approach to Arabic Grammar

(an-nāsu) l-qāsiyatu

qulūbu-hum (Q 39: 22) →

nās-un qulūb-u-hum

qāsiyat-u-n

people-NOM hearts-NOM-their

hard-NOM-IDF

“people whose hearts are hard” (inversion within the relative clause)

→

nāsun qāsiyatun qulūbu-hum

→(determination/naʾt sababī)

an-nās-u

l-qāsiyat-u

qulūb-u-hum

DEF-people-NOM

DEF-hard-NOM

hearts-NOM-their

“the people wiThhard hearts”

→ (deletion of an-nāsu)

(“the hard-of-heart people”) (naʾt sababī) al-qāsiyat-u

qulūb-u-hum

DEF-hard-NOM

hearts-NOM-their

“those whose hearts are hard” (“the hard of heart”)

Typologically, the naʿt sababī structure continues to be hard to capture. Edzard (2011) represents an attempt to compare this type of construction with the stylistic figure enallagé (or hypallagé) adjectivi. Examples include the quotation from Shakespeare’s Hamlet (1, 1, 48), the majesty of buried Denmark, properly meaning “the buried majesty of Denmark,” or the quotation from Schiller’s Der Ring des Polykrates (6), der Schiffe mastenreicher Wald “the ships’ mast-rich forest,” meaning “the forest of mast-rich ships.” A tentative derivation comparable to the foregoing could look as follows (18): (18) Derivation of the surface structure of an enallagé (hypallagé) adjectivi “deep structure”: der Wald der mastenreichen Schiffe extraposition of the genitive: der mastenreichen Schiffe Wald enallagé (hypallagé) adjectivi: der Schiffe mastenreicher Wald Obviously, “deep structure” is understood here in a strictly synchronic sense. The enallagé (hypallagé) adjectivi differs from the naʿt sababī structure in that is cannot be analyzed as the result of the nominalization of a relative clause, but the structurally parallel form is still striking.

7.5 Lexicon Hans Wehr’s (1961) Dictionary of Modern Written Arabic continues to be of utmost importance, also for the reading of Classical Arabic texts. Edward Lane’s (1863) large-scale Arabic–English Dictionary, which transmits the information contained in the Classical Arabic dictionaries such as Lisān al-ʿarab and Tāǧ al-ʿarūs (az-Zabīdī), was followed up in the 20th century by the Wörterbuch der klassischen arabischen Sprache (WKAS) initiated by

The Philological Approach to Arabic Grammar Manfred Ullmann (1970), who decided to start wiThthe letter kāf, that is, at the place in the alphabet, where Lane’s dictionary no longer is exhaustive. So far, the letters kāf and lām have been completed in the large-scale WKAS, but the continuation of the project is not yet secured. Reinhard Dozy’s Supplément aux dictionnaires arabes (1927) still can be considered a useful tool in addition to the previously mentioned works. (For an overview of Classical Arabic lexicography, see Carter 1990.)

7.5.1 Case Study: Quṭrub One of the most impressive and detailed lexicographical studies is certainly Ullmann’s (1976) article on the Arabic “ghost word” quṭrub “wer(e)wolf.” The Lisān al-ʿarab and the Tāǧ al-ʿArūs give a host of meanings, taking up many sources, such as duwaybbatun kānat fī l-ǧāhilīya … laysa la-hā qarāruni l-batata “a little animal that existed in the Jahiliya … which does not ever have a period of rest,” ḏakaru l-ġīlāni “the male ghul” (desert demon), aṣ-ṣaġīru mina l-kilābi “the whelp” (“the little one of the dogs”), al-liṣṣu l-fārihu fī l-luṣūṣīya “the swift thief ” (“the thief who is agile in theft”), as well as others. The term quṭrub is also cited as the nickname of the grammarian Muḥammad ibn al-Mustanīr (cited in Section 7.3.2), who used to visit Sībawayhi early in the morning and to whom Sībawayhi then would say mā ʾanta ʾillā quṭrubu laylin “you are just the night-quṭrub. ” The dictionary closes with the formula wa-llāhu ʾaʿlam “God is all-knowing,” that is, the author and compilator of the dictionary did not know himself. The meaning “wer(e)wolf ” can be derived from the Greek word λυκάνθρωπος, a compound consisting of the elements λΰκος lykos “wolf” and ἄνθρωπος anthrōpos “man,” which underlies the Arabic word quṭrub. In the Qānūn, a medical dictionary by Ibn Sīnā, the term quṭrub is explained as nawʾun mina l-mālanxūliyā “a sort of melancholy,” which is characterized by some of the previously listed qualities, such as roaming around by night and having no rest. In the Greek history of medicine, this disease was known as “lycanthropy.” The word made its way into Arabic via an intermediate Syriac form qanṭropos. Ullmann points out plausibly that the ultimate Arabic form was triggered by the /C1uC2 C3 uC4 / pattern, which holds for animals such as furʿul “young hyena,” qunfuḏ “hedgehog,” and ǧundub “locust.” Of special interest is the semantic transition from a concrete term to an abstract one (designating a malady). While one would have expected an Arabic term for the malady such as dāʾ al-quṭrub, comparable to dāʾ aṯ-ṯaʿlab “alopecia,” the form quṭrub can, according to Ullmann, be traced back to a Greek manuscript, which in a list of maladies had the Greek wording περί λυκανθρώπου peri lykanthrōpou “about the wer(e)wolf ” (instead of expected περί λυκανθρωπίας peri lykanthrōpias “about the lycanthropy”). Hence, the Arabic translators faithfully rendered the term for this disease as quṭrub. Clearly, Ullmann’s analysis constitutes a model example of sound philological work.

7.5.2 Case Study: The “Virgins of Paradise” and the “Luxenberg” Hypothesis In 2000, a Christian Lebanese scholar operating under the pen name of “Luxenberg” set out to re-interpret “obscure” passages in the Quran (Q). The most notorious lexical item in this connection was the noun phrase ḥūr(un) ʿīn(un), approximately “round white-eyed ones” (“houris”), which is commonly understood in Classical Arabic as a metaphor for “beautiful (young) women” and which Luxenberg (2000) tried to reinterpret as a metaphor for “grapes,” apparently wiTha view to providing a more “Victorian” image of the Quran. The common opinion is that ḥūr is to be interpreted as the plural of feminine singular ḥawrāʾ “having eyes in which the contrast between black and white is very intense,” and ʿīn as the plural of feminine singular ʿaynāʾ “wide-eyed.” Against this, Luxenberg suggested interpreting ḥūr as an adjective derived from the Aramaic verb ḥwar “to be white,” and ʾīn as an Arabicized pausal form of Aramaic ʿaynē “jewels.” A philological approach, as already defined, is indispensable for passing judgment on such a dispute. Two quotations from the Quran are central in this context: Q 44:54 and Q 52:20, both featuring the passage wazawwaǧnā-hum bi-ḥūrin ʾῑnin … “And we [God] shall wed them to wide-eyed houris.…” To maintain the assumed meaning “grapes,” Luxenberg emended wa-zawwaǧnā-hum “and we [God] will marry them to …” to warawwaḥnā-hum “and we [God] will provide them with …,” based on the similar consonantal rasm, even though such a form is otherwise unattested in Classical Arabic. Wild (2010) successfully debunks Luxenberg’s proposals, not only by adducing relevant passages in the Quran describing the pleasures of paradise (notably Q 37:48–49, Q 44:51–56, Q 52:19–24, Q 55:56–58, Q 55:62–74, Q 56:15–38, and Q 78:31–34) but also by pointing to the description of the attributes of a banquet in pre-Islamic poetry (19):

The Philological Approach to Arabic Grammar (19) Excerpt (banquet description) of a poem by the pre-Islamic poet al-ʿAʾšā Maymūn wa-miskun wa-rayḥānun wa-rāḥun tuṣaffaqu

“And musk and basil and wine mixed with water,”

wa-ḥūrun ka-ʾamṯāli d-dumā wa-manāṣifu

And big-eyed beauties like statues, and servants,”

wa-qidrun wa-ṭ abbāxun wa-kaʾsun wa-daysaqu

“And a cooking pot, a cook, a goblet, and a plate.”

In sum, a proper philological analysis of this highly publicized issue has to point out at least the following flaws in Luxenberg’s analysis, according to Wild (2010): the unclear definition of a “Syro-Aramaic” substratum in 7thcentury Arabic; the disregard of other important philological witnesses (notably pre-Islamic poetry); and the projection of a Victorian world image (transferred to paradise) on Islam. Other controversial points for which Luxenberg has been criticized include the unsystematic tampering wiThthe consonantal text (rasm), the resulting relativism (i.e., an overproduction of results) due to the previous problem, the disregard of the role of oral transmission of the text, the lack of an explanation for the emergence of the textus receptus of the Quran, and the unclear (selective) and normative definition of “obscure” passages in the Quran.

7.6 Case Study: Philological Analysis of a Judeo-Arabic Document The philological approach can be devoted to individual realms of grammar and lexicon, as in the preceding cases. In the case of documents in a nonstandard variety of Arabic, like Judeo-Arabic, it may be interesting to apply the philological approach wholesale to all of the realms previously treated individually and to attempt to come up with a “mini grammar” of the respective text. The following remarks are based on a recent study by the author on an Iraqi Judeo-Arabic version of the Joseph story, representing the qiṣaṣ al-ʾanbiyāʾ genre (Ben Porat 1924. cf. Edzard, 2012), and cover features of phonology, morphology, syntax, lexicon, and style. The text has many typical features of Iraqi Arabic, in some cases specifically of the “qəltu” dialect, as exemplified in the following. The Hebrew orthography of the text under investigation is phonetic as a tendency, especially as far as the use of vowels is concerned. Due to the loss of the dental fricatives in urban dialects, as also happened in the Baghdad region, the only occasional and unsystematic marking of the dental fricatives /ḏ/ and /ṯ/ is expected. The same holds for the dialectal merger of classical /ḍ/ and /ẓ/. The text off ers rich documentation of the phenomena of tafxīm and tarqīq. Examples of the former process, that is, suprasegmental assimilation with respect to velarization, or advanced tongue root (ATR) include forms like 〈ṭmṣw〉, from ṭamasū “they effaced.” One of the shibboleths of Iraqi Arabic dialects is the indefinite article (or indeterminacy marker) fadd or fard, which can also stand before nouns in the dual or in the plural (e.g., Blanc 1964: 118), for example, in 〈lʾgh frd ʿrby yswq frd nʾgh〉 “he met a Bedouin leading a camel.” Some verb forms found in the text may be interpreted as being clearly indicative of the Jewish Baghdadi variety, such as 〈gbtw〉 “I brought” or 〈nʾdytwk〉 “I called you for help,” where the final -u (i.e., of –tu) is spelled out (cf. Blanc 1964: 107). The hypercorrect use of lam wiThthe suffix conjugation (“perfect”) is noteworthy in the text. Examples include 〈wlm frstw frysh〉 “and I did not became prey.” This example also shows the combination of a hypocorrect construction with a “high” stylistic figure, that is, paronomasia (or figura etymologica— mafʿūl muṭlaq). A further feature is the marking of the definite direct (accusative) object with either li- or hypercorrect ʾilā, such as 〈yḥb ʾl ywsp〉 “he loves Joseph” (e.g., Hary 1991). Definiteness on noun–adjective phrases is usually marked only at the qualifying adjective, for example, 〈tʾg′r ʾl ʿrb〉 “the Arab (Bedouin) merchant.” The Judeo-Arabic text of the Joseph story contains a fair number of colloquial words, some of which are common to the Arabic koiné and some of which are specific to either Jewish or Muslim Baghdadi Arabic or even shared in boThvarieties. The verb 〈 šʾp〉 is the most obvious example of the first category. Likewise, 〈pʾt〉 is used instead of Classical daxala “to enter.” Another shibboleThof Iraqi Arabic, namely, the existence marker aku, is also found occasionally in the text, for example, 〈yʾ ʿrby ʾkw šgrh ʾṣlhʾ tʾbt〉 “O Bedouin, there is a tree with a firm trunk.”

The Philological Approach to Arabic Grammar Stylistically, the oral character of this Judeo-Arabic version of the Joseph story can clearly be associated with the storyteller or ḥakawātī genre. There are several stylistic allusions to the fairy tale genre, as becomes evident, for instance, in the following formulation 〈ywm mn ʾl ʾyyʾ m wsʾʿh mn ʾl zmʾn šʾp ywsp ḥlm〉 “on one day (‘eines sch önen Tages’) at a certain hour Joseph had (‘saw’) a dream.” It is also remarkable that this Judeo-Arabic text destined for a Jewish audience sets out with the basmala and in this way establishes a quasi-Muslim setting (on this issue, cf. Cohen 2007: 20): (20) Example of Muslim textual elements in Jewish context 〈bʾsm ʾllh ʾl rḥʾn ʾl rḥym ʾyyʾhw nʿbdhw wbyh nstʿyn ʿlʾ qwm ʾl ṣ′ʾ lmyn〉 “In the name of God the merciful the Compassionate; Him we venerate and to Him we seek refuge from the evil people”

7.7 Conclusion In principle, the concepts linguistic and philological cannot and should not be separated in the first place. In practice, these two approaches at times have been drifting apart in recent research. Aoun et al. (2010), for instance, provide an exclusively formal linguistic study, without any philological ingredient. However, there is still room for carefully combining insight presented by the native Arab grammarians and modern linguistic theory on one side and insight gained by philological text-based research on the other. Indeed, only the empirical approach based on large diachronic and synchronic data is apt to yield results as presented, say, in the work of Joshua Blau, Werner Diem, Simon Hopkins, Geoffrey Khan, Pierre Larcher, Manfred Ullmann, and Stefan Wild.

References Primary Sources Ben Porat, Josef. 1924. Qiṣṣat Yosef ha-ṣadiq ʿalav ha- šalom. Baghdad: Maṭba ʿat al-Waṭanīya al-Yisrāʾilīya. Ibn Manẓūr, Muḥammad ibn Mukarram. Lisān al-ʾarab Beirut: Dār ʾiḥyāʾ at-turāṯ al-ʿarabā. 1988. Ibn Sīnā. al-Qānūn fī ṭ-ṭibb. Bulaq. 1877. Sībawayhi. Kitāb Sībawayhi. Ed. Muḥammad ʿAbd al-Salām Hārūn. 5 vols. Cairo, 1968–1977; 2nd ed. Cairo. 1977. az-Zabīdī, Muḥammad Murtaḍā. Tāǧ al-ʿarūs min ǧawāhir al-Qāmūs. Kuwait: Maṭba ʿat ḥukūmat al-Kuwayt. 1965. az-Zamaxšarī. Kitāb al-Mufaṣṣal li-z-Zamaxšarī. Opus de re grammatica arabicum. Ed. J.P. Broch. Christianiae: Libraria P. T. Mallingii. 1879.

Secondary Sources Al-Nassir, A. 1993. Sībawaih the phonologist: A critical study of the phonetic and phonological theory of Sībawaih in his treatise “Al-Kit āb.” London: Kegan Paul International. Aoun, Joseph E., Elabbas Benmamoun, and Lina Choueiri. 2010. The syntax of Arabic. Cambridge, UK: Cambridge University Press. Badawi, Elsaid, Michael G. Carter, and Adrian Gully. 2004. Modern written Arabic: A comprehensive grammar. London: Routledge. Bergsträsser, Gotthelf. 1914. Verneinungs- und Fragepartikeln und Verwandtes im Ḳurʾān. Ein Beitrag zur historischen Grammatik des Arabischen. Leipzig: J.C. Hinrichs’sche Buchhandlung. Birnstiel, Daniel. Forthcoming. Selected features of Arabic syntax in the Quran. PhD diss., University of Cambridge. Blanc, Haim. 1964. Communal dialects in Baghdad. Cambridge, MA: Harvard University Press.

The Philological Approach to Arabic Grammar Blau, Joshua. 1965. The emergence and linguistic background of Judaeo-Arabic: A study of the origins of Middle Arabic. Oxford: Oxford University Press. Bobzin, Hartmut. 1992. Geschichte der Arabischen Philologie in Europa bis zum Ausgang des achtzehnten Jahrhunderts. In Grundriß der arabischen Philologie. Band III: Supplement, ed. Wolfdietrich Fischer, 155–187. Wiesbaden: Dr. Ludwig Reichert Verlag. ——. 1995. Der Koran im Zeitalter der Reformation. Studien zur Frühgeschichte der Arabistik und Islamkunde in Europa. Stuttgart: Steiner. Brünnow, Rudolf-Ernst and August Fischer. [1895] 20088 . Chrestomathy of Classical Arabic prose literature. 8th rev. ed. Ed. Lutz Edzard and Amund Bjørsnøs. Wiesbaden: Harrassowitz. Carter, Michael G. 1990. Arabic lexicography. In Cambridge history of Arabic literature: Religion, learning and science in the ʿAbbasid Period, ed. M. J. L. Young et al., 106–117. Cambridge, UK: Cambridge University Press. Cohen, Mark R. 2007. On the interplay of Arabic and Hebrew in the Cairo Geniza letters. In Studies in Hebrew and Arabic letters in honour of Raymond P. Scheindlin, ed. Jonathan P. Decter and Michael Rand, 17–35. Piscataway, NJ: Gorgias Press. Dahlgren, Sven-Olof. 2006. Sentential negation in Arabic. In Current issues in the analysis of Semitic grammar and lexicon II, ed. Lutz Edzard and Jan Retsö, 64–78. Wiesbaden: Harrassowitz. Diem, Werner. 1986. Alienable and inalienable possession in Semitic ZDMG 136: 227–291. Diem, Werner. 1998. Fa-waylun li-l-qāsiyati qulūbuhum. Studien zum arabischen adjektivischen Satz. Wiesbaden: Harrassowitz Verlag. ——. 2002. Translokative Verben im Arabischen. Eine diachronische Studie. Wiesbaden: Harrassowitz. ——. 2006. Philologisches zu den arabischen Papyri der Hamburger Staats- und Universitäts-Bibliothek. ZAL 45: 7– 54. ——. 2011. Arabisch kayfa “wie” als Konjunktion. Ein Beitrag zur Geschichte der arabischen Syntax. Wiesbaden: Harrassowitz. Dozy, Reinhard. 1927. Supplément aux dictionnaires arabes. 2 vols. Leiden: Brill. Edzard, Lutz. 2001a. Grammatical systems in indigenous and in foreign perspective: The case of Arabic. In Indigenous grammar across cultures, ed. Hannes Kniffka, 317–345. Frankfurt a.M.: Peter Lang. ——. 2001b. Sībawayhi’s observations on assimilatory processes and re-syllabification in the light of optimality theory. Journal of Arabic and Islamic Studies 3 (2000): 48–65. ——. 2007. Noun. In Encyclopedia of Arabic language and linguistics, Vol. 2, ed. Kees Versteegh, Associate Editors: Mushira Eid, Alaa Elgibali, Manfred Woidich, Andrzej Zaborski, 422– 428. Leiden: Brill. ——. 2011. Die ʾiḍāfa ġayr ḥaqīqīya, der naʿt sababī und die enallagé (hypallagé) adiectivi: ein typologischer Vergleich. In Orientalistische Studien zu Sprache und Literatur. Festgabe zum 65. Geburtstag von Werner Diem, ed. Ulrich Marzolph, 9–20. Wiesbaden: Harrassowitz. ——. 2012. Linguistic and cultural features of an Iraqi Judeo-Arabic text of the qiṣaṣ al-ʾanbiyāʾ genre. In Mixed Arabic: Diachrony and synchrony, ed. LiesbeTh Zack, 83–94. Leiden: Brill. Endress, Gerhard. 1992. Die wissenschaftliche Literatur. In Grundriß der Arabischen Philologie. Band III: Supplement, ed. Wolfdietrich Fischer, 3–152. Wiesbaden: Dr. Ludwig Reichert Verlag. Fischer, Wolfdietrich (ed.). 1982. Grundriß der Arabischen Philologie. Band I: Sprachwissenschaft. Wiesbaden: Dr. Ludwig Reichert Verlag. ——. 1992. Grundriß der Arabischen Philologie. Band III: Supplement. Wiesbaden: Dr. Ludwig Reichert Verlag.

The Philological Approach to Arabic Grammar Fleisch, Henri. 1958a. La conception phonétique des Arabes d’après le Sirr Ṣināʿ at al-Iʿ rāb d’Ibn Ǧinnī. ZDMG 108: 74–105. Fleisch, Henri. 1958b. Maǧhūra, mahmūsa: examen critique. Mélanges de l’Université Saint-Joseph 35:193–210. Fück, Johann. 1950. Arabiya. Untersuchungen zur arabischen Sprach- und Stilgeschichte. Berlin: Akademie-Verlag. ——. 1955. Die arabischen Studien in Europa bis in den Anfang des 20. Jahrhunderts. Leipzig: O. Harrassowitz. Gätje, Helmut(ed.). 1987. Grundriß der Arabischen Philologie. Band II: Literaturwissenschaft. Wiesbaden: Dr. Ludwig Reichert Verlag. Gensler, Orin. 2003. Object ordering in verbs marking two pronominal objects: Non-explanation and explanation. Linguistic Typology 7: 187–231. Greenberg, Joseph H. 1966. Some universals of grammar with particular reference to the order of meaningful elements. In Universals of language ed. Joseph Greenberg, 577–587. Cambridge, MA: MIT Press. Gruendler, Beatrice. 2006. Arabic alphabet: Origin. In EALL, I. ed. Kees Versteegh et al., 148–165. Guillaume, Alfred. 1955. The life of Muḥammad: A translation of Isḥāq’a Sīrat rasūl allāh. London: Oxford University Press. Hamilton, Alastair. 2006. Arabic studies in Europe. In Encyclopedia of Arabic language and linguistics, Vol. 1, ed. Kees Versteegh, associate ed. Mushira Eid, Alaa Elgibali, Manfred Woidich, Andrzej Zaborski, 166–172. Leiden: Brill. Hary, Benjamin. 1991. On the use of ʾilā and li in Judeo-Arabic texts. In Semitic studies in honor of Wolf Leslau on the occasion of his eighty-fifth birthday, ed. Alan Kaye, 595–608. Wiesbaden: Harrassowitz. Holes, Clives. 1995. Modern Arabic: Structures, functions and varieties. London: Longman. Hopkins, Simon. 1984. Studies in the grammar of Early Arabic based upon papyri datable to before 300 A.H./912 A.D. Oxford: Oxford University Press. Khan, Geoffrey. 2011. Vocalised Judeo-Arabic manuscripts in the Cairo Geniza. In From a sacred source: Genizah studies in honour of Stefan C. Reif, ed. B. Outhwaite and S. Bhayro, 201–218. Leiden: Brill. Lane, Edward. 1863. An Arabic-English lexicon: Derived from the best and the most copious Eastern sources, Book I, Part 1. London: Williams and Norgate. Larcher, Pierre. 1994. mā faʿ ala vs. lam yafʿ al: une hypothèse pragmatique. Arabica 41: 388–415. Luxenberg, Christoph (pen name). 2000. Die syro-aramäische Lesart des Koran: Ein Beitrag zur Entschlüsselung der Koransprache. Berlin: Das Arabische Buch. McCarthy, J. 1991. Guttural phonology. In Perspectives on Arabic linguistics III: Papers from the third annual symposium on Arabic linguistics, ed. Bernard Comrie and Mushira Eid, 63–92. Amsterdam: Benjamins. Nekroumi, Mohammed. 2003. Interrogation, polarité et argumentation: Vers une théorie structurale et énonciative de la modalité en arabe classique. Schenefeld: EB-Verlag. Nöldeke, Theodor. [1896] 20064 . Grammatik des Klassischen Arabisch. 4th rev. ed. Ed. Wolfdietrich Fischer. Wiesbaden: Harrassowitz. Peled, Yishai. 1992. Conditional structures in Classical Arabic. Wiesbaden: Harrassowitz. ——. 2009. Sentence types and word order patterns: Medieval and modern perspectives. Leiden: Brill. Reckendorf, Herrmann. 1898. Die syntaktischen Verhältnisse des Arabischen. Leiden: Brill. Retsö, Jan. 2003. The Arabs in antiquity: Their history from the Assyrians to the Umayyads. London: Routledge.

The Philological Approach to Arabic Grammar Rotter, Gernot. 1976. Ibn Ishâq. Das Leben des Propheten. Tübingen: Horst Erdmann Verlag. Rubin, Aaron D. 2005. Studies in Semitic grammaticalization. Winona Lake, IN: Eisenbrauns. Sacy, Sylvestre de 1831. Grammaire arabe à l’usage des élèves de l’École spéciale des langues orientales vivantes. Paris: Imprimerie royale. Schoeler, Gregor. 2010. The biography of Muḥammad: Nature and authenticity, ed. James Montgomery. Trans. U. Vagelpohl. London: Routledge. Ullmann, Manfred. 1970–. Wörterbuch der klassischen arabischen Sprache. Wiesbaden: Harrassowitz. ——. 1976. Der Werwolf. Ein griechisches Sagenmotiv in arabischer Verkleidung. Wiener Zeitschrift für die Kunde des Morgenlandes 68: 171–184. ——. 1988. Sätze mit lau. Beiträge zur Lexikographie des Klassischen Arabisch 14. Munich: Bayerische Akademie der Wissenschaft en. Versteegh, Cornelis H. M. 1977. Greek elements in Arabic linguistic thinking. Leiden: Brill. Versteegh, Kees. [1997] 20012 . The Arabic language. Edinburgh: Edinburgh University Press. Vollers, Karl. 1906. Volkssprache und Schriftsprache im alten Arabien. Strassburg: Trübner. Wehr, Hans. 1952. Review of Fück 1950. ZDMG 102: 179–186. ——. 1953. Zur Funktion arabischer Negationen. ZDMG 193: 27–59. ——. [1961] 1994. A dictionary of modern written Arabic, 4th ed. Ed. J. Milton Cowan. Wiesbaden: Harrassowitz. Weil, Gustav. 1864. Das Leben Mohammed’s nach Mohammed Ibn Ishâk bearbeitet von Abd el-Malik Ibn Hischam. Stuttgart: Verlag der J. B. Metzler’schen Buchhandlung. Wild, Stefan. 2010. Lost in philology? The virgins of paradise and the Luxenberg hypothesis. In The Qurʾ ān in context: Historical and literary investigations into the Qurʾ ānic milieu, ed. Angelika Neuwirth, Nicolai Sinai, and Michael Marx, 625–647. Leiden: Brill. Willmsen, David. 2010. Dialects of written Arabic: Syntactic differences in the treatment of object pronouns in Egyptian and Levantine newspapers. Arabica 57: 99–128. Lutz Edzard Lutz Edzard, University of Oslo

Arabic Linguistic Tradition II: Pragmatics Pierre Larcher The Oxford Handbook of Arabic Linguistics Edited by Jonathan Owens Print Publication Date: Sep 2013 Online Publication Date: Dec 2013

Subject: Linguistics, Pragmatics, Languages by Region DOI: 10.1093/oxfordhb/9780199764136.013.0008

Abstract and Keywords This article deals essentially with two topics. The first is rhetoric, as one of the two sectors of the basic core of the Arabic linguistic tradition. Since the tradition was not definitively constructed until the postclassical period, Qazwīnī’s Talkhīs (d. 739/1338) is used—the most famous “epitome” of the rhetorical part of Sakkākī’s Miftāħ al-‘Ulūm, which itself is based on the two works of Abd al-Qāhir al-Jurjānī (d. 471/1078), Asrār al-‘Arabiyya and Dalā’il al-’I‘jāz. The second is the intersections of rhetoric with the other sectors of this tradition: linguistics proper, namely, grammar; and not linguistics proper, namely, the theologico-juridical sciences. Keywords: rhetoric, Arabic, linguistics, Talkhi, Qazwini, grammar, theologico-juridical sciences

8.1 Introduction: From the Grammatical Tradition to the Linguistic Tradition THE idea that grammar is nothing more than one of the linguistic disciplines is found in the Arabic tradition itself. We

can cite at least two works in evidence. The first is the Miftāḥ al-ʿulūm (The key to the sciences) by Sakkākī (d. 626/1229). This work is divided into three parts. The first part is dedicated to morphology (ʿilm al-ṣarf), the second to syntax (ʿilm al-naḥw), and the third to the “two sciences of meaning and of expression” (ʿilmā al-maʿānī wa-lbayān), or rhetoric (Miftāḥ, 3:25–27). A little earlier, however, Sakkākī had presented syntax as having as its “complement” (tamām) rhetoric; the first part of rhetoric as having as its “complement” the “two sciences of definition and of argumentation” (ʿilmā al-ḥadd wa-l-istidlāl), in other words logic. Rhetoric in turn was presented as bipartite, to the extent that it is used and practiced in the two arts of prose (nathr) and versification (naẓm), as needing, for the latter, the “two sciences of prosody and of rhymes” (ʿilmā al-ʿarūḍ wa-l-qawāfī), in other words poetics (ʿilm al-shiʿr). Sakkākī indicates that he deals with all of this because it is a matter “of several species of belles-lettres” (ʿiddatʾ anwāʿ al-ʾ adab), “taking each one from the other” (mutaʿākhidha), in other words forming a coherent whole. He explicitly excludes lexicography from his structure, however (ʿilm al-lugha) (Miftāḥ, 2:20–21; 3:1–12).

Arabic Linguistic Tradition II

Figure 8.1 The science of language.

The contents of the Miftāḥ would thus be better described as a veritable encyclopedia of the sciences of language combining grammar, rhetoric, logic, and poetics, with each part presented as a binary structure: the phrase ʿilmā lṣarf wa-l-naḥw appears at 3:33–34; it is parallel, as seen already, to those Sakkākī uses for the other three disciplines. There is no contradiction between these two perspectives. The tripartite presentation is foreshadowed by the quadripartite presentation, in the sense that a distinction is made by the latter between what might be called the basic core and the expansions. The basic core comprises grammar and rhetoric, because these are the only disciplines that have expansions. To be sure, rhetoric is presented as an expansion of one of the two parts of grammar, but, in turn, it has for expansions logic, for the first of its two parts, and poetics for its two parts. The complex relationships among the various parts of this whole can be represented as shown in Figure 8.1. The second work is the Muqaddima of Ibn Khaldūn (d. 808/1406). As its name indicates, it is the “introduction” to his great work of history, the Kitāb al-ʿibar (Book of lessons), Book I of the tome. The Muqaddima itself is divided into six chapters (bāb), which are subdivided into sections (faṣl). The sixth chapter is a veritable encyclopedia of the sciences, which contains a chapter entitled fī ʿulūm al-lisān al-ʿarabī (On the sciences of the Arabic language, no. 45) (Muqaddima, 1055–1070). Ibn Khaldūn enumerates four of them, in this order: grammar (ʿilm al-naḥw); the science of the lexicon (ʿilm al-lugha); rhetoric (ʿilm al-bayān); and the science of belles-lettres (ʿilm al-ʾadab). Comparison between the two works is very instructive. First, the two structures have two disciplines in common, grammar and rhetoric, even though the two authors use different terminologies. Sakkākī uses naḥw in opposition to ṣarf to refer specifically to syntax. In contrast, Ibn Khaldūn uses naḥw in the general sense of grammar. Likewise, Sakkākī uses bayān as the name of one of the two parts of rhetoric, while Ibn Khaldūn uses it to name the entirety. But the presence of these two disciplines within the two structures confirms that they did indeed constitute the “basic core” of the Arabic linguistic tradition [Baalbaki, “ALT I”]. Second, Sakkākī excludes from his structure the ʿilm al-lugha that Ibn Khaldūn, conversely, does include. Even though neither of the two authors explains his choice to include or exclude it, the reason can be inferred on the basis of a remarkable passage in the Sharḥ al-Kāfiya (I:5) by the grammarian Raḍī al-dīn al-Astarābadhī (d. 688/1289) (Larcher 2000). There he defines the specific “objects” of the ʿilm al-lugha, the ʿilm al-ṣarf(taṣrīf, as he calls it), and the ʿilm al-naḥw. He does so through the concept of waḍʿ, which is inherited from falsafa (Greek thesis, Latin impositio, modern “institution”) and which itself in the postclassical period will become the object of its own discipline, the ʿilm al-waḍʿ (Weiss 1976). For Astarābadhī, these objects were instituted either as “determinate expressions” (ʿalfāẓ muʿayyana), which he describes as samāʿiyya (lit. relating to samāʿ “hearing”), dealt with in the ʿilm al-lugha, or a “general rule” (qānūn kullī), through which one knows qiyāsiyya expressions (lit. relating to qiyās or “measure”). If the latter term is translated as “regular,” then samāʿī will be “irregular.” Regular expressions can be simple (mufradāt) or complex (murakkabāt). Regular simple expressions and some regular complex expressions (e.g., the relative adjective or the imperfect verb) relate to the ʿilm al-taṣrīf: the relative adjective and the imperfect verb are formally complex, comprising a stem with a suffix in the first case and a prefix

Arabic Linguistic Tradition II in the second, but do not function any less as units [Ratcliffe, “Morphology”]. The other regular complex expressions (phrases and sentences) relate to the ʿilm al-naḥw. All this can easily be translated into contemporary linguistic terms. A language on the whole comprises a grammar on one hand and a lexicon on the other. Grammar in turn includes a phonology, a morphology (ṣarf/taṣrīf including both phonology and morphology), and a syntax. The ṣarf/taṣrīf and the naḥw deal with the entirety of grammar, so it is only logical that a specific component deals with the lexicon. But the lexicon of a language consists of two parts: a regular part and an irregular part. Of course the regular part of the lexicon—that is, the rules of formation and interpretation of words—is in fact already treated in the ṣarf and, more specifically, the ishtiqāq (“derivation”). Thus, what is left over is the irregular part of the lexicon, which is likely to interest the lexicographer but not someone so concerned about systematization as Sakkākī. Third and last, Sakkākī considers the four disciplines he deals with to relate to an entirety that is the ʾadab (Miftāḥ, 3:12), though Ibn Khaldūn, conversely, considers the ʾadab to be a part of the entirety! This is explained by the difference in perspective of the two authors. Sakkākī is known as the systematizer of rhetoric (Smyth 1995). This is no accident. Rhetoric is central in his structure; it is the only discipline that is an expansion of another discipline, grammar, and that has, with logic and poetics, its own expansions. By including the various disciplines he deals with in an entirety relating to ʾadab, Sakkākī marks out his perspective as essentially literary. Ibn Khaldūnʾs is quite different. He is known as a historian, but his profession was magistrate (qādī). He displays this perspective immediately after naming the four “pillars” (‘arkān) of the “sciences of the Arabic language” (Muqaddima, 1055): Knowledge of them all is necessary for religious scholars, since the source of all religious laws is the Qurʾān and the Sunnah, which are in Arabic. Their transmitters, the men around Muḥammad and the men of the second generation, were Arabs. Their difficulties are to be explained from the language they used. Thus, those who want to be religious scholars must know the sciences connected with the Arabic language. (trans. Rosenthal, abr. ed. 433)

Ibn Khaldūnʾs perspective is clearly hermeneutic. Comparing the two works will thus remind us that the Arabic linguistic tradition has two aspects: one literary and the other hermeneutic. On its hermeneutical side, it thus intersects with the religious (i.e., theologico-juridical) sciences (e.g., fiqh, ʾuṣūl al-fiqh, tafsīr, kalām). This chapter deals essentially with two topics: 1) Rhetoric, as one of the two sectors of the basic core of the Arabic linguistic tradition (for overviews, see Heinrichs 1987, 1998; Halldén 2006; Larcher 2009). Since the tradition was not definitively constructed until the postclassical period, I use the Talkhīṣ of Qazwīnī (d. 739/1338), the most famous “epitome” of the rhetorical part of the Miftāḥ of Sakkākī, which itself is based on the two works of ʿAbd al-Qāhir al-Jurjānī (d. 471/1078), Asrār al-ʿarabiyya and Dalāʾil al-Iʿjāz. The Talkhīṣ owes its fame to the fact that it is a textbook (Smyth 1993), the object of many commentaries and supercommentaries, some of them collected in the Shurūḥ al-Talkhīṣ and widely used, until quite recently, in teaching and also used by Western scholars (Mehren 1853; Jenssen 1998). 2) Given the central nature of rhetoric, its intersections with the other parts of this tradition: one of them linguistics proper (i.e., grammar); the other not linguistics proper (i.e., the theologico-juridical sciences).

8.2 Rhetoric

8.2.1 Structural and Terminological Uncertainty Qazwīnī’s Talkhīṣ calls rhetoric ʿilm al-balāgha and divides it into three parts: ʿilm al-maʿānī, ʿilm al-bayān, and ʿilm al-badīʿ. After briefly presenting the subject of each, he concludes (Talkhīṣ, 36–37): “but many call the entirety ilm al-bayān, whereas some call the first ʿilm al-maʿānī and the other two ʿilm al-bayān and all three ʿilm al-badīʿ” . The first set of terms is, for example, that of Ibn Khaldūn, who gives the name ʿilm al-bayān to both the entirety and the second part of

Arabic Linguistic Tradition II the entirety and the name ʿilm al-balāgha to the first part. It is also Ḍiyāʾ al-Dīn Ibn al-Athīrʾs (d. 637/1239). His work al-Mathal al-sāʾir gives the name ʿilm al-bayān to the entirety, dividing the content into an “introduction” and two “discourses” (maqāla), dealing, respectively, with “technique relating to the lafẓ” (ṣināʿa lafẓiyya) and with “technique relating to the meaning” (ṣināʿa maʿnawiyya). The second set of terms is Sakkākī’s (ʿilmā al-maʿ ānī wa-l-bayān), although he does not give a name to the entirety. I do not know whether the name ʿilm al-badīʿ was ever given to the entirety. But if it was, this might be compared with Ibn Khaldūnʾs remark that some rhetoricians, basically Maghrebi, favored this part, his example being the ʿUmda of Ibn Rashīq (d. 456/1063–1064) (Muqaddima, 1068). I will follow the division and terminology of Qazwīnī, always keeping in mind the structural and terminological uncertainties of the field in the Arabic tradition.

8.2.2 ʿIlm al-Maʿānī The ʿilm al-maʿānī or “science of meanings” in its very name identifies itself as a semantics. In the definition given of it, however, the word “maʿnā,” of which maʿānī is the plural, does not appear, but rather the word paired with it in the Arabic tradition (Talkhīs, 37): “it is a science by which the states of Arabic expression become known, appropriate to the needs of the situation” . Other than lafẓ, the important word here is hāl, which appears twice, once in the singular and once in the plural. Such a definition posits the existence of a correlation between the variation (the “states”) in the expression and the situation (the “state”). It is a semantics, the point of view adopted being semasiological (i.e., going from the expression to the maʿnā), and, more specifically, a contextual semantics. Further elements of an answer can be drawn from the examination of the sections that make up the ʿilm al-maʿānī and the justification for this division (Talkhīṣ, 37–38). It is divided into eight sections: (I) states of the assertive predication; (II) states of the “support”; (III) states of the “supported” (lit. “that which is leaned”); (IV) states of the complements of the verb; (V) restriction; (VI) performative; (VII) conjunction and disjunction; and (VIII) concision, prolixity, and equilibrium. This division is justified by a set of six propositions. The first proposition is as follows: “The utterance, in fact, is either statement or performative, because if its relationship has a referent, to which it is appropriate or not, it is a statement and, if not, a performative.”

The first proposition justifies sections I and VI and suggests that this semantics is primarily a semantics of the utterance. It repeats in fact the classification of utterances into khabar and ʾinshāʾ (Larcher, 1980, 1991), which was established conclusively in the 7th/13th century but which represents the outcome of a long tradition, the maʿānī al-kalām “meanings of the utterance,” to use the title of a chapter of Ibn Fārisʾ Ṣāḥibī (d. 395/1004) (Frank 1981; Buburuzan 1995; Versteegh 2004). If the khabar is defined positively as a referential utterance, in other words an assertion, the ʾinshā’ is here defined simply in a negative way, as a nonreferential utterance. But its name (lit. “creation”) and examination of the sources show that it can be defined positively as a self-referential utterance, identifiable both by extension and by intension with Austinʾs (1962) category of performative (vs. constative). The ʾinshāʾ is at first added to a preexisting classification of utterances into khabar “statement” and ṭalab “request, demand” (= Fr. “jussion” and related adjective “jussive”) designating only juridical performatives (siyagh al-ʿuqūd wa-l-fusūkh “contractual and renunciative formulas”). Then, at a later time, ṭalab was subsumed under ʾinshāʾ. A trace of this history remains in the commentators on Sakkākī. Sakkākī knows only a classification of utterances into khabar and ḍalab. His commentators adopt the new classification into khabar and ʾinshāʾ but subdivide the latter into ḍalabī “requesting, demanding” and ghayr ḍalabī “not requesting, demanding.” If the ‘inshāʾ ḍalabī has the same extension as Sakkākīʾs ḍalab, the ʾinshāʾ ghayr ḍalabī includes, along with the juridical performatives, sometimes called ʾīqāʿī “operatives,” all utterances that are neither assertions nor requests, such as the exclamations (see Section 8.3.5). The second proposition is, “The statement requires a support, a supported, and a predication.”

This alludes to the fact that every utterance, whether statement or not, is a clause (jumla) and every clause a set (literal meaning of the word jumla in Arabic) of two elements (juzʾ) linked by a relationship (nisba) of predication

Arabic Linguistic Tradition II (ʾisnād). Because ʾisnād is the maṣdar of a verb ʾasnada construed with two objects, a direct one and an indirect one introduced by the preposition ʾilā, these two elements are called in Arabic musnad and musnad ʾilayhi and should logically be called predicate and subject. These two terms, however, hardly suit the verb (= fiʿl) and the subject (= fāʿil) of the verb, corresponding to musnad and musnad ilayhi, of the verbal clause, which is a linked clause, in the sense of the Swiss linguist Charles Bally ((1865–1947) (Bally 1965)), where the verb governs its arguments. They are even less appropriate for the terms “topic” (mubtadaʾ) and “comment” (khabar), respectively, the musnad ʾilayhi and musnad of the nominal clause, which is a segmented clause, in the sense of Bally (ibid.). The verb ʾasnada literally means “to lean s.t. on” (ʾilā); the derived passive participles musnad “supported” and musnad ʾilayhi “leaned on” designate the two obligatory parts of a predication, which I will translate as “supported” and “support” (Guillaume 1986; Larcher 2000). This second proposition justifies II and III and shows that this semantics concerns not only the utterance but also its constituents, in their major species. The third proposition states: “The supported can have complements, if it is a verb or an element having the meaning of one.”

This proposition justifies IV and shows that this semantics concerns the minor as well as the major constituents. The fourth proposition is as follows: “Each of the two relationships, predicative and verbal complements, can be made with or without restriction.”

This justifies V. Even though the restriction is presented as bearing on the constituents, whether major or minor, one finds here no less the utterance and even the semantically complex utterance. The fifth proposition states: “Every clause is connected to another, whether coordinated with it or not.”

This proposition justifies VII and shows that this semantics also goes beyond the utterance, concerning the way one clause links with another, in other words, the formally complex utterance or discourse. The sixth proposition is as follows: “The efficient utterance either considerably exceeds what is fundamentally intended, otherwise not.”

If it is too much, there is “prolixity” (ʾiḍnāb); if there is too little, there is “concision” (ʾījāz). If there is no excess, it is “equilibrium” (musāwāt). In this sixth and last proposition, which justifies VIII, the term balīgh appears. This is the adjective corresponding to the verb balugha, whose verbal noun, balāgha, is, within the field of ʿilm, the term for rhetoric in Qazwīnī. Balugha perhaps understood as the stative–resultative voice of a verb of which balagha “arrive, reach” is the active voice. A balīgh discourse is thus a discourse that has achieved its purpose; in other words, it is efficient. The balāgha presupposes faṣāḥa, but not vice versa. Faṣāḥa is the verbal noun of the stative verb faṣuḥa, to which the adjective faṣīḥ corresponds. The elative ʾafṣaḥ that appears in the Quran (28:34) clearly designates the faṣāḥa as Aaronʾs fluency of speech as opposed to Mosesʾ speech impediment. Something of the Quranic usage remains in rhetoric, where the term might be translated as “eloquence.” Before postclassical rhetoric finally got rid of the two terms, faṣāḥa also appears in the titles of works of rhetoric, the most famous being the Sirr al-faṣāḥa of al-Khafājī (d. 466/1074). Outside of rhetoric, faṣāḥa can be used in the sense of “grammatical correctness,” focusing on the matter of case and mood suffixes ʾ(iʿrāb) (on balāgha and faṣāḥa, see Ghersetti 1998). While centering on the utterance, this semantics sometimes deals with sub-utterance elements, taking an interest in its constituents, both major and minor, and sometimes goes beyond it, taking an interest in the connection of utterances among themselves. It is thus not possible to find in the rank of the expression (i.e., the utterance) the characteristics of this semantics. The following sections go into more detail.

Arabic Linguistic Tradition II 8.2.2.1 The Khabar This section begins with a distinction between two uses of a statement: fāʾidat al-khabar (“information provided by a statement”); and lāzim fā ʾidat al-khabar (“what it implies”). The first occurs when a speaker (al-mukhbir) wants to make the hearer know about a state of affairs (al-ḥukm), and the second happens when the speaker wants to make the hearer aware of what the speaker knows (Talkhīṣ, 40–41). Qazwīnī, unfortunately, gives no examples. Sakkākī (Miftāḥ, 72) gives, as an example of the first, Zaydun ʿālimun (“Zayd is a scholar”), when it is said to someone who was not aware of the fact, and, as an example of the second, qad ḥafiẓta al-Tawrāh (“you know the Bible by heart”), when it is said to someone who knows the Bible by heart. It is followed (Talkhīṣ, 41–42) by a second distinction, so famous that in Ibn Khaldū n (Muqaddima, 1065) it becomes one of the symbols of the ʿilm al-maʿānī. This is a distinction of three types of statements: ibtidāʾī (lit. “initial”); ṭalabī (lit. “requesting, demanding”); and ʾinkārī (lit. “denying”). The first is addressed to someone who does not have an idea (khālī al-dhihn, lit. “empty mind”) of the content of the statement. Its name comes from the fact that, purely informative, it is found at the beginning of the discourse. The second is addressed to someone with an attitude of hesitation or questioning with respect to the content. The last is addressed to someone with an attitude of denial with respect to the content. They thus take their names from what constitute reactions (Simon, 1993) to the attitude of the hearer and can take place only in dialogue. Ibn Khaldū n gives grammatical examples: Zaydun qāʾimun (“Zayd is standing”) for the first; ʾinna Zaydan qāʾimun (“Yes, Zayd is standing”) for the second; ʾinna Zaydan la-qāʾimun (“Yes, Zayd really is standing”) for the third. Simply comparing these two paragraphs allows us to further our understanding of what ʿilm al-maʿānī is. In the first case, the meaning of the statement depends not only on the intention (qaṣd) of the speaker, as Qazwīnī has it, but also on what the hearer does or does not know, as noted by Sakkākī. In other words, the calculation of the meaning of the utterance depends narrowly on the situation of utterance and, more specifically, on the participants. This alone suffices to call this semantics a pragmatics, in the sense of the American semiotician Charles Morris (1901–1979). In the second case, the calculation of the meaning is no less pragmatic in nature, but it relies on objective markers that this situation of utterance leaves in the utterance: the reinforcement ʾinna in the second example; and the reinforcements ʾinna and la- in the third. It is this last case that confirms the definition given of ʿilm al-maʿānī, a variation in the expression as a function of the situation of utterance, and explains the interpretation that has been given of maʿānī as maʿānī al-naḥw: a “semantics of syntax” (EI2 , s.v. al-maʿāni wa-l-bayān). The expression maʿānī al-naḥw appears elsewhere, if not in the Talkhīṣ, at least in the other epitome of the Miftāḥ made by Qazwīnī, the ʾīḍāḥ (in Shurūh al-Talkhīṣ, vol. 1:132). It originated in the Dalāʾil of Jurjānī (Heinrichs 1987, Ghersetti 2002). It suffices to read Qazwīnī further to see that this interpretation is reductive. He gives not a grammarianʾs example but a Quranic one (36:14–16), that of the two envoys (identified by the Islamic tradition as the apostles) to the inhabitants of a city (identified as Antioch). After initially being treated as liars, they said, reinforced by a third apostle: ʾinnā ʾilaykum mursalūn “Yes, to you we have been sent” and, after being treated a second time as liars, they said: rabbunā yaʿlamu ʾinnā ʾilaykum lamursalūn “Our Lord knows it: yes, to you we have indeed been sent.” As a result, Qazwīnī believes that the “reinforcement” (taqwiya) of the utterance, while it is “a good thing” (ḥasuna) in the case of khabar ṭalabī, is “obligatory” (wajaba) only in the case of khabar ʾinkārī. The single and double reinforcements of the utterance do not make the difference here between khabar ṭalabī and khabar ʾinkārī but between two retorts, to a first and a second denial. Thus, a one to one relationship cannot be established between interpretation of the utterance and presence or absence of markers in the utterance, since these can be optional. No less, moreover, can it be established that it is always possible to give the hearer a role that is not his, as in Quran 11:37 and 23:27 (“Do not speak to me of those who are unjust: they will be swallowed up”): here ʾinnais not a reponse to an actual question of the hearerʾs but anticipates and forestalls a possible question on his part (“you will ask me: what of the unjust? I reply to you, … ”). Likewise, dealing with the first distinction, the speaker can perfectly do as if the hearer, knowing p (the statement) and knowing that the speaker knows it does not know the two things, for example, by saying to someone who is not praying al-ṣalāt wājiba (“prayer is obligatory”): it is then a third meaning that is engendered, of recall and even recall in order (Shurūḥ al-Talkhīṣ, vol. 1:199). The term “calculate,” emphasized already, is essential because it can be formalized. One would then have, in the

Arabic Linguistic Tradition II case of the first distinction: if speaker A says to hearer B pand that B does not know p, then A causes B to know p; if A says to B p and that B knows p, then A causes B to know not p, but that he knows p; if A says to B p and that B knows both p and that A knows p, then A reminds B of p.

8.2.2.2 Theʾinshāʾ It is a pragmatic calculation of this type that is found in the chapter on the ʾinshāʾ. Qazwīni treats under this name what Sakkākī treats under ṭalab. Like Sakkākī, he subsumes five species under it. But, unlike Sakkākī, he does not take the trouble to show how these five species derive from a single type. This derivation, which has been studied by Moutaouakil (1982, 1990), is basically logical, in both form (Porphyrian tree) and vocabulary. A request requires an object (maṭlūb) that does not exist at the moment of the request (ghayr ḥāṣil waqt al-ṭalab). This object is realizable (ʾimkān al-ḥuṣūl) or not. If it is not, it is a wish (tamannī) (on which see now Zysow 2008). If it is, it is “in the mind” (fī l-dhihn) or “in the external world” (fī l-khārij). In the first case, it is interrogation (istifhām). In the second, it is a matter of “representation” (mutaṣawwar), whether negative (intifāʾ) or positive (thubūt). In the first case, it is an interdiction (naḥy), e.g., lā tataḥarrak (“do not move”), and in the second it is an order (ʾamr), e.g., qum (“get up”) and a vocative (nidāʾ), e.g., yā Zaydu(“Zayd!”). On the other hand, the result is typically pragmatic. For each of these species “expressions” are “instituted,” for example layta for the wish, hal and ʾa for interrogation, etc. But all “can be used in another sense than their own” (qad yustaʿmal fī ghayr maʿnāhu). Each of the five paragraphs of this chapter, then, studies how, beginning with these “primary” (ʾaṣliyya) meanings, a certain number of “secondary” (farʿiyya) meanings “are engendered” (yatawallad) situationally. Just one example (Talkhīṣ, 170–171): for the interdiction only one particle exists, the lā governing the apocopated form, the interdiction being the counterpart of the order, as regards the superiority [of the speaker over the hearer] (istiʿlāʾ). But it can be used for other things than “to require not to do” (ṭalab al-kaff ʾaw al-tark), for example threat (tahdīd), thus when one says to a slave who does not obey: “Do not obey me!” (lā tamtathil ʾamrī, lit. “do not conform to my order”). As specified by Sakkākī (Miftāḥ, 132), in this context, “It cannot be a matter of a requirement of disobedience, if such a thing existed: it is oriented toward something that does not exist …, and what finds itself engendered, aside from itself, is a threat” ( ) (on this concept of “engendering,” Firănescu 2011). It does no harm to recognize here what Searle (1975) calls an “illocutionary derivation.” Similarly, we can recognize more generally in the ʿilm al-maʿānī what Berrendonner (1981) calls a “semantics in Y”: the meaning here appears in fact as the result, symbolized by the stem of the Y, of a calculation operating on two components, symbolized by the arms of the Y, one linguistic (the “institution”) and the other “rhetorical” (the “use”).

8.2.2.3 The qaṭr or restriction: The semantically complex utterance In this section, Qazwīnī studies the utterances that might be called “restrictive” and the various syntactic mechanisms put into effect to express restrictiveness: negation and exception (nafī wa-istithnāʾ), coordination (ʿaṭf), preposing (taqdīm),1 ʾinnamā…. On the semantic level, the classification he proposes “crosses” two distinctions. One is logical. It can either restrict the subject (mawṣūf) to the attribute (ṣifa), e.g., mā Zaydun ʾillā kātibun (“Zayd is nothing but a prose writer,” “Zayd is only a prose writer”), or the attribute to the subject, e.g. mā fʿ l-dāri ʾillā Zaydun (“There is only Zayd in the house,” “Only Zayd is in the house”). The other is pragmatic. Here the restriction is called “singularization” (ʾifrād), “inversion” (qalb), or “specification” (taʿyīn), according to the belief of the hearer to which it reacts. Thus, the first utterance is addressed to someone who believes that Zayd is both (sharika) prose writer and poet. But the utterance Zaydun qāʾimun lā qāʿidun (“Zayd is standing, not seated”), with the syntactic process of coordination (ʿatf), is addressed to someone who believes either the opposite (“Zayd is seated”) or one or the other (“Zayd is seated or standing”).

8.2.2.4 Al-waṣl wa-l-faṣl, or conjunction and disjunction: The formally complex utterance “Conjunction” and “disjunction” are the two types of “connection” of one clause to another. Conjunction is defined as the coordination (ʿaḥf) of the first to the second and “disjunction” as its absence. This last thus corresponds to what is called, in our tradition, asyndesis or parataxis. Generally speaking, what rules “conjunction,” basically, is the semantic and formal homogeneity of “conjoined” utterances, and “disjunction” the semantic and/or formal heterogeneity of “disjoint” utterances. Turning to the details, the first opposition encountered is khabar/ʾinshāʾ.

Arabic Linguistic Tradition II They cannot in fact be coordinated with each other, for example, māta fulānun raḥima-hu llāhu: even if the two utterances are formally declaratives, only the first of them is semantically one (“So-and-so is dead”), the second being in fact semantically optative (“May Allah take pity on him!”). But phenomena are also found that are typically “enunciative,” e.g., Quran 2:14–15 ( ), where the disjunction of Allah yastahziʾu bihim is justified by the fact that “it does not belong to what they say” (laysa min maqūlihim), in other words by the change of utterer. Especially remarkable is the case of “disjunction” called “resumption” (istiʾnāf), because the second clause is to be understood as a response (jawāb) to an implicit question (suʾāl ) suggested by the first, as in the following verse: qāla lī kayfa ʾanta qultu ʿalīlū/saharun dāʾimun wa-ḥuznun ṭawīlū (“ʿHow are you?ʾ he asked me. ʿUnwell! Permanent insomnia and prolonged melancholy!ʾ I replied”); saharun dāʾimun wa-ḥuznu ṭawīlū responds in fact to a question like mā bāluka ʿalīlan (“What maladies do you have?”) or else mā sababu ʿillatika (“What is the cause of your malady?”). We see, from these few examples, that if “conjunction” is defined as a syntactic coordination, then “disjunction” could be interpreted as a semantic coordination, in the sense of Bally (1965): the two disjoint clauses are in the semantic relation of topic to comment and the comment implicitly makes reference to the topic: “He is dead (and, because he is dead,) may Allah take pity on him!”; “(They say that they do nothing but mock, but) it is Allah who mocks them; [I am] sick; (you are going to ask me from what): from permanent insomnia and prolonged melancholy.”

8.2.3 The ʿilm al-bayān If the ʿilm al-maʿānī designates itself by its very name as a semantics, the ʿilm al-bayān appears from the definition given by Qazwīnī to be a stylistics (Talkhīṣ, 236–235): “It is a science by which is recognized the communication of one and the same intention by different means in what concerns the clarity of its meaning” ( ). In the synchrony of postclassical rhetoric, the ʿilm al-bayān represents a point of view simultaneously complementary to and symmetrical with that of the ʿilm al-maʿānī: a point of view that can be called onomasiological, in that it goes from maʿnā “meaning” to lafẓ “form”. But in the diachrony of the discipline, it represents in fact on older point of view. The term bayān, which has Quranic resonances, it formally the maṣdar of the verb bāna–yabīnu “to be distinct.” It appears in the title of the founding work of Arabic rhetoric, the Kitāb albayān wa-l-tabyīn (“to be and to make distinct”) of al-Jāḥiẓ (d. 255/868) (Montgomery 2006). And we have seen ʿilm al-bayān competing with ʿilm al-balāgha as the name of the entire discipline. These “means” are not beyond counting. To count them, Qazwīnī (Talkhīṣ, 236–238) uses the concept, which we have already met, of waḍ ʿ(“institution”), which governs the relation of meaning (dalāla ʿalā maʿnā), lafẓ and maʿnā being called, under this relationship, al-mawḍūʿ (“that which is instituted”) and al-mawḍūʿ lahu (“that for which one institutes it”). This meaning is precisely “institutional” (waḍ ʿiyya), when the expression signifies “the entirety of that for which it is instituted” (tamām mā wuḍiʿ a lahu). It is also called in this case “adequacy” (muṭābaqa). It is “logical” (ʿaqliyya) when the expression means either a “part” (juzʾuhu) of that for which it is instituted or something “external” (khārij ʿ anhu). The first case is called taḍammun (“comprehensiveness,” “inclusion”) and the second iltizām (“implication”). It is only this last that interests the rhetorician. Linguistically speaking, one would say implicit meaning, but the linguistic term should not make us forget its logical origin (Lat. implicitus 〈 implicare). The “expression by which one aims at that which implies that for which it is instituted” (allafẓ al-murād bihi lāzim mā wuḍiʿ a lahu), in other words its implicit meaning (and not its explicit one), is called majāz if there is a “connection” (qarīna) indicating that the explicit meaning is not aimed at, and, if not, kināya (“metonymy”): this is in fact defined (Talkhīṣ, 337) as “the expression by which one aims at that which implies its meaning, with the possibility of aiming at the same time” ( ). Finally, the majāz can be based on a comparison (tashbīh, on which see especially Smyth 1992), hence the three sections of the ʿilm al-bayān: (1) tashbih; (2) ḥaqīqa wa-majāz; and (3) kināya. Here we cannot go into detail on these three sections and must be satisfied with a few remarks. First, majāz and ḥaqīqa do not at all refer, as has often been said, to the literal meaning or proper sense and the figurative meaning, but in fact to the expression used in its literal meaning and to the expression used in its figurative meaning. This confirms the onomasiological point of view (on this contrast, see Heinrichs 1984). Second, to the extent that majāz is opposed to kināya, it refers not to every figurative expression but more specifically to metaphorical expression. Third, because the majāz includes comparison, it refers in fact, for the most part, to figurative expressions based

Arabic Linguistic Tradition II on what there is in common between metaphor and comparison, namely, resemblance (mushābaha). As in our tradition, metaphor is seen as a truncated comparison, see Qazwīni, citing Sakkākī (Talkhīṣ, 330): “He divided the lexical majāz into metaphor and other and defined metaphor as the fact of mentioning one of the two terms of the comparison, aiming, by it, at the other” ( ). Ibn Khaldūn (Muqaddima, 1065–1066), for his part, holds resolutely to two “means,” istiʿāra and kināya, which he differentiates logically. He presents the first as a passage from the “antecedent” (malzūm) to the “consequent” (lāzim), e.g. Zaydun ʾasadun (“Zayd is a lion”: if Zayd is a lion, then he is courageous), and the second as a passage from the consequent to the antecedent, e.g. Zaydun kathīru ramādi l-qudūr (“Zayd has many ashes under his pots”: if Zayd has many ashes under his pots, it is because he is very hospitable). This last example is known from Qazwīnī (Talkhīṣ, 340–341), who includes it in the metonymy of one ṣifa (attribute) for another, but “distant” (baʿīda because it happens “through an intermediary” (bi-wāsiṭa). Qazwīnī reconstructs the chain of inferences leading from one to the other: “many ashes” (kathīr al-ramād), thus “much wood burned under the pots” (kathra ʾiḥrāq al-ḥaṭab taḥt al-qudūr), thus “many cooked dishes” (kathrat al-ṭabāʾikh), thus “many eaters” (kathrat al-ʾakala), thus “many hosts” (kathrat al-ḍīfān). The logical criterion is known from Sakkākī (Miftāḥ, 170) and, following him, Qazwīnī, but criticized by the latter, on the grounds that one cannot be transferred from the consequent, because there is no antecedent! Whatever criterion is employed, the Arab rhetoricians could not get very close to the idea because they had, at base, only two fundamental figures: metaphor and metonymy. In modern Western linguistics, Roman Jakobson (1896–1982) did. As we know, he proposed a correlation between metaphor and similarity, metonymy and contiguity, and, as a result, distributed them on the axes, respectively paradigmatic and syntagmatic, of language (Jakobson, 1956). One last word: Qazwīnī here limits majāz to majāz lughawī (“lexical”). But Sakkākī (Miftāḥ, 166f.) treats both in the ʿilm al-bayān of the majāz ʿaqlī (“logical”). The latter concerns the utterance, notably when its two terms are metaphorical, e.g., ʾaḥyā l-ʾarḍa shabābu l-zamāni “The youth of time [= spring] has revived the earth [= has produced vegetation].” Qazwīnī is not unaware of this, but he treats it in the ʿilm al-maʿānī, section I (Talkhīṣ, 45f.). This suggests that the rank of the expression ends up competing with the point of view. Ultimately, rhetoric appears as nothing but a contextual semantics: of the utterance and its constituents in the context of the discourse for the ʿilm al-maʿānī, of the word in the context of the clause for the ʿilm al-bayān.

8.2.4 The ʿilm al-badīʿ The ʿilm al-badīʿ comes simultaneously from very close and very far. From very close, because, as the third part of rhetoric, it is a recent innovation, due to Badr al-dīn Ibn Mālik (d. 686/1287), the son of the famous grammarian Ibn Mālik (d. 642/1274), in his work on rhetoric called al-Miṣbāḥ. For Sakkākī, it is not yet anything but a simple ornamental tailpiece to ʿilm al-bayān. Qazwīnī gives the following definition (Talkhīṣ, 347): “It is a science through which the manners of embellishing discourse become known, after observing the adequacy [of the expression for what the situation requires] and the semantic clarity” ( ). In this definition not only are the very words of Sakkākī found, but also the memory of a tailpiece (“after”) is preserved, even if, due to the fact of its assumption to the rank of part, a tailpiece no longer to the ʿilm al-bayān, but to the ʿilmā al-maʿānī wa-l-bayān. And from very far, because the term appears in one of the first works of rhetoric that we have, the Kitāb al-badīʿ of the poet (and caliph for a day) Ibn al-Muʿtazz (d. 296/808), who was assassinated the very day of his enthronement. The Kitāb al-badīʿ or book of the “new [style]” takes its name from its polemical aim, namely to show that the style of the poets called “modern” (muḥdathūn), such as Bashsh ār b. Burd (d. 167 or 168/784–5), Muslim b. al-Walīd (d. 208/823), or Abū Nuwās (d. between 198/813 and 200/815), is not so very “new” and that none of its features was not anticipated in the Quran, the traditions of Muḥammad and his companions, and old poetry. As for the rest, the Kitāb al-badīʿ presents itself as a simple catalogue of figures, five basic ones, including metaphor (istiʿāra), to which Ibn al-Muʿtazz adds twelve “ornaments” (maḥāsin) of discourse, in prose (kalām) or poetry (shiʿr), or a total of seventeen figures.

Arabic Linguistic Tradition II Through the centuries, the ʿilm al-badīʿ remained what it had been since the beginning: a tropology. The resemblance of the ʿilm al-badīʿ to what in our own tradition is called “rhetoric restricted to figures” is accentuated by the fact that Qazwīnī, following Sakkākī, divides them into two types (ḍarbān): “semantic” (maʿnawī) and “formal” (lafẓī). In this division the similarity must be recognized to what we call in our tradition “figures of thought” and “figures of expression.” Of the 37 figures named by Qazwīnī, 30 belong to the first type versus 7 to the second. It will suffice here to present the first of each of the two types of figures. The ṭibāq is defined (Talkhīṣ, 348) as “the union of two contraries, that is, of two opposed meanings, in the clause” (al-jamʿ bayna mutaḍāddayn ʾay maʿnayayni mutaqābilayni fī al-jumla), for example taḥsubuhum ʾayqāẓan wa-hum ruqūd (“you believe them awake, even though they are abed”). It is antithesis. The jinās (Talkhīṣ, 388) is defined as the “formal resemblance of two terms” (tashābuhuhumā fī al-lafẓ). According to their degree of proximity, this figure varies from simple paronomasia to repetition pure and simple.

8.2.5 Balāgha vs. khatāba ʿIlm al-balāgha is usually translated as “rhetoric.” This leads to the question of its relationship with what we call “rhetoric” in our own tradition. The question is not empty. Aristotleʾs Rhetoric was translated into Arabic under the name al-khaṭāba. “Rhetoric” reveals itself etymologically to be an “[art] of oratory” (rhetorikè technè). Khaṭāba is the maṣdar of the verb khaṭuba “to be eloquent” (khaṭīb). The eloquent man having been chosen as spokesman of his tribe, the same word designates, by metonymy, the orator, and khaṭaba itself passes from the quality (eloquence) to the activity (office of khaṭīb), whence the reading ⋆ khiṭāba. Genetically, there is essentially no relationship between ʿilm al-balāgha and Aristotelian rhetoric, even if there might have been, marginally, contacts. In the Arabic tradition, in fact, which is heir on this point to a late Alexandrian tradition, Aristotleʾs Rhetoric (and Poetics) are part of the Organon, in other words they are works of logic. Consequently, it is in the framework of falsafa that they are commented on, like the other works of the First Master, by the great falāsifa: al-Fārābī (d. 339/950), Ibn Sīna (d. 428/1037), Ibn Rushd (d. 595/1198) (Aouad 1989, Black 1990, Würsch 1991). Which is not to say that “Hellenizing” works of poetics or rhetoric cannot be found even outside falsafa (see for an overview Larcher 1998a): for the classical period and the Mashriq, we may mention the Naqd al-shiʿr (“criticism of poetry”) of Qudāma ibn Jaʿfar (d. 337/948?) and the Burhān fī wujūh albayān of Ibn Wahb (4th/10th century), initially published under the title Naqd al-nathr (“criticism of prose”) and falsely attributed to Qudāma; for the postclassical period and the Maghreb, the Minhāj al-bulaghāʾ wa-sirāj aludabāʾof Ḥāzim al-Qarṭājannī (d. 684/1285), which is actually, despite its title, a Hellenizing work of poetics, studied as such by Heinrichs (1969). But logic (manṭiq) having become in the 11th century a scholastic discipline, the entire Arabic tradition knows alkhaṭāba and al-shiʿr as excessively abbreviated names for the rhetorical syllogism (al-qiyās al-khaṭābī), i.e., enthymeme, and the poetic syllogism (al-qiyās al-shiʿrī), that is, premisses that are not merely “uncertain” (ghayr yaqīniyya) but actually “producers of imagination” (mukhayyila) (on the poetic syllogism see Schoeler 1983). This clearly shows the double reduction undergone by Aristotelian rhetoric in passing from the Greek world to the Muslim world and, within that, from falsafa to scholasticism. We may note meanwhile the existence in falsafa of a specific development: the theory of the “prophet-legislator.” We may also mention the criticism of the rhetoric and poetics of the philosophers (mutafalsifūn) made by Ibn al-Athīr in the Mathal al-sāʾir (I, 310–312), who quotes the Shifāʾ of Ibn Sīnā. Typologically, there are large differences between ʿilm al-balāgha and Greek rhetoric. Two deserve attention. Greek rhetoric, it is said, is an oratorical art. The ʿilm al-balāgha, in contrast, does not deal with a specific genre, but with all. This explains that the poetics of Sakkākī only deals with strictly technical aspects (meter and rhyme) of poetry. The rest, that is, the basics, the stylistic and thematic aspects, are a matter for ʿilm al-balāgha as they are for the other genres. Even the works that appear to be dedicated to specific genres, such as the Kitāb alṣināʿatayn fi-l-kiṭ atayn fi-l-kiṭāba wa-l-shiʿr (“The book of the two arts: the art of the secretary of the chancellery and poetry”) of Abū Hilāl al- ʿAskarī (d. after 395/1005) actually deal with all of them. Greek rhetoric defines itself as an “art of persuasion,” in other words places at its heart perlocutionary acts (Austin 1962). Conversely, the ʿilm al-balāgha, via the concept of ʾinshāʾ, places at its heart illocutionary acts. This double difference is easily explained if one “recontextualizes” Greek rhetoric and ʿilm al-balāgha. Aristotleʾs Rhetoric is intimately linked to the judicial and political institutions of Athens, exactly, moreover, as his Poetics is

Arabic Linguistic Tradition II linked to the cultural institutions (theater) of the Attic city. Not one of these institutions exists in the Islamic umma. On the other hand, it places one “word” above all the others, which it respects as the word of Allah (kalām Allāh), “revealed” (tanzīl) to Muḥammad, “transmitted” (tablīgh) by him, and transcribed in the Quran. Its addressee is not a spectator, who praises and blames, as in the ceremonial genre of Aristotelian rhetoric, and still less a judge to be persuaded, as in the judicial and deliberative genres of the same. But, once persuaded, he is in fact an interpreter. We are now approaching the hermeneutical side of the ʿilm al-balāgha. Let us note, however, that there exists, in the world of Islam, a form of institutionalized eloquence: this is the sermon (khuṭba) that the preacher (khaṭīb) gives in the pulpit (minbar) on Fridays. His art, which is oratory, obviously bears the same name as Aristotleʾs Rhetoric: al-khaṭāba. This homonymy is the source of much confusion among scholars with insufficient cultural background. To avoid confusion, we call the first, with Heinrichs (1987), “philosophical” rhetoric, and the second, with Larcher (1998), homiletics (for an overview of which see Halldén 2005, 2006 and, for the khuṭbain Jāḥiẓ, Soudan 1992, Avril 1994).

8.3 Intersections

8.3.1 ʿIlm al-balāgha and ʾʿijāz As the title of one of ʿAbd al-Qāhir al-Jurjān īʾs works reminds us, rhetoric is here included in a specifically Islamic context, where it has an apologetic aim: to “prove” (dalāʾil) “the inimitability of the Quran” (ʾiʿjāz al-Qurʾān). The ʾiʿjāz al-Qurʾān has become the object of a considerable literature, among which the most famous work is that of Bāqillānī (d. 403 or 404/1013). As a result, it has also become the object of a considerable literature on the part of Arabists (s.v. ʾiʿjāz in EI2 and Inimitability in EQ, Audebert 1982). Everything that is necessary and sufficient to know on the subject will be found there. The ʾiʿjāz al-Qurʾān was not definitively established as dogma until the 4th/10th century. Its basis is the verses of the Quran called “challenge” (taḥaddī: 10:39; 11:16; 17:90), in which the adversaries of Muḥammad are challenged to produce something similar to the Quran (min mithlihi), what in technical terms is called a muʿāraḍa or “replica.” The fact that the challenge was not met proves the “miraculous” (muʿjiz) character of the Quran, lit. “it renders incapable” of a replica, in other words leaves its adversaries speechless. The concept is thus clearly polemical. As often in Arabic, the term that designates it is only the most important of a series of collocations. It is the collocation and not the term itself that European languages interpret with the word “inimitability (of the Quran).” The mithli of the Quranic text can be interpreted as “similar to the Quran, from the point of view of maʿnā and/or lafẓ.” As a result, we distinguish a “thematic” ʾiʿjāz from a “stylistic” ʾiʿjāz. It is this last that has prevailed. The link between ʾiʿjāz and balāgha(as a quality) is recognized by Rummānī (d. 384/994), the author of one of the first works on the subject that we have: “what is at the highest degree of balāgha is muʿjiz and it has to do with the balāgha of the Qurʾān” ( ).2 As a result, a significant advance was made in the domain of balāgha, i.e. rhetoric. The close connection between rhetoric, in its two basic components, and literature of the ʾiʿjāz did not escape Ibn Khaldūn, even though he, writing long after, attempted to put the relationship in the opposite order (Muqaddima, 1068): The fruit of this discipline is understanding of the inimitability of the Qurʾān. This consists in the fact that the (language of the Qurʾān) indicates all the requirements of the situations (referred to), whether they are stated or understood. This is the highest stage of speech. In addition, (the Qurʾān) is perfect in choice of words and excellence of arrangement and combination. (tr. Rosenthal, abr. ed. 437)

8.3.2 ʿIlm al-balāgha and tafsīr Immediately after Ibn Khaldūnʾs discussion of relation between rhetoric and the ʾiʿjāz al-Qurʾān, he adds (Muqaddima, 1068): “This discipline is needed most by Qurʾān commentators” (

(But he says, “Most ancient commentators (tafāsīr) disregarded it, until Jār-Allāh az-Zamakhsarī (d. 538/1144) appeared” to provide a detailed rhetorical commentary on the Quran. Ibn Khaldūn does not conceal his admiration for this commentary, but he is embarrassed by the fact that its author is catalogued as “heterodox” (ʾahl al-bidaʿ),

Arabic Linguistic Tradition II hence his rejection by most of the “orthodox” (ʾahl al-sunna). He then devotes the entire rest of the chapter on rhetoric to a justification of being at the same time both perfectly “orthodox” and a reader of Zamakhsharī, taking into account the profit that can be drawn from his work for this discipline overall and the ʾiʿjāz al-Qurʾān in particular. In the process, Ibn Khaldūn implicitly reminds us that Zamakhsharī was a Muʿtazilite and that Muʿtazilism, condemned for its thesis called “Qurʾān created (by Allah)” (vs. “uncreated,” i.e. eternal), nonetheless played a considerable role in the elaboration of the dogma of the ʾiʿjāz al-Qurʾān and, as a result, in the development, and also reorientation, of rhetorical studies. If we now turn to the actual introduction that Zamakhsharī wrote to his commentary, we see that he unequivocally adumbrates his point of view. Reviewing all the Quranic specialists, both by background—jurist (faqīh), theologian (mutakallim), narrative expert (ḥāfiẓ al-qiṣ wa-l-ʾakhbār), preacher (wāʿiẓ)—and by form—grammarian (naḥwī), lexicographer (lughawī)—he concludes that (Kashshāf, 16): There is no one among them who can present himself to follow the [Qurʾānic] ways nor anyone who can immerse himself into [Qurʾānic] realities, other than a man who has excelled in two specific Qurʿānic sciences, namely the science of the maʿānī and the science of the bayān.

Here we see an occurrence of the expression, repeated exactly on p. 20, ʿilmā l-maʿānī wa-l-bayān, which EI2 (s.v. al-maʿānī wa-l-bayān) claims first appeared in Sakkākī, nearly a century later. And we also see that, in Zamakhsharīʾs view, there is no rhetoric but Qur anic ….

8.3.3 ʿūlm al-balāgha, ʿilm al-fiqh, and ʿilm uṣūl al-fiqh Less well-known are the relationships between linguistic and juridical disciplines. They are, nonetheless, perfectly well recognized by Ibn Khaldūn, in a general way in the passage quoted in the Introduction above and in a specific way in another passage of the Muqaddima (61). Wishing to distinguish khabar, in the historical sense (the plural ʾakhbār is one of the words for history in Arabic) from khabar in the juridical sense (in this sense khabar is a synonym of hadith), he articulates the difference by means of the linguistic opposition khabar/ʾinshāʾ: the historical khabar is a statement, true or false, but “most legal ʾakhbār are performative prescriptions that the Legislator made obligatory to be put into practice”

(Larcher 1993). The khabar or ḥadith constitutes, after the Quran, the second of the “sources” (ʾuṣūl) of jurisprudence (fiqh) in Islam. It takes its name from the fact that it transmits the Sunna, that is, the entirety of what was said, done, or endorsed by Muh ammad. This transmission takes the following form: ʾakhbaranī (or ḥaddathanī) fulān ʿan fulān ʿan fulān … qāl … (“So-and-so told me after So-and-so, who had it from So-and-so …, as follows: … ”). But since most of what is transmitted consists of speech, the term hadith itself has become synonymous with “said” by Muḥammad. Of course, if one is interested in the mechanism of juridical interpretation of the Quran and the Sunna, one soon discovers that it is rhetoric, in the sense defined above—that is, pragmatics. The ʾuṣūliyyūn, moreover, do not speak of the utterance (kalām), but of the address (khiṭāb). The khiṭāb, for the case where one might not have understood, is defined by the encyclopedist Kafawī (d. 1094/1683) as “the utterance oriented toward another, in order to cause to understand” (al-kalām al-muwajjah naḥw al-ghayr li-lʾifhām) (Kulliyāt, s.v. khiṭāb). ʾIfhām is a direct echo of mafhūm (vs. manṭūq), the name for implicit (vs. explicit) meaning among the ʾuṣūliyyūn, which we met in 8.3.1 in the quotation from Ibn Khaldūn: the juridical meaning of the utterance (Quranic verse or saying of Muḥammad) becomes “intelligible” only when one considers the utterance not for itself, but as an “address” from the Legislator (Allah or his prophet) to the faithful Muslim, who in this context is called mukallaf. Likewise, whereas the other disciplines divide the utterance into khabar and non-khabar, the ʾuṣūliyyūn divide the address into ṭalab and non-ṭalab, a sign of the preeminence of the former over the latter. According to the ʾIḥ kām (I, 91) of ʾĀmidī (d. 631/1233), the ṭalab, according to which he “imposes” (ʾiqtiḍāʾ) to “do” (fiʿl) or “not do” (tark), “categorically” (jāzim) or “uncategorically” (ghayr jāzim), is realized as “obligation” (ʾījāb) and “prohibition”

Arabic Linguistic Tradition II (taḥrīm), “recommendation” (nadb) and “condemnation” (karāhiya). The non-ṭalab either “gives the choice” (takhyīr) between doing and not doing (it is a “permission” ʾibāḥa) or “declares” (ikhbār) that such a thing is valid or not, cause, condition, or obstacle to some other, “to become a (rigorous) duty” (ʿazīma) or “tolerance” (rukhṣa) (Weiss 1992, Larcher 1992). These six legislative acts (sharʿ) constitute the ʾaḥkām sharʿiyya (that is, standards), “prescriptive” (taklīfiyya) for the first five and “ascriptive” (waḍʿiyya) for the sixth (Kafawī, Kulliyyāt, s.v. khiṭāb). Just one example: al-ṣalāt wājiba (“the canonic prayer is obligatory”) is the ḥukm sharʿī that can be derived from a Quranic utterance such as ʾaqim/ʾaqīmū l-ṣalāta (“complete [sg./pl.] the prayer”), which is linguistically an “order” (ʾamr), “rendering obligatory,” juridically, this act. The “prescriptive” aḥkām sharʿiyya—the most important—are organized, as we have just seen, on two levels: the ṭalab and the takhyīr. But the ṭalab level is in turn organized as a “logic square” of deontic modalities: wājib (“obligatory (to do)”) and ḥarām(“forbidden,” understood as obligatory to not do) are opposites; mandūb(“recommended (to do)” and makrūh (“reprehensible,” understood as recommended to not do) are subcontraries. Thus, wājib and ḥarām imply mandūb and makrūh while wājib and makrūh, on the one hand, and ḥarām and mandūb, on the other, are in a relation of contradiction. If we add that the takhyīr is understood as the disjunction “do or not do” and, as a result, the modality that follows from it, that of “permission” (ʾibāḥa), as bilaterally permission to do or not do, the ʾaḥkām sharʿiyya must not be represented on a linear axis, from good to bad, but as an inverted pyramid (Kalinowski 1976). The four quadrants represent the modalities that are a matter of ṭalab, the point is the mubāḥ, and the edges are the relations between it and them (Larcher 1992). Obviously the agreement in number (five) of the ʾaḥkām sharʿiyya “prescriptions” among the ʾuṭūliyyūn, and the species of talab among the rhetoricians, has been noted. The connection between ṭalab and takhyīr among the former finds its parallel among the latter (Talkhīṣ, 169): “the form introduced for an order can be used for something else, such as permission, thus ʿSit beside al-Hasan or Ibn Sīrīnʾ”

That is not the only influence of logic. From primary standards, derived pragmatically, can be derived secondary ones, logically, i.e., by reasoning (qiyās), another source of jurisprudence (at least for those juridical schools that recognize it). Juridical logic has often been distinguished from the logic of logicians by the type of reasoning that is at the heart of each (analogy vs. syllogism), but forgetting that in Arabic they are homonyms. In the postclassical period, the ʾuṣūl al-fiqh recuperated syllogistics but distorted it, as is shown by the following very well known example: al-nabīdh muskir (“wine is an intoxicating beverage”); kull muskir ḥarām (“every intoxicating beverage is forbidden”); al-nabīdh ḥarām (“wine is forbidden”). What makes the specificity of this syllogism is not its form. If we refer to the logical part, and one cannot get more classic than this, of the Miftāḥ of Sakkākī, we will confirm that this is a syllogism of the first figure, one of the two affirmative modes, the analogue of our Darii (except that, following the Arab tradition, the minor premise is stated before the major). What creates its specificity is that it links descriptive and prescriptive utterances: the character as a standard of the major premise (which is a “saying” of Muḥammad) and of the conclusion is attested by the fact that x is ḥarām (“x is forbidden”) can be replaced by the performative ḥarramtu x (“I forbid x”), see ʾĀmidī, ʾIḥkām, I:12 and IV:48 (on juridical logic see EI2 s.v. Manṭiq, Brunschvig 1970, Weiss 1992, Larcher 1992, Hallaq 1994). Indeed, faʿaltu is the most usual form of juridical performatives, the ṣiyagh al-ʿuqūd wa-l-fusūkh (“contractual and renunciative formulas”) of the Arab tradition, that is, utterances used to tie or untie juridical bonds. It suffices, to be convinced, to open one of the great treatises of fiqh of the postclassical period, such as the Bādāʾiʿ of Kās ānī (d. 587/1189). These treatises are organized in two parts: al-ʿibādāt (“worship”), governing the duties of the believer toward the divinity, and al-muʿāmalāt(“transactions”), governing relations among individuals. In this last part, for example in the book of contracts (al-buyūʿ) or the chapter on repudiation (ṭalāq) or manumission (ʾiʿtāq), it is confirmed that the performatives of these acts are in order of priority of occurrence: biʿtu(“I sell”) and ishtaraytu (“I buy”), ʿallaqtu-ki (“I repudiate you”), ʾaʿtaqtuka or ḥarrartuka (“I manumit you or I free you”), etc.

8.3.4 A balāgha integrated into naḥw: Raḍī al-dīn al-Astarābādhī All this is found in grammar itself, in particular in the Sharḥ al-Kāfiya of Astarābādhī (Larcher 1990, 1992, 1998b, 2000, 2007). As its title indicates, it is actually a commentary on the Kāfiya, a brief introduction (muqaddima) to

Arabic Linguistic Tradition II syntax, by Ibn al-Hājib (d. 646/1249). Ibn al-Hājib was also a Malekiteʾuṣūlī, author of two works on the matter, the Muntahā and the Mukhtaṣar, the latter the object of many commentaries. Ibn al-Hājib seems moreover to be the first grammarian to make explicit use of the category of ʾinshāʾ. But although he was ʾuṣūlī, Ibn al-Hājib did not truly make ʾinshāʾ a self-referential and performative conception. He conceived it more readily as the subjective mode of discourse (as opposed to the khabar, objective mode), or again, semiotically, as the mode of signifying “expression” (Ausdruck) a psychological event that the statement “represents” (Darstellung), using an opposition due to Karl Bü hler (1879–1963). In this last conception, one does not state oneʾs intention, one “signals” (tanbīh) it. This is the term, coupled with ʾinshāʾ in his successors, that in the Maḥṣūl (I, 1, 317–318) of Fakhr al-dīn al-Rāzī (d. 606/1209) designates utterances that are neither statements nor requests. Conversely, Astarābadhī, although he was a logician, is not at all a logicist. He dedicates considerable space to the category of ʾinshāʾ. If we gather all the passages where he mentions it, we can actually extract from the Sharḥ al-Kāfiya a veritable “pragmatic theory.” It appears as a diptych, where the category of ʾinshāʾ constitutes the conceptual panel and the performative faʿaltu the formal panel. The elementary propositions can be stated as follows: (1) Kalām is everything the utterance of which constitutes an “act of the utterer” (fiʿl al-mutakallim). The priority given to the semantico-pragmatic criterion over the formal criterion (jumla) first of all permits Astarābadhī to understand the category of ʾinshāʾ as the totality of utterances. He uses ʾinshāʾ in two ways: both in opposition to khabar and ṭalab as the specific name of the juridical performative (Sharḥ al-Kāfiya, I, 8); and in opposition to khabar alone as the generic name for nondeclarative utterances, but subdividing them into ṭalabī (“requesting, demanding”) and ʾīqāʿī (“operatives” = juridical performatives) (Sharḥ al-Kāfiya, II, 221). This double classification confirms that the category of ʾinshāʾ is indeed the product of a generalization from performatives that are both explicit and juridical. But he also continues a reflection that was begun by Ibn al-Hājib on mixed utterances, of the exclamatory type, “susceptible of being both assertive and performative” (yaḥtamil al-ʾikhbār wa-l-ʾinshāʾ) (ʾAmālī IV, 149–150), and then was continued by Ibn Mālik under the name khabar ʾinshāʾī (Sharḥ alTashīl III, 33). Astarābādhī gathers them under the name ʾinshāʾ juzʾu-hu l-khabar(Sharḥ al-Kāfiya II, 93 and 311). There one finds the kam called “assertive” (“How…!”), as opposed to the interrogative kam (“How …?”), rubba (“A little of …!”), the “verbs of praise and blame” (ʾafʿāl al-madḥ wa-l-dhamm), and the “verb of admiration” (fiʿl al-taʿajjub). Astarābādhī thus opens the door to the interpretation of the “element” (juzʾ) khabar, not as posited, but as presupposed. This interpretation is made explicitly, in the commentary in the margin (Ḥal-Kāfiya of the Sharḥ al-Kāfiya (II, 311), by his own commentator ʿAlī b. Muḥammad al-Jurjānī (d. 816/1413) under the name lāzim ʿurfī (“empirical implication”). This last thus characterizes the element Zaydun ḥasanun (“Zayd is good”) in relation to the performative of “admiration” mā ʾaḥsana Zaydan (“How good Zayd is!”). Furthermore, Astarābādhī turns to ascriptive statements, of the type Zaydun ʾafḍalu min ʿAmrin (“Zayd is superior to ʿAmr”). He says in fact that the uttering of this declaration performs an act of tafḍīl. His own commentator (Ḥāshiya, in Sharḥ al-Kāfiya, II, 311) defines this not as “to make superior” (jaʿlu-hu ʿafḍal), but as “to call superior” (al-ʾikhbār ʿan kawnihi ʾafḍal). In other words, he gives the verb faḍḍdala, of which tafḍīl is the maṣdar, not a “factitive” interpretation, but, following the terminology of the French linguist Emile Benveniste (1902–1976), “delocutive” (Benveniste 1958). Astarābādhī finally arrives at purely descriptive statements, of the type Zaydun qāʾimun (“Zayd is standing”), of which he says that the utterance performs an act of assertion (ʾikhbār). The same criterion, of kalām defined by the act of the utterer, moreover allows Astarābādhī to extend the very concepts of kalām and ʾinshāʾ both below and beyond the classical “utterance” defined formally as jumla. Below, as in the case of the ʾasmāʾ al-ʾafʿāl. This is what the Arabic grammarians call “nouns,” the denominal of which (musammā) are verbs, in other words have the same semantic value as them, and which correspond to what we in our tradition call interjections. But whereas the other grammarians paraphrase ʾuff (“Bah!”) and ʾawwah(“Alas!”) with ʾataḍajjaru (“I am distressed”) and ʾatawajjaʿu (“I am distressed”), which they consider to be statements, Astarābādhī paraphrases them with taḍajjartu and tawajjaʿtu, which are explicitly called ʾinshāʾī (Sharḥ al-Kāfiya II, 65). Beyond, as in the case of the pragmatic connectives p lākinna q and p ʾinna q, or p and q are utterances. Of the first (Sharḥ al-Kāfiya, II, 346), Astarābadhī says that “in lākinna, there is the sense of istadraktu ” (wa-fī lākinna maʿnā istadraktu). He describes it using the example jāʾanī Zaydun lākinna ʿAmran lam yajiʾ (“Zayd came to me, but [as for] ʿAmr, did not come”), as carrying out an act of preemptive rectification (istidrāk) by q, of the false

Arabic Linguistic Tradition II conclusion r (“thus ʿAmr came also”) that is in danger of being drawn from p by the hearer, who is aware of the close relationship between Zayd and ʿAmr. Of the second (Sharḥ al-Kāfiya II, 349), Astar ābādhī says that “placed in the middle of a kalām, but being the beginning of a new kalām” (kāna fī wasṭ kalām lākinnahu ibtidāʾ kalām ʾākhar), for example ʾakrim Zaydan ʾinnahu fāḍil (“Honor Zayd: (for) he has merit”), he presents a “justification” (ʿilla) of p by q. In other words, for him, there are here three kalām: the two utterances p and q and the discourse p ʾinna q. The justification, like preemptive rectification, is an act of the utterer, which operate not at the level of the utterance, but at the level of discourse—that is, of the articulation of the utterances among themselves. We may observe that Astarābādhī calls ʾinna-hu fāḍil a “kalām mustaʾnaf” (“resumption”), very certainly in the sense of the rhetoricians (cf. 8.2.2.4), that is, constituting an answer to the implicit question suggested by the preceding utterance: “I say to you p. You are going to ask me why. I answer you q.” The very etymology of French car (“for”) (〈 Lat. quare “why?”) recalls the movement. (2) This “act of the utterer” can be represented by a performative faʿaltu. Formally, this is a jumla. But if we adopt the formalism of the linguistic philosophers (Searle 1969), namely, F(p), where F is an illocutionary force and p a proposition, faʿaltuobviously represents F and not p, a modus, not a dictum. We might just as well adopt the formalism of the logicians, namely f(x), where f is a function and x the argument that this function is going to saturate. If, then, we assimilate F to f, the illocutionary force to a function, faʿaltu appears in the following cases: (1) it is a function that is not going to saturate any explicit argument: this is the case for interjections, which constitute a comment on an implicit topic; and (2) it is a function whose argument is (a) a term n: this is the case of the vocative, or (b) the term n of an incomplete proposition: this is the case of exclamations, or (c) a proposition p: this is the case of classical utterances, or (d) two or more explicit or implicit utterances: this is the case of pragmatic connectives. (3) This representation is either purely semantic or semantico-syntactic, according to whether it does not or contrariwise does play a role in the derivation of sentences. It plays no role in the case of interjections or pragmatic connectives. To say that ʾuffor ʾawwah has the meaning of taḍajjartu or tawajjaʿtu called ʾinshāʾī, in effect says that these expressions really have as their meaning a behavior (taḍajjar-, tawajjaʿ-) of the one who speaks (-tu), but that this behavior is not asserted, but “played” by the utterance of the interjection. It does, on the other hand, play one in the case of the vocative yā Zaydu(“Zayd!”), which Astarābādhī derives from nadaytu/daʿawtu Zaydan (“I call Zayd”) (Sharḥ al-Kāfiya, I, 132). This derivation makes the vocative, on the syntactic level, look like the object complement of an understood verb: the grammarians argue that the vocative always has the marker of the accusative, except in the case illustrated by yā Zaydu, that is, in the case where it refers to an expression that is both simple (mufrad) and definite (maʿrifa) and where an undeclinable ending (- u) appears. The best grammarians themselves recognize that this representation is not entirely satisfactory on the semantic level: it conceals the fact that the vocative transforms a term of reference into a term of address. Whence the remark that this verb is necessarily understood and cannot appear, replaced as it is by the particle yā. It also plays a role in the case of Zaydun qāʾimun haqqan “Zayd is standing, truly,” which Astarābādhī derives from qultu Zaydun qāʾimun qawlan ḥaqqan “I say ʿZayd is standingʾ with a true saying” (Sharḥ al-Kāfiya I, 124). This derivation makes ḥaqqan appear, on the syntactic level, as the “resultative complement” (mafʿūl muṭlaq), of the “specificatory” type (li-bayān al-nawʿ), of an understood verb, which justifies its accusative inflection. And it makes it appear, on the semantic level, as qualifying the speech act (the act of assertion) as veridical, and not the thing said (the fact declared) as true, in other words the equivalent of a sentence adverb, but with enunciative effect. (4) Finally, this representation is abstract. It is abstract in the sense that there does not necessarily exist an explicit performative corresponding to the illocutionary act or that, if it exists, does not necessarily have the form faʿaltu. The first case is represented by the paraphrase of the interjections, which, for Astarābādhī, looks like nothing but a pure invention by the grammarians. The second case is represented by the vocative. Whereas Ibn Mālik understands the verb in the form ʾafʿalu, while considering it a performative (Sharḥ al-Tashīl III, 385), Astarābādhī prefers explicitly, in this role, the form faʿaltu, which he justifies as follows (Sharḥ al-Kāfiya I, 132): The verb has as its object the performance: and therefore it is better to understand it in the form of the past, i.e. daʿawtu or nādaytu, the performative verbs appearing most often in this form

Arabic Linguistic Tradition II

Faʿaltu is thus chosen for its expressive power and is thus equivalent to a real formalization.

8.4 Conclusion With the grammarian Raḍī al-dīn al-Astarābadhī, we attain an extreme degree of sophistication that has no equivalent anywhere except, perhaps, in posterity, essentially Turko-Irano-Indian, of the Kāfiya, the commentators of it keeping an eye on Astarābādhīʾs commentary. Its very sophistication militates against and continues to militate against an appreciation of the work. Nonetheless, if, as done here, we contextualize it, we observe that it simply leads in the same direction, but farther than the entire Arab linguistic tradition in the postclassical period: the direction of a strong and original bond, essentially pragmatic in nature, between the various disciplines, entirely or partially linguistic, that constitute this tradition.

Bibliography Primary Sources ʾĀmidī, ʿIḥkām = Sayf al-dīn Abū l-Ḥasan ʿAlī b. Abī ʿAbī b. Muḥammad al-ʾĀmidī. Al-ʾIḥkām fī ʾuṣūl al-ʾaḥkām. Cairo, Muʾassasat al-Ḥalabī, 1487/1967. ʿAskarī, Kitāb al-ṣināʿatayn = Abū Hilāl al-Ḥasan b. ʿAbd Allāh b. Sahl al-ʿAskarī, Kitāb al-ṣināʿ atayn fī l-kitāba wal-shiʿr, 1st édition, Istanbul, 1320/1902. Astarābādhī, Sharḥ al-Kāfiya = Raḍī al-dīn Muḥammad b. al-Ḥasan al-Astarāb ādhī, Sharḥ Kāfiyat Ibn al-Ḥājib. 2 vols. Istanbul: Maṭbaʿat al-sharika al-ṣiḥāfiyya al-ʿuthmāniyya. 1275/1858 and 1892. [Répr. Beirut: Dār al-kutub alʿilmiyya, s.d.]. Badr al-dīn Ibn Mālik, Miṣbāḥ = Badr al-dīn Ibn Mālik al-Miṣbāḥ fī l-maʿānī wa-l-bayān wa-l-badīʿ, ed. Ḥusnī ʿAbd al-Jalīl Yūsuf. Cairo: Maktabat al-ʾĀdāb, 1989. Bāqillānī, ʾIʿjāz = Abū Bakr Muḥammad b. al-Ṭayyib al-Bāqillānī ʾIʿjāz al-Qurʾān, ed. Aḥmad Ṣaqr, coll. Dhahāʾir alʿArab 12. Cairo: Dār al-maʿārif, 1963. Jāḥiẓ, Bayān = Abū ʿUthmān ʿAmr b. Baḥr al-Jāḥiẓ, Kitāb al-bayān wa-l-tabyīn, ed.ʿAbd al-Salām Muḥammad Hārūn, 4 parts in 2 vol., Cairo, Maktabat lajnat al-taʾlīf wa-l-tarjama wa-l-nashr, 1367/1948. Jurjānī, ʾAsrār al-balāgha = ʿabd al-Qāhir al-Jurjānī, ʾAsrār al-balāgha, ed. Hellmut Ritter, Istanbul, Government Press, 1954. Jurjānī, Dalāʾil al-ʾIʿjāz = ʿabd al-Qāhir al-Jurjānī, Dalāʾil al-ʾIʿjāz fī ʿilm al-maʿānī, ed. Muḥammad Rashīd Riḍā. Beirut: Dār al-maʿrifa, 1402/1982. Jurjāni, Ḥāshiya = ʿAlī b. Muḥammad, al-sayyid al-sharīf, al-Jurjānī al-Ḥāshiya ʿalā sharḥ al-Kāfiya, cf. Astarābādhī. Khafājī, Sirr al-faṣāḥa = Abū Muḥammad ʿAbd Allāh b. Muḥammad b. Saʿīd b. Sinān al- Khafājī al-Ḥalabī Sirr alfaṣāḥa. Cairo: Maktabat Muḥammad ʿAlī Ṣubayḥ, 1389/1969. Ibn al-ʾAthīr, al-mathal al-sāʾir = Abū l-Fatḥ Ḍiyāʾ al-dīn Naṣr Allāh b. Muḥammad b. Muḥammad b. ʿAbd al-Karīm almaʿrūf bi-Ibn al-ʾAthīr al-Mawṣilī, al-Mathal al-sāʾir fī ʾadab al-kātib wa-l-shāʿir, ed. Muḥammad Muḥyī al-dīn ʿAbd al-Ḥamīd, 1 vol. in 2 parts, Cairo, Mustafa al-Bābī al-Ḥalabī, 1358/1939. Ibn Fāris, Ṣāḥibī = Abū l-Ḥusayn Aḥmad Ibn Fāris al-Ṣāḥibī fī fiqh al-lugha wa-sunan al-ʿarab fī kalāmihā, éd. Moustafa El-Chouémi, Coll. Bibliotheca philologica arabica, published under the direction of R. Blachère and J. Abdel-Nour, vol. 1. Beirut: A. Badran & Co. 1383/1964.

Arabic Linguistic Tradition II Ibn Khaldūn, Muqaddima= Walī l-dīn ʿAbd al-Raḥmān b. Muḥammad Ibn Khaldūn. al-Muqaddima, vol. I of Kitāb alʿibar. Beirut: Maktabat al-madrasa and Dār al-kitāb al-lubnānī, 1967. Ibn al-Ḥājib, ʾAmālī = Jamāl al-dīn Abū ʿAmr ʿUthmān b. ʿUmar dit Ibn al-Ḥājib, al-Amālī al-naḥwiyya. Ed. Hādī Ḥasan Ḥammūdī. 4 parts in 2 vols. Beirut: ʿĀlam al-kutub et Maktabat al-Nahḍa al-ʿarabiyya, 1405/1985. Ibn al-Ḥājib, Kāfiya = al-Muqaddima al-kāfiya fī l-naḥw, cf. Astarābādhī. Ibn Mālik, Sharḥ al-Tashīl = Jamāl al-dīn Muḥammad b. ʿAbd Allāh b. ʿAbd Allāh al-Ṭāʾī al-Jayyānī al-Andalusī Ibn Mālik Sharḥ al-Tashīl, ed. ʿAbd ar-Raḥmān and al-Sayyid Muḥammad al-Makhtūn. 4 parts in 2 volumes. Gizeh: Hajr li-ṭibāʿa wa-l-nashr wa-l-tawzīʿ wa-l-ʾi ʿlān, 1410/1990 Ibn al-Muʿtazz, Kitāb al-badīʿ, ed. Ignatius Kratchkovsky, E.J.W. Gibb Memorial Publications, New Series, X, 1935. Ibn Rashīq, ʿUmda= Abū ʿAlī al-Ḥasan Ibn Rashīq al-Qayrawānī al-ʾAzdī al-ʿUmda fī maḥāsin al-shiʿr wa-ʾādābihi wa-naqdihi, ed. Muhammad Muḥyī al-dīn ʿAbd al-Ḥamīd. Beirut: Dār al-Ğīl, 1972. Ibn Wahb, al-Burhān fī wujūh al-bayān, ed. Aḥmad Maṭlūb and Khadīja al-Ḥadīthī, Baghdad 1387/1967. Kafawī, Kulliyyāt = Abū al-Baqāʾ ʾAyyūb b. Mūsā al-Ḥusaynī al-Kafawī. Kulliyyāt al-ʿulūm. Ed. ʿAdnān Darwīsh and Muḥammad al-Maṣrī, 5 vols. Damascus: Wizārat al-Thaqāfa wa-l-Irshād al-Qawmī, 1981. Kāsānī, Badāʾiʿ = Abū Bakr b. Masʿūd al-Kāsānī Kitāb Badāʾiʿ al-ṣanāʾiʿ fī tartīb al-sharāʾiʿ, 7 vols. Cairo: Maṭbaʿat al-Jamāliyya, 1327–1328/1909–1910. Qarṭājannī, Minhāj = Abū l-Ḥasan Ḥāzim al-Qarṭājannī Minhāj al-bulaghāʾ wa-sirāj al-ʾudabāʾ, éd. Muḥammad alĞabīb Ibn al-Ḫawja, 3rd ed. Beirut, Dār al-Gharb al-ʾislāmī, 1986 [first ed. Tunis, 1966]. Qazwīnī, ʾĪḍāḥ= Jalāl al-dīn Muḥammad b. ʿAbd al-Raḥmān al-Qazwīnī al-ʾĪḍāḥ fī sharḥ Talḫīṣ al-Miftāḥ. Cf. Shurūḥ al-Talḫīṣ. Qazwīnī, Talkhīṣ = Jalāl al- dīn Muḥammad b. ʿAbd al-Raḥmān al-Qazwīnī Talkhīs al-miftāḥ fīʿulūm al-balāgha. Ed. ʿAbd al-Raḥmān al-Barqūqī. Cairo, al-Maktaba al-tijāriyya al-kubrā, n.d. Qudāma ibn Jaʾfar, Naqd al-shiʾr,ed. by S.A. Bonebakker, Leiden: Brill, 1956. Rāzī, Maḥṣūl= Fakhr al-dīn al-Rāzī al-Maḥṣūl fī ʿulūm ʾuṣūl al-fiqh. Ed. Ṭaha Jābir Fayyād al-ʾUlwānī. Imām Muḥammad b. Saʾūd University, 1399/1979. Sakkākī, Miftāḥ = Abū Yaʿqūb Yūsuf b. Abī Bakr Muḥammad b. ʿAlī al-Sakkākī, Miftāḥ al-ʿulūm. Cairo: Matbaʿat altaqaddum al- ʿilmiyya, 1348/1929. [répr. Beirut, Dār al-kutub al- ʿilmiyya, n.d.]. Shurūḥ al-Talkhīṣ, 4 vols. Cairo, Maṭbaʿat Īsā al-Bābī al-Ḥalabī. 1937. Zamakhsharī, Kashshāf= Abū l-Qāsim Jār Allāh Maḥmūd b. ʿUmar al- Zamakhsharī al-Kashshāf ʿan ḥaqāʾiq al-tanzīl wa-ʿuyūn al-ʾaqāwīl fī wujūh al-taʾwīl, Cairo, Muṣṭafā al-Bābī al-Ḥalabī, n.d.

Secondary Sources Aouad, Maroun. 1989. La Rhétorique. Tradition syriaque et arabe. In Dictionnaire des philosophes antiques, vol. 1, ed. Richard Goulet, 455–472. Paris: Editions du CNRS. Arnaldez, A. Manṭiq. In Encyclopedia of Islam/Encyclopédie de lʾIslam, new edition, Vol. 6, 427–38. Leiden: Brill. Audebert Claude-France. 1982. Al-Khaṭṭābī et lʾinimitabilité du Coran. Traduction et introduction au Bayān ʾIʿjāz al-Qurʾān. Damascus: Institut Français dʾEtudes Arabes. Austin, John. 1962. How to do things with words. London: Oxford University Press. Avril, Marie-Hélène. 1994. Généalogie de la khuṭba dans le Kitāb al-bayān wa-l-tabyīn de Jāhiẓ. Bulletin dʾEtudes Orientales 46: 197–216.

Arabic Linguistic Tradition II Bally, Charles. 1965. Linguistique générale et linguistique française, 4th rev. ed. Berne: Francke. Benveniste, Emile. 1958 [1966]. Les verbes délocutifs. In Studia Philologica et litteraria in honorem L. Spitzer, ed. Anna G. Hatcher and K. L. Selig, 57–63. Berne: Francke [repr. in Problèmes de linguistique générale, I, 277–285. Paris: Gallimard. 1966]. Bernard, M. Muʿāmalāt. In Encyclopedia of Islam/Encyclopédie de lʾIslam, new edition, 1960–2005, Vol. 7, 25–59. Leiden: Brill. Berrendonner, Alain. 1981. Eléments de pragmatique linguistique. Paris: Minuit. Black, Deborah L. 1990. Logic and Aristotleʾs Rhetoric and Poetics in Medieval Arabic philosophy. Leiden: Brill. Bonebakker, S. Maʿānī wa l-bayān. In Encyclopedia of Islam/Encyclopédie de lʾIslam, new edition, Vol. 5, 904–08. Leiden: Brill. Brunschvig, Robert. 1970. Logic and law in Classical Islam. In Logic in Classical Islamic culture, ed. G.E. von Grunebaum, 9–20. Wiesbaden: Harrassowitz. Buburuzan [Firănescu], Rodica. 1995. Significations des énoncés et actes de langage chez Ibn Fāris. In Proceedings of the colloquium on Arabic linguistics, Bucharest August 29–September 2, 1994, vol. 1, 103–114. Bucarest: University of Bucharest, Center for Arab Studies. Firănescu, Daniela Rodica. 2011. Readings notes on Sakkākīʾs concept of ʿsemantic engendering.ʾ In A Festchrift for Nadia Anghelescu, ed. Andrei A. Avram, Anca Focşeneanu, and George Grigore, 215–233. Bucarest: Editura Universităţii din Bucureşti. Frank, Richard. 1981. Meanings are spoken of in many ways: The earlier Arab grammarians. Le Muséon 94: 259– 319. Ghersetti, Antonella. 1998. Quelques notes sur la définition canonique de balāgha. In Philosophy and arts in the Islamic World, Proceedings of the eighteenth congress of the Union Européenne des Arabisants et Islamisants held at the Katholieke Universiteit Leuven (Septembre 3–Septembrer 9, 1996), ed. U. Vermeulen et D. De Smet, 57–72. Leuven: Peeters. ——. 2002. La définition du khabar(énoncé assertif) dans la pensée rhétorique de ʿAbd al-Qāhir al-Jurjānī. In Studies in Arabic and Islam, Proceedings of the 19th Congress, Union Européenne des Arabisants et Islamisants, Halle 1998, ed. S. Leder with H. Kilpatrick, B. Martel-Thoumian, and H. Schönig, 367–77. Louvain: Peeters. Guillaume, Jean-Patrick. 1986. Sibawayhi et lʾénonciation: une proposition de lecture. Histoire Epist é mologie Langage 8: 53–62. Hallaq, Wael. 1994. Law and legal theory in Classical and Medieval Islam. Aldershot: Variorum. Halldén, Philip. 2005. What is Arab Islamic rhetoric? Rethinking the history of Muslim oratory art and homiletics. International Journal of Middle East Studies 37: 19–38. ——. 2006. Rhetoric. In Medieval Islamic Civilization, ed. Josef Meri, 679–681. New York: Routledge. Heinrichs, Wolfhart. 1969. Arabische Dichtung und griechische Poetik. Ḥāzim al-Qarţājannīʾs Grundlegung der Poetik mit Hilfe aristotelischer Begriffe, Beiruter Texte und Sudien, Band 8. Beirut: Orient Institut & Wiesbaden: Steiner. ——. 1984. On the genesis of the ḥāqīqa-majāz dichotomy. Studia Islamica 59: 111–140. ——. 1987. Poetik, Rhetorik, Literaturkritik, Metrik und Reimlehre. In Grundriss der arabischen Philologie, Band II Literaturwissenchaft, ed. H. Gätje, 177–207. Wiesbaden: Reichert. ——. 1998. Rhetoric and poetics. In Encyclopedia of Arabic literature, ed. Julia Meisami and Paul Starkey, 651–656. London: Routledge.

Arabic Linguistic Tradition II Jakobson, Roman. 1956. Two aspects of language and two type of aphasic disturbances. In Roman Jakobson and Morris Halle, Fundamentals of language, Part II, 55–82. The Hague: Mouton & Co. Jenssen, H. 1998. The subtleties and secrets of the Arabic language: Preliminary investigations into Al-Qazw ī n ī?s Talkhīṣ al-Miftāḥ. London: Hurst. Kalinowski, George. 1976. Un aperçu élémentaire des modalités déontiques. Langages 43: 10–18. Khalafallah, M. Badīʾ; In Encyclopedia of Islam/Encyclopédie de lʾIslam, new edition, Vol. 1, 881–82. Leiden: Brill. Larcher, Pierre. 1980. Information et performance en science arabo-islamique du langage. PhD diss., Université de Paris III. ——. 1990. Eléments pragmatiques dans la théorie grammaticale arabe post-classique. In Studies in the history of Arabic grammar, vol. 2, ed. Kees Versteegh and Mike Carter, 193–214. Amsterdam: Benjamins. ——. 1991. Quand, en arabe, on parlait de lʾarabe… (II) Essai sur la caté gorie de ʾinshāʾ (vs khabar). Arabica 38: 246–273. ——. 1992. Quand, en arabe, on parlait de lʾarabe… (III) Grammaire, logique, rhétorique dans lʾislam postclassique. Arabica 39: 358–384. ——. 1993. Les arabisants et la catégorie de ʾinshāʾ. Histoire dʾune “occultation.” Historiographia Linguistica 20: 259–282. ——. 1998a. Eléments de rhétorique aristotélicienne dans la tradition arabe hors la falsafa. In La Rhétorique d’Aristote: traditions et commentaires de lʾAntiquité au XVIIème siècle(La Baume lès Aix, 10–12 Juillet 1995), ed. Gilbert Dahan et Irène Rosier-Catach, 241–256. Traditions de lʾAntiquité classique. Paris: Vrin. ——. 1998b. Une pragmatique avant la pragmatique: “médiévale,” “arabe” et “islamique.” Histoire Epistémologie Langage 20: 101–116. ——. 2000. Les relations entre la linguistique et les autres sciences dans la société araboislamique. In History of the language sciences, vol. 1, ed. E. F. K. Koerner, Sylvain Auroux, Hans-Josef Niederehe, and Kees Versteegh, 312–318. Berlin: Walter de Gruyter & Co. [English translation: Relationships between linguistics and the other sciences in Arabo– islamic society. In The Early Islamic Grammatical Tradition, ed. Ramzi Baalbaki, 337–348. The Formation of the Classical Islamic World 36. Aldershot: Ashgate Publishing Limited. 2007]. ——. 2007. ʾInshāʾ. Encyclopedia of Arabic Language and Linguistics, vol. 2, Eg–Lan, 358–361. Brill: Leiden. ——. 2009. Mais quʾest-ce donc que la balāgha?. In Literary and philosophical rhetoric in the Greek, Roman, Syriac and Arabic worlds, ed. Frédérique Woerther, 197–213. Europaea Memoria. Hildesheim: Olms. Martin, Richard. 2002. Inimitability. In The Encyclopaedia of the Qurʾān, Vol. 2. ed. J.D. McAuliffe, 526–36. Leiden, Brill. Mehren, A. 1853 [1970]. Die Rhetorik der Araber. Kopenhagen: Otto Schwarz & Wien: Kaiserl. Königl. Hof- und Staatsdrückerei. (Repr. Hidelsheim: Georg Olms Verlag.) Montgomery, James. 2006. Al-Jāḥiẓʾs Kitāb al-Bayān wa-l-Tabyīn. In Writing and representation in Medieval Islam, ed. Julia Bray, 91–152. Muslim Horizons. London: Routledge. Moutaouakil, Ahmed. 1982. Réflexions sur la théorie de la signification dans la pensée linguistique arabe. Rabat: Publications de la Faculté des Lettres et des Sciences Humaines. ——. 1990. La notion dʾʿactes de langageʾ dans la pensée linguistique arabe ancienne. In Studies in the history of Arabic Grammar, vol. 2, ed. Kees Versteegh and Mike Carter, 229–238. Amsterdam: Benjamins. Owens, Jonathan. 1988. The foundations of grammar: An introduction to Medieval Arabic grammatical theory. Amsterdam: Benjamins.

Arabic Linguistic Tradition II Schaade, A. Balāgha. In Encyclopedia of Islam/Encyclopédie de lʾIslam, new edition, Vol. 1, 1012–13. Leiden: Brill. Schoeler, Gregor. 1983. Der poetische Syllogismus. Ein Beitrag zum Verständnis der ʿlogischenʾ Poetik der Araber. Zeitschrift der deutschen morgenl ändischen Gesellschaft 133: 44–92. Searle, John. 1969. Speech acts: An essay in the philosophy of language. Cambridge, UK: Cambridge University Press. ——. 1975. Indirect speech acts. In Syntax and semantics. Vol. 3, Speech acts, ed. P. Cole and J. L. Morgan, 59–82. New York: Academic Press. Simon, Udo. 1993. Mittelalterliche arabische Sprachbetrachtung zwischen Grammatik und Rhetorik: ʿilm al-maʿānī bei as-Sakkākī. Heidelberg: Heidelberger Orientverlag. Smyth, William. 1992. Some quick rules ut pictura poesis: The rules for simile in Miftāḥ al-ʿulūm. Oriens 33: 215– 229. ——. 1993. The making of a textbook. Studia Islamica 76: 99–115. ——. 1995. The Canonical formulation of ʾIlm al–balāghah and al-Sakkākīʾs Miftāḥ al-ʿulūm. Der Islam 72: 7–24. Soudan, Françoise. 1992. Lʾéloquence arabe aux premiers temps de lʾislam dʾaprès le Kitāb al-bayān wa-l-tabyīn dʾal-Ğāḥiẓ, Annales islamologiques 26: 19–46. Versteegh, Kees. 2004. Meanings of speech: The category of sentential mood in Arabic grammar. In Le Voyage et la langue, Mélanges en lʾhonneur dʾAnouar Louca et André Roman, ed. Joseph Dichy and Hassan Hamzé, 269– 287. Damascus: Institut Français du Proche-Orient. Von Grunebaum, G. Bayān. In Encyclopedia of Islam/Encyclopédie de lʾIslam, new edition, Vol. 1, 1147–1150. Leiden: Brill. ——. ʾIʿjāz. In Encyclopedia of Islam/Encyclopédie de lʾIslam, new edition, Vol. 3, 1044–1046. Leiden: Brill. ——. Faṣāḥa. In Encyclopedia of Islam/Encyclopédie de lʾIslam, new edition Vol. 2, 843–846. Leiden: Brill. Weiss, Bernard. 1976. A Theory of the parts of speech in Arabic (Noun, verb and particle): A study in ʾilm al-waḍʿ. Arabica 23: 23–36. ——. 1992. The search for Gods law. Islamic jurisprudence in the writings of Sayf al-Dīn al-Āmidī. Salt Lake City: University of Utah Press. Würsch, Renata. 1991. Avicennas Bearbeitungen der aristotelischen Rhetorik. Ein Beitrag zum Fortleben antikes Bildungsgutes in der islamischen Welt. Berlin: Klaus Schwartz Verlag. Zysow, Aron. 2008. Tamannī. If wishes were…: Notes on wishing in Islamic texts. In Classical Arabic humanities in their own terms: Festschrift for Wolfhart Heinrichs on his 65th birthday presented by his students and colleagues, ed. Beatrice Gruendler, asst. Michael Cooperson, 522–567. Leiden: Brill. (Trans. from French Peter T. Daniels.)

Notes: (1) Taqdīm and taʾkhīr (“postposing”)—that is, order, pragmatically conditioned, of the constituents of the clause— treated in a dispersed fashion in the Talkhīṭ. Conversely, it occupies a special section in the Dalāʾil (83) of Jurjānī, studied by Owens (1988). Ibn Khaldūn (Muqaddima, 1065) includes it as the first theme of his ʿilm al-balāgha (= ʿilm al-maʿānī). (2 ) Quoted by Aḥmad Ṣaqr in the introduction (11) to his edition of the ʿIʿjāz al-Qurʾān of Bāqillānī. Pierre Larcher

Arabic Linguistic Tradition II Pierre Larcher, Université de Provence

Issues in Arabic Computational Linguistics Everhard Ditters The Oxford Handbook of Arabic Linguistics Edited by Jonathan Owens Print Publication Date: Sep 2013 Online Publication Date: Dec 2013

Subject: Linguistics, Computational Linguistics, Languages by Region DOI: 10.1093/oxfordhb/9780199764136.013.0009

Abstract and Keywords This article focuses on the current state of affairs in the field of Arabic computational linguistics. It begins by briefly monitoring relevant trends in phonetics and phonology, morphology, syntax, lexicology, semantics, stylistics, and pragmatics. Then, changes or special accents within formal Arabic syntax are described. After some evaluative remarks about the approach opted for, it continues with a linguistic description of literary Arabic for analysis purposes as well as an introduction to a formal description, pointing to some early results. The article hints at further perspectives for ongoing research and possible spin-offs such as a formalized description of Arabic syntax in formalized dependency rules as well as a subset thereof for information retrieval purposes. Keywords: phonetics, phonology, morphology, syntax, lexicology, semantics, stylistics, pragmatics

9.1 Introduction At a meeting in Doha (Qatar 2011), experts discussed the challenges for Natural Language Processing (NLP)1 applied to (and, if possible, in) Arabic, concerning technologies, resources, and applications in cultural, social, educational, medical, and touristic areas, in the region concerned, for the near future. Interestingly enough, there was a consensus (by majority of votes)2 about more focus on large-scale caption of written Arabic (OCR) in view of preservation and the accessibility of the Arabic and Islamic cultural heritage; the spoken varieties of Arabic in view of the development of all kinds of conversion and answering systems (AS) to and from a standard, speech-tospeech3 (STS) as well as to speech-to-text (STT) and text-to-speech (TTS)4 conversion; and, finally, on multisimultaneous signal processing (subtitling, visualization, and instant translation),5 if possible, with event (EE) or factoid (FE) extraction for information retrieval (IR), document routing (DR), archiving purposes, mass storage, Aboutness suggestions, and different forms of retracing facilities. NLP overlaps to a large degree with computational linguistics (CL), especially when both are applied to standard Arabic or spoken varieties. The former (NLP) usually centers on the interaction between man and machine, deals with “processing” and “automated processes,” and is as exact as possible in nature. It therefore remains measurable and verifiable (NLP is eager for applications). The latter (CL) concentrates exclusively on linguistic theory and language modeling, while using any computational means it can exploit, for an adequate, coherent, and consistent linguistic description or language model. CL is usually characterized as a subsection of artificial intelligence (AI). However, I would like to underline its communicative (action and reaction) perspective against a purely cognitive environment of AI. Moreover, the communicative aspect of CL points to a reference to reality. Even when formalized and in its most abstract and logically implemented form, semantics still remains an open domain. Moreover, I would like to underline two complementary aspects of CL concerning Arabic, one the application of computer sciences to Arabic; and the

Issues in Arabic Computational Linguistics otherArabic linguistics making as much use as possible of computational as well as linguistic means and techniques. The former is more striking, and the latter is more basic. For information on relevant trends in Arabic NLP and CL, we do not need to start from scratch. General trends in NLP are adequately described in Manning and Schütze (2003) and Jurafsky and Martin (2009); 6 for Arabic NLP, see Habash (2010) and the references therein. For CL in general, one should certainly consult Bunt et al. (2004, 2010).7 Soudi et al. (2007) and Farghaly (2010) offer valuable contributions in the field of Arabic CL but also deal with Arabic NLP. Levelt (1974) is a must for formal grammars (and psycholinguistics). On (more linguistically oriented) main and subentries, there is valuable information in Versteegh (2006–2009).8 Needless to say, the Internet is always a good, if not the best, starting point for a literature search. The main topic of interest here is the current state of affairs in the field of Arabic CL. Relevant trends in phonetics and phonology, morphology, syntax, lexicology, semantics, and stylistics and pragmatics will be briefly examined. Then changes or special accents within the field of interest, namely, formal Arabic syntax, will be noted. After some evaluative remarks about the approach of this chapter, it continues with a linguistic description of MSA for analysis purposes as well as an introduction to a formal description. Some early results will be highlighted. Further perspectives are then offered for ongoing research and possible spinoffs such as a formalized description of Arabic syntax in formalized dependency rules as well as a subset thereof for IR purposes. Appendix 1 contains a list of acronyms frequently encountered in NLP and CL. Appendix 2, found only in the online version, gives a glossary of frequently-used terms in NLP and CL.

9.2 Arabic CL Defined as a statistical or rule-based modeling of natural language from a computational point of view,9 CL should always contain a linguistic dimension in the form of a specific theory combined with a descriptive model together with a formal implementation in which that linguistic theory about a natural language or its adequate, coherent, and consistent description is entered in a processing environment for analysis or synthesis purposes. Jurafsky and Martin (2009) adequately describe a modern “toolkit” for CL, but we limit ourselves here mainly to rule-based modeling by means of a relational programming algorithm using a nondeterministic formalism of two interwoven context-free grammars, resulting in a bottom-up unification-based parser for Arabic.10 Levelt (1974) provides a descriptively and didactically good introduction in the field of Formal Grammars and Psycholinguistics. Arabic CL thus combines a linguistic and a computational part. The linguistic part exploits the most recent developments in the field of adequate, coherent, and consistent language description. The formal part tests the (natural or programming) language description concerned with computational viability. Nowadays, linguistic description testing usually takes place in the framework of corpus linguistics (CoL) using large collections of authentic language data, as such serving as a reliable test bed and learning model for refining the linguistic description. The formal part of such a linguistic implementation can be tested using personal computers. Authentic text data contain both stylistic and pragmatic elements. Then we are dealing with a form of semantics, hidden in structured combinations (syntax) of lexical elements (lexicon). Relations and dependencies between elements may be underlined with morphemes (morphology) that should be accounted for in a description of the language concerned. Such a description comprises the inventory of the smallest unit of linguistic description, which is called the phoneme (phonology), or its orthographic, the grapheme (orthography). In this way, authentic (Arabic) data represent single multiple-layer syntax, distinguished in modules only for practical reasons.

9.2.1 Computational Phonetics and Phonology Here the term phonetics indicates the study of the physical properties of the smallest units of linguistic description in the Arabic language (i.e., phonemes), whether for analysis (recognition) or synthesis (generation) purposes.11 At an early stage, this study was extended with the study of its graphic counterpart, the grapheme. Later, research started on the development of all kinds of remedial support such as Arabic Braille (AB), text-to-speech, and speech-to-text conversion, or combinations thereof (for blind, deaf, and dumb). Phonology, on the other hand, is more concerned with the generalized, grammatical characterization of the Arabic phoneme and grapheme inventory of the language system. Computational phonology is the use of models in

Issues in Arabic Computational Linguistics phonological theory. For the description of the typical nonconcatenative root and pattern system of Semitic languages in general and Arabic phonology and morphology in particular, McCarthy (1981) proposes a representation of different layers, further developed by Kay (1987), Beesley (1996), Kiraz (1994, 1997), Kiraz and Anton (2000, 2001), and Ratcliffe (1998) [Ratcliffe, “Morphology”]. More recent developments go in the direction of optimality theory (OT; Prince and Smolenski 2004) and syllabification (Kiraz and Möbius 1998). There already are some specialized studies in this field on Arabic phonology [Hellmuth, “Phonology”]. Müller (2002) adds an NLP flavor with his probabilistic context-free grammars for phonology.12 Computational phonology is the basis of many NLP applications, such as the previously mentioned AS systems and STT, TTS, and STS conversion and others such as Arabic speech recognition (SR), OCR, and text-to-text conversion (TTT) or machine translation (MT).13 In computational phonetics and phonology, we are confronted with terms such as tiers, distinct layers of representation (two-four, three), finite-state automata (FSA or FSM), transducers, programming languages, tables, tagging, ± deterministic, and a few other technical terms. Sometimes, a decisive discussion about progression in the field of research concerned is worded using general (Arabic) linguist terms, but for certain entries in Appendix 9.2 it was necessary to employ less frequently used linguistic terms.

9.2.2 Arabic Computational Morphology As Richard Sproat correctly mentioned (Soudi et al. 2007: viii), Kaplan and Kay (1981) and Kay (1987), in line with Koskenniemi (1983), paved the way for Kenneth Beesley’s (1989) research on Arabic computational morphology, which led to the work of many others as well as to the development of applications in the field of Arabic morphology. One of the pioneers in computational Arabic morphology, Tim Buckwalter, developed BAMA, an Arabic morphological analyzer (Buckwalter 2010). Initially, his research was oriented toward automated corpus-based Arabic lexicology. Later, three lexicons, compatibility scripts, and an algorithm in the feature-rich dynamically typed14 programming language called Perl were combined in a software package for the morphological analysis of Arabic words (Buckwalter 2002, 2004), used, inter alia, for morphological and part-of-speech (POS, part of speech) tagging as well as for syllabification of authentic data in existing Arabic Treebanks15 (Maamouri and Bies 2010; Smrž and Hajič 2010) for morphological or syntactic annotation. It is not surprising that research on Arabic computational morphology is easily adopted, adapted, and incorporated into general approaches to computational phonetics, phonology, and morphophonemics. Al-Sughaiyer and AlKharashi (2004) classify a number of Arabic morphological analyzers (analyzers) and synthesizers (generators) according to the approach employed regarding table lookup, linguistic (two-level, FSA or FSM, traditional applications), combinatorial, and pattern-based approaches. As Köprüand Miller (2009) point out, “Very few of the available systems are evaluated using systematic and scientific procedures.” This is perhaps a bit too harsh a criticism. However, it is always worthwhile to scrutinize and evaluate the advantages and disadvantages as well as the adequacy, coherency, and consistency of a chosen approach. Evaluating 20-odd Arabic morphological analyzers and synthesizers, Al-Sughaiyer and Al-Kharashi (2004: 198, Table 4) mention their algorithm name and type: some “brand” names and even one “Sebawai”; 16 many “linguistics”; and one “rule based.” Smrž (2007: 5–6) qualifies Beesley (2001), Ramsay and Mansur (2001), and Buckwalter (2002) as “lexical” in nature. Habash (2004) calls his own work “lexical-realizational” in nature. Finally, Cavalli-Forza et al. (2000), Habash et al. (2005), Dada and Ranta (2006), and Forsberg and Ranta (2004) are rather “inferential-realizational.” For his ElixirFM, Smrž (2007: 2) emphasizes its implementation within the Prague framework of function generative dependency (FGD) in functional programming (Haskell), contrasting with the dynamic programming (Perl) of Buckwalter (2002) and resulting in “a yet more refined linguistic model.” Partly based on the operational tagging system of Buckwalter’s BAMA morphological analyzer for Arabic, Otakar Smrž developed “description of [Arabic] surface syntax in the dependency framework” (Smrž and Hajič 2010). This brings us to the doorstep between Arabic phonology–morphology and Arabic computational syntax at least as far as the representation of the analysis results in the form of dependency trees is concerned. These results are

Issues in Arabic Computational Linguistics obtained on the basis of a pretagged corpus. The Prague linguists opted for a functional dependency grammar approach. Nonetheless, also for the computational description of Arabic morphology and syntax, a programming language, Haskell,17 is being used.18 There is an important difference between the use of a programming language and a formalism for implementable and operational descriptions of a natural language.19

9.2.3 Arabic Computational Syntax Syntax is the description of the overall organization of a natural language in which different complementary building blocks such as phonology, morphology, lexicology, semantics, stylistics and pragmatics come together to convey a particular message between an A and a B. To describe the general structure of this organization for natural languages in general or for a specific language such as Arabic is the objective of the linguistic part of the description. To find an implementable formal model for such a description is the objective of the computational part of that description.

9.2.3.1 Linguistic Part For a historical overview of language description, I refer to HSK 18.3 (2006). Here we limit ourselves to the century of the dominance of immediate constituency (IC) and the rise of many other linguistic theories and descriptive models such as dependency grammar, of importance or used for the linguistic description of Standard and spoken Arabic varieties. It is evident that the splitting up of a natural language system into its largest and smallest units of linguistic description, as well as the description of mutual relationships and dependencies between these units, form an excellent starting point for any research on the fundamentals of human communication in general, and of the organisation of a specific natural language system in particular (Habash 2010: chapter 6). Computational linguistics (cf. Winograd 1983) started with the annotation (POS tagging) of formal (e.g., parts of speech; word and phrasal categories, sentences, sections, chapters, volumes, and the marking of nontextual insertions); 20 functional elements (e.g., cases, clitics, determiners, heads and modifiers, slots and fillers) in authentic text data (CoL); and continued later in the presentation of derivation trees or labeled bracketing, extracted from this (earlier inserted) information.

9.2.3.2 Formal part For a historical overview of computational language description in general, I refer to Winograd (1983). Here we speak about the current state of syntactic parsing of Arabic text data wherein different steps can be distinguished. Usually, they are labeled with terms such as tokenization, diacritization, POS tagging, morphological disambiguation (Marton et al. 2010), base phrase chunking, semantic role labeling, lemmatization, stemming, and the like (cf. Appendix 9.2; cf. also Mesfar 2010). Most of these processes have been automated by now, but all the existing collections of syntactically analyzed Arabic text data (Habash 2010: section 6.2) such as the Penn Arabic Treebank (Maamouri et al. 2004), the Prague Arabic Dependency Treebank (Hajič et al. 2001), and the Columbia Arabic Treebank (Habash and Roth 2009) have been manually checked. This “forest of treebanks” (Habash 2010: 111) can now be used as learning models for the development of new statistical parsers, evaluating parsers and general Arabic parsers.

9.2.4 Arabic Computational Lexicology Computational lexicology is the branch of linguistics, which is concerned with the use of computers in the study of machine-readable dictionaries (lexicon). Sometimes this term is synonymous with computational lexicography, though the latter is more specifically for the use of computers in the construction of dictionaries (Al-Shalabi and Kanaan 2004).21 Piek Vossen, a well-known computational lexicologist, founder and president of the Global Wordnet Association, worked on the first WordNet project (Fellbaum 1998) and supervised parallel projects such as EuroWordNet (Vossen 1998) and Arabic WordNet (Black et al. 2006; Elkateb et al. 2006). He is thinking in terms of (multi)lingual lexical databases with lexical semantic networks. We come close to a distinction in form, function, meaning, and contextual realization of a lexical entry. Besides this distinction we always have the linguistic and the formal part.

Issues in Arabic Computational Linguistics Linguistic part Relevant here are studies such as those of Dévényi et al. (1993) on Arabic lexicology and lexicography as well as other research with valuable bibliographical references (Bohas 1997; Bohas and Dat 2008; Hassanein 2008; Hoogland 2008; Seidensticker 2008). Moreover, one should include the studies about affixes, features (Dichy 2005; Ditters 2007), or parameters hinting at theta, thematic, or semantic roles.

Formal part On the formal side I would like to mention the tag sets (Habash 2010: 79–85; Maamouri et al. 2009; Diab et al. 2004; Diab 2007; Kulick et al. 2006; Habash and Roth 2009; Khoja et al. 2001; Hajič et al. 2005) used for the annotation of the corpora of Arabic text data as well as the by then enriched corpora (treebanks) from which all kinds of relevant information can be extracted. Here should also be included studies on Arabic semantic labeling (Diab et al. 2007) and Arabic semantic roles (Diab et al. 2008).

9.2.5 Arabic Computational Semantics, Stylistics, and Pragmatics Computational syntax, at the academic level, is still not common practice. Computational semantics, stylistics, and pragmatics are even at a more rudimentary stage,22 not only as far as the Arabic language is concerned but also even for more intensively studied natural languages. It is worthwhile here to refer to the HSK volumes on dependency and valency (HSK 25, 2003–2006), and in particular to contributions of interest for our discussion23 (Owens 2003; Msellek 2006; Bielický and Smrž 2008, 2009). According to Wikipedia: 24

Click to view larger Figure 9.1 An example from the Arabic Propbank (Habash 2010, 115).

Computational semantics is the study of how to automate the process of constructing and reasoning with meaning representations of natural language expressions. It consequently plays an important role in natural language processing and computational linguistics. Some traditional topics of interest are: semantic analysis, semantic underspecification, anaphora resolution, presupposition projection, and quantifier scope resolution. Methods employed usually draw from formal semantics or statistical semantics. In a note on Arabic semantics, Habash (2010: chapter 7) mentions the Arabic Proposition Bank (Propbank) and Arabic WordNet. A propbank (Palmer et al. 2005) is, in contrast with a treebank, a semantically annotated corpus. On the basis of predicate– argument information of the complement structure of a verb so-called frameset definitions (Baker et al. 1998) are associated with each entry of the verbal lexicon (Palmer et al. 2008). The description of the nature, number, and role of the arguments can be as detailed and specific as a linguistic description of the semantics of a language allows. It may be clear that here lies the greatest challenge in the development of adequate, coherent, and consistent parsers for any natural language text data, MSA, and spoken varieties of Arabic included. Figure 9.1 presents information about the linguistic unity of description (S), type of sentence (NP-TPC1), with a verb phrase (VP) as comment. The topic (ARG0) is realised by a noun phrase (NP). The comment, which may be termed “predicate” is realized by a finite transitive verb (PRED) with an implicit subject (ARG0), a direct object noun phrase (ARG1), and a prepositional time adverbial (ARGM-TMP). There is a form of subcategorization at the phrasal level (NP-TPC1, NP, NP-SBJ1, NP-OBJ). Nouns are by subscripts divided into common nouns and subcategories. There is some form of description in terms of functions and categories, but it is not maintained in a consistent and coherent way until the final lexical entries have been reached.

Issues in Arabic Computational Linguistics Arabic WordNet has been mentioned earlier in relation to computational lexicology (§1.4). Here I want to explicitly underline the importance of electronically available collections of text data and nowadays also parallel corpora, for linguistic research (Col = corpus linguistics). In the following section I defend a purely linguistic approach of Arabic language description exploiting as much computational means as possible on the basis of authentic data and within a long-standing linguistic Arabic grammatical tradition. Other studies in the fields of semantics,25 dialogue (Hafez 1991), discourse,26 stylistics (Mitchell 1985; Mohammad 1985; Somekh, 1979, 1981), and pragmatics27 remain mainly theoretical. Little can be found on heuristics. Something like a dialogue act annotation system allowing the ranking of communicative functions of utterances in terms of their subjective importance (Włodarczak 2012) for Arabic has still to be written. Computational semantics has points of contact with the areas of lexical semantics (word sense disambiguation and semantic role labeling), discourse semantics, knowledge representation, and automated reasoning (in particular, automated theorem proving).

9.3 A Formalized Linguistic Description of Arabic Syntax There are, in my opinion, some basic concepts and rules important for long-term linguistic research. The first point is that syntax encompasses a number of subfields, including phonology, morphology, lexicology, semantics, stylistics, pragmatics, and heuristics, together with their respective branches, including the use of computational tools. Moreover, linguistic research should positively and negatively improve the field: positively in the sense of enriching the discipline as well as socially relevant; and negatively in the sense of convincingly demonstrating that a specific approach did not and will not lead to any useful results or meaningful research. A rather important rule is that any account of linguistic research, with some additional information and footnotes, should be readable and understandable foras well as verifiable by colleagues. Finally, scientific linguistic research should not be presented encoded in machine language or a programming language printout or even in PDF form and, moreover, should not be superficial, as are many of the presentations of commercial researchers and product developers. The description of the syntactic structure28 of Standard Arabic, readable and understandable foras well as verifiable by colleagues, may have the form of a hypothesis, to be tested against authentic language data. After refining and renewed testing, this leads to a theory about the syntactic structure of the same layer of data of Standard Arabic as tested in the data. The same approach can be followed by further research on other Arabic text data. Earlier, a listing was made of useful and (moreover) operational computational instruments, including machinereadable resources of all kinds, for the automated linguistic research on Arabic. We discussed the difference between Arabic NLP and Arabic CL, underlining the independence of the linguistic and formal parts in this research, while acknowledging a bias in favor of the linguistic part. Here I will defend an approach to an adequate, consistent, and coherent description of, in this case, MSA for the automated analysis29 of authentic Arabic text data. First, we position this section in Arabic NLP history (9.3.1). Then I say something about linguistic and formal concepts for language description within the Arabic grammatical tradition (9.3.2). Finally, I present a sample of a linguistic (9.3.3) and a formal part (9.3.4) of a description of MSA. Finally (9.4), I say something about perspectives on the basis of options chosen.

9.3.1 Evaluating Remarks about the Approach Opted for When Smrž (2007: 68) says, “The tokens30 with their disambiguated grammatical information enter the annotation of analytical syntax,” we are in the linguistic part of our discussion about computational Arabic syntax. The same is the case with Topologische Dependenz-grammatik fürs Arabische (Odeh 2004). In both, the results of the analysis of some interesting syntactic peculiarities of the Arabic language, processed in a language-independent dependency-oriented environment (Debusmann 2006), are presented in the form of unambiguous, rather very nice tectogrammatical dependency trees on the basis of an analytical representation in the case of Smrž (2007), and, except for the labeling, in almost identical ID (immediate dominancy) and LP (linear precedence) representations 31

Issues in Arabic Computational Linguistics (Odeh 2004).31

Figure 9.2a Tectogrammatic representation of the analysis of a sentence (Smrž 2007b, 73).

The sentence can be interpreted as containing a predicate (PRED) with an agent (ACT), no expressed addressee (ADDR), and an object (PAT). This object consists of two coordinated topics (ID), both further specified by an attributive modifier (RSTR). The second modifier does not have an agent but does govern an object (PAT). A positional apposition (LOC) plays the role of sentence adverbial. The second part of Figure 9.2b lists the Arabic word and its English translation as well as the tags used for the analytical representation (column 1). Column 2 lists the values for some variables used in the analysis. The third column (3) gives the values (in upper case) and representation of the variables (in lower case) I use in my approach of two-level description of the same target language. Odeh (2004: Figure 9.3) presents the ID and the LP representation of a sentence with a finite verb form in first position. The abbreviations in Figure 9.3a are self-evident. Those in Figure 9.3b refer to the topological fields: sentence field (sf) and article field (artf).32 Notwithstanding the vagueness of the dismissal (Smrž 2007: 6), we like to comment on:

Figure 9.2b The analyzed sentence:

‘and

in the section on literature, the magazine presented the issue on the Arabic language and the dangers that threaten it.’ (Smrž 2007b, 72–73).

The outline of formal grammar (Ditters, 2001), for instance, works with grammatical categories like number, gender, humanness, definiteness, but one cannot see which of the existing systems could provide for this information correctly, as they misinterpret some morphs for bearing a category, and underdetermine lexical morphemes in general as to their intrinsic morphological functions.33 This is a correct remark, as far as Ditters (2001) is concerned. I am working with a description in terms of grammatical functions and categories, final lexical entries, dependency relations, and, additionally, an opening toward a description of semantic features as well.34

Issues in Arabic Computational Linguistics Grammatical in this context involves, as said earlier, the phonological, morphological, structural, and lexical modules for language description with rudimentary extensions to semantics, stylistics, and pragmatics as well as. It is necessary to remain understandable for and verifiable by colleagues. Serious semantic extensions are awaiting further computationally more coordinated research under supervision of the linguistic twin part in this kind of research. Let us continue with some words about the outline of formal grammar.

Figure 9.3a Immediate dominance (ID) representation.

Figure 9.3b Linear precedence (LP) representation.

Computational means can be used to test a hypothesis about the linguistic structure of MSA sentences 35 in an efficient but linguistically understandable wording of nonterminals and terminals, applying only a context-free grammar formalism (with room for some additional context-free layers in the second level of description for semantics) but considering the availability of compilers for one, two, or more levels of context-free attribute grammar formalisms. Until now this approach has proved to be promising.

9.3.2 Linguistic and Formal Concepts about Language Description in the Arabic Grammatical Tradition Immediate constituency (IC) has dominated descriptive linguistics for a long time. Dependency grammar (DG) concepts were already (according to Carter 1973, 1980; Owens 1988) familiar to Arab and Arabic grammarians and became a welcome, and needed, addition to an implementable descriptive power for natural languages in general and MSA in particular. Moreover, one can always explore other valuable suggestions. The basis for the description of parts of speech, functions, and categories in phrasal categories in MSA can be found in the Kitāb of Sībawayhi (d. 798).36 Carter (1973: 146)37 calls it “a type of structuralist analysis unknown to the West until the 20th century.” He ends his abstract with: Utterances are analysed not into eight Greek-style “parts” but into more than seventy function classes. Each function is normally realized as a binary unit containing one active ‘operator’ (the speaker himself or an element of his utterance) and one passive component operated on (not “governed”) the active member of the unit. Because every utterance is reduced to binary units, Sībawayhi’s method is remarkably similar to immediate constituent Analysis, with which it shares both common techniques and inadequacies, as is shown. As a first example of such a class of functions, Carter (1973: 151) presents the triad of “grammatical effect” (ʿamal), comprising a “grammatically affecting” (ʿāmil) and a “grammatically affected” ( maʿmūl) component. Similar triads (possibly considered as dependency rules) can be drafted from the other function classes listed. Moreover, the function classes could be arranged in subcategories, for example, accounting for relationships between constituents, sentence types, or word formation. Other function classes deal with phonology and morphology, have discourse functions, or are related to stylistics. Finally, on the basis of Sībawayhi’s comprehensive (exhaustive?) description of Arabic syntax, the function classes could easily be extended with more of this kind of “dependency rules,” for example, with triads accounting for transitivity, or the subject–object– predicate relations.

Issues in Arabic Computational Linguistics Owens’ (1988) historical overview with an apparent DG related perspective is broader. In section 2, all the relevant issues are discussed: constituency (IC) in Arabic theory (§2.9); dependency (DG) (§2.9.2); and dependency in Sībawayhi (§2.9.3). There are also chapters about markedness in Arabic theory (§8) and syntax, semantics, and pragmatics (§9).

9.3.3 Linguistic Description of MSA for Analysis Purposes38 An ideal compromise between IC and DG seems to be, for me, a context-free grammar description of MSA in IC terms accounting for the horizontal sequential order, enriched with a second, context-free grammar level attached to the nonterminals (Appendix 9.2) of the first level, accounting for relationships and dependencies (DG) in the vertical relational order between elements of a constituent or between constituents of one of the two sentence types (nominal and verbal). A third context-free grammar layer, only describing semantic extensions and properties, proved to be locatable (for the moment) within the two-layer context-free grammar frame.39 Working with two semantically and therefore also syntactically different sentence types in MSA, for both types one can distinguish categorial and functional nonterminals, categorial and functional terminals, lexical terminals, dependencies, and relationships between elements of a constituent and between constituents at sentence level. The syntactic consequences of a semantic property of a lexical entry will, of course, be of importance for the analysis and generation of text data of the language concerned.40 The syntax of sentence types in MSA, Sn , and Sv is described by means of alternating layers of categories, functions, and categories until terminals (here: lexical entries) have been reached. In the Sn , we distinguish two obligatory slots: a topic and a comment. The filler for a topic function characteristically belongs to the class of N’s (here: head) or to the category NP’s (optionally also containing modifiers to the head). In the Sv, we are dealing with a single obligatory slot, a predicate. The filler of a predicate function typically belongs to the class of V’s (head) and is usually realized by a VP (optionally also containing modifiers to the head). The comment function in the Sn as well as optional slots, in both the Sn and the Sv, may be filled with entries of the different classes or with different phrasal categories. In line with the first words of the Kitāb of Sībawayhi, as parts of speech we distinguish elements of the open word classes nouns (N) and verbs (V) and the closed class of particles (P).41 Elements of these classes realize the head function in phrasal constituents such as noun phrases (NP), verb phrases (VP), and particle phrases (PP) as well as properties of elements of the word classes (N, V, and P), including morphological, syntactic, and semantic features (valency indications) and some pragmatic ones as well.42 The following section introduces a simple sample implementation.

9.3.4 The Formal Description of MSA For an adequate, coherent, and consistent description of MSA I examined different linguistic theories and models (Ditters 1992) for the best products, implementable in a processing environment for analysis purposes. For example, generalizations in the form of transformations in a transformational generative framework (TG) are too powerful a tool for an overall linguistic description;, when a simple machine like a computer failed to decide in a finite time whether or not a certain TG structure was described in the formal implementation of that linguistic description, I opted for a different formalism (AGFL)43 to implement a description (Ditters 2001, 2007, 2011) in terms of a combination of IC and DG. In contrast to a programming language (dynamic, functional, or relational in nature), a formalism (Ditters 1992: 134) is exclusively for the static, nondeterministic, and declarative description of structures such as programming languages (e.g., Algol-68, Affix- Attribute-Feature- Logical Grammars), also suited for the description of natural languages (e.g., Arabic, Dutch, English, Hebrew, Latin, Spanish). The objective is to test such a hypothetical description of, in our case, lMSA syntax structure on real data, for example, an Arabic text corpus. It is the machine-readable text data that determine whether a match should occur or not. Once tested, corrected, and refined, the hypothesis has become a theory, certainly for the language structure represented in the test bed in the form of a new hypothesis to be tested on other Arabic text corpora. Briefly, AGFL is an interwoven two-level context-free grammar formalism44 with almost context-sensitive properties.45

Issues in Arabic Computational Linguistics On the Chomskian hierarchy of grammars (Levelt 1974) scale, context-free grammars are, for the description and automated testing of natural language descriptions, really rather nonproblematic. As a matter of fact, Chomsky qualified a context-free grammar as an inadequate descriptive tool for natural languages. However, he never showed, as far as I know, any interest in a combination of two (or even more) context-free grammars (with an almost context-sensitive descriptive power),46 enough to describe, for example, most of Standard Arabic, including at least parts of its semantic richness. Furthermore, he never tested, as far as I know, his linguistic hypotheses and theoretical models practically by computational means. A rule-based context-free grammar rewrites one single nonterminal at the left-hand side into one or more nonterminals or lexical entries on the right-hand side. AGFL, successor of EAG (Appendix 9.1), is a formalism for the description of programming languages or natural languages in which large context-free grammars can be described in a compact way. Along with attribute grammars47 and DCGs, AGFLs belong to the family of two-level grammars, the first, context-free level, which accounts for the description of sequential word order of surface natural language elements, is augmented with set-valued features for expressing agreement between constituents and between elements of a constituent as well as linguistic properties (including semantic features). AGFL is implemented in CDL3 and C.48 Notational AGFL conventions include the rewrite symbol (:), the marking of alternatives (;), the separation of sequences (,), the end of a rule (.), the layout of nonterminals and terminals of the first level and of the nonterminals and final values of variables of the second level of description in terms of lower- and uppercase representation. Besides that, there is no longer any capacity problem for the storage of electronic data. Therefore, the choice of names and terminal values for the elements of the first and second level of description of, for example, MSA, may be as linguistically recognizable as one prefers. We use four types of rules within the AGFL formalism: the so-called hyperrules; metarules; predicate rules; and lexical rules: 49 • Hyperrules formally describe the occurrence of elements of word classes, in single or phrasal form. Variation in the sequence of those elements is dealt with by the formalism. • Metarules define the nonterminals of the second level of description to a finite set of terminal values. • Predicate rules describe and, if needed, condition relationships and dependencies between phrasal constituents or between elements of a constituent. • Lexical rules describe final or terminal values of the first level of description, if possible with semantic features, some colocational information (including additional remarks about nonregular and unexpected occurrences in compositional semantics). However, as is well-known, natural languages go further than the addition of the meaning of individual elements for capturing the real meaning of the linguistic data concerned.50 In Figure 9.4, Jaszczolt (2005) illustrates the process of utterance interpretation within the compositional theory of default semantics. It may be clear that the meaning of an utterance does not equal the sum of the meaning of its constituents.

Issues in Arabic Computational Linguistics

Figure 9.4 Utterance interpretation in default semantics (Jaszczolt 2005, 73).

The following presents a sample grammar51 of a two-level context-free rewrite AGFL52 for Modern Standard Arabic: 53 GRAMMARsentence. ROOT sentence(topic). # 1 Meta rules # These define the finite domain of values for non-terminals of the second level of description. # This second level enables accounting for relationships and dependencies. CASE::acc|gen|nom|invar. DECLEN::defec|dipt|CASE. DEFNESS::def|indef. GENDER::fem|masc. HEADREAL::com|pers|prop. MODE::nominal|verbal. MOOD::imper|indic|juss|subj. NUMBER::coll|dual|PLUR|sing. ORDER::topic|elliptic_topic. PERSON::first|second|third. PLUR::explu|inplu. TENSE::perfect|MOOD. TYPES::direc|finalintr|place|timeprep. VOICE::active|passive. # 2 Hyper rules # They describe functions and categories of non-terminals at the first level of description, # until lexical values have been reached. This level enables accounting, in an efficient way, # for the generalization of word order and sentence structure.54 sentence(topic): topic(GENDER,NUMBER), topic comp(GENDER,NUMBER). topic(GENDER,NUMBER): nounphrase(def,GENDER,NUMBER,PERSON,nom|invar); prep(finalintr), np(HEADREAL,def,GENDER,NUMBER,PERSON,gen); bound prep(finalintr) + np(HEADREAL,def,GENDER,NUMBER,PERSON,gen).

Issues in Arabic Computational Linguistics topic comp(GENDER,NUMBER): predicate(MODE,DEFNESS,third,GENDER,NUMBER); np(HEADREAL,DEFNESS,GENDER,NUMBER,PERSON,nom); adjp(DEFNESS,GENDER,NUMBER,CASE); ap; pp. nounphrase(def,GENDER,NUMBER,PERSON,nom|invar): np(HEADREAL,def,GENDER,NUMBER,PERSON, nom|invar). np(HEADREAL,def,GENDER,NUMBER,PERSON,CASE): head(HEADREAL,DEFNESS,GENDER,NUMBER,PERSON,CASE), DEF is(HEADREAL,DEFNESS). predicate(verbal,DEFNESS,third,GENDER,NUMBER): vp(TENSE,PERSON,GENDER,NUMBER). predicate(nominal,DEFNESS,PERSON,GENDER,NUMBER): np(HEADREAL,DEFNESS,GENDER,NUMBER,PERSON,nom), headreal is(HEADREAL). head(HEADREAL,DEFNESS,GENDER,NUMBER,PERSON,CASE): noun(DECLEN,GENDER,NUMBER). noun(DECLEN,GENDER,NUMBER): common noun(DECLEN,GENDER,NUMBER); pers pronoun(GENDER,NUMBER,PERSON,CASE); proper noun(DECLEN,DEFNESS,GENDER,NUMBER). vp(TENSE,PERSON,GENDER,NUMBER): verb(TENSE,VOICE,PERSON,GENDER,NUMBER). # 3 Predicate rules # They describe, or even determine, relations and dependencies between values of elements # of the second level of description by means of the conditioning of specific values. # This type of rules sometimes is called “ empty rules.” # DEF is(HEADREAL,DEFNESS). DEF is(com,DEFNESS):. DEF is(pers,def):. DEF is(prop,def):. #Headrealis(HEADREAL). headreal is(com):. headreal is(pers):. headreal is(prop):.

Issues in Arabic Computational Linguistics # 4 Lexical rules # They describe the final or lexical value of entries in the lexicon. # Adjp(DEFNESS,GENDER,NUMBER,CASE) and ap. “adjp” adjp(DEFNESS,GENDER,NUMBER,CASE) “ap” ap # Common noun(norm,masc,sing), including pronouns and proper nouns. “raǧul”

common noun(norm,masc,sing)

“riǧāl”

common noun(norm,masc,inplu)

“bint”

common noun(norm,fem,sing)

“banāt”

common noun(norm,fem,inplu)

# Perspronoun(fem|masc,sing,first,nom). “ʾanā”

perspronoun(fem|masc,sing,first,nom)

“naḥnu”

pers pronoun(fem|masc,inplu,first,nom)

“ī”

pers pronoun(fem|masc,sing,first,gen|acc)

“nī”

pers pronoun(fem|masc,sing,first,gen|acc)

“nā”

pers pronoun(fem|masc,inplu,first,gen|acc)

# Proper noun(dipt,def,masc,sing). “ʾaḥmad”

proper noun(dipt,def,masc,sing)

“muḥammad

proper noun(norm,def,masc,sing)

“fātima ”

proper noun(norm,def,fem,sing)

# Prep(TYPES). “la”

bound prep(finalintr)

“la”

prep(finalintr)

“pp”

# Verb(TENSE,VOICE,PERSON,GENDER,NUMBER).

Issues in Arabic Computational Linguistics

“kataba”

verb(perfect,active,third,masc,sing)

“kutiba”

verb(perfect,passive,third,masc,sing)

“yaktubu”

verb(indic,active,third,masc,sing)

“yuktabu”

verb(indic,passive,third,masc,sing)

# Remark 55

9.4 Perspectives for Further Linguistic and Formal Research on Arabic Syntax In a joint contribution to the Nemlar conference (Ditters and Koster 2004),56 we explored the potentiality of the existing approach to MSA syntax for other, socially equally relevant, spinoffs of my description of Arabic for corpus-linguistic analysis purposes. I prefer to finish testing the current implementation hypothesis about MSA syntax on Arabic text data. Second, I should like to refine the verified theory by means of a formal description of MSA syntax for generative purposes. Dependency grammar worded implementation could be extracted from such research. Finally, I like to make an, within the AGFL processing environment, implementable subset of DG rules for research on aboutness57 in Arabic text data. Symbol

Meaning

Aboutness

Arabic Braille

ABP

Arabic proposition bank

Automatic correction

AGFL

Affix grammar over finite lattices

Artificial intelligence

Answering system

ASL

Arabic Sign Language

ASR

Automatic speech recognition

Base phrase

CALL

Computer-assisted language learning

CFG

Context-free grammar

Computational linguistics

CoL

Corpus linguistics

Data mining

Issues in Arabic Computational Linguistics DM Symbol DR

Meaning Document routing

EAG

Extended affix grammar

EBL

Example-based learning cf. MBL

Event extraction

Factoid extraction

FGD

Function generative dependency

FSA

Finite state automaton

FSM

Finite state machine

FST

Finite state transducer

Handwriting recognition

IBL

Instance-based learning cf. MBL

Immediate constituency

Immediate dominancy

Information retrieval

Lazy learning cf. MBL

Language modeling

Linear precedence

MBC

Morphological behavior class

MBL

Memory-based learning

MFH

Morphological form harmony

MLA

Machine learning approach

MSA

Modern Standard Arabic

Machine translation

(S)MT

(Statistical) MT

NER

Named entity recognition

NET

Named entity translation

Issues in Arabic Computational Linguistics NET

Named entity translation

Symbol NL

Meaning Natural language

NLP

Natural language processing

OCR

Optical character recognition

Optimality theory

Orthographic normalization

POS

Parts of speech

POS-T

POS tagging

QAS

Question answering system

RBA

Rule-based approach

Statistical approach

SBA

Stem-based approach

Spelling correction

Speech generation

SLA

Supervised learning approach

Signal processing

Speech processing

Speech recognition

STS

Speech-to-speech

STT

Speech-to-text

SVM

Support vector machines

Text categorization

TDT

Topic detection and tagging

Text generation

Text mining

Text processing

TTS

Text-to-speech

Issues in Arabic Computational Linguistics TTS

Text-to-speech

Symbol TTT

Meaning Text-to-text

ULA

Unsupervised learning approach

Selected Bibliography58 Al-Shalabi, Riyad, and Ghassan Kanaan. 2004. Constructing an automatic lexicon for Arabic language. International Journal of Computing and Information Sciences 2: 114–128. Al-Sughaiyer, Imad A., and Ibrahim A. Al-Kharashi. 2004. Arabic morphological analysis techniques: A comprehensive survey. Journal of the American Society for Information Science and Technology 55(3): 189–213. Baker, Collin, Charles Fillmore, and John Lowe. 1998. The Berkeley FrameNet project. (COLING-ACL’98): Proceedings of the University of Montréal Conference, 86–90. Beesley, Kenneth R. 1989. Computer analysis of Arabic morphology: A two-level approach with Detours. In Perspectives on Arabic linguistics III: Papers from the 3rd annual symposium on Arabic linguistics, ed. Bernard Comrie and Mushira Eid, 155–172. Amsterdam: John Benjamins. ——. 1996. Arabic finite-state morphological analysis and generation. In Proceedings of the 16th international conference on computational linguistics (COLING-96), Copenhagen, Denmark, 89–94. Copenhagen: Center for Sprogteknologi. ——. 2001. Finite-state morphological analysis and generation of Arabic at Xerox research: Status and plans in 2001. In Proceedings of the EACL workshop on language processing: Status and prospects. Toulouse, France, 1– 8. Available at: www.xrce.xerox.com/content/download/20547/147632/file/ Bielick ý, Viktor, and Otakar Smrž. 2008. Building the valency lexicon of Arabic verbs. In Proceedings of the 6th international conference on language resources and evaluation (LREC), Marrakech, Morocco. ——. 2009. Enhancing the ElixirFM lexicon with verbal valency frames. In Proceedings of the 2nd international conference on Arabic language resources and tools, ed. Khalid Choukri and Bente Maegaard. Cairo: The MEDAR Consortium. Black, William, Sabri Elkateb, Horacio Rodriguez, Musa Alkhalifa, Piek Vossen, Adam Pease, and Christiane Fellbaum. 2006. Introducing the Arabic WordNet project. In Proceedings of the 3rd international WordNet conference, Jeju Island, Korea. Bohas, Georges. 1997. Matrices, étymons, racines: Éléments d’une theorie lexicographique du vocabulaire arabe. Leuven: Peeters. Bohas, Georges, and Mihai Dat. 2008. Lexicon: Matrix and etymon model. In EALL vol. 3, ed. Kees Versteegh, Associate Editors: Mushira Eid, Alaa Elgibali, Manfred Woidich, Andrzej Zaborski, 45–52. Leiden: Brill. Buckwalter, Timothy. 2002. Arabic morphological analyzer version 1.0. Philadelphia: Linguistic Data Consortium. ——. 2004. Issues in Arabic orthography and morphology analysis, in Proceedings of the workshop on computational approaches to Arabic script-based languages, (COLING 2004), Geneva, Switzerland. ed. M. Farghaly and K. Megerdoomian, 31–34. Stroudsburg: Association for Computational Linguistics. ——. 2010. The Buckwalter Arabic morphological analyzer, in ed. Farghaly, 85–101. Bunt, Harry, John Carroll, and Giorgio Satta (eds.). 2004. New developments in parsing technology. Dordrecht: Kluwer Academic Publishers. Bunt, Harry, Paola Merlo and Joakim Nivre. 2010. Trends in parsing technology. Springer: Dordrecht.

Issues in Arabic Computational Linguistics Carter, Michael G. 1973. An Arab grammarian of the eight century A.D.: A contribution to the history of linguistics. Journal of the American Oriental Society 93: 146–157. ——. 1980. Sibawayhi and modern linguistics. Histoire Épistémologique 2(1): 21–26. Cavalli-Sforza, Violetta, Abdelhadi Soudi, and Teruko Mitamura. 2000. Arabic morphology generation using a concatenative strategy. In Proceedings of the 1st meeting of the North American Chapter of the Association for computational linguistics (NAACL 2000), Seattle, WA, 86–93. Dada, Ali, and Aarne Ranta. 2006. Arabic resource grammar. In Proceedings of the Arabic language processing conference (JETALA), Rabat, Morocco: IERA. Debusmann, Ralph. 2006. Extensible dependency grammar: A modular grammar formalism based on multigraph description. PhD diss., Saarland University. Dévényi, Kinga, Tamás Iványi, and Avihai Shivtiel (eds.). 1993. Proceedings of the colloquium on Arabic lexicology and lexicography (C.A.L.L.): Budapest 1–7 September. Part one: Papers in European languages. Part two: Papers in Arabic. Budapest: Eötvös Loránd University Chair for Arabic Studies. Diab, Mona T. Kadri Hacioglu, and Daniel Jurafsky. 2004. Automatic tagging of Arabic text: From raw text to base phrase chunks. In Proceedings of the 5th meeting of the North American chapter of the Association for computational linguistics/human language technologies conference (HLT-N/LICL04), Boston, MA, 149–152. ——. 2007. Towards an optimal POS tag set for modern standard Arabic processing. In Proceedings of recent advances in natural language processing (RANLP), Borovets, Bulgaria. Diab, Mona, Alessandro Moschitti, and Daniele Pighin. 2008. Semantic role labeling systems for Arabic language using Kernel methods. In Proceedings of ACL-08: HLT, Columbus, Ohio, 798–806. Diab, Mona, Musa Alkhalifa, Sabry ElKateb, Christiane Fellbaum, Aous Mansouri, and Martha Palmer. 2007. Arabic Semantic Labeling. In Proceedings of the 4th International Workshop on Semantic Evaluations (SemEval-2007 18), Prague, Czech Republic, 93–98. Dichy, Joseph. 2005. Spécificateurs engendrés par les traits [±anim é], [±humain], [±concret] et structures d’arguments en arabe et en français. In De la mesure dans les termes. Actes du colloque en hommage à Philippe Thoiron, ed. Henri Bé joint and François Maniez, 151–181. Travaux Centre de Recherche en Terminologie et Traduction (CRTT), Lyon: Presses Universitaires de Lyon. Ditters, Everhard. 1992. A formal approach to Arabic syntax: The noun phrase and the verb phrase. PhD diss., Nijmegen University. ——. 2001. A formal grammar for the description of sentence structure in modern standard Arabic. In EACL 2001 Workshop Proceedings on Arabic language processing: Status and prospects, Toulouse, France, 31–37. ——. 2006. Computational linguistics. In EALL I, ed. Kees Versteegh et al., 511–518. ——. 2007. Featuring as a disambiguation tool in Arabic NLP, in Ditters and Motzki (eds.), 367–402. ——. 2011. A formal description of sentences in Modern Standard Arabic. In Studies in Semitic languages and linguistics, ed. T. Muraoka, A. Rubin, and C. Versteegh, 511–551. Leiden: Brill. Ditters, Everhard, and Cornelis H. A. Koster. 2004. Transducing Arabic phrases into head-modifier (HM) pairs for Arabic information retrieval. In Proceedings of the NEMLAR 2004 international conference on Arabic language resources and tools, September 22–23, Cairo, 148–154. Duchier, Denys and Ralph Debusmann. 2001. Topological dependency trees: A constraint-based account of linear precedence. In Proceedings of the Association for Computational Linguistics. Toulouse, France. 180–187. http://www.aclweb.org/anthology-new/P/P01/P01-1024.pdf Elkateb, Sabri, William Black, Horacio Rodríguez, Musa Alkhalifa, Piek Vossen, Adam Pease, and Christiane

Issues in Arabic Computational Linguistics Fellbaum. 2006. Building a WordNet for Arabic. In Proceedings of the 5th international conference on language resources and evaluation (LREC 2006), Genoa, Italy. Farghaly, Ali, and Karine Megerdoomian (eds.). 2004. Proceedings of the workshop on computational approaches to Arabic script-based languages, (COLING 2004), Geneva, Switzerland. Stroudsburg: Association for Computational Linguistics. —— (ed.). 2010. Arabic computational linguistics. Stanford: CSLI Publications. Fellbaum, Christiane (ed.). 1998. WordNet: An electronic lexical database. Cambridge, MA: MIT Press. Forsberg, Markus, and Aarne Ranta. 2004. Functional morphology. In Proceedings of the 9th ACM SIGPLAN international conference on functional programming, ICFP 2004, ACM Press, 213–223. Habash, Nizar. 2004. Large scale lexeme based Arabic morphological generation. In Proceedings of traitement automatique des Langues Naturelles (TALN-04), Fez, Morocco, 271–276. ——. 2010. Introduction to Arabic natural language processing. San Raphael, CA: Morgan & Claypool Publishers. Habash, Nizar, and Ryan Roth. 2009. CATIB: The Colombia Arabic treebank. In Proceedings of the ACL-IJCNLP 2009 conference Short Papers, Suntec, Singapore, 221–224. Habash, Nizar, Owen Rambow, and George Kiraz. 2005. Morphological analysis and generation for Arabic dialects. In Proceedings of the Workshop on computational approaches to Semitic languages at 43rd meeting of the Association for computational linguistics (ACL’05), Ann Arbor, MI, 17–24. Hafez, Ola M. 1991. Turn-taking in Egyptian Arabic: Spontaneous speech vs. drama dialogue. Journal of Pragmatics 15: 59–81. Hassanein, Ahmed T.. 2008. Lexicography: Monolingual dictionaries. In EALL III, ed. Kees Versteegh et al., 37–45. Hoogland, Jan. 2008. Lexicography: Bilingual dictionaries. In EALL III, ed. Kees Versteegh et al., 21–30. Jaszczolt, Katarzyna M. 2005. Default semantics: Foundations of a compositional theory of acts of communication. Oxford: Oxford University Press. Jurafsky, Daniel, and James H. Martin. 2009. Speech and language processing: An introduction to natural language processing, computational linguistics, and speech recognition. Upper Saddle River, NJ: Prentice Hall. Kaplan, Ronald, and Martin Kay. 1981. Phonological rules and finite-state transducers. In Linguistic Society of America Meeting Handbook, 56th Annual Meeting. New York. Kay, Martin. 1987. Noncatenative finite-state morphology. In Workshop on Arabic morphology. Stanford: Stanford University Press, 2–10. Khoja, Shereen, Roger Garside, and Gerry Knowles. 2001. A tagset for the morphosyntactic tagging of Arabic. In Proceedings of corpus linguistics 2001. Lancaster, UK, 341–353. Kiraz, George Anton. 1994. Multi-tape two-level morphology: A case study in Semitic non-linear morphology. In Proceedings of 15th international conference on computational linguistics (COLING-94), Kyoto, Japan, 180–186. ——. 1997. Compiling regular formalisms with rule features into finite-state automata. In ACL/EACL-97, Madrid, Spain, 329–336. Kiraz, George Anton., 2000. Multi-tiered nonlinear morphology using multi-tape finite automata: A case study on Syriac and Arabic. Computational Linguistics 26: 77–105. —— 2001. Computational nonlinear morphology with emphasis on Semitic languages. Studies in natural language processing. New York: CUP. Kiraz, George Anton, and Bernd Mö bius. 1998. Multilingual syllabification using weighted finite-state transducers. In

Issues in Arabic Computational Linguistics Proceedings of the 3rd ESCA workshop on speech synthesis, Jenolan Caves, Australia, 71–76. Köprü, Selçuk, and Jude Miller. 2009. A unification based approach to the morphological analysis and generation of Arabic. In 3rd Workshop on computational approaches to Arabic script-based languages, Computational Approaches to Arabic Script-based Languages (CAASL3). Available at http://mt-archive.info/MTS-2009Kopru.pdf. Koskenniemi, Kimmo. 1983. Two-level morphology: A general computational model for word-form recognition and production. Publication 11. Helsinki: Department of General Linguistics, University of Helsinki. Kulick, Seth, Ryan Gabbard, and Mitch Marcus. 2006. Parsing the Arabic treebank: Analysis and improvements. In Proceedings of the Treebanks and Linguistic Theories Conference. Prague, Czech Republic, 31–42. Levelt, Willem J.M. 1974. Formal grammars in linguistics and psycholinguistics: An introduction to the theory of formal languages. 3 vols. Den Haag: Mouton. Maamouri, Mohamed, Ann Bies, and Seth Kulick. 2009. Creating a methodology for large-scale correction of treebank annotation: The case of the Arabic treebank. In Proceedings of MEDAR international conference on Arabic language resources and tools. Cairo: Medlar. Maamouri, Mohamed, Ann Bies, Timothy Buckwalter, and Wigdan Mekki. 2004. The Penn Arabic Treebank: Building a large-scale annotated Arabic corpus. Paper presented at the NEMLAR International Conference on Arabic Language Resources and Tools, September 22–23, Cairo. Maamouri, Mohamed, and Ann Bies. 2010. The Penn Arabic treebank, in ed. Farghaly, 103–135. Manning, Christopher D., and Hinrich Schütze. 20036 [1999]. Foundations of statistical natural language processing. Cambridge, MA: MIT Press. Marton, Yuval, Nizar Habash, and Owen Rambow. 2010. Improving Arabic dependency parsing with lexical and inflectional morphological features. In Proceedings of the NAACL HLT 2010 1st workshop on statistical parsing of morphologically-rich languages. Los Angeles, CA, 13–21. McCarthy, John. 1981. A prosodic theory of non-concatenative morphology. Linguistic Inquiry 12: 373–418. Mesfar, Slim. 2010. Towards a cascade of morpho-syntactic tools for Arabic natural language processing. In Computational linguistics and intelligent text processing: Proceedings of the 11th international conference CICLing 2010, Iaşi, Romania, ed. Alexander Gelbukh. 150–162. Berlin: Springer. Mitchell, Terence F. 1985. Sociolinguistic and stylistic dimensions of the educated spoken Arabic of Egypt and the Levant. In Language standards and their codification: Process and application, ed. J. Douglas Woods, 42–57. Exeter: University of Exeter. Mohammad, Mahmud D. 1985. Stylistic rules in classical Arabic and the levels of grammar. Studies in African Linguistics 9: 228–232. Msellek, Abderrazaq. 2006. Kontrastive Fallstudie: Deutsch—Arabisch. HSK 25(2): 1287–1292. Müller, Karin. 2002. Probabilistic context-free grammars for phonology. In Proceedings of ACL SIGPHON, Association for Computational Linguistics, PA, 70–80. Odeh, Marwan. 2004. Topologische Dependenzgrammatik fürs Arabische. Abschlussbericht FR6.2-Informatik, Universität des Saarlandes. Owens, Jonathan. 1988. The foundations of grammar: An introduction to medieval Arabic grammatical theory. Amsterdam: John Benjamins. ——. 2003. Valency-like concepts in the Arabic grammatical tradition. HSK 25(1): 26–32. Palmer, Martha, Dan Gildea, and Paul Kingsbury. 2005. The proposition bank: An annotated corpus of semantic

Issues in Arabic Computational Linguistics roles. Computational Linguistics 31: 71–106. ——, Olga Babko-Malaya, Ann Bies, Mona Diab, Mohamed Maamouri, Aous Mansouri, and Wajdi Zaghouani. 2008. A pilot Arabic propbank. In Proceedings of the Sixth International Conference on Language Resources and Evaluation (LREC’08), ed. By Nicoletta Calzolari, 3467–3472. Marrakech, Morocco. Prince, Alan, and Smolensky, Paul. 2004. Optimality theory: Constraint interaction in generative grammar. Oxford: Blackwell. Ramsay, Alan, and Hanady Mansur. 2001. Arabic Morphology: A categorical approach. In EACL workshop proceedings on Arabic language processing: Status and prospects. Toulouse, France, 17–22. Ratcliffe, Robert R. 1998. The “broken” plural problem in Arabic and comparative Semitic: Allomorphy and analogy in non-concatenative morphology. Amsterdam: John Benjamins. Seidensticker, Tilman. 2008. Lexicography: Classical Arabic, In EALL III, ed. Kees Versteegh et al., 30–37. Sībawayhi (d. 798). Al-Kitāb. 2 vols. Būlāq 1316 A.H. Reprint Baghdad: al-Muṯannān.d. Smrž, Otakar. 2007. Functional Arabic morphology: Formal system and implementation. PhD diss., Charles University, Prague, Czech Republic. Smrž, Otakar, and Jan Hajič. 2010. The other Arabic treebank: Prague dependencies and functions. In ed. Farghaly, 137–168. Somekh, Sasson. 1979. The emergence of two sets of stylistic norms in the early literary translation into modern Arabic prose. Ha-Sifrut-Literature 8(28): 52–57. ——. 1981. The emergence of two sets of stylistic norms in the early literary translation into modern Arabic prose. Poetics Today 2: 193–200. Soudi, Abdelhadi, Antal van den Bosch, and Günter Neumann (eds.). 2007. Arabic computational morphology: Knowledge-based and empirical methods. Dortrecht: Springer. Vossen, Piek. 1998. EuroWordNet: A multilingual database with lexical semantic networks. Dordrecht: Kluwer. Winograd, Terry. 1983. Language as a cognitive process. Vol. 1: Syntax. Reading MA: Addison-Wesley. Włodarczak, Marcin. 2012. Ranked multidimensional dialogue act annotation. New Directions in Logic, Language and ComputationLecture Notes in Computer ScienceVolume: ed. By Daniel Lassiter and Marija Slavkovik, 67–77. Berlin: Springer.

Notes: (1) The Appendix at the end of the chapter lists some abbreviations and technical terms (frequently) used not only in this field in general but also in the current of this contribution, together with some paraphrasing of terms used in this contribution. (2 ) I would like to have had some more attention for equally socially relevant matters like pure linguistics. (3 ) Cf. Bouillon et al. (2007). (4 ) On the program of the annual ALS symposium on Arabic linguistics (2011), more than half of the presentations (17 of 31) dealt with Arabic colloquials, the diglossia situation, and the application of general linguistic theories for the description of Arabic colloquials. Beginning in 1990, this trend can be found in all the issues of Perspectives on Arabic Linguistics. For a while, the Moroccan Linguistic Society had a similar development. (5) If possible, together with a deaf window as well as a form of simultaneous (Arabic) Braille output. (6 ) See also Ali (2003). 7

Issues in Arabic Computational Linguistics (7 ) The end of this chapter offers suggestions for further reading. (8 ) For the main topics of this chapter, see Chenfour (2006) and Ditters (2006). Subtopics and references will be referred to in the body of the text. (9 ) The source here is Wikipedia; see also Ditters (2006). (10 ) For a paraphrase of descriptive terms, see Appendix 9.2. (11) There is a difference in interest between the computational subword (phonetics) and the computational word (phonology) level and beyond. This chapter is concerned with remedial and commercial applications. (12 ) See also Coleman and Pierrehumbert (1997) on stochastic phonological grammars and acceptability. (13 ) Cf. Farghali (2010: chapters 3 and 4) and Habash (2010: chapter 8 and Appendix). (14 ) Here in contrast with statically typed. Computer science presently has four main branches of programming languages: imperative oriented; functional oriented; logical oriented; and object oriented. For our purposes this information will be enough. (15) Wikipedia paraphrases treebank as a text corpus in which each sentence has been parsed, that is, annotated with syntactic structure, which is commonly represented as a tree. (16 ) I appreciate Darwish’s (2002) reference to Sībawayhi in his account of “a one-day construction of a shallow Arabic morphological analyzer.” (17 ) Wikipedia paraphrases: Haskell is a standardized, general-purpose purely functional programming language, with nonstrict semantics and strong static typing. (18 ) Quoting Smrž and Hajič (2010, 140): “these systems misinterpret some morphs for bearing a category, and underspecify lexical morphemes in general as to their intrinsic morphological functions.” I come back on this point while discussing the automated linguistic description of Arabic by means of programming languages or computational formalisms. (19 ) A programming language describes a dynamic and deterministic process. It is dynamic because there is a beginning and a series of steps to be taken leading inevitably to an end. It is deterministic because the computer is explicitly being told from the very beginning, how to start, where to find what it needs for the execution of the program, what to do with it, what the next step will be, and when its activity will come to an end. A formalism also is an artificial, formal language but is designed as a medium for the definition or the description of static structures. Such an approach is declarative because in the formal grammar only structures are defined and described. There is a beginning, a series of rules and an end, but there is no logical link between beginning and end. Not the computer but it is the machine-readable data that determine whether a match should occur or not; that is, the parser is dependent on the input string for deciding whether or not its structure can be recognized as defined or described in the formal grammar. (20 ) For Arabic, Sībawayhi (d. 798, kitāb) described nouns (N), verbs (V) and non-noun non-verb particles (-N-V) as the basic word categories. He also hinted at greater constituents with an element of one of those categories as head, but the labeling into NPs, VPs, and PaPs here is mine. (21) See, for example, also the objectives of The Arabic Language Computing Research Group (ALCRG), King Saud University (http://ccis.ksu.edu.sa/ar/en/cs/research/ALCRG). (22 ) Carter (2007: 27) discusses an earlier form of pragmatics in Larcher’s approach of ʾinšāʾ ( ibid., 28). See also Larcher (1990). (23 ) Cf. Bangalore et al. (2003), Bröker (2003), Fillmore (2003), Hajičovà and Sgall (2003), Hellwig (2003), Hudson (2003), Kahane (2003), Maxwell (2003), Mel’čuk (2003), Oliva (2003), Starosta (2003), Busse (2006), Hellwig (2006), Horacek (2006), and Schubert (2006). (24 ) See also Jurafsky and Martin (2009, section 3) and Eijck and Unger (2010). 25

Issues in Arabic Computational Linguistics (25) The references are slightly dated: Al-Najjar (1984), Bahloul (1994), Blohm (1989), DeMiller (1988), Eisele (1988), Gully (1992), Justice (1981, 1987), Mohammad (1983), Ojeda (1992), and Zabbal (2002). (26 ) Most references are a bit dated but concern colloquial varieties as well as Standard Arabic: Abu Ghazaleh (1983), Abu Libdeh (1991), Alfalahi (1981), Al-Jubouri (1984), Al-Shabab (1987), Al-Tarouti (1992), Bar-Lev (1986), Daoud (1991), Fakhri (1995, 1998, 2002), Fareh (1988), Ghobrial (1993), Hatim (1987, 1989), Johnstone (1990, 1991), Khalil (1985), Koch (1981), Mughazy (2003), Russell (1977), Ryding (1992), Salib (1979), and Sawaie (1980). (27 ) Cf. Dahl and Talmoudi (1979), Ghobrial (1993), Mahmoud (2008), Moutaouakil (1987, 1989), Mughazy (2008), and Suleiman (1989). (28 ) Accounting for all the aforementioned branches of syntax, including an opening to a semantic description of language properties. (29 ) A similar linguistic (and not heuristic thus pragmatic) description for generation purposes is not yet within reach. (30 ) The results of his formal system and the implementation of functional Arabic morphology (Smrž 2007b: 69) are presented in the form of unambiguous dependency trees. (31) Here is not the most appropriate place to initiate a discussion about the processing of a nondeterministic formal description of a natural language, for instance, in CFG terms, and the processing of a deterministic (each programming language) formal description of a natural language, whether or not the results are presented in IC, DG, HPSG, or any other form. (32 ) Duchier and Debusmann (2001) describe a new framework for dependency grammar with a modular decomposition of immediate dependency and linear precedence. Their approach distinguishes two orthogonal yet mutually constraining structures: a syntactic dependency tree; and a topological dependency tree. The former is nonprojective and even nonordered, while the latter is projective and partially ordered. (33 ) It is, unmistakingly, my fault not to have been clear enough to explain the basic principles of my approach to language description: the analysis of a linguistic unit in terms of alternating layers of functions and categories until final (lexical) entries have been reached. (34 ) For an example, see §2.3. (35) Maybe, for Arabic, we are not yet ready to think in terms of a linguistic and implementable paragraph, section, text, volume, and, generally applicable, syntax description of MSA or colloquial varieties of Arabic. (36 ) See for an overview of Arabic literature in general, among others, Sezgin (1967–2000), in particular volume 8 (lexicology) and 9 (syntax). (37 ) Cf. in this perspective also Baalbaki (1979). (38 ) See Appendix for a paraphrase of technical terms used. (39 ) See §2.3. (40 ) I am exclusively working in an analysis perspective of authentic MSA text data. (41) Superscripts to a nonterminal symbol, such as S, N, V, P, NP, VP, and PP point to categorial distinctions at the first level of description. Subscripts point to categorial, functional, or semantic characteristics at the second level of description. (42 ) See §2.3. (43 ) Cf. www.agfl.ru.cs.nl. (44 ) The main advantage of a formal description above a formalism, a computer or a programming language is that

Issues in Arabic Computational Linguistics a linguist is able, after a simple introduction, to read, understand, and comment on your description without first becoming a mathematical, logical, or computational linguist or scientist. (45) It is important to repeat that I exclusively use the formalism for analysis purposes. This means that the description may really be “liberal.” For synthesis objectives it is quite a different story. (46 ) 1 context-free grammar + 1 context-free grammar ≡ 2 context-free grammars ≠ 1 context-sensitive grammar. (47 ) Koster, 1971, 1991. (48 ) CDL refers to Compiler Description Language. CDL and C are, both, imperative programming languages, but in CDL3 the notational conventions are more suited for AGFL-formatted natural language descriptions to be tested. (49 ) For detailed information about probabilistic and frequency accounting properties of the AGFL formalism, I refer to the aforementioned AGFL site. (50 ) For more details see the aforementioned references. (51) A formal two-level context-free rewrite grammar means one describes one and only one nonterminal on the left-hand side and rewrites it into one or more nonterminal or terminal values from the lexicon on the right-hand side. However, this action will take place at the first and second level of description. The second level of description is included between parentheses (()), attached, if desirable this has been considered, to nonterminals of the first level. (52 ) AGFL stands for affix grammar over finite lattices. Cf. www.agfl.cs.ru.nl; Koster (1971, 1991). (53 ) Notational conventions: a hash (#) introduces a comment line; a double colon (::) rewrites the left-hand side of a nonterminal of the second level of description into one or more final values or another nonterminal of the second level of description; a vertical bar (|) separates alternatives on the right-hand side of the second level of description; a colon (:) rewrites the nonterminal at the left-hand side of the first level of description into one or more nonterminals or terminal values; a comma (,) separates successive elements on the right-hand side; a semicolon (;) separates alternatives on the right-hand side; an addition sign (+) tells the machine to ignore all spaces; a dot (.) ends each rule, with the exception of lexical rules. Nonterminals of the second level are written in uppercase. Terminal values of the second level are written in lower case. Terminal values of the first level are enclosed in double quotes (“”). (54 ) The listing of meta-, hyper-, predicate, or empty and lexical rules in reality is slightly longer. (55) In this sample grammar some alternatives of metarule rewritings of nonterminals at the second level of description occur only for elucidating purposes. (56 ) At the AGFL site mentioned as: AP4IR (Arabic [Dependency] Pairs for Information Retrieval). (57 ) Words from the open categories (nouns, verbs, and to a lesser extent adjectives and modifiers) carry the aboutness of a text, the others are in fact stop words. Similarly, only triples whose head or modifier are from an open category carry the aboutness of a text, and any other triples can be discarded as stopwords (Koster, 2011). (58 ) Only titles that appear in the main text (not the footnotes) are included in the bibliography. A complete bibliography can be found in the online chapter of the Handbook. Everhard Ditters Everhard Ditters, Nijmegen (Radboud) University

Sociolinguistics Enam Al-Wer The Oxford Handbook of Arabic Linguistics Edited by Jonathan Owens Print Publication Date: Sep 2013 Online Publication Date: Dec 2013

Subject: Linguistics, Sociolinguistics, Languages by Region DOI: 10.1093/oxfordhb/9780199764136.013.0010

Abstract and Keywords This article provides a critical overview of the application of sociolinguistic principles, methods, and analysis to Arabic data with reference to research conducted over the past three decades or so in various Arabic-speaking societies. It focuses on linguistic variation and change, the major concerns of (variationist) sociolinguistics. The article begins with an outline of the relationship between traditional dialectology and sociolinguistics, the ways dialectological data are incorporated into sociolinguistic analysis, and the benefits of maintaining the link between the two disciplines. Then an outline is presented of the basic principles of the variationist paradigm, which are intricately bound up with sociolinguistic methodology and theory; where relevant, research practices in studies on Arabic are cited. The article then critically reviews the “diglossia” model as an approach to analyzing variation in Arabic. Finally, an alternative and up-to-date model of analysis is given, with case studies from recent research used as illustration. Keywords: sociolinguistic principles, Arabic, linguistic variation, linguistic change, dialectology, diglossia model

10.1 Introduction The beginnings of sociolinguistics as an independent discipline within linguistics is marked by William Labov’s (1963) publication “The Social Motivation of a Sound Change,” in which he showed that the relationship between social and linguistic variables could be studied systematically (see also Fisher 1958). Reporting on his sociolinguistic research on Martha’s Vineyard, Massachusetts, Labov set forth methods of investigating language use in natural social settings within the speech community— hence the popular designation of sociolinguistics as the study of language in its social context. Most sociolinguists would consider the subject of their expertise to be simply a way of doing linguistics that shares with formal linguistics the objective of formulating a linguistic theory to account for the linguistic data. Where sociolinguistics diverges from formal linguistics is in its emphasis that language as a form of social behavior can be studied only in its natural setting, namely, society.1 Sociolinguistics is in various ways connected with most other subdisciplines in linguistics as well as with the other social sciences (especially at the level of interpretation). The most important challenge for sociolinguistics is the study of language change, the subject matter of historical (or diachronic) linguistics. However, whereas the analysis in historical linguistics begins with the outcome of change, sociolinguistic analysis deals with language change from inception through to propagation and completion. In this chapter, I provide a critical overview of the application of sociolinguistic principles, methods, and analysis to Arabic data with reference to the research conducted over the past three decades or so in various Arabicspeaking societies. The focus will be on linguistic variation and change, the major concerns of (variationist) sociolinguistics. I begin with an outline of the relationship between traditional dialectology and sociolinguistics, the ways dialectological data are incorporated into sociolinguistic analysis, and the benefits of maintaining the link

Sociolinguistics between the two disciplines. Then an outline is given of the basic principles of the variationist paradigm, which are intricately bound up with sociolinguistic methodology and theory; where relevant, research practices in studies on Arabic will be cited. The chapter then critically reviews the “diglossia” model as an approach to the analysis of variation in Arabic. In the final part, I set out an alternative and up-to-date model of analysis, with illustration of case studies from recent research.

10.2 Dialectology as a Resource for Sociolinguistics Both dialectology and sociolinguistics focus on linguistic variability in empirically collected data. But, whereas traditional dialectology is more interested in rural communities and, as far as possible, teases out variations in the speech sample by gathering data preferably from the oldest and least mobile members of the community, 2 sociolinguistics is interested in all types of speakers and all types of communities and has devised methods to analyze social as well as linguistic variables and to correlate the social and the linguistic variables in a systematic way. Notwithstanding differences in sampling and analytical techniques, sociolinguistics clearly has its roots in traditional dialectology, and where dialectological data are available sociolinguistic research anchors its findings to previous dialectological findings. To demonstrate the ways dialectological data have benefited sociolinguistic analysis, some examples are cited. In Labov’s very first sociolinguistic research on Martha’s Vineyard in 1962, he noted a particular way of pronouncing the diphthongs /ay/ (“price”) and /aw/ (“house”) with a centralized first element of the diphthongs. Beginning with the records of earlier dialectological studies, he traced the centralized variants from as early as the 16th and 17th centuries, right through to the first decades of the 20th century in various parts of North American English. Importantly for Labov’s Martha’s Vineyard study, The Linguistic Atlas of New England (Kurath et al. 1941) recorded an open, rather than a centralized, pronunciation of these diphthongs for Martha’s Vineyard. On the basis of this record and the earlier historical records, Labov was able to establish that the centralization of / ay/ and /aw/ in Martha’s Vineyard was an old pronunciation whose use decreased during the 1930s and in the 1960s was being revived by a particular group of speakers on the island. The reintroduction of the traditional centralized pronunciation coincided with an influx of wealthy summer visitors from the mainland who were buying up property on the island. Some Vinyarders showed resentment toward these outsiders. Labov showed that centralization of the diphthongs was strongest among the speakers who showed most resistance to outside pressure, thus demonstrating the effect of social factors on linguistic structure. The practice of drawing on earlier records where available has proven to be an invaluable resource, especially since most sociolinguistic studies are investigations in apparent time, where the symptoms of change are gleaned from the patterns of variation rather than from a comparison of the state of language at more than one point in real time. A case in point is the relatively recent discovery of recordings that were hidden in the archives of the New Zealand National Broadcasting Corporation of interviews conducted with the first generation of New Zealand—born Anglophones. Not only have these recordings provided empirical foundations for the study of the formation of New Zealand English (Gordon et al. 2006), but they have also modified previous assumptions regarding a number of innovations in British English, such as the approximate date of the loss of nonprevocalic /r/ in England (see Trudgill 2004). With the exception of a few shining examples, 3 early Arabic sociolinguistics and Arabic dialectology have proceeded almost totally independently of each other, although dialectological descriptions of various Arabic dialects have been available since the second half of the 19th century, and, until the advent of sociolinguistics, dialect descriptions and dialect geography were the only sources of data about modern Arabic dialects. 4 This situation has nonetheless improved considerably especially since the 1990s and in both directions. A particularly interesting and significant case of reanalysis of dialectological data is one recently undertaken by Manfred Woidich and Liesbeth Zack concerning the history of Jim in Egyptian Arabic in which the social history of Egypt as well as insights from sociolinguistic theory and modern dialect geography are incorporated into the analysis (Woidich and Zack 2009). The thesis they examined is whether [g], the Egyptian realization of Jim, is indeed an innovation, as has been suggested by Blanc (1969) and later supported by Hary (1996). They conclude that, contrary to the earlier analysis, Egyptian [g] may be a direct derivation from an older variety of Arabic that had preserved Semitic

Sociolinguistics /g/ (the antecedent to the Arabic Jim) (Woidich and Zack 2009: 56–57). This is a significant conclusion insofar as Egyptian Arabic is concerned and has an impact on the dialect geography of Arabic in general since it places Egyptian Arabic alongside dialects at the periphery of the Peninsula (South Yemen and Oman) in which this relic feature ([g]) continues to exist. 5

10.3 Basic Principles and Methodological Practices in Sociolinguistics Language variation and language change are the most important concerns of sociolinguistics. They represent two interrelated properties of language: the fact of variability and the fact of change. Change is always preceded by variation; it would be inconceivable for form A to change into form B without going through a stage where both forms, A and B, are used variably by members of the speech community. Therefore, understanding the intervening stages of the progression (variation) is a prerequisite to understanding the outcome of the process of change. Sociolinguistics has shown that linguistic variability is not random but is structured by linguistic and social constraints; hence, the first task of sociolinguistic research is to identify and isolate these constraints through analyzing representative empirical data [Holes, “Orality”]. 6 However, obtaining samples from the vernacular, or the way people normally speak when they are not being observed, is not an easy task, and sociolinguistic methodology has gone through a great deal of reformulation to improve this vital part of the procedure.7 In the next section, I consider some of the methodological issues that pertain to Arabic in particular.

10.3.1 Methodological Issues in Arabic Sociolinguistics Since the early days of research in Arabic sociolinguistics, considerable methodological strides have been made toward obtaining data that represent natural speech as far as possible, from recording speech in more than one context (e.g., Owens 1998) to enlisting in-group members of the community as field-workers (e.g., Miller 2005). However, what constitutes the vernacular and how techniques can be improved to elicit natural speech in the context of Arabic-speaking societies are questions that are yet to be addressed directly by sociolinguists working with Arabic data. By way of highlighting some of the important methodological issues in Arabic research, I shall discuss two of the independent variables: how they were isolated and how they were employed in research.

10.3.1.1 Style Generally speaking, stylistic variation is understood to involve variation in the speech of the individual speaker according to the task they are performing (speaking vs. reading), the context (e.g., interlocutors at home vs. work; interacting with family members, work colleagues, superiors), and topic of conversation. 8 In the early Labovian methodology, style was related to the use of more or less standardized speech and different styles that triggered different degrees of standardization were arranged along a scale that ranged from the casual to the most formal styles. Elicitation of the most formal styles has normally been achieved through reading tasks (passage and word lists), where the speakers are said to pay most attention to their speech and hence monitor it carefully. For languages such as English, where status and prestige are attached to linguistic features that are characteristic of the dialects spoken by the higher social classes and often coincide with the features found in the standard variety, the early Labovian methodology of quantifying style and correlating it with linguistic usage has yielded systematic results and interpretations. In this domain, it was found that the more attention speakers paid to their speech the more they approximated to standard or prestigious speech. 9 For a language like Arabic, where the speakers are confined by the normative and prescribed rules of Standard Arabic in reading tasks the Labovian methods of eliciting stylistic variation would not yield a comparable set of data to those obtained from research on English.10 By and large, stylistic variation in Arabic studies has been treated as constituting a binary choice between standard and colloquial features. A number of Arabic studies elicited this particular type of stylistic variation by manipulating the topic of discussion. 11 The underlying principle is as follows. Topics normally discussed in the public domain and in the media are deemed serious topics and thus trigger the use of standard Arabic linguistic features, whereas everyday ordinary topics trigger the use of colloquial features. For example, in Al-Khatib’s (1988) study in Irbid (Jordan), the serious topics included “role of women in society” and “educational policy”; the informants’ linguistic behavior in discussing these topics was considered to reflect a formal style. To elicit casual style, the speakers were encouraged to discuss “lighter topics” such as hobbies and recreational activities and to 12

Sociolinguistics tell jokes. 12 In Al-Khatib’s study, as in others’ that followed this method, the change in “style” thus measures only the use of the standard variant and neglects variation within the nonstandard variants (in variables with more than one colloquial variant). By way of illustration, in Al-Khatib’s (1988) study, the variable (Q) has three colloquial variants, [g], [k], [ʔ], and Standard [q]. Covariation with style is calculated for the Standard variant only: 62%, 100%, 100% in formal style (discussing serious topics), reading passage, and word lists (Al-Katib 1988: 264), respectively. Predictably, in the two reading tasks the speakers used [q] consistently, but in the discussion of serious topics they varied between the standard and nonstandard variants. No statistics are given for the colloquial variants separately, giving the impression that variation at this level is irrelevant or that variation in the use of the colloquial variants (the vernacular) is not constrained by social or stylistic factors. On the contrary, as shown in AlKhatib’s study (as well as in all other studies on Jordanian Arabic), the use of one or the other of the Qaf colloquial variants is highly constrained by the social and contextual (interlocutor) factors, while the use of standard [q] is predictable. Furthermore, the frequency of the occurrence of [q] does not affect the sociolinguistic correlations of the colloquial variants, nor does it affect the course of change in Qaf or the structure of variation in the vernacular. 13

Manipulating modes (speaking versus reading) and topics of discussion are clearly inadequate methods for style elicitation in Arabic vernaculars. 14 The significance of the independent variables, including style, as constraints on linguistic behavior can be interpreted only if they are analyzed in relation to the social context from which they derive; the same variables can mean different things in different societies. Clearly, there is much to be done in this area in sociolinguistic investigations of variation in Arabic. In the next section, I discuss “education” as a widely used but poorly understood social variable in Arabic studies.

10.3.1.2 Education as a Speaker Variable Education has been used very widely as a major sampling criterion in Arabic studies, almost analogously to the way socioeconomic class has been used in studies in North America and the United Kingdom. 15 In most studies, speakers were classified according to “level of education.” The results appeared to show regular and strong correlations between level of education and linguistic behavior. 16 However, these same results are paradoxical, or so they appeared, in that while the group of speakers with the highest educational level showed the highest rate of usage of standard Arabic features for some variables they were in the lead of innovations away from standard features for other variables. For example, in Abdel-Jawad (1981), the highest educational group used standard [q] most. For the second variable, (k), Abdel-Jawad also concluded that the same group standardized their speech most by using [k] rather than the affricated variant [t∫]. In Al-Khatib (1988), the highest educational group used [q] most often, but for the variable Jim they used fricative [Ʒ] rather than the standard affricate variant [dƷ] more than any other group.17 In Al-Wer (1991) the highly educated group used [q] more frequently than the lower educational groups, but they also led all other groups in using three nonstandard innovative features, (θ): [t], (ð): [d] and (dƷ): [Ʒ]. To resolve this prima facie paradox, two issues will be addressed. First, we need to analyze the significance of “education” as a speaker variable in Arabic-speaking societies. Second, the role of the standard variety in variation and change in spoken Arabic needs to be discussed in the light of the empirical data available from variationist research. In the remainder of this section I will discuss the first point, while the second issue forms part of the discussion in Section 10.4. In Al-Wer (2002), I argue that education may be a proxy variable acting on behalf of other less conspicuous variables such as the speakers’ relationships with their local community and the sums and types of relationships they have maintained with the local and outside communities (i.e., their social networks). A further possibility is that the underlying factors are related to their personal ambitions, orientations, and attitudes toward the local community as reflected in their daily activities, future plans for themselves and their children, place of residence in the town (inner-city vs. suburban residence), and participation in local politics (e.g., municipality elections, voluntary social work). The hypothesis that it is not education per se that correlated with linguistic behavior arose as a result of the overall conclusions about the trajectory of linguistic change in three Jordanian cities (Al-Wer 1991). The results showed that change may be in progress with respect to the following features: [θ] 〉 [t]; [ð] 〉 [d];

[ d ] 〉 [ḍ]; [dƷ] 〉 [Ʒ]. The changes altogether represent divergence from the traditional local dialects toward a ˙ 18 nonlocal and at the same time a supralocal norm at the level of the Levant region. In the case of the plain

Sociolinguistics interdental sounds, the changes involve mergers spreading at the expense of differentiated sounds. These two linguistic developments—preference for supralocal variants and spread of mergers—are in keeping with universal tendencies of the direction of change as demonstrated in sociolinguistic research. The local traditional Jordanian variants [θ ð dƷ] are at the same time characteristic features of standard Arabic. The emphatic interdental sound has two counterparts in the standard variety: an interdental sound, as in the local dialects; and an emphatic stop sound.19 Therefore, another way we can summarize the changes in this case is that the educated speakers lead change away from the standard variety or that there is an inverse correlation between level of education and the use of standard features. At the same time, the highest educated group used the glottal stop variant [ʔ] variably with the local variant [g]. Very little variation was found between these sounds, but all of the speakers who used it consistently (5 speakers of 116) or variably (4 of 116) belonged to the highest educational group. At the time of research (1987) there was definitely no change in progress from [g] to [ʔ]; rather, the little variation attested was interpreted as an innovative linguistic behavior that was confined to a few speakers and involved the use of a supralocal feature at the expense of the localized feature [g], since the glottal stop realization is also characteristic of the major city dialects in the region. This then substantiates the findings about the direction of change being toward supralocal features. Predictably, for Qaf, the highest educational group used the highest proportion of lexical items that contained the standard variant [q], which may be explained with reference to their level of education. That is, their functional knowledge of the standard variety is higher than the knowledge of the less educated, and they are therefore able to integrate learned lexical items into their speech; however, obviously no sound change toward [q] (away from [g]) is involved. 20 In sum, the empirical results show that longer exposure to and better knowledge of the standard do not result in convergence toward standard features. On the contrary, where the localized features are identical to the standard features and change is in progress toward supralocal features, the educated speakers diverge most frequently from the standard and localized features and lead the change in the direction of supralocal features. What determines the course (and success) of a language change is not how similar the feature is to a standard form. It therefore follows that proficiency in the standard variety and its determinants, such as formal education, is not the best parameter of speaker categorization. At the same time, since education has been shown to correlate with linguistic usage it must be the case that it denotes something of relevance to linguistic usage. Table 10.1 [θ] variation in Amman [t]

High

59%

284

Middle

34%

164

Low

I have suggested that education may be a proxy variable that acts on behalf of the degree of closeness or openness of individuals’ social networks and the range of contacts they have contracted within and outside the community. The reason that sampling the speakers according to level of education has provided some neat correlations in the past is because it so happened that, in these Arabic-speaking communities that were investigated, the highly educated individuals and the most mobile individuals are the same people. In my research in Jordan, for instance, all of the highly educated individuals in the older age groups (above 50 years) were educated abroad since no institutions of higher education were available in the country; among the younger generations, all of the highly educated (university level) had to leave their hometowns to go to university. To test this hypothesis, I reanalyzed parts of the data collected in 1987 from the city of Sult (Jordan) (Al-Wer 1991) and examined 42 speakers individually with respect to their use of the variable (θ). The linguistic data were correlated with the speakers’ network of contacts inside the town, contact with the community of the nearby capital city of Amman, their backgrounds, activities, and socialization habits. 21 The results of this reanalysis (Al-Wer 2002) provide convincing evidence that amount of contact with the community of Amman (where the use of the innovative feature [t] is at an advanced stage), integration in the local community, and engagement in local issues and future plans are important factors that correlated with linguistic usage. Following are some details about these 22

Sociolinguistics results. 22 The overall statistics for the three educational groups—high (university), middle (high school), and low (elementary)—with respect to their use of the innovative variant [t] in words with etymological [θ] (e.g., θgiil “heavy”) are presented in Table 10.1 (100% = [t], 0% = [θ]). In the first group, the highly educated, the individual scores of the 18 speakers in this group vary considerably. At the polar ends, there was one speaker who used [t] consistently and two speakers who did not use it at all. Most speakers were clustered between 54% and 81% use of [t]. Toward the top end, in addition to the speaker who used [t] consistently, there are two further speakers whose scores are 95% and 81%. These three speakers are most responsible for skewing the group’s score upward. All three of them were educated at Arab universities abroad, two of them lived in the suburbs, one of them was a daily commuter to Amman, and her children also went to schools in Amman. Two of them were close friends and mostly socialized together. All three were very critical of the local community; they neither enjoyed living there nor appreciated the benefits of living in a close-knit community. None of them participated in voluntary work or voted in the local elections. In the middle educational group (14 speakers), 4 young speakers did not use [t] at all, thus maintaining a consistent usage of the local variant [θ]. This subgroup of speakers had known one another since childhood and had been working together in the local post office. Outside work, they had a regular weekly get-together. Their network was based on shared interests, beliefs, outlook, and lifestyle as well as similar social backgrounds. In L. Milroy’s terms, this group represents a (locally oriented) tight-knit social network, which protects and supports its members and acts as a mechanism that enforces the local norms of social behavior, including the local linguistic norm. A social network analysis of these data thus provides important information not only about the use of innovative features but also about the maintenance of local features. 23 In Arab societies in a transitional phase to a modern type of society in economy and lifestyle, education tends to be the main channel through which upward social mobility can be achieved for the majority of the population. Individuals striving to attain higher education tend to expand their outside contacts. Thus, we often find that the speakers in the lowest educational group generally have limited outside contact. In the Sult sample, for example, the speakers in the lowest educational group (10 speakers) were the most conservative, in the sense that they used the local features most often (of all variables not just (θ)). Most of these speakers were also unemployed and rarely traveled outside the town. However, one speaker in this group stands out clearly as an innovator: she scored 72% of [t], 70% of [Ʒ] (for local [dƷ]), and 45% of [d] (for local [ð]). At the time of the research, she was 21 years old and was employed as a caretaker at a college on the outskirts of Sult. Compared with the other speakers in this group, she had more opportunity for access to (trendy) linguistic innovations through daily interactions with the young college students. In a model based on access to the target features and contact, rather than educational level, this speaker would not be exceptional; there is a meaningful correlation between amount of variation in her speech and her social characteristics. Education is clearly a simple, almost ready-made model of speaker classification. However, although it appears to correlate with linguistic usage, its usefulness as a speaker sampling tool in Arabic research may be only temporary as a result of ongoing social change. We might expect that once the developmental phase in these societies is completed and education is not the only channel via which contact and upward social mobility can be achieved, as a speaker variable “education” will run its course of utility.24 The methodological and analytical problems outlined in this and the previous sections are concomitants of an approach that is premised upon “diglossia” and proceeds in a top-down fashion toward analyzing variation in Arabic as an opposition between standard and nonstandard features, thus relegating sociolinguistic research to a confirmatory rather than an exploratory role. The following section discusses the validity and repercussions of this approach.

10.4 The “Diglossia” Model and Its Implications for the Study of Variation in Arabic Almost four decades have passed since the first appearance of studies on variation in Arabic, yet we find that Arabic data play no role in the formulation of generalizations or theoretical frameworks in sociolinguistics. If Arabic data are mentioned in texts, they are cited as rather irksome anomalies. The failure to make an impact is a

Sociolinguistics symptom that these data are gathered, analyzed, and presented in a fashion that is unsuited to the objective scientific study of language. In my view the fundamental problem lies in one of the most pervasive traditions in Arabic sociolinguistic research, which starts with diglossia as a framework for the analysis of variation in Arabic. For one thing, this approach has severed the natural connection between Arabic dialectology and Arabic sociolinguistics and has presented variation in Arabic as a case of opposition between standard and nonstandard features. The problem pointed out in this context is not concerned with investigations whose aim is to look at the ways the diglossic situation is functionalized as a stylistic resource by certain groups of speakers for specific purposes, but with the assumption that this circumstance is the key to understanding variation in Arabic or that this subset of data envelops the structure of variation in Arabic. The most serious misconceptions that emanate from this approach are explained as follows.

10.4.1 Dimensions of Variation Are Uniform Since diglossia is prevalent in all Arabic-speaking societies, taking the diglossic situation as the starting point of sociolinguistic analyses is tantamount to assuming that the structure of linguistic variation in different Arabicspeaking societies is invariant. Thus, there is no need to supply the social context, and the independent variables can be treated as abstract notions that do not relate to the life of the community. One aspect of this practice is the tendency to generalize patterns that were found in a particular community to Arabic in general without reason to believe that this one case is representative. On the contrary, the data show that, in Arabic as in all other languages, linguistic variation and change are structured by an interaction between linguistic and social variables, and both types of constraints are peculiar to each dialect and each community. Importantly, the linguistic constraints on variation are dictated by the respective native Arabic dialects of each community, not by structures or features found in the standard variety. The failure to factor these basic facts into the analysis leads to all sorts of misleading conclusions. For instance, in studies that used samples of speakers from different dialectal background, the linguistic characteristics of the native dialects of the speakers are not factored into the analysis, and the speakers of different Arabic dialects are treated as standing at equal distance from standard Arabic features regardless of their native grammars (the linguistic characteristics of their native dialects). It is thereby tacitly assumed that all native speakers of Arabic have an invariant grammar, that of the standard variety, from which they derive whatever dialectal forms they happen to use in daily interactions. For instance, Abdel-Jawad and Awwad’s (1989) investigation of the use of the interdental sounds, the sample consists of speakers of different dialects from Jordan, Syria, and Egypt. The statistical analysis begins by calculating the average usage of the interdental realizations, which are referred to as standard variants even though all of the local Jordanian dialects have the same sounds. No information is supplied about the dialectal backgrounds of the Jordanian speakers; they are labeled as speakers of urban Jordanian dialects. The use of interdental versus stop sounds in Jordan’s urban centers is variable, but it is constrained by the speakers’ dialectal background: those whose native dialects are indigenous Jordanian dialects or rural Palestinian dialects have interdental sounds; and those whose native dialects are urban Palestinian or urban Syrian do not have the interdental sounds. These important constraints on variation are not factored into the analysis (see Abdel-Jawad and Awwad 1989: 263—265). The same method of analysis is used to make conclusions about the distribution of the use of the variants by sex. It is concluded that “men use the standard variants [i.e., the interdental sounds] more often than women do, whereas women use the socially prestigious forms … more often than men do” (ibid., 264). Here, the problem emanating from the failure to take account of the dialectal backgrounds of the speakers is compounded by using what was in the first place a misleading description (labeling the interdental sounds as the standard variants without accounting for the fact that these sounds are at the same time characteristic of the local dialects) as a basis for the description of a different level of variation: sex differentiation. Naturally, these speakers (men and women) could have simply been maintaining their normal realization of these sounds rather than approximating to or diverging from the standard. More problems arise when interdialectal comparisons are made (Jordanian—Syrian—Egyptian). In these results (ibid., 267, Table 10) the Jordanians were found to use the interdental sounds of all variables more frequently than using the stop or sibilant sounds and more than the Syrian and Egyptian speakers. This is obviously a highly predictable result given that the Jordanians have a native facility toward the use of these sounds, which is unavailable to the speakers of Damascus and Cairo Arabic. Similarly, misleading conclusions can arise by failing to take account of the fact that the social constraints on variation in Arab communities vary quite considerably, even in geographically adjacent and culturally very similar societies.

Sociolinguistics 10.4.2 Standard Features Have the Highest Rate of Approval and Enjoy Unquestionable Prestige This is an assumption without empirical evidence. In fact, the linguistic evidence that is available can be read as an antithesis to this assumption since the empirical data show that many dialectal features are abandoned even though they are identical to standard features. For example, at the level of segmental phonology, in Levantine dialects the standard features [q θ ð dƷ] are abandoned in favor of the nonstandard counterparts [ʔ t d Ʒ]. The standard diphthongal pronunciations of /aw/ and /ay/ are abandoned for the nonstandard (and in some cases supralocal) monophthongal pronunciations /o:/ and /e:/ in Tunisia (Jabeur 1987). At the morphosyntactic level, speakers of dialects that, similarly to the standard, maintain gender distinction in the second- and third-person pronouns, pronoun suffixes, pronominal suffixes, verbal inflections, and nominal endings abandon this system in favor of a system that neutralizes gender in these forms and paradigms. 25 Standard Arabic clearly has a function in Arabic-speaking societies, as the norm used in formal written and spoken domains; it undoubtedly also has a psychological claim on native speakers of Arabic. But it does not play a role, nor does it have a normative effect on the structure of variation in the core domains of phonology, morphology, and syntax, as explained in the next point.

10.4.3 The Trajectory of Linguistic Change is in the Direction of the Standard Features A corollary to approaching variation in Arabic as an opposition between standard and nonstandard features is that in cases of variation where one of the variants happens to be identical to a standard feature and linguistic change is in progress, the trajectory of the change is anticipated to be in the direction of the standard feature. Numerous examples from empirical research can be cited to refute this prediction. For example, many of the traditional dialects in the Gulf region have a glide [y] as a realization of Arabic Jim. The most recent investigation of this feature in Bahrain is by Muna Al-Qouz (2009). This is a particularly good example because in this case we have data in real time from various sources, most importantly from the research by Clive Holes in the 1970s (Holes 1987). Al-Qouz’s investigation thus presents data from Manama 30 years later. In Bahrain, linguistic variation and change is in the first place constrained by the speakers’ sectarian affiliation. There are two major dialects: the more influential Arab dialect, which is numerically in the minority and which has the glide variant [y]; and the less influential Baharna dialect, which has the affricate palatal [dƷ] variant as in Standard Arabic. Holes reports that the Arab group largely does not show variation in the use of Jim (especially in the core dialectal lexical items), but the Baharna speakers replace their traditional (and standard) palatal variant by the glide variant. And 30 years later, one no longer finds segregated towns or neighborhoods; rather, the two groups can be found to reside in all parts of the city, and their children go to the same schools. Importantly, too, in the same period, illiteracy has decreased sharply, and the overall level of education has increased. In other words, there are no social barriers to prevent the two groups from interacting freely, and exposure to the standard variety has extended to larger sectors of the populations of both communities. Al-Qouz finds that the use of the glide variant has become the norm in the speech of the younger Baharna group and that the palatal variant occurs only occasionally, most often fossilized in a few lexical items. The development in Bahrain with respect to this feature is not isolated but seems to be a Gulf-wide development (Holes 1990). In Kuwait, Taqi (2010) shows an identical development among the young speakers from the Ajam group who were found to use the glide variant in place of their heritage feature [dƷ]. The determining factors of the geographical and social diffusion of the glide variant have to do with the social power of the Arab group in the case of Bahrain, and in the case of Kuwait, by the size of demographic representation as well as the social dominance of the dialects which have the glide The similarity or distance from the Standard feature is totally irrelevant in this domain. Many more such examples can be cited from various modern Arabic dialects.

10.4.4 Data from Arabic Contradict Generalizations Insofar as sociolinguistics in general is concerned, the most serious outcome of analyzing variation in Arabic in a top-down fashion is that it has led sociolinguists up the garden path in searching for interpretations of genderdifferentiated patterns. For instance, on the basis of the Arabic data presented in research from the 1970s and 1980s, which claimed that Arab women contradict the pattern found elsewhere by not using standard prestigious features as frequently as Arab men, Labov (2001: 270) writes: There is therefore a widespread reversal of the positions of men and women predicted by Principle 226 in

Sociolinguistics two Muslim-dominated societies [Iran and the Arab World]. Labov also notes and seems to accept the reinterpretation offered by Haeri (1987) that the Arabic data do not in fact contradict the pattern. But he concludes: Principle 2 is a strong and broad generalization that must be explained by women’s social role in the speech community. The cases that depart from the general pattern are from well-defined cultural areas: the Far East … and the Near East. (P. 271) The argument that is commonly advanced to interpret Arab women’s presumed divergent linguistic behavior rests on the presumption that Arab women play a less prominent role in public life, without specifying which women of which Arab society. The important point to stress here is that Arabic-speaking societies are wrongly presented as one culturally and socially invariant unit, where the roles of men and women are polarized and rigidly defined: men are the sole operators in public life and women are the homemakers. Not only does inadequate methodology yield inaccurate data (as demonstrated already), but also such data end up forming the basis of sweeping generalizations that become entrenched and difficult to correct. 27 The interpretation of sex-differentiated patterns of variation in Arabic was revised by Mohammad H. Ibrahim (1986). Additionally, Ibrahim points out a number of problems in the interpretation of Arabic data, which stem from what he identifies as confusing the notions of “standard variety” with “prestigious variety,” and suggests making a distinction between the status of the standard variety as a transnational norm and the spoken dialects that have their own hierarchy of prestige at the local level. Following Ibrahim’s line of argumentation, Niloofar Haeri (1987) offers a more realistic explanation for male— female linguistic differences in Cairo Arabic, which refutes the claim made earlier that sex-differentiated patterns in spoken Arabic contradict the general pattern. I argue (Al-Wer 1997) that the involvement of standard Arabic in sociolinguistic interpretation seems to be based on ideological considerations rather than on empirical data. I suggest there that, by setting the standard variety aside, paradoxical patterns disappear and findings from Arabic regarding gender differentiation fall in line with generalizations in sociolinguistics that have been made on the basis of data from a wide range of languages and societies. The potential for fallacy that lies behind the top-down model of analysis of variation in Arabic is demonstrated by the fact that such a model does not hold up to the empirical data, for as soon as one tries to build in the data the model simply sinks and collapses. A sound analytical model, on the other hand, is one that begins by grounding the linguistic data in the linguistic system from which they are derived (the local dialect) and the social context with which they interact (the local community), as illustrated in the next section.

10.5 The Alternative Model: a Locally Based Analysis and Interpretation A framework of analysis that begins with the locality and the local dialect, sampling the speakers according to independent variables that relate to the community’s history and evolution, identifying linguistic variables that are meaningful and relevant, and deriving the interpretations from the local community’s structure, characteristics, and the daily pursuits of its members is a model that is capable not only of accounting for language use but also of being flexible enough to accommodate, and perhaps uncover, new dimensions of variation. These are the underlying principles of sociolinguistic research in general, which have been recently incorporated in a number of insightful research studies in various localities around the Arab World. The developments seen in handling Arabic data and Arab speech communities also reflect advancements in sociolinguistic research in general.28 In the remainder of this chapter, I will cite four case studies that exemplify some of the new dimensions of variation in modern Arabic societies and identify new linguistic variables.

10.5.1 A New Linguistic Variable in an Old City: (r) in Damascus The first case is an investigation of variation and change in Damascus, completed in 2008 by Hanadi Ismail. 29 Ismail’s analysis of the community’s history in relation to the socioeconomic changes in Damascus shows that, as

Sociolinguistics a result of the physical expansion of the city that began in the early 1970s, new dimensions of linguistic variation emerged that related to two aspects of life in the city: residence in a traditional inner-city district (Shaghoor); and residence in a new satellite suburb (Dummar). These in turn broadly correlate with two different “life modes,” selfemployed and professional, respectively. Table 10.2 Innovative (= non-trill) “r” variants in two Damascus neighborhoods (all correlations statistically significant) Male

Female

Shaghoor Y

29% (406)

23% (407)

4%(294)

5%(449)

1%(380)

3%(393)

19%(422)

27%(499)

21%(414)

10%(393)

4%(379)

20%(427)

Dummar

Note: Y = young; M = middle age; O = old age. Source: Based on Ismail 2007: 207, fig. 9.8. Ismael uses the construct of life mode (Højrup 2003) to explain observed linguistic differences. One of the new variables identified in Damascus is /r/, which is undergoing a process of lenition. Four types of variants were identified. Besides the inherited trill, there are three innovative values: retroflex approximate [ɻ], fricative [ɣ], and aveolar approximant [ɹ]. In all, 47,603 tokens were coded for five phonetic environments. The results of the GoldVarb analysis, adapted for presentation here, are given in Table 10.2. Table 10.2 shows correlations of the use of the most advanced /r/ variant with age, gender, and neighborhood. Statistically, all correlations were found to be significant and the variable of age was found to have the most significant effect, thus indicating a change in progress toward an approximant type of /r/. The table also shows that the locus of the change is the suburb, where the new pronunciation appears in significant proportions in the speech of the middle- and old-age groups as well. The gender effect is particularly interesting: here we notice that, while in the suburb the female speakers have a clear lead, in the inner-city locality of Shaghoor it is the young male speakers that lead the change over all other groups. This particular finding is explained by Ismail in relation to the employment situation; at the time of research, all except one of the women were unemployed, and all of the men were employed in the retail business, which brought them in direct and regular contact with customers from all walks of life and all parts of the city. In addition to identifying a new linguistic variable, the example from Damascus shows how research that is designed to incorporate the social and economic dynamics of the local community is capable of correlating the variables in a meaningful way. These data also provide a well-analyzed example of a male-led linguistic change. 30

10.5.2 New Dimensions of Variation: Style and Dialect Leveling in Casablanca The second example comes from Atiqa Hachimi’s research in Casablanca (2005). 31 In this case, Hachimi shows

Sociolinguistics how growth and increasing heterogeneity in the city’s population have disrupted the old rural–urban dichotomy, the once dominant characteristic of Moroccan city dialects and identities.32 In Casablanca, mass migration has created a new linguistic urban model, represented by the newly formed koines: Fessi (from Fez) and Casablancan. The research investigated the linguistic and social outcomes of the Fessi–Casablancan contact in Casablanca. Hachimi’s research employs the “communities of practice” method for the first time in Arabic research.33 Three linguistic variables were investigated: two phonological, (r) and (q); and one morphosyntactic 2FSG (-i) (which concerns gender neutralization versus the maintenance of gender distinction in the second-person feminine subject clitic). These variables distinguish between the Casablancan and Fessi (Fez) dialects. The results overall show that the speakers use the linguistic forms as a resource in the construction of identities and different styles that are directly related to their lives and daily pursuits in the city. In addition to the old identities, such as “pure Fessi” and their linguistic correlates (maintenance of the heritage variants), the speakers construct hybrid identities that correlate with hybrid linguistic behavior. For instance, for a Fessi to become Casablancan, the heritage variants [ʔ] and [q] are selectively filtered in favor of [g] in [ga:l], and Casablancan trilled /r/ is adopted instead of the Fessi approximant [ɻ]. The overall conclusions show very clearly that in a complex social structure the linguistic variables can be interpreted as stylistic variation reflecting different identities created through combining resources from different linguistic stock. The stylistic shifts in the informant’s speech are dictated not by the diglossic situation or by a shift of task or topic but by a dynamic that is directly related to the everyday realities that inhabitants of the heterogeneous city have to negotiate on daily basis in an attempt to maximize gain and minimize loss in social interactions.

10.5.3 Migrant Groups, New Social Variable, and Linguistic Constraints: (k) in Jeddah The phenomenon of migrant groups in large cities is commonly found all over the Arab World, which provides an exciting area of research on variation, dialect acquisition, and contact-induced linguistic change.34 The research reviewed in this section is by Aziza Al-Essa (2009), which investigated the linguistic developments in the speech of Najdi migrants in the Hijazi city of Jeddah. 35 The investigation included nine phonological and morphophonemic variables. I will focus on the variable (k), which concerns the affrication of [k] in the environment of front vowels and close back vowels (Cantineau 1960, Johnston 1963). Al-Essa’s analysis of this feature in the context of a dialect contact situation shows that the variation that results from the contact situation in Jeddah is quite complex. First, her analysis reveals that the process of affrication of [k] to [ts] differs depending on whether the consonant occurs in the stem, as in mika:n 〉 mitsa:n “a place,” or in the suffix as in ʔumm-ik 〉 ʔumm-its “your (f) mother.” The second variable concerns [k] in the second-person feminine suffix; therefore, it was analyzed as a morphophonemic feature, while affrication of [k] in the stem was analyzed as a phonological feature. Furthermore, while affrication occurs in the phonological variable only in the environment of front vowels, variation in the morphophonemic variable is also present in the environment of close back vowels, such as ʔubu:-ts “your (f) father.” With respect to the phonological variable, that is, [k] in the stem, the results show that affrication is disappearing very fast, with an overall usage of only 6% of a total of 668 tokens. On the other hand, in the case of [-k] in the suffix, the data show much more complex patterns. First, the affricated (heritage) variant shows a considerably higher rate of maintenance, 22%, or nearly four times the frequency of occurrence of the phonological variable. Second, interdialectal forms are used. Al-Essa explains the results for the morphophonemic variable as relating to (a) the complexity of the distribution of the feminine suffix in the target dialect (Jeddah) vis-à-vis the Najdi dialect, and (b) the fact that [ts] versus [k] carries gender information; that is, it has a morphosyntactic function in the Najdi dialect. The complexity of the suffix in the target dialect can be seen in the following schema, which outlines the forms found in both the Najdi and Hijazi dialects:

Najdi

Hijazi

Masc.

-ik

-ak

Fem

-its

ik\C-,

ki\V-

Sociolinguistics Notice that the Hijazi feminine form –ik is identical to the Najdi masculine form; thus, ʔummik is “your (m) mother” in Najdi but is “your (f) mother” in Hijazi. The Najdi dialect relies on the change in consonant to signal gender distinction while the vowels in both forms are identical (phonetically [ǝ] or thereabouts), whereas the Hijazi dialect relies on vowel change to signal gender, and in the case of the feminine ending the form varies morphophonologically according to the preceding sound, -ik before a consonant and -ki before a vowel, thus ʔumm-ik, but ʔabu:-ki. Thus, for the Najdi speakers to accommodate to the Hijazi forms correctly, they would have to (a) deaffricate / ts/, (b) acquire the vowel distinction, and (c) observe the phonological conditioning of the choice between –ik and -ki. These linguistic constraints seem to explain the finding that deaffrication in the suffix lags behind deaffrication in the stem by a wide margin. Additionally, Al-Essa finds instances of both hypercorrection in the target data, for instance -ki used by the Najdi speakers not only postvocalically (as in the target dialect) but also after a consonant, such as maʕki “with you (f),” and neutralization of contrasts, such as ʔubu:k “your (f) father” (Al-Essa 2009: 155–173). With respect to contact, the results showed positive and systematic correlation between amount of contact (quantified according to the index) and use of the target variants. Al-Essa’s research thus provides invaluable innovative methodological and analytical insights for future research.

10.5.4 Koineisation and the Formation of New Dialects: The Phonology of the Feminine Ending in Amman In some cases, the linguistic developments represent not continuation of change in a dialect but the formation of a new dialect from scratch. A case in point is the capital city of Jordan, Amman. As an Arabic-speaking community, its history started only in 1921 when it was declared the capital city of Jordan. From then on, the city’s population grew steadily and at points abruptly, especially between 1950 and 1990. Currently it is home to 1.6 million. The two major groups that constitute the city’s population are the Jordanians and the Palestinians. With such a recent history and no truly native population, Amman naturally does not have a traditional dialect, but a new dialect is in the process of forming. The ingredients that went into the formation of the new dialect are a range of Jordanian and Palestinian dialects.36 In 1998, I launched a project to investigate the process of the formation of the new dialect. So far, 13 linguistic features have been analyzed and shown to be at an advanced stage of focusing. This section cites the data concerning the feminine ending /ah/. This feature concerns the raising of the feminine ending /a/ to /e/. Urban Palestinian dialects and Jordanian dialects differ in the phonology and phonetics of the suffix. Jordanian dialects (central and northern dialects) use /a/ in all environments except after coronal sounds, in which case the ending is raised to a half-open vowel. [Ɛ] (i.e., /a/ is the default variant) /ħilwa/ “pretty,” / ʃara:ka/ “partnership,” / biʃʕa/ “ugly” vs. / madanijjɛ/ “modern or urban,” / fattɛ/ “Fatte, a traditional dish,” / mdʒawwazɛ/ “married,” / sanɛ / “a year,” On the other hand, in urban Palestinian the feminine ending is realized as /e/ except after velarized, emphatic, and pharyngeal sounds where /a/ is used (i.e., /e/ is the default variant). The raised variant is [e] (cardinal vowel 2 or closer). /ħilwe/ “pretty,” / saʕbe / “difficult,” /mixtilfe / “different” vs. / manṭiqa / “area/region,/ruxṣa/ “a licence,” / sulṭa/ “authority,” / ʒa:mʔa/ “university,” / mari:ḍa / “ill.” In Amman, the contact between speakers of the two types of dialects has resulted in the emergence of a fudged form, which combines Palestinian phonology and Jordanian phonetics, that is, /a/ is raised everywhere except after velarized, pharyngeal, and emphatic sounds, and the raised variant is phonetically half-open [ɛ] (cardinal vowel 3). Regional koineization in this case also exerts an influence on the course of developments in Amman, seeing as the urban Palestinian system is identical to the system found in all of the major city dialects in the Levant (identical in phonology but not in phonetics). The previously cited examples represent the diversity of situations that influence linguistic variation and change in modern Arabic. In none of these situations do we see the standard variety play any role in the structure of variation

Sociolinguistics or in the dynamics of the change.

References Abdel-Jawad, Hassan. 1981. Lexical and phonological variation in spoken Arabic of Amman. PhD diss., University of Pennsylvania. Abdel-Jawad, Hassan and Mohammad Awwad. 1989. Reflexes of Classical Arabic interdentals: A study in historical Sociolinguistics. Linguistische Berichte 122: 259–282. Abdel-Jawad, Hassan and Saleh Suleiman. 1990. Lexical conditioning of phonological variation. Language Sciences 12: 291–330. Al-Amadidhi, Darwish. 1985. Lexical and sociolinguistic variation in Qatari Arabic. PhD diss., University of Edinburgh. Al-Essa, Aziza. 2008. Najdi speakers in Hijaz: A sociolinguistic investigation of dialect contact in Jeddah. PhD diss., University of Essex. ——. 2009. When Najd meets Hijaz: Dialect contact in Jeddah. In Arabic dialectology, ed. Enam Al-Wer and Rudolf de Jong, 203–222. Amsterdam: Brill. Al-Jehani, Nasir M. 1985. Sociostylistic stratification of Arabic in Makkah. PhD diss., University of Michigan. Al-Khatib Mahmoud. 1988. Sociolinguistic change in an expanding urban context. Ph.D. diss., University of Durham. Al-Wer Enam. 1991. Phonological variation in the speech of women from three urban areas in Jordan. PhD diss., University of Essex. 1997. Arabic between reality and ideology, International Journal of Applied Linguistics 7:2, 251–265. ——. 2002. Jordanian and Palestinian dialects in contact: Vowel raising in Amman. In Language change: The interplay of internal, external and extra-linguistic factors, ed. Mari Jones and Edith Esch, 63–79. Berlin: Mouton de Gruyter. ——. 2003. New dialect formation: The focusing of -kum in Amman. In Social dialectology: In honour of Peter Trudgill, ed. David Britain and Jenny Cheshire, 59–67. Amsterdam: Benjamins. ——. 2007. The formation of the dialect of Amman. In Arabic in the city: Issues in dialect contact and language variation, ed. Catherine Miller, Enam Al-Wer, Dominique Caubet, and Janet Watson, 55–76. New York: Routledge. Al-Qouz, Muna. 2009. Dialect contact, acquisition and change among Mananma youth, Bahrain. PhD diss., University of Essex. Bergsträsser, Gotthelf. 1915. Sprachatlas von Syrien und Palästina. Zeitschrift des Deutschen Pal ä stina-Vereins 38: 169–222. Blanc, Haim. 1969. The fronting of the Semitic G and the qāl-gāl dialect split in Arabic. In Proceedings of the International Conference on Semitic Studies, Jerusalem, 7–37. Cantineau, Jean. 1936/1937. Etudes sur quelques parlers nomades arabe d’Orient. Annales de l’Institut d’Etudes Orientales (Université d’Alger), II, 1–118; III, 119–227. ——. (1940). Les parlers arabes du Ḥōrān. Atlas. Paris: Klincksieck. ——. 1960. Cours de phonétique arabe: Suivi de notions générales de phonétique et de phonologie. Paris: Klincksieck. Chambers, Jack K. 2003. Sociolinguistic theory. Oxford: Blackwell.

Sociolinguistics Eckert, Penelope. 2000. Linguistic variation as social practice. Oxford: Blackwell. Eckert, Penelope and John Rickford. 2001. Style and sociolinguistic variation. Cambridge, UK: Cambridge University Press. Fisher, John. 1958. Social influences on the choice of a linguistic variant. Word 14: 47–56. Gordon, Elizabeth and Tony Deverson. 1998. New Zealand English and English in New-Zealand. Auckland: New House Publishers. Gordon, Elizabeth, Lyle Campbell, Jennifer Hay, Margaret Maclagan, Andrea Sudbury, and Peter Trudgill. 2004. New Zealand English: Its origins and evolution. Cambridge, UK: Cambridge University Press. Hachimi, Atiqa. 2005. Dialect leveling, maintenance and urban identity in Morocco. PhD diss., University of Hawaii. ——. 2007. Becoming Casablancan: Fessis in Casablanca as a case study. In Arabic in the city: Issues in language variation and change, ed. Catherine Miller, Enam Al-Wer, Dominique Caubet, and Janet Watson, 97–122. London: Routledge. Haeri, Niloofar. 1987. Male/Female Differences in Speech: an Alternative Interpretation. In Variation in Language: NWAV-XV at Stanford: Proceedings of the Fifteenth Annual Conference on New Ways of Analyzing Variation, edited by Denning, Keith M., Sharon Inkelas, John R. Rickford, and Faye McNair-Knox, 173–183. Stanford, CA: Stanford University. ——. 1996. The Sociolinguistic market in Cairo: Gender, class, and education. London: Kegan Paul International. Hary, Benjamin. 1996. The Ğīm/Gīm in colloquial urban Egyptian Arabic. In Israel Oriental studies XVI, studies in modern Semitic languages, ed. Shlomo Izre’el and Shlomo Raz, 154–168. Hûjrup, Thomas. 2003. State, culture and life-modes. Ashgate, UK: Aldershot. Holes, Clive. 1987. Language in a modernising Arab state: The case of Bahrain. London: Kegan Paul International. ——. 1990. Gulf Arabic. London: RoutledgeCurzon. Horvath, Barbara. 1985. Variation in Australian English: The sociolects of Sydney. Cambridge, UK: Cambridge University Press. Ibrahim, Muhammad H. 1986. Standard and prestige language: A problem in Arabic sociolinguistics. Anthropological Linguistics 28: 115–126. Ismail, Hanadi. 2007. The urban and suburban modes: Patterns of linguistic variation and change in Damascus. In Arabic in the city: Issues in dialect contact and language variation, ed. Catherine Miller, Enam Al-Wer, Dominique Caubet, and Janet Watson, 188–213. London: Routledge. ——. 2008. Suburbia and the inner-city: Patterns of linguistic variation and change in Damascus. PhD diss., University of Essex. ——. 2009. The Variable (h) in Damascus: Analysis of a stable variable. In Arabic dialectology, ed. Enam Al-Wer and Rudolf de Jong, 249–272. Amsterdam: Brill. Jabeur, Mohamed (1987). A Sociolinguistic study in Rades: Tunisia. PhD diss., University of Reading. Johnstone, T. M. 1963. The affrication of “kaf ” and “gaf” in the Arabic dialects of the Arabian Peninsula. Journal of Semitic Studies 8: 210–226. Kurath, Hans, et al. 1941 Linguistic atlas of New England. Providence RI: American Council of Learned Societies. Labov, William. 1963. The social motivation of a sound change. Word 19: 273–309. ——. 1970. The study of language in its social context. Studium Generale 23: 66–84.

Sociolinguistics ——. 1972. Sociolinguistic patterns. Philadelphia: University of Pennsylvania Press. ——.. 1990. The intersection of sex and social class in the course of linguistic change. Language Variation and Change 2: 205–254. ——. 2001. Principles of linguistic change: Social factors. Oxford: Wiley-Blackwell. Miller, Catherine. 2005. Between accommodation and resistance: Upper Egyptian migrants in Cairo. Linguistics 43: 903–956. Milroy, James, Lesley Milroy, Sue Hartley, and David Walshaw. 1994. Glottal stops and Tyneside glottalization: Competing patterns of variation and change in British English. Language Variation and Change 6: 327–357. Milroy, Lesley. 1987. Language and social networks. Oxford: Blackwell. Milroy, Lesley and Matthew Gordon. 2003. Sociolinguistics: Method and interpretation. Oxford: Blackwell. Owens, Jonathan. 1998. Neighborhood and ancestry: Variation in the spoken Arabic of Maiduguri (Nigeria). Impact: Studies in Language and Society 4. Amsterdam: John Benjamins. ——. [2006] 20092 . A Linguistic history of Arabic. Oxford: Oxford University Press. Schilling-Estes, Natalie. 2002. Investigating stylistic variation. In The handbook of language variation and change, ed. Jack Chambers, Peter Trudgill, and Natalie Schilling-Estes, 375–401. Oxford: Blackwell. Schmidt, Richard. 1974. Sociolinguistic variation on spoken Egyptian Arabic: A re-examination of the concept of diglossia. PhD diss., Brown University. Shorab, Ghazi. 1982. Models of socially significant linguistic variation: The case of Palestinian Arabic. PhD diss., State University of New York at Buffalo. Taqi, Hanan. 2010. Two ethnicities, three generations: Phonological variation and change in Kuwait. PhD diss., University of Newcastle. Trudgill, Peter 1974. The Social differentiation of English in Norwich. Cambridge, UK: Cambridge University Press. ——. 2004. New-dialect formation: The inevitability of colonial Englishes. Edinburgh: Edinburgh University Press. Woidich, Manfred. 1993. Bilbliographie zum Agyptisch-Arabischen. Ms. Amsterdam, I.M.N.O. Woidich, Manfred and Elisabeth Zach. 2009. The g/ ğ question in Egyptian Arabic revisited. In Arabic dialectology: In honour of Clive Holes on the occasion of his sixtieth birthday, ed. Enam Al-Wer and Rudolf de Jong, 41–60. Leiden: Brill. Wolfram, Walt. 1969. A sociolinguistic description of Detroit Negro speech. Washington, DC: Center for Applied Linguistics.

Notes: (1) For further elaboration on this point, see the discussion in Labov (1970: 30). (2 ) The ideal speaker in traditional dialectological research is described as a nonmobile old rural male (NORM) since much of the very early work in the field typically preferred such speakers. (3 ) A notable exception in this context is the research conducted by Clive Holes (1987) during the 1970s (Holes 1987) in Bahrain, which draws on all available resources from dialectology. (4 ) Of the published material during the early decades of the 20th century the works by Bergsträsser (1915) and Cantineau (1936, 1937, 1940) are worthy of mention. For a comprehensive bibliography on Egyptian Arabic, see Woidich (1993). 5

Sociolinguistics (5) Further important cases of reanalysis incorporating insights from sociolinguistics can be found in Jonathan Owens (2009). (6 ) This is not to say that sociolinguists are not interested in samples of formal or semiformal speech, but such samples normally show irregular patterns as a result of conscious self-correction on the part of speakers and are, therefore, not the ideal source of information to determine the course of language change. On the other hand, samples of unmonitored speech show “more systematic speech, where the fundamental relations which determine the course of linguistic evolution can be seen most clearly” (Labov 1970: 46). (7 ) See Labov (1970: 47) on the observer’s paradox; see Milroy and Gordon (2003) on methodology. (8 ) On the use of “style” in sociolinguistic research, see Eckert and Rickford (2001) and Schilling-Estes (2002). (9 ) This was found to be true for a wide range of variables in North American English (Labov 1972; Wolfram 1969), British English (Trudgill 1974), Australian English (Horvath 1985), and New Zealand English (Gordon and Deverson 1998). (10 ) Notwithstanding the obvious flaws in following this method in research on Arabic, some researchers have in fact resorted to reading tasks and contrasted them with speaking styles along the same stylistic continuum (e.g., Shorab 1982; Al-Khatib 1988), while Abdel-Jawad (1981) had the foresight not to use this method, reasoning that in reading tasks Arabic speakers are confined to choosing standard Arabic pronunciations, which, of course, is correct. (11) Topic was also used by Labov to manipulate style. Elicitation of different styles through change of topic and context (formal speeches delivered on TV or to the public) was used by Abdel-Jawad (1981) in Amman, by Schmidt (1974) in Cairo, by Al-Jehani (1985) in Mecca, and by Al-Amadidhi (1985) in Qatar among others. (12 ) The telling of jokes here seems to mimic Labov’s “danger of death” question, which was meant to elicit narratives that would engage the speaker emotionally. Al-Khatib, however, did not make a distinction within the range of the light topics he used. (13 ) There are sound grounds for discounting [q] in Jordan as a variant of the same phonological unit as the colloquial variants since its occurrence is lexically conditioned to a high degree. (14 ) There seems to be a tacit agreement about the inadequacy of such methods for Arabic; to my knowledge they have not been used in Arabic studies for quite some time. (15) In addition to the previously mentioned studies, Al-Wer (1991) uses level of education as a variable. A notable exception is Haeri (1996), who classifies her speakers in Cairo according to social class, but she also uses education as a variable to refer to “type of schooling.” (16 ) Abdel-Jawad (1981: 268) and Al-Khatib (1988: 130) are so impressed by these results that they even suggest replacing “age” by “education” as an indicator of change in apparent time. Not many linguists would take this suggestion seriously. (17 ) Al-Khatib does not comment on this result although it represents an anomaly within the argument he advocates, namely, that the highest educated group overall standardize their speech most. (18 ) James Milroy and Lesley Milroy introduce the notion of supralocal to sociolinguistic research and use it to explain the results in the Tyneside research with respect to glottalisation—glottal replacement of voiceless stops; see James Milroy et al. (1994). (19 ) The emphatic interdental and stop sounds have a different historical development. In Al-Wer (2003) I argue that, unlike the plain interdental and stop sounds, no mergers are involved in moving from emphatic interdental to the emphatic stop counterpart and that this development should be analysed as a straightforward case of sound change from interdental to stop (rather than merger) in the modern dialects. (20 ) By using these lexical items they do not of course change their phonology in the direction of the standard variety. On the contrary, their phonology was in fact changing in a direction away from the standard variety.

Sociolinguistics Similar findings are found in Holes (1987); Abdel-Jawad & Suleiman (1990); Haeri (1996). (21) Of course this is not an ideal situation as far as methodological rigor goes but was worth exploring. In hindsight, social network should have been integrated and quantified as a method of sampling from the start. (22 ) See Al-Wer (2002) for the full details of these results and their analysis. (23 ) Lesley Milroy’s (1987) research in Belfast in the late 1970s is concerned precisely with maintenance of innercity working-class Belfast linguistic features. To my knowledge, among studies of variation in Arabic, only Muhamad Jabeur’s (1987) study in Rades, Tunisia, employs the social network approach as a sampling technique as well as an analytical tool. (24 ) It is also worth noting that the vast majority of research studies on Arabic use adult speakers only in cases where it was practically possible to use educational level as a way of speaker selection. In more recent studies where children and teenagers are included in the sample the results show them to be the most innovative groups; see, for instance, the results in Al-Qouz (2009) and Al-Wer (2007). (25) Anecdotal evidence points in the same direction: pupils generally dislike Arabic lessons; upward social mobility is not attained by speaking standard Arabic but by adhering to the most prestigious local dialects or by using a foreign language. (26 ) Principle 2 states: “For stable sociolinguistic variables, women show a lower rate of stigmatized variants and a higher rate of prestige variants than men” (Labov 2001: 266). (27 ) Of the mainstream sociolinguistic publications since the revision of the interpretation of sex-differentiated language variation in Arabic, only Jack Chambers (2003: 157—162) revises the case of Arabic correctly. (28 ) We no longer see, for instance, large-scale surveys of the sort conducted by Labov in New York in 1968, but now investigations are smaller and more focused. (29 ) Full details of this research are in Ismail (2008); see also Ismail (2007, 2009). (30 ) Male-led changes are rather hard to come by in the sociolinguistic literature; for a summary, see Labov (1990). (31) Also see Hachimi (2007). (32 ) Also see Miller (2005: 904–906). (33 ) In her study in Detroit, Penelope Eckert (2000) introduces this method into sociolinguistic research; it employs an ethnographic approach and investigates speakers’ behavior closely in their communities of practice (Eckert 2000). (34 ) A particularly interesting research in this field is Catherine Miller’s (2005) study of Upper Egyptian migrants in Cairo, which investigated phonological, morphological, and morphosyntactic adaptations. (35) Also see Al-Essa 2008. (36 ) See Al-Wer (2003, 2007) for details about the process of the formation of the new dialect. Enam Al-Wer Enam Al-Wer, Essex University

Arabic Folk Linguistics: Between Mother Tongue and Native Language Yasir Suleiman The Oxford Handbook of Arabic Linguistics Edited by Jonathan Owens Print Publication Date: Sep 2013 Online Publication Date: Dec 2013

Subject: Linguistics, Languages by Region DOI: 10.1093/oxfordhb/9780199764136.013.0011

Abstract and Keywords This article addresses some long-standing issues in Arabic sociolinguistics. The starting point is the concept of diglossia, which has become the port of entry for any discussion of the semiliquid language situation in the Arabicspeaking world. It first outlines the most abiding criticisms against diglossia and then offers thoughts on these as a prelude to discussing Arabic folk linguistics. It is argued that a folk linguistic perspective should be incorporated in studying Arabic in the social world. This perspective is important in developing an insider understanding of the language that may be at odds with the findings of modern linguistics. To aid the process of developing this perspective, the article adopts the terminology and conceptual frameworks Arabic speakers use in describing their language situation wherever possible—hence, the choice of fusha and ‘ammiyya instead of any of their translations into English, including Classical Arabic and vernacular, which Haeri uses. Keywords: Arabic, sociolinguistics, diglossia, folk linguistics, language, of fusha, ‘ammiyya

11.1 Arabic Diglossia: A Port of Entry into a Semiliquid World THIS paper deals with some long-standing issues in Arabic sociolinguistics. Its starting point is the concept of

diglossia, which has become the port of entry for any discussion of the semiliquid language situation in the Arabicspeaking world. By way of an aside, the concept of izdiwājiyya lughawiyya “linguistic duality” was the subject of discussion and debate in the Levant in the 19th century (Khūri 1991), well before Ferguson’s seminal study of the topic in 1959 and William Marçais’s first use of the term diglossia in 1930. Discussions of this topic in the 1880s were centered on literacy and modernization, but the debates that took place in the emerging press, mainly from the Levant, were inconclusive as to whether diglossia represents an insurmountable hindrance to these societal goals. The fact that we are still debating diglossia and its impact on literacy and modernization in the twenty first century testifies to its durability as a phenomenon and a conceptual tool. Diglossia is too well-known in the literature to warrant restating here. Criticisms of diglossia are also well-known in the literature. But let me outline what I consider to be the most abiding ones and offer my thoughts on these as a prelude to what I want to say about Arabic folk linguistics. First, as expounded by Ferguson, diglossia is said to be impressionistic and fuzzy (Boussofora-Omar 2006). This is true; but this can hardly be regarded as a criticism, not least because Ferguson did not intend his paper (of some fifteen pages) on diglossia to be of an empirical kind similar to the research we are now accustomed to in the quantitative variationist paradigm with its emphasis on statistical correlations. Ferguson aimed at developing the parameters of the concept in a comparative fashion for further application and elaboration, and he has admirably succeeded in doing so judging by the impact of his research. There is a salutary lesson for us in this as Arabic linguists: we should not be afraid to conduct qualitative research, similar to Ferguson’s work on diglossia, because of the ability of this kind of research to tap domains of inquiry which the quantitative method may not be able to delve into owing to the limitations of the correlationst paradigm in its logico-positivist orientation (Suleiman 2011). Ethnographies of language such as Haeri’s (2003)

Arabic Folk Linguistics study of Egypt, with which I will deal later, is an example of what I have in mind. Second, it is often said that Ferguson’s discussion of diglossia created a dichotomy or a binary opposition between what he termed the high and low varieties of the language, the fuṣḥā “standard” and the range of ‘āmmiyyas “dialects” in the Arabic speaking world respectively. The “di-” in diglossia may have created this impression of bipolarity, aided in this regard by Ferguson’s attempt at clarity in conceptualizing the two outer ends of the linguistic situation he was describing at the expense of attending to the liquidity of the middle ground, which has been described as “slippery” and “messy” in the literature (Boussofara-Omar 2006: 630). The truth, however, is rather different. Ferguson was aware of intermediate forms of Arabic, including what Mitchell (1986) over two decades later called Educated Spoken Arabic. This is how Ferguson describes this intermediate form of the language (1959: 10): A kind of spoken Arabic much used in certain semi-formal or cross-dialectal situations [which] has a highly classical vocabulary with few or no inflectional endings, with certain features of classical syntax, but with a fundamentally colloquial base in morphology and syntax, and a generous admixture of colloquial vocabulary. Ferguson’s view of diglossia is, therefore, more nuanced than the “di-” in the term suggests. He is in fact aware that the spectrum between the end points of diglossia is a continuum of some sort that defies strict categorization. Criticisms that Ferguson’s conceptualization of diglossia posited a rigid “categorical” [a better term is categorial] divide between the high and low varieties of the language is, therefore, unwarranted (Mahmoud 1986). In spite of this, the search for the middle ground between the end points of diglossia was driven to some extent by this perceived dissatisfaction with the impression of bipolarity in Ferguson’s treatment of this phenomenon. Badawī’s well-known five “levels” (mustawayāt) of Arabic in Egypt is an attempt at creating a taxonomy that covers the full range of variation between (and involving) the high and low varieties (1973). In a similar vein, Meiseles (1980) recognizes four categories on the diglossic spectrum to overcome the same problem of ill definition. However, the fact that Arabic sociolinguists have not been able to agree on the number of Arabic levels or categories on the diglossic continuum, or on their ontology (whether they are levels/registers/styles or categories of self-contained classification) reflects the semi-liquidity or viscosity of the Arabic language situation at its outer ends and its liquidity in the middle. It further reminds us of the difficulties Ferguson had in drawing up his definition of diglossia. The third criticism is that Ferguson assigned prestige to the high variety only. It is, I believe, this allocation of prestige that underlies the semi-metaphorical description of the fuṣḥā as “high” and the ‘āmmiyyas as “low.” Ibrahim (1986) argued that this assignment of social value to the outer points of diglossia fails to take account of the prestige of some dialectal forms in the ‘āmmiyya domain. While this is true, it is still possible in my view to rescue Ferguson’s notion of prestige from this criticism. It is not difficult to argue in this regard that the prestige of an ‘āmmiyya form “dialect” is normally pitched in relation to other less prestigious ‘āmmiyyas in the orality domain, but hardly ever in my experience to the fuṣḥā as the written and standard form of the language. The weakness in Ferguson’s position is that he treated prestige as a unitary concept that applies to the full gamut of forms in the language indiscriminately, although it is very clear that he was primarily interested in the allocation of prestige across the written/standard versus oral/dialectal divide.

11.2 Arabic Folk Linguistics In spite of these criticisms I want to argue that Ferguson’s concept of diglossia is, to a large degree, still valid in the way he set it out. My argument is based on folk linguistics (Nieldzielski and Preston 2000), by which I mean the range of views and attitudes people have about their language, including its origin and the myths surrounding it, that “allow us to come closer to the overt or covert orientations, assumptions, and hidden ideologies of the community and how these relate to its linguistic repertoire” (Suleiman 2008: 28). These views are sociopsychological in nature: they tell us something about the role language plays in society as cultural capital and as a site of identity conceptualizations that feed into how the linguistic assets of the community are deployed (or not deployed) in education, the media, law, the religious domain and other fields of societal activity and organization, as well as about the various forms of institutional and grass roots resistance to alternative ways of organizing the public sphere linguistically. In modern linguistic theory, folk linguistic views and attitudes may be found to lack the authority and “objectivity” of the findings of modern linguistics, based as it is on empirical investigation and model

Arabic Folk Linguistics building. In spite of this folk linguistics is an important source of information in exploring the role of language in the life of a community. Most Arabic speakers are oblivious of the findings of modern linguistics. This is to some extent true of educationists, whether they are curriculum designers, teachers, inspectors or school administrators. What matters more to these constituencies of professionals in my experience are the socially available conceptualizations of the language, rather than those of modern linguistics, against which they frame their work as practitioners in their respective fields. Tapping these conceptualizations is,1 therefore, important in understanding how the public sphere is configured and why it remains resistant to importations from modern linguistics. Let me illustrate this by exploring the meaning some Arabic speakers attach to the term al-lugha al-‘arabiyya (“the Arabic language”). I have asked fifty illiterate speakers in Amman, Jordan all of whom Muslim women over sixty (it is hard to find Christian women in this age range who are illiterate), what this term means. I chose illiterate women only as informants in my interviews to neutralize, to the extent possible, the effect schooling may have on their answers. A variety of formulations and illustrative examples were provided in response to my question, but these coalesced around an understanding of the term which primarily designates the fuṣḥā, the minority view, or refers to all forms of Arabic including the fuṣḥā, the majority understanding. When the latter dominant meaning is intended my informants framed their answers in terms of a duality involving what professional linguists call fuṣḥā and ‘āmmiyya, which are seen as related but different. My informants are aware of the different spoken varieties and of the different values these may have in society. And they tended to impart greater coherence to the fuṣḥā which they evaluated highly as the language of the Qur’an and literacy.2 I, therefore, contend that their understanding of the term al-lugha al-‘arabiyya is not so very different from Ferguson’s characterization of how diglossia applies in the Arabic language situation. For these illiterate speakers, two forms of Arabic exist: fuṣḥā and ‘āmmiyya. These speakers are aware of intermediate forms of Arabic but they tend not to give them the same category recognition as the primary categories of fuṣḥā and ‘āmmiyya. These intermediate forms are described as different ways of speaking in a manner that corresponds to the notions of style and register in modern linguistics, being in some respects akin to Badawi’s notion of levels above with all its ambiguity. Ferguson’s notion of diglossia does not capture the full complexity of the intermediate forms of Arabic on the fuṣḥā to ‘āmmiyya continuum as these forms are constructed in modern linguistics but, in my view, it corresponds well to the folk linguistic understanding of this continuum in which the prominence of the two ends of this continuum in their semi-liquid state are accentuated at the expense of the fuzzy middle in all its liquidity.3 Because of this I think Ferguson’s diglossia has greater folk linguistic resonance than scientific/empirical linguistic validity, and this goes for his notion of prestige above. This folk linguistic understanding of diglossia is consistent with Mushira Eid’s observation that the “perceived dichotomy [between fuṣḥā and ‘āmmiyya] is deeply ingrained in the collective consciousness of Arabic speakers/writers,” whether in Egypt, which is her primary field of interest, or in other Arabic-speaking countries (2002: 204). It is also consistent with some of the folk linguistic views to which Haeri (2003) refers in her ethnographic study of Arabic in Egypt. In Haeri’s study ‘āmmiyya and fuṣḥā emerge as the bipolar forms of Arabic in relation to which Egyptians orient themselves ideologically and in terms of language behavior. Folk linguistics is important in enabling us to construct an insider or ethnographic perspective of the language situation in the Arabic speaking countries. To begin with, in Arabic folk linguistics the fuṣḥā is treated as indivisible into pre-modern and modern forms, which are sometimes called in Western discussions of the topic as Classical and Modern Standard Arabic respectively. This indivisibility of the fuṣḥā is also an integral part of most expositions of the Arabic language among Arab linguists in the indigenous tradition. Linguists/grammarians in this tradition would not subscribe to the plethora of Arabic varieties that some Western linguists recognize, such as those that Versteegh deploys in his book The Arabic Language (1997): Proto-Arabic, Old Arabic, Early Arabic, Classical Arabic, Early Middle Arabic, New Arabic, Muslim Middle Arabic, Christian Middle Arabic and Judaeo-Arabic. I think most Arabs would agree that there are differences between different manifestations of the fuṣḥā across the ages, but they are very reluctant to treat these differences as the basis for sub-categorization in the body of the language as Versteegh has done. While Arabic speakers are aware of linguistic change, they are more struck by linguistic continuity in their conceptualizations of fuṣḥā. It is in my view important to acknowledge this indigenous or folk linguistic perspective if we are to develop an insider’s understanding of Arabic by accessing the way Arabs construct their language. This insider perspective has its own socio-psychological validity which must be incorporated into any study of the social life of Arabic, even though it may not withstand the scrutiny of empirical linguistics.

Arabic Folk Linguistics Furthermore, folk linguistics enables us to appreciate the double commitment Arabic speakers have to ‘āmmiyya and fuṣḥā and to reconcile some of the attitudinal differences that pervade these commitments, such as the fact that native speakers do not always hold their spoken varieties with the same esteem as the fuṣḥā, even though ‘āmmiyya is their mother tongue and fuṣḥā is not. Even the most ardent supporters of ‘āmmiyya in the Western academy acknowledge this (e.g., Haeri 2003: 38). From a folk linguistic perspective the issue, therefore, is how to conceptualize the relationship Arabic speakers have with the two salient varieties of their language.

11.3 Mother Tongue and Native Language: A New Conceptualization The term mother tongue is used to designate the ‘āmmiyyas in modern linguistics. The ‘āmmiyyas are acquired informally and are the site of personal and community intimacy. Although the ‘āmmiyyas are sometimes committed to writing, the cultural norm is to use them in oral communication in everyday situations [Holes, “Orality”]. The use of “tongue” in designating the ‘āmmiyyas captures the orality of this form of Arabic well. In folk linguistics, the ‘āmmiyyas are not regarded as languages but as dialects. They are also thought to be devoid of grammar (a mistaken view in scientific terms) owing to their being perceived as corrupt forms of fuṣḥā [Retsö, “Arabic?”], [Owens, “History”]. Haeri’s (2003) use of vernacular to refer to the ‘āmmiyyas is, therefore, not completely consistent with this folk linguistic understanding of the Arabic language situation. The term vernacular evokes an emerging language on apar with the rise of the European languages as competitors of Latin in the print, ethnic or national identity domain in post medieval times. In other words, when applied to Arabic the term vernacular ascribes to the ‘āmmiyyas a status that they do not or may not have in comparison with fuṣḥā or in the eyes of those who speak them as mother tongues. It is true that there have been attempts at establishing the dialects as national languages, but to this day none of these attempts has succeeded to either establish these ‘āmmiyyas as widely recognized print or school languages, or even to dent the authority of fuṣḥā in the print or educational domain to any significant degree (Suleiman 2003). Whether a spoken form or variety of Arabic is a language or not is not, therefore, an issue which can be decided by modern linguistics on behalf of those whose medium of communication it is. As Davies rightly reminds us “on sociolinguistic grounds … dialects are dialects of the same language because their speakers claim them to be so, and they are distinguished from languages in terms of power” (2003: 58). So, if ‘āmmiyyas are the mother tongues of Arabic speakers what is fuṣḥā to them? Clearly fuṣḥā cannot be their mother tongue because it is acquired formally through education over a long period of time. Acquisition of fuṣḥā in this way is a rite of passage to literacy; one cannot achieve the status of being literate without schooling in fuṣḥā. In addition, fuṣḥā is a normative code whose inculcation is deliberate. The use of ‘āmmiyya in everyday speech is the default position in Arabic speaking countries; in comparison the use of fuṣḥā in the same domain is marked practice in the same way that the use of ‘āmmiyya in writing is. Clearly Arabs relate to the two ends of their linguistic repertoire differently. To capture this I have elsewhere designated fuṣḥā as the Arabs’ native language to set it apart from ‘āmmiyya as their mother tongue. The distinction between mother tongue and native language needs some theorization that I hope to provide following. In doing so I will follow Blommaert (2006: 243) who makes an important distinction between what he calls (following Silverstein 1996, 1998) linguistic communities on the one hand, and speech communities on the other where “the former are groups professing adherence to the normatively constructed, ideologically articulated ‘standard’ language … and the latter are groups characterized by the actual use of a specific speech form.” Blommaert glosses this distinction by saying that its two partners are not “isomorphic, and the distance between the sociolinguistically definable community and the linguistic-ideologically definable community reveals the degree of hegemony of language ideologies, often resulting in blind spots for sociolinguistic phenomena” (ibid.). Blommaert’s comment on this distinction draws attention to the ideological impregnation of folk linguistics. If we understand language ideology to be the “cultural system of ideas about social and linguistic relations, together with their loadings of moral and political interests” (Irvine 1989: 255), folk linguistics cannot but be an ideological hub in which the beliefs and attitudes of the community must be taken into consideration in any treatment of the social life of language. It would in fact be impossible to produce an ethnography of a language without tapping its language ideology no matter how mundane, obvious or politically problematic this ideology is. Thus, whereas modern linguistics establishes as an article of faith that all languages are equal in the way they fulfill the instrumental needs of their communities, folk linguistics, being ideologically impregnated, does not heed this principle of equality. From a folk linguistic perspective languages are often linked to myths of election which ascribe greater worth to them than their competitors whether in terms of beauty, lexical richness, logicality and so on. For the modern linguist all

Arabic Folk Linguistics languages are structurally and instrumentally equal; in folk linguistics some languages are more equal than others.4 Each view operates within its own domain from which it derives its validity. Treating fuṣḥā as native language and ‘āmmiyya as mother tongue, and linking these two dualities to the concepts of linguistic and speech community respectively, allows us to construct two triadic chains that are not mutually exclusive in describing the Arabic language situation from a folk linguistic perspective.5 The first chain consists of fuṣḥā, native language and linguistic community and corresponds to the high end of Ferguson’s diglossia. The second consists of ‘āmmiyya, mother tongue and speech community and corresponds to the low end of the diglossia spectrum. I am aware that ideology is not exclusively linked with fuṣḥā as native language. ‘Āmmiyya as mother tongue has its own ideological associations in cultural and political terms. However, I have chosen to highlight the ideological in the fuṣḥā in constructing the above categorical distinction because the ideology in this unit of analysis, more than the ideology of ‘āmmiyya, is the subject of hegemonic formulations in folk linguistics insofar as national identity construction is concerned. The above separation of mother tongue and native language may be problematic in modern linguistics because, ordinarily, the two terms are used interchangeably in the literature (Bonfiglio 2010). The overlap between the two is reflected in talk about the native speaker as the native speaker of his mother tongue and of his native language in instrumental terms. However, in the process of conceptualizing these two notions as notational variants of each other differences between them are also acknowledged, the most important of which is the overwhelming orality of the mother tongue in comparison with the written nature of the native language that links it to standardization and schooling. In some cases the empirical distance between mother tongue and native language is not very great; in others it is. It is the fact that empirical distance does exist between the two notions in modern linguistics that allows me to deploy them differently in dealing with the Arabic language situation and, furthermore, to link them to the two notions of speech community and linguistic community. These two communities overlap but they are sufficiently different in category terms to be configured differently in folk linguistic terms. Arabic speakers, according to this, belong to overlapping but different communities that are defined in terms of speech and language, and that correspond to the low and high forms of the language in Ferguson’s diglossia. Consistent with the autonomy of the triadic chains I have established above and their folk linguistic orientation, I propose to designate the mother tongue in Arabic as al-lisān al-umm and the native language as al-lugha al-umm. Furthermore, I am of the view that most theories of language in the social world cannot totally escape the empirical fields that inform their construction, in spite of the claims of universality that may be attributed to them. Most theories of language are in fact restricted theories with aspirational aims to universality. Matthews (1972) made a similar point over three decades ago in developing his views on word and paradigm morphology and the close fit this model has with inflectional languages, using Latin as an example. For this reason the conflation of mother tongue with native language cannot be extended to all language situations without doing some of those situations injustice. Applying to the Arabic language situation theoretical frameworks of restricted nature that do not sufficiently differentiate between mother tongue and native speaker is, in my view, a form of Procrustean and epistemic violence that is similar, in many respects, to enforcing a separation of these two categories on those language situations that do not need it or are best described without it. Whereas in the one situation we would have a case of under categorization: in the other we would commit the error of over-categorization. What we need, therefore, is a different solution that deploys different conceptual horses for different sociolinguistic courses that are in tune with the particularities of their folk linguistic terrain. A native speaker of a language does not, in sociolinguistic terms, have to have full fluency in it as a spoken or written language to qualify as native speaker. As Davies perceptively comments the “politicisation of native speakerness with particular reference to language” enables individuals to “regard themselves (and others) as native speakers for symbolic rather than [instrumental] purposes” (2003: 76). Native speakerness, I wish to argue, does not have to be yoked to the performance-bound notions of authenticity, reliability and representativeness which cognitive and applied linguistics demand of the native speaker to a high level (Coulmas 1981). In folk linguistic terms one can operate with lower levels of performance across these three criteria as long as these levels are linked to a structure of “attitudes and feelings of identity” that define a sense of belonging to a national community in symbolic terms (Davies 2003: 76). An Arabic speaker does not have to have the same facility in fuṣḥā as he has in his ‘āmmiyya for him to be recognized in folk linguistic terms as a native speaker of the former in the same way he is recognized as a native speaker of the latter. This is a deviation from cognitive linguistics which may in fact claim that fuṣḥā and ‘āmmiyya are wired differently in the brain—suggesting

Arabic Folk Linguistics that they may in fact behave as different languages—but it reflects a societal attitude which sociolinguistics must capture if it is to come to grips with the social life of a language, and how this language resonates with those who think they belong to it and it belongs to them [Ryding, “Acquisition”]. I am, however, willing to accept that some Arabic linguists may still feel uncomfortable with my use of native language to designate fuṣḥā, which is often said to be no one’s native language in studies of Arabic (Haeri 2003: 146), in spite of the fact that in the educational field it is referred to in Arabic as the mother language (al-lugha alumm). Fuṣḥā can be legitimately described as an indigenous language in folk linguistic terms, in spite of the fact that some writers have described it as a foreign language in parts of the Arabic speaking world and have accused it of being a force of internal colonialism/imperialism in those parts, for example Egypt (Ahmed 2000). To avoid the unnecessary debate about “indigenousness,” I prefer the term native language in referring to fuṣḥā to capture the Arabic term al-lugha al-umm Arab educationists consistently use to describe it. This term is also preferable to official language that is available to us. Some, if not most, Arab constitutions establish fuṣḥā as “(the) official language of the state” (al-lugha al-rasmiyya), but I find this term wanting in folk linguistic terms. While some of my informants in Jordan were aware that Arabic was the official language of the state, the legal meaning of this term did not resonate with them. Fuṣḥā is also considered as the “national” (al-lugha al-waṭaniyya) or “pan-national” (allugha al-qawmiyya) language in the Arabic speaking countries, but people do not use this identity inflection as their primary compass in orienting themselves towards this form of Arabic, at least in the Middle East (in some countries of North Africa the situation may be different). For many speakers fuṣḥā is equally, if in fact not more, connected to Islam as an extra-territorial phenomenon that cuts across nation-states. To reflect this structure of associations I prefer to stick to native language, which as a term has the ability to hover over all these conceptualizations without discrimination. As a native language fuṣḥā can be indigenous, official, national or pannational separately or together without contradiction.6 . I have also rejected the possibility of treating fuṣḥā as a second language. If we adopt this perspective we would have to designate ‘āmmiyya (in its capacity as mother tongue) as first language. But the use of the first and second language terminology would skew the relationship between fuṣḥā and ‘āmmiyya into a bilingualism framework that clashes with the folk linguistic conceptualizations of the Arabic language situation. None of my informants considered themselves to be bilingual, and I cannot imagine that many Arabs would describe themselves as such in describing their relationship with fuṣḥā and ‘āmmiyya. I have seen many applications for jobs in the Arab world where knowledge of languages is mentioned. Only in two cases did the applicants list their ‘ āmmiyya and fuṣḥā, together with English in one case and English and French in another, as part of their language repertoire. These two cases were the source of amusement by those who evaluated the applications, a very telling reaction. Furthermore, the use of the first and second language terminology favors the instrumental role of language over its symbolic function as the two classificatory principles underlying ‘āmmiyya and fuṣḥā respectively. This may be justifiable on cognitive grounds but it fails to accord with Arabic folk linguistics wherein fuṣḥā seems to be given greater social value than ‘āmmiyya, as Haeri (2003) notes in her ethnography of the language situation in Egypt. Adopting a folk linguistic perspective we might, in fact, be justified to treat fuṣḥā as first language and ‘āmmiyya as second. To avoid the clash of instrumentality and symbolism as principles of classification on the one hand, and cognitive and folk linguistics on the other it would be preferable to drop the terms first and second language in talking about the Arabic language situation in favor of mother tongue and native language. The fact that a conceptual distance exists between mother tongue and native language in the linguistic literature creates a terminological space which allows us to reserve the latter as a designation of fuṣḥā. To this we may add that alternative terms are problematic; they introduce theoretical perspectives which sit uncomfortably or stray outside folk narratives of the language. Native language, while not a perfect term for fuṣḥā because of some of the cognitive loadings it has is, nevertheless, serviceable not least because it can allow the other terminological narratives above to operate in ethnographic accounts of the Arabic language situation.

11.4 Native Language and Mother Tongue: Custodianship versus Ownership To reflect the difference between fuṣḥā and ‘āmmiyya in social terms and how speakers relate to them Haeri applies the concepts of custodianship to the former (which she calls Classical Arabic) and ownership to the latter (which she calls the vernacular) in her ethnographic study of Egypt. The same classificatory schema would

Arabic Folk Linguistics presumably apply to other Arabic speaking countries; Haeri signals this to be the case but she does not explicitly generalize beyond Egypt. Haeri links fuṣḥā to sacredness and ‘āmmiyya to the “mundane,” quotidian or “profane” (ibid. 3, 105)—hence the opposition in the title of her book between sacred language and ordinary people—and uses this distinction as the basis of her claim about custodianship and ownership. Haeri vacillates between ascribing sacredness to the language of the Quran as revelation or to fuṣḥā in all its totality, including the Quran as revelation, but there is a clear sense in her discussion that she favors the second more inclusive application of the term. In my fieldwork in Jordan similar vacillation existed for some informants, but all informants reserved sacredness to the language of the Qur’an only as revelation when questioned about the topic. In my experience ordinary Muslims do not distinguish sharply between Quran and muṣḥaf (text of the Quran as in a printed copy of it), but they do make a clear distinction between the printed word when it is a manifestation of the revelation, and when print is used in ordinary every day domains. One can test the sacredness thesis, therefore, by comparing the reaction to dropping the text of the Quran (muṣḥaf) on the floor and the reaction to dropping a school book, novel or newspaper. In the former case, Arab Muslims would immediately pick up the muṣḥaf, kiss it and ask for God’s forgiveness as if they had committed a sin. I have never seen the same reaction in dealing with other manifestations of the printed word. In these situations Arab Muslims will pick up the book, novel or newspaper but without anything like the urgency or feelings of transgression against the divine they exhibit in the other case. I have in fact used this litmus test, so to speak, to help my informants decide whether fuṣḥā in its totality is or is not a sacred language. It is true that fuṣḥā is treated with veneration, especially in religious contexts, but “veneration” (taqdīs, tabjīl) is not the same as “sacredness” (qudsiyya) in the way Haeri seems to use this term. Haeri articulates the juxtaposition between ‘āmmiyya and fuṣḥā as sacred language in “terms of ownership/custodianship, arbitrariness of the sign, translatability and human versus divine origin” (ibid.: 143). She also says that fuṣḥā as “the language of the Quran continues to separate the sacred from the profane, writing from speaking, and prescribed religious rituals from personal communication with God” (ibid.: 1). As an example of the profane, or “the most profane of activities” Haeri mentions “selling” (ibid.: 105). The arbitrariness of the sign and translatability are linked in Haeri’s exposition. Because Muslims believe that the Quran as God’s revelation is His word verbatim, Haeri argues that the relationship between form and meaning in the Quran is non-arbitrary and extends the same to fuṣḥā in its totality. Thus, while it is possible to translate (translation is always an act of interpretation) the meaning of the Quran into other languages, any such translation cannot be regarded as a Quran because it breaks the non-arbitrary relationship between form and meaning in the revelation. I believe this to be fundamentally wrong. A better way of conceptualizing the relationship between form and meaning in the Quran would be theological/doctrinal inseparability, which is different from non-arbitrariness in the sense this term is understood in the literature following Saussure (1966, originally published in French in 1916). For Saussure the arbitrariness of the sign refers to the lack of a natural or other knowable reason for linking a particular form with a particular meaning or vice versa. According to this interpretation it would be possible to describe the link between form and meaning in the language of the Quran as arbitrary without challenging the sacredness of the revelation. This in my view would apply regardless of whether or not we ascribe to Arabic a “divine” (tawqīf) or “human” (isṭilāḥ) origin. Arabic grammarians have debated this issue centuries ago. The theologians have done the same in dealing with the “createdness” (khalq) of the Quran. The Arabic linguistic and theological traditions contain ample evidence to suggest that the arbitrariness of the sign thesis is compatible with the view of the theological inseparability of form and meaning that underlies the Quran as sacred revelation.7 The untranslatability of the Quran thesis is not a linguistic principle but a primarily theological one. In fact if we accept the arbitrariness thesis as a feature of all languages, including fuṣḥā, we would be able to say that a translation and its original are not the same, no matter which pair of languages is involved. When we translate, we translate meaning not form. And is not this similar to the untranslatability thesis of the Quran? A translation of a poem by the Palestinian poet Mahmud Darwish is not a poem by Mahmud Darwish but a rendering of it in another language, making the original poem as a duality of form and meaning untranslatable if by translation we mean sameness. Translation does not aim to produce sameness but to minimize loss (Hervey and Higgnis 1992), although in the case of such texts as the Quran that are deeply embedded in their cultures translation loss can be extreme owing, among other things, to their reception history. This is why translations of the Quran tend to veer towards foreignization rather domestication (Venuti 1995), wherein the reader is taken to the text in the former and vice versa in the latter. What distinguishes the untranslatability of the Quran from the untranslatability of a Mahmud Darwish poem, therefore, is the theological impregnation of the former and its rootedness in the intellectual history

Arabic Folk Linguistics of Arabo-Islamic culture as integral features of the sacredness of revelation. In addition to the above conceptual problems it is not clear where the sacred ends and the profane begins for Haeri. If selling, as she claims, is one of the “most profane” activities with which fuṣḥā would need to be reconciled (ibid.: 105) in the Arabic speaking world how would one deal with the fact that, in its long history, fuṣḥā was the medium of poetry and other prose compositions that dealt with such activities as homosexuality and wine drinking that unquestionably belong to the domain of profanity? Arabs may object to and deplore these activities or their depiction in poetry on doctrinal grounds, but very few people would do the same because these activities are the topics of poetic compositions in fuṣḥā. Fuṣḥā is both the language of the sacred and the profane (and all that lies between them) in spite of the fact that, attitudinally, it is linked to Islam as faith and culture. These linkages do not make Arabic sacred, contrary to what Haeri says. Furthermore, does not the fact that the vast majority of the Arabic grammarians excluded most of the Prophetic Traditions from their corpora suggest that, outside the strict confines of revelation, sacredness is an extremely tenuous notion when applied to fuṣḥā? If the second most important source of Islamic legal thinking (tashrī‘), representing the “tradition” (sunna) of the Prophet, is excluded from the purview of grammatical description and model building in the early periods of Islam, 8 does not this suggest that Muslims operated with a heavily circumscribed understanding of sacredness in the linguistic domain? Expanding the sacred beyond its limits as Haeri does is not just unwarranted on various grounds, but it also leads to creating the problematic distinction between custodianship and ownership in her description of the Arabic linguistic landscape. Haeri argues that because fuṣḥā is a sacred language, Arabs cannot be its owners because they do not have the right over it (ḥaqq al-taṣarruf, p. 146). She explains this as follows: “if language is considered to be the Word of God then its users are its custodians, not its owners. If so, the right to change it, mould it, translate it, negotiate its boundaries and so on is always contested” (ibid.: 69–70). If we accept this description of the meaning of custodianship, then the speakers of most languages, especially those with a long written tradition, are closer to being their custodians than their owners. Attempts to change a language of this pedigree, mould it and negotiate its boundaries will always be contested. The linguistic literature is full of examples that testify to this. Furthermore, producing new translations of sacred texts, even when these texts are not regarded as the Word of God verbatim, is always the subject of controversy and contestation because of the history of reception of established translations which the new translations seek to replace or supplement. Producing new translations of the Bible in English in the twentieth century has not been free of controversy and contestation on linguistic, theological and aesthetic grounds. The fact that contestation in the case of Quran translations tends to be accentuated because of its status as revelation in form and meaning in doctrinal terms makes this contestation special in terms of degree, not kind. Furthermore, fuṣḥā has undergone considerable changes over time and, as Haeri rightly points out, some Arabs believe that a modern inflection of fuṣḥā has developed that “stands on its own, independent of the fuṣḥā of religion” (ibid.: 21). In most cases, linguistic change takes place surreptitiously with little deliberate interference from the users of the language. The idea of deliberate and directed change in language is the exception not the rule, as the work of language academies testifies (see Cooper 1989). This is true of fuṣḥā and Egypt as it is true of standard French and France. Haeri links the presumed sacredness of fuṣḥā with the long-standing issue of the checkered history of modernization in Arab societies. In doing this, she argues that modernization in the full sense cannot be achieved without replacing fuṣḥā with ‘āmmiyya in all fields in society, including translating the Quran into ‘āmmiyya (ibid.: 148). The following quotation sets some of Haeri’s view on the topic (ibid.: 146): A sacred language cannot become a fully living language unless it loses that status. Since modernisation efforts [in Egypt] began, renovated forms of the language have carried with them at every turn, all the contradictions, accusations, uncertainties and struggles that are inherent in attempts to make a sacred language contemporary. Whether one calls changes to the language “modernisation,” or “simplification” the point is that change necessarily goes against purity and sacredness. Put simply, [fuṣḥā] is not the mother tongue of Egyptians or other Arabs and, not being its “owners,” their rights to the language will remain precarious. The dilemma is that were they to take steps to own it, the language would cease to be sacred. Few if any of these problems would have arisen had fuṣḥā remained simply the language of religion, while other spheres of life would be served by writing in Egyptian Arabic.

Arabic Folk Linguistics Furthermore, Haeri links the dominance of fuṣḥā with a “certain denial of the contemporary self” in Egypt, leading to “an uneasy relation with self ” (ibid.: 148). She also argues that the “status of being a citizen is put in doubt [in Egypt] in the absence of any acknowledgement of that citizen’s mother tongue” (ibid.: 151). In the same vein, fuṣḥā as official language is said to act as an “obstacle to participation in the political realm” and as one reason for the “absence of democracy in Egypt” (ibid.).9 It is true that language, Self, citizenship and democracy are linked in the social sphere, but ascribing some of the tensions and deficits in these domains to the sacredness of fuṣḥā is a step too far, not least because this sacredness is very much in doubt as a property of fuṣḥā in its totality. As we have seen above, the sacredness of fuṣḥā is an exception not the rule of Arabic folk linguistics. We must, therefore, be careful before we inject linguistic sacredness into the analysis of modernization in Egypt by linking it with various ills in society. Calls to replace fuṣḥā by ‘āmmiyya are not new in Egypt. They go back to the nineteenth century when Willcocks (1893) argued that the Egyptians cannot regain their “power of invention” (quwwat al-ihtirā‘) until they use their spoken language as the medium of education and modernization.10 Willcocks, who was ignored and reviled in Egypt, did not talk about the sacredness of Arabic (Suleiman 2004: 62–72). Haeri acts more boldly but it is unlikely that her analysis would resonate with most Egyptians simply because it clashes with the folk linguistic conceptualization of the Arabic language situation. I have dealt with this topic obliquely in my discussion of alShūbāshī’s book (2004) on reforming Arabic grammar, with its provocative title li-taḥyā al-lugha al-‘arabiyya:Yasquṭ Sībawayhi: “Long Live the Arabic Language! Down with Sibawayhi” (Suleiman 2006). Haeri (ibid.: 119, 121), however, is right when she talks about the perilous state of Arabic language teaching in schools and the poor image of Arabic teachers who, she tells us, are referred to disparagingly as the “Grammar Brothers” (al-ihwān al-nahwiyyīn) by analogy with the well-known party of the Muslim Brothers (al-ihwān al-muslimīn).

11.5 Conclusion I have argued in this paper for incorporating a folk linguistic perspective in studying Arabic in the social world. This perspective is important in developing an insider understanding of the language which, it must be acknowledged, may be at odds with the findings of modern linguistics. To aid the process of developing this perspective we will need to adopt the terminology and conceptual frameworks Arabic speakers use in describing their language situation wherever possible; hence my choice of fuṣḥā and ‘āmmiyya instead of any of their translations into English, including Classical Arabic and vernacular which Haeri (2003) uses. Being approximations of the Arabic originals, foreign terms skew the meanings of their Arabic equivalents by injecting into them target language attitudes, ideologies and reception histories that may be absent from the original counterparts. The term vernacular, as I have argued herein, is a case in point. The insider perspective in folk linguistics is important in two important ways. On the one hand, it enables us to understand the deployment of language as a cultural asset in society, together with the structure of feelings and attitudes that inform and accompany this deployment. On the other hand, we can use folk linguistic views to understand the limits of language modernization or, even, to predict the obstacles that may face this modernization in the educational sphere, how to negotiate these obstacles and the ways and means that may be adopted to soften up societal resistance to educational reform. I have also argued in this paper that in spite of the criticisms of Ferguson’s diglossia in modern linguistics, this concept has a great deal of validity in folk linguistic terms. I have, therefore, used this concept in its folk linguistic incarnation to argue for a description of the Arabic language situation in which two conceptual chains operate: (1) fuṣḥā, linguistic community and native language (lugha umm), and (2) ‘āmmiyya, speech community and mother tongue (lisān umm). I believe that these triadic chains capture how the Arabs think about their language situation. This includes the supremacy of language symbolism in the first chain and instrumentality in the second. This analytical framework provides the parameters of a new conceptualization of the Arabic language situation that, while building on Ferguson’s seminal work on diglossia, aims to transcend it. Little work has been done on Arabic from an ethnographic and qualitative perspective. Haeri’s study (2003) is one of a few exceptions in this regard. For this reason it is a welcome addition to the study of Arabic in the social world. In spite of this, however, I disagree with some of Haeri’s main principles and conclusions in describing the Arabic language situation in Egypt, which, in some respects, intersect with the views expressed by writers such as Ahmed

Arabic Folk Linguistics (2000), ‘Awaḍ (1947, 1965), Mūsā (1947), Safouan (2007), and Willcocks (1893). These writers share a vernacularizing perspective whose aim is to establish the ‘āmmiyya of Egypt as its national language. It is not my aim here to deal with the pros and cons of this perspective, but I will highlight three issues with which any future move in the direction of vernacularization will have to contend. First, some Arabic speakers will read vernacularization sociopolitically as an attempt to weaken the cultural and political ties among the Arabic speaking countries. Vernacularization will, more importantly, be further read as a sinister attack whose aim is to undermine the ties which Arab Muslims have with the text of the Quran and the cultural edifice it has given rise to and continues to underpin. These readings will, in turn, raise the issue, at a heightened level in society, as to whether the benefits that may accrue from vernacularization would justify the social and political upheaval it is bound to create. Second, vernacularization will raise extremely divisive issues in status and corpus planning terms (Cooper 1989). These will include the selection of the base variety or varieties for constructing the new vernaculars in Arabic speaking countries; the choice of scripts for rendering the vernaculars in writing; the authoring of new grammars, lexica, style manuals and spelling rules; the production of new curricula for schools and institutions of higher education; and teacher training programs to equip the teachers with the skills of teaching the new languages to students. Finally, the corpus planning consequences of vernacularization will not be cost free: they will require substantial investment in time and effort that may not be justified by the benefits promised by the vernacularizers.

References Ahmed, Leila. 2000. Border passage. New York: Penguin Books. ‘Awaḍ, Luwīs. 1947. Plutoland. Cairo: Matb’aṭ-al-Karnak. ——. 1965. Mudhakkarāt ṭālib ba‘tha. Cairo: al-Kitāb al-Dhahabī. Badawī, El-Sa‘īd. 1973. Mustawayāt al-‘arabiyya al-mu‘āṣira fī Miṣr. Cairo: Dār al-Ma‘ārif. Blommaert, Jan. 2006. Language policy: and national identity. In An introduction to language policy—Theory and method, ed. Thomas Ricento, 238–254. Oxford: Blackwell Publishing. Bonfiglio, Thomas Paul. 2010. Mother tongues and nations: The invention of the native speaker. New York: Walter de Gruyter. Boussofara-Omar, Naima. 2006. Diglossia. In EALLI, ed. Kees Versteegh et al., 610–637. Cooper, Robert. 1989. Language planning and social change. Cambridge: Cambridge University Press. Coulmas, Florian. 1981. The concept of native speaker. In A festschrift for native speaker, ed. Florian Coulmas, 1– 25. The Hague: Mouton. Davies, Alan. 2003. The native speaker: Myth and reality. Clevedon: Multilingual Matters. Eid, Mushira. 2002. Language as choice—Variation in Egyptian women’s written discourse. In Language contact and language conflict in Arabic variations on a sociolinguistic theme, ed. Aleya Rouchdy, 203–232. London: Routledge Curzon. Ferguson, Charles. A. 1959. Diglossia. Word 15: 325–340. Haeri, Niloofar. 2003. Sacred language, ordinary people. New York: Palgrave Macmillan. Hervey, Sàndor and Ian Higgns. 1992. Thinking translation: A course in translation method. London: Routledge. Hussein, Riad F. and Nasir El-Ali. 1988. Subjective reactions towards different varieties of Arabic. al-Lisān al-‘arabī 30: 7–17. Ibrahim, Muhammad. 1986. Standard and prestige language: A problem in Arabic sociolinguistics. Anthropological Linguistics 28: 115–126.

Arabic Folk Linguistics Irvine, Judith. 1989. When talk isn’t cheap: Language and political economy. American Ethnologist 16: 248–267. Kaye, Alan. 1972. Remarks on diglossia in Arabic: Well-defined vs. ill-defined. Linguistics: 81: 32–48. Khūri, Yūsuf Qazmā, ed. 1991. Najāḥ al-umma al-‘arabiyya filughatihā al-aṣliyya. Beirut: Dar al-Ḥamrā.’ Mahmoud, Youssef. 1986. Arabic after diglossia. In The Fergusonian impact: In honour of Charles A. Ferguson on the occasion of his 65th Birthday, ed. Joshua Fishman, Andree Tabouret-Keller, Michael Clyne, Bh. Krishnamurti and Mohamed Abdulaziz, 239–251. Berlin: Mouton de Gruyter. Matthews, Peter. 1972. Inflectional morphology: A theoretical study based on aspects of Latin verb conjugation. Cambridge: Cambridge University Press. Meiseles, Gustav. 1980. Educated spoken Arabic and the Arabic language continuum. Archivum Linguisticum 2: 118–148. Mitchell, T. F. 1986. What is educated spoken Arabic? International Journal of the Sociology of Language 61: 7–32. Mūsā, Salāma. 1947. al-Balāgha al-‘aṣriyya wa l-lugha al-‘arabiyya. Salāma Mūsā li-l-Nashr wa-l-Tawzī‘. (First published in 1945.) Nieldzielski, Nancy A. and Preston, Dennis R. 2000. Folk linguistics. Berlin/New York: Mouton de Gruyter. Parkinson, Dilworth. 1991. Searching for modern Fuṣḥā: Real life formal Arabic. Al-’Arabiyya 24: 31–64. Safouan, Moustapha. 2007. Why are the Arabs not free—The politics of writing. Oxford: Blackwell. de Saussure, Ferdinand. [1959] 1966. Course in general linguistics, ed. C. Bally and A. Sechehaye. New York: McGraw-Hill Company. Silverstein, M. 1996. Encountering languages and languages of encounter in North American ethnohistory. Journal of Linguistic Anthropology 6: 126–144. ——. 1998. Contemporary transformations of local linguistic communities. Annual Review of Anthropology, 401– 426. Suleiman, Yasir. 2003. The Arabic language and national identity: A study in ideology. Georgetown: Georgetown University Press. ——. 2004. War of words: language and conflict in the Middle East. Cambridge: CUP. ——. 2006. Arabic language reforms, language ideology and the criminalisation of Sibawayhi. In Grammar as a window onto Arabic humanism: A collection of articles in honour of Michael Carter, ed. Lutz Edzard and Janet Watson, 66–83. Wiesbaden: Harrassowitz Verlag. ——. 2008. Egypt: From Egyptian to pan-Arab nationalism. In Language and National Identity in Africa, ed. Andrew Simpson, 26–43. Oxford: Oxford University Press. ——. 2011. Arabic, Self and identity: A study in conflict and displacement. New York: Oxford University Press. Al-Shūbāshī, Sharīf. 2004. Li-taḥyā al-lugha al-‘arabiyya: Yasquṭ Sībawayhi. Cairo: Madbūlī (third printing). Venuti, Laurence. 1995. The translator’s invisibility: A history of translation. London: Routledge. Versteegh, Kees. 1997. The Arabic language, Edinburgh: Edinburgh University Press. Weiss, Bernard. 2009. Waḍ‘ al-lugha, In EALL IV, ed. Kees Versteegh et al., 684–687. Willcocks, William. 1893. Lima lam tūjad quwwat al-ikhtirā‘ladā al-miṣriyyīn al-’ān?. Al-Azhar 6: 1–10.

Notes: 1

Arabic Folk Linguistics (1) For a classic paper on the topic see Parkinson (1991) which uses the matched guise technique to investigate reactions to various forms of Arabic. (2 ) See Hussein and El-Ali (1988) for evidence from Jordan that corroborates this view. (3 ) See Kaye’s discussion of diglossia (1972) for the notions of “well-defined” and “ill-defined” which are relevant in this context. (4 ) Sociolinguistics accepts this principle of inequality, often expressed in the different prestige/status values attributed to different languages and to different varieties within a language. (5) This characterization of native language and linguistic community would imply that, for some speakers in Lebanon and North Africa, French may qualify as a native language, albeit the case that French lacks the cultural depth of Arabic and its wider meanings in religious and political terms in the MENA region (see footnote 6 below for further elaboration of this point). See Davies (2003) for the linkage among politics, symbolism and instrumentality in framing the concept of native speaker. (6 ) Although French may be established as “native language” in some countries of North Africa and Lebanon, it is not normally conceptualized as “indigenous,” “official” or “national” in the way these epithets are applied to fuṣḥā. This situation varies from country to country and between different segments of the populations in the same country depending on ethnicity, education and faith. (7 ) Arbitrariness as a property of the linguistic sign is subsumed under the concept of waḍ’ al-lugha (the founding of language) in the Arabic linguistic tradition [Larcher, “ALT II”]. Weiss (2009: 684) characterizes this relationship as follows: “The relationship between … utterances and … meanings was not [considered in this tradition] as a natural or intrinsic relationship. In principle, an utterance could have any meaning. That an utterance had a particular meaning was due entirely to its being assigned to that meaning. The meaning of an utterance had to be learned; it could not automatically be known from the utterance.” An utterance here covers any unit of form and meanings regardless of size. (8 ) The lateness of the texts of the ḥadīths is an important factor in their exclusion from the empirical data of most grammarians. The fact that they were considered as the second source in fiqh wasn’t sufficient to redeem them in the eyes of the grammarian. (9 ) It would be interesting to reflect on this view in the context of the Arab Spring and Egypt’s place and role in it, but this would take us outside the scope of this study. (10 ) The Egyptian Salāma Mūsā championed this solution in the twentieth century. For a discussion of his views see Suleiman (2003: 180–90). Yasir Suleiman Professor Yasir Suleiman is His Majesty Sultan Qaboos Bin Sa'id Professor of Modern Arabic Studies and a Fellow of King's College.

Orality, Culture, And Language Clive Holes The Oxford Handbook of Arabic Linguistics Edited by Jonathan Owens Print Publication Date: Sep 2013 Online Publication Date: Dec 2013

Subject: Linguistics, Languages by Region, Sociolinguistics DOI: 10.1093/oxfordhb/9780199764136.013.0012

Abstract and Keywords This article explores the relationship between linguistic form and function in the varying cultural landscapes of the contemporary Arabic-speaking world, including spontaneous speech, the contemporary electronic media (television, radio, the Internet), cinema, theater, and traditional performed oral literature, which have been revived and “reinvented.” It is shown that the relationship between orality and language in Arabic is complex. The layman’s mental landscape is of a “high,” literary, codified variety of the language strongly identified with a unifying religion (Islam) and a “golden age” of past imperial and literary glories, carrying great cultural prestige; and a “low,” chaotic (often regarded as grammarless) but homely variety associated with domesticity, intimacy, and the daily round. The emotional resonances of the two varieties are and always have been different. Consequently, they have, through the ages, occupied separate functional niches in all linguistically mediated communication, be it speech, writing, song, poetry, cinema, or theater. Keywords: Arabic, linguistic form, function, speech, electronic media, cinema, theater, oral literature

12.1 Speech

12.1.1 Conversation NATURAL, unmonitored talk in Arabic is almost always in a dialectal variety of the language. Arabic dialects are

defined primarily by geography, but there are also distinctions based on social variables within the geographical boundaries of any dialect area and some variables that cross-cut geography. Religion or sect is associated with dialect difference in some places, notably Baghdad and other parts of Iraq (“Muslim” vs. “Christian” dialects and, prior to the early 1950s, “Jewish” dialects; Blanc 1964) and Bahrain (“Sunni” vs. “Shiʿite” dialects1; Holes 1983a, 1983b; [Behnstedt and Woidich, “Dialectology”]). Historical lifestyle is also an important dialect variable and essentially opposes Bedouin-descended dialects to sedentary ones, so that Bedouin dialects of different countries (e.g., Syria, Jordan, Egypt) may have as much or more in common with each other than they do with dialects of sedentary origin in the same country, which are geographically nearer. (Rosenhouse 1984 provides a general characterization of Bedouin dialects.) And, increasingly, a speaker’s response to exposure to nonlocal varieties of the language, whether Modern Standard Arabic (SA or MSA) 2 or other dialects, may be reflected in his speech production, though this phenomenon differs from one Arab country to another3 and is also often related to gender.4 Generally speaking, in cross-dialectal situations, local features tend to be leveled and replaced by those of a regionally dominant dialect, often that of a capital city or of some other variety associated with a socially prestigious group. But the extent to which “dialectal leveling” actually occurs is related not just to the speaker’s degree of exposure to external influences but also to contingent pragmatic factors in the speech context, such as who he is talking to, what is being talked about, the roles of the participants, and the setting and the purpose of the interaction. In more educated speakers’ styles, in addition to vernacular–SA mixing,5 there is often a great deal of

Orality, Culture, And Language hybridization, whereby elements from the dialect are combined with elements from SA. This is noticeable in major sentence constituents such as the verb phrase, where saliently SA lexical elements (e.g., verb stems) are combined with dialectal morphemes expressing categories like tense, number, and negation, and the noun phrase, where, again, saliently SA lexical items may be combined with dialectal morphemes expressing, for example, deixis. As an Egyptian example, compare the following ways of saying “he isn’t accepted”: (1) SA: lā yuqbalu Cairene: ma byit’abalš hybrid: ma byuqbalš It can be seen from this that the dialectal discontinuous negative morpheme, ma …š and b- tense/mood prefix are combined with a saliently SA passive verb stem to form the hybrid form. Examples of the converse process—for example, the combination of vernacular lexical items with SA function morphemes for tense, voice, negation, number, and deixis—do not occur except in the humorous mimicking of the speech of foreigners or of the illiterate’s attempts to speak SA. The kind of mixed and hybridized language illustrated is typical of unscripted conversation in media chat shows and programs involving interviews with politicians, journalists and educated people generally and occurs across the Arabic-speaking world. As noted, the hybridizing of the SA and dialectal systems is by no means random, even if there is a certain degree of fluidity in what occurs, and it appears that the asymmetrical nature of the combinatorial relationships between the lexical and functional morphemes of SA and the dialects previously noted holds good for all Arabic-speaking speech communities. However, the exact nature of the constraints on how the two systems may be combined has yet to be worked out in detail.6 Code switching between Arabic and English has become something of a fashion among the Western-educated youth of several Arab countries. This kind of speech is known as ‘Arabīzī, a portmanteau word coined from ‘arabī (“Arabic”) and inglīzī (“English”), which gave its name to a 52-minute film on the subject made in 2006. ‘Arabīzī involves a constant to and fro between sentences or phrases in the speaker’s dialect and in English (often American accented). Many of those who use this speech style had an English-medium education and work in fields in which English is becoming the global “default” language—international banking, commerce, law, journalism. They often admit to a poor mastery of SA. An example of ‘Arabīzī, taken from the film of the same name, is as follows: ismi ʿayša xālidī, mawlūda bil-kwēt, and I also lived in Egypt when I was younger, u ʿišt arbaʿ snīn fi amrīka fil-jāmʿa…I’m the sports editor of the Daily Star. I love that job; it’s amazing. I was a journalism major so it worked out for me. ‘umrīarba’ u ʿišrīn sana u ana atḥačča ʿArabīzī. [My name is Aysha Khalidi, born in Kuwait, and I also lived in Egypt when I was younger, and I lived for four years in America at university…I’m the sports editor of the Daily Star. I love that job; it’s amazing. I was a journalism major so it worked out for me. I am 24 years old, and I speak ʿArabīzī.] Conversational code switching between Arabic and French is a phenomenon of much longer standing in North Africa (Bentahila 1983; Heath 1989) and occurs with other European languages among émigré groups (see Nortier 1990 for Dutch–Moroccan Arabic). See also Section 12.1.3 on computer-mediated communication (CMC) for examples from Internet chat rooms [Davies et al., “Code Switching”].

12.1.2 Monologue Where the speech is scripted, as in many types of formal monologue (political speeches, sermons in a mosque or church, lectures, media news bulletins), the language variety used is usually SA, but always with some degree of phonological influence from the speaker’s underlying vernacular. However, this statement needs qualification. The purpose of a political speech is often to convince an audience of nonpoliticians, who will have varying degrees of competence in SA. Both these considerations can lead speakers to resort to the vernacular at certain points, either to make themselves clearly understood or to use as a tactic of persuasion (see Mazraani 1997 for evidence from Egypt, Libya, and Iraq). Recounting personal experiences or telling humorous anecdotes to underline a point, for example, or simply showing sympathy with their audience’s concerns, are rhetorical moves that take the speaker off script and into affective and interpersonal domains in which the dialect is strongly favored. Dialect is also often

Orality, Culture, And Language resorted to when some political axiom, couched in SA “officialese,” needs to be explained. The late president of Egypt, Gamāl ʿAbdul-Nāṣir, was a master of this kind of code switching, which he often used quite deliberately to manipulate the emotions of his audience (Holes 1993). Sermons are another type of monologue in which code switching commonly occurs, when, for example, the imam needs to explain or “translate” into the vernacular religious material in the Classical language that the audience may otherwise not fully understand (Bassiouney 2006). However, this is a matter of choice, and some extremely popular religious orators, such as the fiery Egyptian critic of Anwar Sadat, ʿAbdul-Ḥamīd Kišk (d. 1987), derived part of their popularity from their ability to deliver their messages in flawless Classical Arabic, however little some of their audiences may have understood what they were saying. Similarly, public figures who cultivate an image of regal dignity as part of their public persona, such as Sultan Qaboos of Oman or the late King Hussein of Jordan, avoid the use of dialect in public speaking. This is also typical of groups of speakers for whom dialect avoidance is an element in a cultural agenda. The Egyptian literary poet and literary critic Fārūq Šūša, who appears frequently on popular Egyptian radio and TV shows, is one such, and his radio program “Our Beautiful Language” has been running continuously since 1967. In his case, it is part of a personal campaign to persuade the wider public to actively use and engage with Classical Arabic language and literature. Islamic radicals in the public eye, such as spokesmen for the Palestinian group Hamas and the Lebanon-based, Iran-backed Hezbollah, always use SA in their speeches but also even in impromptu media conversations and interviews. Here the reasons seem to be the dogma that CA, the language of the Quran, is the only true form of Arabic and the only one fit to carry their religiously inspired political message. Media news bulletins are overwhelmingly the preserve of SA, but, as al-Batal (2002) shows, dialectal Arabic has made inroads here: the local news on the Christian-owned LBCI channel of Lebanese TV has recently begun to use a form of “vernacularized” SA for its local news bulletins. This, and the increased use of French and English in the channel’s general programming, is seen as an assertion of Christian identity in post– civil war (1975–1990) Lebanon.

12.1.3 Computer-Mediated Communication According to a recent report by Etling et al. (2009), the Arabic blogosphere is organized primarily around countries, with Egypt by far the biggest cluster, followed by Saudi Arabia and Kuwait. There are Levantine–English “bridge” blogs in the eastern Arab World and Maghreb–French ones in North Africa, in which bloggers frequently use the European language alongside Arabic. The bloggers are predominantly young and male, and the most important issues for most of them relate to their personal life and local issues, with religion and human rights featuring strongly. An offshoot of CMC (as in the West) has been the commercially published blog, which has generated at least one best seller. This is Ghāda ʿAbdul-ʿĀl’s ʿAyza Atgawwiz7 (I Want to Get Married), written entirely in Egyptian colloquial Arabic and published as a book in 2008. The author is a 29-year-old Egyptian pharmacist, who claims to be speaking on behalf of the 15 million unmarried Egyptian women aged between 25 and 35 who are being pressured by society to marry, “even though it’s not their fault they are on the shelf.” The book is a selection of the author’s often witty and satirical Internet posts on the subject of marriage and contains apparently personal anecdotes. This blog book was turned into an Egyptian TV serial in 2010 and has even been translated into English.8 The Arabic “speech” of CMC (email, Internet chat rooms, message boards, blogs, social networking sites), insofar as can be judged from the wide variety of orthographic conventions that have arisen, presents a similar picture of variation related to extralinguistic factors. Normally, the Arabic spelling reflects the grammatical structures and vocabulary of the contributor’s vernacular (pronunciation is less easy to tell), though in “conversations” on certain topics, especially Islam, the variety used is usually SA, or the contributor’s best effort to write it. But there are local peculiarities. In Egypt, the use of dialectal Arabic in CMC is particularly prevalent; there is even an “Egyptian Wikipedia,” parallel to the pan-Arab Wikipedia, whose rules forbid the use of any Arabic variety other than Egyptian (effectively, this means Cairene). In particular, some Copts (Christian Egyptians) write only in Egyptian vernacular, thereby foregrounding their “Egyptian identity” and eschewing the use of SA, which they see as part of a “panArab identity” they do not share.9 When, for technical reasons, the Arabic alphabet is not available to the Internet user or in cell-phone text messaging, a romanized “Arabic chat alphabet” is sometimes used. This uses digits and other symbols for certain Arabic letters whose sounds do not occur in Western languages (though there is some variation in the 10

Orality, Culture, And Language correspondences).10 A typical example is: (2) alsalam 3alikom wa ra7mato Allah wa barakatoh which would appear in the normal Arabic script as: (3) and would be phonetically transcribed as: (4) as-salāmu ‘alaykum wa raḥmatu llāh wa barakātuh “Peace be with you and God’s mercy be upon you.” On websites frequented by speakers of North African dialects, there is a great deal of code switching and code mixing (as there is in ordinary speech) between the dialect and French, all elements of which are expressed in romanized script. For example (Arabic elements written here in italic): 11 (5) Je n’ai pu me retenir 7ta yfout le tribunal “I was unable to hold myself back until he entered the court.” rani ndirhaf wa7ed l’institut privée “I am doing it in a private institute.”

12.1.4 Arabic Pidgins An Arabic-based Pidgin is widely used in the Arabian Gulf States and Saudi Arabia as a consequence of the largescale immigration of construction workers, laborers, shop assistants, nannies, maids, and other predominantly unskilled workers from the Indian subcontinent and southeast Asia, beginning in the mid-1960s (Smart 1990; Næss 2008; al-Azraqi 2010). A similar phenomenon has been noted in Lebanon for Sri Lankan maids (Bizri 2005). By 2006, over 80% of the population of the United Arab Emirates (UAE) was non-Arab, the vast majority of whom were from India, Pakistan, Bangladesh, and Sri Lanka (Naess 2008: 21). Many of the men live in construction site work camps, and both the amount and type of linguistic contact these migrant workers have with the native Arabicspeaking population is severely limited. The Gulf Pidgin that has developed has all the classic features of Pidgin languages: compared with the superstrate language, Gulf Arabic, it shows a lack of inflectional morphology, preference for analytical structures, reduced verbal, nominal and pronominal paradigms, commonly a single preposition, and a tendency toward subject–verb–object (SVO) word order (Romaine 1988: 25–31). It is typically used in formulaic and usually fleeting encounters with native Arabic speakers in the workplace and marketplace and, though now seemingly stable, shows no signs of Creolization or of being replaced by Arabic learned as a foreign language, given the legal impossibility of the migrants gaining citizenship, however long they stay, and the social distance between them and the indigenous population [Tosco and Manfredi, “Creoles”].

12.2 Poetry and Song

12.2.1 Poetry Poetry is the paramount verbal art form in all Arab societies and has been so over the course of their recorded history. The earliest poetry that has come down to us (from the 6th and 7th centuries AH) was, as far as we can tell, memorized oral recitation that had to conform to strict schemata of scansion and rhyme to qualify for the designation shiʿr (“poetry”; the basic meaning is “knowledge or perception by means of the senses”). Its social functions in preliterate Arabia were many: as a repository of tribal history; as a vehicle for satire, elegy, praise, and the giving of advice; for describing nature; for expressing love; for celebrating the hunt. The original audience for this poetry was the poet’s fellow tribesmen. After the Islamic conquests, as literacy slowly began to take hold in the towns, the old poetic genres were developed and embellished as new literary influences from conquered territories exerted themselves. Traditional oral poetry continued in the Bedouin milieu but went almost completely unrecorded, except as a reservoir of Bedouin (= “pure”) language for medieval philologists and dictionary makers; serious poetry was a literate and literary art of the city and the court. And, in contrast to the oral tradition that continued to reflect the evolving language of speech, the language of poetry and writing more generally adhered

Orality, Culture, And Language to the fixed rules of CA, which, certainly by no later than two centuries after the Islamic conquests, (i.e., by no later than c. 235/850) was no one’s mother tongue (if it had ever been anyone’s). Occasionally in the medieval period we catch glimpses of this continuing but now subaltern oral tradition. Ibn Xaldūn (1332–1382), the North African philosopher of history, quotes several examples of the oral Bedouin odes of his time, using them to expatiate on the difference between “eloquence” (balāġa) and “grammatical correctness” (’i‘rāb) and condemning the literary critics who confounded the two and who dismissed this poetry’s claims to eloquence on the grounds that its language did not conform to the inflectional system of CA (Rosenthal 1958: vol. 3, 414–415; [Larcher, “ALT II”]). These odes are, like pre-Islamic ones, monorhymed and nonstrophic. Slightly before Ibn Xaldūn, the Iraqi Ṣafīyaddīn al-Ḥillī (1278–1348), presents a treatise (Hoenerbach 1956) on the poetics of the “seven arts” of the dialectal poetry of his era. These poems are different from the Bedouin odes cited by Ibn Xaldūn in that they are strophic and have different rhyme schemes. Some of them were based on models imported from Andalusia; others were developed in Iraq. The modern practice of colloquial poetry continues to reflect these ancient differences: the contemporary so-called nabaṭī Bedouin poetry of Arabia is structurally similar to the Bedouin material quoted by Ibn Xaldūn and to that of the pre-Islamic period. On the other hand, many modern types of dialectal poetry from the urban and rural milieus of Egypt, Lebanon, Syria, and Iraq seem to have developed from the types described by Ṣafīyaddīn al-Ḥillī. Occasional early modern examples have survived: one such is a celebrated 17th-centuryspoof poem by an imaginary Egyptian named Abūš Šādūf (lit. “the man who operates the counterpoise,” a device used for raising irrigation water from the Nile, used for centuries by Egyptian peasant farmers) to which a “learned commentary” by one Yūsuf al-Širbīnī is appended. The poem describes in mockheroic terms the wretchedness, squalor, and ignorance of Egyptian peasant life, adding further ridicule via the device of the fake commentary.12 Today, dialect poetry is composed in every Arab country and is a popular art form in the true sense, being composed by people at very top and at the very bottom of the social pyramid and all points in between. Although originally a purely oral art form and often practiced by the illiterate, composing poetry in dialect is today often also the choice of the educated. The spoken dialects evoke a set of cultural resonances different from those of SA. Because they are the natural way for hundreds of millions of Arabs to articulate their feelings about the ups and downs of everyday life, often in vividly concrete and personal ways, they are the natural poetic idiom for ridicule, satire, and vituperation and, in a region in which a tight rein is kept on the media, hence for expressing dissenting political opinion. This applies to poetry composed in the Bedouin tradition just as much as it does to the urban one. In Egypt, the educated dialect poet Mahḥmūd Bayram al-Tūnisī (1893–1961) composed much colloquial verse that was critical of the Egyptian establishment and the British colonial powers of the early 20th century and suffered decades of enforced exile for his pains. His compatriot of the next generation, Aḥmad Fu’ād Nigm (1929–), who was at the height of his fame in the 1970s at the time of Anwar Sadat’s presidency and darling of the student movement of that time, produced a torrent of linguistically inventive satirical verse on social inequality and political corruption in Egypt. This landed him and his partner, al-Šayx Imām (d. 1995) who put many of his barbs to music, more than once in prison. Their work circulated clandestinely on cassette tapes and faxes, and for a number of years it was illegal in Egypt to own such material.13 Similarly, the Iraqi ʿAbbūd al-Karxī (1861–1946) provides a running poetic commentary in Iraqi Arabic on all aspects of the politics of the Iraq of his time. Like Mahḥmūd Bayram al-Tūnisī in Egypt, he was hounded for it by the authorities of the British mandate and afterwards. His successor and cultural heir was ʿAzīz ʿAlī (c. 1911–1995), who wrote amusing barbed monologues in Baghdadi Arabic on many aspects of social and political life that he set to music and performed on Baghdad radio and television. Today, the exiled Iraqi dialect poet ‘Abbās Jījān is a well-known figure on Arabic satellite television channels, and some of his most memorable compositions, critical of Saddam Hussein’s regime and of the American invasion of Iraq that brought it down, have attained global celebrity by being posted as video clips on YouTube. In a similar vein, the modern Bedouin colloquial poetry of Jordan and Sinai has reinvented itself as a means for making critical comment on a host of social and political issues: where once the focus was on intertribal conflict, nowadays it is more likely to be anti-Bedouin discrimination, police brutality, high taxes, government corruption, the weak leadership of the Arab countries, and the political situation in the Middle East (Holes and Abu Athera 2009). In a lighter vein, in the Gulf tradition there are poems that treat issues such as the pros and cons of marrying a nonArab woman, the lax morals of the young, and the “keeping up with the Joneses” mind-set of some Gulf Arabs (Holes and Abu Athera 2011). Individual poets’ work circulates by a variety of means: word of mouth, on cheap cassettes, even by text messaging, and, in the case of better-known poets, in locally printed books and

Orality, Culture, And Language occasionally the newspapers. But colloquial poetry is finding new audiences well beyond its traditional ones. In the emirate of Abu Dhabi, one of the seven that makes up the UAE, a satellite TV colloquial poetry competition began in December 2006 called šā‘ir il-milyūn (“Poet of the Million”)14 and is now in its fifth season. It has become a huge hit not just with UAE viewers but also more widely in the Gulf and the neighboring areas of the Arab World. Poets aged between 18 and 45, and not just from Arabia and the Gulf but from countries as far away as Syria, Jordan, and Iraq, recite a colloquial poem before a theater audience, which is commented on by a judging panel of experts and voted via text messaging from the viewers of the program, similar to “X Factor,” the UK (and now global) singing competition. The program proceeds on a “knock-out” basis, and the winner receives a banner and a large cash prize. The poems can be on any subject, including politics, but they must be in the colloquial. In the most recent season of the program, a fully veiled Saudi housewife won third prize for a poem criticizing some of the extreme fatwas issued by the religious authorities in her home country, causing much adverse comment and even death threats. The program now has its own dedicated TV channel, as does a parallel one, Amīr aš-Šu‘arā’ (The Prince of Poets), in which poets of the Classical language compete with one another. These televised public displays of poetic talent play to the age-old self-image of the Arabs (and the Gulf Arabs in particular) as a nation of poets. It is one element in the creation of an identity in which the past is being reclaimed and rebranded for the post-oil generations—generations whose adoption of Western lifestyles, cultural preferences, and even “the loss of Arabic” is beginning to alarm Gulf governments.

12.2.2 Song The relationship between song and poetry in Arabic-speaking communities has always been close. Traditional Bedouin oral poetry was often, and occasionally continues to be, sung to the accompaniment of the one-stringed rabāba (“rebec”) and, in rural Egypt, a variety of simple string and woodwind instruments: the simsimiyya (similar to a lyre), the mizmār, and arġūl (types of reed pipe) are still in use. From the earliest times, poems or parts of poems in the Classical language were also put to music.15 This tradition has continued to the present day, and some of the most famous songs of the greatest modern singers are in this category. Most notable is Umm Kulthūm (c. 1900–1975) of Egypt, whose lyrics include many poems in the Classical language by the Egyptian neoclassical poet Aḥmad Šawqī and the Syrian Nizār Qabbānī as well in the colloquial by Egyptian poets such as Maḥmūd Bayram al-Tūnisī (q.v.) and Ṣalāḥ Jāhīn. Her lifelong lyricist Aḥmad Rāmī also wrote for her largely in the colloquial. It is important to note that what we are terming colloquial language in song, as in poetry, is not the same as ordinary speech. In love songs in particular,16 many stock figures (e.g., “the backbiter,” “the criticizer,” “the liar”) are drawn from a poetic repertoire shared for centuries by Classical and colloquial poetry as well as many shared similes and figures of speech. Apart from these features, the main formal difference between the colloquial of song/poetry and ordinary speech is in a somewhat more classicized pronunciation and morphology. The repertoire of the feted Lebanese singer Fayrūz was also partly in SA, partly in the colloquial, but with a strong bias toward the colloquial end of the range, as defined here. It is instructive to examine the types of song in which she used different registers of the language. By and large, songs whose sentiments are personal or love oriented or which describe the concrete details of the real, physical world are in the colloquial; songs that address bigger, grander, and more abstract themes are in the Classical language. A good example of the contrast is provided by two famous songs about Palestine, both of which date from 1971, four years after the end of 6-Day War with Israel. In the first, titled on the album sleeve al-Quds al-‘Atīqa (though pronounced in the song in the colloquial fashion al-’ids il-‘atī’a) “Old Jerusalem,” Fayrūz leads the listener through the streets of the ancient eastern (Arab) part of the city and talks to the shopkeepers about recent shocking events. She receives the gift of a flower vase from them, “the people who wait” (sc. for deliverance) and imagines the nameless Palestinian families who worked for centuries “under the sun, in the wind” to build the city, only to see their houses torn town by “the black hands” that battered down their doors and erected barbed wire fences. All of this domestic, physical description is delivered in succinct Lebanese colloquial in a song lasting three minutes. But in another song titled Zahrat al-Madā’in (Flower of Cities), from the same album, al-Quds fi l-Bāl (Jerusalem on My Mind), the mood and language are quite different. In this song, Fayrūz starts by declaring that she is praying for the sake of Jerusalem, the “city of prayer,” and meditates on the connections of the three monotheistic faiths with it: “our eyes,” she sings, apparently speaking on behalf of all Palestinians, “wander through the colonnades of the temples, embrace the ancient churches, and wipe away the sorrow from the mosques.” This yearning reverie continues with a lilting evocation of “the child in the grotto, whose mother is Mary” and “two faces weeping” (those of Mary and Joseph—Fayrūz is a Christian) but is eventually interrupted by the urgent, martial call of trumpets and painful memories of Jerusalem’s fall at the hands of (Israeli) troops, in which “peace was martyred in the city of peace,” followed by defiant predictions of its

Orality, Culture, And Language reconquest by warriors who will ride atop “terrifying steeds.” The song’s mood then shifts back once more to contemplative reverie and ends with the singer’s fervent hope that peace will once more be restored to the ancient city. The emotional and musical canvas is on the grand scale—the song lasts for eight and a half minutes, with full orchestra and choir—and the words are delivered, as befits the elevated sentiments and religious and nationalist symbolism, in flawless Classical Arabic. The linguistic contrast with the intimacy and domesticity of “Old Jerusalem” could hardly be starker. For a period in the 1970s and 1980s, such was its popularity that “Flower of Cities” became a virtual Palestinian national anthem, a point that underlines that “popular” in the linguistic culture of Arab speech communities does not always equate to “colloquial.” The commercial popular music of the present day, whether Egyptian, Levantine, or Gulf love songs, or songs that carry a social or political “message” in genres such as Algerian rāy (lit. “opinion”) and Tunisian mizwid normally uses the local form of the colloquial. In the case of rāy, this often involves, as in ordinary North African speech, a mix of French and Arabic. A novel feature of the music scene in the Arab world today is the way local musical traditions have become popular outside their original areas, largely as a result of satellite television and the Internet. Gulf popular music, for example, in a style known as xalījī (“Gulf ”), has a following in countries as far afield as Tunisia, despite the difficulty listeners must initially have had in understanding the dialect of the lyrics. But the general homogenization of speech caused by increased contact with and exposure to nonlocal forms of Arabic has had an effect here too. Furthermore, with the passing of a generation of Egyptian superstars like Umm Kulthūm (d. 1975), Abdul-Ḥālīm (d.1977), Farīd al-Aṭraš (d.1974), and Muḥammad ‘Abdul-Wahh āb (d. 1991), the foursome that dominated Arab popular music for most of the 20th century, has come a new and more diverse linguistic and musical era that coincides with the weakening of Egyptian political and cultural domination over the rest of the Arab World. It was the superstar of Algerian rāy, Chebb Khalid (though now, at the age of 50, he has dropped “Chebb” [“youth”] from his stage name) who was invited to sing one of his (nonpolitical) songs at the opening of the 2010 football World Cup in South Africa—which he duly did in Algerian Arabic.

12.3 Cinema, Theater, and Television Drama Drama in the western sense is not an indigenous art form in the Arabic-speaking world. It first appeared in the 19th century as a result of cultural contact with Europe (Sadgrove 1996) and posed an immediate linguistic problem. The impact of any play relies in part on the language of its dialogue: it should be “speakable” and based on an observable or (in the case of historical dramas) imagined social reality. This means that the language should normally have been the colloquial. But for many Arab playwrights, drama was seen first and foremost as a form of “high” art and therefore the natural preserve of CA/SA, which is not the natively spoken form of the language for any Arab. This issue was perceived to be less acute in dramas set in the Arab past, in which the common (though historically false) folk-belief that normal speech back then was indeed in CA could be exploited. An example is Maḥmūd Taymūr’s play Ibn Jalā (A Man of Celebrity), which portrays the career and character of the famous 8thcentury Umayyad governor of Iraq, al-Ḥajjāj ibn Yūsuf, in which the dialogue is in high-flown CA throughout. But what to do about the dialogue in plays set in the present-day and in a specific geographical location? Dialect, and geographically specific dialect at that, seemed to be required for any semblance of sociolinguistic reality. Yet there was hesitation here. In the play “Parents and Children,” a contemporary social drama set in a small Lebanese town written in 1917 by the Lebanese Christian author Mixā’īl Nu‘ayma, the educated characters speak SA (even when drunk) and the uneducated speak dialect, even when talking to each other. The result is artificial and stilted, even comical and absurd (Badawi 1988: 137). A contemporary of Taymūr’s, Tawfīq al-Ḥakīm, tried to come to grips with the conundrum in another way, in a play he wrote in the mid-1950s, al-Ṣafqa (“The Deal”). In it he wrote the dialogue in what he termed “the third language,” a form of Arabic that conformed as far as possible to the syntactic rules of CA/SA on the printed page, avoiding lexical choices that were either saliently local or standard. Because normal Arabic orthography underspecifies the phonological realization of words, this meant that the text could be performed in something akin to the dialect of any group of actors who shared this vocabulary (roughly, that of the eastern Arab World) via the omission of most (unmarked) grammatical inflections and the performing of the short vowels (always unmarked in normal Arabic script) according to their own local dialect. But the result of this experiment was unnatural, an artificial form of Arabic that was neither fish nor fowl, and unlike the real “third language” that educated Arabs speak on a daily basis (see Section 12.1.1). Tawfīq al-Ḥakīm’s idea was quickly abandoned, and writers simply adopted their normal spoken language as the default choice for drama set in the contemporary world. In fact, the “foreignness” of CA/SA as a means of spoken communication has sometimes been

Orality, Culture, And Language exploited for comic effect. In the Egyptian writer Yūsuf Idrīs’s play Gumhūriyyat Faraḥāt (“Faraḥāt’s Republic”), the following farcical exchange takes place between Faraḥāt, a beaten-down suburban policeman, and an equally miserable peasant woman. The exchange occurs when Faraḥāt is taking a statement from her, and she completely fails to understand the SA “officialese” (in bold in the quotation) required of a representative of the state, which he then “translates” for her into Egyptian colloquial Arabic (Idrīs 1981:101): (6) Policeman: yā bitt…halladayki’aqwālun’uxrā? Woman: ’uxrā ēh yā sīdi? Policeman: ‘ayza t’ūli ḥāga tānya, ya‘ni?! Policeman: Now then, girlie…Do you have any further statements to make? Woman: “Further” is what, sir? Policeman: I mean, got anythin’ else you wanna say?! Idrīs made more extended dramatic use of spoken SA in his play al-Laḥẓa al-Ḥariga (The Critical Moment), set in Egypt at the time of the Suez Crisis of 1956. The dialogue is in Egyptian spoken Arabic, except for the words of the British soldiers. In reality, of course, they would have spoken English (perhaps punctuated with some Egyptian “kitchen Arabic” of the “šufti bint” type common among the troops), but Idrīs makes them speak SA, presumably to create a similarly foreign and alienating effect. The actual effect, however, is merely jarring and unnatural. Historical subjects, on the other hand, especially religious ones, have continued to be scripted in SA/CA, as have, in the main, the Arabic translations of the works of classic western playwrights like Shakespeare.17 The same principles apply, a fortiori, to the dialogue of drama in cinema and television, both of which have a much broader popular appeal than the theater. From the 1930s until roughly the 1980s, Arab cinema, and the ever popular TV musalsal, or “drama serial,” often in 13 but sometimes many more parts, were dominated by the Egyptian film and television industries. The setting for the hundreds of films and serials made in this period was almost invariably Egypt—usually Cairo or Alexandria—with Egyptian actors and Egyptian dialogue. This had the side effect of accustoming generations of non-Egyptian Arab moviegoers and television audiences to the rhythms and vocabulary of Egyptian speech at a time when many of them had hardly traveled outside their own town or village.18 The exception has again been historical costume drama, particularly if it has a religious content. The blockbuster film al-Risāla (The Message), an account of the rise of Islam, made in separate Arabic and English versions in 1976 and funded by Libya, used only CA/SA for the Arabic dialogue. Although many of those who saw it would have struggled to understand the dialogue, that did not prevent it being a huge international hit around the Arab (and indeed non–Arabic-speaking parts of the Muslim) world. Since the 1980s, however, the Egyptian film industry has been in relative decline. Although Egyptian films and serials remain popular, there are now many rival sources of production, particularly for the TV market—Syria, Lebanon, Jordan, and the Gulf States—with plots, scenarios, and dialects to match. But where the aim is seen as educative rather than merely entertaining, as in much children’s programming, a simplified form of SA/CA is often used, which can come quite close to the “third language” Tawfīq al-Ḥakīm experimented with in the 1950s. A good example of this is Iftaḥ yā Simsim! (Open, Sesame!), an Arabic version of the U.S.-produced “Sesame Street” that became enormously popular throughout the Arab world in the 1980s and that employed a simplified form of SA for dialogue between the puppet characters.

12.4 Writing Although this chapter is concerned with language and orality, a few remarks will be made here about language levels in prose writing in the modern Arabic-speaking world since in certain text types orality is bound to be an issue. Any Arab creative writer with pretensions to having his work considered as “serious” literature has traditionally composed it in SA. That was axiomatic until perhaps 30 years ago. While it still remains the case that SA is overwhelmingly the vehicle of choice for serious creative prose writing, in recent times, and particularly in Egypt, there has been some experimentation with writing in the vernacular. This had long existed as a marginal and not very respectable phenomenon. In the early 20th century, we find sometimes lurid accounts of “edgy” aspects of

Orality, Culture, And Language urban life, rather like the “penny dreadfuls” of Victorian England. A whole series with the title “Diary of a…,” mostly running to 60 or 70 pages and purporting to be autobiographical, was published in the 1930s, selling for a few pennies on the streets of Cairo. A good example is Muŧakkirāt Naššāl (Diary of a Pickpocket) by one ‘Abdul-‘Az īz al-Nuṣṣ (probably a pseudonym), which supposedly gives the respectable reader an autobiographical insight into the tricks of the petty thief’s trade and is written in a broad Cairene argot. Other vernacular prose works published around the same time sought to amuse: an example is Maḥmūd Bayram al-Tūnisī’s (q.v.) Is-Sayyid wi Marātu fi Bārīs (The Master and His Wife in Paris), which satirizes, largely through their own conversation, the pretensions of middle-class Egyptians traveling to Europe for the first time. But perhaps the most eye-catching example is a recent (1994) novel written entirely in colloquial Egyptian Arabic by the major Egyptian writer Yūsuf al-Qa’īd, titled “Sparrows’ Milk” (a proverbial expression for anything of great rarity, similar to “hens’ teeth” in English), in which the story is a first-person narration by an illiterate peasant woman speaking directly to the reader. She tells what happens when a member of her family finds a million Egyptian pounds in the street and the trials and tribulations that follow from this at first sight extraordinary piece of good luck. The author has denied that he was motivated to write in Egyptian colloquial by any ideological considerations, describing the work as a literary “experiment” and observing that to make the narrator speak SA would have been absurd, given that she and her family are drawn from the ranks of Egypt’s illiterate poor. Rather different considerations motivated the writing of the vernacular novel Nuzūla wa Khayṭ aš-Šayṭān (Tenants and Cobwebs) (1986) by the Iraqi Jewish author Samīr Naqqāsh. The story focuses on a Baghdadi apartment house of the old style during the late 1940s, in which the tenants are a mixture of Muslims, Christians, and Jews—a setup that would have been quite normal in the years before the foundation of Israel in 1948. All the characters tell their story in their distinctive Baghdadi dialect: Muslim, Christian, or Jewish. By the end of the book, the house is in semiruin, and all the tenants have left, save a lone madwoman. The work seems to be an allegory for the 20th-century fragmentation of civil society in Iraq, symbolized by the author’s extraordinary ability to write dialogue in the three dialects of the communities 19 that constituted it. The complex orality of Iraqi society, the passing of which the book seems to mourn, has been one of the casualties of international politics, it seems. As noted already, some categories of written Arabic are produced in a quasi “real-time” interactional context, where it is difficult to draw a line demarcating them formally from speech. But these are not the only contexts in which a nonstandard form of the language is used in writing. Diaries, private messages, notes, personal letters, the captions to newspaper cartoons, and caricatures are all “speech-like” in one sense or another; this tends to be reflected in the non-SA forms used in them. However, it seems that written texts in which different levels of Arabic alternate are becoming commonplace even in commercially published work. As ever, the epicenter of this new development is Egypt. Rosenbaum (2000) dubs this “alternating style” “fuṣḥāmmiyya,” a portmanteau term formed from fuṣḥā, the Arabic term for CA/SA and ‘āmmiyya, the term for colloquial, non-standard Arabic. This style has characteristics that differentiate it from the mixed, “hybridized” speech style exemplified at the beginning of this chapter: first, the fact that SA and the colloquial alternate at the level of phrases and whole sentences in an apparently random fashion (and this has seemingly nothing to do with rhetorical function) and do not form hybrids at word level; second, the frequency of hendiadys—that is, the repetition of a meaning element, often a lengthy phrase, in both codes, one after the other; third, the commutative nature of the alternations—that is, that readers will readily accept fuṣḥāmmiyya texts in which the original SA and colloquial sections have been “translated” into the other code so that the text they are presented with for comment is like a photographic negative of the original.20 This style has been shown to occur in a variety of written text types: magazine articles on fashion or sport but also with current affairs of the day as they impact on the individual reader as well as in humorous and anecdotal stories.21 As Rosenbaum points out, none of these texts involve the transcription of actual speech, whether face to face or computer mediated, but planned and edited written texts that are the result of conscious choice. That is, the code alternation seems to be meaningful in itself and serve a number of purposes, all of which can be seen as examples, in Gumperz’s terms, of “metaphorical” rather than “situational” code switching: expressing emotional commitment to a point of view; emphasizing a statement; being sarcastic, ironic, or flippant, depending on the context. A recent study of the language of three Egyptian newspapers (Ibrahim 2010) shows a similar pattern of code switching in two of them, even in headlines. The Egyptian Revolution of January 25–February 11, 2011, was particularly interesting from the standpoint of how the protesters conveyed their message to the media of the Arab and the wider world. Though the history of how the revolution was organized has yet to be written, it is clear that the “social media”—Facebook, Twitter, and Internet blogs—were a major tool in the early stages. This fact has already given rise to at least one joke—Egyptians are

Orality, Culture, And Language famed throughout the Arab World for their sense of humor— that is widely circulating on the Internet (in the colloquial, inevitably, as are virtually all jokes):

After he died, Mubarak met Nasser and Sadat. “So,” they said to him, “was it poison or a podium?” “No,” he answered bitterly, “it was Facebook.”22 However, once Mubarak’s regime closed down the Internet and the mobile phone networks, as it did about four days into the uprising, Egyptians started appearing in Tahrir Square carrying homemade placards of all shapes and sizes to convey to the world their verdict on their president and government. Some were in rudimentary English, with messages like “Go to Hell,” “Game Over,” and “Get Out.” But many were written in spoken Egyptian Arabic and often inventive and funny. Some examples follow. A photograph of Obama shaking hands with Mubarak has the following speech bubbles: Obama: In my opinion you should write a letter of farewell to the Egyptian people. Mubarak: Why? Where are they going? A placard written in felt tip reads: “The Union of Egyptian Carpenters asks Master Carpenter Mubarak—what kind of glue do you use?” Another: “The woman wants to give birth, but the baby doesn’t want to see you.” Inevitably also, colloquial poetry came to the fore. Egyptian past master Aḥmad Fu’ād Nigm, now in his 80s, led the charge with anti-Mubarak compositions such as Ka’annak mā fīš (“It’s as if you don’t exist”) and ʿAyzīn nugarrab khil’a tanya (“We’d like to try a different face”). A performance of the first poem on YouTube is given by someone who is described as misrītāli‘‘ēnuh, roughly “an Egyptian fed up to the back teeth.”23 What the Egyptians displayed in their use of such orate weapons was what Nigm himself invented a word for in one of his poems: niktalūgia, roughly “joke-ology”: the ability to beat an enemy by joking and mockery, but always with a light heart (khiffat id-dam).

12.5 Concluding Remarks As this chapter has sought to show, the relationship between orality and language in Arabic is complex. The layman’s mental landscape is of a “high,” literary, codified variety of the language strongly identified with a unifying religion (Islam) and a “golden age” of past imperial and literary glories, carrying great cultural prestige; and a “low,” chaotic (often regarded as grammarless), but homely variety associated with domesticity, intimacy, and the daily round. The emotional resonances of the two varieties are and always have been different, and as a consequence they have, down the ages, occupied separate functional niches in all linguistically mediated communication, be it speech, writing, song, poetry, cinema, or theatre. It is undeniable that the 21st century is bringing about a narrowing of this gap, both formally and functionally. This is partly as a result of new technologies, and some new forms of identity are crystallizing to which novel forms of language use are central. But some conservative currents are also going against this tide, a major one being the deliberate public oral use of CA/SA by religious leaders who thereby seek to use it as a symbol that marks them out and harks back to an imagined past that they (and the many who follow them) would like to see return. It should never be forgotten that the call to prayer is precisely a call or that the Quran, Islam’s holy book, was originally an entirely oral recitation of the prophet Muḥammad. Its daily cantillation, amplified through the streets by a thousand mosque loudspeakers, blaring from radios and television sets in every home and at every street corner, and sung or spoken as the preamble to any public event, is still perhaps the most pervasive public expression of Arab orality and the one that immediately impresses itself on the consciousness of any visitor to the region.

References

Orality, Culture, And Language ʿAbdul-ʿĀl, Ghāda. 2008. ‘Ayza atgawwiz (“I want to get married”). Cairo: Dāral-šurūq. ——. 2011. ‘Ayza atgawwiz (“I want to get married”). Available at http://www.wanna-b-abride.blogspot.com/. Abd-el-Jawad, Hasan. 1986. The emergence of an urban dialect in the Jordanian urban centers. International Journal of the Sociology of Language 61: 53–63. Abdel-Malek, Kamal. 1990. A study of the vernacular poetry of Aḥmad Fu’ād Nigm. Leiden: Brill. Al-Azraqi, Munira. 2010. Pidginisation in the eastern region of Saudi Arabia: Media presentation. In Arabic and the media, ed. Reem Bassiouney, 159–179. Leiden: Brill. Badawi, Mustafa. 1988. Early Arabic drama. Cambridge, UK: Cambridge University Press. Bassiouney, Reem. 2006. Functions of code-switching in Egypt: Evidence from monologues. Leiden: Brill. Al-Batal, Mahmoud. 2002. Identity and tension in Lebanon: The Arabic of local news at LBCI. In Language contact and language conflict in Arabic: Variations on a sociolinguistic theme, ed. Aleya Rouchdy, 91–115. London: RoutledgeCurzon. Bentahila, Abdelâli. 1983. Language attitudes among Arabic–French bilinguals in Morocco. Clevedon: Multilingual Matters. Bizri, Fida. 2005. Le Pidgin Madam, Un nouveau pidgin arabe. La Linguistique 41: 53–66. Blanc, Haim. 1964. Communal dialects in Baghdad. Cambridge, MA: Harvard University Press. Davies, Humphrey. 2008. Yusuf al-Shirbini’s brains confounded by the ode of Abu Shaduf expounded (Kitāb hazz al-quhūf bi-Šarḥ qaṣīd Abī Šādūf), vol. 2. English translation and notes. Louvain: Peeters. Etling, Bruce, John Kelly, Robert Faris, and John Palfrey. 2009. Mapping the Arabic blogosphere: Politics, culture, and dissent. Cambridge, MA: Berkman Center for the Internet & Society, Harvard University. Haeri, Niloofar. 1996. The sociolinguistic market of Cairo. London: KPI. Heath, Jeffrey. 1989. From code-switching to borrowing: a case study of Moroccan Arabic. London: KPI. Hoenerbach, Wilhelm. 1956. Die vulgärarabische Poetik al-Kitāb al-’āṭil al-ḥālī wa l-muraxxaṣ al-ġālī des Ṣafīyaddīn ḥillī. Wiesbaden: Franz Steiner Verlag. Holes, Clive. 1983a. Patterns of communal language variation in Bahrain. Language in Society 12: 433–457. ——. 1983b. Bahraini dialects: Sectarian dialects and the sedentary/nomadic split. Zeitschrift fuer arabische Linguistik 10: 7–37. ——. 1993. The uses of variation: A study of the political speeches of Gamāl ‘Abdul-Nāṣir. In Perspectives on Arabic linguistics V, ed. Mushira Eid and Clive Holes, 13–45. Amsterdam: John Benjamins. Holes, Clive and Said Salma Abu Athera. 2009. Poetry and politics in contemporary Bedouin society. Ithaca, NY: Reading. ——. 2011. Nabati poetry of the United Arab Emirates. Ithaca, NY: Reading. Ibrahim, Zeinab. 2010. Cases of written code-switching in Egyptian opposition newspapers. In Arabic and the media, ed. Reem Bassiouney, 23–45. Leiden: Brill. Idrīs, Yūsuf. 1981. Malik al-Quṭn, Gumhūriyyat Faraḥāt (The cotton king and Faraḥāt’s republic). Cairo: Dār Miṣr li lṭibā‘a. Mazraani, Nathalie. 1997. Aspects of language variation in Arabic political speech-making. London: Curzon.

Orality, Culture, And Language Mejdell, Gunvor. 2006. Mixed styles in spoken Arabic in Egypt: Somewhere between order and chaos. Leiden: Brill. Naqqāš, Samīr. 1986. Nuzūla wakhayṭ aš-Šayṭān (Tenants and cobwebs). Jerusalem: Association for Jewish Academics from Iraq. Næss, Unn Gyda. 2008. Gulf Pidgin Arabic: Individual strategies or structured variety? MA thesis, Department of Culture Studies and Oriental Languages, University of Oslo. Nortier, Jacomine. 1990. Dutch–Moroccan Arabic code-switching. Dordrecht: Foris. Al-Nuṣṣ, ʿAbdul-ʿAzīz. 1930. Muŧakkirāt naššāl (Diary of a pickpocket). Cairo: al-Yūsufiyya Press and Bookshop. Al-Qa’īd, Yūsuf. 1994. Laban al-‘asf ūr (Sparrows’ milk). Cairo: Dār al-Hilāl. Romaine, Suzanne. 1988. Pidgin and Creole languages. London: Longman. Rosenbaum, Gabriel. 2000. Fuṣḥāmmiyya: Alternating style in Egyptian prose. Zeitschrift für arabische Linguistik 38: 68–87. Rosenhouse, Judith. 1984. Bedouin Arabic. Wiesbaden: Harrassowitz. Rosenthal, Franz (trans.). 1958. Ibn Khaldûn: The Muqaddima, An introduction to history. 3 vols. London: Routledge & Kegan Paul. Sadgrove, Philip. 1996. The Egyptian theatre in the nineteenth century, 1799–1882. Ithaca, NY: Reading. Smart, Jack. 1990. Pidginization in Gulf Arabic: A first report. Anthropological Linguistics 32: 83–118. Al-Tūnisī, Maḥmūd Bayram. N.d. (first published 1923). Is-sayyid wi Marātu fi Bārīs (The master and his wife in Paris). Cairo: Dār Miṣr li l-ṭibāʿa. Wikipedia. 2013. Arabic chat alphabet. Available at http://en.wikipedia.org/wiki/Arabic_chat_alphabethttp://www.youtube.com/watch?v=jGJ1nlJyFtM

Notes: (1) Strictly speaking, this distinction is only coincidentally sectarian—the basic difference is originally one of geography and genealogy, between groups descended from Bedouin stock who have tribal links with central Arabia (the so-called ‘Arab), who are all Sunni, and those descended from a sedentary coastal, largely farming population (the so-called Bahḥārna), who are all Shi‘ite. (2 ) Grammatically and morphologically virtually identical with Classical Arabic (CA) but with a massively changed vocabulary that reflects the modern world. (3 ) Egyptian speakers, in particular, show great loyalty to their dialect in dialect contact situations. (4 ) Female speakers, as in other societies, have been shown by several studies to lead the way in switches to “prestige” local varieties compared with men (Abd-el-Jawad 1986: 57–61 for Jordanian urban centers) and to propagate a kind of variability that targets non-CA/SA forms (Haeri 1996: 231–232 for Cairo). (5) For many educated North African speakers of Arabic, French replaces SA as the code with which they mix and to which they switch (see Section 12.1.3 on computer-mediated communication, or CMC). (6 ) An attempt for Cairene can be found in Mejdell (2006). (7 ) http://www.wanna-b-a-bride.blogspot.com. (8 ) By the University of Texas’s Department of Middle Eastern Studies under the title “I want to get married.” 9

Orality, Culture, And Language (9 ) I am indebted to Ivan Panovic for this observation. (10 ) See http://en.wikipedia.org/wiki/Arabic_chat_alphabet. (11) In this orthography, “7” stands for the Arabic pharyngeal voiceless fricative (ḥ), “3” for its voiced counterpart (‘), and “9” for the voiceless uvular stop (q). I am grateful to my former student Mary Montgomery for permission to quote these examples. (12 ) Davies (2008) provides an English translation. (13 ) Abdel-Malek (1990). (14 ) The title seems to be a double entendre: the winner originally got a million UAE dirhams (now it’s more) as a prize, but it can also be understood as “poet in a million.” (15) The greatest early compilation is the Kitāb al-Aġānī (Book of Songs) of the poet and musicologist Abūl-Faraj alIṣfahānī (897–972). (16 ) Though renowned for her love songs, Umm Kulthūm also sang many with a political message. (17 ) However, the Egyptian author Muḥammad ‘Uthm ān Galāl (1829–1898) experimented with translating French playwrights of the 17th century, such as Molière and Racine, into Egyptian colloquial Arabic. (18 ) Recorded Egyptian popular music, as we have noted, was also hugely popular and had a similar effect. (19 ) Naqqāsh left Iraq in his teens but seems to have had an ear for and memory of the Baghdadi vernaculars of his boyhood. The work is heavily footnoted since few Arabs—even Iraqis of the present generations—would now understand these communal dialects, which have now largely disappeared from the public arena (disappeared completely in the Jewish case after the emigration en masse of the Jewish population to Israel in the early 1950s). Blanc (1964) provides a detailed linguistic description of the three dialects, but it is not nearly as rich in the detailing of their individual idioms and vocabulary as this work of fiction. (20 ) Shown by an experiment in which Rosenbaum (2000: 78) sought readers’ reactions to a fuṣāmmiyya text he had “reversed.” (21) Rosenbaum (2000: 74) notes plentiful examples in a book of reminiscences titled “Memoirs of a Young Egyptian Washing Dishes in London.” (22 ) The reference here is to the way the two previous Presidents of Egypt met their deaths. It has long been rumoured, though never proven, that Nasser (president 1954–70) died as a result of drinking poisoned coffee. Sadat (president 1970–81) was assassinated by Islamists as he sat on a podium, watching a march past of Egyptian forces. The “bitterness” of Mubarak is because his presidency ended so relatively ingloriously. (23 ) http://www.youtube.com/watch?v=jGJ1nlJyFtM. Clive Holes Clive Holes, Oxford University

Dialectology Peter Behnstedt and Manfred Woidich The Oxford Handbook of Arabic Linguistics Edited by Jonathan Owens Print Publication Date: Sep 2013 Online Publication Date: Dec 2013

Subject: Linguistics, Sociolinguistics, Languages by Region DOI: 10.1093/oxfordhb/9780199764136.013.0013

Abstract and Keywords Geographically, Arabic is one of the most widespread languages of the world, and Arabic dialects are spoken in an unbroken expanse from western Iran to Mauritania and Morocco and from Oman to northeastern Nigeria. Arabic dialects may have millions of speakers, and in some Arab countries the dialect of a politically or economically prominent city plays the de facto role of a standard language, at least with respect to oral communication. Arabic dialectology is closely connected with a number of other disciplines of Arabic linguistics such as historical linguistics and sociolinguistics, including urban linguistics. This article focuses on what may be called “traditional Arabic dialectology,” which refers to the collection of linguistic features in a given geographic area and the study of these features with regard to their distribution in this area to establish dialectal borders lines, transitional areas, core areas, and dialectal continua. Keywords: Arabic, linguistics, language, dialect, traditional Arabic dialectology

13.1 Introduction GEOGRAPHICALLY, Arabic is one of the most widespread languages of the world, and Arabic dialects are spoken in an

unbroken expanse from western Iran to Mauritania and Morocco and from Oman to northeastern Nigeria, albeit with vast uninhabited or scarcely inhabited areas and deserts in between. It is not easy to give the exact number of speakers; estimates from 1999 count 206 million L1 speakers, a figure that today seems too low rather than too high.1 This geographical range is marked by extreme dialectal differences in all fields of phonology, grammar, and lexicon, at times to the extent that different varieties are mutually unintelligible. Arabic dialects2 may have millions of speakers, and in some Arab countries the dialect of a politically or economically prominent city plays the de facto role of a Standard language, at least with respect to oral communication; Cairo Arabic in Egypt and Casablanca Arabic in Morocco (Aguadé 2008: 288a; Caubet 2008: 273b) are relevant cases. Although they have their principal domain in oral communication, dialects are also used for writing and even in some forms of literature (see Aguadé 2006 for Morocco and Rosenbaum 2004; Woidich 2010 for Egypt; see also [Holes, “Orality”]). Arabic dialectology is closely connected with a number of other disciplines of Arabic linguistics such as historical linguistics [Owens, “History”] and sociolinguistics including urban linguistics [Al-Wer, “Sociolinguistics”]. In fact, it constitutes an indispensable prerequisite as it provides these with the necessary data. This article limits itself to what may be called traditional Arabic dialectology (TAD). The subject matter of traditional dialectology is the collection of linguistic features in a given geographic area and the study of these features with regard to their distribution in this area to establish dialectal borders lines, transitional areas, core areas, and dialectal continua. All this can be best made visible as a linguistic landscape by reproducing these features on

Dialectology maps. For more detail, see Behnstedt and Woidich (2005).

13.2 A Glance at History The interest in the regional varieties of spoken Arabic has a relatively long history. In the course of the 19th century, a considerable number of word lists, smaller or larger dictionaries, and practical guides and textbooks appeared as the result of increased possibilities of tourism and scientific research in the Arabic-speaking world. A few more comprehensive treatises on the colloquial had appeared before this time (Alcalá 1505; Dombay 1800; Caussin de Perceval 1833; Ṭanṭāvy 1848). Wallin (1851, 1852) and Wetzstein (1868) provide samples of folk poetry and stories furnished with phonological and factual annotations, which give valuable insights into the Bedouin Arabic of Syria. A real linguistic interest in Arabic dialects, however, and the creation of a discipline “Arabic dialectology” as part of academic Oriental and Semitic studies did not develop until the final quarter of the 19th century, when the first systematic grammars and elaborate dictionaries appeared, which went far beyond previously published works. These were often accompanied by text collections provided with glossaries (e.g., Spitta 1880), which at the same time were of great value for ethnographic and folkloristic studies. The early days of Arabic dialectological studies thus ran parallel to the time when the great enterprises in dialectology, that is, the national projects of the French and German dialect atlases, started. Beside the professional scholars in Arabic and Semitic studies, a considerable number of valuable data of linguistic and ethnographic interest were collected also by archeologists working in Egypt or Iraq, such as Maspéro (1914) and Weissbach (1908–1930). The first attempts at dialect atlases for the Arab world—today historical documents because of the political developments, creation of new states, and movement of populations—were made as early as in 1915 by Gotthelf Bergsträßer (Palestine) and 1940 by Jean Cantineau (Ħawrān). The first half of the 20th century saw many more publications in the field, and, as early as in 1961, Anton Spitaler could observe that a huge amount of material on Arabic dialects was available, so that it was difficult to maintain an overview over the material [135/226]: “Insgesamt verfügen wir heute absolut genommen über ein gewaltiges, nur mehr schwer übersehbares Tatsachenmaterial, das sich über weite Gebiete des arabischen Sprachraums von Marokko bis Buchara erstreckt” “Altogether we have at our disposal a huge amount of primary material extending over a large area of the Arabicspeaking region from Morocco to Bukara, so that an overview of it is difficult” (Spitaler 1961: 133). This amount of data was despite its patchiness sufficient and detailed enough to enable the first comparative overviews by Hans R. Singer (1958) on interrogatives and by Wolfdietrich Fischer on demonstratives (1959), both students of Hans Wehr. Nevertheless, Arabic dialectology was at this point not really an academic profession. Despite the fact that its importance for Arabic linguistic history can hardly be overestimated, it had remained a field to which serious scholars would devote only their Sunday afternoons, as Spitaler used to tell his students. David Cohen, in his “Préface” to Actes des premières journées de dialectologie arabe de Paris (Caubet and Vanhove 1994), states something similar: “Les arabophones méprisaient leurs dialectes, les arabisants la dialectologie.”, “Arabic speakers despised their dialects, Arabicists dialectology.” Arabic studies at that time, following their historical origin, more or less resembled the study of the classical languages Greek, Latin, and Hebrew. Making an academic career in dialectology only was next to impossible for Arabists, as chairs for Arabic and Semitic studies were designed with a far wider profile including Arabic literature, history of the Near East, and Islamic studies. Several developments in the 1960s and 1970s changed this state of affairs to some extent. First of all, the growing political and economic importance of the Arab countries together with the expanded possibilities for traveling, visiting, and researching these countries led to an increased interest in “real Arabic,” that is, the spoken language of daily life or the modern Arabic dialects. Many students would no longer limit themselves to Classical or Modern Standard Arabic for second-language learning but chose to acquaint themselves with the colloquial as a language of daily life as well. Certainly, as a consequence of this, more academic interest for the field was stimulated. Second, the unimagined advances in the technology of speech recording that had begun in the 1950s was not without impact on Arabic dialectology: recording living speech allowed a far more systematic and scientific approach to fieldwork and was less prone to the predilections and deficiencies of the individual researcher. While

Dialectology in earlier times texts could be recorded only by means of dictation—a rather unnatural form of speech—it was now possible to record speech without too much technical effort and cost. Haim Blanc and Wehr were the first dialectologists of Arabic to use a tape recorder in the field (Jastrow 2002: 350). These recordings, as well as the recordings of entire elicitation sessions, when necessary, could be repeated and checked by the researcher or by others for verification in nearly the same way as they had been recorded. This made it possible to embark on projects covering larger areas more systematically and with better scientific methods than ever before. It is beyond any doubt that these technical improvements led to a far higher reliability of the recorded data. Third, descriptive structural linguistics, in particular taxonomic phonology and morphology, which had already been developed in other philologies in the 1930s and 1940s, were gradually adapted by Arabic dialectologists and applied in their works. In particular, Cantineau’s (1960) adaptation of the Prague phonology in several articles from the early 1950s proved to be seminal. His remark that “la dialectologie arabe n’est guère progressiste; les nouvelles techniques de recherche ont du mal à s’y acclimater,” “Arabic dialectology is not at all progressive; new research methods have difficulty becoming established in it” (ibid., 277) proved incorrect for the years to follow, which saw a fruitful adaptation of modern methods of descriptive linguistics. Phonology, in particular, played an important role here. Good examples of strictly taxonomy-oriented grammars including phonology and morphology are provided in the following: for Egyptian, see Harrell (1957); for Moroccan, see Harrell (1962); for Damascus, see Grotzfeld’s (1964) monographs and Ambros (1977); for Mħallami/Anatolia, see Sasse (1971); for Daragözü, see Jastrow (1973); on Maltese, see Schabert (1976). Although formerly a rather intricate phonetic notation with dozens of diacritic marks prevailed in transcription, projecting only an apparent phonetic accuracy, the analysis of the sounds based on phonological principles now led to a limited set of phonemes that were used for transcribing the data and thus allowed for better writability and readablility. To see the difference, one may compare the transcriptions used, for instance, by French dialectologists working on North African dialects, such as W. Marçais and Ph. Marçais, with more recent ones, for instance, R. S. Harrell, D. Caubet, and J. Aguadé. Fourth, up until the 1950s, research concentrated with few exceptions on easily accessible places and areas, a fact deplored by Cantineau (1955). For Egypt, to give an example, only the great cities of Cairo and Alexandria had attracted major scholarly attention, but not the Nile Delta, let alone Upper Egypt or the Oases in the Western Desert. Cairo Arabic was seen as Egyptian Arabic par excellence. Supported by the technical and methodological improvements previously mentioned, fieldworkers endeavored from the 1960s onward to uncover the treasures hidden in many places that had heretofore remained either hardly accessible or neglected for some (other) reason. Very important regional overviews were published that colored in a number of white spots on the map (Iraq, Anatolia, Egypt, Yemen, Syria, Arabian Peninsula, Oman, Sudan, Sahel) and offered a wealth of data both for the dialect geographer and the historical linguist. Already in the 1960s Arabic dialectologists were aware of the fact the existence of many dialects was being threatened. Spitaler (1961: 136/227) quotes Henri Fleisch (Orbis VIII 1959: 386) on the situation in Lebanon justifying his research in rural Lebanon: “L’évolution des parlers se précipite actuellement, elle pénètre partout et tend à un nivellement des parlers qui leur enlève leur originalité. L’urgence de l’enquête était très grande.”, “The evolution of varieties is currently accelerating. It penetrates everywhere and tends towards a leveling of varieties which effaces their originality. The urgency of research is large.” Replacing older, outdated, and sometimes unreliable sources by new ones that meet modern standards and covering the vast areas that have underresearched or not researched at all still belongs to the urgent desiderata of TAD. As a result of political developments and subsequent migrations, many varieties spoken by minorities, above all the religious ones, are in acute danger. In particular, the Jewish dialects are threatened by extinction. The long established tradition of Jewish Arabic studies (Cohen 1912; Brunot-Malka 1940; Cohen 1964) became of prime importance for the recording of these dialects bound for extinction and was taken up again by Blanc (1974), Cohen (1975), Stillman (1988), Jastrow (1990), Mansour (1991), Heath (2002), Rosenbaum (2003), and Yoda (2005). The global phenomenon that local dialects are disappearing in favor of more regionally expanded varieties of a language, as a matter of course, applies to the Arab world too. On one hand, recording these dwindling specimens of speech has become more urgent than anybody could imagine at Fleisch’s time, while, on the other hand, the new media offered by the Internet (blogs, Facebook, Twitter) facilitate the use of the colloquial, albeit further developed in written and supraregional forms.

Dialectology These developments, together with the ever increasing migrations of the population from rural areas to the cities, and from country to country, offer other possibilities for scientific study that go far beyond traditional dialectology and have in the meantime supplanted the latter to some extent: urban dialectology, all types of contact linguistics including diglossia studies, youth language, Pidgin and Creole studies, and other sociolinguistically oriented matters. No wonder that, starting in the 1980s and 1990s with the emergence of sociolinguistics and variational linguistics, many Arabic language researchers such as Jonathan Owens, Gunvor Mejdell, Nilofaar Haeri, Catherine Miller, and Enam Al-Wer directed their interests to these branches of studies of linguistic varieties and developed them into more or less separate fields of activity ([Davies et al., “Code Switching”; Al-Wer, “Sociolinguistics”]). Despite that, the traditional descriptive approach has not been abandoned but has been falling behind somewhat over the last few years (Jastrow 2002, 2008). For all these reasons, the last 50 years saw an unprecedented expansion and increase in fieldwork and in the amount of data accessible to the researchers. Data collections such as monographs, atlases, grammatical sketches, text collections, textbooks, dictionaries, and other linguistic descriptions keep flowing continuously. The mass of data increased again far beyond what had been imagined only some years earlier. This expansion called for a more formal organization. Since both older and younger researchers such as Wolfgang Fischer, Hans. Singer, Otto Jastrow, Hartmut Bobzin, Peter Behnstedt, and Manfred Woidich felt an urgent need for a regular forum for publication and discussion dedicated to Arabic dialectology and linguistics, Jastrow and Bobzin founded the journal Zeitschrift für Arabische Linguistik (ZAL). Its purpose was to make new findings and recent studies accessible to the public as swiftly as possible. The ZAL, the first issue of which appeared in 1978, developed into an important specialized forum for presentation of research and discussion; more than 50 volumes have been published to date. In 1983, the French journal Matériaux Arabiques et Sudarabiques-GELLAS followed. Jastrow started the series Semitica viva in 1987, which in the years to follow hosted many of the most important publications in the field. In 1993, French dialectologists from the renowned INALCO (above all Caubet and Vanhove) took the initiative and convened all colleagues in the field to Paris at a Colloque International, the first conference dedicated to Arabic dialectology. The conference was concluded with the foundation of the Association Internationale de la dialectologie Arabe (AIDA) and thus marked the beginning of a series of nine highly successful conferences to date. The proceedings of these conferences very aptly mirror the most recent developments in the field, both with respect to traditional approaches as well as to other types of Arabic dialectology.3 To conclude, the particular situation in Spain with its Andalusian background and its focus on the Maghreb led to the foundation of another journal with the programmatic title Estudios de dialectologia norteafricana y andalusí (EDNA) in 1996, which since then has also developed into an important forum for documentation and discussion. Maltese, with its rich indigenous linguistic academic landscape, split off from general TAD and formed its own association, L-Għaqda Internazzjonali tal-Lingwistika Maltija (GĦILM) in 2007, which convenes regular conferences.4

13.3 The State of the Art As Owens (2006: 8) rightly puts it, “The modern dialects have an indispensable role in an account of Arabic language history.”5 We would even say that the modern Arabic dialects, their development, and their relation to Classical Arabic (or Old Arabic, whatever one may call it) are the central object of research for Arabic historical linguistics. This gives TAD fundamental importance for any research in Arabic historical linguistics. TAD, therefore, is heavily and primarily fieldwork oriented, not theory driven. In more detail, it aims at the following: 1. Recording contemporary Arabic speech as much as possible from a number of representative community members, all over the Arab world and beyond, wherever a variety of Arabic is spoken 2. Documenting these data and making them accessible for the researcher in various forms of publication: monographs such as atlases, grammars, (Ortsgrammatiken), bidirectional dictionaries, handbooks, collections of texts (of ethnographic, folkloristic interest), and textbooks 3. Describing the variation and differences between one local dialect and another by providing regional and overall comparative descriptions in monographs and articles 4. Classifying the dialects according to synchronic and diachronic criteria: grouping the dialects, clustering,

Dialectology establishing core areas and transitional areas, detecting dialect continua, and showing the linguistic relations between various regions in the Arab world 5. Collecting the older evidence of Arabic dialects as documented in historical records such as historiography, literary works, historical lexical studies, and travel accounts. TAD overlaps here with the study of Middle Arabic; see, for instance, Davies (1981), Lentin (2008), and Zack (2009).

13.4 Presentation of the Data

13.4.1 Atlases and Maps The main goal of traditional dialectology is to document its findings in atlases (see Chambers-Trudgill 1998).6 Very early, before World War I, Bergsträßer (1915) made the first attempt at this type of documentation for what was then Palestine, and Cantineau (1940) followed somewhat later with the adjacent region of Ħawrān, now southern Syria. Both constitute important documents for historical comparison today. It was not until 1961 that Abul-Fadl followed with geographical research on the distribution of phonological and morphological features in the Egyptian province of Šarqiyya in the Nile Delta. Behnstedt and Woidich’s (1983) map on Arabic dialects in Egypt was the first of its kind and was followed by the five volumes of Egyptian dialect atlas (1985–1999); Behnstedt published his atlas on North Yemen in 1985. Arnold-Behnstedt (1993) covers the Qalamūn Mountains in Syria, with a focus on Arabic–Aramaic contacts. Behnstedt’s (1997) Syria atlas is the most comprehensive one so far in terms of number of maps and data. The first two volumes (of a planned total of four) of the Wortatlas der arabischen Dialekte (WAD) by Behnstedt and Woidich appeared in 2011 and 2012, respectively, with 311 full-color onomasiological maps on “Mensch, Natur, Fauna, Flora” and “Materielle Kultur,” each map accompanied by a commentary. The WAD is the first atlas to cover the entire Arab world and provides a survey of the lexical richness and diversity of the Arabic language and its semantic developments. Some regional monographs contain a rather extensive series of maps in their appendices: Arnold (1998) on the province of Hatay (Antiochia, Turkey); and de Jong (2000, 2011) on Sinai. A collection of maps on the terms for animal and body parts resulting from an unfinished survey on Northern Morocco is published in Behnstedt (2005, 2007). Other maps for illustrative purposes can be found in Procházka (1993) on prepositions, Mörth (1997) on numbers, Jastrow (1978–1981) on q∂ltu-dialects in Mesopotamia and Anatolia, and Johnstone (1967) on the Gulf area. Heath (2002) gives an appendix with a series of rather abstract maps on Muslim and Jewish varieties spoken in Morocco. Six simplified maps in Abboud-Haggar (2011) serve as a quick first overview on the geographic distribution of the Arabic dialects; another can be found in Corriente and Vicente (2008). Quite a few publications use maps for illustration: for example, Cantineau (1940a) on Algeria; Fleisch (1974) on Lebanon; Ingham (1973) on southern Iraq and Khuzistan; and Owens (1985) on the Sahel. Vanhove (2009) contains a dialect map of Yemen, and Taine-Cheikh (1998–1999) published seven maps on the distribution of certain “macrodiscriminants” (*q and the interdentals) in the Arabic-speaking world. Projects on dialect surveys have been announced for Northern Israel (Talmon 2002) and Tunisia (Mejri 2002; Sandly 2002)7 but have not materialized so far. Behnstedt and Woidich (2005) give a general introduction to Arabic dialect geography and incorporates a chapter on dialectometry.

13.4.2 Regional Studies The second edition of the Encyclopedia of Islam (1986) offers concise information in the article “ʕArabīya” on “Arabian and North Arabian dialects” (Fleisch) and on the “Western Dialects” (Colin). Març ais (1977) provides a comprehensive survey on the phonological and morphological features of the Western Arabic dialects. To date, there is no equivalent study for the Eastern part of the Arab world. Blanc (1971) gives a concise report on subSaharan Arabic. Fleisch (1974) sketches a number of villages all over Lebanon together with text samples and

Dialectology arrives at a preliminary description of dialect areas. Kaye (1976) deals with Chadian and Sudanese Arabic in light of comparative Arabic dialectology. Jastrow’s (1978) work on the q∂ltu-dialects in northern Iraq and Anatolia had a great impact on the further development of the field. Prochazka (1988) gives the first systematic phonological and morphological survey on the Arabic dialects spoken on Saudi Arabian territory. Arnold (1998) covers the Arabicspeaking regions of Antiochia in Turkey, and Owens (1985, 1993b) reports on Chad and Nigeria. Ingham (1982, 1997) deals with the dialects of northeast Arabia. Palva (1984) classifies the Palestinian and Transjordanian dialects, and Woidich (1996) gives a concise account on the distribution of 41 features in the rural dialects of Egypt. The Arabian Peninsula and Iraq are very competently dealt with in Holes (2006); Al-Wer (2006) tackles the Arabicspeaking Middle East, and Walters (2006) addresses north Africa. Among these regional studies, several should be further listed: Blanc (1953) on North Palestinian Arabic; Blanc (1964), the seminal study on Bagdad and Iraq; Johnstone (1967) on the Gulf area; Grand’Henry (1972, 1976) on Algeria; Diem (1973) and Behnstedt (1985) on Yemen; Ingham (1976) on southern Iraq and Khuzistan; Behnstedt and Woidich (1982) on the Egyptian oases; Doss (1981) on Middle Egypt; Ingham (1982, 1997) on Bedouin Arabic in the Najd and elsewhere on the Arabian Peninsula; Holes (1983) on Baħrayn; Owens (1984) on Libya; Rosenhouse (1984a, 1984b) on the Bedouin in northern Israel; Holes (1989) on Oman; Arnold and Behnstedt (1993) on Qalamūn in Syria; Arnold (1998) on Antiochia; Behnstedt (1998–1999) on Djerba; Heath (2002) on Morocco; Procházka (2002) on çukurova; Heath (2004) on Ħassāniyya; Henkin (2010) and Shawarbah (2011) on Negev; and de Jong (2000, 2011) on Sinai. Regarding the Arabic “Sprachinseln” of Uzbekistan, Vinnikow (1962), Tsereteli (1954, 1956), Fischer (1961), Axvlediani (1985), Chikovani (2008, 2009), and Zimmermann (2009) give us valuable information. On Cyprus we have Borg (2004) and Roth-Laly (2006) and, on Afghanistan, Kieffer (1981, 1985, 2000) as well as Ingham (1994a, 2002). On the newly discovered Arabic dialects in Iran, see Seeger (2002). For the situation in border areas where Arabic is a minority language such as Anatolia, Afghanistan, Eritrea/Djibouti, and Central Asia, one may consult the publications by Owens (2000), Simeone-Senelle (2002), and Csató et al. (2005). Heine’s (1982) book on Ki-Nubi introduced Creole studies to the field of Arabic dialectology, followed by Wellens (2005) and Luffin (2005) [Tosco and Manfredi, “Creoles”]. Corriente, in numerous publications, provided for the systematic analysis and description of the extinct dialect of Andalusia; see Corriente (1977, 1997), and for a short account on its evolution see Vicente (2011).

13.4.3 General and Comparative Studies The first state-of-the-art-reports appeared in the 1950s from the pens of Brockelmann (1954) and Cantineau (1955) and also give a good overview of the available literature. A rather complete bibliography can be found in Fischer (1959), later followed by those in Sobelman (1962), Bakalla (1983), and Eisele (1987). The work of Caussin de Perceval (1833), which is today only of historical interest, can be seen as a first attempt to give a comparative overview of the grammar of Arabic dialects for pedagogical purposes. Nöldeke (1904) analyzes the relationship of the dialects with Classical Arabic and arrives at the conclusion that the former developed from the latter, a view that for good reasons has been abandoned today. Bergsträßer (1928) gives a short historical sketch and some texts within the framework of a handbook of Semitic languages. In his opinion, the dialects did not develop directly from Classical Arabic: “Die neuarabischen Dialekte gehen im großen ganzen auf eine einheitliche Grundform zurück, die im allgemeinen der klassischen Sprache nahestand, in Einzelheiten von ihr abwich” (ibid., 156). In the years to come, and in fact until this day, this has been the scenario many historical linguists accepted, in particular those with a German background, even if Nöldeke’s view has continued to be taken as a starting point for historical discussion.8 Fischer-Jastrow (1980), which incorporates regional sketches and text samples, followed as a comprehensive general account of the situation at the end of the 1970s, but in view of the rapid increase of available data and new insights over the past 30 years it needs to be updated. More recently collected data, but without fundamentally new insights, are offered in Kaye and Rosenhouse (1997), which treats Maltese independently from other Arabic dialects; in Durand (2009); and in Abboud-Haggar (2009). Corriente and Vicente (2008) offer a comprehensive discussion of genesis and classification of Vicente’s modern dialects. Other shorter overview articles, some directed at a more general public, are Jastrow (2002, 2008), Versteegh (2011a, 2011b), and in particular Watson (2011), with a critical discussion of some of the features used to discriminate between Old Arabic and modern dialects. Naturally, general introductions to Arabic linguistics and

Dialectology language history such as Schippers and Versteeegh (1987), Versteegh (1997), Ferrando (2001), and Holes (2004) deal with Arabic dialectology and use its data in the relevant chapters. Other comparative studies focus on particular grammatical topics, describe these, and sketch their historical development. The first ones—Singer (1958) on interrogatives and Fischer (1959) on demonstratives—we owe to the school of Hans Wehr. Blanc (1970) deals with dual and pseudo-dual; Janssens (1972) studies stress and word structure; Czapkiewicz (1975) examines the morphology of the verb; and Diem (1979) looks at the substrate question. Eksell Harning (1980) on the genitive exponent and Retsö (1983) on the passive voice discuss their topics both with respect to morphology and to syntax. Mörth (1992) deals with cardinal numbers from 1 to 10, Procházka (1993) with the prepositions, and Dahlgren (1998) with word order. Isaksson (1998) offers a comparative survey of the pronouns. Brustad (2000) analyzes syntactic structures comparing four dialects under modern criteria independent from traditional Arabic syntax. Watson (2002) compares phonology and morphology of two rather different types of Arabic: Cairene and Ṣanʕāni. Diem (2002) describes the syntax of translocative verbs of some dialects and their historical changes, and Taine-Cheikh (2004, 2009) looks at the expression of future. Versteegh (2004) deals with the interrogatives again, and Procházka (2004) wrestles with unmarked feminine nouns. Vanhove et al. (2009) give an account with grammaticalization of modal auxiliaries in Maltese and Arabic, placing them in a larger theoretical framework with European languages. Aguadé (2011) presents in a concise survey the vowels systems of Moroccan dialects. A different type of resource is represented in the Encyclopedia of Arabic Language and Linguistics (EALL), which could serve as a TAD handbook. It contains 37 grammatical sketches of Arabic dialects and 26 linguistic profiles of Arab countries easily accessible to the scholar, in addition to a number of articles on general issues concerning Arabic dialectology, such as “Creole” (Owens), “Gypsy” (Matras), “Dialect Geography” (Behnstedt), and “Dialects: Classification” (Palva).

13.4.4 Grammatical Descriptions of Individual Varieties Today, we have at our disposal several dozens of systematic descriptions of local dialects in the form of monographs and, above all, hundreds of sketches describing the most important phonological and morphological features of the dialects spoken in various places of the Arab world ranging from major cities to the most remote areas, both for rural and Bedouin Arabic. Astonishingly enough, there is no comprehensive grammar in the form of a monograph on one of the major Bedouin dialects of Saudi Arabia, Ingham (1994b) on Najdi Arabic being the laudable exception. Due to lack of space, only some major monographs published in the recent years will be mentioned here: 9 Palva (1976) on al-Balqā’/Jordan; Cowell (1964), Grotzfeld (1964), and Ambros (1977) on Damascus; Abu-Haidar (1979) on Baskinta/Lebanon; Reichmuth (1983) on the Shukriyya tribe in Sudan; Owens (1984) on Benghazi; Singer (1984) on Tunis; Owens (1993) on Nigeria; Seeger (2009) on Ramallah,; Julien de Pommerol (1999) on Chad; Talay (1999) on the Khawētna/Syria; Werbeck (2001) on Manāxa/Yemen; Wittrich (2003) on Āzəx/Anatolia; Borg and Azzopardi (2005) for Maltese; Woidich (2006) on Cairo; Gralla (2006) on Nabk/Syria; Pereira (2010) on Tripoli; Manfredi (2010) on the Baggāra in Kordofan. The century-old interest of the Maltese in their spoken language produced very early grammatical descriptions and dictionaries, making it to one of the best described and researched varieties of Arabic; see Borg and Azzopardi (2005). In addition to that, there are quite a few PhD and MA theses on Arabic dialects written by native speakers at European and American universities, but many of these are difficult to come by. The degree of comprehensiveness of these grammars differs, as their interest generally is limited to the basic facts of phonology and morphology, while syntax is treated rather marginally. Some exceptions to this point deal with many important syntactic issues: Spitta (1880) on Cairo; Harrell (1962) on Rabat; Cowell (1964) on Syrian; Jullien de Pommerol (1992) on Chad; Caubet (1993) on Fes; Woidich (2006) on Cairo; Naïm (2009) on Ṣanʕā’. A classical structuralistic study of the phonology of Egyptian Arabic is Harrell (1957); a thorough study with a high level of abstractness is Dickins (2007) on the phonology of Sudanese (Khartoum) Arabic. Studies on the syntax of individual dialects include Feghali (1928) on Lebanese, Abboud (1964) on Ħiǧāzi, Bloch (1965) on Damascene, Piamenta (1966) on Palestinian, Denz (1971) on Kwayriš/Iraq, Sieny (1978) on Urban Ħiǧāzi, Watson (1993) on Ṣanɥāni, Vanhove on Maltese (1994), and Eisele (1999) on Cairene. Most important in this respect is Brustad (2000), as it is the only monograph to treat syntactic issues from a comparative perspective

Dialectology and in the light of a fresh modern syntactic approach while not relying on traditional Arabic grammar or a particular modern linguistic school. The latter is an important issue, since in recent years, quite a few articles and monographs on Arabic syntax appeared, but, unfortunately enough, many of these use a particular linguistic framework and are aimed more at serving the further development of syntactic theories than at adding to the knowledge of Arabic.10 Owens and Elgibali (2010) provide a series of articles on structural and pragmatic sources and offer an introduction to information structure as used in spoken Arabic. Of particular linguistic interest are “Sprachinseln” in Uzbekistan, Afghanistan, Turkey, and Iran, which on one hand preserve many old features due to an early split from mainstream Arabic, thus shedding light on earlier linguistic situations in Mesopotamia (Jastrow 2011), and on the other hand show peculiar developments due to their isolation from the core area of Arabic and their contact with other languages. The Creolized versions spoken in Africa (Ki-Nubi/, Juba Arabic/Sudan) are a particularly interesting case for general Creole studies, as they are not based on one of the European languages; see Miller (1983), Prokosch (1986), and Owens (2006) in the Encyclopedia of Arabic Language and Linguistics and also Tosco and Manfredi [“Creoles”].

13.4.5 Dictionaries Arabic dialectologists have a number of dictionaries at their disposal, although many of them are rudimentary and are better classified under the categories “vocabulary” or “glossary” than dictionary. Some are outdated today but have nevertheless been republished without any adaptations (Belkassem 2001), the Georgetown series (see following), and Spiro (1980). The 12 volumes of de Premare (1993–1999) provide a rather comprehensive dictionary for Moroccan. For Maltese, the dictionary by Aquilina in six volumes (2000) (a shorter version is Aquilina 1987–1990) forms an indispensable source for Maltese studies, providing even etymological and comparative information. Others are Landberg (1909) for South Arabia, Spiro (1925) and Hinds and Badawi (1986) for Egypt, Barthélémy (1936) for Syria and the Levant, Taine-Cheikh (1988–1998, 1990) for Ħassāniyya/Mauretania, Piamenta (1990) and Behnstedt (1992–2007) for Yemen, Qafīšeh (1997) for the Gulf area, Jullien de Pommerol (1999) for Chad, Holes (2001) for Baħrayn, Qāsim (2002) for Sudan, Elihay (2004) for Palestinian, Kurpershoek (2005) for Saudi Arabia/Dawāsir, Beaussier et al. (2006) for Algeria, and Chaker and Milelli (2010) for Lebanese. The older dictionaries should be seen as valuable historical documents rather than as reflections of modern language. Still useful, though somewhat outdated today, are the dictionaries of the Georgetown series: Stowasser and Ani (1964) for Syria; Woodhead and Beene (1967) for Iraq; and Harrell (1966) for Morocco. Lentin and Salamé carried out a major project, a comprehensive documentation of Syrian Arabic vocabulary, the [B] having recently appeared as a first letter.11 More in the category “word list,” see Vocke and Waldner (1982) on Anatolia and Jastrow (2005) on Kinderib/ Anatolia. Many text editions, in particular those on Maghrebi dialects, were followed by glossaries with very useful comparative and etymological annotations. Particularly noteworthy examples are, for instance, the monumental Takroûna/Tunisia glossary by Marçais and Guîga (1958–1961) with nearly 4000 pages, Marçais (1911) on Tanger, Brunot (1952) on Rabat/Morocco, and Boris (1958) on Marāzīg/Tunis, which all contain a host of comparative lexical notes as well. For the southern part of the Arabian Peninsula, Landberg’s monumental documentations on Ħaḍramawt (1901) and Daṯīna (1905–1913) are still indispensable, more than 100 years aft er their publication. There are only few dictionaries or glossaries with Arabic as target language. Among these are the bidirectional ones: Sobelman and Harrell (1963) for Moroccan; Clarity et al. (1964/2003) for Iraqi; and Stowasser and Ani (1964) for Syrian. As part of the Georgetown series, these have the advantage of providing rather systematically example sentences illustrating the use of the item listed. Others, mainly intended for practical usage, are Stevens and Salib (2004) and Jomier (1976) for Cairo, Bauer (1957) and Elihai (1985) for Palestinian Arabic, Hillelson (1925) for Sudan, Cohen (1963) and Taine-Cheikh (1990) for Ħassāniyya, and Aguadé and Benyahya (2005) for Morocco. For Maltese see Moser (2005), a rather comprehensive Maltese–German and German– Maltese dictionary.

13.4.6 Lexical Studies The study of etymology and the foreign vocabulary in the dialects has always been a favorite topic for

Dialectology dialectologists and Arab philologists alike. From earlier times we may adduce Vollers’ (1896, 1897) still very useful studies on Egyptian, Almkvist (1891, 1925) on Levantine, and to a certain extent Landberg (1901, 1905–1913), all of which contain rich comparative annotations to other dialects that were already described at that time. Borg (2004) offers an important analysis of the lexicon of the Arabic of Kormakiti/Cyprus with copious references to other Arabic dialects. See further Prokosch (1983) for Turkish loans in Egyptian and Reinkowski (1998) in Baghdad. Kotb (2002) gives a detailed account on Egyptian somatisms, that is, idioms formed with the names of body parts. Special mention should be made of quite a few works of this kind authored by Arab scholars, such as Taymūr (1978–2001) on Egyptian, Frayħa (1947) and Abu Saʕd (1987) on Lebanese, and ʕAbd ar-Raħīm (2003) on Syrian Arabic. Since the existence of the Internet, one may also try one’s dialectological luck in this medium. Many sites, from Saudi Arabia, Ħaḍramawt, Libya, Jordan, and elsewhere, proudly announce lexical peculiarities of remote regions. The researcher is faced with some problems here: (1) the use of the Arabic script, normally not or insufficiently voweled, makes it difficult to discern the correct pronunciation; (2) uncertain origin, copying original sources without references is not uncommon; (3) as these notes are directed from insiders to insiders, the semantic content is often insufficiently described by just giving a MSA equivalent. Checking the validity of this kind of information is thus difficult, and it should thus be handled with prudence.

13.4.7 Text Collections It is a well-established tradition of Arabic dialectologists to document their research not only by means of grammars and dictionaries but also by samples of transcribed texts. Text collecting plays a prominent role in their activities from the very beginning, to the extent that Cantineau (1960: 277) wrote in the conclusion to his article: “on est frappé par la disproportion des résultats: trop de textes, pas assez de grammaires et de dictionnaires.” “One is struck by the disproportionality of results: too many texts, not enough grammars and dictionaries.” His remark is quite to the point, and even more so after speech recording became possible from the 1960s onward. Most grammars, regional overviews, and atlases are furnished with texts or accompanied by text volumes; see, for instance, Peter Behnstedt’s (1997–2000) works on Syria and Otto Jastrow’s (1978– 1981) on q∂ltu-dialects. They not only serve a documentary purpose but also are meant to serve studying the dialect and practicing it. Initially, the interest focused on folklore and popular culture (fairy tales, folk poetry) and paremiology (proverbs), maybe due to the fact that recording speech was not possible yet. The texts had to be noted down by dictation, which means that they had to be present in the memory of the informants so that they could be repeated if necessary. And as story telling played an important role in rural life of that time, this was a relatively simple way to get people to speak. To avoid the highly formulaic language associated with fairy tales and folk poetry, this focus turned to ethnographical issues and oral history in the course of time. Today, these collections offer a wealth of information on urban and rural life in earlier times and are therefore of prime importance not only for the dialectologist but also for the ethnographer and folklorist. For the dialectologist, the problem with these records is that speakers tend to use here an acrolectal type of speech that remains rather descriptive and often features stereotype phrases and does not reflect everyday unmonitored speech, or as Cantineau (1960: 277) puts it: “dans le texte, rarement spontané, il y a autant de celui qui le recueille que de celui qui le dicte.” Therefore, despite the fact that they offer much useful lexical information, texts of this sort are of less value with respect to syntax, phraseology, and pragmatics. Well aware of these deficiencies, dialectologists today prefer to record more personal accounts, life stories, jokes, and similar texts or in any case a good mix of various types of text to provide more space to the speaker for elaborate syntactic constructions, everyday phraseology, rhetorical devices, as is described exemplarily in the introduction to Holes (2005: xviii–xxi). Of course, the observer’s paradox “to monitor unmonitored speech” can never be avoided totally. For linguistic purposes, recorded texts should never be “edited” or “improved” by the transcriber but should be presented with all the deficiencies of natural speech, for example, the Algerian texts given in Bergman (2006), which consciously display the characteristic interferences from French. Large collections of texts are available, both from earlier and modern times: see, for example, Marçais (1911) in Tanger; Rhodokanakis (1911) in Ḏofār; Schmidt and Kahle (1918–1930) for Bīr Zēt in Palestine; Marçais and Guîga (1925) Takroûna in Tunisia; Hillelson (1935) in Sudan; Destaing (1937) Šluħ’s of the Sous (Morocco); Jastrow (1978–1981) in Anatolia and north Iraq; Behnstedt and Woidich (1985–1999) in Egypt; Stewart (1988–1990) in

Dialectology Sinai; Mansour (1991) in Jewish Baghdadi; Palva (1991) in al-Balqā’/Jordan; Behnstedt (1997–2000) in Syria; Jastrow (2003) Kinderib in Anatolia; Luffin (2004) Kinubi in Mombasa; and Bettini (2006) in Syrian Ǧazīra. Text publications as articles in journals and Festschriften run into the hundreds, and for many localities our data come from publications of this kind. Arab scholars, too, are interested in folkloristic issues and published quite a number of collections of folk songs and suchlike. Paremiology is represented in copious collections of proverbs such as the classical ones by Frayha (1938) on Lebanon, Westermarck (1930) on Morocco, Burckhardt (1830) and Taymūr (1970) on Egypt, Goitein (1970) on Yemen, Mahgoub (1968) on Egypt with a linguistic analysis, Nataf and Graille (2002) on Libya, and El Attar (1992) and Lemghūrī (2008) on Morocco. Again, following classical tradition, collecting colloquial proverbs was an activity that attracted Arab scholars as well. In addition, numerous amateur collections are available on the market. Unfortunately, in many of these collections the proverbs are not given in transliteration but in Arabic script, a fact that together with their syntactical and lexical peculiarities makes them unsuitable for many linguistic purposes. As for audio texts, the “Semitisches Tonarchiv” [SemArch] at the University of Heidelberg12 offers a number of recordings together with their transcriptions, which unfortunately cover only a small part of the Arab world. Despite its great success and undisputed benefits, the SemArch seems to have stopped its activities a number of years ago for financial reasons. It is hoped that the new Project “EALL on-line” (Leiden: Brill) will be able to undertake similar activities in nearby future.

13.4.8 Textbooks Arabic dialectologists have always displayed a pedagogical interest, which is evident from quite a number of text books and language handbooks designed for tourists, business people, administrative or military staff, and the like. In fact, writing colloquial grammars started this way with Dombay (1800), Savary (1813), and Caussin de Perceval (1833). Both well-known scholars such as Vollers, Nallino, Ferguson, Mitchell, and interested laymen published to serve this purpose. This tradition has continued, and in recent years Otto Jastrow’s series “Semitica Viva” opened a specialized branch titled “Series Didactica” with Watson (1996) as a first textbook for Ṣanʕāni Arabic. Despite the fact that didactical publications are not recognized by the academic administrations as real “scientific” work, quite a few contemporary scholars developed activities in this field (cf. Holes 1984; Woidich 2000; Bergman 2002). The reason for this pedagogical interest lies in the particular linguistic situation of the Arab world, which creates a specific need for this kind of resources. It was, and still is, quite useful for a non-Arab traveler or resident to learn the local dialect or “real Arabic” of a country he wants to visit or stay in, a language that would be useful in daily life, more so than Standard Arabic, which Arabs themselves have to learn at school and which can be handled properly only by a limited number of well-educated, highly motivated, and trained persons. By speaking a dialect of one of the major cities of the Levant (Beirut, Jerusalem, Damascus), for instance, one can make oneself understood in the whole region: Iraqi (Baghdad) will be helpful in Gulf area and Saudi Arabia; due to the omnipresence of Egyptians, Egyptian Arabic (Cairo) is well understood all over the Arab world, certainly in the East but also to some extent in the West. Since serious dialectological studies begin at the university, textbooks on an appropriate level should be written; examples are Ambros (1998) for Maltese, Woidich and Heinen-Nasr (2006) for Cairene, and Watson (1996) for Ṣanʕāni.

13.4.9 Historical Evidence For the now long extinct dialects of Andalusia we can rely on the comprehensive works of Corriente (1977, 1997); on Sicily, see Agius (1996) and Lentin (2007). In general, historical evidence for a deliberate use of the colloquial in writing, which would allow more insight into earlier stages of the dialects, is scarce. Nevertheless, there are some texts, mostly poetry; see, for instance, Kallas (2007) for Aleppo. Davies (1981, 2005) and Zack (2008) publish and linguistically analyze two important texts on Egyptian Arabic of the 17th century: aš-Širbīnī and al-Maghribī, respectively [Holes, “Orality”]. Vrolijk (1998) on Ibn Sūdūn goes back even further to the middle of the 15th century, while Drozdík (1972) and Woidich (1995) deal with sources from the middle of the 19th century. Significant features of modern Egyptian Arabic are present in the 15th and 17th century; see Woidich and Zack (2008). On Levantine Arabic in the 17th century see Zwartjes and Woidich (2011).

Dialectology 13.5 Interpretation of the Data Collecting, describing, and editing the data of individual dialects for the use of the historical linguist is one side of the coin; the other one is to bring order into the apparent diversity and to cluster the dialects into groups and determine their interrelations. This has to be done by forming either a linguistic hierarchy based on linguistic variables or other ones based on extralinguistic facts.

13.5.1 Linguistic Classification and Dialect Geography As to the linguistic classification, we have to realize here that no generally accepted linguistic variables are available to serve for a linguistic classification of the Arabic dialects as a whole and to say something meaningful when projected onto a map showing their distribution.13 For geographic reasons with deserts and large uninhabited areas and, above all, for a history of continuous movement and settlement of the populations and their mutual influence by contacts, single macrodiscriminants such as the reflexes of *q or the existence of interdentals present themselves in a rather scattered way when projected on the dialectological map of the Arab world, as a glance at the maps in Taine-Cheikh (1998–1999) shows. Nevertheless, using linguistic variables in this way as discriminants is possible, it seems, on a lower level for smaller regions, for instance, when we talk about bukṛa- versus bukṛadialects14 in Middle Egypt (Behnstedt 1979) or about k- versus t-dialects15 in Jemen (Behnstedt 1985: 226, Map 169).16 The macrodiscriminants q/ʔ and g/ǧ in the Egyptian context of the Nile Delta give a rather clear figure on the map, the Cairo-Damiette corridor; see Behnstedt and Woidich (2005, Maps 69 and 70). Linguistic variables thus are adopted in traditional dialect geography by projecting a number of selected features on maps by means of symbols and the subsequent drawing of isoglosses to delineate the areas where these features prevail. When all isoglosses are compiled on one map, these—ideally—form bundles, which are then interpreted as borderlines between different dialect areas, each of them with a core area and a more or less extended transitional area in between depending on the nature of the bundles; see Behnstedt and Woidich (1985, Maps 554–559) for Egypt. The problem of whether some isoglosses or variables should be given more weight than others in this procedure is still unsolved in theory, and the researchers follow their own intuitions. Similar methods have been applied in the regional atlases mentioned already, as aptly described in Palva’s (2006) introduction. The features represented in the bundles of isoglosses can be used for the classification of the dialects and be combined with appropriate extralinguistic features; see Woidich (1996) for Egypt. The “step-method” may be seen as a further methodological development in the classification of dialects that introduces a statistical approach. Neighboring dialects are compared with a set number of variables, all of them considered of the same weight. The number of differences between the two dialects compared is then expressed as a percentage of the total number of the variables. If the percentage is (relatively) low, the dialects belong to the same typological group; conversely, if it is (relatively) high, they belong to different groups. The method was successfully applied to illustrate the continuum of dialects spoken on the northern Sinai littoral, that is, from a Northwestern Bedouin Arabic dialect type to a largely sedentary (rural) dialect type spoken in the eastern Šarqiyya province of the Nile Delta in Egypt (or vice versa); see de Jong (2000, 2011). Another recent and promising way to establish and visualize the subgroups and their interrelations (distance, closeness) is provided by dialectometry, which compares all the data recorded for an area, not only an arbitrary choice according to the predilections of the researcher, by means of more refined statistical methods. As a purely quantitative approach, dialectometry gives all variables the same weight and importance, which makes it more independent on the intuitions of the researcher. It has not yet been applied to Arabic dialects on a larger scale except for a small region, the oases in the Western desert of Egyptian; see Behnstedt and Woidich (2005, Chapter 11).17 This first attempt corroborates the findings of the isogloss method, showing that the dialects of the two oases at the extremes (Baħariyya, Kharga) share more variables with standard Cairene than the two others (Farafra, Dakhla) situated farther away from the Nile valley in terms of traveling distance. For a more detailed discussion of these approaches, see Behnstedt and Woidich (2005, 2006: 586).

13.5.2 Traditional Classifications There is no traditional classification of the Arabic dialects based only on linguistic features, as all rely heavily on

Dialectology extralinguistic (i.e., geographical, social, sectarian, and historical) facts, which are then related to certain linguistic features.

13.5.2.1 Geography All surveys and state-of-the-art reports in their classifications of present-day Arabic dialects rely on geography. This means that the dialects are listed and described mainly within the framework of larger geographical entities such as the Levant, Mesopotamia, and Gulf area, Arab Peninsula with Yemen and Oman, Egypt with Sudan and subSahara, North Africa, and Mauretania (e.g., Fischer and Jastrow 1980; Taine-Cheikh 1998; Corriente and Vicentes 2008; Watson 2011; Versteegh 2011), to which the “islands” Uzbekistan, Afghanistan, Iran, Anatolia, and Cyprus are added. Maltese, as a fully “ausgebaute” written language, is sometimes considered an independent language (Kaye and Rosenhouse 1997). More or less uncontroversial is the subdivision in Western/North African and Eastern dialects with the border running between Egypt and Libya (but see Owens 2003). The linguistic variable most commonly adduced with respect to this dichotomy concerns the paradigmatic leveling that occurred in the imperfect: eastern a-ktib/n-iktib versus western n-iktib/n-iktib-u “I write”—“we write.” But this single feature is far from being conclusive if we want to assign a given dialect to one of these two groups, since there are dialects in Egypt at the Western part of the Delta and in Upper Egypt that have this Western conjugation but that in other respects (stress, syllable structure) clearly belong to the Egyptian, that is, Eastern phylum.18 The Sudanic area, too, is problematic here since the Western type is the norm in Chad, but not in Nigeria or in most of the Sudan. The question as to whether the oases of the Egyptian Western desert should be seen as Western or Eastern (Egyptian) Arabic was answered differently in Woidich (1993a) and Behnstedt (1998), the first being in favor of Egypt and the second in favor of the Maghreb.19

13.5.2.2 History As a variant of the geographic approach but involving a historical fact as well (i.e., the Arabic expansion starting in the 7th century), the subdivision of the Arab linguistic world can be seen in three zones, as introduced by Jastrow (2002: 348). The Arab Peninsula, from where the Arabic expansion started, is considered Zone I, all the territories (Levant, Irak, Egypt, North Africa, parts of Iran) Arabicized due to this expansion are Zone II, and the remaining “Sprachinseln” (Anatolia, Iran, Afghanistan, Uzbekistan, Cyprus, Malta, sub-Saharan Africa) surrounded today by other languages as Zone III. Watson (2011) adopts this view but excludes the southern regions of the peninsula from Zone I, once the stronghold of South Arabic tongues. It is Zone I where we find, according to Jastrow (ibid.), the most archaic dialects today. The question remains: what does “archaic” exactly mean, and how is it defined? See the discussion of allegedly archaic features in Edzard (1998: 142) and Retsö (2003: 116). Zone II “could be called colonial Arabic” with dialects characterized by their innovative features. This is against all experience with other languages forming colonial areas such as English and Spanish. The dialects spoken in the former colonies North America and, respectively, South America are quite homogeneous compared with the respective homelands, a generally recognized fact in dialect geography. Why then the apparent diversity in Arabic “colonial” regions? Can it be attributed only to population movement and contact (Watson 2011), or did it exist before the expansion (Retsö 2000; Owens 2006; [Retsö, “Arabic?”])? The criterion “history” no longer works for Zone III, the language islands; the reason for their being lumped together in a group is the fact that they are surrounded by other languages and separated from zones II and III. In a similar vein, Owens (2006) discusses what he calls pre-diaspora Arabic, that is, Arabic before the conquests “at a time and a place when the ancestral populations were still together” (Owens 2006: 3; [Owens, “History”]), which in one interpretation would correspond to Zone I of Jastrow’s approach. As a “convenient fiction,” he takes Sībawayhi’s approximate date of death (790 CE) as the endpoint of “pre-diaspora” Arabic, thus taking roughly the first 150 years of “diaspora” as “pre-diaspora.” Relating linguistic features reconstructed by means of comparative methods from the modern dialects to this pre-diaspora Arabic, which ended around 790, seems as arbitrary to us as attributing features to the neo-Arabic language type or not. Neither of these two approaches is convincing for the linguistic subgrouping, because they cannot be related to linguistic variables that would justify them.

Dialectology 13.5.2.3 Social/Lifestyle: Sedentary versus Nomadic, Urban versus Rural The division of Arabic dialects into sedentary ones and Bedouin was referred to in Marçais (1938) to explain the Arabization of North Africa. Because of their nomadic lifestyle, Bedouin are considered a group distinct from sedentary people, and since many of their dialects show certain features such as a /g/-reflex of *q, interdental consonants, and feminine forms in the plural, these are considered Bedouin features. This may have been true in the past—the distinction between sedentary and Bedouin speakers with a voiced /g/-reflex of *q is already described by Ibn Sīna in the 11th and Ibn Xaldūn in the 14th centuries (Blanc 1964: 29)—but it would be erroneous to reverse the argument and consider all /g/-speakers Bedouin, let alone nomads.20 Nor can these “Bedouin” features be found with all Bedouin, as in the Sahel region, for instance, interdental consonants have been replaced by dentals (Blanc 1971; Taine-Cheikh 1998; Rosenhouse 2011). So “Bedouin” today is more of a convenient label for a bundle of features and tells us nothing about the present-day lifestyle of the speakers. The same is true for the distinction urban versus rural, which often is related to certain variables, the unvoiced pronunciation of the reflexes of *q, for instance, is considered urban in certain regions (Levant, Morocco). Nevertheless, many villagers today speak urban dialects with a glottal stop, due to the spread of urban speech around the urban centers (Cairo, Damascus) and along trade routes, like in North Africa (Marçais 1938; Singer 1994) between the larger cities and the harbors associated with them. On the other hand, there are numerous cities, the dialects of which show socalled Bedouin features such as the voiced pronunciation of reflexes of *q (Baghdad, Tripoli/Libya, Khartoum, Mecca). The distinction thus says nothing about the whole Arab world and is applicable for smaller regions, Syria for instance, only. On a synchronic level, there is no “urban dialect” delineable by discrete features, each city or town having its own characteristics. It all depends on the history of settlement and migration; the Muslims of Baghdad, for instance, speak their “Bedouin” dialect because they originate from the countryside, which repopulated Baghdad in the Ottoman era (Blanc 1964; Holes 2008; [Retsö, “Arabic?”]).

13.5.2.4 Sectarian: Muslim–Christian–Jewish, Sunnī–Šīʕī, Muslim–Christian, Muslim–Jewish The linguistic situation in Baghdad is a relevant and often cited example of sectarian differentiation, or “communal dialects,” although the situation today no longer exists: the three different religious communities—Muslims, Jews, and Christians—used to speak different dialects. Again, this differentiation harks back to the history of settlement: Muslims originate from the Bedouin population in southern Iraq (from the 17th century CE), whereas Christians have their origins in different cities in northern Iraq (Blanc 1964; Abu-Haidar 1991; Mansour 1991). Due to migration to Israel, the number of communal dialects has been reduced to two. Another well-documented situation (Holes 1983, 1987) exists in Baħrayn where the Šīʕī community speaks a dialect different from their Sunnī compatriots. The Šīʕī dialect is related to sedentary or seminomadic dialects spoken “in an area which extends around the periphery of Arabia proper from Yemen to Oman to lower Mesopotamia” (Holes 1983: 8), whereas the Sunnī dialect corresponds the nomadic type common in the Najd. Again this situation goes back to population movement as the Bedouin tribes arrived there from the Najd in 1782– 1783 only (Holes 2001: xxvvii). For the city of Aleppo, see Behnstedt (1989), where Christians use other dialects than the Muslims (cf. Behnstedt and Woidich 2005). Behnstedt (1998–1999) reports on the rather complicated situation on the island of Djerba/Tunisia with at least three different communal dialects, Muslim, Ibadi, and Jewish, and even Berber in some villages. In general, Jewish communities used and still use Arabic dialects deviating from the Muslim varieties. This is in particular so in the Maghrib; see Stillman (1988) and Heath (2002) for Morocco; Cohen (1912) for Algiers; Cohen (1975) for Tunis; and Yoda (2005) for Tripoli/Libya. For Cairo, see Blanc (1981) and Rosenbaum (2003); for Iraq, see Jastrow (1990a, 1990b) and Mansour (1991).

13.6 Future Prospects Despite a great deal of progress that has been made and a constant flow of new data over the last 50 years, the coverage of the Arabic-speaking area is still thin compared with what we are accustomed to in the dialectological studies of European countries. Talking about “a relative surfeit of information … on modern dialects” (Owens 2005: 274) is far from reality. Egypt, for instance, a country that has been, compared with others, researched quite well

Dialectology over the last decades, still needs closer study in such key areas as Upper Egypt and the Western oases. On the core of the Arab world, the Arabian Peninsula, in particular its northwestern area Ħiǧāz, its southern parts ʕAsīr, Ħaḍramawt, and ʕUmān, we have only very limited knowledge. And the same is true for the Sudan and the subSaharan areas and for Libya, parts of Algeria, let alone the “Sprachinseln,” such as those in Uzbekistan, Afghanistan, and Iran. All these areas still need much fieldwork. Another important point here concerns the validity of data we find in publications from earlier times, often the only ones we have, before dialectology became a more professionalized discipline. A considerable number of these we owe to the efforts of single individuals only, sometimes not even a trained linguist or Arabisant. We cannot be sure that these are free from errors. Redoing fieldwork in places from where we have some older information is not a superfluous task for the future. Systematic research is best done in the framework of projects aiming at drawing linguistic atlases. These should not be limited to Arabic-speaking regions but should also include other languages spoken in the areas under study. To give an example, in Morocco it is highly advisable to extend the research to Berber-speaking areas because of the mutual interferences of the two languages [Kossmann, “Borrowing”]. There are some activities concerning atlases (Israel/Palestine, Tunisia, Morocco), but many more areas have until this day remained unresearched or underresearched (Algeria, Tunisia, Libya, Jordan, Iraq, Saudi Arabia, Sudan). The reasons for this may be seen in facts such as geographical conditions, security problems, and the negative attitude and lack of understanding toward dialect studies prevailing in most Arab countries, which prevents many native Arab scholars from pursuing academic activities in this field. In this respect, much more explanatory work could be done to involve these academics in serious fieldwork in their home countries. Otherwise, it is to be feared that atlases for such countries as Algeria, Libya, Sudan, Saudi Arabia, and Jordan, comparable to Behnstedt and Woidich (1985) and Behnstedt (1999), will never see the light of day. Moreover, atlases should not be confined to regions or states but should cover the Arabic-speaking world as a whole and collect all the data gathered so far to provide us with a means to detect and visualize the migratory processes so important for the understanding of “the history of the linguistic contacts of speakers of varieties of Arabic” (Versteegh 2011b: 549). The recent “Wortatlas der arabischen Dialekte” (Behnstedt and Woidich 2011, 2012) can be seen only as first step in this direction. Regional atlases should, besides the linguistic data, collect and document as much information as possible about the histories of movement and settlement of the population. To a much larger extent than in the European situation, linguistic features were transported by the speakers themselves through physical migration, to mention only the second Arabization of North Africa by the migration of the Banī Hilāl, in addition to diffusion through contact.21 The distributions of many features, therefore, show highly irregular geographical patterns. A prerequisite for future fieldwork is a generally acknowledged but also regularly updated questionnaire to avoid the idiosyncrasies and predilections of the individual fieldworker to make the incoming data better comparable and to ensure an evenly distributed set of data without surprising and unwanted lacunae.22 This questionnaire should not only reflect the traditionally recognized variables of phonology, morphology, syntax, and lexicon but also—at least to some extent—take account of issues relevant in the current linguistic discourse. As a methodological step forward, newer quantitative approaches such as those developed in dialectometry should be applied to the data given in these atlases to arrive at more objective classifications and should be less dependent on the personal views of the individual researchers. As to comprehensive descriptive grammars, there are major cities in the Arab world—Beirut, Mosul, Omdurman, Aden, Masqat, Oran, and Constantine may serve as examples—of which the dialects have not been sufficiently described so far, let alone those of rural areas and of many Bedouin tribes that are still partially or entirely undocumented. What we have at our disposal is very sketchy in most cases, as only the basic facts of phonology and morphology are described. Although syntax is still being treated like an orphan in many a grammar, interesting projects are now being carried out; see, for instance, Isaksson (2009) on circumstantial qualifiers. The same can be said about lexical documentation. There are lots of word lists and glossaries, some of them quite voluminous, which are limited to listing a number of lexemes, in some manner deviating in form or content from mainstream Arabic and that more or less by chance appear in a text or were recorded by the researcher. Here much systematic work is waiting for engaged lexicographers who should try to cover the entire lexicon of a dialect, combining this with in-depth semantic analysis that includes phraseology and pragmatics.23 For regions such as Tunisia, Libya, Oman, and the Arabian Peninsula in general, no substantial dictionaries are available. In other

Dialectology cases, such as the Levant, Iraq, and Algeria, we have reasonable dictionaries, but these are sometimes outdated and need to be replaced or updated. An important point here is the desperate lack of bidirectional dictionaries, which can make the search for an appropriate expression for a given concept a tedious task. With the possibilities of modern databases and the computer, developing glossaries and dictionaries in both directions does not seem an unreasonable request [Buckwalter and Parkinson, “Modern Lexicography”]. These modern facilities make us think of a step further: as there is already a large amount of vocabulary available in the literature, all this could be collected in one large database. Such a resource not only would be very helpful for dialectology itself but also could serve a general philological interest as already stated by Spitaler (1961: 135/226): a “Sammelwörterbuch würde…der weiteren Erforschung des Mittelarabischen die größten Dienste leisten.” “a collective dictionary would be a great service to the the study of Middle Arabic”. As was pointed out already, documentation for Arabic dialects abounds in publications of text collections, which could be used far more efficiently if presented in digitalized form. This means not only that they should be accessible by means of a computer but also that they should be prepared in the necessary way according to a standard tagging protocol to make them fully operable for statistical approaches as developed in corpus linguistics. Such a corpus containing all relevant texts from all regions covered so far would be an invaluable resource for detailed syntactic and phraseological studies, two underdeveloped fields in Arabic dialectology, and particularly helpful for comparison and classification.24 There are individual dialectologists who apply these research methods already (Isaksson 2004; Persson 2008, 2009). More generally speaking, syntactic and semantic research should take the typological and functional frameworks as developed in general linguistics more into account to take part in the contemporary linguistic discussion. The Arabic language in general and Arabic dialects in particular have much to contribute here, but they have remained rather underrepresented so far in the scientific discourse. These last points, if realized, would contradict Jean Cantineaus’s dictum “la dialectologie arabe n’est guère progressiste; les nouvelles techniques de recherche ont du mal à s’y acclimater.” “Arabic dialectology is not at all progressive; new research methods have difficulty becoming established in it” (1960: 277). No doubt, this would be a tremendous task that can be fulfilled only within the framework of full-fledged international cooperation, together with the aṣħāb al-luġa themselves.

References25 Abu-Haidar, Farida. 1991. Christian Arabic of Baghdad. Wiesbaden: Harrassowitz. Arnold, Werner and Peter Behnstedt. 1993. Arabisch–Aramäische Sprachbeziehungen im Qalamūn (Syrien). Eine dialektgeographische Untersuchung. Wiesbaden: Harrassowitz. Behnstedt, Peter. 1985. Die nordjemenitischen Dialekte. Teil 1: Atlas. Wiesbaden: Harrassowitz. ——. 1997–2000. Sprachatlas von Syrien. I. Beiheft, Kartenteil. II. Volkskundliche Texte. Wiesbaden: Harrassowitz. Behnstedt, Peter and Manfred Woidich. 1985–1999. Die ägyptisch-arabischen Dialekte. 5 vols. Dialektatlas von Ägypten. Beihefte zum Tübinger Atlas des Vorderen Orients Reihe B (Geisteswissenschaften) 50/1–5. Wiesbaden: Reichert. ——. 2005. Einführung in die arabische Dialektgeographie. Leiden: Brill. ——. 2011. Wortatlas der arabischen Dialekte. Band I: Mensch, Natur, Fauna, Flora. Leiden: Brill. ——. 2012. Wortatlas der arabischen Dialekte. Band II: Materielle Kultur. Leiden: Brill. Bergsträßer, Gotthelf. 1915. Sprachatlas von Syrien und Palästina. Zeitschrift des Deutschen Palästina-Vereins 38: 169–222. Blanc, Haim. 1964. Communal dialects in Baghdad. Cambridge, MA: Center for Middle Eastern Studies. Brustad, Kirsten E. 2000. The syntax of spoken Arabic: A comprehensive study of Moroccan, Egyptian, Syrian, and

Dialectology Kuwaiti dialects. Washington, DC: Georgetown University Press. Cantineau, Jean. [1955] 1960. La dialectologie arabe. In Études de linguistique arabe. Mémorial Jean Cantineau, 25–278. Paris: Klincksieck. Caubet, Dominique. 1993. L’Arabe marocain. Tome I Phonologie et Morphosyntaxe. Tome II Syntaxe et Catégories Grammaticales, Textes. Paris: Éditions Peeters. Clarity, Beverley E., Karl Stowasser, and Ronald G. Wolfe. 1964. A dictionary of Iraqi Arabic. English–Arabic. Washington, DC: Georgetown University Press. Cohen, David. 1963. Le dialecte arabe ḥassānīya de Mauritanie (parler de la G∂bla). Paris: Klincksieck. Corriente, Federico C. 1977. A grammatical sketch of the Spanish Arabic dialect bundle. Madrid: Instituto HispanoÁrabe de Cultura. Cowell, Mark W. 1964. A reference grammar of Syrian Arabic: Based on the dialect of Damascus. Washington, DC: University of Georgetown. de Jong, Rudolf E. 2000. A grammar of the Bedouin dialects of the Northern Sinai littoral: Bridging the linguistic gap between the Eastern & Western Arab world. Leiden: Brill. Eisele, John C. 1999. Arabic verbs in time: Tense and aspect in Cairene Arabic. Wiesbaden: Harrassowitz. Fischer, Wolfdietrich. 1961. Die Sprache der arabischen Sprachinsel in Uzbekistan. Der Islam 36: 232–263. Fleisch, Henri. 1974. Études d’arabe dialectal. Beirut: Dar El-Machreq. Harrell, Richard S. 1962. A short reference grammar of Moroccan Arabic. Washington, DC: Georgetown University Press. Heath, Jeffrey. 2002. Jewish and Muslim dialects of Moroccan Arabic. London: RoutledgeCurzon. Hinds, Martin and Elsaid Badawi. 1986. A dictionary of Egyptian Arabic: Arabic–English. Beirut: Librairie du Liban. Holes, Clive. 1987. Language variation and change in a modernising Arab state. London: Kegan Paul International. ——. 2005. Dialect, culture, and society in Eastern Arabia. Vol. 2, Ethnographic texts. Leiden: Brill. Ingham, Bruce. 1982. North east Arabian dialects. London: Kegan Paul International. ——. 1994. Najdi Arabic: Central Arabian. Amsterdam: Benjamins. Jastrow, Otto. 1978–1981. Die mesopotamisch-arabischen q∂ltu-Dialekte. 2 vols. Wiesbaden: Harrassowitz. Jastrow, Otto and Wolfdietrich Fischer (eds.). 1980. Handbuch der arabischen Dialekte. Wiesbaden: Harrassowitz. Johnstone, Thomas Muir. 1967. Eastern Arabian dialect studies. London: Oxford University Press. Jullien de Pommerol, Patrice. 1992. Grammaire pratique de l’arabe Chadien. Paris: Karthala. Kaye, Alan S. and Judith Rosenhouse. 1997. Arabic dialects and Maltese. In The Semitic languages, ed. Robert Hetzron, 263–311. New York: Routledge. Landberg, Carlo de. 1905–1913. Études sur les dialectes de l’Arabie Méridionale. Glossaire daṯînois, vol. 2. Daṯînah, vol. 1–3. Leiden: Brill. Nöldeke, Theodor. [1904] 1982. Das Klassische Arabisch und die arabischen Dialecte. In Beiträge zur semitischen Sprachwissenschaft, 1–14. Amsterdam: APA-Philo Press. Owens, Jonathan. 1984. A short reference grammar of Eastern Libyan Arabic. Wiesbaden: Harrassowitz. Owens, Jonathan and Alaa Elgibali (eds.). 2010. Information structure in spoken Arabic. London: Routledge.

Dialectology Palva, Heikki. 1976. Studies in the Arabic dialect of the Semi-Nomadic ∂l-ʕAǧārma Tribe (al-Balqāʔ District, Jordan). Göteborg: Acta Universitatis Gothoburgensis. Prochazka, Theodore, Jr. 1988. Saudi Arabian dialects. London: Kegan Paul International. Procházka, Stephan. 2002. Die arabischen Dialekte der Çukurova (Südtürkei). Wiesbaden: Harrassowitz. Rosenbaum, Gabriel. 2004. Egyptian Arabic as a written language. Jerusalem Studies in Arabic and Islam 29: 281– 340. Singer, Hans-Rudolf. 1984. Grammatik des Arabischen von Tunis. Berlin: de Gruyter. Shawarbah, Musa. 2012. A grammar of Negev Arabic: Comparative studies, texts and glossary in the Bedouin dialect of the ‘Azåzmih tribe. Wiesbaden: Harrassowitz. Taine-Cheikh, Cathérine. 1988–1998. Dictionnaire Hassaniyya-Français. 8 vols. Paris: Geuthner. Vanhove, Martine. 1994. La langue Maltaise. Wiesbaden: Harrassowitz. Versteegh, Kees, et al. (eds.). 2006–2009. EALL. 4 vols. Leiden: Brill. Watson, Janet C. E. 2002. The phonology and morphology of Arabic. Oxford: Oxford University Press. Wellens, Ineke. 2005. The Nubi language of Uganda. Leiden: Brill. Werbeck, Wolfgang. 2001. Laut- und Formenlehre des nordjemenitisch-arabischen Dialekts von Manāḫa. Münster: RHEMA. Woidich, Manfred. 2006. Das Kairenisch-Arabische. Eine Grammatik. Wiesbaden: Harrassowitz.

Notes: (1) Ethnologue: http://www.ethnologue.com/show_language.asp?code=arb. Egypt alone counts more than 84 million inhabitants now, nearly all of them speakers of a variety of Arabic. (2 ) If we consider any variety of a language a dialect, Modern Written Arabic and Classical Arabic are dialects of Arabic as well and should be treated as such. Nevertheless, Arabic dialectology is concerned only with dialects of Arabic that have native speakers. (3 ) See http://www.aida.or.at. (4 ) See http://www.fb10.uni-bremen.de/ghilm/about.aspx. (5) It goes without saying that in this short article not all important publications can be quoted. (6 ) For more details, see the chapter “Die arabischen Sprachatlanten” in Behnstedt-Woidich (2005: 4–7). (7 ) See further a number of articles in the procedings of AIDA 6 (Mejri 2006). (8 ) For a sharp refutation see Sima (2006: 98) and Owens (2006: 9). (9 ) For further titles, we refer here to the relevant articles in EALL. (10 ) Cf. the introduction to Brustad (2000: 2–4). (11) See http://halshs.archives-ouvertes.fr/halshs-00504180/fr/. (12 ) [http://www.semarch.uni-hd.de/index.php43?lang=en] (13 ) This is common for dialects of European languages, like in German, where the variables maken–machen, ik– ich, dat–dass, appel–Apfel, that is, the stops developing to affricates, separate Northern and Southern varieties

Dialectology rather neatly. Within the German context these variables correspond to continuous isoglosses on the map and form the famous “Rhenish fan” (Niebaum and Macha 2006: 107, Map 30). (14 ) That is, insertion of a vowel preceding /r/ in a cluster -vCr(v). (15) According to the initial consonant of the morphemes of the first and second perfect, such as katab-t ∼ katab-k “I wrote.” (16 ) Bailey (1980) uses implicational scales for a classification of Ancient Greek dialects. This has not been done for Arabic dialects yet. One possible implication would be, for instance, *q 〉 g ⊃ *g 〉 ǧ/ž, that is, dialects with a voiced reflex of *q will have an affricated/sibilant reflex of *g. To the best of our knowledge, no dialect has yet been found that falsifies this implication. The Alexandria example found in the texts published in Behnstedt (1980) seems more a case of dialect mixing and a fact of “parole,” not of “langue.” The reverse of this implication *g 〉 ǧ/ž ⊃ *q 〉 g can easily be falsified, since there are several dialects with /ǧ/ž/ and /q/, for instance, the oasis Farafra in Egypt. Taine-Cheikh (1998: 15) points out another implication: dialects with /ʔ/ (glottal stop) for *q will have replaced the interdentals with dentals, that is, *q 〉 ʔ/ ⊃ ṯ 〉 t in, for instance, /* ṯalāṯa/ “three.” There has been no systematic research done so far on implications of this type. Holes (1987) uses, in fact, implicational scales for sociolinguistic variables in his study on Baħrayn. (17 ) Whether dialectrometry can ever be applied to the entire of Arab world is doubtful in view of the density of data and the proximity of the research points it needs. (18 ) This division does not mean, as Owens (2003) apparently assumes, that dialectologists suggest that this paradigmatic leveling developed in North Africa. What is said is that this feature was reimported to Egypt by tribes migrating back to the East. It developed much earlier, maybe not even in Egypt, but on the Arabian Peninsula in “pre-diaspora” times; see next footnote. (19 ) A paradigm of the present tense with the same synchronic structure, that is, one single morpheme for the first person (I, we) and one for the plural, can be found farther to the East in the contemporary northwest Aramaic language of Maʕlūla in Syria (Arnold 1990: 74). There it is the natural outcome of the development of a new paradigm from participles, not a case of paradigmatic leveling as in Arabic. As to the older variety of Galilean Aramaic (Lipiński 2001: 382), only the 1st sg. receives a n-prefix, and Dalman (1905: 213) considers this as a “Plural der Selbstermunterung.” Similar paradigms, though due to different provenance, are thus attested for other Semitic languages. For the discussion of this historical development in Arabic, see Owens (2003), who places its origin in Egypt, and Corriente (2011), who argues for its origin in Yemen. (20 ) H. Blanc’s (1964: 28, emphasis added) statement (based on Cantineau 1939): “The present-day distribution of reflexes of OA /q/ throughout the Arabic-speaking world presents a striking dichotomy: most sedentary populations have a voiceless reflex and all non-sedentary populations a voiced reflex” can today be considered true only for its second part. There are numerous regions in the present-day Arab world, for example, the whole of Upper Egypt, large parts of the Nile Delta, Sudan with a sedentary population speaking voiced /g/, and not a voiceless reflex of *q, due to the settlement of and mixing with Bedouin over the course of history. (21) For some conspicuous examples, see Behnstedt and Woidich (2005, Chapter 5.2). (22 ) An example for Maghrebi dialects is Caubet (2002). (23 ) It is strange that old dictionaries, grammars, and textbooks, some of them dating back to the 19th century, are recycled by publishing houses instead of producing up-to date publications based on fresh and recent research. Only recently (June 2011), Vollers and Burkitt (1895), which is based on the German version by Vollers (1890), was reedited and offered as a textbook for Egyptian Arabic. No doubt, this was an excellent short book in its time, but we are 120 years later; both life and research have progressed. (24 ) A first step in this direction will be taken within the framework of a project titled “Idiomaticity, Lexical Realignment, and Semantic Change in Spoken Arabic,” which Jonathan Owens and Manfred Woidich started recently at the University of Bayreuth. (25) References refer to boldfaced entries in the chapter.

Dialectology

Peter Behnstedt Peter Behnstedt, University of Amsterdam

Manfred Woidich Manfred Woidich, University of Amsterdam

Codeswitching and Related Issues Involving Arabic Abdelali Bentahila, Eirlys Davies, and Jonathan Owens The Oxford Handbook of Arabic Linguistics Edited by Jonathan Owens Print Publication Date: Sep 2013 Online Publication Date: Dec 2013

Subject: Linguistics, Language Contact, Languages by Region DOI: 10.1093/oxfordhb/9780199764136.013.0014

Abstract and Keywords Bilingual speech involving Arabic has been an important source of linguistic research on the language. The greater part of this research has involved Arabic in contact with other languages; in recent years, greater systematic attention has been given to Arabic diglossic speech as well. This article looks at Arabic in contact with other languages and also deals with diglossic speech. It also briefly summarizes the use of secret languages, which has close structural parallels to code switching. Keywords: Arabic, bilingual speech, linguistics, language, diglossic speech

14.1 Introduction BILINGUAL speech involving Arabic has been an important source of linguistic research on the language. The greater

part of this research has involved Arabic in contact with other languages; in recent years greater systematic attention has been given to Arabic diglossic speech as well. This article examines Arabic in contact with other languages and with diglossic speech.1 In addition, we briefly summarize the use of secret languages, which has close structural parallels to codeswitching.

14.2 Issues in Codeswitching Research The systematic study of codeswitching, defined roughly as the use of more than one language within a single piece of discourse or interaction, began to attract the serious attention of linguists in the 1970s and early 1980s. Since then, the lion’s share of research has been concerned with identifying permissible switch points. Early analyses offered highly specific syntactic constraints but were followed by attempts to state purportedly universal constraints, first ones formulated in terms of surface structure (Poplack 1980), later ones evoking underlying syntactic relations (Di Sciullio, Muysken, and Singh 1986, Sankoff 1998), and more recently attempts to account for all possibilities in terms of principles already motivated for the grammars of the two languages, such as MacSwan’s (1999) minimalist approach [Benmamoun and Choueri, “Syntax”]. Once it was noted that there were often clear asymmetries between the roles of the two languages in codeswitching discourse, many models identified one of them as the “base,” “host,” or “matrix” language into which elements of the other language were “embedded” (Joshi 1985; Myers-Scotton 1993a, 2002). At the same time many researchers felt the need to draw distinctions between codeswitching and other phenomena, both the old-established category of borrowing (Heath 1989) and new ones used in the elaboration of their models. Codeswitching has also been approached from other angles. Its discourse and communicative functions were explored by Gumperz (1982) and Bentahila (1983a) and later became the subject of more systematic models like that of Myers-Scotton (1993b). Many studies noted the impact on switching patterns of sociolinguistic variables

Codeswitching and Related Issues Involving Arabic such as language dominance, proficiency, and prestige. A conversational framework is used in Auer (1998), and the use of codeswitching as a literary device has also been examined (Keller 1979; Tessier 1996). Looking back over these developments, we can draw a distinction between two general approaches to codeswitching. On one hand, we have the strictly formalistic models, focusing largely on structure, formulating absolute generalizations, and claiming to identify universal principles; on the other, there are more holistic, interdisciplinary approaches that take a wider view, acknowledging the relevance of many other variables and drawing on insights from fields such as sociolinguistics, psycholinguistics, pragmatics, discourse, and conversation analysis.

14.3 Studies of Codeswitching Involving Arabic The absence of any comprehensive survey of published studies on codeswitching involving Arabic may relate to a number of facts. First, the considerable syntactic, phonological, and lexical differences between the colloquial Arabic varieties used across the Arabic-speaking world [Behnstedt and Woidich, “Dialectology”] may make it difficult for scholars to interpret and compare and data from different communities. Second, the published works report on a diverse range of configurations, covering switching between Arabic and a number of other languages, including French, English, Dutch, Spanish, Hausa, with some studies looking at tri- or even quadrilingual switching (Owens 2005a, 2005b; Edwards and Dewaele 2007), as well as switching between colloquial and standard Arabic (see Section 14.4). Third, the bodies of work published in different languages have tended to remain separate; for instance, discussions of Arabic–French switching published in French frequently ignore similar studies written in English and vice versa. Yet the very diversity of existing studies of codeswitching involving Arabic can be seen as a valuable opportunity for those seeking new and deeper insights into the phenomenon. The availability of detailed studies on different language pairs offers the possibility of comparative studies to explore the impact of the nature of a language on the patterns observed. Moreover, the data available also illustrate a variety of sociolinguistic contexts, from the exFrench colonies of North Africa, where Arabic is the official and majority language, to large immigrant communities in Europe, where it is a minority language but an important community in-group marker, and other cases of more isolated immigrants with no such community support. There is also the possibility of exploring the functions of codeswitching between colloquial and standard Arabic, varieties generally described as being in a diglossic relationship. And the many Arabic-speaking communities that are undergoing sociolinguistic transformations, such as changes in immigration patterns, education systems, ideology, or attitudes, offer the chance to investigate the consequences of such variation on the codeswitching behavior of a community. The wealth of material available for investigation is thus a complicating factor for research on codeswitching involving Arabic but also a stimulating one. Studies where Arabic is one of the languages in a codeswitching configuration are of course of intrinsic interest for what they may reveal about these specific instances of language contact. However, in what follows we will try in particular to assess what these studies have contributed to the theory of codeswitching in general.

14.4 Contributions of Codeswitching Studies Involving Arabic In the extensive debate on syntactic constraints on codeswitching, studies of data involving Arabic have played a number of important roles. First, there are the proposals directly inspired by the analysis of corpora involving Arabic. The earliest studies, in keeping with contemporaneous studies of other language pairs, merely stated ad hoc constraints identifying specific syntactic configurations where switches were or were not observed (Stevens 1974; Abbassi 1977). Bentahila and Davies (1983) attempt to account for nonoccurring switches in terms of subcategorization restrictions. Belazi, Rubin, and Toribio (1994) propose a Functional Head constraint prohibiting switching between a functional head (e.g., a numeral, quantifier, negative particle or modal) and its complement NP or VP, while allowing switching between a lexical head and its complement, such as between a verb or preposition and its object. All these studies were based on data from Arabic–French codeswitching. Boumans (1998), on the other hand, studies Moroccan Arabic–Dutch codeswitching and postulate the monolingual structure approach, a modification of Myers-Scotton’s MLF model, to describe it in terms of insertions of elements from an embedded language into structures provided by a matrix language. Aabi (1999) claims that the possibilities for Moroccan

Codeswitching and Related Issues Involving Arabic Arabic–French and Moroccan Arabic–Classical Arabic switching can be accounted for in terms of his functional parameter constraint, which evokes selectional restrictions to predict possible switch configurations. More recently, Owens (2005a, 2005b) adopts a processing-based account to explain certain patterns in Nigerian Arabic–English and Hausa–English switching, and Ziamari (2009) also looks at the implications of processing factors for Arabic– French switching. However, it must be admitted that none of the models based initially on the examination of codeswitching involving Arabic has been widely tested on other language pairs. On the other hand, evidence from data including Arabic has frequently been evoked in attempts to evaluate models originally based on other data, with significant results, as will be shown. Poplack (1980) is among the first to postulate purportedly universal constraints, with her equivalence constraint, prohibiting switching at points where the two languages do not exhibit similar syntactic structure, and free morpheme constraint, prohibiting switching between root and bound morphemes. However, among the counterexamples to these generalizations offered by other researchers were the very common strings of the type Arabic determiner + French determiner + noun described by Bentahila and Davies (1983) and also attested by many other studies on North African codeswitching. Strings like (1) and (2), which are typical of the discourse of Moroccan bilinguals, respect Moroccan Arabic syntax, where the demonstratives dak, ħad and the indefinite marker waħed require a following definite article but do not conform to French rules. (In the following examples from Bentahila and Davies 1983: 317, French is in italics.) (1) dak

chemise

that

the.F

shirt

“that shirt”

(2) wa ħ ed

liquide

one

the.M

liquid

“a liquid”

This problem led some of Poplack’s colleagues to look at Arabic–French data for themselves, and Nait M’Barek and Sankoff (1988) sought to preserve the equivalence constraint by arguing that these examples should not be classified as code switches at all but as instances of a separate phenomenon termed constituent insertion. In a similar way, faced with the counterexamples to her free morpheme constraint observed in many language pairs, including Arabic–French examples like (3), where a French verb is accompanied by Arabic inflections, Poplack, Sankoff and Miller (1988) sought to deal with them by postulating another category distinct from codeswitching, that of nonce-borrowings: (3)

Codeswitching and Related Issues Involving Arabic

ma-bqa- š

y-fonctionner

not-remain.NEG

3-function

“It stopped working” (Bentahila and Davies 1983: 315)

However, these attempts to protect the supposed universal constraints by simply relabeling problematic examples as something other than codeswitching leave Poplack’s account open to accusations of circularity and of course reduce the generality of its earliest formulation, based on Spanish–English switching. Data involving Arabic have posed still greater problems for Myers-Scotton’s long battle to defend her matrix language frame model as a universal theory of codeswitching. The basis for the earliest versions of this model, which was heavily based on Levelt’s (1989) model of speech production, was the assumption that codeswitching involves the insertion, into a syntactic frame provided by a matrix language (ML), of elements from an embedded language (EL), and that in mixed ML and EL constituents syntactic structures and system morphemes must be provided by the ML (Myers-Scotton 1993a). Over the years, material from language pairs including Arabic has played a significant part in pushing Myers-Scotton to modify all the basic components of her theory: the identification of the matrix language, the definition of system morphemes, and the rules constraining embedded language elements. For instance, she originally identified the matrix language in very concrete terms, as the language contributing the largest number of morphemes to a piece of discourse (Myers-Scotton 1993a). Reactions to this included Bentahila’s (1995) review objecting to her own identification (Myers-Scotton 1993a: 89, 151) of Arabic as the matrix language in isolated Arabic–French fragments like (1) and (4), (taken from Bentahila and Davies 1983: 319) (4) naaḍ

les

privés

arise-3PL

the.PL

private

“The private practitioners rose up (in protest)”

This is depite the fact that she had no access to the conversations from which these fragments are taken and therefore no way of performing a morpheme count. Like Poplack, Myers-Scotton was pushed to take Arabic data into account to improve her model (Myers-Scotton, Jake, and Okasha 1996). Significantly, in a later version the ML is defined only at the level of a complement phrase (CP) and ultimately becomes not a language at all but merely “an abstract frame for the morphosyntax of the bilingual CP” (Myers-Scotton 2002: 66). Moreover, while the original version made the strong generalization that mixed language constituents must respect ML word order and use ML system morphemes, problematic examples including common Arabic–French configurations like those previously cited led Myers-Scotton to weaken these claims. First, she relaxed the rule to allow for examples like (5): (5) kan-t

dak la

semaine dyal ta-y-zuwl-u

les

permis

was-F

that the.F

week of IMPF-3-take-PL

the.PL

permits

“It was that-(the) week where they take away the driving licences” (Bentahila and Davies 1992: 449, cited by Myers-Scotton and Jake 1995: 1012)

Codeswitching and Related Issues Involving Arabic She did this by acknowledging the possibility of an EL island internal to a larger ML + EL constituent. Later she weakened her claim still further by admitting that such EL system morphemes as determiners, demonstratives, and possessive markers are permitted in mixed constituents, thus solving the problem posed by examples like (6): (6) Ça depend de quel degré de connaissance dyal la personne “That depends on the degree of knowledge of the person” (Bentahila and Davies 1998: 38; also cited by Myers-Scotton 2002: 91) But she was then forced to find another explanation for the scarcity of switches for determiners in her own data by introducing both a further constraint, the uniform structure principle, requiring that all constituents satisfy the wellformedness requirements of their language, and a notion of “specific congruence” that offers a loophole: EL system morphemes may occur in mixed constituents if the structures of the two languages are sufficiently similar, even if not identical. In a move reminiscent of Poplack’s argumentation, she also (2002: 8) drew a distinction between what she termed classic codeswitching, that used by speakers proficient in the ML, and composite codeswitching, the latter becoming a convenient dumping ground for problematic data. Nevertheless, even aft er all these modifications, there remain recurrent patterns in Arabic–French switching that no version of the MLF model seems able to account for. For instance, in the speech of highly proficient Moroccan bilinguals (who presumably use classic codeswitching), we commonly find isolated grammatical items from Arabic within otherwise entirely French strings, as illustrated in (7): (7) tu ne vas pas t’amuser chaque fois à former des pelotons à traduire chaque fois dak l matériel “You are not going to bother every time training squads to translate every time that (the) material” (Bentahila and Davies 1998: 38)2 (8) ils contredisent ħ ad la théorie de Darwin pour trois raisons “they contradict this- the theory of Darwin for three reasons” (Bentahila and Davies 2007: 458) If the MLF model’s constraints are weakened yet again to allow even an isolated EL demonstrative (8) or demonstrative + determiner string (7) to occur within an otherwise entirely ML clause/sequence of clauses, then its remaining predictive power seems negligible. Data on codeswitching involving Arabic can thus be seen to have served a useful corrective function, exposing overgeneralizations in purportedly universal models and showing the astonishing lengths to which some researchers will go in defending these models.3 The quest for syntactic universals has clearly led some to quite desperate measures, where data get reinterpreted, reclassified, or oversimplified merely to shore up theoretical claims. However, Owens’s (2005a) material illustrating the complex multilingual discourse of Maiduguri, Nigeria, where switching involves Nigerian Arabic, Standard Arabic, Hausa, and English, suggests that other approaches should also be pursued. Far from seeking to describe this material within some further modification of the MLF, Owens argues that we must recognize that certain types of codeswitching cannot be subsumed under this model; the concept of a matrix language is a very useful one, but seeking to make it a universal model of all codeswitching behavior will only reduce its effectiveness. Still more controversies have come from comparisons of switching patterns between different language pairs involving Arabic. Arabic–Dutch, and Arabic–French data have been compared by Nortier (1990, 1995) and Boumans and Caubet (2000). Interestingly, in contrast to the typical Arabic–French patterns illustrated in (1)–(2), in Arabic–Dutch switching there is instead a strong tendency to use bare Dutch nouns, as in (9), which contrasts with (2): (9)

Codeswitching and Related Issues Involving Arabic

waħed

Ø-bejaardencentrum

one

Ø—old people’s home

“an old people’s home” (Nortier 1990: 199)

Nortier (1995: 89) talks of a “suspension of syntax” in such configurations, Boumans and Caubet (2000: 131) suggest that the second determiner is omitted because it is redundant, and others have suggested various linguistic explanations for the contrast between French and Dutch (the more clitic-like nature of French determiners; Muysken 2000: 83; Versteegh 2001: 223), phonological similarities between French and Arabic determiners (Heath 1989: 35), and the lack of structural congruity between Arabic and Dutch (Myers-Scotton 2002: 126). However, there is a need for caution in seeking to attribute such differences directly to formal features of the languages involved. There are after all significant sociolinguistic contrasts between the North African and Dutch communities concerned (e.g., duration of language contact, stability of the bilingual community, status of the languages and their users). Owens (2005b) argues that factors such as language proficiency and the differing social statuses of the languages involved may explain whether bilinguals opt for the bare noun or the integrated noun patterns, and after exploring the way the two possibilities appear in switching among Hausa, Nigerian Arabic, and English he offers a processing-based explanation, suggesting that uninflected forms are more rapidly processed. In fact, the studies of different corpora involving Arabic may serve to remind us of the importance of looking beyond syntax when seeking to account for the types of switch that occur. For instance, the pragmatic distinction between old and new information is evoked by Owens (2005a), who observes that in Hausa–Nigerian Arabic switching Arabic (L1) tends to be used for the topic and Hausa (L2) for the comment. Ziamari (2009), on the other hand, observes that in her Arabic–French data French (L2) is often used even for topics and suggests that this pattern is characteristic of situations of stable, long-established bilingualism. Carrying this point further, it has been shown that social variables even within a single overall community may be reflected in distinctive switching patterns. Bentahila and Davies (1991, 1992, 1995, 1998) compared switching patterns across two generations of Moroccan bilinguals and noted that the distribution of switch types varied according to their language background and proficiency. Three different switching styles were identified: one used mostly by balanced bilinguals, where there is frequent alternation between clauses in one language and those in the other; a second favored by the younger group who are not as fluent in French, dominated by insertion of French lexis into an Arabic matrix; and a third described as the “leak” style and used by those extremely fluent in French, where discourse almost entirely in French is dotted with occasional fillers and grammatical items from Arabic. Since then similar distinctions have been drawn using different data sets: Jacobson (2000a) endorses the distinction between alternation and insertion with data from other language pairs, and Muysken (2000) adopts a similar distinction between alternation and insertion switching. The “leak” switching style is comparable to patterns described by Lipski (2005) in Spanish– English switching. Another point that has emerged clearly from studies of switching involving Arabic is that, in a stable bilingual community, codeswitching can become a highly norm-governed variety, following conventions specific to the community and fulfilling an important function as an identity marker. Whatever explanation is offered for some of the controversial switches noted already, it is clear that certain patterns have become emblematic markers of a particular group. The NP pattern illustrated in (1), (2), (5), and (8) is a good example. Theoreticians may struggle with what they consider the exceptional nature of such switches, but the fact remains that they are so deeply entrenched in the everyday speech patterns of Arabic–French bilinguals in Morocco that even 4-year-old children are already regularly using them (Bentahila and Davies 1994). On the other hand, members of another community using the same language pair may favor quite different patterns. Sefiani (2003) reports on the Arabic–French codeswitching of 12- to 25-year-olds from the third generation of North African immigrant families settled in Besan Ç on, France. Some of these subjects use only French at home, with their bilingual parents, because they are conscious of their limited proficiency in Arabic, but

Codeswitching and Related Issues Involving Arabic between themselves they use a codeswitching variety to mark their in-group status. Sefiani notes many examples where a speaker inserts, into a French structure where a past participle would be required, the 3PL masculine singular imperfect form of an Arabic verb, this being the form generally used to cite a verb, in the way an infinitive form is used in French or English: (10) elle a

ħlef

she has

swear.PST.3M4

“She has

sworn”

(11) il a

xṭob

la fille mais ça a pas marché

he has

ask-to-marry.PST.3M

the girl but is did not work

“He has asked the hand of the girl but it didn’t work.” (Sefiani 2003:57)

Melliani (2005) reports very similar patterns in the speech of young people of Moroccan origin in the Rouen area. Such examples seem outlandish and amusing to Moroccan bilinguals living in Morocco, who instead are very comfortable inserting French verb roots, complete with Arabic inflections, into Arabic strings, as in (3). Clearly, different configurations can become particularly idiomatic in-group markers in different communities using the same language combinations. A particularly interesting case is that described by Nortier and Dorleijn (2008), who report on what they call Moroccan-flavored Dutch (MLD), an in-group variety that has emerged only in the last few years and that is used in Dutch cities by youths of varied origins (Turkish, Moroccan, Surinamese, and even native Dutch). Apart from a Moroccan accent and certain nonstandard grammatical features, Nortier and Dorleijn report that this variety is also marked by the insertion of Arabic fillers, interjections, and certain grammatical morphemes such as dak, waħed, and the interrogative particle waš, which may remind us of the leak style of Arabic–French switching mentioned already. This adoption of features of Arabic codeswitching style by non-Arabic speakers can be seen as an instance of language crossing (Rampton 1995) and again highlights the symbolic value that codeswitching patterns may acquire: the originally Moroccan features have now become symbolic of a wider group of streetwise urban youths. The role of switching patterns as identity markers may be particularly obvious in public discourse and may be accessible to a wider audience, not merely to in-group members. Interesting examples of this situation may be seen in the use of codeswitching in the lyrics of North African popular songs, both the local genre of rai and that of rap as performed by North African hip-hop artists. This is examined in Bentahila and Davies (2002) and Davies and Bentahila (2006, 2008a, 2008b), where it is argued that incorporating both Arabic and French into the lyrics of a single song may serve to manipulate the audience targeted [Holes, “Orality”]. Codeswitching using the local style may be a strong identity marker and make the lyrics unintelligible to anyone outside the group, achieving an effect of localization and exclusion of the Other. On the other hand, including some French in an otherwise Arabic text may open the song up to a wider audience of outsiders, thereby helping to globalize the message. Williams (2009) makes similar observations about the codeswitching between Arabic and English used by Egyptian hip-hop groups. Jablonka (2009) argues that codeswitching in rai lyrics is used to create a social communicative style, and he also notes the symbolic use of English by some performers, this being another case of language crossing. In fact, the symbolic value of codeswitching involving Arabic in rap lyrics can be set beside similar strategies observed in many different countries and language combinations (Androutsopoulos and Scholz 2002; Sarkar and Winer 2006),

Codeswitching and Related Issues Involving Arabic and there would seem to be much more to investigate in this constantly evolving genre. Finally, a significant phenomenon can be noted, which in one of its two realizations can be regarded as the most extreme form of in-group codeswitching. Secret languages are closely linked socially and structurally to codeswitching. Socially, a secret language can be thought of as an even more restricted in-group variety than is a code-switched variety, used by a closed social group, such as a professional caste, student organization, or marginal occupational groups such as thieves. Structurally, rather than taking the content vocabulary from a language of the wider environment, as happens with codeswitching, speakers of secret language make up the forms to be inserted themselves. Arabic is rich in documented secret languages, as a recent comprehensive overview of the phenomenon (Wolfer 2011) shows. Following a well-defined dichotomy, Arabic secret languages can be divided into argots and ludlings. The former consist of a secret vocabulary, while the latter are purely phonologically manipulations (see Berjaoui 1996 for extensive documentation of the latter). How words are formed in an argot is multifarious and ingenious, but they always reflect in one way or another the immediate environment of their existence and hence are semantically semitransparent; that is, if one knows the local culture well enough, the basis of the secret meaning can often be deduced. For instance, in the “market language” (liγa suu ʔiiye) of Damascus a mihbir is a “rich customer,” a stem IV verb (CA, not Damascene), which is derived from Damascene habra “good red meat.” Hebrew was and still is a favorite source of secret vocabulary among Judeo-Arabic speakers (Wolfer 2011: 29, 34). In the secret language of Koranic school students in northeast Nigeria (termed waris), the local pronunciation of the Koranic word ka-ṣaaħibi,lit. “like the friend of ” means “fish,” the word taken from the Quran 68:48, ka-ṣaaħibi, l ħuut “like the companion of the fish.” In Nigerian Arabic huut is the word for “fish,” here converted to a secret designation via syntagmatic displacement to the adjacent ka-ṣaaħibi, ṣaaħibi itself being one NA word for “friend.” Obviously the association would be transparent only to those who have studied the Quran extensively (Owens and Hassan 2000: 229). Once created, the secret words function in an analogous way to single word insertions in codeswitching. For instance, to say that one has eaten fish today, one inserts the secret word in the appropriate matrix sentence that is provided by Nigerian Arabic: (12) akal-na

ka-ṣaahibi

aloom

ate-we

like-friend

today

“We ate fish today.”

One could equally have substituted English /fiš/ or Hausa /kifi/ here. Ludlings are purely phonological secret languages and hence have no direct analogy in CS. Extraneous material is added by a rule that has no effect at all on the meaning of the base words on which they operate. For instance, the ʕaṣfuuri secret language in Damascus adds a -CV after each syllable of the base word, -V being harmonious with the vowel of the preceding syllable. Thus, to say haati šaay “bring.F” tea one has the following, with the ludling syllables in bold (Wolfer 2011: 10). (13) ha- zaa -ti- zi

š a- zaa -y

Besides their cultural and anthropological value, Arabic ludlings are interesting as evidence in the debate about the minimal unit of morphological analysis in Arabic, consonantal root or voweled stem ([Ratcliffe, “Morphology”]; [Boudelaa, “Psycholinguistics”]). From a quantitative perspective, the vast majority of Arabic ludlings are formed

Codeswitching and Related Issues Involving Arabic by the addition of either syllables, as in (13), or fixed segments of arbitrary length (e.g., women’s talk in Algiers inserting - anga- within each word; Wolfer 2011: 17). Thirty of Wolfer’s documented ludlings fall into this category, as opposed to only three that operate only on the root consonants (e.g., transposing R 1 and R 3 with no change in the vowel). Certainly this evidence speaks strongly in favor of a key role for a voweled stem in word processing.5

14.5 Diglossic Switching, Diglossic Mixing A prominent aspect of Arabic is that it is in contact not only with other languages, the situation underlying codeswitching, but also, as it were, with itself. With marginal exceptions, for instance, Arabic in southern Turkey, Khorasan and in Nigeria, Cameroon and Chad, Standard Arabic is an ever-present variety in the everyday life of Arabs. The well-known linguistic term for this is diglossia ([Suleiman, “Folk Linguistics”]; [Al-Wer, “Sociolinguistics”]). While Arabic diglossia as a concept has an intellectual pedigree as old as that of studies on codeswitching itself, the relation between the two remains to be specified. In an earlier characterization (Owens and Bani Yasin 1987; Owens 2001: 451), diglossic Arabic was conceptualized in terms of borrowing. A different approach has been to apply models of CS to the variety of speech in which SA and CollA are mixed in some fashion. Eid (1982, 1988, 1992) is an earlier example of this approach. Recent studies of so-called diglossic switching have shown that the situation extends beyond borrowing (Bassiouney 2006; Mejdell 2006), though it is equally not clear that classic codeswitching models can be readily applied to it. In this section, a neutral term, mixed Arabic, will be used to describe this variety of Arabic, wiThelements from both SA and CollA.

14.5.1 Contextual Clues There are two factors to be treated here. One is the linguistic structure of the CS, which will be discussed in Section 14.5.2. A second, broadly speaking, is what Gumperz (1982; see also Auer 1998) terms contextual clues. These concern nonlinguistic aspects of the speech situation that impinge on code choice: the social relationships of the speakers, their social roles, conversation type, type of interactional exchange, audience design and occasion, and (problematically it must be said) topic. In a typical multilingual codeswitching situation, CS is associated with the following set of contextual features: (14) Contextual clues: multilingual CS • Social relationship: among equals • Social role: friends, unofficial6 • Conversation type: dialogue • Interactional exchange: two-way· Topic: open • Audience design: flexible according to situation but also often irrelevant (Myers-Scotton, CS as the unmarked choice) • Occasion: spontaneous, unplanned To these background factors we would add that the languages involved in the switching are all part of the normal oral communicative repertoire of the community that engages in the codeswitching. There will be many occasions for speakers to use the languages in a monolingual mode. Turning to diglossic switching or mixing, one may ask whether the contextual factors governing it differ from those when different languages are involved. There are clear examples indicating the answer is “no.” A good illustration of this is Alfonzetti’s (1998) study of Sicilian–Italian switching, in which the contextual factors are essentially the same as those set out in (14). As far as Arabic goes, recent corpora studies (Bassiouney 2006, 2010a; Mejdell 2006) suggest that the background factors are often special. These show Arabic mixing unfolding either in largely monologic contexts or in what can be termed moderated dialogues, media discussions involving two or more individuals. These are two formats where there can be a high degree of interaction between SA and colloquial.

Codeswitching and Related Issues Involving Arabic Bassiouney (2006) and Mejdell (2006) exemplify the monologic style in detail. In one, for instance, a political speech by then president Housni Mubarak (Bassiouney 2006: 245–246) is divided into three parts. Part 1 of the speech is characterized as EgA, part 2 as SA, and part 3 as mixed SA/EgA. The political speech provides clear exemplification of a style of mixing SA/colloquial, whereby one and the same speaker, with input from no one else, is in total control of choosing the different style at his disposal to convey his message. Linguistic studies of the ever-growing domain of moderated dialogues are just emerging (Bassiouney 2010b), even if this genre is perhaps the premier forum where spontaneous SA/colloquial contact can be observed. (15) Contextual clues: Arabic mixing • So

Jonathan Owens - Handbook of Arabic Linguistics.pdf - PDFCOFFEE.COM (2025)